igre y aia i a as 4 is Laas ~*~ =, ty Os, Par bes es Sota ae 2518 f vy. SP ot a aes Pia i es ot h ‘ a: are * . ag: % oA 4 ew ee Ly Ltr hh * 4 + « * ~* “« seca ee ds Okt reat e, wwe aleiettatarevel olevadetarsiatelsleteleisteleele eisie)l\efols le 181 i ilieretin? IS e35 955 cog nannaede > COCR dOeE ened ar Dob Raano Do duoG TIC RU God SOOOnCMUE OD 201 Number 3, May, 1907. Concerning the Intelligence of Raccoons. By L.W. Corr. (From the Department of Psychology of the University of Oklahoma.) With two figures.............cece cece cece cecenee 211 The Ege-Laying Apparatus in the Silkworm (Bombyx mori) as a Reflex Apparatus. By IsapeL McCracken. (From the Physiological Laboratory of Stanford University.) With one HI DUITG aye taretesav aide cloves ehetesToielsl years, oisuslc:stotsleleral sleqerelovcyaretslnrsintchareorahs alata’ sa. vielepetavereiarsrnisiaraia/e 262 A Study of the Choroid Plexus. By Watter J. Meex. (From the Neurological Laboratory of the Unisersityo; Ghicago.)) With nine fipuress s/c ee vale sves amc asisis wae aieeaieie’> Hale we 286 Number 4, July, 1907. The Tactile Centers in the Spinal Cord and Brain of the Sea Robin, Prionotus carolinus. By C. Jupson Herrick. (Studies from the Neurological Laboratory of Denison University, No. AKT.) Wath fifteen figures 2 <:...0¢.5 do cles vee inue emer ae ener An Experimental Study of an Unusual Type of Reaction in a Dog. By G. van T. Hamitton. (From the McLean Hospital, Waverley, Mass.) With two figures.............0seeceeee The Normal Activity of the White Rat at Different Ages. By James Rotimn StonaKer. (From the Physiological Laboratory of Stanford University.) With eight figures..............+. Editorial, Concilium: Bibliographicum.. +... ... Jie. sis sis enue clans aos +s. ogc eons eee Literary Notices <3: Siw. /astac ecne ieee sane Bigtiae Oc Dame eee ae 6 « anise nee eee Number 5, September, 1907. The Homing of Ants: An Experimental Study of Ant Behavior. By C.H. Turner. (From the Zodlogical Laboratory of the University of Chicago.) With Plates II-IV and one figure in EEE CERES g aurra haseie ce Siateraicmce tino beards aaGhs Sais ha ehnne eRe uae Sols 2 ee The Behavior of the Phantom Larve of Corethra plumicornis Fabricius. By E. H. Harper. (From the Zoélogical Labortary of Northwestern University.) With five figures........... Literary Notices a err Number 6, November, 1907. The Nerves and Nerve Endings in the Membrana Tympani. By J. Gorpon Witson. (From the Hull Laboratory of Anatomy of the University of Chicago.) With Plate V............. A Study of the Diameters of the Cells and Nuclei in the Second Cervical Spinal Ganglion of the Adult Albino Rat. By Suinxisu1 Hatar. (From the Wistar Institute of Anatomy and Biology.) > With four fiplires: «2. «oc. vawdee uote oddest nos apse ae Anomalies of the Encephalic Arteries among the Insane. A Study of the Arteries at the Base of the Encephalon in Two Hundred and Twenty Consecutive Cases of Mental Disease, with Special Reference to Anomalies of the Circle of Willis. By I. W. Bracxsurn. (From the Government Hospital for the Insane, Washington, D.C.) With eleven figures ........... Editorials. ssc Ski's alos cGinx vier ana sd eg/ak sitios wma oleae eebatet ada aans a een Professor Gore on the Doctrine of the Neurone. Neurological Terminology. The International Zoclopical Congress...) ...icas vs eaeov ss oneness oes a= 1.08 ee Tsiteray- NOtCES ssccr sss cic s'5-<,cerujeleve ois Save Sera rarer ouelnTotagera te folevevetevaan er resaneicterstsr ses cyaes ote tera iv 307 329 342 360 364 367 435 457 459 469 SUBJECT WAND AUTHOR INDEX. VOLUME XVII. Author’s names are in small caps. References to subjects and authors of original articles are marked with an asterisk. * Asn, of rat in relation to age, 342. *Ameiurus, taste buds of, I. *gustatory paths of, 68. Animal behavior, section of the zodlogical congress, 524. *Ants, behavior of, 367. *Arteries, encephalic, of the insane, 493. Bere J.M. Mental Development, 98, 457. Bean, R. B. Negro brain, 184, 208. *Behavior, of ants, 367. *of Corethra, 435. *of dog, 329. *of lower organisms, 93. *of raccoon, 211. *of silkworm, 266. *of Volvox, 99. Bett, J.C. Depth perception, 189. Reactions of crayfish, 98. Bentiey, I. M. Intensity of sensations, 191. Brancut, L. Aphasia, 458. *BiackBuRN, 1. W. Anomalies of encephalic arte- ries in insane, 493. *Bombyx, egg-laying apparatus in, 262. Bosr, J.C. Plant response, 527. *Brain, of codfish, 67. of negro, 208. ees E.H. Total reactions, 191. Cannon, W. B. Bilateral vagotomy, 187. Cartson, A. J. Refractory period of heart, 187. Carpenter, F. W. Flight of birds, 98. and Main, R. C. Migration of medullary cells, 185. CassirerR, R. and Oprenueim,H. Encephalitis, 458, *Choroid plexus, 286. Chromotropism, 89. *Circle of W11tIs in the insane, 493. Crark, L. P. and Herter,C.A. Nervous diseases, 533- *Codfish, brain of, 67. Corr, L. J. Image-forming eyes, 193. Phototropism, 193. *CoirE, L. W. Intelligence of the raccoon, 211. *Communis system, 4. *Concilium bibliographicum, 360. *Corethra larve, behavior of, 435. Craig, colony for epileptics, 98. | D Socsaan U. and Sytvester,C.F. Electric organ of fishes, 366. Dawson, J. Reactions of Physa, 186. Dean, B. Chimeroid fishes, 98. Deraparre, E.B. Visual, pressure images, I9I. DieutaFre, L. Nasal fosse, 210. *Discrimination in raccoons, 226. Diseases, Nervous, 533. Dopps, G. S._ Brain of salamander, 366. *Dog, reactions of, 329. Driescu, H. Autonomy of life, 366. Drew, G. A. Habits, anatomy, and embryology of Pecten, 98. Dunn, E. H. Innervation of leg of Rana, 184. Boxes: L. Anatomy of cerebellum, 457. Brain of Myxine, 210, 364. Editorial, 360, 519. Encephalitis, 458. Essicx, C. R. The hind-brain, 185. EycresHyMer, A.C. Habits of Necturus, 366. *Eye-spot, of Volvox, 108. Fosse» brain of cod, 67. *taste buds of, 1. *tactile centers of, 307. Froriep, A. Origin of eye, 458. *Funicular nuclei, of codfish, 75. Ge Brain of, 67. Gacr, S.H. Glycogen in nerve cells, 185. *Ganglion, cells of spinal, 469. Gtaser, O.C. Movements of Ophiura, 198. Gotci. On the neurone doctrine, 519. Grasset, J. Insanity and responsibility, 98, 365. *Gustatory organs, I. *paths, 68. Guyer, M. F. Animal micrology, 98, 210. Hrs, G. V. Reactions of a dog, 329. Haraitt, C. W. Behavior of sea-ane- mones, 198. Behavior of annelids, 199. *Harper, E.H. Behavior of Corethra larvae, 435. Harrison, R.G. Peripheral nerves, 98. Transplanting limbs, 181. *Harar, S. Spinal ganglion, 469. Hatuaway, J. H. Eighth cranial nerve, 183. Hawkes, O.A.M. Nerves of Chlamydoselachus, 458. Herms, W. B. The bay shrimp, 366. *Herricx, C. J. Brain of codfish, 67. *Brain of sea-robin, 307. Herrick, F. H. Instincts of birds, 194. Herter, C. A. and Crarx, L. P. Nervous dis- eases, 533. Hopeg, C. F. and Detuncer, O. P. Structure of Ameba, 186. Hrpuicxa, A. Anatomy of orang skull, 98. Cranial fosse, 366. MAGES, in the raccoon, 249." *Imitation, in raccoon, 232. InceonieEros, J. Musical language and hysteria, 98. *Instincts, of ants, 378. *Intelligence, of ants, 367. *of raccoons, 211. Jers H.S. Behavior of lower organisms, 93. Reactions of starfish, 190. Jupp, C. H. Psychology, 532. Yale psychological studies, 458. Kate C. U. A. Anatomy of brain, 458. Kerr, A. T. Brachial plexus, 183. Lect= F. L. Taste buds of Ameriurus, 1 *Learning, in ants, 382. *in dog, 329. *methods of, in raccoon, 212. Lez, F.S. Fatigue, 186. Treppe, 186. Levi, E. Anatomy of brain, 457. Levi, G. Anatomy of brain, 458. ‘Linpsay and Buaxiston. Physicians’ visiting list, 98. Linton, E. Habits of Fierasfer, 210. *Locomotion, of Volvox, 112. M Auestes C. N. Rotation of eye, 189. *McCracken, I. Egg-laying apparatus. of silkworm, 262. *Masr, S. O. Reactions of Volvox to light, 99. Maxwe tt, S.S. Chemical stimulation of the cor- tex, 98, 365. Mayer, A.G. Pulsation in animals, 200. Meap,G.H. Animal perception, 189. *Meex, W. J. Choroid plexus, 286. Mettuvs, E. L. Frontal lobe of monkey, 183. Me zzer, S. J. and Aver, J. Section of the vagus, 186. *Membrana tympani, nerve endings in, 459. *Memory, in ants, 412. *in raccoon, 225. *Mental disease and anomalies of arteries, 493. Mercatr, M. M. Salpa and the vertebrate eye, 98. Mertter, L. H. Clinical physio-pathology, 366. Minxrewicz, R. Chromotropism and phototro- pism, 89. Monrog, W. S. Incipient fatigue, 191. Monrcomery, T. H. Scientific thought, 98. Morris and McMurricu. Human anatomy, 457. Noareet History, of Volvox, 101. *Nerve cells, diameter of, 469. *Nerve endings, in membrana tympani, 459. *Nerves, on membrana tympani, 459. Nervous diseases, 533. *Nervous system, of Ameiurus, I. *of codfish, 67. *functions of, 201. *nomenclature of, 522. *Neurological terminology, 522. Neurology; at the zodlogical congress, 524. Neurology, cases, 366. Neurone, doctrine of, 519. Neuropathology, 210. *Nuclei, of nerve cells, 469. QO essen, H. and Cassrerer, R. Encepha- litis, 458. *Organs, of taste, I. *Orientation, of Volvox, 115. *theory of, 436. Pare: G. H. Senses of Amphioxus, 197. Migration of pigment, 366. and Mercatr,C.R. Reactions of earthworm. 66. Paton, : Reactions of vertebrate embryo, 457. Peart, R. Variation in Ceratophyllum, 366. Perers, A. W. Chemistry of cell, 210. Phototropism, 89. *of Corethra, 444. *of Volvox, 99. Plant response, 527. Porter, J. P. Variability in spider webs, rgr. Porter, W. T. VWasomotor reflexes, 186. *Prionotus, nervous system of, 307. Psychology, Jupp’s “General Introduction,” 532. *of ants, 367. * *of dog, 329. *of raccoon, 211. Psycho-physiology, 91. RR coon, intelligence of, 211. *Rat, activity of, in relation to age, 342. *cells of, in spinal ganglion, 469. reactions of, 530. *Reactions, of ants, 367. *of Corethra, 435. *of dog, 329. *of raccoon, 211. of rat, 530. *of Volvox, 99. to light, 89. to color, 89. Reaction-types, 329. Reese, A. M. Habits of alligator, 194, 458. Response, in plants, 527. *Right-handedness, in dog, 335. Rirey, I. W. Mental diseases, 191. Rowe, S. H. Habit and idea, 189. S22 F. R. Model of fiber paths, 182. *Sea robin, tactile centers of, 307. SrasHoreE, C. E. Tone-psychology, 191. *Sense organs, of Ameiurus, I. Sensations, kinesthetic and organic, 530. SHEPARD, J. F. Cerebral circulation during sleep, 189. SHERRINGTON, C. S. Functions of the nervous system, 201. *Silkworm, egg-laying apparatus of, 262. Sirvester, C. F. Electric organ of fishes, 185. *SLoNAKER, J. R. Activity of rat, 342. Smattwoop, W. M. Branchiobdella, 98. Smitu, B. G. Habits of Cryptobranchus, 197. SmitH,G. Eyes of gasteropods, 98. Stocxarp, C. R. Median eye in fishes,1gr. Srratron, G. M. Ocular movements, 189. Streeter, G.L. Inter-forebrain commissures, 183. "Tac centers, of Prionotus, 307. *Taste buds, 1. *Temperature, reactions of Volvox to, 161. *Terminal buds, 3. Vil *Terminology, Neurological, 522. Terry, R. J. Skeleton of Amblystoma, 210. *Threshold of stimulation in Volvox, 163. *Touch, in sea robin, 307. Tower, W. L. Evolution of beetles, 210. *Trial and error, in Corethra, 442. *Tropism, chro- and photo-, 89. *of ants, 370. *of Corethra, 436. *theory, 436. Turner, C.H. Behavior of ants, 367. U Vor lobes, of codfish, 67. vaN GEHUCHTEN, A. WALLER’s law, 365. *Volvox, reactions of, to light, 99. \ X J attenperc, A. Acustic tract of dove, 366. Watson, J.B. Reactions of rat, 189, 530. *WEBER’s law, in Volvox, 171. Weysse, A. W. Animal behavior, 98. and Burcess, W. S. Histogenesis of retina, 98. Witper, B. G. Anatomical nomenclature, 210. Brain nomenclature, 98. Shark brains, 192. *Wixson, J.G. Nerve endings in membrana tym- pani, 459. Woopworth, R. S. Sense perception, 191. Wricut, R. Anadidymus of chick, 210. NDERHILL, F. P. and Menpet, L. B. Adap- tation in salivary glands, 187. Yous, K. Brain anatomy, 98. Yerkes, R. M. Functions of the ear, 186. Behavior of dancing mouse, 190. Vision of mouse, 194. al . . OGLOGICAL congress, animal behavior and neurology at, 524. Z, The Journal of Com eee coal and een VotumE XVII JANUARY, | 1907 NuMBER I ON THE PLACE OF ORIGIN AND METHOD OF DISTRI- BUTION OF TASTE BUDS IN AMEIURUS MELAS. BY Fale DANDACKE: (Associate Professor of Zodlogy, Ohio State University.) Wirn Pirate I aNp Four Ficures IN THE TEXT. CONTENTS. LOSEN XOPDLUKCN OO) igian cota doco cepa aeons e Had eos cers Sn eo cian MEOW c Copan omdcomuancqccD Ont I tan Jeb AOE S) OSC gnc Re ISS EAD EM De ar Sain 1008 ths cae Goon a toon camo nocdnersacd caaicoS 2 (G)eeliactelomenminal Duds: o50.cn06 occ pobeUUDUL MC Ome MD NacC 61 INTRODUCTION. The present investigation was undertaken with the object of ascertaining, if possible, at what time and in what order taste or terminal Keds appear in the anterior oral cavity and on the outer body surface of Ameiurus melas as compared with those situated in the pharynx. 2 ‘fournal of Comparative Neurology and Psychology. The siluroids, owing to the enormous number of taste buds scattered over practically the whole of the body surface, seemed the most favorable group for such a study and Ameiurus melas in particular was selected on account of the very complete descrip- tion of the distribution of the gustatory fibers which has been given for this type by C. J. Herrick (01). Owing to the fact that all taste buds wherever located are inner- vated by communis fibers and that those in the endoderm of the pharynx appear in some forms before buds appear in the ecto- derm, it was thought that buds would be found to spread from endodermic into ectodermic territory. This has proven not to be the case in Ameiurus and if a study of less specialized or more primitive forms shows this suppo- sition to be the general rule, a view which has been advanced by JOHNSTON (05a), the condition in Ameiurus will have to be explained on the basis that the gustatory system is so highly specialized that phylogenetic relationships have been materially modified or obscured. While the results obtained do not show a derivation of ecto- dermic from endodermic buds nor the reverse, as suggested by CoLE.(’00, p. 320), still a careful study of the time of appearance and the direction of spreading has brought to light a number of interesting facts, which, when related to the nerve supply, throw light on the problem of the distribution of ectodermic buds, if they do not answer the question as to how these two groups are related to each other primitively. i HISTORICAL: The attempt to correlate the distribution of taste buds with the distribution of gustatory fibers in the various cranial nerves is rendered possible by two somewhat distinct lines of research. ‘The first culminated in the complete isolation, both structurally and functionally, of the taste or terminal buds from all other cutaneous sense organs. [he second culminated in what is commonly called the theory of nerve components in which we have the isolation of the various components of the cranial nerves, such as the gus- tatory, the lateral line and general cutaneous, based on a differ- ence in the size of their fibers, and the isolation of their central endings or nuclei in the brain from each other as well as on the difference in types of sense organs supplied. Lanpacre, Taste Buds of Ameiurus. 3 In the brief historical outline which follows no attempt is made to give a digest of the earlier literature on the cranial nerves of fishes, much of which is confusing and still more of which is of little value for the present purpose, owing to the fact that it de- pends for its analysis of the cranial nerves on the method of gross dissection, a method which is totally inadequate where it 1s pos- sible to separate the various components only by a microscopic study of serial sections. The review of the theory of nerve components makes no claim to completeness either, it being the writer’s intention to present the salient points of that theory as far as they would be of value for the present paper, which is concerned solely with the com- munis system. For a comparison of the results of the micro- scopical analysis of the cranial nerves with the earlier attempt by gross methods the reader should consult the papers of STRONG, HERRICK, JOHNSTON, KincsBury, CoLe and others. (a) Taste or Terminal Buds.—The taste buds were first dis- covered on the palatal organ of the carp in 1827 by WEBER (27) and their function correctly inferred. In 1851 Leypic (’51) dis- covered the taste or terminal buds on the outer skin of fishes but gave a somewhat faulty description. In 1863 F. E. ScHULZE (63) described the same organs and distinguished structurally between the sensory and supporting cells and inferred their func- tion to be the same as similar buds found inside the mouth. ‘The same author showed in 1870 that the terminal or taste buds differ in structure from the lateral line organs or neuromasts of what- ever form in that the sensory cells of the taste buds are elongated cells and pass down from the surface of the epithelium entirely to the basement membrane, while the sensory cells of the lateral line organs and of all superficial organs related to them are pear- shaped and do not reach the basement membrane. MERKEL (80) in 1880 confirmed these structural differences, but confused the subject by attributing the same function, namely, touch, to both terminal or taste bale and neuromasts or lateral line and related organs. In 1904 C. J. Herrick (04) demonstrated by experiments that the function of the terminal buds is undoubtedly gustatory, and in addition showed that the cat fishes, at least, can locate sapid substances by the sense of taste and can learn to distinguish between gustatory and tactile stimuli, although ordinarily these 4 fournal of Comparative Neurology and Psychology. are probably used in conjunction in locating food. ‘These experi- ments, reinforced by the complete isolation of the gustatory fibers and gustatory centers microscopically, leave no doubt that we have here a sensory system quite distinct functionally and struc- turally from any other cutaneous sense organs. As to the exact time and place of appearance and the direction of spreading of the buds, less work seems to have been done. Autts (’89) calls attention in Amia to the time of appearance of what he takes to be terminal buds and describes briefly the manner of spreading posteriorally from the anterior parts of the head back to the body. He describes (p. 509) and figures the terminal buds as appearing as whitish lines, usually parallel with the lateral lines of the head, that later break up into individual buds. The serial arrangement disappears as the buds become more numerous. JOHNSTON (05a) states that buds are present in the branchial region only ofthe Ammoceetes stage of Petromyzon but that they are present in the skin also of the adult. Corregonus and Cato- stomus sp. were studied also in serial sections, and the taste buds of the pharynx found to be much more numerous and highly de- veloped than elsewhere. Buds were found in the cesophagus and on the roof and floor of the mouth but in both places were smaller than those in the pharynx. No buds were found on the skin of the body except in a specimen taken some days after hatching. These facts, 7. ¢., the earlier appearance of the pharyngeal buds, their larger size, and the fact that all taste buds are innervated by communis fibers, inclined JoHNsTON to the opinion that buds have spread from the endoderm to the ectoderm. (6) The Communts S ystem.—The need of the application of histological methods in the study of peripheral nerves and the substitution of components as physiological and morphological units instead of nerves may be traced to GASKELL (’86,’89). The two-root theory of BeL_ had dominated the study of spinal and cranial nerves from the time BELL’s law was enunciated in 1810 until GasKELL proposed what is called the four-root theory. Ac- cording to GASKELL a typical spinal nerve contains four roots, somatic sensory, somatic motor, splanchnic sensory and splanchnic motor. STRONG (95), following a suggestion of H. F. Osporn (88) that it might be possible to analyze the cranial nerves on the basis LanpacreE, Taste Buds of Ameiurus. 5 of structural differences in their components and peripheral dis- tribution, made the first successful attempt at such an analysis on the sensory portions of the cranial nerves of the tadpole of the common frog. As a result of this analysis StRoNG found that the sensory portions of the V, VII, VIII, LX and X nerves could be resolved into three components. 1. The general cutaneous component, characterized periph- erally by having medium sized fibers and free nerve endings, and centrally by ending in the spinal fifth tract which is a continua- tion of the dorsal horn of the cord. 2. The acustico-lateralis component, characterized peripherally by having coarse fibers and innervating the ear and lateral line and related organs, and centrally by ending. j in the tuberculum acusticum. 3. The communis component, Fea peripherally by innervating unspecialized mucous surfaces and specialized or- vans in the form of terminal or taste buds, by having fine fibers, and centrally by ending in the lobus vagi and lobus facialis, which are really one nucleus morphologically. STRONG’s work was followed and confirmed in all essential details by that of Herrick on Menidia (’99), on Gadus (’00), on Ameiurus (01); and by Cote on Gadus (’98) and on Pleuro- nectes (’01); and by CoGuiLt on Amblystoma (02) and on Tri- ton (06); and by JoHNnsTon on Acipenser (’98) and on Petromy- zon (’05), and others, so that we have a thorough knowledge of the components of the cranial nerves based on a careful micro- scopic study of serial sections for the cyclostomes, teleosts and amphibia. In all these papers the necessity of correlating the peripheral distribution of the nerve with its central endings and the unraveling of the nerve throughout its whole course has been kept in mind and accomplished with an unexpected degree of success. A study of the central endings of the cranial nerves in the medulla and their homologies with the centers of the four roots of GasKELL has been made particularly by KincsBury (95a, 97) and by JoHNsTON (’98, ’O1, 02a), in addition to the papers men- tioned above, while the larger problem of the morphology of the vertebrate head has been attacked by JoHNsTON (’05b) from the functional standpoint and in accord with the work on nerve com- ponents. Since taste buds are innervated wherever situated exclusively 6 Fournal of Comparative Neurology and Psychology. by communis fibers, we may confine our attention to this com- ponent. As mentioned above, all communis fibers end ina mor- phologically single center, those from the ninth and tenth nerves in the lobus vagi and those from the VII nerve in the lobus facialis, which correspond in the position in the medulla to the visceral sensory or region of CLARKE’s column inthe cord. All communis fibers running out through the VII nerve trunk and through what are commonly designated as the fifth rami come from the genicu- late ganglion of the VII nerve, while those running out through the ninth and tenth nerves come from ganglia situated on those nerves. “The communis fibers coming from the geniculate gan- glion enter the brain anterior to the position of their nucleus and pass back as the fasciculas communis or fasciculus solitarius of mammals to the facial lobe. The communis component consists of unspecialized fibers sup- plying mucous surfaces and specialized fibers supplying taste buds situated in both ectoderm and endoderm. ‘These specialized and unspecialized divisions cannot be accurately separated cen- trally although Herrick (’05) has worked out the reflex gusta- tory paths both ascending and descending in the cyprinoids and siluroids. ‘This is the first successful attempt to trace definitely the secondary and tertiary gustatory paths in any vertebrate, and from being the least known system centrally, the communis sys- tem, at least in teleosts, 1s one of the best defined systems both peripherally and centrally of any of those contained in the cranial nerves. The communis fibers running out through the [X and X nerves are fairly constant in their distribution. The IX nerve may send fibers to the surface of the body, as in Menidia, but the distribu- tion seems usually to be confined to mucous surfaces as in Ameiurus. The communis fibers arising from the geniculate ganglion, however, show the greatest diversity in the number of rami through which they pass to the surface and in the extent and variety of the surface innervated. “They may pass out through only one ramus, as in Petromyzon (JOHNSTON ’05), or two as in Rana (STRONG ’95), or three as in Amblystoma (CoGHILL ’02) and ‘Triton (CoGHILL ’06) and Amia (ALLIS 97) or five as in Pleuronectes (Core °o1), Gadus (Herrick ’00), Menidia (HERRICK ’99), or even as many as twelve rami in Ameiurus (HERRICK ’O!). LanpacreE, Taste Buds of Ameiurus. 7; Since both ectodermic and endodermic buds are innervated from the geniculate ganglion and it shows such a diversity 1 in the num- ber of rami of the V and VII nerves used, and since the interest in the relation of ectodermic to endodermic buds centers in the distribution of these nerves, a table has been arranged to show these relations. TABLE I. Table showing the rami of the V and VII nerves which carry communis fibers in types in which they have been worked out fully. In this table rami carrying communis fibers are indicated by (x). In the Amphibia and Petromyzon the presence of the ramus or truncus is indicated by (”), and when a ramus is known to be absent it is indicated by (—). The rami as outlined here are fairly constant for the Teleosts and Amia and the main rami are also present in the Amphibia and Petromyzon, although no attempt has been made to analyze these and determine their homologies. The ramus oph. profundus is probably not present in Teleosts, and on the homology of the ramus max. acces. and ramus max., see Cocuitt ('01). 2 ENT eee ou icret ieee, | Sere SHS o ney ecmlpaa leis] (vo ol eile also ee ta eal ete Pe = a a Ta) ee ices ice ee poet ED | cae (ey | | ‘Ramus GPbe PLO eyeseststersysrel ste) acs lei >=) | |) at id Mi ReaMUSOLICUSAVI Hee reer rte tate = eke oe a) fave x Ral eves WIL aos conusmee eee on c ool) ee eke a exe | aS ae | Gr) ANAT RIE oldontts owe cies b Olndn oS an Ddao ee | IREEMAUIS{O PAN SU pon Var arctebteasiey 2-3-1 e/e)ore) kee x x x lige Y u RAMAMSIO PUSH ps VUE yest ietls.F-1\0 21> = = | Far eae Ka ag NL eMinGusanira-OVpit,, woeienciess sere «cls se oe catyee as | | ERATEUNYGHETL ATIC © Voeatat oncteren eta cterereys echoes wis | ad} we Se Mian dt exer oye ect ies vce ache ts | x | | | ae WY | Mimic eintep ve es arte eetote on emt eters shite cr 0) x | x x x | RamMUsinax VAP reer orien) ans a tee CNS at AVE a tai secre ree heieyeeo ten eaewnccrent's y=? seis lex | | Mifaxe mes. 0Vi gs Sess evclatere eit c/abte lore sas Peeler || | (Nami sim axed COONS srersiete deac Se peosfers i ee x | | | va Ramus buce ville! we ee see hse ic tyes 4 | (ees ui Ui Pa ge IBucnextera Wile ea tay lok aie. oes IB ICHIMCELAN eaten acta reahetere ere ae ie 2 |Rernits PHELAN, 644 ooonaooosoudoeo oe exer | x >a |e ell Wee Sa Nae x | x Rename yoallprasis WOE con So cenese case Di hes Pe W ge || by Hees | x pimncuspiyonandlonerrne tee rie telat cicesencie = | | ee ate ISAS Oe} WUD 3 Son oaeeogyoosduee — RamaisphyOid evs) Wills aya2 = efecsy eis 2 o1< = o's ae | x | A aoa bang uf Ramusimiand. ex.) Vl tyryars sees scree atl| x | | i fs Us | Ramusimand asimteilileews an toeey-tan ncaa: St a te | 5. Cal| Apel > < | x 8 Fournal of Comparative Neurology and Psychology. Although the work on nerve components has shown several of the names commonly used to designate these rami to be inap- propriate in a number of types, in order to get some basis on which a comparison of the number of rami carrying communis fibers could be made, the current nomenclature has been used. ‘The table is open to criticism, of course, in this regard, but it does bring out the fact that of the rami usually attributed to the V and VII nerves a great variation exists in the number carrying communis fibers. In the case of Petromyzon, one branch of the hyomandibular trunk carries communis fibers and JOHNSTON does not state, so far as lam aware, its homology with corresponding rami of the teleosts. In the case of Rana, I have catalogued only two rami as bearing communis fibers, namely, the ramus pala- tinus and the ramus mand. int. VII, but the ramus palatinus near its peripheral distribution forms anastomoses with the ramus oph. V and with the ramus mand. int. VII, so that there are really four rami in the frog carrying communis fibers, although they do not enter these rami near their ganglia. The teleosts, owing to their close relationship and the constancy of these rami, furnish the best basis for comparison. . Of the fourteen rami usually found in teleosts, two, the ramus mand. int. VII, and ramus oper. sup. VII, are absent in Ameiurus, and of the twelve remaining, ten contain communis fibers and these fibers are absent from ahe ramus oph. sup. VII and ramus buccalis VII only. An examination of Menidia shows only five rami carrying communis nerves; the same is true of Gadus, but while these two types have the same number of rami carrying communis fibers the rami are not identical. Menidia has com- munis fibers in the ramus mand. int. VII, which is absent in Gadus while Gadus has communis fibers in the ramus mand. int. V which is present in Menidia but contains no communis fibers. The distribution of the taste buds on the ectoderm is correlated more or less closely with this variation in the rami carrying com- munis fibers, although peripheral anastomoses sometimes ma- terially modify the Getnibacen of taste buds from what we should expect to find, judging fromthe number of rami carrying communis fibers at a point near the ganglia. The same variation is found in Amphibia. Amblystoma and Triton each have three rami which carry communis fibers that run out with these rami from near the ganglia, while the frog has only two. LanpacreE, Taste Buds of Ameiurus. 9 These differences cannot be explained on the basis of a larger or smaller number of communis fibers in a particular ramus in various types, such as the teleosts, for while no enumerations have been made, it is altogether probable that Ameiurus not only con- tains more rami bearing communis fibers, but contains more com- munis fibers in a given ramus than either Gadus or Menidia or Pleuronectes, if we may judge from the number of taste buds on the surface in a given area. Of course, we would have to except from this general rule the nerves supplying the palatal organs in the cyprinoids. Judging by the variable number of rami of the V and VII nerves used by communis fibers in the teleosts, it would seem that the constancy of these rami in this group is not main- tained primarily by the communis system, but has to be accounted for by the phylogenetically older and more constant general cuta- neous system and in some cases the lateralis system of fibers. In an attempt to explain how this variation came about it Is 1m- portant to keep in mind the proximity of the ganglia of the com- munis, general cutaneous, and lateralis components in the tri- gemino-facial complex, which seems to furnish a clue to the way in which communis fibers have used them in some types and have not used them in others. In Ameiurus we have an extreme case of this usurpation where ten out of twelve rami carry communis fibers. Since the brain center is constant in position and mor- phologically single in all these types, and since fibers grow both peripherally and centrally from the ganglion cells, it would seem that our attention should be given largely to the ganglion as a source of variation. The question of the relation of the fibers to the taste buds will be taken up under the general survey of the oral and cutaneous groups. 2. MATERIAL AND METHOD. In order to determine the place of first appearance and the rate and manner of distribution of the taste buds, a number of series of young Ameiurus melas were taken and the total number of buds enumerated and their locations tabulated. The most complete series and the one from which most of the tabulations were made was found in a nest at Sandusky, Ohio, on July 1, 1905, at 3.30 p.m. Both parents were on the nest and the eggs were apparently just being deposited and fertilized. The nest was conveniently located near the edge of a pond so IO Ffournal of Comparative Neurology and Psychology. that the eggs could be removed with a pipette at desired intervals without disturbing the male parent. As segmentation had not begun when the nest was found, no eggs were removed until the following day at 3 o’clock, when the blastoderm was found to cover about half the yolk. Forty-nine hours after the nest was found eggs were removed and as the embryo was sufficiently developed to render the presence of taste buds possible, series were taken at intervals ranging from four to fourteen hours up to the eighth day, and on the ninth day a thirty-one hour series was taken when the eggs in the nest were exhausted. After ex- amination it was found that in the oldest embryo of this series buds were not present back of the operculum on the body and later series were taken the following year. Since no nests were found in which the eggs were being fertilized, so that the exact age could be determined, these embryos were arranged accord- ing to measurement rather than age. The oldest embryo of the first series was 9.4 mm. in length, and the later series were selected so as to complete this, embryos being cut which measured II, 14%, 154, 174 and 20} mm., respectively. In the first series five lots were taken for each stage and fixed in the following fluids, (a) HERMANN’s, (6) FLEMMiNG’s, both the weak and the strong solution, (c) a chromo-aceto-osmo-platinic mixture, and (d) ZENKER’Ss fluid. A number of stains were tried but HEIDENHAIN’s iron hema- toxylin after ZENKER gave the best results. ‘This stain gives a sharp differentiation to the taste buds, sometimes staining the sense cells and sometimes the supporting cells. In the ear and lateral line organs it usually stains the pear shaped sensory cells quite black, thus rendering the separation of the neuromasts from taste buds quite easy. Every alternate series beginning with the earliest was cut, but no buds were found until series K’ (113 hours). ‘This is about 24 hours before hatching, the incubation period being about six days. The earliest buds to appear are of course very immature, but little or no difficulty was experienced in recognizing them as soon as the first rudiments appeared. The only organs with which they could possibly be confused would be the teeth in the oral and pharyngeal cavities and the neuromasts in the cutaneous regions. ‘The teeth appear first as invaginations of the Malphigian LanpacreE, Taste Buds of Ameiurus. Il layer of the epidermis, while the taste buds always begin as an evagination of the same layers. On the outer surface the same distinction exists between the taste buds and the neuromasts. All neuromasts are in the form of more or less well defined de- pressions of the epidermis at the time taste buds appear, while the taste buds, as inthe mouth and pharynx, begin as elevations. Even in cases where the neuromast does not begin as a well de- fined depression, the presence of two layers of cells, the outer one of which is rounded and lies quite near the surface and will later become the pear shaped sensory cell so characteristic of neuro- masts, makes the separation of these two types of organs easy. In order to avoid any possibility of confusion all aie lateral line organs of all the series in which taste buds occur and all the neuromasts present in the earlier stages have been located, and at the time the first taste buds appear all the placodes of the lateral] line organs are present and accounted for. In this connection it may be of interest to note the difference in time of maturing of these two types of organs. Perfectly formed lateral line organs with’ the sensory cells having the same shape as in the adult and with their free borders exposed can be found as early as series N, 138 hours. This is about the time of hatching and 1s probably correlated with the effort of the young fish to right and orient themselves, which occurs about this time, although no notes were taken of the exact time when this occurs; nor were any notes taken of the exact time at which the young orient themselves in the nest after they can maintain an erect position. ‘This orientation 1s quite a characteristic phenomenon of larve of very young cat fish, however. On the other hand, no taste buds are found in my series that could be considered mature until much later, probably in series QO, 174 hours. It is much more difficult to determine when taste buds mature on account of the irregularity in the staining of the sensory cells in the taste buds. There are, however, in series QO, both in the oral cavity and in the pharynx, buds that resemble mature buds of the oldest embryo studied in all essential details except size. The later maturing of the taste buds is probably correlated with the reduction in amount of yolk and the begin- ning of feeding on food secured from the surrounding water. As to the relative time of maturing, there is no apparent dif- ference between the oral, pharyngeal and cutaneous regions. Ma- Lo Fournal of Comparative Neurology and Psychology. ture buds appear in the pharynx, and in the oral cavity, and on the maxillary barbule, at the same time as nearly as I can deter- mine. Nor does the difference in size between the buds in the three regions which JOHNSTON (05) finds in Catostomus and Cor- regonus exist in Ameiurus. ‘[’his was to be expected after it was found, as will be shown later, that buds arise in the oral and pharyngeal cavities simultaneously. Series were all cut 7 microns in thickness and all tabulations are based on serial sections, since nothing could be made of the distribution of taste buds from surface views, largely on account of the amount of pigment i. che epidermis. Since the question as to whether a given group is spreading forward or backward from the position in which the buds first appear in a preceding series depends upon the actual length of the specimen compared with the limits of the groups in the two series, a table is given showing the exact age and the age increment, and the average length in all fixing fluids, of the specimens of each age with the average increment in length of each age. TABLE II. Table showing ages in hours after fertilization of series K’ to U and the average lengths in mm. of these embryos in all fixing fluids and the increments in age and length in each case. EMBRYO. AGE. | Ace Incr. | LENGTH. | Lenorn Incr. Ke 113 | 8 BE73 1G, 120 | 7 6.55 .82 M 128 | 8 6.36 09 N 138 fe) 7.06 -70 OF 146 8 Ti 33 27 O’ 155 9 7.54 21 JE 163 8 7-59 05 1@) 174 II TOD) 23 R 183 9 8.33 51 S 199 16 8.50 | 17 a 213 14 8.70 | 20 U 244 | 31 9.40 70 There are three possible sources of error to be taken into con- sideration in attempting to determine the exact position of a group of buds located on structures that are not segmental and the man- ner of their spreading. (1) All embryos probably do not shrink uniformly. Although a table prepared, but not given in this paper, shows a rather remarkable uniformity in the various fluids, Lanpacre, Taste Buds of Ameturus. 13 still some variation may be due to this factor. Greater shrinkage in an embryo in which a group of buds had appeared in an earlier period would make the group seem to spread forward if measured in sections from the anterior end. (2) All the earlier series were not taken directly from the nest; for the first three days two lots were taken each day and intermediate series were taken from those brought in at a previous trip. Embryos reared in the lab- oratory from the earlier stages until the time of hatching prob- ably do not grow so fast and undoubtedly do not become so thoroughly pigmented as those reared in the nest, but this prob- ably does not materially affect the rate of differentiation of taste buds in the series under discussion, since the intérvals were too short. (3) The most puzzling question that arises in attempt- ing to determine whether a group of taste buds, not located on structures segmentally arranged, is actually spreading in a given direction arises from the unequal growth of various regions of the body. When we give the limits of a group of buds in various ages in sections counted, as must be done sometimes, from the anterior end of the body, we assume that the anterior end of the body is a fixed point. ‘This of course is not true. A group of buds might show if measured in this manner a well defined spreading back- ward and still not be spreading backward at all, but be simply increasing in a given area which owing to the growth of that area and of areas in front of it, comes to lie further from the anterior end; or it might be moving bodily back with reference to other structures near it. The branchial apparatus, for instance, owing to its functional importance elongates much more rapidly than other structures occupying similar segments of the embryo and comes to occupy many more segments in the adult than in the embryo (JOHNSTON ’05b), and the anterior taste buds situated on the proximal hyoid and suspensorium move backward almost as rapidly as the posterior buds, showing a backward movement of these two structures, while on the other hand some portions of the nasal group of buds remain practically stationary. For practical purposes in determining how fast and in what direction a group is spreading the best means seems to be to reduce the older embryo to terms of the younger in length. ‘That is, if a group of buds extended twice as far from the anterior end in series B as in A and series B had increased less than twice as much in length as A we would be warranted im thinking that the group 14 Fournal of Comparative Neurology and Psychology. had actually spread back and had not simply moved with the area:-on which it was situated. Both these conditions are found, as will be shown in the analysis of the groups. ‘This rule seems to be particularly applicable to the solid parts of the head and body, rather than to the gills, where the spreading can be deter- mined segmentally much more accurately. 3. THE ORAL, PHARYNGEAL AND CUTANEOUS GROUPS. In order to ascertain whether taste buds appear first in areas that are undoubtedly ectodermic or in areas that are endodermic, buds were classified roughly into three groups: (2) The oral group, comprising buds lying on the anterior portion of the mouth, both roof and floor, and extending as far back as and including the proximal hyoid arch and suspensorium. (b) The pharyngeal group, comprising buds lying on the roof and floor of the pharynx, on the gills, and extending as far for- ward as and including the ventral portion of the hyoid arch and extending posteriorally into the cesophagus. (c) ‘The cutaneous group, including all buds lying on the dorsal and ventral portions of the head and on the barblets, operculum and posterior portions of the body back of the operculum. This group is, in the anterior portion of the head at least, a continua- tion of the anterior oral group, and is catalogued separately for the purpose of a preliminary survey only. In tabulating the buds of the barblets it was not practicable, of course, to place them in the sections in which they were found, since they might lie in almost any position, nor in the sections in which the proximal portions lie, since the base varies in size at different ages and always occupies a number of sections; nor was it possible to ascertain accurately their number in later series, since they are very numerous. The barblet buds are not in- cluded in Table II, since the main object of this tabulation is to show the place of first origin and the rate of progression from this point. ‘They will be given in separate tables and discussed fully. The buds of the barblets should be added to the total of Table II to ascertain the exact number of buds in a given series. From Table II it will be seen that buds appear simultaneously in the anterior portions of the mouth (really on the inner surface of the lips and breathing valve, which cannot be distinguished at this age) and in the pharynx. The buds of the pharyngeal ~ LanpacreE, Taste Buds of Ameturus. £5 group appear simultaneously on the first, second and third gill arches. No buds appear on the skin of the outer surface of the head until in series N, 25 hours after they appear in the oral cav- ities and the pharynx. ‘The first to appear on the outer surface of the head spread continuously from those just inside the lips. TABLE III. Table showing gross distribution in oral, pharyngeal and cutaneous groups. AGE IN No. or Bups 1 | No. or Bups In | | No. or Bups In Empryo. Inc. Inc. | Inc. | Hours. | Mourn. | PHARYNX. | SKIN. | KZ rel | 4 [sant 10 | 10 — — L 330 | 10 6 ite) ° — — M 128 | Pai II 12 2 — ae N 138 2 3 47 35 14 Ws O 146 | 56 4 69 22 30 16 O’ 155 65 9 80 IX | 23 7 | 12 163 84 | 19 102 22 | 34 II Q 174 94 10 147 45 36 2 R 183 86 8 201 54 | 35 I | | S 199 113 | ag] | 210 | 9 | 49 14 AE 213 146 |ni99 352 | ee | 72 | 23 U 244 | 188 | 42 | 477 ‘ 125 | 117 45 In che case of the Poel barnes mete an exception to this statement might be taken. In discussing those buds later they are catalogued as appearing in K’; but, since the lateral portions of the upper lips are continuous with the inner and lateral por- tions of the maxillary barblet, it is impossible to state positively, owing to the relation of the barblet to the lips, whether they are inside the mouth or on the outer surface of the head. “The above statement 1s correct for all cutaneous buds aside from the doubt- ful ones on the maxillary barblet. From the table it will be seen also that the total number of buds in the oral cavity is greater than that in the pharyngeal for the first four series excepting in K’; from series O’ on the total number of buds in the pharynx 1s greater than in either the oral or cutane- ous groups excepting the maxillary barblet buds and after series QO is greater than both of them combined. This inequality may possibly disappear after the cutaneous group secures its full com- plement of buds, which is much later than any series studied. The relation of the oral and cutaneous buds will be discussed 16 fournal of Comparative Neurology and Psychology. later under the analysis of the smaller subdivisions of which these are composed. So far as Tables II and III are concerned, the cutaneous group may be considered as an extension of the oral buds out over the upper and lower lips and so to the head and body. It is almost impossible to tell whether a bud situated on the anterior edge of the lip is to be counted as oralor cutaneous. For these particular areas buds above the dorsal boundary of the oral epidermis of the upper lip were counted as on the top of the head and the same arbitrary rule was applied to those of the lower lip. And while the appearance of buds on the body is by discontin- uous groups, as will be shown later, we may consider them for the present as simply extensions of the buds within the mouth with which they are continuous at that point. The simultaneous appearance of the two primary groups and their wide separation, one lying in the ectoderm of the extreme anterior portion of the oral cavity and the other in the endoderm of the pharynx, renders the exact determination of the anterior limits of the pharynx unnecessary so far as this type 1s concerned, at least. The anterior limits of the pharyngeal group are definite and do not move forward, while the posterior limit of the oral group moves steadily backward (see Table IV) until it reaches the anterior limit of the pharyngeal group; and in the absence of definite knowledge of the limits of the pharynx it is much more probable that oral buds, as will be seen later, may spread into endodermic territory than that the reverse process takes place. Except for the spreading of the pharyngeal group back into the cesophagus, it is much more constant in its position than the oral and cutaneous groups, which in addition to spreading back in the mouth also spread out from the lips and finally cover nearly the whole surface of the body. These buds of the oral and cutaneous groups which are inner- vated by the gustatory fibers from the geniculate ganglion show a much greater adaptability to the functional needs of the organ- ism than those of the pharynx which are innervated by the ninth and tenth nerves. These three groups seem to represent three more or less well defined functional groups of which the oral and cutaneous are much more closely related to each other structurally and func- tionally than are either of them to the pharyngeal; and it is in the LanpacreE, Taste Buds of Ameiurus. 17 functional needs of the organism, 7. ¢., in the appearance of the buds simultaneously on areas where they could be stimulated at the same time, that we shall have to look for an explanation of the place of appearance and manner of spreading of the various groups of buds. TABLE IV. Table showing the extent of the oral, cutaneous and pharyngeal groups given in sections of 7 microns each. ORAL. PHARYNGEAL. | CUTANEOUS. Empryo, Limits 1N_ | Lencruin | Limits IN | LENGTH IN | Limits in.) Lenoru 1N SEc. | SEC. SEG | Src | SEc. SEc. K’ 10-15 | 5 68-73 | 5 | — “= L 3-21 | 18 116-141 25 = = M 3-41 38 1OS$-121 16 = == N 3-65 | 62 go-165 73 2-36 34 O 2-73 71 go-167 77 8-59 51 O’ 2-67 65 67-173 106 5-38 33 1p 2-76 74 88-180 g2 2-45 | 43 O 2-70 68 75-192 117 2-60 58 R 2-122 120 86-207 121 2-44 | 42 Ss 2-153 131 93-227 134 2-64 | 62 ae 2-154 | 132 96-248 152 | 2-67 | 65 U 2-174 172 93-301 208 | 2-68* | 66 *The opercular buds catalogued in Table VIII, C, for the 9.4 mm. embryo (U) are not included in this summary. The relations of these groups to each other and particularly their limitations are shown in [Table IV, where the anterior and posterior boundaries of the groups are given in sections of 7 mi- crons each. From this table it will be seen that the isolation of the oral and pharyngeal groups is quite marked in the earlier series and even up to series R they do not overlap, but in this series the two groups overlap. ‘Those of the oral group occupy the roof and dorso-lateral portion of the palate and extend pos- teriorly beyond the anterior end of the pharyngeal buds which occupy the floor and sides of the pharynx. In the regions in which the overlapping occurs the oral group includes in this table all buds of the anterior and posterior pala- tine groups, the latter including the proximal hyoid and suspen- sorium buds. All these groups have a certain degree of contin- 18 Fournal of Comparative Neurology and Psychology. uity as they spread backward and all are innervated by the seventh nerve in common with buds in the anterior oral regions. The pharyngeal group includes, in the region of the overlap, buds on the floor of the pharynx, on the distal portion of the hyoid and on the gills. “These are all innervated by fibers from the ninth and tenth nerves, and, with the exception of the distal portion of the hyoid, appear in segmental order from the first gill back- ward. The oral group begins at the extreme anterior end of the mouth and in series U extends back to section 174, the most posterior buds of the oral group at this time being situated on the proxi- mal portion of the hyoid. The pharyngeal group begins at section 68 in K’ and appears in 116 in L and its anterior buds then remain constantly near the 89th section in the remaining series. The most striking excep- tion is in O’, in which the group extends forward to section 67. The most anterior buds of the pharyngeal group as catalogued here always lie in the median ventral line in front of the union of the hyoid with the copula and the overlapping described above is due to the growth of the oral group backward over the pharyn- geal and not to the growth forward of the pharyngeal, since the anterior pharyngeal buds lie grouped about the 8gth section and the oral move back from the 15th to the 174th section. In series S, T and U the cesophageal buds are included with the pharyngeal. If we subtract 3 sections from S, ten from [ and 26 from U, it will give the exact limit of the pharyngeal group. This makes the total in the three groups, respectively, 131, 142 and 182 sections. The preponderance of taste buds in the pharynx where there are 477 in U as compared with 188 in the oral group is not accom- panied with a corresponding elongation of the pharyngeal areas as compared with the oral areas in the early stages of Ameiurus. Even in U, where the areas occupied by oral and phary ngeal buds are nearly equal, being 172 and 182, respectively, the number of buds in the pharynx is more than two and a half times as great as that in the oral group. A comparison of the posterior limit of the oral and cutaneous groups shows a much slower rate of progression in the cutaneous group. In series N, for instance, where the cutaneous buds first appear, the posterior limit of the oral group 1s section 65, while LanpacreE, Taste Buds of Ameturus. 19 the posterior limit in U is 174. In the cutaneous group the cor- responding limits in N and U are 36 and 68. This slower rate of progression, accompanied as it is by the later appearance of the cutaneous buds, probably indicates the phylogenetic relation of these two groups to each other. The oral group would be functionally important regardless of the condition of the eye, while the appearance of the cutaneous buds is probably associated with the reduction of the relative functional importance of the eye in seeking food, which is so characteristic of Ameiurus. The spreading of buds from the oral areas to the cutaneous areas does not involve a very great change, since the two areas are really continuous on the lips and both lie in ectoderm and both are sup- plied by communis fibers from the geniculate ganglion of the seventh nerve. The simultaneous appearance of buds in the oral and pharyngeal areas is, however, much more striking, since one lies in ectoderm and the other in endoderm. siege difference might be minimized by lessening the radical distinction which is usually made between. sale denn and ectodermic areas, since endoderm is always derived from ectoderm, if it were not for the fact that buds lying in ectodermic areas are supplied by com- munis nerves. JOHNSTON’s suggestion that buds spread from endodermic into ectodermic territories, if found to be true in more generalized types, would lessen the dithculty of explaining this curious condition. Since taste buds do not spread from endoderm into ectoderm in Ameiurus, some other explanation of their appearance there must be sought which will explain the conditions here, as well as in other types. In the following sections an analysis of the vari- ous subdivisions of the three principal groups mentioned has been made to ascertain if possible what factor has determined the time and place of appearance of taste buds and in particular to see if the distribution was correlated with or controlled by the nerve supply. In this analysis four things have been kept in mind: (a) Bo determine the time of first appearance of taste buds situated in different areas but innervated by fibers from the same nerve, 7. ¢., to see if buds not parts of the same functional unit but having a common innervation appeared at the same time. (2) To deter- mine if buds that are a part of the same functional unit but have their fibers from different nerves appear at the same time. _ If this 20 Fournal of Comparative Neurology and Psychology. proves to be the case, the functional unit is the prime factor in determining the time of the appearance and direction of spread- ing of groups. Neither of these questions can be answered, of course, except in groups of buds, the smaller subdivisions of which have a single innervation. (3) To determine, if possible, whether buds appear first on the peripheral or proximal distribution of anerve. ‘This ought to throw some light on the manner in which gustatory fibers reach the areas they innervate. (4) To deter- mine if buds appear in any definite order on structures segmentally arranged, such as those in the gill region. The description of the relations of the various areas to the nerve supply is based, as mentioned before, on the account of the communis system of Ameiurus by C. J. HERRICK (or), mainly, although not entirely for the pharyngeal regions. “The reader is referred to that paper for a fuller description of these nerves; only so much of it is incorporated here as is deemed necessary to make the correlation clear. The term “functional unit,” as used in the following descrip- tion, is applied to groups of buds, discontinuous at the time of their appearance, which from their position on the body would seem to function as units, that is, be stimulated in unison. ‘They are characterized, first, by having areas between them and other groups devoid of buds at the time of their appearance, although they may become continuous later with these groups; and, second, by differences in time of appearance, this difference varying from a few hours up to one hundred or more for adjoining groups which may later become continuous. ‘The units are composed of smaller subdivisions closely related to the number of nerves sup- plying the unit. These show the same manner of spreading as the larger unit of which they are part. 4. THE DORSAL LIP AND MAXILLARY BARBLET GROUP. This group includes buds lying on the upper lip both inside the lip and on the outside, buds lying on the maxillary barblet and on the dorsal breathing valve. The first buds to appear in this group are situated on the lateral portion of the upper lips, the dorsal breathing valve and the region of the premaxillary teeth and on the base of the maxillary barblet. Buds appear on the base of the maxillary barblet of K’ (118 hours). As mentioned before, these may be really internal lip buds and not on the outer LanpacreE, Taste Buds of Ameturus. a surface. It is not possible to distinguish between the buds on the lip, on the breathing valve and on the region of the premaxillary teeth. Teeth are not present at this stage and the premaxillary bone, being a membrane bone, is not present and the breathing valve is not yet differentiated from the lip. The breathing valve can, however, be distinguished easily from the roof of the mouth at this stage. The epithelium of the lip and breathing valve is thick, while that of the roof of the mouth is thin. ‘This enables one to separate these two structures from the roof of the mouth, which receives buds much later. All these areas receive buds simultaneously and all are innervated by communis fibers running out through the ramus maxillaris (excepting one twig from ramus mand. V to the barblet), the buds of the maxillary barblet and the skin of the upper lip just in front of the maxillary barblet and con- tinuous with it being supplied with the lateral branch of that nerve and those of the lateral parts of the upper lip and the region of the premaxillary teeth by the mesial stem of the same nerve. The buds of the maxillary barblet are supplied by three twigs from the same nerve. Herrick (01), in his description of the distribu- tion of the ramus maxillaris, does not mention the upper breathing valve and I infer that it is included in the description of the inner- vation of the lateral lip region and the premaxillary region. Table V shows the number of buds and the number of sections occupied in this group for series K’ to P. TABLE V. Table showing time of appearance, rate of increase and length of area measured in sections occupied by the dorsal lip and maxillary barblet groups. Under each age the first column gives the number of buds present, the second column the extent indicated by the section numbers between which they are found. eee KY L 7 M i N i O (Oy x | °°) ees 8] cee, | 8 ea | aces | cae Lip and breathing valve. .| ;: | ee 3 | asl oe 3720 9 | 37-21 a 2-40 38 | a Ila lide coooucea ne 2 26 13 | 36-54] 11 | 43-58) 18 | 40-74] 22 | 21-59] 34 | 31-73 This area, although innervated by several nerves, would un- doubtedly function as a unit. ‘The only exception to this state- ment that might be taken is in regard to the maxillary barblet, but as stated above its first buds appear at the base of the anterior IF Fournal of Comparative Neurology and Psychology. surface where they are continuous with the inner and lateral por- tion of the lip. And when buds appear later distally on the bar- blet, they are accompanied by other buds farther back on the body which would increase the area innervated from the lip region, that is, the long maxillary barblet bearing buds on its distal por- tion is equivalent to enlarging the cutaneous group by spreading over the head and body. This group is isolated in position from buds on the top of the head by an area devoid of buds between it and the nasal group which appears later and from the anterior palatine group by the later appearance of that group as well as by the different histo- logical character of the epithelium on which they are situated. The buds on the maxillary barblet supplied by the ramus mand. V cannot be separated from those supplied by the ramus maxil- laris. here is no reason, however, for supposing that they do not appear along with the other maxillary buds in point of time rather than with buds appearing later in other areas which are innervated by the ramus mand. V. The almost invariable rule is for buds to appear with their functional group regardless of the various nerves supplying them. 5. THE VENTRAL LIP AND BARBLET GROUP. This group includes buds lying inside the mouth on the lower lip and mucosa covering the mandible and on the outside cover- ing the outer anterior surface of the mandible and on the mental and post-mental barblets. All these buds are placed under one head because those lying on the anterior portions of the lips are continuous both with those inside the mouth and with those under the lips on the outside of the body. ‘They probably comprise two groups, however, at least functionally. Those lying inside the mouth would function with those buds occupying a similar position on the upper lip and breathing valve and the region of the premaxillary teeth, while those lying under the lower jaw and on the mental and post-mental barblets would function with the buds situated on the maxillary barblet and those of the nasal group. Structurally this group is isolated, both those lying inside the mouth and those outside, from groups lying farther back by well defined areas devoid of buds at the time the posterior groups appear and by the later appearance of those groups. Lanpacre, Taste Buds of Ameturus. 23 (a) Buds of the lower lip inside the mouth on the lower breath- ing valve and on the mucosa of the mandible. In separating the buds of the extreme anterior portions of the lip into those inside the mouth and those outside the mouth the same arbitrary rule was followed as in the case of the upper lip. Buds lying below the ventral border of the mucosa of the lip were tabulated as on the outside of the body. It is not possible to separate buds of this subdivision which lie on the lip from those on the mucosa of the mandible. Meckel’s cartilage extends almost into the extreme anterior portion of the lip and there is no perceptible difference between the epithelium of the lip and the breathing valve and that of the mandible. Table VI shows the number of buds in this subdivision up to series O’, where there are 24. The backward movement of the buds up to this time is not marked, although present; in later stages buds extend much farther posteriorly and in the 20} mm. embryo they occupy four-fifths of the total extent of Meckel’s cartilage. TABLE VI. Table showing time of appearance and extent of subdivisions A and B in the lower lip and the two ventral barblet groups. A. Mucosa. B. Sxin. lc. Tee eae D. ee eee SeeEEe No. oF | re No.or| Ex- | No. or | Ex | No. or | Bups,| tent. | Bups, | TENT. Bove Tent, | Bups. | EXTENT. L I 20 I 27 — == = — M 7 30-41 == ff = — N 201 927-05 2 27 7 ABeS5 | a7 | 63-75 O 21 17-70 8 17-31 10 36-48 Io | = 43758 Of ay || ae 66 6 19-30 9 47-61 12 SS = OF ‘The taste buds on fie mucosa of the ae are innervated by twigs from the internal branch of the ramus mand. V. Those of the middle of the lower lip and lower breathing valve are inner- vated by the lip twig of the same nerve. ‘This nerve also gives off two twigs farther back for the mental and post-mental bar- blets which acquire buds somewhat later. Buds on the edge (“lateral portion”) of the lower lip are inner- vated by the external branch of the ramus mand. V along with those throughout the whole length of the outer surface of the 24 Fournal of Comparative Neurology and Psychology. mandible. ‘The buds of the middle of the lip cannot be separated from buds of the lateral portion of the lip even in the earlier stages, although innervated separately by the internal and exter- nal branches, respectively, of the ramus mand. V. They would undoubtedly function as a unit. If we compare the posterior limit of the dorsal lip and breathing valve group (Table V) with the anterior limit of the lower lip and mandible group (Table VI), it will be seen that there is no over- lap. ‘The posterior limit of the former group for series K’, L, M and N are sections 15, 16, 20 and 21, respectively, and the anterior limit of the latter group for series L, M and N are 20, 30 and 27, respectively. he ventral group underlies the dorsal in later series, however, and its failure to do so in,the earlier series is due probably to the elongation of the upper parts of the head on account of the rapid growth of the brain. ‘This is equalized later by the growth of the lower jaw to fit into the upper. The posterior limit of buds on the mucosa of the mandible moves backward, as stated above, until in series N it comes to the region occupied by the most anterior buds of the anterior pal- atine group, and in series O’ reaches the anterior buds of the mid-ventral pharyngeal group which, however, had appeared first in an earlier series (N). The spreading of the buds back from the anterior portion of the mandible both on the mucosa and on the skin, as will be shown presently, toward the posterior or proximal portion and the man- ner in which the nerve runs along the mandible from the proxi- mal toward the distal portion is conclusive evidence that in these two cases the more peripheral buds supplied by these nerves receive fibers before the proximal buds do. The functional need of the organism, represented by the more or less continuous spreading of the buds from the lips out over the surface of the body and back in the oral cavities, seems to furnish the key to the explanation of the methods of spreading rather than the anatomical arrangement of the nerves. Buds rarely or never in the oral and cutaneous groups appear on the shortest twigs or the proximal distribution of a nerve first, but always on the distal or longest portions, since this maintains the continuity in the anterior-posterior method of spreading. (b) Buds lying on the skin of the mandible outside the mouth, and on the mental and post-mental barbules (groups B, C and D). Lanpacre, Taste Buds of Ameiurus. 25 From Table VI it will be seen that buds appear first on the skin of the mandible (group B) in series L where there is one bud. None are present in M and two in N. In series N buds appear on the mental and post-mental barbules, but there are a few sec- tions in O’ and P lying between the buds on the mental barbules and those of the lower lip devoid of buds. Notwithstanding this slight discontinuity, there seems to be no doubt of the pro- priety of including the barblet buds with the anterior mandibular or lip group, since buds situated farther back on the lower jaw do not appear until series U, when buds appear on the posterior portions of the operculum. Buds which would be most likely to be confused with these in later stages are those of the posterior mandibular divisions of the post-orbital and opercular group, which do not appear until the It mm. embryo, so that the group 1s quite homogeneous in distribution and time of appear- ance compared with any other group with which it might be con- fused. Functionally this group probably serves the same pur- pose for the lower jaw that the buds of the nasal group do for the dorso-lateral portion of the head. Buds of the lower Jaw, for in- stance, have extended as far back as section 30 in O’, while those of the nasal group have extended to the 38th section. It is not possible to make a comparison of the mental and post- mental barbules with the maxillary and nasal on account of the pliable character of the structures. The posterior limit of the anterior mandibular group moves posteriorly slightly in the earlier series until it reaches a point just back of the post-mental barblet, but in series U the pos- terior buds of this group lie in section 53, so that there is prob- ably up to this time little actual spreading. ‘This must occur in later series, however, since according to Herrick the buds of this group supplied by the ramus mand. V occupy the whole extent of the mandible. The innervation of buds on the lateral portion of the lip and mandible is by the external branch of ramus mand. V, while that of the mental and post-mental barblets is from the internal branch of that nerve along with the middle of the lower lip, the lower breathing valve and the mucosa of the mandible. It is of interest to note in this group, first, the slight discon- tinuity in the appearance of mental and post-mental barbule buds, since they represent the posterior extension of the lower 26 Fournal of Comparative Neurology and Psychology. lip group and have their innervation in common with that, and, second, the fact that the buds on the mandible supplied by the ramus mand. ex. V appear {first on the anterior portions of the mandible; that is, on the distal distribution of the nerves, and later spread back to the posterior portion of the mandible at a time corresponding to the appearance of buds in post-orbital and opercular group. 6. THE NASAL GROUP. A second group of taste buds isolated by position and by time of appearance includes those lying in front of the nasal barblets, on the skin at the base of the nasal barblet, about the nasal opening and those between the nasal barblet and the eye. All these buds are innervated by the ramus oph. sup. V (WorK- MAN, 00). his group can be broken down into four, or possibly five, subdivisions separated by differences in position of the first buds appearing and by a very slight difference in time of appearance, although the group as a whole is quite homoge- neous in both ties respects. This group lies in the region where the anterior and posterior nasal openings are formed, and it is possible that the subdivisions in which this group appears are due in some measure to the presence of this organ, although at least three of them are indicated by divisions of the nerves supplying them. ‘The nasal opening is an antero-posterior slit still open through- out its whole length in series L, but closed in series M. ‘The ae mation of two nasal openings from a single slit-like opening 1s accomplished by the approximation of the dorsal and ventral lips of the slit in the middle region, leaving an opening at either end around which the nasal cones appear later. ‘The nasal barblet occupies a position slightly posterior to the middle of the dorsal lip in the embryo but comes to have a more posterior position in the adult quite near the posterior opening. The subdivisions of the group are as follows: (A) A group in front of the nasal barblet,on the top of the head, and extending forward toward the snout and innervated by one of the last four large divisions of the ramus oph. sup. V. (B) A group lying on the nasal barblet and innervated by the remaining three of oe last four divisions of the ramus oph. sup. V. Lanpacre, Taste Buds of Ameiurus pA (C) A group in front of the eye and back of the nasal barblet and over the supra-orbital line and innervated by the first four main branches of the ramus oph. sup. V. (D) A group lying under the line of closure of the nasal slit whose _ innervation Rohe nie does not give, or at least does not distin- guish from the other buds about the nasal opening. (EZ) A group in front of the eye and under the supra-orbital line and probably innervated by some of the fibers supplying group C. TABLE VII. Table showing the number of buds at different ages in the various subdivisions of the nasal group and their extent as indicated by the section numbers between which buds are found (sections 7 microns thick). rae Aan: Cc i. Wiig ee obeeee E EMERYO:);a can) | | ae aes | ERNE FS?) ee | No. Extent. | No. | No. | Extent. | No. | Exteni. | No. | Exrenr N =e = I I 33 I 36 = = O = = 3 2 32-40 Dy SAO meena || = (OY 2 19-23 2 2 30-38 I 33 Vn ay P 2 15-22 4 2 30-40 | 2 42-45 | — | = Q 3 19-23 5 2 35-43 2 49-60) 6): ae re R 3 16-22 5 2 31-43 I 44 — = s 5 17-32 5 2 33-60 4 hacp | eee 64 AR 5 15-21 6 2 41-50 5 S5S50 i Sele 67 v 9 19-34 25 2 73-89 9 _48- ~68 | IOI The aie columns of Table VII show ae Bat ee of Bae Aa the double columns give the extent of groups expressed in sections. It will be seen from the table that three of these subdivisions appear in N and one additional in O’. The presence of two buds on the skin in front of the nasal barblet, group A, indicates that possibly the later appearance of this sidliieioant is an individual variation. It is possible that series E which lies in front of the eye and below the supra-orbital line belongs to D, though I have catalogued it separately, and that the sections in which these buds are located indicate the posterior extent of that group. The first buds in this group appear 25 hours later than the first buds of the preceding group and are isolated from them in position. All the buds of this group lie about the nasal opening and on the dorso-lateral portions of the head, while those of the preceding eroup are on the lip and maxillary barblet. ‘The posterior buds of the lip and barblet group lie farther from the anterior end of the 28 ‘fournal of Comparative Neurology and Psychology. head than the anterior buds of this group in the later series owing to the more rapid growth posteriorly of the lip and barblet group, but this is not true at the time of the appearance of the nasal group. While the nasal group lies dorsal to the maxillary group, it also rep- resents the posterior spreading of the lip group since it les farther back on the longitudinal axis of the body at the time of its appear- ance. Later, however, the nasal group probably represents func- tionally the extension of the lip group to the dorsal portion of the head. Subdivision C of this group is very peculiar in that it contains but two buds up to series U. ‘The position of these buds through- out the earlier series 1s so nearly constant and their separation from each other remains so nearly the same that there can be little doubt that they are the same buds in all series. “The constancy of these two buds throughout so many series is still more striking when we consider the fact the area between the nasal barblets and the mid- dorsal portion of the eye of the adult is innervated according to Workman (’00) by communis fibers running out through four well defined divisions of the ramus oph. sup. V and by several small twigs. Comparing group C with group A, we find practically the same stationary condition in that group, so that we shall have to consider the areas on which these buds are situated as growing slowly or else the spreading of the buds 1s practically wanting and does not correspond closely with the number of nerve twigs supply- ing these two subdivisions. The gustatory fibers evidently do not run to the surface simultaneously through nerve twigs which inner- vate practically the same area. The whole group is characterized during the earlier stages of Ameiurus by its practically stationary condition. ‘The back- ward movement in all the series except U can be accounted for mainly by the growth of the embryo. Group C, while stationary in the earlier series, in the 11 mm. embryo spreads back as far as the middle of the dorsal portion of the eye and it was puzzling to know whether to include some of these buds lying over the eye in the nasal group or in a division of the post-orbital group lying behind and above the eye and innervated by the ramus max. acces. However, that group appears along with the remainder of the post-orbital and opercu- lar group; and, while the posterior extension of the nasal group comes close to the anterior buds of the post-orbital group, still LanpacreE, Taste Buds of Ameiurus. 29 there is an area devoid of buds between the two groups. This is the only case in which two groups which later become con- tinuous are at all difficult to separate at the time of their first appearance. 7. THE POST-ORBITAL AND OPERCULAR GROUP. / A fourth group comprises buds lying back of the eye mainly, and in front of the posterior edge of the operculum. ‘The various divisions of this group arise almost simultaneously, but receive their nerves from a variety of sources. ‘lwo other groups, the cerebellar and the occipital groups, belong here as far as their position on the longitudinal axis of the body is concerned; but since they appear much later and have in addition their nerves in common with the body buds, they will be described with those of the body. The post-orbital group can be divided into several smaller subdivisions as follows: (A) Buds lying behind, under and above the eye and inner- vated by fibers from the ramus max. acces. and the ramus mand. ex. VII. (B) Buds lying on the mandible and distributed along the mandibular lateral line canal from lateral line organs three to eight and innervated by the ramus mand. ex. VII. (C) Buds lying on the operculum, those inthe region of the pre-operculum bone being innervated by the ramus “mand. ex. VII, while those lying on the posterior and dorsal portions of the operculum in the region of the post-frontal and squamosal are innervated by the ramus oticus. (D) Buds lying on the branchiostegal rays and innervated by the ramus hyoideus. TABLE VIII. Table showing the number and time of appearance of buds in the post-orbital group. A B | Cc D EmMBryYo. Post-MAND. | BRrRANCHIOSTEGAL Optic Group. Oper. Group. GROUP. | Group. 9.4 mm — — II | oo 11.0 mm. 6 8 19 | I 30 Fournal of Comparative Neurology and gia Subdivision A does not appear until in the 11 mm. embryo, where there are six buds present, one under the eye, four on the dorsal and posterior portion of the cornea and one behind the eye. Of the ten buds present in series A in the 142 mm. embryo, eight lie on the upper and posterior portion of the eye and two beta the eye, none being present under the eye. This subdivision is distinctly isolated in position from other members of the group to which it belongs. It is, however, not so sharply separated from the posterior buds of fhe nasal group as adjoining groups are usually separated. Even in this case, however, there is an area of 20 sections between the anterior buds of group A and most posterior buds of the nasal group. In addition, the posterior buds of the nasal groups are above the anterior half of the eye and the first buds of the orbital sub- division appear under and behind the eye. The first buds to appear in this subdivision are situated on the border of the cornea, some of them actually lving on the transparent portions of that structure. Without raising any question as to the relation of the acces- sory maxillary -nerve to the ramus maxillaris, it may be of interest to note that the taste buds innervated bythe ramus max. lie on the lateral portions of the upper lip, on the maxillary barblets, and on the skin of the upper lip, just anterior to the maxillary barblet at the extreme anterior end of the embryo and appear more than 130 hours sooner than those lying about the eye and supplied by the ramus max. acces. The second subdivision B appears first in the 11 mm. em- bryo, where there are at least eight buds. It is not present in series U (244 hours). “The whole mandible, as mentioned above in the adult (HERRICK, ’or), is supplied with two nerves; first, in common with the lower lip and mental and _ post- ienel bar- bules, by the ramus mand. V, external branch, and secondly, by the ramus mand. VII, a branch of the ramus hyomandibularis. This last nerve supplies the mandible from the region of the third of the main lateral line organs back to the region of the seventh, or possibly farther. In the earlier stages, however, the groups supplied by these two nerves are quite distinct. “The anterior mandibular group spreads back from the lips and men- tal and post-mental barbules and appears much earlier, while the posterior group comes in with the remaining divisions of LanpacreE, Taste Buds of Ameiurus. 31 the post-orbital group in point of time. Even in the 11 mm. embryo there is an area of 30 sections between them quite free from buds. The posterior mandibular group buds are distributed as fol- lows with reference to the mandibular lateral line organs. In the 11 mm. embryo, four lie between the third and fourth organ, and four between the fourth and fifth, while in the 142 mm. em- bryo, where there are 11 buds, six lie between the third and fourth organ, and five between the fourth and fifth. In the 174 mm. embryo, not shown in the table, there are 13 buds, six lying between the third and fourth organs, five lying between the fourth and fifth organs, and two between the fifth and sixth organs; compared with the lateral line organs, we have an evident spreading from anterior to posterior. It is difficult, in later stages, to separate the posterior mandi- bular buds from the opercular groups, but since the mandible does not extend farther back than section 131 in the 9.4 mm. embryo andthe first buds in the opercular group appear in section 166, which is between the seventh and eighth mandibular lateral line organs, it 1s altogether probable that ae separation of these two groups is represented by this difference. In the 11 mm. embryo the mandible does not extend beyond section 184, and the last bud on the mandible lies in section 129 and the first of the opercular group in 160. Herrick (or) describes the recurrent twig of the mandib. ex. V as supplying the skin over the dentary, articular and quadrate bones. It is possible that the first two buds of the opercular group, which lie in sections 160 and 163 behind the end of the mandible, and 30 sections in front of the remainder of the opercular group, may belong to the quadrate group, although he does not say specifically t that the recurrent nerve supplies taste buds. The posterior mandibular division at the time of its appear- ance is quite distinct from the anterior mandibular division anteriorly and from the posterior opercular division posteriorly.” Its innervation is undoubtedly at this time solely from the ramus mand. ex. VII. It also shows, as mentioned above, a well de- fined progression from anterior to posterior, spreading from the second and third mandibular lateral line organs to the region between the third and fourth and later to Site between the 32 Fournal of Comparative Neurology and Psychology. fourth and fifth. In later stages it 1s joined and probably over- lapped by the posterior spreading of the anterior mandibular group. The third group (C) includes buds lying on the operculum posterior to the preceding group. ‘The buds lying in the region of the pre- operculum are innervated by the fibers arising from the main stem of the ramus hyomand. just before it divides into the ramus mand. ex. VII and ramus hyoideus. The buds lying on the dorsal and posterior portions of the operculum near the post-frontal and squamosal are innervated by the ramus oticus. This group appears somewhat earlier than those of the two pre- ceding subdivisions, there being 11 buds in series U, 9.4 mm. ‘Two of them are situated on the dorsal posterior portion of the operculum between the first and second organs of the main lateral line and nine along the area of the mandibular lateral line canal. Of these nine one is between the seventh and eighth mandibular lateral line organ and eight are between the last mandibular lateral line organ and the first main lateral line organ and all lie below the line of the canal. Of the 19 buds lying on the opercle in the 11 mm. embryo, two lie between the fifth and sixth mandibular lateral line organs. ‘These two buds have been incorporated here but they may belong to the quadrate group mentioned above. Of the remaining 17, five lie between the sixth and seventh mandibu- lar lateral line organs, and six lie between the seventh and eighth, while two lie between the eighth mandibular and the first main lateral line organs and four lie between the first and second of the organs of the main lateral line. Of the six buds appearing in the opercular group in the 143 mm. embryo, four lie between the seventh and eighth mandibular organs and two between the last mandibular organ and the first main lateral line organ. There seem to be no buds on the posterior dorsal portion corre- sponding to those between the first and second main organs in the 11mm. embryo. ‘The deficiency of buds in this group in the 14? mm. embryo is quite striking. In Herrick’s Ameiurus paper (’01) the ramus oticus is de- scribed as innervating the dorsal portions of the operculum and the post-frontal and squamosal bones. ‘There seems to be no way to separate this area from that of the buds appearing on the pre-operculum, since membrane bones are not definitely Lanpacre, Taste Buds of Ameiurus. 23 formed enough to ascertain their exact boundaries and the mandibular lateral line curves dorsally in the posterior portion of the operculum and buds distributed along this line have the same general direction in spreading, so that it is not possible to give these boundaries separately. However, in the 11 mm. em- bryo three buds are found on what | take to be the post-frontal and squamosal regions, and in the 142 mm. embryo only one bud is found in this location. A fourth division (D) of this group comprises buds lying on the branchiostegal membrane. ‘There is only one bud in this position in the 11 and 14% mm. embryos. ‘This group is inner- vated, as mentioned before, by ramus hyoideus. The post-orbital and opercular group as a whole is isolated structurally, as mentioned, by well defined areas devoid of buds lying between it and the groups situated anterior to it. _ Its isola- tion in time of appearance 1s still more marked. Group (A) which approaches in position most nearly the nasal group can hardly be considered at the time of its appearance as continuous with that group and the remaining subdivisions are still further isolated from preceding groups. As to the homogeneity of this group functionally, in the absence of experimental evidence and apparently of the possibility of obtaining such evidence, we must fall back upon the isolation in position and in point of time as our only evidence at present. Its innervation by communis fibers running out through five different nerves, while not different in principle from the ante- rior groups (notably the dorsal and ventral lip groups), still shows the greatest diversity of any of the oral or cutaneous eroups. One interesting fact in connection with this group is that it occupies practically its whole territory at the time of its appear- ance and of course little evidence of its spreading backward is given, although in both the mandibular and ventral opercular portions there is evidence that the group spreads posteriorly. 34 Journal of Comparative Neurology and Psychology. 8. THE CEREBELLAR, OCCIPITAL AND BODY GROUPS. TABLE IX. Table showing the time of appearance and number of buds in the cerebellar, occipital and body groups. Empryo. A. CEREBELLAR. B. Occrrira. C. Bopy. U 9.4 mm. — | Fe II mm. — | -— 2 143 mm. = = 174 mm. — = 4 20+ 5 10 14 This group of buds includes (A) those lying on the cerebellum and innervated by the meningeal twig of the ramus. lat. acces. (B), those lying over the occipital region and innervated by the twigs of the ramus lat. acces. given off before that nerve leaves the skull; and (C) buds on the body innervated by the ramus lat. acces. here is no overlapping of this eroup (C) with the pre- ceding group, the body buds being entirely separate from buds of the posterior operculum until in the 20} mm. embryo, when they become continuous. It will be noticed from the table that the cerebellar and occi- pital groups appear simultaneously in the 20} mm. embryo. It is not possible to separate these two groups at this stage, the cerebellar being continuous with the occipital. Doubtless, if a series had been cut between the 174 mm. embryo, where they are not present, and between the 20} mm. embryo, they would be found to be distinct groups. ‘The separation has been made in cataloguing, by including all buds in front of the posterior portions of the cerebellum in the cerebellar group, and all those back of that point in the occipital group. The failure of the cerebellar and occipital groups to appear simultaneously with the opercular or post-orbital group, with which they are related in position, is difficult to explain. They have the same innervation practically in Ameiurus that the body buds have and it was thought at first that possibly they followed the same rule that is so often illustrated in the groups anterior to this, of the appearance of buds on the peripheral distribution of the nerves sooner than on the proximal. However, buds on the body appear first on the anterior segments and spread from there posteriorly, and the same occurs, as we shall see later, in the case of the cesophageal buds. Lanpacre, Taste Buds of Ameturus. 35 It is possible that the irregularity of these two groups can be explained by the fact that they lie in close proximity to the post- orbital group which is large and appears much earlier. The functional needs of this portion of the body being thus well sup- plied, these two groups are not of so much importance as the body buds which occupy the regions back of the operculum alone. However, the innervation of He region under discussion is com- plicated and very difficult to work out, according to Professor HER- RICK, so that there may be some facts about the innervation which would clear up the difficulty, and there is a slight possibility that we may have to deal with nothing more than an individual varia- tion, since only one series of each embryo has been catalogued. The limits of the body group were determined with reference to the lateral line organs and to the free posterior border of the operculum; that is, the point at which it was detached from the body dorsally in the section. TABLE X. Table showing the number and position of the buds on the body with reference to the organs of the main lateral line of the body, the free posterior border of the operculum and the distance in sections from the posterior buds of the opercular group. | No. or SEC. FROM OrercutuM DeracHED First Bopy Bups No. oF Empryo. Post, OPERCULAR B AT SECTION, APPEAR AT SECTION, WDSc | Bups. ie pee TE NS : II mm. 318 between 1.1. org. 3 and4 | 412 between I.1. org. 3 and 4 68 2 14 ?mm | 412 between 1.1. org. 3 and 4 | 427 between 1.1 org. 3 and 4 129 2 17 +mm | 422 between ll. org. 2 and 3 | 461 between 1.1. org. 3 and 4 552 4 20.4mm | 612 between |.1. org.3 and 4 | 591 between 1.l. org. 3 and 4 3 14 | The first body buds lie in each series between the third and fourth main lateral line organs. ‘The lateral line organs seem to be constant in position Stak reference to each other and to the posterior free border of the operculum, so that in locating these buds by the organs we avoid difficulties arising from the irregular growth of the head as compared with the rest of the body which might arise if they were located by sections alone. Of the four buds found in the 17} mm. embryo, two are found between lateral line organs three and four, as are the two found in each of the earlier series of the 11 mm. and 14? mm. and the other two are found behind the fourth lateral line organ. 36 Fournal of Comparative Neurology and Psychology. Fic. 1. Series L, 120 hrs. D, dorsal lip and max. barbule group. JV, ventral lip group which at this stage contains only one bud that could be considered as on the outer surface. Fic. 2. Series O, 146 hrs. V, ventral lip buds. M.B, mental and post mental buds. D’, dorsal lip and maxillary barblet group. B, C, D, subdivisions of nasal group. Subdivision A is not repre- sented on this chart. It is so closely related to the anterior nasal opening that it is not easy to separate it from buds lying on the nasal barblet. Fic. 3. SeriesU,244hrs. V.L., Ventrallip. Ba, mental and post mental barblet buds. D. L., Dorsal lip and maxillary barblet group. A, B, C, D, E, subdivision of nasal group. C’and C”, ventral and dorsal subdivisions of the opercular group. Subdivisions A and B are not present in this stage. Fic. 4. 20% mm. stage. At this stage the various subdivisions of the nasal group are continuous with each other and reach back to the eye. The dorsal lip and maxillary barblet group is continuous dorsally with the nasal group and the ventral lip and barblet group is continuous posteriorly with the posterior mand. subdivision (B) of the opercular group and dorsally behind the angle of the mouth with the dorsal lip group. None of these buds are shown on the chart except (B). B, the posterior mand. division of the opercular group. C and C’, the ventral and dorsal portions of the opercular division. D, the branchiostegal buds. Subdivision A, whichlies behind and under the eye,is omitted. A’, B’, the cerebellar and occipital buds. C”, the body buds. Lanvacre, Taste Buds of Ameturus. 37 38 Fournal of Comparative Neurology and Psychology. Of the 14 buds found in the 20} mm. embryo, one is found between the third and fourth main lateral line organ, one on the skin over the fourth organ, eight are found between the fourth and fifth organ, and four between the fifth and sixth organ. This body group shows a well defined progression posteriorly as measured by the lateral line organs, moving back from the area between the third and fourth to that between the fourth and fifth and later to that between the fifth and sixth. The nerves supplying these buds run out from the ramus lat. acces. through spinal rami, which are, of course, arranged segmentally, but no further evidence of a segmental arrangement can be detected than that noted in the table, and J am inclined, of course, from a study of the anterior oral and cutaneous buds to con- sider the segmental arrangement of buds here to be purely sec- ondary; that is, they are segmental simply because they are innervated by nerves that are segmental on account of having used segmental spinal nerves as a means of reaching the surface. The question as to whether these buds spread out from the gills or spread back from the post-orbital group will be only mentioned here, and will be discussed more fully with the pha- ryngeal buds. Their continuity in time of appearance is cer- tainly with the post-orbital group, and their innervation is from the seventh nerve, along with the post-orbital group and buds lying farther forward about the anterior portions of the head. The only two ways in which continuity in distribution could be established with the pharyngeal group, would be for the pharyn- geal buds to spread out from the under side of the operculum to the body or else from the bases of the gills down to the ven- tral surface of the isthmus and thus to the body. It will be seen later that neither of these movements takes place. We must con- clude from the time of the appearance, the continuity in distri- bution and the innervation that they are related to other buds on the head and operculum and are part of the general spread- ing of buds from the anterior to the posterior portions of the body eee They represent the posterior extension of a function- ally more or less homogeneous group of buds enabling Ameiurus to ascertain the location of sapid substances when the stimulus reaches portions of the body other than the mouth and the pharynx. The spreading of buds from the place of appearance on the LanpacreE, Taste Buds of Ameiurus. 39 body between the third and fourth lateral line organs in a pos- terior direction can only be interpreted as a spreading from the proximal toward the distal distribution of the nerves, since the nerves arise anterior to the areas innervated. ‘This it will be recalled is in sharp contrast with the condition usually found in the areas lying anterior to the origin of the nerves. The location of the cutaneous groups is shown in text figures I to 4, p. 37. Q- THE ANTERIOR PALATINE GROUP. This group consists of buds lying on the anterior roof of the mouth. It has quite definite boundaries structurally and_ its mucous membrane is quite sharply differentiated anteriorly from that of the lip and breathing valve. ‘The lip and breath- ing valve have a thick mucous membrane, while that of the palate is quite thin as late as series U (244 hours), so that the separation of these groups is quite easy, without considering the area devoid of buds between them, and the difference in time of appearance. In the 20} mm. embryo, however, the mucous membrane of the anterior roof of the mouth is quite as thick as that of the breathing valve, but usually stains lighter. Even if this histological difference were not present, it would be quite easy to determine the anterior boundaries of this group, since the breathing valve always has a free border posteriorly and medially and there is, as mentioned above, no continuity in the distribution of buds from the valve to the palate. This group is supplied exclusively by the ramus palatinus VII, while the following group is supplied by the ramus pala- tinus posterior. ‘The posterior boundaries of the anterior palatine group are not so definite as the anterior boundaries but there is always in the earlier series and even up to series U (244 hours), a well defined area devoid of buds between the anterior and pos- terior palatine groups. This group appears first in series N (138 hours), when four buds are found arranged symmetrically on either side of the me- dian line. In fact, this whole group up to the time when there are 20 buds present in U is entirely symmetrical, the same num- ber of buds lying on either side of the median line. Up to series R (183 hours), there are only four buds. From this time on 40 Fournal of Comparative Neurology and Psychology. they increase, as shown in Table XI, and are very numerous in the 20} mm. embryo. TABLE XI. Table showing the number of taste buds in the anterior palatine group and the limits of their distribution expressed in sections. Empryo. No. or Bups. ExtTENT 1N SECTIONS. N 138 hrs. 4 55-65 O 146 “ 4 43-73 Oe aise 4 56-67 E163) 5 4 58-76 ie (ee 4 . 56-70 IRE riba = 4 | 46-66 S199) < 8 56-98 23a 16 55-104 Ur2g.5 6 20 57-110 It will be seen from the table that the areas occupied as rep- resented by sections are quite constant, ranging from sections 43 to 58 for the anterior boundary and from 65 to 110 for the posterior boundary. After the number of buds increases be- yond four, the posterior limit of this group moves back slightly, but up to this time there is undoubtedly little or no spreading in this group. ‘Fhe posterior limit of this group in U brings it into contact with the anterior buds of the posterior palatine group. I0. THE POSTERIOR PALATINE GROUP. This group consists of (A) buds lying on the posterior wide palate, (B) on the proximal parts-of the hyoid arch and (C) on the suspensorium, 1. ¢., the hyomandibular, quadrate and symplectic. This group, like the preceding, is innervated by a single nerve, the ramus palatinus posterior VII. This group is isolated when it first appears from other groups on the roof of the mouth in position, there being 40 sections between the posterior limits of the anterior palatine group and the anterior buds of this group and about ten sections between the posterior buds of this group and the buds on the roof of the pharynx in the region of the first gill arch. Lanpacre, Taste Buds of Ameturus. 4 TABLE XII. Table showing time of appearance of buds and extent occupied in the posterior palatine group. A. PALaTte. B. Hyorp. c. Se UeEes oe: EMBRYO nant No. or Bups. |Extent.| No: or Buns. | Extent. No. or Buns. | Extent. —_ _-- — _ — — oom —|— ~ — {= R 4 108-122 I 105 = = S 4 123-140 2 | 129-153 I 144 a 7 104-150 6 | 132-154] I 157 U 27 110-170 10 | 152-174 6 180-193 (A) From this table it will be noticed that the buds lying in the roof of the wide palate posterior to those innervated by the ramus pal. VII are found first in R at section 108. ‘The poste- rior boundary of this group moves back from section 122 in R to 170 in U, which brings it into contact with the roof buds lying over the first gill arch. While the anterior boundary of this group varies slightly, still the main movement of the group 1s backward from the point of origin, 7. e., from the position of the peripheral buds innervated toward the more proximal buds. (B) The buds on the proximal portion of the hyoid arch appear first in series R and spread from the ventralmost point of appearance back up toward the proximal. I have taken as the ventral limit of the hyoid group receiving the innervation mentioned, the point where the hyoid cartilage becomes i incorpor- ated with the operculum, as distinguished from that portion lying in the isthmus and constituting the distal or ventral por- tion and receiving its nerve supply from the post-trematic divi- sion of the ninth. I have grouped the buds of the proximal hyoid supplied by the ramus pal. posterior VII with the oral group in order to show the relation of that group to those buds and because it has the same innervation; but it is altogether probable that the pos- terior palatine group, the hyoid and suspensorial groups with those of the distal hyoid and the floor of the pharynx form a functionally complete group, since they occupy the same area in the longitudinal axis of the embryo. As mentioned above, the posterior palatine group undoubtedly spreads from the distal toward the proximal areas innervated, but in the case of the hyoid and suspensorial groups the evidence is not so clear. In the hyoid group the first bud 1s found in R at 42 ‘fournal of Comparative Neurology and Psychology. section 105 and the last in U at section 174, but the anterior limit of the group also moves back from 105 in R to 152 in U; and in the case of the suspensorial buds the same thing occurs. ‘The one bud found in S lies in section 144, while the first bud in U lies in section 180. The only inference from this, barring the loss of buds, which is altogether improbable, is that these struc- tures are moving bodily away from the anterior end of the em- bryo. .In the roof buds, as mentioned above, the anterior limit is not so constant as in the anterior palatine group; still the first buds in R and the first in U are within two sections of each other while the posterior limit moves backward much more rapidly than the relative increase in length of the embryo, which 1s from 8.33 mm. to 9.40 mm. II. SUMMARY.OF THE ORAL AND CUTANEOUS GROUPS. TABLE XIII. Table showing the relative time of appearance of the various buds in the cutaneous, oral and pharyn- geal groups. Empryo. a3 CuTANEous. Orat, PHARYNGEAL, 1n Hours. K’ | 113 Dorsal lip and barb. I, 2 and 3 gill arch buds. buds. L 120 Ventral lip and barb. buds. M 128 N 138 Nasal buds | Ant. pal. buds. Ventral pharyng. buds. O 146 4 gill arch buds. O’ 155 Dorsal pharyng. buds. Dis. hyoid buds. Le 163 Q 274 R 183 Post. pal. buds. 5 gill arch buds. S 199 CEsoph. buds. at 213 U 244 | Post.-orb. and opercu- lar buds. Oi se lle Sooo ooSa0 Body buds. 14} mm. 174mm. hoy Tilia lo eanvancops Cerebellar and occipital buds. | LanpacreE, Taste Buds of Ameiurus. 43 In summarizing the result of the study of the oral and cuta- neous groups two facts are patent, first, the spreading of buds is always from anterior to posterior, and second, the spreading is always by discontinuous groups. - Table XIII shows the time of appearance of the cutaneous, oral and pharyngeal groups. From Table XIII it will be seen that all the subdivisions of the pharyngeal group are present long before the cerebellar and occipital buds have appeared, and are present some hours before the body buds appear. This is to be explained probably by the homogeneous character of the pharyngeal area and by its limited extent as compared with the cutaneous area. There are insufficient data to make a comparison of the ontogenetic order in which taste buds innervated by the various rami of the V and VII nerve appear in Ameiurus and other types. If we arrange the nerves of Ameiurus in an order corresponding to the order of appearance of the groups of taste buds it will give the following arrangement: TABLE XIV. Table showing the order in which the groups of taste buds appear as related to the various rami of the V and VII nerves which innervate them. 113 hrs. | Dorsal lip and barb. buds innervated by R. max. V 120} co Ventral lip and barb. buds innervated by R. mand. V rigye) AY Nasal buds innervated by R. oph. Sup. V 138) Ant. palatine buds innervated by R. pal. VII OA ce Post. palatine buds innervated by R. pal. post. VII api, Post.-orbital and oper. buds innervated by R. max. acces. V? R. mand. ex. VII R. hyomd. (prox. twigs) R. oticus R. hyoideus 11 mm. _ Body buds R lat. acces. 204 mm. Cerebell. and occip. buds innervated by Meningeal twigs and Proximal twigs of R. lat. acces. Whether we look upon the taste buds as appearing fortuit- ously and later being connected with their gustatory nerves or take just the opposite view, which seems more in accord with the evidence, and look on the nerve fiber as taking the initiative and producing the bud on reaching the surface, the order in which the nerves are arranged here would seem to indicate the 44 ‘fournal of Comparative Neurology and Psychology. relative times at which communis fibers are found in these rami of the V and VII nerves. The spreading of buds from anterior to posterior and their appearance in detached groups have been sufficiently emphasized in the discussion of these groups. ‘The real significance of these two facts when brought into relation with the nerve supply will be discussed later. In connection with this table, it is of interest to note that these groups are not determined solely by the nerve supply, although in some cases groups are indicated by the nerve supply, as in The two palatine groups. ‘There are, however, enough of these groups which either have several nerves supply ing the group or else branches from the same nerve supplying two groups to indicate that the fundamental fact is the anterior- posterior spreading. The appearance of buds on the peripheral distribution of the nerve earlier than on the proximal distribution, which is of so frequent occurrence, seems to be subservient to the anterior-pos- terior spreading, since it 1s reversed in the case of the body buds which lie posterior to the origin of the nerves. The smaller subdivisions of the units or groups are deter- mined in almost all cases by the number of the nerves supply- ing these subdivisions and by the fact that they appear nearly simultaneously. “T’hey are characterized briefly by being slightly discontinuous but usually simultaneous in time of appearance. The units, on the other hand, are never continuous with ad join- ing groups at the time of appearance, and never simultaneous with the time of appearance of adjoining groups. ‘They differ from the smaller subdivisions of which they are composed in the character of the nerve supply also. Some, as the anterior and posterior palatine groups, have a single nerve supply, while others, as the opercular, have as many as five nerves, some of them from the VII and some from the V.. Still other groups are innervated by nerves, some of whose branches innervate other groups far- ther forward. The above facts seemed to the writer to furnish a clue, not only to the origin of ectodermic from endodermic buds if such prove to be the phylogenetic method, but also to the innervation of ectodermic buds by communis fibers. The whole difficulty lies in the latter condition, for if we can explain the innervation of ectodermic buds by communis fibers the difficulty in the LanpacreE, Taste Buds of Ameturus. 45 derivation of ectodermic from endodermic buds, if it exists, dis- appears. An attempt to explain the relation of ectodermic to endoder- mic buds should be in harmony with the following facts. First, buds situated in both endodermic and ectodermic areas are supplied by communis fibers. ‘This statement is confirmed by all the workers on nerve components. ‘Those of the oral and cutaneous areas in Ameiurus are supplied exclusively by fibers from the geniculate ganglion. Second, the number of nerve rami through which the genicu- late ganglion sends communis fibers varies greatly in various aquatic types (see historical sketch, p. 7, Table I), ranging froma few or even only one as in the larva of Petromyzon (JOHNSTON ’05) to nearly every ramus of what are commonly designated as the V and VII nerves in such types as Ameiurus. ‘Third, the ontogenetic method of increasing gustatory areas in Ameiurus is by detached groups, spreading from anterior to posterior. ‘This indicates that the number of nerve rami through which the geniculate ganglion sends fibers increases from fhe earlier stages of Ameiurus to the later, so that the total number of rami carrying gustatory fibers is not complete until more than five days after hatching. The assumption involved in the last statement, 1. e., that the appearance of the taste bud indicates the time at which the nerve supplying it reaches the surface needs verification for taste buds in Ameiurus, but has been shown to be true by Szymonowicz for (’95) tactile corpuscles and later (96) for Granpry’s and Hergst’s corpuscles. ‘These struc- tures, according to this author, appear as the result of the growth of the nerves to the areas where they are developed and the same is probably true for taste buds. The assumption is fur- ther strengthened by the fact that on section of the gustatory nerve in rabbits (Semr Mryer ’96), the taste buds revert into ordinary epithelium. The experiments of Lewts (’04) in trans- planting the optic cup, and producing a lens in new areas 1s particularly interesting in this connection. An apparent objec- tion to the assumption is the disappearance of gustatory fibers with the loss of superficial taste buds in terrestrial vertebrates. This objection is greatly minimized, if not entirely negatived, by the results of experiments on sectioning nerves (RANSON 06, and his review of the literature), and the consequent retrograde 46 fournal of Comparative Neurology and Psychology. degeneration of sensory fibers and the loss of ganglion cells. The disappearance of specialized communis fibers in terrestrial ani- mals might be accounted for on some such basis as the above. Fourth, the varying extent of the body covered by taste buds in different types and the varying number of nerve rami through which gustatory fibers from the geniculate ganglion run in dif- ferent types indicate a very great degree of variability in this ganglion. Any variation in the number of sensory fibers must of course be traced to the variability 1 in the ganglion from which they come. As to the variation in the number of taste buds, it seems hardly probable that they could have increased in num- ber by varying fortuitously and later have been connected with gustatory nerves, since this would involve the origin and devel- opment of a definitely constructed sense organ, in types having a peripheral and central nervous system, that would have to exist as such without a function until connected with its appropriate nerve. Whatever may have been the phylogenetic method of origin of specialized communis fibers supplying specialized sense organs, the method in higher types seems to be that the special- ized communis fiber produces its appropriate organ on reaching the surface. As to the possibility of variations in ganglia in general, the work of Harpesty (’05), Harar (’o2) and Ranson (06) ts interesting. Ranson, in particular, has shown that the same spinal ganglion may vary by as much as 21 per cent of the total number of cells in the smailer ganglia in white rats of the same age. All three of these authors have called attention to the fact that ganglia always contain a large excess of cells over the num- ber of fibers coming from these ganglia, which would seem to indicate that we have here a structural basis for variation in the number of functional cells and fibers. The explanation of the methods of increasing gustatory areas supplied by the V and VII rami mentioned above, may be stated provisionally as follows: The geniculate ganglion varies in the number of gustatory fibers it sends to the surface in various aquatic types and in different ages in the same type, as in Ameiu- rus. Some of these fibers on reaching the surface produce taste buds, whether in the ectoderm or endoderm. ‘The functional needs of the organism determine the direction and manner of spreading, The various rami of the V and VII nerves which Lanpacre, Taste Buds of Ameturus. 47 carry gustatory fibers in Ameiurus and do not in less specialized types are looked upon as routes through which these fibers reach the surface and explain to some extent the discontinuous method of spreading. Finally, the disappearance of gustatory fibers in terrestrial vertebrates from the rami of the V and VII nerves which bore such fibers in aquatic forms may be explained as a process of retrograde degeneration. This hypothesis seems to be in accord with known facts and certainly does away with the difficulty of deriving ectodermic from endodermic buds. Spreading of buds from endoderm into ectoderm in the strict sense of the word does not occur in Ameiurus or in any other type so far as the writer is aware. Buds appear in the ectoderm in detached groups. ‘There is a possibility that buds in ectodermic territory may actually spread into endodermic territory, as was mentioned above, in the case of the posterior palatine group. ‘This, however, 1s probably peculiar to Ameiurus and has no bearing whatever on the question as to where taste buds first appeared phylogeneti- cally. In regard to this problem, the evidence seems to be in favor of Jounstow’ s hypothesis (’05, ’06) that buds in primitive forms appear first in endodermic territory, since taste buds are always supplied by communis fibers which are visceral in their relationship as far as their central nuclei are concerned. ‘The hypothesis of CoLE (’00, p. 320) that taste buds arose first in the ectoderm and spread into the endoderm seems to be nega- tived by the visceral character of the communis system. Nei- ther the hypothesis of CoLE nor JOHNSTON seems to the writer to be tenable, if by spreading is meant continuous spreading; for this method does not occur in Ameiurus, and, as mentioned above, probably not in other types. A more fundamental ques- tion, and one whose solution would include to a large extent that of the place of first appearance, is involved in the source of the specialized portions of the communis system. Have specialized communis or gustatory fibers innervating taste and terminal buds arisen simply through the differentiation of unspecialized communis fibers? If they have, we should expect taste buds to arise first in the endoderm, as JOHNSTON suggests. However, before any definite conclusion is reached we must know the fate of that portion of the geniculate ganglion which is supposed to be derived from the epibranchial placode. It 48 Fournal of Comparative Neurology and Psychology. is conceivable that the specialized portions of the communis sys- tem may be derived from that portion of the geniculate ganglion which came from the epibranchial placode. In that case its specialized character could be traced to the fact that it was derived froma specialized sensory area before it became buried and a part of the ganglion. The question of where buds appeared first phylogenetically becomes then of secondary importance, since they might appear at any place where these specialized communis Ahers reached the surface. I2. THE GROSS DISTRIBUTION OF THE PHARYNGEAL GROUP. The pharyngeal group of buds, like the oral, appears on K/’ (113 hours), where buds appear on the first, second and third gill arches simultaneously. The smaller number, two, on the third arch indicates, however, that if series had been taken at small intervals preceding K’, they would have been found on the first gill arch first, then on the second, and then on the third. This probability is strengthened by the fact that buds on the fourth gill arch do not appear until in O and on the fifth or posterior demi-branch until in R: TABLE XV. Table showing the gross distribution of the taste buds on the distal portion of the hyoid arch, the gill arches, the roof and floor of the pharynx and in the esophagus. Emeryo. | Hyorp. | 1 Gitt.| 2 Grxt.| 3 Gixt.| 4 Git.) 5 Gitt.| Froor.| Roor. CEsoru. K’ | Sak 4 a | | | = L — | 4 Ae a 2: | = M — 4 8 | — — aa N = 14 15 4 == = 14 = = O = 22 18 9 4 eos 16 = = O’ | 2 16 17 ee die te) — 16 4 = 12 2) 20 25 20 | 10 | — 17 8 — 1@) Om Z9 30 2 | Fis ae 16 28 | = R II 38 39 29 25 2 Ze Na 35: Ni = S 10 32 38 32 23 5 24 44 2 T 14 55 55 48 | 37 6 40 92 5 U rey lists 66 6Gunienao 15 69 12 22 In the earlier series the first buds to appear lie on the free portions of the gill arch, that 1s, not on the point of attachment, although they are quite near the median ventral line, and do not reach to the dorsal portion of the gill, but appear here later, so Lanpacre, Taste Buds of Ameiurus. 