The Rat-Trypanosome, Trypanosoma lewisi, in its Relation to the Rat-Flea, Cerato-phyllus fasciatus

In this memoir we give a detailed account of investigations which have occupied us intermittently, with many interruptions, during the past five years. When we undertook this task our object in view was to work out as fully as possible the life-history and mode of transmission of a trypanosome, so that in at least one species of these important parasites its relation to the invertebrate host might be as thoroughly known as, for instance, the relation of the malarial parasite to the mosquito, thus furnishing a standard with which the life-histories of other species of trypanosomes might be compared and contrasted as they become known. How far we have succeeded in our task must be left to our readers to judge.

The species which we selected as the subject of our investigations was Trypanosoma lewisi, the common parasite of rats which is apparently of world-wide distribution. This species offers many advantages for such a study. It is common in London and easily procured when required; its vertebrate host, the rat, is a mammal of small size, the domesticated variety of which lives well and breeds rapidly in confinement, is inoffensive, and is easily handled; its invertebrate host, the rat-flea, is also easy to keep in captivity and is extremely prolific, and it is of a size which, though it increases to some extent the difficulties of manipulation, has a great advantage that the material to be searched and studied microscopically is confined within a small compass¹; and finally, by no means the least of the advantages of working with T. lewisi is the fact that it is non-pathogenic to its natural host and cannot live at all in human blood.

Since there is no difficulty whatever in obtaining the vertebrate host in abundance, either in the clean (i. e. noninfected) or infected condition, our first care was to obtain a stock of the invertebrate host, the flea. This we succeeded in doing from rats trapped in the open near Mr. Gurney’s Laboratory at Sutton Broad, Norfolk. Fifty specimens of Ceyatophyllus fasciatus were obtained in this way in the autumn of 1908, and were kindly identified for us by the Hon. N. 0. Rothschild, and with these fleas a breeding-cage was stocked and a flea-farm started. The cages used were of the type used by the Plague Commission, as figured in the ‘Journal of Hygiene,’ vol. vi, pl. iv. A rat was kept in the cage to feed the fleas, and they were left to themselves. Early in 1909 one of us (E. A. M.) went to Rovigno for some three months, during which time the fleas were left to breed under the care of an assistant, whose duties consisted of attending to the rat and replacing it if it fell ill or died. When the cage was examined after Easter it was found to be swarming with fleas, and our work began in May, 1909. We have worked ever since then with the fertile progeny of the original fifty fleas from Norfolk, and have never added further to our stock from without. The fleas breed so fast that it is often necessary to keep their numbers down, otherwise they take too much blood from the rat and affect its health. Fresh breeding-cages have also been started, and during the greater part of the time that we have been at work we have kept two cages constantly going, one in which the fleas are fed always on a clean healthy rat, and another in which an infected rat is always kept. We shall refer to these two cages as the non-infected and the infected breeding-cages respectively. As will be shown below, the stock of fleas with which we have worked all along was fortunately quite free from any natural infection with leptomonad or other flagellate parasites. Thus we have been saved from a fertile source of confusion and error, since we can be quite certain that any flagellates found in our fleas are stages of T. lewisi and nothing else.

Although we cannot claim that in our work we have solved completely every problem presented by the transmission of the trypanosome and its development in the flea—a result which probably no man could achieve in a life-time—we think it now fitting that we should publish such results as we have obtained, after having done as much as we were able to do in the time and under the circumstances. We claim at least that we have not jumped to our conclusions; our note-books contain not only the records of many experiments, but also of the dissection and examination of over 1,600 fleas, and we have over 700 drawings of stages of the development of the trypanosome, from which those given in this memoir are a selection. It would, indeed, have been easier for us to have written a plausible and apparently complete account of the development of T. lewisi, full of positive statements, after one year of our work than it is after five years, during which we have been forced by the logic of facts to abandon or modify many of our earlier conclusions or beliefs.

It is our pleasant duty at this point to express our thanks to those of our friends and colleagues to whom we are indebted for assistance. To Dr. Woodcock and Miss Robertson we are grateful for much advice, friendly criticism, and valuable suggestions. Our work could not have been carried out, certainly not in the time at all events, without the assistance of Miss Rhodes, who has not only drawn all the illustrations with a skill to which it is quite unnecessary for us to draw the reader’s attention (since the figures speak for themselves), but has also relieved us of a large part of the wearisome drudgery of searching through the microscopic preparations. Mr. George Kauffmann has been most helpful in every part of the investigation, not only assisting in making preparations, examining rats, and other similar duties, but more especially in carrying out intelligently and enthusiastically all the details of the experiments, in which his extraordinary skill and resourcefulness in controlling the wayward flea were invaluable. Dr. D. J. Reid has given us the benefit of his skill and experience in microphotography. From our colleagues of the Lister Institute, Dr. C. J. Martin, the Director, and Mr. Bacot, who have been themselves engaged in studying the transmission of plague by fleas, we have had many valuable hints and help in various ways. To each and all of these we desire to express our cordial thauks and gratitude.

The fleas collected for dissection and examination were thrown, or allowed to hop, on to the surface of a small quantity of salt-citrate solution¹ placed in a suitable glass capsule. The fleas are quite helpless on the surface of the liquid, and each flea that it is required to dissect can be picked off the surface of the liquid and transferred to a small drop of the same solution on a slide for further operation.

