— ~ ee ee 2 a ae. == OA OA, - = . = * Te FO Pixp > oN a a eS Et I ors: Oe er SS, Ms ode eS Ce ws 58 oe ess e2. pe. Sv & Se 4/ Piette, Se ‘S eae: etae - AS ee Wet ey ol « $4 ‘ ~~ f a os *~ oe f a3 aw A ey wee <7 tthe: < ’ + athe Crane: tats eaters fetes eo were @. ® 4 2 * bee * f%, % Phas 4 ve oe, i * 94%, Sok + 48. $a4ney: 1% oF gate +59, WK * * . a. TNS v.46 3 tte 7” * A a ss fect gee bigt in aet Tepito dierts Sats athe Get ‘ $'5 sO Pee . ‘ 4 nattiys hid} Sua ia ane ANA A THE JOURNAL OF COMPARATIVE NEUROLOGY - EDITORIAL BOARD Henry H. Donatpson ApDoLF MEYER - The Wistar Institute : Johns Hopkins University J. B. JoHNsTon Ouiver S. Strona University of Minnesota Columbia University C. Jupson HERRICK, University of Chicago Managing Editor THIS VOLUME IS DEDICATED TO PROFESSOR CAMILLO GOLGI VOLUME 30 DECEMBER, 1918-AUGUST, 1919 PHILADELPHIA, PA. THE WISTAR INSTITUTE OF ANATOMY AND BIOLOGY CONTENTS No. 1. DECEMBER, 1918 Otor LARsSELL. Studies on the nervus terminalis: Mammals. Forty-nine figures....... 1 Epwarp Puetrs Aus, Jr. The ophthalmic nerves of the gnathostome fishes......... 69 D. A. Rutnenart. The nervus facialis of the albino mouse. Fourteen figures......... 81 Krvoyasu Marur. On the finer structure of the synapse of the Mauthner Cell with especial consideration of the ‘Golgi-net’ of Bethe, nervous terminal feet and the ‘nervous pericellular terminal net’ of Held. Fifteen figures....................... 127 No. 2. FEBRUARY, 1919 FRONTISPIECE. Portrait of Professor Camillo Golgi — Wiuiiam F. Auten. Application of the Marchi method to the study of the radix mesencephalica trigemini in the guinea-pig. Thirty-five figures.................... 169 Hovey Jorpan. Concerning Reissner’s fiber in teleosts. Ten figures.................. 217 Rosert 8. Exuis. A preliminary quantitative study of the Purkinje cells in normal, subnormal, and senescent human cerebella, with some notes on functional locali- maui s LWwOtmures Hn ONG Charten'.:). i... ., Va.kr: ka eeed eee ye uke os. b ae eee 229 None APRS Kiyoyasu Marutr. The effect of over-activity on the morphological structure of the ene sa, MUUT COR tim Uren i 2), ). 2 at) SV 2s tues 4 Aalgtigisls « s. « Smuaye eye abby cates eee 253 O. VAN DER Srricut. The development of the pillar cells, tunnel space, and Nuel’s spaces in the organ of Corti. Eighteen figures............ “ft, SANG. SiS cE, «eee 283 No. 4. JUNE Howarp Ayers. Vertebrate cephalogenesis. IV. Trnasformation of the anterior end of the head, resulting in the formation of the ‘nose.’ Twenty-six figures........... 323 Lesiiz B. Arry. A retinal mechanism of efficient vision. Two text figures............ 343 D. OGatTa AND SwALE VINCENT. A contribution to the study of vasomotor reflexes. RUT TEES. 1.5 aa Vax os AS x Siete Boze os ser nase hiec Soe McaN 68. 17s + eS RBA Fg Ae we a No. 5. AUGUST Suicreyvuk1 Komine. Metabolic activity of the nervous system. III. On the amount of non-protein nitrogen in the brain of albino rats during twenty-four hours after SCANT y's, AE a Rese IRLT ey ere SPS, Ba re A oa oP cake. Us NIRS oF HOE peg STATS 397 JAMES Stuart Puant. Factors influencing the behavior of the brain of the albino rat are TeR LLG? Sa eh, Name. ONY Ae ar Ge ooh. Semen Sines ny at ale eB Stone e Sicane 411 O. LarsELL. Studies on the nervus terminalis: Turtle. Sixteen figures............... 423 C. G. MacArruur anv E. A. Dotsy. Quantitative chemical changes in the human brain rime ree ey Cli,” sk RIMM Shay", Facial et suse cscs vip Piso eaten dean ea lew eed Wa vk oan o ORS 445 ie he tet ee yr f wif a Sn lodges ia oan att * Ayia we > ' F ‘ i (Pode a he” TO CAMILLO GOLGI PROFESSOR OF PATHOLOGY AT THE UNIVERSITY OF PAVIA, SAVANT AND CITIZEN, GUARDIAN OF PUBLIC HEALTH AND STUDENT OF HISTOLOGY, TO WHOM THE WORLD IS INDEBTED FOR A METHOD WHICH GAVE A DEEPER INSIGHT INTO THE ARCHITECTURE OF THE CENTRAL NERVOUS SYSTEM—THIS VOLUME OF THE JOURNAL OF COMPARATIVE NEUROLOGY IS DEDICATED AS A TOKEN OF HIGH ESTEEM FOR THE SCIENTIST AND THE MAN WHO, HONORED AND FULL OF YEARS, NOW WITHDRAWS FROM HIS PROFESSORIAL RESPONSIBILITIES THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, No. 1 DECEMBER, 1918 Resumido por el autor, Olof Larsell. Estudios sobre el nervio terminal. Mamfferos. El autor describe con algtin detalle el nervio terminal en el gato, buey y mulo,,dando también descripciones mds breves de dicho nervio en el caballo, perro, ardilla y en el hombre. En el buey, mulo y ardilla se describe este nervio por primera vez. En los mamiferos esta’ formado principalmente por fibras del simpatico; las del fasciculo principal del nervio tienen con las que se distribuyen periféricamente una relacién semejante a la que existe entre las fibras preganglionares y postganglionares. Esta semejanza esta aumentada por la estructura de los racimos ganglionares y por la presencia de cestas pericelulares en muchas de las células ganglionares. También existen redes intercelu- lares y células tfipicas del simpatico. Por la arteria cerebral anterior y susramas y por el plexo vascular del tabique nasal se distribuyen fascfculos de fibras mielinicas y amielfnicas proce- dentes del tronco principal y ganglios. Estas fibras terminan en las paredes de los vasos sanguineos por medio de termina- ciones nerviosas sensitivas y motrices. Hay algunas pruebas de que las terminaciones nerviosas libres en el epitelio nasal estan relacionadas con el nervio terminal, pero la presencia indudable de fibras del trigémino en el plexo del tabique, junto con las pro- cedentes del terminal, no permite una afirmacién rotunda sin previo trabajo experimental. Estas terminaciones, Junto con las de ciertas células ganglionares parecen indicar la existencia de un componente sensitivo en el mervio, distinto de las fibras aferentes del simpdtico. Parece claro que la inervacién del érgano de Jacobson por parte del nervio terminal es incidental y secundaria. El autor expone la posibilidad de que el nervio terminal represente una divisién del sistema del simpdtico, rela- cionada con el cerebro anterior. Translation by Dr. José Nonidez, Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 9 STUDIES ON THE NERVUS TERMINALIS: MAMMALS! OLOF LARSELL Department of Anatomy, University of Wisconsin FORTY-NINE FIGURES CONTENTS If. JEN ys IGE) ae ia eae ieee ees ey eho Shey 6. riryic’'3'eecit tact che Bea Ree Rote cea bey d Ge acer 3 SIS GOMES SINC TLATNL IT 2 222 aye eee ote te gees elo, hs Or eke 8 sie es hate oa? 3 Conditions in the different classes of vertebrates..................... 5 APSE NP UEV GE UIACU 7510's: 2 cha cyank bs a. 4.3 3 ace, se uaeie oars Lames at la hts Rey Us 12 MeiienlaleanGemevhod ses. *.42:.7 casero Cee eee ere oe ee eter. 12 iL, Aad Tvsanvsy iGiaemhar bE) Oru CMtno ann onas oacosdassocntescusgcaavor 16 istolo orca lie | techs cls oeh ool hide ce en Se eee a ome 26 iynes at.gamg lion Celle... s25:.. xy. 18 tee bleed ake eee ede ENS oi 26 Kiberstand stibernnebwOrksic.30-,.... 0tadtonsee teeter eee ae 33 INET Ve RU CIEL AULOMS foes sccce cio sie Bare aeicce © oe MMe ts Bete 37 Hehe menvus terminalis Of; thie bee. .y 1... 55 scl: «ss » pele cee elas r= 43 uaa sical s seme tte. BASS OU 28h Re ee ee 48 Sin chuLe of menverOundlesiq cots ccios os (eee eee 48 INGIVcaLeLminatlOnsiyes piycinck ies cake acs» seein ae ee ee 50 3. The nervus terminalis of the mule and the horse.................. Bl AU oVevaen0ll (ein Nae eae s e GA SoA can NES Go, ee A A Se 8 oe c 51 LIST OLO OTC AIM nahn. arth ates et Soa ora io > SA Es ee 55 TRIDENT EL aoa le eee Re hac ee ee a Rey. ae 57 4. The nervus terminalis of the dog, the squirrel, the human, and of embpryosrof pig, sheep, andirabpit fica: < $2 4. Veen eee tle 61 Se eer LEMIAA ATV CATEG COTITMA CTI Bi (cc x os) pci e/si yak cvaaal gl wale «e's Sieldbd a Pabede had oe BM 62 hs TEI OUT CHATS 01620 eS ec ie denM IR ERIE ok eu Ue a nM PRs ely ae Pr 64 > I. INTRODUCTION Discovery and naming. The cerebral nerve now known as the nervus terminalis first began to attract the attention of morphol- ogists in 1894. In that year Pinkus described in the dipnoan fish, Protopterus, a hitherto unrecognized nerve of the forebrain. 1 Contribution from the Zoological Laboratory of Northwestern University, William A. Locy, Director. 4 OLOF LARSELL This nerve had previously been figured by Fritsch in the sela- chian, Galeus, in 1878, and had been mentioned in 1893, by C. L. Herrick in the urodele amphibian, Necturus. The observations of Fritsch and Herrick, however, were merely incidental, in con- nection with other work, and the significance of the observations — escaped them. After these anticipatory glimpses the nerve remained unno- ticed until Pinkus described it in Protopterus in 1894. In 1895, in a more extended paper, Pinkus figured and described it as ly- ing ventral to the olfactory nerve and extending caudad over the ventral surface of the forebrain to the recessus praeopticus, _its peripheral terminations being in the olfactory sac. Thus, although first figured in selachians, it was first described with sketches, in the Dipnoi. Shortly afterward, Allis (97) described and figured in the ganoid fish, Amia, a strand which he traced centrally to the bulbus olfactorius and peripherally to the olfactory capsule. He considered this strand to be homologous with the nerve de- scribed by Pinkus. In his comments regarding the possible function of the nerve, he suggested that it might be of sympathetic type. This is interesting in view of the position taken by Brook- over and others and of the demonstration of the presence of sympathetic fibers in the bundle of the nervus terminalis of mammals. Now appeared the first study of the embryological history of the nerve (Locy, ’99) together with a description of its adult condition in Squalus acanthias. In this form the nerve was described and figured as possessing a compact ganglion, as con- nected with the brain in the fissure between the lobes of the tel- encephalon, and as distributed anteriorly to the lateral part of the olfactory capsule and entering between the folds of the nasal epithelium. It was claimed that the nerve arises from the neu- ral crest before the appearance of the olfactory fibers. At that time Locy considered as doubtful its homology with the nerve described by Pinkus, and provisionally designated it as a median ‘accessory olfactory strand.’ In 1903, after observing the same nerve in six genera of selachians, Locy reversed his earlier opin- NERVUS TERMINALIS: MAMMALS 5 ion and concluded that the nerve of selachians is homologous with that described by Pinkus in Protopterus and later by Allis in Amia. The same author in his paper of 1905, based on the ex-. amination of twenty-seven species of adult selachians and the puibsveluees history of the nerve in Squalus, proposed the name of ‘nervus terminalis’ for this new nerve. This name has been generally adopted. In the interval Sewertzoff oy had described the nerve in embryos of Ceratodus, finding it ganglionated and terminating peripherally in the mucous membrane of the anterior nasal chamber—not in the sensory epithelium. On account of its point of central connection with the brain he suggested for it the name of ‘nervus praeopticus.’ This nerve had also been cited in the adult of Ceratodus by K. Firbringer in 1904, as well as by Bing and Burckhardt in 1904 and 1905. Conditions in the different classes of vertebrates. Since 1905 a considerable literature has accumulated regarding the nervus terminalis. Its presence.has been demonstrated in all cases of vertebrates except the cyclostomes (and possibly the birds) and various suggestions regarding its function have been made from time to time. Inasmuch as the present paper deals with the nervus terminalis chiefly in mammals, it is not necessary to enter into a review of the rather extensive literature of the nerve in the lower verte- brates, But, since the nerve presents some modifications and some erdries: in the mammals, it is advantageous both for comparison and for discussion of results, to have a brief state- ment of the chief structural features which have been observed in other classes of the phylum. ‘The nervus terminalis appears to be more generalized in the fishes, especially in the selachians, and in discussing its relations in mammals it would be a mistake to disregard the findings in the lower vertebrates. Fishes: a) Selachians. In a paper on the telencephalon of the selachians, Johnston (’11) shows that typically the nervus ter- minalis enters the brain substance near the recessus neuroporicus internus. He remarks: ‘“‘Some evidence has appeared recently (Burckhardt, ’07, p. 340; Brookover, ’10) that the nervus ter- 6 OLOF LARSELL minalis is a mixed nerve, containing in some fishes peripheral sympathetic fibers distributed to the blood-vessels. These ef- ferent fibers make their exit in a dorsal root (N. terminalis) as the viscero-motor fibers typically do in the spinal region of lower vertebrates.”’ Belogolowy (12), from a study of young selachian embryos, concludes that the nervus terminalis is derived from the terminal portion of the neural crest. This was also claimed by Locy (99) and (’05 a). \ McKibben (’14) studied the histological structure of the gan- glion terminale of Mustelus by intravitam methylene-blue stain- ing. He found the great majority of the cells to be multipolar and ‘‘few if any bipolar cells.”” Landacre (16), from observa- tions on Squalus embryos, holds that the terminalis is combined with the olfactorius as a possible general cutaneous component of the latter. . b) Ganoids. Through the studies of Brookover (’08 and ’10) we have very complete reports of the terminalis for the ganoid fish Amia. He concludes that the ganglion of the terminalis arises from the olfactory placode a little later than do the fibers of the olfactory nerve. He doubts its mdependence and is disposed to consider it a part of the olfactory system. In histological observations he finds ganglion cells chiefly of sym- pathetic type. He also finds fibers along the blood-vessels and a connection between the terminalis fibers and the ciliary gan- glion, and suggests that the circumstantial evidence leads one to ascribe to it a vasomotor function, in part. The same author gets similar results from Lepidosteus (14) where the ganglion terminale appears to arise from the olfactory placode after the formation of the olfactory fibers. ‘‘The disposition of the cells in Lepidosteus in a more compact central and a diffuse peripheral ganglion allows of its falling quite naturally into. the morpho- logical relations of the typical autonomic system.” This shows the tendency toward interpreting the terminalis as sympathetic in nature (or at least as containing sympathetic fibers) which becomes more marked in studies of the mammals. NERVUS TERMINALIS: MAMMALS ri c) Teleosis. The presence of the nervus terminalis in bony fishes was first reported by Sheldon and Brookover (’09) in the carp (Cyprinus carpio). Sheldon (’09) independently takes up the central course of the nerve. The tract is composed of un- myelinated fibers. Numerous scattered ganglion cells were found on the ventromedial aspect of the olfactory nerve, from some of which coarse fibers were traced to the olfactory epithel- ium where they were distributed with the olfactory nerve fibers. Centrally the fibers for the most part decussate at the anterior commissure, but no exact nuclear connection could be found. Brookover and Jackson (’11) studied the development of the terminalis in Ameiurus, and also its adult relations by means of the silver-impregnation methods. They find the nerve to be closely related in its development to the olfactory nerve and are inclined to consider it a part of this rather than as an inde- pendent nerve. They point out the proximity of fibers of the nervus terminalis to blood-vessels, but find only a single in- stance where the blood-vessel definitely appeared to be inner- vated by terminalis fibers. A vasomotor functon is suggested. Amphibia. C. Judson Herrick (’09) found in larval and adult frogs a bundle of unmyelinated fibers corresponding so. closely to the nervus terminalis of the fishes in its central course that he considered it to be homologous with the latter. He was unable from his material to determine peripheral terminations. The nerve is not exposed as in selachians. It runs along the ventral border of the olfactory nerve and becomes imbedded in the brain substance just caudad to the glomerular formation. Within the brain substance it passes caudally (in one case showing arboriza- tions in the lamina terminalis) and the fibers cross in the middle part of the anterior commissure. McKibben (’11) traces the course of the nervus terminalis in Necturus and a number of other tailed amphibians. As in the frog, it is mainly imbedded in the brain substance. The principal central distribution is to the preoptic nucleus. The central bundle undergees a partial decussation in the anterior commis- sure, but groups of direct fibers extend further backward from this point, giving off branches at intervals. The continuity of 8 OLOF LARSELL these fibers with those of the nervus terminalis was not clearly demonstrated. The fiber groups were lost in the brain substance without evidence of definite terminations, though the fascicles reach backward into the mesencephalon (hypothalamus and interpeduncular region), a condition not yet noted in other forms. Peripherally are ganglion cells with fibers going appar- ently to the nasal capsules. C. Judson Herrick (’17), in connection with other studies, has completely confirmed in Necturus the findings of McKibben. Reptilia. Among reptiles the nervus terminalis has been found by Johnston (’13). In a paper, which embraces also a consideration of the nerve in pig, sheep, and human embryos, he describes the terminalis in embryos of the turtle (Emys lutaria). In Emys the nerve emerges from the rostral end of the median wall of the brain hemisphere, caudal to the olfactory bulb. From this point ‘‘It descends over the medial surface of the bulb and olfactory nerve and bears clumps of ganglion cells at several points of its course. It comes into close relation with the dorsal division of the olfactory nerve, but is distinguished from it.’ Peripherally the fibers of the nervus terminalis are distributed with those of the dorsal. division of the olfactory nerve to the extreme lateral portion of the nasal sac, which is interpreted by Johnston to correspond with the vomeronasal organ of mammals. According to McCotter (’17), the dorsal division of the olfac- tory nerve of the turtle is to be considered homologous with the vomeronasal nerve of mammals, and the dorsal portion of the formatio olfactoria of the bulb, as illustrated in Johnston’s fig- ure 11, corresponds to the accessory olfactory bulb of mammals. Birds. Whether or not the ganglion cells observed by Ruba- schin (’03) in the chick embryo represent cells of the nervus termi- nalis is problematical. This writer describes a ganglionic mass related on the one hand to the trigeminus nerve and on the other to the olfactory mucosa. Axones from this mass were traced to the Gasserian ganglion. Two types of cells were found in the ganglionic knots: 1) bipolar cells resembling those of the inter- vertebral ganglia, and 2) multipolar cells with numerous proc- NERVUS TERMINALIS: MAMMALS 9 esses, of which one in each case enters the ‘ramus olfactorius nervus trigemini.’ Cells of this type are relatively few in num- ber. These observations have been interpreted by some writers to indicate the presence of the nervus terminalis in birds. Mammals. Since 1905 the nervus terminalis of mammals has been dealt with in no less than nine scientific memoirs. In some cases it has been confused with the fibers of the vomeronasal nerve (Devries, ’05; Déllken, ’09), but in most cases the fibers of the terminalis are distinguished from those of the vomeronasal. Notwithstanding these investigations, the nervus terminalis in the mammals is very imperfctly known and its relations are obscure. The first published notice of this nerve in the mammals was made by DeVries in 1905. He found in the human fetus of three to four months a transitory ganglion which he regarded as cor- responding to the ganglion of the nervus terminalis, and which he designated ‘ganglion vomero-nasale.’ He also found similar conditions in the guinea-pig. His assumption that the vomero- nasal nerve of mammals represents the nervus terminalis of selachians and other fishes is not substantiated by more recent work. Dollken (09) studied embryonic stages of rabbit, mouse, guinea- pig, pig, and human. His account of the central connections of what he describes as the nervus terminalis is extended. He finds roots which enter the brain and reach the cortex, the gyrus forni- catus, the hippocampus, and the septum pellucidum. The pe- ripheral distribution he describes as being by four or five strands ‘to the vomeronasal organ. It seems clear from his description and figures that he is dealing almost entirely, if not completely, with the vomeronasal nerve, which Read (’08) and McCotter (12) have clearly differentiated from the olfactory fibers proper in mammals, and McCotter (’17) in the turtle and the frog. In a paper already referred to in connection with the nervus terminalis in reptiles, Johnston (’13) also describes the nerve in embryos of pig, sheep, and human. In pig embryos he finds the root of the terminalis entering the brain at the ventral end of the fissura prima. ‘The fibers are traceable for some distance within 10 OLOF LARSELL the brain toward the anterior commissure. The peripheral course of the nerve is described as being in the wall of the sep- tum nasale, along which it passes by several strands to the wall of Jacobson’s organ and to a small area of the nasal sac immedi- ately adjacent. In the human embryo Johnston found essentially the same central relations as in the pig, but the peripheral distribution is by a network of nerve bundles in the nasal septum. He concludes that ‘‘the evidence at present in hand seems to establish beyond doubt the presence in all vertebrates of a re- ceptive component in the nervus terminalis supplying ectodermal territory. This component is derived either from the terminal part of the neural crest (Johnston, ’09b; Belogolowy, 712) or from the olfactory placode (Brookover, ’10). The nerve is dis- tributed to the nasal mucosa, or to a specialized part of it, the vomeronasal organ.” McCotter (713) demonstrated by dissection the main central bundle of the terminalis in the adult dog and eat. He also found the typical ganglion cells of the nerve distributed along the vomeronasal strands. No differentiation of fibers from those of the vomeronasal nerve was obtained by the staining methods used. The application to the problem of a modified pyridin-silver technique by Huber and Guild (’13), served to clearly differenti- ate the terminalis from the vomeronasal nerve. These investi- gators used rabbit fetuses and young rabbits. They were fortunate in securing a differential stain which made it possible to follow the fibers of the two sets of nerves individually. They conclude that this nerve is not a component of the olfactory and vomero-nasal com- plex, but an independent nerve, with central connections by means of several small roots to the ventro-mesial portion of the forebrain, caudal to and independent of the olfactory stalk, and courses in the form of a loose plexus along the ventro-mesial surface of the olfactory bulb, reaching the nasal septum and the mesial surface of the vomero-nasal nerve, which nerve it follows to the vomero-nasal organ, and is further distributed to the septal mucosa anterior to the path of the vomero- nasal nerve, in which region especially it is joined by terminal branches of the trigeminus, mainly from the naso-palatine bundles. NERVUS TERMINALIS: MAMMALS Nail Numerous ganglionic masses of various sizes are found. One group of relatively large size located near the most caudad bundle of the vomeronasal nerve a short distance from where the latter leaves the accessory olfactory bulb is regarded as the ganglion terminale of authors. These groups of ganglion cells present the appearance of small sympathetic ganglia, and the authors state that the nerve fibers have more the appearance of sympathetic and preganglionic fibers than of neuraxes and dendrites of sen- sory neurones. A comparison of these cells with the cells of the Gasserian ganglion of the same animals revealed.the fact that the terminalis cells are of smaller size. Distribution of terminalis fibers to blood-vessels and septal mucosa was considered probable from the observations, but the authors hesitated to assert such distribution because of the com- mingling of fibers from the trigeminus with those of the nervus terminalis. Johnston (’14) describes the central relations of the terminalis in the adult human, in the horse, porpoise, and the sheep. Nu- merous rootlets were found in some, especially in the horse, while in other forms only two or three are enumerated. In the horse, a large, compact ganglion is described. In the other mammals the ganglionic masses are described as being smaller but more numerous, and more or less scattered along that portion of the nerve which it was possible to examine. Brookover (’14) independently of Johnston discovered the cen- tral portion of the nerve in the brain of the adult human, and reached substantially the same conclusions as to central connec- tions as did Johnston, namely, that its intracranial course lies over the middle of the gyrus rectus and appears to enter the brain substance in the region of the medial olfactory striae. Both central and peripheral distribution in the human fetus is described by McCotter (’15). He indicates the peripheral course in the nasal mucosa, where it resembles in general the conditions found by Huber and Guild in the rabbit. Centrally the majority of the fibers form a single strand. The latest paper which has come to notice has appeared since the greater part of the observations recorded in the present ar- i OLOF LARSELL ticle were made. In this paper Brookover (’17) describes the peripheral distribution ofthe terminalis in the nasal septum of the human fetus at full term. This material was prepared by the pyridin-silver method. A large plexus of fibers is found anastomosing over the nasal septum deep to the main arteries. The writer states that this network ‘‘is so large that it may be considered as hypertrophied as compared to the known development in other mammals, without apparently increasing the central root.’’ There are indications of a sympathetic chain connection with the spheno- palatine nerve and ganglion. It seems clear that in selachians, Dipnoi, ganoids, teleosts, Amphibia, reptiles (turtle at least), and mammals there is com- mon to all a nerve with central connections with the brain near the embryonic anterior neuropore, and having a primary periph- eral distribution to some part of the lining of the nasal cavity. II. DESCRIPTIVE PART Material and methods. ‘The original design of this investiga- tion was to make a comprehensive analysis of the nervus termi- nalis in the various classes of vertebrates. More difficulties of analysis and interpretation are encountered in mammals than in the other groups, so that the present contribution is limited in its scope to the conditions found in certain mammals, with the expectation of extending these observations in a subsequent paper to other classes. The studies were carried on in the Zoological Laboratory of Northwestern University from 1915 to 1918, inclusive, under the direction of Prof. William A. Locy, to whom I express my sense of indebtedness for helpful advice and criticism. My thanks are also due Prof. 8S. W. Ranson, of Northwestern University Medical School, for valuable suggestions. The mammalian material used consists of the following: 1 longitudinal and 1 transverse series of 10 mm. kitten embryos, fixed in 10 per cent formalin and stained with haematoxylin. 1 sagittal series through forebrain and nasal septum of kitten one day old, treated by the pyridin-silver process. NERVUS TERMINALIS: MAMMALS 13 1 sagittal series through forebrain and nasal septum of kitten two weeks old, pyridin-silver method. 2 sagittal series of nasal septum of kittens two weeks old, pyridin-silver method. 1 sagittal series through forebrain and nasal septum of kitten one day old, fixed and decalcified in Zenker’s fluid and stained by the Weigert method. 2 septal mucosae of kitten one day old, stained with methylene-blue. Numerous series of pig embryos at various stages, stained with haematoxylin or with Mallory’s connective tissue stain. In addition to these sections, the following materials were also studied according to the method indicated: Numerous dissec- tions of kittens, of puppies, and of adult dogs and cats were made. The method described by McCotter, by which the head is fixed in’ Miiller’s fluid to which acetic acid has been added, was used with good results. Fixation in 10 per cent formalin, followed by decalcification in 10 per cent nitric acid, was also adopted in many instances. A light surface stain with borax-carmine was found advantageous in differentiating the nervus terminalis from the neighboring tissues. Frozen beef brains were obtained from the Chicago Stock Yards, and were found to be very favorable for the dissection of the delicate strands of the nervus terminalis. These refriger- ated brains were allowed to thaw in a solution of 10 per cent formalin at room temperature. Besides being relatively easy to dissect, this material responded well to the gold-chloride tech- nique and gave excellent histologic preparations. A number of beef fetuses of 110 mm. to 140 mm. greatest length were also dissected. These had been preserved in formalin. Supple- mentary studies were made on beef material from which the meninges and a portion of the brain beneath the region of the nervus terminalis were removed and fixed in 1 per cent osmic acid, immediately after the brain was taken from the cranial cavity. Most of the material thus obtained was still warm when fixed. It was made possible to obtain this through the courtesy of Swift & Company. The brain of a full-term mule, freshly removed and fixed in 10 per cent formalin, was obtained. Another mule fetus of 121 mm. greatest length was also studied. The brain, in situ, and the nasal septum of an adult horse was obtained through the 14 OLOF LARSELL courtesy of the Chicago Veterinary College. This was fixed in 10 per cent formalin and decalcified in nitric acid. A number of dissections of pig, rabbit, and sheep embryos were made. Most of these were treated according to the method given by Prentiss (715). Opportunity to examine twelve human brains was obtained through the courtesy of Prof. 8. W. Ranson. It was attempted to ascertain if the relations of the fibers of the nervus terminalis to the cerebral blood-vessels is the same in the human as in the cat, the mule, and the ox. The nerve was identified in five of the brains, but no definite light was obtained on the point in question. ABBREVIATIONS ant.cer.art., anterior cerebral artery art.w., arterial wall ax., axone az.cyl., axis cylinder bi.c., bipolar cell bl.v., blood-vessel bu.olf., bulbus olfactorius bu.olf.ac., accessory olfactory bulb cen., centripetally cer.hem., cerebral hemisphere co.ant., anterior commissure c.pr., central process cri.pl., eribriform plate d.n.ter., dorsal main bundle of nervus terminalis in nasal septum fi., nerve fiber gn., main ganglion (‘ganglion termi- nale’) of nervus terminalis gn’., accessory ganglia in.pl., intracranial plexus of nervus terminalis lam.ter., lamina terminalis m.n.ter., median main bundle of nervus terminalis in nasal septum my., myelinated nerve fiber my.sh., myelin sheath n.eth.ant., anterior ethmoidal nerve n.op., optic nerve no.R., node of Ranvier n.ter., nervus terminalis n.vom., hervus vomeronasalis olf.fi., olfactory fibers op.chi., optic chiasma or.vom., vomeronasal organ per., peripherally ; p.pl., peripheral (septal) plexus of ner- vus terminalis p.pr., peripheral process pr., nerve process R., fiber of Remak r1jr2.r3,r4, central roots of nervus ter- minalis ram., ramus r.dor., ramus of dorsal main bundle of nervus terminalis resp.epith., respiratory epithelium r.med., ramus of middle main bundle of nervus terminalis r.ven.,ramus of ventral main bundle of nervus terminalis sp., spiral un.c., unipolar cell unmy., unmyelinated nerve fiber v.n.ter., ventral main bundle of nervus terminalis NERVUS TERMINALIS: MAMMALS 15 Turtle embryos and young of Amia were prepared by various methods and a number of dissections were also made. For nerve terminations the gold-chloride method was em- ployed, according to the modification given by Hardesty. Meth- ylene-blue was tried, but with less success in demonstrating the sensory terminations to be described, although it brought out Fig 1 Dissection of nervus terminalis of kitten of two weeks, illustrating plexiform appearance of nerve, as seen through the binocular microscope. The blood-vessels, along which the majority of the nerve strands course, are not represented. the motor endings quite clearly. Both sensory and motor ter- minations were shown in pyridin-silver preparations. A modification of the pyridin-silver method given by Huber ° and Guild (’13) was used with good results. This modification consisted essentially in lengthening the periods during which the preparation was kept in the various fluids. Briefly summar- ized, the procedure was as follows: Ammoniated absolute alcohol 16 OLOF LARSELL (after injection with same) seven days; decalcification in 7 per cent nitric acid; washing; 80 per cent, 95 per cent, and absolute alcohols with 1 per cent ammonia added to each, ten days alto- gether, to insure thorough dehydration; pyridin four days; silver nitrate, after washing, ten days; four per cent pyrogallol in 5 per cent formalin two days. All of these fluids except the last, were changed several times. In several of the preparations the strength of the silver solution was varied, beginning with a solu- tion of 2 per cent for several days, then reducing to 1 per cent, to 0.75 per cent, and finally back to 2 per cent. The results in the way of details of structure and of staiming the finer fibers, which were obtained by this modification were superior to those given by’the unmodified procedure. Fig. 2 Right nervus terminalis from same kitten of which the left nerve is shown in figure 1. Removed and mounted entire. X 32. 1. The nervus terminalis of the cat The nervus terminalis of the cat is found on the medial side of the olfactory stalk. The main trunk runs parallel with the ventral border of this stalk, between the fissura prima of the forebrain and the vomeronasal nerves. As shown in figure l, which represents the mesial aspect of the forebrain and nasal septum of a kitten of two weeks, the nerve is connected with the brain by three strands (r},r?,r?). They follow closely parallel to blood-vessels of small calibre, which enter the brain near the fis- sura prima. When these vessels were cut at their points of en- trance into the brain, the nerve strands also became detached. This was due to the minuteness of the strands which it was not found possible to sufficiently disentangle from the connective tis- NERVUS TERMINALIS: MAMMALS 17 sues surrounding both vessels and nerves. Sections (figs. 3 and 23) indicate that the larger strands at least enter the brain inde- pendently of the blood-vessels. A fourth strand (r4), which joins the nerve trunk, appears not to enter the brain. This strand was traced caudally for a short distance along the anterior cerebral artery, but became so atten- uated by separating into minute bundles of fibers that it was not possible to follow the divisions far. In one of the specimens examined, a kitten one day old, cells similar to ganglion cells were observed along the course of the roots for a little distance within the brain (fig. 23). An elongated ganglionic swelling (fig. 1, gn’) is shown rostrad to the point where the nerve strand es the anterior cerebral artery unites with the main trunk of the nerve. Further rostrad a larger ganglion (gn) is seen in close proximity to the most cau- dal of the three principal vomero-nasal bundles. Between these two ganglionic masses the main trunk of the nerve breaks up into a plexus of nerve strands (a, b,c). Many of these follow the larger blood-vessels of the region and send twigs into their walls. Three of the largest strands of the plexus con- verge distally, uniting with vomeronasal bundles. A number of strands, finer than any represented in the figure were found, but were torn in dissection. They were composed of relatively few fibers each, and uniting with the larger strands, formed a loose plexus over the medial surface of the olfactory bulb, as shown in figure 3. The more ventrally located (c) of the larger pimlles divides into two strands which unite, one with the ventral bundle of the vomeronasal nerve, the other with a more dorsally lccated bundle of the same nerve. i While the three strands (a, b, c), already noted, diverge at various angles from the principal axis of the nerve, the other divisions do not depart so widely. As shown in figure 1 (e), they form a secondary plexus which reunites, with the exception of one small strand, into the ganglionic swelling (gn) already noted. The single bundle which continues rostrally from this ganglion crosses two of the vomeronasal bundles to become en- THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL, 30, NO. 1 OLOF LARSELL 18 -umpukg *09}4TY PlO-syooM-omy Jo pvoy Ot} JO 4aed pie auvd uo uo pozoford stjeurUL1e} SnALoU ey} JO snxe]d ‘onbruyoe} TOATIS MIOJ YSNOLY} SUOTIIOS [V}IIGVS JO SOLOS & WLOdF [BIUBIOVIPUT OY} JO UOTPNAYSMOSOY & ‘SI du) NERVUS TERMINALIS: MAMMALS 19 cased within the same connective tissue sheath which surrounds the third, more rostrally situated, vomeronasal bundle. The small strand which fails to reunite with the plexus appeared. to have been torn from its course distally along a blood-vessel which passes between the dorsal surface of the olfactory bulb and the cerebral hemisphere. Essentially the same relations to the vomeronasal nerves and to the cribriform plate were found in a dissection of a half-grown kitten, not figured. In this specimen the left olfactory bulb was removed, and it was attempted to trace strands of the terminalis into the nasalseptum. Six strands which clearly belong to the nervus terminalis were present. Four of these became related to the vomeronasal nerves, and one of these four remained suf- ficiently separated from the dorsal bundle of this nerve so that its course could be followed distinetly through the cribriform plate. Most of the strands became enclosed by the sheaths of the vomeronasal bundles in such a manner that it was not possible to distinguish them from the strands of the latter nerve in their course peripherally by the method of dissection. Two of the strands which passed more dorsally did not converge with the vomeronasal bundles. One of these passed through one of the more dorsally situated foramina of the cribriform plate, together with a large olfactory bundle, and its course on the nasal septum was traced for some distance. ‘The other continued dorsally and became attached to an artery which lay in the furrow between the olfactory bulb and the cerebral hemisphere. Figure 3, which represents a graphic reconstruction of the terminalis plexus between the vomeronasal nerves and the point where its roots enter the brain, shows essentially the same relations. This figure represents a composite of thirteen sections of the region of the forebrain and nasal septum of a kitten two weeks old, prepared by the pyridin-silver method. It supple- ments figure 1 by showing the finer strands of the plexus to which reference was made, and by bringing out numerous small ganglia which could not be seen in the dissection represented in figure 1. A comparison of figure 3 with figure 2, which repre- sents an in toto amount of the right nervus terminalis of the same 20 OLOF LARSELL kitten from which figure | was drawn, is interesting in showing the same manner of distribution of the larger strands as is seen in the reconstruction. It also indicates, somewhat more clearly, the position of the ganglion cells along the main trunk of the nerve. This specimen differs from the one illustrated in figure 3 in that the main ganglionic mass (gn) has fewer ganglion cells than are present in the corresponding ganglion (gn, fig. 3) of the other kitten of the same age. More numerous cells, however, are scattered along the nerve trunk, so that the total number is approximately the same in the two specimens, if the smaller ganglia, not observed in the dissected animal, are left out of consideration in both. The finer strands, which radiate in various directions from what may be designated the central bundle, follow along or soon reach, blood-vessels of various sizes. Many similar strands, consisting of but three or four fibers could not be seen with the low magnification of the projection apparatus used in plotting the figure, and are not included in this reconstruction. These, if represented, would make the plexus much more intricate, especially in its rostral part, and would cause it to extend further rostrally over the olfactory bulb than is figured. The course of the nervus terminalis in the nasal septum was followed to best advantage in methylene-blue preparations, fixed in ammonium-picrate and mounted in a mixture of ammonium picrate and glycerine. The silver preparations brought out more clearly the finer strands, but the distortion of the septum produced by this technique made it difficult to follow the general course of the various branches by the method of reconstruction. The nasal septum of a kitten one day old was removed and was kept moistened for forty minutes in a 0.25 per cent solution of methylene-blue in physiological salt solution. The preparation was examined from time to time until a differentiation was ob- served between the main bundles of the vomeronasal nerves and the smaller bundles which course parallel to them and which had previously appeared to be part of them. These smaller bundles assumed the blue color characteristic of this stain, while the vomeronasal nerves remained practically unstained. After NERVUS TERMINALIS: MAMMALS 21 fixation over night in ammonium-picrate, the mucosa was re- moved from the bony septum and was mounted whole. This was done with the mucosa from both sides of the septum. The two sides showed essentially the same. picture, but the right side was somewhat clearer, and is illustrated in figure 4. In this specimen the vomeronasal nerves (n.vom.) consist of three principal bundles in their proximal course on the septum. About midway toward the organ of Jacobson, which could not be removed with the mucosa without too great danger of injury to the latter, two of these bundles divide into secondary strands. These strands continue to the vomeronasal organ. Parallel with the vomeronasal bundles and in close proximity to them for some distance are the main strands of the nervus terminalis. These strands pass through the cribriform plate, as previously shown in the dissections and as verified by pyridin-silver prepa- rations, in company with the vomeronasal bundles. They con- tinue parallel with them for some distance (fig. 4, n.fer.) and then divide, forming an intricate plexus in the deeper part of the mucosa (fig. 4, p.pl.). Comparison with silver preparations of the nasal septum makes it evident that only a portion of the plexus was stained in this specimen. This portion was derived chiefly from the most dorsal (d.n.fer.) of the principal terminalis strands present. The median of these strands (m.n.fer.) gives off some small twigs of fibers which anastomose with the main trunk of the dorsal bundle, and more rostrad it breaks up into branches, one of which (r.med.) forms a portion of the plexus. A larger branch of this median bundle continues parallel with the median branch of the vomeronasal bundle, but could be fol- lowed for only a short distance rostrally. The most ventral bundle of the terminalis (v.n.ter.), which is also the largest, was lost distally because of the idiosyncrasy of the stain. A small twig (r. ven.) given off in the more proximal part of its course passes beneath the ventral vomeronasal bundle and is soon lost in the mucosa ventral to this bundle. A large branch (r.dor.) from the dorsal terminalis bundle also courses ventrally, but this could not be traced beyond the dorsal ramus of the ventral vemeronasal bundle. All other branches which were stained OLOF LARSELL N N ‘02 X “QuNoU sjoY AA “UTe4S anjq-ousAyPL “Plo ABp ouo u0zTy B jo wNgdes [eseU UO ST[VUTUTIOZ SNAIOU Jo uorngqiaysip jeroydiueg fF “Bi NERVUS TERMINALIS: MAMMALS a coursed toward that part of the mucosa which lay dorsal to the vomeronasal bundles, where the greater part of the plexus is located. Examination of the figure indicates that many of the nerve strands follow quite closely the paths of the blood-vessels represented by broken lines. Owing to the complexity of the vascular network, only the larger of these vessels were seen clearly enough in the preparations to make it possible to trace their courses. No ganglion cells or small ganglia were seen, but this was laid to the peculiarity of the stain. Silver and Wei- gert preparations revealed the presence of such cells in the nasal septum, but in much smaller numbers than are indicated in the rabbit by Huber and Guild (7138) or in the human by Brookover (17). The ganglion clusters are, however, very numerous in- tracranially, especially on the mesial sides of the olfactory bulb. While it seems likely that most of the plexus formed by the nervus terminalis on the nasal septum was seen, it is doubtless true that the rostral part of the septum, which unfortunately did not take the stain, also contains a continuation of this plexus. Pyridin-silver preparations indicate this beyond question in the cat, and it has been shown to be true in the rabbit by Huber and Guild, and in the human by Brookover, in the papers above cited. The observations of the olfactory region of the mucosa are more dubious. In the methylene-blue preparations, the region in which olfactory fibers were present was stained a diffuse dark blue-green, which made it impossible to see any portion of the plexus if it were present. The pyridin-silver material shows oc- casional fibers in this region which may belong to the nervus terminalis, but this cannot be stated with any degree of certainty. It is possible that they are fibers from the anterior ethmoidal nerve, the main trunk of which lies in close proximity to many of the fibers found. . So far as the methylene-blue material indicates, there is no connection of the nasal plexus of the terminalis with either the anterior ethmoidal or the nasopalatine branches of the trigemi- nal nerve. The silver preparations, however, showed such a confusion of trigeminal and terminalis fibers in the rostral end 24 OLOF LARSELL of the septum that it seems certain that there is some comming- ling of fibers from the nasopalatine nerve in the terminalis plexus. This was also indicated, although somewhat less clearly, in Weigert preparations. The Weigert material showed very clearly that fibers from the anterior ethmoidal nerve take some cen. Fig. 5 A small cluster of cells slightly posterior and ventral to the ganglion (gn.) of figure 3, illustrating some of the types of ganglion cells, together with myelinated and unmyelinated fibers. Pyridin-silver technique. X 825: part in the formation of the peripheral plexus of the nervus terminalis. As shown in figure 22, which represents a portion of this plexus in a kitten one day old, a few myelinated fibers are present. Some of these were traced into a strand of the anterior ethmoidal nerve, which lay in close proximity to the NERVUS TERMINALIS: MAMMALS 25 portion of the plexus figured. This nerve shows development of the myelin sheaths, while the intracranial portion of the nervus terminalis of this specimen (fig. 23) gave no indication of myelin sheaths as brought out by the Weigert treatment. It does not seem likely that such sheaths would be formed in the peripheral portion of the nerve at an earlier date than they are formed in the part of the nerve nearer the brain. It is therefore assumed that they are fibers belonging to the already myelinated anterior ethmoidal nerve. So far as this material indicates, the central roots of the ter- minalis, which are easily followed in the sections to their points Fig. 6 Two bipolar eells and some of the nerve fibers from periphery of main ganglion (fig. 3, gn.) of the nervus terminalis in kitten of two weeks. Pyridin- silver technique. X 1266. of entrance into the brain, are composed entirely of unmyelinated fibers. In kittens of two weeks, myelinated fibers are found in the intracranial plexus, although in these also no clear evidence of such fibers was found among the strands which enter the brain. To avoid as far as possible the entrance of fibers from the fifth nerve as a factor, the greater part of the histological studies to be described was confined to the intracranial plexus and ganglia of the terminalis. Histological. The plexiform character of the nerve in the cat, together with the structure of the ganglion cells to be described 26 OLOF LARSELL in the mule, and the relation of the nerve to the adjacent blood- vessels shown in cat, mule, and beef, suggested strongly that the nervus terminalis is composed, at least in part, of sympa- thetic fibers. There remains the possibility, which could not be adequately tested in the available equine or bovine material, that there may be a general or special sensory component, in addition to the motor and sensory sympathetic fibers which were found. It was accordingly deemed advisable to make as thorough a study of the terminalis ganglia and of the fibers con- nected with them in the cat as the material available would permit. The fact that the nerve is situated in a position so difficult of access, together with its small size, and the further circumstance of its relation to the cribriform plate, made necessary a technique permitting of decalcification, so that nerve and ganglia might be studied in situ. For this reason, chiefly, the pyridin-silver method, as previously described, was employed. There was considerable variation in different parts of the same section in the intensity and clearness of the impregnation. It was also found that, in general; the cells of the smaller ganglionic clusters were much better differentiated from the background than were those in the more crowded larger ganglia. Because of this fact the majority of the cells figured are from the small clusters of cells, but for comparison, considerable attention was paid to the large ganglionic mass (figs. 1, 2 and 3, gn.) which appears to correspond with the ‘ganglion terminale’ of authors. Types of ganglion cells. Figure 5 represents a typical small ganglionic mass which had its position in the meninges covering the mesial side of the olfactory bulb. It lay slightly caudad to the ganglion terminale and a little more ventrally. This clus- ter was similar to numerous others scattered throughout the plexus. Such clusters of cells are usually situated at the meet- ing point of several small strands of fibers which converge from various directions. The group of cells figured represents only a portion of this particular ganglionic mass. The remaining portion was to be seen in the next section of the series. NERVUS TERMINALIS: MAMMALS 27 It will be noted that both myelinated (my.) and unmyelinated (unmy.) fibers are present. The axis cylinders of the myelinated fibers were stained a darker orange color than were the myelin sheaths surrounding them. The processes of the unmyelinated fibers were quite black and stood out distinctly. One of the latter (a), which comes from the direction of the central connec- tion of the nerve, divides into two smaller fibers which in turn subdivide into terminations with small varicosities, and which form simple pericellular baskets on two of the nerve cells shown. Several classifications, both of ganglion cells and of sympa- thetic cells, have been made by investigators, notably by Cajal (05) Dogiel (08), and Ranson (12) for the former; and by . Cajal (05), Carpenter and Conel (14), Dogiel (’96), Michailow (11), and others for the latter type. There is considerable in- dividual variation among the ganglion cells observed in the ter- minalis clusters, and they come within one or the other of these classifications. Still, for purposes of description it is convenient to designate the types observed according to the number of proc- esses they possess as unipolar, bipolar, and multipolar. A few binucleated cells were seen in the cat, but aside from the num- ber of nuclei, they resembled the other cells of the several varie- ties and will not be treated separately. (a) Unipolar cells. Most of the unipolar cells observed re- semble those usually considered characteristic of the spinal ganglia. The body of the cell (figs. 5, 10, 18) is ovoid or spheri- cal, with a rather large nucleus. The single process, which usu- ally stained brown near the cell body, becomes darker as it assumes a smaller diameter in its course away from the peri- karyon. In those cases in which it was possible to follow it for any distance, it divides into two processes, one directed in the general direction of the central connection of the nerve, the other peripherally as respects this connection. No marked dif- ference in size of these two divisions was noted, although that which appeared to be the central process was usually slightly smaller in diameter. It must be understood that only the gen- eral direction of the course of these fibers is indicated, because of the various directions different strands of the plexus assumed. 28 OLOF LARSELL While myelinated fibers were present in the same bundles which included processes from the unipolar cells, in no case was a myelin sheath found in connection with processes from such cells. Three distinct sizes of unipolar cells were observed. The predominating size is represented in figure 5 (un.c.), which also Fig. 7 Bipolar cell and accompanying unmyelinated fiber on the wall of a blood-vessel between the cerebral hemisphere and the olfactory bulb. Kitten two weeks old. Pyridin-silver technique. X 450. Fig. 8A Characteristic small bundle of myelinated and unmyelinated fibers, with a single bipolar cell in its course. Fig. 8B Portion of a myelinated fiber more highly magnified, showing a node of Ranvier, and some large varicosities. Both from kitten of two weeks. Pyridin-silver technique. Figure 8A, X 660; figure 8B, X 950. indicates very clearly the bifurcation of the single process of the cell. Figure 5 (wn.c’.) also shows one of the smallest of the uni- polar cells seen. The process of this cell could not be followed for any great distance and no bifurcation was seen. The proc- ess was directed peripherally. Attention may again be directed to the pericellular basket surrounding this cell. The relation of the cell to the nerve fiber of which this basket is the termination, NERVUS TERMINALIS: MAMMALS 29 together with the peripherally directed process of the cell, sug- gests that it pertains to the sympathetic system. Its small size favors this interpretation. The largest unipolar cell observed (fig. 10) was also the only one of this type found which was binucleated. The process was of large diameter and did not stain so dark as did the majority of fibers. It was not possible to follow it beyond its point of entrance into the bundle of smaller black fibers of which it became a part. The large size, both of cell and of cell process, resembles somewhat the large unipolar cells found by Carpenter (12) in the ciliary ganglion of the sheep. . Unipolar cells were not numerous and were found only in the smaller ganglia. Whether their presence in the larger ganglia was hidden by the crowded condition of these could not be de- termined. It seems, however, that the latter are composed principally of multipolar cells, with fairly numerous bipolar cells near their peripheries. (b) Bipolar cells. Cells of this type are quite numerous, both in the smaller ganglia and in the larger ones. A number of such cells isolated from other nerve cells were also found. Figure 6 represents two typical bipolar cells from the periphery of the largest ganglion (fig. 3, gn.) of the left nervus terminalis of a kitten of two weeks. The processes of these, which are of large size, took only a brownish tinge from the impregnation. A few fibers of small size which stained black were present in the im- mediate neighborhood of these cells and are indicated in the figure, as are also some large unmyelinated fibers. These are doubtless fibers of similar bipolar cells. A few relatively small bipolar cells were found, one of which is shown in figure 5 (67.c.). The processes of these were quite slender and stained black. They were followed for some distance, but no conclusive evi- dence as to their terminations was obtained. The isolated bipolar cells above noted were found in the course of small strands of fibers, between the nodal points where several such strands converge (fig. 8). A few were found on the walls of blood-vessels or near them. One of these is illustrated in fig- ure 7, in which also is shown an unmyelinated fiber of small size 30 OLOF LARSELL Figs. 9 to 15 Ganglion cells from intracranial clusters of nervus terminalis of kitten two weeks old. Figures 12 and 13 were drawn from cells which lay in the trunk of the nerve between its points of entrance into the brain wall and the main portion of the plexus; figures 14 and 15 were drawn from cells which lay near the center of the main ganglion (gn.) and the other figures were drawn from cells in various portions of the plexus. Figures 9, 10, 11, and 14 magnified 990 times; figures 12 and 13 magnified 1020 times, and figure 15 magnified 675 times. Pyridin-silver technique. NERVUS TERMINALIS: MAMMALS ol which runs parallel with the processes of the cell. The position of this cell was well within the crevice between the olfactory bulb and the cerebral hemisphere, on the wall of a blood-vessel which passes between the bulb and the hemisphere. The processes of the cell were stained rather lightly by the silver and it was not possible to follow them far. Their course so far as visible was parallel with the wall of the artery. Two slender processes may be seen to issue from the cell, in addition to the polar processes of much larger size. Because of these slender offshoots the cell should possibly be classed with those of multipolar type, but it is included with the bipolar cells because of its greater similarity in other respects. In figure 8A is shown a cell whose position was on the medial surface of the olfactory bulb, posterior and ventral to the main ganglion of the terminalis. The larger process (p.pr.) is directed toward the ganglion, while the very slender process from the opposite pole of the cell is turned in the general direction of the central connection of the nerve with the brain, although the group of three fibers of which it forms one, turns at right angles to the principal axis of the terminalis plexus. ‘This process was followed for some distance, but it was lost in the plexus centripetally. (c) Multipolar cells. The multipolar type of cell predomi- nates in the ganglia. These cells vary in size from the relatively small ones shown in figure 16, to the large cells drawn to the same scale represented in figures 9, 11 and 17. The number of processes varies. The cells illustrated in figures 12 and 13, which were found on the main trunk of the nerve, show but three processes. The majority of multipolar cells included in the small ganglia have at least five offshoots. Typical examples of such cells are shown in figures 11 and 16. The axone could not always be determined with certainty, but in the cells shown in figures 11, 12, 14, and 15 the process marked az. appeared to be the axone. In each of these cases it was directed centripetally. That which appeared to be the axone of the cell shown in figure 13 (ax.) was directed peripherally. This cell lay in the course of the main trunk of the nerve. cen. JA ioe LEE Y, \7 Fig. 16 Small ganglionic cluster and intercellular network from portion of intracranial network of nervus terminalis near lower posterior border of olfactory bulb. Kitten two weeks old. Pyridin-silver technique. X 990. Fig. 17 Large multipolar cell with extracapsular network from intracranial plexus of nervus terminalis of kitten two weeks old. Pyridin-silver technique. x 990. 32 NERVUS TERMINALIS: MAMMALS oe Very few binucleated cells of the multipolar type were ob- served. These, as illustrated by the one represented in figure 9, were similar in every other respect to other multipolar cells. Fibers and fiber networks. A few of the individual cells were surrounded by a reticulum of delicate fibers which suggest an extracapsular network (fig. 17). The capsule itself was not easy to see in such cases, but the position of the capsule nuclei within the network seems to justify the interpretation given to these reticula. In the example illustrated, a small number of threads extend from the region of a neighboring bipolar cell and take part in the formation of the pericellular plexus described. The main fiber from which these threads have their origin, runs par- allel to the larger process of the bipolar cell, and appears to ter- minate in a spiral (fig. 17) about this process. The fibers form- ing this network are intertwined in the most confusing manner in every direction. They are of very small size. ‘ Intercellular networks also were found in those parts of the preparations where the impregnation was most favorable. A peculiarity of the impregnation revealed itself in the fact that those parts of the sections which showed the fibers most clearly did not serve so well to differentiate the outlines of the cell bodies and of the large processes from the perikaryon. The intercellular networks were found in every case at nodal points where a number of fiber strands converge. Usually the ganglion cells enclosed by such a network were of small size, but occasionally a larger cell was also included. In the example il- lustrated (fig. 16), which was situated near the lower margin of the posterior portion of the olfactory bulb, four strands of fibers converge about a small group of cells. Many of the individual threads may be followed from one strand through the cell cluster and into one of the other converging strands, without any ap- parent connection with the cells. The majority of fibers were lost in the network. A few may be seen to connect with nerve cells of the cluster. This intercellular network resembles to a considerable degree structures of a similar nature found by Dogiel (’95) in the di- gestive tract of the dog. It bears an even more striking resem- THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 1 34 OLOF LARSELL ape NCE CD KC ioe “ ESO Ren AFR E3 ! j ae 7 Be eS ZR Fig.18 Bundle of fibers with two cells in its course, and showing the charac- teristic method by which small strands of fibers leave the larger bundles. Kitten two weeks old. Pyridin-silver technique. X 675. Fig. 19 A somewhat isolated strand of myelinated fibers and a single unmye- linated thread of relatively large size as compared with the centrally directed process of the nerve cellshown. Kitten two weeks old. Pyridin-silver technique. x 990. NERVUS TERMINALIS: MAMMALS 30 blance to networks found by Ranson and Billingsley (718) in the cervical ganglion of the dog. An attempt was made to analyze the various strands of the plexus, in order to determine, if possible, the types of component nerve fibers. As previously stated, both myelinated and un- myelinated fibers are present. Of the latter type, both varie- ties, namely, fibers of Remak and naked filaments, are abundant. While for short stretches some of the strands appear to be com- posed exclusively of one type or the other (figs. 5, 6, 8, 19), the rule is mixed strands. The fibers of Remak predominate as to number when the entire plexus is considered. They are particu- larly numerous near the main ganglion, and appear to include the majority of fibers which enter this ganglion. In figure 20 are shown two converging strands consisting prin- cipally of Remak’s fibers, which approach this ganglionic mass. The bundle resulting from their union was one of the most com- pact of the entire plexus. Naked filaments of several sizes are also included in it. One of the most slender of these (fi) is seen to approach the nerve cell represented in the figure and to follow what appeared to be the axone of the cell, to end on the perikar- yon in a simple pericellular termination. A similar strand, but with fewer naked fibers, is shown in figure 21, which was drawn from near the ventral border of the posterior part of the olfac- tory bulb. This bundle of fibers runs parallel with one of the arteries of this region. An offshoot (ram.) consisting of four or five threads leaves the main strand and passes to the wall of a branch of the artery. Two fibers which show no neurilemma sheath leave the strand to pass to the wall of the main artery. These strands so closely resemble others, the terminations of which are described below, that it seems certain that they represent fibers which ramify to form the type of nerve-endings shown in figures 24 to 29. ' Whether or not the fibers of Remak terminate in the sensory end- ings represented in figures 28 and 29 could not be determined with certainty. It seems unlikely, in view of the fact that the fibers leading to the sensory terminations show myelin sheaths near their endings. The naked filaments show considerable 36 OLOF LARSELL variation in size, the finer fibers greatly outnumbering those of coarser diameter. The myelinated fibers are relatively few in number. They usually occur in strands of three to five filaments, accompanied C3her RK NZ es SOSA WN ge Sis) sre Ais ators PRAT SA art.w. 21 Fig. 20 A bundle of fibers from near the dorsal margin of the olfactory bulb, caudad to the main ganglionic mass. The strand divides centrally. Kitten two weeks old. Pyridin-silver technique. X 600. Fig. 21 A bundle of fibers from near ventral margin of olfactory bulb in about the same vertical plane as figure 20, showing relation of some of the fibers to a blood-vessel. Kitten two weeks old, Pyridin-silver technique. X 600. NERVUS TERMINALIS: MAMMALS 37 by naked threads. In some cases these naked filaments repre- sent fibers which have lost their myelin sheaths (fig. 8A, az. cyl.). Many of them show very large varicosities (fig. 8, A and B), the most pronounced of which are often found near the nodes of Ranvier. These varicosities appear to have been pro- duced by unequal shrinkage of the axis cylinder, probably dur- ing the process of fixation of the tissue. At other points, as shown in the figures, the cylinders are extremely attenuated. Some show spiral formations of the axis cylinder within the myelin sheath (fig. 8A, sp.), In diameter the intracranial myelinated fibers varied from 1.5 yu to.2.6 yu. In the septal plexus of the terminalis, myelinated fibers are found intermingled with the unmyelinated threads (fig. 22). As previously stated, these belong, in part at least, to the tri- geminus. It is possible that the myelinated fibers of the intra- cranial plexus are related to those found in the septal plexus, and may therefore be trigeminal in orig. This does not seem likely, but can only be adequately tested by degeneration experi- ments, which the writer hopes to perform. The roots which enter the brain, in both of the extra-uterine stages of growth of the cat in which this point was examined, appear to be composed exclusively of unmyelinated threads (fig. 23). Fibers of Remak predominate, but a few naked filaments are mingled with them. Nerve terminations. There are present in the cat two kinds of nerve terminations in the walls of the cerebral blood-vessels, which are connected with fibers from the nervus terminalis. A third type consisting of free endings in the epithelium of the nasal septum appears also to be related to this nerve. The nerve terminations in the walls of the anterior cerebral artery and its branches, for convenience of description, will be designated as type I and type II. Type I (figs. 24 to 27) consists of delicate varicosed fibers which penetrate the muscular walls of the ‘blood-vessels from the nervous plexus surrounding these vessels. At varying depths in the muscular layer, the fibers which penetrate the arterial coat ramify into very fine arborizations which pass between the smooth 38 OLOF LARSELL muscle cells and end on the latter. The nerve fibers which end in this manner are in every case unmyelinated. ‘They show very slight typical varicosities. The twigs of the terminations are Fig. 22 Portion of septal plexus showing the presence of myelinated fibers. Midway between principal vomeronasal foramen and rostral end of nasal septum. Kitten one day old. Weigert technique. X 230. Fig. 23 Central roots of nervus terminalis at point of entrance into brain wall. Kitten one day old. Weigert technique. X 75. also varicosed. It will be noted from the figures that the man- ner of distribution of these terminations varies considerably. Those represented in figures 25 and 27 end in relatively short, NERVUS TERMINALIS: MAMMALS 39 stout twigs. Others (figs. 24 and 26) have long, very delicate branches, which sometimes continue their course parallel with the plane of the fibers from which they spring, sometimes di- art.w. 25 Fig. 24. Motor nerve termination from muscular coat of anterior cerebral artery of kitten two weeks old. Gold-chloride technique. 