49 that there is a movement of the buds from the ventral half of the gill arch dorsally toward the roof of the pharynx. ‘The appearance of buds on the roof in series O’, where four buds are present, is due to this spreading of the gill buds to the area over the gill arches. Buds appear much earlier on the floor of the pharynx, where they are situated on a prominent ridge exactly in the me- dian line. These mid-ventral buds appear much larger in their immature stages than any other buds in the embryo, but at their later stages are perfectly typical. ‘Their apparently larger size is due to the fact that the ridge itself, which is segmented, may be very easily mistaken for the taste buds. This mid-ventral series, situated as it is on a prominent ridge, is well defined in its situation and there is no difficulty in sepa- rating it from adjoining lateral areas except in the later series R, S, ‘I’ and U, where the buds appear on the fifth arch. Here it is dificult to separate the two series on account of the enlarge- ment of the copula by the appearance of the inferior pharyngeal teeth and also on account of the fact that the cartilages of the. fifth arch lie parallel to the copula. Buds appear in the cesophagus first in series S, where there are two, and increase to 22 in U. In the oldest series 204 mm. there are about 60. Leaving out of consideration for the present the hyoid group and those buds on the floor of the pharynx, the most evident thing about this table is the progression from an anterior to a posterior position of buds on the gills and in the cesophagus. There is here a segmental or branchiomeric order of appear- ance arranged in the following manner. Buds probably ap- pear first on the first gill, then on the second, then on the third, and then on the fourth gill arches, and finally appear in the cesophagus. The dorsal and ventral groups, as will be seen later, correspond quite closely in their segmental arrangement with the gill buds. This order probably represents not only the ontogenetic but the phylogenetic method of appearance as well. The segmental appearance of buds in the pharynx is in sharp contrast with the method by which they appear in the oral and cutaneous groups. The apparent exception to the segmental appearance of the buds in serial order in this group is furnished by the hyoid arch and the ventral pharyngeal buds. The hyoid arch lies anterior 50 Fournal of Comparative Neurology and Psychology. to the gill arches but does not acquire buds until in series O’, which is much later than the time at which buds first appear on the gills, and corresponds in time with their appearance on the fourth gill arch. In discussing the posterior palatine division it was suggested that possibly that group belonged, at least functionally, with the pharyngeal group, since Its position corresponded in the longi- tudinal axis to the anterior gill buds. If now we compare the position of the posterior palatine group, and of the proximal and distal hyoid groups, and the mid-ventral group, we find the fol- lowing relations as indicated in Table XVI, where the areas covered by these groups are given in sections. The anterior and posterior sections of each group marking its boundaries. TABLE XVI. Table showing extent of the anterior palatine, posterior palatine, proximal and distal hyoid, and mid-ventral floor group of the pharynx. EMBRYO. Ant. PAL. Post. Pat. Prox. Hyorp. Dist. Hyori. | Froor. N 55-65 = re we | 90-143 O 43-73 i iz pie go-158 oS 56-67 a = 75-79 | "67=aha 1 58-76 | = = 103-109 88-178 Q 56-70 | a= = 76-102 | 75-157 R 46-66 | 108-122 105 86-104 86-206 S 56-98 123-140 129-153 93-126 | 93-205 it 55-104 104-154 132-154 105-126 96-217 U 57-110 110-170 152-174 IOZ-150 =| + 93-283 The distal hyoid group fits in with the proximal hyoid group, the anterior and posterior palatine and the mid-ventral group, to form a structurally homogeneous group of buds on the roof, floor and sides of the area in which the hyoid lies. The relations are somewhat different in series O’, P and Q, where the posterior palatine and proximal hyoid buds have not yet appeared, from that in series R, S, IT and U, where those groups are present and serve to cover the roof, floor and sides of the pharynx completely. The condition of the buds in U will serve to explain these relations most easily. In U the posterior palatine group extends from section I10 to 170; the distal hyoid group from 103 to 151 and the proximal hyoid from 152 to 174, making the whole extent of the hyoid group from 103 to 174, which corresponds quite closely with Lanpacre, Taste Buds of Ameturus. 51 the area on the roof occupied by buds of the posterior palatine group, which is from section 110 to 170. On the floor of the pharynx we have the mid-ventral group, beginning at 93 and extending back to the beginning of the cesophagus. ‘lhe mid- ventral buds, of course, furnish the ventral portion for the gills as well as for the hyoid arch. Fig. 1 on Plate I will make these relations somewhat clearer. It will be seen from this diagram that the distal hyoid buds really represent a continuation of the anterior oral group rather than the beginning of the pharyngeal groups. The distal hyoid group 1s a part of the posterior palatine group structurally, that 1 is, in time of appearance, and doubtless func- tionally also, for it is dificult to see how buds lying on the roof, floor and sides of the pharynx could fail to be stimulated at the same time, since they lie in practically the same area measured on the longitudinal axis of the embryo. ‘The innervation of the distal portion of the hyoid is from the ventral post-trematic ‘branch of the ninth which enters it at the point of union with the copula and runs proximally. The posterior palatine and proximal hyoid group, on the other hand, are innervated by the r. palatinus posterior VII and the mid-ventral buds in front of the union of the first gill arch with the copula are supplied by the same nerve as the dis- tal hyoid group. Probably no better illustration of whaes is meant by functional groups or areas in the appearance of taste buds could be given. This area is composed of four subdivisions having definite anterior and posterior boundaries and is so situated that it would seem to function as a unit. Portions of it are innervated from the seventh and other portions from the ninth nerve. The hyoid arch in particular has these two nerves supplying it, the former going to its dorsal or proximal portions and the latter to its distal or ventral portions. ‘The distal hyoid buds represent the posterior continuation of the oral group in point of time and not at the beginning of the pharyngeal group. In series O, P and Q the dorsal buds are represented by the anterior palatine group , only, the ventral buds by the mid-ventral group and the lateral area by the distal hyoid group, there being a vacant area on the proximal hyoid which later is occupied by that group of buds. On the roof the posterior palatine is inter- ‘52 ‘fournal of Comparative Neurology and Psychology. polated between the anterior palatine and the buds lying on the roof of the pharynx at the point where the first gill arch joins the roof. I3. CERATOBRANCHIAL AND ROOF GROUP. This group consists of buds situated on the ceratobranchial bar, the pharyngeal cartilages and the roof of the pharynx. The dorsal side of each ceratobranchial is innervated by the dorsal branch of the post-trematic division of the ninth or tenth nerve; the dorsal side of the first gill arch by the dorsal branch of the ninth and the second, third and fourth by the same branches of the second, third and fourth divisions of the tenth nerve. The posterior demi-branch behind the fourth gill slit is innervated by the ventral pharyngeal branch of the fourth division of the tenth nerve in Menidia. As stated above, the first buds to appear on the gill are situ- ated on the distal portions of the ceratobranchials of gills 1, 2 and 3- From this point they spread proximally, or dorsally, until in series O’ they reach the pharyngobranchials. Table XVII shows the time of appearance of buds on the pharyngo- branchials and the roof of the mouth. TABLE XVII. Table showing time of appearance and extent of buds on the pharyngobranchials, the extent indicated by the section numbers between which buds are found. Ist PHARYNGOBR. | 2d PHARYNGOBR. 3d PHARYNGORR. 4th PHARYNGOBR. EMBRYO. | Extent. | No. EXTENT. No. | Extent. | No. | EXTENT | | | O’ 4 | 126-132 4 147-132 —_— —_— — | —- P Z 140-142 | 2 156 4 166-171 ae QO 132-139 9 | 149-158 2 164-166 | 2 180-185 In series O’ and P all buds are found on the pharyngobran- chials and there is no overlapping, the groups being quite dis- tinct; but in Q there are 28 buds present in the roof of the pharynx, only 18 of which are catalogued and shown in the table, the remaining ten being scattered between the pharyngo- branchials. In Q the first, second and third pharyngobranchial groups overlap, since the cartilages do, so that in Q we have a LanpacreE, Taste Buds of Ameturus. 53 continuous series of buds running from section 132, where they first appear, back to the cesophagus, but in the two preceding series this is not the case. In the region of the dorsal pharyngeal teeth it is not possible to separate buds lying on the fourth and fifth pharyngobranchials (or rather epibranchials, since there is only one cartilage present for both these arches), and the mucous membrane is disarranged by the presence of numerous teeth. From the table it will be seen, however, that the buds appear on the pharyngobranchials in the same order as on the gills; that is, on the first, then on the second, then in P on the third, and in O on the fourth. No buds are found on the outer sides of the gill arches on the areas occupied by the gill filaments and no buds spread from the dorsal end of the ceratobranchials or the roof of the mouth in the region of the pharyngobranchials out to the inner surface of the operculum up to a period at least as late as the 20} mm. embryo. In this embryo there are buds on the inner surface of the anterior portion of the operculum which belong with the mandible, and buds also extend back on the inner side of the hyoid belonging to the proximal hyoid group, but there is a wide area from this point back to the posterior free border of the operculum totally devoid of buds, so that unless they appear later than any of my series there is no evidence that buds reach the outer body surface from the gill region. If they should prove to be found in older embryos to spread back from the mandibular and proximal hyoid groups to the inner side of the operculum beyond the point I have indicated, it would furnish no evidence that buds move from endodermic to ectodermic territory, since these areas are continuations of ectodermic groups and have their innervation from the seventh nerve and not from the ninth and tenth. It should be noted that the spreading here is from anterior to posterior, so that it would be much more likely that buds spread from ectodermic to endodermic territory, provided the anterior limit of the pharynx is cephalad of the area occupied by the proximal hyoid and suspensorium. Buds lying on the floor of the pharynx on either side of the median line are innervated by the dorsal branches of the post- trematic branchiomeric nerves in serial order. ‘These nerves 54 fournal of Comparative Neurology and Psychology. also innervate the buds of the ceratobranchials. No buds 176 /49 175° . 17 30 Fic. 1. Shows the relative extent measured in sections of 7 microns each of the various divisions of the anterior palatine, posterior palatine, pharyngeal, proximal and distal hyoid and mid-ventral buds on the longitudinal axis of the embryo. The numbers indicate the sections bounding of the various sub-divi- sions. The diagram is drawn to scale from series U. Fics. 2 and 3 are drawn to scale also, and show the extent of attachment of the hyoid and gills to the copula and the extent» of the median ventral ridge on which buds are situated and the segmentation of thisridge. Fig. 2 is plotted from series N, and Fig. 3 from series U. A, the extent of attachment of the hyoid and gills in each figure. B, the extent of the median ventral ridge and its segmentation corresponding to the areas of attach- ment of the gill. The numbers on the diagram represent the limits in sections. fourth FiKt4 205 228 214 124 127 146 178 170 (175 177 188 /92 1/97 200 206 209 2/6 2/8 A fet oe my Aaa ‘i Sea's 7 ‘ ry d A STUDY OF THE VAGAL LOBES AND FUNICULAR NUCLEI OF THE BRAIN OF THE CODFISH. By C. JUDSON HERRICK. (Studies from the Neurological Laboratory of Denison University, No. XX.) Witn Eicut Ficures. INTRODUCTION. The peripheral and en organs of taste have received more or less careful study in three distinct groups of teleostean fishes in which taste buds are known to occur plentifully in the outer skin; viz.: the cyprinoids (carp, etc.), the siluroids (Ameiurus and other catfishes) and the gadoids (cod, tom-cod, hake). The structure of the cutaneous taste buds of the carp was described by Lreypic in 1851 and more accurately in 1863 and 1870 by F. E. Scuurze, who correctly inferred their function. ‘Their gustatory function was subsequently demonstrated physiologically in Ameiurus and various gadoids (HERRICK ’04), and it was shown that either tactile or gustatory stimuli alone may be correctly localized in the outer skin, though ordinarily both senses cooperate in the locating of the food. The same reaction may follow either a tactile or a gustatory stimulation of the outer skin or the simul- taneous stimulation of both kinds of sense organs in the same cutaneous area. The distribution of the nerves of touch in the skin of fishes is very accurately known and the general arrangement of these nerves is nearly constant throughout the phylum. ‘The innervation of the cutaneous taste buds has also been studied in the three types here considered—exhaustively in Ameiurus (HERRICK ’o1) and less completely in Gadus (Herrick ’oo) and Carassius (HERRICK "04, p. 249). The nerves for the cutaneous taste buds spring from the communis root of the facial nerve, without exception in Ameiu- rus and with but small exception in Gadus. ‘The same facial root supplies also taste buds in the anterior part of the mouth. 68 Fournal of Comparative Neurology and Psychology. The central gustatory paths have been investigated in cyprinoids and siluroids (HERRICK ’05). In view of the demonstration, in the works cited in the preceding paragraphs, of the essential simi- larity in the structure, innervation and function of the cutaneous taste buds of cyprinoids, siluroids and gadoids, we should expect their central connections in the group [eet mentioned to resemble those described for the two former groups. But the facts seem to be quite the contrary, and the primary purpose of this inquiry was to determine the extent and if possible the functional signif- cance of these differences. This naturally led to a study of the tactile centers also in Gadus, for the underlying problem in both this and my previous study (’06) is the central relations of those tactile and gustatory nerves which innervate the same areas of skin and have independent local signs but common pathways of motor discharge in the reaction. THE CENTRAL GUSTATORY PATHS OF AMEIURUS AND GADUS. In siluroids and cyprinoids practically all the nerves supplying taste buds in the outer skin, lips and palate terminate in a single huge nucleus which forms a dorsal protuberance on the medulla oblongata, the facial lobe, while taste buds of the pharynx and gill region are innervated from the vagal lobe farther back in the oblongata. In the gadoid fishes a facial lobe has been described and figured by several previous authors; but my examination of the brain of the cod shows that the facial lobe does not exist in the form described by these authors. Nevertheless | find the peripheral distribution of the gustatory root of the facial nerve of Gadus is practically the same as in Ameiurus. What, then, are the central connections of the cutaneous taste buds in Gadus? Taste buds are scattered over the whole body of Ameiurus but most abundantly on the barblets, and experiment shows that these are very sensitive to both tactile and gustatory stimulation. ‘They are constantly used for the exploration of the bottom and the maxillary barblets are actively waved about whenever a savor dif- fuses through the water. In the gadoids, particularly the tom-cod and the hee: the free filiform rays of the pelvic fin function in a similar fashion and are likewise richly supplied with end-organs of both taste and touch. Herrick, Brain of the Codjfish. 69 In Ameiurus the facial lobe, which receives all the gustatory nerves from the outer skin, gives rise to two chief secondary tracts. One passes upward into the mid- brain and one downward toward the spinal cord. The first of these accompanies the ascending cerebral tract for taste from the vagal lobes and is feebly developed orabsent inthecod. It will therefore receive no further considera- tion here. The descending path terminates, as shown in Fig. 1, in the primary tactile correlation center of the brain, the funicular nuclei, and farther caudad in the spinal cord. From this corre- lation center a common motor pathway runs out by way of the funiculus ventralis for the innervation of all the somatic muscles of the body. Now remembering that the body, considered as a reflex mechan- ism comprises two primary systems (the somatic system for reaction of the body as a whole to external stimuli, and the visceral system for correlation of the internal parts by reaction to visceral stimul1), we may summarize the reflex connections of the taste buds of Ameiurus as follows: The primary gustatory center is differen- tiated into a visceral region (vagal lobe) for taste buds lying in mucous membranes and a somatic region (facial lobe) for taste buds lying in the outer skin. ‘The somatic character of this latter region has been secondarily acquired. It was without question phylogenetically derived from the visceral center. Both centers send an ascending tract to a common mid-brain nucleus. ‘Their descending paths are strikingly different. The visceral center (vagal lobe) makes reflex connections only with the visceral muscu- lature of the jaws, gills, cesophagus, etc. “The somatic center (facial lobe) connects broadly with the funicular nuclei and there the gustatory stimuli from the skin are correlated with tactile stimuli from the same cutaneous areas and from this common sensory correlation station the motor pathways go out to the soma- tic muscles. There is-also a direct path from the facial lobe to the motor V nucleus which innervates the muscles of the barblets, these being waved about when stimulated either by taste or touch. The preceding description applies with but slight modification also to the carp and other cyprinoid fishes. Now, in the cod there is no well defined facial lobe. Gustatory fibers from the mouth and from the outer skin terminate in the vagal lobe, a condition which was entirely inexplicable to me after my demonstration physiologically that the taste buds on the fins - 70 Journal of Comparative Neurology and Psychology. of the gadoids and on the barblets of Ameiurus function similarly as organs of taste, that the fishes can localize pure gustatory stimuli applied to these organs and that ordinarily both taste and touch cooperate in finding food by means of them, though the two sensation factors may be experimentally isolated by training. Accordingly, I have reéxamined the vagal lobes and their con- nections in the cod and find that the gustatory nerves of the somatic type, those from the outer skin, and the gustatory nerves of the visceral type, from the mucous membranes, do have distinct cen- tral connections, though in quite a different way from that found in the catfish. The facts are these (c7. Fig The vagal lobe of the cod is internally divided by a longitudinal septum into median and lateral lobules of about equal extent (Figs. 3 and 4). The median lobule is primarily visceral and receives the gustatory roots of the LX and X nerves, as in other fishes. The gustatory root of the facial nerve, which carries prac- tically all of the fibers from taste buds in the outer skin and a smaller number of fibers from the mucous membrane of the ante- rior part of the mouth (ramus palatinus VII, etc.), terminates in both lobules, but chiefly in the lateral one. While I have not been able to demonstrate that the facial fibers which reach the median lobule are from visceral buds within the mouth, yet this 1s very probable. In any case, it is clear that the lateral lobule is prima- rily, if not exclusively, the terminal nucleus for the gustatory fibers of the somatic type; that 1s, it 1s the physiological, if not the mor- phological, equivalent of the facial lobe of siluroids and cyprinoids. The two important papers on the brains of teleosts published in the GrGenpaur Festschrift (Goronowirscu 796, and Harter 96) give figures of the medulla oblongata of Lota vulgaris, a form closely allied to Gadus. Both of these authors were so dominated by the endeavor to reduce all cranial nerves to segmental units of the spinal type as to vitiate to some degree their observation. Their two figures of the brain of Lota bear very little resemblance to one another, or to the brain of Gadus, which is somewhat remarkable in view of the fact that both authors were studying the internal courses of the nerve roots microscopically. Hater shows (Plate I, Fig. 6) a series of four pairs of enlargements in the vagal region, the first belonging to the IX nerve, the others to the vagus. The figure of Gorono- witscu (Plate I, Fig. 3) is evidently much more accurate as to externals, but is in part incorrectly inter- preted. He figures a lobus facialis (my lobus vagi lateralis), to which he correctly traces the communis VII root; a lobus glossopharyngei (my lobus vagi medialis), to which he correctly traces the IX and X nerves; a lobus vagi impar (my visceral commissural nucleus), to which he traces vagus fibers; anda lobus vagi which corresponds to my somatic commissural nucleus. The error in the latter point is due to his failure to recognize the distinction between the somatic and visceral regions of the oblon- gata. With this point in mind it is impossible to confuse the vagal lobes with the somatic com- missural nucleus. Herrick, Brain of the Codfsh. 71 The subdivision of the vagal lobe of Gadus into median and lateral lobules is mentioned in the catalogue of the Museum of the Royal College of Surgeons of London (Burne ’o2. p. 93) but unfortu- nately inthe absence of a knowledge of the internal structure Mr. Burne here was led by external appear- ances and by the previous figures of Lota by Goronowirtscu (’96) into error in the interpretation of this region Jn the two dissections of the brain of the codfish there figured and in the accompanying text the true vagal lobes are termed facial lobes and the somatic commissural and funicular nucleus complex of my descriptions is termed lobus vagi and is said to give rise to the sensory roots of the vagus nerve, Examination of the internal courses of these nerves at once corrects this mistake. The dissections of the codfish brain figured in T. J. ParKer’s Zodtomy (’oo, p. 125) show the vagal lobes very small with no designation. The “lobi posteriores,”’ which might be mistaken for the vagal lobes, are the tubercula acustica, The excellent series of transections of the codfish brain given by Kappers (’06) extends back only to the cephalic end of the vagal lobe. Fig. 4 of the present paper is from a section taken a short distance caudad of Fig. xcix of Plate VII of Kapprrs’ memoir. In Fig. 3 I present a photograph of this region of the brain of a large codfish (Gadus morrhua), whose internal connections are as shown in Fig. 2. From the figure and description of GoronowiTscH, I have no doubt that the relations are essentially similar in Lota. The central tracts leading from the median lobule are the same as those from the vagal lobe in other fishes; and caudad this lobe merges into the visceral commissural nucleus of Cajat in the typical way. The lateral lobule, which receives only facialis fibers, is wholly unlike the facial lobe of the siluroid and cyprinoid fishes in its secondary connections. It contributes few, if any, fibers to the long ascending secondary gustatory tract, this tract being derived from the median lobule (Fig. 4). More- over, there does not arise from any part of the vagal lobe a large, clearly defined descending secondary gustatory tract, like that from the facial lobe in Ameiurus and Cyprinus, entering the dorso-lateral fasciculus for the funicular nucleus region and spinal cord. The lateral lobule is directly continuous caudad with the funicular nucleus region. A considerable tract of delicately medullated fibers accumulates on the ventro-lateral border of the lateral lobule and passes directly back to enter the cephalic end of the funicular nucleus where it joins the substantia gelatinosa Rolandi. My Wiegert sections do not show positively whether these fibers are ascending or descending, but I assume the latter by analogy; for this tract seems to correspond with the descending secondary gustatory tract from the facial lobe of Ameiurus. But in the cod this is a small and unimportant tract. The chief secondary connection of the lateral lobule 1s by a very strong and heavily medullated tract which passes from its entire extent Into the ventral commissure and enters the ventral funiculi of the same and the opposite side. These appear to be descend- 72 Fournal of Comparative Neurology and Psychology. ing fibers. Some of them may pass over into the opposite tractus bulbo-tectalis (lemniscus), but it 1s not possible to be certain on account of the confusion in the ventral commissure of the internal arcuate fibers from the lateral vagal lobule with those from the tuberculum acusticum. ‘This connection through the ventral commissure puts the lateral vagal lobule into connection with the long conduction paths of the somatic system, and thus directly connects the primary center for cutaneous taste buds and the ven- tral cornua of the spinal cord, from which the muscles of the fins and body are innervated. ‘These relations are shown diagram- matically in Fig. 2 The representatives of the Ostariophysi which I have studied (Ameiurus, Cyprinius, Catostomus, etc.) agree in possessing a distinct center (the facial lobe) for all of the fibers from cutaneous taste buds, the secondary connections of this center being partly with visceral motor and partly with somatic motor centers via the funicular nuclei. ‘The fact that Gadus accomplishes a somewhat similar result by the different and more direct method of separating at the start the facialis root fibers into those for visceral and soma- tic centers is another illustration of the distinctness of the Ostario- physi from other teleostean fishes. ‘The end result is similar in the case of the cod and the catfish. Peripheral areas of skin may receive both tactile and gustatory stimulation simultaneously. The fish reacts to the composite stimulus by a single movement of the body adapted to reach and seize the food object. ‘The tactile path in both cases leads to the funicular nuclei, and thence to the somatic muscles. “The somatic gustatory path in the catfish leads to the tactile correlation center (funicular nuclei), whence it reaches the somatic muscles by the same tracts as the tactile. In the cod the somatic gustatory path passes directly from the primary center to the somatic motor cen- ters in the ventral cornu without interruption in the tactile centers. In searching for the explanation of this difference two lines of inquiry are at once suggested. First, are there any mechanical necessities of cerebral striehine sufficient to account for them; and, second, do the habitual modes of reaction to external stimull, 1. ¢., the habits of the animals, suggest an explanation. I believe that both of these factors have operated. In the first place, why do the somatic gustatory nerves of the cod end in a specialized part of the vagal lobe and those of Ameiu- Herrick, Brain of the Codfish. 73 rus in a separate lobe in front, the lobus facialis ? Why should not the cutaneous gustatory nerves of Ameiurus end likewise in the vagal region? ‘hat there is no mechanical impediment to the necessary enlargement of the vagal lobe is evident from the fact that in the carp the vagal lobe suffers much greater enlargement to provide a terminal nucleus for an increased number of taste buds within the mouth. In Ameiurus many, though by no means all, of the cutaneous taste buds are in the head. ‘hese areas of skin receive their tactile innervation from the trigeminus nerve. Now, I have found in this fish two centers of correlation between the nerves of touch and taste from the skin of the head. One is the funicular nuclei, already referred to. The other is in the facial lobe, in whose deeper layers trigeminus root fibers have been found to end. ‘The demand for a correlation center in front of the vagal lobe is proba- bly the motive which in Ameiurus has drawn the facialis gustatory center cephalad of the vagal lobe, thus providing also an imme- diate path from end organs of both touch and taste to the motor centers of the barblets and jaws. In the gadoids, on the other hand, some of the most important areas of distribution of cutaneous taste buds are on the fins, which are freely moved about in the exploration of food objects. ‘These receive their tactile innerva- tion from spinal nerves which enter behind the vagal lobes and, therefore, the motive for a forward movement of the somatic gusta- tory center to correlate with the corresponding tactile nerves does not exist to so high a degree, or possibly has been counteracted by stronger spinal tactile impulses associated with gustatory stimuli on the fins. But this explanation still leaves unaccounted for the short-circuit- ing of the somatic gustatory path in the cod by which it passes under the tactile centers without connection with them and reaches the motor centers directly. “he peculiar feeding habits of the cod may explain this arrangement (HERRICK 04). The body taste buds of Gadus and its allies are most abundant on the filiform pelvic fins, and these are the organs most used in the detection of food, serving a purpose closely similar to that of the barblets of Reacieras: In Ameiurus the somatic reaction consequent upon contact of a barblet with food is a lateral turning of the whole body by a single movement to reach the food object. 74 ‘Fournal of Comparative Neurology and Psychology. But in Gadus the movement is quite different, since the food stimulus is under the center of the body when it is detected. Before the food can be taken, the fish must check the forward movement and back up until the mouth has reached the object. This involves a very precise movement of the pectoral fins in par- ticular, and if the prey be living it must be very rapidly done. These features, taken in connection with the more active life of the gadoids in general, are sufficient to account for the short-circuiting of the reflex path between the gustatory root of the facial nerve and the ventral cornu of the spinal cord, so that a gustatory stim- ulus on the fins alone may cause the reaction promptly without the cooperation of the tactile centers. We have then, in summary, the’ following striking series of structural adapations correlated with the appearance of taste buds in the outer skin of fishes. Such buds occur in fishes gen- erally in the mouth and on the lips, the latter being innervated by the facial nerve. LaNpDacRE has shown in a recent embryo- logical research (’07) that in Ameiurus the cutaneous buds appear first in the region of the lips and then progressively farther caudad. It is probable that this was also the order of their appearance phylogenetically, a supposition which is supported by the course of the branches of the facial nerve which supply these cutaneous buds in the adult. As these facial gustatory nerves increased in importance, especially those from the barblets, cen- tral correlation was required with the tactile and motor centers for the barblets in the region of the trigeminus. ‘This anatomical connection finally caused the cutaneous gustatory center to move forward from the vagus region into the facial, and a further corre- lation was effected between the gustatory and tactile centers by means of the descending secondary gustatory tract from the facial lobe to the funicular nuclei. In the gadoids the fins, particularly the free pelvic fin rays, serve as motile organs of tactile-gustatory sensation. ‘The gusta- tory innervation is as before through the facial nerve, but the tactile through spinal nerves which enter the brain behind the vagal lobes. Accordingly, the somatic gustatory center does not migrate forward, but remains stationary, and the secondary gusta- tory path passes from it directly to the motor centers of the spinal cord instead of first to the tactile correlation center. This pro- Herrick, Brain of the Codfish. 75 vides a shorter path from taste buds on the fins to the muscles which move the fins and the body as a whole. THE COMMISSURA INFIMA AND FUNICULAR NUCLEI OF GADUS. The analysis of the region of the commissura infima Halleri and funicular nuclei is much more difhcult in Gadus than in some other types where the visceral and somatic elements of this com- plex are more highly differentiated, as in Ameiurus. For the typical arrangement and nomenclature of these parts the reader is referred to my recent paper (’06) on the centers.for taste and touch in Ameiurus and to Fig. 1 of the present article, which summarizes the chief conclusions reached in that inquiry. The somatic division of the commissura infima of Gadus is large and heavily medullated; the visceral division is unmedullated. The commissural nucleus is large and the somatic part is more extensive than the visceral. The median lobule of the vagal lobe passes directly back into the visceral commissural nucleus, which occupies the mid-dorsal line caudad of the vagal lobes (Figs. 3 and 5). ‘This nucleus con- tains no medullated fibers; it does, however, contain many small cells and a dense neuropil of unmedullated fibers, many of which cross the median line, forming the most cephalic part of the com- missura infima. ‘The nucleus ambiguus lies below it and gives rise to motor roots of the vagus. ‘The most caudal sensory root of the vagus is seen in Fig. 5 entering the caudal tip of the vagal lobe laterally of the commissural nucleus. In the Weigert sections of young fish here examined (the specimens were about 7 cm. long) no vagus root fibers are seen to enter the commissura infima. In this, I confirm the statements made for Gadus by KAPPERS (06), who also worked with Weigert preparations. It is, however, by no means clear from my preparations that no unmedullated termini of these root fibers cross in the commissure. In fact, the appearance of the sections strongly suggests that this is the case, as I have also found it in both Ameiurus and Cyprinus. Immediately behind the last sensory root of the vagus the soma- tic commissural nucleus fills the wide space embraced between the spinal V tracts and their nuclei of the two sides, and is composed of dense neuropil, large cells and medullated fibers in very complex 76 Fournal of Comparative Neurology and Psychology. formation (Figs. 3and 6). Just cephalad of the level of this figure there is a large transverse band of thick medullated fibers which connects the substantia gelatinosa of one side with that of the other or possibly with the opposite commissural nucleus. ‘There is shown in this figure a broad medullated connection between the commissural nucleus and the homolateral spinal V nucleus and formatio reticularis, and also fascicles of commissural fibers in the commissural nucleus. A very short distance caudad of the level of Fig. 6 the commis- sural nucleus greatly expands and merges laterally with the spinal V nucleus and funicular nucleus and ventrally with the formatio reticularis (Fig. 7). his complex area may also contain a por- tion of the visceral sensory commissure and nucleus, though no — part of it can be recognized as such. It 1s not possible to analyze this area into its component parts on the basis of microscopical appearances in Weigert sections, as ‘I have done in Ameturus. Short tracts of medullated and unmedullated fibers pass through it in all directions, many crossing the median line. A vestige of the nucleus ambiguus extends as far back as the level here figured under the commissural nucleus. ‘The relations shown in this figure continue essentially unchanged far back into the spinal cord, where the area in question gradually shrinks in size and passes over into the dorsal cornua. The first spinal nerve is a fusion of two or more nerves. The dorsal roots enter the complex area just mentioned, which at the level shown in Fig. 8 is designated cornu dorsalis. At the level of the origin of the: second spinal nerve (which joins the first in the brachial plexus) the relations are similar, though the dorsal cornu complex is much smaller and in its dorsal portion the true dorsal cornu is structurally well defined, with a small but distinct funic- ulus dorsalis laterally of it. At the level of the dorsal root of the third spinal nerve the dorsal area of gray matter 1s still further reduced, the dorsal cornu and funiculus are still more distinct and the other portions of the dorsal gray complex are reduced to a small median vestige. “The ventral ramus of this nerve also effects connection with the brachial plexus for the innervation of the pectoral fin. The fourth spinal does not enter the brachial plexus. The pelvic fin is innervated chiefly from the ramus ventralis of this nerve and by a smaller twig from the fifth spinal. Herrick, Brain of the Codpish. ii The dorsal cornua at the levels of the fourth and fifth spinals are reduced to the meager dimensions commonly seen in teleosts. The cross-section of the spinal cord at this level 1s almost com- pletely filled with very large medullated fibers, showing that long conduction paths are here more important than short reflex con- nections. ‘The dorso-lateral fasciculi, in particular, are large and heavily medullated. In this respect the cod resembles the eel and other fishes with highly developed body musculature, in striking contrast with the sluggish cathsh. Even the cyprinoids, like the carp and the gold-fish, have far smaller longitudinal spinal tracts. At the cephalic end of the spinal cord (third spinals to vagal lobes) the same compact formation of the long tracts is evident as farther caudad, save in the dorsal cornu and funicular nucleus region. The dorsal funiculi disappear in the most caudal part of these nuclei and the dorso-lateral fasciculus sends large tracts into them for their entire extent, suffering corresponding reduction in size cephalad. A considerable proportion of this fasciculus, however, passes farther cephalad to terminate inthe oblongata. ‘The ventro- lateral fasciculi also decrease greatly in size in the funicular nucleus region, or more properly expressed, they increase as they pass caudad under the funicular nuclei by accretions from this region. ‘The lateral fasciculi (tracts midway between dorsal and ventral funicull) also suffer considerable diminution in this region; but some strong bundles of these fibers pass directly cephalad from the spinal cord into the brain as the tractus spino-tectalis, to be greatly augmented in the region of the tuberculum acusti- cum by the tractus bulbo-tectalis. We conclude, then, that in Gadus the region of the funicular and commissural nuclei is, as in other types of fishes, a correlation center for all tactile impressions from the skin and their motor responses. ‘The pectoral and pelvic fins of the gadoids are particu- larly delicate tactile organs and the dorsal cornua of the anterior end of the spinal cord have been enlarged and intimately related to the funicular nuclei and somatic commissural nucleus to serve these sense organs. This process has been carried to a much greater extreme in the gurnards (Irigla, Prionotus, etc.). But ane taste buds located on these fins in the gadoids, as we have seen above, are not centrally connected with this tactile correlation center, as they are in the siluroids and cyprinoids, but effect an 78 Journal of Comparative Neurology and Peychology. independent and more direct connection with the ventral horn cells and other motor centers by way of the ventralfuniculi. For summary of these connections, see p. 74. Finally it is a pleasure to acknowledge my indebtedness to the. U.S. Bureau of Fisheries for the specimens of young codfish upon which the histological part of this paper is based, and to my col- league, Professor FRANK Carney, for assistance in the Prepakas tion of the illustrations. Denison University, December 22, 1906. Herrick, Brain of the Codfish. 79 LITERATURE CITED. Burne, R. H. 702. Descriptive and illustrated catalogue of the physiological series of comparative anatomy contained in the museum of the Royal College of Surgeons of England. Second Edi- tion. London, Vol. 2, pp. 90-93. Goronowitscu, N. 796. Hatter, B. 96. Der Trigeminofacialis-Complex von Lota vulgaris. Festschr. f. Gegenbaur, Vol. 3. Der Ursprung der Vagusgruppe bei den Teliostiern. Festschr. f. Gegenbaur, Vol, 3. Herrick, C, Jupson. ’99. The cranial and first spinal nerves of Menidia. Fourn. Comp. Neur., Vol. 9. ’oo. A contribution upon the cranial nerves of the codfish. Fourn. Comp. Neur., Vol. 10. ’o1. The cranial nerves and cutaneous sense organs of the North American siluroid fishes. Fourn. Comp. Neur., Vol. 11. ’o4. The organ and sense of taste in fishes. Bulletin of the U. S. Fish Commission for 1902. ’os. The central gustatory paths in the brains of bony fishes. ‘fourn. Comp. Neur. and Psych., Vol. 15. ’06. On the centers for taste and touch in the medulla oblongata of fishes. Journ. Comp. Neur. and Psych., Vol. 16. Jounston, J. B. . ’o6. The nervous system of vertebrates. Philadelphia, P. Blakiston’s Son & Co. Kaprers, C. U. Ariens, 06. The structure of the teleostean and selachian brain. Psychol., Vol. 16. Fourn. Comp. Neur. and Lanpacre, F. L. > Leypic, FR. Agile Parker, T. bi Scuutze, F. 63. 70. ? 07. oo. On the place of origin and method of distribution of taste buds in Ameiurus Melas. Fourn. Comp. Neur. and Psychol., Vol. 17, No. t. Ueber die aussere Haut einiger Siisswasserfische. Zeits. wiss. Zool., Vol. 3. JEFFERY. A course of instruction in zodtomy (Vertebrata). New York, The Macmillan Co. EE Ueber die becherférmigen Organe der Fische. Zeits. wiss. Zool., Vol. 12. Ueber die Sinnesorgane der Seitenlinie bei Fischen und Amphibien. Arch. f. mik. Anat., Vol. 6. = 8o Fournal of Comparative Neurology and Psychology. Fic. 1. Diagram of the gustatory and tactile paths in the medulla oblongata of Ameiurus, as seen from above. The gustatory connections are shown on the lower (left) side of the figure, the tactile on the upper (right) side. The gustatory centers are bounded by the fine dotted lines, the tactile centers by the broken lines (short dashes). The neurones of the commissural nuclei are not shown in the dia- gram. The position of the visceral commissura infima and its nucleus is indicated by two vertical crosses (+) behind the vagal lobes; that of the somatic commissure and its nucleus by a single oblique cross ( ) farther caudad between the two median funicular nuclei. Only the long secondary tracts are indicated. The short secondary and tertiary tracts from both visceral and somatic centers to the for- matio reticularis, etc., are omitted. The data from which the diagram is constructed will be found in two previous papers (HERRICK, ’o5 and ’o6), where cross sections and other figures of the oblongata of Ameiurus are given. Fic. 2. Diagram of the gustatory and tactile paths in the medulla oblongata of Gadus morrhua, as seen from above. The plan of the diagram is the same as that of Fig. 1, both gustatory and tactile centers being bounded by fine dotted lines. As before, the position of the visceral commissura infima and its nucleus are indicated by two vertical crosses (+) and that of the somatic commissure and its nucleus by oblique crosses (X X X). Herrick, Brain of the Codfish. 81 rx. sens. V (aleral funicular nucleus \ ynedian funicular nucleus \\ weinal Vv nucleus NN fase. dorso-laterals via Ae J). SP SENS. . "spot rx.cutaneus Super/or i Secondary: gustatory. nucleus — rs Bee A 1 He ae z f - e Z , = = ie Bs ey nt, asc ae funie dorsalis / = 7 4°. rx.sens X Jf Fonte. ventt rxsens IX (§ #ase dorso-lat. \ / ‘ ESC SEGIUI Te ep ears Ie. rx sens. VII. AME/IURUS EG te rx. sens. V rx culaneus vag! _& secundus V// N.gsp. mot. n. Sp. Sehs. fase. dorso-/at., J SUPES/IOL ; secondary qustatory mUCIEU Ss. Is SE Cornu dorsalis’ tr spino-tectalis’ , x sens. V// 7x. sens. 1IX+X funiculus ventral’s GAD US. fase. laterals BiG. 22 82 Journal of Com parative Neurology and Psychology. Fic. 3. A photograph of a dissection of a brain of a large specimen of Gadus morrhua. X 1.3. The cerebellum and a portion of the tubercula acustica have been dissected away, the cut surfaces ’ being indicated by parallel shading on the accompanying outline. The dissection exhibits the relations of the lateral and median vagal lobes and the commissural nuclei. For the internal connections of these structures compare Fig. 2 and the following figures of cross sections. Fic. 4. Transverse section through the medulla oblongata of Gadus morrhua. 35. The section is taken through the middle of the vagal lobes and illustrates the relations of their median (visceral) and lateral (somatic) lobules. Vagus root fibers are seen entering the median lobule on its lower border. Secondary facialis tracts pass from the lateral lobule to the ventral funiculi and from the median lobule to the ascending secondary gustatory tract. Dorsally of the latter is the fasciculus dorso-lateralis, containing the spinal V tract, tracts between the oblongata and the spinal cord and probably the tr. spino-cerebellaris. Fics. 4 to 8 are taken from a single series of transverse sections of an entire fish 7 cm. long, fixed in FLeMMING’s stronger fluid and stained by the method of We1cert. The external form of this young brain does not differ materially from that of the very large adult shown in Fig. 3. cerebe//um \ tubercu/um acust. lobus vagi lat Jobus vag/ med. SNUG. commIss. V/SC. rk.vagl sen tr secundus = f74) ae / (ey SaXew yaa fasc. dorso- lal... ayy / nuc. ambiguugr— 777 asc. sec. gusk¥r_ formatio\etic. / tr spino-tertalis-” — | funiculuy ventrali§--/) 84. Fournal of Comparative Neurology and Psychology. Fic. 5. Transverse section 0.4 mm. caudad of the last, passing through the caudal tip of the median division of the vagal lobe at the point where it passes over into the cephalic end of the visceral commis- sural nucleus. The caudal end of the lateral division of the vagal lobe may be seen two sections farther cephalad in the area here designated substantia gelatinosa. X 35. Fic. 6. Transection a little farther caudad through the beginning of the somatic portion of the com- missura infima. The vagal lobes and visceral commissural nucleus lie farther cephalad and the corre- sponding region is here occupied by the somatic commissural nucleus (cf. Fig. 3). The substantia gela- tinosa has begun to enlarge and a little farther caudad (Fig. 7) has expanded into the spinal V nucleus. Short tracts pass between the commissural nucleus and the substantia gelatinosa and formatio reticularis, and secondary tactile paths pass from all of these regions to the ventral cornu and, as internal arcuate fibers, into the ventral commissure. 35. Herrick, Brain of the Codfish. asc. sec. gust. tr.__ IX. Vag! Sensor —_ AW rx.vagl motor“ fasc. dorso-latZ—--- nuc. amblguus—-— formatio reticularis-~ tr spino-tectalis_-—~ funiculus ventra/srs—- som. commysSura pucleug, ae: commissufka /ntin ay : / Zs 86 Fournal of Comparative Neurology and Psychology. Fic. 7. Transection a short distance farther caudad. The area designated funicular nucleus con- tains in addition to that structure the somatic commissural nucleus and probably also a spinal prolonga- tion of the visceral sensory column, though the latter cannot be separately distinguished. The spinal V nucleus is so intimately related to the same area that no line of demarcation can be found between them. The whole area seems to function as a unit. The lateral funicular nucleus is not separately developed. The most cephalic ventral root fibers of the first spinal nerve appear in this section. SG Fic. 8. Transection farther caudad which passes through a dorsal and a ventral rootlet of the first spinal nerve. The cornu dorsalis is considerably larger than in the remainder of the spinal cord and still includes a vestige of the funicular nucleus complex and commissura infima. X 35. Herrick, Brain of the Codfish. funicular nuclfus__Zv spinal V nucleuge -/at. fasc. dorso sec.tactile tra 11. spinalis BGS 72 GaSe CHROMOTROPISM AND PHOTOTROPISM. Because of the obvious importance of the factswhich MINKIE- wicz claims to have discovered and of the stimulating value of his statements it has seemed worth while to print in this ‘fournal a translation of the two brief notes on responses to chromatic stim- ulation which he hzs recently published.t.| The whole of the first note appears below; in the case of the second note the introduc- tory paragraph is omitted in the translation.—THE EDITORS. Because of striking contradictions in the generally accepted theory of SacHs and Logs, to the effect that the most refrangible rays of the spectrum are alone active in the phototropism of animals and plants and that their action is the same as that of white light, and certain facts well established by P. Bert and Lussock for Daphnia, which is attracted especially by the yellow-green, and by WiEsNER for plants, which present two extremes of tropic action (first to the blue-violet, second to the infra-red, the action of the yellow being nil), I have given special attention in the course of my researches upon the tropisms and instinct to the tropic action of the chromatic rays. Some of my results follow: 1. The larve of Maia squinado (zoea) recently hatched present, as is well known, a strongly marked positive photo- and heliotropism. I have shown that they are at the same time very sensitive to chromatic rays, that they are directed constantly toward the rays of the shortest wave length, that is toward the violet, and in its absence toward the blue. They distinguish thus all the visible rays. The reaction is almost instantaneous; all the larvz dash like a flock of birds toward the most refrangible rays as soon as they are placed under their influence. This phenomenon has taken place not only in horizontal glass tubes but also in vertical ones no matter what the distance of the most tropic region from the surface of the water. 2. Nemerteans of the specias Lineus ruber behave in an entirely different man- ner. They are strongly negative in the presence of diffuse white light and at the same time they all direct themselves toward the rays of the greatest wave light, that is, toward the red, and in its absence, toward the yellow, etc. Thus far everything seems to agree with the theory of Lors. ‘The positive pho- totropism of the larve coincides with the strongest positive action (attractive) of the violet part of the spectrum; the negative phototropism of Lineus with the strong- est negative action (repellent) of the violet part. And yet these phenomena are not necessarily bound together. 1 Minxtewicz, Romautp. Sur le chromotropisme et son inversion artificielle. Comptes rendus de l'académie des sciences, Paris. Nov. 19, 1906. Le role des phénoménes chromotropiques dans étude des problemes biologiques et psycho-physiologiques. Comptes rendus de l’académie des sciences, Paris. Dec. 3, 1906. go Fournal of Comparative Neurology and Psychology. 3. One may bring about artificially with the nemerteans the inversion of the tropisms in the presence of chromatic rays while preserving the negative sense of their phototropism with reference to white light. a. Placing Lineus in a solution composed of from 25 to 80 cc. of distilled water to 100 cc. of sea water, I obtained on the following day this inversion with absolute certainty. Lineus while remaining negative with reference to white light now directs itself toward the most refrangible rays of the spectrum, just as it had previously directed itself away from them. The result of the inversion is that the phototropism, which remains negative, is here absolutely separated from the chromotropism, of which the sense is changed. Every chromatic ray has a specific action and at the same time the action of white light is not a simple resultant of a mechanical fusion of the actions of all the possible rays of the spectrum. I must remark further, that I have not as yet found, in spite of long continued researches, a single means of transforming the negative phototropism of Lineus into positive phototropism by agents either chemical, osmotic or thermic. ‘Thus, for example, the animal remains negative until its death in the presence of white light whatever the concentration of the sea water. b. ‘The inversion of the chromotropism of the nemerteans appears the second day, continues in general two days and disappears the fourth, the animal becoming again normally erythrotropic. This seems to me to prove that the nature of chro- motropism is not an absolute function of such orsuch vital medium but a function of the physiological state of the organism, a fact which agrees with the observa- tions of LoEB concerning the changes in heliotropism at different periods of life. c. There is one fact which confirms further this point of view, namely, that my Lineus after having lived for two or three weeks in my solutions and presenting consequently their normal chromotropism (erythrotropism) change anew when one transfers them into pure sea water and become purpurotropic (direct themselves toward the violet). But this is not all. d. The inversion of the chromotropism is not produced immediately and it also does not disappear all at once. There are stages when the animal still © erythrotropic (normal) ceases to distinguish green from yellow. ‘There are others when though indifferent to green and yellow it is already purpurotropic. ‘These stages of tropic blindness with reference to the middle parts of the spectrum last several hours and thus one can easily observe them on the second and the fourth day. ‘There should exist still two stages in the passage from erythrotropism to purpurotropism and inversely, during which the animal is completely indifferent in the presence of colored rays, that is, is achromotropic—either because it is equally influenced by all the chromatic rays or because it is entirely insensitive. ‘This I have not yet been able to observe, for these stages are of very short duration. In a second note the author points out the following bearings of his discovery upon the problems of general biology and of the psycho-physiology of vision. Minktewicz, Chromotropism and Phototropism gl A. GENERAL BIOLOGY. 1. One may find animals chromotropic with reference to the middle rays of the spectrum, each ray having its own specific action. Indeed Daphnia, according to Paut Bert and J. Lusgock, is xantho-chlorotropic, a fact which 1s absolutely incompatible with the theory of Lore (90). 2. Purpurotropism is not necessarily connected with positive phototropism nor erythrotropism with negative phototropism. One may find animals which behave in the reverse manner. This is proved by the excellent experiments of ENGELMANN upon unicellular organisms (1882 and 1883). According to him Paramecium bursaria, which is positively phototropic, avoided the green and the blue and directed itself toward the red. Likewise Navicula (a diatom) ceased movement in darkness and in the green. 3- Organisms should exist which, while being positively or negatively photo- tropic are not at all chromotropic (total tropic blindness; that is achromotropy), as Lineus is during the transitory stages of inversion. 4. There are organisms like the plants studied by J. WrESNER (’79) which are insensitive to certain rays of the spectrum (partial tropic blindness; that is, the axanthotropy of Vicia sativa). 5. It follows from the vertical tube experiments upon the zoea larve that the chromatic rays may have a certain influence upon the vertical distribution of aquatic animals, granting the unequal absorption of the different rays of the spectrum, the greatest for the red, the least for the blue (W. Sprine, H. For, E. Sarasin, ForEL). B. PSYCHO-PHYSIOLOGY, The experiments which I described in my preceding note lead me to believe that in studying the biology of the lower animals one should seek the indices which may point the way to an explanation of the complex and difficult phenomena of vision. 1. Chromotropism being independent of phototropism, it seems to me that the perception of white light may be a primitive phenomenon, much more simple than it is generally believed to be and independent of chromatic perception (this is corroborated by the well known experiments of A. CHARPENTIER, and also by the historic fact that in the best theory of vision, HERING was forced to admit the existence of a special white-black substance). 2. It seems superfluous and futile to seek for the solution of the problems of chromatic vision in the hypothetical creation of different nerve fibers (YoOuNG- HELMHOLTZ), of various pigment granules (A. Pizon), or of different chemical sub- stances (HERING), endowed with specific sensitiveness for different rays of the spectrum. One should rather ask whether there are not different physiological states in the same living substance which give rise to these complex phenomena of chromatic vision, as in Lineus the different states artificially induced bring about all the successive and transitory stages of chromotropism. [Experiments of PER- GENs and further of Lopato (’00) upon the chemical phenomena in the retina, the intensity of which varies according to the action of different rays of the spec- trum.] 92 Fournal of Comparative Neurology and Psychology. 3. Thusis it not possible and profitable to compare color-blindness (Daltonism), in general, and the different cases of achromotropy, partial or total, of the xantho- chloro- ; — tropic of xantho- Lineus in certain stages of the artificial inversion of its chromotropism ? [Personal experiments of W. NaGeEL (’o1) upon the artificial transitory blindness for red induced by santonin.] 4. Finally, is it not in this direction that one should look for the explanation of the consecutive colored images (six according to C. Hess (’00) or even more) after a short chromatic stimulation, if one bears in mind the succession of stages in the artificial inversion of chromotropism in my Lineus? tropic blindness of plants (WIESNER) and of the indifference LITERARY NOWICES: Jennings, H.S. Behavior of the Lower Organisms. Columbia University Biological Series, X. New York, The Macmillan Company. 1906. pp. xiv and 366. Price $3.00. The custom which the zodlogical faculty of Columbia University has followed for some years of having an annual series of special lectures on some particular topic in biology, given either by one of its own members or by some distinguished biologist from elsewhere, has indirectly resulted in the production of a very notable series of books. With a single exception these books have been written by men who by birth or adoption are American workers. ‘The series as a whole is unques- tionably representative of the best in American biological work. This being the case, it is altogether fitting that the volume here under consideration should have a place in the series. ‘The author’s brilliant investigations on the behavior of lower organisms, begun about ten years ago, have attracted wide attention, both here and abroad, and further can truly be said to have distinctly influenced the trend of Ameacan biological work in the period during which they have been appearing. In no small degree is the present activity in the field of animal behavior in this country a result of the stimulus of Dr. JENNINGS’ pioneer work. To characterize the present book in a single sentence it may be said to be a care- fully condensed statement from a unified standpoint of the important objective results which have been obtained by students of the behavior of lower organisms, together with a critical analysis and discussion of the significance of these results and their bearing on certain specific and general problems of biology. Roughly two-thirds of the space is given to the descriptive account of behavior, and the other third to discussion. ‘The descriptive portion of the work is not encyclopedic in character. Much which might have been inserted, and which doubtless would have been by a less careful and critical author, has been omitted. The reader is spared the mass of unimportant and trivial detail which is usually supposed to be the necessary accompaniment of the thorough and systematic development of a scientific subject. From this statement, however, it is not to be concluded that the treatment of behavior in the descriptive portion of JENNINGS’ book is in any way sketchy or lacking in thoroughness. Rather, the fact is that it is an unusually nice sense on the author’s part of what really 7s important both in his own work and the work of others, for a thorough general survey of the subject, which has kept the oe core eet eecley oar and va vial derail.) 4c the outstart it should be said that the scope of the work is not precisely indicated by the title. To have been exactly indicative of the contents the title should have read, “‘The Behavior of Certain Lower Organisms.” ‘The descriptive portions of the book include full and con- nected accounts of behavior of representatives of only two large groups of organ- isms; namely, (1) unicellular animals and plants (especially the bacteria), and (2) coelenterates. A brief account is given of certain features of the behavior in repre- sentatives of other invertebrate groups, notably the Echinodermata, Platyhel- minthes (Planaria) and the Rotifera. We are told in the preface that: “As origi- nally written, this descriptive portion of the work was more extensive, including, 4 Fournal of Comparative Neurology and Psychology. besides the behavior of the Protozoa and Ccelenterata, systematic accounts of behavior in Echinoderms, Rotifera, and the lower worms, together with a general chapter on the behavior of other invertebrates. “The work was planned to serve as a reference manual for the behavior of the groups treated. But the exigencies of space compelled the substitution of a chapter on some of the important features of behavior in other invertebrates for the systematic accounts of the three groups men- tioned.” Every student of behavior will regret the necessity for this omission. The descriptive portion of the book is subdivided as follows: Chapter I deals with the behavior of Amoeba and Chapter II with the behavior of bacteria. The next eight chapters are devoted to the infusoria, the forms on which the greater part of the author’s own investigations have been made. ‘The behavior of Paramecium is taken as the type and described in considerable detail. Chapter III includes a description of the normal movements of Paramecium, special emphasis being laid on the adaptive characters of the spiral path followed in the swimming. The mechan- ism of the “avoiding reaction,” by which term is designated the reaction originally called by the author the “motor reflex,” is fully described. A final section describes the method by which “positive” reactions are brought about. From this chapter as a foundation the author proceeds to a more detailed discussion of special features in the reactions of Paramecium to different stimuli, Chapter IV including the account of reactions to mechanical, chemical, thermal and photic stimuli and the orienting reactions to water currents, to gravity and to centrifugal force. Chapter V is devoted mainly to an excellent account of the reactions of Paramecium to electricity. It also includes a short section on the subject of trichocyst discharge. Chapter VI concludes the account of the behavior of Paramecium, and deals with some of the more general features. ‘The first section treats of the behavior under two or more stimuli, especial attention being paid to the interference of the contact reaction (thig- motaxis) with the reactions to other stimuli. Variability and modifiability of reac- tions are next considered. It is clearly demonstrated that the behavior is not of a fixed character, but subject to relatively wide modifications under certain circum- stances. ‘The behavior during fission and conjugation is briefly described. Follow- ing this there is given a very interesting “composite reconstruction” of the daily life of Paramecium. Finally we have a brief discussion of certain general features of the behavior in this organism. The next four chapters deal with behavior of other infusoria according to the same general plan as is followed in the account of Paramecium, but of course with less detail as to individual organisms. An account of the “‘action systems” of flagellates, holotrichous, hypotrichous and heterotri- chous ciliates is followed by descriptions of the reactions of representatives of these groups to mechanical, chemical and thermal stimuli. The reactions of infusoria to light receive very full treatment. The general conclusion is that “reactions to light occur in the infusoria in essentially the same way as do the reactions to most other stimuli through the avoiding reaction; that is, by the method of trying move- ments in different directions. The cause of reaction is a change in the intensity of light, primarily that affecting the sensitive anterior end.” ‘There is a brief account of the reactions of infusoria to gravity and centrifugal force. Chapter IX is devoted to an account of reactions to the electric current, and a critical discussion of the various theories of the electrotactic reaction which have been proposed. None of these is found to be entirely satisfactory: Chapter X completes the account of the behavior of infusoria: the topics discussed in it are modifiability of behavior, behay- Literary Notices. 95 ior under natural conditions and food habits. The next two chapters are given to the behavior of lower Metazoa. The behavior of ccelenterates receives very full treatment. Summing up his conclusions from the study of behavior in these relatively simple multicellular organisms the author says: ‘‘Comparing the behavior of this low group of multicellular animals with that of the protozoa, we find no radi- cal difference between the two. In the ccelenterates there are certain cells—the nerve cells—in which the physiological changes accompanying and conditioning behavior are specially pronounced, but this produces no essential difference in the character of the behavior itself.” The treatment of other metazoan forms occupies but a single chapter and is topical rather than systematic in form. ‘This brings us to the end of the descriptive portion of the book. The objective point of view is very carefully and consistently maintained in. these descriptive chapters. Though treating of a variety of topics, they have never- theless a very definite unity. This comes about through the marked emphasis which the author puts upon certain features which he has found to be common to the behavior of all the lower organisms so far investigated. Of these common fea- tures the following are the ones on which greatest stress is laid: 1. Organismschange their behavior(i. ¢., react) in response to changes in exter- nal conditions. ‘‘The most general external cause of a reaction is a change in the conditions affecting the organism.” 2. The character of the behavior of an organism under any particular set of external conditions is not determined solely by those external conditions but to as great or even to a greater degree by internal conditions (the “physiological state”’ of the organism). 3. The behavior of all organisms is, generally speaking, of a markedly adap- tive character. 4. The most usual mechanism by which this adaptiveness in behavior is brought about is that of the “‘trial and error” method of response to stimuli. “We find behavior largely based on the process of performing continued or varied move- ments which subject the organism to different conditions of the environment, with selection of some and rejection of others.” ‘ 5. Behavior is in general not a fixed or stereotyped character, but instead is capable of varying degrees of modification under different circumstances. On the whole modifications so induced are of an adaptive character. In the opinion of the reviewer the clear and conclusive demonstration that the features enumerated are common to the behavior of all the lower organisms hitherto investigated in detail is a great achievement of the book. ‘The thing most lacking in animal behavior work hitherto, has been a unified standpoint leading to the elucidation of the features of behavior which were general (that is, common to a wide range of organisms). Turning now to the theoretical portion of the book, we have first a short chapter (XIII) comparing the behavior of unicellular and multicellular organisms. ‘The author’s general conclusion is that there are no differences of fundamental character in the behavior of the Protozoa and Metazoa. He is strongly inclined towards the view as to the function of the nervous system which has been upheld by Logs; namely, that it has no exclusive functions not primitively common to all protoplasm. The next chapter is devoted to a convincing destructive criticism of the “tropism theory.” The author’s views on this subject are so well known that special com- 96 Fournal of Comparative Neurology and Psychology. ment regarding them is unnecessary. How any person of a scientific habit of mind can still uphold the general validity of the local action theory of tropisms in the face of the facts which have been brought out by JENNINGs showing just how uni- cellular organisms actually do react to “directive” stimuli, passes all understanding. In this chapter a brief section is given to a discussion of the various systems of ter- minology and nomenclature which have been devised for use in animal behavior work. ‘The following sentences demand quotation: ‘‘o the present writer, after a long continued attempt to use some of the systems of nomenclature devised, descriptions of the facts of behavior in the simplest language possible seems a great gain for clear thinking and unambiguous expression. If investigators on the lower organisms would for a considerable time devote themselves to giving in such simple terms a full account of behavior in all its details, paying special attention to the effect of the movements performed on the relationof the organism to the stimulating agent, this would be a great gain for our understanding of the real nature of behav- ior and some theories now maintained would quickly disappear. Less attention to nomenclature and definitions, and more to the study of organisms as units, 1n their relation to the environment, is at the present time the great need in the study of behavior in lower organisms.” Who does not heartily agree ? Chapter XV asks the question: ‘“‘Is the behavior of lower organisms composed of reflexes?” The answer may be gathered from the following sentences: ‘The behavior of Paramecium and the sea-urchin is reflex if the behavior of the dog and of man is reflex; objective evidence does not indicate that there is from this point of view any fundamental difference in the cases.” The next three chapters contain a searching analysis of the phenomena of behavior in lower organisms. ‘The general result of this analysis is summed up by the author as follows: “The three most significant features of behavior appear to be (1) the determination of the nature of reactions bythe relation of external conditions to the internal physiological processes, and particularly the general principle that interference with these processes causes a change in behavior; (2) reaction by varied and overproduced movements, with selection from the varied conditions resulting from these movements—or, in brief, reaction by selection of overproduced movements; (3) the law of the readier resolu- tion of physiological states after repetition. The first of these phenomena pro- duces the regulatory character of behavior. The second and third furnish the mainsprings for the development of behavior, the second being constructive, the third conservative.”’ In Chapter XIX are set forth the author’s views regarding the development of behavior. Space is lacking for a full consideration of the argument on this subject. The principal factors which make for progressive development of more effective (i.e., adaptive) behavior in the individual are held to be (1) the selection of varied movements as a general method of behavior, and (2) ‘‘the law in accordance with which the resolution of one physiological state into another becomes readier and more rapid through repetition.” It is shown in detail how these factors might lead in several ways to progressive development in behavior. To account for progressive evolution of the race inrespect to behavior the author, after rejecting the inherit- ance of acquired characters as unproven, falls back on the principle of natural selection operating according to the method which has been called “‘organic selec- tion.’ In reading over this section one cannot escape the feeling that the author only adopts natural selection to account for race progress in behavior because there Literary Notices. 97 is no other general principle even formally adequate at hand, and further that he is quite cognizant of the fact that in our present state of knowledge of the phenom- ena of behavior the explanation of evolution in this field given by natural selection is a purely formal and artificial one. Chapter XX deals with the question of consciousness in lower organisms in a highly interesting way. ‘The conclusion reached is that: ‘All that experiment and observation can do is to show us whether the behavior of lower organisms 1s objec- tively similar to the behavior that in man is accompanied by consciousness. _ If this question is answered in the affirmative as the facts seem to require, and if we further hold, as is commonly held, that man and the lower organisms are subdivisions of the same substance, then it may perhaps be said that objective investigation is as favorable to the view of the general distribution of consciousness throughout ani- mals as it could well be. But the problem as to the actual existence of conscious- ness outside of the self is an indeterminate one; no increase of objective knowl- edge can ever solve it.”’ The final chapter has for its heading “ Behavior as regulation, and regulation in other fields.” ‘The importance as a general regulatory principle of the selection from a variety of activities, those lines of activity which lead to a minimum inter- ference with the physiological processes of the organism, together with the preserva- tion or fixation of adaptive activities through the law of the readier resolution of physiological states, is emphasized. ‘This chapter has been published elsewhere in essentially its present form. A bibliography and index complete the volume. ‘The book is very completely and well illustrated. On the whole, one finds very little in the book to criticise. Undoubtedly there are many who will not fully agree with JENNINGs in some of his interpretations and conclusions, but technical discussion of views opposed to those of an author falls properly within the scope of the special memoir, not that of the review. He who searches this book for errors in statement of fact or of principle, or for evidence of deficient knowledge of what has previously been done in the field covered, or for slips in logic or diction, will spend his time fruitlessly. One fault of omission should be corrected in a later edition. “Throughout the book there is no indication of the scale to which the figures are drawn, and hence there is no way for one not already familiar with the organism discussed to know anything about their relative sizes. Of course this matters not at all to the biologist, but as things stand at present, the “Jay” reader of the book must inevitably go away with the impression that Chilo- monas is a veritable giant among Protozoa altogether surpassing in size the mediocre, not to say diminutive, Bursaria. ‘This, however, is a matter of detail. In general one has only praise for the book. It is a contribution of a high order of merit to biological literature. It is valuable to the specialist as a careful and thorough sum- mary and digest of the present state of knowledge in the field of which it treats, and as a unified setting forth of the author’s matured opinions regarding the broader aspects of the problems of animal behavior. ‘To the general reader it presents an authoritative, clear and most interestingly written account of a side of natural his- tory which has hitherto, for the most part, lain entirely outside his ken. We heartily wish for it the large circulation which it certainly deserves. RAYMOND PEARL. 98 Journal of Comparative Neurology and Psychology. BOOKS AND PAMPHLETS RECEIVED. Baldwin, James Mark. Mental development in the child and the race. Methods and processes. Third Edition, Revised. New York, The Macmillan Co. 1906. Grasset, J. Demifous et demiresponsables. Paris, F. Alcan, Editeur. 298 pp. 1907. Ingegnieros, J. Le langage musical et ses troubles hystériques. Etudes de psychologie clinique. Paris, F. Alcan, Editeur. 208 pp. 1907. Guyer, Michael F. Animal micrology. Practical exercises in microscopical methods. The Univ- ersity of Chicago Press. 240pp. 1906. Maxwell, S. S. Chemical stimulation of the motor areas of the cerebral hemispheres. Reprinted from the fournal of Biological Chemistry, Vol. 2, No.3. 1906. Yagita, K. Ueber die Verinderung der Medulla oblongata nach einseitiger Zerstérung des Strick- k6rpers nebst einem Beitrag zur Anatomie des Seitenstrangkernes. Reprinted from Okayama- Igakkwai-Zasshi (Mitteilungen der medixinischen Gesellschaft zu Okayama), No. 201. 1906. Hrdlicka, A. Anatomical observations on a collection of orang skulls from western Borneo; with a bibliography. Reprinted from the Proc. U. S. Nat. Mus., Washington, Vol. 31. 1906. Weysse, A. W.and Burgess, W.S. Histogenesis of the retina. Reprinted from the Am. Nat., Vol. 40, No. 477. 1906. Smith, Grant. The eyes of certain pulmonate gasteropods, with special reference to the neurofibrille of Limax maximus. From the Bul. Mus. Comp. Zool., Harvard College, Vol. 48, No.3. 1906. Harrison, Ross Granville. Further experiments on the development of peripheral nerves. Reprinted from dm. Fourn. Anat., Vol. 5,No.2. 1906. Dean, Bashford. Chimaroid fishes and their development. Publications of the Carnegie Institution, No. 32. 1906. Drew,G.A. The habits, anatomy and embryology of the giant scallop (Pecten tenuicostatus Mighels). University of Maine Studies, No. 6. 1906. Metcalf, M. M. Salpa and the phylogeny of the eyes of vertebrates. Reprinted from Anat. Anz., Vol. 29. 1906. Carpenter, F.W. An astronomical determination of the heights of birds during nocturnal migration. From The Auk, Vol. 23, No.2. 1906. Montgomery, Thos. H. The esthetic element in scientific thought. Reprinted from the Academy of Science, Vol. 8. 1905. Bell, J. Carleton. Reactions of the crayfish. Reprinted from Harvard Psychological Studies, Vol. 2. 1906. Smallwood, W.M. Notes on Branchiobdella. Reprinted from the Biol. Bul., Vol.11, No.2. 1906. Wilder, B. G. Some linguistic principles and questions involved in the simplification of the nomen- clature of the brain. Reprinted from the Proc. Am. Physiological Assoc., Vol. 36. 1905. Weysse, Arthur W. Notes on Animal behavior. Science, N. S., Vol. 19, No. 495. 1904. Lindsay and Blakiston. The physician’s visiting list for 1907. Philadelphia, P. Blakiston’s Son & Co. Thirteenth Annual Report of the Craig Colony for Epileptics at Soneya, N. Y. The Journal of Comparative Neurology and Psychology VotumE XVII MARCH, 1907 NUMBER 2 LIGHT REACTIONS IN LOWER ORGANISMS. ik. VOLVOX GLOBATOR: BY SO), IMAI. Professor of Biological Sctence at Hope College, Holland, Michigan. Wirth Firreen Ficures. CONTENTS. ise Ibetagaxelorqaloynl 6.5 dersic.4 oO OED SEO C a ROEM OEAC AAR Tacos ae Codon oanndqsdag Ors au or 99 25) AN[ETRICALISER GOIN”) Coen one ad senna dae Hac eM eBnBOnow Ee Sambo ccOuCK sao eno Ao AO UUGOR Sou ue 101 Gio. “NianllGulids Gigs OCU BOAO eC Coe ao aee aromas Gene ono unin ccp.onCoaccodoor Usooh 102 Ame EUINGtOnS OL Esye=spO bitter sea tarsi 2) sha) els ce lel cle ialekehe te ot aeb enero Nel aNcls ak-eae- Ay halen Col netFelaee 108 fo. GENOA « vo.cuic OURS Uo dao tM aeRO MEE Oe nro om pod bb op ooud modo veme SDs 112 Ga Onientation=— General DisGUSSIOMiA 2 tele seo e = aos © ely sieee eles tol men etene era ee Tanna a mele nteRees eect eters 115 a. Effect of Internal Factors on Orientation... ..........0- cece eee teste cette eee 119 b. Effect of Change in Light Intensity on Orientation .......- 20+ s cece eee eee e eee I21 c. Effect of Gravitation on Orientation. ..... 0.0.10. eevee tet ene 122 d. Effect of Contact Stimuli and Rotation on Orientation .......-.+++0ee eee veer eee eee 128 7- Orientation to Light from two Sources. .........-0 +02 seer eee eee eee ere eet eee eee 131 8. Orientation in Light Graded in Intensity.........- 22. 2+ sso teeter tee eee eee eee 136 Om Onientatiom of Sep ments ter ari o-ion a)~ oe = leat tae ees ete en a Rodeos ye 144, TOs Mechanicsion Orientation merece tarts ser so-y= 15 i= siaisiege dened ste eteta Neder ieee eres iter eet tet Ne 145 Hip NeactlomoniNerative COlomesamier ts st. iley- sri «le che spcaeteveestal=ehete eed eve te tehed Reto eh heard sastaner 1 154 12 Catise of Change im Sense Of REACHIOW «1. 4¢ oa «ier elem oe acetate iets hts cielo Ye dad= apr 157 13. Effect of Temperature on Change in Sense of Reaction ..........+--. 0s sees e eee eee eens 161 14. Effect of Mechanical Stimuli on Change in Sense of Reaction...............-+-+ +e esse sees 162 Hiya, “INorCesovol lie: cain cobosenn oo uae MOCoOUpASOmortoouTGtadcoc00500 onounpenaoovoocouOmDnne 163 iy ‘Ohare occu duh neosouonaee obo s AaneoebunoSOT ooo can ACodcsAGetanconssabcouc 164 17. Reactions on Reaching the Optimum in a Field of Light Graded in Intensity ............... 166 Tithe, Wiis Cen oA Sos Son aR OMA eB eOaOnane oooh ocoGna DU GudcommmocoooDeROTdey 171 iG)” SUUNIEISTG oc aBb ode cuud old Adee OR ED DA ete doo pre rmod.co cop ccutadnoneLyecédenaqcooobude 176 Aer, slo Hofar ooh2 ace gin bn ane sods pee ne ASME or Oon s aonddunbouccRmoGes op uooEd OoouobaSoD Se 179 I. INTRODUCTION. Since the discovery of Volvox by LEEUWENHOEK, over two hundred years ago, it has been studied in detail by many investi- gators. Nearly all noted the effect of light on the direction of 102 Journal of Comparative Neurology and Psychology. smaller ones, especially such as were still within the mother colo- nies, appeared quite normal in color. Intense light evidently causes some change in the chlorophyl. The specimens used in the experiments performed at Harvard University were collected in various small ponds located some little distance west of Cambridge. Some of these ponds are arti- ficial, having formed in clay pits; others are apparently natural, being located in low, swampy land. All of the ponds contained numerous aquatic plants, and the water in them was stagnant but clear and not foul. The material used in the work done at Hope College was collected in ponds connected with a very slug- gish river which runs through a marsh directly north of the city of Holland. Colonies of Volvox were found sparsely scattered here and there along almost the entire shore line of nearly all the ponds. Ina few spots, however, they were so numerous that the water appeared green, and in these places they could readily be collected in great numbers. There are two well defined species of Volvox, globator and aureus (KHRENBERG =minor STEIN). In the ponds near Cam- bridge practically all the colonies belonged to the species globator; but in the ponds north of Holland the two species were found about equal in number. ‘They were usually found intermingled, but in a few places I found only globator and in one place nothing but minor. After colonies of Volvox have been in the laboratory from 12 to 24 hours they become inactive, and no longer respond readily to stimuli, and are therefore not satisfactory for experimental work. This makes it necessary to collect frequently. An abundance of material close at hand is consequently almost a requisite for experimental work on this form. In the following experiments, the specimens usually were collected early in the morning and used the same day. 3- STRUCTURE. Since the discovery of Volvox by LEEUWENHOEK nearly every naturalist has had something to do with the study of this exceed- ingly interesting organism. Most of these investigators laid great- est stress on the structure, but in spite of all this work there are still two questions with regard to structure, concerning which there is some doubt. One 1s the location of the eye-spot with reference Mast, Light Reactions in Lower Organisms. 103 to the colony as a whole, the other, the variation in form of the vital portion of the individuals composing the colony. Since these structures are of considerable importance in the study of light reactions, I shall take up the structure of Volvox rather more in detail than otherwise would be necessary. ‘The following descrip- tion is the result of a review of the literature on this subject, sup- plemented by my own observations. Volvox varies in form from approximately spherical to ovoid. The smallest free swimming colonies can scarcely be seen with the naked eye, while the largest are nearly, if not quite, one millimeter in diameter; KLEIN (’89, p. 143) gives 850n, HANscIRG (’88, p. 101) 8004, KIRSCHNER en 700, and Focke (’47) 11004. Some of the investigators found Volvox globator to be larger than Volvox minor, while others found the opposite to be true. KLEIN gives 800y as the diameter of the largest colonies of V. globator and 850y as that of the largest V. minor. HanscirG gives 800n as the diameter of the fone? ga 460pn as that of the latter. In my own collections I found V. globator in general much larger than V. minor. I did not, hemes make any accurate measurements with reference to this point. The colonies of both species are composed of numerous individ- uals, each of which consists of one cell. KLEIN (88, p. 146) found from 200 to 4400 individuals in various colonies of V. minor and from 1500 to 22,000 in V. globator. The individuals consist of a central portion, composed largely of protoplasm, and a thick hya- line layer which surrounds the central portion. ‘The central por- tion will be referred to asthe zooid in the future description. The hyaline layers of contiguous cells usually appear continuous, one with the other, but WiLi1aMs (53) demonstrated that they are limited by cell walls. Iwas not able to see these in living colonies of V. minor, but could see them very distinctly in a few spore-bearing colonies of V. globator, especially at the anterior end. ‘The hyaline layer is much thicker in V. minor than in V. globator and the zooids are much more nearly spherical in the former than in the latter, in which they are in general quite angular. ‘The difference in the shape of the zooids forms the chief distinguishing characteristic of the two species. The cells in the colonies are arranged side by side so as to form a wall enclosing a cavity. In V. minor the hyaline layer is figured by MEYER (95, p. 227) as extending nearly to the middle of the colony, thus leaving only a very small central 106 = -fournal of Comparative Neurology and Psychology. the surface, and a conical portion which projects from near the middle of the flattened portion into the hyaline layer almost to the surface of the colony. At a point about halfway between the anterior pole and the equator of the colony, the altitude of the projection is about twice as great as the diameter of its base. ‘The ratio between these dimensions becomes gradually greater as one proceeds farther from the anterior end, until at the posterior end | the altitude is four to five times as great as the diameter. It will thus be seen that the distal end of the conical projections gradually extends farther out as one proceeds from the anterior end to the posterior (see Fig. 2). The only reference to the variation in Fic. 3. View of the anterior end of a colony of Volvox minor, showing the location of the eye-spots. Z, zooids; e, eye-spots; p, protoplasmic fibers connecting the zooids. form of zooids in the same colony is found in OverrTon’s article (89, p. 70), and he states only that the projection (Schnabel) is longer in the neighborhood of injured places (“In der Nahe von verletzten Stellen verlanget sich der Schnabel’). The projection is nearly circular in outline at the base, but it becomes considerably flattened toward the distal end, so that a cross section near this end is elliptical in outline. ‘The zooids are so arranged in the colony that one of the flattened surfaces of the projections faces the anterior end and the other the posterior. Viewed from either of the flattened surfaces, the outline of the distal end forms nearly a straight line, at either end of which is found the attachment of a flagellum. ‘The flagella are five or S1X times as long as the diameter of the zooids. OveERTON (’89, p. 72) Mast, Light Reactions in Lower Organisms. 107 says they are about 3 apart and 25y long. ‘The eye-spot is situ- ated on the surface of the projection which faces the posterior end of the colony. It is found but a short distance from the free end, between the points of attachment of the flagella. If the projections are short or absent and the zooids nearly spherical the eye-spots are still located in the same relative position as they are in zooids containing long projections, 1. ¢., they face the posterior end of the colony. Tie becomes very evident in viewing a colony from the anterior end. Under such conditions it is clearly seen that the eye-spots are situated on the surface of the zooids farthest from the middle of the anterior end, as repre- sented in Fig. 3. Nearly all the investigators, who have worked on the structure of Volvox, figure the eye-spot as situated on one side of the zooids near the outer surface, but only one, OveERTON, describes and figures it in such a way that its position with reference to the colony as a Whole 1s made clear. Overton (’89) in Taf. 4, Fig. 26, and Taf. 1, Fig. 3, clearly represents the eye-spot as being located near the outer anterior surface of the zooids andsays,p.114: “Sehr bemerkenswerth erscheint, dass, wie bei einstellung auf einen Meridiankreis des Volvox Stockes sich ergibt, die Augen flecke (wenigstens bei V. minor) bei allen Zellen derjenigen Seite anlie- gen, die dem vorderen Pole am nahsten liegt.” During the first few days in August, 1905, I examined 30 speci- mens of V. globator and 50 of V. minor, with special reference to the location of the eye-spots and found that in all but one of these, they were unquestionably located on the outer posterior surface of the zooids. Furthermore, I gave the problem of locating this structure to three of my pa sGkea in October, 1905. “These stu- dents had never seen Volvox before and knew nothing about any work done on it. All of them concluded that the eye-spots face the posterior end of the colony. When they took up the problem they knew that these organisms are usually positive in their light reactions. I had given them the term, eye-spot, and it was clearly evident that they assumed that this structure functioned in direct- ing the organisms toward the light, and consequently expected to find it on the anterior surface of the zooids, for they were all much surprised to find it on the opposite surface. It is, therefore, safe to conclude that OVERTON’s observation was wrong. The eye-spots in Volvox are brownish in color and lenticular in I10 Fournal of Comparative Neurology and Psychology. all phototactic under others. Davenport (’97, p. 188) proved certain species of Amceba to be negative to ieee and it is well known that Stentor coeruleus responds very definitely to stimula- tion by light. It is said that the Chytridium swarm spores have an orange colored oil globule at the base of the flagellum which may function as an eye-spot, but in the four organisms mentioned last there are no structures which appear as though they could take the place of these organs. It is, therefore, evident that we have organisms without eye-spots which are sensitive to light but as far as I know there are none with these structures that are not sensitive. 4. WaceER (oo, Pl. 32, Fig. 2) represents the flagellum in Euglena viridis as indirectly connected with the | eye-spot, in that it has an enlargement which lies immediately over the concave surface of this structure as represented in Fig. 4. Ihe eye-spot is supposed to absorb the blue of the spectrum and in some way to stimulate the enlargement on the fla- gellum. 5. The fact that the eye-spots are larger and more highly colored at the anterior end of Volvox than at the posterior, that _Fic. 4. Side view of antee they lose their color and become smaller rior end of Euglena viridis, after =. : WaGen; @cyeapat f dae in the absemeevon lieht, andethat theysane Re ae enlargementinflagel- situated near the distal end of projections um; c.v., contractile vacuole. ; which become longer as one proceeds from the anterior end toward the posterior, aad thus expose the eye- spots to more light, indicates that these structures function in light reactions. In view of the evidences presented above in favor of considering the eye-spot as a light recipient organ, and in view of the fact that there is nothing in the structure or location which indicates that it could not function in light reactions or that it has any other function, it appears safe to conclude that EHRENBERG’S idea with reference to the function of the eye-spot is correct. WacER (’00) suggests three ways in which the eye-spot in Euglena viridis may function in light reactions: (1) It mayabsorb light and thus produce a change in the movement of the flagellum; (2) it may merely prevent the light rays from reaching one side of Mast, Light Reactions in Lower Organisms. III the enlargement in the flagellum, while the other side is exposed, and thus produce a difference in light intensity on opposite sides of the enlargement; or (3) it may cut off the light from one side of the sensitive portion of the anterior end of the organism when it 1s not oriented, and thus produce unequal illumination on opposite sides of this end. It seems impossible to test the suggestions of WAGER experi- mentally, but it may be possible to arrive at a tentative conclusion concerning the matter, from what we know about the structures and reactions of these organisms. JENNINGS (’04, p. 54) shows that Euglena swims in a spiral course with the larger lip constantly farthest from the center of the spiral. In thus swimming the longitudinal axis never points toward the source of greatest illumi- nation, so that when the organism is oriented the side of the ante- rior end containing the larger lip is always more shaded by the eye-spot than that containing the smaller lip (see Fig. 4). From this it seems evident that the eye-spot in Euglena does not func- tion in accordance with WaGEr’s third suggestion. ‘That it does not function in accordance with this suggestion in Volvox 1s still more clearly evident, for here the eye-spots are located near the posterior surface in the projection of the zooid, so that if this pro- jection 1s sensitive to light there certainly is no possibility of oppo- site sides being equally stimulated when the organism 1s oriented, for under such conditions the shadow of the pigment granule falls on the posterior surfaces while the anterior surface 1s fully exposed to the light. With reference to the second suggestion, it is prob- ably true that the eye-spot does prevent the light from reaching one side of the enlargement in the flagellum in Euglena, but by referring to Fig. 4 it will be seen that the difference in light inten- sity thus produced on opposite sides of the enlargement must be practically the same when the light strikes the organism nearly parallel with the longitudinal axis, as it 1s when it strikes it an at angle from the side containing the smaller lip. If this be true, there 1s no change in stimulation when the organism 1s slightly thrown out of orientation. It therefore does not seem probable that the eye-spot in Euglena functions in accordance with the second suggestion of WacER. In Volvox we know of no enlarge- ment in the flagellum such as that found in Euglena, and if there were one, or some other similar structure, the criticism offered above with reference to Euglena would hold here also. 7 a ne Py ie tee Ae Jolt 114. ‘Fournal of Comparative Neurology and Psychology. in the direction of rotation is caused by contact stimuli at all it must be by contact stimuli along the sides of the colonies. Volvox colonies were subjected to such stimuli by laying a glass slide into an aquarium containing filtered water about 3 mm. deep, so that the edge of the slide made an angle of about 45 degrees with the rays nee light. When the colonies moved toward the source of light and came in contact with the slide, the point of contact was not at the anterior end but some little distance from it. After being thus stimulated they immediately turned from the slide making an angle of about 95 degrees with their previous course. Then they gradually turned toward the source of light again and thus continued along the edge of the slide making a zig- zag path. In following along the edge in this way they frequently came in contact with the lide before they were perfectly oriented and were consequently stimulated at a point further from the anterior end than usual, sometimes about midway between the two ends. In all these reactions the direction of rotation was seldom changed. It is therefore clear that a single contact stimulus on the aide of a colony, which does not obstruct forward progress, does not cause reversal in the direction of rotation. In the experi- ment just referred to a small portion of one of the upper corners of the slide was slivered off, making an incline on which the water became gradually more shallow onal: at the upper end, it was not deep enough for the larger colonies to swim without difficulty. As the colonies worked up this incline, they came in close contact with the glass and the direction of rotation was frequently changed. It may then be concluded that continuous contact stimulation on the sides causes reversal in the direction of rotation, providing the contact is such that considerable resistance is offered to for- ward motion. But why should contact stimuli on the anterior end, which pre- vents forward motion, not cause reversal as well as similar stimuli along the sides? Considering the structure of the organism in question, it seems probable that rotation is brought about largely by an oblique stroke of the cilia along the side and that those at the ends have little if anything to do with it. Now it seems reason- able to assume that when a certain proportion of these cilia on the sides meet considerable resistance they all strike in the opposite direction and thus produce reversal of rotation. When the ante- rior end is in contact with an object the cilia along the sides are of Mast, Light Reactions in Lower Organisms. 115 course free, and if it is the action of these cilia which causes rota- tion we should not expect a change in the direction of rotation when the anterior end is stimulated. As stated above, we find reversal in the direction of rotation frequent in water containing numerous small particles. What is the cause of this? ‘This is probably due to particles becoming entangled in the cilia and obstructing their free movement, thus causing a change in the direction of rotation. While we have thus found that reversal of rotation is largely caused by external agents, it is unquestionably true that it depends to some extent upon the condition of the organism itself, for under similar external conditions difference in the frequency of reversal was repeatedly noted. 6. ORIENTATION—GENERAL DISCUSSION. It is well known that if Volvox is subjected to light of moderate intensity, it swims toward the source of light; but if the light 1 inten- sity is high, as e. g., direct sunlight, it aaecleg in the opposite direc- tion. In casually studying such movements it appears as if the course of the colonies in either direction were nearly parallel with the light rays, and investigators have, in general, assumed this to be true. HoLmes (03, p. 320), writes: “It 1s easy to deter- mine that Volvox orients eelt and that very accurately, to the direction of the rays of light. If specimens of Volvox are taken into a dark room and exposed to the light from an are lamp they travel towards the light in almost a straight course, swervin remarkably little to the one side or the other. ‘They will often travel a foot without deviating as much as a quarter of an inch from a perfectly straight course.” In studying the effect on the direction of movement, of difference in light intensity on opposite sides of a Volvox colony, I accident- ally Riccavered that, contrary to Homes’ conclusion, Volvox very seldom orients “accurately to the direction of the rays.” “The colonies do, of course, swim toward or from the source of light in a general way; but movement parallel with the rays is quite the exception. In swimming toward a source of light the colonies may deflect not only to ‘ie right or left but also up or down. De- flection up or down will be discussed under the effect of gravi- tation on orientation (p. 122); deflection to either side will be taken up in connection with the description of the following experiments. 116 Fournal of Comparative Neurology and Psychology. Most of these experiments were performed in an apparatus which I have called a “light grader.” I have given a detailed descrip- tion of this apparatus in another paper (Mast 06, p. 364). The Fic. 5. Fic. 5. A vertical section of the light grader. The lens (a) which is a segment of a cylinder has its longitudinal axis lying in the plane of the section; b, stage; c, Nernst glower; d, non-reflecting back- ground; e, mirror; f, light rays; g, opaque screens. Distance from glower of lamp to stage, one meter. Fic. 6. Stereographic view of light, lens, and image; a, lens; b, field of light produced by the image of the glower (c); d, opaque screen which lies flat on lens and contains a triangular opening which causes a gradation in the light intensity of the field (b). important features of the apparatus will be readily understood, however, by referring to the accompanying figure. Mast, Light Reactions in Lower Organisms. rt 7 An aquarium! 15 cm. long, 8 cm. wide and 8 cm. deep inside was placed at one of the principal foci in the light grader, which was in a horizontal position and so arranged that the Nernst glower at the other principal focal point was vertical. The light rays passed through the aquarium practically parallel with each other and the bottom of the aquarium and perpendicular to the sides. An opaque light screen, containing a rectangular slit 10 mm. wide and 13 cm. long, was fastened to the side of the aquarium nearest the light, so that the lower edge of the slit was on a level with the bottom and the ends of the slit were 1 cm. from either end of the aquarium. Filtered water was poured into the aqua- rium to such a depth that its surface was above the upper edge of the slit and was consequently in darkness. Since the rays were practically parallel with the bottom and perpendicular to the sides of the aquarium, it 1s evident that reflection from the water surfaces was practically eliminated. ‘The light which passed through the aquarium was largely absorbed by the wall of the dark room which was over seven meters from the light grader, and since this was the only light which entered the room it 1s safe to conclude that the direction of movement of Volvox in the aqua- rium was influenced only by rays direct from the Nernst glower. By placing an opaque screen containing a triangular opening over the cylindrical lens in the light grader, a field of light 1s produced which becomes gradually less intense from one end to the other (Mast ’06, p. 364). If Volvox is allowed to swim toward the source of light in such a field it is evident that one side of the colonies will be more strongly illuminated than the opposite, and if difference in light intensity on the two sides, regardless of . ray-direction, determines the direction of movement we should expect the organisms to move at an angle with the direction of the rays of light. This was found to be true, as will be shown later (pp. 136-141). e first series of experiments made to ascertain the effect of difference in intensity on orientation was performed in the light grader, arranged as described above, by carefully introducing about one hundred colonies into the aquarium at a fixed point J ' The aquarium was made of the best plate glass obtainable, accurately cut and ground, and cemented with Canada balsam boiled in sufficient linseed oil to give it the desired consistency. This cement proved very satisfactory, much more so than any other of several tried. Balsam in xylol is good but it becomes so brittle on drying that it breaks readily. Linseed oil prevents this. 118 ‘fournal of Comparative Neurology and Psychology. near the side farthest from the light. In some conditions, the colonies thus introduced, proceed across the aquarium nearly parallel with each other, spreading but little, frequently not more than five or six millimeters. In other conditions, however, they spread out as much as three or four centimeters. By laying a straight wire on the aquarium and constantly keeping it over the middle of the group of colonies as they proceeded on their way, the average course was quite accurately ascertained. ‘The course under most conditions, although at a decided angle with the rays, was remarkably straight; but under some conditions it curved con- siderably as the organisms approached the side of the aquarium nearest the light. [he paths produced and the direction of the rays were transferred to paper by means of a miter square; and thus the angle of deflection was recorded for future reference. Between August 20 and 30, 1904, seventy-three paths observed under the conditions described above were recorded, and many more were observed. In nearly all of these cases the colonies deflected strongly to the left, frequently making an angle of 45 degrees with the light rays, and rarely less than 5 degrees. This deflection to «he left was brought out in a striking way, by putting the Volvox colonies into the aquarium a few centimeters to the right of the left edge of the light area. When introduced at this point, they soon reached the plane between light and shadow and passed into the dark area without any apparent change in their course. After they had traveled in the dark region some little dis- tance they rose, deflected more sharply to the left, and frequently made a small circle or spiral and entered the light region again. This deflection to the left in the light area was, of course, thought to be due to the fact that the light was graded in intensity, the more intense end of the field being to the right. It was accident- ally discovered, however, that similar deflections were produced when no lens was used, and later it was found that when the screen over the lens was inverted so as to make the left end of the field the more intense instead of the right as in the previous experiment, the Volvox colonies still deflected to the left. It was therefore clear that this deflection was not due to difference in light intensity on opposite sides of the organisms. One hundred and one additional paths were observed and recorded between July 18 and August 3, 1905. Some of these observations were made in the light grader; others outside. In Mast, Light Reactions in Lower Organisms. 11g many cases a single colony was selected and its course studied, in place of that of a number of colonies in a group. To my surprise, I found that whereas during the preceding season, 1904, Volvox colonies deflected, with scarcely an exception, to the left, they now deflected to the right more often than to the left. We shall consider the cause of this somewhat in detail later. In all I have records of 174 paths, only a few of which were observed in light of uniform intensity. Seventy-eight of them deflect to the left from 2 to 45 degrees; seventy-five deflect to the right from 2 to 45 degrees; and only twenty-one are found in the area between 2 degrees to the right and 2 degrees to the left, and very few of these are parallel with the rays. In these experiments, however, only deflections to the sides were recorded; it is important to note that marked deflection up or down was also to be observed. It becomes clear then, that the colonies which appeared to be moving nearly parallel with the rays when seen from above, were in all probability slowly ascend- ing or descending as they proceeded toward the source of light. eine cause of deflection—the inability to orient accurately —is complicated. ‘The direction of movement in Volvox is affected by internal as well as by external factors. ‘The effect of some of these factors on orientation or deflection will be discussed under the following headings: (a) Effect of internal factors on orienta- tion; (0) Effect of light intensity on orientation; (c) Effect of gravitation on orientation; (d) Effect of contact stimulation and rotation on orientation. a. Effect of Internal Factors on Orientation. If a number of Volvox colonies, varying in size, are put into the aquarium at the same time and allowed to swim horizontally toward any concentrated source of light, it will be seen that the larger colonies, especially such as contain numerous daughter- colonies, soon collect along the right side of the group, and the smaller ones, and such as contain only very small daughter-colonies along the left side. In some experiments there was “such striking difference between the deflection of different colonies in a group that two distinct columns were formed, which moved across the aquarium at quite a definite angle with each other. The mght column in such cases invariably contained most of the larger colonies, and the left most of the smaller ones. 120 Fournal of Comparative Neurology and Psychology. On July 20, 1905, the paths of two such diverging columns, observed in light of uniform intensity, were recorded. ‘The one containing the larger colonies deflected to the right, making an angle of nine degrees with the light rays, while the one containing the smaller colonies deflected to the left, making an angle of fifteen degrees. Both columns, however, sometimes deflect to the right or to the left of the light rays. Deflection then varies with different periods in the life of the colonies; but it also depends upon the physiological state of the organisms, as is shown by the following observations. ie the morning, after being in the aquarium all night, Volvox colonies were repeatedly found lying on the bottom, apparently perfectly quiet. They were in a state which may be termed dark rigor. When light is thrown on them while they are in this condi- tion, they do not respond at once. After a time, however, they begin to swim about, slowly at first, without orienting; but soon, more rapidly, until they become normally active, and move toward the light. Apparently there 1s a certain chemical change neces- sary to bring the organisms out of the state of dark rigor into such a condition that they can respond readily to light; and this change appears to be induced by light. The production of carbon dioxid in darkness suggests itself as the probable cause of dark rigor. When a colony, after having been in darkness all might, first begins to respond to light, it moves toward the surface of the water ane deflects strongly einer to the right or left as it proceeds toward the source of light. But if it be made to cross the aquarium several times in succession, it is found that the deflection gradually decreases until it has traveled 25 to 30 cm. ‘Then it reaches an apparently stable condition; and on the following trips it takes a fixed course which may be at almost any angle with the light rays, but is usually at an angle of from 5 to to degrees. Such reactions were ob- served many times, mostly in experiments performed for other purposes. The following detailed experiment is typical. On August 7, 1905, Volvox was collected at 6. 15 a. m. and left in total darkness until 8.30 a. m., at which time the colonies were still moving about, but very slowly. One of them was put into the aquarium in the light grader in a light intensity of nearly 400 candle meters. ‘This colony moved about irregularly at first and deflected strongly to the right, but it soon became more active and moved quite rapidly toward the light. On its first trip across Mast, Light Reactions in Lower Organisms. 121 the aquarium it deflected to the right 17 degrees, on its second trip 15 degrees, and on its third trip 11.5 degrees. The following thirty trips were made with so little deviation from 11 degrees that it could not be measured. ‘The experiment was closed at 10.45 a. m., two hours and fifteen minutes after it was begun. It appears, then, that when the internal factors have become stable, and the external factors are not changed, the angle of deflec- tion remains constant. b. Effect of Change in Light Intensity on Ortentation. In general a decided increase or decrease in light intensity causes an increase in deflection. [his seems to be connected with the fact, pomted out by Hormes (’03, p. 321), that in low or high light intensity the colonies are not strongly positive. On August 3, 1905, the relation between the course of a given colony and the ray direction was obtained in a light intensity of 400 candle meters and also in an intensity of 20 candle meters. In the higher intensity, the deflection to the left was found to be 1 degree; in the lower intensity 11 degrees. ‘The course was ascertained by letting the colony across the aquarium three times in succession in the lower intensity, then three times in the higher, then twice in the lower, and finally twice in the higher. The light intensity was reduced by cutting off part of the light with a screen, which contained a narrow slit, placed close to the Nernst glower, and so arranged that the slit was perpendicular to the glower. Neither the light nor the aquarium had to be touched in decreasing or increasing the light intensity, so the ray direction was unquestionably the same under both conditions. ‘There was remarkably little variation in the angle of deflection in all the trips made across the aquarium in either light intensity. There can thus be no question about the accuracy of these observ- ations. [his experiment was repeated a few days later with similar results. “The colony selected, however, deflected to the right instead of to the left, as the one in the first experiment had done. ‘The deflection in the second experiment was studied in three different light intensities: 20 candle meters, 400 candle meters and nearly 2000 candle meters. ‘The highest intensity was produced by a carbon arc. ‘The angle between the light rays and the course taken by the colony was found to be 12 degrees in 20 candle meters intensity; 2 degrees in 400, and 40 degrees in 2000. 122 ‘fournal of Comparative Neurology and Psychology. Moderate increase or decrease in light intensity does not appear to affect the degree of deflection, e¢. g., the path of a given colony in a light intensity of 400 candle meters was found to be so nearly the same as that of the same colony in an intensity of 100 candle meters that the difference could not be measured. From numer- ous experiments, it appears that in order to influence deflection, the increase or decrease in intensity must be great enough to affect the positiveness of the organism; that is, the intensity must be decreased to somewhere near the threshold or increased to near the optimum. Now the threshold and optimum in different colonies, and in the same colony under different conditions, vary extremely. It is therefore to be expected that the effect of variation in intensity on deflection varies much. ‘This was found to be true experi- mentally. The above discussion on the effect of change in light intensity on deflection might lead one to assume that all Volvox colonies could be made to move parallel with the rays, if the proper light intensity were used. ‘This, however, was not found to be true. To bring about such a reaction, not only the proper light intensity is necessary, but the organisms must also be in a certain physio- logical state. Immediately after taking colonies from darkness or very intense light in which they have been for some time, they are in such a condition that no light intensity was found in which they travel parallel with the rays. And many colonies under various other conditions could not be made to swim parallel with the rays. In the above discussion deflection up or down is not considered; by parallel we mean merely without lateral deflection. Bs Effect of Gravitation on Orientation. If Volvox is killed in formol and then transferred to water, it gradually sinks to the bottom, showing that its specific gravity is greater han one. When first dropped into the water there 1s, of course, no indication of orientation; the longitudinal axis of the different colonies point. in all directions, but as they sink, it is soon found that their axis becomes approximately perpendicular, Lowen the colonies become oriented with the anterior end up. Such orientation is especially marked in organisms which contain num- erous daughter-colonies, but it is apparently accidental or absent in those without. Since the colonies are dead this orientation can be brought about only by a difference in the specific gravity of the Mast, Light Reactions in Lower Organisms. 