LIBRARY 
 
 OF THK 
 
 UNIVERSITY OF CALIFORNIA. 
 
 BIOLOGY 
 LIBRARY 
 
 G Class 
 
 
ATLAS OF NERVE CELLS 
 
ATLAS OF NERVE CELLS 
 
 BY 
 
 M. ALLEN STARR, M.D., PH.D. 
 
 PROFESSOR OF DISEASES OF THE MIND AND NERVOUS SYSTEM, COLLEGE OF PHYSICIANS AND SURGEONS, MEDICAL DEPARTMENT OF 
 
 COLUMBIA COLLEGE; CONSULTING NEUROLOGIST TO THE PRESBYTERIAN AND ORTHOPEDIC HOSPITALS 
 
 AND TO THE NEW YORK EYE AND EAR INFIRMARY 
 
 WITH THE CO-OPERATION OF 
 
 OLIVER S. STRONG, A.M., PH.D. 
 
 TUTOR IN BIOLOGY, COLUMBIA COLLEGE 
 AND 
 
 EDWARD LEAMING, M.D., F.R.P.S. 
 
 INSTRUCTOR IN PHOTOGRAPHY, COLLEGE OF PHYSICIANS AND SURGEONS 
 
 WITH FIFTY-THREE PLATES AND THIRTEEN DIAGRAMS 
 
 V OF THE 
 
 UNIVERSITY 
 
 OF 
 
 Neto 
 
 PUBLISHED FOR THE COLUMBIA UNIVERSITY PRESS BY 
 
 MACMILLAN AND CO. 
 
 AND LONDON 
 1896 
 
 All rights reserved 
 
GENERAL 
 
 COPYRIGHT, 1890, 
 BY MACMILLAN AND CO. 
 
 XortoooU Prtas 
 
 J. 8. Cuihlng & Co. Berwick <i Smith 
 Norwood Mui. U.S.A. 
 
PREFACE 
 
 IT is the object of this atlas to present to students and teachers of histology a series 
 of photographs showing the appearance of the cells which form the central nervous system, 
 as seen under the microscope. These photographs have been made possible by the use of 
 the method of staining invented by Professor Camillo Golgi of Pavia. This method has 
 revealed many facts hitherto unknown, and has given a conception of the structure and con- 
 nections of the nerve cells both novel and important. In the light of these facts it has 
 been necessary to discard many of the views previously taught by anatomists, and to ' revise 
 some of the physiological and pathological data supposed to be fundamental. 
 
 The nervous system is now known to be composed of a vast number of independent 
 units, called neurons, which consist of a cell body with two varieties of branches called den- 
 drites and neuraxons. The cell bodies vary in size, shape, and appearance ; their dendrites, 
 formerly known as protoplasmic processes, present great differences in form, length, and man- 
 ner of subdivision ; their neuraxons, formerly called axis cylinder processes and believed to 
 have no branches, are now known to give off many little collateral offshoots as important 
 as the main trunk. 
 
 The arrangement of these neurons varies greatly in different parts of the nervous system. 
 In the spinal cord they are collected into groups arranged in a long cylindrical column. In 
 the cerebral axis they are scattered among the various nerve tracts as well as collected into 
 separate groups. In the basal ganglia they are gathered into large masses separable into 
 divisions. In the cortex of the cerebrum and cerebellum they are spread out into thin but 
 very extensive layers containing a great variety of cells. 
 
 The interrelation of these neurons is also a subject of importance which recent researches 
 have demonstrated satisfactorily for the first time. The old theory that the processes of 
 adjacent cells joined together, forming everywhere a fine network of nerve fibres within the 
 gray matter, has been discarded. For the method of Golgi has shown that each cell is an 
 independent entity, its branches and subbranches of both varieties preserving their identity 
 from origin to ending, interlacing, it may be, with those of other cells, as the branches of 
 trees in a forest may interlace, but really as distinct and separable from each other as are 
 those trees with their twigs and leaves. 
 
 175644 
 
vi PREFACE 
 
 In the most recent text-books of neurology and in the atlas of Golgi these facts have 
 been shown by drawings and diagrams. But all such drawings are necessarily imperfect and 
 involve a personal element of interpretation. It has seemed to me, therefore, that a series of 
 photographs presenting the actual appearance of neurons under the microscope would be not 
 only of interest but also of service to students. The Golgi method lends itself very readily 
 to the photographic process, for the cell with its dendrites and neuraxon is stained black 
 upon a light yellowish ground, and thus is capable of giving a sharp picture. In the prepa- 
 ration of this atlas I have had the cooperation of Dr. O. S. Strong, who has cut and 
 stained the specimens, and of Dr. Edward Learning, whose skill in photography has made this 
 work possible. Dr. Strong has been able to produce remarkably successful sections of the 
 various parts of the nervous system, both brain and spinal cord, and has made some valuable 
 modifications of Golgi's methods. He has contributed a section upon the technique contain- 
 ing many original and important suggestions. In the art of photographing microscopic speci- 
 mens Dr. Learning has been particularly successful. It can be readily imagined that the 
 difficulties of obtaining a clear picture focussed in one plane upon the photographic plate are 
 at times almost insuperable, the microscopist ordinarily bringing various planes into his vision 
 by the aid of the fine adjusting screw of the instrument. By care in the selection of speci- 
 mens, by ingenious contrivances to insure a perfect focussing, and by the use of various 
 methods adapted to each emergency, Dr. Learning has succeeded where others have failed. 
 He has contributed a section of much value upon the photographic technique. The photo- 
 graphs have been reproduced in a painstaking manner by Mr. Edward Bierstadt, whose 
 process of artotyping has been selected after a careful comparison with other methods of 
 reproduction; and it can be justly said that they show every detail of the original photo- 
 graphs. 
 
 In preparing this atlas I have not attempted to write an exhaustive account of nerve 
 histology, but rather to present a brief review of the essential facts which can be demon- 
 strated by the aid of the Golgi stain, and to show how these facts aid in the knowledge of 
 nervous action. There are other methods of investigation, notably the method of Nissl, which 
 cannot be demonstrated by photography, and which reveal facts of equal importance, but these 
 must be sought elsewhere. I may be permitted, however, to point out that this atlas is based 
 mainly upon preparations from the human nervous system ; that it not only includes the 
 spinal cord, cerebellum, and brain cortex, which have been studied by Golgi, Cajal, Van 
 Gehuchten, Retzius, and Lenhossek, but also presents original studies of the corpora quadri- 
 gemina, optic thalamus, and lenticular and caudate nuclei, and is thus quite complete in its 
 scope. It is my intention at some future time to issue another volume which will include 
 .the peripheral nerves and their terminations and the organs of sense. 
 
 M. ALLEN STARR. 
 
TABLE OF CONTENTS 
 
 -*- 
 
 PAGE 
 
 THE HISTOLOGICAL TECHNIQUE By O. S. STRONG, Pn.D I 
 
 THE PHOTOGRAPHIC TECHNIQUE. By EDWARD LEAMING, M.D. ............. 12 
 
 THE SPINAL CORD. Plates I. to XIII 15 
 
 THE CEREBELLUM. Plates XIV. to XXI 34 
 
 THE CORPORA QUADRIGEMINA. Plates XXII. to XXIV 46 
 
 THE OPTIC THAI.AMUS. Plates XXV. to XXX 51 
 
 THE CORPORA STRIATA. Plates XXXI. and XXXII 60 
 
 THE CEREBRAL CORTEX. Plates XXXIII. to LIII 62 
 
 vii 
 
LIST OF PLATES 
 
 PLATE FACE PAGE 
 
 I. HUMAN SPINAL CORD; SECOND DORSAL SEGMENT. WEIGERT STAIN 15 
 
 II. HUMAN SPINAL CORD OF EMBRYO, EIGHTH-MONTH. GOLGI STAIN 20 
 
 III. LARGE CELL OF THE ANTERIOR HORN OF THE SPINAL CORD OF A FCETAL PIG 21 
 
 IV. LARGE CELL OF THE SPINAL CORD OF A FCETAL PIG 22 
 
 V. LARGE CELL OF THE ANTERIOR HORN OF THE SPINAL CORD OF A CHICK EMBRYO 23 
 
 VI. A GROUP OF MOTOR CELLS IN THE ANTERIOR HORN OF THE SPINAL CORD OF A FCETAL PIG .... 24 
 
 VII. LARGE CELLS OF THE MEDIAN PORTION OF THE SPINAL CORD OF A HUMAN EMBRYO, EIGHTH-MONTH ... 25 
 
 VIII. LARGE POLYGONAL CELL OF THE SPINAL CORD OF A HUMAN EMBRYO, EIGHTH-MONTH, IN LONGITUDINAL SECTION 26 
 
 IX. SECTION THROUGH THE POSTERIOR COLUMNS AND BASE OF THE POSTERIOR HORN OF THE ADULT HUMAN SPINAL 
 
 CORD 27 
 
 X. THE POSTERIOR HORN OF THE SPINAL CORD OF A HUMAN EMBRYO, EIGHTH-MONTH, SHOWING THE SUBSTANCE OF 
 
 ROLANDO ......_ 28 
 
 XI. SECTION THROUGH A POSTERIOR SPINAL GANGLION AND SPINAL CORD OF A CHICK EMBRYO, SEVENTH-DAY . . 29 
 
 XII. LONGITUDINAL SECTION THROUGH THE SPINAL CORD OF A CHICK EMBRYO, SEVENTH-DAY 30 
 
 XIII. NEUROGLIA CELLS IN A LONGITUDINAL SECTION OF THE SPINAL CORD OF A NEW-BORN CHILD 33 
 
 XIV. PURKINJE CELL OF ADULT HUMAN CEREBELLUM 34 
 
 XV. PURKINJE CELL OF ADULT HUMAN CEREBELLUM 35 
 
 XVI. PURKINJE CELL OF ADULT HUMAN CEREBELLUM 36 
 
 XVII. MOLECULAR LAYER OF THE CEREBELLUM, SHOWING STELLATE CELLS 37 
 
 XVIII. MOLECULAR LAYER OF THE CEREBELLUM, SHOWING STELLATE CELLS WITH BASKET FIBRES 38 
 
 XIX. SMALL GRANULE CELLS IN THE GRANULAR LAYER OF THE CEREBELLUM 39 
 
 XX. SECTION THROUGH THE CORPUS DENTATUM OF THE CEREBELLUM . . 41 
 
 XXI. SECTION THROUGH THE CEREBELLAR CORTEX, SHOWING NEUROGLIA CELLS AND THEIR VERTICAL FIBRES . . 45 
 
 XXII. SECTION THROUGH THE ANTERIOR CORPUS QUADRIGEMINUM, SHOWING THE VARIOUS LAYERS OF WHITE AND GRAY 
 
 MATTER 46 
 
 XXIII. THE CORPUS QUADRIGEMINUM. LARGE POLYGONAL CELLS OF THE SECOND LAYER 49 
 
 XXIV. THE CORPUS QUADRIGEMINUM. SMALL AND TRIANGULAR CELLS AND FINE PLEXUS OF NERVE FIBRES IN THE 
 
 THIRD LAYER 50 
 
 XXV. MEDIUM-SIZED STELLATE CELLS OF THE OPTIC THALAMUS, MEDIAN NUCLEUS 54 
 
 XXVI. LARGE STELLATE CELLS OF THE OPTIC THALAMUS, LATERAL NUCLEUS 55 
 
 XXVII. LARGE POLYGONAL CELLS OF THE OPTIC THALAMUS, VENTRAL NUCLEUS 56 
 
 XXVIII. LARGE POLYGONAL CELLS OF THE OPTIC THALAMUS, VENTRAL NUCLEUS 57 
 
 XXIX. SMALL TRIANGULAR CELLS OF THE NUCLEUS HABENUL/E OF THE OPTIC THALAMUS 58 
 
 XXX. SECTION THROUGH THE MARGIN OF THE OPTIC THALAMUS, SHOWING THE ENTRANCE OF ITS FIBRES INTO THE 
 
 INTERNAL CAPSULE 59 
 
 XXXI. LARGE RECTANGULAR CELLS OF THE CORPUS STRIATUM . . 60 
 
 XXXII. SMALL TRIANGULAR AND STELLATE CELLS OF THE CORPUS STRIATUM 61 
 
x LIST OF PLATES 
 
 L 
 
 PLATE FACE PAGE 
 
 XXXIII. SECTION THROUGH A CONVOLUTION OF THE HUMAN BRAIN, FROM AN EMBRYO, EIGHTH-MONTH .... 62 
 
 XXXIV. THE SUPERFICIAL LAYER OF THE CEREBRAL CORTEX OF THE HUMAN EMBRYO, SHOWING CAJAL CELLS ... 64 
 XXXV. THE SUPERFICIAL LAYER OF THE CORTEX, HUMAN EMBRYO, EIGHTH-MONTH, SHOWING CAJAL CELL ... 64 
 
 XXXVI. THE SUPERFICIAL LAYER OF THE CEREBRAL CORTEX, HUMAN EMBRYO, EIGHTH-MONTH, SHOWING CAJAL CELLS . 64 
 
 XXXVII. THE SUPERFICIAL LAYER OF THE CORTEX, SHOWING TRIANGULAR CAJAL CELL 65 
 
 XXXVIII. THE SECOND LAYER OF THE HUMAN CORTEX OR LAYER OF SMALL PYRAMIDAL CELLS 66 
 
 XXXIX. THE SECOND LAYER OF THE HUMAN CORTEX OR LAYER OF SMALL PYRAMIDAL CELLS 66 
 
 XL. PYRAMIDAL CELLS OF THE ADULT HUMAN CEREBRAL CORTEX 67 
 
 XLI. SECTION THROUGH THE SECOND AND THIRD LAYERS OF THE HUMAN CORTEX 68 
 
 XLII. PYRAMIDAL CELLS OF THE CEREBRAL CORTEX, HUMAN EMBRYO, EIGHTH-MONTH 68 
 
 XLIII. PYRAMIDAL CELLS OF THE CEREBRAL CORTEX, HUMAN EMBRYO, EIGHTH-MONTH 68 
 
 XLIV. MEDIUM-SIZED PYRAMIDAL CELLS OF THE THIRD LAYER OF THE CORTEX, HUMAN EMBRYO 68 
 
 XLV. LARGE CELL OF GOLGI'S SECOND TYPE. IN THE THIRD LAYER OF THE CORTEX 68 
 
 XLVI. LARGE CELL OF GOLGI'S SECOND TYPE, IN THE THIRD LAYER OF THE CORTEX 68 
 
 XLVII. LARGE PYRAMIDAL CELLS OF THE THIRD LAYER OF THE CORTEX 70 
 
 XLVIII. A MARTINOTTI CELL OF THE THIRD LAYER OF THE CORTEX 71 
 
 XLIX. THIRD LAYER OF THE CORTEX, SHOWING IRREGULAR POLYGONAL CELLS WHICH LIE AMONG THE PYRAMIDAL CELLS 71 
 
 L. FUSIFORM CELL OF THE FOURTH LAYER OF THE CORTEX 72 
 
 LI. A SECTION THROUGH THE HIPPOCAMPUS OF A HUMAN EMBRYO, EIGHTH-MONTH 76 
 
 LII. THE VARIOUS LAYERS OF THE CORTEX OF THE HIPPOCAMPUS 77 
 
 L1II. THE CELLS OF THE CEREBRAL CORTEX, SECTIONS AT VARIOUS REGIONS (HAMMARBERG) ..... 78 
 
 LIST OF FIGURES 
 
 FIGURE PAGE 
 
 1. TRANSVERSE SECTION OF THE SPINAL CORD 16 
 
 2. DIAGRAM TO SHOW THE TRANSMISSION OF SENSORY, MOTOR, AND ASSOCIATION IMPULSES THROUGH THE SPINAL 
 
 CORD '..... 18 
 
 3. DIAGRAM TO SHOW THE REFLEX ACTION OF THE SPINAL CORD . 24 
 
 4. DIAGRAM OF THE TRACTS IN THE SPINAL CORD . 31 
 
 5. THE NEUROGLIA CELLS AND FIBRES OF THE SPINAL CORD (LENHOSSEK) 33 
 
 6. DIAGRAM OF A SECTION THROUGH THE CEREBELLAR CORTEX 39 
 
 7. DIAGRAM TO SHOW THE CONNECTIONS OF THE CEREBELLUM 42 
 
 8. DIAGRAM OF THE CORPORA QUADRIGEMINA ANTERIOR AND THEIR CONNECTIONS 48 
 
 9. THE OPTIC THALAMUS (v. MONAKOW) 52 
 
 10. DIAGRAM OF THE CELLS OF THE CEREBRAL CORTEX 72 
 
 n. THE PROJECTION TRACTS OF THE CEREBRAL CORTEX 73 
 
 12. THE ASSOCIATION FIBRES OF THE CEREBRAL CORTEX 73 
 
 13. THE TERMINAL FIBRES WITHIN THE CORTEX (ANDRIEZEN; .74 
 
-,r THE 
 
 DIVERSITY 
 
 OF 
 
 THE HISTOLOGICAL TECHNIQUE 
 
 BY OLIVER S. STRONG, M.A., PH.D. 
 
 TUTOR IN BIOLOGY, COLUMBIA COLLEGE 
 
 THE Golgi methods of staining nerve cells and their processes are known as impregnation 
 methods in which metallic salts are precipitated in or around the elements, thus bringing them 
 into clear relief. The nerve cell and its processes appear as a black silhouette upon a yel- 
 lowish or whitish ground. Golgi has employed both silver and corrosive sublimate methods. 
 It is the former only which have been used in the preparation of the specimens shown in 
 this atlas. 
 
 In the silver methods the tissue is first hardened in a solution of potassium bichromate 
 and then brought into a solution of silver nitrate. If the tissue is in the proper stage of 
 hardening, the silver is deposited (probably principally in the form of silver chromate) in and 
 around certain of the nervous elements. As the remainder of the tissue remains unstained 
 the sharp contrast thus afforded enables one to trace the finest ramifications of the nervous 
 elements impregnated. 
 
 It is a peculiarity of these methods that all the elements present in the tissue are very 
 rarely impregnated at one time, and those elements which are thus picked out by the stain 
 are more or less different in each preparation. Owing to this peculiarity, the various nerve 
 cells with all their processes can be distinguished separately. Were all the cells and processes 
 present stained at once, there would be simply an inextricable mass of black-stained cells and 
 fibres. 
 
 The silver methods may be subdivided as follows : 
 
 (a] Golgis long method, in which potassium bichromate alone is used in hardening the 
 tissue before placing it in the silver nitrate. This process requires nearly a month to harden 
 the tissue properly. 
 
 (b] Golgi 's rapid method, in which a mixture of potassium bichromate and osmic acid is 
 used to harden. Here the time required is reduced to a week or less. 
 
 (c] Golgi's mixed method, in which the tissue is placed first in potassium bichromate for 
 from a few days to a month and then in the osmic-bichromate mixture. 
 
2 ATLAS OF NERVE CELLS 
 
 For all these methods we are indebted to Professor Camillo Golgi, of Pavia. The rapid 
 method is the most important, and is the one which is now so extensively used. It is espe- 
 cially applicable to embryonic tissue, to nerve endings, etc. 
 
 In addition to the above, the writer has devised two modifications which are quick and 
 yet avoid the use of osmic acid, which is an expensive reagent and difficult to manage ; viz. : 
 
 (d) The lithium bichromate method, in which lithium bichromate is used instead of potas- 
 sium bichromate. The results resemble those of the long method, but the time of hardening 
 is reduced from nearly a month to two days. 
 
 (e) The formalin-bichromate method, in which a mixture of potassium bichromate and 
 formalin ( = 40% solution of formaldehyde) is used in hardening. This also requires little 
 time to harden. It is more certain than (d}. Both (d) and (e) are especially applicable to 
 adult brains. 
 
 The actual method of procedure in making the preparations for the atlas will now be 
 given. Certain of the details were rendered necessary by the nature of the material and by 
 the fact that the preparations were to be photographed. 
 
 It will be most convenient to deal first with the preparations made from human embryos 
 and by means of the rapid method, as these constitute the bulk of the preparations used. 
 
 The first essential is that the material be tolerably fresh. All the material which yielded 
 good results was obviously, from its general appearance, in good condition. Exact data are 
 not available, but in general human embryos of seven and eight months, if kept cold, probably 
 remain in fairly good condition for at least 24 hours after death. 
 
 The hardening process. If the entire embryo is at hand, two courses are open : the 
 brain may first be injected in situ with hardening fluid, or may be immediately removed and 
 suitable pieces placed in the fluid. Injection was tried on one brain only, but the results war- 
 rant its further trial. In this case the fluid used was potassium bichromate 5% i vol. + osmic 
 acid 2 % i vol., the solution being made especially strong in osmic acid to allow for its sub- 
 sequent dilution in the tissue. About 100 cc. were injected into each carotid, and, when the 
 brain was opened an hour afterward, brown spots throughout the tissue showed where the 
 fixation had begun. The result was that in places groups of cells lying close to large 
 blood-vessels were impregnated, proving that the injection aided by infiltrating the tissue with 
 larger quantities of bichromate thus securing the formation of more silver chromate in these 
 places. Had the whole tissue been uniformly fixed in this way, which would have required 
 a much larger quantity of fluid, the result would probably have been still better. 
 
 After the injection is completed and an hour or so allowed for fixation to take place, 
 the brain is removed and suitable pieces placed in the hardening fluid. These pieces should 
 
THE HISTOLOGICAL TECHNIQUE 3 
 
 not exceed \ cm. in thickness whether from an injected or from a fresh brain. This size is 
 rendered necessary by the fact that the fixation by the hardening fluid extends inwards slowly, 
 and because the diffusion of the silver cannot take place properly beyond a certain depth. 
 This does not exclude the impregnation of large pieces, provided they are of the necessary 
 thinness. Care must also be taken to cut the pieces to be impregnated as accurately as pos- 
 sible in the plane in which it is desired to section them subsequently ; e.g. if it be desired to 
 make sections displaying the whole length of the pyramid cells of the cortex, the cuts should 
 be made perpendicular to the surface of the cortex; if the protoplasmic expansions of the cells 
 of Purkinje, they should be made at right angles to the convolutions of the cerebellum, etc. 
 If this precaution be not observed, much of the thin pieces impregnated will be wasted by 
 cutting off corners with oblique cuts, etc. Furthermore, the impregnation is liable to vary 
 at different depths from the cut surfaces of the piece, being especially full in details nearest 
 the surface ; hence it is obvious that sections passing parallel to these cut surfaces give 
 better and more uniform fields. 
 
 The surface of impregnated pieces is usually covered with silver chromate, which mars 
 the beauty of the preparations, and may also obscure details of structure lying near the 
 surface. To avoid this, it has been recommended to cover the surface of the pieces, before 
 hardening, with gelatine (Sehrwald) or with celloidin or blood (Cajal). Gelatine, however, 
 interferes with the impregnation. Egg albumen might be adapted for this purpose, and 
 should be smeared upon the surface (not necessarily upon the cut surfaces, however) either 
 before or during the hardening. 
 
 The hardening fluid usually employed is potassium bichromate 3! % 4 vols. + osmic 
 acid i % i vol. Instead of this a mixture was sometimes used practically the same as that 
 recommended by Berkley; i.e. osmic acid 2% 16 vols. + potassium bichromate 5 % 84 vols. 
 It is difficult to determine whether this be better than the ordinary mixture. Possibly its 
 use results in the impregnation of a greater number of elements, owing to the fact that 
 there is more bichromate of potassium in the tissue available for the formation of silver 
 chromate. 
 
 Another requisite to success is that a sufficient quantity be used. It is not easy to 
 give an absolute rule regulating the proportion between the amount of tissue and the quantity 
 of fluid. The pieces may first be placed in a small quantity of the fluid. This soon 
 becomes turbid, and after half an hour or an hour should be changed for a larger quantity 
 of fresh fluid. This should be examined now and then, and if at any time it should be 
 found to be turbid, or to have lost, or nearly lost, its characteristic smell of osmic acid 
 (accompanied often by a darkening of the fluid), it should be changed for fresh fluid. As 
 
4 ATLAS OF NERVE CELLS 
 
 an example, three pieces cm. thick, and about 3 cm. by 4 cm., after being in a small 
 quantity for an hour, were placed in a dish containing 100 cc. of the hardening fluid, which 
 had to be renewed the next day. This probably represents the maximum quantity of tissue 
 which could be hardened in such a quantity of fluid. 
 
 If the pieces have flat surfaces which lie quite closely applied to the bottom of the 
 dish, a small quantity of absorbent cotton should first be placed in the dish, so that the 
 hardening fluid may have easy access to all sides. 
 
 The crucial point in the method lies in the duration of hardening. How long the tissue 
 shall be hardened depends upon (a) the kind of impregnation aimed at, (b) upon the kind 
 of tissue, (f) upon the stage of development of the tissue, whether adult or embryonic, (d) upon 
 the kind of animal, and (e) upon the temperature in which the hardening takes place. 
 