The examinations of the fleas were usually conducted by both of us acting in concert. Oue of us worked with the dissecting- microscope, extracted the parts of the flea required, placed them on slides, covered them with glass slips, and handed them to the other, who proceeded to search them carefully through under a microscope, using dry lenses of fairly high magnification (Zeiss D or apochromatic 4 mm.). In some cases one of us worked entirely alone, but it is difficult for one person to carry out satisfactorily both the dissection and examination of the flea; the various parts of the digestive tract often require prolonged and careful searching to find the flagellates, and if the operator be working single-handed, one preparation may dry up while he is searching through another.

For the dissection of the flea² the following apparatus was used: A pair of fine needles mounted in wooden handles, a fine pair of forceps, and a dissecting-microscope, besides slides and coverslips. The needles used were sharpened on a hone, one to a sharp point, the other to a flat, chisel-like edge with rounded corners. The pointed needle was the more useful for holding, the flat-edged needle for cutting. The dissecting-microscope used was a Zeiss binocular, with No. 4 eyepieces and the paired objectives F⁶⁵. It is also necessary for the dissector to have at hand an ordinary microscope armed with a low power, since it is often difficult to distinguish the minute’ organs of the flea under the dissecting-microscope; the intestine, for instance, if severed from its connections might easily be confounded with a portion of a Malpighian tubule.

In the following paragraphs is described the method of procedure for making what may be considered an exhaustive examination of a flea for trypanosomes; it was not always necessary, however, to attempt so much, nor is it claimed that the entire operation was always successfully carried out, since both our knowledge of the flea’s anatomy, and our skill in extracting the organs required, advanced considerably during the progress of our investigations.

The flea, as stated above, is picked up with a fine pair of forceps, holding it by its head, and placed on a slide (slide 1) in a drop of salt-citrate solution. The first operation is to cut off the head, which is not always easy if the flea be a lively one, in which case it is best to asphyxiate or drown the flea partially by holding it under water with the forceps for a short time. To decapitate the flea, hold it still by pressing the pointed needle across the thorax, and with the flat-edged needle cut across the head in the region of the eyes. The severed head may then be removed to another slide (slide 2), covered with a cover-glass, and the contents of the proboscis examined; but as the proboscis was never found to contain trypanosomes we ceased to trouble about it in our later studies.

It is frequently the case that the flea has its rectum filled with fæces or with partially digested blood, and when this is so it happens commonly that the rectum empties itself by a violent contraction at the instant that the head is severed (sometimes also eggs are extruded); or if the evacuation does not take place at this point in the proceedings, it is very difficult to avoid squeezing out the contents of the distended rectum during the subsequent operation of opening the abdomen. In cases where fæces are thus extruded the body of the flea is removed at once to another slide (slide 3), and the fæces left on slide 1 are covered with a slip and examined.

Through the integument of the flea the stomach can be seen lying ventrally in the anterior of the abdomen, and often the rectum can be seen at the hinder end in the dorsal region. The ventral posterior and dorsal anterior part of the abdomen is seen to be occupied by a whitish mass, most conspicuous in the female, and consisting chiefly of the reproductive organs.

The next stage in the proceedings is to open the body of the flea. This is done near the hinder end, at about the level of the fourth or fifth tergite of the abdomen. The body is held still with the pointed needle, with which the thoracic region is pressed down or speared, and with the flat-edged needle the body-wall is cut through dorsally and ventrally in the region indicated, and the hindermost segments of the abdomen gently detached in such a way as to separate the integumental portions without rupturing or tearing the internal organs. It is especially important, if it be desired to examine the contents of the body-cavity, that the digestive tract should not be in any way torn or punctured. By holding the anterior part of the body and pulling gently on the detached hinder part, the gut can be stretched out and seen in nearly its full length; the stomach, usually containing a greater or less amount of more or less digested blood, is seen projecting from the anterior part of the body, the rectum is contained in the detached hinder part, and stretching between the two is the intestine like a delicate white filament, exposed in its whole length, but more or less obscured by the fat-body, Malpighian tubules, and generative organs, especially by the large ovaries in the female; these organs render the female flea much more difficult to dissect, in spite of its larger size, than the less-encumbered male. The generative organs and as much as possible of the fatbody are now pulled out on to the slide and cut off from the body, care being taken not to injure the gut. The carcase of the flea, with the hinder part hanging on by the still intact intestine, is now removed to another slide (slide 4), and the extracted contents of the body-cavity on slide 3 can be covered with a slip and passed on for examination; but so far as stages of T. lewisi are concerned, it is superfluous to do so, since they are never found in the body-cavity unless the gut has been punctured or ruptured.

The next step is to divide the digestive tract into two parts, thereby severing completely the hinder part of the body from the fore-part. This is done at the point at which the Malpighian tubules are given off at the junction of the stomach and intestine, the region which represents the transition from the mid-gut, lined by endoderm, to the hindgut or proctodæum, lined by ectoderm. The Malpighian tubules are four in number in the flea; two of them run forward a short way on the wall of the stomach right and left, attached to it by fine tracheal tubes, and then turn backwards again with a sharp, elbow-like bend towards the dorsal side of the body; the other two tubules run backwards parallel to the intestine and alongside of it towards the hinder end of the body. The posterior pair of the tubules are also bent on themselves towards their distal extremities, but not so regularly as the anterior pair. The gut is cut across with the flat-edged needle at the point of origin of the tubules, and if this be performed accurately one pair of tubules (the anterior pair) remains attached to the stomach, the other pair to the intestine; sometimes, however, all four tubules remain attached to one or other of these organs. The hinder part of the body, with the intestine and rectum, is now removed to another slide (slide 5). The stomach is then pulled backwards out of the anterior part of the body on slide 4, and with it come out also, continuous with its anterior termination, the proventriculus and the oesophagus, these two parts representing the embryonic stomodæum, lined by ectoderm, while the stomach represents the whole of the embryonic midgut. The proventriculus is lined by a thick chitinous cuticle prolonged into stiff, curved, pointed spines, densely planted and forming, apparently, a straining apparatus; it is approximately globular in form and usually contains blood. The oesophagus is a delicate tube, its walls composed of the chitinous cuticle internally and a delicate network of muscles externally; it generally performs active movements, twisting from side to side, when freshly extracted.

