THE JOURNAL OF ANIMAL BEHAVIOR VOLUME 6, 1916 EDITORIAL BOARD Madison Bentley Edward L. Thorndike University of Illinois Teachers College, Columbia University Gilbert V. Hamilton Santa Barbara, California Margaret F. Washburn Samuel J. Holmes Vassar College The University of California Walter S. Hunter Jo hn B - Watson The University of Texas The Johns Hopkins University Harvey A. Carr The University of Chicago William M. Wheeler Editor of Reviews Harvard University Robert M. Yerkes, Harvard University Managing Editor Published Bi-monthly at Cambridge, Boston, Mass. HENRY HOLT & COMPANY 34 West 33d Street, New York G. E. STECHERT & CO., London, Paris and Leipzig, Foreign Agents Entered as second-ciass matter March 7, 1911, at the post-office at Cambridge, Boston, Massachusetts, under the act of March 3, 1879 CONTENTS OF VOLUME 6, 1916 Number i, January-February Lashley, K. S. The color vision of birds, i. The spectrum pages of the domestic fowl 1-26 Rau, Phil and Nellie. The biology of the mud-daubing wasps as revealed by the contents of their nests 2 7~^3 Myers Garry C. The importance of primacy in the learn- ing of a pig 64-69 Wheeler, William M. The marriage-flight of a bull-dog ant (Myrmecia sanguinea F. Smith) 7°~73 Wheeler, William M. Jean-Henri Fabre 74-8o Financial statement for 191 5. Number 2, March-April Sayle, Mary Honora. The reactions of Necturus to stimuli received through the skin 81-102 Fletcher, John M., Cowan, Edwina A., Arlitt, Ada H. Ex- periments on the behavior of chicks hatched from alcohol- ized eggs 103-137 DuPorte, E. Melville. Death feigning reactions in Tychius picirostris 138-149 I Torrey, Harry Beal. The physiological analysis of behavior 150-159 Turner, C. H. Notes on the feeding behavior and oviposi- tion of a captive American false spider. ( Eremobates formicaria Koch) 160-168 Number 3, May-June Johnson, H. M. Visual pattern-discrimination in the verte- brates — IH. Effective differences in width of visible striae for the monkey and the chick 169-188 Johnson, H. M. Visual pattern-discrimination in the verte- brates — IV. Effective differences in direction of visible -triae for the monkey and the chick 189-204 lie 7 H- CONTENTS iii Johnson, H. M. Visual pattern-discrimination in the verte- pages brates — V. A demonstration of the dog's deficiency in de- tail-vision 205-22 1 Burtt, Harold E. A study of the behavior of the white rat by the multiple-choice method 222-246 Churchill, E. P., Jr. The learning of a maze by gold-fish. . 247-2=55 Kempf, Edward J. Two methods of subjective learning in the monkey Macacus rhesus 256-265 Notes Hunter, Walter S. Titles of behavior papers 266 Number 4, July-August Yerkes, Ada W. Comparison of the behavior of stock and inbred albino rats 267-296 Wenrich, D. H. Notes on the reactions of bivalve mollusks ' to changes in light intensity: image formation in pecten. . 297-318 Notes Goodale, H. D. Note on the behavior of capons when brooding chicks 319-324 - Harris, J. Arthur. The habits in oviposition of the beetle Bruchus 325-326 Peterson, Joseph. Tone and noise perception in the white r at 3 2 7-329 Montane, Louis (C. S. Rossy, Translator). A Cuban chim- panzee 330-333 Vincent, Stella B. The refractive power of lens and fluid media of the mammalian eye 334 Number 5, September-October Walton, A. C. Reactions of Paramoecium can-datum to light. 335-340 Buddenbrock, W. von. A criticism of the tropism theory of Jacques Loeb 341-366 Notes Rau, Phil and Nellie. Notes on the behavior of certain soli- tary bees 367-37° Bingham, Harold C. Setting reactions of bird dogs to turtles 37^-373 Announcement 374 iv CONTENTS Numbkr 6, November-December Taliaferro, W. H. Literature for 1915 on the behavior of pages the lower invertebrates 375~3^> 2 Turner, C. H. Literature for 1915 on the behavior of spiders and insects other than ants 383-399 Wells, M. M. Literature for 1915 on ants and myrmecophils 400-406 Vincent, Stella B. Literature for 1915 on the behavior of vertebrates 407-418 Carr, H. A. Donaldson's " The Rat " 419 Herrick, C. Judson. Child's " Senescence and Rejuvenes- cence " and " Individuality in Organisms " 420-425 Donaldson, Henry H. Herrick's " An Introduction to Neu- rology " 426-432 Yerkes, Robert M. Hobhouse's "Mind in Evolution".... 433 Note Rau, P. A correction 434 Subject and Author Index VOLUME 6 Original contributions are marked by an asterisk (*) * A lcohol, effects on behavior of chicks, n io3. Allen, G. D. Reactions of Planaria, 375, 381. Andrews, E. A. Behavior of Follicu- lina, 375, 381. *Ant, marriage flight ,of, 70; literature on, 400. *Ape, record of birth, 330. Arey, L. B. Orientation in amphi- oxus, 411, 417. *Arlitt, A. H. Behavior of chicks, 103. Awati, P. R. The apple sucker, 386, 395. Back, E. A. Fruit fly, 386, 395. Baege, M. H. v. Animal psychol- ogy, 412, 417. Banks, N. Distribution of insects, 392, 395. Barnard. Feeding of butterfly, 395. Barrows, W. Reactions of spider, 380, 395. *Bees, behavior of solitary, 367. *Beetle, habits of, 325. * Behavior, of chicks from alcoholized eggs, 103; ^analysis of, 150. *Bingham, H. C. Reactions of bird dogs, 371. "Bird, color vision of, 1 ; ^'effects of alcohol on, 103. Bittner, L. H. Tropisms, 375, 381. Blackman, M. W. Life history of Pityogenes, 383, 396. Blair, K. G. Luminous insects, 386, 396. Bonner, W. Wintering of ants, 400, 406. *Brain weight, in relation to behavior, 267. Bromley, S. W. Asilids, 388, 396. Briin, R. The orientation of ants, 401, 406. *Buddenbrock, W. v. The tropism the- ory, 341; function of statocyst, 376, 381; the tropism theory, 376, 381. "Burtt, H. E. Behavior of white rat, Buttel-Reepen, H. Color sense of bees, 386, 396. */" > apon, behavior of, 319. [^ Carr, H. Donaldson's " The Rat," 419. Carter, L. A. The cyst of amoeba, 376, 381. Gary, L. R. Sense organs of jellyfish, 376, 381. Chapman, R. N. Life history of Agril- lus, 386, 396. Chemical senses, literature on, 408. "thick, behavior of capon toward, 319; *vision in, 169, 189. Child, C. M. Development of starfish, 377, 381; " Senescence and Rejuvenescence," 420; " Individuality in Organisms," 420. 'Chimpanzee, born in Cuba, 330. *Churchill, E. P. Learning in gold fish, 247. Cleare, L. D. Behavior of butterfly, 388, 396. Coburn, C. A. Behavior of pig, 414, 418; behavior of crow, 416, 417. Cole, L. W. Experiments with rac- coons, 412, 417. *Color vision, in birds, 1 ; literature on, 386. Commensalism, literature on, 394. *Cowan, E. A. Behavior of chicks, 103. Coward, T. A. Behavior of blackbird, 412, 417. Crawley, W. C. Marriage flight of ants, 401, 406. Crozier, W. J. Reactions of sea cu- cumber, 377, 381. Davis, W. T. An ant's nest, 401, 406. *Death feigning in Tychius, 138. Delayed reaction^ 412, 413, 417, 41S. *Dermal senses of Necturus, 81. De Voss, J. C. Color blindness of cats, 409, 417. VI INDEX Diguet, L. Behavior of spider and beetle, 394, 396. Dodson, J. 1). Habit formation in kit- ten, 41:5, 417. *Dog, vision in, 205; r< action (d turtles, 371. Donaldson, II. II. "The Eat," 419; Jlerriek's "Neurology," 420. Donisthorpe, J. K. British ants, 401, 406. Dunn, L. H. Behavior of Dermacenton, 387, 396. "DuForte, E. M. Eeactions of Tychius, 138. Ecology, literature on, 392. Elberfeld, horses of, 412. Elliott, J. S. The nest of ants, 402, 406. Emery, C. Natural history of ants, 402. 406. Essenberg, C. Habits of back-swim- mers, 383, 396 ; habits of water water-strider, 384, 396. Evolution, of mind, 433. : "Eye, measurements of, 334. *r?abre, Jean-Henri, life of, 74. 1 "Financial statement for 1915, 81. *Fish, learning in, 247. *Fletcher, J. M. Behavior of chicks, 103. Flint, W. P. Cornfield ants, 402, 406. Fly. behavior of, 391. Forbes, S. A. Applied entomology, 392, 396. l'risch, K. v. Color sense of bees, 386, 396. Frobawk, F. W. Hibernation of but- terfly, 391, 396. Fuller, C. Termites, 386, 396. Funkhouser. W. D. Life history of I anduzea, 386, 397. ( A anson, E. Color blindness of cats, J 409, 417. Carrey, W. E. Behavior of earth- worm, 377, 381. Gerhardt, \Y. Behavior of crickets and locusts, 386, 397. Girault, A. A. Geotaxis, 383, 397. behavior of bird-bug, 388, 397. *Goodale, H. D. Behavior of capons, 319. *O amt formation, in pig, 64; Jl *in goldfish, 247: •in monkey, 256; *in rat, 267 : literature on, 412. *Habits, of wasp, 27; ^mating, in ant, 70; *0f bees. 367. Haddon, K. Feeding of glow-worm, 388, 397. Eadwen, S. Warble flies, 395, 397. Hancock, J. L. Katydids, 386, 397. Hargitt, C. \V. Behavior of butter- Hies, 383, 397. *Harris, J. A. Habits of beetle, 325. 'Hearing, in rat, 327; literature on, 407. Hegner, E. W. Behavior of beetles, 387, 397. *Herrick, C. J. Child's "Senescence" and "Individuality," 420; " Introduction to Neurology," 426. Hewitt, C. C. Behavior of fly, 391, 397. Hibernation, of insects, 391. llirsch, G. C. Biology of gastropods, 377, 381. Hirshberg, L. K. Animal and human behavior, 412, 417. Hobhouse, L. T. " Mind in Evolution," 433. Howlett, F. M. Eeactions of fruit- flies, 383, 397. Hubbert, H. Habit formation in rat, 413, 417; errors in maze, 414, 417. *Hunter, W. S. Titles of behavior papers, 266; hearing in rat, 407, 417; delayed reactions, 412, 417; experiments with raccoons, 412, 417. Hutchison, E. H. Eesistance of para- moecium, 377, 381. Hyde, I. H. Effect of light on Vorti- cella, 378, 382. *Tnbred rats, compared with outbred, 1 267. *Insects, study of wasps, 27; *death feigning of, 138; "habits of, 325, 367; literature on, 383 ; distribution of, 392. *Instinct, brooding, 319; mating, 386 ; maternal. 387, 411; feeding, 388. Invertebrates, literature on, 375. Johnson, G. E. Tropisms, 375, 381. *Johnson, H. M. Pattern discrimi- tion in vertebrates, 169, 1S9, 205; vision in animals, 409, 417. Jones, B. J. Ants and fruit blight, 402, 406. INDEX vn Jordan, H. Movements of Muctru, 378, 382. Kanda, S. Geotropism in animals, 378, 382, 411, 417. Kellogg, J. L. Ciliary mechanisms of LamelUbranchs, 378, 382. *Kempf, E. J. Learning in monkey, 256. Kennedy, C. H. Life history of dragon flies, 386, 397. Kepner, W. A. Food reactions of Pla- naria, 378, 382. Kincaid, E. Effect of color on butter- fly, 384, 397. Knab, F. Commensal ism in insects, 394, 397. i Krecker, F. H. Orientation of Ephe- meridae, 383, 398. Kiihn, W. Hibernation of snail, 378, 382. * f ashley, K. S. Color vision of birds, L i Letisimulation, literature on, 394. *Light, reactions to, 98 ; •■reactions to, in mollusks, 297; ""'reactions to, in paramoeeium, 335. . Lloyd, J. T. Behavior of caddis worm, 389, 398. *Loeb, J. Tropism theory, 341 ; reactions of hydroid to spectrum, 378, 382. Lovell, J. H. Feeding of insects, 389, 398. Lyon, H. Behavior of flea, 388, 398. Mammals, literature on, 407. Mann, W. M. Insects fromHayti, 403, 406. Marcovitch, S. Biology of berry in- sects, 388, 39S. Mast, S. 0. Effects of spectral colors, 379. 382; behavior of fishes, 411, 417. *Maze, for study of fish, 247; "experiments with rats, 268. McAtee, W. L. Flight of caddis flies, 395, 398. McDermott, F. A. Reaction of house fly, 383. 308. Melndoo, N. E. Smell in Coleoptera, 384, 398. McShane, H. Transmitter of fever, 395, 398. •Mollusks. image formation in. 297. "Monkey, vision in, 169, ISO; ^learning in, 256. *Montane, L. Cuban chimpanzee, 330. Moore, A. R. Behavior of earthworm, 377, 381. Morrill. A. \V. Harvester ants, 403, 406. Muller, G. E. Trained horses, 112,417. *Multiple-choice method, 222; multiple-choice method, 414, 418. "Mvers, G. C. Learning process in pig, 64. 'JUatural history, of bee, 367. 1\| *Necturus, reactions of, 81. *Nest building, in wasp, 27. Neurology, Introduction to, 426. Nicholson, C. Hibernation of wasp, 391, 398. Noyes, A. A. Behavior of caddis worms, 390, 398. "/"Vviposition, in beetle, 325. ♦paramoeeium, reactions of, 335. 1 Paul, J. H. Regeneration in Crus- tacea, 379, 382. Pearse, A. S. Habits of fiddler crab, 380, 382. Pecker, S. Modification of Colpoda, 380, 382. "Pecten, image formation in, 297. Pemberton. Fruit fly, 386, 395. "Perception, auditory, in rat, 327. *Peterson, J. Auditory perception in rat, 327. Phipps, C. F. Behavior of amphipods, 380, 382. Pieron, H. Objective psvehology, 416, 417. Pieron, M. H. Instinct, 416, 417. *Pig', learning in, 64. Polimanti, O. Behavior of Planaria, 380, 3S2. Powers, E. B. Movements of fishes, 408, 417. *Oat, behavior in, 222, 267; IV *hearing in, 327 ; life history, etc., by Donaldson, 419. Ran, P. Behavior of wasp, 386, 398; correction of specific name, 434. *Rau, P. and N. Biology of wasp, 27 : "behavior of bees, 367. Redfield, E. S. P. Grasping organ of Dendrocoelum, 380, 382. Rich. A. Food reaction- of Planaria, 378, 382. Richardson. C. H. Breeding habits of wasp, 388, 399. vni INDEX Root, F. M. Life history of Podophyra, 380, 382. *Rossy, C. S. Cuban chimpanzee, 330. Safir, S. R. Habil formation in li.l- dler crab, 380, 382. Sanders, G. E. Habits of Xylena, 388, 399. Savlc M. H. Reactions of necturus, 'si. Schneider, K. C. Trained horses, 412, 417. Schwartz, B. Habit formation in fid- dler crab, 380, 382. Schwarz, E. Habits of Catocala, 386, 399. Seffrin, L. Smell in the dog, 409, 417. Senses, literature on, 380. Severin, 11. II. P. and II. C. Reactions of Tachinidae, 387, 399. Shannon. R. ( '. Svrphid fly, 388, 399. Shelfi.nl. V. E. Distribution of in- sects. 392, 399. movements of fishes, 408, 417. Shepherd, W. T. Intelligence of chim- panzee, 412, 417; intelligence of mammals, 412, 418. Skinner, H. Behavior of house fly, 391, 399. Smell, literature on, 409. Smith. II. S. Termites, 403, 406. Snyder, T. E. Inseets and forests, 403, 406. "Spider, behavior of, 160. Spreier, C. Effect of light on Vorti- cella, 378, 382. Stellwaag, E. Color sense of bees, 386, 399. Sturtevant, A. H. Sex recognition, 3S7, 390. u lrich, J. L. lis. Effort in learning, 414, 189, 205; aliaferro, W. H. Behavior of in- vertebrates. 375; 1 VI reactions of Rhabdocoeles, 378, 382. Theobald. Aphids, 404, 406. *Torrey, H. B. Analysis of behavior, 1*50. tropisms, 375, 3S1. Touch, in necturus, 82. Tropisms. literature on, 383. 411. *Tropism theory, criticism of, 341. Tschermak, A. v. Vision in fishes, 410, 418. *Turner, C. H. Feeding of false spider, 160; behavior of ant-lion, .".S3. 399; •behavior of insects. 383; habits of cricket. 386, 399; mating of ants, 404, 406. ♦TTertebrates, vision in, 169, V literature on, 407. "Vincent, S. B. Measurements of mam- malian eye, 334; "behavior of vertebrates, 407; rat and maze problem, 415, 418. ■'Vision, in birds, 1 ; "in vertebrates, 169, 189, 205; "in rats, 279; *in pecten, 297 ; '""mechanism of, 334; literature on, 409. *fTTalton, A. C. Reactions of para- W moecium, 335; delayed reactions in dogs, 413, 418. : \Yasp, biology of, 27. Wastenays, H. Reactions of hydroid to spectrum, 378, 382. Webster, F. M. Migration of butter- fly, 395. 399. "Weevil, death feigning in, 138. Welch, P. S. Reactions of Lumbricil- lus, 381, 382. "Wells, M. M. Literature on ants, 400; reactions of fishes, 408, 418. "\\ 'enrich, D. H. Reactions of mollusks, 297. "Wheeler, W. M. Marriage flight of ant, 70 ; *life of Fabre, 74; food of glow worm, 388, 399 ; ant cocoons, 405, 406. White, G. M. Behavior of trout, 412, 418. Wilhelmi, J. Grasping organ of Tri- chids, 381, 382. Williams, C. D. Behavior of pea thrips, 386, 399. *yerkes. A. W. Behavior of rat, 267. 1 Yerkes, R. M. Color vision in dove, 410, 418; instinct in monkey, 411, 418; behavior of pi,?, 414, 418; behavior of crow, 416, 417; *Hobhouse's " Mind in Evolution," 433. 'etek, J. Behavior of mosquito, 383, F 399. JOURNAL OF ANIMAL BEHAVIOR Vol. 6 JANUARY-FEBRUARY No. 1 THE COLOR VISION OF BIRDS I. The Spectrum of the Domestic Fowl K. S. LASHLEY Johns Hopkins University Evidence bearing upon the color vision of birds has come, thus far, chiefly from three sources; tests for the Purkinje phenomenon, discrimination tests with pigments, and photo- electric studies of the retina. Hess ('07) advanced the first evidence for the Purkinje phenomenon in birds. He placed white rice grains on a matt-black background, illuminated them with a spectrum, and placed fowls before them. The birds, light adapted, pecked first from the region of the orange through the red to the limit of the human spectrum, then in the yellow to blue-green, never in the blue or violet. Dark adapted and with a spectrum of low intensity they began to peck in the region of the yellow or orange-yellow and reached somewhat farther into the blue-green but never into the blue. Similar behavior was observed in the pigeon. The point of greatest stimulating value seemed to shift from the longer to the shorter wave-lengths with darkness adaptation. These results also led Hess to conclude that the spectrum of the day-bird is shortened at the violet end. In view of Watson's proof ('15) that the spectrum of the fowl is fully as extensive as that of man the validity of such a method as that of Hess becomes questionable. Katz and Revesz ('07 and '09) reported experiments with color-papers and stained rice grains in which fowls, when dark adapted and in dim light pecked more frequently at grains reflecting the shorter wave-lengths than when light adapted and 2 K. S. LASHLEY in bright light. A somewhat more conclusive test was made by Hess ('08) in measurements of the pupillary opening. The max- imum contraction of the iris was produced by longer light waves in light adaptation than in darkness adaptation. A preliminary report of the writer's experiments was given by Watson ('14). They are recorded at length in the following pages. Yerkes ('15) using a similar method, obtained evidence of the change in the relative brightness values of the red and green with change of adaptation in the ring-dove. Students of color discrimination in birds have usually em- ployed color-papers or dyes with no certain control of intensity or saturation. Porter ('06) records tests with the English sparrow and cow-bird in which red, yellow, green, blue, and two shades of gray paper were used. The birds learned very readily to distinguish between these. Rouse ('06) used almost identical methods with the pigeon and secured like results. Katz and Revesz ('07) trained chicks to pick out colored rice grains from among gray ones. As Hess ('12, p. 21) points out for the work of Katz and Revesz, all these results might have been obtained with completely color blind animals. Hess ('12) reports tests with rice grains stained to match the Seebeck- Holmgren wools. Fowls distinguished the reddish grains from the greens and grays of this series, which were confused by a red-green blind man. From this Hess concludes that the fowl is sensitive to the wave-length of the red and green rays. The uncertainty of conclusions as to the color vision of animals drawn from analogy with the condition of brightness vision in color blind men has been pointed out by Frisch ('14) and the evidence of Hess himself for a shortening of the fowl's spec- trum in the blue-green tends to invalidate his evidence for color vision. The careful work of DeVoss and Ganson ('15) shows in a practical way the justice of the most severe criticisms that have been urged against the use of color-papers in the study of animal vision. Rouse ('05), by recording the rate of respiration in the pigeon after stimulation with light, was able to show an increase in rate progressive with change from the longer to the shorter wave-lengths. He was not able to find like changes with alter- ation in the intensity of light of constant wave-length. The tests with wave-length and intensity were not carried out under THE COLOR VISION OF BIRDS 3 like conditions, however, and the results obtained are not com- parable. Katz and Rev6sz ('09) advanced evidence for color fatigue in the fowl after exposure to red which seems to prove some differential action of wave-length. Yerkes ('15) showed the ability of the ring-dove to distinguish spectral red from green but did not control the brightness of the lights. Finally the study of the action currents of the retina gives evidence of differential sensitivity to wave-length in day birds, Kohlrausch and Brossa ('14) obtaining characteristic types of galvanometer deflection for different wave-lengths irrespective of intensity. The correlation between action currents and color vision has not been made however. All this evidence offers good ground for the presumption of a true color vision in day birds, yet the results are in no case, absolutely conclusive. The question of the nature of the bird's spectrum, the number of primary colors, has not been approached experimentally. Hess describes the color vision of the fowl variously, as like that of a normal man, and as like that of a man looking through a piece of orange glass. He takes the latter view because of the supposed shortening of the spectrum at the violet end. Watson's data upon the threshold of the fowl make this view untenable. Beyond the probability that the bird is capable of some sort of reaction to wave-length, no definite conclusion can be drawn from the studies reviewed here. The details of spectral sensitivity, contrast, after images, color adaptation, and the role of color vision in the normal activities of birds are still open problems, promising to tax objective methods to the utmost. The present paper offers further evidence for the existence of color vision in the fowl, in the form of data upon the relative stimulating effect of different wave-lengths upon the light and dark adapted eye, the ability of the fowl to react upon the basis of wave-length, and the appearance of relatively abrupt changes in the stimulating value of different parts of the spectrum. The experiments were begun in 1912, but, owing to delays incident to reorganizing the apparatus and to pressure of other work, they have progressed but slowly. The chief problem and the method of attack I owe to the suggestion of Professor J. B. Watson who has also given generously of his time in the readjustment and control of the apparatus. I am indebted, 4 K. S. LASHLEY likewise, to Dr. A. R. Middleton for assistance in equating the energies of the stimulus lights. TECHNIQUE Game Bantam cocks of somewhat impure breed have been used in all the tests, differential reaction to the stimuli being obtained by the Yerkes discrimination method with food, or food and punishment, as the stimulus to learning. Concerning the suitability of the fowls for experiments little more com- plimentary can be said than that they fit the apparatus. They show most of the peculiarities recorded by Yerkes for the dove and by Tugman for the sparrow, plus sexual reactions which sometimes make work impossible. The Yerkes-Watson light apparatus, with some slight mod- ifications for ease of manipulation, provided the visual stimuli. This apparatus gives two fairly pure monochromatic bands of which the wave-length, energy, and saturation are under ac- curate control (Yerkes and Watson, '11). These bands are reflected from similar plaster surfaces, 1 by 10 cm., exposed in the alleyways of the discrimination box. Three surfaces are used, one light being exposed constantly upon the middle sur- face, the other being shifted to the right or left in unison with lateral movements of the experiment box, thus effecting the interchange of the relative positions of the lights. This technique introduces a source of error which must be controlled carefully. The primary distinction to be made in tests for color vision is between intensity and wave-length but reactions to other characters of the stimuli and experimental procedure must be guarded against. With the stimulus lights under complete con- trol the discrimination method offers five other groups of stimuli which may lead to deceptive results. These are listed below together with the method of eliminating them in the final trials, in part as recommended by Yerkes and Watson (op. cit. p. 84 ff). THE COLOR VISION OF BIRDS 2. Cues from the noise and shifting position of the apparatus. 3. Unequal illumination of the pas- sages due to the reflection of the shorter wave-lengths from black surfaces. 4. Differences in the pattern of the stimulus plates. 5. Light leakage in the apparatus and dark room. 1. Cues from the movements of the a. The experimenter was probably experimenter. never visible to the birds, certainly not in control experiments. b. White lights were substituted for the monochromatic ones. a. White lights as above. b. Interchanging the monochromatic bands so that all movements of adjust- ment were reversed. c. Pretense of shifts during training. a. The apparatus was built to avoid this. (Yerkes and Watson 1. c.) b. White light was substituted alter- nately for each of the monochromatic bands. a. Plates interchanged. b. Lights interchanged. c. White lights as above. a. Tests under 2 above. b. The stimulus plates were illumi- nated directly by small Tungsten lamps, with and without interposed color niters. These tests were all employed in the study of red-green dis- crimination and the more important also in all other tests except those upon the form of the spectrum. They gave always nega- tive results: occasionally the chicks were detected in attempts to follow the movements of the experiment box, but they never learned to distinguish between a real shift in its position and the pretended shift which was made at every fourth or fifth trial. The results of the other controls limit the reactions definitely to the illumination of the stimulus plates in every important test. They will not be given in detail except where they bear upon brightness vision. Adaptation was controlled in the following way. All tests were carried out in a completely darkened room with the ex- periment box enclosed in a black hood. When the chick made a correct choice a 4 ca. p. lamp 18 inches above the food dish was turned on until the chick picked up a bit of food, usually for less than one second. The chick was then returned to the starting box in darkness. Where light adaptation was desired, a 40-watt lamp with etched globe was held in the starting box for 30 seconds before each trial. That this produced a thorough brightness adaptation was shown by the following comparison 6 K. S. LASHLEY of the chick's time of reaction and the time required by the experimenter to adapt to the stimulus lights, 1/45 of the standard intensity. The experimenter's adaptation was always lighter than that produced by looking directly at the 40-watt lamp for 10 seconds. Chick Experimenter 22 seconds 20 seconds 14 " 16 4 6 18 " 10 6 4 14 " 20 2 3 " 6 " 2 " 3 * 3 When darkness adaptation was desired the chick was exposed to light for only the moment when he pecked at the food. Be- fore the experiments he was kept in the dark for one hour, this being the time required for complete darkness adaptation (Hess '07). THE PURKINJE PHENOMENON Experiments were begun with two Bantam cocks, A and C. They were first tested with white light and were found to be markedly photopositive, invariably going to an illuminated plaster surface. Two. white lights of different intensity were then introduced. The chicks showed a tendency to choose the brighter of the lights but their discrimination was very inaccurate and when the dimmer of the lights was well above their threshold they no longer selected the brighter: the reaction was evidently to the light as an object. An attempt was then made to increase the accuracy of their discrimination. Food was given at the brighter light and no food at the other. Training under these conditions was continued from October 23 until January 31 with ten trials per day. During this time it was found impossible to get the chicks to distinguish accurately between white lights of relative in- tensities less than ten to one (absolute intensities of 18 and 1.8 ca. m.). They showed a slight preference for the brighter of two lights of intensities three to one (18.0 and 6.0 ca. m.) but never made ten successive reactions without error. I believe that the ratio of three to one represents very nearly the difference limen of the fowl under the conditions of these experiments. It is possible, however, that with a stronger THE COLOR VISION OF BIRDS 7 motive than hunger the apparent limen might be much lessened. The statements of other investigators imply that the chick is very strongly photo-positive and will choose the brighter of two lights of almost equal intensity. As this was found to be untrue, long training in brightness discrimination was necessary in order to prepare the animals for the following experiments upon the Purkinje phenomenon. The training was continued for 500 trials after fairly accurate brightness discrimination had been established. Many changes in the brightness of the lights, position of the experiment box, noise of rotating sectors, illumi- nation of the experiment box, etc. were introduced to accustom the animals to changing experimental conditions. When reactions to brightness had become automatic mono- chromatic lights were substituted for the white lights. Those first introduced were Red 650/^ and Green 520^/*. The inten- sities of the two beams were equated with each other by the method and with the standard of Pfund 1 ('12, Yerkes and Wat- son, '11). a. The fowls, light adapted, were offered a choice of the stimulus plates with these lights at standard intensity, food being given at every trial. They chose — Green Red Chick A 12 9 Chick C 19 11 Average 63% 37% The green had perhaps the greater stimulating value under these conditions but the difference in brightness for the chick was evidently not very great. b. The chicks were next dark adapted for one hour and the illumination of the stimulus plates was cut down to 1/90 of the standard intensity. Chick C was tested under these conditions. He chose — Green Red 13 1 93% 7% At this energy the green had a much greater stimulating value. ^his standard has not been expressed in other terms. The stimulus patches illuminated at the standard energy have a luminosity which I estimate for 580 ufi. at 2 candles per square meter. In the tests for the Purkinje phenomenon the lights were equated with the Pfund standard after reflection from the stimulus patches. * In all other tests the energies of the beams were equated before reflection from the diffusing surface. 8 K. S. LASHLEY Control experiments indicated that the red was below the chick's threshold and that he was reacting to it as to darkness. The intensity of the lights was next increased to 1 /45 of the standard and Chick A was offered a choice of them. He chose — Green Red 9 1 90% 10% Furthermore, when offered a choice between the red and dark- ness he chose the red in 8 of 10 trials showing that the red was above his threshold. At low intensities the green has a much greater stimulating value than the red for the dark adapted chick. c. With the energies of the beams at 1/90 standard, Chick A. light adapted, was introduced into the discrimination box. He chose — Green Red 10 100% 0% As a control he was offered a choice between the red and dark- ness. He chose the red in 10 consecutive trials. The red used was below the threshold of the experimenter when light adapted and a record of the time of adaptation of the experimenter and the time of reaction of the chick shows that the chick also re- quired a slight darkness adaptation before the red became visible to him. His accurate choice of the green at this energy was undoubtedly due to the fact that the choice was made before the red came above threshold. d. The energies of the beams were increased to 1/9 of the standard. Chick A, dark adapted for one hour, chose — Green Red 5 5 50% 50% The energies were immediately #educed to 1/18 and he chose — Green Red 3 3 50% 50% Under the same conditions Chick C chose — Green Red 4 6 40% 60% When the energy is as great as 1/18 of the Pfund standard THE COLOR VISION OF BIRDS 9 the greater stimulating value of the green is lost, irrespective of the degree of adaptation. The results are thus far valuable only as illustrating sources of error in tests for the Purkinje phenomenon. In tests a and d no evidence for a greater stimulating effect of either light was obtained; in b and c the results are due, in all probability, to the reduction of the intensity of the red to threshold value and not to the process of adaptation. It seems probable that two factors may produce spurious evidence of the Purkinje phenomenon in experiments with animals. 1 : While two lights of unequal energy may not produce a reaction upon the basis of brightness at high intensity, a reduction of their intensity may bring the weaker below threshold while the other is still visible. From data obtained during training with white lights it is clear that when two lights differing greatly in stimulating value are exposed together the reaction to the less intense may be lost while it still persists when this stimulus is exposed alone or with another of equal stimulating value. In work of this sort it seems necessary, then, to distinguish two types of threshold; a lower threshold determined by the absolute sensitivity of the eye, and a higher one, which might be termed the threshold of attention, due to the interference of the stim- uli. 2: With lights of unequal intensity differing in brightness at high intensity by less than the limen of the animal the action of the Weber-Fechner law might produce a perceptible differ- ence following a reduction of the intensity of the lights. These factors, in all probability, would have produced the ♦results obtained in all the recorded tests for the Purkinje phe- nomenon if a band of white light of properly graded intensity had been used instead of the spectrum. I am inclined to think •that the methods of Hess, Katz and Revesz, Yerkes, and my own recorded above are too crude to give a proof of the Purkinje effect and that the results obtained are due rather to one or other of these secondary phenomena. The records of the Pur- kinje effect in man are so conflicting that an analogy between man and other animals seems hardly profitable. The resem- blance in the distribution of cones of the day bird's retina to the human fovea, where probably the phenomenon does not appear (Nagel '11), is sufficient, however, to demand extreme caution in interpreting results in this field. 10 K. S. LASHLEY With the uncertainty of the discrimination method it seems necessary either to introduce some technique which will elim- inate fluctuations of attention or to test for the relative stimu- lating effects of monochromatic lights when the energy remains constant and only the state of adaptation changes. In testing for the Purkinje phenomenon I did not realize the necessity for avoiding threshold stimuli and so carried out only a few tests with constant light energies. Contrary to the statements of Hess the red never seemed to have a greater stimulating value than the green of equal energy either for the light or for the dark adapted chick. (The energy of the red which he used was probably much greater than that of the green.) By a series of tests an energy relation was determined at which the red was just brighter than the green for the light adapted chick. This was at red 1/6, green 1/45 of the standard. When the energy of the red was decreased to 1/18 there was no choice, when it was increased to 2/9 it was chosen accurately and without hesitation. With red 650 h/j,, intensity 1/6, and green 520 pp, intensity 1/45, Chick C, light adapted, chose — Red Green 11 2 84% 16% He was dark adapted for one hour, then chose- Red Green 4 6 40% 60% Light adapted immediately he chose — Red * Green 3 1 Chick A under the same condition became erratic, developed a position habit, and gave no results that can be interpreted. The few records of Chick C indicate that there is a shift in the point of greatest stimulating value of the spectrum from the longer to the shorter wave-lengths with increasing darkness adaptation. The number of trials is too small for certainty. As there was the danger of association of one or other color with food work with red and green was dropped at this point and the apparatus was arranged to give yellow 590 pp. and blue-green 500 pp. at equal energies. Both lights were cut THE COLOR VISION OF BIRDS 11 down to 1/90 of the standard; at this energy both are above the chick's light adapted threshold. Chick A, fully light adapted was placed in the experiment box. He chose — Green Yellow 7 7 Chick C was dark adapted for one hour, then offered the lights. He chose — Green Yellow 7 5 Light adapted immediately afterward, he chose — Green Yellow 4 5 At this intensity neither of the colors has a markedly greater stimulating value than the other. At the same energy the red drops below the light adapted threshold. A point was determined at which the yellow was just brighter than the green for the light adapted chick. This was yellow 4/90, green 1/90. Light adapted, Chick A chose — Green Yellow 3 21 After dark adaptation he chose- Green Yellow 4 2 Chick C was tested in the same way but for some unexplained reason showed a complete breaking down of his reactions; even failing to choose between light and darkness. This lasted for several days and Chick A also became erratic so that the tests could not be continued without retraining for brightness. Such experiments, with food at every trial, make heavy demands upon the retention of earlier habits and it is surprising that the reaction persisted as well as it did. The apparent shift in brightness for the yellow and green, like the last test recorded for the red and green, indicates a Purkinje effect of adaptation. The determination of energies of the monochromatic lights having equal brightness values for the chick (the brightness equation point) is of interest when compared with the respective thresholds for the same wave-lengths and is of primary im- portance for tests for color discrimination. The values that I 12 K. S. LASHLEY have determined are very rough but they indicate that the relative stimulating value for the light adapted eye of mono- chromatic bands of high intensity is proportional to the threshold values of the same wave-lengths with light adaptation, as de- termined by Watson ('15). SENSITIVITY TO WAVE-LENGTH As the primary object of these experiments was to determine whether or not the chick is sensitive to differences of wave- length and as six months had been consumed already in the preliminary tests it was thought best to abandon the latter and test for color discrimination. Red 650 w. and green 520 /*/*. at equal energies were used as stimuli. Preference tests were of course impossible with chicks A and C. 2 Learning was rapid and after 200 trials the chicks were judged ready for control tests. The control tests described on page 5 limited the reaction to the illumination of the stimulus plates. Tests to distinguish between reaction to energy and to wave-length were made by altering the brightness relations of the saturated lights, and by diluting them. As determined in the Purkinje tests the brightness equation point for 650 pp. and 520 ^/t is some- where near the energy relations of 8 to 1. The energy relation of the stimulus lights was varied in both directions from this ratio as described in table 1. Tests with the colors diluted were made as indicated in table 2. The results of these tests, summarized in table 3, show that a reversal of the brightness values of the red and green did not seriously disturb the accuracy of the discrimination. The errors made appeared when the energy of the positive light was much reduced, irrespective of whether this was red or green. The dilution tests give fairly clear evidence of discrimination except in the case of Chick C with dilute red. 2 The behavior of some other birds in tests for color preference seems worth recording. Training for red-green was begun with five hens but not completed. Four of these showed no color preference for the lights at the brightness equation point for light adaptation. The fifth under the same conditions chose the green in preference to the red in 20 successive trials. This behavior suggests either an acquired color preference or color-blindness. Unfortunately, this bird was killed by rats before her vision could be tested. THE COLOR VISION OF BIRDS 13 TABLE 1 Tests for red-green discrimination with varied brightness relations of the stimulus lights. The energies at which the red was known to be brighter are printed in italics. In the remaining tests the brightness value of the green was equal to or greater than that of the red. Energies are expressed in decimals of the Pfund standard. Chick A (Positive to red) Chick C {Positive to green) Energies Energies No. of trials No. of errors No. of trials No. of errors Red Green Red Green 1.0 1.0 5 1 1.0 1.0 5 1.0 .022 5 0. 1.0 .022 10 2 1.0 .044 5 1.0 ■ 044 5 1.0 .088 5 1.0 .088 10 3 1.0 .177 5 1.0 .177 5 1.0 .355 5 1.0 .355 5 1.0 .500 5 1.0 .500 5 1.0 1.0 5 1.0 1.0 5 .022 1.0 9 3 .044 1.0 10 1.0 .022 6 1.0 .022 10 3 .022 1.0 10 3 .022 1.0 10 2 .088 1.0 10 i TABLE 2 Tests for Red-Green Discrimination with Varied Saturation of the Stimulus Lights Condition of stimuli Chick A No. of trials No. of errors Chick C No. of trials No. of errors Both lights at standard energy Green much diluted Bright white light substituted for green Both lights at standard energy Red much diluted Red diluted until its quality was lost . . . Both lights at standard energy White light of low intensity for red .... Same white for green White light of high intensity for red . . . Same white for green Both lights at standard energy Two white lights of equal brightness . . . 5 10 5 10 10 7 5 8 10 5 10 5 10 1 3 4 1 3 5 5 6 5 5 10 5 12 6 6 5 6 10 1 1 4 'o 4 3 2* * A careful examination was made for secondary criteria which might account for this choice which tends to invalidate all the results obtained with Chick C. Nothing could be discovered by the dark adapted human eye. The test was re- peated on the following day, when ten trials with five errors were obtained. The above record was probably a chance variation. 14 K. S. LASHLEY TABLE 3 Summary of First Tests for Red-Green Discrimination Chick A (Positive to red) Chick C (Positive to green) Condition of stimuli No. of trials No. of errors No. of trials No. of errors Red 1 : Green 1 35 29 16 ♦ 10 10 7 13 25 20 5 10 1 6 1 4 7 13 ToD. ToD. 4 31 20 20 6 10 18 16 20 15 10 1 Red below 0.1: Green 1 3 (green brighter) Red 1 : Green below 0.1 5 (red brighter) Green dilute: Red 1 Red dilute: Green 1 4 Red greatly diluted : Green 1 White light for red 7 White light for green 1 Controls White light of equal energy 7 Red 1 : Darkness ToD. 5 Green 1: Darkness ToD. 1 The tests in which white light was substituted for each of the lights in turn are interesting as indicating the relative im- portance of the negative and positive stimuli in determining the reaction. Both chicks confused the white with the positive stimulus (table 3) and distinguished it readily from the negative. Punishment had been used in this case with the negative light which here appears to be most clearly recognized. In later experiments without punishment conditions appear to be re- versed, the negative light being most frequently confused with the white. This is in accord with the view that punishment is the more efficient factor in learning. The tests also evidence either color discrimination or accurate perception of brightness. The use of the lights singly gives further evidence of the same sort, since darkness was chosen fewer times with the positive color than with the negative (table 3). The work had to be dropped at this point. The tests recorded make it highly probable that the birds were reacting upon the basis of wave-length, but as the energies of both red and green were never altered simultaneously, the possibility that an ex- tremely accurate perception of absolute brightness furnished the basis for reaction was not completely eliminated. It seemed THE COLOR VISION OF BIRDS 15 advisable therefore to repeat the experiment using still more thorough controls. After a long delay, while smoked wedges were substituted for the Aubert diaphragms used in equating light energies, the work was resumed. Three cocks, A (now three years old), and two younger ones, D and E, were trained for discrimination with red 650 w. and green 520 m*. uncalibrated, just as given by the carbon arc spectrum. Chick D was trained to avoid the red (middle stimulus plate) and choose the green; Chicks A and E to choose the red. After about 400 trials the chicks began to react perfectly. They were given an additional hun- dred trials to make the reaction more nearly automatic — then, as the lights could not be equated immediately, a series of rough control tests for brightness and secondary criteria was begun, partly to accustom the chicks to a changing problem and partly as a test of their sensitivity to wave-length. Three types of tests were employed to distinguish between brightness and wave-length; 1, first one, then the other light was reduced to threshold intensity while the other remained at the full en- ergy of the spectrum; 2, a white light of constant intensity was substituted for each of the colored lights in turn; 3, each light was exposed alone with one passage completely dark. Table 4 shows the records of the chicks in the first of these tests. It will be noted in the results of this experiment that reduction of the intensity of the negative color produced no disturbance of the reaction while reaction of the positive color led to many errors. Had the reaction been due to the relative brightness of the stimulus patches Chick D should have been disturbed when the red was reduced, as were A and E, since this must then have represented the condition in which the difference in brightness approached threshold. Instead of this, however, Chick D reacted inaccurately only under conditions where A and E remained undisturbed. This seems to limit the effective stimulus either to the wave-length or to the absolute intensity of the positive stimulus patch. 16 K. S. LASHLEY TABLE 4 Reactions to Red and Green Lights of Estimated Threshold Value , Intensity of lights Chose Red Green R. G. Chick A. (Positive to red) Just above threshold Well above threshold Full intensity Full intensity Full intensity Just above threshold 47 20 10 18 Chick E. (Positive to red) Just above threshold Well above threshold Full intensity Full intensity Full intensity Just above threshold 35 14 15 10 1 Chick D. (Positive to green) Just above threshold Full intensity Full intensity Just above threshold 23 9 12 The second group of tests, made with Chicks D and E gives additional evidence against a reaction to the relative brightness of the stimulus patches. A very intense white light (the image of a Nernst glower directed upon the stimulus patch by a 2 -inch lens of 6-inch focus) was first substituted for the red light, the green light being left at full intensity somewhat diluted by light reflected from the white stimulus patch. Under these conditions Chick E (positive to red) chose — ■ White 11 Chick D (positive to green) chose — White 3 Green 1 Green 13 The white was then substituted for the green and the red was restored at full energy. The records of the birds were — Chick E. Chick D. White 1 4 Red 9 1 The intensity of the white light was many times that of the monochromatic so that the brightness relations of the stimulus patches were certainly reversed in the two tests. The prob- ability of reaction to absolute brightness is also reduced by the dilution of the monochromatic lights by light reflected from the white patch. For me this reflected light had greater bright- ness than either of the monochromatic lights. To avoid a reaction to secondary criteria given by the shifting THE COLOR VISION OF BIRDS 1? of the lights and experiment box the monochromatic lights were arranged so that both could be projected upon the same stim- ulus patch. They were exposed in irregular order with the other patch illuminated with the intense white light. Under these conditions Chick E chose — Red 6 White 6 when r and w were exposed together. Green 1 White 5 when g and w were exposed together. Chick D chose — Red 2 White 4 Green 5 White under the same conditions. The white light was immediately reduced to threshold intensity and the tests repeated. Chick E chose — Red 8 White 1 Green 3 White 8 Chick D chose — Red 2 White 5 Green 5 White The change in the intensity of the white light with the resultant change in the dilution of the monochromatic points to a sen- sitivity to wave-length. The complete change in the technique eliminates a reaction to secondary criteria. The conclusions from the third set of tests are based upon the time of reaction to a single monochromatic light: the time required for the chick to go from the door of the starting box to the food dish, a distance of 3 feet. The results of the tests with the three chicks are given in table 5. During these tests the intensity of the lights was varied from 1/18 to the full energy of the spectrum. The birds, although instinctively photo- positive, showed a marked delay in advancing to the negative light, irrespective of its intensity. Unless the brightness differ- ences of the monochromatic lights for the chicks are extremely great this experiment is sufficient to eliminate reaction to the absolute intensity of the light. These tests are given in some detail as illustrating a fairly satisfactory technique when the energy and brightness values of the stimulus lights are unknown. Such methods have the disadvantage of introducing many disturbing elements; in such tests the animals ultimately become confused and give con- flicting results. 