123 anterior and the posterior half of the body, and since this orienta- tion to gravity is definite only in specimens containing daughter- colonies it is evident that the daughter-colonies, located as they are mostly in the posterior half of the body, render it heavier than the anterior half. The specific gravity, of living Volvox is also greater than one. If the colonies become inactive > they sink to die bottom, and it is undoubtedly due to this that they are frequently found lying quietly on the bottom of the aquarium after being in darkness all night. The fact that the specific gravity of living colonies is greater than one and that the posterior end of those which contain daughter-colonies or spores 1s heavier than the anterior end, has an important bearing on orientation to light and the direction of motion. It is owing to the difference in weight of the two ends, that the anterior end turns up, if for any reason the forward motion of a colony ceases. In this position the colonies are frequently found in very dim light, apparently hanging in the water motionless. If they become active while in this position, they swim upward. Such activity may be induced by light so dim that the organisms do not orient. ‘The degree of activity in light of low intensity, without doubt, depends upon the physiological state of the organ- ism, for it was frequently noticed that many colonies did not become quiet in darkness, and several times after exposure to darkness for as long as four or five hours, a large majority was found at the surface of the water, apparently clinging to the surface film. If horizontal movement of Volvox colonies toward a given source of light is observed from the side instead of from above, as was customary in the experiments described in the preceding pages, it can be clearly seen that the longitudinal axis of most of the speci- mens forms a decided angle with the bottom of the aquarium, that is, the posterior end is lower than the anterior. ‘This angle varies from zero to go degrees. Contrary to the observations of KLEIN (89, p. 169), it was found to be larger in organisms which contain numerous daughter-colonies and spores than in those which do not contain these structures. It is therefore in all probability caused by the difference in weight of the two ends. The angle which the axis anaes with the bottom of the aquarium varies also with the light intensity. “The more strongly positive a given colony is, the smaller the angle; but the positiveness of 124 ‘fournal of Comparative Neurology and Psychology. Volvox depends upon the light intensity, as was shown above (p. 121). Light, therefore, under the seconditions, tends to keep the axis aren, while gravitation tends to keep it vertical. In traveling horizontally toward a source of light, then, the axis of Volvox is not parallel with its course, but if the light 1s suddenly decreased in intensity, as was repeatedly done, the colo- nies change their course and start in the direction in which the axis points. ‘This seems to indicate clearly that they tend to travel in a direction parallel with the longitudinal axis. Now when they are strongly positive the axis becomes nearly horizontal and they consequently tend to move horizontally toward the source of light, but the force of gravity keeps pulling them down so that when ‘the colonies are strongly positive they move toward the light very near the bottom of the aquarium. This was observed many times. If they are oriented in a beam of light thrown through the aqua- rium at some distance from the bottom, they soon sink out of the region of light into the darkness, but as soon as they get into the Sear region gravity causes their longitudinal axis to pale a vertical position and they swim upward again, unless darkness produces inactivity and thus causes them to sink slowly to the bottom. Thus they were frequently seen, while swimming across the aqua- rium, to pass from light down into darkness and ees into the light again several times. ais the specimens are not strongly positive the inclination of the axis toward the horizontal is not great, and they therefore tend to swim toward the surface. This upward ten- dency may be just sufficient to compensate the effect of gravity, and if so, the colonies appear to be moving parallel with the rays when viewed from the side. Under these conditions specimens were frequently seen to sw.m across the aquarium with very little deflec- tion upward or downward. In summing up, we find that when the colonies are strongly posi- tive to light, the deflection to the side is reduced to a minimum, but owing to the effect of gravitation the downward deflection is marked; and when they are not strongly positive the deflection to the side is marked, while the vertical deflection may be practically zero. Thus it becomes evident that accurate orientation in hori- zontal movement is indeed exceptional. If gravitation tends to keep the longitudinal axis of Volvox verti- cal with the anterior end directed upward, and light tends to keep it parallel with the rays with the anterior end directed toward Mast, Light Reactions in Lower Organisms. 125 the source of light, and if the colonies tend to travel parallel with the axis, we should expect them to move parallel with the rays, when the rays are vertical and the source of light is above. ‘This was found to be approximately true, as is shown by the following experiment. On August 8, 1905, the plate ¢ glass aquarium was nearly filled with filtered water and put upon the stage of the light grader which was so arranged that the rays were vertical (see Fig. 5). Anumber of colonies were then put into the aquarium with a pipette and set free near the bottom in a beam of light, which was uniform in intensity and two and one-half centimeters square in cross section. After swimming upward to the surface of the water, some of the colonies wandered out into the shaded region. ‘These could readily be forced to swim down again by reflecting the beam of light upward through the aquarium slightly to one side of the illuminated area produced by the light direct from the glower. The reflected beam could be made vertical by tipping the light grader so that the direct beam of the light made an angle of about 10 degrees with the vertical. In this way movements both upward and downward were studied. In swimming up Volvox was found to travel very nearly parallel with the light rays, taking a spiral course, which was in some instances at least 2 mm. te In thus traveling upward, it could be clearly seen that the anterior end described a larger circle than the posterior, which in many colonies appeared to go almost in a straight line. ‘The anterior end appeared to swing about the posterior as a pivot. While a large majority of the colonies trav- eled nearly parallel with the rays, “there were a few which deflected considerably, some to such an extent that they passed out of the beam of light before reaching the surface of the water. That the movement parallel with the rays was due to the harmonious inter- action of gravitation and light, and not to especially favorable conditions of light intensity, was demonstrated by the course of a certain colony in traveling upward toward the source of light parallel with the direction of the force of gravity, and then again in movement perpendicular to this force. When moving parallel with the direction of the force of gravity, the colony observed did not deflect more than one degree in making several trips up through the water in the aquarium, but in moving perpendicular to this force in the same aquarium and in the same light intensity, this same colony deflected 30 degrees to the right. 126 ‘fournal of Comparative Neurology and Psychology. In swimming downward there is no evidence of a spiral course, the path, however, 1s much more irregular than in swimming up- ward; colonies on their way down were frequently seen to swerve to one side as if about to turn and go in the opposite direction. Gravitation, as has been stated, tends to keep the longitudinal axis vertical with the anterior end up, but the light from below, under the conditions of the experiment, tends to orient the organisms with the anterior end down. It is the interaction of these two opposing directive forces which brings about the swerving reaction and the irregularity 1 in the downward course. If the light is weak its direc- tive influence is not as strong as that of gravitation, and many colo- nies may be seen oriented with the anteriorend up. “The downward movement of specimens in this position is very slow compared with that of those with the anterior end directed down. ‘This 1s evidently the result of the effect of gravity and a tendency to swim upward, 1. €., in the direction which the anterior end faces. The rate of movement varies greatly in different colonies under the same external conditions. It is, however, in general, much faster toward a source of light with the force of gravitation than against it. [his is shown by the following results. The time required for each of three specimens to swim downward 8 cm. toward a source of light, in a given intensity, was found to be 40 seconds for one, 32 seconds for another, and 30 seconds for the third. That required to swim up toward the light in the same intensity, was 100 seconds, 80 seconds, and 66 seconds, respectively, an average of 48 seconds longer to swim upward 8 cm. than to swim the same distance downward. It is very probable that the activity of Volvox in swimming upward is just as great as it 1s in swimming downward and that the difference in rate is entirely due to its specific gravity. Insumming up this whole matter we find: (1) That Volvox tends to move in a direction parallel with its longitudinal axis; (2) that gravity tends to keep this axis vertical, ith the anterior end up, but owing to stimulation by light the organisms tend to orient with the anterior end facing in the direction of strongest illumination ; (3) that Volvox trav els * very nearly parallel with the rays in moving up toward a compact source of light, but that it very rarely moves parallel with the rays in swimming downward or horizontally toward a source of light; (4) that in reacting to light it almost always deflects upward or downward, or to the right or left, and Mast, Light Reactions in Lower Organisms. W27 that these deflections depend upon the light intensity and the physiological conditions of the organisms; (5) that it deflects most in moving horizontally when its axis is most nearly vertical and that the axis becomes most nearly vertical when the organism is not strongly positive. In swimming downward toward a source of light, the deflections are clearly due to a tendency of the organism to orient in the direc- tion of the force of gravity with the anterior end directed upward. In swimming horizontally it is clear that the downward deflection is due to the specific gravity of the organism, andthe upward deflec- tion to the tendency to swim parallel with the axis. “The cause of lateral deflection in such movement is, however, not so evident. Colonies swimming horizontally toward a single source of light, tend, as stated, to take a position such that the axis is parallel with the rays and the zooids on all sides are equally illuminated. _ If the organisms are strongly positive, the axis is nearly horizontal, so that if they turn to the right or left, one side immediately becomes shaded and thus causes a reaction which tends to keep the direc- tion of movement parallel with the rays. But if the colonies are not strongly positive, the axis is more nearly vertical, and while they are in this position there isalre: dy a difference in light intensity on opposite sides, so thzt if the organism now turns to the right or to the left, this intensity difference is only slightly changed. ‘There is consequently but little cause for reaction and therefore nothing to prevent movement at an angle with the rays. Since lateral deflection has been observed to be greater the more nearly vertical the axis, it seems probable that this is a valid explanation of the cause of such deflection. But how 1s it that a colony can repeat- edly travel across an aquarium, making the same angle with the light rays each time; or that when the position of the source of light is changed, after it has started on a course at a given angle with the rays, it changes its course until the new one has the same angle? The only explanation I have to offer is the following: If colonies in water only a few millimeters deep, are simultaneously and equally illuminated from above and from below, they do not move in straight lines but in curves, frequently making continuous complete circles to the right or to the left. They therefore seem to have a tendency to swim in curves. I am unable to account forthis. But if it 1s true, their path (as seen from above) in travel- ing horizontally toward a source of light, must be the resultant of 128 Journal of Comparative Neurology and Psychology. the directive force of the light and the tendency to swim in curves. This would necessarily result in movement at an angle with the light rays. ‘The size of this angle would depend upon the relative efficiency of the directive force of the light and the tendency to swim in circles. If the organisms are strongly positive, the direc- tive force of the light is strong compared with the tendency to move in curves and the angle becomes small. But if they are not strongly positive, the directive force of the light is relatively weak and the angle becomes large. The theoretic results thus formu- lated are in accord with the experimental results described in the preceding pages. d. Effect of Contact Stimulation and Rctation on Orientation. When colonies of Volvox come in contact with the side of the aquarium nearest the light and the rays are perpendicular to this side, many of them soon begin to drift to the right along the glass wall, and in a short time a large majority are found in the right hand corner of the aquarium nearest the source of light. ‘This movement to the right takes place in a field of graded light as well as in light of uniform intensity, and it is apparently as marked if the intense end of the field is to the left as it 1s if this end is to the right. ‘Thus the organisms were frequently seen to move along the wall toward the right, on the one hand, into regions gradually decreasing 1 in intensity nel they passed 1 into darkness and, on the other, into regions gradually increasing in intensity until they became negative. “he movement to hie right along the wall takes place, with much greater regularity, however, in specimens containing large daughter-colonies or spores than in young colo- nies. Indeed it is doubtful whether more of the young colonies turn to the right than to the left after they reach the wall of the aquarium. At any rate shortly after the introduction of a group containing both large and small colonies, practically all the large colonies, together ah some small ones, have gathered in the right hand corner, some small ones have collected in the left hand corner, and a few of both kinds usually remain scattered along the entire side. What is the cause of this movement to the right along the wall? After reaching the wall the colonies ordinarily remain with the anterior end in contact with it for some little time, but sooner or Mast, Light Reactions in Lower Organisms. 129 later the posterior end begins to settle, the longitudinal axis becomes nearly vertical, and the organism begins to swim upward along the wall, deflecting to the right. ‘The angle of deflection varies greatly. Some colonies travel nearly parallel with the bottom at once; others swim nearly straight upward. During the time that the anterior end is in contact with the wall, the colonies usually rotate counter-clockwise without reversal, and rotation in this direction frequently continues during the whole process of turning and moving to the right. It is, therefore, clear that the drifting to the right along the wall is not due to change in the direction of rotation. After the axis becomes nearly vertical the colonies sometimes remain in close contact with the wall but continue to rotate counter- clockwise without moving forward, and thus roll along the wall to the left. Frequently after thus moving along the wall a short dis- tance, the anterior end turns to the left and the organism begins to swim forward, but still continues to roll on the wall. Ties rolling along the wall, together with the effect of gravity, soon carries it to the bottom of the aquarium, where it apparently be- comes lodged in the angle between the bottom and the side. Here it remains for a time, but sooner or later works its way out, usually by swimming back from the wall a short distance, after which it turns and soon comes in contact with the wall again. A colony may, as is clear from what has just been said, turn either to the left or to the right after reaching the wall, but many of those which turn to the left are prevented from continuing on their course by the effects of rotation and gravitation, as explained above; and since those which turn toward the right are not thus prevented from continuing, the result is, of course, a general drifting of the colonies in this direction. But as a matter of observation a much larger proportion of colonies turn to the right than to the left shortly after they reach the wall, so that general movement to the right cannot be primarily brought about by the prevention of con- tinuous movement to the left. Neither can it be due primarily to the direction of rotation, for many colonies were repeatedly seen to deflect to the left in swimming across the aquarium toward the source of light, and then to the right, after coming in contact with the wall, without changing the direction of rotation. It seems then that the tendency to turn to the right after reaching the wall must be due primarily to contact stimuli. As evidence in support of this view I present the following experiments: 130 Fournal of Comparative Neurology and Psychology. On August 10, 1905, between 20 and 30 colonies were put into the aquarium into an intensity of 21 candle meters. When the rays were parallel with the bottom the group spread very little and swam across the aquarium nearly parallel with the rays. But when the glower was lowered so that the rays passed up through the glass bottom of the aquarium, making an angle of 25 degrees with it, the group spread out considerably and the majority deflec- ted quite sharply to the right. The largest colonies were found along the right side of the column and the smallest along the left, under both conditions. It is doubtful whether the smaller colo- nies changed their course after the position of the glower was changed, but the larger ones certainly did. Later more definite results were obtained by experimenting with a single colony. The specimen selected was of medium size and contained quite a num- ber of rather small daughter-colonies. When the rays were par- allel with the bottom this colony deflected three degrees to the right, but when the light was below the level of the bottom and came up through it so that the rays made an angle of 25 degrees with it, the organism deflected 19 degrees in the same direction. In ascertaining these deflections the colony was allowed to cross the aquarium a few times first with the rays parallel with the bottom, then with the rays at an angle of 25 degrees with it, then again with the rays parallel with it, and finally, with the rays at an angle of 25 degrees. ‘Uhe deflection during the various trips under each condition, was nearly constant. It is therefore certain that the increase in deflection was not due to a possible change in the physiological condition of the organism. Neither was it due to difference in light intensity, for the strength of illumination was nearly the same under the two conditions one the experiment, and deflection is not much affected unless there is very marked change in the intensity of the light (see p. 122). In moving toward the light in rays parallel with the bottom, the axis of this colony was at an angle of about 12 degrees with the bottom. The organism moved near the bottom of the aquarium so that the posterior end appeared to be slightly in contact with it. But when the light came from beneath at anangle of 25 degrees the axis of the colony was nearly horizontal and the organism moved so near the bottom that the cilia must have come in close con- tact with it. As the specimen thus swam across the aquarium the axis could be clearly seen to swing at short intervals, from a posi- Mast, Light Reactions in Lower Organisms. 131 tion nearly’ parallel with the general direction of motion to a posi- tion nearly perpendicular to it. This swinging of the axis, it is thought, was due to contact with the bottom and counter-clock- Wise rotation, owing to which the posterior end seemed to roll to the left more rapidly than the anterior. This appeared to turn the anterior end of the axis sharply to the right, and since the colo- nies tend to move parallel with their axis, it would cause deflection to the right. Some such reaction must be at the basis of the deflec- tion to the right when the organism 1s in contact with the vertical wall nearest the light. It may also explain why the larger colonies are found to deflect more to the right than the eineller. since the specific gravity of the two is different. I have discussed the cause of the movement of Volvox to the right along a vertical wall at some length because of its importance in the study of the reactions of the colonies in aggregating in regions of optimum intensity in graded light, which will be taken up later. 7. ORIENTATION TO LIGHT FROM TWO SOURCES. In the preceding pages we have conclusively demonstrated that while Volvox moves in general toward a given source of light, it seldom travels parallel with the rays, excepting when they are ver- tical, and it swims upward. But while the colonies do not usually swim parallel with the rays they still orient in a definite way. ‘That is, if a colony is swimming at a given angle with the rays and the source of light 1 is moved, it so changes the direction of motion that its course again makes the same angle with the rays that it did before the position of the source of light was changed. What is the cause of orientation ? OrtTmaNNs, as has been stated (p. 100), came to the conclusion that difference in light intensity is the principal cause of orientation of Volvox, but he presented no direct evidence in favor of this view, and his indirect evidence is based upon experiments which have since been proved to be defective. HoLMeEs was not able to explain orientation by assuming difference in light intensity on opposite sides of the organism to be the cause, and he 1s inclined to believe that it is due to the direction of the rays. He writes (03, p. 324): “It seems not altogether improbable that light in passing through a nearly transparent organism like Volvox exercises a directive effect upon its movements, in a similar way, whatever it may be, to that produced by the current of electricity. 132 ‘fournal of Comparative Neurology and Psychology. The direction of the ray may be the important factor in orientation irrespective of difference of intensity of light upon different parts of the organism, as has been maintained by Sacus for the photo- tropic movements of plants. I am not ready to adopt the theory of Sacus, but I feel that it is a viewthat is not entirely out of court.” The following experiments on the movement of Volvox when exposed to light from two different sources, and on the orientation of Volvox in light graded in intensity seem to me to settle this ques- tion conclusively. On August 18, 1904, a single 222 volt Nernst glower was fixed in a vertical position 70 cm. from the middle of the plate glass aquarium, so that the lower end of the glower was level with the bottom of the aquarium and the rays perpendicular to the side at a point 4cm.fromoneend. A single 110 volt glower was arranged like the 222 volt glower, but in such a position that the light rays were perpendicular to the end of the aquarium at the middle and, therefore, perpendicular to the rays from the 222 volt glower at a point 4 cm. from the end, and half way between the two sides, as represented in Fig. 7. The 222 volt glower was stationary, but the 110 volt glower could be moved to any desired distance from the aquarium. ‘These glowers were both carefully screened so that the only light which escaped passed through a rectangular slit a trifle larger than the glower. ‘The side and end of the aqua- rium facing the glowers was also screened, with the exception of an opening one centimeter wide and six centimeters long, at the bottom of the aquarium, as indicated in Fig. 7. The aquarium contained thoroughly filtered water 1.5cm.deep. ‘Thus, practically all reflection from the sides of the aquarium and the surface of the water was eliminated. The direction of movement of Volvox was ascertained, first with the 222 volt glower exposed alone, then with both glowers exposed, the 222 volt glower 66 cm. from the side, and the 110 volt glower 24, 49, 99, and 199 cm. from the end of the aquarium. In order to ascertain the direction of motion under the various light conditions, a considerable number of colonies were carefully dropped into the corner of the light area farthest from the glowers. Among the specimens used in this experiment there were about as many that deflected to the right as to the left, so that when one glower only was exposed the center of the group of colonies moved across the aquarium practically parallel with the light rays. Sev- e Mast, Light Reactions in Lower Organisms. 133 eral trials were made under each light condition and each path, as recorded in the table below and in Fig. 7, is the average of sev- eral such trials. “There was, however, surprisingly little variation in the direction of motion of different groups when subjected to the same light condition. ‘The light intensity was measured with care. Both glowers were on the same circuit so that variation in voltage could not have affected markedly the relative intensity of the light from the two sources. There can thus be no question about the approximate accuracy of the experiments, the results of which will be readily understood by referring to Table I, in connection with Fig. 7. TABLE I. if Il. ; II. 82.4 candle meters. .© candle meters. o degrees. 82.4 candle meters. 6.0 candle meters. 9 degrees. 82.4 candle meters. 23.5 candle meters. 25 degrees. 82.4 candle meters. 89.0 candle meters. 47 degrees. 82.4 candle meters. 318.8 candle meters. 59 degrees. Table I represents the effect of light from two sources on the direction of movement of Volvox. Column 1 gives the light inten- sities at the middle of the light area in the aquarium, which were produced by the 222 volt glower under the five different condi- tions. Column-ir gives light intensities produced by the 110 volt glower, and column 111 the angles between the rays produced by the 222 volt glower and the course taken by the organisms under the different light conditions. In these five experiments the direction of the rays from the two sources of light was practically constant, but the direction of movement of the Volvox colonies varied 50 degrees. ‘This varia- tion was certainly not primarily due to any influence of the ray direction; for when the relative intensity of light affecting different sides of the organism was changed the orientation changed, though the direction of the rays remained the’ same. It can, therefore, be considered fully demonstrated that difference in light intensity on different sides of the colonies may determine orientation inde- pendently of the direction of the rays. Additional proof of this con- clusion will be given later, in experiments of a different character. This conclusion is not in harmony with the dictum of Logs, repeatedly expressed in a recent work (1905), in which he writes, “Tt is explicitly stated in this and the following papers that if there are several sources of light of unequal intensity, the light with the strongest intensity determines the orientation and direction of 134. ‘fournal of Comparative Neurology and Psychology. motion of the animal. Other possible complications are covered by the unequivocal statement, made and emphasized in this and the following papers on the same subject, that the main feature in all phenomena of heliotropism is the fact that symmetri- cal points of the photosensitive surface of the animal must be struck by the rays of light at the same angle. It is in full harmony with this fact that if two sources of light of equal intensity and WOR \\ \ is AAG Yidédédddddédbléda Fic. 7. Representation of the direction of movement of Volvox when subjected to light from two sources. a, plate glass aquarium 8 cm. long and 8 cm. wide; b, 222 volt Nernst glower, 66 cm. from aquarium (distance from aquarium constant); c, 110 volt glower, (distance from aquarium variable); d, screen; e, point of introduction of Volvox; f, direction of light rays; 1, 2, 3, and 4, courses of Volvox exposed to light from both glowers: 1, with 110 volt glower 199 cm. from aquarium; 2, with 110 volt glower 99 cm. from aquarium; 3, with 110 volt glower 49 cm. from aquarium; 4, with 110 volt glower 24 cm. from aquarium; xy, course of Volvox when exposed to light from glower b only; y—z, course when exposed to light from glower c only. distance act simultaneously upon a heliotropic animal, the animal puts its median plane at right angles to the line connecting the two sources of light. ‘This fact was not only known to me but Mast, Light Reactions in Lower Organisms. 135 had been demonstrated by me on the larve of flies as early as 1887, in Wurzburg, and often enough since. ‘These facts seem to have escaped several of my critics” [p. 2]. ‘When the diffuse daylight which struck the larvae [Musca larve] came from two windows, the planes of which were at an angle of go° with each other, the paths taken by the larve lay diagonally between the two planes. his experiment was recently published by an American physiologist as a new discovery to prove that | had overlooked the importance of tae intensity of light!’ (p. 61-62). “The direction of the median plane or the direction of the pro- gressive movements of an animal coincides with the direction of the rays of light, if there is only a single source of light. If there are two sources of light of different intensities, the animal is oriented by the stronger of the two lights. If their intensities be equal, the animal is oriented in such a way as to have symmetrical points of its body struck by the rays at the same angle’ (p. 82). “Attention need scarcely be called to the fact that if rays of light strike the animal [larva of Limulus polyphemus] simultaneously from various directions, and the animal is able to move freely in all directions, the more intense rays will determine the direction of the progressive movements” (p. 268). It is evident without further discussion that the reactions of Volvox do not fit the statements by Lor, given in the above quo- tations. Upon what experimental evidence does he base these statements? [hose with reference to orientation when the ani- mals are subjected to light from two or more sources are based largely, if not entirely, upon the following observations: (1) “When the diffuse daylight which struck the larvae (Musca larve) came from two windows, the planes of which were at an angle of go° with each other, the path taken by the larva lay diagonally between the two planes.” (2) “Hawk moths were brought into a room with the single window at one end, and a petroleum lamp at the opposite end. It was found that, as twilight came on, the moth flew to the window, or to the light, according to the relative intensity of the one or the other at the point where the moth was liberated.” In the first place I am unable to understand how the direction of rays can be ascertained in diffuse daylight coming through a window; and in the second place, it is certainly not difficult to see that an object placed between two windows, or between a 136 fournal of Comparative Neurology and Psychology. window anda petroleum lamp, in an ordinary room, is illuminated by light rays striking it from every conceivable direction, for light aden such conditions is reflected from practically all surfaces in the room as well as from those outside. Under the conditions of the experiments cited above, then, the larvae and moths were not exposed to light from two sources but to light from an infinite number of sources, and the direction of the rays was not known. How then, can it be concluded from the results of these and simi- lar experiments (1) “That if there are several sources of light of unequal intensity, the light with the strongest intensity deter- mines the orientation and direction of movement of the animals;”’ (2) “that symmetrical points of the photosensitive surface of the animal must be struck by the rays of light at the same angle;” and (3) “that if two sources of light at the same intensity and distance act simultaneously upon a heliotropic animal, the animal puts its median plane at right angles to the line connecting the two sources of light ?” Let it be clearly understood that in the criticism of LoEB’s con- clusions, I do not wish to intimate, that because the reactions of Volvox or any other organism do not take place in accord with those conclusions, they necessarily cannot hold for the organisms Lors worked with. I do, however, wish to state and emphasize that in my opinion his experimental results as quoted above, do not warrant his conclusions, even for the animals worked on, much less for all organisms which orient in light. ‘The experiments upon which Logs bases his theory of orienta- tion to a single source of light will be discused later (see p. 142). 8. ORIENTATION IN LIGHT GRADED IN INTENSITY. The reaction of Volvox to light from two sources varying in relative intensity seems to me to prove conclusively that orienta- tion is determined by the relative intensity of the light on opposite sides of the organism, while there is no evidence that the direction of the rays has anything to do with orientation in this organism except in so far as it may affect the relative light intensity on opposite sides. If, however, difference in light intensity on oppo- © site sides of a colony can be produced with the rays of light approx- imately parallel, and such intensity difference affects the direc- tion of motion, the verdict must be considered final. Mast, Light Reactions in Lower Organisms. 137 By means of the light grader referred to several times in the preceding pages, I was able to subject colonies to rays which were nearly parallel but decreased in intensity from one end of the field to the other, so that when the longitudinal axes of the colonies were parallel with the rays, one side was more strongly illuminated than the other; and I found that this intensity difference did affect the direction of motion, as will be shown in the following detailed account of the experiment. The light grader was so arranged that the Nernst glower was vertical and the rays and the long axis of the lens horizontal. ‘The * plate glass aquarium was so placed that the rays were parallel with the bottom. Now by fastening over the lens a screen, which contained an opening in the form of two truncated triangles with their apices in contact, a field of light: was produced which was of high intensity at either end and gradually became lower toward the middle. “Iwo methods were used in ascertaining the direction of movement in such a field of light. In the first method a large number of colonies were taken up in a pipette and half of them introduced into the aquarium near the side farthest from the glower at a fixed point some distance from one end of the field, and the other half in a similar place near the opposite end. ‘Thus the organisms in one group as they swam across the field were more intensely illuminated on the right side, while those in the other group were more intensely illuminated on the left side. In the second method a single colony was selected and allowed to cross the aquarium toward the source of light several times, first near the right end of the field so that the lower light intensity was to the left and then near the left end of the field so that the lower light intensity was to the right. ‘This alternating process was continued until the path in the two different positions was definitely established. ‘The angles of deflection were read and recorded as described on p. 117. Those obtained by the first method may be found in Table II and those by the second method in Table III. ‘The negative numbers indicate deflection to the left of the ray direction and the positive to the right. Table IV represents the effect of difference in light intensity. on deflection in graded light. ‘The course taken by the colonies was obtained by studying the reactions of single colonies, just as in the experiments of Table III. ‘This table shows that an increase 138 TABLE II. Light Intensity 333-£ candle meters. Angle of deflection with strongest illumination to the left. ~_ Umum Angle of deflection with strongest illumination to the right. _ a Mum Average difference TABLE III. Light Intensity 333-: candle meters. Angle of deflection with strongest illumination to the left. || WWuH (oy Sp an al _ Angle of deflection in 142 candle meters of light. Strongest illu- mination to the left. Gass 10 Angle of deflection with strongest illumination to the right. Average difference TABLE IV. Angle of deflection in 380 candle meters of light. Strongest illumination to the left. Average difference Fournal of Comparative Neurology and Psychology. Difference in angle of deflection ° (Se a CANE NYPH HYD HON YD DAA HO KNW AMFUH An n Difference in angle of deflection 3.1 degrees. Difference in angle of deflection Dib 5 13 degrees. Mast, Light Reactions in Lower Organisms. 139 in light intensity from 142 candle’ meters to 380 candle meters causes an average decrease in deflection of 13 degrees. By referring to the above tables and text figures it will be noted: (1) that Volvox, in swimming horizontally toward a source of light, seldom moves parallel with the rays. There is striking individual variation in the angle of deflection, the variation in these experiments being from 16 degrees to the left to 24 degrees to the right; (2) that in a field of light graded in intensity there 1s a tendency to deflect toward the brighter end of the field, an Fic. 8. Graphic representation of the total average difference in deflection due to difference in light intensity on opposite sides of the colonies, as indicated in Tables II and III. a, plate glass aquarium 8 cm. wide and 15*cm. long; }, light rays; c, c’ points where the colonies were introduced; d, average course with the region of highest light intensity to left; e, average course with strongest illumination to the right. Light intensity at (f) the middle of field 57.12 candle meters. From the middle the intensity gradually increased toward either end where it was 442.68 candle meters. Intensity at c, 327 candle meters, at c’, 263 candle meters. average of over 15 degrees under the conditions of these experi- ments; (3) that the degree of deflection in a field of light graded in intensity depends upon the strength of illumination, it being greater in a low light intensity than in a high one. A decrease in intensity from 380 candle meters to 142 candle meters without 140 ‘fournal of Comparative Neurology and Psychology. change in the grade of intensity caused an average increase in deflection of 13 degrees. Cause of Deflection Toward the More Strongly Illuminated Side in Graded Light.—If a colony of Volvox deflects to the right in light of uniform intensity it will deflect more in a field of light graded in intensfty, provided the more highly illuminated end of the field is to the right, but not as much if this end is to the left. ‘his fact is clearly expressed in Fig. 8. Under the condi- tions of the experiments described above, this difference in deflec- - tion must have been primarily due to one of three factors: (1) difference in total light intensity under the two conditions; namely, with the more highly illuminated end of the field to the right and with this end to the left; (2) refraction or reflection as the light passes through the aquarium; (3) difference in light intensity on opposite sides of the colony. A discussion of these three factors follows. 1. We have demonstrated (see Table IV) that an increase in light intensity, without change of grade, causes a decrease in deflection. Now, as represented in Fig. 8, the colonies, as they deflect in crossing the aquarium with the brighter end of the field to the right, gradually pass into regions of higher light intensity, but when the brighter end of the field is to the left, they gradually pass into regions of lower intensity. ‘This consequently tends to cause a decrease in deflection under the former conditions and an increase under the latter, but the angle of deflection is greater under the former condition than under the latter. “The difference in deflection under the two conditions, therefore, cannot be due to the higher light intensity to which the organisms are exposed when the more strongly illuminated end of the field is to the right than when it is to the left. 2. As the light passes through the glass wall of the aquarium and the water in it, some is reflected and some refracted thus producing lateral rays. ‘This reflection and refraction cannot be entirely eliminated even with the utmost precaution. May not these lateral rays have been of sufficient intensity to cause deflec- tion toward the brighter end of the field as was found to be true in case of OLTMANNS’ apparatus 2 Orrmanns (92) produced a field of light graded in intensity by placing a hollow prism filled with a mixture of gelatine and India ink between the source of light and the aquarium. He assumed that the rays in the aquarium were all perpendicular to the wall facing the source of illumination. This, however, is not true, for the particles of ink in the prism disperse the light before it gets into the aquarium. Mast, Light Reactions in Lower Organisms. 141 The field of light in which the colonies were exposed in the above experiments was high in intensity at either end and low in the middle. In such a field of light it is clear that an organism swimming toward the light in the middle is stimulated alike on both sides, since the lateral rays necessarily come in equal numbers from both ends of the field. Consequently the direction of motion cannot be influenced by these rays. But if the organism in travel- ing toward the light swims nearer one end of the field than the others, the lateral rays might influence the direction of motion. If, however, the lateral rays do affect the direction of motion under such conditions, we should certainly expect to be able to detect it when all lateral rays on one side of a colony swimming toward the source of light are eliminated by shading the entire portion of the field either to the right or to the left of the colony. I repeated the above experiments many times with a portion of the field thus shaded, but was unable to detect any effect on the angle of deflec- tion. It must therefore be concluded that the difference in deflec- tion, represented in columns 1 and 11 of Tables II and III, was not caused by lateral rays. ‘THE DIRECTION OF MOTION IN VOLVOX EXPOSED TO LIGHT IS CONSEQUENTLY REGULATED BY THE RELATIVE INTENSITY OF THE LIGHT ON OPPOSITE SIDES OF THE COLONIES REGARDLESS OF THE DIRECTION OF THE RAYS. Cause of the Effect of Change in Intensity U pon the Degree of Deflection 1 in Graded Light. —The difference in intensity of illumina- tion on opposite sides of the colonies exposed in the light grader under the conditions of the experiments just discussed, can readily be calculated. The light intensity was 442.6 candle meters at either end of the field, from which it gradually decreased toward the middle, where it was 57 candle meters. ‘The distance from the middle to either end was 60 millimeters. We have therefore a change of 385+ candle meters in 60 millimeters or 6.4 candle meters per millimeter. The largest colonies are nearly a milli- meter in diameter and the average light intensity to which they were exposed was about 333 candle meters. In the largest speci- mens, then, one side was exposed to an intensity of about 330 and the other to an intensity of about 336 candle meters. If WEBER’s law holds true, as we have good reason to believe, (see p. 171), we should expect this difference in intensity on Oppo- site sides to be more effective in weak light than in strong and we should consequently expect a greater deflection in regions in the 142 ‘fournal of Comparative Neurology and Psychology. field where the light intensity is low than in those where it 1s high. As is clear from Table IV this was found to be true. But since we have demonstrated (p. 121) that deflection in light of uniform intensity can be increased either by decreasing or increasing the intensity, it may be maintained that the difference in deflection recorded in Table IV is due to the difference in light intensity in the field regardless of difference in intensity on opposite sides of the organisms. It must be remembered, however, that deflection in a field uniformly illuminated, is increased only if the intensity is decreased to a point near the threshold or increased to a point nearthe optimum. In the experiments just referred to, the inten- sity, in all probability, was far F F below the optimum and above the threshold, so that it is not likely that mere reduction in Ss K illumination affected the deflec- K ae tion to,any considerable extent. ——— The difference between the de- ——= gree of deflection in 142 candle A meters and 380 candle meters of light, graded in intensity, must, therefore, have been due to the greater effect of the differ- ences in light intensity on oppo- site sides of the organism when exposed to weak light than when exposed to strong. ‘The experimental results recorded in Table IV therefore support our previous conclusion, that the direction of motion in Volvox 1s regulated by the relative intensity of light on opposite sides of the colonies. Lors, however, as is well known, asserts that orientation 1s caused by the direction of the rays tegardless of the difference in light intensity. He bases his assertion largely on the results in the three following experiments on Porthesia larva (LOEB ’05, p. 25-28). “Experiment 3.—IThe test tube is placed perpendicular to the plane F of the window, and at the beginningof the experiment the animals are collected at the window side Bof thetest tube.”’> Now if the half near the window is covered, the animals soon collect at A. ‘As soon as they emerge from the box K into 4 they turn about, Fic. 9. After Lozz, 1yo05, p. 25, Fig. 1. Mast, Light Reactions in Lower Organisms. 143 direct their heads toward the window, move to the edge of the pasteboard and remain at the boundary between the covered and uncovered portions of the tube at 4, and especially at the top of Fic. 10. After Lorr, 1905, p.27, Fig. 2. the test tube. “The remarkable thing is that they are not distributed evenly over the whole brightly illuminated part of the test tube. ‘The explana- tion is as follows: As soon as the animals near the window at B are covered by the pasteboard, the weak rays of light reflected from the walls of the room fall upon them. The animals follow the paths of these rays and arrive at the uncovered portion of the tube’’ [Italics ours]. Experiment 4.—The larve were found to move to C, toward the win- dow FF. \\imethe test cube 6, shaded as represented in the figure, the light intensity is lower than in the test tube 4, not Experiment 5.—Vhe anim to B into the diffuse day- light. ‘They pass from the direct sunlight into diffuse daylight without even attemp- ting to return into the sun- light. In these, as inother experi- ments of Loers referred to on p. 135, the animals were exposed to light, the ray direc- tion of which must have been exceedingly complicated, since the light was diffused before it reached the tube in which the animals were. shaded, but the larve go to C. als move from direct sunlight at A E We BI "IGF B is S Fic. 11, After Loxrs, 1905, p. 28, Fig. 3. Moreover, the walls of the tube caused still further diffusion by refraction and reflection. How, then, could it be ascertained in any of these experiments whether the animals moved in the direction of the rays or not? 144 ‘fournal of Comparative Neurology and Psychology. In Experiments 4 and 5, the animals moved from a region of higher light intensity to one of lower. Now from this the author concluded that difference in intensity does not cause orientation, for if 1t did, the animals, being positive, would remain in the region most highly illuminated. In discussing the effect of difference in light intensity it is necessary to define the sense in which this is meant. ‘There is a vast difference between the difference in light intensity in a given field and tae difference in intensity on different areas of the surface of a particle in the field. For example, hold an opaque piece of paper in direct sunlight so that the rays strike it at right angles and you will find almost an infinite difference in the light i intensity on the two sides, but remove the paper and you will find that the intensity difference in the field is actually infini- tesimal. It is evident then that an organism can move from regions of higher to regions of lower light i intensity in a field pro- duced by apparatus arranged as represented in Figs. 10 and 11, and still have the anterior end constantly more highly illuminated than the posterior. Loes evidently did not recognize this in the experi- ments cited above, for he accepts the theory of SAcHs, who (87, p. 695) defines his position very clearly, as follows: “I came to the conclusion that in heliotropic curvatures, the impor- tant point isnot at all that the one side of the part of the plant is illuminated more strongly than the other, but that it is rather the direction in which the rays pass through the substance of the plant.” In moving toward the window in the test tubes arranged as represented in Figs. 10 and 11, the anterior end of the animal was very likely more highly illuminated than the posterior. On the assumption that difference in intensity on the surface of the organism causes orientation, the larvae would consequently be expected to move toward the window. I can, therefore, see nothing in these experiments which in any way indicates that difference in light intensity on the surface of the body, regardless of the Hedin of the light rays, is not the cause of orientation. Q. ORIENTATION OF SEGMENTS. In working on Volvox it was noticed that colonies with various portions missing still appeared to respond to stimulation by light. Such colonies were most frequently found after heavy rain storms or other rather violent disturbances. On July 28, 1905, a colony Mast, Light Reactions in Lower Organisms. 145 was found, in which the anterior end and a narrow portion of the side extending nearly to the posterior end, were missing. This segment oriented quite definitely. In swimming horizontally toward a source of light it moved approximately parallel with the rays, deflecting but little. When exposed to light from two sources of equal intensity, it took a course about midway between them. If the light from one of the sources was cut off after the segment had thus oriented, it continued on its original course for a few millimeters, then changed the direction of motion until it was orientedonce more. Its light reactions in general were like those of intact colonies, but the path of this segment instead of being straight as is true in case of entire colonies, was in the form of a spiral. This was evidently the result of the mechanical effect of the gap inthe sideznd rotation on the longitudinal axis. The reactions of many other segments of colonies were studied later. Most of these segments were made by cutting the colonies in pieces. In performing these operations a considerable number were put under a cover glass which was then carefully pressed down until the colonies split open. Under these conditions they usually split at the posterior end, but sometimes at the side. By inserting a needle ground to a knife-edge, the wall could be cut in any direction desired without much difficulty. It was found that segments of practically all forms and sizes responded to stimulation by light, but owing to their form and the effect of gravitation, many could move only in small circles, and were unable to orient. It can be stated definitely, however, that zmong segments of various forms and sizes, such as are produced by cutting the colo- nies in half, either parallel or perpendicular to the longitudinal axis, respond in general like whole colonies, with the exception that most of the segments take a spiral course, the width of which depends upon the form of the segment. It is thus clear that a colony of Volvox can orient when the anterior or the posterior end or one side is missing. A theory of orientation must be broad enough to explain not only the reactions of entire colonies but also those of any segments. I0. MECHANICS OF ORIENTATION. JENNINGS (’04, p. 32-62) found that Stentor cceruleus and Euglena viridis orient by means of motor reactions when exposed 146 ‘fournal of Comparative Neurology and Psychology. to light. If stimulated they turn toward a structurally defined side regardless of the direction of the rays or difference in light intensity on opposite sides of the organisms. If they fail to become oriented by a single motor reaction they repeat the reaction, turning successively in different directions, until they turn in the right direction; this direction they hold and thus become oriented. The process of orienting in these organisms is, therefore, strictly on the trial and error basis. In Volvox, taking a colony as a whole, there is no evidence of motor reactions, nor is there any hit or miss method about its orientation. It makes no mistakes in the process. If exposed to light it turns toward the source of light without error. What sort of mechanism has this organism, by means of which it can thus regulate the direction of its motion ? Fic. 12, After Hormes, 1903, p, 325. A colony of Volvox may be conceived to turn in its course by decreasing or increasing the backward stroke of the flagella on one side or the other, or by using the flagella on either or both ends as rudders, or even by directing the stroke of these flagella 1 in such a way as to turn the organism. But since the organisms orient when either the posterior or the anterior end is missing, and prob- ably also when both ends are missing, it is clear that the flagella on the ends do not function primarily in changing the direction of motion. Such changes must, therefore, bethe results of inequality of the strokes of the flagella on opposite sides. What then is it that causes the strokes on opposite sides to become unequal f HoiMEs (03, p. 325) after concluding that it cannot be caused by difference in light intensity on opposite sides, suggests the Mast, Light Reactions in Lower Organisms. 147 following explanation. “The orientation of the colony may be Becouneed for, if we suppose that the eye-spots are most sensitive to light striking them at a certain angle such as is indicated in the diagram by the lines a—b and e—f. If rays of light enter the colony in the direction of the lines a—b and c—d somewhat obliquely to the long axis, 4—P, the flagella of the cells repre- sented on the upper side of the diagram would beat more vigorously and accelerate the motion of eat side of the organism. ‘The opposite cell being struck by rays in the direction c—d would be less stimulated, and, as the flagella would beat less strongly than those on the other side of the colony, the organism would swing about until its long axis is brought parallel with the rays when, being equally stimulated on both sides, it would move in a straight course towards the light. We do not have to suppose that each cell makes a special effort to orient itself at a particular angle to the rays, but that it is so organized that the effective beat of the flagella is most accelerated by light striking the cell at a certain angle. If the cells were most stimulated by light falling upon them at such an angle as would result if the rays diverged from a spot in front of the colony and in line with its long axis the con- ditions for orientation would be fulfilled. Since the eye-spots in all the cells face the anterior end of the colony this supposition appears very probable. ‘The foregoing explanation of the orienta- tion of Volvox may or may not be the true one, but it enables us to see a significance in the peculiar arrangement of the eye-spots in this form and is consistent with the results of the experiments we have described.” Is it also consistent with the results of the experiments described in the preceding pages ? In the first place the eye-spots, upon the arrangement of which Hommes places considerable importance in his theory, are not so situated that they all face the anterior end; quite the contrary, they face the posterior end of the colony, as pointed out on p. 107; and in the accompanying diagram by Hormes they should be on the side of the zooids nearest the: end P, instead of on that nearest the end 4. ‘They do, however, probably function as light recipient organs, as already stated (p. 108). Let us then assume that the zooids are influenced by the direction of the rays as Ho.LMEs sug- gests, even if the eye-spots do face the posterior end of the colony, and see if the theory fits our experimental results. 1. It was clearly demonstrated (p. 139) that if specimens of 148 ‘fournal of Comparative Neurology and Psychology. Volvox be exposed to parallel rays of light so that there is a differ- ence in intensity on opposite sides of the organisms when the longitudinal axis is parallel with the rays, they do not move directly toward the source of light but deflect toward the side most highly illuminated. In accordance with HoLMEs’ theory we should expect them to move parallel with the rays under these conditions. 2. Howmes states that the condition for orientation, according to his theory, would be fulfilled “if the rays diverge from a spot in front of the colony in line with its long axis.” If this be true, we should certainly expect the conditions for orientation also to be fulfilled, if the rays converge from two luminous points in front of the organism and if “the eye-spots are most sensitive to light striking them at a certain angle” we should expect the organisms to move toward a point nearly, if not exactly, midway between the two sources of light regardless of their relative intensity. But it has been demonstrated (p.133) that if Volvox colonies be exposed to light from two sources of unequal intensity, they orient and swim toward a point nearer the more intense source. It 1s, there- fore, evident that the explanation of orientation in Volvox, sug- gested by HoLMEs, is not consistent with the experimental results which I have presented. I have demonstrated beyond a reasonable doubt that the dif- ference in intensity on opposite sides of Volvox modifies its direc- tion of motion regardless of the direction of the light rays, and since the direction of motion is changed by difference in the effective stroke of the flagella on opposite sides, it must be differ- ence in intensity which influences the stroke of the flagella. But Ho.MEs, as stated above, concluded that the reaction of Volvox cannot be explained upon the assumption that difference in inten- sity on opposite sides of the body causes the flagella to beat with unequal vigor. Upon what does he base this conclusion and wherein fed the fallacy of his argument ? I can present his line of thought best by quoting verbatim (03, p.. 221-399): “Let us consider a Volvox in a region of sub- optimal stimulation and lying obliquely to the rays of light. If it orients itself to the light the backward stroke of the flagella, 7. e. the stroke that is effective in propelling the body forward must be more effective on the shaded side than on the brighter side. ‘This may conceivably occur in the following ways, which, however, Mast, Light Reactions in Lower Organisms. 149 amount practically to the same thing: the diminished intensity of light on the shaded side of the body may act as a stimulus to the backward p-ase of the stroke, or decrease the efhiciency of the forward phase of the stroke of the flagella; or the light on the brighter side of the body may inhibit the backward phase or increase the forward phase of the stroke of the flagella; In any case, if the organism is passing into regions of ever-increasing intensity of light, we should expect its rate of speed would be lowered. If theorientation 1s affected by a shading of the side away from the light it would follow that in a region in which the shading were less the speed of the travelling body would be diminished. If the parts of the body which are most shaded are the parts where the effective beat of the flagella is the strongest, then, as the organ- ism passes to a point where the illumination on both sides of its body is increased, its rate of transit would be diminished. If we suppose that the forward stroke is most stimulated, or the back- ward stroke most inhibited on the brightest side of the body we should expect that with more illumination the more inhibition there would be, or the more the backward phase of the stroke would be increased, and the rate of locomotion would likewise be reduced. If we imagine a machine in the form of a Volvox colony and provided on all sides with small movable paddles so adjusted that when they come into regions of diminished light as the machine rolled through the water their effective bee would be increased, it is clear that such a machine might orient itself to the direction of the rays and travel towards the source of illumination, but its rate of locomotion would be diminished the brighter the light into which it passed. We may conceive the light to increase or decrease the backward or forward stroke of the paddles in any way we please and we cannot explain how such a machine can orient itself and go towards the light and at the same time move through the water more rapidly as it comes into regions of greater illumination.” It is evident that the crux of this whole argument is the relation between rate of movement and light intensity. “This relation was worked out in detail by Hoimes (’03, p. 323) with the following results: “It was foundthat, as the Volvox travelled towards the light, their movement was at first slow, their orientation not precise, and their course crooked. Gradually their path became straighter, the orientation to the light rays more exact and their 150 ‘fournal of Comparative Neurology and Psychology. speed more rapid. After travelling over a few spaces (centi- meters), however, their speed became remarkably uniform until the end of the trough was reached.”” Unfortunately, HoLMEs does not give the length of the trough, but he says the distance over which there 1s a marked increase of speed 1s considerably less than the space over which the speed is nearly uniform. Ho.tMEs concludes from these results that the increase in rate of speed is due to increase in light intensity and consequently that orientation cannot be due to difference in intensity on opposite sides of the organism, because if it were, the backward stroke of the flagella would have to be more effective on the side in the higher light intensity than on the side in the lower, and this would cause the organism to turn from the source of light instead of toward it. Are these conclusions correct ! If the increase in rate of speed is due primarily to increase in light intensity, one would certainly not expect the rate to become uniform after the colonies have traveled a few centimeters in the trough, nor would one expect it to increase if the colonies are exposed to light of a given intensity for some time. But HoLmes states that the rate does become uniform, and I frequently observed that if relatively quiet colonies in an aquarium containing water a few millimeters deep, are illuminated from above, they oradually become more active. Since, under these conditions, they cannot move toward the source of light, it is evident that this‘increase in activity is not due to increase in light intensity. It is very probable then, that the increase in rate of movement is more dependent upon the time of exposure to light than upon the increase in inten- sity. Moreover, Homes states that orientation is more exact after the colonies have traveled some little distance, 7. ¢., after the rate has become nearly uniform. It must, therefore, be least exact when the increase in rate of speed is greatest. If this be true, it follows that the factors which regulate rate of speed are quite different from those which regulate orientation. We have demonstrated that difference in light intensity on opposite sides of the colonies modifies the direction of movement. And since the factors which regulate the direction of motion and those which regulate the activity of the colonies are different, we may conclude, from this point of view, as well as from what has gone before, that the increase in the rate of speed is not primarily due to increase in light intensity. Such being the case, the argument of HoLMEs Mast, Light Reactions in Lower Organisms. 51 cited above cannot be valid, for it is based upon the supposition that increase in speed in alone: is due to increase in light intensity. We shall refer to this question again (p. 153). Ifa colony which is not oriented turns toward the source of light, it is clear that the stroke of the flagella on the shaded side must be more effective in driving the organism forward than that on the illuminated side. ‘This may be conceived to be caused directly by the difference in light intensity on opposite sides, or indirectly in that a Volvox colony may possibly act as a lens and thus cause the light on the side opposite that most highly illumi- nated to become most intense; or, since the zooids are intimately connected by protoplasmic strands, it is not impossible that impulses produced by excessive photic stimulation may be trans- mitted to the opposite side and result in action there. At any rate, it 1s undoubtedly true that these strands serve to transmit impulses from zooid to zooid, and thus bring about coordinate action. It was found, as previously stated, that segments, e. g., halves produced by cutting specimens parallel to the longitudinal axis, orient essentially like normal colonies. Such segments, however, cannot act as lenses, nor can impulses originating on one side be transmitted to the opposite side. The last two of the possible explanations suggested, therefore, must be abandoned, and it must be concluded that the unequal effect of the stroke of the flagella is due directly to difference in light intensity on opposite sides of the organism. But this unequal effect of the stroke on opposite sides may be caused, as HoLMEs pointed out, by an increase in the backward phase of the stroke on the shaded side, or a decrease in the same phase on the illuminated side or a decrease in the forward phase on the shaded side, or an increase in this phase on the illuminated side. Can it be ascertained which of these is the cause of the difference between the effect of the stroke of the fla- gella on the shaded sides and that of those on the illuminated side of the colonies ? If the light intensity of the field is suddenly decreased while colonies of Volvox are swimming horizontally toward it, they stop forward motion, the longitudinal axis takes a vertical position due to the effect of gravity, and then the colonies swim slowly upward. It is not at all difficult to find specimens in which this upward swimming is just sufficient to overcome the effect of gravity, and 152 ‘fournal of Comparative Neurology and Psychology. under such conditions they appear to be hanging in the water motionless. ‘They are, however, rotating on their longitudinal axis. Ifnow the light intensity, to which these apparently motion- less organisms are “exposed, is increased they soon begin to turn toward its source; but in so doing they swim upward, as repre- sented in the accompanying diagram. In thus swimming upward and horizontally toward the source of light, it is clear fat the effect of the backward stroke of the flagella i increases both on the shaded side and on the illuminated side, for both sides move forward. But the shaded side moves farther than the illuminated side, consequently the increase in the effect of the backward stroke must be greater on the former than on the latter. ‘The difference in the effect of the stroke of the Fic. 13. Diagram representing the reaction of a Volvox colony when the light intensity is suddenly changed. a, outline of colony; b, longitudinal axis; c, light rays; d, point in the course where the light is suddenly decreased; e, point where it is suddenly increased; f, course taken by colony. In continuing from e, the side of the colony facing the source of light travels over a shorter distance than the shaded side. Consequently the backward stroke of the flagella on the latter side must be more effective than that of those on the former. flagella on opposite sides which results in orientation of positive Volvox colonies is, therefore, due to a greater increase in the back- ward stroke of the flagella on the shaded side than of those on the illuminated side. If the light thrown upon apparently motionless colonies is quite intense, they frequently may be seen to sink 4 or 5 mm. immedi- ately after the light is turned on, but while they are sinking this short distance, they apparently become acclimated and soon turn toward the light, and at the same time swim upward, just as described above. During the time in which these colonies sink they continue to rotate in the same direction as before. The Mast, Light Reactions in Lower Organisms. 153 sinking must then be due to a decrease in the effect of the backward stroke of the flagella on all sides, and this decrease 1s due to an increase in light intensity. But when the colonies turn toward the source of light, and at the same time swim upward, it 1s evident that the increase in light intensity must cause an increase in the backward phase of the stroke of the flagella on all sides, for if this were not true there could be no upward motion. ‘The side nearest the source of light, however, passes over a shorter distance than the opposite side, as will readily be seen by referring to the dia- gram, and therefore the increase in the effect of the backward phase must be greater on the latter than on the former. But the light intensity 1s greater on the former than on the latter (a paradox). When the light intensity in the field is increased the effect of the backward phase of the stroke of the flagella may be increased or decreased on all sides. If itis increased the effect is most marked on the side in lowest light intensity. Furthermore, if the light is strong the colonies turn toward its source more rapidly and do not swim upward so far and thus make a sharper curve than when it is weak; but the stronger the light the greater the difference between the intensity on the shaded and that on the illuminated side. It, therefore, follows that the greater the differ- ence in intensity on these sides, the greater the difference in effect of the backward phase of the stroke of the flagella, the effect being greatest on the side least illuminated. Teas considerations sup- port the conclusion arrived at above, 7. ¢., that the factors which regulate the activity of the colonies, as a aliales are different from those which regulate the direction of motion. We have thus demonstrated that while orientation is due to difference in light intensity on opposite sides of the colonies, it is brought about in positive specimens by the flagella striking backward with greater effect on the side in lowest light intensity than elsewhere. I suggest the following explanation of this: First, it must be remembered that the organism constantly rotates on its longitudinal axis. If then a colony is so situated that one side is more highly illuminated than the opposite, it is clear that the zooids will constantly be carried from a region of higher to a region of lower light intensity, and vice versa. They are thus subjected to constant changes in strength of illumination. As stated above, the flagella strike backward with greater vigor on the shaded side than on the opposite one and, therefore, it is 154 fournal of Comparative Neurology and Psychology. evident that as the zooids reach the region of lower light intensity, in other words when the light intensity to which they are subjected decreases, they increase the effect of the backward stroke of the flagella, 7. ¢., they attempt to turn toward a structurally defined side (the side facing the anterior end of the colony). ‘This is pre- cisely what Euglena does when it passes from a region of higher to one of lower light intensity, 7. ¢.,-1t turns toward a structurally defined side, the larger lip. The individuals in a colony then respond with a motor reaction induced by change in light intensity; they react on the same basis as do Euglena, Paramecium, Stentor and other unicellular forms, in their trial and error reactions, but owing to the way in which they are inter-related, and to the rota- tion of the colony on the longitudinal axis, this reaction of the zooids causes orientation in the colony as a whole, without error. This explanation of orientation in entire colonies holds also for orientation in segments. As previously stated, only those seg- ments orient which have such a form that they can rotate. As they rotate the cut surface constantly faces the center of the spiral, so that if the axis of the spiral is not directed toward the source of light, the outer surface where the zooids are situated is alternately turned toward the light and away from it. ‘Thus the zooids are carried from regions of higher to regions of lower light intensity and vice versa, and the motor reaction is induced just as it is in entire colonies. Orientation in negative colonies can be explained 1 in precisely the same way as that in positive ones, assuming merely that in this condition the zooids respond with the motor reaction when they pass from lower to higher light intensity instead of when they pass from higher to lower (as i is true when the organisms are posi- tive). The backward stroke then becomes most effective on the side most highly illuminated. II. REACTION OF NEGATIVE COLONIES. Volvox becomes negative when exposed to light of a certain intensity. ‘[he intensity, however, varies greatly in different colo- nies and in the same colony under different conditions. RapL ('03, p. 103) concludes his discussion on the difference between positive and negative phototropism with the following paragraph: “Ich glaube nun, dass der Unterschied zwischen positivem und negativem Phototropismus 4hnlich wie beim Menschen nicht ein Mast, Light Reactions in Lower Organisms. 155 Wntersehied in der Orientierung, sondern nur in der Lokemotion ist; dass das Tier in beiden Fallen gegen die Lichtquelle orientiert ist, jedoch nicht gleiche Muskeln spannt.”’ This explanation will not hold for Volvox or Euglena, for both of them turn the anterior end from the source of light when they are negative. When Volvox colonies are negative they orient in all essentials as they do when positive, except that they direct the anterior end from the source of light. In swimming horizontally from a source of light they seldom move parallel with the light rays. If the position of the light 1s changed after they have oriented, they change the direction of motion until the course again bears the same relation to the ray-direction it did before. If exposed to -light from two.sources, so arranged that the rays make a definite angle with each other, they move from a point between the two. If one source is more intense than the other, the point from which they move is nearer that source. These facts and others are established by the following experi- mental results, which are presented in graphic form (Fig. 14). By referring to path @ it will be seen that the colony introduced at m was positive to light from the three glowers as well as to that from the arc, but that it became negative after swimming toward the arc for a short distance from c, turned about and moved across the aquarium to c’. That is, at the end of the experiment the colony was negative to a much lower light intensity than at the beginning. The arc was approximately 250 candle power. It was I5 cm. from the point where the organism became negative. The light intensity at this: point was therefore 11,111 + candle meters. But the colony was still negative after beaae crossed the aquarium, a distance of nearly 8 cm., or nearly 23 cm. from the arc, 7. e., in an intensity of 4726 + Sadie meters, which is 6385 + candle meters less than the intensity in which it first be- came negative. Similar results are represented in path B and the paradoxical nature of the results is even more striking than in the case of path 4. Unfortunately, the distances between the sources of light and the aquarium, in this exposure, were not recorded. The colony which produced path B was positive to the light from the arc when first put into the aquarium at c, but after moy- ing toward the source of light a few centimeters, it became negative, 156 Fournal of Comparative Neurology and Psychology. turned about and moved in the opposite direction. When it reached c’ the glowers were exposed and the colony promptly changed its direction of motion and proceeded on a course directed from a point between the two sources of light. ‘This point, how- ever, was much nearer the arc than the glowers, the light fromthe Le Fic. 14. The lines 4 and B represent the course taken by single colonies as seen in water 2 cm: deep in the plate glass aquarium, e (the paths are represented in approximately accurate propor- tions); g, a group of three 222 volt Nernst glowers in a vertical position; a, carbon arc; f, direction of light rays; d, opaque screens; n n’, path with glowers exposed and arc shaded ; cc’, path with arc exposed and glower shaded; ¢c’ n, path with both glowers and arc exposed. former being much more intense than that from the latter. When the light from the arc was cut off at 1, the colony was found to be negative to the comparatively weak light from the glowers. It consequently changed its course and moved from this source; but Mast, Light Reactions in Lower Organisms. S97 after continuing about 3 cm. it became positive, turned about and moved toward the glowers to n’, and probably would have con- tinued farther had it not been prevented from doing so by the wall of the aquarium. It will be noticed that the point n’, where the colony was still positive at the end of its course, was about 3 cm. nearer the glowers than n, where it proved to be negative, and nearly 7 cm. nearer than the point where it changed its course from negative to positive. That 1 iS, the organism was positive at n’ in a much higher light intensity than oer in which it was negative at 7 aah at the. point where it changed from negative to positive. 12. CAUSE OF CHANGE IN SENSE OF REACTION. The results presented above demonstrate that Volvox may be either negative or positive in a given light intensity. ‘This will be brought out more clearly later where it eal be shown that Vol- vox, In certain conditions, is negative to light of all intensities to which it responds at all. Since a Volvox colony may be either positive or negative in the same environment, it is clear that the transformation from positive to negative or vice versa must be due to some internal change. This change, whatever it may be, is induced by light. It is dependent upon nthe intensity and also upon the time of exposure, as 1s shown by the fact that when specimens are exposed to intense light they may be positive for a time and then negative to a much lower inten- sity than that in which they were positive when first exposed. Weak light tends to induce the change which causes the colonies to become positive, whereas strong light tends to induce the change which causes them to become negative. Some photosynthetic process in chlorophyl bearing organisms, suggests itself as the probable condition upon which the sense of reaction depends. It might be assumed that the organisms are positive when a given amount of synthesized substance, such as carbohydrates, proteids, or fats, 1s present, and negative when this amount is decreased. ‘This assumption fits fhe observed reaction in that such substances are formed in the presence of light, and in that they disappear in darkness, being either further synthesized to form protoplasm, or, perhaps, directly oxidized. But the short time and the slight change in light intensity necessary to produce a change in the sense of reaction is entirely inadequate 158 Journal of Comparative Neurology and Psychology. to cause the formation and destruction of photosynthetic sub- stances, such as those mentioned above. ‘The inversion of the sense of reaction, therefore, cannot be due to a photosynthetic process. May it not be due to the effect of light on the chemical equilibrium of some other substance ! One of the more important results of recent investigation in physical chemistry is the establishment of the fact that substances in chemical equilibrium are dynamic and not static, as had for- merly been supposed. If, for instance, alcohol be added to acetic acid, it is well known that water and ethyl acetate will be formed; but it is also true that if water be added to ethyl acetate, the form- ation of alcohol and acetic acid results, that is, the former reaction is reversed. When the reaction in both of these cases has reached a state of equilibrium, there is a certain amount of each of the following substances present: Alcohol, acetic acid, ethyl acetate, and water, and this amount remains constant; but the reaction continues; alcohol and acetic acid react to form ethyl acetate and water just as fast as ethyl acetate and water react to form alcohol and acetic acid. ‘These reactions are expressed as follows: C.H,OH +HOOC- Cher. Coo C.H,+10. This indicates that two reactions are taking place simultaneously in opposite directions. ‘The relative amount of substance indi- cated in the two members of an equation representing equilibrium in chemical reaction, depends upon the nature of the substances and the environment, 1. e.; the temperature, pressure, etc. If, for instance, the temperature oe compounds in equilibrium be raised, the equilibrium will be destroyed and the reaction in one direction will take place faster than that in the other. When equilibrium is again restored the relation of the amounts of the different sub- stances will no longer be the same as it was at the lower tempera- ture. If the temperature 1s lowered, the rate of motion will increase in the opposite direction. JONES (02, p. 514) states this as follows: ‘The effect of a rise in temperature is to favor the formation of that system which absorbs heat when it is formed. . . Increase in pressure diminishes the volume and, therefore, favors the formation of that system which occupies the smaller volume.” Reversible chemical reactions were formerly supposed to be quite exceptional, but it is now known that they are not. JONES (02, p. 481) writes: “We must regard chemical reactions in general as reversible.”’ Mast, Light Reactions in Lower Organisms. 159 No work, as far as I know, has been done directly on the effect of change in light intensity on equilibrium in chemical reaction; but we know that light does affect many chemical reactions, and since we must regard chemical reaction in general as rever- sible, it seems reasonable to assume that the relative amount of different substances present in a mixture is dependent upon the light intensity, provided the chemical reaction between the sub- stance is at all affected by light. [his means that substances in chemical equilibrium in one light intensity will not be in equi- librium in another, so that the direction in which the reaction takes place faster depends upon the light intensity. To explain reversal in the sense of reaction on the basis of chemical reactions induced by light let us assume: (1) That Vol- vox contains substances X and Y, the chemical reaction between which is regulated by the intensity of light; (2) that a sub-optimum intensity favors the formation of calgantes represented by X and a supra-optimum those represented by Y’; and (3) thatthe colonies are neutral in reaction when there are Y substances in one member of the equation and X in the other; positive when one member con- tains (X +) substances and the other (Y —), and negative when one contains (X —) andthe other (1! +). Can the change in sense of reaction as represented in paths 4 and B, Fig. 14, p. 156, be explained on the basis of these assumptions ? The colony which produced path 4 was positive when put into the aquarium at n. In accordance with our assumption it, there- fore, contained (X +) and (Y —) substances. ‘The intensity at n was relatively low so that the chemical reaction favored the formation of compounds represented by X. ‘This may be expressed thus (X +) S(V -), indicating that the reaction toward X takes place Pigue than that toward Y. The increase in the X and decrease in the Y substances continued until a state of equilibrium was attained or the organism reached n’ and c, where the light from the glower was turned off and that from the arc turned on, and the colony was thus exposed to light of supra-optimum inten- sity. Why did it not then turn from the source of light at once? According to our assumption, because it contained (X ar) and (Y —) substances. But since the colony was in a supra-optimum intensity, the chemical reaction favored the formation of Y sub- stances at the expense of X, represented thus (X +) & (V -). As soon as this reaction had continued far enough so that (X +) 160 Fournal of Comparative Neurology and Psychology. was decreased to X and (Y —) increased to Y, the colony became neutral. ‘The point where this took place is represented in the path by the sharp curve. But why did the colony not remain neutral? Because it was ina supra-optimum light intensity and, therefore, in accordance with our assumption, X continued to decrease and Y to increase, X¥2<&Y resulting in (X —) and (Y +) compounds which caused the organism to become negative and it remained so to the end of its course. Had the aquarium been wider it would have reached a point at which it would have been neutral in an optimum light intensity. If the reactions are regu- lated as assumed, it would have reached this point as follows: (X —)S (1 +) expresses the condition of the colony as it pro- ceeded from the source of light toward c’, but as the intensity decreases the rate of formation of X increases and that of Y de- creases until the colony reaches the point of optimum intensity, when the rate in opposite directions is equal (X —)=@(Y +). The organism, however, is still negative at this point, since it con- tains eg —) and (Y 7) substances, and it therefore proceeds into a region of sub-optimum intensity, where (X =) increases and (Y +) decreases (X sae e ah). This results in X and Y\sub- stances and the colonies consequently become neutral. The chemical reaction, however, continues to favor the formation of X, since the light is sub-optimum, and this soon results in (X +) and (Y 2 substances, which causes the organism to become positive. It therefore turns and proceeds toward the source of light again, but owing to the accumulation of (Xx +) and (f —) substances, it passes the region of optimum intensity before it becomes neutral, and therefore becomes negative again. It may be conceived to thus pass back and forth several times, like a pendulum, before being neutral in the optimum region. In accordance with our assumption, the conditions of the colony in producing the path B could be represented as follows: (X +)S(V -—) from c to the beginning of the curve; (X) S(1) atthe point m the curve nearest the arc; (X —)S(V +) from this point to n; (X —)S(¥ +) from n tothe beginning of the curve beyond; (X) and its external medullatedlamina. The bundles running to this nucleus represent the main tract to the cortex, for from the lateral surface of the ventro-lateral nucleus a large bundle of well-medullated fibers passes to the cortex of the lower part of the posterior central gyrus. The sensory path is almost entirely broken in the lateral nucleus. The dorsal segment of the medial lemniscus, Foret’s Bath, passes to the centre médian of Luy and to the cup-shaped nucleus. The motor path is medullated only down to the cerebral peduncle. It makes connections with the lenticular nucleus, the hypothalmic nucleus and the substantia nigra. The model shows that the motor and sensory paths are parallel throughout the brain stem. ‘The sensory path is always dorsal to the motor. In the medulla, the paths are-adjacent, in the pons, they are separated somewhat by the cells of the nuclei pontis; in the midbrain and hypothalmic region, the substantia nigra and hypothalmic nucleus come in between the two paths, while in the thalamus and subcortical regions, they are adjacent. In the thalamus, the nuclei that contain medullated fibers from the medial lem- niscus are the lateral nucleus, the centre médian of Luy, and the cup-shaped nucleus. The lateral geniculate body and the pulvinar have medullated optic fibers, while the medial and anterior nuclei contain no medullated fibers whatever. ‘The fascicu- lus retroflexus stands out very clearly, however, passing through the lower border of the medial nucleus. Development of the Inter-forebrain Commissures in the Human Embryo. By GrorceE L. STREETER, Wistar Institute of Anatomy, Philadelphia. A morphological study of the corpus callosum, anterior commissure and the commissure of the hippocampus, based on a series of wax-plate reconstructions of human embryos varying from 80 to 150 mm. in length. All three structures cross the median line in that portion of the brain wall developed from the lamina terminalis. In 80 mm. embryos, the corpus callosum consists of a round bundle of fibers lying directly on the commissure of the hippocampus, representing the condition found in non-placental animals. The succeeding growth consists in the lengthening of the fornix and caudal migration of the hippocampal commissure, the latter remaining in close relation to the caudal end of the corpus callosum, which, in the meantime, by increase in number of fibers has extended anterior to the anterior commissure and posterior so as to deck over the region of the third ventricle. The formation of a cavity in the septum lucidum occurs in embryos of about 95 mm. The anterior or olfactory division of the anterior commissure does not enter the olfactory bulb but is traced to the cortex dorsal to the bulb. The Relations of the Frontal Lobe in the Monkey. By E. Linpon ME tus, Anatomical Laboratory, ‘fohns Hopkins University. The Terminal Distribution of the Eighth Cranial Nerve in Man. By Joseru H. Hatuaway, Cornell University Medical College, Ithaca, N. Y. Statistical Studies of the Brachial Plexus in Man (A Preliminary Note). By AsraM T. Kerr, Cornell University Medical College, Ithaca. These studies are based on the record of 175 plexuses dissected and drawn by students, mostly in the Johns Hopkins Medical School, some in Cornell. The records were verified in order of time by Drs. ELtiINc, BARDEEN, and Kerr. This 184» fournal of Comparative Neurology and Psychology. note deals with the type of plexus according to the origin and combination of branches. Seven types are recognized: A. With outer cord formed from the 4th to the 7th nerves inclusive, the inner from the 7th to the gth, the posterior from the 4th to the 8th, occurred in .57 per cent of cases; B, like A except posterior cord which was formed from the 4th to the gth in 1.71 per cent of cases; C, with outer cord from 4th to the 7th, inner from 8th to the gth and posterior from 4th to oth, in 58.28 per cent of cases; D, like C except outer cord from 4th to 8th in 2.28 per cent of cases; F, with outer from 5th to 7th, inner from 8th to gth, posterior from 5th to gth, in 29.77 per cent of cases; G, like F, except the 5th sends a branch to the 4th, in 6.85 per cent of cases; H, like G, except the outer cord from the 5th to the gth, in .57 per cent of cases. No record was made of the relation of the roth spinal nerve to the plexus. Attention was also called to the outer head of the ulnar nerve which when searched for will be found in over 50 per cent of the cases. A Racial Peculiarity in the Temporal Lobe of the Negro Brain. By Roserr BENNETT? Bean, University of Michigan. Measurements were made of 236 temporal lobes, 54 from white brains and 182 from negro brains. Six measurements were made from fixed points on each tem- poral lobe, at three levels, two at the base (antero-posterior and transverse), 2 at I cm. toward the pole, and 2 at 5 mm. from the tip of the temporal lobe. At each level, the lobe is smaller in the negro, more nearly approaching the white in size at the base, and diverging from the white in size as the pole is approached. The widest divergence i is found in the transverse measurement about the middle of the temporal lobe. There is a variable difference with a mean of about 5 mm. in each measurement. ‘The temporal lobe of the negro is more slender than that of the white, narrower transversely, and more pointed at the extremity, the length being about the same in each race. The differences in general conform to the other characteristics previously described in the negro brain. Supplementary Report regarding the Innervation of the Leg of Rana virescens. By Euizasetu H. Dunn, Department of Anatomy, University of Chicago. A partial report upon this material was made at the 1905 meeting of the Associa- tion. The histological examination of the muscle on the operated side in which all the efferent neurones were degenerated, shows but a slight change from the normal. ‘The cross striations were not obliterated and the nuclear staining was unaltered. The staining with acid dyes was slightly less deep in the muscles of the operated side. The bulk of muscle seemed unchanged. As the counting was done upon osmic acid material, only medullated nerve fibers entered into the enumerations. On the intact side, both afferent and efferent fibers were present. On the operated side, the efferent fibers had been eliminated, and the count was of afferent fibers destined for the skin and for the muscles. Among both the afferent and the efferent fibers, splitting occurred. ‘This splitting was in the main trunks of the various segments of the leg. In both classes, the proportion of splitting fibers increases progressively from the thigh to the foot. A greater amount of splitting occurred among the purely afferent fibers of the operated side than among the mixed nerves of the intact side. Hence the propor- tion of splitting afferent fibers is greater than that of splitting efferent fibers. “This splitting seems to occur in both the cutaneous and muscular afferent fibers. This result carries with it some important physiological corollaries, as for instance that T he Mid-Winter Meetings. 185 some afferent neurones must have two “local signs”’ according to the point at which they are stimulated. In considering the distribution to the various segments, correction must be made for the splitting fibers. When this is done, it is found that the unit area of skin received the same num- ber of afferent fibers in either the thigh, shank or foot. In a similar way, a study of the muscular afferent fibers shows that the muscles are uniformly innervated by muscular afferent fibers according to the unit weight of muscle. The Electric Organ of Astroscopus as Compared with that of Other Fishes. By Cuar es F. Sirvester, Princeton University. Concerning a New Ganglionic Mass of the Hind-brain, the Corpus Ponto-bulbare. By Cuartes R. Essick, Student of Medicine, ‘fobns Hopkins University. A ganglionic mass accompanied by a layer of myelinated fibers is found over- lapping the restiform body just caudal to the dorsal cochlear nucleus. It forms a direct lateral process or extension of the ganglion mass of the pons. It is con- stantly present in all human brains, though in some brains it reaches a greater size than in others. Its relations and general characteristics are constant. It makes its appearance on the ventro-lateral surface of the pons near the emerging root bundles of the trigeminal nerve and extends backward, passing between the roots of the facial and acoustic nerves. It continues caudalward passing dorsal to the glossopharyngeal nerve and ends on the dorsal surface of the restiform body, forming part of the lateral boundary of the fourth ventricle. It may end as a prom- inent tongue-shaped mass or may spread out as a thin coating of the restiform body that is only discernible microscopically. The Migration of Medullary Cells into the Ventral Nerve-roots of Pig Embryos. By F. W. Carpenter and R. C. Matn, University of Illinois. In sections of pig embryos 11 mm. long, medullary cells were observed appar- ently migrating from the neural tube in company with the fibers of the ventral nerve roots. ‘These cells have been found just inside the external limiting mem- brane in an intermediate position half in and half out of the neural tube and in the base of the nerve root just outside the limiting membrane. A few sections show continuous lines of medullary cells, touching end to end and reaching from the nidulus across the boundary of the tube into the proximal part of the nerve root. These migrant cells do not appear to be directly connected with the embry- onic nerve fibers. A few were observed undergoing mitotic division. Evidence of a similar migration of medullary cells has been seen in sections of a cat embryo. In these medullary elements escaping from the neural tube, we recognize the ‘indifferent cells” of SCHAPER. Such cells remaining in the medullary wall become either supporting elements (neuroglia cells) or nervous elements (neurones). ‘Those which escape, in part at least, probably contribute to the formation of the sheaths of SCHWANN, which are supporting in function. Whether any migrant indifferent cells become the nerve cells of sympathetic ganglia we are at present unable to say. ‘ Glycogen in the Nerve Cells of the Brain and Spinal Cord of Larval Lampreys and in the Central Nervous System of the Amphioxus from Naples. By Smmon H. GacgE, Cornell University. 186 Fournal of Comparative Neurology and Psychology. At the ninth annual meeting of the American Phystological Soctety, the follow- ing papers of special interest to neurologists were presented: Functions and Structures in Ameeba proteus. By C. F. Hopce and O. P. DEL- LINGER. The material of this study consisted of Amceba which had been killed in vari- ous ways and stained 1m toto, and of series of sections stained by different methods. The function of contraction is effected by a meshwork of fibrillz, these being woven into heavy trabeculae with wide intertrabecular spaces in the interior (endosarc), and united to form a fine web over the exterior (ectosarc). “This mechanism is all that can be demonstrated to account for movements: locomotion, ingestion, egestion, contraction of vacuole, division of nucleus and of entire body, and internal circulation. ‘This tissue forms, without essential differentiation, the outer mem- brane (ectosarc), trabecul:e, wall of contractile vacuole and of all food vacuoles, nuclear membrane and nuclear reticulum. Movements are coordinated, but no differentiation of conducting fibrilla has been clearly demonstrated. ‘This tissue, when supplied with necessary nutrient substances, must possess the function of growth. ‘The function of digestion is mediated in all animals by gland cells, charac- terized by zymogen granules. The only structure in Amoeba which is definitely granular is the nucleus. Sections show nuclear granules apparently passing out of the nucleus into the food vacuoles. “These two mechanisms, together with a circu- lating fluid, account for all the functions of the animal, reproduction being undif- ferentiated from growth, and respiration, excretion and circulation being effected by movements of the whole body and supplemented by a similar action of the con- tractile vacuole. Except as noted above, sections reveal no difference between ectosarc and endosarc. Physiological Reactions of Physa. By J. Dawson. Vasomotor Reflexes. By W. T. Porter. The Cause of Treppe. By F.S. Lrg. Methods of Studying Fatigue. By F. S. Lee. The Functions of the Ear of the Dancing Mouse. By R. M. YERKEs. Both the static and the acoustic functions of the ear of the dancer differ markedly from those of the common mouse. Orientation and equilibration are fairly good. There is no evidence of turning dizziness. “The whirling movement which is char- acteristic of the race appears as soon as the young mouse is strong enougl to stand. It is somewhat more pronounced in the female than in the male, and it occurs chiefly toward evening. With respect to this movement there are three well defined groups of dancers: those which almost always whirl toward the right, those which whirl toward the left, and those which whirl now one way now the other. At present, I have no satisfactory evidence of the inheritance of the tendency to whirl in a certain way. Direct and indirect methods of testing acoustic sensitiveness have given nega- tive results in the case of the adult, but the young dancer responds to certain sounds for from two to five days during the third week of life. This brief period of sen- sitiveness to sound is preceded by a marked change in behavior. The Effect of Section of One Vagus upon the Secondary Peristalsis of the Gisopha- gus. ByS. J. MELTzER and J. AUER. The Mid-Winter Meetings. 187 On the Alleged Adaptation of the Salivary Glands to Diet. By F. P. UNDERHILL _ and L. B. MENDEL. The experiments reported corroborate the current statement that the saliva of dogs and cats is devoid of amylolytic properties. Ne1Lson and TERRY (American Fournal of Physiology, vol. 15, 1906, p. 406) have announced that dog’s saliva frequently will digest starch; and they claim to have increased the amylolytic power by appropriate (carbohydrate) feeding. An inspection of their data, how- ever, indicates that the amylolytic activity is not very pronounced when compared with that ordinarily observed in human saliva. We have failed to note similar results in ordinary dogs and to obtain adaptation in animals maintained ona _ special diet rich in starchy foods. Special attention was devoted to the conditions under which the digestion trials were conducted. After the presentation of this paper Dr. NerLson stated that he had additional evidence obtained since the publication of his paper with Dr. Terry, of adaptation of the salivary glands in man as well as in dog. Some Observations on the (Esophagus after Bilateral Vagotomy. By W. B. Can- NON. On the Mechanism of the Refractory Period in the Heart. By A. J. Cartson, University of Chicago. One series of the experiments was directed toward determining whether the refractory state of the heart is a property of the heart muscles or the nervous tis- sue or both. ‘The other series aimed to determine the degree of refractory state exhibited by the different parts of the vertebrate heart. If there is a causal con- nection between automatism and a refractory state in the sense of an absolute inexcitability, we ought to find this condition in the primum movens of the heart, the sinus, or in mammals, the mouth of the great veins; and we would expect a less degree Bi refractory state or even none at all in parts of the heart not automatic, for example, the tortoise ventricle or the apex of the frog ventricle. J. Lhe Tissues of the Heart Concerned in the Property of the Refractory State. 1. ‘The automatic heart ganglion of Limulus exhibits the typical refractory period of the heart or a state of reduced excitability during systole. 2. As long as the heart ganglion is in physiological connection with the heart muscle, the heart muscle and nerve plexus exhibit a condition of reduced excit- ability at the beginning of systole. ‘This result is obtained from all regions of the heart. A stimulus strong enough to produce an extra beat by acting on the heart muscle and nerve plexus fails to produce any visible effect when sent through the same tissues, at the beginning of the normal systole. 3. Do the Limulus heart muscle and motor nerve plexus exhibit a systolic refractory state after being severed from the ganglion? ‘The results of the experi- ments were not conclusive, mainly because of the difficulty of getting a series of contractions of absolutely uniform amplitude from the Limulus heart after the nerve cord is removed. 4. Is the refractory state in the vertebrate heart a property of the heart muscle apart from the intrinsic heart ganglia and nerve plexus? (1) It is needless to say that this question has not so far, and perhaps never can be, attacked by direct experiments. RoHDE and ScHuLtz have attempted to settle this question by studying the action of chloral hydrate on the heart. But in chloral hydrate nar- 188 Fournal of Comparative Neurology and Psychology. cosis a systolic refractory period in the sense of diminished excitability is in evidence as long as the heart retains excitability and contractility. (2) Nerve tissue prob- ably dies sooner than muscular tissue when circulation and nutrition are stopped. If the refractory state depends on the nervous tissue in the heart, the excised or dying heart ought to exhibit a condition of diminished or abolished refractory state in the later stages while it still retains some excitability and contractility. But even in the last stages of dying the ventricular tissue or tissues retain the prop- erty of systolic refractory state in the sense of diminished excitability. (3) The sodium chloride rhythm of the ventricular apex is probably idio-muscular. In case the refractory period is a property of the nervous tissue alone we might expect a diminution or abolition of the systolic refractory state in this rhythm. But fresh ventricular strips exhibit practically as marked refractory state in the sodium chloride rhythm as in the normal rhythm. It is therefore evident that the question whether heart muscle when tsolated from the intrinsic nervous tissues exhibits the property of refractory state to greater degree than skeletal and smooth muscle 1s still an open one, since the facts bearing on the question can be interpreted either way. In Limulus the heart ganglion exhibits the typical refractory period of heart tissue, and as this is a characteristic of at least many ganglion cells in the central nervous system of vertebrates, it is probable that the ganglion cells in the vertebrate heart possess a systolic refractory state similar to that of the Limulus heart ganglion. Il. Lhe Degree of Refractory State in the Heart of Different Animals, and in the Different Parts of the Heart of the Same Animal. 1. ‘The ventricles of higher vertebrates do not exhibit the same degree of refractory state. A strong induction shock sent through the ventricle at the beginning of systole produces a super- maximal beat in the frog, toad and salamander, but not in the tortoise. In the case of the latter the strong induction shock diminishes the amplitude of the beat. But inasmuch as the inhibition of a phenomenon is just as much evidence of irrit- ability as the augmentation of it, it is obvious that even the tortoise ventricle is excitable at the beginning of systole. ‘The refractory state in the heart is therefore a condition of diminished excitability and not a state of absolute inexcitability. 2. The degree of refractory state is not necessarily the same in the different parts of the heart of the same animal. In one tortoise (Cistudo) supermaximal beats are readily produced in the sinus venosus by stimulation at the beginning of systole, while the ventricle responds to the same stimulation with a diminution of the beat. 3. There is probably no causal connection between the property of automatism and the property of refractory state, for the following reasons: (1) The property of refractory state is exhibited by tissues that do not have the property of auto- matism under normal conditions—the apex of the frog and tortoise ventricle, nerve centers or ganglion cells in the central nervous system, the mammalian intes- tines (Magnus), the nerve plexus and heart muscle of Limulus. (2) Some tissues that are normally automatic do not exhibit an absolutely refractory state, but only a condition of greatly diminished excitability—the hearts of invertebrates, the hearts of many vertebrates. (3) In the same heart the parts possessing the great- est degree of automatism may exhibit a less degree of refractory state than the part of the heart not automatic. The Mid-Winter Meetings. 189 At the fifteenth annual meeting of the American Psychological Association, the following papers in the fields of comparative psychology and sense physiology were read: The Photography of Ocular Movements. By G. M. Stratton. The Rotation of the Eye during Fixation and in Movement. By C.N. McA isteEr. Studies in Binocular Depth Perception. By J. C. BELL. Some Results of Experiments on Cerebral Circulation during Sleep. By J. F. SHEPARD. This was a preliminary report on volume reactions during sleep while the sub- ject was lying down. ‘Two subjects were used. For part of the records the influ- ence of movements was eliminated by placing the subject’s head in a swing. With the first subject, the volume of the brain and of peripheral parts increases as he goes to sleep, and decreases as he awakes. ‘There is often a temporary fall of the brain volume preceding the more marked rise which shows itself as sleep becomes deeper. ‘There is a prominent breathing wave in the records from both brain and periphery. In this wave, the fall in the circulation record very nearly corresponds to an inspiration, the rise to an expiration. Stimuli that disturb but do not awake the subject cause a temporary increase in breathing in both chest and abdomen, a fall of volume of the brain and peripheral parts with comparative elimination of the breathing wave therein. While the subject is sleeping soundly and there are apparently no distinct stimuli acting, one often finds a more or less rhythmic repetition of such changes, analogous to the TRAUBE-HERING wave. There is always some evidence of this wave, and the changes in brain and periphery are always parallel. If discussed in detail, these statements would require some modification. With the second subject the results have not been so definite. ‘There is usually a distinct increase of volume of the brain when he goes to sleep, and in several cases there is a marked fall with awakening. The TRauBE-HERING wave in the volume and in the breathing is not so prominent, but is still present; and its rela- tions are the same. The breathing wave in the brain curve, on the other hand, often seems to follow the depth of breathing, and to be larger while the subject is awake. ‘The variations may be due in part to the fact that the subject was more nervous, and never slept very soundly nor very long during the experiments. The Difference between a Habit and an Idea. By S. H. Rowe. The Relation of Imitation to the Theory of Animal Perception. By G. H. Mean. Kinesthetic and Organic Sensations: ‘Their Role in the Reactions of the White Rat tothe Maze. By Joun B. Watson. The work here reported grew directly out of the experiments carried out some years ago by SMALL at Clark University. In our experiments a maze very similar to the one used by SMALL was adopted. ‘The floors and sides of the galleries, how- ever, were made of four inch boards instead of wire netting. Very gentle male rats were used in all crucial experiments. An accurate record was kept of the forma- tion of the maze association, consisting of the variations in time, number of Igo Fournal of Comparative Neurology and Psychology. errors,and so on. The normal behavior in learning the maze was first observed. Nineteen normal rats were used for this purpose. Tests were then made to determine the principal sense organs used in learning the maze. Normal rats which had previously learned the maze in the light, were tested with the maze in the darkness. All the rats with one exception, ran the maze perfectly in the dark. Normal animals were then allowed to learn the maze in the dark from the beginning. ‘Their records were quite normal. It was found that animals trained to the maze in the light can run the maze almost perfectly after both eyeballs have been extirpated. Likewise untrained rats, with both eyeballs removed, can learn the maze in normal time. Animals without olfactory bulbs learn the maze in normal time, two of our animals making phenomenal records. ‘These anosmic animals after being trained to the maze in the light are not at a loss if forced to run the maze in the dark. From further experiments, which we shall not report in detail, it becomes evident that neither auditory sensations nor cutaneous sensations set up in the drum mem- brane by the changes in the pressure of the various air columns in the maze play a part in the formation of this association. Likewise it isimprobable that the discrimi- nation of the correct turns in the maze is effected by means of differences in the contact values of the various parts of the floor of the maze. ‘The vibrissz, which are probably used in the beginning of the association, are not used after the association has been thoroughly established; for if after their removal a short period of adapta- tion is allowed the rats in their living cages, no disturbance in their reactions is noticeable when they are forced to run the maze. ‘The correct turns are likewise not made upon the basis of any possible difference in the temperature in the various places in the maze. Differences in the pressure of the air columns do not form the basis of the discrimination for changing the air pressure does not disturb the rats. Iti is of 1 interest to note that the simple turning of the maze through the angles of 45°, go’, etc., badly disturbed both the normal and the defective animals for the first three or four trips after the change was made. ‘The method adopted in this case was as follows: The animals, e.g., were allowed to learn the maze with the entrance south; after they had pecan thoroughly familiar with it in this position the entrance was placed east. Now, although the relations of the turns within the maze are exactly the same as with the entrance south, the animals are confused. No explanation of this 1s offered at present. Summarizing the results obtained from this series of experiments, we may say that neither visual, auditory nor tactual sensations furnish the animal with the cue for making the correct turns in the maze. ‘The final conclusion is that kinzesthetic and organic sensations are the principal sensory factors in this association. A suggestion as to how the kinzsthetic sensations are “controlled” is offered. 