 (a) According to the rule laid down by Cajal for the cords of embryo chicks, and by 
 Kolliker and Lenhossek for the human embryonic cord, the different elements are impreg- 
 nated, according to the degree of hardening, in the following order: (i) Neuroglia (2-3 days), 
 (2) Nerve cells (3-5 days), (3) Nerve fibres and collaterals (5-7 days). The times given apply 
 to the human cord. This is only true in a very general way. Both neuroglia and 
 nerve fibres usually appear in any preparations in which impregnation has taken place, irre- 
 spective of the duration of hardening. One factor of uncertainty, often overlooked, is caused 
 by differences in the ratio between the bulk of tissue and quantity of hardening fluid. In 
 proportion as the former is increased, the fluid becomes less efficient during the hardening, 
 and thus a dish containing a certain quantity of tissue and hardening fluid may really be 
 in a more advanced stage of hardening than a greater quantity of tissue placed in the same 
 amount of hardening fluid for the same period of time. 
 
 If the tissue be underhardened, the result is often that it is filled, after impregnation, 
 with a diffuse red precipitate of silver chromate. The effect of overhardening is usually a 
 complete absence of any precipitate of any kind in the tissue or, possibly, the formation of 
 numbers of clean, sharply defined black crystals. As the tissue progresses toward the maxi- 
 mum period of hardening, the tendency is for fewer elements, perhaps, to be impregnated, 
 but those that are stained are brought out with greater cleanness. The tendency is also 
 for the tissue to be more uniformly impregnated at all depths. This is probably due to 
 the fact that the tissue is hardened more uniformly throughout when left in the hardening 
 fluid longer. 
 
 (b) Respecting the kind of tissue, it is within the scope of the present work to deal 
 with the central nervous system only, and no very definite rule can be laid down, owing to the 
 factors just mentioned. I am inclined to believe that the cortex of the cerebellum requires 
 
THE HISTOLOGICAL TECHNIQUE 5 
 
 the briefest period of hardening, and that the cord possibly does not require quite the same 
 length of time that the basal ganglia and cerebrum do. 
 
 (c) The stage of development of the tissue is an important element in the duration of 
 hardening. In general, the older the tissue, the longer the period of hardening. Embryonic 
 tissue is much more favourable than adult tissue. Good impregnations of nerve fibres cannot 
 be looked for in tissue in which the fibres are medullated. It is not entirely true that the 
 axis-cylinder of a medullated nerve fibre will never impregnate, but such fibres are decidedly 
 unfavourable. In an eight-months human embryo, for example, many of the axis-cylinders of 
 the pyramidal tracts, in some of the preparations, were apparently impregnated, while the 
 medullated tracts of the cord and medulla had simply the faint brown stain due to the osmic 
 acid used in hardening. This indicates that for the best study of the cord and medulla, 
 younger material must be used than is necessary for the higher centres. The medullation 
 of fibres is not the only cause of the difference between embryonic and adult tissue, however, 
 for some cells impregnate better in embryonic tissue. This is possibly due to the different 
 consistency of the embryonic tissue and also to the fact that its cells are smaller. 
 
 (d) The kind of animal used is another factor to be taken into account. Cajal has 
 shown that in the mouse the elements of the cortex are very embryonic at birth, and that 
 the most favourable age is from the 8th to the 25th day after birth. In larger mammals 
 (e.g. the rabbit), he shows that this period is thrown back to between the ist and 8th days 
 after birth. This interesting tendency is still further exemplified in man ; for, as the atlas 
 shows, the cortical cells are well differentiated, though perhaps not in all respects completely 
 developed, one and two months before birth. Consequently, also, the period of hardening 
 would exceed that required for smaller mammals (with shorter gestation periods) of the same 
 age relative to birth. 
 
 (e) A higher temperature naturally accelerates the hardening. The temperature of the 
 room in which these impregnations took place rarely varied much from 21 C. 
 
 The best period of hardening in general for an eight-months human embryo, as can be 
 seen from the table below, is six days, though good results in the cortex were obtained before 
 this and also in tissue hardened eight days. Six days is also the best period for the medulla 
 and basal ganglia. While cells may be obtained in the cortex before six days of hardening, 
 six days is necessary to bring out the tangential fibres and Cajal cells. It is obvious, how- 
 ever, that it is best not to depend entirely upon any one period of hardening, but to transfer 
 pieces at various intervals of time into the silver nitrate. 
 
 Impregnation. After the hardening, the next step is .to bring the tissue into the solution 
 of silver nitrate. The solution used is a f % in water, but a | % or a i % solution may be 
 
6 ATLAS OF NERVE CELLS 
 
 employed without producing any difference in the results. In other lines of work I have 
 even used a 4 % and a 10 % solution, the effect apparently being to increase, sometimes, the 
 number of elements impregnated with a tendency to lose, to some extent, cleanness of 
 impregnation. 
 
 The pieces are brought immediately into the silver solution, and the first solution used may 
 be one that has been used in a previous impregnation. The pieces are simply rinsed in this 
 first solution, which need merely be sufficient to cover them. A copious precipitate is imme- 
 diately formed, and then the fluid is poured off and thrown away. When a second quantity 
 is poured upon the tissue, it usually clouds up slightly. This is again poured off, and the 
 pieces are now brought into the final solution. This may again become discoloured, especially 
 if the pieces be of some size, and should then be changed. The quantity of the solution in 
 which the pieces finally rest should be considerable, even double the quantity of osmic-bichromate 
 used in hardening. When brought into the final fluid, the pieces are placed upon absorbent 
 cotton to facilitate the diffusion of the silver nitrate into the tissue. 
 
 The impregnation begins very soon. Golgi says it may begin immediately at the surface 
 and be completed in favourable cases in 2-3 hours. It is generally completed in 12-24 hours. 
 When the pieces cannot be cut at once they should be kept in the silver solution, and this 
 should be changed whenever it shows a yellowish or greenish tinge. Tissues can be thus 
 kept in silver for several weeks or even longer, though it is wise to cut them sooner. 
 
 Usually both hardening and impregnation are done in the dark. This is not necessary 
 however. The light plays no part in the impregnation. When the pieces are to be kept in 
 the silver for some time, it is better to keep them in the dark. At any stage in the pro- 
 cedure they may be brought into the light temporarily for examination, etc., without any fear 
 of injury. 
 
 Besides the precipitation of silver chromate, and possibly of other silver compounds in 
 the tissue, another change effected by the silver, to which I have not observed any allusion, 
 is the clearing up of the blackening produced by the osmic acid in the hardening. 
 
 As above noted, one difficulty with the Golgi silver methods, is the slow and feeble 
 penetration of the osmic-bichromate (in the rapid method), and also of the silver nitrate, 
 so that impregnations are limited to a certain distance from the surfaces of the pieces 
 besides being often more unequal than is desirable. The writer sought to overcome the latter 
 and increase the diffusion of the silver by adding sulphates to the silver solution. The sulphates 
 of sodium and zinc when thus added did not interfere with the impregnation, but whether 
 they produced the desired effect to any extent is questionable. Their presence in some of the 
 silver solutions given below need not be regarded as one of the requisites to obtaining the 
 
THE HISTOLOGICAL TECHNIQUE 7 
 
 impregnation. The manner in which the impregnation is facilitated when the tissue is 
 placed in the silver by anything that stimulates diffusion, is shown in preparations by the 
 fact that wherever a crack or a break in the piece exists, in that vicinity the greatest number 
 of elements are impregnated. The addition of formic acid to the silver nitrate solution (one 
 drop to 200 cc.) as recommended by Cajal, has not proved of any advantage. 
 
 Another modification of the silver solution may be mentioned here. It is at times desirable 
 to double-stain preparations in order to bring out the unimpregnated elements. This may be 
 done by a quick staining of the sections in any suitable dye dissolved in 95 % alcohol. 
 Another method consists in adding the stain to the silver solution, the requirements being 
 that the stain has a neutral reaction and does not affect nor is affected by the silver nitrate 
 solution. Some of the anilines, e.g. acid fuchsin, and a neutral carmine solution, fulfil these 
 requirements. This gives of course an in toto stain. 
 
 Cutting and Mounting. Preparatory to cutting, the piece is transferred directly from the 
 silver solution into 95 % alcohol. It may be left in this an hour, or longer, according to its 
 size. The alcohol should be changed several times to free the tissue as far as possible from 
 the silver nitrate. Pieces of tissue impregnated by the rapid Golgi method cannot be 
 infiltrated with celloidin, as the prolonged immersion in alcohol, etc., is liable to injure the 
 impregnation ; nor, indeed, is it necessary, as sections can be readily cut of the desired thick- 
 ness without imbedding. Consequently the piece is placed for a few minutes in a thin and 
 then in a thick solution of celloidin, to give it support and to gum it upon the microtome 
 block. After placing it upon the block, the celloidin is hardened by immersion in chloro- 
 form for a few minutes. This should not be too prolonged, or the tissue will crack or 
 become friable. 
 
 The sections are cut in 95 % alcohol and kept in the same until cleared. They 
 should be cut as a rule from 50 /A to 100 /A thick. Thinner sections cut the nerve 
 cells and their processes into small pieces. Small sections can be removed from the knife 
 with a brush ; with larger ones, however, the slide must be brought up under the edge 
 of the knife and the section gently washed or pulled down upon it. The upper surface 
 of the knife should be as nearly level as possible, to avoid an accumulation of alcohol at the 
 edge. If it be found impossible, on account of its size, to cut the unimbedded piece 
 without considerable tearing, painting the surface of the block with a medium thick solution 
 of celloidin must be resorted to. Not much time need be lost by this procedure, especially 
 as sections are then more quickly removable from the knife. As the knife is pushed back over 
 the block after a cut, the surface of the block is wiped dry by the knife, and the celloidin 
 can then be applied to it with a brush. While the last section which has just been brought 
 
8 ATLAS OF NERVE CELLS 
 
 upon the slide, if it be desired to mount them serially as described below, is being placed 
 in position, the film of celloidin will dry sufficiently. Alcohol can then be applied and the 
 next section cut. 
 
 For clearing, oleum origani cretici is preferred. It clears readily from 95 % alcohol, 
 does not make the sections brittle, and sections may be kept in it for a considerable time, 
 even a day or so, without injury. It is better before mounting to wash them briefly in xylol. 
 The xylol allows the balsam to set more rapidly, and is less liable to affect the preparations 
 than is the origanum brought over into the balsam with the section. On the other hand, 
 xylol tends to make the sections crumpled and brittle. 
 
 In the handling and mounting, several methods may be employed. The sections may be 
 brought from the origanum upon the slide, the oil blotted off, and a drop of balsam placed 
 upon the section, the mount then being watched to see that the balsam does not run off. 
 This has the disadvantage that a small quantity of origanum remains in the section, usually 
 not enough to prevent the balsam from drying properly. Another mode of procedure is 
 to bring the sections from origanum into xylol, and thence into a dish of balsam. The 
 diffusion of balsam takes place while in this, and consequently when the section is lifted out 
 and placed upon the slide, the balsam does not tend to run off. The disadvantage here is 
 that sections of the central nervous system tend to become crumpled and brittle. 
 
 The best method, one which preserves the sections in serial order, and keeps them flat 
 for photographing, is the following. The sections are brought from the knife and arranged 
 on the slide in 95 % alcohol, and, after the slide is filled, the alcohol is blotted off and a 
 thin solution of celloidin is quickly run over the slide and drained off. After the film has 
 set by slightly drying, absolute alcohol is quickly run over the slide and drained off, thereby 
 facilitating the subsequent clearing. Care must be taken not to dissolve the film. The slide 
 is then immediately placed in origanum, where it may remain while the next slide is being 
 filled. It is then placed successively in two dishes of xylol to remove the origanum. If the 
 slide clouds up in xylol, it has not been completely dehydrated. Absolute alcohol may then 
 be again run over the slide, which may then be again cleared in origanum. A slight opacity 
 outside the sections will usually clear up after mounting. For the mounting medium a solu- 
 tion of damar in xylol is used. 
 
 The sections are mounted without a cover slip, and to do this properly is not as simple 
 as would at first sight appear. The sections should be covered with a layer of balsam, which, 
 when dry, should be even and as thin as possible and yet cover the sections. If a thin 
 layer be applied at first, it runs off in spots, leaving the sections dry, and requiring constant 
 watching, which interferes with other work. All this may be obviated by placing the slide, 
 
THE HISTOLOGICAL TECHNIQUE 9 
 
 after being brought from xylol, on a level table and flooding it with as much of a rather 
 thick solution of damar as it will hold without overflowing. It can then be covered over 
 so as to protect it from dust, and left at least half an hour and longer if convenient. Dif- 
 fusion of the balsam into the celloidin and sections will then be completed, and the superflu- 
 ous balsam may be drained back into the bottle. Just how much to leave on the slide 
 depends upon the thickness of the balsam and must be determined by experience. After 
 the superfluous balsam has been removed, the slides are put in an oven (5O-55 C.) over night 
 to harden the balsam. One of the essentials to a good preservation of preparations, to avoid 
 yellowing, etc., is that the balsam be quickly and completely hardened. 
 
 Besides the rapid method, certain of the preparations were made either by means of 
 Golgi's mixed method or by means of what might be termed modifications of Golgi's long 
 method, introduced by the writer. 
 
 Plates III., V., and VI. are taken from preparations made by means of Golgi's mixed 
 method. Here the cord was placed in Miiller's fluid (potassium bichromate alone answers 
 as well) for one week, then for several days in osmic-bichromate, and finally in silver nitrate. 
 This method was not used very extensively. It appears to be especially adapted to bringing 
 out the cells rather than the fine plexus of fibres, and is particularly applicable to adult 
 tissue. Cajal recommends it especially for the demonstration of the cells and transverse 
 fibres of the molecular layer of the cerebellum in young mammalia. It was the method 
 most strongly recommended by Golgi himself. This method, furthermore, permits the tissue to 
 be kept, as Golgi points out, for from two to thirty days in bichromate, from which, at con- 
 venient times, pieces can be removed and placed into osmic-bichromate and thence into silver 
 nitrate. Whether this could equal the rapid method with embryonic material has perhaps 
 been hardly sufficiently tested, but the quick fixation of the rapid method would appear to 
 be especially desirable with the more delicate and less firm embryonic tissue. 
 
 The long Golgi method has not been employed in preparations shown in this Atlas. 
 The impregnations made by the long method differ from those made by the rapid method in 
 the fact that they possess greater stability in strong alcohol, in which they may remain for a 
 considerable period without deterioration. The method is especially indicated for the adult 
 cerebrum and cerebellum, but one of the two modifications given below is probably easier. 
 
 It has been found by the writer that lithium bichromate hardens tissue much more 
 rapidly than the potassium salt, and, by employing the former, the period of hardening may 
 be cut down to a few days. 
 
 Since the introduction of formalin for hardening, Hoyer, Jr., discovered that material kept 
 in it could be afterwards impregnated like fresh material, and that the osmic acid could in 
 
10 ATLAS OF NERVE CELLS 
 
 this case be dispensed with. The writer found subsequently, but independently, that fresh 
 tissue could be placed immediately in a mixture of formalin and potassium bichromate, and 
 thence brought into the silver nitrate solution. Thus formalin may be substituted for osmic 
 acid, but it has to be used in much larger proportion. 1 
 
 The best proportions for the fluid and periods of hardening for this method have not 
 yet been sufficiently worked out. The addition of from 2\ % to 30% of formalin to 32 % to 
 
 5 % of bichromate, and a period of hardening for 18 hours or upwards, have all been recom- 
 mended. In fact, considerable latitude in both respects appears to be allowable. The results 
 are similar to those of the long Golgi method, which it also resembles in the fact that 
 pieces of impregnated tissue, or sections therefrom, can be left in strong alcohol for a con- 
 siderable period, days or even weeks, without deterioration. It is not improbable that this 
 will prove the best method, when properly developed, for the adult central nervous system of 
 mammals, but it is not as well suited for embryonic tissue as the rapid method of Golgi. 
 Its defect lies in its deficient hardening power. With adult tissue, where the osmic-bichromate 
 mixture penetrates very poorly, and also overhardens the periphery, this defect is an advantage. 
 Another advantage of this method is, that the favourable period for impregnation is not, 
 apparently, passed through so rapidly in the hardening. If brains are injected through the 
 carotid arteries with a solution strong in formalin (e.g. formalin i vol. + potassium bichromate 
 10 % sol. i vol.), the best results of both hardening and subsequent impregnation may be 
 obtained. 
 
 The following table will give further details as to how the majority of the preparations 
 here reproduced were made. By " used " is meant a solution which had been previously 
 used but remained in good condition. As a rule, it is better, however, to use a fresh solution. 
 By " osmic-bichromate " is meant a solution consisting of potassium bichromate 3^ % 4 vols. + osmic 
 acid i % i vol. 
 
 PLATE. HARDENING FLUID ; TIME OF HARDENING. SILVER SOLUTION. 
 
 III., V., & VI. Miiller i wk., osmic-bichromate several days Silver nitrate |% 
 
 VII., VIII., IX., ) (Zinc sulphate 2% i vol. ) 100 cc. -\- 2 drops 
 
 - Berkley's solution (used) 2 da. 20 hrs. 
 
 6 X. ) (Silver nitrate 2% i vol. ) formic acid. 
 
 1 The substitution of formalin for osmic acid in the hardening was first made by the writer during the summer of 1894, 
 and the fact announced in a note to his article "The Cranial Nerves of Amphibia," dated July, 1894, which appeared in the 
 "Journal of Morphology," Vol. X. No. I, January, 1895. A communication was read before the New York Academy of Sciences 
 Jan. 14, 1895, an abstract of which appeared in the " Anatomischer Anzeiger" for March 15, 1895 (Bd. X. No. 15). In the 
 number for June 13, Dung records similar observations, the March number of the "Anreiger" not having come into his hands 
 when his communication was sent off. In the number for July 19, Lachi calls attention to the fact that this mixture was 
 recommended by him in the "Monitore Zoologico," 1895, Anno. VI. No. i, and also by Isola in the "Bullettino della R. 
 Accademia Medica di Genoa," 1895, No. 2, and in the " Gazzetta degli Ospedali," February, 1895. 
 
PLATE. 
 
 XI. 
 
 THE HISTOLOGICAL TECHNIQUE 
 
 HARDENING FLUID ; TIME OF HARDENING. 
 Berkley (used) 2 hrs., osmic-bichromate 50 hrs. 
 
 II 
 
 XII. 
 
 XIII. 
 
 XV. & XVI. 
 
 XVII. & XVIII. 
 
 Osmic-bichromate + 2\% of formalin (unnecessary) 18 hrs. 
 
 Osmic-bichromate (used) 4 da. 15 hrs. 
 
 Lithium bichromate 2\<J 3 da. 3 hrs. 
 
 Potassium bichromate 5% '+ 5% of formalin, 22 hrs. 
 
 XX., XXII. to ) 
 
 > Osmic-bichromate 6-J- da. 
 XXXII., inc. \ 
 
 XXI. 
 
 XXXIII., XLII. to ) 
 
 Osmic-bichromate 2 da. 
 
 Osmic-bichromate 5 hrs., Berkley's solution 6 da. 
 XLIV., inc. > 
 
 XXXV. & XXXVI. Osmic-bichromate 5 hrs., Berkley's solution 6 da. 
 
 XXXVIII. 
 
 Osmic-bichromate 2 hrs., Berkley's solution 2\ da. 
 
 XXXIX., XLV. to) 
 
 \ Osmic-bichromate 7 da. 20 hrs. 
 XLVIII.,inc., L. \ 
 
 XLI. 
 
 LI. 
 
 Osmic-bichromate 2| da. 
 Osmic-bichromate 6 da. 4 hrs. 
 
 ,. , 
 
 i vol. ) 
 
 SILVER SOLUTION. 
 (Silver nitrate 2% i vol. 
 ( Sodium sulphate 2 % 
 
 Silver nitrate i <f c 
 
 Silver nitrate \% 
 
 Silver nitrate i % 
 
 Silver nitrate i % 
 
 Silver nitrate | % 
 Silver nitrate |% 
 
 Silver nitrate f % 
 
 (Silver nitrate 2% i vol. ) 200 cc. -f- 2 drops 
 ( Zinc sulphate 2 % i vol. ) formic acid. 
 
 Silver nitrate f % 
 
 Silver nitrate f % 
 { Silver nitrate 4% i vol. 
 ( Zinc sulphate 10% i vol. 1 
 
 ( Silver nitrate 2 % i vol. ) 200 cc. + 3 drops 
 
 x > 
 
 ( Zinc sulphate 2% i vol. ^ formic acid. 
 
 The Golgi method, like all special methods, has its peculiar defects, which are liable to 
 lead to misinterpretation. If, however, familiarity be acquired with the various appearances 
 it gives, the careful investigator will not be misled by irregular precipitations, etc., any more 
 than he is in other directions when using other special methods. Its great value in the 
 investigation of nervous histology cannot but be admitted by all. 
 
12 ATLAS OF NERVE CELLS 
 
 THE PHOTOGRAPHIC TECHNIQUE 
 
 BY EDWARD LEAMING, M.D., F.R.P.S. 
 
 INSTRUCTOR IN PHOTOGRAPHY, COLLEGE OF PHYSICIANS AND SURGEONS, COLUMBIA COLLEGE. 
 
 IN photomicrography it is essential that the optical parts of the apparatus used, termed 
 the system, should have their centres in the same straight line, and that this system, together 
 with the camera, should be so mounted as to be as rigid and free from vibrations or accidental 
 displacement as possible. These conditions were fulfilled by the apparatus used, which was 
 the large photomicrographic outfit of Carl Zeiss of Jena, in the laboratories of the College 
 of Physicians and Surgeons. 
 
 This consists of two heavy iron stands, upon one of which rests the camera, supported 
 by wheels, so that it can be rolled backward and forward without being displaced laterally ; 
 thus the operator is enabled to observe the object directly through the microscope without 
 altering the centring of any of the parts of the system. The second stand supports a broad 
 table, upon which is the microscope, inclined to the horizontal position, and an optical bench 
 of two parallel metal rails, upon which are fastened by set screws the accessory parts of the 
 system. These vary with the work, but are usually coloured light niters for altering the white 
 light to a monochromatic light, or screens for otherwise modifying it, diaphragms for centring 
 and for limiting the size of the light pencil, condensers, and reflecting mirror. The electric 
 arc used was so made as to fit the optical bench and be mechanically centred both laterally 
 and vertically. The weight of the two stands insured rigidity, and the mechanical adjustments 
 of the different parts were easily made, so that a beam of light from the electric arc or 
 from the heliostat would pass through the optical centres of the system of lenses and be 
 projected upon the centre of the ground glass screen of the camera. Where possible, sunlight 
 reflected from the speculum of an heliostat was the illuminant, and was used in preference 
 to other sources of illumination ; but, as in these latitudes during the winter season there 
 are many dark and cloudy days, it was found necessary to resort frequently to the electric 
 arc, which, as recently improved, gave results that left little to be desired. 
 
 It would seem at first, as the Golgi stain is a metallic impregnation and opaque, that 
 the use of isochromatic plates and a colour screen would have been unnecessary, but those 
 
THE PHOTOGRAPHIC TECHNIQUE I3 
 
 familiar with the Golgi preparations know that the tissue surrounding the silver impregnated 
 cells and processes frequently has a decidedly yellow colour, which more or less rapidly darkens, 
 and as it was desired to reproduce the effect of the cells and their processes upon a nearly 
 or quite colourless background, it was found advisable to use both colour-sensitive plates and 
 an orange-yellow screen. In previous work it had been the custom to use as a colour screen 
 a glass cell containing a solution of one of the Tropasolins. The use of such a cell is not, 
 however, unattended with difficulties ; it is somewhat troublesome to make up anew each time, 
 as so doing involves the use of the spectroscope if accuracy be desired ; it is subject to change 
 from dust, from evaporation, and in a short time from precipitation of the dye from its aqueous 
 solution. In lieu of the glass cell, therefore, an unexposed lantern slide plate was fixed in 
 sodium thiosulphite and well washed ; when dry it was placed upon a level surface, gelatine- 
 coated side upwards ; upon this was poured a strong solution of Tropaeolin in alcohol, which 
 was allowed to evaporate protected from dust ; Canada balsam was then poured over the dyed 
 plate and a cover glass applied. The plate thus prepared was allowed to harden under pressure 
 in a vertical position. A darker coloured screen was made by putting two dyed plates face 
 to face in the same manner. The screens so made have proved very convenient, show scarcely 
 perceptible fading or change during two years of constant use, and do away entirely with all 
 the trouble incident to glass troughs and fluids. 
 