The two pairs of salivary glands are situated in the anterior region of the abdomen right and left of the stomach. Each gland has the form of a simple oval pouch, the wall of which is composed of a single layer of large cells with very large nuclei. From each gland comes off a duct, which, after running a short distance, unites with the similar duct of the other gland of the same pair. (In one instance we have seen the two glands of one side of the body fused into one, but with their ducts quite separate; on the other side of the body there was a pair of distinct glands in theii’ normal relations). The common duct of each pair of glands then passes forwards alongside the gut through the thorax into the head, where it meets and joins with the corresponding duct from the other side of the body. The common salivary duct then runs a short distance and opens into the proboscis, doubtless on the hypopharynx as in other insects. The salivary ducts are recognisable at once under the microscope by their trachea-like structure, being lined by a thick cuticle which has ring-like thickenings; the rings are, however,somewhat irregular and easily distinguishable from the very even and regular spiral thickening of the wall of a tracheal tubule. Externally to the cuticular lining the tubule is covered by an investing layer of protoplasm, of uneven thickness in different parts and containing fairly large nuclei at irregular intervals. The ring-like thickenings of the cuticular lining become less marked as the ducts approach their point of junction, and cease altogether before they unite; the cuticular lining being quite smooth in the common duct and for short distances in the paired ducts.

Not infrequently the salivary glands come out with the stomach when it is pulled out; more usually, however, they do not do so, but remain in situ. In such cases the anterior part of the body is removed to another slide (slide 6), and the stomach, left on slide 4, is teased up, covered, and handed on for examination.

Now the dissection of the hinder part of the body, on slide 5, is proceeded with, in order to extract and separate the intestine and rectum. The rectum, situated dorsally to the accessory reproductive apparatus, penis or receptaculum seminis, is a fairly large pear-shaped organ, the stalk of the pear terminating in the anus. The slender intestine joins the rectum at its broad end, and in this region are situated the six conspicuous rectal papillæ, remarkable and very characteristic structures, the presfence of a single one of which makes it easy to recognise even a small fragment of the rectum. Behind the papillæ the rectum has a thin wall, to which the crithidial stage of the trypanosome, when present, is usually found attached, sometimes in vast numbers. In its anterior part, the region of the papillæ, the rectum has only circular muscle bands, between which are wide interspaces. In the hinder region, behind the papillæ, there are both circular and longitudinal muscle-bands; the latter can be traced forward to just behind the papillæ, at which point each band becomes rapidly narrowed to a tendon-like fibre, and at the same time the striations of the muscle disappear. The tendinous continuations can be traced forwards, in the living condition, for some distance, but we have not made out the exact points of their insertion.

The intestine is characterised by a continuous coat of ringlike muscle-bands, with interspaces, arranged very regularly external to the epithelium. When the edge of the intestine is focussed under the microscope, the layer of circularly-disposed muscle-libres is seen in optical transverse section like a string of beads. The intestine is frequently seen to be performing active peristaltic movements, and it may be thicker in some parts than in others, owing to the contraction of the muscles.

The rectum must be dissected carefully out of the hinder part of the body, so that it remains on the slide, free from all the adjacent organs or chitinous plates of the integument. The easiest way to do this is to make an obliquely longitudinal cut with the flat-edged needle so as to sever the ventral-anterior half of the hindmost segments, together with the genitalia, from the dorsal-posterior half containing the rectum and anus. The genitalia can then be removed and the rectum extracted without much difficulty. It requires some care to separate it from the anus without injuring it. When this has been accomplished, all unnecessary debris is cleared away. If it be desired to make separate examinations of the intestine and rectum, the intestine is cut off as close as possible to its junction with the rectum. To effect this it is best to spear the rectum with the pointed needle and make the cut with the flat-edged needle; or the operation of cutting off the intestine may be performed before the rectum has been dissected out from the hinder part of the body. In either case, the intestine is removed to another slide (slide 7), and both rectum and intestine, on their respective slides (5 and 7) are teased up, covered, and passed on for examination. It is not difficult to tear the rectum into several pieces with the needles, but it is not so easy to tease up the intestine; it is too slender to make sure of splitting it lengthways, except by good luck and more or less accidentally, and it is necessary as a rule to content oneself with cutting it transversely into two or three short pieces, the contents of which are generally squeezed out during the process.