3 This paper is to appear soon as a monograph supplement to the Psychological eview. Habit Formation in the Starfish. By H. 8S. Jennings. An account of experiments showing that by a course of training the starfish may be induced to use habitually a certain pair of rays on which to turn in the righting reaction. ‘The habit lasted in certain cases three or four days. Modifiability of Behavior in the Dancing Mouse. By R. M. YERKEs. Visual discrimination tests show that the dancer avoids a disagreeable stimulus The Mid-Winter Meetings. IgI after about one hundred experiences. This modification of behavior occurs more quickly in the male than in the female. It persists several weeks. Labyrinth tests are serviceable in the study of the dancing mouse only when the avoidance of an unfavorable condition is demanded. Neither escape from con- finement nor the obtaining of food furnishes satisfactory motives for the following of a labyrinth path. The animal can find its way readily in a simple labyrinth without the guidance of sight, smell and touch. ‘Thus far my experiments indicate the superiority of the female in the acquirement of labyrinth habits. Further Study of Variability in Spiders. By J. P. Porrer. Results indicative of the variability of the instinctive behavior of spiders were reported. The Effect of Distraction upon the Intensity of Sensation. By I. M. BEnTLey. Some Contributions to Applied Tone-psychology. By C. E. SEASHORE. Total Reactions. By E. H. CamMERon Non-sensory Components in Sense Perception. By R.S. Woopwortu. The familiar “staircase figure’? presents a problem which may be stated in the following terms: What is the difference in conscious content between seeing the figure as the upper side and as the lower side of a flight of stairs? Previous work has shown that the difference cannot be ascribed to the stimulus; and inquiry by the author showed that there was no detectable sensory imagery present in con- sciousness that could serve to differentiate the two appearances. ‘The two appear- ances have, not different sensory qualities, but what may be called different “ per- cept qualities. ” Other non-sensory percept qualities are found in the subjective grouping of dots, in a subjective rhythm, in perception of size and distance, and in perception of chimps! A percept is not properly described as a synthesis of joie sation and image, for the image is often not present when the percept is perfectly clear and definite. It is better to call the percept simply a mental reaction to sen- sory stimulus, and to recognize that a reaction, as a new event, probably has a quality of its own. ‘This point of view is borne out by brain physiology, especially by cases of word-deafness, etc., in which there is a loss of mental content, without a corresponding loss of sensory consciousness. Visual Pressure Images: Their Nature and their Relations to the Visions due to , Mescal and other Drugs. By E. B. DEJ.ABARRE. Indications of Incipient Fatigue. By W.S. Monroe. BENJAMIN Rusu, M. D., On Mental Diseases. By I. W. Rivey. F “Section F. (Zoélogy) of the American Assoctation for the Advancement of Science and the American Soctety of Zodlogists held joint sessions for the reading of papers, during which the following papers, among others, were read: The Artificial Production of a Single Median Eye in the Fish Embryo by Means of Sea-water Solutions of Magnesium Chlorid. By CHaries R. StockarD, Columbia University. Fundulus embryos when developed in certain strength solutions of MgCl, in 192 “fournal of Comparative Neurology and Psychology. sea-water form a large single median eye. This condition is comparable to the one eyed human monsters known as Cyclops, Cyclopia or Synophthalmia. The single eye results from an antero-medio-ventral fusion of the elements of the two optic vesicles at an early developmental stage. This fusion is more or less complete in the different embryos. The large compound optic cup induces the formation of a single lens. This lens is formed from ectoderm different in position from that of the normal lens forming region. ‘The lens is abnormally large in size as is also the optic cup, and the size of the former varies directly with that of the latter. It is probable that there 1s no localization of lens forming substance in the ectoderm of the fish embryo. This inter-relationship in the development of the optic cup and lens is interestingly compared with the processes of development in the amphibian eye as shown by recent experiments. Mixed sea-water solutions of MgCl, and NaCl also cause the one eyed condition. Since such a defect is characteristic of the MgCl, action when used in sea-water solutions one must infer that the Mg constituent in the mixture is responsible for the result. On the Place of Origin and Method of Distribution of Taste Buds in Ameiurus melas. By F. L. Lanpacre, Obio State University. This paper appeared in Pallas last issue of this Fournal. The Central Reflex Connections of Cutaneous Taste Buds in the Codfish and the Catfish: An Illustration of Functional Adaptation in the Nervous System. By C. Jupson Herrick, Denison University. The substance of this paper appeared in the article by the same author in the last issue of this Fournal. Some Little-known Shark Brains, with Suggestions as to Methods. By Burt G. Wiper, Cornell University. This paper continues that of which an abstract was printed in Science for May 26, 1905. Now first, so far as I know, are shown the brains of Heterodontus (Ces- tracion) and Pristiophorus. With the former the cerebrum and cerebellum resem- ble those of the “acanth” (Squalus acanthias), indicating an antiquity little if any greater. Notwithstanding certain ectal resemblances of the two dentirostral genera, Pristis, the “‘saw-ray” and Pristiophorus, the “‘saw-shark,” their brains differ markedly, the latter being the more primitive. Their inclusion within the same family or even the same division would seem to me an error only less in degree than would be their combination with Xiphias, Polyodon and Psephurus as “‘Rostrata,”’ or than was GUNTHER’S association of Ganoids and Selachians as “‘ Palzichthyes,”’ aptly characterized by GIL as a “‘piece of scientific gaucherie.”” Upon encephalic grounds I think Pristiophorus and Scymnorhinus should be excluded from the Squalide, and Sphyrna from the Carchariide. The brain of each selachian genus is, I think, recognizable, but I am less certain as to family forms. The Notidanoid or Diplaspandylous type is well marked, and includes Scymnorhinus. At present the rays cannot be distinguished from che sharks in any such simple way as, e. g., the Anura may be from the Urodela by the secondary fusion of the olfactory bulbs. Perhaps, ‘in no shark is the prosocele so nearly obliterated as T he Mid-Winter Meetings. 193 it seems to bein all rays. Inno ray do the cerebral protrusions remain unconjoined as in some sharks; but, paradoxically, in no ray is there, as in several sharks, so nearly a complete obliteration of the evidence of their primary independence. Under “Methods” may be enumerated: (1) The need of well-preserved brains of all species; (2) maintaining the natural contours, especially of thinner parts, by injecting the preservative into the cavities; (3) making solid injections of the cavities; (4) exposing brains with a “shoe-knife,’ ’ obliquely shortened; (5) explor- ing with the ‘‘syringotome” or canaliculus knife; (6) the use of sheets of uniform size, Say 35 x 45 cm., upon which, in a manner permitting change, are drawn out- lines of the animal and of its characteristic parts, especially the brain; such sheets may be arranged and rearranged upon the wall so as to facilitate research and exposition to small classes. An Experimental Study of the Image-Forming* Powers of Various Types of Eyes. By Leon J. Coxe, Rhode Island Agricultural Experiment Station, Kingston, RL. The responses of certain phototropic animals to two areas of light of different size, but of equal intensity, were used as criteria in drawing inferences as to the image-forming powers of their eyes. To one side was a ground-glass, lighted from behind, which gave an evenly illuminated area 41 cm. square. ‘To the other side was practically a point of light; but at the position midway between them, where the experiments were performed, the intensities of the two lights were equal. Eyeless forms (the earthworm was used) turned practically an equal number of times toward each light, showing no power of discriminating between them. Animals with “‘direction eyes” were but little better in this respect (e. g., Bipalium, Oniscus, larva of Tenebrio). On the other hand, animals with well-developed “compound eyes” (Vanessa, Ranatra) and “‘camera eyes”’ (frogs) discriminated readily, posi- tive animals turning much more often to the large light, and negative animals more often to the small. This discrimination was taken as evidence of image- formation by the eyes. Frogs (Acris gryllus) with the skin covered but eyes exposed reacted like normal frogs; without the use of the eyes their responses corresponded to those of the earthworm. We have thus a physiological test of the image-forming powers of the eyes, and in these experiments it corroborated in the main inferences which would be drawn from a study of the structure of the eyes in question. The Influence of Direction vs. Intensity of Light in Determining the Phototropic Responses of Organisms. By Leon J. Core, Kingston, R. J. The large land planarian, Bipalium kewense, was the principal animal experi- mented with. Its responses were first tried to shadows from a light directly overhead, 7. ¢., non-directive. It was then tested in a partial shadow, a strip of less intense light in an area of more intense illumination. In this case all the light came from one direction, namely, horizontally from one side. Although strongly negative, the worms would crawl directly toward the light in the partial shadow rather than turn out into the greater intensity. A similar result was obtained with the earthworm (Allolobophora foetida). In these experiments Bipalium and Allolobophora appeared to respond to intensity alone, regardless of the direction of the impinging light. 194 ‘fournal of Comparative Neurology and Psychology. The Sense of Vision in the Dancing Mouse. By Ropert M. Yerkes. Harvard University. That brightness vision is fairly well developed in the dancer is shown by its ability to discriminate blacks, greys and whites. Color vision is extremely poor. ‘There is some indication of the discrimination of red and green and of red and blue, but none whatever of blue and green. All my experimental tests as well as my observations of the habits of the mouse support the conclusion that such visual guidance as is received results from stimulation by brightness differences. There are many reasons for believing that the red end of the spectrum is much lower in brightness value for the mouse than for man. ‘The general behavior of the dancer and the results of form, brightness and color tests show that vision is not very impor- tant 1n the life of the animal. The Breeding Habits of the Florida Alligator. By AtBert M. Reese, Syracuse University. The habits of the alligator were studied during parts of three summers in the Everglades, in the swamps of central Florida, and in the Okefenokee Swamp. The time of laying is the month of June, usually during the second and third weeks. The nests, which are built on the bank near the caves of the alligators, vary con- siderably in size, and consist of a very compact mass of damp, decaying vegetation. They probably serve more as a means of keeping the eggs moist and at a constant temperature than as a means of heating them. ‘The average number of eggs in a single nest is about thirty, forty-eight being the greatest number found in one nest. The eggs are so closely packed in the nest that it seems hardly possible that the young alligators, on hatching, should be able to dig their way out; it is possible that the female who laid the eggs may hear the noise made by the young before hatching and may dig them out of the nest before they suffocate. ‘The period of incubation is probably about eight weeks, and sometimes is found to have begun before the eggs are laid, so that eggs taken directly from the oviducts may contain well advanced embryos. ‘There is considerable variation in the size of the eggs, the variation in long diameter being greater than that in short diameter. “The average long diameter of the four hundred eggs measured was 73.742 mm. ‘The average short diameter was 42.588 mm. Analysis of the Cyclical Instincts of Birds. By Francis H. Herrick, Western, Reserve University. The behavior of wild birds is primarily determined by a number of commanding instincts of ancient origin. ‘These cardinal instincts are of two kinds, namely: (1) continuous instincts, which are needed for the preservation of the individual, such as preying, fear, concealment and flight, and (2) cyclical instincts, which are necessary for the maintenance of the race. By cyclical instincts we mean those discontinuous, recurrent impulses which attend the reproductive cycle, and which may be described as parental instincts. The cyclical or parental instincts as a rule recur with almost clock-like precision, in spring or summer, with repetitions within the breeding season in certain species. They are modified by the continuous instincts, such as fear, and the instinctive behavior as a whole is liable to modification at every point by intelligence. Neglect- ing such changes for the present, we will briefly analyze the cyclical instincts, reserving details and tabular statements for a fuller presentation. The Mid-Winter Meetings. 195 The reproductive cycle is made up of a series of terms, representing discrete acts or chains of actions in a definite succession. Eight or more terms may be recog- nized, many of which, such as brooding and feeding the young, are recurrent within the series. The cycle may be graphically represented by a series of tangent circles, each one of which stands for a distinct sphere of influence, or subordinate series of related impulses, named and numbered as follows: (1) Spring migration; (2) courtship and mating (often attended by song); (3) selection of nesting site and building nest (often accompanied by the fighting instinct); (4) egg-laying; (5) incubation—including care of eggs, such as shielding, rolling, cleaning and cover- ing (fear often completely blocked by brooding instinct); (6) care of young in nest, subject to the following analysis: (a) feeding young, including capture and treatment of prey, return to nest (pause), call-stimulus, testing reflex response of throat, watching for reflex response (pause); (b) inspection of young and nest; (c) cleaning young and nest; removal and disposition of excreta; (d) incidental care of young and incidental behavior in this and other terms of cycle—brooding, shielding or spreading over young whether sitting or erect, bristling and pufhing, preening, gaping, stretching and yawning, guarding and fighting; (7) care and incidental education of young when out of nest; guarding, feeding, play, and other instinctive acts; (8) fall migration. Beginning at 2, 3 or 4, according te circumstances, the cycle may be repeated once or oftener within the season. The coordinated instinctive responses of the young begin in the sixth term, and are mainly as follows: (6) Initial responses at moment of hatching or shortly after, including grasping movements of limbs, elevation of head, opening of mouth, and the swallowing reflex in response to contact of bill of old bird or of food in deep part of throat; characteristic actions in muting following feeding, in response to the attitude of inspection in adult; call-notes, pecking, and gaping, stretching, and spreading in response to heat, flapping, fear and flight; (7) calling (teasing), follow- ing, crouching and hiding, play, imitation, preying and flight; (8) fall migration. The formula of the reproductive cycle given above is a composite, which with slight changes will apply to most of our common wild birds. In the most aberrant cases of behavior, where the parental instincts have been reduced to a minimum as in the cow buntings and the megapodes, the cycle ends abruptly at term 5, and in the cowbird there is no attempt to either build a nest or to conceal the eggs. The Blending and Overlap of Instincts. By Francis H. Herrick, Western Reserve University. There are many anomalous actions or peculiarities of behavior in wild birds which have not been satisfactorily explained, although certain of them have been long known. Some of the eccentricities of conduct referred to are the following: (1).Repair of the old nest or the building of a new one at the close of the breeding season; (2) omission of nest building, and dropping of eggs on the ground; (3) leaving young to perish in nest, and starting on migration; (4) offering strings or other objects to young in the place of food; (5) building more than one nest including the “‘cock nests’’ of marsh wrens; (6) rebuilding on the same site, producing super- imposed nests or nests of from two to four “‘stories” “‘to conceal” foreign bodies, such as the cowbirds’ eggs in the nests of vireos and warblers. All of these curious actions receive much light, and in most cases are satisfactorily explained by what we shall call the blending or overlapping of instincts. As shown 196 Fournal of Comparative Neurology and Psychology. in the previous paper, the wild bird commonly passes through a cycle of instincts which mark the breeding season. ‘This cycle is made up of eight or more terms, which follow in serial order, and some of which are recurrent. Normally the bird passes from center of influence I to center 2, 3, and so on, to the end of the cycle. There is little overlap or blending, the bird remaining under the influence of a given instinct or series of instincts, such as nest building, incubation, or feeding the young until its instinct in any given direction has been satisfied, before entering a new sphere or being swayed by new impulses. When the correlation or attune- ment is perfect the instincts of mother and child fit like lock and key. Like clocks beating synchronously the instincts of mother and child are generally in harmony, but one of the clocks occasionally gains or loses, stops or runs down; one term is liable to be weak or to drop out altogether, so that there is an overlap or a gap in the series which may be serious. On the other hand, one term may be unduly strength- ened, like nest building or incubation, and a preceding or following term corre- spondingly weak. In all such cases there are eccentricities of conduct, which, if not fatal to the young, are very puzzling to the naturalist. Most wild birds normally pass one reproductive cycle in the season; a certain number, however, begin, but do no complete a second cycle; further, many like the robin and bluebird not only begin but complete a second and even a third cycle within the breeding period. The repair of the old nest in autumn by fish hawks or eagles is not done “in anti- cipation of spring,” and implies no more intelligence than the building of the original nest. tas simply the recrudescence of the building instinct, due to che beginning of a new reproductive cycle which is never finished. Leaving the young to perish in the nest in autumn is brought about by the scamp- ing of the cycle at the other end. The migratory impulse overlaps and replaces the parental instinct. An adult robin has been seen to offer a string to its fully grown young, and try to cram it down the throat of the fledgling. Later, the old bird flew with the string into atree. ‘This was the result of the overlapping of two reproductive cycles, or of the last term of one cycle, and the first term of a succeeding cycle. The bird was alternately swayed by opposing impulses, now being impelled to gather nesting material, when she picked up the string, now by parental instinct to feed her young, when she tried to serve it, and again possibly by the instinct of building when she flew with the string into a tree. Building more than one nest can be accounted for by excessive development of the building instinct, or by the influence of fear repeatedly interrupting the cycle, together with attachment to nesting site, but the discussion is too long for this abstract. The rebuilding of nest on nest, giving rise to the wonderful storied structures sometimes produced by the summer yellow bird, or vireo, when plagued by the cowbird, so that the foreign egg is buried out of sight, is not an illustration of reason, as commonly suggested, but the curious result of a pure instinct. “The reproductive cycle is broken by fear, and a new one is begun, and in these rare cases the old nest is retained as a site to be build upon. Instead of having two supernumerary nests, both of which may contain eggs, as in reported cases of the phoebe, we have a series of superimposed nests. The new nest is not built to conceal the cowbird’s egg, although it does so perfectly, any more than the addition of new materials to the The Mid-Winter Meetings. 197 osprey’s nest in the fall is in the nature of repairs, although it answers this purpose admirably. ‘The nest is built because the bird is at the opening of a new cycle, and is impelled by the building instinct. Many confirmatory facts could be given. The herring gull will not only bury an egg, in rebuilding on its old site the nest, when its cycle has been interrupted by fear, but will bury its dead young which it treats as so much nesting material. The Interrelation of Sensory Stimulations in Amphioxus. By G. H. Parker, Harvard University. To weak acid solutions and other like mixtures the anterior end of Amphioxus was found to be most sensitive, the posterior end less so, and the middle trunk region least sensitive. To the pressure of a camel’s hair brush, the middle region was less sensitive than the two ends which, however, were not distinguishable one from the other by this method of stimulation. To a current of warm water (40° C.) the anterior end was most sensitive, the middle less, and the posterior end least. “There were no reactions to a current of cold water (2°C.) To a fine pencil of strong sun- light, previously passed through water to eliminate heat, the anterior end including the “eye spot’’ was not sensitive, the region immediately behind the ‘“‘eye spot” was most sensitive, the posterior region slightly less so, and the middle region least so. The distribution of sensitiveness to light corresponds to the distribution of the pigment cups in the central nervous organ and these cups are without doubt the mechanisms concerned with the reception of light. The distributions of the other classes of sensitiveness are in mutual agreement, and, from the nature of their stimuli, these classes are doubtless represented by integumentary nerve terminals. To what extent these classes are independent may be inferred through the effects of exhaustion. After the tail of Amphioxus has been repeatedly stimulated with weak acid, the animal ceases to respond to this stimulus but is still normally sensitive in that part of its body to heat or to mechanical stimulation. In a similar way after exhaustion to mechanical stimulation or to heat stimulation, the particular part of the body experimented upon is still sensitive to the other classes of stimuli. Exhaus- tion to light stimulation has no effect upon the sensitiveness to the other classes of stimuli. “These observations lead to the conclusion that light, heat, mechanical and chemical stimuli are received by physiologically separate mechanisms and that these mechanisms are located in the skin except in the case of light, whose receptive organs are the pigment cups in the central nervous organ. The Habits and Life History of Cryptobranchus allegheniensis. By BERTRAM G. SmirH. (Introduced by Dr. O. C. GLasrr.) The adult Cryptobranchus has its dwelling place in a cavity or cavern under a large rock, in swift and shallow water. The animal seldom comes out during the daytime, except during the breeding season. ‘The eggs are laid and fertilized during the first two weeks of September. ‘They are deposited in the usual dwelling-place ofthe animal. About 450 eggs are laid by a single female. Fertilization is external as in fishes; no spermatophores are formed. After the eggs are deposited they are usually guarded for a time by the male, who fights and drives away other hell- benders which attempt to eat the eggs. The male himself eats some of the eggs, but on account of the slowness of his digestion is unable to eat more than a small pro- portion, hence his presence is in the main protective. In defending the eggs the 198 Fournal of Comparative Neurology and Psychology. male is merely guarding his own food-supply: the origin of the brooding habit in this case seems to be the feeding habit. The eggs hatch about six weeks after fertiliza- tion. Thenewly hatced larva is about 25 mm. long, and has a large yolk sac. Larve kept in the laboratory for two months after hatching retain a remnant of the yolk sac, and refuse food. Year-old larva are 6-7 cm. long, and retain the external gills. Larvae two years old are about 12 cm. long and the external gills are greatly reduced. Sexual maturity is attained with a length of about 34 cm. and probably requires three or four years. Movement and Problem Solving in Ophiura brevispina. By O. C. GLaser, University of Michigan. 1. Ophiura brevispina moves in practically all of the ways possible to a pen- taradiate animal. 2. Its behavior in removing obstructions from its arms is not perfected by practice under ordinary conditions. 3. PREYER’s conclusion that Ophiurans are intelligent is not substantiated by this study; for not only is it impossible to demonstrate “‘resolution” or improve- ment, by the method that he employed, but the assertion that an animal is intelligent because when stimulated it performs varied movements until some one of these brings about cessation of the stimulus, leads into difficulties, for these animals often perform in instantaneous succession movements that fail for the same reason. Ophiura, moreover, hardly ever executes a single movement, but usually a consider- able number. Each of these on PREYER’s view results in learning, but it is impossi- ble without striking evidence to the contrary, to believe that Ophiurans can learn half a dozen things at the same time. If some of all the movements performed at a certain instant are “‘correct,”’ the case is farther complicated in that some of all the things which the animal learned fall into the category of successes, some into the category of failures. 4. The reason why Ophiura brevispina does not improve under ordinary cir- cumstances is probably due to its versatility. This animal can perform a sur- prising number of movements. Of all these some are better fitted to meet a cer- tain difficulty than others, but a considerable number will serve the purpose. Where the number of solutions to a problem is large, it is not surprising that no particular method of solution shall be perfected, viz.,that resolution should not occur. Notes on the Behavior of Sea-Anemones. By Cuas. W. Harcirt, Syracuse University. The paper discussed the aspects of behavior of several species of sea-anemones studied both under natural conditions and those of the laboratory. ‘The points chiefly under observation had reference to the behavior of these creatures under the influence of light. So far as known few details along this line have been recorded. At least three species of anemones were found which showed very evident reactions to photic stimuli: namely, Eloactis (Halcampa) producta, Sagartia modesta, and S. leucolena. Of these two are tube-dwelling, burrowing in the sand near tide lines, and forming rude tubes or burrows through the adhesive secretions of the ectoderm. S. leucolena is occasionally found in similar habitat, though chiefly adhering to rocks or among colonies of ascidians, or sponges, on piles of docks, etc. Experiments showed that the first two species are most sharply T he Mid-Winter Meetings. 199 responsive to light, and this sensory sense is located chiefly in the tentacles and oral regions of the body. S. leucolena, while less sensitive, is yet evidently so in strong light. Exposed to direct sunlight it quickly closes up into a hemispherical mass, or creeps over the edge of the rock or shell into shaded portions of the aquarium. In its native haunts it may be found protruding its crown of tentacles from a crevice while the body is hidden. Sagartia lucia is a free living species found abundantly almost everywhere, on rocks in open pools, or on floating fucus, and freely exposed to direct sunlight, action of waves, etc. Of similar habit is Metridium marginatum. Neither of these species seems in the least degree responsive to photic stimuli. Under a strong beam of sunlight reflected directly upon them for ten minutes they showed no response whatever. These facts, together with others as to food habits, etc., render it quite certain that their behavior is due to several factors, and that in response to light there is an evidence of adaptation involving varying physiological conditions, of which the burrowing habit is one of several expressions. Further Observations on the Behavior of Tubicolous Annelids. By Cuas. W. Haraitt, Syracuse University. Following up the work done on these animals and reported elsewhere, the writer has extended the observations to aspects of behavior other than those already recorded. Three points are concerned in the following observations: First, a study of behavior under natural conditions of environment. This has been possible in quiet pools near low tide lines. Experiments on Hydroides dianthus with shadow stimuli, or light intensity of varying degree, under these conditions have confirmed in all essentials those made last year. Experiments as to tactile responses showed considerable variations as compared with the former series. ‘This may be attributed to the fact that specimens living under these conditions become more or less inured to similar stimuli from the actions of waves which naturally buffet them almost constantly. Second, experiments on the relative sensory acuteness of specimens from deep water, about twenty fathoms compared with those from shallow waters, one to three or four fathoms. In cases tested there were shown a definite preponderance of positive reactions among the latter, and a corresponding preponderance of nega- tive responses in the former. Third, a comparative study of the aspects of behavior shown in the growth of colonies taken from shore waters, subject to the action of waves, and those from quiet waters of bays, etc., shows an unmistakable variability in the aspects of the tubes, which clearly indicates environmental adaptation. Furthermore, specimens growing in an environment, such as marly bottom, or silt, or other similar condition, show the same evident response of adaptation. On the other hand, specimens growing along shore lines, or on rocky bottoms, show likewise the unmistakable response natural to such condition. Not a single colony among hundreds along the shore lines showed any free and vertical tubes. Likewise specimens dredged from muddy bottoms showed the erect and vertically directed tubes which would bring the animals above the obstructing mud. Any careful consideration of the facts would hardly fail to convince one that no single factor, such as heliotropism, or geotropism, or any other tropism alone, was adequate for their explanation. 200 8 ‘fournal of Comparative Neurology and Psychology. Rhythmical Pulsation in Animals. By ALFrep G. Mayer. Experiments made at the Tortugas Marine Laboratory of the Carnegie Instt- tution upon Cassiopea, Salpa, Lepas, and the loggerhead turtle give results as follows: Rhythmical pulsation can be sustained only when a strong stimulus is counter- acted by an inhibitor, so that the pulsating organism is maintained at or near the threshold of stimulation, in a state analogous tothat of unstable equilibrium, thus allowing weak internal stimuli to produce recurrent movement. In the lower marine animals the NaCl, calcium, and potassium of the sea-water combine to form a powerful stimulant, which if unchecked would produce only sustained tetanus, but the magnesium overcomes this effect by its anesthetic (diastolic) influence. The pulsating organs of terrestrial animals are also stimulated by optimum combinations of NaCl, with potassium and calcium, and this is held in check by a definite proportion of magnesium. A Ringer’s solution resembles this optimum combination of NaCl, calcium and potassium, and is only a stimulant, not an inorganic food. It must be counter- balanced by magnesium in order to enable it to sustain pulsation indefinitely. In Cassiopea any paralyzed strip of sub-umbrella tissue, cut in the shape of a closed circuit, will remain indefinitely i in rhythmical pulsation, if once a contraction wave be genes in the circuit. Everytime this wave returns through the circuit of tissue to the place whence it started, it is re-stimulated and sent forth anew, and being thus reinforced at each return it is sustained indefinitely. In the scyphomedusa, Cassiopea, the diffuse nervous or epithelial elements of the sub umbrella transmit the pulsation stimulus to which the muscles respond by contraction. The peripheral muscular layer of the wall of the loggerhead turtle’s heart is the only part actively concerned in the rhythmical movement, and the internal cavernated mass of the heart’s tissue may be removed without checking the pulsation. This peripheral part of the muscular wall of the heart tends to maintain itself in pulsa- tion very much as will circuits made of the sub-umbrella tissue of Cassiopea. The pulsation-stimulus acts solely upon the peripheral muscular layer of the heart’s wall, the inner cavernated tissue remaining passive. The above is a brief review of Publication No. 47, of the Carnegie Institution of Washington, “Rhythmical Pulsation in Scyphomeduse,” 1906. LITERARY NOTICES. Sherrington, Charles S. The Integrative Action of the Nervous System. New York, Charles Scribner's Sons. 1906. Pp. xvi-git. 85 Figs. $3.50. The material of this book was presented as the second series of the Yale Univer- sity Mrs. Hepsa Ely Silliman memorial lectures. Professor SHERRINGTON, who by his brilliant researches on the structure and functions of the nervous system has proved himself a master investigator, has gathered together in the ten lectures which constitute this volume a large number of the most important facts of nerve physiology. But, further than this, he has pre- sented the results of his investigations, for the book is essentially the product of his own and his students’ researches, in an interesting manner and has skillfully pointed out their meaning. Physiologists, students of animal behavior and psy- chologists alike recognize the value of the author’s investigations, and of his interpretations of his results. As the title suggests, the lectures deal primarily with certain of the relations of the nervous system to the activities of the organism; they are preéminently important contributions to the comparative study of behavior. Indeed, “The integrative action of the nervous system” might well be read as a sequel to JENNINGS” masterly discussion of the behavior of organisms which either totally lack a nervous system or possess a very simple one, for SHERRINGTON deals with the nature and relations of the reflex in higher animals, and with the regula- tion of behavior by the nervous system. Because of the striking individuality of the author’s style and his relatively new terminology the reader is likely to get an over-emphasized impression of the original- ity of the book. Even old, familiar facts as expressed in it at first appeal to the reader as new discoveries. Nevertheless the work is original, unusually so, not alone in its materials but even to a greater degree in method of presentation and point of view of interpretation. For most readers the careful study of SHERRING- ToNn’s book will mean a new and research inspiring conception of the role of the nervous system, of the place of the study of activity, and of the relations of the physiology of the nervous system to psychology. My first plan was to make this review an abstract of the volume, but I soon found that anything like an adequate summary of the lectures would demand an unreasonable amount of space. I shall therefore attempt, instead, after describing by title several lectures, to give a vivid impression of the nature and value of the book by selecting for presentation certain of the most important points made by the author; and while thus describing the materials of the work, I shall attempt to exhibit the interesting style and terminology of the author by the use’ of his own words in illustrative quotations. Anent the terminology of the book, it may be observed that it is the best example of the application of something like the objective terminology of BEER, BETHE and von UEXKULL that has ever appeared in physio- 1The Behavior of the Lower Organisms. 1906. 202 Literary Notices. logical literature. It is reasonably self-explanatory, simple and direct and as employed by the author it serves the excellent end of enabling him to speak of objective and subjective facts without confusion. Lectures I, II and III, codrdination in the simple reflex, discuss the nature of the structural and the functional units of the nervous system (the neurone and the nerve-arc) and of the interconnection of the various parts of the body through the integrative action of the nervous system. The phenomena of refractory phase and inhibition are dealt with at length and in a most illuminative manner. Lecture IV, interaction between reflexes, points out the artificiality of the conception of the simple reflex, the features of reflex-arc functions which result in the harmonious compounding of reflexes and the existence of two important classes of reflexes, the allied and the antagonistic, the mutual relations of which provide us with the mani- fold phenomena of facilitation and inhibition. Lectures V and VI, compound reflexes, are devoted, the first to simultaneous combination of reflexes, the second to successive combination. In connection with simultaneous combination it is made clear that PFLUGER’s laws of spinal irradiation are contradicted by many of the facts of the author’s investigations. Under the subjects, irradiation and “reflex pattern” the main features of the simultaneous compound reflex are thoroughly discussed. In the lecture on successive combina- tion chain-reflexes receive attention. The current theories of inhibition are examined. Noci-ceptive (pain) nerves are shown to bring about reflexes which are prepotent. Lecture VII, reflexes as adapted reactions, is an excellent discus- sion of the purposes of reflexes and of the conditions which reveal and conceal the same. Lecture VIII, some aspects of the reactions of the motor cortex, in addi- tion to giving with delightful clearness the topography of the motor cortex of the chimpanzee, the orang-outang and the gorilla, suggests the relation of the motor cortex to receptors (sense? organs), more especially to distance receptors. Lecture 1X, the physiological position and dominance of the brain, adds to a splendid résumé of the results discussed in the previous lectures, a convincing array of facts in support of the contention that the cerebrum is preéminently the ganglion of the distance-receptors, the cerebellum the ganglion of the proprio-ceptive (that is, the internal as contrasted with the surface-receptor) system. Lecture X, sensual fusion, consists of a comparison of the nervous integration of movement and of sensation by reference to results of the study of certain visual phenomena. Of the three prominent points of interest from which the functions of the ner- vous system may be studied, metabolism, conduction and integrative action, this series of Jectures deals almost exclusively with the last. This integrative or inter- connecting activity of the nervous system which serves to make of the complex multicellular organism a functional unit of a high degree of efficiency is in large part the coérdination of parts of the body by reflex action. Structurally the unit of the nervous system is the neurone, functionally it is the nerve-arc, which consists in its simplest form of a receptor and an effector which are linked by a conductor. The series of functional stages in the action of the nerve-arc may be termed initiation, conduction and end-effect. So far as integrating activity is concerned the reflex- arc is the unit of mechanism, the reflex-act the unit of reaction. And a reflex-act, according to the author, is a reaction “in which there follows on an initiating reac- tion an end-effect through the mediation of a conductor, itself incapable either of the end-effect or, under natural conditions, of the inception of the reaction”’ (p. 6). Literary Notices. 2.03 The simple reflex is an abstraction for in reality no reflex exists independently; there is always coérdination. This codrdination is of two sorts; simultaneous and successive; the former gives origin to the reflex-patterns, the latter to the chain- reflexes or shifting patterns by which the parts of the body are constantly brought into adaptive relation to one another and to the changing environment. Briefly and pointedly the author discusses the rdle of each of the elements of the reflex-arc. The chief function of the receptor is selective excitability; ‘it lowers the threshold of excitability of the arc for one kind of stimulus, and heightens it for all others” (p. 12). Point by point the characters of nerve-arc conduction are exhibited in the light of the reflexes of the spinal dog. Nerve-are conduction differs from nerve-trunk conduction in eleven important respects: latent period, after- discharge, relation of rhythm of end-effect to rhythm of stimulus, relation of inten- sity of stimulus to end-effect, summation, irreversibility of direction of impulse, fatigability, variability of the threshold, refractory period, dependence on blood-cir- culation, susceptibility to drugs (p.14). The first three lectures present the results of intensive research concerning each of these eleven features of nerve-arc conduc- tion, together with interpretations of the results in terms of adaptive reactions. Many of the differences between the conduction of the reflex-arc and the nerve trunk SHERRINGTON thinks are to be referred to the inter-neurone region, the synapse. Irreversibility of the direction of conduction is due, possibly, to the fact that the synaptic membrane is more permeable in one direction than in the other (p. 42). In the discussion of refractory phase, we have an excellent illustration of the way in which the author gives life to his book by pointing out the significance of his facts. ‘The scratching-reflex, in order to secure its aim, must evidently consist of a succession of movements repeated in the same direction, and intervening between the several numbers of that series there must be a complemental series of movements in the opposite direction. Whether these two series involve reflex contractions of two antagonistic muscle-groups, respectively, in alternate time, | would leave for the present. The muscle groups or their reflex-arcs must show phases of refractory state during which stimuli can not excite, alternating with phases in which such stimuli easily excite. Evidently this is fundamental for securing return to the initial position whence the next stroke shall start. “The refractory phase secures this. By its extension through the whole series of arcs it prevents that confusion which would result were refractory phase in some of the arcs allowed to concur with excitatory phase in others” (p. 63). And further, in the same connection, to show the value of inhibition and the reasons for the existence of central as well as peri- pheral inhibition, he states that the scratch-reflex “‘is but one of several reflexes that share ina condominium over the effector organ—the limb. It must, there- fore, be possible for the scratch-reflex taken as a whole, to be, as occasion demands, replaced in exercise of its use of the limb by other reflexes, and many of these do not require clonic action from the limb—indeed, would be defeated by clonic action. It would not do, then, for the peripheral organ itself to be a clonic mechanism. The clonic mechanism must lie at some place where other kinds of reflex can preclude the clonic actuator from affecting the peripheral organ. Now such a place 1s obviously the central organ itself; for that organ is, as its name implies, a nodal point of meeting to which converge all the nervous arcs of the body, and among others all those which for their several ends have to employ the same mechanical organ as 204 Literary Notices. does the scratch-reflex itself. It is, therefore, only in accord with expectation that the seat of the refractory phase of the scratch-reflex lies where we trace it, in the central nervous organ itself, and somewhere between the motor neurone to the muscle and the receptive neurone from the skin” (p. 65). In the principle of the final common path, SHERRINGTON has found an explana- tion of many of the complex relational aspects of reflexes. Each receptor has its private path; these come together in the so-called internunctal paths, which in turn unite in the final common path. Results of the necessity for the use of the same final common path by a number of receptors are: (1) That receptors which have different or opposite end-effects are forced to use the path successively, not simul- taneously. “The result is this or that reflex but not both together” (p. 117). Thus the final common path prevents the harmful interference of antagonistic re- flexes. (2) That receptors which have similar or harmonious end-effects reinforce one another in the use of their final common path. ‘“‘If, while the scratch-reflex is being elicited from a skin point at the shoulder, a second point distant, e. g., 10 cent. from the other point but also in the receptive field of skin, be stimulated, the stimulation at this second point favors the reaction from the first”’ (p. 120). Between the two classes of receptors, the extero-receptors, which supply the surface of the body and the proprio-receptors, which lie in the depths of the organ- ism, important relations of reinforcement and inhibition are shown to exist. ‘Thus, in the case of the flexion-reflex, “the receptive field includes not only reflex-arcs arising in the surface field, but reflex-arcs arising in the depths of the limb. Com- bined, therefore, with an extero-ceptive area, this reflex has, included in its receptive field, a proprio- ceptive field. The reflex-arcs belonging to its extero-ceptive and proprio-ceptive components cooperate harmoniously together, and mutually rein- force each other’s action. In this class of cases the reflex from the muscle-joint apparatus seems to reinforce the reflex initiated from the skin” (p. 131). Accord- ing as a reflex initiated by a given receptor exalts or depresses another simultaneously or successively occurring reflex it may be spoken of as excitatory or inhibitory. In this mutual relation of interference we have, according to SHERRINGTON, an expression of the fundamental importance of the principle of the common path. For every convergence of afferent neurones furnishes a condition for the interference of their reflexes or, in other words, it constitutes a mechanism of co6rdination. The author, far from making hasty or ill-supported general statements concern- ing these matters, builds his structure of facts with skill, system and rare insight to the point at which his conclusions, and in many cases his interpretations as well, are forced upon the reader. I do not wish to give the impression that the book is a highly speculative discussion of the integrative action of the nervous system. It is first of all an account of experimental study of the subject, and secondly an exceedingly valuable discussion of the meaning of the facts which are available. In connection with his investigation of the phenomena of irradiation it 1s note- worthy that SHERRINGTON has discovered a number of important inconsistencies between observed facts and the laws of spinal action as formulated by PFLUGER (p. 161). And it is also significant that in calling attention to important evidences of the inadequacy of the so-called laws of reflex action he does not attempt to formulate laws after his own observations but with the wisdom of an investigator who realizes that we are working in the infancy of nervous physiology describes the appearances which he has observed, and, so far as his observations go, states the Literary Notices. 205 rules which certain classes of reflexes follow. These descriptions are given in terms of the reflex figure, a most convenient scheme for the presentation of the total visible effect of a given stimulus or stimulus complex. “Thus at any single phase of the creature’s reaction, a simultaneous combination of reflexes is in existence. In this combination the positive element, namely, the final common paths (motor neurone groups) in active discharge, exhibits a harmonious discharge directed oy the dominant reflex-arc, and reinforced by a number of arcs in alliance witht =. * But there is also a negative element in this simultaneous combination of reflexes. The reflex not only takes possession of certain final common paths and discharges nervous impulses down them, but it takes possession of the final common path whose muscles would oppose those into which it is discharging impulses, and checks their nervous discharge responsive to other reflexes. * * * In this way the motor paths at any moment accord in a united pattern for harmonious synergy, cooperating for one effect” (p. 178). A number of reflex figures, with their respec- tive nervous patterns, are lesenbied a in detail. A good illustration of the author’s aptitude for indicating the utility and developmental conditions of activities is furnished by the following fragment from the discussion of immediate induction. “If a parasite in its travel produces excita- tion which is but close below the threshold, its progress is likely to so develop the excitability of the surface whither it passes that the scalptor-reflex will be evoked. In the skin and the parasite respectively we have, no doubt, two competing adapta- tions at work. It is perhaps to avoid the consequences of the spatial spread of the “bahnung” that the hop of the flea has been developed” (p. 184). In the competition of reflexes for the use of a common path dominance 1s deter- mined in large measure by four factors: spinal induction, relative intensity of the stimulus, relative fatigue, and the functional species of the reflex. Each of these factors hue been investigated by the author and receives adequate consideration in the lecture on the successive combination of reflexes. Concerning fatigability and its excuse for existing, he writes in his characteristic style: “The waning of a reflex under long maintained excitation is one of the many phenomena that pass in physiology under the name of ‘fatigue.’ It may be that in this case the so-called fatigue is really nothing but a negative induction. Its place of incidence may lie at the synapse. It seems a process elaborated and preserved in the selective evolu- tion of the neural machinery. One obvious use attaching to it is the prevention of the too prolonged continuous use of “a common path” by any one receptor. It precludes one receptor from occupying for long periods an effector organ to the exclusion of all other receptors. It prevents long continuous possession of a com- mon path by any one reflex of considerable intensity. It favors the receptors tak- ing turn about. It helps to insure serial variety of reaction. The organism, to be successful in a million-sided environment, must in its reactions be many-sided. Were it not for such so-called “fatigue,” an organism might, in regard to its recep- tivity, develop an eye, or an ear, or a mouth, or a hand or leg, but it would hardly develop the marvelous congeries of all those various sense organs which it is actually found to possess” (p. 222). Under the subject, species of reflex, it is convincingly argued, in the light of a wealth of experimental evidence, that reflexes initiated by noci-receptors (such as are especially adapted to the reception of nocuous or harmful stimuli) are pre- potent. Such stimuli do not demand specialized sense organs adapted to a particu- 206 Literary Notices. lar form of energy and possessing for the adequate stimulus a low threshold, but instead it is to the advantage of the organism to have in certain naked nerve endings themselves, which are called by SHERRINGTON the noci-ceptive nerves, a form of receptor which is capable of responding to a wide range of different kinds of stimuli. The noci-ceptive function would be cramped, as the author puts it, by the specializa- tion of an end-organ. The lecture on adapted reactions is of special interest to students of behavior and of psychology since in it the author considers the purposes of reflexes, the nature of spinal shock and its relation to the brain, the local sign of reflexes, and the rela- tions of reflexes to emotion and its expressions. - We may not do more than note a few of the main results of this lecture. With reference to the purposive aspect of reflex action the author writes: “‘In the flexion-reflex of the hind limb excited by noxuous stimuli, e. g., a prick of a faradic current, the limb itself is drawn up—if weakly, chiefly by flexion at the knee; if strongly, by flexion at hip as strongly as at knee. At the same time the crossed hind limb is thrown into action, primarily in extension, but this is soon followed by flexion, and alternating extension and flexion is the characteristic result. The rate of hie alternation is about twice a second. That is to say, the foot which has stamped on the thorn is drawn up out of way of further wounding, and the fellow hind limb runs away; and so do the forelegs when, —which is more difficult to arrange, owing to the height of the necessary spinal transection—they also are included, fairly free from shock, within the ‘spinal’ animal” (p. 240). Although it is well known, it may not be amiss to mention the fact that the James, Lance and Sercr notion that emotion is the result of certain organic pro- cesses, vaso-motor, visceral, cutaneous, has been tested experimentally by SHERRINGTON and found to lack satisfactory support. ‘There can be no doubt that SHERRINGTON effectually did away with the possibility of influence of the organic process to which the above writers had ascribed an all important role in the production of an emotional state, but in the opinion of the reviewer, 1t may fairly be objected that he has not conclusively proved that the emotion-expressing reflex figures, which he observes after spinal transection, are not the result of changes induced in the central nervous system by these same organic processes previously to the spinal transection. In an attempt to meet this possible objection the author writes: ‘‘But it is noteworthy that one of the dogs under observation had been deprived of its sensation when only nine weeks old. Disgust for dog’s flesh could hardly arise fromthe experience of nine weeks of puppyhood in the kennel” (p. 265). Surely, however, during that time the puppy may have ac- quired the experience of the disagreeable, perhaps it need not be that of the taste of dog’s flesh. Investigation of the effects of stimulation of the cortex of the anthropoid apes and of the influence of strychnine and tetanus toxin upon the reflexes induced by stimulation of the motor cortex has brought into clear light the fact that excitation and inhibition constantly accompany one another. Under certain conditions inhi- bition is converted into excitation and serious confusion results. The disorders brought about by tetanus and by strychnine poisoning “work havoc with the coordinating mechanisms of the central nervous system because in regard to certain great groups of musculature they change the reciprocal inhibitions, normally assured by the central nervous system, into excitations. The sufferer is subjected Literary Notices. P07 to a disorder of coordination which, though not necessarily of itself accompanied by physical pain, inflicts on the mind, which still remains clear, a disability in- expressibly distressing. Each attempt to execute certain muscular acts of vital importance, such as the taking of food, is defeated because from the attempt results an act exactly the opposite to that intended. The endeavor to open the jaw to take food or drink induces closure of the jaw, because the normal inhibition of the stronger set of muscles—the closing muscles—is by the agent converted into excitation of them” (p. 298). SHERRINGTON’s work has discovered the existence of two systems of innervation controlling two sets of musculature. “These systems are named by him the phasic and the tonic reflex systems. “The phasic system exhibits those transient phases of heightened reaction which constitute reflex movements; the tonic system main- tains that steady tonic response which supplies the muscular tension necessary to attitude” (p. 302). If receptors be classified as those which receive impressions which are referred beyond the organism (SHERRINGTON’S distance-receptors) and those whose im- pressions are referred to the organism (proprio-receptors), it is to be noted that the cerebellum is the main ganglion of the proprio-ceptive system, while the cerebrum stands in a like relation to the distance-receptors. ‘““It is the long serial reactions of the “distance-receptors” that allow most scope for the selection of those brute organisms that are fittest for survival in respect to elements of mind. The “‘dis- tance-receptors”’ hence contribute most to the uprearing of the cerebrum. One of the most important of the groups of proprio-ceptive organs is that of the laby- rinth, and these together with those other receptors whose chief function is the con- trol of attitude have as their center of reference the cerebellum. “The distance- receptors, and therefore the cerebrum, are the chief inaugurators of reaction; the proprio-ceptors, and therefore the cerebellum, control the habitual taxis of the skeletal musculature. In the lecture on sensual fusion it is shown that the fusion of sensation does not follow the rules which have been established for the fusion of reflexes. ‘‘The cerebral seats of right- eye and left-eye visual images are thus shown to be separate (referring to an experiment previously described.) Conductive paths no doubt interconnect them, but are shown to be unnecessary for visual unification cf the twoimages. ‘The unification of a sensation of composite source is evidently asso- ciated with a neurone arrangement different from that which obtains in the synthesis of a reflex movement by the convergence of the reflexes of allied arcs upon its final common path” (p. 383). Finally, it should be mentioned that SHERRINGTON firmly believes in the right of psychology to existence and in the mutual helpfulness of physiology and psy- chology. Their workers should, in his opinion, give close attention to one another’s results. Of comparative psychology he says: ‘“‘Despite a protest ably voiced by fv. UEXKULL, comparative psychology seems not only a_ possible experimental science but an existent one’ ” (p- 307). A few months ago in reviewing Jennincs’ “ Behavior of the Lower Organisms’ I saw good reason to characterize it as the most important book on animal behavior that had ever been written. To that statement of my opinion I may now add that “al ' Four. of Phil., Psy. and Scientific Methods. Vol.3, p. 658. 1906. 208 Literary Notices. SHERRINGTON’S book is an equally important analysis of behavior from the side of nerve physiology. JennrnGs has indicated the essential features in the behavior of the lower organisms and the problems which are presented to the student of the evolution of organic activity; SHERRINGTON has shown a way to the scientific study of the behavior of the vertebrates and has made an important contribution to our knowledge of certain forms of activity in the higher animals. ROBERT M. YERKES. Bean, Robert Bennett. Some Racial Peculiarities of the Negro Brain. American Fournal of Anatomy, vol. 5, No. 4, pp. 353-432, With 16 figures, 12 charts and 7 tables. September, 1906. To quote the author’s own summary of his work, this is “‘an effort to show by measurement of outline drawings of brains in different positions, by composites of these outlines, and by actual drawings from individual brains that there is a difference in the size and shape of caucasian and negro brains, there being a depression of the anterior association center and a relative bulging of the posterior association center in the latter; that the genu of the corpus callosum is smaller in the negro, both actually and in relation to the size of the splenium; and that the cross section area of the corpus callosum ts greater in relation to brain weight in the caucasian, while the brain weight of negro brains is actually less. The amount of brain matter anterior and posterior to the fissure of Rolando is roughly estimated, but other points of possible difference, as in the gyri, the insula, the opercula, the Affenspalte, the proportions of white and gray matter, and the cerebro-cerebellar ratio are necessarily omitted in this study.”’ Only those who have attempted to institute exact comparisons between the brains of representatives of different human races can fully realize how great a debt of gratitude we owe to Dr. BEAN (and to Professor Matt and Dr. Hrpiicka, who suggested the investigation) for this simple and graphic method of exhibiting the contrasts in form and in the proportions of various parts of the brain in negroes and people of European extraction. Many writers have called attention to individual points of racial difference in the brain and some investigators have attempted to indicate these differences in figures; but hitherto no one has given us so comprehen- sive a means of expressing in exact measurements the distinctive features of the brain as a whole and the relative size of those parts which exhibit distinctively racial and sexual variations in form and magnitude. The method adopted is very simple and perhaps even crude—as, in fact, every attempt at expressing in figures the distinctive characters of such a complicated organ as the brain in a large series of examples is bound to be—yet it admirably serves the purpose for which it was invented, 7. e., to present in a numerical form the broad contrasts in the form and proportions of the brain in different races and SEXES. The memoir is based chiefly upon the results obtained by the measurement of 152 brains, of which 103 were “American negroes” and the rest “American caucasians.” All the measurements were made from an arbitrary axis proposed by Dr. Mart—a line passing in the mesial plane “ just above the anterior commissure and just below the splenium.”” By measuring the distance to the surface of each hemisphere along radii drawn from the center of this line at angles of 60° and 120°, respectively, to the anterior half of the axis in three planes—vertical sagittal, Literary Notices. 209 horizontal and midway between these two, 7. e., a plane making an angle of 45° with the horizontal—it is possible to obtain three pairs of measurements in each hemisphere, by which the relative development of the frontal and parietal association areas can be compared the one with the other, as well as with those of the other hemisphere and of other brains, and can be expressed in an arbitrary numerical form which lends itself to statistical treatment. As an example of the results obtained by the use of this method I might quote some figures taken from Table Ila (p. 366). In a series of 34 brains of white men and 43 black men, in which the average length of MaLv’s axis is the same (168 mm.), the average ciceniee of the center of the left frontal area (the place where the 60° radius cuts the surface in the plane 45° above the horizontal) from the center of the axis is 70 mm. in the whites and only 66 mm. in the blacks, whereas the center of the left parietal area (measured along the 120° radius) is 71 mm. in the whites and 73 mm. in the blacks. The time will soon come when we must attempt to estimate the exact area and volume of gray matter in each different histological area of the cerebral cortex in a series of brains of different races. This possibility has become opened up by the discovery that in perfectly fresh human brains the difference in the color and texture of the different cortical areas is quite recognizable by the naked eye, so that each area can be cut out and by snipping away the white matter (with a scissors under water) the cortex may be spread out in one plane and its exact extent esti- mated. This method, however, is so exceedingly difficult and exacting that it will be a long time before a large series of records can be obtained. Until this is done the results obtained by the much simpler method devised by Dr. Matt will provide us with information of the utmost value. In the absence of any absolute measurements of the extent of the cortical areas, the exact size of the surface of the corpus callosum and its various parts as exposed in mesial sagittal section is the surest guide to the relative development of the cerebral cortex and its parts that we possess at present. It is, therefore, particularly instructive to note that the separation of the races exhibited in Dr. BEAN’s table of the relative proportions of the anterior and posterior halves of the corpus cal- losum, and of its genu and splenium, is much sharper than that shown in the cruder measurements of the radial distances of the frontal and parietal areas from an arbitrary point. The results obtained by Dr. BEAN in his comparison of the left and right hemi- spheres (pp. 37-373) are particularly interesting, seeing that they agree with the evidence yielded by other modes of investigation and give numerical expression to the differences. The fact that a “smaller posterior association center” 1s found “‘on the left side of the caucasian”’ (p. 371) is a very instructive observation when it is recalled that as a general rule the left occipital region is much more pithecoid than the right because the visual cortex has been pushed backward toward the mesial surface by the parietal expansion to a much less extent than in the other hemisphere. The fact that there is “‘a more marked racial difference on the night side than on the left” (p. 372) 1s borne out by my own observations that in all human brains the left parieto-occipital region shows a tendency toward a simpler and more ape- like conformation than the nght and that in negro brains there is much less asym- metry in this area than there is in non-negro races. ‘This implies a much greater 210 Literary Notices. dissimilarity between the right than the left sides of the brain in blacks and whites. For Dr. Bean’s interpretation of the psychological significance of his results, the reader is referred to the original. G. ELLIOT SMITH. Guyer, Michael F. Animal Micrology. Practical Exercises in Microscopica) Methods. University of Chicago Press. 1906. 240 pp. $1.75 net. This work is intended as a laboratory manual in microscopic anatomy and embry- ology for college classes. While the work is much more complete than the pam- phlets of laboratory outlines usually furnished to the histology classes of the medical schools, it is by no means designed to replace the larger standard works of reference on the microscope and microscopical methods. It does present in very clear form a judicious selection of methods, including an excellent untechnical account of the microscope and its optical principles, adequate for the undergraduate course in histology. The author has also succeeded very well in his attempt to include the minor cautions (usually omitted from the manuals on methods) necessary to enable the beginner to avoid failure or correct it. In an appendix is given an extensive table of the more important tissues and an approved method of preparation for each. The nervous tissues have not been emphasized in this book, for they are adequately treated in Harpesty’s “ Neurological Technique,” though the neuro- logical methods necessary in a course in general histology are briefly given. é. Jo B. BOOKS AND PAMPHLETS RECEIVED. Edinger, L. Ueber das Gehirn von Myxine glutinosa. Aus dem Anghang zu den Abhandlungen der K6nig. Preuss. Akademie der Wissenschaften vom Jahre 1906. Berlin. 1906. Terry, Robert J. The nasal skeleton of Amblystoma punctatum (Linn.). Reprinted from Transac. Acad. Sct. of St. Louis, Vol. 16, No. 5. 1906. Wright, Ramsay. An early anadidymus of the chick. Reprinted from Trans. Roy. Soc., Canada, Second Series, Vol. 12, Section 4. 1906. Dieulafe, Leon. Morphology and embryology of the nasal fosse of vertebrates. Translated by Hanau W. Loes, St. Louis. Reprinted from Annals of Otology, Rhinology and Laryngology. 1906. Peters, Amos W. Chemical studies onthe cell and its medium. Part I. Methods for the study of liquid culture media. Reprinted from Am. Four. of Physiology, Vol. 17, No.5. 1907. Tower, W.L. An investigation of evolution in chrysomelid beetles of the genus Leptinotarsa. Publications of the Carnegie Institution of Washington, No. 48. 1906. Linton, Edwin. Note on the habits of Fierasfer affnis. Am. Nat., Vol. 41, No. 481. 1907. Wilder, B. G. Anatomic nomenclature: an open letterto Professor Lewertys F.Barxer. Re- printed from Science, N. S., Vol. 24, pp. 559-560, Nov. 2, 1906; also the same reprinted from American Medicine, N.S., Vol. I, Nov., 1906, p. 459. Neuropathological Papers from the Harvard University Medical School. 1905. Correction.—In the last number, in the List of Books and Pamphlets Received, B. G. WitpER’s paper, ‘Some Linguistic Principles,” etc., should have been credited to the American Philological Association. - The Journal of oe Siete and psygioregy VotumE XVII MAY, 1907 NUMBER 3 CONCERNING THE INTELLIGENCE OF RACCOONS. BY Ibi Wilco (COVED (Professor of Psychology, University of Oklahoma.) Witu Two Ficures. CONTENTS. LO CUCL OTM ERP Cea EST tee hoPore. tos cos REM goat acc te te Mn eee NL cad ih aoatidletn Suz abet leeds 211 Wearnin cqeOpeleaSewbla SEMIN OSE pars) «ae yeu. hs catepare ee totetener= cae es cuenta uclarcnes te eeiate aUetsiay aioe a eke pe, MV Rer rye wpb aSE@HUIa CS oeWey sey © cect nd oc foc oie se ciel-os cans heres eh wals, Ghee TE tos enero ee 225 DD rctcral TALI Mma ae eee eset ay ete: ssc ohsac tessa ha tee ae es ee OTE MC ee ae OE eee 226 LERRIURAY RORY” 9, Such cee SE BSc ER Ce eR eC ERE ee ileal de eae i Re 232 ealnincaiLompbempbut Mbtouphvan Acts «hoch. eae ayer eee ele elses sete eae 235 Onitheveresencesots Mental Tima ges «aio.cieyeo ates = seis oo ouctercyood ie epee Sele fey ccthe e eesTsiey wi negssea tere 249 SUMAN Masai Pore sseictore setae cae ate Se cvocaperes ba-w ai tae so P apt eat Sarte oe Spear aida ee ele eee a epee amie 248,261 INTRODUCTION. This paper is a report of experiments which were made with the raccoon, Procyon lotor, to determine what type of associations it is able to form, the complexity and permanency of its associations, and to ascertain whether mental images and a tendency to imitate are present in this animal. ‘The paper as originally planned con- tained also observations on the senses, instincts and habits of raccoons, together with comparisons of their behavior with that of other mammals under similar experimental conditions. “Uhese observations, as well as most of the tables on “ Learning to release fastenings,” have necessarily been omitted from this article. I am greatly indebted to Dr. R. M. Yerkes for valuable criticisms of the experimental results and for many suggestions for the prepara- tion of this paper. In all I have had six young raccoons under observation, four males and two females. ‘The descriptions refer in the main to the four individuals which were received July 3, 1905. With these individuals experiments were made twice each day during 212 “fournal of Comparative Neurology and Psychology. the remainder of the summer and almost daily from September to the following May. Subsequently, series of new experiments and repetitions of old ones were given at irregular intervals. Dur- ing each series of experiments, however, the successive tests were made on consecutive days, so that the conditions of hunger and fatigue might be as nearly uniform as possible. ‘The four rac- coons must have been, at the time I secured them, about eight weeks old. Comparisons with the other two, Nos. 5 and 6, whose age | definitely knew, make the above estimate fairly accurate. The four raccoons are designated by the numbers 1, 2, 3, and 4 respectively. ‘The reader should remember that No. 4 alone is a female. When no ambiguity results, I shall use the word “ani- mals’? as a synonym for the word raccoons, or, in other connec- tions, for the expression “‘dogs and cats.” “This usage, however, does not imply the opinion that different mammals are alike psy- chologically. It was my purpose to use tests so similar to those already used in the case of other animals that I might learn by comparisons the place of the raccoon in the scale of mammalian intelligence. ‘This purpose was rather strictly adhered to except in the experiments to test visual discrimination and the presence of visual images. I am not aware that the card-showing device and my method of using it have been employed by other investigators. LEARNING TO RELEASE FASTENINGS. The method employed in the experiments with fastenings was that used in the laboratory by THORNDIKE! in his study of cats and dogs. ‘The peculiar facility of the raccoon in the use of his forepaws and his tendency to investigate objects by touch suggested at once that he might learn readily to operate simple fastenings. Before proceeding to a description of the fastenings used, and the tabulated records, I may say that I do not believe that my raccoons can fairly be called “victims” of experimental conditions. As long as they continued to suckle, or until August 30, 1905, they were fed from a bottle twice each day until fully satished. During August, bits of apple, lumps of sugar and water were added to their *Tuornpike, E. L. Animal Intelligence. Psychological Review Monogr. Suppl., vol. 2, no. 4. 1898. Coie, Intelligence of Raccoons. DTG. bill of fare. From August 30 to the present time the animals have been fully fed and given fresh water once each day. They had to work hard for their food and it is possible that their growth may have been retarded slightly, but I think that this was not the case, for a table of their weights during the period shows a fair increase with age. The year-old raccoons apparently are not quite full erown. ‘The animals have been kept in a room 14 by Io feet, in which they could climb about on several rolls of poultry wire which were hung on the walls. “They had there a nest of hay. Two windows, in which they frequently sat, were open in summer. They could climb about, and they were frequently let out to follow me over an open field, climb nearby trees, or play about the house. Since their eighth week they had never experienced any other environ- ment. No one of the four ever showed a tendency to pace up and down in the windows barred (?) with poultry wire. Raccoon No. 6, however, did this by the hour, and if chained by the neck he would continue to pace to and fro at the end of his tether. This is often observed in captive raccoons. ‘That no restlessness ever appeared in the four would seem to be evidence of their general contentment and of nearly normal conditions in their unusual environment. The animals worked well and, although they possibly might have formed certain associations more rapidly under the stimulus of what THORNDIKE calls ‘utter hunger,” I believe that my results indicate approximately their normal rate of learning. ‘The cases of slow work, due to approaching satiety, were noted and valued accordingly. Readers of the tables will note that the loss of time due to fatigue or satiety is small in many cases, for the ani- mals often work merely for the sake of working, or, more probably, playing. Too great hunger results in much eagerness to secure food and this seems invariably to prolong the time of escape from the experiment box. This is to be observed in the case of the first trial each day for each animal. Description of Fastenings.—In the following descriptions the dimensions of the boxes are given in inches; length, breadth and height being stated in order. The doors varied much in size and in their position in the front of the box. Some were high in the front, others low; some were in the middle from right to left, others at one side. Since none of these variations delayed the animal’s attack on the fastenings, I soon ceased attempting to construct 214 “fournal of Comparative Neurology and Psychology. uniform doors. Some doors were hinged at the bottom, others at the right or left side. ‘“Uhis variation also seemed to have little effect on the animal’s work, for, after experience in one or two boxes, he seems to attack fastenings rather than doors, unless as happened once, shaking the door would release it. In that case, the door was attacked in about one-half the total number of trials. My difficulty with single or double fastenings was not in mak- ing one sufficiently easy for the raccoon to operate, but rather in making one difficult enough. Box 1. 20” x10"x 13”. This box had a door in front, hinged at the right (looking outward), and fastened by a button at the left. The door opened outward the instant the button was turned to a ver- tical position. This box had solid sides and back but the front was made of upright slats 14 inches apart. This fastening is very similar to Kinnaman’s’ A 1, and it corresponds to THorNnpike’s Box C, save that the door did not drop inward. Had the raccoon’s forepaw been bruised or even rapped sharply by the falling door he would have hesitated to open it again. Box 2. 14” x 13” x 26”. This box had a door in the front six inches from the bottom. It swung outward, was hinged at the right and fastened by a vertical bolt at the top. To this bolt was fastened a cord which passed over a pulley, then down through the top of the box and ended in a loop which hung near the side of the door. Pulling down on the loop raised the bolt and allowed the door to swing open. This box, which we may call “Loop at front,” is comparable with THornpixr’s Box A, “O at front.” Box 3. 26” x 14” x 14”, had entirely closed sides with the exception of the front, which was made of vertical slats about one inch apart. On a level with the floor of the box and in the middle from right to left was a front door. This door was hinged at the left and fastened at the right by a horizontal bolt, to which was attached a string which ran in a horizontal position parallel with the front of the box but outside it. The door could be opened by reaching through between the upright slats and clawing the cord which was attached to the bolt. A loose piece on top of the box enabled the experimenter to put the raccoon through the top and then to close the opening by replacing the piece. This is designated in the tables as ‘1st put-through box,” because of the kind of learning it was designed to test. I have compared it with THornpike’s Box E, ‘‘String outside.” Box 4. 14” x13” x26”, “‘Loop at back.’? This was similar to Box 2, ‘“‘Loop at front,’’ with the addition of a second pulley at the back of the top of the box. The cord passed over both pulleys so that the terminal loop hung in the back of the box. The door was six inches above the floor of the cage. This box is comparable with Tuornpike’s Box B, “O at back,” save that the string could not be clawed where it passed along the top of the box. The only way to open the door, therefore, was to pull down- ward on the loop. Box 5. 14” x 13” x 26”, ‘‘2d put-through box,” two fastenings. This was Box 2 with a button added. To open the door it was necessary both to pull the loop and turn the button. Either might be done first. A door was also added at the side of the box through which the experimenter could push the animal into the cage or through which the raccoon could walk into the box. This side-door extended down to the floor. Two raccoons, No. 4 and No. 3, were put through the acts necessary to open the door, the other two were not put through. I have compared this with THorNp1Ke’s Box J, “double.” Box 6. 14” x 13” x 26”, two fastenings, was Box 5 except that the loop was now hung in the center. This change was made to test whether the raccoons ‘‘would claw at the place where the loop had been,” whether this arrangement would change the order in which acts were performed, and whether they would associate this loop with the loop in the other position. Box 7. 32” x 20” x 20”, two fastenings. Both the sides and the top of this box were made of slats so as to admit light. The door in the middle of the front, hinged at the bottom, swung, or rather fell, outward when the fastenings were released. The latter consisted-of a button at the right of the door and a bolt at the top operated by a loop in the back part of the box. The raccoons went into the box through a door opposite the front door. This box was much larger than the preceding ones so that the *Kinnamay, A. J. Mental Life of Two Macacus rhesus Monkeys in Captivity. American Fournal Psychology, vol. 13, pp. 98-148, 173-218. 1g02. CoLe, Intelligence of Raccoons. 215 relative positions of loop and button were changed. The object was to see whether these changed posi- tions would delay escape or whether the fastenings would be at once attacked as if recognized in the new positions. This box also has been compared with Tuornpike’s Box J, ‘‘double.’”’ Box 8. 32” xe20” x 20”, three fastenings. This was Box 7 with an added loop. Thus we had loop 1 at the left side of the back part of the box, loop 2 at the right side of the back, and button 1 at the right side of the door. I have compared this with THornpike’s Box L, which consisted of ‘‘A (O at front), D (string), I (lever).” It is also comparable with Kinnaman’s F 31. Box 9. Four fastenings. This was Box 8 with an added button atthe left side of the door, ‘‘ button 4 9? Box 10. 26”x13”x14”. The ends and back of this box were entirely closed. The top and front were closed with slats only. In the middle of the front was a door hinged at the left. The door was fastened with a thumb-latch which could be released with slight pressure. The bar of the thumb latch would fall back in place unless the door was pushed out a little. This is comparable with THornpikr’s Box G, ‘‘Thumb-latch.” Box 11. Five fastenings. This was Box 9 plus the thumb-latch which had been learned singly. There were, therefore, 2 buttons, 2 pulleys, and 1 latch. The latter had to be operated last lest its bar fall back into the catch. Box 12. Six fastenings. A third bolt was added to Box 11 but the cord from it extended to a treadle or platform which extended across the right end of the box. Depressing the raised end of this treadle released the bolt. Box 13. Seven fastenings. This was Box 12 with the addition of a herizontal hook at the left side of the door. For convenience I used the following notation in recording: 1=button 1, 2=button 2, 1/=loop 1, 2’ =loop 2, 5=thumb-latch, 6=treadle, 7 =hook. Box 14. Hook. 26” x13” x14”. A door hinged at the right and fastened with a horizontal hook was placed in the middle of the front. The animals were put into the box through a door in the back. I have compared this with Kinnaman’s Box 12, ‘‘Horizontal hook.” Box 15. 50” x 20” x 20”, one fastening. Imitation. This box was divided into two equal com- partments. A door at the back admitted an animal to either compartment and a door in the partiton allowed me to change a raccoon from one compartment to another. ‘The right compartment only had a door in front, which was fastened by means of an old-fashioned barn-door latch. This consisted of a wooden bolt which might be pushed to and fro from either side of the door by means of a pin which passed through the bolt and through the door. Pushing the bolt to the right unfastened the door and it could then be pushed open. The plan was to place a raccoon in the closed compartment and let him see another open the door and get out. After this had been done many times the observer was to be let into the other compartment in order that I might observe whether he had learned by seeing the other open the door. All sides and the partition were made of poultry wire so that I might count only those times the imitator apparently saw the act performed and so that he could readily see the performance. Box 16. Imitation. This was Box 15 plus a second latch placed below the first one. This was a dificult box to open because pushing either latch to the left fastened the door. In the early trials of course the animals pushed the latches first to one side, then to the other. : Box 17. 10” x 10” x 4”. Imitation. This box had solidly closed sides. A three inch square was sawed out of the top and replaced to close the opening. Round holes at the corners of this square enabled the raccoon to claw it out and he could then reach into the box and get food. The animals secured food by getting into this box, instead of getting out of it. Box 18. 36” x 24” x 14”. Varying means to an end. An opening was made in the center of the top large enough for the raccoon to go in and get food. This opening could be closed and fastened. The box, which had no bottom, rested on a foundation of a single row of bricks. Removing a brick enabled the animal to crawl through the foundation. The object was to see whether the animal would change promptly from one opening to the other when the opening through which he had been going was closed. If so, perhaps there was some notion of apple-in-box instead of the imageless coupling of a fixed set of muscular movements with a fixed sense impression of the box. Box 1g. 21” x 183” x 20”, two fastenings. A door in the middle of the front, hinged at the bottom, was fastened by a bolt at the top, operated by a loop inside the cage. It was also fastened by a stick leaning against it from the outside. In addition it had to be pushed open. This is the same as THoRN- pIKe’s Box J. Box 20. 21” x18}” x 20”, one fastening. Same as Box 19 except that the bolt was removed. Thus the door was fastened only by a stick leaning against it from the outside. Box 21. 203” x11” x 114", three fastenings. This box had a door in the middle of the front, hinged at the bottom and fastened by a lever at each side and also by a wooden plug which was thrust obliquely 216 ‘fournal of Comparative Neurology and Psychology. into a hole inthe door frame. The box had entirely closed sides. The raccoons were taught to go into this box to get food. At first both levers had to be pushed up but later they were arranged so that considerable force would push them downward. The plug was very dificult to draw. I have com- pared this with THornpIKe’s Box K. Observations on Reactions to Fastenings. —The raccoons learned very readily to perform a certain act in a particular situation. This learning is doubtless of the trial and error type, yet when a latch has been operated a few times there 1s probably present in the animal’s mind a distinct memory image of the act, including a memory of its difhculty. Experiments with colored cards, to be described later, gave evidence in supportof this opinion. At first I supposed are as was true in [THORNDIKE’s work with cats, the raccoons would be found to learn chiefly from the stimulus of hunger. As already stated, however, they soon showed a ten- dency to unfasten latches and set themselves free from the mere pleasure of performing the act. This motive was not strong enough to overcome the discouraging difficulties of a box of six or seven fastenings, but the tables show so-called “play trials” for all boxes of fewer fastenings. The term “play trials” means, then, that though the animal unfastened the latches and escaped from the box, he refused to eat or drink milk on coming out or at best merely tasted the milk and turned away from it. Generally this work was deliberately done, but often rapidly. It seemed, therefore, that this tendency to be occupied was the motive for some of the raccoon’s normal learning and careful records were kept of all play trials. In Table I, I have indicated the cases that were certainly play trials with italics, but the cases which were certain to the observer were fewer than the actual number, for as the raccoon’s hunger was gradually allayed he worked partly for the mere pleasure of doing the work and partly from hunger. This is shown in the longer times taken to escape toward the close of each day’s work. When the animal showed any eagerness for food, the reaction was recorded as a hunger trial even though play trials had preceded it. ‘The tables show that this was unusual, the rule being that play trials began only when hunger began to be satished. Even when using the most complicated fastenings I did not employ “utter hunger.” I usually gave the raccoons considerable food after I had finished the day’s experiments. In several trials with the raccoons, when they were young, I was able to get one to work, which otherwise would not do so, by bringing Coxe, Intelligence of Raccoons. 217) another near the door of the cage. As they grew older this was of no use. Under ordinary experimental conditions the motives from which a raccoon learns are, therefore, hunger, an apparent desire to be occupied, called by several writers curiosity, and in the young, loneliness.* One may ask, were not the play trials actuated by a desire to escape from the narrow confines of the box? I cannot say so with certainty, for all four raccoons would go into a box willingly enough unless it took prolonged work to escape. In that case it was difficult even to put them in, and they developed a tendency to snap at the experimenter’s hand before he could withdraw it from the box. Evidently the memory of previous hard work to escape was the cause of this resistance, for with easy fastenings the animal would re-enter the box time after time and then delib- erately work the latches as a part of an aimless activity which included toying with loose objects, reaching out with, the forepaws through the slats or trying to pull dust or straws into the cage. A change of food from meat to sugar at this moment would often stimulate the animal to escape instantly. Without some such stimulus as this the animal might not come out of the box when the door swung open or it might come out very slowly. Reluctance to re-enter a box being in direct proportion to the difficulty of its fastenings, | can but believe that the raccoons felt no sense of confinement in a box which they knew how to open very quickly. At any rate their behavior toward re-entering easy boxes was the exact reverse of that toward re-entering difficult ones. The conditions which prevented quick working of the mechan- isms and consequently delayed the forming of an association were too great eagerness due to hunger, approaching satiety and dis- traction of attention. (1) Hagerness.—In most cases the first attempt each day, or each half-day, required more time than succeeding attempts even though the animal had operated the mechanism quickly many times before. ‘The eagerness seemed in most cases to amount to great excitement. In the first trialthe animal seemed to fall back on primitive impulses. It made many ineffective movements. In the second trial each day it seemed to depend on memory, and often made but one movement for each latch. *I do not mention the motives of pain, danger, etc., as they were not employed in this study. 218 “fournal of Comparative Neurology and Psychology. (2) Approaching satiety usually, but not always, inhibited uick work. The animals seemed to form associations more rapidly when their work was deliberate. (3) Distraction of attention inhibited all work. The animals never seemed to work a latch as a purely reflex performance. Consequently, I never could get them to claw where a loop had been or when the door was open, as THORNDIKE’S cats did. ‘The nearest approach to this occurred with Box 3. This box had its fastening at the right of the door (looking outward), while Box 1 had a button at the left and the loop of Box 2 hung at the left side of the door. The doors of Boxes 1 and 2 swung to the right, the door of Box 3 swung to the left. No. 4 clawed four times (first, second, third and tenth trials) at the left side of the door (125 experiences in preceding two boxes). No. 3 went to the left side of the door the first four trials in the morning and the first, second and fourth trials in the afternoon of his first day’s experience in this box (200 trials in preceding boxes). A third raccoon, No. 2, clawed twice at the left side of the door after 132 experiences in the preceding boxes. Therefore, after six days of work with latches at the left side of the door, seven is the maximum number of times an animal went to that side of the door in the new box, and four the maximum number of times an animal clawed at that side. In all future new boxes the animals seemed to pick out the new latch and work directly at that, as if experience led them to attack movable objects within the box, or else objects which gave a click or other sound when operated. Only the buttons were noiseless. These facts, with others to be mentioned, indicate, I think, that the raccoon’s learning to operate a latch includes something more than the mere mechanical coupling up of a certain instinctive act with a given situation. In each day’s or half-day’s work, there was usually a slow suc- cess due to eagerness, several rapid ones due to hunger without too great eagerness, and finally several reactions, ae gradually became slower, in which the stimulus was but idle more than a native desire: to be occupied. Most of the latter are recorded as “play trials. I give in Table I the time in seconds for the first forty trials in each of seven boxes. As these results are typical those with other fastenings are omitted. Where the results obtained with other boxes are mentioned in the text the records are quoted with CoLe, Intelligence of Raccoons. 219 sufficient fullness, | think, to verify the deductions made from them. A light vertical line following a figure indicates the termination of one-half day’s experiments, a heavy line indicates the termina- tion of a day’s experiments. Unoccupied spaces preceding figures indicate the number of times a raccoon was put through the act of operating a fastening, for example, No. 4 was put through Box 5 four times the first half day. The records show that putting through is not a great obstacle to the raccoon’s learning as it seems to be in the case of cats. ‘The times were originally taken in sec- onds and fifths, but in this table the nearest whole number of sec- onds is given. For example, No. 2’s first two records in Box 1 were 45.6 and 41.8 instead of 46 and 42 seconds. It we take the average of the times required for all first and second trials with single fastenings for the raccoons not put through the act, we find that they stand approximately in the ratio of 3 to (Boxes I, 2, 3, 4). KINNAMAN’S results show that the male monkey’s first and second trials in the “ Button,” “Vert. Hook,” “Bolt,” and “String and Nail” boxes (I omit the T-Latch Box, as its first time is unusually long), when averaged are approxi- mately in the ratio 2 to 1 (strictly 156:74). After having been trained in seventeen boxes the monkey oe the average time of second trials to one-fourth that of first trials. Among THornpIke’s records for cats there are many failures on second trials, and he insists on the extreme gradualness of the formation of associations in the animals. In rapidity of form- ing associations with single fastenings the raccoons, therefore, stand next to the monkeys. Had we records for four monkeys instead of one, the ratios would probably be still more nearly the same. The table shows that often a raccoon may operate a fastening quickly two or three times, after which there follow immediately longer times. ‘The case seems to be like that of a man who may find a house in the city once by fortunate accident, but only after he has had to search for it does he know where it is in relation to its surroundings. The time records for the raccoons show greater and more nu- merous variations than those for cats, or even those for monkeys. Perhaps this is due to the fact that the cats were utterly hungry and the rhesus monkeys did not exhibit play trials. 220 ‘fournal of Comparative Neurology and Psychology. TABLES Time in seconds for trials | to 40. Box FAastTeninc, Raccoon, 1 5 10 15 TOS SMOG Sciccowesaass No.1 27 56 4 24 8 11 5 \360 10: 10 29 41 40° 44) Speed IROxe Ie BUtLOMe reece ar No.2 46 42 309 F | 5 39) 53) 2545 40129 23 15 “glume Box, Buttons cies No.3 Iz 12 12 75 10 60 330; E 60|80\ 57° 51 20) 27)\>xoumeg IBORe Ti, ISMAOIE Sor oosa Dos] No.4 300: Jo) %65-345 195 42 10 94 Bo 12 5 10 Io 20 —Scammny Box 4, Loop at back ...... INOS i eS GCS Vy i oh, (UN RE hey Gy Bie Box 4, Loop at back....... Noyzi nse 207 12, 50) YO 3G se QO) 2, 0)) 62) Box 4, Loop at back...... No. 3 TGP LHe UES) Re) it Box 4, Loop at back ...... No. 4 Box 5, 2d Put-through Box. No.1 46 25 15 4 3 Qo) 4 5 6). §. Se 2 ee Box 5, 2d Put-through Box. No.2 51 10 26 24 9 BI igee (4) Gea) 2. 2 Box 5, 2d Put-through Box. No. 3 20 40k 17. 8 4: 909| 110 7 eee Box 5, 2d Put-through Box. No. 4 4 Bie ul 1 nt eT Smell Box 9, Four Fastenings..... INOwt 4524 1G) Onna Box 9, Four Fastenings..... INGOs 2s Sy SE i) 9 Box 9, Four Fastenings.... . No. 3 20 16 12 25 12° 16496 qo 4a g28 Box 9, Four Fastenings..... No. 4 on > On Ww Box 10, Thumb-latch..... . Nost « FOF OF. 6 2B WeGrso.i73. 058 7 G28 oe ee Box 10, Thumb-latch. ..... INOa2n QOUS3) 7) 2A eae, we TUT et eee Box 10, Thumb-latch... .. Nonaags 7.) 9: 9n Saas oiluibue ta 2) “0 a 92) 8. ae eee Box m4, Hooke 4. seis Nos tee 1 Gf SO apo IG (a Box A pELOO Ks ay atlte te INS, T/A) ey iG) 7, 9 9] dG VR ep Bow ta, ooke: 2 sc zcie. ne: Nor snag 24) 16. 062896) Wom eee 2 <1 a S|) eee Boxaraveloole semen cerea No. 4 ZR 620 623i AG Box 13, Second Attempt*.. No.1 11 165 86 106 43 72 85 45 24 29 26 22 89 48 28 | 22 Box 13, Second Attempt... No.2 35 20 13 14 661 186 38 37 24 24 35 44 15 45 64 29 Box 13, Second Attempt... No.3 31 746 296 482 601 223 56 39 37 22 27 49 32|19 20 57 Box 13, Second Attempt.. No.4 71 12 27 38 24 17 33 19 15 139 34 35 14 21 29| 31 *In the first attempt with Box 13 all the raccoons failed on the “hook,” They were therefore tried with the hook alone in Box 14. Coxe, Intelligence of Raccoons. 22 TABLE I—Continued. 25 30 35 me Gf Sinies tga agen 25 )| 6 28: a. al W915 94h 30" IQUUSR ON SOMO” 6) Se n6w 75 4 18. Se 4) TS SG. TUB ORNS OMNI ae Gay rai fal OH a SS OI 5 @ iH Bemezemioteee <4 eSae sO) 7 a0) 25) ON 5 5. zlian 2 LT Ss (oil AR Ail > le a ee Ze aga Sh 5 408. OW Ca 20") 5) 8 ae ae Ta Sada Le a2 ye Lr Se ec wrillaren Myke 2 sin” 8 21 75 149 23 263| 29 16 Ag 5 4851s 34 22 01 OOO AS AS. 4a m7 Or 3) 4) 5) gee 4G ete pera Se La iG) 1G 38 oo) eae als Vat oh FO) eter ele Renew x oe ATS. Uae: 1g Pa ge ang ee Jig pS PR I ea a N22 | Sh syd eG Va TROMmOreLGn On. Oo 14|. 9 Me ly a gu 8 ae se igen tanty, 6. is) 9 7. 22 8 12 ig ST oR OB Ray OG die = af ii te, “Th Ah TOW 2h tO 985) 17 15 17 6 6 © § gq ml pe int 1| i ft i rT 2B CI ue a Pe a DEENA Siesta iy. Dey a Oy eee ee ge ee Wrz by ot Zee veo" TOPE DEG Ta vite Gime Caan GE URE Gas Casa 18 GAAS ek Sk a Sam Ce Gy = “GIN GIN 7 ae ellen Weal err? ae) et tg CP Rr eo ees ot Ty Sey By IRS pe ae ereaee Ciyrae ay Cer Meh Oee | ay yeas es en aK To) Ve est Ew T) 2). se aione, ate nie aimed aes ier 18| 226 154 19 49 21 25|386 35 97 26 50 Ta SMA uno 2A STG tg wl eet wee Oe E22 ee ONeOmmTR Luge 45 61 27 | 8 18 31° 180. s5- 22,93 39055 133] 19) 34 3 i ws 2 ag ey a Te AP 1G wy oy) BO te) ul ty 222 “fournal of Comparative Neurology and Psychology. Complexity of Associations.—It was my purpose to compare the complexity of the associations a raccoon 1s able to form with those formed by monkeys. I therefore combined new fastenings with those already learned until in Box 13 [ came very near the limit of their abilities. In Box 21, I also tested the animals’ ability to operate three hitherto untried fastenings and I changed the plan from coming out to be fed to going into the box for that purpose. The animals all succeeded in learning to work seven fastenings: namely, two buttons, two bolts lifted by a pull on each of twoloops hung in different parts of a large box, one thumb-latch, one bolt raised by the animal’s mounting a platform and a horizontal hook placed at the left side of the door. ‘The thumb-latch had to be worked last. “The raccoons thus learned a combination of seven latches. “The rhesus monkey did the same, but no doubt the rac- coons were given more trials. ‘The average time required for the first trial in boxes of two to seven fastenings, inclusive, is 65 sec- onds, for the second trial, 44 seconds. KINNAMAN! gives 25.5 as the average of first trials in similar combinations and 16.5 as the average of second trials. ‘Thus, for both raccoons and monkeys working with groups of fastenings, the time required for the sec- ond success is but two-thirds the time for the first. “The anoma- lous case of raccoon No. 3 in Box 13 in which he required only thirty-one seconds for the first trial, and seven hundred forty-six for the second, has not been included in the average, for to include it would have been to let one peculiar case be equal to eighteen ordinary cases. His third trial is two hundred ninety-six seconds. This case of No. 3 in Box 13 is an example of the fact already mentioned that a raccoon may operate a mechanism quickly once or twice before his actual learning begins. In boxes of two to seven fastenings there is almost no tendency to follow a routine order in undoing them. Occasionally a def- nite order may appear one day and another the next in the same box, but neither is followed very closely. Several hundred experi- ences in Boxes 12 and 13 failed to establish a definite order. The raccoon often seems to begin with the first fastening which attracts his attention. With more than four fastenings each ani- mal showed a tendency to forget a certain one of them, for exam- ple, one button or one loop throughout the day’s training, or per- haps for two successive days. “The case seems not unlike that *Amer. Four. of Psychol., vol. 13, p. 118. Cote, Intelligence of Raccoons. D2 of a man who makes the same mistake each time in adding a long column of figures. Not only was no routine order followed, but often the raccoons worked one or more fastenings more than once, never, however, was a latch operated after the door was open. The temptation is strong to say that the raccoon has no memory of having already released a latch, since he operates it a second time. I think this interpretation of the animal’s conduct incor- rect, however, for in Boxes 11 to 13 inclusive, in which a thumb- latch had to be released last, it was operated almost twice as many times as any other fastening. ‘The animals would work one or two latches, then try the thumb-latch, and so on. This indicates that the animal had a distinct association of the opening of the door immediately after the depression of that latch, that is, the per- ‘ceptual factor of the opening door was a part of the association. No such perception was possible after the release of any other of the latches so they were worked at random because in the past each of them had resulted in the opening of the door. If the door did not open they were worked again. Surely it is asking too much of the animals to expect them to know that each latch when released partly unfastens the door although the door does not move until the last latch is worked. Such a view would demand of the animals either reasoning or a human being’s knowledge of bolts and pulleys. His association, his idea, if he has one, is that the last act opens the door. It is noticeable, too, that the two buttons and the hooks which could be seen to be out of the way when unfastened were not operated a second time nearly so often as the loops and platform which presented no perceptible change of place after having been depressed. Consequently it seems quite as fair to argue that the raccoon pulls a loop a second time because no desired result perceptible to bim followed the first pull, as to urge that he pulls it a second time because he has no memory of having pulled it the first time. Since, therefore, so far as the animal can see, only the final act opens the door and gains the reward of food, the conditions of the experiments were probably quite unfavorable to the acqui- sition of a fixed order in performing the acts. Combination locks, the second element of which could not be unfastened until the first had been operated did not entirely obviate this difficulty in experiments with monkeys (see KINNAMAN, p. 124), So It remains to add to this device some means of making the effect of releasing 224 “fournal of Comparative Neurology and Psychology. each fastening perceptible to the animal. It might at least be arranged that each act should bring the animal nearer to the food. Until this has been done we cannot confidently assert that an animal cannot learn to perform a series of acts in a fixed order. In Box 21, which had three hitherto untried fastenings and in which the plug was extremely difficult to draw, all the raccoons failed in their first and second attempts. The average time required for their first success was 132 seconds, for the second, 85 and for the third, 37. Some failures followed the third trial in the records of all except No. 3. The records in this box serve to show the rate of learning of raccoons compared with the more slowly formed associations of cats. All the raccoons showed a tendency to abbreviate their acts. ‘They would merely turn toward a loop without clawing it or make a slight motion toward it without touching it. Only rarely did one of the raccoons press down two buttons simultaneously. In Box 12, kowever, raccoon No. 3 was observed to try to pull a loop while standing on the platform whose depres- sion raised another bolt. The next day he succeeded several times and finally settled down to doing both these acts at the same moment. A few days later No. 4 had also combined these two acts, and thereafter she did both simultaneously in about one-half the trials. “he other two raccoons never combined these acts. Often the thumb-latch and one button would be worked simul- taneously; but this, we believe, was a mere physical convenience, since the animal could press on the latch with one forepaw and depress the button with the other without changing the position of its body. Variability.—\ have shown that in a series of acts no routine order was established. Was there variation in the method of per- forming the act? Box 14 was fastened with a horizontal hook which could not be raised with the paw and was therefore very difhcult for the animals to learn. All except No. 2 lifted it with the nose; he did the act with his teeth for thirty trials and only twice the first half-day with his nose, and six times the second half- day, up to the twentieth trial. That time he raised the hook with his nose and continued to do so thereafter. He was escaping by means of the mouth reaction in the average time of two and one- half seconds, so he had fully mastered the mechanism before changing thus abruptly to the muzzle reaction. All of the rac- Coxe, Intelligence of Raccoons. 225 coons turned a button once or twice with the nose in early trials then settled down to working it with the paw. In acts so difh- cult to learn that the animal had to be put through them, there was no change from the act put through to one accidentally hit upon. The raccoons were observed to operate fastenings with either the right or left paw or with both at once. We may say in general that the first successful act was not always stamped in because it was not always the most convenient. Sooner or later the more convenient was substituted for the more awkward performance, and the change was sometimes abrupt. We cannot say, of these animals, therefore, that a given situation has power fatally to evoke the formerly successful act. No. 2’s behavior at least was entirely unpredictable. Wherever else in psychology we find the employment of two different means to the same end we account for it by means of an image or notion. But we may speak of this later. MEMORY FOR FASTENINGS. As I built up combinations of fastenings from those which the raccoons had already learned, it was not possible to give memory trials for single fastenings with a time interval sufficiently long to find the limit of their power to remember such acts. Intervals of three or four days or of two weeks showed no appreciable for- getting. After completing work with Box 13, however, I allowed an interval of one hundred and forty-seven days to elapse. ‘This box had seven fastenings and was very difficult for the raccoons to master. At the end of this period No. 3, No. 2 and No. 1 were again tried in this box. Only the firstesucceeded in working all the fastenings and releasing himself. He undid the seven fasten- ings and came out of the box in 34, 28, 131, and 182 seconds, suc- cessively. “The other two worked nearly but not quite all the fastenings, the horizontal hook being most frequently missed. This period, therefore, may be regarded as very near the limit of the raccoon’s memory for the most complex motor associations he is able to form. It seems likely that No. 3’s superior memory for this box was due to the extreme dithculty he encountered in mastering it. No. 2 had had more trials in Box 13 than No. 3 and he is fully as intelligent an animal, yet No. 3, whose difficulties were very great at first, reached the extremely low minimum time 226 ‘fournal of Comparative Neurology and Psychology. of seven and two-tenths seconds and remembered the combina- tion better. Were raccoon No. 3 a human being, we should have no hesitation in saying that he had to give closer attention to the mechanism in order to learn it. If the learning were nothing more than the formation of a habit, No. 2, who had had more experiences with the combination, should have been superior in operating it after a long time interval. Additional memory tests will be described in connection with the tests of discrimination. DISCRIMINATION. Visual Discrimination.—In the tests of visual discrimination no attempt was made to determine whether the raccoons distinguished colored objects by differences in color or by differences in bright- ness. In fact, the greater number of trials required to distinguish two colored objects as compared with the number required to dis- tinguish white from black is, in so far, evidence that the animals were reacting to brightness alone and that the diminished differ- ence in brightness rendered discrimination more difficult. The tests for discrimination of colored objects presented in succession led naturally to a test for the presence of visual images and this question was investigated rather than that of color-vision. I hope in the future to test color-vision. Meanwhile, where colors are named in this and succeeding sections it will be understood that colors exclusive of brightness differences are not implied. In the first tests a modification of the apparatus used by K1nna- MAN in his study of the color perception of monkeys was employed. Two ordinary drinking glasses were covered on the convex sur- face with papers of different colors. Of one pair, one glass was covered with white paper, the other with black; of another pair, one was covered with red, the other with green. ‘The black and white papers were of Milton Bradley manufacture and were of the same intensity respectively as his black and white Maxwell disks. The red and green also were the Bradley standard colors. In the experiments a bit of food was placed in one glass and the glasses were then brought into the view of the animal and placed side by side on the floor, from six to thirty inches apart in different trials. An assistant set the raccoon free facing the two glasses. The animal came to the glasses and secured the food. He was returned to the assistant, food was put in the same glass CoLe, Intelligence of Raccoons. 227 as before and their positions from right to left were reversed. This reversal was made in each successive trial at first, then the feeding glass was left in the same place twice in succession, then three times, so that neither position nor illumination should influ- ence the choice. The usual distance between the two glasses was six or eight inches, for beyond this distance the animal seemed to get his eyes fixed on one of the glasses and to go straight to that one. His reaction was influenced by the direction of his gaze at the moment he was set free. ‘This seems unusual, yet it appeared regularly whenever the glasses were placed from twelve to thirty inches apart. “The distance from the point at which the rac- coon was set free to the glasses was eight feet. No. 4 and No. 1 were tried with black and white. Food was always placed in the black glass. No. 4 was given 25 trials the firs: day, 68 the second and 50 the third; No. 1 was given 25 trials first day and 100 the second. ‘Thus both were practically perfect toward the close of the second day’s test. TABLE. A, No. TI. No. 4. No. of trials. Black. White. Black. White. I-10 4 6 5 5 11-20 5 5 4 6 21-30 6 4 5 5 31-40 4 6 5 5 41-50 5 5 6 4 51-60 3 7 8 2 61-70 (ty 4 9 I 71-80 7 3 9 I 81-go 10 fe) 9 I gI-100 9 I 10 ° JOI-IIO 10 ° 10 ) I1I-120 10 ° 121-130 10 ° 131-140 10 ° After five days without practice No. 4 in fifty trials went directly to the black forty-five times. On the third, seventh, tenth, four- teenth, twentieth and twenty-second trials she went to the white. Two days later No. 1 was perfect in fifty trials, and after an interval of five days in 46 out of 50. ‘The animals, therefore, learn to discriminate black from white in from seventy to ninety trials. No. 2 and No. 3 were tried with red and green glasses. Food was placed in the latter. No 2 was given approximately 120 228 “fournal of Comparative Neurology and Psychology. trials per day for five days; No. 3 approximately 140 trials each day for 5 days. ‘The number of trials per day varied slightly with the degree of the animal’s hunger. TABLE Wii. No. 2. No. 3. No. of trials. Green (right). Red (wrong). No.of Trials. | Green (right). | Red (wrong). I-100 52 48 I-I00 54 46 101-200 50 50 101-200 54 46 201-300 51 49 201-300 ie 47 301-400 68 32 301-400 64 36 401-500 84 16 401-500 52 48 501-595 87 or gi 1970 8 501-600 55 45 601-690 75 or 834% 15 It is evident from these tests that many more trials are required to learn to distinguish red from green, than to discriminate black from white. As already stated, this may be evidence of a response to difference in brightness alone. At this point I devised a “card displayer’’ by which the two colors could be shown in succession instead of simultaneously; it was also necessary to arrange the experiment so that it could be carried on by one person (Fig. 1). The front of the card-displayer consisted of a board twelve inches high. A round pin or pivot on which two levers could be turned was inserted in a hole near the lower edge of the board. In the upper ends of these levers colored cards were fastened so that raising one of the levers to a vertical position displayed red, for example, raising the other displayed green. During one test red would be on the forward lever one inch in front of the other, dur- ing the next test on the rear lever. “The animal could not, there- fore, react to the position of the cards. On account of the difh- Coxe, Intelligence of Raccoons. 229 culty of one person alone having to display the colors, feed the raccoon and keep the record, my notes are not perfectly reliable in respect to the exact number of times required for the mastery of each pair of colors. Consequently when a pair of colors seemed to have been mastered, each animal was given a final test of either twenty-five or fifty trials, an experienced assistant keeping the record. If out of twenty-five trials there were at least twenty- three correct reactions, or out of fifty trials at least forty-five, I assumed that the colors were discriminated. In most of these final tests the raccoon never failed to react to the right (food) card and never attempted to react to the wrong one. ‘lhe response demanded of the animal was that he mount a box 24 feet high by means of another 15 inches high which served as a step to the first, when the food-card was displayed, and that he refuse to go up when the other card appeared. If he started up but returned at once after a second look at the no-food card, the reaction was recorded as correct. If he did not come back immediately from the lower box, the reaction was recorded as incorrect. On the other hand, he was required to go to the top of the two steps when the food card was displayed. According to the above standard the animals learned to dis- criminate the following pairs of cards of different colors and inten- sities. No. 4. Black-white, Black-yellow, Black-red. No. 3. Black-white, Black-red, Black-blue, Black-yellow, Black-green, Blue-yellow. No. 2. Black-white, Black-red, Black-blue, Black-yellow, Black-green, Blue-yellow. No. 1. Black-white, Black-red, Black-blue, Black-yellow, Black-green, Red-green. By this method also it always required many more trials for the discrimination of red from green, or blue from yellow than for the discrimination of black from white, or of black from the colors. The female, No. 4, though given many trials, did not succeed in discriminating red from green, nor blue from yellow, hence in this case the brightness difference seemed too slight to serve as a means of distinguishing the colors. Further evidence that No. 1 distinguished cards of different colors and intensities is given on p- 256. ‘The fact of his discrimination of the series white, orange, blue, from the series blue, blue, blue, whether each series was 230 ‘fournal of Comparative Neurology and Psychology. shown alternately or twice in succession is mentioned on p. 257, and the reactions of all three males to similar series are recorded on p. 259. ‘These tests of visual discrimination may be regarded merely as experiments preliminary to the test for visual images. One peculiarity in the behavior of the raccoons should be emphasized. When they were discriminating well their eyes were never more than 18 inches from the colored cards, more often within a foot of the cards and still more often within three inches, 1. e., the animal took a position with his forepaws on the front board of the card displayer and looked intently for the card to appear. I have never seen the animal look at the card from a distance of several feet and respond to it. “This shows the difh- culty of such discrimination, and it may indicate that the distance for perceiving color or brightness is extremely short. If this be so, the apparent inability to see two glasses when placed 30 inches apart and at a distance of eight feet from the animal is accounted for. Discrimination of Sounds.—I\ endeavored to ascertain the ability of the raccoons to discriminate a high from a low tone and to form the association of being fed at the sound of the high note. The response expected of the raccoon was that he mount the high box to be fed on hearing the food signal. While pure tones should have been used it was, for practical reasons, impossible to do so. I therefore sounded the highest note, A,, possible with an ordi- nary A French harp or harmonica, then the lowest, A”. For the first few trials the hand was extended toward the high box when the food signal was given and the animal fed when he climbed upon the box. When this aid was withdrawn it was found that No. 1 was practically perfect in responding to the high tone and in refus- ing to respond to the low one. No. 2 had not mastered the asso- ciation. His record after the first few trials in which the hand signal aided him is as follows: GTABEE IV. No. 2. No. 2. High-tone, food signal. Low-tone, no-food signal. Right. Wrong. Right. Wrong. I-50 37 13 I-50 34. 16 51-100 44 6 51-100 38 12 101-130 27 3 101-150 48 2 150-200 50 fo) Coxe, Intelligence of Raccoons. ee The animals had completed the visual discrimination tests before they were tried with this pitch discrimination. Discrimination of Forms.