THE JOURNAL OF ANIMAL BEHAVIOR VOLUME 4, 1914 EDITORIAL BOARD Madison Bentley University of Illinois Harvey A. Carr The University of Chicago Editor of Reviews Gilbert V. Hamilton Montecito, California Samuel J. Holmes The University of California Walter S. Hunter The University of Texas Herbert S. .Jennings The Johns Hopkins University Edward L. Thorndike Teachers College, Columbia University Margaret F. Washburn Vassar College John B. Watson The Johns Hopkins University William M. Wheeler Harvard University Robert M. Yerkes, Harvard University Managing Editor Published Bi-monthly at Cambridge, Boston, Mass. HENRY HOLT AND COMPANY 34. West 33d Street, New York U. K. STKCHEltT & CO.. London. Paris and Leipzig Foreign Agents Entered as second-class matter March 7, 191 1 at the post-office at Cambridge, Boston Massachusetts, under the act of March 3, 1879. d CONTENTS OF VOLUME 4, 1914 Number i, January-February P A.GKS Victor E. Shelford and W. C. Allee. Rapid modification of the behavior of fishes by contact with modified water. 1-30 Victor E. Shelford. Modification of the behavior of land animals by contact with air of high evaporating power. 31-49 Robert M. Yerkes and Chester E. Kellogg. A graphic method of recording maze-reactions 5°-53 John B. Watson. A circular maze with camera lucida attachment 56-59 Helen B. Hubbert. Time versus distance in learning. . . . 60-60 W. T. Shepherd. ( hi sound discrimination by cats 70-75 H. M. Johnson. A note on the supposed olfactory hunt- ing-responses of the dog 76-78 Number 2, March-April Eupha Foley Tugman. Light discrimination in the Eng- lish sparrow 79-109 Harry Beal Torrey and Grace P. Hays. The role of ran- dom movements in the orientation of Porcellio scabcr to light 1 10-120 Wallace Craig. Male doves reared in isolation 121-133 H. M. Johnson. Hunter on the question of form-percep- tion in animals 134-135 Harold C. Bingham. A definition of form 1 36-141 Sergius Morgulis. The auditory reactions of the clog studied by the Pawlow Method 142-145 Number 3, May-June John H. Lovell. Conspicuous flowers rarely visited by insects .• 147-175 Robert M. Yerkes. The Harvard Laboratory of Animal Psychology and the Franklin Field Station 176-184 Charles A. Coburn. The behavior of the crow, Corvus Amcricamts, Aud 185-201 W. C. Allee and Shiro Tashiro. Some relations between rheotaxis and the rate of carbon dioxide production of isopods 202-214 W. S. Hunter. The auditory sensitivity of the white rat. 215-222 .T//2. CONTENTS iii Henry H. P. Severin and Harry C. Severin. Behavior of pages the Mediterranean fruit fly (Ceratitis Capitata JVied.) towards kerosene 223-227 Number 4, July- August Lee Raymond Dice. The factors determining the vertical movements of Daphnia 229-265 Raymond Pearl. Studies on the physiology of reproduc- tion in the domestic 'fowl. VII. Data regarding the brooding instinct in its relation to egg production 266-288 Osv. Polimanti. On the thele-perception of sex in -silk- worm moths 289-292 K. S. Lashley. A note on the persistence of an instinct. 293-294 Number 5, September-October G. Y- Hamilton. A study of sexual tendencies in monkeys and baboons 295-318 H. M. Johnson. Visual pattern-discrimination in the vertebrates — I. Problems and Methods 3 I 9~339 H. M. Johnson. Visual pattern-discrimination in 'the vertebrates — II. Comparative visual acuity in the dog, the monkey, and the chick 340-361 Sergius Morgulis. Pawlow's theory of the function of the central nervous system and a digest of some of the more recent contributions to this subject from Pawlow's laboratory 362-379 R. P. Cowles. The influence of white and black walls on the direction of locomotion of the starfish 380-382 Number 6, November-December S. J. Holmes. Literature for 1913 on the behavior of the lower invertebrates 383-393 C. H. Turner. Literature for 19 13 on the behavior of spiders and insects other than ants 394-413 Stella B. Vincent. Literature for 1913 on the behavior of vertebrates 414-438 Walter S. Hunter. Pycraft on the infancy and courtship of animals 439-441 Walter S. Hunter. H. Volkelt's " Uber Die Vorstel- lungen der Tiere " 442-445 Subject and Author Index VOLUME 4 Original contributions are marked by an asterisk ( * ) * X llee, W. C. Behavior of fishes, 1 ; £\ *rheotaxis in isopods, 202 ; rheotaxis in isopods, 383, 391, 396. * Animal Psychology, laboratories for, 176. Amoeba, reactions to food, 388. Amphioians, literature on, 416, 423 Annelids, reactions of, 385. Assagioli, R. Thinking horses, 436. Association, literature on, 407. * Audition, see hearing. Austen, E. E. Fly and health, 406, 409. Babak, E. Color vision in frogs, 417, 436. *Baboons, sex tendencies in, 295. Bailey, V. Life zones, 405, 409. Baldasseroni, V. Animal intelligence, 436. Balss, H. Chemical sensitivity in crustaceans, 383, 391. Bancroft, F. W. Heliotropism in Eu- glena, 383, 391. Barber, H. S. Life-history of beetles, 409. Barrows, W. B. Instincts of bittern, 427, 436. Baunacke, W. Function of the stato- cyst, 384, 391. Bee, color vision in, 397. Beetle, behavior of, 400. Benard, G. Behavior of beetles, 400, 407, 409; acrobatic feats of insects, 407, 409. Bergtold, W. H. Behavior of house finch, 427, 436. Biddle, E. Hibernation of butterfly, 404, 409. *Bingham, H. C. Definition of form, 136; visual perception in chick, 419, 436. Bird, H. Habits of Oligia, 406, 409. Bird, literature on, 419, 422. Bishopp, F. C. Biology of tick, 402, 404, 406, 409. Blocloek, B. Insects and disease, 406, 409. Bohn, G. Memory in lower organ- isms, 384, 391." Brauns, H. Behavior of aphids, 406, 409. Breed, F. S. Development of instinct, 426, 438. Brierley, W. B. Life history of Lep- losphaeria, 409. Brocher, F. Respiration in insects, 409. Browne, F. B. Life history of beetle, 409. Brues, C. T. Distribution of stable flies, 405, 406, 409. Brundin, M. Light reactions of am- phipods, 384, 391. Brunelli, G. Behavior of hermit crabs, 384, 391. Buddenrock, W. v. Function of the statocyst, 385, 391. Burrell, A. C. The giant midge. 406, 410. Buttel-Reepen, v. Thinking horses, 436. Buttrick, P. L. Breeding habits of mosquitoes, 399, 410. *r* amera lucida, for maze, 56. \_j Car, L. Locomotion of infusoria, 385, 391. *Cat, hearing in. 70; *instmcts of, 293; learning in, 428. Chaine, J. Termites, 410. Champion, G. C. Behavior of scirtes, 406, 410. Chapman, T. A. Behavior of Agriades, 406, 410. *Chartometer, for maze, 58. Chemical sense, literature on, 423. Chemotropism, literature on, 394. *Chiek, visual acuitv in, 340. Chubb, E. C. Habits of spiders, 403, 410. Claparede, E. Trained horses, 434, 436. Claude, D. Jumping spiders, 405,410. Clementi, A. Functions of the ner- vous system of diplopods, 385, 392. INDEX Coad, B. R. Habits of mosquito, 399, 410. *Coburn, C. A. The behavior of the crow, 185. Cockle, J. W. Behavior of Bombus, 410. Coelenterates, modifiability in, 387. Cockerell, T. D. A. Behavior of hem- iptera, 405, 410. Cole, L. J. Locomotion in the star- fish, 385, 392. Cole, L. W. Behavior of raccoons, 415. Collinge, W. E. Behavior of Collem- bola, 406, 410. Comstock, J. H. Spider book, 404; silk of spiders, 400, 410. Cooke, \Y. W. Bird migration, 428, 436. Copeland, M. Olfactory reactions, of fish, 423, 436; of newt, 423, 436. Copepoda, the behavior of, 386. Courtship of animals, 439. *Cowles, 11. P. Behavior of starfish, 380; habits of Crustacea, 385, 392. Crab, behavior of, 384. *Craig, W. Behavior of doves, 121 ; stimulation of egg laying, 427, 436. Cros, A. Behavior of Hymen'optera, 406, 410. *Crow, behavior of, 185. Crustacea, habits of, 385, 390; vision in, 388, 392; acquired color response in, 390; functions of antenna, 390. * P\ aphnia, movements of, 229 ; L/ color responses of, 386. Davis, J. J. Life cycle of Lachno- sterna, 410. *Dice, L. B. Movements of daphnia, 229. Disease, in relation to insects, 406. ft Distance records for maze, 60. Doane, R. W. Behavior of beetle, 402, 410; insects and disease, 410. "Dog, olfactory responses of, 76; *auditory reactions of, 142; ^visual acuity in, 340; learning in, 428. *Doves, behavior of, 121. Dubois, R. Reactions of Echinoderms to light, 392. Dutt, G. R. Life history of Hymen- optera, 410. t} cology, literature on, 405. It Ely, C. R. Feeding habits of Cleonus, 402, 410. Emotions, literature on, 398. Erhard, |l. Color responses of daph- nia, 386, 392. Ettinger, M. Thinking horses, 436. Ewing, H. E. Hibernation of lady bug, 404, 405, 410. I ^ abre, J. H. Social life of insects, 399 ; habits of spiders, 403, 405, 410. Fasten, N. The behavior of Copepods, 386, 392. *Field Station for animal behavior, 176. *Fish, behavior of, 1; literature on, 417. Fiske, W. F. Behavior of Glossina, 406, 410. "Flowers, visited by insects, 147. Folbort, G. V. Inhibitory conditioned reflexes, 377. *Form perception, in animals, 134, 136. Franz, S. I. Right and left handed- ness in monkey, 425, 436. Franz, \. The behavior of snails, 386, 392; the value of phototaxis, 386, 392. Frideman, S. S. Further contributions to the physiology of differentia- tion of external stimuli, 376. Frisch, K. v. Color responses of Crus- tacea, 386, 392; color vision in insects, 397, 410; color vision in fishes, 417, 436. Frolich, F. W. Light and color vision in Octopus, 387, 392. Frost, C. A. Habits of Diptera, 403, 410. Frohawk, F. W. Life history of Ar- gyimis, 406, 410. *FnuVfly, behavior of, 223. Fulton, B. B. Habits of crickets, 399, 412. C^ ee, W. The behavior of leeches, J 387, 392; modifiability in the sea anemone, 387, 392. Geotaxis, in daphnia, 229 . Geotropism, literature on, 394. Gerhardt, U. Sex behavior of crick- ets, 398, 410. Gillette, C. P. Behavior of populus, 402, 410. Vl INDEX Girault, A. A. Habits of insects, 405, 410; letisimulation of beetle, 407, 410. Green, E. E. Humming of midges, 407, 411. Gregarines, movements of, 385. < I regg, F. M. Behavior of raccoons, 415, 436. Guyenot, E. Behavior of Drosophila, 400, 411. Habits, of Crustacea, 385, 390; breeding of Nereis, 388; literature on, 425. Hadwen, S. Tick paralysis, 406, 411. Haggertv, M. E. Behavior of apes, 425, 437. "Hamilton, G. V. Sexual tendencies in monkeys, 295. Haenel, H. Thinking horses, 431, 437. Harte, t'. R. Flight of moth, 405, 411. Hartman, C. Habits of bee, 400, 411; habits of wasp, 407, 411. "Harvard Laboratory of Animal Psy- chology, 176. *Havs, G. P. Orientation of Porcellio, * 110. "Hearing, in cats, 70; *in the dog, 142; :: iu the white rat, 216; literature on, 421. Heinrich, R. Light reactions of in- sects, 394, 411. 'Hen, brooding instinct of, 266. Hentschell, H. Aquaria for insects, 408, 411. Heredity, of savageness and wildness, 425*1 Hess, C. The color sense of animals. 388, 392; vision of amphibians, 417, 437. 1 1 ei ins. Transmission of poliomye- litis, 406, 412. Herriek, G. W. Scale insects, 406, 411. Hibernation, literature on, 404. Hodge, C. F. Movements of flies, 405, 411. : "Holmes, S. J. Literature on behavior of invertebrates, 383 ; light- reactions of dermestidae, 394, 411; orientation to light, 395, 411. Homing, literature on, 407. Horse, performance of trained, 431, 432. Iluhbert, H. B. Time versus distance, 60. "Hunter, W. S. Hearing in white rat, 215; form perception, 397, 411; delayed reactions, 435, 437; *on infancy and courtship, 439; "ideas in animals, 442. in animals, 442. deas, Imitation, in parrot, 422. *T, Imms, A. D. Behavior of Indian in- sects, 411. "Infancy, of animals, 439. Infusoria, locomotion of, 385. "Insects, in relation to flowers, 147; ^perception of sex in, 289; light reactions of, 394; "literature on, 394; acrobatic feats of, 407. *Instinct, of male doves, 121; "brooding, 266; *of silk worms, 289; ^persistence of, in cats, 293 ; *of monkeys, 295 ; methods of studying, 393 ; mating literature on, 398; maternal, 399; nest-building 399; defensive, 402; procuring, 402; literature on, 406, 425; development of, 426 ; ^social, 439. *Invertebrates, literature on, 383 ; color sense of, 388. *Isopods, rbeotaxis in, 202. Jennings. Insects and pellagra, 406, 411. Jerofeeva, M. N. Electrical irritation of the skin of the dog, 378. *Johnson, H. M. Smell, in the dog, 76 ; "form perception in animals, 134; "visual pattern discrimination, 319; "visual acuity in animals, 340; audition in dogs, 415, 421, 437. Just, E. E. Breeding habits of Ne- reis, 388, 392. Katz, D. Vision of night birds, 420, 437. Kchschkowsky, K. Reactions of ver- mes to electricity. 388, 392. Kellogg, C. E. Graphic method for maze, 50. Kepner, W. A. Reactions of Amoeba to food, 388, 392. Kin-. W. V. Biology of tick, 402, "404, 406, 409. INDEX vn Klein, F. Behavior of Chrysomela, 411. Knab, F. "Forest Malaria." 406,411. Kawlbersz, G. J. v. Reactions of iso- pods, 388, 392. Krall, K. Trained horses, 432, 437. Kupelwciscr. 11. Color responses of Crustacea. 386, 392. *T abyrinth, see maze. \_j Ladd- Franklin, C. Color vision in bees, 397, 411. "Lashley, K. 8. Persist ence of an in- stinct, 293 ; imitation in parrot, 422, 437; development of young monkey, 425, 437. "Learning, time and distance in. 60; literature on, 428. Leech, modifiability in, 387. Leplat, G. The eye of birds, 421, 437. Letisimulation. literature on, 407. "Light, orientation oi Porcellio to, 110; *reactions to, of starfish, 380; reactions to, 384, 389; reactions to, in crnstacea, 390; response to, by snails, 391; orientation to, 395. Lillie, F. R. Breeding habits of Ne- reis, 388, 392. Linstow. Diet of caterpillars, 402, 411. "Literature for 1913, on invertebrates, 383; *on spiders and insects, 394; "on behavior of vertebrates, 414. Locomotion, literature on, 405. *Lovell, J. H. Flowers and insects. 147; a vernal bee, 411. Lntz, A. Forest Malaria. 411. MacCnrdy. Reactions of starfish to light, 389, 392. MacKenzie, W. Thinking horses, 431, 437. Mammals, literature on, 414. Mangin, M. Thinking horses, 437. Mangold, E. Animal hypnotism, '428, 430. Mansion, J. Insects living in formol, 411. Mathews, A. Habits of Gfatn/marus, 390, 393. Matula, J. Functions of antenna of lobster, 390. 393. "Maze, graphic method for, 50; "circular, with camera lucida, 50; "time and distance records for, 60. McGraw, K. \Y. Orientation to light, 395, 411. Mclntvre, .1. L. The rule of memory, 431, 437. McPheeters, C. A. Behavior of rac- coons, 415, 436. Meijere, J. C. A. Habits of Orthop- tera, 406, 411. Memory, in lower organisms, 384; literature on, 407, 428. Menegaux, M. A. Educated horses, 437. Metalnikow, S. Choice of food by paramoecia, 389, 392. "Methods of studying vision, 340. Migration, literature on, 405. Mitzmain, M. B. Insects and disease, 406. 411. "Modifiability, in fishes, 1; "in land animals, 31; in coelenterates, 397; in leeches, 387. Mollusca, habits of, 389, 390. ■.Monkey, sex tendency in, 295; "visual acuity in, 340 ; development of, 425. Moore, a. R. Phototropism of D'ki/)- tomus, 389 ,392. Morgan, A. C. Behavior of beetles, 407, 411. Morgan, A. H. Behavior of Mav-flies, 398, 402, 411. "Morgulis, S. Reactions of the dog, '142; "Pawlow's theory of function of ner- vous system, 362. Morse, E. S. Habits of solenomya, 389, 392. Mosquito, behavior of, 395. Movements, Brownian, 390. Mrazek, A. Locomotion of branchi- pus, 389, 392. Newell, W. The rice weevil, 39S. 402, 411; letisimulation of weevil, 407, 411. Nichols, M. L. Habits of bee, 400, 412; homing of bees, 407, 412. Niewenglowski, G. H. Transmission of malaria, 406, 412. Octopus, vision in, 387. "Olfaction, see smell. 76. "Orientation, of Porcellio. 110. Orton. J. H. Natural history of lim- pet, 390. 392. O'Shea, M. V. Educated horses, 437. Vlll INDEX Paramoecium, choice of'food by, 389. Parker, <;. II. Chemical sense of vertebrates, 423, 437; hearing in fishes, 424, 437. Parrott, R. J. Habits of crickets, 399, 412. •Pawlow, The Method of, 142. •theory of function of nervous sys- tem, 362. *Pearl, it. The brooding instinct, 266. Pearse, A. S. Habits of Crustacea, 390, 393. Phillips, J. E. Bird migration, 427, 437. *Phototaxis, in daphnia, 229; its significance, 386. Pictet, A. Hibernation of moth, 404, 412. Plate, L. Thinking horses, 437. Plessner, 11. Reactions of the star- fish, 390, 393. *Polimanti, 0. Behavior of silkworm moths. 289. Popovivi-Baznosanu, A. Behavior of spiders, 406, 412. Porcupine, behavior of, 429. Protozoa, movements of, 390. Przibram, H. Functions of antenna of lobster, 390, 393. Przibram, K. Movements of protozoa, 390, 393. *Pycraft on the infancy and courtship of animals, 439. Phototropism, literature on, 391. * O at, hearing in, 215. 1\ Pan, P. and N. The biology of mantis, 399, 402, 412. homing of wasps, 407. 412. Revesz, G. Vision of night birds, 420, 437. Reese, A. M. Reactions of newt, 437. *Reflex, the salivary, 362. Regen, J. The stridulation of crick- ets, 407, 412. Reiff, W. Light reactions of insects, 394, 412. •Reproduction, the physiology of, 266. *Rheotaxis, in isopods. 202. Rheotropism, literature on. 396. Riley, C. F. C. Responses of toads, 416, 437. Rojanski, N. A. Materials to the physiology of sleep. 377. Roubaiid. Behavior of wasps, 401. Runner, G. A. Behavior of beetles, 407, 411. Sacked, L. W. Behavior of porcu- pine, 429, 437. Savitch, A. A. New materials for the study of nutritive reflexes, 377. Sawyer. Transmission of poliomye- litis, 406, 412. Schneider, K. C. Thinking horses, 437. Schwantke, C. Thinking horses, 437. Scitz, A. Vision in insects, 397, 412. Sekera, E. Habits of nemertian, 393. •Severin, H. H. P. and H. C. Behavior of fruit fly, 223, 402. *Sex, reactions of doves to, 121 ; *behavior in fowls, 266 ; •perception of, 289; •behavior in cats, 293 ; *behavior in monkeys, 295. Sexton, E. W. Habits of Gammarus, 390, 393. *Shelford, V. E. Behavior of fishes, 1 ; *behavior of land animals, 31 ; reactions of fish, 423, 437. Shephard, J. F. Development of in- stinct, 426, 438. "Shepherd, W. T. Sound discrimina- tion by cats, 70. Sherman, A. R. Habits of hawk, 427, 438. "Silkworm, perception of sex in, 289. Skinner, H. Hibernation of house fly, 404, 412. •Smell, in the dog, 76; *in fruit lly. 223; literature on, 423. Smith, L. W. The biology of stone fly, 398, 412; biology of Perla, 402. 412. Snail, the behavior of the, 386; light response of the, 391 ; mating behavior of the, 391. Social relations, 384. "Sparrow, light discrimination in, 79. Speech, literature on, -407. "Spider, literature on, 394; behavior of, 404. •Starfish, behavior of, 380; locomotion in, 385 ; reactions of, to light, 389; vision in, 390. Statocyst, the function of, 387. Slander, H. The biology of Lysman- tria, 406, 412; behavior of caterpillars, 407, 412. Stevens, H. C. Acquired reactions in crab, 390. 393. Strand. E. Biology of Diapalpus, 406, 412. INDEX Reactions of snail, IX Szymanski, J. S 391, 303; methods of studying instinct, 393 ; habit formation in dog and cat, 428, 4.>8. I^aliaferro, W. H. Reactions of amoeba, 388, 392. "lashiro, 8. Rheotaxis in isopods, 202. Tchecotareva, O. M. Further contri- butions to the physiology of con- ditioned inhibition, 370. Technique, literature on, 408. Theocritova, U. P. Time as stimulus of salivary gland, 370. Thigmotropism, literature on. 396. Thompson, J. A. Thinking horses. 438. "Time records for maze, 60. *Torrey, H. B. The orientation of Por- cellio, 110; trials and tropisms, 391, 393. Townsend, C. H. T. Disease producing organisms, 400, 412. Trogardh, 1. Chemotropism of insects, ■ 394, 412. Tropisms, 386 ; of insects and worms, 391; and trials, 391, 393. Tugman, E. F. Light discrimination in the sparrow, 79. *Turner, C. H. Literature on behavior of spiders and insects, 31)4 ; behavior of roach, 398, 412; acrobatic feats of insects, 407, 413; memory in roach, 408, 412. Unsicker, migration of moths, 405, 413. Urban, C. Life history of beetle, 413. Vassiljev, P. N. Differentiation of thermal stimuli by dog, 376. Vermes, habits of, 388. "Vertebrates, vision in, 319; literature on, 414. Vesme, C. Thinking horses, 438. Vestal, A. C. Distribution of grass- hoppers, 405, 413. ^Vincent, S. B. Literature on behavior of vertebrates, 414. Vision, in the sparrow, 79; *and form perception, 134, 136; *in insects, 147, 397; *in the crow, 185; "pattern discrimination. 310, 340; *acuity of, 340; literature on, 414; *in the starfish, 380; color responses of crustacea, 380 ; color responses of daphnia, 386; light and color, in octopus, 387 ; the color sense of animals, 388; color response in crabs, 390; in starhsh, 390; response to color in infusoria, 392; of fishes, 417. Yolkelt. Ideas in animals, 442. Walker, E. M. Sex adaptation, 398, 413; life zones, 405, 413. Warrington, Y. Cause of disease, 400. 400. Wasp, behavior of. 401. Watson, J. B. Maze and camera lu- cida, 56; development of young monkey, 425, 437; Watson, J. B. and M. I. Vision in rodents, 438, 414. Webster. Feeding habits of Gypona, 402, 413. Weiss, H. B. Odor preferences in in- sects, 394, 413; behavior of mosquitoes, 395, 413; thigmotropism of mosquito, 396, 404, 413; death feigning of weevil, 407, 413. Wells. B. W. An acrobatic fly, 407. 413. Wheeler, W. M. Behavior of solitary wasp, 401, 413; behavior of lx>es. 403, 406, 413. Wiegge, C. Thinking horses. 438. Williams, C. B. Behavior of Baphidia, 406, 413. Winslow, C. E. A. Infant paralysis. 406, 413. Wodsedalek, J. E. Reactions of der- mestidae, 413. *"\7erkes, R. M. A laboratory of Ani- 1 mal Psychology, 176; heredity in rats, 425, 438. "Yerkes, R. M. and Kellogg, C. K. Graphic method for maze. 50. Young, E. The eves of snails. 301. 393. Zetek, J. Flight of mosquitoes, 405. 413; methods for study of insects, 40s. 413. Zimmermann. K. Habits of Galathei- den, 391. 393. • JOUR NAL OF ANIMAL BEHAVIOR Vol. 4 JANUARY-FEBRUARY, 1914 No. 1 RAPID MODIFICATION OF THE BEHAVIOR OF FISHES BY CONTACT WITH MODIFIED WATER VICTOR E. SHELFORD AND W. C. ALLEE From the Zoological Laboratory of the University of Chicago. Page I. Introduction 1 II. Conditions and Methods of Study 3 III. Experimental Results 6 IV. Interpretation of Results 22 V. Bibliography 29 I. INTRODUCTION At the present time animal behavior is being studied from three or four more or less independent points of view: (1) the point of view of tropisms and reflexes or the study of specificities of behavior, (2) the point of view of positiveness or negativeness to environmental factors — an aspect of ecology or the interpre- tation of the relation of animals in their normal environments, (3) the point of view of speed or vigor of reactions and of reflexes in relation to the so-called physiological states and their modi- fication — an aspect of physiology, because such states are due to physiological changes such as a change in the rate of meta- bolism, and (4), the point of view of the modification of behavior by repetition of action or repeated stimulations — an aspect of psychology. Such a separation of view points can be only pro- visionally made as all are probably resolvable into physiology. Most workers combine two or more of these points of view, so that one type or aspect of behavior acts as an index of another. The first two view points noted deserve some comment. Specificities of behavior may be defined as those peculiarities of action which characterize species, genera, or even larger groups. 2 VICTOR E. SHELFORD AND W. C. ALLEE The senior author, for example, is able to distinguish some species of tiger beetles (about one-half inch long) occurring near Chicago, by their peculiarities of flight. Again, ornithologists depend much upon characteristic movements to distinguish species of birds. The details of action in the more stereotyped reflexes of Protozoa and other of the lower invertebrates, such as the backing and turning of Paramecium, are characters of species or of groups of species. These specificities have been much studied by some zoologists, but, like other specific char- acters, serve chiefly as material for the study of modification, and as characters to be used in cross breeding. The point of view of ecology is the one of most recent develop- ment. It considers all phases of physiology that are related to the life of the animals in their natural environments. The behavior aspects which have been developed center around the following questions: (1) Do animals select their habitats ? (2) Is the behavior of the same species different under different conditions ? (3) Is there community of behavior among ani- mals of the same or similar habitats ? (4) To what factors of the environment do animals respond and what is the degree of the response to the different factors ? A relatively small number of investigations have answered the first three questions in the affirmative, for the particular cases studied. Studies on the effect of particular factors of the environ- ment have been made, but usually with very small animals and under conditions which made accurate measurement and control of the experimental factors difficult. One of the most neglected aspects of physiology and behavior is the reaction of larger ani- mals to the different factors involved in the surrounding medium. From several points of view the importance of investigations at this point seemed sufficient to more than justify an attempt to determine whether or not fishes react to differences in the dissolved content of the water which they inhabit and whether or not different species differ in their reactions. The data relating to these two questions have been organized and pub- lished ( Shelf ord and Allee, '13). Apparent rapid modification of the behavior of the fishes by repeated contact with the treated water was quite characteristic of their reactions and aroused interest sufficient to cause us to go over the results of one hundred experiments from the point of view of modification. This in- MODIFICATION OF THE BEHAVIOR OF FISHES 3 volved re-counting of movements and recalculation of time, etc., which, together with the examination of a large amount of liter- ature has delayed the preparation of this aspect for several months. II. CONDITIONS AND METHODS OF STUDY The experiments were conducted under conditions as nearly uniform as possible with respect to all factors except the amount of the various solutes used. (For detailed plans and figure of the apparatus, see Shelf ord and Allee, '13, pp. 225-229). Two galvanized iron boxes, 120 cm. long by 14 cm. deep by 20.5 cm wide, with screen partitions 5 cm. from the ends, making the in- Control :^ + ii +-* Experiment Tap__ 49^ Ta ^co 3 30 .2 00 r 3 10 4 :- 4 35 _I0 ' Figure I The upper two quadrangles show the ground plan of the experimental tanks with the positions of the central drain and the lights under which the fishes went when passing to and fro. The positions of the lights are indicated by the crosses and that of the drain by the double broken line near the center. The broken lines near the ends indicate the positions of the screen partitions. The tanks are rep- resented a little less than one-twentieth actual size. The ends A receive tap water and AM treated water. Below the tank used for the experiment the method of recording movements is indicated. A portion of the record sheet is shown about twice natural size, and immediately below the reduced tank so that the marks dividing the tanks into sixths are opposite the corresponding dividing lines of the ruled paper. The movements of two mud-minnows during the first five minutes of expt. 49A are indicated with the time of the principal movements, in minutes and seconds after the beginning, shown in figures. The graph is an enlarged copy of the original . The solid line indicates that the two fishes were moving together; the broken line shows the movement of a single fish. 4 VICTOR E. SHELFORD AND W. C. ALLEE side length 110 cm., were painted dull gray and covered with yellow sand while still adhesive. Water was allowed to flow in at both ends at the same rate (usually 600 cc. per minute) through tees made from iron pipe, the cross bar of which contained a number of small holes. The cross bars of the tees rested on the bottoms of the tanks behind the screens. The drains were transverse tubes with their lower sides made of screen, located near the top and opening outside the boxes. The water flowed in at the ends and drifted toward the center at the top and flowed out through the drain. We found no evidence that fishes react to the slight current thus produced. Since each half of the tanks held about twelve liters, it required twenty minutes to fill them or to replace all the water in one of the halves. Both tanks were enclosed under a black hood, side by side as shown in the plan (Fig. 1) and were placed about ten centi- meters apart. Two four candle power incandescent lights were fixed above the center of the two halves, i.e., above a point midway between the screen partition and the center drain. The light was thirty centimeters above the surface of the water, which was ten centimeters deep. The lights above a given tank illuminated the outer wall of the other tank (see Fig. 1), while the inner wall of the same tank cast a shadow throughout its entire length. The two tanks were identical longitudinally, but the shadow was reversed with respect to points of the compass. The room was darkened during the experiments, which were observed through openings in the hood above the lights. Fishes do not usually note objects separated from them by a light. The fishes not accustomed to aquaria sometimes showed fright and behaved erratically when first put into the tanks, but all such experiments were thrown out. The main stock of fishes was kept in the laboratory in glass-sided aquaria during the period of experimentation. In this way they became accustomed to an aquarium, to the presence of moving objects, and to variously placed lights. The purpose of the experiments was to test the reactions of the fishes to a difference in the water in the ends of one of the experi- mental tanks. Water differing as little as possible from that in which the fishes usually live was introduced at both ends of the other tank in most experiments (control) . ' Treated water from the device already described (Shelford and Allee, '13, p. 214), was MODIFICATION OF THE BEHAVIOR OF FISHES 5 introduced at one end of the experimental tank, while water differ- ing as little as possible from that in which the fishes normally live was introduced into the other end. Various kinds of treated water were used as follows: (1) boiled water — oxygen, nitrogen, carbon dioxide, and bicarbonates in part removed; (2) water with varying amounts of carbon dioxide added ; (3) boiled water with oxygen added (either against tap or boiled water at the other end) ; (4) boiled water with either carbon dioxide, acetic acid, or ammonia added. Various combinations of these factors were also tried and nitrogen was added in a few experiments. When the difference between the solutes at the two ends of the tank was not great, we found by chemical tests that the central portion of the tank was a gradient between the character- istic waters introduced at the two ends. Usually the end thirds were essentially like the inflowing water. When the difference in concentration was great the region of the gradient was pro- portionally longer and the ends with the inflowing concentra- tions were accordingly shorter. When the difference in con- centration was very great the entire tank was gradient. During the experiments the two authors worked together. Three fishes were placed in each of two dishes containing enough water to barely cover them and set above the tanks. When all was in readiness and at a time agreed upon the two lots of fishes were emptied into the centers of the tanks. Marks on the sides divided the tanks into sixths. The fishes nearly always swam back and forth apparently exploring the tanks. The movements of the fishes were recorded graphically as shown in figure 1 . For this purpose sheets of ruled paper were used. Three vertical double rulings correspond to the center and two ends of the tanks. Two pairs of single rulings divided the space between two primary rulings into three equal parts and the entire distance from right to left into six parts. Distance from right to left was taken to represent the length of the tanks; vertical distance to represent time which was recorded in minutes and seconds at the center. The width of the tanks was ignored. The graphs on the following pages are copies of the originals with the time corrected to scale. Before or after the experiment, the headings of the sheets were filled in with data regarding the kind, size, and number of fishes, their previous history, the conditions in the tanks, concentra- tions of the solutes and other significant data. Details of the 6 VICTOR E. SHELFORD AND W. C. ALLEE reflexes of the fishes and notable peculiarities of behavior were recorded in full at the left of the graph. The fishes were observed continuously for from twenty to ninety minutes. In many cases they were caught at the end of such a period of observa- tion in a small hand net and replaced in the small dishes. Ob- servers then changed places and transferred the fishes from one tank to the other. Thus the control fishes of the first experiment were observed in the experimental tank and the experimental fishes were observed in the control tank for a time equal to the first test. III. EXPERIMENTAL RESULTS Most of the fishes studied reacted negatively to various concen- trations of carbon dioxide; to little oxygen; to boiled water with the removed oxygen restored; and to boiled water with acetic acid or carbon dioxide added (see table 3, p. 20). The behavior of the fishes when giving a negative reaction usually possessed prominent features. They tried the modified water a number of times and then began to turn back in the lower and lower concentrations of the gradient, or to spend shorter and shorter time in the modified water with each visit. Usually these modifications did not show a uniform gradual decrease in time spent in the modified water or an increasing tendency to turn in lower and lower concentrations of the gra- dient which extended throughout the experiment. The response was rhythmic rather than cumulative. This may be seen in the charts especially in Expt. 10, chart 1, p. 9, and in controls 78 and 83, chart 2, p. 11. A number of successive trials of the modified water resulted in either a lessening of the time spent in that water or in turnings in the gradient, or both. This led to spending more time in the untreated water. After some time in this water there was again a tendency to enter the treated water with the same results as before. That is, there was a rhythm of reaction, which, while not perfect, was present to a recognizable degree in the majority of cases. The number of trials of the modified water at the beginning of the experiment was in most instances inversely proportional to the degree of stimulation as indicated by the special activities, except where the concentration was great enough to cause " staggering," or other abnormal reaction. The control fishes on the other hand, went back and forth quite symmetrically. MODIFICATION OF THE BEHAVIOR OF FISHES The results of the experiments on each species were tabulated for the purposes of this paper as shown for the river chub in table 1, and for the golden shiner in table 2. TABLE 1 Showing the varying speeds of the modification of the behavior of the river chub {Hybopsis kentuckiensis Raf, habitat, small, clear streams) associated with different kinds of modified water. The ratings given in column 3 represent degree of avoidance of the kind of water given first in column 2. The ratings were ob- tained by averaging per cent of time in the two halves and turnings from the two halves. The rating is 100 when all turnings were from one-half and all the time was spent in the opposite half, and when time and turnings were equally divided between halves. These figures are taken directly from table 20, p. 256 of the pre- ceding paper (of which see pp. 254-57 for further details*). The control rating for the species is +0 based upon eleven controls. c -y c ^TJ-2 •O CO i_, C b0'-£ C C C_ CD 03 T3 -M . Pa O '•3 c £? - & r 1 111 SB'S w 2 8 .£ u £ CD 03 £ £ 2 tt Q Q H 2 2 H H 2 C0 2 in boiled vs. boiled 97 21 6 6 1.0 2 C0 2 in boiled vs. tap. 94 7 47 6 6 1.8 1 CO 2 (weak) in tap vs. tap 1 5-8 1 1 3.0 110 30 2 CO2 (strong*) in tap[ 87 vs. tap j 18 8 6 1.3 169 510 3 Boiled vs. tap 67 7 2 9 9 3.7 95 221 2 Boiled vs. boiled plus oxygen 34 10 6 6 4.3 182 91 2 Boiled plus oxygen vs. tap 33 2 6 6 3.0 137 85 2 Boiled plus nitrogen plus oxygen vs. boiled plus oxygen. 26 6 Nitrogen gradient 4 cc. Totals and aver- 48 40 18.1 693 737 ages 83% 2.6 * Corrections on p. 255 lines 12 and 13 for "tap water or water nearest like that in which they had been kept" read, more modified water. 8 VICTOR E. SHELFORD AND W. C. ALLEE Chart I Showing modification of fish behavior by contact with boiled water, and with boiled with acetic acid added and with boiled water with ammonia added. Distance between the small vertical lines adjoining the scales at the top cor- responds to the length of the tank. The scales represent time in minutes divided into ten-second periods. The horizontal distances in the graphs represent the portion of the length of the tanks traversed by the fishes, the obliqueness of the line as measured by vertical distance and the vertical portions of the graph lines rep- resent respectively the time required to move the distance and the time spent in resting or moving crosswise. The numbers above the generic names at tht top represent the number of fishes used. When all or two of the fishes moved together a solid line occurs. The movements of single individuals are shown by broken lines. Double pointed arrows above the tracings indicate the distance occupied by the gradients. When the concentration was high the entire tanks were gradient and the secondary gradient is represented by the lighter double pointed arrows. A corresponding portion of the controls is likewise indicated. The kind of water introduced at the end indicated by the words "tap" and "boiled" the former being the kind of water in which the stocks here discussed were kept. The oxygen con- tent of the water is given in cc. per liter in the vertical wordings where the amount of added solute is indicated also in cc. per liter excepting acetic acid which is given in grams per liter. For statistical purposes (Shelford and Allee '13) the proportion of time in the two halves, the number of turnings in the gradient are used as data. In experiment 10 the fishes are shown to have entered the boiled water three times during the first two minutes, spending about one-half of their time there. At the end of the two minutes, they began turning back occasionally. This con- tinued until the end of ten minutes when turning became the rule with more or less rhythmic entrance into the treated water. The experiment lasted forty minutes but the remaining ten minutes showed nothing different. When emptied into the control tank the fishes came to rest in one end and remained there for the first twenty minutes, a common reaction when the water is like that from which they were taken. At the end of the twenty minutes the fishes began moving back and forth in a symmetrical manner. Experiment 13 shows the reactions of sun fishes which are representative of the reactions of the fishes studied, to boiled water. The graph of experiment 41 shows a reaction to acetic acid comparable to that given to carbon dioxide. The reaction to ammonia might well be that of a control. Since three fishes were used it was not always possible to distinguish the different individuals in the experiments and in working over the graphs. In the case of the gregarious species, all three nearly always moved together. With non-gregarious fishes it was nearly always possible to distinguish the different individuals in any one group of invasions of the stimulating water. Individuality was lost only in the periods of rest between invasions. After the graphs were made, the tracing of each fish was followed with red, blue or green ink, the individuals being followed so far as possible, but where individuals were lost during a period of rest each tracing was continued without conscious reference to what the individuals had done pre- viously. This was due to the fact that the tracings were made as a basis for calcu- lating time spent in the two ends and before any discussion of modification was in mind. The individuals are then, where not clearly distinguished, treated in a chance fashion, and should show the same average result as the actual movement of the individuals. The comparisons of species and factors are valid because such errors as occur are present in all the experiments and we believe that any errors arising from failures to distinguish individuals are of minor significance. MODIFICATION OF THE BEHAVIOR OF FISHES 9 CHART I a Experi- c g ment 10 g x 3 . « O Hybopsis d Tap Boiled gControl 10 g X 3 K O Hybopsis d Tap Tap c Experi- c g, ment 13 g >> 3 •»• O Lepomis o Tap Boiled [j Experi- h = ment 41 ^ oo 3 ,_; O Abramis d Tap Boiled "' Control, * I 30 & 41 g £ 3 fr d Abramis o Tap Tap o po g Experi- ^S < ment 36 g « O Abramis o^ Tap us o *■• Boiled 10 VICTOR E. SHELFORD AND W. C. ALLEE Chart II Showing modification by contact with boiled water, with boiled water plus oxygen to balance that of the tap water, and with tap water plus carbon dioxide. For further explanation see Chart I. In experiment 78, Hybopsis did not react to the effect of boiling with the oxygen factor eliminated but did react to the boiled water. At the end of about two minutes after all the fishes had tried the boiled water several times, the turnings in the gradient began, and the same type of modi- fication was again shown (see Chart I). Ambloplites reacted to both the boiled water and the boiled water with oxygen added. Detailed study of these fishes from the side in the boiled water while in glass boxes showed that the respiratory movement was increased but that other activity was depressed, this species being an exception in the matter of depression. However, in both parts of experiment 83, after the fishes had tried the water in both ends several times they began to turn back and make shorter stays in the modified water. Experiment 53 shows the reaction where sufficiently high concentration of carbon dioxide was used to produce death in less than an hour. Under these conditions the fishes did not turn back until two had tried the high concentration. The third fish turned with the other two without entering the high concentration, a thing which takes place normally in a gregarious species and thus can hardly constitute a real exception. The graph is typical of the whole series of experiments until the end of eight minutes when the movements become erratic due to the effect of the carbon dioxide upon the fishes. In a later experiment with only a slightly larger amount of carbon dioxide the fishes ceased to react properly after a short time and turned upon their backs. MODIFICATION OF THE BEHAVIOR OF FISHES 11 CHART II Arrows indicate that the fishes were driven Oi Experi- a -= g, ment 78 g, 3 f? 3 « <; o Hybopsis o Boiled Control 78 3 Hybopsis Tap Boiled oi a Exp. 83 | «jO Am b so Boiled Tap -a >> -a "Control 83 B Ml , ° u , M >> Amblo- >> O plites o Boiled Tap Experi Oi O a 50 oi ment 53 >:^ O Notropiso b Tap Tap rfc C0 2 Control 53 3 Notropis Tap Tap 12 VICTOR E. SHELFORD AND W. C. ALLEE Chart III Showing modification by contact with boiled water with 50 cc. per liter of car- bon dioxide added. For further explanation see Chart I. In these experiments the fishes tried the gradient of the entire tank a number of times and then either remained in one end except when driven as indicated by arrows, or turned back often before the center was reached. For example Hybopsis (experiment 56) entered the modified water once or twice and then came to rest in the low concentration, invading the high only when driven with the exception of a single excursion which followed a disturbance. The graph of Ameturus shows one of the very few cases in which the fishes turned back before the strongest stimulus had been encountered. Still the graph brings out the same general fact of modification. The gradient in experiment 76 was established with some difficulty as the nitrogen was only 93% of the gas available the rest being oxygen. Boiled water had to be used at both ends and sufficient oxygen added at one end to balance the oxygen added with the nitrogen at the other. Even here, with the factors involved somewhat in doubt, the same modification is suggested. MODIFICATION OF THE BEHAVIOR OF FISHES 13 CHART III Arrows indicate that the fishes were driven — 00 Experi- >o S ment 56 c5 fc 3 2 O Hybopsisd Tap Boiled ^Control 56 g. r"> 3 >> O Hybopsisd Tap Tap a Experi- < ~ > ft ment 61 o 2 g oAme: Tap ft ', J 14 E E j.-t::-:-^ "3= = : i = = -=-^ V =4l. ^ r ; = - 1 = z i £ ' \ \ - z. I E E urus o Boiled ^Control 61 ft ^> 2 >. O Ameiurusd Z Experi- ^ [7 merit 66 ~ O Hybopsis d Boiled + .\o B. ftControl 66 ft O Hybopsis o Tap Tap :x::::ir in H 14 VICTOR E. SHELFORD AND W. C. ALLEE Chart IV Showing the establishment of apparent preferences by groups of individuals of Abramis. In experiment 1A an apparent preference already existed for the right hand end. An apparent preference for the left hand end was established and broken again by introducing boiled water as indicated. In experiment 84 the graphs of experiment and control are very similar during the first ten minutes. The apparent preference for the left hand end of the control tank was modified and apparently broken by repeated drivings indicated by arrows, and by confining the fishes in the avoided end during the nineteenth minute. The control fishes of experiment 58 established an apparent preference for one end after ten minutes and for reasons unknown. For further explanation see Chart I. MODIFICATION OF THE BEHAVIOR OF FISHES 15 CHART IV Arrows indicate that the fishes were driven Experiment 1 and 1A Experi- Control 84 Control 58 Control Experiment 1 and 1A Experi- 1, 1A ment 84 One space equals 20 sec. Boiled + Tap Tap Tap Boiled Boiled Tap Boiled C0 2 Tap Tap Tap Tap M-X t« — 1 I L = : : : E~ " ,,r 1 j "^ "- L = = L ""\ 16 VICTOR E. SHELFORD AND W. C. ALLEE TABLE 2 Showing the varying speeds of modification of the behavior of the golden shiner {Abramis crysoleucas Mit. Habitat, stagnant ponds.). The rating of the controls + 11 is based on 15 controls. For further details see table 1. , C X'- 1 arbon grad- en "to T3 bo a invasions of modi- - before first turning 9 ^ c Factor or factors, avoided water first n amount of nd of gradient n amount of c t each end of . per liter > '■3 a S-. (LI X> CO 13 ■> CO econds in a , \ m beginning t odification conds in unti m beginning t of modificatio ^ 73 ifference i at each e per liter ifference i dioxide a ient in cc Id . . -*J 3 3(C! me in s water fro sign of m me in se water fro evidence X Q Q H Z Z H H C0 2 in boiled vs. boiled . . 85 21 6 6 1.8 63 71 CO2 in boiled vs. tap .... 91 7 47 6 6 3.6 52 268 CO2 (weak) in tap vs. tap 60 5 3 3 2.0 43 62 CO2 (strong) in tap vs. tap J 65 6 6 3.5 66 96 Acetic 12 grms. per L. in boiled vs. 3.3 grms. per L. in tap 85 7 2 2 2 3.0 215 485 Boiled vs. tap 75 7 2 9 4 4.0 202 262 Boiled vs. boi.ed plus 2 . 75 10 3 p Boiled plus O2 vs. tap 39 2 3 ? Boiled plus NH3 85 cc. per L. vs. tap 1 cc. per L 8 7 2 2 Totals and averages . . . 40 27 67% 17.9 2.98 641 1244 The behavior of Hybopsis is further illustrated in chart 1, Expt. 10, the data from which are included in the table. It will be noted that in the reaction recorded in the chart the three fishes first turned back after two invasions of the boiled water and that turnings occurred more often than entrances after fifteen trials of the boiled water. It will also be noted that the entrances of the boiled water toward the end of the experiment occurred only after some time had been spent in the untreated water. In this case there was less activity in the control as MODIFICATION OF THE BEHAVIOR OF FISHES 17 the fishes remained in one end for thirty minutes and then (see below graph of Expt. 13) began to go back and forth more slowly than in the experiment. Two turnings at the center are shown by the control fishes and these are opposite in direction. This shows that the fishes sometimes turn in the absence of solutes. A comparable result with boiled tap water is recorded in chart 2, Expt. 78 (control). Here the turnings are less prominent and shortened stays in the boiled water were followed by resting in the tap water. In the part of Expt. 78, where the amount of oxygen was the same at each end of the gradient there is no evi- dence of modification, but simple disturbance is indicated. This shows that the fish sense the effects of boiling even when there is a normal amount of oxygen in the boiled water. In chart 3, Expt. 56, very rapid modification is indicated, and after the first few trials of the modified water the fishes stayed in the tap water end, except when driven out (indicated by arrows). In the same chart, Expt. 66 shows an apparent modification due to the introduction of an atmosphere of nitrogen and oxygen. This atmosphere had an oily odor which may have affected the results obtained. Expt. 13 of chart 1 and Expt. 83 of chart 2 show the type of mild negative reaction with indications of slight modification given by the sunfishes and basses tried. The graph of Expt. 53, chart 2 shows the symmetrical type of control most common with Notropis and a case of somewhat erratic action in very strong carbon dioxide after it has time to seriously affect the fishes. The graphs of Expts. 41 and 36 of chart 1 show the reaction of Abramis to acetic acid and ammonia in boiled water. On encountering the acetic acid the fishes often gave a definite reflex — the " backing-starting " reaction to be described on page 24. The reaction to acetic acid in boiled water compares very favorably with that to carbon dioxide in boiled water. Such a reaction is shown in Expt. 84, chart 4. The ammonia experiment shows the failure of the animals to react negatively to this factor, although the backing-starting reaction occurred more often than in the acid. The fishes died in the ammonia. The case of Abramis demands special attention as this is the only species studied in which the groups of individuals developed apparent preference for unknown reasons. In this species this 18 VICTOR E. SHELFORD AND W. C. ALLEE peculiar trait seems to be well developed. It avoids mild stimuli by rhythmically shorter stays in the modified water and stronger stimuli by turning back in the gradient. The general behavior when turnings are being given is shown in table 2. In our earliest experiment with Abramis we obtained no results because the fishes remained in the tap water and so failed to encounter the treated water. It was noted also that when these fishes were left in the tanks they tended to stay in one end or the other. This led to the following experiment: Three individuals were placed in each tank with the same kind of water flowing into both ends of each. After both groups had developed an apparent preference for one end, the fishes selected for the control of the experiment to follow were disturbed until they tended to go back and forth. At the same time a shadow was thrown over the end of the tank selected for the experiment opposite to that in which the fishes were staying. The experimental fishes were driven into this shadow several times and soon developed an apparent preference for the shaded end of the tank. The fishes to be used for the experiment were left in the shadow for about two hours. The fishes to be used for the control were driven from end to end several times during this period. The apparatus was then arranged for the introduction of treated water which necessitated the removal of the shadow. Even after the shadow was removed and the apparatus was disturbed, the fishes per- sisted in their apparent preference for the end which had been shaded (Edinger, '01; Holmes, '11). Boiled water was intro- duced in the end in which the fishes were staying. The results of this part of the experiment are shown in Expt. 1-A, chart 4. In the control, the time was divided between the two halves in the ratio of 41 to 59 for the hour of the observation. During the same period the experimental fishes stayed in the boiled water end during the first 29 minutes, when they began going back and forth. They clearly stayed a little longer in the tap water with each excursion. After making five trials of the tap water in a trifle more than ten minutes the fishes came to rest in the tap water end and remained there until the completion of the hour's observation. The boiled water introducer was then placed in the newly ' preferred ' ' end and similar results were obtained modified only by the additional activity of a juvenile individual. In Expt. 84, chart 4, the same species showed a comparable MODIFICATION OF THE BEHAVIOR OF FISHES 19 reaction. The control fishes avoided the end corresponding to the carbon dioxide end of the experiment as clearly as did those in the experimental tank. When they were confined in the avoided end the apparent preference for the other end was broken in the control, but only strengthened in the experiment. In the control of Expt. 58, Abramis established an apparent preference for the end avoided in Expt. 84. Since Abramis does not rest on the bottom, this reaction might be thought to be parallel to the resting of other fishes (chart 1, Expt. 10), were it not for the fact they that choose one end after visiting both many times, while others usually come to rest after the dash which follows their being poured into the tank. The behavior of Abramis in these controls was similar in many respects to the avoidance of carbon dioxide and acid given by this fish and others in the experiments. That is Abramis sometimes reacted positive- ly to one end when tap water was running into the two ends from the same pipe at the same rate. The fishes do not appear to have any special tendency to rest near objects. They may sense the current at the end, but we have no evidence that such is the case. These apparent preferences of Abramis demand further ex- perimentation for their analysis. From the work so far they apparently do not belong in the same category as the rest of the reactions described here and may be due entirely to associative memory. Although if this be true the associated elements are at present unknown. A general summary of the data on modification is contained in tables 3 and 4. In table 4, the various species are arranged roughly in the order of their sensitiveness. Later work by Mr. M. M. Wells has shown that the data in columns three and four are not accurate, because the time to loss of correlation or until "staggering" occurs is rather indefinite and difficult to determine. He uses the time until death. However, the data represent relative sensitiveness in a general way. The time required to accelerate the respiratory movements in low oxygen is likewise difficult to determine, but the data presented are more reliable than that in the two preceding columns. The remaining data are concerned with the number of trials before evidence of modification was given by turning. This is a definite criterion and does not seem open to serious criticism. Reading 6, 7, 8 20 VICTOR E. SHELFORD AND W. C. ALLEE and 9 from top to bottom, we note that generally speaking the less sensitive fish show the greatest number of trials of the modified water before giving evidence of modification. In columns 10 and 11, we note that the percentage of individuals showing modi- fication is greatest in the more sensitive species and that when all are reduced to terms of 100 per cent, of modification the most sensitive species show the smallest number of trials before turning. The time in the two kinds of water is variable, but usually greatest in the unavoided water. Reading the lines from left to right we note that the sensitive- ness of the fishes is different for different stimuli. TABLE 3 A list of the species used; the size of the individuals; physiological relations to the factors used; number of turnings and time spent in each half before modi- fication was apparent. 1 2 3 4 5 6 7 8 9 10 11 12 13 u u c CO CO CU c CU^-v CU^i 0> cd ca O grf s» •* s s Ui CU X CU x CO CJ cu O. J2 X t3a J2 X a) ■So cU c s CO O z z CO 13 ■3 a 5 •a . cu^g 4J CTJ So bo & d ca E{£ cuO S O So SPECIES (Individuals used, E c cu "3 co \~ C£) — c ■go u O u cj ■a'S '"2 +J > O CO +J • c >.2 z «— 1 cC •3 E Ho JO CO •a «j CU > > a — a "8 > '" CO chiefly juveniles) c j= CO 1*4 O CO CO « go is O" 1 ■ s CO ro oj: »— ' *-> O CO W CO OE cj OcC m -5 CO „E a a c .2 co CO > <■> 5 cu O CU X E c 3.2 be s '5 CO CU Eh E" XI ■*-* be c CU J CO*""* u > CU C7) — — .5* 2 §| CJ ? CU in U CU •0,0 C CU cj CU CU -3 . CU k - co«= §? c CU if) c O C CJ C CU Micropterus dolomieu Lac 8- 9 20 40 6 6 100 6.0 257 215 Notropis cornutus Mit 5 9 60 376 45 1.0 0.8 1 100 1.0 97 108 Hybopsis kentuckien- sis Raf 7-10 50 355 50 1.8 1.8 3.7 2.3 100 2.3 251 108 Ambloplites rupestris Raf 5- 9 180 340 55 1.0 1.5 5.4 2.6 58 4.5 200 119 Catostomus commer- sonii Lac 8-11 300 60 1.3 2.3 1.8 60 3.0 95 110 Abramis crysoleucas Mit 6-15 400 60 3.6 3.5 4.0 3.5 67 5.8 268 87 Etheostoma zonale Stor 3- 5 150 20 20 Lepomis cyanellus Raf 4- 6 150 300 2.3 3.5 2.5 2.8 60 4.6 146 165 Ameiurus melas Raf. Low O2 13-15 350 5 Ameiurus melas Raf. High Oo 13-15 10-13 1800 270 2.3 19.6 2.3 40 5.7 53 106 30 Umbra limi Kirt 81 Totals 11.0 1.8 15.8 2.2 41.2 6.8 22.3 2.8 585 65 32.3 4.0 159 Averages 104 MODIFICATION OF THE BEHAVIOR OF FISHES 21 In general the fishes were found by other criteria to be most sensitive to high carbon dioxide in boiled water, to high carbon dioxide in tap water, and to boiled water (low oxygen). It will be noted that the fishes showed modification with fewest trials of the type of water to which they are generally most sensitive. There are several exceptions to this, but the averages of all the species show this very clearly. TABLE 4 Showing the speed of modification due to different conditions Experimental factors. Avoided conditions precede " vs" -a "o C 6 u Acetic acid in boiled vs. tap C0 2 in boiled vs. tap a *-> w > as C 6 . u a > T3 *o 6 "a -a -a '0 pq a a +j > 6 w "E -a "6 CQ as NH 3 (strong) vs. ; NH 3 (weak) Strong Higk 2 Stock Low 2 Stock Boiled pi tap plus No. of species tried . . 2 1? 8 8 5 9 1? 6 6 2 No. of individual trials. Grand total, 230 12 4 38 45 9 35 23 33 27 4 Per cent of individual trials showing mod- ification by turning 100 100 71 97 100 60 61 51 55 Ave. No. invasions of stimulating water before turning (modification) .... 1.4 2.7 1.8 2.1 2.1 4.4 3.8 4.7 5.5 Ave. sees, spent in modified water before turning (modification) .... 30 147 30 119 62 205 86 99 174 Ave. sees, spent in untreated water before turning (modification) .... 