1020. Fig. 25 Motor nerve termination from branch of anterior cerebral artery of kitten two weeks old. Gold-chloride technique. > 1020. Fig. 26 Motor nerve termination from one of the vessels on the mesial sur- face of the olfactory bulb (in pia mater) of kitten two weeks old. Pyridin-silver technique. X 1425. Fig. 27 Motor nerve termination from one of the blood-vessels of the nasal septum of kitten one day old. Pyridin-silver technique. X 1425. Fig. 28 Sensory nerve termination from anterior cerebral artery of kitten two weeks old. Gold-chloride method. X 1020. Fig. 29 Sensory nerve termination from a small vessel in the meninges near the olfactory bulb of kitten two weeks old. Pyridin-silver method. X 1425. verge from it at various angles. These terminations appear to be similar to those found by Huber (’99) in the cat, and consid- ered by him to be motor endings. 4() OLOF LARSELL The terminations represented in figures 24 and 25 were stained by the gold-chloride method. Those shown in figures 26 and 27 are from pyridin-silver preparations. Similar endings were also found by the molybdenum methylene-blue process, in the an- terior cerebral artery and its branches, from which vessels all of the preparations were made. The type II endings are strikingly different in appearance from those of type I. As shown in figures 28 and 29, the termina- tions are by somewhat spindle-shaped structures, composed of short, thick branches from the main fiber. These rami end with terminal knobs. In the gold-chloride preparations (fig. 28) a spindle-shaped clear space appears to be enclosed by the short processes which are derived from the nerve fiber. In the silver material no such clear space is evident, although the general contour of the termination is the same as in the gold preparations. The pyridin-silver slides showed the presence of delicate myelin sheaths on the fibers leading to these terminations. Such sheaths are not clearly evident in the gold chloride material of the cat, although similarly prepared slides of the corresponding blood- vessels of the beef indicate their presence in that animal. No capsules are present around these terminations in any of the animals in which they were examined. The methylene-blue staining did not clearly demonstrate this type of endings, al- though suggestions of them were visible by this method also. In general appearance these end-organs resemble to some ex- tent the corpuscles of Ruffini, but are much smaller. Both in shape, however, and in the absence of a capsule, they bear a stronger sunilarity to a type of sensory ending found by Dogiel in the heart of the cat (Dogiel, ’96, fig. 2, D). Smirnow de- scribes terminations of somewhat similar appearance in the atrial endocardium of the cat (Smirnow, 95, fig. 6), and Michailow (08) has also described non-capsulated sensory terminations in the myocardium. In the anterior cerebral artery and its branches of the cat and of the beef, they lie not in the loose connective tissue, but seat- tered at various levels in the muscular coat itself. NERVUS TERMINALIS: MAMMALS 41 So far as the writer is aware, similar structures have not heretofore been described in the walls of the cerebral blood- vessels. Huber (’99) noted myelinated fibers along the walls of the cerebral arteries of the cat, and considered them to be sen- resp. epith. 3| Fig. 30 -Free nerve terminations in respiratory portion of nasal mucosa of kitten one day old, also a portion of the nervous plexus. Pyridin-silver tech- nique. > 600. Fig. 31 Portion of the septal plexus of the nervus terminalis of kitten one day old, showing one of the fibers bifurcated and ramifying into slender twigs which appear to have been cut off at their ends. Pyridin-silver technique. X 600. sory as distinguished from the unmyelinated motor fibers which he found in company with them. The compact form of this type of endings, differing to so marked a degree from the other type which has been described as motor, and closely resembling sensory terminations in other 42 OLOF LARSELL parts of the vascular system, seems to justify the assumption that they are sensory in function. The third type of ending to which reference was made was found among the epithelial cells hLning the mucosa of the nasal septum. These endings consist of very delicate arborizations which pass between the columnar cells of the epithelium and approach the surface of the membrane (fig. 30). They are ter- minal twigs of fibers which appear to be unmyelinated. These fibers, as shown in the figure, approach the epithelium in small strands of three or four fibers to spread out at its base, where the terminal threads which form the free end fibers are given off. No.varicosities or end-knobs were seen. Such terminations are present in both sensory and respiratory regions of the septal mucosa and in the epithelium of the vomero- nasal organ. Similar endings, but with varicosities or end- knobs, have been described and figured in these regions by von Brunn (’92), von Lenhossék (’92), Retzius (92), Cajal (94), Read (’08), and others. Most of these writers tend to ascribe them to the trigeminal nerve, although von Lenhossék suggests the possibility that they represent olfactory fibers whose cells of origin do not have the same position as others, but le within the centripetal olfactory tract, enclosed in the course of the olfactory bundle. Figure 31 represents what appears to be the centripetal con- tinuation of a fiber which gives rise to free endings such as those just described. This figure was drawn from a section which lay just below the epithelium, in a sagittal series through the nasal septum. As shown in the figure, the fiber (fi) divides at its extremity into four slender twigs which appeared as if they had been cut near their tips. Centripetally this fiber unites with a similar one (fi’). The nerve process of which these fibers are branches is part of a small bundle (p.pl.) which forms a portion of the terminalis plexus of the nasal septum shown in figure 4. While the fibers which terminate in the manner indicated re- semble in size and distribution those of the nervus terminalis, there remains the-possibility that they are the continuation of the more delicate threads which are present in the nasopalatine NERVUS TERMINALIS: MAMMALS 43 and anterior ethmoidal branches of the trigeminal nerve. As already noted, there is a commingling of fibers of this nerve and of the terminalis, and fibers of the trigeminus enter the bundles which constitute the septal plexus of the terminalis. The intri- eacy of this plexus made it impossible to follow any individual fiber very far. Accordingly it was not possible to determine with certainty whether the free terminations of the septal mucosa are from terminalis fibers or from the trigeminal nerve. Other fibers of larger size are also present in the mucosa. These terminate on or near the septal glands, and show varicosities on their finer twigs. They also enter into the plexus of the termi- nalis to some extent. They are given off from the fifth nerve. It is usually stated that the septal glands are innervated by the trigeminus, and these fibers appear to be the ones by which this is accomplished. 2. The nervus terminalis of the beef The nervus terminalis of the beef, as shown in figure 32, lies median to the olfactory nerves, between the meninges and the ventral brain surface. Running parallel with it are branches of the anterior cerebral artery. In the specimen figured the greater portion of the left nerve is a compact bundle, while the right nerve is composed of several strands for the greater part of its length. In the numerous brains examined there was con- siderable variation in the relation of the different strands which by their union form the main nerve bundle. This variation was found not only when comparing one brain with another, but, as just indicated, on comparing the two nerves of the same speci- men. In. some cases the strands were independent up to a short distance from the forward margin of the hemispheres, in other cases the bundle was formed much further caudad. For the greater part of its course the portion of the nerve pres- ent in the specimens was covered by the meningeal membranes. It emerges to the outer surface of the arachnoid coat at about the point where the cerebral hemispheres begin to curve upward. At this point the nerve was always compact in a single bundle. 44 OLOF LARSELL This bundle after emerging lay free on the surface of the arach- noid. The appearance of the free end indicated that it had been stretched and broken in removing the brain from the cranial cavity. Several ganglionic swellings are visible, the largest one in the brain from which the figure was drawn, at the point (fig. 32, gn.) where the nerve crosses the artery. Just caudad to this ganglion, as more clearly shown in figure 33, the bundle divides into two strands. A short distance rostral to this ganglionic bg arse eS ant.cerart. (a Fig. 32. Ventral view of anterior portion of the beef brain, showing the nervus terminalis and a portion of the anterior cerebral artery and some of its branches. Nerve strands traced along arteries to points marked z. mass, a similar strand is given off from the main trunk. These strands follow the branches of the anterior cerebral artery, as in- dicated in figures 32 and 33, ramifying to the secondary branches of these vessels, and at intervals they give off fine twigs which penetrate into the muscular walls of the arteries. Several of the strands were traced continuously to the points marked «x in figure 32, where the respective arteries dipped into the fissures. On the same brain, and on others, similar strands were found on all of the arteries examined in this region of the Fig. 33 Portion of left nervus terminalis and related vessels shown in figure 32, enlarged to show the distribution of strands to the walls of the blood-vessels. 45 46 OLOF LARSELL brain, both on the ventral surface and in the sagittal fissure. In all cases in which their connection could be determined, when followed toward the ventral surface of the brain such strands led to the principal branches of the anterior cerebral artery, and connected with the main nerve bundle of the terminalis. When the overlying connective tissues were successfully removed the white nerve strands stood out clearly against the reddish brown of the blood-vessels. The larger bundles were followed with comparative ease even with the unaided eye, and were not easily torn. It is unlikely that all of the finer bundles were left intact. The relatively long stretches on some of the arteries shown in figure 33 where no twigs are represented probably indicate areas where they were inadvertently torn in dissection. Those shown in other parts of the vessels could be raised upon the point of a needle and stretched in such a manner as to clearly show that the finer twigs into which they ramify enter the walls of the blood-vessels. Caudally, the principal roots which by their union form the nerve trunk, follow along the larger vessels as far as the latter could be traced without cutting into the brain. In most of the specimens examined the internal carotid artery had been severed so close to the brain in removing the organ that it was not pos- sible to determine with certainty that any of it was present. There seems little doubt, however, that the continuation of strands from the main trunk of the terminalis unites with the plexus surrounding the internal carotid. A branch from the main bundle was also followed along the posterior ramus of the anterior cerebral artery as far as the genu of the corpus callosum, and similarly on other rami of this vessel nerve strands were observed. — Evidence of direct connection with the brain was difficult to obtain. Delicate branches from the nerve strands on the arteries were found occasionally to enter openings in the anterior per- forated space. These branches were extremely difficult to dis- . entangle from the mass of small blood-vessels, connective tissue, and elastic fibers among which they were found. Many ap- peared to be related to the larger vessels which enter the brain NERVUS TERMINALIS: MAMMALS 47 substance in this region. In the material examined, only one relatively large strand (fig. 32, c) was seen to enter the brain. This compared in size with the strands which enter the mule’s brain (figs. 40, 41, 42). Although the peripheral relations in the beef could not be determined in the available adult material, dissection of a num- ber of ox fetuses of 110 mm. to 140 mm: greatest length brought ge Py Fig. 34 Part of the nasal septum and the cerebral hemisphere of fetal ox of 121 mm. greatest length, showing the nervus terminalis, the vomeronasal nerves, and a portion of the vomeronasal organ. This drawing was combined from sev- eral dissections and is to that extent diagrammatic. ™X 4. out the fact that in the beef, as in the other mammals studied, the terminalis passes through the cribriform plate in company with the vomeronasal nerve, and doubtless spreads out on the septum in the characteristic plexiform manner observed in other mammals. As shown in figure 34, which represents the relations in a fetus of 121 mm. greatest length, the main nerve bundle divides 48 OLOF LARSELL on approaching the cribriform plate into three strands, each of which unites with one of the bundles of the vomeronasal nerve. No differentiation between the terminalis and the vomeronasal nerve could be seen after the two had joined. Since they are separate in the cat, rabbit, horse, and human, there can be little doubt that differential staining methods would show the two as distinct in the septal region of the beef also. Attention may be directed at this point to the two ganglia which are visible on the Fig. 35 Longitudinal section of two nerve strands accompanying anterior cerebral artery of the beef, showing myelinated fibers. Formalin fixation, iron- hematoxylin stain. Sections 10 yu. X 450. Fig. 36 Transverse section of relatively large nerve strand parallel to anterior cerebral artery of beef, showing myelinated fibers. Formalin fixation, iron-hae- matoxylin stain. Sections 10 yu. X 450. main central bundle of the nerve in the fetus and to the two roots at its central end. These roots appeared to enter the brain substance near the fissura prima. Histological. The composition of the main nerve bundle in the adult, and of the principal strands which run parallel with the several branches of the anterior cerebral artery, was deter- mined by the study of sections. As shown in figure 36, which represents a cross-section of one of the larger strands on the internal frontal branch of the anterior artery cerebral, sixteen NERVUS TERMINALIS: MAMMALS 49, very delicate myelin sheaths are present, as brought ‘out by iron-haematoxylin staining. More slender nerve strands in the same preparation showed a smaller number of myelinated fibers. In longitudinal sections (fig. 35) the myelin sheaths are shown to continue without interruption for considerable stretches. They measure from 1.5 » to 2 » in diameter. Sections of the main trunk peripheral to the ganglionic masses revealed a much larger number of myelin sheaths. In the sec- tion illustrated (fig. 37) which was stained by the Weigert method, sixty-four delicate sheaths are visible. Attention may also be called in this connection to the many fasciculi, fifteeen in num- ber, which enter into the formation of the larger bundle. Fig. 37 Transverse section of main trunk of nervus terminalis of beef, show- ing distribution of myelinated fibers. Formalin fixation, Weigert stain. 360. Great care was exercised to prevent decolorizing any of the sheaths during the process of differentiation, when those meth- ods were employed which required caution. In view of the par- tial degeneration which had taken place in some of the sheaths, it is possible that this process had reached a stage in some fibers at which the staining methods employed were no longer effective in differentiating the myelin. It is believed, however, that the majority of myelinated fibers were stained. The remaining por- tion of the nerve strand was assumed to be made up of unmye- linated fibers. These observations were subsequently confirmed in material which was fixed in osmic acid shortly after the animal was slaughtered. THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL, 30, NO. 1 50 OLOF LARSELL An attempt was made to analyze those roots which penetrate into the brain, using the osmic-acid material. The few roots which were successfully isolated were of small size. They were removed and mounted whole in glycerin. One or two myelin sheaths were observed in each case, the remainder of the root being evidently composed of unmyelinated fibers. They thus resemble, except in size, the strands on the arterial walls. art.w. my.sh; 38A 398 Fig. 388 A and B- Motor terminations in muscular layer of anterior cerebral artery of beef. A illustrates the typical appearance, B showed slight terminal knobs. Gold-chloride technique. X 733. Fig. 39 Aand B_ Sensory nerve terminations in wall of anterior cerebral artery of beef, illustrating typical spindle-like outline and myelinated fibers. A repre- sents a typical termination, B shows two smaller ones attached to a common branch. Gold-chloride technique. X 733. Nerve terminations. The frozen beef brains responded well to the gold-chloride treatment, and terminations of the twigs which enter the muscular walls of the blood-vessels were found. These nerve endings, as in the cat, are of two types. Type I (fig. 38, A and B) consists of delicate varicosed fibers which ramify and run for varying distances among the unstriated muscle cells. NERVUS TERMINALIS: MAMMALS 51 These fibers are unmyelinated, and correspond in appearance with those previously described in the cerebral vessels of the cat. They are present at various levels of the muscular coat of the arterial wall, as shown in cross-sections. They send their twigs between the smooth muscle cells in the typical manner of motor terminations in this type of muscle. As previously stated, it seems probable that they should be regarded as motor endings. The terminations which have been designated as Type II are, as in the cat, strikingly different in appearance (fig. 39, A and B) from those of Type I. The nerve fiber leading to these is mye- linated, as indicated by a reddish cylinder surrounding the pur- plish-black of the central axis. The myelin sheath in most cases continues almost to the spindle-shaped termination. The axis, after emerging from the sheath, divides into two or three main rami which give off a varying number of short processes. These processes terminate invariably in rounded knobs. The portion of the spindle-shaped organ which was not occupied by these processes appeared nearly clear or had a slight bluish tint and was slightly granular. It stood out in strong contrast to the purplish red of the surrounding muscle cells. The spindles in the beef showed considerable variation in size. Of those measured, the smallest were 4 » in diameter and 12 u in length. The largest were 6 u to 7 uw in diameter and 35 » in length. The majority appeared to be about 5 » by 25 u to 30 uw. Figure 39 illustrates one of the largest observed (A) and also’ an example of the smallest size (B) drawn to the same scale. It will be noted that the two end-organs represented in figure 39 B are the terminations of what appeared to be a branch of the larger fiber shown in the figure. A node is seen in the myelin sheath, through which the branch to these terminal organs passes. In a few instances it was possible to follow the myelinated fiber to the external surface of the artery and for a little way beyond. 3. The nervus terminalis of the mule and the horse The mule. As represented in figure 40, the nervus terminalis of the mule runs parallel to the olfactory tracts, between them and the sagittal fissure of the brain. The nerve in both the mule 52 OLOF LARSELL and the horse, as in the beef, may be easily seen with the unaided eye. As shown in the figure, the left nerve had two compact ganglia. The right nerve had a single ganglionic mass of larger size. The larger of the two ganglia of the left nerve, which was the one more rostrally situated (figs. 40 and’ 43), was about 2 mm. long and 4 mm. in diameter. The smaller ganglion was of about one-half the volume of the other. The peripheral ends of both nerves were frayed and had the appearance of having been broken, doubtless where they sub- divided into strands similar to those observed in the cat and in the horse. The anterior ends of the olfactory bulbs had been torn off in removing the brain from the cranial cavity, so it was not possible to study the relation of the terminalis to the vomero- nasal nerves in this specimen. The main bundles of the nerve lay outside the pia mater as far caudally as several millimeters ros- trad to the ganglia. Here they pierced the pia and in their fur- ther course centrally lay between this and the brain surface. Immediately caudal to the posterior ganglion of the left nerve, and slightly more caudad to the single ganglion of the right nerve, the main bundle divides into ‘a number of strands or rootlets. For convenience, the description will be confined to the left nerve, but in all essential respects, unless indicated to the con- trary, the same statements apply to the right nerve also. As shown in figures 40 and 48 the main trunk of the nerve is formed by the union of three strands. Two of these, the mesial one and the lateral (figs. 40 and 41), follow a course closely parallel to two of the branches of the anterior cerebral artery, and give off delicate rami to them. These twigs enter the walls of the vessels in the same manner as has already been described in the OX. The relation of the medial of the three roots to another branch of the anterior cerebral artery is shown in figures 40 and 41 at a. Figure 40 represents the left hemisphere of the brain as viewed from the medial side. In figure 41 a portion of the same surface is represented on a larger scale, as seen through the binocular microscope. It will be noted that strand a turns so as to run nearly at right angles to the main nerve bundle and follows the NERVUS TERMINALIS: MAMMALS 53 course of the artery. The nerve strand soon divides into two smaller strands, one of which had been inadvertently cut in. separating the hemispheres. The severed portion may be seen on the stump of the artery, which it was ascertained belongs to the right hemisphere of the brain. a co.ant. nopt. Se: lamter. Fig. 40 Mesial view of left hemisphere of full-term mule fetus, showing the nervus terminalis. Slightly enlarged. The middle root of the left nerve, as also shown in figures 40 and 41, runs toward the perforated area in front of the lamina terminalis. At first it is roughly parallel with the anterior cere- bral artery, then turning upwards, it gives off a small ramus which enters the brain through a perforation in the furrow of the small suleus (posterior parolfactory) shown in the figure. Another ramus of slightly larger size is given off a little further caudad. This, and the caudally continuing main strand, also enter the 54 OLOF LARSELL co.ant. lam.ter. Fig. 41 Central connections of the left nervus terminalis of full-term mule fetus, illustrating relations to neighboring blood-vessels and points of entrance into brain substance. X ca. 5. Fig. 42 Central relations of right nervus terminalis of mule fetus. X ca. 9. NERVUS TERMINALIS : MAMMALS 55 brain substance, the one slightly rostrad to the sulcus previously mentioned, the other at a point about 2 mm. anterior to the lamina terminalis, and somewhat more than midway between the anterior commissure and the ventral border of the brain. As already stated, essentially the same conditions are found on the right side of the same specimen (fig. 42). Here, however, only two roots were found which entér the brain substance. The larger one (b) enters through the same opening as does one of the blood-vessels of medium size. This was the only case observed in the mule where a nerve strand of the terminalis and a blood- vessel enter the brain together. Other rootlets enter very near the openings through which other blood-vessels pass. The main nerve trunks on both sides are very clearly formed by the union of the roots described, with the possible addition of other smaller ones which may have been torn and lost during the dissection. The left nerve was removed from its attachments to make possible a closer study of its structure. As represented in figure 43, the three principal roots merge intothesmallerganglion. Ros- trally from this ganglion the nerve is compactly enclosed in a sheath of connective tissue. A few millimeters rostrad from the small ganglion is the larger one previously noted. Continuing forward from this point, the nerve remains as a compact bundle as far as the place where it had been broken in removing the brain from the cranial cavity. A very slender blood-vessel (fig. 48, 6/.v.) winds around the nerve for the greater part of its course. At intervals this vessel gives off branches which penetrate into the nerve bundle. Histological. Sections of the ganglia stained with thionin reveal a considerable variety in the form of the ganglion cells. The sections were cut 18 » and 20 u thick, so that the processes could be followed in many instances for some distance. As il- lustrated in figures 44 to 47, all variations of type from simple bipolar, to cells having at least five processes occur. Numerous binucleated cells (fig. 47) were observed. Of the 292 ganglion cells which were found in the single ganglion of the right nerve, twenty-seven were binucleated, or 9.25 per cent. Sections cut at 10 » and stained with haematoxylin showed the same types of cells in the two ganglia of the left nerve. rminalis of mule fetus at full term, removed from its Fig. 43 Left nervus te connections. Only the ganglia, with portions of the main trunk and of the cen- tral roots are represented. 56 NERVUS TERMINALIS: MAMMALS o7 Some attention was given to the arrangement of the chroma- tophile granules in the cells. Carpenter and Conel (’14) have pointed out the characteristic peripheral arrangement of this substance in ganglion cells of the sympathetic system, and hold this to be a distinguishing mark which differentiates them from cells of the cerebrospinal system. The granules are said to be scattered throughout the body of the perikaryon in ganglion cells of the central system. No very satisfactory results were obtained with the material available. Many of the cells (figs. 44 and 45) are seen to have a somewhat peripheral arrangement of the granules, while the other cells figured do not give suffi- cient indication of such distribution of granules to be noticeable. It should be stated, however, that the sections stained with thionin, which would affect the Nissl bodies, were too thick to be favorable for such a study. An effort was made to demonstrate myelinated fibers, if present, and to learn, if possible in the large central roots of the terminalis of the mule, the relative number of such fibers to the number in the main peripheral trunk. The osmie acid, iron- haematoxylin, Weigert, and Stroebe methods were each tried, with variations, a number of times. No success was had in demonstrating myelin sheaths. In some of the preparations delicate fibers of small diameter were visible, but it was not possible to trace them continuously through more than three or four sections of a series. In the central roots from two to four such sheath-like structures were present in some of the sections. In others none could be seen. The peripheral bundle showed as many as sixteen in certain sections. The results of this part of the study were on the whole negative. Either the myelin sheaths are not well developed in the nervus terminalis at this stage of fetal life of the mule or the formalin preserved material did not respond to the methods employed for their demonstration. The Horse. The nervus terminalis of the horse brain examined (fig. 48) consists centrally of four principal strands which unite near the base of the olfactory stalk to form a broad, flattened nerve trunk. This trunk continues rostrally as a compact bundle as far as the posterior part of the .bulbus olfactorius, receiving 58 OLOF LARSELL a number of delicate strands in this part of its course. Many of these strands were traced caudally to neighboring blood-vessels. The shrunken condition in which these vessels were found, how- ever, did not favor an attempt to find twigs, such as were ob- served in the beef, which pass into their muscular walls. At the region where the olfactory stalk swells to form the bulbus olfac- torius, the main trunk of the terminalis divides into five rela- tively large strands and a number of smaller ones. A few milli- meters caudally of this point a single strand is given off, which appears to pass laterally and beneath the olfactory stalk. The five larger strands continue rostrally, anastomosing with the finer strands and with one another, and thus form a plexus very similar Figs. 44-47 Ganglion cells from ganglion terminale of full-term mule fetus. Formalin fixation, thionin stain. > 550. to that observed in the eat (figs. 1 and 3) and in the-dog (fig. 49), About 3 mm. rostrad to the point where these strands assume a separate course, and lying in the path of the largest of the five, is a large flattened ganglion (fig. 48. gn.) of irregular outline. This ganglion receives fibers also from other strands of the plexus than the one in whose apparent course it is placed, so that it lies in the midst of the plexus. A much smaller ganglion (gn’,) is present caudally, on the main trunk midway between the larger ganglion and the point of union of the central roots. Rostrally of the main ganglion, two of the larger strands enter the sheath of connective tissue which encloses the vomeronasal NERVUS TERMINALIS: MAMMALS 59 nerve in its passage through the cribriform plate. One enters dorsally of the latter nerve trunk and the other on its ventral side. The vomeronasal nerve differs in the horse examined from the condition found in the other mammals studied in that it cerhem. ( ll Hi ( , yi Fig. 48 Ventrolateral view of forward portion of the brain and of part of the nasal septum of the horse, showing the relations of the nervus terminalis. The olfactory bulb was removed to bring into view the ganglion terminale and the intracranial plexus, from the angle at which the dissection was made. does not divide into the characteristic number of -bundles until it has passed through the bony plate as a single compact trunk, surrounded by a heavy sheath. A relatively large strand of the terminalis penetrates this sheath immediately distal to the point 60 _ OLOF LARSELL of emergence of the vomeronasal bundle from its foramen. This strand assumes an independent course, running dorsally a little way, then divides into a number of smaller strands whose gen- . eral course is rostrally. These become so attenuated by re- peated division that it was not possible to follow ‘their finer ramifications by dissection. Apparently they form a portion of a plexus similar to that found in the cat and in other mammals. aN cerhem. bu.olf. bu.olf ac. Fig. 49 Anterior portion of cerebral hemisphere and a portion of nasal septum of a puppy two weeks old (estimated) showing nervus terminalis. The course of the nerve strands in the septum was followed to some extent by the expedient of stripping off the thick mucosa from the bony septum and examining its deeper surface. This required a minimum of dissection, as the nerve strands lay for the most part in the deeper portion of the mucosa. The proxi- mal portions of the first two divisions of the vomeronasal nerve lay in rather deep furrows in the bony septum, and the strands of the terminalis which accompany the larger nerve through the cribriform plate, accordingly, emerge from the sheath very close to the periosteum of the bony septum. In addition to the rami NERVUS TERMINALIS: MAMMALS 61 of the bundle just noted, other strands were observed in the sep- tum which also appeared to form a part of the terminalis plexus. Returning to the other intracranial strands which were noted distally of the ganglion terminale, the most dorsal one was somewhat torn in dissection, but appeared to have passed be- tween the olfactory bulb and the cerebral hemisphere. The re- maining fibers appear to pass through the cribriform plate in company with bundles of olfactory fibres, ventral to the vomero- nasal foramen. Such fibers could not be found distally of the plate. Possibly the strands in the septum which were noted in connection with those which emerge in company with the ner- vus vomeronasalis are distal continuations of the strands now under discussion. The connections in the bony plate might easily have been inadvertently injured to such an extent that the strands immediately distal to the plate were completely destroyed. 4. The nervus terminalis in other mammals The observations of the nervus terminalis of the dog, squirrel, and human, and in embryos of the pig, sheep, and rabbit, were less extended than those described in the preceding pages. So far as they were carried, they agreed with the findings in the other animals studied. The nerve was found in the typical re- lation to the vomeronasal bundles, and Bases through the eribriform plate in company with these. Only one ganglion was found in the dog. This is situated near the vomeronasal nerve, and is long and fusiform (fig. 49). In the squirrel also (not figured) one large ganglion was present at the point where the main trunk of the terminalis splits into strands similar to those found in the eat and the horse, which accompany the larger bundles of the vomeronasal nerve through the bony plate. A number of smaller strands, which appeared to correspond with the intracranial plexus found in the other animals, were observed. It may be noted that in the squirrel the accessory olfactory bulb lies on the dorsolateral side of the bulbus olfactorius so as to be invisible when the brain is viewed from the medial side. 62 OLOF LARSELL As already stated, the attempt to find strands from the termi- nalis to the blood-vessels of the human brains was not suecess- ful. Nerve strands similar to those present on the anterior cerebral artery of the beef and the mule were found and small twigs were observed to enter the walls of the corresponding vessels of the human material. Efforts by the gold-chloride and Bielschowsky methods to demonstrate nerve terminations in these vessels were not satisfactory with the material available. The studies on the embryonic material did not serve to reveal any features not already known. | Ill. SUMMARY AND COMMENTS 1. The nervus terminalis of mammals is made up in part, at least, of sympathetic fibers, and its ganglionic clusters contain sympathetic cells. The wide distribution and large number of fibers of the peripheral plexus (as noted by Brookover), in com- parison with the small size of the central connections, resemble the relation of preganglionic fibers to the postganglionic fibers of the sympathetic system. This resemblance is strengthened by the occurrence of pericellular baskets on many of the ganglion cells of the intracranial clusters. 2. Two types of neurones are present in the terminalis, namely, 1) sensory and 2) motor. 3. Some of the sensory fibers end in the muscular walls of the anterior cerebral artery and its branches by a type of nerve ter- mination hitherto undescribed in cerebral blood-vessels. 4. There is some evidence that free nerve terminations in the epithelium of the septal mucosa and of Jacobson’s organ are also connected with afferent neurones of the nervus terminalis. For the present it is assumed that these free sensory terminations belong to a sensory component of the nervus terminalis which is distinct from sympathetic afferent fibers which have terminations in the walls of the blood-vessels. The type of nerve endings and their position in the mucosa would seem to indicate that this component is part of the general visceral afferent system. The early embryonic history of the nerve in ganoids and selachians might, however, point to a rela- NERVUS TERMINALIS: MAMMALS 63 .tionship with the special visceral afferent group. As previously noted, Brookover states that the origin of the nervus terminalis in the ganoid fishes studied by him is from a portion of the ol- factory placode. Locy, in describing its early development in Squalus, attributes its origin to the neural crest, but states also that “The new nerve has at first a fusion (placode) with the thickened surface epithelium, located just above the shallow de- pression that marks the beginning of the olfactory pit. This connection between the surface epithelium and brain-wall, con- sists of a group of closely packed cells in which I have failed at this early stage [6 to 8 mm.] to recognize fibers.’ If this pla- code described by Locy be homologous with that found by Brookover, the embryonic evidence in the two groups on which these writers worked points to an origin which in part at least corresponds to that of other special sensory ganglia in the head region. In connection with the free nerve terminations described, this embryonic origin of the nerve from a placode suggests the conclu- sion that there is a sensory component of the terminalis which is _ distinct from the sensory fibers which terminate in the walls of the blood-vessels. The neuroblasts which have their origin in the neural crest might well give rise to the sympathetic cells ' described and figured in the present article, and which have been described in other groups of vertebrates than the mammals in the nerve under consideration. The free nerve terminations in the mucosa do not, however, seem to fit in with the view that this sensory component belongs to the special visceral system. 5. Many physiologists find insufficient evidence of vasomotor control of the cerebral blood-vessels. Wiggers (’05, ’08) and Weber (’08) have presented experimental support of the histo- logical evidence. Both conclude that there is direct physiological proof of nerve control over the cerebral vessels. Weber, more- over, finds indications that there is an accessory vasomotor center further rostrad in the brain than that situated in the bulb. This observation is suggestive, in connection with McKibben’s findings in urodeles, of terminalis tracts extending to the inter- peduncular region, with a probable center in the neighborhood of the preoptic nucleus. 64 OLOF LARSELL 6. If connection of the nervus terminalis with such a center. in the forward part of the brain should be demonstrated, the forebrain should be included in the list of those divisions of the central nervous system which are directly related to the sympa- thetic system. 7. The evidence now at hand points to the conclusion that the nervus terminalis of mammals is functional. Its relatively small size in this group as compared with its development in selachians, may indicate that its functional importance is reduced. 8. Innervation by the nervus terminalis of the vomeronasal organ, which appears to occur to some extent, is merely incidental. IV. BIBLIOGRAPHY ALEXANDER, W.T. 1875 Bemerkungen iiber die Nervender Dura Mater. Archiv fiir Mikros. Anat., Bd. 11, 8S. 281. Aus, E. P. 1897 The cranial muscles and the cranial and first spinal nerves in Amia calva. Jour. Morph., vol. 12, no. 3. Batoau, C. 1860 Das Jacobson’sche Organ des Schafes. Sitzungsb. d. Kais. Akad. d. Wissenschaften in Wien, Bd. 42, 8. 449. BaRDEEN, C. R. 1903 The growth and histogenesis of the cerebrospinal nerves inmammals. Am. Jour. Anat., vol. 2, pp. 231-257. BawvEN, H. H. 1894 The nose and Jdcobson’s organ with especial reference to . Amphibia. Jour. Comp. Neur., vol. 4, pp. 117-152. Beprorp, E. A. 1904 The early history of the olfactory nerve inswine. Jour. Comp. Neur., vol. 14, pp. 390-410, BeLocotowy, G. 1912 Studien zur Morphologie des Nervensystems der Wir- belthiere. I. Die Entwickelung des Nervus terminalis bei Selachiern. Bull. Soc. Nat. Moscou, vol. 25. Bina, R., AND BurcKHARDT, R. 1904 Das Centralnervensystem von Ceratodus forsteri. Anat. Anz., Bd. 25, S. 588-599. 1905 Das Centralnervensystem von Ceratodus forsteri. Semon Zoél. Forsch., in Jenaische Denkschr., Bd. 4. BrRooKxoverR, Cuas. 1908 Pinkus’s nerve in Amia and Lepidosteus. Science, N.S., vol. 27, p. 918. 1910 The olfactory nerve, nervus terminalis and preoptic sympathetic system in Amia calva, L. Jour. Comp. Neur., vol. 20, pp. 49-118. 1914a The development of the olfactory nerve and its associated ganglion in Lepidosteus. Jour. Comp. Neur., vol. 24, pp. 113-130. 1914b The nervus terminalis in adult man. Jour. Comp. Neur., vol. 24, pp. 131-135. 1917 The peripheral distribution of the nervus terminalis in an* infant. Jour. Comp. Neur., vol. 28, pp. 349-360. BRookover, Cuas., AND Jackson, T. 8S. 1911 The olfactory nerve and the ner- vus terminalis of Ameiurus. Jour. Comp. Neur., vol. 21, p. 237. NERVUS TERMINALIS: MAMMALS 65 Casa, 8S. R. 1893 Neue Darstellung vom histologischen Bau des Centralner- vensystems. Arch. fiir Anat. u. Phys., Ant. Abtheil., p. 393. 1905 a Tipos celulares de los ganglios sensitivos del hombre y mami- feros. Trab. del Lab. de Inves. Biol. de la Univ. de Madrid, T. 4, p. 1. 1905 b Las celulas del gran simpatico del hombre adulto. Trab. del Lab. de Inves. Biol. de la Univ. de Madrid, T. 4, p. 79. 1911 Histologie du Systéme Nerveux de L’Homme, T. 2, p. 670. CARPENTER, F. W. 1911 The ciliary ganglion of birds. Folia Neurobiologica, Bd. 5, 8. 738-754. 1912 On the histology of the cranial autonomic ganglia of the sheep. Jour. Comp. Neur., vol. 22, p. 447. CarPENTER, F. W., AND Cone, J. L. 1914 A study of ganglion cells in the sym- pathetic system, with special reference to intrinsic sensory neurones. Jour. Comp. Neur., vol. 24, p. 269. CarPEnTER, F. W., ann Main, R. C. 1907 The migration of medullary cells into the ventral nerve-roots of pig embryos. Anat. Anz., Bd. 31, S. 303-306. : Cote, Franx J. 1896 The cranial nerves of Chimaera monstrosa. Proc. Roy. Soc. Edinburgh, vol. 21, pp. 49-56. Devries, E. 1905 Note on the ganglion vomeronasale. K. Akad. van Weten- schappen te Amsterdam, vol. 7, p. 704. Doaiet, A. S. 1895 Zur Frage iiber den feineren Bau des sympathischen Ner- vensystems bei den Sdugethieren. Arch. fiir Mikros. Anat. und Entwick., Bd. 46, 8S. 305-344. 1896 Zwei Arten sympathischer Nervenzellen. Anat. Anz., Bd. 11, S. 679-687. 1898 Die sensiblen Nervenendigungen im Herzen und in den Blutge- fissen der Siugethiere. Arch. fiir Mikros. Anat. und Entwick., Bd. 52, S. 44-70. 1908 Der Bau der Spinalganglien des Menschen und der Siugethiere. Jena. DoéuixeNn, A. 1909 Ursprung und Zentren des Nervus terminalis. Monatsschr. f. Psych. u. Neur., Erg. Heft, Bd. 26, S. 10. Fritscu, G. 1878' Untersuchungen iiber den feineren Bau des Fischgehirns. Berlin. Firprincer, K. 1904 Notiz iiber einige Beobachtungen am Dipnoerkopf. Anat. Anz., Bd. 24, S. 405-408. Herrick, C. Jupson 1903 On the morphological and physiological classification of the cutaneous sense organs of fishes. Amer. Nat., vol. 37, pp. 313- 318. 1909 The nervus terminalis (nerve of Pinkus) in the frog. Jour. Comp. Neur., vol. 19, p. 175. 1916 Introduction to neurology. Philadelphia. 1917 The internal structure of the midbrain and thalamus of Nectu- . ‘rus. Jour. Comp. Neur., vol. 28, p. 236. Herrick, C. L. 1893 Topography and histology of the brain of certain reptiles. Jour. Comp. Neur., vol. 3, p. 124. Howe, W. H. 1915 Text-book of physiology, 6th ed., p. 632. Philadelphia. THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, No. 1 66 OLOF LARSELL Huser, G. Carut 1897 Lectures on the sympathetic nervous system. Jour. Comp. Neur., vol. 7, pp. 73-145. 1899 a Observations on the innervation of the intracranial vessels. Jour. Comp. Neur., vol. 9, pp. 1-25. 1899 b A contribution on the minute anatomy of the sympathetic ganglia of the different classes of vertebrates. Jour. Morph., vol. 16, pp. 27-90. 1913 The morphology of the sympathetic nervous system. XVII International Congress of Medicine, London. Huser, G. C., anp Dewitt, Lyp1a_ 1898 A contribution on the motor nerve- endings and on the nerve-endings in the muscle spindles. Jour. Comp. Neur., vol. 7, pp. 169-230. Huser, G. C., anp Guitp, Stacy R. 1913 Observations on the peripheral dis- tribution of the nervus terminalis in Mammalia. Anat. Rec., vol. 7, pp. 253-272. Jones, W. C. 1905 Notes on the development of the sympathetic nervous sys- tem in the common toad. Jour. Comp. Neur., vol. 15, p. 113. JounstTon, J. B. 1906 The nervous system of vertebrates. Philadelphia. 1909 The morphology of the forebrain vesicle in vertebrates. Jour. Comp. Neur., vol. 19, pp. 457-539. 1911 The telencephalon of selachians. Jour. Comp. Neur., vol. 21, pp. 1-118. 1913 Nervus terminalis in reptilesandmammals. Jour. Comp. Neur., vol. 23, pp. 97-120. 1914 The nervus terminalis in man and mammals. Anat. Rec., vol. 8. KeEIBEL AND Matt 1912 Human embryology, vol. 2. Philadelphia. KO6O.uikEeR, A. 1896 Handbuch der Gewebelehre des Menschen, Bd. 2, S. 835. Kuntz, ALBERT 1911 The evolution of the sympathetic nervous system in ver- tebrates. Jour. Comp. Neur., vol. 21, pp. 215-236. 1913 a The development of the cranial sympathetic ganglia in the pig. Jour. Comp. Neur., vol. 23, pp. 71-96. 1913 b On the innervation of the digestive tube. Jour. Comp. Neur., ‘vol. 23, pp. 173-192. 1914 Further studies on the developmert of the cranial sympathetic ganglia. Jour. Comp. Neur., vol. 24, p. 235. Lanpacrg, F. L. 1910 The origin of the sensory components of the cranial ganglia. Anat. Rec., vol. 4, pp. 71-79. 1916 The cerebral ganglia and early nerves of Squalus acanthias. Jour. Comp. Neur., vol. 27, p. 19. Laneuey, J. N. 1900 The sympathetic and other related systems of nerves. Schifer’s Text-book of Physiology, vol. 2, pp. 616-696. 1903 The autonomic nervous system. Brain, vol. 26, pp. 1-26. Locy, W. A. 1899 New facts regarding the development of the olfactory nerve. Anat. Anz., Bd. 16, S. 273-290. 1903 A new cranial nerve in selachians. Mark Anniv. Volume. 1905 a On a newly recognized nerve connected with the forebrain of selachians. Anat. Anz., Bd. 26, 8. 33-63 and 111-123. 1905 b A footnote to the ancestral history of the vertebrate brain. Science, N. S., vol. 27, pp. 180-183. NERVUS TERMINALIS: MAMMALS 67 Matong, E. T. 1913 Nucleus cardiacus nervi vagi and the three different types of nerve cells which innervate the three different types of muscle. Am. Jour. Anat., vol. 15, pp. 121-127. McCorrer, R. E. 1912 The connection of the vomeronasal nerves with the accessory olfactory bulb in the opossum and other mammals. Anat. Rece., vol. 6, p. 299. 1913 The nervus terminalis in the adult dog and cat. Jour. Comp. Neur., vol. 23, p. 145. 1915 A note on the course and distribution of the nervus terminalis inman. Anat. Rec., vol. 9, p. 243. 1917 The vomeronasal apparatus in turtle and frog. Anat. Rec., vol. 12. McKissen, P. S. 1911 The nervus terminalis in urodele Amphibia. Jour. Comp. Neur., vol. 21, pp. 261-309. 1914 Ganglion cells of the nervus terminalis in the dogfish (Mustelus canis). Jour. Comp. Neur., vol. 24, pp. 437. Micuaitow, 8S. 1908 Die Nerven des Endocardiums. Anat. Anz., Bd. 32. 1910 Uber die sensiblen Nervenendapparate der zentralen sympath- ischen Ganglien der Siugethiere. Jour. f..Psych. u. Neur., Bd. 16, 8. 269. 1911 Der Bau der zentralen sympathischen Ganglien. Internat. Monatsschr. f. Anat. u. Phys., Bd. 28, S. 26. Miuumr, L. R., unp Dani, W. 1910 Die Beteiligung des sympathischen Ner- vensystems an der Kopfinnervation. Deutsches Archiv f. klin. Med., Bd. 99, S. 48-107. : Norris, H. W. 1913 The cranial nerves of Siren lacertina. Jour. Morph., vol. 24, p. 254. Pinxus, Fevrx 1895 Die Hirnnerven des Protopterus annectens. Morph. Arb. (G. Schwalbe), Bd. 4, 8. 275. 1905 Uber den zwischen Olfactorius und Opticusursprung des Vor- derhirn (Zwischenhirn) verlassenden Hirnnerven der Dipnoer und Selachier. Arch. Physiol. Jahrb., sup. Heft. 2, 8. 447. Prentiss, C. W. 1915 Text-book of embryology. Philadelphia. Ranson, 8. W. 1912 The structure of the spinal ganglia and of the spinal nerves. Jour. Comp. Neur., vol. 22, pp. 159-175. 1915 The vagus nerve of the snapping turtle (Chelydra serpentina). Jour. Comp. Neur., vol. 25, p. 301. Ranson, 8. W., anp Binutinastey, P.R. 1918 The superior cervical ganglion and the cervical portion of the sympathetic trunk. Jour. Comp. Neur., vol. 29, no. 4. Reap, Erriz A. 1908 A contribution to the knowledge of the olfactory appara- tus in the dog, cat and man. Am. Jour. Anat., vol. 7, pp. 17-47. Ruspascuin, W. 1903 Uber die Beziehungen des Nervus trigeminus zur Riech- schleimhaut. Anat. Anz., Bd. 22, 8. 407. SHEtpon,-R. E. 1908 The participation of medullated fibers in the innervation of the olfactory mucous membrane of fishes. Science, N. 8., vol. 27, no. 702, p. 915. 68 OLOF LARSELL SHEeLpon, R. E. 1909 The nervus terminalis in the carp. Jour. Comp. Neur. vol. 19, p. 191. 1912 The olfactory tracts and centers in teleosts. Jour. Comp. Neur., vol. 22, pp. 177-389. SHELDON, R. E., AND Brookover, CuHas. 1909 The nervus terminalis in tele- osts. Anat. Rec., vol. 3, p. 257. SHERRINGTON, C. 8S. 1911 The integrative action of the nervous system. New Haven. Sewertzorr, A. N. 1902 Zur Entwickelungsgeschichte des Ceratodus forsteri. Anat. Anz., Bd. 21, p. 606. Smirnow, A. 1895 Uber die sensiblen Nervenendigungen im Herzen bei Am- phibien und Saugethieren. Anat. Anz., Bd. 10, 8. 737-749. Smitu, G. Exitior 1895 Jacobson’s organ and the olfactory bulb in Ornitho- rhynchus. Anat. Anz., Bd. 11, S. 161. STREETER, G. L. 1912 The development of the nervous system. 4. The sym- pathetic nervous system. Human Embryology, Keibel and Mall, vol. 2, pp. 144-154. von Brunn, A. 1892 Die Endigung der Olfactoriusfasern in Jacobson’schen Organe des Schafes. Arch. f. Mikros. Anat. u. Entwick., Bd. 39, S. 651. von Lenuoss&éK, M. 1892 Die Nervenurspriinge und Endigungen im Jacob- son’schen Organ des Kaninchens. Anat. Anz., Bd. 7, 8. 628. , 1911 Das Ganglion ciliare der Végel. Arch. f. Mikros. Anat. u. Entwick., Bd. 76, S. 745. Weser, Ernst 1908 Uber die Selbstindigkeit des Gehirns in der Regulierung seiner Blutversorgung. Arch. f. Phys. u. Anat., Phys. Abteil., S. 457. Wiaasmrs, C. J. 1905 Action of adrenalin on the cerebral vessels. Amer. Jour. Physiol., vol. 14, p. 452. 1907 The innervation of the cerbral vessels as indicated by the action of drugs. Amer. Jour. Physiol., vol. 20, p. 206. 1908 Some vasomotor changes in the cerebral vessels obtained by stimulating the carotid plexuses. Amer. Jour. Physiol., vol. 21, p. 454. Resumido por el autor, Edward Phelps Allis Jr. Los nervios oftalmicos de los peces gnatostomos. En Polypterus sale del crdneo, con las fibras del lateral que van a los nervios oftalmico superficial y bucal lateral, un fascf- culo de fibras del comiin, formandose un ganglio con estos dos fasciculos, colocado en posicién dorsal con relacién al ganglio cutaneo general del trigémino. De este ganglio comun-lateral parten fibras del comin, las cuales van a parar al ganglio cutdneo general, que no envia fibra alguna al primero. El oftdlmico superficial se origina en su totalidad en el ganglio lateral comtin y contiene fibras del lateral y comtin, pero en pequefo ntimero, careciendo probablemente de fibras del cutdneo general. El llamado trigémino-oftaélmico profundo es en realidad un ramo oftalmico profundo puesto que nace de un ganglio profundo e independiente, formado por una raiz profunda independiente; aparentemente contiene solo fibras cutdneas generales. Posee ramas frontales y nasales que son homdlogas de las mismas ramas del nervio oftalmico de‘’los vertebrados mds superiores. Kste ultimo nervio es también por esta causa un nervio profundo y no una parte del trigémino. En los Dipnoos la distribuci6én es, en apariencia, las misma que en Polypterus. En los Holé- steos hay un ramo oftdlmico superficial que contiene fibras del lateral, comun y cutdneo general y una porcién oftalmica pro- funda, pero el ramo oftalmico profundo, cuando existe, esta de- generado. En ciertos Teleésteos (Gasterosteus) existe una dis- tribucién semejante a la mencionada en los Holdsteos, pero en la mayor parte de ellos hay un ramo oftdlmico superficial de com- posicion variable, faltando el ramo oftdlmico y la porcién oftalmica profunda. Transiation by Dr. José Nonidez, Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 9 THE OPHTHALMIC NERVES OF THE GNATHOSTOME FISHES. EDWARD PHELPS ALLIS, JR. Menton, France There has long been, and still is, confusion in the terms em- ployed to designate the ophthalmic nerves of the gnathostome fishes. ‘These nerves were formerly considered to be the equiva- lents of the ramus ophthalmicus trigemini of higher vertebrates (Stannius, ’49), and as in many fishes, one of them lies super- ficial to the other they were called the rami superficialis and profundus trigemini. The term profundus was, however, fre- quently given to two distinctly different nerves, one of which runs forward between the superior and inferior divisions of the nervus oculomotorius and then ventral to the nervus trochlearis, while the other runs forward dorsal to both those nerves. The former nerve is alone properly called a profundus and it is typ- ically found in the Elasmobranchii. A nerve that, in certain fishes, connects the basal portion of this profundus nerve with that of the superficialis was called the portio ophthalmici profundi. Later, it was found that those nerve fibers of the superficialis nerve that innervate the organs of the supraorbital laterosensory line all issue from the medulla as an apparent part of the root of the nervus facialis, and as this was considered to be incontro- vertible evidence that they belonged to the latter nerve, that part of the ophthalmicus superficialis that was formed by them was called the ramus ophthalmicus superficialis facialis. The re- maining fibers of the superficialis nerve were still considered to belong to the trigeminus, and were usually called the ramus oph- thalmicus superficialis trigemini, but as, in the Holostei and Teleostei, they lie deeper than the lateralis fibers, they were fre- quently called the ramus ophthalmicus profundus trigemini, and 69 70 EDWARD PHELPS ALLIS, JR. apparently considered to be the homologue of the similarly named nerve of the Elasmobranchii. This latter nerve of the Elasmobranchii was still called the ramus ophthalmicus profun- dus trigemini, but there was a growing opinion that it belonged to a cranial segment next anterior to that of the trigeminus. It was then still later found that communis fibers might also form part of the ophthalmicus superficialis, and that, like the lat- eralis fibers of that nerve, they issued from the medulla as an apparent part of the root of the nervus facialis. Consistency then evidently demanded that these communis fibers also be included in the ophthalmicus superficialis facialis, but as that term had come to mean a purely lateralis nerve, the communis fibers were either still relegated to the ophthalmicus super- ficialis trigemini or a new term, truncus supraorbitalis, was given to the entire ophthalmic nerve, the term ophthalmicus superficialis facialis still being employed to designate the later- alis fibers only of the nerve. The ramus ophthalmicus superficialis of fishes thus came to be considered to be a nerve formed by the secondary juxtaposition of fibers derived from two adjacent segmental nerves, the gen- eral cutaneous fibers of the nerve being derived from the nervus trigeminus and the lateralis and communis fibers from the ner- vus facialis. This schema of the nerve still, however, left un- accounted for those fibers that were known to be derived, in cer- tain fishes, from the portio ophthalmici profundi, and although this portio, as an independent nerve, had only been described in a few fishes, there was no apparent reason for assuming that it did not also exist in many, if not all, of those fishes in which the profundus and trigeminus ganglia have completely fused with each other. This confusion of terms and complication of conditions led me, in my work on the mail-cheeked fishes (Allis, ’09) to readopt the term, ophthalmicus superficialis trigemini, first given to this nerve, and to call its lateralis and communis fibers the lateralis and communis trigemini. A name of some sort had to be given to the nerve, and this one seemed to me to be the “‘single name already current” that Herrick (’09) has later suggested should be OPHTHALMIC NERVES OF GNATHOSTOME FISHES 71 selected for each so-called composite nerve, and it certainly had valid claim to priority over the one that he had himself earlier employed, namely, truncus supraorbitalis. Furthermore, it was at that time, and still is, my opinion that it has by no means as yet been definitely established that the lateralis and com- munis fibers found in the ophthalmicus superficialis do not be- long definitely to the trigeminus, their centers of origin having simply fused with, or become contiguous to, those of similar fibers that belong to the nervus facialis. The conditions in Polypterus, described immediately below, certainly favor this conclusion, but they at the same time further complicate the choice of a proper name for the nerve. In Polypterus there is a profundus ganglion which, in a 75-mm. specimen examined in serial transverse sections, is extracranial in position and lies wholly anterior to, and independent of, the trigeminus ganglion. The root of this ganglion traverses a fora- men that lies anterior to the foramina for the roots of the nervus trigeminus, enters the medulla slightly anterior to the general cutaneous root of the latter nerve, and is, so far as can be deter- mined from my somewhat imperfect and unsatisfactory sections, composed exclusively of general cutaneous fibers. From this profundus ganglion a typical ramus ophthalmicus profundus arises, and also either a single nerve, which immediately bifur- cates, or two independent nerves, these latter one or two nerves forming the portio ophthalmici profundi shown by van Wijhe (82) in his figure of this fish. The branches of this ramus and portio all join the ramus ophthalmicus superficialis, and, accom- panying it and its branches, but in no way fusing with them, are distributed mainly to tissues on the dorsal surface of the anterior portion of the head. The ramus ophthalmicus superficialis arises from a ganglion formed on two bundles of fibers, one of which contains all the lateralis fibers that go to the rami ophthalmicus and buccalis lateralis, while the other is an intracranial branch of the communis root of the nervus facialis. These two roots of this ganglion, which thus quite certainly contain no general cutancous fibers, issue together from the cavum cerebrale cranii, and the ganglion formed on them lies dorsal to the ganglion 72 EDWARD PHELPS ALLIS, JR. formed on the general cutaneous root of the trigeminus. Com- munis fibers are sent from this lateralis-communis ganglion to the general cutaneous one, but no fibers can be traced from the latter to the former ganglion. The ramus ophthalmicus super- ficialis, which arises wholly from the lateralis communis ganglion and contains both lateralis and communis fibers, thus certainly contains but few, and probably no general cutaneous ones. The ophthalmicus profundus must then supply most, and prob- ably all, of the general cutaneous fibers that go to that part of the dorsal surface of the head that is supplied, in most teleosts, exclusively by branches of the so-called ophthalmicus super- — ficialis trigemini, and the latter nerve is wholly wanting in Polypterus. The ophthalmic nerves of Polypterus thus are: a nerve that I should call the ramus ophthalmicus superficialis trigemini, but which, according to currently accepted views, would be called the ramus ophthalmicus superficialis facialis; and a ramus ophthal- micus profundi, which has ramuli frontalis and nasalis that are, respectively, the portio ophthalmici profundi and ramus ophthal- micus profundus trigemini of current descriptions. If the oph- thalmicus superficialis were to abort, as the fibers that form it always do in all land vertebrates, there would remain a nerve that would quite unquestionably be the homologue of the oph- thalmic nerve of Hoffmann’s (’86) descriptions of embryos of rep- tiles, for the radix longa of my 75-mm. Polypterus, which arises from the profundus ganglion, is the ramus ciliaris of Hoffmann’s descriptions of reptiles, which latter nerve fuses, for a certain distance, with the ramus nasalis to form the ramus nasociliaris. The opinion long ago expressed by His (’87, p. 398), that the ramus nasalis, or nasociliaris, of mammals corresponds to the so-called ramus ophthalmicus profundus trigemini of selachians is thus confirmed, as is also my conclusion, in my work on Mus- telus (Allis, ’01, p. 299), that the ramus ophthalmicus profun- dus of Potypterus and the portio ophthalmici profundi of gan- oids are, respectively, the homologues of the nasal and frontal branches of the ophthalmic nerves of higher vertebrates. There would then be no so-called ramus ophthalmicus superficialis tri- OPHTHALMIC NERVES OF GNATHOSTOME FISHES 73 gemini in these latter vertebrates, which is in accord with Pinkus’s (94) conclusion that the presence of this nerve in the Amphibia is very doubtful, and with Norris’s statement (’13, p. 292) that, in Siren lacertina: ‘“‘It is questionable whether any general cu- taneous fiber should be considered as a constituent part of the dorsal, or supraorbital division [of the truncus supraorbitalis] of Siren.”’ The ramus ophthalmicus profundi of Polypterus thus quite certainly being the homologue of the so-called ophthalmicus trigemini of man, the introduction of a proper and uniform ter- minology becomes a somewhat radical proceeding, for it evi- dently requires a renaming of the nerve in man. The conditions in those fishes, other than Polypterus, in which either a ramus ophthalmicus profundus trigemini, or a portio ophthalmici profundi, has been described, may now be consid- ered, and I have, furthermore, examined the conditions in Gas- terosteus, Cottus, and Clinocottus, in. which fishes there is an anterior portion of the ascending process of the parasphenoid that occupies the position of the pedicel of the alisphenoid of Amia. ‘The reason for examining these latter fishes was the con- viction that, if there were a portio ophthalmici profundi, it would lie anterior to the above-mentioned anterior process of the parasphenoid, for that is the relation that the radix profundi " of these fishes has to that process; and, conversely, if no branch of the profundus, sent to the superficialis, were found in that posi- tion, it would be, in my opinion, conclusive evidence that the portio ophthalmici profundi was wanting in these fishes, and hence presumptive evidence that it was also wanting in those of the Teleostei in which the process is not found. In Protopterus, Pinkus (’94) describes a ramus ophthalmicus profundus trigemini, but no portio ophthalmici profundi. The first three branches of the ophthalmicus profundus are small, and, running forward dorsal to the nervi oculomotorius and troch- learis, become associated with a ramus ophthalmicus superficialis facialis. The ophthalmicus profundus then separates into two nearly equal portions, one of which is shown, in the figure given, running forward dorsal to both divisions of the nervus oculomo- torius, and the other ventral to them, the dorsal branch then 74 EDWARD PHELPS ALLIS, JR. apparently passing ventral to the nervus trochlearis and joining the first three small branches of the nerve. The relations of this profundus nerve to the oculomotorius thus differ radically from those in Polypterus, but it seems probable that its first three branches correspond to the frontal branch of the nerve of the latter fish, and the two larger ones to the nasal branch. A small nerve is said to arise from that part of the trigeminus ganglion from which the ophthalmicus profundus has its origin, and to immediately join the ophthalmicus superficialis facialis, and Pinkus doubtfully calls it the ramus ophthalmicus superficialis trigemini. Comparison with Polypterus would, however, indi- cate that it is a persisting remnant of the communis component only of the ramus ophthalmicus superficialis of the latter fish. In Ceratodus there is a bar of cartilage that represents the pedicel of the alisphenoid (Allis, 714), and the ophthalmicus pro- fundus of Greil’s (13) descriptions of embryos of this fish issues from the cranium anterior to that bar. The nerve is then shown, in one of Greil’s figures (l.c., fig. 8, pl. 55), separating into two branches, one of which is evidently a portio ophthalmici profundi and the other a typical so-called ophthalmicus profundus tri- gemini. The ophthalmicus superficialis of this figure, called by Greil the ophthalmicus superficialis trigemini in another figure (fig. 4) on the same plate, receives no branch from what is ap- parently the general cutaneous’ ganglion of the trigeminus, the profundus thus supplying, as in Polypterus, all the general cu- taneous fibers sent to the dorsal surface of the anterior portion of the head. | In Amia I have fully described the nerves here concerned, without, however, definitely determining their components (Allis, 97). It is, however, probable that the ophthalmicus superficialis contains lateralis, communis and general cutaneous fibers, and it receives an important portio ophthalmici profundi from an independent profundus ganglion, the root of the latter ganglion issuing from the cranium anterior to the pedicel of the alisphenoid. A delicate nerve that arises from the anterior end of the profun- dus ganglion was considered by me to be a greatly degenerated ramus ophthalmicus profundus. OPHTHALMIC NERVES OF GNATHOSTOME FISHES iD In Lepidosteus, van Wijhe (’82) describes both a ramus oph- thalmicus profundus trigemini and a ramus ophthalmicus super- ficialis trigemini. In a 75-mm. specimen of this fish I find the so-called ophthalmicus superficialis trigemini composed of later- alis, communis and general cutaneous fibers, the lateralis fibers forming a bundle which lies somewhat above the communis and general cutaneous ones. A radix profundi arises from the me- dulla anterior, and close to the general cutaneous root of the tri- geminus, and issues from the cranial cavity by an independent foramen. A profundus ganglion forms on this root, and from it a radix longa, a ramus ciliaris longa, and a portio ophthalmici profundi arise. There is no perceptible trace of a ramus oph- thalmicus profundus. The portio ophthalmici profundi runs upward and joins and fuses with the communis and general cu- taneous components of the ophthalmicus superficialis, as shown but not index-lettered in Luther’s figure of this fish (13, fig. 1), but as there is no pedicel to the alisphenoid of this fish (Allis, 09) the relations of the nerve to that element of the skull are, as in Polyp- terus, undefined. The conditions in this embryo thus show that the so-called nervus ophthalmicus profundus of Landacre’s (712) descriptions of a 10-mm. embryo of this fish is probably a portio ophthalmici profundi and not a ramus profundus; and this nerve accordingly cannot be, as Landacre concludes in a later work (16, p. 27), a nerve comparable to the ramus ophthalmicus pro- fundus of the Selachii. In Acipenser: van Wijhe (’82) describes a so-called ophthalmi- cus profundus trigemini, but as this nerve is said to run forward dorsal to all the muscles of the eyeball, and is shown in his figure lying dorsal to the nervus trochlearis, it must be either a portio ophthalmici profundi, a ramus ophthalmicus superficialis tri- gemini, or both those nerves combined. Which one it is cannot be told either from his descriptions or from those of Gorono- witsch (’83), for the profundus ganglion is completely fused with the trigeminus ganglion. In a 40-mm. specimen of Gasterosteus acus, in which fish there is an anterior portion of the ascending process of the parasphenoid that replaces the pedicel of the alisphenoid (Allis, in press), the 76 EDWARD PHELPS ALLIS, JR. ophthalmic and maxillomandibular branches of the trigeminus issue from the cranium posterior to that process, but two small - branches that arise from the intracranial portion of the trigemi- nus ganglion run forward in the cranial cavity and issue from it anterior to the process. One of these two branches is the radix longa. The other separates into two parts, one of which is the ramus ciliaris longa and the other a portio ophthalmici profundi. In a 37-mm. specimen of Cottus aspera, and in a 40-mm. speci- men of Clinocottus, in both of which fishes there is also a process of the parasphenoid that replaces the pedicel of the alisphenoid (Allis, ’09), there is a radix profundi which issues from the cra- nium anterior to that process and separates into a radix longa and ramus ciliaris longa, but there is neither portio ophthalmici profundi nor ramus ophthalmicus profundus. In each of these three fishes the profundus ganglion is so completely fused with the trigeminus ganglion that, while its general outlines can be readily determined, it cannot be told whether or not there is any exchange of fibers between the two ganglia. The conditions nevertheless show that where there is both a portio ophthalmici profundi and a process of the parasphenoid that corresponds to the pedicel of the alisphenoid, the portio issues from the cranium antericr to that process. Where there is neither pedicel of the alisphenoid nor corresponding process of the parasphenoid, as in most of the Teleostei, it might be assumed, as already stated, that the portio still existed, but the conditions in the three fishes above described, and those in Scomber (Allis, ’03), where there is an independent profundus ganglion, but neither portio pro- fundi nor ramus ophthalmicus profundus, tend to show that both these nerves have wholly disappeared in most of the Teleostei. In a 22-mm. embryo of Squalus acanthias, Landacre (’16) finds both a so-called ramus ophthalmicus profundus and a general cutaneous nerve which he considers to be the ramus ophthalmi- cus superficialis trigemini, but the manner of origin of this latter nerve from the trigeminus ganglion is not incompatible with its being a portio ophthalmici profundi. There is, how- ever, an anterior prolongation of the profundus ganglion that OPHTHALMIC NERVES OF GNATHOSTOME FISHES 77 strongly suggests this portio. Of this process Landacre says that it “is evidently the remains of the structure which Neal identifies as a persistent connection of the ganglion with the ec- toderm, and which Scammon identifies as the utrochlea process, i.e., the remains of the connection of this ganglion with the neural crest.”’ The conditions in the several fishes above considered thus show that there are two distinctly different nerves that may supply the general cutaneous fibers that are distributed to the dorsal surface of the anterior portion of the head. One of these nerves is the so-called ramus ophthalmicus superficialis trigemini, the other what I have called the ramus ophthalmicus profundi, the frontal branch of the latter nerve being the So-called portio ophthalmici profundi. The trigeminus one of these two nerves always issues from the cranium posterior to the pedicel of the alisphenoid, or posterior to a corresponding process of the para- sphenoid, while the profundus always issues anterior to that pedicel or process, and I consider these peripheral relations to these structural elements to be as definite and positive evidence of the segments to which the nerves belong as are the facts of development and the central origins of the nerves. The relative importance of these two ophthalmic nerves varies greatly in different fishes, as does also the relative importance of the frontal and nasal branches of the ramus ophthalmicus pro- fundi, and it would seem as if the frontal branch alone ‘of the latter nerve might be the serial homologue of the entire ophthal- micus trigemini. In the Selachii, where there is apparently both a ramus ophthalmicus profundi and. a ramus ophthalmicus trigemini, the portio ophthalmici profundi is wholly wanting, unless it be represented in some part of the so-called ophthalmicus trigemini. In the Holostei and certain of the Teleostei there is an ophthalmicus trigemini and a portio profundi, but the ramus ophthalmicus profundi, if*present at all, is a small and degenerate nerve (Amia). In certain others of the Teleostei (Scomber) there is an ophthalmicus trigemini, but neither ramus ophthalmicus profundi nor portio profundi. In Polypterus, and probably also in Ceratodus, there is a ramus ophthalmicus pro- 78 EDWARD PHELPS ALLIS, JR. fundi, with ramuli nasalis and frontalis, but no evident trace of an ophthalmicus trigemini, excepting as it is represented in the so-called ophthalmicus superficialis facialis; and this, with the further absence of the ophthalmicus superficialis facialis, is apparently the condition found in all higher vertebrates. Herrick (’09) says that’ the facialis nerve of primitive verte- brates was a branchiomeric nerve, supplying a gill-bearing seg- ment and containing at least four components. It is not said that one of these components was a general cutaneous one, but that seems understood. In the mandibular arch similar condi- tions must certainly have existed, and probably also in a pre- mandibular arch supplied by the nervus profundus. In any event, it is certain that the general cutaneous tissues of the region innervated in Polypterus, Ceratodus, and the Selachii by the nervus profundus are innervated, in the Holostei and Teleostei, either by both the portio profundi and the ophthalmicus tri- gemini, or by the latter nerve alone. There must then have been, phylogenetically, either an actual change of innervation of these tissues in one of these two groups of fishes or the tissues innervated, in one of the groups, by one segmental nerve, must have degenerated, with the related nerve, in the other group, and there have been replaced by tissues primarily innervated by the nerve of an adjacent segment. The presence, in Amia, of a degenerate ramus ophthalmicus profundi would seem to exclude the possibility of assuming that the remaining fibers of the latter nerve have simply, in the Holostei and Teleostei, ac- quired a different and more favorable course than that followed by them in the Selachii. But however this may have been, it is evident that, of the several fishes above considered, Polypterus, the Dipneusti, and the Elasmobranchii alone present actual conditions of these nerves that could have led to those found in higher vertebrates, for that a nerve once so degenerated as the ramus ophthalmicus profundi of the Holostei and Teleostei would have been redeveloped and perpetuated seems improbable. Palais de Carnolés, Menton, France August 26, 1918. OPHTHALMIC NERVES OF GNATHOSTOME FISHES 79 LITERATURE CITED Auuts, E. P., Jr. 1897 The cranial muscles, and cranial and first spinal nerves in Amia calva. Jour. Morph., vol. 12, Boston. 1901 The lateral sensory canals, the eye-muscles, and the peripheral distribution of certain of the cranial nerves of Mustelus laevis. Quart. Journ. Microse. Sci., vol. 45, London. 1903. The skull, and the cranial and first spinal muscles and nerves in Secomber scomber. Jour. Morph., vol. 18. 1909 The cranial anatomy of the mail-cheeked fishes. Zoologica, Bd. 22 (Hft. 57), Stuttgart. 1914 The pituitary fossa and trigemino-facialis chamber in Ceratodus forsteri. Anat. Anz., Bd. 46, Jena. i (In press), The myodome and the trigemino-facialis chamber of fishes, and the corresponding cavities in higher vertebrates. Goronowitscu, N. 1888 Das Gehirn und die Cranialnerven von Acipenser ruthenus. Ein Beitrag zur Morphologie des Wirbelthierkopfes. Morph. Jahrbuch, Bd. 13, Hfts, 3/4, Leipzig. Greit, A. 1913 Entwickelungsgeschichte des Kopfes und des Blutgefasssys- ‘temes von Ceratodus forsteri. Zweiter Teil: Die epigenetischen Er- werbungen wiihrend der Stadien 39-48. Jenaische Denkschriften, Bd. 4, Jena. Herrick, C.J. 1909 The criteria of homology in the peripheral nervous system. Jour. Comp. Neur., vol. 19, Philadelphia. His, W. 1887 Die morphologische Betrachtung der Kopfnerven. Archiv f- Anat. u. Physiol., Anatom. Abtheil., Leipzig. Horrman, C. K. 1886 Weitere Untersuchungen zur Entwicklungsgeschichte der Reptilien. Morph. Jahrb., Bd. 11, Leipzig. Lanpacre, F. L. 1912 The epibranchial placodes of Lepidosteus osseus and their relation to the cerebral ganglia. Jour. Comp. Neur., vol. 22, no. ile 1916 The cerebral ganglia and early nerves of Squalus acanthias. Jour. Comp. Neur., vol. 27, Philadelphia. Luruer, A. 1913' Uber die vom N. Trigeminus versorgte Muskulatur der Ganoiden und Dipneusten. Acta Soc. Scientiarum Fennicae, Tome 41, No. 9, Helsingfors. Norris, H. W. 1913 The cranial nerves of Siren lacertina. Jour. Morph., vol. 24, no. 2, Philadelphia. Pinxus, F. 1894 Die Hirnnerven des Protopterus annecten. Morphol. Ar- beiten, Bd. 4, Jena. Srannius, H. 1849 Das peripherische Nervensystem der Fische. Rostock. Wie, J. W. van 1882 Uber das Visceralskelet und die Nerven des Kopfes der Ganoiden und von Ceratodus. Niederl. Archiv fir Zoologie, Bd. 5, H. 3, Leiden. THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. | Resumido por el autor, Darmon Artelle Rhinehart. El nervio facial del ratén albino. La porcion motriz del nervio facial del rat6én de distribuye sobre la musculatura facial e hioidea. El nervio intermedio se compone de fibras aferentes y eferentes, y una parte de las Ultimas forma una raiz separada. El ganglio geniculado contiene los cuerpos celulares de las fibras aferentes, perteneciendo estas células al tipo bipolar. El nervio petroso superficial mayor lleva fibras aferentes y eferentes al ganglio esfenopalatino, que no recibe fibra aleuna del trigémino; sus fibras terminan en la glindula de Stenson, glandula lagrimal, glandula del tabique y vasos san- guineos de la nariz y en el paladar. El autor no ha podido determinar exactamente la inervacién de las glindulas palatinas y botones gustativos. La cuerda del timpano envia fibras a los ganglios de las glindulas submaxilar y sublingual y termina en la lengua; lleva las fibras gustatorias de los dos tercios anteriores de dicho 6rgano y probablemente desempena otras funciones. El autor hallé cerca de mil células ganglionares, de relaciones y funcién desconocida, en el trayecto de los nervios de una de las mitades de la lengua. Las fibras cutdneas procedentes del ganglio geniculado terminan en la piel del meato auditivo ex- terno, piel de la oreja, y es posible también que inerven parte de la membrana del timpano; estas fibras forman el ramo cutdneo-facial, rama independiente de’ facial. Las fibras del ramo auricular del vago forman una parte de este nervio y fibras del cutdneo-facial se distribuyen por el ramo auricular del vago. El nervio petrosomenor superficial falta en el rat6n. Translation by Dr. José Nonidez, Columbia University. AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 2 THE NERVUS FACIALIS OF THE ALBINO MOUSE D. A. RHINEHART Anatomical Laboratory of the Medical Department of the University of Arkansas FOURTEEN FIGURES CONTENTS MURR ERSE RMS e's. so ccs cine Svs eee acim eel Se ewe sce ceases babe ek 81 ee Ie ep TPICE IOUS S).545 8 5 koe et EE eae Oe ER «hE iks GAs EME Silay 82 Perea n Gee RReRNer WS, LACIAlIS..... 0. sisi coseoremomhereniel Mirtle xe.’ er \ Ce: MTenTym Fig. 3 Sagittal section through the nervus facialis and the ganglion geniculi. 60 » lateral to figure 2. X 166. N.Int.e.). Followed centrally from this position, it passes pos- teriorly and medially for a short distance across the surface of the pons. At the ventral edge of the spinal root of the trigemi- nal nerve it enters the pons and passes dorsally and medially, parallel with, but more lateral and posterior to the emerging motor part of the facial nerve. At the level of the ascending part of the facial nerve the compact bundle becomes broken up and the fibers scattered, so that they cannot be followed farther. 92 D. A. RHINEHART Peripherally, the efferent root of the nervus intermedius passes around the anterior border of the motor part of the facial nerve, under the rest of the nervus intermedius, through the medial border of the geniculate ganglion and into the great super- ficial petrosal nerve. It is certain that only a few, if any, of its fibers are connected with the cells of the geniculate ganglion. This part of the nervus intermedius is present in all my series and undoubtedly contributes efferent fibers to the great super- ficial petrosal nerve. In the material used in this work it was impossible to tell what proportion of the fibers of the nervus intermedius were medul- lated and what non-medullated. All of the axis-cylinders are small in diameter and the medullated sheaths must be very thin. Weigner (’05) has given an excellent description of the fibers of the nervus intermedius of the ground-squirrel. In teased preparations stained with osmic acid he found that most of them are very fine with very delicate myelin sheaths (2 u in diameter), that a few are larger (5 uw), and that the nerve con- tains nonmedullated fibers. The fibers of the nervus inter- medius of the mouse are probably similar in size and character. Weigner describes numerous anastomoses between the nervus intermedius, the vestibular nerve, and the facial nerve, which are characterized by the presence of ganglion cells, scattered and in groups, along the anastomosing bundles. These ganglion cells, he states, are similar to those of the geniculate ganglion and furnish additional centers of origin for nervus intermedius fibers. In the mouse, no ganglion cells are present central to the dorsal part of the ganglion and no true anastomoses were observed. That the afferent fibers of the nervus intermedius terminate in the anterior extremity of the nucleus of the fasciculus solitarius or a group of cells lying anterior to it seems well established. All of the available papers dealing with this point in lower ver- tebrates state this to be true. Van Gehuchten (’00) found this to be the termination of these fibers in rabbits. The efferent fibers to the submaxillary and sublingual glands arise from a nucleus in the formatio reticularis at the level of the motor facial nucleus. This nucleus was called the nucleus salivatorius NERVUS FACIALIS OF ALBINO MOUSE 93 by Kohnstam (’02) and has been described by Yagita and Ha- yama(’09). Herrick (’16) calls it the nucleus salivatorius superior. The efferent fibers of the great superficial petrosal nerve arise from other cells in the same region (Yagita, ’14). In the mouse, three distinct bundles of nerve fibers continue the nervus intermedius peripheral to the ganglion. From the anterior angle of the ganglion the great superficial petrosal nerve emerges, while from its lateral angle two bundles of fibers pass peripherally in a common sheath with the motor fibers of the facial nerve. Before the genu is reached these bundles lie along the antero-ventral side of the facial nerve, the smaller of the two lying ventral to the larger (figs. 9 and 10, N.Ch.Ty., R.Cut.N.F.). In ‘sagittal series medial to the genu these bundles are always more or less separated by delicate connective tissue septa from each other and from the motor fibers of the facial nerve. Beyond the genu they lie along the lateral side of the nerve (fig. 10). In this position the septum is not as distinct as it is nearer the gan- glion. They are well separated in some series, while in others an indentation along their medial side is all that indicates a division into two parts. The ventral and smaller of these two bundles becomes the chorda tympani, the dorsal and larger contains fibers which are distributed to the skin of-the auricle and form a nerve which I have called the ramus cutaneus facialis. With the exception of the efferent fibers of the great super- ficial petrosal nerve, it was impossible in this material to tell the exact relation of the fibers of the nervus intermedius to the cells of the geniculate ganglion. That some of the fibers are connected with the cells of the ganglion and others pass through without interruption seems well established. | Cutting the chorda tympani in the middle ear results in a degeneration of about four-fifths of the cells of the geniculate ganglion (Amabilino, ’98; DeGaetani, ’06). Nissl degeneration of nerve cells in the brain stem after cutting the fibers to the submaxillary and sublingual glands is proof that the chorda tympani contains fibers which do not have their fibers in the ganglion (Kohnstam, ’02; Yagita and Hayama, ’09). 94 D. A. RHINEHART There is also considerable evidence for the presence of both afferent and efferent fibers in the great superficial petrosal nerve. By staining the geniculate ganglion of the mouse by the Golgi method von Lenhossék (’94) found fibers which pass directly from the nervus intermedius into the great superficial petrosal nerve. From this he erroneously concluded that the great superficial petrosal is a motor nerve for the supply of the levator veli palatini and levator uvulae muscles. After cutting the great superficial petrosal nerve in dogs, Yagita (’14) found that about one-twelfth of the cells of the geniculate ganglion show typical Nissl degeneration, and that there were degenerated cells in the formatio reticularis of the same side of the pons extending from the middle to the upper third of the facial nucleus. Weigner (’05) mentions a bundle which passes directly from the facial nerve to the great superficial petrosal through the human geniculate ganglion. In the ground-squirrel he describes a bundle between the nervus intermedius and the great super- ficial petrosal nerve which has no connection with the ganglion cells. He could not determine, he states, whether the fibers of this bundle are processes of the scattered cells in the nervus inter- medius or of those in the great superficial petrosal nerve. It is safe to conclude, therefore, that both the chorda tympani and the great superficial petrosal nerve contain afferent fibers whose cell bodies are located in the geniculate ganglion, and effer- ent fibers which pass through the ganglion without connection with its cells. BRANCHES OF THE NERVUS FACIALIS 1. Nervus petrosus superficialis major It has been shown above that the nervus petrosus superficialis major arises within the cranial cavity from the anterior pointed extremity of the ganglion geniculi, and that it is composed of fibers whose cell bodies are located in the ganglion, and others which form a separate efferent root of the nervus intermedius. From its origin this nerve passes anteriorly for a short distance | along the lateral side of the ventral surface of the ganglion semi- NERVUS FACIALIS OF ALBINO MOUSE 95 lunare. After a short course it bends at almost a right angle and passes medially, ventral to the nervus trigeminus, and ventral to the internal carotid artery and the sympathetic plexus which surrounds it. At a point just medial to the artery the nerve bends anteriorly and is joined along its medial side by the nervus petrosus profundus, the two uniting to form the nervus canalis pterygoidei. A. Nervus petrosus profundus. The nervus petrosus profun- dus, in the mouse, is formed by two small bundles of fibers from the internal carotid plexus. It does not have a separate course for the fibers join the great superficial petrosal nerve immediately after leaving the plexus and while still in relation to the artery. The fibers composing this nerve can readily be followed to the interior of the superior cervical sympathetic ganglion. Along the nerve of the pterygoid canal they occupy, at first, a position on its medial side Farther anteriorly, however, they become intermingled to such an extent with those from the great super- ficial petrosal nerve that the fibers from the two sources cannot be separately identified. B. Nervus canalis pterygoidet. In mice, the internal carotid artery enters the cranium and the nerve of the pterygoid canal leaves it through a slit-like fissure between the tympanic and periotic bones (eustachian aperture). From this fissure the nerve passes anteriorly and slightly medially along the ventral surface of the body of the sphenoid bone. One of the pharyngeal muscles (probably the salpingo-pharyngeus), the pharyngeal opening of the auditory tube and a mass of glands lie ventral to it. At the anterior border of the opening of the auditory tube the muscle and glands disappear and the nerve lies between the bone and the mucous membrane of the nasal cavity. At this place the pterygoid process extends ventrally to the soft palate, the nerve lying medial to it where it fuses with the body of the bone above. In this position the nerve extends anteriorly for some distance finally passing laterally through a foramen into the most posterior part of the orbit. The nerve of the pterygoid canal is, at first, flattened dorso- ventrally becoming circular in outline more anteriorly with the THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, No. 1 ’ 96 D. A. RHINEHART fibers more loosely arranged. All of its fibers are very fine. There are no ganglion cells either along the great superficial or the deep petrosal nerves and only a few along the nerve of the pterygoid canal. With the exception of one elongated microscopic ganglion about the middle of its course, these are single and widely separated. In one series a few fibers were given off from the nerve of the pterygoid canal to the mass of glands lying dorsal to the audi- tory tube. Similar fibers could not be found in any other series, nor were there other branches from this nerve. In addition to the sympathetic fibers to the sphenopalatine ganglion by way of the nervus petrosus profundus and the nervus canalis pterygoidei, others from the internal carotid plexus reach the ganglion by way of the nervus abducens. ‘These fibers join the nervus abducens as two bundles slightly anterior to the point of union of the two petrosal nerves.’ They leave the nervus abducens as four small bundles, two of which join the ophthalmic nerve and the other two the sphenopalatine ganglion just posterior to its middle (figs. 4 and 7, bundles 6 and c). Koch (’16) mentions the presence of sympathetic nonmedul- lated fibers from the cavernous plexus in the nervus abducens of the dog. These leave the nerve more anteriorly to pursue an in- dependent course. Piersol (’13) states that branches of the sphenopalatine ganglion have been described joining the nervus abducens. C. Ganglion sphenopalatinum. It will be shown later that none of the fibers of the sphenopalatine nerves end in the spheno- palatine ganglion. This ganglion belongs, therefore, more to the facial than to the trigeminal nerve. Most of the fibers from it, however, are distributed as constituents of the palatine nerves. For this reason, it has been found necessary to study the ganglion itself and the nerves connected with it in any way. The sphenopalatine ganglion in the mouse is an elongated mass of ganglion cells lying between the medial side of the max- illary nerve and the medial wall of the orbit. It begins as a small accumulation of cells along the ventral border of the nerve of the pterygoid canal immediately after that nerve enters the NERVUS FACIALIS OF ALBINO MOUSE 97 orbit. Anteriorly, the ganglion gradually increases in size, the cells completely surrounding the nerve. This continues until the anterior limit is reached where it ends in a blunt extremity. The posterior half of the ganglion is flattened laterally and is, in cross-section, an elongated oval with the long axis extending Fig. 4 Transverse section through the most posterior part of the orbit show- ing the nervus oculomotorius, nervus trochlearis, nervus ophthalmicus, nervus maxillaris, nervus abducens, the posterior part of the ganglion sphenopalatinum and the origin of the nervus palatinus posterior. By referring to figures 7 and 8 the connections of the nerve bundles marked with small letters and arabic numerals will be seen. Figures 4, 5, and 6 are from the same series. X 90. vertically (fig. 5). Near its middle the ganglion changes its shape and becomes flattened dorsoventrally (fig. 6). At this point a ventral bend in the ganglion together with its change in shape imperfectly separates it into a posterior and an anterior part. 98 D. A. RHINEHART Medially the sphenopalatine ganglion is separated by the bone from the nasal cavity. Laterally it is related to the max- illary nerve throughout its entire extent, the sphenopalatine nerves extending ventrally along its lateral side at about its middle. The sphenopalatine ganglion of mammals is described as receiv- ing the sphenopalatine nerves from the maxillary nerve and as giving off branches which are distributed to the mucous mem- brane of the nose, mouth and pharynx, and branches to the orbit. All of the nerves which are usually mentioned in connection with the ganglion, with the exception of the pharyngeal branch, are represented in the mouse by similar nerves. a. Nervi palatinus posterior et medius. The posterior and middle palatine nerves arise from the lateral side of the maxillary nerve widely separated and distinct from both the sphenopalatine ganglion and the sphenopalatine nerves. They will be described, therefore, before the sphenopalatine nerves. The posterior palatine nerve is represented by one or two small nerves coming from the lateral side of the maxillary nerve at the level of the middle of the pesterior part .of the sphenopalatine ganglion (fig. 4, N.Pal.P.). It is separated by the maxillary nerve from the sphenopalatine ganglion, and arises from a part of the maxillary nerve which gives off branches for the supply of the teeth and gums of the upper jaw. The posterior palatine nerve receives two small bundles of fine fibers from the ventral border of the sphenopalatine ganglion (fig. 4, d,e). These pass laterally, ventral to the maxillary nerve, and join the posterior palatine immediately before it enters the * canal in the bone through which it passes to the palate. They are so small that in some series it is difficult to follow them even by using the high power of the microscope. Unsuccessful at- tempts were made to count the fibers they contain. I believe it safe to conclude, however, that they do not exceed thirty-five or forty in number. The middle palatine nerve is represented by a single smaller branch which takes origin from the ventral side of the maxillary nerve. This nerve passes inferiorly where it, too, enters a canal NERVUS FACIALIS OF ALBINO MOUSE 99 in the bone and extends to the mucous membrane of the palate. In some series it was possible to follow a minute bundle from the sphenopalatine ganglion: into it; in other series this could not be done. When present this bundle is even smaller than the ones joining the posterior palatine. The termination of these nerves will be discussed later. Fig.5 Transverse section through the nervus maxillaris, the ganglion spheno- palatinum, and the nervisphenopalatini. 810. anterior to figure 4. X 90. b. Nervi sphenopalatini. The sphenopalatine nerves take origin from the ventromedial part of the maxillary nerve at about the middle of the sphenopalatine ganglion (fig. 5, bundles a, b, c, d, e, g, 7, and part of f.). The bundles of fibers present in these nerves vary in different series. In the one represented in the graph, figure 7, and from which figures 4, 5, and 6 were drawn, there were six distinct bundles at their point of origin. 100 D. A. RHINEHART Almost immediately their arrangement becomes so complicated that it cannot be easily described. ‘This arrangement is, how- ever, accurately represented in the graph (fig. 7, N.Sp.). A greater part of five of the six bundles assist in forming the an- terior palatine nerve and its posterior inferior nasal branch, and the greater part of the other bundle assists in forming the naso- palatine nerve. From these six bundles there are only two small subdivisions that enter the sphenopalatine ganglion, joining it well toward its anterior end and close to the point of origin from the ganglion of the fibers which enter the nasopalatine nerve. In transverse series these bundles can be followed through the ganglion into the nasopalatine nerve, which is probably their fate in the other series. ce. Nervus palatinus anterior. The anterior palatine nerve is the largest of the nerves in this region. The bundles from the maxillary nerve which form it pass medially and anteriorly across the ventral border of the sphenopalatine ganglion. While in this position they are joined by four bundles whose fibers come from the interior of the sphenopalatine ganglion. The fibers from the ganglion are smaller than those from the maxillary nerve and represent about one-third of the fibers of the anterior palatine nerve. From the ventral border of the sphenopalatine ganglion the anterior palatine nerve passes ventrally through a canal to reach the mucous membrane of the palate. In this canal one bundle separates from the nerve, and as soon as the mucosa is reached, takes a posterior direction. The remainder of the nerve passes anteriorly just lateral to the middle of the hard palate. The posterior inferior nasal branch of the anterior palatine nerve is made up of two small bundles of fibers from the spheno- palatine nerves (fig. 6, a, 6) and four smaller bundles of fine fiber from the sphenopalatine ganglion (fig. 7). This nerve extends anteriorly along the lateral wall of the nasal cavity, the smaller fibers soon leaving it to enter a mass of glands in the lateral wall of the nose (gland of Stenson). The fibers of larger size end in the mucous membrane of the lateral wall of the nasal cavity. NERVUS FACIALIS OF ALBINO MOUSE 101 d. Nervus nasopalatinus. One bundle of the sphenopalatine nerves passes obliquely forward and medially across the inferior surface of the anterior part of the sphenopalatine ganglion, finally reaching the medial border of the anterior end of the gan- glion (fig. 5, bundlea). In this position it receives a large contri- bution of fine fibers from the ganglion and becomes the nasopala- J s, 3 * $8 20% o90 20. 3 S9ggsee Zoos Reese Oe GarL2039 26 6®. %% 8 01 Oe Cae ® Fig. 6 Transverse section through the nervus maxillaris and the anterior extremity of the ganglion sphenopalatinum showing the origin of the nervus naso- palatinus and two of the orbital branches of the ganglion. 600 » anterior to figure 5. X 90. tine nerve (fig. 6). The fibers from the ganglion form a little more than half of this nerve. From its origin the nasopalatine nerve extends medially along the posterior wall of the inferior meatus of the nose to reach the posterior and ventral part of the septum. From this point its course is anterior along the ventral part of the septum. Through- 102 D. A. RHINEHART out its extent branches are given off, some of coarse fibers to the mucous membrane, and others composed of finer fibers from the sphenopalatine ganglion which pass into the glands of the sep- tum. The septal glands are numerous in the region posterior to the vomeronasal organ. Finally, reduced in size and composed only of fibers of larger size, the nasopalatine nerve passes through the incisive foramen to reach the roof of the mouth. e. Other branches. In addition to the fibers from the spheno- — palatine ganglion described above, there are several other small nerves either arising or ending within it. One small nerve ex- tends from the ganglion posteriorly along the medial side of the semilunar ganglion (fig. 7). Several other small bundles extend from the sphenopalatine ganglion dorsally and laterally to join the ophthalmic nerve. (These are not shown in the graph, fig. 7.) The origin, the direction or the endings of these nerves could not be determined. Several small nerves take origin from the anterior part of the sphenopalatine ganglion and pass dorsally along the medial wall of the orbit. Some of these join the nasociliary nerve, and through it finally terminate among the gland ducts of the an- terior and inferior part of the lateral wall of the nasal cavity. ‘Others accompany a blood-vessel into the medial part of the lacrimal gland, while one other joins a small branch of the maxil- lary nerve which terminates in the nasolacrimal duct. Other branches from the anterior end of the ganglion join the arteries in the neighborhood and accompany them in their peripheral courses. . f. The nerve supply of the palate. The nerve supply of the palate was investigated to determine, if possible, the relationship between the nerves of the taste-buds, the nerves to the palatal glands, and the facial nerve. Although no positive conclusions were reached, certain features are of interest and will be recorded here. From the ventral ‘end of its bony canal the anterior palatine nerve extends anteriorly, in company with.a medium-sized ar- tery, as far forward as the nasopalatine canal. Throughout its entire extent it sends branches medially and laterally to supply NERVUS FACIALIS OF ALBINO MOUSE 103 one-half of the palate. Anterior to the nasopalatine canal the mucous membrane of the medial part of the palate is supplied by the terminal branches of the nasopalatine nerve, that of the lateral part by a branch of the maxillary nerve. Branches from these nerves are abundant under the transverse ridges of the palate and particularly so around the nasopalatine canal. In the tunica propria the nerve fibers form rather wide meshed plexuses from which individual fibers can be followed into the epithelium. No encapsulated nerve endings were found. In this areaof the palate there are no taste-buds or glands. From the anterior palatine nerve many fine fibers can be followed along the walls of the blood-vessels, presumably to end in their muscle fibers, although this termination could not be positively identified. The graph (fig. 8) shows the distribution of the middle and posterior palatine nerves and the posteriorly directed branch of the anterior palatine nerve to the posterior part of the hard pal- ate and the soft palate. The transverse line X shows the place of union of the hard and soft palate. The graph includes one- half of the palate, the right margin representing the median line and the left margin the lateral edge. In this connection it must be remembered that the anterior part of the palate shown in the upper part of the graph is wider than the posterior part, thus accounting for the apparent preponderance of nerves in the upper part of the graph. The small circles show the approximate locations of the taste-buds, and the broken line represents the limits of the palatal glands. In the series represented in the graph, there were twenty-five taste-buds, twelve in the posterior part of the hard palate, and thirteen in the soft palate. In other series the number is ap- proximately the same and the arrangement is similar. They are. most numerous along the posterior part of the hard palate and gradually decrease in number posteriorly. The nerves of this region can be followed from their emergence from the bony foramina to their termination. Those supplying the epithelium can be traced to the basement membrane and some of the fibers were seen to pass among the epithelial cells. 104 D. A. RHINEHART Fig. 7 A graph, made according to the method devised by Prof. A. G. Pohl- man, showing the connections of the nervus petrosus superficialis major and the ganglion sphenopalatinum. A graph of this sort is made by selecting that part of a series which is to be used, and marking out-on one edge of a piece of milli- meter plotting paper a set of stair steps for each row of sections on the slides, representing each section by a single step. Nerves or other structures are repre- sented in the graph by broken or solid lines, dots, etc., and as these structures con- tinue through the series the lines or dots are continued on the paper, passing through a millimeter for each section. By continuing this and representing changes in position, branchings, anastomoses, etc., by changes in the lines, al- most any structure which passes for some distance through a series can be graph- ically shown. The graph shown as figure 7 represents certain nerves and ganglia and their connections in slides 16 to 24 of a tramsverse series of one half of a mouse head. The numbers on the left of the graph are the numbers of the slides and the rows on each slide. Each section is represented by a single step in the stairs. As an illustration, the ganglion geniculi is present in section 1, row 1, slide 16, and is shown at the lower right part of the graph. The ganglion ends and the nervus petrosus superficialis major begins in the last section in that row. This nerve continues through the second into the third row where it bends medially, this being shown by a bend in the line representing the nerve. In section 8, row 1, slide 17, and in section 2 of the second row the nervus petrosus profundus joins the nervus petrosus superficialis major to form the nervus canalis pterygoidei. The nervus canalis pterygoidei continues through the series until row 3, slide 21, is reached where the ganglion sphenopalatinum begins, the ganglion being repre- sented by an unshaded area continuing the course of the nerve. Bundles of sympathetic fibers and branches of the ganglion sphenopalatinum are represented by broken lines, cranial nerves and their branches by solid lines, and ganglia by unshaded areas along or within the nerves. Figures 4, 5, and 6 were drawn from sections of thé same series used in making the graph, the loca- tion of these sections is indicated by the broken transverse lines. The small letters along the lines indicate bundles of nerves which are shown and similarly labeled in the figures. Fig. 8 A graph showing the distribution of the a ee and middle palatine nerves and a part of the anterior palatine nerve to one half of the soft palate and the posterior part of the hard palate, made from the same slides which were used in making the graph shown as figure 7. The small circles indicate the ap- proximate locations of the taste-buds and the broken line the limits of the pala- tal glands. The arabic numerals indicate branches of the nerves which are shown under the mucosa of the palate in figures 4 and 5. The broken transverse line marked x indicates the place of junction of the hard and the soft palate. For further explanation see the text. ’ 105 ALBINO MOUSE FACIALIS OF NERVUS 1 Sy Nt K iD NY is aN a dy . =O S Lae N SS \ NX eS \ NS Ne \ \ Is = W ~ \\ . i : Wout i ¥y i f 1 im 1 Fw AT =! Hatt |e aes EEO: | 4 reer = liner! a = e Pa le ) | Pa X ries eels Lm io) N.Max .R Inf.Na. 23 106 D. A. RHINEHART On these nerves no encapsulated nerve endings were found. The nerves of the glands are not so well demonstrated. Bundles of fibers can be traced until they break up into smaller bundles or | isolated fibers among the gland alveoli. The nerve supply of the taste-buds, with one exception, could be identified as coming from one or the other of the palatine nerves. In the series represented in the graph the posterior palatine nerve is represented by two relatively large nerve bundles. These emerge through separate canals in the bone. One of them is directed anteriorly and supplies a small part of the lateral side of the hard palate, the other passes posteriorly to supply a limited area of the hard palate and one entire half of the soft palate. The middle palatine nerve is distributed to the lateral part of the hard palate anterior and lateral to the place of emer- gence from its canal. Two bundles of the anterior palatine nerve are shown, one passing medially and the other posteriorly, to supply an area of the hard palate medial to that supplied by the other nerves. Practically the entire mass of the palatal glands and all but a few of the taste-buds are located in the area sup- plied by the posterior. palatine nerve. In the nerves connected with the sphenopalatine ganglion there are two varieties of fibers, those of large size from the maxillary nerve, and those of small size from the sphenopalatine ganglion. In the nasopalatine nerve and in the posterior inferior nasal branch of the anterior palatine the fine fibers are grouped and leavethe nerves as separatebranches. Thesebranches enter glands and are assumed to contain glandular fibers although they are not stained within the glands and could not be followed to their final termination. In the posterior branch of the anterior palatine and in the middle and posterior palatine nerves the fine fibers are inter- mingled with those of larger size and terminate in company with the larger fibers. Except certain of the fine fibers which enter the walls of the blood-vessels the endings of the finer fibers could not be determined. In the area of the palate supplied by the posterior palatine nerve the fine fibers presumably supply the palatal glands, the muscle fibers in the walls of the blood-vessels, NERVUS FACIALIS OF ALBINO MOUSE 107 and the taste-buds, if these structures receive their innervation from the facial nerve. From what has been said above, however, concerning the minute bundles of fine fibers from the sphenopala- tine ganglion to the posterior palatine nerve, it does not seem possible that they are numerically sufficient for the supply of all of these structures. The only other possible source for their nerve supply is from the sympathetic or the trigeminal nerve. The evidences from comparative anatomy do not materially assist in clearing up this problem. In petromyzonts, Johnston (08) found that the maxillary nerve supplied the roof of the mouth. The palate of bony fish (Herrick, ’99, ’00, ’01) is sup- plied by the ramus palatinus VII. In the amphibians Coghill (01, 702) and Norris (’08, 713) describe anastomoses between the ramus palatinus VII and branches from the fifth nerve, the re- sulting nerves being distributed to the roof of the mouth, the teeth and the nasal capsule. In none of these forms is there a sphenopalatine ganglion present, although Johnston (’08) de- scribes ganglion cells in the roof of the mouth in cyclostomes, and Norris (08), in Amphiuma means, mentions ganglion cells at cer- tain points on the anastomoses between the ramus palatinus VII and ramus ophthalmicus profundus V. Norris, in Siren lacertina (13) also describes branches from the ramus es an VII to the vessels of the roof of the mouth. In this connection it is interesting to note that Herrick (16, p. 248) says, ‘‘Unlike the visceral. sensory system, however, its (referring to the gustatory apparatus) peripheral fibers have no connection with the sympathetic nervous system and the reac- tions may be vividly conscious.” If this statement be literally true, then the taste-fibers of the palate must come from the trigeminal nerve. This is hardly probable, for Herrick (’01) has shown that in the siluroid fishes, where the taste-buds are very numerous, none of them are directly supplied by the trigeminal nerve. If this statement means that taste-fibers have no con- nection with sympathetic ganglion cells, then it is possible for the taste-buds of the palate to be supplied by fibers from the facial nerve which pass through the sphenopalatine ganglion without interruption. 108 D. A. RHINEHART In conclusion it can be said that, in the mouse, the epithelium of the palate is supplied by fibers from the maxillary nerve, and that the muscle fibers in the walls of the blood-vessels are sup- plied by fibers from the sphenopalatine ganglion. It does not seem possible for the taste-buds and glands of the palate to be supplied by fibers from the sphenopalatine ganglion. The evi- dences that they are supplied by the trigeminal nerve is equally inconclusive. In the absence of other direct anatomical obser-_ vations on this problem in mammals it must, for the time being, remain an unsettled question. Until recently the nerve supply of the m. tensor veli palatini and the m. levator uvulae was believed to come from the facial nerve. In the mouse the branch from the mandibular nerve to the internal pterygoid muscle passes ventrally around the sta- pedial artery and, after supplying the internal pterygoid, sends one branch into the levator veli palatini and another which passes ventral to the otic ganglion and terminates in the tensor tympani muscle. The tensor veli palatini and the levator uvulae muscles are supplied by a branch from the accessory portion of the spinal accessory nerve which sends fibers into these muscles and terminates in the muscles of the pharnyx. 2. Nervus stapedius The first branch of the facial nerve peripheral to the genicu ate ganglion is the nervus stapedius which arises in the dorsal wall of the tympanic cavity. In the mouse this is undoubtedly a motor nerve. The fibers composing it come from the medial side of the facial stem (figs. 10, 11, and 12, N.Stap.). They unite into a small bundle at the junction of the medial and dorsal borders of the facial nerve and pass into the stapedius muscle, to terminate after the manner of motor nerves elsewhere. There are no ganglion cells along its course, nor does it contain fine fibers. Weigner (’05), in the ground-squirrel (Ziesel), describes a ganglion at the place of origin of the nervus stapedius and states that it contains many fine fibers similar to those of the nervus NERVUS FACIALIS OF ALBINO MOUSE 109 intermedius. He also reports that cutting the facial nerve where it passes across the cochlea did not result in a degeneration of all of the nervus stapedius although there is a degeneration of the entire facial nerve distal to the cut. He offers no explanation for this, other than the implied one that the nervus stapedius is composed mostly of fibers arising in the microscopic ganglion located at its place of origin from the facial nerve. sifieg fife My sf BN yaa se bON Vest. Fig. 9 Sagittal section through the nervus facialisat the lateral edge of the ganglion geniculi and medial to the genu externum showing the two bundles of fine fibers which pass from the ganglion peripherally in the nervus facialis. Same series as figures 1, 2, and 3. 225 u lateral to figure 3. x 166. Fig. 10 Transverse section through the nervus facialis just posterior to the genu externum and through the origin of the nervus stapedius. The two bundles of fine fibers are shown along the lateral side of the nerve. X 166. Remembering that the mouse and the ground-squirrel both belong to the order rodentia, an order of mammals in which there is little variation in anatomical structure, I am unable to account for the differences in this nerve in the two animals. 110 D. A. RHINEHART 3. Nervus chorda tympani The chorda tympani arises from the ventral side of the facial nerve along the posterior part of the upper wall of the tympanic cavity between the origin of the nervus stapedius and the anas- tomosis with the ramus auricularis vagi. It is formed by the more ventral and smaller of the two bundles of fine fibers which come from the interior of the geniculate ganglion and pass pe- ripherally as a part of the facial stem (figs. 9, 10, 11, 12, and 138, N.Ch.Ty.). It has been shown above that it contains fibers which are processes of cells in the geniculate ganglion and others which pass through the ganglion without interruption. In the mouse, soon after its origin, the chorda bends medially and anteriorly and passes through a fissure-like opening into the tympanic cavity. Here it extends anteriorly in a shallow groove along the upper part of the lateral wall, and then along a groove in a spicule of bone which projects into the cavity. At the apex of this spicule the nerve crosses a small gap to reach the medial surface of the head of the malleus, across which it passes lying ventral to the attachment of the tendon of the tensor tympani muscle. In the anterior part of the cavity it passes along the medial side of the anterior process of the malleus, accompanying that process into the fissure between the tympanic and periotic bones (petrotympanic fissure) through which it extends into the infratemporal fossa. In the infratemporal fossa the chorda tym- pani passes medially and ventrally posterior to the emerging mandibular nerve. After a course of varying extent it joins the posterior aspect of the lingual nerve. . Just before joining the lingual there is usually an anastomosis with a bundle of fibers from another source. This is, in a ma- jority of my series, a small branch arising from the mandibular nerve. In one series a branch from the auriculotemporal nerve joins it; in another a branch from the lingual joins it, the two then uniting with the lingual; in one series there is no anastomo- sis of any sort. All of the branches which anastomose with the chorda tympani are composed of fibers of larger size than those in the chorda. When followed centrally they become lost in the trunk of the mandibular nerve. NERVUS FACIALIS OF ALBINO MOUSE is In none of the series is there a connection of any sort with the otic ganglion. Weigner (05) mentions an almost constant anastomosis be- tween the lingual and mandibular nerves in man, the chorda joining the lingual just distal to it. He also describes a com- munication from the great superficial petrosal nerve to the fibers in the facial which form the chorda, and states that there are many scattered ganglion cells along the chorda tympani both before and after its origin from the facial. In the mouse no such communication can be identified. With the exception of one series, in which there is a small ganglion at its place of origin from the facial nerve, no ganglion cells are present in the chorda. A. The nerve supply of the tongue and the salivary glands. In one transverse series the nerves in the tongue were carefully fol- lowed. The results of this study will be briefly presented. The hypoglossal nerve is composed of fibers of uniform size re- sembling closely those in the motor portion of the facial nerve. Immediately after entering the tongue it divides into two branches which run anteriorly, one close to the septum and the other farther laterally. From these, small branches are given off in all directions, many of which were followed to their termination in motor nerve endings in the muscle fibers of the tongue. The glossopharyngeal nerve is composed of fibers of medium and small size resembling those of the lingual nerve. It enters the tongue between the hyoid bone and the thyroid cartilage and passes anteriorly for a short distance along the side of the base of the tongue. In this part of its course there is, in all my series, a ganglion almost as large as the otic and having the his- tological characteristics of a sympathetic ganglion. Just ante- rior to this ganglion the nerve breaks up into five bundles which enter the tongue and spread out in a fan-shaped manner. The most medial of these passes medially and posteriorly and supplies the mucous membrane of the tongue as far as the larynx. The next passes medially and anteriorly and supplies the dorsum of the tongue in the region of the single circumvallate papilla. This branch sends a number of fibers into the base of the papilla, which, together with a similar contribution from the opposite THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 1 En? D. A. RHINEHART side, supply this structure. These fibers form a plexus exter- nal to the valley and within the core of the papilla. From these plexuses fibers supply the taste-buds and: the epithelium of the papilla and the surrounding mucous membrane. In two series a microscopic ganglion is present within the papilla. The most lateral of the five branches of the glossopharyngeal nerve extends anteriorly along the lateral side of the tongue and terminates in the foliate papillae. The other two bundles are distributed to the tongue over an area extending as far forward as a line connecting the anterior limits of the foliate and cir- cumvallate papillae. The lingual nerve, because of the fibers in it from the facial nerve, was carefully studied. From one transverse series a graph of the course, branches, and distribution of this nerve within the tongue was made, and all the branches, as far as pos- sible, followed to their termination. From its union with the chorda tympani the lingual nerve passes medially, ventrally, and anteriorly to reach the tongue. It lies between the internal and external pterygoid muscles, then between the internal pterygoid and the mandible, and finally between the mylohyoid and the side of the tongue. It curves around the ducts of the submaxillary and sublingual glands, turns anteriorly lateral to the genioglossus, and inclines dorsally into the substance of the tongue. A few branches are given off from the lingual nerve before it enters the tongue. At the ventral edge of the internal pterygoid and under the mylohyoid muscle a number of small branches are given off which pass posteriorly to the submaxillary and sublin- gual glands. At the ventral edge of the internal pterygoid a larger branch arises which passes medially to supply an area of the mucous membrane of the cheek opposite the folliate papil- lae. This branch sends fibers into the epithelium, into a mass of glands, and into a number of taste-buds (eight in the series studied). Along the side of the tongue, in relation to the gland ducts, one relatively large and a number of smaller branches are given off which pass forward in company with the ducts and supply the mucous membrane of the floor of the mouth andthe gum behind the lower incisor teeth. NERVUS FACIALIS OF ALBINO MOUSE 113 In the tongue the lingual nerve passes forward midway be- - tween its lateral edge and the median septum. One branch is given off immediately .after the nerve enters the tongue. This passes dorsally and posteriorly to supply the area in front of that supplied by the glossopharyngeal nerve. The remaining branches arise irregularly and supply all of the tongue in front of this area. Some of the branches extend laterally, some to the ventral surface, but. the larger and most numerous branches pass to the dorsal surface. In the ventral part of the tongue a few small branches join those of the hypoglossal nerve, and a few others end in an undetermined manner among the muscle fibers. The majority of the branches, however, could be followed to the mucous membrane. In the series studied the taste-buds are located in both waiis of the valley around the circumvallate papilla, in the foliate papillae, in a small area of the cheek opposite the foliate papillae, and irregularly scattered over the dorsal surface of the tongue in the area supplied by the lingual nerve. There are no scat- tered taste-buds in the area supplied by the glossopharyngeal nerve nor along the sides or the ventral surface of the tongue. In the area supplied by the lingual nerve there are thirty-six taste-buds in one-half of the tongue. These are widely separated posteriorly and become more numerous as the tip of the tongue is approached. Each of them is placed near the surface of the epithelium on the top of a flat tunica propria papilla. One or two bundles of nerve fibers from the lingual nerve were followed into the bases of the papillae on which the taste-buds are located. In the papillae these fibers break up into a number of branches and form an intricate complex of very fine fibers. From these plexuses some fibers pass into the taste-buds. Other fibers pass into the surrounding mucous membrane, so that the taste-bud and a considerable area of the mucous membrane is supplied from the plexus in each papilla. The most striking feature of the nerves within the tongue is the large number of ganglia and ganglion cells along them. The majority of these cells are in clumps or microscopic ganglia along the nerve bundles. Along the glossopharyngeal nerve there are 114 D. A. RHINEHART six of these ganglia, along the lingual nerve they are far more numerous, being forty-one in number. Along the hypoglossal nerve there are scattered smaller ganglia and isolated cells. Fib- ers from the lingual, the glossopharyngeal and even the hypo- glossal nerve can be followed into these ganglia to end in an unde- termined manner among the ganglion cells. A count of the cells in these ganglia along the lingual and hy- poglossal nerves was made. Only those in the ganglia in which there is a definite nucleolus were counted, so that the figures obtained are less rather than more than the actual number present. In this one series, in one-half of the tongue, there were 1079 cells, 957 along the lingual and 122 along the hypoglossal nerve. While the presence of ganglion cells in the tongue has long been known, I have failed to find a reference as to their number or their significance. They are usually dismissed with the state- ment that they are sympathetic cells. I am of the opinion that much will be added to the knowledge of the nerve supply of the tongue when the central connections, the endings of the fibers, and the functions of these ganglion cells have been worked out. In this work nothing has been found other than in support of the generally accepted view of the nerve supply of the tongue. That the taste-buds are supplied by the chorda tympani has long been known. The course of the taste-fibers into the brain was, however, for a time, an unsettled question. From clinical cases evidences have been deduced in support of every possible pathway for these fibers into the brain stem. Many of these are included in the review of the literature in the articles by Cushing (’03), Weigner (’05), and Sheldon (’09). The carefully conducted experiments, of Cushing (’03) have done much to prove that the taste-fibers enter the brain over the nervus intermedius. In the mouse the anatomical evidence supports this view. No connection, such as was found by Weig- ner (’05) between the chorda tympani and the great superficial petrosal nerve is present, and the entire absence in all my series of a communication from the facial nerve to the tympanic plexus excludes even the possibility that the taste-fibers reach the brain except through the nervus intermedius. NERVUS FACIALIS OF ALBINO MOUSE 115 Cushing (’04) has presented very convincing clinical eviderce that the chorda tympani supplies the tongue with certain forms of common sensation: After trigeminal neurectomy he found that sensations of pain and temperature and tactile sensations are absent in the area supplied by the lingual nerve. There re- mained, however, the ability of the patient to appreciate the presence, the general location, and the movement of a piece of cloth or a cotton swab across this area. The nerve supply of the submaxillary and sublingual glands in the mouse corresponds closely to that given by Langley (’90) and Huber (’01) for the dog and cat. In the mouse the sublingual or the retrolingual gland is located lateral and ventral to the sub- maxillary and is pure mucous in type while the submaxillary is pure serous. The nerves to these glands leave the lingual as a number of branches (five to eight) arising deep to the mylohyoid muscle ard rather widely separated. These nerves come into relation with the ducts and pass with them into the glands. Along these rerves there are several small ganglia. In the series most care- fully studied there were five of these, two sending their fibers along the submaxillary duct, one along the sublingual duct, while the other two send fibers along the ducts into both glands. Within the submaxillary gland there are two large ganglia and ' several smaller ones. The ganglion cells in these ganglia are so ‘ numerous that if each is supplied with a basket-work from the chorda tympani as described by Huber, each chorda fiber must divide and terminate in relation to several cells. 33. After the cutaneous branch of the facial and the auricular branch of the vagus have separated from the facial stem, each nerve contains fibers from the other. Because the greater pro- portion of the fibers of one nerve come from the facial, it is con- sidered as a branch of that nerve, the other for the same reason. being considered as a branch of the vagus. 118 D. A. RHINEHART 0 N.to TymMemb. BL ONCh-Ty. N.to Dia, ies N.to Sty-Hy, 13 Fig. 13 Outline drawings of every other section through the nervus facialis in a horizontal series, beginning above and passing ventrally through the anas- tomosis with the ramus auricularis vagi, to show the mixing of the cutaneous facial fibers with those from the vagus. The motor facial fibers are unshaded, the fibers from the ganglion geniculi are black, and the vagus fiber cross-hatched. X 100. NERVUS FACIALIS OF ALBINO MOUSE 119 In different series there are some differences in the arrange- ment of these nerves. In one series the ramus auricularis vagi is composed of a single large bundle, in another the ramus cu- taneus facialis is composed, after its origin, of two bundles, and there are apparently some differences in the number of fibers which interchange. However, the essential arrangement is constant and as described above. Immediately after its origin the ramus cutaneus facialis passes from the medial to the lateral side of the posterior auricular nerve. From here it extends anteriorly and laterally above the external auditory meatus until the point of the attachment of the cartilage of the auricle to the side of the cranium is reached. At this point it enters the auricle, extending upward and anteri- orly along the medial side of that part of the auricle which bounds the pouch-like concha laterally (figs. 14A and 14B). It gives off branches in this region which supply the skin and hair follicles. Other branches pass dorsally beyond the concha and supply the skin of the posterior third of the lateral surface of the auricle. The ramus auricularis vagi sends a small branch to supply a part of the external auditory meatus and the tympanic membrane, this branch containing a few fibers from the facial (fig. 13D). Otherwise the nerve supply of the tympanic membrane is the same as that given by Wilson (’07, 10-11). After the origin of the branch to the tympanic membrane the auricular branch of the vagus passes along the cranial side of the concha. It supplies this area of skin and its terminal branches pass dorsally beyond the concha to supply the anterior two-thirds of the lateral surface of the auricle (fig. 14). These two nerves have been followed with great care from their origin to their place of ending. Each, in that part of its course where it is related to the auricular cartilage, passes between the cartilage and the skin, a location where there are no muscle fibers. Branches of each have been followed to their final end- ing and have been found to end in a plexiform manner immedi- ately under the epithelium or around the hair follicles. It is safe to conclude that they are both common sensory in function for the supply of the skin and hair of the auricle. 120 D. A. RHINEHART The anatomical and clinical evidence for the presence of gen- eral cutaneous fibers in the nervus facialis is meager in amount and inconclusive. In petromyzonts, Johnston (’08) describes gen- eral cutaneous fibers arising from the geniculate ganglion and passing to the ventrolateral surface of the head below and behind Cervical-coe Vogus “#” Facial x%* - 14 Fig. 14 A A diagram of the connections and distribution of the nervus facialis of the albino mouse, based largely on a flat reconstruction made by projection of sagittal sections on to a median plane. 14B_ A diagram of a coronal section through the auricle to show the areas of skin suvvlied bv the vagus, facial and cervical nerves. the orbit. In bony fish, Herrick (’99, ’00, ’01) found a general cutaneous component in the facial nerve for the region of the operculum, the fibers, however, being derived from the Gasserian ganglion. In the amphibia, Norris (’13) found a general cutane- ous component in Siren, and Herrick (’14) described fibers from the geniculate ganglion in Amblystoma larva which enter the ‘NERVUS FACIALIS OF ALBINO MOUSE 13t spinal V tract, this tract being considered as a tract whose nucleus is generally cutaneous in function. Van Gehuchten (’98) found that a few cells of the geniculate ganglion degenerate after cutting the facial nerve at its exit from the facial canal. In an abstract of an article by DeGaetani (’06) sensory fibers in the ramus temporofacialis are, mentioned. Weigner (’05) describes fibers of small caliber, presumably in- termedius fibers, in the facial nerve of man distal to the origin of the chorda tympani. The chief clinical evidence for the presence of such fibers in the facial nerve has been presented by Hunt (15). By a method which he calls the herpetic method and which is based on the fact that herpetic vesicles on the skin are caused by pathological proc- esses involving the ganglion from which the cutaneous fibers arise, he attributes the supply of certain parts of the auricle, including the concha, to the facial nerve. One of the probable pathways to the skin for these fibers which he mentions is through the auricular branch of the vagus. In the different cases reported the areas showing the herpetic eruption vary to some extent. If homologous conditions are found in man to those here reported in the mouse this variation can probably be accounted for by the mixing of the vagus and facial fibers at the point of the anastomosis. In the mouse there is no evidence that either the inner or middle ear receive fibers from the facial nerve. A communicating branch from the facial nerve to the tympanic plexus is not present in any of my series. 5. Nervus auricularis posterior and the nerves to the stylohyoid and digastric muscles The posterior auricular nerve arises from the lateral side of the facial nerve just distal to the origin of the ramus cutaneus (figs. 11 and 12,R.Aur.P.). It passes dorsally behind the auricle and is distributed to the muscles posterior to and along the cranial side of the auricle. It is composed of fibers from the motor part of the facial and is a purely motor nerve. A similar area of skin of the auricle is supplied by sensory fibers from the second and third cervical nerves. 122 D. A. RHINEHART From the most posterior part of the convexity of the facial nerve, and just distal to the anastomosis with the ramus auric- cularis vagi there arise from the motor part of the facial nerve two small motor nerves which are distributed to the posterior belly of the digastric and the stylohyoid muscles (figs. 11, 12, and 13G, N. to Dia., and N. to Sty-hy.). SUMMARY AND CONCLUSIONS The facial nerve of the mouse corresponds very closely to that of other mammals and man, and more closely resembles the glosso- pharyngeal than any of the other cranial nerves. It consists of two parts: one part made up of those fibers which arise inthe motor nucleus and form the motor part of the nerve, the other part being formed by the nervus intermedius and its peripheral continuation. The motor part of the facial nerve supplies the stapedius, the stylohyoid, the posterior belly of the digastric, the auricular muscles, and the superficial muscles of the face including those of the vibrissae (special visceral efferent component; Herrick, ’16, p. 146). ; The nervus intermedius is composed of afferent fibers having their cell bodies in the ganglion geniculi, and efferent fibers which have no connection with the cells of the ganglion. The ganglion itself is of the cerebrospinal type of ganglia and is composed of unipolar ganglion cells. The nervus intermedius has three branches, the great super- ficial petrosal nerve, the chorda tympani, and the ramus cutaneus facialis. The first two contain both afferent and efferent fibers the third is composed entirely of afferent fibers. The efferent fibers which enter the great superficial petrosal nerve form a separate efferent root of the nervus intermedius. The great superficial petrosal nerve enters the sphenopalatine ganglion. Sympathetic fibers from the superior cervical sympa- thetic ganglion reach the sphenopalatine ganglion by way of the deep petrosal nerve and the nerve of the pterygoid canal, and through the nervus abducens. Because the termination of these fibers in the sphenopalatine ganglion has not been determined, the function of each set of fibers is uncertain. NERVUS FACIALIS OF ALBINO MOUSE 123 Fibers from the sphenopalatine ganglion terminate in the gland of Stenson, in the medial part, at least, of the lacrimal gland, in the glands of the nasal septum, and along the blood- vessels of the nose and in the palate (general visceral efferent ‘component). If fibers of the great superficial petrosal nerve supply the taste-buds of the palate, which, in the mouse is by no means certain, they probably pass through the sphenopalatine ganglion without connection with the ganglion cells. Thenerve supply of the glands of the palate could not be determined. The afferent fibers in the chorda tympani carry taste impulses from the anterior part of the tongue (special visceral afferent component). The efferent fibers of the chorda tympani termi- nate in the ganglia in connection with the submaxillary and sub- lingual salivary glands, each fiber ending in relation with more than one ganglion cell. There is also some clinical evidence that the chorda tympani contains afferent fibers which carry impulses of certain kinds of common sensation from the tongue. The afferent fibers of the ramus cutaneus facialis terminate in the skin of the external auditory meatus, in the skin of a part of the auricule and possible a part of the tympanic membrane (gen- eral somatic afferent component). Fibers from the ramus auric- ularis vagi are distributed as a part of this nerve, and cutaneous fibers from the facial nerve are distributed through the ramus auricularis vagi. There are several unsolved problems in connection with the anatomy and function of the mammalian facial nerve which have been suggested by this work. Among these may be mentioned: 1) The presence and the termination of cutaneous fibers in the facial nerve of larger mammals and man. 2) If such be found present, the central connections of the cutaneous fibers. 3) The termination of the afferent and efferent facial fibers and the sympathetic fibers of the nervus canalis pterygoidei in the sphenopalatine ganglion and the function of each set of fibers 4) The nerves within the tongue, especially the termination of the fibers and the significance of the large numbers of ganglia and ganglion cells along the nerves. It is hoped that in the near future the time and the material will be available for the working out of some of these problems. 124 D. A. RHINEHART - BIBLIOGRAPHY AMABILINO, B. 1898 Sui rapporti del ganglio geniculato con la corda del tim- pano e col faciale. Il Pisani, T. 19. Abstract in Jahresbr. tiber d. Fortschr. d. Neur., 1898, p. 219. Cocuity, G. E. 1901 The rami of the fifth nerve in amphibia. Jour. Comp. Neur., vol. 11, pp. 48-60. 1902 The cranial nerves of Amblystoma tigrinum. Jour. Comp. Neur., vol. 12, pp. 205-289. CusHinG, Harvey 1903 The taste-buds and their independence of the nervus trigeminus. Johns Hopkins Hosp. Bull., vel. 14. pp. 71-78. 1904 The sensory distribution of the fifth cranial nerve. Johns Hop- kins Hosp. Bull., vol. 15, p. 213. De Gaerant, L. 1906 Del Nervo intermediario do Tienes e della corda del timpano. Le Nevraxe, T. 8. Abstract in Zentralbl. f. nor. u. path. Anat., Bd. 4, 8. 