Finally there remain the salivary glands, on slide 6, in the portion of the carcase consisting of the thorax and fore-part of the abdomen from which the gut has been extracted. The salivary glands, as has been stated above, are lodged in the fore-part of the abdomen beside the stomach, and it is generally by no means difficult to extract them when the stomach has been removed. To do this it is best first to spear the thorax with the pointed needle, then insert the flat-edged needle into the abdominal cavity from behind, and rake out gently the contents of the abdomen. The salivary glands sometimes come out fairly clean, but more often they are embedded in fat-body, tracheæ, etc., from which they must be carefully freed as much as possible. In such cases they are sometimes a little difficult to detect under the dissecting microscope, but theii’ position may be traced by their long, thread-like ducts. They are much smaller in the male flea than in the female. Another method which sometimes succeeds better in extracting the glands is to pull on the integument of the thorax with one needle and on that of the abdomen with the other. The body-wall then often tears across at the junction of the thorax and abdomen, and the salivary ducts are seen at once stretched out between the two. By continuing to pull the thorax forwards, the glands may be pulled out of the abdominal cavity and are seen hanging on to the back of the thorax, from which it is not difficult to detach them. By this method the glands may often be obtained very clean and free from encumbering fat and other tissue. When the glands have been extracted, other débris is cleared away and the coverslip is put on. The glands are very soft and are crushed immediately by the weight of a coverslip if there is no other tissue under it; but for examination of their contents this is not a disadvantage.

In the foregoing paragraphs we have given a detailed account of a full examination of the flea, such as we practised in the earlier periods of our investigation. But when it became evident to us that the trypanosome, during its development in the flea, never strays beyond the limits of the digestive tract proper, we were able greatly to curtail the ritual of the examination and to omit entirely theproboscis,. body-cavity, and salivary glands. It is also unnecessary, as a rule, to separate the intestine and rectum in the dissection. Consequently, our later examinations were reduced to (1) th© excluded fæces, if any, on the slide on which the flea was decapitated; (2) the stomach, on a second slide; and (3) the rectum and intestine, on a third.

It was no part of our task to make a special and detailed study of the anatomy of the flea, but a few points observed by us incidentally in our dissections may be noted briefly here.

The nervous system, of which some beautiful dissections were made in this laboratory by Major Christophers, I.M.S., consists, as in insects generally, of (1) the brain or supra-œsophageal ganglion-complex, sending off the peri-œsophageal connectives which pass on either side of the oesophagus to connect with the foremost of (2) the three large thoracic ganglia, joined by connectives to form a series which passes on into (3) the abdominal chain of ganglia. It is a very difficult operation to dissect out the brain and the first two thoracic ganglia, but it happens very frequently that in the ordinary dissections of the flea the third thoracic ganglion and the abdominal chain of ganglia are exposed entire and in continuity. It is then seen that the abdominal chain consists of a series of small ganglia terminated posteriorly by a larger ganglion; and further that in the male there are seven smaller ganglia, in the female only six, in the abdominal chain. The larger hindmost ganglion, from which nerves are sent off to the genitalia and rectum, evidently represents a fusion of several ganglia equivalent to the more anterior smaller ganglia. Consequently it is seen that the concentration and fusion of ganglia at the hinder extremity of the ventral chain has proceeded a step further in the female than in the male.

The genitalia consist, in the male, of a conspicuous pair of testes, situated dorsally in the abdomen, and a pail of filamentous glands (prostates?) not unlike Malpighian tubules at first sight, but of slightly smaller calibre, and differing entirely in histological structure. There is no separate seminal vesicle, but each testis is a tightly convoluted tubule, the lower end of which is dilated to contain the ripe spermatozoa. Ducts from the testes and prostates unite to form a median Ductus ejaculatorios, which opens into a large penis of very complicated structure. In the female the two ovaries occupy practically the same position as the testes, but take up much more space and extend forwards to the most anterior limits of the abdomen. Each ovary consists usually of four egg-tubes or ovarioles, but in one specimen that we have mounted as a permanent preparation there are five ovarioles on each side. The ducts of the ovarioles unite to form the paired oviducts, and these unite in their turn to form the common oviduct. Ventral to the common oviduct lies the unpaired receptaculum seminis, consisting of a brown, chitinous capsule of a peculiar shape. The main body of the capsule is spherical, but gives off a curved, horn-shaped diverticulum, ending blindly. The horn-shaped portion has its concave curve turned towards, and connected by striped muscles with, the spherical part of the capsule. A slender duct of great length, and much convoluted near its origin, arises from the spherical part of the capsule, and runs back to open probably into the distal extremity of the oviduct or into the genital vestibule. The spherical part of the capsule and duct of the receptaculum are surrounded with unicellular glands, thickly clustered round the capsule and the convoluted portion of the duct, but thinning out and becoming smaller towards the distal end of the duct. The receptaculum, dissected out, stained and mounted for the microscope, is a singularly beautiful object. It usually contains a dense mass of spermatozoa.

The heart is frequently seen in dissections at the hinder end of the body as a delicate filament, which by its own contractions twists and lashes itself about. Under the microscope it appears a delicate tube, beset towards the hinder end by the pericardial cells which are attached to it on either side, right and left, and are crowded together towards the hinder end, but occur more sparingly towards the middle region and are wanting in the anterior third of the heart. The ostia appear to be confined to the posterior region of the heart, but we have not made out their exact number or arrangement. For the pericardflil cells, see Minchin (1910).