18 K. S. LASHLEY TABLE 5 Time of reaction to one light alone. The time is that required by the chicks to go from the door of the starting box to the food dish, a distance of three feet. Red light alone at varied intensity No. of trials Average time No. of errors (Darkness chosen) Chick A. (Positive to red) 8 12 5 8 8 14 1 . 45 sec. 2. 12 sec. 30.60 sec. 7.25 sec. 16.37 sec. 5.60 sec. Chick E. (Positive to red) Chick D. (Positive to green) 4 Green light alone at varied intensity Chick A 3 (Refused to advance in three trials) Chick E 2 Chick D At the end of these tests the lights were equated in energy, the chicks were retrained for a few days to increase the accuracy of their reactions and then tested with the lights under accurate control. The energies of the lights were varied widely in both directions from the brightness equation point and were likewise varied in absolute intensity at different brightness ratios. The results of the tests with the three chicks are given in table 6. At the end of the experiments the chicks were offered two white TABLE 6 Tests for red-green discrimination with varied brightness relations of the stimu- lating lights. Arranged as table 1. Chick A. Chick E. Chick D. (Positive to red) (Positive to red) (Positive to green) Energies No. of trials No. of Energies No. of trials No. of Energies No. of trials No. of R G errors R G errors R G 2.0 1.0 15 1 2.0 1.0 20 2.0 1.0 20 1 1.0 1.0 68 8 , 1.0 1.0 20 4 1.0 1.0 20 1 1.0 .5 10 .5 1.0 5 .5 1.0 10 1 1.0 .25 10 .5 .11 10 2.0 .11 12 1 1 .0 .10 10 1 .25 .5 10 1 Ini ured by Chi ckE. 1 .0 .0? 10 .25 .11 20 2 2.0 ■05 10 1 .11 1.0 15 2.0 .08 5 .11 .11 9 1 2.0 .15 10 2.0 .11 15 .44 1.0 15 2.0 .08 10 .22 1.0 20 2 .22 .22 10 1 THE COLOR VISION OF BIRDS 19 lights of equal intensity exposed under the same conditions as the monochromatic. Chick A chose the stimulus patch to which he usually reacted positively 7 times, the other 8 times. Chick E chose the central patch 6 times, the other 4. Chick D could not be tested. A was further offered a choice of two white lights of relative intensities 1 to 90. He chose the brighter 7 times, the other 3, an accuracy less than that with any in- tensity of monochromatic lights. These tests, with the other controls mentioned in the discussion of technique limit the efficient stimulus to the illumination of the stimulus plates. Summarizing table 6 we find that with the red certainly brighter than the green the chicks made 3 errors in 72 trials or 4.1 per cent, error. In 325 trials with the green brighter or equal to the red in brightness they made 23 errors or 7.7 per cent. Further, in trials where the intensity of both lights was greatly altered there was no reduction in the accuracy of choice, so that the possibility of reaction to the absolute brightness of either stimulus is eliminated. As between brightness and wave- length, then, these experiments speak conclusively for the latter. The poss bility of other characters of the stimulus lights them- selves furnishing the basis of discrimination will be considered after the work with yellow and blue light has been recorded (page 20). The same three fowls were trained with yellow at 588 PP. and blue-green at 500 pp. with energies equal. Chicks A and E were trained to choose yellow, D to choose green in order to take advantage of the previous training. Strangely enough the chicks were confused by the green at 500 pp, not greatly different from 520 pp. to my eyes, and failed to discriminate in the first trials. They had previously reacted to 520 pp when white light, darkness, or yellow were substituted for the red. However, after 120 trials they gave less than ten per cent, error and were considered ready for control tests. Tests for the Purkinje effect had previously shown that the following bright- ness values held for these wave-lengths for the moderately light adapted eye of the fowl. Yellow Green 4 1 yellow brighter. 2 1 yellow perhaps brighter. 1 1 neither brighter. 1 45 green certainly brighter. 20 K. S. LASHLEY Tests at these values were carried out with the results shown in table 7. The ratio of 45 to 1 gives uncertain results owing TABLE 7 Tests for yellow-green vision with varied brightness relations of the stimulus lights. Figures in italics show where the yellow is brighter for the chick. Energies Chick A. Chick D. Yellow Green Trials Errors Trials Errors 1.0 1.0 14 2 20 4 .01 1.0 20 2 20 2 .11 2.2 14 1 .05 1.0 10 2.0 i.o 29 6 12 1 1.0 1.0 10 1 to difficulty with the apparatus: a reduction of the yellow below .1 of the standard brings it near the chick's threshold 3 and a greater energy of the green than 2.2 could not be obtained with the present arrangement. The tests are therefore not as extensive as I desired. However, chicks long trained to bright- ness discrimination failed to distinguish this red and green even at the energies of 1 to 1 and the tests with variations in both directions from this ratio seem adequate to eliminate a reaction to brightness. As in the case of red and green the effective stimulus is restricted to wave-length. The evidence for color vision thus far presented seems to prove that the chick can distinguish between monochromatic lights of any intensity between threshold and the Pfund standard, irrespective of the brightness or saturation. The effective stim- ulus is the wave-length. The question remains, is the discrim- ination based upon retinal sensitivity to differences of wave- length or upon some entoptic phenomenon. The difference in visual acuity in light of different wave- length is well known, and, though it seems to be largely a matter of relative brightness it may furnish a basis for discrimination. Whether a like variation in acuity occurs in the color blind I cannot discover from the literature. Allied to this phenomenon is that of diffraction or dispersion within the eye. At low intensities the red and yellow stimulus patches have for me sharp outlines while the green of equal brightness lacks sharpness 3 The absolute intensities of the lights in this experiment are not comparable with those of the Purkinje tests. The introduction of the smoked wedges cut down the energy of the spectrum so that the equation of the lights with the Pfund standard had to be made before they were reflected from the plaster surfaces instead of afterward as in the first experiment. THE COLOR VISION OF BIRDS 21 and seems to be surrounded by a luminous halo. It seemed possible that the birds were reacting upon the basis of the sharpness of the visual objects rather than upon their color, either as a result of diffraction or of serious chromatic aberration. Tests were made at different times during the experiments with a view to controling these factors and gave the following results. 1. Wide variations in the size and shape of the stimulus plates were without effect. 2. High intensities and dilutions, reducing the apparent diffraction of the shorter wave-lengths and giving sharp out- lines for me to both stimulus plates did not effect the reaction of the birds. 3. Intense white light with imperfect color filters, completely altering the secondary phenomena for me, were discriminated as promptly as the spectral lights. 4. The evidence given in the following section for abrupt changes in the reaction value of the spectrum is difficult to explain by these phenomena. The visual acuity of the fowl determined by Johnson ('14) seems to eliminate the question of any great chromatic aberration in the chick, and the foregoing tests make a reaction upon any other basis than retinal sensitivity highly improbable. THE CHARACTER OF THE SPECTRUM With the establishment of sensitivity to differences of wave- length the problem of the difference limen for wave-length in different parts of the spectrum at once suggests itself. A thorough solution of this problem would require months of training with careful controls of brightness at every stage. The Yerkes- Watson apparatus is not well adapted for such exper- iments as it cannot be arranged readily to give lights from adjacent parts of the spectrum. The following tests, serve to give a clue to the form of the spectrum of the bird, particularly with reference to the existence of regions of widely different reaction value separated by relatively short intervals of the spectrum. Chick A, positive to 650 pp. and negative to 520 pp. was offered red 650 and yellow 580 pp. He chose the red six times in seven trials and showed little confusion. The red was evi- 22 K. S. LASHLEY dently different in reaction value from the yellow. The element of intensity was not eliminated here but since the chicks had been reacting so well to wave-length and not at all to brightness differences I think it unlikely that the discrimination was made here or in the following tests upon the basis of brightness. Chick E, trained like Chick A, was given 7 trials with 650 /*/*. and 520 \x>\i, making no error. The red was then shifted toward the shorter wave-lengths with the following results. Character of longer waves for the human eye Red. Red-orange. Orange-yellow. Yellow. Position habit. Yellow-orange. Secondary evidence of discrimination. This indicates that there is a change in the reaction value of the spectrum between 580 and 590 mm. This is the region in which the yellow assumes an orange quality for man. Chick D, positive to green, was first given 5 trials with the red and green as in training, making no error. The test was continued as follows: Wave-length, mm Chose Wave-length, /uju. Chose "Red" Green "Red" Green 650 520 7 610 520 5 600 520 6 580 520 3 3 590 520 3 Red "Green" Red "Green" 650 520 5 650 600 5 Reactions prompt. 650 630 Position habit developed at once. This indicates a change in the reaction value of the spectrum between 600 and 630 ^. This is the region of the transition from red to orange for man. Chick A was given the following tests. Wave-length, nu. Chose "Red" Green "Red" Green 650 520 5 610-590 520 5 590-565 520 . 5 1 (The reaction here with green and yellow was given upon the basis of the green as negative color, the chick going to darkness in preference to green.) 550-530 520 2 2 Refused to advance. THE COLOR VISION OF BIRDS 23 This suggests a change in the character of the light between 530 and 565 a 1 /*, corresponding roughly to the change from yellow to green for man. After training with yellow and blue-green Chick D was given the following tests. Wave-length, pi.fi. Chose "Yellow" Green "Yellow" Green 580 500 1 6 535 500 Right position habit. 565 500 5 565 500 3 (1/18 standard) There is a change between 535 and 565 w. This agrees with the results obtained from A. The change in the energy of the yellow was made as a slight control of thejprightness element. Chick A was offered green 530, violet 460 \ni. The use of the yellow-green was due to the difficulty of obtaining lights nearer together in the spectrum. In an earlier experiment after train- ing with red and green Chick A had been offered a choice between white and red, white and blue, and red and blue (impure color niters) . He chose accurately red and blue in preference to white, red in preference to blue and all in preference to green 500 fifi. I expected him to choose the violet in preference to the green and refused him food at the latter. Instead,' however, he chose the green five times in succession, then, getting no food, de- veloped a position habit. He was then fed at the green and immediately chose it five times in quick succession. Chick D was offered a choice between 500 and 460 w. He showed no choice and developed a position habit immediately. It is rather difficult to interpret the results of these two tests with the assumption of identical color divisions for the chick and man. In the beginning of training with yellow and blue the shift from 520 to 500 w. seemed to confuse the chicks (page 19). If we suppose that there is a marked change in the re- action value of the spectrum between these wave-lengths, there seems to be a general agreement in the results; for Chick A 530 M/i and 580 /x/x. resembled each other more than they did 500 At/*, and 460 A 1 /* perhaps resembled 500 A 4 /*; for Chick D 500 w and 460 a 1 /* were more nearly alike than the original 580 /j>/j> and 500 pp. The accord of the animal's behavior with this interpretation justifies the assumption, pending more accu- 24 K. S. LASHLEY rate investigation, that there is a marked change in the reac- tion value of the spectrum between 500 and 530 up with no marked change between 500 and 460 pp, i. e., the region corre- sponding to the blue and violet for man begins at a longer wave- length for the chick. To summarize this part of the work, the accuracy of the discrimination of the fowls is affected by changes in the position of the stimulus lights through the following intervals and it is suggested that the wave-lengths lying between each pair of these intervals, which are probably shorter than the tests indicate have a fairly uniform reaction value. 630 — 600 mm. 590 — 580 " { 565 — 550 " j 565 — 535 " j 520 — 500 " "1540 — 470 " THE RELATIVE REACTION VALUE OF INTENSITY AND WAVE-LENGTH In field studies of animals where the conditions of illumination are not under control it is difficult to determine whether the animal, even if known to have color vision, is reacting to color or to brightness. The distinction is of importance in many biological problems where the necessity for natural conditions makes an actual test of the efficient stimulus impossible. The only clue to it in such cases is given by the relative ease with which habits of reaction to the two attributes of the light stim- ulus are formed. The data upon the rate of learning in my experiments gives some indication of the relative efficiency of wave-length and intensity in light stimuli. a. Birds already trained in brightness discrimination, when trained with red and green of different brightness values for them (red 650, green 520 pp. at equal energies) changed quickly to reaction upon the basis of wave-length, without any alteration in the brightness values of the lights which might have destroyed the habit of reacting to brightness. After training with colors the association with wave-length was dominant even when the brightness values of the stimuli were so greatly different as to call out the instinctive brightness preference in untrained birdss b. The conditions under which training for brightness and color discrimination have been carried out are not strictly THE COLOR VISION OF BIRDS 25 comparable, but some unrecorded experiments by Dr. Middleton and myself upon the fowl's difference limen for brightness indi- cate that where threshold intensities are not involved habits of reaction to brightness are established with much greater difficulty than are those of reaction to wave-length. From the evidence even of a. alone I believe that the field experimenter may feel confident that, if the birds show a differ- ential reaction to colored objects not differing enormously in brightness for him, their reactions are made upon the basis of wave-length. DISCUSSION One fact the experiments seem to have brought out clearly: The fowl is sensitive to differences of wave-length irrespective of intensity and its sensitivity is greater than, or at least different from, that of any form of partial color blindness in man. Less certainly it appears that for the fowl the spectrum is divided into areas of widely different reaction value, of which there are not less than five, probably more. It is not probable that all these occupy areas of the spectrum corresponding to the more conspicuous divisions for man. There is also some slight evi- dence for a Purkinje effect but this needs verification. The proof of color vision in the fowl should make possible the more ready correlation of the so-called physiological methods of studying sensitivity with other behavior methods. The prob- ability that the specific action-currents of the retina following stimulation with different wave-lengths are an index to a more general color sensitivity is increased. The lack of such specific action-currents in night birds furnishes a final test of the inter- pretation of retinal action currents. Experiments upon the color vision of owls are in progress. The value of such methods as those of Rouse and Babak is also enhanced by a verification of some of their results by other methods. The reflex responses offer an index of sensitivity more accurate and convenient than the discrimination method, but they must be correlated with other activities of the animal before they can furnish a basis for attack upon more general problems. The probability of a high degree of sensitivity to wave-length in other day birds suggests anew the problems of organic and sexual selection in birds, problems which have largely fallen into disrepute of late. The prominent role played by sensitivity 26 K. S. LASHLEY to wave-length as compared to brightness offers the possibility of field and experimental studies of these much discussed questions where the experimenter may feel reasonably confident that his results are due to color perception. The conclusion of Katz and Revesz that the Purkinje phe- nomenon cannot be explained by the process of adaptation in the rods does not seem to be justified since neither the Purkinje phenomenon nor the absence of visual purple has been estab- lished for birds. It seems too early yet for a satisfactory corre- lation of structure and function. Further studies of adaptation upon birds with different retinal composition will doubtless help to clear up the question of the function of the retinal elements. LITERATURE CITED DeVoss, J. C. and Ganson, Rose. Color Blindness of Cats. Journal Animal 1915. Behavior, 5, 115-139. Hess, C. Ueber Dunkeladaptation und Sehpurpur bei Hiihnere und Tauben. 1907. Arch. /.. Augenheilk., 57, 298-316. 1907. Untersuchungen liber den Lichtsinn und Farbensinn bei Tagvogeln. Ibid., 57, 317-327. 1908. Untersuchungen tiber das Sehen und tiber die Pupillenreaktion von Tag und von Nachtvogeln. Ibid., 59, 143-167. 1912. Vergleichende Physiologie des Gesichtssinnes. W inter stein' s Handbuch der vergleichenden Physiologie. Bd., 4, Jena. Johnson, H. M. Visual Pattern Discrimination in Vertebrates. Journal Animal 1914. Behavior, 4, 340-361. Katz, D. and Revesz, G. Ein Beitrag zur Kenntnis des Lichtsinns der Htihner. 1907. Nachr. d. K. Ges. d. Wiss. zu Gottingen, Math.-physik. Kl, 406-409. 1909. Experimentelle-psychologische Untersuchungen mit Huhnern. Zeitschr. f. Psychologie. Abt. 1., 50, 93-116. Kohlrausch, A. and Brossa, A. Die Photoelektrische Reaktion der Tag- und 1914. Nachtvogelnetzhaut auf Licht verschiedener Wellenlange. Arch. f. Physiol., 5, 421-431. Nagel, W. Helmholtz. Handbuch der physiologische Optik. Bd. 2, 207. Leipzig. 1911. Pfund, A. H. Application of the Selenium Cell to Photometry. Physical Review, 1912. 34, 370-380. Porter, J. P. A preliminary Study of the Psychology of the English Sparrow. 1904. Amer. Jour. Psych. 15, 248-271. 1906. Further Study of the English Sparrow and other Birds. Ibid. 17 248-271. Rouse, J. E. Respiration and Emotion in Pigeons. Jour. Comp. Neurol. Psychol. 1905. 1J" 494-513. Watson, J. B. Behavior. Holt, New York. 1914. 1915. The Spectral Sensitivity of Birds. Papers from the Tortugas Labora- tory. Carnegie Inst. Pub. 211. Yerkes, R. M., assisted by Eisenberg, A. M. Preliminaries to a Study of Color- 1915. vision in the Ring-dove, Tutor risorius. Jour. Animal Behav. 5, 25-44. Yerkes, R. M. and Watson, J. B. Methods of Studying Vision in Animals. 1911. Behav. Monographs. Vol. 1, No. 2, iv+90. THE BIOLOGY OF THE MUD-DAUBING WASPS AS REVEALED BY THE CONTENTS OF THEIR NESTS PHIL RAU AND NELLIE RAU Saint Louis, Mo. With five plates This paper is primarily a study of the contents of the nests of three species of mud-daubing wasps. Before entering upon a discussion of how the nests are provisioned and how the young fare in their struggle against adversities and enemies, it will be well to become familiar with the nests themselves, their structure, their sites, decorations, etc. Fossorial Hymenoptera are roughly classified, according to their habits, in four groups: 1. Those which build no special receptacles for their young but are either parasitic or subparasitic, or take advantage of the abodes of other insects, holes, etc.; 2. Builders of cells in the form of pottery constructed of clay mixed with water or saliva and dried; 3. Excavators of burrows in the ground; 4. Makers of tunnels in wood or the stems of plants. 1 The three species of Fossorial wasps herein dealt with belong to the second group of this classification, those wasps which construct cells of clay or mud as receptacles for their young and its provisions. The mothers themselves do not use these structures for their own shelter, but go elsewhere for the night. All three of these species are spider ravagers. First we have the pipe-organ type of mud nests (see figs. 1, 2, 5) built by Trypoxylon albitarsis Fab., the shiny black wasp with beautiful white toe-tips. Then we have the familiar mud nests which at first sight look like a clod of earth; these are made by two species belonging to the subfamily Sceliphroninae, Sceliphron (Pelopoeus) caementarium Drury, a pretty black wasp trimmed up in yellow (this wasp is commonly known by the boys in my neighborhood as "yellow-legs"), and the beautiful steel-blue wa sp, Chalybion caeruleum Linne. The nests of these 1 Sharp, D. Insects, Pt. II. Cambridge Nat. Hist., p. 90. 28 PHIL RAU AND NELLIE RAU two species are so similar as to be indistinguishable (figs. 3,4, 6, 18). Some authors have placed these two wasps in the same genus, probably chiefly because of their similar habits of nesting, but I have found that there is some difference in the spinning of the pupal case by the larvae of the two species, and Mr. Rohwer writes to me that he believes that the two species "should be retained in different genera for besides a difference in coloring there is a shortening of the petiole in C. caeruleum and certain other differences which indicate that it is of a different group." Thus we have convergence of habit of nest-building in three distinct genera. NIDIFICATION OF S. CAEMENTARIUM AND C. CAERULEUM During the sunny days of summer one may see many of these mud-daubers coming to the edges of streams and puddles, criti- cally selecting mud of exactly the right consistency and literally standing on their heads biting out chunks of it and carrying it to some distant shelter. There they fashion it into the familiar cells illustrated in figs. 3, 4, 6, 18. The size and shape of these pellets which they carry may be seen in fig. 13. These were dropped by insects taken at such sources, and are exact size. The smaller ones however are probably incomplete balls dropped by wasps which were interrupted during the gathering. It is generally thought, and all the treatises (excepting the Peck- hams' 2 ) state that this mud is mixed with the saliva in the mouth of the insect. I do not wish to discredit this statement, for I have no proof to the contrary, but since it is generally made on supposition, I think it would be permissible for us here to suppose also that the amount of saliva used by an insect in a day would be many times the size of the little body producing it, especially when the wasp sometimes constructs a whole cell in an hour and a half, or on other occasions when she is building incessantly, whole days at a time, or reinforcing the nest, as shown by the thick walls in figs. 7, 9 and 10. In these two species and in T. albitarsis as well, the young from eggs deposited in the early fall feed and go through their metamorphosis during the winter and emerge as adults the next 2 Instincts and Habits of Solitary Wasps, p. 178: "The wasp adds nothing to the mud, depending upon its drying for the necessary firmness, and if by some accident the rain strikes it the whole becomes soft and falls to pieces." Plate I Fig. 1. Pipe-organ nest of T. albitarsis. Reduced. Fig. 2. Same, back view. Reduced. / THE BIOLOGY OF THE MUD-DAUBING WASPS 29 May or June. Here at St. Louis we have both summer and winter broods, and I have a faint suspicion that those nests which harbor the winter brood of 5. caementarium and C. caeru- leum are built more massive than the summer cells. This question is worth investigation to determine whether the wasps are endowed with the instinctive power of apprehending the approaching season. When one sees the enormous thickness of some of the walls (e.g., figs. 7, 9 or 10) in contrast with the thin walls and partitions often found in the summer nests (fig. 20), one is almost tempted to attribute this to something other than individual traits, although we have found an enormous amount of individuality expressed in the work of 5. caementarium. One finds, for instance, much variation in the careful or slipshod manner in which load after load of mud is applied to the nest. Fig. 4 shows how precisely each mouthful is sometimes applied to form half a ring on the cell, while fig. 8 reveals the careless way in which another mother applied her plaster, without any regard for size, shape or security. I am sure that a nest made as loosely as this would be an easy mark for Hymenopterous parasites of the genus Melittobia 1 or even Anthrenus larvae when minute, had not this mother made up for her carelessness by thickly daubing mud over the outside of the nest. This photograph shows the details of the contours after the loose reinforcement was carefully scraped off. The nests of 5. caemen- tarium and C. caeruleum usually have additional mud daubed over the outside, no doubt for the purpose of strengthening the structure and increasing its warmth. Some are decorated on top of this with whole pellets stuck here and there over the surface — not flattened or spread as usual (figs. 3, 18). This serves no utilitarian purpose, so far as we can see, and is not generally done, but the very fact that it is sometimes done, shows that individual differences in manipulation exist, and to account for the origin of this difference leads one into fanciful speculation. It certainly seems farfetched to say that some females have developed an aesthetic taste. Better would it be to say that this habit of decorating is a vestige from the time when one or both species, living in milder climate, made a crude nest entirely of small round pellets carelessly stuck together. Later when climatic changes or migration northward occurred, 3 They can pass through holes .013-inch in diameter. 30 PHIL RAU AND NELLIE RAU those few individuals which builded strong survived, until now only this vestige remains of what may have been a former habit. An idea of the amount of effort expended in the construction of a, nest can be gained by examining the cross-sections illustrated in figs. 7 and 9, and the longitudinal section in fig. 10. The material used in the nest is usually native clay, gray, red or various shades of brown, or even black loam. The Peckhams describe several cells made of white plaster, and I have several nests taken from the walls of a brick-yard stable which have some cells made of brick-dust and the remainder of black mud. Sometimes even the pellets of mud, which make up a single cell are clearly from different sources. The favorite nesting-place of these species is the rafters of barns or sheds, especially when they are high and well out of reach of intruders. In areas untouched by man the overhanging rocks and bluffs are favorite nesting-places. One also finds the mud nests on the back of torn or hanging wall-paper, in the folds of horse-blankets, old carpet or grain-sacks hanging in out-buildings, in the sleeves and folds of hanging coats, on old umbrella ribs, on strands of hay or corn-husks or in small spaces between the rafters, but in every such case the nest conforms to its immediate surroundings. It frequently happens that the nests built in close quarters are very beautiful, since the builder cannot obliterate her art here by daubing mud promiscuously all over it. The nests built in folds of cloth or on strands of hay are usually more asymmetrical than those built flat upon a board or an overhanging rock, and conform to the folds of the drapery. It was once my good fortune to gather a bushel of mud nests from some freight cars side-tracked in the city. This suggests a mode of the dissemination of the species not heretofore con- sidered, and may well be a large factor in their distribution, since the Peckhams and others think that the wasps do not usually migrate far from the place of their birth. On one occasion, when searching for mud nests, I was taken to a farm-house where the nests were abundant in the upstairs rooms. The house had not been occupied for the two years previous to that summer and insects had made themselves at home in it. The nests were abundant upon the walls and window-casings of the upper rooms and many wasps were busily Plate II _>il 4?' ^ 6 Fig. 3. Nests of S. caementarium or C. caendeum, with decorations of mud pellets. Natural size. Fig. 4. One-celled nest of mud-dauber. Natural size. Fig. 5. Pipe-organ nest of T. albitarsis. Reduced. Fig. 6. A common form of nest of S. caementarium or C. caendeum. Natural size. THE BIOLOGY OF THE MUD-DAUBING WASPS 31 coming and going. But when I proceeded to gather them, they all proved to be cells from previous years, and I found that the good new nests were only on the ceiling, just out of reach. It seems that the wasps built their nests wherever their fancy dictated so long as they were unmolested, but when people occasionally moved about in the room they promptly chose sites higher up, out of the range of disturbance. The two species of wasps, Sceliphron caementarium and Chaly- bion caeruleum make mud nests which are very similar in appear- ance. The species of the builder is ascertained with accuracy only by finding the dead pupae or adults in the cells or by noting the species of the adults as they emerge. Occasionally the nests of both are decorated with pellets. The only constant distinction which I have been able to discover lies in the differ- ence in the structure of the cocoon; that of 5. caementarium is smooth, glossy and brittle while that of C. caeruleum is the same but covered with a webby mesh. 4 The occasional occur- rence of both species of wasps emerging from one colony of cells does not necessarily mean that an erring mother has de- posited her egg in the wrong nest, although it would seem to us almost impossible for a mother to find her own nest among hundreds of others as we sometimes see them massed in the lofts of large barns. (In our barn, 643 were gathered and about three times as many remained). The phenomenon is easily explained by. the fact that the mud-dauber's nest occasionally occurs on top or along side of the pipe-organ nests, and on several occasions we have found them plastered to paper wasps' nests. In fact one curiosity which we have is the nest of these two species of mud-dauber and a pipe-organ nest all subjoining a large paper nest. We can see that probably the mother seldom commits the error of laying her egg in another's nest, but when choosing a site on which to build she sometimes regards the architecture of her sisters the same as the side of a barn. NIDIFICATION OF T. ALBITARSIS Trypoxylon albitarsis makes the Pipes of Pan, or as they are more frequently called the pipe-organ nests. Fig. 5 shows the nest as it usually occurs; the short tier is in course of con- struction. T. albitarsis does not daub the nest all over with 4 Details in Psyche, Vol. XXII, p. 62-63. 32 PHIL RAU AND NELLIE RAU mud when it is completed, thereby hiding its artistic architec- ture, but it, like the other two species, smooths the interior of each cell carefully. The holes at regular intervals in the nest are made by the emerging adults, and the white spots are the hardened chalky substance which the insects emit from their bodies immediately after they emerge. This white sub- stance is emitted by 5. caementarium and C. caeruleum also, but in the form of many minute pellets discharged before emergence. The larva of T. albitarsis spins a very light web about the walls of its cell; just inside this it constructs its cocoon, black very strong and brittle. This it probably makes of the excre- ment which it clears from its alimentary tract after feeding is completed, and utilizes for a cocoon by throwing it all over itself. This is then kneaded into shape by dextrous movements of the body; it then hardens and makes a very comfortable abode. 5 These nests are usually built against some flat surface; some- times the back side is lined with mud, and sometimes the board on which it is plastered serves as a back wall. Here in fig. 2, we have a view from the rear of one such nest built without a back wall, showing cells, partitions, pupal cases, spiders, empty cells and two cells in course of construction. In this nest are three instances of the emerging insect opening its way into an adjoining cell instead of to the outside; if the adult belonging in the latter cell had already emerged, this wasp could leave by its exit, but if not it must certainly die imprisoned, because instinctively the insects can open their way through only one wall. In contrast to this flat form I have seen about a dozen nests built on hanging corn-husks. Here without a substantial foundation the cells attained a perfectly cylindrical shape, as thick on the back as on the front. They afforded sufficient protection in all respects, however, for they normally brought forth healthy adults. On only one occasion have I been able closely to observe the details of the building operations of T. albitarsis in progress. 5 1 have tried to observe this process by placing larvae about to pupate in vials. But it seemed they were unable to work on the smooth glass, for they produced only thick ribbons of the black material, but no pupal case. Plate III *■*> r .- ■ JT Fig. 7. Cross section of mud-dauber's nest showing the thickness of the walls and the relative position of the cells. Natural size. Fig. 8. Nest of mud-dauber, showing the details of the architecture after the reenforcing layers of mud had been removed. Natural size. Fig. 9. Cross section of nest. Slightly reduced. Fig. 10. Longitudinal section of the mud-dauber's nest, showing the size of the cells and the thickness of the walls. Slightly reduced. THE BIOLOGY OF THE MUD-DAUBING WASPS 33 This nest that I was so fortunate to discover in course of con- struction was in an accessible position so one could easily watch the methods of building. It contained two complete tiers and the third was being added. The wasp returned with a mouthful of mud, but our presence disturbed her and she flew away. Only on her fourth return was she content to settle down to her work without heeding our presence. She would remove the load of mud from her mandibles to her front legs and apply it to the structure in its proper place and then smooth and work it down with her head. Her mud puddle could not have been far off, for she required only from one to three minutes to make the round trip and bring her load of mortar. As usual, the male remained, in the tube that was being con- structed while the female brought the mud and continued the building. The male sometimes poked his head out of the cell to meet or greet his spouse, in a way very similar to the habit which we have seen in the smaller species of Trypoxylon (7\ clavatum). A long pipe was first constructed. This was then filled for a certain distance with spiders and an egg laid with them and a partition put in, making a cell out of this section of the pipe. This process was repeated until the entire pipe was divided into cells, and then a second tier was made beside it. Since most of these nests are built vertically, with the openings downward, one wonders what prevents the spiders falling out while the cell is being filled, the egg laid and the mud applied for the partition. I have wondered whether the male did not in some way perform this office. It was soon necessary for me to leave, so we captured the female and took down the nest. We found the male 'way up in the topmost corner of the unfinished tier. The two older tiers were complete in every way, properly partitioned and sealed, and each cell contained spiders and a young wasp. Dr. William H. Ashmead 6 says:"Walsh was the first to record the habits of Trypoxylon albitarsis, which usually selects the deserted cells of a mud-dauber (Pelopoeus) in which to nidificate, provisioning its cells with spiders. I can confirm this statement of Walsh's from personal observations, as I have not only ob- tained them from the old cells of Pelopoeus but also from those of Chalybion caeruleum. I have also bred T. clavatum from the same mud-dauber's cell." 6 The Habits of Aculeate Hymenoptera, II., Psyche, Vol. VII. p. 45. 34 PHIL RAU AND NELLIE RAU While I can agree with Mr. Ashmead in regard to obtaining adults of T. clavatum from the nests of Pelopoeus or Chalybion, I cannot agree to the statement of both himself and Mr. Walsh that T. albitarsis uses the old cells of both these species. Many hundreds of adults have emerged from pipe-organ cells in my possession. In addition I have invariably seen these wasps at work on these nests, so I can hardly think of them as using old cells of other wasps. One often finds nests parts of which, and especially the middle parts, are two, three or even four tiers high, one built on top of the other, as shown in fig. 1, where two layers exist and a third is being commenced. Whether one mother constructs tier upon tier, or whether a second mother builds her nest on top of the first I do not know. This condition occurs almost too frequently to be attributed to mere accident. Nevertheless the occurrence of nests in this form brings out an interesting point in the instinct of the emerging wasp. Fabre 7 carried on some experiments with the mason-bee Chalicodoma muraria, in which he found that instinctively the animal could bore out of only one earthen covering, and while apparently it had the physical ability to emerge from an extra covering it would rather die in its prison-house than make the extra exertion to escape. But how fares it with these wasps whose mothers or aunts build tier upon tier, making it necessary for the emerging wasp to bite through more than one wall before gaining its freedom ? It is interesting here to note that the insects seem to have some faculty of discerning the front side of their cells, just as the larva of the blue wasp or the yellow- legs has the faculty of righting itself about in its cell as it reaches maturity so that it always pupates with its head toward the exit. In T. albitarsis nests of only one layer of cells, one seldom sees an error in choosing the proper side for exit. It has been my good fortune to obtain a few nests which were two or more layers deep, and to study their contents to learn whether these wasps did plod faithfully on until they gained their freedom, or gave up at the first defeat. I have little doubt that there is sufficient strength in those mandibles to penetrate several walls, so the question is merely one of instinct. One nest of two layers was placed flat upon a table so that the emerging insects 7 The Mason-Bees. Tr. by A. T. De Mattos. Chap. II. Plate IV 9 12 - I " 15 13 18 17 Fig. 11. A deserted mud-dauber's nest, the cells used by 0. cordata and the openings resealed with waxy material. Natural size. Fig. 12. Pupal cocoons of T. albitarsis after having been used by 0. cordata; a, complete pupal case with old opening resealed with wax; b, waxy plug removed, showing hole made by emerging 0. cordata; c, part of cell opened to show pupal case of 0. cordata; d, cell of T. albitarsis opened to show partitions and plug of waxy substance. Natural size. Fig. 13. Mud balls taken from the mandibles of mud-daubing wasps. Fig. 14. Longitudinal section of mud-dauber's nest showing the pupal cases of 0. cordata between the waxy partitions. Natural size. Fig. 15. Pupal cases of 0. cordata taken from deserted cells of mud-daubers' nests. Fig. 16. Pupal cases of T. clavatum. Fig. 17. Old nest of mud-dauber, with two mud partitions, and mud plug in old opening made by 0. lignaria. Fig. 18. Nest of mud-dauber, showing decorations of mud pellets. Reduced one-half. THE BIOLOGY OF THE MUD-DAUBING WASPS 35 could not gain their freedom by way of the under side, which was partly open; the table merely replaced the board wall upon which it had been built. Fourteen wasps escaped in the normal way from the top layer and the periphery of the lower layer. Seven adults in the central cells could not escape, but each one bored through the wall and entered the adjacent cell, where they were found dead. They followed no special direction in boring out of these lower cells; one had bored through the front wall, and the other borings were equally divided on either side. In one such dungeon three dead prisoners were found, the original inmate of the cell and the neighbor from either side. Had each one of these seven mature insects had the instinctive courage or energy to push on through one more wall, all would have escaped. In another nest we found where this same in- ability had brought death to one insect, the only one in the nest that was so situated that its exit led into another cell. In a third nest the same was true for two insects. We cannot call this defective instinct, but only simple instinct, for in the insects' normal experience they should have but one wall to penetrate in order to gain their freedom. One would like to say at least that it is wonderful that the emerging wasp knows how to direct its exit towards the ligr^t, but even this is not always the case. I have another record of an eight-celled, one-story nest in which three individuals had bored through the side-wall into their neighbor's cell and died there, instead of breaking through their own roof to free- dom. So even this instinct of working out toward the light is sometimes defective. PARASITES AND LESEES Often one finds other species occupying the nests or preying upon the wasps at home. Among these are: Parasites Cuckoo-bees, Trichrysis tridens 8 Lep. and T. parvula* Fab. Dipteron, Toxophora americana Guerin. Dipteron, Spogostylum fur O. S. Dipteron, Argyromoeba obsoletum Loew. Hymenopteron, Melittobia. 8 Identified by Mr. S. A. Rohwer. 36 PHIL RAU AND NELLIE RAU Museum pest, Anthrenus scrophulariae L. Mutillid wasp, Sphaerophthalma scaeva Blake. House-renters Bees, Osmia cordata Robt. and 0. lignaria Say. Eumenid wasp, Ancistrocerus (Stenancistrocerus) fulvipes Sauss. Eumenid wasp, Ancistrocerus camestris Sauss. Wasp, Trypoxylon clavatum Smith. Wasp, Trypoxylon cocker ellae Rohwer. Wasp, Pseudagenia mellipes Say. Wasp, Pseudagenia adjuncta Banks. Meal-worm larva, Tenebrio obscurus. Small click-beetle. Lepidopterous pupa, Apatela, possibly radcliffei. Mites and spiders. PARASITES Perhaps the most destructive of these is MelittobiaJ The parasites enter the mud nests by making holes about .013-inch in diameter, enter the pupal cases in the same way and lay their eggs. The larvae eat the contents of the cell (usually in the prepupal stage), and escape as adults, leaving the cocoon empty. These adults bore holes through the neighboring walls and enter other cells and the life cycle is repeated. That these parasites bore into cells promiscuously is evidenced by the fact that their borings are also found in empty cells. Considering their large broods and the rapidity with which they develop, and their ability when ready to migrate, to bore into the mud cells, it is surprising that even one larva should escape them. They usually attack the organism in the prepupal stage, boring into the cell and then into the pupal case. They do not eat the pupal case, nor do they mutilate it, but leave it intact ex- cepting a few small punctures, and containing a quantity of fine debris — probably their own shedding-skins. The Dipteron which is directly parasitic upon both 5. caementarium and C. caeruleum is Argyromoeba obsoletum, Loew. 10 These flies are by no means rare. The larva is found in the wasp's pupal cocoon; this shows that the parasite attacks the 9 Identified by Dr. L. O. Howard. 10 Identified by Mr. F. Knab. Plate V CM CM § ^'*) CO CM 10 CM ids o SJO 0) *L <• ,--^:> « WUte JB& CM 3- 1 "O ~ ~J5 > *^ rj C x: K ^ '-> O M .is co o o .x: %-. co -a .„ c - c '- cu ^ CX O ° u "O o CU CL) o •r: ij cu N 03 3 — 03 2 SO . "5 5 ■• £ ~ 5 ■ a jg -^ . IS- - 1 ~ c Q.3 r O M ?-« S >- >>8 £ m ■; n 3 3 a; ■ 2 C « ■■a £^ £-<-3 o u& UP" fin CU -l-i rj CU CU > — X- O 03 +-> e ^■O a) * C ■" "2^3 "n n-S ° > D. O T3 to "^ 3 v* cu n CO co ac co » O -C 3 J2 t^ X3 rt O T3 i~ O '-3 eg ^«§^ ^^ ^'.r-3 d <3 O 03 +j 3 5 ■S < "3 u Z a h CO "3 ■a Cd •o o o bo bO c '> 'Sc CO "3 u 6 Z •o o & •v CI CO J£ ° bo %« CD °o « Z £ o •a 22 ~a CI v> "3 u o Z •a .c •a jy "w 01 CO B 3 CO "3 o z ■a c « T3 .22 "cd U CO C 3 jn 3 ^_ °E d 2 tu id > .2 0) •0,0 S = cd u ■=83 w 3 H en cj G Z eg cd > .2 CO 1*2 1 = •=-c .2 83 "3 3 O CO o5 d* Z id ci •o .c +j CO 0) 3: cd d 0. 