—For experiments in the discrimina- tion of forms and sizes the card-displayer already described was used. Cards of different forms or of different sizes were substi- tuted for cards of different brightness and color. In the tests for form discrimination the animal was fed when a square card 6 x 6 inches appeared and not when a circular one 6 inches in diameter was shown. If the animal formed an association between the square card and food so that he went to the top of the high box to be fed when that card was shown and refused to go up when the circular card appeared, we may say that he discriminated. M purpose was to test the discrimination of two objects widely dif- ferent for the human eye, not to test the delicacy of discrimination. The results for form discrimination given by No. 2 and No. 1 appear below. Both animals had already discriminated differ- ently colored cards by this method, so that attention to the cards was well established and the form test proved to be very easy. ADIER We No. I. No. 2. Square. Circle. Square. ‘ Circle. Right. Wrong. Right. Wrong. Right. Wrong. Right. Wrong. I-50 38 12 35 15 3 7 41 9 51-100 47 3 47 3 42 8 39 II IOI-150 48 2 44 6 As a matter of fact, these two cards differed in size as well as in form, but for sensation (barring judgment) I thought this circle to have more nearly the value of the square than one of exactly equal area. However, anyone who will compare two circles with radit of 3 and 33%, inches respectively will, I think, find their visual difference very slight. Discrimination of Sizes.—No. 2, No. 3, and No. 4 were tested in the discrimination of sizes by the method used in the form dis- crimination. “Iwo square cards 64 x 64 and 44 x 44 inches were used. ‘They were first shown alternately, then in varying order. The rapid learning which occurred is due to much previous train- ing in brightness discrimination by this method. It is evident that each animal began to form the association within the first fifty trials, and that learning not to respond to the small card pro- ceeded more slowly than learning to go up when the large card 232 fournal of Comparative Neurology and Psychology. appeared. The cards were not shown simultaneously, but in suc- cession. ‘Thus, remembrance of the card just shown was required for a successful response. On presenting the larger card the ani- mal was fed, if he climbed to the top of the large box. TABLE VI. No.2. No. 3. No. 4. Large. Small. Large. Small. Large. Small. Right. Wrong. Right. Wrong. | Right.Wrong-Right.Wrong. Right.Wrong. Right-Wrong. E50; 2 AT) 8 44 6 29y 4 20. 33." MET On A317 35: (5 51-100 45 5 22) 18 48 YO i 51-100 43 7 Ary 429 IOI-I150 47 3 yo) a 44 es te | 202-250 47 3 23a ma I5I-200 39 II 28," p12 47 Bi S4y LG 252-200 44 6 25) 25 201-250 49 I 47 g | 20-1250 49 2 27 za 251-300 48 2 48 2 IMITATION. Experiments were made to test whether the raccoons imitate one another and whether they would come to perform an act from seeing the experimenter do it. Briefly, I found that the animals not only do not imitate one another, but that they do not pay the slightest attention to one another except when playing, or fighting, or when biting each other gently for the sake of mutual scratching. I give an example of the experiments for the sake of criticism. The method may be inadequate. Experiments arranged so as to attract the animal’s attention to the thing to be learned may still reveal imitation. The raccoons did, in two forms of experiment, seem to acquire an impulse to do an act from seeing me doit. In one, the act was so easy that the evidence is almost worthless, but in the other the act was so difficult that it would seem to be evidence for either imitation or the presence of ideas or both. In other cases, how- ever, the animals failed to learn from seeing me operate a mechan- ism. No. 1 had learned to open Box 16, whose door was fastened by two horizontal wooden bolts, primitive barn-door latches. Throwing both of these to the right released the door; throwing one or both to the left fastened the door. The box was a difh- cult one to open, for having once thrown a latch to the right the chances were that the raccoon’s next movement would throw it to the left. Coe, Intelligence of Raccoons. 233 The box had two compartments separated only by a partition of poultry wire. ‘The imitator facing this partition was near the door so that it was possible for him to see the work of No. 1, in opening it. No. 4 first failed in three minutes. She was then put in the imitator’s compartment while No. 1 opened the door 18 times. No. 4, however, did not see him do it. She was put in the same compartment with No. 1 and still I could never be cer- tain that she saw his acts. She was then held and saw the experi- menter open the door during several series of ten trials each. She continued to fail when left to try alone. Subsequently, I held the animal so that he certainly saw the work of the raccoon he was to imitate. When the door was opened both came out and were fed. No. 4, No. 3 and No. 2 did not learn to open the door from see- ing No. 1 do it or from seeing the experimenter do it. As a further test of imitation | taught No. 2 to claw the small block out of the opening in the top of Box17. No. 1wastien given opportunity to learn by imitating No. 2. He did not watch No. 2’s work. He was then held so that he could not fail to see it. After this he followed No. 2 to the box each time and soon learned to dive into the box as soon as No. 2 pulled out the block and get the food before No. 2 could do so. Left to open the box for him- self, he did not even goto it. Hewas then held and saw the experi- menter remove the block three times. ‘Then he began to claw at the block while it was being removed. He did this twice more and then was perfect in the performance of the act. No. 3 failed after four minutes to remove the block though he clawed at it somewhat. Apple was then placed in the box and he was loosed just in time to see the experimenter remove the block. He reached in and got the apple. ‘This was repeated ten times. He then clawed out the block instantly though it had been put in tightly. No. 4, however, did as well with no chance to imitate. Evidently the act is too easy to learn to be of much value as a test of imitative abilit The card-displayer, however, afforded a more difficult task than I would have planned for the animals deliberately. After having had some six weeks of experience in distinguishing a black from a white card and in distinguishing complementary colors, each of the four raccoons developed a tendency to reach over the front board of the apparatus and claw up the colored cards. ‘This tendency was encouraged and finally they would claw up the right (food) 234 ‘fournal of Comparative Neurology and Psychology card and go to tle high box to be fed, or, having clawed up the wrong (no-food) card they would claw it down. ‘The cards could not be seen until they had been lifted up and they were difficult for the animal to raise. ‘Therefore there were many errors. So far as imitation is in question, the important point is that the raccoons did begin to do, or try to do what they had seen done by the experi- menter. Before they began this they had learned to watch the cards and the movements of the trainer’s hands very closely indeed. Therefore, the animals either imitated or else from their impatience to see the right card come up there sprang the idea that they them- selves might make it come up. ‘This, however, may be all there is in intelligent imitation. I stimulated their impatience by moy- ing the cards slowly, and the clawing soon began. ‘The whole problem, in the case of these animals, may be one of attracting their attention to the thing to be done. Perhaps seeing a thing done often enough will set free in them an impulse to do it just as being put into a box will arouse an impulse to go into it. An important question to ask is, What free impulses is the animal capable of acquiring? ‘Thus far we have at least two: an im- pulse to enter a box into which it has always been lifted; and an impulse to claw up color cards which it has previously merely seen raised. Such impulses must accompany ideas acquired from the experience of being lifted in and of seeing the card raised. This card-displayer test of imitation has an advantage over those with latches, inasmuch as the animal did not at first fail. He simply passed from seeing a thing done to doing it himself. Since the raccoons do seem to develop a tendency to do an act they see done by an experimenter, it seems possible that were one raccoon made dependent on another for all his food he might de- velop a tendency to imitate the food-getting acts of the other. There is good reason to doubt, however, whether even a young raccoon can be taught to watch another. The animal’s /ife depends upon his finding and getting food before another of his kind gets it, not with that other or ajter him, for nature puts but one bit of food in a place for raccoons and | should say also for chicks, dogs and cats. The bone must be seized and escaped with before another gets it, if another animal be near. Hence nature puts a premium on attention to the bone and punishes with hunger any tendency to watch another animal getting food. ‘Therefore, I think it unlikely that imitation of another will ever appear in Coxe, Intelligence of Raccoons. 235 these animals in connection with the food-getting impulse. Almost all experiments so far employed in laboratories have depended on hunger as a stimulus. Perhaps a new motive should be searched for to test the presence of imitation. Such an opinion certainly seems warranted by the behavior of raccoons. [| think the same is true of dogs, cats, and chicks. In monkeys, however, KINNAMAN (p. 121) elicited two examples of undoubted imitation of one rhesus by another, in connection with food-getting, and apparent cases of “instinctive imitation’? were numerous. May this difference not be attributed to the fact that monkeys’ live in groups or droves and search for stores of food rather than for single bits as the raccoon does? LEARNING FROM BEING PUT. THROUGH AN ACT. The evidence for ‘)HORNDIKE’S most far-reaching conclusions concerning the mental life of cats and dogs seems to be based on their behavior in experiments in which they were put through the act to be learned. In view of his conclusions it would seem highly important that this question be tested carefully for as many of the higher animals as possible. On page 67 of ‘Animal Intelligence” THorNnpIkeE says: ‘‘A cat has been made to go into a box through a door, which is then closed. She pulls a loop and comes out and gets fish. She is made to go in by the door again, and again lets herself out. After this has been done enough times, the cat will of her own accord go into the box after eating the fish. It will be hard to keep her out. The old expla- nation of this would be that the cat associated the memory of being inthe box with the subsequent pleasure, and therefore performed the equivalent of saying to herself, “Go to! I will go in.” The thought of being in, they say, makes her goin. The thought of being in will not make her goin, For if, instead of pushing the cat toward the doorway or holding it there, and thus allowing it to itself give the impulse, to innervate the muscles, to walk in, you shut the door first and drop the cat in through a hole in the top of the box, she will, after escaping as many times as in the previous case, not go into the box of her own accord. She has had exactly the same opportunity of connecting the idea of being in the box with the subsequent pleasure. Either a cat cannot connect ideas, representations, at all, or she has not the power of progressing from the thought of being in to the act of going in. The only difference between the first cat and the second cat is that the first cat, in the course of the experience, has the impulse to crawl through that door, while the second has not the impulse to crawl through the door or to drop through that hole. So though you put the second cat on the box beside the hole, she doesn’t try to get into the box through it. The impulse is the sme qua non of the association. ‘The second cat has everything else, but cannot supply that. These phenomena were observed in six cats, three of which were tried by the first method, three by the second.” On p. 73 he writes: ‘‘Presumably the reader has already seen budding out of this dogma a new pos- sibility, a further simplification of our theories about animal consciousness. The possibility is that animals may have no images or memories at all, no ideas to associate. Perhaps the entire fact of associa- tion in animals is the presence of sense-impressions with which are associated, by resultant pleasure, certain impulses, and that therefore, and therefore only, a certain situation brings forth a certain act.” So definite and convincing is his evidence for this failure to learn by being put through an act in the case of dogs and cats, that 236 “fournal of Comparative Neurology and Psychology. I supposed at the outset that my experiments to test this hypothesis in the case of raccoons would be few and perfectly confirmatory of ThorNDIKE’s view. But the behavior of the raccoons on the second and later days of my experiments soon indicated that this confirmation might not be forthcoming. It will be recalled that similar experiments of ‘THORNDIKE’s® on monkeys were incon- clusive, and that the monkeys experimented with by KINNAMAN could not be handled. I took pains, therefore, to handle the young raccoons as much as possible, and they showed no objection to it for many months. ‘Then one refused to be handled. On the second day of my experiments with the female, No. 4, in Box 1 (button), and much to my surprise, she turned, on the thirty-third trial, and went quickly back into the box. She opened the door in six seconds, came out, was fed for a moment from the bottle and then immediately re-entered the box. Now possibly the reader is saying, “Yes, this is the phenomenon observed by ‘THORNDIKE in cats which were pushed toward the door or held near the door.” ‘This is not the case. ‘This young raccoon had been picked up by the nape of the neck, lifted quickly through the door and dropped on the floor of the box. When thus held the four legs of the animal hang down limp as they do in the case of a kitten carried in the mouth of its mother. ‘This fact makes this method of holding and lifting the animal most convenient. ‘There was no innervation of her own muscles. Four days later when tried in this box she went in on the second, third and fourth trials. No. 3, the second raccoon tried in this box, went in himself on the twenty-second, twenty-third and twenty-fourth trials. He also had been lifted into the box on the preceding trials. On the forty-fifth and from the forty-seventh to the fifty-frst trials, inclu- sive, he re-entered the box. On the fifty-second trial he started back but turned at the door and did not go in again that day. Subsequently, he went in regularly until his hunger began to be satished. During his last eighty-five trials in this box he re-entered it of his own accord eighty-two times. On the seventy-first trial, and several times thereafter, he was held near the door to make him go in but this was not done with any one of the animals until they had gone in spontaneously frequently enough to show that it was an established part of the reaction. Moreover, the hold- 5 Toornpixe, E. L. The Mental Life of the Monkeys. Psych. Rev. Monogr. Suppl., vol. 3, no. 5. 1gol. Cote, Intelligence of Raccoons. 227 ing was done simply to make them go in when their hunger was partially satished. Up to that time they were eager to go in, after having done so several times. These remarks apply, to boxes with from one to four or five fastenings. In connection with experiments with Box 13, I several times whipped No. 3 to make him go in, for the box was very difficult to unfasten. ‘This was done, however, only after he had gone into the box repeatedly. If this behavior is to be used as evidence of the presence of ideas, then the reluctance of the animals to enter the boxes when they were not hungry, and when the box was difficult to unfasten is quite as significant as the fact of their getting in spontaneously at fitst. No. 2 started back into Box 1 on the eleventh trial and went back into the box on the thirty-ninth, fifty-sixth, sixty-seventh, sixty-ninth and seventieth trials. “The next afternoon he went into the box and came out to be fed before [ could close the door. I fed him a little, and he went back. After this the usual thing was for him to go into the box when hungry. Until after the seven- tieth trial nothing was done to encourage No. 2 to go back. My object was to see whether the animal would turn and go in instantly entirely of his own accord. I did not even wait for “hin to go 1n; unless he returned promptly to the box, he was lifted into it. No. 1 went into the box first on the fifty-seventh trial. After that he was held at the door six times and went in. ‘Thereafter he went in regularly. These results are radically different from those obtained by THORNDIKE in his experiments with cats. Since four raccoons exhibited this reaction, it is safe to conclude that any raccoon which has been lifted into a box and allowed to come out and be fed will sooner or later go in of his own accord, and further that he will goin before the one-hundredth trial and probably before the seventy-fifth trial, as my four animals did. ‘The behavior of these animals forces one to believe that it dawns on the animal that he can hurry the matter of getting food by rushing back into the box and coming out again. ‘The association here irivolved not only what the animal had done but also something which had been done to it. It may very well be doubted, however, whether lifting the animal about taught it anything. I should say rather that it had an image of the interior of the box as the starting point of the.food- getting process and an idea of going back to recommence the pro- 238 “fournal of Comparative Neurology and Psychology. cess. his idea lost all motive power as soon as hunger was allayed. This difference between the behavior of the raccoons and the cats, occurring as it did with Box 1, led me to modify the succeed- ing experiments so as to test further the animal’s ability to learn without innervating its muscles. In Box 2 the door was placed six inches above the floor to see whether this would prove to be an obstacle to going back into the box. No. 4 did not begin to go in of her own accord until the fifty-first trial, but she did so very often thereafter. No. 3 went into the box on the first trial, that is, before he had ever been put into it, notwithstanding the difference between the positions of the doors and the size of the boxes. One may explain this behavior, which occurred often afterward, either by association by similarity or by inability to distinguish the differences between the two boxes. My opinion is that the open door at once suggested the usual act of going in. Probably it was the same door to the raccoon. This, however, is a crude association by similarity. “‘Similarity is partial identity.”” The differences are entirely unnoticed. No. 2 re-entered the box on the second trial; No. 1 on the fifth. Box 3 was arranged to test further this difference between raccoons and cats. In the first place an opening in the top of the box was covered only by a loose piece of board and the plan was to put the raccoons into the box through this opening, to see whe- ther they would learn this indirect way of entering. Then No. 4 and No. 3 were put through the act of opening the door. This was done by holding the animal, taking its paw and placing it on the string then pressing it down until the bolt was withdrawn and the door opened. No. 2 was not put through the act and No. I was not worked in Box 3 A low step was placed at the end of Box 3 to enable the animals to climb more easily to the top of the box. The order of procedure was as follows: ‘The raccoon came out of a door in the front, was fed, went around to the end of the box, mounted by the step, to the top of the box and dropped through the opening into the box. We may discuss first the act of going in. On the seventeenth trial No. 4 went in. On the eighteenth she was held on the box and went in. On the nineteenth she climbed upon the box. On the twenty-first she was put on the box and went in, and so on to Core, Intelligence of Raccoons, 239 the twenty-eighth trial, in which she may have been helped by the motion of the experimenter’s hand in the direction of the opening. No. 3 after the fifth trial went in when placed on the box. On the eighteenth trial he re-entered the box from the floor of the room, and later he went by way of the step and the end of the box. On the twenty-seventh trial he climbed up over the front of the box and dropped into the opening in its top, thus sub- stituting a direct for a roundabout way. No. 2 went in when put on the box after the eighth trial. From the twelfth trial he went in when put on the step, and from the twenty-second trial he went in from the floor of the room. In the extract from “Animal Intelligence” already, quoted ‘THORNDIKE says, “So, though you put the second cat on the box beside the hole, she doesn’t try to get into the box through it.’”’ This description certainly does not suit the behavior of raccoons. Having shown that raccoons learn to go into a box by being dropped in through a hole in the top, we have yet to answer the question, will the raccoon learn to operate a fastening, to per- form a complicated act, by being put through the motions neces- sary to do the act? In order to make trial of this I decided first to put two of the animals through the act of opening cages and let two of them learn it by trial. If the average time of the first success for those put through should be shorter than the average time for those not put through, it would be fair to conclude that the putting through facilitated learning. In order to make the evidence especially strong I selected for most of the putting through experiments the two raccoons which, up to this time, had shown themselves slowest in learning the mechanisms, Nos. 3 and 4. It seems to me, therefore, that much weight must attach to the averages. he average time required for the first success in each of eleven boxes by the animals which were put through the act is 41.6 seconds; by those not put through go.2 seconds or more than twice the former average. ‘The results are shown in ‘Table VET. In Table VIII the animals which were put through fazled to escape by their own unaided efforts, but succeeded after being put through. [| have other instances of this. It will be seen that in two of the eleven boxes the averages favor those not put through. Box 3 shows an average of 85 sec- onds for those put through and of but 26 seconds for those not 240 ‘fournal of Comparative Neurology and Psychology. TABLE VII. Numser or Times | Time oF First AVERAGE FOR | AVERAGE FoR Pur Turoucu. Success. THOSE PUT | THose Nor Put THROUGH. THROUGH. | Box 3. Single No. 4 | 8 | 162 sec. No. 3 5 9 85 sec. 26 sec. No. 2 none | 26 Box 4. Single No. 4 22 48 No. 3 fe) 7 No. 2 none 150 77 | et No. 1 | none 62 Box 5. Double | No. 4 4 4 No. 3 4 20 No, 2 | none | 57 = * No. 1 none 46 Box 6. Double - | No. 4 none 16 No. 3 | 7 5 No. 2 none 12 7 ‘ No. 1 none IO Box 7. Double | No. 4 | 6 21 | No. 3 6 2 | No. 2 | none 22 44 | 139 No. 1 none 296 Box 8. Triple | No. 4 6 | 24 No. 3 6 25 No. 2 | none 62 ea 39 No.1 none 16 Box. Four | Latches No. 4 6 16 No.3 6 20 No. 2 none 29 a8 2 No.1 none 45 Box Il. Five Latches No. 4 | 6 30 No. 3 | 6 | 39 34 a No. 2 | none 36 No. 1 none 12 Box 12. Six | Fastenings No. 4 6 134 No. 3 6 23 | 8 - No. 2 none 364 “ 2 No.1 none 34 Cote, Intelligence of Raccoons. 241 TABLE VIII. FAILED BY Numeer Times Time ReQuireD AVERAGE FOR AVERAGE FOR Own Errorts | Put THrouGH. For First Tuose Put (TuoseNot Pur AFTER. SuCCESs. THROUGH. THROUGH. Box 10. Thumb- | latch No. 1 840 sec No. 1 go No.1 75 3 | 6sec. |] No. 1 120 I | Gn os r 10 sec. No. 4 6 | 18 J No. 2 none Zhe) 30 sec. Box 14. Hceok No. 4 600 9 | 231 \ No. I 1920 5 | 3 J iat No. 3 none 545 il Ae No. 2 none | 75 f put through. This is due to No. 4 alone, however, for No. 3 (put through) made a record of g seconds, while No. 2 (not put through) made a record of 26 seconds. Box 11, with its five fastenings, gave an average of 34 seconds for those put through, as against 24 seconds for those not put through. ‘This is due to INoo 4s remarkably short time, 12 seconds, on the first trial, and to No. 3’s difficulty in learning the box. The full record of No. 3s learning makes Box 11 give rather conclusive evidence in favor of putting through. After ‘being put through six times No. 3 suc- ceeded with Box a1 eighteen times consecutively. The next morning he failed after twenty-five minutes to pull one of the loops, though he worked the other fastenings. Although he was evi- dently already hungry and worked hard to escape he was left untried for two hours to see whether increased hunger would help him. When next tried he failed after eleven minutes, was put through ten times and succeeded in forty seconds, then in forty- four, then in twenty-nine, and soon. The next morning he failed after seven minutes, was put through six times, then pieeeeded in twenty-three seconds. The next morning he failed in five minutes, was put through twice and then succeeded in thirty sec- onds. ‘The next morning he failed on the same loop, was put through once and succeeded always thereafter, steadily reducing his time. Now this is, in many respects, the poorest record for No. 3 or for any other of the raccoons and I fancy the reader 1s 242 ‘fournal of Comparative Neurology and Psychology. saying, “This one case outweighs all your averages. Do you not see that the animal was hindered rather than helped in learning by being put through?” The answer is, of course, why should not his eighteen consecutive successes, unaided after the first six trials, have stamped in the reaction? Each morning he failed once more (he almost always failed on loop 1, a fastening he had already learned). He failed, also, no matter how long I waited for him. But he never failed immedzately after he had been put through, and each of his successes following the putting through was quick. ‘The fact is the box was very complex for him, he would forget a fastening, be put through, then not fail again that day. The next day the difficulty would reappear. He was very slow to learn this box, but remembered it longer than did any of the others. ‘The point I would emphasize is simply that putting through after a failure certainly and always resulted in making the next trial a success. It seemed, as we say of human beings, to refresh his memory. Would he have failed as frequently and during so many days had he been forced to learn by trial and error, not obtaining food at all until he succeeded, be it a day or a week? I think he would not have failed as frequently after the first success. No doubt the putting through caused him to depend upon it. I do not believe that putting through has nearly so much stamping-in power as a self-innervated movement. It has not for man. A man may be told how to make a shot at billiards but only practice in making the shot will fix it. A player having made the shot once, as directed, may at that time succeed. In later trials he will make it sometimes very awkwardly. So with our animals. Often the first success does not require the longest time either for those put through or for those which innervate their own muscles. ‘These short first times and longer later ones are sufhciently frequent to show a marked difference between the learning of dogs and cats and that of raccoons. I think, finally, that putting through helps a raccoon to succeed in trials imme- diately following the experience of being put through, and that this is a mental effect. It establishes a transient association. Trial and error forms more stable and permanent associations—a reflex affair simply. The description of No. 3’s learning in Box 11 should make it clear that the averages in that box deserve but little weight. They differ by only ten seconds. But, however that table be counted, Coxe, Intelligence of Raccoons. 243 the averages and the number of individual cases in favor of learn- ing by being put through are too widely different from those against It to be ascribed to chance. ‘The experiments from which these data are taken were continued for three months. One who observed the experiments closely might state what appears at first to be a very strong argument against our conclu- sions. For example, in Box 5, which had two fastenings, | called the order in which the animal was put through the two “direct,” the other possible order “reverse.”’ No. 4, who was put through, did the acts in the reverse order roughly two-thirds of the times, and No. 3 probably three-fourths of the times. . I did not succeed in establishing the order in which [ had put the animal through. ‘This is a serious objection until we compare it with the behavior of the animals which were not put through. Let us examine the records of No. 2 and No. 1 for July 19, 1905, calling that order, “direct” in which theanimal attacks the fasten- ings for three consecutive times. No. 2, on the morning of that date, followed the direct order eleven times after he established it. In the afternoon he did it in the reverse order thirteen out of fourteen trials. I will quote from my record of his work the next morning. “Direct, reverse, direct, reverse, direct, reverse, reverse, reverse, direct, reverse, direct, reverse, reverse, direct, reverse, re- verse, reverse.’’ ‘T’his is typical. On the morning of July 19, No.1 did the act eleven times in the direct order and five in the reverse. In the afternoon twice in the direct and seven times in the reverse order. ‘The fact is the raccoons never mechanize the order of their performances into a settled routine. ‘Therefore if at this point of study I were asked the question, Did the animals perform the same act you put them through? I should answer, they did and they did not. ‘They were put through most of these acts with one forepaw. ‘They did the act with that paw, with the other forepaw and with both forepaws and exactly the same is true of those who learned the fastenings by trial and error. ‘The question, however, may be changed to, Can the animal be made to learn the act you put him through and to employ no other? Yes, it hap- pens that this can be easily done with these animals. A more decisive test of the value of putting through would be, of course, one which answers the question, Does the raccoon, by being put through an act, learn to operate a mechanism which it had failed to learn by its own efforts ? 244 ‘Journal of Comparative Neurology and Psychology. No. 1 in his first work in Box to failed because he worked two slats loose and kept attacking them. He first failed in fourteen minutes; was put through, then failed in one minute and thirty seconds; was put through and failed in one minute and fifteen seconds and was put again through. He then did the act in six seconds. He afterward failed once, was put through and did the act in six seconds. ‘This is, as shown above, the usual condition. The putting through helps to a quick success but does not insure permanency unless repeated more times than a reflex perform- ance. Box 14 was most difficult for it was fastened with a horizontal hook which had to be lifted vertically, and the raccoon cannot well lift an object vertically with his paw unless he can stand directly above it. ‘There remained but three possible ways to lift the hook, namely, with the teeth, with the nose, or with the back of the head. The latter was done but three times in all; I think this was because in this case the animal could not see the hook become free and fall. It was really a quick and convenient way of lifting the hook. While No. 3 and No. 2 succeeded in this box, the other two rac- coons failed. No. 4 failed after ten minutes of steady clawing. She was put through ten times by lifting the hook with her nose. She then lifted the hook with her nose after three minutes fifty-one seconds, again in sixty-one seconds, then in twenty-three, then in five, four, five and one seconds successively. Before being put through No. 4 did not attack the latch directly. It was a black hook, the box was of rather dark wood and all preceding latches had been more conspicuous both in position and color. After being put through she worked directly at the latch. If one objects that iN. 4 should have succeeded in less than three minutes, I can only reply that the hook was a difficult fastening, that this is the first time the raccoons had to learn to work with the nose and that I am quite willing to grant that little or no ski// comes from putting through. Finally, let me add that No. 4 always worked the latch with fre nose, by the act she had been put through. No. 2 also failed on the horizontal hook. To make it a certain failure I waited thirty-two minutes while he worked steadily. I put him through five times by raising the hook with his nose. He then dueceeded:i in three and four- enh seconds, then in seven and two-tenths, and so on. Supported by the averages of the table above, these two examples make it certain that raccoons can learn Core, Intelligence of Raccoons. 24.5 an act from being put through it, even though they have failed to learn it by their own efforts. My own opinion is that No. 1 learned the exact act by being put through. No. 4, it is true, may have learned only the placeto attack. ‘To urge this objection, however, amounts to saying that the animal must have got some idea or image of the place or hook from being put through, for surely no reflex act is established in an animal whose muscles are limp. Had I not held my hand beneath her muzzle she would have let it hang down and it would not have raised the hook. So in this case especially the act of putting the animal through with uninner- vated muscles gave her a motive or impulse to innervate the mus- cles. Personally I should judge that the hook lifting with a click and noisily falling, not more than an inch in front of the raccoon’s eyes, was fully as well attended to as the place of attack. No.1 also did not vary once from the act of lifting the hook with his nose. This is important when we compare it with the work of those not put through. I record in Table IX the trials and methods of lifung the hook of No. 2 and No. 3. TABLE Ix. No. 2. ist half-day. 2d half-day. 1. Done with mouth. 1. Done with mouth. 2. Done with mouth * 2. Done with side of nose. 3. Done with mouth 3 to 1rincl. Done with mouth 3. Done with mouth. 12. Done with nose. 4. Done with mouth. 13 to1sincl. Done with mouth. 5. Done with mouth. 16 to 18 incl. Done with nose. 6. Done with mouth. 19. Done with mouth. 7. Done with nose. 20. Done with nose, and always so thereafter. 8. Done with mouth. Total with mouth, 30 times. 9g. Done with mouth. Total with nose, 8 times. 10. Done with mouth. 11. Done with mouth. No. 3. 12. Done with nose. Done with foreleg (not paw). Done with head. Done with head. Done with head. Done with nose, and continued in this way. 13 to 18 incl. Done with mouth. mp Wb It is evident, therefore, that the best method for the animal is to lift the hook with its nose. I have now shown why we should change the question, Does the animal learn the act you put him through? to the question, Can he be made to do so? If the act which he is put through is the one which will remain the easiest and most convenient for him throughout the tests, irrespective of 246 ‘Journal of Comparative Neurology and Psychology. his position 1n the box, he will never vary from it. If not, he will employ your act when his position makes it convenient and he is looking at the latch you began with. He will also vary from it very often but not a whit more often than a raccoon not put through will vary from the act he seems to establish in his early trials. Moreover, an animal may begin a new way sometimes after a hun- dred or more trials, for example, No. 4 combined acts 1 and 6 in Box 13 after several hundred trials; No. 3 combined them much earlier; No. 2 after mounting the platform in Box 13 many times took to lifting it with both paws. When it was dropped the jerk in addition to the weight of the platform would raise the bolt. This was an awkward method and, while it occurred almost con- secutively during three days’ work and now. and then for some time longer, it was gradually relinquished. It would seem that enough experimental evidence has been presented to show that the raccoons do learn without innervating their own muscles. But the opposite condition as found by ‘THORNDIKE 1n cats, namely, that they learn by “trial and error” only, has been made to support so important conclusions concern- ing the mental life of animals, that I shall risk taxing the reader’s patience with a further recital of experimental tests. Se aoe In these experiments I used the card-showing device already described, but I placed a lever holding a color on the front side of the apparatus so that the animal might learn to lift it himself. This could be done either by the nose or the paws. It was easiest at the beginning of the ascent to raise the lever with the nose but hard to elevate it thus completely. It required a vigorous toss of the head to make the lever reach the point where it would not fall back. On the other hand, while difficult to start with the Co.e, Intelligence of Raccoons. 247 paws, it was easy to finish the ascent by that method. The dis- advantages of these ways of lifting the lever were, therefore, nearly equal. Can one raccoon be taught to lift the lever with his nose, another with his paws, thus proving that more than one kind of reaction can be taught the animals by putting them through? And can they be put through the acts enough times to establish the habit in addition to what has been described as an apparent mental effect which is readily forgotten ? All the raccoons were first given trials to determine whether they would hit upon the act of raising the lever. All came to the apparatus and watched it closely. Previous experience had aroused their interest. None, however, lifted the lever nor did I expect it, for the animal was free alone in the room, and while all the individuals clawed at colors back of the board none had ever done so in frontof it. On May 24, 1906, No. 1 was put through sixty times by lifting the lever with his nose. He was then given an opportunity to do so unaided. If he failed within approximately thirty seconds, the lever was raised and he was fed. ‘This was done fifteen times. The seventh time he came and looked at the card when it was down, on the ninth and fourteenth he pushed at it with his nose, and on the eleventh with his paws. Although he had not mas- tered the performance he seemed to have made a slight approach toward it. ‘The next day he was put through ninety-five times and succeeded instantly in performing the act on the ninety- sixth. On the ninety-seventh, ninety-eighth, one hundredth, one hundred-seventh and one hundred-tenth trials he failed, but he succeeded in all the other trials between the ninety-sixth and one hundred-tenth (fourteen in all), and always after the one hundred- tenth. On July 11, without practice in the meantime, he was per- fect in the performance of the act. No: 9*learned@as, did) No. a, although more slowly. He did not fail after his first success, although at first, for ten trials, he lifted the lever only half way up. These ten attempts I rewarded but later he was forced to give the lever a strong toss to get food. In the cases of both No. 1 and No. 3 the movement they were put through was the one they fol- lowed uniformly. On June 1, 1906, No. 2 was put through the act of lifting this lever with his paw one hundred times. He then did the act with his paw forty-five times in succession. On June 7, he did it fifty 248 ‘fournal of Comparative Neurology and Psychology. times without error and on July 11 he was still perfect although without practice meanwhile, As No. 2 pursued one method throughout, and No. 1 and No. 3 the other, it cannot be said that our only crucial tests consist of reactions with the nose which are so forced and unnatural as to be poor evidence. As to the establishment of a habit, the records are ambiguous. No. 2 was as perfect after being put through one hundred times as he would have been after eight or ten accidental successes. “The other two, with more experience, were less perfect. “They did not learn the toss I gave the lever, but expected food for raising it only a trifle and letting it drop back. Withholding the food brought the complete reaction. I must explain how the raccoons showed that they expected food after the abortive performances. On raising the lever the animal stood with his forepaws on top of the front board to be fed. After every abortive effort he would take this position, then, as food was not forthcoming, he would drop to the floor and dive under the lever again with his nose. All the animals added to this reaction the act of clawing the lever down into the horizontal position so that it might be raised again. ‘The experimenter merely had to feed the animal each time the lever was raised, and the work thus became very rapid. Let us summarize this long section: (1) All the raccoons began, of their own initiative, to run back into boxes into which they had hitherto been lifted. (2) All learned to go up to the top of a box and drop through a holeinto the box after having been lifted into the box repeatedly. (3) All, after having learned to go to the end of a box, up a step and thence to the top of the box, by being lifted through these several stages of the ascent, learned to abbreviate the act by climbing directly up the front to getto the hole in the top of the box. (4) No matter how well the animal had learned the through- top reaction, if the front door, out of which he had just come, was not closed behind him he would dodge back through that as the quickest way to re-enter the box. (5) All four raccoons learned to undo a fastening by being put through the act. They did not in general duplicate the act they were put through, but neither did they in general duplicate the act of their first success, or of their first three consecutive successes, when these were attained by their own efforts. Coe, Intelligence of Raccoons. 249 (6) ‘They could be made to duplicate the exact act they were put through by the arrangement of apparatus so that other acts were more difficult. The duplication was then perfect in all trials. (Gapeclite average time required for the first success after being put through is very much less than the average time for the first success by trial and error. ‘This was true in nine out of eleven boxes and with the color showing device. (8) Finally, the animals learned acts by being put through them which they repeatedly had failed to learn when unaided. In all these cases the act was a duplicate of that which they had been put through. ON THE PRESENCE OF MENTAL IMAGES. It would seem that nine-tenths of the experimental evidence for the absence of ideas in dogs and cats comes from their inability to learn from being put through. The experiments were almost identical with some of those described above. If inability thus to learn is evidence against the presence of ideas, then ability to do so should be equally strong evidence for it. We are, therefore, already embarked on the discussion of the presence of ideas in raccoons. It seems to me that animals which, so far as we know at present, are utterly unable to learn save by innervating their own muscles must be devoid of ideas or at least “of a stock of images which are motives for acts.”’ This conclusion of ‘THORN- DIKE’S 1s, I think, of the utmost value to those who experiment with animals, and the evidence against it in the case of cats is meager in the extreme. ‘Therefore, I must first urge the reader to compare point by point the behavior of cats and raccoons in put-through experiments, and to note the radical difference at every point. We may now consider what further evidence of the presence of mental images is furnished by the raccoons and what behavior of theirs seems to show a lack of images. Recognition of Objects—Some of the observations of this are commonplace enough. First, on the fifth day after I received the raccoons one of them climbed to a box, then to the top of a barrel on which the bottle of milk had been placed. When lifted down he at once repeated the performance. A day or so later, another, 250 Fournal of Comparative Neurology and Psychology. No. 3, recognized the bottle at a distance of two feet and went to it. He was given milk, and went to the bottle again. ‘This bot- tle was small and round and was almost completely covered by the hand of the experimenter when he was feeding a raccoon. Were it the whole situation the animal was reacting to, why should he not have come to me, the source of all his food, instead of mak- ing for the bottle as soon as he saw it. The act in No. 3’s case was far too definite to be an accident. I think that he recognized the red rubber nipple. All of the animals now go directly to the bottle if it is set down at all, so it must be hidden. I varied the experiment by lowering the bottle into the room through a win- dow when the raccoons were lying at rest in a remote corner. Within a minute all were clinging to the bottle and struggling to get at the nipple. Next I lowered a small piece of wood, the size of the nipple and wrapped with red cloth to appear like the nipple. All came to it. Two tried to suck it. At first they seemed unable to distinguish it from the nipple. Perhaps this indicates that they do not rely greatly on the sense of smell. When tried thus again they merely played with the piece of wood. When working with Box 12 (six fastenings), No. 4 refused to go into the box. She was switched twice to make her do so. After this, showing her the whip would make her go in. A case of direct searching for the bottle may now be mentioned. On being released from the large cage in which they were confined during experiments, No. 4 went directly toward the corner of the room where she had some days before found the bottle. The total distance was ten feet, but she could not have seen the bottle until she came around a box within two feet of it. All the other raccoo:s were seen to do this. i : No. 3 was reluctant to go into a complicated box and he formed the habit of biting when I attempted to lift him into it. | held him and thrust a finger down his throat, then whipped him. For five days afterward he would growl, snap at and retreat from me though still on good terms with my assistant, Mr. ERwin. Forgetting.—After three days without practice in Box 2, No. 4 seemed almost to have forgotten how to work it. There was no directness in her movements and her time records were poor. A period no longer than three days should show no influence of this sort on a well established reflex, and all records agree, I believe, in indicating that a period of some weeks or months would not Coe, Intelligence of Raccoons. 251 suffice to show any great falling off in the skill of dogs and Cats, In complicated boxes all the raccoons had periods of forgetting one fastening only. Sometimes this fastening was forgotten during two or three days. Often my notes read asfollows. “Aug. 5, Ox Ll, Dolly (ves No: a forgot loop 2 today in three out of four consecutive trials. Jack (7. e., No. 1), forgot button 1 almost invariably except when he pulled loop 1 first. In those cases he turned button r next.” Does this not give us an important dis- tinction between a reflex and an association’? ‘The reflex has but one cue, an association many. Jack did not forget button 1 when he pulled loop 1 first. This had become partly habit because I had built the box up from two fastenings and when it had two he usually pulled the loop first, then turned button 1. Later when I was reviewing Box 13 after periods of one hundred forty-four and one hundred forty-seven days respectively, each of the animals (except No. 3) failed on some particular latch or two latches, not on all, nor on one latch in one trial and another in another. If they had settled down to a routine order of working this box, I venture to say that not one of them would have failed after two hundred days or longer. The recall r) in this case would have been, like that of a boy in swimming for the first time since the prceedine summer, perfect. No. 3’s Bvork gives evidence of this. He alone gave a pair of duplicate performances in Box 12 (six fastenings). Thus, on August 18, in the twelfth trial he worked the fastenings in the following order: (12) 5-2’ -2—-5-6 & 1’ —5—1-57 . (13) 5-2’ -2-5-6 & 1’—5-1-5. The eighth and ninth trials were almost duplicates and there were other partial duplicates. After one hundred forty-seven days without practice No. 3 alone escaped from Box 13, which was Box 12 with one added fastening. [ attribute his success entirely to the superior mechanization of his performance. KINNAMAN says of monkeys, “When the group consisted of two or three fastenings the monkey soon adopted a regular routine which he rarely failed to follow.” I infer from this that the monkey did not do so with more than three fastenings. The raccoons did not with only two fastenings. I have show that No. 2 followed 76 & 1/” means both latches simultaneously. 252 ‘fournal of Comparative Neurology and Psychology. one order predominantly in the morning, and the reverse in the afternoon of the same day, and in general with two fastenings the two orders appeared alternately. Consequently I should say that the monkeys are superior to the raccoons as habit-formers. The raccoons operated almost as complex mechanisms, but they could not reduce the performance to routine. Finally, if the natural history books are to be believed, trappers, in order to ensnare the raccoon depend not on his habits but on his instincts. The trap is not put where he habitually enters the stream, for he enters all along it; instead a bright swinging object 1s hung over the trap so that in reaching for it he steps on the trigger. Finally: a corollary of the proposition that there are two types of learning, namely, learning by trial and error and learning by means of ideas, should be that selene are two types of forgetting, distinguished especially by their time intervals. ‘This, our records seem to show when compared with those for dogs and cats. Variability—In addition to having no fixed order for groups of fastenings, the raccoon changes his method of reacting to a single fastening. I have shown that which paw he uses depends on his position with regard tothe latch to be unfastened (c}. p. 225). As already stated, No. 2, who had learned perfectly to lift the horizontal hook with his mouth finally changed to lifting it with his nose. Finally, raccoons which have done two acts separately hundreds of times may suddenly come to do them simultaneously. No. 3 and No. 4 depressed the platform and pulled loop 1 in Boxes 12 and 13 at the same time. Others occasionally worked the thumb-latch and a button at once, and sometimes two buttons were depressed simultaneously. Association by Similarity—Ilf a latch similar to another be added to a group of fastenings, but in a different place, it may be attacked and worked first. I cannot say certainly that this 1s untrue of a dissimilar fastening for, while it was not the fact with the horizontal hook, the wooden plug or the platform as a plat- form, it would probably have occurred with the thumb-latch had I not first used it singly. The vertical cord leading down to the platform was jumped at directly and vigorously pulled by No. 1 as soon as he saw it, as if he thought it another loop; later he learned to jump upon the platform. He also worked a second barn-door latch before the first one, with which he was familiar. The second wastwo inches above the first. All the animals would Cote, Intelligence of Raccoons. 253 pull a second loop with one direct pull, though it was in a different part of the box, and the same is true of a second button placed on the opposite side of the door from the first. Now, no one can meet the argument that this is not the noticing of similarity but a failure to notice differences and | have said above that Jack attacked the platform cord “as if he thought it another loop.” We may ask, however, ey he did not so attack the horizontal hook or the wooden plug? It will not do to reply that they were in new places. It may answer to say that they were more difficult, but even then they should have been attacked, by one of the rac- coons first, even if unsuccessfully. All this is no doubt inconclu- sive. I may say, however, that the raccoons did not give the same experimental warrant for this dialectic reply that cats do. That is, unlike cats, they did not paw at the place where the loop had been nor did they claw at the loop or button when the door was open. I tried moving the loop from place to place in the boxes. Not once even did they claw where it had been; instead they attacked it at the new place with one direct movement. | removed from the box one loop and then another. Each of the raccoons would come to the place where the loop had hung and look u through the slats in the top of the box. Once I had left the loop lying on the top. It was seen by the raccoon, clawed back into the box, and then pulled. With each of the boxes I tried leaving the door open. The raccoons came directly out with no move- ments in the direction of the fastenings. Reluctance and Expectancy.—All of the raccoons when hungry were eager to re-enter a box of two, three, or four fastenings. They could escape from these quickly. But they were very reluc- tant, even when hungry, to enter a box of five, six or seven fasten- ings. “The small piece of meat they received as a reward seemed to have its effect eclipsed by the memory of the difficulty of escape. I regularly had to put them in Box 13, though they knew the way. Sometimes they resisted strongly by laying hold of the sides of the door and sometimes by snapping at the hand of the experimenter at the moment they were dropped into the box. No raccoon would willingly re-enter a box of from one to four fastenings after his hunger was satished. One may say that in this case the sensations of satiety and weariness did the work, yet no one who saw the animals resist being put into a box failed to credit them with a rather distinct memory of the difficulty of escape. 254 ‘fournal of Comparative Neurology and Psychology. In common with other animals, the raccoon expects food when he has done the thing which usually brings him food. All would come and look up at me on escaping from the box. I tried to test this at some length. An inclined plane of poultry wire was made and No. 3 was fed when he climbed to the top of it. After a few trials the plane was extended twenty-eight inches. When he reached the former terminus he stopped and looked at the experi- menter. The plane was again lengthened with the same result. He always failed to go beyond the old point on the first trial, but on the second he would pause at that point, look about and then go on. Varying Means to the Same End.—While No. 3 was on top of Box 16 a piece of apple was dropped through the top. He started down the back side of the box to get it. The door at the back was closed so he came slowly around the long cage into the front door, through which he had never entered before but through which he had escaped, and got the apple. This was repeated. He was fifteen seconds coming around the cage, but three or four seconds were wasted in trying to reach through the wire to get the apple. When this was repeated again, he was twenty seconds coming around but did not reach through, as the apple was too far from any side of the cage for him to reach it. In this case he certainly would have entered the cage at the back had the door been open. As it was closed, he came around to the front door. Further- more, in the third repetition, he would have tried to reach the apple through the wire had it not been too far away. In the above examples No. 3 may have seen the apple all the time, though this is doubtful. Box 17 was such that food placed in it could not be seen from outside. It was ten inches square and four inches high, with closed sides. A three inch square was sawed out of the top. This piece could be put in place again thus closing the opening. A staple was fixed in the center of the piece sawed out, so that it might be clawed out and away from the open- ing, which it fitted closely. The plan was to throw bits of apple into the box through the opening in its top, allow the raccoon to reach in and get the apple several times, then cover the opening with the piece which had been sawed out. After this had been done with No. 4 she instantly clawed out the block. She seemed to work as if actuated by a thought of apple in the box. It was not done by random clawing, Core, Intelligence of Raccoons. 255 nor could she smell or otherwise perceive the piece of apple in the box. Her work was based entirely on the former experiences of having found apple in the box when open. No. 3 clawed directly at the block but failed to dislodge it; tried once more and succeeded. I now varied the experiment by using Box 18. ‘This was large and had a square sawed out of the top large enough to permit the animal to crawl through the top. No. 1 secured food twenty-five times by going through the top. The box had no bottom and instead of resting directly on the floor it rested on a row of bricks. Removing one of these made an opening under the lower edge of the box through which the raccoon might crawl. ‘The opening in the top was now closed and nailed fast. No. 1 was freed, went to the top of the box and tried to claw out the block. He then walked about the room then tried the block again. He then went to the opening made by removing the brick, stopped a mo- ment, then crawled in. ‘Total time, 100 seconds. ‘Thus No. 1 learned to go in by either the upper or lower opening, but when the one through which he had been fed last was closed, he would hesi- tate a moment, then go to the other. Finally the side opening was closed and over the opening in the top was placed end-wise a cylinder eighteen inches high made of a roll of poultry wire. No. 1 was freed. He walked around the roll of wire which thus fenced in the opening in the box. He then climbed up the outside of this roll and down the inside of it, into the box. The time from his release until he entered the box was thirty seconds. No. 4 went directly into the lower opening at the first trial; time, seven seconds. No. 2 failed to go in through the roll of poultry wire. Thus all but he turned almost directly from the opening, which they knew and found closed, to another opening and entered through that one. They could not see the apple for it was dark inside the box, nor could they smell this particular piece of apple for the room was full of the odor of apple. It seems to me that they must have retained an image of “apple-there.”’ I should not urge this point, however, if I did not think that the following experiments givé substantial evidence of the presence of visual images. In concluding the description of experiments in discriminating cards of different colors and intensities, [ pointed out that success- ful reactions demand that the raccoon compare a color which has 256 Fournal of Comparative Neurology and Psychology. just disappeared with one now present. Either this or else the animal must keep track of the number of times the colors are shown, going up every other time when the colors appear alter- nately, every third and fourth time when they appear by twos , etc. This second explanation violates the law of parsimony, of course, and was eliminated by the fact that I secured series of perfect reactions when the colors were shownat random. When I changed for the first time from alternate showing to twos, or from twos to threes, there resulted confusion and errors quite sufficient, I think, to show that the animals distinguish one movement from two or two from three (a species of counting). As this was of no avail when the colors were shown at random, I also believe that the animal must have retained some sort of image or visual impression of the absent color and reacted to it. The experiment now took a form which shows this more clearly. Spontaneously several of the raccoons had been clawing the “no-food” card down and sometimes they-clawed the“‘food” card up. Finally,all but No. 4 became fairly proficient in this. I quote one of the records made by No. 1 (Jack), after much training. TABLE Trial. Green. Red. Remarks. I * am a ne ce He clawed green up and was fed. 2 * ay PH Rone roots He clawed green up and was fed. 3 * Sar! UEP rag ta tian He clawed green up and was fed. 4 * Tae tl” Soapbsckoels He clawed green up and was fed. 5 * = Le apoeuet.o He clawed green up and was fed. 6 a ty J paooshanae Put red up, looked at it, then put it down, then put green up. if - ait 4 UA Saateosteck Put red up, then put green up. 8 * Se hs ain ceniareds Put red up, then put it down. 9 * ae We eee AS ore Leaving red up, he put green up. 10 * a Ee EN fateace Put red up and left it up, then put green up. II S 55 85-7555 05.55 Ss. Os 53 7s Os ROn dts bab, 0- Av.,' Tnpemoesees Series 10. Moths not mated, decapitated and stimulated. TABLE III. No. 1. No. 2. | No. 3. | No. 4. Issued Sep. 8 Issued Sep. 8. Issued Sep. 8. | Issued Sep. 8. Decap. Sep. 8 a.m. Decap. Sep.8 am. Decap. Sep. 8 a.m. Decap. Sep. 8 a.m. Died Sep. 19. Died Sep. 18. | Died Sep. 15. Died Sep. 16. Stimulated. Ne o: Stimulated. pose Stimulated. peso Stimulated. pies | eggs. eggs. eggs. | eggs. Sep. 8 p.m. 4 | o | Sep. 8 p.m. 4 o | Sep. 8 p.m. 4 6 | Sep. 8 p.m. 4 8 8 p.m. 9 ° 8 p.m. 8 4 8 pm.9| 13 | 8 p.m. 9 3 9) p:m=s9) | 3 8 pm.g | Io 8p.m.g.15 3 | 9 pm.9| 12 I2 p.m.10 | 10 Gupte Zale | | 12 p.m. 3 13 am. 8 | 16 12 p.m.9) 14 g9pm.2| 8 13 a.m. 10 | 8 T2speiero) |S 1G) Joven eh 115) TABLE I1—Continued. No. 5. No. 6. | No. 7. Issued Sep. 8. Issued Sep. 8. | Issued Sep. 8. Decap. Sep. 8 a.m. Decap. Sep. 8 a.m. Decap. Sep. 8 a.m. Died Sep. 19. Died Sep. 13. | Died Sep. 18. No. of No. of| No. of Stimulated. eggs. Stimulated. eggs. Stimulated. eggs. Sep. 8 p.m. 4] 12 | Sep. 8 pm. 4] 10 | Sep. 12 p.m. 25 Qup.m; 14) |," 0 Spm. 5 |, 6. | 13 p-m. 4] 20 9 p.m.9| II 3 p.m. 9 6 13 pm. 9| 15 1d pee, el) snr ugh fopiene, Gye | 14 p.m. 4 4 I2 p.m. 9 9 17 p.m. 9| 18 | McCracken, Egg-laying Apparatus of Silkworm. 275 Series 11. Moth mated, afterward decapitated and stimulated at intervals. No. i Issued Sep. 4 a.m. | Mated Sep. 4 p.m. No. 2. Issued Sep. 4 a.m. Mated Sep. 4 p.m. TABLE IV. | | No. 3. Issued Sep. 4 a.m. | Mated Sep. 4 p.m. | No. 4. Issued Sep. 4 a.m. Mated Sep. 4 p.m. Issued Sep. 8 a.m. Mated Sep. 8 a.m. Decap. Sep. 8 p.m. Issued Sep. 8 a.m. Mated Sep. 8 a.m. Decap. Sep. 8 p.m. 5. Issued Sep. 10 a.m. Mated Sep. 10 a.m. Decap. Sep. 10 p.m. Died (?) Died Sep. 15 Died Sep. 20. | | Stimulated. Hae Stimulated. toy Stimulated. Yeo ge | eggs eggs. | eggs. Sep. 8 p.m. 5.30 6 | Sep. 10 p.m. Wy 2s Sep. 10 p.m. 9.30 24 8 p.m.7.30 17 11 p.m. 8 44 II p.m. 7 12 8 p.m. 8.30 6 Ii p.m.8.10 44 Taps 7-055 8 9 p.m. 9 8 Ir p.m.8.25 | 59 Hig ps7 1ON | SO 12 p.m. 9 15 13 p.m. 7 12 Il p.m.7.30 | 56 (all fertile) 13 p-m.7-15 | 13 Tit jones, | || ee 13) Pas. 7:30"), 155 II p.m.8.15 | 34 | (all fertile but two) | (allfertile) | Decap. Sep. 5 p.m. | Decap. Sep. 5. Decap. Sep.5 pm. _Decap. Sep. 5 p.m. Died Sep. 18. Died Sep. 18. Died Sep. 16. Died Sep. 14. No. | | No. No. | No. Stimulated. of | Stimulated. | of Stimulated. of Stimulated. of eggs. eggs. eggs. epgs. Sep. 6 p.m. 3 4 Sep. 5 p.m. | 10 | Sep. 9 a.m. II 6 Sep. Io p.m. 8.30 | 6 6 p.m. 3.10 |204 9 a.m. atl rams ries || h7 | 10 p.m. 8.35 | 30 6 p.m. 3.45 | 52 IO p.m. 21 ga.m. 11.45 | 63 | 10 p.m. 8.40 | 12 6 p.m. 3.55 | 33 Il pm. 16 9 a.m. 12 33 To p.m. 8.45 | 20 6 p.m. 4.05 | 22 T2e pms ||P 9 p-m. 12.15 | 45 II p.m. 5.30 | 13 6 p.m. 4.15 | 39 13 p.m. fe) 9 p.m. 12.20 | 41 12 p.m. 8.30 | 6 6 p.m. 4.35 | 27 (all fertile) 9 p.m. 12.30 | 15 12 p.m. 8.45 | 12 6 p.m. 4.45 40 . 9 p-m.12.45 | 22 | (all fertile) | 6 pm.s5 | 25 | | 10 p.m 9 | 6 p.m. 5.15 | 39 | II p.m 63 | 6 p.m. 5.30 | 7 15 p.m 13 | Opperman 5-315) | 7 | | | - 6 pom. 5.45 5 | | | 6p.m.7 \>3 | Total (all fertile) 507 Total (allfertile) [349 | TABLE IV—Continued. No. 5 No. 6. No. 7. 276 Fournal of Comparative Neurology and Psychology. Inspection of ‘Tables III and IV shows that the number of eges oviposited 1 in each group; that 1s, after each successive stimu- lation, is also variable, being, however, on the average no less variable on the first day than several days later. The strength of the stimulus, as employed, not being measur- able, the connection between this and the response was not deter- mined but is conceivably a potent, though probably not a controll- ing, factor in determining the character of the response. The same relative variability prevailed from day to day and from group to group upon the same day as to number of eggs in a group in both normal and decapitated (stimulated) moths, due prob- ably to variability in strength of stimulus and other unknown con- ditions. On the whole, unmated moths oviposited fewer eggs at any one time than the mated moth, as in the normal insect. ‘This suggests a connection between activity of the reproductive mech- anism and the presence of sperm in the female receptacle. ‘The average time rate of egg placing in the decapitated moth varies from group to group upon the same day and from day to day as the following examples show. SERIES II. A issued Sept. 22, a. m., mated Sept. 23, p. m. 4, decapitated Sept. 24, p. m. 1.30. Time rate of ovipositing after stimulation. Ist group, Sept. 24, p. m. 5.35-—14, 6, 0, 8, 20, 7; 21, 11, 20, 20, 8, 7, 8,175 05. 85 05°05.375 07. Av. Of l per 15) Sec: 2d proup, Sept. 24, p. m. 5.45—3, 21, 11, , 25, 10, 10, 12, 7, 8, 9, 10, 15, 11, 24, 8. Av. of I per 12 sec. 3d proup, Sept. 24, p. m. 8.15—4, 11, 23, 30, 20; 7, 13, 15, 14, 11. Av.of 1 per 14 Sec. B issued Sept. 4, a. m., mated Sept. 4, p. m., decapitated, Sept. 5, p. m. 1.30. Time rate of ovipositing after stimulation. Ist group, Sept I0, a. m. 8.30—3, 16, 25, 6, 14, 9, 3, 25, 16, 27, 6, 20, g, 11, 8, Io, Oy 205155, 1557752 Moise as Pls Uap Srl 2oe gee eeev OF © per 1O'sec. ad group, Sept. 10, a. m. 8.35—58, 7, 13; 15, 9, 13, 9, 25» 135 79, 15,9: Av. of I per I5 sec. 3d group Sept. 10, a. m. 8.40—18, 8, 18, 7, 20, 33,17, 15, 9, 16, 19, 14,13. Av. of I per 17 sec. 4th group, Sept. 10, a. m. 8.45—20, 7, 29, 17, 10, 13, 14, 14, 14, 14, 15, 6, 27, 10, g. Av. of I per 14 sec. 5th group, Sept. 11, p. m.—20, 6, 12; 27, II, 14, 9, 35, 29, 25, 39. Av. of I per IQ sec. ; 6th group, Sept. 12, a. m.—3, 39, 37, 266, 3, 26, 17, 27, 35, 23,24, 24. Av. of 1 per 24 sec. McCracken, Egg-laying Apparatus of Silkworm. 277 C issued Sept. 4, a. m., mated Sept. 4, p. m. 4, decapitated Sept. 5, p. m., 1.30. Time rate of ovipositing after stimulation. Ist group, Sept 9, 11.357, 5, 11, 14, 6, 7, 6, 45 5545 45 5 9» 5» 6, 16, 5, 6, 6, 6, 5, 555» 95 5» 524542 5s 9, 524595 7> 419,429, 1- Total, 34 eggs, 1 per sec. 2d group, Sept. 9, 11.45—2, 3, 3,45 5» 145 5s 7s 7s 4 4 52 42 95 5» 45 5s 5s 39 39 45 5» Oe Syed Ss 55 155) Op: To Ss Os Onihs Oo Ss Os Haya 5n 55) Sates Hs Puno) Seponse aaa. 5s 4, 4, 10,3, 3, 10,5,6. Total, 62 eggs. Av., 1 per 6\sec. 3d group, Sept. g, 12 m.—2, 3, 3, II, 4, 5, 6, 59 45 45 3555 55 59 5545 45 135 55 45 5> es l4wG. 3254, 7553 4,,5. Lotal, 30 epps. “Ava, 1 perosec, 4th group, Sept. 9, 12.15—3, 2, 6, 6, 5, 6, 7, 8, 45 3545 5s 3943 41 4145 4 5» 18, 5, 7; 72 72 53 7s 53 39 6, 31, 4, 6, 9, 11, 5, 6, 36, 5> 8, 8, 7> 6, 18, 18, 57- Total, 44 eggs. v., I per Q sec. 5th group Sept. 9, 12.20—4, 9, 13, 7s 7» 5» 5» 5» 5» 5» 9 9; 5s 5s 5s 5s 59 5» 5» 9, 10, Bos LO. Ga tige S, O, 125.3, LO, 25, 10, 11,45 255.9, 745.0127, 20. » Detal, 42 epes: saw. I per sec. 8. 6th group, Sept. 9, 12.30—3, 4, 23, 26, II, 10, 11, 14, 14, 4,27, 4, 5,32,6. Total, T5eegs. Av., If per 13 sec. Gch eroup, Sept. Ti, 4 p. mi—6, 15, 22, 15, 454, 1054505, 52109455) 2051 55 oO (intermittent groups not timed), 4, 9,7, 5,6, 5, 4,6, 29, 4, 7, 6, 5, 6, 33, 25 5,9, 11, 6, 6, 4, Q; 55 Q, 3> 7> 6, IQ, 55 55 6, 9, g, 55 8, 4, 8, 3> 5> 6, 5> 13, 5. Total, 61 eggs. Piva, ot, Pet See. IOth| proup, Sept 11, 4.15 p. m.—4, 13; 14, 7, 555 57> lo 445 Sls 10s 4m foe ece Total, 13 eggs. Av., 1 per 26 sec. The percentage of the variability fluctuates, it does not progress from day to day as one might expect if one assumed a gradual decrease of tone in the system from day to day. Such an assump- tion appears to be unwarranted with reference to the reproductive system. ‘The following fact lends additional evidence to the unwarrantability of this assumption with reference to the ovi- positor. After the placing of the full quota of eggs, as in Nos. 1 and 3 in Table IV, stimulation still brings about extension of this organ and normal egg placing movements. In one particular case a moth issued July 10 and produced, after decapitation, and through a series of stimulations, 314 eggs in twelve groups, covering a period of seven days. ‘The last eggs were oviposited July 17. Thereafter until July 22, when the moth died, at each stimulation (stimulation was repeated daily) the ovipositor was extended, remaining so for from twenty to twenty-seven seconds before being withdrawn. ‘This was the rule in stimulated, decapitated moths after the ovariole tubes were known to be empty. In insects that had been mated previous to decapitation, the eggs were all or nearly all fertile, whether these were oviposited immediately or not for ten or twelve days after decapitation. ‘This 278 Fournal of Comparative Neurology and Psychology. was evidenced by the gray appearance assumed a few days after ovipositing—a color characteristic of the monovoltin egg, or egg destined to lie over the winter before issuing. An occasional non- fertile egg occurs. This frequently occurs also in masses of eggs oviposited under normal conditions. That the full quota of eggs were not procured in every case, as the tables show, was due merely to the fact that the moth was not submitted to a sufficient number of stimulations. In each case where stimulations were continuous, an unexpectedly large number of eggs were placed, for example Table IV, Nos. 1 and 3. In the former (No. 1) frequent stimulations occurred throughout Sept- ember 6, two days after the moth had issued and twelve days before death ensued. In the latter (No. 2) frequent stimulations occurred throughout September 9g, five days after the moth had issued, then two stimulations twenty-four hours apart, and not again for ninety-six hours, twenty-four hours before death ensued, like results following in each case. These tables, therefore, with the supplementary data, show again that the reproductive mechanism 1s perfect even for eight or ten days after decapitation, or as long as the moth lives. The sperm also retains vitality for that length of time within the body of the female. In Series 12, as in Series 9, the moth was first decapitated and either immediately afterward or from one to three days afterward, mated. Mating takes place several days after decapitation with the same ease as before decapitation. In every case, eggs followed a few seconds after stimulation. “lable V shows a series of indi- viduals that were protected from external stimulus for several days after decapitation (from one to fourteen) and_ subsequently mated. Inspection of the table shows that while eggs are produced as successfully upon the fifteenth day after issuing as upon the first, those produced after the thirteenth day are apt to be unfertilized. Either the vitality of the sperm has ceased, the mechanism for the passage of sperm into vagina has ceased working, or there may be an imperfect action on the part of the micropyle in the fifteen day old eggs. As to whether this was true exclusively in moths mated after decapitation was not sufficiently tested. “The cement gland secretion also gives out about this time and eggs are there- after apt to be non-adhesive. McCracken, Egg-laying Apparatus of Silkworm. aie Series 12. Moth decapitated, afterward mated and stimulated at various intervals. TABLE V. No. I. No. 2. No. 3. No. 4. * Issued Sep. 4 a.m. Issued Sep. 4 a.m. {ssued Sep.3.a.m. | Issued Sep. 4a.m. Decap. Sep. 4 a.m. Decap. Sep. 4 p.m. Decap. Sep. 4 p.m. | Decap. Sep. 4 p.m. Mated Sep. 5 p.m. Mated Sep. 5 p.m. Mated Sep. 4 p.m. | Mated Sep. 5 p.m. Died Sep. 14. Died Sep. 15 Died Sep. 19. | Died Sep. 20. Stimulated. ‘Eggs. Stimulated. Eggs. Stimulated. Eggs. Stimulated. | Eggs. Sep. 12 p.m. 8.30) 9 | Sep. 12 p.m. 8.45 | 96 Sep. 12 p.m. 9 20 | Sep. 18 a.m. 18 12 p.m. 8.45) 20 17 p.m. 8.50 | 3 17 p.m. 3 (not fertile) 12 p.m. 8.50| 52 I2 p.m. 9 25 18 p.m. 10 (one not fertile) (all fertile) (The last ten not fertile) TABLE V—Continued. No. 5. No. 6. No. 7. Issued Sep. 4 a.m. Issued Sep. 6 a.m. Issued July 6 Decap. Sep. 5 pm. | Decap. Sep. 6 a.m. Decap. July 6 a.m. Mated Sep. 5 p.m. | Mated Sep. 6 p.m. Mated July 9 a.m. Died Sep. 18. | Died Sep. 12. Died July 15. Stimulated. | Eggs Stimulated. Eggs. Stimulated. Eggs. Sep. 17 a.m. 8c | Sep. 7 a.m. 10 | July 10 p.m.3 16 17 p.m. 32 IO a.m. 20 Iopm.4 | 10 (all fertile) | Teed |e IO p.m. 4.30 | 23 | | ny) Eien, || i 10 p.m. 4.45 | 38 | | 21am. | 6 10 Pm. 5-15 | 13 | (The last 6 not IO p.m. 5.30 5 fertile) II a.m.9.15 | 4 ‘ | | Seite pms selon 136 II p.m. 3.45 | 14 | | Ii p.m. 4.18 | 68 | 12 a.m. 11 12 | 12 a.m. 11.20) 14 | 13 am.11.50 0 =o ae A nal SLs: Ser ce | Total (allfertile) | 353 | In No. 7 only was an effort made to induce an ovipositing of the full quota of eggs. A later dissection of this individual (353 eggs having been deposited in twelve groups, covering a period of six 280 ‘fournal of Comparative Neurology and Psychology. days) showed but four eggs left in the body and these were well back in the ovariole. In Series 13 both head and thorax were removed, either before or after mating. Such an abdomen, carefully protected from external stimulus, produced no eggs, nor is there any extension of the alluring glands. The average length of life, however, 1s reduced to five days. With the dethoraxed abdomen, persistent efforts upon the part of the male failed to perfect a mating. ‘The difficulty seemed to be partly at least a mechanical one due to the continuous projec- tion of the ovipositor in egg laying position. ‘The tone of the muscles moving the ovipositor was apparently lost, for no stimu- lation was sufficient to induce its total withdrawal. No test was applied to determine whether or not the tone of the sphincter muscles surrounding the copulatory pore was also lost. In Series 14 the thorax was removed from the moth after mating had taken place. In every case, as shown in Table VI, applica- tion of stimulus was followed by a group of eggs. The size of the group was as variable as in the normal or decapitated moths, but no moreso. ‘The length of life averaged five days, as in the dethor- axed, non-mated moth. The ovipositor moved from side to side now with the same pre- cision as in the normal or decapitated insect, but was rarely entirely withdrawn beneath the anal plates (frequently partially with- drawn). In the absence of the legs, the body was dragged from side to side simply by the action of the muscles of the abdomen. No attempt was made to obtain the full quota of eggs under this condition, but there was no reason to believe that it might not have been obtained by persistent stimulation. The factor seeming to control the functioning of the ovipositor was determined with moths in the dethoraxed condition. While the decapitated insect experiences no difficulty in righting itself when placed upon its back, the dethoraxed moth, when so placed, makes absolutely no effort to right itself. “This reflex has appar- ently disappeared with the thorax. In a stimulated abdomen lying in this position (that is, upon its back), the ovipositor 1s extended for its full length. ‘The sides of the abdomen are slightly drawn in as when an insect is making a forced effort to oviposit. (This movement has been observed in both normal and dethoraxed moths.) The egg is ready to be expelled, but the ovipositor fails McCracken, Egg-laying Apparatus of Silkworm. 281 Sertes 14. Moth mated, afterward dethoraxed and stimulated at various intervals. TABLE VI. No. ¢. | No. 2. | No. 3. No. 4. Issued Sep. 12. Issued Sep. 23. Issued Sep. 26. Issued Sep. 27. Mated not. Mated Sep. 23. Mated Sep. 26. Mated Sep. 27. Dethor. Sep. 12. Dethor. Sep. 24. Dethor. Sep. 28. Dethor. Sep. 28. Died Sep. 15. | Died Sep. 29. Died Sep. 30. Died Oct 2. Stimulated. ‘Eggs: Stimulated. Eggs. Stimulated. Eggs. Stimulated. | Eggs. Sep.13am.9 4 | Sep.25a.m. | 42 Sep. 29 p.m. 2 | 24 |Sep. 28 p.m. 2 5 I14a.m.9 | 10 26am. | 20 | 29p.m.2.15 | 44 28 p.m. 2.15 | 22 14 p.m. 25 (all fertile) 29 p.m. 2.30 | 34 28 p.m.2.30 | 8 (all fertile) — | | 29 p.m. 3 | 24 28 p.m. 2.45 | 16 | (all fertile) (all fertile) \ TABLE V1I—Continued. No. 5. No. 6. No. 7. Issued Sep. 28. Issued Sep. 23. Issued Sep. 23. Mated Sep. 28. Mated Sep. 28. Mated Sep. 23. Dethor. Sep. 29. Dethor. Sep. 29. Dethor. Sep. 24. Died Oct. 4. Died Oct. 3. Died Sep. 29. Stimulated. Eggs Stimulated. | Eggs. Stimulated. —_ Eggs. Sep. 29 p.m. 94 Sep. 29 p.m. 262 Sep. 24 pm.2 aur 30 p.m. 20 30 p-m. 20 2A 2L-Or 22 Oct ripe 30 @ct.” X jp:m: 25 24) p-m..2-05 |) 79 (all fertile) (all fertile) 24pm.8 | 39 24 p.m.8.30 20 (all fertile) to open, and no egg appears. If, now, the sensory hairs that cover the outer surface of the ovipositor be barely touched with a pencil, or if a fiber of cotton held by a pair of pincers is brought into con- tact with these hairs, the lips of the ovipositor immediately open and an egg is pushed out. ess: Each such contact brings forth an In an insect in normal position, the sensory surface of the 282 Ffournal of Comparative Neurology and Psychology. ovipositor is brought into contact with the surface upon which the insect rests. [his serves as a stimulus for the action of the lips. Unfortunately, no experiments were tried to determine whether the nature of the material affected the functioning of these lips. It would be interesting to know whether the lips would open if these sensory hairs came in contact, for instance, with another egg. If not, we could understand how it was that one egg is rarely placed, either under normal or experimental condi- tion, upon another. A study of the action of the ovipositor in this regard might also throw light upon the “selections” of various insects of certain plants only for placing their eggs—these plants being the particular food to which the larve of their kind isadapted. Series 15 was a further reduction of the nervous system. The first abdominal ganglion lies in the second abdominal seg- ment. The commissure connecting it with the second abdominal ganglion lies close to the ventral surface of the body. ‘This com- missure can be snipped in two through the ventral wall of the body with but little loss of blood. With the first abdominal gan- glion severed from the rest of the nervous system, egg placing fol- lowed, as before, upon stimulation, the rate of egg placing not being materially lowered. By snipping through the suture sepa- rating the third and fourth abdominal segments, the second abdominal ganglion was severed from the posterior part of the nervous system. Again stimulation resulted in egg placing. Snipping through the suture separating fourth and fifth abdomi- nal segments severed the third abdominal ganglion from the fourth and last abdominal ganglion. As before, egg-placing took place with no more hesitancy than in the intact nervous system. ‘That severance of the commissures in insects thus operated upon had actually taken place was later verified in several cases by dis- sections. Upon severance of the nerve connections between the last abdom- inal ganglion and the ovipositing apparatus, stimulation failed to bring about movements in the ovipositor. That this was not due to loss of blood or disturbance of the respiratory apparatus incident upon the incisions, is evidenced by the fact that the ovipositing apparatus (egg tubes, ovipositor, etc.) with its controlling ganglion may be completely severed from the body, and eggs will yet pass down through the egg tube and out between the lips of the ovipos- itor, if the parts are kept normally moist. McCracken, Egg-laying Apparatus of Silkworm. 283 The following examples give the time rate of egg placing under the several conditions enumerated above of a single representative moth. Moth issued Sept. 21, decapitated Sept. 23, 11.36 a. m. Stimulated immedi- ately. dime rate et ovipesiting—o, 16, 26, 13) 12,10, Bs 7, LO, 12, 1anl5.22,0,9-5 15 eggs. Av. of I per 12 sec. Dethoraxed Sept. 23, 4.25 p. m. and stimulated immediately. Time rate of ovipositing—7, 17,17, 36,15,9,9,21. Seggs. Av. of 1 per 16sec. First abdominal ganglion severed 4.30 p. m. and stimulated immediately. Time rate or ovipositing—7, 6, 8, 8, 9,19, 17, 7, 22, 9, 17, 21, 22, 24, 20, 21, 24, 20. 18 eggs. Av. of I per 15 sec. Second abdominal ganglion severed, 4.35 p. m. and stimulated immediately. ‘Himerate of ovipositine—8, 8, 11, 8;14, 12, 11, 8515,27, 23.95 275 BOs tO. 25 eggs. Av. of I per 13 sec. Third abdominal ganglion severed, 4.40 p. m. and stimulated immediately. Time rate of ovipositing—9o, 5, 7, 16, 34, 19, 6, 31, 15, 13,25,7,24. 13 eggs. Av. of I per 16 sec. The coérdination of movement in various segments, after each operation, is progressively lost. The ovipositor, throughout, makes the same effort to.avoid placing one egg upon another, by moving from side to side. With abdominal ganglia intact, this effort is successful. It becomes, noticeably less and less so as each abdominal ganglion 1 is severed, until when but two remain, mus- cular cooperation of the segments is so far reduced that the com- bined efforts of the last three abdominal segments is not sufficient to pull the body around and the eggs (after the first three or four, which are placed side by side) are piled one upon another. The results from the last series of operations show the high degree of independent activity exhibited by the controlling center of the reproductive apparatus, namely the last abdominal gan- glion, and the coérdination of the functions of this ganglion with those of the preceding ganglia in the ventral chain. Hence the general opinion that “each segment of a segmental animal may be regarded as a simple reflex animal” is only a part of the truth. Betue found that certain reflexes are located in each thoracic ganglion for the corresponding segment, as already cited. Each thoracic and abdominal ganglion in the silkworm is so organized that the reflexes carried on in the last segment of the abdomen are accompanied by a set of reflexes in other segments 284 ‘fournal of Comparative Neurology and Psychology. of a kind altogether different from that of the initial reflex. The functioning of the ovipositor is perfected by the movements in space of the legs, or in the absence of these, of the anterior seg- ments of the abdomen. Direct Stimulation of the Commissure.—In direct stimulation of the ventral commissure with an electric current, which is quite possible in silkworms, the response is not so determinate. Then only an occasional egg follows upon stimulation. ‘The difference in response to direct stimulation and integumental stimulation may be due to the fact that in integumental stimulation only those neurones are affected that are in direct connection with the ovi- positing machinery promoting movement, whereas in direct stimu- lation through the commissure, all the neurones are similarly stimulated, the inhibitory as well as those promoting movement. SUMMARY 1. The vitality of the silkworm moth, as measured by length of life and capacity of the reproductive system to function, is not impaired by removal of the head (supra- and sub-cesophageal ganglia). Length of life is shortened, but functioning of the reproductive system Is not impaired by removal of thorax (thoracic ganglia). 2. [he mating instinct is dominant in unmated moths and in some way correlated with an inhibition of the ovipositing mechan- ism. ‘This inhibition is reduced in the mated moth, functioning of the ovipositing mechanism then dominating. ‘The inhibition is reduced in decapitated moths, as evidenced by the early response to stimulation. 3. Effectual mating takes place several days after decapitation of the female. It is impossible with the dethoraxed female. 4. Inthe presence of the brain or brain and thorax, reaction of the ovipositor to external stimulus is resisted. In the absence of the brain, or brain and thorax, spontaneous movements cease, but reaction of the reproductive mechanism to external stimulus 1s prompt and efficient. 5. The effectual stimulus is pressure and contact, this being a stimulus to which the normal silkworm moth is constantly sub- jected under normal conditions. 6. Thetime duration of the reflexes centered in the last abdom- McCracken, Egeg-laying Apparatus of Silkworm. 285 inal ganglion 1 is much more limited in the decapitated or dethor- axed insect than in the normal insect. Repetition of stimulation causes repeated responses with no evidence of fatigue or decrease of functional activity until death ensues. ‘There 1s neither awk- wardness nor feebleness exhibited, neither is there evidence of aug- mentation. 7. Response of the ovipositor is initial and endures after all eggs have been oviposited, therefore entirely independent of the rest of the reproductive apparatus. 8. The posterior abdominal ganglion is the controlling center of the reproductive mechanism and exhibits a high degree of independent activity. g. ‘The reflexes of movement connected with the three anterior abdominal and the thoracic ganglia are coordinate with the reflexes of the ovipositing apparatus connected with the posterior abdom- inal ganglion. In conclusion, it is apparent that while an intact nervous system is necessary for a prompt, efficient and continuous functioning of the reproductive system, in the absence of the head or head and thoracic ganglia, external stimulus brings about prompt, efficient and complete though not continuous reaction. In the absence of head or head and thorax, there is no apparent augmentation of the reflexes. This reaction is efficient if the posterior abdominal ganglion alone is intact. Corresponding ganglia in the anterior abdominal segments are not only reflex centers but centers of cor- relation as well. A STUDY: OF THE CHOROID PLEXUS: BY WALTER J. MEEK. (From the Neurological Laboratory of the University of Chicago.) With Nine Ficures. With the exception of scant references in the standard anat- omies, very little literature treating of the choroid plexuses is available to the general student. ‘The literature dealing with the subject directly is not as full as one might wish, and in fact many questions of interest concerning these structures still remain unanswered. The object of the present paper is to review the subject briefly to date, and to present some results of the author’s own investigations. The writer’s attention was called to the choroid plexuses by Dr. SuinkisH1 Hatatr at the University of Chicago. Dr. Hatat had noted that the position of the nucleus in the cells of the foetal plexus was apical, while in the adult it was central. The problem as first undertaken was to determine the time of , this shifting of the nucleus, but as the work proceeded, it seemed best to prepare a brief study of the entire subject. TECHNIQUE AND MATERIAL. Fixation of tissue is at best an unsatisfactory process, and doubly so when structures as delicate as the choroid plexuses are concerned. Slight differences of osmotic pressure cause shrinkage or swelling of the cell. Mechanical injury must also be avoided. The number of fixing fluids adapted to the plexuses 1s somewhat limited. “The following were used, and the order indicates approx- limately their relative merit. Bouin’s fluid. Carnoy’s solution. Acetic sublimate. Potassium bichromate-corrosive sublimate solution. Kopsch’s fluid. nAPw iS) Lal For small animals the best results were obtained by fixing the entire brain. In the case of larger forms, the lateral ventricles MEEK, Choroid Plexus. 287 were opened to allow an easy access of the fixative. Often the plexus was left untouched on the floor of the ventricle while the overlying nervous matter was trimmed away. ‘These devices all avoid mechanical injury, and they also preserve the position of the plexus in the ventricle. Only in the largest brains were the plex- uses removed and fixed separately. The period of immersion in the fixative is important. When an entire brain is fixed the average time required by the fixative may be allowed, but if the plexuses are exposed the length of time should be reduced to about one-third. In the case of Carnoy’s solution, about 20 minutes gave by far the best results. The usual methods of embedding in parafin and sectioning were employed. Sections were cut 3 and 4 micrathick. Many different stains were used in the hope of differentiating various structures. The most satisfactory results were obtained with iron haematoxylin followed by acid fuchsin. Other stains, copper-chrome hama- toxylin,! toluidin blue, erythrosin, VAN GiEsoNn’s acid fuchsin, and EuRLIcuH’s triacid stain were used as controls, and for the sake of comparison. On account of their accessibility and their size, the plexuses of the lateral ventricles were used exclusively. So far as known, the plexuses of the other ventricles are precisely the same in lunes and structure. Mammalian material was studied from the fol- lowing forms: albino rat, rabbit, guinea pig, cat, dog, sheep and man. GENERAL MORPHOLOGY. The choroid plexuses of the lateral ventricles are due to an ingrowth of the pia mater pushing the mesial wall of the hemi- spheres into the ventricles. ‘The arachnoid is not supposed to be present, although the plexus is but a fringe of the velum interposi- tum, into the structure of which the arachnoid does enter. “The neural wall is of course preserved, but consists only of a simple epithelium. ‘The plexuses then are thin laminz covered with an epithelium, beneath which is a connective tissue stroma containing an extraordinarily rich network of blood vessels. In many animals, the laminz are smooth, but in others, they are covered with projecting villi. Between these two extremes 1 The formula for this stain was kindly given by Dr. Benstry of the Department of Anatomy, Univer- sity of Chicago. 288 ‘Journal of Comparative Neurology and Psychology. are to be found all the intermediate gradations. ‘The guinea pig, mouse and rat possess plexuses that are smooth. Fig. 1 shows the cross section of a lateral plexus of a one day old rat. It will be noted that the surface is not entirely regular. There are pro- jections and prolongations of the folds, but the typical villi are absent. In the rabbit, the laminz are still more irregularly folded, but they are not villous. Villi are scarce in the chicken, duck and pigeon, but more abundant in the hog, while they reach a con- siderable development in the horse, ox, and especially among por- poises, crocodiles, and some of the selachians (PETTIT ’02~03). In the sheep, villi are numerous along the free edge of the plexus, but they are thick and short. In man, the villosities are also found but the type is somewhat intermediate. IMMAMURA (02) states that the human plexus con- sists of two parts, a villous and a villus- free portion. ‘There seems, however, to be much variation in this respect. The surface of the plexuses is much greater than one would suppose. FAIVRE (54) has estimated it for the human plexuses in the lateral ventricles. He considers the average length as 7 centi- meters, and the width as 2 centimeters. By estimating 40 villi to each square cen- Fic. 1. Crosssection throughthe tjmeter, he calculates that the surface right ventricle of a one day old rat. Would be increased about four times, Magnification 150. a, Plexus in 5 : ; making a total of 112 square centimeters for the two plexuses. We might add that since the folds hang freely in the ventri- cles, and are covered on all sides by the same epithelium and villi, that twice this number, or 224 square centimeters, is more nearly the total area of the free surface. The blood supply of the plexuses in man is well known. ‘Two- thirds is supplied by the anterior choroid branch of the internal carotid which enters the plexus at the anterior end of the descend- ing cornu. ‘The remainder is supplied by the postero-lateral cho- roid artery, a branch of the posterior cerebral. These arteries | break up into arterioles, the largest of which are visible to the naked eye. After passing through the network of capillaries, the ventricle; 5, ependyma; c, nervous tissue. MEEK, Choroid Plexus. 289 blood returns through the choroid vein to the V. cerebri interna which joins its fellow and forms the V. cerebri magna or vein of GALEN. It is generally believed that the choroid plexuses are largest in the embryonic state, and that their volume diminishes as the brain reaches its full development. LorpeR (’04) states that in the human feetus the plexuses are largest from the third to the sixth month, after which they decrease in size. Possibly a more accu- rate statement would be that after the sixth month they are out- grown by the other parts, and the decrease in size is only propor- tional. The His-Z1EGLER model for the three months human Fic. 2. Cross section through the brain of an embryo rat, showing the origin of the choroid plexus. Magnification 500. foetus, shows the plexuses as large swollen glandular organs oc- cupying practically all of the space in the lateral ventricles. Along with this presentation, has grown the idea that the plexuses furnish some kind of a fluid food necessary for the growth of the brain during the foetal period. Accordingly, they have sometimes been called the “cerebral placenta.”’ Whatever the facts may be in man,? it seems that the above description does not apply to all forms. Fig. 2 shows the begin- ? A recent reconstruction of the encephalon of a human embryo (length 22.8 millimeters, age approxi- mately 2 months) by Dr. Ewinc Tay or, of the Universtiy of Pennsylvania, shows the plexuses of the lateral ventricles as much smaller and more folded than they appear in the His-ZircLer model of three months. A similar difference is shown by the model of the encephalon reconstructed from a slightly larger embryo (length 30 mm., approximate age 55 days) by Dr. G. L. Srrerrer of the Wistar Institute of Anatomy. In view of these results, the relations exhibited by the His-ZiecLer model can hardly be considered as typical. 290 ‘fournal of Comparative Neurology and Psychology. ning of the plexus in the albino rat. ‘The exact age of this embryo unfortunately is not known, but the neural walls are thick and the ventricles large. In a two centimeter embryo rat, which has about the same brain development as a three months human foetus, the plexus has elongated somewhat more than in Fig. 2, but it is not essentially different. From this time to birth, there seems to be a steady growth of the plexus, until the structure assumes the appearance of Fig. 1. At no time is there any evi- dence that the organ is enlarged or distended, or that it fills any- thing like all of the ventricular space. MICROSCOPIC HISTOLOGY. I. Examination of Living Tissue.—From a freshly killed adult albino rat, the plexuses may be removed and examined in cerebro- spinal fluid, or in normal salt solution. Often under these con- ditions the blood will continue to flow through the capillaries for several minutes. ‘This circulation is due to a gradual shrinkage of the tissues and the weight of the cover slip, both of which aid in forcing the blood out of the ruptured capillaries. In case the arterioles are not clogged, this circulation may be in a reverse direction, since resistance would be less toward the open arterioles. Not a great amount of detail can be obtained in this way, still it serves as a control for the sections subjected to reagents. Under low power, the plexuses of the rat appear as a thin, semi- transparent membrane, which is threaded by an immense number of capillaries. “Uhe edges are somewhat sinuous, and these curves are closely followed by the blood vessels, so that the appearance is not unlike the scalloped edge possessed by many leaves. Under ahigh power, it is readily seen that the lamina consists of an upper and a lower layer of epithelial cells, with the blood vessels embed- ded between. Details are plainest at the edges. ‘The cells are cuboidal, and can best be located by their nuclei, which appear as shadowy discs. The cell boundaries can be made out only with much difficulty. Occasionally on a flat surface, the ordinary mosaic appearance of the cell walls may be brought into focus. The apical plate or marginal zone of the cell is visible as a thin refractive line. ‘The apical end of the cell is curved with the con- vexity outward. At times, the lateral cell walls may be detected, but not usually. ‘The cytoplasm appears finely granular through- MEEK, Choroid Plexus. 291 out. Very small refractive globules may be seen within tt. No pigment has been noted in the rat’s plexuses. The most careful examination has failed to show the presence of cilia 1m the adult. These are present in younger forms but it seems reasonably certain that cilia are not present in the adult. At the same time, cilia are often noticed in the adult on the ependyma, which has been torn from the ventricular walls. “The ependyma, in certain por- tions, is ciliated at all ages. The microscopic appearance of the plexuses of other forms is similar to that of the albino rat. In most forms, particularly the guinea pig, dog and rabbit, the marginal or apical zone of the epithelial cells seems more specialized. The rabbit’s plexus 1s remarkable in being studded with clear drop-like spaces within the cells. ‘These will be discussed more at length in the next section. II. Examination of Stained Ma- terial. Under the influence of fix- atives and stains, many other details of the cells are brought out. At best, there must be some distortion, but judging from the appearance of the fresh tissues it is believed that the most successful prepara- tions area fair representation of the structures in the fresh condition. adult rat’s plexus. Magnification X 1500. Fig. 3 shows the cross section a, Capillary; b, connective tissue. of a small loop of an adult rat’s plexus. The two capillaries represented, show about the two extremes in size, the smaller being about 12 micra in diameter. The capillaries consist of a delicate endothelial intima, with elon- gated nuclei. ‘This intima is strengthened by connective tissue cells and fibrils. FinpLay (99) considers this as an adventitial coat. Between this adventitial and the epithelium are more connective tissue cells and their processes. In young animals, lymphocytes are occasionally noticed. The epithelium covering the plexus of the rat is composed of cubical cells, which in cross section average about 10-12 micra in width, and 8-10 micra in height. The basal wall is rather poorly defined. Often it is difficult to seperate the cell wall from Fic. 3. Cross section of a portion of an 292 ‘fournal of Comparative Neurology and Psychology. the connective tissue beneath. ‘The lateral walls of the cells are also obscure, but they may usually be identified by a hazy line. Under a low power, the cytoplasm appears finely granular and homo- geneous, but with higher magnification, it is easily seen that it is finely reticular. In the meshes are deeply staining granules, which are often collected into small irregular masses. For the most part the reticulation is more pronounced near the periphery. Neither vacuoles, clear spaces of any kind, nor pigment granules have been found in the plexuses of the rat. ‘The nuclei of the epithelial cells are oval or circular in outline, and centrally located in the adult. They average about 6 micra in diameter. Nucleoli are frequently present. ‘The free edge of the cell is slightly con- vex. It consists of a thin apical plate or cuticle, which is not very apparent. Embryologically the epithelial cells of the plexuses are derived from the inner layer of the neural tube which also produces the ependymal cells, and which is called the ependymal zone. The ependymal cells and the epithelial cells of the plexuses are there- fore parts of the same layer. ‘This may be plainly seen by tracing the plexus back to where its epithelial covering joins the ependyma lining the ventricle. Here the two types of cells pass into each other by an easy gradation. Fig. 2 illustrates this for the embryo, and the condition is not essentially different in the adult. It is generally conceded that the epithelium of the plexus has lost all vestiges of the neuroglia, which normally underlies the ependyma. CaTOLa (02), however, reports that he has found glia fibers by the WEIGERT method. So far as known, his observations have not been confirmed. ‘The epithelial cells have also lost all pro- jections from the base, a characteristic of so many ependymal cells. There has been some discussion as to whether or not the epithe- lium of the plexuses consists of a single layer. Harcket (60), Hexpr (74), DEJERINE (’95) and KOLLikER (’96) describe only asingle layer. LuscHKka (’55), however, recognizes several layers with transitional forms. HarcKeL admits that in pathological cases the cells may greatly increase in number. Frnpray (’99) reports that there may be from one to four, or even more layers, and that the cells change in character as they pass from the basal layers toward the surface. He finds this condition in the sheep, calf, ox and man, and believes it occurs too frequently to be explained as pathological. MEEK, Choroid Plexus. 293 The writer’s own results are decidedly in favor of there being but a single layer of the epithelial cells. In about too plexuses examined, there has never been the slightest suggestion of strati- fication. This refers, of course, to normal tissue. Pathological proliferation would be both likely and possible. No such cases, however, were noted. It may be that observers have been misled by oblique sections of some of the villi. It is not possible to have all of the tissue when it is covered with villosities, at right angles to the plane of section. In such a case, an oblique section might lead one to believe that the tissue had several layers of cells. Intercellular spaces have been found in ependymal tissue by OpersTEINER (or), Renaur (99) and Stupnicka (00). Re- NAUT’S idea is that the spaces are filled with a delicate cement sub- Fic. 4. Cross section of a portion of an adult rabbit’s plexus. Magnification % 1500. 4, Micro- somes simulating basal ‘bodies; b, capillary with blood corpuscles; c, clear space (fat droplet); d, marginal zone. stance, while SrupnicKa holds that they are true lymph spaces. Srupnicka has found similar spaces between the epithelial cells of the choroid plexuses in the case of the shark, Notidanus cin- ereus. Mrtan (04) and Perrir and Grrarp (’02~’03), on the contrary, believe that normally the cells are closely appressed. They have noticed that such spaces increase in number as the result of poor fixation and post-mortem changes. They therefore regard them as due to changes in the cytoplasm, occasioned by faulty fixation, or post-mortem alterations. We have observed intercellular spaces in but a single specimen, that of a sheep’s plexus, which had been fixed with the brain in formalin, the ven- tricles not having been opened. In freshly fixed tissue, there was not even a suggestion of spaces between the cells. There is no reason for doubting that the observations on the ependymal cells 294 ‘fournal of Comparative Neurology and Psychology. are correct, but the intercellular spaces do not occur between the epithelial cells of the plexuses, at least in the forms here studied. Gold preparations show that nerves are present in the plexuses in the vicinity of the large blood vessels. BENEDIKT (’73 and ’74) described nerves in the plexuses of the fourth ventricle, and thought he traced them into the vagus. Finpay (’99) found nerve fibers in the plexuses of man and the calf, by using a modi- fication of HELLER’s method. ‘These fibers are probably vaso- motor fibers to the blood vessels. Fig. 4 shows the cross section of a small irregularity in the plexus of the adult rabbit. In regard to capillaries, connec- tive tissue, and endothelium, it exhibits no particular difference from the plexus of the rat. The cyto- plasm is somewhat more plainly reticular. Nucleoli are more evident, and often there is more than one present. ‘The rabbit’s plexus is especially character- ized by the presence of circular or oval clear spaces. [hey are from 2-6 micra in diameter, and really much more numer- ous and noticeable than Fig. 4 could in- dicate. [he contents are dissolved out by alcohol and xylol. In the ordinary stained sections they show as clear round Fic. 5. Section through an adult or oval areas. [he contents are of a Fabbie's plexus alter staining ms fatty natutey for whey veadily staimmeniee micacid. The protoplasm is faintly : : : : granular, and the droplets are stained OSE acid we Sudan Ht. Fig. 5 ise Gece. Measures cross section of a rabbit’s plexus treated with osmic acid. It gives a fair idea of the number of the droplets and their position in the epithelial cells. The basal droplets are the smallest. “Toward the apex of the cell, they gradually increase in size. Sometimes one may be seen lying half within, half without the cell. This shows that they are expelled through the top wall. During this process, the nucleus remains entirely unchanged, but rarely it 1s pressed to one side by the droplet. There is no evidence, how- ever, that the discharge of the droplet is attended by the death of the cell. ‘These fatty droplets have not been found in the nor- mal plexuses of any other form studied. Lorper (’04) speaks of small globules in the plexus of the guinea pig that stain with MEEK, Chorord Plexus. 295 osmic acid, but he does not intimate that they are anything like the large droplets found in the rabbit. As before noted, these clear spaces can be seen in the fresh tissues even before the blood has ceased circulating. For this reason, it is not believed that they can be due to any error in technique or to post-mortem processes. Vacuoles have been mentioned in a general way as occurring in the typical cells of the plexuses, but to our knowledge, nothing similar to these clear areas has been described. It is not believed that they represent the chief secretion of the cells, since they have not been found in the other forms examined, and it is therefore best to consider them as of secondary importance. A second feature in the rabbit’s plexus is the development of the modified structures at the apex of the epithelial cells. In the rat, the marginal zone is at best but a double contoured line. In the rabbit, however, it is wider, and composed of filaments placed perpendicularly to the surface of the cell, and embedded in some kind of a matrix. This gives the cells the ap- pearance of ciliation, but this cannot be confirmed by ciliary movements in the fresh tissue. The structure is ~ "6 6 Three cells from a dog’s what VIGNON (or) describes as the plexus. The animal was killed with “‘bordure de brosse”’ or filamentous plateau. At the base of the filaments are cytoplasmic microsomes, which take up the stain and simu- late basal bodies. “Terminal bars may be seen in cross section at the corner of the cells. Fig. 6 shows the epithelium of a dog’s plexus. The stain is rather diffuse, and the reticulations show poorly, but it differs from the rabbit’s in no particular way, except by the absence of the clear spaces or droplets. The epithelial cells of the guinea pig’s plexus are somewhat peculiar in having a great many nucle- oli. There are usually two or three, and often fouror five. In other respects, the plexus is similar to that of the dog, as shown in Fig. 6. The preceding description has referred entirely to the adult plexuses of the rat, sheep, rabbitanddog. If alate foetal or a new- born specimen be examined, striking differences will be noticed. Fig. 7 is a section from a one day old rat. The epithelial cells differ from those of the adult in three particulars: Shape, stair.- illuminating gas, and the cells are in a resting state. Magnification X 1800. 296 Journal of Comparative Neurology and Psychology. ing power and location of the nucleus. A comparison with Fig. 3 from the mature rat, shows that in the one day form the cells are narrower and deeper. The long diameter is perpendicular to the base of the cell, while in the adult it is parallel with it. Many measurements give the following averages. Cells from the one day old plexus are 13 micra in height, and 8 micra in width. Cells from the adult are 9 micra in height, and 11.6 micra in width. This would indicate that the young cells were a trifle the smaller. An interesting question here is: How does growth take place? The extent of the plexus is greater at maturity than it is at birth, and it is therefore necessary to account for the extension of the epithelial layer to cover the larger area of the adult plexus. A partial answer is contained in the fact shown above, that the adult epithelial cells are about one-third wider than those of the foetal type and so cover about twice the area. This is possibly sufhcient to account for growth from the foetus to the adult. Growth might, how- ever, take place m another way by a egular mitotic division of the epithelial cells. So far as known this has not been observed in epithelial cells of the plexus Fic. 7. Loop of a plexusfrom 1n late stages of growth, but it has been aone day old rat. The cells are reported as occurring in the ependyma. high and narrow, and the nucleus The location of the nucleus in the one Te paren ate fn: day form is apical, while in the adult it Hone acer. if is central or basal. In the former the nucleus is removed from the base of the cell by three-fourths the vertical diameter of the cell; in the latter, it is removed by one-half the vertical diameter. The contents of the cell in the one day plexus are well nigh unstainable, at least by ordinary methods. ‘The basal part of the cell shows a few radiating lines of microsomes, but it is for the most part clear. ‘The nucleus stains deeply and so diffusely, that in only a few cells is its structure visible. Under favorable con- ditions of staining, reticulations are visible. Between the nucleus and the apex of the cell, the cytoplasm stains somewhat diffusely, but a reticulation with granules in the meshes can be made out. Meek, Chorotd Plexus. 297 The free margin of the cell shows double contoured lines. Asthe rat grows older, the cells gradually assume the adult type. Plex- uses four days old show the nuclei removed by two-thirds the diameter of the cell from the base, and by the seventh day the position of the nucleus is similar to that in the adult. A change in the staining reactions of the cells occurs at the same time. The primitive columnar condition of the epithelial cells is therefore retained until extra-uterine life is well begun. Whether this means that the cells do not function until this time cannot be said. It would certainly indicate that their work could not be the same as in the adult. This rapid change in the epithelium after birth seems to occur in the plexuses of other forms than the white rat. We have observed it in the rabbit and cat, and judging from STUDNICKA (00, Plates XXXII, XXXIIT), it occurs in man also. SECRETORY PHENOMENA. Thus far we have studied the plexus in what might be termed its resting stage. Careful examination shows that in any adult plexus there are always some cells that have the appearance of secretory activity. It is to this phenomenon and its relation to the production of the cerebro-spinal fluid that we would now direct attention. For many years there has been a suspicion that the choroid plexuses secreted, or at least aided in the secretion of the cerebro- spinal fluid, but the idea, until recently, remained without much supporting evidence. Even as late as 1901, CHARPY (01) stated that the origin of the cerebro-spinal fluid was unknown. Although the possibility had been mentioned the century before, it was E. Fatvre (54) who first definitely stated that the plexuses were probably concerned in the production of the fluid. LuscHKa (55) published an important monograph in which he produced evidence to show that the cerebro-spinal fluid could not be a transu- date, and that it must be considered as a secretion produced by the membranes of the brain. For many years no further evidence appeared. In 1897 PauL CLaissE and Cuaries Levi (’97) reported a case of internal hydrocephalus, in which there was hypertrophy of the plexuses. There was a great amount of granulation, and the veins were dilated and gorged with blood. Kincspury (’97), while working 298 ‘fournal of Comparative Neurology and Psychology. on ganoid fishes, came to the conclusion that the columnar cells of the membraneous portions of the brain must be of use in the elaboration of the cceliotymph. He was led to this opinion by the many evidences of secretion shown by the cells. FINDLAY (99), GALEOTTI (’97) and STUDNICKA (’00) all believed the plex- uses secretory in function, and based their conclusions on the pres- ence of secretory globules which will be discussed later. Cavaz- ZANI ('93) advanced negative evidence, by showing that lympho- gogues did not affect the flow of the cerebro-spinal fluid. He therefore concluded that it was a secretory product of the epithelial cells. To secure cerebro-spinal fluid, Cavazzani (99) made use of a cerebro-spinal fistula between the first cervical vertebra and the occipital bone. CapPpELLeTI (01) used this fistula to study the effect of certain drugs on the flow of the fluid. He found that ethyl-ether and pilocarpin increase the flow, while atropin and hyoscyamin retard the flow, sometimes even to the point of sus- pending it. Of course these substances all have a vasomotor action, but pilocarpin and atropin are supposed to affect secre- tory cells directly. Perrrr and Grrarp ('02~'03) went a step far- ther by administering drugs, such as pilocarpin, muscarin and ether to animals, and then removing the plexuses to find whether there was any evidence of secretion. ‘Their results have amounted almost to a demonstration of the origin of the cerebro-spinal fluid. The writer has made experiments along the same line as Cap- PELLETI and Petrir and Grrarp. No claim is made for origin- ality, but it is hoped that a brief description of the results may y be of interest both as corroborative testimony and also because some of the forms used had not been studied by the investigators just mentioned. The canula described by Cavazzani for his cerebro-spinal fistula is a rather complicated instrument. In practice it was found that with care a simple glass canula with a short rubber tube was satisfactory, at least for demonstrating the action of drugs. The musculature is cleaned away and the dura mater over the space between the atlas and the occipital bone laid bare. A small puncture is made and the canula inserted. ‘The elasticity of the dura is enough to press the tissue against the tube, and thus avoid loss of fluid or entrance of blood. In etherized dogs, the insertion of the canula is followed by a rush of fluid. This is due to a secre- MEEK, Choroid Plexus. 2.99 tion under the influence of the ether, and an accumulation of the fluid. A flow of about two drops per minute develops. If 1 per cent pilocarpin now be injected through the femoral vein the secretion becomes more marked, and may more than double in amount. An injection of atropin will abolish the secretion entirely. It is realized that the account just given does not contain suf- ficient data. Many experiments should be made before definite statements are submitted, regarding the rate or amount of flow, or the time in which it develops. It is hoped that these omissions may be supplied in some succeeding paper. ‘The important fact, however, seems perfectly clear, and that is, that the secretion of the cerebro-spinal fluid is accelerated by pilocarpin and retarded or abolished by atropin. . Microscopic examination has been made of the plexuses of dogs and rabbits that had been under the influence of ether for 20 min- utes, and alsoof rabbits, guinea pigs, and rats injected with muscarin. The latter drug was most useful when diluted to 1-500. ‘Two injections were made from 15-20 minutes apart, and 15 minutes later the animal was killed, and the plex- uses removed. ‘The injections were Fic. 8. Three cells from an madeswith avhypodemmic syringe, and sub z>Pits plexus alee ng TEESE etic was given. tion of muscarin. eee ‘$ P : X 1500. a, Marginal zone which inghestatyniiscatm did not give de- 4c pecome a thin pater ine, cisive results, but inthe case of rabbits clear spaces toward the apex; ¢, and guinea pigs the results were defin- _ basal granular zone. ite. Often as many as two-thirds of the cells showed evidences of secretion. Fig. 8 shows cells from a rabbit’s plexus after treatment with muscarin. Normally, the epithelial cells in the rabbit are about 6 micra high (see Fig. 4) but here the height has increased to 12 micra. A differen- tiation into two zones, a basal granular, and an outer clear, is suggested, but it is not quite so well marked as in the fig- ures of Petrir and Grrarp. ‘The granulations, however, are always heavier and more compact toward the base of the cell. Clear spaces begin to appear toward the top, and rarely does the stainable cytoplasm extend to the upper cell wall. Masses of larger granules are common in the upper part of the cell where 300 Fournal of Comparative Neurology and Psychology. the lines forming the reticulations cross. The nucleus remains globular with a clear outline, in fact is not distinguishable from that of the resting cell. “The two things most striking about these modified cells are their great increase in height, and the appearance of so much clear space at the distal side of the nucleus. The apical structures of epithelial cells according to VIGNON (or) take no part in the functional activity of the cell. In case of the extrusion of droplets, they may be pierced, but there is no change in their structure during the process. In the case of the rabbit, where there are fatty droplets, which we consider a secre- tion of secondary importance, we can find no evidence of any change in the marginal zone of the cell. ‘The globules evidently pass out without anything more than rupturing the plateau, and this soon reunites. In the case of the typical secretion, however, the evidence seems to be that the marginal zone of the cell is Fic. 9. Cells from a dog’s plexus after ether anesthesia. Secretory changes are to be noted Magnification X 1800. a, Marginal zone; b clear areas toward apex; c, fat droplet. modified. A comparison of Fig. 4 with Fig. 8, both from the rabbit, will show that the marginal zone of the resting cell has been transformed into a simple, thin, granular wall. We do not believe that the apical wall of the cell disappears during the secre- tion, but it is evidently considerably modified. Fig. 9 shows secreting cells from a dog subjected to ether anzs- thesia for 30 minutes. Fatty droplets have not been found in the normal epithelium of the dog’s plexus, but it is seen that they appear after use of ether. ‘The third cell from the left shows a large droplet which has compressed the nucleus. As in the pre- ceding case, the cells are twice as high, the outer part of the cell contains many clear areas, and the marginal zone has been trans- formed into a simple granular line. The mechanism of secretion, not only in the choroid plexus, but in general, has been a much discussed subject. In this connec- MEEK, Choroid Plexus. 