138 226 167 219 66 438 100 99 346 22 VICTOR E. SHELFORD AND W. C. ALLEE Table 4 shows for the types of modified water shown in detail for all the species and individuals (230 in all). An examination of the table shows the same general relations as have been brought out for the particular species. The average time spent in the two kinds of water is least in the water producing the modification in all cases save one and in that the time in the untreated equaled that spent in the modified water. IV. INTERPRETATION OF RESULTS The phenomenon discussed and illustrated by the graphs on the preceding pages is clearly one of rapid modification. The behavior of the fishes was different after from one to four entrances into the end where the stimulation was greatest. This modifi- cation is indicated by the two types of behavior suggested above, viz.: (1) Turning in weaker and weaker parts of the gradient; (2) By making shorter and shorter stays in the modified water. This last type of behavior is not considered in the tables, since it cannot be readily tabulated because of its rhythmic nature. In general, fishes swim about either continuously or period- ically when in normal water and under uniform light conditions. This tendency under the unnatural conditions of the experi- mental tanks is strong and though some species may rest for con- siderable periods in the control tank, they periodically move from end to end. The fishes move crosswise of the tanks but this was not considered, since it appeared to be a minor matter because of the narrowness of the tanks. Furthermore, it bore no relation to the experimental conditions other than to prolong the time spent in a particular part of the tank. The modification, which will be chiefly discussed, consisted of breaking the tendency to swim to the end of the tank and of substituting turnings at points of weaker and weaker concentra- tions of the experimental factor or factors. The tendency to pass to the end of the tank as opposed to turning nearer the center may not be markedly strong, for all the fish occasionally turned and swam back or swam about in circles without crossing the center (see chart 1, column 3, last ten minutes of the control of Expt. 10). Still, on account of the small size of the tanks, this tendency is apparently almost as strong as the tendency of fish to swim anywhere when not especially stimulated. No doubt it is about as strong as the tendency of Mobius' pike MODIFICATION OF THE BEHAVIOR OF FISHES 23 (Holmes, '11) to strike its nose against the glass partition, and from the standpoint of the fishes in relation to nature, modifi- cation by contact with stimulating water is of a more significant type. Any explanation of the modification demands first a clear statement of the problem. With figure 1 before us, this state- ment together with a discussion of various possible explanations will be made. One explanation that may be advanced is that the fishes were depressed by the modified water and thus tended to stop swimming forward on entering it and so finally came to rest in the normal water. This is very clearly not the case, although it is suggested by some of the graphs because the cross movements of the fish are not indicated. With one ex- ception, the modified water was stimulating. The exception was a temporary depression of activity of the rock bass due to lack of oxygen. In a few experiments, the rock bass entered the boiled water and, being depressed, stayed there a half hour or more. Finally, however, they began to move back and forth and selected the tap water end. Another explanation is that the fishes are stimulated by the modified water and thus move out of it more quickly and spend more time in the untreated water. This is clearly what hap- pened in many cases. The fishes rushed forward more rapidly when they encountered the stimulating conditions, and upon reaching the end, turned and moved out quickly. Still, this does not explain the turning in the gradient which took place more often than the simple acceleration. Neither does it explain what was also sometimes true, namely, that fishes did not show the acceleration until they had encountered the treated water a number of times. A third explanation that may be advanced is this: When the fishes had been exposed to the low oxygen water or to water containing much carbon dioxide for a long enough time to affect the oxygen or carbon dioxide content of much or all of the blood and thus affect the nervous system as a whole, the fishes began to turn back. The change in gas content caused a change in the physiological state of the fishes, so that they were more sensitive to the surrounding medium. In connection with this explana- tion, and the preceding one as well, certain facts brought out in the experiments should be noted. When dropped into water 24 VICTOR E. SHELFORD AND W. C. ALLEE containing 150 cc. or more of carbon dioxide and only lcc. of oxygen per liter, very striking evidences of stimulation appear almost instantly. In 47 cc. of carbon dioxide per liter in boiled water nearly all the fishes showed evidence of stimulation at once. They would start gulping before reaching the end of the tank on the first entrance of the solute. In only a few of the cases was this delayed as long as 50 seconds. In from 20 to 60 cc. of carbon dioxide per liter in tap water the fishes showed similar stimulation in from two to ten seconds after entering the high concentration. In boiled water some fishes showed increased activity, gulping within ten seconds, but such manifestations were frequently delayed for nearly a minute and were quite variable in intensity. In acetic acid the evidences of stimu- lation were similar to those in carbon dioxide and in ammonia some of them were noticeable. The fishes undoubtedly sense the solutes upon entering them. For this, they give evidence by the following activities: A definite reflex was often given by Abramis, Notropis, Hypopsis and Lepomis the first time they entered the modified water. The fish suddenly stopped, backed quickly a few millimeters and then started ahead again, often repeating the reflex before going farther forward. In the earlier paper, we called this the backing- starting reaction. This may be due to stimulation of the nostrils. Sheldon ('09, p. 278) states that stimulation of the nostrils of the dog fish resulted in a quick jerk of the head. There was ac- celeration or increased vigor of movement of fins, tail or body which began at once or after a very short time. Sheldon found that the application of solutions to these parts caused them to be moved. The opercles were lifted, the lower jaw protruded, or the mouth moved in a manner characterized as coughing, gulping or yawning. Sheldon found that stimulation of the mouth or spiracle gave rise to violent gulps. In our experiments these reactions occurred singly or in combination. The time necessary to produce them was variable, but depended upon the strength of the stimulus, which confirms further observations by Sheldon ('09). As further evidence of the quick sensing of the stimuli, Hybopsis turned back the first time the gradient was tried in nine cases, Notropis in three cases, Ameiurus in four cases, Umbra in one case, and Abramis in one case. With a single exception carbon MODIFICATION OF THE BEHAVIOR OF FISHES 25 dioxide was the factor thus avoided. This is good evidence that fish ascertain the condition of the water by peripheral sense organs or otherwise. The modifications appearing after a number of trials must be due to increased sensitiveness to the modified water or to associative memory or to both. The physiological explanation which may be made for the increased sensitiveness with increased exposure to water high in carbon dioxide is simple, and is based largely upon the relations of organisms to carbon dioxide. The arterial blood of dogs and horses (Hill, '06) has been shown to contain 330-550 cc. per liter of carbon dioxide, free and combined. The free carbon dioxide is about 20 cc. per liter, so in most of the concentrations used more carbon dioxide would be taken up (Hill, '06, p. 533) and its removal from the blood and tissues was undoubtedly hindered in all cases. Since carbon dioxide is constantly produced inside the fish's body the effect of increased concentration on the out- side would become greater with repeated trials of the carbon dioxide water. Waller ('96) has shown that carbon dioxide in small amounts increases the irritability of nerves. Hill ('09) states that similar results have been obtained with micro "organ- isms.* A stimulating concentration of carbon dioxide is generally recognized among physiologists. Because of the increase in inter- nal carbon dioxide brought about by entrance into the carbon dioxide water, the fishes tended to become more sensitive with repeated entrances and hence to turn back in weaker concentra- tions. After spending some time in the weaker carbon dioxide of the tap water end of the gradient tank, they partially recover and tend to resume their usual movements. This brings them again into the modified water and the process is repeated, hence we have a rhythmic invasion of the carbon dioxide water as shown in chart 1, Expt. 10. As was noted in table 3, the respiratory center is stimulated and the respiration movements increased in carbon dioxide and in low oxygen. This requires more time than the reflexes which follow the sensory impressions (Westerlund, '06). In carbon dioxide the increased respiratory movements occur within a few seconds after the reflex movements. The same is true in low *The sensitiveness of fishes to carbon dioxide probably increases with starvation. A stock of fishes kept during the winter of 1911-12 without food showed markedly low resistance in that they lost their equilibrium in from 30 to 50 cc. of carbon dioxide per liter. 26 VICTOR E. SHELFORD AND W. C. ALLEE oxygen, where respiratory changes have been definitely timed (table 3). Our evidence indicates that peripheral stimulation is not of supreme importance in the regulation of breathing, because the fishes reacted definitely to a change in the water before the breathing was affected. Bethe ('03) proposed the hypothesis that the breathing rate of fishes is regulated directly by stimulation from the periphery, particularly by stimulation of the mucous membrane of the mouth and gills. This hypothesis has been widely tested (Baglioni, '10 and citations), but the more recent experimentation (Reuss, '10) seems to show that the breathing of fishes is regulated indirectly as in the higher vertebrates, although the suggestion of Kuiper ('07) that both the automatic center and reflex stimulation are concerned has much evidence in its favor. The important thing in the work of the followers of Bethe from the standpoint of this paper is the establishment of quickly working sense perception in the mucous membranes of the gills and mouth. The negative reaction and modification of reaction of fishes to acid may be explained in a manner similar to that presented for carbon dioxide. Winterstein ('11), Signorelli ('10) and Quagliarello ('11) and others report that it is the acidity of the blood that affects the respiratory center and similar results are nearly always obtained with carbon dioxide and other acids in reversing reaction to light, etc. (Mast, '11 and citations). Ac- cordingly the increased sensibility in acetic acid and in carbon dioxide probably have a common explanation. It is well known that an insufficient supply of oxygen leads to the formation of lactic acid rather than carbon dioxide as the end product of respiration. Fletcher ('98), Fedman and Hill ('11) and Araki ('91) report that lactic acid production bears some inverse relation to oxygen supply. Signorelli ('10) found that lactic acid directly affects the respiration center. When the fishes remain long enough in the low oxygen to affect the amount of oxygen in the blood and tissues, the presence of lactic acid probably results and the plasma tends toward acidity just as when acids are used directly. We thus infer that in- creased sensibility and the resulting modifications are due to acidity just as in the other two cases. Ammonia is present in the blood of mammals and appears to bear some relation to the carbon dioxide (Hopkins and Dennis, MODIFICATION OF THE BEHAVIOR OF FISHES 27 '11). The failure of the fishes to react to ammonia accords with the known effects of the drug. According to Cushny, when ammonia or its common salts are absorbed by the blood, it is not rendered more alkaline but the ammonia is rapidly changed to urea and excreted. The effects of the kation on the common frog is to paralyze the terminations of the motor nerves. While the effect upon the sensory ending appears not to have been investigated, paralysis appears to have occurred in the fishes studied. Frogs and mammals usually die from ammonia poisoning in tetanic convulsions as did the fishes used in the experiments. It may be noted that the stimuli which give rise to the modifi- cations most quickly are those commonly encountered by fishes in nature. Ammonia, which is rarely encountered in any considerable quantity, did not give rise to a modification and not even to an avoiding reaction. Again the darters are swift stream fishes, depending upon mechanical conditions to main- tain themselves in a suitable environment. They rarely encoun- ter carbon dioxide or low oxygen and failed to react to them at all quickly, although they were affected by both. Advan- tageous reactions appear to be confined to stimuli commonly encountered in the normal life of the animal. By this we mean merely to imply that whatever the processes of origin and sur- vival may have been in detail (Mathews, '13), there is correlation between the conditions of existence and types of irritability (Henderson, '13). It thus seems probable that associative memory does not necessarily play any role in the process of modification des- cribed. Since the experiments were conducted with a view to eliminate any possible effects of learning, only incidental evidence was acquired. The treatment of the stock of fishes was as fol- lows: There were a number of individuals of each species in the aquaria, and individuals were drawn at random for each experi- ment. A given series of experiments were run and the fishes were returned to the aquaria and not used for experiments until several days had elapsed. In nearly all cases the same fishes were used only by chance. When it was necessary to repeat an experiment on the same day, different fishes were used except in one or two cases not included here. Apparently, in experiments thus conducted, evidence of reten- 28 VICTOR E. SHELFORD AND W. C. ALLEE tion can be of two kinds. These are indicated by the following questions: After entering the modified water a number of times and turning back, did the fishes often turn back before reaching the gradient ? When transferred to the control after exposure to the gradient did they turn back, or show a preference for either end of the control ? The first type of behavior could be shown only where the grad- ient was confined to the central third of the tanks. Since some species turn in any part of the tanks it is necessary to select a particular species which does not show this trait. Hybopsis was our best example of this for normally they went back and forth symmetrically in the controls. The graphs indicate that in the boiled water experiments this species turned back often before the gradient was reached, which raises the question as to whether they associate the center drain or difference in lighting with stimulating water ahead. Fishes are able to form associations (Mobius fide Holmes, '11 and many others). There appear to be two ways in which associations formed in the experimental tanks could be carried over to the control tanks. One is through the kinaesthetic sense, the other through differences in illumination of the sides of the body when approaching the drain from the different direc- tions. The data which we have upon this question comes from comparing the responses in the first and second halves of the double experiments described on page 6. It is obvious that if the fishes were able to associate the dif- ferences in illumination upon the sides of the body when ap- proaching the drain, with increasing stimulation further on in the gradient and were able to retain this after being dipped out of the water with a net and placed for a short time in very different surroundings, they should show an apparent preference for the end of the control tank opposite that in which they spent most time in the experiment, (cf. Fig. 1). An examina- tion of the records of twenty-eight controls indicates that the fishes usually showed some difference in their relations to the two ends of the control tanks, but in these exchanged controls less than half of the fishes show an apparent preference for the end that should have been favored if the light was depended upon. This is as it might be if the kinaesthetic sense were depended upon for the reaction, but the number of experiments MODIFICATION OF THE BEHAVIOR OF FISHES 29 is too small to make the slight difference more than a chance one. The comparison of the reactions in the experiment with those of the succeeding controls only serve to emphasize former findings that the solute is the chief guide in the reactions of the fishes. Special investigation would be necessary to determine whether or not associative memory plays any role, but if so, it might be due to an association of increasing stimulation with stronger stimulation further on in the gradient. There is a more or less distinct rhythm of the reactions in the gradients and the view of associative memory applied to this phase of the reaction would call for a make and break of associations so rapid as to cast doubt upon this being the entire explanation. And, since the increased sensibility due to repeated stimulation is explain- able otherwise, the assumption of learning, as applied to this as- pect of behavior, is as unnecessary as it is questionable. Still, while we have thus separated the type of modification here des- cribed from associative memory by the use of the usual criteria, we do not mean to imply that the processes involved are neces- sarily fundamentally different. IV. ACKNOWLEDGMENTS AND BIBLIOGRAPHY The authors are indebted to Dr. Harvey A. Carr, Dr. Tashiro and to Mr. M. M. Wells for suggestions during the preparation of the manuscript. BIBLIOGRAPHY For further citations, see Shelford and Allee, 1913. Araki, T. Ueber die Bildung von Milchsaure und Glycose im Organismus bei 1891. Sauerstoffsmangel. Zeit. f. Physiol. Chem., XV, p. 335. Baglioni, S. Zur Vergleichenden Physiologie der Atembewegungen der Wirbel- 1910. tiere. Ergeb. der Physiol, Vol. 9, pp. 90-137; 1911, Vol. 11, pp. 526-597. Bethe, A. Algemeine Anatomie und Physiologie des Nervensystems. Leipzig. 1903. Cushny, A. R. Pharmacology and Therapeutics, 10th edition, pp. 498-501 Phila- 1910. delphia. Edinger, L. Haben die Fische ein Gedachtniss? Allg. Zeiiung. (Trans in Smith- 1899. sonian Rep., 1899, p. 375.) Fedman, I. and Hill, L. The influence of oxygen inhalation on the lactic acid 1911. produced during hard work. Jour. Phys., Vol. 42, pp. 439-443. Fletcher, W. M. The survival respiration of muscle. Jour. Physiol., Vol. 23, 1898. pp. 10-99. Hopkins, R. and Dennis, N. Interrelation of the Ammonia and Carbon dioxide 1911. Content of the Blood. Jour. Bio. Chem., Vol. 10, pp. 407-415. Holmes, S. J. Evolution of Animal Intelligence. New York. 1911. 30 VICTOR E. SHELFORD AND W. C. ALLEE Henderson, L. J. The Fitness of the Environment. New York. Hill, L., (ed.) and Others. Recent Advances in Physiology and Biochemistry. 1906. London. 1909. Further Advances in Physiology. London. Kuiper, T. Untersuchungen uber die Atmund der Teleostier. Pfliigers Arch., 1907. Vol. 117, pp. 1-107. Mast, S. O. Light and the Behavior of Organisms. New York. 1911. Mathews, A. P. Adaptation from the Point of View of the Physiologist. Amer. 1913. Nat., Vol. 47, pp. 90-104. Quagliariello, G. Influenza delle injezione endovenose di acido cloridrico sulla 1911. respirazione. Arch, di Fisiol, Vol. 9, pp. 477-484 (fide Baglioni '11.) Reuss, H. Die Wirkung der Kohlensaure auf Atmung der niederen Wirbeltiere, 1910. in besonderen der Fische. Zeit f. Biol., Vol. 53, pp. 555-587. Sheldon, R. E. The Reactions of Dogfishes to Chemical Stimuli. Jour. Comp. 1909. Neurol., Vol. 19, pp. 273-331. Shelford, V. E. and Allee, W. C. The Reactions of Fishes to Gradients of dis- 1913. solved atmospheric Gases. Jour. Exp. Zool., Vol. 14, pp. 208-266. Signorelli, E. Influence de l'acide lactique sur la fonction du centre respiratoire. 1911. Arch. ital. d. Biol, Vol. 55, pp. 119-128 (fide Baglioni '11). Westerlund, A. Studien uber die Atembewegungen der Karausche mit beson- 1906. derer Ruchsicht auf den verscheidenen Gasgehalt der Atemwasser. Skandinav. Arch. f. Physiol., 1906, pp. 263-280 (fide Baglioni '10). Waller, A. D. On the Influence of Reagents on the Electrical Excitability of 1896. Isolated Nerve. Brain, Vol. 19, pp. 44-67. Wells, M. M. The Resistance of Fishes to different Concentrations and Combi- 1913. nations of Oxygen and Carbon dioxide. Biol. Bull., XXV pp. 323-347. Winterstein, H. Die Regulierung der Atmung durch das Blut. Pfliigers Arch., 1911. Vol. 138, pp. 167-184. MODIFICATION OF THE BEHAVIOR OF LAND ANI- MALS BY CONTACT WITH AIR OF HIGH EVAPORATING POWER VICTOR E. SHELFORD Hull Zoological Laboratory, University of Chicago I. INTRODUCTION The rapid modification noted in the case of fishes, by the author and Dr. W. C. Allee is likewise shown by various land Amphi- bians and Arthropods, used in experiments designed to test the sensibility of different terrestrial animals to variations in evaporating power of air. Some animals of supposedly lower organization than the fishes, showed modifications similar in OH ri T"rJ Ot~ PT* II. MATERIAL AND METHOD The following species were studied: the yellow margined milliped (Fontaria corrugate Wood), ground beetles (two species of Pterostichus) , the wood frog {Rana sylvatica LeC), the red backed salamander (Plethodon cinereus Gr.) — all fro'm moist forest habitats; and the common toad (Bufo lentiginosa), the small digger wasp (Microbembex monodonta Say), the bronze tiger beetle (Cicindela lecontei Hald), and the sand spiders (Geolycosa wrighti Em and pikei Marx) — all from dry sand dunes. The animals were put into small cages across which air was forced through three narrow slits. The ground plan of the cages is indicated in Fig. 1. The covers of the cages were of glass, the fronts opposite the slits of screen. Gradients of evaporating power were secured by passing air of different relative humidities, or different temperatures, or by passing it at different velocities across the different thirds. Thfe device for thus controlling the rate of evaporation was designed by the writer and Prof. E. O. Deere of Bethany College. The statistical and environmental aspects of the one hundred experiments performed together with details of the methods were published elsewhere (Biol. Bull., June, 1913). Nearly seventy-five of the experiments were of such a character as to bring out the modification phenomenon. The tracings of the movements were drawn to a minute and 32 VICTOR E. SHELFORD second scale at the time of the observation (Fig. 1) and those presented here are original drafts. The different individuals could not be distinguished in a few cases, but since this error entered into all the experiments the results are constant for this piece of work. (For further discussion of this point, see Shelf or d and Allee, '14, p. 8). III. EXPERIMENTAL RESULTS To illustrate the method of obtaining and recording the data used further on, we present three charts. In Chart I, Expt. 71, the reaction of the red-backed salamander to increased evapo- ration in the right hand third of the experimental tank, is shown. Figure I Showing the ground plan of the experimental cages in their relative positions, the hood which covered and separated them is not indicated. E is the experimental cage; W, the section used for wet air; M, for the air supplied directly from the pump; D (dry), H (warm), and R (rapid flow), stand under the section where the highest rate of evaporation was maintained. The crosses indicate the positions of the 1 c.p. lights; the arrows the direction of the flow of air. The screen portions of the cage are represented by the broken lines. C, is the control cage, similar to the experimental in every way except the kind of air supplied. Below this is shown the control record of an experiment during the first three minutes. The ruling of the paper used corresponded to half minutes and the figures were written in at the center. The graph is about two-thirds natural size and the cages about one- seventh. The increased evaporation is due to a rapid flow of ordinary air. The salamanders tried the region of highest evaporation re- peatedly during the first fifteen minutes and then began to turn back when the rapid flow was encountered. We note that one individual turned back the first time it encountered the rapid flow. The control individuals came to rest practically where they were placed and moved only a little throughout the experiment. MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 33 The toads gave the same reaction to similar conditions. They tested the region of high evaporation repeatedly during the first ten minutes and then began to turn back and make short stays in that section. The control is a little less symmetrical than in most cases; for a more symmetrical control, see A, Chart IV, Expt. 42. Both species show a preference for low evaporation at the end of the period of experimentation. Chart II, Expt. 75, shows rapid modification of the behavior of the millipeds. (Fontaria). In the experiment, definite turnings began at the end of nine minutes, when each animal had entered the section of high evaporation. The control is a characteristic symmetrical graph for the species. The graph of the spiders (Geolycosa) is difficult to interpret and is peculiar, due to the fact that the spiders are very quick and pugnacious, so that it is hardly possible for three of them to be in the same third of the cage at the same time. Still, even this graph appears to indicate what is true of single individuals namely, an avoidance of the moist and medium air. This avoid- ance does not begin until the end of eight minutes. The control is typical of the species and shows fairly symmetrical distribution of the spiders. Chart III shows the reaction of the wood frog to gradients of evaporating power produced in three different ways. The graph of Experiments 60 and 70 shows the type which results from acceleration of movement or of movement due to mere stimulation. In Experiment 60, one of the frogs was placed in each third. Their commonest reaction to evaporation is to crouch close to the substratum. If, however, the evaporation continues, they finally hop, apparently at random. In the case of two of the frogs in Expt. 60, the hopping was in the direction of the lowest rate of evaporation where the frogs remained for an hour or more after the observations here graphed were ended. Experiment 70. shows first the stimulation of one frog leading to a series of hops in the direction of lower evaporating power, of another frog a little later in the direction of higher evaporating power which in turn led to stimulation, resulting in hops in the opposite direction. These graphs do not indicate modification. When compared with the next graph, where temperature was used to increase evaporating power, we note a striking difference (com- pare the graphs of Expts. 60 and 70 on the one hand and 73), 34 VICTOR E. SHELFORD Chart I Showing the modification of the behavior of the red backed salamander (Ple- thodon) and toad (Bufo) through a gradient of evaporation due to rapid movement of ordinary air in one section, the other two sections being supplied with respec- tively moist and ordinary air at the standard rate of flow (12 liters per minute, a velocity of .08 meters per second). The rapid flow was approximately .65 meters per second, or about 8 times the standard. The temperature is given in degrees centigrade. Immediately above the three columns of the ruled paper between the time scales is given the evaporation from the Livingston evaporimeters in hundredths of cc. during a period of 20 minutes immediately before and after the experiment. Distance from right to left represents the movements of the animals in the cage (lengthwise) and time is indicated by vertical distance as measured by the scales. The broken line is used where two or three animals move together. In the controls the large S's indicate that still air was used in the three sections. MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 35 CHART I Plethodon — Am Current 24°C Bufo— Air Current— 24°C Experiment 71 Control Experiment 68 Control 36 VICTOR E. SHELFORD Chart II Showing modification of the behavior of millipeds (Fontaria) and spiders (Geoly- cosa) through contact with an evaporation and temperature gradient. The lower temperatures given are for the left and center sections and the higher of the right hand one. For further explanation see Chart I. MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 37 CHART II Fontaria — Am Temperature 24° and 29°C Experiment 75 Control Geolycosa — Air Temperature 23° and 28°C Experiment 79 Control 23°C 38 VICTOR E. SHELFORD which is, I believe, the difference between a graph in which modi- fication of at least one individual is indicated and one in which there is little or no evidence of modification. In Expt. 73, clearly one individual tried air of highest power, then turned back with the second trial, and repeated the same after a brief stay in the moist air. The others appear to have avoided the middle section except for one trial of the air of high evaporating power, after which all came to rest in the moist air. The control of 73, when compared with the rest of the controls indicates the greater activity which commonly occurs in the air that is moving enough to raise the evaporation above the optimum for the animals. Evidence of modification, as in the case of the fishes (Shelf or d and Allee, '14), may be of the following types: (a) an animal may begin turning back after entering the stimulating air a number of times, (b) it may spend shorter and shorter periods of time in the modified air with each entrance; (c) after entering the grad- ient and turning back in it, an animal may begin to turn back before the change of air conditions is encountered, indicating retention; (d) after having experienced differences in one tank it may remain in one end and turn back from the other when there is no difference between the two ends. The second two types of modification may indicate learning while the first two do not. Table 1 gives the data on these questions by species and factors. The kinds of experiments in which there was no evidence of modi- fication, are omitted. For example, the experiments with Fon- taria and wind gave no good evidence of modification and only experiments with reactions to dry and heated air are included. Where the reactions were strongly negative the animals some- times turned back the first time they encountered the stimulating air (Chart I, Expt. 71; Chart III, Expt. 73). All of the species, except Pterostichus which was used for only one experiment, did this when the. evaporation was great. Such turnings indicate that the animals sense the strong stimulii the first time they en- counter them, i.e., before their sensibility has been increased by repeated contacts with the stimulating air. We note the figures showing the number of times this happened, in the twelfth column of the table. Usually the animals showing modification entered the stimulating air a number of times before beginning MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 39 CM in "3 H c o s[Oi}uo3 jo joqmnfvj CO CM CM O CM CM CM *-> i-h ; tC -*CM ti 5 ^J- i-l juaipBjS oj Suipuod CM ■* -S3JJ03 3uu3juno3 -r cm cm 00 -H io -ua 3Joj3q psujnj. 4J CO juaipBJS 3uua}uno3 ^TCMCM 00 CM ■ CMC-] ■ - co lO co i— < c o u -us aaojaq psujnj, CM CM CO i—i eg i— I « 3UIIJ jsjy 3qi paujnx 05O»H c > CM -CM • ■ • ^ .... j M ■CM 4. C > CM rH ,-H .H-D o e 1) c -~ }S3u3iq ui 3iuix t-H i— 1 CI —( CM CM ! a t e ) g&£8 UOIJBJOdBA3 -*l-0C0 a ) 00 00 O to CM CO if cot> t£ c- to to o t~t>t^ c cotooocoooco - rtCM ^ e 1 CT) CO p-1 JS3MOI UI 3U1IX i-H LO ~* CI ) CO-H CM CM ; c CMW P ct CV CI 1 rH CM ! cd 01 uoijBoyipoui 3JOj3q LOCMtO - < oocootooo c CM CO t^ If ) CO CJ> sjeu} jo jaquinu 33ej3av i-HCO-H CI ] i-i CM-HCM--HCM ! CN n CM^H C! Cl 1 CM r- ca 03 UO!}B3UipOUI SUIMOUS OOtOO tocoo ■si O O CO o o o oocoooo IN OC COCO 00 00 X ) O CO ) O 01 i SIBnptAipUI JO }U33 J3J .—I ci J i-l i— » i— 1 i— * rH K cc CI 1—1 CO t> s[BnpiAipui jo jsquinN 00 CO CO c 4- to to to CO CO CO > 1 C CM to a: ] CO ■ CO tO sjusuiuadxs jo jaquin^ tD CM CM : CMCMCM — ih ■ l> TfCM t£ 4- J 4- i 00 • CM to O LO 33U3J3JJip 3jnjEJ3dUI3X ■ -o ■ ■ 1—1 in ■<* 3jri}BJ3duiax COM CM CM CM CM CM CO CO CO CO CO CM CM CM CM CM CM CM CM CM CM CN Cv u — U w muQ 3 rt , c ^ > c ■» CO—iOO CO— 'O LOLO £ cote OC u- X LfU^LO [> o to t> t> ■ F t>t> c^ c 1— < CN CM o a a CO CO CO CO- • 4^ W CM CM CM CMCMCM ^- , CM CM C\ c\ J o -o 5 o -ooo ■ S oo CI r c c Cf J £ : C 4J re . . a. — "co ■ c c > « K o ■ « [C n n I4H co c £ bi o 2 5 C ) Is 3 > CO a hi en cu tn ) 01 13 s - > 5J Ui 0) > k4 CU > 5 < H S < .S-a 6 CO ' < Z < | ! 1 1 I I I I I I I I I I I I I I I I I I I I I II |J I I I I I I I I 1 1 I I I I I I I I I I I I I I f I | | f I | [ I I I I 1 1 I I I I I I I I I I 'll I I I I I I I I I I I II I I l]l I 1 1 l]l | I I I II | | I I I I I | | I I I I I | I I I I I | I I I I I | I I I I I | | | I I I I 1 | |l I I I I |l I I I I |l I I I I ; I I II I | I I I I I | I I I I I I I I I I I I I I I I I I I I I I II I u O CM z w 05 K Z P H U — 1 05 i < -1 a < j O < > E \A CO X < W < O o CM CM 05 < 05 Q o < > >< < z < 42 VICTOR E. SHELFORD to turn back. In the ninth column, we note that the average number of such trials varies from 1 to 3.2 It will be noted, also, in the eighth column that the percent of the individuals tried, which showed modification, varied from 33 to 100. In most of the cases where no modification was indicated, the animals reacted to the higher rate of evaporation from the first or remained in* the air of low evaporating power without encountering the higher; hence they do not indicate that the capacity for modifi- cation is not present under favorable conditions. It will be further noted from the tenth and eleventh columns that the length of time spent in the air of low evaporating power is nearly always greater than the time spent in the highest evaporating power (the time spent in the central third is omitted and is usually small). The experiments were not planned to test the ability of the animals to form associations. In the main, the work was con- ducted in a manner intended to prevent the formation of asso- ciations. There was a considerable stock of most of the animals and in nearly all cases the individuals used in a given experiment were not used again for some days. Still, after the gradient had been encountered a number of times, turning before it was encountered may indicate that the stimulation by the air was associated with the approach to the light or with the screen- covered side on the right (Fig. 1). In the case of Plethodon cinereus there were ten more turnings before the gradient was encountered than in corresponding positions in controls and with the exception of Pterostichus, the number of turnings in the experiment before entering the gradient exceeded that for the corresponding positions in the controls. To test the possibility of association formation, four readings (A, B, C and D, Experiments 42 and 43, Chart IV) of the behavior of three toads were made. "A," (Chart IV) is a typical symmetrical control which was observed first. The same toads were then transferred to the experimental cage, and reading begun within five minutes. Here they began showing some avoidance of the air of the high evaporating power at the end of six to eight minutes. At the end of a twenty-minute observation, they were removed for a few (less than five) moments and then returned to the same cage (C) where they showed avoidance of the dry air in one and two minutes. At the end of the twenty- MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 43 minute period (C), the toads were transferred (as before) to the control cage. After a few moments they began to show a preference for the same end that was preferred in the preceding period. While controls sometimes show asymmetry, this is the only one out of ten which shows any such peculiarity. In the table, the items marked B, C, D, under Bufo, repre- sent two experiments and the control which followed them. In the first experiment (Column 9) the number of trials before turning back was 2.6 and the time over 200 seconds in the highest and lowest evaporation. In the second experiment the trials were reduced to 1.0 and the time to a little more than 80 seconds. In the control observation which followed, they showed a prefer- ence for the end corresponding to the moist air with less trials and with less time than was required in the first experiment. Toads have no tendency to come to rest in one end, or in contact with the walls. The two sides of the cage were not exactly alike and the toads could perhaps sense the screen on their right or the lights in front when approaching the end in which the high evaporation occurred in the experiments. The one experi- ment suggests that the toads may have associated the lights ahead or the view through the screen with the dry air further on (see Fig. 1). IV. INTERPRETATION OF RESULTS Protoplasm and the plasmas of organisms possess a definite mechanism for maintaining approximate neutrality. ' Neu- trality is quite as definite, quite as fundamental and quite as important a characteristic of the organism as its temperature, or osmotic pressure, or in fact anything else we know" (Hender- son, '13a). 'Within wide limits of amount any acid or base may be poured into the organism and the reaction will not vary," nor will it vary if such be produced by the organism. In the preceding paper we noted that presenting acid either externally or internally produced increased sensibility or in- creased irritability. By this we do not mean to suggest that any particular degree of acidity was probably attained except where staggering occurred, but rather that a tendency toward acidity occurred during the adjustment of the neutrality mech- anism. Just what the relation of the ammonia to the neutrality mechanism is in the case of fishes and other vertebrates is not 44 VICTOR E. SHELFORD Chart IV Showing modification of the behavior of the toads (Bufo) as indicated by experi- ments following each other closely and by the actions of the toads under control conditions immediately after the second experimental reading was' ended. MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 45 CHART IV Bufo Dry Air— Possible Retention of Modification A B C D Control 23°C 42 Experiment 23°C Experiment 23°C 43 Control 23°C J20^ 46 VICTOR E. SHELFORD clear, for while it doubtless causes a tendency to alkalinity, its paralyzing effects are probably due to specific action upon the protoplasm. It should be noted that alkalies, including am- monia, as well as acids, usually excite animals and are almost as potent as acids in causing reversal of reaction (Loeb, '06, Mast, '11), increased irritability, etc. It appears then that a disturb- ance of neutrality in either the direction of acidity or of alkalinity may be expected to give such results as we have discovered in connection with the reactions of fishes to gradients. The nega- tive reactions here described are to conditions which can prim- arily either dilute or concentrate the plasma either in the peri- pheral sense organs or in the organism as a whole. Did this result in acidity or in alkalinity or did it disturb some other regulatory mechanism ? While a probable answer to this ques- tion could be presented and discussed, we reserve it until special investigation, now about to be undertaken, may have been com- pleted. It is, however, more than probable that the increased sensibility to air of high evaporating power is due to some dis- turbance of neutrality, brought about by concentration. We may note, also, that if associations are formed, they go hand in hand with and can hardly be distinguished from the other type of modification. In other words, there is no reason to assume that associative memory is essentially different or stands apart from the type of modification here described. In fact, it appears that since all excitation and all activity increase the out- put of carbon dioxide the neutrality mechanism may be involved in all associative processes. Henderson ('13a) says it is gradually becoming clear that all physico-chemical conditions in proto- plasm — alkalinity, osmotic pressure, colloidal swelling, chemical equilibrium, temperature — are interdependent and that car- bonic acid and acid base equilibrium are among all these things probably the most important variable. Thus it seems probable that many of the simpler problems of associative memory must be referred to the biochemist for solution. The zoologist and the psychologist can by their present methods do little more than describe the phenomena of modification and the best results can come only from general cooperation. The need for cooperation between the psychologist and the naturalist is even more evident than between psycho- logist and biochemist. MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 47 In all of the cases here discussed, except the reactions of the wood frog to dry and moving air, and in most of the other ex- periments, rapid modification took place. All the stimuli used are those commonly encountered by the animals experimented upon, in their natural environments. It is impossible to deter- mine, however, in the case of animals collected and brought into the laboratory, how much of this ability to avoid disadvantageous stimuli has been acquired during the life of the individual. We note that plants, which respond to environmental changes by variations in growth form, show a similar relation to the various stimuli. Those commonly encountered in nature usually call forth advantageous responses, while those not usually encountered call forth responses which are often decidedly disadvantageous. Here this quality could not have been acquired by the individual in its life experiences and is accordingly to be considered as an innate capacity. It is, therefore, a fairly safe assumption that the capacity to avoid disadvantageous stimuli, and the capacity for advantageous modification, are innate also. Such capacities appear to be common. The difference in the speed with which frogs learn to avoid distasteful food as compared with the time required for them to learn a maze or the presence of a glass plate is a further example of the difference between reactions to natural and unnatural stimuli and problems. As a further example, we note the difference between the apparent "stupidity" of the frog in failing to learn not to snap at a fly when the punish- ment was the pricking of the upper jaw by a pin or needle (Holmes, '11) as compared with the rapid learning to discriminate between the kinds of food presented when certain kinds brought punishment through the taste organs (Schaeffer, '11), the more usual channel for disturbing results of food taking. The raccoon is slow with a man's problem, such as a latch and door problem ; what could he do with a forest and hollow tree problem the first time it was presented to him ? What could a man do with a forest and hollow tree problem ? That with some such problems he errs widely and loses his way often, we are well aware. The use of problems which the animals concerned could never by any possibility have encountered is legitimate, and indeed a necessity if the effect of experience is to be eliminated in animals not bred for experimental purposes. Still, is not the degree of intelli- gence shown by an animal, with its first problem of a kind 48 VICTOR E. SHELFORD common in the experience of the species, a better test of its intelligence in terms of what is so named in our own species, than its ability to solve an entirely unnatural problem ? Why should the intelligence of a monkey be indicated any more by its ability to operate locks, than a man's, by his ability to balance himself upon the swaying branches of a tree ? The comparisons may seem crude, indeed they are so, but the matter in point seems well borne out by the evidence suggested above. The error in rating the intelligence of animals solely, either upon the basis of problems which they normally encounter in nature, or those never encountered is not small. Indeed, experimental psychologists have, to a considerable degree, abandoned at- tempts to rate the intelligence of animals. It is still, however, a common practice among zoologists. To understand the be- havior of his animals to the best advantage, the experimental psychologist, should have first-hand naturalistic knowledge of them. The naturalist is even more in need of a knowledge of experimental methods and results. It appears to one not prim- arily interested in either, that the experimental psychologist should be a naturalist and the naturalist an experimental psy- chologist. Theoretically, the explanation of the apparent difference between the relations of animals to natural and unnatural problems is very simple. Direct evidence of recent years tends to show that natural selection does not usually, at least, operate upon structural characters in the manner formerly supposed. On the other hand, a rapidly increasingly body of experimental data shows that animals survive or perish under severe conditions in accordance with their irritability and other physiological characters. Irritability is the chief mechanism of survival. Accordingly, where natural selection has been able to operate, responses to stimuli are commonly advantageous; where it has not, they are often disadvantageous and sometimes clearly detrimental. Two other points are noteworthy. The sensitiveness of the animals used, to slight differences in rate of evaporation shown, makes errors in laboratory experiments easily possible. Air current in controls were practically abandoned because some of the amphibians detected differences in the different parts of the control cage which could not be detected with the hand, making MODIFICATION OF THE BEHAVIOR OF LAND ANIMALS 49 great care necessary in the preparation of each current control. The modification of behavior in gradients here discussed, has been made evident solely by the graphic method of recording the movements of the animals. It is quite possible that appli- cation of similar methods to reactions to gradients generally, may show that such modification is the rule, from the more simply organized animals upward. ACKNOWLEDGMENTS AND BIBLIOGRAPHY The writer is indebted to Dr. Harvey A. Carr and Mr. M. M. Wells for suggestions during the preparation of the manuscript. All other literature consulted is cited in Shelf ord and Allee, '14. Holmes, S. J. Evolution of Animal Intelligence. Holt. 1911. Schaeffer, A. A. Habit Formation in Frogs. Jour. Animal Behavior, Vol. I, pp. 1911. 309-335. Shelford, V. E. and Allee, W. C. The Reactions of Fishes to Gradients of Dis- 1913. solved Atmospheric Gases. Jour. Expt. Zool, Vol. XIV, pp. 207-266. Shelford, V. E. and Allee, W. C. Rapid modification of the Behavior of Fishes 1914. by Contact with Abnormal Water. Jour. Animal Behavior, Vol. IV, pp. 1-30. Shelford, V. E. Reaction of Certain Animals to Gradients of Evaporating Power 1913. of Air, with a Method of the Establishment of Evaporation Grad- ients by V. E. Shelford and E. O. Deere. Biol. Bull. XXV, 79-120. Henderson, L. A. The Regulation of Neutrality in the Animal Body. Science 1913a. N. S., Vol. XXXVII, pp. 389-395. 1913. The Fitness of the Environment. New York. Loeb, J. Dynamics of Living Matter. 1906. Mast, S. O. Light and the Behavior of Organisms. New York. 1911. A GRAPHIC METHOD OF RECORDING MAZE-REACTIONS ROBERT M. YERKES AND CHESTER E. KELLOGG From the Harvard Psychological Laboratory One figure For nearly twenty years the maze or labyrinth has been em- ployed by students of animal behavior. Both apparatus and procedure have been improved steadily during the last decade, but even to-day we lack an intelligently standardized form of maze and a reliable method of recording the several important aspects of the subject's reaction. We propose, in this paper, to describe a method of recording maze-reactions which should greatly increase the value of the results obtained in experiments with the maze. We shall not attempt to describe a type of maze which promises to be worthy of standardization, but instead shall limit ourselves to a brief discussion of methods of observation. Experiments with the maze offer opportunities for the measur- ing of several aspects of reaction. Especially important among the data obtainable are (1) time of reaction; (2) distance; (3) number of errors; and (4) nature and distribution of errors. Prior to the devising of the method herein described, it has been prac- tically impossible for even the highly practiced observer to obtain accurate measurements of all of these features of reaction. Indeed, with a rapidly moving subject like a rat or a mouse, it has been impossible, during the first few trials, to obtain with accuracy any other measurement than that of time. This is obviously quite as unnecessary as it is unfortunate, for we have good reason to believe that distance and error data are in many experiments more important than time data. Because of our conviction that a variety of data should be obtained in every maze experiment and that all measurements should be made with a reasonable degree of facility and accuracy, we have made it our business to attempt to devise a method which shall enable an experimenter to record the various aspects of reaction directly and graphically. 50 GRAPHIC xMETHOD OF RECORDING MAZE-REACTIONS 51 When the idea of using a graphic method of recording maze- reactions occurred to us, it was immediately suggested, through correspondence, to Professors J. B. Watson and Madison Bentley. The former, feeling the immediate need of such an improvement in the technique of maze experiments, promptly devised and con- // H D IL Mi IV IL Figure 1 Diagrams of four types of apparatus for obtaining graphic records of maze-reactions. I. Apparatus for the direct method: Z, maze; L, lens; D, drawing surface; H, hood. II. Apparatus for simple reflection method: M, mirror. III. Apparatus for double reflection method of Watson: M and M 2 , mirrors. IV. Apparatus for double reflection method of Kellogg. 52 ROBERT M. YERKES AND CHESTER E. KELLOGG structed what may be termed the camera lucida apparatus. This has been in use for several months in the Psychological laboratory of the Johns Hopkins University and is reported by Professor Watson to work satisfactorily. In our search for simple, inexpensive, reasonably convenient and adaptable means of obtaining the desired data of reaction, we have happened upon the four methods or devices which are now to be described. Figure 1 is a diagrammatic representation of these several devices. Since they are not of precisely the same value, we shall point out the chief merits of each in des- cribing them. It was our aim to project, in some convenient manner, an image of the maze and of the reacting subject upon a record sheet which should bear a diagram of the maze. Upon this record sheet the experimenter may accurately trace the path of the ani- mal, while, at the same time, keeping a record of the time of reaction. From the graphic record of the route taken by the animal, the distance and error data may be read. We shall designate the four methods as the direct method (Fig. 1, I); the simple reflection method (Fig. 1, II); the double reflection method of Watson (Fig. 1, III) ; and the double reflection method of Kellogg (Fig. 1, IV). I. The direct method. This is the simplest device which we have been able to imagine. Above the maze, Z, of Fig. 1, I, is placed either a simple or a compound lens, L, and at the proper distance above it, a plate of glass, D, conveniently framed in a drawing table, and enclosed by a hood, H. Upon this plate of glass, a thin sheet of paper bearing a plan of the maze is placed. As the observer looks down upon the record sheet, he sees an image of the maze and of the reacting subject, and at will he may trace with pen or pencil upon the record sheet the path followed by the subject. This method has the important advantages of being extremely simple, inexpensive, and adaptable. It gives a reversed image of the maze, but this is no considerable disadvantage. The chief disadvantages of the method are its inconvenience in connection with large mazes because of the great distance nec- essary between maze and drawing board. But even with very large mazes, the method may be used to advantage if a vertical distance of twenty to thirty feet is available. This arrangement GRAPHIC METHOD OF RECORDING MAZE-REACTIONS 53 is likely to prove most practicable where two rooms, the one above the other, are available for maze experiments. The writers have tested the method only with very small mazes. II. The single reflection method. The device for this method, as shown in Figure 1, II, consists of a mirror, M, which is placed above the maze, Z, and which causes an image of the maze to fall upon the lens, L. This image is focused upon a record sheet at D. As in the case of the direct method, the drawing board is hooded in order that the experimenter may work in dim light and thus be able to see, clearly, both the alleys of the maze and the moving animal. In comparison with the former method, this is somewhat more expensive. It yields a completely re- versed image and it may be used for even very large mazes. Its chief defects are the inconvenient inclination of the drawing surface, at one end of which the observer must work. In this laboratory we have thoroughly tested the method and find it to work satisfactorily. A little practice enables the observer to follow a rapidly moving rat or mouse and to obtain records which yield accurate measurements of distance, time and errors, even in the early experiments with a given subject. III. The double reflection method of Watson (camera lucida method). Two mirrors are used in this apparatus together with a lens and drawing board. The arrangement of these parts is shown in Fig. 1, III. This apparatus has the disadvantage of being more expensive by reason of an additional mirror than the preceding method, and it is also placed at a slight disadvantage because it supplies an image of the maze which is reversed from right to left. To counterbalance these disadvantages, we may mention the following obvious advantages: (1) the more con- venient position of the drawing surface; (2) the removal of the experimenter to a considerable distance from the maze; and (3) the adaptability of the apparatus to spatial demands in room or laboratory. A more detailed account of this method is given by Professor Watson on p. 58. IV. The double reflection method of Kellogg. This differs from method III in that M 2 is placed below the lens, and the image falls upon the record sheet from below, as in method I. Disturbing shadows cast by the hand of the experimenter are thus avoided. The image obtained by this method is completely 54 ROBERT M. YERKES AND CHESTER E. KELLOGG reversed and the apparatus, as a whole, is quite as adaptable as is Professor Watson's. General discussion of methods. The above devices for obtaining graphic records of maze-reactions yield less satisfactory results than would a good photographic device, and we recommend them simply because they are less expensive in construction and operation. All are so simple that detailed description is needless. We shall, however, in concluding this article, call attention to certain important points which experience in the use of the graphic method has brought to our attention. In the first place, although it is perfectly possible to get along with a simple lens, especially if one is working with small mazes, a much more satisfactory image may be obtained by the use of a compound lens. Second-hand portrait lenses are available and wholly suitable, but even such a lens is likely to be much more expensive than a simple lens. Each of the four devices which we have described has its obvious advantages and disadvantages, and it is clear that choice of a method should depend upon the relative importance of these in each particular case. On the whole, it is likely to be more convenient for most experimenters to have their drawing board slightly inclined toward them. This is possible in methods III and IV. Method II necessitates the use of an inclined drawing board, but unfortunately the observer must sit at one end of this board and work in a somewhat awkward position. So far as the position of the drawing surface is concerned, methods III and IV would seem slightly more satisfactory than methods I and II. In those devices in which the light falls upon the record sheet from above, the shadows cast by the experimenter's hand and pencil are disturbing, sometimes rendering it difficult to follow accurately a swiftly moving animal. Other things being equal, it is therefore preferable to have the light come from beneath the drawing surface, as in methods I and IV. In the first few trials with a given animal, it is extremely important for the experimenter to be able to change record sheets quickly, since the animal is likely to traverse the alleys of the maze rapidly and repeatedly. If the image comes from above the record sheet, it is possible to have the sheets made up in the form of a tablet or block with two edges glued. The GRAPHIC METHOD OF RECORDING MAZE-REACTIONS 55 tablet having been properly oriented, the experimenter may at any moment tear off a record sheet and continue his tracing almost uninterruptedly. This method may be made to work satisfactorily even when a printed diagram of the maze appears on each record sheet, for the orientation of the block may be kept constant. When the image falls upon the record sheet from below it is necessary to use rather thin paper and to have the drawing board so arranged that the sheets fit neatly and may be quickly placed in position. Although we have tried only methods I and II in this laboratory, we are inclined to believe that it is more satisfactory on the whole to have the image come from below the record sheet. Especially in the first trials with a given animal, the time required is likely to be long, and the experimenter should be able to make his observations without undue discomfort or fatigue. We recommend that as soon as the experimental . device has been selected and properly adjusted, a zinc etching, which exactly reproduces the image of the maze as it falls upon the drawing board, be made, and that this be used in the prepara- tion of blank record sheets. For, although a diagram of the maze is not absolutely essential, it has considerable value in connection with the early trials and sometimes prevents errors in the reading of records of later trials. It is extremely laborious and wasteful of time to draw the diagrams by hand, and if hundreds or thousands of record sheets are to be used, the cost of a zinc etching and of printing the sheets will be trivial in comparison with the value of the experimenter's time. As appears from the above discussion, we are not in a position to recommend any one of the four methods over the others without careful consideration of the type of maze which is to be used, of the character of the lens, and the location of the ap- paratus. But we are fully convinced that all investigations with the maze should make use of some graphic method of record- ing reactions. There can be no doubt that the data previously obtained from maze experiments have only a fraction of the value which they should have, and that because of the inaccuracy and incompleteness of the records. A CIRCULAR MAZE WITH CAMERA LUCIDA ATTACHMENT 1 JOHN B. WATSON The circular maze shown in Fig. 1 is made with wooden base and aluminium walls. The base is 150 cm. in diameter and 4 cm. in thickness, and is constructed as follows: Two wooden discs 150 cm. in diameter and 2 cm. in thickness are first sawed out. These two discs are finally glued together. Before glueing, however, the upper disc is marked off into a series of concentric circles. The diameter of each of the circles is as follows, be- ginning with the outermost one: 140 cm., 120 cm., 100 cm., 80 cm., 60 cm., 40 cm., and 20 cm. These circles are then sawed out upon a band saw. The width of the saw is so chosen that it is just equal to or slightly larger than the thickness of the alu- minium sheets used for the walls. After sawing, the disc as a whole is converted naturally into a series of concentric rings. These are fastened down to the lower disc with hot glue and screws. The lower surface of the base is thus solid, while the upper surface shows a series of grooves into which the aluminium walls may be slid. Soft aluminium bought in rolls is used for the latter. The height of the aluminium is 18.5 cm., the thickness, .8 mm. The aluminium is unrolled and cut into the proper lengths. Each strip is cut just 10 cm. shorter than the length of the cir- cular groove into which it is to be fitted. This gives an opening into the alley. By means of this arrangement it is possible to slide the aluminium around in its groove and thus to place the entrance in any desired position. Fig. 1 shows clearly the construction of the maze, the number of alleys, the placing of the entrances, and the radial stops. 