3 -- a C 01 u da Z •0 ca CI •v •J3 '% CO "3 u CI . CO o£ do. Z •0 cd 0) •a ■4-» '5 _co 3 CO *^ "3 °5 . ■O V N '■k-t '35 cd u cd CO '55 cd u cd a u d bo C "> cd J3 CO "3 U CI > cd bo 4-» cd ■S -*J CO 3 CO O-o w cd = T3 « °8 1912 552 447 51 15 2 18 1 1 6 2 4.0 5, D 81% 1913 203 112 15 11 1 2 2 1 12 10 1 17 12 7, D 55% 1914 527 413 60 1 1 24 1 6 9 7,0 5, D 79% Total 1282 972 126 27 1 2 4 20 37 12 13 26 14 28 76% 62 PHIL RAU AND NELLIE RAU How cheering it is to find that 972 of these, or 76 per cent, gave forth good adults in the spring after wintering in the cells. In 126 cases (10 per cent.) did the mothers err in filling the cell or fail to oviposit; 27 cells (2 per cent.) were sealed empty, but in some of these cases the sealing partition was so placed that the room was so small that a growing larva could not possibly have had room to develop anyway. We have never yet noticed Sceliphron committing this blunder of constructing her cells too small to permit the full development of her young. There was only one cell that was unsealed containing a few spiders; this case may well have been due to the sudden death of the mother. In only 4 cases can death in the larval stage be attributed to insufficient food, while 20 others died in the larval stage despite the fact that much food remained untouched in their respective cells. The deaths in the prepupal, pupal and adult stages were 37, 12 and 13 respectively, lower per cents than occurred in the other species. These insects were preyed upon by some of the parasites of Chalybion and Sceliphron but not to so great an extent; 26 cells were parasitized by Melittobia and 14 showed that the Anthrenus beetle had played havoc; 28 cells were divided between a Dip- terous parasite and an Ichneumon belonging to the tribe Ophio- nini, indicated in the table by "D" and "O." Thus elimination in all the stages of development is far less in T. albitarsis than in the other species. So here we have three species so similar in morphology that we cannot discern any differences which would give one or the other the least advantage in the struggle for existence. Like- wise their life history, their habits of living and their habits of nesting are so similar that it would appear to us that they have equal chances of surviving. And yet there exists this remarkable difference in the survivability of these species, and even more strange it is that the species which has the fewest numbers in this vicinity is the one which thrives far better than the others. Had the variation studies, for which the material was first intended, been carried to a conclusion, we might thereby have located some tendencies which would account for the comparative vigor or weakness in these species, but I think it more than likely that in pursuing variation problems THE BIOLOGY OF THE MUD-DAUBING WASPS 63 purely, we would have been sidetracked on some feature of the insect's anatomy which is entirely irrelevant to its ability to work out its own salvation. But if now the variation studies could be taken up in the light of the preceding paragraphs, they might lead us to profitable problems of the existence or nonexistence of correlation between variation and elimination, or in other words, to add one more word to the hypothesis of selective elimination. THE IMPORTANCE OF PRIMACY IN THE LEARNING OF A PIG GARRY C. MYERS Brooklyn Training School for Teachers During the summer of 1911 the writer performed some ex- periments on an eight-week old pig. While there are many obvious imperfections in the experiments the records emphasize an important aspect in the learning process, namely, the tenden- cy of the first of a series of habits to possess a relatively high persistency. The apparatus was very simple. When not fed the pig was shut in a pen X from which a door Y led into a lane L 23 feet in length. Near the end of the lane was an opening O which led into the general pen P whose measurements were 10 x 12 feet. Five feet from either side of P was a board K 7 inches high, parallel to the lane. This board was so high that the pig could not naturally see over it. In either of the corners of P farthest from O and equally distant from it was placed a food receptacle which could be transferred from one corner (A) to the other corner (B). JS' In the first instance the pig was released at Y and found his way into A where he soon found the food, morning noon, and PRIMACY IN THE LEARNING OF A PIG 65 evening of the first day. On the second and third feedings he ran directly to A without hesitancy. Each time, as soon as his meal had ended, he was shut up in X. The food always, was put into the receptacle before the pig was released so that he could not see it until he had crossed K. Second Day Fed every 2 hours. Food was put regularly into B. First feeding. When released the pig ran directly to A as on the first day, not finding food he rambled about much dis- turbed for several minutes. Finally he ran a little farther to the right than before and found the food at B. Second feeding. The same procedure as in the first, but in about half the time. Third feeding. He ran as before, directly into A; then immediately to B. Fourth feeding. Same as third feeding, but in less time. Each time when he had consumed the food at B he would return to A looking for food there. He was never fed to satiety. Fifth feeding. Before he was released, a partition C, 8 inches high was placed mid-way between the two feeding places. He entered the A side directly then, with almost no hesitancy, jumped over the partition toward B. There he was so dis- turbed he did not get the food at B but jumped out over K. After a number of quick false movements he jumped back over K to B and ate the food. There was no more running to A after the food was eaten. Sixth feeding. Jumped directly into the A side, came out over K and with no false moves turned to the right and jumped over K to B. Seventh feeding. Same as sixth. Third Day (Still fed at B) First feeding. He made a bee-line from O to A, back over K, to the right, and to B over K. Second feeding. He made a bee-line for A but only raised his head and looked over K, paused a few moments, turned and jumped over K near partition, into B side. Third, fourth, and fifth feedings. Made a bee-line from O to B. 66 GARRY C. MYERS Fourth Day He was fed only three times, morning, noon and evening, when he always made a bee-line from O to B. ► Fifth Day (Feeding place changed to A) First feeding. He made a bee-line for B, searched for food with restless movements, jumped out over K, then scarcely moving to left at all, jumped over K to A. Second feeding. He jumped into B, then with practically no pause, over partition to A. Third feeding. Started for B side, but looked over, paused, then into A. In going to B side, however, his pathway diverged a little toward the partition, a decided modification of old path- way from O to B side. Fourth feeding. Same as third but divergence greater, i. e., nearer partition. Fifth feeding. He ran directly toward A side but very close to partition, paused without looking over, and jumped directly into A side. Sixth and Seventh feedings. Made a bee-line to A. Sixth Day (Feeding place changed to B) First feeding. He made a bee-line to A but jumped over partition quickly to B, as if planned, with no appreciable pause. Now another board of 6 inches was added to partition C, making it 14 inches high. Second feeding. He made a bee-line to A but this time he came out of A over K, then to right into B side. Third and fourth feedings. Same as second. Fifth feeding. In bee-line to A side but near partition looked over, then turned and jumped into B. Sixth feeding. He ran directly into B side, but near par- tition. No pause. Seventh Day (Morning. Feeding place changed to A) First feeding. He ran in a bee-line into B out over K, and into A. Second feeding. He ran directly to B side near partition looked over and very quickly jumped into A. Third feeding. Bee-line to A. PRIMACY IN THE LEARNING OF A PIG 67 Seventh Day (Afternoon. Feeding place changed to B) First feeding. He ran directly to A looked over cautiously, then jumped into B. Second feeding. He ran to B side, near partition, looked over and jumped into B side. Third feeding. Directly over B side near partition. Fourth and fifth feeding. Same as third. Sixth and Seventh feedings. Bee-line from O to B. Eighth Day (Morning. Feeding place changed to A) First feeding. In bee-line toward B, looked over K, and put fore feet over K, then drew back and jumped into A. Second, third, fourth, and fifth feedings. Bee-line from O to A. Eighth Day (Afternoon. Feeding place changed to B) First feeding. Bee-line for A, quickly out and into B. Second feeding. Directly to A side, looked over, then into B. Third feeding. Same as second. Fourth feeding. Directly to B side near partition and into B. Fifth feeding. Bee-line from O into B. The following figures show the number of times the pig was fed A and B respectively and the number of right and wrong trips to each. No. times Wrong No. times Wrong fed in A trips to A fed in B trips to B First Day 3 Second Day 7 7 Third Day 2 5 Fourth Day 3 Fifth Day 7 4 Sixth Day 5 6 Seventh Day (Morning) 3 2 Seventh Day (Afternoon) 1 7 Eighth Day (Morning) 5 1 Eighth Day (Afternoon) . 3 5 18 18 33 7 While in all, the pig was fed 18 times in A and 33 times in B, he wrongly entered B (i. e. when the food was placed in A), 7 times and wrongly entered A 18 times or, 33.3 per cent, and 54.5 per cent of the responses, respectively, were wrong. It took only the first feeding to establish a direct pathway from O to A, but after two repetitions of this trip it took ten G8 GARRY C. MYERS trials before the subject learned to go directly from O to B. After going directly to B three successive times it took only five trials to learn to go directly from O to A again. To make the next transition of the pathway (from OA to OB) it took six trials. Hence, while primacy tends to persist throughout it decreases somewhat with time. The modification in the learning is worthy of note. The excited, random movements so obvious in the earlier new situ- ations which successively appeared with the alternation of feed- ing places, decreased as the experiment progressed. On the other hand the responses to the later situations were character- ized by hesitancy and by attempts to use the eyes to help deter- mine, before jumping over the board K, which was the right way to go. To illustrate, on the twelfth feeding which was the second feeding of the third day, after having gone from A to B eight successive times, he paused at K at a point on a straight pathway to A from O, raised his head high enough to look over K, then turned and jumped over K near the partition into the B side. Likewise at the third and fourth feeding of the fifth day, after making two wrong trips to B, he paused at K, looked over and turned and jumped into A. His pathway, moreover, instead of being in a straight line from O to B diverged a little toward the partition, and on the succeeding trip diverged still more, so that the next time he jumped directly into the A side but near the partition. At the fifth feeding of this same day, instead of looking over K, he ran to it in line with A, paused, and jumped into the A side. These gradually shifting pathways seem to indicate the resultant of two antagonistic impulses. It must be remembered here that although the subject often took the wrong pathway from O, eventually he always found the food; but in finding the food after first entering the wrong side the subject never went back to O for a new start, but sought the food by the shortest way he could find from where he was. No doubt the transitions from one pathway to another, from O, could have been completed with fewer errors if the corrections could have been made from O. Even then, theoretically, there would have been some tendency to make the wrong trip again and to return to O rather than to go to the food directly. In any event, the mere recency of the appeal of the stimulus where it once was, does not so much account for the tendency to con- PRIMACY IN THE LEARNING OF A PIG 69 tinue to take the wrong pathway as the persistency of the impulse to go from O by the pathway that last was successful from O to the food. It is interesting to note that the pathway taken at the first feeding of each morning was always the same as the pathway taken last on the previous day. An example from the writer's experience of an attempt to break a certain habit of forgetting may help explain the prin- ciple involved. On two successive evenings he forgot to ex- tinguish the gas light in the cellar after he had made the fire. On the third evening as he closed the cellar door behind him, on entering the kitchen he remembered that he had left the light burning and he at once returned to extinguish it. For about a score of evenings he went through exactly the same process of forgetting and correcting. Just as soon as the thought of the light suggested itself when the fires had been made, the proper reaction to the light was elicited. Again, it is not so much the attraction of a stimulus that determines a certain reaction as it is the precedence and strength of that reaction to that stimulus. The results of this experiment are significant in showing the tremendous force and persistency of the first of a series of habits. Of course humans are not pigs but fundamentally the methods of learning for pigs and people are about the same. Therefore, these records suggest the gravity of the "first impression," and emphasize the importance of correct reactions at the outset in any kind of learning. It is to be regretted that the time for each reaction was not recorded and that some device was not provided whereby the exact part of K over which the pig jumped in his search or the food could be determined, and that the experiment was not continued for a much longer time. While these results are not at all conclusive they probably suffice to warrant further study along the same lines on small children as well as on animals. THE MARRIAGE-FLIGHT OF A BULL-DOG ANT (MYRMECIA SANG UINEA F. SMITH) WILLIAM MORTON WHEELER During a recent visit to Australia I had an opportunity to study in the field the habits of a number of species of the large ants popularly known as ' bull-dog ants," belonging to the very primitive genus Myrmecia. This genus comprises some sixty described species, subspecies and varieties, all confined to Australia and the island of New Caledonia. They bear about the same relation to the more specialized and more modern ants that the Marsupials of Australia bear to the placental mammals of other continents. Many of the species are among the largest of ants, several of them are beautifully colored and sculptured, and all of them sting or both sting and bite severely. In his latest work on the Ponerinae in the "Genera Insectorum" Emery divides the genus Myrmecia into three subgenera, Myrme- cia sens, str., Pristomyrmecia and Promyrmecia. I have found that the species of these three groups differ greatly in nesting and other habits. A detailed account of my observations will be published later in connection with taxonomic descriptions of the various forms. Here my remarks will be mainly confined to one of the largest and most aggressive species, Myrmecia sanguinea F. Smith, an ant which must be widely distributed in Australia, since it has been recorded from Queensland, New South Wales, South and Western Australia and Tasmania. I found it very abundant at Salisbury Court, near Uralla in northern New South Wales, where, through the courtesy of the government entomologist, Mr. W. W. Froggatt, I was able to spend a few days at a laboratory recently established for the study of the blow-flies that have acquired the unfortunate habit of injuring living sheep. In the open forest surrounding the laboratory there are many large nests, of sanguinea, mounds of earth of a more or less conical shape, varying from three to five feet in diameter at the base and one to two feet in height. The surfaces of these THE MARRIAGE-FLIGHT OF A BULL-DOG ANT 71 mounds are covered with a thin layer of bits of charred wood, dead leaves, twigs and pebbles collected by the ants. The entrance to the nest is a rather small hole, scarcely more than half or three quarters of an inch in diameter and usually situ- ated on one of the slopes near but not at the summit. Each nest contains about 200 to 500 workers. These are subopaque, rich brownish red in color, with the gaster shining black, and vary considerably in size, from 12 to 28 mm. The females measure 26 to 28 mm., the males 18 to 20 mm. The smallest workers, not exceeding 12 to 15 mm. in length, function as door-keepers and are usually found stationed just within the nest-entrance, with their long, scissor-like mandibles directed outward. When the nest is disturbed these small workers are the first to sally forth, followed by others of the same or medium size, and it is only somewhat later that the huge and formidable individuals, measuring 25 to 28 mm. advance to the attack with wide open jaws and threatening sting. This behavior is the reverse of that observed in other ants (e. g. in species of Cam- ponotus, Colobopsis, Atta, Pheidole, etc.), the major workers or soldiers of which act as door-keepers and are among the first to rush to the defence of the colony. Unlike the puny, small-eyed ants of our northern latitudes, the bull-dogs can clearly discern objects at a distance of several feet with their great, prominent, abundantly facetted eyes. They lose no time in running about in all directions hunting for the intruder, but with ferocious unanimity make directly for him. The extraordinary tenacity with which they hold on with their mandibles to any moving object that comes in their path is, of course, responsible for their popular name, but no one has been able to suggest a polite epithet that will do justice to the virulence of their sting. At the time of my visit to Salisbury Court, during the last week of November, there were no winged males or females in the nests, though there were plenty of larvae and a small number of worker pupae. This is rather surprising because the sexual forms of most of the ants of New South Wales are to be found in the nests during late October and early November. Mr. Froggatt expressed the opinion that the males and females of the various species of Myrmecia do not mature till January. This opinion has been confirmed in a letter recently received, in which he describes a remarkable marriage flight of sanguined 72 WILLIAM MORTON WHEELER in a locality very near the one in which I studied the nests. Fully to appreciate his description, the reader must bear in mind what I have said about the size of the females of this ant. They, of course, have powerful stings, like the workers, though the males are stingless and have feeble mandibles. Mr. Froggatt's account runs as follows: "On January 30th, after some very hot, stormy weather, while I was at Chevy Chase, near Armidale, N. S. W., I crossed the paddock and climbed to the top of Mt. Roul, an isolated, flat-topped, basaltic hill, which rises about 300 feet above the surrounding open, cleared country. The summit, about half an acre in extent, is covered with low "black-thorn" bushes {Busaria spinifera). I saw no signs of bull-dog ant nests till I reached the summit. Then I was enveloped in a regular cloud of the great winged ants. They were out in thousands and thousands, resting on the rocks and grass. The air was full of them, but they were chiefly flying in great numbers about the bushes where the males were copulating with the females. As soon as a male (and there were apparently hundreds of males to every female) captured a female on a bush, other males surrounded the couple till there was a struggling mass of ants forming a ball as big as one's fist. Then something seemed to give way, the ball would fall to the ground and the ants would scatter. As many as half a dozen of these balls would keep forming on every little bush and this went on throughout the morning. I was a bit frightened at first but the ants took no notice of me, as the males were all so eager in their endeavors to seize the females." Except for the great size of the participants, this nuptial flight presents an exact picture of occasional flights of some of our common Myrmicine ants, especially of Myrmica scabrinodis Nyl. That such enormous swarms of Myrmecia as the one described must be of rare occurrence, is evident from the state- ment of such a keen observer as Mr. Froggatt that he has "never before seen more than a dozen winged bull-dog ants of any species together." I find, however, a brief description by Tepper 1 of what must have been a very similar scene. He describes a nuptial flight of one of the large species of Myrmecia 1 Observations about the Habits of Some South Australian Ants. Trans. & Proc. Roy. Soc. S. Austr., 5, 1882, pp. 24-26, 106-107. THE MARRIAGE-FLIGHT OF A BULL-DOG ANT 73 (probably sanguined or pyriformis), early in April in South Australia, as "rather a formidable affair, owing to many hun- dreds of the large creatures (the female above an inch in length while alive) flitting about one's head, all armed with a sting about a quarter of an inch in length, while the shrubs near the nest were covered with scores of pairs and single ones." The observations of Tepper and Froggatt prove conclusively that the species of the Ponerine genus Myrmecia celebrate a regular marriage flight like all the ants of the other taxonomic subfamilies, except the species with wingless males or females, and that these flights occur during January in northern New South Wales or a few months later in the more southern and colder portions of Australia. This season corresponds, of course, to our autumn months, which are likewise the nuptial season of some of our species of Lasius (L. claviger Roger, brevicornis Emery, etc.) We may also infer from the accounts of the two Australian observers that each female Myrmecia, after fecunda- tion, loses her wings in the same manner as other ants, except certain parasitic species, enters the ground and establishes a small colony without the assistance of workers of her own species. I am able to show that this is actually the case. On September 19, 1914, I found under a stone in one of the deep sandstone canyons near Katoomba, in the Blue Mts. of New South Wales, a fine dealated female of Myrmecia tricolor Mayr occupying a little cavity in the soil and engaged in caring for about a dozen small larvae. This little incipient colony was very similar to those just established by our common car- penter ants (Camponotus pennsylvanicus DeGeer and novebora- censis Fitch) under the bark of old logs. That the most primi- tive of existing ants should found their colonies in precisely the same manner as the most highly specialized species, is not without interest. JEAN-HENRI FABRE William Morton Wheeler Through the death of Jean-Henri Fabre on October 11th, 1915, the world has lost its greatest entomologist, a man who combined in an extraordinary degree the gifts of a virile and penetrating observer and those of a literary artist of high dis- tinction. During the greater portion of a life of poverty, ex- tending over a period of 92 years, he ceaselessly devoted himself to an intensive study of insect behavior and to the recording of his observations in such fascinating language that Victor Hugo styled him the " Homer of the insects." Like the life of his countryman Latreille, who preceded him as the " prince of entomologists," Fabre's life was uneventful. His biography has been written by a sympathetic admirer, C. V. Legros, and rendered into English by another admirer, Bernard Miall, but from many passages scattered through Fabre's great work', the " Souvenirs Entomologiques " it is possible to glean an even more illuminating and intimate knowl- edge of his powerful individuality and of his methods of working and thinking. He was born of humble peasant parents on December 22nd, 1823, in the hamlet of Saint Leon, in the part of the Provence known as the Haute-Rouergue. Through dili- gent application to the classics, physics, chemistry and mathe- matics in the rather mediaeval schools of his day he prepared himself to become a teacher. At 19 he entered on this profes- sion in the College of Carpentras and in 1850 accepted a posi- tion as professor in the lycee of Ajaccio, Corsica, at a salary of £72. Here he met the naturalist Moquin-Tandon, who seems to have had an important influence in determining his career as a biological investigator. Even as a boy, however, Fabre had been greatly interested in insects, so that Moquin-Tandon probably only helped to reveal to him his innate aptitude for observation and experimentation. He realized that he had a genius for observing small animals, and from that time forth, JEAN-HENRI FABRE 75 like Socrates, he implicitly obeyed the voice of his daemon almost to the hour of his death. Falling ill with malaria at Ajaccio he was compelled to return to France, and in 1853 was appointed assistant professor of physics at the lycee of Avignon. This post he held for nearly 20 years (till 1871), without advancement ^ and with a salary not exceeding £64! During this period he made some of his most important obser- vations. The written accounts of his work, contributed to the " Annales des Sciences Naturelles ' include a study of the habits of the solitary wasp Cerceris and of the cause of the long conservation of the beetles on which it feeds (1855), notes on the life-history of Cerceris, Bembex and Sitaris (1856), fol- lowed by his classic memoir on the hypermetamorphosis and habits of Sitaris (1858) and studies on the role of adipose tissue in the urinary secretion of insects (1862). In 1871 he left the lycee of Avignon to devote the remainder of his life to the study of instinct in insects. He moved to Serignan, a hamlet near Orange, not far from Avignon. Here he lived henceforth and worked as a hermit and here he died, in a little cottage on a plot of ground called the " harmas," a beautiful description of which is given in the opening chapter of the second series of " Souvenirs." 1 During the early years of his residence at Serignan he was compelled to devote much time to writing text-books on natural history subjects for the purpose of keeping the wolf from the door. It will probably be found that these little books were the forerunners of the modern " nature books." He also continued to contribute scien- tific articles to the " Annales des Sciences." Two of these, on the habits and parthenogenesis of some bees of the genus Halic- tus (1880) and one on the repartition of the sexes in the Hymen- optera (1884) are of unusual interest. But the great work accomplished at Serignan is embodied in the ten volumes of the " Souvenirs." Anything like an adequate review of this monumental work would require much time and labor. Only the entomologist who has endeavored to work out complicated insect life-histories will fully appreciate Fabre's powers as an observer and will not be greatly surprised to learn that during the course of years he wore a groove in the stone floor of his translated by A. T. de Mattos in the "Life of the Fly," Dodd, Mead & Co. 1913. 76 JEAN-HENRI FABRE laboratory by walking around his table. The complete eluci- dation of some of the life-histories, like that of the sacred scara- baeus, required observations extending over a period of nearly forty years. The newspapers and magazines have made us familiar with the romance of the closing years of Fatjre's life. Mistral, the Provencal poet, Maeterlinck and Rostand are said to have dis- covered Fabre and to have called the attention of the world to his destitute condition. According to a well-known French magazine, 'In 1910 he was revealed to the people; a group of litterateurs and savants conceived the idea of offering this modest, almost unknown man a plaque to perpetuate the memory of his work. Two years later his ninetieth birthday was cele- brated by a ceremony at which the Institut was represented, and somewhat later the President of the Republic paid him a visit." ' Why," asks Le Gros at the close of his account of the celebration of 1910, " at this jubilee of the greatest of ento- mologists, was not a single appointed representative of ento- mology present ? " And he goes on to say: ' The fact is that the majority of those who 'amid the living seek only for corpses', according to the expression of Bacon, unwilling to see in Fabre anything more than an imaginative writer, and being themselves incapable of understanding the beautiful and of dis- tinguishing it in the true, reproached him, perhaps with more jealousy than conviction, with having introduced literature into the domains of science." This is an unfair statement of the case. Fabre has long been known to naturalists and especially to entomologists and many of them, from Darwin to the Peck- hams and Forel have referred to his work in terms of the greatest admiration. It is only the litterati and general public who have just discovered Fabre, and it is not difficult to account for this belated appreciation. Insects are so peculiarly organized and offer to the casual observer so few points of contact with the general trend of human interests that even the magic style of a Fabre failed to elicit a widespread desire to know about their activities. But when a great writer like Maeterlinck announced that, " Henri Fabre is one of the greatest and purest glories in the present possession of the civilized world, one of the most erudite naturalists and the most marvellous poet in a modern and truly legitimate sense of the word," and added that Fabre JEAN-HENRI FABRE 77 was one of the profoundest admirations of his life, people who had never devoted five minutes of their lives to studying an insect, began to" sit up and take notice. Whenever there is complaint of the neglect of a genius by the world, it is well to scrutinize the behavior of the genius. If we do this in Fabre's case we shall have little difficulty in ac- counting for the neglect from which he suffered, both on the part of the general public and the scientific fraternity. Fabre and his publisher are undoubtedly responsible for much of the popular neglect. The title of the great work, " Souvenirs Ento- mologiques," is inept, to say the least; the ten volumes were unattractively printed and inadequately illustrated and the arrangement of the articles in the series might have been much better. All rights of translation, even of extracts, were, moreover, rigidly withheld till very recently. When Fabre be- came more generally known these faults were corrected by the publication of selected essays in more pleasing volumes and under more appropriate titles, such as "La Vie des Insectes " and ' Moeurs des Insectes," an arrangement which has been followed in the English translations now displayed in all our bookstores. The reserved and unsympathetic attitude of entomologists towards Fabre was very largely due to the fact that he was a crotchety and opinionated recluse, who seems never to have made the slightest attempt to enter into friendly personal or epistolary relations with other entomologists, who never men- tioned and probably never read the work of his contempora- ries, who lost no opportunity of holding up to ridicule some of the most important entomological studies, such as insect tax- onomy, and who repeated investigations that had been made by others, without intimating and evidently without knowing that such investigations had long been known to the ento- mological world. He failed to realize that entomologists are more human than the objects of their studies and that " wer Liebe ernten will, muss Liebe saen." It is not surprising, there- fore that there was no afflux of entomologists to Serignan to celebrate his jubilee. While their presence might have been commendable, we can hardly blame them, under the circum- stances, for staying at home. Another reason for the attitude of entomologists towards 78 JEAN-HENRI FABRE Fabre is to be found in his peculiar views concerning instinct, views that were prevalent enough in the early part of the nine- teenth century but are singularly foreign to the psychology and theoretical biology of the present day. He not only declined to accept the doctrine of evolution but vigorously attacked it in more than one of his essays, although many of his criticisms so far overshoot the mark that one reads them with amazement. It must be remembered, of course, that Fabre was nearly 36 years old when the ' Origin of Species " appeared. He was not, therefore, like the naturalists of the present generation, suckled, so to speak, at the breasts of evolutionary doctrine, and his life-long lack of contact with biological speculation kept him from viewing the phenomena of instinct from a genetic stand- point. But even in his chosen field, the study of instinct, he confined himself to a comparatively circumscribed group of phenomena. He worked only on a series of insects selected from his immediate environment and for certain peculiar reasons took little interest in the social species, (ants, social bees and wasps) which are, nevertheless, abundant in southern France. His studies on these forms are limited to an essay on the amazon ant (Polyergus -rujescens) and a few essays on wasps (Vespa vulgaris). He seems to have borne a grudge against the ants because they so often entered his breeding cages and killed the insects with which he was experimenting. There were also other and more weighty reasons for this neglect of the very insects which naturally suggest a genetic interpretation of instinct. Fabre believed that instinct manifests itself in its purest form in the solitary species. He was, moreover, greatly impressed by its fixity and mechanical aspect, and his rigid training in physics, chemistry and mathematics and his keen analytical ability prob- ably biased him in favor of views which have grown more and more repugnant to modern biologists. He had a strong tendency to schematize his observations and to ignore the variability of instinct. This tendency has been pointed out by several obser- vers and is most clearly marked in his classic work on the solitary wasps. Although Fabre suffered in the estimation of many biologists on account of the theoretical views which he elaborated and very stubbornly upheld through the course of a long life in voluntary isolation from the great current of biological thought, JEAN-HENRI FABRE 79 we are bound to confess that in some very important matters his vision was clearer than that of his contemporaries. He alone realized the great significance of the study of animal behavior at a time when other biologists were absorbed in purely morpho- logical work. No better proof of this statement can be given than the concluding paragraphs of the first essay in the second series of the " Souvenirs " published in 1882. I quote from de Mattos' translation (pp. 2.6 and 27) : ' Laboratories are being founded, at great expense, on our Atlantic and Mediterranean coasts, where people cut up small sea-animals, of but meagre interest to us; they spend a fortune on powerful microscopes, delicate dissecting instruments, engines of capture, boats, fishing crews, aquariums, to find out how the yolk of an Annelid's egg is constructed, a question whereof I have never yet been able to grasp the full importance; and they scorn the little land- animal, which lives in constant touch with us, which provides universal psychology with documents of inestimable value, which too often threatens the public wealth by destroying our crops. When shall we have an entomological laboratory for the study not of the dead insect, steeped in alcohol, but of the living insect; a laboratory having for its object the instinct, the habits, the manner of living, the work, the struggles, the propagation of that little world, with which agriculture and philosophy have most seriously to reckon ? " To know thoroughly the history of the destroyer of our vines might perhaps be more important than to know how this or that nerve-fibre of a Cirriped ends ; to establish by experiment the line of demarcation between intellect and instinct; to prove, by comparing facts in the zoological progression, whether human reason be an irreducible faculty or not; all this ought surely to take precedence of the number of joints in a Crustacean's an- tenna. These enormous questions would need an army of work- ers, and we have not one. The fashion is all for the Mollusc and the Zoophytes. The depths of the sea are explored with many drag-nets; the soil which we tread is consistently dis- regarded. While waiting for the fashion to change, I open my harmas laboratory of living entomology; and this laboratory shall not cost the ratepayers a farthing." Not only was Fabre the first to realize the full importance of a scientific study of animal behavior but he was the first 80 JEAN-HENRI FABRE consistently to apply the experimental method to the investi- gation of the animal mind. The " Souvenirs" abound in ac- counts of experiments, performed for the purpose of elucidating the nature of instinct, not the less illuminating and conclusive because they were carried out with crude, home-made appa- ratus. It is as instructive as it is humiliating to read his results and to reflect on the mountains of complicated apparatus in our modern laboratories and the ridiculous mice in the form of results which only too frequently issue from the travail of " research." Another valuable service of Fabre consisted in his calling attention to the fact that the applications of zoology to human welfare must be based on an accurate knowledge of animal behavior. This has been tacitly assumed by economic ento- mologists, but neither they nor the modern behaviorists have sufficiently emphasized the fact that we cannot hope to control animal depredations or to compel animals to contribute to our well-being and wealth without an exhaustive knowledge of such apparently remote phenomena as the animal sensations, reac- tions and instincts. Fabre is quite explicit in this matter. Though he lived as a recluse from the scientific world, he re- mained in intimate contact with the life of the Provencal peas- ant and had no doubts concerning the important bearing of his own work on such a fundamental industry as agriculture. Financial Statement for the Journal of Animal Behavior, December 2, 1914 to December 1, 1915 (Volume 5) receipts Balance from 1914 $191.88 Receipts from sales of complete volumes and odd numbers 719.75 Receipts from advertising 25.00 Gifts and contributions toward the cost of illustrations and tabular material 292.00 Interest 19.41 From Behavior Monographs for office ex- penses, etc 75.51 $1,323.55 expenditures Cost of manufacturing and distributing vol- ume 5 . .. $1,111.41 Office expenses, including postal and express items 208.00 1,319.41 Balance on hand $4.14 JOURNAL OF ANIMAL BEHAVIOR Vol. 6 MARCH-APRIL No. 2 THE REACTIONS OF NECTURUS TO STIMULI RECEIVED THROUGH THE SKIN MARY HONORA SAYLE University of Wisconsin CONTENTS PAGE Introduction.- 81 I. The Reactions to Touch 82 II. The Reactions to Chemicals 84 (a) Acids 84 (b) Alkalies and Alkaline Earth Salts 91 (c) Alcohol 94 (d) Clove Oil and Turpentine 96 (e) Summary 96 III. The Reactions to Heat 96 IV. The Reactions to Light 98 V. Discussion 100 VI. Summary 101 Bibliography 101 INTRODUCTION The mud puppy, Necturus maculosus (Rafinesque), of the fresh waters of North America has long been of interest on account of its primitive structure and its systematic position near the base of the amphibian line, but its reactions have been little studied. For over two centuries it has been known that fishes possess various highly specialized sense organs in the skin, and later work along this line has been done by Herrick ('03), Parker ('09, '10), Sheldon ('09), and Reese ('12). During recent years the reactions of amphibians have been studied by Torelle ('03), Parker ('03), Cole ('07), and others, but upon Necturus very little has yet appeared. Reese ('06), and Pearse ('10) have worked upon the reactions of Necturus to light and heat. Reese 82 MARY HONORA SAYLE found that though all parts of the body could be stimulated by white light, the head was the most sensitive region. He also worked with light passed through "red" and "blue" glass, and found that slow responses were thus induced. In his heat experiments he found Necturus to be sensitive to considerable changes in temperature. Pearse showed that Necturus is nega- tively phototropic and that it comes to rest in shaded areas. Both the skin and eyes act as photoreceptors, and the stimula- tion of either brings about negative reactions. The present study endeavors to ascertain the sensitiveness of the general body surface of Necturus to touch, chemicals, heat, and light. The experiments were performed on four individuals of aver- age size, though only three of them were used for the light and heat experiments. Just before each experiment the individual to be used was transferred from the aquarium tank into a zinc tray, three by one and one-half feet. The transfer was made by grasping each animal just behind the front legs. The method of measuring movement employed in the experiments was a rather simple one. A coarse wire screen that covered the tray was divided off with twine into inch squares. It was possible in this way to observe and at the same time measure the move- ments of Necturus beneath the screen. When not being used for the experiments the animals were kept in running water in an aquarium tank, six by three feet, in the vivarium of the Biology Building, University of Wis- consin. The tank was divided into four compartments, one for each animal. Bricks and strips of wood were used to make suitable dark nooks like the natural haunts of Necturus. This work was accomplished under the direction of Professor A. S. Pearse, for whose helpful suggestions and encouragement it gives me great pleasure to express my appreciation. REACTIONS TO TOUCH The whole outer surface of Necturus is open to stimulation from a deforming pressure, that is, it is sensitive to touch. Animals were tested for touch sensitiveness with wood, glass, and a soft brush. A blunt piece of wood the size of an ordinary pencil was touched gently to various regions of the body. When any REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 83 portion was stimulated, there was a forward movement of two or three inches. The nostrils, gills, and tail were the most sen- sitive regions. Stimulation with a glass rod induced similar reactions as were noted for wood. Stimulation with a soft brush, however, induced no locomotion. The tail and feet were drawn close to the body when stimulated; the gills closed, and when the nostrils were touched with the brush, there was at first a backward jerk, but further stimulation produced no reaction. In stimulating the ventral side of the body, the mouth region and the tip of the tail were found to be most sensitive. When either of these regions was touched, locomotion followed, but the response was slow. The gills and feet upon stimulation were drawn close to the body. No other regions of the ventral side were sensitive. Experiments were next performed to ascertain if the nerve endings fatigued by chemical stimuli would react to tactile stim- uli. Necturus was stimulated over its entire body with ten per cent hydrochloric acid until it ceased to respond. A blunt point was then used to induce tactile response, but very little reaction followed. The same experiment was tried after sul- phuric and nitric acid of the same strength. It was found that there was more response to tactile stimulation after Necturus was fatigued by hydrochloric than after sulphuric or nitric acids. In performing the touch experiments several points worthy of note were observed: 1. Necturus is nowhere as sensitive to a blunt as to a sharp instrument. 2. When the mouth region was stimulated the head was jerked backward. 3. The regions of greatest sensitiveness are the nostril region, gills and tail. *4. When the gills were touched their movement ceased, but when the wood, glass or brush was removed their motion went on at an increased rate. 5. When the tail was stimulated it was curled up. It was most sensitive at the tip. 6. When the hind legs were stimulated they were drawn back- ward under the body, while the front legs started to walk. 7. The dorsal region between the legs was the least sensitive of all. 84 MARY HONORA SAYLE REACTIONS TO CHEMICALS For testing the effect of chemicals on the skin, the surface of the body of Necturus was divided into seven regions, selected mainly for the different responses that result from their stimu- lation. They were the regions of the nostrils, head, gills, pectoral girdle, pelvic girdle, back, and tail. With each solution used all the regions were tested with each concentration, and the time given for reaction was always the same, fifteen seconds. If no response was made during the period, it was recorded as no action. The location of the regions are shown in figure 1. i;.i..ls H««J iu» e "&Xl Fig. 1. Outline of Necturus maculosus, showing the regions stimulated in the experiments The substances used in the work were hydrochloric, sulphuric, nitric and acetic acids; potassium hydroxide, ammonium hydrate, magnesium chloride, magnesium sulphate, alcohol and clove oil. In the experimental work all of these solutions were gradually diluted until the limit of reaction was reached. Sufficient time was given between tests at different degrees of concentration and with different substances to eliminate after effects. The solutions were applied by means of a pipette and were ejected slowly with the tip of the pipette held about two milli- meters from the skin. In all cases the animal was completely covered with water during the experiment. (a) The Reactions to Acids In tables I to XI are given the reactions of Necturus to the four acids used. Necturus responds to acid stimuli by locomo- tion or by a movement of the regions stimulated. If the acid is strong, locomotion takes place; if it is weak, local response REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 85 is induced. Necturus is most sensitive to nitric acid at all concentrations and hydrochloric gives the weakest response, except acetic. The reactions were found to become weaker and fewer as the acid was diluted. The gills were in nearly all cases the most sensitive region, with the nostrils, head, and tail following in order. Stimulation of the nostril, with a ten per cent solution usually produced an expulsion of air from the nostrils and mouth in addition to locomotion. The limit of reaction for hydrochloric, sulphuric and nitric acids was 0.015 and for acetic 0.01 per cent. TABLE I* Reactions to 10% Hydrochloric Acid Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I B9 B F 14 X F15 F3 Squirm Squirm Switch tail F9 II B F 14 B F 15 X F 14 Squirm Squirm Squirm Switch tail F 15 III F 15 B7 F 14 F3 Squirm Squirm Switch tail IV B F 7 B7 F 8 F 13 F7 Switch tail F17 Average 11.2 12.7 14 3.2 10.2 * Explanation of the Tables. — For convenience, several signs have been employed in the tables. They indicate as follows: F, forward; B, backward; X, animal turned around; O, no reaction. A letter not followed by a number indicates a very slight movement, such as a jerk. The numbers I, II, III and IV indicate the four individuals under experiment. The reactions are noted for seven regions of the body, as indicated at the top of each table, making a total of twenty-eight reactions of four individuals in each concentration. The reactions are measured as inches of locomotion, as described elsewhere, and the average is for total move- ment, backward and forward, for each region stimulated. 