184, 1907. Herrick, C. Jupson 1899 The cranial and first spinal nerves of Menidia; a contribution upon the nerve components of the bony fishes. Jour. Comp. Neur., vol. 9, pp. 153-455. 1900 A contribution upon the cranial nerves of the codfish. Jour. Comp. Neur., vol. 10, pp. 265-316. 1901 The seantal nerves and cutaneous sense organs of the North American siluroid fishes. Jour. Comp.-Neur., vol. 11, pp. 177-249. 1914 The medulla oblongata of larval Amblystoma. Jour. Comp. Neur., vol. 24, p. 361. 1916 An introduction to neyrology. Philadelphia, Pa. Houpeir, G. Cart 1901 Observations on the innervation of the sublingual and submaxillary glands. Jour. Exp. Med., vol. 1, p. 281. Huser, G. Cart, and Guitp, Stacy R. 1913a Observations on the peripheral distribution of the nervus terminalis in mammalia. Anat. Ree., vol. (Ds 258" 1913 b Observations on the histogenesis of protoplasmic processes and of collaterals, terminating in end bulbs, of the neurones of the peripheral sensory ganglia. Anat. Rec., vol. 7, p. 331. Hunt, J. Ramsty 1915 The sensory field of the facial nerve; a further contri- bution to the symptomatology of the geniculate ganglion. Brain, vol. 38, p. 418. Jounston, J. B. 1908 Additional notes on the cranial nerves of petromyzonts. Jour. Comp. Neur., vol 18, p. 369. Kocu, 8. L. 1916 The structure of the third, fourth, fifth, sixth, ninth, eleventh, and twelfth cranial nerves. Jour. Comp. Neur., vol. 26, pp. 541-552. KounstamMM, Oscar 1902 Der Nucleus salivatorius chorda tympani. Anat. Anz., Bd. 21, S. 362-363. Lancuiey, J.N. 1890 On the physiology of the salivary secretion. Jour. Phys., Vole ap eli2a: Lenuossték, M. von 1894 Das Ganglion geniculi nervi facialis und seine Ver- bindungen. Beitrige zur Histologie des Nervensystems und der Sinnesorgane. Wiesbaden. NERVUS FACIALIS OF ALBINO MOUSE 25 Norris, H. W. 1908 The cranial nerves of Amphiuma means. Jour. Comp. Neur., vol. 18, pp. 527-568. 1913 The cranial nerves of Siren lacertina. Jour. Morph., vol. 24, p. 285. Prersot, Geo. A. 1913 Human anatomy. Philadelphia, Pa. Prenzo, R. 1893 Uber das Ganglion geniculi und die mit demselben zusammen- hiingenden Nerven. Anat. Anz., vol. 8, p. 738. Ranson, 8. W. 1912 The structure of the spinal ganglia and the spinal nerves. Jour. Comp. Neur., vol. 22, pp. 159-177. Rerzius, G. 1880 Untersuchungen iiber die Nervenzellen der cerebrospinal Ganglien und den iibrigen peripherischen Kopfganglien mit besonderer Riicksicht auf die Zellenausliufer. Arch. f: Anat. u. Phys., Anat. Abt., S. 369. SHEevLpon, R. E. 1909 The phylogeny of the facial nerve and the chorda tym- pani. Anat. Rec., vol. 3, p. 593. Van GruucHTEN, A. 1898 Recherches sur l’origine réelle des nervs craniens. Jour. de Neur., T. 9. 1900 Recherches sur la terminateon centrale des nerfs sensibles peri- pheriques. 1, Le nerf intermédiaire de Wrisberg. Le Nevraxe, T. 1. 1906 Anatomie du systéme nerveux de l’homme. 4th edition, Lou- vain. Vincent, 8S. B. 1913 The tactile hair of the white rat. Jour. Comp. Neur., vol 23, pp. 1-39. Weiener, K. 1905 Uber den Verlauf des Nervus intermedius. Anat. Hefte, Bd. 29, S. 97-163. Witson, J. G. 1907 The nerves and nerve endings in the membrana tympani. Jour. Comp. Neur., vol. 17, p. 459. 1910-11 The nerves and nerve endings in the membrana tympani of man. Am. Jour. Anat., vol. 11, p. 101. Yaaita, K. 1910 Experimentelle Untersuchungen iiber den Ursprung des Ner- vus facialis. Anat. Anz., Bd. 37, S. 195. 1914 Einige Experimente an dem Nervus petrosus superficialis major sur Bestimmung des Ursprungsgebietes des Nerven. Folia Neurobio- logica, Bd. 8, S. 361-382. Yaarra, K., anv Hayama, 8S. 1909 UberdasSpeichelsekretionzentrum. Neurol. Centralbl., Bd. 28, S, 738-753. Resumido por el autor, Kiyoyasu Marui. Sobre la fina estructura de la sinapsis de la célula de Mauthner, con especial menci6n de la ‘‘red de Golgi’”’ de Bethe, los ‘‘piés nerviosos terminales”’ y la ‘‘red terminal nerviosa pericelular”’ de Held. A pesar de numerosas investigaciones la estructura de Ja sinapsis de una célula nerviosa no esté completamente aclarada y todavia se necesita una investigacién cuidadosa y exacta, basada en material adecuado. El autor ha estudiado la fina estructura de las sinapsis de la célula de Mauthner de los peces éseos, empleando diferentes métodos, e intenta explicar la natu- raleza de la ”’ red de Golgi’’ de Bethe, su relacién con las estruc- turas nerviosas pericelulares, y la estructura de los ‘‘pies ner- viosos terminales.”’? ‘También intenta dar una explicaci6n acerea de la existencia de la ‘‘red nerviosa pericelular” de Held y la teoria de los contactos o continuidad de las: neurofibrillas. Los resultados de estas investigaciones pueden resumirse del siguiente modo: 1) La red de Golgi es de naturaleza gliar. 2) La parte no medulada de las fibras nerviosas est’ envuelta por una vaina de tejido gliar cuya fina estructura se desconoce. 3) La red nerviosa pericelular terminal no es realmente una red sino una imagen producida por el tefido simultaneo de la red de Golgi. 4) La teoria de los contactos es una imposibilidad histolégica; la continuidad de las fibrillas intra y extracelulares pudo compro- barse claramente. 5). Los piés terminales no son- 6rganos especificos de contacto, sino puntos del trayecto de las fibras axOnicas en los cuales tiene lugar una disolucién de dichas fibras. 6). No ha podido observar el autor estructura reticular en los piés terminales ni en la célula de Mauthner. Translation by Dr. José Nonidez, Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, DECEMBER 9 ON THE FINER:* STRUCTURE OF THE SYNAPSE OF THE MAUTHNER CELL WITH ESPECIAL CONSID- ERATION OF THE ‘GOLGI-NET’ OF BETHE, NERVOUS TERMINAL FEET AND THE ‘NERVOUS PERICELLU- LAR TERMINAL NET’ OF HELD KIYOYASU MARUI Sendai, Japan Neurological Laboratory of the Henry Phipps Psychiatric Clinic, Johns Hopkins Hospital, Baltimore, Md. FIFTEEN FIGURES INTRODUCTORY NOTE The problem of the synapse and the transmission of stimulus from one nerve cell to another has been a topic of study for more than a century; we can, however, say in the retrospect that it was really impossible to come to any safe result on this subject before the discovery of Golgi’s method. Although recent stud- ies by many investigators by means of different methods (Apathy, Bethe, Held, Cajal, Bielschowsky) have brought forth many a valuable contribution, we are not yet fully enlightened in many respects. On the basis of their exploration by means of the Golgi method, Golgi and his pupils came to the hypothesis that the gray matter of the central nervous system contains a con- tinuous mesh-work, which consists of the telodendria of the axone - and of the initial collaterals of the cells of the first type and of the collaterals from tracts. Forel (9), Cajal (12, 13), Kélliker (9), Retzius (26), and others denied the existence of this net- work on the strength of their observations. They declared that the Golgi net is a false net-work, which in fact is a kind of felt- work made of the dense processes of neighboring nerve cells. 127 THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 1 128 KIYOYASU MARUI The embryological study of the nervous system supported this idea, inasmuch as it was believed that the growth of the axis- cylinder from the cell-body of an embryonic ganglion cell could be pursued for some distance. Waldeyer (9) summed up all the facts obtained by means of the Golgi technic in 1891 and formulated the ‘neurone theory,’ in which he took for granted that the nervous system is composed of numerous nerve ele- ments, which are independent anatomically as well as geneti- cally. This theory implied the so-called ‘contact theory; ac- cording to the latter, the telodendrion of the neurone terminates with free endings and the conduction of a stimulus from one neurone to another takes place by means of contact between the nerve endings and the nerve cells. Apathy (9), who demonstrated by his own method the neuro- fibrils in clear and sharp pictures, assumed that the neurofibrils at certain points of the central nervous system cross the border of the nerve cell and through abundant splitting and anastomoses with the neurofibrils from the adjacent nerve cells form a real net-work (‘neuropil’). Bethe (7), who by means of his molyb- den method stained the neurofibrils in the nervous system of the vertebrate, found on the surface of the ganglion cell as well as its dendrites a fine net-work of irregular configuration, which he called ‘Golgi’s net’ in honor of the discoverer. He claimed an analogy of this net structure with the ‘neuropil’ of Apathy in the invertebrate, and interpreted it as the connecting link between the ganglion cells and the nerve fibers, which come afar from other cells. He claimed to have found that on the one hand the nerve fibers go over to this pericellular net-work with their telodendria and on the other hand the intracellular neurofibrils enter this same net structure. These theories of Apathy and Bethe cannot therefore be harmonized with the ‘neurone doc- trine’ ‘n which the nerve fibers are considered as non-anastomos- ing. It is to be added here that a pericellular net-work con- tinuous with the nerve fibers was described also by Auerbach (2, 3), Semi Meyer (23) (24), and Held (17, 18). But the first two stood on the standpoint of neurone and contact theory, while _ FINER STRUCTURE OF SYNAPSE 129 the last on the basis of his investigations accepted his doctrine of double interneuronal continuity. According to Held, the pericellular ends of axis-cylinders are characterized by the loos- ening of their axospongium and the densely embedded ‘neuro- somes’ and the so constituted axis-cylinder-ends (‘Achsencylin- derendfliche’) are connected with the protoplasm of the nerve cell, which is covered by them, so closely, that there exists no contact but a continuity between the two. Moreover, Held assumed that the different axis-cylinders, which enter together in the nervous cover of a ganglion cell, do not lie isolated side by side, but are combined into a continuous net-work, which he called the pericellular nervous terminal net (perizellulire ner- vose Terminalnetze’). He believed that also by means of Golgi’s method he could obtain this ‘pericellular terminal net,’ which covers widely the cell-body as well as its processes and seemed to receive numerous axis-cylinders to its beams. When Held (’97; 17) demonstrated for the first time this silver-impregnated mesh, he identified it at first with the Golgi net, which Golgi produced with the same technic and interpreted as a neurokeratin cover of the nerve cell. Later on (17) he changed his opinion and then considered the Golgi net as not identical with his nervous net. He asserted that the ‘Golgi net,’ which was demonstrated by Golgi, Semi Meyer, and then by Bethe; was probably not of a nervous nature and denied the direct connection of this net- work with the nervous elements. On the contrary, he was of the opinion that two kinds of net-work—the Golgi net and the peri- cellular nervous terminal net—occupy the surface of the nerve cell, and that alternately. He transferred his ‘neurosome-con- glomerations’ into the mesh of the Golgi’s net and claimed to have found that these conglomerations-are connected into a net- work by minute anastomosing bridges, which extend between them, over or beneath the beams of the Golgi net. He also fig- ured the direct connection of the axis-cylinders with these neuro- some-conglomerations. As regards the formation described by Auerbach, Held assumed that he observed the same structure. Ramon y Cajal (’03, 12) summed up the results of his inves- 130 * KIYOYASU MARUI tigations by means of his own technic and still adhered to the contact theory. According to the Spanish author, Apathy’s ele- mentary grating (‘Elementargitter’) does not exist; the ‘neuropil’ of the latter is not a three dimensional net-work, but is to be regarded as a plexus composed of delicate processes of ganglion cells. We should emphasize here that Retzius (26) could not confirm the ‘Elementargitter’ even in Apathy’s own prepara- tions. In opposition to Bethe’s hypothesis of the independence and the individuality of the neurofibrils, Cajal claimed to have observed a neurofibril reticulum in the nerve cells, which is said to be especially clear at the surface of the cell and around the nucleus. The Golgi net did not appear in his new preparations; moreover, his study by the Bethe’s technic showed him a direct conjunction of the Golgi net with the ‘Fillnetz’ of Bethe, which the latter interpreted as a coagulation product. On the strength of these facts Cajal came to the conclusion that the Golgi net is nothing but. an artificial product formed by the precipitate from the lymph in Obersteiner’s pericellular space. He denied also the transition of the axone-fibrils into the intracellular fibrils; accord- ing to him, the axones end on the surface of nerve cell in the form of knobs, and there exists always a thin layer of protoplasm, which is free from neurofibrils at the edge of the cell between the knobs and the intracellular neurofibrils. He considered this fact as a new argument of the contact theory. Held’s (19, 20) ob- servations by Cajal’s method, however, brought to light two different things, namely, the reticular fibrillous structure of the knobs themselves and the continuity between the axone-fibrils and the fibril net-work of the ganglion cell. Holmgren (22) confirmed these findings of Held. Cajal (13) also figured later the reticular structure of the knobs. By means of his method, Bielschowsky (8, 9, 10) reached a conclusion similar to that of Bethe; in the greater majority of cell types he could confirm Bethe’s description of the isolated course of neurofibrils. In the other types, however, he demon- strated clearly a net-structure of the fibrils, which Bethe also ad- mitted in some types of the nerve cells. Bielschowsky could not | | | ie FINER STRUCTURE OF SYNAPSE 131 acknowledge Cajal’s supposition of a constant existence of a net- structure, and he attributed this to a fault of Cajal’s method. With regard to the central endings of axones he formed an idea that the contact concept must come from the imperfect staining of the structure referred to. Like Held, he recognized the recticu- lar structure of the knobs and also believed he could demonstrate neurofibrils entering the cell so as to constitute a continuity of the neurones. In addition to this fibril continuity Wolff empha- sized recently the existence of the plasmatic continuity between nerve fibers and nerve cells. Bielschowsky and Wolff (11) ob- served also a pericellular net-work, which resembles the ‘Golgi net.’ They considered that it is formed by the mutuai com- munication of the fibrillous and plasmatic substances of the nerve fibers and stated their opinion that it is probably a structure identical with the Golgi net. According to the accurate investigation by Held (18, 19), the Golgi net is directly continuous with the ‘Fillnetz’ and they are both of glious nature. Economo (15), Paladino (25), and others confirmed this. As I mentioned above, Held (18) supposed the existence of two kinds of pericellular net-work. Economo (16) agreed with him. But this problem is not at all altered because Heidenhain (16) described recently that he would consider the Golgi net an artefact rather than a glious element. With regard to the pericellular terminal net, he admitted so far only that occasionally there exist sling-like loops within the arborization of one and the same branch of nerve fiber. Despite many investigations, the theory of contact still opposes the doctrine of the neurofibril continuity. Hedenhain (16) said that one can undoubtedly deny the latter, declaring that the figures by Held (17, 19, 20), Holmgren (22), Bielschowsky (9), and Wolff (27) prove nothing at all and that with others (Cajal (12, 13), Dogiel, (14), Retzius (27), Lenhossek (cited in 16) ) he sought in vain the continuity of nerve fibers and the cells in question. In 1915 Bartelmez (4) studied the Mauthner cell of teleosts, which had been investigated by Mayser' and Becarri (5), and 1 Zeitschrift f. wissenschaft. Zoologie, Bd. 36, 1882. 12 KIYOYASU MARUI he further analyzed the elements of the peculiar synapse of this giant cell. Lately I had the opportunity to do some experimental pathological study on this wonderful cell and its synapse, the results of which will appear soon. Careful and thorough investi- gations led me, however, to many interesting conclusions in re- gard to the structure of the synapse and the mode of conjunction of nerve cells; I will therefore describe these results in the present paper with special considerations of the condition in higher ver- tebrates. I beg to express here my gratitude to Dr. Adolf Meyer for his constant help and frequent advice in this work. MATERIAL AND METHODS OF STUDY The present work is based upon the investigation of serial sections of brains of adult Ameiurus nebulosus and Carassius auratus. The fish were decapitated, bled, and the brains were dissected out carefully but quickly, and placed promptly in different fixatives; the methods of preparation used are as follows: At first I mention the toluidin-blue preparation, in which brains were fixed in 95 per cent alcohol for twelve to twenty-four hours. Par- affin sections 8 to 10 uw thick were stained in 1 per cent warm aqueous solution of toluidin blue, differentiated in alcohol and sometimes coun- terstained with eosin. ‘ Other brains were fixed in formol-Zenker fluid twelve to twenty-four hours, cut at 8 yw in paraffin, stained in a saturated solution of thionin in a 1 per cent aqueous solution of carbolic acid and counterstained with eosin after differentiation in alcohol. For Cajal preparations brains were placed twelve to twenty-four hours in alkaline alcohol (96 per cent alcohol with 1 per cent ammonium hydroxid). The fixed brains were then rinsed with 95 per cent alcohol, placed in distilled water till they sank to the bottom of the container and then they were placed in the silver-bath of 37° to 40°C. for ten to fourteen days, according to the size of the pieces. As silver-bath I used a 1 per cent silver-nitrate solution for the first seven to ten days, with one change, and then a 2 per cent solution with one change, in which the pieces were kept three to four days. They were then rinsed with distilled water and developed twelve to twenty-four hours in 1 per cent pyrogallic acid solution in 5 per cent formalin, again rinsed with water, dehydrated and cut in paraffin 5 to 10 w thick. This al- ways gave good results; in my experience the Mauthner cells are very apt to show shrinkage space after fixing in acetic alcohol for a short time (Bartelmez) (4), which is unfavorable for the study of the synapse. I next applied the Levaditi method for spirochaeta pallida to the fish brain. The brains were fixed in 10 per cent formol twelve to FINER STRUCTURE OF SYNAPSE 133 twenty-four hours, rinsed with water, placed in 95 per cent alcohol for twenty-four hours, and then placed in distilled water, till they sank to the bottom and they were then put into the silver-bath for five to seven days (in a 1 per cent silver solution for three to four days and in a 2 per cent solution for two to three days). For the reduction of the impregnated silver a 2 per cent pyrogallic acid solution in 5 per cent formol was used. ‘The sections were cut at 5 to 10 yu. This preparation offered excellent pictures of the synapses and led to many interesting findings, which will be described later. The Bielschowsky method of silver reduction also yielded nice pic- tures of the intracellular neurofibrils and thesynapse. The brains were fixed in 10 per cent formol for twenty-four hours, rinsed with water, dehydrated and cut in paraffin at 5 to 10 uw. The paraffin was removed with xylol and alcohol and the sections were thoroughly washed with distilled water and placed in a 2 per cent silver solution for two to three days. The sections were treated on the slide, following exactly the directions of Bielschowsky. Some of the series were counter- stained with eosin. For Heidenhain preparations, brains were fixed in Zenker’s fluid or in formol-Zenkr fluid. Some brains were put into the latter, after being fixed first in 10 per cent formol for twelve to twenty-four hours. The sections of 5 to 8 » were stained in the iron-hematoxylin of Hei- denhain. This method gave not only a clear picture of the cell-body and the synapse, but also a blue stain of the myelin sheaths that was sometimes very advantageous. In addition to these preparations I prepared also a few series with Held’s neuroglia. stain, Weigert’s stain, and Mallory’s stain. In all I studied ninety-seven series of normal fish brains in the different methods, classified as follows: (1). 5 series of Ameiurus ee Fixed in 95 per cent alcohol, and stained 5 series of Carassius auratus { with toluidin-blue and sometimes eosin. (2) 5 series of Ameiurus nebulosus | Fixed in formol-Zenker fluid, stained with 5 series of Carassius auratus | thionin-eosin. (3) 5 series of Ameiurus nebulosus 11 series of Carassius auratus J \ Cajal preparation. (4) 14 series of Ameiurus pees i : Levaditi’ : 15 series of Carassius auratus eae eethod (5) 10 series of Ameiurus nebulosus 11 series of Carassius auratus (6) 5 series of Ameiurus nebulosus 4 series of Carassius auratus (7) 2 series of Carassius auratus Held’s neuroglia stain. (8) 2 series of Ameiurus nebulosus Mallory’s stain. (9) 2series of Ameiurus nebulosus Weigert’s stain of myelin sheath. Bielschowsky’s stain. Heidenhain preparation. 134 KIYOYASU MARUI NEUROFIBRIL STRUCTURE OF THE MAUTHNER CELL It is not the purpose of the present work to study the posi- tion, general relations, and size of this wonderful cell. My main purpose lies, on the contrary, in the further investigation of the finest structure of this giant cell. I merely refer here to the works of Mayser, Beccari (5), Herrick,? Bartelmez (4), and others in regard to those points. As far as the internal morphol- ogy of the Mauthner cell is concerned, it would not do to ignore the condition of the intracellular neurofibrils in the study of synapse, since, as I mentioned above, the relation between extra- and intracellular neurofibrils is not clearly decided yet. Though Bartelmez and others described the neurofibril structure of this cell, I will add the results of my study here, as there are many points not sufficiently clear. The Cajal, Levaditi, and Bielschowsky preparations were chiefly used in my study of the neurofibril structure of the Mauthner cell. By means of the first two methods the neuro- fibrils were exceedingly clear in many cases, but in other cases they proved very uncertain in,their results. The cell was now and then stained merely a diffuse yellow or brown, without any neurofibrils being differentiated in it, and in other cases the ground substance of the cell was also tinged more or less inten- sive yellow or brown, so that the neurofibril structure did not present itself clearly enough. In comparison with them, the Bielschowsky technic offered always excellent results; I shall therefore speak mainly of the Bielschowsky preparations, besides giving consideration to the best preparations of Cajal and Le- vaditi. Owing to the wealth of neurofibrils in the Mauthner cell, the sections must be thin in order that the course of the in- dividual neurofibril could be followed distinctly. The results of my investigation on the neurofibril structure in Mauthner’s cell harmonize in main features with those of Bethe (6, 7), Bielschowsky (8, 9, 10), and Economo (15), in so far as the neurofibrils do not form a real net-work in the cell. On the 2 Journal of Comparative Neurology, vol. 24, 1914, p. 343. FINER STRUCTURE OF SYNAPSE LS) strength of his study by means of his method, Bethe came to the conclusion that the neurofibrils pass generally through the cell- body without any net-formation. With regard to this, how- ever, the cells of the Ammon’s horn and the Purkinje’s cell re- mained uncertain to him. As far as the cells of the spinal gan- glia and of the lobus electricus of Torpedo are concerned, he was quite sure that there exists a net-work in them. According to Bielschowsky, who worked with his method, the neurofibril structure corresponds in essential features to that of Bethe. Eeconomo also demonstrated neurofibrils which do not anasto- mose with the other fibrils. Like Bethe, he found also fibrils which run isolated from one branch of one dendrite to another and are not continuous with those in the cell-body. As figure 1 shows, the neurofibrils run straight or winding, but smoothly through the cell-body and, generally speaking, par- allel to the axis of Mauthner’s cell, without any anastomoses for a long distance. Some of the neurobrils can even be followed through the whole length of Mauthner’s cell, without any bifur- cation of neurofibrils. The latter occurs, after all, only now and then, and this strongly suggests that the neurofibrils do not form a real mesh-work in the cell. On this point I certainly agree with Bethe (6), who argued that bifurcation would appear oftener, if there were really a net-work in the cells. Hconomo (15) suggested that there probably occurred merely a sticking together of neurofibrils, which lie side by side in the cells. In the neighborhood of the nucleus the course of the fibrils is slightly more irregular, but I could not observe any anastomosis between the neurofibrils. At the starting point of the cell processes the fibrils radiate forth into the cell-body and at this point certainly nothing of a net structure or even of a branching of single fibrils could be noticed, which, with Economo (15), we would here ex- pect in a real net-structure. From the processes and from the interior of the cell the neurofibrils stream into the axone hillock, and they could be followed pretty far into the latter. All these findings speak no doubt against the existence of a neurofibril net-work in Mauthner’s cell. 136 KIYOYASU MARUI Raméon y Cajal (12) emphasized on the basis of his study by his own technic a reticular structure of the neurofibrils in the gan- glion cells. He distinguished two kinds of fibrils—the primary and the secondary neurofibrils; described the existence of a super- ficial and a deep neurofibril net-work, and he stated that the parallel fibrils of the dendrites enter into correspondence with both the net-works. He declared, moreover, that the fibrils of the axis-cylinder originate from both these reticula. Retzius (26), v. Lenhossék, and Held (19, 20) produced by means of Cajal’s method similar results as Cajal, while others (Marinscu, v. Ge- huchten, and others cited in 15) maintained a certain reserve on this point, at least for some cell types. Bielschowsky (8), who studied the motor cells of the anterior horn by Cajal’s method, pointed out the disadvantages of the latter, through which a false picture of a reticulum might be brought out. Wolff (28) expressed an opinion concerning the spatial relation be- tween the honeycomb of the protoplasm and the neurofibrils and declared that the neurofibrils always lie in the wall of.the former. According to Economo, the neurofibrils through simul- taneous impregnation of this’wall of the honeycomb pattern may look as if they were connected by cross-beams so as to form a net-work. — As far as the anastomotic connections between the neurofibrils are concerned, Wolff (28) did not want to speak of a real net- work; on the contrary, he assumed theoretically a plexus forma- ‘tion of the neurofibrils. Heidenhain also, on the basis of the theoretical considerations, assumed a metaplexus in the cell. In the present work, however, I will not enter further into theoreti- cal considerations, but restrict myself merely to the microscop- ically visible things. I came to the conclusion that the neuro- fibrils in Mauthner’s cell form no net-work at all. Above all I wish to emphasize here that I could not observe any neurofibril reticulum in this cell even in my best preparations with Cajal’s and Levaditi’s methods (fig. 2). FINER STRUCTURE OF SYNAPSE 17 THE ‘GOLGI NET’ OF BETHE IN THE SYNAPSE OF MAUTHNER’S CELL : Careful investigation of the synapse of this giant cell by means of different methods revealed a most characteristic net-work, which covers the cell-body as well as its processes like a basket and also fills the ‘axone cap’ (figs. 3 and 4). There is no doubt that this net-work is identified with the structure, which was described for the first time by Golgi (’93) and later called the Golgi net by Bethe. This net-work was constantly demonstrated in Mauthner’s cell and was most beautiful in the Levaditi preparations; as far as I know this technic has never been applied before for the study of this structure. In Ameiurus brain the net-work appeared in a brown or dark color, whereas in Carassius it was stained in a brown to yellow color. As another advantage of this technic it must be emphasized that the nerve fibers in the synapse were also demonstrated as clean cut: dark brown fibers simultaneously ‘with the Golgi net, especially clear in Carassius. In the latter the Golgi net substance appeared mostly yellow and the con- trast with the stain of the nerve fibers was so striking that it threw very good light on the problem of the mutual relation be- tween both structures, which had been discussed by different authors and still had remained undecided. In both Ameiurus and Carassius the ‘Fiillnetz’ of Bethe was also stained yellow in this method. In the Heidenhain preparations the same structures were also demonstrated, especially clearly in formol-fixed material; but here the Golgi net could only be observed around the cell and in the ‘axone cap.’ On the surface of the Mauthner cell I could hardly see this structure, owing to the fact that the cell-body and the Golgi net appeared in too closely similar color (fig. 5). In the Heidenhain preparations (fig. 7) and the thionin-eosin preparations of formol-Zenker material one could hardly recog- nize the Golgi net-work, unless he has impressed the picture upon him in the above-mentioned preparations. In the Cajal preparations the Golgi net and the ‘Fiillnetz’ were not observable. 138 KIYOYASU MARUI I will first describe my findings in the Levaditi preparations. On the dendrites and on the part of cell-body which is not covy- ered by the ‘axone cap’ the Golgi net is a single layer; that is to say, one layer of a net surrounds the cell and the dendrites. Each nodal point of this net-work is connected with the cell sur- face by means of a beam, which is attached to the cell surface perpendicularly, with a slightly extended basis (fig. 4): The ‘axone cap’ appears in a triangular shape in Ameiurus sections and is filled with more or less numerous layers of the Golgi net; the nodal points of the net layer, which lies closest to the cell surface, are connected with the latter by a Golgi net beam each. The meshes of the net-work are irregular and of variable size; but, generally speaking, the mesh is smaller the nearer we come to the cell. In Carassius the ‘axone cap’ has a round shape and here it shows a dense conglomeration of the Golgi net substance. I believe I can compare this heap of the Golgi net substance in the ‘axone cap’ with that conglomeration of the Golgi net structure which Bethe (7) found on the Purkinje cell of the cerebellum and in other parts of the central nervous system. Most meshes are five or six cornered, but there are many four-cornered and some seven- and eight-cornered ones. . At the nodal point of the mesh-work three beams join together as a rule; the nodal points, at which four mesh beams unite together, arerarer. The nodal points of the mesh-work in both the ‘axone cap’ and on the cell surface are a little thickened, and in my preparations of Caras- sius brain some of them contrast clearly as deeply brown stained round points with the yellow or light brown impregnated Golgi net beams. These points might well be interpreted as the cross- sections of the nerve fibers in. the synapse, as is to be described more precisely below. Unlike Bethe (7), who denied the direct connection between the Golgi net and the ‘Fillnetz,’ I could find the direct continuity of both the net-works, as has already been claimed by Held (18, 21), Economo (15), and others. The ‘Full- netz’ pervades the whole gray and white matter and fills the space between the myelin sheaths, surrounding the latter with its net- work. At the nodal points and also in the beams of this mesh- _—, FINER STRUCTURE OF SYNAPSE 139 work we find in this preparation brown- or black-stained points of different caliber, which are certainly the cross-sections of nerve fibers. The ‘Fiillnetz’ is less sharply marked, looks lighter than the Golgi net; the mesh itself is larger and more irregular and the mesh beams are thicker than those of the Golgi net-work. Contrary to the statement of Bethe (7), I could observe very distinctly that the Golgi net beams are connected with glia nuclei by means of somewhat extended bases, and also with the walls of capillaries. In the Heidenhain preparations of formol material I could confirm similar relations of the Golgi net and the ‘Fillnetz,’ as described above (fig. 5). The nodal points of the net-work also were found a little thickened here and there. The only thing which is different is that the nervous elements here are stained in a color similar to that of the Golgi net, and the distinction between the nervous elements and the Golgi net becomes natu- rally difficult. In both the above-mentioned preparations I observed indis- putably the direct transition of neurites into the Golgi net beams, as is shown very clearly in figures 5, 6, and 11, although I do not mean by that at all that the Golgi net is of nervous nature. To this question of the nature of the Golgi net and its relation to the nerve fibers I will return later. As already remarked, Held (18) demonstrated the net-like formation on the different kinds of the central ganglion cells, after Golgi had. described it for the first time. Held at first identified his net-like formation with Golgi’s net-work and char- acterized it as a dense net-work with coarse nodal points, formed by the fusion of arborized nerve fibers in the gray matter. He therefore described it as the ‘pericellular nervous terminal net.’ Besides Held, the net-like framework of the ganglion cell was also described by Semi Meyer (23, 24), Auerbach (2), and Bethe (7). Auerbach described the terminal nervous net around the ganglion cell on ground of his own method. Bethe, who demon- strated his Golgi net with the help of his molybden technic, con- cluded in harmony with Held, Semi Meyer, and Auerbach that 140 KIYOYASU MARUI it is probably of nervous nature, because he believed to have found in the first place that the ends of the neurites are directly connected with the Golgi net and in the second place that the neurofibrils of the ganglion cells are combined with the nodal points of the Golgi net-work. But Held (18) raised the ques- tion whether the different authors have really seen the same structure. He said: ‘‘There is no doubt that certain net-struc- tures of the Golgi preparations and the methylene-blue picture of Semi Meyer and Bethe’s net-work are identical. What char- acterizes them is above all a certain monotony of the net figure and the stiffness of the net beams and further the nodal points of the mesh are generally not considerably thicker than the beams themselves.’”? He then declared that Semi Meyer had — seen something different, because Meyer’s figures offered no proof of the nervous nature of his net-work, as it was stained isolated and without any relation to the nerve fibers. As regards Bethe’s opinion, Held assumed that neither Golgi nor Bethe had fur- nished the necessary evidence. According to him, the statement of Golgi? that the neurofibrils of the gray matter unite with the net-work around the nerve cells proves nothing, because in Gol- gi’s method the impregnation can jump from the branching nerve fibers to the thoroughly disconnected net structure and create a false connection between them. On account of the same reason, Held (18) denied the conclusiveness of his own Golgi prepara- tions, in which the close relation between the nerve endings and the net-work was distinctly visible (fig. 8, table 13, 1902). Nor did the figures and the statement of Bethe about the transition of neurites into Golgi net beams seem trustworthy to him, al- though he was not able to give any special reasons for this. As far as the findings of Auerbach are concerned, Held thought that Auerbach had observed a similar structure and had inter- preted it essentially alike. On the basis of these considerations, Held (18) came to the conclusion that on the surface of certain ganglion cells of vertebrates there exist two different kinds of net- 8 Jahresbericht v. Merkel u. Bonnet, 1894. Cited in Held’s 1902 (18). FINER STRUCTURE OF SYNAPSE 141 work—the Golgi net and the nervous pericellular terminal net— which are quite distinct in their relation to other elements of gray substance and in their functions. Held furnished as an argument for this hypothesis the fact that the nodal points of each net-work are distinct from each other. He figured round or star-shaped formations in the mesh of the Golgi net, connected among each other by delicate threads, which lay above or beneath the beams of Golgi net. He further claimed, pointing to the figures and statement of Bethe (7), that some round or oval specks, which are visible here and there as the contents of some meshes, might well be interpreted as the residue of discoloration of those star-shaped formations, which remained stained in his own preparations. Held (17) stated formerly that the surface of the ganglion cells are covered by the variably densely scattered and variably large protoplasm pieces, which have granular structure and are connected among each other-with fine threads so as to form a net-work. At that time he spoke of them as the neurosome conglomerations on ac- count of their thick granulations, and he identified them with the nodal points of his nervous terminal net, which he demonstrated by means of the Golgi method. Later he (18) identified these neurosome conglomerations with the contents of the Golgi net meshes referred to. Bethe (7) tried to argue that the neurosome conglomerations of Held must be regarded as the broken prod- ucts of his Golgi net. Against this opinion, however, Held (18) furnished the proof that his neurosome conglomerations could be pretty clearly demonstrated in sections, which were prepared for Bethe’s molybden method II by means of other stains, whereas in the alternating sections the intact Golgi net could be demonstrated by Bethe’s stain. In the following I wish to describe my results in Mauthner’s cell in comparison with the findings of other authors in other cells. My observations led me to many interesting conclusions, which could not be brought into conformity with the findings of other authors without herein wanting to generalize those re- sults in the fish brain immediately. In the first place, attention. 142 KIYOYASU MARUI must be called to the fact that in my Levaditi preparations and Heidenhain preparations of formalin material the meshes of the Golgi net were free from any formation or granule, which could be homologized with the neurosome conglomeration of Held. As the figures (6 and 11), produced from a Levaditi preparation of Carassius, show, the cell-body is covered by a net-work, whose nodal points are considerably thicker than the beams and appear as deeply brown-stained spheroidal points. This net-work passes directly into the net structure of the ‘axone cap,’ which is itself continuous with the ‘Fiillnetz’ of Bethe. There can be no doubt that this net-work is the Golgi net. The meshes of this net-work are free from any further material, as the figures clearly indicate. In the axone cap, too, we cannot find any content in the mesh of the net-work. Above all, stress must be laid in these figures (6 and 11) upon the fact, that numerous nerve fibers, which have lost their myelin sheaths at the border of the ‘axone cap,’ run toward the surface of the cell, always keeping their place along the beams of the net-work. Moreover, we find the nodal points of the net-work here and there showing round brown and thick points, which could well be interpreted as the cross-sections of the nerve fibers, running along the beams of the net-work. Now and then we find a cross-section of a nerve fiber, which looks as though it lay isolated in the mesh of the Golgi net, but through careful observation with continual rota- tion of micrometer screw we realize that it is connected with the neighboring nodal points of the net-work by means of delicate yellow threads. On the surface of the Mauthner cell we also, find occasionally in the mesh of the Golgi net a certain granule, which might perhaps be homologized with the neurosome con- glomeration of Held. This granule, which is seemingly an iso- lated round particle, but in reality a star-shaped structure with radiating beams, does not lie in the mesh of the net-work, but is » found in a level different from that of the neighboring net beams, and with careful observation it is, at least at times, found to be connected with the Golgi net by means of the radiating beams. In the Heidenhain preparations of formol material a similar a FINER STRUCTURE OF SYNAPSE 143 condition of the synapse was in evidence and the close relation between the nerve fibers and the Golgi net-work was very clear, as is shown in figure 5. In the Heidenhain preparations of formol-Zenker material the synapse appears denser than in the previous preparations in sections of the same thickness, but one can after some practice easily find the similarity of the net-work in the cell circumference and in the ‘axone cap.’ In figure 7, produced from this preparation, we observe here and there in the meshes of the Golgi net around.the Mauthner cell a star- shaped granule which is found oftener here than in the previous - preparations. It is obvious that these granules correspond to Held’s neurosome conglomerations, which he located in the mesh of the Golgi net and interpreted as the nodal points of his ‘peri- cellular terminal net.’ As was noted in previous preparations, these granules do not lie exactly in the mesh of the Golgi net, so far as my observations go. We cannot observe both these granules and the surrounding Golgi net beams with equal distinctness in one and the same focus of the microscope. By careful observation with slight rotation of the micrometer screw we could ascertain here as well that these granules are connected by means of their radiating threads with neighboring nodal points of the Golgi net. Al though I noted also, as Held (18) did, that now and then the radiating threads do not pass over the neighboring nodal points, but reach the adjacent granules, passing either over or beneath the net beams, ‘this should not be wondered at in the three- dimensional net-work. On the ground of these considerations, I am inclined to assume that these granules correspond to nodal points of the Golgi net, which lie in different levels and give the false idea that they are within the mesh of the Golgi net. Even in the thionin-eosin preparations a similar structure was recog- nizable, although it is much more difficult to perceive than in the previous preparations. Held (18) figured the mutual relation between the neurosome conglomerations; but it is not indicated clearly enough in his figures (7 a, b, 11, 12, 1902). His Golgi figures (3, 4) showed the THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL, 30, NO. 1 144 KIYOYASU MARUI situation more clearly; but his interpretation seems to be wrong in so far as he wanted to explain these figures as his ‘nervous pericellular terminal net,’ which according to his opinion les alternating with the Golgi net on the surface of the nerve cell. I shall come back to the argument on this point later. As men- tioned, Held (18) denied Bethe’s opinion concerning the direct transition of the neurites into the Golgi net; also Economo (15) figured such a picture (table 3, fig. 25), but he remained in agree- ment with Held. I have already remarked that my prepara- tions (figs. 5, 6, and 11) showed indisputably the close relation between the Golgi net and the neurites; I could even demonstrate pretty often delicate deep brown impregnated nerve fibers in the beams of the Golgi net, as shown in figures 3, 6, and 11, which arrive at the surface of the celi. Bethe (7) enumerated five points in his paper which according to him speak in favor of the con- nection of neurites with the Golgi net and in favor of the nery- ous nature of the Golgi net. With regard to his second argu- ment, that the Golgi net is always dense and fine in places of the central nervous system where numerous axis-cylinders ramify, I would mention that in the axone cap of Mauthner’s cell, where abundant axone-fibers form a dense plexus, the Golgi net struc- ture offers a very thick conglomeration. As far as Bethe’s third argument goes, concerning the probability of demonstrating the neurofibrils in Golgi net beams, I can explicitly assert that I could without doubt find the nerve fibers in the latter, as I stated already. I believe I have settled this point, on which Bethe (7) was not sure himself. At the same time I must em- phasize that I could not confirm Bethe’s statement, that the neurofibrils form a net-work in the net beams; as I shall discuss in the following chapter, I do not find any net-work of the nerve fibers in the synapse of Mauthner’s cell. Although these argu- ments of Bethe, as Held (18) declared with good reasons, give merely indirect proofs for the nervous nature of the Golgi net, I think they indicate at least that the neurites have something to do with the Golgi net. I do not mean by that, however, at all, that the Golgi net is a nervous structure. Now one might FINER STRUCTURE OF SYNAPSE 145 raise objection against me that the nerve fibers, which I have demonstrated in the beams of the Golgi net, might be neuroglia fibers. Concerning this I should raise the following points: in the first place, I could not, like Bartelmez (4), find any neuroglia fiber by means of specific neuroglia methods, and in the second place, I was able to follow the nerve fibers far back to the part where they have myelin sheaths. As quoted above, the connection of nerve fibers with the Golgi net was stated by Golgi; also Held himself figured such points from his Golgi preparations (fig. 8, table 13, 1902). But he immediately denied the demonstrative power of these findings on account of the drawback of the Golgi technic mentioned above. I wonder why Held was so severe in his criticism concerning Golgi’s method, when the picture was unfavorable to his opinion, whereas he avails himself of it to its full extent, if it is convenient for him. Held (17) formerly demonstrated (figs. 5 to 8, 1897) his ‘pericellular nervous terminal net’ around a certain kind of nerve cell. Later, however, he altered his interpretation of these figures so that only his figures 7 and 8 can hold their ground even in his extended critical consideration. If one, however, com- pares Held’s figures 5 and 6 with 7 and 8, one can hardly see why in the figure 6 and in the light part of the figure 5, which showed the impregnation of the Golgi net, he should later have seen evidence of a nervous net-work. To me all these figures, espe- ‘cially 6 and 7, look very much alike; in the light part of the figure 5 the nodal points of the net-work are not thickened, but to my mind one should not be surprised at this in a method which is so inconstant in its results. Besides, it must be emphasized here that Held (17) did not rely on satisfactory evidence, when he set up his ‘pericellular nervous terminal net.’ He considered the pericellular net-work as the nervous-one merely for the reason that he found the connection between the net-work and the ar- borization of the nerve fibers and the thickening of the nodal points of the former. The relation between the Golgi net and the nerve fiber arborizations has been settled, I believe, through my present investigation. Moreover, despite the characteriza- 146 KIYOYASU MARUI tion of the Golgi net by Held, I have found a thickening of the nodal points of the Golgi net, although I would not deny the fact that the nodal points sometimes, especially in my Levaditi’s preparations of Ameiurus, were not particularly thickened, when the impregnation of the nervous elements was not suf- ficient. According to my opinion, however, this mark of dis- tinction is not of considerable value and the variation must be regarded as the consequence of variable impregnation and methods of demonstration. It would be necessary to add here that in my Heidenhain preparations the nodal points appeared always thickened. On ground of these considerations, all the figures of Held seem to be pictures of the Golgi net, which is connected with nerve fibers and possesses the thickening on its nodal points; the latter might well be regarded as the cross- section of the nerve fiber or the so-called terminal foot of Held, which I will describe and discuss in the following chapter. The figures of Held’s publication (02; 3, 4) might come under the same point of view. Anybody who would compare those figures of Held with mine (figs. 6 and 11) would note immediately that they demonstrate thoroughly, similar situations. I therefore came to the conclusion that on the surface of the Mauthner cell there exists a single net-work—the Golgi net—and that the endings of nerve fibers do not lie in the mesh of the latter, as it was supposed by Held, but the nervous elements of the synapse are related to the Golgi net and reach the cell surface at the ° same points where the Golgi net beams are attached to the cell surface. Held seemed to have been misled in his hypothesis on account of his one-sided interpretation of Golgi and neurosome preparations and his belief in the glious nature of the Golgi net. As far as the nature of the Golgi net is concerned, the inter- pretation of Golgi,t Cajal (12), and Heidenhain (16), and Wolff (27), who wanted to see in it a neurokeratin structure, an arti- fact by coagulation, or then the superficial part of the honey- comb structure of the nerve cell, must immediately be denied. 4 Cited in Plasma u. Zelle, Heidenhain, Bd. 1, a, p. 911. FINER STRUCTURE OF SYNAPSE 147 Also the hypothesis of Bethe (7) who considered it as the nervous structure, I should deny positively with Held (18, 21), Economo (15), and others. Held discovered in both the white and gray matter of the brain a glia’ reticulum which provides myelin fibers with neuroglia sheaths and he identified this with the ‘Fiillnetz’ of Bethe; he also demonstrated that the beams of the glia reticulum pass into those of the Golgi net, increasing thereby the density of its substance. Afterwards he added his further observation that the neuroglia cells accompanying the ganglion.cells form the Golgi net of the latter. Economo found the direct connection between the Golgi net beams and the gla nucleus by pediform appendages. All these findings I could es- tablish in the synapse of Mauthner’s cell; there is no doubt now that the Golgi net is of glious nature; it must be added here that also Schiefferdecker,? Paladino (26), Besta,® and Alzheimer admitted its glious nature. What, then, is the relation between the nervous elements and the Golgi net structure? As I remarked already, the myelin fibers are enclosed in glia sheaths of, net-like structure, which is con- tinuous with the diffuse glia reticulum. As far as I know, the relation of the unmedullated part of the nerve fibers to the glia tissue is not known very clearly. Held (19) said that, only oc- casionally and rather unevenly by means of Bethe’s molybden technic, he obtained pictures in which the unmedullated fibers are enveloped in a delicate Golgi net. He added also that these sheaths of unmedullated nerve fibers are connected with the diffuse delicate glia reticulum by means of projecting spikes. This description of Held comes pretty near to my own findings in this work (nerve fibers in the beams of the Golgi net, cross- sections of nerve fibers at the nodal points of the net-work), which indicate that the unmedullated part of the nerve fiber is also enclosed in glia sheaths. I am very sorry that I could not establish the finer structure of these glia sheaths in the present 5 Heidenhain, Plasma u. Zelle, Bd. 1, a, p. 911. 6 Riv. di. patol. nerv. e ment., T. 16, p. 604. 7 Nissl’s Arbeiten, Bd. 3. 3, p. 412-421. 148 KIYOYASU MARUI investigation; but on the basis of my findings I should be able to declare, I believe, that they accompany the neurites to their ends, connecting each other by means of delicate beams and thus forming the Golgi net in the axone cap as well as on the cell surface. Also the minute cap dendrites described by Bartel- mez (4) are enclosed in the yellow net beams of the Golgi net. Bielschowsky (18) also demonstrated the Golgi net by his method and he could even observe in harmony with Bethe (7) | that nerve fibers stream on many points into the beams of the Golgi net. Further, he and Wolff (11) expressed themselves about the relation between Bethe’s net-work and the ‘terminal nervous net,’ which they demonstrated as follows: The mesh-formation in our picture is not as regular as that which Bethe’s technic yields. Also the net beams appear in ours more deli- cate than those in Bethe’s. Still we consider it is probable that both methods demonstrate in this the identical structure. The difference might depend upon the fact that in Bethe’s technic the plasmatic component of the net-work and in ours, the fibrillous component, be- comes more manifest in the preparations. As Bethe (7) insisted, the beams of the net-work are not homogeneous but delicate fibrils are recognizable in them, which are continuous with the intracellular neurofibrils. : Now, if one compares this statement with my description, it becomes highly probable that the so-called “‘plasmatic compo- nent’ of the net-work of Bielschowsky and Wolff is to be inter- preted as the Golgi net substance. As will be described in the fol- lowing chapter, in some of my Bielschowsky preparations the sharp and intensely impregnated endingsof neurites in the synapse of Mauthner’s cell were seen connected with each other by pale stained beams forming a net-work. In the eosin counterstain of this preparation the sharp and dark stained components of the net-work were covered by more or less thick red sheaths and they were interconnected by red-tinged bridges so as to form a net-work (fig. 15). It was extremely interesting to find that the red-stained component of the net-work was directly contin- uous with the glia reticulum and with the glia nucleus. This picture is to some extent equaled by my findings in the Levaditi FINER STRUCTURE OF SYNAPSE 149 preparations. Bielschowsky and Wolff (11), Held (20), and others took for granted that the transition of the axone-fibers into the nerve cell consists of axone-plasma and fibrils. I agree with this opinion. But I believe that the component of the net-work, which appeared red in eosin counterstain, would be the Golgi net substance, to some extent at least. The statement of Bielschowsky and Wolff (11) that the simultaneous existence of a glious pericellular net-work is compatible with their previous findings, can no longer be maintained, on ground of my above- described considerations. THE NERVOUS TERMINAL FEET AND THE NERVOUS TERMINAL NET OF HELD AND THE NEUROFIBRIL CONTINUITY The connection of neurones by means of special structures (nervous terminal feet) was first discovered by Held (17, 18), confirmed by Auerbach (2), and popularized by Ramon y Cajal (12, 18). Through the further investigations of many authors by means of Cajal’s and Bielschowsky’s methods many valuable contributions were added to this interesting and important prob- lem. The question of the nervous pericellular terminal net and the neurofibril continuity also has long been the subject of dis- pute among the histologists. We are not as yet enlightened thoroughly on these questions. I shall now go over these questions on the basis of my inves- tigation on the Mauthner cell. In my Cajal preparations the nervous elements of the synapse were demonstrated almost exclusively as clean-cut deep brown fibers; the glia nuclei were the only things impregnated distinctly besides the nerve fibers. Figure 8 is reproduced from a preparation of Carassius; the lat- eral dendrite of Mauthner’s cell is enveloped here in a sheaf of unmedullated nerve fibers. Each nerve fiber is provided with a thickening, which is to be homologized with ‘bontones de Auer- bach’ of Cajal and lies more or less close to the surface of the dendrite. Besides this, some fibers have one or multiple thick- enings on their way to the cell and sometimes present the pic- ture of a string of pearls. There is no doubt, that these are iden- 150 KIYOYASU MARUI tical with the ‘bontones de trajecto’ of Cajal. Both these ‘bon- tones’ are, generally speaking, spindle-shaped or spheroidal and variously large. As far as my observations reached, there is no particular difference in structure between these two kinds of ‘bontones.’ They are now and then impregnated massively or as if punched; most of them show, however, the splitting up of the axone fiber into several delicate neurofibrils in them, as figure 8 x, indicates very distinctly. Some of them appear in the .preparations in cross-section and then they show the cross- sections of these delicate fibrils in them (fig. 8, xx). I must emphasize here that I did not find any net structure in the ‘bon- tones,’ as is described by many authors; the one, as we find in figure 8 at xxx looks as though it had net structure, but I believe it is by no means a real net-work; on the contrary, it shows merely the splitting up of the axone fiber into multiple delicate fibrils. Besides these large ‘bontones’ we find many minute rings, lying close to the surface of the dendrite, and these rings are continuous with very delicate fibers. The latter come from the other fibers as their ramifications or from the end of the large ‘bontones’ (fig. 8, xaxx}. It is quite obvious that these rings are similar to those structures which were described and figured by Cajal (13). Now, the nerve fibers, which possess the above-mentioned qualifications, pass sometimes very near the surface of .the dendrite, and thus remind us, especially in the tangential sections, of fibers coming into contact with the cell surface of the dendrite by means of their ‘bontones de trajecto,’ as was claimed by Cajal (13). But on careful observation, es- pecially on examination of profile pictures, we can easily con- vince ourselves that there is no contact between them; at least I can state definitely that there are many of these ‘bontones de trajecto,’ which lie quite remote and free from the surface of the cell (fig. 8). On the ventral dendrite of Mauthner’s cell I found a similar condition in the synapse. In the ‘axone cap’ we see abundant unmedullated nerve fibers, which, as far as my investigation went, 1orm a piexus of nerve fibers. Beccari (5) used the term ‘canes- FINER STRUCTURE OF SYNAPSE Lol tro,’ to express the condition of nerve fibers in the synapse; but he did not give any statement about the behavior of the fibers with reference to each other. We find nerve fibers of multi- farious directions and ramifications of the fibers are very often observed; but the anastomosis formation between them and even the net formation of the nerve fibers could not at all be ob- served in the ‘axone cap.’ The nerve fibers of the axone cap showed also the ‘bontones’ and the minute rings. The whole condition of the synapse is, generally speaking, similar to that of the lateral dendrite. Cajal (13) stated and figured that the endings of nerve fibers come in contact with the cell surface not only by his ‘bontones de Auerbach’ and the minute rings, but also by means of the ‘bontones de trajecto.’ His figure 9 indicates that the latter lies quite close to the cell membrane of the ganglion cell. Economo (15) claimed to have observed that the axis-cylinder enters into connection with the nerve cell by multiple terminal feet, run- ring on the surface of the ganglion cell. Heidenhain (16) de- clared further that most of the nerve cells carry a very thick pelt of end-knobs. As already described, there was no evidence in my preparations of the connection between the cell surface and the ‘bontones de trajecto.’ At least I can state that I have found many of these ‘bontones’ quite free and remote from the cell surface. The figures of Cajal (13) and Economo (15) do not prove that they are connected with the latter. I have the idea that they present merely an optical illusion, caused by the nerve fibers with those ‘bontones’ passing quite near the nerve cell. Cajal (12) at first characterized his ‘bontones’ as spheroidal or elongated structures, impregnated solidly or at the most punched. He regarded them as the foundation of his contact theory, as he believed to have found that the nerve fibers end on the cell surface with those structures. Dogiel (14) also de- scribed ring- or net-shaped terminal structures and remained a partisan of the contact theory. According to Held’s (19, 20) observation with Cajal’s method, the neurofibrils of his terminal feet form a net-work and communicate directly with the neuro- 1a KIYOYASU MARUI fibril net-work of the nerve cell in two different ways. In one group of the terminal feet the net-work is connected with the intra- cellular neurofibrils by means of a feebly stained neurofibril net- work, embedded in a homogeneous substance, and in the other group delicate neurofibrils enter the cell body from the terminal feet in a radial direction, also surrounded by homogeneous ma- terial. Holmgren (22) distinguished also two types of terminal feet; his first type was of annular shape and his second type showed a tiny net-work, which is connected directly with the neurofibril net. of the nerve cell. Later Cajal (13) also figured and described net-shaped terminal and transitional knobs and minute ring-shaped structures, which he believed to come into contact with the cell surface. Heidenhain (11) declared even that these are nothing but the net corpuscles, which appear in the periphera' nerve endings. In my Cajal preparations, how- ever, the net-like structure of these nerve endings, as remarked, did never come to my observation. Some of the terminal feet looked merely granular in my preparations, owing to a fine silver precipitate; occasionally I observed in them a false net figure caused by the irregular distribution of the latter. But the pic- tures which are shown in figure 8 exclude beyond doubt the net figure of the ‘bontones.’ I wonder if the terminal feet possess . really the net structure, as it was claimed by Held (19, 20), Holmgren (22), Cajal (13), and others. I suppose that this net figure comes from artificial admixtures of the impregnation of the honeycomb structure of the neuroplasm, which occurs now and then in Cajal’s technic. To this question I shall return later. | As far as my observations go, the terminal feet must not be regarded as the definite ends of the axone fibers, as was assumed by Cajal and others. In my Cajal preparations I observed very often that single or multiple delicate neurofibrils come from the end of the terminal feet and advance toward the cell surface, embedded in a homogeneous substance and entering the cell body (fig. 12). I am very sorry to confess, however, that I could not demonstrate definitely the neurofibril continuity in FINER STRUCTURE OF SYNAPSE 153 these preparations, as the intracellular neurofibrils were often not impregnated quite distinctly in them. So my finding corre-— sponds to that type of the terminal feet of Held, in which the fibrils were followed in radial direction from the terminal feet into the cell body; the connection by means of a fibril net-work did not come to my observation. Figure 9 demonstrates the condition of the synapse in the Bielschowsky preparation; on the surface of the axone cap, we see a number of nerve fibers going into the region of the ‘axone cap;’ Weigert and Heidenhain preparations show that they lose their myelin sheaths just at the border of the ‘axone cap.’ Some of these nerve fibers have each a large club-like expansion here and before the loss of their myelin sheaths. According to my experience, this phenomenon is more noticeable in Ameiurus than in Carassius; in regard to the significance of this finding I am not able to say anything definite as yet. On their way to the surface of the cell the unmedullated fibers have single or multiple spheroidal or spindle-shaped swellings; sometimes the fibers look like a string of pearls showing many of these swel- lings. There is no doubt that these enlargements are identical to the ‘bontones de Auerbach’ and the ‘bontones de trajecto’ of — Cajal. In my preparations these nodes came out partially dif- fusely black and partially punched as Bielschowsky (18) described them. Besides these I could find also ‘bontones’ with splitting of the nerve fibers into numerous delicate fibrils (fig. 13), just like those which were described above in my Cajal preparations. The net-work, as was described by Wolff (27) from his Bielschow- sky preparations, I could not demonstrate at allin them. In my Bielschowsky preparations I could not find any place, where the contact between the ‘bontones de traecto’ and the cell surface takes place, as was described and figured by Cajal. It is quite obvious in my figure 9 that at least there are many of those ‘bontones’ which lie quite remote and free from the cell surface. From the peripheral end of the terminal feet single or multiple delicate fibrils come out and proceed toward the cell surface, surrounded by the homogeneous substance, and enter the cell 154 KIYOYASU MARUI body communicating directly with the intracellular neurofibrils. Figure 10 demonstrates clearly that the neurofibrils from the terminal feet enter the cell body. In this figure it is also shown that the intra- and extra-cellular neurofibrils communicate with each other; one might raise the objection against me, that the fibers which enter the cell-body might be the ‘cap dendrites’ of Bartelmez (4). On this point I can give the following arguments: first, some of the fibers could be traced back to the part, where they are to be enveloped in myelin sheaths, and, second, it is connected with the cell by means of the terminal foot (fig. 10). Terminal ramification of nerve fibers occurs very often within the ‘axone cap;’ between the ends of individual nerve fibers neither anastomosis formation nor even a simple net formation of nerve fibers came to my observation in the ‘axone cap’ oron the cell surface. Through careful examination I could always isolate the sharply and intensely dark stained nerve fibers from each other. In the previous chapter I already stated that through the simul- taneous impregnation of the Golgi net substance a false picture of a nervous net-work becomes observable in the synapse. Ac- cording to my experience, the,Golgi net substance is to a certain extent antagonistic in its staining reaction to the nerve elements in the Bielschowsky method; when the nervous elements are brought out sharply and dark enough the Golgi net substance remains unstained or is stained quite feebly, whereas the latter is impregnated more or less intensely when the former is stained more feebly. I also stated above that by means of the eosin counterstain the Golgi net substance is demonstrable in red color and is connected directly with the glia reticulum and glia nucleus on the one hand and the nervous elements on the other hand. Figure 15 shows that the nervous elements are covered with a more or less thick layer of the red-stained substance; also the terminal feet are surrounded by the same substance. This pic- ture corresponds to that of the Levaditi preparation, which was described in the previous chapter. After this description of my results in both the Cajal and Bielschowsky preparations I go over the discussion of the ques- FINER STRUCTURE OF SYNAPSE 155 tions referred to. As remarked, my findings in the Cajal and Bielschowsky preparations are in opposition to those of Held (19, 20), Cajal (13), Holmgren (22), Wolff (27), and others, in so far as I demonstrated in the terminal feet merely the splitting of the nerve fiber into fine fibrils instead of a net structure. Now it is remarkable that some of the above-mentioned authors (Ca- jal, Held, and others) assume also a net structure of the intra- cellular neurofibril, while others (Wolff, 27) deny the net struc- ture in the cell body, at least in many kinds of nerve cells. I could not find any real net structure in either the Mauthner cell or in the terminal feet. It appears extremely interesting to me that in Wolff’s (27) figures Biitschli’s honeycomb structure was stained in the cell body as well as in the dendrites. Economo (15) supposed that the net figure of the terminal feet might de- pend upon the simultaneous impregnation of the Biitschli struc- ture. Also Auerbach (2), who was at first a partisan of the contact theory, declared that the neurofibrils do not form a reti- culum in the terminal feet,. but that one, two, or three delicate fibrils go radially into the cell body, embedded in the ground substance of the terminal feet. Ramon y Cajal (12, 13), Dogiel (14), Retzius (27), Heiden- hain (16),and others despite the repeated argumentation of the antagonists Held (19, 10), Holmgren (22), Bielschowsky (9), An- toni (17), Auerbach (3), and others) remained firm on the stand- point of the contact theory. According to Held, Bielschowsky, and others, however, the theory of the contact must result from the imperfect impregnation of the structure in question; they observed, as remarked, that the fibrils of the terminal feet go into the cell body and enter into relation with the cell fibrils. I could also confirm the neurofibril continuity in Mauthner’s cell in the Bielschowsky preparations. The ‘bontones de Auerbach’ of Cajal evidently must not be regarded as the contact organs, in which the nerve fibers come to their ends, but are rather tobe interpreted as the stations in the course of nerve fibers, where the modification of the substance takes place, which was claimed by Bielschowsky (8). With the supposition of the latter, how- . 