2 This maze offers several desirable things in work of this character: in the first place it can be used on the unit plan, in that only the home box and the surrounding segments need be ■ From the Psychological Laboratory of The Johns Hopkins University. ' The base as a whole may then be sawed into half or quarter sections for con- venience in shipping or storing. Indeed, it is easier to build the base in half sec- tions. When set up the sections are locked together and placed horizontally upon a wooden framework. The material should be well seasoned. A thorough coat- ing of boiled linseed oil should be applied. 56 CIRCULAR MAZE WITH CAMERA LUCIDA ATTACHMENT 57 used where a very simple maze is desired. The addition of other segments then merely increases the complexity in an, at present, unknown ratio. The coefficient of increasing complexity could be determined by allowing one group of animals to learn the maze in its simplest form, another in its next most complex, etc. Figure 1. General view illustrating camera-lucida maze 58 .JOHN B. WATSON Secondly, the ease with which complications can be intro- duced makes the maze very desirable. This is brought about by the flexibility in the arrangement of the entrances and radial stops. Fig. 1 shows the maze just as it was used by Miss Hubbert in the work which she reports on page 60. The camera lucida attachment is easily installed; it is simple and permanent. Had it not been for Professor Yerkes' insis- tence upon the necessity for having some recording device for the movements of the animals, it is doubtful if this attachment would have been made in its present form. He suggested to me two years ago that some such device would be desirable and that he had certain plans for its construction. Before the publica- tion of his work it became necessary to have an exact record of Figure 2. Chartometer the stages of the acquisition of the maze habit, since Miss Hubbert wished to undertake a comprehensive study of the difference in the acquisition of habits of animals of different ages. Accord- ingly, I went ahead independently and finally constructed the apparatus which is shown in Fig. 1 . A large plate glass mirror (Mj, 91 cm. wide by 121 cm. in length, was placed at an angle of 45° directly over the center of the maze. This mirror was strapped by small clamps to the edge of the supporting framework. At a certain distance from this mirror a second mirror (M), 60 cm. by 75 cm., is placed at an angle of 45 ° above the maze and at such a distance from M x that the light reflected downward from M falls outside of the maze. Below M, and in the path of the light reflected from it, is placed a single achromat (L), 6 cm. in diameter and 50 cm. focus. The lens is placed in a barrel and the barrel is attached to a wooden disc 30 cm. in diameter. This board is attached CIRCULAR MAZE WITH CAMERA LUCIDA ATTACHMENT 59 to an iron collar which slides freely up and down the rod CR. This gives a very easy means of adjusting the size of the image, focusing, etc. Below this first disc will be found a second disc similar in character and controlled in the same way. A pad of circular paper is laid upon this disc. 3 A reduced image (IM) of the maze appears upon this paper. Extraneous light is excluded by means of a soft black flannel curtain attached to the disc which supports the lens (L). As may readily be seen from the figure, the maze must be illuminated quite highly in order to produce a clear image. The illumination is obtained by means of six lights placed symmetrically around the maze and by one light in the center of the maze. The six lights on the periphery are supplied with opaque half shades, the light in the center of the maze with a similar opaque circular shade. These shades are of aluminium, blackened on the upper surface. The floor of the maze is covered with imported white linoleum. This serves to reflect the light upward to M,, thence to M. Passing through the lens the rays are brought to a focus at IM. The ratio between the maze and the image is 6.4 to 1. The image appearing at IM is extremely clear when proper precautions are used to sensitize the eye. Even the smallest mouse can be seen quite clearly. The movements of the animal are traced upon white paper with a soft pencil. In the early stages of learning several sheets of paper are used on each animal at any given trial in order to avoid a too complicated tracing. Each sheet is marked with the number of the animal, the number of the trial and the serial number of the tracing. The length of the lines so traced is meas- ured by means of a chartometer furnished by Eugene Dietzgen and Co. Keuffel and Esser furnish a similar and somewhat better instrument. This instrument is surprisingly accurate even in measuring lines which are tortuous in their course. The error in measuring the length of the charted lines is about one per cent. Fig. 2 shows the chartometer actually employed. 3 It is convenient to cut out several sheets upon a disc cutter and to stamp a hole 13 mm. in diameter in the center of each for the reception of a stud 13 mm. in diameter and 1 cm. height placed in the center of the board IM. TIME VERSUS DISTANCE IN LEARNING 1 HELEN B. HUBBERT The present investigation is concerned with the factors of the total time consumed and the total distance run in the learning of the maze by rats. The maze used in obtaining the records presented below was designed by Professor Watson. He des- cribes the maze elsewhere in this issue of the Journal (p. 56). The records were taken in terms of the time consumed in running from the point of entrance to the food box, and of the total distance traversed during this time. Timing was done by means of a continuous stop-watch registering one-fifth seconds. The watch was started the moment the animal left the starting box (S. B., Fig. 1, p. 57) and was stopped at the moment of entrance into the food box in the central compartment of the maze. The distance record was obtained by tracing the move- ments of the rat upon soft white paper with a very soft pencil. The tracing so obtained was then measured by means of a chart- ometer which had been calibrated. Calibration showed that the instrument had an error of about one per cent. As has been stated by Professor Watson, the ratio of the size of the image to that of the maze is as 1 to 6.4. Consequently, the distance in cm. obtained with the chartometer must be multiplied by 6.4 in order to obtain the actual distance run by the animal. The values given in the table represent the actual distance covered by the rats. For example, in trial 31 (p. 163), the dis- tance accumulated by the chartometer was 92.69 cm. Con- verting this we have 92.69 cm. x 6.4=593.2 cm., as the actual distance run. Figure 2 shows the actual tracing, figures 1 and 3 the diagrammatic representation, of the paths traversed by two rats in the trials indicated. The records of the total time consumed (T.) and of the total distance (D.) run were taken on 27 rats, 14 males and 13 females. The animals began the problem when 35 days old. They were born and reared in the laboratory, consequently all were tame and accustomed to handling. The rats had been fed in the maze 1 From the Psychological Laboratory of the Johns Hopkins University. PO TIME VERSUS DISTANCE IN LEARNING 61 Figure 1. Schematic representation of pathway traversed by rat No. 28 on his 2nd trial in the maze. Four sheets were used in the actual tracing. Time, 2419.8 seconds. Distance, 1579.2 cm. Figure 2. Actual pathway traversed by rat No. 26 on the 62nd trial in the maze. Time, 31 seconds; Distance, 630.4 cm. G2 HELEN B. HUBBERT for a week before the experiment began, but during feeding they were strictly confined to the food box. Two trials per day were given each rat, and at the end of the second trial the animal was allowed to eat in the food box for from three to five minutes. Milk-soaked bread was used as the incentive throughout. Until the animals were 45 days old they were allowed to eat food in the cage for from three-quarters Figure 3. Schematic representation of pathway traversed by rat No. 26 on the 78th trial in the maze. Time, 5.8 seconds. Distance 448 cm. of an hour to one hour after each day's run. The rats were run every day and as nearly as possible at the same hour every day, since it was found that rats accustomed to being run at night did not react well if forced to run in the daytime. The problem was considered learned when the rat, for six consecutive trials, went straight to the food box without stopping or turning aside from the true pathway, i.e., when all excess movements had been eliminated. No time limit was set, but as a matter of fact it was found that most of the rats made such runs in six seconds or less. The shortest perfect run was 4.2 seconds, TIME VERSUS DISTANCE IN LEARNING 63 TABLE 1 Trial No. Rats Average Average 1 Trial No. Rats Average Average No. Running Time Distance cm. No. 1 Running Time Distance cm. 1 27 467.0 4216.1 59 18 15.9 607.7 2 27 627.7 3736.1 60 18 10.6 550.9 3 27 413.8 3147.8 61 18 10.1 519:5 4 27 158.5 1866.9 62 18 9.9 558.0 5 27 129.9 1573.6 63 18 14.5 553.5 6 27 186.6 1719.2 64 18 12.9 581.3 7 27 79.2 1164.1 65 18 10.1 551.7 8 27 68.8 1300.2 66 17 8.4 518.2 9 27 48.1 925.6 67 16 10.5 562.2 10 27 64.9 1169.0 68 16 8.8 505.0 11 27 40.3 1029.8 69 16 6.6 473.4 12 27 62.9 1240.7 70 16 22.5 496.6 13 27 59.4 925.1 71 15 7.7 484.3 14 27 77.7 1022.5 72 15 8.4 489.4 15 27 86.9 874.6 73 15 8.7 518.0 16 27 25.5 868.4 74 14 18.4 519.3 17 27 20.8 671.5 75 11 12.9 416.0 18 27 51.0 1148.1 76 10 26.5 526.7 19 27 33.1 772.0 77 8 11.2 480.8 20 27 35.4 1032.7 78 6 44.6 706.1 21 27 24.2 739.9 79 5 42.3 798.7 22 27 48.2 1010.4 80 5 36.4 508.2 23 27 19.6 649.5 81 5 8.6 448.0 24 27 24.5 789.0 82 5 147.7 788.5 25 27 27.7 754.7 83 4 28.1 600.0 26 27 26.1 756.5 84 4 359.8 1468.8 27 27 71.6 776.6 85 4 55.9 794.4 28 27 16.9 633.3 86 4 6.6 451.2 29 27 36.7 641.9 87 4 19.7 513.6 30 27 25.1 790.8 88 4 6.3 448.0 31 27 31.8 593.2 89 4 13.9 556.8 32 26 25.1 734.8 90 4 7.4 526.4 33 25 100.5 830.5 91 3 47.7 780.8 34 25 30.3 821.6 92 3 15.8 669.9 35 25 19.6 559.5 93 3 24.1 805.3 36 25 31.2 840.5 94 3 6.1 448.0 37 25 19.9 520.6 95 3 52.2 1339.7 38 25 20.5 674.1 96 3 7.2 509.9 39 25 27.3 569.6 97 3 24.7 904.5 40 25 28.9 879.7 98 3 7.0 448.0 41 25 57.3 704.0 99 3 10.1 460.8 42 25 18.2 616.7 100 3 6.1 460.8 43 25 9.0 505.0 101 3 6.9 448.0 44 25 13.7 603.8 102 3 5.3 448.0 45 25 12.3 497.5 103 11.2 448.0 46 24 19.0 716.4 104 5.4 448.0 47 23 37.1 683.4 105 8.0 537.6 48 23 14.7 574.3 106 4.8 448.0 49 22 24.1 581.2 107 8.0 ' 448.0 50 22 24.4 706.5 108 12.0 566.4 51 21 15.7 534.3 109 5.2 448.0 52 21 12.6 514.7 110 5.4 512.0 53 20 14.1 521.6 111 5.4 448.0 54 20 9.9 500.8 112 5.2 448.0 55 19 8.4 506.8 113 4.2 448.0 56 19 13.3 545.2 114 4.8 448.0 57 19 19.6 588.3 115 5.8 448.0 58 19 10.4 562.5 116 4.0 448.0 64 HELEN B. HUBBERT the longest, 14.2 If a rat remained in the maze for 15 minutes without reaching the food box he was taken out and replaced in the entrance box for a second attempt. Distance and time were recorded in the same way as for a successful run, i.e., when an animal worked for 15 minutes on his first trial, and failed, and then attained success after eight minutes on his second attempt, the total time of his first trial would be 23 minutes. Distance was treated in a similar way. In the Table 1 given above, the time record and distance record of all individuals at work upon the problem at any given trial were separately averaged, e.g., in trial 1, 27 animals were used, the time and distance records were taken and then averaged separately, giving one point on the time curve and one on the distance curve respectively. Once the problem had been learned by an animal it was taken from the group. As may be seen in Curve 1, plotted from Table 1, the number of animals at work is steadily decreasing. The striking similarity between the time and distance curves bears out the contention of Watson and of Hicks that the time record is a good index of progress in learning. It will be recalled that this position has been severely criticised by Washburn 2 and by Yerkes 3 . The similarity of the curve contour, and the close correspondence of the maxima and minima is apparent. As Hicks has pointed out, certain differences between T. and D. appear in the early trials. The drop in time in the first nine trials is 89.5 per cent, while that in distance is 78.5 per cent. A partial explanation for the increased percentage of the time drop may be found in the rat's behavior during the first few runs in the maze, when he often crouches against some partition and refuses to run for three, six, ten and sometimes even fifteen minutes. In such cases the time average increases enormously, while the distance average remains practically unchanged. Perhaps a better method of procedure would have been to deduct from the time record the time spent in absolute quiet, but this would lead into difficulties of standardization, making necessary an arbitrary decision as to how much time shall elapse without movement on the part of the rat before deduction is justifiable— a pitfall similar to that encountered in computing errors, and one « Washburn, M. F., Jour. Comp. New. and Psy., 1907, Vol. 17, p. 532. 3 Yerkes, Robert M., Jour Phil, Psy. and Sci. Methods, Vol. IV, p. 585.. TIME VERSUS DISTANCE IN LEARNING 65 we are anxious to avoid. It is this same system of taking records, however, which largely accounts for the apparent dis- crepancy in the time curve from the 81st to the 86th trial. Rat No. 22 became erratic, increasing his time record enormously here, while his distance record changed but little. Thus in the 400 350 Q 3> s 300 250 20a 150 100 50 U37 Timet — 1 1ur,if-5sec DisTancif---) 1umt-5dm-50cm 90 100 Hi Curve I. Relation of time to distance. The figures above the curve indicate the number of rats running at each trial. Failures are counted in. The first point in the time curve is 350, the second (not shown in curve) 627. Plotted from the figures in Table 1. 82nd trial, where the first decided rise occurs, his time record was 715 seconds, while the combined time of the other rats running was only 23.6 seconds. Had No. 22 not been running the average time would have been 5.9 seconds instead of 147.7 seconds, while the average distance would have been 448 cm. 66 HELEN B. HUBBERT 3.50 300 ^50 instead of 788.5 cm. 1 So in the 84th trial. Here the recorder's note-book states that after the first spurt, which occupied a little over two minutes, the same rat, No. 22, refused to run and crouched near the far partition in alley 1 for the remainder of the 15-minute norm. Again throwing out his record, we have a time average of 5.3 seconds instead of 359.8 seconds and a distance average of 4,480 cm., instead of 1,468 cm. In the 86th trial, he steadied down and the throwing out of the record here gives 5.4 seconds no Curve II. Plotted from Table 2, which is the same as Table 1, except that failures are eliminated. instead of 6.6 for time and 452.3 cm. instead of 45.1 cm. for dis- tance. The small curve, Curve III, shows the effect of throwing out the record of rat No. 22. The similarity of the time and distance curves again becomes apparent. An additional factor which must be considered in respect to both time and distance, when interpreting the latter part of Curves I and II, is the small number of rats then running. The best ones had dropped out 4 It will be remembered that the minimum time may be taken as seven seconds while the minimum distance is 448 cm. TIME VERSUS DISTANCE IX LEARNING 67 and at the 81st trial only five were running, while at the 86th trial only four remained. Where only a few animals are running individual differences are sure to alter the curve contour. With the exception of the discrepancy shown in the first eight trials and the apparent discrepancy in the final trials, the time and distance curves show great similarity in contour. Table 2 shows the effect of throwing out both time and distance records in each case of failure on the part of individual animals. It was found that the figures were identical with those of Table 1 , except for the trials indicated. For the sake of comparison the original records (Table 1) for the corresponding trials are repeated. Examination of Curve II, plotted from these figures, shows a somewhat steadier curve and in general a closer ap- proximation between time and distance. The same explanation V\A Curve III. To show the effect of eliminating the records of Rat No. 22 from the 81st to the 86th trial. which was offered in the case of Curve I for the discrepancy from the 81st to the 86th trial holds here. TABLE 2 ' Failures Averaged Failures Eliminated Trial No. No. Average Average No. Average Average Rats Time Distance cm. Rats Time Distance cm. 1 27 467.0 4216.1 24 325.7 3203.2 2 27 627.7 3736.1 25 257.1 2373.9 3 27 413.8 3147.8 25 236.9 2553.7 5 27 129.9 1573.6 26 98.4 1279.7 6 27 186.6 1719.2 26 95.3 1379.7 15 27 86.9 874.6 26 53.3 822.7 27 27 71.6 776.6 26 32.4 700.5 33 25 100.5 830.5 24 23.0 648.0 41 25 57.3 704.0 24 20.3 595.1 84 4 359.8 1468.8 3 5.3 448.0 08 HELEN B. HUBBERT In an admirable discussion of the values of curves of learning 5 Miss Hicks says (p. 141): " The total distance criterion presents so many difficulties as to render it impracticable for ordinary work. One difficulty lies in the matter of taking records ac- curately. The rats, after a few trials, run so rapidly that it is extremely difficult for one person to observe and record at the same time. To do this, it is necessary to mark off the maze into small segments and commit to memory some scheme of representation so that records can be jotted down in a purely automatic manner. The work of transcribing this record into distance terms and computing the same is very laborious. Eliminating these practical difficulties, the distance criterion is in some ways an ideal one. (Italics mine.) There can be no divergence of practice as to what shall be omitted or included and results obtained by different experiments upon the same maze will be strictly comparable." (Page 154.) ' The distance and error criteria are fundamentally alike. The distance curve is the better representative of the progressive approximation of the act towards automatic accuracy. It portrays all the details of this eliminative process and it approximates the ideal of uniformity and regularity of descent. However, it is impractic- able from the standpoint of recording and manipulating the data." These practical difficulties in recording and manipulating the data have been overcome, at least where small animals are the subjects used in the maze. The total distance record can be obtained accurately. Its desirability has, so far as we know, never been questioned. The error curve has often been used in lieu of the distance curve, but it has been criticized as practically valueless because of the difficulties encountered in standardizing an " error." 6 The prevalent practice of omitting all total and partial returns from the error record, and of making no attempt to evaluate varying degrees of error gives a curve which is not only worthless but false." 7 With this criticism on the customary methods of obtaining the errors committed by an animal we are in hearty sympathy. It is far better, both from the standpoint of convenience and that 5 Hicks, V. C, The relative values of the different curves in learning. Jour. Animal Befiav., Vol. 1, pp. 138-156. 6 Watson J. B., The behavior of noddy and sooty terns. Carnegie Pub., No. 103, p. 249, note 1. 7 Hicks. Ibid., p. 156. TIME VERSUS DISTANCE IN LEARNING 69 of accuracy, to base comparisons upon the time alone than to consider these so-called errors. As to which type of record is best, time or distance, it seems wise to await a more complete study of the question before deciding. CONCLUSIONS I. It is possible to chart the path of the run of an animal through the maze and to measure accurately the total distance covered in that run. II. Time and distance curves are so similar in character when normal animals are tested that it is impossible to state which is the better criterion of learning. We are sadly in need of a close analysis of just what time and distance curves mean. In deter- mining the activity at different ages, the difference between blind and normal animals, between anosmic and normal animals, etc., there still remains the possibility that the distance and time curves will show great disparity. It is not fair, however, to maintain this until there is some factual evidence in its favor. ON SOUND DISCRIMINATION BY CATS W. T. SHEPHERD, Ph.D., Waynesburg College This paper is a report of experiments which were made on cats, to ascertain their ability to discriminate sounds; viz. : differences of musical notes of different pitch, and differences of intensities of " noise." If an animal forms an association between a certain musical note and food, so that he reacts in a definite manner to that note in order to obtain food, and does not so react to other notes, we may infer that he discriminates that note from the other notes. In a similar manner a cat may give evidence that it distinguishes one " noise " from another " noise " of a different intensity. Kalischer 1 has reported experiments with dogs on auditory discrimination, in which his animals showed ability to discrimi- nate pitch. His method was to sound a certain note on an organ, or on an harmonium, whereat the dog should react in a definite manner, such as by springing up and snapping at a piece of meat which the experimenter held out in his hand. Selionyi, - using the " Pawlow method " on dogs, obtained evidence of discrimi- nation by them of the tones of an organ, of organ pipes, and of the sounds of two whistles. Selionyi, as well as Kalischer, was chiefly interested in sound discrimination from a physiological standpoint. In experiments on auditory discrimination in raccoons made by the present writer, reported by Cole, 3 evidence of discrimination of musical pitch by these animals was obtained. In experiments which I made in 1909 on auditory discrimination in Rhesus monkeys, satisfactory evidence of discrimination by the latter animals of pitch, and also of " noise," was obtained. 4 Johnson, 5 in tests on sound discrimination in dogs, lately made at i Kalischer, O., Eine neue Horprufungsmethode bei Hunden. Sitz. d. Kgl. Ak. d. Wiss., X, 1907, 204 ff. « Selionyi, G. P., Contribution to the study of the reactions of the dog to auditory stimuli. St. Petersburg, 1907. 3 Cole, L. W., Concerning the intelligence of raccoons. Jour. Cornp. Neur. and Psy., 17: 236. * Shepherd, W. T., Some Mental Processes of the Rhesus Monkey. Psy. Rev. Mon. Sup. No. 52, 1910, 26 ff. » Not published at this writing, Dec. 17, 1912. 70 ON SOUND DISCRIMINATION BY CATS 71 the Johns Hopkins University, got some positive results. How- ever, I am informed in a verbal report of the experiments by Dr. J. B. Watson, that an average of only about 60 per cent, of correct responses was obtained. At the time the experiments herein reported were begun, one of the animals was eighteen months old. Its only previous training had been in some tests on discrimination of articulate sounds, completed about nine months previously to those ex- periments. The animal appeared to be of medium intelligence. The other cat, the mother of the first, was about four years old. It had received the same training in sound discrimination as the other animal, and at the same time. The latter cat was of average intelligence. Both were ordinary grey house cats. In the experiments on pitch discrimination to be reported in this paper, I used the same plan as I had employed in the ex- periments with the raccoons and with the monkeys, and which is somewhat similar to the method Kalischer employed. The cat was placed in a cage 66 centimeters in height. A wire netting formed the front, and also the top of the cage. The experimenter sat at a distance of about a meter from the cage and sounded a certain note on the instrument used. At this note, in the case of the first animal, Pet, the cat was to rear up with its paws on the front of the cage and look up through the top of the cage for food to be given it. At the other notes, it was not so to react. In the case of the second animal, Mary, on account of its inactive habits, as shown in previous experiments, it was deemed a suffi- cient response if it merely looked up at the top of the cage for food, when the ' feed " note was sounded. But whether or not a positive response was obtained, the animals were fed when the " feed " note was sounded. Ten seconds were allowed for a response. In experiments 1 and 2, an ordinary harmonica, A, was used; in experiment 3, a Bradbury piano; the procedure in experiment 4 will be explained under that head. In each test, care was taken not to give by looks, movements of the hands, or in any other manner, any cue to the proper response. In order that the animal might not react to the mere rhythm of the sounds, they were sounded in an irregular order. 1. Discrimination of a Difference of Two Octaves of Pitch on a Harmonica. A-3, Food Note; A-l, Non-food Note. Pet. — In the first day's experiment, this animal showed 72 W. T. SHEPHERD indications of forming the association ; in thirty trials 6 it responded to A-3, the food-note, fourteen times, and to A-l two times. Its first correct reaction was in the eleventh trial, the next in the twelfth, next the fourteenth, then the seventeenth, and in an increasing degree throughout that day's tests. On the second day, in fifteen trials, or in forty-five trials in all, Pet had per- fected the association. In fifteen trials on that day it responded to A-3 fourteen times with one doubtful response, and to A-l one time with three doubtful. In two trials that day, the fourth and the seventh, the animal got up at A-l, but got down at once and appeared, from its looks and actions to know that it had made a mistake. In ten test trials four days later, the cat reacted to A-3 ten times and to A-l none. Mary.— The older animal showed no definite indication of forming the association until the third day. On that day, in ten trials, it responded to A-3 eight times and to A-l five times with two doubtful responses. After this, Mary continued to improve, and on the seventh day had perfected the association. In twenty trials that day, or in ninety trials in all, the animal responded to A-3 nineteen times, and to A-l, five times, with two doubtful. At the twentieth trial, when I sounded A-l, Mary went to "washing her feet," and appeared, by her actions, to know that no food was to be gotten at that note. In two ad- ditional day's tests, Mary responded to A-3 twenty times in twenty trials each day, with five and four wrong responses res- pectively on the eighth and ninth days. This cat never entirely inhibited the tendency to react to A-l. 2. Discrimination of a Difference of One Octave of Pitch. Notes A-2 and A-l on a Harmonica. Fed at A-2. Pet. — In the first day's experiment, in ten trials, the animal gave no positive response. I had handled it rather roughly on the previous day, and the cat appeared uneasy and afraid of me. Also it did not seem hungry. On the following day, in twenty trials, or in thirty trials in all, Pet reacted to A-2 eighteen times and to A-l two times. It did not respond to A-2 in the first trial nor in the eleventh. The animal reacted to A-l (wrong) in the third and fifth trials only. When it got up at A-l in the fifth trial, it got down at once, and gave every indication 6 That is in thirty trials of each auditory stimulus. ON SOUND DISCRIMINATION BY CATS 73 of being aware that it had made a mistake. In a test of ten trials on the following day, Pet made no mistakes. 3. Discrimination of the Difference of Two Octaves of Pitch on a Piano; i.e., between F-l, Bass Cleff, and F-2, Treble Cleff. Fed at F-l. Not fed at F-2. Pet. — On the first day of this experiment, in twenty trials, the cat reacted to A-l thirteen times and to A-2 three times. Its wrong responses to A-2, were in the first, second and fifteenth trials. On the second day of the' experiment (three days later), in twenty trials, or with forty trials in all, Pet responded to A-l twenty times and to A-2 none. In the first trial on that day, the animal got up slowly at the food-note, as if in some doubt what to do. In the remaining nineteen trials correct responses by the cat were prompt. 4. Discrimination of Noise. The animal was placed in the cage as in the preceding experi- ments on pitch discrimination. The sound apparatus consisted of a wooden box, 18 x 11 x 10 inches, and a slat 13 x 4 x tV inches, fastened to the top of the box by a leather hinge. By rais- ing the free end of the slat and suddenly letting it go, it struck the top of the box and made a sound varying in loudness with the force with which it struck. To give sounds of different degrees of intensity or loudness, two sticks, one 2\ inches in length, the other 4+ inches, were separately used and placed perpendicular to the box, under the free end of the slat. By pressing on the slat near the hinge and suddenly removing the shorter stick, the slat would strike the box and produce a sound of noticeable intensity, and by using the longer stick in a similar manner, a louder noise was made. The same pressure, as nearly as possible, was exerted on the slat in both cases. By rearing up and looking through the top of the cage when the louder noise was made, the animal was to show its discrimination of the louder and lesser noises. It was fed at the louder noise and not fed at the lesser. The noise apparatus was manipulated at the closed side of the cage, so it was not possible the reactions were to stimuli other than the sounds. The noises were made in an irregular order. Pet. — On the first day of the experiment, in twenty trials, the animal reacted to the food-sound, the louder noise, eight times, to the lesser noise four times and with three doubtful responses. 74 W. T. SHEPHERD In the third trial (the first correct response), in the eighth and in the fifteenth, though giving the correct response to the stimulus, the cat's reactions were slow, as if in some doubt what to do. On the second day, in twenty trials, or forty trials in all, Pet reacted to the louder noise twenty times and to the lesser noise none. In the second and fourth trials it started to respond to the lesser noise but stopped. On the next day, being tested in fifteen trials, the animal made no mistakes. We may regard the results of these experiments as positive. In the experiments on pitch discrimination, the criticism may be offered that the experimenter should have sounded the notes out of the animal's sight. This is true, but as I had no assistant, it was not practicable. However, I was careful not to give the cats any cue to the correct response, by any difference of attitude when the food notes were sounded and when the other notes were played. So it does not seem possible to attribute the reactions of the animals to anything else than the association of a certain note with food-getting, and the consequent discrimina- tion of that note from the other notes. Moreover, several observed incidents in the course of the experiments, such as getting up at the wrong note and getting down at once, strengthen this conclusion. Furthermore, the looks and actions of the ani- mals, to an unbiassed observer, would indicate such clear discrimi- nation of the notes. We conclude, therefore, that eats, or at least some cats, discriminate musical pitch, and also discriminate noises of different degrees of intensity. It will be noted that in the experiment in which two individuals were tested, i.e., in the discrimination of a difference of two octaves of pitch on a harmonica, while the younger animal formed the association in 45 trials, the older animal required 90 trials to perfect the association. Again, as compared with the ability of raccoons, in similar tests, of the discrimination of the difference of two octaves of pitch on a harmonica, while the cats took respectively 45 and 90 trials, the two raccoons tested required 100 and 150 trials respectively to form the association. In exactly similar experiments which the writer made on two Rhesus monkeys, one individual formed the association in 30 trials, and the other in 40. Though, from the fewness of the individuals used in these different experiments, we are not warranted in drawing final conclusions as to the comparative rapidity of the ON SOUND DISCRIMINATION BY CATS 75 formation of such associations in cats, raccoons and monkeys, the indications, however, point to the conclusion that Rhesus monkeys form such associations with somewhat more rapidity than cats, and with considerable more facility than do raccoons under similar experimental conditions. NOTES A NOTE ON THE SUPPOSED OLFACTORY HUNTING- RESPONSES OF THE DOG H. M. JOHNSON Nela Research Laboratory, National Lamp Works of General Electric Co., Nela Park, Cleveland Ohio. Our present inability to measure or control olfactory stimuli may have some bearing on the fact that the dog's sense of smell has been avoided by careful experimental investigators. A number of problems in this field are highly interesting even though they do not lend themselves to treatment by quantitative methods. The supposed olfactory responses made by the dog in hunting are especially puzzling. At present the only available data are purely anecdotal, and these are meager, conflicting and untrustworthy. Various assertions have been made as to the ability of the trained hunting-dog to trail his quarry. The belief is wide- spread that the bloodhound can follow a trail over 24 hours old without back-tracking. Let us accept provisionally a seemingly conservative statement; that a fox-hound can follow a three- hour-old trail of a rabbit without back-tracking, and assume that this is very near the limit of the dog's ability. A casual examination suffices to show the difficulty of explaining the dog's behavior. Many hunters have said in effect that the dog follows the trail in the direction taken by the rabbit because the tracks made recently excite more intense smell-processes than do the older tracks. This explanation is not satisfactory. Suppose that in each of a series of tracks, a, b, c, etc., a like quantity of the same single sme 1-substance had been deposited by the rabbit; that the tracks had been made one second apart, and that a was made three hours ago. It is evident {changes of chemical composition being excluded) that the smell-substance is greatest in quantity when first deposited. It becomes dissipated in time so that in this case there is barely enough left in the track a to affect the dog. If the smell-substance is deposited in a gaseous state, its 76 OLFACTORY HUNTING RESPONSES OP THE DOG 77 diffusion could be represented by one of the well-known " curves of decay." The absolute intensity of the stimulus (i.e., the amount of odorous substance present in the track at a given moment of time) may, within limits, be formulized : Log S t = Log S — kt, wherein S equals the amount of the substance first deposited, t the time which had elapsed since the deposit was made, and k a constant function dependent on conditions of temperature, pressure, etc. In the case under consideration the stimulus-intensity at the track a is nearly zero when it is presented to the dog. The absolute difference of stimulus-intensity at a, b and c would have to be extremely small, since the difference in the respective values of t is of the order of one part in nearly 11,000. Further: even this difference between a and b would exist only if they were simultaneously presented. Since the dog is supposed to be following the trail of the rabbit, for him to be affected by even a part of the difference between a and b it is necessary that he travel faster than did the rabbit in making the tracks. If the dog travels at the same rate as did the rabbit, when he reaches b its intensity will be just equal to that of a when a was passed. Moreover, in actual practice other difficulties arise. Suppose the rabbit has run from moist ground to dry ground. The smell- substances are diffused more rapidly under conditions of relatively small humidity that under conditions of greater humidity. The stimulus-intensity of the recently made tracks on dry ground could thus be less than those made earlier on the wet ground. In such case our assumption fails to explain the dog's failure to show confusion. But the dissipation of the smell-substance may be a complex process. For instance, it may be deposited, not in a gaseous state, but as a liquid or solid. In such case vaporization must precede diffusion. Vaporization, conditions being constant, proceeds at nearly a uniform rate in the open air. The amount of substance present in a gaseous state might thus be as great at a very advanced stage of dissipation as at an earlier stage. Since the substance to be odorous must be gaseous, we are not warranted in assuming that the stimulus-intensity is greater at a recently made track than at one made earlier, unless we know that all the smell-substance in the later track has been vaporized. 78 H. M. JOHNSON There may be other factors such as chemical changes by which the deposited substance becomes odorous, etc., but con- sideration of them only increases the presumption against the intensity-difference theory. It has been suggested also that the dog may have an acute olfactory sensitivity to the form of the tracks made by his quarry and follow the trail from heel to toe. Certain features of the dog's behavior certainly indicate that he is very sensitive to differences of spatial position of olfactory stimuli. Another sug- gestion is that the smell-substances deposited by the different parts of the foot or body may differ specifically in stimulating quality, and that the dog is affected by this difference. Assuming either of these suggestions as a complete explanation of the dog's hunting behavior would require us to expect a bloodhound striking a man's trail at right angles, to back-track if the man had walked backward instead of forward across the field. Dr. P. W. Cobb has suggested a simple hypothesis; that the dog's sense of direction may be due to the trailing of ground smell-substances. For instance: the smell-substances affecting a dog trailing a man who had crossed a mint-bed might be (1) ground + man; (2) ground + man -f mint, the mint being intense; (3) ground + man + mint, the mint-smell-substances diminishing rapidly in the direction the man had taken. The hypothesis impresses the writer as being valuable, although it does not afford a complete explanation of the facts as variously alleged. The value of careful field-tests should be apparent. The question may well be raised whether the hunting-behavior of the dog is really an olfactory response. A comparison of the field- behavior of anosmic dogs and normal dogs of the same litter and of a hunting breed, such as the beagle-hound, should prove highly interesting. It would be well worth while to ascertain as a beginning what responses a good hunting dog actually makes when introduced to trails the time and direction of which had previously been ascertained. The effect of numerous disturbing factors which could be introduced, some of which have been suggested above, ought to be quite interesting. It is to be hoped that some one with proper facilities and ample training may become interested enough to make an experimental in- vestigation in this field. JOURNAL OF ANIMAL BEHAVIOR Vol. 4 MARCH-APRIL, 1914 No. 2 LIGHT DISCRIMINATION IN THE ENGLISH SPARROW EUPHA FOLEY TUGMAN, A.M. From the Psychological Laboratory of Indiana University CONTENTS PAGE I. Introduction 79 (1) Statement of the problem. (2) Previous experimental work. (3) History of the investigation. (a) Preliminary experiment. II. Method 83 (1) Care of the birds. (2) Method of handling the birds. (3) Description of the apparatus. (4) Experimental procedure. (5) Calibration of the lights. III. Results 89 (1) Threshold of discrimination. (a) Comparison with human threshold. (2) Method of learning. (a) Tables and results for each bird. (b) Comparison of results. (3) Incidental results. (a) Individuality. (b) Influence of former experiences. (c) Position habit. (d) Mental instability. (e) Relation of time to the failure or success of choice. (f) Effect of 48 hour interval between successive series. (g) Persistence of stimulation. IV. Questions suggested by this investigation 107 I. INTRODUCTION (1) Statement of the problem. — This investigation was pursued for the purpose of determining (1) the threshold of brightness discrimination in the house (English) sparrow, and (2) the behav- ior which the sparrow exhibits and the habits which it forms in learning to make such discrimination. 80 EUPHA FOLEY TUGMAN (2) Previous experimental work. — Considerable work has been done in experimenting on the learning methods and capacities of animals, from amoeba to man. Only recently, however, have investigators begun to study the delicacy and completeness of the sensory equipment of animals. Of most importance in con- nection with the present investigation is the work of Breed, 1 Cole, 2 and Bingham 3 on chicks, and Porter 4 ' 5 on birds. Breed 1 used the discrimination method in experimenting with chicks. His results show that his chicks could discriminate between black and white, different colors, and two objects of different size. He made form tests also, but his chicks gave negative results. Cole 2 used the discrimination method in studying " the rela- tion of strength of stimulus to the rate of learning in the chick." His results seem to indicate that when discrimination is easy the number of trials necessary for learning is less than when the discrimination is difficult. Bingham 3 experimented on size and form perception. His chicks were punished by electric shocks when they made a wrong choice, and were rewarded with " food, light, warmth, and com- panionship " when they discriminated properly and were thus able to reach their nest box. He found that " the chicks' thresh- old of difference in size perception lies between one-fourth and one-sixth when the diameter of the standard circle is 6 cm." He holds that " earlier experimenters on the chick's perception of forms have failed to eliminate all possible conditions for dis- crimination other than the factor of form. * * * Reactions to optical stimuli which have been interpreted by observers as indicating form discrimination are probably made on the basis of unequal stimulation of different parts of the retina. * * 1 Breed, F. S. The Development of Certain Instincts and Habits in Chicks. Behavior Mono., vol. 1, No. 1, Nov. 1, 1911. 2 Cole, L. W. The Relation of Strength of Stimulus to Rate of Learning in Chicks. Journal of Animal Behavior, vol. 1, No. 1, 1911. Page 111. 3 Bingham, H. C. Size and Form Perception in Gallus Domesticus. Journal of Animal Behavior, vol. 3, No. 2, 1913. Page 65. 4 Porter, J. P. A Preliminary Study of the English Sparrow and Other Birds. Amer. Jour, of Psych., vol. 13, 1904. Page 313. 'Porter, J. P. Further Study of the English Sparrow and Other Birds. Amer. Jour, of Psych., vol. 17, 1906. Page 248. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 81 » A chick can acquire a perfect circle- triangle reaction, but con- trol tests show that it has no general idea of circularity in con- trast with triangularity." Bingham's conclusion is that " the order of importance of factors in the chicks vision is size, brightness and general illu- mination, and form." Porter's 6 work is more directly related to the present inves- tigation since he experimented with English sparrows. He first studied their method of approaching food, and found that they alight some distance from the food and then hop to it. He then studied their rate of learning to open a latch on a food box and of finding their way through a maze. He found that the spar- rows learned quite rapidly by the trial and error method. They exhibited some ability to profit by experience, also to some extent by imitation. Porter then tested the number sense of the English sparrow in the same way that Kinnaman 7 did with monkeys. The birds could not count, but showed some sense of position. When the food glass was covered with colored papers of the standard yellow, blue, red, and green, the birds were able to distinguish the colors. Various forms of food boxes were then used and the food was placed in one of them, the position of which was changed from time to time. This test was made on one bird only — a female. She was unable to dis- tinguish the forms. In his later work Porter 8 experimented with English sparrows, vesper sparrows, a cowbird, and a pigeon, attempting to compare the rates and methods of learning of the different birds. In learning the simple maze the vesper sparrow was the slowest. There was little difference between the others. All showed good memories, the cowbird being best. With the puzzle box the sparrow learned most rapidly. Porter then tested the birds for discrimination of three horizontal black lines on a card from a blank card; a card marked with a black diamond from a blank card; and two marked cards from each other. Both the Eng- • Porter, J. P. A Preliminary Study of the English Sparrow and Other Birds. Amer. Jour, of Psych., vol. 13, 1904. Page 313. ' Kinnaman, A. J. Mental Life of Two Macacus Monkeys in Captivity. Amer. Jour, of Psych., vol. 15, 1902. « Porter, J. P. Further Studv of the English Sparrow and Other Birds. Amer. Jour, of Psych., vol. 17, 1906. Page 248. 82 EUPHA FOLEY TUGMAN lish sparrow and the cowbird learned to distinguish these de- signs, but the sparrow showed some superiority in being able the more quickly to unlearn an old habit and to learn a new one. In discriminating colors all the birds did about equally well, except the sparrow which showed superiority in the case of blue. (3) History of the investigation. — (a) Preliminary experiment. This investigation was begun in the Psychological Laboratory of Indiana University on February 8, 1911. Twenty-six house sparrows were obtained but seventeen soon died. The experi- ment was begun with the nine remaining birds. Two birds were tested to see if they could be trained to choose the darker of the two stimulus areas. One was given 38 trials but chose the darker side only six times or 14% of the time. The other bird was given 32 trials and it chose the darker side 5 times, which is 14%. Then both birds died. The other seven birds were allowed to choose the brighter side which seemed to be the natural tendency. But the experiment was not continued very long as these birds also died. However, the few results obtained seemed to show that all the birds were able to discrim- inate the wide differences in intensities used. The preliminary experiment was important in that it showed wherein the apparatus and method needed revision. One of the most important changes made in the apparatus was in the method of producing a motive. In the preliminary experiment the floor of the discrimination chamber was covered with par- allel copper wires one centimeter apart. The wires were con- nected so as to form an interrupted circuit in connection with the induction coil and key. This was the plan used by Yerkes 9 in his experiments with the dancing mouse. But the sparrows rested with their feet in a sort of arched position with only the claws touching the wires. Consequently it was not easy to shock them. It was necessary to construct perches as described in a later section of this paper, and as used in the final experiment. In October, 1911 the writer resumed the experiment with four birds, — two males and two females. The work continued until June, 1912. • Yerkes, R. M., 1907. The Dancing Mouse. The Macmillan Co., N. Y. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 83 II. METHOD (1) Care of the birds. — The writer found it very difficult to keep sparrows alive in captivity. The birds were kept in a large cage before a large, open window in the experiment room. They were provided with an abundance of Spratt's mixed bird seed, fish bone, and clean water. The adjustable floors of the cage were cleaned frequently and covered with coarse sand. But for some unknown reason many sparrows died, usually in spasms. The experimenter found that the best plan was to cage the sparrows some weeks before beginning an experiment. If they died it was usually very soon after being placed in confinement. If they survived the first few weeks it was reasonably safe to begin an experiment with them. The four birds used in the final experiment kept in good physical condition until near the end of the investigation. One died in spasms shortly before the end of the experiment, and two others shortly after. The cause of death was not apparent. The fourth bird is still alive, after a confinement of one and a half years. (2) Method of handling the birds. — In the preliminary experi- ments the writer tried handling the birds. This seemed to make them wilder, instead of taming them. To avoid the necessity of handling the birds the writer constructed a box 24 x 19. 5 x 15 cm., and covered it with wire mesh. It was provided with a handle, and with a door at the end, and was used to transfer the birds from the cage to the apparatus. Against the back wall of each half of the cage was an inside adjustable wall of wire which, when pulled forward, forced the bird out at the door at the front, and into the portable box. Then the door of the box was closed and the bird carried about at will by the experimenter. This seemed to be a very good method of hand- ling the birds as it did not frighten them after the first few times. (3) Description of the apparatus. — The apparatus used in the investigation consisted of the Yerkes-Watson brightness appa- ratus, 10 and an experiment box modelled after the one described by these authors. 11 10 Robert M. Yerkes and John B. Watson. Methods of Studying Vision in Ani- mals, pp. 17-24. 11 Ibid, pp. 24-25. 84 EUPHA FOLEY TUGMAN The discrimination box is shown in Fig. 1 as that portion of the apparatus below the line EE 1 and marked DB. It is built of wood and blackened inside and out with several coats of lamp black and oil so as to make it a dull black. Very little light is reflected from the sides of the box. The box consists of an entrance chamber (C, Fig. 1) which is 5 x 4.5 x 19.5 cm.; the discrimination chamber is 54.5 x 46 x 19.5 cm., and is Figure 1. Discrimination Box. Openings indicated by D, chambers and alleys by C. Route of bird could be either C, D,, C, D 2 , C 2 , D 3 , C 3 , D 4 , C, or C, Di, C„ Dt,, C 5 , D 6 , Co, D 7 , C. Discrimination made after passing through D. Cross lines in discrimination chamber indicate perches 2cm. above floor. Split brass tubes, stuck on wooden core with paraffine, form broken circuit which could be closed by bird grasping perch in alighting. divided at the end next the light box by a partition 30 cm. long, into two identical compartments, Ci and C. Alleyways, C. and Ca, which are 86 x 8 x 10 cm. connect C, and C* respectively with the compartments Ca and Ce, both of which open into the entrance chamber C. These various compartments are separated by sliding doors D,, D 5 , D 8 , D<, D 6 , Do and D?, operated by LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 85 means of a system of cords and pulleys leading to the front end of the apparatus. Two centimeters above the floor of the large discrimination chamber and four centimeters apart are perches across the box. The perches were made of three-eighth inch 22 gage brass tubing, oxidized and slit longitudinally into two halves. The two halves of this tubing were placed on either side of a .5 inch wooden core and held in place with paraffine. One-half of each perch is connected to one terminal and the other half to the other terminal of a Williams' Dial induction coil whose primary is in circuit with two dry batteries. A hand key is placed in the circuit and the secondary coil shifted to a position so that when a bird is resting on a perch it may be shocked by closing the circuit at the key. It is natural for the bird to hop from one perch to another and grip the perch with the feet. So when the feet are moist they can be shocked very effectively. A wet pad is kept in the entrance chamber C (Fig. 1) to keep the bird's feet moist. When dry the horny epidermis serves to protect them from the electric discharge. The whole discrimination box is covered with .5 -inch wire cloth, not shown in Fig. 1. The chamber C, and the near por- tion of Ci and C4 are covered with black velvet paper which prevents the bird from seeing the experimenter and the experi- menter from seeing the bird until after the bird has discriminated and hopped to either Ci or C4. (4) Experimental procedure. — Before beginning the experi- ments each bird was left in the discrimination box for twenty- four hours. All the doors were left open so that the bird could thoroughly acquaint itself with the apparatus. The use of arti- ficial light for illuminating the stimulus areas made it necessary to conduct all the tests in a dark room. The birds were kept in the dark room so that they would not be excited by being moved from one room to another and also would become ac- customed to the darkness and to the noise of the induction coil. A small electric light in the room was turned off during each single test and then turned on again as soon as the bird had made its choice of the stimulus area and passed on into the alley. On succeeding days (Sundays excepted) each bird was given a series of 15 trials. A trial consisted in requiring the bird to 86 EUPHA FOLEY TUGMAN discriminate between the two stimulus areas in order to return to the nest box without receiving a shock. Two of the birds, Male V and Female VI, were required to choose the darker area and the other two birds, Male IV and Female V, were required to choose the brighter area. The two groups of birds were other- wise experimented upon in identically the same manner. Throughout the whole investigation the same order in shifting the lights was given to all four birds. The standard light was shifted to one side or the other at frequent and irregular inter- vals. The order was such that the standard light was on the left side the same number of times as on the right side. Table I gives the relative position of the standard light during the first 225 trials. The other positions were similar to the ones shown in Table I. When the standard intensity (.098 c.p.) was in the position L (Fig. 1) the letter L was used. That means the standard was on the left. When the standard was on the right (at R, Fig. 1), the letter R was used. Since two birds were trained to go to the darker side L means for them that the darker light was on the left side and the brighter on the right. In the positions marked R the darker light was on the right and the brighter on the left. TABLE I Position of Standard Light for the First 225 Tests Series 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Preference L RRRRLLRLL 1 R RLLRLRRLRRLRLR 2 L LRLRLLRRRLLRLL 3 L RLLLRRLRRLRLLR 4 L RRLRRLLRRLRLRL 5 L RLLRRLRLRRRLRL 6 L LRLRRLLRRRLRLL 7 L LRLLRLRRLLRRLR 8 L R R L R R L R L R L R L R L 9 L RLRRLRLLRRLRLL 10 R R L R L R L L R L R L R R L 11 R LRLLRLRLRLLRRL 12 R R L L R L R L R R L L R L R 13 L R L R L L R R L R L L L R R 14 L R R L L R L R R L R L R L R 15 R L R L L R L R L L R R L R L The sources of light used were two 25 watt, 110 volt Mazda lamps, which were as near alike in quality of light as the experi- menter was able to judge. The apertures in the stimulus adapter were 6 cm. in diameter. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 87 The lamps were occasionally interchanged to prevent the birds from forming any habits due to difference in the quality of the light which the observer was unable to detect. The discrimina- tion box was frequently washed out as a precaution against the bird getting any "cues" or reflections due to the waste material which collected. The procedure for each test was as follows: The bird was placed in the entrance chamber (C, Fig. 1) on the wet pad. The light in the room was then turned off and the induction coil started. After waiting several seconds to accustom the experimenter's eyes to the darkness, the door Di was opened and the bird allowed to pass into the discrimination box. (The entrance chamber was covered with black velvet paper and consequently was always dark. Hence the bird's eyes were accustomed to the dark.) Having entered it the bird could return to the entrance chamber only by passing through to the right or to the left of the partition and then through either door D 2 or D 5 and thence to C by way of the narrow alleys C 2 or C 6 through C 3 or Co and through doors D. or D? into the entrance chamber. After passing through the door D, the bird nearly always hesitated several seconds and frequently several minutes, before making a choice between going to one or the other of the stimulus areas. If the bird chose correctly it was allowed to pass on into the alleyway and back into the entrance chamber. But if it made a wrong choice the circuit was closed at the key and the bird shocked. The birds learned very soon that when they received an electric shock they must go back and around the partitions to the other side. In the earlier experiments the birds would often stop just beyond the door D,. In such cases the experimenter gave them instantaneous shocks which caused them to move on and thus force them to discrim- nate. But after they learned what they were expected to do and discrimination became established, they would hesitate only a few seconds before making a choice. In the later experiments it was seldom necessary to shock them except when they made a wrong choice. A series of correct choices for two consecutive days — 30 trials — was counted as correct discrimination and the difference in intensity of the standard and the variable light was decreased. At the beginning of the investigation the standard light was 88 EUPHA FOLEY TUGMAN placed at 38 cm. from the stimulus area, thereby giving it a measured intensity of .098 c.p. The variable light was placed at 238 cm. from the stimulus area, thereby giving it an intensity of .0025 cp.. After two consecutive days of correct choices the variable light was moved closer to the standard, which was always in the same position. Thus the intensity of the variable was increased each time, and the difference between the standard and the variable decreased in the same ratio. Each shift in the variable light necessitated the bird learning the problem again. It usually required a shorter time than the original problem, yet each shift was really a new problem to the bird. The differ- ence between the standard and variable was decreased step by step until the bird was unable to discriminate between the two areas. Then the difference between the standard and the vari- able was increased step by step until the bird was able to dis- criminate between the two areas. The threshold of discrimi- nation was taken as the difference between the least discrim- inable difference in the descending series and the least discrim- inate difference in the ascending series. Full records were kept during the progress of ' each series, showing the relative positions of the standard light at each trial, the time consumed in making each choice, the success or failure of the trial, and full data regarding the actions of each bird during each trial. The characteristics of the individual birds, methods of learning, etc., and the tables and results are taken from these records. (5) Calibration of the lights.