86 MARY HONORA SAYLE TABLE II Reactions to 0.1% Hydrochloric Acid Individual No. Nostrils Head Gills Tail I B 1 B2 F 14 II B4 B2 F 14 F6 III B3 B2 F 12 IV F12 Switch tail Average 2 1.5 13 1.5 TABLE III Reactions to 0.09% Hydrochloric Acid Individual No. Nostrils Head Gills Tail I F2 Head moved Head moved II Bi Head moved III F3 Head moved F4 IV F4 F3 F3 Curled tail Average 2.2 .75 1.7 REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 87 TABLE IV Reactions to 10% Sulphuric Acid Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I F24 B13 F24 F 13 Switch body Switch body Switch tail F10 II B2 B F 14 Moved in place F14 Switch body F7 X Tail curled III F 14 F4 F24 F14 Switch body F14 F14 IV X F 15 B15 F 14 F10 F10 F7 F12 Average 13.7 14 15.5 12.7 4.2 6.6 9 TABLE V Reactions to 0.1% Sulphuric Acid Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I B5 B4 F6 F7 F4 F3 F5 II B3 F2 Head moved B2 Moved in place III B F8 Moved in place F7 F6 F4 F4 IV B F7 F7 X F7 F5 F5 Average 5.7 4.7 5.2 4.7 3 3.5 88 MARY HONORA SAYLE TABLE VI Reactions to 0.09% Sulphuric Acid Individual No. Nostrils Head Gills I - B F3 F3 II B 1 III B F2 F2 IV B2 B2 Average .75 1.2 1.7 TABLE VII Reactions to 10% Nitric Acid Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I B4 F 10 F22 F16 F 14 F 14 Switch tail F 14 II B7 F8 F 14 F14 F8 F12 Tail curled III B5 F20 F 14 F15 F 10 F 15 F 14 IV B5 F 15 F22 F10 F8 F10 F 14 Average 5.2 13.2 18 13.7 10 12.7 10.5 REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 89 TABLE VIII Reactions to 0.1% Nitric Acid Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I B4 F6 F4 F2 0* F3 II B4 F4 F7 F2 Fl F5 F6 III B3 F 10 F6 F 10 F5 F3 F3 F4 IV Bl F6 F 10 F3 F3 F4 F2 F4 Average 8.5 6 5 3 2 2.5 4.2 TABLE IX Reactions to 0.09% Nitric Acid Individual No. Nostrils Head Gills I B2 Moved in place F5 II B2 Moved in place F6 III B3 F4 IV Bl Moved in place F2 Average 2 4.2 90 MARY HONORA SAYLE TABLE X Reactions to 1% Acetic Acid Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I B3 F5 Bl F6 II B3 F6 F3 F6 F3 F 1 Tail curled III B2 F 12 B2 F6 B6 B6 F3 IV B F 14 F 10 F22 F10 F4 Tail curled Average 8.7 5.5 10 4.7 2 TABLE XI Reactions to 0.05% Acetic Acid Individual No. Nostrils Head Gills I Bl Moved head F2 11 B2 F3 Fl Closed gills III Bl F2 Fl IV Bl Fl F2 Average 2 1 1.2 REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 91 (b) The Reactions to Alkalies and Alkaline Earth Salts The reactions of Necturus to alkalies are given in tables XII-XV. Potassium hydroxide and ammonium hydrate did not produce as definite responses as did the acids. The different regions of the body varied in sensitiveness, but the gills were always the most sensitive region. The reactions of the gills to one per cent potassium hydroxide were greater than to ten per cent nitric acid, but the other regions of the body were less sensitive to potassium hydroxide or ammonium hydrate. The limit of reaction for both was a 0.01 per cent solution. It is apparent from inspecting tables XVI-XIX that Necturus is more sensitive to magnesium sulphate than to magnesium chloride. The nostril region was the most sensitive to magne- sium sulphate. At 0.05 per cent magnesium chloride the nostril region only was sensitive, while at the same per cent Necturus was sensitive to magnesium sulphate in the regions of the nos- trils, head and gills. Necturus will not react to either chemical at 0.01 per cent. TABLE XII Reactions to 1% Potassium Hydroxide Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I Switch body F 14 B F7 Switch body F22 Switch bodv F12 Switch body F14 Switch body F14 Switch tail II F 12 B F7 Switch body F22 F7 F9 F9 Switch tail III B F 14 B F 12 Switch body F22 Switch body F12 F9 F12 F12 IV F13 B F8 Switch body F 14 Switch body F10 F8 F7 Switch tail Average 13.5 8.5 20 10.2 10 10.5 3 92 MARY HONOR A SAYLE TABLE XIII Reactions to 0.05% Potassium Hydroxide Individual No. Nostrils Head Gills I B2 F4 Fl F4 II F 12 Moved in place F2 III F 10 Head moved Head moved IV F6 F5 F7 Average 8.5 1.5 3.2 TABLE XIV Reactions to 1% Ammonium Hydrate Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I X B6 F4 F20 F4 F6 F7 F5 II F9 B2 F6 F5 F2 F10 F10 III F 12 B F 10 F8 F12 F7 F5 F5 IV F 10 X F 12 F12 F7 F8 F10 Average 7.7 5.5 16.5 8.2 5.5 7.5 7.5 REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 93 TABLE XV Reactions to 0.075% Ammonium Hydrate Individual No. Nostrils Head Gills I B2 X F2 II B2 F2 Fl III IV B2 F8 B3 X Average 3.5 1.2 .75 TABLE XVI Reactions to 2% Magnesium Chloride Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I F 14 F 10 F 22 F8 F8 F6 F7 II F9 F 14 F 10 F10 F5 F5 III B2 F6 F 12 F 18 F8 F3 F7 F10 1 IV F5 B5 B5 F2 F4 F8 Average 9 10.2 12.5 7.7 4.5 5.5 6.2 TABLE XVII Reactions to 0.05% Magnesium Chloride Individual No. Nostrils I F2 II B3 III Moved head IV F2 Average 1.7 94 MARY HONORA SAYLE TABLE XVIII Reactions to 2% Magnesium Sulphate Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I F22 F7 F6 F5 X F4 F3 II B F26 F8 F 10 FIO FIO F6 F6 III F 10 F 12 F22 X F9 F4 F5 IV B F14 F20 F 12 F18 F6 F7 X Average 18 11.7 12.5 8.2 6.2 5.2 3.5 TABLE XIX Reactions to 0.05% Magnesium Sulphate Individual No. Nostrils Head Gills I Bl F2 F2 F2 II B2 F2 Bl F2 F3 III Moved head Moved head Moved head IV ■ B 1 Bl • F2 Fl Average 2 2 1.5 (c) Reactions to Alcohol A study was made of the reactions of Necturus to alcohol at fifty per cent and ten per cent. Tables XX and XXI show these reactions. There was no reaction with one per cent. When nostrils were stimulated with either of the effective solu- tions there was a greater response than with any other region REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 95 of the body. The gills and head were less sensitive. All regions responded when stimulated with a ten per cent solution and all regions except the pelvic and back were sensitive to ten per cent. Necturus is less sensitive to alcohol than to any- other chemical used, as indicated by the stronger solution neces- sary to induce reaction. TABLE XX Reactions to 50% Alcohol Individual No. Nostrils Head Gills Pectoral Girdle Back Pelvic Girdle Tail I B4 F18 F 14 B2 F 12 F4 F2 F10 F3 II B6 F9 B4 F 12 F10 F3 F8 Tail curled III B3 F20 F 12 B3 F 12 F18 F7 F8 F2 IV Bl F 10 F 12 F 12 F10 F5 F7 F7 Average 15.5 11.7 14.2 10.5 4.2 8.2 3 TABLE XXI Reactions to 10% Alcohol Individual No. Nostrils Head Gills Pectoral Girdle Tail I F4 F3 F4 F2 F3 II B F 10 F8 F7 F5 F4 III IV F 12 F8 F10 F4 F7 Average 6.5 4.7 5.2 2.7 3.5 96 MARY HONOR A SAYLE (d) Reactions to Clove Oil and Turpentine Drops of clove oil were ejected upon the skin of Necturus with no results, except a slight response in the region of the nostrils. When Necturus, however, raised its body out of the water and thus carried some of the oil upon its skin, every portion was sensitive. Cotton was saturated with turpentine and applied to Nec- turus with no result, whether the animal was in or out of water. (e) Summary of the Chemical Reactions 1. Necturus is more sensitive to nitric than to hydrochloric or sulphuric acid. It is least sensitive to acetic. Responsive- ness, is therefore, roughly proportional to degree of electrolitic dissociation. 2. Necturus is more sensitive to potassium sulphate than to potassium chloride. 3. Necturus is less sensitive to alcohol than to the acids, alkalies, or alkaline earth salts, as indicated by the high per cent of alcohol necessary to induce reactions. 4. The skin of Necturus is sensitive to clove oil in the air but not when under water. Whether in or out of the water, Necturus does not react to turpentine. THE REACTIONS TO HEAT Several experiments were performed with three individuals to determine the thermic reactions of Necturus. When taken from the tank where the water was 10° C. and put into water at 30° C. the three animals made the most vio- lent struggles, beginning almost immediately and lasting until they were completely exhausted. The violence of these strug- gles was quite remarkable, though less pronounced in some cases than in others. A very short stay in the warm water was sufficient to completely exhaust all the individuals, so that they turned ventral side up; but when returned to water of moderate temperature they soon recovered. Removal from water at 10° C. to 25° C. induced locomotion also. The gills were spread out and moved rapidly. The nos- trils were often lifted above the water and air was expelled. At 18° C. and 15° C. the animals remained quiet, but the REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 97 gills were moved rapidly. -Reese ('06) in a similar experiment found two of the five Necturus under experiment gave loco- motor responses at 18° C. When put into water at 4° from 30°, the animals showed the same responses as noted by Reese ('06) — that is, marked activity, walking rapidly up and down the tank. The reactions of certain regions of the body of Necturus to hot water was tested. A pipette that ejected 5 C.C. of water at a time was used and tests were made in water at 10° C. When the gills were stimulated with water at 70° C. the animal walked forward thirty inches, switching the body from side to side. The head, nostrils and tail were somewhat less sensitive. Their stimulation with water at 70° C. resulted in a forward movement of from eight to twelve inches. Water of the same temperature ejected on the back, caused only two or three inches of movement. The same regions of the body were tested in a like manner with 30° C, with no response. Local heat stimulation was tested in another manner. A long rubber tube was run through the tank of water containing the animals. At one end of the tube was a funnel; the other end was open and hung outside of the tank. This arrangement allowed water to be poured in at the funnel, run through the tube and empty outside of the tank. To secure local response to heat, the tube was placed in contact with that region of Necturus to be stimulated. The tank water during this test was 0° C. and the water poured through the tube was 96° C. The tube was placed in contact with the body and the water was poured in at the funnel. When the nostrils, head and gills were stimulated the animal drew away from the tube im- mediately. This showed that direct response resulted from the heat radiating from the tube containing the heated water. Scarcely any response was induced from any other region of the body. Another experiment was performed in which one end of a large pan, three by one and one-half feet, containing water was placed over a lighted gas burner and the other end on a cake of ice. The temperature of the heated end was 40° C, the cold end was 3° C, and the middle 10° C. Three animals were placed in the pan. They walked up and down several times, and finally rested in the cold end. When in the hot end they 98 MARY HONORA SAYLE splashed and struggled, but several times remained in the heated portion and tried to get out of the pan. In many cases they stayed in the hot end for some time and burned themselves over and over again, before finally reaching the middle or cold end of the pan. From the experiments described above it is evident that Necturus is sensitive to considerable changes in temperature. It is apparent that Necturus could not live successfully in water above 18° C. and that life in water at 30° C. would prob- ably be impossible. THE REACTIONS TO LIGHT The experiments performed to determine the sensitiveness of the skin of Necturus to light were carried on in a dark room, the temperature of which ranged from 75°-70° F. A 144 candle- meter Nernst lamp was used in all the experiments. The first experiment tested the reactions of Necturus to light and shadow, the lamp being at the side one foot from the tank and a movable screen interposed in such a way that one-ha-lf of the tank was in shadow and the other half in light. Three animals were successively introduced. At 23.4 candle-meters, all three immediately went to the dark end of the tank. When the screen was changed to the opposite half of the tank the animals again moved into the shaded area. At 5.7 candle- meters each animal moved about for some time, equally in the light and dark areas of the tank. At 2.5 candle-meters the response was slow. After three minutes all three animals moved their heads and about one-half of their bodies into the shaded area. When the screen was changed to the opposite half of the tank the animals did not attempt to move again into the shaded area but remained in the light. With the intensity at 1 candle-meter the response was very slow. After five minutes all three moved their heads into the dark and remained in that position. Definite reactions in determining the relative sensitiveness of the body region of Necturus were brought about by illuminat- ing small areas of the skin. The apparatus was in the same position as for the experiments first described, except that the screen was arranged in such a manner that a band of light (144 candle-meters above the water at the center of the tank) REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 99 could be suddenly thrown on different regions of the body through a hole in the screen, three-eighths of an inch in diam- eter. Three individuals were used for this experiment and all of them behaved in essentially the same manner. After an animal had remained quiet in the dark for five minutes it was suddenly illuminated and a following behavior noted. The reactions observed agree with those of Pearse ('10, p. 169) who performed similar experiments with Necturus. When the light fell on the tail the animal moved forward, but when it was allowed to fall on the head the movement was usually backward. The various regions of the body exhibit great varia- tion in the time between stimulation and response. The reac- tion times of the regions subjected to the ray of light are shown in table XXII. TABLE XXII The Reaction Times in Seconds of Various Regions of the Body to Light, 144 C. M. Individual No. Head Pectoral Girdle Back Pelvic Girdle Tail I 19 90 162 159 153 II 22 130 203 180 120 III 15 180 240 120 120 Average 18.6 133.3 201.6 153 131 As Pearse ('10, p. 170) found the head most sensitive in normal Necturus while the tail was most sensitive in individuals with eyes removed, the greater sensitiveness of the head is due undoubtedly to stimulation received through the eye. When the eye only was stimulated by a ray of light (144 CM.) through a hole in the screen one-eighth of an inch in diameter, each animal moved away from the light in about thirty seconds, thus indicating the greater sensitiveness of the eye over any other region of the body. As Pearse ('10) says: "It is evident that Necturus is nega- tively phototropic and that it comes to rest in shaded areas. Both the eyes and skin are photoreceptors, and the stimulation of either brings about negative reaction." 100 MARY HONORA SAYLE DISCUSSION Let us try to picture Necturus in its daily life on the bottom of a lake or slow flowing stream. It frequents quiet waters from four to eight feet deep where a clean, sandy bottom is fairly well covered by vegetation. During the day it rests quietly beneath boards, logs or stones. One is rarely fortunate enough to get a glimpse of it; it is extremely shy and disappears at the slightest disturbance of the water, such as that caused by the approach of a boat. In walking, the diagonally opposite legs move in unison. When disturbed it swims swiftly, with a vigorous lateral motion of the broad and powerful tail, the feet being held closely against the body. Necturus never swims long distances, at most a few yards, then seeks concealment either in the mud or beneath some object. If it is undisturbed it usually rests with its head protruding from beneath a rock. The animal thus presents a curious appearance with its ruby gills moving gracefully to and fro. When one is disturbed the gills change from bright red to a dusky color and are at once drawn down tightly against the neck. During the night Necturus moves from place to place. If it wanders into a warm area it becomes restless and walks away quickly into cooler water. It may approach dead vegetation that is giving off acids and alcohol. Such an environment drives Necturus away immediately. If it swims into an area containing an excess of salts or alkalies it quickly withdraws. Thus Necturus moves about, being directed hither or thither by the various changes in the chemical constitution or temper- ature of the water, till it comes into cool, pure water. Here it seeks crustaceans, insect larvae, small fish and earthworms, which constitute its food. From the consideration of the facts of behavior one is natur- ally led to ask what the artificial conditions are which best suit the needs and instincts of Necturus. In aquaria it avoids the sunlight, and always seeks concealment. A tank well sup- plied with dark sheltered nooks is therefore best. Necturus can not live in warm water, and hence an aquarium should be sup- plied with running water and kept cool. Everything should be kept free from fungus, for a Necturus infected with Sapro- legnia will not live long. When Necturi are kept in aquaria they are frequently ob- REACTIONS OF NECTURUS TO STIMULI THROUGH SKIN 101 served to thrust their snouts above the water, open the mouth wide, and then return to the bottom where they soon expel the air, both through the gill-slits and from the mouth. It would thus seem that, while the branchiae are the chief means of respiration, the lungs play considerable part. The great tenacity of life showed by mud-puppies is a matter of frequent comment. They are able to live for months with- out food and may be easily revived after being left for three or four hours out of water. After severe mutilation they re- cover and regenerate lost parts. Notwithstanding this great vitality they seem to fall easy victims to Saprolegnia, especially if there is a slight abrasion on the skin. SUMMARY 1. The skin of Necturus is everywhere sensitive to tactile stimulation. The regions about the nostrils, gills and tail are most sensitive and the back the least sensitive. 2. Necturus is sensitive to chemical stimuli over the entire body surface, the reactions being characteristic for the different regions stimulated. The gills are usually the most sensitive region, with the. nostrils and head following in order. 3. Necturus is more sensitive to nitric than to hydrochloric or sulphuric acid. It is least sensitive to acetic acid. It is more sensitive to potassium sulphate than to potassium chloride. 4. When any region of the body is fatigued for a given chem- ical, it rarely responds to tactile stimuli, although it usually reacts to other kinds of chemical stimuli. 5. Necturus is sensitive to considerable changes in tempera- ture. All parts of the body are sensitive to hot water (70° C), but the gills, head and nostrils are the most sensitive regions. 6. Necturus is negatively phototropic and comes to rest in shaded areas. Both the eyes and skin are photoreceptors and the stimulation of either brings about negative reactions. The head and tail are the most sensitive regions. BIBLIOGRAPHY Copeland, M. The Olfactory Reactions of the Puffer or Swellfish, Speroides 1912. maculatus (Block and Schneider). Jour. Exper. Zooi, 12, 363-368. 1913. The Olfactory Reactions of the Spotted Newt, Diemyotylus viridescens (Rafinesque). Jour. Animal Behav., 3, 260-273. Eyci.eshymer, A. C. The Habits of Necturus maculosus. Atner. Nat., 40, 123- 1906. 136. 102 MARY HONORA SAYLE Herrick, C. J. On the Morphological and Physiological Classification of the 1903. Cutaneous Sense Organs of Fishes. Amer. Nat., 37, 313-318. Hurwitz, S. H. The Reactions of Earthworms to Acids. Proc. Amer. Acad. 1910. Arts and Set., 46, 67-81. Miller, W. S. The Lung, Vascular System, and Brain of Necturus maculatus. 1900. Bull. Univ. Wis., Sci. Ser., 2, 199-246. Pearse, A. S. The Reactions of Amphibians to Light. Proc. Amer. Acad. Arts 1910. and Sci., 45, 162-206. Parker, G. H. Influence of the Eyes, Ears, and other Allied Sense Organs on 1910. the Movements of the Dogfish, Muslelus canis (Mitchell). Bull. Bur. o) Fisheries, 29, 45-57. 1909. The Integumentary Nerves of Fishes as Photoreceptors and their Sig- nificance for the Oririn of the Vertebrate Eyes. Amer. Jour. Physiol., 25, 77-80. Reese, A. M. Food and Chemical Reactions of the Spotted Newt, Diemyctylus 1912. viridescens. Jour. Animal Behav., 2, 190-208. 1906. Observations on the Reactions of Cyptobranchus and Necturus to Light and Heat. Biol. Bull, 11, 93-99. Sheldon, R. E. The Reactions of the Dogfish to Chemical Stimuli. Jour. Comp. 1909. Neur. and Psych., 19, 273-311. EXPERIMENTS ON THE BEHAVIOR OF CHICKS HATCHED FROM ALCOHOLIZED EGGS JOHN MADISON FLETCHER, EDWINA ABBOTT COWAN AND ADA HART ARLITT From the Callender Laboratory of Psychology and Education of H. Sophie Newcomb Memorial College of Tulane University During 1912, in the Biological Laboratory of H. Sophie New- comb College of Tulane University, experiments were carried on relative to the effect of subjecting chicken eggs to alcohol- ization before hatching. It was discovered possible to hatch from eggs so treated chicks which were apparently of normal physical structure. When a technique had been evolved which attained this end the experiment was taken up and carried on by the Department of Psychology in an attempt to discover whether there was any characteristic modification of behavior which could be attributed to the effect of the alcohol on chicks hatched from alcoholized eggs. The experiments extended from October, 1913, to June, 1914, and from March, 1915, to May, 1915. In the course of the work five sets of eggs were hatched. The first, fourth and fifth hatchings were from mixed stock; the second was in part pure bred Buff Orpington and part pure bred White Leghorn and the third was of Rhode Island Red stock. The eggs numbered seventy in each hatching except the third, when only sixty eggs were used. Of each of the first four sets of eggs twenty-five were not tampered with. The remainder of the eggs, just before being placed in the electric incubator, were treated in the fol- lowing manner: A hole was made in the air chamber of the egg with a dissecting needle. Into the air chamber was injected, by means of a hypodermic needle, five drops of 95% ethyl alcohol. In some cases the air chamber was found to be too small to hold five drops so some eggs had less than five drops. None had more. After the alcohol was in, the hole was sealed with hot sealing wax. Of the last hatching of eggs ten were 104 FLETCHER, COWAN AND ARLITT left untampered with, ten were merely pierced with the dis- secting needle and sealed, twenty-five were injected with five drops of distilled water in place of the alcohol, and twenty-five were injected with alcohol as usual. The incubator was always divided into compartments which prevented the chicks from the different groups of eggs from mixing as they hatched. As the chicks were taken from the incubator they were marked with pigeon markers having numbers on them and from that time on were kept together and treated in exactly the same way except in the case of a few which were isolated as a check on the imitation factor. For purposes of convenience the chicks which were hatched from normal eggs will be designated " nor- mal;" those hatched from alcoholized eggs will be designated " alcohol;" those hatched from eggs with distilled water in them will be designated " water;" and those hatched from eggs with holes in them will be designated " holes." In observing the behavior of the chicks their reactions were divided into two groups, inherited and acquired. The inherited reactions studied were, reactions to light, pecking and drinking reactions, and reactions to height. The acquired reactions were those involved in choice between two visual stimuli presented by means of the Yerkes apparatus, as modified by Breed, 1 and those involved in learning to run mazes. Promptness and direc- tion of response to light, speed and accuracy of the pecking and drinking reactions, and the height from which a chick would jump were used as criteria of the types of behavior of the first group. The number of trials necessary before making ten correct choices and the time spent in the mazes were criteria of the second group of reactions. The instinctive reactions of alcohol chicks differed little, if at all, from those of normal chicks except in the case of reac- tions to height. After the fifteenth day the alcohol chicks jumped from greater heights than did the normal. Alcohol chicks were, as a group, slower in reacting within the mazes and had to be given more trials before they made ten correct choices in the Yerkes apparatus. The water chicks and the hole chicks were tested only for pecking reactions and mazes. Their behavior resembled much more nearlv that of the alcohol 1 Breed, F. S. The Development of Certain Instincts and Habits in Chicks. Behav. Mono?,., vol. 1, no. 1, 1911. CHICKS HATCHED FROM ALCOHOLIZED EGGS 105 chicks than that of the normal chicks. The different reactions will be taken up in detail in the following pages. We are happy to have this opportunity to express our lasting gratitude to President B. V. B. Dixon for hearty co-operation in the experiment and to Dr. Irving Hardesty and Dr. Gustav Mann, of the Medical College of Tulane University, to whose suggestions for technique we owe much. Our greatest debt is to Miss Ellinor H. Behre, of the Department of Biology, whose biological resourcefulness and unmeasured helpfulness in the carrying on of the work made it possible. To Professor J. A. Lyon, of the Department of Physics, we are indebted for the apparatus used in the experiments with reactions to light. To Professors Angell and Carr of Chicago University we are also indebted for valuable suggestions. EXPERIMENTS WITH INHERITED REACTIONS Reactions to Light The apparatus used to test the first reactions to light of the first set of chicks consisted of a hooded light placed behind a cardboard screen in such a way that a single horizontal strip of light 5 mm. in width was visible between the screen and the table, affording a single, highly localized stimulus. The experi- mental table stood in a dark room used for photographic pur- poses. No light was permitted to enter the incubator from the nineteenth to the twenty-fourth day of the hatch. On the twenty-fourth day the incubator was taken to the dark room. Fourteen chicks, seven normal and seven alcohol, (the extent of the hatch) were removed from the incubator one at a time, numbered and marked with cloth markers and placed one at a time on the experimental table eighteen inches from the stimulus. Of the seven normal chicks only one re- acted definitely. This chick walked towards the light, stopped, went on towards the light, stopped, then went straight to the light streak and tried to nestle in a corner of the screen. Four of the seven normal chicks walked away from the screen. Of the seven alcohol chicks, five went to sleep after making a few random movements, one stood still screaming and one ran about the table, finally running off and up the experimenter's arm. Light was thrown on the heads of the first two chicks which were tested but elicited no response. . After this test this apparatus was discarded because so much 106 FLETCHER, COWAN AND ARLITT light from the stimulus was reflected on the walls of the room. The following apparatus was found eminently satisfactory and was used throughout the rest of the series of experiments with light. It consisted of a one-candle power incandescent light connected in series with a rheostat and so arranged that by varying the resistance of the circuit the light could be adjusted from zero to full intensity. The stimulus was found to give best results when adjusted to about one-half full intensity. The apparatus was on a table covered by a cloth of dull black ma- terial having a minimum refractive power. The experiments were conducted at night. At the time of the second hatch all light was excluded from the incubator from the nineteenth to the twenty-fourth day. On the evening of the twenty-fourth day ten chicks, five normal and five alcohol, were removed from the incubator one at a time, numbered and marked with aluminum pigeon markers and placed on the table, eighteen inches from the stimulus. They were so placed that their heads were turned slightly away from the light. Three alcohol and two normal chicks reacted positively. Turning and facing the stimulus, they walked toward it. The other chicks walked diagonally or directly away from the light. Two normal and one alcohol chick walked off the table. On the following night the same procedure was repeated with the ten chicks previously tested and eleven others, four normal and seven alcohol, which had been too weak to be handled on the first night. Of the ten previously tested, eight reacted positively. No. 36, which had reacted slowly and Nos. 4, 6 and 35, which had not reacted at all, walked directly to the light as soon as placed on the table. Of the four normal chicks not tested before one reacted positively to the stimulus, walking directly to the light and placing its bill against it; the other three walked away from the light. Of the alcohol chicks not tested before three immediately moved toward the light, one after an interval of three minutes, and the other three walked around the table, going sometimes towards, sometimes away from the light. The twenty-one chicks were tested again on the third day. Normal chick No. 3 and alcohol No. 34, which had reacted positively on the first and second days, did not react on the third; normal No. 4 reacted on the second, but not the first and third days; alcohols No. 21 and No. 33 reacted on the CHICKS HATCHED FROM ALCOHOLIZED EGGS 107 third, but not on the second day; normal chicks Nos. 2, 6 and 10, and alcohol chicks Nos. 36, 22, 27 and 35 reacted positively- through the third day; normal chicks Nos. 11, 8, 7 and 1, and alcohol chicks Nos. 27, 26, 31, 28 and 24 did not react posi- tively at any time. The apparatus and procedure employed with the third hatch- ing of chicks was the same as that used with the second hatch- ing but these chicks were not removed from the incubator until the evening of the third day. At the end of the first day a partition was placed in the incubator in such a manner that the chickens hatching later would not mingle with those already hatched. Only those which had hatched the first day were tested. Of the six alcohol and eleven normal chicks tested, two alcohol chicks reacted positively, the other fifteen chicks did not react at all, wandering aimlessly about the table or stand- ing still and screaming. There is nothing in these results or in the behavior of the chicks, as they moved toward or away from the light stimulus, to indicate the presence of a phototropism either positive or negative. There seems to be a highly variable instinctive ten- dency to react positively to a single light stimulus of very weak intensity and small area. This tendency may appear for the first time on the first or second day and rarely on the third day. When it does appear before the third day it may not persist through that day. This reaction failed to show any characteristic difference between normal and alcohol chicks. Pecking Reaction The pecking reactions of the first set of chicks are not re- corded here as the cloth markers which distinguished normal from alcohol chicks dropped off the evening of the fourth day. These chicks were not used for further experiments. The following method was used with the second set of chicks. On the morning of the third day after hatching the chicks were placed, one at a time, on a table covered with a smooth, black cloth. On this cloth were scattered several bits of light colored grain. The time before the first peck and the accuracy of the pecking reaction was carefully noted. For the first test with each chick a record was kept during the first five minutes of its pecking activities. For succeeding tests the record was kept for three minutes. It was found convenient to use the Breed 108 FLETCHER, COWAX AND ARLITT method of recording reactions, i.e., to use "1" to denote peck- ing at but missing grain; "2", pecking at and hitting grain; "3", pecking, hitting, picking up and dropping grain; "4", peck- ing at, striking, picking up and swallowing grain, or the com- plete reaction. On the occasion of the first test the alcohol chicks with the exception of Nos. 22 and 29, pecked after longer intervals than did the normal chicks. Alcohols Nos. 28 and 21 did not peck at all. Alcohols Nos. 33, 39 and 34 went to sleep a number of times before they pecked at the grain. All of the alcohol chicks were quieter than the normal, made fewer random move- ments and showed little or none of the tendency to explore, so apparent in all of the normal chicks. They did, however, peck with accuracy equal to that of the normal chicks and their records show they improved with equal rapidity. Alcohols Nos. 33 and 34 made only 24% and 20% of "4" reactions respectively on the third day. On the eleventh day No. 34 made 51.1% and No. 33 made 44.4% of "4" reactions. One of the normal chicks, No. 10, made a better record than any made by alcohol chicks, making 78.9% of "4" reactions on the eleventh day; but the records of Nos. 8, 6 and 1 were not so good as those made by some of the alcohol chicks. No. 7 had 35.2%, No. 6, 44.4%' and No. 1, 47%, of "4" reactions on the eleventh day. The following table shows the comparative rate of progress of normal and alcoholic chicks of the second group. TABLE 1 Chicks 1, 2, 6 and 10 Were Normal Chicks. Chicks 21, 22, 29, 32, 34 and 35 Were Alcohol Chicks Day 3 Day 4 Reaction Reaction No. of Chick 12 3 4 12 3 4 1 1 2 4 2 2 6 8 5 2 10 3 3 2 Average 8 5 2 1/2 2 1/2 3 1/2 2 21 22 1 2 5 3 29 4 9 7 5 32 8 7 6 4 2 3 5 5 34 5 5 2 35 1 9 3 Average 2 3/5 4 4/5 6 2/5 3 2/5 2 3 5 5 CHICKS HATCHED FROM ALCOHOLIZED EGGS 109 TABLE 1- -Continued Day 5 Day 8 Reaction Reaction No. of Chick 1 2 3 4 1 2 3 4 1 1 1 13 1 2 4 4 11 1 13 11 6 5 5 4 1 1 1 10 3 4 12 Average 4 1/2 4 1/2 7 1/2 1/4 1 1/2 7 3/4 6 1/4 21 1 1 1 3 7 5 22 5 3 2 2 11 12 29 4 5 4 4 7 14 32 2 5 7 6 8 2 10 6 34 4 3 9 3 11 3 35 4 6 5 Average 3 1/5 3 2/5 4 3/5 3 1 1/3 1 1/2 8 2/3 7 1/2 Day 10 Day 1] L Reaction Reaction No. of Chick 1 2 3 4 1 2 3 4 1 1 11 8 1 8 8 2 5 18 5 11 6 3 4 12 4 6 8 10 6 21 4 15 Average 1 6 1/2 14 3/4 1 1/4 5 3/4 10 1/2 21 11 15 1 9 6 22 1 8 8 1 4 29 3 19 2 8 5 32 1 1 10 8 4 6 6 34 1 8 8 1 8 10 35 5 4 5 1 12 3 Average . No. of Chick 1 2 6 10 1/6 1 1/3 7 1/3 10 1/2 Time Until First Average Time Until First Peck 1' 20" 1' 30" 2' 10" 30" 1' 22 1/2" No. of Chick 21 22 29 32 34 35 Average 1 1/2 7 1/2 5 2/3 Time Until First Peck 15" 30" 2' 29 1 12" 5' 9' 50" 3' 36.9" The same apparatus and procedure was repeated with the third set of chicks with approximately the same results. The highest record in this group as in the second was made by a normal chick, the lowest by an alcohol, but the records of the other chicks differed little if at all. The following table (Table 110 FLETCHER, COWAN AND ARLITT 2) presents the rate of improvement of the pecking reaction of normal and alcohol chicks. TABLE 2 Chicks 2, 3, 5, 6, 8, 15 and 16 Were Normal Chicks. Chicks 21, 22, 24, 29, 31, 33 and 34 Were Alcohol Chicks Day 3 Reaction No. of Chick 12 3 4 2 12 3 6 5 3 10 5 6 5 2 8 13 6 8 15 16 5 9 6 5 Average....... 6 3/4 3 1/2 6 1/4 7 1/4 21 7 6 9 1 22 24 29 31....'..'.'.'..'.. 13 6 3 33 34 Average 3 1/2 9 1/2 7 1/2 2 ' * Day 5 Reaction No. of Chick 12 3 4 2 3 12 3 5 5 14 5 3 3 14 6 8 3 7 4 7 15 3 16 12 5 2 5 Average 3 3 5/6 2 1/2 7 21 9 4 17 22 4 5 12 24 29 7 9 4 2 31 4 4 13 33 1 11 3 13 34 5 2 16 Average 11/3 7 3 2/3 12 1/6 Day 4 Reaction 1 2 3 5 9 1 1/2 7 7 6 7 5 5 1/2 7 9 6 4 2 9 2 5 2 16 3 6 2 2/3 9 Day 6 Reaction 1 2 3 4 3 3 6 6 3 5 2 3 8 6 8 8 5 18 16 22 19 22 25 15 4 1/3 6 2/3 22 5/6 19 13 3 19 4 3 5 16 5 15 8 4 1/4 13 1/4 CHICKS HATCHED FROM ALCOHOLIZED EGGS 111 TABLE 2- —Continued Day 8 Day 10 Reaction Reaction No. of Chick 1 2 3 4 • 1 2 3 4 2 4 4 14 1 5 13 3 6 7 7 1 2 8 11 5 1 5 14 1 7 12 6 5 9 6 3 7 15 8 4 5 16 1 6 14 15 ' o 7 11 7 3 12 16 3 4 15 1 7 15 Average 3 4/7 5 6/7 116/7 1/7 2 2/7 6 1/7 13 1/7 21 3 6 14 2 5 15 22 3 4 17 24 14 4 14 14 3 14 29 6 6 9 2 5 31 1 2 4 18 7 4 14 33 5 4 16 4 3 16 34 6 6 17 1 4 7 9 Average 1/6 6 5 14 2/3 1/7 4 6/7 4 12 6/7 Day 11 Reaction No. of Chick 1 2 3 4 Time Until First Peck 2 3 3 14 17' 4 7 3 7 1 20 2' 10 '/ 5 4 4 14 4' r 6 " 2 o 18 38' 4' V 8 4 2 19 51 .6" 15 5 3 20 1' 28' r 16 7 4 14 Average 4 4/7 3 1/7 17 9' 56.9" 21 1 4 20 3' 34' t 22 3 3 19 4' r 24 13 5 13 15' t 29 3 3 14 4' 31 5 4 16 2' 20' r 33 3 2 19 4' 34 ?, 3 19 8' r Average 4 : 2/7 3 3/7 17 1/7 1' 52 ti The fifth hatching of chicks, which consisted of seven normal, four hole, seven water and five alcohol chicks, were also tested for pecking reactions. The method differed slightly from that of the preceding tests in that a record was kept of the first 112 FLETCHER, COWAN AND ARLITT twenty reactions in each test rather than of the total reactions for a certain length of time. No record was made of the time which elapsed before the first pecking reaction in the first test. In all other respects the method for this group was like that for the preceding groups. The following table shows the rate of improvement in peck- ing accuracy of each chick tested: TABLE 3 Chicks 1, 2, 3, 4, 6, 10 and 14 Were Normal Chicks. Chicks 20, 21, 23 and 25 Were Hole Chicks. Chicks 41, 43, 46, 47, 48, 51 and 54 Were Water Chicks. Chicks 71, 72, 74, 80 and 81 Were Alcohol Chicks Day 2 Day 3 Reaction Reaction No. of Chick 12 3 4 12 3 4 1 14 1 5 2 6 6 8 3 10 3 7 4 2 6 1 11 4 2 14 6 1 9 4 6 10 4 1 15 14 8 1 11 Average 11/2 8 1/6 12/3 8 2/3 6 11/2 12 1/2 20 13 1 6 21 4 4 2 10 23 7 6 7 25 1 8 1 10 8 6 6 Average 12/3 6 1/3 3 9 10 1/2 3 1/2 6 41 6 3 11 43 12 4 4 46 9 3 4 4 47 7 8 5 48 8 1 11 51 1 4 5 10 54 7 5 5 3 Average 3 2/5 6 4 1/5 6 2/5 7 1/2 4 1/2 8 71 3 10 4 3 72 1 6 13 74 80 9 3 8 81 7 5 2 6 1 2 1 16 Average 3 2/3 7 2 7 1/3 1/2 5 1/2 2 12 CHICKS HATCHED FROM ALCOHOLIZED EGGS 113 TABLE 3 — Continued Day 5 Day 6 Reaction Reaction No. of Chick 12 3 4 12 3 6 10 Average 21 23 Average 43 46 Average 71 72 5 74 Average 5 Day 7 Reaction No. of Chick 1 2 2 3 4 6 10 14 2 Average 2/3 20 21 25 Average 41 1 43 46 47 1 48 1 4 51 Average 2/3 2 71 72 74 81 Average 8 8 2 1 1 3 10 10 15 9 7 2 2/3 1 2/3 112/3 6 1 1 11 17 8 2 1 14 5 5 7 19 8 1 2 1/2 13 4 4 2 9 5 10 7 15 8 2 8 10 Day 10 Reaction 1 2 3 4 4 1/2 3 17 4 7 9 7 13 1 8 11 1 13 6 5 13 5 14 1/3 2 9 1/3 8 2/3 1 19 3 4 13 4 16 1 19 1 1/2 4 14 1/2 10 9 4 13 3 6 14 13 6 9 5 1 3 6 10 • 10 2/3 6 2/3 1/3 2 1/3 8 1/3 9 2 7 11 3 17 4 7 9 7 13 7 13 2 5 2/3 12 1/3 114 FLETCHER, COWAN AND ARLITT TABLE 3 — Continued 19 14 10 15 15 10 9 3 7 1/3 7 12 9 1/2 Day 13 Reaction No. of Chick 12 3 2 1 4 1 5 14 3 7 Average 11/3 4 1/3 14 1/3 20 2 3 Average 2 3 41 1 9 47 4 7 48 9 8 Average 4 2/3 8 80 13 81 1 7 Average 1/2 10 Day 14 Reaction No. of Chick 12 3 3 6 1 10 Average 1/3 21 23 1 25 Average 1/3 43 46 1 48 51 2 Average 1 71 1 8 11 72 1 19 74 1 19 Average 1/3 3 1/3 16 1/3 Day 16 Reaction 1 2 4 16 3 16 6 14 4 1/3 15 1/3 1 19 13 6 2 18 5 1/3 14 1/3 3 17 3 16 4 14 4 7 9 3 1/3 15 2/3 4 7 9 CHICKS HATCHED FROM ALCOHOLIZED EGGS TABLE 3— Continued 115 No. of Chick Day 17 Reaction 1 2 3 4 Day 19 Reaction 1 2 3 4 2 3 4 6 10 14 1 1 1 9 9 2 10 10 17 2 2 2 15 10 9 3 8 9 Average 1 6 2/3 12 1/3 2 11 1/3 6 2/3 20 21 23 25 2 7 11 1 2 2 8 7 6 11 11 12 Average 2 7 11 12/3 7 11 1/3 41 43 46 47 51 5 7 6 13 9 3 2 2 13 9 8 4 9 10 Average 2 1/2 6 1/2 11 2 1/3 10 7 2/3 72 74 80 81 3 14 11 3 9 1 3 7 17 12 Average 1 1/2 12 1/2 6 1/2 5 14 1/2 No. of Chick Day 20 Reaction 1 2 3 4 Day 21 Reaction 1 2 3 4 2 3 4 6 10 14 1 1 1 4 3 4 15 16 15 3 1 2 10 9 6 7 10 16 Average 1 3 2/3 15 1/3 2 8 1/3 11 20 21 23 25 2 4 14 2 2 5 8 5 15 10 13 2 4 14 1 1/3 6 12 2/3 116 FLETCHER, COWAN AND ARLITT 41 1 2 17 43 1 13 6 46 3 6 11 47 6 6 8 48 4 6 10 51 1 2 5 12 3 2/3 4 2/3 112/3 1/3 2 8 9 2/3 71 1 4 15 72 2 18 74 3 2 15 80 6 3 11 81 1 4 15 3 1/2 3 1/2 13 1 1/3 2 2/3 16 Day 24 Day 27 Reaction Reaction No. of Chick 1 2 3 4 1 2 3 4 5 7 12 11 TABLE 3 — Continued Day 20 Day 21 Reaction Reaction No. of Chick 12 3 2 1 6 4 3 6 1 3 Day 24 Reaction 1 2 2 3 4 2 6 10 14 1 Average 2 20 21 23 25 Average 41 2 43 46 47 2 4 48 2 3 51 Average 11/3 3 71 72 74 80 5 81 5 Average 5 8 11 6 2/3 11 1/3 1 19 1 19 2 16 6 9 8 6 5 2/3 10 9 5 6 10 8 6 8 1 1/2 7 6 13 3 13 10 11 1/2 10 7 13 2 2/3 7 1/3 10 9 8 10 10 1 6 13 1 1/3 7 2/3 11 1 14 5 3 1 7 2 1/3 8 5 12 12 9 2/3 CHICKS HATCHED FROM ALCOHOLIZED EGGS 117 The Effect of Imitation on the Pecking Reaction Some authors, notably Lloyd Morgan, 2 and Holmes 3 , are of the opinion that imitation plays a large part in the perfecting of the pecking reaction. Holmes states that imitation is an important factor in the education of young birds. The follow- ing passage may be quoted from Lloyd Morgan: "A hen teaches her little ones to pick up grain and other food by pecking on the ground and dropping suitable materials before them, while they seemingly imitate her action in seizing the grain." Thorndike, 4 is of the opinion that the pecking instinct is practically perfect at birth. Watson, 5 believes that the im- provement in the speed and accuracy of the pecking reaction is due solely to practice. In support of his view he quotes the results of Breed's 6 experiments with five-day old chicks which had been prevented from previous pecking. These chicks pecked with the accuracy of twenty-four-hour old chicks and required two days' practise before their accuracy was up to standard. In order to determine what role, if any, imitation played in the perfecting of the pecking reactions of our normal and alcohol chicks, three normal and three alcohol three-day old chicks of the second set were placed in separate pens. These pens were so constructed that each chick was kept completely isolated from and out of sight of the rest. The chicks were kept in these pens for eleven days and tested on the same days and by the same method as the group from which they were taken. Alcohols Nos. 21 and 28 were among the six chosen as they had not pecked at all when under observation on the third day. Two chicks, alcohol No. 28 and normal No. 7, died when seven days old. The pecking reactions of the four remaining chicks improved as rapidly as those of the group from which they were taken. Alcohol No. 21 made a higher record than was made by any other alcohol chick in spite of the fact that it did not peck at all while under observation on the third day and did not see any other chick peck at that time or any time 2 Morgan, Lloyd. Animal Behavior., p. 189. 3 Holmes, S. J. The Evolution of Animal Intelligence, p. 254. 4 Thorndike, E. L. Animal Intelligence, 1911, p. 160. 5 Watson, J. B. Behavior: An Introduction to Comparative Psychology, p. 141. 6 Breed, F. S. Maturation and Use of an Instinct. Jour, of Animal Behav., 1913, 3, 274. 118 FLETCHER, COWAN AND ARLITT thereafter until replaced with the other chicks on the eleventh day. The following table gives the record of the pecking reactions of these chicks during their period of isolation: TABLE 4 . Chicks 4 and 8 Were Normal Chicks. ( Chicks 20 and 21 Were Alcohol Chicks No. of Chick Day 3 Reaction 1 2 3 4 Day 5 Reaction 1 2 3 4 4 8 4 3 2 2 1 2 1 2 7 4 7 Average 3 1/2 2 1 1/2 1 1/2 7 4 7 20 21 2 5 3 6 9 7 13 Average 1 4 7 1/2 10 No. of Chick Day 8 Reaction 1 2 3 * 4 Day 10 Reaction 1 2 3 4 4 8 4 2 6 12 10 11 5 5 9 7 2 7 Average 3 9 10 1/2 5 8 4 1/2 20 21 1 8 3 13 19 1 5 15 12 1 Average 1/2 5 1/2 16 1/2 10 6 1/2 No. of Chick Day 11 Reaction 1 2 3 4 4 8 4 8 7 9 6 Average 2 7 1/2 7 1/2 20 21 7 8 10 10 7 1/2 10 CHICKS HATCHED FROM ALCOHOLIZED EGGS 119 Drinking Reaction A watch glass containing a few drops of water was used as a stimulus for the drinking reaction. This glass stood on a square of smooth, white paper, ten by ten inches. The chicks were placed one at a time on the paper, six inches from the glass, and a record made of the time before they drank and of their manner of doing it. The drinking reactions of twenty chicks of the first set were observed, seven on the third and thirteen on the fourth day. Only one of the seven three-day old chicks drank and that one apparently found the way by accident, first walking into the glass and then drinking. Five of the thirteen four-day old chicks, three alcohol and two normal, walked to the glass and drank, one drinking at once, the others after a short interval. Three, one normal and two alcohol, afterwards performed the drinking reaction on the smooth, white paper and the edge of the glass. One chick pecked at the water, then swallowed and immediately performed the drinking reaction. The others were placed close to the glass after a long interval during which they had run about the table and the water was ruffled as in the Breed 7 experiment. The chicks then drank. The same apparatus and method was used with the second set of chicks. The twenty-two chicks of this set found the water by accident, walking into the watch glass, or pecking at it and then drinking. One alcohol chick, No. 32, walked into the glass and immediately made the drinking reaction on the white paper on which the glass stood, walked out of the glass on to the paper, then walked back into the glass and went through the movements of the drinking reaction, scratching in the water at intervals. The touch of the water on the chick's feet had evidently served to start the drinking reflex. There was no difference in the behavior of alcohol and normal chicks of either set. Reaction to Heights Thorndike 8 , experimenting with reaction to heights with chicks ninety-four hours old, found the chicks hesitated for a longer and longer interval before jumping, as the height of 7 Breed, F. S. The Development of Certain Instincts and Habits in Chicks. Behav. Monog., vol. I, no. 1, 1911. 8 Thorndike, E. L. Animal Intelligence, 1911, p. 159. 