 The optical apparatus consisted of a Zeiss' Abbe condenser specially made for photography 
 of i N. A. and achromatic, and of the following objectives : a Zeiss 2 mm., 3 mm., 4 mm., 
 8 mm., and 16 mm. apochromatic lenses, a 35 mm. projection lens of Zeiss, a ^-in. Bausch 
 and Lomb, a i-inch Ladd of London, and a 3-inch Ross of London. No eyepiece was 
 employed with the dry objectives, the camera length being 140 cm., which gave a magnifica- 
 tion of from 25 to 190 diameters; with the oil immersion apochromats 2 and 3 mm. a projection 
 ocular No. 4 was used and a camera length that gave from 1000 to 500 diameters by stage 
 micrometer measurement. It was found when the image was thrown upon the screen that no 
 matter how carefully the light was centred there would remain diffraction spectra, and the 
 phenomenon known as halation manifested itself in this peculiar way, the stain being an 
 impregnation of the chromate of silver and absolutely opaque, the light would be refracted 
 around the edges of the cell bodies and large processes, and would appear on what should 
 have been a perfectly black image as little tufts and jets of white (plates Nos. XIV. and 
 XXXIV. are examples of this being among the first taken). In order to obviate both these 
 difficulties a plate of finely ground glass was inserted, with its plane at right angles to the 
 optical axis of the system, between the condenser and source of light, and placed so as to be 
 within the focus of the condenser when both condenser and objective were focussed upon the 
 
14 ATLAS OF NERVE CELLS 
 
 
 
 object; this gave a softly, evenly illuminated field, particularly free from the defects mentioned 
 above, and with the sources of illumination used the sun and the electric arc the prolongation 
 of the exposure was not sufficient to cause inconvenience from casual vibrations. All development 
 was carried on with a single solution hydrokinon developer, composed of hydrokinon, 5 gr., 
 sulphite of soda 20 gr., carbonate of potash 15 gr., to the ounce of water. When fully developed 
 the plates were fixed in a solution of sodium thiosulphite, rinsed thoroughly and placed in a 
 saturated solution of alum for five or ten minutes, washed in running water from one to two 
 hours and allowed to dry spontaneously. The plates were not retouched in any manner, as it 
 was most desirable that the finished prints should be entirely photographic and should be as 
 little affected by the personal equation of the operator as possible ; it was for this reason also that 
 the artotype process of Mr. Bierstadt was chosen for the reproductions. This is the well-known 
 gelatine process which is used under various names in this country and in Europe, and may be 
 briefly described as follows : The negative is placed in contact with a glass plate coated with 
 gelatine containing- bichromate of potash (bichromated gelatine) and then exposed to light, as in 
 ordinary photographic printing. Gelatine is hygroscopic and unaffected by light, but bichromated 
 gelatine is rendered insoluble and incapable of absorbing moisture by the action of light in direct 
 proportion to the length of time the light action continues. The light passing through the 
 transparent portions of the negative renders the bichromated gelatine beneath non-absorbent, the 
 partially opaque parts of the negative allow a varying light action according to the degree of 
 density of the silver deposit, and the opaque parts prevent any action of light on the bichro- 
 mated film ; as a result, therefore, the negative is reproduced in all its details of light and 
 shadow by the varying power of absorption of the bichromated gelatine. The action of the 
 light having continued for the proper length of time, the plate is removed from beneath the 
 negative and placed in cold water, in the dark, until all the bichromate of potash has been 
 dissolved out ; it is then allowed to dry and is ready for the press. The plate is then firmly 
 fixed in a horizontal position in a common lithographic printing-press, with the gelatine coating 
 uppermost, and saturated with water, which is absorbed by the gelatine, as stated above, in a 
 degree varying inversely as has been the amount of action of the light ; the surplus water is 
 then removed and a roller charged with printer's ink is rolled over it in all directions. The 
 ink used is a greasy ink and will adhere to the non-absorbent portions of the plate, but will 
 be repelled by the moist portions either wholly or partially, according to the amount of moisture 
 they retain. When sufficiently inked, a sheet of paper is laid upon the plate and receives the 
 pressure, as in ordinary printing, the ink is transferred from the plate to the paper, and the 
 result is a finished print. 
 
HUMAN SPINAL CORD; SECOND DORSAL SEGMENT. WEIGERT STAIN. X 25 DIAM. 
 
 PLATE I. 
 
THE SPINAL CORD 
 
 THE SPINAL CORD 
 
 THE general appearance of a transverse section through the human adult spinal cord is 
 shown in Plate I. The level is at the upper part of the lumbar enlargement. This section 
 was stained by the Weigert haematoxylin method. The diagram (Fig. i) will assist in the 
 explanation of its chief features, and the details of structure are shown in Plates II. to XI. 
 
 The H-shaped arrangement of the gray matter of the cord, and the fact that the gray 
 matter is everywhere surrounded by white columns, which are darkly stained, is clearly marked. 
 It is evident that the anterior horns form the larger part of the gray matter ; that their 
 contour is irregular, owing to the massing of cells into groups, which cause a projection of 
 the gray into the white matter; and that they do not reach the surface of the cord. In the 
 diagram three cells (a) are shown representing three groups of anterior horn cells. These 
 cells are motor in function, presiding over the reflex movement and nutrition of the muscles, 
 each group representing a single muscle, or a group of muscles whose action is conjoined. 
 Many fibres are seen in the plate to issue from the anterior horns of the cord and to pass 
 through the anterior column to the surface. These are the anterior nerve roots. Their course 
 is shown in the diagram. In their course while in the gray matter they give off fine branches 
 which end about other cells in the vicinity. 
 
 The posterior horns of the cord are smaller than the anterior horns ; they nearly reach 
 the surface of the cord and present a club-shaped appearance at their extremities where the 
 substance of Rolando is situated. The cells of these horns are too small to appear in the 
 plate, but are shown in the diagram. It will be noticed that two of these cells (V) send their 
 neuraxons into adjacent columns of white matter, while a third (c) is confined to the posterior 
 horn itself. At the base of each posterior horn, on its median side, a group of cells is to be 
 seen in the plate ; this is the Clarke column of cells. It is indicated in the diagram by a 
 single cell (d) whose neuraxon is seen to pass outward to the periphery of the cord where 
 it turns upward in the direct cerebellar column. 
 
 In the gray matter of the cord, scattered irregularly through it, between and within the 
 anterior and posterior horns, there lie many cells whose existence is better shown in Plate II. 
 
16 
 
 ATLAS OF NERVE CELLS 
 
 These are indicated in the diagram. Their neuraxons pass out into the adjacent white columns 
 of the cord, where they bifurcate, one division turning upward, the other downward. These 
 neurons have as their function to associate the action of various segments or levels of the 
 cord. They are termed Strangzellen by German writers, who classify and name them according 
 to the column into which their neuraxon passes. They may be called the intrinsic cells of the 
 cord. Lenhossek 1 affirms that some of these cells appear to send out a neuraxon which 
 divides into two parts, each of which goes to a separate column. In the diagram it will be 
 seen that these intrinsic cells (e) send neuraxons into all the various columns. One set of 
 these cells should be especially noticed, which send their neuraxons upward in the periphery 
 
 VII 
 
 IX 
 
 VIII 
 
 FIG. i. 
 FlG. I. Diagram of a transverse section of the spinal cord. 
 
 On the right side the columns of the cord are shown and the fibres entering the gtay matter from these columns. /, anterior median column. 
 //, antero-lateral column. ///, lateral limiting layer. IV, ascending antero-lateral tract of Gowers. V, direct cerebellar column. VI, pyramidal 
 column. VII, Lissauer's column. VIII, column of Burdach. IX, column of Goll. 
 
 The posterior nerve roots are shown on the right side of the diagram and their various methods of termination in the gray matter. /, fibre 
 entering Lissauer's tract. 2, fibre entering posterior horn, j, fibre entering posterior horn and terminating deep within it. 4, fibre entering 
 column of Burdach. 5, fibre passing to root zone of column of Burdach and sending the collateral fibre to the anterior horn. 6, fibre entering 
 root zone and sending collateral to the Clarke column of cells. 7, fibre entering root zone and passing by way of the gray commissure to the 
 opposite side of the cord. S, fibre entering posterior horn through root zone having come up from a lower level. 
 
 On the left side the various cells of the gray matter are shown, a, motor cells with motor nerve roots passing out of the cord, b, intrinsic 
 cells of the posterior horns: one lies on the margin, is a "border cell" ; the other lies deep within the horn, they send neuraxons into the 
 posterior and lateral columns respectively, c, intrinsic cell of the posterior horn ; Golgi's second type, d, cell of the column of Clarke with its 
 neuraxon passing outward to the direct cerebellar column, t, intrinsic cells of the anterior and median gray with their neuraxons passing into the 
 antero-lateral column. /, intrinsic cell in median gray sending its neuraxon to the ascending tract of Gowers. g, commissural cells in the median 
 gray sending their neuraxons to the other side of the cord. A, sensory cell sending neuraxon to opposite column of Gowers. 
 
 1 Lenhossdk, Der feinere Bau des Nervensystems. 1895. 
 
THE SPINAL CORD I7 
 
 of the antero-lateral column in the division of it known as Cowers' tract (h). These are sup- 
 posed to be sensory in function and to transmit sensations of pain and temperature. There 
 are other cells in this median region of the gray matter which send their neuraxons across 
 the median line through one of the commissures to the opposite side of the cord (g). These 
 are known as commissural cells, and serve to transmit sensory impulses across the cord. Many 
 of the cells of this median gray matter are concerned in the reception and transmission of 
 sensations received through the posterior nerve roots. 
 
 The gray commissure uniting the two halves of the gray matter of the cord has in its 
 centre the central canal. This is seen in Plate II. to be lined with cylindrical epithelium. 
 The white commissure lies in front of the gray and contains many decussating fibres. 
 
 The entrance of the posterior nerve roots is well shown in Plate I. The nerve fibres 
 enter the cord at the apex of the posterior horn ; some pass directly into the horn, traversing 
 the substance of Rolando, but the majority enter the horn from its median surface, passing 
 through what is known as the root zone of the posterior column on the way. Some of these 
 fibres are seen to end in the gray matter near the column of Clarke. Others may be traced 
 toward the anterior horn. The actual direction and destination of these posterior roots are 
 shown in the figure (see also Fig. 2, p. 18, and Fig. 3, p. 24). The anterior fissure of the 
 cord containing blood-vessels, and the posterior septum, are also shown in the plate. 
 
 The white matter is made up of longitudinal fibres whose cut ends appear in the plate 
 as fine round white dots. In a normal cord thus stained there is no evident distinction 
 between the various columns of the white matter, though a posterior, lateral, and anterior 
 portion can easily be made out. The various columns as determined by pathological investigation 
 are as follows, and their situation is shown in Figs, i and 2. 
 
 I. Anterior median column contains fibres from the motor cortex of the hemisphere which 
 have come through the pyramid of the medulla and have descended without decussating. These 
 fibres pass into the anterior gray matter and terminate in a fine brush-like expansion about 
 the cells of the anterior horn. They transmit impulses of voluntary motion. 
 
 II. Antero-lateral column surrounding the anterior horn of the cord contains fibres passing 
 in both directions, which arise from cells lying in the anterior horn. These fibres pass from 
 the cells into the column, then bifurcate, turning both upward and downward, pass a long or 
 short distance through the column, giving off collateral branches in their course, then reenter 
 the anterior gray matter, and terminate, as do also the collaterals, in brush-like expansions 
 about the cells of the anterior horn. They transmit impulses of association which harmonize 
 the action of motor cells at various levels. This column is traversed by the anterior nerve 
 roots. 
 
i8 
 
 ATLAS OF NERVE CELLS 
 
 III. Lateral limiting layer, a layer of white fibres 
 adjacent to the median and posterior gray matter, con- 
 tains short fibres of association, whose origin is from 
 cells in the median portion of gray matter and whose 
 course is similar to those in II. Their termination 
 is about the cells of the same region at different levels. 
 This layer is traversed by many fibres from the other 
 lateral columns, entering and leaving the gray matter. 
 
 IV. Ascending antero-lateral tract, or column of 
 Gowers, lying on the surface of the cord, contains 
 fibres whose origin is in cells situated in the median 
 gray matter. These fibres traverse the other columns 
 and turn upward in this column, passing into the 
 antero-lateral field of the medulla and through the 
 formatio-reticularis of the pons into the internal cap- 
 sule. They transmit sensations of pain and tempera- 
 ture upward to the brain from the cord. Among 
 these fibres in this column are also descending fibres 
 which come downward probably from the cerebellar 
 hemisphere, but do not form a separate column. 
 
 V. Direct cerebellar column, lying on the surface 
 of the cord behind the last named column, contains 
 fibres which arise from the cells of the column of 
 Clarke, and after traversing the other lateral columns, 
 turn upward in this column, and pass to the cortex 
 
 FlG 2 of the cerebellar hemisphere. They transmit sensations 
 
 F.;. 2. - Diagram to illustrate the transmission of sensory, of equilibrium> ( See also Fig. 7 , p. 4 2.) 
 motor, anil association impulses through the spinal cord. 
 
 A, B, C, are three levels of the cord. /, II, ///, are sensory nerve roots entering the posterior root zone at level, B, bifurcating, ascending 
 and descending, and sending collaterals into the gray matter at many levels. Three such collaterals are shown at A, B, C, ending in the posterior 
 horn, and opposite median gray. The collaterals shown in diagrams I and 3 are not shown here. The collateral which crosses to the opposite side 
 terminates about the cell, f, which sends its neuraxon into the column of Gowers, G, and thence upward to the medulla. At level C a collateral 
 ends about sensory cell, A, which sends its neuraxon also into G. The collateral entering the base of the posterior horn terminates about the 
 column of Clarke on the same side. For the sake of distinctness the cell, if, of the column of Clarke is shown on the opposite side with its fibre 
 entering the direct cerebellar column, D, and ascending to the cerebellum, x, descending cerebellar tract sending its collateral fibres into the median 
 gray, e, intrinsic cell of the anterior horn at level, C, sending its neuraxon into the antero-lateral column, z, where it ascends, giving off col- 
 laterals, z, at levels B and A, and proceeding upward to the medulla. ', intrinsic cell at level, A, sending its neuraxon into the antero-lateral 
 column, y, where it ascends and descends, sending its collaterals, y, y, into the anterior horns of levels B, C. P, pyramidal tract giving off collaterals 
 which pass to the anterior horn of the cord and terminate in brushes about the motor cells. The motor cells, in, n, o, are shown for the sake of 
 distinctness on the opposite side, v, anterior median column with collateral entering the anterior horn and terminating in brushes around the 
 motor culls. 
 
THE SPINAL CORD Ig 
 
 VI. Lateral pyramidal tract contains fibres from the motor cortex of the opposite hemis- 
 phere which have come through the pyramid of the medulla and have decussated in the 
 medulla. These fibres descend and give off many collaterals, which, like their terminals, 
 turn inward and forward and terminate in a plexus of fine fibrils about the cells of the 
 anterior horn. They transmit voluntary motor impulses. 
 
 VII. Lissauer's column is a small collection of white fibres of short course derived from 
 the posterior nerve root, which lies on the apex of the posterior horn. These fibres bifur- 
 cate on entering the column and pass up and down for a short distance, finally entering the 
 posterior horn, to which they give also many collaterals. The fibres terminate in brush-like 
 expansions about the cells of the posterior horn. 
 
 VIII. The column of Burdach consists of three divisions: (i) the root zone near the pos- 
 terior horn, through which pass many of the posterior root fibres which bifurcate on entering the 
 zone and turn up and down, giving off numerous collaterals which enter the posterior horn 
 through the root zone ; (2) the median portion of the column which contains long fibres passing 
 from the posterior nerve root up to the cuneate nucleus of the medulla ; (3) a peripheral zone 
 containing other long fibres passing chiefly upward, but finally turning inward to the posterior horn. 
 Mingled among these three divisions are many association fibres whose origin, termination, and 
 collateral destination are in the posterior horns. These fibres are all sensory in their function. 
 
 IX. The column of Goll consists wholly of long fibres which have entered the cord in 
 the posterior nerve roots of the sacral, lumbar, and lower dorsal regions and pass upward to 
 the nucleus gracilis of the medulla. They transmit sensory impulses. 
 
 The constituents of the posterior nerve root require to be mentioned, a number of 
 different destinations being shown in the figures. These root fibres all bifurcate on entering 
 the cord, their divisions turning upward and downward. The downward division is short, and 
 soon turns into the gray matter, terminating in a brush-like expansion either in the posterior 
 or median gray matter about the intrinsic cells of the cord. The upward division is long, 
 and passes a varying distance toward the medulla, some fibres extending up the entire length 
 of the cord. Both divisions give off numerous collaterals which terminate like their true 
 ends. Some collaterals pass directly into the posterior gray horn and end there ; others pass 
 to the Clarke column of cells ; others pass by way of the posterior commissure to the other 
 half of the cord to terminate in the median gray matter; others pass directly through the 
 median gray to the anterior horns ending near the motor cells. These last are the paths 
 of reflex action. Impulses entering the cord thus are transmitted to the vicinity of the 
 intrinsic cells of the cord, which then take them up and send them on to the brain by 
 various tracts lying in the different columns already described. 
 
20 ATLAS OF NERVE CELLS 
 
 Plate II. shows a transverse section through a human foetal cord at the eighth month 
 stained by the Golgi method. The ratio of the white to the gray matter is very small at 
 this period of development. The cells of the gray matter are, however, already formed, and 
 their processes are well developed, but their separation into groups is not yet very manifest. 
 It will be noticed that the cells are irregularly scattered throughout the anterior horn and 
 the central gray matter, and extend back into the posterior horns to some extent. It will 
 be noticed that the cells are large in comparison with the size of the cord ; that they have 
 very numerous branches which pass in all directions, forming an interlacing mass of fibres 
 which touch but do not unite. In the commissures of the cord it will be noticed that 
 there are many of these fibres crossing from one side to the other. The substance of 
 Rolando is seen to be devoid of cells. The various cells of the cord may be classified as 
 follows : 
 
 I. Root cells, lying in the anterior horn or median portion of the gray matter, whose 
 neuraxons pass out of the cord chiefly through the anterior nerve roots, but also in part 
 through the posterior nerve roots. 
 
 II. Intrinsic or columnar cells, lying in all parts of the gray matter, whose neuraxons 
 do not leave the cord, but pass into the columns, bifurcate, and after traversing some extent 
 of the cord reenter the gray matter and end. Several varieties of these cells are described 
 according to the destination of their neuraxons. (a) Cells of a single column ; the neuraxon 
 passing into one column only. (b) Cells of two columns ; the neuraxon dividing, one branch 
 going to one column, the other to another. (c] Commissural cells ; the neuraxon crosses the 
 middle line and ends in the gray matter, or turns up or down in a column, finally entering 
 the gray matter again. (d) Cells which are both commissural and columnar; the neuraxon 
 divides and one part crosses while the other enters a column on the side of the cell. 
 
 III. Cells of Golgi's second type; 1 the neuraxon divides and subdivides in the gray 
 matter and does not leave it to enter the roots or columns. 
 
 1 C. Golgi, Sulla fina anatomia degli organi central! del sistema nervoso. Milano, 1885. Untersuchungen iiber d. feinere Bau 
 d. Nervensystems. Atlas u. Text. Jena, 1894. 
 
* 
 
 * 
 
 . ; < - 
 
 . 
 
 HUMAN SPINAL CORD OF EMBRYO EIGHT MONTHS. GOLGI STAIN, SHOWING CELLS WITH 
 BRANCHES. X 45 DIAM. 
 
 PLATE 
 
OF THE 
 
 UNIVERSITY 
 
 OF 
 
LARGE CELL OF THE ANTERIOR HORN OF THE SPINAL CORD OF A FOETAL PIG, SHOWING 
 NUMEROUS DENDRITES, AND THE NEURAXON ISSUING FROM THE LOWER END OF THE CELL, AND 
 GIVING OFF A COLLATERAL AT A RIGHT ANGLE IN ITS COURSE. X 190 DIAMETERS. 
 
 PLATE 
 
THE SPINAL CORD 21 
 
 Plates III. to VIII. show the large cells of the gray matter of the spinal cord both in 
 transverse and in longitudinal section. These cells are of various shapes with numerous 
 branches. They measure from 70 to 130/^1. There are long narrow cells whose branches 
 have two principal directions, as shown in Plate III. There are oval or polygonal cells with 
 branches running in many directions, as shown in Plates IV. to VIII. In all these cells it 
 is possible to distinguish a large body, which by other methods of staining can be shown to 
 contain a nucleus and nucleolus, and two varieties of branches characteristic of all nerve 
 cells ; viz. : dendrites and neuraxons. 
 
 THE DENDRITES are branches which come off from projections of the body of the cell. 
 They are thick near the cell, but diminish in size as they pass outward. They present 
 numerous offshoots in their course, branching like the limbs of a tree. The size of the dendrite 
 varies in its course, and it may present a varicose appearance. The surface of the dendrite is 
 not smooth, but presents a fine moss-like or furry appearance. The dendrites and their branches 
 terminate in free ends which usually have a little knob-like swelling upon them. Some of the 
 dendrites are very long. Others are quite short and terminate near to the cell. The dendrites 
 are made up of protoplasm similar to the cell body, and hence were formerly called protoplasmic 
 processes. It has been thought by almost all authorities ( Ramon y Cajal, 1 Van Gehuchten, 2 
 Lenhossek) that they collect the impulses going to the cell, from the fine neuraxons of other 
 cells, whose terminals reach the gray matter, and hence are cellulipetal in function. This is the 
 view generally accepted. It has been thought by Golgi that they have a purely nutritive function 
 and play no part in the action of the cell, being in relation with the blood-vessels at their 
 extremities and thus collecting nutrient substances for the cell, just as the roots of a tree 
 nourish its trunk. This view has, however, not met with any general acceptance, for Lenhossek 
 has shown that the dendrites develop long before the blood-vessels, and do not depend upon 
 these latter for their direction or distribution. It was formerly believed that after many branch- 
 ings the dendrites of one cell united with those of other cells, forming thus a network of fine 
 fibres throughout the gray matter (Gerlach). Golgi refuted this theory, and it is now 
 known that no such network exists ; and that while the dendrites may interlace and probably 
 touch one another, as do the adjacent leaves and branches of the trees in a forest, they never 
 really join one another, and are thus as independent of each other as are the trees. Thus 
 each nerve cell with its dendrites must be considered as a separate entity. 
 
 The cells, however, are in a close functional relation with one another though not joined 
 
 !S. R. Cayal, Les Nouvelles Ide"es sur la Structure du Systeme Nerveux. Paris, 1894. Also Internationale Monatschrift 
 f. Anat. u. Phys. Bd. VI., VII., VIII., and La Cellule, Tome VII. and IX. 
 2 Van Gehuchten, Le Systtme Nerveux de 1'Homme. Lierre, 1893. 
 
22 ATLAS OF NERVE CELLS 
 
 together, and it is supposed that an activity in one may excite activity in others, through the 
 medium of neuraxons and dendrites either by accidental contact, or, more probably, by the 
 induction of impulses in the one from the other; just as an electric current in one wire will 
 set up a current in independent wires about it. 
 
 Plate IV. shows a large motor cell of the anterior horn of the spinal cord of a foetal 
 pig, with numerous dendrites. The interlacing of the dendrites is particularly evident in the 
 plate, where many dendrites belonging to cells not shown in the plate intertwine with those 
 whose origin is shown. Gerlach believed that the sensory nerve fibres after entering the cord 
 became continuous with the ends of the dendrites, and that thus sensory impulses are sent 
 into nerve cells. But this theory has been abandoned, and the free termination of all den- 
 drites must be accepted. There is no continuity between nerve fibres and dendrites. There 
 is no real network of nerve fibres in the nervous system. The sensory impulses come in, as 
 we shall soon see, by nerve fibres which terminate in free ends like brushes, and their brush- 
 like expansions may surround the nerve cells or may interlace with their dendrites, but do not 
 necessarily come into contact with them. In Plate V. just below the cell such a free ter- 
 mination is to be seen, there being little knobs on each terminal filament. 
 
 THE NEURAXON is the second variety of branching process of the nerve cell. It was 
 formerly known as the axis cylinder projection or the functional process of the cell. The 
 neuraxon may come off directly from the body of the cell or from a small pyramidal projec- 
 tion of the body as in Plate IV., or it may arise from one of the large dendrites near the 
 cell body. The characteristics of the neuraxon which serve to distinguish it from the den- 
 drites are its uniform size from its origin onward, the possession of a myelin sheath giving 
 it a smooth surface, and its lack of many branches. It was formerly supposed that the axis 
 cylinder process never branched. This, however, is not so. The neuraxon gives off branches 
 called collaterals, but these differ greatly from the branches of the dendrites. In the first 
 place, the branches of the neuraxon are usually given off at right angles to its direction; in 
 the second place, they are always very small and fine as compared with the neuraxon. There 
 is often a slight swelling of the neuraxon at the point of origin of a collateral, as may be 
 seen in Plates IV. and VI. In Plates III. to VIII. numerous neuraxons can be seen with fine 
 collaterals coming off from them. They are seen to interlace with the dendrites, but their 
 smaller size, regular smooth contour, and method of giving off branches, nearly at right angles, 
 serve to distinguish them from dendrites. Golgi considers that it is only through the neu- 
 raxon that the cell performs its function, but this, as already stated, is not accepted. Van 
 Gehuchten and others have agreed that all neuraxons are cellulifugal in function, conveying 
 impulses outward from the cell, but this too is not fully established. 
 
^^^^ 
 
 LARGE CELL OF THE SPINAL CORD OF A FOETAL PIG, SHOWING NUMEROUS DENDRITES WITH 
 BRANCHES, AND THE NEURAXON WHICH PROJECTS UPWARD. INTERLACING NEURAXONS OF OTHER 
 CELLS WITH COLLATERALS ARE SEEN BELOW THE CELL, AND MANY DENDRITIC BRANCHES OF OTHER 
 CELLS. X 190 DIAM. 
 
 PLATE IV. 
 