In a former publication one of us (E. A. M., 1910) has described some parasites found in our stock of fleas. The most important was a form to which the name Malpighiella refringens was given, occurring, as the generic name implies, in the Malpighian tubules of the flea. Since that time this infection seems to have died out entirely incur fleas, and we have not seen any Malpighiella in the fleas dissected by us for the last three years or more. Why this parasite should have died out in our fleas it is impossible to say, but it may be remarked that no conditions could possibly be more favourable for contaminative infection from flea to flea (whether from adult to adult, or larva to larva, or adult to larva, or vice versa) than those in our breeding cages, where vast numbers of fleas in all stages of development are herded together in a confined space. Consequently the disappearance of Malpighiella in our cages rather indicates that the fleas do not acquire infection with this parasite by the contaminative method.

In the publication referred to, numerous yeast-like bodies were described and figured from the digestive tract of the flea. Since then we have found organisms of this kind abundantly in smears of the salivary glands (text-fig. 24, p. 642).

In the larvæ of fleas that we have dissected and examined from our cages we have found the gregarine Agrippina bona (Strickland, 1912).

The cysticercoids of tapeworms are found not infrequently in the fleas. Nicoll and Minchin (1911) described two species of cysticercoids from our fleas, representing Hymenolepis diminuta and another species of the same genus. We have found the same two species frequently, and also have in our possession specimens of a third species not identified. The cysticercoids appeal sporadically, and are sometimes quite common for a period, and then are not found again for a long time. This uncertainty in their occurrence is quite intelligible, since their appearance must be caused by the introduction into the cage of a rat infected with tape-worms, which doubtless infects a large number of the larvæ that later become adult fleas.

The point upon which we wish to lay special stress is the absence in our stock of fleas of any flagellate parasites, and more especially of the leptomonad described, by Swingle (1911) under the name Herpetomonas (Leptomonas) pattoni. We have been at great pains to convince ourselves upon this point. In the first place we dissected at various times about eighty¹ fleas from the non-infected breeding-cage without finding any flagellates of any kind in them, while flagellate parasites occur in a very large percentage of those known to have fed upon infected rats, though not in all, since the trypanosome often fails to establish itself in the flea, and even when the insect has been fed on a rat with trypanosomes swarming in the blood, they often disappear completely from the digestive tract of the flea within twenty-four hours of its having fed.

We give here three tables (A, B (1), and B (2)) showing the results of dissections of fleas from our stock which had been put upon infected rats, and so had had the chance of acquiring an infection of T. lewisi. Fleas do not always feed, however, when given the opportunity to do so, especially in cold weather,² and if the fleas are dissected and examined within twenty-four hours after having been put on the rat (the fleas in all cases having been kept hungry for two or three days previously to being put on), it is quite easy to distinguish those which have been fed from those which have not availed themselves of the opportunity of doing so. In this way useful controls are obtained for determining whether the fleas contained any flagellate infection before being used for putting on the infected rat.

From Table A it is seen that of 289 fleas which were put on infected, rats, 92 (31’83 per cent.) had not fed and contained no flagellates, 167 (57’79 per cent.) had fed and contained T. lewisi, and 30 (10’38 per cent.) had fed but contained no flagellates.

In addition to these negative data we have had the opportunity of comparing our stock of fleas with another stock which was actually infected with Leptomonas pattoni. When one of us (E.A.M.) was in Paris in January, 1913, he was very kindly presented by Dr. E. Chatton, of the Pasteur Institute, with some living fleas (Ceratophyllus fasciatus) from a stock infected with Leptomonas pattoni. These fleas were brought back to London and a fresh breeding-cage colonized with them. The fleas were left to breed for a year, during which time the rat in the breeding-cage was changed frequently, but none of the rats put in acquired any trypanosome-infection. When the fleas were examined at the beginning of 1914, they had multiplied enormously, and were found to contain Lepto-monas-infections. We did not keep any exact records of oui’ dissections of the Leptomonas-fleas, but, roughly speaking, about 50 per cent, of the fleas contained teeming infections. The leptomonads appear to establish themselves in the fleas as readily as does T. lewisi, perhaps more so, since, as will be seen from Table B (1), barely more than 14 per cent, of our stock of fleas contained swarming infections when exposed permanently to infection with T. lewisi in the infected breeding-cage, and Table B (2), if we count only those known to have fed on an infected rat, not less than six, not more than fourteen days previously, gives but a slightly higher percentage (21·19).

We may conclude, therefore, from a comparison of our stock of fleas with those bred from Dr. Chatton’s stock infected with the leptomonad, that, had our stock also been infected with leptomonads, we should not have failed to find fleas containing leptomonads in those fed on clean rats in the first place, and secondly, that the percentage of fleas containing flagellates would have been far higher than is shown by our tables, in fleas exposed to infection by T. lewisi.¹

We shall have occasion, when describing the developmental cycle of the trypanosome in Part II below, to relate how the trypanosome penetrates into the epithelial cells of the stomach of the flea and goes through a process of multiplication within them. It is a necessary preliminary, therefore, to understanding the effects of the parasites that we should preface our description of their development by some remarks upon the structure and contents of the flea’s stomach; and in the following section we give an account of our observations upon these matters, without claiming to have added anything to the scientific knowledge of insect histology.

The histology of the digestive tract of insects has been the subject of numerous memoirs, and its general characteristics are very well known. It would be beyond the scope of this memoir to attempt to discuss this subject in detail or to cite the very copious literature dealing with it; but of recent works we may refer more especially to the very excellent monograph of Léger and Duboscq (1902), who have studied the intestinal epithelium of Tracheata from the same point of view as ourselves, that is to say, with the object of describing the changes produced in the epithelium by parasites (gregarines) attacking the cells, None of the insects studied by Léger and Duboscq, however, were of blood-sucking habit and the stomach-epithelium of the flea differs in a number of points from any of the epithelia described by the French authors.

