Explanation of Plate XXXI (figs. 1-5):

'Fig. 1. Contents of experimental chamber that had remained 72 hours in the peritoneal cavity of the rabbit. Five large amoeboid plasma-cells, with altered red corpuscles and apparently dead leucocytes. Outlined with camera lucida. Apochromatic oil immersion and ocular No. 4. Zeiss. Prepared over osmic vapour.

Fig. 2. Contents of a chamber for 18 hours in the peritoneal cavity (rabbit); near the centre of the chamber. Fibrin filaments, leucocytes, red corpuscles, and an ill-defined granular mass forming a nodal point in the fibrinous network-the beginning of a "cell-islet." Outlined under camera. Similar method of preparation, and similar magnification to preceding.

Fig. 3. Fragment of inflammatory membrane formed within a chamber placed for three days in the subcutaneous tissue (guinea-pig). Islets and groups of islets scattered through the membrane. Zeiss, Obj. A, Oc. 2. Osmic acid solution, and Ehrlich's logwood.

Fig. 4. Contents of same chamber as in Fig 1. Close to the opening of the chamber. Five plasma-cells, one of them continuing a leucocyte within a large vacuole. Magnification and method of preparation as in Fig. 1.

Fig. 5. Contents of same chamber. Two plasma-cells and two red corptiscles; the plasma-cells are indistinguishably united with fine filaments of fibrin in their surrounding, some of which are given in the figure. Osmic acid vapour. Zeiss, apochr. system, oc. No. 2.' (575) 

Figs. 1 and 4-5 in text:

'The preparations gave an almxiost bewildering number of examples of the infinite variation in shape of the large amoeboid plasma-cells, which also varied very considerably in size, and as to granules. The body of the cell was for the most part plate-like, being in many instances extended into so thin a film that its exact limit was hard to determine, especially when, as occasionally happened, the granules of the cell-body were less pronounced towards the periphery. Some idea of the wide diversity of outline exhibited by individual cells may be gathered from our figures. Cf Figs. 1, 4, 5, 6, 7 and 8, Plates XXXI. and XXXII.' (558)

'In the specimens obtained from chambers that had rested for seventy-two hours in the subcutaneous tissue of the guinea-pig, we found individuals among the plasma-cells, which showed wellmarked vacuolation, Figs. 1, 4, 5, Pl. XXXI. For the most part the matter within the vacuole was a granular debris that furnished no sufficient clue as to its nature. But in a few it was indisputable that the vacuole contained, more or less altered but still perfectly easily recognisable, a leucocyte or red blood corpuscle. In Fig. 4 is shown the appearance presented by one of these cells. A large vacuole contains a somewhat faintly stained body, which is finely granular and indistinctly nucleated. It is a little smaller than is the nucleus of the plasma-cell itself. Fine threads seemed to pass from the sides of the vacuole across the cavity to the substance of the included leucocyte. Taken with the context afforded by examination of other cells in the neighbourhood we believe that this and other similar instances were examples of leucocytes lying in vacuoles in the plasma-cells. Many stages of ingestion could be found. Cf. Figs. 1, 4, 5, Pl. XXXI. Simple approximation, the hollowing out of a little bay in the side of the plasma-cell into which the leucocyte was as it were drawn, partial inclusion, total inclusion - all these were exemplified.' (559)

Figs. 1 and 5 in text:

'Contiguous plasma-cells or even those a little distance apart were often connected together by their processes (Figs. 1, 5, 7 and 8, Plates XXXI. and XXXII.). The bands of connection might be short thick arms or long gossamer threads of protoplasm. By similar arms and threads the cells seemed to adhere to the most diverse objects in their surrounding. The surface of the cover-glass, a filamllent of fibrin, a hair, a fibre of cotton, a lump of the cement fastening the sides of the chamber together, all afforded points to which the processes from the plasma-cells would cling (Figs. 14 and 15).' (560)

Fig. 1 in text:

'Here must be mentioned another sign of degeneration in the leucocytes examined in these chambers [note: 'Kuss, Paris, 1846. Paget, Surgical Pathol. p. 151.']. Many of them showed the triple and multiple nuclear bodies that are universally regarded as evidence of the lethal disintegration of the nucleus - as Fleming names it, the "fragmentation" of the nucleus. On the other hand the cell-body of the leucocyte was not granular or fatty, but fairly evenly though deeply tinted by the osmium. These points are seen in Fig. 1, Plate XXXI.' (556-557)

Figs. 2-3 in text:

'Eighteen Hours.- In chambers removed after the appearance of fibrin within them, but before the stay within the body had exceeded eight and forty hours, it was usual to find a number of areas in which leucocytes were present in much greater numbers than elsewhere. Fig. 2, Pl. XXXI.

The tendency to collect to certain points which the leucocytes evinced in even very early specimens was more marked in these later preparations. About the nodal points of the fibrinous network crowds of them were present. The ouitlying individuals were frequently arranged in lines along the converging filaments of fibrin. The older within certain limits these films of coagulum the more obvious the aggregation of the leucocytes into certain groups. For convenience on account of their prominence. and apparent importance in subsequent stages we have been accustomed to refer to these groups shortly as the cell-islets. Cf. Fig. 3, Pl. XXXI. They are little collections of cells, occurring constantly, scattered about in the thin cellular membranes which grow over and within the glass chambers.' (557)

Explanation of Plate XXXII (figs. 6-13):

'Fig. 6. Giant cells from chamber 72 hours in the peritoneal cavity (rabbit). Zeiss apochr. system, ocul. 5. Osmic acid vapour.

Fig. 7. Plasma-cells from same preparation which furnished Fig. 6.

Fig. 8. " Cell islet " from inflammatory film obtained in a chamber left eight days in the subcutaneous tissue of the guinea-pig, At the margin it is united to outlying plasma-cells. Zeiss oil, oc. 4. Osmic acid vapour.

Fig. 9. Young cicatricial tissue of anastomosing branched cells, some of which are represented under the higher magnification in Fig. 17. From a thrombosed artery (syphilis) near the centre of the thrombus. Zeiss A, oc. 3. Logwood. Preparation kindly shown us by Dr Seymour Sharkey.

Fig. 10. "Cell islet" from inflammatory membrane obtained from chamber five days in the peritoneal cavity of the rabbit. Osmic acid solution. Zeiss oil imm. and oc. 2.

Fig. 11. Mass of blood-cells (? clot) surrounded by fibroblastic cells, and invaded by them at four places. Inflammatory membrane from chamber eight days in subcutaneous tissue. Magnification as in preceding, and prepared in similar manner.

Fig. 12. Fusiform plasma-cell (fibroblast) surrounded by a fibrillated material which forms a thread-like band of connective tissue. Zeiss oil and oc. 4. Osmic vapour. From chamber 10 days in subcutaneous tissue.

Fig. 13. Similar but larger and thicker fibrous band from same preparation. Similar preparation and magnification.' (575-576)

Figs. 6-8 in text:

'The preparations gave an almxiost bewildering number of examples of the infinite variation in shape of the large amoeboid plasma-cells, which also varied very considerably in size, and as to granules. The body of the cell was for the most part plate-like, being in many instances extended into so thin a film that its exact limit was hard to determine, especially when, as occasionally happened, the granules of the cell-body were less pronounced towards the periphery. Some idea of the wide diversity of outline exhibited by individual cells may be gathered from our figures. Cf Figs. 1, 4, 5, 6, 7 and 8, Plates XXXI. and XXXII.' (558)

Fig. 6 in text:

'There were present also in chambers of eighteen hours', twentytwo hours', twenty-six hours', forty-eight hours', and seventy-two hours' standing, as also in others of older date containing well formed granulation tissue, many giant cells (Fig. 6) - huge multi-nucleate cells, that obviously in many instances were cell-fusions. ' (560)

Figs. 7-8 in text:

'Contiguous plasma-cells or even those a little distance apart were often connected together by their processes (Figs. 1, 5, 7 and 8, Plates XXXI. and XXXII.). The bands of connection might be short thick arms or long gossamer threads of protoplasm. By similar arms and threads the cells seemed to adhere to the most diverse objects in their surrounding. The surface of the cover-glass, a filamllent of fibrin, a hair, a fibre of cotton, a lump of the cement fastening the sides of the chamber together, all afforded points to which the processes from the plasma-cells would cling (Figs. 14 and 15).' (560)

Figs. 8-9 in text:

'In membranes of ten, fourteen, and even eighteen days' growth, not all the cells nor even the majority were spindle-shaped. A vast number were triradiate, and multiradiate; some had but one process; very few were rounded. Many recalled to mind the branched fixed corpuscles of the cornea. Long tapering branches united cell to cell, not only the cells of one plane one with another, but the cells of different planes also (Figs. 8, 9 and 17). A meshwork of infinite variety and complexity was thus established. But in all these examples of plasma cells in the stable as well as in the previously described labile forms, the granular nature of the cell substance and the clear oval nucleus were characters never lost.' (563)

Figs. 8 and 10 in text:

'in the specimens of more than forty-eight hours' duration, the plasma-cells begin to apply themselves to the islet-groups of leucocytes. Cf. Figs. 8 and 10. They surround the leucocytes. The islets come to consist of a central portion made up of leucocytes, and an outer zone of large and granular plasma-cells. In this way the islets seem to increase rapidly in size. Neighbouring islets appear to become merged together.' (561)

Fig. 11 in text:

'It was among the plasma-cells of the fringe of the islets that we noticed the earliest regularly fusiform cells, the immediate precursors of fibrous elements in the new tissue. It is true that plasma-cells of an irregular spindle-shape were observable not rarely among even the earliest of the plasma-cell swarm entering the chamber. But in those instances the outline was probably but one of many which the amoeboid cell successively assumed, and generally it was not of the same character as the regularly fusiform type prevailing among these plasma-cells in the outskirts of an islet. In that latter the majority of the cells lay in lines concentrically set about a core of ill-stained, broken-down matter that composed the centre of the mass. Cf. Fig. 11, Pl. XXXII. The fusiform fibroblasts began in fact the encapsulation of the débris of the breaking-down blood cells, &c.' (526)

Figs. 12-13 in text:

'Older specimens revealed further progress in the formation of a fibrous-tissue membrane. After a stay of eight days, or ten days, or fourteen days in the subcutaneous tissue in many instances the islets consisted of plasma-cells alone. The leucocytes had disappeared. The pigmented remnants of the red blood corpuscles were much longer traceable. In many places along certain lines the spindle-shaped cells had become attenuated, and formed distinct bands and often long and delicate cords (Figs. 12, 13). In many places in the tenth day specimens, and in some of the eighth day ones an inter-cellular substance showing fibrillation exists (Fig. 12).' (562)

Explanation of Plate XXXIII (figs. 14-19):

'Fig. 14. From chamber five days in subcutaneouis tissue. Plasma-cells adhering to a cotton-fibre. Osmic vapour and carmine. Zeiss apochlr. oil and oc. 4.