OF 
 
 Vc, f 
 
OF THE 
 
 UNIVERSITY ,, 
 
 OF >*> 
 

 LARGE CELL OF THE ANTERIOR HORN OF THE SPINAL CORD OF A CHICK EMBRYO, SHOWING NU- 
 MEROUS DENDRITES WITH SUB-DIVISIONS A FINE TERMINAL BRUSH OR ARBORIZATION, THE END OF A 
 COLLATERAL IS SHOWN JUST BELOW THE LOWER OF THE TWO DENDRITES PASSING TO THE LEFT. 
 X 190 DIAM. 
 
 PLATE V. 
 
THE SPINAL CORD 2 3 
 
 There are two varieties of neuraxons which were first distinguished by Golgi. These 
 differ in their course and destination. (i) From some cells the neuraxon proceeds in a long 
 course, giving off collaterals, but preserving its identity till it reaches its destination. It may 
 end in another distant part of the nervous system; or in a muscle, which it reaches by way 
 of the motor nerve trunk; or in the skin which it reaches by way of a sensory nerve; 
 or in some organ of the body. When a cell has such a long neuraxon, it is brought into 
 relation, by means of it, with some distant and different part of the body. It is such a 
 neuraxon which proceeds from the cells of the anterior horns of the cord, enters the nerve 
 root, and goes to the muscles, forming a motor nerve fibre. It is such a neuraxon which 
 proceeds from the Clarke column of cells in the base of the posterior horn, and ascending 
 the cord, in the direct cerebellar column, ends in the cerebellum. Golgi classed all cells 
 with such long neuraxons together as his "first type" of cells. He affirms that all such 
 cells are of motor function, but this hypothesis is no longer accepted. This type of cells is 
 easily recognized. It is shown, e.g., in Plates IV. and V. in the spinal cord and in Plates 
 XL II. and XLIII. in the cerebral cortex. 
 
 (2) There is a second kind of neuraxon which has a short course and divides and sub- 
 divides soon after leaving the cell, thus forming a plexus of fine fibres in the gray matter, 
 and never proceeding in any particular direction to a great distance. Golgi classes cells pos- 
 sessing this form of neuraxon together as a " second type" An example of this type is shown 
 in Plates XLV. and XLVI. He affirms that they are sensory cells; that the neuraxons and 
 their branches from such cells unite with one another, forming a fine network or plexus of 
 fibres, and he states that the sensory nerve fibres arise from or terminate in this network. It 
 has been shown, however, by Cajal that no such true network exists. It is known, further, that 
 cells with such neuraxons are to be found in all parts of the nervous system, not particularly in 
 the posterior horns of the cord ; hence the sensory character of this type of cell is now denied. 
 No essential functional distinction between cells of the first and second type can as yet be admitted. 
 Cajal has shown that the sensory fibres do not join any such network as Golgi describes, but 
 end in free extremities, and he considers these cells of Golgi's second type as having a sort of 
 associative function. There is but one neuraxon to each cell throughout the nervous system. 
 
 The usual termination of the neuraxon in the nervous system is by breaking up into sev- 
 eral fine filaments, forming a terminal brush, each filament ending freely with or without 
 a little terminal knob. Such a termination has already been shown in Plate V. The col- 
 laterals also have free ends with brush-like expansions. Thus the cell preserves its identity 
 in its neuraxon as in its dendrites. It has a wide possibility of distribution of its impulses 
 through the neuraxon, but it never comes into direct continuity with another cell. 
 
ATLAS OF NERVE CELLS 
 
 Plate VI. shows a group of neurons in the anterior horn of the spinal cord, the intricate 
 interlacing of both dendrites and neuraxons being clearly seen. 
 
 It has already been stated that the function of these neurons is motor, each cell group 
 
 controlling a single muscle. Muscles act in response 
 to two forms of nervous stimulation, viz., reflex and 
 voluntary, and the mechanism, of these different 
 forms of movement requires separate consideration. 
 Reflex motion is easily understood by the aid 
 of the diagram, Fig. 3. In this figure three levels 
 of the spinal cord are shown, A, B, C, each con- 
 taining two motor cells m, n, o, in the anterior horns, 
 and from these motor cells the motor nerve roots 
 issue. Sensory nerve roots I., II., III., IV., are 
 seen to enter the posterior columns and to bifur- 
 cate, turning both up and down. The downward 
 branch of I. ends in B; the downward branches of 
 II. and III. end in C. All send their upward 
 branches to higher levels than those shown. All 
 give off collaterals at all levels, which pass into the 
 gray matter. Some of these collaterals pass forward 
 to the cell of the anterior horn of the same side 
 and terminate in brush-like expansions about the 
 motor cell. A few collaterals pass to the median 
 gray matter and terminate in brush-like expansions 
 about the commissural cell (c}, whose neuraxon 
 passes to and terminates about the motor cell of 
 the opposite side. A sensory impulse entering the 
 cord at the level B, through the posterior nerve 
 roots II. or III., is sent (i) to the motor cell, , on 
 the same side, and (2) if sufficiently intense to 
 excite the intermediate cell c, to the motor cell of 
 
 A, B, and C, three levels of the cord. 7, 77, 777, and 
 
 IV, posterior nerve roots entering the cord, bifurcating, ascend- the Opposite side also. The Sensory impulse also 
 ing and descending, and sending their collaterals and terminations , , , , , /-* i . 
 
 passes down the cord to level C, where it may 
 
 into the gray matter. These terminals end in brushes around the 
 
 motor cells m, , o. Some terminals end in brushes around the excite (3) motor Cells, O and O, the latter (4) through 
 
 commissural cells, <-, which in turn communicate impulses to 
 
 the motor cells of the opposite side, m, n, o. the medium of commissural cell c. It also passes 
 
 FIG. 3. 
 FIG. 3. Diagram to illustrate the reflex action of the spinal cord. 
 
A GROUP OF MOTOR CELLS IN THE ANTERIOR HORN OF THE SPINAL CORD OF A FOETAL 
 PIG, SHOWING NUMEROUS COARSE AND FINE DENDRITES INTERLACING WITH ONE ANOTHER, AND 
 MANY FINE NEURAXONS GIVING OFF COLLATERALS. X 190 DIAM. 
 
 PLATE VI. 
 
OF 
 
 .. 
 
LARGE CELLS OF THE MEDIAN PORTION OF THE SPINAL CORD OF AN EIGHT MONTHS HUMAN 
 EMBRYO IN LONGITUDINAL SECTION. A PORTION OF A WHITE COLUMN IS SEEN TO THE RIGHT 
 OF THE CELLS WITH ITS LONGITUDINAL FIBRES AND COLLATERAL BRANCHES FROM THEM PASSING 
 INTO THE GRAY MATTER. X 190 DIAM 
 
 PLATE VII. 
 
THE SPINAL CORD 25 
 
 up the cord to level A, where it reaches and excites the (5) motor cells m and (6) m. It 
 also passes upward to the basal ganglia, giving rise there to automatic acts, like the cry 
 of pain (7); or. to the brain cortex (8); producing there a conscious perception of the 
 sensation. The latter are not represented in the diagram. 
 
 The greater the intensity of the sensation, the more marked are its effects in motion. A 
 slight sensation may cause a slight movement on the same side limited to the muscles moved 
 by the cells lying at the level of the cord which the sensory impulse reaches. A very severe 
 sensation may throw into activity the motor cells of both sides at all levels of the spinal cord 
 above and below the level at which it enters. The simple knee jerk is an illustration of the 
 former, the intense rigidity of all the muscles caused by the pain of a lumbago is an illustra- 
 tion of the latter. And all degrees of reflex action between these extremes are possible. 
 Any break in the course of these impulses by a defect in the reflex arc will cause a sus- 
 pension of reflex action. Hence neuritis by destroying the nerve fibres in the nerve trunk ; 
 posterior sclerosis by destroying the nerve fibres in the root zone of the column of Burdach ; 
 myelitis by destroying the nerve fibres in their passage through the gray matter; and anterior 
 poliomyelitis by destroying the motor cells in the anterior horn, cause a cessation of reflex 
 activity. 
 
 Plate VII. shows a longitudinal section through the spinal cord of a human embryo of 
 eight months, and demonstrates not only the appearance of the large cells of the median 
 region with their dendrites, but also the fine plexus of interlacing nerve fibres, dendrites, and 
 neuraxons in the gray matter, seen on the left; and the longitudinal fibres of a white column, 
 seen on the right. From the fibres of this white column, several collaterals can be seen to 
 come off at right angles and to pass into the gray matter. Around the middle cell in the 
 plate, a few fine terminal filaments of such collaterals can be distinguished with free ends. 
 
26 ATLAS OF NERVE CELLS 
 
 Plate VIII. shows one of the intrinsic cells lying in the median gray matter near to the 
 white column. The large size of its body, the great number of its dendrites, and their rich 
 branching are very evident. The neuraxon is not visible. To the right of the plate a longi- 
 tudinal section through the white column is seen. This column is made up of fibres having 
 the characteristic appearance of axis cylinders, as contrasted with dendrites. Fine collaterals 
 are seen to come off from some of these fibres, at right angles, and to enter the gray 
 matter. These form the means of communication between the fibres of the columns and 
 the cells of the gray matter. They terminate in free extremities. 
 
 To the left of the plate several cells are to be seen with fine radiating branches entirely 
 different from the nerve cells and branches. These are the neuroglia cells, which form the 
 supporting framework of the spinal cord. 
 
LARGE POLYGONAL CELL OF THE SPINAL CORD OF A HUMAN EMBRYO, EIGHT MONTHS, IN 
 LONGITUDINAL SECTION, SHOWING NUMEROUS BRANCHES. TO THE RIGHT OF THE PLATE ARE LON- 
 GITUDINAL FIBRES OF A WHITE COLUMN WITH NUMEROUS COLLATERALS GIVEN OFF AT RIGHT 
 ANGLES AND PASSING INTO THE GRAY MATTER. TO THE LEFT OF THE PLATE NUMEROUS NEUR- 
 OGLIA CELLS ARE SHOWN. X 190 DIAM. 
 
 PLATE VIII. 
 
* N 
 
 SECTION THROUGH THE POSTERIOR COLUMNS AND BASE OF THE POSTERIOR HORN OF THE 
 HUMAN SPINAL CORD SHOWING THE CELLS OF THE COLUMN OF CLARKE WITH NUMEROUS DEN- 
 DRITES AND AXIS CYLINDER PROCESS. IN THE CELL IN THE CENTRE OF THE PICTURE THE NEU- 
 RAXON IS SEEN TO PASS FORWARD AND THEN TURN OUTWARD IN ITS COURSE TOWARDS THE DIRECT 
 CEREBELLUM COLUMN X 190 DIAM. 
 
 PLATE IX. 
 
T-HE SPINAL CORD 
 
 27 
 
 While the cells of the anterior horns are collected into distinct groups, those of the 
 posterior horn are usually scattered irregularly through the gray matter. There is, however, 
 one distinct group of cells found at the base of the posterior horn on its median side at 
 certain levels of the cord, viz. the Clarke column of cells. This extends from the mid-lumbar 
 region upward to the lower cervical region. It has also been called the dorsal group of 
 Stilling. A few cells are grouped in the same location in the lower sacral region (Stilling's 
 sacral nucleus), and in the upper cervical region (Stilling's cervical nucleus). The cells of 
 this group are from 45 /i to 90 p in size. Under some stains they appear to be pear- 
 shaped, with but one projecting process. But the Golgi stain shows them to have both 
 dendrites and a neuraxon. 
 
 Plate IX. shows this group of cells, one of which is successfully stained. It is seen to 
 the left of the centre of the plate. It has four dendrites, and a long fine neuraxon which 
 passes forward, and then turns outward through the gray matter. The neuraxons of these 
 cells pass through the median gray matter and through the lateral column to its surface, 
 where they turn upward, and traversing the entire length of the cord, forming the direct 
 cerebellar column, turn outward in the inferior peduncle of the cerebellum, and finally reach 
 that organ (cell d and column V. in Fig. i, on page 16; also i in Fig. 7, page 42). The 
 probable function of these cells is to receive from the posterior root fibres those sensory im- 
 puls'es which are necessary to our appreciation of equilibrium, and without which the nice 
 adjustment of balance is impossible. These impulses are transmitted from these cells to the 
 cerebellum by the direct cerebellar column. 
 
28 ATLAS OF NERVE CELLS 
 
 The posterior horn of the spinal cord presents a peculiar structure differing from that 
 of the remainder of the gray matter. This is evident in Plates I. and II. and is more 
 clearly shown in Plate X. It has a peripheral portion and a central portion. The 
 peripheral portion is known as the substance of Rolando, and about its posterior part is a 
 zone of somewhat different appearance known as the spongy layer. The substance of 
 Rolando extends through the entire length of the cord and upward through the medulla 
 and pons, being there in close relation with the termination of the sensory branches of the 
 fifth nerve. It appears, therefore, to have a sensory function. This substance resists many 
 staining materials, and it is only within a short time that its actual structure has been 
 determined by Cajal and Lenhossek. It has a granular homogenous appearance, but they 
 have found that it contains many very small polygonal cells, with dendrites and a neuraxon, 
 the latter coming off from the dorsal side and passing into the lateral column of the cord, 
 where it divides and passes up and down as do the other longitudinal fibres of the cord. 
 About these cells there is a mass of interlacing fibres of most minute calibre, too fine to 
 be stained by ordinary methods. 
 
 In the spongy zone, outside the Rolandic substance, lie much larger cells, whose long 
 axis is parallel with the periphery of the cord, and whose dendrites and neuraxons enter the 
 posterior and lateral columns. 
 
 The central portion of the posterior horn is made up of a coarser mesh of fibres and 
 of numerous posterior nerve root fibres which traverse it to reach the central gray matter. 
 It contains many of the intrinsic cells already described. 
 
 Plate X. shows the entire posterior horn of the cord, including both the substance of 
 Rolando and the spongy zone. The large cell seen near the surface of the cord (at the 
 bottom of the plate), whose long axis is parallel with the surface of the cord, and which 
 has numerous processes passing both into the posterior and lateral columns, and directly into 
 the posterior horn, is the type of cell described by Cajal, and known as the border cell. 
 These cells are supposed to be association cells, and their neuraxon appears to pass uniformly 
 into the posterior columns. The general homogenous structure of the substance of Rolando 
 appears in the plate, but it is seen to be traversed by numerous fibres, which enter the 
 posterior horn from the posterior nerve root. These pass directly to the plexus of nerve 
 fibres which lies deep within the horn near to its base. 
 
THE POSTERIOR HORN OF THE SPINAL CORD OF A HUMAN EMBRYO, EIGHT MONTHS, SHOW- 
 ING THE SUBSTANCE OF ROLANDO WITH FIBRES PASSING THROUGH IT, THE SPONGY ZONE AND 
 A LARGE BORDER CELL ON THE PERIPHERY. X 190 DIAM. 
 
 PLATE X. 
 
OF THE 
 
 UNIVERSITY ] 
 
 or 
 
SECTION THROUGH A POSTERIOR SPINAL GANGLION AND SPINAL CORD OF A CHICK EMBRYO, 
 SEVENTH DAY, SHOWING CELLS IN THE POSTERIOR SPINAL GANGLION WITH TWO PROCESSES, ONE 
 PASSING OUTWARD IN THE SENSORY NERVE, THE OTHER PASSING INWARD AND ENTERING THE SPINAL 
 CORD X 190 DIAM. 
 
 PLATE XI. 
 
THE SPINAL CORD 
 
 The spinal cord is made up not only of the neurons thus far described, whose cells lie in 
 it, and of the neuraxons which come down from cells in the brain, but also of a large number 
 of neuraxons which enter it from the posterior spinal ganglia. The researches of His ' have 
 shown that in these ganglia there develop cells which send out two processes. Plate XI. shows 
 such a ganglion from a chick embryo, containing a number of cells, each with two processes. One 
 process proceeds outward in the posterior spinal nerve, and passes to the skin or to some 
 internal organ. The other process proceeds inward in the posterior spinal nerve root and 
 enters the spinal cord. The latter process is usually smaller in its calibre than the former 
 one, as proven by Retzius. 2 In a later stage of development the body of the cell appears 
 to be pressed to one side, so as to lie not directly in the line of the branches, and then 
 the two branches become fused together, so that in an adult the appearance of the cell is 
 changed, and it looks like a pear with the stem divided and passing in two opposite 
 directions. There has been some discussion as to which of the two branches of these 
 cells is to be called a dendrite and which a neuraxon. As a matter of fact, no distinctive 
 differences are to be seen in the branches, and it is better to regard this type of cell as 
 different from the spinal neurons, and to maintain that these cells have two neuraxons and 
 no dendrites, especially as both branches receive a medullary sheath on leaving the cell. 
 It is known that sensations are conveyed from the skin to the spinal cord along these 
 branches, and it is evident, therefore, that the distal branch conveys the impulse to the cell, 
 and the proximal branch conveys it from the cell. It is not impossible that the cell has 
 little to do with the function of sensation, and that its only function is to maintain the 
 nutrition of its processes. It is certainly true that if either process is separated from its 
 cell it atrophies in its entire length. 
 
 1 W. His, Arch. f. Anat. u. Phys. Anatom. Abtheil., 1887. 2 Retzius, Biol. Untersuchungen, Stockholm, I. 1890. 
 
30 ATLAS OF NERVE CELLS 
 
 The neuraxons which enter the cord from the spinal ganglia pass in partly at the apex 
 of the posterior horn, and partly through the column of Burdach in the portion adjacent to 
 the posterior horn, hence known as the root-zone of the posterior column. On entering the 
 cord each neuraxon divides into two parts, which turn at an angle of 150 to the original 
 direction of the neuraxon, making a Y-shaped division ; and these pass up and down the 
 cord. Plate XII. shows this peculiar division of the fibres, as seen in a longitudinal section 
 of the cord through the root-zone. The fibres which pass downward appear to be short, 
 rarely extending downward more than three centimetres. They pass together in a bundle 
 which lies in the antero-lateral part of the column of Burdach, near the posterior horn, and 
 has been termed the comma-shaped bundle, from the shape of its cross-section. Those which 
 pass upward are of indefinite length, many in fact extending all the way to the medulla. 
 Both divisions give off collaterals which enter the gray matter. In a long organ like the 
 cord, which receives 31 pairs of nerves, it is evident that the higher the level, the greater 
 the number of these fibres in a cross-section ; for in the cervical region are to be found 
 fibres from every level below it. It has been found that a definite order of arrangement of 
 these long ascending fibres takes place. As each successive nerve root comes into the root- 
 zone, from below upward, it enters near to the posterior horn, and displaces the nerve fibres, 
 already ascending, inward and backward toward the posterior septum. Hence in the cervical 
 cord the fibres from the cervical nerves lie next the posterior horn, those from the dorsal 
 cord are further inward, and those from the lumbar and sacral cord are crowded against the 
 posterior septum. 
 
 The arrangement of columns in the cord and their connections with the various cells are 
 shown in the diagram (Fig. 4). A, B, and C are three levels of the spinal cord. I., II., and 
 III. are three sensory nerve roots entering the cord at level B. They are seen to bifurcate, 
 to send their branches downwards and upwards to levels 4 and C, and to send collaterals 
 into the posterior horns at different levels, some of which terminate about the intrinsic cells 
 of the same side (/t) t and some of which cross in the gray commissure, to terminate about 
 the intrinsic cells of the opposite side (/). From these cells a neuraxon proceeds outward to 
 the antero-lateral ascending tract of Gowers (G), which passes upward to the brain. Cell d, 
 representing the Clarke column of cells, is also shown at the three levels, on the right side 
 only, and its neuraxon ascending the cord in the direct cerebellar column (D). Collaterals 
 from the posterior nerve roots terminate around these cells, as has been already shown in 
 Fig. i, page 16. These collaterals are shown on the left side only in this figure for the 
 sake of clearness. It is thus evident that impulses entering the spinal cord by the pos- 
 terior nerve roots are not only distributed to the gray matter of the cord itself, as shown 
 

 LONGITUDINAL SECTION THROUGH THE SPINAL CORD OF CHICK EMBRYO, SEVENTH DAY, SHOW- 
 ING THE ENTRANCE OF THE POSTERIOR NERVE ROOTS AND THEIR T SHAPED DIVISION INTO AS- 
 CENDING AND DESCENDING FIBRES. X 190 DIAM. 
 
 PLATE XII. 
 
THE SPINAL CORD 
 
 in Fig. 3, page 24, representing the reflex action 
 of the cord, but are also sent upward to the brain 
 by three separate tracts ; namely, the posterior col- 
 umns of the side on which they enter, the direct 
 cerebellar column, and the antero-lateral tract of 
 Cowers on the opposite side. 
 
 The diagram also exhibits the course of vol- 
 untary impulses from the brain downward to the 
 motor neurons of the cord. These impulses, as 
 already described, descend both in the pyramidal 
 tract (P) and in the anterior median column (v). 
 From the descending fibres of these columns fine 
 collaterals come off, which enter the anterior horns 
 of the cord, and terminate in fine terminal brushes 
 about the motor nerve cells. For the sake of clear- 
 ness, these cells (m, n, o) are shown on one side 
 only. These collaterals are seen to come off at all 
 levels of the cord. The voluntary impulses descend- 
 ing in these columns and thus reaching the anterior 
 motor cells excite them to activity, and the result 
 is a voluntary contraction of the muscle to which 
 their neuraxons pass. 
 
 The diagram also shows the existence of in- 
 trinsic cells in the cord (/, e\ with their neuraxons 
 passing in the antero-lateral column (y, z), and 
 ascending and descending, giving off collaterals which F I G . 4 . 
 
 , , i r . i , , FIG. d. Diagram to illustrate the transmission of sensory, 
 
 pass into the anterior gray matter of other levels 
 
 motor, and association impulses through the spinal cord. 
 
 A, B, C, are three levels of the cord. /, //, ///, are sensory nerve roots entering the posterior root zone at level B, bifurcating, ascending 
 and descending, and sending collaterals into the gray matter at many levels. Three such collaterals are shown at A, B, C, ending in the posterior 
 horn, and opposite median gray. The collaterals shown in diagrams I and 3 are not shown here. The collateral which crosses to the opposite side 
 terminates about the cell /, which sends its neuraxon into the column of Cowers, G, and thence upward to the medulla. At level C a collateral 
 ends about sensory cell A, which sends its neuraxon also into G. The collateral entering the base of the posterior horn terminates about the 
 column of Clarke on the same side. For the sake of distinctness the cell, </, of the column of Clarke is shown on the opposite side with its fibre 
 entering the direct cerebellar column, D, and ascending to the cerebellum, jr, descending cerebellar tract sending its collateral fibres into the median 
 gray, e, intrinsic cell of the anterior horn at level C, sending its neuraxon into the antero-lateral column, z, where it ascends, giving off col- 
 laterals, 2, at levels B and A, and proceeding upward to the medulla. i, intrinsic cell at level A, sending its neuraxon into the antero-lateral 
 column, y, where it ascends and descends, sending its collaterals, y, y, into the anterior horns of levels B, C. P, pyramidal tract giving off collaterals 
 which pass to the anterior horn of the cord and terminate in brushes about the motor cells. The motor cells, /, , o, are shown for the sake of 
 distinctness on the opposite side, v, anterior median column with collateral entering the anterior horn and terminating in brushes around the 
 motor cells. 
 
32 ATLAS OF NERVE CELLS 
 
 than the one from which they start, i is such a cell, on the right side of the cord at 
 level A, with its neuraxon y in the right antero-lateral column shown as passing downward ; 
 and e is such a cell, at level C in the left anterior horn, with its neuraxon passing upward. 
 At x in the drawing, the descending fibres of the antero-lateral tract, supposed to come from 
 the cerebellum, are seen with collaterals entering and terminating in the median gray portion 
 of the cord. 
 
 \ 
 
 MEDULLA OBLONGATA 
 
 The cells of the medulla do not differ essentially from those in the spinal cord, except- 
 ing in their relative position to the tracts of white matter. The numerous motor cells giving 
 an origin to the motor cranial nerves, are quite similar to the motor cells of the anterior 
 horns of the spinal cord. Throughout the medulla there lie numerous intrinsic cells which 
 are undoubtedly association cells in function, and combine the various complex acts, which are 
 automatically performed by the medulla oblongata. These resemble the intrinsic cells of the 
 cord. Small sensory cells are also found in the substantia gelatinosa of the fifth nerve 
 nucleus of the medulla, similar to those found in the posterior horns of the spinal cord. 
 
 All these cells have numerous dendritic branches, and an axis cylinder, which passes 
 upward and downward to the brain or spinal cord, giving off in its course numerous collat- 
 erals. The arrangement of these fibres in the medulla is extremely intricate, and does not 
 concern us here. They are clearly traced by Bechterew in his " Leitungsbahnen im Gehirn 
 und Ruckenmark," and by Van Gehuchten in his "Systeme Nerveux de l'Homme." The histo- 
 logical structure which is especially peculiar to the medulla is the olivary body. This is made 
 up of a thin layer of nerve cells, which layer is highly convoluted, and which is disposed 
 upon the convex surface of the body and around a central mass of fibres, which appear to 
 issue from the olivary body, on one side only. The structure of the olivary body is identical 
 with that of the corpus dentatum of the cerebellum, shown in Plate XX. The olives are 
 connected by means of fibres with the cerebellum, and seem to be in intimate relation with 
 it, as shown in Figure 7, page 42. 
 