156 KIYOYASU MARUI ever, who took for granted that here the dissolution of the nerve fibers occurs as a consequence of the loss of the perifibrillar cement substance, I cannot agree. On the contrary, I assume that the dissolution of the fibers is to be regarded as the result of the accumulation of the perifibrillar neuroplasm. Bartelmez (4) described in the ‘axone cap’ of Mauthner’s cell two kinds of endings—‘free endings’ and ‘knob endings,’ which latter are in contact with the cell surface; besides these he mentioned on the lateral dendrite ‘club endings.’ As far as my investigation went, the ‘free endings’ of Bartelmez are very hard to accept; though I found in my preparations nerve fibers which do not reach to the cell surface, it is always probable that they appeared in section. Nor can I find any essential difference between the endings in the ‘axone cap’ andin the lateral dendrite, as already described. Moreover, he did not state the structure of these endings in detail; he figured as ‘pericellular net’ (figs. 12, 13) a minute solid or ring-shaped structure, which is perhaps identical with the ring-shaped ending apparatus described by Cajal. Above all I must emphasize that I found the neurofi- bril continuity on the cell surface as well as on the dendrites; the ‘plasma membrane,’ which Bartelmez described on the sur- face of his club endings, must not be regarded as the last end of the nerve fibers. Held (18) assumed a ‘pericellular nervous terminal net,’ which exists on the cell surface alternating with the Golgi net, as re- marked before. In the previous chapter I showed that there is only one net-work on the cell surface formed by both the Golgi net and the nervous elements. Now, in my Cajal preparations, which do not show the Golgi net substance, I was not able to find any net-work in the synapse. I showed that Held’s figures of his neurosome and Golgi preparations do not argue for his hypothesis. It appears extremely interesting to me that in his Cajal preparations the deeply stained terminal feet are not at all (19) connected by bridges or at the most connected among each other by feebly (20) stained bridges. Moreover, it must be remembered here that in the neurosome preparations of Held FINER STRUCTURE OF SYNAPSE LST the ‘terminal feet’—the nodal points of his terminal net—were chiefly granular, whereas the beams between them appeared free from granules. Auerbach (2, 3) also assumed the existence of the nervous terminal net, whose nodal points coincide with his ‘terminal knobs;’ but his figures prove nothing. Besides, it must be emphasized that his net-work possesses a three-dimensional character not only on the cell surface, but also in certain other parts of the central nervous system—the substantia gelatinosa and the molecular layer of the cerebellum. The existence of a nervous net-work of this sort, which would penetrate the gray matter of the central nervous system, was denied by many au- thors (Cajal, (13) Kdlliker (cited in 9), Retzius (26), etc.). Even Held (20) admitted that by means of Cajal’s technic he could observe merely the indication of the extension of the ‘peri- cellular nervous terminal net.’ The figures of Holmgren (22) also prove nothing. Concerning the pericellular net-work, which is found now and then in Bielschowsky preparations, I expressed my idea before, that there might be a false picture caused by the simultaneous impregnation of the Golgi net. This consideration will perhaps be strengthened by the description and the figures of Wolff (27). He figured on the surface of the cells as well as between the latter a net structure, which shows a striking resemblance to Bethe’s net-work and is mainly stained more feebly than the terminal feet, which latter are directly continuous with the former. In the text he stated that the axone fibrils are not naked, but are enveloped in a mantle of bubble-like structure, which goes over into the honeycomb structure of the cell border. On the ground of his findings and Gegenbauer’s intercellular bridge theory, Wolff declared that Bethe’s pericellular net-work is nothing but the impregnated honeycomb structures of the cell border, the neuroplasmatic anastomoses and the perifibrillary mantles. That Bethe’s net-work is not, however, to be regarded as the honeycomb wall, but as a glious structure, I have already stated.: 158 KIYOYASU MARUI SUMMARY 1. In the ‘axone cap’ and on the surface of Mauthner’s cell a Golgi net structure was very distinctly demonstrated by means of Levaditi’s method; in Heidenhain and other preparations a similar net-work was brought out. The Golgi net is of glious nature and is in close relation to the nervous elements in the synapse. According to the results of this study, the unmedul- lated parts of the nerve fibers are also enveloped in a sheath of glious tissue; the finer structure of this glia sheath is as yet un- known. | 2. The hypothesis of Held concerning the existence of two kinds of net-work on the cell surface—a Golgi net and a ‘peri- cellular nervous terminal net’—is denied; there exists, as far as my observations go, only one net structure, which is formed by both the nervous and the glious tissues. The so-called pericel- lular nervous terminal net is not to be regarded as a real nervous net-work, but to be considered as a picture produced by the simultaneous stain of the Golgi net-work. 3. The contact theory is a histological impossibility. The terminal feet cannot be regarded as the specific contact organs, but as the points in the course of axone fibers, where the dissolu- tion of fibers takes place. The continuity of the intra- and extracellular neurofibrils is very clearly demonstrated. 4. The intracellular neurofibrils do not form any reticulum in Mauthner’s cell; nor did the net structure which was described by many authors in other animals, ever come to my observation in the nervous terminal feet. The nerve fibers showed merely a splitting into numerous delicate fibrils in them. 10 11 13 14 18 19 FINER STRUCTURE OF SYNAPSE 159 LITERATURE CITED Antoni, N. 1908 Die Frage von einer neurofibrilliren Kontinuitit im Zentralnervensystem der Wirbeltiere. Folia neuro-biologica, Bd. 2 (Sammelreferat). AUERBACH, L. 1899 Das terminale Nervennetz in seinen Beziehungen zu Ganglienzellen der Zentralorgane. Monatschift f. Psychiatrie u. Neurologie, Bd. 6. 1904 Extra sowie intrazellulire Netze nervéser Natur in den Zentral- organen von Wirbeltieren. Anatomischer Anzeiger. Bartetmez, G. W. 1915 Mauthner’s cell and the nucleus motorius teg- menti. Jour. Comp. Neur., vol. 25. Beccart 1907 Ricerche sulle cellule e fibre del Mauthner e sulle loro con- nesioni in pesci ed anfibii. Arch. Ital. Anat. e. Embr., vol. 6. Betue, A. 1898 Uber die Primitivfibrillen in den Ganglienzellen vom Menschen und anderen Wirbeltieren. Morphologische Arbeiten, Bd. 8. 1900 Uber die Neurofibrillen in den Ganglienzellen von Wirbeltieren und ihre Beziehungen zuden Golginetzen. Arch. f. mikroskop. Ana- tomie, Bd. 55. Birtscuowsky, M. 1904 Silberimpriignation der Neurofibrillen.. Journ. f. Psycholog. u. Neurolog., Bd. 3. 1905 Die histologische Seite der Neuronenlehre. Journ. f. Psycholog. u. Neurolog., Bd. 5. 1908 Uber die fibrillare Struktur der Ganglienzellen. Journ. f. Psy- cholog. u. Neurolog., Bd. 10. BretscHowsky, M., anv Wourr, M. 1904-1905 Zur Histologie der Klein- hirnrinde. Journ. f. Psycholog. u. Neurolog., Bd. 4. Casat, S. R. 1903 Consideraciones criticas sobre la teoria de A. Bethe acerca de la estructura y conexiones do las células nerviosas. Tra- jabos de Laboratorio de investigaciones biologicas de la Universidad de Madrid, T. 2, cited in Heidenhain, Held, etc. 1908 L’hypothése de Mr. Apathy sur la continuité de cellules nervenses entre elles. Anatomischer Anzeiger, Bd. 23. Doatrt, A. 8S. 1904 Uber die Nervenendigungen in den Grandryschen und Herbstschen Kérperchen im Zusammenhange mit der Frage der Neuro- nentheorie. Anatomisch. Anzeiger, Bd. 25. Economo, D. 1906 Beitriige zur normalen Anatomie der Ganglienzelle. Arch. f. Psychiatrie u Neurologie, Bd. 41. HEIDENHAIN, M. 1911 Plasma u. Zelle. Hexp, H. 1897 Beitrige zur Struktur der Nervenzellen u. ihre Fortsitze. Arch. f. Anat. u. Physiolog., Suppl., Anat. Abt. 1902 Uber den Bau der Grauen u. Weissen Substanz. Archiv f. Anat. u. Physiol., Anat. Abt. 1904 Zur weiteren Kenntnis der Nervenendfiisse und zur Struktur der Sehzellen. Abhandlungen d. mat. -phys. Kl. d. Kogl.-sichs. Ges. d. Wissenschaft, Bd. 30. THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 1 160 KIYOYASU MARUI 20 1905 Zur Kenntnis einer neurofibrilliren Kontinuitét im Centralner- vensystem der Wirbeltiere. Arch. f. Anat. u. Physiolog., Anat. Abt. 21 1903 Uber den Bau der Neuroglia und iiber die Wand der Lymphge- fisse in Haut u. Sehleimhaut. Abh. d. math.-phys. cl. d. Kégl- Sichs Ges. d. Wissen., Bd. 28. HoutmaGren, F. 1905 Uber die sog. Nervenendfiisse (Held). Jahrbiicher f. Psych. u. Neurolog., Bd. 26. 23. Meyer, 8. 1896 Uber eine Verbindungsweise-der Neuronen. Nebst Mit- teilungen iiber die Technik, und die Erfolge der Methode der subkutanen Methylenblau-injektion Arch. f. mikroskop. Anat., Bd. 47. 24 1897 Uber die Funktion der Protoplasmafortsitze der Nervenzellen. Berichte d. kégl.-siichs. ges. d. Wissen. math.-phys. Kl., Bd. 47. 25 Pauapino, G. 1913 Continuity in the vertebrate nervous system and the mutual and intimate connections between neuroglia and nerve cells and fibers. Review of Neurolog. and Psych., vol. 11. 26 Rerztus, G. 1908 The principles of the minute structures of the nervous system as revealed by recent investigations. Croonian lecture. Pro- ceedings of the Royal Society, vol. 80. 27 Wourr, M. 1904-1905 Zur Kenntnis der Heldschen Nervenendfiisse. Journ. f. Psycholog. u. Neurolog., Bd. 4. bo bo All figures are taken from preparations of Mauthner’s cell. The photomicrographs were not retouched at all. An achromatic Zeiss ocular no. 4 and a Zeiss oil-immersion ;'s were used. In figures 1 and 8 the length of bellows was 100 em. and in all the others it was 60 em. The figures 11, 12, 13, 14, and 15 were drawn using the Abbe camera lucida with slight rotation of the mi- crometerscrew. (Apochromatic Zeiss ocular no. 4, oil-immersion ;'s, tube length 20 cm.) In those drawings except 12 and 13 the glia structure is presented in a gray, and the nervous structure in a dark color. It is tobe added here that figure 7 came from a fatigue preparation. ae - ~ +> ol ae po ph _e Par - 4% Pet ee ek , Me tee Fig. 1 Carassius auratus. Bielschowsky preparation. Fig. 2 Carassius auratus. Levaditi preparation. 161 % Fig. 3 Ameiurus nebulosus. Levaditi preparation. Fig. 4 Ameiurus nebulosus. Levaditi preparation. 162 Fig. 5 Ameiurus nebulosus. Fig. 6 Carassius auratus. Heidenhain preparation (formol material). Levaditi preparation. 163 Fig. 7 Ameiurus nebulosus. Heidenhain preparation (formol-Zenker mate- rial). Tig. 8 Carassius auratus. Cajal preparation. 164 Fig. 9 Ameiurus nebulosus. Bielschowsky preparation. Fig. 10 Ameiurus nebulosus. Bielschowsky preparation. 165 12 13 Tig. 11 Carassius auratus. Levaditi preparation. lig. 12 Carassius auratus. Cajal preparation. Fig. 13 Ameiurus nebulosus. Bielschowsky preparation. 166 Fig. 14 Carassius auratus. Levaditi preparation. - Fig. 15 Carassius auratus. Bielschowsky preparation with eosin counter-stain. 167 hare ‘alll ; bal cits al Ft . AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE APPLICATION OF THE MARCHI METHOD TO THE STUDY OF THE RADIX MESENCEPHALICA TRIGEMINI IN THE GUINEA-PIG WILLIAM F. ALLEN Department of Anatomy of the University of Oregon Medical School, Portland, Oregon THIRTY-FIVE FIGURES CONTENTS RRC IGEN 92 AE tt a2 cc cha 'e's gible epsea eres aise Fee hee Rares; « ‘: eee 169 Feewioweor tue lmperapOre 1608. 6 A somewhat diagrammatic view of a sagittal section of the spinal cord (neural tube) near the middle of the body. A connection between Reissner’s fiber and the fibrils of ependymal cells is simulated. The cell in the lumen of the tube is helping to form, or to repair (probably the latter) Reissner’s fiber. 60 days. X 750. 7 Sagittal section of spinal cord near the middle of the body, showing Reissner’s fiber apparently free from the walls of the neural tube. The dorsal wall is not shown in its whole thickness. 100 days. X 750. 8 Sagittal section in the region of the sinus terminalis. X 325. 9 Posterior end of the section shown in figure 8, at a slightly different focus and more highly magnified. Several cells are seen to be in process of meta- morphosing into the posterior end of Reissner’s fiber; others, apparently, are just entering the lumen of the neural tube. 750. 10. Sagittal section through the sinus terminalis, showing the terminal plug. There is only one cell in process of metamorphosing into the Reissner fiber, and this is, apparently, in continuity with the connective tissue lying beyond the posterior end of the neural tube. X 325. ABBREVIATIONS a, anterior fbrl."’, cross-sections of fbrl.’ cl. e’end., ependymal cell 7., internal limiting membrane cl. R., cells in process of forming Reiss- _zvlr., sheath of fbrl.’ ner’s fiber nl. e’end., ependymal nucleus coag., coagulum nl., nucleus co’ms. p., posterior brain commissure’ nl. n., nucleus of nerve cell d., dorsal obt. trm., terminal plug of neural tube ex., external limiting membrane p., posterior fbr. p., fibers of posterior commissure par. n., wall of neural tube fbr. R., Reissner’s fiber sb-co’ms., sub-commissural organ fbrl., fibrils from deep (proximal) ends tis. co’nt., connective tissue of ependymal cells v., ventral fbrl.’ fibrils from superficial (distal) ends of ependymal cells 4 hes Poe elie, i Ae? A REISSNER’S FIBER IN TELEOSTS PLATE 1 HOVEY JORDAN tis.cont >< A 8) ae De i uA PAGER Oey yoo LI TIIS é ——— fe Tok. is ~Z tis.cont at ie” _forR, PES CERIRE RA) AT GLEE: Resumido por el autor, Robert Sidney Ellis. Estudio preliminar cuantitativo de las células de Purkinje en el cerebelo normal, subnormal y senfl del hombre, con algunas notas sobre la localizacién funcional. En el presente trabajo se compara el ntiimero de células de Purkinje en el cerebelo normal, subnormal y senil del hombre. El autor ha hecho correcciones para las. diferencias de volumen de los cerebelos estudiados, de modo que los valores consignados representen el ntimero de células existentes en dreas equivalentes, tomadas como unidad. Ha contado las células de seis dreas de cada mitad del cerebelo, determinando de este modo el ntiimero de células de Purkinje que existen normalmente en areas equiva- lentes del cerebelo de tamafio normal. Comparando con la norma asi establecida ha podido comprobar que el nimero de células de Purkinje es muy deficiente en el cerebelo de los idiotas-e im- béciles. Del mismo modo, en el cerebelo del viejo se observa también una disminucién progresiva en el ntimero de dichas células. Existe cierta correlacién entre estas deficiencias celu- lares y las deficiencias en la coordinacién motriz. Translation by Dr. José F. Nonidez. Columbia University AUTHOR'S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JANUARY 6 A PRELIMINARY QUANTITATIVE STUDY OF THE PURKINJE CELLS IN NORMAL, SUBNORMAL, AND SENESCENT HUMAN CEREBELLA, WITH SOME NOTES ON FUNCTIONAL LOCALIZATION ROBERT 8. ELLIS The Wistar Institute of Anatomy and Biology and The Training School at Vineland, New Jersey TWO FIGURES AND ONE CHART INTRODUCTION! In the spring of 1916, while examining the cerebellum of a ‘general paralytic, the writer was first impressed by the fact, familiar perhaps to most neuropathologists, that in this disease there is often a disintegration and disappearance of a large number of the Purkinje cells, leaving, however, the basket of fibers which normally surrounds them. Over a year later, while ex- amining the cerebellum of a microcephalic idiot, the same scarcity of Purkinje cells was observed, with the difference, however, that the section did not show the same evidence of the cells having become reduced in number by disintegration; the empty pericellular baskets were not found as in the case of paresis; it seemed, rather, that through some defect of development the normal number had never been present. This difference presented an interesting problem and it seemed worth while to make a more careful quantitative study of the Purkinje cells in different types of cerebella. In order to get 1 The writer is indebted to the staffs of the Vineland (N. J.) Training School and of St. Vincent’s Hospital (Philadelphia) and to several physicians in other hospitals for assistance in securing the material used in this study. He is also indebted to the staff of the Wistar Institute, especially Drs. Greenman, Donald- son, and Hatai, for their encouragement and assistance in numerous ways during the course of the investigation: to all of these he wishes to express his appreciation. 229 230 ROBERT S. ELLIS a fair basis for comparison, a number of cerebella were studied and the relative frequencies of cells noted. In some of the cases the cells appeared to be almost uniformly distributed and with few large spaces between them; others showed losses similar to the two cases already mentioned. Among the cerebella examined was one of a man who had died at about the age of sixty-five years after a protracted illness, and this, too, showed a distinct loss of cells. So from this preliminary set of observations it seemed clear that the number of Purkinje cells is variable under different conditions. It is well known that in paresis, in extreme old age, and in low grades of feeble-mindedness there is ordinarily a considerable degree of deficiency in motor coérdination. The question con- sequently arose, how far is it possible to find differences in the number of cells that will account, partially at least, for the observed differences in behavior? The writer’s primary interest at the time of taking up this investigation lay in the question of the anatomical basis of mental defect, and it seemed not improbable that a careful study of the Purkinje cells might throw some light on one of the most evident deficiencies found in such cases. The human motor mechanism is much more highly developed than that of lower forms, especially with reference to speech, hand movement, and the maintenance of equilibrium while standing or walking. Mental defectives generally show less motor control along these lines, and it is desirable that we know as far as possible the neural basis for such lack of codrdination. A further reason for making a study of the cerebellum in such cases is found in the fact that a number of writers, especially Tredgold (’03) and Bolton (’03, 710) in England, have em- phasized, perhaps unduly, the importance of the frontal lobe of the cerebral cortex as the area particularly affected in amentia. It accordingly seemed worth while to determine whether the brains of aments show defects in other parts, such as the cere- bellum, which is not generally associated with intelligent reactions as such. QUANTITATIVE STUDY OF THE PURKINJE CELLS 231 In order to determine the nature and extent of the variations in the Purkinje cells, and especially to determine the differences between normal and subnormal cerebella, a careful study of the numerical distribution of these cells has been made, the results of which it is the purpose of this paper to present. MATERIAL For the best results in studying the present problem it would be desirable to have cerebella in perfectly normal condition to be used as the standards with which the subnormal ones are compared. These normal cerebella should be from individuals in the prime of life, who were known to have had good physique and good motor control and who had died either from accident or from some acute disease which did not cause a disintegration of nervous elements. Pathological material is to be had in abundance, but to secure any number of approximately normal cases is next to impossible. In the present study, consequently, it has been necessary to use as our norms cerebella from the museum collection of The Wistar Institute. No claim is made that this material is ideal; it is hoped, however, that sufficient care and caution have been used, both in the procedure followed and in the interpretation of results, to prevent serious errors arising from the material employed. To obtain norms, the cerebella of three adult whites and of five adult negroes, all males (thirty to forty-two years of age), were sectioned and a preliminary study was made to determine whether or not they had suffered any loss of Purkinje cells. In nearly every case there was at least some maceration and some loss of these cells, but four of the number, three negroes and one white, showed very slight losses, and these were accordingly used as the material for determining the norm. These are the cases listed in table 1. The other four cases were rejected because microscopical examination clearly showed that cells had disintegrated and dropped out, and it would consequently have been impossible to determine from them the number of cells normally present. THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 2 232 ROBERT S. ELLIS Little or no information of importance regarding the clinical histories of these cases was available. They were simply cases that had come to autopsy and their brains had been preserved as a routine measure. Even so, however, the results from the four should give us a close approximation to the conditions present in the average cerebellum as it appears in the hospital population—a group probably somewhat below the average for the community at large in the development of the nervous system. Nine cerebella from mental defectives, who had died at the Vineland Training School, were selected from the collection in The Wistar Institute Museum and a study was made of each case. Several of these cases have been described by Goddard (Feeble-mindedness, New York, 1914), and his case numbers, preceded by ‘G,’ enclosed in parentheses, are given after the serial numbers which these brains have on the records of The Wistar Institute. The histories of these defectives, all whites, and listed in table 2 and table 4, are as follows: 14880, male, age 23 years; alcoholism and insanity in family; three miscarriages before the birth of this child; he had had convulsions, marasmus at one year, measles at four years; small in size, head small; poor motor power, poor gait, poor speech; no will power; right handed; fair memory. 15144, male, age 34 years; height, 6 ft. 3 in.; brain weight 1619 grams; supposed insanity and feeble-mindedness on father’s side; no deaths in infancy in family; had whooping-cough, measles, and pneu- monia; walked at three years; head, bullet-shaped; right side not developed; dragged feet, could thread needles and tie shoes; talked little; could not count. 15145, male, age 24 years; brain weight 1491 grams; normal (?) heredity; first child dead; had measles and whooping-cough; epileptic; rs on right hip; weak heart; normal electric reaction; grip, rt. 0-It. 0. 15214, male, age 26 years; brain weight 1065 grams; father alcoholic; five children died young; first child a ‘lunatic,’ followed by two mis- carriages; two other children mentally defective; this, the sixth child; probably congenitally luetic; spasms at nine months; walked at three years; talked at four years; had measles and whooping-cough, ab- scesses, scrofula, Pott’s disease; hunchback, T.B. of vertebrae; weak heart; was apparently normal until nine months; not strong enough to work; grip, rt. 18-It. 22; could dress self; learns quickly, forgets soon. This case is presented by itself in table 4. QUANTITATIVE STUDY OF THE PURKINJE CELLS Das 15250 (G. 285), male, age 43 years, mental age three; nervous family; epileptic, had measles and whooping-cough, brain fever at two years; side of face deformed from atrophy of jaw; stooped shoulders; left arch fallen; irregular heart; poor vision, eye fatigue and headache; no abnormal movements; exaggerated reflexes; grip, rt. 21-It. 23. 15297 (G. 203), male, age 26 years, mental age eight, apparently normal heredity; two deaths in infancy and two other children feeble- minded; physicians believed defect due tio congenital lues; did barn and garden work, grip, rt. 33—It. 31. 15299 (G. 195), male, age 36 years, mental age two; brain weight 335 grams; father probably feeble-minded; this case the first of seven conceptions; two later ones being miscarriages; had whooping-cough and scarlet fever; large head, left scapula higher than right, left testicle undescended; fallen arches; knock-kneed; did no work; grip, rt. 15- lt. 17; could not talk, but would swear freely. 15310 (G. 258), male, age 18 years, mental age four; brain weight 1051 grams; ancestry possibly neuropathic; this case Mongolian in type; Wassermann positive; knock-kneed and stoop-shouldered: died of T.B. after extreme wasting away; left testicle undescended; normal reflexes; ate well and played well; left-handed; grip about .18 for each hand; did not alternate feet in climbing stairs; speech poor; subject to headaches. 15320 (G. 327), female, age 20 years, mental age one; mentality of ancestry doubtful; had large tumor in anterior part of the right hemi- sphere of the cerebellum; began to grow very weak at eighteen years; walking became poor, and a year later had difficulty in swallowing; helpless in both hips and hands, left hand especially weak; pupil reflexes lost, other reflexes exaggerated. Subnormal infants. For purposes of comparison, the cerebella of six infants, four negroes and two whites, were studied. These cases are of the type found in charity hospitals where nearly all are below normal and where a very large percentage are illegiti- mate. Out of twenty such cases, where the brain weights were taken, only two brains weighed as much as the average for their age. Consequently, though the point could not be determined individually, we are safe in assuming that they represent a distinctly subnormal group of the population. Senescents. The cerebella of five cases of senescence were studied. Enough is known of each of these to make it fairly certain that the changes observed were due primarily to old age rather than to other causes. The age, color, and sex of these are given in table 5, which shows the results of the cell counts. Paresis. While making this study, the cerebellum from a 234 ROBERT S. ELLIS case of paresis was sectioned before the cause of death was known, and it seemed desirable to include it as an example of what may happen in this disease. It is from a white man who was of average physique and who died at the age of thirty-three years. The data are given in table 6. In the course of this investigation careful counts have been made on parts of about forty cerebella and sections from about twenty-five others have been less carefully examined. The twenty-five cases on which the figures given in this paper are based are believed to be typical of the classes under which they are listed. PREPARATION OF MATERIAL The cerebellum was removed from these brains, weighed, and the weight in grams recorded. In cases where the specific gravity had changed materially from being for several years in alcohol, the specific gravity was determined and the brain weight was corrected to the normal weight for a cerebellum of that volume. No change of consequence was observed in the specific gravity of those cerebella which had been preserved in formalin. Consequently, no corrections were made on the weights as taken. After weighing, three’ blocks were taken from each cerebellum. The entire vermis was removed and cut so that a sagittal section could be made of it entire. Then blocks were cut which would give sections nearly through the middle of each hemisphere and at right angles to practically all the folia. The plane of these was so selected as to cut the lobus biventer, or paramedianus (Bolk), on the under surface, to pass through the dentate nucleus, and to cut the anterior dorsal edge of the hemisphere near the vermis. A section so made shows for comparative purposes the areas of each hemisphere to which different functions have been assigned by Bolk (05), Rynberk (’07, ’12), et al. Figure 1 shows the plane in which the sections were taken, and figure 2 the appearance of such a section and gives the localization pattern advanced by the above writers. ‘The blocks of hardened tissue were cut about 5 mm. thick. Their maximum length and maximum breadth were then meas- QUANTITATIVE STUDY OF THE PURKINJE CELLS 235 ured in millimeters and the measurements recorded on the cards with the weights. The procedure in dehydrating and embedding the blocks from material, preserved in 10 per cent formalin was as follows: Alcohol 50 per cent + Formalin 6) per) cent; < « éc4%.). 0c. oe cece 2 hours Alcohol 70 per cent + Formalin 4 per cent.................... 15 hours Aleohol 90 per cent + ‘Formalin 2 per cent:................... 8 hours Ploomorre wer Cent: «..\) -.'skih a Meats ere Eek Lhe co tee 24 hours Alcohol + Ether, equal parts..... oe eS OF ee Se 24 hours eI tO. DEM CONL, 2 ¥ ay. mace einai ivclaeie's) 015 ete: 2 to 4 days UL ING SCA 0 ae ae nO iE te okt 1 oR a 6 hours SEED 4 58S Se eR ER ee he os at Larter ae eee 1 hour Benzole-paraffine....... a Sif, 20h bye Reg aught eee ea ire 18 hours at 40°C. BACAR CRETE Gisic1s cc cs ciecsa cls: craie este RIG aoe 3 to 6 hours at 52-58°C. Fig. 1 Human cerebellum seen from above. The line S.S. marks the plane of the section which is shown in figure 2. In the experience of the writer, old formalin cerebella tend to macerate easily, and the molecular layer especially shows a tendency to break away from the internal granular layer. Where this occurs there is always the possibility that some of the Purkinje cells will drop out during the process of cutting and staining. When the above procedure is followed, however, this maceration is not usually found and good preparations are secured. The common practice’ of washing formalin material in water before dehydrating makes the molecular layer more likely to break away from the rest of the cerebellum, and it should consequently be avoided. Material preserved in alcohol was treated as above with the exception that the blocks were put at once in 90 per cent alcohol without formol. ; 236 ROBERT 8S. ELLIS After embedding, the sections were cut on a rotary microtome at 25 u, fixed to the slide with Meyer’s albumen, and stained with carbol-thionine and eosin. Better results are sometimes secured with the cerebella of infants if Delafield’s haematoxylin and Orange G are used instead. Care must be taken that the sections be not treated with absolute alcohol, as this dissolves the parlodion. Equal parts of absolute alcohol and chloroform may, however, be safely used (King, 10). The most satisfactory fiber preparations from old formol material were secured by Bielschowsky’s pyradine method. With a modification of this the pericellular basket fibers have been clearly shown in cerebella that had been in formol for thirteen Fig. 2. Section through a hemisphere of the human cerebellum in the plane shown in figure 1. The designations are explained on page 237. years. If the pyradine is not thoroughly washed out before putting into silver, and a weak solution of silver nitrate, 0.5 to 0.75 per cent, is used and changed frequently, the fibers will be impregnated with silver in one or two days. The sections are then treated as usual. These fiber preparations give valuable assistance in determining the character and causes of the deficiency in cells. After staining, the sections on the slide were measured and the measurements recorded on the cards with the original measurements of the blocks as cut from the cerebellum at the time of weighing. From these two measurements the percentage of shrinkage is calculated. This will be referred to later. QUANTITATIVE STUDY OF THE PURKINJE CELLS Zoe METHOD OF STUDY The purpose of this study, as has been stated above, was to compare the number of Purkinje cells in different types of brains, and also to find what numerical differences, if any, exist between different areas of the same cerebellum. Reasoning by analogy from what is known of the cell losses in the cerebrum, it seemed not improbable that some areas might be found more subject to degeneration than others, and, in view of the localization theory of Bolk, it appeared all the more desirable to make a comparison of the different areas to which different functions have been credited. For this purpose each hemisphere has been divided ~ into six areas,? as shown in figure 2. 1. Lobus anterior (Bolk); head area; the region anterior to the sulcus primarius (S. pr.) 2. Lobulus simplex (Bolk); neck area; from the sulcus primarius to the sulcus postclivalis (S. pel.) » 3. Lobulus semilunaris superior; arm area; from the sulcus postclivalis to the sulcus horizontalis magnus (S.h.m.) 4. Lobulus semilunaris inferior; arm or leg area; from the sulcus horizontalis magnus to the sulcus pregracilis (S. prg.) 5. Lobulus gracilis; leg area; from the sulcus pregracilis to the suleus postgracilis (S. psig.) 6. Lobulus biventer; leg area (?); part posterior to the sulcus postgracilis, exclusive of the amygdala. In order to secure exact and strictly comparable measurements of the relative numbers of cells in these different areas and of the relative numbers in the same areas of different cerebella, the following method has been used: _ The sections were projected at a magnification of 30 diameters by means of the Edinger projectoscope, and tracings made in pencil of the line of Purkinje cells to be counted. The length of this line as traced was measured in millimeters by means of a map measurer. This divided by 30 gave the actual length of ? The first two designations (1-2) are given according to Bolk ’05 while the remainder (3-6) are those used in Quain’s Elements of Anatomy, vol. 3, Neu- rology, part 1, eleventh edition, 1908. 238 ROBERT S. ELLIS the line on the slide. The number of cells in the line was then counted under the microscope, only those cells showing the nucleolus being counted, with the exception that when the cells were so disintegrated that the nucleolus would not appear in the preparation, the cells were counted if they appeared to belong properly to the section. Dividing this total by the length of the line in millimeters gives the number of cells in a line 1 mm. long on the slide, which, as has been stated, is for a section 25 uw thick. To correct this for shrinkage during dehy- dration and embedding, the value thus obtained for 1 mm. on the slide has been multiplied by the square of the percentage obtained by dividing the sum of the length and breadth of the section on the slide by the sum of the length and breadth of the block as first cut from the cerebellum. This gives the number of cells found in a line of Purkinje cells 1 mm. long and 25 yu thick, this being really a surface 25 » by 1 mm. in the cerebellum as weighed. For comparable quantitative results this correction is neces- sary because it is evident that when the cerebellum shrinks, the line of Purkinje cells shrinks also, and it is necessary to correct both for the change in the length of the line and for the change in the thickness of the section on the slide—it being evident that the 25u thickness of the section represents a greater thick- ness in the cerebellum as weighed. The sum of the length and breadth of the section have been used rather than one dimension alone, because careful measurements show that the section is compressed in breadth to some extent when cut on the micro- tome. The sum of the two measurements consequently gives a more accurate indication of shrinkage. The number of cells found in a line 1 mm. long and in a section 25u thick would afford a satisfactory unit for comparing different areas of the same brain; it would not do, however, asa unit for com- paring different brains. If the molecular layer of the cerebellum should be removed, leaving the cell bodies of the Purkinje cells intact, it is evident that these would extend in a much convoluted sheet over the surface of the remaining part of the cerebellum. The area occupied by the Purkinje cells in a large brain is thus larger than that occupied by them in a small brain. The sur- QUANTITATIVE STUDY OF THE PURKINJE CELLS 239 faces of similar solids vary, we know, as the squares of the cube roots of their volumes. A priori we should perhaps not be safe in assuming that large and small cerebella are exactly similar solids, it being possible that small cerebella might be more convoluted than large ones so that the disparity in surfaces might to some extent at least be removed. Careful examination, however, of large and small cerebella— where the latter are from cases at least one month old—does not justify this a priori assumption. Individual differences do exist, but size has not been found to alter materially the convolution pattern after birth (Berliner, ’05). Furthermore, our results based on the assumption that large and small cerebella are similar solids will be found to agree with theoretical expectations. In order, then, to compare areas which represent the same fraction of the total number of Purkinje cells in different cere- bella, it is necessary to take areas which vary as the squares of the cube roots of their volumes. As a convenient indication of volume we have used the weight in grams of the fixed cerebellum, this having been corrected, if necessary, for change in specific gravity. A constant fractional part of each cerebellum, an equivalent unit area (HUA), has been secured by taking a line of Purkinje cells as many millimeters long as the numerical value obtained, by taking the square of the cube root of the cerebellar weight in grams—this being from a section cut 25, thick and corrected for shrinkage as already explained. If, then, we let N=the number of cells per EUA L=the length of the line of cells as projected in millimeters M=the magnification C=the total number of cells counted S:=the sum in mm. of the length and breadth of the block as first measured S.=the sum in mm. of the length and breadth of the section on the slide W =the weight of the cerebellum in grams after fixation then Wy foekes = W3 n=(C+5)x(Z pak 240 ROBERT S. ELLIS or CM ey 7 N= == as W2 L x S. x 3 As an example we may substitute the values for Case 15035, Area 3, left hemisphere (table 1). N = xe) meas 2220 89.1 = 6.47 x 0.86 X 25 = 138 RESULTS OBTAINED FROM COUNTS Neurohistologists who have made a study of the Purkinje cells are familiar with the fact that these cells are not evenly distributed, but tend to be more or less irregularly spaced, and this seems to be particularly true for the human cerebellum. In the bottom of sulci they are not normally so numerous as they are at the summits of the folia. It would not be far wrong, in fact, to say that they increase in frequency as one passes from the bottom of the sulcus to the summit of the folium. This is probably to be interpreted as incident to the phenomena of growth. The bottom of a sulcus represents an arrest in growth; the summit of the folium is the last to fill out and develop. It is, then, not surprising that a greater number of Purkinje cells should appear in the region characterized by the greatest amount of growth change. Considerable variation will be found in: the frequency of Purkinje cells, even in adjacent folia. This is especially true of defective brains, less so of normal ones. This makes it necessary to count several hundred cells in order to get a satisfactory average for any given area. Each value given in the tables in this paper is based on an average of about 400 cells, and it is believed that by taking a number so large the effects of acci- dental selection have very largely been eliminated. If perfectly satisfactory normal cerebella had been used as the basis for the norms given in table 1, it seems certain that QUANTITATIVE STUDY OF THE PURKINJE CELLS 241 all the values for the different areas would be somewhat higher than those given there. It has seemed wiser, however, to use the values actually observed rather than to attempt the more or less dangerous expedient of guessing at the extent to which these have fallen below the norm. The number of Purkinje cells in equivalent unit areas (HUA) of the two hemispheres of the four normal cerebella is shown in table 1. From table 1 it appears that the cells are. closer together in the lobus anterior (area 1) and that they become progressively less numerous in the anteroposterior direction, with the exception that in the lobus biventer, area 6, the number is slightly greater than in area.5. . In the lobus biventer the cells are usually more numerous in the posterior part. This is not shown in the table because the two parts are considered together; the fact is of importance, however, in relation to some of the changes found in subnormal and senescent cerebella, for in our cases the anterior part of this lobe has sufféred the greatest amount of degeneration. TABLE 1 Number of Purkinje cells per HUA in normal cerebella. Areas as in fig. 1. R=right hemisphere, L=left hemisphere W.I.No.| RACE | yu ,| AREA 1 | arpa 2 | anea3 | arma d | area 5 | arma 6 | Toran fee apeoit [2B (AE |p os a0 Gg et ee oa Et ee len Wee eae. idea uae esa aS Pet ae wae | EF. | Blog| allan AVIOLO SOx on ch te 162=8 | 185+2 | 135+10| 182+14) 125+6 | 127+5 Peta Note. The numeral preceded by + equals one-half the average difference between the right and the left hemispheres. 242 ROBERT S. ELLIS Differences between right and left hemispheres. A glance at the table shows that in most cases there is a considerable difference between the values for the right and left hemispheres. In three cases the values for the-right hemispheres are larger, while in one case the value for the left is the larger. Unfortunately, it is not known whether these cases were right or left handed, and it is consequently impossible to show here a positive correlation between a greater number of cells and superior unilateral skill. As a suggestion, however, it is not improbable that we have here three right-handed individuals and one left-handed. In any case it may be noted that the differences between the right and left hemispheres are greatest in areas 3 and 4, the supposed arm and hand areas, and that they are least in area 2, the neck area. On purely a priori grounds this is what is to be expected if there is a correlation between number of cells and fineness of muscular coordination. The greatest difference between the right and left — hemispheres should be in the hand and arm areas, because it is - in the two hands that there is normally the greatest difference in motor skill. In the neck area there is obviously less reason for expecting one side to be more developed than the other. How- ever, further studies must be made here on cerebella from cases with known histories before this question can be answered with any assurance of correctness. SUBNORMAL CASES In order to compare the relative frequency of Purkinje cells in the cerebella of subnormal individuals, the histories of which have been given on page 232f, with normal ones, counts have been made on areas 1, 3, 4, and 6 of the right hemisphere and on areas 3 and 4 of the left hemisphere. These results are presented in table 2. It is obvious at a glance that in this group of mental defectives the number of Purkinje cells is notably deficient. ‘The con- stancy of the results and the wide difference between the values for the normal and those for the defective cerebella as shown by’ the fact that the latter are only 63 to 78 per cent of the former make it impossible to assume that accident, chance, or. ee a ee le aa ae QUANTITATIVE STUDY OF THE PURKINJE CELLS 243 TABLE 2 Number of Purkinje cells per HUA in subnormal cases. Designations as in table 1 W.I. NO. Fae AREA 1] AREA 3|AREA4/AREA6| TOTALS R OF: be 980:| 98 sel FORE 14880 i 102 | 100 ‘ 598 R Bee 126 122" 100K TT se L 114 | 102 fdee R 123 85 Baur lage sl 15145 i 95 | 105 630 R 508) e109.) 127 15250 { L 108 | 108 + \ 660 R 135 93 89 95 |\ 52 15297 my 38 97 f 597 R 109 92 79 47 D) 15299 L 95 81 \ 503 R 85 | 100 94 98 |). 15310 * 77 73 f 527 R 93 80 95 67 |\ 2 15320 { i 85 75 f 495 PNG OTA Ce etc A Pie Cs oe eal ss Me 105 96 94 97 Per cent of normal as in table 1......... 63 iil “al 78 personal equation have played any considerable part in de- termining the differences observed. Moreover, the number of cases is large enough to make it practically certain that typically a decidedly smaller number of Purkinje cells is present in the low grades of feeble-minded individuals. Here, then, we have an anatomical deficiency as a basis of the observed deficiency in motor codrdination. To this should be added the fact that many of the cells counted were atrophied and showed con- solidation and degeneration of the nucleus to such an extent that they could hardly have been of any functional value. Many 244 ROBERT S. ELLIS other cells, although not so far degenerated, showed clearly the early stages of the process. With many of these very im- portant elements of the neuromuscular mechanism lacking, and with many more in a degenerating condition, it i is inevitable that poor coérdination should result. After the cell counts for the different areas were completed, the clinical histories of the subnormal cases were consulted and an effort was made to see how far the variations in the different areas would give evidence either for or against the localization theory of Bolk. Too much must not be expected from such a comparison, because the case histories were not made with a view to their being used in this manner and they are conse- quently incomplete. It will be interesting, however, to consider briefly the facts in each case. No. 14880 shows as compared to the normal the lowest number .of cells per EUA in the head area. The personal history shows that he had a small head and poor speech. It is consequently not improbable that the entire head musculature was under-developed and poorly coérdinated. The record says he was right handed. The. cell count shows a slightly greater number of cells in the left hemisphere. It will be shown later in discussing senescence and paresis that in right-handed individuals the right hemisphere suffers more loss than the left. The case history indicated postnatal degeneration, and this, if true, would account for the smaller number of cells in the right hemisphere. No. 15144. The history shows a bullet-shaped head and very poor speech. He was able, though, to thread needles and tie his shoes, which shows that his hands at least were capable of difficult motor coddinations. The cell counts show the head area very low, but the arm and hand area nearly normal. This agrees well with the locali- zation theory. No. 15145. The history shows epilepsy—a degenerating disease—an abscess on the right hip, and a failure in the grip test. The cell counts show the arm and hand areas lower than the others, with the right side lower than the left, this being due probably to degeneration. No. 15250. The record shows poor speech, atrophy of the jaw, and eye fatigue. The cell counts show the head area the lowest. The grip test shows the left hand slightly stronger and the cell count for area 3 agrees with this. No. 15297. The record shows a mental age of eight, which, as far as the cerebellum is concerned, would indicate a higher development of the head area with respect to eye movement, speech, and facial expression. The cell count for area 1 is the highest of the eight cases listed. This agrees with theoretical expectations. The record for grip QUANTITATIVE STUDY OF THE PURKINJE CELLS 245 shows the right hand stronger; the cell count for area 3 agrees with this, but for area 4 is opposite to it. No. 15299. This is a low-grade case, unable to talk or work; also he was knock-kneed. ‘The cell counts for both hemispheres are very low, and this is especially true in area 6, the leg area (?). A large part of the anterior part of the lobus biventer in the right hemisphere was completely atrophied. The grip test shows both hands about equally weak and there is little difference between the cell counts for the two hemispheres. No. 15310. The record shows that he was left handed and that a Wassermann test was positive. This probably accounts for the extreme loss of cells in the left hemisphere through degeneration. The grip tests shows the right hand stronger and the cell count shows more cells in the right hemisphere. No. 15320. The low cell counts here agree well with the low men- tality and the general lack of motor power and coérdination. The tumor in area 6 of the right hemisphere probably accounts for the loss of ability to walk well after the age of eighteen. Also it is important to note that many of the remaining Purkinje cells were in the process of degenerating. General motor deficiency was consequently inevitable. In view of the shia? meagre clinical data available, it was not expected that a very high degree of correlation would be found between the reported defects in motor codrdination and the numerical deficiency in Purkinje cells. Furthermore, a state- ment of the relative number of cells, if taken alone, is not neces- sarily a complete indication of functional efficiency; for the cells, although present, may be so far degenerated that they are without functional value. It is consequently interesting to find that in the cases considered there is a considerable difference, even in cell number, between the normal and the subnormal and that the losses in cells agree very well with Bolk’s locali- zation theory. The question naturally arises as to whether this deficiency in cells is due to agenesis or to some toxin or other agency present during intra-uterine life, or to postnatal disease or injury, acting on the cells already formed. To throw some light on this problem, counts were made on area 3, the arm area, of both hemispheres of six low grade in- fants. These may safely be said to be of subnormal ancestry; also they can hardly be said to have had the most favorable nutritional conditions during fetal life. How the number of cells in these cases compares with the normal is shown in table 3. 246 ROBERT S. ELLIS TABLE 3 Number of Purkinje cells per EUA in subnormal infants. Designations as in table 1 AREA 3 Ww.tI. NO. RACE AGE 18? L months 14428 B 1 120: 105 E22 W 5 126: 115 14250 W. 3350) 106: 120 H23 B 5 94: 103 14215 B 9 110: 102 14427 B 36 LOA ae Average.:.... Mi Oracle cs, 08,3 fd, ce Gace TEES Rea eR 110 Normalan mene eos ot cd score. 2 Rita) Re ee i Pe aD 2 ote Isis BRemcentiol normal as inevapler) 15 eee eeeeen ee a eens 82 The sections do not show that cells have degenerated and dropped out, and yet in every case the number of cells is dis- tinctly below normal.. The cause of the deficiency must then have acted antenatum. Whether this is to be charged to agenesis or to early. destruction is, however, not easily deter- mined. ‘The writer’s opinion, based on the evidence at hand, is that the cell deficiency is due primarily to agenesis. Beyond this it is not possible to determine from the present study the etiology of the conditions observed. In some cases, however, mental and motor deficiency appears to be due to causes which operate postnatum. This is shown in case no. 15214. The cell counts as compared with the normal are shown in table 4. These values in table 4, the reader will note, are almost normal. Yet the case is a distinctly subnormal one., The clinical history shows that the ancestry is free from feeble-mindedness as far as known; also it shows that the child was normal until nine months of age. At the time he began to have spasms and to show evidences of subnormality, due perhaps to the delayed effects of congenital lues. The cerebellum is of normal weight, but the —E——————— QUANTITATIVE STUDY OF THE PURKINJE CELLS 247 TABLE 4 Number of Purkinje cells per HUA in case No. 15214, an example of postnatal arrest of development. Designations as in table 1 w.I. NO HEMISPHERE ARBA 1 AREA 3 AREA 4 AREA 6 R 159 124 106 143 pet { L 147 151 ATEN AC ee ger}. eve ayer 159 1385+12 128 +23 143 ISG aes ae eee 162=8 135+10 1382+14 127-5 cerebral hemispheres are much below weight, which further indicates that the cause of the arrest took effect after birth. An examination of the cells reveals on the anatomical side at least a reason for the defect. Many cells are atrophied and are so far degenerated that they cannot have any functional value. Other cells are less degenerated and some are apparently normal, but others have completely disintegrated and have been ab- sorbed. This is clearly a case of postnatal degeneration, but even so, it represents the exception rather than the rule among subnormal individuals. The entire case history shows this. In this as in the other cases, then, we find that in low-grade mental defectives there is a distinct deficiency, either numerically or cytologically, in a large percentage of the Purkinje cells. SENESCENCE It is a familiar fact that very old people usually suffer a con- siderable loss of motor control. The results of counts made on five cases of senescence of different ages are given in table 5 and show how the cells drop out with increasing age. The results of table 5 are shown graphically in chart 1. The value for the normal is based on table 1 and the starting-point for the drop in the curve is placed somewhat arbitrarily at the age of forty years. In connection with this curve it is interesting to note that it is based on the cerebella of two men of superior mentality, on one negro male autopsied in a general hospital, and on one white male and one white female dying at extreme old age in hospitals THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, No.2 248 ROBERT S. ELLIS TABLE 5 Number of Purkinje cells per HUA in senescent cerebella. Designations as in table 1 HEMI- W.I.NG.| RACE| SEX AGE | ueRE| BEA 1 AREA 3 AREA 4 AREA 6 Total Noma 42? 162+8 { ee f Hig 2725 14464 | B. | M. | 62 { - it an Bs a } 667 16107 | W. | M. | 65 { x es ee 3 oe \ 591 14340 | w. | M. | 79 : ot a os a } 500 E27 | w.| F. | 94 x & x ei ee \ 462 14544 | w. | M. |100 - ae * = mi \ 403 for the insane. The uniformity of the curve is consequently surprising. Much speculatidn might be based on such results as these, but here it will suffice to call attention to two points: first, the average loss in area 1, the head area, as compared with the values from table 1, is relatively the greatest; second, in areas 3 and 4, the right hemisphere suffers more than the left. PARESIS It was not originally intended to include any cases of paresis in this study, but as the cerebellum of one case was prepared before the cause of death was known, and as it is of interest because of its similarity in cell losses to the other types of cases presented, it seemed worth while to include it for purposes of comparison. The results of the cell counts for this case are presented in table 6. Here again we find the head area very low and the right hemisphere much lower than the left. This was a right-handed man and the cell losses have been greatest on the right side. QUANTITATIVE STUDY OF THE PURKINJE CELLS 249 The figures for cell losses do not of course fully indicate the loss in functional efficiency of such cerebella. Many of the cells that remain are in the process of disintegration and there- fore many of those included in the cell count probably have little or no capacity for functioning. NUMBER OF PURKINJE CELLS “iS ae RE RRA ao 1 a a i 2 mmiamisietstet ele yaa) sia SEA SESE Saisie arias eb eee e eet DSRS RECGGS OS Oo Cece ee SESS 5 SG00S 00s aso U Ae SS : ERO GRS BEG Ioo. s HUSSRSSR TOE SE ss ios = tees SCGSERNS DOTS C IIe ee BSE ou hacia eae betta boon BA ae ee ) G0 Ee es Noe SUT UteRResoec cose eee ja a Dm CESAR SSS Ie iN oe 2S aE SIT a {oe (62S RSS oo Oe ese aes - SCRE n SNE) oO Sdn ZOBUSEESEES a3 /c5 oye ens Toft halal de UL) a a a cy ead Ve O08 re a I Ls LO oor comes So. FOS e100 Chart 1 Showing the decrease in the number of Purkinje cells with advanciny age. Based on the totals of the four areas given in table 5. The dots show the relation of the six observed values to the graph as drawn. THE VERMIS Sections of the vermis were prepared in all cases and a few preliminary counts were made to determine whether or not it would be profitable to make counts on all cases. The results for the tuber vermis in five cerebella—two normal, one senescent, and two subnormal—are given in table 7, and for comparison, the averages of area 4 of both hemispheres are given also. No significant. differences in the relations found in the' two sections appear from this table. The sections of the vermis from the other cases were examined under the microscope with- out counting, and as it seemed clear that they would not differ materially from what had been found in the hemispheres, no further counts were made. 250 , ROBERT Ss. ELLIS TABLE: 6 Number of Purkinje cells per HUA in a case of paresis. Designations as in table 1 W.I. NO. AGE AREA | AREA 3 AREA 4 AREA 6 ‘ R 107 99 68 90 area ee { ie 133 118 Normalan. cte ec ee ir cies: 162+8 135+10 132+14 127+5 COMPARISON WITH THE CEREBRAL CORTEX The writer has had occasion to make cell counts on a number of areas in the cerebral cortex, and has found, in agreement with most others who have conducted investigations along this line, that there is ordinarily a distinct deficiency in cells in cases of amentia. Among those authors who may be mentioned here are, for example, Hammerberg (’95), Roncoroni (’05), Tredgold (03), and Bolton (’08). The motor area of the cortex in several cases of paresis and also of senescence has been carefully examined, and it is interest- ing to note that in all of these there has been found a deficiency in Betz cells. How general-this particular loss is, the writer does not know; it is, however, important to recognize that disintegration takes place in the cerebrum in a manner similar to that found in the cerebellum. It is therefore not probable that the motor deficiencies observed in the cases studied have been due solely to deficiencies in the cerebellum. Probably in the majority of cases a defective cerebellum is accompanied by a defective cerebrum, and vice versa. It would be beyond the scope of the salar paper to attempt a more detailed statement of the character of the cell changes in these atypical brains; enough has been done, however, if we have succeeded in pefablahine the numerical differences found in these different types of cases. QUANTITATIVE STUDY OF THE PURKINJE CELLS O51 TABLE 7 Number of Purkinje cells per EUA in the tuber vermis, compared with the average for area 4 NORMAL SENESCENT SUBNORMAL W.I. No. | W.1I. No. | W.I. No. | W.I. No. | W.I. No. 14485 15073 16107 15310 15320 PT DCI VETS) a1. ios eee ee 157 137 112 83 91 PATO A Beat ssi sidc natheteac tee ernee 138 119 87 84 85 SUMMARY AND CONCLUSION The main purpose of this paper was to show the numerical differences in Purkinje cells in normal, subnormal, and senescent cerebella. | From the data submitted it is evident that in cases of extreme mental defect due to agenesis or to the early action of toxins during intra-uterine life, there is an evident deficiency in the number of these cells. Similar reductions in the number of cells due to various causes are found in senescence (and paresis). In the subnormal cerebella the evidence indicates that the normal number of cells has never been present in a developed form. In the senescent (and paretic) cases, however, the small number is due to disintegration. The anterior lobe of the cerebellum shows the greatest de- ficiency in cells in both the subnormal and’ senescent cerebella. The biventral lobe shows the greatest variation in both types of cases. In some cerebella it shows the greatest loss of cells and in others the least loss. The differences between the two hemispheres in respect of cell number average less in subnormal cerebella than in normal ones. This probably has a relation to the differences in the degree of unilateral dexterity found in normal and subnormal individuals, i.e., normal people are usually more distinctly right (or left) handed than are the subnormal, who tend to be more ambidexterous. ae ROBERT §S. ELLIS The deficiency in cell number affords in large measure an explanation of the motor defects found in subnormal individuals. It shows, furthermore, that in idiocy and in imbecility we may expect to find the whole brain defective rather than the frontal lobes only, while the higher grade of defectives (morons) proban show very slight deviations from the normal. Further studies in this field on better material with better- known clinical histories in which is included a study of blood- vessels, nerve fibers, neuroglia, and a cytological study of all the important types of cells are necessary to bring out the more detailed differences between these types of cerebella. For a review of the literature on the clinico-pathological study of the cerebellum with a detailed report of a single case see Archambault (’18). LITERATURE CITED ARCHAMBAULT, LASautite 1918. Parenchymatous atrophy of the cerebellum. Jour. of Nerv. and Ment. Disease, vol. 48. ~ > BERLINER, K. 1905 Beitrige zur Histologie und Entwickelungsgeschichte des Kleinhirns. Archiv. f. mikr. Anat., Bd. 66. : Botx, L. 1905-07 Das Cerebellum der Siugetiere. Petrus Camper, Neder- landsche Bijdragen tot de Anatomie, Bd. 3 u. 4. Botton, J.S. 1903 The histological basis of amentia and dementia. Arch. of Neurology, vol. 2. 1910-11 A contribution to the localization of cerebral function based on the clinico-pathological study of mental disease. Brain, vol. 33. Gopparp, H. H. 1914 Feeble-mindedness. New York. HAMMARBERG, ©. 1895 ‘tudieniiber Klinik und Pathologie der Idiotie. (Trans. from Swedish into German by W. Berger and pub. by S. £. Henschen). Upsala. Kine, Heten D. 1910 The effect of fixatives on rats’ brains. Anat. Rec., vol. 4. ‘ Roncoront, L. 1905 Lo sviluppo degli strati molecolari del cervello, ete. Arch. di Psichiat., vol. 26. RynBerk, G. Van 1907-08 Die neuren Beitrige zur Anatomie und Physiologie des Kleinhirns der Siiuger. Folia Neuro-biologica, Bd. 2 1912 Weitere Beitrige zum Localizations-problem im Kleinhirn. Folia Neurobiologica, Bd. 6, Supplement. 1 TrepcoLty, A. F. 1903 Amentia. Arch. of Neurology, vol. 2 1914 Mental deficiency. 2nd Ed. London. * ee er AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MARCH 17 THE EFFECT OF OVER-ACTIVITY ON THE MORPHO- LOGICAL STRUCTURE OF THE SYNAPSE KIYOYASU MARUI Sendai, Japan Neurologial Laboratory of the Henry Phipps Psychiatiic Clinic, Johns Hopkins Hospital, Baltimore, Md. FOURTEEN FIGURES INTRODUCTION Investigations on the histological manifestations of the nerve cell in fatigue have long been familiar to us. Numerous authors have brought contributions to this important and interesting subject. If we take for granted, as Sjoeval declared, that the al- teration of nerve cell in tetanus is to be regarded as the effect of activity, the number of references becomes even more abundant. It would suffice to point to those works of Hodge (14, 15, 16), Vas (33), Mann (23), Lugaro (29), Sjoeval (82), Dolley (17, 18, 19), and many others. There is almost a complete agreement on the point, that over-activity causes appreciable histological alterations in the nerve cell. Kocher (19), who studied the same subject in our laboratory, could not, strangely enough, find any qualitative and quantitative changes in the histological characters between the fatigued and resting nerve cells. Recently Saito undertook an exploration in the same line in this clinic, but the results are not published yet. Compared with the abundant researches on the neurocyto- logical manifestations, there has been no attempt made to inves- tigate the histological alteration of the synapse in fatigue, as far as I know. Not only in regard to over-activity, but also in other pathological conditions, the histopathological changes of the synapse have been the topic of very few investigations. And no wonder, for the structure of the synapse had not yet been con- clusively demonstrated even in the normal condition, despite 253 254 KIYOYASU MARUI many investigations by different authors (Golgi, Bethe, Held, Bielschowsky, Auerbach, and many others). The lack of our knowledge concerning the pathological manifestations in the synapse and especially of experimental pathological data seems to be dependent on the difficulty of technique on the one hand and inaccessibility of adequate material on the other. In my recent publication (24), I described more minute structures of the synapse of the Mauthner cell of teleosts (bony fish). Per- haps I have not made the structure of the synapse clear beyond discussion; however, my results have been sufficiently definite for me to attempt experimental work on the pathological con- dition, especially in fatigue. Careful and thorough investigation on this topic led me to very interesting results, which throw light on the histological condition of the synapse in functional activity, although it was forced far beyond the physiological limit. In the present paper I will describe mainly the manifesta- tions in the synapse as it was not my purpose to investigate the neurocytological alterations of the Mauthner cell itself. But wherever notable manifestations of the cell body appear, they will be described as well as the findings of the synapse. ‘To Dr. Adolf Meyer I beg to express my high appreciation here for his help and kind suggestions in this study. MATERIAL AND METHODS OF STUDY In the present investigation Ameiurus nebulosus was the ma- terial of experimentation; Carassius auratus, which offered me many interesting results in my previous work, proved to be un- favorable material, owing to the fact that it is weak and often dies before it comes to the utmost exhaustion. It was always at- tempted in this experiment to work with strict control of the nor- mal structure of the Mauthner cell as well as its synapse, and care was always taken to avoid the formation of artefact as much as possible. As a resting non-fatigued control, which was provided in each experiment, I used a fish of about the same size which was kept in a resting condition during the experiment of the other fish. The body length of the fish used varied from 5 to 7 inches. EFFECT OF OVER-ACTIVITY ON SYNAPSE 250 The experiment consisted in forced activity of the fish, carried to the most advanced stage of fatigue; the experimental fish was placed in a jar 7.5 inches in diameter and about 8.5 inches in depth and then the water in the jar was continually stirred by running water, which came with high pressure from a faucet. The fish swims in the stirred water, trying all the time to hold its equilibrium. According to Bartelmez (3), the Mauthner cell participates in the equilibratory reflex; so it was supposed that in this experiment the Mauthner cell would be forced into con- tinuous activity. The duration of the experiment varied re- markably in each fish (from 24 to 98 hours). At first: the fish swims actively and is able to hold its equilibrium very well, but gradually it gets tired and in the final stage of the experiment it is deprived of the ability of balancing, so that it is moved pas- sively tumbling around in the stirring water. By this time if we stop the running water for a moment, the fish would lie on its back or on one side, showing no attempt to maintain its upright pos- ture. As the sign of utmost exhaustion, I chose a test, which con- sisted in holding the fish upside down by its tail in the water as well as in the air. In the most advanced stage of exhaustion the fish did not flap at all even in this test. The fish was then decapitated and bled, the brain was quickly but carefully dissected out and fixed in 10 per cent formalin, for- mol-Zenker fluid, and alcohol (95 per cent), respectively. The resting control fish was killed at the same time and the brain was manipulated in .quite the same way as the fatigued brain. It must be emphasized also that the material was always fixed fresh and that material from fish which died was not examined. From the material thus obtained, the following preparations were made with the same technique as was described in my recent publication: 1. Thionin-eosin preparation (1 series each of normal and fatigued brains). 2. Toluidin-blue preparation (5 series each of normal and fatigued brains). 3. Heidenhain preparation (6 series each of normal and fatigued brains). 256 KIYOYASU MARUI 4. Levaditi preparation (14 series each of normal and fa- tigued brains). 5. Bielschowsky preparation (6 series each of normal and fatigued brains). 