^— The light sources and the stim- ulus areas were calibrated near the beginning of and frequently during the investigation with a Lummer-Brodhun photometer against a Tungsten light, standardized at 1 c.p. in Hefner units. The photometer readings were first taken with the light sources 38 cm. 12 back of the ground glass stimulus area. After the in- tensity of the stimulus area was calculated with the light source at this position, the intensities of the stimulus areas were cal- culated for the other positions of the light source. The intensities shown in Table II are calculated from the 12 The first experiments were made before this reading was taken. It was un- fortunate that the light was not set nearer the glass and thus have given a reading of I.e. p. instead of the decimal .098. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 89 readings taken on June 1, 1912, which was during the time that the birds were making the most difficult discrimination. TABLE II The figures in the first column indicate the number of discriminations required. The second column shows the c. p. of the standard light, calculated in Hefner units which remained the same throughout the experiment. The third column indicates the c. p. of the variable light and the last column shows the difference in intensity between the standard and variable. Intensity of Intensity of Discriminated ition Standard Variable Difference 1 .098 .002 .096 2 .098 .003 .095 3 .098 .004 .094 4 .098 .006 .092 5 .098 .008 .090 6 .098 .010 .088 7 .098 .015 .083 8 .098 .019 .079 9 .098 .024 .074 10 .098 .043 .055 11 .098 .046 .052 12 .098 .048 .050 13 .098 .062 .036 14 .098 .065 .033 15 .098 .068 .030 16 .098 .073 .025 17 .098 .076 .022 18 .098 .081 .017 19 .098 .085 .013 20 .098 .089 .009 III. RESULTS (1) Threshold for discrimination. — Table III shows the four sparrows' threshold of discrimination for the standard intensity .098. The second column gives the least difference between the intensities of the standard and the variable illuminated areas TABLE III Showing the threshold of brightness discrimination for each of the four birds. Bird Least Discriminable Difference, Descending Order Least Discriminable Difference, Ascending Order Estimated Threshold Male IV Female V Male V Female VI .013 .036 ! Discriminating at .03\ \ when he died / .022 .017 .033 ^022 .015 .035 ^022 90 EUPHA FOLEY TUGMAN which each sparrow was able to discriminate in the descending series. The third column gives the least difference which each sparrow was able to discriminate in the ascending order. The last column gives the average of these two which is taken as the threshold of discrimination for the standard .098 c.p. (a) Comparison with human threshold. The experimenter, using the same apparatus and methods, investigated light dis- crimination in three human subjects in order to compare the results with those obtained with the birds. Two of the human subjects were required to choose the darker and one the brighter of the two stimulus areas. The subjects are spoken of as A, B, TABLE IV Comparison of brightness threshold for the four sparrows and the three human subjects. Group Subject Estimated Threshold of Discrimination for the Standard Intensity=.098 c.p. Bird Male IV Female V Male V Female VI .015 c.p. .035 " .03 (died) .022 c.p. Man A B C .013 c.p. .009 " .013 " and C. B gave very much better results all through the inves- tigation than did either of the others and he responded to the darker of the two lights. C did not give very good results. He responded to the darker of the two lights. C was partially color blind and yet he was always seeing yellow or red in the lights. He said that the color in the lights confused him. This fact may have caused his poor results. The experiment with the human subjects was interrupted many times and often several days would elapse between successive series. This un- doubtedly had some effect on the judgments. As the observers' time was limited, the work was not done as thoroughly or as completely as in the case of the birds. The thresholds are not absolute but are estimated from the limited data at hand. Table IV gives the estimated thresholds for both the birds and the LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 91 human subjects. There seems to be little doubt that for the particular intensity the human subjects have a very much smaller threshold of discrimination than do the birds. However, Male IV could discriminate almost as fine a difference in intensity as did two of the men. When the difference between the two lights became very small two of the subjects found that they were simply guessing and not really discriminating. Yet their results appear very good. So even from the results it is very difficult to determine just what their thresholds of discrimina- tion are. (2) Methods of learning. — The data which this investigation gives regarding the learning process jof sparrows is fully as im- portant as the light which it throws on the threshold of visual discrimination. More problems were of course raised than were answered, but much light is thrown upon many of the factors which condition animal learning and which in turn condition the sensory threshold which an animal may acquire. One of the most striking facts is the very large number of trials neces- sary to bring the animal to the threshold. The three animals for which the threshold was determined averaged 2420 trials each. For the discrimination of the lowest threshold they aver- aged 480 trials each; one of them discriminating only after 615 trials. This bird was trained daily, Sunday excepted, from April 8th until May 31 before she made a record of perfect choices for two days in succession. (a) Tables and results for each bird. The results for each bird were tabulated in a regular form which stated the intensity (in Hefner units) of both the standard and the variable lights for each separate position; the difference between the intensity of the standard and the intensity of the variable; the date of each series; the number of each series; the number of right choices; the number of wrong choices; the number of times the wrong choices occurred when the standard was on the left; the number of times the wrong choices occurred when the standard was on the right; and the percentage of error for each series. Male IV was allowed to choose the brighter of the two illumi- nated areas. The intensity of the standard was .098 c.p. and at the beginning of the experiment the intensity of the variable was .002 c.p. After each successful series of 30 choices the light was shifted. 92 EUPHA FOLEY TUGMAN On November 21st the difference in intensity between the standard and the variable lights was shifted from .036 c.p. to .022 c.p. This seemed to be too large a step and confused the bird, so on November 28th the difference was placed back at .036 c.p. and the bird soon relearned that discrimination. Then when the difference in intensities was reduced in smaller steps the bird learned to discriminate the lights with differences much smaller than .022. In the descending series the least discrim- inable difference in intensity was .013 c.p. When the intensity was reduced to .009 c.p. the bird seemed wholly incapable of discrimination and very soon fell into the position habit. When the intensity was increased to .013 c.p. the experimenter was unable to break up the bird's position habit though it had dis- criminated the lights at this position in the descending series. Finally the lights were moved still farther back to .017 and after 14 days the experimenter succeeded in breaking up the position habit. The bird was again able to discriminate the illuminated areas. Hence the position midway between .013 and .017 c.p. is taken as the least discriminable difference. That position gives the difference in intensity between the standard and the variable as .015 c.p. Of the 599 wrong choices the bird made 223 when the brighter light was on the right and 376 when it was on the left. Female V did not give as good results. This the observer attributed to the fact that she was always so frightened that she would never hesitate long enough to make a discrimination. She apparently had little position habit but simply went to one side or the other in a confused manner. She made the neces- sary 30 consecutive correct choices with the difference in intensity .036 c.p. But when the position was shifted giving a difference in intensity of .030 c.p. she was not able to discriminate even after 540 trials. So the variable light was moved back, in- creasing the difference to .036 c.p. and after 600 trials she finally made 30 correct choices. Hence the least discriminable differ- ence for Female V is .033 c.p. Out of 434 wrong choices 305 were made when the brighter light was on the right and 129 when the brighter light was on the left. For Male V and Female VI the problem was very much more difficult. They had to overcome their natural tendency to go LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 93 always to the brighter light, and learn to choose the darker light. It took Male V 165 trials and Female VI 270 trials to learn the problem. Table V shows the record made by the two sparrows in learning the problem. The first column gives the number of each series of 15 trials. The second and fifth columns give the average percentage of error (series divided into small TABLE V Showing errors made in learning the first discrimination by the two birds which were trained to choose the darker light. The middle column in each group shows the percentage of errors made in each test. In the column to the left of the middle, the period of learning is divided into thirds and the average percentage of error is calculated for each third. In the column to the right is given the percentage of error for each half of the learning period. Despite occasional lapses as shown by increased percentage of error for single days, each of the later groups of days shows a decided lowering of the error percentage. This shows the gradual aspect of the learning process. MALE V FEMALE VI Number : Average of per cent Series of Error Percentage of Error Daily Series Average per cent of Error Average per cent of Error Percentage of Error Daily Series Average per cent of Error 1 2 3 4 45 60 40 40 40 40 33 42 40 13 13 42.1 50.6 42 42 60 40 60 60 40 40 20 66 6 26 20 13 26 20 44.8 5 6 7 8 38.7 15.4 33. 9 10 11 12 13 5.2 16.5 19.6 14 15 16 ■ 17 18 groups) for Male V and Female VI respectively. The averages show a decided decrease in the percentages of error made by the birds during the learning of the problems. The third and sixth columns show the percentage of error for the daily series of 15 trials each, for each of the. birds. The fourth and last columns give respectively the average percentage of error for the first 94 EUPHA FOLEY TUGMAN half and the last half of the trials required to learn the prob- lem. The figures also show a decided decrease in the percent- ages of error as the birds gradually learn the problem. The curve (Fig. 2) is plotted from these results. The ordi- nates indicate the percentage of error and the abscissa, the number of the series — each series consisting of 15 consecutive trials. The solid line ( ) is the error curve for Female VI and the broken line ( ) for Male V, while learning to go to the darker side. The table and curve show that the sparrows varied in their daily record as well as in the time required to learn completely the problem. The table shows also that the 130- So to •70 so / \ / \ / \ / \ 10 30 L— « i • 1 *• ^ \ \ \ 10 i to l t — \y 1 ! / / / V ' \ \ \ I i H . < 3 a i 1 1 1 1 3 f 4 1 J 1 1 1 l S 1 1 z c Figure 2. Curves showing percentage of errors made by two birds, Male V ( ) and Female VI ( ) while learning to choose the darker of two illuminated areas. Despite occasional lapses the learning appears to be a gradual process. birds exhibited great instability even until the end of the experi- ment. Male V made one wholly perfect series of 15 trials — the ninth series. Female VI did not make a perfect daily record until the 30 correct choices were made in succession. The progress of Male V toward the threshold was very slow. The observer felt that the bird really could discriminate the lights long before it gave perfect tests. But it was more hasty in making a choice than Female VI, with less attention and effort at discrimination. When Male V died he had been given 1515 tests and was working with the difference in intensities at .03 c.p. When Female VI had been given 1515 tests she was working with the lights at the discriminable difference .017 c.p. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 95 Female VI had reached a point much nearer the threshold than had Male V after the same number of trials. Male V made 295 wrong choices, 184 when the brighter light was on the right and 111 when it was on the left. The observer had expected Male V to give better results than any of the other birds because he was in the same cage with the other birds and in the same room where the experiments were conducted for two months and a half before he was experimented upon. He did learn the prob- lem more quickly than Female VI, but then fell behind her after the first change in the intensity of the lights. Male V died before the experiments were finished so his threshold of discrimination was not reached. Female VI was a very satisfactory bird with which to work. After she once learned to choose the darker of the two illumi- nated areas, she would always stop just a few seconds outside the door (D,, Fig. 1), look to one side and the other and then hop on calmly to the light she had chosen. She was never excited but always slow in her movements. She worked grad- ually toward the threshold. When the variable light reached 43 cm. with the standard at 38 giving a difference of intensity of .022 c.p. she discriminated immediately. The variable light was moved to 42 cm. thus giving a difference in intensities of .017 c.p. At this position she was given 720 trials, with 24 the percentage of error. But she apparently could not discriminate well enough to give two days of perfect trials. So the variable light was shifted back to 43 cm., giving a difference in the in- tensities of the lights of .022 c.p. At this point it took 450 trials before she gave 30 perfect trials in succession, while in the descending series she gave 30 perfect trials after having made only one wrong choice out of the 30 preceding trials. She had evidently been so puzzled with the lights when the difference in intensity was only .017 c.p. that she had either forgotten the problem or had formed the habit of not trying to discrimi- nate. Whatever the cause it took her 450 trials to learn the problem which she had learned in the descending series in 30 trials. Out of a total of 414 wrong choices 262 were made when the darker light was on the right and 152 when it was on the left. Hence the least discriminable difference for Female VI was .022 c.p., a difference of 5 cm. between the position of the stand- ard and the variable light. 96 EUPHA FOLEY TUGMAN (b) Comparison of results. Table VI shows a summary of the results for each bird. The first column gives the difference in intensity (c.p. Hefner units) between the standard and the variable lights in the decreasing series. The third column gives the number of trials it took each bird to learn to discriminate the lights at each position. The fourth column gives the number of correct choices and the sixth column gives the number of incorrect choices. The last gives the percentage of error of each bird for each intensity. TABLE VI Summary of results for each bird. Male IV and Female V chose brighter light Male V and Female VI chose darker light. Male V began discrimination with lights nearer together than did the other birds. Difference in Intens- ity (Hefner Units) between the Stand- ard and the Variable Light Bird Number of Trials Right Choices Wrong Choices Per cent •of Error .096 Male IV.... Female V. . . 30 30 30 30 .095 Female VI. . 240 150 90 37 .094 Male IV.... Female V . . . Female VI . . 30 45 15 30 42 13 3 2 6 13 .092 MaleV Female V. . . Female VI . . 75 90 30 70 84 30 5 6 15 6 .090 .\ Male IV.... Female V. . . Female VI . . . 30 30 15 30 25 14 5 1 16 6 .088 Male IV.... Female V . . . Female VI . . 30 30 75 30 30 65 10 13 .083 Male IV.... Female V. . . MaleV Female VI . . 30 30 165 15 30 30 113 14 52 1 31 6 .079 MaleV 30 30 .074 Male IV.... Female V . . . MaleV . Female VI . . 30 15 30 30 28 10 30 30 2 5 6 33 LIGHT DISCRIMINATION IN THE ENGLISH SPARROW TABLE VI— Continued 97 Difference in Intens- ity (Hefner Units), between the Stand- ard and the Variable Light Bird Number of Trials Right Choices Wrong Choices Per Cent of Error .055 Male IV.... Female IV. . MaleV Female VI . . 30 135 420 30 30 85 329 23 50 91 7 37 21 23 .052 MaleV 210 146 64 31 .050 MaleV 180 151 29 16 .036 Male IV.... Female V . . . MaleV Female VI . . 90 45 120 15 66 41 86 14 24 4 34 1 27 8 28 6 .030 Male IV.... Female V . . . Female VI . . 30 540 285 415 229 125 56 23 20 Back at .033 Female V . . . 600 496 104 17 .025 Male IV.... Female VI . . 15 30 14 1 6 .022 Male IV.... Female VI . . 90 30 66 29 24 1 27 3 .017 Male IV.... Female VI . . 225 720 181 557 ■ 44 163 19 22 Back at .022 Female VI . . 450 .013 Male IV.... 315 239 76 24 .009 Male IV.... 390 341 49 12 Back at .013 Male IV.... 360 306 54 15 Back at .017 Male IV.... 240 165 75 31 The following figures are made from this table. Figures 3 and 4 show the number of trials each bird required to learn to discriminate the light at the given intensity. The abscissae indicate the differences of intensity produced by the standard light and the various positions of the variable light. The ordi- nates indicate the number of trials required to learn the dis- crimination. Figure 3 gives the results for Male IV and Female V, both of which were allowed to choose the brighter of the 98 EUPHA FOLEY TUGMAN two stimulus areas. The solid line (- -) is the curve for Male IV and the dashed line ( ) is the curve for Female V. Figure 4 gives the results for Male V and Female VI. The solid line (- ) represents the results for Male V. The abrupt ending is due to the bird dying at this point in the in- vestigation. The dashed line ( ) represents the results for Female VI. Both Male V and Female VI, as stated before, were trained to choose the darker of the two illuminated areas. The curves for Male IV and Female V (Figure 3) are not widely different until the threshold was reached for Female V. They hoc As cent Si —K 1 DP.M rend TM}5 Serie \ H50 00 s ene c ) , JDe seen I'lTHjf Ser es iiO i i i Soo 1 1 t 1 i tso I ^0 3SCJ / 300 / / i 260 1 1 1 \ koa ) I 1 1 1 1 no \ 1 T)e£ d> 1 1 1 \ 100 iS \ / 1 1 1 io !'''- / \ \ J — — — - — ---/ '-' •0« Figure 4. Number of trials necessary for Male V (- -) and Female VI ( ) to learn each discrimination. The difference in intensity is shown on abscissae. Number of trials is represented on ordinates. Curve for Female VI shows large number of trials necessary when difference in intensity is small. shows that as the discrimination became more difficult, the bird required more trials to learn the discrimination. But when the difference in intensity of the illuminated areas was reduced to .009 c.p. Male IV was unable to learn the discrimination even after 390 trials. The difference in intensity was increased to .013 c.p. At this position he finally succeeded in learning the discrimination after 315 trials. But in the ascending series he was not able to learn to discriminate correctly even after 360 trials. The difference in intensity was increased to .017 c.p. 100 EUPHA FOLEY TUGMAN and after 240 trials he succeeded in giving 30 correct choices. The observer felt that the bird could discriminate the areas when the difference in intensity was .013 c.p. But when he was on the .009 intensity he acquired the position habit which lasted through the 306 trials at the next position and almost until the end of the investigation. Figure 4 represents the results of Male V and Female VI. The curves are more irregular than those for Male IV and Female V. The curve of Female VI is more nearly like those of Male IV and Female V than is the curve of Male V. The sudden drop in both curves indicates the fact that both birds learned to choose the darker of the two illuminated areas. After Female VI learned the first discrimination she had no trouble until the difference in intensity became small. She readily discriminated a difference in intensity of .022 c.p., but when the difference was reduced to .017 c.p. she was incapable of discriminating the areas even after 720 trials. The difference in intensity was then increased to .022 c.p. and after 450 trials she succeeded in dis- criminating the lights. Male V learned to go to the darker of the two illuminated areas more quickly than did Female VI, but his other results were never so good as those of Female VI. The observer felt that the discrimination was really not difficult for the bird but that many of his wrong choices were due to fright. He often did not even attempt to discriminate the lights. He would go to one light and if he failed, would hurry around to the other side and into the alley way. He was beginning to give better results when he died. (3) Incidental results. — The author has included under this heading some of the most interesting facts brought to light during the experiment with the birds. (a) Individuality. As shown in the previous discussion, the sparrows exhibited distinct individual differences in their rate of learning. This was also true of their general behavior and of their method of attacking the problem. The individual birds also varied their course of procedure at intervals during the experiment. All the birds were quick in learning the apparatus and its various alleyways. When they chose the wrong light and were LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 101 shocked they very soon learned to turn and go around to the other light. They were slower to learn to enter the alleyways because they were dark and the bird avoids dark places. But even this natural tendency was overcome in a surprisingly short time. The directness or indirectness of approach and entrance into the discrimination box varied in the different cases. Two of the birds would hesitate several seconds or even minutes before entering the discrimination box. But the other two seemed eager to go through the experiment. Male IV was calm and deliberate in his movements. He would hop into the doorway (Di, Fig. 1) as soon as it was opened. He would sit in the door- way or just outside for several seconds, turning his head quickly to one side or the other. (All the birds looked at the lights with their heads turned to one side or the other. They never seemed to look at the lights with both eyes at the same time.) Then he would hop calmly on, to whichever light he had chosen. He never seemed excited and seldom required shocking. When shocked it required a heavy discharge to produce any effect. Male V was always very much excited. For a long time he would dart out to one side or the other without attempting to discriminate the lights. Suddenly he changed his method and would hesitate a long time just beyond the door (D, Fig. 1). There he would sit, often for several minutes. Sometimes the observer would have to shock him before he would move. But usually after a long hesitation he would dart on to one light or the other. The observer felt that as a usual thing he did not rely upon visual discrimination. His judgments were never very satisfactory. At the beginning of the experiment Female V would always hesitate in the doorway, discriminate between the lights and hop on. But later she acquired the habit of darting out to one light or the other the very instant the door was opened. For a long time it seemed absolutely impossible to break up this habit. Frequently there would be days when the bird seemed to show some improvement but in general she gave very poor results. Finally the observer decided that she did not have such good sight as the other birds and that she had really reached her threshold of discrimination. The difference in the intensity of the lights was increased a little and after a short time the 102 EUPHA FOLEY TUGMAN bird began to discriminate them again. The previous behavior was evidently due to inability to make the discrimination. Female VI was more calm than any of the other birds. Her behavior was even more satisfactory than the behavior of Male IV. After Female VI learned the first problem, to choose the darker of the two lights, she would move about very slowly and deliberately. She would hesitate in the doorway of the entrance chamber for a few seconds, then hop out into the discrimina- tion box. She would hesitate several seconds looking a while to one side, then turning her head and looking to the other side. Presently she would hop on to the side she had chosen and into the alleyway. She was always slow making a choice and gave good results all through the experiment. The nervous birds gave the poorer results. Cole 13 thought the chicks which were most sensitive to the electric shock learned more rapidly under the influence of weak stimuli. The most sensitive sparrows did not learn most quickly. The two birds which gave the best results, Male IV and Female VI, both required a heavier shock than the other two birds. It was not necessary to shock them frequently. But when it was necessary, the observer had to give them almost twice the strength of current that was given the other birds. As stated before, the final results differed widely. Male V and Female V failed to discriminate lights with which the other birds had no difficulty. The observer thinks that Male V and Female V were inferior to the other two birds in acuteness of vision and prob- ably in mental capacity also. (b) Influence of former experiences. At first the birds were greatly influenced by their former experiences. They would tend to respond in the same direction as in the immediately previous test, provided they had not received a shock. If they had just received a shock they would nearly always go to the other light in the following test. But after the birds learned the problem their judgments were founded on visual discrimina- tion and they were not guided much by previous experience except when they acquired the position habit. (c) Position habit. The thing that gave the experimenter the most trouble was the tendency of the birds to get the posi- 13 Cole, L. W. The Relation of Strength of Stimulus to Rate of Learning in Chicks. Journal of Animal Behavior, Vol. 1, No. 1, 1911. Page 111. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 103 tion habit. The habit might appear at any time and the ob- server always had to be alert lest the habit get firmly fixed. If it did become fixed it was very difficult to break. It is evident the sparrow forms the habit of choosing by position much more easily than the habit of choosing by visual discrimination. The form of position habit which appeared most frequently was that of alternating from one side to the other without regard to the illumination. But the observer found that the birds learned also to go twice to one side, then twice to the other side, etc. A few times when the order required three choices in succession to one side, the birds would invariably want to go three times to the other side. The experimenter planned to try more com- plicated orders and see if the birds could be trained to learn the habit. Lack of time prevented her carrying out this part of the investigation. (d) Mental instability. The sparrows exhibited three quite distinct types of mental instability; all of which were sources of considerable trouble to the observer. At times the birds would go to one side or the other apparently without trying to discriminate. This is the first type. The second type of mental instability was the persistence in going to one side. When they would persist in this the experimenter found it quite impossible to do anything with them. The third type might be termed stubbornness or stupidity. When they would get stubborn or stupid they would simply refuse to move. They would pay no attention to an electric shock. The observer would force them around through the apparatus though the results obtained on such days were useless. The experimenter noticed that all the birds seemed to have these periods of mental instability though they did not all have them at the same time. One day a bird would record an almost perfect series and the very next day a series of failures. It was noticeable that such stupidity nearly always succeeded a period of unusually good work on the part of the bird. The bird usually recovered from this stupidity as quickly as it had come on. Then the bird would continue the usual method of procedure as though nothing unusual had happened. A few times they re- covered during a day's series and the last few results would be very satisfactory, while the first of the series were failures. The question which came to the mind of the experimenter was 104 EUPHA FOLEY TUGMAN whether or not these periods of stupidity or stubbornness come at regular intervals. But she was unable to decide from the data at hand. (e) Relation of time to failure or success of judgment. The human subjects thought that the hasty judgments were more often correct than the deliberate ones. The observer, as stated before, kept account of the time required to make each choice. Then she averaged about 500 of these times for each bird. The 500 times were taken as representative of the total data. They include the time required to make the choice for each of a day's series selected at intervals throughout the whole experiment. These times were averaged — the time of the correct choices in one column and the time of incorrect choices in another' col- umn. The results are shown in Table VII. TABLE VII Relation of time of choice to failure and success of judgment. Average of 500 choices taken at random throughout the experiment. Correct choices required less time than incorrect ones except in case of Female V, all of whose choices were very rapid. Male. IV Time of correct choices average 28 seconds. Time of incorrect choices average 30 seconds. Female V Time of correct choices average 18 seconds. Time of incorrect choices average 17 seconds. Male V Time of correct choices average 31 seconds. Time of incorrect choices average 47 seconds. Female VI Time of correct choices average 1 min. 20 seconds. Time of incorrect choices average 1 min. 49 seconds. Female V made the quickest choices and gave the poorest results. Male IV gave the best results and made rather hasty judgments. Female VI gave the next best results and was the slowest of them all in making her choices. It is noticeable that, with one exception, the time required to make the choice was shorter for the correct choices and longer •for the incorrect choices. This coincided with the opinion of the human subjects. However, there were too few birds experimented upon to draw any definite conclusions as to the relation of the time required to make a choice and the success or failure of the judg- ment. This is a most interesting problem for further investi- gation. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 105 TABLE VIII All records for the day before and the day after a forty-eight hour interval. The day before was usually Saturday; the day after Monday. The average for the day after is in every case greater than that for the day before the long interval. Male IV Female V Male V Female VI. Per Cent of Error Per Cent of Error Per Cent of Error Per Cent of Error Saturdays Mondays Saturdays Mondays Saturdays Mondays Saturdays Mondays 6 6 60 40 90 90 13 6 13 33 46 10 40 6 6 20 40 20 13 30 13 6 6 53 33 26 13 6 13 46 60 60 26 20 26 13 40 33 40 26 20 33 33 26 26 46 26 26 6 26 46 20 26 46 26 33 26 20 33 26 13 20 6 53 33 13 20 13 13 20 26 26 20 40 20 20 26 26 26 40 40 40 6 26 26 33 33 46 46 26 20 46 13 33 13 6 26 20 46 13 13 20 33 20 20 26 26 33 20 26 13 13 13 26 40 53 26 20 40 53 33 20 40 6 33 33 20 6 40 13 13 26 20 46 40 33 33 40 26 20 40 46 20 33 26 26 53 46 20 13 13 40 53 13 26 20 46 33 20 26 33 46 20 26 13 20 46 6 26 13 33 20 53 33 13 26 20 40 40 53 26 40 13 53 33 13 20 26 33 33 13 20 6 13 26 26 40 • 6 6 26 6 13 26 Ave. Ave. Ave. Ave. Ave. Ave. Ave. Ave. 26.97 28.8 15.81 21.35 20.91 21.91 15.73 18.73 Difference between Saturday and Monday record. 1.83 5.54 1.00 3.00 106 EUPHA FOLEY TUGMAN (f) Effect of 48 hour interval between successive series. Table VIII shows the effect of 48 hours interval between successive series. As stated before, one series per day was given each bird except on Sunday and an occasional holiday. So this table gives the percentages of error for the series on the day preceding the holiday and the percentage of error for the day succeeding the holiday. In a few cases the position of the variable was changed over the holiday and so these cases were not considered. For Male V the average percentage of error for Saturday and days preceding holidays is 26.97, while for the day following the 48 hours' rest, his percentage of error is 28.8 or an increase of 1.83%. For Female V the average percent of error for Satur- day, etc., is 15.81% and for Monday the average is 21.35%, which shows an increase of 5.54%. Male V has an average per cent of error on Saturday of 20.91% and on Mondays an average per cent of error of 21.91%, which is an increase of just one per cent. Female VI has an average per cent of error on Saturdays of 15.73% while on Mondays her average per cent of error is 18.73%, which shows an increase of three per cent. To recapitulate: For Male VI Mondays show an increased percentage of error of 1.83%; Female V, 5.54%; Male V, 1%; and Female VI, 3%. Thus each of the birds shows the effect of an extra 24-hour interval between two series, by an increase in the average percentage of error for the following day. (g) Persistence of stimulation. It has been the experimenter's observation that the sparrows do not retain the effects of stimu- lation very long at a time. Frequently, as stated before, the birds would get the position habit of alternating from one side to the other. One day, when this occurred, the series was inter- rupted as the experimenter was called out of the room. Upon returning to the experiment the observer noticed that the bird hesitated before choosing between the lights. He then chose the same light as in the test preceding the interruption. Then in the following tests he alternated from one side to the other as before. So the alternation was in regular order with the ex- ception of the one break due to the interruption. After that when a bird acquired the habit of alternating from one side to the other, the observer would stop the experiment several minutes. In practically every case the bird would hesitate in mak- ing a choice in the test following the interruption and usually LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 107 this caused a break in the order. It looked as though he had forgotten which light he had chosen before. It would be inter- esting to experiment further in detail upon this question. But from the very limited material at hand, the observer is inclined to think that the birds do not retain visual impressions very- long at a time. IV. QUESTIONS SUGGESTED BY THIS INVESTIGATION. The question arose as to whether or not the birds would have reached their threshold of discrimination in less time if the inter- mediate steps were omitted. That is to say — train the bird to dis- criminate the lights with wide differences in intensity. Then make the difference very small and see if the bird could not learn to make the discrimination in less time than was required to pass through all the intermediate steps. The observer thinks that a great number of the intermediate steps might be omitted in the earlier part of the experiment when the differences in the intensities of the lights were large. Female VI and Male V were trained to choose the darker of the two lights. Female VI learned to choose the darker of the two lights when they were 162 cm. apart, which made the darker of the two lights very dim. Male V learned to choose the darker of the two lights when they were 60 cm. apart and learned the problem in a shorter time than did Female VI. It might have been better to have omitted all the work with the wide difference in intensity for Female VI and begun where Male V began. The results of Male V might have been due to the fact that the darker of the two lights was not so dark as it was for Female VI. It may have been easier for the bird to learn to choose a light of medium intensity, than it was for Female VI to learn to choose a light of extremely low intensity. However, the small steps seemed to be essential as the birds neared the threshold. In one case the observer de- creased the difference in intensity too much by one step. The bird seemed absolutely unable to make the discrimination after it had been at the problem a long time. But when the lights were shifted back to their former position and then the differ- ence in intensity decreased step by step, the bird experienced no difficulty in making even finer discriminations. So it seems that the tests at intermediate intensities were essential and that practice was an important factor in learning. 108 EUPHA FOLEY TUGMAN Another question was raised. When perplexed for a long time does the bird forget or unlearn the problem? The observer is inclined to think that it does. Male IV learned to discriminate the lights when the differ- ence in intensity was .017 c.p. after 225 tests. The difference in intensity was decreased to .013 c.p. and he learned the dis- crimination in 315 tests. The difference was again decreased to .009 c.p. The bird was completely* perplexed. Each time he went to the right side. 390 tests were given him but he could not learn to discriminate so small a difference in intensity. The variable light was shifted back so the difference in intensity was .013 c.p. He had correctly discriminated this difference be- fore in 315 tests. But now he continued to go every time to the right side. After 360 tests the difference in intensity was again increased — this time to .017 c.p. He had previously learned this discrimination after 225 tests. The bird still persisted in going to the right side. The observer finally decided that the bird had forgotten the problem. So the difference in intensity was made extremely large. The bird correctly discriminated the lights in 15 tests. The lights were then shifted back to .017 c.p. and the bird took 240 trials in learning the discrimination which had only required 225 trials in the descending series. The results therefore, seem to indicate that the bird was perplexed so long that he really forgot or unlearned the problem. Another question which suggested itself was whether or not the birds would learn the position habit if the order was very complicated. They soon acquired the habit of alternating from right to left. A few times the experimenter observed that the birds learned to go twice to one side and twice to the other and three times to one side and three times to the other. These latter cases might have been merely accidental. It would cer- tainly be an interesting problem to try various orders and see if the bird could learn them. Also, could the bird acquire the position habit if it did not have the light to guide it, i.e., if the lights were of equal intensity? The observer is very much interested in the question of whether or not these periods of stubbornness or stupidity, which all the birds seem to have occasionally, occur periodically. She was unable to tell from the data at hand. LIGHT DISCRIMINATION IN THE ENGLISH SPARROW 109 Before concluding the author wishes to express her deepest appreciation of the valuable suggestions, criticism and encourage- ment of Dr. M. E. Haggerty, who so kindly suggested and directed this experiment. She is also greatly indebted to Dr. Haggerty, Mr. William O. Trapp and Mr. George H. Hyslop, who were the human subjects in this investigation. THE ROLE OF RANDOM MOVEMENTS IN THE ORIENTATION OF PORCELLIO SCABER TO LIGHT HARRY BEAL TORREY and GRACE P. HAYS Reed College, Portland, Oregon 1 In his admirable paper on " The Selection of Random Move- ments as a Factor in Phototaxis," Holmes 1 has given great significance to random, that is, spontaneous, non-directive movements in the orientation of earthworms, blow fly larvae, and leeches, to light. As he carefully watched the movements of these organisms under the influence of light, it " soon devel- oped that what seemed at first a forced orientation, the result of a direct reflex response, is not really such, but that the orien- tation which occurs and which is often quite definite is brought about in a more indirect manner by a mode of procedure which is in some respects similar to the method of trial and error fol- lowed by higher forms." The organism becomes oriented by following up those random movements which bring them away from the source of light. While our experiments on the larvae of an undetermined species of blow fly and on a species of earthworm (Allolobophora sp.) materially lessen for us the importance of random move- ments as a factor in the orientation of these organisms to light, our conclusions are in complete accord with Holmes' view that the type of reaction he describes " differs from Jennings' ' motor reflex ' by which many of the so-called tropic reactions are produced in the Protozoa." This difference has little signifi- cance for Mast 2 who believes that " the only difference between the orienting reactions in the two classes of animals mentioned is that the unicellular forms studied by Jennings always turn toward a structurally defined side, while the metazoa investi- Journal of Comparative Neurology and Psychology, 1905, No. 15, p. 18. Light and the Behavior of Organisms, 1910, p. 51. 110 THE ORIENTATION OF PORCELLIO TO LIGHT 111 gated by Holmes are not thus limited in their direction of turning." In thus minimizing a difference to which Holmes has explic- itly called attention, Mast may have missed a cardinal point in Holmes' illuminating discussion. The direction of the random movements of the blow fly larvae as observed by Holmes is not predictable so far as it bears no definite relation to the source of light. The direction of the movements of Euglena, an organ- ism in which the 'motor reflex' plays an important part in its orientation to light, is predictable, since it does bear a defi nite relation to the source of light. In the one case, the orient- ing movements, made at random, are not controlled, as to direction, by the light; in the other case, the orienting move- ments are definitely controlled, as to direction, by the light. In the former, selection operates among so-called trial move- ments; in the latter, in so far as the movements are controlled or forced by an external agency, the method of trial is excluded. This difference, then, is of no little significance in an attempt to determine — as this paper is attempting to determine for certain organisms — the actual value of the orientation hypo- thesis that rests upon the assumption of trial movements. The fact that some authors do not distinguish between ran- dom movements and directive movements forced by the en- vironment has been a source, of some confusion in the literature of animal behavior. Further confusion has centered about the conception of symmetrical stimulation repeatedly emphasized by Loeb and recently reaffirmed by Parker. 3 Investigators of the orienting reactions of non-symmetrical protozoa or symmetrical organisms such as rotifers and worm larvae that swim, like the protozoa, in spiral courses, have had difficulty in seeing the applicability of this conception to their material. That the conception is applicable, however, to the behavior of such organ- isms as Euglena, though not in the form apparently anticipated by some of its critics 4 a recent paper 5 has attempted to show. And its applicability to the orientation of earthworms and blow Journal of Animal Behavior, 1911, No. I, p. 461. 4 Mast, 1910, p. 85. Torrey, Science, No. 38, p. 873. 112 HARRY BEAL TORREY AND GRACE P. HAYS fly larvae to light has been convincingly discussed by Parker in the paper just mentioned. These conceptions of symmetrical stimulation and of forced directive movements have long characterized the tropism hypo- thesis, whatever other attributes it may be said to possess; and they appear to be quite inconsistent with the conception of orientation by the selection of trial reactions. There should be little danger of confusion, then, in designating as tropic reac- tions not only the very gradual turning movements that may or may not be connected with tonic contractions accompanying constant stimulation, but also the more abrupt and angular turning movements composed of a series of forced shock reac- tions, all in the same general direction, that we have repeatedly observed' in the orientation of Euglena to light. Both extremes are, in fact, represented in the behavior of Euglena, which will be considered in another paper. Whether they also represent two different mechanisms of orientation is a question for the future to decide. 6 That they do not involve the selection of random movements there appears to be no doubt. In the following account of the reactions of Porcellio scaber, it will be seen that although random movements are common they can readily be distinguished from the forced movements that occur in definite predictable directions in response to dif- ferential stimulation of symmetrically situated photoreceptors. But such phototropic movements not only exist; they are large factors in the orientation of Porcellio to light. This is true also for Allolobophora sp. and the larvae of an undetermined blow fly. Porcellio scaber, a species of sow-bug, or wood louse, very common on the Pacific coast, is a typical symmetrical isopod with a pair of compound eyes set far apart in the head segment, and two pairs of antennae, of which the second antennae are conspicuous tactile organs, restlessly active during locomotion. The subequal walking appendages and the body in general are also sensitive to contact stimuli. Of other sense organs it is unnecessary now to speak. • Since this was written, a paper by Dr. F. W. Bancroft, in the Journal for Ex- perimental Zoology for November, 1913, appears definitely to have settled the question, for Euglena at least, in the affirmative. THE ORIENTATION OF PORCELLIO TO LIGHT 113 During, the day Porcellio is usually found under stones, logs, rubbish, in dark cellars, and various other sorts of cover from the light of the sun. Correlated with this habit is a definite negative phototropism. In our first experiments, this phototropism was more or less masked by large individual differences in sensitiveness to light, and the apparent indifference of many individuals to light com- ing from incandescent bulbs placed directly in front of them. Later we discovered that the locomotion of many such indif- ferent individuals could be controlled with great definiteness by holding an incandescent bulb behind them, as they marched over a dead black table top, and moving it to one side or the other. Under these conditions — Mazda bulbs of both 25 w. and 60 w. were used— the organisms would move away from the light with the precision of a boat answering the helm. They could be guided in circles, in spirals, in courses that were di- rected, now to the right, now to the left, at the will of the experimenter. That the eyes were the organs responsive to light was demon- strated by blinding them with a mixture of charcoal and glue. Individuals with the right eye blinded reacted to light from the left only; when the left eye was blinded light from the right was alone effective; when both eyes were blinded the individ- uals thus treated were indifferent to light from any direction. Porcellio responds not only to changes in the direction of light. Exposure to light stimulates into activity animals that in darkness are quiescent; though sudden changes in intensity of illumination may produce inhibitory effects. Individuals vary considerably in their responses to these and all other types of stimulation. Marked differences may exist between individuals of the same size and apparently the same age; also between the reactions of the same individual at different times. Age differ- ences are frequently connected with different reactions. Very young, unpigmented individuals are more responsive to direc- tive stimulation than old. It is the rare exception for them to fail to respond, although adults are not uncommonly refractory. To sudden changes in intensity of light, however, old react at least as sensitively as young. In this connection the following case may be cited. 114 HARRY BEAL TORREY AND GRACE P. HAYS A female with a full brood pouch was placed in a Petri disk, round and round which she proceeded to move in the light of a 25 w. tungsten bulb. Many times when she was facing the light, the latter was turned off. Invariably she came at once to a dead stop. Only occasionally when the light was turned off while she was facing away from it would she react similarly; being obviously less responsive in such cases. Sudden increases of intensity, (i.e., when the light was turned on) always pro- duced definite inhibition of locomotion. One of the brood of this female responded but rarely to sud- den increases of intensity when facing the light, not at all to sudden decreases and never while going away from the light. The fact that young are more readily directed in locomotion by light while they appear to be at least no more sensitive than adults to sudden changes in intensity of light, suggests the possibility of two mechanisms governing the two types of reac- tion. There is a wide variation in the responsiveness of adults to sudden changes in intensity, however. The problem pre- sented here will be investigated further. 5 Though the eyes of Porcellio are sensitive to light, their power of forming images is approximately very small. Totally blind individuals avoid obstacles with the ease of normal individuals. When the second antennae of either are removed, however, they often bump squarely into obstructions, avoiding them only after contact through legs or body. The importance of the second antennae is emphasized by their constant activity dur- ing locomotion, when, by a rapid succession of tappings on the substratum, and wavings in the air, they explore the region immediately to the front. The usual random movements that are made by the anterior end of the earthworm and blow fly larvae are in Porcellio restricted to these mobile antennae. Since the head segment of Porcellio does not move perceptibly from side to side, it is only necessary to amputate the second antennae to eliminate what correspond to the usual random or trial movements in earthworm and blow fly larvae. Such an operation was made in several cases. It was soon found, however, that, with or without the second antennae, Porcellio responded to photic stimulation under the conditions THE ORIENTATION OF PORCELLIO TO LIGHT 115 of our experiments with unequivocal, definite, tropic reactions. So the operation was discontinued as useless. In later experi- ments on blow fly larvae and Allolobophora, the same definite tropic reactions were observed. 6 For the sake of clearness it should be pointed out not only that " random movements ' and ' trial movements ' are ex- pressions not always used in the same sense, but that appar- ently spontaneous random movememts may be controlled to some extent by the environment. The exploring movements of the second antennae of Porcellio are largely initiated and regulated by internal conditions; this is evident especially when environmental conditions remain constant. A slight change in the texture of the substratum, however, may produce marked changes in its behavior; in the absence of the antennae, contact differences may make themselves effective upon the path of locomotion through the legs or body. It is a truism that the behavior of an organism is a resultant of the responses to all simultaneously acting stimuli. A movement initiated from within, when the organism is exposed to various contact stimuli, may frequently be modified if not entirely inhibited by them. The same may be said of movements initiated from without. It happens, therefore, that so-called " trial " movements in Porcellio and blow fly larvae and earthworms vary their char- acter and intensity with circumstances. They may be so aug- mented by external stimuli as largely to obscure the tropic reactions which under other conditions are readily perceived. The source of the external stimulation may, however, be very inconspicuous. This was especially true in the case of a blow fly larvae that had been traveling away from the light in a direct course with very slight lateral movements of the anterior end. Suddenly the anterior half of the body was lifted and swung from side to side, up and down, in irregular movements of large amplitude that continued for several seconds. The cause of this change in behavior was finally discovered in a bit of filament from the paper substrate that had been picked up and was adhering to the anterior end. For the time, these vigorous " trial " movements, initiated probably by internal con- ditions but owing their conspicuous characters to contact stim- 116 HARRY BEAL TORRE Y AND GRACE P. HAYS uli, effectively masked the heliotropic movements so apparent under other conditions. Similar pronounced movements were frequently seen when a larva, crawling out over the edge of the glass plate on which it was being observed, would free the anterior third or half of its body. It would then wave this free portion about much in the manner of a leech. Dryness of the substratum may produce similar effects. Such behavior suggests the probability that even the small random or trial movements of the anterior end that ordinarily accompany locomotion are controlled — their amplitude, perhaps being determined — to some extent by contact stimuli. 7 It is possible then, to distinguish between random movements that have no connection with photic stimulation, and movements that Mast calls trials, but are conditioned by photic stimula- tion. For convenience in further analysis, it will be desirable to distinguish between two groups of reactions thus conditioned. In the one may be placed reactions to high intensities of light, such as direct sunlight; in the other, reactions to lower inten- sities. All of these reactions are regarded by Mast as trial movements similar to the avoiding or shock reactions of the lower organisms. The reactions of the second group — however we may view them as " trials " — do indeed resemble those reac- tions of such a form as Euglena, that are in the same general direction with reference to the source of light. The reactions of the first group, however, occur either toward or away from the source of light. They are non-directive with reference to the source of light. This distinction is emphasized by our observations on earth- worms and fly larvae. When light was allowed to fall from the side upon the extended anterior end of either of these forms, the first movement of the anterior end was for certain intensi- ties of light away from the latter, whether directed toward or away from the light, when exposed. 8 To eliminate as far as possible all non-directive reactions from the behavior of Porcellio to light, in order to discover any directive, tropic movements of orientation that might be present, we adopted two very simple methods. The first consisted in THE ORIENTATION OF PORCELLIO TO LIGHT 117 exposing sensitive individuals suddenly to lateral illumination. The individual to be observed was placed on a smooth dead black ground, in a dark room. When its orientation had been accurately determined by means of a 25 or 60 w. tungsten bulb a few inches behind it, away from which it was moving, or a distant light in the ceiling, another tungsten bulb of either 25 or 60 w. and at different distances varying between 20 and 40 cm., was suddenly turned on, so that its light should strike the animal from the side at an angle as near ninety degrees as possible. Sometimes at the instant the lateral light was turned on, all other lights were extinguished; at other times, they were not. In both cases, the direction, with reference to the lateral light, of the first movement of the organism out of its course was determined. These experiments, simple as they were, gave results that were strikingly definite and convincing. Almost invariably- the first movement was away from the lateral light. The reaction was sharper, on the whole, when light came from the side only. To the 60 w. light, at 40 cm., the response was more definite than to the 25 w. light at the same distance. But the reaction was unmistakably negative within the limits of variation of lighting and distance mentioned. A significant feature of the results was the ease with which they were obtained and the simplicity of means employed. It must be remembered that all individuals are not equally sensitive to light. But the consistency with which many indi- viduals turned away from. the light, whether the latter was on one side or the other, left no room for doubt that the reaction was forced in a definite direction. 9 The second method of experimentation, equally simple, was determined by the fact that many individuals responded more readily to light coming from behind than from in front of them. The following series of observations taken one afternoon are not selected, but indicate the reactions of the first individuals tested. The lamps used in these experiments gave a source of light 4 to 5 cm. in diameter. This fact it is important to keep in mind when considering the definiteness of the responses of Porcellio for the smaller angles of incidence recorded in the tables. For 118 HARRY BEAL TORREY AND GRACE P. HAYS instance, at 70 cm. from the organism the light used possessed an angular diameter of 4°; at 36 cm., 7°; at 50 cm., 6°; at 15 cm., 16°. I. A 25 w. tungsten bulb gave the light at approximately 70 cm. from the animal. The latter was a medium-sized adult. Since the sexes respond similarly to light, no account was taken of sex in this and the following experiments. Having deter- mined the orientation of the animal by means of a 60 w. bulb behind it, this bulb was turned off as the 25 w. light was flashed upon it, from in front, striking the eyes of the animal so as to make an acute angle with the axis of the body. Trial 1, Light 35° to left ; response to right. 