The wall of the stomach consists of the following principal layers, counted from within outwards (Pl. 39, fig. 126): (1) the lining epithelium, (2) a layer- of circular muscle fibres, and (3) a layer of longitudinal muscle fibres. In addition to these layers, which are very easily seen, there are to be found also, though by no means in every section, flattened epithelial cells external to, and in contact with, the lining epithelium, between it and the circular-muscle-layer; they occur sparingly and far apart, but appear to represent an integral and perhaps primitive constituent of the wall of the mid-gut. Similar flattened cells are found here and there on the Malpighian tubules, tire wall of which is similar in histological composition to the stomach-wall if the latter be imagined as reduced to the lining epithelium and the flattened cells alone, without the muscle-layers,

We are concerned here only with the lining epithelium of the stomach, but it may be mentioned in passing that the circular musclefibres occur as bauds or separate rings with considerable intervals between them, and consequently do not appear in every transverse section of the stomach. The longitudinal muscles are also separated from one another by intervals, a fact at once apparent in the transverse section, in which the muscles are seen cut across and in which it can further be seen in a well-preserved section, that each longitudinal muscle-fibre is connected to its neighbours by a delicate membrane, appearing as a fine line running between each adjacent pair of musclebands and forming a delicate sheath or investment round the whole stomach. The circular muscle-layei-is continued on into the intestine, where it forms a continuous investment without intervals between the bands; the longitudinal muscles end at the pylorus posteriorly and start anteriorly behind the proventriculus, which has its own system of musculature, running for the most part in oblique bands arranged symmetrically right and left. Each muscle-band in the stomach-wall is a single, transversely-striated fibre, in which an occasional elongated nucleus is seen, embedded in a small quantity of protoplasm.

The following account of the epithelium and contents of the stomachapplies, unless otherwise stated, to ections of the stomach fixed with Flemming’s fluid and stained with iron-hæmatoxylin ollowed by Lichtgriin-picric (see the following section dealing with technique). If such a section through a number of stomachs—all taken from a batch of fleas dissected at the same sitting and at the same interval of time after having fed on the infected rat, all preserved in the same way, all stuck on the same slice of liver, cut and stained simultaneously—be examined even in the most cursory manner, very considerable differences are seen between the stomachs in one and the same microtome-section. These differences affect both the epithelium and the contents. The epithelium varies, in the first place, in the form of the cells, from flattened to columnar, and secondly, in the staining reactions of the cells. The contents of the stomach, that is to say the blood-débris, vary greatly in colour, staining in some cases deep opaque black, or less deeply in various shades of grey, in other cases, however, bright yellow.

The variations in the form of the epithelial cells are to be ascribed to the differences in the degree to which the stomach is dilated by the ingested blood. In a gorged flea the distension of the stomach stretches the epithelium until the cells become thin and flattened; but when the flea is hungry, or has taken in a small quantity of blood, or when the quantity ingested has become reduced by digestion and absorption, the epithelium resumes what may be considered its normal columnar form. Every gradation between the flattened and columnar conditions can be found in different sections or in different parts of one and the same section.

The variations in the staining reactions of the epithelial cells depend, in the first place, on the age or senescence of the cells. It is a matter of common knowledge that the lining epithelium of the mid-gut of insects is continually being thrown off and regenerated. The ordinary epithelial cells do not multiply and no mitoses are ever found in them; the centres of regeneration are the so-called epithelial crypts, each representing morphologically a small diverticulum of the epithelium in which the approximation of the cells usually obliterates the cavity and produces a solid, bud-like mass of cells (Text-fig. 2, a and Pl. 44, tigs. 314,316). When a flea’s stomach, containing a certain amount of ingested blood, is plunged into a fixative, the epithelial crypts are very easily seen with a hand-lens or with the naked eye as little opaque white spots in the semi-transparent stomach-wall, very conspicuous against the reddish-brown background of the stomach-contents. In the sections it is common to find mitoses in such a crypt, especially towards its fundus (Text-fig. 2, a). As the cells multiply they are pushed upwards to the general level of the epithelium and outwards from the crypt to replace the old epithelial cells which, having degenerated, are cast off from the wall into the lumen of the stomach, and are digested there.

The young, freshly-regenerated epithelial cells have the cytoplasm clear, staining light-grey, and are relatively poor in granulations; the older cells, on the contrary, have the cytoplasm full of granules that stain very deeply, until finally the whole cell, including its nucleus becomes black and opaque. Consequently, the epithelia of different stomachs show very varied appearances. A recently regenerated stomach will show clear epithelium all round, and, according to the time that has elapsed since regeneration, there may be no detached cells in the lumen of the stomach, or there may be a certain number of detached black cells, or there may be still, here and there, isolated black cells or patches of such cells in situ in the epithelium or in process of being cast off from it. On the other hand, a stomach which is about to be regenerated shows very dark epithelium all round, and in places this may be in process of rejection and replacement from the crypts, in which the cells have clear cytoplasm. The condition of the epithelium may vary in different parts of the same stomach, and from what we have observed we have gained the impression that the regeneration proceeds usually from before backwards, so that the anterior part of the stomach is further advanced in-degeneration or regeneration, as the case may be, than the posterior region. We find in our preparations all possible conditions of the epithelium in different stomachs in one and the same microtome-section, and we have not been able to establish any definite relation between the feeds of the flea and the regeneration of the epithelium, but we have not paid sufficient attention to this point to be able to state positively that no such relation exists; the differences seen in the epithelia of flea-stomachs examined at the same interval of time after feeding may be due to inequalities in the rate at which digestion proceeds.