301 tion, we may quote some observations on the vesicular theory. LuscuKa (’55), FINDLAY (99), STUDNICKA (00) and GALEOTTI (97) have all observed small globules within the epithelial cells of the plexus, and on the free surface of the plexus itself. “These they have taken as evidence of the vesicular secretion by the epithelial cells. Prerrrr and Grrarp (’02-03) call attention to the fact that these globules increase in number with the time the tissue remains tated: and also with the slowness of fixation. These globules are analogous to the sarcode globules which are known to be due to post-mortem changes. These globules have been reported for fresh tissue, but they must have been due to mechanical injury. If any one will crush a fresh plexus under the microscope, he can see these structures form before his eyes. It would seem that in the choroid plexuses fatty droplets may be extruded in a way prescribed by the vesicular theory, but the nor- mal secretion passes through the marginal zone in another way, possibly by some kind of canalicull. The chemical composition of the cerebro-spinal fluid has often been advanced as a proof of its being a secretion and not a transu- date. Hariipurron (’89) has shown that it is peculiar and dif- ferent from the lymph and blood serum in its proteids and in the presence of a reducing substance. ‘The proteids are a globulin coagulating at 75° C. and albumoses and peptones. ‘The reduc- ing substance HaLLiBuRTON identified as pyrocatechin, but this identification does not seem to have been confirmed, later investi- gators reporting the presence of glucose. Inthe amount of sodium and potassium salts, there is but little difference from the blood serum. The function of the cerebro-spinal fluid and its circulation are of interest, but related only indirectly to our subject, so we shall discuss them very briefly. The general idea has always been that the fluid is a nutritive solution to nourish the nerve cells with which it may come in contact. Sprna (’01) has shown that absorp- tion of the cerebro-spinal fluid may take place, since fuchsin injected into the subdural spaces appears in the jugular vein. He concludes that the absorption is proportional to the pressure exerted on the fluid. ‘The plexuses themselves, however, can scarcely be resorptive organs, for the blood pressure in the capil- laries is too great to admit of it. Still it is not inconsistent to believe that the cerebro-spinal fluid may be absorbed to some 302 Fournal of Comparative Neurology and Psychology. extent by the walls of the ventricles, and thus reach the true ner- voustissues. LEwaNnpowsky (’01) finds that an amount of strych- nine so small that when injected into the blood there is no result, causes violent spasms if injected into the subdural spaces. From this he argues that the fluid bathes the nervous tissue and carries the poison directly to the nerve cells. “here can be no doubt that the fluid is in intimate relation to the nervous tissues, nor that small amounts of poison might reach the nerve cells more surely by way of the subdural channels, than by the systemic blood stream, still the fluid can scarcely be distinctively nutritive. It is too poor in proteid. One function that it undoubtedly has 1s the purely mechanical one of supporting the nervous walls, pre- venting friction, and serving as a water bed for the brain and cord. This is probably its main function. Prrrir and Grirarp have suggested that it may also serve as some kind of an internal secre- tion, but for this suggestion no evidence has been presented. Having been secreted, the cerebro-spinal fluid accumulates in the ventricles and their continuations, and in all the spaces beneath the dura. The connection between the ventricular cavities and the subdural spaces has been thought to be by means of the fora- men of MacENDIE and the two lateral foramina of the fourth ventricle. According to Mrzian (’04), some doubt seems to have been thrown on the presence of these foramina by the recent work of Cannieu and Gentes. ‘They consider the above mentioned foramina due to post-mortem changes, and in a number of cases the foramina could not be demonstrated. It is possible, then, that the manner of communication between the ventricles and the subarachnoid spaces is not yet definitely known. Mi ian (’04) and CaTHELIN (’03) hold that the fluid reaches the lymphatic system from the subarachnoid spaces by means of the arachnoid sheaths around the nerves, the Pacchionian bodies and the peri- vascular lymphatic sheaths. SUMMARY. The animals used in this study were the albino rat, guinea pig, cat, dog, sheep and man. ‘The best microscopic preparations were made from material fixed in Bouin’s or CarNoy’s solution, and then stained in iron haematoxylin and fuchsin. As a control, examination was made of fresh material either in cerebro-spinal fluid or in physiological salt solution. Meek, Chorotd Plexus. 303 The choroid plexuses are due to the invagination of the neural wall by the pia mater. They are thin lamine with an epithelial covering, derived from the neural wall, a connective tissue stroma, and a rich supply of blood vessels. Villi are absent in the rat, guinea pig, and mouse, but they are present in man, horse and ox. Villi are poorly developed in most birds, but they are especi- ally well developed in the selachians and the crocodile. In area, the plexuses have a greater free surface than would be supposed. ‘This free surface is quite large enough to account for the secretion of the cerebro-spinal fluid. In the case of the albino rat, the plexuses increase in size grad- ually, from their first appearance, until they reach their maximum growth, and at no time do they seem to be relatively enlarged, or to fill all of the ventricular space. Microscopic examination of the fresh plexuses show that they are semi-transparent membranes, covered with ill defined cubical cells. The cytoplasm is finely granular. Cell walls and nuclei are dimly visible. Stained material shows that the capillaries have an endothelial intima, which is strengthened by connective tissue. The epithe- lial cells in the albino rat are 8-10 micra high, and 10-12 micra wide. The nuclet are oval or circular and after birth located basally or centrally. The cytoplasm ts reticular with granules in the meshes. The free or apical edge of the cell 1s convex, and consists of a cuticle which ts little developed 1n the rat, but is more highly differentiated in the rabbit and dog, where it appears as a filamentous plateau. Inno case have cilia been found on these cells after birth. The epithelial cells of the plexuses have lost all basal projections which are characteristic of the ependymal cells from which they were derived. The epithelium consists of a single layer, and the cells are closely appressed without intervening intercellular spaces. In the case of the rabbit, many of the epithelial cells contain droplets of fat. T hese increase 1n size as they approach the apex of the cells, and appear to be extruded without causing a destruction of the cell from which they come. Vasomotor nerves to the blood vessels have been reported by BENEDIKT (’73) and FinpLay (99). The epithelial cells of embryonic and young forms, differ from 3Ttalics are used to indicate the results which are new, or to which the author believes he may have made some contribution. 304. ‘fournal of Comparative Neurology and Psychology. those of the adult, in being more columnar in shape, less easily stained, and in having the nuclei located nearer the cell apex. Soon after birth the cells rather rapidly assume the adult form. In the rat this is accomplished by the seventh day. Only recently has definite proof been offered that the plexuses are concerned in the secretion of the cerebro-spinal fluid. ‘This proof rests on the following facts: The administration of ether, pilocarpin or muscarin, increases the flow of the fluid. Atropin or hyoscyamin retards the flow. After the administration of muscarin or ether, if the plexuses be removed, the epithelial cells will be found very much modified. The apical portion has increased in height, and become clear, while the basal portion has remained granular. In other words, they show changes characteristic of typical secreting cells. The fact that the flow of cerebro-spinal fluid is increased by injection of drugs that usually stimulate secretion in the epithelial cells, and that the cells themselves show secretory changes, justifies the conclusion that the fluid is secreted by the choroid plexuses. The ependymal cells may have their part in the production of the cere- bro-spinal fluid, but it is certainly a minor one compared to that of the choroid plexuses. Additional evidence that strengthens this conclusion is the occur- rence of hypertrophy of the plexuses in certain cases of hydro- cephalus. ‘The fact that the fluid differs from the serum or lymph is also in favor of the view that it is a secretion. BIBLIOGRAPHY. BENEDIKT. ’73. Ueber die Innervation des Plexus choroideus inferior. Canstatt’s Archives. CapPeEL_etl, E. ; ’or. L’ecoulement du liquide cérébrospinal par la fistule céphalorachidienne en conditions normales et sous ’influence de quelques médicaments. Archiv. Ital. de Biol., Vol. 36, Pp- 299-302. Cavazzanl, A. ’93. Sur la circulation du liquide cérébro-spinal. Archiv. Ital. de Biol., Vol. 18, p. 475. Cavazzanl, E. ’99. Die Cerebrospinalfistel. Centralblatt fiir Physiologie, Vol. 13, p. 345- Catota, G. o2. Sulla presenza di neuroglia nella struttura du plessi coroidei. Riv. di Patol. Nerv. e Ment., No. 9. CHARPY. ’o1. Traité d’anatomie de Poirier et Charpy. t. ili, p. 147. Meek, Chorotd Plexus. 305 CaTHELIN. ?03. Presse Médicale, No. 90. DEjERINE. ’95. Anatomie des Centres Nerveaux. Faivree, E. ’54. Recherches sur la structure du conarium et des plexus choroideus chez "homme et chez les animaux. Comptes Rendus del’ Acad. Sciences, Vol. 39, pp. 424-427. Finpiay, J. W. ’99. The choroid plexuses of the lateral ventricles of the brain. Brain, Vol. 22, pp. 161-202. Ga eortl, G. ’97. Studio morfologico e citologico della volta del diencefalo in alcuni vertebrata. Rrv. di Pat. Nerv.e Ment. Vol. 12, pp. 480-517. HAECKEL. 60. Zur normalen und pathologischen Anatomie des Plexus choroideus. Schmidt’s Fahr- biicher. HE pt. ’74. Ueber die Kalkoncretionen in den Plexus choroideus des menschlichen Gehirns. Thesis. Ha.ureurton, W. D. 89. Cerebro-spinal fluid. ourn. Phys., Vol. 10, pp. 232-258. Immamura, S. ’02. Beitrage zur Histologie den Plexus choroideus des Menschen. Arbeit aus dem neurolo- gischen Institute an der Weiner Universitat. Herausgegeben von Professor H. OBER- STEINER, Fasc. 7, pp. 272-280. KOcuiker. ’96. Handbuch der Geweblehre des Menschen. Kincspury. ’97. The encephalic evaginations in ganoids. Journ. Comp. Neurology, Vol.7, pp. 37—44- Lorprer, Maurice ’o4. Sur quelque points de l’Histologie Normale et Pathologique des Plexus Choroides de Vhomme. Archiv. de Méd. Expérimentale, Vol. 16, p. 473. LuscHKka. ’55. Die Adergeflechte des menschlichen Gehirns. LEWANDOSKY. ?or. Zur Lehre von der Cerebrospinal-fliissigkeit. Neurol. Centralblatt, p. 497. Mian. 04. Le Liquide Céphalo-rachidien. Paris. 1904. OBERSTEINER, H. ’or. Anleitung beim Studium des Baues der Nervosen Centralorgane im gesunden und kranken Zustande. Pertir A. er J. Grrarp. ’o2~’03. Sur la Fonction Secretoire et la Morphologie des Plexus Choroides. Archiv. d’ Anat. Mic., Vol. 5, pp. 213-264. Renaut, J. ?99. Traité d’Histologique pratique. t. ll, pp. 339 et 719. SrupnickA, A. K. ’oo. Untersuchungen tiber den Bau des Ependyms der Nervésen Centralorgane. Anat. Aeft., Vol. 15, pp. 303-429. 306 = ‘fournal of Comparative Neurology and Psychology. Spina, A. or. Untersuchungen tiber die Resorption des Liquors bei normalen und erhéhtem intercra- niellem Druck. Neurol. Centralblatt, Vol. 20, p. 224. VIGNON. ’or. Réchérches de Cytologie générale sur les Epitheliums. Archiv. de Zool. Ex. et Gén., Vol. 9. p. 371. The Journal of Com es bl eesti and eaniee VotumE XVII - yULy, 1907 NUMBER 4 tHE TACTIEE CENTERS IN THE SPINAL CORD UAND BRAIN OF THE SEA ROBIN, PRIONOTUS GAROLINUS' L- BY C. JUDSON HERRICK. (Studies from the Neurological Laboratory of Denison University, No. XX1.) Wirnu Firteen Ficures. The nervous system of the gurnards (notably the European genus Trigla) has long been recognized as exhibiting points of special interest from the standpoint of functional differentiation. The brains of these fishes differ but little externally from the usual teleostean type, but the cephalic end of the spinal cord exhibits a series of segmentally arranged dorsal enlargements (“‘accessory lobes,’’ Ussow) which receive huge nerve roots from the specially modified free rays of the pectoral fins. Since the demonstration by MorriLy (’05) that in the allied American genus, Prionotus, the function of the free rays and their specially modified nerves is purely tactile and that neither gustatory sensa- tion nor specially modified end-buds of any description are found upon them, the neurological interest of these fishes is enhanced. For it now appears that these lobes are simply enlargements of the dorsal cornua of the spinal cord and their associated fiber tracts and that they therefore exhibit an extraordinary develop- ment of the unspecialized somatic sensory system uncomplicated by any other modifications. Through the kindness of Professor I. A. Fietp of Western Maryland College I have received a number of brains, with a portion of the spinal cord attached, of Prionotus ‘carolinus from which serial sections were cut for me in the transverse and hori- zontal planes and stained by the method of WeicEerT by my 308 =“fournal of Comparative Neurology and Psychology. former pupil Mr. P. S. McKissen, to whom I am also greatly indebted for assistance in other ways. Morritu has figured an excellent dissection of the organs here under consideration in Prionotus carolinus, and also the histo- logical details of the peripheral terminations of these modified spinal nerves; but the central connections of these nerves have not been analyzed microscopically so far as I know. In Fig. 1 is given a sketch of the external form of one of my alcoholic specimens as seen from the left side. The spinal cord caudad of the so-called accessory lobes is considerably larger than is usual among the teleosts, the enlargement being confined chiefly to the dorsal and lateral parts (Fig. 2). The ventral funiculi and ventral commissure are as usual. ‘The tractus spino-tectalis (fas- ciculus lateralis, Mayser, the tract which is continuous ceph- alad with the lemniscus of my nomenclature in former papers) is well-defined, but not greatly enlarged, immediately behind the accessory lobes; farther caudad it can with difhculty be distin- guished from the adjacent fasciculi propri. ‘The latter region is enlarged and filled with fine medullated longitudinal tracts (fundamental lateral tracts). Farther laterally the ventro-lateral and dorso-lateral fasciculi are still more greatly enlarged, the fibers being larger and with denser sheaths, especially dorsally. The dorsal cornu and dorsal funiculus, which are very small in most telecosts, here comprise nearly one-half the total cross-section of the spinal cord. ‘This region is made up of small bundles of medulated fibers separated by dense masses of unmedullated fibers or neuropil. Strong tracts pass obliquely laterally and ventrally between these masses and the dorso-lateral fasciculus and more diffuse fibers, chiefly unmedullated, to the fasciculi propril (formatio reticularis) and dorsal commissure. At inter- vals, also, bundles of medullated fibers pass along the extreme external surface between the dorsal and ventral funiculi. “These are uncrossed. The longitudinal tracts in the dorsal funiculi are mostly short paths. Individual bundles as a rule do not long remain distinct, and medullated tracts leave them at frequent intervals to enter the fasciculus dorso-lateralis, where they appear to turn caudad, for the latter fasciculus increases in size and compactness for a time as it passes backward. Ultimately, however, it blends with the adjacent fasciculus ventro-lateralis and fasciculus proprius, Herrick, Tactile Centers of Prionotus. 309 as in teleosts generally. The tracts which I here term dorsal funiculi are apparently chiefly descending secondary fibers from the accessory lobes. It is doubtful whether either here or in the region of the accessory lobes there is any considerable true funic- ulus dorsalis as that term is used in higher vertebrates, though the scattered bundles of root fibers and secondary tracts which permeate the dorsal cornu and, farther cephalad, the accessory lobes in a general way correspond with the dorsal funicull. As the sections are followed toward the head, at the level of the fourth spinal nerve the grey matter of the dorsal cornu becomes still more greatly enlarged, the sensory root of this nerve terminating in the center of this dense mass of neuropil. Head- ward of the fourth nerve the dorsal cornu becomes more compact with bundles of well medullated fibers scattered within the neuro- pil, the whole complex occupying the dorsal third of the cross- section of the spinal cord. Large, medullated tracts of descend- ing fibers run from the dorsal cornu into the dorso-lateral fascicu- lus and fasciculus proprius of the same and opposite side, the latter crossing in the dorsal commissure. Both dorsal and ventral commissures are very small in this region, the latter especially being no larger than usual in fishes with spinal cords of the usual form. As we approach the last (sixth) accessory lobe the medul- lated fiber tracts in the dorsal cornu increase in number, many of these fibers crossing in the dorsal commissure, which increases in size, and the grey matter accumulates chiefly at the dorso- lateral border, where it is overlapped by the accessory lobe, with which it becomes continuous a little further cephalad. Meanwhile the dorso-lateral fasciculus has retained its original form and position, as have all of the more ventral structures; but the tract occupying the position of the dorsal funiculus is repre- sented by a small area of compact medullated fibers bounded laterally by the dorsal cornu and accessory lobe, with both of which it 1s connected by transverse medullated fibers. A short distance farther headward the dorsal cornu fuses with the lateral border of the accessory lobe, the median portion of the lobe being free from the underlying tissue. The accessory lobes vary somewhat in appearance in different specimens. Morriii follows Ussow (’82) and enumerates six of these lobes. Their number in Prionotus might be counted as one or two more or less depending upon whether account is taken 310 “fournal of Comparative Neurology and Psychology. of some of the shallower furrows. That they are really enlarge- ments of the dorsal cornu and not adventitious structures is shown by their continuity caudad with these cornua, and by their rela- tions with the dorsal roots which supply tactile sensibility to the free pectoral fin rays, and also by their secondary connections. These outgrowths evidently arose from the dorso-lateral border of the spinal cord and then from broad pedicles at this point spread in mushroom form both mesially and laterally (Fig. 3). The enormous dorsal root of the third spinal nerve enters the lateral border of the fourth, fifth and sixth lobes and terminates within them. Secondary tracts in large heavily medullated bundles pass from these accessory lobes ventro-laterally to descend in the fasciculus dorso-lateralis. The greater part of this tract caudad of this point is apparently derived from the sixth lobe, only a small pro-— portion of the descending fibers of this tract coming from the spinal cord farther toward the head. Very large tracts of medullated fibers pass in the fasciculus dorso-lateralis between the fifth and sixth lobes, which are separated by a wide and deep furrow (Fig. 4). The dorsal commissure contains massive bundles of medullated decussating fibers in the inter-lobarspaces as well as in the regions of the lobes. These fibers run between the lobes and the opposite ~ dorsal and dorso-lateral tracts. As we pass cephalad under the fourth and fifth accessory lobes, the relations of the funiculus ventralis are little modified; the dorsal and ventro-lateral fascicles are somewhat enlarged; and the dorso- lateral very greatly so. This enlargement is due partly to fibers which run between the fifth and sixth lobes and also to large bundles of root fibers, which pass toward the head to end in the fourth and fifth lobes. These lobes, like the others, send tracts of medullated fibers downward to the ventral cornua and ventro- lateral fasciculi of the same side. Between the fourth and third lobes (Fig. 5) the dorso-lateral funiculus shrinks to small dimensions—scarcely larger than in some other fishes, like Ameiurus. The tract which I have termed the funiculus dorsalis, however, maintains its large size. From these relations of the dorso-lateral fasciculus, which is evidently the chief path of communication between these tactile centers of the spinal cord, it appears that the fourth, fifth and sixth lobes, which receive the dorsal root of the third spinal nerve, constitute a Herrick, Tactile Centers of Prionotus. 311 functional unit, being more closely related with each other than with any structures above or below them. Within the third lobe the fasciculus dorso-lateralis rapidly increases again in size and a lateral protrusion of this lobe receives the dorsal root of the second spinal nerve (Fig. 6). Between the second and third lobes the dorsal root of the second spinal nerve enters and distributes to both of these lobes. Between them the dorso-lateral fasciculus also increases to its maximum size, composed largely of dorsal root fibers of the second spinal nerve, some of whose fibers extend forward to the first lobe. The first and second lobes are somewhat smaller than the others and are but indistinctly separated by a wide shallow groove. Ata point near the cephalic end of the second lobe the grey mass of the lobe (dorsal cornu), which behind this point is separated like the other lobes from the lobe of the opposite side by a deep dorsal fissure, fuses across the median line (Fig. 7). “This fusion, which at first involves only the second lobe, a little farther forward becomes overlapped by a similar fusion of the first lobe, the second lobe being thrust far forward embedded under the first (Fig. 8). In this way there are formed two commissural grey masses, a dorsal connected with the first lobe, and a ventral connected with the second lobe, both lying dorsally to the ventricle and both containing cells and medullated and unmedullated commissural fibers. ‘hese median grey masses constitute the somatic commis- sural nucleus and the fibers which decussate associated with them the somatic element of the commissura infima Halleri. The com- missure is continuous caudad with the dorsal inter-lobar commis- sure already described. It is considerably enlarged at its cephalic end just behind the membranous roof of the fourth ventricle (Fig. 11). [he somatic commissural nuclei in WEIGERT preparations resemble closely in structure the accessory lobes with which they are directly connected. The first accessory lobe functions also as the nucleus of the spinal V tract. It is related with the fasciculus dorso-lateralis and with the scattered tracts which I have designated funiculus dorsalis and so may be regarded as sharing some of the functions of the nucleus funiculi, though the latter structure is separately represented in a well defined nucleus associated with the fascic- ulus dorso-lateralis. 312. “fournal of Comparative Neurology and Psychology. The dorsal root of the first spinal nerve is small, entering the caudal end of the first accessory lobe. From the cephalic end of the second lobe a short and broad tract of medullated fibers passes ventrally into the formatio reticularis and dorso-lateral fasciculus (Fig.g). Here most of the fibers seem to end in the large fasciculi proprii which fill this region and pursue only a short course head- ward or tailward. This tract is merely a concentration of a similar connection found in all of the lobes. The formatio retic- ularis, which under the other lobes is composed mostly of alba, here becomes somewhat more loosely aggregated, with large fasci- culi propri scattered through a grey field. The grey matter is nowhere extensive. As the first accessory lobe merges with the medulla oblongata it shrinks in size and becomes enveloped dorsally by the fibers of the very large spinal V tract. Meanwhile the number of fibers in the dorso-lateral fasciculus diminishes, some passing into the formatio reticularis and others disappearing into a mass of grey which occupies the whole of the median part of the. fasciculus (Fig. 10). This mass is the nucleus funiculi:. From this nucleus a strong tract of heavily medullated internal arcuate fibers passes ventro-mesially, decussating in the ventral commissure, to reach the ventral funiculus and the ventral cornu of the opposite side at the level of the origin of the first ventral root of the first spinal nerve. As we pass forward into the oblongata, just before the level of the vagal lobe is reached, the last vestige of the first accessory lobe disappears and the somatic commissural nucleus shows no longer any distinction between its dorsal and ventral portions. Massive medullated commissural tracts pass between the funicular nuclei of the opposite sides, greatly augmenting the most cephalic part of the commissura infima. There are also broad connections between these nuclei and the formatio reticularis of the same side and the ventral funiculi of the opposite side via the ventral com- missure. As we reach the level of the vagal lobe, more than half of the fasciculus dorso-lateralis has terminated in the funicular nucleus and this nucleus itself is rapidly replaced by the vagal lobe (Figs. 11 and 12). A considerable portion of the dorso-lateral fasciculus continues cephalad into the oblongata, apparently without inter- ruption in the funicular nucleus region. ‘These are probably Herrick, Tactile Centers of Prionotus. 213 ascending correlation fibers of the tactile system, since this tract is larger in this fish than in any other which I have examined and there is no enlarged structure in the oblongata from which its fibers may arise. here are also added to the ventral side of the ventro-lateral fasciculus large heavily medullated tracts which arise from the funicular nucleus and pass cephalad into the oblon- gata. [hese are believed to be in the main ascending correlation tracts for the same reason as those last mentioned. Some of these tracts enter the dorso-lateral fasciculus by way of the adjacent formatio reticularis. Probably there are important descending secondary tracts from the funicular nucleus in the lateral and dorso-lateral fasciculi, but they are not separately distinguishable. Contrary to my expecta- tions, neither the funicular nucleus nor any of the accessory lobes send large secondary tracts to the ventral funiculi. The very large internal arcuate tracts from the funicular nucleus nearly all terminate in the ventral cornu immediately adjacent. Nor do any large numbers of these secondary fibers ascend in the fasciculus lateralis (of Mayser, “lemniscus,” HERRICK), as was to have been expected by analogy with other fishes. In short, the enlarge- ments represented in the accessory lobes represent a sensori-motor mechanism for the free rays of the pectoral fins, and these modifh- cations do not extend far beyond the limits of the segments of the spinal cord directly involved in the innervation of these fins. The modifications of the spinal cord behind the “lobes” is more evident than that of the medulla oblongata in front of them. ‘This doubt- less 1s correlated with the fact that the sensory stimulation of the free rays 1s more apt to call forth swimming movements of the trunk and tail musculature than any cephalic reaction. The specialized apparatus of these free rays, then, 1s adapted for reflexes of only the simplest order. A study of the series of excellent figures of Lophius given by KappeErs (06) reveals in the funicular nucleus region an arrange- ment very similar to that of Prionotus. The somatic sensory centers are much enlarged and are designated nucleus Rolandi and, farther caudad, lobus sensib. Probably these masses of grey include both my funicular nuclei and the spinal V nucleus. ‘The commissura infima is no doubt chiefly somatic, the visceral centers being small in this fish. The accessory lobes, as we have already seen, are enlarged 314 ‘fournal of Comparative Neurology and Psychology. dorsal cornua. The dorsal root of the first spinal nerve, which terminates in the first lobe, is, however, not greatly enlarged. The prime motive, then, for the development of this lobe is not the modification of the pectoral fin, but, rather to serve as the terminal nucleus of the spinal V tract, which is of very great size in Prionotus. In this fish the spinal V nucleus is a direct derivative of the dorsal cornu. The funicular nucleus is not closely related to the dorsal cornu (the grey matter of the “lobes”’), but rather with the ad- jacent formatio reticularis. In both of these respects Prionotus agrees very closely with the morphological relations in Ameiurus (Herrick ’06), although the structural details are very different. The spinal V tract, as in other fishes, can be separately followed by reason of the heavier medullation of its fibers, into the spinal cord beyond the first spinal nerve, a remnant of it being recognized in Fig. g. As we pass forward into the oblongata from the funicular nucleus region, the first accessory lobe (dorsal cornu + nucleus of spinal V tract) disappears before the funicular nucleus, and the latter 1s enveloped dorsally by the spinal V tract, laterally by the dorso- lateral fasciculus and ventrally by the latter tract and massive bundles of the formatio reticularis alba, which farther forward become incorporated into the dorso-lateral fasciculus (Fig. 11). When the level of the vagal lobe is reached-the funicular nucleus has entirely disappeared, its place being occupied by a very small substantia gelatinosa Rolandi. Here the spinal V tract receives a strong accession of descending fibers from the general cutaneous root of the vagus (Fig. 12), and there is added to the dorso-lateral fiber complex on its ventral side the ascending secondary gus- tatory tract from the vagal lobe. Whether there is a descending secondary gusatory tract from the vagal lobe represented in the fasciculus lateralis my preparations do not enable me to state. I find no clear evidence of it. Secondary vagus fibers clearly pass in large numbers to the formatio reticularis of the same side and others cross as internal arcuate fibers in the ventral commissure (Fig. 12). These are of fine caliber, like the other secondary vagus fibers, and can therefore be distinguished from the other and coarser fibers of this commissure which are derived from the substantia gelatinosa and tuberculum acusticum and belong to the secondary somatic sensory system. ‘They arise from the whole extent of the vagal lobe, which is not obviously differentiated into Herrick, Tactile Centers of Prionotus. 315 various functional regions. ‘They pass into the opposite formatio reticularis for the most part, apparently to effect connection with the dendrites of the nucleus ambiguus. Unlike the somatic com- missural fibers, they do not appear to enter either the funiculus ventralis or the fasciculus lateralis. In this they differ from the secondary fibers which enter the ventral commissure from the lateral division of the vagal lobe in Gadus. In the latter species I have found (’07) that the cutaneous taste buds are innervated from a different part of the vagal lobe from those taste buds which lie in the mucous membranes of the mouth and phary nx, the secondary connections of the cutaneous, or “somatic,’’ gustatory center passing through the ventral commissure to the opposite somatic motor centers by way of the funiculus ventralis. Priono- tus, however, 1s not known to possess taste buds in the outer skin, nor does it react to gustatory stimuli applied to the skin (cf. HER- RICK '04, p. 264), and since the secondary gustatory path from the vagal lobe to the opposite funiculus ventralis in Gadus is a special adaptation called forth by the cutaneous taste buds, we are not surprised to find that the commissural secondary gustatory tract terminates differently in Prionotus. ‘The primary gustatory cen- ters of Prionotus are not large, though their peripheral and central connections are of the typical form. ‘The crossed secondary con- nection in the ventral commissure, though present in other fishes, is relatively larger in this species, a peculiarity which I am not able to explain. The vagal lobes are extended caudad into the visceral com- missural nucleus under the commissura infima Haller: in the typical teleostean way, though this nucleus is small and not clearly separated from the much larger somatic commissural nucleus (Fig. 11). The visceral commissural nucleus receives sensory root fibers from the vagus and sends a broad unmedullated second- ary tract to the formatio reticularis. “The visceral partof the com- missura infima 1s diffuse and chiefly unmedullated. The tubercula acustica are very large, but I find no indication of the commissural fibers which pass between them in the com- missura infima, such as appear in Conger. The general composition of the fasciculus dorso-lateralis proxi- mally of the funicular nucleus has already been indicated. This tract,as I here define it, includes the whole dorso-lateral fiber com- plex and therefore more than is comprised in the “ dorso-lateraler 316 Journal of Comparative Neurology and Psychology. Strange” of GoronowirTscH (’88, ’96). The “System 7’ of this author’s descriptions comprises chiefly the ascending and descend- ing secondary gustatory tracts. [he descending fibers of the spinal V and spinal (general cutaneous) vagus roots occupy the most dorsal partof the cross-section of the dorso-lateral fasciculus. This tract is very large and more heavily medullated than the other fibers of this complex, so that its course can easily be separ- ately followed toits terminus in the first accessory lobe of the spinal cord. The more ventral bundles of this fasciculus in the oblon- gata comprise tracts (probably mainly ascending) which pass between the first accessory lobe and the adjacent formatio reticu- laris alba and the funicular nucleus on the one hand and the for- matio reticularis of the oblongata farther proximally, on the other hand. ‘There are other tracts in this fasciculus which pass be- tween the oblongata and lower regions of the spinal cord, but they are so confused with the shorter tracts that it is not possible to follow them separately for their entire length. As the dorso-lateral fasciculus passes under the vagal lobe, it receives on its ventral side the fine fibered ascending secondary vagus tract in the way characteristic of teleosts in general (Fig. 12). Farther forward, at the level of the origin of the sensory [X root (Fig. 13), the fasciculus becomes very compact and deeply em- dee in the substance of the oblongata under the massive tuber- culum acusticum. Here the ances: chief elements mentioned attain about equal proportions in the area of the cross-section of the fasciculus, viz: the spinal V tract dorsally, the ascending secondary gustatory tract ventrally and between them fibers - the fasciculus lateralis of mixed character, probably mainly of the fasciculus proprius type, putting the accessory lobes of the spinal cord and the funicular nucleus into relation with the medulla oblongata. he greater part of the latter fbers break up in the formatio reticularis under the tuberculum acusticum. Tractus Cerebello-spinalis.—There is, however, one important tract in the complex last mentioned, whose relations are more clearly brought out in these sections than in those of any other fish which I have examined. Its fibers are heavily medullated and on the average of greater diameter than those of the spinal V tract, with which they are closely associated. ‘There is no possibility of confusing them with the latter fibers, nor with any others of the fasciculus lateralis complex, all of the remainder of these being Herrick, Tactile Centers of Prionotus. a7 smaller and more lightly stained in WEIGERT preparations. ‘This tract can be clearly followed between the granular layer of the body of the cerebellum dorsally of the superior secondary gusta- tory nucleus (Fig. 15) and the spinal cord in the funicular nucleus region. It retains its individuality perfectly as far,back as the vagal lobe (Figs. 13 and 14); but in this region its outlines become somewhat confused with those of the spinal V tract and other elements of the dorso-lateral fasciculus. It can, however, be separately distinguished, though its outlines are not clearly defined, as far as the funicular nucleus (Figs. 11 and 12). I have no doubt that it extends far back in the spinal cord in the fasciculus dorso-lateralis below the accessory lobes, as figured by EDINGER (96, p. 61, Fig. 30) for Trigla, though I am not able to follow its fibers separately. Whether the direction of conduction is ascend- ing or descending, I am not able to state. In other fishes where I have found a spinal cerebellar tract it lies farther laterally and ventrally in the oblongata. Except in Ameiurus, where | have described it (06, p. 413) under the name tr. spino-cerebellzris. I am unable to account for its association with the spinal V tract in Prionotus, nog even to assert positively its homology with the more lateral tract of other fishes. CONCLUSION. The six “accessory lobes” or dorsal swellings of the cephalic end of the spinal cord of Prionotus are simple enlargements of the dorsal cornu evoked by the highly differentiated tactile organs on the free finger-like rays of the pectoral fins. The first spinal nerve is but little enlarged, but the dorsal roots of the second and third are greatly so and give rise to the second to sixth accessory lobes. The first lobe receives the first spinal nerve, very large spinal V and spinal X tactile roots and secondary fibers of the fasciculus pro- prius type from the accessory lobes farther caudad. ‘The second spinal nerve terminates in the second and third lobes, which are broadly connected by short secondary tracts in the fasciculus lateralis. ‘The third spinal nerve terminates in the fourth, fifth and sixth lobes, which are closely bound together by massive short secondary tracts, as are the second and third. But between the third and fourth lobes these short tracts (fasciculi proprii) are 318 “Fournal of Comparative Neurology and Psychology. relatively very feebly developed. From the sixth lobe massive tracts run caudad in the spinal cord, which is much larger than usual among teleosts, though the remaining nerves are not greatly enlarged. ‘This enlargement is seen in the dorsal, dorso-lateral and ventro-lateral tracts, including the funiculi propri, but does not involve the ventral funiculi. This would imply that reflex movements of the trunk musculature habitually follow tactile stimulation of the free pectoral fin rays, an inference which is sub- stantiated by observations made on the living fish. But as we pass toward the head from the first accessory lobe, there 1s evident very little modification of the central nervous system due to the enlargements of the spinal cord. These facts show that the reflexes connected with the free pectoral fin rays are of the sim- plest type, not involving extensively the higher cranial centers. The relations described in the preceding paragraph seem to be merely a special case under the general rule formulated by SHER- RINGTON (’06), where he says (p. 58), “Broadly speaking, the degree of reflex spinal intimacy between afferent and efferent spinal roots varies directly as their segmental proximity. aeons The spread of short ‘spinal reflexes in many instances seems to be rather easier tailward thanheadward. * * * ‘Taken generally for each afferent root there exists in immediate proximity to its own place of entrance in the cord (e. g., in its own segment) a reflex motor path of as low a threshold and of as high potency as any open to it anywhere.” It is possible, too, that if an analogous relation holds in the spinal cords of higher vertebrates, it may have some bearing on the further fact brought out by SHERRING- TON (p. 241), that “spinal shock appears to take effect in the aboral direction only.” Associated with the first and second accessory lobes are the large somatic commissural nucleus and commissura infima. A Share distance headward from this lobe there is a very highly developed median funicular nucleus, the lateral nucleus not being differentiated. This nucleus sends massive medullated tracts into the commissura infima. The secondary connections of these extensive somatic sensory centers at the lower end of the oblongata are mainly with the adya- cent formatio reticularis, ventral cornua and dorso-lateral fas- ciculus. No unusually large numbers of long tracts descend in the ventral funiculi or ascend in the crossed fasciculus lateralis Herrick, Tactile Centers of Prionotus. 319 (tractus spino-tectalis or “lemniscus”). ‘This again illustrates the essentially local character of the chief reflex paths associated with the tactile system. ‘There is a cerebellar connection of the funic- ular nucleus region which is stronger than in any other teleost which I have examined. The visceral sensory centers of Prionotus are rather poorly developed, though the visceral commissura infima and commis- sural nucleus have the typical teleostean relations. LITERATURE CITED. Epincer, L. Vorlesungen iiber den Bau der nervésen Centralorgane des Menschen und der Thiere. 5 Ed. Lerpzig. 1896. GoronowitTscu, N. Das Gehirn und die Cranialnerven von Acipenser ruthenus. Morph. Fahrb., vol. 13. 1888. Der Trigeminofacialis-Complex von Lota vulgaris. Festchr. f. Gegenbaur, vol. 3. 1896. Herrick, C. Jupson. The organ and sense of taste in fishes. Bul. U. S. Fish Commission for 1902. 1904. The central gustatory paths in the brains of bony fishes. Fourn. Comp. Neurol. and Psych., vol. 15, no. 5. 1905. On the centers for taste and touch in the medulla oblongata of fishes ‘fourn. Comp. Neurol. and Psych., vol. 16, no. 6. 1906. A study of the vagal lobes and funicular nuclei of the brain of the cod-fish. Journ. Comp. Neurol. and Psych., vol. 17, no. 1. 1907. Kaprers, C. U. A. The structure of the teleostean and selachian brain. ‘fourn. Comp. Neurol. and Psych., vol. 16,) 1091-6 1900. Moraitt, A. D. The pectoral appendages of Prionotus and their innervation. ‘fourn. Morph., vol. 11, no. 1. 1895. SHERRINGTON, C. S. The integrative action of the nervous system. New York. 1906. Ussow. De la structure des lobes accessoires de la moelle épiniére de quelques poissons osseux. Arch. de biol., tom. 3. 1885. 320. © fournal of Comparative Neurology and Psychology. Fic. 1. The central nervous system of Prionotus carolinus seen from the left side. Drawn from an alcoholic specimen which had been fixed in potassium bichromate and designed to illustrate the relations of the spinal nerve roots to the six ‘‘accessory lobes” of the spinal cord (/ob. 1 to /ob.6). The cranial nerve roots are indicated with Roman numerals; t. ac., tuberculum acusticum. X 3. Fics. 2 to 15. A series of transections through the spinal cord and medulla oblongata of an adult Prionotus carolinus, from sections stained by the method of Weicerr. All are drawn to the same scale. X 12. Fic. 2. Section between the fourth and fifth spinal nerves. Areas of unmedullated fibers and neuro- pil are stippled. The enlarged dorsal cornu fills nearly the whole of the dorsal part of the spinal cord. Fic. 3. Section through the middle of the sixth accessory lobe (dorsal cornu), showing the entrance of fibers of the dorsal root of the third spinal nerve. The fibers marked desc. sec. tracts are secondary tactile tracts which descend from the sixth lobe and enter the funiculus dorsalis and fasciculus dorso- lateralis farther caudad. The more medial of the tracts thus designated extends backward to become continuous with “‘funic. dors.” of Fig. 2. Fic. 4. Section taken between the fifth and sixth accessory lobes, including the most cephalic tip of the sixth lobe. This also disappears, a few sections farther forward, leaving the dorsal cornu practically unrepresented in the section. The dorso-lateral fasciculus is very large, containing, in addition to root fibers of the third spinal nerve, massive medullated tracts between the fifth and sixth lobes. Fic. 5. Section taken between the third and fourth accessory lobes. In comparison with Fig. 4, taken between the fifth and sixth lobes, note the great reduction of the dorso-lateral fasciculus. The large size of this fasciculus between the fifth and sixth and fourth and fifth lobes is due to the facts that the third spinal root passes by way of this tract into each of these three lobes and that very large second- ary tactile tracts pass between them. No root fibers pass between the third and fourth lobes and rela- tively few secondary fibers. This indicates that each spinal root supplying the free fin rays is an inde- pendent reflex mechanism within the spinal cord. Herrick, Tactile Centers of Prionotus. 321 TU) C “GOrS == COFANU AOrs,<--\— Gere fasc:dorsolat—~\ com. dorsa@/rs_- tr spino-tecta//s_ 2-4 fasc. ventro-lat.— C¥ae cornu yentralis_z Bs funse. ventralis—_ as acc. lobe b~— fase. dorso-/atera/l — a commiss. dorsal_ =a Se SSS = tr spino-teclal’s \ gp os fyuniculus dorselis>4 acc. lobe 6.- tx. dors. I//, spinal ~ DESC. SEG TFACTS = Commis. dorsal_= tr. spino-tecta lis ) commis. ventral Tasc. dorso-latera/ lasc. dorso-larer.— funte dorsalis~\ Comm /sS. d0r$- => so tr spino-tectal JS Commis. vent-7 TX VEIT as Sp: /. 322 Fournal of Comparative Neurology and Psychology. Fic. 6. Section taken through the caudal end of the second lobe, including a portion of the dorsal and ventral roots of the second spinal nerve. The dorso-lateral fasciculus at this level is composed partly of root fibers and partly of massive secondary tactile tracts between the second and third lobes. Secondary tactile tracts pass from the grey matter of the lobe to this fasciculus and to the ventral cornu and fasciculi proprii adjacent. Fic. 7. Section through the cephalic end of the second lobe, showing overlapping first lobe and the caudal end of the somatic commissural nucleus (comm. nuc.) and contained dorsal commissure, which from this point toward the head may be termed the somatic commissura infima. Fic. 8. Section .225 mm. farther toward the head through the somatic commissural nucleus (comm. nuc.). This nucleus is divided into dorsal and ventral parts, each with a fascicle of medullated commis- sural fibers. A vestige of the second accessory lobe surrounded by medullated secondary tactile fibers (cf. Fig. 9) is embedded in the substance of the first lobe. Fic. 9. Section .225 mm. farther toward the head, showing dorsal and ventral portions of the somatic commissural nucleus and their commissures (designated com. inf. d. and com. inf. v., respectively). The more dorsal fibers, designated sec. t. tr.,are secondary tactile tracts from the cephalic end of the second accessory lobe (cf. Fig. 8); the more ventral ones are short tracts of similar character passing between the first and second lobes (fasciculi proprii). Fic. 10. Section through the cephalic end of the first lobe and somatic commissural nucleus, the latter still showing the division into dorsal and ventral portions. The spinal V tract occupies the dorsal part of the section and is sending terminal filaments into the first lobe, which thus serves as nucleus spinal V as well as dorsal cornu for the first spinal nerve (cf. Fig. 9). ‘The nucleus funiculi appears at this level and receives terminals of the dorso-lateral fasciculus, embedded within whose fibers it lies. Large secondary tactile tracts spring from the funicular nucleus and cross in the ventral commissure (sec. t. tr. cruc.), terminating mainly in the adjacent ventral cornu. Some turn caudad in the ventral funiculi, and probably a smaller number join the tractus spino-tectalis (tr. sp. tect.) to form the fasciculus lateralis, or lemniscus, though the latter are not numerous. Uncrossed secondary tactile tracts (sec. t. tr. rect.) of the fasciculus proprius type enter the formatio reticularis from both the spinal V nucleus and the funicular nucleus. The nucleus ambiguus (uc. amb.) appears at this level and the canalis cen- tralis begins to dilate into the fourth ventricle. Fic. 11. Section through the visceral commissural nucleus and nucleus funiculi. The first accessory lobe and spina V nucleus lie farther caudad and the fasciculus dorso-lateralis is terminating in the nucleus funiculi. The most cephalic part of the somatic commissura infima (com inf.) is shown passing between the funicular nuclei. Ventrally of it is the visceral commissural nucleus which contains no medullated commissural fibers, though diffuse unmedullated tracts cross the median line. At this level this nucleus receives the most caudal sensory vagus root (rx. sens. vagi.) The motor vagus root arises from the nucleus ambiguus farther ventrally. Medullated secondary tactile tracts arising in the nucleus funiculi cross in the ventral commissure (sec. t. tr. cruc.) to reach the ventral cornu and ventral funiculi. From this level forward the tractus cerebello-spinalis (tr. cereb. sp.) can be distinguished from the spinal V and adiacent tracts. Herrick, Tactile Centers of Prionotus. g28 BCCuLONE: ean acc lobe 2— fase. dors-lat ~ tunic. dorsal. Ss ) comm. nuc. com. dors. 5 f tr spino-ted 4 “ye A UnIC. dors. 1X. ventt Sp. 24 rx. darsal sp. f GHC, HONG | acc.lobe Ae F Zia Tasc. dorsofar Wa Sp lect nue. funic.. spinal V tract. ace. /obsa |. & fase. dors-latl nuc, tunic. ht comm. nue.— NUC. amber) he tr sp tecke: 0 aay \ [eee sect lr cru€. SEG LULL CS pFeu! Ge: Ki V/SC. Commy It. cereb-sp, ie TA Sens vag, SS 324 “fournal of Comparative Neurology and Psychology. Fic. 12. Section through the vagal lobes. The section is slightly inclined, the left side being farther toward the head and including the extreme caudal tip of the tuberculum acusticum. On the right side the general cutaneous root of the vagus is seen entering the spinal V tract dorsally of the gustatory vagus root for the vagal lobe. The small amount of grey among the fibers of the dorso-lateral fasciculus is an extension from the funicular nucleus (cf. Fig. 11). Fic. 13. Section through the tuberculum acusticum to illustrate the composition of the fasciculus dorso-lateralis at the level of the origin of the [IX nerve. Dorsally are the spinal V tract and the tractus cerebello-spinalis; ventrally is the ascending secondary gustatory tract. The fibers between, designated fasc. dorso-lateralis, are fasicuclus lateralis fibers of mixed character, between the oblongata and the spinal cord. Herrick, Tactile Centers of Prionotus 325 tuberculum acyst Jobus vagi--{- rx gust vagn— nuc. ambiquus\. t “Vrs i \ Seth : yp r @ x culaneus vag/ ir spino-lectalis iN —~ Ser TX gust vag! \ 2 \fasc dorso-lat EX mol. vagr \ Sec, gustarory ‘tract j “fuberculuin acusfticum \ lateralis X. IX SENS. Vi rx. faterasi/s spinal V [ract _trcere bel/o- spinalts /r spino-lecl- =--fasc. dorso-/aterals sec.gustatory tr “ire sens, 1K 326 = “fournal of Comparative Neurology and Psychology. Fic. 14. Section at the level of the origin of the trigeminus. The elements of the dorso-lateral fasci- culus as shown in Fig. 13, have disappeared in the oblongata farther caudad, for the most part, except the tr. cerebello-spinalis and the secondary gustatory tract from the vagal, lobe. Fic. 15. Section through the body of the cerebellum, illustrating the relations of the tractus cerebello- spinalis in the granular layer. Herrick, Tactile Centers of Prionotus. 229 cerebellum 0-cerebeYX/Qr/s 3 a lv Cae acustico-cerebhe//aris V. a a ay ‘ ESC rice ene wen eae . . . CVAT \y HOG mot. V ormatio relic d/aris— IA ( Te Seynes Ve ebES/O-spinalis Sr fecto> spinalis 14. optic Voie re Ait, aa ; tr tecto-spmalie!/) D7 . ulo-cerebellaki's \ V, | | j; ‘ : Wee - i No if x Z “2. ORG gustators te \7t. tecto-spinalis rectus Tr Spino-tectalis at AN EXPERIMENTAL STUDY OF AN UNUSUAL TYPE OF REACTION IN A DOG. BY Gis WONINT TDS IBUAIMUOE ITO MN 5 Mine (McLean Hospital, Waverley, Mass.) Wirth Two Ficures. These experiments were undertaken in the interests of a prob- lem suggested by such instances of animal behavior as are with difhculty, or not at all, interpretable in terms of instinct or of associative memory, and which may be ascribed to accident or not, according to the sympathies and viewpoint of the observer. “lie practical difficulties in the way of limiting, regulating and repeating the stimuli that seem to produce unusual adjustments render the always objectionable apparatus necessary to a sufhciently critical investigation of such a problem; but negative conclusions drawn from results obtained by apparatus experiments cannot properly include a denial of the possibility that such animals as habitually solve their problems by the “trial and error” method of reaction (where instinct does not serve) may, under exceptional circumstances, display a type of reaction which stands higher in the scale of modifiability of behavior. Simplicity of an experi- mental situation calls for like simplicity of reaction to it on the part of the animal; and any artificial complication of an already artificial situation is apt not to be in line with his general reactive tendencies. It is to be regretted that, owing to the requirements of my method, I was unable to use more than one animal throughout the 600 experiments; but the efficiency of this method, and the suggestiveness of the results obtained by its use will justify, I hope, the report that is to follow. ‘The subject, a bull terrier of mixed breed, was about four months old when he came into my possession. At that time he knew no tricks whatsoever, had never been trained, and was not used to people and houses. Since then he has had no tuition except what has been necessary to over- 330 ‘fournal of Comparative Neurology and Psychology. come certain objectionable habits. Although he has done the usual number of things that are ascribed to “reason” by the uncritically sympathetic, his everyday life has afforded no beha- vior that instinct, associative memory or accident will not explain. Early in his career he was familiarized with a simple apparatus which was equipped with four suspended blocks of wood, one of which, if clawed, would release a door which led to food. These blocks were placed in various positions in the cage, alter- nately attached to the door-release, and labeled in some manner. Thus, colored cards were scattered about the cage, and a white one was always placed on the block that was attached to the door- release. “This was merely for the purpose of making experimental situations a part of his daily life; and especially, to insure that later they would not excite fear or aversion. To a playful, well fed puppy all this meant a good time, food, objects to be chewed up, and the presence of his master after an all day absence, whether he succeeded or not in getting out of the cage. If he did succeed many added and unusual pleasures awaited him, and no suggestion of penalty for failure entered into the game. This preparatory work extended over several months, and in the end afforded me a subject so well fitted for the formal experi- ments that the reaction-value of the experimental situations was far greater, I think, than is usually obtainable by ordinary experi- mental methods. Description of apparatus.—(1) A wooden frame, 4 feet wide, 5 feet long, 21 feet high was covered with coarse wire netting, as was also the lid. (See Fig.1.) (2) A door (4), 1 foot high, 26 inches wide, fitted with a spring to pull it open when the but- ton (B) which held it closed was turned. (3) Four wooden pedals (D) which were passed through slots in the rear base- board, and hinged to a floor railing 11 inches to the rear of this. These pedals were held in a slanting position (anterior ends directed upwards) by means of wire springs. Very slight pres- sure upon their anterior ends was sufficient to nove “hen toward the floor. (4) Four strings, each of which was attached to a pedal at the point of its emergence from’ the cage; from there the individual string was carried upward to the series of horizontal rings that carried all four strings to a trigger (C) at the left side of the cage. The trigger itself, when sprung, released a spring, which in turn pulled “the button aside, thus releasing the door. Hamitton, Unusual Reaction of a Dog. 331 By hooking a given string to the trigger the pedal from which it came was thus attached to the door-release. (5) Four “pedal cards” (F), which were simply heavy wooden boards, 4 inches wide by 11 inches high; they were furnished with two legs each, which fitted into slanting holes in the pedals in such a manner as to give them an upright position when in place. Each pedal card had a different colored paper pasted upon its anterior surface, viz: black, red, green and yellow. (6) Four pairs of odor cards (G); these were small wooden paddles with cotton tacked to their ee ee Fig. 1. Diagram of the experiment cage; 4, spring door; B, button with trigger string attached; C, trigger with one pedal string attached; DDDD, pedals, one of which is shown separately with string and suspension spring; E, sign board; FFFF, pedal cards, one of which is shown separately, with legs to fit into the holes in the pedals, and with odor card (G) suspended from its anterior surface. anterior surfaces. Each of these was kept saturated with a drug having a distinctive odor, viz: asafcetida, lupulin, castor oil and beef extract. (7) Four wooden sign boards (F£), 11 inches wide, 3 feet long. Their anterior surfaces were covered with papers matching in color those of the cards. (8) One white sign board and pedal card, and three plain pedal cards. Experiments with the white sign board and white pedal card.— In this series of experiments the different pedals were attached, one at a time, and in varying order, to the trigger, the attached pedal always bearing the white pedal card, and the unattached 332 ‘fournal of Comparative Neurology and Psychology. pedals bearing plain pedal cards. ‘The white sign board was kept suspended above the row of pedals throughout this series of experi- ments. [he animal was given 160 trials, 20 a day. In calculat- ing results an error was recorded for each attempt to escape by striking an unattached pedal. If the animal struck the attached pedal first during a given trial a “correct first choice” was record- ed for that trial (see table). Time measurements proved to be of absoultely no value, since the animal sought the pedals as his means of escape from the start, and was so dexterous in passing from one pedal to another that there was no essential difference between the time of escape when there were no errors, and when there were several. Besides, as there was no pressing desire to escape, conditions external to the interests of our problem deter- mined the length of his stay in the cage. He always came out directly the shock of the discharged trigger was felt, but often he wandered about the cage contentedly before attempting to escape. As a rule he was fed before experiments were begun; the only in- centives for escape which the experimenter supplied were a few much gnawed bones and a commending caress. It has been my experience that a well fed, contented dog, if undistracted by other sense-impressions than those of the experimental situation, is apt to inspect his immediate environment quite thoroughly, and to act with more appearance of caution than has been reported by pre- vious experimenters with these animals When put into the cage for his first trial the dog went directly to the attached pedal, struck it, and came out. Of course his previous experience easily ac- counts for his having struck at the first projecting object that came his way, and [ do not doubt that his immediate success in getting at the attached pedal was accidental. When he was put in for the second trial he went at once to the pedal that had just let him out and struck it, then started for the door; the white pedal card was now on another pedal, and that pedal alone was attached. But the animal did not seem to take the directing pedal card into account, for he returned to the unattached pedal again and again, striking it more forcibly each time, and always running to the door after each attempt. After ten such attempts he sat down by the door and began to whine. He was spoken to reassuringly, after which he returned to the pedals and inspected them. After some hesitation he struck the attached pedal, thus releasing himself. His behavior during the remainder of the 160 trials showed nothing Hamitton, Unusual Reaction of a Dog. 333 of particular interest beyond a progressive decrease of errors, and an increase of correct first choices. ‘The experiments were dis- continued after sufficient data had been obtained for purposes of comparison with those to be obtained from the more complicated experiments about to be described. Experiments with colors, colors and odors, and odors alone.— At the 161st experiment the white sign board and white pedal card were discarded. ‘The green sign board was hung in position, the green pedal card was placed in p 1 (the first pedal to the left), the black pedal card in p 2, the yellow pedal card in p 3, and the red pedal card in p4. ‘There were sixteen different combinations of this sort, so arranged that each color was represented upon each “pedal attached” once in sixteen times, and upon each “pedal unattached”’ three times in every sixteen. Each of the four pedals was attached four times in a series of sixteen experiments. An irregular order of attaching pedals and placing pedal cards was followed in order to prevent the animal from learning a sequence; but care was taken that no one pedal should be attached twice in succession, that no one color should occur as the sign of “pedal attached” twice in succession, and that every pedal card should be shifted to a different pedal after each experiment. At the 481st experiment the colors were reinforced by odors: red by lupulin, black by asafcetida, green by castor oil, and yellow by beef extract. ‘Thus, if the yellow pedal card (now reinforced by the beef extract odor card, suspended from its anterior sur- face) were on the attached pedal, the yellow sign board was hung out with a beef extract odor card attached to it. At the 581st experiment the colors were abandoned, and only odor cards and odor sign boards were used. ‘These were attached to plain sign boards and pedal cards, identical in form with those that bore the colors. It can be seen from the above that an adequate reaction to the situation required the animal first to seek the sign board, then to inspect it, and finally, to strike the pedal bearing the only card that would afford him the same odor or visual stimuli that he got from the sign board. In this connection it may be said that it was impossible to determine positively whether or not he was ever guided by the colors as such, or even by their differences in light intensity, although the latter increased rapidly in the order from black to yellow. It is probable that the colored papers had each 334 “fournal of Comparative Neurology and Psychology. a distinctive odor, since the animal not only looked at them, but sniffed them carefully before reacting. It is to be remembered that so long as distinctive stimuli were afforded, and so long as the sign boards gave the same stimuli as their appropriate pedal cards, the principle remains the same. Color discrimination as such 1s not a part of our problem, and colors were used only for the sake of a possible reinforcing value. Per White pedal card Sant ETFOrs Color pedal cards and sign boards. Same witha goe Cent only. odors. alone. EERE FECCEEECECHNE EEE pA CERCA CAAT ER aap Hite aera ea MBG Rees thee Zee h ewes) ~ee nL BRaNe a a ra Bz aba NG's Ane as [i EVs |_| BaAahAN SCE S CE ee Eeea or Rae aN igagaescacesage cela hae Fig. 2. Explanation of Table of Curves. Each of the dots connected by the closed lines represents the percentage of correct first choices in 20 trials; each of the dots connected by the broken lines repre- sents the total number of errors in 20 trials. Where stars (**) are used in place of dots, sets of experiments in which electricity was used are indicated. The heavily shaded horizontal line upon which the figures ¢25” and ‘‘30” stand, is the average chance line mentioned in the text. = a Our results demand two general analyses: (1) An analysis of errors, 7. e., of the number of times that unattached pedals were struck, and of the factors determinative of these inadequate reac- tions; (2) an analysis of the correct and incorrect first choices. It may be well to remind the reader that the pedal which the animal struck first during a given trial is recorded as his first choice for that trial. Hamitton, Unusual Reaction of a Dog. 335 Analysis of errors.—This scarcely need be given in detail, since a large number of errors during a given trial seemed to be due variously to, (1) mere playfulness and inattention, (2) a persistent effort to escape by returning (during the same trial) many times to the pedal or pedal card that had proved successful at the 1m- mediately preceding trial, or (3), when electricity was finally intro- duced, to fear and a wild desire to escape. Analysis of correct and incorrect first choices of pedal_—TYhere was such a varying degree of attention to the sign boards and pedal cards, and the formation of so many misleading associations occurred, that a large number of factors determinative of the first choice of pedal is disclosed by the analysis of results. While it seems an almost hopeless task to attempt a complete and ade- quate classification of these factors, the following will serve, I think, to throw some light upon our problem. (1) As factor “1” I wish to refer to several closely related fac- tors which can be conveniently grouped together. After the animal had been successful with a certain pedal he was apt to return to it and strike it first for many successive trials. At times this seemed to become a kind of habit with him; he would go to his favorite pedal at once on entering the cage, strike it, and then, failing to get out, inspect the apparatus in a very leisurely manner, and make his second choice. A preference for p 3 and p 4 was undoubtedly due to the fact that he is “right handed.” In clawing things from a table, and in all other acts that call for the use of a single forepaw he almost invariably uses the right forepaw. ‘This is due neither to tuition nor to any injury or discoverable deformity, and has,been amply demonstrated by appropriate experiments.’ The following table, supplemented by the analysis that follows, will indicate to what extent the com- plex factor ‘‘1”” accounts for the dog’s first choices: First choice of the same pedal for 2 successive trials......... Pil 7 pa iE spi —18) op a at First choice of the same pedal for 3 successive trials......... Di — Tt p27 ps pa 7 First choice of the same pedal for 4 successive trials......... D2— 2) psi—— 2) pia 1 First choice of the same pedal for 5 successive trials......... iS Pa Ti gai a joe A » First choice of the same pedalfor 6 successive trials......... P3= 1 First choice of the same pedal for 7 successive trials......... plI=I p2= 1 First choice of the same pedal for 13 successive trials.......-.- P4 I First choice of the same pedal for 16 successive trials......... p= i 1See Batpwin, Mental Development, Methods and Processes. Second edition, p. 67. 1903. Also Ernst Weser, Ursachen und Folgen der Rechtshandigkeit. pp 10-13. 1905. It is possible that my dog is exceptional in this respect. 330 ©=fournal of Comparative Neurology and Psychology. Obviously, factor “1”? does not account for all the 266 first choices giveninthistable. From this total we must subtract (a) the 88 first choices that started the 88 series; (b) the 49 correct nee choices that occurred in the course of the series (although it is a fair inference that factor “1”? accounts for many of this latter group, such an inference cannot be drawn with certainty, since the animal often terminated a series by choosing. the attached pedal, only to return to oe favorite pedal of the series thus inter- rupted at the next trial); (c) the 42 first choices that occurred as the second member of a series when the initial first choice of that series was successful (such reactions are tabulated under factor 2”); and (d) the 25 times when the animal might have been fol- lowing up a successful color or odor in making his second succes- sive first choice of a given pedal. To summarize: (1) there are left 266 —88 —49 —42 —25 = 62 first choices that cannot be accounted for on the assumption of a transiently formed association between the desired result and an immediately preceding successful pedal, color, or odor, or to the influence of the directing sign board. ‘These 62 first choices—all of ate incorrect—may ibe ascribed to factor “I.” ) In ror trials the animal made (incorrect) first choice of the So that had just let him out. (3) In 1o2 trials he made (incorrect) first choice of the pedal bearing the color or odor of the immediately preceding attached pedal. This occurred in 83 (25.9 per cent) of the 320 experi- ments with colors alone; in 15 (15 per cent) of the 100 experiments with colors reinforced by odors; and in 4 (20 per cent) of the 20 Ree with odors alone. ) Out of the total 440 experiments in which colors, colors a ee and odors alone were used, he made correct first choice of the attached pedal 122 times (27.7 per cent). It is here that we must look for any light that the experiments may have thrown upon our problem. The question at issue is, “ Did the presence of the directing sign board increase the adequacy of the animal’s reactions to ae situations in which it entered?” Inferences drawn from a mere numerical tabulation of results must, of course, be qualified by the character of these 122 objectively correct first choices. A great deal depends on whether his choice of the attached pedal appeared to be haphazard, * ‘hit-or-miss, or to be attended by such preliminary activities as would have made for HamMILTON, Unusual Reaction of a Dog. 337. greater adequacy of reaction. But the relatively slight influence of the experimenter’s personal equation on tabulations of results, and its inevitably distorting influence on any report of the animal’s apparent intention, must, so far as one kind of certainty 1s con- cerned, give precedence to mathematically derived inferences. First of all, then, we must take into account the fact that had the animal been uninfluenced by any associations, mere chance would have enabled him to make correct first choices in approx- imately 25 per cent of the total number of trials. Or, had he always struck the same pedal first, or the pedal bearing the same color or odor, he would have made 25 per cent correct first choices. Now, the 53 incorrect first choices unaccounted for by any of the above-mentioned factors, and the 62 incorrect first choices ascribed to factor “‘1”’ (the latter, since they often came in series of more than four each) can most conservatively be considered to be due to nothing that impaired the animal’s chances to make the average chance 25 per cent correct first choices. The case is quite different with the 101 incorrect first choices ascribed to factor “2,” where the animal chose the pedal that had just released him; of necessity any choice due to the influence of such an association would be an incorrect one, thus materially low- ering his chance of attaining the average 25 per cent correct first choices. But mere chance would have enabled him to strike first the pedal that was last attached in rio of the trials. This would be a fatal objection to any attempt to demonstrate that the animal was misled by such associations were it not for the fact that the choose-first-the-last-successful- -pedal reactions came, not at irregu- lar intervals, but in definite and prolonged series. ‘Thus, if p 3 had just released him, he would strike it on reéntering the cage, and p I proving to be the successful pedal, the next time he would make an unsuccessful attempt to escape by striking p 1, and so on, until some new determinative of first choice became effective. alae same is true of the 102 incorrect first choices ascribed to factor ce 519 It now becomes clear, I think, that to have made 122 correct first choices the animal must have been influenced by the sign boards; otherwise, the effects of the misleading experience-deter- minants to first choice would have been reflected in a considerably lower than the average chance per cent (25 per cent) of correct first choices. 338 «= fournal of Comparative Neurology and Psychology. Passing now to a consideration of the character of the adequate reactions, it should be said that so far as observation of the indi- vidual trials was concerned, the majority of correct first choices might have been due to mere chance. But I was able to record 23 incidents like the following. The animal inspected the sign board, pedal cards, and pedals, turned his head to look at me, wagging his tail (my position was at a point 20 feet directly in front _of the door), and then, confining his attention exclusively to the sign board, devoted some time to sniffing its surface. Following this he went up and down the row of pedals, sniffing their cards; then pausing at an unattached pedal, he raised his paw as if to strike, but desisted, withdrawing the uplifted paw slowly, with- out having struck the pedal. Again he returned to the sign board and sniffed its surface, following which he passed down the line of pedals, sniffing their cards as he went, until he came to the at- tached pedal (v7. e., the one affording him the same stimuli as the sign board); this he struck and was released. ‘There was not a single failure when he behaved in this manner. | An examination of the correct first choice curve will show that there was but little, if any, improvement at the end of the 440 experiments with sign boards. This agrees with the fact that the kind of behavior just described never became a fixed mode of reaction. Usually, after he had once reacted so adequately, he quickly chose the previously successful color or odor, thus falling into error. It is interesting that the odor reinforcement was of so little value to him. That particular series of experiments would have been continued much longer had it not been for the fact that the animal finally learned a very effective and easy manner of finding the attached pedal. Starting at the third or fourth pedal to the nght, he would pass to the left, striking the pedals in succession until the shock of the discharged trigger sent him scurrying to the door. If he started at p 3, and p 4 was attached, he would go down the line to p 1 in the usual manner, then walk directly to p 4 and strike it. Reference to the curves shows that in the third from the last set of 20 experiments he made 29 errors, and struck the attached pedal first in 5 of the 20 trials. “That he made only 29 instead of the “average chance” 30 errors, is due to the fact that in switch- ing from the method of starting at p 4 (a method which he pur- sued for 13 successive trials), to the method of starting at p 3, he Hamirton, Unusual Reaction of a Dog. 339 avoided making one error. It was to break up this habit that electricity was introduced as a penalty for errors. It seemed to be effective during the first 20 trials of its use, but after that all previously acquired fondness for the experiments was replaced by fear, and I found myself working with. an animal that reacted like THORNDIKE’S hungry cats and dogs.’ Lioyp Morcan defines a psychological process as “the middle term between results of complex stimuli from the environment on the one hand, and the results of complex reactions to that envi- ronment on the other hand.’ Instead of stating our problems in the interests of hypothetical interrelations of these “middle term”’ processes, and instead of making our experimental and clinical observations subservient to problems so stated, it seems desirable to pursue a method of studying animal behavior which will keep us more closely in contact with the facts accessible to us. Such a method is realized, I believe, in the clinicalt and experimental study of reaction-types. It is true that partially objective methods have been followed in this field so far as the higher vertebrates are conerned, but the divorcement from middle term speculative demands has been more apparent than real. Otherwise, there would be no appeal to “criteria of conciousness; no catering to hypothetical modes of mental elaboration of sense-data; tre in short, no need of psychological inferences, im our interpretations of animal behavior. Animal behavior affords data for the solution of a great and comprehensive problem: Starting with the assumption that from the lowest forms of life to human life, there is an ever increasing adequacy of adjustment to complex environments, and that the adequacy (in the sense of complexity) of adjustment implies a cor- responding complexity of effective inner elaboration intervening between reception of stimuli and reaction to them, the general > Tuornpike, E. L., Animal Intelligence. Psychological Review Monograph supplement, vol. 2, no. 4, Pp. 30-32. 1898. 3 Liroyp Morean. Comparative and Genetic Psychology. Psychological Review, vol. 12, Pp. 79- 1905. 4 For the lack of a better term ‘‘clinical” is used here to indicate the kind of observation that is employed by the psychiatrist in his studies of insane patients. It is generally recognized by clinical psychiatrists that the academically trained psychologist (if he lack adequate clinical knowledge of insan- ity) is greatly hampered in his experimentation with patients by his lack of clinical checks upon his work. The case is quite analogous where experimental work with animals is not supplemented by prolonged and extensive observation of the subjects,dealt with. ° See Yerkes, Animal Psychology and the Criteria of the Psychic. ‘fournal of Philosophy, Psychol- ogy and Scientific Methods, vol. 2, pp. 141-150, 1905, for a viewpoint which recognizes the value of psychological inferences. 340 ©“fournal of Comparative Neurology and Psychology. problem becomes far simpler, and loses none of its importance if it have to deal only with the possibility of establishing the pres- ence or absence of a continuity of the third terms of the formulz, stimuli—inner elaboration—reaction. Attempts to conduct experi- ments for the sake of gaining knowledge of the intimate work- ings of the “inner elaboration” seem only to retard and unnec- essarily to complicate the problem. If the complexity of a given situation be definitely known, and if there be only one “ mostgade- quate” reaction possible to that situation, these, and not inferred psychic processes will enable us to give the reaction its continuity- position. Of course, other objectively determined factors must enter into consideration, such as the relative influence of instinc- tive equipment and of experience. Returning to an interpretation of the results obtained from the experiments just reported, it is required of us, then, to isolate such reaction-types as appeared there. Of these, two stand out more or less distinctly. (1) During the first 160 experiments, when only the white pedal card was used, the animal showed an increasing adequacy of reaction. Here, amongst several constant factors, the most important experience-determinant (the white pedal card), was the only one that made one situation different from another; but in spite of its changing pedal relations, the white card, in its direct sense-value to nose and eyes, was a simple, concrete, per- sistently recurring experience-determinant. Innumerable obser- vations, both in the laboratory and in everyday life, have shown that dogs’ experience-determined reactions are usually thus con- ditioned, and that final unfailing and perfect adequacy of reaction to such factors requires their frequent recurrence in situations in which the animal is placed. (2) In the last 440 experiments, where the directing sign board was used, the important experl- ence-determinant was constant in principle, but not in direct sense- value. A perfectly adequate reaction to the situation required that, as the result of previous experience with a concretely dis- similar situation, the dog be so influenced by the stimuli derived from the sign board as to strike first and only the pedal bearing a card which afforded him the same kinds of stimuli. A given sit- uation was repeated only once in sixteen times, and no simple associations could be of any positive assistance to him in guiding him to the attached pedal; on the contrary, they only served to mislead him. Hamitton, Unusual Reaction of a Dog. 341 Although such behavior cannot, if my analysis be correct, fall under the head of accidental happenings, and therefore stands out as an actual reaction-type, it did not tend to increase in frequency. Whether the capacity for occasional adjustments of this degree of complexity is a necessary part of the canine endowment, or merely an individual trait, it strongly suggests the possibility of there being still higher reaction-types higher in the phylogenetic scale. The reader will be better oriented as to the nature of the views and problems just expressed if a brief reference be made to the fact that they are due, in part, to certain investigations in the field of psychiatry. From the extremely inadequate reaction-types observed in certain psychoses, to the relatively adequate reaction- types of relatively normal individuals, there is unbroken contin- uity. KRaEpeELin,’ with his keen insight into abnormal reactions, establishes this continuity by means of transitional clinical pic- tures, which lead us from the abnormal to the normal. And in spite of his objections to the radical degeneracy doctrines of Lom- BRoso and the Italian school of positivists, he makes frequent references to the fact that many of the reactions occurring in the “ degenerative psychoses” strongly suggest the reactions of primi- tive peoples. On the assumption of degenerative reversion in certain psychoses, there at once arises the ‘possibility that the ap- parent gap between human and animal adequacy of adjustment may be bridged by means of studies of reaction-types of the highest of the animals below man on the one hand, and of studies of reac- tion-types in savages and the degenerative insane on the other hand. The writer wishes to acknowledge the obligations under which he has been placed by his colleague, Dr. R. C. KELL, who ren- dered much valuable assistance in the construction of apparatus, the drawing of figures, and the actual carrying out of experi- mental work. Kraepetin, Psychiatrie, If Band, siebente Auflage. 1904. THE NORMAL ACTIVITY OF THE WHITE RAR AS DIFFERENT AGES. BY JAMES ROLLIN SLONAKER. (Stanford University, California.) Wiru Ercur Ficures. While observing Dr. Watson’s experiments to test the ability of white rats at different ages to learn new tricks,’ it occurred to me that possibly their greater susceptibility to education at a cer- tain age might be due, to some extent at least, to the fact that they were more active at that age than at any other, and that they were thus able to try more avenues of approach to their food in a given time. With this thought in mind, I planned to test their volun- tary activity at different ages to ascertain how the age of greatest activity compared with that at which they were the most capable of education. Three preliminary experiments were performed in the Neuro- logical laboratory of the University of Chicago, and other experi- ments are now being carried on in the Physiological laboratories of Stanford University. In this connection I wish to thank Dr. Henry H. Donatpson for his ready cooperation and assistance in providing necessary materials for the apparatus and Dr. Watson for his aid in constructing the apparatus. The apparatus employed was very similar to that used by Dr. C. C. SrewarT? in his experiments on different animals. Several improvements have been made which will be described in a later paper. The apparatus consists of two essential parts: first, the revolving 1 Watson, J. B., Animal Education;An Experimental Study on the Psychical Development of the White Rat, Correlated with the Growth of its Nervous System. The University of Chicago Press. 1903. 2 Srewarr, C.C., Variations in Daily Activity Produced by Alcohol and by Changes in Barometric Pressure and Diet, with a description of Recording Methods. Am. Fourn. Physiol., vol. 1, pp. 40-56. 1898. . SLONAKER, Activity of the Rat. 343 cage and its accessory parts for obtaining and transmitting a cer- tain kind of action; second, the recording apparatus composed of a clock which accurately records the number of revolutions made, and the writing lever and kymograph which graphically show the distribution of the activity during the day and night. As each rat is in a separate cage and each cage has its own recording clock and writing lever these records may be directly compared at any time. 3 Revolutions 10000 8000 6000 2000 Days 7 9 10 20 25 Fic. 1. Curves representing the total activity of each rat as indicated by the number of revolutions recorded by the clocks. All the experiments were carried on in a large basement room facing the north and west. Frosted windows furnished a subdued and almost uniform light during the daytime. At night street lights cast in a very dim light, but this was not sufficient to enable a person to avoid coming in contact with obstacles when entering the room. ‘The temperature was fairly constant during the day and varied between 60° and 70° F. At night it was much ccoler but never lower than 40°, even in extreme cold weather. 344 “fournal of Comparative Neurology and Psychology. 2) The rats were fed and watered about 8 a.m. each day. At this time the clocks recording the number of revolutions were read. They were also read late in the afternoon between § and 6 o’clock. In this manner the activity for daytime and night time was easily ascertained. Further details will be described in the discussion of the different experiments. Revolutions 600 500 40u 300 2u0 | 100 ovorr trans teense Days 5 10 15 20 25 Fic.2. Curves of the average number of revolutions of each rat per day based on the data in Table I. EXPERIMENT I. For Experiment I four healthy rats of different ages were selected. ‘Their ages at the beginning of the experiment were as follows: No. 1, 30 days; No. 2, 60 days; No. 3, 71 days; No. 4, 266 days. Rats of different ages were taken in order to ascertain at once whether there was any marked difference in activity due to a discrepancy in age. ‘The experiment lasted 25 days, beginning February 1 and ending February 25. The curves of Figs. 1 and 2 and the tabulations in Tables [| and II show the condensed results of the experiment. While these show a great difference, they cannot be wholly relied upon SLONAKER, Activity of the Rat. 345 as being characteristic for the age which they represent. Indi- vidual variation plays an important role, as will be seen later on. The results, however, can in general be considered as fairly typical for the different ages, as later experiments will show. The curves in Fig. 1 indicate the total activity of each rat as represented by the number of revolutions recorded by the clocks. The curves pass through ordinates erected on the base line at points corresponding, to the fifth, tenth, sixteenth, twenty-first and the twenty-fifth days of the experiment. As can be readily seen the curves of No. 1, No. 2, and No. 4 run fairly close together, while that of No. 3 is very different. The total amount of activity as represented by the number of revolutions is: No. 1, 2224; No. 2, 2040; No. 3, 10,740; No. 4, 1640. TABLE I. Average number of revolutions for a period of 24 hours at different ages. | Number | Number None Number Number | - of Reyo- . |of Revo- . of Revo- -_ of Revo- . of Revo- et | Age in Age in Pao Age in Age in : Age in 3 AT. | | lutions. ? lutions. lutions. lutions. | lutions. | Days | Days Days Days Days | Average : Average Average Average || Average lof 5 days || |of 5 days of 6 days| of 5 days) of4adays 1} 35 86 | 40 87 || 46 | 86 Bins regs ul Mesh, nme II 65 20 || 70 105 76 104 81 75 85 103 Ill 76 Gs) | Sie ee 7 87 609 g2 503 96 | 484 IV 271 80 | 276 (Sis |i ese 84 28s Or 291 28 greatest activity. If this were true, No. 2 would have been a close second, as it was only eleven days younger. It only shows the great individual variation which occurs and must be considered. Table I is a tabulation of the average number of revolutions per day on the same days of the experiment as Fig. 1. These tabula- tions are plotted as curves in Fig. 2. From these it would appear that No. 3 reached its greatest average daily activity on the six- teenth day of the experiment. It was then 87 days old. After thatage the number of daily revolutions steadily decreased. Nos. 1 and 2 increased slightly. No. 4 shows a gradual decrease from an average of 80 revolutions on the fifth day to 28 on the twenty- fifth day of the experiment. One should not infer that the curves would continue to rise or decline as the case may be had the experi- ment continued. 346 Fournal of Comparative Neurology and Psychology. TABLE II. Representing four single days activity at the ages indicated. | | Number | Total | Number | Longest Shortest | Longest Shortest ide Age in | Revolu- | Number | Periods — of es Period of | Period of Days. tions Revolu- daily | Activity Activity | Rest in Rest in daily. | tions. | activity. in | in minutes. minutes. | | minutes. | minutes. I 31 72 130 9 12 | 5 180 10 II | 61 Sistas 24 10 120 Le 300 fe) we | 72 28 | 48 8 20 a 390 20 IV | 267 236 256 fe) 12 I | 300 20 I 36 rigee || 660 6 20 I 300 7° II 66 igo | 250 8 40 I 305 15 Tl 77 3640 |)” Gaz 13 20 I 240 30 IV 272 nes al 516 14 fe) I 130 30 It 48 TA ies | Age: iB 50 10 480 120 II 78 S2 el sgy 8 7° | ait 300 15 III 89 338 | 6830 13 7o 2. 340 15 IV 284 52 | 1244 7 60 ie 300 go I 56 40 2224 4 15 2. 700 180 II 86 110 2040 | 7 20 Se 360 60 Ill 97 294 | 10740 | ed 100 Tis 300 30 I\\/ 292 16 | 1640 | 10 55 ie 190 8o The tabulation in Table II is intended to show the work, the number of periods of rest and activity, etc., during four of the days of the experiment. ‘The table is self-explanatory and needs no further comment. From this it is observed that the rats become more regular in their activity toward the end of the experiment. The daily records show this much more prominently than 1s seen in this tabulation. The reason for this is no doubt due to their becoming accustomed to the environment. This experiment shows that the very young rat is more active than the old one and that somewhere between these two extremes the period of greatest activity is to be found. According to these results the period of greatest activity for No. 3 was at the age of 87 days. ‘The period of greatest activity for Nos. 1 and 2 had not yet been reached. No. 4 was decidedly on the decline from the very beginning of the experiment. ‘The marked discrepancy between No. 2 and No. 3, which are so nearly the same age, is certainly due to individual variation. SLONAKER, Activity of the Rat. 347 EXPERIMENT II. On the twenty-fifth of February the rats of Experiment I were replaced in stationary cages where they remained for fifteen days, At the end of this time their ages were 70, 100, 111 and 306 days, respectively. They were then returned to the revolving cages from which they were taken. ‘This was done primarily to give a Revolutions Days 5 10 Nl SIC ne SOS ONE 4045) mn 30 57 Fic. 3. Curves of total activity of each rat at different ages as represented by the recording clocks for a period of 57 days. The age of each rat at the end of this time was: No. 1, 127 days; No. 2, 157 days; No. 3, 168 days; and No. 4, 363 days. different series of ages from those of Experiment I and incidentally to test their memory. In the former experiment each rat had learned a way peculiar to itself of reaching the food box, of getting down to the wheel and of entering the nest box. 348 = Fournal of Comparative Neurology and Psychology. After an interval of fifteen days it was interesting to note that when they were first returned to the revolving cages they appeared completely at home and within two minutes each had sought the entrance to the nest box in the same manner as in Experiment I. In going to and from the food box each showed its peculiar tricks. This showed conclusively that the duration of memory for these Revolutions 7000 6000 5000 4000 5000 2000 1000 A OIDD Days 5 10 15 20 29°55 501) 795 40 45 50 55 57 Fic. 4. Curves of daily activity of each rat for the same period as in Fig.3. For simplicity the average of each five days is taken. rats was at least fifteen days. Since the experiment was not intended to follow along this line no further tests were made. In regard to their activity there was a very noticeable change. No. 1 was exceedingly active, while No. 4 was very slow and showed decidedly the effects of advanced age. No. 2 and No. 3 showed a degree of activity approaching that of No. 4. The marked differences in activity are best seen in the curves SLONAKER, Activity of the Rat. 349 of Figs. 3 and 4. Fig. 3 represents the curves of total activity from the beginning of this experiment, March 13, to the end, May g, a period of 57 days. Fig. 4 represents the curves of daily activity. For simplicity the curves pass through the ordinates erected on each fifth day of the experiment, eiche representing the average of the number of revolutions for the preceding five days. By comparing the curves of Fig. 3 and those of Fig. 1 4 marked difference is noticed. No. 3 was the most active in Experiment f, but now it occupies third place. No. 1, whose position was second in the former experiment, now far surpasses all the others. When we compare the ages of these individuals at which they were the most active we find that No. 3 did the most work when it was between 85 and 95 days old, No. 1 (Fig. 4) at the age of 110 to 127 days, No. 2 somewhere between 100 and 120 days, while No. 4 never seemed to show much change from day to day. In other words, No. 4 had passed the age of greatest activity before the beginning of Experiment I. Individual variation no doubt accounts for these differences in the age of greatest activity. If the average were taken it would bring the most active period as indicated by greatest number of revolutions, at about theage of 105 days. Taking the data we now have, a hypothetical curve representing the curve of activity from birth to death from old age could be constructed. Such a curve would show a gradual increase in activity to about the age of 100 days after which it would begin to fall and would finally reach the base line at death.’ But owing to the individual variation manifested in these experiments, such a curve could not be relied upon as being correct. In order to construct such a curve Experiment III was begun. EXPERIMENT III. The foregoing experiments show the need of a curve represent- ing the average of a number of individuals subjected to the same conditions. ‘That this might prove successful the rats should be the same age, as closely related as possible, and all subjected to the same food and environment. Accordingly, a litter of eight rats was selected. At the age of 25 days they were practically 3 The age which rats living in the above conditions would attain has not yet been ascertained. 350 ©‘fournal of Comparative Neurology and Psychology. uniform as to size and general appearance. ‘Their weights, how- ever, varied somewhat as can be seen by consulting Table III. Owing to the fact that the young cannot very successfully be weaned beforé 25 days, the early activity could not be definitely ascertained. But from observation of the young previous to this age one perceives that they move very little except when they begin to crawl out of the nest between the fifteenth and the twentieth day. The amount of activity therefore previous to the beginning of this experiment is very little. At the age of 25 days these eight rats were placed in eight separate cages. Four of the cages were the revolving ones used in the former experiments. The other four cages were of the ordinary stationary type in which rats are usually reared for laboratory purposes. ‘This arrangement was made in order to see what effect, if any, the voluntary exercise of those in the revoly- ing cages might have on the rate of growth and the longevity. 23.2] 64.5 41. TABLE III. Weights of rats from the same litter at different ages, showing sex and gain of weight of each. The averages of those in the stationary cages are for the two females only. = aj as 8 ale i, \h oun eel 2 Seal Sr allpiales Cee ltee eS (oe Goa lpg ates e gh gl S/R eo [Be] 3 (Gs) SEES Rat Sex, |= 0|.8 0 = (2) ce) a Borg ee 48) cs 3 js GS) A Cat eal a all es 2 NO. st +s oe Sey el 2 ws) ow Sores) ele el) se ihe] a leo] |e sl s 9 | D C 3 o a vo BalB<| 6 /F 2] 6 |B 2| 6 |B 2| 5 | <| S —_——_———_ = ee i: | | bo I | Female | 24.1] 58 33-9 86 28 93 Tal el23 (ae 3Ouy 46 13 ee g, II | Female 23-5, Roots) i) | ev 27-5 93 g | 113 | 20 |132 19 PO III | Female 25-5) 64.5) 39 | 85 20.5 104 | 19 | 121 | 17 146 | 25 a IV | Female| 22.3] 55.8 | 33.5 86 aoe || Gis) UP Ze bey ey hes) ae Ave age 23.8] 58.9 34.8 85.2] 26.5 95 9.7) 117. 5)|| 22a54Or Gil 2Onm bp V | Male 26.3, 76.3 50 |L14 Plein] | 13, |163 36 191 28 5 e. VI | Female|-22 |-58 | 36 | 8 |27 | 9¢° | to |xz2 | a7 jrzqg | 32 a 5 VII | Male 23-8 73 | 49.2105 Bee UTS) etige | mon 43 188 =| 27 n VIII | Female! 24.7| 71 46.3 93 ZI TOs 12 rg 18 |138 15 Ave -age 1 89 245 TOO. i Ti | ELT sal) Dsig uae Tes For convenience these rats were numbered from one to eight. Nos. 1, 2, 3 and 4 were placed in the revolving cages and 5, 6, 7 SLONAKER, Activity of the Rat. 351 and 8 in the stationary cages. [he experiment was started May g and stopped July 8, extending over a period of 60 days.‘ he rats were all fed the same kind and approximately the same amount of food each day. ‘The food given to each was not weighed but was measured in a fairly accurate manner. Any slight variation in amount would have no effect because more food was given than they could eat during the time elapsing before they were fed again. ‘They were fed about 8 o’clock each morn- ing. This consisted in washing the drinking cups, filling with clean water, removing scraps of food not eaten and putting in fresh. An abundance of cracked corn was always provided so that food was always at hand. In order to determine the rate of growth each rat was weighed at certain intervals during the experiment. ‘lable III indicates the sex, the weights at different ages and the gains. It is readily seen that No. 5 and No. 7 soon surpass the others in weight. It is also noticed that No. 5 is the heavier of the two at the first weighing and that this relationship obtains throughout the experi- ment. ‘These two were males, the other six being females. ‘The females run fairly close in their weights. But here again it 1s seen that those that were the heaviest at 25 days of age are among the heaviest when the final weights were taken at the age of 85 days. It thus appears that the start which they get while nursing is main- tained. Since the males soon surpass the females in weight, they were not taken into consideration in computing the averages. The average of those in the stationary cages is therefore an average of only two. In the revolving cages it is an average of four. Table PED shows some very interesting results. In the first place, the average weight of those in the revolving cages is greater than the average of the females in the stationary cages both at the beginning and at the end of the experiment. However, at the ages of 45, 56 and 61 days those in the stationary cages surpass the others. At the age of 73 days they are the same. Just what this means one can not say. It is probable, however, that the averages of those in the stationary cages are based on too few individuals and for that reason can not be relied upon. The *Tt was the intention that this experiment should continue throughout the normal life of the rats, but owing to my change of residence to California it had to be terminated prematurely. 352 ‘Fournal of Comparative Neurology and Psychology. differences are so slight that no conclusions regarding the advan- tage or disadvantage in growth can be reached. Experiments now in progress will be able to determine this question. Owing to the early termination of the experiment, nothing can be said regarding the effect on their longevity. When these rats were first weaned and placed in revolving cages their graphic records showed that their activity was more or less distributed over the entire 24 hours of each day. With the exception of the feeding time, when all were active, there was no regularity in their activity. This is easily perceived by consult- ing the graphic record in Fig. 5 which represents 24 hours’ activity at the age of 36 days. The interrupted line below indicates the hours as marked by the electric clock. The line representing the activity of each rat is indicated by the appropriate number at the left. Where the record is a straight line the cage was stationary 8am 9 10 11 12 1 PN. 2 3 4 5 6 7 8 9 10 It 12 1AM. 2 3 4 5 6 7 Fic. 5. The graphic record of the activity of each rat at the age of 36 days for a period of 24 hours as recorded on the kymograph paper. The broken lines indicate periods of activity, the straight lines periods of rest. and the rat was to all appearances resting. Occasionally the interval between revolutions was so great that the individual turns can be made out in the records. But since the kymograph paper moved very slowly—about four inches per hour—the records more often appear as a solid band with occasional interruptions of rest. This record shows the irregularity in the periods of activity. It also shows that the rat is resting much more of the time than it is active. In using the term resting I do not wish to be interpreted as meaning that the rat was asleep. As a matter of fact immediately after feeding they begin at once to carry their food into the nest boxes. This accounts for the regular and general activity at 8 a.m. They no doubt eat this food as they desire during the day. Since the water cannot be carried inside, they are forced to come SLONAKER, Activity of the Rat. 353 out to the water boxes. This I think accounts for the occasional short runs of often only a few revolutions of the cage. As the rats grow older, they become more and more like their wild gray relatives in that they are almost wholly nocturnal in their habits. They also become much more regular in their periods of activity and rest. With the exception of a short period of activity at the feeding time (8 a.m.) and an occasional short run at other times during the day, they spend the entire day time within their nest boxes. They are presumably asleep during the greater part of this interval. June 25 CA eee re I i Py a EE ST eS Lf | aren? = eras a Ea | ee T T T T T I ss, Ts a T T Sa.m. 9 10 at 12 IP.M. 2 3 4 5 6 7 8 9 10 a 12 LAM. 2 3 4 5 4] June 26 i a os so ft) sss 2-1 0 1h | Mi \Gumrempers Set || ES tT June 27 Fic. 6. The graphic records representing the activity of each rat at the ages of 70, 71 and 72 days as traced on the kymograph paper. The broken lines indicate periods of activity, the straight lines periods of rest. The time in which they did the great bulk of their work varies somewhat. In the winter time when the days were short they began their running at an earlier hour. As the seasons changed and the days became longer the time when they began their ac- tivity gradually shifted to a later hour. In every case this time appeared to coincide, in a general way, with the first deep shadows of night. Earlier darkness caused by a storm or cloudy sky did 354 fournal of Comparative Neurology and Psychology. not seem to cause any marked advance in the time of beginning their activity. In other words, the periods of activity seem quite regular and are not materially affected by incidental circum- stances. In this experiment the period of activity extended between the hours of 8 p.m. and 4 a.m.—a period of eight hours. The remainder of the 24 hours, with the exceptions noted above, was devoted to rest. This is beautifully shown in Fig. 6 which represents the graphic records for threesuccessive days, extending from 8 a. m., June 24, to 8 a. m., June 27. The ages of the rats during this iain ee 70, 71 and 72 days. By comparing these records with those of Fig. 5 one is impressed with the marked regularity of activity at this “later age. They began their work with a noticeable degree of punctuality. Rat No. 4 was especially prompt and usually began within a few minutes of 8 o’clock each evening. {maim 4 io a Ee NE Se Se ee een — |) — eee + +t a a 8a.M. 9 10 ir 2 TPM. 2 3 4 5 6 7 8 9 io Ie 12 IAM. 2 3 4 5 6 7 8 Fic. 7. The graphic records of the activity of each rat for a period of 24 hours at the age of 77 days, showing the effect of continuous light. The electric lights were left burning during the entire night. To what is this nocturnal activity due? Is it due to the absence of light or to some other cause? In order to see if light had any effect on their activity the electric lights were left burning during the night of July 3. The graphic record for that period is shown in Fig. 7. A great difference is observed between this record and those of Fig. 6. The lack of regularity and the great reduction in the amount of activity are very noticeable. ‘The marked difference in activity is also seen in Table IV, which illustrates the average number of revolutions for day and night. On the nights of July 3 and July 6 the lights were left burning and the readings the next morning at 8 o'clock are indicated. I| can give only two sugges- tions for the difference between these two nights’ work. Enon midnight on during the night of July 3-4 there was an almost continual roar and racket from the explosion of giant firecrackers, SLONAKER, Activity of the Rat. 355 cannon, etc., in celebrating independence day. ‘This may have had some effect in retarding their activity. Or on the second night the rats may have become somewhat accustomed to the light from their former experience and were reasonably active in Spite of it. From this only one conclusion can be drawn; that is, light does seem to have an influence on the time of the rat’s activity. The wild gray rat has by natural selection become accustomed to seek its food at night to escape its enemies. Its various organs have been modified to fit it for this nocturnal habit. The most promi- nent organ that has been changed is the eye. ‘The very large pupil and the great predominance of rods in the retina fit it especially for perceiving objects in dim and diffuse light. Bright light would not only be blinding but very probably painful. _ In the white rat, owing tothe absence of pigment in the eye, this effect of light would be more marked. This I think is the most instrumental cause for the nocturnal activity in the white rat. TABLE IV. Representing the number of revolutions of the revolving cages during the night time and the day time of a number of days. The lights were left burning during the entire night on July 4 and 7 to determine what effect light would have on the activity. Average Number of Revolutions Average Number of Revolutions Pare during night, 6 p.m. to 8 a.m. |* during day, 8 a.m. to 6 p.m. June 25 2383 42 2 4233 60 2 5870 | 94 2 4085 (?) 2 4175 57 3° 4753 135 July 1 82147 20 2 5313 405 3 7182 51 4 464 9c9 5 6512 332 6 $430 105 ie 2013 ; 133 In ae to show more clearly ave actual difference in activity between day and night work Table IV has been constructed. 356 “Ffournal of Comparative Neurology and Psychology. ‘These figures are the averages of the four rats. Slight fluctua- tions in the averages occur. I can not give the cause of this difference. It certainly cannot be due to their food, for they were fed the same during the time represented in this table. The fluctuations might be due to changes in barometric pressure. | have not as yet demonstrated this. ‘The cause, therefore, cannot at present be given. It is very noticeable that in general when there is a great reduc- tion in the average amount of work during the night the following day shows a marked increase. In other words, they seem to prefer to do about so much work each 24 hours and if this amount is not done during the night they are more active the following day. A better idea of the amount of work may be gotten if the num- ber of revolutions is converted into distance. A little less than 1175 revolutions of each cage are equal to one mile. From this we see that the average nightly run at the age of 70 days is about five miles. The average number of revolutions during the day time at the same age is only about one-tenth of a mile. Many individual cases far surpass this average amount of work. The greatest run which [| have observed for a single night (14 hours and 45 minutes) was 16,516 revolutions, or a distance of fourteen and one-tenth miles. Such a run is usually followed by a notice- able reduction in activity the following day. By close and quiet observation one could see that activity was apparently performed from the mere love of it. The rats would frisk about, jump and play, then start the wheel going and run it for a number of turns without a stop. After a few seconds, or minutes’ rest they would start it again with renewed vigor. ‘This playful attitude was especially noticeable at the age of 50 days. As they grow older this activity which was displayed in frolicking and investigation gradually assumed the form of turning the wheel. So that by the time the rat reached the age of 80 days most of its spontaneous activity was manifested in revolutions of the cage. I have seen them run as many as 100 revolu- tions In a minute without a single stop. ‘This is equivalent to a twelfth of a mile The records in Fig. 6 show that in many cases if periods of rest occurred they were extremely short. In one case (No. 4, last record) there appears to be continuous activity for two and a half hours before a rest. ‘This appearance, I think, is due to the slow movement of the paper. I have never yet seen SLONAKER, Activity of the Rat. 357 a rat run continuously but a few minutes. They may be active, however, for hours with only very short periods of rest. The curves of the total activity are seen in Fig. 8. For the first twenty days after the experiment started there was little or no variation in the curves and they coincided with the average curve. From this time on individual variation appears. A marked difference is thus seen in the total number of revolutions. It 1s Revolutions 140000 120000 100000 80000 60000 40000 20000 Days 45 5O) 55° 56 Fic. 8. Curves representing the total activity of rats from the same litter as indicated by the number of revolutions of the revolving cages. The average curve is shown in heavy line. y also noticed that each curve very closely resembles in general appearance the average curve. Whether these curves would finally have approached each other or would have diverged more had the experiment continued can not be predicted. The following table (Table V) gives the average daily activity throughout the experiment. The average is computed for each 358 “fournal of Comparative Neurology and Psychology. fifth day and is based on the records of the preceding five days. Some very interesting things are brought out in this table. At first the environment was new to the rats and everything was strange. No attempt was made to show them the way to the food nor to the nest box. As a result they did much more work at first than later when they had become accustomed to their surroundings. This accounts for the apparent decrease in activity from an average of 121 revolutions on the fifth day to 26 on the fifteenth day. From this time on there is an almost constant increase in the daily activity. It is especially noticeable toward the termination of the experiment. TABLE V. Average number of daily revolutions. May May May | May | June | June | June June | June | June July | July Date. 15. 20. 2b 30. 4: 9. 14. 19. 24. 29. 4. 8. Agein Days 30 35 | 40 45 50 55 60 65 7o 75 80 84 — | | Average Number Daily 121 58 26 | 62 52 242 | 886 | 1065 | 2111 | 4239 | 5313 | 5917 Revolu- tions This experiment coincides so far as it goes with the results obtained in the former experiments. It shows that the period of greatest activity, as measured by the number of revolutions, for these four rats has not been reached at the age of 84 days. Whether the average age of greatest activity would be greater or less than that determined by Experiments I and II can not be determined from these results. It could only have been ascer- tained by continuing the experiment. ‘The experiment now under way should make this clear. CONCLUSIONS, 1. White rats of different ages show a marked difference in in their activity. 2. The very young rat and the old rat are each noticeably SLONAKER, Activity of the Rat. | 359 inactive. Somewhere between these two extremes we find the age of greatest activity. 3. These experiments indicate that the age of greatest activity, as represented by the number of revolutions of the revolving cages, ranges between 87 and 120 days. 4. Owing to the marked individuality exhibited, a correct curve of activity from birth to death cannot be constructed from the results of one rat or of several individuals of different ages. 5. A curve representing the normal activity from birth to death due to senility must be based on the records of a number of individuals of the same age and subjected to the same con- ditions. 6. The white rat is affected by light. ‘This, I think, is mainly due to the structure of the eye and to certain tendencies which they inherited from their wild ancestors. 7. From these preliminary experiments no correlation can be made between the age at which they are most active and the age at which they learn most rapidly. EDITORIAL. CONCILIUM BIBLIOGRAPHICUM. The revision of the Anatomical Bibliography of the Concilium Bibliographicum at Zurich, as announced in their latest publica- tion, is a matter of importance to anatomists in general and to neurologists in particular. Very few anatomists in America are so situated as to be able to search out the literature even within their own specialties; fewer still are fond of such work even when facilities for it are amply provided. And yet failure to study the literature of his subject, especially the current literature, from the whole world is fatal to thoroughly broad work in any field of mor- phology. The path of the neurologist is peculiarly difficult, for his sub- ject is not only the most intricately complex of all of the morpho- logical specialties, but any detail of an inquiry may develop the most unexpected relations with remote parts of the field and even with far distant departmients of inquiry. While, therefore, the neurologist is more in need of bibliographic assistance than are other morphologists, the problem of producing a servicable bib- liography 1s immensely more difficult. In the past the writer has found the anatomical bibliography of the Concilium Bibliographicum of great service in spite of its con- spicuous defects. The Concilium, though founded and main- tained by a morphologist, has hitherto been compelled by the exigencies of its organization to devote its first energies to other departments and anatomy has remained frankly neglected. ‘The zoological, physiological and other bibliographies have now so far progressed that the Concilium is able to take up the anatom- ical bibliography in earnest. ‘The revised Anatomical Conspec- tus has been elaborated after nearly ten years of practical experi- ence with the scheme on a more modest scale, and in our opinion this document should receive the careful attention of every work- ing anatomist. Editorial. 