Fig. 15. From chamber eight days in subcutaneous tissue. Plasma-cells adhering to a hair, which had accidentally been allowed to get into the wound. Zeiss obj. D, oc. 2. Osmic acid solution and haematoxylin.

Fig. 16. Inflammatory membrane from chamber eight days in the abdominal cavity; taken from a tenuous portion of the membrane. Four fibroblasts, in a film which is composed of an extremely irregularly arranged network of filaments resembling fine fibrin threads. The processes from the cell-body are continuous apparently with the fibrils of the matrix. Osmic acid vapour and haematoxylin. Zeiss apochr. oil imm. and oc. 4. Outlined with camera lucida.

Fig. 17. Stellate fibroblasts and two leucocytes from same preparation as Fig. 9, more highly magnified. Zeiss apochr. oil and oc. 4. Outlined with camera lucida.

Fig. 18. The modified Ziegler chamber; the sketch shows the actual size employed.

Fig. 19. Portion of the chamber seen edgewise, showing the opening between the cover-glasses. Enlarged 12 times.' (576) 

Figs. 14-15 n text:

'Contiguous plasma-cells or even those a little distance apart were often connected together by their processes (Figs. 1, 5, 7 and 8, Plates XXXI. and XXXII.). The bands of connection might be short thick arms or long gossamer threads of protoplasm. By similar arms and threads the cells seemed to adhere to the most diverse objects in their surrounding. The surface of the cover-glass, a filamllent of fibrin, a hair, a fibre of cotton, a lump of the cement fastening the sides of the chamber together, all afforded points to which the processes from the plasma-cells would cling (Figs. 14 and 15).' (560)

Fig. 16 in text:

'Each individual cell was of a discoid or fusiform figure, and granular, with a large clear nucleus. The edge of the disc was thin and often deeply scalloped; it merged, under all methods of staining used by us, at certain points quite imperceptibly, in a tenuous film which composed the bulk of the membrane proper. When fixed with osmic acid and after-stained with haematoxylin (Ehrlich's), this membrane is shown to contain, if not to be entirely made up of, a feltwork of filaments, like filaments of fibrin. These cross in every direction in the plane of the membrane, without prominent arrangement in any one particular sense. The individual filaments vary a good deal in size. Fig. 16, P1. XXXIII.' (561-562)

Fig. 18 in text:

'Two circular cover-glasses, each 5/8 of an in. in diameter and .006 of an in. in thickness, were fastened together so as to form a little flat glass chamber, in the manner employed by Ziegler. A strip of tinfoil placed between them at their edge along 11/12 of their circumference was cemented by shellac on each face to the corresponding surface of the cover-glass. The tiny chamber thus formed had therefore between the two ends of the strip of tin-foil an opening into the interior. The tin-foil first employed was 1/10 mm. thick; that thickness was inconvenient, as the depth of the chamber was then too great for higher powers of the microscope to explore. Tin-foil 1/20 mm. in thickness was subsequently employed. With this thickness membranes were obtained between the cover-glasses that made very satisfactory microscopical specimens. Fig. 18, Plate XXXIII.' (552-553)

Explanation of Plate VIII:

'Fig. 1. Sub-maxillary gland of rabbit. Fresh state. After a moderate amount of saliva from pilocarpin injection. The ductule- and transition-cells are darkly granular; the alveolar-cells shew an outer clear zone.

Fig. 2. Sub-maxillary gland of rabbit. Few hours after feeding. Osmic acid 1 per cent. acid, 2 1/2 hours, absolute alcohol 20 hours, mounted in dilute glycerine. The darker patches mark the transition- and ductule-cells.

Figs. 3, 4 and 5. Infra-orbital gland of rabbit. Treated with osmic acid and subsequently with alcohol. In figs. 4 and 5 the granules are rather too large; in the actual specimens, they appear more as darker spaces in a lighter network. In figs. 3 and 4 the nuclei are rather too prominent.

Fig. 5. Resting gland.

Fig. 4. Gland after moderate amount of secretion from pilocarpin.

Fig. 3. Gland after protracted secretion from pilocarpin, sympathetic also stimulated.

The actual appearance of the specimens cannot be very accurately represented by the lithographic process; the shaded portions of the figs. including the nuclei should be perfectly smooth and homogeneous.' (279-280)

Fig. 1 in text:

'Both in hunger and digestion, the appearance of this [sub-maxillary] gland [in the rabbit] in the fresh state is less constant than in the parotid... Nevertheless, as the result of many observations, I feel confident that the number of granules in the alveolar-cells diminishes during activity just as it does in the parotid. When the cells at the beginning of an experiment were granular throughout, a thinly granular periphery, or a non-granular outer zone, made its appearance, after continued secretion from pilocarpin injection, or sympathetic nerve stimulation; when the cells at the beginning of an experiment had an outer clear zone, the clear zone became larger. (P1. VIII. Fig. 1.) The alveolar-cell granules are less highly refractive and somewhat smaller than those of the parotid.' (269)

'The sub-maxillary has... one very characteristic point, the transition-, and at any rate some of the ductule-cells are crowded with granules much larger than those of the alveolar-cells (Pl. VIII. Fig. 1).' (270)

Fig. 2 in text:

'if, after two hours treatment with osmic acid, the gland is washed with dilute spirit or water and placed in alcohol about 75 per cent. for twenty-four hours and then sections cut, the appearances are markedly different; scattered about, are dark, deeply stained patches which at once catch the eye (Pl. VIII. Fig. 2); these are the transition-cells and ductules; there is now a much greater equality of staining between these and the ducts; the transition-cells may be somewhat lighter than the ducts and the ductules somewhat lighter than either, all being much darker than the alveolar cells. The nuclei of the alveolar-cells are much more conspicuous.' (271)

'Nussbaum [note: 'Arch. f. Mik. Anat., Bd. xvi. s. 543, 1879.'] suggests that the transition-cells which I describe are different from the darkly-stained cells described by him; a comparison of the figure 1 of his first communication with mine [note: 'op. cit.' ['Journal of Physiology, Vol. I. p. 69, 1878.']], and of the figure 1 of his second communication [note: 'Arch. f. Mik. Anat., Bd. xv. s. 119, 1878.'] with my fig. 2, Plate VIII., adjoined to this Paper will, I think, shew that we are dealing with the same cells.' (272-273)

Figs. 3-4 in text:

'During activity, the outer portion of each alveolus begins to stain evenly (Fig. 4), at first without much alteration in the nuclei or in the inner portions of the cells. Later, the nuclei become larger, spherical, and travel towards the centres of the cells; they are then much less distinct; as the outer zone encroaches on the inner zone, the network look of the latter becomes less and less apparent, so that it may be represented only as a few scattered dots (Fig,. 3). At the same time the lumen becomes more obvious and stretches out somewhat between the cells.' (275)

Fig. 5 in text:

'In the resting gland the alveoli are throughout unevenly stained (Fig. 5, Pl. VIII.). The nucleus is irregular, and lies in the peripheral portion of the cell. When looked at with a high power the network appears ligrht and the spaces dark, as if the network were the substance between the granules which are seen with a lower power; ordinarily the granules seen under a low power are described as being the nodal points of the network seen with a high power. At present, however, I am not prepared to discuss either this point or that of the relations of the granules normally seen with those visible after reagents.' (274)

Explanatio nof Plate XXX (figs. 1-8):

'Fig. 1. Alveoli of fresh sub-maxillary gland of dog, irrigated with 5 p.c. NaCl., showing mucous granules streaming out of an alveolus (a).

(b) Bleb of swollen, transparent mucin with granules in it.

(c) Other alveoli with mucous granules.

(d) Demilune cells with smaller, fainter proteid granules. Magnified about 480 times.

Fig. 2. Alveolus of sub-maxillary gland of dog, showing the pseudonetwork which is sometimes seen a short time after a piece of gland has been mounted.

Fig. 3. Alveolus of orbital gland of cat, in 5 p.c. NaCl, slightly pressed, showing the limits of the cells as clear lines.

Fig. 4. Isolated mucous granules of the orbital gland of the dog.

(a) Swollen in 1 p.c. NaCl.

(b) In 5 p.c. NaCl, showing faint filament sometimes seen to join and connect the granules, probably in consequence of their having stuck together.

(c) Shrunken in 20 p.c. NaCl.

Fig. 5. Isolated cells from sub-maxillary gland of dog, a day after death.

(a) Cells in 5 p.c. NaCl; two of the cells are focussed to show the nuclei, the nucleus is often completely hidden by the granules.

(b) Cells isolated in 5 p,c. NaCl, and then irrigated with 1 p.c. HCI.

(c) Demilunes in 5 p.c. NaCl.

Fig. 6

(a) Mucous cells sub-maxillary gland of dog irrigated with water, showing different sized network in the cells according to the amount they swell up; the network is here represented as more distinct than it usually is on addition of water. Nuclei rather refractive and homogeneous in appearance.

(b) Mucous cell of sub-maxillary gland of dog, isolated in 5 p.c. NaCl, and then irrigated with 1 p.c. HCl, showing fat globules in the meshes of the network. 

(c) Mucous cell of sub-maxillary gland of cat, irrigated first with dilute ammonia, then with 1 p.c. HCI, showing bleb of transparent mucin issuing from the cell.

(d) Mucous cell of orbital gland of dog, two days after death. Isolated in 5 p.c. NaCl. and irrigated with 1 p.c. HCI, showing bulging of mucous substance between the bars of the network.

(e) Mucous cells from orbital gland of cat, one day after death. Isolated in 2 p.c. NaCl, a few granules only are left in the cells.

(f) Isolated nuclei of mucous cells of sub-maxillary gland of dog.

Fig. 7. Cells from orbital gland of dog left for three days in 2 p.c. chloral hydrate.

(a) Cell irrigated HNO3 5 p.c., showing wide-meshed inner network.

(b) Cell irrigated with water, showinig fine-meshed network on the surface of the cell.

Fig. 8. Ducts of sub-maxillary gland of the dog.

(a) Without addition of fluid the granules are not always so distinct as they are represented here, but may become more distinct on irrigation with 5 p.c. NaCl.