In addition to what appears to be the normal process of senile decay, in which the cells take up the iron-hæmatoxy]in stain very deeply and become black and opaque, we have observed a second mode of degeneration, which we are inclined to ascribe to the action of the trypanosomes, since in all cases where it occurs in our preparations there are trypanosomes to be found in the stomach, and frequently in the degenerated cells themselves. In this second type of degeneration the black-staining granules in the cell diminish in quantity, without, however, disappearing entirely, while the cytoplasm of the cell stains yellow (Pl. 39, fig. 133; Pl. 40, fig. 140); hence we have generally referred to this condition in our notes as “yellow necrosis.” In all the stomachs in which we have found it the blood-debris is also stained yellow, and it is often very difficult to make out the precise boundary of the necrosed cell-body, or to distinguish the cells from the débris when they lie free in it (Pl. 39, fig. 125), except by the presence of the nucleus and. of a certain number of black granules in the cytoplasm of the necrosed cel]. Indeed, our first impression was that the yellow colouring matter of the blood had in some way penetrated into the cell and stained its cytoplasm, but there can be no doubt that this idea is au illusion and that the yellow colour, both of the blood-débris and of the necrosed cells, is due to the picric acid in the Lichtgriin-picric staining combination, though it is, of course, possible that the substance, whatever it may be, which stains yellow in the blood-débris may have become infiltrated into the dead cells and given them their peculiar staining properties.

The variations in the staining reactions of the contents of the stomach are more difficult to explain. There appear to be two types of staining after the use of the iron-hæmatoxylin-Lichtgrün-picric combination, one in which the contents stain in various shades of grey up to black, with a greenish tinge, and one in which they stain a bright lemon-yellow. It is not possible to bring these two types of staining into one series, as there is no transition between them; the grey-black and the yellow types occur side by side in different stomachs in one and the same microtome-section, and each stomach shows either the one or the other condition through the whole series of sections. So far as our observations permit us to generalise, the grey-black series represent the normal stages of the digestion of the blood; the yellow reaction appears to be due to some abnormal condition.

The blood ingested by the flea is very soon affected by the digestive action of the stomach, and the red corpuscles cease to be recognisable within a few hours after digestion. In the middle period of digestion, that is twenty-four hours, or thereabouts, after feeding, the blood has become thick, viscous and brick-red in colour, and contains immense numbers of irregularly shaped grains of all sizes, but for the most part coarse and large. Towards the end of digestion, forty-eight hours or so after feeding, the stomach contents are fluid and watery, dark brownish-black in colour, and the grains are much diminished in number and in size.

In the sections, taking first the grey-black series, the blood in the earlier phases of digestion (eighteen to twenty-four hours) usually consists of densely-packed grains and spherules, varying in size from very coarse to very fine, and staining intensely black. Between the grains there is visible a coagulated albuminous matrix, stained greenish with the Lichtgriin-picric combination. The stomach-contents fill the whole section and adhere closely to the epithelial cells, penetrating down between them when they have the columnar form, but in the centre of the section there is generally a clear patch, circular in outline, which is seen to owe its clear appearance to the fact that coarse grains are absent, and it consists only of the albuminous matrix with finer granules. Hence the digestion, or more probably the passage backwards towards the rectum of the indigestible remnants, of the blood-débris appears to proceed from the centre of the section— that is to say, from the axial region of the stomach—towards the periphery.

As the digestion proceeds, the grains in the débris become smaller and stain less deeply; consequently the stomach-contents stain grey, in varying shades of darkness, while the matrix still shows the greenish hue. In stomachs thirty-six hours after feeding the contents of the stomach are generally greatly diminished in quantity, and are absorbed in the centre of the lumen, leaving a clear space of variable form, while round the periphery the greenish-grey débris adheres close to the epithelium. Leucocytes, especially the polymorphonuclear forms, can be recognised in the blood twenty-four hours after feeding, but at thirty-six hours we have not found them. Owing to the lighter tints of the stomach-contents at thirty-six hours, the trypanosomes free in the blood-débris can be seen more easily, in contrast with the earlier state of affairs.

In the yellow stomachs the contents appear at first almost uniform, but on close examination they are seen to consist of closely-packed granular substance, all of which, both granules and matrix, is coloured by the picric acid in the staining combination used. The first point that strikes one immediately is that the contents in such stomachs are large in quantity and fill the whole stomach, or show but a slight amount of absorption towards the centre of the section, even at thirty-six hours, when the contents of the grey-black stomachs are considerably diminished. The epithelium of the yellow stomachs may vary from the flattened to the columnar form, but the normal cells stain grey or black, in sharp contrast with the yellow contents.

It seems obvious from these data that the yellow stomachs represent an abnormal condition; we have endeavoured, not very successfully, to find a relation between this condition and either the presence of trypanosomes, on the one hand, or the state of the epithelium on the other.