. 
 
 f^ OF THE 
 
 ; UNIVERSITY.!! 
 
 OF 
 
f 
 
 ' 
 
 * 
 
 NEUROGLIA CELLS IN A LONGITUDINAL SECTION OF THE SPINAL CORD OF A NEW BORN 
 CHILD. X 190 DIAM. 
 
 PLATE XIII. 
 
THE SPINAL CORD 
 
 33 
 
 Plate XIII. shows a number of cells of the neuroglia in the cord. The neuroglia is 
 admitted by all authorities to be a sort of supporting framework extending through the entire 
 nervous system, giving support to the neurons. It was formerly supposed that the neuroglia 
 was a sort of connective tissue, but the researches of Golgi and His * have shown that it 
 originates in the embryo from the ectoderm, the same layer which develops into the nervous 
 system. The neuroglia develops from the ependymal cells lining the central canal, which send 
 out long fibre-like offshoots extending to the periphery of the cord, and forming a sort of scaf- 
 fold (Fig. 5). These are shown 
 on the left side of the figure. 
 They do not enter the posterior 
 horns or columns. In addition 
 to this scaffold there develop in 
 the cord spider-like cells with long 
 projections, and these are shown 
 in Plate XIII. The enormous 
 number of these cells, and their 
 grouping in the fcetal cord is 
 shown in Figure 5, taken from 
 Lenhossek. He has shown 2 that 
 these cells develop from the epi- 
 thelial layer of the embryo, like 
 the nerve cells themselves. A 
 large number of these glia cells 
 cluster about the central canal of 
 
 the cord, forming a dense interlac- 
 
 FIG. 5. 
 
 FlG. 5. (Lenhossek.) Neuroglia cells and fibres of the spinal cord of a human embryo 
 
 14 cm. long. 
 
 On the left half the long, straight fibres developing from the epithelial cells lining 
 the central canal are shown. On the right half the large number of neuroglia cells devel- 
 network of 2"Ha fibres known P' n 8 within the cord and sending their fibres through it are seen. 
 
 as the gelatinous substance of Rolando. No special function can be assigned to this region. 
 These glia cells appear to play a considerable part in the pathological processes which go on 
 in the cord. Their abnormal increase gives rise to gliosis in the gray matter, and to sclerosis 
 in the white matter. 
 
 1 His, Arch, fur Anat. u. Phys., Anat. Abth., 1890, s. 103. 
 
 2 M. v. Lenhossdk, Verhandl. d. Anat. Gesellsch., 5 Versammlung, 1891, Anat. Anzeiger, 1891, s. 193. 
 
34 ATLAS OF NERVE CELLS 
 
 THE CEREBELLAR CORTEX 
 
 THE cortex of the cerebellum is, by reason of its numerous fine convolutions, almost as 
 extensive in area as that of the cerebrum. It is made up of two distinct layers of gray 
 matter, between which lie the large Purkinje cells, which form the chief characteristic of this 
 organ. Everywhere beneath these layers of gray matter the nerve fibres collect into white 
 masses, which pass in various directions, bringing the cerebellum into relation with other parts 
 of the nervous system. 
 
 Plate XIV. shows the existence of the two cortical layers, and the appearance of the 
 Purkinje cells (P) l in the adult human brain. 
 
 The first, or superficial, layer of the cortex has been named the molecular layer. 
 
 The second, or deep, layer of the cortex is known as the granular layer. 
 
 The Purkinje cell body lies between these layers, and its dendrites pass into the super- 
 ficial layer, while its neuraxon passes into the deep layer. 
 
 The body of the Purkinje cell " is round or oval, as shown in Plates XIV. and XV. 
 There are several large protoplasmic extensions of the body which give off numerous branches 
 soon after leaving the body, and these branches present the extraordinarily rich sub-branching 
 which is peculiar to the Purkinje cell. It will be noticed that all these dendritic branches 
 have a rough surface. This is owing to the presence of fine thorn-like excrescences close 
 together on the branch. A high power of magnification shows that these excrescences have 
 clubbed extremities, not being pointed as thorns are. The excrescences have been called 
 gemmules. This extraordinary mass of branches of the Purkinje cell has been termed its 
 arborization. But their arrangement is not exactly like that of a tree, for it is found that 
 these branches all lie in one plane, a plane transverse to the folding of the convolution. 
 Hence if the convolution is cut in a longitudinal plane, the cells and branches are seen only 
 as a thin vertical line. The natural position of the branches may then be compared to the 
 artificial position given to the branches of a plant which has been pressed in a herbarium. 
 As the vast majority of the fine nerve fibres of the molecular layer of the cerebellum have 
 a horizontal and longitudinal direction, it is evident that they can pass through and between 
 
 1 These letters refer to Figure 6. on page 39. 
 
PURKINJE CELL IN ADULT HUMAN CEREBELLUM, SHOWING NUMEROUS BRANCHES IN THE MO- 
 LECULAR LAYER AND ALSO A LONG AXIS CYLINDER PROCESS WITH TWO COLLATERALS. THIS 
 PROCESS PASSES THROUGH THE GRANULAR LAYER ON ITS WAY TO THE WHITE MATTER. BASKET 
 FIBRES ARE SEEN TO THE RIGHT X 190 DIAM. 
 
 PLATE XIV. 
 
Of THE 
 
 UNIVERSITY 
 
 OF 
 

 PURKINJE CELL IN ADULT HUMAN CEREBELLUM. SHOWING NUMEROUS BRANCHES IN THE MOLEC- 
 ULAR LAYER AND THE BEGINING OF AN AXIS CYLINDER PROCESS PASSING INWARD INTO THE GRANULAR 
 LAYER. X 190 DIAW. 
 
 PLATE XV. 
 
THE CEREBELLAR CORTEX 35 
 
 the branches of the Purkinje cells, without becoming tortuous. All the dendrites of the 
 Purkinje cells end in free terminations, as is well shown in the plates. The majority of the 
 branches extend quite to the superficial margin of the convolution. 
 
 The neuraxon of the Purkinje cell, well shown in the plate, enters the granular layer 
 from the side or base of the cell, and passes through it to enter the white matter. In its 
 passage it gives off several collaterals, which turn backward and ascend into the molecular 
 layer, where they become longitudinal in their direction, branch, and interlace with the 
 dendritic branches of the cell. Cajal supposes that by means of these collaterals the inter- 
 action of numerous cells is secured. The neuraxons of the Purkinje cells form the chief 
 constituents of the white tracts of the cerebellum which issue from the cortex. They pass in 
 various directions which will be described later. 
 
 To the right of the plate a horizontal fibre is to be seen in the molecular layer, giving 
 off a branch which in turn divides into several branches. This is one of the so-called 
 baskets of the cerebellum a brush-like or basket-like mass of fibres which surrounds the 
 body of the Purkinje cell. These are also shown in Plate XVIII. 
 
 Plate XV. shows another Purkinje cell of the adult human brain with its arborization. 
 The gemmules upon the dendrites are well shown in this plate. The size of the cell is 
 70 p.. This plate displays the wonderful manner in which the dendrites divide and subdivide 
 until the final terminal filaments are reached. The body of this cell is distinctly pear- 
 shaped, and the neuraxon is clearly seen coming out of the base on one side, but can only 
 be traced a short distance. It has recently been affirmed by Semi Meyer ' that the gemmules 
 are artifacts of the Golgi method, he having failed to demonstrate their existence by the methylin 
 blue method of staining. This assertion appears to be inconsistent with the uniformity of their 
 appearance in all specimens and with the regularity of their arrangement, both on the dendrites 
 of the Purkinje cells and on the dendrites of the pyramidal cells of the cerebral cortex. They 
 must be considered as actual structures. 
 
 1 Arch, fur Micro. Anat., Oct. 1895. 
 
36 ATLAS OF NERVE CELLS 
 
 Plate XVI. shows another Purkinje cell with its branches. The tree-like appearance of 
 the branches is well marked in this plate, and the thorn-like appearance of the gemmules 
 which cover the branches. It will be noticed that the arrangement of the branches conforms 
 to the situation of this cell at the bottom of a sulcus, and that the dendrites are strictly 
 limited to the molecular layer, and do not enter the granular layer at any point. It is only 
 by the Golgi method of staining that the remarkable number and complexity of the divisions 
 of these dendrites has been made known. 
 
 The function of the Purkinje cell is undoubtedly to preside over the equilibrium of the 
 body. When it is considered that the act of balancing requires a nice adjustment of 
 muscles throughout the entire body, that every movement even of eyes and head implies a 
 shifting of the centre of gravity, which must be compensated for when one is standing, by 
 the greater or less contraction on one side or the other of the muscles of the trunk and 
 limbs, and when, moreover, the necessity of constant variations .of this adjustment in the 
 motions of the arms, in ordinary gesticulations, in walking, running, dancing, etc., which 
 must be controlled by a competent unconscious mechanism is appreciated, it becomes a 
 matter of less astonishment that the function of these cells is one of the most important in 
 the nervous system, and that their complexity of structure is so great and their connections 
 with other parts are so numerous. 
 
' 
 
 PURKINJE CELL OF ADULT HUMAN CEREBELLUM, SHOWING LARGE NUMBER OF BRANCHES WITH 
 SUBDIVISIONS WHICH OCCUPY THE MOLECULAR LAYER OF THE CORTEX. THE PECULIAR THORN LIKE 
 EXCRESCENCES ON THESE BRANCHES ARE WELL SHOWN. X 190 DIAM. 
 
 PLATE XVI. 
 

 
MOLECULAR LAYER OF THE CEREBELLUM SHOWING STELLATE CELLS. X 190 OIAM. 
 
 PLATE XVII. 
 
THE CEREBELLAR CORTEX 37 
 
 THE MOLECULAR LAYER OF THE CEREBELLUM 
 
 The cells of the molecular layer of the cerebellum are of two varieties. First, the small 
 stellate cells (S)\ secondly, the cells with fibres which enter into the basket-like formation 
 about the Purkinje cell (B) (Korbzellen). 
 
 Plate XVII. shows the small stellate cells of the molecular layer. They have irregular 
 polygonal bodies with three, four, or more dendrites, and an axis cylinder. The size of these 
 cells is only 10 to 15/u,. The long axis of the body is horizontal in the deeper portion of 
 the molecular layer, but may be directed in any direction in the more superficial part. The 
 dendrites are comparatively short. They divide and subdivide, the majority taking a horizontal 
 direction. The neuraxon is long and gives off numerous collaterals. It appears to pass ver- 
 tically in the molecular layer only, not entering the granular layer. Its final destination is 
 uncertain. It is impossible to distinguish between the dendrites and the neuraxon in the plate, 
 but a higher magnifying power shows that the longest of the fibres is the neuraxon. Some 
 of the collaterals of the neuraxon can be traced. They usually take a horizontal course after 
 leaving the neuraxon. There are numerous horizontal and vertical fibres shown whose cells 
 are not visible. These together make up a complex interlacing mass of fibres in this molec- 
 ular layer. 
 
 The second variety of cells in the molecular layer is the cell with basket expansion (It). 
 These lie chiefly in the deeper part of the layer. They are small cells, polygonal in shape, 
 which give off numerous dendrites and a long neuraxon. This differs from the neuraxon 
 of other cells in being thick and irregular in size. From it there arise numerous collaterals, 
 which take a downward course and end in brush-like expansions, which appear to surround 
 the body of a Purkinje cell. The name " basket cell " is given to these cells because the 
 ends of their collaterals appear to enclose the body of the Purkinje cell as in a basket. It 
 would be erroneous to suppose that the fibres intertwine as do the straws of a basket. The 
 size of these cells is about 20 /u,. A doubt has arisen that the basket fibres are true neu- 
 raxons; in fact, some authorities have supposed the basket to be a glia structure. Cajal, Golgi, 
 Van Gehuchten, and Kolliker, 1 however, agree with the view that the baskets are made up 
 of fine nerve fibres. 
 
 1 Kolliker, Handbuch d. Gewebelehre des Menschen, 6te Aufl., 1893. II. Bd., iste Halfte. 
 
38 ATLAS OF NERVE CELLS 
 
 Plate XVIII. shows the molecular layer of the cortex, its intricate plexus of nerve fibres, 
 and a few of the small polygonal cells which give rise to basket fibres. Cajal and Kolliker do 
 not distinguish these cells sharply from the small polygonal cells, previously described as the 
 first variety of cell. Cajal calls them all stellate cells. It is evident, by comparing Plates 
 XVII. and XVIII., that in appearance there is little difference between the first and second 
 variety of cell. The chief difference lies in their size and in the termination of the fibres. 
 It is extremely difficult in any one preparation to show the basket-like expansion (<$) about the 
 Purkinje cell, especially in the adult human brain. In this plate, however, one such expan- 
 sion is well shown, and two others are really present, though the undivided fibres are not easily 
 distinguished in this focus. The fibres become enlarged, diverge, and then again converge, 
 forming almost a spherical basket. These baskets lie on the margin between the molecular 
 and granular layers, hence appear, to project downward in the plate into the granular layer. 
 It can readily be imagined that these fibres surround the body of the Purkinje cell. 
 
m 1 1 
 
 MOLECULAR LAYER OF THE CEREBELLUM SHOWING STELLATE CELLS WITH BASKET FIBRES, 
 AND FINE INTERLACING FIBRES. X 190 DIAM. 
 
 PLATE XVIII. 
 
uTiA*3 
 
 OF THE 
 
 UNIVERSITY, 
 
 OF 
 
SMALL GRANULE CELLS IN THE GRANULAR LAYER OF THE CEREBELLUM WITH SHORT DEN- 
 DRITES AND WITH A FINE AXIS CYLINDER PASSING INTO THE MOLECULAR LAYER. THE LONGITUD- 
 INAL FIBRES OF THE MOLECULAR LAYER ARE SHOWN IN THE UPPER PORTION OF THE FIGURE. 
 
 X 190 DIAM. 
 
 PLATE XIX. 
 
THE CEREBELLAR CORTEX 
 
 39 
 
 THE GRANULAR LAYER OF THE CEREBELLUM 
 
 The second, or granular, layer of the cerebellar cortex has about the same thickness as 
 the molecular layer; viz. i millimetre. It is made up of very small cells, 5 to 10 /* in 
 diameter (G), which are polygonal and have a number of very short dendrites. These cells 
 are very difficult to stain. 
 
 Plate XIX. shows three such cells of various shapes and their short, crenated dendrites, 
 which terminate near to the cell in a club-shaped extremity. Kolliker has seen these ends 
 terminating in a fine brush of fibres. Cajal has shown that these cells possess a neuraxon, 
 
 FIG. 6. 
 FIG. 6. Diagrammatic representation of a section through the cerebellar cortex. The cells are reproduced from the plates. 
 
 /, molecular layer. //, granular layer. ///, white matter. P, Purkinje cell with its neuraxon, /, entering the white matter. S, small 
 stellate cells of molecular layer. B, large stellate cells with basket fibres, b. These basket fibres surround the body of the Purkinje cell shown 
 in dotted outline. G, cells of the granular layer with long, straight neuraxon, g, ascending to the molecular layer, and there bifurcating to become 
 tangential fibres. These fibres run at right angles to the plane of section of the plate. M, moss-like termination of white fibres, m, entering the 
 cerebellum from without. //, large Golgi cell of the second type with dendrites in both granular and molecular layers and neuraxon dividing and 
 subdividing within the granular layer. /, terminal filaments and fibres, ', entering the cerebellum from without and ending around the branches of 
 the Purkinje cells. 
 
 which is directed toward the molecular layer (g), and after entering it divides in a T-shaped 
 pair of branches, which pass as horizontal and longitudinal fibres in this layer. In the plate 
 one such neuraxon is seen ascending from the middle cell toward the molecular layer. This 
 
4 o ATLAS OF NERVE CELLS 
 
 plate also shows the large number of horizontal fibres in the molecular layer. The granular 
 layer is made up almost completely of these cells, which are very numerous, lying closely 
 packed together. 
 
 Among them are other cells of larger size, first described by Golgi, multipolar, with many 
 dendrites and a neuraxon, which divides and subdivides soon after leaving the cell, forming 
 an extensive plexus through the molecular layer (ff). These cells belong to Golgi's second 
 type of cell. They are not shown in the plate, but are figured in the diagram (Fig. 6, //). 
 The dendrites of these cells enter the molecular layer, but their neuraxon, with its numerous 
 subdivisions, lies in the granular layer. 
 
 The mutual arrangement of the various cells of the cerebellum may best be understood 
 by the aid of the diagram (Fig. 6). 
 
 The termination of fibres entering the cerebellum in the granular and molecular layers 
 has been investigated by Cajal. He describes in the granular layer a number of branching 
 fibres with short, divergent extremities which have a particularly thick appearance with furry 
 surface, and these he has called the moss-like fibres, since their appearance under the micro- 
 scope resembles a collection of moss-stems (M). This appearance was thought at first to be 
 an artifact, but Van Gehuchten and Retzius have confirmed their existence. This moss-like 
 terminal structure is confined to the granular layer, and seems to bring the impulses coming 
 into the cerebellum through these fibres into relation with the small polygonal cells of this 
 layer. 
 
 The termination of fibres entering the molecular layer of the cerebellum is different from 
 that seen in the granular layer. These fibres, after entering the molecular layer, appear to 
 branch very much in the same manner as the cells of Purkinje branch, and the branches of 
 these fibres run parallel with and quite near to the branches of the Purkinje cells, so that 
 there appears to be a second arborization in this layer, enabling the impulses reaching this 
 layer to pass into the Purkinje cell. 1 Cajal has compared this arrangement to that of ivy 
 climbing up a tree, the Purkinje cell being the tree, and these branches the ivy. 
 
 1 KSlliker does not agree with this view, but thinks the branches are less complex than those of the Purkinje cells and that 
 these fibres come rather into contact with the basket cells. 
 
SECTION THROUGH THE CORPUS DENTATUM OF THE CEREBELLUM SHOWING THE CONVOLUTED 
 APPEARANCE OF THE GRAY MATTER AND THE BUNDLES OF WHITE FIBRES SWEEPING AROUND AND INTO 
 THE CONVOLUTION. THREE LARGE CELLS OF THE CORPUS DENTATUM WITH NUMEROUS DENDRITES 
 ARE SHOWN X 190 DIAM 
 
 PLATE XX. 
 
THE CEREBELLAR CORTEX 41 
 
 THE CORPUS DENTATUM OF THE CEREBELLUM 
 
 Deep within the cerebellum, surrounded on all sides by white fibres, there lies, in each 
 hemisphere, a thin, convoluted mass of gray matter, known as the corpus dentatum. This is 
 made up of nerve cells and of their branches, and appears to be in close connection with the 
 white matter around it. In fact, many fibres in all the peduncles appear to terminate in or 
 to arise from the corpus dentatum, though the chief mass of fibres issuing from it pass into 
 the superior cerebellar peduncle. 
 
 Plate XX. shows the appearance of a portion of this body; the convolutions of the gray 
 matter, with the bundles of nerve fibres sweeping around them, or penetrating into them ; 
 the large polygonal cells which lie near the surface of one convolution ; their dendrites coming 
 off entirely upon the inner side of the cell, i.e. upon the side toward the cavity of the con- 
 volution and branching in every direction ; and the plexus of nerve fibres everywhere present 
 throughout the gray matter. The size of the cells is about 35 /t. They are known to have 
 a neuraxon, but it has not been possible to follow it out or to determine its direction or 
 terminations by the Golgi method. The method of degeneration, however, proves that the chief 
 destination is the red nuclei of the tegmentum, the majority of the fibres decussating with those 
 of the opposite side. It is probable that a part of the fine plexus of nerve fibres in the 
 corpus dentatum consists of terminal filaments of neuraxons coming to this body from a distance, 
 chiefly from the red nuclei of the tegmentum and from the spinal cord. 
 
 There are several small groups of cells forming nuclei within the cerebellum, lying in the 
 portion covering the roof of the fourth ventricle, and hence known as the tegmental nuclei, 
 the cells of which resemble those of the corpus dentatum. 
 
 It may be added that the olivary bodies and the interolivary nuclei of the medulla are 
 composed of cells of identical appearance and structure with those of the corpus dentatum. 
 
 THE CONNECTIONS OF THE CEREBELLUM 
 
 The origin, course, and termination of the various tracts of fibres which enter and leave 
 the cerebellum and which make up the white matter beneath its cortex may best be under- 
 stood by the aid of the diagram (Fig. 7). 
 
 The cerebellum is connected with the other parts of the nervous system by three 
 peduncles. 
 
 I. The inferior peduncle of the cerebellum contains fibres which come from 'several 
 sources, viz. : 
 
42 ATLAS OF NERVE CELLS 
 
 (a) The direct cerebellar column of the spinal cord. The fibres are the neuraxons of 
 the cells of Clarke's column as shown in Fig. 4, page 31. They pass to the cortex of the 
 cerebellum, some of them decussating in the part of the cerebellum which lies above the 
 pons with similar fibres from the other side. No attempt is made in the diagram to show 
 this decussation (i). 
 
 FIG. 7. 
 FIG. 7- Diagram to show the connections of the cerebellum. 
 
 SC, spinal cord. G, column of Goll. B, column of Burdach. MED, medulla oblongata at sensory decussation. O, olive. G, nucleus 
 
 gracilis. C, nucleus cuneatus. V, fifth nerve. Pons, pons Varolii. VIII, eighth nerve and its nucleus. F, fillet. Py, pyramids. CRUS, crus 
 
 cerebri. KN, red nucleus of tegmentum. CD, corpus dentatum of cerebellum. / to Jj, various tracts entering cerebellum from spinal cord, 
 medulla, pons, and crus. For description see text. 
 
THE CEREBELLAR CORTEX 43 
 
 (6) The descending cerebellar fibres. . These arise from the cortex of the cerebellum and 
 pass down the cord mingled with other fibres in the antero-lateral tract. They end in the 
 median gray matter of the cord. Marchi has shown that these fibres are numerous (2). 
 
 (c) The nuclei gracilis (G) and cuneatus (C) of the medulla on both sides, in which the 
 columns of Goll and Burdach end, send a few fibres to the cerebellum. Those on the 
 right side send fibres directly into the right restiform body and thence to the right hemi- 
 sphere of the cerebellum (3), (4). Those on the left side send fibres across the median line 
 of the medulla in the sensory decussation. These then pass forward in the raphe, pass around 
 the surface of the right pyramid and olivary body (O) and join the restiform body, and thus 
 reach the cerebellum (5), (6). These end partly in the corpus dentatum and partly in the cortex. 
 
 (d) The olivary body (O) and interolivary nuclei of each half of the medulla send a 
 large tract across the median line which enters the inferior peduncle and terminates about 
 the corpus dentatum (7). 
 
 II. The middle peduncle of the cerebellum contains the fibres which make up the great 
 mass of transverse fibres in the pons Varolii. 
 
 (a) There is an important bundle of fibres connecting the nucleus vestibuli of the 
 medulla (in which the portion of the acoustic nerve which comes from the semilunar canals 
 ends) with the cerebellum. These lie deep in the mass of fibres of the middle peduncle, 
 and pass to the cortex and nuclei of the vermiform lobe (8). 
 
 (S) There are some fibres which connect the cerebellar cortex and the fillet as shown by 
 the degenerative method, but in which direction they pass is uncertain, probably upward (9). 
 
 (c) Those fibres which arise from the gray masses of the pons pass outward to the 
 cerebellar cortex. These have close relation with the fibres which descend to the pons from 
 the frontal lobes of the brain and thus make a continuous functional tract between the frontal 
 lobe of each side and the cerebellar hemisphere of the opposite side, the crossing occurring 
 in the pons (10). 
 
 (d) Those fibres which arise from the cortex of the cerebellum pass to the gray masses 
 of the ventral part of the pons, from which many new fibres arise which join the pyramids 
 and pass to the medulla or join the descending fibres which pass into the antero-lateral 
 columns of the cord. Some of these fibres decussate in the pons, but the majority end on 
 the side from which they arise (u). 
 
 III. The superior cerebellar peduncle contains two sets of fibres passing in opposite 
 directions between the corpus dentatum and the red nuclei of the tegmentum (12), (13). The 
 majority of these decussate under the corpora quadrigemina, but all do not, hence each corpus 
 dentatum is connected with both red nuclei. 
 
44 ATLAS OF NERVE CELLS 
 
 In the superior cerebellar peduncle are a number of fibres which go beyond the red 
 nucleus and enter the optic thalamus. 
 
 Thus it is evident that the cerebellum has intimate relations with all other parts of the 
 nervous system, a fact which its principal physiological function, that of adjustment of the 
 equilibrium, would naturally presuppose. 
 
 If we review the facts thus far stated, it is evident that there are several ways in which 
 the impulses sent to the cerebellum may awaken a response. 
 
 The impulses which reach the moss work of the granular layer set in action the granule 
 cells ; these in turn, through their fibres which enter the molecular layer, may set in action 
 either the Purkinje cells through their dendrites or the stellate cells ; if the latter are set in 
 action, they may arouse the Purkinje cells by way of the baskets, and then the Purkinje cell 
 may send its impulse outward by its long neuraxon. The impulses which reach the molecular 
 layer by the ivy-like arrangement of fibres come into direct contact with the Purkinje cell, or 
 may act first on the stellate cells and then on the Purkinje cell by the basket fibres. 
 