(b) Showing alteration which may take place a short time after the piece of gland has been mounted. Early stage of striation.

(c) Irrigation with water, granules swell up and give a vacuolated appearance to the duct cells.' (456-457)

Fig. 1 in text:

'On irrigating a slightly teased fragment of a fresh gland with NaCl 5 p.c., it is seen that the contents of the alveoli as they stream out inito the fluid, form first a clear bleb, closely beset with granules (cf. Fig. 1); then the bleb swells and the granules become arranged in lines parallel to the surface of the bleb.' (437)

Fig. 2 in text:

'Some of the alveoli usually undergo changes different from that described above. In some, the granules become less and less distinct, and the substance between them becomes more and more distinct, so that the cells appear at first sight to contain a network with small meshes, viz. 8 to 12 lumen to basemnent membrane (Fig. 2, Pl. XXX.). On closer examination the apparent network can here and there be seen to be simply the optical section of the small amouint of substance existing between the granules, but in most cases it is not sufficiently distinct to justify any conclusion with regard to it.' (435-436)

Fig. 4 in text:

'On continuing the irrigation with 5 p.c. NaCl, the groups and rows of granules are gradually washed away from the edge of the specimen... A delicate connecting filament can sometimes be observed between two granules (cf. Fig. 4 (b)) as they are being rolled over. Since no such connecting filament between the granules can be seen inside the cells, it probably indicates that the granules here and there stick together where they come in contact.' (438)

'In 20 p.c. and in stronger solutions of sodium chloride, sometimes also in weaker, the granules are apt to be elongated or irregular in shape instead of being spherical (cf. Fig. 4 (c)); they are more apt also to cling together in small groups; in some of these groups the separate granules may be barely distinguishable.' (438)

Figs. 5 and 7 in text:

'The outline of the cells on irrigation with dilute acids, is in optical section more or less obviously beaded (Fig. 5 (b) and Fig. 7 (a)); there are usually twelve to fifteen smrrall swellings in the length of the cell, in some cases the outline of the cell appears to be discontinuouis in optical section, the limiting layer of the cell looking as if it were a perforated membrane. Similar small swellings have been mentioned by Schiefferdecker as occurring in cells treated with Muller's fluid. Now and then, and especially after treatment with chloral hydrate, the beading can be followed into a small-meshed network in the limiting layer of the cell (Fig. 7 (b)).' (444)

Fig. 5 in text:

'A mucous cell isolated in 2 to 5 p.c. sodium chloride appears as a mass of highly refractive spherical granules held together by a barely visible cell-substance (cf. Fig. 5 (a)); it is more or less columnar, never pear-shaped or globular, as it is after isolation in chloral hydrate; it does not show a cell-process; the nucleus lies close to the basal end of the cell.' (439)

'Demilune cells isolated in salt solution are more or less distinctly granular, the granules being one-half to one-third the size of those in the mucous cells (Fig. 5 (c)). lhe cells are usually in groups of two or more, the cell-outlines are very indistinct and often not visible, they may however be brought out by adding acetic acid and ferrocyanide of potassium and by various other reagents.' (440)

'An isolated mucous cell on irrigation with one of the abovementioned reagents swells up, becoming pear-shaped or globular... As the granules disappear from the cells a network comes into view, and the cell-outlines become sharply contoured. In the isolated cells the network has wide meshes, the number of meshes being as a rule 4 to 6 in the length of the cell (Fig. 5 (b)); the meshes are not quite equal in size, and the fineness of the fibres and the size of the nodal points varies slightly with the strength and nature of the reagent. In cells which are not isolated, the meshes of the network are generally smaller and less distinct, there may be 9 to 12 in the length of the cell.

Fig. 5 (b) shows the changes produced in the isolated mucous cells, Fig. 5 (a) by the addition of 1 p.c. hydrochloric acid. A similar change is produced by many other reagents'. (443)

Figs. 6-7 in text:

'The scattered proteid particles spoken of above stain deeply with methylene blue and with most other colouring matters. A not inconsiderable number of them are spherical; in isolated cells-especially of the orbital gland of the dog treated with hydrochloric acid-similar small proteid granules are seen attached to or forming part of the network (Figs. 6 (d), 7 (a)).' (445)

Fig. 6 in text:

'In the glands of the dog I have generally found the granules numerous for two days after death; but in one or two observations on the orbital gland of the cat, whilst the cells isolated, immediately after death, in 5 p.c. NaCl were densely granular throughout, the cells isolated two days after death varied widely in granularity, some of them containing three or four granules only, scattered in the transparent mass of the cell; these cells thouah less angular than those in which the granules were preserved were not much swollen, but their outline was doubly contoured, differing thus from the cells in which the granules were for the most part unaltered (cf. Fig. 6 (e)).' (439-440)

'Notwithstanding the very considerable swelling of mucin which takes place in the isolated cells on irrigation with HCI 1 p.c., it is remarkable how seldom a plug of mucin is seen to project from the free end of the cell. I have only seen this in the fresh gland, when a piece has been irrigated first with dilute ammonia and then witlh hydrochloric acid (cf. Fig. 6 (c)).' (445-446)

'When a gland has been kept for one day to two days in NaCl 5 p.c. the cell-network and the limiting layer of the cell are not brought out so distinctly by acids. The limiting layer is the first to become indistinct; in this stage the cells of the orbital gland not infrequently show a network between the peripheral meshes of which the clear mucous substance of the cell bulges (Fig. 6 (d)).' (452)

Fig. 7 in text:

'Chloral hydrate 2 to 5 p.c. Irrigated with this, the mucous granules swell up and become very pale; the cells swell and the granules in them cannot be distinguished. This fluid is commonly used for obtaining isolated cells; when a small piece of gland is placed in it the outermost cells burst, and their contents form a ropy mass. This remains for several days, but is gradually changed into an irregular granular deposit. Isolated cells are best obtained in about three days; the cells are then much swollen, most show with greater or less distinctness a wide-meshed network in their interior (Fig. 7 (a)); it may be made still more distinct by irrigating the specimen with 1 p.c. osmic acid; sometimes the meshes near the nucleus are smaller and the bars of the network thicker than elsewhere. In some of the mucous cells generally those in which the mucin is most swollen, and the inner network indistinct - there is a very marked fine-meshed network in the limiting layer of the cell (Fig. 7 (b)); the meshes vary somewhat in size; they are usually five or six sided. In such cells the limiting layer of the cell, apart from the network, is extremely faint; here and there fibres run from this network to the wide-meshed internal network. I use the phrase network in the limiting layer of the cell simply as descriptive of the appearances under the microscope, I shall later discuss whether it really is a network, or whether it consists of ridges on the limiting layer caused by the pressure exercised by the granules as they swell. ' (450)

Fig. 8 in text:

'The ducts of the lobules in the fresh gland have a slightly yellowish tint, this is much more marked in the sub-maxillary than in the orbital gland. The outer and larger portion of the ducts, occasionally almost homogeneous, is usually finely and indistinctly granular (cf. Fig. 8 (a), Pl. XXX.). The striation which is so obvious after most methods of hardening is not seen in the fresh gland; but it usually becomes more or less apparent when the specimen is kept. The granules for a time become more obvious, then they swell, and as they do so, beconme paler, arranged in rows, and often elongated, at the same time the cellsubstance between them becomes more distinct; owing to these two changes the ducts acquire a striated appearance (cf. Fig. 8 (b)). Occasionally the duct cells, besides containing small fat globules chiefly in the neighbourhood of the nucleus, contain also large fat globules about the size of the nucleus; these are not uncommon in the submaxillary gland of the cat.' (436)

'The granules of the duct cells become larger and much paler, so that the ducts often look densely vacuolated, the cell-substance between the apparent vacuoles becomes very distinct (Fig. 8 (c)).' (445)

Explanation of Plate XXIX (figs. 9-12):

'Fig. 9. Section of the liver of an Embryo Cat, 15 mm. in length. (x 400 diam.)

a. Capillary blood space.

b. Endothelial wall of capillary blood-vessel.

c. A liver tubule.

d. Lunmen of a liver tubule cut transversely.

Fig. 10. Portion of an osmic acid preparation of the liver of the Rat. (x about 400 diam.)

a. Spaces occupied by the network of blood capillaries cut longitudinally.

b. Similar spaces cut transversely.

c. Transverse section of a bile capillary lying between two liver cells.

d. A similar bile capillary lying at the angle where three liver cells meet.

Fig. 11. A similar preparation of the liver of the Mole. (x about 400 diam.)

a. Spaces for blood capillaries.

b. Section of a bile capillary lying between four cells. A sinmilar one is seen near it, lying between three cells.

Fig. 12. Portion of a section of a cochineal stained specimen of the liver of the Pig, from the peripheral part of a lobule. (x about 400 diam.)

a. Nucleated wall of a capillary blood-vessel.

b. Liver cells showing a collection of granules where two or three cells are contiguous.' (427-428)

Figs. 9 and 12 in text:

'There is nothing in the known history of the development of the liver which opposes the idea that the gland may have arisen from a solid mass of hypoblast cells formed at the extremity of the primitive hepatic diverticulum and which growing and forcing its way into the surrounding mesoblast, becomes itself interpenetrated by ingrowing mesoblast cells. By the development of these latter into blood capillaries the hypoblastic cell mass would become broken up into a network of solid anastomosing rods, in which secretion channels would subsequiently form, and which would be more or less fine according to the completeness with which the ingrowth of developing blood-vessels took place...

An examination of the liver of the types we have described in the light of our hypothesis will show that all the appearances can be explained by it. We can find no other satisfactory interpretation of the arrangement seen in the liver of the lamprey (Plate XXVII. Fig. 1). A comparison of a section of the eel's liver with that of a mammal (Plate XXVII. Fig. 2 and Plate XXIX, Fig. 12) shows clearly that a more intimate subdivision by capillaries in the case of the latter would account for the differences, and at the same time it is seen that the relative magnitude of the blood capillaries in the two cases is very different - those of the eel can scarcely be called "capillaries". That this is the true difference in the two cases is clear from a comparison of a section of the liver of an embryo mammal with that of the adult and with that of the eel (Plate XXVII. Fig. 2, and Plate XXIX., Figs. 9 and 12). On examining a section of the liver of the newt, the appearances seen, making allowance for the different sizes of the cells and smaller details, are more like those of the mammal's liver than that of any other of our types - (Plate XXVIII. Fig. 6) - There has been in this case a penetration of the mass of liver cells sufficiently intimate to have left but three or four rows of cells to form the tubules, and at the same time the total vascularity of the organ, as shown by the size of the blood spaces, is not so great as in the case of the eel or the mammal.' (423-424)

Explanation of Plate XXVII (figs. 1-4):

'Fig. 1. Portion of a section of the liver of the Lamprey. (x about 400 diam.)

a. Blood capillary with nucleated endothelial wall.

b. Blood capillary cut transversely. Around it is seen a radial arrangement of the elongated liver cells.

c. Solid hepatic cylinder with loosely arranged cells.