6. Cajal preparation (5 series each of normal and fatigued brains). Besides these I used the following stains in the present work: 7. Scharlach stain (5 series each of normal and fatigued brain). The fish brain was fixed in 10 per cent formalin twenty-four hours, rinsed with water, cut at 154 with the freezing micro- tome. The sections were stained in saturated solution of Schar- lach R in 70 per cent alcohol, rinsed with distilled water, counter- stained with diluted Ehrlich’s hematoxylin solution, washed again, and mounted in glycerin. As there is only one pair of Mauthner cells in a fish brain, it was rather hard to get the sec- tions in which the Mauthner cell is found. 8. Mallory preparation (5 series each of normal and fatigued brain). In this preparation the brain was fixed first in 10 per cent formalin twenty-four hours and then placed in formol-Zenker fluid twenty-four hours. Paraffin sections 5 to 8 thick were stained with a diluted solution of Mallory’s hematoxylin. 9. S.-fuchsin-light green stain (5 series each of normal and fatigued brains). Brains were first fixed in 10 per cent formalin twenty-four hours and then placed for eight days in the chromic acid acetic acid mixture. Paraffin sections of 54 were treated as follows: 1) Remove paraffin from the sections and pass to 96 per cent alcohol. 2) Place sections for one hour in a saturated aqueous solution of S-fuchsin in the incubator at 58°C. 3) Wash twice with distilled water, till no more stain comes out of the sections. 4) Dip the slides in motion 10 to 20 seconds in the following solution: Saturated alcoholic solution of picric acid 30 Aqua destillata 50 5) Rinse carefully twice in water. 6) Place the slides in a sat- urated aqueous solution of light green twenty minutes. The sec- tions were then washed with water, dehydrated very quickly and passed into xylol. EFFECT OF OVER-ACTIVITY ON SYNAPSE 257 In all 122 series of normal and fatigued brains form the basis of the present article. ON THE INTERNAL MORPHOLOGY OF THE MAUTHNER CELL AND ON THE MINUTE STRUCTURE OF THE SYNAPSE Under this heading I will make a few notes on the internal mor- phology of the Mauthner cell (figs. 1 and 2) and on the structure of the synapse, which are necessary as the foundation of the fol- lowing statement and were not yet described in my recent pub- lication. Nissl bodies are distributed evenly through the cell body and bases of the dendrites, leaving free only the axone hil- lock. They are relatively small as compared with those in the motor cells and very numerous, as Bartelmez (3) stated. The Nissl substance is found in the shape of variably long striae and is arranged generally parallel to the contour of the cell body and in part also to the surface of the nucleus. The remaining stain- able substance is irregularly scattered and is more or less short; some of it is spheroidal. The spindles were found especially on the surface of the cell and in the larger dendrites. The so-called nuclear caps did not come to my observation. The axone hillock is entirely free from stainable substance and marked off by a tol- erably sharp-curved plane from the granular protoplasm of the cell body and shows at its margin a layer of especially fine gran- ules. The nucleus of the Mauthner cell differs in no essential from the typical nuclear structure of the nerve cell. As was precisely stated in my recent communication, the synapse of the Mauthner cell is penetrated by the Golgi network, which is formed by the arborization and the reunion of the deli- cate processes of the neuroglia cells in and about the ‘axone cap.’ It was also accepted that the Golgi network is to be attributed to that category of the neuroglia tissue, which Held (24) termed as the reticular glia tissue formed by the somewhat modified plas- ma of the glia cell. Furthermore, I paid special attention in that paper to the histological structure of the nervous elements of that synapse and the relation of the latter to the Golgi net. We may therefore pass directly to the description of the finer structure of the glia cells themselves and the condition of the capillaries in this synapse. uu 4, +, of EA EFFECT OF OVER-ACTIVITY ON SYNAPSE 259 The neuroglia nucleus is surrounded by a variously wide bor- der of protoplasm, which latter sends protoplasmatie processes in different directions. The process itself ramifies and becomes more and more delicate, until it passes into the beams of the Golgi net or becomes attached to the wall of a capillary. This picture could very clearly be observed in the Levaditi preparations, Heidenhain preparations of formalin material, Mallory and acid- fuchsin-light green preparations. In the Heidenhain and thionin- eosin preparations of formol-Zenker material a similar condition was demonstrated, although it was not so clear as in the above- _ mentioned preparations. The protoplasm of glia cells was brought out favorably in the thionin-eosin preparations, Heidenhain preparations, Mallory and acid-fuchsin-light green preparations. There are two different types of neuroglia cells; in one type the protoplasm is very scanty, so that the glia cell shows a small round cell body, while in another type we find a large mass of protoplasm around the nucleus. The neuroglia nuclei are sometimes connected with each other not by means of the Golgi-net substance, but by a variously broad mass of granular protoplasm. I should interpret this as the result of amitosis which is observed now and thenin thesynapse. It must be positively emphasized here, however, that all the glia cells of the synapse belong to one reticulum and that there is no cell indi- vidual among them in normal brains. The structure of the glia nucleus hardly calls for a description except that sometimes it shows evidences of amitosis, as Bartelmez (3) also stated. Cap- illaries are found here and there in and about the synapse of the Mauthner cell. They do not offer anything particular in their The figures 1 to 8 are the unretouched photomicrographs taken from different preparations of both the control and the fatigued Ameiurus brains. In figures 2 and 5 an apochromatic Zeiss ocular no. 4 and Zeiss objective D were used; others were taken with the same ocular and a Zeiss immersion js. The length of bel- lows was 60 em. in all the photomicrographs. The figures 6 to 14 were drawn from different preparations of fatigued Ameiurus brains, using the Abbe camera lucida. (Zeiss apochromatic ocular no. 4, Zeiss oil-immersion ;'z, tube length 20 em.) Fig: 1 Toluidin-blue preparation (alcohol material) (control fish). Fig. 2. Thionin-eosin preparation (formol-Zenker material) (control fish). Fig. 3 Toluidin-blue preparation (alcohol material) (fatigued). 260 KIYOYASU MARUI structure and consist of endothelium and adventitia with very few nuclei. As already remarked, the adventitia of the capillary is connected with the bases of the reticular beams; but I was not able to distinguish the perivascular limiting membrane of Held as a membrane separated from the adventitia, so that I could not find the so-called perivascular lymph space between them in normal fish brains. THE HISTOLOGICAL MANIFESTATIONS OF THE MAUTHNER CELL IN FATIGUE To recapitulate and discuss the results of other authors here would go beyond the purpose of the present work; my remarks will be restricted to my main results of investigation. The cell body of fatigued cells was found either in the state of turgescence (figs. 3 and 4) or of shrinkage (fig. 5); in the former case the cell border was convex between the dendrites and the den- drites appeared shorter, whereas in the latter the cell border was concave and the dendrites looked longer. I agree with the opin- ion of Vas (33), Mann (23), Lugaro (21), Pugnat (26) and Holm- gren (17), that the enlargement.of the cell body is to be considered as the manifestation of activity and the shrinkage as that of ex- haustion. In this way the results of Hodge (14, 15, 16), which de- viate from those of others, might well be interpreted. Dolley (7, 9) described the fluctuations of the size of the cell body in the course of activity, but I. have a little doubt about his statement that in later stages the absolute size of the cell body increases steadily to the end. The alteration of the Nissl substance mani- fested itself in more or less advanced stages of chromatolysis, as was described by Vas (33), Lambert (20), Mann (23), Lugaro (21, 22),and many others, and thereby the cytoplasm was stained vari- ously deeply (figs. 3, 5). The Nissl bodies were found in a state of fragmentation, shortening, and irregular distribution; in the advanced stage of chromatolysis they were reduced to fine gran- ules or even to a homogeneous substance. Sometimes I observed the central beginning of the chromatolysis, as was stated by Vas (33), Mann (23), and, in tetanus, by Sjoeval (32). EFFECT OF OVER-ACTIVITY ON SYNAPSE 261 nt s - t Ls tle tre AY, A n> os Sas Se Aaa Ait pte < Mere Pes televise 3 QOS ARE Rae eK Fig. 4 Toluidin-blue preparation (alcohol material) (fatigued). Vig. 5 Thionin-eosin preparation (formol-Zenker material) (fatigued). 262 KIYOYASU MARUI The nucleus was sometimes located in the center of the cell body, but in many cases it was situated more or less eccentrically, as was described by many others—Vas (33), Lambert (20), Sjoeval (32), Holmgren (17). Morphologically, the nucleus showed now and then no particular change, but in other cases it was found either swollen or shrunken. The swollen nucleus appeared large and round and showed a smooth nuclear membrane, whereas the shrunken nucleus showed irregular shape and a crenated mem- brane. Vas (33) observed always an enlargement of the nucleus, while Hodge (14, 15, 15, 16) found regularly the shrinkage of the latter. Lugaro (21), Mann (23), Pugnat (26), and Holmgren (17) declared on the basis of their study that activity causes turges- cence followed by shrinkage in exhaustion. Sjoeval (32) denied the pathological significance of this finding for the reason that he observed those nuclei also in the cell which showed no change in particular. I think those manifestations of the nucleus are to be attributed to different stages of activity; Dolley (9) also declared that the nucleus shows fluctuations of size in the course of activity. The nucleolus was found mostly eccentric in the nucleus, al- though this was the case also in some resting nerve cells. It was sometimes observed swollen (fig. 4); in other cases it showed an irregular shape, oblong or angular (fig. 5). Mann (23), Lugaro (21), Luxemburg (22) observed the enlargement of the nucleolus during activity; it disappears later in exhaustion (Mann (23), Luxemburg (22) ). Goldscheider and Flatau (12, 13), and Sjoeval (32) confirmed this in tetanus. Mathes (25) observed the nu- cleolus of angular shape as I did; the objection, that it might be the deceptive appearance caused by the granules above or be- neath the nucleolus, could not come in question in my cases. As a most peculiar manifestation of the nucleus figure 5 was presented; the nucleus is very large compared with the size of the cell body, the nuclear membrane is marked sharply and the nucle- olus itself is small and shows an ellipsoid shape. Besides that there is a deeply blue-stained substance of triangular shape and the remaining space of the nucleus is filled with compact acido- phile substance. The nucleus shown in figure 6 might come under EFFECT OF OVER-ACTIVITY ON SYNAPSE 263 Fe at 4 7 Figs. 6 and 7 Levaditi preparation (fatigued). 264 KIYOYASU MARUI the same category of nuclear change; we observe here besides the nucleolus a rod-shaped substance which shows a staining reac- tion similar to that of the nucleolus, although I am not sure about this, as it is from a Levaditi preparation. As far as I know, no such picture of the nucleus has been described before; figure 8 of Dolley’s (7) publication demonstrates a cell, the nucleus of which shows a deeply stained spheroidal substance besides the nu- cleolus, but Dolley did not mention anything about that in the text or in the description of the plate. Whether this kind of mani- festation of the nucleus is to be regarded as the alteration of double nucleolus, which latter is met not infrequently, or can be attrib- uted to a special appearance of the nucleus in fatigue, I can- not tell. Holmgren (17), Sjoeval (32), and others found the stainable substance massed about the nuclear membrane, and forming either an irregular or a complete ring. I also observed the same phenom- enon in many cases (fig. 3); the nuclear membrane was out- lined by a delicate blue-stained line or a large mass of stain- able substance in a different portion of its circumference, giving the picture of a half-moon. Sjoeval and Holmgren interpreted this as a restitution phenomenon of the tigroid substance; Dolley (7) also regarded it as a sign of greater nuclear activity. I agree with the opinion of these authors. Holmgren (17) observed besides, that the nucleolus and the nuclear granulation emigrate from the nucleus into the cell body. The emigration of the nucleolus never came to my observation; the case, however, from which figure 4 was reproduced may indicate the emigration of stainable sub- stance from the nucleus into the cell body. The nucleus as well as the cell body is swollen in this case, and the nucleolus is also ex- tremely swollen, and we find many blue-stained granules going out of the nucleus into the cell protoplasm. On the other hand, I found also the accumulation of the acidophile substance in the nucleus (fig. 5). On the basis of these findings I should agree also with Holmgren, who came to a conclusion that a mutual inter- change of substance takes place between nucleus and cell proto- plasm in activity. On the ground of Richard Hertwig’s doctrine of nucleus-plasm ratio, Dolley (6, 7, 8, 9) measured the size of cell EFFECT OF OVER-ACTIVITY ON SYNAPSE 265 body and nucleus of nerve cells in activity and divided the cells. into many stages of alteration. As I did not undertake the meas- urement of the cell body and nucleus, I will not go further into details of Dolley’s work; but his method of division is, as he him- self admitted, an arbitrary one, and I found many cells, which can- not be assigned to any of his stages. Furthermore, I am afraid that in his interpretation of things, facts are linked with hypo- thetical considerations which are not directly observable. I will be satisfied in the present work, if I can make sure that the Mauthner cell, the synapse of which is the material of this study, manifests appreciable changes in fatigue. THE HISTOLOGICAL MANIFESTATIONS OF THE SYNAPSE IN FATIGUE In these experiments which consist in forced activity, although it goes far beyond the physiological limit, attention was direct- ed from the first, not only to the nervous constituents of the synapse directly, but also and especially to the manifestations in the glia tissue, which shows the changes of the functioning nerve tissue in an indirect way. It was hoped that through the study of the histological manifestations in fatigue some light would be thrown upon the problem, concerning the function of the neurog- lia cells in the state of physiological activity. I shall first de- scribe the findings in the synapse of the Mauthner cell in fatigue and later go over to the consideration of the significance of the manifestations. A. Manifestations in the Pericellular Reticular Structure of the Synapse The gha reticulum of the axone cap and of the cell surface pre- sented itself in fatigue in a more or less advanced stage of devia- tion from its normal configuration. In the Levaditi preparations, which bring out the net figure very clearly and sharply in a dark brown or black color, the alteration of the net configuration is most distinctly noticeable. Figures 3 and 4 of my recent pub- lication (24) demonstrate the normal condition of the glia reticu- lum in the Levaditi preparation. The Golgi network, which is visible on the cell surface (3) as well as in the axone cap (4), 266 KIYOYASU MARUI stands out sharply, and the net beams are demonstrated in rigid dark lines. In many eases of fatigue this net configuration appears more or less irregular and less sharply marked. The net beams present themselves swollen and thick here and there; in other places they are thinner than usual and in some places even broken up. The substance of the net beams looks loose and less compact and in the extreme state of decay the net beams are reduced to variably large amorphous corpuscles and look not unlike silver precipitates distributed irregularly in the synapse. 3 Figure 7 was reproduced from the Levaditi preparation of fatigue case; it shows the surface section of a slightly shrunken Mauthner cell and the cell surface is covered partially with the Golgi network. The reticular structure of the synapse stands in this case in a more advanced state of alteration. The spheroidal or star-shaped structures, which correspond to the nervous ter- minal feet, are demonstrated very distinctly in the nodal points of the Golgi network (especially clear in the lower part of the fig- ure). Now, the net beams connecting these terminal feet with each other are in part marked tolerably sharply, but most of them are swollen and are not clean cut. Some others look, on the con- trary, very thin and loose or even broken up. Here and there we find the terminal feet, which lie isolated on the cell surface, as a consequence of the breaking up of the radiating net beams. The reticulum of the axone cap (upper part of the figure) is in a similar state of alteration; the net beams are extremely swollen and loosened and appear thick. The reticular figure is partly well preserved, although we find here and there the decay of the net beams. Figure 8 was produced from another case prepared by means of Levaditi’s method; the Golgi network in the axone cap as well as on the cell surface shows essentially similar changes. The net figure here and there is preserved tolerably well, but the meshes are irregular as compared with those in the resting condi- tion. In other parts of the synapse the net beams show a more or less advanced state of alteration and some beams are really re- duced to amorphous black fragments. EFFECT OF OVER-ACTIVITY ON SYNAPSE 267 The above described findings were repeated in a different inten- sity In many cases of fatigue by means of Levaditi’s method, although there were several cases with negative findings in the synapse. It must be emphasized here that in the resting control animal the pericellular reticular structure was always brought out clearly and sharply. Of the manifestations of the glia cells in the synapse I shall give a more precise description later. As espe- Fig. 8 Levaditi preparation (fatigued). cially worthy of note here, I want to discuss the relation between the glia cells and the reticular glia structure. The protoplasm and the protoplasmatic processes of the neuroglia cells increase and swell in mass, as will be mentioned below. Some processes show their relation to the reticulum even in their swollen condi- tion, but some of the processes are no longer connected with the glia reticulum. Some of them fall to pieces so that we find pro- toplesm masses of different shape and size around the glia cells. 268 KIYOYASU MARUI Some glia cells even lie freely in and about the synapse; to this I shall come back again. In the thionin-eosin preparations of the formol-Zenker material it is very difficult to find such delicate alterations of the reticular structure. In a slight alteration it is almost impossible to distin- guish the difference between the resting condition and the fatigue case, so far as the net figure is concerned. In the most advanced stage of alteration, however we are able to find similar manifesta- tions as those described in Levaditi preparations, although it is not so easy to see as in these preparations. Figures 2 and 5 were reproduced from thionin-eosin preparations. In figure 2, which demonstrates the resting condition, we find regularly arranged dark points around the cell, which evidently represent the net beams of the pericellular reticulum. In the fatigued cell (fig. 5) we find around the cell body a number of similar dark points, which are, however, scattered without any order at the cell periphery. The microscopic observation revealed clearly a change of peri- cellular network similar to that described in Levaditi prepara- tions; the net beams were in part swollen and others broken up. The findings of the glia cells I shall describe later. In the Heiden- hain preparations and other preparations a similar manifestation of the reticular structure was observable, although it is not so clearly demonstrable as in the Levaditi preparations so that we ean find the alteration only in a far advanced state. B. Manifestations of the neuroglia cells in the synapse Among the neuroglia cells of the synapse of the Mauthner cell I found in many cases Alzheimer’s amoeboid glia cells (figs. 9, 10, 11, 12, and 14). These amoeboid glia cells, as characterized by Alzheimer (1), show a large protoplasmatic cell body with a spe- cial morphological structure and small dark nuclei, and in typical Figs. 9, 10, and 14 Thionin-eosin preparation (fatigued). Fig. 11 Acid-fuchsin-light green preparation (fatigued). Fig. 12 Mallory preparation (fatigued). Fig. 13 Scharlach R stain (fatigued). 269 270 KIYOYASU MARUI forms they have striking resemblance to an amoeba. Besides the typical forms I found many others which do not resemble amoeba. Before I pass to the description of the amoeboid glia cells, I will pay some attention to the manifestations of glia cells which go hand in hand with the appearance of the former. Although not so numerous, we find glia cells in every stage of regressive and pro- eressive change (figs. 9, 10, and 14). The regressive nuclei appear sometimes extremely swollen and pele and at other times they show a zigzag shape and deep stain. The homogeneous stain of the nucleus and protoplasm, the breaking up of the nu- cleus into spherules or small masses are other histological prop- erties of the regressive nuclei. Furthermore, I observed in the same sections production of youne amoeboid glia cell; karyokinesis of the glia nucleus was found now and then, and amitosis of the nucleus, observable occasionally in the physiological condition, seems to appear oftener in fatigue preparations (fig. 6). In young amoeboid glia cells the shape of the cell body is simple and we find sharply marked protoplasm around the nu- cleus. In older cells, however, the protoplasm grows larger and sends processes of irregular shape in different directions. The process itself has at first a simple shape, but later it shows a more or less complicated shape; sometimes I observed that the gha cells send the processes to nerve fibers or capillaries, holding the latter between their ramifications. Generally speaking, the amoeboid glia cells were found relatively more numerous near the blood-vessels than in the other parts of the synapse. In young amoeboid glia cells the protoplasm is first quite homogene- ous, but in the further course of life many kinds of manifestation become noticeable in the cell body. In the Mallory preparations (fig. 12) the protoplasm of large amoeboid glia cells contains variously large vacuoles; the size of the vacuoles varies considerably, but generally speaking they are not very large in my preparations. The number of the vacuoles is also variable with the s ze and age of the cell. The content of the vacuoles is quite clear in my preparations; I assume that these vacuoles are lipoid cysts, the content of which was extracted in the process of embedding. In my thionin-eosin preparations. EFFECT OF OVER-ACTIVITY ON SYNAPSE 271 (fig. 10) and fuchsin-light green preparations I could also find those vacuoles. Besides these I observed in the Mallory prepa- rations dark violet or blue-stained granules in the cell protoplasm. There is no doubt that these granules are identical with the methyl-blue granules of Alzheimer (1). This kind of granules varies considerab y in size, but in any one cell they are in general of similar size as Alzheimer described. With the production of this kind of granule the loosening of the protoplasm structure of the amoeboid glia cells takes place. In the acid-fuchsin-light ereen preparations, in which the methyl-blue granules cannot be brought out, the cell body shows a granular or bubble-like ap- pearance in this stage. At the same time marked changes be- come noticeable in some nuclei; they are stained either homo- geneously deeply or remarkably pale. Even the neuroglia cells in the synapse, which are not in possession of the proper attributes of an amoeboid glia cell, but have long narrow processes instead of a large protoplasm mass, sometimes display alterations; then the processes look as though they were dissolved into granules, which show the same staining reaction as the methyl-blue granules. In the acid-fuchsin-light green preparations (fig. 11) the amoe- boid glia ‘cells were brought out very clearly; in the evenly. green-stained cell protoplasm the Alzheimer fuchsinophile gran- ules appeared as large red spherules. The number of these gran- ules varied in different cells; some large and old cells have gran- ules scattered through the whole protoplasm. The sizes of the granules are almost equal and the considerable size distinguishes these granules from the fine fuchsinophile granules, which are only occasionally observable in the normal glia cells. As already re- marked, I found also in this preparation vacuoles of different size in the cell body of the amoeboid glia cells. The contents of these vacuoles in my preparations were always clear; the large lipoid cysts with yellow substance described by Alzheimer (1) did not come to my observation. The Alzheimer light green granules were not observed either in my preparations. In my thionin-eosin preparations (fig. 10) of formol-Zenker ma- terial I found also a number of typical as well as atypical amoeboid THE JOURNAL CF COMPARATIVE NEUROLOGY, VOL. 30, NO. 3 272 KIYOYASU MARUI glia cells, which were met more frequently around the blood- vessels than in the other parts of the synapse. Large and old cells showed vacuoles of different size in their bodies. Some of these glia cells showed another kind of granules, which were dem- onstrated by means of the thionin-eosin stain in a characteristic metachromatic or more or less blue-violet color. The granules fill the cell body as well as the processes; the size is variable, but in any one cell they are of almost equal size. The shape of these granules is round or that of irregular lumps. Another charac- teristic of this kind of granules is that in the illumination by elec- tric light they are especially beautifully observable. In the space around the blood-vessels and also in other parts of the synapse I often noticed a group of these granules; this is to be interpreted as the section of a cell or its process, bearing this kind of granules. As far as my observation went, these granules do not lie freely in the tissue. What is the nature of these granules? Reich (27, 28, 29, 30) demonstrated in the Schwann cells of the peripheral nerve fibers rod- or comma-shaped fairly large granules (7-granules), which were brought out in a characteristic metachromatic stain by means of thionin, toluidin-blue or kresyl-violet, and he identified these granules with the protagon of Liebreich on account of the similarity of the staining reaction and of the solubility in warm alcohol (45°) and in warm ether. He found, moreover, that they are soluble also in warm xylol, that they are not at all stainable in acid stains, and also that they are especially beautifully ob- servable in the illumination by electric light. Later, in certain pathological conditions, Alzheimer (1) dem- onstrated in the neurogolia cells granules which gave a char- acteristic metachromatic basophile stain by means of toluidin- blue and thionin; he identified these granules with the 7-granules of Reich and called them metachromatic basophile granules, al- though the granules differed somewhat from the z-granules mor- phologically. Idid not test the properties of solubility of the gran- ules, which I observed in the glia cells; but on the basis of their staining reaction and-morphological characteristics I as- sume that they are identical with the metachromatic basophile. EFFECT OF OVER-ACTIVITY ON SYNAPSE Sie granules of Alzheimer. Whether these granules consist of pro- tagon or not, is another question; recent studies raised doubt against the real existence of protagon as a uniform substance. (Rosenheim, Tebb, Thudicum, cited in (18)). I should add here that I always used bergamot oil instead of xylol in the process of paraffin embedding. Alzheimer (1) declared that he did not find, or he found at the most only indications of, these granules in the amoeboid glia cells; but as far as my observation went, I found a number of amoeboid gha cells with this kind of granules. The finding that these gran- ules are observable in the cells of blood-vessels, as will be de- scribed later, and in the glia cells around the blood-vessels relatively more numerously, and the fact that in Scharlach stain fat drops are found in those cells, as will be related below, make one assume that they are transported toward the blood-vessels and that these granules give rise to the production of fat as Alzheimer did. Reich assumed that the appearance of this kind of granules has a relation to the decay of the myelin sheath; according to Alz- heimer (1), it is not, however, necessary for the appearance of this kind of granules. As the neuroglia cells of the synapse of Mauth- ner’s cell lie mostly at the border of the axone cap, where the nerve fibers lose their myelin sheaths, I cannot decide this question from my own observation. Moreover, as the granules appear only in fatigue, it is probable that they have something to do with a pathological nutrition condition of the nerve tissue; the fact that they are a catabolism product is acceptable because they are transported toward the blood-vessel, and it is also probable that they are changed into fat, just like the other catabolism products. On the basis of the above-described facts, it is quite clear that in fatigue a number of amoeboid glia cells are produced in and about the synapse, which carry different kinds of catabolism prod- ucts in their cell body as well as in their processes. As already repeated, I found these amoeboid glia cells relatively more numerous around the capillaries in and about the synapse. It was also remarked that many a conglomeration of metachromatic basophile granules was demonstrated around the blood-vessels 274 KIYOYASU MARUI with a different .size and shape (fig. 10). This was also the case with the methyl-blue granules and the fuchsinophile granules. All these granules were embedded in the evenly and lightly stained ground substance, which is to be interpreted as the sec- tion of the cell body or the process of the glia cell. As far as my observation reached, these granules did not le free in the space around the blood-vessel or in the tissue. The cells of the adventitia of blood-vessels in the normal brain show very little protoplasm around their nuclei; but in fatigue cases, when a number of amoeboid glia cells are present around the vessels, they show a larger mass of protoplasm. In the thionin- eosin preparation I observed occasionally the metachromatic basophileg ranules in them. Scharlach R (fig.13) revealed many fat drops in the latter. It must be added here that in the neuroglia cells around the vessel and in the synapse fat drops were demonstrated within the protoplasm. According to my obser- vation, there was, however, very little fat lying in the tissue or in the space around the vessel. THE ‘FUELLKOERPERCHEN’ OF ALZHEIMER After the description of the changes of the glious reticulum and the glia cells of the synapse, one more manifestation, which goes hand in hand with the appearance of the amoeboid glia cells is to be mentioned. I have already described that in sections in which a number of typical amoeboid glia cells were observable, many a protoplasm mass of different size and shape appears in the peri- vascular space and in the other parts of the synapse. These pro- toplasm masses were found to be sometimes homogeneous and sometimes granular and they occurred usually near the amoeboid glia cells. Some of these masses must be regarded as the section of the cell body or the process of an amoeboid glia cell (fig.10) ; but some of them show evidently that they were cast off from the body of the amoeboid glia cells (fig. 15). It was also related above that the processes of the glia cells, which are not in possession of the proper attributes of an amoeboid glia cell are broken up in their substance and are reduced to fragments. I assume that I EFFECT OF OVER-ACTIVITY ON SYNAPSE BID can homologize these protoplasm pieces with the ‘Fuellkoerper- chen’ of Alzheimer (1); as far as the origin of these corpusceles is concerned, I agree with Alzheimer, in so far as the reticular beams of glia tissue swell and are loosened in their substance so that’ fi- nally they fall here and there to pieces. The probability of a post- mortem alteration cannot come into consideration; it must be emphasized here that the fish brains were always fixed in their _fresh condition. THE RELATION OF THE AMOEBOID GLIA CELLS TO THE GLIA RETICULUM As stated before, the glia reticulum of the synapse of the Mauth- ner cell in fatigue was found in a more or less advanced state of deviation from its physiological configuration. The net figure appeared less sharp and the meshes were found more irregular; the net beams were observed either swollen or lessened in thick- ness. The substance of the net beams appeared loosened, and in the state of extreme deterioration the net beams were here and there broken up and reduced to fragments so that in some parts of the synapse the net figure was no longer in evidence. The question now arises whether these manifestations of the peri- cellular reticulum are to be attributed to artefacts caused by the process of preparation of the sections or to be claimed as ante- mortem phenomena caused by the over-activity. The follow- ing circumstances speak explicitly for the latter; first, I got the same results in different kinds of preparations; second, I did not find any case of resting control fish, in which a similar picture of the reticulum was observed, and, third, I noticed a number of amoe- boid glia cells with various catabolism products and the so-called ‘Fuellkoerperchen. It must be emphasized here that the nu- clei of the glia cells of the synapse showed not only regressive, but also progressive changes. Buscaino (5), Rosental (31), and others studied the postmortem appearance of the amoeboid glia cells. Wohlwill (35) declared that through the postmortem decay of neuroglia tissue glia cells take occasionally the amoeboid appearance, and the occurrence of the methyl-blue granules and the ‘Fuellkoerperchen’ does not 276 KIYOYASU MARUI always indicate the previous existence of amoeboid glia cells. In the present work all the fish brains, both normal and fatigue, were placed in the fixing solutions in their fresh condition, and the like- lihood of postmortem production of the amoeboid glia cells cannot at all come under consideration. Attention must especially be called to the fact that in all my control preparations not a single amoeboid glia cell did come to my observation in and about the synapse. What would then be the mechanism of the breaking up of the glia reticulum? It is very hard to answer this question definitely. Hisath (10), who studied and demonstrated the protoplasmatic glia structure very well by his own method, suggested that in the sec- tions in which amoeboid glia cells occurred either in the marrow only or in both the marrow and the cortex, the glia cells with deli- cately arborized protoplasm processes do not come to observation. Alzheimer (1) also made the same observation; instead of the glia cells with arborized protoplasm processes, he found by means of the same method glia cells with a little increased plasm with- out any process or those with large cell body carrying different kinds of granules. He took for granted that with the appearance of the amoeboid glia cells the other glia structure also sustain some alteration. On another occasion Alzheimer (1) made the observation that in a cortex, which showed no glia cells with pro- toplasmatic processes by means of Mallory’s method, the. Golgi method revealed such cells with processes. On the ground of this finding, he carefully expressed his opinion, declaring that the processes did not here go into decay or were not withdrawn by the cells, but merely did not show the affinity to Mallory’s hematoxylin. Now, as already remarked, the processes of the glia cells of the synapse arborize and unite into a uniform glia reticulum in the synapse. This condition can be beautifully demonstrated in the Levaditi preparations. In fatigue a number of the glia cells are converted into the amoeboid glia cells and they lie free from the reticular structure of neuroglia tissue; and in this case attention must be called to the fact that some other glia cells in the synapse still show their relation to the reticulum and that I could observe a EEE EFFECT OF OVER-ACTIVITY ON SYNAPSE vk almost every stage of the dissolution of an amoeboid glia cell from the diffuse glia reticulum. Near the amoeboid glia cells the ‘Fuellkoerperchen’ or the fragments of the Golgi net beams were also observable, as already stated. So I came to the conclusion that in the extreme stage of fatigue. a number of amoe- boid glia cells are produced and are set free from the reticulum. Jakob (18) recently described in his article on secondary degenera- tion the detachment of the glia cells from the diffuse glia reticu- lum. Of course this process took place only slightly and on a small scale in my material, but the mechanism would here be the same. At the same time the substance of the reticular struc- ture becomes looser, and finally some of the beams undergo disso- lution and fall to pieces, so that the above-described alteration of the reticulum takes place. What effect the fixing solutions have here on the more or less loosened reticular beams, I can not say; but I believe that the mechanism of the breaking up of the glia reticulum could well be interpreted by thé above- described facts. As far as I know, the studies of the pathological snes of the pericellular retictaltin are very few; the studies of Eisath (10) and Alzheimer (1) were not especially directed toward the peri- cellular reticulum, but merely to the glia reticulum in general. Besta (4) investigated the behavior of the pericellular reticu- lum in certain pathological conditions, but his description does not explain the mechanism of the deterioration clearly enough and he did not mention the appearance of the glia cells at all. As far as the manifestation of the structure in question in fatigue is concerned, this investigation is unique. THE BIOLOGICAL SIGNIFICANCE OF THE AMOEBOID GLIA CELLS AND THE CONCLUSION FROM THE RESULTS | As far as my observation went and so far as the histological ’ technique used brought out the facts, over-activity caused no def- inite change in the nervous structure of the synapse of the Mauth- ner cell. Considering the nature of the experiment, one should not be surprised at this; moreover, the structures in question are so fine and the results of the technique for those structures are 278 * KIYOYASU MARUI sometimes so inconstant that one should be very careful to attrib- ute any pathological significance to any slight histological mani- festation in those structures. The appearance of the amoeboid glia cells may indicate, however, that some catabolism process takes place in and about the Mauthner cell. Alzheimer (1), who investigated thoroughly the amoeboid glia cells and the catabolism processes in the nerve tissue, said about the cases in which a number of amoeboid glia cells were found without any finding on the side of nervous structure, that some catabolism products which escape the microscopical demonstration at the present time would be produced on account of the disturbance of the nutrition of the nerve tissue. The finding of the amoeboid glia cells in the synapse might show that in over-activity some catabolism products are produced as the effect of pathological nutrition condition or owing to dilapidation of the nervous structure, not demonstrable at the present time. These would stimulate the formation of the amoeboid glia cells to serve as scavengers. That postmortem production of the amoeboid gla cells cannot come under consideration, I already remarked. Rosenthal (31), who wanted, to interpret the appearance of amoeboid glia cells with methyl-blue granules as a sign of necro- biosis of neuroglia tissue, regarded the formation of the amoeboid glia cells with fuchsinophile granules as that of increased scay- enging activity; the amoeboid cells with these granules were also found in fatigue, as remarked. According to Wohlwill (34), different kinds of diseases which show the amoeboid glia cells have edema as their common cause. ‘The question whether over-activ- ity causes edema in the region of the synapse and a swelling process of the glia cells can come under consideration or not, must be left undecided here. So far I may assume that in over- activity a catabolism process in a wide sense takes place in the synapse, which comes under the fourth category of catabolism processes of Alzheimer (2). But I cannot state definitely whether ° the catabolism products come only from the synapse or from both the synapse and the cell; the latter appears to me more probable. EFFECT OF OVER-ACTIVITY ON SYNAPSE 279 The question whether the amoeboid glia cells come from newly produced glia cells or from those which were present in the synapse, is very hard to answer; but the finding of many amoeboid glia cells in spite of very little increase of the glia cells may in- dicate that at least some of the old glia cells give rise to amoeboid glia cells. As far as the transition of one kind of catabolism prod- uct into another within the protoplasm of the amoeboid glia ‘cells is concerned, I cannot add much to the description of Alz- heimer (1). The fact that we find the amoeboid glia cells with different granules and lipoid substance relatively more numerous around the blood-vessels than the other parts indicates that the catabolism products are assimilated into different granules by the amoeboid glia cells and carried to the blood-vessels and de- posited in the cells of blood-vessels as fat and later are gradually disposed of. In my experiments I could not find a single case in which the picture of the ‘neuronophagia’ of the Mauthner cell was observed; this might be interpreted to mean that the alteration of the cell body did not go so far in over-activity. Dolley (8) described cell death as the effect of over-activity; it is rather strange to me that no attention was directed to the changes of the glia cells in his article. SUMMARY Careful investigation of the cell body as well as the synapse of the fatigued Mauthner cell in Ameiurus revealed a number of in- teresting findings, which can be summarized as follows: 1. The cell body was found either swollen or shrunken; the turgescence was regarded as the result of over-activity and the shrinkage as that. of exhaustion. 2. The Nissl substance was in a more or less advanced stage of chromatolysis and thereby the cytoplasm was stained vari- ously deeply. On the border of the nucleus a mass of stainable substance was observed as the restitution phenomenon of the Nissl bodies on the side of the nucleus. It was also accepted that a mutual interchange of substance occurs between nucleus and protoplasm in activity. 280 KIYOYASU MARUI 3. The nucleus was also either swollen or shrunken; in the swollen nerve cell it was sometimes removed to one side of the cell body. The nucleolus was also found swollen sometimes and other times shrunken and it was of angular or otherwise irregular shape. . . 4. In the nervous structure of the synapse no definite altera- tion could be brought out; but the synapse showed a number of amoeboid glia cells with methyl-blue, fuchsmophil, and meta- chromatic basophile granules. Also fat drops were demonstrated in the glia cells and in the cells of the blood-vessels. 5. The reticular glia structure of the synapse appeared in many cases of fatigue in more or less advanced deviation from its normal configuration and even broken up in sonte parts; this was interpreted as the result of the detachment of the amoeboid glia cells from the reticulum, as also the effect of the loosening and dissolution of the net beams. 6. The appearance of the amoeboid glia cells showed that some catabolism process occurs in the synapse as the effect of patho- logical nutrition conditions in fatigue. Ld EFFECT OF OVER-ACTIVITY ON SYNAPSE 281 LITERATURE CITED 1 Auzueimer, A. 1910. Beitrige zur Kenntnis der pathologischen Neuro- glia und ihrer Beziehungen zu den Abbauvorgiingen im Nervengewebe. Nissl’s und Alzheimer’s Histologische u. Histopathologische Arbeiten, 3.3. 1913 Uber die Abbauvorgiinge im Nervensystem. Referat auf d. VII Jahresvers. d. Ges. Deutscher Nervenirzte in Breslau, Sept., 1913. Deutsche Zeitschr. f. Nervenheilkunde, Bd. 50, 1914. 3 BartTeLMEzZ, G. W. 1915 Mauthner’s cell and the nucleus mortorius teg- menti. Jour. Comp. Neur., vol. 25. 4 Besta, C. 1910 Sul modo di comportarsi dei plessi nervosi pericellulari in: aleuni processi patologici del tessuto. Riv. d. pat. nerv. e ment., 15. 5 Buscatno, V. M. 1913 Sulla genesi e sul significato delle cellule ameboidi. Riv. di patol. nerv. e. ment. 18. (Referat, Zeitschr. f. ges. Neurol. u. Psycht., Referate u. Ergebnisse 8, 1914.) 6 Dotury, D. H. 1909 The morphological changes in nerve cells resulting from over-work in relation with experimental anemia and shock. Journ. of Med. Research. bo 7 1910 The neurocytological reaction in muscular exertion. Ameri- can Journ. of Physiology, 25. 8 1911 Studies on the recuperation of nerve cells after functional activity from youth to senility. Journ. of Med. Research, 24. 9 1913 The morphology of functional activity in the ganglion cells of the crayfish, Cambarus virilis. Arch. f. Zellforgchung, 9. 10 Ersatu, G. 1911 Weitere Beobachtungen iiber das menschliche Nerven- stiitzgewebe. Arch. f. Psychiatrie u. Nervenkrankheiten, Bd. 48. 11 Eve, F. C. 1896 Sympathetic nerve cells and their basophile constituent in prolonged activity and repose. Journ. of Physiology, vol. 20. 12 GoLDScHEIDER UND Fxiatrau 1898 Normale und Pathologische Anatomie der Nervenzellen. Berlin, 1898. 13 1898 Weitere Beitriige zur Pathologie der Nervenzellen. IV. Mit- teilung. Uber Verainderungen der Nervenzellen bei menschlichem Te- tanus. Fortschritte der Medizin, 1898. 14 Hopaez, C. F. 1892 A microscopical study of changes due to Ainetional ac- tivity in nerve cells. Journ. Morph., vol. 7. 15 1894 A microscopical study of the nerve cell during electrical stim- ulation. Jour. Morph., vol. 9. 16 1895 Changes in ganglion cells from birth to senile death. Obser- vation on man and honey bee. Journ. of Physiology, vol. 18. 17 Hotmeren, F. 1900 Studien in der feineren Anatomie der Nervenzellen. Anatomische Hefte, 15. 18 JaAxos, A. 1912 Uber die feinere Histologie der sekundiren Faserdegenera- tion in der weissen Substanz des Riickenmarks (mit besonderer _ Beriicksichtigung der Abbauvorgiinge). Nissl’s u. Alzheimer’s His- tolog. u. Histopatholog. Arbeiten, Bd. 5 19 Kocurr, R. A. 1916 The effect of activity on the histological structure of nerve cells. Journ. Comp. Neur., vol. 26. 282 KIYOYASU MARUI 20 Lampert 1893 Note sur les modifications produites par l’excitation elec- trique dans les cellules nerveuses des ganglions sympathiques. Soe. de Biolog. (Cited in (32) ). 1 LuGaro Cited in (82). 2 Luxempure 1899 Uber morphologische Verinderungen der Vorderhorn- zellen des Riickenmarks wihrend der Tatigkeit. Neurolog. Central- blatt. 23 Mann 1895 Histological changes induced in sympathetic motor and sen- sory nerve cells by functional activity. Journ. of Anatomy and Phys- lology. 24 Marur, K. 1918 On the finer structure of the synapse of the Mauthner cell ; with especial consideration of the Golgi net of Bethe, the nervous ter- minal feet, and the ‘nervous pericellular terminal net’ of Held. Jour. Comp. Neur., vol. 30. 25 Matures 1898 Riickenmarksbefunde bei zwei Tetanusfiillen. Deutsche Zeitschr. f. Nervenheilkunde. 26 Pucnat 1898 Des modifications histologiques de la cellule nerveuse dans ses divers etats fonctionels. Bibl. Anat. (Cited in (32). 27 Reicu, F. 1905 Uber die feinere Struktur der Zelle des peripheren Nerven. Allgem. Zeitschr. f. Psychiatrie. oR) 28 1907 Diskussion zu dem Vortrage von Alzheimer. Allgem. Zeitschr. f. Psychiatrie. 29 1907 Uber den zelligen Aufbau der Nervenfaser auf Grund mikro- histiochemischer Untersuchungen. Journ. f. Psych. u. Neurolog., Bd. 8. 30 1905 Zur feineren Struktur der Zelle des peripheren Nerven Allg. Zeitschr. f. Psych., 62. 31 RosenrHaL, S. 1913 Experimentelle Studien tiber amoeboide Umwandlung der Neuroglia. Nissl’s u. Alzheimer’s Histolog. u. Histopatholog. Ar- beiten, Bd. 6. Referat: Zeit. f. d. ges. Neurolog. u. Psychiat., Referat u. Ergenbnisse, 8, 1914. 32 Ssonvatyt, E. 1903 Die Nervenzellenverinderungen bei Tetanus und ihre Bedeutung (im Anschluss an einen Fall von menschlichem Tetanus). Jahrbiicher f. Psych., 23. 33 Vas, F. 1892 Studien iiber den Bau des Chromatins in der sympathischen ganglienzelle. Arch. f. mikroskop. Anatomie, Bd. 40. 34 WouLwiLL, F. 1914 Uber amoeboid Glia. Virchow’s Archiv f. patholog. Anat., 216. 35 Woutwit., F. Diskussion zu Alzheimer (2). ' be a rh G H NEDA Ages 5. .- “aed : | ied f ; 4 ye me tee it ars ' ; te + f, ies ' hae Sly a) ware ¢ A ART hie alt ee eee 0) 4) oT ena nT RS ee Waa , 1D of Resumido por C. Judson Herrick, por el autor, O. Van der Stricht. El desarrollo de las células de los pilares, el espacio del ttinel y los espacios de Nuel del 6rgano de Corti. El espacio del ttinel se desarrolla alrededor del fasciculo del nervio espiral, el cual camina entre las porciones nuc!eadas de las células internas y externas de los pilares. En su origen es una hendidura intercelular cuyo contenido liguido es elaborado por el citoplasma vacuolar de las células de los pilares y se vierte en el espacio adyacente. Algunas partes de este protoplasma secretor sufren un proceso de citolisis, de tal modo que la hendidura crece y su contenido liquido aumenta en cantidad a sus expensas. El autor describe con detalle el desarrollo ulterior de las células de los pilares y sus cabezas. El primer espacio de Nuel aparece en forma de una hendidura longitudinal situada entre los pilares externos y las células ciliadas externas y en su interior se acumula el liquido segregado por los pilares externos. En las superficies laterales de estos Ultimos se proyectan vesiculas de seerecciém claras, las cuales experimentan un proceso de citolisis y liquefac- cidn. Las células externas de los pilares verifican na emigracién embrionaria desde la primera fila de células ciliadas externas hacia las células internas de dichos pilares. El autor describe el desarrollo del segundo, tercero y cuarto espacio de Nuel. El contenido liquido del ttinel y el del primer espacio de Nuel se mezclan a través de hendiduras existentes entre los pilares exter- nos y comunican con las de los segundo, tercero y cuarto espacio de Nuel. El liquido de todos los espacios de Nuel esta separado de la endolinfa del conducto coclear por los techos, muy delgados, de estos intersticios. Tales estructuras realizan, indudablemente, la propagacion de las ondas vibratorias desde la membrana basilar hasta la membrana tectoria, contenida en el canal coclear. Translation by José F. Nonidez Columbia University 1: eis “ig elf a & ete Sa eda ethavie: yer te tobeeieotis rn 3 eat eyaeth STI We oH 7 (1% ! lana Pppdetiger goibisnsd' At ia Canidae He dian : Peal iain 0 ; sa TG BS, it: <9 ae leM: ci 3 Teteshig p id ‘onbbhnOht inc oj 1 odcial Abin aul ul Se egauisoes die 12 pes me bas Pagid sl ha ‘ting ACliy ta Seo a aE ee SG ikea ade orale Wt ft sepsis ey babies) pi Xu . a ae alo: ‘otitoonaliiys pert eyes He Ione SUL AT siriayt itis a) ae epei herp aL OUTS OG iy we Ht Sip upg oe utile ete B10? Ser beokeniety LT 4 i “nano bie cel Sun Bebe aint dreary ; be titre Resumido por el autor, Leslie B. Arey. Un mecanismo retinal para la vision eficiente. Las células visuales y el pigmento retinal de muchos vertebra- dos inferiores exhiben sorprendentes movimientos a la luz y en la oscuridad. Los experimentos comunicados previamente han establecido que la rapidez de estos cambios ha sido muy exagera- da. La validez de lasuposici6n, también muy extendida, respecto a su umbral de sensibilidad extremadamente bajo, ha sido in- vestigada después por el autor. Tal determinaci6én es importante en vista de otra suposicién discordante que considera que los cambios en la posicién de las células visuales a la luz intensa y a la difusa favorecen la visién de los conos y bastones, respectiva- mente, mientras que los movimientos correspondientes del pig- mento retinal también aumentan mecdnicamente la eficiencia visual. En otras palabras, si los beneficios reputados como adaptativos se derivan de tales cambios fotomecanicos, las re- acciones a la luz difusa deben ser esencialmente idénticas con las que se sabe ocurren en la oscuridad total. Esta suposicién sin embargo, es en su mayor parte gratuita. Las reacciones de estos elementos bajo la acci6én de la luz de intensidad graduada prueban que el umbral de estimulacién es notablemente elevado. En general, el maximo de reaccién hacia la luz aparece primero en una intensidad luminosa que permite justamente la lectura de los caracteres de imprenta ordinarios. De aqui que la su- puesta sensibilidad fética elevada de las células visuales y pig- mento retinal no queda probada, mientras que las condiciones mecénicas para una visién de la penumbra, tedricamente mas eficiente, se establecen sobre una base experimental. Translation by José F. Nonidez Columbia University AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, MAY 1 A RETINAL MECHANISM OF EFFICIENT VISION! LESLIE B. AREY The Anatomical Laboratory of the Northwestern University Medical School TWO TEXT FIGURES PRELIMINARY It is a well-established fact of retinal physiology that the visual cells and retinal pigment in many of the lower vertebrates exhibit striking movements in light and in darkness (’15 a, 716), although in man and other mammals such changes apparently are slight (’15 b). The pigment cells have processes, probably fixed, which inter- digitate with the visual elements and in which the pigment granules migrate to and fro (figs. 1 and 2). The visual cells likewise modify their positions due to the contractility of the so-called myoid, which is that portion of the inner member of the rod or cone between the ellipsoid and the external limiting membrane. The extensibility of the myoid is variously de- veloped, but in some animals the extremes of change may be as one is to ten (’16). There is represented in figures 1 and 2 the condition of the retina, with respect to these changes, as it is characteristically found in a fish, the common horned pout, Ameiurus. In dark- ness (fig. 1) the pigment withdraws toward the chorioid, thus exposing the visual rods and cones; of the two types of visual elements the cone is extended, whereas the rod is so retracted that its ellipsoid lies close to the external limiting membrane. In light (fig. 2) the appearance is reversed; the pigment now 1 Contribution No. 69, March 3, 1919. The experimental data were obtained at the Fairport Biological Station while a guest of the United States Bureau of Fisheries. 343 344 _LESLIE B. AREY Fig. 1 The effect of total darkness on the position of the retinal pigment and of the visual rods and cones of the fish Ameiurus nebulosus. The pigment is withdrawn toward the chorioid; the cone myoid elongates, the rod myoid shortens. Fig. 2. The effect of bright, diffuse daylight on the same retinal elements of Ameiurus. The pigment moves forward toward the external limiting mem- brane, thereby masking the visual rods and cones. At the right of the figure the positions of these visual cells are indicated—rods elongated, cones shortened. ABBREVIATIONS chr., chorioid; mb.lim.ex., external limiting membrane; my.bac., rod myoid; my.con., cone myoid; pd.cl.pig., base of pigment cell; scl., sclera; st.nl.ex., exter- nal nuclear layer; st.pig., pigment layer. oa se — ee a A RETINAL MECHANISM OF EFFICIENT VISION 345 pushes outward to mask the visual cells, while these latter oceupy mutually reciprocal positions—rods elongated, cones shortened. Such positional changes have been interpreted as of use in furthering efficient vision. It seems logical that the masking pigment of the light phase would serve to protect the delicate visual elements from the overstrong influence of light (Chiarini, 06), while the insinuation of such pigmented processes between the individual visual elements effects for the latter a certain degree of optical isolation by acting as an absorbent of dis- persed rays refracted from neighboring cells (Garten, ’07). From the standpoint of sensory reception, the sharpness of the retinal image is in this way enhanced. The withdrawal of the pigment in dim light might be thought to involve a response which allows the, visual cells to utilize all the weak light available. Furthermore, there is good reason to believe that the cones are concerned with bright-light vision, the rods with dim-light or twilight vision. Hence a shortening of the cone in bright light, drawing it down nearer the source of illumination, while the rod at the same time elongates and is thus moved out of the way, would appear to be a useful maneuver (Herzog, ’05; Exner and . Januschke, ’06). The converse procedure in dim light—by which the rods are shortened and the highly refractile cones, with their lens-like ellipsoids no longer masked by pigment, are length- ened—would be equally advantageous (Garten, ’07; Arey, ’15 a). For the detailed applications of these apparently adaptive responses the reader is referred to Garten (’07), who gives an extended consideration of the correlations within the vertebrate classes between the morphology, optical qualities, and distribu- tional ratios of the rods and cones on the one hand, and, on the other, their movements together with the migrations of the retinal pigment. Interesting and logical as these speculations may be, they nev- ertheless lack a sound experimental basis. It is certain that the movements as summarized in a preceding paragraph occur re- spectively in daylight and in darkness; but since responses in total darkness cannot be useful in the manner suggested, it is obvious that in order to derive the reputed adaptive benefits 346 LESLIE B. AREY from such photomechanical changes, the responses in dim light must be identical with those demonstrably occurring in the dark. This assumption, however, is largely gratuitous. The supposition of an identity between the set of responses ensuing in dim light and in darkness is further complicated by certain conflicting statements and beliefs. There is a general impression current that the retinal pigment and visual cells react to the slightest traces of light; this is reflected in the statements of many workers who have feared their results would be im- paired unless the strictest precautions were observed. On the basis of actual experimentation, however, the earliest observations are encountered in the writings of Angelucci (90), who reported that five minutes of candle light caused the frog’s cones to be highly shortened, whereas the pigment remained as in darkness. On another page, nevertheless, he records that after twenty minutes of twilight the pigment assumed the light position, but the cones are influenced to a less degree! Somewhat later, Pergens (’97) wrote that after a five hours’ exposure to colored lights (red, yellow, green, and blue) of an intensity such that colors could be distinguished by an observer after one minute of dark adaptation, the cones of the white fish, Leuciscus rutilis, are strongly retracted. The weakest response was reported from the blue—a result, however, not in agreement with Herzog’s (’05) later findings on the frog. The latter worker found the blue-violet most effective, although it should be added that the duration of exposure employed by him was only two minutes. In a further communication (’99) Pergens confirmed his earlier conclusion that the pigment migrates least extensively in red light (equal intensities being used), but mod fied his previous belief regarding the inefficiency of the blue to provoke cone retraction. Exner and Januschke (’05) performed some experiments, which, unfortunately, are not trustworthy as evidence. Speci- mens of the fish Abramis brama were exposed during the late afternoon, the experiment continuing through the period of failing ight and terminating at dusk. Examination showed the cones to be in the position characteristic of light. A RETINAL MECHANISM OF EFFICIENT VISION 347 In a few experiments Garten and Weiss (’07) found that light in which colors could not be recognized, acting for five or more hours, caused the cones of the frog to assume a position inter- mediate between that characteristic of bright light and of total darkness. A limited pigment migration was reported as well. These same workers made observations on the fish Abramis brama, contained in a dish lighted by reflection from the cover. Two grades of illumination were chosen: in one colors held within the container could not be distinguished; in the other they were recognizable. According to the results given, in the first grade the cones were maximally retracted in nine retinas and partially soin seven. In the second grade the cones were found shortened in all the eight retinas used, whereas the pigment exhibited no noteworthy change except in the sector constituting the ventral one-third of the eye. The foregoing statements reveal the following conditions: Angelucci’s several pronouncements in the same publication, if not actually contradictory, at least serve to befog the issue. The use of colored lights by Pergens was unfortunate; moreover, the reliability of certain of his conclusions is questionable. Exner and Januschke’s experiments were so ill devised as to furnish no crucial evidence. The results of Garten and Weiss suggest an extreme sensitivity of the cones to low light intensities, whereas the pigment patently has a higher threshold. Finally, there exist no data concerning the responses of the rods to weak light. It is clear that if the visual elements and retinal pigment are as highly sensitive to mere traces of light, as often has been held, they can assume no useful positions in the ordinary dim light of rod vision, while the utility of a response evoked under conditions of virtual darkness will still await an explanation. Accordingly, it was with the intent of learning the true conditions that this investigation was undertaken. 348 LESLIE B. AREY PROCEDURE Essential to success in a determination of this kind is the choice of appropriate experimental animals. Previous experi- ence with a variety of forms led to the selection of two fishes and the frog. The cones of the golden shiner, Abramis crysoleucas, have large, conspicuous ellipsoids and are so highly mobile that the light-adapted myoid can shorten to one-tenth its maximum length in darkness (16). The cones of the common grass frog, Rana pipiens, are easy to observe, but show a narrower range of movement, the limits of extensibility being as one is to four. The rods of the horned pout, Ameiurus nebulosus, not only are of exceptional size in comparison with the usual diminutive ele- ments of fishes, but they also undergo changes in length in the ratio of one to ten (16); the value of Ameiurus for experimenta- tion of this sort cannot be overestimated. There is a further inherent advantage in the animals chosen, inasmuch as their visual cells remain at a uniform level during the characteristic positional changes. The retinal pigment of ail three animals exhibits extensive movements: in darkness it is confined to a narrow zone at the bases of the pigment cells, whereas in the light it migrates nearly to the external limiting membrane. Temperature is an additional factor which must be carefully controlled, although the necessity of this has been recognized only recently (16). In dark-adapted fishes a temperature near the freezing point brings about a maximal contraction of the cone myoid, such as is characteristically associated with the action of light, while raising the temperature to the limit which is compatible with life proportionately elongates the myoid. Light, however, is so much more effective than temperature that the latter factor does not enter as a complication in the light adaptation of cones. The relatively slight quantitative effects of temperature acting in darkness on the position of the rods and on the distribution of pigment may also be disregarded. In the frog, on the contrary, high and low temperatures evoke a maximal expansion from the pigment during dark adaptation, complete contraction being obtainable only at an intermediate grade of about 15°C.; the cones, however, are shortened at the A RETINAL MECHANISM OF EFFICIENT VISION 349 upper temperature limits alone. From these statements it fol- lows that to obtain significant results from experimentation upon frogs a temperature of about 15°C. must be maintained, while with fishes ordinary summer temperature of 22°C., or higher, is favorable. | Experiments were conducted in a large, windowless room into which weak daylight of a non-directive nature could be introduced from a second room; the latter, in turn, received its light directly from windows located at the end far from the single communi- cating door. Animals were tested under five? different conditions of illumination: 1) total darkness; 2) light in which the presence of objects could just be determined; 3) light of an intensity which allows the certain identification of bright colors; 4) light which just permits the reading of ordinary journal print; 5) bright, diffuse daylight. Exposures lasted two or three hours or more. As experience proved, these seemingly rough criteria of light intensity are sufficiently accurate for the purposes required; with a little practice such grades can be kept fairly uniform. Permanent preparations of Perenyi-fixed, paraffin-infiltrated sec- tions served as a basis for study. To further brevity and clearness, the results obtained from experimentation will be condensed to mere summaries. EXPERIMENTATION A. Retinal pigment 1. Frog. In total darkness, and in light of just sufficient strength to allow objects or colors to be discerned, the pigment lies in a narrow stratum near the chorioid. When the illumi- nation is increased just sufficiently to allow the reading of ordi- nary print, the pigment, for the most part, becomes expanded, in some cases completely, in others in a zone only three-fourths the maximal breadth. 2 For Ameiurus another intensity—one which enables ordinary print to be easily read—was also utilized. 350 LESLIE B. AREY 2. Ameiurus. The pigment of this fish is at first more sensi- tive than that of the other animals studied, migration being distinctly initiated in many (or perhaps most) individuals at an intensity by which the presence of objects can be detected. At the next grade (colors distinguished), expansion is well ad- vanced, although the pigment does not extend the maximal dis- tance, for it is dense at the cell bases, but sparse distad. In light in which one can just read, the expansion is nearly com- plete, but does not become maximal until the illumination is sufficient to allow easy reading. 3. Abramis. The effect of light is not apparent until it is of such a strength that colors can be recognized. At this intensity the pigment in about half the individuals was essentially in a condition of greatest contraction; the remainder showed the pig- ment well started, but not extended at most more than half the way to the external limiting membrane. Owing to a sudden failure in my supply of animals at the end of the season, I secured but few observations on the effect of light which makes reading possible; the pigment in those animals studied, however, was in a state of partial expansion only, so that it appears safe to con- clude that stronger illumination is necessary to allow expansion to proceed to completion. B. Visual cells 1. Cones. The cones of Abramis and the frog remain fully elongated, in the characteristic dark positions, until the illumina- tion is so increased that colors are recognized. At this intensity the cones of Abramis show distinct indications of incipient re- traction, while those of the frog are less than half their former length. When illuminated sufficiently to allow reading, the frog’s cones shorten maximally; the cones of the few Abramis studied?’ were likewise greatly shortened, but not completely so. ® As on the tests on the retinal pigment of this animal, further experimenta- tion would have been desirable, but the supply of available material suddenly ceased. A RETINAL MECHANISM OF EFFICIENT VISION 351 2. Rods. Until an intensity is used at- which printed matter can be read, the rods of Ameiurus remain in the typical short- ened condition of darkness; at about this grade, however, they apparently begin to elongate slightly. Any considerable degree of lengthening must first appear only in stronger light. DISCUSSION AND CONCLUSION These results, as compared with the findings of Garten and Weiss (compare p. 347), indicate that there exists a rather lower degree of photic sensitivity in the visual cells and retinal pig- ment than they maintain. Nevertheless, it must be apparent to all, not only that in the subjective choice of arbitrary grades of light a variable personal factor enters, but also that the de- scription of stages thus chosen cannot be expressed in accurate terms. It is possible, of course, for any one person to standardize these grades fairly well for his own experiments; on the other hand, the determination of a critical intensity, such as that in which ‘colors can be distinguished,’ depends on. one’s individual acuity in color discrimination, on the brightness of the test colors, and on the decision as to whether these colors are to be just distinguishable with intense scrutiny or to be identified with ease. In any event, it appears that my first grade of illumination (that in which objects were discernible) was un- doubtedly of lower intensity than the weakest employed by Gar- ten and Weiss' (compare p. 347); it lay far below the point where colors cease to be recognizable, this latter constituting their lowest grade. Moreover, it is not impossible that the intensity of light to which their animals were actually subjected was higher than Garten and Weiss supposed. Their animals, contained in a ‘spacious basin,’ received light (electric) reflected downward from the cover. After the experimenter had accustomed his eye to total darkness for five minutes, the condition of illumina- tion was judged by looking downward into the dish at colors placed directly over the surface of the water. As to whether this is a method which tends toward underestimating the light THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 4 352 LESLIE B. AREY conditions actually obtaining in the basin, the reader may decide for himself. There is one further circumstance which on casual considera- tion might be held responsible for the quantitative divergence between my results and those of Garten and Weiss. They con- tinued their experiments in most cases for five hours, while my determinations lasted on the average perhaps three hours. It seems plausible that the long-continued action of very weak light might register an effect not manifest in shorter periods. That this possibility is not operative in the cases under consid- eration follows from certain other observations of Garten and Weiss. They report maximal cone retraction in the fish after the weakest grade of light employed had acted in one series for three hours and in another crucial series for one and one-half hours. Garten also records that in light too weak to distin- guish colors by, a shortening of the cones occurred (‘eintritt’) in one hour. The facts developed in this investigation may for conve- nience be consolidated into the following statement: although re- sponses of the visual cells and retinal pigment may be initiated at lower intensities of light, an approach to.a maximal response is first elicited at an intensity which permits the reading of ordi- nary print. This signifies that the threshold of stimulation of the visual cells and retinal pigment ts high; or, in other words, the assumed great photic sensitinty of these elements 1s disproved. Furthermore, since the responses in weak light are substantially identical with those in darkness, the mechanical conditions are present for a theoretically more efficient dim-light and bright- light vision, as postulated (compare p. 345). SUMMARY The threshold of stimulation of the visual rods and cones and of the retinal pigment, at which they exhibit their characteristic photomechanical changes, is high. The alleged great sensitivity of these elements to light of ex- tremely low intensity is consequently disproved. . A RETINAL MECHANISM OF EFFICIENT VISION 353 Although responses may be initiated at lower intensities, in general an approach to a maximal light response is first elicited at an intensity which makes ordinary print legible. Since the responses in dim light approximate those in total darkness, the mechanical conditions are present for a theo- retically more efficient dim-light and bright-hght vision than would otherwise obtain. This increased efficiency depends upon the assumption by these elements of correlative advan- tageous positions. LITERATURE CITED Ancetucci, A. 1890 Untersuchungen iiber die Sehthatigkeit der Netzhaut und des Gehirns. Untersuch. zur Naturlehre d. Menschen u. d. Thiere (Moleschott), Bd. 14, Heft 3, S. 231-357. Arey, L.B. 1915a The occurrence and significance of photomechanical changes in the vertebrate retina—an historical survey. Jour. Comp. Neur., vol. 25, no. 6, pp. 535-554. é 1915 b Do movements occur in the visual cells and retinal pigment of man? Science, N.S., vol. 42, no. 1095, pp. 915-916. 1916 The movements in the visual cells and retinal pigment of the lower vertebrates. Jour. Comp. Neur., vol. 26, no. 2, pp. 121-201. Curarini, P. 1906 Changements morphologiques qui se produisent dans la rétine des vertebrés par l’action de la lumiére et de l’obscurité. Deuxieme partie. La retine des reptiles, des oiseux et des mammi- féres. Arch. ital. de Biol., T. 45, fasc. 3, pp. 337-352. Exner, S., unp JanuscHkKE, H. 1905 Das Verhalten des Guanintapetums von Abramis brama gegen Licht und Dunkelheit. Ber. d. k. k. Akad. d. Wissensch. zu Wien, Math.-naturw. Kl., Bd. 114, Abth. 3, Heft 7, S. 693-714. 1906 Die Stibchenwanderung im Auge von Abramis brama bei Lightverinderung. Sitzungsb. d. k. k. Akad. d. Wissensch. zu Wien, Math.-naturw. K1., Bd. 115, Abth. 3, S. 269-280. Garten, S. 1907 Die Verinderungen der Netzhaut durch Licht. Graefe- Saemisch-Handbuch der gesammten Augenheilkunde, Leipzig, Aufl. 2, Bd. 3, Kap. 12, Anhang, 130 S. Garten, S., UND Werss, R. 1907 [Results incorporated in Garten, ’07, pp. 39 and 40.] Herzoc, H. 1905 Experimentelle Untersuchung zur Physiologie der Beweg- ungsvorgiinge in der Netzhaut. Arch. f. Anat. u. Physiol., Physiol. Abth., Jahrg., Heft 5 u. 6, S. 413-464. Percens, E. 1897 Action de la lumiére colorée sur la rétine. Ann. Soc. roy. d. sci. med. et nat. de Bruxelles, T. 6, pp. 1-38. 1899 Vorgiinge in der Netzhaut bei farbiger Belichtung gleicher Intensitait. Zeitschr. f. Augenheilk., Bd. 2, S. 125-141. Resumido por el autor, Swale Vincent. Contribucién al estudio de los reflejos vasomotores. La estimulacién de los nervios sensoriales en perros aneste- siados con étero cloroformo produce generalmente un aumento en los movimientos respiratorios, cuando la narcosis no es muy fuerte o cuando no se emplea el curare. Este aumento de movi- mientos respiratorios produce una disminucién de la presién sanguinea y cuando son pronunciados no se puede obtener ningtin reflejo presor, aun estimulando fuertemente. Con nar- cosis fuerte o compresi6n del cerebro no se produce la disminucién de la presién sanguinea debida a esta causa. La interferencia mecdnica con la circulacién es la causa de esta disminucién, puesto que se elimina abriendo el torax. Este fendmeno complica seriamente los experimentos sobre los reflejos vasomotores. En los perros anestesiados con éter o cloroformo o con el cerebro comprimido, una débil estimulacién de los nervios produce genera- Imente una disminucién, y una fuerte estimulacién un aumento en la presién sanguinea. La frecuencia de la estimulacién ejerce efectos sobre el reflejo; cuando es rapida se obtiene un aumento, y con menor rapidez una disminucién. Los nervios mas volu- minosos responden de un modo mas marcado al estimulo que los menores. La estimulacién de la piel, mtisculos e intestino origina generalmente una disminuci6n en la presién sanguinea, pero si se estimula la piel de un modo violento y extenso se pro- duce un aumento en dicha presién. Bajo la accién de la morfina y el curare, por el contrario, un aumento en la presién tiene lugar generalmente, aunque con la morfina la estimulacién débil puede producir una disminucién de la misma. La influencia de las glindulas endocrinas sobre los reflejos vasomotores no es clara (vedse sin embargo Pearlman and Vincent ‘Endocrin- ology,” en prensa). El] cambio de reflejo en la estimulacién de los nervios somaticos se produce principalmente por los efectos sobre los vasos sanguineos del drea espldcnica. Translation by José F. Nonidez Columbia University ES AUTHOR’S ABSTRACT OF ''HIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICF, JUNE 2 A CONTRIBUTION TO THE STUDY OF VASOMOTOR REFLEXES D. OGATA AND SWALE VINCENT Physiological Laboratory, University of Manitoba, Winnipeg, Canada NINETEEN FIGURES CONTENTS PRETORIA te oe cele eke oon eee te EE See ee mee eee 355 2. The influence of respiratory movements upon blood-pressures............ 357 3. The effect of the strength of the stimulus upon vasomotor reflexes....... 351 4, The influence of the frequency of stimulation upon vasomotor reflexes... 364 5. The effects upon vasomotor reflexes of stimulating nerve trunks of dif- ferent categories (sensory, motor, and mixed nerves) and of different SUAS ON arent cabs cede PEN EE TSS CIEE RATE IN COTES IT ORT Ca Io Ehao a oes fies Oe 366 6. Vasomotor reflexes from nerve terminations..............6.....00e0eeees 370 7. The influence of the ductless glands upon vasomotor reflexes............ 374 8. The question as to which vascular areas are constricted or dilated on central stimulation of somatic merves.....) 20.2680. oe AOS. 375 MESES ISIS AE Yes AMRIT My in Witw' fe bay Pp Se Ae © Tan's te reece Nhe) Ree eR 376 I, INTRODUCTION General blood-pressure is affected reflexly by central stimula- tion of various sensory nerves (reflex vasomotor action). This subject has been studied already by a number of authors. A complete list of the older investigations may be found in Tiger- stedt’s Lehrbuch der Physiologie des Kreislaufs,“! papers by Asher! and Bayliss? in Ergebnisse der Physiologie, and in Nagel’s Handbuch der Physiologie (Hofmann"). The history up to November, 1914, is given by Vincent and Cameron. As to more recent important investigators of this problem, we may refer to Porter,2°—*4 Martin, 22-2 Ranson,**—** Gruber,*? and their respective co-workers, also to Domitrenko® and Hunt.'*:16 Even among these recent investigators there seems to be con- siderable difference of opinion as to what may be regarded as the usual or normal response to afferent impulses. Thus Porter 355 356 D. OGATA AND SWALE VINCENT and Quinby* say: ‘‘It is sometimes urged that in shock the blood-pressure falls instead of rising on stimulation of afferent nerves. This abnormal reaction was observed in several of our experiments.”’ This statement clearly involves the assumption that a rise is the normal effect, though it is recognized that the fall is a not very unusual occurrence. Vincent and Cameron*® seem to be of opinion that the usual effect of stimulating the central end of the cut sciatic nerve is a rise, and the fall due to a pure vasomotor reflex is rather rare. So also Hunt. On the other hand, Martin and Lacey” having observed regularly a definite drop in blood-pressure by weak stimulation and a rise by far more strong stimulation, became doubtful of the truth of the generally accepted doctrine that pressor responses are the normal results of sensory stimulation. These differences of opinion must be due to some factor or factors other than the strength of stimulus. Besides the fac- tors most usually considered, such as different modes of stimu- lation, different nerves, different conditions of the same nerve, different narcotics, drugs, etc., there are two important consid- erations recently brought forward which unmistakably affect the vasomotor reflexes or complicate the problem of their elucidation. In 1915 Vincent and Cameron“ called attention for the first time to a fall of blood-pressure caused by increased respiratory movements. ‘They write: ‘While anaesthesia is fairly complete the effect of stimulating the central end of the cut sciatic nerve is a pure and distinct rise. As the effect of the anaesthetic be- gins to pass off, the effect of stimulation will be a rise of blood- pressure followed by a more or less pronounced fall. Respira- tory movements will now be found to have been markedly in- creased, and the extent of the fall of pressure appears to be at any rate proportional to the violence of the respiratory activity.” Martin and Lacey” investigated the influence of the inter- ruption of the primary current at widely varying rates, but failed to notice any effect, as also did Hunt™ in his earlier work. Only quite recently it was clearly pointed out by Gruber’ that with the same strength of stimulus, pressor and depressor results VASOMOTOR REFLEXES 357 were obtainable by varying the rate of stimulation from 1 to 20 - stimuli per second. This was later incidentally confirmed by Hunt.!* Thus, for the investigation of the complicated problem of vasomotor reflexes, it became very necessary to investigate each possible factor separately. In this way only would one be able to answer correctly for the normal vasomotor response. The present investigation was undertaken for the purpose of studying some of the factors separately, of confirming previous investigations, and of trying, if possible, to reconcile contradic- tory views as to the conditions which determine any particular vasomotor response. We beg to acknowledge our indebtedness to Mr. John Car- michael for his valuable assistance in all our experiments. 2. THE INFLUENCE OF RESPIRATORY MOVEMENTS UPON BLOOD- PRESSURES It is well known that the respiratory center can easily be affected by central stimulation of sensory nerves. Thus Howell'* writes in his Textbook of Physiology that ‘‘stimulation of any of the sensory nerves of the body may affect the rate or the amplitude of the respiratory movements.’ But no mention is made of the influence of these movements upon blood-pressure. The same applies to other text-books, except Starling’s,?* in which we find, ‘‘The increased respiratory movements will also aid the venous circulation and have a similar effect in increasing the systolic output,” which would necessarily bring about a rise of blood-pressure. But, ‘‘A constant and immediate result of exaggerated respiratory movements is a fall of blood-pressure,”’ and not a rise, as Vincent and Cameron pointed out. They found that ‘‘the extent of the fall of pressure appeared to be at any rate largely proportional to the violence of the respiratory activity,” that the fall of blood-pressure was ‘‘brought about by performing rapid artificial respiration by compression of the thorax,” that ‘‘deep voluntary breathing in the case of the human subject produced a regular and pronounced lowering of the blood-pressure,” that ‘‘the more widely the thorax is opened 358 D. OGATA AND SWALE VINCENT the more the fall of pressure tended to become replaced by a rise,’ and that the effect of artificial respiration was ‘‘a rise, and : not a fall, when the animal was under curare,”’ i.e., when a stop was put to the spontaneous respiratory movements. Thus the fall of blood-pressure as a result of increased respiratory move- ments seems to have been sufficiently established. By the majority of previous observers curare was thought to be an indispensable drug in the study of the problem of vasomotor reflexes, with or without any consideration of its action on the vasomotor center itself. But we know that narcotics and other drugs are not always free from influence upon these reflexes, as pointed out by various previous investigators,':9? and, there- fore, in experimental work they should be reduced to as few as possible or altogether eliminated (Vincent and Cameron). The change in character of the respiratory movements, especially their increase, becomes thus an almost unavoidable complica- tion in the study of vasomotor reflexes when curare is not used and the narcosis is not deep enough. If this complication be left out of consideration, erroneous conclusions may be reached. Since the appearance of Vincent and Cameron’s paper several writers have referred to the influence of the increased respiratory movements. Unfortunately, they are not in complete harmony with one another. Ranson and Billingsley****> say, ‘‘ With stronger stimulation the greatly increased respiratory movements _ may no doubt play an important part in the drops in blood- pressure,’ but Gruber and Kretschmer® write that their “‘ex- periments do not support Vincent and Cameroh’s theory that the fall in blood-pressure is brought about by movements of respi- ration which interfere with the heart’s activity.’”’ This latter statement seems to deny definitely the respiratory role upon blood-pressure. Vincent and Cameron did not positively deny that there is a true vasomotor fall of blood-pressure under cer- tain conditions as a result of central stimulation of afferent fibers. But they insisted, and rightly, too, that many apparent vasomotor falls are really due to increased respiratory move- ments. We shall see later that the fall of blood-pressure with weak stimuli is a commoner occurrence than Vincent and VASOMOTOR REFLEXES 359 Cameron were inclined to believe. At any rate, the matter is so importan! that we have repeated the experiments to test the effects of respiration on blood-pressure. _ We have stimulated electrically the central stump of several cut nerve trunks (saphenous, tibial, peroneal, sciatic, ulnar, and median) with various strengths of stimulus. When the narcosis (with ether or chloroform) was not deep enough, the respiratory movements were always increased by strong stimulation. The most frequent response of the blood- pressure to stimulation, e.g., of the sciatic nerve, may be illus- trated by figure 1. With weak stimulation there is practically no increase of res- piratory movements either in amplitude or in frequency, and the blood-pressure is either a fall or a fall followed by a more or less marked rise. When stronger stimuli are applied, the re- spiratory movements increase either in amplitude or in fre- quency, or in both, and the blood-pressure rises, instead of falling, and is followed by a marked fall. The rise of blood- pressure increases usually in proportion to the development of the strength of stimulation. In figure 2 the anesthesia was made much deeper with the same animal as in figure 1, and stimuli of several strengths were applied to the same nerve. There is very little increase of respiratory movements on each stimulation, and the response of blood-pressure is also small in degree. The latter, as seen from the figure, is either a fall or a rise according to the strength of stimulation, and the marked fall after the rise which is observed in figure 1 simultaneously with the increased respiration cannot be seen. This suggests at once that the marked fall accompanied by remarkably in- creased respiratory movements might be ascribed, at any rate, mainly, to the influence of the latter movements caused by sen- sory stimulation. Moreover, under brain compression it is not very difficult to stop the respiratory movements entirely, and in this case only very strong stimulation will initiate spontaneous respiratory movements. Under these conditions a fall of blood- pressure of the same character as that observed with an in- 360 D. OGATA AND SWALE VINCENT ereased respiration never occurs. Thus it would not be un- reasonable to assume that figure 2 shows real vasomotor reflexes, even though weak, not complicated by the increased respiratory movements, while figure 1 represents the vasomotor reflex masked by the effects of increased respiration. It is often very difficult or almost impossible to obtain any rise of blood-pressure when the respiratory movements are very violent. In those cases a marked fall is the only result of cen- tral stimulation of afferent fibers. How these increased respiratory movements affect the blood- pressure was very carefully investigated by Vincent and Cam- eron. After pointing out several possible causes, they came to the conclusion that this fall is due to direct mechanical interfer- ence with the heart’s action and with the return of the blood to the heart. In order to confirm this theory, we opened the thorax in the middle line as did Vincent and Cameron, and found that the falls disappeared. The contrast is clearly shown in figures 3 and 4. In these two cases the same nerve of the same animal was stimulated with the same strength of stimulus, in figure 3 in the intact animal and in figure 4 with thorax open. In addition to opening the thorax we cut both vagi, both phrenici, and as many intercostals as possible on both sides, without obtaining very different results from those obtained by merely opening the thorax. In a very few cases a marked fall of a similar character to that due to the increased respiratory movements, was observed in animals with thorax wide open. It was, however, soon dis- covered that this fall was produced by compression of the in- ferior vena cava by the heart which became more freely movable than before through opening the thorax. The heart fell back upon the soft-walled vein, and thus diminished the flow of blood to the right heart. But it is certainly true that by means of almost pure vaso- motor reflex, i.e., without any or with very little increase of respiratory movements one can obtain a marked fall preceded by a rise, as shown in figure 5. VASOMOTOR REFLEXES 361 Therefore we do not conclude that such a fall of blood-pressure is always produced by increased respiratory movements. But the important point for us at present is the undoubted fact that increased respiratory movements can and do cause a fall of blood-pressure, and that this fall can be easily eliminated by opening the thorax sufficiently wide. These observations, along with various others quoted above from the paper by Vincent and Cameron both on animals and on human subjects, confirm fully their statement as to the occur- rence of a fall of. blood-pressure brought about by increased respiratory movements, and probably explain the nature of this fall. We believe that the increased respiratory movements caused by sensory stimulation form a very important com- plication which has often led to misunderstanding of the true vasomotor reflexes. Gruber and Kretschmer, as mentioned before, deny this re- spiratory effect upon blood-pressure. They used a slow rate of stimulation and the fall of blood-pressure was the usual effect. But the fall is generally thought to be a result of weaker stimu- lation, and they do not deny that the increased respiratory movements to a certain degree cause a fall of blood-pressure when the stimulus is strong enough as to produce them. Our experiments were made on thirty-three dogs. 3. THE EFFECT OF THE STRENGTH OF THE ey UPON VASOMOTOR REFLEXES After repeated experiments by numerous investigators, the generally accepted view as to the effect upon the vasomotor reflexes of different strengths of stimulus seems to coincide with Knoll’s!® original statement, i.e., that a depressor effect is usu- ally the result of a weak stimulation, while a pressor effect fol- lows, as a rule, a stronger stimulation. Reid Hunt" pointed out that weak stimulation was one of the methods of obtaining a reflex fall of blood-pressure, and Vincent and Cameron noticed the same fact. Among more exhaustive investigations on this point we should refer to those by Porter,?°-*4 Martin,”~**4° and their respective 362 D. OGATA AND SWALE VINCENT co-workers. The former writer seems to regard a rise of blood- pressure as the normal vasomotor response, while the latter holds a different view. Martin and. Lacey’s”? experiments were con- ducted on cats either under brain pithing, decerebration or brain compression, or under ether or urethane. The nerves stimulated were the sciatic, radial, median, ulnar, and saphenous. The results of their experiments were very definite. ‘‘In every one of the experiments the stimulation was repeated many times over a range of stimuli from the threshold value to three or four times the threshold. Well-marked drops of pressure followed all such stimulations,’ save in one exceptional case. Thresholds for pressor reflexes were much higher than those for depressor reflexes. Thus the experiments of these workers support Knoll’s statement. Our own experiments consisted in stimulating various nerves (sciatic, tibial, peroneal, saphenous, median, ulnar, and vagus) with induction shocks on dogs under ether, chloroform, and brain compression. As to the method, we have to mention that the different effects of weak and strong currents, respectively, were satisfactorily attained by means of sliding the secondary coil up to or away from the primary, but that on many occasions more than one battery was used to obtain a stronger stimulus. The rate of stimulation was 388 to 54 in a second. The fall of blood-pressure due to the increased respiratory movements being taken into consideration, the main results of our experiments may be summarized as in the following table: FALL WITH Ste Peay eS) SAME SS UNE ASEM AIR NOES RISE WITH AND STRONG errata STRONG STIMULATION STIMULATION Bithvercets crs See St Re ee 20 2 4 Chloroforms: ts. eRe eee eee, 14 4 4 Brain Compression .«.405...¢4.04 00n ee Ve con 12 2 0 Totaly AS ee ok 46 8 8 The term ‘fall’ in the table comprises also a fall followed by a rise and ‘rise’ also a rise followed by a fall. alt VASOMOTOR REFLEXES 363 In forty-six cases out of sixty-two in total, weak stimulation produced a fall or a fall followed by a rise, and strong stimula- tion caused a rise or a rise followed by a fall. A typical response is shown in figure 6. The animal was under ether and the thorax was very wide open in the middle line in order to eliminate the disturbance from increased respiratory movements. Figure 7 shows a similar response under chloroform. In the remaining sixteen cases the response was either a fall or a rise through all strengths of stimulation which we used, and the different effects with weak and strong stimulation were not observable. Thus it does not seem to us unreasonable to conclude that weak stimulation of the central stump of the cut nerve produces usually a fall of blood-pressure and a strong stimulation produces usually a rise. From these conclusions it may naturally be understood that from the threshold of stimulation up to a certain point the fall of blood-pressure increases with the development of the strength of stimulus, and then the fall gradually decreases until a neutral point is reached, where the vasoconstriction and dilatation just counterbalance each other, and finally the rise appears, which increases usually with the increase of the strength of stimulus, but cannot continue very long, since powerful stimuli would elicit vigorous reflex movements of the animal and obscure the true vasomotor reactions unless indeed the animals were deeply under curare. As we have been unable so far to find any at- tempt by previous investigators except Stiles and Martin‘? to describe this rather peculiar course of vasomotor responses, we though it worth while to emphasize it in this place (fig. 8). In our experiments we have employed also stimuli of other kinds than electric induction shocks, namely, mechanical, ther- mal, and chemical. In this series thirty-eight out of sixty-seven stimulations were effective, and of these thirty-five caused a fall of blood-pressure and only three produced a rise. As the cali- bration of these stimuli was not so practicable as with induction shocks, we cannot draw any very positive conclusions, but we 364 D. OGATA AND SWALE VINCENT are inclined to believe that a greater number of pressor re- sponses could be obtained if we could improve the method of stimulation so that the sensory fibers might be stimulated more strongly. It thus appears from our experiments that the depressor effect of weak stimuli is much more common than Vincent and Cam- eron thought, though these observers were careful not to deny its occurrence. Reid Hunt,'* in a recent paper, seems to have had the same difficulty that Vincent and Cameron encountered in obtaining the depressor effect of weak stimulation, which he ascribed to the different frequency of stimulation they employed. The fact that a weak stimulation of a sensory nerve causes, as a rule, a reflex fall of blood-pressure and a strong stimulation a reflex rise, together with the statement of Bayliss* that the orthodox effect due to the stimulation of the depressor nerve (nerve of Cyon®) can be converted into a rise by the action of strychnine, led us to inquire whether a pressor response could be obtained by strong stimulation of the depressor nerve. So far as our results inform us, neither such a strong current as would injure the nerve nor inductién shocks up to eighty per second frequency could reverse the depressor response. The response to the stimulation after injection of strychnine was sometimes increased and sometimes decreased, but the reversal of the re- sponse did not appear in our experiments even with a dose which caused general convulsions on weak stimulation. 4, THE INFLUENCE OF THE FREQUENCY OF, STIMULATION UPON VASOMOTOR REFLEXES That the frequency of stimulation has a certain effect upon vasomotor reactions seems to have been known to the older in- vestigators. In 1883 Kronecker and Nicolaides” noticed that the vasomotor centers could more easily be affected by changing the frequency of stimulation than by changing its strength. They write: “‘One can never attain such a strong vasoconstriction by increasing the intensity of the stimulating current as by increas- ing the frequency of the current of moderate ‘ntensity.”’ We have not been able to consult the original paper of these writers. —— - VASOMOTOR REFLEXES 365 From a reference in Ergebnisse der Physiologie by Asher,! it is not clear whether the stimulation was directly upon the nerve centers or reflexly through afferent nerves. — But the credit of pointing out clearly that the frequency of stimulation has an effect upon vasomotor reflexes must be ascribed to Gruber.’ This writer remarks: ‘‘That summation takes place with rapid rates of stimulation is undisputable, but it does not seem probable where the strength is more than 400 times threshold that the phenomenon of summation can explain the different effect obtained with these rates of 1 per two seconds and 20 per second interruptions.’’ The similar effect of fre- quency of stimulation was afterward proved incidentally by Reid Hunt,!* who considers it convenient to use the infrequent rate of stimulus to obtain a reflex fall of blood-pressure. Our experiments on this subject have been carried out on dogs with rates of stimuli of 1, 2, 5, 10, 20, 40, and 80 per sec- ond upon various nerves, under chloroform and curare or under brain compression. ‘Though our results were not so conclusive as those obtained by Gruber (in fifteen out of forty stimulations similar results to his were obtained), still we do not hesitate to ascribe an important rdle to the frequency of stimulation. Ac- cording to Martin’s” investigations, the intensity of stimula- tion in Z-units is directly proportional to that of the current in the primary circuit. We arranged the apparatus in such a way as to get a current of a certain strength and one ten times stronger as we desired. With the former current we obtained a fall by stimulating five times per second, and a distinct rise by stimulating ten times per second, while with the latter current we observed a fall with the rate of stimulus one per second and a rise with five per second stimulations. A selected record is shown in figure 9, where one and the same nerve was stimulated with the same intensity but with different frequency. Much more remarkable were the rates of stimulation at which the maximum pressor response was reached. 366 D. OGATA AND SWALE VINCENT 7 RATE OF STIMULI PER SECOND Number of experiments whose maximum pres- sor response was reached at the rate of stimuli mentioned above......0......5..0.....1 O} OF Of 4 138" Mies As is seen from the table, in 78.9 per cent the maximum re- sponse is reached between twenty to forty per second stimula- tion, and in one-third at the rate of forty per second. Beyond these points the effect increased only in four cases. This phe- nomenon may be seen a!so in figure 9. Kronecker and Nicolaides?® observed the fact that the effect of stimulation of the vasomotor centers increased with the fre- quency of stimuli up to twenty to thirty per second, but not beyond this point. Tur” also pointed out that the effect of stimulation of the lingual nerve increased until the stimuli reached forty per second, beyond which, however, the effect diminished. These observations coincide fairly well with our own. , 5. EFFECTS UPON VASOMOTOR REFLEXES OF STIMULATING NERVE TRUNKS OF DIFFERENT CATEGORIES (SENSORY, MOTOR, AND MIXED NERVES) AND OF DIFFERENT SIZES According to the investigations of some authors, different nerves, apart altogether from the depressor nerve, respond dif- ferently to central stimulation. Hofmann," in Nagel’s Hand- buch, says: ‘‘There are single nerves, which for the most part (glossopharyngeal) are depressor, and others which are exclu- sively (splanchnic) or preponderatingly (sciatic, facial, infra- orbital, cervical nerves) pressor.’’ Vincent and Cameron studied the effect of stimulating the main trunk of the sciatic, as well as its common peroneal, lateral cutaneous, and purely muscular branches, the saphenous, median of axilla, the hypoglossal, the glossopharyngeal, the superior laryngeal, and the vagus. But the different nerves all produced similar or comparable results on the blood-pressure. They were strongly tempted to the hypothesis that an equivalent stimulation of a roughly equal number of afferent fibers will yield similar reflexes. ~~ ee ee eee he a deal ~ VASOMOTOR REFLEXES 367 Our experiments also have led to the conclusion that there is no essential qualitative difference between the various nerves subjected to stimulation (sciatic, tibial, peroneal, median, ulnar). A possible exception may be made in the case of the saphenous. It may be that there is a greater tendency to a fall on stimu- lating this nerve than in the case of others. Whatever the nerve may be, whether a purely sensory nerve, as the saphenous; a mixed nerve, as the sciatic, peroneal, tibial, ulnar, or median, or a purely motor nerve, such as a muscular branch of the fe- moral, the course of response to weak and strong stimulation was in most cases a fall and a rise as already described in section 3. A few examples are quoted in tabular form where a selected purely sensory nerve and a purely motor or a mixed nerve ap- parently of the same size were stimulated in turn under entirely similar conditions. Sensory and motor nerves compared BRANCH OF RIGHT FEMORALIS ARRANGEMENTS OF STIMULATION RIGHT SAPHENOUS 1 battery coil at 30 em. 15 seconds...) —4 mm. Hg. 0, mm. Hg. 1 battery coil at 25 em. 15 seconds...| —10 mm. Hg,| —1, +2 mm. Hg. 1 battery coil at 20 cm. 15 seconds...| —10, +2 mm.Hg.| —4, +2 mm. Hg. 1 battery coil at 15 em. 15 seconds...| +4, —12 mm. Hg.| +4, 0mm. Hg. 1 battery coil at 10 em. 15 seconds...| +6, —14mm.Hg.| +8, —2 mm. Hg. 1 battery coil at 5 em. 15 seconds...| +16, —18 mm. Hg. | +14, —12 mm. Hg. 1 battery coil at 0 cm. 15 seconds...| +22, —14 mm. Hg. | +26, —14 mm. Hg. (From experiment 47. Under brain compression.) ; The sign ‘—’ means a pure fall of blood-pressure, ‘+’ a pure rise, and ‘—, +’ or ‘+, —’ a mixed response, namely, a fall followed by a rise or a rise followed by a fall, respectively. Sensory and mixed nerves compared ARRANGEMENTS OF STIMULATION RIGHT SAPHENOUS RIGHT PERONEAL 1 battery coil at 30 em. 15 seconds... 0 mm. Hg. 0 mm. Hg. 1 battery coil at 25 em. 15 seconds...) —6, O:mm. Hg.|.=—1, 0 mm. Hg. 1 battery coil at 20 em. 15 seconds...| +2, —10 mm.Hg.| —4, +4 mm. Hg. 1 battery coil at 15 em. 15 seconds...| +6, —12 mm. Hg.| +2, —6 mm. Hg. 1 battery coil at 10 em. 15 seconds...| +8, —12 mm. Hg. | +14, —10 mm. Hg. 1 battery coil at 5 em. 15 seconds...| +22, —10 mm. Hg. | +20, —14 mm. Hg. 1 battery coil at 0 em. 15 seconds. ..| +22, —12 mm. Hg. | +22, —20 mm. Hg. (From experiment 47. Under brain compression.) THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 4 368 D. OGATA AND SWALE VINCENT With the increase of the strength of stimulus the respiratory movements also increased, though not very markedly, and therefore some of the falls following the rise on stronger stimu- lation might have been more or less due to this complication. But in the main it seems that the purely sensory nerves have somewhat lower threshold than other kinds of nerves (fig. 10). Whether this is due to a large number of afferent fibers con- ‘tained in the sensory nerve than in those of the other kinds of the same size could only be decided by more numerous experi- ments and more elaborate methods than those we have employed, as, for example, the measurement of resistance of each nerve and more satisfactory methods of controlling the intensity of stimulation in each case. in connection with the problem as to different kinds of nerves we have studied the influence of the size of the nerve upon vaso- motor reflexes. The hypothesis of Vincent and Cameron is quoted at the beginning of this section. A similar problem was taken up also by Stiles and Martin,‘? who compared the effect of stimulating two nerve paths at the same time with that of exciting each. by itself. They found that “stimulation of two afferent paths at the same time has often a more marked vaso- motor effect than the stimulation of either path alone with an equivalent strength of current. The degree of summation was only moderate.’’ This shows that the stimulation of a larger number of afferent fibers will produce often a more marked effect than that of few fibers. We stimulated two nerves of the same category but of different sizes separately one after another under conditions as similar as possible, a different number of afferent fibers being assumed to be present in the nerves of different sizes. The results may be represented as follows, page 369. These few examples show that the results were not very con- clusive. We can say only so far with some confidence that when the responses were in the same sense, i.e., when the fall or the rise was the result of corresponding equivalent stimulations, the reflex change of blood-pressure was on the whole more marked with the nerve of larger size than with those of smaller size (fig. 11). Oi ia la ae alice i eal Mixed nerves compared with each other VASOMOTOR REFLEXES ARRANGEMENTS OF STIMULATION 1 battery coil at 30 cm. 1 battery coil at 25 cm. 1 battery coil at 20 cm. 1 battery coil at 15 em. 1 battery coil at 10 cm. 1 battery coil at 25 cm. 1 battery coil at 30 cm. 1 battery coil at 15 cm. 1 battery coil at 10 cm. (From experiment 47. Sensory nerves compared with each other 15 seconds. .. 15 seconds. .. 15 seconds. .. 15 seconds. .. 15 seconds... 15 seconds... 15, seconds... 15 seconds... 15 seconds... RIGHT SCIATIC 0 ba —14, mm. Hg. 4 mm. Hg. 0 mm. Hg. 12, —22 mm. Hg. 18, —16 mm. Hg. RIGHT SCIATIC —2, —6, 369 RIGHT PERONEAL 2 mm. Hg. 4 mm. Hg. 16, —18 mm. Hg. 22, —22 mm. Hg. Under brain compression.) ARRANGEMENTS OF STIMULATION 1 battery coil at 25 cm. 1 battery coil at 20 cm. 1 battery coil at 15 cm. 1 battery coil at 10 cm. 1 battery coil at 5 cm. 1 battery coil at O cm. 1 battery coil at 25 cm. 1 battery coil at 20 cm. 1 battery coil at 15 cm. 1 battery coil at 10 cm. 1 battery coil at 5 em. 1 battery coil at O cm. (From experiment 48. 10 seconds...| 10 seconds... 10 seconds... 10 seconds... 10 seconds... 10 seconds... 10 seconds... 10 seconds... 10 seconds... 10 seconds... 10 seconds... 10 seconds... RIGHT SAPHENOUS 0 5) =18, iho e118; +4, mm. Hg. 0 mm. Hg. 0 mm. Hg. 0 mm. Hg. 0 mm. Hg. 0 mm. Hg. RIGHT SAPHENOUS 0 a se) =a +4, +2, mm. Hg. 0 mm. Hg. ,0 mm. Hg. 0 mm. Hg. 0 mm. Hg. 0 mm. Hg. Under brain compression.) 0 mm. —10, 0 mm. —8, 4 mm. 8, —6 mm. 12, —10 mm. RIGHT TIBIAL —2, 2mm. —2, 0 mm. 10, —8 mm. 20, —16 mm. SAPHENOUS A BRANCH OF THE RIGHT 0 mm. Hg. es Lg. +2, =" 0 mm. Hg. 0 mm. Hg. 0 mm. Hg. 0 mm. Hg. A BRANCH OF THE RIGHT SAPHENOUS These conclusions coincide with the experience of Stiles and Martin and lend some support to the hypothesis of Vincent and Cameron. It may not be amiss to add to these conclusions that in few cases when the stimuli were very strong the smaller nerve re- acted more vigorously than the larger one, which phenomenon may perhaps be explained partly by different resistances of dif- ferent-sized nerves to currents of similar strength. 370 D. OGATA AND SWALE VINCENT Thus, so far as our experiments go, we are inclined to conclude provisionally that among nerves of different categories there are no essential qualitative differences of response, and the greater the number of afferent fibers stimulated, the more marked is the response of blood-pressure within a limited range of strength of stimulation. 6. VASOMOTOR REFLEXES FROM NERVE TERMINATIONS Several investigators have stimulated the nerve terminals instead of the nerve trunk itself. When we apply a stimulus to a surface such as the skin we should bear in mind that we may actually be stimulating either the end-organs alone or these structures as well as the nerve fibers, according to the mode of stimulation. Any physiologi- cally appropriate stimulus, though mild, applied to the end- organs would give rise to more highly effective impulses than inappropriate ones. Thus the study of vasomotor reflexes in response to stimulation of the sense organ with its most ap- propriate stimulus is highly desirable. But even with other kinds of stimulation we may learn much that is valuable, be- cause any stimulus which plays a part in our normal daily life comes usually through the end-organs on the outer or inner surface of the body, and not by way of exposed nerve trunks, as in the foregoing experiments. The skin, the mucous membrane of the nose, muscles, the in- testine, and other abdominal organs were employed frequently by previous investigators. We have selected the skin, muscles, and the intestine as representative of regions containing differ- ent modes of nerve endings. The stimuli used were mechanical (incision, scratching, pinching, kneading), thermal (hot or boil- ing water and cold water or lumps of ice), chemical (10 per cent solution of sulphuric acid), and electrical (induction shocks of various strengths). The animals (dogs) were under ether, chloroform, or brain compression. The results of a first series are presented in the following tables: ~ VASOMOTOR REFLEXES Mechanical stimulation of the skin - : STIMULATED r ANESTHESIA Scere MODE OF STIMULATION L. inn. thigh | Pinching ithersss.i-. ages: | R. inn. thigh | Scratching R. inn. thigh | Incision | L. inn. thigh | Pinching R. inn. thigh | Scratching | Abdomen Chloroform....... Over saphen- | Incision . ous, ulnar, | and sciatic [| nerves Brain compres- { L. inn. thigh | Scratching Gl ey Se eee 8 R. inn. thigh | Incision SNAEPCRR AEN: 5s SU Sos sts cus ORE e SoMee Ques eee via See Thermal stimulation of the skin L. inn. thigh | 65°C.—boiling water i A aie { L. inn. thigh | Ice—15°C. water L. inn. thigh | Boiling water Chloroform....... { meee jing ice Brain compres- L. inn. thigh | Boiling water BLODLEE satiate. L. inn. thigh | Ice POMEL See Sch ath ich Reta Bini Sys oie 2 ate alk sath Sita cues (ark ARs hin 0 0s Electrical stimulation of the skin UES cad ote separ L. inn. thigh | Strong induction shocks Chloroform......... L. inn. thigh | Strong induction shocks Brain compression..| L. inn. thigh | Strong induction shocks 371 REFLEX RESPONSE OF BLOOD-PRESSURE chet | Fall | Rise 0 2 0 2 5 0 0 6 0 1 1 0 2 3 0 3 4 0 0 2 0 0 3 0 8 | 26 0 0 5 0 3 0 0 2 3 0 0 1 0 0 1 1 2 0 0 Fie a) 1 8 5 0 4 2 0 4 2 0 8 4 D. OGATA AND SWALE VINCENT As is clear from the tables, almost every stimulation in this series, produced a reflex fall of blood-pressure, and no signifi- cant qualitative difference is observable either with different modes of stimulation or with different methods of anaesthesia or with different portions of the skin. That this statement is applicable almost without any modifi- cation to the results of stimulation of muscles and intestine will immediately be understood from the tables on page 373. Thus it is fairly clear that the stimulation of nerve terminals in the skin, muscles, and the intestine produces usually a reflex fall of blood-pressure, as was reported by Vincent and Cameron. But the threshold of stimulation for the nerve-terminals in the skin is very much higher than that for the exposed nerve- trunks. Thus a stimulus which is to be reckoned a strong one for the exposed nerve-trunk is to be considered a weak one for the surface of this skin. This fact explains the previously de- scribed results. So far the effects have all been those of a weak stimulation, namely, a fall of the blood-pressure. If, now, we take steps to secure a considerably greater amount of stimulation by simultaneous scratching of large areas in dif- ferent regions, it is not difficult to satisfy oneself that the same general law applies for the nerve-terminals as for one exposed nerve-trunk. Thus, if we scratch a limited area with a mod- erate degree of vigor, we get a fall, while more violent applica- tion of the instruments to a large area, will give a rise (fig. 19). In the last section we compared the effects of stimulating two nerves of the same category but of different sizes, and showed that the nerves of greater size usually surpass those of smaller size in their power of evoking vasomotor reflexes, and referred to Vincent and Cameron’s hypothesis that the number of af- ferent nerve fibers is an important factor. In our stimulation of nerve endings, as a rule, we could only apply the stimulations to a small portion of the surface. Now the nerve fibers spread widely from the nerve trunk, and the stimulation of a nerve would be equivalent to the stimulation of the entire surface to which the nerve is distributed. In other words, the stimula- tion of a small portion, e.g., of the skin, corresponds to that of a ee ee rr SS ANESTHESIA Chloroform....... Brain compres- BION: ELLER oe VASOMOTOR REFLEXES Mechanical stimulation of muscles STIMULATED MUSCLE R. add. mag. R. sartor. R. sartor. L. semitend. R MODE OF STIMULATION Scratching Scratching Kneading Scratching . add. mag. or 1 semitend. | Scratching Stimulation of the intestine Small intestine Interior surface intestine Interior surface intestine Interior surface intestine Small intestine Small intestine Small intestine Interior surface intestine Interior surface intestine Interior surface intestine Interior surface intestine of of of of of of of small small small small small small small Kneading Induction shocks Pinching ' Boiling water ap- plied Distension Kneading Kneading Induction shocks Scratching Boiling water ap- plied Ice piece applied 373 REFLEX CHANGE OF BLOOD- PRESSURE 0| 1,0 0| 2)0 0} 3/0 3| 0,0 4/ 0|0 7 | 15) 0 0; 3,0 1) 01.0 0; 1,0 £7). 28 0; 40 0; 21 0; 4,0 Of 22 0/ 1/0 0; 1,0 DA iaied 4 2 | 20) 3 374 D. OGATA AND SWALE VINCENT _ small number of sensory nerve fibers. If the stimulation of a few fibers be the equivalent of a weak current, the fall of blood- pressure caused by stimulating the nerve terminals may be ascribed to the fact that we are stimulating only a few fibers. Under morphia and curare a rise of blood-pressure is more easily obtained than a fall on stimulation of nerve-endings. But under morphia at any rate it is not difficult to obtain a rise with a strong stimulus and a fall with a weak one (fig. 19). Gaskell’s’? discovery that in mammals ‘‘A large dose of curare will remove both the contraction of the muscle and the dilatation of its blood-vessels upon stimulation of the nerve,’’ may possi- bly account for the greater tendency towards a rise when the animal is under this drug. The paralysis of the vasodilator nerves by curare seems to necessitate the taking of certain precautions in interpretation of the results of experiments. It seems probable that if it were found possible to increase very considerably the energy and extent of the stimulation in the cases of kneading of muscle and of the intestine, we should have to record a rise of pressure instead of the fall with which we are familiar. 7. THE INFLUENCE OF THE DUCTLESS GLANDS UPON VASOMOTOR REFLEXES The extracts of some of the ductless glands (adrenal body, thyroid, and pituitary) have been alleged to affect the vaso- motor irritability on one way or another.!7:43.12.27.28 Since the results of the previous investigations are not conclusive, we thought it might be worth while to investigate the matter again. It is to be feared that our experiments are not much more convincing than those of previous workers on this subject. The change of blood-pressure (augmentation or diminution) due to the injection of the extracts of these glands is an undesir- able complication. In cases where an augmented blood-pres- sure is the result, as with adrenin and pituitrin, the decreased pressor reaction to the stimulation of a nerve may most properly be ascribed to the diminished response of the already more or VASOMOTOR REFLEXES 375 less contracted blood-vessels, or possibly to the additional con- traction of the blood-vessels of the small areas other than those previously affected by the drug. But the comparison of the vasomotor reflexes before and after the injection of adrenin (‘‘adrenalin” Parke, Davis & Co.) seems to show that the pressor reflex is slightly decreased in the latter case. The results of Hoskins and Rowley were similar and more definite. The elimination of the function of the supra- renal glands by tying them off gave no clear results.* The injection of thyroidin (Parke, Davis & Co.) and the extir- pation of both thyroid glands do not appear to have any distinct influence upon vasomotor reflexes. Pituitrin (Parke, Davis & Co., surgical) showed scarcely any significant results. All these experiments were performed on dogs under brain compression for the purpose of excluding any influences from the increased respiratory movements and those of anaesthetics and other drugs. 8. THE QUESTION AS TO WHICH VASCULAR AREAS ARE CON- STRICTED OR DILATED ON CENTRAL STIMULATION OF SOMATIC NERVES The fall of blood-pressure produced by stimulation of the de- pressor nerve is effected chiefly by dilatation of the splanchnic area,” though, as Bayliss has shown, the vessels of the limbs, head, and neck also partake in the relaxation. The latter writer showed also that the rise of blood-pressure on stimula- tion of the central stump of the splanchnic (?) nerve was, for the most part, due to the constriction in the splanchnic area. The reflex rise of blood-pressure due to the stimulation of the *That is to say, when the nerves to the limb are intact. In the denervated limb there is a very important difference according to whether or no the ad- renal bodies are eliminated. Mr. Pearlman and myself have recently found that when the central end of the sciatic is stimulated in such a way as to give a pressor response the intact limb follows passively the blood-pressure while the denervated limb constricts. After removal of the adrenal bodies the dener- vated limb also dilates. These results are explained more fully in a paper about to be published in ‘Endocrinology.’—S. V. 376 D. OGATA AND SWALE VINCENT sensory nerves of the skin, too, depends mainly on the constric- tion of the blood-vessels in the same area,** and Hofmann" writes: ‘‘The rise of blood-pressure on stimulation of sensory nerves is produced by the constriction of the blood-vessels of abdominal organs as in asphyxia. At the same time the blood- vessels of the brain, skin, and muscles dilate, as a rule, and an increase of the volume of limbs takes place.”” Thus the splanch- nic area plays a principal part in the reflex changes of blood- pressure on stimulation of the somatic as well as the splanchnic nerves. We have made some experiments on this point, and can con- firm the above statements. The dogs had both vagi cut and were under morphia and curare or brain compression, and the sciatic or saphenous nerve was stimulated with induction shocks. The volume changes of the limbs (hind and fore) and of the abdominal organs (small intestine, kidney, and spleen) were recorded. The rise of blood-pressure, when sufficiently high, was always accompanied by a remarkable diminution of the volumes of abdominal organs and a pronounced dilatation of limbs (fig. 18). The pronounced fall of blood-pressure with weak stimulation when the animal was under brain compression was seen to be accompanied by a distinct increase of the volume of the intes- tine. Thus it appears clear that a reflex rise and fall of blood- pressure on stimulation of a somatic nerve (sciatic, saphenous) is brought about chiefly by constriction and dilatation of the blood-vessels in the splanchnic area. 9. SUMMARY 1. In dogs under ether or chloroform, stimulation of sensory nerves (saphenous, tibial, peroneal, sciatic, ulnar, and median) causes usually increased respiratory movements when narcosis is not profound or curare is not employed. These increased move- ments produce a fall of blood-pressure, and when they are very violent, one cannot obtain any pressor reflex even with a strong stimulation. When much increased respiratory movements are prevented by very deep narcosis or brain compression, fall of VASOMOTOR REFLEXES Yi | blood-pressure due to this cause does not occur. Mechanical interference with the circulation as a result of the increased movements of the thoracic walls seems to be the main cause of this fall, since it can be eliminated by opening the thorax. When this complication is not taken into careful consid- eration the results of vasomotor experiments are liable to be misinterpreted. 2. In dogs under ether, chloroform, or brain compression, a weak stimulation of the central end of the cut nerves (sciatic, saphenous, tibial, peroneal, median, ulnar, and vagus) pro- duces usually a fall and a strong stimulation, a rise of blood- pressure. With a gradual increase of the strength of stimulus up from the threshold, the reflex fall of blood-pressure first increases, then decreases, and gradually becomes converted into a rise, passing through a neutral point. We have failed to obtain a pressor effect by the strongest stimulation of the de- pressor nerve of Cyon. 3. The frequency of stimulation has an effect upon vasomotor reflexes. With a rapid rate of stimulation a rise is obtained and with a slow rate of stimulation in many cases a fall of blood- pressure. Of the different rates of stimulation we employed (one to eighty per second), the maximum pressor response is reached at twenty to forty per second. 4. No essential qualitative difference was found among vari- ous nerves (sciatic, tibial, peroneal, median, ulnar, branch of femoral nerve) subjected to stimulation. The saphenous nerve has a greater tendency to give a fall than those above mentioned. A purely sensory nerve seems to have a somewhat lower thresh- old than other kinds of nerves. Between nerves of the same category but of different sizes, the larger one produces usually a more marked response within a limited range of the strength of stimulation. 5. When the animal is under ether, chloroform, or brain com- pression, stimulation (mechanical, thermal, chemical, and elec- trical) of nerve terminations, such as those in the skin, muscles, and the intestine, causes a fall of blood-pressure in the great majority of cases, but violent or extensive stimulations of the 378 D. OGATA AND SWALE VINCENT skin produce a rise. Under morphia and curare, on the con- trary, a rise is a usual response, due clearly to a specific pharmaco- dynamical influence of these drugs. But under morphia a weak stimulus will produce a fall. 6. The influence of the ductless glands (adrenal, thyroid, and pituitary) upon vasomotor reflexes is not clear.* The injection of adrenin, thyroidin, and pituitrin and tying off or extirpation of the glands produced in our experiments no distinct effect. 7. The reflex change (fall or rise) of blood-pressure on stimu- lation of the somatic nerves (sciatic, saphenous) is produced chiefly by the dilatation or constriction of the blood-vessels in the splanchnic area, as in the cases of the stimulation of the splanchnic and the depressor nerve. *See footnote, page 375. Oowrwnds © oO NI 10 it 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 VASOMOTOR REFLEXES 379 BIBLIOGRAPHY Asner, L. 1902 Ergebnisse der Physiologie, Jg. 1, Abt. II. Bayuiss, W. M. 1906 Ergebnisse der Physiologie Jg. 5, S. 319. 1908 Proc. Roy. Soc., B., vol. 80, p. 353. BRUNTON AND TUNICLIFFE 1894 Journ. Physiol., vol. 7, p. 373. Cron, E. pe 1900 In Richet’s Dictionnaire de Physiologie, p. 774. DomitrrEenKo, L. F. 1912 Dissert., Odesse, p. 312. (Physiol. Abstracts, 1917, vol. 2, p. 30). GaskELL, W. H. 1916 The involuntary nervous system. p. 90. Grouper, C.M. 1907 Amer. Journ. Physiol., vol. 42, p. 214. GruBer, C. M., anp Kretscumer, O. 8S. 1918 Amer. Jour. Physiol., vol. 46, p. 222. GrUTzNER AND HripENHAIN 1878 Arch. f. d. gesammt. Physiol., Bd. 16, (quoted from Stiles and Martin) .*4 Hormann, F.B. 1909 In Nagel’s Handbuch der Physiologie des Menschen, Boel Sais: Hoskins, R. G., anD Rowtey, W.N. 1915 Amer. Journ. Physiol., vol. 37, fie ak 7A le Howe..t, W. H. 1915 Text-book of physiology, p. 688. Hunt, R. 1895 Journ. Physiol., vol. 18, p. 406. 1918 Amer. Journ. Physiol., vol. 45, p. 197. 1918 Amer. Journ. Physiol., vol. 45, p. 231. Kepinow 1912 Arch. f. exper. Pathol., Bd. 67, 8. 247 (Zentralbl. f. Physiol., 1913, Bd. 27, S. 129). KLEEN, 1887 Skand. Arch. f. Physiol., Bd. 1, 8. 247 (quoted from Reid Hunt). Kwnoitu 1885 Sitz. Acad. Wiss., Wien, Math. Naturur. Kl., Bd. 92 (iii), S. 448 (quoted from Reid Hunt). ’ Kronecker, H., anp Nicotarpes, R. 1883 Du Bois-Reymonds Arch. (quoted from L. Asher ).? Lupwia anp Cyon 1866 Ber. d. Siichs. ges. d. Wissenschaften (quoted from Bayliss, Journ. Physiol., 1893, vol. 14, p. 303). Martin, E. G. 1912 The measurement of induction shocks, p. 34. Martin, E. G., anp Lacey, W. H. 1914 Amer. Journ. Physiol., vol. 33, p. 212. Martin, E. G., anp MenpenHALL, W. L. 1915 Amer. Journ. Physiol., vol. 38, p. 98. Martin, E. G., anp Stizes, P. G. 1914 Amer. Journ. Physiol., vol. 34. 1916 Amer. Journ. Physiol., vol. 40, p. 194. OswaLp, A. 1916 Arch. f. d. gesammte Physiol., Bd. 164, S. 506-582. (Physiol. Abstracts). 1916 Arch. f. d. gesammte Physiol., Bd. 166, S. 169-200 (Physiol. Abstracts). Porter, W. T. 1908 Boston Med. and Surg. Journ., vol. 158. 1910 Amer. Journ. Physiol., vol. 27, p. 276. 1915 Journ. Physiol., vol. 36, p. 418. 380 42 43 D. OGATA AND SWALE VINCENT Porter, W. T., AND Storey, T. A., 1907 Amer. Journ. Physiol., vol. 18, Peale ee AND Turner, A. H. 1916 Amer. Journ. Physiol., vol. 39, idea en AND QuinBy, W. C. 1908 Amer. Journ. Physiol., vol. 20, ep AND BILLINGSLEY, P. R. 1916 Amer. Journ. Physiol., vol. 42, p. 16. 1917 Amer. Journ. Physiol., vol. 42, p. 16. Sotmann, T., ano Piucuer, J. D. 1910 Amer. Journ. Physiol. xxvi, p. 233. Sraruine, E. H. 1915 Principles of human physiol., p. 1006. Stewart, G. N., anp Larrer, W. B. 1913 Arch. Int. Med., vol. 11, p. 365. Stites, P. G., anp Martin, E. G. 1915 Amer. Journ. Physiol., vol. 37, p. 102. TicmrsTeDT, R. 18 3 Lehrbuch der Physiologie des Kreislaufs (quoted from Asher,? p. 349). Tur 1898 Hermann’s Jahresb., S. 61 (quoted from Nagel’s Handbuch). VincEnT, S., AND Cameron, A. T. 1915 Quart. Journ. Exper. Physiol., vol. 9, p. 48. beget Sake he a ~ aie, os ‘ PP Pe ee a vs Re Clk Sk (. ws et ee FUL . Neat wa ea tat. 1. “BY apes bs | shy case hs : Te. tie de Vy. ‘ Deiat | ee " wo gfe Sh aeeley TE “d BOP ae Pa tagae ' 7 = > i basthovarty we sue wavny ip { ‘ pate PLATE 1 EXPLANATION OF FIGURES Fig. 1 The effect of increased respiratory movements upon vasomotor re- flexes. Bitch. 9 kilos. 10/5/1918. Ether. The left sciatic nerve was stimu- lated at intervals, the stimulus increasing from left to right. Upper curve, respiratory movements. Lower curve, blood-pressure. Base line is that of zero pressure, with periods of stimulation. The height of the blood-pressure in mm. Hg. is indicated by the em. measured out and numbered. Time in sec- onds. For further explanation see text. Fig. 2 Increased respiratory movements prevented by very deep narcosis. Same bitch as in figure 1. For explanation see text. Fig. 3 Effect of increased respiratory movements upon blood-pressure. Bitch. 14 kilos. 30/5/1918. Ether. Thorax intact. The left sciatic nerve was stimulated. The result is a marked fall of blood-pressure. Fig. 4 Effect of increased respiratory movements upon blood-pressure pre- vented by opening the thorax. Same bitch as in figure 83. Thorax wide open in the middle line. The same nerve was stimulated with the same strength of stimulus as in figure 3. The result is a marked rise of blood-pressure. iS) ~ bo VASOMOTOR REFLEXES D. OGATA AND SWALE VINCENT mt Hi mm ' wits mi i m fi Ne f “enh matte nt ly , a A Hy a ue m | HA rN il it 80 ee bt, Fesnia ' aint sale ahead Ly | i ie yet Hemi dei Ea Sala Naty at a NT a A a UU dt Dd i \ va lal g) NNT iii INCOME TCTNNT ONO TS 11) 9 FATPIFIANITEN TORT AROMAS DT Da — _@ L Tint | bs obisigiclnalasinstlelutiaau cua diw aianiann, SSOSTUETTTAITTGSTUCUOQONUOUUOTECUUOTONUOUUICNTUOUONOTONOUOTOTOCIUTUONUCOTOOTUNNTON OMT TCU MTTT TIM TT ——— a=. id PLATE 1 LUELINUATOUFODURONOUUESTIUCUIVIUITON ENON IPTUUTITIUUTTIUTVOUUTUNNUWUDIVUTUCUUOIISUNOGUEERE PLATE 2 EXPLANATION OF FIGURES Fig. 5 A marked fall of blood-pressure which is apparently not due to in- creased respiratory movements. Dog. 12 kilos. 9/5/1918. Ether. The left sciatic nerve was stimulated. Fig. 6 An example of vasomotor reflexes upon weak (left) and strong (right) stimulations. Same bitch as in figure 3. Thorax was very wide open in the middle line to eliminate the influence from increased respiratory movements. A weak stimulation caused a fall and a strong stimulation caused a rise of blood- pressure. Fig. 7 Vasomotor reflexes under chloroform. Bitch. 7 kilos. 28/5/1918. Thorax wide open. A weak stimulation produced a fall (left) and a strong stimulation produced a rise (right) of blood-pressure. Fig. 8 Effects of weak and strong stimuli, respectively, under brain compres- sion. Dog. 18 kilos. 15/8/1918. Brain compression and artificial respira- tion. Right ulnar nerve stimulated. The fall of blood-pressure increased at first with the development of the strength of stimulus and then passed over to a rise crossing a neutral point. , Fig. 9 Effect of frequency of stimulation upon vasomotor reflexes. Bitch. 15 kilos. 18/7/1918. Chloroform and curare. Right saphenous nerve was stimulated. The frequency employed 1, 2, 5, 10, 20, 40, and 80 per second, re- spectively, from left to right. The one per second stimulation caused a fall, the two per second stimulation showed practically no effect, and the other stimu- lations produced a rise. The maximum pressor response was reached at forty per second stimulation in this case. Fig. 10 Stimulation of a sensory (saphenous) and a motor (a branch of the femoral) nerve. Dog. 9 kilos. 10/10/1918. Brain compression and artificial respiration. Stimulation of a sensory nerve gave a more pronounced fall than that of a motor nerve. Fig. 11 Stimulation of nerves of the same category but of different sizes. Same dog as in figure 10. The stimulation of a larger nerve (sciatic) produced a more marked response than that of a smaller nerve (peronea!). 386 VASOMOTOR REFLEXES D. OGATA AND SWALE VINCENT HUUESUUUUUUUNCUUEUOVODUCOVEUSUNUUCETUUNTIUICUUUS SUSUNCTENUCTTSUTIVUN TT RUST TONTTTTN|] UU UUUUUTUUTUNvEUUOWOSHOUUCLITTNCUTCD [NTT TIE AAA AAA AV AAA AAA VY VV VAAL AOTARATANTIVV HURT TR RAD LALLA cement | “Hiya CUTNLAVUVEOUUUUOUUBUDUNDOUUCUDAUOTUSUHUECUOUECUUUVONICUUNNUUOUNNTNUIUSVETUUSTUCUUNUNTITNOUITISHNUUEUEUUTSNTICITY I TUCE EN : » iil wl iti ill 5 IUVOVI OUUUUCUUNUUTU VE CVUVAUUUUUCUNDUDTUTEITIUUUUCUECUUUUCQUUUCNIIUENUUEUOUCNUCOUNUNUNUUSUUURIUCNIIUIUUCUEUUNUNEULONIL, iN , AN) ui | ne a aaa : mM \ 4 m ee cance es cn ll ‘inmgnasl hy iin uty Tl HN Yn eT —" ine io a — 4 ———— e. ga ———— 1 — — —o L = i ‘ ! n i 1 Lili OOTPTETETTR TTT ERTOT PUTT OTT ITTOTOOTOTENTTTTTT TETTCTT TN NURUTTTRTTOTTITITUPOOTUNTTTTTTTITTTCCOTTN OT OTTTTTTTTTTTOTTTOTOTTTTTTTTTTTRTTeTTaTaTOTTaTTTTTCCTTiTTTTTOTONTTNUCTCTTTTTTTTOT ONT OTTTTTTTTINVTD OTONNTTOITTTICOTTLTOCNCKUTUTTOTITUTTIAKRUN UNUCACOTICUTULNTUDOCOVOURULCLUUORUOUCMCONCOOUITUUUUTUON USE OUNUITTTITOO TON EMnnEN COOITOTITinee® CVUNUORTUIT Tran COUHNCURNTTULULDIUCOUCECDSERONCOOTeLIIULCrOanoR FUOUUTUOVUTTOTONVCUOUNMUNENUTNEDONUUGy PLATE 3 EXPLANATION OF FIGURES Fig. 12 Seratching of the skin. Dog. 12kilos. 9/5/1918.' Ether. A marked fall of blood-pressure. Respiratory movements practically unaffected. Fig. 13 Application of heat (boiling water) on the skin. Dog. 10 kilos. 25/10/1918. Brain compression and artificial respiration. A fairly marked fall of blood-pressure. Fig. 14 Electrical (strong) stimulation of the skin. Dog. 11 kilos. 14/5/1918. Ether. A very marked fall of blood-pressure. Respiratory move- ments affected very slightly. . Fig. 15 Scratching of muscle (right sartorius). Bitch. 10 kilos. 22/10/1918. Ether. a 4) nite 2 abs ,) aes ’ ’ evs Say Be i ' Wire mda sy +) ‘ota aT etsy fu : ay a TT ak yah v4 ora® wi Ait Sika i NAL OF COMP: esuenox0en vou. 30, No. 5 Resumen por el autor, James Stuart Plant. Instituto Wistar de Anatomia y Biologia. Factores que influyen en el comportamiento del cerebro de la rata albina en el liquido de Miiller. El cerebro de la rata albina sufre un cambio tipico cuando se ‘‘fija”’ en el liquido de Miiller, compuesto de 2 por ciento de bi- cromato potdsico y 10 por ciento de sulfato sédico. El cerebro aumenta rapidamente de peso, al que sigue una pérdida de peso lenta y continua, hasta que al cabo de setenta y cinco dias, al pesar el cerebro se comprueba que pesa de 20 a 30 por ciento mas que el cerebro fresco. 1. La edad es la principal condicién que influye en esta reaccién del liquido de Miller. Los cerebros de ratas viejas aumentan mds en peso y retienen este aumento durante los setenta y cinco dias. 2. El peso inicial del cerebro, 0 sea su tamafio, es la condicién de mas importancia después de la apuntada mas arriba. Del mismo modo que con los cerebros de la misma edad, los cerebros mds ligeros aumentan mds de peso durante la primera parte de su permanencia en el liquido de Miller. Esta diferencia disminuye gradualmente y desa- parece al cabo de los setenta y cinco dias. 3. La edad pr6xi- mamente semejante (entre ciertos limites) tiene casi el mismo valor determinante que la igualdad de edad. 4. El sexo es un factor sin importancia, del mismo modo que la composicién hereditaria (relacién dentro de un tronco determinado). Translation by José F. Nonidez Carnegie Institution of Washington AUTHOR'S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JUNE 30 FACTORS INFLUENCING THE BEHAVIOR OF THE BRAIN OF THE ALBINO RAT IN MULLER’S FLUID - JAMES STUART PLANT Neurological Laboratory of The Wistar Institute of Anatomy and Biology The brain of the albino rat, placed for a period in Miiller’s fluid, exhibits a typical change. In the course of time it not only hardens, but also markedly increases in weight. There is a rapid increase to a maximum in about one week’s time, after which there is a slow, steady loss until the seventy-five day weighing, at which time the brain weighs 20 to 30 per cent more than when fresh. It was thought that changes in this typical curve might be induced in the brains of rats previously anes- thetized for prolonged periods, and it was hoped that this cri- terion would be more delicate than the microscopic or analytic tests which had, so far, failed to demonstrate a change. The work was done at The Wistar Institute of Anatomy during the academic year 1913-1914. PLAN The main question of the effect of the anesthetic on the typ- ical curve remains unanswered. From the start, however, it was recognized that various factors influenced the reaction of ‘control’ brains to Miiller’s fluid—factors which are inherent in the material. It is these factors and their influence on the reac- tion with which the present paper deals. PROCEDURE The. brains studied belonged to ‘stock’ albino rats. The animals were killed with ether and the brains quickly removed with every care not to damage them. They were immediately weighed and then suspended in 50 ec. of Miiller’s fluid. The 411 412 JAMES STUART PLANT whole was kept in a black cardboard case in a dark closet. Subsequent weighings were as follows. The brain was removed from the solution and placed for about ten seconds on a dry piece of filter-paper. The string by which it had been suspended was during this time removed. The brain was then placed on a watch-glass and immediately weighed. It was returned to the Miiller’s fluid as quickly as possible. The watch-glass was then weighed. Reweighings were carried out at 24 hours, and at 7, 14, 30, and 75 days after killing the rat. On completion of the weighings the percentage of water in the brain at the final weigh- mg was determined. This procedure involved placing the brain, immediately after its last weighing, in a small glass vial of known weight. This was kept in a drying oven (temperature, 97°) for one week. On removal, the vial was cooled in a desic- cator at room temperature and weighed. The Miiller’s fluid used was made up in 1000 ce. lots. To 25 grams of potassium bichromate c. p. and 10 grams of sodium sulphate c. p. was added 1000 cc. of distilled water. Time was given for dissolving the salts and, after thorough agitation, the solution was divided into two-500 ec. lots and kept for one month before being used. In every instance ‘pairs’ of brains were fixed in fluid from the same bottle. The Miiller’s fluid was always kept in a dark closet. No attempt was made to control the temperature during the reaction of fixation other than that all specimens were kept in the same dark closet at room temperature. Thus the results may be considered as comparable. Necessarily our original results are in fechas of absolute weights and absolute gains. In the presence of so diverse initial weights it seemed, however, best to state all gains in weight as percent- ages of the original weight. This makes the data comparable. Corresponding to this, all statements of the relation of one brain’s gain to that of another are in terms of a percentage of the percentages of gain of the heavier brain. This leads to higher figures, in the relations of the gains, than would be the case were the actual differences between the absolute gains stated. ALBINO RAT BRAIN IN MULLER’S FLUID ALS OBSERVATIONS If we consider the brains of pairs (rats of the same age, sex, and litter, i.e., as similar as possible), there appears a very dis- tinct tendency for the brains of older rats to gain more in Miiller’s fluid than do those of the younger rats. Table 1 presents fifteen pairs arranged on the basis of their age. In ten of the fourteen possible comparisons the brains of older rats gain more in twenty- four hours. Also in ten of the comparisons the older brains TABLE 1 Percentages of gain of pairs of albino rat brains arranged according to age and weighed at intervals from twenty-four hours to seventy-five days AVERAGE PERCENTAGES OF GAIN OF BOTH BRAINS ese eh IN MULLERS FLUID SEX AGE 1 2 24 hours 7 days 14 days 30 days 75 days days grams grams °) 52 1.465 1.368 19.0 28.2 26.2 23.5 22.9 °) 55 1.615 1.578 20.8 o2.1 28.1 26.8 25.4 of 57 1757 1.671 20.6 31.2 26.0 25.9 25.7 of 59 1.635 1.519 21.2 Bout 29.9 27.0 26.7 eh 61 1.834 1D 18.0 32.2 28.0 25.9 25.0 g 61 1.699 1.581 20.0 29.5 21.0 24.1 23.0 fof 62 1.490 1.477 18.6 31.8 28.3 25.3 24.4 fof 62 1.662 1.656 19.3 28.2 24.6 23.1 22.4 g 62 1.497 1.496 20.5 31.7 29.1 25.7 25.3 of 62 1.831 1.699 20.8 32.8 30.8 28.0 7.8 °) 64 1.677 1.606 20.9 32.0 28.6 26.2 250 fof 67 1.651 1.587 21.7 32.6 29.7 26.6 26.4 °) 72 ‘1.610 1.492 27.6 33.6 30.7 27.9 26.7 of 160 1.791 1.752 25.1 37.2 35.8 32.8 32.2 ey 218 2.008 1.824 28.3 40.0 38 .6 315) 47/ 34.6 gain more in seventy-five days, though this does not in every case involve the same comparisons as were favorable to the older rats at the twenty-four hour weighing. Table 2 presents a summary of the data of table 1. The averaged figures for the youngest three pairs and for the oldest three pairs are given. The data show that age is a very important factor in the reac- tion of the brain of the albino rat to Miiller’s fluid. Brain weight increases as a function of age, and there exists between these two characters a very high coefficient of correla- 414 JAMES STUART PLANT TABLE 2 Percentages of gain of averaged entries of table 1 AVERAGE PERCENTAGES OF GAIN IN INITIAL WEIGHT M@LLERS FLUID PAIRS AVERAGED AGE 24 7 14 30 75 5 1 < hours | days days days days days .