SECTION THROUGH THE CEREBELLAR CORTEX SHOWING GLIA CELLS AND THEIR VERTICAL 
 FIBRES FORMING A FRAME WORK. X 190 DIAM. 
 
 PLATE XXI. 
 
THE CEREBELLAR CORTEX 
 
 45 
 
 Plate XXI. shows the appearance of the framework of neuroglia which gives support to 
 the cerebellar cortex with its cells. The glia cells lie between the two cortical layers, and 
 send outward to the surface numerous fine fibres which form a framework in which the nerve 
 elements lie. 
 
46 ATLAS OF NERVE CELLS 
 
 THE CORPUS QUADRIGEMINUM ANTERIOR 
 
 THE corpora quadrigemina are four small masses of gray matter lying above the crura 
 cerebri and behind the optic thalamus, and are the smallest of the basal ganglia. They are 
 made up of alternate layers of white and gray matter. They receive numerous bundles of 
 nerve fibres from various parts of the nervous system, and they send out numerous bun- 
 dles of nerve fibres to the cortex. The corpora quadrigemina posterior receive fibres from 
 the acoustic nerve and from the fillet. They send fibres outward to the temporal region of 
 the brain. 
 
 The corpora quadrigemina anterior receive fibres from the optic nerve and from the 
 fillet, and send fibres upward to the occipital cortex of the brain. These ganglia, therefore, 
 are to be regarded as junctions in the passage of nervous impulses, where some impulses 
 are switched off to side tracks, while others pursue their course along the main line. 
 
 Plate XXII. shows a transverse section through the right corpus quadrigeminum anterior, 
 its free surface being seen above, and the aqueduct of Sylvius being seen below to the left. 
 It will be seen that everywhere through this section there is a fine interlacing plexus of nerve 
 fibres. It is also evident that there are several concentric layers to be seen. 
 
 1. There is a layer of glia cells, which are stained deeply in the plate. 
 
 2. The layer of superficial white fibres. In the plate the deeper mass of these nerve 
 fibres only is stained, the surface portion not having taken the stain. The fibres making 
 up this layer come from the optic nerve, their cells of origin lying in the retina. They 
 terminate in the corpora quadrigemina anterior in fine brush-like expansions and free endings. 
 
 3. The superficial gray matter. This forms the layer beneath the superficial white 
 matter. It is narrow toward the median line, but broader along the middle of the body. 
 It contains numerous cells, large in size, of various shapes, with extensive branches. These 
 cells are better shown in Plate XXIII. 
 
 4. The deep layer of white fibres. Some of these are seen in the plate to have a 
 direction from the median line outward. The majority, however, pass in an antero-posterior 
 direction, and are cut across in this section. They radiate widely into the superficial gray 
 matter, and also enter the deeper gray matter. Many of these fibres are supposed to be 
 
\ 
 
 SECTION THROUGH THE ANTERIOR CORPUS QUADRIGEMINUM SHOWING THE VARIOUS LAYERS OF 
 WHITE AND GRAY MATTER. THE AQUEDUCT OF SYLVIUS IS SEEN TO THE LEFT AND BELOW. X 29 DIAM. 
 
 PLATE XXII. 
 
CORPUS QUADRIGEMINUM ANTERIOR 47 
 
 derived from the occipital cortex, being the termination of axis cylinders which arise from 
 the cells there. Others come from, or pass to, the optic thalamus, many crossing the median 
 line in the posterior commissure. 
 
 5. The deep gray matter forming the innermost layer is made up of numerous nerve 
 cells small in size. These do not appear to have any direct communication with the 
 visual apparatus, but are in connection with the fibres of the lateral fillet and of the red 
 nucleus. 
 
 6. The gray matter surrounding the aqueduct of Sylvius has a homogeneous appearance 
 like that surrounding the central canal of the spinal cord, the so-called gelatinous substance 
 of Rolando, and is made up chiefly of neuroglia. 
 
 Von Monakow 1 has shown that when the optic nerve is divided, the superficial white 
 layer and the superficial gray layer undergo atrophy. He has also shown that when the 
 occipital cortex is extirpated, the deep white matter undergoes atrophy, and the superficial 
 gray matter is also slightly affected. Hence, he believes that the visual tract consists of 
 two segments, each of which has a double system of fibres. The first segment extends from 
 the retina to the corpus quadrigeminum anterior (also to the corpus geniculatum externum 
 and optic thalamus). Many of the fibres in this segment arise from cells in the retina, 
 and terminate in a fine network in the corpus quadrigeminum. Others, much fewer in 
 number, arise from the cells in the corpus quadrigeminum and terminate in the retina. The 
 second segment extends from the corpus quadrigeminum (corpus geniculatum externum and thala- 
 mus) to the occipital cortex. Many of the fibres in this segment arise from the cells of the corpus 
 quadrigeminum and terminate in a brush-like expansion in the occipital cortex. Others arise from 
 the pyramid cells of the cortex, and terminate in free extremities about cells of the corpus quadri- 
 geminum. The facts established by the investigation of degenerations after lesions, either experi- 
 mentally produced in animals or found in man, support this arrangement of double tracts, not 
 only in the visual apparatus, but also in the auditory apparatus, and in fact in all the sensory 
 tracts. Hence they must be accepted as proven. A similar arrangement is present in all the 
 tracts entering and leaving the optic thalamus, as will be seen when this ganglion is studied. 
 In the diagram (Fig. 8) these double tracts with their respective origins and terminations 
 are shown. 
 
 The function of the anterior corpora quadrigemina is to coordinate the visual impulses which 
 result in ocular movements whether automatic or voluntary. When it is remembered that the 
 eyes will follow a light either automatically during unconsciousness, or voluntarily because of an 
 act of attention, it becomes evident that a double tract, one from the retina, the other from the 
 
 1 C. von Monakow, Archiv f. Psychiatric, Bd. XX. s. 714. 
 
4 8 
 
 ATLAS OF NERVE CELLS 
 
 cortex, is necessary to put 
 in action the coordinating 
 visual apparatus. Hence 
 the double tract established 
 anatomically is a necessity of 
 physiology. But a mech- 
 anism is also necessary 
 to correlate the sensory im- 
 pressions with the motor 
 nuclei of the ocular muscles. 
 This is found in the cells 
 lying in the superficial and 
 deep gray matter of the 
 corpora quadrigemina, among 
 the brushes and cells already 
 described, which send their 
 neuraxons downward to the 
 III., IV., and VI. nerve 
 nuclei, reaching the latter 
 through the posterior longi- 
 tudinal bundle. These con- 
 nections are shown in the 
 diagram, Fig. 8. 
 
 FIG. 8. 
 FIG. 8. Diagram of the corpora quadrigemina anterior, CQA, showing their connections. 
 
 On the right of the figure the superficial and deep masses of gray matter are shown, pulv, pulvinar of the optic thalamus. /, posterior 
 nucleus of the optic thalamus lying between the corpus geniculatum externum, fge, and the corpus geniculatum internum, fgi. fC, internal 
 capsule. /", fillet. KN, red nucleus of the tegmentum. FED, peduncle of the cerebrum. sn, substantia nigra. OT, optic tract. X, optic 
 chiasm. //, optic nerve. /, 2, fibres from the retina to the pulvinar of the optic thalamus. /, centripetal. .?, centrifugal. 7 and 8, fibres between 
 the optic thalamus and the occipital cortex, 3 and 4, fibres between the retina and the corpus geniculatum externum. 9 and 10, corresponding 
 fibres to the occipital cortex, j and 6, fibres between the retina and the corpus quadrigeminum anterior. // and 12, corresponding fibres to the 
 occipital cortex. For the sake of distinctness the decussating optic fibres only are traced. 13, cell of the superficial gray matter of the CQA sending 
 fibre to the nucleus of the third nerve (/6). 14, cell of the deep gray of CQA sending fibre to third nerve nucleus, rj, cell of the deep gray of 
 CQA sending fibre into the fillet. 77, fibre from the red nucleus terminating about 14. :8, fibre from the fillet terminating about 13. OL, occipi- 
 tal lobe of the brain, with its cortex, containing both cells and terminal brushes of the visual tract. 
 
>* OF THE 
 
 UNIVERSITY 
 
THE CORPUS QUADRIGEMINUM. LARGE POLYGONAL CELLS OF THE SUPERFICIAL GRAY LAYER. 
 X 190 DIAM. 
 
 PLATE XXIII. 
 
CORPUS QUADRIGEMINUM ANTERIOR 49 
 
 Plate XXIII. shows the large nerve cells of the superficial gray layer of the corpus quad- 
 rigeminum anterior. This layer appears to contain the cells which are peculiar to the corpora 
 quadrigemina. 
 
 Two cells are shown in the plate quite different from each other. The larger cell is 
 about 60 /A in size. It has six protoplasmic extensions, from each of which several dendrites 
 are given off. These have numerous branches and radiate widely from the cell, evidently 
 passing to a considerable distance from it. There is no neuraxon visible, but other specimens 
 from this region show that these cells possess a long neuraxon which passes out of the body 
 and assumes an antero-posterior direction, entering either the optic tract or the fillet. 
 
 The second cell shown is a triangular cell with numerous dendrites which have a very 
 large calibre, a long course in various directions, and give off a few branches in their course. 
 The neuraxon of these cells enters the visual tract and passes to the occipital cortex. 
 
 Through the plate the enormous plexus of nerve fibres is evident, and many axis cylin- 
 ders can be seen easily separable from the dendrites by their finer calibre and straighter course. 
 Many have varicosities, but very few give off collaterals. The general direction in which the 
 majority are passing is from without inward. This- plexus belongs to the visual apparatus which 
 has been already described. 
 
ATLAS OF NERVE CELLS 
 
 Plate XXIV. shows a portion of the deep white layer of the corpus quadrigeminum. It 
 demonstrates the inextricable interlacing mass of nerve fibres which appear to be passing in 
 every direction. Among these fibres, numerous cells are seen of varying size. The majority 
 are triangular with a long apical process pointing outward and downward. These cells have 
 numerous dendrites which give off branches. The direction of their neuraxons is downward 
 and inward to the III. nerve nucleus. They are smaller than those in Plate XXIII., 
 being from 20 to 30 /u, in size. Through the mass of fibres there appear to be numberless 
 fine neuraxons with collaterals, and many of them seem to have free ends. Tartuferi 1 and 
 Amaldi * have studied the corpora quadrigemina by the Golgi method, but I have not had 
 access to their writings. 
 
 The corpus quadrigeminum posterior contains cells and fibres quite similar in appearance to 
 those shown in Plates XXIII. and XXIV. It has a relation to the auditory apparatus quite 
 homologous to that which the anterior body has to the visual apparatus. The fillet conveys 
 the auditory impulses to the corpus quadrigeminum posterior and to the corpus geniculatum 
 internum, and to the posterior nucleus of the thalamus, and from these centres a second tract 
 conveys them outward to the temporal region of the cortex. Von Monakow has shown that 
 in both of these tracts fibres pass in both directions as in the visual tracts. 
 
 1 F. Tartuferi, Sull' anatomia minuta d. eminenze bigemine anterior! dell 1 uomo, Archivio ital. p. 1. malettie nervose; 1885. 
 * P. Amaldi, Rivista sperimentale de Freniatria; 1892. 
 
THE CORPUS QUADRIGEMINUM. SMALL AND TRIANGULAR CELLS AND FINE PLEXUS OF NERVE 
 FIBRtS IN THE DEEP LAYER. X 190 DIAM. 
 
 PLATE XXIV. 
 
THE OPTIC THALAMUS 5I 
 
 THE OPTIC THALAMUS 
 
 THE optic thalamus consists of a large mass of gray matter, egg-shaped, which lies upon 
 the floor of each lateral ventricle, its upper and inner surfaces being free, its outer surface 
 lying against the internal capsule, and its lower surface being upon the base of the brain 
 and upon the tegmentum of the crus cerebri. Its connections with the other parts of the 
 brain are made partly through its base, and chiefly through the internal capsule. Upon 
 section either horizontally or vertically, it is evident to the naked eye that bundles of white 
 fibres enter it from the capsule, forming apparent divisions between the masses of gray matter. 
 These white bundles have been called the laminae medullares, and the various divisions of the 
 gray matter formed by the presence of these laminae have led authors to describe a number 
 of different nuclei in the thalamus. The most recent and most accurate description of these 
 nuclei has been given by von Monakow. 1 He describes the following separate masses of 
 gray matter in the thalamus, but admits that there is no absolute separation between these 
 nuclei at some parts of their circumference. That they are, however, distinct from one 
 another has been established by pathological investigations. Although very little is known 
 with regard to the function of these nuclei at present, von Monakow has shown that each 
 nucleus is in anatomical relation with a definite area of the cortex, and hence the separate 
 function of each nucleus can be in part deduced from the facts known regarding the function 
 of the area of the cortex to which it is joined. 
 
 The position of these nuclei in the thalamus is shown in Fig. 9. 
 
 The various nuclei are as follows: 
 
 i st. The tuberculum anterius (to). This is a distinct hillock of gray matter lying upon 
 the upper free surface of the thalamus far forward and near to the median line. It consists 
 of a number of large nerve cells clustered together. The neuraxons of these cells pass 
 downward to the corpus mammillare at the base of the brain, forming the so-called bundle of 
 Vicz d'Azyr. They there meet the fibres of the fornix, which come from the hippocampus, 
 and thus bring the tuberculum anterius into anatomical relation with the cortex of Ammon's 
 horn. 
 
 1 C. von Monakow, Archiv f. Psychiatric, Bd. XXVII. s. 640; 1895. 
 
ATLAS OF NERVE CELLS 
 
 2d. The median nucleus of the thalamus (med) lies behind and below the tuberculum 
 anterius, forming a large mass of gray matter which von Monakow divides into two portions, 
 a median (med, a) and a lateral (med, b] portion. This occupies about one-half of the 
 thalamus in extent, from before backward, being bounded on its lateral surface by the lateral 
 nucleus, and upon its basal surface by the ventral nucleus, and tegmentum of the crura. 
 Von Monakow describes the inner half of this median nucleus as made of larger cells than 
 the outer half, and the two halves are separated from one another by a distinct lamina of 
 white matter. The median nucleus is in anatomical and functional relation with the frontal 
 portion of the Island of Reil and with the second and third frontal convolutions. When 
 
 O T 
 
 
 O T 
 
 FIG. 9. 
 FIG. 9. Frontal sections through basal ganglia to show the nuclei of the optic thalamus. (After von Monakow.) 
 
 A, section at junction of posterior and middle thirds of the thalamus. B, section at junction of middle and anterior thirds of the thalamus. 
 
 OT, optic thalamus. lat, lateral nucleus. med, median nucleus. a, its median division. *, its lateral division. vent, ventral nucleus. 
 
 H, ganglion habenulae. in, substantia nigra. FED, cerebral peduncle. //, optic nerve. la, anterior tubercle. Int Cap, internal capsule. 
 LN, lenticular nucleus. /, lenticular loop. 
 
 these parts of the cortex are extirpated in animals, the nucleus atrophies, and when they are 
 defective in man, this nucleus is defective in development. The cells of the nucleus are shown 
 in Plate XXV. 
 
 3d. The lateral nucleus (lat). This consists of a large mass of gray matter extending 
 from the anterior extremity of the thalamus backward to the pulvinar and lying against the 
 internal capsule, from which it receives numerous fibres. It is made up of large cells which 
 are shown in Plate XXVI. It is in anatomical relation with the central convolutions. 
 
 4th. The ventral nucleus (vent). This nucleus lies beneath the median and lateral 
 
THE OPTIC THALAMUS 53 
 
 nuclei and extends from the anterior backward to the posterior limit of the thalamus, lying 
 very near to the lower portion of the internal capsule. Von Monakow divides this into four 
 sub-nuclei, whose boundaries from one another are very indistinct, but whose independence is 
 assured by the results of degeneration. The anterior part of the ventral nucleus is in 
 anatomical relation with the frontal lobe of the brain in part. The remaining three parts of 
 the ventral nucleus are in relation with the portion of the cortex which lies near the fissure 
 of Sylvius, namely, the operculum, both central convolutions, and the gyrus supramarginalis. 
 The cells of this nucleus are shown in Plates XXVII. and XXVIII. 
 
 5th. The pulvinar of the thalamus is the large collection of gray matter lying free upon 
 its posterior and superior surface, and manifestly connected with the optic tract (pulv in 
 Fig. 8). This lies behind the lateral nucleus and above the posterior part of the ventral 
 nucleus. It is in direct relation with the occipital lobe of the brain. Upon its posterior 
 and under surface two distinct hillocks of gray matter can be made out by the naked 
 eye ; namely, the corpus geniculatum externum (cge) and internum (cgi), Fig. 8. The corpus 
 geniculatum externum is directly related to the occipital cortex, as are the pulvinar and anterior 
 corpus quadrigeminum. The corpus geniculatum internum is in relation with the first and 
 second temporal convolutions, as is also the posterior corpus quadrigeminum. 
 
 6th. The posterior nucleus (pn in Fig. 8). A small mass of gray matter belonging to 
 the thalamus lies beneath the pulvinar between the two corpora geniculata. This nucleus is 
 connected with the cortex lying between the temporal and occipital areas. 
 
 The ganglion habenulae is a small collection of cells forming a distinct hillock on the median 
 surface of the thalamus and giving origin to the fibres which pass backward to the pineal gland. 
 The cells of this nucleus are shown in Plate XXIX. 
 
 Von Monakow affirms that the tracts connecting these various nuclei with their respective 
 portions of the cortex are made up of two sets of fibres. One set has its origin in the 
 cells of the thalamus and its termination in a brush-like expansion about the cells of the 
 cortex. The other set arises from the pyramid cells of the cortex and ends in brushes 
 within the thalamus. Such a double set of fibres has already been shown in Fig. 8, con- 
 necting the pulvinar with the occipital cortex. To establish a further relation between the 
 impulses passing over these double tracts, von Monakow believes that in the thalamus there 
 lie many cells of Golgi's second type, an assertion which the investigations of Marchi confirm. 
 
 There are several varieties of cells in the optic thalamus, and these are shown in Plates 
 XXV. to XXIX. 
 
 Plate XXV. shows a form of cell present in many parts of the thalamus, but especially 
 characteristic of the median and lateral nuclei. These cells have a spherical body from which 
 
54 ATLAS OF NERVE CELLS 
 
 numerous dendrites come off. The number of dendrites varies from four to ten. No one 
 dendrite is very long, but each gives off very numerous fine branches which radiate in all 
 directions from the cell body, giving an appearance which seems to justify the name, which 
 I propose, of stellate cells. These dendritic branches are short and terminate in free extrem- 
 ities in the thalamus. The interlacing of numerous adjacent dendrites with one another 
 forms a fine plexus of fibres in the gray matter. The neuraxon of the stellate cells arises 
 from the body of the cell and passes off in a straight direction, giving off very few collat- 
 erals. The direction of the neuraxon varies in different cells even in the same nucleus. There 
 appears to be no general mass of neuraxons passing together in bundles that can be traced. 
 Yet the existence of the laminae medullares in the thalamus proves that such bundles do pass 
 through the masses of gray matter, and it is probable that the neuraxons gather into bundles, 
 leaving the thalamus through the internal capsule. Marchi l has studied the thalamus by the 
 Golgi method, but his monograph is quite incomplete; he found only two kinds of cells, large 
 and small,' and he does not distinguish the nuclei of the thalamus from one another. 
 
 1 V. Marchi, Sulla struttura del thalami optici, Rev. Sper. di Fren.; 1884 and 1887. 
 

 
 k'. f^^^l 
 t 
 
 MEDIUM SIZED STELLATE CELLS OF THE OPTIC THALAMUS WITH VERY NUMEROUS DEN- 
 DRITES, FROM THE MEDIAN NUCLEUS. X 190 DIAM. 
 
 PLATE XXV. 
 
OF THE 
 
 UNIVERSITY. 
 
 OF 
 
< > 
 
 V 4 
 
 
 LARGE STELLATE CELLS OF THE OPTIC THALAMUS WITH NUMEROUS DENDRITES FROM THE 
 LATERAL NUCLEUS. X 190 DIAM. 
 
 PLATE XXVI. 
 
THE OPTIC THALAMUS 55 
 
 Plate XXVI. shows a cluster of large stellate cells from the lateral nucleus of the thal- 
 amus. They are about 50 /A in diameter. They resemble in appearance the medium-sized cells 
 of the median nucleus, but are distinctly larger. They are to be found scattered irregularly 
 throughout the lateral nucleus ; but a special group of these cells appears to be constant in 
 its inner part, near the dorsal part of the median nucleus. These stellate cells are, however, 
 found along the free margin of the thalamus, m the tuberculum anterius and in the pulvinar, 
 and also occasionally single cells are seen in the internal capsule among the fibres. The 
 neuraxon is difficult to follow and may come off on any side of the cell. It takes a very 
 devious course. Marchi affirms that there are to be found both of the Golgi types of cell in 
 the thalamus. It has not been possible for me to confirm this statement, as the difficulties 
 in the way of following the neuraxons for any distance are insuperable. 
 
ATLAS OF NERVE CELLS 
 
 Plate XXVII. shows a second variety of cell found only in the ventral nucleus of the 
 optic thalamus, and hitherto undescribed. These cells are placed quite regularly at definite 
 distances from one another, so that the nucleus presents a very different appearance under 
 the microscope from other parts of the thalamus, in which irregular distribution of the cells is 
 the rule. These cells are very large, from 50 to 60 p, in diameter. They have a large body, 
 irregular in shape, polygonal as a rule, never fusiform, and from this body numerous dendrites 
 are given off in all directions. These dendrites are very slender in comparison with the size 
 of the cell, but are very long. They appear to take a tortuous course, and branch less freely 
 than the dendrites of the stellate cells. Some of the dendrites are seen to possess the furry 
 surface, which is most marked in the Purkinje cell branches. But there is no uniformity in 
 the direction of these dendrites. The neuraxon, one of which is well shown coming off from 
 the lower side of the cell in the upper right corner of the plate, passes out of the body of the 
 cell on any one of its sides, and gives off few collaterals. There is no uniform direction in 
 the course of the neuraxons of these cells. 
 
LARGE POLYGONAL CELLS OF THE OPTIC THALAMUS WITH NUMEROUS DENDRITES, FROM 
 THE VENTRAL NUCLEUS. X 190 DIAM. 
 
 PLATE XXVII. 
 
v l>:M 
 
 OF THE 
 
 UNIVERSITY 
 
 OF 
 

 LARGE POLYGONAL CELLS OF THE OPTIC THALAMUS WITH VERY NUMEROUS DENDRITES FROM 
 THE VENTRAL NUCLEUS X 190 DIAM. 
 
 PLATE XXVIII. 
 
THE OPTIC THALAMUS 57 
 
 Plate XXVIII. shows the cells of the ventral nucleus of the thalamus. The nucleus is 
 bounded above by a distinct band of nerve fibres, entering the thalamus from the internal 
 capsule and sweeping inward toward the median nucleus. It is bounded on its outer and lower 
 surface by the fibres of the internal capsule. Hence it is more distinct as a separate nucleus 
 than any other of the thalamic masses. It is also distinguished from other nuclei by the size 
 and appearance of its cells, as already described. The cells of Luys' body, which lies not far 
 from this nucleus among the longitudinal bundles of the internal capsule, resemble these cells. 
 
58 ATLAS OF NERVE CELLS 
 
 Plate XXIX. shows a group of cells of smaller size than those already shown and chiefly 
 triangular in shape, which is found on the median border of the thalamus in its posterior 
 portions. These cells are not only grouped together, but are scattered along the median 
 surface of the thalamus in a thin layer over a considerable area. The large group is the 
 ganglion habenulae, shown as a distant hillock (/t) in Fig. 9, A. This ganglion gives rise to 
 a set of fibres which pass backward to the level of the posterior commissure, forming the pillars 
 of the pineal gland. It also gives origin to a band of fibres which passes downward and 
 backward between the red nuclei of the tegmentum to a small mass of gray matter lying 
 between the crura. This band is the fasciculus retroflexus of Meynert. It is shown in Fig. 8. 
 
 The majority of the cells of the ganglion are triangular, with three chief dendritic branches, 
 but some are polygonal, with many dendrites. The neuraxon is very fine, and has few, if 
 any, collaterals. The majority of the neuraxons turn downward and probably pass into the 
 fasciculus retroflexus. In the plate the free median border of the thalamus is seen to the left. 
 
SMALL TRIANGULAR CELLS OF THE NUCLEUS HABENULAE OF THE OPTIC THALAMUS. 
 
 PLATE XXIX. 
 
SECTION THROUGH THE MARGIN OF THE OPTIC THALAMUS SHOWING THE ENTRANCE OF ITS 
 FIBRES INTO THE INTERNAL CAPSULE. TWO FUSIFORM CELLS ARE SEEN WHOSE LONG AXIS IS 
 PARALLEL WITH THE CAPSULAR FIBRES. THE MAJORITY OF AXIS CYLINDERS APPEAR VARICOSE AND 
 A FEW ARE SEEN TO GIVE OFF COLLATERALS. X 190 DIAM. 
 