Fig. 2. Part of a section of the liver of the Eel. ( x 400 diam.)

a. Transverse section of a larger vessel (probably radicle of hepatic vein).

b. Liver tubules cut transversely.

c. Lumen of a liver tubule.

d. Liver tubule in longitudinal section.

e. Nucleated endothelial wall of a capillary blood space. 

Fig. 3. Section of an osmic acid preparation of the liver of the Frog. (x 400 diam.)

a. Blood corpuscles lying in a capillary vessel.

b. Outer lightly stained zone of the cells of the liver tubules.

c. Inner black-stained granular zoite around the lumen of the tubules.

Fig. 4. Section of a logwood stained preparation of the liver of the Frog. (x about 400 diam.)

a. Sections of the network of spaces for blood capillaries.

b. Liver tubules in transverse sectioln.

c. Lumina of liver tubules cut longitudinally.' (426-427)

Fig. 1 in text:

'Our first impression was, that there were in this liver large lumina with cells grouped round them though not in a single layer; but further examination made it clear that these supposed lumina are really blood channels and that between these blood channels the liver cells are disposed in solid anastomosing cylinders, the central cells of which are loosely arranged, with intercellular spaces occupying what may be viewed as the lumen of a potential tube. (Plate XXVII. Fig. 1.) The relative proportion of cell-cylinder to blood-vessel is extremely large, and between any two blood-vessels the number of layers of cells varies from two to five or six. The cells of which the intervascular cylinders are composed naturally fall into two classes, distinguished by their form and arrangement;- (a) those immediately adjacent to the blood-vessels, which have the form and arrangement of a columnar epithelium disposed radially round the blood channels, and each measuring 5µ in width and about 15µ in length, (b) those situated within the cell cylinders, which are generally polyhedral in form, some few being slightly elongated and each measuring about 5µ in all diameters. The latter are disposed in a sponay manner and in hardened specimens show channels running between them.' (414)

Figs. 1-2 in text:

'There is nothing in the knownw history of the development of the liver which opposes the idea that the gland may have arisen from a solid mass of hypoblast cells formed at the extremity of the primitive hepatic diverticulum and which growing and forcing its way into the surrounding mesoblast, becomes itself interpenetrated by ingrowing mesoblast cells. By the development of these latter into blood capillaries the hypoblastic cell mass would become broken up into a network of solid anastomosing rods, in which secretion channels would subsequiently form, and which would be more or less fine according to the completeness with which the ingrowth of developing blood-vessels took place...

An examination of the liver of the types we have described in the light of our hypothesis will show that all the appearances can be explained by it. We can find no other satisfactory interpretation of the arrangement seen in the liver of the lamprey (Plate XXVII. Fig. 1). A comparison of a section of the eel's liver with that of a mammal (Plate XXVII. Fig. 2 and Plate XXIX, Fig. 12) shows clearly that a more intimate subdivision by capillaries in the case of the latter would account for the differences, and at the same time it is seen that the relative magnitude of the blood capillaries in the two cases is very different - those of the eel can scarcely be called "capillaries". That this is the true difference in the two cases is clear from a comparison of a section of the liver of an embryo mammal with that of the adult and with that of the eel (Plate XXVII. Fig. 2, and Plate XXIX., Figs. 9 and 12). On examining a section of the liver of the newt, the appearances seen, making allowance for the different sizes of the cells and smaller details, are more like those of the mammal's liver than that of any other of our types - (Plate XXVIII. Fig. 6) - There has been in this case a penetration of the mass of liver cells sufficiently intimate to have left but three or four rows of cells to form the tubules, and at the same time the total vascularity of the organ, as shown by the size of the blood spaces, is not so great as in the case of the eel or the mammal.' (423-424)

Fig. 2 in text:

'The cells of the eel's liver are of medium size, and are granular throughout with a condensation of granules towards the lumen, which is distinct though small. The number of rows of cells forming the wall of the tubule is generally five or six, but in some cases is four. The diameter of the tubule averages 25µ, but is not quite uniform, (Plate XXVII. Fig. 2).' (412)

Figs. 3-4 in text:

'The cells of the frog's liver are large and clear. They have granules on their inner borders, next the lumen, which feature is brought out very clearly by treatment with osmic acid, but is also visible in good logwood-stained specimens, one of which is figured in Plate XXVII. Fig. 4. Langley [note: 'Langley, loc. cit.'] states that in summer a hungry frog has granules equally scattered throughout the cells of its liver, and that during winter there is a marked inner granular zone. We can confirm this, and the specimen figured in Plate XXVII. Fig. 3, is from a winter frog, which on account of the distribution of the granules is the one best suited for tracing a lumen. In an osmic acid preparation the anastomosing liver tubules are rendered very plain by the darker staining of the blood in the capillaries, and they have the appearance of a lightly stained mass with the division into cells not very obvious, but with a well marked inner zone of black granules. In a logwood or cochineal stained specimen the number of rows of cells is seen to be usually five.' (415)

Explanation of Plate VII, figs. 1-9:

'Fig. 1. Diagram of the Cardiac apparatus and gills of Octopus, with the cardiac and respiratory nerves.
Vc. = Vena Cava. KY. = Kidney Vein. BH.= Branchial Heart. G. = Gill. LA. = Left auricle. RA. = Right auricle. V. = Ventricle. CA. = Cephalic Aorta. Ga. = Genital artery. Va. = Visceral artery. SF. = Veins from sides. P. = Pleural ganglia. S. = Stellate ganglia. Vg. = Visceral nerve. C. = 1st Cardiac Ganglion. Cr. = 2nd Cardiac Ganglion. B. = Branchial Ganglion.

The dotted lines indicate the parts covered by glands.

Fig. 2. a. Muscle fibre of the ventricle of Octopus. (x 500.) Alcohol and Hematoxylin.

Fig. 2. b. Portion of a muscle fibre of the ventricle of Sepiola. (x 500.) Gold chloride.

Fig. 3. a. and 3. b. Plasma cells from the julletion of the kidney and auricle of Pterotrachea. (x 500.) Osmic acid and Picrocarmine.

Fig. 4. Ganglion-cell from the cerebral ganglia of Pterotrachea. (x 500.) Osmic acid and Picrocarmine.

Fig. 5. Plasma cell from the cutis of Pterotrachea. (x 500.) Osmic acid and Picrocarmine.

Fig. 6. Connective-tissue cells; a. from the auricle, b. from the cutis of Pterotrachea. (x 500.) Osmic acid and Picrocarmine.

Fig. 7. Diagram of the branchial and cardiac nerves of Aplysia.

LV., RV. = Left and Right Visceral nerves. Vg. = Visceral ganglion. Gen. = Genital nerve. O. = "Olfactory organ." rn. = Left branch of the nerve to the gill. nb. = Branch to the pericardium. G. = Gill. P. = Pericardium. A. = Auricle. V. = Ventricle. a. = Aorta.

Fig. 8. a. Plasma cell on auricular muscle of Helix.

Fig. 8. b. Plasma cell from the connective tissue membrane between the loops of the genital duct of Helix. (Zeiss. D. Oc. 4.)

Fig. 9. Diagram of the cardiac nerves of Helix.

SO. = Supracesophageal ganglia. Sub. O. = Subcesophageal ganglia. Lv. = Left visceral nerve. Os. = Ovisperm duct. K. = Kidney. P. = Pericardium. A. = Auricle. V. = Ventricle. a. = Aorta.' (338-339)

Fig. 1 in text:

'The species chosen as most convenient for study was the common Poulp, Octopus vulgaris, and to it the following pages will be almost exclusively confined.

The cardiac apparatus of Octopus (Fig. 1, Plate VII.) consists of a pre-branchial or venous, and a post-branchial or arterial section. The former collects the blood from the veins and drives it through the gills; the latter receives the aerated blood fromn the gills and propels it through the bodv. The former section may be divided into the following parts:-

Vena Cava.

Kidney Veinis.

Branchial Hearts.

The latter consists of :-

Auricles.

Ventricles.

The Vena Cava (Fig. 1. VC.) is a straight thlin-walled tube, which runs down the ventral side of the liver, parallel to the intestine, towards the heart.

At the level of the front end of the ventricle it divides into two branches, which being covered with glands considered to have a renal function, may be conveniently distinguished as the Kidney Veins (KV.)

Each Kidney Vein, curving out towards a gill, is continued into a bulged oval body known as the Branchial Heart (BH.), their cavities being separated by a pair of valves at the junction. The Branchial Hearts have been usually held to be muscular organs devoted to propelling the blood through the gills; but that this view is not entirely correct will be shewn by the description of their minute structure given below. From each Branchial Heart a thick-walled "branchial artery" leads to the gill (G). The efferent vessels of each gill form a number of veins which soon unite to form a tuLbular auricle (LA. RA.), thin-walled but stouter than the Vena Cava. The two auricles open into the medianly situated, fleshy, somewhat globular ventricle (V), from the cavity of which the blood is prevented from returning by a pair of valves at each auriculo-ventricuilar orifice. 

Three efferent vessels issue from the ventricle, the most important being the great cephalic aorta (CA.), which starting from the posterior dorsal right region of the ventricle curls round to run to the head. Medianly and posteriorly is given off a small genital artery (G a.); and from the anterior ventral edge a small bulb gives rise laterally to a fine artery to each auricle and to a somewlhat larger visceral artery (V a.) medianly to the intestine.' (263-264)

'The whole cardiac system is supplied by a pair of nerves from the pleural ganglia in thie head... Their relations to the neighbouring organs and directions for finding them have been accurately given by Fredericq... Their branches and ganglia need however a more detailed description (Fig. 1). On issuing from the skull, each nerve appears as a double cord from which various small branches are given off. The cords on either side end in a small ganglion lying under the liver, from which go nerves to the body wall and the columnar muscle inserted into the mantle, while a single main trunk is continued downwards towards the heart. Just in front of the auricle it dilates into a ganglion, which may be called the 1st Cardiac Ganglion (C). From this issue a fine nerve to the generative duct, a nerve which enters the auricle, gives off a branch there and then passes through to the ventricle, and lastly a stout nerve which runs dorsally to the auricle down to the branchial heart, where it is connected with a ganglion-the 2nd Cardiac Ganglion (Cr). From this ganglion go nerves to the substance of the branchial heart and a short way into the kidney vein, but the main trunk proceeds to the gill, at the base of which it expands into the Branchial Ganglion (B). No other ganglia are revealed by dissection on the visceral nerves or their branches...