2, 15° ; animal stopped, wavered, and turned to right. " 3, " 70° " " ; animal stopped, then turned to right. " 4, " 60° " " ; response to right. 5, 10° " ; animal stopped, moved forward, then to left (toward light). " 6, " 5° " " ; same as 5. " 7, " 3° " " ; same as 5. " 8, " 5° " right; animal turned to left. These trials show a tendency in the organism to turn away from the light, the direction in which the turn is made depending upon the' position of- the light and the angle at which it strikes the eyes; there is a stronger tendency to turn to the left than to the right, but this is overcome when the light from the left strikes the eyes at an angle with the body axis of 15° or more. The same tendency to turn more readily to one side than to the other is seen in the next series; though here the organism turns more readily to the right. II. Another individual. Lights as in Series I. Trial 1, Light 10° to left, 36 cm. distant 2, 3, 4, 5, 6, 7, 8, 35° 5° 12° 5° 5° 45° 30° 36 right, 36 36 70 left, 70 right, 70 70 Response to right, distant. Response to right, distant. Response to right, distant. Response to right, distant. Response to right, distant. Response to right, distant. Response to left, distant. Response to left. As in the first series, the organism turns away from the light, either to the right or left, when light strikes it at an angle greater than a certain magnitude, in this case between 12° and 30°. When the light strikes it at an angle of 12° or less, the organism THE ORIENTATION OF PORCELLIO TO LIGHT 119 turns toward the light in the definite turning movement that ultimately carries it away from the light. The two series of trials just presented suggest a difference in .the sensitiveness of the two eyes to light. Tests of each individual by means of a light shining upon it from behind, fully supported this view. The first individual was guided without fail to the left when the light came from behind at a "small angle to the right ; but the same individual did not respond with such definiteness to light coming from behind at a similar angle to the left. These statements will apply equally well to the second individual, if the directions are reversed. III. Another individual, young, unpigmented. Lights as before. response to left, response to right, response to right, response to right, response to right, response to left, wavered, then left, wavered, then left, response to left, response to left, toward light, then left, response to right, response to left, response to right, response to left, response to right, stopped, then to left, response to right, wavered, then to left, response to left, wavered forward. This series brings out the fact that although the individual responds to light as an approximately symmetrical animal, its reactions lose precision when the light rays fall upon it from the front at very small angles (e.g., five degrees or less) with the axis of the body. The following record of another individual bears directly upon this point. Preliminary tests showed that this individual, almost symmetrically sensitive to light, responded toward the right a bit more readily than toward the left. A 60 w. Mazda lamp was used, about 15 cm. in front of the animal, a given number of degrees of arc to the right or left as the case might be, Trial 1, Light 25° to right, 36 cm. distant ' 2, tt 3° " left, 50 " distant ' 3, it QO i* <* 50 tt distant ' 4, « t 5° " " 50 i i distant ' 5, tt 5° " right, 50 n distant ' 6, a 5° " " 50 n distant ' 7, n go « « 50 1 1 distant ' 8, 1 1 5° " " 50 n distant ' 9, a 15° " " 50 tt distant ' 10, tt 5° " " 50 tt distant ' 11, tt en face, 50 " distant ' 12, " 5° " " 35 a distant ' 13, tt 5° " " 50 a distant ' 14, a 10° " left, 50 " distant ' 15, a 8° " right, 50 tt distant ' 16, a 10 " left, 50 a distant ' 17, a 5° " " 50 " distant ' 18, i t 10° " " 50 1 1 distant ' 19, tt en face, 50 a distant ' 20, 1 1 5° to right, 50 a distant ' 21, a 5° " " 50 tt distant 120 HARRY BEAL TORREY AND GRACE P. HAYS Light 5° to right; 5 trials. Responses to right, 2; left, 3. " 5° ' ' left; 6 " ** «t c < o ' 1. " 10° ' ' right; 6 " 2 ' ' 4. " 10° ' ' left; 5 " " 4 • ' 1. " 15° ' ' right; 7 " ' ' 7. " 15° ' ' left; 6 " " 6 ' ' 0. " 20° ' ' right; 6 " " ' ' 6. " 20° ' ' left; 5 " " 5 ' ' 0. It appears from these observations that while the initial locomotor response might be toward the light in a small per- centage of cases, such responses occurred only when the rays of light made an angle of less than 15° on right or left with the body axis. This is not surprising when one remembers the large angular diameter of the source of light in this experiment. Be- yond 15° the response was consistently away from the light. Further, in the few cases when the response was at first toward the light, the animal continued to turn toward the same side until it ultimately moved away from the light. These excep- tional cases, then, only emphasized the negative phototropism of Porcellio. SUMMARY 1. Reasons are given for considering every orienting reaction phototropic whose direction is predictable in that it bears a definite relation to the source of light. Euglena viridis, species of blow fly larvae and earthworms, and Porcellio scaber exhibit reactions of this type, which is not satisfactorily interpreted by. the method of trial. 2. Porcellio is easily guided in any desired direction by chang- ing the direction of light falling on it from behind. 3. The first locomotor movement made by Porcellio, when exposed suddenly to light striking it at an angle of 90° with the major axis, was away from the light. 4. The same pronounced negative reaction followed sudden exposure to light from the front at angles between 90° and 15°. 5. When exposed suddenly to light coming from the front at angles less than 15°, Porcellio moved with less consistency away from the light; but the reactions were, on the whole, markedly negative. This lack of consistency was referred partly to the relatively large angular diameter of the source of light, partly to demonstrable inequalities in the sensitiveness of the two eyes of certain individuals to light. MALE DOVES REARED IN ISOLATION WALLACE CRAIG The University of Maine Eight of my Blond Ring-doves 1 have been reared in isola- tion, being removed from their parents after the age of weaning but long before the age of maturity, and being brought to maturity in cages where, though they could sometimes hear other doves, they could never touch nor see them. The original intention was to rear each dove out of ear-shot as well as out of sight of all others of its species; but since this would require the keeping of each dove in a separate building, with a quarter- mile or more between buildings, it was found to be impractic- able. Of the eight doves reared in isolation, the present article will give the history of three males, Jack, No. 22; Billy, No. 23; and Frank, No. 30, and brief mention of the fourth male, No. 39. Jack, No. 22. Hatched July 17, 1907. Removed from his parents August 17, his 32nd day. Throughout the autumn and early winter this bird cooed very little. But about the first of February there began a re- markable development of voice and social behavior. The dove was kept in a room where several men were at work, and he directed his display behavior toward these men just as if they belonged to his own species. Each time I put food in his cage he became greatly excited, charging up and down the cage, kahing 2 and bowing-and-cooing to me, and pecking my hand whenever it came within his cage. From that day until the day of his death, Jack continued to react in this social manner to human beings. He would bow-and-coo to me at a distance, or to my face when near the cage; but he paid greatest atten- 1 For a general account of the social behavior and life-history of this species, see Craig, W., The Expressions of Emotion in the Pigeons. I. The Blond Ring-Dove {Turtur risorius). Jour. Comp. Neurol, and Psychol., 1909, Vol. 19, pp. 29-82. 2 " Kah " is the name I use for the well-known cry, sounding like a laugh, which has won for this dove its specific name risorius. 121 122 WALLACE CRAIG tion to the hand — naturally so, because it was the only part with which he daily came into direct contact. He treated the hand much as if it were a living bird. Not only were his own activities directed toward the hand as if it were a bird, but he received treatment by the hand in the same spirit. The hand could stroke him, preen his neck, even pull the feathers sharply, Jack had absolutely no fear, but ran to the hand to be stroked or teased, showing the joy that all doves show in the attentions of their companions. Growing up in isolation from all compan- ions of his own species, he gave himself completely to the com- panionship of human beings. July 7, 1908, when Jack was almost a year old, I put an end to his isolation. I tested him (and Billy also) with birds of different kinds, to see if he would choose his own kind. The results were positive, but I have discovered a possible flaw in the conditions of the experiment. When I have pigeons of several species, as I have not now, I shall repeat this experiment on species recognition after rearing in isolation. July 12, 1908, I placed Jack's cage beside the cage of dove No. 19, a virgin female, in order that they might become ac- quainted. They had seen each other a little during the pre- vious few days. When the cages were placed side by side, each dove at once showed excited interest in the other, and the female repeatedly gave signs of a desire to mate with the male. July 14, I let Jack into the cage of the female, by opening the doors between the cages, this being the first time since his infancy that Jack has come into contact with another dove. He went into her cage without hesitation, and soon began to peck and chase her. He had long been accustomed to pecking my hand, but now when he made his first peck at a dove and his bill closed on the feathers, he stopped in evident surprise and did nothing more for a few seconds. But ever after this first experience, he tugged and shook the female dove's feathers as an old male does. On this first day of contact his attitude toward the female was that of cruel pugnacity, never showing any considerable tenderness or eros. I was obliged to close the doors, preventing contact of the two birds, out of mercy to the female. Next morning the two were lying as near together as they could, in their separate cages, in apparent love and friendship. Never- MALE DOVES REARED IN ISOLATION 123 theless, after four more days of such acquaintance, when on the afternoon of July 18th I again let him into her cage, he was again cruel to her, though he did also nest-call to her. I closed the doors between them after ten minutes. A main reason to be given for Jack's cruelty to the female, is that he regarded me, the human being, in some degree at least, as his mate; the female dove was, therefore, in so far regarded as an interloper, to be attacked and driven away. But from July 14th, when Jack first came in contact with a dove, he began to divide his attentions between human beings and doves. He ceased to kah and bow-and-coo to me; though until the day of his death he remained as tame as ever, and al- ways pecked the hand that was put in his cage. In regard to the sexual reaction, Jack's behavior was most remarkable. He never showed this reaction at all, so far as I observed, until he came into contact with another dove. And then he showed (at first) no tendency to unite with that dove. But it appeared nevertheless that the dove had aroused his sexual impulse; for on July 18th, when food was put in his cage, Jack pecked the hand that was putting in the seed-cup, then assumed that peculiar erect posture which precedes copu- lation, jumped on the hand, and began to go through the move- ments of copulation, an act he had never done before. "July 19, 10:40 A. M. I open the doors, letting male into cage of female. He chases her savagely, bites, pulls feathers. After one minute he goes to nest-calling, but soon savagely chases her again. " He then bows-and-coos a great deal to me. Also, he makes a curious little flight upward, and a feint as if to alight on back of female. I suspect that he is seeking copulation, but is seek- ing the hand as his stimulus; so I put hand in the cage, and find that my surmise is correct. He does not tamely jump on the hand, he flies on it; then he begins to go through the move- ments of the sexual reaction. Soon as this was put beyond question, the hand was taken away from him." The male then went back to the nest site, sounded the nest- call, allowed the female to come to him and caress him, preened her head a little in return, and thus worked himself up to another crisis of excitement. When the crisis came, he again made not the slightest attempt at intercourse in the normal manner, but 124 WALLACE CRAIG made frequent flights upward in the direction of the female, or in other directions, exhibiting high excitement and some be- wilderment. ' When he flies up he always hovers, sometimes over the female, sometimes over he knows not what." Some- times he tries to reach me, sometimes not. After such an abor- tive attempt at venting his passion "he chases the female as savage- ly as ever, or more so, even jumping on her back. Then they nest-call again. Then he chases her again." They reach the stage of attempting to bill, which stimulates the male so that he dashes up again in his passionate, hovering flight. " Thus they repeat and repeat." July 20, both forenoon and afternoon, I let the birds come together for a time, with the same results. July 21, I let the birds come together, and put a nest in the cage. Once when the male was in the nest and the female at the other end of the cage, he ' nest-calls to her a few times, then suddenly he lifts himself and glares at her, chases and worries 'her a long time, even pulling feathers out. Four times during this onslaught he makes one of his peculiar passionate flights; the first three times the flight was toward female, as if with vague notion of alighting on her; the second time he did alight on back of her neck but did nothing more. The fourth time, in contrast, was a flight high (nearly two feet?) in air." The same day the female began to sit in the nest, preparing to lay. This fact checked the male somewhat in his activity toward her; as it does every male. But the effect on this male was interesting, thus: " Male jumps toward female, then turns toward me, then pecks female, then runs toward me. Gives it up. Soon at it again; drives female off nest and pecks her many times, then tries to get to me, thus back and forth for long period. Apparently it was his failure to reach me that drove him each time to the female; then his contact with her restimulated him so that he wanted to come to me." The first experience with a nest I shall describe in another article, dealing with many birds. The first egg was laid July 22. Under the influence of the nest, the egg, and the sitting female, Jack gradually succumbed to the brooding impluse and ceased to show erotic activity. All through his brooding he showed a tendency to come off the nest toward any human MALE DOVES REARED IN ISOLATION 125 being who came near, partly in friendship, partly in anger in defense of nest. But even in showing hostility toward us, Jack reacted toward us, not as he would toward other enemies such as dogs and cats, but with the behavior which a normal dove would show toward intruding members of its own species. The eggs were of course infertile. At the end of the brood- ing period I separated the pair, and kept Jack in isolation again through the winter. When the spring awakening came upon him (spring of 1909), he directed his display at first to no fixed and definite objects, but a little coaxing started him bowing- and-cooing to human beings, and soon he cooed to us a great deal, though not, I think, with quite the same earnestness as in his first spring season when he had had no experience with a dove mate. July 5, 1909. For the first time since the previous year, I placed him where he could see another dove — this time dove No. 20, an old, experienced female. The next day he began to show eagerness, evidently erotic, to reach human beings. On July 11th he tried to copulate with the hand. On July 15th, on two occasions, I saw him trying long and hard to accom- plish the sexual act on one of the perches of his cage. July 22, 9:15 A. M. -12:30 P. M. I opened the doors, allow- ing Jack into the cage of the female. He made no attempt to copulate with her; but at 11:30 A. M. I saw him, by himself, trying as on July 15th to accomplish the sexual act on a perch. 2:15 P. M. Again I open the doors. Soon I see the pair billing. The female, an old experienced bird, takes the lead, assuming the copulation posture many times, but the male does not mount. 2:35. They try again. The male mounts, but fails to ac- complish the act. 2:43. They try again. The male does not mount. 2 :50. The male tries to get out toward me. 2 :58. After long preliminary the male mounts, but too far back and to one side, and he soon dismounts. They bill again, then give it up. 3:07. After billing male begins to go through the sexual reaction on the perch. The female interrupts him by commenc- ing again her begging reaction. 3:10. The male tries to get out toward me. 126 WALLACE CRAIG 3 .35. They try again, not successfully, I think. After fur- ther preliminary reactions, the male makes a slight attempt to perform the sexual act on the perch, but soon desists. 3 :50. They make an attempt which is apparently successful. After this the male quickly learned to copulate with the directness and efficiency which characterize experienced doves. He went through the process of incubating the eggs and brooding the young. And in March, 1910, he fathered another brood. But he was always liable to leave the eggs or the young whenever a human being came into the room. He was, there- fore, a very poor sitter and a poor brooder, and his young were not well fed. Since I was hard pressed for room to keep my birds, I felt I could not keep a bird which was worthless as a parent: I gave Jack to the University of Maine, and he was killed and mounted for the museum. Billy, No. 23. Hatched September 23, 1907. Father re- moved October 1. Young removed from mother October 29, his 37th day. In order to test whether the development of voice in the young dove is at all due to exercise of the voice, I endeavored to prevent this bird from kahing and cooing. I kept him in a room by himself, with a brick apartment building between him and my other doves, and with the room darkened, so far as possible, at night. Kept so for months, he was far more silent than other doves, but he did coo a little, prompted evidently by internal stimuli. On January 6th I took him to a room in the University of Chicago where he could hear one other Ring- dove (Jack), and sometimes Common Pigeons; still he cooed but little. The comparative lack of vocal exercise did not, in any way that I could observe, retard or impair the development of his voice. His display behavior appeared very suddenly; so far as I observed, it appeared within three days, March 2, 3 and 4. Billy gave himself to human companionship as heartily as Jack had done, losing all fear of human beings, and showing all the signs of excitement and joy in our presence. After his long period of isolation, Billy was introduced to other birds on the same day and under the same circumstances as Jack (see page 122). MALE DOVES REARED IN ISOLATION 127 In 1908 I did not give Billy an opportunity to mate, as I did Jack, but kept him in a cage by himself. He could always hear other doves about him, but most of the time he was unable to see them. He continued as familiar as ever with his human companions. Even through the autumn he bowed-and-cooed to us whenever he was enticed to do so. His spring awakening began about the 20th of January, and became just as intense as that of the year before, for when his spring fever was at its height he cooed almost incessantly from the time the window shades were rolled up in the morning until the lamps were ex- tinguished at night. After three months of excitement, however, he seemed to be tiring out. In the month of May he became much more quiet, and toward the end of that month he acted as if he wanted to sit. We gave him a little straw, and he tried to make some use of it, so a few days later, about May 29, we gave him a nest containing an egg. He took quickly to the nest and sat faith- fuly on the egg all day, leaving it only to roost each night. June 1. Fearing that his health may surfer from lack of exercise, I decide to put an end to his sitting. So at 1 :30 P. M. I quietly remove the egg from under him. He sits on uncon- cernedly. June 4. Though the egg was removed three days ago he still continues to sit. For the first day or two he sat on the empty nest, but now he sometimes sits on the floor, hooking his bill around little pebbles or such objects and pushing them under him as 'f they were eggs. He is as savage as a broody hen. June 11. We took the nest out some days ago, but he con- tinues to try to sit. He is still insanely combative. June 18. I bring Billy into the company of another male dove (Frank), allowing him not only to see the other dove, but to come into contact with him and fight. This puts an end to his tendency to sit. Billy was not given opportunity to mate until October 8, 1910, when he was more than three years old. On October 8 and 9 I allowed him to enter the cage of female No. 19 (now a • bird of considerable breeding experience) , whose cage had long been beside his for preliminary acquaintanceship. I watched the behavior of the pair continuously (closing the cage door between them whenever I could not be present), but I kept out ] L'S WALLACE CRAIG of sight myself, in order that Billy might not be distracted by my presence as Jack was. October 8. Billy was very cruel to the female. October 9. Little or no cruelty. He responded to the female with mating and nesting behavior. Several times he showed sexual excitement and a desire to fly on something, yet no ten- dency to mount the female, until — 3:22 P. M. Female comes in again' and flies up beside male, and, on her initiative, they bill two or three times. She then takes the copulation posture, maintaining it steadily, close beside the male and parallel to him. This evidently gives the male just the needed stimulus, for after just the normal pause he mounts. He mounts, however, obliquely across her body, and goes through his sexual reaction in that sidewise position, not effecting union with the female. On October 10th I was not able to be with the birds, and by mistake I left the door open so that they had free access to each other. During that day Billy evidently learned to mate in a manner almost normal; but for a long time (for years, and I think in some degree to the present day) he persisted in a habit of omitting the preliminary ceremony of billing, flying without warning on the back of the female. Billy now has had much experience, not only with a mate and young but also with a small flock of doves; his attention has thus been drawn strongly toward his own species, people have been kept away from his cage to some extent, and he has practically given up his abnormal attachment to human beings. For a long time he continued to react with, more or less excite- ment to our presence, especially to our hands, but now no such tendency is noticeable. During his almost three years of isola- tion he developed a most truculent disposition, partly perhaps because he was teased by some persons — this is probably one reason why he so readily gave up human companionship. After a long period of peaceful life with mate and young, his disposi- tion has become very mild. He is a good sitter, brooder, and feeder of young. Frank, No. 30. Hatched July 24, 1908. Put in isolation about November 10, his 110th day (a late date, due to lack of facilities). MALE DOVES REARED IN ISOLATION 129 The history of Frank is in general like that of Jack and Billy, but with many differences of detail. Jack and Billy had been kept in a well-heated room, where several students were at work daily, both forenoon and after- noon. But Frank was kept in a cold room, in the climate of Maine, where he saw no one except myself, and saw compara- tively little even of me. These conditions probably account for the fact that Frank was for a long time a silent and shy bird. He never bowed-and-cooed, so far as I observed, until April 8 or 9, which was just after a warm wave had struck the locality, bringing, as it happened, a great wave of migrant birds. On April 10th Frank bowed-and-cooed a good deal. But after a few days he became quiet, and I did not hear this display coo from him again until about May 8. In bowing- and-cooing he always stood at the same point on his perch, facing toward a certain corner of the room, and thus was prob- ably directing his display to some object, though I did not discover what that object was. He never directed his bows to me until a change came over him which I shall now recount. Since the bird was uncomfortably shy and afraid of human beings, I began about the last of April to starve him mildly and compel him to feed from the hand. He quickly learned to take his seed in this way, and he always jumped on the hand — but not in a friendly manner, often with a few sharp pecks or a blow of the wing. But on May 1 1th, after jumping on the hand as usual he stood still a few seconds and then, quite unexpect- edly, he gave the sexual reaction of the male. As soon as the bird had performed this act for the first time, his whole bearing and demeanor changed so markedly that he looked like a different individual. Before, I had mistaken him for a female. Now, his form, his pose (tending toward the charging attitude), his movements, and the glare of his eye betokened the male. He kahed and bowed-and-cooed to the hand, and pecked it in amorous fashion, whereas before he had always pecked in an unfriendly manner. And he allowed the hand to preen his neck and even pull the feathers. From that date on until the next change in him (June 17th) Frank exhibited almost daily the sexual tendency, but he grad- ually ceased to bow-and-coo, and he relapsed largely into the demeanor of an immature bird. 130 WALLACE CRAIG June 17. I put his cage beside that of Billy, thus allowing Frank to see another dove for the first time since he was put in isolation. The sight of the other bird and the sound of his voice at once wrought a change in Frank like the change he had temporarily undergone on May 11th, but in this case the change was far greater, and was permanent. He was now in a few moments transformed from the meek young bird of indeterminate sex into the strong, aggressive adult male. I should not have recognized him as the same bird. He seemed to become so much larger than before, that it was hard to believe there was not an actual increase in size. Now he not only bowed-and-cooed, ut- tered the kah of excitement, and charged up and down the cage, but after a half-hour of such display he assumed the nest-call attitude and gave the nest-call coo, which I had never known him to do before. June 18 and 19. I allowed these two males to come together. They fought with might and main, and Frank worsted Billy. After these experiences with another dove, Frank readily and persistently bowed-and-cooed to my face and to my hand, as he had not done before. He continued for a long time, even after he was mated, to jump on the hand that fed him, so per- sistently that he was a nuisance. But though he jumped on the hand he did not show sexual behavior toward it, not after his first contact with the feathers of another dove on June 18th. July 5. I placed his cage beside that of a female dove. July 22, I opened the doors, allowing the two to come together. Frank showed a gradual leading up to the perfect mating beha- vior, similar to that of Billy, but more rapid. The most inter- esting feature was, that Frank sometimes turned from the female dove to bow-and-coo to me. Since then he has had much experience with doves, and has shown chiefly normal behavior. He maintained for a long time, as stated above, a habit of jump- ing on the hand. In 1910 he was taken from his cage and put, with other doves, in a large room where he ceased to come much into contact with people's hands, but came naturally into prox- imity with our feet; he developed that year a habit of bowing- and-cooing to one's shoe and then jumping on the shoe. This habit persisted in 1911, when I noticed that he reacted to tan shoes just as to black shoes, and he showed in many ways that he was reacting to the human being, even though his attention MALE DOVES REARED IN ISOLATION 131 was given chiefly to the shoe. Last year and this year C1913) Frank has been kept in small cages which are so arranged that the hand does not need to be put into the cage to put seed in; he has therefore had no contact with hands or shoes, and he has seen much more of doves than of human beings : he has largely, though not entirely, given up cooing to human beings. When I come near his cage, he still shows a desire to get out to me, and jealousy of other doves in my presence. 3 But he is a successful mate and a good, sitter, brooder, and feeder of young. Dove No. 39. Hatched July 14, 1910. Put in isolation September 26, his 75th day. This dove took to human com- panionship as did the other three. He has not yet been al- lowed into contact with his own species, but has been used for an entirely different experiment which is not yet completed. SUMMARY AND CONCLUSIONS The history of these doves reared in isolation covers a wide range of behavior, and many points of interest, from which I select the following. These conclusions will be confirmed and amplified in other articles, one of which will treat of female doves reared in isolation. 1. Four male doves were reared (after weaning) in isolation, each being unable to see any dove companions. 2. All these doves were for a long time very quiet. In the case of Frank especially, the masculine display behavior did not appear at all until he was socially stimulated; then the display behavior appeared so suddenly as to transform the bird in a few moments. All four doves exhibited more or less of this sudden development of behavior under the influence of new social factors in the environment. 3. The various notes uttered by this species, and all accom- panying expressive movements, developed in perfect form in isolated individuals, showing that young doves do not need to « August 18, 1913. Today, due to an accident, Frank escaped from his cage. I followed him with an open cage in which I sprinkled tempting seed, but could not induce him to enter. So I carefully approached him from below, and gently raised my hand to catch him. But he, seeing my hand come to him thus, began to show some of the old fascination for the hand, and after several seconds he jumped on my palm. I quickly put my other hand over him, and he was caught. 132 WALLACE CRAIG learn the sounds 4 or gestures of their species by copying older doves. The vocal and gesture reactions are thus, in their motor aspect, very completely and definitely fixed by the innate organization of the nervous system. 4. On the other hand, the innate sensory inlets leading to these reactions must be very indefinite or flexible. For the doves give their cries and their gestures, now to one sense- object, now to a very different object, according to their ex- perience. The four doves herein described, before they were allowed access (as adults) to their own species, gave their social reactions to human beings. 5. The three individuals which, after maturing in isolation, were allowed into the presence of their own species, associated with the other doves with every mark of eagerness and satis- faction. 6. These three gave up their intimate friendship for human beings. But they gave it up slowly and gradually, showing interesting divisions of attention between human companions and dove companions. If they had been encouraged to do so, they would probably have continued indefinitely to display to human beings; and even without special encouragement they remain, probably for life, exceeding tame, unafraid of the human species. This is one example of the importance and lasting influence of the dove's early impressions. 7. The sexual reaction of the male is, in its motor aspect, very completely and definitely provided in the innate nervous organization. But the innate sensory inlet to this reaction is not a complete sensory inlet; it is supplemented by experience. 8. The object to which the dove directs his social behavior becomes a symbol, in some cases it might even be called a fetich, to which he clings tenaciously, and to which he attaches a great complex of reactions. With all four of these doves, the human being became such a symbol ; especially the human hand, and in the case of one dove (Frank) the shoe. 9. When a dove performs an instinctive act for the first time, it generally shows some surprise, hesitation, bewilderment, or even fear; and the first performance is in a mechanical, reflex 4 It is true that each of my isolated males could hear other doves, but not such as to serve as a copy. For Jack and Billy could hear only each other; Frank could hear only the faint sound of doves cooing in a distant room; and No. 39 could hear no male dove but only a few females. MALE DOVES REARED IN ISOLATION 133 style, whereas the same act after much experience is performed with ease, skill, and intelligent adaptation. Thus even those acts which do not show improvement by the formation of asso- ciations, show improvement by facilitation. 10. In "the case of an act in which instinct plays the greater part, and learning by experience the lesser part, especially when the result of experience is merely facilitation, the improvement in the performance of the act may be so rapid that it quickly leads to perfection. To detect the influence of experience or practice, one must observe from the very first performance of the act. To observe the very first performance of the social activities of the adult, one must rear the animal in isolation; and then allow it, while under close observation, to come into contact with another animal. NOTES ' HUNTER ON THE QUESTION OF FORM- PERCEPTION IN ANIMALS H. M. JOHNSON Assistant Psychologist, Nela Research Laboratory, National Lamp Works of General Electric Company In a recent interesting communication, Mr. Hunter 1 calls attention to the need of sharper distinction between the study of form-discrimination and that of pattern-discrimination. He presents the thesis that animals below man and children between certain ages "have only a more or less crude pattern vision," and are unable to discriminate forms. Mr. Hunter asserts that there is no means of testing the validity of his belief unless the surroundings of discriminable forms be changed, since the form is " seen " with its surroundings and hence must be considered as " part of a pattern." Even if no other objects are in the visual field, the stimulus-object ' is seen surrounded by the more or less irregular outline of the field of vision, and so is again part of a pattern." As a means of controlling the surroundings, he pro- poses that after form-discrimination has apparently been estab- lished, the alleys of the Yerkes experiment-box leading to the stimulus-forms be enclosed with hollow cylinders or hollow triangular prisms. Thus, he says, ' it should be possible to demonstrate experimentally whether the subject was reacting to the ' forms ' or to the ' patterns.' " I am not clear as to two points raised in reading Mr. Hunter's article. First, with reference to his proposed method of control : Changing the enclosures of the alleyways would probably intro- duce new olfactory stimuli, and if the animal should have to touch any of the walls, the change would certainly introduce new tactile stimuli. The introduction of any new stimulus- > Hunter, Walter, S.: The Question of Form-Perception. This journal, vol. 3, 1913, pp. 329 ff. 134 FORM-PERCEPTION IN ANIMALS 135 factors frequently works serious disturbances. Quite recently Mr. Lashley 2 reported disturbance from the first source. Mr. and Mrs. Watson 3 obtained disturbance from both the first and second factors. Other instances might be enumerated. Suppose an experimenter should obtain failure to discriminate after making such a change as Mr. Hunter suggests. Is there any means of deciding whether the disturbance resulted from the change of " pattern " or from the simultaneous introduction of other novelties ? Secondly, with regard to the necessity of any control, which Mr. Hunter assumes: If a human observer place his eye at the exit of the " home-compartment " of the Yerkes box, will not a given form appear against quite different backgrounds and behind quite different foregrounds according as it occupies the right and the left positions respectively ? To the writer it does. Since the stimulus-form is as effective in one setting as in the other it would seem that we are justified in saying that the animal is reacting to the constant form difference and disregarding the variable pattern-difference of the stimuli; using the term pat- tern-difference in Mr. Hunter's way. » Lashley, K. S.: Visual Discrimination of Size and Form in the White Rat. Ibid., vol. 2, 1912, pp. 310 ff. » Watson, John B. and Watson, Mary I.: A Study of the Responses of Rodents to Monochromatic Light, Ibid., vol. 3, 1913, pp. 1 ff. The disturbance referred to is not reported, having occurred in the preliminary work. The writer received the information directly from the authors and refers to it with their permission. A DEFINITION OF FORM HAROLD C. BINGHAM Ellsworth College Regarding separate studies in form perception by Lashley and myself, 1 the following criticism has been urged: 2 " Both series of experiments referred to above are concerned with patterns, not forms." ' In problem boxes such as those described by Lashley and Bingham . . . the animal tested is con- fronted not by two "forms " corresponding to the configurations of the opal glass, but by such designs as are suggested in figure 1. Figure 1. Reprinted from Jour. Animal Behavior, vol. ^3, no. 5, p. 331. The squares drawn in the figure represent the rectangular tunnels down which the animal goes in making his responses. What the animal sees is a triangle or a circle each in more or less of a square setting." In substance, the form is not without a per- ceptible environment and, therefore, is strictly a pattern. Accepting the definition of pattern and conceding for the mo- ment the definition of form urged by Hunter, I maintain that, « Lashley, K. S. Visual Discrimination of Size and Form in the Albino Rat. Jour. Animal Behavior, 1912, vol. 2, No. 5. Bingham, H. C. Size and Form Perception in Gallus Domesticus. Jour. Animal Behavior, 1913, vol. 3, No. 2. » Hunter, W. S. The Question of Form Perception. Jour. Animal Behavior, 1913, vol. 3, No. 5, pp. 330-1. 136 A DEFINITION OF FORM 137 under the conditions as described in my paper, 3 the visible stimuli presented to the animals for discrimination were forms not patterns. On page 66 I state that the whole apparatus was set up in a dark-room. I follow this with an explanation of the only sources of illumination. On page 98 appears the plan of con- trolling all light factors. Now, in these conditions of control, there is to be found a refutation of the point which Hunter seeks to establish. The rectangular tunnel, to be sure, remains, but the perceptibility of the environment is wholly changed, if not destroyed. That the animals could not see the environment is attested by the fact that they were frequently observed to walk blindly into the confining walls. Not all of the time was the environment " darkened," but the control tests were always Figure 2 made to determine whether or not, among other factors, setting was a factor in discrimination. Figure 1 does not accurately illustrate the condition of the stimulus areas. With the introduction of a screen between the general illumination and the electric boxes and with the re- duction of the intensity of the source lights, a condition similar to that illustrated in figure 2 appears. In the compartment where the triangle appears, the source light fails to illuminate the corners of the tunnel, and so the perceptible portion of the setting changes to a sort of circular form as in diagram 1 . About the circular stimulus the visible setting is more nearly a perfect circle as is shown in diagram 2. Even if my apparatus offered a possibility of pattern discrimination, my plan of control would have made so variable the patterns confronting the animal that they never could have served as a basis of discrimination. . Hunter has apparently missed one of the essential features of the apparatu s which was used in my study. The dark-room • op. at. 138 HAROLD C. BINGHAM apparatus allows the experimenter to control the conditions of setting by means of artificial illumination. His criticism would be valid for similar experiments conducted in natural and un- controlled light. With the visual discrimination uninfluenced by setting, the perception could not have been of patterns. Another feature which Hunter has overlooked is one of method. Referring to figure 1, he asks: 4 " If an animal is trained on dia- grams 1 and 2, is it any wonder that he breaks down when con- *fronted by diagrams 2 and 3 ? ' Assuming now that the animal actually sees ' ' a triangle or a circle each in more or less of a square setting," no explanation is offered for the breaking down of the discrimination when merely the size of the form was changed, i.e., when the triangle of 1 was a circumscription or an inscription of the circle of 2. (Witness table 8, series 12, 13 and 14, March 21-22, and series 10, March 28; also table 9, series 5, April 21). 5 Now in these tests the patterns remained the same except in size, but the reactions changed from a high percentage of cor- rectness to a relatively low percentage. Besides this mis-statement of conditions there is an obvious lack of agreement in the matter of defining " form." The so- called " abstract sense " in which I have used the term has called forth objections. In my study of form perception I was not concerned with genetic phases of the problem. My task demanded an answer to the . question : Does the chick perceive forms ? 6 Consequently, it makes no difference whether or not the conception of form, to which I have given expression, is the result of development. One might consider that phase of the subject, but in my problem I was justified in determining whether or not the chick perceives form in this " abstract sense." Now if our animals fail to perceive circularity and triangularity as such, there are several principles that we should not lose sight of. In the first place, we should not try to excuse our animals nor become over-dogmatic in theorizing about extraneous, or even allied problems. We should accept as a fact the conclusion to which the evidence points. Moreover, we should seek to determine and define just what • Op. Cit., p. 331. ' Join. Animal Behavior, 1913, vol. 3, No. 2, pp. 106 and 109. • The task would have taken on other complexities had positive results been secured in the initial problem. A DEFINITION OF FORM 139 elements our animals do perceive. In this task we need not speculate on the question whether such elements are logical or genetic precedents of form perception. To avoid confusion, we should avoid the application of a multiple meaning to the same terminology. We should not attempt to simplify our definition of form so that this factor may be included in the animal's stock of perceptual experiences. Finally, if we find that our animals have a power of discrimi- nation which approaches form perception, but which is not form perception in the strict sense of the term, we should adopt a terminology to fit the special case; we should not enlarge the conception of the term " form " to cover the special case. Perhaps " a more or less crude pattern vision " is the nearest approach to form perception that animals possess. At any rate, Hunter has done well in calling attention to the distinction between patterns and forms. However, our definition must not stop here. Two forms may be identical, but different in " shape." This would be the condition in Lashley's study. He used two identical forms in that both were rectangles 2 mm. by 60 mm. They differ in this respect : one is extended laterally thirty times as far as its vertical extension, while the other is extended ver- tically thirty times longer than laterally. Now this is a difference in " shape " of two identical forms. Miss Washburn, in reviewing my study, 7 has failed to make this distinction between form and shape. She says : ' Bingham's chicks discriminated between a circle and a triangle when the apex of the triangle was on top, but since this discrimination broke down when the circle was presented with a triangle whose base was uppermost, the chick failing to choose the triangle, Bingham concludes that the chick was not reacting to form difference, but to ' the unequal stimulation of different parts of the retina.' The reviewer would conclude rather that the chicks were not possessed of an abstract idea of triangularity. A triangle with apex up is a different form from a triangle with apex down: the two have in common only the abstract quality of three-sidedness. The perception of form, as distinct from an abstract idea of form, is based precisely on the unequal stimu- lation of different parts of the retina." 7 Washburn, M. F. Recent Literature on the Behavior of Vertebrates. Psycho- logical Bulletin, 1913, vol. 10, No. 8, p. 320. 140 HAROLD C. BINGHAM It is not to be denied that a triangle with vertex up differs from a triangle with vertex down. But we can scarcely say that they are two different forms. They are both triangles; yes, more than that: they are equilateral triangles. Where they differ is not in form but in shape. When the extended base of the triangle is so placed as to stimulate the region of the retina which was formerly stimulated by the vertex of the triangle, a condition occurs similar to that pointed out regarding Lashley's ' forms:" the forms remain identical, but the lines of maximum and minimum extension have interchanged. This fact led me to conclude in my paper 8 that the apparent reactions to forms are the result of keen perception of size differences. I might have said they are due to perception of shape differences. The inversion of the triangle causes certain particular size changes — vertex or point interchanged with base or line — which causes a change in shape, but no general change of size since the area remains con- stant. Similarly the factor of triangularity remains constant and the form is unchanged. Not ' the perception of form," therefore, but the perception of shape ' is based precisely on the unequal stimulation of different parts of the retina." Our definition, then, as separate from the distinction between forms and patterns, must draw a line between forms and shapes. Referring to the retinal area stimulated, there is form which is general, e.g., triangle. But there is a particular feature about this general distribution of light — it is equilateral, or isosceles, or right angled — viz., shape. Forms are identical when their areas are equal and their general retinal distribution is similar. Shapes are identical when all extensions of the identical forms are equal and in the same relative directions. Thus, the area remaining constant, either or both form and shape may change. The form remaining constant, the shape may change. Change in form must always cause change in shape. Subsequent studies in this field should not fail to consider the factors of ' shape ' and ' ' pattern ' in their relation to form perception. Whatever system of control is adopted, such possible disturbances as these factors must be considered and, as far as possible, eliminated and isolated. Unquestionably my final test for form discrimination by shifting the position of the form was a severe one. Surely the factor of shape was a disturbing in- 8 Op. Cit., p. 110. A DEFINITION OF FORM 141 fluence. If, with all possible disturbing factors properly con- trolled, this test of shifting fail, form perception in the strict sense of that term can scarcely be said to prevail. I have shown that the discrimination of patterns was impossible in my study. There was a possibility of discrimination on the basis of two other factors. One of the remaining factors was form : the other has been arbitrarily termed shape. The inverted triangle possessed a different shape but an identical form as com- pared with the upright triangle. The high percentage of correct- ness in reactions changed to a relatively low percentage with the inversion of the triangle. Obviously, then, form was not the basis of choice. THE AUDITORY REACTIONS OF THE DOG STUDIED BY THE PAWLOW METHOD SERGIUS MORGULIS Biochemical Laboratory of Columbia University, New York From the time the first review 1 of Pawlow's ingenious method in animal psychology was published in this country great progress has been made by the numerous students of Professor Pawlow which puts an entirely new aspect on the psychology of the dog. Unfortunately, we are not in a position at this moment to offer a further extensive summary of the results obtained by this method since 1909, but a brief review of the recent paper of Usiewitch, 2 concerning the auditory faculty of the dog may prove of interest to American investigators. It will not be amiss to state succinctly for the benefit of those not familiar with the first article referred to, the principle of Pawlow's method, the minute analysis of the animal reactions performed with its aid and some of the broad generalizations regarding nervous activity deduced from those analyses. It is a matter of common experience that the salivary reflex may be actuated by the mere thought or even remote suggestion of a. delectable article, but it remained for Pawlow's unusual acumen to recognize in this trivial fact the means provided by nature for penetrating the hidden workings of the animal's psychology. The presence or absence of the salivary reflex informs the investigator of the organism's reaction to a given stimulus. The method, therefore, possesses all the advantages of being strictly objective, i.e., quite independent of the observer's interpretation or " personliche Ueberzeugung," as the Germans name it. Starting with the idea of the salivary reflex, it was a relatively simple matter to determine the flow of saliva both quantitatively 1 Yerkes, R. M., and Morgulis, S. The Method of Pawlow in Animal Psychology- Psychol. Bull., vol. 6, pp. 257-273, 1909. * Usiewitch, M. A Physiological Investigation of the Auditory Capacity of the Dog. Bull, St. Petersburg Military Medical Academy, Vol. 24, pp. 484-502; Vol. 25, pp. 872-891, 1912 (Russian). 142 THE AUDITORY REACTIONS OF THE DOG 143 and qualitatively by a special adaptation of the Pawlow fistula method. The duct of the parotid gland is exposed by an incision of the cheek and a permanent fistula or outlet to the exterior is made. The saliva is collected in a tube where it can easily be measured. Through persistent training the salivary reflex may become coupled with any desired stimulus which is frequently applied while the secretion of saliva is called forth by feeding the animal a powder consisting of meat and sugar. After long continued repetition the application of the particular stimulus alone is sufficient to cause a normal flow of saliva. This indirectly- produced salivary reflex is what Pawlow terms a " conditioned reflex," and the success of the analysis of reactions is based upon the absolute specificity of the latter. The ability of the animal to differentiate between gradients of various stimuli is measured by the changes registered in the fundamental salivary reflex. What the experimenter achieves by patiently adhering to a prearranged plan happens in nature continuously. The world of the individual is two-fold in its make-up. Some of its elements act on the animal's receptors causing sensations by directly stimulating them, others exert an influence, thanks to a more or less temporary association with one of the fundamental or unconditioned reflexes. In the function of the higher centers Pawlow distinguishes, therefore, two mechanisms; the mechanism of receptors (" analysers " in Pawlow's terminology) which is for sifting out and selecting from the mass of external stimuli and transforming them into nervous processes of purposeful reactions; the other mechanism is that of the transitory asso- ciation or interlocking of the phenomena of the outside world with the organism's responses. The latter is the mechanism of the conditioned reflexes in all its complexity. One of the most important discoveries in the investigation of the conditioned reflexes is the fact that every receptor at first enters into temporary association with the salivary reflex by its most generalized activity, its more refined and subtle faculty of differentiation being involved only gradually and by a very slow process. The intensity of an illuminated area becomes the cause of a conditioned reflex much sooner than the shape of that area. Likewise, when the central portion of the receptor mech- anism — -which may be either in the brain or in the spinal cord — 144 SERGIUS MORGULIS is destroyed or injured this particular receptor loses its ability to form conditioned reflexes except by its primitive and general- ized function. Animals whose optic centres have been injured can still form associations between stimuli of various light inten- sities and the salivary reflex, but not with stimuli from special groupings of light and shadow. As regards the auditory reactions of the dog it has been dis- covered by the conditioned reflex method (Selionyi. Elliasson, Tichomirov, Babkin, Burmakin) that its auditory faculty is much greater than that of man. The dog perceives J of a tone and appreciates tones of a frequency of vibration which is entirely beyond human reach. It was also discovered that the dog has an absolute memory for sounds, which probably, but very few of the most gifted musicians possess. Usiewitch's particular problem has been to study the dog's reaction to an intermittent auditory stimulation with a view to determining its ability to differentiate time intervals. It is hardly necessary to describe his technique, which is essentially the same as already described in the review alluded to. The intermittent stimulation was produced by means of a metronome. The subject of these experiments, a large healthy dog which never had been used for similar tests before, was found to be totally indifferent to the metronome so far as its salivary . reflex was concerned. By persistent training a conditioned reflex has been established to the stimulation with 100 oscillations per minute of the metro- nome. The stimulation of intermittent sounds of such frequency called forth 6-10 drops of saliva every time. The interval between successive oscillations was then modified, the moment of the disappearance of the conditioned salivary reflex indicating the lowest limit of differentiation. Without going into any details of this most interesting investigation or quoting actual data, I will say that the dog could sharply distinguish the short- ening of the interval by less than A~A of a second. Indeed with the well developed reflex to the stimulation of 100 beats per minute a change of the rate to either 96 or 104 beats was im- mediately reacted upon by a marked diminution or even complete cessation of the flow of saliva. Furthermore, Usiewitch brought out some very significant points with regard to the intermission between tests with the THE AUDITORY REACTIONS OF THE DOG 145 established, standard stimulation, and some unusual stimulation. Thus, the dog is able to differentiate distinctly between 104 and 100 beats (standard) if the new rate is tested 10, 15, 45, or 60 minutes later. The differentiation is less certain after 18 hours of intermission and vanishes completely after a lapse of 4'5 hours. In responding to various intermittent stimuli of unaccustomed frequency a remarkable regularity and uniformity stamps the results. Applied immediately after stimulation with the standard rate of oscillation it produces a distinct depressing effect on the salivary secretion in the first trial. During the subsequent few trials the conditioned salivary reflex increases considerably, then again diminishes to complete disappearance in further tests. -These facts are very important because they offer a clue to the analysis of the phenomenon of inhibition. This review purports to bring once more before the attention of American investigators the enormous value of this purely objective analytical method in animal psychology and to stimulate an active interest in the subject which should soon lead to a systematic investigation of the reactions of various animals by this method. JOURNAL OF ANIMAL BEHAVIOR Vol. 4 MAY-JUNE, 1914 No. 3 CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS i JOHN H. LOVELL, Waldoboro, Maine There are many cultivated flowers adapted to winged pollinators, which are rarely visited by insects although they are of large size and display the most brilliant hues. Among the species enumerated by Plateau as illustrations are the red geranium (Pelargonium zonale Willd., hybrid Lepidopterid flowers from Southern Africa), the scarlet sage (Salvia splendens Sellow, ornithophilous, from Brazil), the cardinal flower (Lobelia car- dinalis L., ornithophilous, from North America), and the splendid gaudy flowers of Passiflora incarnata L. (probably ornithophi- lous, ^ from North America) . s Other neglected flowers employed by Plateau for experimental purposes were Lilium candidum L. (hawk-moth flowers), Passiflora adenophylla Masters (?), (prob- ably a hybrid), Oenothera speciosa Nutt. (hawk-moth flowers), Pisum sativum L. (almost invariably self-fertilized, probably introduced from Western Asia into Europe in prehistoric times) , > Pelargonium zonale Willd., Clematis Jackmanni Jack, (hybrid pollen flowers), and Petunia hybrida Hortul. (hybrid, the South American species are ornithophilous?). That anthophilous birds and insects have played an important part as pollinators in the phylogenetic history of the flowers enumerated, in the 1 The pollination of green or inconspicuous flowers has been considered by the writer in an earlier paper. Am. Nat., 46:83-107, 1912. ' In Alabama Trelease saw the flowers visited by humming-birds. Knuth, Paul, " BliUenbiologie," 3: 510. » Plateau, F., " Les insectes et la couleur des fleurs," L'Annee Psychologique, 13:72. « De Candolle, A., "Origin of Cultivated Plants," p. 329. 