 PLATE XXX 
 
THE OPTIC THALAMUS 59 
 
 Plate XXX. shows the outer margin of the optic thalamus adjacent to the internal capsule. 
 The object of the plate is to demonstrate the ' enormous number of fine neuraxons issuing 
 from the thalamus and entering the internal capsule. The thalamic fibres are shown in the 
 plate as passing from above downward and toward the left. The internal capsule fibres are 
 shown passing across the plate below. Among the latter, two long fusiform cells are seen 
 near the thalamus, with very long branches which subdivide at some distance from the cell. 
 Such cells are seen along the outer side of the thalamus in all sections and even in the 
 laminae medullares within the thalamus. They form a fourth variety of thalamic cell. The 
 fibres from the thalamus turn in a spiral manner on entering the capsule and become parallel 
 with the fibres of the capsule, but such turning cannot be shown by a photograph, which is 
 necessarily in one plane only. 
 
 Von Monakovv has proven by experimental researches on animals and by studies in pathol- 
 ogy in man that the neurons in the thalamus must be considered as made up of two cat- 
 egories: (i) cells which lie in the thalamus and send out neuraxons which pass to and terminate 
 in the cortex; (2) neuraxons of cortical cells passing into the thalamus and terminating in 
 brush-like expansions about its cells. By means of these two sets of neurons the mutual 
 relation of the nuclei of the thalamus and the cortical areas, already described, are maintained. 
 He believes that the mutual relations of these neurons is secured by the intermediate action 
 of cells of Golgi's second type lying in the optic thalamus. I have not included such cells 
 in the diagram, Fig. 8. There are undoubtedly many other fibres leaving and entering the 
 thalamus to and from the lower levels of the central nervous system, as well as those which 
 enter it from the optic tracts. The latter are clearly connected with the pulvinar of the 
 thalamus and with the corpora geniculata. There is no reason to believe that the other 
 nuclei of the thalamus have any anatomical relation with the optic tracts. The relations known 
 to exist between the optic nerve, pulvinar, and occipital cortex may, however, be taken as in- 
 dicative of the relation between other parts of the thalamus and other subcortical and cortical 
 structures. These have been shown already in the diagram of the visual tract, Fig. 8. 
 
 It is probable that the thalamus is the organ in which many automatic acts are coordi- 
 nated under the stimuli of various sensory impressions there united and brought into mutual 
 relation. Its exact function is, however, undetermined. 
 
60 ATLAS OF NERVE CELLS 
 
 THE CORPUS STRIATUM 
 
 THE corpus striatum of each hemisphere, a single mass of gray matter on the base of 
 the brain in its anterior part, is divided above into two parts, the caudate and lenticular nuclei, 
 by the passage downward of the fibres of the internal capsule. The caudate nucleus has a free 
 upper surface forming a part of the floor of the lateral ventricle. The lenticular nucleus is 
 surrounded on all sides by white matter, being bounded externally by the external capsule, in- 
 ternally by the internal capsule, and below by the lenticular loop. It is subdivided into an 
 external body or putamen, a median portion, the lobus intermedius, and an internal portion, 
 the lobus pallidus, by two bands of white fibres coursing through it from the internal capsule 
 to the lenticular loop. These relations are shown in Fig. 9, B, on page 52. 
 
 Microscopical examination of the caudate and lenticular nuclei shows that they are essen- 
 tially alike in structure. They are made up of gray substance containing two varieties of cells; 
 very large rectangular cells and very small triangular or polygonal cells. These cells are 
 scattered indifferently through the mass of gray matter and are never collected into groups. 
 Plate XXXI. shows the large variety of cells characteristic of the corpus striatum. They are 
 70/1 long and 10 /u, to 15/11 broad. The body of the cell is long and narrow with a swell- 
 ing in its middle where the nucleus lies. At the extremities of the cell body dendrites are 
 given off, either one or two in number, and these appear to turn at right angles to the long 
 axis of the cell soon after leaving it. They give off very few branches and have a very long 
 course, being often traceable for a great distance through the gray matter and even into the 
 internal capsule. The neuraxon may arise from the body of the cell, but usually arises from 
 one of its protoplasmic prolongations near the body. It usually turns soon after its origin 
 toward the internal capsule, but does not uniformly arise from the side of the cell toward the 
 capsule. The majority of the neuraxons in the caudate nucleus enter the capsule from its 
 under surface. Those from the lenticular nucleus either pass into the lenticular loop by way 
 of the laminae, dividing the nucleus into its three parts, or else enter the capsule after travers- 
 ing the lobus pallidus. 
 

 I 
 
 LARGE RECTANGULAR CELLS OF THE CORPUS STRIATUM. THESE CELLS ARE FOUND ONLY 
 IN THIS REGION. X 190 DIAM. 
 
 PLATE XXXI. 
 
^Wx 
 
 OF THc 
 
 ,'ERSITY 
 
 
 OF 
 
Of THE 
 
 ( UNIVERSITY ) 
 
SMALL TRIANGULAR AND STELLATE CELLS IN THE CORPUS STRIATUM X 190 DIAM. 
 
 PLATE XXXII. 
 
THE CORPUS STRIATUM 6 1 
 
 Plate XXXII. shows the second or smaller variety of cells found in the corpus striatum. 
 These may be stellate cells quite like those of the thalamus, of which one is shown in the 
 plate ; or they may be very small triangular cells, with protoplasmic prolongations coming off 
 at right angles, or at acute angles to each other. Several such small cells are shown in 
 the plate. The majority of the dendrites of these small cells run in an antero-posterior 
 direction ; hence in the section shown, which is a frontal section, they are cut off, and 
 cannot be traced far. They have few branches, and are not long. The neuraxons of these 
 cells come off in all directions, and cannot be followed. It has been asserted by Marchi 
 that cells of both the types of Golgi are present in the corpora striata. It is evident in 
 both plates that a fine plexus of neuraxons is present throughout the gray matter, and 
 numerous collaterals can be seen coming off from these fibres. 
 
 No definite relation between the corpus striatum and the cortex has been demonstrated, 
 though Kovalewski l believes that fibres enter the outer surface of the putamen from the 
 corona radiata, and external capsule, and end there, establishing such a relation. 
 
 The function of these ganglia is undetermined. 
 
 1 Sitzb. d. K. Akad. d. Wissensch. im Wien, Bd. LXXXVI. iii. Abt., December, 1882. 
 
6 2 ATLAS OF NERVE CELLS 
 
 THE CEREBRAL CORTEX 
 
 GENERAL TOPOGRAPHY 
 
 PLATES XXXIII. to LIII. show the structure of the cerebral cortex. This is the most 
 intricate portion of the nervous system. It has formed the subject of numerous careful 
 studies, but there remain many points which are still obscure, and there is as yet much 
 uncertainty regarding the arrangement of its cells and the direction of its fibres. The 
 methods of Golgi have, however, been particularly successful in demonstrating the constituent 
 parts of the cortex, in revealing the existence of numerous varieties of cells, in tracing the 
 division and destination of their branches, and in unravelling the tangled mass of fibres 
 which permeate it everywhere. The description here given rests largely upon the study of 
 the facts presented by Golgi, Cajal, and Retzius, and illustrated in the plates. 
 
 Plate XXXIII. shows a section through the cortex of the Rolandic region of a human 
 embryo of eight months. It gives a general idea of the topography of the cortex at a 
 period in the development when the structure is simple, and the various layers can be easily 
 distinguished. It is evident that there are several layers in the gray matter. In this plate the 
 superficial layer is so deeply stained that little of its structure can be seen, although some fibres 
 parallel to the surface, so-called tangential fibres, can be made out in the lighter parts. The 
 characteristics of this layer are shown in Plates XXXIV. to XXXVII. Beneath the super- 
 ficial layer a comparatively clear region is seen having a striated appearance, because 
 traversed by long fibres at right angles to the plane of the surface. These fibres are really 
 the long, slender, upright apical extensions of the pyramidal-shaped cells which form the 
 second and third layers. A careful inspection, however, will show here and there in this 
 clear zone some small cells with little apices extending upward toward the surface. And 
 when the layer of large pyramidal cells is studied, it will be noticed that these cells lie at 
 different levels, a number of smaller ones being nearer to the surface than the very large 
 ones which first catch the eye. A number of deeper lying cells are visible at the right of the 
 plate. There is really no sharp dividing line between the layers of small and large pyramid 
 cells; and although modern authors divide them into two distinct layers, the plates demon- 
 strate that they intermingle. Beneath the layer of cells is another fairly clear zone occupied 
 
uk 1 W-1 I .' 
 
 SECTION THROUGH A CONVOLUTION OF THE HUMAN BRAIN, FROM AN EIGHT MONTHS EM- 
 BRYO. THE VARIOUS LAYERS OF THE CORTEX ARE SHOWN AND THE WHITE FIBRES PASSING TO- 
 GETHER BENEATH THE CORTEX. X 29 DIAM. 
 
 PLATE XXXIII. 
 
THE CEREBRAL CORTEX 63 
 
 chiefly by the basal or axis cylinder processes of the pyramid cells which sweep downward, 
 forming the mass of parallel fibres which make up the white matter under the cortex. But 
 it is quite evident that interlacing with these fibres, and passing at right angles to them, 
 are many other fibres ; and although the cells from which they arise are not here shown, it 
 can easily be believed that these cells and fibres form a fourth layer. After these plates 
 have been studied, some statements will be made regarding the arrangement of the cells in 
 different layers in the various regions of the brain. 
 
6 4 
 
 ATLAS OF NERVE CELLS 
 
 THE FIRST, OR SUPERFICIAL LAYER 
 
 Plates XXXIV. to XXXVII. show the superficial layer of the cortex. This is also known 
 as the molecular layer, from its punctate appearance under the carmine stain and from its 
 supposed lack of nerve cells. The modern methods of staining have demonstrated the existence 
 of a peculiar type of cell in this layer and also the existence of a fine interlacing mass of 
 fibres. These cells having been first described by Cajal are now known by his name ; but 
 Retzius l has studied them more thoroughly in the human brain, and to him we owe most of 
 our knowledge of these structures. There are several varieties of Cajal cells, all of which are 
 shown in the plates. 
 
 Plate XXXIV. shows two of these cells, one fusiform, one triangular. The fusiform cell 
 has an oval body, whose long axis is parallel to the surface of the convolution ; and two 
 long protoplasmic processes, called stalks by Cajal, which pass out nearly straight in the hori- 
 zontal plane. These stalks, whose surface is smooth and not covered with gemmules, give off 
 at right angles numerous fine filamentous fibres. These appear to extend vertically toward the 
 surface, and at their ends little knob-like terminations are often seen. They sometimes divide 
 or give off fine ramifications, which in turn become horizontal in their course. These fibres 
 do not leave the surface layer of the convolution. The end of the stalk terminates in a fine 
 fibre, which is very long, and extends a great length horizontally through this layer. The 
 second cell in the plate is a rectangular, diamond-shaped cell. From this cell the stalks come 
 off at right angles to each other, two being horizontal, two being vertical. From both hori- 
 zontal stalks fine fibres arise, which take a vertical course, and the right-hand stalk finally 
 turns upward. From the lower pointed process of the cell a neuraxon comes off which gives 
 off a collateral, then divides and becomes a tangential, fibre. This is the course taken by the 
 neuraxons of all these Cajal cells, as has been clearly demonstrated by Retzius. They do not 
 leave the superficial layer, but become tangential fibres. The plate shows the large number 
 of tangential fibres. These appear somewhat thicker than in the subsequent plates, owing to 
 a difference in the photographic method adopted. 
 
 Plate XXXV. shows a large triangular Cajal cell with two long stalks, both giving off 
 
 numerous fine vertical branches which pass upward to the surface of the cortex. One of the 
 
 branches becomes a horizontal fibre and can be traced for some distance. The neuraxon is 
 not shown. 
 
 1 Retzius, Biologische Untersuchungen, Bd. III.; 1894. 
 
I 
 
 THE SUPERFICIAL LAYER OF THE CEREBRAL CORTEX OF A HUMAN EMBRYO SHOWING CAJAL 
 CELLS, TANGENTIAL FIBRES, AND FINE INTERLACING VERTICAL FIBRES. X 190 DIAM. 
 
 PLATE XXXIV. 
 
THE SUPERFICIAL LAYER OF THE CEREBRAL CORTEX OF AN EIGHT MONTHS HUMAN EMBRYO, 
 SHOWING TRIANGULAR CAJAL CELL WITH ITS NUMEROUS SECONDARY BRANCHES. X 190 DIAM. 
 
 PLATE XXXV. 
 
THE SUPERFICIAL LAYER OF THE CEREBRAL CORTEX OF AN EIGHT MONTHS HUMAN EMBRYO, 
 SHOWING TWO CAJAL CELLS, ONE FUSIFORM, THE OTHER POLYGONAL. X 190 DIAM. 
 
 PLATE XXXVI. 
 
' 
 
 
 
 THE SUPERFICIAL LAYER OF THE CORTEX SHOWING TRIANGULAR CAJAL CELL WITH VERTI- 
 CAL BRANCHES TERMINATING IN KNOBS. ALSO DEEP TANGENTIAL FIBRES WITH THEIR COLLATER- 
 ALS. X 190 DIAM. 
 
 PLATE XXXVII. 
 
THE CEREBRAL CORTEX 65 
 
 Plate XXXVI. shows a fusiform cell (0 1 ) and a polygonal cell (c) of the Cajal type in the 
 superficial layer of the cortex. The polygonal cell has several dendrites coming off in various 
 directions and dividing into smaller branches, which diverge. It also has a single neuraxon, 
 very small in diameter, which passes downward into the surface layer and becomes horizontal 
 and gives off collaterals, which also take a horizontal course. These fibres never turn down- 
 ward to enter the deeper layers. These cells differ in all respects from glia cells, which 
 Golgi has shown to exist in large numbers in this layer. 
 
 In all the plates it is evident that there are numerous horizontal fibres in this superficial 
 layer. These are well marked in its deepest portion, and are shown with great distinctness 
 in Plate XXXVII. Into the surface layer there are seen to pass (in Plates XXXVIII. and 
 XLII.) the termination of the apical processes of the pyramidal cells. These have the typical 
 dendrite appearance, being covered with granules, and branching and dividing as they ascend. 
 These dendrites appear to end in free extremities in the superficial layer, there coming into 
 contact with the nerve fibres already described. 
 
 Plate XXXVII. shows the triangular variety of Cajal cells (6). It also shows the fine 
 interlacing fibres of this layer. The origin of some of these fibres has already been described. 
 There are others which require to be noticed. It has already been stated that the neuraxons 
 ascending to the cortex from the spinal cord, the medulla, the cerebellum, and the basal 
 ganglia terminate in fine brush-like expansions in the cortex. It is in this layer of the cortex 
 that many of these terminal brushes lie. Hence this layer of the cortex certainly receives 
 impulses of many kinds from various parts of the lower nervous centres. It also receives 
 impulses from other regions of the cortex by means of fibres whose terminal expansions have 
 been traced to this layer. Thus the fibres of association from adjacent or distant pyramidal 
 cells end here, as do also many of the collateral fibres coming off from the neuraxons of 
 pyramidal cells in the layer just beneath the superficial layer, and also the neuraxons of 
 certain cells which pass directly to this layer; viz. the Martinotti cells. 
 
 If, as Cajal and Van Gehuchten believe, it is the function of the branches of the apical 
 processes of the pyramidal cells of the second and third layers of the cortex to collect 
 impulses and convey them to their cells, it is evident that in the superficial layer of the 
 cortex they can receive such impulses from the most diverse and far separated parts of the 
 central nervous system. 
 
 1 These letters refer to the cells shown in the diagram Fig. 10 on p. 72. 
 
66 ATLAS OF NERVE CELLS 
 
 THE SECOND LAYER, OR LAYER OF SMALL PYRAMIDAL CELLS 
 
 Plate XXXVIII. shows the second layer of the cortex or the layer of small pyramidal 
 cells; and Plates XXXIX. and XL. demonstrate the characteristic features of the cells of 
 this layer. In Plate XXXVIII. the surface layer is seen, but its cells and fibres, excepting 
 a few deep tangential fibres, are not shown. Just beneath these deep tangential fibres a 
 number of small pyramid cells (d) are evident with their short apical processes, really a part 
 of the body of the cell, running up toward the superficial layer. The size of these cells is 
 from io/t to 12 p.. In one or two cases the apical process can be seen to bifurcate, and its 
 two branches can be seen to diverge and end in the superficial layer. Other apical processes 
 coming up from cells lying more deeply, and not shown, can also be seen to take the same 
 course. The small pyramid cells are seen to have a few dendrites coming off from their 
 basal corners and a true neuraxon coming from the base which can be followed for some 
 distance downward, and in some cases can be seen to give off collaterals. The majority of 
 the dark broad black lines in this plate are capillary blood-vessels. 
 
 In the lower part of the plate the dense mass of nerve fibres interlacing and running 
 in all directions is noticeable. All through this layer it is evident that fine neuraxons as 
 well as apical projections of deeper cells are making their way toward the surface layer, many 
 of them giving off collaterals in their course. 
 
 Plate XXXIX. shows this second layer under a higher power of magnification and brings 
 out the structure and appearance of a few of its cells and the enormous mass of its fibres. 
 Plate XL. shows the characteristics of its cells alone. 
 
 The cells have a triangular shape with narrow base and long apex (e). The body gives 
 off a few branches which pass outward, usually downward, from the two basal corners of the 
 cell. These have the characteristic features of dendrites, tapering as they pass on, having 
 an irregular surface, giving off a few branches at an acute angle, which branches in turn 
 divide and give off secondary branches. The general direction of these dendrites is diagonal 
 to the vertical fibres of the layer. The neuraxon also comes off from the base of the cell, 
 but rarely from a corner, usually in the middle. It differs from the dendrite in being 
 smooth, fine, of uniform calibre, and giving off its finer collaterals at right angles to its 
 course. Cajal and Golgi believe that many of the collaterals turn in their course and ascend 
 through the second layer to the superficial layer. It is evident in the plate that this layer 
 
THE SECOND LAYER OF THE CEREBRAL CORTEX OR LAYER OF SMALL PYRAMID CELLS. THE 
 STRIATED APPEARANCE OF THIS LAYER IS DUE TO THE LARGE NUMBER OF FIBRES COMING UP 
 FROM THE THIRD LAYER. THE BIFURCATION OF THE APICAL PROCESSES OF THE PYRAMID CELLS 
 BEFORE ENTERING THE SUPERFICIAL LAYER IS WELL SHOWN. VERY SMALL PYRAMID CELLS WITH 
 NUMEROUS DENDRITES ARE SEEN AT DIFFERENT PORTIONS OF THIS LAYER. X 80 DIAM. 
 
 PLATE XXXVIII. 
 
THE SECOND LAYER OR LAYER OF SMALL PYRAMIDAL CELLS OF THE HUMAN CORTEX, SEV- 
 ERAL SMALL CELLS AND VERTICAL AXIS CYLINDERS PASSING THROUGH THIS LAYER ARE SHOWN. 
 X 190 DIAM. 
 
 PLATE XXXIX. 
 
PYRAMIDAL CELLS OF THE ADULT HUMAN CEREBRAL CORTEX. THE SHAPE OF THESE CELLS 
 THE LONG APICAL PROCESS PASSING UPWARD TO THE SUPERFICIAL LAYER, AND THE NUMEROUS 
 SHORT BASAL DENDRITES ARE SHOWN. X 190 DIAM. 
 
 PLATE XL. 
 
THE CEREBRAL CORTEX 67 
 
 contains many fine neuraxons whose course can be traced to the surface layer though their 
 origin is not seen in the plate. They present a very much finer appearance than the apical 
 processes of the cell near to which they lie. 
 
 Plate XL. shows the small pyramidal cells (/) of the second cortical layer. These are 
 more distinctly triangular than those shown in Plate XXXIX., and have longer apical processes 
 as they lie at a deeper level. The number of branches from the apical process is larger. 
 
 The apical process or stalk of the pyramidal cell is for some distance almost as large as 
 the cell body, and it is many times as long, but gradually becomes narrower and lies straight, 
 and finally divides like a dendrite into branches. In its ascent toward the surface it con- 
 stantly gives off branches which pursue the same direction or which turn upward soon after 
 passing away from the stalk. In Plate XL. such branches of the apical process are very clearly 
 seen. As the apical process becomes more slender, it often appears to present a varicose 
 appearance, and such varicosities are also seen in its branches. 
 
68 ATLAS OF NERVE CELLS 
 
 THE THIRD LAYER, OR LAYER OF LARGE PYRAMIDAL CELLS 
 
 Plate XL I. gives a general view of the second and third layers of the cortex, or the 
 layers of pyramidal cells. These two layers are only to be distinguished from one another by 
 the size of their cells, the deeper cells being much larger and reaching from 15 //, to 40 /x in 
 diameter. There is no actual boundary between the two layers, and Cajal prefers to consider 
 the two as one. In fact, the cells are small, intermediate, and large, as they lie deeper. 
 Their apical processes or stalks reach up to the superficial layer. The cells are seen to lie 
 at various levels, yet they are so nearly together as to make quite a distinct stratification of 
 the cortex. The characteristics of the cells in the deep or third layer are the great size of 
 the cells, the length of their apical process, and the greater number of the vertical fibres 
 which give this layer a striated appMfSfee. The numerous dendrites of the cells, running out 
 from their bases, and also the '^n^lraxoiW with collaterals, can be seen in the plate. The 
 
 '-- f; * | jcp RH\>'^ 
 
 special features of this layer are *eesHto better advantage in the plates which follow. 
 
 Plate XLII. shows very clearly a group of intermediate-sized pyramidal cells in a human 
 embryo of eight months. 
 
 The cells of the second layer did not stain, hence the course of the apical process can 
 be traced to the superficial layer. 
 
 The characteristic triangular appearance of these cells is to be noticed ; their dendrites 
 coming off at the corners and subdividing; their long apical projection running far up into 
 the second layer, giving off branches as it ascends, and finally splitting up into two parts, 
 which diverge and enter the superficial layer of tangential fibres ; and lastly, their axis cylin- 
 der process, coming out of the base, passing downward and giving off one or two collaterals 
 in its course. 
 
 It is quite evident that the apex projection has a rough surface covered by gemmules or 
 little spike-like or thorn-like excrescences. These appear to be better marked on this process 
 than on the other protoplasmic processes, though they can be seen on them as well. On 
 several of the apical processes, extending through the plate, varicosities are seen. Andriezen 1 
 is inclined to attach a pathological significance to these, but Berkley 2 considers them as nor- 
 mal. Their significance is not known. In this plate and in Plates XL 1 1 1. and XL VI. a few 
 
 1 W. L. Andriezen, On Some of the Newer Aspects of the Pathology of Insanity. Brain, Vol. XVII., p. 548; 1894. 
 
 2 H. ]. Berkley, A Theory of the Causation of Permanent Dementia. The Medical News; Nov. 9, 1895. 
 
SECTION THROUGH THE SECOND AND THIRD LAYERS OF THE HUMAN CORTEX SHOWING BOTH 
 SMALL AND LARGE PYRAMIDAL CELLS LYING BETWEEN NUMEROUS VERTICAL FIBRES- X 190 DIAM. 
 
 PLATE XLI. 
 
\ 
 
 1 
 
 , 
 
 
 
 PYRAMIDAL CELLS OF THE CEREBRAL CORTEX, EIGHT MONTHS HUMAN EMBRYO. THE LONG 
 APICAL PROCESS PASSING UPWARD TO THE SUPERFICIAL LAYER, THE NUMEROUS SHORT BASAL 
 PROCESSES AND THE NEURAXON COMING OUT OF THE BASE ARE SHOWN IN THIS SECTION; IT 
 IS POSSIBLE TO SEE COLLATERAL BRANCHES COMING OFF FROM THE NEURAXON. X 190 DIAM. 
 
 PLATE XLII. 
 

 PYRAMIDAL CELLS OF THE CEREBRAL CORTEX, EIGHT MONTHS HUMAN EMBRYO. THE SHAPE 
 OF THESE CELLS, THE LONG APICAL PROCESS PASSING UPWARD TO THE SUPERFICIAL LAYER, THE 
 NUMEROUS SHORT DENDRITES AND THE NEURAXON WITH ITS COLLATERALS ARE SHOWN IN THIS 
 SECTION. X 190 DIAM. 
 
 PLATE XLIII. 
 
MEDIUM SIZED PYRAMIDAL CELLS OF THE THIRD LAYER OF THE CORTEX IN HUMAN EMBRYO 
 SHOWING LONG APICAL PROCESS ENTERING SUPERFICIAL LAYER AND FINE NEURAXON COMING C 
 OF THE BASE. X 190 DIAM- 
 
 PLATE XLIV. 
 
LARGE PYRAMIDAL CELL OF GOLGI'S SECOND TYPE IN THE THIRD LAYER WITH LONG APICAL 
 PROCESS AND NUMEROUS FINE DENDRITES. THE NEURAXON DIVIDES AND SUBDIVIDES BELOW THE 
 CELL BODY. 
 
 PLATE XLV. 
 