From the pleural ganglia there also runs on either side a stout nerve ending in tlle stellate ganglion (s) and containing the motor fibres for the mantle; so that the gills and heart are connected by nervous structures with the motor organs of respiration.' (267-268)

Fig. 2 in text:

'The fibres of the ventricle are of an elongated fusiform shape, with long tails, but are shorter and thicker than those of the auricle. Like the others they have no sarcolemma. Thus far they resemble the ordinary smooth muscle cells of vertebrata, but in addition they possess a fine but regular transverse striation very like that of the branchial heart, the fibres of which they closely resemble. Careful focussing also shows in the fibre indications of a granular core of a different nature to the outer zone. Both of these features are best seen in osmic acid or alcohol preparations (Fig. 2a) [note: This axial column of apparently less differentiated substance has been noticed by most observers of molluscan muscles. Thus Ranvier (Traite technique d'Histologie, p. 851), figures a gold chloride preparation of the retractor muscle of Helix pomatia showing distinctly the "cordon protoplasmique"; Leydig and Kölliker saw it in the buccal muscles of Gasteropoda, and H. Muller and F. Boll in the branchial hearts of Cephalopoda. (See F. Boll, Arch. f. mikr. Anat. Bd. v. Suppl. Hft. p. 28).'].' (266)

'while the ventricular fibres of Eledone resemble those of Octopus, those of Sepia present a bolder striation, and in Sepiola the difference is still greater. In this animal the fibres are much larger, and the average distance between the striations is 3.3m.; the dark bands are very sharp, but narrow; and the margin of the fibre is bulged out opposite each broad clear disc (Fig. 2. b). The central core is much more distinct than in Octopus, occupying about one third of the diameter of the fibre, and staining very deeply with gold chloride, while the outer zone remains clear. It also shows a beaded outline corresponding with the striations. [note: 'Transverse striation has also been noticed - chiefly in hearts or the buccal mass - by many observers, a list of whom is given by Boll (loc. cit.) and by F. Darwin (Journ. of Anat. a. Phys. Vol. x. Part Ii. April 1876, p. 506). Dogiel (Arch. f. mikr. Anat. Bd. 14) has also seen it in various hearts, (Pecten, Anodon, Aplysia, Helix) and Haller (Gegenbaur's Jahrbuch, 1883) in the hearts of Fissurella and Haliotis. Margo (Wien. Sitzusngsb, Bd. 39) has described the black bands in the striated shell muscle of Anodon as doubly refracting.'].' (266 [note 266-267])

Figs. 3-5 in text:

'Two kinds of elements however occur [in Pterotrachea] which might be mistaken for nerve cells, and which I believe have been taken for such in other Molluscs. The first of these are large, moderately granular, often roundish or oval cells, with a not very distinct oval nucleus placed near the periphery of the cell, and in which a nucleolus is sometimes visible. A distinct capsule is not present. These cells occur in great abundance on the walls of the rhythmically contractile kidney, and a few are usually to be found in the auricle [note: 'Although in teased preparations of auricle these cells were often found, I am inclined to doubt whether they occur normally in the auricle itself, and to think that in these cases they have appeared from adhering fragments of the closely connected kidney, which owing to the transparency of the tissues it is difficult to cut clearly off at the boundary. In preparations where the auricle was cut through at a distance from the kidney no such cells were found. At the same time it is possible that they may occasionally wander from the kidney to the auricle. They are also found in the pericardial wall.']. I have not observed them in the ventricle. Although the approximately oval form (Fig. 3 (a) Pl.VII.) is perhaps the commonest shape, yet it is by no means constant. In many cells short blunt processes or pseudo-podia are seen (b) which suggest the power of amoeboid movement, and occasionally a single such process may occur, causing some resemblance to a ganglion cell with the stump of a nerve fibre. The number and variations of the processes however oppose this idea, and a comparison of such a cell with a true ganglion cell (Fig. 4) at once shews a number of differences in appearance. Further these cells are never found in connection by fibres with either nerve or muscle, and are shewn to be very loosely applied to the tissues by the fact that in teased preparations large numbers of them become entirely detached and float like the blood corpuscles in the fluid in which they are mounted. Lastly, they occur not only in the parts mentioned, but in the connective tissue all over the body, as is well seen in preparations of the cutis, where they (Fig. 5) abound. These facts point to the conclusion that they are wandering connective tissue cells; and we may almost certainly identify them with the "plasma cells" which have been shewn by Brock [note: 'J. Brock. " Untersuch. ii. d. interstitiellen Bindesubstafiz der Mollusken." Zeit. f. Wios. Zool. Bd. 39. 1883.'] to be so characteristic of the connective tissue of some other Molluscs (Aplysiadae, Pulmonatae).' (321-322)

Fig. 6 in text:

'The second form of element which might be considered nervous [in Pterotrachea] is of much smaller size and more delicate structure. These cells again I found only where connective tissue was abundant. They are granular branched cells with a usually indistinct nucleus. Occasionally the branches may be reduced to two, and there then appears some resemblance to a small bipolar ganglion cell (Fig. 6 (a)). These cells however are nothing more than the ordinary connective tissue cells, which give off processes to form fibres and are found everywhere throughout the connective tissue (Fig. 6 (b)). A comparison of the two figures makes plain the identity of those of the heart with those of the cutis. These also are described and figured by Brock in the connective tissue of Aplysia as stellate connective tissue cells, and very similar cells by Haller as occurring 'in small quantity in the auricle of Fissurella and as forming a second type of ganglion cell.' (323)

Fig. 7 in text:

'Apart from the improbability of the existence of apolar ganglion-cells in the heart, the incorrectness of Dogiel's drawing of the heart and the adjacent nerves causes great doubt as to the accuracy of his interpretation of these cells.

In his figure (Taf. V. a. Fig. 15) of the "branchial" ("visceral," Spengel) ganalion he entirely overlooks the olfactory organ to which the right nerve runs, and he represents the left nerve as of equal thickness and as running straight to the auricle. But instead of such a short thick left nerve, there come off from the left half of the double "visceral " ganglion two distinct long nerves. Of these one supplies the generative duct, while the other runs on to the gill. Here it divides into two, the left one running on straight towards the hind end of the gill while the right curls round anteriorly and appears to end in the pericardium near the origin of the auricle (Fig. 7).' (325)

Fig. 8 in text:

'In the auricle of the Snail may be found upon the muscle bundles [note: 'Both the auricle and ventricle are formed of a meshwork of bundles of striated muscle fibres, that of the ventricle being denser and thicker.'] a number of plasma cells which are, there is little doubt, identical with Dogiel's "apolar!" ganglion cells. Some of these cells of an oval or pear-shaped form present a considerable resemblance to ganglion cells, and if only these were observed might be considered such. But when others of varying shapes, and others in process of division are seen, such an idea becomes untenable; and when the striking similarity of these cells with the plasma cells of the connective tissue is noticed, no doubt can remain of the identity of the two (Fig. 8).' The various stages of division also which are met with, from a group of two or three contiguous cells to a nest of closely packed small ones, point clearly to the true nature of these elements. True ganglion cells I believe do not exist in the Snail's heart.' (327)

Fig. 9 in text:

'The heart... receives nerves both at the auricular and aortic ends [note: 'A similar double nerve supply has been described by Haller, loc. cit. as existing in Muricidae and Fissurella.']. (Fig. 9.) The discovery of this nerve to the heart at once necessitated a reconsideration of the results obtained by previous workers, and the question of its function became important.' (328)

Explanation of Plate XVII (figs. 1-4):

'The sections are all carefully drawn from osmic acid preparations of the nerves of man.

Fig. 1. Transverse section of rootlet of IIIrd cranial nerve to show degenerated ganglion. (Zeiss A, Oc. 4.)

Fig. 2. Three transverse sections of IVth cranial nerve. (Zeiss A, Oc. 2):

a. Section near exit of nerve from valve of Vieussens, showing degenerated tissue arranged so as to form a sheath around the functional medullated nerve-fibres.

b. Section of nerve farther away from point of exit, showing the formation of the degenerated galglion.

c. Section of nerve peripheral to the ganglion. The degenerated tissue has almost entirely disappeared.

Fig. 3. Transverse section of VlIth cranial nerve to show degenerated ganglion. (Zeiss A, Oc. 2.)

Fig. 4. Longitudinal section of a rootlet of IlIrd cranial nerve to show the structure of the degenerated material. (Zeiss D, Oc. 2.).' (209)

Figs. 1-2 in text:

'In Pl. XVII, Fig. 2, I give three sections of the IVth nerve selected out of the whole series. The section (2a) is through the root of the nerve close to its exit from the valve of Vieussens, and it shows that the nerve fibres are surrounded by a thick sheath of peculiar connective-tissue-like material; while section (2c) taken more towards the periphery shows that this sheath of peculiar tissue no longer exists. An examination of the sections between 2a and 2c shows that the manner of disappearance of this thickened sheath is very peculiar, it does not remain on the outside of the nerve fibres but forms roundish masses of the same -peculiar material in between the nerve fibres themselves as is seen in (2b), so that the sections with a low power possess a remarkable resemblance to a section through a ganglion. At the peripheral side of this ganglion- like formation these round masses of fibrillar-like material cease somiewhat abruptly and the sections appear as in 2c. Again in the IlIrd nerve the same formation is seen; here we have a nerve composed of a great numnber of rootlets and the sections show that the ganglion-like structure makes its appearance on the separate rootlets before they join together to form the conjoint root. Very striking indeed is the sight of a number of these rootlets each presenting in a greater or less degree the appearance represented in Fig. 1, PI. XVII. Here too we see another point of resemblance between this structure and a spinal ganglion. In the latter it is of common occurrence to find that the formation of the ganglion has thrown many nerve fibres out of their previously parallel course, so that on transverse section of the ganglion many fibres are cut in a more or less longitudinal direction; the same deviation is found, as is seen in Fig. 1, in many of the large, fibres of the IIIrd nerve, which are thrown out of their direct course in consequence of the formation of these rounded fibrillar masses in a way precisely similar to what occurs in the spinal ganglion.' (167)

Fig. 2 in text:

'In the Elasmobranch, as for instance the dog-fish... we find in the infundibular region a membranous sac known as the saccus vasculosus, the membranous roof of the 4th ventricle is more extensive than in the mammal, the choroid plexuses are very conspicuous as is shown in the diagram Fig. 2, P1. XVIII.' (194)

Explanation of Plate IV (figs. 1-8):

'I am indebted to the kindness of Mr G. Turner for the figures of this plate. Figures 2, 3, 4, 5 were drawn from photographs taken by him of the microscopic sections. 