| grams | grams First three entries......... 55 | 1.612) 1.539) 20.1 |) 30.5" | 2628) | 25 4a Last three entries........- 150" |* £803) 17689) 27-0") 3720 35.07) Sane tion. If we make a comparison of the individual brains of the respective pairs involved in table 1, however, we may study the effect of initial brain weight in animals of like age. The data are given in table 3. In place of the percentage of gain of the lighter brain there is entered at the several columns marked ‘Per cent deviation of 2’—under ‘Time in Miller’s fluid’—only the relation of that percentage to the percentage of gain of the heavier brain. Of the fifteen pairs involved, it will be noted that in twelve the lighter brain gains more in the first twenty- four hours (represented by a + in the second column). Also in twelve of the fifteen pairs, thé heavier brain later ‘catches up’ —that is, the relative gain of the heavier brain is greater at seventy-five days than at one day. ‘This phenomenon is clearly demonstrated in the ‘averages’ at the bottom of the table. The results may be summarized as follows: 1 DAY 7 pars | 14 pars | 30 pays | 75 pays Average difference.................. +3.9} 41.5] +1.3) 41.2} +0.2 Standard) deviation.....-...-2 one +6.1 =—-3-9)| = 50),| 522) Soe Ul Thus it appears that the lighter brain gains more in the early part of the stay in Miiller’s fluid, but that this difference prac- tically disappears at the seventy-five-day weighing. It is to be noted that the standard deviations are lowest at the seven-day weighing. This seems to represent the period of least indi- vidual variation. As this represents the time of maximum increase, we may consider that as the more stable period in the a ALBINO RAT BRAIN IN MULLER’S FLUID 415 curves and think of the rise and fall in percentage of increase as periods more subject to individual variation. While increasing age presupposes increase in brain weight, it is apparent that these two age and brain weight—act as oppos- ing factors in the determination of the reaction of the brain to Miiller’s fluid. That is, the older brains (these are heavier) TABLE 3 The effect of initial brain weight—albino rat—on the percentage of gain of paired brains. INITIAL WEIGHT Pairs arranged according to age. Deviation of the lighter brain TIME IN MULLERS FLUID. PERCENTAGES OF GAIN eg Reem 24 hours 7 days 14 days 30 days 75 days 1 2 Per Per Per Per Per cent de- cent de- cent de- cent de- cent de- viation viation viation viation viation of 2 of 2 of 2 of 2 of 2 days | grams | grams Q 52 |1.465)1.368) 19.3)— 2.8/27.9] +2.1/26.8) —4.2/24.1!|— 4.9/23.5|— 5.7 Q 55 |1.615]/1.578} 20.4)+- 3.6)31.7| +2.5/27.7| +2.6125.8)/4 8.2/25.1/4+ 2.1 rot 57 |1.757/1.671) 19.5)4+11.1/80.5) +4.1/25.5} +4.1/25.2/+ 5.1/24.8/4+ 6.8 ot 59 |1.635)1.519) 20.5)+ 6.5/33.0) +0.2/30.3) —2.2/27.4;— 2.4/27.2|— 4.0 Q 61 |1.699/1.581) 18.8)4+13.3)/28.7| +6.0/26.2) +8.8)23.3}/+ 7.0121.6/+12.9 fof 61 |1.834!1.715) 17.9|+ 1.3/31.5| +4.2/27.5| +4.1125.4/+ 3.8/24.7/4+ 1.8 of 62 |1.490/1.477| 18.8)}— 1.8)382.4| —3.4/29.3) —7.2/26.3/— 7.3/25.0/— 4.8 Q 62 |1.497|1.496) 20.5/+ 0.3/31.4) +1.7/28.7| +2.1/25.9)— 0.7/25.1/4+ 2.1 of 62 |1.662/1.656} 19.2)+ 0.9/27.4) +6.0/23.7| +7.2/22.6)/+ 4.4/22.2)4+ 4.5 of 62 {1.831}1.699) 18.9/+19.6)31.5) +8.4/29.4) +9.8)26.7/+10.3)26.3/+11.4 2 64 |1.677/1.606) 21.1)— 1.2/32.6| —3.5/29.4| —4.9/26.8|/— 4.3]26.2|/— 4.3 rot 67 |1.651/1.587| 21.6/+ 0.8/33.3) —4.6)80.5) —5.1/27.3}— 4.7/28.0)/—11.2 2 72 |1.610/1.492} 26.9)/+ 5.2/33.3} +1.3/80.4| +1.7/28.3]— 2.3/27.4/— 5.2 ot 160 |1.791/1.752| 25.0/+ 0.3/38.0| —4.0|/85.3| +2.4/32.6/+ 1.1/32.5|/— 2.2. rot 218 |2.008]1.824) 28.0/+ 2.0/40.1) +1.2/38.5| +0.9/34.8/4+ 4.9/84.7)/— 0.8 VETAD Cote sod Ws + 3.9 +1.5 +1.3 + 1.2 + 0.2 Standard deviation....... + 6.1 +3.9 +5.0 + 5.2 = 6.5 gain more than do the younger ones; yet the lighter brains gain more than do the heavier ones if we can eliminate age as a fac- tor, as was done in table 3. The curve of increase in weight may be considered as capable of solution into at least two curves —expressing these two factors. Age appears to be by far the more potent factor. 416 JAMES STUART PLANT A further study was made of fifty-nine brains arranged accord- ing to increasing brain weight but without regard to sex or litter. In table 4 these are arranged according to initial brain weight in three age groups. Within these grouns two phenomena are apparent (shown in the averages under the vertical column, ‘Percentage difference from the following value’). These are the early greater gains for the lighter brains (this does not hold clearly for the ten brains of the youngest group where there is practically no difference); and the fact that at the seventy-five- day weighing the lighter brains show relatively a less percentage of increase than they do at the twenty-four-hour weighing. Since these facts are just those which determine the curve when brains of rats of the same age, sex, and litter are compared, we may conclude that in the reaction of the brain to Miiller’s fluid: 1. Sex is negligible. 2. Inherited composition is negligible. 3. Approximate similarity of ages (the range being limited) may be considered as having the same effect as though the ages were identical. The data on the percentage of water—in the last column of table 4—will be discussed later. A group of four brains—all belonging to young rats—was sub- jected to an additional procedure. The brains, immediately upon removal, were separated into cerebrum, cerebellum, stem, and olfactory bulbs. Each part was then treated as were the ‘whole brains of the other series. The data are given in table 5 in this way, that that percentage of the whole brain weight represented by the weight of each part at each weighing is re- corded. The figures for the four brains show but slight varia- tion, and table 5 therefore presents only the averages of the four. The relative weights of the various parts undergo considerable change in Miiller’s fluid, but this change is mainly consummated in the first twenty-four hours. Thus we may assume from this study that, while the relations of the various parts are altered in the fixing solution, the length of time, after the first twenty-four hours, during which the parts are subjected to this treatment, is | a matter of minor import. ALBINO RAT BRAIN IN MULLER’S FLUID 417 TABLE 4 The effect of initial brain weight—albino rat—on the percentages of gain of brains arranged in age groups, regardless of sex or litter . 38 AVERAGE PERCENTAGES OF GAIN ak — NuM- | AVER as Ses 3 INITIAL WEIGHT a “| 24 Qo & = mn (crams) — /PCxses | ace | FOURS ace Pal || 30 75 [8 oe 38 ges days | days | days | days Bae z Ee tale BSS$8| 53 4 4 ow 50 to 60 days 1.35-1.40 1 52 | 18.7 |— 2.8] 28.5 | 25.7 | 22.9 | 22.2 |— 5.6] 79.3 1.40-1.50 1 52 | 19.3 |—11.1] 27.9 | 26.8 | 24.1 | 23.5 |—11.2] 79.3 1.50-1.60 3 53 | 21.7 |+ 5.8} 33.2 | 29.2 | 27.5 | 26.5 |+ 1.2] 80.1 1-.60—1.65 2 o7 | 20.5 |— 2.2) 32.3 | 29.0 | 26.6 | 26.2 |— 3.0} 79.9 1.65-1.70 2 58 | 20.9 |+ 3.2) 32.8 | 28.8 | 27.7 | 27.0 |+ 8.7] 80.3 1.70-1.80 il 57 | 19.5 SOR) | 2020) | 2oeen eas 80.3 Average.1.58 10 55 | 20.5 |— 0.9 31.7 | 28.1 | 26.3 | 25.7 |— 2.1) 80.0 60 to 70 days 1.40-1.50 7 64 | 21.7 |+ 1.9) 32.9 | 29.8 | 26.6 | 25.7 |+ 0.5] 79.7 1.50-1 .60 5 65 | 21.3 |— 8.6) 32.1 | 29.1 | 26.2 | 25.6 |— 9.3] 79.3 1.60-1.65 6 69 | 23.3 |+12.3) 34.1 | 31.5 | 28.8 | 28.2 |+10.6} 80.2 1.65-1.70 |} 11 68 | 20.7 |+ 2.9) 31.5 | 28:5 | 26.0 | 25.5 |4+ 3.1] 79.7 1.70-1.80 5 65 | 20.1 |+ 8.2) 31.6 | 28.2 | 26.4 | 24.7 |— 1.8} 80.1 1.80-1.90 3 63 | 18.6 Se e209: leave tala 2 80.7 Average.1.65 | 37 66 | 21.1 |+ 3.9} 32.3 | 29.2 | 26.6 | 25.9 |4+ 1.8] 79.8 180 to 240 days 1.65-1.70 1 189 | 28.7 |+ 8.5} 42.4 | 39.2 | 36.1 | 35.4 |+ 9.3) 81.6 1.70—1.80 4 191 | 26.5 |+ 2.8) 38.2 | 36.2 | 33.2 | 32.4 |+ 3.2) 80.5 1S 5 36.4 | 32.4 | 31.5 |+ 6.1} 80.5 1.9 2 33.0 | 30.4 | 29.7 78.4 Average.1.83 | 12 | 213 | 25.9 |+ 6.1) 37.9 | 36.0 | 32.7 | 31.8 |+ 5.6} 80.2 *The percentages were obtained originally by the use of values carried to three places. These have now been reduced to one-place numbers and there are therefore some apparent discrepancies in the percentage-difference columns. These differences, however, are not significant. 418 JAMES STUART PLANT TABLE 5 Averaged percentage weight relations of the parts of four albino rat brains during the course of their reaction to Miiller’s fluid WEIGHT OF PERCENTAGE| PERCENTAGE! PERCENTAGE SL SEE NEE REPRE- TIME IN AGE emer a pet ace x een REPRE | sENTED BY | MULLER’S SrsAGe OF (| gmepnn a | smn) E097 | pues ee a2, 1.610 Ga. 17 17.40 13.80 Sialos" Initial QV27 63.10 17.68 15.08 4.14 24 hours PAU 64.82 IZ PALL 14.33 3.63 7 days 2.094 63.77 ieaiG 14.72 39.5 30 days Difference between per- centages at initial and 24-hour weighing.......| —2.07 +0.28 +1.28 +0.51 Difference between per- centages at 24-hour and 30-day weighing....| +0.67 +0.08 —0.36 —0.39 WATER RELATIONS We have studied the water relations after seventy-five days in fifty-nine whole brains and after thirty days in the parts of three brains. There is evidently, in’the reaction of the brain to the Miiller’s fluid, a deposition of salts in the brain tissue. This is shown in table 6. Part A deals with the fifty-nine whole brains; Part B with the parts of three brains (all belonging to young females). The deposition of salts at the end of seventy-five days in the whole brain is from 3.9 per cent to 4.3 per cent of the total water of the brain despite the fact that the salts in Miiller’s fluid are present in a concentration of but 3.5 per cent. This shows a deposition of salts in the tissues. If, in addition, there is some diffusion of solids from the brain to Miiller’s fluid— and, from inspection, this appears to be the case—the percent- age of salts deposited must be even higher than that indicated by the figures given. The final percentage of water-in a given brain at the seventy- five-day weighing is only slightly greater than that of a fresh brain belonging to a rat of the same age, sex, and litter. In view of the 20 to 30 per cent net increase in weight in Miller’s eS ALBINO RAT BRAIN IN MULLER’S FLUID 419 TABLE 6 Part-A. Water relations at the seventy-five day weighing of fifty-nine whole brains (see table 4) UNDER 60 | 6070 120 | over 120 DAYS DAYS DAYS Average fresh brain weight.......00. 0.0. uss. see ek 1.582 | 1.651 | 1.826 Percentage of water (from Donaldson, ’16)..... etacs. 79.1% | 78.9% | 78.1% Calculated amount of water represented in fresh LOLS, ae pce ale le ARS Aaa aia th ls Ph, 9 atta 2 ad 0 ge 1.252'| 1.303 | 1.426 Hime bean meigha? /). LILES A Oe A 1.988 | 2.078 | 2.407 Buel amount, of water.<..°.:lc.se.hand.d poueseer 1.590 | 1.659 | 1.932 Percentage of water—observed....................... 80.0% | 79.8% | 80.2% Increase in weight due to water, gms................. 0.338 | 0.356 | 0.506 Increase in weight due to salts, gms.................. 0.068 | 0.071 | 0.075 Percentage of salts in total increase in weight.........| 16.6% | 16.7% | 13.0% Percentage of salts in the total water in brain........ 4.29, \ 4.3% | 3.9% Part B. Water relations at the thirty-day weighing of parts of three brains : OLFAC- ‘ CEREBRUM STEM dene ria Average fresh brain weight................. 1.097 | 0.292 | 0.235 | 0.058 Percentage of water (from Donaldson, ’16)...| 80.0% | 76.1% | 79.7% | 82.3% Calculated amount of water represented in UTE SUnS Ter PS 6) RS ae ee ne 2 ea 0.878 | 0.222] 0.188} 0.048 RRC Ges. ote hae he eet Atos oo ak 1.403 | 0.386 | 0.329] 0.078 Final amount of water....................... 1.139 | 0.301 | 0.266 | 0.067 Percentage of water—observed...............| 81.2% | 77.9% | 80.8% | 85.6% Increase in weight due to water, gms........ 0.261 | 0.079 | 0.078 | 0.019 Increase in weight due to salts, gms.......... 0.044 | 0.016 | 0.015] 0.001 Percentage of salts in total increase in weight (for totaljbrain:14)8%)). si... edn. ad. 2 14.5% | 16.4% | 16.4% | 4.9% Percentage of salts in total water in the part RRR VGA ORDA 5 V5 aan. o>, paminne Be ele nd «oes 3.9% | 5.0% |. 5.8%. | 125% fluid, this seems a striking fact, though where the initial per- centage of water is so high, it is evident that it takes a relatively large difference in the absolute water content to markedly affect the percentage value. In the three brains divided into their parts the salts are de- posited in the following percentages after thirty days: 420 JAMES STUART PLANT Cerebelltm ss (275-24 a tae ie ee ee ee te is os nee ee 5.8 per cent Stemmid aac. Ai LB ER oF ks 5.0 per cent Ceréebruim::coic os, Lean et et AR ee Be Ia ee 3.9 per cent OLRCHOTY DULDS. 2 aide Ie Sth se sian aie eeie ols Maen peee 1.5 per cent Average cor whole Braiienrts seks: sacs: elt eee ++. 4.3 per cent With the exception of the cerebellum, the percentage of salts deposited is in direct relation with the proportion of myelin in the part involved. As none of the parts were washed, it may well be that the interstices of the cerebellum were the site of large deposits of salts, a physical factor which may account for this anomalous result. CONCLUSIONS General reaction of the brain to Miiller’s fluid, or type curves. The brain of the albino rat undergoes a typical change when ‘fixed’ in Miiller’s fluid (2.5 per cent potassium bichromate and 1 per cent sodium sulphate). There is a rapid increase in weight followed by a slow, steady loss until at the seventy-five- day weighing the brain weighs 20 to 30 per cent more than when fresh. , Factors affecting this reaction, or components of the type curve. 1. Age is the main condition controlling this reaction to Miiller’s fluid. The brains of older rats gain more and retain this higher relative gain throughout the seventy-five days. 2. Initial brain weight, or size, is the condition of next impor- tance. As between brains of like age the lighter brains gain more during the earlier part of their stay in Miiller’s fluid. This difference is gradually lessened, and it disappears at the seventy- five-day weighing. Thus, while age and initial brain weight are highly correlated, they constitute factors which, when taken alone, influence in opposite ways the reaction of the brain to Miiller’s fluid. The early greater increase of the smaller brain of two rats of the same age may be due to one or both of the following factors: a. If we consider the brain as a sphere and the fluids as pene- trating at a fixed rate, then in the smaller brain a slightly greater ALBINO RAT BRAIN IN MULLER’S FLUID 421 proportion of the brain will be penetrated—that is, swollen by the fixing fluid—at any given instant in the early part of the reaction. This would give a more rapid enlargement in the smaller brain. b. The smaller brain has a higher percentage of water which might make diffusion more rapid. If this is a controlling factor, the matter of a higher percentage of water must be of more immediate importance than is that of a lesser percentage of myelin. 3. Approximately similar age (the range being limited) has nearly the same determining value as equality of age. 4, Sex is a negligible factor, as is also inherited composition (relationship within a given strain). FINAL WATER RELATIONS The increase in weight is due mainly to the taking up of water, but the percentage of salts deposited in the fixed tissues is much greater than that in the fixing fluid. With the exception of the - cerebellum, the deposition of salts is proportional to the myelin present in the part of the brain. RELATION OF RESULTS The curve of reaction of ‘control’ brains to Miiller’s fluid is evidently of such constancy in character as to make it a satis- factory criterion for a judgment as to alterations produced in the brain by various experimental procedures. It seems that in problems involving changes not available to such microscopic or analytical tests as we have, this reaction might furnish a valuable means of study. It appears from the foregoing results that when tests are made for the experimental modification of the response of the brain to Miiller’s fluid, it is necessary to have the test and control brains of the same age and, in those cases in which the brains differ in weight, to allow sufficient time for the compensation of this difference. 422 JAMES STUART PLANT LITERATURE While various observations bearing upon this problem have been made, none have employed exactly our experimental pro- cedure. Various solutions of potassium bichromate have been used—Donaldson (94), Fish (’93), and King (’10)—but a study in the changes in the reaction to Miiller’s fluid has not been made. A similar study was carried out by Hrdlicka (’06) in which various formalin preparations were used as fixing reagents. The effect of age upon this reaction has been discussed in a gen- eral way—Donaldson (94) and King (’10)—but it seems that the extent of its control over the reaction has not been previously stated. BIBLIOGRAPHY Donatpson, H. H. 1894 Preliminary observations on some changes caused in the nervous tissues by reagents commonly employed to harden them. Jour, Morph., vol. 9. 1916 A revision of the percentage of water in the brain and in the spinal cord of the albino rat. Jour. Comp. Neur., vol. 27. Fisn, P. A. 1893 Brain preservation with a résumé of some old and new methods. Wilder Quarter-Century Book, Ithaca. Hropuickxa, A. 1906 Brains and brain preservatives. Proc. U. S. Nat. Mu- seum, vol. 30. Kine, H. D. 1910 The effects of various fixatives on the brain of the albino rat, with an account of a method of preparing this material for a study of the cells in the cortex. Anat. Rec., vol. 4. FS ei i 9s , | betapich Vitel ae . PO HI TT 44. Waray. ' ait ths vi rds fae v4 | a 1 . . ¥ » . aa TG) fa) hnareaiz ns meen te hee. eae : j ry L ‘ j ay. t Pi koe ites aii) ai “aN Stage MAE sbi alivats any ‘ip et Rosa as at Toh Birtley MA a Lier tho: SORE: Pray a ay gat. He : Me aby ios. SEK bof onl Niel ip) ete ee % oe anigodta ait) ie | (CP sea a DIYARG deb Aomericda 6 PE Re arsesley LNg bak hora toon toh aaheepigiaige eile, aah ep icaNE Shag | bi aiahirsbtion “f nay ba ie LEE Ss oie +.) ae dost ¥e oe a is i bet rt “ay vila gE ACs RG Gea te bis y plaka oe | ya i) so potas i oti Rin eae cha ry rp Aa arial aang NF se oe anti Lehn Hs iaygeyeal ep rd Halal i He on Satine dyif; xan fa ane dress). vel Yh at ahaa Hen Bee coal ipod: selsb-Aattinoserey MRA’ Faxkdeame. a ale nye eines ip FD MO ee oha't, iia fay Au teal wveaglbnt, EE ERE IE ant aati gihent AiGhoy age es al i, ad (is dah | LMA to Pinta fac) Vigne S Ky Resumen por el autor, Olof Larsell. Universidad de Wisconsin. Estudios sobre el nervio terminal: la tortuga. El autor describe las relaciones periféricas y la distribuci6én del nervio terminal en el embrién de la tortuga. Un plexo del nervio, situado en el tabique nasal, esta’ intimamente mezclado con un plexo de la rama oftdlmica del nervio trigémino. El autor compara las células ganglionares del nervio terminal, en lo referente al tamafio y caracteres morfoloégicos generales, con las células de los ganglios sensorio y del simpdtico de los mismos embriones. Las células del nervio terminal presentan una seme- janza sorprendente, tanto en,el tamafio como en la forma, con las células halladas en los ganglios del simpatico, especialmente las de los situados en la regién cefdlica. Las células de los racimos ganglionares del nervio terminal no pueden diferenciarse de células emigrantes que se presentan a lo largo de la rama oftalmica del trigémino. Todo esto indica que el nervio terminal esta relacionado con el sistema del simpatico. Translation by José F. Nonidez Carnegie Institution of Washington AUTHOR’S ABSTRACT OF THIS PAPER ISSUED BY THE BIBLIOGRAPHIC SERVICE, JUNE 30 STUDIES ON THE NERVUS TERMINALIS: TURTLE! O. LARSELL Department of Anatomy, University of Wisconsin SIXTEEN FIGURES The present contribution was begun as part of a comparative study of the nervus terminalis in several groups of vertebrates. The work had not progressed far before it was found advisable to confine attention to one group at a time, so that the greater portion of the present report embraces the results of observa- tions made since the studies on the nerve in mammals by the author (718) was published. The somewhat extensive literature of the nervus terminalis was reviewed in the previous article, and only two papers which have appeared on the subject during the past year will be men- tioned briefly. hese papers are by Van Wijhe (’18) and Ayers (719). Van Wijhe’s paper reviews much of the literature of the nervus terminalis in the various groups of vertebrates briefly and homol- ogizes the nerve with one he noted a number of years ago (’94) in Amphioxus, which he termed at that time the ‘ner'vus apicis.’ He states: “‘Before the homologue of the profound ophthalmicus there is in Amphioxus still another nerve which supplies the utmost point of the snout. On account of this and because it arises from the morphological fore-end of the cerebral ventricle I called it the nervus apicis.”’ Ayers (719), in continuing his studies of Cephalogenesis, begun long ago, has found the nervus terminalis (Van Wijhe’s ‘nervus apicis’) in Amphioxus, and calls attention to its large size as 1 Contribution from the Zoological Laboratory of Northwestern University, William A. Locy, Director, and from the Anatomical Laboratory, University of Wisconsin. 423 THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, No. 5 424 O. LARSELL compared with the olfactory nerve in that form. He finds it also in Cyclostomes and states that in Bdellostoma the nerve presents an intermediate stage, as respects its size and relations, between Amphioxus and the selachians. He calls attention to the distinction between the vomeronasal nerve, which he terms the ‘nervus septalis’ and the olfactory nerve, and suggests a new classification of the cranial nerves in which the nervus terminalis would be number J, and, with the ‘nervus septalis’, would be added to the list of twelve cranial nerves usually recognized. The nervus terminalis is considered a sensory nerve which has to do with a group of chemical sense organs, and is related physi- ologically to the vomeronasal (his septal) nerve. The conclusions are reached from a study of Amphioxus and cyclostomes. Doctor Ayers believes that in higher forms the nerve has undergone con- siderable modification due to changes in head structure. I am under great obligation to Doctor Ayers for opportunity to read his manuscript prior to publication and for permission to make use of his observations. He also afforded me opportunity to read Van Wijhe’s paper, which I had not previously seen. It is a pleasure to express here my sense of indebtedness to him. My acknowledgments are also due Prof. William A. Locy, of Northwestern University, under whose direction the general problem was originally begun and who has since continued his interest. MATERIAL AND METHODS Embryos of the painted turtle (Chrysemys marginata) were used. Most of these had been fixed in a formol-bichromate- acetic fluid, some in formalin of 10 per cent, others in Tellyes- niczky’s fluid, and a few living embryos were obtained and pre- pared by the Cajal and the Vom Rath methods. Stages beyond 10- to 1l-mm. carapace length had become chitinized to such an extent in the rostral region that intact serial sections could not be obtained. Chiefly for this reason, the present contribution is confined to a description of the nervus terminalis in embryos up to 11-mm. carapace length (about 17 mm. greatest length. at a NERVUS TERMINALIS: TURTLE 425 Numerous dissections of embryos and of newly hatched turtles were made with the aid of the binocular microscope. The head was split slightly to one side of the midsagittal plane, and the soft parts were then sufficiently removed to expose the nerve and its adjacent structures. The embryos sectioned were cut in the sagittal plane or trans- versely, and were stained by various methods. The most generally satisfactory stain for older stages was found to be iron-hematoxylin, but some of the most instructive series were ob- tained by overstaining with Delafield’s hematoxylin, followed by a counterstain of saturated aqueous orange G to which two drops of glacial acetic acid were added for each 50 cc. of stain. The serial sections studied were-as follows: 1 series 6-mm. embryo, sagittal, stained with iron-hematoxylin. 2 series 6.3-mm. embryo, transverse, stained with hematoxylin and Congo red. 3 series 7.5-mm. embryo, transverse, stained with hematoxylin and Congo red. 1 series 8-mm. embryo treated by the Cajal method, cut sagittally. 1 series 9-mm. embryo, sagittal, stained with hematoxylin and Congo red. 1 series 9-mm. embryo, sagittal, treated by Vom Rath method. 1 series 9.5-mm.-carapace-length embryo, stained with hematoxylin and Congo red, sagittal plane. 2 series 10-mm.-carapace-length embryos, sagittal, stained with iron-hema- toxylin. 1 series 10.5-mm.-carapace-length embryo, sagittal, stained with hematoxylin and erythrosin. _ 1 series 11-mm.-carapace-length embryo, sagittal, stained with hematoxylin and Van Gieson’s stain. 1 series 1l-mm.-carapace embryo, sagittal, stained with hematoxylin and orange G. ‘DESCRIPTIVE The nervus terminalis in the turtle has its origin by several small roots from the ventromesial surface of the forebrain, just caudad to the olfactory bulb. It can be demonstrated by dis- section in suitably prepared material. Figure 1 represents a dissection of an embryo of 11-mm. carapace length, showing the left nervus terminalis and its relation to neighboring structures. In the specimen figured the rootlets were not evident until some of the overlying brain tissue had been removed by brushing. By this process three roots were demonstrated. In other dissections but two roots were brought to light, usually after some brushing. O. LARSELL 426 NERVUS TERMINALIS: TURTLE 427 In sections of corresponding stages, cut in the sagittal plane, two or three roots were observed to enter the brain substance, but their fibers could be followed within the brain for only a short distance (fig. 2). On following these roots distally from the brain, they are seen to unite (figs. 1 and 2) and a ganglionic swelling was invariably found just beyond their point of union. In the dissection figured it will be noted that the two more dorsal roots unite to form a single short trunk before entering the ganglion, while the ventral root enters the ganglion directly. In sections a number of ganglion cells (figs. 2 and 3) may be observed in this mass, but in none of the embryos examined did this ganglion appear to have as many cells as others located more rostrally, especially the one marked gn. (fig. 3). Rostrally from the ganglionic mass which lies at the junction of the rootlets, the nerve continues as a compact trunk as far as the most anterior part of the olfactory bulb, midway between the Fig. 1 Dissection of head of turtle embryo of 11-mm. carapace length to show the nervus terminalis and its relation to neighboring structures. The left lateral half of the brain is viewed from the mesial aspect. XX ca. 15. Fig. 2 Central roots of the nervus terminalis at their point of junction with the brain. Turtle embryo of 1l-mm. carapace length. Hematoxylin and orange G stain. X 410. Fig. 3 Reconstruction of the right nervus terminalis of turtle embryo of 1l-mm. carapace length, stained with hematoxylin and orange G. The nervus terminalis is represented as lying on the medial surface of the vomeronasal nerve for all of that part of its course which is parallel to the latter. Part of its course, however, as shown in figure 1, is lateral to the vomeronasal nerve. X ca. 16. ABBREVIATIONS bl.v., blood-vessel m.ob.ven., ventral oblique muscle bu.olf., olfactory bulb m.r.ant., anterior rectus muscle cer.hem., cerebral hemisphere n.olf., olfactory nerve bundles cr.cav., cranial cavity n.oph., ophthalmic branch of V nerve gn., main ganglionic mass (ganglion 7.ter., nervus terminalis terminale) of nervus terminalis n.vom., nervus vomeronasalis gn.’, accessory ganglion terminale na.ant., anterior naris gn.c., ganglion cells ’ na.po., posterior naris gn.cl., ganglionic clusters olf.epith., area of olfactory epithelium mes., mesencephalon op.chi., optic chiasma m.ob.do., dorsal oblique muscle ret., retina 428 O. LARSELL dorsal (vomeronasal) bundle and the main bundle of olfactory fibers proper, as shown in figure 1. At this point it turns to follow these bundles, passing between them in such a manner’ that it could not be traced further by the method of dissection. Sections, however, reveal the further course of the main bundle of the nerve and indicate the presence of numerous ganglionic cells scattered along its trunk as it passes over the mesial surface of the olfactory bulb. These cells form clusters of various sizes. One of the largest of these clusters is no doubt indicated by the swelling (fig. 1, gn.’) shown in the dissection. Another and larger ganglionic mass is shown (fig. 3, gn.) at the point where the ter- minalis passes lateral to the vomeronasal bundle. This corre- sponds to the position usually occupied by the largest cluster of cells in the majority of the embryos which were sectioned. From the position of this ganglion distally the nerve could not be followed further by the method of dissection, because its strands became too intimately mingled with those of the olfactory and vomeronasal nerves. Fortunately, however, in some of the series of the older stages studied, a differential stain was obtained by the method previously described, so that the terminalis bundles could be distinguished from those of the other two nerves and, further rostrad, from the fibers of the ophthalmic branch of the V nerve. This differentiation was aided by the fact that the olfactory and vomeronasal nerves appear as compact bundles of wavy fibers with few nuclei scattered among them. The strands of the nervus terminalis are smaller and much less compact and present relatively numerous sheath nuclei as well as larger gan- glionic cells. Where the strands of the trigeminus were inter- mingled on the septum, they also had a characteristic appearance, ° apparently due to the process of myelination, as well as to dif- ferential staining. ‘These characteristics, however, could not be noted with any degree of certainty in the smaller tracts, so that the reconstruction represented in figure 3 indicates only the larger bundles and their grosser ramifications. In some of the preparations the nerve was seen to divide in- tracranially at the ganglion gn. into two strands, which, however, reunited to form a compact trunk before the nerve left the brain NERVUS TERMINALIS: TURTLE 429 cavity. This condition is illustrated in figure 7. There were some indications of much finer strands also in this region, but they were not sharply enough differentiated from the olfactory strands to justify inclusion in the figure as part of an intracranial plexus of the terminalis. After emerging from the cranial cavity, the nervus terminalis is composed of several well-marked strands which continue par- allel with the vomeronasal bundles, mesial and in part dorsal, to the latter. A short distance from the point of emergence of the nerve, its strands begin to form a plexus over the nasal septum. Lack of silver preparations of the older stages made it impossible to follow this plexus for any considerable distance, especially in its more rostral part, where it becomes more complex due to entrance of fibers from the trigeminus. It was very evident that both the nervus terminalis and rami from the ophthalmic branch of the V nerve take part in the formation of a plexus on the nasal septum. Clusters of ganglionic cells (fig. 3) were scattered throughout this plexus. Along the vomeronasal nerve there was a nearly continuous mass of cells from the point where the nerve turned ventrorostrally at the bulbus olfactorius, to the point where the more profuse spreading out of the septal plexus began. A some- what similar arrangement of cells along the vomeronasal nerve was found by Johnston (’13) in Emys, but apparently the cells were not so numerous as in Chrysemys. A fortunate Cajal preparation of an 8-mm.-total-length embryo gave a very clear demonstration that the trigeminus forms a more important portion of this septal plexus than might have been an- — ticipated. As shown in figure 4, which represents a reconstruc- tion from fourteen serial sections cut in the sagittal plane, the trigeminal portion of the plexus is formed by the ramifications of the ophthalmic nerve. The fibers were stained quite uniformly brown or black. They could be followed to the gasserian gan- glion, with the beautifully stained cells of which they united. The nervus terminalis in this preparation is represented by a few clusters of cells and some yellowish fibers. The cells could not with certainty be distinguished from the mesenchymal cells, 430 O. LARSELL NERVUS TERMINALIS: TURTLE 431 but sections of embryos of approximately the same stage which were stained by other methods, indicate that the clusters are composed of ganglionic cells, which from their position no doubt belong to the nervus terminalis. Some details of this plexus of the ophthalmic nerve are illus- trated in figures 5 and 6. As shown in the figures, relatively small strands of fibers meet at nodal points, from which the indi- vidual fibers are redistributed to bundles diverging at various angles. No ganglionic cells could be observed about these nodal points at this early stage. In older stages, however, cell clusters of various sizes are numerous at the points where the ophthalmic nerve ramifies on the septum, and are found as far forward (figs. 3 and 7) as the most rostral part of the septum. It seems likely that the more rostral of these cells correspond to clusters of sym- pathetic cells described and figured by Willard (’15) in Anolis. The observation of Rubaschin (’03) of a ‘ganglion olfactorii nervi trigemini’ on one of the branches of the ophthalmic nerve in the chick appears also to be related. At various points the branches of the trigeminus and of the terminalis become so intimately related that the two cannot be told apart, and the smaller strands of the plexus which continue from these points, appear to contain fibers from both nerves. As shown in figure 7, which represents a reconstruction from nine serial sections, from which the finer strands of the plexus are omitted, the nervus terminalis anastomoses with one of the larger branches ‘of the ophthalmic nerve just dorsal to the vome- ronasal nerve. At the point of anastomosis is a large ganglionic cluster. The two nerves had the characteristic different appear- ance, previously noted, proximal to their point of union, but their Fig. 4 Reconstruction of the septal plexus of the ophthalmic branch of the trigeminal nerve of a turtle embryo of 8 mm. greatest length, prepared by the Cajal method. The figure was reconstructed from sections 79 to 92 of the series. The numerals indicate the sections from which the adjacent structures were projected. For the sake of simplicity, section 79 is indicated by the numerals 1, section 92 by 14, and intervening sections accordingly. X 80. Fig. 5 Portion of the plexus from section 81, at point marked 3* in the pre- vious figure, to show details of structure. > 385. Fig. 6 Portion of the plexus from section 88 of the series, at point marked 5* in the reconstruction, to show detail. XX 385. : 0. LARSELL ~, olf epith: NERVUS TERMINALIS: TURTLE 433 strands could not be told apart distal to this point. The relation of the two nerves and something of their appearance are repre- sented more highly magnified in figure 8, but this does not show the slight although clearly apparent differentiation of staining which the preparations themselves reveal in the larger bundles. The trigeminal fibers appear coarser and more compactly collected into bundles than do those of the terminalis, but in the smaller strands these characteristics are not apparent, probably because of the small number of fibers composing them. The olfactory and vomeronasal bundles, which are also rep- resented in the figure, resemble each other in being composed of very delicate fibers with a characteristic wavy appearance which could not be well represented in the drawing. Huber and Guild (713) in the rabbit and Brookover (717) in the human found indications of such an anastomosis of terminalis and trigeminal strands on the nasal septum, and the pres- ent writer (’18) partially demonstrated it in the cat. It is rather striking that three nerves, the terminalis, the olfactory (Read, ’08), and the trigeminus, should each form a plexus on - the nasal septum. Two of these plexuses overlap to a marked degree. The vomeronasal nerve is not plexiform, except as the large bundles composing it branch and reunite to a slight extent in their course to the vomeronasal organ. In this connection a statement from the previously cited paper of Ayers is of interest. He states: ‘‘The nasal chamber in man therefore contains the surface distribution of these three (terminalis, septalis, olfac- torius) cranial nerves as well as the surface terminations of in- vading branches of a fourth and more recent cranial nerve, the trigeminus.”’ Fig. 7 Reconstruction from ten sections (175 to 185) of the series from which figure 3 was reconstructed, to show anastomosis between rami of the nervus terminalis and of the ophthalmic branch of the trigeminal nerve. 44. Fig. 8 Reconstruction from four of the sections (sections 182 to 185) in- cluded in figure 7, to show at higher magnification the anastomosis of one of the rami of the ophthalmic plexus with a bundle of the nervus terminalis. This figure also gives some idea of the appearance of the fiber bundles. X 180. 434 O. LARSELL GANGLION CELLS An effort to reach some conclusion as to the character of the ganglion cells of the terminalis in the turtle embryos was made by comparing them with cells of other ganglia, cranial, spinal and sympathetic. The gasserian, sphenopalatine, and ciliary ganglia were studied in the head region. The spinal ganglia, beginning with the first thoracic, and the sympathetic chain ganglia were studied in the body region. Measurements were made of the nuclei with the aid of an ocular micrometer, the results of which are indicated in table 1. The cells measured were taken at random. The only selection exercised was in measuring nuclei TABLE 1 AVER- NUM- | SIZE OF|SIZEOF| AGE u BER | LARG- | SMALL-| SIZE OF POSITION OF CELLS “anes (Gea Rea! SS REMARKS URED | CLEUS | CLEUS | NUM- BER Periphery of spinal | 48 | 12.3 | 8.2 | 9.8 | Many unipolar. gangha Central part of same | 52 8.2 | 5.3] 7.3 | Bipolar, approaching unipolar ganglia condition Above combined 1005) 223. 9|\ 523 Gasserian ganglion 52 | 11.4! 7.0] 9.0 | Two sizes present, but small cells not many Ciliary ganglion, 50 | 12.4 | 7.0|9.0 | Nuclei large in proportion to large cells entire cell Ciliary ganglion, 22 7.9| 5.3 | 6.4 small cells Above two combined | 72 | 12.4} 5.3 | 8.3 Sympathetic chain | 50 8.1 | 5.3 |6.2 | Five nuclei larger than 6.7 u ganglia Sphenopalatine 50 8.115.838 | 6.74 ganglion N. terminalis, periph- | 50 8.8 | 5.3 | 6.72 | Some of these cells probably eral clusters belonged to clusters related to the trigeminus N. terminalis, central | 30 8.4 | 5.3 |6.7 | Of the 80 nuclei measured in root clusters . both peripheral and cen- tral clusters three were larger than 8.1 » and three were smaller than 5.5 xz. NERVUS TERMINALIS: TURTLE 435 which were spherical or nearly so. In the case of some cells, especially many in the terminalis ganglia, the nuclei were so elongated that it was necessary to measure the greater and the lesser diameters and take the mean of the two. All of the measurements and the drawings of the ganglion cells represented were made from a single embryo of 10-mm. carapace length, stained with iron-hematoxylin. This was done for the sake of uniformity, although the statements hold in general for all the embryos examined which were sufficiently advanced to show any pronounced differentiation of the various types of nerve cells. In drawing the figures the outlines of the cells and nuclei were traced with the aid of the camera lucida, and the same com- bination of lenses was employed in each case, so that the figures represent directly the variations in form and size of the various types. Comparison of embryos at different stages of development in- dicated that in embryos of 10-mm. total length, the spinal gan- glion cells were on the average somewhat larger than those of the sympathetic chain ganglia. There was, however, but slight difference in the size of the individual cells within the spinal ganglia. The sympathetic chain ganglion cells had still much the appearance of indifferent cells. In the gasserian ganglion of embryos at this stage of development, many of the cells were larger than the spinal ganglion cells of the same embryo. In embryos of 9.5 to 11l-mm. carapace (15.5 to 17 mm. greatest length) there is a marked difference between the size of the largest sensory ganglion cells and those of the sympathetic chain ganglia. The latter (fig. 12) are of pretty uniform size, but the spinal ganglia and, in less marked degree, the cranial ganglia, showed two fairly distinct sizes of cells. The larger type in the spinal ganglia (fig. 9) was found near the periphery of the ganglion. Many had already reached the unipolar con- dition, but others showed various transitional stages from the primitive bipolar cells. Of one hundred cells measured from three different ganglia, one in the thoracic region, one in the lumbar, and one in the sacral, the largest nucleus had a diameter of 12.3 u and the O. LARSELL 436 NERVUS TERMINALIS: TURTLE 437 smallest was 5.3 » in diameter. The average size of the entire one hundred nuclei was 8.5 uw. These cells were divided into two groups, those with a nuclear diameter greater than 8.2 u and those whose nuclear diameter was less than this, down to 5.3 » which was the smallest found. While this division was somewhat arbitrary, there was sufficient ground for it in the character of the cells, aside from their size, to make it appear justifiable. The smallest cells (fig. 10) had considerably less cytoplasm surrounding the nucleus, both relatively and ac- tually. Only various stages of the bipolar condition were ob- served in these smaller cells. They were found closely packed together in the central portion of the ganglion, while the larger cells were nearer the periphery. As indicated in the table, the fifty-two cells which showed a nuclear diameter less than 8.2 u had an average diameter of 7.3 uw. The forty-eight cells whose nuclear diameter was greater than 8.2 « were found to have an average nuclear diameter of 9.8 yu. The smaller cells (fig. 10) may correspond to the small ganglion cells found in the spinal ganglia of mammals by Dogiel (’08), Ranson (712), and other workers, but it seems likely that some of them at least represent cells of the larger type which have not Fig. 9 Cells from the peripheral portion of the first thoracic spinal ganglion of a turtle embryo of 10-mm. carapace length. Iron-hematoxylin stain. > 500. Fig. 10 Cells from the central portion of the same ganglion from which figure 9 was drawn. Turtle embryo 10-mm. carapace length, iron-hematoxylin stain. x 500. Fig. 11 Cells from the gasserian ganglion of turtle embryo of 10-mm. cara- pace length. Iron-hematoxylin. X 500. Fig. 12 Cells from the third thoracic sympathetic chain ganglion of turtle embryo of 10-mm. carapace length. Iron-hematoxylin. X 500. Fig. 13 Cells from the ciliary ganglion of turtle embryo of 10-mm. carapace length. Cells of smaller type indicated at A, and the larger type at B. Iron- hematoxylin. X 500. Fig. 14 Cells from the sphenopalatine ganglion of turtle embryo of 10-mm. carapace length. Iron-hematoxylin. X 500. Fig. 15 Cells from the peripheral cell clusters of the nervus terminalis of a turtle embryo of 10-mm. carapace length. Iron-hematoxylin. X 500. Fig. 16 Cells from the central root clusters of the nervus terminalis of a turtle embryo of 10-mm. carapace length. Iron-hematoxylin. > 500. 438 O. LARSELL yet reached the same degree of development. This view appears to be favored by the extremely crowded condition of these cells within the ganglia. This would appear to result in a reduction both of the amount of space and of nourishment which the indi- vidual cell may obtain, thus retarding its growth. The fact that the larger cells were found near the periphery of the ganglia, where there would appear to be space for greater expansion of the cells during growth as well as more abundant nourishment, may account for their larger size at this stage of development. There were also cells of intermediate size between the two groups, but these were not so numerous as the cells of either group. In the gasserian ganglion (fig. 11) of the older stages the number of small cells was not so great as in the spinal ganglia, but here also some were present, as the figure indicates. The sympathetic chain ganglia of the older stages, as in the younger, contained cells of rather uniform size and appearance (fig. 12). Two or three processes were observed on most of the cells which were examined on this point. The nuclei were smaller than those of the small spinal ganglion cells, showing an average size for fifty cells of 6.2 u, as compared with 7.3 » for the latter type. The relative amount of cytoplasm surrounding the nucleus ap- peared to be about the same in the two types. No difference in the size of the peripherally located cells, as compared with those situated nearer the centers of the sympathetic ganglia, was observed. The ciliary ganglia showed two groups of strongly contrasting cells. Without entering into a detailed description of this ganglion or attempting to review the large amount of litera- ture which has accumulated concerning it, the present pur- pose will be served by calling attention to Carpenter’s (’06) excellent study of it in the chick and adult fowl. He finds two distinct sizes of cells. The smaller cells were arranged in a definite group on the dorsal side of the ganglion. This group composed about one-third of the entire mass. A similar group of small cells (fig. 13, 4) was present in many of the turtle embryos, but not in all, studied by the present NERVUS TERMINALIS: TURTLE 439 writer. In all cases, however, whether or not arranged in a distinct group, small ganglionic cells were found in the ganglion. These differed not only in size, but also in form, from the larger more typical cells. Carpenter considers the large cells in the chick to be derived from the midbrain and to have migrated along the oculomotor nerve to the ciliary ganglion. The small cells he believes to be sympathetic cells which have migrated forward along the ophthalmic nerve. Fifty of the large cells were measured. The largest had a nuclear diameter of 12.4 yu, the smallest of 7 u. The average nuclear diameter of the fifty cells was 9 ». While the size of these nuclei approached that of the large spinal ganglion cell nuclei, and the largest found in the ciliary ganglion even ex- ceeded the largest observed in the spinal ganglia, the amount of cytoplasm surrounding the nuclei of the ciliary ganglion cells was considerably less, as may be seen by comparing figures 9 and 13, B. The actual size, therefore, of the ciliary ganglion cells was somewhat less than that of the large spinal ganglion cells. Twenty-two cells of the smaller type were measured. These were found in several adjacent sections, forming a distinct mass in the embryo on which the measurements were made. The largest of these cells had a nuclear diameter of 7.9 uw. The smallest was 5.3 », and the average of the twenty-two was 6.4 u. As figure 13 (A) indicates, there were also differences in the form of the cells and in the amount of cytoplasm surrounding the nucleus, as compared with the large cells. . There remain the cells of the sphenopalatine ganglion, which form a relatively small cluster. These cells are of quite uniform size and structure (fig. 14) with spheroidal, eccentrically located nuclei. Most of them were in the primitive bipolar stage of dif- ferentiation. Except that there was less individual variation among these cells than among those found in the clusters of the nervus terminalis, to be described, the sphenopalatine cells may be said to resemble the terminalis cells more closely than any of the other ganglionic cells studied. Of fifty sphenopalatine cells which were measured, the largest had a nuclear diameter of 8.1 ». The smallest was 5.3 », and the average for the fifty nuclei was 6.74 yu. THE JOURNAL OF COMPARATIVE NEUROLOGY, VOL. 30, NO. 5 440 O. LARSELL With these measurements as criteria, a comparison may be at- tempted between the ganglionic cells of the nervus terminalis and these other cells of well-recognized function, with the purpose in mind of throwing more light on the relationship of the nervus terminalis. Many of the terminalis cells (figs. 15 and 16) have a peculiar elongated form not met with elsewhere in the ganglia which were subjected to observation. This is well illustrated in some of the cells represented in figure 15. Only two processes were observed in any of these cells, and these were in most cases continuous with the longer axis of the cell. In view of the findings of McKibben (714) in the dogfish and of the various authors who have made observations on the ganglion cells of the terminalis in the mammals, it seems likely that this bipolar condition is developmental in the turtle. - The nuclei were elongated in many cases and many were also considerably distorted otherwise, as the figures indicate. This was more frequently the case in the peripheral clusters than in those within the cranial cavity. These conditions made the terminalis cells more difficult to measure, so that the results obtained represent the mean of several measurements on many of the nuclei. In order to determine the difference in size, if any, between the cells which were located in the ganglionic swelling near the central roots of the nerve and those along the olfactory and vomeronasal nerves outside the cranial cavity, the measurements were tabulated separately for the two groups. Thirty cells from the centrally located ganglion (fig. 16) indicated an average nuclear diameter of 6.72 uw, the largest being 8.4 u, and the smallest 5.3 « (table 1). The fifty cells of the peripheral clusters (fig. 15) which were measured indicated an average nuclear diameter for the entire number of 6.7 , practically the same as that of the centrally located cells. The largest nucleus of the peripheral clusters was 8.8 » in diameter, the smallest was 5.3 uw. alek.ot | 255.2 310.3 0.178 154765) 29. 15S 2505. 228.9), Gb.05 73.67 0.000 0.000 | 0.169 1.590 | 2.266 13.479 | 27.207 0.027 0.321 W017 | Selssul 2744 19.718 19.621 0.099 1.154 | 7.453 | 9.906} 5.666 | 40.058 | 50.604 0.304 2.951 | 17.774 | 25.896 | 25.8638 | 138.23 171.02 0.644 4,736 | 21.439 | 31.973 | 30.730 | 91.3805 103.65 0.263 2.106 | 11.116 | 13.485 | 9.811 13.892 | 24.363 0.171 1.440 | 5.362 | 6.228} 5.180 11.633 11.431 0.434 3.546 | 16.478 | 19.663 | 14.991 25.525 | 35.794 0.0005) 0.0060} 0.0207; 0.0627} 0.0620) 0.3870) 0.3880 0.0044} 0.0333) 0.1800} 0.2230} 0.3310) 0.6051) 0.4973 0.0026) 0.0214) 0.1007} 0.1456) 0.0814; 0.1898} 0.2623 0.0002) 0.0024) 0.0192} 0.0270) 0.0514; 0.0236) 0.1011 0.0077} 0.0631); 0.2706} 0.4583) 0.5248 1 555 1.2487 483 ADULT 67 YEARS 1297.9 1020.0 278.0 81.93 21.081 17.296 34.120 154.27 98.00 12.639 13.018 25.657 0.3453 0.7938 0.1767 0.0365 1.3525 484 C. G. MACARTHUR AND E. A. DOISY TABLE 12—Continued FETUS FETUS CHILD CHILD CHILD ADULT ADULT ADULT 3 7 1 3 8 21 35 67 MONTHS MONTHS MONTH MONTHS MONTHS YEARS YEARS YEARS Lipin phos- phorus .| 0.0075} 0.0643) 0.3747) 0.5020) 0.6206) 2.0980} 3.2550) 3.2360 Protein phos- phorus .| 0.0043} 0.0102) 0.0319} 0.0533) 0.0591} 0.1775) 0.1871} 0.1782 Organic phos- phorus .| 0.0044) 0.0333} 0.1639) 0.2969} 0.1600) 0.1489} 0.6054) 0.1262 Inorganic phos- phorus .| 0.0096} 0.0738} 0.3758} 0.5415) 0.2926) 0.7241) 0.6356) 0.7691 Total phos- phorus. ..| 0.0258] 0.1821) 0.9463) 1.2836) 1.1323} 3.9585} 4.6831) 4.3095 Figure 1 was plotted from the data in this table relating to the earlier period of growth. CHEMICAL CHANGES IN HUMAN BRAIN 485 TABLE 13 Whole brain: Milligrams added per day eae | Cate | Seoees | spine | Pacis | ore 3-MONTH | 7_yonTH | 1-monTH | 3-MONTH | 8-MONTH we REIS FETUS CHILD CHILD cuitp | 2/-YEAR Whole brain...............| 190.0 | 848.0 | 3764.0 | 2127.0 | 501.6 | 131.2 RRP on ca eos, con's al. 174.0 | 766.0 | 3270.0 | 1763.0 | 417.0 | 113.0 PA, ees eis sd = axa ciao oa 15.3 82.3 494.0 | 364.0] 84.6 18.2 PHOspHatids,. ..... 3... 2 1.98 | 10.80 85.3 34.9 | 26.3 5.4 WEreDrOsIdes... 20.5... i 6. 0.0 0.0 1.88 23.7 4.07 1.4 BON SIIOS, 424.005 06-0}. O90 2.45 7.73 35.4 2.99 2.2 REO RUEE Ob.0: 5. $6.0 Sun s.0 oe « 1.10 8.79 70.0 40.9 0.74 4.4 Marana cs cc) eee | a 162 er ab Tae ts 4 Watal proveins. 2.0.5.3 6..2) U7216 1 9Sae 185.6 | 175.6] 38.5 5.1 Organic extractives......| 2.81 | 15.4 100.1 38.7 7.2 | —0.6 Inorganic extractives.... 1.90 | 10.6 43.6 14.4 5.2 | —0.3 Total extractives........| 4.81 | 26.0 143.7 53.1 | 11.4 | —0.3 Lapin.sulphur.........<...| 0,006) 0,05 0.16 0.70} 0.08 0.04 Protein sulphur. ....3..: 0.049) 0.24 1.07 1.55} 0.61 0.0 Neutral sulphur.......... 0.029) 0.16 0.88 0.75} 0.01 0.0 Inorganic sulphur........| 0.002} 0.02 0.19 0.13) 0.11 | —0.01 Doral sulphur: ....22.......)" 0.086) 0:47 2.30 3.13} 0.81 0.03 Lipin phosphorus........ 0.083} 0.47 3.45 2.12} 1.01 0.27 Protein phosphorus...... 0.048) 0.081 0.24 0.36; 0.09 0.01 Organic phosphorus...... 0.049) 0.24 1.45 2.221 0.0 | —0.02 Inorganic phosphorus....| 0.107) 0.54 3.36 0.93) 0.17 0.03 Total phosphorus.......... 0.287; 1.30 8.50 5.62) 1.27 0.29 A part of these data are plotted in graph 2. TABLE 14 Whole brain: Average percentage increase per day 3-MONTH | 7-MONTH FETUS FETUS 1-MoNnTH | 3: MONTH | 8-MONTH Wihole brainer (4 tascey- ces erick ores Dao igs 0.88 0.26 0.028 Witenes cere eee oer emcee ae UAT 0.88 0.23 0.04 Solids 4:2 Satie scien ae eee ane 2.4 her 0 0.36 0.046 Phosphatids) ae OR Pp eae (LC | 1.59 0.62 0.093 | 0.014 Proteins. . RA Rees A Ae ES he ee a 1.42 0.66 0.087 | 0.007 Organic SEU Teh eee lee 1.52 0.32 0.046 | 0.0 Inorganic Pein Verne ROTI ieearie hig! sa? 1.28 0.25 0.067 | 0.0 WotalextTactivieszecns. ch cece cecere|) NuoO 1.44 0.29 0.048 | 0.0 Calculated from data in table 11. The average number of milligrams added _ per day was divided by the average weight for the given period, instead of the weight at the beginning of the period, as is usually done. When there are rapid changes in weight, this method is not as accurate as that used in table 12. It is believed that the temporary rise in growth at about the seventh month of fetal life is due to the method of calculation. The curve marked extractives (c) in graph 3 was plotted from the shove data. 486 SUBJECT AND AUTHOR INDEX CTIVITY of the nervous system. IIT. On the amount of non-protein nitrogen in the brain of albino rats during twenty-four hours after feeding. Meta- DO Lee eabinee tes ree aia eile wots: Siesta sila n'occ oc elevoe 397 Albino mouse. The nervus facialis of the.... 81 ——— rat in Miiller’s fluid. Factors influ- encing the behavior of the brain of the... 411 ——— rats during twenty-four hours after feeding. Metabolic activity of the nerv- ous system. III. On the amount of non-protein nitrogen in the brain of...... 397 Auten, Witu1AM F. Application of the Mar- chi method to the study of the radix mesencephalica trigemini in the guinea- PARE MMOOPEE RE che Pek cicl atc ne'a sie p cuatlacs fapanione sieges 69 Axis, Epwarp Pueps, Jr. The ophthal- mic nerves of the gnathostome fishes..... 69 Arry, Lresytiz B. A retinal mechanism of iT TAT a CV) OS eee eee ee eee 343 Ayers, Howarp. Vertebrate cephalogenesis. V. Transformation of the anterior end of the head, resulting in the formation of the “IEEE = SoReal AO ee Oe ae neers 323 EHAVIOR of the brain of the albino rat in Miiller’s fluid. Factors influencing — RUA st tint gl oaiels sels serrate sta ry lieinels 411 Brain during growth. Quantitative chemical Charnes AN The MUMIAN: oo oso a,< cies eas oe les 445 ——— of albino rats during twenty-four hours after feeding. Metabolic activity of the’ nervous system. III. On the amount of non-protein nitrogen in the.............. 397 —— of thealbinoratin Miiller’s fluid. Fac- tors influencing the behavior of the...... 411 ELLS in normal, subnormal, and senes- cent human cerebella, with some notes on functional localization. A prelimi- nary quantitative study of the Purkinje. 229 ——— tunnel space, and Nuel’s spaces in the organ of Corti. The development of the PUNT: ARG E cS Be RD ie Ae eR 283 Cell with especial consideration of the ‘ Golgi- net’ of Bethe, nervous terminal feet and the ‘nervous pericellular terminal net’ of Held. On the finerstructure of the syn- apse of thesMawphner, ccc nes esate seen 127 Cephalogenesis. IV. Transformation of the anterior end of the head, resulting in the formation of the ‘nose.’ Vertebrate...... 323 Cerebella, with some notes on functional lo- ealization. A preliminary quantitative study of the Purkinje cells in normal, sub- normal, and senescent human............ 229 Changes in the human brain during growth. Quantitative chemical..................-.- 445 Chemical changes in the human brain during ptowthn. . Quantitativesss. ec. tere. 445 Corti. The development of the pillar cells, tunnel space, and Nuel’s spaces in the OTE OL es are olatarsraisteie a UNtay e Oteele sete 283 | Be aes rec none of the pillar cells, tun- nel space, and Nuel’s spaces in the Organon Corti, Thess ives. o le 283 Doisy, E.A., MacArruur, C.G.,and. Quan- titative chemical changes i in the human brain during @rowth. i046 ).02i5..<0 0 foes ts 445 FFICIENT vision. A retinal mechanism OLR shea ask, PUES ee Te Oe aL 343 Evuis,.Rosert §. A preliminary quantita- tive study of the Purkinje cells in normal, subnormal, and senescent human cere- bella, with some notes on functional lo- CHlazatOM etc. css chemother eee 229 ACIALIS of the albino mouse. The ner- WLI Ss een Rite ene fad cetepete ate nters ik oriore ie ane 81 Factors influencing the behavior of the brain of the albino rat in Miiller’s fluid......... 411 Fiber in teleosts. Concerning Reissner’s..... 217 Fishes. Theophthalmic nerves of the gnatho- SUOWMG Hs alae seve insa piesa velpia'a ecaraverg oe eine ala aiata ates Fluid. Factors influencing the behavior of the brain of the albino rat in Miiller’s.... 411 Formation of the ‘nose.’ Vertebrate cepha- logenesis. IV. Transformation of the anterior end of the head, resulting in the.. 328 Functional localization. A preliminary quan- titative study of the Purkinje cells in nor- mal, subnormal, and senescent human cerebella, with some notes on............. 229 (cena fishes. The ophthal- NIG TOE VERIOL HE <2 5 aero ccc nat eens sears 69 Golgi. Frontispiece. Portrait of Professor Canalo sine oki Peete t sei e niece aes Pore ee ree 168 ‘Golgi-net’ of Bethe, nervous terminal feet and the ‘nervous pericellular terminal net’ of Held. On the finer structure of the synapse of the Mauthner cell with es- pecial consideration of the................ 127 Growth. Quantitative chemical changes in the human: Pram Gurney. «. sate vinias eee 445 Guinea-pig. Application of the Marchi method to the study of the radix mesen- cephalica trigemini in the................ 169 EAD, resulting in the formation of the ‘nose.’ Vertebrate cephalogenesis. IV. erppetormation of the anterior end of Held. sym the finer structure of the synapse of the Mauthner cell with especial con- sideration of the ‘Golgi-net’ of Bethe, nervous terminal feet and the ‘nervous pericellular terminal net’ of.............. 127 Human cerebella, with some notes on func- tional localization. A preliminary quan- titative study of the Purkinje cells in normal, subnormal, and senescent........ 229 ORDAN, Hovey. Concerning Reissner’s GHEMUBIEEIGOSER hes :s Sou casa sy leccmien a ie 217 OMINE, Suicryuxt. Metabolic activity of the nervous system. III. On the amount of non-protein nitrogen in the brain of albino rats during twenty-four INOUTH/AIter LEGGING: |... .. ss. secs eelerns =e 397 ARSELL, Otor. Studies on the nervus terminalis: Mammals................... Studies on the nervus terminalis: GUTTAG Ra ere a ei ac Giclee Sayeiace Ws te etal 423 487 488 Localization. A preliminary quantitative study of the Purkinje cells in normal, sub- normal, and senescent human cerebella, with some notes on functional............ 229 acARTHUR, C. G., and Dowy, E. A. Quantitative chemical changes in the human brain during growth.......... oo ET Mammals. Studies on the nervus terminalis. 1 Marchi method to the study of the radix mes- encephalica trigemini in the guinea-pig. Applicationiofthe.2. icc. ce se asee-- 2 169 Maruti, Kryoyasu. On the finer structure of the synapse of the Mauthner cell with es- pecial consideration of the ‘Golgi-net’ of Bethe, nervous terminal feet and the ‘nervous pericellular terminal net’ of Held. 127 ——— The effect of over-activity on the mor- phological structure of the synapse....... 253 Mauthner cell with especial consideration of the ‘Golgi-net’ of Bethe, nervous termi- nal feet and the ‘nervous pericellular termi- nal net’ of Held. On the further struc- ture of the synapse of the.............. vas, 127 Mesencephalica trigemini in the guinea-pig. Application of the Marchi method to the Bud yO the TAGIX.. 50. <2a5 +e eos 169 Metabolic activity of the nervous system. III. On the amount of non-protein ni- trogen in the brain of albinorats during twenty-four hours after feeding........... 397 Morphological structure of the synapse. The effect of over-activity on the............. 253 Mouse. The nervus facialis of the albino..... 81 Miiller’s fluid. Factors influencing the be- havior of the brain of the albino rat in... 411 ERVES of the gnathostome fishes. Ophthalmic vecses15 occ ve tee tee a eer 69 Nervoussystem. III. Ontheamount of non- protein nitrogen in the brain of albino rats during twenty-four hours after feed- ing. Metabolic activity of the........... 397 ——— terminal feet and the ‘nervous peri- cellular terminal net’ of Held. On the finer structure of the synapse of the Mauthner cell with especial consideration of the ‘Golgi-net’ of Bethe............... 127 Nervus facialis of the albino mouse. The.... 81 terminalis: Mammals. Studiesonthe 1 —— turtle. Studies on the................ 423 Net’ of Held. On the finer structure of the synapse of the Mauthner cell with especial consideration of the ‘Golgi-net’ of Bethe, nervous terminal feet and the ‘nervous pericellular terminal....................+. 127 Nitrogen in the brain of albino rats during twenty-four hours after feeding. Meta- bolic activity of the nervous system. IIT. On the amount of non-protein............ 397 Non-protein nitrogen in the brain of albino rats during twenty-four hours after feed- ing. Metabolic activity of the nervous system. III. On the amount of.......... 397 ‘Nose.’ Vertebrate cephalogenesis. IV. Transformation of the anterior end of the head, resulting in the formation of the... 323 Nuel’s spaces in the organ of Corti. The de- velopment of the pillar cells, tunnel space, and GATA, D., and Vincent, SwALe. A con- tribution to the study of vasomotor reflexes. 21, 5. o552 sane eeyl eee Peek ted 355 Ophthalmic nerves of the gnathostome fishes. ests 1S ie Re ek = see nee 69 Organ of Corti. The ge pele of the pillar cells, tunnel space, and Nuel’s BDACEBIN CUE. : hot te.s eee tre ue ah 283 Over-activity on the morphological structure of thesynapse. The effect of............ 253 INDEX ILLAR cells, tunnel space, and Nuel’s spaces in the organ of Corti. The de- velopment of the, .......25<-2:s)e+- oon . 283 PLANT, JAMES Stuart. Factors influencing the behavior of the brain of the albino ratin Miller's fuid.:.. 2s,.-+- 7a 411 Purkinje cells in normal, subnormal, and senescent human cerebella, with some notes on functional localization. A pre- liminary quantitative study of the....... 229 R225 mesencephalica trigemini in the guinea-pig. Application of the Marchi _ Inethod to the study of the............. 169 Ratin Miiller’s fluid. Factors influencing the behavior of the brain of the albino....... 411 Rats during twenty-four hours after feeding. + Metabolic activity of the nervous system. III. On the amount of non-protein ni- trogen in the brain of albino............. 397 Reflexes. A contribution to the study of VASOMOLOL:... asks csen ashe +s cee eee 355 Reissner’s fiber in teleosts. Concerning...... 217 Retinal mechanism of efficient viaion. A.... 343 RuineHart, D. A. The nervous facialis of theialbino MOuses..4. hi >- seen eee PACE, and Neul’s spaces in the organ of Corti. The development of the pillar cells; tunnel e710 250 ee oe eee 283 Spaces in the organ of Corti. The develop- ment of the pillar cells, tunnel space, and Nuelist. cece ou bee mak wean eee 283 Structure of the synapse of the Mauthner cell with especial consideration of the ‘ Golgi- net’ of Bethe, nervous terminal feet and the ‘nervous pericellular terminal net’ of Held: On the: finer. s3.b-<-ee tee eee 127 Structure of the synapse. The effect of over- activity on the morphological............ 253 Synapse of the Mauthner cell with especial consideration of the ‘ Golgi-net’ of Bethe, «nervous terminal feet and the ‘nervous pericellular terminal net’ of Held. On the finer structureiof the) 7.2 /2-.- oes 127 The effect of over-activity on the mor- phological structure of the................ 253 System. III. On the amount of non-protein nitrogen in the brain of albino rats during twenty-four hours after feeding. Meta- bolic activity of the nervous.............. 397 ELEOSTS. Concerning Reissner’s fiber LTE 2 . teae = folk. aac: ee 217 Terminalis: Mammals. Studiesonthenervus 1 turtle. Studies on the nervus......... 423 Terminal net’ of Held. On the finer struc-. ture of the synapse of the Mauthner cell with especial consideration of the ‘Golgi- net’ of Bethe, nervous terminal feet and the ‘nervous pericellular..............-.20 127 Transformation of the anterior end of the head, resulting in the formation of the ‘nose.’ Vertebrate cephalogenesis. IV.. 323 Trigemini in the guinea-pig. Application of the Marchi method to the study of the radix mesencephalica..-......c.. 1c. ocetee 169 Tunnelspace;and Nuel’sspacesin the organ of Corti. The development ofthe pillarcells 283 Turtle. Studies on the nervusterminalis.... 423 AN DER STRICHT, O. The develop- ment of the pillar cells, tunnel space, and Nuel’s spaces in the organ of Corti. 283 Vasomotor reflexes. A contribution to the SHUdVIO! So. cu ol Ss he ees La ee 355 Vertebrate cephalogenesis. IV. Transforma- tion of the anterior end of the head, re- sulting in the formation of the ‘nose’:... 323 Vincent, SWALE, Ocata, D., and. * 48 + K) . *, 4 stats 4 * i . < ys - ot « ts aes” eat. 4 2 Pe 55 i state Satitete at ‘% 4 re ae " Rot Oe © ee ee ARS S ey AA ? ¥ os 56 ~ 4s riers tate. a ae +s * #5 a a $a°,* Ay 24 »~ & be Perry toes, ‘ . ee a ta .*. PX) $$ 2-4 1 .