361 We have used the Concilium Bibliographicum scheme of classi- fication, not only for bibliographies, but also for many other kinds of cataloguing (such as microscopic slides, etc.) and find it as simple and easy of application as any practicable scheme could be expected to be. Some of the details of the classification are in our opinion unfortunate, but the principle on which it 1s based 1s sound and is the only practicable principle for a card bibliography. And the card bibliography is the practicable one for a working anato- mist. This does not imply that it will run itself. Any scheme of anatomical classification which is comprehensive enough to have more than local usefulness must be complicated. At least this is true of the neurological sections, where unless a minute sub- division of topics is employed the vast number of diverse entries will very soon become a useless mass of inaccessible material. The complication is inherent in the subject matter. ‘This means that intelligent and continuous attention must be given to the bibliography from the start or it will break down in use, no matter how perfect may be the system. If only each investigator could take the time to elaborate his own bibliography the problem would be much simpler, but each system so devised 1s likely to be useful to its inventor alone. If in the course of the further division of labor in scientific work a general bibliographic scheme is to be adopted, the individual must subordinate his own point of view somewhat. In spite of its defects, this bibliography 1s practicable. Moreover it is sufficiently flexible to enable each individual to adapt it to his own needs and still have the enormous advantage of the prompt tssue of titles already printed on cards, with the privi- lege of subscription to such topics only as interest him. It enables one for a nominal sum to purchase information which every inves- tigator needs, but will not and usually cannot procure for himself, and to keep it in available form. Inquiry usually develops the fact that those who find the system impracticable have never spent nalf an hour in serious study of the theory and detail of its organ- ization. The writer has kept his neurological bibliography on library cards catalogued by the Dewey numbers -for about a decade. After subscribing to the neurological series from the Concilium (the cost of which has been trifling) these cards have been slipped 362 ‘fournal of Comparative Neurology and Psychology. into their places as they arrived and now completely overshadow the original list, into which they were incorporated without dis- turbance of the arrangement. And my complete bibliographic list on any subject is together, so that if I wish to assemble all of my available titles on, say the facial nerve, instead of searching through many volumes of periodically published lists, I simply pick out the cards between two guides and the whole list is before me. The same series of cards may also be used as a catalogue of one’s own library without disturbing their position in the file. By marking one of the upper corners of every card which bears a title which is represented in the library with a colored wafer or a rubber stamp, inspection of the card instantly tells whether the article is in the library and the shelf number may be added, if necessary. The Concilium, as is well known, is not a commercial enter- prise; nor is it necessarily a rival of the various other bibliographies now serving the scientific public. ‘The monthly and annual lists are valuable in their way, but in our opinion cannot replace the card catalogue. And now that the anatomical bibliography of the Concilium has been thoroughly established and enlarged, we bespeak for it the hearty support of anatomical laboratories and libraries and remind our readers that standing orders for all cards on individual topics can be placed. For instance, the neurological cards alone will be sent independently of the rest of the bibliog- raphy. But the cost of the whole anatomical bibliography is at present (author and subject catalogue) only about nine dollars per year. This revision differs from the original anatomical classification chiefly by the addition of more subdivisions of the leading topics, thus in no way disturbing the placing of the old cards. ‘There are only two places where the old numbers are changed. ‘The first is a transposition to secure a more logical position of the general introductory divisions. In this case the original numbers occupied by these subjects are left vacant, so that confusion is not likely to occur. The second change involves the substitution in the neurological bibliography for two relatively unimportant entries of a division for Tectonics (including the course of fibers and gray substance) and a division for Localizations. But few old cards are involved in this change and these the Concilium has decided Editorial 363 to reprint. ‘[he change strengthens the weakest point in the old classification, a point where it quite broke down in actual prac- tice, and, with the very great amplification of details throughout the neurological bibliography, has made a great improvement in these sections. In the preparation of the neurological bibliog- raphy the Concilium has had the active and very valuable codpera- tion of Professor v. Monakow. Cuapwet LITERARY. NOVICES. Edinger L. Ueber das Gehirn von Myxine glutinosa. Aus dem Anhang zu den Abhandlungen der Kénigl. Preuss. Akad. der Wissenschaften vom Jahre 1906. Berlin. 1906. 36 pp., 3 plates. The results are based largely upon the new fiber method of BrELscHowsky, which however, EDINGER finds it necessary to modify for Myxine by leaving the heads to be studied much longer in the silver solution—as long as 30 days. The method gives a histological picture which is more like that of certain invertebrates than of other vertebrates, differing even from Petromyzon and Amphioxus. Espe- cially in the forebrain fi thalamus, there is a much smaller number of cells and fibers and the fibers are finer than in any other vertebrate. The tissue between these cellular elements appears in these preparations as a very fine fibrous reticulum. In the discussion of the morphology of the forebrain EDINGER distinguishes two divisions of the cerebral hemispheres, (1) a ventral, the hyposphzrium, which receives the olfactory nerves and contains internally the striatum and nucleus of origin of the tania thalami; and (2) a dorsal, the epispherium, containing the pal- lium. On the basis of his more recent work (and also that of KappERs which was published in January, 1906, in this ‘fournal), he has receded from ‘his former posi- tion on the morphology of the pallium in fishes and adopted the view of C. L. HEr- RICK and STUDNICKA that the fish brain does not consist exclusively of hyposphz- rium but contains at least the beginnings of an episphzrium, though in an abnor- mal position. ‘The obscurity regarding the ventricles of the forebrain of Myxine is at last satisfactorily cleared up by the discovery of actual cavities or definite epithelial vestiges in all of the places where other vertebrates possess forebrain ventricles. The cerebellum is said to be totally absent, a statement which suggests an inter- esting field for inquiry in the exact relations of cells and fibers in the somatic sen- sory centers of the oblongata from which the cerebellum of other vertebrates has quite certainly been derived phylogenetically. The plates include an excellent series of cross-sections through the medulla oblongata, but the author has been able to unravel but few of the fiber complexes and to identify few of the structures, remarking, “‘to cumber the literature with a new description whose basis is no better established than that of earlier authors would be to no purpose. . . . Doubtless it will be possible later, when the fish oblongata is better known, to study more closely and to identify the relations of the fiber tracts shown for the first time in these very accurately drawn figures.” This illustrates very clearly the opinion long held by the reviewer, that the fundamental plan of structure of the nervous system can best be discovered by an exhaustive comparative study of a large number of types in which the systems in question are very diversely developed rather than by exclusive attention to primitive or generalized species. Having discovered the fundamental vertebrate plan of each functional system, by the comparison of forms where it attains maximal and minimal develop- ment, then this schema can be read back with ease into the primitive and unspecial- Literary Notices 365 ized types, which before were incapable of analysis. In other words, the phylogeny should be read backward as well as forward. Only in this way can the study of generalized types yield its best fruits. Ca je-t Van Gehuchten, A. La loi de Waller. Le Névraxe, vol. 7, fasc. 2. 1905. In this lecture, delivered at the University of Utrecht, Professor van GEHUCH- TEN has summarized the recent experimental work on WALLER’S law, and sketched the history of the conflict which has waged about the question of so-called retro- grade degeneration. As is well known, he has studied exhaustively the phenomena of central degeneration, which he finds sometimes to occur, though more tardily than the degeneration of the peripheral portion of the nerve. ‘This is not, however, a cellulipetal process; it is initiated in the cell body. Van GEHUCHTEN would reformulate the law of WALLER thus: When a central or peripheral nerve tract is severed, the peripheral portion always degenerates. “The behavior of the cen- tral portion depends upon the intensity of the reaction of the cells of origin to the lesion. If the lesion is not so severe as to cause the death of these cells, the fibers of the central portion remain intact. Otherwise the atrophy of the cells of origin induces the secondary degeneration of the central portion of the nerve. Van GEHUCHTEN criticises the current conception of nervous degeneration. It is not a process signalizing the death of the nerve. Quite the contrary, it 1s a process of reorganization, a regulatory phenomenon. Eyen the destruction of the axis cylinder and the fragmentation of the myelin take place only in an evniron- ment of living tissue, and the proliferation of the nuclei of the sheath of SCHWANN is distinctly a step in regeneration, which may or may not come to anything, depend- ing on the other conditions of the tissue. Ce. [ee Maxwell, S.S. Chemical Stimulation of the Motor Areas of the Cerebral Hemispheres. Fournal Biol. Chem., vol. 2, no. 3. 1906. We quote the author’s summary: (1) Substances applied to the surface of the cortex either give no indication of stimulation, or do so after so long an interval that they would have time to act osmotically or by diffusion upon the underlying white matter. (2) The white matter of the motor areas can be stimulated chem- ically by the calcium precipitates and by barium chloride in solutions isosmotic with the blood serum. ‘The response to such stimulation is very prompt, occurring within a few seconds at most, after application of the solution. The same sub- stances when applied in the same concentration to the cortex give no result at all or only after an interval of some minutes. (3) Solutions of high concentration can stimulate the white matter by osmotic action very promptly and effectively. (4) When solutions are injected into the gray matter but not so deeply as to reach the white matter no evidence of stimulation is seen. ‘The gray matter is apparently devoid of irritability to chemical and osmotic stimulation as well as to mechanical and electrical stimulation. Grasset, J. Demifous et demiresponsables. Paris, F. Alcan. Pp. 297. 1907. In this work the author takes up the consideration of those individuals who are not insane or mentally deficient in the ordinary sense of these terms. An attempt 3606 =Ffournal of Comparative Neurology and Psychology. is made to show that individuals cannot be classed solely as responsible and irre- sponsible, but that there is a class that has degrees, one may say, of responsibility. The individuals making up this class are mentally ill, and may sooner or later become insane; and they, like the insane, may be cured. The demifous is one characterized by an enfeeblement of the higher mental faculties and an uncon- trolled functional hyperactivity of the lower psyche (p. 130). In opposition to this generalization it is pointed out that these individuals are often mentally bril- liant and have contributed much to literature, science and art, as well as taking a considerable part in politics, etc. The treatment and the legal standing of the individuals must be considered individually. S.. "Tye BOOKS AND PAMPHLETS RECEIVED. Wallenberg, Adolf. Die secundare Acusticusbahn der Taube. Reprinted from Anatom. Anzeiger, Bd. 14, no. 14. 1898. Wallenberg, Adolf. Der Ursprung des Tractus isthmo-striatus (oder bulbo-striatus) der Taube. Reprinted from Neurol. Centralbl.,no.3. 1903. Wallenberg, Adolf. Neue Untersuchungen iiber den Hirnstamm der Taube. Reprinted from Anatom. Anzeiger, Bd. 24, nos. 5, 6, 13, 14. 1903-1904. Wallenberg, Adolf. Die basalen Aeste des Scheidelwandbiindels der Vogel (Rami basales tractus septo-mesencephalici). Reprinted from Anat. Anzeiger, Bd. 28,nos.15 and 16. 1906. Dodds, Gideon S. On the brain of one of the salamanders (Plethodon glutinosus). Reprinted from Univ. of Col. Studies, vol. 4,0. 2. 1907. Dahlgren, Ulric and Sylvester, C. F. The electric organ of the stargazer, Astroscopus (Brevoort). A new form of electric apparatus in an American teleost. Preliminary paper. Anat. Anz., Bd. 29, no. 15. 1906. Hrdlicka, A. Measurements of the cranial fosse. Reprinted from the Proc. U. S. Natl. Museum, vol. 32, pp. 177-232. Washington. 1907. Eycleshymer, Albert C. The habits of Necturus maculosus. Reprinted from Am. Naturalist, vol. 40, no. 470. 1906. Parker, G.H. The influence of light and heat on the movement of the melanophore pigment, espe- cially in lizards. Reprinted from the Fourn. Exp. Zoél., vol. 3, no. 3. 1906. Parker, G. H. and Metcalf, C.R. The reactions of earthworms to salts: A study in protoplasmic stimulation as a basis of interpreting the sense of taste. Reprinted from Am. Fourn. of Physiol., vol. 17, no. 1. 1906. Driesch, Hans. Analytische und kritische Erganzungen zur Lehre von der Autonomie des Lebens- Reprinted from Biol. Centralblatt, Bd. 27, nos. 2 and 3. 1907. Cole, Leon J. An experimental study of the image-forming powers of various types of eyes. Reprinted from Proc. Am. Acad. Arts and Sci., vol. 42,no. 16. 1907. Herms, Wm B. An ecological and experimental study of Sarcophagide with relation to lake beach débris. Reprinted from Fourn. Exp. Zoél., vol. 4,n0.1. 1907. Herms, Wm B. Notes on a Sandusky Bay shrimp, Palemonetes exilipes Stimson. Reprinted from Ohio Naturalist, vol. 7, n0. 4. 1907. Pearl, Raymond. Variation and differentiation in Ceratophyllum. Publ. of the Carnegie Institu- tion of Washington, no. 58. 1907. Mettler, L. Harrison. Clinical physiopathology. The need of a new classification of diseases of the nervous system. Reprinted from Fourn. Am. Med. Assoc., vol. 48, pp. 664-669. 1907. Neurographs. A series of Neurological Studies, Cases and Notes. Edited by William Browning, M.D. Vol.1,no.1. Brooklyn, N.Y. 1907. The Journal of ee perce and eno logy VoitumeE XVII SEPTEMBER, 1907 NUMBER 5 THE HOMING OF ANTS: AN - EXPERIMENTAL STUDY OF ANT BEHAVIOR. BY Ch LURNER: WitH Ptates t-ty, AND ONE FiGureE IN THE Text. CONTENTS. INTRODUCTION. IRC COUT ecserescicae, ais tytn Srisen Ap ic os etrenes acts eee nee ote Bene soern a oesee se aunts yenone ae cles 367 Mistoni@al AR es vit tin cn mys cia actos esates sfoterrconlnleg meena srale ae Ewe mink eidiecoees Seiepouwisserceheeee ss 368 JNA aah ye latte OL ea emnd Aor CE A o.oo tinea SG Does esA Mo ME Oo OO ARID oO D ICO UC Oi Er 369 Ee PERIPUERTMON LS ONG VWI ORES MSI stetsvatavar 3 ant ant ant. col of exp | exp. I | 8 10 12 16 69 Stage with incline 4 on left side. 24 I 2 4 6 100 Do. eh Pe I ° ° 60% Incline 4 on right, inc. B on left. A || 6 ° ° fi Do. revolved the stage through an angle of 180°. 5 1085 ° ° 150f Do. 6 8 | 4 13 20 Do. 7 1260 oi) | ° Do. ; Sar iit | I | zy 4 A vertical gap of 3 mm. between | | base of incline B and island. 9 2880 aes || 9 af Do. but 16 c.p. lamp near incline A, incline B on opposite side. 10 | ° 2 | Br | 8 8 40° ) Do. but 16 c.p. light placed near | | | incline B. II ae i ries ° 60 | Incline B removed, light placed | | where incline B had been. 12 | 1271 ° fo) § | Incline 4 on right, incline B on | | left, 16 c.p. light near B. ie al ° | Avy || | 4 264 | Do. but 16 c.p. light is placed near | | val 14 PeCene |g n heey 16 gI 155** | Do. but placed dark chamber on stage. * Over 100 pupe were used. Eight were carried down incline 4, probably all by the same worker. The remainder were carried down incline B. + All were carried down incline B. t The pupz were all carried down incline 4. ° The workers were, at the beginning, much disturbed. The first three pupe were carried down incline B. By that time a line of workers had begun to pass down incline A. After that a a pupa was carried down incline B. § All of the pupe were carried down incline B. g] At first an occasional worker went down incline 4, while a line of workers was passing down incline B. Gradually the number going down incline 4 increased until there were lines of workers moving down both inclines. ** The majority of the workers began almost at once to store pupe in the dark chamber. One worker soon found the way to the nest with a pupa, and after that made regular trips. It was not until all of the pupe had been removed from the open and placed in the dark room that a line of workers began to convey pupz to the nest. Turner, Homing of Ants. 393 TABLE V. Formica fusca var. subsericea Say. Series I. A B Cc D 13 F G 3 Exp. aa Time Time | Time Time Total Rewnke. ae her of ist of 2d of 3d of ant time ant. ant. ant. col. of exp. exp I I 2 3 3 23 Incline 4 on left side. 2 5 5 5 10 Do. 3 ° 3 4 Do. 4 ° ° ° 13 Do. 5 3 fe) | fo) 27 Do. 6 ° ° | ° 30 Do. 7 240 I I 17* | Incline 4 on the right, incline B | on the left. 8 53 5 10 33 Do. but both stage and incline new. 9 6661 II 13 15 18 65 Incline 4 on the right, no other incline used. 10 I 6 Do. II 5 2 Do. i) 33 5 307 | Incline 4 on the left, incline Bon the right. 13 2400 14 54 Incline 4 on the right, incline B not used. 14 4 16 Do. * Three pupe were carried down incline 4, the rest down incline B. The workers moving down incline B rushed along, the ones that went down incline 4 moved very slowly. + The pupz were all carried down incline B. 394 TABLE SVI. Fournal of Comparative Neurology and Psychology. Formica fusca var. subsericea Say. Series 2. D E Time Time of 2d of 3d ant. ant. 5 7 2 2 42 43 4 35 29 31 3 4 7 2 24 26 2 4 3 = 1%) Time of ant col. | | | | | | G Total time | of exp. 43 28 28 95 69 Remarks. es Incline 4 on the left. Do. 2 stages, one higher than other; incline A leads from 1 to 2, incline B from 2 to the island. Do. Do. Do. but incline C substituted for A and incline A placed to lead from stage 1 to island. * At first the ants stored all of the pupa in the crevice between incline 4 and stage number one. After that had been accomplished, they began carrying them to the nest. + This was conducted at night, by incandescent electric light; all of the other experiments of the series were conducted by daylight. ¢ The pupz were carried up incline C to stage number two, across stage number two to incline B, down incline B to the nest. No pupe were carried down incline 4. Turner, Homing of Ants. 395 TABLE VII. Myrmica puncttventris Rog. Series I. A B c D E Fo pee ie Time Time Time Time | Total v Exp. i) since fast fod ead Hens pone Remarks. no. last ee e = : ant ant ant. col. of exp exp I | GL | TeXe) 122 152 192 | Inc. A onright side. 2 214 8 ia kL Omnarn Band Olean LOS Do. 3 818 2 2m 2 Ey co Do. 4 155 ar 4 | | o | 45* | Incline 4 on the right; incline B | on the left. 5 26 ° Laake 39f | Do. 6 13 I | 6 30t | Incline 4 on the left; incline B on | the right. 7 3142 meal 6 | 7 aig: || Dies 8 25820 § | New stage, inc. A on left; placed | | on new island. 9 1440 | | Same apparatus as in ex. 8. 10 1521 1364** | | 1395 Same stage; incline 4 on right; | dark chamber on top of stage. * One pupa was conveyed down incline B, all the others down incline 4d. One worker ascended incline B to the stage and carried a pupa down incline 4. + One pupa was conveyed down incline B, the remainder down incline Yale t All of the pupe were conveyed down incline B. ° A very large majority of the pupe were carried down incline B to the nest. One pupa was carried down incline 4 to the nest and several were carried down incline 4 and stored under the base of the incline. § Up to bedtime the pupe had not been carried to the nest. Thus the ants were lost for over 12 hours. ** In the course of half an hour all of the pup# had been stored in the dark chamber. There they were left, while the workers, one by one, straggled back to the nest. After the lapse of 21 hours and thirty minutes, the pupz were still in the dark chamber. I then removed the dark chamber and placed workers from the nest on the stage. IV. IMPRESSIONS THAT INFLUENCE HOME-GOING ANTS. I have endeavored to show that ants find their way home by virtue of something which they acquire by experience and retain; in other words, that they acquire from their environment impres- sions which influence their home-going. _[ now propose to examine the nature of these impressions. Most recent students of ants write as though these impressions were composed solely of olfactory elements. One group of writers, represented by Brrue, claims that the home-going of ants 1s the result of olfactory reflexes; another, represented by WasMANN, claims that olfactory percepts are important. Scattered through the literature are passages which indicate that some authors are not fully satisfied with this view. LLuBBocK 396 =“fournal of Comparative Neurology and Psychology. (81, p. 262) showed that, under certain conditions, Lasius niger will move a short distance along an unscented path. WasMANN (or), although believing that the footprints possess an odor-shape which enables ants to tell which way to go, gives several examples of ants going a short distance along paths that have not been scented with their trail. VIeLMEYER (00), in his study of Lepto- thorax unifasciatus, claims that, when near the nest, light and shadows assist the ants in finding their way, and more than twenty years before that LuBBock had stated that “In determining their courses ants are greatly influenced by the direction of the light.” PIERON (’04), although admitting that odors play an important role in the He of ants, was influenced by his discovery that ants sometimes move along paths that have not been scented by trails to conclude that odors play no part in guiding them home. Ac- cording to him, their home-going is controlled by a tactile sense and muscular memory. As far as I can understand Pirron, his hypothesis 1s as follows. On the outgoing trip the muscular movements made induce in the nervous mechanism a condition which enables the ant to return to the nest, in a reverse order, over the same pathway by which it journeyed forth. ‘That tactile impressions are, and that muscular impressions may be, factors in the impression that guides ants harmonizes with my experiments; but that the muscular movements play any such role as is here indicated seems improb- able. P1ERON’s hypothesis implies that an ant returns toits home in practically the path by which it went out. I have conducted a large number of experiments which show conclusively that this is not always the case. 1. I have already mentioned the case of the ant which would drop, with a pupa in its jaws, from the stage to the island; but which, on returning from the nest, would wander about the island until I presented a pair of forceps. It would then mount the for- ceps and rest quietly thereon until I had placed the tip of the for- ceps on the stage. It would then step off and pick up a pupa and take a flying leap from the stage to the nest. 2. I had a specimen of Formica fusca var. subsericea Say, which regularly descended from the stage to the island on the under side of the incline, but on returning to the stage this ant always moved along the upper side. 3. Another specimen regularly descended to the island by way Turner, Homing of Ants. 397 of the incline and returned to the stage up the bottle by which the stage was supported at its center. ‘This it did until half the pupz had been conveyed to the nest. 4. On several different occasions, where I had attached a second incline to a stage from which ants had been conveying pupz down another incline to the nest, I have noticed that an ant would occasionally ascend to the stage by the second incline, pick up a pupa and then pass down the other incline to the nest. During the first part of my experiments I was not looking for data along this line, but towards the close I devoted five hours an evening for two weeks to an investigation of the problem. For- mica fusca var. subsericea served as a subject. I placed a number of pupe and a marked worker on a stage from which an incline led to the island. I found that in each case the ant had to learn the way, not only from the pupz to the incline and down it to the island and thence to the nest, but that it had also to learn the way from the nest to and up the incline. And it usually took a much longer time to learn the way from the nest to the stage than it did to learn the way from the stage to the nest. Nea if PIERON’S hypothesis were true, the muscular memory of the ant should have carried it back to the stage in practically the same path by which it had passed from the stage to the nest. But such was not the case. On leaving the nest, die ant would wander first in one direction and then in another, often returning to the nest. It acted as though hunting for something it could not find. At times it would fail entirely to find its way back. 6. In my experiments on the role light plays in the home-going of ants, I gathered some data on this point. With the light i in a certain position, the marked ant was allowed to learn thoroughly the way to and from the stage to the nest. ‘Then the light was placed on the opposite side of the stage. “The ant was always much disturbed by the change and it always took it a long time to find the way to the nest. But, having reached the nest, if PrIERON’s hypothesis be correct, its muscular memory should have guided it immediately back along the path by which it reached the nest. The ant, however, always hada hard time finding the way back to the stage and often it failed completely. Experiments with Odors.—Having observed that ants can find their way about without eyes and without antenna, PIERON con- cludes that odors play no part in the home-going of ants; but, as 398 “Ffournal of Comparative Neurology and Psychology I have already stated, this no more proves that normal ants do not utilize odors in their home-going than does the fact that blind men are able to find their way about demonstrate that normal men do not use visual sensations in their journeys. Most writers lay great stress upon the odor of the trail, and FIeELDE claims that a special organ (the ninth joint of the flagel- lum) exists for the detection of the track-odor. If I interpret my experiments aright, it is not the scent of the trail merely, but the odor-peculiarities of the pathway as such, that form a part of the impression which enables ants to pass home. ‘The readiness with which ants react to any strange odor that 1s added to their pathway lends support to the view that the odors of the pathway itself form a part of the psychic impression of home-going ants. PIERON himself says that ants traveling in a common road are arrested by unexpected odors; they flee from odors of their enemies and cross with little difficulty scents of vegetable origin. “To my mind, this statement of PIERON’s supports the view just presented. To test this matter in the laboratory, several experiments like the following were tried. A colony of Formica fusca var. subser- icea Say was divided and each nest placed in the same relative position on different Lussock islands. On each island was placed a cardboard stage from which an incline led to the island. Each of these inclines led from the same relative position on the stage and in the same direction. On each stage was placed a large number of pupz and a marked worker. After each worker had thoroughly learned the way to and from the stage to the nest, the experiment proper was performed. ‘The ant that was carrying pupz from stage number one was the subject of the experiment proper. The ant on stage number two being used, in the manner hereafter mentioned, as a control. Special inclines were made by placing across the middle of the white slip used for ordinary inclines a transverse band, three-fourths of an inch wide, of some odoriferous substance. “The chemicals used for this purpose were xylol, oil of cedar and oil of cloves. When the ant on stage number one had become so well acquainted with the way home as not to be disturbed by the substitution of a new incline of the same kind for the old, the old incline was replaced by one of these special inclines. In each case the ant was very much disturbed, but in the case of both the xylol and cedar oil, after a while, the ant began to go back and forth across the band TurNER, Homing of Ants. 399 conveying pup to the nest. In a short time after the first few trips had been made across the band, the ant would be making its journeys to and from the nest as rapidly and regularly as down the old incline. ‘Then a new special incline or a plain incline of the ordinary kind was substituted for the one just used and that one transferred to the control stage. In each case the animal used for control was much disturbed, which demonstrated that, from the ant’s standpoint, the transverse band of volatile substance was still on the incline. Evidently that odor or better that volatile substance had become for ant number one a familiar element of the homeward path. ‘This experiment then shows that the vola- tile chemical peculiarities of the path form a portion of the impres- sion experienced by ants on their journeys (Figs. g, 18). Experiments with Tactile Stimuli.—Vhat tremors affect the home-going of ants is evidenced by the fact that a comparatively slight tap on the stage will cause an ant to halt and a severe jar will so disturb it as often to make it necessary for the ants to re- learn the way home (Fig. 12, D). A number of experiments were performed for the purpose of discovering whether ants are affected by the surface character of the pathway. In this case a worker was trained to go down and up a smooth black incline. After it had reached the condition where it was not disturbed by the substitution of another smooth black incline for the old one, a black incline with a velvety sur- face was substituted for the smooth black one. Myrmica punc- tiventris Rog. was much disturbed by the change (Fig. 10, g), but Formica fusca was not (Fig. 18, M; 14, P). It seems legitimate to assume that tremors and jars probably give kinesthetic stimuli and that the velvety surface gives a tactile stimulus. Such being the case, kinesthetic impressions probably form part of the mental furniture of all ants examined by me and at least some ants have tactile impressions. Ex periments with Optic Stimul1.—To see if optical 1 impressions are received by ants two kinds of tests were conducted: experiments with ants working in concert, and with marked individuals. In experiments of the first type a cardboard stage from which a cardboard incline led to the island was used. A 16 c.p. incan- descent electric lamp was placed, sometimes near the side to which the inclined plane was attached, and sometimes near the opposite side. After the ants had thoroughly learned the way 400 ‘fournal of Comparative Neurology and Psychology. home, a new incline was attached to the side of the stage which was opposite the one to which the old incline was attached. If after a lapse of five minutes no ants went down this second incline, conditions were considered right for the test. The light was now transferred to the opposite side of the stage. In each case the halting movements of the ants showed that they were disturbed (Fig. 6). In most cases, some of the ants would finally go down the new incline and in a few cases, after the lapse of several min- utes, all of them would go down the new incline. ‘These experi- ments were tried on all of the ants used in the experiments on heliotropism (p. 371). Similar experiments were conducted with marked individuals of the species, Formica fusca var. subsericea Say, Myrmica punctiventris Rog. ‘The results harmonized with those derived from experiments conducted with ants working in concert (Fig. 14, B; Fig. 17, B, H, M); only, in almost every case, after a greater or less lapse of time, the ant would usually find its way down the old incline to the nest; and after a still greater lapse of time find its way back to the stage. Fig. 16 illustrates how ants are disturbed by altering the position of the light better than a multitude of words. From the beginning to B the ants had been working by daylight, the light coming from two windows, one on the south and one on the west. From B to C the ants were working at night, the illumination being furnished by electric lights and coming from the northeast. From C to the end, the condinere were the same as from 4 to B. To determine whether this effect was due to the intensity of the light, to the direction of the rays or to heat, the following series of experiments were conducted with Formica fusca var. subsericea Say. That the effect described above was not due to heat, was proved in the following manner. A cardboard stage was arranged with its left side connected to the table by a cardboard incline. At the right and left of this stage was placed a heat-filter, consisting of a tall museum jar 34 cm. x16 cm. x 7 cm., filled with cold distilled water. At the beginning of the experiment, a 32 c.p. incandes- cent lamp was placed behind one of these filters (usually behind the one near the incline). After the ants had traveled the path long enough to make the trips regularly and rapidly, the lamp was transferred to a point behind the opposite filter. In every case Turner, Homing of Ants. 401 the workers were much disturbed in the manner stated above. Since the heat had been excluded, it is evident that the disturbance was the result of some form of light stimulus. ‘This was repeated with seven different colonies of Formica fusca. To determine whether the intensity or the direction of light was the determining factor, a different kind of experiment was used. The stage and incline were arranged as before. Sometimes the heat-filters were used, but more often they were not. ‘To furnish illumination, four different candle powers (4, 8, 16 and 32) of incan- descent electric lamps were used, one at a time, in a darkened room. At the beginning of the experiment a lamp of a certain candle power was placed near the side of the stage to which the incline was attached. After the ants had thoroughly learned the way home, a different candle power was substituted for the first. After the lapse of a few more minutes this lamp was transferred to the opposite side of the stage. Shortly it was returned to its for- mer position. A few minutes later a different candle power was substituted for this. ‘This performance was repeated over and over again until each lamp had thus been used one or more times. It was found (Fig. 17) that substituting a lamp of one candle power for one of a different power had no disturbing effect on the actions of the ants; but that by any marked change in the angular position of the light, no matter what the candle power, the ants were very much disturbed. ‘This warrants the conclusion that when light is present the direction of its rays plays a prominent role in the home- going of ants whose eyes are normal. LuBBOCK over twenty-five years ago stated this fact, but his observation has been either overlooked or ignored by recent conti- nental writers. Experiments with Colored Pathways.—TVhe purpose of the experi- ments now to be described was not to determine whether ants have color vision, but simply to ascertain whether the color of the path plays a part in the home-going of ants. In these experi- ments the usual stage was employed; but the inclines were com- posed of colored papers of practically the same texture. Each incline was composed of one color. Of the color series I used, red, yellow, green, blue, purple; of the brightness series, white and black. When an ant had learned the way down a certain colored incline so well that it was not disturbed by the substitution of another of the same color, an incline of a different color was 402 ‘fournal of Comparative Neurology and Psychology. substituted for the old one. ‘This was repeated until all of the colors had been used at least once. All ants were not affected in the same way by these changes. Myrmica punctiventris was slightly affected each time a change i in color was made (Fig. 15); another ant (sp. !) was disturbed by a change from black to white, but not by changes from one member of the color series to another. Formica fusca was not usually disturbed by any changes made but occasionally (Fig. 14) it was slightly disturbed. This shows that changes in color of the pathway do not disturb some ants at all’ whereas other ants seem to be slightly affected by such changes;’ while yet others are affected by changes in brightness but not by changes in hue. Experiments with Auditory Stimulit.—This section is not in- tended to be an exhaustive discussion of the auditory sense of ants. It is, however, an attempt to collate our knowledge on the subject, to give additional experimental data and to harmonize the con- flicting views. About a century ago St. FarGeEau, in his Hist. Nat. des Hyme- nopteres, asserted positively that ants hear (LuBBocK ’81, p. 221). But Huser (Nat. Hist. of Ants), Foret (Fourmis de la Suisse) and Lussock (’81) conducted experiments to test the power of ants to hear and each decided that they were deaf to sounds that fall within the human auditory range. Sir JoHN Luppock’s experiments were especially well planned. He used sounds produced by a dog whistle, a violin, the human voice, a shrill penny pipe, and a full set of tuning forks. These experiments were tried both upon ants that were carrying pupz home and upon ants confined to paper bridges. In no case did he get the slightest response to any of the sounds made. Since, however, he had discovered in ants what he thought was an audi- tory organ, he presumed that it was probable that ants both heard and produced sounds; but that they were tones that fell outside of the human auditory range. The negative results of the experiments of these three author- ities have caused most people to believe that ants cannot hear. 5 Lupgock (Joc. cit., p. 198) performed an experiment which showed that the ants studied by him did not notice the color of the path. He placed some ants on a narrow bridge, which was supported by pins with their bases in the water. On this bridge he projected a spectrum and noticed that the ants acted as they did on the plain bridge. 6 Tt is not improbable that these disturbances may have been due to a slight difference in the texture or odor of the different papers. Turner, Homing of Ants. 403 Yet, if they cannot hear, it is hard to understand why so many ants are provided with organs for producing sound. Davip SHARP (93) has proved conclusively that the Dorylidz, the Pon- eride and the Myrmicidz possess stridulating organs, each con- sisting of a file on the anterior portion of the tergum of the third abdominal somite, which file is rubbed against the roughened underside of the secona abdominal ring. He did not find any such organ in either the Camponotidz nor the Dolichoderida and he was doubtful of its presence in the Cryptoceride. LANpotIs (67) had seen such an organ about fifty years ago in PONERA. SHARP'S work was strictly morphological, but physiological evidence that ants can make audible sounds is furnished by WrRouGHTON (92), Foret (Fourmis de la Suisse), WAsMANN (gt) and JANET (’94), each of whom has heard sounds produced by ants. WROUGHTON listened to Cremastogaster rogenhoferi, Foret to Camponotus ligniperdus, JaNerT to Myrmica rubra L. and ‘Tetramorium cespitum L. and WasMANN to Myrmica rugi- nodis. So far as I have been able to ascertain, WASMANN (’91) 1s the only European who has produced any experimental evidence against the negative results of Husper, Foret and Lussock. He experimented with Formica rufa, which was confined in a Lussock nest, the floor of which was covered with one mm. of dirt. Whenever he scratched on the glass with a needle, he found that the ants made responsive movements which demonstrated they were affected by the stimulus. Experience has taught me that the slightest touch upon any portion of my JANET nests 1s almost certain to be responded to by some movement of the ants within. It is to be regretted that WAsMANN did not produce his sounds in some orher way than by scratching on the nest; for, since there is a possibility that the scratching produced a slight tremor, a fair critic is bound to admit that the positive evidence furnished by this experiment is not of sufficient weight to offset the negative results of the three men mentioned above. About eight years ago an American, LE Roy D. WELD (’99) performed a carefully planned series of experiments, which seem to prove that Cremasgaster lineolata, Lasius Americanus and Aphenogaster sp. can hear sounds that are audible to man. Some of these experiments were conducted under conditions that apparently precluded the possibility of jars from the sounding body reaching the nest by any other medium than the air. ‘The 404 “fournal of Comparative Neurology and Psychology. sounds used were produced by a steel bar, a tin whistle, a wooden whistle, a middle A tuning fork and a milled disk rotating against the edge of a card. FieLtpe and Parker (’o4) claim that ants do not react to aerial sound waves from a piano, violin, or GALTON whistle; but only to vibrations that reach them through some solid body. Recently [ have performed a series of experiments upon For- mica fusca var. subsericea Say, and a variety of Formica sanguinea Latr. For these experiments the ants were housed in JANET nests, the covers of which were composed of orange colored glass. During the day, the experiments were never begun until the sun was several hours high; and at night, the experiments were not begun until the electric lights had been shining for at least three hours. ‘These precautions were taken to eliminate the disturbing effects of a change in illumination. ‘To lessen the possibility of jars from the sounding body reaching the nest through any medium other than air, the nests were placed on a layer of cotton batting half an inch thick. For the same reason, the legs of the table on which the nests rested were placed on thick wads of cotton batting. The sounds used were produced by the Galton whistle, organ pipes and the human voice. Sometimes the whistle or pipe was held near the nest and in other cases a short distance away. The pipes were always held in such a position that the air when expelled from the pipe could not impinge directly upon the nest. These experiments were conducted during the winter and early spring. They showed conclusively that each of the two species of ants mentioned was sensitive to vibrations of the air which to the human ear would be sounds. I obtained responses to notes as high as 4138 vibrations per second and as low as 256 vibrations per second. The responses, in the form of zigzag movements, were usually slight for pitches higher than 3000 vibrations per second and sometimes slight for other pitches; but, to most pitches under 3000 vibrations per second, the ants usually responded in a pronounced manner, usually darting about as though much excited. After the ants had been subjected to the sounds for a long time they seemed to become fatigued, failing to respond in the least to pitches which would call forth pronounced responses from fresh ants. ‘The fertile females seemed to be more intensely excited by the tones than the neuters. Several series of experiments were performed. Details of four typical series are recorded in the following table. Turner, Homing of Ants. 405 TABLE VIII. Series I. VIBRATIONS. ForMICA FUSCA. | ForMICA FUSCA. ForMICA SANGUINEA. 3480 | Slight Slight Response | Marked | Slight Marked 3906 P | g Slight 4138 None | None None Series 2. VIBRATIONS. ForMICA FUSCA. ForMICA FUSCA. ForMICA SANGUINEA. 256 Marked Slight Marked 3168 Response Slight Slight 3480 Very slight Very slight Very slight 512 Marked Marked Marked 316 Slight* Slight ? Slight ? 880 Response Response Response 3906 Slight Slight ? Slight ? 4138 Very slight None Very slight 4645 ? None ? 256 Slight None Marked 512 Slight Slight Slight Series 3. VIBRATIONS. ForMICA FUSCA. ForMICA FUSCA. ForMICA SANGUINEA. 1760 Marked Response Slight 1056 Marked Slight > 880 Slight Slight None? 512 Marked Very slight Marked 362 None? None? Slight? 256 None None Slight? 3168 Response Response None 2816 Slight Slight Response 2640 Response Slight Slight 2280 Marked Marked Marked 2112 Response Response Response 1980 Response Response Very slight 1900 Marked Slight Response Series 4. VIBRATIONS. Formica FuscA. | ForMICA FUSCA. ForMICA SANGUINEA | FoRMICA FUSCA. 512 Slight Slight Slight None 880 Very slight Very slight Slight? | None 256 ? ? | Marked None 1056 Marked Marked | P None 3168 Slight ? None? Slight? None Slight Slight ? | ? None 512 Marked | Response Marked None 362 Marked | Slight ? Slight None 256 Marked | Slight Marked None * The fertile females gave a marked response. 406 “fournal of Comparative Neurology and Psychology. EXPLANATION oF Taste VIII. Column 1. The number of vibrations per second. Column 2. A colony of Formica fusca var. subsericea Say, containing three fertile females and about two hundred neuters. Column 3. A colony of Formica fusca var. subsericea Say, containing about two hundred neuters and no fertile females. Column 4. A mixed colony of Formica sanguinea Latrl. and Formica fusca. This colony contained about two hundred neuters, over two-thirds of which were fuscas. No females were present. Column 5. A few workers of Formica fusca var. subsericea Say, the antenne of which had been removed two days before. Some may think the response of these ants was the result of tactile stimuli caused by shaking of the nest by the sound wave. To me it does not seem possible for the nests to have vibrated in response to each of so wide a range of pitches. However, to meet that objection, the following experiment was devised. | obtained some felt cloth two mm. thick and placed two layers of it in the bottom of each of the living chambers of a JANET nest. This nest rested on cotton placed on the table the legs of which rested on wads of cotton. After the ants had become accustomed to their carpeted floor, various sounds were made. ‘This experi- ment was tried upon both Formica fusca var. subsericea Say and Formica sanguinea Latr. It is important to note that even with the false felt bottom to stand on, the ants could not reach the top of the nest with their antennz. In each case the ants responded to the sound in the same manner as has been described above. Since the felt would prevent nest tremors reaching the ants, the response, 1t seems to me, must have been to air vibrations which the human ear would sense as sounds. At present I have housed in a JANET nest a small colony of Camptonus herculeano-ligniperdus. ‘This colony consists of eight winged females and about twice as many neuters. Usually one of the neuters mounts guard in the outer doorway. Whenever sounds similar to those mentioned above are made, this guard shows marked evidence of being disturbed. Under such condi- tions it makes agitated movements with its antennz and often snaps with its jaws right and left. If the noise is continued, the guard is apt to rush back into the nest. Undoubtedly, artificial colonies of each of these species of ants respond in a pronounced manner to atmospheric vibrations which to the human ear would be sounds. When, however, I tested ants that were moving about outside the nest, I obtained no such marked reactions. Often I could Turner, Homing of Ants. 407 detect no response whatever. But usually there would be a slight movement of the antennz, or a slight but sudden acceler- ation of speed, or else a halting movement. Sometimes the ant would stop still. All of these movements were slight, usually so slight that they might easily be overlooked by an observer who was not expecting a response of some kind. Even with single individuals confined to a test-tube which was suspended by a string (a method much used by WE Lp), I some- times obtained no response. Indeed, I seldom received more pronounced responses than a slight movement of an antennae, or a twitch of one or more legs, or a more tense attitude of the body. In most cases, however, I did get these slight movements, and, occasionally, I observed wide sweeping movements of the antennz or other easily observed bodily movements. The pronounced positive results obtained by WELD and my- self do not seem to harmonize with the equally pronounced nega- tive results obtained by Huser, Foret and Luspspock. One cannot believe that the American ants are physiologically as much unlike the European as these antagonistic results would seem to indicate. It seems to me that these contradictory results can be harmonized in the following manner. LuBBOCK’s experiments were made either upon ants that were carrying pupz home, or else upon ants that were confined to paper bridges. As has been stated above, my experiments show that under such conditions the ants usually do not respond at all to sounds, or else react with movements so slight that they might easily be overlooked by one not expecting movements of the kind made. I am fully convinced that experiments conducted upon European ants colonized in JANET nests, would yield the same positive results obtained by WELD and myself. To harmonize FIELDE and PaRKER’s (’04) work with mine 1s not an easy task. We agree that, up to about 4000 vibrations per second, ants respond to a long range of vibrations which a human ear would sense as sounds. ‘Uhey claim that these vibra- tions are responded to only when received through a solid medium; while my experiments seem to show that they are responded to when received through the air. Yet two of the species used by me (F. sanguinea and F. fusca var. subsericea) were also used by them. ‘They used the piano, the violin and the Galton-whistle; I used organ pipes and the Galton-whistle. It seems to me that 408 “fournal of Comparative Neurology and Psychology. I have taken even more precautions to preclude the possibility of the vibrations reaching the ants through a solid medium than they did. The only precaution they took was to rest the nest upon a thick layer of paper. I carpeted my nests with two layers of thick felt, placed a thick layer of cotton between the nest and the Lupspock island and thicker wads of cotton beneath the legs of the table upon which the island rested. ‘The lowest note of the whistle used by FiELDE and PARKER was 10,000 vibrations per second, which is far above the highest pitch to which ants respond. The whistle I used ranged from 3480 to 51,000 vibrations per second. If the sound of the piano and of the violin were common in the room where the ants were kept they may have become too familiar to arouse responses. My ants were kept in a room fac- ing a paved street. The noises of this street did not disturb them. The more | meditate on FIELDE and PARKER’s paper, the more I am inclined to believe that the lack of harmony between their results and mine is due to the difference in our technique. In their experiments the same note was sounded ten times in slow succession; in mine each note was sounded continuously or else in rapid succession for one minute or longer. I thought that, if ants can hear, the occasional production of a note might simply attract attention, while a continuous rapid repetition of the same would arouse some form of visible motor response. In my own experiments a note repeated a few times in slow succession would often cause no response, yet a prolongation of the sound or a con- tinuous rapid repetition of the same would soon produce marked responses. If this contention be valid, we would expect FIELDE and PaRKER’s technique to yield nothing more than an occasional response. And this is the result they obtained; for on page 643 (Joc. cit.) they say, “‘ Now and then an ant would seem to respon to a given note, but in every case repetitions of the experiment gave a negative result. Responses to light within and without the nest exhibit differ- ences similar to those observed for sound. If light is admitted into a nest, the ants at once are much disturbed and show it by vigorous movements. Yet in the outer world, so long as the direction of the rays remains the same, ants are scarcely, if at all, affected by changes in the intensity of the light. Both light and sound are almost constant factors of the external world, but are TurRNER, Homing of Ants. 409 rare phenomena in the confines of the nest. With ants, as with us, it seems that unusual stimuli cause unrest. Experiments on Direction and Distance.—LuBBock (81, p. 260) devised an experiment which shows that ants have accurate impressions of direction in a horizontal plane. ‘This experiment, a portion of which I shall presently quote, has been repeated by me, with confirmatory results, on most of our southern species of ants. LuBBock writes as follows,‘‘I then accustomed some ants (Lasius niger) to go to and fro over a wooden bridge, 6, c, tosome food. When they had got quite accustomed to the way, I watched when an ant was on the bridge and then turned it around, so that the end 6 was at c, andc at b. In most cases the ant immediately turned around also; but even if she went on to b or c, as the case may be, as soon as she came to the end of the bridge she turned around. I then modified the arrangement, placing between the nest and the food three similar pieces of wood. Then when the ant was on the middle piece, I transposed the other two. To my surprise this did not at all disconcert them.”’ Lussock next tried a different experiment. By means of a pin through its center, he pivoted a disk of cardboard to the middle of the table. At one corner of the table he placed some food. When the ants had come to know their way so that they passed straight over the paper disk on their way from the nest to the food, he moved the disk around with an ant on it, so that the side that had been towards the nest was now towards the food. As in the above experiment, the ants turned round with the paper. He also repeated it with a table arranged to revolve in three concentric sections. ‘The result was the same as above. I think some of my experiments warrant the assertion that ants have an impression of vertical as well as horizontal direction. In my experiments (excepting those on color and tactile phenom- ena) the stage and incline were made out of the same material, often being cut from the same piece of cardboard. The apparatus was so adjusted that when the ant reached the union of stage with incline, only two changes in movement were necessary in order to take the next lap of the journey; the ant must turn so as to receive the light rays on a different side than before, and it must move obliquely downward instead of moving in a_ horizontal plane. ‘The apparatus was also so adjusted that if the ant turned at the parting of the ways and moved horizontally along the stage in 410 ‘fournal of Comparative Neurology and Psychology. the direction of the nest, it would receive the light rays on the same side of the body as it would in passing down the incline. The only new factor that entered the life of the ant that moved down the incline rather than horizontally along the stage towards the nest was the experience of moving downward. Vertical changes in direction must affect ants differently from horizontal changes in direction, otherwise there is no reason why the ant should pass regularly down the incline rather than horizontally along the stage towards the nest. Again, after an ant had thoroughly learned the way down and up the incline, it would often not take the trouble to go all the way to the union of stage and incline, but, reaching the edge of the stage at a greater or less distance from its union with the incline, the ant would reach downward over the edge; if it could not touch the incline it would move nearer the junction of stage and incline and again reach downwards. If necessary, it would move nearer still. As soon as the ant could touch the incline, it would step off of the stage and move obliquely downward to the nest. To test this matter further, the following experiment was de- vised. Underneath the incline of one of my stages was placed an adjustable support made by sticking a pin vertically into the cork of asmall bottle. Since the lower end of the incline was free while the upper was attached to the stage, by moving this adjustable support more or less towards the base of the incline, a vertical gap of any desired size could be formed and maintained between the base of the incline and the island. After the ants had thoroughly learned the way to and from the nest, a small vertical gap, which was gradually increased until the ants on the island could no longer touch the base of the incline with their antennz, was made between the foot of the incline and the island. Now so long as the ants could touch the incline with their antenne, they would stretch upward until their forefeet touched the incline, then mount the incline and go to the stage. Ants coming down the stage acted the same way, only they stretched downward rather than upward. When, however, the base of the incline could no longer be touched by their antennz, the ants would come to the place where the foot of the incline had been, elevate the front part of their bodies as much as possible and reach upwards with their antenne. Ele- vating the base of the incline still more, I placed along the side of the path and parallel to it a stack of clean microscopic slides, one Turner, Homing of Ants. 4II high. This stack of slides was placed near enough to the dhe “for the ants, by stepping across a narrow horizontal gap, to pass easily from it to the incline. At once the ants mounted the stack of slides and went up the incline. Does not this indicate that the ants have an impression of verti- cal direction? Does it not also indicate that they have an impres- sion of distance? Otherwise, what makes them stop at that place and feel upward for the incline? It will not do to say that they stopped because the scented trail terminated there, for, even if it were not true that tracks of those ants cross that place 1 in all direc- tions, it has already been shown that the termination of a scented trail does not impede the progress of ants moving in a direction that is thoroughly known. Nor will it do to say that they halted and acted that way because they sensed the incline above them; for they do not react that way towards bodies held above them on other parts of the stage, nor do untrained ants react that way when they pass underneath the incline. In each of these test cases, ants may reach upward with the antennz and then pass on; but they do not return to the spot over and over again and reach up as though they were hunting for something that was lost. This view is supported by yet ies experiments. In another connection, I have mentioned the case of a worker ant that learned to come to a certain point on the island, mount my forceps and be conveyed to the stage. After this mode of conduct had been thoroughly learned, the ant, on leaving the nest, would immedi- ately go to this position on the island and meander in a small circle until my forceps were presented. It was not an uncommon thing for an ant that knew the way up and down a certain incline, occasionally to miss arriving at the foot of the incline. In such cases, the ant would not, as a rule, roam here and yon, but would usually turn quickly and hunt for the incline. If it missed it again, it would go back to some point in its path, often as far as the nest opening itself, and, taking a new start, arrive at the incline without further trouble. “This mode of conduct not only indicates that ants use certain land- marks as such, a subject which will be referred to in the next sec- tion, but it also seems to harmonize with the views stated in this section. Another common occurrence which puzzled me not a little may, perhaps, be best explained in this connection. In some 412 “fournal of Comparative Neurology and Psychology. experiments planned to solve a problem not discussed in this paper, two stages were used, from each of which an incline of the same kind led to the island. In one case the incline was attached to the right side of the stage and in the other tothe left. Incline num- ber two was nearer the nest opening than was incline number one. An ant that had thoroughly learned the way up and down incline number one, would occasionally start up incline number two. Sometimes it would continue on to the stage, but more frequently, when only part of the way up, it would turn about and return to the island and continue on to incline number one; either directly, or else after having first retreated to some point farther back on the trail. “lhe same thing sometimes happened where two inclines were attached to opposite sides of the same stage. From a psychological point of view, there were only two differ- ences between incline number one and incline number two; incline number one was scented with the footprints of the ant, while incline number two was not; and incline number one was farther from the nest than incline number two. Since it has been shown that the substitution of an unscented incline for one scented by the tracks of ants is not a disturbing stimulus to ants that are moving in a well known path, the difference in the distance traveled seems the only thing that could have caused the inhibition. The grouping of all these data in one section does not imply that they are of the same psychic order. Some may be simple and others complex. Some may even be the derivatives of others mentioned here. ‘lo discuss such issues is not the purpose of this section. Its sole aim is to show that the psychic impression that confronts the home-going ant is not a simple olfactory stimulus, as most writers seem to suppose, but that it is a complex impres- sion composed of contributions from the olfactory (topochemical), visual, tactile and kinesthetic and auditory senses. V. HAVE ANTS ASSOCIATIVE MEMORY? ‘“ By associative memory, I mean the two following peculiarities of our central nervous system: First, that processes which occur there leave an impression or trace by which they can be reproduced even under different circumstances than those under which they originated. . . . . Vhesecond peculiarity is, that two processes which occur simultaneously or in quick succession will leave TuRNER, Homing of Ants. 413 traces which fuse together, so that if, later, one of the processes 1s repeated, the other will necessarily be repeated also. By asso- clative memory we mean, therefore, that mechanism by means of which a stimulus produces not only the effect which corre- spond to its nature and the specific structure of the stimulated organ, but which produces, in addition, such effects of other causes as at some former times may have attacked the organism, almost or quite simultaneously with the given stimulus” (LoEB "02, pp. 213 and 214). As to the criteria of memory, the same author writes (bid, p. 218): “It will require more observations than we have made at present to give absolutely unequivocal criteria. For the present, we can say that if an animal can learn, that is, if it can be trained to react in a desired way upon certain stimuli (signs), it must pos- sess associative memory. ‘The only fault with this criterion is that an animal may be able to remember (and to associate) and yet not yield to our attempts to trainit. . . . . The fusion or growing together of heterogeneous but by chance simultaneous processes is a sure criterion for the existence of associative mem- ory.” If we consider the experiments herein described in the light of these criteria of Lors, we must certainly conclude that ants have associative memory; for, as has been shown, ants learn by experi- ence, retain what they learn in a way that can be recalled by the proper stimuli, and they can be trained to do certain things. Some psychologists do not agree with Logs that the mere ability to learn predicates memory. ‘There is a method of learning depending on repeated blundering efforts with fortuitious suc- cesses that are gradually selected which Ltoyp Moreaw calls the method of trial and error. Such a mode of learning, some say, does not indicate the existence of memory. Memory predicates the existence of ideas which are associated. In learning by trial and error, no association of ideas is involved; we have simply assimilation of a sense impression or impressions with an impulse. To use the words of THORNDIKE (Joc. ct., p.71) “ The ground work of animal associations is not the association of ideas, but the asso- ciation of idea or sense-impression with the impulse.” HosuouseE (or), who experimented on the same animals as ‘THORNDIKE, differs from him in his conclusions. He thinks that the average laboratory psychologist has gone to extremes in his 414 ‘fournal of Comparative Neurology and Psychology. conclusions which the experiments do not warrant. He reminds us that, “A dog may show not merely highly developed hunting instincts, but real cleverness in the adaptation of past experience when it is a question of catching a hare, but may also be an intol- erable dullard about openingabox. . . . . To test an animal’s intelligence by mechanisms, seems to be about on a par with gaug- ing the nature of a man’s intelligence by certain puzzles, in which, as 1s well known, many able men are, indeed, dullards . : What Mr. THoRNDIKE’s experiments prove so far is, not that cats and dogs are invariably educated by the association process, that is by habituation, but, on the contrary, that at least some cats and dogs conform, in at least one point, to the method of acquisitions by concrete experiences . . . . . In some cases they not only merely learn to meet a given perception with a given motor reac- tion, but also to combine and adapt their actions so as to effect physical changes which, as they have learned, aid them in gain- ing their ends.”’ Hosnouse further reminds us that habits are generally formed gradually by many repetitions and that, where the act is the result of habituation, the time curve should gradually descend. Indeed, Tornpike himself lays stress upon this same point; for he says (/oc. cit., p. 45): “And if there were in these animals any power of inference: however rudimentary, however sporadic, however dim, there should have appeared among the multitude some cases where an animal, seeing through the situation, knows the proper act and does it, and from then on does it immediately upon being con- fronted with the situation. “There ought then to bea sudden verti- cal descent of the time curve. Of course, where the act resulting from the impulse 1s very simple, very obvious, and very clearly defined, a simple experience may make the association perfect and we may have an abrupt descent in the time curve without needing to suppose inference. But if in a complex act, one found such a sudden consummation in the associative process, one might well claim that reason was at work. . . . . The gradual slope of the time curve, then, shows the absence of reasoning. ‘They represent the wearing smooth of a path in the brain, not the deci- sions of a rational consciousness.’ If a gradual descent of the time curve be a sure test that a crea- ture has learned by the method of trial and error and if an abrupt descent is an indication of perceptual learning, we must conclude TurneER, Homing of Ants. 415 that the ant’s mode of learning is not the trial and error but the perceptual mode and that they have memory (Figs. 8, 10, 12, 18). I freely confess, however, that this seems to me a very unsatis- factory criterion. For this criterion to be of any real value, we would need some way to determine, from the animal’s standpoint, what is complicated and what is not; and there is no known way. There are, however, other considerations that lead me to believe that ants have memory. In memory wé usually find a complex impression composed of contributions from more than one sense organ. In the section that precedes this, I have tried to show that home-going ants have a complex impression which is composed of contributions from several sense organs. Now the existence of this complex impression does not necessarily prove the existence of memory, but, it does demonstrate the existence of one of the usual physiological accompaniments of memory. ButTTeL-REEPEN (00) chloroformed some bees and found that - they were no longer able to find their way home. ‘This he con- sidered a proof of memory. How, he asks, could they forget unless they had a psychic content to forget. Is it not reasonable to consider cases of ‘forgetting’ that cannot be attributed to injuries received nor to retardation due to fatigue as evidences of memory? In my experiments, I frequently met cases of what might be called lapses of memory. Sometimes this would happen on the stage, at others on the island. “The ant would move about at random as though it had forgotten the way. ‘These lapses would increase the time ordinarily required for the trip, from one to three minutes. ‘he subsequent trips were made as rapidly and as regularly as before. I have also met with cases of what might be called mistaken identity (Fig. 7). Recall the experiment with two stages, from each of which an incline descended to the island—the case when the ant occasionally went up the wrong incline—and you have a case of this kind. Then, too, I have noticed cases which I think reveal ants as using a thing or things as a means of accomplishing a certain end, rather than responding to it as an end. ‘Take the experiment mentioned in the previous section, where a high vertical gap was present between the foot of the incline and the island. The ants were coming repeatedly to the same point and reaching upward for something beyond their reach. In that place, a stack of clean glass slides was placed a little to one side of the path. At once the 416 “fournal of Comparative Neurology and Psychology. ants climbed the slides and mounted the incline. Similar stacks of slides placed just to one side of the path at other points along the trail were not thus mounted by the ants. ‘The road by which the ants had learned to reach the stage from which the gap now separated them, had no vertical walls that needed to be climbed. This wall of slides was not placed across the path, but a little to one side of it. Immediately, the ants mounted it and thereafter used it as a means of passing from island to incline and from incline to island. Then, too, it seems to me that ants use light as a means to an end. In a former section, an attempt was made to prove that light is not for ants a tropic stimulus, yet, by a repetition, under control, of one of LuBBock’s experiments, I have shown that the direction of the rays of light does have a guiding influence. In each of those experiments, I used a cardboard stage, which was connected with the Lussock island by means of a single incline. In some experiments, I placed the light (a 16 c.p. incan- descent electric lamp) near the side to which the incline was at- tached, in other cases the light was placed on the opposite side. There was no other light in the room. After waiting untrl the ant had thoroughly jeacned the way down the incline,’ I trans- ferred the light to the opposite side. Invariably, as shown by its movements, the ant would be very much disturbed by the change. Now to my mind, this disturbance seems due to the fact that the ants use the direction of the rays of light as reference data. In other words, the light is responded to as a means to an end and not as itself an end. If this contention be valid, then ants learn, not by the method of trial and error, but perceptually, and they have associative memory. As the result of careful observation, | am convinced that ants use such things as irregularities of the surface, edges of flat surfaces, the edges of shadows, etc., as reference data. On several occasions, after a marked worker had been carrying pupz for a long time, it was imprisoned from one to several hours and then returned to the stage. In some cases the ant failed com- pletely to find the way to the nest, in a few others it went to the nest immediately; but in most cases, it took some time to find the way to the nest. Usually the time required to re-solve the problem was 7 When the ant was not disturbed in its movements by the substitution of a new incline of the same kind for the old, the ant was judged to be thoroughly acquainted with the way. This was the test I always used ip such cases. Turner, Homing of Ants. 417 shorter, often much shorter, than was needed by the ant the first time such a problem was presented. Aside from the differences of time required for different individuals to reach the nest, there were marked individual differences in their conduct. I noted five different modes of response: (1) An ant would take a pupa to the nest and remain therein instead of returning to the stage. (2) An ant would pass unburdened to the nest and remain there- in instead of returning to the stage. (3) An ant would carry a pupa to the nest or to some point on the island and then return to the stage and make repeated trips. (4) An ant would pass, unburdened, to the nest, and then immediately return to the stage for a pupa and thereafter make regular trips. (5) Occasionally, a worker would, pass, unburdened, to a point near the nest, then return to the stage for a pupa and thereafter make regular trips. Such marked individual differences in conduct were noted be- tween members of the same colony, under the same external environmental conditions. -All of these individuals, at the time of their imprisonment, were thoroughly acquainted with the way between the nest and the stage. ‘The difference cannot be attrib- uted to fatigue, for they had been resting from one to several hours. It seems to me that these facts harmonize rather with the memory hypothesis than with either the “assimilation of idea to impulse”’ or to the “determination of stimulus’ hypothesis. If either of the two latter hypotheses is true, the following is probably a fair statement of the factors that guided my ants to and from the stage to the nest. ‘The sensing of the pupa on the stage causes the ant to pick it up and move in a certain direction. ‘The direction of the rays of light, etc., cause it to move to the point of union of stage and incline. Contact with that point, etc., cause it to turn and pass down the incline, receiving the light on a differ- ent side than before. At the base of the incline, contact with the zinc, etc., Cause it to pass to the side of the nest, where other stimuli cause it to go into the nest, where yet other stimuli cause it to lay down its pupa. Now all of these several stimuli act on it in a reverse way and cause it to return to the stage. After the “assim- ilation of impulse and idea” or the “determination of stimulus” has been made, the animal moves along just as automatically as an animal in a tropic response. Each stimulus is responded to as an end in itself and not as a means to anend. Now, if I interpret my experiments aright, none of the ants examined acted this way. 418 “fournal of Comparative Neurology and Psychology. They acted as though getting the pupz under some shelter, pre- ferably the nest, was the end in view, and as if all of their responses to stimuli were but means put forth to accomplish that end. If either of these hypotheses applies to ants, a worker meeting a pupa anywhere in the open air, unless carried on by the momen- tum of other stimuli, should respond to it in the same way that it would had it encountered the pupa on the stage. Now what are the facts in the case? Whenever I place a pupa in the track of such a worker, as soon as the pupa was sensed it was picked up by the worker unnl Crean eulle nest; either back along the path just traveled or first to some reference point and thence to the nest. One might suppose that this is what would have happened had the ant encountered the pupa on the stage, for there is a general belief that the ant picks up the pupa, turns about at once and returns in its own tracks to the nest. With some species of ants, especially when they are working in concert, it is hard to decide whether this is the case or not; but even a casual observation of an individual of Formica fusca var. subsericea Say, working alone, will convince anyone that such is not the case. With the pupz scattered here and there over the stage, the ant must hunt for them, and each pupa is approached over a different, more or less sinuous, line. When the pupa is reached, the axis of the ant’s body may make almost any angle, from zero to one hundred and eighty degrees, with the line that leads from the pupa to the union of stage and incline. Hardly any two of these positions are alike. Whether or no, on picking up the pupa, the ant turns about at once is a secondary matter, dependent upon the position of the ant’s head at the time. The species just mentioned rarely, if ever, returned in its own tracks from the pupz to the union of stage and incline. What the ant really does is to pick up the pupa and move off in such a direction as to keep the light rays falling on that portion of the body which it has learned from other trips, must receive the rays of light if it is to reach the union of stage and incline. ‘This may not take it direct to the junction but it will take it to some familiar point and from there it will take a bee- line to the junction. In the test experiments, I took especial pains to place the pupz in such situations that, to get to the nest, the ant must move in such a direction as to receive the rays of light on a different side of the body and at a different angle than would have been the case had the pupa been encountered on the stage. And Turner, Homing of Ants. 419 yet the ant always went home without a moment’s hesitation. ‘To say that the pupa does not give the ant a stimulus to make a cer- tain definite movement, but any movement necessary to take it home, is equivalent to saying that the ant responds to the stimulus as a means to an end. There is yet another observation which deserves mention in this connection. In all my experiments with ants working in concert, when the pup had all been removed from the stage the ants would thoroughly search the stage and incline and then retire to the nest. Why did the ants stop going to the stage Surely the same stimuli existed in the nest as formerly; the only new factor is that there are no more pupe on the stage. ‘This cessation of movement cannot be attributed to fatigue, for they act in the same manner after removing a small lot of pupz as they do after removing a pile many times as great. It would be equally fallacious to assume that the pup on the stage influence the ants directly from afar, and that the lack of pupz to furnish an immediate stimulus is the cause of the cessation; for ants roaming over the island in search of pup are not reered by the pupez. Many a time [| have seen specimens of Lasius, and Formica pass repeatedly back and forth beneath the stage without being attracted by the pupz on the stage; I have known Formica to ascend the incline and pass to within three centimeters of the pupz and return to the island without sensing them. ‘To test whether such ants were in a con- dition to be affected by the sensing of pupz, I have placed pupz in front of them as soon as they reached the island. Immediately such pupz were picked up and carried to the nest. Nor would it be consistent to assert that the stimulus-to-return did not affect ants that returned unburdened to the nest; for after the pupz were removed, each ant always made from one to a few round trips; and, furthermore, it frequently happened, when the stage was well supplied with pupz, that an ant would pass, unladen, to the nest, and then immediately, or after a few moments, resume its period- ical round trips. I fail to see how these facts can be harmonized with either the assimilation hypothesis of THORNDIKE or with the determination of stimulus hypothesis, but they do harmonize with the assump- tion that ants have associative memory. It is not claimed that ants exhibit a high grade of intelligence in finding their way about. In my experiments, the first trip to the nest was always the result 420 “fournal of Comparative Neurology and Psychology. of happening to succeed after making many blunders. So far this accords with the observations made by all students of vertebrate animal behavior. During the first few trips, however, aided by memory, certain associations of ideas are formed which persist and enable the ant to orient itself. “To men of THORNDIKE’S con- viction, this is a discordant note, but it accords perfectly with Hopxouse’s view concerning vertebrates. Although this does not necessarily make the ant a self-conscious creature, it makes it much more than a mere reflex machine. It is worth noting that in ascribing associative memory to ants, I am in perfect accord, not only with WaAsmMann, but also with ee a stanch advocate of the tropisms as Logs. In his chapter “The distribution of associative memory in the animal kingdom”’ ae p. 224), Lores says, “Although I heartily sympathize with BETHE’S reaction against the anthropomorphic conception of animal instincts, I yet believe that he is mistaken in denying the existence of associative memory in ants and bees.” After the space occupied in attempting to show that the “ assim1- lation” and “determination of stimulus’’ hypotheses do not ade- quately account for the conduct of home-going ants, I fear that by “associative memory” I may be amicunderstood to mean some high type of rational life. ‘Therefore, | wish to emphasize the fact that I use the term as Lor uses it in the quotation with which this section opens. VI. DOES DIVISION OF LABOR AMONG ANTS SIGNIFY MUTUAL COOPERATION ? In the course of my experiments many examples of division of labor have been noticed. JI have known one set of workers to store the pupz under the incline, or in a dark chamber, or on the stage in the open, while yet other workers conveyed them from these places to the nest. I have seen one set of workers lay pupe down on the island, while a different set carried them into the nest. I have even seen cases where one set laid the pupz down in the feed chamber, from which another set carried them into the nest proper. I have noticed similar examples of division of labor in most of the species experimented upon. It was with Prenolepis imparis, however, that I took pains to investigate it carefully. Here we have one set of workers bringing pupz to a certain spot Turner, Homing of Ants. 421 whence they are carried further by another set of workers. Each set does its share, and between them the work is quickly accom- plished. ‘Thus stated, these facts suggest the thought that these codperating ants must have some kind of mutual understanding. I am constrained, however, to add a statement which will give a different interpretation to the act. In all of the cases of cooper- ation observed by me, the pup were carried into the nest by workers which, roaming from the nest, by chance discovered them. The workers that stored the pupz under the incline, those that hid pupe in the dark chamber, and those that piled them in the open on the stage, were ants which, being unable to find the way home disposed of the pupe in these special ways. ‘Those that laid them on the island simply brought the pupe thus far and laid them down. ‘They did not go to the nest, nor did they touch with their antennz any of the ants that subsequently carried these pupz to the nest. In all of the cases mentioned, some worker or workers, roaming from the nest, discovered the pupz and carried them home. ‘The only observations that did not accord completely with the above statement, were those few cases in which the workers from the stage deposited the pupa in a group of ants that was swarming against one side of the nest. In that case, 1t was impossible to ben sure that some of those did not assist the stragglers from the nest in carrying the pupe. Even in that case, I could detect no evidences of communication between the ants from the stage and those from the nest. “These observa- tions warrant the assertion that each case of division of labor observed by me was the result, not of mutual understanding, but rather of accidental cooperation. On re-reading MoccripGe’s (RoMANES ’92, p. 98) account of ants dropping leaves for other ants to pick up, bene Si(ibid tp: 99) account of “regular depots for their provisions,” and BELY’ S (tbid., p. 99) account of ants casting down burdens for other ants to carry, I cannot help but feel that had these special cases been observed from the beginning, they would have proved to be of the same type as those described above. Any one acquainted with the habits of ants who reads MocGrinGE’s account of the cooper- ative dismemberment of a grasshopper by ants, is certain to be reminded that any large insect dropped in the midst of ants 1s sure to be independently attacked from all sides by ants which, acting as individuals, gnaw and pull away from the object. 422 ‘fournal of Comparative Neurology and Psychology. Mocecrince (loc. cit., p. 98) tells us that the grasshopper was “too large to pass through the door, so they tried to dismember it. Failing in this, several ants drew the wings and legs as far back as possible, while others gnawed through the muscles where the strain was greatest. [hey succeeded at last in thus pulling itin.” It is a pity this account is so meager. If there were only two ants to each member, one pulling it out and the other gnawing at the place of greatest strain, we would have an undoubted case of intel- ligent cooperation; whereas, if there were a large number of ants attacking each member, the fact that some happened to gnaw at the point t of greatest strain does not prove intelligence. To repeat, there is nothing in the anecdote, as recorded, to show that this was was not a case of accidental rather than of mutual cooperation. A case cited by RoMaNEs (’92, p. 99) is so remarkable that I quote it in full. “In Herr GREDLER’s monastery, one of the monks had been accustomed for some months to put food regularly on his window sill for ants coming up from the garden. In con- sequence of Herr GREDLER’S communications, he took it into his head to put the bait for the ants, pounded sugar, in an old ink stand, and hung this up by a string to a cross piece of the window and left it hanging freely. A few ants were in the bait. They soon found their way out over the string with the grains of sugar and so their way back to their friends. Before long a procession was arranged on the new road from the window sill along the string to the spot where the sugar was, and so things went on for two days, nothing fresh occurring. But one day the procession stopped at the old feeding place on the window-sill and took the food thence without going up to the pendant sugar jar. Closer observation revealed that about a dozen of the rogues in the jar above were busily and unwearyingly carrying the grains of sugar to the edge of the pot and throwing them over to their comrades down below.”’ Remarkable? Yes; but before passing judgment, let us recall a few results of experiments. It is a common occurrence to have different ants of the same species doing different things at the same time. [| have had a portion of the ants of a colony busy conveying pup into an artificial nest, while the remainder were industriously excavating holes in the sand. Likewise, in my cardboard stage experiments, I have observed at one and the same time, one set of ants storing pupz under the incline, another set Turner, Homing of Ants. 423 conveying pupz to the nest, and yet another set piling the pupz in the center of the stage. Time and again, when ants were busy conveying pup from the stage to the nest, I have dropped pupa in the path of ants returning from the nest. I have tried this experiment with all the species of ants considered in this paper; and, without excep- tion, the pupa was picked up and carried into the nest. In the light of these facts, let us consider Herr GREDLER’s account. In this case the ants had learned the way from the sugar, by way of the string, wall, window-sill and wall to their nest. Now had a small number of these ants begun to pile the sugar on the edge of the bottle, from which most of it would surely have fallen to the window-sill, the case would have been similar to the one where one portion of one of my colonies conveyed pupe from the stage to the nest, while another portion piled the pupz at a certain place on the stage. But the falling of a portion of the sugar to the window-sill, which was a portion of the pathway of the ants returning from the nest, would introduce a modifying variation. As soon as a returning ant encountered a grain of sugar, it would pick it up and carry it to the nest. And if the grains fell from the rim of the bottle fast enough, we should soon have no ants taking the trip up the string. ‘hus yrauld be produced a behavior which simulates mutual cooperation. Since Herr GREDLER failed to observe the stages by which the contin- uous line of ants was transformed into two distant yet cooperating lines, this anecdote has no value as evidence proving that ants rationally cooperate. To sum up the results of this section, it is quite probable that division of labor of the type mentioned above is of rather common occurrence among certain ants; but until new data are forthcom- ing, we must consider all such cases as coincidences, rather than as proofs of rational cooperation. VII. CONCLUSIONS. 1. In their journeys, the movements of ants are not tropisms. 2. Ants are not as slavishly guided by the scent of their foot- prints as is usually believed, for all of the species examined by me could be trained to pass over at least twelve inches of an unscented path. ‘This discovery furnishes an easy means of investigating many problems in ant psychology. 424 ‘fournal of Comparative Neurology and Psychology. 3. Lussock was right when he said, “In determining their course ants are greatly influenced by the direction of the light.” 4. [The color of the pathway has no, or little, effect on the home-going of ants. ‘There are a few doubtful cases where the hue may have had some effect. “There are many in which pro- nounced changes in the brightness of the pathway seem to affect the ants. 5. In their wanderings, ants are influenced by olfactory (z.e., topochemical), optic, auditory, kinesthetic and tactile stimult. 6. Ants seem to have fairly definite impressions of direction in both horizontal and vertical planes, and also impressions of distance. 7. [hey are enabled to take long round trips by learning by experience and retaining what they thus learn. 8. ‘They have associative memory. g. Such cases of division of labor as RomaNEs—quoting from MoceripGE, Lespes, Bett and Herr GReEDLER—describes in his ““ Animal Intelligence,” are to be looked upon as cases of coinci- dence rather than as examples of mutual cooperation. 10. In their home-goings, ants display marked individual variations. 11. They are not guided by a homing instinct. ‘While conducting these experiments, | have made many observations, unrecorded in the body of the text, which show that WHEELER is right in emphasizing the high development of the female, for the winged females often take part in the regular duties of the nest. I have had them learn the way home from new situa- tions and assist the workers in carrying the pupz home. 13. [he males seem unable to solve even the simplest problems. 14. [he major workers of Pheidole, which ERNEstT ANDRE claims function as soldiers and do not take any active part in the ordinary work of the nest, frequently assist the workers in making excavations and, occasionally, assist in conveying pupe from one place to another. | have never noticed one continue to carry pupx for any considerable length of time. ‘This statement is based upon numerous observations which were omitted from the body of the text on account of lack of space. 15. Ants are much more than mere reflex machines; they are ee creatures guided by memories of past individual (onto- genetic) experience. Turner, Homing of Ants. 425 LITERATURE CITED. Berue, A. g8. Durfen wir den Ameisen und Bienen psychische Qualitaten zuschreiben? Abs. in Amer. Nat., vol. 32, pp. 439-447- ’oo. Noch einmal tiber die psychischen Ruditaten der Ameisen. Arch. f. d. ges. Physiol. (Pfliiger), 79 Bd., Hft. 1-2, pp. 39-52. ’o2. Die Heimkehrfahigkeit der Ameisen u. Bienen zum Theil nach neuen Versuchen. Biol. Centralbl., 22 Bd., no. 7, pp. 193-2153 no 8, pp. 234-238. BucuNner,L. ?80. Mindin Animals. Trans. by Annie Besant. London. Butret-REeren. ’oo. Sind die Bienen Reflex-maschinen? Experimentelle Beitrage zur Biologie der Honig- biene. Bzol. Centralbl., Bd. 20, pp. 97-109, 130-144, 177-193, 209-224, 289-304. CraparEDE, E. ‘ ’03. The Consciousness of Animals. Intern. Quart., vol. 8, pp. 296-315. Ernst, CHRISTIAN. ’o5. Einige Beobachtungen an kunstlichen Ameisennestern. Biol. Centralbl., Bd. 25, pp. 47-51; Bd. 26, pp. 210-220. Fieve, A. M. ’or. Further Study of an Ant. Proc. Acad. Nat. Sct. Phil., vol. 53, pp. 521-544. ’03. Experiments with Ants Induced to Swim. Jbid., vol. 55, pp. 617-624. Fietpr, A. M. anp Parker, G. H. ’o4. The Reaction of Ants to Material Vibrations. Ibid., vol. 56. Foret, AuG. ’o1. Die Psychischen Eigenschaften der Ameisen und eniger anderen Insecten. Tageb/. v. Intern. Zool. Congr. no 3, pp. 5-6. ’03. Ants and Some of Their Instincts. Monzst, vol. 14, pp. 33-66; pp. 177-194. Hosguovsse, L. T. or. Mind in Evolution. New York. The Macmillan Co. Huser, P. *10. Recherches sur les Moeurs des Fourmis Indigénes. Paris. Janet, Cu. ’94. The Production of Sound Among Ants. Abs. in Amer. Nat., vol. 28, pp. 270-271. Lanpors, H. ’67._ Die Ton- und Stimm-apparate der Insecten in anatomische, physiologiger und akustis- cher-Beziehung. Zezt. f. wiss. Zool., Bd. 17, pp. 105-186. Loes, JAcQUuEs. ’oz. Comparative Physiology of the Brain and Comparative Psychology. New York, G. P. Putnam’s Sons. ’06. The Dynamics of Living Matter. New York. Lupgock, Sir JouN. 81. Ants, Bees and Wasps. London. Kegan, Paul and Co. oth Edition. Moraan, C. Lioyp. ?oo. Animal Behavior. London. Edw. Arnold. Pieron, H. ’o4. Du réle sens musculaire dans lorientation des fourmis. Bull. Inst. Gen. Psyc. Paris. T. 4, pp. 168-187. Porter, J. P. ’04. A Preliminary Study of the Psychology of the English Sparrow. Am. four. of Psychol- ogy, vol. 15, pp. 313-346. Romangs, G. J. ’9g2. Animal Intelligence. New York. D. Appleton and Co. SANForRD, E. C. ’03. The Psychic Life of Fishes. Inter. Quart., vol. 7, pp. 316-333- SHarp, D. ’93. On Stridulation in Ants. Trans. Entom. Soc. London. 1893, pt. Il, pp. 199-213. Tuornpike, E. L. ’98. Animal Intelligence. Psychological Review, Monog. Suppl., vol. 2, no 4. VietMeYeR, H. ’oo. Beobachtungen iiber das Zuriickfinden von Ameisen (Leptothorax unifasciatus Str.) zur ihrem Neste. I/l. Zeit. f. Entom., Bd. 5, pp. 311-313- 426 “fournal of Comparative Neurology and Psychology. WasMann, E. ’g1. Zur Frage nach dem Gehorsvermogen der Ameisen. Bzo/. Centralb/., Bd. 11, pp. 26-27. Warson, JNo. B. ’03. Animal Education. Univ. of Chicago Press. We Lp, Le Roy. ’99. The Sense of Hearing in Ants. Science. n.s., vol. 10, pp. 766-768. WrouGuton, Rogert. 92. OurAnts. Four. Bombay Nat. Hist. Soc. Yerkes, R.M. ‘703. The Instincts, Habits and Reactions of the Frog. Psychological Review. Monog. Suppl. no. 17; Harvard Psychological Studies, vol. :. EXPLANATION OF FIGURES. In all of the figures the successive ordinates represent minutes of time. In Figs. 1 to 6 (see text, p- 382) each abscissa represents an experiment in the series tabulated and each ordinate the time in minutes that elapsed in that case from the beginning of the series until the first pupa was carried to the nest. These experiments were conducted on ants acting in concert, the purpose being to see if ants retain what they gain by experience. In the remainder of the figures the abscissa represent successive trips to the nest. When a line terminates in an arrowhead, that indicates a failure. In all curves any time less than a quarter of a minute is considered zero. EVAR Eaiie Fic. 1. Learning curve of Myrmica punctiventris Rog., acting in concert, showing how easily it is affected by even slight changes in the environment, and that it has not a homing instinct. I. June 21,7:30a.m. Apparatus, a cardboard stage connected on the right to the island by means of a cardboard incline (4). 2. Do. 2:45 p.m. Same apparatus as’before. The ants have had a rest of two hours and thirty- five minutes. 3. June 22, 11:08 am. Same apparatus as before. The ants have had a rest of twenty hours and twenty-three minutes. 4. Do. 2:35 p.m. Same apparatus as before. The ants have had a rest of one minute. 5. 3:56p.m. Same apparatus as above. The ants have had a rest of twenty-six minutes. 6. Do. 4:48 p.m. Same stage as before, but the scented incline is placed on the left side of the stage and a new unscented incline (B) placed where the scented incline had been. The pupe were all carried down the new incline. No larva were carried down the old incline; one worker started down the old incline and then returned to the stage. Occasionally, a worker would ascend the old incline from the island. The ants had rested one minute. 7. June 23, 10:40 a.m. Same apparatus as before. Workers carry the pupe down incline B to the nest; other workers convey pupe down incline 4 and store them under its foot, thence stragglers from the nest convey them tothenest. The ants had rested twenty-nine hours. 8. July 11, 9:15 am. Anew cardboard stage, a new incline and all (even the nest) was placed on anew island. The figure indicates a failure but does not indicate how great. Those ants were lost for over twelve hours. At 9:50 p.m. when I left the laboratory, they were resting quietly on the stage. Sometime before 8:30 the next morning they found the way to the nest (see text, p. 379). g. July 13,7:54.a.m. Same apparatus as above. The ants had rested a day and had had a chance to become acquainted with the island. 10. July 24, 9:15 a.m. A cardboard stage from which one incline passed to the island. A dark chamber opening upon the stage and also upon the incline, was placed over the top of the incline. All the pupz were stored under the dark chamber, none were carried to the nest. The workers went to the nest and left the pupz on the stage. The ants had rested eleven days and seventy-four minutes. Fic. 2. Learning-curve of Prenolepis imparis Say, showing that the homing of ants is not a chemo- tropism and that ants can learn a new way home when the way they once knew has been removed. I. June 24, 7:38 a.m. A new cardboard stage with a new incline (4) attached to the left side. 2. Do. 2:46 p.m. Same apparatus as above. The ants had rested three hours and forty-six minutes. 3. Do. 3:10 p.m. Same apparatus as above. The ants had rested ten minutes. 4. Do. 3:20p.m. Same apparatus as above. The ants had rested four minutes. TurNeER, Homing of Ants. 42.7 5. Do. 3:40p.m. Same apparatus as above, only Ladded incline B to the right side of stage, so that now incline 4 is on the left and incline B on the right. The ants carried the pupe down 4, none were carried down B. 6. Do. 3:58 p.m. Same apparatus as above. The ants had rested seventeen minutes. 7. Do. 4:20p.m. Same apparatus as before, only inclines d and B were made to change places. All the pupe were carried down B, none were carried down the scented incline d. The ants had rested one minute. 8. Do. 4:40 p.m. Same apparatus as above. No pupz carried down A. The ants had rested one minute. g. Do. 5:06p.m. Same apparatus as above. The ants had rested one minute. Io. June 26, 7:12 a.m Same apparatus as above. All the pupe were carried down B, a few unburdened workers occasionally promenaded on 4. The ants had rested one day, thirteen hours and forty-nine minutes. 11. Do. 7:42 a.m. Used the same stage but turned it bottom side up, thus making practically a new stage. The same inclines were used in the same position, 4 on the right and B on the left. All the pupe were carried down B, none were carried down A. The ants had rested two minutes. (An unbur- dened worker went down B to the nest one minute after the ants were placed on the stage.) 12. June 26, 8:11 a.m. Same stage, but inclines 4 and B are both removed and a new incline C placed where 4 had been; this left no incline at the place where the ants had been descending and an unscented incline on the other side. The ants pick up pupae and go immediately to the place where B had been. Failing to find B, they become much confused. They return over and over again to that point and reach down as though hunting for something. Finally they carry the pupe down C to the nest. The ants had rested twenty-nine minutes. 13. Do. 9:42 a.m. Same apparatus as above. Several workers convey pupe to where B had been. The ants had rested two minutes. 14. Do. 10:09 a.m. Same apparatus as above. The ants had rested two minutes. 15. Do. 10:30a.m. Same apparatus as above. The ants had rested one minute. 16. Do. 11:02 a.m. Same apparatus as above, only I added a new incline D to the left side of the stage. The majority of the pupe were carried down C; only two pupe were carried down D. The ants had rested four minutes. 17- Do. 11:24 a.m. Same apparatus as before, only C and D were interchanged, so that we now had the unscented path D in the path the ants had been traversing. The majority of the pupe were carried down D, only three were carried down C. The ants had rested nine minutes. 18. Do. 6:55 p.m. Same apparatus as above. The ants had rested one hour and two minutes. 19. Do. 7:18 p.m. Same stage, but I replaced C by D and placed a new incline E where D had been. The majority of the pupe were carried down D. The ants had rested two minutes. Fic. 3. Learning-curve of Prenolepis imparis acting in concert, showing lapse of memory, the effect of light rays, interruptions of the path, etc. I. June 27, 9:42 a.m. A new cardboard stage with new incline A descending from its left side to the island. 2. Do. 10:55 a.m. Same apparatus as above. The ants had rested two minutes. 3. Do. 11:20a.m. Same apparatus as above. The ants had rested one minute. 4. Do. 12:15 a.m. Same apparatus as above. The ants had rested thirty-five minutes. . Do. 12:30 p.m. Same apparatus as before, only there is a vertical gap of 1 mm. between the foot of the incline and the island. The ants had rested four minutes. 6. Do. 1:15 p.m. Same apparatus as before, but the vertical gap is increased to4mm. The ants had rested ten minutes (see text, p. 410). 7- Do. 2:35 p.m. Same apparatus as before, only the 4 mm. gap was horizontal instead of verti- cal. The ants had rested forty-two minutes. 8. June 28, 10:55 a.m. Same apparatus as before only the ants must cross a horizontal gap of 4 mm. to a pile of slides and down them to the island. The delay was caused by the gap. They went down to the gap and there hesitated, many returning to the stage. This was unexpected, since only yesterday they had been trained to cross a horizontal gap of that kind. The ants had rested twenty- two hours and twenty minutes. g. July 10,7:16a.m. A cardboard stage from which an incline ascends to thenest. The ants had rested thirteen days, twenty hours and eleven minutes. 10. Do. 7:55a.m. Sameasabove. The pupe had rested four minutes. 11. Do. 8:46 a.m. Used same stage and incline but rotated the stage through 180°. Incline in same position. This gave an unscented path from pupe to the incline. The ants went direct to the incline. The ants had rested one minute. 428 “fournal of Comparative Neurology and Psychology. 12, Do. 9:30a.m. Same stage, substituted incline B for A and placed incline 4 on the opposite side of the nest. The ants had rested five minutes. (At the beginning a worker went, unburdened, up incline B to the nest, but did not return to the stage.) 13. July 25, 7:30 a.m. A stage from which incline A descends from the left side to the nest. It was illuminated by a 16 c.p. incandescent lamp placed near to the off-side of the incline. The ants had rested fifteen minutes. Fic. 4. Learning-curve of Formica fusca var. subsericea Say, acting in concert; showing the effect of frequent changes of environment. I. July 1,8:30a.m. A cardboard stage with an incline attached to the left side. 2. Do. g:00a.m. Sameas above. The ants had rested ten minutes. 3- Do. 9:10a.m. Sameas above. The ants had rested one minute. 4. Do. 9:33 a.m. Same as above. The ants had rested one minute. 5. Do. 10:33 a.m. Same as above. The ants had rested three minutes. 6. Do. 11:00a.m. Same as above. The ants had rested one minute. 7- Do. 3:30p.m. The same stage used in the above, but a new incline B is substituted for incline A. Ais then preached to the opposite side of the stage. The majority of the pupz were carried down incline B. At first, occasionally, a worker would carry a pupa down A, but in each case that worker moved much slower than those moving down B. The ants had rested four hours. So DO 4:40 p.m. A new stage and anewincline. The ants had rested one hour. g. July 6, 11:24 a.m. A new stage with a new incline on the right side (the other inclines had been on the left side). The ants had rested four days, fourteen hours and thirty-four minutes. 10. Do. 12:28 p.m. Same apparatus as above. The ants had rested one minute. 11. July 6, 2:25 p.m. Same apparatus as above. The ants had rested five minutes. 12. Do. 3:00 p.m. Same stage, but incline 4 was placed on the opposite side of the stage and a new incline B placed where A had been. The pupe were all carried down B to the nest, occasionally a worker would carry a pupa partly down 4 and then return to the stage. The ants had rested thirteen minutes. 13. July 10,7:55 a.m. A new stage with a new incline attached to the right side of the stage. The ants had rested four hours and twenty-five minutes. Fic. 5. Learning-curve of Prenolepis imparis Say acting in concert, showing unexpected variations in the mode of response. . Lem Uy re sgt asm cardboard stage with an incline A attached to the left side. 2. Do. 10:46a.m. Same apparatus as above. The ants had rested one minute. a2 Do: 8:05 p.m. Same apparatus as above. The workers had rested over nine hours. 4. July 10,6:30a.m. Sameapparatus as above. The ants had rested eight hours and ten minutes. Ga Dor 7:55 a.m. Same apparatus as above. The ants had rested ten minutes. 6. Do. 9:35 a.m. Same stage, a new incline B is substituted for 4 and 4 is placed on the opposite side of the stage. The ants had rested ten minutes. Fic. 6. Learning-curve of Prenolepis imparis Say acting in concert, showing the effect of the direction of the rays of light and the result of placing a dark chamber over the top of the incline. 1. July 9, 7:06 a.m. A cardboard stage with incline 4 on left side. Zr 8:16 a.m. Same apparatus as above. The ants had rested one minute. go Dios 9:55 a.m. Same apparatus as before. The ants had rested one minute. 4. Do. 10:45 a.m. Same apparatus only the stage was revolved through an angle of 180°, but the incline was left in the same position. Thus there was an unscented path from the pupe to the incline. The ants had rested six minutes. 5. July 10,6:30a.m. Same apparatus as before. The ants had rested twenty hours. 6. Do. g:08 a.m. Same as above only the stage was revolved through 180°, 4 was left in the old place. Thus there was an unscented path from the pupe to the incline. 7- July 11, 6:28 a.m. Same apparatus as above. The ants had rested twenty-one hours. 8. Do. 7:21 a.m. Same apparatus as above. The ants had rested fifty minutes. g. July 12,7:46a.m. A new cardboard stage with a new incline leading down to the island. A19 c.p. incandescent lamp was placed near the incline. The ants had rested one day and twenty-five minutes. 10. Do. 8:53 a.m. Same apparatus as before. The ants had rested three minutes. 11. July 13,7:11 a.m. Stage with incline B on the right. 16 c.p. lamp on the same side as B. The ants had rested forty-nine minutes. Turner, Homing of Ants. 429 12. Do. 8:05a.m. Same apparatus as above, only a dark chamber opening upon stage and incline was placed over the top of the incline. One set of workers conveyed pupe to the dark chamber; another set, stragglers from the nest, conveyed them from the dark chamber to the nest. The delay represents the time that elapsed before a straggler from the nest discovered the pupe in the dark chamber. The ants had rested nine minutes. 13. July 16, 7:39 a.m. Same as above. The workers had rested two days, twenty-one hours and fifty-nine minutes. 14. Do. 2:07 p.m. Same apparatus as above. The delay was caused because they stored all of the pup in the dark chamber before any were carried into the nest. The ants had rested four hours and a half. Fic.7. Learning-curve of an individual Formica fusca var. subsericea Say, showing lapse of mem- ory, etc. Apparatus. Two cardboard stages, 7 and 2; inclines D and E leading to the island from each, respec- tively. The feet of the inclines were side by side. Each was attached to the right side of the stage. The marked worker was placed on stage 2 (see p. 412). A. Ontrips 6 and 7 the pupa was taken from stage 1 to the nest. B. Ontrip 12, do. C. The worker ascends and descends incline D, then ascends and descends incline E, then ascends to stage 2 and carries a pupa to the nest. D, On trip 36 the pupa was taken from stage I. E. On trips 43-45, do. F, Ontrip 50, do. hee oa - ‘ . ie a XVII, Plate II- Vol. Journal of Comparative Neurology and Psychology. Sea EgaESHEUS SSSn50 50 5050S55555EReTEEE EES etter fe 70 ue Y !1 | cag. 7 i “we 4 eal, % sae x Turner, Homing of Ants. 431 JIL AMIEID, IOUT Fic. 8. Learning-curve of an individual Myrmica punctiventris, showing its reactions to colored pathways. Apparatus, cardboard stage with a red incline leading down to the island; illuminated by diffuse daylight. While the worker was on the stage a new red incline was substituted for the old. Do. Green for red. Do. Purple for green. Do. Blue for purple. Do. Yellow for blue. White for yellow. Do. Red for white. Do. White for red. Do. Smooth black for white. Do. Black with velvety feel for smooth black, wea Th RO SA is) ° Fic. 9. Learning-curve of an individual Formica fusca var. subsericea Say, showing its reactions to odors encountered in its path. Apparatus, a cardboard stage with an incline passing down to the island. On this stage the pupe and a marked worker were placed. The worker had made several trips before the first trip plotted here. A, While the worker was on the stage, I substituted for the incline one with a transverse band of xylol across its middle. B. Lreplaced the old incline. C. While the worker was on the stage I substituted the incline with the transverse band of xylol. D. While the worker was on the stage the incline with the xylol band was transferred to the control stage and allowed to remain there eight minutes. The ant used for control was much disturbed. Dur- ing the eight minutes it mounted the incline several times, but would not cross the band. E. The incline with the xylol band is reattached tostage 1. During the eight minutes that have elapsed the marked ant has made several trips down an unscented incline which was attached to the stage when the incline scented with xylol was removed. F. The incline with the xylol band is transferred to the control stage and causes the same disturb- ance as before. Fic. 10. Learning-curve of a Myrmica punctiventris, showing its reactions to light rays and to tactile stimuli. Apparatus, a cardboard stage with an incline A passing from left side to the island. Near the left side of the stage was placed a 16 c.p. incandescent lamp. The room was darkened. a. While the worker was on the stage the 16 c.p. lamp was transferred to the right side of the stage. b. While the worker was on the stage, the incandescent lamp was extinguished and the curtain of a window thrown up so as to illuminate strongly the left side with daylight. c-d. Experiment was interrupted for six minutes. d. The same stage, but illuminated uniformly by diffuse daylight instead of by an incandescent lamp in a special position. A smooth black incline leads from the stage to the island. e. While the worker was on the stage the white incline was substituted for the black. f- Do. Smooth black incline for white. g. Do. Black incline with velvety feel for the smooth black. Fic. 11. Learning-curve of an individual Formica fusca var. subsericea Say, showing its reaction to a colored pathway. Apparatus, a cardboard stage with red incline leading down to island, illuminated with diffuse day- light. a. While the worker was on the stage a new red incline was substituted for the old. The paper on this incline was somewhat wrinkled. b. Do. Green for red. Fic. 12. Learning-curve of an individual Formica fusca var. subsericea Say, showing effect of a jar upon its behavior, etc. Apparatus, a cardboard stage with a white incline leading from its right side to the island. A heat- filter composed of a tall rectangular museum jar filled with water was placed near both the right and 432 ‘fournal of Comparative Neurology and Psychology. left sides of the stage. Behind the heat-filter on the right was placed a 32 c.p. incandescent lamp. Excepting for this light the room was dark. A. Placed a new white incline on the opposite side from the first incline. B. While the worker was on the stage the 32 c.p. lamp was placed behind the heat-filter on the side where the second incline was. C. While the worker was in the nest the light was replaced on the same side as the first incline. D. While the worker was on the stage I substituted a new white incline for the old. In doing so the stage was jarred considerably. E. Do. White for black. F. Replaced the worker on the stage. G. While the worker was on the island, I substituted a white incline for the black. H. Do. New white for the old. Fic. 13. Learning-curve of an individual Myrica punctiventris, illustrating one of those occasional cases in which it took the ant much longer to make the second trip to the nest than it did to make the first. Apparatus, two cardboard stages, each with an incline leading down to the island. Pupz are placed on both stages. The marked worker is placed on stage number one. Vol. XVII, Plate III. Journal of Comparative Neurology and Psychology. B Ly TO ATT TT N TTY I ASCE ER 000S8a Ove Rey. a Eig. IO Le Lh LUe, ueguaue Z | SS90088 cs | Ges PAG 12 Fig. é i aa , y ; Asia, eh ay ; li = 4 . : eas ae ty "A Rs ene . * f Yh cae id a ), branch from limbus towards manubrium. Zeiss comp. oc. 4; obj. 16 mm. Fic. 2. Nerves (7) entering at inferior posterior quadrant of right membrana tympani of dog, from external auditory meatus. mm, position of limbus; x, zonular plexus. Zerss comp. oc. 4; 0bj.8 mm. Fic.3. Pars flaccida (p.f.) and pars tensa (p. 2.) of dog, external surface, after malleus dissected off from within. 7, nerves entering from external auditory meatus; g.p. ground plexus; p, plica membrane tympani; c, branching cells in epidermis which readily stain with methylene blue. Zr1ss comp. oc. 4; obj. 8 mm. Fic. 4. Right membrana tympani of cat and adjacent part of middle ear, viewed from within. Through the transparent membrana tympani, nerves (”) are seen entering from external auditory meatus all around the periphery, breaking up at limbus (/) into zonular plexus with branches, of which only a few are shown (.c.), to mucous membrane of middle ear and (.t.)to membrana tympani. With this low power no branch can be distinguished passing from tympanic cavity to membrana tympani. ZxIss 0¢. 