148 JOHN H. LOVELL lands where they are or were endemic, will not be questioned by any orthodox floroecologist. But manifestly when they are cul- tivated in widely separated stations, under the most diverse conditions, there is a strong probability that in many localities their normal pollinators will be entirely absent or extremely rare; while the flowers themselves modified both in form and function by artificial selection and hybridization may cease to remain equally attractive, e.g., double flowers may be devoid of both nectar and pollen. On the other hand why should we expect common Hymenoptera and Diptera frequently to visit flowers from which they can not legitimately obtain nectar, and to which they are not beneficial ; or why should we look for diurnal insects as common visitors to crepuscular flowers ? One of the advan- tages of reciprocal adaptation between flowers and their polli- nators is the exclusion of injurious and useless forms. But Plateau assumes that all bright-hued flowers, according to the theories of M tiller and Knuth, no matter what their manner of pollination, should frequently be visited by diurnal insects. The rarity of insect visitors to many beautiful flowers with very showy colors, he remarks, places the biologists of the school of Hermann Muller in a singularly embarrassing position. « He summarizes his views as follows : ' My observations establish the truth of the fact, well-known though not sufficiently insisted upon, of the existence of many plants with flowers formed on the entomophilous type and pre- senting large dimensions as well as brilliant colors, which attract almost no -diurnal insects. It follows that the attractive role, or, as it is often called, vexillary role of the forms and colors of floral envelopes is either nul or of very little importance. Causes of attraction other than colored surfaces are necessary to bring pollinators to flowers and to lead them to return again after a first visit; they are an odor, which is agreeable to the insects, and a sweet liquid, which permits them to appease their hunger and provide food for their young. * Unfortunately for the general acceptance of Plateau's con- clusions, they are not of universal application, but are controverted 6 Plateau, F., " Recherches experimentales sur les fleurs entomophiles peu visiters par les insectes rendues attractives au moyen de liquides sucres odorants," Mem. de V 'Acad. roy. de Belgique, 2me ser., 2:5, 1910. • hoc. cit., pp. 51-2. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 149 by the characters of various natural flowers. The cornflower, {Centaur ea Cyanus L.), Gentiana acaulis L., and several other gentians have conspicuous nectariferous flowers, which are visited by numerous insects although they are devoid of scent. Bees frequently gather pollen from poppy flowers, which are not only nectarless but possess a faint unpleasant odor. From the wind- pollinated, purple flowers of the elm, which are both nectarless and odorless, honey-bees in immense numbers sometimes procure pollen for early brood-rearing; while many other anemophilous species are also valuable to the bee-keeper as sources of pollen. Nor is it stated that there are many conspicuous flowers, which are neglected by insects notwithstanding they are strongly odoriferous, as the sweet pea, Lilium candidum, and varieties of Pelargonium, which have the entire plant pleasantly scented. Finally, if a flower is rich in nectar, it may be both inconspicuous and odorless and yet receive numerous visits. According to Fritz Muller, there is in South Brazil a species of Trianosperma which is visited very abundantly all day long by Apis mellifera and species of Melipona, although the flowers are scentless, greenish and to a great extent hidden by the foliage.' It is thus apparent that the visits of insects in large numbers are not dependent on the presence of an agreeable odor. But, assuming the validity of his conclusion that bright coloration is without significance because certain conspicuous flowers are commonly neglected by insects, Plateau performed a long series of experiments, in some instances introducing honey and in others odoriferous sweet syrups into neglected flowers with the result that in most cases insects were attracted, often in large numbers. In his earlier experiments of 1897, he em- ployed only honey diluted with water. When a small quantity of this mixture was placed on the handsome flowers of Pelar- gonium zonale, Phlox paniculata and Anemone japonica, it was speedily discovered by numerous Diptera and Hymenoptera. Similar results were obtained with greenish or dull-colored flowers. The vexillary organs are, therefore, asserted to be of little or no importance. « Knuth considered these experiments of no value since " they i Muller, H., " Fertilization of Flowers," p. 270. » Plateau, F., "Comment les fleurs attirent les insectes," 3me part., Bull, de V Acad, roy. de Belgique, 33:27-37, 1897; 4me part., loc. cit., 34:604-10, 1897. 150 JOHN H. LOVELL only prove that the odor of honey exercises a great power of attraction which has long been known. It is only necessary to place honey anywhere to secure the immediate appearance of numerous insects which are fond of it."» To this criticism, Plateau replied: ' Quelle pauvre argumentation! Knuth ne s'apercoit pas qu'il me donne pleinement raison. En effet, s'il a suffi de l'introduc- tion d'un peu de miel dans des fleurs habituellement negligees pour y amener les Insectes, c'est que l'eclat des corolles ne compte guere et que le perfum de la substance que ces animaux recher- chent avidement a constitue seul l'excitant determinant leurs actes. J'avais done demontre ce que je voulais demontrer."i° Plateau's conclusion that certain conspicuous flowers, which are devoid of nectar and pollen, or nearly so, are neglected because insects fail to notice their colors, it is believed, can readily be shown to be fallacious. The flowers are neglected not because they escape attention, but because anthophilous insects have learned from experience their inability to procure food materials from them. They do not neglect them entirely, but visit them occasionally, > > although they do not often repeat their futile visits since ' ' memory appears to replace both odor and color as the directive stimulus of first importance. > = In his experi- ments with odoriferous essences, that is, odors without a sweet syrup, Plateau recognized the fact that if they are employed alone a Hymenopteron or Dipteron entering the corolla and finding nothing will not return again. 13 This statement is 9 Knuth, Paul, " Handbook of Flower Pollination," translated by J. Ainsworth Davis, 1:206. " Plateau, F., " Recherches experimentales sur les fleurs entomophiles," etc., p. 8. " This statement will be supported later by a large number of observations. 12 Coulter, Barnes and Cowles, " Textbook of Botany," 2 (Ecology by H. C. Cowles) :850. On the memory of honey-bees cf. Forel, A., "Ants and Some Other Insects," translated by W. M. Wheeler, p. 28; and on the memory of place in bees cf. Buttel-Reepen, H.V., "Are Bees Reflex Machines," translated by M. H. Geisler, pp. 19-39. In the autumn of 1912 I placed a dish containing fragments of comb honey in a secluded spot nearly surrounded by a steep bank and willow bushes. A few bees were brought to the honey and it was soon visited by a large number. After they had been fed several times the dish was removed and everything left as at first. Two weeks later I examined the place but failed to discover a single bee. The weather was, moreover, growing colder and they were no longer flying freely. I now placed on the same spot as previously another dish of comb honey; and two hours afterwards I found it swarming with bees. During two weeks they had evidently kept this locality under constant surveillance, inspecting it from time to time, although there was nothing to attract their attention. i» Plateau, F., " Recherches experimentales," etc., p. 10. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 151 equally applicable to color. Neither color nor odor separately or together will attract insects continuously, if they can obtain no spoil. Plateau was equally mistaken in supposing that the addition of an agreeable odor is indispensable; for it is only necessary to introduce a solution of sugar and water, which is odorless, to bring insects to the flowers in great numbers, as will be shown experimentally. In the absence of accessible food materials pleasantly scented flowers will not be visited more frequently than would be the case if they possessed only bright coloration. Insects will not repeatedly visit an inflorescence because they experience an aesthetic pleasure. This is well shown by Lathyrus odoratus L., or the sweet pea, which, notwithstanding its strong fragrance and brilliant hues, is very rarely sought by insects, because the nectar is inaccessible to nearly all of them. An ample, available food supply will alone secure continued and frequent visits of insects to flowers. Since it can be shown, therefore, that an inflorescence can be rendered very attractive to insects without the addition of an odor, it logically follows from Plateau's own method of reasoning that conspicuousness is beneficial. When Plateau introduced honey into certain selected flowers, they received two allurements, an agreeable odor and a sweet liquid food, which sharply distinguished them from the flowers left in their natural state. In effect, the flowers containing honey became distinct physiological varieties. Color and odor were not brought, therefore, into competition on equal terms; the flowers in their natural state possessed only color and form, while those into which honey was introduced possessed color, form, an agreeable odor and a liquid food. Manifestly, the latter flowers were given the greater advantage, and it is unfair to conclude that because they received the greater num- ber of visits, odor was essential and color was of no significance. Throughout Plateau's experiments, the presence of the vexillary organs was a source of error. As he had assumed that they were of no value, it is difficult to understand why he did not remove the floral envelopes, when the flowers would of necessity have been compelled to depend wholly on the odoriferous liquid food. Finally, to have made the competition impartial, an odor- 152 JOHN H. LOVELL less syrup should have been introduced into the empty flowers. The experiments were, therefore, not well adapted for the pur- pose intended and the results obtained, as interpreted by Plateau, are misleading. In another series of experiments, Plateau unsuccessfully attempted to draw insects to flowers by means of the odoriferous essences of lavender, thyme, sage, and mint. "The Labiatae are habitually much visited by bees and I hoped in giving the preference to essences extracted from these plants to see bees and allied insects drawn to the flowers." Essences of orange and bergamot were also employed. But the attrac- tion proved very small or non-existent. Certain essences as thyme and sage were feebly attractive, while mint was even repellent. •« Knuth makes the following comment: " From these experi- ments it follows that solutions of odoriferous plant extracts, which ought to attract insects, do not do so."'* Plateau subse- quently attributed the failure of the flowers to attract insects to the too violent and medicinal odors of the extracts employed; they never possessed the delicate perfume of the plants from which they were extracted. In a new series of experiments undertaken in the spring and summer months of the years 1907-9, instead of odoriferous essences, odoriferous liquid foods, which it had been previously ascertained were attractive to insects, were intro- duced into neglected flowers. The sweet liquids employed were anisette I* (essence of anise, syrup of sugar and diluted alcohol), the cooked juice of cherries, syrup of cassonadd' (syrup of brown sugar to which a few drops of rum had been added), and syrup of Angelica ■ > (syrup of cane sugar flavored with a strong aromatic essence obtained from the petioles of Angelica officinalis). Fifty- five experiments were performed with these syrups, but descrip- tions of only a part of them were published, those being selected "Plateau, F., "Comment les fleurs attirent les insectes," 5me part., Bull, de VAcad. roy. de Belgique, 34:872-5, 1897. "Knuth, P., " Handbook of Flower Pollination," 1:207. » A bee-keeper in California reported that he found essence of anise very useful in attracting swarms of bees to empty hives, while another bee-keeper in Ohio did not find it of much value. Gleanings in Bee Culture, 40:482. i' This is somewhat similar to the mixture used in " sugaring " for moths. Psyche, 19:195. ■ 8 The tender stalks are preserved in sugar and sold as a confectionery. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 153 which most strongly sustained his views, while a few particulars were given in regard to his other experiences. •» There will be described in the present paper a few of the more interesting experiments performed by Plateau on relatively large and brilliantly colored flowers seldom visited by insects, following which will be given the observations of the writer on similar flowers. Among the familiar species selected by Plateau was a purple-flowered variety of Clematis Jackmanni Jack., a hardy perennial vine widely cultivated both in Europe and Amer- ica. The flowers are nectarless, but bees obtain from them a small amount of pollen. A vine of C. Jackmanni superba is described by Plateau as covering a wall three meters in height and displaying many hundred magnificent blue-violet flowers, which are said to have been wholly ignored by insects. On a very warm clear day anisette was introduced into eleven flowers, near each other, and constituting a group by themselves. In the hour following, they were visited by fourteen bumblebees and six flies belonging to the family Syrphidae. In four instances bumblebees examined adjacent flowers which remained in their natural condition. The facts related by Plateau are not called in question; but it should be noted again that the ungarnished flowers possessed only conspicuousness and pollen, while the' eleven flowers contain- ing anisette possessed conspicuousness, pollen, an agreeable odor and a sweet liquid; evidently color was not here brought directly into competition with odor. Let us now endeavor to determine whether the purple flowers are as completely neglected by insects as Plateau supposed; and whether insects can not be induced to visit them in large numbers without the addition of an agreeable odor! The purple-flowered Clematis on which my observations were made was a small vine bearing only eleven flowers wholly or partially expanded. The flowers were of large size, pale purple, nectarless, and odorless. As regards brilliancy of coloring and number, they were at a great disadvantage compared with the inflorescence described by Plateau. They were very frequently examined during the entire period of blooming. » Plateau, F., " Recherches experimen tales sur les fleurs entomophiles peu visitees par les insectes rendues attractives au moyen de liquides sucres odorants," Mem. de I' Acad. roy. de Belgique, 2me sen, 2:1-55, 1910. 154 JOHN H. LOVELL On June 11, 1912, a warm clear day, a honey-bee was observed at 12:35 p. m., gathering pollen, also a wild bee which flew away so quickly that it could not be determined. The honey-bee visited four or five flowers before returning to the hive. A few minutes later a second and third honey-bee came for pollen; and during the succeeding hour one or two workers were constantly visiting the flowers for this purpose. One of them remained for a long time, and the loads of purple pollen in the pollen-baskets were plainly visible. Two females of Halictus craterus came for pollen. A bumblebee inspected the flowers, but did not alight. A small undetermined bee flew from flower to flower apparently looking for pollen. At 1:35 p. m., there were no insects on the flowers; but a little later a small species of Halictus, and also a female of the larger Halictus craterus arrived and removed all the pollen remaining available. On three other occasions a female Halictus craterus was seen collecting pollen, which in one instance colored purple the under side of the abdomen and the brushes on the posterior legs. No attempt was made to capture any of the bees since this would have lessened the normal number of visits. The nectarless flowers of Clematis were not, therefore, entirely neglected by insects; but were visited by a number of bees suffi- ciently large to remove all the pollen they produced, and to have effectively pollinated the stigmas had they been in a normally receptive condition, and as this is all that is required, additional visits would have been of no advantage. The sterility of the flowers is not thus due to the absence of pollen-carriers as Plateau supposed. The flowers should be examined immediately after anthesis before the pollen has been removed; since Plateau makes no mention of the pollen he probably did not observe whether it was removed or not. = » I inspected the flowers many times without finding any insects, and it is easy to understand how a casual observer might gain the impression that they were entirely neglected. Plateau's failure to discover insects on the flowers in their natural condition may have been partly due to an insufficient number of observations, partly to location, and partly to the absence of suitable species of bees. Bumblebees are not well adapted for gathering the scanty supply of pollen, and prob- « Cowles has suggested that Plateau failed to see the earlier visits of his insects, Cowles, H. C, " Insects and Flower Colors," Bot. Gaz., 39:70, 1905. CONSPICUOUS FLOWERS RARELY VISITED BY INSCETS 155 ably seldom make the attempt. After the pollen has been entirely removed there is, of course, no reason why bees should continue their visits. In an earlier paper I have shown that flowers fre- quently visited by bees were almost entirely deserted when the corollas were removed ; there is, therefore, good reason to believe that the purple sepals of Clematis attract the attention of insects. I next proceeded to place on a few flowers an odorless sweet liquid for the purpose of ascertaining whether they would not be visited by bees in large numbers. White granulated sugar dissolved in equal parts of water yields an odorless and colorless syrup, as is admitted by Plateau. ^ June 16 and 17 were cloudy, rainy days, but the 18th was fair. At 8 o'clock a. m., a small quantity of syrup of sugar was placed on three flowers. No visitors were observed until 9:15, when two females of Halictus craterus began feeding on the syrup ; five minutes later there was a honey-bee at the syrup and a female of Halictus craterus gathering pollen. Sugar syrup was now placed on a fourth flower. At 10:00 o'clock there were three honey-bees and one female H. craterus feeding on the syrup, a second female H. craterus on a flower without syrup, and a third hovering in the air. Ten minutes later a honey-bee left a flower on which there was syrup and flew to two empty flowers; but, after carefully examining their centers and finding nothing, it returned to the flower on which it had previously been at work. The bees were compelled to learn by experience which flowers contained syrup and which did not. I replenished the supply of syrup from time to time as it was consumed, and at 12:15 p. m., there were seven honey- bees sucking on the flowers. On the morning of June 19 I again put syrup of sugar on the flowers, and presently three or four bees were at work. It seemed needless to continue the experiment further, for the bees came from my apiary and it was only a question of time and of supplying the syrup in sufficient quantity to have attracted them in great numbers. During the latter part of this experiment there were eighteen flowers in bloom. Plateau's assumption that the flowers would not be visited unless they were given an agreeable odor was shown to be wholly erroneous; the addition of an odorless sweet liquid secured the visits of insects in far greater numbers than were observed by him. « Plateau, F., " Recherches experimentales," etc., p. 19. 156 JOHN H. LOVELL Another common flower selected by Plateau for experiment was the edible garden pea, Pisum sativum L. The flowers are rarely pollinated by insects, and self-fertilization is almost in- variable. It was for this reason selected by Mendel for his celebrated experiments in hybridization. He says: "Among more than 10,000 plants which were carefully examined there were very few cases where an indubitable false impregnation had occurred." » During four summers, however, Muller frequently saw the flowers visited by both sexes of Megachile pyrina, and the females both sucked nectar and collected pollen.^ Plateau's observations were confined to walking on two occasions through many cultivated fields of peas, in one of which he saw a single Bombus agrorum. Plateau introduced anisette into a dozen, or, on one day, two dozen flowers of Pisum sativum growing in his garden, which were carefully observed for from one to three hours on five days. The anisette was renewed_ each day. Twenty visits were made by species of Bombus and Megachile] and ten visits by flies and small bees which could not possibly effect pollination. Plateau attributed the small number of insects attracted by the odor- iferous liquid food to frequent interruptions by rain. The flowers of the common garden pea are rarely visited by insects, not because they are nearly odorless and the coloration is of no value, but because of the difficulty of depressing the carina. This species no longer exists in the wild state; and, according to De Candolle, was probably introduced into Europe from Western Asia. *« Muller says: " In its original home the pea no doubt adapted itself to some strong and at the same time diligent and skillful species of bee, which could easily depress the carina, and was plentiful enough in ordinary weather to act as the regular fertilizing agent. Under such conditions, the advantages of firm closure would outweigh the disadvantages. In our climate the pea fails to find bees adapted to its flowers, and it would be much better for it under these altered conditions to have its flowers less firmly shut." »» During the summer of 1912, I saw the flowers of the garden "Bateson, W., "Mendel's Principles of Heredity," p. 342. Bateson is of the opinion that Thrips may be a source of error. » Muller, H., " Fertilization of Flowers," p. 214. » De Candolle, A., " Origin of Cultivated Plants," p. 329. « " Fertilization of Flowers," p. 214. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 157 pea visited a few times by females of Bombus fervidus only; but in other seasons I have occasionally observed honey-bees endeav- oring to find nectar in the flowers. The visits of the bumble- bees were made in the legitimate way, but I was unable to approach near enough to determine whether the carina was actually depressed or not. In each instance, the bee visited only three or four flowers, probably because it experienced difficulty in obtaining the nectar which was not abundant. In this connection, it is a matter of surprise that Plateau passes over the flowers of the sweet pea, Lathyrus odoratus L., without mention. This species belongs to the same family as the garden pea, to which it is closely allied in form and structure, though differing in details. Although the blossoms have a strong and pleasant odor suggestive of honey in addition to the most brilliant hues, it is yet more sparingly visited by insects than the garden pea. According to Plateau, the nearly scentless flowers of the garden pea require an agreeable odor to attract insects; but the fragrance of the sweet pea, which is so pleasing that any effort to improve it would be as futile as the proverbial attempt to paint the lily, does not give the inflorescence any permanent advantage over that of the garden pea. If the absence of insects from the garden pea shows that the influence of its coloration is of no significance, then it may be inquired does not the absence of insects from the sweet pea prove that both color and odor are of no importance ? Bees neglect to visit the sweet pea frequently not because these two allurements are of no benefit, but because they have learned from experience that they can not obtain nectar. To attract numerous visits, both the garden pea and the sweet pea require an available food supply. Place a honey-bee on one of the wings of the sweet pea, and it is at once apparent that it is neither large enough nor strong enough to depress the carina. Repeated examinations of the flow- ers continued through several weeks of the summer of 1912 failed to reveal a single visit by any species of bee. But by September 22, the autumnal honey-flow from the golden-rods was over, and the honey-bees were at liberty to give more attention to the few other flowers still remaining in bloom. On this date I repeatedly saw honey-bees alight and examine the flowers of the sweet pea, but they made no attempt to depress the keel. One probed diligently between a wing petal and the 158 JOHN H. LOVELL keel, while another sought for nectar under the calyx lobes, at one time standing on the back of the standard. None of their efforts proved effectual. Neither can any of our Maine bumblebees depress the carina. On September 26 I saw a female Bombus fervidus visit illegiti- mately twenty flowers in succession. Standing sideways on the flower, clinging to one of the wings and the calyx, she inserted her tongue in a crevice between the standard and a wing petal. Subsequently she robbed many other flowers of their nectar in the same way. The nectar was also obtained in a similar manner by a worker of Bombus consimilis.* Until the summer of 1912 I did not suppose that any of our indigenous bees could properly pollinate the flowers; but on August 17 and September 15 and 22, a female leaf -cutting bee, Megachile latimanus, was observed to visit the flowers legiti- mately. She manifested so little fear that I was able to watch her movements at close range. The stigma protruded for a long distance, touching the abdominal scopa on one side and on the other the brush of hairs on the tibia of the posterior leg. Both brushes were thickly covered with pollen. In England, also, according to Punnett, a species of Megachile is able to depress the carina. =e Muller saw only Anthidium manicatum sucking on the flowers. Neither color alone in the garden pea nor color and odor combined in the sweet pea will induce frequent visits, if nothing is to be gained thereby; but, if an odorless sweet syrup is placed on the flowers, bees will resort to them in large numbers. On the morning of August 16 I placed syrup of sugar on a number of sweet pea blossoms. Three times during the afternoon I found a worker of Bombus consimilis feeding on the syrup — ■ probably the same bee in each instance. On the 17th I renewed the supply of syrup, and at about 12:30 p. m., a honey-bee dis- covered it; an hour later there was three honey-bees. Before the close of the afternoon, four honey-bees and two bumblebees were sucking the syrup, or flying about the flowers to which it *Bombns consimilis Cr. is doubtless correctly regarded as a synonym of B. vagans Sm., but as the local specimens agree exactly with a set of the three forms of B. consimilis obtained from the Ac. Nat. Sci. Phil, the name has been permitted to stand in this paper. » Punnett, R. C, " Mendelism," p. 188. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 159 adhered in small drops. It is evident that they must have oc- casionally inspected the blossoms, or they would not have discovered the colorless and odorless liquid. By frequently replenishing the syrup, an indefinite number of bees might have been attracted. There was sugar syrup on about ten flower clusters. An available and abundant food supply is required, therefore, to secure numerous and continued visits. Let us now inquire whether similar results can not be obtained in the case of the garden pea, Pisum sativum. On a clear and moderately warm morning (July 31, 1913), at 8:00 o'clock, about forty flowers of this species were dipped in sugar syrup, a few, small drops of the thin, colorless and odorless solution adhering to each corolla. The garden was in a secluded location, which had not been planted previously for many years, and was nearly surrounded on two sides by a tall cedar hedge. During the half hour following, a honey-bee inspected the flowers on another row of peas, but failed to find the flowers garnished with sugar syrup. At 8:40 a. m., a white-banded wasp, Vespa con- sobrina Sauss, was also seen examining the flowers on another row of peas, and presently, more fortunate than the bee, it came to the flowers on which there was sugar syrup. For the larger part of the day this wasp, and a little later a second wasp of the same species, worked diligently gathering the sweet liquid. I recorded many of their visits, but it would be tedious to relate them in detail. At 9 : 10 a. m., a honey-bee was observed inspecting ungarnished flowers of the garden pea; it alighted on the carina and then sought unsuccessfully to reach the nectar through the side of the flower. Ten minutes later a honey-bee discovered the flowers with syrup, and subsequently it continued to return to them at intervals until 10:20 a. m., when I closed the experiment. It met with many disappointments as it often examined un- garnished flowers. The pea blossoms were also visited by a yellow-banded wasp, Vespa germanica Fab. At 4:00 p. m., I found both species of Vespa still resorting to the flowers. On August 2, a hot, clear day, at 12:30 p. m., forty flowers of the garden pea were supplied with sugar syrup, which was almost immediately found by a honey-bee and a Vespa consobrina. At 12:45, a second honey-bee and a Vespa germanica came to 160 JOHN H. LOVELL the flowers. In another part of the garden a female Megachile melanophaea (one of the larger leaf-cutting bees), was observed to visit ungarnished flowers in the normal way. At 1:15 o'clock there were two honey-bees, two Vespa consobrina and the small pale blue butterfly, Lycaena pseudargiolus, sucking syrup from the flowers; and fifteen minutes later one honey-bee, two V. consobrina and two V. germanica. The visits continued until 2:45 p. m., when I closed the experiment. The number of visits by bees and wasps received by the flowers of the garden pea garnished with sugar syrup, during the time they were under observation, was much greater than I had ex- pected. Under the conditions I should not have been surprised had there been no visits by Hymenoptera. On the night preceding August 2 there had been much rain, and the following morning was very foggy, so that the leaves of the pea vines at noon were covered with small drops of water, which could not be distinguished from drops of sugar syrup. The bees made many fruitless visits to flowers without syrup and also to flowers on the wrong row. But both bees and wasps soon learned to confine their attention chiefly to the end of the row with garnished flowers. There were many small Syrphid flies, as well as larger flies, flitting about among the foliage of the pea vines. Although they not infrequently came to the flowers on which there was sugar syrup, but little importance was attached to their visits, as evidently they might be largely the result of chance. One or two smaller bees belonging to the genera Sphecodes and Prosopis were also among the visitors. But the larger aculeate Hymen- optera, whose visits are manifestly purposive, were regarded as much better adapted for observation than small, little specialized insects. It was conclusively shown that an available food supply, without the addition of an agreeable odor, would induce numerous visits of honey-bees and social wasps to the odorless flowers of Pisum sativum. " The many horticultural varieties, known under the name of Petunia hybrida and cultivated in all gardens, have resulted, as is well understood, from crossings between P. nyctaginiflora Juss. and P. violacea Lindl. They offer this very interesting peculiarity, from the point of view of the present work, of receiv- ing no visits from the domestic bee, notwithstanding the bril- liancy and dimensions of their beautiful, infundibuliform, white, CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 161 rose, violet, or purple flowers."" Plateau, however, observed visits by many bumblebees, and species of Diptera belonging to the genera Eristalis and Syrphus. Plateau employed in his first experience a large group of Petu- nias, surrounded by other plants, as Tagetes paiula and Scabiosa atropurpurea, attractive to bumblebees, flies and butterflies; while among the Petunias there was a single stalk of Borago officinalis which alone was visited by honey-bees. On a clear but cool August morning, at 9:30 o'clock, he introduced the odoriferous juice of cooked cherries into six flowers near the stalk of borage. At 3:30 p. m., of the same day, the honey-bees discovered the cherry juice and entirely abandoned the borage flowers for the Petunias. During an hour there were fourteen arrivals, each individual visiting many of the garnished flowers, and rarely a few of the empty flowers. Essentially similar results were obtained in Plateau's other observations on Petunias. The two common species of Petunia endemic to South America have long narrow tubes, are strongly scented in the evening, and are either adapted to crepuscular Lepidoptera or are orni- thophilous ; in either case we should not expect to find honey-bees among their legitimate pollinators. The hybrid forms of cul- tivation, moreover, are destitute of nectar; and even if it were present the throat of the corolla is so obstructed by the filaments and style that it would be inaccessible to them. Plateau asserted that an odoriferous syrup was required to attract visits by honey- bees, but it can readily be shown that the presence of an odorless, sweet liquid will render their visits very numerous. A medium sized group of single-flowered Petunias of various colors was sel- ected for my observations. On July 31, 1913, there were only two flowers in bloom, into both of which I introduced sugar syrup. A bumblebee inspected both flowers but overlooked the syrup. On the 2nd there were two fully expanded flowers, and one which had wilted and closed. A honey-bee examined all three, and remained a long time in one of the open flowers. As the sugar syrup had evap- orated, the supply was renewed. The honey-bee returned and thirty minutes later was still visiting the flowers. On the follow- ing day a female Bombus consimilis was a visitor. 2' Plateau, F., " Recherches experimentales, etc." Mem. de V Acad. roy. de Bel- giqae, 2me ser., 2:46, 1910. 162 JOHN H. LOVELL On August 10, I introduced sugar syrup into nearly all the expanded flowers. Vespa consobrina was a constant visitor throughout the day, and subsequently Vespa germanica was also observed on the inflorescence. ^ At 2:30 p. m., a honey-bee appeared and continued its visits for half an hour. The day following was very cold and windy for mid-summer; but the 12th was clear and warm. At 9:05 a. m., I introduced sugar syrup into the expanded flowers. A honey-bee was soon at work, and by 11:00 o'clock the number had increased to three; at 12:45, there were four honey-bees and a V. consobrina; at 2:35 there were five honey-bees and a V. consobrina; and at 6:00 p. m., the wasp and six honey-bees. The number of flowers in bloom was about thirty-five. The weather continued fair on the 13th, and in the morning I found four honey-bees on the flowers. A new supply of sugar syrup was provided, and by 9:10 a. m., there were twelve honey-bees at work. Manifestly, it was needless to continue the experiment further. Thus, without the addition of an agreeable odor, but merely by intro- ducing a supply of an odorless, colorless syrup the visits of honey-bees were induced in great numbers. Although sugar syrup was not again introduced into the flowers, on August 14, 15 and 16 I saw honey-bees examining the inflo- rescence, doubtless remembering their former experience. On September 2, a honey-bee alighted on two flowers and examined others; by this time most, if not all, of the flowers into which syrup had been introduced had wilted. Bumblebees were also seen to visit the flowers occasionally, but not finding nectar, they did not remain long. There were many small Diptera flying about the foliage of the Petunias, but little or no significance was attached to their visits. A small bee of the genus Halictus also alighted on the corollas. Pelargonium zonale Willd., says Plateau, is one of the more noteworthy forms of plants with very brilliant flowers, which are almost wholly ignored by insects; the beds of scarlet Pelar- goniums, commonly called red geraniums, of which there are a profusion in public gardens, permit us to establish this fact each year. A large bed of Pelargonium zonale displayed more than fifty umbels of scarlet flowers; into three umbels on the left side « For the determination of the specific names of these wasps I am indebted to Mr. S. A. Rohwer. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 163 of the bed Plateau introduced the cooked juice of cherries, and in two umbels on the right side anisette was used. Immediately many flies belonging to the families Muscidae and Sarcophagidae, and later two Syrphidae and three wasps were attracted to the odoriferous liquids. The clusters which remained in their natural state are said not to have received a single visit. A large plant of Pelargonium zonale, of the variety called " General Grant," produced in my garden during the larger part of the summer of 1912 numerous bright scarlet umbels. The nectaries had disappeared and the stamens were largely petaloid so that the flowers yielded neither nectar nor pollen; notwith- standing frequent inspections no insect visits were observed during the larger part of the season. On September 23, at 1 :00 p. m., odorless sugar syrup was introduced into two umbels near the center of the plant. From the 23rd to the 26th, no insects found the syrup, which was renewed from time to time as it evaporated. The 26th was warm and clear, and in the afternoon I saw a honey-bee inspect a cluster of flowers near the ground, but it did not alight. The weather continued fair on the 27th, and at 7:00 a. m., there were no insects on the flowers; but at 9:00 o'clock there were, at least, a dozen honey-bees feeding on the syrup, which was speedily consumed. There were six other fully expanded umbels on which there was no syrup, and it was interesting to note how the bees searched them again and again in their efforts to find more of the edible liquid. Two other umbels with a few buds partially open were also carefully exam- ined. Their attention at first was entirely confined to the gaudy flowers, but later they discovered some of the liquid, which had dripped on a few leaves, and removed it. Their number continued to increase so long as I supplied the syrup. Later they flew to a bed of Portulaca grandiflora Lindl., to the inflorescence of which they had never before been seen to pay any attention, and inspec- ted flower after flower but seldom alighted. >» Evidently the bees had learned from past experience to asso- ciate the presence of nectar with conspicuousness, and though they had never found any food in these particular flowers, they had no doubt continued to occasionally inspect them, as in the 29 During a part of the time this experiment was in progress one of the colonies in my apiary was allowed to remove the honey from a few partially filled combs; and it subsequently occurred to me that this probably stimulated the bees to search the flowers more diligently for nectar. 164 JOHN H. LOVELL single instance observed on the 26th, when a bee inspected an empty umbel but failed to visit those containing syrup. After they had found syrup on two of the umbels, they examined all the others very thoroughly, and also other flowers in the garden previously neglected. They discovered the syrup on the flowers long before they did that which had dripped on a few leaves, and the discovery of the latter was incidental to their visits to the flowers. The bright coloration was clearly an advantage in this instance in enabling honey-bees in large numbers to find the odorless sweet syrup. Obviously highly specialized bees are much better adapted for the purpose of such an experiment than the common flesh-flies observed by Plateau. Plateau made many additional experiments in the course of which he introduced odoriferous syrups into the flowers of Lilium candidum L., Passijlora adenophylla Masters, CEnoihera speciosa Nuttal, Linum perenne L., and Convolvulus septum L., with the result that insects in variety were attracted. But it is unnecessary to consider his experiences further since insects in large numbers may also be attracted to conspicuous, neglected flowers by means of an odorless sweet liquid. Since Plateau knew that sugar syrup was odorless it is natural to inquire why he failed to employ it in control experiments. On four occasions he did introduce syrup of sugar into the flowers of Lilium can- didum, in three instances into two flowers and in one instance into six flowers. He says that, as he foresaw, syrup of sugar without odor did not show any power of attraction. >° But a small number of Diptera, as Syritta pipiens, Melanophora roralis, Anthomyia radicum and Calliphora erythrocephala, did find the syrup and profit by their discovery. No information is given as to the length of time the flowers were under observation. The number of visits received, however, was about the same as in the case of Polygonum Convolvulus, when anisette was added to eight groups of flowers on a very warm clear day. Certainly the list of Diptera recorded gave promise that many visits would have been received had the supply of syrup been continued for a longer period. Lilium candidum is a campanulate flower two or three inches long adapted to pollination by hawk-moths, and it is easy to understand that some time might elapse before the deeply concealed syrup was found by Hymenoptera. No s° " Recherches experimentales sur les fleurs entomophiles," etc., p. 19. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 165 mention is made of the use of sugar syrup in any other control experiments, an omission which can hardly be regarded as excusable. It seems desirable, therefore, in this connection to give a few additional instances observed by myself, where the introduction of sugar syrup resulted in frequent visits of bees. A group of Zinnia elegans Jacq., in my garden, was almost wholly neglected by insects. On the morning of August 16, I introduced syrup of sugar into several capitula, renewing the supply the fol- lowing day. During the forenoon of the 17th, a honey-bee examined the ray flowers of two ungarnished capitula, and then, coming to a head, containing syrup, sucked for a short time. Later a worker of Bombus consimilis found the syrup. At 12 :30 p. m., there were on the flowers two honey-bees and two worker bumblebees, Bombus consimilis and B. terricola. At 3:30 p. m., there were seven honey-bees and one bumblebee on the flowers — there was syrup in a dozen capitula. The honey- bees also examined the capitula which remained in their natural condition. The experiment was now discontinued. Three days later, on August 20, a honey-bee, undoubtedly one of the for- mer visitors, examined many capitula; evidently it remem- bered its previous experience. The brilliantly colored flowers of the scarlet runner, Phaseolus multiflorus Willd. var. coccineus Lam., contain nectar; but owing to the difficulty of depressing the carina, are much neglected by insects. Occasionally in this locality females of Bombus fervidus visit the flowers legitimately. I have also seen a honey- bee for several hours fly from flower to flower inserting its tongue in the opening beneath the standard, and apparently able to reach a very small quantity of the nectar. On the morning of August 16, I put sugar syrup on a few corollas, and during the afternoon there were always from four to six bees on the flowers. They also inspected flowers on which there was no syrup. On the 17th, I renewed the supply of syrup and the bees continued their visits during the entire day. Honey-bees have not sufficient strength to depress the carina and obtain the nectar normally; but if the nectaries are punc- tured they will then visit the flowers in great numbers. Every year the scarlet runner is under cultivation in my garden, but I have never known bumblebees to bite holes in the flowers except 166 JOHN H. LOVELL in 1908. On August 14 of that year, the vines were in full bloom, and there were present many workers of Bombus terricola, which perforated the flowers as fast as they matured — so far as I could discover not a single blossom escaped. The holes were all on the under side of the calyx on the left hand side, which may be explained by the fact (also observed by Muller'i) that the more powerful bees almost invariably alight on the left ala. The honey-bees promptly discovered the holes and used them most diligently for extracting the nectar. There was no pretence on the part of either honey-bees or bumblebees of making nor- mal visits. The absence of bees from the flowers of the scarlet runner does not, therefore, prove that their brilliant hue is of no advantage, or that an agreeable odor is required, for it is only necessary to render the nectar easily accessible by punctures to induce the visits of bumblebees and honey-bees in great numbers. The correlation existing between the accessibility of nectar and the number of honey-bees present is also most instructively shown by the inflorescence of red clover, Trijolium pratense L. The flowers are pollinated chiefly by bumblebees, which are frequent visitors, and in their absence are largely sterile. An historical illustration is the well-known experience of the agri- culturists of New Zealand, in which country at the time of its discovery there were neither honey-bees nor bumblebees. In consequence the yield of seed did not become commercially profitable until in 1855, when about one hundred bumblebees were imported from Europe. 32 The nectar of red clover is secreted at the base of a tube a little over 9 mm. in length, where it is beyond the reach of the tongue of the honey-bee. This has occasioned much regret among bee-keepers, for the flowers not only secrete nectar very freely but the quantity is not greatly affected by weather con- ditions. Repeated attempts have been made to develop a strain of red clover bees, but the gain in tongue length has invariably 31 Mliller, H., " Fertilization of Flowers," p. 216. Both honey-bees and bumble- bees almost invariably alight on the left ala. The reason for this is that the spirally coiled carina closes the entrance beneath the standard on the right hand side. Usually the alae stand apart, but when one occasionally overlaps the other, honey- bees alight on the center. Bumblebees visit the flowers of the common, garden bush beans in a similar manner. "Knuth, P., " Handbook of Flower Pollination," translated by J. R. Ainsworth Davis, 2:292. Jarvis, P. D., " Bumblebees that Fertilize Red Clover," Rep. Ent. Soc. Ont., 36:128. Graenicher, S., " New Zealand's Experience with the Red Clover and Bumblebees," Bull. Wis. Nat. Hist. Soc, 8:166. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 167 proved only temporary. Under normal conditions, then, honey- bees do not frequently resort to the red clover fields; but occa- sionally in very dry weather the floral tubes become so short that large yields of honey are obtained. Two or three times during the last thirty years at Borodino, N. Y., red clover has been a very valuable source of honey; and one season fully sixty pounds, on an average, to a colony was secured. ^ An apiarist in Michigan reports that in one year his bees stored 500 pounds of pure red clover honey as surplus." The black bees stored none, the hybrids only a little, while the bulk of the 500 pounds was gathered by Italian bees. The length of the tongue of the common black bee is 6 mm., of the pure Italians, not over 7 mm., while that of the hybrids is intermediate. Thus there was pre- sented the singular spectacle of fields of red clover visited by thousands of Italian bees, while the black bees were absent. Had the drought shortened the corolla tubes another millimeter the nectar would have been accessible to black bees, and they, too, would have been present. But undoubtedly the most remarkable illustration ever recorded of the relation of rainfall to the length of the corolla-tubes, and consequently of the accessibility of the nectar to honey-bees, was observed by an apiarist at Medina, Ohio. Of two apiaries belonging to him one is located near Medina, and the other two miles north of that city. A few years ago (1906) there was a drouth at the north bee-yard, and the floral tubes of the red clover were so much shorter than usual that honey-bees were able to reach the nectar. When one of the farmers began to cut his field of red clover that season, the cutter knives of the mower stirred up so many bees that they attacked the horses and their driver. So numerous and pugnacious were they that it looked as though they would prevent anyone from cutting off their supply of honey. Singularly enough at Medina and the south bee-yard, there was an abundance of rain. Here, when he went over a big field covered with a luxuriant growth of red clover scarcely a bee could be found. The corolla- tubes were so long that the bees could not obtain the nectar, and consequently, there were none on the clover heads. Thus two bee-keepers, living only a few 33 Doolittle, G. M., " Honey from Red Clover," Gleanings in Bee Culture, 34:993. 34 Hutchinson, W. Z., " Red Clover," The Bee- Keepers' Review, 21:342. 168 JOHN H. LOVELL miles apart, might have arrived at diametrically opposite con- clusions as to the value of red clover as a honey plant. 35 It is clear that the presence or absence of honey-bees in large numbers on the flowers of red clover is not determined by the color or odor, but by the accessibility or inaccessibility of the nectar. Drouth may not render the nectar accessible more than once in ten years, but when it does happen, the bees promptly avail themselves of the opportunity. Evidently they must inspect the flowers each season, but, finding no booty, they do not often repeat their visits. The utter inconsistency with the facts of the claim that the absence of insects from certain con- spicuous flowers proves that bright coloration is of no advantage and that an agreeable odor is a necessity, could not be better shown than in the instance where the Italian bees were able to obtain the nectar and the black bees were not. The flowers of alfalfa, Medicago sativa L., a leguminous plant very extensively cultivated in the west for forage, offers very similar phenomena. In the irrigated regions of California and Colorado, nectar is yielded so abundantly that alfalfa surpasses all the other local honey plants in importance, even the famous purple, black and white sages of the former state. But in Kansas, for example, the results are strikingly different. In the Western part of the state along the river bottoms the flowers can usually be depended on for nectar during most of the season, while around Topeka, bees only occasionally visit the bloom. A bee- keeper who has lived in Eastern Kansas for thirty-five years states he has never seen a bee on the flowers, or known of a pound of alfalfa honey being produced in that section. >* Where alfalfa, then, secretes nectar freely the vast acreage is constantly the resort of millions of bees; but in localities where it is nectarless, "Root, E. R., "Red Clover as a Honey Plant," Gleanings in Bee Culture, 34: 990. The three apiarists cited in this article are .careful observers and recognized authorities on bee-culture. Buttel-Reepen has remarked: " It seems to me that the biological knowledge concerning Apis mellifica which has been gained by practical bee-keeping has scarcely entered scientific literature ... In proof of this there are the vague, defective assertions which are found in the newest editions of scientific works." "Are Bees Reflex Machines," p. 1. » Root, E. R., " Bee-keeping in the Semi-arid Regions of Oklahoma, Kansas and Nebraska," 41 :345. In the eastern states of North America, white clover, Trifolium repens L., is the foremost honey plant, and the domestic bee stores from its bloom annually hundreds of tons of an excellent, white honey; but in France and Switzer- land it yields no appreciable quantity of nectar and one may travel several kilo- meters and rot see a bee on it. " White Clover in Europe," Am. Bee Journal, 53:331. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 169 their visits are so rare that the flowers appear to be entirely deserted through a long series of years. Honey-bees do not usually visit the flowers legitimately, but procure the nectar through a crevice in the side. An excellent illustration on a scale of great magnitude showing that honey-bees are guided by the memory of past experience in gathering nectar is furnished by the honey-flow of buckwheat, Fagopyrum esculentum Moench., which Buttel-Reepen describes as follows : " If colonies stand in buckwheat, the flight is lively in the mornings until ten o'clock; then it lessens, and is entirely quiet for the greater part of the day, beginning vigorously again the next morning. The buckwheat nectar flows only in early morning; so, as the nectaries dry up, the bees fly out a couple of times and then discontinue their vain flight. In spite of the shimmering sea of flowers, in spite of the strong fragrance, only a few bees may usually be found after ten o'clock in the buck- wheat field." »» The period of time during which the flowers of buckwheat secrete nectar varies in different localities. In this region the bees continue to work on them, according to observations made by a young friend of the writer, until about 12 :30 p. m. Their visits then quickly decrease in number until about 1 :00 o'clock, when they cease entirely. But for an hour or more afterwards, the bees may be seen occasionally flying from blossom to blossom, pausing, however, for only an instant, as they apparently dis- cover at once that the flowers are now nectarless. At Delanson, N. Y., buckwheat yields nectar most abundantly between 9:00 o'clock in the morning and 2 :00 p. m. A bee is seldom at work on it much earlier or much later, notwithstanding there are hun- dreds of colonies of bees in the vicinity. In parts of the west, buckwheat is a more uncertain honey plant than in the east and in some years the flowers fail to become nectariferous, when they are almost wholly deserted by bees. 3 . Again a sudden shower fol- lowed by a fall in temperature may bring the buckwheat harvest to an abrupt and premature close in August, when ordinarily it would continue into September. Such an interruption of the » Buttel-Reepen, H. V., "Are Bees Reflex Machines ? " translated by Mary H. Geisler, p. 29. » Root, A. I. and E. R., " The A B C and X Y Z of Bee Culture," p. 71. 170 JOHN H. LOVELL honey season occurred at Delanson in 1906. For several days a hive on scales had shown a gain of eight pounds a day; but dur- ing the night of August 24 there was a light shower and a de- cline in temperature of 1 1 degrees F., after which the hive on scales did not show a gain of half a pound any day that fall. The bees immediately ceased visiting the flowers, and in countless thousands attempted to rob each other and the honey house. 39 Owing to the intermittent nature of the flow of nectar, bees are more irritable during the buckwheat harvest than during that of any other plant. The time of the flight of the bees thus always coincides with the period of active secretion of nectar, or if the flowers are nectarless they neglect them almost entirely. The preceding experiments and studies of honey plants show that honey-bees learn from observation and are guided by the memory of past experience. Flowers rich in accessible food supplies receive numerous visits, but if for any reason the flow of nectar suddenly ceases the bees immediately discontinue their visits. If the inflorescence of a plant species yields abundant nectar in one locality but is devoid of nectar in another, even though only a few miles intervene, the flowers in the former place will be frequently visited and in the latter deserted. But honey-bees do occasionally visit and examine conspicuous flowers from which they can not obtain food materials, and it is upon this premise that the argument of the present paper is based. A. priori reasoning alone would lead the floroecologist, who believes that conspicuousness is an advantage to flowers to this conclusion, thus Campbell remarks that "it is safe to say that no showy flower is entirely destitute of insect visitors." <« Much evidence has already been adduced in support of this statement, but it is desirable to give additional observations, made especially for this purpose. The casual observer will often fail to discover a single visitor, and may easily conclude that they never attract the attention of insects; but long continued investigation proves this to be a mistake. The variegated flowers of the Sweet William, or bunch pink, (Dianthus barbatus L.), display the most vivid shades of crimson and scarlet; and, as the name indicates, exhale a pleasant fra- il' Alexander, E. W., " Buckwheat as a Honev Producer," Gleanings in Bee Culture, 35:394. "Campbell, D. H. ( ' 'Plant Life and Evolution," p. 227. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 171 grance. They are adapted to pollination by butterflies and day-flying moths. The nectar, while not abundant, is sufficient in quantity to yield a sweet taste to the tip of the tongue; and it lies at the bottom of a calycine tube 15 mm. long, far beyond the reach of honey-bees. Previous to July 11, 1912, I failed to record a single bee visitor. On this date I saw a honey-bee inspect several clusters of flowers, but it never actually alighted, although flying close to the inflorescence. On the 23rd, a honey- bee visited a few flowers. At about 11:00 a. m., August 6, a warm clear day, two and at one time three honey-bees were observed on the flowers. They were carefully watched for ten minutes, and one of them vainly endeavored, standing in various positions; to reach the nectar by thrusting its tongue down the center of the flower. Others probed between the petals, even looking under the corolla. An hour later, a bee was still found on the clusters; at intervals, wasps and flies also examined the flowers. Observations extending through the entire season show that the flowers are very far from being wholly neglected by Hymenoptera and Diptera, although a few inspections might readily lead to this belief. The flowers of the bee larkspur (Delphinium elatum L.), which are normally pollinated by bumblebees, have so long a spur that the nectar is wholly inaccessible to honey-bees. In my garden they are very rarely visited by insects of any kind. On the morning of July 11, a honey-bee after visiting one or two flowers, desisted from its useless efforts. On July 24, in the afternoon, a honey-bee visited several flowers in an unsuccess- ful attempt to find nectar. It pushed its tongue as far as possible into the mouth of the spur, and also looked for nectar under the upper perianth segment. On August 4, a bee inspected two blue floral leaves, which had fallen from a flower to the green foliage, thus showing that a single detached petal could gain its attention. On July 16, a large moth poised before several flowers and obtained the nectar without difficulty; in the evening the white center contrasts so strongly with the blue ground color that the attention of crepuscular Lepidoptera might easily be gained. During the summer of 1910, no insects were seen to visit the flowers of the pansy (Viola tricolor L.). By October 1, nearly all the wild and cultivated flowers had perished, but a few 172 JOHN H. LOVELL pansies still remained in bloom. October 7 was cold and rainy, but the day following was clear, warm and calm, and at 10 a. m., a honey-bee spent more than ten minutes on the pansy flowers searching for nectar. Two Syrphid flies {Eristalis tenax) were also flying from flower to flower looking for pollen, but making no attempt to find nectar. On the afternoon of the 10th, a worker of Bombus consimilis and a male of B. jervidus were examining the flowers for nectar; and on the 11th a worker of B. consimilis and a white butterfly. Thus the pansies are not so much neglected as at first appeared probable, but in the absence of more desirable flowers are frequently visited by insects. On the morning of October 28, 1912, two honey-bees were examining the larger, neutral flowers of Hydrangea paniculata Sieb., but they soon learned that they were nectarless and passed over to the smaller, perfect flowers. On July 16, a species of Megachile visited two flowers of the climbing honeysuckle (Lonicera Periclymenum L.), a hawk-moth flower; but its stay was very brief, as it could not reach the nectar. It then flew to another moth flower {Oenothera biennis L.), which was closed. Finding no opportunity to get flower food it returned to the honeysuckle; but meeting with no better success than on its previous visit, it abandoned that part of the garden altogether. In the evening, while the hawk-moths were industriously at work on the honeysuckle flowers, they repeatedly inspected large, red roses blooming on a bush a few feet away. The roses are pollen flowers and devoid of nectar, but the hawk-moths were compelled to learn this fact by direct examination. Another pollen flower is the poppy, but before the anthers dehisc honey-bees may often be seen searching for nectar at the base of the petals. Honey- bees have likewise been observed looking for nectar under the calyx segments of flowers belonging to the Labiatae. Further examples that honey-bees occasionally examine carefully flowers, which are commonly neglected, might be multiplied indefinitely; but sufficient instances have been given for the purpose of the present paper. It has been shown that such visits are actually made, and that they are infrequent because the bees remember their inability to obtain flower food. Nevertheless, in the aggregate they do waste much time in fruitless visits to a great variety of flowers, which for one reason or another CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 173 yield no booty; but this loss is reduced to a minimum by their ability to learn from experience. They are able to store up in their brains, as described by Forel, various sense impressions of color, form and spatial position, by which their movements are subsequently guided and which prevent them from indefinitely making useless visits. ' It results, therefore, from the unanimous observations of all the connoisseurs that sensation and percep- tion, and association, inference, memory and habit follow in the social insects the same fundamental laws as in the vertebrates and ourselves."^ But he adds that "the above mentioned faculties are manifested in an extremely feeble form beyond the confines of the instinct-automatism stereotyped in the species." ^ In closing this paper it is desirable to remind the reader that the visits of bees to flowers are, of course, often influenced by other factors besides the characters of the flowers, as temperature, rainy or foggy weather, the numb.er of insects in the locality, and especially by the blooming period of common plants very rich in nectar. During the honey-flow from the more important honey plants, bees restrict their visits very closely to a single species, and there is no occasion nor would it be for their advantage to pay attention to flowers containing little or no nectar. Plateau himself noticed that when the apricots expanded their flowers, the Hymenoptera abandoned the violets, and he was forced to discontinue his experiments with artificial flowers. <> During a honey-flow the entire force of field bees of each colony is largely governed by a common impulse, and their attention maybe fairly termed obsessional. The hives may then be opened and the honey exposed with scarcely any danger of robbing. Buttel- Reepen tells of a bee-keeper who placed a dish of honey over his strongest colony during the buckwheat honey-flow, and after eight days of good forage the bees had not touched the honey, although it was pure. ** Manifestly, under these conditions » Honey-bees will not visit bright-hued pieces of paper or cloth, whether large or small, attached to a line and suspended over a bed of flowers, or crude floral groups painted on large screens or walls, because they are not deceived by these objects, or images, any more than ourselves. Cf. Plateau, F., " Le Macroglosse," Mem. Soc. ent. de Belgique, 12:141-80, 1906. " Forel, August, "Ants and Some Other Insects," translated by William Morton Wheeler, p. 21. "Plateau, F., " Les fleurs artificielles et les insectes," Mem. de I' Acad. roy. de Belgique, 1:24, 1906. « Buttel-Reepen, H. V., "Are Bees Reflex Machines," p. 27. 174 JOHN H. LOVELL* small groups of conspicuous, nectarless flowers, and even those containing nectar, will be likely to be passed over unheeded. CONCLUSIONS Entomophilous flowers are usually characterized by the pos- session of either bright coloration, or odor, or both, although apparently to some extent the two qualities are mutually ex- clusive. Both allurements are useful in attracting the attention of insects; but the absence of either conspicuousness, or odor, or both, will not necessarily cause a flower to be neglected if it con- tains an ample supply of pollen or nectar. But under similar conditions, small, green, odorless flowers, even if rich in nectar, will not be discovered as quickly as nectariferous flowers, which are conspicuous or agreeably scented. <> On the other hand, the possession of both color and odor will not ensure frequent visits in the absence of available food materials. The experiments afford no evidence that bees visit flowers for the purpose of experiencing an aesthetic pleasure. Insects, especially bees, occasionally examine the neglected, conspicuous flowers of cultivation; but, obtaining no food mat- erials, or very little, they do not often repeat their visits. Many neglected flowers are pleasantly scented, and the addition of another agreeable odor is neither necessary nor beneficial. When odoriferous fruit syrups are introduced into conspicuous flowers, commonly neglected, a group of miscellaneous insects, especially Diptera, will be attracted; but the inference that, therefore, color is no advantage and that an agreeable odor is required is fallacious. For the introduction of an odorless syrup into similar flowers will induce insect visits in large num- bers; also when flowers, with the nectar inaccessible to honey- bees and, consequently, seldom visited by them, have the nectaries artificially punctured, or the floral tubes shortened by drouth, they are then visited by bees in countless thousands without the addition of either an agreeable odor or a sweet liquid. Flowers which in one. locality freely secrete nectar and are visited by numerous insects are sometimes in other localities nectarless and almost entirely neglected. Insects, therefore, perceive the colors and forms of neglected flowers, and the rarity of their «Lovell, John H.. " The Pollination of Green Flowers," Amer. Nat., 46:83-107, 1912. CONSPICUOUS FLOWERS RARELY VISITED BY INSECTS 175 visits is the result of their memory of the absence of food materials, not because the flowers lack an agreeable odor, which is often not the fact. The flowers into which Plateau introduced odoriferous sweet liquids were thus artificially converted into distinct physio- logical varieties. Since flowers possessing conspicuousness, an agreeable odor, and a liquid food were opposed to flowers pos- sessing only conspicuousness, it is clear that color was never directly brought into competition with odor — the latter was invariably given the advantage. Colors and odors attract the attention of insects, but bees in their visits to flowers, previously examined by them, are guided largely by the memory of past experience; they are able to associate different sense impressions and unconsciously make analogous inferences. THE HARVARD LABORATORY OF ANIMAL PSYCHOL. OGY AND THE FRANKLIN FIELD STATION ROBERT M. YERKES With two figures It is now fifteen years since the Director of the Harvard Psychological Laboratory, Professor Hugo Miinsterberg, made a place for experimental work on the psychology of infra-human organisms in his laboratory. In 1899, two rooms in Dane Hall were assigned to students of animal psychology, and under the direction of the writer, three investigations were conducted. To meet the needs of an increasing number of workers, an additional room was made available in 1902. In December, 1905, the laboratory equipment; together with all experimental work in psychology, was transferred to a newly and specially planned and constructed laboratory in Emerson Hall. Here, five rooms, in addition to the Instructor's office and a large amount of space in an unfinished attic, were available for work with animals. The following account of the facilities afforded for this work is quoted from a description of the Harvard Psychological Laboratory, published in 1906 :i ' Several rooms are fitted up with special reference to the investigation of the various forms of organic movement, animal behavior and intelligence. As one result of several investigations in animal psychology already pursued here, the laboratory has a considerable number of devices for testing and making statistical studies of the senses and intelligence, methods of learning and emotional reactions of animals. "Adequate provision is made for the keeping of animals in a large, well-lighted, and well-ventilated corner room. Instead of having aquaria built into the room, an aquarium-table eighteen feet long has been constructed to support movable aquaria of various sizes. Whenever it is desirable for the purposes of an investigation, any of these aquaria may be moved to the research- ■ The Harvard Psychol. Studies, vol. 2, p. 35. 176 THE HARVARD LABORATORY OF ANIMAL PSYCHOLOGY 177 room of the investigator or to such quarters as the special con- ditions of the experiment demand. ' The vivarium-room contains, in addition to provisions for water-inhabiting animals, cages of a variety of forms and sizes. The largest of these cages, six and a half feet high, six feet wide, and four feet deep, may be used for birds, monkeys, or any of the medium-sized mammals. Cages for rabbits, guinea-pigs, and other small animals are arranged in frames which support four double compartments. Similarly, small cages suitable for mice, rats, and other small rodents are in supporting frames which carry four of the double cages, each of which is removable and may be carried to the experimenting-room at the convenience of the experimenter. ' Ina large unheated room above the main laboratory are tanks for amphibians and reptiles. These tanks, since they can be kept at a low temperature during the winter, are very convenient and useful for frogs, tortoises, and similar hibernating animals." Work progressed satisfactorily in these quarters until the spring of 1913, when the introduction of experimental work in Educational Psychology, rendered desirable a redistribution of space. During the summer of 1913, the unfinished fourth floor of Emerson Hall previously referred to was developed, in accord- ance with plans prepared by the writer, as a laboratory of animal psychology. The floor plan of this new laboratory is presented in the accompanying figure 1. Ten rooms, in addition to an office for the director of the work, are now at the service of students of animal psychology. Of these rooms, several were especially planned and have been at least partially equipped for definite lines of inquiry. Thus rooms 40 and 41 have been built about the Yerkes and Watson apparatus for the study of the several aspects of vision in animals. Preliminary studies of vision by simpler rough and ready methods are conducted in other rooms of the laboratory, or at the Field Station described below, and the more elaborate apparatus is used only for accurate and thorough-going investigations. By means of our varied visual equipment, it is possible to study color, intensity, size, form, and distance perception with a degree of exactitude which heretofore has been exceptional in connection with studies of animal behavior. Room 42 is equipped with the Watson circular maze and the 178 ROBERT M. YERKES □ 2 □ n: u > L/l CI I < >> bfi O "o .c o >» 1/1 a O >> "O ■a +-> U 3 >. i- o +-> 5 o < f n > 03 ffi >- £ cr □ n h- -»-> < o rr CS r i a LU o n < E 1-5 w « § fc THE HARVARD LABORATORY OF ANIMAL PSYCHOLOGY 179 Yerkes and Kellogg graphic record device. The latter enables an observer to obtain accurate records of distance and errors, in addition to those of time, in all maze experiments. Thus, the value of the maze-method is 1 trebled. This improved apparatus demands stability, and, although it may readily enough be moved from room to room, it is eminently desirable to have a suitable place reserved for it, so long as the maze method maintains its present importance and promise as a comparative method and offers so many obvious possibilities of improvement. The rooms numbered 43, 44 and 45 are daylight rooms as is also 42, which may be employed as occasion demands. At present, two of them are used for studies of problems of heredity in rats and mice. Later, the Hamilton insoluble problem multi- ple choice apparatus and the Yerkes soluble problem multiple choice apparatus will be installed in this group of rooms. These devices demand a special recorder-room. It is our purpose to install the recorder for both outfits in one room while placing the respective reaction devices in separate rooms. These two sets of multiple choice apparatus will render possible in this laboratory or at the Field Station (since we propose so to construct the apparatus that it shall be readily movable) the study of ideational reactions, in a variety of animal types, in such wise as to furnish directly comparable data of reaction. The line of dark-rooms numbered 46, 48 and 49, is especially convenient because it may be used either in sections or as a whole. A supply of compressed air is delivered to room 49, and it is in- tended that in this room, in conjunction with room 48, there shall be installed apparatus demanding air under constant pressure for varied studies of olfaction and audition. A store room, number 47, provides adequate space for supplies in the shape of food stuffs, bedding or litter, small cages, and packing or transportation boxes. Storage space for larger apparatus and materials is afforded by a room to which entrance is given by the doorway indicated in room 42. Finally, room 50 is the " animal living room " of the laboratory. The floor of this room is water proof so that cages and aquaria may be thoroughly washed and the floor flushed at need. In this vivarium are set cages for a variety of vertebrates. At pre- sent, the laboratory is supplied with cages especially designed for mice, rats, guinea pigs, rabbits, cats, monkeys and birds. 180 ROBERT M. YERKES A large aquarium table, upon which any desired form of aquarium may be placed, provides for the housing of amphibians and fishes. The writer's students' training course in animal psychology is conducted in a class-room and lecture room on the third floor of Emerson Hall. The space of the laboratory on the fourth floor is, therefore, wholly available for research. The rooms of the new laboratory are supplied with water, gas, compressed air, and a variety of electric currents. The latter are conveniently delivered from boards located in each room. In every room there are available 110 volt direct and alternating currents, as well as currents from Edison storage batteries which are located in the battery room of the main labora- tory. A conveniently placed and well constructed switch board (S. B. of Figure 1) in the corridor of the laboratory, provides for the distribution of these storage currents. This board is fitted with miniature Weston switch board voltmeter and am- meter, and with taper plugs. Realizing the extreme need for apparatus in animal investi- gations which shall, in a large measure, eliminate the experi- menter from the situation to which the animal is expected to respond, the writer, in planning this new laboratory, has attempted so to arrange spaces that automatic setting, actuating and re- cording devices may readily be placed in rooms adjoining those in which the animal is responding. Heretofore, the majority of students of animal behavior have deemed themselves com- petent and able to observe and record accurately the doings of their subjects. That this, however, is not the case is clearly proved by numerous instances of misobservation and mis- interpretation of reactions. We have, for example, twice dis- covered in this laboratory that dogs which were presumably responding to a definitely arranged experimental situation were actually responding to certain unconscious movements of the experimenter. The only safe and sure way to avoid such risks is to provide mechanical recorders which shall at least enable the experimenter to separate himself widely from his reacting subject. We have striven for flexibility and adaptability in this new laboratory of animal psychology while arranging for the devel- opment, in designated spaces, of specific forms of apparatus. So far as the conduct of experimental work under highly con- THE HARVARD LABORATORY OF ANIMAL PSYCHOLOGY 181 trollable and reasonably controlled conditions is in question, the laboratory, with its instrumental equipment, is excellent. But in addition to the ever present need of the development of new methods and the opportunity for the advantageous in- stallation of new apparatus, the writer has felt as a still more urgent and important need, the supplementation of the laboratory by facilities for field work. It would appear to be self-evident, yet the attitude of many experimental students of animal behavior seems to contradict the statement, that every student of animal life should be familiar with the objects of his interest in nature as well as in the laboratory; that he should possess, as a basis for evaluating the results of experiments, intimate knowledge of the instincts, habits, temperaments, and habitat of whatever type of organism he happens to be using for experimental purposes. The writer is fully convinced that naturalistic observation, or field work, should be held alike by naturalists and experimentalists as of equal importance with experimental observation, and should be regarded as an indispensable supplement to the latter. There are naturalists, to be sure, who decry all observation of animal behavior made under experimental conditions, whether within or without the walls of a laboratory, and there are experimental- ists who deny the value of naturalistic work, or ignore it. But surely the last decade has furnished abundant proof of the unprofit- ableness of these attitudes. We propose, so far as is possible, in connection with our laboratory studies of animal behavior, to attempt to unite the naturalistic and the experimental points of view and methods. The Harvard Psychological Laboratory is particularly for- tunate in having the use of a field station in Franklin, New Hampshire, at which naturalistic studies on any organism which will thrive in a temperate climate may be pursued. This station consists of a tract of about one hundred and fifty acres of hill land, of which about half is wooded. The elevation is fourteen to fifteen hundred feet. There are numerous springs and a brook on the tract. Two sets of old farm buildings are available for such needs as arise. This tract, which is constituted by two old farms, was purchased by the writer in the years 1911 and 1912 to serve both as a summer home and as a reservation which might, as seemed desirable, be used for studies in animal behavior 182 ROBERT M. YERKES It is proposed that this private field station shall meet two keenly felt needs of the Harvard Laboratory; the one, that of a suitable place for purely naturalistic field work; and the other, that of a similarly suitable place for the conduct of laboratory investigations which cannot well be continued during the summer in Cambridge. We may consider, first, the second of these needs. There are frequently in progress, in the Harvard Laboratory, researches on heredity or on problems which demand long ex- perimental training, the interruption of which, during the summer vacation, entails serious loss. It is often impracticable to attempt to continue such investigations throughout the sum- mer in Emerson Hall, for even if the investigator is willing to work there, it usually means a serious sacrifice on his part of opportunity for rest and recreation through a change of scenes. The Franklin Field Station, it is hoped, will result in the saving of considerable time to certain investigators, since there it should be possible to continue work uninterruptedly throughout the summer, while at the same time the investigator may profit by the change from city to country and the chance to combine experimental and naturalistic studies in animal behavior with the recreations of a mountainous country. It is by no means intended that all of the investigations con- ducted in the Harvard Laboratory shall be transferred to the Field Station. Instead, only a few should or can, to advantage, be so transferred. But of primary importance, as contrasted with its value as a place for transferred experimental investigations, is the oppor- tunity which the field station offers for naturalistic work. In and about the Cambridge Laboratory, favorable opportunities for training students to observe animals carefully, critically, and at the same time sympathetically, in their native habitats, are rare. And the writer has observed, in many otherwise admirable students of the biological sciences, a tendency toward the acquisition of a narrow minded attitude toward experi- mental observation, which blinds them to the value of nature- study. It is hoped that at Franklin something may be done for at least a few students of animal behavior to counteract this tendency and to train them to become enthusiastic and reliable naturalists as well as skilled experimentalists. There is no obvious reason why, at the Field Station, any one Figure 2. Views of the Franklin Field Station for the study of animal psychology. THE HARVARD LABORATORY OF ANIMAL PSYCHOLOGY 183 of scores of invertebrates and vertebrates should not be observed under conditions of varying degrees of freedom. The country is already rich in animal life. Indeed, the Pemigewassett Valley, in which the station is located, is well-known to orni- thologists because of the abundance of birds. It will undoubtedly prove feasible, as occasion arises, to import organisms for study. The station is at present available during the months of June, July, August and September. It is at this time that work in Cambridge can least satisfactorily be conducted. The climate at Franklin is healthful and agreeable. For a few days in July the heat is at times oppressive, and before the end of September, frosts are likely to chill the enthusiasm of the field worker and to encourage his return to the city laboratory. Only a small group of observers can be received at the Field Station during the summer. Each individual is responsible for his living expenses, and for the present at least, he must be res- ponsible, also, for such expenditures as the conduct of his work demands. The station, as stated above, is the property of the writer, and is by him, in his capacity as director of studies in animal behavior at Harvard, placed at the service of a selected group of investigators during the summer. Behavioristic work was initiated at the Franklin Field Station in the summer of 1912, by a study of habit -formation in earth- worms, conducted by Ada W. Yerkes and the writer. This was a continuation of work begun previously in the Harvard Labora- tory. Two investigations were pursued during the second season (June to October, 1913) by Mr. C. A. Coburn and the writer. Of these, the one was a study of the transmission of savageness and wildness in mice, and the other, a naturalistic and experi- mental investigation of the behavior of the crow. The first of these was transferred for the season from the Harvard Labora- tory. The second was a new inquiry which indeed could be con- ducted to advantage only at the Field Station. Both of these investigations prospered most encouragingly during the season, and we confidently expect and hope that they may be continued during the coming summer. Mr. Coburn, in a paper which appears in this number of the Journal (p. 185), has given a preliminary account of the results of certain of his experiments with crows. The naturalistic data which we obtained are reserved 184 ROBERT M. YERKES for later presentation in connection with observations which we hope to make next summer. This paper on the crow initiates a series of contributions from the Franklin Field Station which should, in invaluable ways, supplement our studies from the Harvard Laboratory. THE BEHAVIOR OF THE CROW, CORVUS AMERICANUS, AUD. CHARLES A. COBURN From the Harvard Psychological Laboratory and the Franklin Field Station, Franklin, New Hampshire Many years ago, Henry Ward Beecher remarked that if men were feathered out and given a pair of wings, a very few of them would be clever enough to be crows. This statement represents in a general way the opinion of the mental ability of the crow held by many students of bird life. The literature, both early and late, abounds with anecdotes depicting the in- tellectual superiority of the crow over other birds. During the last two decades investigations have been made, by the United States Department of Agriculture and several state boards of agriculture, to determine whether the battle waged by the farmer against the crow is justified. The results of these studies tend to show that the value of the crow to the farmer by its destruction of injurious insects, mice and other rodents, more than compensates for the injury it does to the growing crops. These studies have also provided interesting data on the habits and mental characteristics of the crow. The data, derived in this manner, in no way contradict the general impression. It is, in general, indicated that the crow is very intelligent, supremely cautious and suspicious. Forbushi states that, in his opinion, it naturally is neither very cautious nor suspicious, but bold and fearless. Its apparent traits have been acquired by force of necessity. The reason for his statement is that on the Pacific Coast, especially during the early period of settlement, the crows were extremely bold and unsuspicious. No definite study of the mental ability of the crow was made until 1910, when James P. Porter > used three crows in his in- vestigation of intelligence and imitation in birds. His results ■ Forbush, E. H. Useful Birds and their Protection. Published under the direc- tion of the Massachusetts State Board of Agriculture. 1907. 2 Porter, James P. Intelligence and Imitation in Birds: A Criterion of Imita- tion. Amer. Jour, of Psychology, 1910, vol. 21, pp. 1-71. 185 186 CHARLES A. COBURN did not put the crow on a higher plane of intelligence than sev- eral other birds, especially the English sparrow. In co-operation with Professor Robert M. Yerkes, an inves- tigation of the intelligence of the crow was begun in June, 1913, at the Franklin Field Station. Work was continued until late September. It is planned to continue the investigation in suc- ceeding summers under the favorable conditions of the station. The first summer's work included a general study of the habits and development of the bird (to be reported after addi- tional data have been obtained) and a preliminary examination of its ability to discriminate brightnesses, sizes and forms. It soon became apparent that the adaptation of an apparatus and method to the extremely wary and suspicious nature of the crow was a more difficult task than had been anticipated. This was accomplished after much experimenting with different methods of procedure and many changes in the apparatus. By the time both method and apparatus were fairly well adapted to the characteristics of the crow, the summer was well gone. Our results are only approximations to the crow's discrim- inating ability. They are of value, however, in that they indi- cate certain important tendencies. A comparison of the results obtained during the first weeks with those obtained the last few days clearly shows the effect of improvement in method. Two crows were used in these experiments. They were taken from a nest near the Field Station on the 6th of June. They were then, probably, about two weeks old. Number 1, a male, was larger and better developed when caught. When full- grown it was larger and bolder and less easily frightened than the female, Number 2. For two or three weeks after they were caught, the young birds were fed earthworms, with an occasional bit of cooked cereal. Gradually this diet gave way to various kinds of meat, bread soaked in milk, cracked corn soaked in water, and table scraps. The development, care, and feeding of young crows, will be discussed in a later paper. When the two crows were about nine weeks old, they were able to fly a short distance and to eat alone. They were so tame that they recognized the voice of the experimenter and would come when called, perch on his arm or shoulder, and eat from THE BEHAVIOR OF THE CROW 187 his hand. This friendliness was shown to no other person, and an entire stranger would frighten them very much. Four other crows were obtained from Pennsylvania, but they were too wild for use in the investigation. The building, in which the experimenting was done, was divided into two compartments, each 10 feet by 12 feet. One of these served as a roost and feed-room. Adjoining this room was a fly, 24 by 10 by 8 feet high, made of chicken wire. The crows could fly direct from the roost to a perch in the far end of the fly. The other compartment, which served as an experiment- room, was set off from the feed-room by a partition of chicken wire and a burlap curtain. The curtain could be pulled aside when experiments were not in progress, thus allowing a free circulation of air. The apparatus used was a modified form of the discrimina- tion-box used by F. S. Breed » and later by L. W. Cole « in their studies of the reactions of chicks to visual stimuli. The following description is intended to give only the essential points of the apparatus. For a more detailed account, reference may be made to the reports of Breed and Cole. The entrance-chamber was a movable box 18 by 16 by 14 inches deep. The top, bottom, and three sides were of one-half inch boards. The fourth side was covered with wire netting, one-fourth inch mesh. In each end were openings, 7 inches by 9 inches, with horizontal slide doors. Leaving the entrance-box, the crow entered the discrimina- tion-chamber. This was 16 by 19 by 13 inches deep. The top was of wire, one-fourth inch mesh. Opening directly into this chamber were two chambers, 18 by 19 by 13 inches deep. The tops of these chambers were of wood as were also the sides and floors. The exit from each of these chambers was 7 inches by 9 inches, with horizontal slide doors. They opened directly into two exit -boxes similar to the entrance-box. The front ends of the stimulus-chambers were formed by a three-stimulus plate-shifter sliding in wooden tracks. For a minute descrip- tion of this shifter, the reader is referred to the papers of Breed 3 Breed, F. S. Reactions of Chicks to Optical Stimuli. Jour, of Animal Be- havior, 1912, vol. 2, pp. 280-295 * Cole, L. W. The Relation of Strength of Stimulus to Rate of Learning in the Chick. Jour, of Animal Behavior, 1911, vol. 1, pp. 111-124. 188 CHARLES A. COBURN and Cole. The stimulus plates used in the experiments on size and form, were the standard plates devised by Yerkes and Watson s for their brightness vision apparatus and are described in detail in their paper. The floor, walls, and top of the discrimination-chamber and the two stimulus-chambers were painted a dark gray. This rendered the two stimulus-chambers alike in every way except with respect to the desired difference in optical stimuli, namely, that of brightness, size, or form. Care was taken throughout the work to see that this was the only means by which the crow could choose the correct path. The exit doors were operated by a system of cords. A cur- tain was suspended from the ceiling at the rear of the apparatus. The experimenter, standing behind the curtain and looking through a small peep-hole, could observe the behavior of the crow while in the apparatus and open and close the exit-doors without being seen by the crow. Late in the summer, two swinging gates of wire were sus- pended between the discrimination-chamber and the two stimu- lus-chambers. These gates also were operated by cords. At the beginning of a test they were drawn up to the ceiling of the discrimination-chamber. The purpose of these gates was to prevent the crow from returning into the discrimination-chamber after it had made a wrong choice. During the experiments on brightness discrimination, the apparatus faced a north window. With the beginning of the tests of size discrimination, it was shifted to face a larger south window. In this position, it remained during the rest of the season. The ability of the crow to detect a slight change in the situa- tion, together with its wary and suspicious nature, made it impossible to choose a method of procedure at the beginning and to adhere to it rigidly throughout the period of work. The method used at the beginning was evolved during the prelimi- nary trials, when the first indications were received of what the crow might reasonably be expected to do. Various changes were made in this initial method until a reasonably satisfactory one had been developed. * Yerkes, R. M., and' Watson, J. B. Methods of Studying Vision in Animals. Behavior Monographs, 1911, vol. 1, no. 2, p. 23. THE BEHAVIOR OF THE CROW 189 For several days previous to the first preliminary series, the crows were compelled to enter the discrimination -chamber in order to get their food. For this purpose the apparatus was placed before a small door in the partition separating the feed- room from the experiment-room. At first, the pan containing the food, was placed just inside the entrance door. Then, grad- ually, it was placed farther back until the crows were required to go through the discrimination-chamber, and the one or the other of the stimulus-chambers, into the exit -boxes. After a few days, they did this with no apparent fear. The first preliminary tests were given on July 16th. The crows were then about nine weeks old. The standard stimulus plates had been removed from the stimulus shifter, leaving square openings, 12 cm. by 12 cm. Opal flashed glasses were placed in the slides immediately before these openings, so the illumination of the two chambers was the same. The apparatus was adjusted with the entrance-box before the small door in the partition between the feed-room and the ex- periment-room. When one of the crows had entered this box to get the bit of food placed therein, both doors were closed and the entrance-box was then placed before the entrance to the discrimination-chamber. The door leading to the discrimi- nation-chamber next was opened and the crow allowed to enter. The exit doors being open the crow could proceed to one of the exit-boxes and obtain food. The exit and entrance-boxes were now exchanged and the crow given another trial. Both crows were much frightened by being confined in the entrance and exit-boxes. After two days, with nine such trials, they became somewhat calmer during the experiments. The exit doors were now closed and the crows allowed to enter the discrimination-chamber, go to one of the stimulus-chambers and there wait until the exit door was opened. This new situation, especially the opening of the exit door, frightened them as much as being shut in the entrance or exit-boxes had at the beginning. In the first trial they could not be induced to enter the discrim- ination-chamber until the exit doors were opened as before. However, after eight trials with the doors closed, they had lost much of their fear. In these seventeen trials, Number 1 went eleven times to the right and six times to the left. Number 2 went every time to the right. . 190 CHARLES A. COBURN When the crow chose the correct path, it was always rewarded with a bit of food, — a small piece of mouse, frog, or other meat. If it chose incorrectly, it received no food and was required to remain three or four minutes in the exit-box, which had been previously darkened by a cloth thrown over the wire side. The dislike of crows to remain in a darkened chamber was utilized also by the gradual darkening of the entrance-box when the crows hesitated too long before entering the discrimination- chamber at the beginning of a test. This never failed to cause them to leave immediately. There were, therefore, at least two motives for correct choice, namely, the desire for food and the dislike of the darkened box. The latter can be considered a constant factor, for they reacted to the darkened box as strongly at the end of the summer as they did at the beginning. Care was taken throughout the experiments to keep the fac- tor of hunger constant. It was impossible to do this at all times, and it is highly probable that the results in many cases were materially affected by the change in this factor. In the beginning, two series of five tests each were given per day. The times for the beginning of these series varied slightly, but as a rule they were 7 :30 A. M. and 1 :00 P. M. The crows, with this number of tests, would still be hungry at the end of the series, so the number of tests per series was raised to ten and the amount of food given at the end of each correct choice was lessened. It soon became apparent that the crow, in this case, was confined too long. After the seventh or eighth test, it usually busied itself more with getting out of the apparatus than with choosing the correct path in order to get food. On this account, three series, (7:30 A. M., 12:00 M. and 4:00 P. M.), of eight tests each were given per day. Finally the number of trials in each series was changed to five, and this seemed to be the best solution of the problem, as the crows were sufficiently hungry three times a day to be eager to get food. In the major- ity of cases, they were still hungry at the end of a series. The time required for the five tests was rarely over ten minutes, and the crows, as a rule, did not become restless in this time. As a rule, one crow was given all the trials of a series before the other was caught. In a few series, the crows were given alternate tests. This was not conducive to the best results, for the crow, waiting in the entrance-box until the other completed THE BEHAVIOR OF THE CROW 191 the test, would become so restless that in many cases it would begin to throw itself against the woven wire side of the box. By the time its turn came, the desire to escape from the box had entirely overcome the desire for food, and, as a result, it would rush through the test and recommence its struggle to free itself. If, by chance, it made a correct choice, the food would not be noticed. The results of each series of tests were kept on record sheets similar to those used by H. C. Bingham • in his study of the perception of size and form in the chick. In addition to a record of the correct and incorrect choices, the time required for the choice and a sketch of the path of the crow were also recorded. In the study of brightness perception, the apparatus remained as in the preliminary series except that the stimulus areas of the stimulus-chambers differed in intensity. This difference was obtained by the use of more or less opaque substances, namely, black cardboard, milk glasses, and paper. These were placed over the opal flashed glass of one of the stimulus areas. The slides, which held the plates of opal flashed glass before the stimulus areas, were large enough to admit also the card- board, milk glasses, or sheets of paper. Black cardboard was first used. Since it allowed no light to pass, the illumination of the stimulus area before which it was placed was practically zero. The crows, in the trial series, had become partially accustomed to stimulus areas of an intensity produced by light passing through but one thickness of opal flashed glass. Consequently in the brightness experiments, they avoided the darkened chamber. The chambers were darkened in no regular order, but in ten or twenty tests, one chamber would be darkened as many times as the other. After fifteen tests with each crow, the cardboard was ex- changed for two milk glasses, then later for one milk glass and finally for one sheet of paper. The difference in the intensity of the two areas in this last case was comparatively slight. With care it could be distinguished by the human eye. Table 1 shows the results of these tests. • Bingham, H. C. Size and Form Perception in Galhis domesticus. Jour, of Animal Behavior, 1913, vol. 3, no. 2, pp. 65-113. 192 CHARLES A. COBURN TABLE 1 Intensity Discrimination Correct choices Date No. of tests Crow No. 1 Crow No. 2 • Conditions of Discrimination Cardboard and opal flashed glass — Opal flashed glass July 19 5 5 5 " 19 5 4 5 " 20 5 5 5 Two milk glasses and opal flashed glass — Opal flashed glass July 21 5 5 5 " 21 5 5 5 "22 5 4 5 One milk glass and opal flashed glass — Opal flashed glass 5 5 2 5 3 3 5 4 4 5 4 4 5 5 4 5 3 4 5 4 5 5 5 5 5 4 5 5 5 5 5 5 4 One sheet of paper and opal flashed glass — Opal flashed glass July 29 5 4 3 " 30 5 5 3 " 30 5 5 3 " 31 5 5 3 " 31 5 5 4 Aug. 15 5 4 "15 5 5 These results are but roughly indicative of the crows' ability to distinguish differences in illumination. Accurate measure- ments of the birds' visual acuity was not the aim of our ex- periments. The chief value of these experiments on the discrimination of intensity is the demonstration of the ease with which the crow is able to adapt itself to experimental conditions and to solve accurately one variety of problem. With the beginning of the experiments on size discrimination, the apparatus was so shifted that the front end was immediately before a large south window. In this position it remained dur- ing the season. The only other change was the insertion of the July 22 a 23 u 23 a 24 a 24 a 25 u 25 a 26 a 27 <; 28 u 28 THE BEHAVIOR OF THE CROW 193 standard stimulus plates in the stimulus shifter. Difference in the illumination of the stimulus areas was eliminated. A 5 centimeter » circle versus a 2 centimeter » circle was chosen for the beginning of this study. The correct exit was indicated by the larger circle. This change in the conditions of discrimination naturally threw the crows into confusion. They refused to enter the dis- crimination-chamber unless forced to do so by the darkening of the entrance-box. If this were done, they would pass to and fro before the two stimulus-chambers, but they would not enter far enough into either of them for the exit doors to be opened. The series of the first two days had to be interrupted on account of the crows' fright. On the third day no attempt was made to work. During the day the crows were fed somewhat less than the usual amount of food. The next morning (August 5th), they were tried with a 9 centimeter circle versus a 5 centimeter circle. By this change the illumination of the stimulus-cham- bers was made to approximate that to which the crows had become accustomed in the experiments on the discrimination of intensity. Their hunger, on this day, was great enough to overcome in large measure their fright. The results of this, and the remaining series on size discrimination are given in Table 2. After one series with the 9 centimeter versus the 5 centimeter circle, a 2 centimeter circle was substituted for the 5 centimeter circle. The crows' behavior now became practically normal. The only significant difference from previous reactions was a greater hesitation in choosing. Before finally entering a cham- ber, they would often pass to and fro several times before the two stimulus-chambers, again and again starting to enter one chamber only to back out and go to the other. As appears in the table, crow no. 1 made twenty correct choices in succes- sion, while crow no.. 2 succeeded in choosing correctly eighteen times in twenty. This sudden return of calm and controlled reaction and the high percentage of correct choices, were due probably to the fact that the illumination of the stimulus- chambers through the 9 centimeter and the 2 centimeter circles was closely similar to that in the experiments on intensity discrimination. ' Stimulus plates will be designated by the diameter or the side. 194 CHARLES A. COBURN It seems probable that the birds were simply choosing the more highly illuminated stimulus-chamber, which, in every case, was also the one presenting the larger stimulus area. That they did not continue to use this cue is proved by experiments in which the large stimulus area, and irregularly the small one also, were darkened by placing one thickness of milk glass over the opal flashed glass. This enabled the experimenter in some tests to present two stimulus areas differing in size and intensity of illumination. Now the chamber illuminated by the larger circular area was the more intense, and now the one illuminated by the smaller area. Had the crows attempted to depend upon the illumination of the chambers, or on the relative intensities of the stimulus areas, instead of on their size, they certainly would have been confused. As a matter of fact, the change influenced markedly neither their behavior nor their percentage of correct choices. The experiments on the perception of size were continued for twenty-five days. The results (Table 2) show that the crows TABLE 2 Perception of Size Correct choices Date No. of tests Crow No. 1 Crow No. 2 Conditions of Discrimination 5 centimeter — 2 centimeter circle Aug. a 2 3 Crows frightened. Abandoned series. Crows frightened. Abandoned series. 9 centimeter — 5 centimeter circle Aug. 5 5 3 9 centimeter — 2 centimeter circle 4 Aug. tt 6 6 10 10 10 10 5 centimeter— 2 centimeter circle 8 10 Aug. u « a tt a 7 7 8 8 9 9 10 10 10 10 10 10 10 9 9 7 5 10 5 centimeter — 3 centimeter circle 10 10 9 10 9 9 Aug. tt tt 10 11 11 10 10 10 8 8 9 9 7 10 THE BEHAVIOR OF THE CROW 195 TABLE 2 — Continued Correct choices Date No. of tests Crow No. 1 Conditions of Discrimination Crow No. 2 3 centimeter- -2 centimeter circle Aug. 12 10 9 7 " 13 8 4 4 " 13 10 6 7 " 13 8 8 5 " 14 8 6 6 " 15 7 6 6 " 15 9 7 8 " 15 8 5 5 " 16 8 7 5 " 16 8 8 8 " 16 8 6 5 " 17 8 6 4 5 centimeter- -3 centimeter circle Aug. 17 8 8 6 " 18 8 8 8 " 19 5 3 3 (Left habit) " 19 6 4 6 " 20 10 4 (Left habit) 8 " 20 10 8 9 " 20 10 7 10 " 21 10 9 10 " 21 10 5 (Left habit) 9 " 21 5 3 5 " 22 8 7 7 6 centimeter- -3 centimeter circle Aug. 22 8 8 7 5 centimeter- -3 centimeter circle Aug. 22 8 6 8 " 23 8 3 (Left habit) 8 " 23 8 6 7 " 24 10 9 6 " 24 10 10 9 " 25 10 9 9 " 25 10 9 9 improved surprisingly little with practice. The percentage of correct choices with the 5 centimeter versus the 3 centimeter circle was as low during the last few days of the training as it was on August 10th and 11th when they were first required to distinguish between these circles. Throughout these experiments, the behavior of the crows while working was very erratic. Some days they worked slowly and carefully. Sudden noises, such as those caused by the opening or closing of an entrance or exit door, did not greatly 196 CHARLES A. CO-BURN disturb them. The results on these days of calm steadiness, as a rule, showed an increase in the number of correct choices. On other days, their behavior would be practically the opposite. While still in the entrance-box they would walk impatiently to and fro before the woven wire side of the box. When the en- trance door was opened, they would often start several times to enter only to turn back into the entrance-box. When they finally did enter, they would rush to one of the exit doors, and, in a crouching attitude, wait until it was opened. On these days, great care had to be taken in opening and closing the doors for an unusual noise or sudden movement would greatly increase their excitement. During this behavior they were very likely to develop a position habit. Series, in which this excited behavior resulted in a considerable number of incorrect choices, have been noted in the tables. The ability of the crow to pass directly from one set of cir- cles to another with no great difference in the number of cor- rect choices (see Table 2), was further tested by a series of experiments, the results of which appear in Table 3. In these experiments, the attempt was made to determine whether the crows were reacting to a certain specific stimulus, or whether they were reacting to it because of its relation to another stimulus. For instance, if the 6 centimeter and the 4 centimeter circles were presented, and the crow trained to react positively to the 6 centimeter circle, would it continue to do so when the 6 centimeter circle was presented with a 9 centi- meter circle, or would it, instead, choose the larger area in each instance? As in the preceding series the crows were trained to choose the larger of two circles. When they had gained the ability to choose correctly, they were given ten trials with a different pair of circles. During these ten trials, they were rewarded after each test, regardless of the correctness or incorrectness of the reaction. A reaction was considered correct if the crow chose the larger circle. These series are designated, in Table 3, " rela- tive reactions." The training series which preceded the relative series of August 26th are given in Table 2. The results of these experiments indicate fairly clearly the relativity of the crows' reactions. Especially is this true of crow no. 1. For example, on August 24th and 25th, when the THE BEHAVIOR OF THE CROW 197 3 centimeter circle was presented with the 5 centimeter circle, the crow reacted to the 3 centimeter circle thirty-seven times negatively and. three times positively. On August 26th, the 3 centimeter circle, displayed with the 2 centimeter circle, was reacted to positively in every case. The results for crow no. 1 with the 6 centimeter circle when displayed with the 4 centi- meter and the 9 centimeter circles, on August 26th and 27th, were almost as decisive. TABLE 3 Reactions to Relative Sizes of Circles Correct choices Date No. of tests Crow No. 1 Crow No. 2 Relative reactions, 3 centimeter — 2 centimeter circle Aug. 26 5 5 5 " 26 5 5 3 Training series, 6 centimeter — 4 centimeter circle Aug. 26 10 7 9 Relative reactions, 9 centimeter — 6 centimeter circl Aug. 27 5 4 2 " 27 5 5 3 Training series, 6 centimeter — 4 centimeter circle Aug. 27 8 2 ' 5 "28-10 6 8 " 28 8 5 3 (Righ habit) "29 5 4 .4 " 29 5 4 3 " 29 6 5 5 Relative reaction -, 3 centimeter — 2 centimeter circle Aug. 30 5 5 2 " 30 5 5 4 Tra ! ning se.ies, 6 centimeter — 4 centimeter circle Aug. TO 5 5 2 Relative reactions, 9 centimeter — 6 centimeter circle Aug. 31 5 2 4 " 31 5 4 3 Only one day intervened between the conclusion of the tests of the relativity of the reactions and the beginning of experi- ments to determine the ability of the crow to distinguish circles from triangles, squares and hexagons. ' With the beginning of this study of form perception the experimenter became more convinced than ever that the results, obtained in the previous experiments, did not truly indicate the 198 CHARLES A. COBURN crows' intelligence. A new form of reaction now developed. When either of the crows had made an incorrect choice and the exit door was opened, showing a dark exit-box, instead of enter- ing as they hitherto had done, they would whirl about and quickly go to the other exit and there wait, even for five or ten minutes, until the door was opened. This behavior naturally tended to lower the percentage of correct choices. The experimenter first tried to overcome this difficulty by having the exit-box illuminated until they had entered it. Crow no. 2 would always enter the box under these conditions, but crow no. 1, after a few trials, refused to enter either box unless there was a bit of food in view. To meet this difficulty, the gates, described on page 188, were constructed. When the crow entered the wrong stimulus- chamber, the exit door was opened and at the same moment the gate between that chamber and the discrimination-chamber was dropped, thus preventing the crow from escaping to the other exit. The dropping of the gate tended to frighten them somewhat, so they always quickly entered the exit-box, which was again darkened as in the early experiments. The effect of this improvement in the apparatus on the behavior of the crows appears in the results of Table 4. The crows had been given one hundred and six tests for their ability to distinguish a 6 centimeter circle from an 8.081 centi- meter triangle. During these trials no appreciable increase in the percentage of correct choices had been made. Immediately after the gates were brought into use, improvement commenced and thereafter the majority of the choices were correct. Crow no. 2 did not make quite as high a percentage of correct reac- tions as did crow no. 1. This was probably because no. 2 seemed to be more frightened by the dropping of the gate. If an incor- rect choice was made early in a series, there was a tendency, on the part of no. 2, to avoid that stimulus-chamber during the remainder of that series. The 6 centimeter circle, the 8.081 centimeter triangle, the 5.317 centimeter square, and the 3.29 centimeter hexagon are of equal area. The last thirty tests were with figures unequal in size. The 6 centimeter and the 9 centimeter circles each possess a greater area than the 3 centimeter hexagon, whose area, in turn, is almost twice as great as that of the 3 centimeter circle. THE BEHAVIOR OF THE CROW 199 TABLE 4 Discrimination of Form Correct choices Date No. of tests Crow No. 1 Crow No. 2 Conditions of Discrimination 6 centimeter circle — 8.081 centimeter triangk i Sept. 2 6 5 3 u 2 5 4 3 a 3 5 3 4 a 3 5 4 3 a 3 5 4 3 a 4 5 4 2 « A 5 5 2 it 4-8 45 39 32 it 8 5 5 2 (Right habit) a 8 5 3 4 a 9 5 3 4 a 9 5 3 5 a 10 5 3 1 (Left habit) a 10 5 (Began using gates) 3 5 a 11 5 5 4 a 11 5 5 5 it 11 5 3 5 u 12 5 5 4 it 12 5 5 5 It 12 5 5 5 6 centimeter circle — 8.081 centimeter triangle (Inverted) Sept. 13 5 5 6 centimeter circle — 5.317 centimeter square 5 Sept. 13 5 5 5 u 13 5 5 5 it 14 5 5 3 (Left habit) It 15 5 5 6 centimeter circle — 4.243 centimeter square 5 Sept. 15 5 5 5 a 15 5 4 5 it 16 5 5 5 6 centimeter circle — 3.29 centimeter hexagon Sept. 16 5 4 4 6 centimeter circle — 3.00 centimeter hexagon Sept. 16 5 5 5 " 17 5 5 4 3 centimeter circle — 3.00 centimeter hexagon Sept. 17 5 5 3 " 18 5 5 2 (Left habit) 9 centimeter circle — 3.00 centimeter hexagon Sept. 18 5 5 4 " 18 5 5 5 200 CHARLES A, COBURN The intensities of the stimulus areas and the general illumi- nation of the chambers were varied in these tests by the use of milk glasses as described on page 194. The only visual factor which was constant during the thirty trials, was that of form. It is evident, therefore, that this was the cue which enabled crow no. 1 to make a perfect record in these series. Lack of time prevented further work on the perception of form. The last two days of work were devoted to a further study of size discrimination. The purpose was to obtain, if possible, more conclusive evidence of the crows' ability to dis- tinguish sizes, and, incidentally, to learn if the improvement in the method (introduction of gates) would increase the percentage of correct reactions to differences in size. Thirty tests were given, the results of which appear in Table 5. During the first series, the crows appeared to be confused by the sudden change in the problem presented to- them. They worked rather slowly and quietly, but their choices were not made with the usual defmiteness. It was evident that they (especially crow no. 1), did not clearly appreciate what was required of them. TABLE 5 Discrimination of Size Correct choices Date No. of tests Crow No. 1 Crow No. 2 Conditions of Discrimination 5 centimeter — 3 centimeter circle Sept. 19 10 4 7 " 19 5 5 4 5 centimeter — 4 centimeter circle Sept. 19 5 5 5 5 centimeter — 4.5 centimeter circle Sept. 20 5 4 4 " 20 5 5 5 In the second series, the indefiniteness and hesitation in their behavior were lacking. In every case no. 1 went quickly and directly to the correct exit. Crow no. 2 made a mistake in the first test. Its decisions, however, were made clearly and definitely thereafter. This clear-cut, decisive type of reaction continued, with both crows, during the remaining tests, even when the discrimination was between the 5 centimeter and the THE BEHAVIOR OF THE CROW 201 4.5 centimeter circles. If quickness of choice can be taken as a measure of the ease of discrimination, it is probable that the crows are capable of distinguishing much smaller differences. The crow deserves its reputation. It is an exceptionally interesting subject for the behaviorist and worthy of his greatest skill. As has been indicated earlier in this report, it is planned to observe systematically crows at the Franklin Field Sta- tion, both in the field and in the laboratory, in order that a reasonably complete and reliable description of their behavior may be given. Because of the division of labor among a number of observers, it will be necessary to publish reports from season to season instead of reserving all materials for a monograph. The present paper is indicative of some of the chief character- istics of the bird, and suggestive of experimental difficulties. Another season should prepare us to report on the habits, in- stincts, and development. SOME RELATIONS BETWEEN RHEOTAXIS AND THE RATE OF CARBON DIOXIDE PRODUCTION OF ISOPODS W. C. ALLEE Thompson Biological Laboratory, Williams College AND SHIRO TASHIRO Laboratory of Biochemistry and Pharmacology, the University of Chicago For several years one of us has been working upon an analysis of the rheotactic reaction of the isopod, Asellus communis, Say. In the course of this work it became evident that certain con- ditions known to affect animal metabolism likewise regularly affected the rheotactic reaction of isopods. Thus it was found that low oxygen tension, high carbon dioxide tension, chlore- tone, potassium cyanide, lowered temperature, sudden extreme increase of temperature, starvation, and fatigue decreased the percentage of positive rheotactic responses given. On the other hand caffein, increased oyxgen tension, and a gradual increase of temperature had the opposite effect. (Allee, '12, '13.) When the rate of metabolism of isopods was determined by their resistance to potassium cyanide (Child, '13, '13a; Allee, '14; also page 206.) isopods giving a high percentage of positive rheotactic responses in a circular current had the highest rate of metabolism. Those with a high per cent, of negative re- sponses were second, while those with a low positive response and with either the negative or indefinite reaction dominating had the lowest rate of metabolism. For a number of years the other of us has been working upon a method for determining with analytical accuracy the minute amounts of carbon dioxide given off in the metabolism of nerve fibers. (Tashiro, '13, '13a, '14.) The apparatus devised will detect and measure with accuracy 0.000,000,1 gram of carbon dioxide. For those who are not familiar with the new method for determining carbon dioxide we may say that the quantitative method depends upon determining the minimum amount of gas 202 RHEOTAXIS OF ISOPODS 203 which will give the first precipitate of barium carbonate when introduced into a chamber in which a perfectly clear drop of barium hydroxide is exposed. It was previously found by work with known amounts of very dilute carbon dioxide that the minimum amount of carbon dioxide which gives the first pre- cipitate is 0.000,000,1. gram. Thus, by determining the mini- mum number of cubic centimeters of a gas from a respiratory chamber of known volume, we can calculate very accurately the amount of carbon dioxide given off by the tissue or animal under observation. The details of the method are as follows: An animal is left in the respiratory chamber (15 cc. capacity) for a respiration period of ten minutes. Then this air is driven into a gas pipette. After cleaning and washing the apparatus one cubic centimeter of this gas is introduced into the barium hydroxide chamber but gives no precipitate; .25 cc. more is introduced with no result; .25 cc. more gives a precipitate. The total volume introduced, 1.5 cc. is the minimum volume that will give the first precipitate. Since 0.000,000,1 gm. of carbon dioxide is the minimum amount which gives the first precipitate, it is certain that 1.5 cc. of respired air must contain 0.000,000,1 gm. of carbon dioxide. 