120 FLETCHER, COWAN AND ARLITT the boxes on which they were placed increased, finally refusing to jump from boxes thirty-nine inches in height. We undertook to determine the height at which our chicks would refuse to jump and whether this height differed for normal and alcohol chicks. The subjects were six seven-day old chicks, three normal and three alcohol, and seven fifteen-day old chicks, two normal and five alcohol. The chicks were placed on stands above the box in which the other chicks were kept. The distance from the top of the stand to the floor of the box increased gradually from 10.7 cm. to 171.5 cm. The alcohol chicks w r ere not placed on stands of greater height than 171.5 cm., as it was feared that they might be injured in jumping to the box. The chicks were allowed to rest and feed for thirty seconds after each return to the box. There was little difference in the reactions of normal and alcohol six-day old chicks (see Table 5), but in the case of the fifteen-day old chicks the difference in behavior of the normal and alcohol chicks was marked. The normal chicks hesitated for longer and longer intervals as the height of the stands in- creased, waiting for over five minutes before jumping from a stand 106 cms. above the box. Though repeatedly pushed they would not jump from stands above this height. The alcohol chicks jumped with little or no hesitation. No. 32 would not jump from a stand 134 cm. in height, but Nos. 21, 28 and 22 jumped with almost no hesitation from stands 171.5 cm. above the box. The following tables present in detail the results of the ex- periments with heights: TABLE 5 Chicks 1, 2 and 10 Were Normal Chicks. Chicks 27, 35 and 34 Were Alco- hol Chicks. The Fact That the Chick Jumped is Indicated by +. The Fact That the Chick Refused to Jump is Indicated by — . Chicks Six Days Old Height No. of Chick 10.7 22.3 59.6 74 84.6 95 119.6 130 134 143 1 + + + + + + + + — — 2 + + + + + + + — 10 + + + + + — 27 + + + + + + + + - 35 + + + + + + + + — 34 + + + + - CHICKS HATCHED FROM ALCOHOLIZED EGGS 121 TABLE 6 Chicks 4 and 8 Were Normal Chicks. Chicks 21, 22, 28, 29 and 32 Were Alcohol Chicks. The Fact That the Chick Jumped is Indicated by +. The Fact That the Chick Refused to Jump is Indicated by — . Chicks Fifteen Days Old Height No. of Chick 10.7 22.3 74 89.8 95.5 106 134 158.3 171.5 4 + + + + + — 8 + + + + + — 21 + + + + + + + + + 22 + + + + + + + + + 28 + + + + + + + + + 29 + + + + + + + + + 32 + + + + + + — ACQUIRED REACTIONS Experiments on Behavior in Mazes During the course of the experiments with mazes three mazes of varying complexity were used which will be designated as Maze 1, Maze 2 and Maze 3. Maze 3 was used only with the fifth hatching of chicks, which contained hole and water chicks as well as normal and alcohol ones. The time which the animal spent in the maze was the criterion of learning. So far as the general behavior of the chicks in the mazes was concerned there seemed to be one main difference which was, however, quantitative rather than qualitative. The alcohol chicks were on the whole less prone to react to the maze situa- tion with general activity. An ordinary normal chick, when placed in a maze, will run about and attempt to get out. It will normally continue this running until it does get out and subsequent trials are made shorter by a gradual reduction of this running about. In other words, the process of learning the maze is like that of any other animal so placed. The alcohol chicks, as well as chicks raised from eggs which had been tampered with in other ways, that is, the three different kinds of abnormal chicks, did not seem to be inspired by this desire to get out of the maze. The situation did not seem to call forth the reaction of running about as it did with the normal chicks. The chick when placed in the maze might go promptly to sleep and sleep for some five minutes. It might then wake up and suddenly dart forward. If it brought up in a blind alley it might either go to sleep for another period, trying to the experimenter's nerves, or it might turn and go into the 122 FLETCHER, COWAN AND ARLITT correct path and so out. This sort of procedure naturally made the time record of the abnormal chick much longer than that of the normal chick. But there are many modifying points which must be considered when discussing this difference in the learning process involved in learning mazes. In the first place this time difference was not one which persisted through- out the learning process. The results will show that it is con- fined mainly to the first third of the number of trials used. After this point the time records of the abnormal chicks do not vary from those of the normal. On this account we are very doubtful whether this can be called a genuine difference in the learning process of normal and abnormal chicks. Another consideration is the fact that this lack of reaction while preva- lent among the abnormals was not invariable among them and was occasionally found among the normals. A glance at the tables will show that there were chicks among the alcohols, among the water and among the hole chicks whose records were as good or better than the average of the normal chicks. And there were individuals among the normals who showed the same lack of interest during the first trials in the maze that the abnormal chicks showed. And the final consideration is that the fact that a chick, either normal or abnormal, exhibited this sluggish type of behavior in a maze on one day did not necessarily mean that it would exhibit this same type of beha- vior in a different maze on another day. Again a glance at the individual records will confirm this. The experiments with the mazes seemed to indicate, then, that tampering with an egg in the ways we used would cause the chicken subsequently developing from it to be subject to fits of sluggishness and inactivity which might temporarily impede its progress in acquiring a new co-ordination. They did not in any way demonstrate the inability of such a chick to acquire new co-ordinations, nor did they even demonstrate that this sluggishness is an inevitable result of the tampering, or a constant characteristic of the chick in which it appears or a characteristic which does not sometimes appear in normal chicks. The ordinary chicken raiser who remarks that a certain chicken ' looks droopy and may be sick " is well acquainted with the type of behavior or lack of behavior we have been trying to describe. The following pages present the maze results in detail: CHICKS HATCHED FROM ALCOHOLIZED EGGS 123 Maze No. 1 Maze No. 1 (which perhaps should not properly be called a maze) was a straight path between two boards 55 cms. in length. It stood on a table below which was the box in which all the chicks were kept during the tests. One end of the path was closed. The open end was even with the edge of the table just above the box. During the tests of chicks 2, 6, 7, 9 and 34 the box was 79 cms. below the table edge. The wire netting cover of the box was made to lean against the table in such a way as to form an inclined plane from the table edge to the box. The chicks showed such reluctance to leave the table under these circumstances that the box was raised to within six inches of the table and the inclined plane done away with. The chick to be tested was placed at the back of the maze and the time it took to find its way back to the box recorded. Each chick was given ten trials in succession. An interval of one minute was allowed to elapse between the end of one trial and the beginning of another. The chick was sometimes placed in the maze by one experimenter and sometimes by another to correct any initial direction error which might occur. The following table gives the results for Maze 1 : TABLE 7 Chicks 1, 2, 3, 4, 5, 6, 7, 8 and 9 Were Normal Chicks. Chicks 27, 29, 32, 33, 34 and 35 Were Alcohol Chicks Trial No. of Chick 12 3 4 5 1 7' 15 " 2 6' 45 " 3 1' 9.4" 4 1' 28.8' 5 12 " 6 15.6" 7 1' 15.6" 8 23.8" 9 3' 51.6" Average 2' 30+" 27 27 " 29 30.6" 32 32.2" 33 8' 35.4" 34 12.4" 35 11' 23 " Average 3' 36+" 21 " 2' 27 " 5 " 4.8" 26.2" 2' 7 " 38 " 18 " 20.6" 8.4" 10 " 10 " 30.8" 18.4" 7.2" 9 " 8' 8.8" 6.8" 5.2" 3.6" 15.6" 52.6" 5.2" 5 " 11.6" 5.2" 3.2" 21.2" 38.8" 36 " 24 " 14 " 14 " 27.8" 10.2" 11.2" 1' 14+" 44 + " 12 + " 10+" 9 " 8" 9 " 11.2" 8.4" 4.2" 4.4" 3.6" 6 " 7 " 3 " 4.2" 8 " 1' 20 " 2' 16.4" 8 " 7 " 10.4" 7.4" 8' 5 " 2' 1 " 55 " 46 " 1' 27+" 37 + " 33 + " 14+* 124 FLETCHER, COWAN AND ARLITT TABLE 7 — Continued Trial No. of Chick 1 2 3 4 5 6 7 8 9 Average 27 29 32 33 34 35 Average 8 10 10.5" 9.8" 9 " 11 " 2.6" 3 " 15 " 9 " 9.6" 8+" 4 " 3 " 9 " 8 " 6.8" 29 " 9+" 6.5" 21 " 12 " 13.6" 2.8" 7 " 18 " 4 " 9.4" 10 + " 2.8" 4.6" 8 " 15.6" 5 " 15.2" 8+" 11.5" 28 " 10.8" 13.2" 2 " 5 " 26.6" 4.2" 24 " 13+" 4 " 4.2" 16 " 17 " 6 " 24 " 11+" 3.8" 18.8" 6 " 12.4" 2.8" 4.8" 8.6" 4.2" 9 " 7 + " 5 " 2.8" 8 " 8.2" 10.8" 27 " 10+" 4 " 45.2" 12.4" 9.4" 3.4" 4'. 4" 23.2" 11.6" 13.6" 14+" 3.2" 3 " 9 " 23. " 10.6" 16 " 10+" Maze No. 2 Maze 2 was a simple choice maze, consisting of one straight alley with two slightly shorter alleys running at right angles to it. The right hand alley ended blindly. The left hand alley was open, permitting the chick to jump to the box in which its companions were kept. The maze stood on a table, the edge of which was six inches above the box. The subjects of the first experiment with this maze were six normal and six alcohol chicks of the second set. The pro- cedure was the same as that used with Maze 1. The difference in the behavior of normal and alcohol chicks was more pronounced in these tests with Maze 2 than in any of the other tests. It was thought at the time that this was due to the increased difficulty of the maze problem which em- phasized differences in learning capacity. But later results with Maze 2 and with the still more complicated Maze 3 did not bear this hypothesis out. In these tests with Maze 2 the normal chicks with one exception, found the right path and thereafter chose it. One chick chose the wrong alley on the second trial and went half way to the end, then turned and chose the right path. The normal chicks which went to the box slowly did so because they stopped and pecked at the bright nail heads with which the strips where the alleys joined were studded. CHICKS HATCHED FROM ALCOHOLIZED EGGS 125 All of the alcohol chicks made two or more wrong choices after the first trial. Several tried to peck their way through the side and two tried repeatedly to fly out at the back though they had several times found the path to the box. There is not one instance of this kind of behavior in the case of the normal chicks of this set. Two of the alcohol chicks went to sleep and had to be prodded repeatedly. No. 27, whose record in Maze 1 was as good as that made by the normal chicks, found the right path, chose it three times in succession, made one wrong choice in each of the three fol- lowing trials, went straight to the box in the seventh, eighth and ninth trials, then hesitated a long time before making a choice on the tenth trial. Alcohol No. 35 failed to find its way out of the maze. The first trial it was slowly pushed to the entrance at the end of twenty-five minutes. After an interval of thirty seconds, during which it was allowed to feed with the other chicks, it was re- placed in the maze. At the end of twenty -five minutes, during which it had repeatedly gone to sleep, it was replaced with the other chicks. The following table gives the results in detail: TABLE 8 Chicks 1, 2, 3, 4, 6 and 8 Were Normal Chicks. Chicks 21, 22, 28, 29, 32 and 35 Were Alcohol Chicks Trial No. of Chick 1 1 4' 41.2" 2 4' 20 " 3 32.2" 4 6 " 6 9' 1 " 8 3' 22 " Average 3' 40+" 21 1' 59 " 22 2' 30 " 28 39.9" 29 20' 23 " 32 19' 35 Failed 40.2" 11.2" 2 " 1.8" 42 " 13.6" 13 " 2.2" 18.6" 3 " 4 " 3.9" 5 " 4 " 2.2" 2 " 1' 35 " 3' 52 " 1' 13 " 2' 10 " 33 " 10 " 4 " 3.6" 38+" 45+" 16+" 23+" 16 " 20 " 4 " 5 " 32.8" 50.6" 15 " 24 " 8.6" 16.4" 16 " 6 " 6' 42 " 30 " 17 " 20.2" 5 " 15 " 4.4" 3.6" Average 5' 34 + " 1' 32+" 26+" 11 + " 11+" 126 FLETCHER, COWAN AND ARLITT TABLE 8 — Continued Trial No. of Chick 6 7 8 9 10 1 2 " 3 " 1 " .9" 1 " 2 2.8" 1.4" 2.4" 1 " 2 " 3 3 " 4" 5 " 18.6" 2 " 4 2 " 2.4" 5 " 4 " 6 2' 2.5" 4 " 6.8" 2.8" 3.4" 8 5.8" 4.4" 2 " 5 " 5 " Average 23+" 3+" 3+" 5 + " 2 + " 21 6 " 4 " 10.4" 5 " 2 " 22 18 " 4 " 4.8" 9.4" 4.4" 28 15 " 9 " 11.8" 5 " 5.8" 29 5 " 18 " - 5 " 13 " 4.6" 32 3 " 2.8" 3 " 6.4" 2 " 35 Failed Average 9+" 7+" 7+" 7+" 3+" The second experiment with this maze was made on the chicks of the fifth hatching, consisting of six normal, four hole, six water and five alcohol chicks. The procedure was like that used with Maze 1. The following table gives results in detail: TABLE 9 Chicks 2, 3, 4, 6, 10 and 14 Were Normal Chicks. Chicks 20, 21, 23 and 25 Were Hole Chicks. Chicks 41, 43, 46, 47, 48 and 51 Were Water Chicks. Chicks 71, 72, 74, 80 and 81 Were Alcohol Chicks. Trial No. of Chick 1 2 2 " 3 9 " 4 58 " 6 4' 1.5" 10 4 " 14 30 " Average * 57 + " 20 1' 48 " 21 35 " 23 15 " 25 5' Average 3' 24+" 3 " 3.5" 5 " 5.5" 6.5" 7 " 15 " 8 " 10.5" 4 " 14 " 3 " 13 " 4' 47 " 19 " 31 " 2 " 2.5" 3.5" 4.5" 39 " 4.5" 6 " 8.5" 12+" 18+" 51+" 17+" 10 + " 8 + " 10 + " 10 " 15 " 4.5" 3.5" 21 " 12.5" 14.5" 11 " 31 " 15 " 10 " 5 " 11.5" 26.5" 4 " 4.5" 6 + " CHICKS HATCHED FROM ALCOHOLIZED EGGS 127 TABLE 9— Continued Trial No. of Chick 41 43 46 47 48 51 Average 71 72 74 80 81 Average No. of Chick 2 7.5" 3 22 " 4 7 " 6 8.5" 10 6.5" 14 3.5" Average 9+" 20 6 " 21 33.5" 23 22 " 25 1'27 " Average 47+" 41 10.5" 43 12 " 46 11 " 47 17 " 48 2.5" 51 7.5" Average 10+" 71 15 " 72 1' 17 " 74 8 " 80 4.5" 81 3 " Average 21+" ick 1 2 3 4 5 1' 3 " 8' 5 " 1' 35 " 55 " 7.5" 17.5" 2' + 10 " 12 " 1.5" 6 " 43 " 12 + " 14 " 10' 42 " 47 " 17 " 22 " 1' 2 " 2' 12+" 4 " 8 " 11 " 9' 2 " 22 " 1' 57+" Trial 12 " 18 " 25 " 22 " 5 " 7 " 16+" 9 " 5 " 16 " 8 " 8.5" 9+" 5 " 4 " 23 " 12 " 4 " 35 » 13 + " 12 " 6 " 9 " 5 " 10 " 8+" 3.5 " 4 " 21 " -« 10 " 2.5" 13 " 9+" 10 " 47 " 9.5" 3.5" 7 " 15+" No. of wrong turns 6 7 8 9 10 13 " 5.5" 3.5" 7 " 4 " 3 " 6+" 4 " 11 " 11 " 7 " 8+" 3 " 3 " 6.5" 13 " 2.5" 1' 13 " 15+" 6.5" 8 " 3 " 18 " 5 " 3 " 7+" 6.5" 24.5" 17 " 4 " 4 " 15 " 20 " 2.5" 25 " 11+" 5 " 11.5" 3 " 7 " 3 " 6 " 5 + " 3 " 44 " 45 " 9.5" 27+" 8 5 19 4 3 7 " 7+" 20 10 17 3 5 4' 11+" 12.5" 3 " 4 " 53+" 13 1'40 19 3 5 13 " 13.5" 3 " 18 " 7.5" 6 " 10+" 3 " 24 " 7 " 4 " 9 + " 9 " 4 " 17 " 9 " 2.5" 8 " 8+" 21 " 7' 6 " 8.5" 5 " 15 " 1 1 2 b o i ■> i i 2 3 2 1 28+" 1' 15+" 128 FLETCHER, COWAN AND ARLITT Maze No. 3 Maze 3 consisted of a series of short, straight alleys arranged in the following pattern: a a was the point at which the chick was placed in the maze. b was the open exit where the chick emerged at the edge of the table holding the maze and ran down a short incline into the chick pen ten inches below the table edge. The chicks used with this maze were those of the fifth hatch- ing. Each chick was given thirty successive trials without interval. The problem appeared to be as difficult as it could be and still be solved by the normal chicks. But even here the hole, water and alcohol chicks were at a disadvantage only during the earlier trials. Their records for the last ten trials are as good as those of the normal chicks. Certain of the chicks of this hatching made little or no im- provement in the course of the tests with Maze 2. These were chicks 10, 21, 51, 71 and 72. Chick 21 died before the tests with Maze 3 were begun. Of the others the only one which did not make a record which showed the normal rate of im- provement in Maze 3 was chick 10, a normal chick. Chick 43 was the only other chick whose behavior was unusual in both mazes. This chick was amazingly sluggish during the early trials in Maze 2. Its record for that maze is entirely normal, however, in its improvement as the trials progressed. This sluggishness was not apparent in Maze 3 but the time, record increases instead of decreasing gradually as the trials progress. CHICKS HATCHED FROM ALCOHOLIZED EGGS 129 Chick 43 was a water chick. Alcohols 71 and 72, who made very bad records in Maze 2, both made excellent records in Maze 3. Water 51, although slow in Maze 3, showed a dis- tinctly normal rate of improvement. If consistent inability to decrease the time in running a maze is any indication of inability to acquire a new co-ordination then our results would indicate that the only chick we had which showed such inability was one hatched from an egg any unusual characteristic of which we were neither responsible for nor cognizant of. The following table gives the results for Maze 3. TABLE 10 Chicks 2, 4, 10 and 14 Were Normal Chicks. Chicks 20, 23 and 25 Were Hole Chicks. Chicks 41, 43, 46, 47, 48 and 51 Were Water Chicks. Chicks 71, 72, 74, 80, and 81 Were Alcohol Chicks. Trial No. of Chick 1 2 3 4 5 6 2 2' 15 " 38 If 30 a 38 a 25 " 31 // 4 42 " 1' 2 It 12 i/ 50 a 44.5" 27 // 10 10 " 11 It 28 n 17 a 10.5" 10 // 14 •. 1' 1.5" 41 If 4 ii 16 a 27 " 21 It Average 1' 2+" 38 " 18+" 30+" 26+" 22+" 20 13' 23 33 " 25 1' 7 " Average 4' 53+" 1'+ 26+" 5' 9+" 1' 20+" 48+" 32.5" 52 " 14' 16 " 31 " 13 a 55 " 18 " 45 " 1' 55 " 1' 4 a 1' 34 " 8.5" 28.5" 1' 36 " 1' 7 a 41 34 43 33 46 1' 19 47 3' 20 48 28 51 1' 9 a a 27 " 2' 44 " 6' 46 a 43.5" 2' 24 » 8 " 25 " 29 n 16 " 16 n 41.5" 2' 1.5" 44 a 1'37.5" 47 ii 43 " 17 " 38 a 25 " 42 ii 25 " 2' 30.5" 50 a 33.5" 7 ii 2' 50 " 29 " 6' 11 a 20 " 9 ii Average 1' 13+" 52+" 1' 24+" 2' 37+" 39+" 44+" 71 33.5" 72 5' 30 " 74 17 " 80 48 " 81 22 " Average 1' 30+" 1' 54 + " 1' 28+" 1' 28+" 1' 8 + " 1' 35 + " 10.5" 25 If 5 a 27 " 11.5" i 7 ,i 5' 16 It 5' 57 a 3' 12.5" 3' 28 " 53.5" 43 If 51 a 25.5" 3' 34 " 8.5" 5 ft 12 a 9 " 4.5" 14 " 51 It 18 a 1'28 " 41 " 130 FLETCHER, COWAN AND ARLITT TABLE 10 — Continued Trial No. of Chick 2 4 10 14 Average 33+" 20 42 " 23 59 " 25 18.5" Average 39+" 41 1' 1 " 43 13.5" 46 20 " 47 36 " 48 34 " 51 11.5" Average 29+" 71 21 " 72 1' 1 " 74 2' 5" 80 5 " 81 3' 32 " Average 84+" 7 8 9 10 11 12 1' 21 ft 43 If V 11 It 13 " 30 It 18 // 20 If V 52 It 37 tf 58.5" 1' 2. 5" 27 // 11 It A If 9 if 3 " 4 If 12 It 23 It 40 It 13. 5" 8.5" 9 It 20 It 49 + " 43 + " 23- 28 + " 32 + " 47 + " 34+" 20 + " 18 + " 16+" 26 + " 12+" 20+" 28+" 1' 56+" 31+" 19 + " 46 it 7 " 9 " 7.5" 12 // 49 It 1'45.5" 40.5" 25 " 1' 12 a 36 It 30 " 6 " 5 " 13 a 32 + " 15 " 7 " 6 " 7 It 7 " 9.5" 14.5" 24.5" 34 If 20.5" 59 " V 6 " 12 " 19 It 25 " 25 " 1' 29 " 23.5" 33 It 20 " 10 " 15 " 12.5" 11 tf 20 " 22 " 18 " 18 " 21 If 3' 2 " 45 + " 11 " 8 " 58 a 16 " 4 " 35 " 55 " 30 a 1' 1 " 58.5" 16.5" 12 " 16 a 18.5 " 19 " 6 " 7.5" 4 a 43 " 8 " 1' 13 " 1' 7' 53 a 17 " 1' 19 " 33+" Trial 14 No. of Chick 13 2 40 " 4 7 " 10 4 " 14 29 " Average 20+" 20 1' 27 " 23 30 " 25 22 " Average 46+" 1' 12+" 15 16 17 18 39 If 39 3' 8 It 25 18 If 5 6 tf 7' 52 50 If 43 If V 23 " 30 It 22 If 1'36.5" 24 ft 28 It 10 " 15 It 14 It 2' 31 " 1' 2 + " 2' 15 29 + " 26 + " 1' 25+" 11.5" 38.5" 2' 54 If 10' 56.5" 39 " 3' 4 " 22.5" 35 It 19 " 27 " 23 " 25 " 20 It 21.5" 5.5" 28 + " 1' 16 + " 3' 52+" 23+" CHICKS HATCHED FROM ALCOHOLIZED EGGS 131 TABLE 10— Continued Trial No. of Chick 13 14 15 16 17 18 41 6.5" 7 " 8 " 7 " 8 " 7 " 43 16 " 29 " 24 " 15 " 12.5" 13 " 46 13 " 15 " 16 " 15 " 21 " 19 " 47 30 " 55 " 13.5" 38 " 1'23.5" 39.5" 48 14 " 4 " 21 " 5 " 27 " 11.5" 51 1'26 " 17 " 1' 51 " 25 " 16.5" 46.5" Average 25+" 21+" 32 + " 17+" 28+" 22+" 71 5 " 19 " 4 " 4" 16 " 4 " 72 55 " 39.5" 47 " 38 " 53.5" 22 " 74 15 " 19 " 51 " 29 " 42 " 14 " 80 3 " 3 " 6 " 6 " 4 " 3.5" 81 32 " 30 " 2' 6 " 1' 36 " 19 " 16 " Average 22+" 22+" 46+" Trial 34 + " 26+" 11+" No. of Chick 19 20 21 22 23 24 15+" 15 " 5 " 26+" 2 45 " 4 1' 4.5" 10 40 " 14 30 " Average 37+" 20 6 " 23 48 " 1' 25 19.5" Average 24+" 41 16 " 43 59 " 46 8 " 47 2' 25 " 48 5 " 51 24 " Average 42+" 1' 7+" 71 5 " 72 43 " 74 15 " 80 13 " 81 14 " Average 18+" 13+" 21 " 16.5" 58.5" 15 " 15 " 21.5" 21.5" 6 " 9 " 9 " 12 " 16 " 19 " 6.5" 7.5" 9 " 5 " 22 " 6 " 5 " 14+" 26+" 9+" 18+" 12+" 13+" 40+" 20+'' 51+" 9+" 9 « 9 a 7.5" 9 42 a 24 a 29 " 29 5 a 4 a 4 " 40 26+" 1' 11.5" 10 a 8 " 6 a 5 a 1' 33 " 33 a 25.5" 44 a 1' 5 it 13 " 16 a 58. " 1' 14 a 18 " 17 it 15 " 1' 16 a 1'36 n 6 " 6 a 4 " 5 a 19 u 3' 23 " 2' 41 a 12.5" 1'55 a 14 a 35+" 3 " 3 " 11 " 3 a 18 " 21 " 46 " 44 " 45 a 30 " 26 " 11 " 9 " 1'23 a 14 " 3.5" 7.5" 4 " 4 a 5 " 16 " 14 " 10.5" 6 a 16.5" 16+" 17+" 28+" 16 + " 132 FLETCHER, COWAN AND ARLITT TABLE 10— Continued Trial No. of No. of Chick 25 26 27 28 29 30 errors 2 9 " 20 " 18 " 57.5" 49 " 36 " 21 4 24 " 39 " 1' 8 " 23 " 28 " 10 " 26 10 28 " 43 " 26 " 1' 24 " 53 " 7 " 14 14 11 * 5 " 27 " 15 " 9 " 12 " 19 Average 18+" 26+" 34+" 44 + " 34 + " 16 + " 20 7.5" 22 " 29 " 32 " 7 " 30.5" 13 23 43 " 51 " 18 " 31 " 4 " 35 " 20 25 5 " 36 " 24 " 6 " 6 " 21 " 20 Average 18+" 36+" 23 + " 23 + " 5 + " 28+" 41 5 " 7 " 6.5" 11 " 9.5" 6 " 8 43 21 " 50 " 30 " 32 " 53 " 41 " 23 46 12.5" 29 " 28 " 1' 3 " 23 " 9 " 21 47 8 " 19 " 19 " 29 " 23 " 11 " 12 48 5 " 6 " 4 " 4 " 10.5" 12.5" 16 51 1'43.5" 1' 33 " 1' 13 " 2 " 36 " 2' 13 " 8 Average 25+" 34+" 26 + " 23 + " 25+" 35 + " 71 6 " 3 " 10.5" 2.5" 3 " 5.5" 14 72 1' 21 " 47 " 1'38 " 55 " 45 " 1' 3.5" 29 74 10 " 19 " 26 " 11.5" 23 " 8 " 14 80 5 " 5 " 4.5" I'll " 14 " 24 " 11 81 8 " 5 " 4 " 7 " 4 " 4 " 18 Average 22 " 15+" 28+" 29+" 17+" 21 + " EXPERIMENTS WITH SENSORY DISCRIMINATION The apparatus used for testing the ability of the chicks to make sensory discriminations was the Yerkes visual choice appa- ratus. Only the two middle alleys were used, the other two being shut off by wooden partitions. Stimulus screens with openings in them large enough to permit the chick to pass through were slipped into the grooves at each end of each alley. The screens were made of photographic mounting card- board and two of them were light grey and two dark grey in color. The exit path in which the light grey screens were placed was always open and the ' chick choosing that path pro- gressed directly out of the apparatus to the chicken pen and the society of the other chicks. The exit path characterized by the dark screens was always closed at the far end and the CHICKS HATCHED FROM ALCOHOLIZED EGGS 133 wires crossing the floor of the path were electrically charged so that the chick choosing that path received a slight shock and did not get out of the apparatus until it had gone back and found the other exit. The light screens were placed alter- nately in the left and right hand alley. The chick to be tested was placed in the release box at the back of the apparatus and the wire screen which led to the paths was not lifted until the chick crowded against or pecked at it. No chick was ever left struggling in the release box. The chick was always allowed to find its way in time into the correct alley and out of the apparatus but its record for the trial depended on its first Choice of exit paths. Each chick was given ten preference trials with both exits open and the wires not charged. The one chick which showed a preference was not used in these experiments. Chicks of the third hatch- ing were used with this apparatus. Six normal and five alcohol chicks were each given three trials a day for twenty -five days. Immediately after this normal No. 5 and alcohol No. 24 and six days later normal No. 13 and alcohol No. 33 were given thirty additional trials in succession. The records for the last fifteen trials at the end of the first seventy-five trials are as follows: one normal chick made 73^% correct choices; two made 60% correct choices; one made 46|% correct choices; one made 40% correct choices; one alcohol chick (No. 24) made 80% correct choices; two made 46f% correct choices; one made 33^% correct choices, and one made 26f% correct choices. The records for the last ten of the thirty additional trials given four of the chicks are as follows: normal No. 5 made 100% correct choices; normal No. 13 made 90% correct choices; alcohol No. 24 made 60% correct choices; and alcohol No. 33 made 40% correct choices. Chicks of the fourth hatching were also used in this appa- ratus. These chicks were trained in pairs, each pair being com- posed of a normal and an alcohol chick. One normal chick was given a trial. Then an alcohol chick was given a trial. Then the normal chick was retested and the alcohol chick re- tested and so on. This procedure was continued until one of the chicks had made ten consecutive correct choices. Tests with that pair were then discontinued and tests with a new pair begun. Normal No. 14 and alcohol No. 31 were each given 101 trials. 134 FLETCHER, COWAN AND ARLITT Normal No. 14's percentage of correct choices during the last ten trials was 100; alcohol No. 37's was 30. Normal No. 18 and alcohol No. 30 were each given 75 trials. During the last ten trials the normal chick's record was 100% correct choices and the alcohol's record was 50% correct choices. Normal No. 15 and alcohol No. 34 w T ere each given 67 trials. The normal chick made a record of only 70% correct choices during the last ten trials whereas the alcohol chick made a record of 100%. The alcohol chicks plainly formed the association more slowly than did the normal chicks and there were alcohol chicks which seemed unable to form the association in the time at our dis- posal for training them. But as in the experiment with the mazes there were individuals among the alcohols which did as well as normal chicks and there were normal chicks which did as badly as the alcohol chicks. INHIBITION OF AN INSTINCTIVE REACTION The apparatus used in this experiment was a large square of dark grey paper over which was placed a pane of glass the exact dimensions of the paper. On the paper under the glass were scattered bits of egg yolk. Ten chicks of the fourth hatch- ing, five normal and five alcohol, w r ere used in this experiment. The group of normal chicks was placed on the glass for five minutes each day and the total number of pecks of all five chicks was recorded. The chicks were then replaced in the box and the same procedure repeated with the group of alco- hols. These tests were repeated for twenty-five days. There was no appreciable difference in the behavior of the two groups. The alcohol chicks learned to inhibit the reaction with approx- imately the same rapidity as the normal chicks. The following table gives the number of pecks each group made on each of the trial days. TABLE 11 Day 1 2 3 4 5 6 7 8 9 10 11 12 13 No. of pecks made by normal chicks 4662121000122 No. of pecks made by alcohol chicks 4661301040001 CHICKS HATCHED FROM ALCOHOLIZED EGGS 135 TABLE 11 — Continued Day 14 15 16 17 18 19 20 21 22 23 24 25 26 No. of pecks made by normal chicks 0121242511001 No. of pecks made by alcohol chicks 1033253403010 INHIBITION OF AN ACQUIRED REACTION It will be remembered that the chicks of the fifth hatching were used in Maze 2. This maze was a reversible maze so constructed that after it had been used with the exit at one side it could be turned over and present the same problem, except that the exit would be at the other side, requiring the chick to turn in the opposite direction from that in which it had formed a habit of turning. Each chick of the fifth hatch- ing was given ten trials in this maze (for record see Table 9) and immediately thereafter the maze was reversed and the chick given ten more trials. The chicks which failed successfully to inhibit the previously acquired reaction were normal chick No. 2 and alcohol chicks Nos. 74 and 80. Chicks Nos. 10, 21, 51 and 72 made very poor records but these chicks, it will be re- membered, had never acquired the proper reaction in the maze during the first ten trials. They were equally unsuccessful when the maze was reversed. These four chicks, it will be noted, represent the four different groups of chicks obtained in this hatch- ing. In the matter of inhibiting an acquired reaction these results seem to give no pronounced advantage to any group of chicks. The following table gives the results for Maze 2 reversed: TABLE 12 Chicks 2, 3, 4, 6, 10 and 14 Were Normal Chicks. Chicks 20, 21 and 25 Were Hole Chicks. Chicks 41, 43, 47, 48 and 51 Were Water Chicks. Chicks 71, 72, 74, 80 and 81 Were Alcohol Chicks Trial No. of Chick 12 3 4 5 2 8.5" 18 " 8.5" 10 " 7.5" 3 1'34.5" 26.5" 10.5" 10 " 19 " 4 2 " 2 " 4 " 3 " 3.5" 6 6' 42. 5" 17 " 3' 28 " 39 " 26 " 10 16.5" 3 " 13 " 4 " 14.5" 14 4 " 2 " 3 " 2.5" 2 ' Average 1' 28+" 11+" 41+" 11+" 12+" 13G FLETCHER, COWAN AND ARLITT TABLE 12— Continued Trial No. of Chick 1 20 V 3 " 21 51.5" 25 12' 57 " Average 4' 57+" 41 27.5" 43 9 " 47 1' 3 " 48 4' 51 " 51 25 " Average 1' 23+" 71 20 " 72 30 " 74 12.5" 80 5 " 81 7' 12 " Average 1' 38+" 37 " 13 " 15 " 13 " 37 " 15.5" 33 " 2' 54 " 2' 45 " 1' 54 " 8 " 21.5" 1' 19+" * 47+" 18+" 1' 9+" 28 " 9 " 29 " 8 " 8 " 5 " 12 " 24 " 10 " 34 " 18 " 16 " 10 " 10 " 4 " 16 " 30 " 13 " 6.5" 8.5, 17+" 14+" 13+" 14+" 8 " 3 " 5 " 6 " 21 " 45 " 17 " 8' 15 " 6 " 5.5" 5 " 11 " 13 " 15.5" 7 " 17 " 1' 4.5" 14 " 11 " 5 " 22+" 16+" 9+" 1' 46+" Trial No. of Chick 6 7 8 9 10 2 : ... 19 " 7 " 19 " 8 " 31 " 3 15 " 6 " 10 " 12.5" 9.5" 4 3 " 4 " 2.5" 3.5" 3 " 6 8 " 6 " 5 " 1' 3 " 13.5 10 4.5" 15.5" 15.5" 6.5" 25 " 14 3.5" 2 " 2.5" 5 " 2 " Average 8+" 6+" 9 + " 16+" 14+" 20 13 " 28 " 14.5" 17.5" 10 " 21 35 " 21.5" 3' 10 " 31 " 46.5" 25 6.5" 45 " 5.5" 5.5" 5 " Average 18+" 31+" 1' 10+" 18+" 20+" 41 13 " 23 " 20 " 7.5" 19.5" 43 14 " 10 " 32 " 8 " 9 " 47 20 " 8 " 19 " 4 " 8 " 48 12 " 17 " 9 " 4 " 12 " 51 36 " 16 " 43 " 1'40 " 37.5" Average 19+" 14+" 24+" 24+" 17+" 71 4 " 5 " 9 " 3 " 4 " 72 6' Became so inactive the experiment had to be dropped. 74 3.5" 8 " 12 " 8 " 12.5" 80 13 " 6 " 17 " 23 " 4.5" 81 4.5" 4 " 3.5" 20 " 10 " Average 1' 15 + " 5+" 10 + " 13+" 7+" No. of wrong turns 10 3 2 2 3 3 4 6 2 6 2 1 5 1 1 CHICKS HATCHED FROM ALCOHOLIZED EGGS 137 CONCLUSIONS 1. Chicks are neither positively or negatively phototropic. Some show what appears to be an instinctive positive response to light on the first but not on the second and third day; some on the second but not on the first and third day; a few not until the third day and some not at all. 2. Imitation plays no part in the perfecting of the pecking reaction of either normal or alcohol chicks. 3. It is possible by tampering with eggs before the hatching of chicks to modify the behavior of the chicks in the following ways; on the fifteenth day they do not show the caution in jumping from heights which chicks hatched from normal eggs show all the time and which the chicks hatched from abnormal eggs show earlier than the fifteenth day; they are less likely than normal chicks to react to a maze situation with general motor activity and consequently are slower in beginning to acquire the consequent motor co-ordination though they may eventually acquire it in the same number of trials required for a normal chick; -they are slower in acquiring an association .depending on sensory discrimination in the Yerkes visual choice apparatus; the instinctive reactions of pecking and drinking apparently remain unmodified; the processes of inhibition of both instinctive and acquired reactions seem to be alike in normal and abnormal chicks. 4. It is not possible to find any difference between the in- stinctive reactions of pecking or between the modifications of the acquiring of reactions in the maze in the case of chicks raised from eggs into which alcohol has been injected by our method, and chicks raised from eggs into which distilled water has been injected or eggs which have had the shell perforated and sealed. That is, alcohol, per se, seems to have no specific effect beyond an effect which can easily be explained by malnu- trition during hatching and which may be as readily produced by other agents as by alcohol. DEATH FEIGNING REACTIONS IN TYCHI US PICIROSTRIS E. MELVILLE DuPORTE Biological Laboratories, Macdonald College (McGill University), Canada The behavior of the lower organisms is of importance, not only for the intrinsic interest attached to the study of the subject but also for the light which it may throw on the more complicated reactions of the higher organisms. Any contribu- tions, therefore, which will add to our knowledge of any phase of animal behavior or confirm results already obtained in this field, which has not been greatly exploited, will be of some interest to the student of comparative psychology; and for this reason the writer is publishing the following brief notes on observations incidentally made while studying the habits of Tychius picirostris, an European weevil observed this year for the first time in this country doing appreciable injury to clover. The habit of feigning death is found in a large number of animals but is most marked in certain insects. This habit among insects is of interest not only to the student of behavior but also to the collector and to the applied entomologist. The entomological collector often takes advantage of this reaction in collecting beetles and other insects which fall to the ground in a death feint when a sudden jar is given to the plant on which they are feeding, and this practice of jarring has been used effectively in the control of certain insect pests such as the May beetles {Lachno sterna spp.) and the plum curculio (Conotrachelus nenuphar Herbst), especially prior to the advent of arsenical sprays. Methods of Producing the Feint. The weevils congregate in large numbers on the leaves of the clover in the spring, and on the flower heads later in the season. If a plant attacked by these weevils is jarred, the majority of the insects will fall to the ground. Their action upon falling shows considerable individual variation; most of those which fall on their backs DEATH FEIGNING REACTIONS IN TYCHIUS PICIROSTRIS 139 will assume the death feigning attitude for a few seconds, and, on recovery, turn over and either scramble away immediately or remain quiescent for a brief period before moving away. A few of those which fall on their backs will not feign death but will turn over and move away at once. Of those which fall to the ground right side up some will remain quiet for a short time with their legs in the normal resting position, others will run away immediately on reaching the ground but very few will assume the typical death feigning attitude. In the laboratory the feint was obtained in one of three ways : by dropping the insect on the desk, by pinching it lightly with a forceps or by merely turning it on its back. The last method was usually successful in evoking the death feigning response and was generally employed. After remaining in the feint for a variable period the insect relaxes its legs and struggles to turn over. On turning over, or on being turned over the insect scrambles away, but while it is running it may be induced to feign a second time by again turning it on its back, and within certain indefinite limits to be indicated below this process may be repeatedly successful. While struggling on its back the insect may be induced to feign death by tapping it on the sternum. If, however, it is in the normal position, with its back turned up it will not feign death on being tapped on the back; on the contrary a quiescent insect thus treated will become active. The intensity of the shock seems to have no effect on the production or duration of the death feint, as an insect which has been merely turned over, or dropped from a height of six inches remains in the death feint as long as one dropped from a height of eight feet or more. It was also found that if the feint could not be elicited by overturning the insect or by drop- ping it from a height of six inches or thereabouts, it was im- possible to produce it by dropping from greater heights up to eight feet, or even by lifting with forceps and forcibly dashing to the table. It was found impossible to obtain the death feigning reaction with some individuals. As the insects experimented on had matured some time previous to the period of these investiga- tions and had not been under constant observation from the 140 E. MELVILLE DuPORTE time of eclosion it is not possible to say whether these individuals always reacted negatively, or from what cause they had lost the power of response. Death Feigning Attitudes. The attitude typically assumed by the insect during the death feint is not a simulation of the atti- tude of the dead insect. In death the femora are always ex- tended more or less perpendicularly from the ventral surface of the body, and are never pressed against the sternum, even if the insect is killed when in the feigning attitude. The tibiae and tarsi are flexed in such a way that the latter are often more or less intertwined and always quite close to each other. In the death feigning attitude the legs are all pressed closely against the sternum and the tarsi are never very close to each other. Figure 1. — Death-feint of Tychius picirostris. A and B, lateral and ventral views of the death-feigning attitude. C, lateral view of the insect after death. X21 The fore femora extend straight out laterally, pressed against the presternum, the tibiae are bent back against the ventral face of the femora and the tarsi are bent forward and outward. The middle femora also extend laterally and slightly forward, the tibiae bend caudo-mesad, their apex being situated in front and to one side of the meta-coxae. The tarsi run either straight or obliquely backwards. In the hind legs the femora are stretched obliquely backwards, the tibiae bent forward almost parallel to DEATH FEIGNING REACTIONS IN TYCHIUS PICIROSTRIS 141 the femora and the tarsi bent backwards. The rostrum is almost perpendicular to the body and the antennae are folded together behind the rostrum. Often the insect becomes quiescent on being stimulated, with- out assuming the attitude described in the foregoing paragraph. In such cases it may remain perfectly still with the legs out- stretched or partially and asymetrically flexed. This attitude I have termed the partial feint and it will thus be referred to when necessary to distinguish it from the typical feint. Some- times the insect when stimulated assumes the typical death feign- ing attitude, but after a while it slowly relaxes its legs without becoming active. Usually but not invariably the partial feint is manifested after the insect has been repeatedly stimulated. It is therefore probable that the failure to flex the legs is due to muscular fatigue. Variability of the Feigning Response. The feigning response to tactual or shock stimuli is not stereotyped as there exists a considerable range of variability in the duration of the feint, not only in the different individuals but in the same individual in successive feints. In some cases too, as stated above, I found it impossible to evoke the feigning response by any means. The reaction of mutilated specimens was also not constant owing no doubt to slight differences in the sensitiveness of the nervous system which would influence the intensity of the shock effects of mutilation. While any general statement made is based on the results of the reaction of the majority of individuals, the number failing to give the typical response was in some cases considerable, and for this reason special attention is called to the fact here. As to the exact cause of this variability I am unable to give a positive explanation. Probably, however, it is not wholly due to the inherent individuality of the specimens but partly to the varying "experiences" of the individuals. If the insects had been kept under the same conditions from the time of their eclosion it is probable that the reactions would be more nearly uniform, or where a wide variability exists there would be an evident explanation. Duration of the Death Feint. The duration of the feint at room temperature is usually very short, the insect recovering almost invariably before the expiration of a minute. The 142 E. MELVILLE DuPORTE longest feint obtained in these experiments lasted for seven minutes and twenty seconds. This case was unique, but oc- casionally feints lasting over a minute or even as long as two minutes were obtained. There seems to be no definite relation between the durations of successive feints. They are not grada- tional nor do they vary proportionately in any two individuals examined. TABLE I Duration in Seconds of Ten Successive Feints of Seven Individuals 1 2 3 4 5 6 7 8 9 10 Average B 15 11 22 15 15 10 5 3 10 10.6 C 5 10 15 14 6 11 9 18 10 7 10.5 D 12 7 440 36 31 10 277 2 4 32 85.1 E 10 8 20 7 15 48 15 3 20 14.6 F *0 125 10 13 40 25 8 30 28 20 29.9 G 30 60 6 6 10 35 25 20 5 3 20.0 I *0 *0 13 t4 16 f80 18 15 15 15.1 Average . . . 10.3 31.6 73.7 13.0 17.3 22.1 59.9 12.6 9.7 15.3 26.5 * Failure to respond to stimulation. f Partial feint. Effects of Repeated Stimulation on the Death Feint. Experi- ments were conducted to determine whether the death feint could be induced indefinitely by repeated stimulation. It was found that the feint could not be elicited indefinitely, but the number of times that it could be elicited before the insect ceased to respond varied considerably with different individuals. In some cases the insect would feign death only a few times and then refuse to respond further, and in no case was I able to elicit more than forty successive feints. The results of other workers may be of interest in this connection. ' Holmes found in the case of ten Ranatras successively put into death feints that these were continued without interruption from 9 a. m. to 5 p. m., when the last specimen refused to feign longer. The Severins found it possible in the case of Belostoma to induce feints successively for a total of five hours. The responses of the curculio (Conotrachelus nenuphar Herbst.) were very much less pronounced than was the case of these forms. Feints could not be elicited successively for a period of greater length DEATH FEIGNING REACTIONS IN TYCHIUS PICIROSTRIS 143 than two hours, fifty-three representing the largest number of feints successively produced in a single individual." l The results obtained with Tychius picirostris did not differ markedly from those of Gee and Lathrop with the plum cur- culio, but the period during which the feint could be succes- sively elicited was not so great, varying from a few minutes to somewhat over an hour. Repeated stimulation results in a reduction of the average duration of the feints, although the duration of one of the later feints may be greater than that of some occurring earlier in the succession. The following table gives the average dura- tion of the first ten feints of four individuals as compared with that of all succeeding feints. In computing the averages, cases in which the insects did not respond are omitted. If these were included the average duration of the later feints would be much lower. TABLE II Average Duration (in Seconds) of Earlier and Later Feints D F G I First ten feints 85.1 17.6 29.5 18.6 20.0 5.7 27 4 Feints after tenth 9 5 There is a certain irregularity to be noted in the later feints. With few exceptions, in cases where a long succession of feints was obtained, the first several attempts to produce the feint were successful. Later, however, the number of unsuccessful attempts increased until finally the point was reached at which the feint could no longer be elicited. When the limit of successful response was reached the method of producing the shock stimulus was changed to discover whether this would have any effect in inducing the feint. It was found that in most cases either the insect did not respond, or it re- sponded in the first few trials only. Only in one case did I succeed in obtaining a long succession of feints by changing the method of evoking the feint after the insect had failed to respond to the first means employed, and in this case seventy- six per cen t were partial feints. 1 Gee and Lathrop. Death Feigning in Conotrachelus nenuphar Herbst. Annals Entomological Society of America, 1912, 5, 391-399. 144 E. MELVILLE DuPORTE Another question which suggested itself was whether the insect would recover, its sensitiveness if allowed to rest. It was found that after one or two hours' rest the reaction was in general similar to that obtained by changing the method of producing the shock, except that there was a larger number of fairly long successions. The history of the three following cases will serve to illus- trate more definitely some of the points discussed in this section : Dl. Left for an hour after the fourteenth feint. On resuming the experiment, the first four trials were successful; of the six next, three were successful and after this there was a long series of failures with an occasional success. The method of inducing the feint was then changed and the insect dropped on its back instead of being merely turned over. In the first eight attempts there were five successes but after this, repeated endeavor could elicit no response. El: The feint was produced immediately by the first seven attempts. The fifteen following elicited the feint eight times, and after that the insect feigned death only twice in twenty trials. The method of producing the shock was changed and the insect feigned three times in the first ten trials. I. In the first forty-five trials there were thirty-three suc- cesses and after this a long succession of failures. The insect was left for an hour and the trials were then resumed. The first three attempts were successful, but after these twenty successive attempts elicited no response. A change in the method of stimulating the feint resulted in the production of several partial feints but few T perfect responses. During the feint the muscles of the legs are very tensely contracted and Gee and Lathrop suggest that muscular fatigue eventually occurs as a result of this contraction, thus explain- ing in part the failure to elicit indefinite response. Effect of Mutilation on the Death Feint. Several specimens were placed in the death feint and while in this condition por- DEATH FEIGNING REACTIONS IN TYCHIUS PICIROSTRIS 145 tions of some or all of the legs were snipped off. In most cases the insect remained in the feint, either making no response or slightly twitching the legs. Sometimes, however, the insect -became active on being dismembered. In some cases in which the insect was in a partial feint the shock of removing the tarsi induced a perfect feint. Removing the antennae of the feigning insect almost invari- ably met with no response, the weevil remaining perfectly quiescent. Individuals with the antennae and legs mutilated could always be induced to feign again. Snipping off the beak always resulted in renewed activity on the part of the subject except in one case, when the four tarsi, both antennae and the rostrum were successively removed without any manifestation of activity. It was found impossible to induce the feint until some time after the mutilation, when the shock effects of the operation had passed away. Cutting away the tip of the abdomen always caused the insect to cease feigning. Insects thus mutilated, however, could usually be induced to feign again, though often with more difficulty than in the case of the normal individuals. Insects divided between the thorax and abdomen or between the pro- and meso-, or meso- and meta-thorax always responded by coming out of the death feint. The hinder portion could not again be induced to feign death. In fact it did not live very long after being cut off. The anterior portion, however, gave the death feigning response but the threshhold of stimu- lation was raised by the mutilation for it was necessary to drop it to the table several times, or to tap it several times on the sternum before the death feint could be elicited. In this con- nection it may be advisable to state that only individuals which reacted readily previous to mutilation were used in these tests. Decapitation invariably results in a feverish activity of the limbs. The legs are worked to and fro, and the tarsi rubbed against each other. At room temperature it was found im- possible to obtain the feigning reaction with decapitated indi- viduals. This is probably due to the shock effects of the very violent disturbance in the central nervous system caused by the mutilation, and not to any direct control of the supra-oesoph- ageal ganglion over the death feigning reaction. 