43 Obj. a. Fic. 5. External surface of membrana tympani of monkey (Macacus rhesus) after removal of sur- rounding parts, but still adherent to its bony attachments. The manubrium is seen through the mem- brane and the capitulum through the thin bone of external auditory meatus. Nerves (7) are seen enter- ing over pars flaccida and posterior plica, and at limbus. a, principal artery to membrana tympani; pp-, plica posterior. Zeiss oc.; obj. ao. Fic. 6. Part of left membrana tympani of monkey (Macacus rhesus), showing nerves entering at upper segment and forming ground plexus (g.p.); a—a’ indicate line of principal artery which divides at a’ to send branches to each side of manubrium. Zetss oc. 4; obj. 8 mm. Fic.7. Pars flaccida of dog immediately above plica posterior. Two nerves (” and n’) faintly medullated at first which end in interlacing arborizations in fibrous tissue under cuticular epidermis. ZEIss oc. 4; obj. 2 mm. Fic. 8. The form of branched-ending seen most commonly under cuticular layer of pars tensa in the monkey. The nerves are nonmedullated. Zeiss. oc. 4; obj. 2 mm. Fic. 9. Nerve (n) dividing and forming plexus in connective tissue of pars tensa of dog. Zetss oc. 4; obj. 8 mm. Fic. 10a and b. Two varieties of nerve endings in fibrous tissue of pars tensaof dog: (a) Nerve (1) divides into two branches each of which, after making several corkscrew turns, again divides to sur- round a cell which appears in every respect similar to the neighboring connective tissue cells. (b) Nerve (1) gives off branches which coi! round each other or the parent stem forming a long narrow end- ing. ZeIss comp. oc. 4; obj. 2 mm. Fic. 11. Cuticular layer of pars tensa of monkey, superior posterior quadrant. (a) Subepithelial plexus with endings in epidermis; (6) Nerve branching and interlacing with subepithelial plexus. ZerIss oc. 4; obj. 2 mm. | eh. ae by re he oe AND PSYCHOLOGY. VOL. XV AKOEN @ CO, BALTIMORE K. Hill del. a he arent Seam wat Si Des tUDY OF THE DIAMETERS OF THE CEELS AND NUCLED VIN VEE “SECOND CERVICAL, -SRINAL GANGLION OF THE ADULT ALBINO RAT. BY SHUNGTSEM: EVAMy Ales he: (Associate in Neurology, The Wistar Institute.) From The Wistar Institute of Anatomy and Biology, Philadelphia. With Four Ficures. INTRODUCTION. It is generally believed that the spinal ganglion contains sey- eral types of nerve cells which can be morphologically differen- tiated from one another. ‘The varieties of cells whose existence in the ganglia have been repeatedly confirmed are: (1) cells with a T- or Y-shaped division of the processes. Such cells are consid- ered to be most abundant and to be both large and small in size. (2) DoctEt’s cells of second type, multipolar cells, and (3) multi- polar cells which resemble in shape and structure sympathetic ganglion cells. “The most complete classification is based on the study of methylene blue preparations. In a general way the pres- ence of the several varieties of cells in the ganglia may also be demonstrated in ordinary paraffine sections treated with any of the basic dyes followed by a counter-stain. In such preparations one can easily distinguish cells of different sizes as well as those exhibiting different arrangements of the stainable substance. These two characters, size and arrangement of stainable substance, have been used as a criterion by several investigators in order to classify these cells. By this method Lucaro (’96) distinguishes in the dog five different varieties of the spinal ganglion cells, LENHOSSEK (96) in the human spinal ganglion distinguishes three varieties, Cox (98) in the spinal ganglion of the rabbit, two main varieties, and the author (’or) using the same criterion has distinguished three varieties in the spinal ganglion of the albino rat. It is my 470 Fournal of Comparative Neurology and Psychology . intention later to analyse in detail all these classifications and at the present moment it is merely necessary to call attention to the fact that in the spinal ganglion several varieties of cells have been distinguished from one another. Can all these cells of different varieties be considered as belonging to a single class or are there really several types of cells composing the spinal ganglion? In other words, a frequency of distribution of all these cells based on their sizes' should give us more than one mode if there were more than one type of cell involved. If but one mode appears we have good ground to conclude that all these cells, though differing in size as well as in structure, may be considered from the stand- point of size, as members of a homogeneous population. ‘The differences in structure are for the moment neglected and must form the subject of a special study. MATERIAL AND TECHNIQUE. For the present investigation the second cervical spinal ganglion of the adult albino rat was employed. ‘The second cervical gan- glion was purposely selected since through the investigation of Ranson (’06) we have already some numerical data in regard to this particular nerve. The second nerve with ganglion was removed from right side of a healthy male having a body-weight of 194 grams and was fixed with osmic acid. Following the usual procedure the sections of the ganglion were cut 12 micra thick and mounted in series. Three sections from the middle of the entire series of 80 sections and three sections from midway between the middle and end on both sides, thus making altogether nine sections, were chosen. ‘These nine sections were selected for the measurement of the cells and nuclei on the assumption that the cells of the different sizes were uniformly distributed and consequently that the nine sections would adequately represent the total cell “population” of the ganglion. The measurements obtained from each cell and its nucleus were recorded on a separate card. In every case two maximum diameters at right angles to each other were determined for both cell-body and furclewe, by means of the ocular micrometer. The ' Although this point could be tested also from the standpoint of the structure, nevertheless it is very dificult to obtain numerical data in terms of the structure, suitable for biometric treatment. Hatat, Spinal Ganglion Cells of Rat. 471 values of the two diameters thus obtained were multiplied together and the square root of the product was called “calculated diam- eter” of the cells and nuclei. Of course every section of a cell which possessed a distinct nucleus and nucleolus was measured from the nine sections and altogether 1108 such cells were found. The 1108 cells and nuclei thus measured were arranged according to the magnitude of the “calculated diameters” and the fre- quencies of the variates were determined as shown in the corre- lation table (see p. 490). For grouping the variates I have selected 2 micra as the unit for the cell-body and 0.65 of a micron for the nucleus. As will be seen later, small differences in the value of the unit do not produce any significant change in the final results, and therefore it is advisable to select some integral number for convenience in computation. ANALYTICAL CONSTANTS AND FREQUENCY DISTRIBUTION. I shall first discuss the frequency distributions of the cell-bodies and nuclei. The fundamental analytical constants necessary for such discussion are given in the following table: TABLE I. | | . | ; Cell-body. Nucleus. _ || Cell-body. | Nucleus. IND (OPY | Le 5 | ere | Skewness. ment. 1108 1108 | 4081 .024 | .1734+.024 | tes ye Uy] 9.5254 | Modal divergence | 1.3558 .006 | .4918+ .002 [4 446.9891 233-3303 || Standard deviation | 6.6448-+ .952 ee 8457: .026 | fi 5486 1731 | Mean 23 -3356+1.346 | 10.9523 .037 | | Bo 3 6687+ .099 | 3.58894 .099 | Mode 21.9798 | 10.4605 /B, 7407+ .052 | .4161-+.028 | Coef. of variation |28.4749+0.4398 16.8521-+0.2482) B2-3 .6687 5889 | Coef. of correlation .8616+ .006 Ki — .3084 6585 || Lower and upper 7.5844 | || ends of ranges 60.9693 | Ke —1.5179 8248 } Type of curve I | IV I | It has been shown by Pearson (’05) that in order to fit a given distribution of frequency to a Gaussian probability curve 472 “‘fournal of Comparative Neurology and Psychology. the following conditions, within the limit of random sampling, must be fulfilled: V BB + 3) = ©) By — 6 8, = 9 By examining the analytical constants given in Table I, it is seen that all those constants for both the cell-bodies and nuclei are considerably greater than zero even when their respective Nie V 8, =0; 2 — 3 =0; 0; and d, modal divergence =o. 210 200 1g0 180 170 160 150 140 130 120 IIo 100 go 80 70 60 50 40 30 20 s fe) 8 12) 16520 24 28 32 36 40 44 48 micra. Fic. 1. Frequency polygon and fitted curve for variation in the diameters of the cell-bodies. probable errors are considered. ‘This at once leads to the con- clusion that in both cases the frequency distribution can not be represented by the normal curve. Furthermore a considerable deviation of those constants from zero, 7. e., skewness, as well as modal divergence, indicates that they can never be represented by any other symmetrical curve since the deviation in excess and Hata, Spinal Ganglion Cells of Rat. 473 defect are not equally probable. It is therefore evident that n order to represent the data in hand we must find a curve which is able to represent the odds against any given deviation. It has also been shown by PEARSON (’95, ’or) that the assign- ment of a given distribution of frequency to any one of the six 250 240 230 220 210 200 190 180 170 160 150 140 130 120 110 100 4-75 6 8 TO) 12) 14) 16) 3) emuicra: Fic. 2. Frequency polygon and fitted curve for variation in the diameters of the nuclei. types of his skew curves depends on the value of the analytical constants x,; k,3 8, and @,. As is shown in Table I, in the case of the cell-body we have the following relations: KOs Kk, << O) anda, . 10 474 “fournal of Comparative Neurology and Psychology. These three conditions satisfy PEARSON’s skew frequency curve of Type 1, while for the nuclei we have ky > OF 2 0558, > Qo andig 0 and)< 41, which calls for PEARson’s curve of Type 4. The frequency distributions and their fitted curves are shown graphically in Figs, 1 and 2. The equations for the curves are: For the cell-bodies (Type 1)? Me a (: een sie (: e x bie ar iran 6.2286 62.3049 origin at mode. For the nuclei (Type 4) y = 13.0825 (Cos al e11.20380 origin at 7.4341 micra, X = 1.0002, aan 0 Examining Figs. 1 and 2 we see at once that the theory and observation do not agree at all well. ‘The theoretical curve in both cases considerably underestimates the observed ordinates for the smaller values of variates x, and overestimates the same for the larger values of x. The degree of deviation between the observed and theoretical curves is most pronounced at or in the neighborhood of the mode. ‘This unexpected results forced the writer to reinvestigate the following points: 1. Since the spinal ganglion contains various sized cells it may be possible that these cells are not uniformly distributed from ? Original formula of Type 1 is given by my m2 M », Mo yan (1+*) () , where yy =“, my, limg 2 a, ao b Gams m, + my a DP (m, + mz + 2) : IT (my, + 1) I" (mm, + 1) and for Type 4 2 y= yo (Cos) v6 h ae , where y, ==. — a Hatrat, Spinal Ganglion Cells of Rat. 475 section to section. In other words, some sections may contain relatively more cells having larger or smaller diameters, therefore the nine sections selected might not give a proper representation of the entire cell population and therefore the 1108 cells measured might not constitute a real random sampling of the entire popula- tion. 2. The disagreement between the theory and observation may be due to an improper selection of the unit for grouping the variates. If so, it may be improved by taking some other unit. 3. The two curves may agree more closely if the uncorrected or raw moments about the mean were used in determining various analytical constants. 4. The spinal ganglion cells may not represent a homoge- neous population, but a mixture of various groups of elements. If so, dissection of the frequency curve into several components should give a better agreement. The question contained in point 1 has been answered in the following way: the nine sections were divided into three series, each represented by three sections. For the series 1, one section was taken from the middle and one from the midway between the middle and the extremes on both sides, while for the series 2, the three sections which lie to the right of the three sections of the series 1 and for the series 3 those which lie toward left. The percentage values of these three series just mentioned were plotted separately, the same unit of course being used for each series. Comparison shows that these three curves agree with each other in every minor detail, and therefore with the original curve, too. [his means that the cell-bodies of different sizes are uniformly distributed, otherwise the three curves should not agree so closely. Te fore the 1108 cells here measured can be regarded as giving a true representation of the entire cell-population. ‘The disagree- ment found between the theoretical and observed curves is con- sequently not due to a lack of uniformity in distribution. The question contained in point 2 was also answered by tak- ing three different units for the cell-bodies; 1, 1.8, 2 micra; and one different unit for the nuclei, 1 micron, and comparing the new results with those already obtained. It was found that these variations in the units did not make any significant alteration in curve. his proves then the difficulty is not due to an improper choice of the unit. 476 “fournal of Comparative Neurology and Psychology. As to point 3, I have treated the data for the cell-bodies, using the uncorrected moments, but without effecting any improvement on the results. Finally I have tried to split the observed frequency curve (point 4) into two normal curves according to the method given by Pearson (’94). After laborious calculations it was found that the present data can not be split. The reason for this conclusion is omitted since it needs an elaborate mathematical presentation. The result shows however that there is not the slightest indica- tion of separate groups in the cell-population. - Therefore the cause of disagreement must depend on other conditions than those already enumerated. After failing to obtain in this way a reasonable explanation for the considerable deviations between the observed and theoretical Fic. 3. Diagram of the spinal ganglion cell containing nucleus and nucleolus. curves, it occurred to me that the explanation might be found in the method of sectioning and measurement. In order to make clear the relations existing within the ganglion let us suppose that 8000 spinal ganglion cells of various forms (from spherical to oblong) are thoroughly mixed in an ovoid receptacle. This is then cut into 80 slices of equal thickness. ‘The entire series is sampled by taking three slices from the middle and three slices from the midway between the middle and extremes on both sides. Thus nine slices are selected for examination. ‘The slices of the ganglion which we have examined in this way contained 1108 cells. Under these conditions the knife cuts the individual ganglion cell Hatat, Spinal Ganglion Cells of Rat. 477 in various planes. In some cases the knife makes a right angle with the longest axis (dc, Fig. 3) of the cell-body and some cases with the shortest axis (ab). In the remaining cases the angles will always be less than ao°. It must be remembered that the 1108 cells counted are those acl contain both nucleus and nu- cleolus, and therefore it is assumed that the knife always passed through, the approximate center of the cell-body. The chance that the knife will make a right angle with shortest axis* must however be very small compared with a failure. Whichever plane we cut, as long as the knife passes through the center, the diameter (ab) is constant and therefore the product of the diame- ters varies directly with the changes in the longer. But (cd) is the maximum diameter and its length diminishes as the axis moves from the original position fon the axis (ab). As soon as it reaches (ab) it becomes minimum. As was stated already, there are more chances for the knife to pass through somewhere between the two points (a) and (c) than to cut through (c) itself. If there- fore we determine the two diameters of the cell from the cut surface and the square root of their product is taken as the mean or “cal- culated diameter” of the given cell, the final value thus obtained willoften be less than it should really be, because 1” ab x dcis always > V ab Xx xy, where (xy) is any arbitrary line between (a) and (c). But as was stated already, the sections of the smaller cells are more nearly circular in outline. ‘Therefore the mean diame- ters thus obtained may represent nearly the true value in the case of the smaller cells but less nearly, in the case of the larger cells which have become ovoid. From this it will be clearly seen that the frequency curve of the diameters of the ganglion cells based on the square roots of the product of the observed diameters can not represent the true frequency. As to the range of the diameters, the maximum “calculated diameters” found may be considered to be the true value since there is at least one chance that the knife could pass through the longest axis while on the other hand the observed minimum diameter may be somewhat less than the true minimum diameter since there 1s a tendency for the diameter of those cells which are in any degree ovoid to be made smaller. Consequently if the values ‘ 3 The maximum size of the cell is obtained only when the knife makes go° with the axis (ab), pro- vided it also passes through the center of the cell. 478 ‘fournal of Comparative Neurology and Psychology. found for the small cells were nearly right but those found for the large cells were less nearly right it is evident that we should expect to find more cells showing smaller diameters than are actually present in the nine slices examined. ‘This leads at once to the conclusion that the ordinates representing the cells of the small diameters must be compounded of two heterogenous elements; the true small cells plus those cells which are artificially made smaller by the method of section. In the same way the ordinates for the larger cells will be compounded of the measurements on the larger cells minus those whichwere artificially made smaller. Thuswe shall find an excess of cases towards the smaller value of x and a deficit towards the larger value of x. This is just what we have ob- served. It was found, when the observed polygon was compared with the theoretical curve that the latter considerably underesti- mated the observed ordinates which correspond to the smaller values of x and overestimated the observed ordinates which corre- spond to the larger values of x. “The facts mentioned above indi- cate that the deviations of the observed polygon from the theoretical curve are mainly, if not entirely, due to the method of section. The greater observed excess found in the neighborhood of the mean or mode Is interesting, since it may be assumed that up to the neighborhood of the mean or mode the cells remain nearly spherical while beyond this region the increase in size is accom- panied by a change in shape. If my argument is correct, then we should expect the greater percentage deviation of the two diameters to appear more fre- quently towards the larger abscissal values. Although it would seem at first easy to test such hypothesis by taking the percentage deviation from the averages corresponding to the different abscissal values, nevertheless in practice we meet considerable difficulties. Any one who is familiar with the sections of the spinal ganglia prepared by the usual fixation methods, will recall that there are present a number of small cells with very unequal diameters. Such cells are most abundant along the periphery of the sections. The general outline of such cells is either rectangular, instead of curved, or the opposite boundaries are represented by nearly par- allel lines. ‘The cause of the deformation may be attributed toa shrinkage of the capsule of the spinal ganglion itself. On account of the presence of such deformed cells a mere comparison of the . Hatat, Spinal Ganglion Cells of Rat. 479 percentage deviations of the two diameters is unsatisfactory. Although it may not be a conclusive test, yet remembering that the spherical or nearly spherical cells should occur more frequently either towards the negative side or at the mode than towards the positive side, a determination of the relative frequency of such spherical cells, or the cells with nearly equal diameters, may be employed. Under the circumstances I think this is the only feasi- ble method of testing this point. From an examination of original data, it has actually been found that such spherical or nearly spherical cells diminish with the increasing calculated diameter and increase with a diminishing calculated diameter. Even without any further test we cannot doubt from the theoretical standpoint that the method of section diminishes the diameter of the large cells, thus artificially increasing the frequencies of the small cells. If this fact just mentioned is accepted, the conclusion follows that the theoretical curves may be considered as satisfactory rep- resentations under the circumstances and also may be considered much truer representations of the frequency distributions of the cell-bodies and nuclei than that shown by the actually observed data. Since the curve of Type 1 has limited range in both directions, we find from the constants that lower limit of range = 7.58444 and upper limit of range = 60.9693, while the observed limits are 7.8 micra and 47.4 micra, respectively. We see therefore that the theoretical lower limit agrees very closely with the observed, while the upper limit in the theory is consider- ably higher than that of the observed. But that this upper value may not be entirely improbable is indicated by my previous work (02) on the spinal ganglion cells where I find cells in the fourth cervical spinal ganglion of the adult albino rat as large as 52.7 micra. ‘This figure just given is the average for the three largest cells observed, therefore one or two individual cells must be still larger. Nevertheless it is not necessary to assume that these cells are the largest which could be found. ‘This fact indicates that there is a tendency at least to approximate the values given by the theoretical curve. 480 “fournal of Comparative Neurology and Psychology. MEAN, STANDARD DEVIATION, AND COEFFICIENTS OF VARIATION. 3(V. }) In the equation 4 = See where 4 represents the mean, it will be clearly seen that the absolute value of mean (4) varies directly according to the greater or smaller number of frequencies associated with the smaller or greater values of VY, as long as “n,” the total number of variates, is constant. We have demonstrated above that the number of the observed frequencies of V for both cells and their nuclei cannot be considered as the true frequency owing to’the method of section. ‘The true frequencies for the smaller values of V should be the observed frequency minus those cells which have been transferred from the group of large cells, while for the larger values of V it should be observed frequency plus those cells which have been thus transferred. Consequently the mean values actually found for the cell-bodies and nuclei must be consideredas smaller than they should actually be. However we cannot determine at the present moment how large the true mean values should be, owing to the difficulty of determining the number of the cells and nuclei which are assumed to have been transferred. On the other hand, the values for the standard deviation and for the coefficients of variation in the present case should be smaller than those found, since following an increase in the frequencies towards the larger values of V the resulting frequency distribution would become more regular than they are shown to be by the ob- served polygons and consequently the mean square deviation would become smaller. Diminution in the mean square deviation causes a reduction in the value of the standard deviation and conse- quently in the value of the coefficients of variation. As a matter of fact, we found the value for the standard deviation as well as the coefficients of variation decidedly larger when compared with apparently more variable characters. For example PEARL (’05) found the coefficient of variation in Paramecium from 8 to g per cent and in Arcella 10 per cent (PEARL and Dunsar, ’03) while in the present case that of the cell-body is as high as 28 per cent and that of the nucleus 17 per cent. Although we have not as yet any available data with which directly to compare our own, nevertheless our own values appear too great when they are compared with the coeflicients of variation obtained from the measurement of highly variable organs like the weight of the Hatat, Spinal Ganglion Cells of Rat. 481 liver (21 per cent, GREENWOOD, 704), weight of the body (10 per cent, PEARSON, ’97), weight of the heart (18 per cent, GREENWOOD, 704), etc. I therefore corrected the values of the mean, standard deviation, and coefficients of variation, assuming that the theoretical curves represent more nearly the true distribution of frequencies. On employing the values of the theoretical ordinates there was found for cell-bodies, mean, 28. 5948 micra, standard deviation, 14.8824 micra and coefhcient of variation, 18.36 per cent; while for the nuclei, the mean was 13.0535 micra; and standard deviation, 1.7929 micra; the coefh- cient of variation being 13.73 per cent. When these corrected values are compared with uncorrected ones we find an increase of 3 micra for the mean in both the cells and the nuclei, and a reduction by Io per cent in the coefficient of variation in the case of the cells and a reduction by 4 per cent in the case of nuclei. These corrected values appear to be the more probable, and are the best we can obtain until some further means of correcting the raw observations have been found. CLASSIFICATION OF THE SPINAL GANGLION CELLS. The unavoidable modification in the size of the spinal ganglion cells due to the method of sectioning as here described suggests a revision of the classification of the cells so far as it depends on their observed sizes. It has been mentioned already that using the size of the cells and the arrangement of the stainable masses as criteria, several investigators have attempted to classify the cells composing the spinal ganglion. ‘Three such classifications pro- posed by Lucaro, LENHossEK and Cox will be presented in detail. Lucaro (’96) distinguishes in the dog five different varieties of the spinal ganglion cells: 1. Large cells with delicate, closely packed stainable masses which are distributed uniformly throughout the cell-body. Around the nucleus are large stainable masses closely packed. ‘The nu- cleus is large and clear and is provided with a nucleolus. ‘These cells appear to be numerous. 2. Clear, medium-sized cells with irregularly formed small and large stainable masses which are large atthe periphery. Even here we see that individual masses are not isolated but are united 482 ‘fournal of Comparative Neurology and Psychology. together by fine processes. ‘The nucleus is clear and possesses anucleolus. ‘These cells are most numerous. 3. Small, dark cells with numerous small stainable masses lying in the region of the nucleus. ‘The ground substance becomes _diffusely stained. The nucleus also stains diffusely and con- tains two or more nucleoli. ‘These cells rank third in point of number. 4. Small or medium sized clear cells with large stainable masses which are present in small numbers and connected with each other by processes. ‘The nucleus frequently possesses more than one nucleolus. ‘These cells are not numerous. 5. Large clear cells with long drawn out masses which are continuous with one another and which arrange themselves in concentric lines around the nucleus. ‘These last cells present a laminated appearance like the cross section of an onion. ‘These cells are least numerous. LeNuHosséK (’96) in the human spinal ganglion distinguishes three varieties. 1. The first variety consists of cells with a light staining ground substance only. ‘These, which are the largest cells, have a pale ground substance and less numerous, loosely arranged stainable masses, which are most dense around the nucleus. 2. To the second variety belongs coarsely granular cells (grobscholligen Zellen), the appearance of which depends on the arrangement of the stainable substance, and most of the cells in the ganglion belong to this variety. ‘These cells are of medium size, but sometimes small and rarely very large. 3. The third variety contains small cells which have a peculiar internal structure. ‘These cells stain darkly because of the den- sity of the ground substance. Cox (’96) distinguishes in the spinal ganglion of the rabbit two main varieties. 1. One variety contains larger or smaller irregular masses of stainable substance, which do not show a distinct concentric arrangement. ‘The cells of this variety may be either large or small. 2. The other variety contains large, irregular masses of stain- able substance arranged concentrically. It will be clearly seen from the description given by these authors that there exist some structural characters common to both large Hata, Spinal Ganglion Cells of Rat. 483 and small cells. That is to say, some small cells have characters possessed by large cells and therefore size is the only means of dis- tinguishing two forms. ‘There is however another group of the small cells (third variety of both LuGARo and LENHOsSsEK) which exhibit still different structural characters. “They are much darker in appearance owing to a strong afhnity for staining reagents. The arrangement of the stainable substance is irregular and indis- tinct. The cell-outline is irregular. Thus there is no question as to the presence of the two kinds of the small cells which differ in both structure and shape from each other. ‘The entire series of small cells which exhibit a resemblance to the large cells were considered by Cox, LENHosséK and LuGaro as early formed cell elements which persist in the spinal ganglion as such in small size. I have however just shown that a considerable number of the large cells are made smaller artificially by the method of sec- tioning. One would therefore expect to find a number of the small cells similar in structure to the large cells, except that the arrangement of the stainable substance may differ slightly accord- ing to plane of section. The cell-outline of the majority of the “artificial” small cells should be nearly spherical, unless they are distorted. ‘Therefore the existence of small cells with the internal characters of the large cells can be explained readily on the assump- tion that they are in part if not entirely those large cells modi- fied by the method of sectioning. I therefore conclude that a majority of these cells with the eee of the large cells do not preéxist as such and that consequently the conclusions of LuGaRo, LrENHOSSEK and Cox are to this extent misleading. While the writer was engaged in the study of the structure of the spinal ganglion in the albino rat (Harat ’o1) the following groups of cells were recognized and described. ‘The one group is larger in size and stains lightly with eosin or erythrosin, while another group is smaller in size and stains deeply with eosin or erythrosin. Still a third group which, although it agrees in stain- ing reaction as well as in an irregular outline with the small cells, nevertheless differs in the arrangement of the stainable substance and in size. ‘The size is slightly larger on the average than that of the small deeply staining cells but much smaller than large cells. It now seems better to consider the group intermediate in size as a variety of the small cells rather than as a distinct type. The fol- lowing are the reasons for this conclusion: 484 Fournal of Comparative Neurology and Psychology. 1. The intermediate sized cells agree with small cells in two important characters, the cell-body stains deeply with eosin or erythrosin and the cell-outline is irregular. 2. Since the arrangement of the stainable substance is rather unstable its difference has significance only when other characters also differ from any other given group under consideration. If the recently proposed hypothesis by Scotr (’05) that the stain- able substance is identical with the zymogen granules of the pan- creas turns out to be true, then the size and form of the granules as well as their distribution may vary considerably according to the functional condition of the cell. Consequently we have in the spinal ganglion two forms of cells; one which stains deeply and the cell-outline of which is irregular. Such cells are usually small in size. The other the cell-body of which stains lightly the cell-outline being regular. Such cells range from small to large in size. It must be remembered how- ever that the entire cell-population when they are grouped accord- ing to their sizes grade from smallest to largest without showing any interruption. ‘This means of course that there is no definite demarcation line to divide the large from the small cells or vice versa. Asa matter of fact, the cells which stain lightly and which also exhibit regular outlines are by no means constant in size. This is also true for the group of the cells which stain deeply and which exhibit an irregular outline, although they are more uni- formly small. For this reason, the size of the cell-body is not a proper criterion by which to classify them. The writer has adopted the staining reaction of the cell for classification because, as has been mentioned above, the ganglion cells fall readily into one of the two classes: that is (a) those with a deeply stained cell-body with irregular cell-outline and (b) lightly stained cell-body with regular cell-outline. Although we should expect to find intermediate forms, which must always be present, nevertheless grouping by this method 1s more definite and practical than by the size and is much simpler than by those proposed by other investigators. In addition, these different histological characters are undoubtedly associated with different physiological states (HarTatr ’O1). I have chosen from Nrsst’s nomenclature two terms by which to designate the two groups of the ganglion cells with the idea that they may aid description. Hatal, Spinal Ganglion Cells of Rat. 485 1. Pycnomorphic cells, those cells which appear darker owing to a stronger afhnity to the staining reagents. ‘The cell-outlines are irregular. Such cells are usually small in size. 2. Apyenomorphic cells, those cells which appear pale owing to weaker affinity for the staining reagents. ‘The cell-outlines are regular, being either spherical or oblong. Such cells range from small to large and include those which are made artificially smaller owing to the method of sectioning. Attention is called to the fact that the above classification does not modify our views concerning the existence of three histological varieties of cells recognized in the spinal ganglion (see p. 1) but merely shows that these varieties do not distinguish themselves by their diameters in such a way as to form separate groups under this method of examination. ON THE CORRELATION BETWEEN CELL-BODY AND NUCLEUS. The intimate physiological relations existing between the cell- body and the nucleus suggest that there may also exist a definite size or mass relation between these two structures. Generally speaking in the growing spinal ganglion cells (Harat ’o1), the cell-body grows much faster than the nucleus. It was my object to determine this mass relation, between the cells and nuclei and if possible to find some mathematical expressions by which such relation could be concisely stated. The correlation table (Table II, p. 490) furnishes us all the data necessary to determine sucha relation. The table shows the range of variates in one character corresponding to that in the other. The coefficient of correlation would be then a numerical expression of the occurrence of the several values of x in one character in association with the several values of y in the other. PEARSON gives the formula for obtaining the coefhicient of correla- tion in the following form: a Geseya) pee aes AOS 11. Dy 05 Using the above formula, the value of r (coefficient of correla- tion) was found to be 0.8616 +0.0055. ‘This shows that the size of the cell-body is highly as well as positively correlated with the size of the nucleus. Therefore we infer that the larger cell-body is associated with larger nucleus, and vice versa. We can also 486 “‘fournal of Comparative Neurology and Psychology. find from the correlation table the diameter of the nucleus corre- sponding to the any given diameter of the cell-body. The value of the nucleus thus obtained is however affected by a variable probable error owing to insufficient number of observations com- bined with a random sampling. We therefore need to find the most probable values from the observed data, or the characteristic equation which can best represent the data with minimum error. We have two kinds of characteristic equations, linear and non- linear. Whether or not a given expression can be best represented by the linear or non-linear Scharaderrionc equation is of the utmost importance, and it is necessary to determine which equation applies to our present data. Prarson (’04) has introduced a new constant, 7, called the correlation ratio and this is used to test the linearity of the regression. The correlation ratio accord- ing to Pearson is the ratio of the variability of the means of the arrays of one correlated character to the total variability of that character and is shown in the following formula: The constant 7 has the same value as the coefficient of correla- tion when the regression is perfectly linear. If the regression is not linear 7 will be greater than r. Then evidently 7~r 1s a measure of the approach of the regression to linearity. I have calculated the value of 7 by the formula given above and found that when this value is compared with the coefficient of correlation the former is significantly greater than the latter as is shown in the following: 7 — r = .g267—.8616 = .0651 However, the difference between the value of these two constants will in practice deviate more or less from zero. It is therefore necessary to find whether or not the difference found between the two constants is significant. Recently BLAKEMAN (’05) has given methods of obtaining the probable error of various functions of w—r. If we let C=r>r an approximate formula for the probable error of ¢, 7.e., E is ae Va cae I ty 0.6745 9 VS V14(—9)?-C—P)? | Col Hata, Spinal Ganglion Cells of Rat. 487 Applying this formula it was found that C= 1104 -0125 Thus the difference is certainly significant and data demand a non-linear characteristic equation. I have applied PEARSON’s method of parabola (’04, ’05) to the present data and obtained very satisfactory results as will be seen later. “The general formula of parabolas of any order is as follows: I~ Jo cae fe} + & G)+ é3 ¢) i ee where / is a half range of variates and ¢, are the constants to be determined from the observed data. I found that for the present Lt OLE SS CES ee EON 21 23) 250527 0 293 E SB eS Susi S9 Ata Sa ee eon ce Fic. 4. Probable diameter of the nucleus for given diameter of the cell-body. Sd0000 observed; ——————,, calculated. data the parabola of the second order makes a very close fit to the observed means of the arrays. The smooth curve in Fig. 4, where the observed and calculated results are graphically repre- sented, was plotted from the following equation: x a 2 | ¥y = 12.2939 {1.0252 + 3564 (*)-0738 (3) f 488 ‘fournal of Comparative Neurology and Psychology. As will be seen from Fig. 4, the two curves agree very satis- factorily. It is therefore concluded that there certainly exists some definite mass relation between the cell-body and nucleus and its relation is mathematically expressed by the parabolic formula of the second order, as given. Since the regression is not linear but is best represented by parabola we may say that gain in the diameter of the nucleus following increase in the diameter of the cell-body varies in every stage and although the curvature is not pronounced from the nature of the parabola, the diameter of the nucleus is relatively greater in the small cells than in the large cells. For example, when the volume of the cell-body is compared with that of the corresponding nucleus, the following relation is found: In the cell-body whose diameter is g micra the volume of the same is 2.64 times that of the nucleus, while in the cell-body whose diam- eter 1s 47 micra its volume is 25.34 times that of the correspond- ing nucleus. ‘This fact indicates, as was mentioned already, a predominant growth of the cell-body over that of the nucleus. This gives us a method for comparing at some future time the relations in the small cells in the adult ganglion with that of the small cells having the same size in the immature ganglion. CONCLUSIONS. We see from the preceding observations that: 1. The method of section modifies the true frequency distributions of the cells and nuclei when their diameters are considered. 2. Under the circumstances the skew curves of Type 1 for the cell-bodies and that of the Type 4 for the nuclei may be considered the best and most reasonable representation of the frequency distribution of the diameters. 3. The theory that the entire group of small cells with the structural characters of the large cells represents un- changed small cellsis probably erroneous in view of the unavoidable modification of the large cells by the method of section. 4. The diameters of the nucleus and that of the cell-body are highly and positively correlated (r = 0.8616). 5. There exists a definite mass relation between cell-body and nucleus, and the diameter of the nucleus corresponding to any given diameter of the cell-body is best represented by a parabola of the second order. 6. The spinal ganglion cells in a given ganglion may be considered as a homogeneous group, so far as the size is concerned. 7. Spinal Hartat, Spinal Ganglion Cells of Rat. 489 ganglion cells are classified into two groups according to their structural characters: (a) Pycnomorphic cells, those cells which appear dark owing to a stronger affinity to the staining reagents; and the cell-outline of the same is usually irregular. Such cells are usually small in size. (6) Apycnomorphic cells, those spherical or oblong cells which stain lightly and have cell-outlines which are regular. Such cells range from small to large in size. ‘These two groups however grade into one another 490 APPENDIX. TABLE II. Correlation between cell-bodies and nuclei. Fournal of Comparative Neurology and Psychology. = | ) -g |&| “i | wel Zag | Cell-bodies | ~ | | { 8-10ft | 2 | 10-12 | | 12-14 | Pera |e 16-18 | 18-20 | 20-22 | | 22-24 ew 24-26 26-28 | 28-30 30-32 (re seast | 34-36 36-38 | 38-40 40-42 42-44 44-46 46-48 Totals. ....| 2 BegOe 5-95 | | | | | | | Bee) Ss terete eee lot Se is [6 | |r Joo |a|aAlo lala lala om) tn | 1 | else der Te eet ata || a he ih RS |S) 8/3/28 eels [8 [ele |e ls |e era cca catch healer an staged bnagy ale hy ares | | SS SS SSS OS 1 | 2, |) 93 1| | } a | }6)6)4] 3 2 cas Na La a a (a a! |ai5| S| el ee “ae git Seal Seale ES GA Bel 14, 16 a 0 ie 2 | 8 || 12 44, 22 eee I | | 4] 3] 55| 50] 20) 4] 2| | | 6 | 1 37) 76| 36) 25) 11 PH Mime i — — — | Eo) Te eapegs E716). Ei! | 7| 20} 22] 25 14| 14] Ale her 8] 7 14) 26) 14. Bi ae I II| 10 7 3 | 2 reaeaeasl eee aiceal I AMID lcs a ns lea eel Roel eel | 8| 8} 10) 13} - 8 5 ated I 5) 5] 5| 11] 6 1] 2} 9) 4) 9) 4) 1 lee bb oD Teg)! x} 6) 'G=|.4 iota en) Suan lest Res | Cs | | I 1} 10] 3 | 2 | Pee ea leae a | | I) I | mee 7 alles | ae | | a En el OSES EH el eed ae eT | eae =amlrae al at ol cand SR ina ARR ctl ee a 7 | 9 15 | 35) 43,181 240/136 124 97) 7036 52 Rees 15.70-16.35 ‘ 1635-17 .00 17 .00-17.65 108 Seach ae lelelelelelels a Be 3 Bie: 2 1108 Haral, Spinal Ganglion Cells of Rat. 4gI BIBLIOGRAPHY. BLAKEMAN, le ’o5. On tests for linearity of regression in frequency distributions. Biometrika, vol. 4, pp. 3325390: Cox, W. H. *98. Der feinere Bau der Spinalganglienzellen des Kaninchens. Anatomische Hefte, Abth. 1, Bd. 10 pp. 75-103. GREENWOOD, M. ’og. A first study of the weight, variability and correlation of the human viscera, with special reference to the healthy and diseased heart. Biometrika, vol. 3, pp. 63-83. Haral, S. ’or. The finer structure of the spinal ganglion cells in the white rat. ourn. Comp. Neurol., vol. 11, no. I, pp. I-24. oz. Number and size of the spinal ganglion cells and dorsal root fibers in the white rat at different ages. Ibid., vol. 12, no. 2, pp. 107-124. ’o5. A note on the significance of the form and contents of the nucleus in the spinal ganglion cells of the foetal rat. Ibid., vol. 14, no. 1, pp. 27-48. Lenuoss&x, M. von. ’96. Ueber den Bau der Spinalganglienzellen des Menschen. Arch. f. Psychiat. u. Nerven’k., Berl., Bd. 29, pp. 346-380. Luagaro, E. ’96. Sulle alterazioni delle nervosi dei ganglia spinali. Rev. di. pathol. nerv. e ment., Firenze, vol. 5, p. 457. Peart, R. anp Dunzar, Francis J. ’03. Variation and correlation in Arcella. Ann. Report of Michigan Acad. Sci. for the year 1903, pp. 202-204. Peart, R. ’o7. A biometrical study of conjugation in Paramecium. Biometrika, vol. 5, pp.2, 13-297. Prarson, K. ’94. Contributions to the mathematical theory of evolution. I. On the dissection of fre- quency curves. Phil. Trans. Roy. Soc. London, vol. 185, A, pp. 71-110. ’96. Mathematical contributions, etc. Skew variations in homogeneous material. Phil. Trans., 186 A, pp. 343-414. ’97. The chance of death. London. ’o1. Mathematical contributions, etc. Supplement to a memoir on skew variation. Phil. Trans., 197 A, pp. 443-459. ’oz. On the systematic fitting of curves to observations and measurements. Biometrika, vol. I, pp. 265-303, and vol. 2, pp. 1-23. ’ ’os. ‘“‘Das Fehlergesetz und seine Verallgemeinenungen durch Fechner und Pearson.” A rejoinder. Biometrika, vol. 4, pp. 169-212. ’05. Mathematical contributions, etc. On the general theory of skew correlations and non- linear regression. Drapers’? Comp. Research. Memoirs, Biometric series, 2, Pp- I-54, London. Ranson, S. W. ’06. Retrograde degeneration in the spinal nerves. Four. Comp. Neurol. and Psychol., vol. 16, no. 4, pp. 265-293. Scort, F. H. ’o5. On the metabolism and action of nerve cells. Brain, Part 111 and 112, pp. 506-526. 4 P \¢ ) f Pees ; ie hy ihe be < ANOMALIES OF THE ENCEPHALIC ARTERIES AMONG THE INSANE. A SsLuUDY OF THE ARTERIES AT THE BASE OF THE ENCEPHA- LON IN TWO HUNDRED AND TWENTY CONSECUTIVE CASES OF MENTAL DISEASE, WITH SPECIAL REFERENCE TO ANOM- ALIES OF THE CIRCLE OF WILLIS. BY I. W. BLACKBURN, M.D. Pathologist to Government Hospital for the Insane, Washington, D. C., Professor of Morbid Anatomy in the Medical Department Georgetown University, and in the Medical Department of the George Washington University. Witn ELeven Ficures. Minor anomalies of the encephalic arteries are so common that their study and observation have been mainly confined to modi- fications of the circle of WiLLis and the principal basal trunks. Slight deviation from the normal is very common in my autopsies and probably in most cases unimportant, though it is claimed that anomalies of the basal trunks are more common in the insane, to quote BERKLEY, indicating some defect in the “molding of the vital clay.” BERKLEY admits, however, that comparative statis- tics are wanting, while he states that in four cases out of sixteen the arteries showed anomalies, a very large proportion. It would seem evident that developmental defects in the arteries would in most cases be compensated for by modifications of other trunks; and that under normal conditions functions would be preserved, but under diseased states and possibly unusual circulatory disturb- ances the abnormality might be of serious importance. ‘The actual etiological relationship of arterial anomalies to mental diseases other than to defective developmental states, in which it may be inferred, is hard to establish; but the relation they bear to gross cerebral disease, and tosurgical operations upon the cervical arteries is easily demonstrated. It must be borne in mind in the study of these anomalies that 494 “fournal of Comparative Neurology and Psychology. developmental defects should be considered apart from the changes produced by arterial disease. In most cases this is not difficult, and in many instances the acquired condition is clearly inadequate to produce the anomalies present. In order better to comprehend the significance of the arterial anomalies of the circle of Wr1Lis and the great basal trunks it 1s necessary to refer to the normal condition of the circulation as in- dicated by the diagram, Fig. 1. Normally the carotid system and the vertebral system are separated by a balance, or meeting of the two currents, in the middle of the two posterior communicating arteries, while at the same time the circulation of the two cerebral hemispheres is virtually distinct on account of a normal balance which exists in the anterior communicating artery, and by the direc- tion of the blood current in the first part of the posterior cerebral artery. Under normal conditions of current and pressure it must be extremely improbable for any part of the cerebrum to receive even temporarily any blood supply but from the proper vessels. ‘The separation of the cerebellar circulation from that of the cerebrum under normal conditions is equally distinct, the current from the carotid system being opposed by that in the basilar artery. ‘The pontine and upper cerebellar arteries have a more common blood supply, being in keeping with the fusion of the vertebrals into the basilar, though the two posterior inferior cerebellar vessels must normally receive their blood supply from the corresponding verte- brals. Anastomoses between the cerebellar arteries are also more common than in the cerebral trunks though usually the obstruction | of a cerebellar artery results in softening of its proper area. In abnormal developmental conditions this regulation of circu- lation is of course disturbed and the direction of the current in a vessel may be permanently reversed, while in disease of the vessels these anomalies may greatly add to the gravity of the situation. As a basis for this study I have noted the conditions found in 220 consecutive cases of mental disease; and, as a matter of some additional interest I have given the age, sex, pathological condition of the vessels, and the form of mental affectionin each case. I must claim, however, as the foundation of my opinion, the examina- tion of over 2220 cases rather than the smaller number given as the basis for the statistics. I have also given somewhat in detail three cases of anomalies of unusual interest, not included in the list, and have introduced a BLacKBURN, Anomalies of Encephalic Arteries. 495 number of diagrams of associated anomalous conditions which are forthe most partself-explanatory. | would, however, call attention to the number of these in which surgical ligation of the internal carotid, or obstruction of the vessel by thrombosis would endanger the whole hemisphere. It will be noted in these diagrams, and in a number of the cases in the list, that anomalous conditions are frequently associated. In the sixteen cases diagramed this was strikingly manifested, and at the same time there is a certain correlation between some of the anomalies, notably in enlargement of the anterior cerebral of one side and small size of the opposite vessel; enlargement of a pos- terior communicating artery with underdevelopment of the corre- sponding posterior cerebral; and in increased size of the opposite vertebral when one of these is undeveloped. In the list of cases, 227 were included—Nos. 1968—2194—but in some of these the brain was not examined, leaving 220 consecutive cases in which the vessels were studied. “This number and order of cases I think may be taken as fairly representative of the rela- tive frequency of arterial anomalies among the insane, and, though the number is too small for valuable statistics on this question, it may be taken as showing the frequency in the several forms of mental disease. In the 220 cases Sey 65 showed no anomalies, slight dispar- ity in size of the paired vessels being disregarded, and no anomalous conditions being noted unless presenting a deviation beyond the normal variation. Professor WINDLE, who studied the vessels in 200 cases among those presumably sane, found 76 cases normal, and disregarding slight difference in the size of the paired vessels he found 119 cases normal in number and arrangement. As I have included the cerebellar arteries in my list, while WINDLE studied only those of the circle of WILLIs, this must be taken into consideration in com- paring the two final results. As I shall have to make frequent reference to the statistics of WINDLE, for comparison with my own, I would here acknowledge my indebtedness. SUMMARY. Internal Carotid Artery.—The internal carotid artery has been quite constant in size and development in my cases. It is all the 4960 ‘fournal of Comparative Neurology and Psychology. more remarkable that in several reported cases, a large anomalous branch has been found arising from the cavernous portion of the vessel and turning abruptly backward, joining the basilar. A case of this interesting anomaly is given somewhat in detail in Case No. 1926, Figs. 7 tog. A case of direct union of the carotids with- out the intervention of an anterior communicating artery 1s given by Mircuett and Dercum; this I have not met with, though direct fusion of the anterior cerebrals, a closely allied condition, has been seen a number of times. Occasionally some disparity in size of the two vessels is seen, but it was deemed noteworthy in only one case. Middle Cerebral Arteries —The middle cerebral artery does not enter into the circle of Witiis, though the circulation in it is markedly influenced by anomalies of its component vessels. A few abnormal branches were noted, and in one case the inferior external frontal branch was absent and its place was taken by a recurrent vessel from the anterior cerebral. As a result of this a large thrombotic softening of the Sylvian region did not affect the third frontal convolution and speech was preserved. Cases have been recorded of abnormal origin of the vessel itself, and of origin of the post communicating artery and the anterior choroid from this artery, but I have found no case which could not be otherwise explained. Anterior Cerebral Arteries —The most common anomaly of the anterior cerebral arteries is the enlargement of one artery and a corresponding small size of the opposite artery. As the result of this the main blood supply comes through the enlarged vessel, the anterior communicating artery is enlarged, and both anterior cerebrals appear to arise from one carotid. Sometimes the ante- rior communicating artery remains distinct, but it may be com- pletely merged into the anterior cerebrals and the remnant of the small undeveloped anterior cerebral join it at an acute angle. Nos. 2057, 2063, 2066, 2099 and 2137 represent this type of anomaly. Another modification of these arteries is their fusion into a common trunk from the site of the anterior communicating artery onward to the vicinity of the genu callosi, where they commonly again form the two anterior cerebral trunks with a normal distribution. In this condition there is no proper anterior communicating artery. Sometimes the anterior cerebral of one side is very small, in some the vessels are about equal in size. BLackBuRN, Anomalies of Encephalic Arteries. 497 In a few instances the two arteries have fused anterior to the anterior communicating artery leaving a triangular opening be- tween the vessels. Nos. 2073, 2077, 2139 and 2141 show this type of anomalous development. Other anomalies of these arteries are occasionally met with. The vessels may both arise from one carotid as distinct trunks, one of which may be joined by an undeveloped vessel represent- ing the opposite artery, or this vessel may be completely absent. Sometimes both vessels arise from one carotid by a single trunk which may or may not be joined by an abortive branch represent- ing the opposite vessel. “This condition must, however, be care- fully analyzed as it is not essentially different from the conditions represented by Nos. 2004 and 2066; especially if the vessels divide at the situation of the anterior communicating artery. The important surgical complications which would arise in these anomalous developments of the anterior cerebral arteries will be readily comprehended by reference to the accompanying diagrams of combined anomalies. In my 220 cases the anterior cerebrals were fused into acommon trunk in seven cases; see examples Nos. 2073, 2139 and 2141. This condition seemed at first sight to be more frequent as one of the vessels was not infrequently quite small and sometimes prac- tically impervious, so that the vessels appeared as one trunk arising from the carotid. Nos. 2004, 2057, 2063, 2066, 2073, 2099 and 2137 show this type of anomaly. In the list ten cases showed the right artery small and the blood supply coming from the opposite carotid; in six, the left was the defective vessel. In one case the right artery was unusually large while the left was normal; and in five instances a large recurrent branch originating commonly from the anterior cerebral at its junction with the anterior communicating turned backward and supplied the third frontal gyrus and the insula. WINDLE found eight cases of fusion of the anterior cerebral arteries; and two instances of absence of one vessel, the place of the missing trunk being taken by a small branch from the opposite carotid. I have not regarded these cases as absence of the vessel in question, but as instances of underdevelopment with compen- satory enlargement of the opposite trunk and the anterior com- municating artery. See Nos. 2066, 2077, 2099. Anterior Communicating Artery.—The anterior communicating 498 “fournal of Comparative Neurology and Psychology. artery may be absent, extremely small, double, or treble, or it may be formed of two vessels which join before junction with the - anterior cerebral of the opposite side then forming a Y-shaped vessel. When double the anterior is usually the larger trunk. In a number of cases a median anterior cerebral artery arises from this vessel. ‘This trunk is sometimes as large as or even larger than the true lateral trunks, and in some cases it curves around the genu, extends backward along the callosum, and finally divides into two branches which supply the two medial surfaces of the - quadrate lobules. In such cases the true anterior cerebral arteries are distributed to the anterior portion of the medial surfaces, and to the orbital gyri. In normal conditions and in cases of complete fusion of the anterior cerebrals, absence of the anterior communicating artery is of slight importance; in conditions of thrombosis of the carotid its absence may be of serious import. In my list this vessel was absent in only two cases; it was unusually small in five; large in two; double in thirteen cases; formed a Y-shaped vessel in one; and gave off a large median anterior cerebral in two instances.!' In quite a large number of cases this vessel was markedly modified by changes in the anterior cerebral trunks as mentioned above; it was also absent or greatly modified in cases of fusion of the anterior cerebrals, as in Nos. 2073;; 2077, 2099, 2141 and 2169; In WrinDLE’s cases this vessel was double in fourteen; triple in three; formed a Y in six cases; in six instances it was absent in fusion of the anterior cerebrals; in two it was present under like conditions, as in No. 2139; in two cases he found the vessel absent on account of there being but one anterior cerebral artery; and in one case in complete fusion of these arteries. “There were nine instances of a median anterior cerebral artery. Posterior Communicating Arteries —Yhe commonest anoma- lous condition of the posterior communicating arteries is enlarge- ment of one or both of these vessels. In almost every case enlarge- 1 Tt is probable that had a more careful search been made this artery would have been more frequently found. Since the above was written an excellent example of this vessel has been found. The artery took its origin from the anterior communicating, and proceeded as a single trunk with a few small branches to the callosum, to about the anterior border of the precuneus, when it divided into two equal sized trunks going to supply both of the quadrate lobules and adjoining portions of the paracentral lobules. The true anterior cerebral vessels were a little smaller than usual and were distributed to the - lower and anterior medial surfaces of the hemispheres. BLacKBURN, Anomalies of Encephalic Arteries. 499 ment of the posterior communicating arteries 1s correlated with small size of the posterior cerebral trunks at their first part or origin from the basilar. When but one of these arteries is en- larged it is most commonly the right. Quarn found the right artery enlarged in 5.5 per cent of the cases examined; the left in 4.5 per cent; both in 2 per cent. In all these cases of enlargement of the posterior communicating arteries the main blood supply to the posterior cerebral territory comes from the internal carotid artery, but I think it a mistake to regard even high degrees of this anomaly as instances of origin of the posterior cerebrals from the carotid system. Often the posterior cerebral is represented by a mere thread, sometimes this may be impervious or even absent, but in none of my cases have I been convinced of the carotid origin of the posterior cerebral. HyrtTLe reports a case in which the middle cerebral was given off by the posterior cerebral a condition not so readily explained, but it must be extremely rare. The posterior communicating artery is not infrequently wanting, fails to join the posterior cerebral, or ends in a few filiform branches about the crus. QUAIN gives it as absent on the right side in 4.5 per cent; on the left in 6.5 per cent; and on both sides in 1.5 per cent. The origin of the posterior communicating arteries from the middle cerebral rather than from the internal carotid | deem unimportant, as the former vessel is merely a continuation of the latter without definite dividing line, though the posterior communi- cating arteries arise inside of the origin of the anterior choroids, the latter arising from the terminal part of the internal carotid. The posterior communicating artery of the left side was abnor- mally small in seven of my cases; that of the right side in three; both unusually small in four cases. The vessel was totally absent on the right side in one case. The right vessel was enlarged twenty-six times; the left sixteen times; and both were enlarged in twenty-three cases. In all of these cases the corresponding pos- terior cerebral arteries were below the normal in size. In several instances there were slight and inadequate anastomoses with the basilar through small and abnormal branches representing the posterior cerebral arteries. Nos. 2099, 2159 and 2177. WINDLE found these arteries normal in 175 of his cases. In twenty-eight cases the right was the larger; in fifteen it was the left; both were abnormally small in seven cases. He gives these 500 ‘fournal of Comparative Neurology and Psychology. arteries as absent in twenty-five cases; thirteen times the left, the right nine times and both three times. The author states that a partial anastomosis with the basilar was secured in some of these by means of small twigs in the interpeduncular space. It seems possible that some of these may have represented undeveloped posterior cerebral and posterior communicating arteries, .as in INe@:.2177. Posterior Cerebral Arteries—The most common abnormal con- dition of the posterior cerebral arteries is small size of the vessels at their first part before joining the posterior communicating arteries. ‘[his, as mentioned above, is the condition which leads to the conclusion that the vessels arise from the carotid. In some cases the vessel joins the posterior communicating artery by a fine practically impervious thread, it may join at an unusual place, or may fail to anastomose with the vessel at all. In case of undevelopment of this vessel it is almost invariably compensated for by enlargement of either the posterior communicating, or the anterior choroid, but in obstruction of the internal carotid of the same side the effects are disastrous. See Case No. 1909. In my list the posterior cerebral of the right side was noted as smaller than normal in fifteen cases; the left in ten, and both in twenty-one cases. The right artery was absent in one case, and in two instances the main distribution was supplied by an enlarged anterior choroid artery. Nos. 2163 and 2177. WINDLE claims that in twenty-four of his cases this vessel orig- inated from the carotid artery; eleven on the right side; nine on the left; and four on both sides. I have not been able to satisfy my- self that this was so in any of my cases, as it seems to me that all may be regarded as compensating enlargements of the posterior communicating and anterior choroid arteries. In one of WINDLE’s cases and one of my own the third nerve was divided by this artery. Anterior and Posterior Chorotd Arteries.—The anterior choroid artery is fairly constant but may be represented by mere threads. In two instances among my cases I found this vessel in part taking the place of the posterior cerebral. Nos. 2163 and 2177. The posterior choroid arteries sometimes arise from the posterior cere- brals; sometimes from the superior cerebellar arteries. “They are Leia irregular in their development but were not carefully studied. BiackBurRN, Anomalies of Encephalic Arteries. 501 The Basilar A rtery.—This artery, formed originally by coales- cence of the two vertebrals, sometimes contains a septum in its interior, and occasionally shows an incomplete fusion, leaving an opening in the vessel as seen in Case 2009 and in Fig. 2. ‘The vessel is sometimes joined by a large anastomotic branch from the carotid artery. See Case 1926. Inmy list of cases the basilar showed partial separation into its embryonic components in two cases. It was markedly curved in most cases of enlargement of one of the vertebral arteries, the convexity of the curve being opposite to the enlarged vessel. Nos. 2004, 2063 and 2066. Anomalies of origin of the cerebellar arteries arising from the basilar were common. Superior Cerebellar Arteries—Vhe superior cerebellar arteries were quite constantly represented, sometimes being duplicated. This occurred on the right side four times; on the left twice; and on both sides in two cases. Occasionally this vessel sends branches to reinforce the anterior inferior cerebellar when it 1s ill devel- oped. Anterior Inferior Cerebellar Arteries.—This vessel is quite vari- able in its place of origin from the basilar. ‘The vessel was duphi- cate in eight cases; one or the other was absent in seven; and in five instances it sent large branches to take the place of the pos- terior inferior cerebellar artery when this vessel was small or absent. [here seemed to be a constant correlation between the two inferior cerebellar arteries, so that in eighteen cases the two had a common origin in a single trunk arising from the basilar. Posterior Inferior Cerebellar Arteries.—This vessel was absent on the right side in ten cases; on the left in six. In these cases the place of the vessel was usually taken by branches from the anterior inferior cerebellar. In a few cases the artery took its origin from the basilar just above the junction of the vertebrals. Vertebral Arteries—One vertebral artery, according to QuAIN, more frequently the left, is sometimes much smaller than the other. The right vertebral was abnormally small in twenty-one cases; the left in twelve cases; both were small intwo instances. An unusu- ally large right vertebral with the left about normal in size was noted twice; the two vessels united by a transverse trunk in two cases. The posterior inferior cerebellar artery was large and received prac- tically all of the blood from the cervical portion of the vertebral, in five cases; three times on the right; twice on the left. In these 502 “fournal of Comparative Neurology and Psychology. and a few other cases the vertebral was almost or quite inpervious, but in no instance was either vessel absent. T he Anterior Spinal Artery.—The anterior spinal artery usually shows slight variability in its trunks of origin, depending upon the size of the vertebrals. Occasionally the vessel takes its origin from a short transverse trunk joining the two vertebrals. The posterior spinal arteries have not been studied. Associated Anomalies.—In many cases the anomalous condi- tions were so associated that several were present in the same case. I have, therefore, introduced diagrams of a number of the most interesting cases, and have described three somewhat in detail.’ Pathological Conditions of the Arteries —It was a matter of some surprise and interest to find that no less than 148 of the 220 cases showed some degree of arterio-sclerosis. However, as 138 of the patients were over sixty years of age this may be accounted for. In this connection it is well to bear in mind that arterial thrombosis, one of the most common accidents of arterio-sclerosis in the aged, is most apt to result seriously in these anomalous states of the arteries. BULLEN claims that arterial anomalies are most frequent in paresis; in thirteen cases of this disease in my list, six showed slight anomalous conditions, while some were nor- mal in arrangement and not diseased. Of the cases studied seventy were of senile dementia; forty, chronic dementia; fifteen, dementia precox; fifteen, epileptic dementia; nineteen, chronic melancholia; fourteen, chronic mania}; thirteen, paresis; eight, manic depressive insanity; seven, organic dementia; four, terminal dementia; three, acute mania; three, imbecility; three, acute confusional insanity; two of toxic insanity; two, acute melancholia; and one each of acute confusional insan- ity and acute febrile delirium. It is perhaps well to state that under a more modern system of classification some of the above diagnoses of the mental diseases, made when the patients were admitted to the hospital, would be materially changed. Case of Anomalous Circle of Willis. No. 1926. A. C.; white; male; aged 78; soldier; nativity, Ohio. Mental disease, senile dementia; duration uncertain. Autopsy partial, only the brain being examined. * Sixteen of these diagrams, Figs. 3-6, are included among the descriptions of the forty cases abstracted from the complete list of 220 cases. They are, however, sufficiently clear to be understood without reference to the descriptive list. BLackKBuRN, Anomalies of Encephalic Arteries. 503 Skull thick and dense; symmetrical; sutures and bones normal. Over inner surface of dura of the convexity is a thin neo-membrane of internal pachymenin- gitis. Brain. Weight, 1420 grams. Slight general shrinkage over the convexity; veins prominent; arterial branches tortuous. ‘The arteries at the base are sclerotic and present the following anomalies. 1. Anterior communicating artery double. 2. At the junction of the upper and middle third of the basilar artery it is joined by a large vessel fully equal to the normal basilar in size, which comes from the left internal carotid just after its emergence from the carotid canal in the temporal bone. This abnormal artery has pierced the dura mater just outside of, and pos- terior to the posterior clinoid process, and dissection shows that it comes from the internal carotid at the junction of the intra-osseous portion with the cavernous por- tion of the vessel. The branch is not at all in the bony canal nor does it pierce the dorsum sellz, though it lies at its origin quite close to the posterior clinoid process. The aberrant vessel curves abruptly backward and inward joining the basilar, which is so altered as to appear a part of the anomalous trunk. ‘The large size of the abnormal branch has diverted the basilar from its course and evidently the main blood supply to the posterior cerebral arteries, and the superior cerebellars, came through the anastomatic vessel from the left carotid, the other portions of the basilar and the whole vertebral system being small. The sixth nerve has the nor- mal relation to the carotid artery and the anomalous branch pierces the dura at its anterior and inner side. The other cranial nerves are not in any way interfered with. Complete dissection of the carotid artery outside of the cranium is not per- missible. Two or more minute arteries arise from the anomalous vessel and are distributed to the dura in the vicinity. 3. The vertebral system presents marked deviation from the normal. ‘The left vertebral is unusually small and the main blood supply from below entered a large trunk which represents both the posterior inferior cerebellar artery and the anterior, no trace of the latter arising at the usual place from the basilar. This common trunk is distributed to the territory of both vessels represented. ‘The right vertebral artery is about the usual size and passes without change of size but with marked curvature into the basilaris. ‘There is no artery on this side corresponding to the posterior inferior cerebellar. 4. The basilar artery is so changed that the limits of the vessel can only be determined by the junction of the vertebrals and the division into the posterior cerebrals. At about the junction of the middle and lower thirds of the basilar a large trunk is given off which takes the place of the anterior inferior, and posterior inferior cerebellar arteries, the exact reverse of the condition on the left side. ‘The middle portion of the basilaris is small and on the left side gives off the internal auditory artery and some pontine vessels. ‘The upper part of the vessel, scarcely recognizable as such, gives off the superior cerebellar arteries, divides into the poste- rior cerebral arteries, and receives the anomalous carotid branch, as above described. The posterior cerebrals give off the posterior choroid arteries and both are dis- tributed as usual. 5. The posterior communicating arteries are rather small but their origin and destination are normal. The anterior choroid arteries are normal in origin and distribution and so far as can be determined the other arterial trunks show no anomalies. (Figs. 7, 8 and 9.) 504 “fournal of Comparative Neurology and Psychology. Thrombotic Softening of an Entire Hemicerebrum. No. 1909. O. B.; female; white; aged 69; nativity, U. S. Mental disease, chronic melancholia with secondary dementia. Six days before the patient’s death she had a general convulsion from which she did not fully regain consciousness, followed the next morning by another, with hemiplegia and loss of speech; after this she remained comatose until she died four days later. Synopsis of Autopsy.—Arteries at base of brain sclerotic and tortuous; posterior communicating artery of left side unusually large; posterior cerebral of the same side small and impervious; anterior communicating artery small and impervious; On the right side the vessels were of normal caliber except the posterior communi- cating which was rather large. A clot had formed at the upper end of the left inter- nal carotid artery, and as result of the above peculiarities of the circulation the whole left hemicerebrum had undergone acute softening. The hemisphere was deeply reddened, extremely soft, and the cerebral veins were engorged even to their smallest visible branches. It was in fact, practically a hemorrhagic infraction of the whole hemisphere. ‘The sinuses were not obstructed. The softened hemisphere was generally swollen, presenting a marked contrast with the right, which was moderately atrophied. (Fig. 10.) Anomalous Condition of the Basal Artertes. No. 1954. M. J. B.; female; white; aged 64; nativity, U. S. Mental disease, chronic dementia. The patient had a history of cerebral hemorrhage occurring about six years before her death. ‘This resulted in partial paralysis of the right side, from which she suf- fered when admitted to the hospital. Her last illness was a sudden attack of left hemiplegia with evidences of disturbance or irritation of the motor region of the left side of the brain. Synopsis of Autopsy—The arteries at the base of the brain were extremely sclerotic, calcified and anomalous. ‘The left anterior cerebral was large; the ante- rior communicating artery large; and the right anterior cerebral was small, thus the main blood supply to the anterior cerebral region came from the left carotid. Both posterior communicating arteries were unusually large; both posterior cerebral vessels small, especially the right. The left superior cerebellar artery double; left vertebral artery as large as the basilar; the right quite small, and the posterior inferior cerebellar, received nearly all of the blood brought by the vertebral. A small aneurism had formed on the right middle cerebral artery at the point of origin of the principal branches, and a large intra-cerebral hemorrhage had occurred on this side from one of the lenticulo-striate arteries but the actual vessel could not be determined. A horizontal section of the brain revealed an enormous hemorrhage of the right side which had completely destroyed the basal ganglia and capsules and torn its way into and filled the ventricles. On the left side there was a small brownish lesion in the internal capsule just posterior to the angle, which was supposed to be the remains of the former hemorrhage. ‘The brain was cedematous and soft, and the perivascular spaces were greatly enlarged; some general atrophy of the convolutions of the surface. Other conditions found not important in this connection. BLACKBURN, Anomalies of Encephalic Arteries. 505 It would seem probable that in this case we had both the formation of the aneu- rism and the hemorrhage in part dependent upon arterial overstrain consequent to the lack of anastomoses between the right side and the left. The enlargement of the right posterior communicating artery seems to be a direct compensation for the small size of the posterior cerebral, but the small size of the right anterior cere- bral and the current in the anterior communicating artery would effectually pre- vent relief to overstrain in this direction. Certainly the left carotid had a much freer outlet for its contents than the right, but at the same time owing to the com- pensation present the two hemispheres probably received a normal supply of blood until the complications due to arterio-sclerosis supervened. (Fig. 11.) CONDITION OF THE ENCEPHALIC ARTERIES IN FORTY OF THE TWO HUNDRED AND TWENTY CASES. Case 1973. R. M.; colored; male; aged 61; paresis. Arterio-sclerosis. Right anterior cerebral artery small, and the larger portion of the blood came from the left side. From the anterior communi- cating artery onward, the left anterior cerebral artery forms a large trunk which afterwards divides opposite the genu callosi into two callosal arteries. Left posterior inferior cerebellar artery absent; right small. Case 1974. E. N.; white; male; aged 76; chronic dementia. Arterio-sclerosis. Right posterior communicating artery very large; the corresponding posterior cerebral small; post choroid artery arises from the right posterior communicating artery. Left posterior communicating artery rather large; posterior cerebral of this side small. Anterior communicating artery small and practically impervious. Right vertebral artery very small and separates into two branches, one of which forms the posterior inferior cerebellar artery; one of which joins the basilar. Case 1980. L. F.; colored; female; aged 40; manic-depressive insanity. The right posterior communicating artery large and furnishes the main blood supply to the posterior cerebral region, this artery being quite small. The left anterior cerebral artery is larger than normal and sends its blood supply to the opposite artery through an enlarged anterior communicating artery. The right anterior inferior cerebellar is double; the opposite artery comes off the basilar by a common trunk with the posterior inferior cerebellar, and supplies its region, the proper artery being very small. Case 1985. W. E.; white; male; aged 65; senile dementia. Arterio-sclerosis. Left posterior communicating artery very large and is distributed to the inferior temporal regions. ‘The posterior cerebral artery of this side does not join the former, but is distributed to the parts around the crus and the choroid plexuses. ‘The left anterior cerebral artery rather small, the main blood supply com- ing from the opposite artery. Case 1986. D. L.; colored; male; aged 95; senile dementia. Arterio-sclerosis. Both posterior communicating arteries very large; post cerebral arteries extend forward and join them at an acute angle. Anterior communicating artery absent. Right superior cerebellar artery double. Case 2000. L. T.; colored; male; aged 82; senile dementia. Arterio-sclerosis. The right anterior cerebral artery extremely small both anterior cerebral arteries being supplied from the left side. Anterior communicating artery double. Right posterior communicating artery very large and the posterior cerebral joins it about its middle as a small vessel. (Fig. 3.) Case 2004 W.L.; colored; male; aged 50; chronic dementia. Arterio-sclerosis. Right anterior cerebral very small at its origin, both anterior cerebral arteries appear to come from the left side. No true anterior communicating artery exists. After turning around the genu callosi the anterior cerebral arteries are distributed as usual. Right posterior communicating artery small; left larger than normal. Right vertebral artery very small; posterior inferior cerebellar artery of this side represented by a small thread and an impervious branch from the common trunk with the anterior inferior cerebellar artery. On the left side a double anterior inferior cerebellar artery is present, and a double superior cerebellar. (Fig. 3-) Case 2009. S. W.; white; male; aged 66; chronic epileptic dementia. Senile arterio-sclerosis. Both vertebral arteries, about equal in size, join at the upper end of the medulla, and again separate, making a fenestra in the basilar about one-half an inch in length. The right posterior inferior cerebellar artery is normal in position, the left arises at the first point of junction of the vertebrals. The two anterior inferior cerebellar arteries arise from the right and left division of the basilar respectively. 506 “fournal of Comparative Neurology and Psychology. Case 2011. J. W.; colored; female; aged 36; dementia from tumor of the brain. Slight arterio- sclerosis. The left anterior cerebral artery crossed mainly to the right through an enlarged anterior communicating artery, joining with the opposite artery in a common trunk as far as the genu callosi, after which this trunk sends branches to both hemispheres. From the anterior communicating artery onward the left vessel is represented by a small branched vessel. Case 2014. E. B.; colored; female; aged 71; senile dementia. Slight arterio-sclerosis. The left vertebral artery is extremely small, the posterior inferior cerebellar artery taking nearly all the blood that comes through the vertebral artery. Right posterior inferior cerebellar artery very small and a large vessel from the basilar takes its place. Above this vessel is the proper anterior inferior cerebellar artery. Right superior cerebellar artery double. Case 2016. J. C.; colored; female; aged 61; senile dementia. Arterio-sclerosis of the small-vessel type. The anterior communicating artery has a Y-shape. Both posterior communicating arteries are large and furnish the main blood supply to the posterior cerebral territory, these vessels being small. Right posterior inferior cerebellar artery absent, its place being taken by branches from the anterior inferior artery. Case 2018. J. H.; white; male; aged 76; senile dementia. Arterio-sclerosis. Right anterior cerebral small; left large. On each side from the anterior cerebral artery a branch turns abruptly backward and is distributed to Broca’s convolution in part. The right anterior inferior cerebellar artery large and supplies the posterior inferior cerebellar territory, this vessel being inadequate. Case 2034. A. W.; colored; male; aged 76; senile dementia. Arterio-sclerosis. Left anterior cerebral artery small. Both posterior communicating arteries are very large and the corresponding posterior cerebral arteries are small at their origin: Both anterior inferior cerebellar arteries are large, and are sent in part to the posterior inferior cerebellar region, these vessels being rather small. Anterior inferior cerebellar of right side is duplicated. : Case 2037. M. G.; white; male; aged 74; chronic melancholia. Arterio-sclerosis. Anterior cerebral of left side small. Left posterior communicating artery very large and supplies the posterior cerebral region as this artery is very small. Anterior inferior cerebellar artery of right side extremely small and its territory is supplied by branches from the superior cerebellar which is unusually large. Case 2039. J. W.; white; male; aged 50; chronic dementia. Arteries not diseased. The right anterior cerebral artery is small and the main blood supply comes from the opposite vessel. The anterior inferior cerebellar arteries are large and supply the posterior inferior cerebellar regions. Case 2041. M. H.; white; female; aged 74; chronic dementia. Arterio-sclerosis. The vertebral arteries are joined by a transverse vessel about one-fourth inch from their junction with the basilar, and from this vessel arises the anterior spinal artery. Left posterior communicating artery rather large. Right anterior inferior cerebellar artery supplies the posterior inferior cerebellar region. Case 2057. A. K.; white; male; aged 77; chronic melancholia. Arterio-sclerosis. Both posterior communicating arteries large; posterior cerebral arteries small. Right anterior cerebral artery large and sends the main blood supply to the opposite side through a large anterior communicating artery, the left anterior cerebral being very small. Anterior inferior cerebellar artery practically absent on left side. (Fig. 3.) Case 2063. J. B. C.; white; male; aged 65; senile dementia. Arterio-sclerosis. Right anterior cerebral artery small at its origin; left, enlarged and sends the main blood current to the right side through a large anterior communicating artery. Left vertebral artery very small, right unusually large and enters without change of caliber into the basilar. (Fig. 3.) Case 2066. L.D.; colored; male; aged 34; paresis. Slight arterio-sclerosis. Left anterior cerebral artery large and furnishes the opposite side with blood through a large anterior communicating artery, the right anterior cerebral being very small at its first part. Left vertebral artery small; right about the size of the basilar. (Fig. 4.) Case 2070. C. J.; colored; female; aged 57; chronic dementia. Arterio-sclerosis. Anterior communicating artery double and from the anterior vessel arises a large artery which supplies the corpus callosum—a median anterior cerebral. Both posterior communicating arteries are very large, and the posterior cerebrals are very small. Case 2073. A.S.; colored; male; aged 25; dementia precox. No disease of arteries. Left anterior cerebral artery small and distributed mainly to the inferior surface of the frontal lobe; right artery forms a large trunk which divides an inch posterior to the genu of the callosum and is distributed as usual. Left posterior communicating artery rather large. (Fig. 4.) Case 2077. A. M. M.; colored; female; aged 72; senile dementia. Arterio-sclerosis. Right anterior cerebral artery small; anterior communicating artery also small; left anterior cerebral artery forms a large trunk which extends as far as the genu callosi where it divides into two large branches, one of which supplies each median surface. Both posterior communicating arteries are large; posterior cerebral arteries small. (Fig. 4.) BLACKBURN, Anomalies of Encephalic Arteries. 507 Case 2092. H.F.; colored; female; aged 65; senile dementia. Slight arterio-sclerosis. Both pos- terior communicating arteries very large and supply the posterior cerebral region, the posterior cerebral of the right side being represented by a small thread which was impervious, and the left is a small trunk which curves around the crus but does not join the posterior communicating artery. Case 2094. S. S. T.; white; male; aged 65; chronic mania. Arterio-sclerosis. The vertebral arteries are connected by a transverse trunk of considerable size, from which arises the anterior spinal artery. Right anterior inferior cerebellar artery absent, and branches from the posterior inferior cerebellar take its place. Case 2097. P.McC.; white; male; aged 74; chronic dementia. Arterio-sclerosis. Posterior com- municating artery of left side very large; the posterior cerebral very small; posterior communicating artery of opposite side very small, posterior cerebral very large. Anterior communicating artery double. Right posterior inferior cerebellar artery large; anterior inferior cerebellar small. Case 2099. M. D.; white; female; aged 71; senile dementia. Arterio-sclerosis. Right anterior cerebral artery very small; the left is very large, divides into two trunks which correspond to the two anterior cerebral arteries, and there appears to be no true anterior communicating artery. The right posterior cerebral artery apparently arises from the carotid, and is only united to the basilar by a very small branch of a rudimentary vessel which is mainly distributed to the crus and interpeduncular space. (Fig. 4.) Case 2137. C. N.; white; male; aged 66; senile dementia. Arterio-sclerosis. Right anterior cerebral artery very small, its place being taken by the opposite artery which supplies the two vessels through an enlarged anterior communicating artery. Right posterior communicating artery large, posterior cerebral small. Left posterior inferior cerebellar artery absent, its place being taken by a large branch from the anterior inferior cerebellar artery. Case 2139. R. W. E.; white; male; aged 48; chronic epileptic dementia. Arterio-sclerosis. The two anterior cerebral arteries join into a common trunk just anterior to the anterior communicating artery leaving a triangular opening between the vessels. The two arteries then form a large vessel about one-half an inch in length after which they again divide and the upper vessel forms a callosal artery, while the other branch is mainly distributed to the medial surface anteriorly. On the left side, opposite the junction of the anterior communicating artery with the anterior cerebral a three branched trunk arises, the anterior branch taking the place of the anterior cerebral artery of this side, the middle branch distributed to the orbital surface, and the posterior branch goes backward to supply Broca’s convolution and the insula. The right posterior inferior cerebellar artery is absent, the anterior inferior being rather large. (Fig. 5.) No. 2139; all of the branches not represented. Case 2141. J. C.; white; male; aged 80; senile dementia. Arterio-sclerosis. The two anterior cerebral arteries join directly into a single trunk without an anterior communicating artery. This trunk extends to the genu callosi where it again divides into two vessels which are distributed normally. A large branch which arises from the junction of the vessels supplies the orbital and medial regions of the left frontal lobe and is possibly the representative of the left anterior cerebral artery. The right posterior communicating artery is very large, corresponding posterior cerebral artery very small at its origin from the basilar. Left superior cerebellar artery is doubled. (Fig. 5.) Case 2153. R.W.; white; female; aged 66; senile dementia. Arterio-sclerosis. Anterior commu- nicating artery very small andimpervious. Left anterior cerebral does not lie close to the genu of the callosum but is distributed to the orbital, and outer anterior medial surfaces of the left side; the opposite artery furnishes a single callosal artery, only one being present. Both posterior communicating arteries are large and supply the territory of the posterior cerebral arteries, these vessels being small at their origin. (Fig. 5.) Case 2159. P. J.; white; male; aged 66; senile dementia. No disease of arteries. Right posterior communicating artery large; posterior cerebral artery of this side practically absent being joined to the former artery by a small impervious branch. Posterior communicating artery of left side enlarged and the posterior cerebral very small. The basilar, both vertebrals, and the posterior inferior cerebellar arteries are quite small. Anterior communicating artery absent. (Fig. 6.) Case 2160. J.S.G.; white; male; aged 66; chronic mania. Slight arterio-sclerosis. Left vertebral as large as the basilar; right small and the posterior inferior cerebellar artery receives nearly all the blood brought by the lower part of the vertebral artery. Left posterior communicating artery unusually large. Small aneurism at origin of right anterior choroid artery. Case 2163. A. H. T.; white; male; aged 68; senile dementia. Arterio-sclerosis. Posterior com- municating artery of left side represented by an extremely small vessel. The left anterior choroid artery is unusually large; it lies at its first part deeply in the hippocampal fissure and branches supply the choroid plexus; it finally emerges and forms the parieto-occipital and the calcarine artery. The posterior cerebral of this side is distributed mainly to the inferior temporal region and the crus. On the right 508 ‘fournal of Comparative Neurology and Psychology. side a small aneurism is situated at the origin of the posterior communicating artery from the carotid, and one at the junction of the right anterior cerebral and anterior communicating artery. (Fig. 6.) Case 2169. A. J. L.; white; male; aged 72; epileptic dementia. Arterio-sclerosis. Right anterior cerebral artery large, left small; and distributed mainly to the lower medial surface. A large median cerebral artery takes the place of the left anterior cerebral, and after it passes the genu it is distributed to the left upper medial surface and forms a callosal branch. Both posterior communicating arteries are enlarged; left posterior cerebral artery quite small. (Fig. 6.) Case 2172. F. J. L.; white; male; aged 64; senile dementia. Arterio-sclerosis. Right anterior cerebral artery sends a large callosal branch which afterwards divides and supplies both quadrate lobules, while another branch of this artery forms the callosal branch of the left side. Right vertebral artery small; left continuous with the basilar and of the same size. Case 2176. P. C.; white; male; aged 66; senile dementia. Marked arterio-sclerosis. Posterior communicating artery of left side large; posterior cerebral small, mainly distributed to velum inter- positum, crus, and a small branch which joins the posterior communicating artery about one-half inch from its origin. The posterior communicating artery of this side, therefore, chiefly furnishes the posterior cerebral region. Case 2177. D. P.; colored; male; aged 63; senile dementia. Arterio-sclerosis. Right posterior cerebral artery very small, and joins the posterior communicating artery by a small impervious branch. This artery is distributed mainly to the crus, and a large artery corresponding with the anterior choroid runs backward in the fissura hippocampi and after supplying the choroid plexus takes the place of the posterior cerebral artery being distributed in the same way. A small branch from this anomalous vessel anastomoses with the posterior choroid artery. The right posterior inferior cerebellar artery arises from the basilar; the opposite artery arises by a common trunk with the anterior inferior cerebellar, Fig. 6. Case 2179. J. F.; white; male; aged 69; senile dementia. Advanced arterio-sclerosis. Both posterior communicating arteries are large and furnish the region of the posterior cerebral arteries. The left posterior communicating artery communicates with the basilar artery by a small impervious branch; the right anastomoses with the superior cerebellar by a small pervious branch; both posterior cerebral arteries being practically absent. The left anterior inferior cerebellar artery is very small. Case 2184. W.T.; white; male; aged 57; chronic epileptic dementia. No arterial disease. Both posterior communicating arteries are small and are distributed mainly to parts around the crura. The left anastomoses with the posterior cerebral artery by a small branch, the right has no connection with the posterior cerebral artery. The nutrient arteries of the interpeduncular space are larger than usual, otherwise the posterior cerebrals are normal. Case 2189. M. J.; colored; female; aged 65; senile dementia. Arterio-sclerosis. Both posterior communicating arteries small; both posterior inferior cerebellar arteries arise in common with the anterior from the basilar. Small impervious trunk connects the two anterior cerebral arteries posterior to the anterior communicating. BLACKBURN, Anomalies of Encephalic Arteries. 509 LITERATURE CITED. BERKLEY. Mental Diseases. 1900. BRownine. The Cerebral Circulation. Reference Hand-Book of the Medical Sciences, vol.2. 1901. Buen. Post-Mortem Examination of the Brain, etc. ournal of Mental Science, vol. 36. 1890. DeErcum. A Collection of Anomalies of the Circle of Willis. eurnal of Nerveus and Mental Disease, vol. 14. 1889. Lucas, E. Essai historique, critique et expérimental, sur la circulation, arterielle du cerveau. Thése de Paris. 1879. WILDER. : Encephalic Nomenclature. Arteria Termatica. New York Medical Fournal, vol. 41, 1885, and Journal of Nervous and Mental Disease, vol. 12, 1885. WINDLE. On the Circle of Willis. Reports of the British Medical Association for 1887, New York Medical Fournal, vol. 2, 1888, and Fournal »f Anatomy and Physiology, 1887-1888. 510 “fournal of Comparative Neurology and Psychology. Fic. 1. Diagram showing the normal direction of the blood current in the basal arteries, and how completely separated are the carotid and the vertebral system by the balance in the middle of the posterior communicating arteries. At the same time in normal conditions the two hemispheres are almost as distinctly supplied, owing to the balance in the anterior communicating artery, and the direction of the current in the posterior cerebral arteries. The cerebellar circulation is not so distinctly separated on the two sides and the anastomoses between the cerebellar arteries are more complete than those of the cerebrum. Fic.2. Diagram showing the most common variations in the circle of Wits, etc. (4) Enlargement of one anterior cerebral artery with corresponding small size of the opposite artery; (B) double anterior communicating artery; (C) common trunk made by junction of the two anterior cerebral arteries, with division at the genu callosi; (D) enlargement of one, or both posterior communicating arteries which then go to supply the region of the posterior cerebral arteries which are abnormally small; (£) doubling of the superior cerebellar artery; (F) partial doubling of the basilar artery; (G) common trunk giving off the anterior, and posterior inferior cerebellar arteries from the basilar; (H) common trunk giving off these two arteries from a vertebral artery; (J) small size of one vertebral artery with corresponding large size of the opposite artery; (K) large size of a vertebral artery before giving off the posterior inferior cerebellar artery. Fic. 3. Anomalous arrangements of arteries shown by Cases 2000, 2004, 2057 and 2063. BLACKBURN, Anomalies of Encephalic Arteries. 511 memes ie A: Cereors ant. — bes 8 A-Spinals anterior — orm mor “rau | ha OG a i qi ty a SK o We Al : Es = aT ey yi sucog nna ¥ yf als Z = 3 af j z z Zane > ; Es 3 > ACER MEDIA > . <1 Frc: 2. No. 2063. 512 ‘fournal of Comparative Neurology and Psychology. Fic. 4. Anomalous arrangements of arteries shown by Cases 2066, 2073, 2077 and 2099. Fic. 5. Anomalous arrangements of arteries shown by Cases 2137, 2139, 2141 and 2153. BLACKBURN, Anomalies of Encephalic Arteries. 513 V No. 2066. No. 2077. No. 2099. Fic. 4. L& : No. 2137. No. 2139. No. 2141. Fic. 5. 514 ‘fournal of Comparative Neurology and Psychology. Fic. 6. Anomalous arrangements of arteries shown by Cases 2159, 2163, 2169 and 2177. Fic. 7. Case 1926. Semidiagrammatic sketch of the circle of Wit.ts viewed from the left, partly in profile, showing the point of origin of the anomalous branch from the internal carotid artery and its junc- tion with the basilar. By the origin of the basilar branches it will be seen that the basilar artery is represented by three portions which have been designated upper, middle, and lower portions, and that the vessel is very irregular in caliber owing to the junction with the abnormal branch, and the unusual origin of the cerebellar arteries. BLACKBURN, Anomalies of Encephalic Arteries 515 oN O © No. 2163. No, 2169. No. 21 Wife Fic. 6. No. 2159. A Conebetts interior. PorTerion, A. Vertebrate Pa) A CARO got A . scavann©n 277 Hi|= » WEL Wile ai = 12 fe Vi-« 3 z A: Basilarts upper porlion- ree > osterior b) Sante A Corepelti inh P Sh SEES A Ceredrd myedio eS Pic. 7: 516 ‘fournal of Comparative Neurology and Psychology. Fic. 8. No. 1926. Diagram of the cerebral arteries viewed from above, as they lie at the base of the skull. The cavernous portions of the carotid arteries are exposed, and the dotted line shows the position of the anomalous carotid branch. Fic.9. No.1926. Diagram of the arrangement of the arteries of the base of the brain in Case 1926, viewed from below as they lieon the brain. The point of origin of the anomalous branch and its junction with the basilar artery are shown. Fic. 10. Case 1909. Thrombotic softening of the left hemicerebrum from a thrombus, situated in the intracranial portion of the left carotid artery; the anterior communicating artery and the posterior cerebral artery of the same side being small, partly obstructed by arterio-sclerosis, and insufficient-to maintain the circulation. Thrombosis of the right carotid would not have so resulted on account of the large size of the posterior cerebral and the normal size of the posterior communicating artery. Fic. 11. Case 1954. Set of cerebral arteries showing almost all of the common anomalies, with their compensatory relationships. 1, The anterior communicating is large and modified by enlargement of the left anterior cerebral; 2, left anterior cerebral large, right small; 3, both posterior communicating large, corresponding post cerebral arteries small, especially the right; 4, left superior cerebellar artery double; 5, right vertebral artery small, and the posterior inferior cerebellar receives nearly all of the blood brought by the vertebral; 6, left vertebral artery as large as the basilar. An aneurism on the right middle cerebral artery at the origin of its principal branches. BLackBuRN, Anomalies of Encephalic Arteries. 517 oy Fic. 8 Fic. 9. Fic. Io. ae, a ay Le wa : y Os DER a oe Oe | ea Sa 7 P my 4 < ts > ve yy a TA ¥ at i Ph UUs e 8D eigen Lee FaMe Aa : 4 . Pees ra te ene ee ae Pape eee hac | CLA gL Gd Pea ess ey mini aaa ( ae vba 4 ol AL EDITORIAL. PROFESSOR GOLGI ON THE DOCTRINE OF THE NEURONE. The address delivered by Professor Gouct on the occasion of his reception of the Noe prize, which was printed in full in Archivio di Fisiologia for March of this year, 1s of more than passing inter- est to neurologists, physiologists and psychologists. As is well known, the discoverer of the method of silver impregnation of nervous tissue was led to conclusions regarding the structure of the nervous elements which are widely at variance from those reached by the majority of the able students who have followed him in the use of his own methods. The neurone theory, as built up by the labors of His, Foret, CaJAL, VAN GEHUCHTEN and others and formulated by WALDEYER, has dominated all recent work in neurology, both in morphology and in physiology. And now we are again inquiring how far this dogma was accepted uncritically because of its illuminating sim- plicity and how far it is adequate as a basal concept. With still newer methods we are reaching still different points of view. We therefore welcome this review by Professor Goxet of the whole problem of the neurone doctrine, containing the fully matured opinions of his ripest years. Professor Goutal insists first that 1f the word neurone is used at all, it should be applied in the sense proposed by its inventor, WALDEYER, the concept implying, (1) the embryological, (2) the anatomical, and (3) the physiological, independence of the nerve unit. He brings out evidence in vertebrates for a continuous network of nervous processes, as opposed to their anatomical independence and also opposed to the doctrine of polarization of the neurone. ‘The existence of the rete nervosa diffusa is in fact, as it has been from the beginning, the chief point in Professor Go.et’s consistent opposition to the widely prevalent views based on the neurone theory. 520 ‘fournal of Comparative Neurology and Psychology. Taking up first the alleged independence of the nerve elements in embryological origin, attention is called to the fact that the classical results of His upon which this teaching is so largely based were reached by the use of ordinary histological methods and that the application of the silver methods reveals a much more complex structure. Besides evidences brought out by the Goicr method of intercellular connections in very early stages, we have the ac- counts of the pluricellular origin of nerve cells, and of the origin of nerve fibers by cellular concatenation, both of which, if confirmed, would be contrary to the doctrine of the neurone. ‘The results of the most recent studies on nerve regeneration are also interpreted by Professor Gorter in harmony with the conclusion that the em- bryological independence of neurones is unproved. (\ \ at Mh || a hea iy AY iM as | Ay ene 3: ee ~ ‘ SG Ges ees fe TAS OR) ie va oy a ien jee ce esis se x “y Fic. 1. Structure of the fascia dentata of the hippocampus, as figured by Professor Gotct. From Archivio di Fisiologia, vol. 4, fasc. 3, March, 1907, p. 202, by permission of the publishers. On the question of the cellular independence of the adult neu- rone the well known views of the author are reafhrmed and refer- ence is again made to the overshadowing importance of the rete nervosa diffusa as a morphological and physiological entity, the Editorial. 521 full import of which other neurologists have been slow to appreciate. This rete, which is the most important -structure in putting the various elements of the nervous system into functional relation- ship, Got! insists must be regarded as a network whose elements are in protoplasmic continuity. And this, he maintains, is at variance with the fundamental postulate of the neurone doctrine. This matter of terminology is in our opinion of minor significance, but the question of fact involved is of the greatest importance; for the truth of Gotcr’s contention 1s now generally recognized, in so far at least as the recognition of the great importance of the neuropil, or rete nervosa diffusa. Whatever may prove to be the truth regarding the anatomical structure of the neuropil, or Punktsubstanz, neurologists are giving more and more attention to it as the medium of interneuronic com- munication. Indeed it is probable that by far the most important nervous functions take place here rather than in the cell bodies. SHERRINGTON’S doctrine of the synapse illustrates the extreme fruitfulness of this conception when applied in concrete physiologi- cal problems. Gotcr maintains that all of our physiological knowledge 1s opposed to the idea of physrological independence of neurones within this reticulum, but that a given rete must function more or less as a whole. Such a rete is found in the granular layer of the cerebellum and in the deeper layers of the cerebral cortex. In the fascia dentata of the hippocampus we have the mechanism in question in diagrammatic form (see the accompanying figure). Between the cortical cells and the fmbria the fibers break up into a reticulum. “Everything in this relation speaks in favor of the cumulative action of the cells of the whole stratum of the fascia dentata and against any individual action whatever of the cells - themselves.” The evidence in favor of the existence and functional impor- tance of such a rete nervosa diffusa as Professor Gotct describes, at least in some parts of the central and peripheral nervous system, is growing. At the same time, whether or not we call the elements which compose this rete neurones, it remains true that the concep- tions comprised under the term “neurone doctrine” are by no means valueless. ‘The doctrine itself in its original form has been of the greatest assistance in the solution of knotty problems of nervous structure and its usefulness is by no means outgrown, 522 “fournal of Comparative Neurology and Psychology. even though the form of the concept may ultimately be greatly modified. Making the best use of all of these aids to research, let us strive to keep an open mind and avoid a slavish devotion to dogma, whether new or old. NEUROLOGICAL TERMINOLOGY. The appearance of Dr. Lewettys F. Barker’s manual of Anatomical Terminology' suggests some reflections upona subject which was formerly actively debated, but which has by common consent apparently been left to one side in recent years. A decade ago the editors of this ‘fournal remarked (vol. 7, p. 168), “The unification of our nomenclature is to be accomplished, if atvall by a process of survival of the fittest among competing terms at the hands of our working anatomists rather than by legislative enact- ment. Yet the international discussions now in progress may do much to furtherthis end.”’ In the interval which has elapsed since those words were written anatomists have devoted themselves to the prosecution of research, in America at least, with unparalleled vigor, and, though for the most part silent on questions of ana- tomical terminology, it is evident that they have not been heedless of them. The American usage has evidently tended more and more to take its departure from the BasLe terms of the Nomen- clature Commission of the German Anatomical Society (commonly referred to as the BNA terms), and we are therefore indebted to Dr. Barker for the publication of their list with vernacular Eng- lish equivalents in parallel columns. This little manual 1s almost a necessity on the desk of every anatomist who has not the original German list at hand for ready reference. To many neurologists the selection of the terms of this list does not seem to have been always happy; but the adoption of the terms in their entirety is not binding on any one and nothing but practical use in the manifold complexity of actual research and didactic conditions can demonstrate which of the terms have actual survival value and which ones will ultimately give way to better ones. It is gratifying to remember, in this connection, that the Basle system is not designed as a finality, but as one stage merely 1Barxer, Lewettys F. Anatomical Terminology, with Special Reference to the [BNA]. With Vocabularies in Latin and English and Illustrations. Philadelphia, P. Blakiston’s Son & Co. 1907. $1. Editorial. 523 in the development of a more uniform nomenclature. Much remains for future revisions, even in the case of the parts enumer- ated in these tables, to say nothing of future additions to the list. Already a few minor changes have been widely adopted, such as the substitution of dorsal and ventral for posterior and anterior. The naming of encephalic nuclei and fiber tracts opens new prob- lems in nomenclature, which are assuminglarge proportions. ‘The tracts should without doubt be named from their termini, the nucleus of origin preceding the terminal nucleus, much as muscles are named; but to make this plan practicable we must have shorter names for many cerebral structures. “Thus by common consent the word bulb has supplanted rhombencephalon and medula oblongata as a component of names of fiber tracts con- nected with this part of the brain. The most distinctive feature of the BNA neurological terms is the consistent way in which they have been built up on the embryological foundation laid by the researches of Professor His upon the devel- opment of the human brain. While this is an element of great strength, it is also an element of weakness, even from the stand- point of human neurology. For the natural subdivision of the human brain must be based quite as much on phylogeny as upon its ontogenetic development. And when the application of the terms chosen by Professor His to the lower vertebrates is attempted the system breaks down in many parts. ‘The path of the compara- tive neurologist is peculiarly difficult and it is hoped that at the next revision of the Basle terms the neurological list may be very carefully worked over with this standpoint in mind. THE INTERNATIONAL ZOOLOGICAL CONGRESS. The seventh International Zoological Congress held its meetings in Boston, August 19 to 24. The Congress, as judged by the papers which were offered a the attendance at the meetings, was eminently successful. ‘The program clearly indicates that the interest of biologists at present centers about the problems of heredity and animal behavior. We present below a list of the neurological papers which were announced on the programs of the various sections, and a report of the meetings of the section of Animal Behavior, which the organizing secretary of the section, Professor H. S. JennincGs, has prepared for this ‘fournal. F. E. Borezar Epicellulare Nervenscheiben in Tastkérperchen. A. Meyer The Homologies of the Mesial Wall of the Cerebral Hemisphere of Vertebrates. J.B. Jounston The Phylogenetic History of the Somatic Sensory Column of the Vertebrate Brain. B.G. Wiper Provisional Phyletic Schema of the Selachian Prosencephal (forebrain). H. H. Donatpson The Nervous System of the American Leopard Frog, Rana pipiens, Compared with that of the European Frogs, Rana esculenta and Rana temporaria (fusca). E.H. Dunn Some Neurometric Findings from the Peripheral System of Rana virescens. R. J. Terry The Morphology of the Pineal Region in Teleosts. S. P. Gace Changes in Form of the Forebrain of Human Embryos during the First Eight Weeks. H. W. Norris The Cranial Nerve Components of Amphiuma. H. F. Nacutries Lateral Line System of Polyodon. F. E. Borezar Localization und Arten der Geschmacksorgane bei Vogeln. Microscopic prepara- tions. Y. F. Guprernatscn Ueber Geschmacksorgane bei Wassersdugern. Microscopic preparations. S. ApArHy New Facts and Critical Notes about Neurofibrille. H.V.Neat The Development of the Ventral Nervesin Squalus. PartII. Ventral Cranial Nerves. C. Van Bampexe Considérations sur la genése du névraxe, spécielment sur celle observée chez le Pélobate brun (Pelobatis fuscus Wazgl.). R. G. Harrison Outgrowth versus Primary Continuity: A Study of the Development of the Nerve Fiber in Living Specimens. W.M. SmaAttwoop anp C.G. Rocers Physiological Studies on Molluscan Nerve Cells. W. M. SMattwoop anv C.G. Rocers A Cytological Study of Invertebrate Nerve Cells. O. Zur Strassen Animal Behavior and Development. L. Route L’origine de la notochorde et du neuraxe chez les larves. B. G. Witper The Taxonomic Value of the Brain in the Sharks. B.G.WitpEer The Educational Uses of Certain Vertebrates and of Vertebrate Brains. A. Perrunkevitcu Sense of Sight in Spiders. E. Yuna Le sens de l’humide chez les Mollusques. THE FOLLOWING DEMONSTRATIONS WERE ANNOUNCED. . J. Terry Neuroglia syncytium in a Teleost. . F. Gupernartscu (for F. E. Botezat) Epicellular Nerve Disks in Touch Corpuscles. . F. Gupernatscu (for F. E. Botezat) Taste-buds in Birds. F. GupernatscH Organs of Taste in Aquatic Mammals. . A. Locy The Nerve Terminals in Selachians. . J. Terry Model of Brain of Opsanus (pineal region). Be The International Zoological Congress. 525 A. Petrunkevitcu Images in the Spider’s Eyes. H. F. Nacurries Lateral Line System and other Features of Polyodon. J. Warren Paraphysis and Pineal Region in Necturus maculatus, Lacerta muralis, and Chry- semys marginata. B.G. Witper Photographs of Brains. R. M. Yerxes Apparatus for the Study of Animal Behavior. R. G. Harrison Preparations Illustrating the Development of the Nerve Fiber. ANIMAL BEHAVIOR. The International Zodlogical Congress had for the first time at its seventh session in Boston a section of Animal Behavior. ‘That the recognition of this subject was justified is shown by the fact that this section was one of those having the fullest program and the best attendance among those of the Congress, as well as by the character of the communications made. ‘The section was in session throughout the entire time set apart for sectional meetings, though this was by no means the case with all of the sections of the Congress. The program, given below, contained many papers which marked real and important advances in our knowledge i in this field. Animal behavior is a subject of such popular interest that there is of course danger of its degenerating into dilettantism; of its reverting to the anecdotal and generally unsystematic condition which characterized it some years ago. It is therefore gratifying to note the thorough experimental and analytical character of most of the papers presented. ‘There is nothing in biology which lends itself better to studies of the origin, development, and essential nature of the characteristics of living things than the activities with which the subject of behavior deals; but no- where else is a thoroughly scientific method more indispensable. The section of Animal Behavior united with the section of Comparative Physi- ology for its sectional address, which was given by Professor Jacques Logs, on the Chemical Character of the Process of Fertilization. A peculiar feature of the program is the absence from it of any papers by foreign members of the Congress, save in the case of two papers whose authors were not present. ‘This of course shows the prominent part taken by America in developing in a scientific way the field of animal behavior; but the real reason for it 1s doubtless mainly the fact that the students of animal behavior abroad are largely drawn from among the physiologists and psychologists. It was perhaps too much to hope that these would come so far for a ZoOlogical Congress. The program as presented was as follows: L. T. Hosnouse The Importance of Animal Psychology for the Theory of Evolution. E.G.Spautpinc Postulates and Results in Treating the Problem of Conduct. S. J. Hormes The Relation of Behavior to Formative Processes. T. B. Rozertson The Genesis of Protoplasmic Motion. S.SmirH The Educability of Parameecium. S.O. Mast Light Reactions in Volvox. L. J. Core The Influence of Intensity versus Direction of Light in Determining the Direction of Movement of Bipalium and Allolobophora. C. W. Harairr Behavior of Tubicolous Organisms. A.G. Mayer The Annual Swarming of the Atlantic Palolo. H.S. Jennincs Feeding and Defensive Reactions in the Starfish. H. S. Jennincs_ Features in the Behavior of the Starfish, illustrating the Basis for the Attribution by Older Authors of Intelligence to Lower Animals. A. L. Herrera On the Comparative Behavior of Insects and Winged Fruits. 526 ©“fournal of Comparative Neurology and Psychology. V.E.Suetrorp The Behavior of the Tiger Beetles and its Relation to their Variation and Distribu- tion. B. G. Witper Two Contrasted Modes of Feigning Death Adopted by the Same Beetle. W.B.Herms The Reaction of Sarcophagid Fly Larve to Light. . Turner Do Ants Form “Practical Judgments” ? .RercHarp Color Vision, Association, and Warning Color in Coral-Reef Fishes. .Porrer A Comparative Study of Birds with Respect to Intelligence and Imitation. . Herrick Organization of the Gull Community: a Study of the Communal Life of Birds. . Berry An Experimental Study of Imitation. . Yerkes Visual Reactions of the Dancing Mouse. .Cote Behavior of Raccoons under Experimental Conditions. .Witper The Habits of the Coati. Living Animal. .WitpeR Habits of the Domestic Cat not Commonly Recognized or Interpreted. . Witper Examples of the Power of Association with the Cat and Dog. . Hamitron An Experimental Study of an Unusual Type of Reaction in a Dog. Diagrams. TO PS nt QRH R EAPO O faa Pads Aes? > * OTs < « > rate tates tetieley a, at etek, dk SEIS a a ae oy tats’ t Rey ae? De, om a Pies ratat,