Fig. 1 shows the connection between the posterior root ganglia and the sympathetic chain in the tortoise. Sy, sympathetic chain; the numbers 2-7 are placed against the corresponding thoracic nerves. The line l shows the direction of the series of parallel sections taken through the roots, ramus communicans, and root ganglion of the 5th thoracic nerve.

Figs. 2, 3, 4, 5 are copies of 4 sections out of this series and show the formation of ganglion cells on the sympathetic nerves, a, b, c, and the ultimate amalgamation of these ganglion cells with those of the root ganglion.

Fig. 6. Section of that branch of the 2nd thoracic nerve which forms the communicating branch to the 1st thoracic nerve and also helps to form the ramus visceralis of the ganglion stellatum. (From a photograph taken with lens 1/3 in.)

V. Small fibred portion of nerve which becomes the internal branch or ramus visceralis in connection with the ganglion stellatum.

A. Large fibred portion which becomes the external branch or branch of communication with the brachial plexus.

Fig. 7. Section of spinal accessory close to the ganglion jugulare vagi. (From a photograph taken with lens 1 in.)

V. Small fibred portion which becomes the internal branch or ramus visceralis in connection with the ganglion trunci vagi.

A. Large fibred portion which becomes the external branch and communicates with the cervical plexus.

Fig. 8. Upper part of spinal cord of dog cut in two in order to show the arrangement of the upper cervical nerve roots and those of the neural segment immediately above them.

1, 2, 3. Anterior and posterior roots of the corresponding cervical nerves.

V. Vagus nerve.

H. Hypoglossal nerve.

Ac. Spinal accessory nerve.' (79-80).

Fig. 1 in text:

'The connection of the fibres of the ramus visceralis with the cells of the ganglion on the posterior roots of the thoracic nerves is most clearly visible in the case of the tortoise. In this animal the ramus visceralis does not spring from the ventral branch of the spinal nerve as in mammalia, but arises directly from the ganglion on the posterior root. In Fig. 1, PI. IV., I give a representation of the connection between the sympathetic ganglia and the thoracic nerves from the 2nd to the 6th inclusive. As is seen, the ramus communicans directly connects each posterior root ganglion with its corresponding sympathetic ganglion. Owing to the inisertion of the lurmbo-caudalis muscle the rami communicantes of the 4th, 5th, 6th thoracic nerves at first pass out of the posterior root gangrlion in the direction of the roots of the nerves, and then turn over the muscle to reach the sympathetic chain.' (61-62)

Figs. 1-5 in text:

'In consequence of this arrangement, as is seen in Fig. 1, Pl. IV., a series of sections made for the purpose of following the spinal roots into the ganglion will at the same time enable us to trace the ramus communicans into the ganglion. In Figs. 2, 3, 4, 5, Pl. IV., I give drawings made from photographs of four sections taken from a consecutive series through the roots and spinal ganglion of the 4th thoracic nerve after staining with osmic acid. The figures show clearly how ganglion cells are formed in the ramus visceralis independently of those formed round the spinal roots, how these sympathetic ganglion cells increase in number as the posterior root ganglion is approached, and, finally, how they pass into and are lost in among the cells of the posterior root ganglion itself. Such a series makes it certain that in this case some of the nerve cells of the posterior root ganglion are connected with the fibres of the ramus visceralis.' (62)

Fig. 6 in text:

'The white ramus communicans of the 10th spinal nerve (2nd thoracic) is large and is composed of many separate bundles; the largest of these bundles arises directly-from this small branch of communication with the brachial plexus, and not from the main part of the 10th nerve. In Fig. 6, Pl. IV., I give a section of this communicating branch after it has left the main stem of the 10th (2nd thoracic) nerve. As is seen, it is composed of two parts, the one, A, chiefly containing large medullated fibres; the other, V, chiefly containing very fine medullated; as we trace the series of sections further and further from the origin of the nerve, we find the two portions separate more and more from each other'. (60)

Fig. 7 in text:

'In Pl. IV. Fig. 7, I give a picture of a section of the spinal accessory just before it reaches the ganglion jugulare of the vagus. The roots of the medullary portion of the accessory and of the vaaus were carefully hardened in situ with osmic acid, the whole with a portion of the medulla oblongata removed, imbedded in paraffin, and "ribbons" of consecutive sections made through the whole of the nerve roots from the medulla oblongata up to and beyond the ganglion trunci vagi. The whole series of sections was mounted in order, every nerve fibre was well stained and remained on the slide in the exact position it occupied when imbedded. In Pl. II. Fig. 10, the arrangement of the fibres when imbedded is reproduced.

The figure shows that the spinal accessory is divided into two distinct portions, the one (A) composed of large medullated fibres with a few isolated medium sized ones among them, the whole of that portion being remarkably free from connective tissue; the other portion (V) composed mainly of the smallest medullated fibres among which are a few large ones imbedded in a conspicuous matrix of connective tissue.' (10)

Fig. 8 in text:

'An examination of the relative positions of the roots of the nerve group immediately above the 1st cervical nerve (see Fig. 8, Pl. IV.) demonstrates the presence of two sets of rootlets which correspond in position to the anterior and posterior rootlets of the 1st cervical nerve, the only difference being that they are not so directly opposite to each other: a slight shifting of position which is easily accounted for by the alteration in the direction of the spinal axis due to the opening out of the central canal, and the formation of the 4th ventricle.' (74)

Explanation of Plate XII (figs. 1-11):

'The figures with one exception were drawn under a zeiss D and ocular 4. The outlines were filled in with the Camera Lucida and an attempt made to colour the drawings like the particular specimen under examination.

Fig. 1. To show the first segmentation of the myeline in degeneration. Each segment completely enclosed in myeline, with a central piece of the axis cylinder. Experiment XX. 2 days. R. S. middle of forearm. Osmic acid and haematoxylin.

Fig. 2. To show the secondary fragmentation of the myeline (extreme case). Exp. XX. 2 days. L S. middle of forearm. Osmic acid and haematoxylin.

Fig. 3. To show the remnants of the disintegrating axis in the segments of myeline. Exp. XIII. 4 days. R. S. 3 in. below wound. Gold and haematoxylin. Nucleus stained blue in haematoxylin.

Fig. 4. The same. Exp. IIl. 7 days. R. S. Below wound 2 in. Picric acid and fuchsin stain.

Fig. 5. The same. Exp. III. 7 days. R. S. Below wound 2 in. Picric acid and fuchsin stain.

Fig. 6. To show the more rapid fragmentation of the myeline in the neighbourhood of the nuclei. Exp. III. 7 days. Gold and haematoxylin. Nucleus stained blue.

Fig. 7. To show the multiplication and migration of the nuclei of the sheath and greater absorption near nuclei. Exp. III. 7 days. Gold anld haematoxylin.

Fig. 8. The same. Exp. III. 7 days. Gold and haematoxylin. Fig. 9. An instance of apparent indirect division of nucleus. The progressing fragmentation of myeline. Exp. III. 7 days. Gold and haematoxylin.

Fig. 10. (Drawn under obj. C. oc. 4). To show fragmentation of myeline especially at nuclei, and the material containing fragments of myeline which fills the fibre. Exp. III. 7 days. Gold and haematoxylin.

Fig. 11. To show progressive absorptionl of myeline; the proliferation of the nuclei; the position of the nuclei with reference to the balls of myeline; and the collapse of the empty sheath. Exp. XI. 9 days. R. S. middle of forearm. Gold and haematoxylin.' (402)

Fig. 1 in text:

'The segments first formed in degeneration, unlike the normal appearance of the segments of Lantermann seen in fibres after the action of osmic acid, have a complete envelope of the myeline, as shown in Fig. 1, and the small interspaces between them are filled with a colourless material. In each of these segments is a piece of the original axis cylinder, at this time apparently unchanged. Many of the older histologists have asserted that the axis remains unaffected during the degeneration. On the contrary, even at this early stage the breaking of the myeline into segments is accompanied by, or causes a simultaneous breaking of the axis.' (371)

Figs. 2-8 in text:

'Shortly after the first cylindrical segments of myeline and axis are formed an irreaular fragmentation occurs in these segments in most parts of the fibre. The segmetnts break up into smaller or larger irregular pieces or balls, an extreme case of which is shown in Fig. 2. Very frequently in these irregular masses of myeline remnants of the axis cylinder can be clearly distinguished in all stages of disintegration as shown in Figs. 3, 4, 5. It should be stated that some of the large segments frequently persist in different portions of a fibre long after tlle other segments have broken into small pieces and become partially absorbed,-the process is very irregular. It is at the time that this breaking up of the cylindrical segments into smaller fragments becomes apparent that the increase in the size of the nuclei of the sheath and the growth of the protoplasm surrounding these rnuclei become clearly marked. In fact this secondary fragmentation is always visible first in the neighbourhood of the nuclei as shown in Figs. 6, 7, 8. The large rounded nuclei lie in the middle of the fibre, and close to them are the small drops or balls of myeline. For this reason we believe that one cause of this secondary fragmentation is to be found in the absorption which takes place under the influence of the nucleus and its suirrounding protoplasm.' (372)

Fig. 6 in text:

'As evidence of the actual absorption which is beginning to take place one finds, at this time, say the seventh day, in the fibres stained with osmic acid or gold, some of the balls of myeline in the neighbourhood of the nuclei left colourless by the stain, see Fig. 6. In the later stages the same fact may frequently be noticed at different parts of the fibre and always most clearly near the nuclei. See Fig. 11.' (372-373)

Figs. 7-8 in text:

'An interesting fact in connection with the inultiplication of the nuclei is the way in which they migrate. In the beginning, of course, there is a single nucleus to each internode. At the time the secondary fragmentation of the myeline is fully under way one frequently finds a number of nuclei in the space which an internode would occupy. Sometimes they are in pairs as though from a recent division, but in other cases one or more large masses of myeline will be found between. See Figs. 7 and 8. This latter appearance has been used to support the theory that there are several nuclei present in the internode in a normal fibre, but that they are hidden by the myeline. Such a view as this it is not necessary to consider at present. The only explanation of the appearance described that seems reasonable is that after division the nuclei migrate or may migrate to some distance and start the process of absorption at a new place.' (373)