146 E. MELVILLE DuPORTE The results obtained with Tychius picirostris agree in general with those obtained by other workers. Influence of Temperature on the Death Feint. Heat. — Insects feigning death were placed on their backs on a warm iron* plate, the temperature of which was not high enough to be unbearable to the tips of one's fingers. Without a single excep- tion the insects immediately became active and attempted to turn over and escape. It was impossible to produce the feint by dropping the insect on the hot surface. When the plate became cooled to a temperature slightly above that of the room, the feint, however, could be induced without difficulty. Insects dropped several times alternately on the warm plate and on the surface of the table feigned death in the latter case but would not on the heated surface. Cold. Several individuals were placed in a refrigerator regis- tering a temperature of 11° C. The feint could be very readily elicited and the duration was much greater than at room tem- perature. The duration of the initial feint varied from 55 seconds to 4 minutes and 35 seconds. The variability of the duration of successive feints was as marked as at room tem- perature, as may be seen from the following figures, which give the duration (in seconds) of the first five feints of one individual : 55, 40, 140, 385, 50 Other individuals were tested on a plate of thin glass placed on a block of ice. The feint lasted for periods ranging from nineteen minutes to upwards of an hour. At the end of an hour those which had not recovered spontaneously were removed to the table, where they immediately became active. On first being placed on the cold plate the insect becomes somewhat restive but does not actively struggle. After a short while the legs slowly and gradually relax, the degree of relaxation varying with the different individuals. In this condition the legs are never pressed against the body but their attitude resembles to some extent that of the limbs of the dead insect. Cold alone, without mechanical shock, can produce a condi- tion similar to that of the death feint. If an insect struggling on its back is placed on the cold glass plate it will after a while cease to struggle, become quiescent and assume a position similar to the feigning insects. This raises the question as to whether DEATH FEIGNING REACTIONS IN TYCHIUS PICIROSTRIS 147 at low temperatures some other factor than those concerned in the production of the death feint does not come into play. As stated above it was found impossible at room temperature to induce the feint in decapitated insects. Near the freezing point, however, the feint was evoked without difficulty. Under the microscope a slight twitching of the legs could be observed in the decapitated insect during the feint. On being removed to the table the insect resumed the twitching movements. The anterior portion of weevils severed across the thorax also feigned death readily, and the duration of the feint was longer than at room temperature. It will thus be seen that cold lowers the threshhold of stimu- lation and reduces the shock effects of mutilation. These results agree closely with those obtained by Gee and Lathrop on Conotrachelus nenuphar Herbst and by Holmes on Ranatra. Effect of Chloroform and Ether on the Death Feint. Insects feigning death were introduced into the vapor of chloroform and of ether. Chloroform seemed to have little effect on the duration of the first feint, but of fourteen individuals tested five could not be induced to feign a second time. The others with one exception could only be induced to feign death twice and this one exception was anaesthetised during the fourth feint. TABLE III The Effect of Chloroform on the Duration (in Seconds) of Feint, and on the Number of Successive Feints of Fourteen Individuals 14 1 2 3 4 5 6 r-r 1 8 9 10 11 12 13 2,3 15 20 *15,4 4,9 15 30 11.14 20 8, 14 2,2 14,6 17 13,7 20,3 * Partial feint. In three cases, stimulated by pinching with forceps while lying on their backs, it was found that the insects would draw up their legs to the feigning attitude, but relaxed them and resumed their struggles on being released from the forceps. In the case of ether the results were different. A longer succession of feints could be obtained, but the ether had the 148 E. MELVILLE DuPORTE effect of reducing the duration of the feints. While the number of successive feints was higher than was the case under the effect of chloroform it was not as high as was obtained under normal laboratory conditions. Only one individual feigned sufficiently often to be killed by the fumes of the ether. TABLE IV Shows the Average Duration in Seconds of the First Feint in a Large Number of Individuals Under Normal Conditions and Under the Influence of Physical and Chemical Agents Normal Heat Cold (near 0° C.) Chloroform Ether 14.2 2020 13.3 5 The Nature of the Death Feigning Instinct. The death feigning instinct occurs in many widely different animals, even among certain birds and mammals. The advantage of the instinct to the possessor is a doubtful one. In the case of insects and other animals, the color of which resembles that of the natural environment, the power of feigning death is an additional safe- guard against the prying eyes of the enemy. But in many cases the death feint is not only useless but may be positively injurious, so that the instinct cannot wholly be regarded as an advantageous adaptation, nor can its origin and subsequent evolution be accounted for on the grounds of usefulness to its possessor. The instinct undoubtedly has its roots in some fundamental physico-chemical reaction of zooplasm. Holmes believes that in the amphipods the death feigning instinct has its foundation in the thigmotactic reactions common among these animals. The Severins also advance the theory that the reaction in Nepa and Belostoma is thigmotactic in nature. The results obtained with Tychius picirostris, in the writer's opinion, corroborate the theory that the physico-chemical reaction responsible for the manifestation of the death feint is of the same nature as that which calls forth the thigmotactic response of many insects and other animals and also of such plants as the Mimosas and other " sensitive plants." It must be borne in mind, however, that the statement that the death feigning instinct is thigmo- DEATH FEIGNING REACTIONS IN TYCHIUS PICIROSTRIS 149 tactic in nature is an explanation which in turn needs to be explained, and a positive explanation is not yet forthcoming. It is probably due, however, to some disturbance in the balance of the metabolic processes causing the excessive elaboration of some substance or substances which incite the contraction of the leg muscles. The feigning reaction may perhaps be resolvable into two components, one manifested in the flexing of the leg appen- dages, the other in the quiescence of the insect. The fact that it is possible to induce a partial feint in which the insect is per- fectly quiet while the legs are not flexed and, in some cases, after treatment with chloroform, to induce the flexing of the legs unaccompanied by the quiescent state, suggests that these two actions may be the result of separable causes, but the data at hand are too slender for a definite pronouncement on the subject. Attention should also be called to the fact that the reaction is segmental and not controlled by the supra-oesophageal gang- lion. The fact that a decapitated insect will not react at room temperature is in no way antagonistic to this view, because the removal of some of the most important trophic centres must necessarily cause some disturbance in the metabolic processes of the nervous system, and the intensity of the stimulus neces- sary to bring about a definite reaction will differ from that necessary under normal conditions; in other words, the thresh- hold of stimulation may be either lowered or raised. One result of this disturbance is manifested in the pendulum and rhythmic movements of the legs of a decapitated weevil, and it it is obvious that this movement cannot coexist with the death feigning reaction as, from their natures, they are mutually exclusive. It will be remembered, however, that cold will so lower the threshhold of stimulation that the decapitated insect will respond readily at low temperature. THE PHYSIOLOGICAL ANALYSIS OF BEHAVIOR* HARRY BEAL TORREY Reed College The facts of behavior make a manifold appeal. Among them all, those which seem especially to attract us pertain to the intricate game of human relations. Whether they take the form of premeditated words or unconscious deeds, elaborate rhetorical statements or direct emotional reactions, whether they suggest a soaring or a grovelling imagination, they challenge our powers of interpretation; they entice us into a search for their meaning. Attentive to their significance for us, we set up social standards, we develop moral codes, and we encourage a science of conduct that shall help us to formulate the courses of action we call ideals. We do all this that we may be the gainers thereby, somehow, somewhere, in the immediate present, the near or the distant future. Our motive is personal. We follow an impulse to self-preservation which reveals itself in all the races of man- kind. Few individuals, indeed, are so thriftless as to fail to take some such thought, crude and sporadic though it may be, for the morrow. Explanations of some sort all but this exceptional and degenerate few must have. For primitive man, spirits and demons, naive myths and legends built on accidental circumstance suffice. For the children of civilization in large part, they suffice also. Nothing more dramatically proclaims the essential similarity of the peoples of the earth, however diverse in degree of savagery and sophistication, than the con- servatism of their mythologies. It is but a few years since a most distinguished British statesman solemnly defended against scientific criticism the authenticity of the unclean spirits that entered into the Gadarene swine. Today whole nations are passionately associating their Gods and all the angels with their several contributions to the inconceivable violation of civilization that is now debauching Europe. 1 Read at Berkeley, California, before Section F, A. A. A. S., August 3, 1915. THE PHYSIOLOGICAL ANALYSIS OF BEHAVIOR 151 II These strange anachronisms are in striking contrast with cur- rent biological theory. Darwin's masterly argument for natural selection as a mechanical theory of adaptation banished forever from scientific speculation the doctrine of special creation accord- ing to Design, and established in its place an interest in the genesis of organic forms and functions according to natural law that has been of inestimable value in the interpretation of the world as we find it. With the publication of the "Origin of Species" appeared new aspects of fitness, new conceptions of conduct, a new and more practical ethic. The problems of behavior gradually assumed new forms under the influence of the com- parative method. New light was thrown on questions of human right and wrong by investigations of the behavior of the lower organisms. Objective studies made head against pre-Darwinian types of speculation and the doctrine of natural selection was applied with increasing freedom to the elucidation of all sorts of organic response. The adoption of a mechanical substitute for the crude teleo- logy which natural selection displaced, marked an enormous advance in methods of investigation at the same time that it gave new meaning to problems of behavior. No mechanical theory of selection, however, is competent to explain the origin of the differences between organisms, their parts, or their ac- tions, on which selection must be based. Such a mechanism of survival is insufficient because it represents results rather than processes. Even in the extreme form in which it appears in the theory of germinal selection, it remains external to the units selected. It is a logical, not a physiological mechanism. Ill It is to the analysis of physiological mechanisms that we turn, then, for further light upon the problems of behavior. At once questions of meaning take a new form. The signifi- cance of behavior, and its development, cease to be of imme- diate concern. For the objective of the physiologist is control of physiological processes, which is at once the pragmatic test of understanding and a practical necessity for the righteous life. It is as a physiologist, then, that I invite your further atten- » 152 HARRY BEAL TORREY tion to the problems of behavior. That we may see the fruits of physiological analysis against a background which will throw them into significant relief, it will be desirable first of all to consider the influence of selection theory on the physiological field. Allusion has already been made to Weismann's application of selection theory to the problem of variation, which as the theory of germinal selection marks its extreme development in this direction. In the field of behavior it has appeared in its most extreme form in connection with the interpretation of the activities of the lower organisms. Here it is found in signifi- cant association with the method of trial. The method of trial was first used in this field as a measure of the learning capacity of the higher animals. As such it has been eminently successful. In the obscure and baffling field of psychology it has encouraged the establishment of objective standards and the prosecution of vigorously scientific investiga- tion. With its extension to the behavior of organisms in gen- eral, certain difficulties have arisen that have more frequently been subjects of dispute than sources of common understanding. Let me illustrate. An interesting parallel to the learning process by which a cat is taught to escape from a puzzle box is the behavior of Paramecium in the presence of an obstruction in its path. The unsuccessful trials that accompany the learning process of the cat are represented in Paramecium by the motor reflexes that follow contact with the obstacle. These reflexes are the means by which the organism is adjusted to the situation. Passing the obstruction is evidence of successful adjustment. To what extent is the behavior of Paramecium a key to the behavior of organisms in general, especially to that peculiar type in many of the lower organisms that is characterized by the direction of the reaction with reference to the source of stimulation ? Is the motor reflex an essential factor in tropic responses ? IV Of the answers which have been given to these questions, the affirmative alone is of present concern. According to this view, the final orientation of a phototropic organism to light THE PHYSIOLOGICAL ANALYSIS OF BEHAVIOR 153 is selected from among several positions occupied at random. Not only are individual responses the material for selection, but those individuals also whose initial responses are especially favor- able to the rapid attainment of success; that is, the fit survive. Now there can be no more doubt of the development of types of behavior with the aid of selection than the development of structural variations with similar aid — and no less. Scanty and uncertain as is the experimental evidence for natural selection, I do not wish to deny that it may have had large influence in organic evolution of whatever kind. This is far from agreeing, however, that the precise responses of organisms to what for the sake of brevity may be called directive stimuli, have been effected, in all or in most cases by selection, whether of individual reactions or of individuals them- selves. On the contrary, evidence is accumulating in favor of the view that organisms respond typically without trials; that what have been called, by a figure of speech, trials, are actually definite responses to stimuli that are neglected by the observer; that behavior is the resultant of many stimuli of which the directive stimulus is but one; that, in fact, the tropic response is rigidly determined as to direction by factors which complete analysis may be expected in all cases to bring to light. V The serious difficulty with a thoroughgoing application of selection theory to behavior is that it so emphasizes certain types of reaction as to divert attention from others, especially from the analysis of those physiological states on which it is generally recognized behavior depends. Nothing illustrates this fact more clearly than the criterion of stimulation that was introduced into studies of behavior with the method of trial. Let us return again to Paramecium. How shall we determine when this organism is under stimulation ? It is obvious that our criterion will depend on our reason for ascertaining the fact. If we are interested primarily with selection among such motor reflexes as Paramecium executes upon contact with ob- structions, a sufficient criterion for pur purpose might well be the reflex itself, however pronounced, that would bring the organism into a recognizably new position with reference to 154 HARRY BEAL TORREY the source of stimulation. And it would continue to be suffi- cient regardless of the agency of stimulation — whether contact or chemical, light, gravity, galvanic electricity — provided only that the motive for it remained the same. It has been insisted, indeed, that in the orientation of the micro-organism Euglena to light, its swervings from the original path are typical evidences of stimulation, being regarded, in fact, as motor reflexes. It follows naturally from such a criterion that when organ- isms such as Paramecium and Euglena, pursue unswervingly a direct course, they are in an unstimulated condition. That this conception, which is entirely compatible with the motive from which it springs, is quite useless in physiological analysis, will not be difficult to show, if it is perhaps not already apparent. It is characteristic of the reaction of Paramecium to a con- tinuous galvanic current that it moves toward the cathode, or anode, as the case may be, along a line of force, under very obvious compulsion. From beginning to end of its course toward the cathode, for instance, the cilia at the cathodal end of the organism beat constantly in a direction the reverse of that in which they beat when swimming freely in the absence of the galvanic current. It is as obvious that they are being subjected to constant stimulation as it is that the organism maintains a constant course. How can such behavior be reconciled with the criterion of stimulation established for the use of the selectionist? It cannot. But the difficulty may be composed by disre- garding it; by asserting, in fact, that since galvanic electricity has never been experienced by Paramecium in a state of nature; since it has never either directly or indirectly provided material for selection; the galvano-tropic response is of no serious im- portance to the student of behavior who is busy with problems of genesis and survival. Such a procedure avoids still another embarrassment, namely, the radical modification of the definition of stimulus that served well enough for the reactions of various organisms to sudden changes in the intensity of what may be called the stimulating medium. When Stent or passes from a shaded to a brightly illuminated area, it executes a motor reflex at the boundary of the shadow much as it would react to contact with a solid object. Reactions of this type have long been recognized and ascribed to a capacity of the organism commonly known as THE PHYSIOLOGICAL ANALYSIS OF BEHAVIOR 155 unterschieds empfindlichkeit or differential sensibility. They are of the nature of trigger effects, depending on the release of internal energy by a sudden change of potential in the stimu- lating medium. It is this change in potential or intensity rather than any addition of energy from without that represents the stimulus in all such cases. Can this conception of stimulus be applied to tropic responses as well ? It is noteworthy that probably the most typical reactions of this type are executed by various organisms when subjected to sudden changes of potential in the galvanic current. Parame- cium, for instance, responds to the making and breaking of the current by characteristic motor reflexes. And this is quite re- gardless of the fact that the galvanic current is unknown to Paramecium in a state of nature and could have played no part in the evolution of its behavior. The responses of Paramecium to the galvanic current are, accordingly, of two types. And these it is quite impossible to reduce to the common denominator of the motor reflex. How is it then with the responses of photosensitive organisms to light, which fall similarly into two groups ? For light, unlike the galvanic current, belongs to the world of nature. It is only exceptionally a laboratory product. Hence its influence on the behavior of organisms cannot be disregarded. Can these two types be reduced to one whose key is the motor reflex in one form or another ? Are the remarkably direct and precise turning movements that characterize tropic reactions resultants of the selection of random movements of this type ? Or are they compelled by the constant application of energy, as is Paramecium in the constant galvanic current ? There can be no doubt that a successful attempt to resolve the difference here presented, in terms of a single fundamental factor, would greatly simplify the problem. But, assuming the existence of such a factor, the view that its application to tropic responses would involve selection of random movements does not square with the fact that various organisms, among them Euglena, the earthworm (Allolobophora) , the sow bug (Porcellio) , and house fly larvae have exhibited unmistakable evidence of responding to light by reactions whose direction can be definitely predicted. 156 HARRY BEAL TORRE Y VI The crux of the matter, however, for the physiologist, lies in the third question. Here the problem appears under a purely physiological guise. It is formulated in physiological terms. It invites attack from the side of physics and chemistry. It suggests a measure of energy transformations. If there is a definite relation between the application of energy to an organism and its behavior, it might be expected to appear in an examination of the evidence for constant stimulation. For the sake of brevity but a few of the recent contributions to this end will be considered, including certain results as yet unpub- lished. They have been chosen so as to outline the gradual emergence of demonstrable fact from a crowd of observations whose real significance could only be surmised in the absence of later experimentation. We may begin with the phototropic behavior of Euglena. It has frequently been observed that when Euglena is exposed to light from two sources the direction of its locomotion may be most delicately adjusted to their intensity as well as to their position, and experimentally controlled with remarkable preci- sion by modifying these two factors. These facts have been interpreted in various ways. To some they strongly suggest, though they do not demonstrate, the presence of constant stimulation. Two years ago, experiments that involved the most careful and detailed observations of the behavior of Euglena that have ever been published, led Bancroft to the conclusion that the reactions of this organism to light are not only of two distinct types but governed by two distinct mechanisms. Of especial significance was the observation, which has been confirmed in my own laboratory, that when Euglena swims directly toward the source of light in its characteristic spiral path, the width of the spiral changes with the distance of the organism from the light. In this orientation, the photosensitive area is pre- sented constantly to the source of light, not intermittently as would be the case were the light falling on the organism from one side. Constant illumination produces a constant response. The variation of the width of the spiral with the intensity of THE PHYSIOLOGICAL ANALYSIS OF BEHAVIOR 157 the light thus constantly playing upon the photosensitive area leaves no room to doubt that the organism is as surely in a condition of constant stimulation as is Paramecium when exposed to a constant galvanic current. Its orientation is forced. Can the relation between the photic energy applied to the organism and the reaction of the latter be expressed in quanti-. tative terms ? Though this has not yet been accomplished for Euglena, experiments have shown that the relation can be ex- pressed for several other organisms in terms of the law origin- ally offered by Bunsen and Roscoe as a formulation of the effect produced by light upon silver chloride. According to this law, stated in more general terms than those originally employed, the effect of the light radiation is proportional to the product of the intensity of the light and its duration, that is, to the total amount of photic energy supplied. The first application of the Bunsen-Roscoe law to biological phenomena appears to have been made by the plant physiolo- gist Blaauw. The results of his investigations on oat seedlings (Avena) and a fungus (Phycomyces) published in 1909, are in striking conformity with the theoretical expectation. It will not be possible here to consider either the details of his experi- ments or their technic. It must suffice to say that in a series of twenty-six experiments on Avena, he varied the duration of the light from 1-1000 of a second to 43 hours, and its intensity in inverse ratio so that the energy delivered should be a constant quantity; and that in spite of this enormous range in the values of these factors, the extreme experimental variation from the average among the entire twenty-six measurements, was but 28 per cent. When it is remembered that the response of an organism to a given stimulus is necessarily complicated by the influence of various other stimuli whose effects may not even be recognized, much less accurately determined, I believe we shall be willing to admit that Blaauw' s experimental results demonstrate the applicability of the Bunsen-Roscoe law to the phototropism of plants. Later experiments have shown the applicability of this law to the phototropism of animals also. Loeb has obtained evidence of its existence in a hybroid {Eudendrium). Ewald has found it to apply to the eye movements of Daphnia under appropriate 158 HARRY BEAL TORREY stimulation. According to Patten there is evidence of it in the behavior of blowflies. And one of my students, G. P. Hays, appears to have demonstrated that the phototropism of the fruit fly Drosophila conforms to it also. This case will be briefly sketched. The imagoes of Drosophila are positively phototropic. The method of experimentation consisted essentially in subjecting a given number confined in a large test tube, to diffused light from a source whose area could be readily controlled by dia- phragms. The flies were first aggregated at the end of the tube away from the light, and then exposed to it, from darkness, for a given number of seconds. The number of flies entering during that time the fifth of the tube which lay nearest the light was adopted as the criterion of the effect of the light. Drosophila is so sensitive to light that it was found necessary to use very low intensities in order that the responses obtained at different exposures might be determined with a sufficient degree of accuracy. Five intensities, multiples of 1, were used, with exposures, varying from 16 to 80 seconds, calculated to deliver the same quantity of photic energy in all five cases. Every day ten trials w r ere made at each intensity, and the average number of flies affected was computed. This was repeated on five succes- sive days with five different sets of flies and the results aver- aged. These averages agree excellently with Blaauw's figures, varying from 1.44 to 2.1. Under the conditions, this is a satis- factorily constant result. I do not see how it can be inter- preted on any theory that neglects the quantity of photic energy involved. If it appears from the facts that have been thus hastily sum- marized that tropisms are not interpretable in terms of random movement and selection processes, must we then conclude that they possess no fundamental characteristics in common with reactions of the non-directive type ? By no means. The essen- tial relationship of these superficially different kinds of reflexes is to be found in the physiological processes that they represent, processes that are evident also in the phenomena of growth, development, differentiation, and all other manifestations of life. As physiological reflexes, tropisms exercise their fascination for THE PHYSIOLOGICAL ANALYSIS OF BEHAVIOR 159 the general physiologist, because they offer certain peculiarly favorable opportunities for physiological analysis. And I think we may say that all reflexes must be thus analyzed if they are to be subject to the control that is demanded by human beings for whom conduct is the one problem of perennial interest and concern. NOTES ON THE FEEDING BEHAVIOR AND OVIPOSI- TION OF A CAPTIVE AMERICAN FALSE SPIDER (Eremobates formicaria Koch) C. H. TURNER Summer High School, St. Louis, Mo. INTRODUCTION Eremobates formicaria is a false-spider, a member of a peculiar group of Arachnids (Solpugidae), in which there is a ten-jointed abdomen and a segmented thorax, and in which the first pair of legs is modified to serve as feelers, so that each individual seems to have three pairs of legs and two pairs of pedipalpi. These are rare invertebrates and are found only in the warmer parts of the world. In our country they occur only in the southern and western portions. We know almost nothing about the habits of our indigenous forms. During the past quarter of a century no original observations upon the behavior of our American forms have appeared. All that has been published upon their activities prior to that time has been condensed by Putnam (16) in the following quotation: — " Regarding the habits of the American Solpugidae comparatively little is known. I have myself collected five individuals of D. pallipes in different parts of Colorado. They were always found, one at a time, in a small cavity in the earth, under a stone or ' buffalo chip.' No other animal life was visible in their vicinity, and they appeared to be quite pugnacious. In my haste to transfer them to alcohol I failed to make any particular observations upon their behavior and mode of life. " At the meeting of the Academy of Natural Sciences of Philadelphia, held No- vember 7th, 1871, Prof. Cope stated, on the authority of Dr. Gerhung, that it (pallipes) was common in (Denver) houses, and was an enemy and destroyer of the Cimex lectularius (bed-bug). In captivity it showed a preference for them as food, and crushed them in its short chelae preliminary to sucking their juices. " Upon sending me two specimens of D. Californicus, Mr. D. Cleveland wrote that they were found at San Diego, in October, under a board near together in a small, funnel-shaped sand hole, with nothing else near them in the way of life, and that they were belligerent. Rev. Green writes that one morning at Silver City, New Mexico, he found a small Goleodes dead and half buried away on the top of his candle. On another occasion, while lying on his bed one very hot night, a large Galeodes ran across his body and down one of his legs, but escaped before he could capture it. Regarding the Texas species (D. suljurea and C. geniculata), Mr. G. W. Belfrage writes as follows: ' They are true nocturnal species, running OVIPOSITION OF A CAPTIVE AMERICAN FALSE SPIDER 161 with great activity about houses at night, and are attracted by the light, and some- times to sugared trees. They are rare, but appear more common in rainy seasons, and only in dark and damp nights. They are probably solitary, and although I have examined everything in the locality where they have been taken for years, I have never found a single one in the day time. When attacked they raise and swing their formidable palpi and show fight, but they are perfectly harmless.' " From Guanajuato, Mexico, Dr. Eugene Duges writes as follows: ' Here there is a Ghtvia which the vulgar call Genisaro, and make out to be excessively ven- omous, which it is not the least in the world.' " Dr. Edward Palmer, speaking of the specimens collected by him in Mexico, says: ' The Solpnga marked San Louis Potosi was running in a garden, the one from Bledos was taken out of my bed in the night, while the other was found under a piece of volcanic rock.' " Dr. R. A. Phillippi, writing from Santiago, Chili, and speaking_ of the two species described by Gervais (C. morsicans and M. variegata) says: " ' They are very common in the streets of Santiago, running with great swiftness in the sun, so that they are called ' aranas del sol,' (spiders of the sun); their bite is said to be very painful, but I know of no personal experience of my friend or my own, unless one case when a young daughter of mine having been bitten in the finger, experienced severe pain, fever, and had the whole forearm swollen for two days, but I do not know whether she was bitten by a Galeodes or by Lathrodectus jormidabilia.' " Meager as these accounts are, so far as I have been able to ascertain, they contain all of the first-hand knowledge we have of the habits of the American false-spiders. THE BURROW* Knowing that the female of an Indian false-spider (17, 11) burrows into the ground to lay her eggs, I confined my captive in a glass jar containing about four inches of dry, compact soil. Although it was broad daylight, she proceeded at once to exca- vate a burrow. With her jaws (chelicerae) and her second pair of legs, she dug out the soil and scraped it backwards. When a pile of dirt had been thus accumulated beneath her body, she shoved it away in the following manner. The ventral sur- face of each chelicera is flat and on a level with the equally flat ventral surface of the thorax (Figs. 2 and 3). The jaws are on the upper anterior (dorso-anterior) portion of each cheli- cera (Fig. 1) and the lower anterior (ventro-anterior) surface of each chelicera is almost truncate (Fig. 3). With the body appressed against the ground and the ventro-anterior surface of the chelicerae pressing against the pile of dirt and the blades of her jaws above it, the solpugid moved forward, pushing the dirt ahead of her. These movements were repeated over and over again until the burrow had been completed. *The false spider which supplied the data for this paper was sent to me alive, from New Mexico, by Mrs. Maud Tanter. 162 C. II. TURNER The floor of the completed burrow sloped inward, forming an inclined plane of about thirty degrees. Having completed her burrow she plugged the opening with soil and remained therein the remainder of the day. This false-spider is nocturnal in her habits. Each night she constructed a new burrow, in the manner mentioned above; but there were variations in both the direction of the burrow and its length. Sometimes, after reaching the bottom of the jar, the cavity turns and continues upward almost to the surface; at other times, by means of an elbow, the burrow is extended in some other direction. At night, on leaving the burrow to forage, she usually closes the mouth with dirt. In about seventy per cent of the cases the burrow was closed during the day. In its habit of excavating a new burrow each night, our New Mexican Eremobates formi- caria differs from the Indian species of Galeodes studied by Hutton (11), which occupies the same burrow throughout the breeding season. FEEDING ACTIVITIES Daily a large number of live insects were placed in the jar of this solpugid. At night, after closing her burrow, she would proceed to slay and eat. She is, indeed, voracious; feeding ravenously and almost indiscriminately upon small moths, small beetles, nymphal locusts, small locusts and grasshoppers, lace- wing flies, crickets, small bugs, caterpillars, flies, gnats, etc. Some she chases down; others she captures by resting quietly until they come within striking distance of her jaws. On the evening of July 30th, between half past nine and half past ten, she captured and devoured four moths, four small beetles, one lace-wing fly and several gnats. When I left at half past ten, she was lying in wait for other prey. With the exceptions mentioned later on, this greediness was exhibited nightly. In 1843, Hutton (11), in discussing a large false-spider (Gale- odes) which he captured in India, says: ' This species is ex- tremely voracious, feeding at night upon beetles, flies and even large lizards. * * * A lizard three inches long, exclusive of the tail, was entirely devoured. * * * The only parts uneaten were the jaws and parts of the skin." J. H. Comstock (6) doubts the correctness of this statement. In 1911 he replies: "Captain Hutton states distinctly that the Galeodes observed by him con- Fig. 1. Dorsal view of the cephalothorax of the false-spider Eremobales formi- caria, Koch. Each division of the scale shown at the bottom is one millimeter. Fig. 2. Ventral view of the cephalothorax of the false-spider Eremobales formi- caria, Koch. The scale is the same as in figure 1. Fig. 3. Lateral view of the cephalothorax of the false-spider Eremobates jormi- caria, Koch. The scale is the same as in figure 1. Fig. 4. Eggs of the false-spider Eremobales formicaria, Koch, photographed twenty-four hours after they were laid. OVIPOSITION OF A CAPTIVE AMERICAN FALSE SPIDER 163 sumed an entire lizard except the jaws and parts of the skin. Other instances in which solpugids are supposed to have eaten their prey are given by Rev. J.J. Wood, in his ' Natural History Illustrated,' and quoted by Murray. Still it is believed that solpugids take only liquid food, which they suck from the bodies of their victims." Since the jar in which my solpugid was confined is only five inches in diameter, frequently the arachnid captured her food near enough to the side of the jar for me to focus my hand lens upon her jaws. Evening after evening, magnifying glass in hand, I have watched those jaws while the false-spider devoured insect after insect. And then, when the arachnid had cast the remains aside, I have fished them out with a crooked wire. Our Eremobates formicaria is not a mere imbiber of juices; she consumes all of her victims except the hard chitin. * Her first pair of large appendages are powerful chelae, each blade of which is armed with stout teeth (Fig. 3). These cheli- cerae are so articulated to the head that, as the blades move dorso-ventrally like those of scissors, the chelicerae move for- ward and back with a saw-like movement. While the right chelicera is moving forward the left is moving backward. When a small insect is captured by the chelicerae, this scissors-saw- like movement of the jaws soon reduces it to a shapeless mass of chitin, while the contents of its body are being consumed. When the Eremobates formicaria encounters a large insect, it usually grabs it, on the ventral side, just back of the head. Once the jaws have secured a hold, the creature is usually doomed. The insect may squirm and struggle and jump, but the false-spider retains its hold. I have seen a locust jump back and forth across the jar several times without getting rid of its. antagonist. In a comparatively short time the scissors- saw-like movements of the jaws make a breech in the integu- ment of the victim. If the thorax is large enough, the jaws are inserted and their scissors-saw-like movements enable the arachnid to soon pulpify and devour much of the contents of the body. After the removal of all of the contents that can be secured in this manner, the lower blade of each jaw is placed on the outside and the upper blade on the inside of the body of the victim, and the scissors-saw-like movements continue. At the same time the carcass is moved from side to side and 1G4 C. H. TURNER back until the chitin has been scraped clean. When the thorax is too small for the jaws to be inserted, from the beginning, the lower blade of each jaw is placed on the outside and the upper on the inside of the body. After the body has been disposed of, the fleshy parts of the legs and even the bases of the wings are subjected to these scissors-saw-like movements of the jaws and passed from side to side between them, until all edible parts have been removed. If the false-spider is not disturbed while eating the carcass left behind is usually nothing but a shapeless mass of chitin. Even when disturbed the result is often the same. Normally this false-spider feeds only on living prey. With dead insects on all sides, I have seen it rest quietly for more than an hour, waiting for living prey. However, it is possible to induce it to eat recently killed insects, if they are properly manipulated. Two large crickets that had been dead about an hour were placed in the jar with the solpugid. One of these was manipulated in such a manner that it moved about and occasionally touched the false-spider. The movements did not resemble in the least those of a living cricket; they were simply jerky and somewhat gentle. Finally the false-spider grabbed the cricket by its posterior end (which end was moving forward) and devoured it in the manner mentioned above. In due time nothing was left but one leg and a shapeless mass of chitin. Although she came in contact with the other dead cricket several times, never once did she even attempt to eat it. I found two living things that my specimen would not eat: hairy caterpillars and spiders. When first confronted with a hairy caterpillar, she sprang to the attack, snapping viciously. At each snap she secured a jaw full of hairs, but the larva was unhurt. She then wiped her chelicerae on the soil, as though trying to clean them. After the first few nights she no longer snapped at hairy caterpillars, but, when they came near her, she would back or rush away. Along with the insects obtained by sweeping I placed live spiders in the jar. She never touched one of them in my presence. Each morning I found prac- tically all of the spiders alive. Whether this immunity of the spiders is due to their elusiveness or to an inherited bias against them on the part of the solpugid is uncertain. Although these false-spiders are nocturnal in their habits and spend much of the day asleep — reclining on either the side, OVIPOSITION OF A CAPTIVE AMERICAN FALSE SPIDER 165 the back, or the sternum — if food strays near them, they rise, slay and eat. Occasionally the burrow was left open in the day-time. On such occasions, on dumping live insects into the jar, some were sure to fall into the burrow. Immediately the false-spider would begin to eat. THE EGGS On the twenty-sixth of July the burrow of this false-spider was built against the glass of the jar in such a manner that practically all of the interior was visible. That evening the solpugid did not come out to forage, and on the following morn- ing she was resting quietly in the bottom of the burrow. Some- time during the day she laid a batch of milk-white, spherical eggs (Fig. 4). Each egg was about 1.7 millimeters in diameter. At six on the evening of the twenty-seventh the eggs were in the bottom of the burrow and the false-spider was resting about halfway between them and the mouth of the burrow. At nine that evening she was busy plugging the mouth of the burrow with soil, taking pains not to include dead insects. Off and on several dead moths got into the dirt she was shoving into the burrow. Each time she caught the moth in her chelicerae and dragged it, backwards, away from the burrow. Later in the night, after having gluttonized, she excavated a new bur-^ row in another part of the jar. For the next two weeks she excavated a new burrow each night. On August the eighth no part of her burrow was exposed, hence it was impossible to see what was transpiring inside. Up to a few minutes after nine that evening she had not appeared on the surface. This was past her usual time of appearing, and since for several days I had been expecting her to lay again, I began to wonder if she were ovipositing. Little by little the soil was carefully removed until the solpugid was exposed in her burrow. She had not laid. Soon after this disturbance she crawled to the surface and moved about sluggishly; but did not feed. I watched her until about eleven o'clock and then went to bed. By the next morning she had oviposited. There were a few milk-white, spherical eggs in the bottom of an ex- ceptionally short burrow; and many more were strewn on the ground near to and on its closed mouth. In the course of a 166 C. H. TURNER few days a few of these eggs had turned yellow and had begun to shrivel into angular shapes. The night following the laying of these eggs, the false-spider came to the surface and fed. The next morning she was resting in a depression of the ground. For several nights thereafter she fed as usual; but did not construct any more burrows. She spent the day in some depression of the ground. On the night of the twelfth she fed heartily; on the morning of the thirteenth she was dead. In both color and size the eggs of our American Eremobates formicaria are similar to those of the Indian Galeodes studied by Hutton (11, 15), but the habits of the mothers are not iden- tical. After laying her eggs in the burrow, the Indian Galeodes rests quietly among them. After they have hatched, she guards the burrow and prevents any intruder from harming her chil- dren. Our American Eremobates formicaria digs a new burrow each night, and lays a second batch of eggs before the first has had time to hatch. Nothing in her behavior suggests that she guards her young from danger. A CONCLUSIONS 1. The American Eremobates formicaria constructs its burrow in the same manner as does the Indian Galeodes studied by Hutton; but instead of using a common burrow throughout the breeding season, the female constructs a new one almost every night. 2. The female of this species is nocturnal in her habits, re- maining in her burrow all day. On retiring to her burrow and in departing therefrom, she usually closes the mouth with dirt. 3. She feeds voraciously upon a large variety of insects, which she secures by chasing them, or by lying in wait until they come to within striking distance of her powerful jaws. 4. Like the Indian Galeodes studied by Hutton, this false- spider eats solid food. Aided by the scissors-saw-like move- ments of her powerful jaws she pulpifies and devours all parts of the captured insects except the chitin. vS. Normally this species feeds upon live insects; but she may be enticed to feed upon recently killed insects that are arti- ficially induced to move. Ability to move seems the main OVIPOSITION OF A CAPTIVE AMERICAN FALSE SPIDER 167 attribute by means of which she distinguishes the living from the not-living. 6. Like the Indian Galeodes she deposits her milk-white, spherical eggs in the bottom of a burrow; unlike the Indian species she closes the mouth of the burrow and leaves the eggs to their fate. 7. She lays, at intervals, more than one batch of eggs a season. 8. Judging from the mother's behavior in captivity, the young are allowed to. shift for themselves ; the Indian Galeodes described by Hutton protects her young. REFERENCES* 1. Banks, N. Synopsis of American Invertebrates. IX. The Scorpions, Sol- 1900. pugids and Pedipalpi. Amer. Nat., XXXIV, 421-427. 2. 1902. Papers from the Hopkins Stanford Galapagos Expedition, 1898- 1899. VII. Entomological Results. (6) Arachnida with Field Notes by Robert E. Snodgrass. Proc. Washington Acad., IV, 49-86, pis. MIL 3. Bernard, H. M. Are the Solpugidae Poisonous ? Nature XLVI, 223. 1892. 4. 1897. Wind Scorpions. A Brief Account of the Galeodidae. Science Pro- gress, New Series, I, 317-343. 5. Birula, A. Bemerkungen ueber die Ordnung der Solifugen. Annuaire Mus. 1903. St. Petersb., VIII, 391-416. 6. Comstock, J. H. The Spider Book, pp. 32-39. 1911. 7. Cook, A. J. Nature and Habits of Pliny's Solpuga. Nature, LVIII, 247 1898. and 320. 8. Faussek. Defensive Attitudes of Scorpion and Soufugid. Biol. Unlersuch- 1906. ungen in Tramskaspien. St. Petersburg, Zap. Russ. Georg. Obsc, XXVII, 1 + 1-192 + 1, 4 Taf. (In Russian. I have not seen this article.) 9. Fischer. Habits of Galeodes. Bombay Nat. Hist. Soc, XX, 886. 1911. 10. Heymons, R. Biologische Beobachtungen an Asiatischen Solifugen nebst 1901. Beitragen zur Systematik derselben. Abh. Ak. Berlin, Anh. i, 65 pp. 11. Hutton, Captain Thomas. Observations on the Habits of a Large Species 1843. of Galeodes. The Ann. and Mag. of Natural History, XII, 81-85. 12. Lonnberg, E. Some Biological Observations on Galeodes and Buthus. Ofv. 1899. Ak. Fork, LVI, 977-983. 13. Lucas, H. On the Eggs of Galeodes. Bull. Soc. Ent. Fr., X, LVII-LVIII. 1891. 14. Marx, G. Contributions to the Knowledge of the Life History of Arachnida. 1891. P. E. Soc. Wash., II, 252. * In this list an attempt has been made to include all papers that have been written upon the habits and life-history of American Solpugidae; and papers on the habits of foreign forms that have appeared during the past twenty-five years. References to earlier papers will be found in Putnam's (16) article. 168 C. H. TURNER 15. Pocock, R. I. On the Genera and Species of Tropical African Arachnida 16 17 1897. Putnam, J. 1883. Pocock, R. 1898. 