15 The total respiratory chamber must have contained ■ x 0.000,000,1 or 10 x 10 -7 grams carbon dioxide. Not all of the work to be reported in this paper was done in this detailed quantitative manner since in some cases only comparative results were needed. For the comparative work a piece of apparatus called a Biometer> (Tashiro, '13.) was used. This consists essentially of two chambers of equal size which are prepared for a determination in exactly the same manner and the animals to be tested for relative carbon dioxide pro- duction are inserted. At the start each chamber contains a perfectly clear drop of barium hydroxide. The chamber which first shows a precipitate of barium carbonate and which later shows more precipitate evidently has had carbon dioxide pro- duced at a higher rate than the other chamber. Hence it is easy to find which of two isopods is producing the greater amount of carbon dioxide. ' Quantitative determinations can also be made with the Biometer by using one chamber as a respiration chamber and the other for the determination. 204 W. C. ALLEE AND SHIRO TASHIRO THE PROBLEM With this more refined method of obtaining an insight into the relation between physiological states and animal behavior the following three lines of inquiry were prosecuted during the time at our disposal at Woods Hole during the summer of 1913: 1. What is the relation between the rate of carbon dioxide production of isopods and their resistance to relatively strong solutions of potassium cyanide? 2. What is the effect of the calcium ion upon carbon dioxide production and rheotaxis in isopods? 3. Is there a relationship between the variation in carbon dioxide production and the rheotactic reaction of isopods? THE STOCK Isopods from a series of collections from small fresh water ponds near Woods Hole, Massachusetts, were used in these experi- ments. They were all Asellus communis, Say, and were about half grown. The isopods came from silt and debris bottomed ponds and were kept under comparable conditions in the laboratory. METHODS The isopods were tested for their rheotactic reaction in a circular current the bottom of which was covered with wax. The responses of an individual isopod for ten successive minute reaction periods were taken as a fair indication of the rheo- tactic tendencies of the animal. The isopods were judged to give a positive reaction when they went against the current for over half of the minute's reaction period. They were considered to give a negative reaction when they moved with the current for over half of their minute's trial and indefinite when their movements gave no indication of being regulated by the direc- tion of the current. The approximate distance covered by each reaction was recorded and will be found in a standardized form in the tables under the head of efficiency in the current. In general the movements of highly positive isopods are more vig- orous than those of negative isopods, which in turn are more vigorous than those giving an indefinite reaction. (For further details see Allee, '13.) When the carbon dioxide output was to be tested the isopod was dipped into water free from carbon dioxide, dried momen- RHEOTAXIS OF ISOPODS 205 tarily on filter paper and placed on a cover glass with no water added except that which clung to the animal. An amber ring about three millimeters high was placed around the isopod and this was covered with wire gauze to prevent extended movement of the animal while in the respiration chamber.* This precau- tion was the more effective since isopods are strongly positively thigmotactic and will rest quietly for extended periods when they are able to place their bodies in an angle of their container. When comparisons were made in the Biometer, isopods of approximately the same size were selected in order to guard against a greater production of carbon dioxide due to greater mass. The isopods were left in the respiration chambers for only ten minutes during the quantitative work but in the quali- tative results, tested in the Biometer, the isopods were left as long as thirty minutes. The fact that isopods survived nine daily tests and showed no ill effects of the handling demonstrates there is little danger to the animal in such treatment. THE RELATION BETWEEN CARBON DIOXIDE PRODUCTION AND RESISTANCE TO POTASSIUM CYANIDE The work upon the relation between carbon dioxide output and the resistance of isopods to potassium cyanide was quali- tative only and was carried on to ascertain whether or not the resistance of isopods to the cyanide is a safe index of their met- abolic activity. Child ('13, '13a) found that in Planaria the susceptibility of animals or pieces of animal to 0.001 mol. solu- tion of potassium cyanide varied in general with the rate of metabolism. Estimation of carbon dioxide production on in- dividuals and pieces (of the same species) made at Dr. Child's request by Tashiro with the aid of his Biometer showed that carbon dioxide production ran parallel with susceptibility to potassium cyanide, and so afforded a confirmation of Child's conclusion concerning the relation between susceptibility and the rate of metabolism. 2 We appreciate the roughness of this method for checking spontaneous muscular movement, but it was out of consideration for us to devise an automatic recorder for bodily activity such as is necessary in order to make accurate metabolism experiments with mammals. The principal source of error in determining the amount of carbon dioxide given off by isopods lies in the fact the spontaneous muscular movements were not under complete control. With the device described above and by constant, careful observation with a hand lens of the animal during the experiment, we convinced ourselves that we had sufficient control of the move- ments of the animals to answer the needs of our experiments. 206 W. C. ALLEE AND SHIRO TASHIRO Allee has conducted a series of tests to find whether or not Child's cyanide resistance method would apply to isopods and has found decided evidence that the application is possible. (Allee, '14.) As a final check upon this work we repeated with isopods the tests of the relation between carbon dioxide produc- tion and cyanide resistance that had been made earlier for Planaria. The detailed method of procedure was as follows: Two easily distinguished isopods of equal size, whose rheotactic reac- tion had been previously tested by Allee, were placed in the Biometer by Tashiro and the relative speed of carbon dioxide output was determined. Immediately upon the removal of the TABLE 1 Showing the relation between the rheotactic reaction, carbon dioxide production and resistance to 0.001 mol. potassium cyanide. The carbon dioxide output was compared in the biometer and the total reactions of two animals thus compared are shown in each horizontal division of the table. Rheotactic reaction in percentage of total *-> a Carbon dioxide output compared me in ours tes i o o number of trials >> c c •r! >- Survival ti .001 mol. KCN. in h and minu c s: M c o + — 00 92 60 20 20 2.1 more 2:10 c? 6.5 89 70 30 2.0 less 3:10 d» 6.0 86 100 2.1 more 2:00 d 1 5.0 87 100 2.2 less 2:30 9 7.0 90 80 10 10 2.1 more 2:00 c? 6.5 91 80 10 10 2.0 less 2:30 9 5.5 95 100 2.6 more 5:30* c? 7.0 94 20 80 2.2 less 4:45 c? 5.0 30 70 20 10 more 2:20 o 71 4.5 169 40 2) 40 less 3:10 tf 5.0 84 1 • 40 50 less 2:55 c? 6.0 171 60 40 more 1:35 tf 5.5 * See discussion, page 207. RHEOTAXIS OF ISOPODS 207 isopods from the respiration chamber they were placed in an Erlenmeyer flask in 0.001 mol. solution of potassium cyanide and their survival time was ascertained by Allee. It should be noted that the experimenter on the determination of the carbon dioxide output was ignorant of the behavior of the animals, thus eliminating any prejudice for the determinations. The results of experiments of this character are shown in table 1. The table shows that in ten of the twelve isopods tried the evidence from the survival time ran parallel with that of the carbon dioxide production, that is, the isopods giving the more carbon dioxide had the shorter survival time in the cyanide. The table also shows that where there was a difference in the rheotactic reaction of the animal tested, the more carbon dioxide was given by the isopod that gave the higher percentage of positive rheotactic reactions. (Cf. page 213; also Allee, '14.) In the case of isopods No. 94 and 95 where the isopod that gave off less carbon dioxide lived a shorter time in the cyanide the experimental records show that No. 95 moved more in the respiration chamber than did No. 94 and also that it was two millimeters longer. Either of these factors might account for the discrepancy. If the potassium cyanide resistance in this case is taken as the true index of the metabolism it will be noted that the animals living longer, i.e., having the lower rate of metabolism gave the highest percentage of positive rheotactic reactions. This apparent contradiction will be discussed later (page 211). Since in S3% of the cases tried the carbon dioxide produc- tion tallied exactly with the resistance to potassium cyanide and in the other 17% of the cases the apparent exception is capable of reasonable explanation, it seems safe to conclude that so far as carbon dioxide production is concerned the resis- tance of isopods to potassium cyanide is a safe index of the metabolic activity of the animals. THE EFFECT OF THE CALCIUM ION UPON RHEOTAXIS AND CARBON DIOXIDE PRODUCTION In connection with experiments on irritability Tashiro has studied the effects of inorganic salts upon tissue metabolism. Whatever the mechanism of the effect of such salts on tissues 208 W. C. ALLEE AND SHIRO TASHIRO TABLE 2 Showing the effect of calcium chloride upon carbon dioxide production and rheo- taxis in isopods. The survival time in potassium cyanide is added for comparative purposes. The isopods were first tested for rheotaxis, then two of approximately the same size were taken for determination of their carbon dioxide output in the biometer. The one of these that gave the least carbon dioxide was taken as a con- trol, its rheotactic reaction was again tested and it was allowed to stand in water to which it was accustomed while the other was treated. The second isopod, the one giving the most carbon dioxide, was placed in a 0.16 mol. solution of calcium chloride until the positive rheotactic tendency was mark- edly decreased. Immediately afterward the carbon dioxide production of the two was again compared in the biometer. Isopod No. 30 Isopod No. 169 Rheotaxis test, 11:55 A. M. Temp. 20 Rheotaxis test, 12:25 P. M. Temp. 20 50%+, 50%—; Efficiency, 2.1 90%+, 10%—; Efficiency, 2.1 Tested in Biometer 1:47-2:00 P. M. Tested in Biometer 1:47-2:00 P. M. Temp. 23.5 Temp. 23.5 Less C0 2 than No. 169 More CO 2 than No. 30 Rheotaxis test 2:00 P. M. Put in 0.16 Mol. CaCh 2:05 P. M. 70%+, 20%— 10% oo; Efficiency, 2.25 Rheotaxis test 2:07 P. M. 80%+, 20%— Efficiency, 1.6 Rheotaxis test 2:27 P. M. 40%+, 20% oo, 40%o; Efficiency, .95 Taken from CaCl 2 3:43 P. M. In CaCh 36 minutes Tested in Biometer 3:44-3:57 P. M. Tested in Biometer 3:44-3:57 P. M. Mere C0 2 than No. 169 Less CO s than No. 30 Survival time in 0.001 Mol. KCN Survival time in 0.001 Mol. KCN 2 hours, 20 minutes 3 hours, 10 minutes d\ 4.5 mm. long cT , 5.0 mm. long Isopod No. 171 Isopod No. 84 Rheotaxis test 12:25 P. M. Temp. 20 Rheotaxis test 12:00 M. Temp. 20 10%+, 90%—; Efficiency, 2.4 30%+, 60%—, 10% oo; Efficiency, 2.6 Tested in Biometer 2:49-3:05 P. M. Tested in Biometer 2:49-3:05 P. M. Little CO 2 given off Little C0 2 given off. Less C0 2 than No. 84 More C0 2 than No. 171 Rheotaxis tested 3:50 P. M. Put in 0.16 Mol. CaCl, 4:02 P. M. 60%+, 40%—; Efficiency, 2.0 Rheotaxis tested 4:07 P. M. 10%+, 40% 00, 50%o; Efficiency, 0.9 Taken from CaCl. 4:27 P. M. In CaCh 25 minutes Tested in Biometer 4:35-5:12 P. M. Tested in Biometer 4:35-5:12 P. M. More CO, than No. 84 Less C0 2 than No. 171 Survival time in 0.001 Mol. KCN Survival time in 0.001 Mol. KCN 1 hour, 35 minutes 2 hours, 55 minutes <5\ 5.5 mm. long cT , 6.0 mm. long RHEOTAXIS OF ISOPODS 209 may be, it was clearly proven (Tashiro, '13), that inorganic salts which affect physiologic states of the nerve equally modify metabolism as measured by the carbon dioxide production. With Dr. Lingle* he has further extended the study of the effects of calcium and sodium ions upon tissue metabolism upon iso- lated pieces of heart tissue of turtles. Allee* has spent considerable time upon the effects of certain inorganic salts upon the rheotactic reactions of isopods and found among other results that calcium chloride caused animals that were highly positive to a water current to become much less positive. That the calcium rather than the chlorine ion is responsible for these results is shown by tests with a number of other inorganic chlorides some of which increase while others decrease the positiveness of the rheotactic reaction of isopods. From his work on tissue metabolism Tashiro suggested that the calcium chloride in some way caused a decrease in the met- abolic activity of the isopods. In order to test this the crucial experiments were made the results of which are exhibited in table 2. Although only two comparisons were made yet the results are so diagrammatic and are so fully in accord with the previous experience of both authors that they may fairly be taken as giv- ing a truthful picture of the conditions under consideration. In brief the experiments* were as follows: Two isopods of approximately the same size were tested for their relative rate of carbon dioxide production in the Biometer. The isopod hav- ing the lower rate of carbon dioxide output was taken as a con- trol and was again tested for the rheotactic reaction and then left in conditions to which it was acclimated while the other was treated. The second individual, which had the higher rate of carbon dioxide production was placed in a 0.16 mol. solution of calcium chloride until the tendency to give a positive rheo- tactic reaction was markedly reduced. Then the rate of carbon dioxide production of the two was again tested in the Biometer. In both pairs tested the isopod with the higher rate of carbon dioxide production at the first test in the Biometer had also given the higher percentage of rheotactic responses, but after being treated with calcium chloride for 25-36 minutes it came to be less positive in its rheotactic reaction, and also gave less 3 Unpublished results. 210 W. C. ALLEE AND SHIRO TASHIRO carbon dioxide and was less susceptible to potassium cyanide than the control individual. In other words the calcium chlo- ride (0.16 mol.) decidedly decreased the rate of metabolism of the isopods and also reduced their tendency to give a positive rheotactic reaction. TABLE 3 Showing the quantitative daily determination of the carbon dioxide output of isopods Nos. 102 and 12. Isopod No. 12 was tested quantitatively for carbon dioxide production in the biometer; isopod No. 102, in a new apparatus especially devised for quantitative work. Column 3 gives the capacity of the respiratory chamber; column 4 shows the number of cc. of respired air which first gave the precipitate of BaCO; column 5, the amount of CO 2 given by the isopod in ten minutes, for method of calculation see page 203. Date Temperature in degrees C. Volume of respiratory chamber Minimum cc. giving Ppt. of BaCO Amount of CO given by isopods during 10 minutes in gms. 1 2 3 4 5 Iso- pod No. Aug. 14 1- 23 102 22.5 12 15 102 25 12 1.5 102 .75 12 10 x 10~ 7 102 33 x 10~ 7 15 23 23 15 25 1.6 .4 9.3 xlO -7 62.5 xlO -7 16 23.5 23.5 15 25 .55 .5 27.2 xlO -7 50 x 10~ 7 17 24.5 24 15 25 1.8, .4 12.2 xlO -7 62.5 xlO -7 18 24 24 15 25 1.4 2.0 10.7xl0 _/ 12.5 xlO -7 * 19 Nodet 2rminatk >n.f 20 20.5 20 15 25 7.1 1.6 12.1xl0 -7 15.6 xlO -7 21 22.5 20.5 15 25 1.4 .7 10.7 xlO -7 35.7 xlO -7 22 22.5 22.5 15 25 .5 1.1 30 x 10 -7 22.7 xlO -7 * Determination in some doubt due to lack of facilities for running a duplicate determination. t No determination because of lack of carbon dioxide free air which is required n preparing the apparatus for a determination. THE RELATION BETWEEN DAILY VARIATION IN CARBON DIOXIDE PRODUCTION AND THE RHEOTACTIC REACTION It has been observed that the rheotactic reaction of isopods varied to a considerable degree even when the animals were kept under approximately identical external conditions. (Allee, RHEOTAXIS OF ISOPODS 211 '13.) In order to analyze this behavior it is necessary to follow the variations in the daily metabolism which obviously cannot be done by the cyanide method because that depends on the death point of the animals. Daily determination of the carbon dioxide output of the isopods in connection with a daily test of the rheotactic reaction proved feasible. The carbon dioxide determinations were made by Tashiro in the manner already given (page 203). The rheotactic reactions were tested by Allee and neither knew the results of the other until the end of the tests. The results obtained are listed in tables 3 and 4. From the results exhibited in table 4 it is seen that with both isopods No. 102 and No. 12 six of the seven changes of carbon dioxide production and the rheotactic response run in a parallel direction. This means that with these two isopods 86% of the variations in carbon dioxide output and rheotactic reaction were similar. The amount of variation is not always proportionate but it should be remembered that the isopods were able to move to a limited degree in the respiration chamber and that this caused an increase in the carbon dioxide production that was not controlled. Also there is a possible error of about 5% in the method of ascertaining the sign of the rheotactic reaction. (Allee, '12.) In view of these considerations the experimental results are about all that could be expected and are certainly more exact than any previous observation on the correlation of the behavior of animals upon their metabolic rate or phys- iological state. Incidentally the table shows an agreement in the direction of variation of carbon dioxide production and the oxygen ten- sion in the water in which the isopods were kept in 66% of the cases. The variations in the rheotactic response and oxygen tension agree in 73% of the cases. This seems to be good evidence that all three of these factors are more or less closely related. The evidence here presented also makes it apparent that each individual isopod has in all probability a different rate of metabolic activity from that of any other isopod, (cf. Allee, '14) and farther that it is not a fixed standard of metabolism that accompanies a high degree of positiveness in the rheotactic response but rather a relative rate. Thus on the average iso- pod No. 102 (length 8 mm.) gave off over twice the amount 212 \V. C. ALLEE AND SHIRO TASHIRO TABLE 4 Showing the relation between daily variations in carbon dioxide production and rheotactic reactions of isopods Nos. 102 and 12. The amount of carbon dioxide given in ten minutes may be obtained in grams by multiplying the numbers given in the second column by 10- 7 . In the third column + indicates the results are what would be expected from our present knowledge of the rheotactic reaction if there is a direct relation between the rate of positiveness in the water cur- rent and the rate of carbon dioxide production; — indicates the opposite state of affairs. Two oxygen tensions are given; the first being that from which the isopods were taken before their rheotactic test; the second represents that in which they were placed after the carbon dioxide determination. Between testings the isopods were kept in one liter Erlenmeyer flasks which were full of water and tightly corked. Isopod No. 102 O V l +-1 •-I > .a,a Rheotactic reaction in 3 U U •2 CO > .£ <_> were now possible: (1) The rats may be unable to hear c' 512 either because of a shortening of the scale similar to the probable shortening of the rat color spectrum, or because of an inability to hear tone at all, i. e., complete tonal deafness, or because of a tonal island. In either of the three cases, we have a sensory defect. Furthermore, it would be a defect common to all of the rats tested. (2) The ignoring of c' 512 may be due to a factor of attention as Dr. Weidensall has pointed out in her report at the American Psychological Association. 1912.* Our dilemma here is the same as the one which confronted the Watsons, » e.g., when their rodents were shown to be ignoring red in the red-green discrimination. The solution in either case must come from evidence drawn from tests made on mere sensitivity, i.e., the discrimination of a stimulus from its absence. In view of this the following tests were made. Three of the rats of set one, were each given 10 trials daily, in an attempt to set up an association between the tone and turning to the right and between absence of tone and turning to the left. 410, 520 and 350 trials were given, but the association Was never set up. These three rats had formed the original discrimination of noise vs. tone in 310, 370 and 520 trials respec- tively. In all save the third rat, therefore, the original discrimi- nation was set up in fewer trials than given in this control. I have not regarded this as entirely conclusive, however, because as a result of previous training the rats were all ignoring the tone. This habit (if habit it were) may have persisted. A new set » There is a third possibility, viz. : the rats may have been unable to discriminate the tone from its auditory background. It was impossible to carry out the tests in a sound-proof room, so this possibilitiy has not been rigorously excluded. However, the rats were accustomed to what auditory stimuli did occur, and as far as the ex- perimenter was concerned, the tone dominated over all other sounds save in excep- tional cases that rarely occurred. * Weidensall, Jean. A Critique of the Discrimination Test. Psych. Bull., 1912, 9, pp. 57-58. « Watson, J. B. and Watson, Mary I. A Study of the Responses of Rodents to Monochromatic Light. Jour. Animal Behavior, 1913, 3, pp. 1-14. 218 WALTER S. HUNTER (II) of six untrained rats was now chosen. To this group was added one rat from set I, not included in the above three. These seven rats were given ten trials daily. Table I gives the results. Not only did these rats not learn to discriminate the tone from its absence, but the data indicate that they reacted as poorly at the close of the 700 tests as at the beginning. The objection may be made that the number of trials was insufficient ! Such a criticism I should regard as valid only if the animals were slowly improving in accuracy. This was not the case. Particular attention should be drawn to the case of Rat No. 5. This rat had learned to discriminate " noise from tone " within 400 trials. The above table shows that when tested on mere sensitivity to the tone, there was no improvement in accuracy even at the end of 700 trials. TABLE I. The learning processes of the rats of set 11. The numbers stand for the trials in each successive fifty that were correct. Rats Trials 5 13 14 15 16 17 18 50 . . 24 26 22 19 22 22 23 100 . . 27 17 25 22 27 27 28 150 . . 28 24 21 17 24 24 27 200 . . 34 26 24 24 24 28 26 250 . . 30 24 22 25 23 30 25 300 . . 25 24 29 26 27 25 27 350 . . 28 23 24 23 28 26 27 400 . . 31 23 27 21 18 22 28 450 . . 25 26 25 22 25 24 28 500 . . 28 23 25 27 21 25 19 550 . . 24 19 28 24 23 22 22 600 . . 28 23 22 25 27 25 23 650 . . 30 24 25 23 24 28 24 700 26 21 26 25 26 24 24 The present work is of interest when compared with that of Johnson on pitch discrimination in dogs. At the present writing his complete results have not appeared; but the preliminary report « shows that the dogs could not discriminate between middle C and the E above. The possibility that the dogs can- not hear these tones is not considered, although the data are in harmony with such a view. 6 Johnson, H. M., " Some Experiments on Pitch-Discrimination in Dogs." Psych. Bull., 1912, 9, p. 59. THE AUDITORY SENSITIVITY OF THE WHITE RAT 219 Comparative psychologists are agreed, I believe, that, with the higher animals, the ability to associate a stimulus with a simple response shall be the criterion of sensitivity to that stimulus. Exceptional cases may occur, as is suggested by the studies of Yerkes upon the hearing of the frog. ■> Other factors being equal, however, where an animal can learn an association with one auditory or visual stimulus, inability to do so with another auditory or visual stimulus is to be taken as evidence of lack of sensitivity. Granting this, the above data prove either that the rats cannot hear c' 512, under the conditions of the present experiment, or (note 3 has some reference here) that their sensitivity is extremely slight. The analogies between the present conclusion and that reached by other students with respect to color vision in animals are both striking and instruc- tive. It is to be noted in particular that tone and color cor- respond to periodic vibrations and that noise and the white- black series correspond in general to heterogeneous vibrations. The ability to react to periodic ether vibrations is apparently a late acquisition in animals. Why then not expect, a priori, the same to be true in sound, particularly when periodic vibra- tions seem to be more artificial and hence rarer (at least in the habitats of non-musical animals) than heterogeneous ones? Further comment upon this must await a later presentation of data. In addition to the above crucial evidence on the sensitivity of the white rat to c' 512, much other material bearing upon the same problem has been accumulated. All of this, while not in and of itself decisive, is in harmony with the conclusion above drawn. (1) All rats of set I ignored the tone and reacted on the basis of noise and the absence of noise. There must be, then, some fundamental difference in the effect of noise and tone on the rat. Otherwise we should expect individual differences to appear. (2) Six rats of an untrained set (III) all failed to discrim- inate a very intense sounding of c' from a faint sounding of the same tone. These two intensities may be described with sufficient accuracy as follows: One was as intense as could be secured by striking the fork. The other was approximately of i Yerkes, R. M. " The Sense of Hearing in Frogs." Jour. Comp. Neur. and Psych., 1905, 15, pp. 279-304. 220 WALTER S. HUNTER the same intensity as a tone produced by a solid rubber ball of 102 gm. striking the fork after a free fall of 25 cm. The conditions of testing here were the same as described above. Five trials daily were given, save at certain periods with two rats, with punishment and reward. Between 575 and 800 trials were given. There was no more evidence of discrim- ination at the last than at the first. This problem of intensity discrimination was begun simultaneously with the work on noise and tone before I suspicioned that the rats were insensitive (or very slightly so) to the tone in question. (3) Miss Alda Barber of this laboratory is studying local- ization of sounds in rats. The standard stimulus, tapping upon wood, is well localized. The following controls have been used with significant results: (a) Tapping with the rubber end of a lead pencil on the resonance box of c' is localized with normal accuracy. This gives a noise predominately of a 512 v. s. pitch. (b) A tuning fork c' sounded steadily is completely ignored. (c) The same pitch blown upon an organ pipe as an interrupted (tooting) tone is also ignored. It is not known yet whether or not special training will overcome these last two failures. (4) Watson » obtained reactions from white rats with the Galton whistle. I have never secured an unambiguous response to tone, although violent starts are often made to the slightest noises. A few times I have thought that reactions occurred. These may well have been to the noise accompanying the whistle tone. Tests have been made with organ pipes and Edelmann- Galton whistles. At least 30 rats have been tested in this labo- ratory both when they were awake and when they were asleep, when they were nervous and responded to slight noises readily and when they were not nervous. A few tests were made upon some twenty rats at the University of Chicago in the summer of 1913. The Edelmann whistle was used throughout its range, but no reactions were observed. Professor R. E. Carter of the University of Kansas witnessed these latter tests and agreed in the findings. Possibly it is true that rats are sensitive to each others squeaks, but who is to say whether these are more tone than noise? This type of test is to be carried further. At present this •Watson, Jno. B. " Kinaesthetic and Organic Sensations." Psych. Rev. Mon., 1907, 8, pp. 53-54. THE AUDITORY SENSITIVITY OF THE WHITE RAT 221 method of " general response " has convinced me more strongly merely that there is a fundamental difference for the rat in noise and tone. (5) It will be recalled that only three rats of set I were tested immediately upon sensitivity to tone vs. no-tone. The other three rats were tested as follows: In place of hand claps, the following noises were each substituted for five trials from time to time: (a) rattling of paper; (b) dropping sunflower seed on tin; (c) scratching on wood; (d) drumming on the table with the fingers ; (e) rubbing two pieces of board together ; (f) hissing through the teeth; and (g) rattling of nails in a glass. Pitch, volume and quality varied greatly, but a rough attempt was made to keep the intensity values equal to that of the tone. (See above, page 216.) The rats responded to all of these stimuli as accurately as to the regular stimulus of hand claps, i.e., never below 80% correct. Rat No. 5 failed to react cor- rectly to noises e and f, i.e., although repeatedly tested, he never made more than from 55% to (at most) 70% correct. This was the rat tested later with set II, on tone vs. no-tone, with negative results. In view of those tests, his failure to re- spond correctly to the two noises is to be explained on the basis of their dissimilarity to the standard noise rather than upon their likeness to c' 512, which tone this rat seems not to hear. Tests were also made in which each of the following tones were substituted for the original c' on the fork on enough occa- sions to be sure that the reactions were not due to chance: (1) c'" 2048 on the fork; (2) c' 512 v. s. on the organ pipe, sounded steadily; (3) No. 2, sounded interruptedly, i.e., in toots; (4) c" 1024 on the organ pipe, sounded steadily; (5) No. 4 sounded interruptedly; and (6) f 341.3 on the organ pipe, sounded steadily. With no exception, the rats reacted to these tonal stimuli as to the original tone which had been sounded steadily, i.e., they ignored them. There are many suggestions as to in- terpretation which arise from these results. The points that can be definitely stated are these: (1) All of the tones given were for some reason very different from the noises. (2) This difference was not the fact of smoothness, i.e., lack of inter- ruptedness. This point seems conclusively proved, because on the same day with the interrupted noises were given trials with the interrupted tones, yet the rats paid no more attention to 222 WALTER S. HUNTER the tones than if they had been continuously sounded. Inas- much as the animals reacted in the same manner to all of the noises, it is certainly a striking fact that none of the tonal stimuli given were classed as noises. Further experimentation alone will determine whether sensory defect is the reason for neglecting all of the tones given here. Such an explanation certainly seems necessary for the lack of sensitivity to c' 512. ^ At the present, the above work is being extended in three directions: (1) Search is being made throughout the pitch scale for a tone to which the rats will respond. Both continuous and interrupted tones will be used. When an effective tone is found, the original problem will again confront us, viz., can rats hear noise and tone as distinct experiences. (2) Miss Bar- ber's work will probably throw added light upon the question of relative sensitivity to noise and tone. (3) Mr. A. C. Scott is beginning tests which are expected to emphasize the relations between vision and hearing with respect to the learning pro- cesses involved. One problem with which he will deal will be this: Is the simultaneous presentation of stimuli, such as is used in visual discrimination, more favorable to learning than a successive presentation of stimuli, such as must be used in auditory discrimination work? One additional matter needs comment. So far as I have been able to ascertain there are no published studies on the anatomy of the white rat's ear. I am supported in this state- ment by several eminent authorities. In view of the results above presented, it is at least possible that careful anatomical studies might throw light upon the structural basis for the perception of noise and tone. » Johnson's work >» has appeared since this paper went to press. On pp. 44^5, the author reports negative results at the conclusion of 150 trials on mere sensitivity to a tuning fork of 256 d. v. The suggestion is made from this that in ordinary noises the dog may reach only to high overtones. '» Johnson, H. M. Audition and Habit Formation in the Dog. Behavior Mono- graphs, 2, no. 3, 1912. BEHAVIOR OF THE MEDITERRANEAN FRUIT FLY (CERATITIS CAPITATA WIED.) TOWARDS KEROSENE HENRY H. P. SEVERIN, Ph.D. and HARRY C. SEVERIN, M.A. The following observations on the behavior of the Mediter- ranean fruit fly toward kerosene were made in the field in Manoa Valley, situated on the outskirts of the city of Honolulu, Hawaii. On account of the abundance of rainfall in this valley, the kero- sene traps used in our experiments consisted of pans three and one-half inches in depth and twelve inches in diameter. Each pan was wired to the lower branches of a fruit tree (Fig. 1). -■fib* v^'" ■'•^^f- ?^?vf- e. m** '■- *•<>, Figure 1. Pan containing kerosene wired to the lower branches of a fruit tree, to trap the Mediterranean fruit fly. Enough kerosene was poured into each pan to cover the' bottom so that in case of a heavy rain, the kerosene might overflow directly to the ground and not injure the tree. After a light or moderate rain such traps are probably just as effective as when there is no precipitation, for the oil floats on the surface of the water. The first problem which presented itself was to determine whether the color of the pan containing the kerosene made any 223 224 HENRY H. P. SEVERIN AND HARRY C. SEVERIN difference in the number of Mediterranean fruit flies caught. Five white, three black, one blue and seven orange-colored pans were wired to the branches of orange, lemon, grapefruit and guava trees. From the results of our catches in the various pans, it was evident that the number of fruit flies captured was not influenced by the color of the pans. Moreover, it is highly probable that the sense of smell is the determining factor in attracting these insects to the kerosene. In our second experiment we endeavored to ascertain in what particular kind of fruit-bearing tree of an orchard the pest would be captured in largest numbers with the kerosene traps. Accordingly, one pan was wired to the lower branches of a common guave tree (Psidium guayava pomiferum), nine pans were fastened in nine different navel orange trees (Citrus auran- tium), and one pan was placed in a Java plum tree (Syzygium jambolana). All of the pans used in this experiment were enam- eled .white, because most insects caught in the oil were more conspicuous against such a background. The following table shows the number of fruit flies taken at intervals of three to four days for a period of eighteen days in the kerosene traps attached to the three different kinds of fruit trees: TABLE 1 Number of Mediterranean Fruit Flies Captured in Kerosene Traps Placed in Guava, Orange and Plum Trees. Trees One guava Nine navel orange One Java plum Four days catch 75 1155 398 Four days catch 33 749 207 Three davs catch 25 715 213 Three days catch 16 422 60 Four days catch 35 1093 295 Eighteen days catch 184 4134 1173 Average capture per day in 1 trap 10 25 65 Total number of males captured 5461 Total number of females captured 30 5491 BEHAVIOR OF THE FRUIT FLY 225 It is evident from this table that the attraction of the Mediter- ranean fruit fly to the kerosene was confined almost entirely to the male sex. Female flies were present in this orchard because hundreds were caught by sweeping with an insect net among the fruit trees. Trapping the pest with kerosene was carried on for a period of eight months in the Hawaiian Islands in con- nection with other experiments and the results show that of every one thousand fruit flies captured only three on an aver- age were females, the remainder being males. A dissection of some of the flies captured in the kerosene trap wired to the Java plum tree showed that the alimentary canal was filled with the blue juices of the plum. The ripe plums were seriously infested by the maggots of different species of Drosophilidae and the juices were exuding from the punctured and bruised fruit. Mediterranean fruit flies and some of these plums were placed in a breeding jar and frequently the Trype- tids were seen feeding on the plum juices. The reason why more specimens were captured in the kerosene trap wired to the Java plum tree finds its probable explanation in the fact that the plum is more attractive to the pest then the common guava or navel orange, or possibly because the plum juice was more available than the juices of the less bruised guavas and oranges. The Mediterranean fruit fly was often captured in kerosene traps wired to trees that were not bearing fruit and also near fruits in which the pest has not been reported to breed. Kero- sene traps were fastened repeatedly in mango trees (Mangifera indica) months before the mango season was on, and in every instance fruit flies were trapped. One trap was wired in an isolated bread fruit tree (Artocarpus incisa) which at that time bore very hard, green fruit, and in four days twenty-nine male Mediterranean fruit flies were caught. A kerosene trap was also placed in a clump of mulberry shrubs (Morus nigra) bearing ripe fruit, and in four days twelve male flies were taken. In the last catch the flies may have been attracted by the ripe fruit, the juices probably serving as food material for the adults. The interesting part of the last two catches rests in the fact that the pest under consideration does not breed in the fruits of these trees. In all probability the reaction of the male Mediterranean 226 HENRY H. P. SEVERIN AND HARRY C. SEVERIN fruit fly to kerosene is not in any way connected with the feed- ing habits, but too much emphasis, however, should not be attributed to any of these experiments, for the distance that kerosene will attract the flies is not known. Weinland (2, page 847) claims that the sphere of influence of kerosene ' is limited, being possibly fifty feet or so, varying with the wind, freshness of kerosene, etc." Howlett (1, page 414) of India says the distance at which the fruit flies (Dacus zonatus Saund.) are able to perceive the smell of citronella oil "is doubtful, but seems to be considerable; half a mile is probably not extravagant an estimate if the wind is favorable." The behavior of the Mediterranean fruit flies was occasion- ally observed in the neighborhood of the kerosene traps. In some instances fruit flies remained at rest on the inside of the pans for long periods of time as if stupefied by the volatile parts of the oil. In other cases, the flies would walk along the inside of the pan for a time, then take wing and fly up to a neigh- boring leaf or twig, or in their apparently dizzy, zigzag flight over the surface of the oil they would plunge into the kerosene and generally cease all activity noticeable to the naked eye in less than half a minute. It certainly is peculiar that the Mediterranean fruit fly plunges into the kerosene to its own destruction. The flies may be at- tracted to the oil as a result of a chemotaxis due to one or more hydro-carbons or to the impurities of the petroleum oils, such as the sulphur constituents or nitrogenous products. Small quantities of sulphides are detected by the human nose and it may be possible that the minutest traces are perceived by the fruit flies. Furthermore, sulphides have recently been found within the bodies of insects. Again, the hydro-carbons of the oil may act as an anesthetic, and stupefy the insects whenever they remain within its influence. It is known that the volatile parts of gasoline, for instance, have a stupefying effect upon animals. According to a scientist connected with the Standard Oil Company, cases are on record where men, who had opened barrels of gasoline, were suddenly overcome by the fumes and plunged "head-first" into the oil. Large gasoline tanks which have been recently emptied are dangerous for men to go into, and require about twenty-four hours of ventilation before they are safe for a human being to enter. BEHAVIOR OF THE FRUIT FLY 227 Why should enormous numbers of male fruit flies and only a few females be captured in certain oils? Concerning the be- havior of Dacus zonatus towards citronella oil, Howlett (1, page 413) writes: ' Since the -reaction was confined to the male sex and did not appear to be in any way connected with feeding habits, it seemed most reasonable to suppose that the smell might resemble some sexual odour of the female which in natural conditions served to guide the male to her." This is, in substance, a view which we also expressed to a number of entomologists and mentioned in a paper read before the Agri- cultural Seminar in Honolulu on January 11, 1912, to explain the behavior of the male Mediterranean fruit fly towards kero- sene. Howlett believes that "the smell is in all probability perceived by means of the antennae," for, after he had carefully amputated these "at the base of the second joint," none of the mutilated insects were attracted to the oil of citronella. If it is true that kerosene gives off an odor which resembles that emitted by the female fruit flies to attract the opposite sex, then how would the fact be explained that a few females are usually caught in the oil? We would have to assume that the specialized sense organs present in the males to locate the females are absent in the latter. We would then be forced to conclude that the females were not attracted to the kerosene, but came within the sphere of influence of the oil by accident, became stupified and dropped into the oil. There is, of course, the possibility that the reaction of the male Mediterranean fruit fly towards some volatile part of the petroleum oils may be a positive chemotaxis "not representing the sexual smell of the female," a possibility to which Howlett also calls atten- tion in the behavior of Dacus zonatus toward citronella oil. BIBLIOGRAPHY 1. Howlett, F. H. The Effect of Oil of Citronella on Two Species of Dacus. 1912. Trans. Ent. Soc. London, pt. II, pp. 412-8. 2. Weinland, H. A. The Present Fruit Fly Situation and some Results of the 1912. Hawaiian Campaign. Cat. State Comm. Hort., Mon. Bull. I, No. 11, pp. t45-852. JOURNAL OF ANIMAL BEHAVIOR Vol. 4 JULY-AUGUST, 1914 No. 4 THE FACTORS DETERMINING THE VERTICAL MOVEMENTS OF DAPHNIA LEE RAYMOND DICE From the Zoological Laboratory of the University of California CONTENTS PAGE Introduction. Part I. Experimental 231 Phototaxis in Relation to: Light Intensity 231 Temperature 235 Chemical Content of Water 236 Mechanical Stimulation 236 Time of Day 237 Geotaxis in Relation to: Light Intensity 237 Temperature 242 Chemical Content of Water 244 Mechanical Stimulation 245 Time of Day 245 Locomotor Activity in Relation to: Light Intensity 245 Temperature 246 Chemical Content of Water 248 Mechanical Stimulation 248 Time of Day 249 Thermotaxis 249 Chemotaxis 249 Pressure - 249 Part II. Discussion 250 Observed Vertical Movements in Daphnia 250 Vertical Movements Caused by: Changes in Light Intensity 253 Changes in Temperature 255 Mechanical Stimulation 257 Changes in Chemical Content of the Water 258 Aging of Individuals ' 258 General Features of Behavior in Daphnia 259 Reversal of Geotaxis in other Plankton Animals 262 Summary 263 Literature 264 230 LEE RAYMOND DICE INTRODUCTION Vertical movements of the plankton Crustacea have been noted by many observers. The published results show that there is great variation in the movements made by different species, and that the same species in different lakes shows great differences in its vertical migrations. These facts indicate that there is a considerable number of factors which operate together to determine the movements. This paper is presented with the hope that it will help to make clear the relative importance of the factors determining the movements of a common fresh water entomostracean . This study has been carried on under the direction of Dr. S. J. Holmes to whom the author is indebted for constant advice and criticism. To Professor E. P. Lewis of the Department of Physics thanks are due for information concerning the absorption of ultra-violet rays in passing through the atmosphere and through water. Dr. A. C. Chandler assisted in taking observa- tions in some of the longer experiments. In the following experiments only one species of crustacean was used, Daphnia pulex De Geer, which was found abundantly at times in the lakes and ponds about Berkeley. The specimens used were obtained from several artificial ponds on the Uni- versity Campus. Many were kept for some time in glass aquaria in the laboratory before being used for experimentation in order to accustom them to laboratory conditions; others were used immediately after being collected from the pond in order to ascertain how far the behavior had been modified by the arti- ficial conditions. It has been observed that daphnids in different stages of their life history show differences in their vertical movements. To avoid this complication only adult individuals were used. In all these experiments the animals were kept in water taken from the same aquarium as the animals unless a statement is made to the contrary. In the experiments on geotaxis the jars were filled completely and the top sealed by a ground glass stopper to exclude the possibility of a difference of oxygen content between the top and the bottom. To avoid the possibility of chance in the animals remaining constantly at one end or the other of the dish, the ends were occasionally reversed or the contents of the dish stirred up. MOVEMENTS OF DAPHNIA 231 The experiments were performed between August, 1913 and April, 1914 in the Zoological Laboratories of the University of California at Berkeley. PART I. EXPERIMENTAL PHOTOTAXIS IN RELATION TO LIGHT INTENSITY The individuals of Daphnia pulex are normally positively phototactic and under conditions similar to those they would meet in nature it is difficult to obtain from them a negative response. In weak light or moderately strong light most of them are strongly positive. To the light from a 100 watt Mazda lamp at a distance of 30 centimeters they remained strongly positive throughout an exposure of nearly six hours. (Exp. 1.) Exp. 1 Phototaxis in weak light January 29, 8.30 A.M., 18 Daphnia pulex from diffuse daylight placed in a glass dish 28 cm. long marked off into five transverse divisions. Exposed in darkroom to horizontal light from 100 watt Mazda lamp at 30 cm. Temperature 17.5° C. Positive Negative end end Divisions I II III IV V 9.00 A.M. *14 1 1 2 9.15 15 1 1 1 9.45 15 1 2 10.50 15 1 2 11.50 15 2 1 1.00 P.M. 14 1 3 2.00 13 3 1 1 2.20 13 2 1 2 Averages 14.2 1.4 0.4 1.2 0.8 Another experiment (Exp. 2) shows that to the weak light from a 50 watt Edison lamp at 50 centimeters the animals remain positive during a continuous exposure of 60 hours. Probably they would remain positive indefinitely to these in- tensities. It appears that they are more strongly positive in the stronger light. * The figures refer to the number of individuals in each division. 232 LEE RAYMOND DICE Exp. 2 Phototaxis in weak light March 23, 4 P.M., 25 Daphnia pulex placed in a 28 cm. long glass dish marked off into five divisions. Exposed in darkroom to horizontal light from a 50 watt Edison lamp at 50 cm. distance. Temperature 17.5° C. Positive Negative end end Divisions I 11 III IV V Mar. 24, 4 P.M. 20 1 1 3 8 17 2 2 2 2 12 14 1 5 2 3 Mar. 25, 4 A.M. 10 2 3 5 5 8 13 1 5 4 2 12 M. 14 1 3 2 5 4 P.M. 10 3 5 4 3 8 10 2 6 3 4 Mar. 26, 8 A.M. 13 5 1 1 5 Averages 13.4 1.9 3.4 2.7 3.5 These daphnids, at temperatures of 20° to 23° C, are gen- erally nearly neutral to diffuse daylight (Exp. 3), to the light from a 15 ampere electric arc (Exp. 4), or to sunlight lacking the ultra-violet rays (Exp. 5). Experiment 3 shows a rather large variation in the distribution of the animals during the experiment which is probably due to the changes in light in- tensity corresponding to changes in the position of the sun. Exp. 3 Phototaxis in diffuse daylight March 19, 8 A.M., 22 Daphnia pulex from laboratory aquarium placed in a 28 cm. long glass dish marked off into three divisions. Exposed to diffuse light from a south window covered by a white screen. Light rays from the room were ex- cluded by a black lined box. Temperature 23° C. Positive end Negative end Divisions I II III 8.20 A.M. 6 5 11 8.45 6 6 10 9.30 9 3 10 10.00 8 3 11 11.00 8 7 7 12.00 M. 9 8 5 1.00 P.M. 9 6 7 2.00 9 7 6 3.00 9 8 5 4.00 11 7 4 5.00 10 9 3 6.00 13 6 3 Averages 8.9 6.3 6.8 MOVEMENTS OF DAPHNIA 233 Exp. 4 Phototaxis in arc light January 29, 18 Daphnia pulex which had been exposed to the light of a 100 watt Mazda lamp for four hours were placed in a glass dish 28 cm long marked off into five divisions. Exposed them to the horizontal light from a 15 ampere electric arc at a distance of 40 cm. The light was passed through 8 cm. of water to cut out the heat. Positive Negative end end Divisions I II III IV V 2.30 P.M. 7 1 1 1 8 2.40 7 2 2 2 5 2.50 6 1 3 8 3.00 8 1 3 6 3.10 7 1 3 2 5 3.20 5 2 3 2 6 3.30 4 6 2 6 4.00 9 3 2 1 3 4.30 2 1 3 4 8 5.25 8 1 2 2 5 Averages 6.3 1.2 2.6 1.9 6.0 Exp. 5 Phototaxis in sunlight April 13, 2 P.M., 13 Daphnia pulex from laboratory aquarium placed in 18 cm- long glass dish marked off into five divisions. Exposed to sunlight with ultra, violet rays cut off by one sheet of window glass. Temperature kept at 21° to 22° C. Positive Negative end end Divisions I II III IV V 2.05 P.M. 4 1 2 4 2 2.15 2 1 4 2 4 2.20 1 2 2 2 6 2.25 3 3 2 1 4 2.30 1 2 3 3 3 2.50 2 4 2 1 4 3.30 5 1 2 2 4 3.40 4 3 1 5 3.50 7 2 1 3 4.00 8 • 1 2 1 1 Averages 3.7 2.0 2.0 1.7 3.6 In these experiments it is shown that at temperatures of from 17° to 22° C. these daphnids are normally positive to weak electric light or to moderately strong light. They are more strongly positive to moderately strong light than to very weak light. The positive phototaxis to weak light persists for seem- ingly indefinite periods. To light intensities stronger than weak diffuse daylight these animals are indifferent and do not become 234 LEE RAYMOND DICE negatively phototactic to any intensity of ordinary light at these temperatures. It has been shown by Moore (1912) that Daphnia pulex may be made negatively phototactic by the ultra-violet rays from a mercury vapor arc lamp. The ultra-violet rays in sunlight are nearly all absorbed in passing through the atmosphere. How- ever, experiment 6 indicates that the quantity of these rays reaching the surface of the earth at sea level (altitude of Berkeley about 100 feet) is sufficient to have a slight effect in causing negative phototaxis. In passing down through the water of a lake the ultra-violet rays will be very rapidly absorbed. In lakes at high altitudes, which receive more of these rays than lakes at sea level, and in lakes whose water is very clear the effects the negative phototaxis caused by these rays in the bright part of the day might operate to keep the daphnids down a short distance from the surface. Exp. 6 Effect of ultra-violet rays on phototaxis April 13, 2 P.M., 13 Daphnia pulex from laboratory aquarium placed in an 18 cm. long glass dish marked off into five divisions. Temperature 21° to 22° C. Exposed to sunlight with the ultra-violet rays cut off by one thickness of window glass. Positive end Negative end Divisions I II III IV V 2.05 P.M. 4 1 2 4 2 2.15 2 1 4 2 4 2.25 3 3 2 1 4 2.50 2 4 2 1 4 2.51 Removed window glass 2 1 10 2.52 6 7 2.53 2 2 3 6 2.55 2 2 2 7 3.10 2 4 7 3.20 3 2 4 4 3.25 Replaced window glass 4 2 2 2 3 3.30 5 1 1 2 4 Averages ' without ultra-violet rays 3.3 2.0 ' 2.2 2.0 3.5 Averages with ultra-violet rays: 1.8 0.2 1.0 3.2 6.8 MOVEMENTS OF DAPHNIA 235 Frisch and Kupelweiser (1913) state that after a short expos- ure Daphnia magna and less evidently Daphnia pnlex become indifferent to the light and show no decided phototaxis so long as the intensity remains the same. When the light intensity is decreased the animals become temporarily positive while on increase of intensity they become temporarily negative. After a few minutes exposure to a new intensity the daphnids again become indifferent to the light and uniformly distributed through- out the dish. It has been shown above that in weak light or in light of moderate intensity Daphnia pulex remains constantly positive. Within the range of these intensities I have been unable to reverse even temporarily the phototaxis by sudden change of intensity. However, within the limits where the light intensity is strong enough so that the daphnids are normally indifferent, sudden changes of intensity produce the effects noted by the above mentioned authors. The tendency to these effects is probably present on any sudden change of intensity, but in weak light this tendency is overpowered by the normal phototaxis. It does not seem likely that these temporary changes of phototaxis can be of much importance in causing the vertical migrations of Daphnia. PHOTOTAXIS IN RELATION TO TEMPERATURE Loeb (1906) found that daphnids which were indifferent to light at 19° C. became positive when the temperature was re- duced to 11° C. When the temperature was raised to 25° C. they became again indifferent or weakly negative. My experi- ments show the same results. Daphnids indifferent to sunlight at 20° C. show positive phototaxis to the same light at low temperatures. This first becomes pronounced at about 12° C. Animals positive to diffuse light showed a slight reduction of the positive phototaxis on heating to 32° C. In diffuse light to which Daphnia pulex was positive at 22° C. Yerkes (1900) found no change of phototaxis on raising the temperature, although this was carried to the point where all the daphnids died. In weak or moderately weak electric light I have been unable to detect a variation in phototaxis on any amount of heating. We may then conclude that decrease of temperature causes a tendency to increase of positive photo- taxis while increase of temperature causes a tendency to decrease 236 LEE RAYMOND DICE of positive phototaxis, but that in weak light these tendencies are overpowered by the normal positive phototaxis. The temperature at which these animals become indifferent to light varies with the light intensity. Below 15° C. they seem to be positive in any intensity. To weak light they seem posi- tive at all temperatures. Above 25° C. in very strong light there may be a slight negative phototaxis, but if the ultra- violet rays are screened off this seems never to be at all strong. PHOTOTAXIS IN RELATION TO THE CHEMICAL CONTENT OF THE WATER Oxygen: No observable change in phototaxis was produced by changing daphnids from tap water containing much oxygen to the same kind of water from which the oxygen had been driven off by boiling for a half hour, or by the. reverse change. Carbon dioxide: Loeb (1906) has found that carbon dioxide will make neutral daphnids positive to light. Moore (1912) has further shown that this substance will make negative daph- nids positive. This chemical then has a tendency to render Daphnid pulex positively phototactic. Food: No observations are at hand bearing on the effect of food substances on phototaxis. Waste products: A very large number of daphnids were placed in a small vessel and exposed to weak light. They remained positive to this light until the excess of organic waste products in the medium grew so great that all died. The tendency of carbon dioxide to produce positive phototaxis seems to be the only factor in the relationship between the chemical content of the water and phototaxis that is likely to be of importance in the normal movements of Daphnia. PHOTOTAXIS IN RELATION TO MECHANICAL STIMULATION Yerkes (1900) states that some individuals of Daphnia pulex may be made temporarily weakly negative to light by mechan- ical stimulation, such as picking them up in a pipette, but with many individuals he failed to obtain this result. I have been unable to produce negative phototaxis by any amount of gentle or rough handling, either by picking the animals up in a pipette or by shaking the dish. MOVEMENTS OF DAPHNIA 237 PHOTOTAXIS IN RELATION TO TIME OF DAY Daphnids exposed in the darkroom to weak horizontal light from a 50 watt Edison lamp at 50 cm. distance, with observa- tions taken each hour for 30 hours, gave no evidence of a change in phototaxis corresponding in any way to a daily period. Neither could such a change be detected in daphnids exposed for 24 hours to the same weak light before beginning the series of obser- vations. In no experiment is there found the least evidence for believing in the existence of any sort of a rhythm in phototaxis independent of the direct effect of recurrent external changes. GEOTAXIS IN RELATION TO LIGHT INTENSITY When daphnids which have been kept some time in darkness are suddenly exposed to light of any intensity coming from a horizontal direction there is a tendency for them to go to the bottom of the dish (Exp. 7). Conversely, when daphnids which have been exposed to light of any intensity are placed in com- plete darkness they show a tendency to rise to the top of the dish. A tendency to positive geotaxis is also found on changing from weak light to stronger light (Exp. 8), or from moderately strong light to the light from an e