Fig. 9 in text:

'By the 7th day a very active proliferation of the nuclei of the sheath has begun. The increase in their number is very striking. Our methods of hardening were not such as to show the method of multiplication of the nuclei, though we often found nuclei showing a dumb-bell form, i.e., an elongated nucleus constricted in the middle as though multiplying by direct division. There can be little doubt that the division is indirect, as we sometimes found, even after hardening in Mueller's solution, nuclei like that pictured in Fig. 9 which evidently represents a badly preserved mitosis.' (

Figs. 10-11 in text:

'From the 7th day to the 14th day the process of absorption of the balls of myeline with their contained fragments of axis cylinder goes on actively, yet quite irregularly. Fig. 10 from a nerve after 7 days shows very well the breaking-up of the myeline and the formation in between, especially at the nuclei, of an apparently liquid substance in which are contained numerous fragments of the old myeline. As the absorption progresses and the fragments of myeline become smaller and less numerous the direct participation of the nuclei in the process becomes more evident. The nuclei are much more numerous and are found clustered in and about the remaining balls of myeline as shown in Figs. 11 and 12, representing a degeneration of 9 and 14 days respectively. One often sees bits of the myeline partially imbedded in a nucleus, and this appearance is found from this time on well into the later stages of regeneration, as long, in fact, as any of the myeline remiains unabsorbed. After 14 days absorption has gone so far that long stretches may be found, as shown in Figs. 12 and 13, in which only small fragments of myeline are present. At such places the fibre consists of a homogeneous, apparently liquid substance lying in the old sheath, and of many nuclei, often in pairs or groups, the latter giving indication of an active proliferation. Yet at this time, 14 days, and even later, one sees many fibres in which the absorption has lagged behind the condition of what may be considered a typical fibre of this period. In one and the same fibre places will be found in which absorption has made rapid progress in spots, all the myeline having disappeared, while in other spots the large cylindrical segments have suffered scarcely any change. Examples of this are pictured in Figs. 14 and 15. However irregularly the process may go on the final outcome is the complete absorption of the remnanits of the old myeline and axis; though as we have said before, balls of the myeline may be found in certain fibres long after this period, even at the time when fully formed new fibres have beeu produced. As the absorption proceeds the old sheath collapses more or less. It seems at first to contain a liquid material with some debris of the old myeline, but this too finally disappears and the beginning of the actual process of regeneration is inaugurated by the formation of new protoplasmic material around the numerous nuclei contained in the fibres.' (373-374)

Explanation of Plate XIV (figs. 29-42):

'Fig. 29. Cross-section of human ulnar peripheral end 6 1/2 months after section, union with central end not made, To show cross-section of bundles of embryonic fibres and the appearance of .the nuclei (only one bundle filled in). Gold and haematoxylin.

Fig. 30. The same. Stained in haematoxylin alone.

Fig. 31. Cross-section of pelipheral ulnar 4 weeks after division. After suture but before complete regeneration. To show the resemblance of the nuclei and their nucleoli to cross-sections of young nerve fibres. Exp. XVII. 4 weeks. L. S. Gold and haematoxylin.

Figs. 32, 33. Embryonic fibres in peripheral ulnar of rabbit, 7 weeks after section without union to central end. To show apparent attempt at the formation of myeline. Exp. XXVI. Osmic acid and haematoxylin.

Figs. 34, 35. To show the apparent beginning of a myeline transformation in the protoplasm of the embryonic fibres. Exp. XXVIII. 3 weeks. From branch to flexor profundus. Osmic acid and haematoxylin.

Figs. 36, 37. The first, to show the isolated drops of newly-formed myeline, the second, to show the varicose tube formed by their union. (Portions of the same fibre.) Exp. XXVIII. 3 weeks. Just below wound. Gold and haematoxylin.

Fig. 38. To show the isolated drops of newly-formed myeline and the processes which unite them. Exp. XXVIII. 3 weeks. Just below wound. Osmic acid and haematoxylin.

Figs. 39 and 40. To show the same and the staining of the balls of myeline in the haematoxylin. Each of the fibres shows a few fragments of old myeline unabsorbed lying near the nuclei. Exp. XXVIII. 3 weeks. Just below wound. Osmic acid and haematoxylin.

Figs. 41 and 41a. The same. To show discontinuous formation of the myeline tube. Exp. XXVIII. Below wound. Osmic acid and haematoxylin.

Fig. 42. The same. Exp. XXVIII. Below wound. Osmic acid and haematoxylin.' (403-404)

Figs. 29-31 in text:

'Our longest experiments upon a nerve cut but not sutured, and in which reunion was prevented by removing a piece of suitable length, were experiment XXVI., upon a rabbit examined 70 days after the operation, and XXII., central ulnar of dog, 75 days. In addition, however, we have had the opportunity of examining a portion of the peripheral end of a human ulnar nerve which had been cut accidentally six and a-half months previously. We owe the opportunity to the kindness of Dr T. A. Mc Graw, who in an operation for secondary suiture removed small portions of the central and peripheral ends and sent them to us in Mueller's solution. Figs. 25, 26, 27, 28 are from drawings of specirmens prepared from the peripheral end of this nerve after treatment with the potash gold stain of Freud and Böhmer's haematoxylin. They show that regeneration had advanced to the stage of the formation of embryonxic fibres, but that there were no signs of an axis cylinder or myeline sheath. Similar preparations were obtained from the animals upon which section without suture was made. Cross sections of the peripheral end of the human nerve are shown in Fig. 29, stained with gold and haematoxylin and in Fig. 30, stained in haematoxylin alone. The drawing in Fig. 29 was made from a portion of the nerve in which some of the nuclei show a central nucleolus in cross section, simulating the appearance of. new nerve fibres with a central axis. This is better shown in Fig. 31, taken from the ulnar nerve of a dog after sutture but before regeneration.' (376-377)

Figs. 35-41 in text:

'Whatever theory of the immediate cause of the formation of the myeline may be the true one, there can be no doubt that it is first found as disconnected drops. These may afterwards become united by slender processes to form a bead-like string which sooner or later elongates to an even tube, or the drops mnay first elongate to form cylindrical segments which eventually unite to form continuous, delicate tubes of myeline. Both of these processes, with the intermediate stages, are shown in Figs. 36 to 55 better than we could describe them in words. What becomes of the numerous nuclei scattered along the embryonic fibre it is not possible to say, other than that they disappear. In some cases, as in Figs. 38, 39, 42, they disappear rapidly as the myeline tube is formed, while in other cases they persist for a much longer time, the newly-formed myeline tube winding in and out among them in a very beautiful manner (see Figs. 37, 51, 55). We can only suppose that they disappear by absorption, as their nutritive relations with reference to the protoplasmic contents of the fibre become less and less important. With reference to the nodes and internodes of Ranvier, it is evident that no simple hypothesis, such as the development of each internode from a single cell, will fit the facts as they appear in regenerating fibres. The developing internodes and nodes are plainly shown in Figs. 51 to 55, but why the ends of the internodal tubes do not fuse together is difficult to explain.' (379-380)

Figs. 36-41 in text:

'After the formation of the embryonic fibres the additional steps necessary for a complete regeneration consist in the production of a new myeline sheath and a new axis cylinder... Figs. 36, 37, 39, 40 and 41 give an idea of how the myeline first appears. As shown in these figures, it appears first as irregular deposits in the protoplasm of the embryonic fibres, and usually first in the immediate neighbourhood of the nuclei. Delicate prolongations of the myeline are often seen running from one small mass of the myeline to another, and eventually these latter become connected together, forming a varicose tube, shown in various stages in Figs. 36 to 51. With reference to the immediate origin of these deposits, we must confess ourselves in doubt. In many cases we obtained specimens such as are shown in Figs. 34 and 35, in which clear bead-like spaces are seen forming in the protoplasm. The general arrangement of these spaces strongly suggests that they mark the beginning of the transformation of the protoplasm to myeline; a step farther and the chemical change will be such as to give the usual staining with osmic acid or gold as seen in Figs. 36-38, etc. On the other hand, the first deposits of myeline, especially when stained in osmic acid, often bear a striking resemblance to the nuclei, both in shape and in the fact that they show a rather distinct staining with the haematoxylin. (See Figs. 38, 39, 40, 42, etc.) There is no doubt that as the myeline forms many of the numerous nuclei scattered along the fibre disappear.' (378-379)

Explanation of Plate XV (figs. 42-55):

'Fig, 43. To show varicose appearance of newly-formed tube of myeline. Exp, XXVIII. 3/4 in. below wound. Osmic acid and haematoxylin.

Fig. 44. The same. Exp. XXII. R. S. near wrist. Osmic acid and haematoxylin.

Fig. 46. To show the formation of the myeline tube without varicosities. Exp. XXVIII. 3/4 in. below wound. Osmnic acid and haematoxylin.

Fig. 47. To show union of isolated drops of myeline to form a tube. Exp. XXVIII. Osmic acid and haematoxylin.

Fig. 48. To show discontinuous formation of myeline sheath. Exp. XXVIII. Osmic acid and haematoxylin.

Fig. 49, The same. Exp. XXVIII. Osmic acid and haematoxylin.

Fig. 50. To show the persistence of balls of old myeline as yet unabsorbed and their position with reference to the nuclei. Exp. XXVIII. Osmic acid and haematoxylin. Nuclei stained in haematoxylin.

Fig. 51. To show the formation of a node of Ranvier. Exp. XXVIII. Osmic acid and haematoxylin.

Fig. 52. To show the newly-formed myeline tube lying in the embryonic fibre. Exp. IX. 3 weeks. From dorsal cutaneous branch. Gold and haematoxylin.

Fig. 53. The same. Exp. IX. L. S. Dorsal cutaneous branch. Gold and haematoxylin.

Fig. 54. To show the formation of nodes of Ranvier. Exp. IX. L. S. Dorsal cutaneous branch. Gold and haematoxylin.