18. Say, Thos. 1823. of the Order Solifugae, with Notes Upon the Taxonomy and Habits of the Group. Ann. Nat. Hist., XX, 249-272. D. Solpugidae of America. Davenport, Iowa, Proc. Acad. Sci., Ill, 249-310. I. The Nature and Habits of Pliny's Solpuga. Nature, London, LVII, 618-620. Account of Expedition from Pittsburgh to the Rocky Mountains in 1819-20, under command of Major Stephen H. Long, Phila- delphia, vol.11. JOURNAL OF ANI MAL BEHAVIOR Vol. 6 MAY-JUNE No. 3 VISUAL PATTERN-DISCRIMINATION IN THE VERTEBRATES— III EFFECTIVE DIFFERENCES IN WIDTH OF VISIBLE STRIAE FOR THE MONKEY AND THE CHICK H. M. JOHNSON Nela Research Laboratory, National Lamp Works of General Electric Company In an experiment reported some time ago 1 I determined the width of the individual members of a regular system of striae necessary to enable three animals to distinguish the field as striate at a given distance and under the experimental condi- tions prescribed. The work reported in this paper was done on the same animals, in an attempt to ascertain what difference in width of the members of two regular systems of striae, both of which the animals can distinguish as such, is necessary to effect discrimination between the two systems. The discrimi- nands are two interchangeable circular fields, 6 cm. in diameter and equal in brightness and color. The distribution of energy through the visible spectrum approximated that of a tungsten lamp operated at a specific consumption of 1.25 watts per candle. In the work on the chickens the mean brightness of the fields was 12.24 candles per square meter; in the work on the monkey, 6.67 candles per square meter; and in a supplementary test made on two human subjects, 6.24 candles per square meter. The striae composing the system on the positive field are coarser and fewer in number than those composing the system on the negative field. The animals were fed in the food-compartment 1 Johnson, H. M. Visual pattern-discrimination in the vertebrates. II. Com- parative visual acuity in the dog, the monkey and the chick. This journal, vol 4, 1914, pp. 340-361. 170 H. M. JOHNSON indicated by the positive field, and were punished by an electric shock for attempting to enter the food-compartment indicated by the negative field. In an earlier communication 2 I described at some length the optical instruments used, and the methods of preparing the stimuli and of training the animals. 3 I followed these methods rigidly in the present work except in three par- ticulars. It became necessary, when the differences in width between the members of the two systems of striae became small, to make the adjustments by hand, employing the micrometer screw for the purpose, instead of using the string and pulley mechanism. Further, in this work it is unnecessary that each animal be required to compare the test-fields at the same minimal distance as that prescribed for another animal, in order to make the results comparative. I therefore eliminated the stops in front of the alleys A 1 and A 2 of the Yerkes box (shown in figure 6 of the last article cited) which I had used in the work on the chicks. This reduced to 50 cm. the minimal distance at which comparison could be made without a choice being registered. The actual distance usually employed by the birds seemed to be between 50 cm. and 60 cm. If the difference in width of the members of the two systems was large as compared with the least effective difference for the individual, the animals tended to choose without comparing except possibly from the position they happened to occupy in the home-compartment when the exit-door was raised and the test-fields exposed. I retained the plate glass partition formerly used in the work on the monkey, in order to limit his movements. In this work he occasionally thrust his head into one opening in this partition and withdrew it without choosing that alley. Such behavior was relatively infrequent, and in such cases I did not exclude the responses. In practically all the presentations in which he compared the two fields before choosing, he inspected them successively with 2 Johnson, H. M. Visual pattern-discrimination in the vertebrates. I. Prob- lems and methods. This journal, vol. 4, 1914, pp. 319-339. 3 The original drawing for Figure 1 in that article was lost or destroyed after it had been mailed to the printer. A second drawing had to be prepared hastily and was used. It contains an error which I beg the reader, in the interest of clear- ness, to correct. The acute angle I, II, in the figure should be lettered o, and the obtuse an^le I, II, should be lettered p 1 instead of p. The system of right lines bisecting the obtuse angle should be lettered III. On page 330, in the phrase, " the lines III bisect the angle I, II (o)," the symbol o 1 should be substituted for the symbol p. My responsibility for the error is limited to carelessness in copy and proofreading. PATTERN-DISCRIMINATION IN VERTEBRATES 171 his eyes in or quite near the plane of the windows in the parti- tion — 60 cm. from the test-field. In all cases I recorded as the alley "chosen" the one into which the animal first stepped. The animals automatically and instantly registered their choices by breaking the circuit through a signal lamp as they stepped on to the floor of the alley. The third change is the introduction of a shallow copper tray into the Yerkes box, between the exit- door and the alleys A 1 and A 2 . This tray contained a moist- ened felt pad 4 on to which the animal had to step before he could enter the alley. This tended to minimize the variability of the resistance of the animal's feet. It is somewhat more satisfactory than the means employed in my earlier work — that of soaking the animal's feet before beginning the daily series of trials. RELATIVE EASE OF LEARNING FOR INDIVIDUAL ANIMALS The animals used were a young male Capuchin monkey and two Indian gamecocks. They were described in paper No. 2 of this series, and designated severally as Monkey 2 and Chicks 1 and 2. Since that report was published, Dr. P. W. Cobb refracted the eyes of the birds by skiascopy. The static error 4 Breed had used a similar device in his work on the chick. Some such precau- tion is indispensable to good results if punishment with electric shock is employed. The resistances of dry feet and moist feet are of different orders of magnitude. If an animal is put into the box with dry feet and hands, and if during the series the latter become moist, as from perspiration or contact with urine or wet food, the increased effect of a steady line current may work great disturbance. In my work on audition in dogs I found this factor troublesome. Miss E. M. Smith reports a similar experience. The unsatisfactory results reported by some other experi- menters may be partly due to failure to take this precaution. Watson (Behavior, p. 60) asserts that some animals " are extremely resistant even to fairly high cur- rents. The rabbit is not disturbed in the smallest degree by a current which is unbearable to the human being." It should be remembered that thick tufts of long hair cover the soles of the rabbit's feet, forming an excellent insulation when dry. When moistened, the water acts as a conductor. The rabbit will then react violently to a current which is not unpleasantly strong when received by the human subject through the moistened finger-tips. I recently demonstrated this fact in a simple experiment on the rabbit. However, even with the best attempts to keep the resistance of the animal's feet constant, the variability is fairly large — as great, I have been told, as 25%. It is impossible to keep the animal's feet free of dirt and grease, which interfere with good contact with the electrodes. It there- fore seems unnecessary to use such extreme care to control the steadiness of the line current as some students have done. I have obtained satisfactory results by passing A.C. from the city circuit through the primary coil of a Zimmermann inductorium, connecting a bank of lamps in series with the coil and in parallel with each other. Small changes in intensity of shock can be made by changing the position of the secondary coil with reference to the primary. This method has one advantage over the use of D.C. with an interrupter, in that the noise of the latter, which is often a source of disturbance, is dispensed with. 172 H. M. JOHNSON under mydriasis for Chick 1 was 0.75 D. hyperopia in the right eye and 1.25 D. hyperopia in the left eye. For Chick 2 the error was between 0.25 D. and 0.50 D. hyperopia in both eyes. No astigmatism was discoverable in either bird. After the ani- mals had recovered from mydriasis, Dr. Cobb repeated the tests, but, since the birds were continually changing their ac- commodation, the results were quite indefinite. In the emme- tropic human eye an hyperopia as great as that found for Chick 2 usually exists during mydriasis, owing to lack of tone in the ciliary muscle. It seems fair, therefore, to regard this bird's eyes as practically emmetropic, since no error was discoverable without mydriasis. Chick 1 failed to establish a perfect or highly accurate habit in 300 trials, although the width of the members of the positive system of striae was three times as great as that of the members of the negative system. At this point I abandoned the attempt to train him. Later on I gave him the problem of discriminat- ing between two regular systems of striae the members of which were equal in width, but differed in direction by 90°. He failed to learn this problem also. When I introduced a 2 to 1 differ- ence in width in addition to the difference in direction, he estab- lished a perfect habit in 150 additional trials, and maintained discrimination until the relative width of the members of the two systems was reduced to a 5 to 4 relation. The bird's record over 1100 trials indicated that a difference in either width or direction is ineffective if it is presented alone, but that either difference is effective if it is made sufficiently large and is pre- sented with a maximal difference in the other characteristic. This fact raised some questions of great interest, which I was prevented from studying by an injury to the bird's left eye. I shall, therefore, omit a detailed report on the work done on Chick 1. Chick 2 and Monkey 2 learned the problem in less than 100 trials each. BASIS OF DISCRIMINATION In attempting to train an animal to respond to a difference between two stimuli either of which may be made the variable, it is important to know whether the animal is reacting on the basis of relative difference, or by choosing or rejecting a familiar stimulus. If in work on this problem, the animal can be trained PATTERN-DISCRIMINATION IN VERTEBRATES 173 invariably to compare the test-fields and choose the coarser system regardless of the absolute width of its members, it would seem feasible to adopt the method of right and wrong cases, presenting in haphazard order large and small differences in both directions from the standard, and to treat the results statistically. The advantages of using such a method in pre- ference to the cruder methods on which we have thus far had to rely are obvious. TABLE 1 Daily Record of Chick 2 Width (mm.) te of striae on Number of Da Positive Negative Trials Correct 1914 field field given responses i Remarks Feb. 13 2.23 0.92 10 8 Not worked on days for which a 14 2.23 0.92 10 10 no record is given u 15 2.23 0.92 10 9 a 16 2.23 0.92 10 9 a 17 2.23 0.98 10 8 u 18 2.23 0.98 10 8 If 19 2.23 0.98 10 9 ft 20 2.23 0.98 10 8 a 21 2.23 0.98 10 8 a 22 2.23 0.98 10 9 « 23 2.23 0.98 10 9 a 24 2.23 0.98 10 10 a 25 2.23 0.98 10 10 a 26 2.23 0.98 10 10 u 27 2.23 1.00 10 10 ti 28 2.23 1.02 10 8 Disturbed by punishment Mar. 1 2.23 1.02 10 10 a 2 2.23 1.08 10 8 a 3 2.23 1.08 10 10 it 3 2.23 1.11 10 9 it 4 2.23 1.11 10 9 it 5 2.23 1.11 20 15 Greatly excited by punish- ti 9 2.23 0.11 10 10 ment it 11 2.23 0.11 10 10 Responses very slow it 12 2.23 0.11 10 10 Responses very slow it 13 2.23 0.11 10 10 More active ti 14 2.23 0.11 10 10 Active ti 15 2.23 0.11 10 10 Active ti 16 2.23 0.11 5 5 it 16 2.23 1.04 5 5 tt 16 2.23 1.11 10 10 it 17 2.23 1.08 5 4 tt 17 2.23 1.11 5 5 tt 17 2.23 1.15 10 10 tt 18 2.23 1.11 5 5 tt 18 2.23 1.15 5 5 tt 18 2.23 1.18 20 18 tt 19 2.23 1.15 5 5 ft 19 2.23 1.18 5 4 174 H. M. JOHNSON TABLE 1— Continued Width (mm.) :e of striae on Number of Dal Positive Negative Trials Correct 1914 field field given responses Remarks Mar. . 19 2.23 1.22 20 16 <( 20 2.23 1.15 5 4 (i 20 2.23 1.22 5 5 (i 20 2.23 1.28 20 11 Taken as threshold It 20 2.23 0.92 10 10 Given as test of bird's con- (( 21 2.23 1.11 10 10 dition a 21 4.33 2.23 10 3 fChose familiar field against \ punishment a 22 4.33 2.23 10 4 a 23 2.23 1.11 10 8 « 23 4.33 2.23 10 4 |Chose familiar field against \ punishment It 24 4.33 2.23 10 2 It 25 2.23 0.92 10 9 a 25 1.95 0.92 10 10 u 25 1.73 0.92 10 8 a 26 1.95 0.92 5 5 it 26 1.73 0.92 5 5 a 26 1.56 0.92 5 5 U 26 1.42 0.92 10 7 ii 27 1.56 0.92 10 9 u 27 1.42 0.92 5 5 u 27 1.30 0.92 10 7 a 28 1.42 0.92 5 5 a 28 1.30 0.92 10 5 Taken as threshold Apr. 5 2.60 0.98 10 9 « 6 2.60 1.04 10 9 a 7 2.60 1.04 10 8 a 7 2.60 1.11 10 9 a 8 2.60 1.11 5 5 a 8 2.60 1.20 5 5 It 8 2.60 1.30 10 9 it 9 2.60 • 1.39 5 4 It 9 2.60 1.44 5 4 u 9 2.60 1.53 10 10 a 9 2.60 1.73 10 6 Taken as threshold a 15 3.12 1.04 10 10 a 16 3.12 1.04 10 10 It 17 3.12 1.04 10 10 It 18 3.12 1.11 5 5 It 18 3.12 1.20 5 5 It 18 3.12 1.30 5 5 It 18 3.12 1.42 5 4 U 19 3.12 1.42 10 10 a 20 3.12 1.42 5 5 It 20 3.12 1.56 10 8 a 21 3.12 1.56 10 8 a 30 3.12 1.56 20 15 May 1 3.12 1.56 10 8 U 2 3.12 1.56 10 8 It 3 3.12 1.56 10 9 a 3 3.12 1.73 10 8 u 4 3.12 1.30 5 5 u 4 3.12 1.81 10 7 Taken as threshold u 5 3.12 1.30 5 5 PATTERN-DISCRIMINATION IN VERTEBRATES 175 • TABLE 1- -Continued Width (mm.) :e of striae on Number of Dal Positive Neeative Trials Correct 1914 field field given responses Remarks May 5 3.12 1.90 10 6 u 5 3.12 1.04 5 5 a 5 2.60 1.04 5 5 a 6 2.23 1.04 5 5 a 6 1.95 1.04 10 6 Greatly excited by punish- a 7 2.23 1.04 10 8 Still excitable [ment a 11 2.23 1.04 10 9 a 12 2.23 1.04 10 10 u 12 1.95 1.04 10 10 a 12 1.73 1.04 10 8 It 13 1.73 1.04 5 It 13 1.56 1.04 10 8 U 13 1.44 1.04 10 6 Taken as threshold It 13 3.12 1.04 5 5 Taken as test of bird's con- u 14 1.30 0.74 10 10 dition u 14 1.20 0.74 20 16 u 14 1.11 0.74 10 8 u 15 1.30 0.74 10 10 a 15 1.04 0.74 10 6 Taken as threshold TABLE 2 Daily Record of Monkey 2 te Width (mm Positive .) of striae on Negative Number of Da Trials Correct 1914 field field given responses Remarks June 7 1.561 0.780 10 6 Very inactive it 8 1.561 0.780 10 4 a 9 1.561 0.780 10 7 a 10 1.561 0.780 10 7 u 11 1.561 0.780 10 8 Compared in 9 tri< u 12 1.561 0.780 10 9 a 13 1.561 0.780 10 10 It 14 1.561 0.780 10 10 a 16 1.561 0.780 10 10 Not worked on a 17 1.561 0.822 10 10 which no record a 18 1.561 0.867 10 10 It 19 1.561 0.908 10 10 a 20 1.561 0.976 10 10 it 21 1.561 1.041 10 10 u 22 1 . 561 1.115 10 10 u 23 1.561 1.201 10 10 u 24 1.561 1.301 10 10 u 27 1 . 561 1.301 10 9 a 28 1.561 1.301 10 4 u 29 1.561 1.301 10 7 a 30 1.561 1.301 10 7 July 1 1.561 1.301 10 5 a 1 1.561 0.976 5 5 u 2 1.561 0.976 5 5 a 2 1.561 1.115 5 5 days for 176 H. M. JOHNSON TABLE 2— Continued Width (mm e Positive .) of striae on Negative Number of Dat Trials Correct 1914 field field given responses Remarks July 2 1.561 1.301 10 8 Taken as threshold u 3 2.602 1.561 10 6 a 4 2.602 1.561 10 7 it 5 2.602 1.561 10 8 a 6 2.602 1.561 10 10 u 7 2.602 1.561 10 10 u 8 2.602 1.561 10 10 a 9 2.602 1.561 10 10 a 10 2.439 1.561 10 10 a 11 2.439 1.561 10 10 u 12 2.296 1.561 10 10 a 13 2.296 1.561 10 10 u 13 2.168 1.561 10 10 a 14 2.168 1.561 10 9 a 14 2.001 1.561 10 10 a 15 1.904 1.561 10 9 a 16 1.904 1.561 20 19 a 17 1.904 1.561 10 9 a 17 1.815 1.561 10 10 a 18 1.815 1.561 10 10 a 18 1.734 1.561 10 6 a 19 1.815 1.561 10 9 a 20 1.815 1.561 10 9 a 21 1.774 1.561 10 7 a 24 1.774 1.561 10 7 Taken as threshold Aug. 10 0.780 0.446 10 6 In bad condition since July a 11 0.780 0.446 10 4 Refused food [24 it 12 0.780 0.446 10 9 U 13 0.780 0.446 10 7 « 14 0.780 0.446 10 10 « 15 0.780 0.446 10 10 <( 16 0.780 0.446 10 10 It 17 0.780 0.520 10 9 (i 18 0.780 0.520 10 9 a 19 0.780 0.538 10 10 (i 20 0.780 0.558 10 10 u 21 0.780 0.578 10 8 u 22 0.780 0.520 3 3 It 22 0.780 0.578 2 2 a 22 0.780 0.600 10 10 it 23 0.780 0.520 2 2 it 23 0.780 0.558 3 3 a 23 0.780 0.624 10 10 u 24 0.780 0.520 2 2 a 24 0.780 0.558 3 3 it 24 0.780 0.650 10 10 it 25 0.780 0.520 2 2 u 25 0.780 0.558 3 3 « 25 0.780 0.679 12 7 a 26 0.780 0.520 1 1 u 26 0.780 0.558 2 1 it 26 0.780 0.600 5 5 it 26 0.780 0.650 3 3 « 26 0.780 0.655 10 8 PATTERN-DISCRIMINATION IN VERTEBRATES 177 TABLE 2— Continued te Width (mm Positive .) of striae on Negative Number of Da Trials Correct 1914 field field given responses Aug . 27 0.780 0.600 10 9 ti 27 0.780 0.661 10 9 a 28 0.780 0.600 10 9 tt 28 0.780 0.667 10 10 u 29 0.780 0.600 10 10 a 29 0.780 0.673 10 5 a 30 1.040 0.780 10 10 a 31 0.918 0.780 10 10 sept. 1 0.918 0.780 10 10 ii 1 0.909 0.780 10 9 a 2 0.918 0.780 10 10 a 2 0.897 ' 0.780 10 10 tt 3 0.918 0.780 10 9 tt 3 0.890 0.780 10 7 a 4 0.650 0.520 10 10 it 5 0.624 0.520 10 10 « 6 0.624 0.520 10 10 ii 6 0.615 0.520 10 8 a 7 0.624 0.520 10 8 a 7 0.610 0.520 10 7 tt 8 0.624 0.520 10 9 a 8 0.605 0.520 10 7 it 9 0.520 0.390 10 9 a 10 0.520 0.400 10 9 it 11 0.520 0.411 10 10 u 13 0.520 0.422 10 9 tt 13 0.520 0.434 10 10 ii 15 0.520 0.446 10 9 u 15 0.520 0.459 10 7 it 16 0.520 0.446 10 8 a 16 0.520 0.467 10 8 it 17 0.520 0.434 10 8 tt 17 0.520 0.473 10 9 tt 18 0.520 0.434 10 10 tt 18 0.520 0.479 10 7 tt 19 0.520 0.434 10 10 it 19 0.520 0.488- 10 6 tt 20 0.459 0.390 20 17 tt 21 0.459 0.390 10 10 it 21 0.446 0.390 10 8 ii 22 0.459 0.390 10 10 ii 22 0.441 0.390 10 8 ii 23 0.459 0.390 10 10 ii 23 0.436 0.390 10 9 it 25 0.459 0.390 9 9 ii 25 0.431 0.390 11 10 « 27 0.459 0.390 10 10 ti 27 0.427 0.390 10 7 a 28 0.459 0.390 10 10 ii 28 0.422 0.390 10 10 ii 29 0.459 0.390 10 10 ti 29 0.417 0.390 10 9 tt 30 0.459 0.390 10 10 a 30 0.413 0.390 12 8 Remarks Taken as threshold Taken as threshold Taken as threshold Taken as threshold Taken as threshold 178 H. M. JOHNSON TABLE 2— Continued te iVidth (mm .) of striae on Number of Da Positive Negative Trials Correct 1914 field field given responses Oct . 1 0.459 0.390 10 9 u 1 0.409 0.390 10 7 u 2 0.390 0.325 10 10 u 2 0.390 0.339 10 7 II 3 0.390 0.339 10 9 a 4 0.390 0.339 10 10 u 5 0.390 0.339 10 10 u 5 0.390 0.347 10 9 a 6 0.390 0.339 10 10 tt 6 0.390 0.355 10 8 u 7 0.390 0.339 10 8 u 7 0.390 0.358 10 10 it 8 0.390 0.339 10 10 u 8 0.390 0.361 10 7 a 9 0.390 0.339 10 10 u 9 0.390 0.365 10 9 a 10 0.390 0.339 10 10 if 10 0.390 0.368 10 10 u 11 0.390 0.339 10 9 a 11 0.390 0.371 10 7 it 12 0.390 0.339 10 9 ii 12 0.390 0.371 10 7 ii 13 0.347 0.312 20 19 u 17 0.347 0.312 20 15 ii 19 0.372 0.312 10 8 ii 19 0.347 0.312 10 7 lov. 11 0.390 0.312 20 17 ii 12 0.390 0.312 20 15 ii 13 0.390 0.312 20 15 ii 14 0.390 0.312 15 12 ii 15 0.390 0.312 15 15 u 16 0.434 0.312 20 16 ii 17 0.422 0.312 20 20 ii 18, 0.411 0.312 10 10 it 18, 0.400 0.312 10 9 it 19' 0.390 0.312 20 17 it 20 0.381 0.312 20 20 it 21 0.372 0.312 10 10 it 21 0.363 0.312 10 10 it 22 0.354 0.312 10 10 ti 22 0.346 0.312 10 9 tt 23 0.346 0.312 10 10 it 23 0.339 0.312 10 8 it 24 0.346 0.312 10 8 u 24 0.335 0.312 10 9' a 25 0.346 0.312 10 9 it 25 0.332 0.312 10 9 ti 26 0.346 0.312 10 9 tt 26 0.328 0.312 10 9 it 27 0.346 0.312 10 10 it 27 0.325 0.312 10 10 a 28 0.346 0.312 10 8 u 28 0.321 0.312 10 8 it 29 0.346 0.312 10 9 Remarks Taken as threshold Distracted bv noise outside PATTERN-DISCRIMINATION IN VERTEBRATES 179 TABLE 2 — Continued te Width (mm Positive .) of striae on Negative Number of Dai Trials Correct 1914 field field given responses Remarks Nov . 29 0.321 0.312 10 5 Taken as threshold u 30 0.312 0.260 20 16 Dec. 1 0.312 0.260 10 10 It 1 0.312 0.264 10 10 a 2 0.312 0.269 10 10 It 2 0.312 0.274 10 10 a 3 0.312 0.279 10 10 u 3 0.312 0.284 10 10 a 4 0.312 0.289 10 9 U 4 0.312 0.294 10 9 a 5 0.312 0.284 10 9 a 5 0.312 0.298 10 9 it 6 0.312 0.28-4 10 10 It 6 0.312 0.300 10 8 it 7 0.312 0.284 10 9 a 7 0.312 0.304 10 8 Taken as threshold It 8 0.312 0.284 10 10 a 8 0.312 0.306 10 5 a 8 0.312 0.284 10 8 fControl test to determine \ basis of choice a 8 0.284 0.260 10 10 a 9 0.244 0.223 10 9 It 9 0.240 0.223 10 10 a 9 0.236 0.223 10 9 a 10 0.244 0.223 6 6 It 10 0.234 0.223 10 8 a 10 0.232 0.223 15 11 a 11 0.244 0.223 10 9 it 11 0.232 0.223 10 7 Taken as threshold u 11 0.223 0.199 10 9 u 12 0.223 0.199 10 9 a 13 0.223 0.199 10 9 u 13 0.223 0.200 10 10 u 14 0.223 0.200 10 10 It 14 0.223 0.203 10 10 a 15 0.223 0.203 10 10 a 15 0.223 0.208 10 9 u 16 0.223 0.203 10 9 It 16 0.223 0.210 13 9 u 18 0.223 0.203 10 10 a 18 0.223 0.210 10 8 it 19 0.223 0.203 10 10 a 19 0.223 0.210 10 8 Taken as threshold it 19 0.195 0.173 10 9 it 20 0.195 0.173 10 9 u 20 0.192 0.173 10 7 it 21 0.195 0.173 10 8 a 21 0.192 0.173 10 9 It 22 0.200 0.173 10 9 u 22 0.190 0.173 10 7 u 23 0.200 0.173 10 9 It 23 0.190 0.173 15 11 Taken as threshold 1915 Mar. 4 0.780 0.520 20 19 u 5 0.780 0.538 10 10 180 H. M. JOHNSON TABLE 2— Continued te ^Vidth (mm.) of striae on Number of Da Positive Negative Trials Correct 1914 field field given responses Kan •. 5 0.780 0.558 10 10 « 6 0.780 0.577 10 10 a 6 0.780 0.600 10 10 a 8 0.780 0.624 10 9 a 8 0.780 0.650 10 10 a 9 0.780 0.678 10 10 a 9 0.780 0.709 10 9 a 10 0.780 0.709 10 8 a 10 0.780 0.723 10 8 « 11 0.780 0.678 10 8 a 11 0.780 0.729 10 8 « 12 0.780 0.678 10 9 (i 12 0.780 0.731 10 10 (( 13 0.780 0.678 10 10 U 13 0.780 0.743 10 8 a 14 0.780 0.678 10 10 a 14 0.780 0.750 10 7 a 15 0.780 0.678 10 10 a 15 0.780 0.750 10 7 « 16 0.975 0.780 10 9 a 16 0.918 0.780 10 10 a 17 0.867 0.780 10 9 a 17 0.848 0.780 10 9 a 18 0.867 0.780 10 10 a 18 0.839 0.780 10 9 u 19 0.867 0.780 10 9 u 19 0.830 0.780 10 9 u 21 0.867 0.780 10 10 u 21 0.822 0.780 10 9 a 22 0.867 0.780 10 10 n 22 0.813 0.780 10 10 a 23 0.867 0.780 10 9 a 23 0.805 0.780 10 9 u 25 0.867 0.780 10 9 a 25 0.805 0.780 10 9 a 26 0.867 0.780 10 9 a 26 0.796 0.780 10 6 a 27 0.867 0.780 10 9 « 27 0.796 0.780 10 6 Remarks Taken as threshold Taken as threshold If the reader will now examine the daily records of Chick 2 and Monkey 2, shown in tables 1 and 2 respectively, he may observe that the chick was usually disturbed when he was re- quired to avoid a field which he had recently been in the habit of choosing. The records for 21 to 24 March, 1914, exhibit this fact strikingly. Monkey 2 tended to respond in this way in the earlier stages of the work, but later overcame the ten- dency. On December 8, 1914, I made a control test which tends to make the fact clear. In the week immediately pre- PATTERN-DISCRIMINATION IN VERTEBRATES 181 ceding this test I had required him to choose a field each stripe on which was 0.312 mm. wide, and to reject a field each stripe on which was 0.284 mm. wide. He chose correctly in 48 out of 50 trials. In this control test I presented the 0.284 mm. system ten times with a system each member of which was 0.312 mm. wide, and ten times with a system each member of which was 0.260 mm. wide. The animal was required to choose the coarser system at each trial, thus choosing the 0.284 mm. system ten times and rejecting it ten times in the same daily series. The results are shown in table 3. TABLE 3 Width (mm .) of striae on Relative position Field chosen Positive Negative of positive by the ?rial field field field animal Remarks 1 0.312 0.284 Right Right Compared before choosing 2 0.312 0.284 Right Right a it a 3 0.312 0.284 Left Left a it tt 4 0.284 0.260 Left Left tt tt it 5 0.284 0.260 Right Rieht tt (i tt 6 0.284 0.260 Right Right a a u 7 0.312 0.284 Left Left u it a 8 0.312 0.284 Left Left a tt u 9 0.284 0.260 Rieht Right u a it 10 0.284 0.260 Right Right a tt a 11 0.312 0.284 Left Left a tt tt 12 0.312 0.284 Right Right a tt it 13 0.312 0.284 Right Right it tt it 14 0.284 0.260 Left Left a it it 15 0.284 0.260 Left Left a a tt 16 0.284 0.260 Right Right a a tt 17 0.312 0.284 Right Left Chose without comparing 18 0.312 0.284 Right Left it a tt 19 0.284 0.260 Left Left a tt tt 20 0.284 0.260 Left Left a a It The animal was punished for choices 17 and 18. He had usually shown a preference for the left-food-compartment. In trials 17, 18, 19, and 20 he inspected the left field only. The field presented on that side at these trials was the field which he had correctly chosen in trials 14, 15 and 16. At trials 1 to 16, inclusive, he inspected both fields before choosing, and in every case he selected the relatively coarser system, regardless of the absolute width of its members. I believe that this animal could have been adapted by training to study by the standard method of right and wrong cases. Chick 2 gave no such promise. 182 H. M. JOHNSON The tendency of Chick 2 to respond by choosing or rejecting the familiar stimulus differs from the behavior of some birds used by Bingham, 5 which chose the larger of two circles and avoided the smaller without regard to the absolute size of the larger, and without regard to its previous association with re- ward or punishment. The comparison is worthy of mention as it seemingly points to an individual difference. 6 Bingham, how- ever, gives no details which indicate how nearly unanimous or how consistent his birds were in manifesting this form of behavior. AN INCIDENTAL OBSERVATION One incidental feature of the behavior of Monkey 2 seems deserving of special mention. As was remarked above, I used a plate glass partition across the entrances of Alleys A 1 and A 2 of the Yerkes box. This partition contained two rather small holes through which the animal had to squeeze himself in order to enter the alley. On two occasions — June 16, 1914, and March 14, 1915, I neglected to insert this partition before giving the first trial of the daily series. On each occasion the animal refused to enter either alley. When I looked into the box to ascertain the cause of his delay in responding I found him grop- ing in large semi-circles with his hands near the plane in which the glass partition belonged, and uttering frequent vocal excla- mations. It was necessary to recall him to the home-com- partment and insert the partition. It appeared from this be- havior that he had become habituated to disregarding the par- tition as a visual object. RELATIVE EASE OF DISCRIMINATION In tables 4 and 5 I have summarized the values taken as "thresholds" for the two animals. The reader may see how these values were obtained by referring to tables 1 and 2, in which the animals' daily records appear. Whenever there 6 Bingham, H. C. Size and form perception in Gallns domesticus. This journal, vol. 3, 1913, pp. 65 ff. 6 Watson (Behavior p. 367) refers to the difference between the behavior of my chick and those of Bingham's as indicating that " this (Bingham's) observa- tion cannot be confirmed." This interpretation is not mine. Had all the birds been worked on the same problem I should not have considered that the behavior of one bird was predictable from the behavior of a few others. But the two prob- lems are so different that there is little basis for comparison. PATTERN-DISCRIMINATION IN VERTEBRATES 183 seemed room for doubt whether the animal's errors were due to the magnitude of the stimulus-difference or to disturbance from some other cause, I presented a larger difference at a number of trials in the same daily series with the smaller difference. If all or most of the incorrect choices were made at the small stimulus-difference, I concluded that discrimination at that stim- ulus-difference was becoming difficult; if the animals' percentage of correct choices was low for the large stimulus-difference as well, I assumed that the source of disturbance was extraneous. The values obtained on the two animals are not strictly in- tercomparable. In work on the monkey I reduced the stimulus- difference by smaller gradations than I could use in the work on the chick, owing to a limitation of the optical instrument by which the field was formed. As the angle of rotation of the gratings over each other becomes very small, as is the case where the width of the visible stripes thus formed is large, a very slight change in the angle of rotation makes a large differ- ence in the width of the visible stripes. In the work on the monkey a larger angle and smaller gradations could be em- ployed, since he was sensitive to much smaller widths than was the chick under similar conditions. I took for "threshold-differ- ences" for the monkey the differences at which his average per- centage of correct choices most nearly approximated 75. For the chick I took the stimulus-difference at which the first break- down of discrimination not apparently due to disturbance from other causes occurred. This procedure is open to criticism in that I did not ascertain to what extent the bird could be made to overcome his uncertainty by continued training. At the time, however, this bird did not react well under punishment. As soon as the stimulus-difference became relatively small he usually refused to inspect both fields and adopted a position- habit. I recognized this defect in the procedure, and in a later piece of work attempted to settle the question. 7 I found that this bird could be made to improve to a limited extent after very long continued training. The degree of improvement which I was able to elicit is not sufficiently large to invalidate the results herein presented as the rough approximations of the 7 Johnson, H. M. Visual pattern-discrimination in the vertebrates. IV. Effec- tive differences in direction of visible striae for the monkey and the chick. To appear in this journal. 184 H. M. JOHNSON limits of the bird's discriminative ability which I regard them as being. It will be seen that the threshold- values for the chick vary irregularly from 33% to 42% of the width of the striae on the standard field. The variations may be explained by assuming that discrimination was difficult throughout this region of stimulus-differences. If the bird received punishment several times in close succession shortly after the difference approached this region, he "gave up" earlier, and yielded a larger "thresh- old" than if his errors were more widely distributed. Table 5 also shows a great disparity between the "upper" and the "lower" threshold values given by the monkey when the width of the striae on the standard field was 0.520 mm. This fact pointed to a large effect of practice. The magnitudes of the thresholds obtained at the smaller values of the standard stimulus are so much lower than those obtained in the earlier stages of the work at the larger stimulus- values, that it was necessary to make a control test to discover if these differences were not due to the effect of training, instead of being a function of the absolute w T idth of the striae. This test was made between March 4, 1915 and March 27, 1915. It shows quite clearly that the differences were due to the effect of training. The results suggest strongly that if training had been continued sufficiently long after the full effect of practice had been obtained, the values for all the difference-thresholds where the striae on the standard field were over 0.3 mm. wide would have borne a rela- tion to the absolute width of the members of the standard sys- tem analogous to Weber's law for brightness. This relation probably does not hold for absolute widths below 0.2 mm., under these experimental conditions, since such fine systems become increasingly hard for the monkey to distinguish as striate. Due to pressure of other work and the small likelihood of Monkey 2 living through many more months, I did not feel justified in carrying this exploratory study farther at the time. Should a similar study ever be made, it would seem advisable to select fewer points at which to determine the animal's thres- hold, and to give a large number of presentations — several hundred, at least — of a number of differences in either direction from each point. In work on another problem with this animal I found such procedure quite fruitful. PATTERN-DISCRIMINATION IN VERTEBRATES 185 TABLE 4 Threshold Conditions for Chick 2 Width of striae on Difference per cent width on standard Mean of upper and lower Positive Negative field field field thresholds *2.23 mm. 1.28 mm. 42 1 . 30 " *0.92 a 41 *2.60 " 1.73 a 33 *3.12 " 1.81 u 42 1.44 " *1.04 it 38 1.04 " *0.74 « 40 Remarks TABLE 5 Threshold Conditions for Monkey 2 1 . 774 mm. *1.561 " 0.890 " *0.780 " 0.610 " *0.520 " 0.413 " *0.390 " 0.321 " *0.312 " 0.232 " *0.223 " 0.190 " 0.796 " *0.780 " *1.561 mm. 1 . 301 " *0.780 " 0.673 " *0.520 " 0.479 " *0.390 0.371 *0.312 0.304 *0.223 0.210 *0.173 *0.780 0.750 14 17 14 14 17 8 5.9 4.9 2.9 2.6 4 5.8 9.8 2 3.8 15.5 14 11.5 5.4 2.8 4.9 Note training-effect 2.9 Note training-effect * Standard stimulus. COMPARISON WITH HUMAN SUBJECTS A comparison of the relation between the absolute stimulus- value (expressed in terms of width of striae on the standard field) and the minimal effective differences for the animals with that obtained for human subjects is of some interest. Accord- ingly I determined this relation for two human subjects by means of the method of limits, using the same visual conditions as for the animals. The magnitudes of the thresholds obtained on the human subjects by the method of limits are not to be com- pared with those obtained by the discrimination-method on the animals. A threshold-value obtained by the method of limits 186 H. M. JOHNSON is generally smaller than one obtained under the same conditions by the method of right and wrong cases, of which the discrimi- nation-method is a special and rough adaptation. The relation between the absolute value of the standard stimulus and a series of thresholds obtained by either method ought to vary in the same way. This comparison between the two sets of data is the only one which I wish to be made directly, although the values obtained for the monkey and for the human subjects indicate that their sensitivities are of the same order. The two human subjects were Dr. A. G. Worthing (W), a member of the research staff of this laboratory, and Mr. B. E. Shackelford (S), of the University of Chicago, late Brush fellow in this labo- ratory. Both observers are physicists possessed of a high degree of skill in optical pyrometry — one of the most difficult types of photometric measurements. Perhaps it should be remarked that both observers considered the visual conditions in the present experiment quite trying. An hour or more was usually required for making ten paired readings, which constituted a single sitting. The results are summarized in table 6. TABLE 6 Width of striae on standard field (mm.) 0.780{W" °- 390 {s Y ' 0.260|^' 0.195 fW. (s. Upper threshold in mm. 0.026 0.019 0.007 0.014 0.009 0.006 0.006 0.009 Lower threshold in mm. 0.020 0.012 0.009 0.013 0.008 0.008 0.006 0.008 Mean threshold per cent standard 3.0 2.1 2.1 3.5 3.3 2.7 3.1 4.7 Mean variation per cent mean threshold 581 45/ 28\ 35/ 40\ 27/ 49 34 Mean threshold for the two observers 2.6 2.8 3.0 3.9 The relation between absolute width and magnitude of the threshold for the animals and the two human observers is shown graphically in the accompanying figure. The average threshold for the two observers at each stimulus-value was taken merely for convenience in plotting on the scale selected. SUMMARY A very large effect of practice was found in the work on the monkey, which indicates the desirability of modifying the dis- crimination-method for use on the higher mammals. PATTERN-DISCRIMINATION IN VERTEBRATES 187 When the full effect of practice has been obtained, Monkey 2, under optimal conditions, can distinguish differences in width of striae of less than 3%. These values are of the same order of magnitude as those obtained by the method of limits on two human observers possessed of unusual skill in photometry. Chick 2 ceased to discriminate when the difference in width of striae was reduced to a value between 33% and 42%. The relatively poor results of the chick were not due to errors of refraction, as both his eyes were emmetropic. The discriminative ability shown by the monkey is on the average roughly ten times as great as that shown by Chick 2. His visual acuity, however, is only four to five times as good as that of the same bird. This disparity suggests that differ- ence of width between two systems of visible striae constitutes a more difficult basis of discrimination for the chicken than the mere presence or absence of the striae. The fact that Chick 1 did not learn the problem, although width-difference was effec- tive for him when presented with an ineffective difference in direction, lends support to this belief. Nothing in the above work indicates that Chick 2 might not have yielded a lower threshold had training been sufficiently prolonged. A later experiment, however, indicates that the chick's susceptibility to improvement under prolonged training is not sufficiently large to affect the order of difference between his results and the monkey's which appears in this work. For all the subjects, both human and animal, the relation between absolute size of detail and effective difference in size approximates an analog of Weber's law. For the chick, familiarity was a more effective stimulus- characteristic than relative size of detail, and the bird never overcame the tendency to respond on that basis without con- tinued retraining. The monkey eventually learned to respond on the basis of relative size. His results suggest that he is adaptable to study by a method of much greater precision than the method which was actually employed. In conclusion I wish to thank Dr. P. W. Cobb for the work of refracting the eyes of the animals, and for making the photo- metric determinations for me; also, Dr. A. G. Worthing and Mr. B. E. Shackelford for their cheerfulness in undertaking the tedious observations. 188 H. M. JOHNSON Least effective difference per cent standard a A \ \ \ \ / / 4 1 I 1 \ ( ? \ \ \ > / \ a : <\ i A to !u / \ ^ \ X h — c4 i -u- or s 2 -a 5 re ■§ c s-. re 7s"*- 1 O o 2 « S 03 ai -^ U 2 u o +j T3 U > "o CO cu H f 1 CU CO S« o CNOJ 03 >>>,£ CU CD 3 -*^2 §§o SSB b i< ii o o o ^-t V4-. %-t CO CO CO CU CU CD re 03 re > > > T3 X3 *C 'o'c'o CO CO CO CU CU CU >- Ui l-c +-> +J +J S c G re re re CU CU CU > > > l_ 1_ L_ •_ 3 3 3 3 uuuu VISUAL PATTERN-DISCRIMINATION IN THE VERTEBRATES— IV EFFECTIVE DIFFERENCES IN DIRECTION OF VISIBLE STRIAE FOR THE MONKEY AND THE CHICK H. M. JOHNSON Nela Research Laboratory, National Lamp Works oj General Electric Company The purpose of the experiments described below was to ascer- tain how small a difference in direction of two regular systems of visible striae will condition the responses of animals possess- ing eyes of different types. In this work I used the "discrimi- nation method" described by Yerkes and Watson, 1 feeding the animal in a compartment of the training box indicated by the "positive" test-object, and punishing him for attempting to enter the food-compartment indicated by the "negative" test- object. A detailed description of the method of training and of preparing the stimuli has appeared in an earlier report. 2 In the present work the discriminands were two circular fields, 6 cm. in diameter and 6.24 candles per square meter in bright- ness. The energy-distribution in the visible spectrum approx- imated that of a tungsten lamp operated at a specific consump- tion of 1.25 watts per candle. Both fields were covered with alternate dark and bright striae equal in width and separation. The chicks and the monkey had previously demonstrated their ability to react with perfect accuracy under the experimental conditions to the presence or absence of the striae from the test-field. The striae on the positive field were horizontal; those on the negative field deviated from the horizontal direction by a definite and variable amount. In the beginning of the train- ing the difference was 90°, and was decreased by small grada- tions after the animal had formed the habit of invariably choos- ing the food-box indicated by the positive test-field. Yerkes, R. M. and Watson, J. B. Methods of studying vision in animals. Behavior Monographs, vol. 1, no. 2, 1911. 2 Johnson, H. M. Visual pattern-discrimination in the vertebrates. I and II. This journal, vol. 4, 1914, pp. 319-361. 189 190 H. M. JOHNSON THE ANIMALS USED The four animals used in the present study were described in the earlier report just cited, and were therein designated severally as Dog 1, Monkey 2, and Chicks 1 and 2. Dog 1 failed to learn the problem. This experiment is a part of a special study of that animal's vision. His results are interesting only when considered in connection with those obtained in the other work, so I shall reserve them for a separate report. 3 Chick 1 also failed to learn the problem in 800 trials. He had a large refractive error (1.5 D. hyperopia) in the right eye, and was also very easily excited by punishment or by unusual noise, so that he would refuse to work in the experiment -box. Event- ually he was blinded in the left eye during a fight, and I did not resume experimentation on him. On account of lack of space I shall not include his daily records in this report. Chick 2, both of whose eyes were emmetropic, had distinguished as striate a system each member of which was 0.78 mm. wide at a distance of 60 cm. from the eye; the width under threshold- conditions for Monkey 2 being 0.163 mm. at the same distance. Chick 2 had also yielded a difference threshold for width of visi- ble striae of 33% to 42%, and Monkey 2, a series of thresholds as low as 3% under optimal conditions. In the course of the work some serious interruptions occurred. In September, 1914, Chick 2 acquired a severe white diarrhoea, and became greatly weakened. He finally refused to eat, and all the masculine characteristics of his behavior disappeared. I transferred him to an out-door yard where he had plenty of range. He showed improvement in a few days, and made a complete recovery in three months. Meantime his comb grew very large and depended over the left eye, largely occluding the visual field on that side. I amputated the comb March 6, 1915, and resumed retraining on the 10th. On the 20th he became engaged in a fight and was nearly killed. Other uncon- trollable conditions then intervened and prevented the resump- tion of work until June 16, 1915. Monkey 2 — a frail and probably tubercular little animal — also sickened May 21, 1915. He refused food and in two days be- came too weak to stand alone. I moved him to a room where 3 Johnson, H. M. Visual pattern-discrimination in the vertebrates. V. A demon- stration of the dog's deficiency in detail-vision. To appear in this journal. PATTERN-DISCRIMINATION IN VERTEBRATES 191 he could be kept in direct sunlight most of the day, and used forcible feeding on milk and raw eggs for ten days, during which time his condition began to improve rapidly. He became quite wild as a result of such handling, and I dared not resume work with him until June 15, 1915. On August 1 following, he fell from a steam-pipe six feet from the floor and sustained a par- tial fracture of the right femur and the left tibia. He made a fair recovery, but is still occasionally excited by unusual handling. While these accidents render the results somewhat ragged in appearance, they occurred after the preliminary training had been completed. They probably do not affect the validity of the results. RELATIVE EASE OF LEARNING The daily records for the two animals appear in tables 1, 2-A and 2-B. They show a striking difference in the relative ease with which the animals learned the problem. Chick 2 required 585 trials, distributed over 58 days, to establish a habit of highly accurate choosing. It should be noted, however, that the bird showed consistent improvement in less than 400 trials. The average percentage of correct responses for the 195 trials given between July 12 and August 17 is 87. The monkey, on the other hand, learned the problem in the first daily series of 20 trials, during which he made only two incorrect responses. In learning the first problem — that of plain vs. striate discrimi- nation, the chicken required only 214 trials for establishing a perfect habit, and the monkey required 312. TABLE 1 Record of Monkey 2 Width of striae on both fields: 0.780 mm. Mean brightness of both fields: 6.24 candles per sq. meter Direction of striae on positive (standard) field: horizontal Deviation from horizontal of striae on No. of No. of Date negative trials correct 1915 field given responses Remarks May 18 —90° 20 18 Incorrect choices, trials 1 and 9 " 20 —90° 20 20 Not worked on days for which no record is given June 15 —90° 30 26 Very inattentive first 10 trials " 16 —90° 20 20 " 17 —90° 20 19 192 H. M. JOHNSON TABLE 1— Continued Deviation from horizontal of striae on No. of No. of Date negative trials correct 1915 field given responses June 18 —90° 20 18 « 19 —90° 20 20 H 20 —90° 20 20 u 21 —90° 20 20 It 22 —90° 10 10 u 22 —75° 10 10 u 23 —90° 10 10 u 23 —60° 10 10 it 24 —90° 10 10 it 24 —45° 10 10 « 25 —90° 10 10 « 25 —30° 10 10 it 26 —90° 10 10 it 26 —15° 10 2 u 27 —90° 10 10 it 27 —30° 10 10 u 28 —30° 10 8 It 28 —25° 10 10 <( 29 —30° 10 10 It 29 —20° 10 5 It 30 —30° 10 9 It 30 —20° 10 7 fuly 1 —30° 10 10 (1 1 —20° 10 9 u 2 —30° 10 10 « 2 —20° 10 9 It 3 —20° 5 5 u 3 —15° 15 15 a 6 —15° 10 10 It 6 —10° 10 7 u 7 —15° 10 10 It 7 —10° 10 6 u 8 —14° 10 9 It 8 —13° 10 10 it 9 —13° 20 13 It 10 —13° 20 16 it 11 —15° 10 9 u 11 —13° 10 9 « 12 —15° 20 18 It 13 —15° 5 4 it 13 —12° 15 14 u 14 —15° 5 4 u 14 —12° 15 15 it 15 —15° 5 5 It 15 —11° 13 12 It 16 —15° 5 5 It 16 —10° 15 14 It 17 —15° 5 5 It 17 —9° 15 14 It 18 —15° 5 4 " 18 —8° 15 13 Remarks Very inattentive and slow Up to this series most choices were made without comparison; ten- dency to compare from today on Only records from today on are sum- marized in table 4 Removed after trial 18 PATTERN-DISCRIMINATION IN VERTEBRATES 193 TABLE 1 — Continued Deviation from horizontal of striae on Date negative 1915 field July 19 " 20 " 20 " 21 " 22 " 22 " 23 " 24 " 24 " 25 " 26 " 26 " 29 29 30 30 31 1 Aug. a a u a a u a a a a a a 41 a a u u u it u a u u U u a u ■a 8 9 10 10 11 11 12 12 13 13 14 14 15 15 16 16 17 17 18 19 20 21 21 22 23 24 24 25 25 26 27 28 —8° —8° —7° —7° —7° —6° —6° —6° —5° —5° —10° —5° +30° +25° +20° + 15° + 15° + 15° +20° +20° +20° + 15° + 19° + 18° + 17° + 16° + 15° + 14° + 14° + 13° + 13° + 12° + 12° + 11° + 11° + 10° + 10° +9° +8° +7° +8° +8° +8° + 10° +20° + 15° + 10° + 12° + 12° + 12° No. of trials given 20 5 15 20 5 15 20 5 15 20 5 15 10 10 10 10 20 20 20 20 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 20 20 20 20 20 20 20 6 1-1 10 10 20 20 10 No. of correct responses 18 4 15 18 4 13 19 4 12 15 5 9 Remarks 10 10 10 7 15 14 18 18 10 8 10 10 10 9 8 9 10 9 10 9 8 9 9 9 17 18 19 14 17 14 14 4 14 10 7 16 20 10 Records from today to Aug. 9 inclusive not summarized in table 4 Sustained severe injury in fall just following this series Still quite lame; responses slow Condition improved More active Active Very active Was frightened during this test; be- came wild and remained so for 3 weeks 194 H. M. JOHNSON TABLE 1 — Continued Deviation from 1 lorizontal of striae on No. of No. of Date negative trials correct 1915 field given responses Aug. 28 + 11° 10 9 " 29 + 11° 20 17 " 30 + 11° 10 9 " 30 + 10° 10 7 " 31 + 10° 20 17 Sept. 1 + 10° 20 15 " 2 + 10° 20 19 3 + 10° 20 12 6 + 10° 22 17 7 + 10° 20 14 9 + 15° 20 20 " 10 + 15° 20 20 " 11 + 15° 20 19 " 12 + 15° 20 20 " 13 + 15° 20 20 " 14 + 12° 20 19 " 15 + 11° 20 20 " 16 + 10° 10 7 " 17 + 12° 20 15 " 18 + 12° 10 10 " 18 + 10° 10 10 " 19 + 12° 10 10 " 19 + 10° 10 10 " 23 + 12° 20 19 " 24 + 10° 20 20 " 25 +9° 20 18 " 26 +9° 20 20 " 27 +9° 20 17 " 28 +9° 20 19 " 29 +8° 20 20 " 30 +7° 20 19 Oct. 1 +7° 20 14 7 +7° 20 17 8 +7° 20 15 9 +6° 20 15 " 10 +6° 20 18 " 11 +6° 20 18 " 12 +6° 20 20 " 13 +5° 20 20 " 14 +5° 20 18 " 15 +5° 20 18 " 16 +5° 20 17 " 17 +4° 20 16 " 18 +4° 20 15 " 19 +4° 20 16 " 20 +4° 20 15 " 21 +4° 20 17 " 22 +3° 20 14 " 23 +3° 20 12 " 24 +3° 20 17 " 25 +3° 20 17 " 26 +3° 20 16 a 27 +2° 20 12 Remarks Retraining considered necessary- Distracted by noise outside PATTERN-DISCRIMINATION IN VERTEBRATES 195 TABLE ] —Continued Deviation from horizontal of striae on No. o No. of Date negative trials correct 1915 field given responses Oct. 29 +2° 20 17 " 30 +2° 20 16 *ov. 2 +2° 20 16 3 +2° 20 14 4 + 1° 20 12 5 + 1° 20 11 6 + 1° 20 11 7 + 1° 20 12 8 + 1° 20 15 Remarks TABLE 2-A Record of Chick 2 Mean brightness of both fields: 6.67 candles per square meter Positive field Negative field Width Angular Width Angular of deviation of deviation No. of No. of Date striae from striae from trials correct 1914 (mm.) horizontal (mm.) horizontal given responses May '21 1.56 0° 1.56 90° 10 5 « 22 1.56 0° 1.56 90° 10 3 u 23 1.56 0° 1.56 90° 10 7 a 24 1.56 0° 1.56 90° 10 7 tt 26 1.56 0° 1.56 90° 10 7 a 27 1.56 0° 1.56 90° 10 6 a 28 1.56 0° 1.56 90° 10 7 a 29 1.56 0° 1.56 90° 10 6 June : 7 1.56 0° 1.56 90° 10 6 a 8 1.56 0° 1.56 90° 10 7 tt 9 1.56 0° 1.56 90° 10 6 u 10 1.56 0° 1.56 90° 10 6 a 11 1.56 0° 1.56 90° 10 . 