 I 
 
 - 
 
 
 A LARGE CELL OF THE THIRD LAYER WITH ITS NEURAXON COMING OUT FROM THE BASE 
 AND DIVIDING INTO TWO BRANCHES WHICH SUBSEQUENTLY DIVIDE. THIS IS AN EXAMPLE OF THE 
 SECOND TYPE OF GOLGI CELL. X 570 DIAM. 
 
 PLATE XLVI. 
 
THE CEREBRAL CORTEX 69 
 
 fine straight fibres, with bead-like swellings at regular intervals, are to be seen extending through 
 the layer. These are thought by Retzius to be neuroglia fibres. 
 
 Plate XL 1 1 1. shows another group of pyramidal cells (g) with the same characteristics as 
 those in Plate XLII. In this plate the axis cylinder process or neuraxon of several of the 
 cells is well shown. It can be seen to issue from the base of the cell, and to pass 
 downward toward the white matter in a wavy course, and to give off little collaterals here 
 and there, some of which appear to turn backward, as if to ascend. Cajal affirms that some 
 collaterals do turn upward and pass into the superficial layer to ramify among the horizontal 
 fibres. In this plate, also, a varicose neuroglia fibre is to be traced. It will be noticed that 
 in the layer beneath the pyramids, where no cells are stained, there are many more fibres 
 of horizontal direction than in the second layer. 
 
 Plate XL IV. shows the characteristics of the pyramidal cells already described. The 
 neuraxon of the cell in the centre of the plate is well shown, and the very great difference 
 in its appearance from that of the dendrites is to be noticed. A small collateral is seen to 
 come off from this neuraxon near to the lower limit of the plate. 
 
 Plate XLV. shows another pyramidal cell of medium size from the third layer (k). It 
 is given in order to ' demonstrate the branches which often come off from the apical process 
 and ascend at its side. It also shows the divisions and subdivisions of the dendrites given 
 off from the body of the cell, and the tendency to varicose swellings along these dendrites, 
 particularly at the point at which a branch is given off. 
 
 In Plate XLVI. a single pyramidal cell is shown with its axis cylinder or neuraxon 
 under a very high power of magnification. This is seen to run from the base and to divide 
 into two branches; each branch then divides, and from these branches smaller collaterals are 
 given off. This cell is one of the cells of Golgi's second type, in which the axis cylinder 
 has no long course or definite destination, but loses itself in a plexus of very fine nerve 
 fibres. It was impossible to focus both cell and fibre upon the plate at once, with this very 
 high power, hence the cell is not clearly shown. 
 
7 o ATLAS OF NERVE CELLS 
 
 Plate XLVII. shows three of the large pyramidal cells of the cortex (g). In this plate 
 the existence of thorn-like excrescences upon the apical process of the cell is clearly shown. 
 These are the so-called gemmules, and it will be noticed that many of them have club-shaped 
 extremities. These gemmules are similar to those already seen on the branches of the 
 Purkinje cells of the cerebellum. Their exact significance is unknown, but Berkley considers 
 them of great functional importance, and has shown J that in degenerative diseases of the brain 
 of toxic origin, they are the first part of the cell to suffer, the degenerated cell having a 
 swollen apical process without gemmules. In this plate the division of the apical process and 
 the divergence of its branches before they enter the superficial layer are well seen. 
 
 To the right of the plate a long straight neuraxon is seen passing toward the superficial 
 layer. This may be taken as one of the terminal fibres of which so many come up from 
 the cord and basal ganglia and end in the cortex. These, of course, pass between and 
 interlace with the other systems of fibres already described, and make up a considerable part 
 of the white matter. The terminal filaments of these fibres and the terminals of their 
 collaterals enter into close relation with the cells of this layer, and they may thus bring in 
 impulses from a distance, which, being received in the pyramidal cells, set up new responses 
 in the form of motor or sensory or mental activity. 
 
 1 The Medical News; Nov. 9, 1895. 
 
LARGEST PYRAMIDAL CELLS OF THE THIRD LAYER OF THE CORTEX SHOWING LONG APICAL 
 PROCESS WITH GEMMULES AND NUMEROUS BASAL DENDRITES. X 190 DIAM. 
 
 PL'ATE XLVII. 
 
 OF TMC ^A 
 
 ( LfMVERS'TY) 
 'W r'.>s^,X 
 
A MARTINOTTI CELL OF THE THIRD LAYER OF THE CORTEX SHOWING THE AXIS CYLINDER 
 FROM THE APEX AND ASCENDING TOWARD THE SUPERFICIAL LAYER GIVING OFF A 
 LLATERAL. OTHER AXIS CYLINDERS WITH COLLATERALS ARE SEEN IN THE SECTION THE STEL- 
 LATE CELL TO THE LEFT IS A GLIA CELL. X 190 DIAM. 
 
 PLATE XLVIII. 
 
or THE 
 UNIVERSITY 
 

 I 
 
 
 
 THIRD LAYER OF THE CORTEX SHOWING IRREGULAR POLYGONAL CELLS WITH MANY DEN- 
 DRITES WHICH LIE AMONG THE PYRAMIDAL CELLS. X 190 DIAM. 
 
 PLATE XLIX. 
 
THE CEREBRAL CORTEX 
 
 Plate XLVIII. shows the existence of another type of cell found in the third layer of 
 the cortex, and first described by Martinotti (m). This variety of cell gives off an axis 
 cylinder from its apex which ascends toward the superficial layer 'of the cortex, a direction 
 exactly opposed to that of the neuraxon of the pyramidal cells. This axis cylinder gives off 
 collaterals which pass laterally into the adjacent substance. These cells have numerous 
 dendrites which come off from the lower part of the cell body and pass downward, dividing 
 and branching, and furnished with gemmules like the apical process of the pyramidal cell. 
 
 The cell to the left of the plate is a glia cell. The numerous branches of such a cell 
 and its spider-like appearance are apparent. The broad black lines are capillaries. 
 
 Plate XLIX. shows several cells of irregular type which are found in the third layer of 
 the cortex. They differ in shape from the pyramidal cells. They do not have an apical 
 process, but have numerous long, slender dendrites, and the neuraxon is not easily separated 
 from the dendrites. It is possible that these cells belong to Golgi's second type of cell 
 whose neuraxons divide and subdivide within the cortical layers (). 
 
ATLAS OF NERVE CELLS 
 
 THE FOURTH LAYER, OR LAYER OF 
 POLYGONAL CELLS 
 
 Plate L. shows the fourth or deepest 
 layer of the cortex. This layer is made 
 up of polygonal or fusiform cells, whose 
 general direction is at right angles to the 
 pyramidal cells of the third layer. One 
 such cell is shown in the plate, a fusi- 
 form cell (/), its long axis being hori- 
 zontal. It has two large protoplasmic 
 processes covered with gemmules and 
 giving off branches. It has a large 
 neuraxon which passes almost horizontally 
 toward the left, giving off collaterals, and 
 then turns downward. This cell is rather 
 larger than the majority of the cells of 
 this layer. But other cells of the layer 
 are triangular, rectangular, or polygonal (q), 
 hence the latter name has been selected 
 as descriptive of them. Cajal says that 
 a stalk is often wanting in these cells, 
 but when present it varies greatly in its 
 direction, and never turns downward or 
 upward or terminates in the superficial 
 layer of the cortex. The neuraxons of 
 these cells turn downward into the white 
 substance. 
 
 FIG. 10. 
 
 FIG. 10. Diagram of the cells of the cerebral cortex. The cells are repro- All anatomists have distinguished this 
 
 duced from the plates. 
 
 7, superficial layer, a, fusiform, />, triangular, c, polygonal cells of Cajal. 
 
 7/, layer of small pyramids. d, smallest, e, small, / medium-sized pyramids with their neuraxons descending to the white matter, giving off 
 collaterals in their course. 
 
 III, layer of large pyramids, g, largest (giant) pyramidal cells, k, large pyramidal cell with very numerous dendrites. All pyramidal 
 cells are seen to send long apical processes up to /. m, Martinotti cell with descending dendrites and ascending neuraxon. , polygonal cells. 
 
 IV, deep layer. /, fusiform cell, q, polygonal cell. 
 
 V, the white matter containing the neuraxons from pyramidal cells, d, e, f, g, and from cell of the deep layer g. r, neuroglia fibre. 
 
* 
 
 FUSIFORM CELL OF THE FOURTH LAYER OF THE CORTEX; ITS LONG AXIS DIRECTED HORI- 
 ZONTALLY WITH NUMEROUS DENDRITES AND A NEURAXON PASSING TOWARD THE LEFT. X 190 DIAM. 
 
 PLATE L. 
 
THE CEREBRAL CORTEX 
 
 73 
 
 deeper layer of the cortex from the other layers, and have considered its cells as association 
 
 rather than projection cells. The fibres of 
 
 this layer are seen to pass in all possible 
 
 directions, intermingling with the neuraxons 
 
 of the other layers, which descend through 
 
 the layer to enter the white matter be- 
 neath. 
 
 An attempt has been made to put 
 
 together the various elements of the cor- 
 tex here described in a scheme or diagram 
 
 (Fig. 10) which may convey to the mind a 
 
 picture of their mutual relation. Each va- 
 riety of cell which has been described and 
 
 photographed in the plates is reproduced 
 
 and is given its proper size and position 
 
 relatively to the other cells. The course of 
 
 the fibres within the cortex is also shown, 
 
 with their probable termination so far as 
 
 actually determined. Fibres r are supposed 
 
 to be glia fibres forming a part of the 
 
 glia framework supporting the cortex. The 
 
 various systems of fibres entering and terminating in the cortex are not shown. 
 
 The neuraxons of all these cells are 
 shown, making up the white matter beneath 
 the gray, and giving off in their course 
 collaterals which pass in various directions. 
 The various neuraxons making up the 
 white matter beneath the cortex have been 
 classified by anatomists in accordance with 
 their final destination. 
 
 There are (i) projection fibres, which pass 
 through the centrum ovale into the internal 
 capsule and thence to some of the masses 
 of gray matter in the basal ganglia, me- 
 dulla, or spinal cord. These bring the 
 
 FIG. ii. 
 
 FIG. II. The projection tracts joining the cortex with lower nerve centres. 
 Sagittal section showing the arrangement of tracts in the internal capsule. 
 
 A, Tract from the frontal lobe to the pons, thence to the cerebellar 
 hemisphere of the opposite side. B, motor tract from the central convolu- 
 tions to the facial nucleus in the pons and to the spinal cord; its decussation 
 is indicated at K. C, sensory tract from posterior columns of the cord, 
 through the posterior part of the medulla, pons, crus, and capsule to the 
 parietal lobe. D, visual tract from the optic thalamus (OT) to the occipi- 
 tal lobe. E, auditory tract from the corp. quad. post, to which a tract 
 passes from the VIII. N.. nucleus (f) to the temporal lobe. F, superior 
 cerebellar peduncle. G, middle cerebellar peduncle. H, inferior cerebellar 
 peduncle. CA 7 , caudate nucleus. CQ, corpora quadrigemina. Vt, fourth 
 ventricle. The numerals refer to the cranial nerves. 
 
 FIG. 12. 
 
 FIG. 12. The association fibres. 
 
 A, Between adjacent convolutions. B, between frontal and occipital 
 areas. C, between frontal and temporal areas, cingulum. D, between frontal 
 and temporal areas, fasciculus uncinatus. E, between occipital and temporal 
 areas, fasciculus longitudinalis inferior. CN, caudate nucleus. OT, optic 
 thalamus. 
 
74 
 
 ATLAS OF NERVE CELLS 
 
 cortex into relation with the lower levels of the nervous system, the various parts of which 
 
 have been already studied. They are shown in Fig. n. 
 
 There are (2) association fibres, which pass from one part of the cortex to other parts, 
 
 thus bringing various regions of the cortex into mutual relation. They are shown in Fig. 12. 
 
 There are (3) commissural fibres, which 
 pass by way of the corpus callosum from 
 the cortex of one hemisphere of the brain 
 to the opposite hemisphere, to terminate 
 in the cortex. 
 
 It is not at all impossible, in the light 
 of recent discoveries, that a neuraxon from 
 a nerve cell may become a projection fibre 
 and give off collaterals, which may become 
 association or even commissural fibres. It is 
 certain that the human brain is superior 
 to that of all other brains in the number 
 of its association fibres which bring about 
 a perfect interaction of all its parts. The 
 termination of these association fibres ap- 
 pears to be in all of the various layers 
 of the cortex, thus bringing any cell into 
 relation with many cells of different situa- 
 tion and type. 
 
 It has been stated already that the 
 cortex is the terminal station of a very 
 large number of fibres which reach it from 
 other parts in the projection, association, 
 and commissural tracts. These fibres are not 
 
 Pol. P. 
 
 w. 
 
 shown in the diagram (Fig. 10), for the sake 
 
 FIG. 13. (Andriezen). Cortex of human brain showing the nerve fibre sys- 
 tems and plexuses (combined Weigert's and Golgi's methods). 
 
 ,*, clear zone (free of nerve fibre). M P, molecular plexus (Exner's), Qf dearnesS) but they are s ' nO wn in Fig. 1 3 
 in the molecular layer. A sir, ambiguous cell stratum. Subm P, sub-molec- 
 
 ular plexus. Gt PP, great pyramidal plexus. Pol P, polymorphic plexus. from a design by AndrieZCn ' which demon- 
 ic, white substance. 
 
 strates their complexity. They enter the 
 
 cortex from the white matter and terminate in brush-like expansions about the cells of all 
 the various layers. At some layers the intricate meshes of these fibres are very thick. The 
 
 1 Brain, Vol. XVII. p. 629; 1894. 
 
THE CEREBRAL CORTEX 
 
 75 
 
 thickest meshes are about the bodies of the largest pyramid cells and of the fusiform 
 cells of the deep layer. Many fibres ascend to the superficial layer, where a larger mesh 
 is shown to exist about the Cajal cells. There are terminal ends of fibres at all parts of 
 the cortex, and as Andriezen has pointed out these are non-medullated at their tips, where 
 they can come into contact with the dendrites and apical processes of the cortical cells. An 
 attempt has been made by Cajal l to trace the probable course of impulses through the 
 various cortical layers, but such an attempt is too hypothetical to be at present considered. 
 
 1 Cajal, Les nouvelles Idees sur la structure du Systfeme Nerveux, p. 66. Paris, 1894. 
 
76 ATLAS OF NERVE CELLS 
 
 THE HIPPOCAMPUS 
 
 The hippocampus is a peculiar structure which has attracted the attention of all anato- 
 mists who have studied the brain. It lies deep upon the base on the inner surface of the 
 tempero-sphenoidal lobe, its free inner surface projecting into the descending horn of the 
 ventricle. The peculiar appearance of the anterior portion known as Ammon's Horn is due 
 to a folding of the convolution upon itself in such a manner that the outer layer of one 
 portion of the cortex, being folded backward, comes in contact with the outer layer of another 
 portion of the cortex, and the free extremity is rolled outward in a sort of spiral curve. This 
 peculiar configuration is well shown in Plate LI. 
 
 It is not necessary to give a very detailed account of this structure, as its physiological 
 significance is not yet known. It is much more fully developed in the lower animals than 
 it is in man. It is easily possible to distinguish a number of layers in the cortex, and these 
 may be enumerated as follows (from right to left in the figure): 
 
 First. A layer of epithelium upon which lies the choroid plexus in the descending horn 
 of the lateral ventricle. This forms the outer boundary of the alveus. 
 
 Second. The second layer is the layer of white matter, really made up of the neuraxons 
 issuing from the base of the pyramidal cells, but deflected by the curving of the convolution 
 into a zone of horizontal fibres. This is known as the Alveus. 
 
 Third. A thin layer containing numerous polygonal cells which are similar to those of 
 the fourth layer of the cortex. This has been called the stratum oriens. 
 
 Fourth. A layer of pyramidal cells of various shapes and sizes having many dendrites and 
 long apical processes which divide and branch as they enter the fourth layer. This corresponds 
 to the third and second layers of other cortical regions fused together. 
 
 Fifth. A layer made up of numerous fine fibres passing in all directions, and of many 
 cells. This is really the superficial layer of the cortex, and is fused with a superficial layer 
 of the other part of the cortex, which is folded upon this layer; so that this broad layer 
 represents the union of two superficial cortical layers. It has been named the stratum lacu- 
 nosum. It contains many Cajal cells and cells of Golgi's second type. 
 
 Sixth. A layer of pyramidal cells with their bases directed away from the layer of fine 
 fibres ; thus lying in exactly an opposite direction to the cells of the fourth layer. 
 
 Seventh. A layer of white matter made up of the neuraxons coming out of the base of 
 these pyramidal cells. The fibres issuing from the seventh layer gather into a mass and 
 
A SECTION THROUGH THE HIPPOCAMPUS OF AN EIGHT MONTHS HUMAN EMBRYO, SHOWING 
 VARIOUS LAYERS OF THE CORTEX. X 29 DIAM. 
 
 PLATE LI. 
 

 THE CORTEX OF THE HIPPOCAMPUS. THE LONG PYRAMIDAL CELLS, THE SMALL PYRAMIDAL 
 CELLS AND THE INTERLACING LAYER OF FINE FIBRES ARE ALL EVIDENT. X 190 DIAM. 
 
 PLATE Lll. 
 
THE CEREBRAL CORTEX 77 
 
 sweep around the free extremity of gray matter, and thus arrive upon its surface, there passing 
 beneath the layer of epithelium. Hence the layer of white matter which we have described 
 as the second layer of the hippocampus is really made up not only of the neuraxons issuing 
 directly into it from the first set of pyramids described, but also contains fibres coming from 
 the second set of pyramids by a long spiral curve. 
 
 It is possible in the plate to follow the layer of pyramids all the way around the curve 
 of the cortical fold, and thus to become convinced that this layer of pyramids is really a single 
 layer folded upon itself, and hence necessarily lying with all the bases outward. 
 
 Plate LII. shows a section through the hippocampus at a higher power. At the bottom 
 of the plate the layer of white substance is seen, with a few fibres passing chiefly in a 
 horizontal direction. Above this layer the pyramids lie with their long apical processes, 
 similar in appearance to those which we have already studied. These apical processes are 
 seen to divide and enter the layer of fine fibres, where they interlace with the fine horizontal 
 fibres of the superficial layer of the cortex. These fine fibres are quite easily visible in 
 the plates. In the ordinary cortex the superficial layer is comparatively thin, and the apical 
 processes of the pyramids diverge and become parallel with the surface soon after their entrance 
 into it; but in this region it is evident that the layer of tangential fibres is broad, and that 
 there is less tendency of the apical processes to become horizontal. The interlacing of the 
 two sets of fibres in this layer is well demonstrated in the plate. In the plate a few small 
 cells can be seen in the superficial layer. These have been termed stellate cells by Cajal. 
 They have numerous dendrites which branch and interlace with the other fibres of this layer 
 adding to its complexity. Dejerine 1 describes other fibres entering and branching in this layer 
 which are terminal collaterals from the cells of the other layers. It will be noticed that no 
 distinction has been made between the layer of small pyramids and the layer of large pyra- 
 mids in this description of the hippocampus. No such distinction is possible, because the 
 cells are intermingled. 
 
 1 Anatomie des Centres Nerveux, p. 722. Paris. 1895. 
 
78 ATLAS OF NERVE CELLS 
 
 In the description of the cortex given above it has been described as consisting of four 
 layers. It is easily possible to distinguish everywhere in the cortex the four types of cells 
 described in these layers; viz. Cajal cells, small pyramidal cells, large pyramidal cells, polygonal 
 cells. But the relative number of these cells varies greatly in different regions of the brain, 
 and their arrangement in layers also varies. Thus few authors have agreed as to the num- 
 ber of layers of cells to be described, and it is erroneous to suppose that any single scheme 
 corresponds to all parts of the brain. Bevan Lewis has shown some of the differences exist- 
 ing in the cortical structure in different regions (see Quain's "Anatomy," loth edition). But the 
 most careful .study was made by the late Dr. Carl Hammarberg of Upsala, in " Studien iiber 
 Klinik und Pathologie der Idiotic nebst Untersuchungen iiber die Normale Anatomic der 
 Hirnrinde," and his plates are reproduced in Plate LI II. 
 
 Hammarberg's drawings of sections through* various parts of the cortex made to a scale 
 have proven that great variations in number and arrangement of the cells exist, that the frontal, 
 parietal, central, occipital, temporal, and median portions of the brain as well as the gyrus 
 fornicatus and the hippocampus have distinctive characteristics, and that no description of one 
 portion holds good for all. It is beyond the scope of this work to describe these variations, 
 which are rather in the mutual relation of such cells as have been shown than in any 
 peculiarities of cell structure. It has seemed necessary, however, to observe that the four- 
 layer type of structure is by no means universal, and should not be taken as typical of cor- 
 tical structure in man, however universal in lower animals. 
 
 Plate LI 1 1. shows sections from eight different regions of the adult human cortex, drawn 
 to a scale, and demonstrates the varying thickness of the cortex and the varying arrangement 
 of the cells in different parts. The first four sections are from the anterior regions of the 
 brain, and present the appearances supposed to characterize the motor cortex. They are 
 from the superior frontal (Fig. i); anterior central (Fig. 2); third frontal, posterior to the 
 ascending limb of the fissure of Sylvius (Fig. 3) ; third frontal, anterior to the ascending 
 limb of the fissure of Sylvius (Fig. 4). The second four sections are from the posterior regions of 
 the brain, and show the arrangement of cells prevailing in the sensory areas of the cortex. They 
 are from the posterior central (Fig. i); superior temporal (Fig. 2); superior parietal (Fig. 3); 
 and superior occipital (Fig. 4). It will be noticed that the larger pyramidal cells are found more 
 constantly in the motor region, and that the contrast between layers of large and small cells is 
 more apparent in the sensory region. It is as yet impossible to assign any separate functions to 
 the various kinds of cells. Hammarberg's studies, however, demonstrate conclusively that the 
 degree of mental power depends directly upon the number and perfection of development of these 
 cortical cells, for in the brains of idiots the cells are few in number and imperfectly developed. 
 

 , 
 
 I 
 
 1 
 
 * " 
 
 ' 
 
 y 
 
 i 1 
 
 i H: 
 
 j <* 
 4 i * 
 
 "i'" 
 
 
 
 
 
 t 
 
 1 
 
 
 
 i 
 
 ..^ 
 
 - 
 ;i 
 
 
 
 
 
 Si 
 
 
 
 q , 
 
 
 
 
 i 
 
 >- 
 
 y 
 
 F^j ^*'" 
 
 
 i 
 
 A * 
 
 B 
 
 jfn^p 3 * 
 
 xL" 
 
 r i- 
 k 
 
 
 * 
 
 
 , v 
 
 . 
 
 f 
 
 1 
 
 1 
 
 K 
 
 ^~ / 
 
 n 
 
 1 ; 
 
 
 
 
 i. . 
 
 , 
 
 - 
 
 -!* i^ 
 
 - 
 
 R '- 
 
 * 
 . 
 
 
 
 
 J, L '-- 
 
 
 J> T 
 
 5 * A 
 
 
 
 ' 
 
 
 
 
 
 I 
 
 
 i 
 
 ^ 
 
 
 
 v 3r 
 
 1 
 
 
 
 
 
 , I z: 
 
 - - 
 
 
 /i. , 
 
 
 W ^ 
 
 -.. . 
 
 
 
 
 
 
 : 
 
 
 IT 7 " z; 
 
 \ 
 
 ... 
 
 
 
 "T 
 
 
 ! y. i- !- 
 
 -* 
 
 
 4- 
 
 j *r*r 
 
 
 . . i 
 
 : ) 
 
 1 ^i 
 
 ;/ 
 
 7 ~ 
 
 ii i 
 
 
 
 
 
 
 T 
 
 1 
 
 
 2 x 
 
 * \ 
 
 
 
 
 
 4 
 
 
 ,*" A. 
 '' ^fefcj 
 
 
 
 
 
 
 
 
 
 ' 
 
 r 
 
 1 
 
 
 . 
 
 
 1 
 
 
 j 
 Lj 
 
 
 ^K 
 
 / 
 
 ; " 
 
 
 
 
 r v- 
 
 ' A 
 
 : 1 J * j_ 
 
 i 
 
 . 
 
 ' 
 
 ^ ff 
 
 H __ ^ T-T rv 
 
 
 
 ^ 
 
 i- 7* 
 
 f ' *-- 
 
 
 
 
 '4 
 
 - J 
 
 
 ' 
 
 < 
 
 
 .-*. 
 
 -J 
 
 'f ' '' 
 
 
 
 - 
 
 
 . 
 
 
 i 
 
 ^L 
 
 
 | ' 
 
 
 
 
 
 
 
 * I t- 
 
 
 
 u 
 
 
 
 . 
 
 
 f 
 
 =H 
 
 SECTIONS THROUGH EIGHT DIFFERENT PARTS OF 
 THE BRAIN CORTEX TO SHOW THE VARIOUS ARRAN- 
 GEMENT OF CELLS. 
 
 HAMMARBERG 
 
 PLATE Llll. 
 
 > 
 
UNIVERSITY OP CALIFORNIA L1BRAKY 
 BERKELEY 
 
 THK 
 
 demand may be renewed if 
 "i'> of loan period. 
 
 
 on ' made before 
 
 
 OCT 
 
U C BERKELEY LIBRARIES 
 
 Ilimillllllllllir 
 
 I!