Fig. 55. To sbow newly-formed myeline tube with node of Ranvier and persistent nuclei. Exp. IX. L. S. 3 weeks. At wound. Gold and haematoxylin.' (404-405)

Figs. 42-55 in text:

'Whatever theory of the immediate cause of the formation of the myeline may be the true one, there can be no doubt that it is first found as disconnected drops. These may afterwards become united by slender processes to form a bead-like string which sooner or later elongates to an even tube, or the drops mnay first elongate to form cylindrical segments which eventually unite to form continuous, delicate tubes of myeline. Both of these processes, with the intermediate stages, are shown in Figs. 36 to 55 better than we could describe them in words. What becomes of the numerous nuclei scattered along the embryonic fibre it is not possible to say, other than that they disappear. In some cases, as in Figs. 38, 39, 42, they disappear rapidly as the myeline tube is formed, while in other cases they persist for a much longer time, the newly-formed myeline tube winding in and out among them in a very beautiful manner (see Figs. 37, 51, 55). We can only suppose that they disappear by absorption, as their nutritive relations with reference to the protoplasmic contents of the fibre become less and less important. With reference to the nodes and internodes of Ranvier, it is evident that no simple hypothesis, such as the development of each internode from a single cell, will fit the facts as they appear in regenerating fibres. The developing internodes and nodes are plainly shown in Figs. 51 to 55, but why the ends of the internodal tubes do not fuse together is difficult to explain.' (379-380)

Fig. 42 in text:

'the first deposits of myeline, especially when stained in osmic acid, often bear a striking resemblance to the nuclei, both in shape and in the fact that they show a rather distinct staining with the haematoxylin. (See Figs. 38, 39, 40, 42, etc.) There is no doubt that as the myeline forms many of the numerous nuclei scattered along the fibre disappear.' (379)

Explanation of Plate XVI (figs. 56-69):

Figs. 56, 57, 58, 59. From nerve stained in picric acid atid haematoxylin. Exp. XIX. 7 weeks. R. S. J in. below wound. 59. An embryonic fibre. 58. Newly-formed fibres with myeline and axis. 57 and 56. Newly forming fibres, myeline as a varicose tube. Axis also present but not stained in the interior of the swellings of myeline. Shows how quickly after the formation of the myeline the axis grows down from above.

Fig. 60. To show the branching of the axis cylinder where an old fibre passes into several new fibres, central end of ulnar. Exp. XXlI. Cut for 75 days and union with peripheral end prevented. Picric acid and haematoxylin.

Fig. 61. To show what seems to be the outgrowth of an axis cylinder from an old fibre toward a new one. New myeline lies in the embryonic fibre as a continuous delicate sheath. Exp. XVII. 4 weeks. R. S. Gold and haematoxyiin.

Fig. 62. To show the formation of several new fibres in the degenerated portion of a single old fibre at the central end of wound. Exp. XVII. 4 weeks. L. S. wound of median ulnar suture. Osmic acid and haematoxylin.

Fig. 63. The sanme. Human ulnar nerve. Central end of nerve 6 1/2 months after injury. No union with peripheral end. Gold and haematoxylin.

Fig. 64. To show junction of old with new fibre at the central end. Exp. XVIII. R. S. wound. Osmic acid and haematoxylin.

Fig. 65. The same. Human ulnar. Central end 6 1/2 months after section and no union with peripheral end. The continuation of the protoplasm of the "embryonic fibre" into the hypertrophied protoplasm surrounding the nuclei of the old fibre, and the fragmentation of the end of the old myeline to be specially noted. Gold and haematoxylin.

Fig. 66. To show junction of old and new fibres, and the new fibre with its thin myeline sheath lying in the protoplasm of the "embryonic fibre." Exp. XVII. R. S. wound. Gold and haematoxylin.

Fig. 67. To show junction of old and new fibres. Central end of human ulnar, 6 1/2 months after injury and no union with peripheral end. The fragmentation of the myeline and the hypertrophy of the protoplasm round the nuclei of the old fibre to be noted. Gold and haematoxylin.

Fig. 68. The same.

Fig. 69. Cross-section through the bulbous enlargement of central end of human ulnar, 6 1/2 months after injury and no union with central end. Shows the increase in the endoneural connective tissue as well as in the nerve fibres. The fibre marked + shows cross section of new fibres in same sheath with old. Compare with figs. 63, 65.' (405-406)

Figs. 56-57 in text:

'when the new myeline is evidently just formed no definite axis can be demonstrated by the gold stain. With the picric acid and haematoxylin stain, on the contrary, the newly formed axis is clearly seen, even at that early stage in the formation of the myeline tube when it exists as a string of bead-like swellings (see Figs. 56 and 57). It follows then that though the myeline sheath probably begins to form before the axis cylinder can be distinguished, the latter appears shortly afterward, before the new fibre has gone far in its development.' (380-381)

Fig. 61 in text:

'Very many apparent examples of this branching of the axis were found in this specimen, but owing, to the intricate way in which the fibres were twisted and the possibilities of deception arising therefrom, the connection of the old and new axes could not be satisfactorily followed. The example given in the drawing, and some others, were, however, quite distinct and seem to us to give fairly satisfactory histological proof that in regeneration the new axis cylinders are outgrowths from the axes of the uninjured fibres of the cenitral end. Fig. 61 gives an apparent example from another experiment of the outgrowth of the axis.' (382)

Figs. 62-69 in text:

'We have not followed all the stages of degeneration and regeneration in the central end with the same care as in the peripheral end; but the stages we have examined have convinced us that the processes are practically identical in the two ends. The myeline and axis disintegrate and are absorbed for a certain distance; an embryonic fibre is formed from the new protoplasm arising from the nuclei, and in this a myeline sheath is first formed into which an axis cylinder penetrates as an outgrowth from the end of the old axis. Various examnples of this formation of a new fibre within the sheath of the old are shown in Figs. 62 to 72. In many cases, in the central end, when union was not made or when difficult union was made as in cross sutures, an old fibre was found to terminate in a bunch of two or more new fibres (see Figs. 62 and 63), usually coiled round one another so that they could not be disentangled.' (382)

Figs. 63, 65 and 67 in text:

'Still another interesting fact is shown by the teased preparations of the central end of the same nerve, and that is, that the degenerative changes in the central end, when union is not made, apparently keep on progressing centripetally at a slow rate. Figs. 63, 65 and 67 give good illustrations of this fact.' (383)

Figs. 63-65 and 69 in text:

'Cross sections of the central stump of the human ulnar nerve operated upon by Dr Mc Graw for secondary suture, six and a-half months after injury, confirmed the results which were obtained by teasing (see Fig. 69). The section was rnade through the bulbous enlargement of the central stump. At the level of the section no normal medullated fibres were found, though occasionally a cross section of a smaller fibre with some remnant of the axis was seen. In other places a bundle of small fibres was found of the same area as one of these enlarged fibres, and at still other spots intermediate stages were seen showing an enlarged fibre surrounded by small new fibres in the same sheath. In this case the bulbous enlargement was undoubtedly caused by an increase in the nerve fibres as well as in the epineural connective tissue. If the cross section described is compared with the teased preparations, Figs. 63, 64, 65, made from the central end of the same nerve the way in which an old fibre makes connection with a bundle of new fibres lying in the same sheath will be more readily understood. One can understand from the teased preparations how in the cross section a portion of the myeline of an old fibre may be obtained surrounded by a number of newly-formed fibres in the same sheath. ' (383)

Figs. 64 and 65 in text:

'Similar preparations were obtained from the central end of the ulnar, in a dog, which had been severed 75 days before the examination was made, and had not been allowed to unite with the peripheral end. In some of the specimens from this latter nerve the mode of union of the axis cylinder in the newly regenerated fibre with the axis in the old fibre is clearly shown (Figs. 70, 71, and 72). Fig. 71 is particularly instructive when compared with Figs. 64 and 65. The new axis cylinder is seen to escape the swollen end of the old fibre and to penetrate the inyeline some distance beyond this point in order to reach the old axis. Fig. 72 shows the end of an old axis cylinder enlarged and sending out a new axis.' (383-384)

Explanation of Plate XVII (figs. 70-76):

'Figs. 70, 71, 72. To show the connection between the axis cylinder of the old and the new fibre. Compare 71 with figs. 64, 65. Exp. XXII. Central ulnar after 75 days, no union. Picric acid and haematoxylin.

Fig. 73. To show the formation of two embryonic fibres in a fibre of central end near the wound, connecting a distal piece which has not degenerated. Central end of human ulnar 61 months after injury. Gold and haematoxylin.

Fig. 74. To show intercalated pieces of newly formed fibres without degeneration of distal end. Exp. XVII. L. S. 4 weeks. Wound. Osmic acid and haematoxylin.

Fig. 75. The same, from central end of ulnar (Exp. XXII.) 75 days after section, shows the axis cylinder in the intercalated piece and its connection on each side with the intact axis of old fibre. Picric acid and haematoxylin.

Fig. 76. The same. Exp. XXII. Central end of ulnar after 75 days. Picric acid and haematoxylin.' (406)

Figs. 70-72 in text:

'We have not followed all the stages of degeneration and regeneration in the central end with the same care as in the peripheral end; but the stages we have examined have convinced us that the processes are practically identical in the two ends. The myeline and axis disintegrate and are absorbed for a certain distance; an embryonic fibre is formed from the new protoplasm arising from the nuclei, and in this a myeline sheath is first formed into which an axis cylinder penetrates as an outgrowth from the end of the old axis. Various examnples of this formation of a new fibre within the sheath of the old are shown in Figs. 62 to 72. In many cases, in the central end, when union was not made or when difficult union was made as in cross sutures, an old fibre was found to terminate in a bunch of two or more new fibres (see Figs. 62 and 63), usually coiled round one another so that they could not be disentangled.' (382)

'Similar preparations were obtained from the central end of the ulnar, in a dog, which had been severed 75 days before the examination was made, and had not been allowed to unite with the peripheral end. In some of the specimens from this latter nerve the mode of union of the axis cylinder in the newly regenerated fibre with the axis in the old fibre is clearly shown (Figs. 70, 71, and 72). Fig. 71 is particularly instructive when compared with Figs. 64 and 65. The new axis cylinder is seen to escape the swollen end of the old fibre and to penetrate the inyeline some distance beyond this point in order to reach the old axis. Fig. 72 shows the end of an old axis cylinder enlarged and sending out a new axis.' (383-384)

Figs. 73-76 in text:

'In conclusion we wish to speak briefly of an appearance often seen in the central end of the nerve, whether or not union is made with the peripheral end, an appearance which has been frequently described by other observers, but which has not been explained, as far as we are aware. This phenomenon consists in the intercalation of a segment of one or even two new fibres between the ends of an old fibre which has not undergone degeneration. (See Figs. 73, 74, 75, 76.) The noteworthy thing is the apparent exception to the general rule that the distal portion of a severed fibre always degenerates.' (384)