7 a 12 1.56 0° 1.56 90° 10 7 a 14 1.56 0° 1.56 90° 10 6 a 15 1.56 0° 1.56 90° 10 5 tt 16 1.56 0° 1.56 90° 10 5 a 17 1.56 0° 1.56 90° 10 9 a 18 1.56 0° 1.56 90° 10 9 a 19 1.56 0° 1.56 90° 10 7 a 20 1.56 0° 1.56 90° 10 6 tt 21 1.56 0° 1.56 90° 10 5 tt 22 1.56 0° 1.56 90° 10 8 it 23 1.56 0° 1.56 90° 10 7 a 28 1.56 0° 1.56 90° 10 4 it 29 1.56 0° 1.56 90° 10 7 it 30 1.56 0° 1.56 90° 10 7 Julv 1 1.56 0° 1.56 90° 10 7 it 2 1.56 0° 1.56 90° 10 7 it 3 1.56 0° 1.56 90° 10 7 196 H. M. JOHNSON TABLE 2- A— Continued Aug. Positive field Negative field Width Angular Width Angular of deviation of deviation No. of No. of Date striae from striae from trials correct 1914 (mm.) horizontal (mm.) horizontal given responses Lily 4 1.56 0° 1.56 90° 10 8 " 5 1.56 0° 1.56 90° 10 9 ■ 6 1.56 0° 1.56 90° 10 7 " 7 1.56 0° 1.56 90° 10 5 " 8 1.56 0° 1.56 90° 10 6 " 9 1.56 0° 1.56 90° 10 8 " 10 1.56 0° 1.56 90° 10 6 " 11 1.56 0° 1.56 90° 10 8 " 12 1.56 0° 1.56 90° 10 6 " 13 1.56 0° 1.56 90° 10 10 " 14 1.56 0° 1.56 90° 10 9 ■ 15 1.56 0° 1.56 90° 10 10 " 16 1.56 0° 1.56 90° 10 9 " 17 1.56 0° 1.56 90° 10 8 " 18 1.56 0° 1.56 90° 10 8 " 19 1.56 0° 1.56 90° 5 3 " 21 1.56 0° 1.56 90° 10 7 " 22 1.56 0° 1.56 90° 10 9 " 23 1.56 0° 1.56 90° 10 10 " 24 1.56 0° 1.56 90° 10 8 12 1.56 0° 1.56 90° 10 9 " 13 1.56 0° 1.56 90° 10 10 « 14 1.56 0° 1.56 90° 10 8 " 15 1.56 0° 1.56 90° 20 17 " 16 1.56 0° 1.56 90° 10 10 " 17 1.56 0° 1.56 90° 10 9 " 18 1.56 0° 1.56 90° 10 10 " 19 1.40 0° 1.40 90° 10 9 " 20 1.30 0° 1.30 90° 10 10 " 21 1.20 0° 1.20 90° 10 9 " 23 1.11 0° 1.11 90° 10 8 « 24 1.11 0° 1.11 90° 10 7 " 24 1.04 0° 1.04 90° 10 9 " 25 1.56 0° 1.56 90° 6 3 " 25 1.11 0° 1.11 90° 5 5 " 25 0.98 0° 0.98 90° 10 8 " 26 1.30 0° 1.30 90° 3 3 " 26 1.11 0° 1.11 90° 5 5 " 26 0.92 0° 0.92 90° 15 7 " 27 1.30 0° 0.10 0° 5 5 " 27 1.40 0° 0.10 0° 10 10 " 27 1.04 0° 0.10 0° 10 9 " 29 1.30 0° 0.10 0° 10 10 " 29 1.04 0° 0.10 0° 5 5 " 29 0.98 0° 0.10 0° 10 8 " 30 1.11 0° 0.10 0° 5 5 " 30 0.98 0° 0.10 0° 10 8 " 30 0.92 0° 0.10 0° 10 10 " 31 1.04 0° 0.10 0° 5 5 " 31 0.98 0° 0.10 0° 10 9 " 31 0.87 0° 0.10 0° 10 6 PATTERN-DISCRIMINATION IN VERTEBRATES 197 TABLE 2-B Record of Chick 2 Width of striae on both fields: 1.561 mm. Mean brightness of both fields: 6.24 candles per sq. meter Direction of striae on positive (standard) field: horizontal Deviation from horizontal of striae on No. of No. of Date negative trials correct 1915 field given responses Remarks June 16 —90° 10 10 Retraining records of 10-19 March tt 17 —90° 20 13 omitted « 21 —90° 20 18 Not worked on days for which no a 22 —90° 20 19 record is given a 23 —90° 20 20 u 24 —90° 20 17 a 25 —90° 20 19 a 26 —90° 10 3 it 27 —90° 10 9 u 28 —90° 20 20 it 29 —90° 20 19 a 30 —90° 20 16 July 1 —90° 20 18 a 2 —90° 20 19 a 3 —90° 20 18 it 6 —90° 20 15 a 7 —90° 20 20 tt 8 —80° 20 20 a 9 —75° 20 20 a 10 —70° 20 20 a 11 —65° 20 16 a 12 —65° 20, 19 a 13 —60° 20 17 a 14 —60° 20 20 tt 15 —55° 20 18 u 16 —50° 20 18 a 17 —45° 20 17 a 18 —45° 20 18 a 19 —40° 10 6 Removed after tenth trial a 20 —40° 15 9 Very excitable u 21 —60° 20 16 a 22 —60° 15 7 u 23 —90° 20 14 Very excitable; responses slow tt 26 —90° 20 16 a 27 —90° 15 14 Very slow tt 28 —90° 20 19 Slow a 29 —90° 20 17 u 30 —90° 5 5 It 30 —60° 15 14 Active a 31 —60° 20 18 Active Aug. 1 —60° 10 10 a 3 —60° 10 9 a 3 —55° 10 8 a 6 —55° 10 10 tt 6 —50° 10 8 u 7 —50° 20 17 a 8 —50° 10 9 a 8 —45° 10 7 4 198 H. M. JOHNSON TABLE 2-B^Continaed Deviation from ] horizontal of striae on No. of No. of Date negative trials correct 1915 field given responses Remarks ^ug. 9 —50° 10 9 9 —45° 10 9 " 10 —50° 5 5 " 10 —45° 15 11 " 11 —50° 10 8 " 11 —45° 10 10 " 12 —45° 10 10 " 12 —40° 10 10 " 13 —40° 20 18 " 14 —40° 20 16 " 15 —45° 5 3 Very excitable " 16 —45° 10 10 " 16 —40° 20 12 " 16 —50° 5 5 " 19 +90° 20 18 " 20 +75° 20 19 " 22 +60° 20 20 " 23 +55 c 20 19 " 24 +50° 20 19 " 25 +50° 20 20 " 27 +50° 10 9 Rather slow " 27 +45° 10 10 " 28 +45° 20 18 " 30 +45° 30 24 " 31 +45° 10 8 " 31 +40° 10 7 ;pt. 2 +40° 20 17 3 +40° 20 19 6 +40° 20 16 Refused to work 4th an 7 +40° 10 9 7 +35° 10 7 9 +35° 20 18 " 10 +35° 20 17 " 11 +35° 20 18 " 12 +30° 20 13 " 13 +30° 20 15 " 14 +30° 20 18 " 15 +30° 20 18 " 16 +30° 20 17 " 18 +25° 20 13 " 19 +25° 20 15 " 23 +25° 20 13 " 24 +25° 20 16 " 25 +25° 20 14 " 27 —40° 20 17 " 28 —40° 20 13 " 29 —40° 20 16 " 30 —40° 20 14 ct. 7 —40° 20 10 PATTERN-DISCRIMINATION IN VERTEBRATES 199 AN ALTERNATIVE METHOD OF TESTING VISUAL ACUITY A very convenient measure of an animal's visual acuity is the angle subtended at the eye by a single dark or bright stripe in a system which the animal can just distinguish as striate at a given distance. In the earlier experiments mentioned above, I obtained this value by training the animal to distinguish a striate field from a plain field at a distance of 60 cm. from the eye, and then reduced the width of the striae on the positive field until discrimination ceased. An alternative method con- sists in training the animal, as I did in the present work, to discriminate between a horizontal and a vertical system of equal width, and then gradually reducing the width of the striae in both systems simultaneously until the animal ceases to dis- criminate at the given distance. The work of Casteel 4 suggests this possibility, although Casteel did not attempt to control the factor of distance or to ascertain the limits of the animal's sensitivity. A priori, we should not expect the results yielded by the two methods to differ greatly, provided the animal's eyes were free from astigmatism in the vertical or horizontal direc- tion, and provided otherwise that in both cases the animal is responding by choosing the positive field, rather than by reject- ing the negative field. If experimental data confirm this ex- pectation, and if the second problem should prove as easy for an untrained animal to learn as the first, the time required for learning the first problem might as well be saved. I decided to test the practicability of this method on Chick 2. The daily results appear in table 2-A, between the dates of August 18 and August 26, 1914. From August 27 to August 31 inclusive I repeated the test by the first method used — that of plainstriate discrimination, which is designated as method No. 1, and the other as method No. 2. In these tests, the minimal distance between test-field and eye at which comparison could be made without a choice being registered was 60 cm. The results are summarized in table 3. The values given under the heading " Width of striae (mm.)" are the widths of striae on the positive field for method No. 1, in which the striae on the negative field were invisibly small, and the widths of the striae on both fields in method No. 2, in which the difference in direction was 90°. 4 Casteel, D. B. Discriminative ability of the painted turtle. This journal, vol. 1, 1911, pp. 1 ff. 200 H. M. JOHNSON TABLE 3 Method No. 1 Method No. 2 Width of No. of No. of No. of No. of striae trials correct trials correct (mm.) given responses given responses 1.56 t , 106 94 1.40 10 9 1.30 5 5 13 13 1.20 10 9 1.11 25 25 30 25 1.04 20 19 10 9 0.98 30 25 10 8 0.92 10 10 15 7 0.87 10 6 t , ( , These results indicate that the second method is feasible, and that uncertainty in choosing begins at nearly the same absolute stimulus- value whichever method is used. The relative ease of learning for the chick is decidedly in favor of method No. 1. Chick 1 failed to learn to discriminate when trained by method No. 2, but established a perfect habit in 443 trials when trained by method No. 1, despite numerous disturbing factors. Chick 2 learned by method No. 1 in 214 trials, as against 585 trials for method No. 2. The results given in table 3 should not be taken as indicating with accuracy where the bird's stimulus-threshold lies, as the training was interrupted too soon. I did not wish to risk spoiling the bird by frequent punishment for work on the main problem. DIFFERENCE-THRESHOLD FOR DIRECTION The term "threshold" is properly used to designate the mean between effectiveness and ineffectiveness of a stimulus or difference between stimuli, in producing a response. In the present discussion this mean is assumed to have been reached when the stimulus-difference is effective in half the total of a fairly large number of presentations or " trials." If the differ- ence were wholly ineffective, the animal should be expected, on the theory of probability, to respond correctly in 50% of a large number of trials. If the difference were always effective, 100% of the responses should be correct. The stimulus-differ- ence at which the percentage of correct choices is 75 is therefore taken as the animal's " threshold." The accuracy of such a determination is indicated by the consistency of the results. The measure of consistency used herein is the mean variation PATTERN-DISCRIMINATION IN VERTEBRATES 201 of the percentage of correct choices for the daily series of trials from the average percentage of correct responses for the total number of trials at a given stimulus-difference. In much of the past work on the special senses of animals carried on by means of training methods, the experimenters have trained the animals to respond to a given difference between the stimuli, and have then reduced the difference until the animals cease to discriminate. When a break-down occurs, the experimenters frequently discontinue the training after a few additional trials, and take the stimulus-value or stimulus-differ- ence at the breakdown as the animal's " threshold." In some earlier work I followed this procedure, having observed that in some instances further training confused the animal and caused it to sulk. In the present work I have endeavored to find the degree to which such confusion can be overcome by continued training. 5 The results obtained on the chicken show a training- effect which is important, though not relatively large; while the final results obtained on the monkey yielded a threshold of a different order of magnitude from the stimulus- difference at which the first break-down occurred. As appears in table 1, in the records from August 22 on, the monkey showed a strong tendency to break down at a stimulus-difference of 8° to 10°; but he overcame this tendency after continued training, and eventually yielded a threshold in the neighborhood of 2°. Had I given the monkey a considerably larger number of trials at the stimulus-differences of -13°, " -10°, +10°, +8° and +7°, after the full effect of training had been obtained, the average percentages of correct responses could be plotted against the stimulus-differences on a consistent and fairly smooth curve. The results of Monkey 2 are summarized in table 4, and those of Chick 2 in table 5. In compiling the results for each stimulus-difference, I observed the following procedure : Having obtained the average percentage of correct responses for the total number of trials, I took the average percentage of correct responses for each daily series, and obtained the mean variation of these daily averages from the average for the group. Since 20 trials compose the standard daily series, I weighted the 5 My adoption of this course is the result of an extensive discussion carried on with Professor Knipht Dunlap some two years ago, regarding the validity of the discrimination-method, as the latter is usually employed. 202 H. M. JOHNSON variations from the mean of the individual daily averages ac- cording to the number of trials in the individual daily series referred to 20 as a basis. The mean variation of the averages of the daily series from the general average is expressed in units of the latter in tables 4 and 5, and it gives an indication of the consistency and hence the reliability of the results. TABLE 4 Record of Monkey 2 Width of individual striae on both fields: 0.780 mm. Mean brightness, both fields: 6.24 candles per sq. meter Striae on positive (standard) field: horizontal Deviation from Average horizontal of No. of No. of percentage striae on trials correct of correct negative field given responses responses —15° 95 89 93. 7=b 2.9 —14° 10 9 90.0 —13° 60 48 80. 0± 7.5 — 12° 30 29 96. 7± 3.7 —11° 13 12 92.3 —10° 40 32 80. 0± 7.5 —9° 15 14 93.3 —8° 40 35 87. 5± 1.9 —7° 40 37 92. 5± 3.7 —6° 40 36 90. 0± 3.3 —5° 50 36 72. 0± 6.0 + 15° 130 126 96. 9± 3.5 + 14° 30 29 96. 7± 4.3 + 13° 20 18 90. 0± 0.0 + 12° 140 127 90. 7=fc 6.0 + 11° 80 73 91. 1± 3.0 + 10° 228 185 81. 2± 8.3 +9° 100 92 92. 0± 4.4 +8° 100 84 84.0±11.2 +7° 100 79 79. 0± 8.8 +6° 80 71 88. 7± 6.9 +5° 80 75 93. 7± 6.7 +4° 100 79 79. 0± 3.3 +3° 100 76 76. 0=b 8.8 +2° 100 75 75. 0± 8.0 + 1° 100 61 61. 0± 5.6 PATTERN-DISCRIMINATION IN VERTEBRATES TABLE 5 Record of Chick 2 Width of individual striae on both fields: 1.561 mm. Mean brightness, both fields: 6.24 candles per sq. meter Striae on positive (standard) field: horizontal 2(tt Deviation from Average horizontal of No. of No. of percentage striae on trials correct of correct negative field given responses responses —90° 400 345 86.2±10.0 —80° 20 20 100. —75° 20 20 100. —70° 20 20 100. —65° 40 35 87. 5± 7.5 —60° 130 111 85. 4± 8.7 —55° 40 36 90. 0± 3.3 —50° 90 79 87. 7± 2.6 —45° 110 94 85. 4± 6.4 —40° 195 141 72.3±11.4 +90° 20 18 90. +75° 20 19 95. +60° 20 20 100. +55° 20 19 95. +50° 50 48 96. 0± 2.7 +45° 70 60 85. 7± 5.8 +40° 80 68 85. 0± 5.0 +35° 70 60 85. 7± 4.3 +30° 100 81 81. 0± 8.8 +25° 100 71 71 0± 5.2 AN APPARENT EFFECT OF " CONTRAST " After I had obtained the threshold for deviation in the posi- tive direction for Monkey 2, I gave him 40 trials — 20 trials each on the 9th and 10th of November — in which the striae on the positive field were horizontal and those on the negative field deviated from the horizontal by 5° in the negative direction. Only 12 responses out of the 40 were correct, the animal per- sistently tending to choose the negative field against punish- ment. In this connection it should be noted that when the difference in direction is small and when the striae on one field are horizontal, if the two fields are regarded simultaneously, the striae on both fields appear to be inclined from the horizontal, but in opposite angular directions. For about six weeks previous to these control tests, the striae on the positive field, although really horizontal, appeared to the human observer to deviate in the negative direction, owing to the influence of the other field, whose striae were inclined slightly in the positive direction. If 204 H. M. JOHNSON we assume a similar "contrast-effect" for the monkey, we have a sufficient explanation of his behavior in these control tests. SUMMARY Three animals — a monkey and two chickens — previously trained under comparable conditions, showed a striking dif- ference in their individual ability to acquire a perfect habit of response to difference in direction in two systems of visible striae. The monkey perfected his habit in the first series of 20 trials. One chick failed to learn the problem and another chick required 58 days and 585 trials to perfect the habit. The work demonstrated the practicability of an alternative method of testing visual acuity, but suggested that for some animals the problem is so difficult as to make the method orig- inally used by the author more desirable. The monkey's difference-threshold for direction of elements of a pattern lies between 2° and 5°; and the chick's threshold between 25° and 40°. The magnitude of the thresholds within these limits for either individual depends in large part on the effect of training. The training-records, as well as the final results, suggest that the characteristic of direction in visual objects has much more significance for the monkey than for the chicken. The relative improvement brought about by training is very much greater for the monkey than for the chicken. This fact suggests that other differential factors than relative development of the sense-organ play an important part in discrimination of this type. The point is emphasized by the fact that the mon- key's visual acuity, as shown in an earlier study, is between four and five times as good as that of the chick; while his sen- sitivity to difference in direction is from eight to twelve times as great. VISUAL PATTERN-DISCRIMINATION IN THE VERTEBRATES— V A DEMONSTRATION OF THE DOG'S DEFICIENCY IN DETAIL-VISION H. M. JOHNSON Nela Research Laboratory, National Lamp Works of General Electric Company The experiments reported in this paper grew out of certain questions of especial personal interest to the author. These questions are of minor importance as compared with some gen- eral questions which the work involves, but as the former really determined the course of experimentation it seems well to men- tion them. In the years between 1910 and 1912 the writer, then working at the Johns Hopkins University, acquired some dogs which had been rendered temporarily blind by having their eyelids scraped at the edges and stitched together during the first week of puppyhood. This operation caused the eyelids, which normally separate at about the ninth day, to grow tightly together; and they remained in this condition until they should be separated by a slight operation. While these animals were blind 1 I trained each of them to open three "problem-boxes" by a single movement of the head, the teeth or the paw, within a minimal time (two seconds or less) and without making any useless movements. The manner of arranging and conducting these experiments is open to criti- cism, but the results indicate clearly that the blind dogs became quite as proficient as normal dogs trained under identical con- ditions, and that the number of trials necessary to the acqui- sition of skill is not consistently greater for blind dogs (Nos. 7 and 8) than for a normal dog (No. 6) of the same litter and trained in the same way. Also, the blind dogs, as well as the normal dogs, showed practically no loss of skill after sixty days of rest. 1 Johnson, H. M. Audition and habit-formation in the dog. Behavior Mono- graph, no. 8, Cambridge, Mass., Henry Holt & Co., 1913. 205 206 H. M. JOHNSON After each of the blind dogs had learned to open three boxes in this way, the eyelids were parted by an operation. All the animals except Dog 7 showed an opacity of the cornea, usually slight in degree and extent, but varying in different individuals. Dog 7 did not exhibit an opacity. None of these animals showed marked changes of behavior after the operation. I then required Dogs 7 and 8 to learn three more problem-boxes each. They exhibited slightly greater proficiency in acquiring skill than they had shown while they were blind, but the improvement is due — ■ at least in part — to the effect of practice, and is considerably less than that brought about wholly by practice in the case of the normal dog. 2 These experiments demonstrated that the dog can get along surprisingly well without vision in making complicated adjust- ments. Other experiments indicated that even normal dogs did not react to some objects as if the latter were visually per- ceived, provided the objects were stationary. While the foregoing experiments suggested that the dog may make little use of vision as we ordinarily think of it, they left the question open as to the kind and degree of vision which the dog may have. In 1911-12, I made an attempt to test the dog's sensitivity to difference of outline in visual objects. I used a single animal, designated as Dog 5 in the report cited. I made the tests in the psychological laboratory of the Johns Hopkins University, under the direction of Professor John B. Watson, using the Yerkes-Watson standard apparatus 3 and the discrimi- nation-method recommended by these authors. The test-objects used were a circle 6 cm. in diameter and its (approximately) equivalent square. The animal was fed at the food-box under the square. He learned in about 1000 trials to choose the square invariably, but its brightness (and hence its luminous intensity) was four times that of the circle. When the bright- ness-difference was eliminated, the animal immediately ceased to show any preference for the square, and he did not make any consistent improvement in 600 trials. It became necessary to discontinue the work at this point. Shortly afterwards the 2 The records of dog 6 (normal) and of dogs 7 and 8 (first blind and later with the eyelids separated) are intercomparable. The records of dogs 1 and 2 and of dog 5 are not. 3 Yerkes, R. M. and Watson, J. B. Methods of studying vision in animals. Behavior Monograph, no. 2, Cambridge, Mass., Henry Holt & Co., 1911. PATTERN-DISCRIiMINATION IN VERTEBRATES 207 animal was accidentally killed. Meantime I had acquired no information regarding the refraction of his eyes. Consequently I did not feel justified in publishing the report of the work as containing anything decisive or important. It is of some in- terest when considered in connection with some later experiments described below. Three years ago I became interested in the question of the relative visual sensitivity of animals possessing different types of retinal structure. The retina of the dog's eye is fairly typical of the mammals below the primates. There is no fovea, 4 rather indefinite "sensitive areas" and numerous opaque nerve-fibers which must be considered as blind spots. Such a retinal struc- ture is characteristic, in the main, of mammals below the pri- mates. It seemed fair to select the dog as a representative of this class. I chose the domestic chicken as representative of animals possessing a definite sensitive area but no fovea, and the Capuchin monkey as representative of animals possessing a fovea centralis, temporal to the entrance of the optic nerve. I also felt impressed by the apparent superiority of stimuli such as Casteel 5 used, over stimuli differing only in outline, for a test of the question in which I was particularly interested. I selected 6 as the stimulus to be chosen by the animal, a cir- cular field, 6 cm. in diameter, covered by a system of horizontal dark and bright stripes of equal width. The number of stripes on the field and the width of each stripe could be varied by insensible gradations. The field to be rejected was covered by a similar system of striae, but the members were so numerous and so fine that the eye would not resolve them at the minimal distance at which the animal could make his choice. The stim- uli differed in no other respect. The sensitivity of the animal was to be measured by the angle subtended at the eye by the separation of the upper edge of one bright stripe from the lower edge of the bright stripe just above it. The reader should con- sult the original report for details of these tests. The daily results obtained for Dog 1, a pure-bred English 4 Slonaker, J. R. A comparative study of the area of acute vision in verte- brates. Journal of Morphology, vol. 13, no. 3, 1897. 5 Casteel, D. B. Discriminative ability of the painted turtle. This journal, vol. 1. 1911, pp. 1 ff. 6 Johnson, H. M. Visual pattern-discrimination in the vertebrates. I and II. This journal, vol. 4, 1914, pp. 319-361. 208 H. M. JOHNSON bull-terrier, are shown on page 352 of that report. When the width of each stripe on the "positive" field was 5.2 mm., and the distance from the eye was 60 cm., the animal chose the positive field 36 times in 50 presentations or "trials". In the last 20 trials, 18 choices were correct. Believing at the time that he had learned the problem, I reduced the width of the stripes. On examining his subsequent daily records with these, I became convinced that the improvement was accidental, and due to a change in the animal's "position-preference" as a time when his right-left order of choosing happened to fit the order of presen- tation quite well. When the width of each stripe was 3.9 mm. the animal chose correctly in 127 cases out of 200; when it was reduced to 3.72 mm., 112 out of 175 choices were correct. 7 These results did not show anything conclusive regarding the dog, except that working near the limit of the instrument I had not made a stimulus-difference large enough to be effective in 50% of the cases. (This condition would be met when 75% of a large number of responses were correct.) The general method followed in these experiments had yielded quite definite and positive results in the work on the monkey and the chickens. The individual stripes composing the posi- tive system had subtended a very large visual angle throughout all the w T ork on the dog. The failure to obtain positive results on the dog raised several questions regarding the work. Early in the training the animal had learned to choose the correct food-box by testing the electrodes at the entrance-alleys for electrical charge. When I eliminated this factor he showed great disturbance. It seemed advisable to use another indi- vidual, which had not been thus disturbed during the training. I selected a pure-bred female beagle-hound, four months old, for this purpose. Beginning May 26, 1914, I gave her ten trials daily for 90 days, using the method described in the work on Dog 1, the monkey and the chickens. She showed no consistent pref- erence at any time for the positive field, each stripe on which was 5.2 mm. wide. The minimal distance between test-field and the eye at which a choice could be made was 50 cm. A detailed re- port of her work would be without value in view of facts which we later ascertained, and which I shall mention below. 7 After the original report was prepared I resumed the work at this stimulus- value, and obtained a lower percentage of correct choices over 150 additional trials. PATTERN-DISCRIMINATION IN VERTEBRATES 209 It now became necessary to determine whether the negative results which the dogs gave were due to the inappropriateness of the conditions under which the latter had worked. The first matter to be settled was that of errors of refraction in the dog's eyes. This examination should have been made be- fore the experiments were begun, as is made evident by the results. My colleague, Dr. P. W. Cobb, was kind enough to make the examination and to assume responsibility for the results. He administered atropin to ■ the animals for about a week, until the pupils ceased to respond to light. He then made a careful skiascopic examination. The results obtained during mydriasis were very consistent. Dog 1 showed about 0.25 to 0.50 D. hyperopia during mydriasis. (This condition is characteristic of the emmetropic human eye.) No astigmatism was discoverable. The examination without mydriasis yielded variable results, but none of them indicated the presence of a refractive error which the animal did not overcome by accom- modation or by an analogous process. We assumed, therefore, that this dog's eyes were for practical purposes emmetropic. The correction for Dog 2 was +1.75 D. sphere with +0.75 cylinder, the axis lying in the horizontal meridian. The neces- sary correction was the same for both eyes. A week later Dr. Cobb attempted to test the animals for extent and range of accommodation, but got no consistently positive results. It seemed useless, therefore, to continue experimentation of Dog 2, since even the best accommodation could hardly have overcome so large a refractive error, and since it was inconvenient to equip her with spectacles. At this time I was working some other animals on the prob- lem of discriminating between a vertical system of striae and a horizontal system differing from the former only in direction. The problem had proved itself quite easy for some of the animals. I had also found it feasible to test an animal's visual acuity by making the stripes on both fields rather coarse, and training the animal to choose the food-box indicated by the horizontal system. The next step was to reduce the width of the stripes in both systems simultaneously, until the members became so numerous and their width and separation so small that the animal could not distinguish the horizontal system from the vertical system at the given distance. I decided to test Dog 1 by this method. 210 H. M. JOHNSON In this work I used the same experiment-box and test-objects and the same general technique as are described in the former report. 8 The mean brightness of each test -field was 6.2 candles per square meter. The width of each stripe in either system was 3.902 mm. The animal had to choose with the eye 40 cm. or more from the test-field, and he was fed in the food-box indicated by the horizontal system. I gave him 20 trials each day. In 140 trials the animal chose correctly only 75 times, and showed no tendency to improve-. There seemed reason to question whether he was capable of discriminating between purely visual objects, and of demonstrating the fact by acquiring a perfect habit. Professor Mast, with whom I discussed the question, urged me to settle the point. On June 24, 1915, I gave the animal 20 trials in which only the positive test-object was pre- sented. He was fed in the box at which it appeared. All 20 choices were correct, and there was no hesitation at any time. On the following day I reintroduced the negative' field, but stopped it down, so as to introduce a difference in area in addi- tion to the difference in pattern. Hitherto, each field had been limited by a circular diaphragm 6 cm. in diameter. Over the diaphragm limiting the negative field I now slipped a smaller diaphragm, the opening of which was concentric with that of the former. The animal's daily records for the various differ- ences in area appear in table 1. The results show that the dog is able to discriminate when the difference of area presented is between the ratios of 2.25 to 1 and 1.44 to 1. On July 24, August 1, August 14 and August 31 I made control-tests, in which the width of the striae on both fields was so reduced that the fields appeared of uniform brightness to the human eye. As will be seen in the table, these tests showed that the dog's behavior was not affected by the presence or absence of the stripes. After the animal had ceased to discriminate, when the stimuli was a 6 cm. circle and a 5 cm. circle, I retrained him for six days, beginning August 30, 1915, requiring him to choose the 6 cm. circle and to reject the 3 cm. circle. Both fields were of uniform brightness. His average percentage of correct choices for the 120 trials was 85.8. I then reduced the brightness of the 6 cm. circle by introducing an absorption- 8 In the present work I placed a copper tray, holding a moistened felt pad, between the exit-door of the home-box and the alleys enterin? the food-boxes. This tended to reduce the variability of the resistance of the animal's feet to a minimum. PATTERN-DISCRIMINATION IN VERTEBRATES 211 o o Q O a S o o w ►J •Mean brightness, both fields: 6.2 candles J per square meter C/3 u a S Not worked on days for which no record is given + — field field 3.9 3.9 0° 90° 6 5 S9SUOdS9J ;09JJOD 'OJsj; U9AI3 sjbu; '-ojsi + — field field 3.9 3.9 0° 90° 6 4 S9SUOdS9J P9JJO0 •OJSJ U9AI§ SJBU} '-OM + — field field 3.9 3.9 0° 90° 6 3 S9SUOdS9J ;D9JJO0 'O^ U9Al3 SJBUJ '-Oft field field 3.9 3.9 0° 90° 6 2 S9SUOdS9i P9JJO0 - 0^[ U9AI§ SJBU} '-OJSJ CO CO -tf CO 00 i— 1 i— ( i-H i-H I— 1 ooooo N t^O5?DO5 ooooooooooooooooooo CV]C^C\]C^MC\1WC^C\1C\)C\]C^C\]CnJC\]0}Cn]CMC\] Width of striae (mm.) Deviation from horizontal. Q Lf3CDt^ooaiO--HiMcocDt^ooaiO'-HCMco'*Lncr>t>ooa^o fi! 3 3 3 3 j^S 3S33333333333333 »— 1 •— > 212 H. M. JOHNSON a .s s a g o Q b, O a K O o w < en CD lJ "O Q c O . • o o . re u i£E hE CM gee gs CD U_05 O _ CI U_H en fcow £ fco •a •C o o *j *j +j . en S tJ*» tness, both fi e meter en re £ eady for fu ae reduced ae restored ake presen und presen hapman pr eady for fu ae reduced ge present. ae restored ndahl prese eady for fu ae reduced >-'£ be p u'C n brigh ir squar ,u3-w i- +j siderec thof s thof s R. Ke A. H. F. W. siderec th of s r . E. B thof s ,. 0. G siderec thof s ]CN1CN!OJ i s<^°^ S9SUOdS9J inoooxtoNao Koj oco i-l rH rH iH H rH rH N 1 eCOOCT. pgxioo 'ojnj; ,2CT. U9AI§ +|«o^ COOCOOOO SJBU} ■OJsT CMCV]CNICN1CNICN!CMCM iSo 1 '^ S9SUOdS9J 00C100O Md • o cm T— 1 |— 1 I— 1 1— 1 1 OOOC>C •-!-— i ^f cr> o oo cr. o <— i oj oo ■<* m .c re 2 +J CMCNJCMCNCMO3CNCNCN1C0C0 r— 1 r— 1 ,— 1 ■— I r— 1 T— 1 •S"> re Q >»- 3 ►-a ~ 3 - - - < PATTERN-DISCRIMINATION IN VERTEBRATES 213 s s CO o u O +J J-i CU O H-> -^ g ,0 - £ t/j rH rn CS O a; O — Ih T3 - CD at, O I) 3 l-i 0) cu S-i tj a cu CD ro cu S-i : = *-> -J ■ 09 O >*i 3 cu 5 a ~> t> 7; > m i— i 4-1 ~2 Vh n i- +-" _ -t-> , CI J CT2 u S ^ C cu ,4 o . ^ n ■=> ■s* '.J .i>cr>Ln< •—i i— I CM •— I r—i r— 1 i— ( i i CM' OOOOOOCOOOO t>0~. O CM CO ^t< LO t>- 00 OlOr- KMCOaiOxPLOCDOOO) »-H ^H i— I -rr LO O CTi O (DLON OO i-H tt co r^^Loco ^ « 2;lo«3 ^2 Oi £ -tfCO'— i^LOO^LOCO^COCOLOCMCO-tfO .— 1 CM & co^oicolococoloocoooloooocoS OJ oo S. 10 1.2.3 I— 1 ^h cm rO'-H co •— i NC5 CO ■— i i— i i— I CM >-H r- < ■—( CM CO-— 1 CO --H CO CO CO ■— 1 i— i CO COCO'— 'CO-— li— i ^co 1-H 1— I .— t S.9 7.8.9 CO iri LO S. 8 4.5.6.7.8 ^ t^ "^ "<* O LO ^f t> ■>* ^^^^oo^O'^rti^t'odLd^^t'LOLO^^ 1 oq CO LO "^ CO Tf S. 7 5.6.7 LO O LO LO LO LOCOLOLOcbLOLOLOLOCOLOLOcdLOLOLOLO 2.3.4.5 6.7.8 CM LO S. 6 6.7.8 CO 00 CO CO00 COCO CO COOCOCOOOOOCOOOOOCOCOCOCOCOCOOOCO LO CO CM ** 1— 1 CO 00 S. 5 2.3.4.5.6 CM CDC) CM CM CO CO (M CM CM CM CM COLriLO^CMCOCMCMCMCMLOCO^'LOCMCMCM c-. 00 CO CM S. 4 7.8.9 ooc>t^t>-o6cxiooc^t>-odt>t>t>t>t>cj5C^ od co LO t> S.3 3.4.5.6.7 CO CO CO t- O COCOCO CO ^t< CO CO-^CO'*COCOLOM , COLOCOCOCOCOCOt>CO LO PC CO CO CO 00 00 ooooooooooooooaJooocooooooocoooooo CT5 00 CO 00 S. 1 1.2.3 .— 1 <— 1 CM CO ■— 1 I - 1 CO' - 1 1— If- H .-H .-H coco>— ICMCMCM-— icmco<-|'-hcmcM'-icmcO'-h CO CN CM 1— 1 No. of trials ooooooooooooooooo — r ) co — i o cc t -- x ~. ~ — r i c- : — i r. -o i - 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 i— iOOCTiO'— < CM CO "^ LO CO o I— t 1 T—t Date >1 "^OOCJiS OCMCO^ts LOS COS 00003 O __j .— 1 I— 1 t-H i-H .— 1 .— 1 t— 1 r— I ,— 1 CM 1 1 o CM Reten- tion Sept. 9 230 HAROLD E. BURTT Ratio of RtoW QCOQCOQQO CO O CO CO LO t>- CO Q p CO. O CO O O LO CO LO CO CO CM. f-h " 00 1> cmo 21 ^ 2^^ m ^ f-H tt FHCOIMCO-HrHTt (£) 00 COCO00 ^ ^2 Oi £ Oit^00t>OiOiCDC^C^CDCOC^C^-<-rCMF-HCM-H Oi 0. OiO-oot>oot^ooioot>t>ooooc^t> t> CO LO t-H § 00 CO 00 LO ■* LO -f* LO ■<* CO LOC^LDl>CO Tt-F* ^j< O tooot>-ooi>'^'*-Fd < LnLO^t < '^ , oO'*oo-FH , 'Fi , '^ , Tt i -rti CO t> CD LO CO CO 03 O K 2 co to LO LOLOLOCO LOLD LOLOLOLOLO LO LO cococ^c^Lot>-[^i-ocDcot>coc^LOLocoLDcoLOLo ^CO CN CM TABLE II Rat D, Outbred Mal to CO 00 CD C-- CD CD CO CO CO CD 0-'cOC^odo6t>COododcDCOCDCDCOCDCOCOCODCD LO ■>* CO CM 1—1 f-H CD Lfi ■>* CO CM CM W M CO ^- ^T CM CM CM CM LO CM LO CO ^i*-* CO CM CM CM CM c OlOCDCMLOCOCDLOCOCMCO'<*COCMCMCMCMCMCMCM Oi 00 t> CO K O C/i H J t/) U CO C. 00 t> CO t> oi t> c^c^oo i>- c^ t> t^ t> t> t> ede^t^ Oi Oi t~- 00 oi t> o> t> 00 00 Oi 00 00 00 1>- a-, a 00 1> 0. 06 t> CD LO 0# CO cd 'A in co CO LOCOCO CO CO CO CO -^ ^ COCO CO -^^cocot^c^c-^cococo^^cococococo^tcoco LO fO CO CM CO 05 06 00 cc cr.ee 0000 000c 00 00 O- O- Oi 00 Oi O- 00 Oi Oi 00 00 O. 00 Oi 00 00 00 00 00 00 Oi 00 t> CD F-H CO co 1-H f-H F-H CO CM c 2 en c 00 00000000000000000 F-HCMCO^tLOCDt^OOO. 0-— iCMCO-tfLOCOC^OCO. 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 I 1 1 F-HCJCO^rLOCDC^OCOOF-HCMCO-F*LOCO!>OCO. 1 F-H CD -4-1 c3 Q >> ^FtCONOOC.OlCO'H/3 LO 3 CDs t^OCO-3 Os ^h 1^ f-H f-H f-H 1— 1 f-I ^H .— 1 f-H CM CM CM WHITE RAT AND THE MULTIPLE CHOICE METHOD 231 w c OS a, w ►J < ~ a CQ g < ~ < « o cfi H ,J 13 fa Ratio of RtoW 1:9.00 1 :3.00 1:0.54 1:0.54 1:0.43 1:0.43 1 :0.54 1:1.00 1:0.43 1:0.33 in CM o »— 1 £ en J£5 t — ococo i>loco 2 CM M rH LO 23 «^N 2lOC^ g 00 £ OS cr. coco^fCMiocococo'^LOcocoM'coo CM tt >-H'-H'*c^coooLnt>t>t>coLoi s -c^cot>-2 Oco ^CM i— t — 1 CM -— ' ^H CM ■— i CO N CM i— i ,— I COCO'— < -h CM >— i CO ■— 1 »— 'CO-— i CM >— i i— I i— t CM •— I 1.2.3.4.5 6.7.8.9 I— 1 CM CO LO 00 en .00 CnoOI>l>t^tX^C^t>C^t>C^t>t>-I>C^O CO - LO S. 8 4.5.6.7.8 cd^t co Tf o6 ^ ooLooo-*TfT} CO CO COo6cOCOt> 00 CD 00t>00t^00COCOCOCOCOCOt>COCOt>COCO LO CO cm t-H S. 5 2.3.4.5.6 CM CM CM CM CM CM CM CM COCO^f" CO COCO CM CM CO CM CO CM CM COCOCOCM"st < CMLO-^CMCO'*"sf , COCOCOCOCM G5 od o CO S. 4 7.8.9 o 00 cr. c- t> c- O ooodocooc^t>t>-c^ooc^i>-t^oc^t>t^i^ cr. 00 1>^ CO LO <* CO S. 3 3.4.5.6.7 CO CO LO CO LO CO CO COCOt> t^cO c^cocococo coco C-~t>C^COCOt~-'fCO"'*^ , -'*-<*TfCO-3 , -*CO LO CO M en c/3 °° oocoxocooocooocococooooooocoooocc cr. oo S. 1 1.2.3 r—l i-H i— i i— i ^n CM CM i— i CM CM CO CO CO CM •— i i-l i-H ^h COCOCm'cMCMCOCM^— i — i CM CO "-< <-h CM ■— i<-h CO CM No. of trials ooooooooooooooooo •— iCMcO'^ , Locot>cocr. c-— iojco^lococ^ 1 1 1 1 1 1 1 1 1 1 !! 1 1 1 1 1 ^cmco^iocooooct. o^cmco^loco o E. BrilTT Ratio of RtoW 1:0.66 1:0.66 1:1.00 1:1.50 1:1.50 1:1.00 1:0.81 1:1.50 1:1.50 1:1.85 1 :0.66 1:1.50 1:1.00 1:0.66 1:0.33 1:0.25 1:.033 1:.043 1 :0.66 1:0.15 1:0.11 1 :0.05 © i-H £ ■<* •* 2^ £- ^ cr. co cd 22 °o ^ m ^ >n ■* in x ^ ■* cm r-n o « coco 2-f «m S^r^ c- S3 xincc S 2 S^« g 2 2 o £ t -r co -r co i^mmc^cMcoco^coTi<-'!j<-3— icmi— 1.-1.-10 o X <£HO^)><0^1> xmmxX^"*XTr'COCDCDCMinCOt>Xt>CT>Xt>C^COaiCr. 00 Ci Ci O 2. o S. 10 1.2.3 .—I i— 1 •— I CM •— 1 t— 1 ^H i— 1 >— 1 rtHHMH •— ' CO CM CM i— iNNr- i .— 1 CM i— 1 i— 1 •— i CO COOOCOCOCM ■HNMM'HM'HNMMNNCO'HNiHiHNr^NlHlHNrHiHrHrHrH^I^I CM^ •co H S.9 7.8.9 co in m S. 8 4.5.6.7.8 -<* ■<* m -^ -* -* in rr xlo t^ oo oo x^ tj< Tj"* Tt 1 tj< Tt> rt t> oo ■* ^tin-^ t* -5j«T?^Lnin^u^in^incDt>^a5-^mmtDmt^cDt>t>mmm5DCDmc£)mmmmi>m ^CO CM S.6 6.7.8 CD CD CO CD CD CO CD C~^ CO CO CO CO CO CO CD C-^ X CDCD00C^COCOC^t>COCOCDCDCD00t-COt>C^COCDCOCOCDCDt>COCOCDCDCO LO CO CM .—4 S. 5 2.3.4.5.6 CM CM CO CM CM CM CM CM CM CM CM CM-* CM CM CO CM OJ COCO COCOCMCM CO CM CO CM CM CO CM"* — T CD m CO XX^CCCM-tf^XXCMLOX^XCMXCMCMCMCMCMCMCOCMCMCMXLnCMCM X CO S. 4 7.8.9 t"- t> CT> t> OO t^O t> XOOOOOOOOOOOOOOOOOOOOOCOOOOOOOOO X o co m •<* "* S. 3 3.4.5.6.7 CO CO -tf com m CO CO t>COC0COC0COC0 CO coco COCOmoOCO tJ< CO CO ■* t> O t> CO t> O COCO t> CD"* COCO CO CO COCOCOCO CO CO CO coco coco m ■* X 00 0000X0000 XG1XXXXXXXXXXXXXXXXXXXXXXXXXXXX Oi x o co S. 1 1.2.3 «— 1 i-H .— 1 H ,— 1 ,— 1 i— 1 I— < •— 1 CICO^-— iCM ^CM^hCMCMCM.— '^^.-.CMCM^h ^- i-H ,_i 00 CM No. of trials 28328 8gS882S8SS8£8882S8§88R88828gS!8 — CM CO -T LOCDNXacHcim^LOCONSSo^NCO^LOCONSaO^WCO^ o 1— 1 1 I-H Date EJ&SSgJ 3 g 3^- NCOS -tfCOOs Ms CX= CMX^s ma CDs t> X CT> s * Os Ha CM CM WHITE RAT AND THE MULTIPLE CHOICE METHOD 233 on 10 new settings to check the possibility of memorizing the particular settings rather than actually solving the problem. One of the rats did the check series perfectly, two others with only one mistake, and the other with two mistakes. It is evident that all formed the habit of selecting the compartment at the extreme right. A number of interesting individual tendencies were manifested in the course of the experiment. A's method of reaction was usually as follows: passing down the right wall of the reaction- chamber to compartment No. 1, she turned leftward and followed along close to the compartments until she reached the first open door. In many instances she turned from the right wall before reaching compartment No. 1, with the result that she entered a door too far to the left. This factor was of course more opera- tive with settings comprising doors at the right end. It is evi- dent from table I that settings 1, 3, 5 and 10 yield a larger number of incorrect first choices than do the others. With settings farther toward the left the first approach to the doors brought the subject more often in front of a closed door and hence the first open one encountered was correct. On Sept. 9, over seven weeks after the discontinuance of A's training, she was given 10 trials with the original settings and made 8 correct choices out of 10. Evidently the habit was still strong. D manifested a tendency entirely different from that of A. He ran along the left wall of the reaction-chamber and across in front of the open doors until he reached a closed one, where- upon he turned back into the last open one. This turning back appeared to be caused by the touch of the vibrissae on the closed door rather than by the sight of it. The turning was always in the same direction, — leftward or toward the doors. D ex- perienced more difficulty than the other rats with setting No. 2, i. e., doors 8 and 9. Going down the left wall he often entered compartment No. 9, whereas he normally turned at that point. As he did poorly also on the other settings involving No. 9, it would appear that he was reacting by tactual rather than visual criteria, — habitually going along the wall until he encountered an obstacle and then turning across, etc. The directness with which No. 9 was entered in contrast with any other door was noticeable. A month after the training ended, although sick and able to move only slowly, he performed two trials correctly 234 HAROLD E. BURTT on the first two settings, employing his customary method of reaction. J manifested, like A, a tendency to go down the right wall of the reaction chamber to compartment No. 1 and then in front of the closed doors to the first open one. He was often, how- ever, premature in turning to the left thus missing the first open door, and he often lost the latter part of his habit and passed several open doors without entering. As contrasted with A, he was less ready in recovery from mistakes. Both solved the problem in 170 trials, but whereas with A there were 17 trials in which more than one incorrect choice was made before the correct one, there were 31 such trials with J. It is to be remem- bered that J was inbred and A outbred. C was extremely rapid in her motions, rushing toward the compartments the moment she was released from the entrance box. In her correct trials the total time from entrance-box, through the compartment and alleys to the food was fre- quently 5 seconds. She often ran along the back alley so rapidly that she slipped and fell at the turn into the side alley. This tendency to hasten was perhaps instrumental in her failure to solve the problem in less than 350 trials. She usually appeared to take her orientation from door No. 1, going thence leftward to the correct door. The settings comprising doors at the left proved easy by this method. The most noticeable thing about her behavior was the suddenness with which she sometimes became thus oriented. Frequently when pausing and looking in the wrong door, or smelling about in almost any part of the reaction-chamber she would suddenly dash to door No. 1 and then across into the correct one. It is possible that this is the same phenomenon mentioned by Watson, 3 where a trained rat, placed in a part of the maze other than the entrance, runs about at random and suddenly darts off correctly, having passed over a "kinaesthetic unit" which arouses a proper sequence of kinaes- thetic impulses. Problem 2, which is definable as the second mechanism from the left end of the group, proved insoluble for the two rats C and B in 800 and 900 trials respectively. For the settings of this problem the total number of open doors is 50 with 10 of course correct. Hence the probability of Watson, J. B. Behavior. New York, 1914, p. 218. WHITE RAT AND THE MULTIPLE CHOICE METHOD 235 CM w D3 O K W ►J < w fa Q W « ffl W CQ < U H < cm OS o b CO H -J co W o £ 8 co CO 8 CO CO 88 co co 'oS »* LO CI r-H CM OiCi LO tt i—\ r-H 1— 1 1 — 1 HH r-H £ 00 o r— 1 o o CO f=cn r-H tt CM co r-H co CM r-H CO £ 00 O oo O r-H o O 00° Oj CD Oi OS 00 t> »2 ti CM r-H CM O r-H r-H r-H r-H O r-H r-H r-H r-H CM CO coo 00 o o NNN t> t>- o i— i c/i CO LO 3.4.5 6.7 4.5.7 o co O CO LO CO LO O LO co OLO OCOCO .4.5.6 .4.5.6 .6.5.6 t> CO- LO co_ LO od CO LO' CO CO CO CO CO TP CO ^* LO00 00 OOOOOi 00 CO t> t^OO CO ^F -tf ■^ CO ^ Oi CO CM CO r-H CO CM CO CO CO r-H ■o CO (/) CM CO r— 1 i— 1 TH< CO CO r-H r-H CO CM CM CM CM'* CO CO CO CM CO r-H ^fCM ^f Tfl *tf CM r-H r-H 1— 1 r-H r-H ■* -**& CM ^*r-H oo 00 oo 00 00 00 o LO 00 TfOON o CO tr- ee 00 o o 00 oooo LO CO C/} co CONCO i— I CD LO LO LO CD LO r-H o- co ■** COLO 00 CM ooc^— 1 00 oo t> 00 00 00 CO^J* LO LOLO "* O "* 'tf COCO ^t< ^ M*'* "* CO C/J CO CM ■* "tf CO LOCM COCO^ coco^ CO Tf CM ■* CM CM CM CMLO^* ■<* CM LOCO CM CMLO CM LO LO ^LO CO CM CM CO CO CM r-H CM r-H CO CO CO CO CO CM CM CM U) i—l i— i CO CO CM r-H CO CM CO CM CM r-H CM CM CM CM CO CM CM r-H r-H CO i— 1 CO r-H CO CO CO r-H CO CO CO r-H CO r-l r-l CO O o t- LO 00 o co CO O co LO o o cb CD O t> CO- CO LO t> CD en LO ^ CO O ** o CO LO o LONCO CD LO ^ LO ^ "tf ■^ CO 'CO co LO ^P 000 ^00 00 oo 00 00 OOO CO CO O LO LO LO LO m CO LO LOCO^t; OLO^fr-HLOLO^fCM "* LO^ "4 1 in CO CM r-H r-H LO CM^f CO 1.3. 2.4.1 6.2.3 -HTfCOr- 1 ^f ^ CO r-H LO MCOCMCMCOCOCMCM^ LO 00 CM r-H LO co • CM ^r-<' CO-- MO LO CO CDC O r-H CO r- 1 CD CD C M CM r-H r-H CO CO CO co CM CD o CO CO CO CO CD CD CD CON vO o ■* LO LO LOLO CO LO co CO CM COLO "*CO CO LO CO CM CM CM CM CM CO CO CD LO 3.4 2.3.4 4.5 ^CO CO CMLO LO CM CO CO LO tJ" CM LO ■^ LO CM -^ CO CO -# CM LO CO CM CM "tf t>t>co t>o-^ o CO ^ t> CO CM CM CO CO CO «* r-H co CO CO CO CO r-H CO CO CO CO C/3 CM CO CM CM CM r-H CM CO CM CM CM CO CM CO ■** r— I r-H r-H Tf CO^CM r-H r-H r-H co CO CO ^f-* 00 - 00 o OOC o oS o 00 00 00 Ui o ot> o 00 00 00 oo 00 00I> 00 00 00 ot> OCT o o o O o t^NN t>00 00 tr- oo CJ! (TO to o r-H o CM o CO o o LO o co OOO c-oocn OOO Or-HCM r-H r-H r-H o CO o r-H o LO r-H oo coo r-H r-H 6 2 -r-> 1 1 r-H 1 CM 1 CO 1 1 LO I l l CONM 1 1 1 CJiO r-H >-H r-H 1 CM r-H 1 CO r-H 1 1 1 LOCO r-H r-H 1 0) *-> 3 >> ■gco O CM a 00 CM a en CM S 3 3 OS r^ CO CO < 3 LO 3 3 236 HAROLD E. BURTT (M s w -j e c 05 c 2 «— i •w M <1> w a - 5? < '«*• ■*** S a w 3 Ph Q W > 05 32 2 < U H < 0- 05 O u. C/J H J 3 ;/) W K o £ 8 8 00 CO CO CO Q Q CM CO CO CO O © CO CO CO CO 8 CO CO 8 CO CO "3 2 -^ CO CM CM CM ■**■ •"* OJ CM -tf CM Oi CM K « »— ( i— i i— 1 r— 1 i-H i-H f— 1 t— 1 i — 1 . — i •—< i-H - 1 - 1 £ CM co os £3oOO00 00 o t> to o 00 i-H o 04 00 . — i O CO CO CO CM CM co co *tf co Ol co £ 00 a> loo-. OlNNNOO 00 o ooo 00 o cr-a> o o 04 CM i-H LO^h —iCO CO CO CM CM co CO CM CM CO ^H^H CO CO o CO o o td Tf O tOLO ^coo tO LO cot> O LOi— 1 LOtOtO c/3 1/3 >■* CO 00 .-HtO O OOLOOOO O00OOC000OO 00 LO 00 00 oo oc to oo OO 00 00 to LOCO OOCO^f o ood o o 00 CO CO "<* CM CO CM CO CO a> M CO Tf OJ r-i CM CO coco ■<* Cfl CSJ OJ co CO rH *-< CO CO Tf CO •-! ^H CO CO CMOJ ■-H CO T* rj< ■** CM Tf i— 1 "tf CM ■** CM CO CM^tf ^* CO i-H 7—1 CM ^f co oo CO oj 1 ^ 00 oo o 00 00 OO oooooo oo oc oo oo coco OO oo ■- 1 o ^to co 00 ^LO (T. LO - Tf 00 00 LO *# ■<*•* o «# ■<* ■<* LO COLO ■* to co CM CO CM^}* Tj< t^cO ■* ^ "* CM CM *tf LO ^ -* LO ■* LOLO CO CO LO ^f LO LO LO LOLO LO -* COTt< CO LO tO co CM CM CM Ol OJ CO 7; i-H CM CM CO —i CMCOCMCM OJ CO CM CM -H OJ CMOJ CM r-l CO CO CM ■— i CO CM CO i-h CO CO COCO co .-H coco CO CO NN NN o LO oq o to O td o to^t to to LO Tf t> LO LO o to to 1* C/3 LO 00 to -rf -^ tO'* •* LOO LO o oo o o ■<* oo LO t>- C^ 00 00 t> 00 00 00 t> oo oo oo 00 oooo 00 oo to