In the yellow-staining stomachs which we have studied we have found trypanosomes to be present in the stomach in every case except one, and in that case there were attached clumps of crithidial forms immediately behind the pylorus, showing that the stomach-phase was over. But on the other hand, we have found the grey-black condition of the contents in well-infected stomachs also, showing at least that, if the yellow condition is in any way due to the parasites, they do not always produce that effect. On the other hand, in those cases in which we have found no trypanosomes at all, either in the stomach or outside it, the contents are always in the grey-black condition. A significant circumstance is, perhaps, the fact that we have only found the “yellow necrosis “of the cells in stomachs with yellow-stained contents.

As regards the condition of the epithelium, we have found the yellow condition of the contents associated (1) with epithelium black all round and in process of being cast off, or (2) with epithelium mostly clear, but with black patches of cells in situ or detached; in one such stomach the first condition is found in the anterior half, the second in the posterior. We can state, therefore, that in our experience the yellow-stained contents occur only in stomachs about to be regenerated, or in process of regeneration, or very recently regenerated. But, again, we have found the grey-black condition in stomachs that appeared also to have undergone regeneration very recently, which makes it difficult to correlate this condition with the process of regeneration. The question of the significance and cause of the yellow-staining condition of the stomach-contents must be left an open question at present; the data to hand do not suffice for drawing decisive conclusions, and it would lead us too far to attempt further investigations upon this problem. On the whole, however, it seems at least probable that we are dealing with an abnormal state of the digestive processes towards which the trypanosomes are a contributory cause, if not the sole one.

As already stated, the different conditions of the stomach-contents described above are those seen after staining with iron-hæmatoxylin and Lichtgriin-picric. After the use of Giemsa’s stain the colour of the contents differs considerably in different cases.

Most of our sections stained with Giemsa were fixed with Maier. In those in which the trypanosomes were best stained and show the flagella clearly and sharply the grains and spherules of the débris are coloured for the most part orange-pink, especially in those stomachs in which the digestion of the blood is further advanced (figs. 109,113, Pl. 38); in the earlier stages of digestion many of the ‘larger grains and masses in the débris are stained deep purple, making the contents of the stomach more opaque. In one of our series preserved in Flemming, consisting in all of seven slides, the first six were stained with the iron-hæmatoxy-lin-Lichtgrün-picric combination, the seventh with Giemsa’s stain; on this seventh slide there are sections of four stomachs, two of which, on the other six slides, show grey-black contents, while the remaining two have the contents yellow in colour. In the Giemsa-stained slide the blood-débris shows a coloration very different after Flemming to that which it shows aftei Maier, being stained a bluish-green tint. The stomachs of the yellow type are slightly more blue in tint than those of the grey-black series, but otherwise the difference between them is but slightly marked.

Having now described the chief variations in the conditions of the stomachs and their contents, or at least those differences which are obvious upon the most cursory inspection of the sections, it remains to give a more detailed account of the epithelial cell. In any given stomach the cells show great individual variation in form and structure, but, nevertheless, it is not possible to divide them into distinct classes. There are no special glandular or secreting cells, as described by Léger and Duboscq in other insects, and all the cells of the general epithelium of the stomach of the flea are to be regarded as equipotential, the differences visible between them being merely the expression of varying physiological conditions in relation to their changing environment, on the one hand, or to their constitutional vigour or senescence, on the other. Hence it is possible to give a generalised description of the cells, beginning first with the normal, healthy cell and dealing afterwards with the changes it undergoes in the process of degeneration.

The epithelial cells are produced, as already stated, in the “crypts of regeneration,” which have been described in various insects by Léger and Duboscq. In the flea these structures appear usually as solid, budlike cell-masses that often project outwards from the wall of the stomach to a considerable extent (Pl. 44, figs. 314, 316) beyond the level of the muscle-layers, which pass on either side of them. Internally the crypts do not rise up beyond the general level of the epithelium. The closely-packed cells of the crypts show distinct limits, and do not form a syncytial mass of protoplasm, as described by Léger and Duboscq (1. c., p. 410) in the larva of Anthrenus verbasci, for example. At the fundus of the crypt mitoses are often found, sometimes in two cells simultaneously in the same crypt; in other cases all the nuclei are in the resting state. Doubtless the crypts have periods of active multiplication, alternating with periods of repose, as in other insects. The crypts are often seen to be marked off from the general epithelium by slender dark cells, the “cellules de recouvrement” described by Léger and Duboscq (1. c., Pl. II, fig. 2, c.r.; p. 388). The crypts appear to have the monopoly of cell-production in the stomach of the flea. We have not found basal cells, “cellules de remplacement,” in the general epithelium.

By multiplication and increase in their numbers the cells are pushed outwards on all sides from the crypt to take their place in the general epithelium (Pl. 44, fig. 316, and Text-fig. 2a). The young epithelial cells seen in the immediate neighbourhood of the crypts are columnar cells, roughly rectangular in form, and generally about twice as high as they are broad. The lateral boundaries of the cell are approximately parallel, and each cell is in contact with its adjacent neighbours for its whole length. The free, apical surface of the cell is convex, and on this side is developed a very distinct, thick border, at first covering only the upper surface of the cell, which projects like a dome towards the lumen of the stomach.

The further development in the form of the cell consists in an extension of the upper free surface, brought about by the cells becoming free and separated from one another at their sides, first at their apices and then downwards along almost the whole length of the side of the cell, till finally each cell is connected with the adjacent cells only by a narrow isthmus at its base (Text-fig. 2b). As the cell becomes free the border develops also on the exposed surface, so that, instead of being confined to the apex of the cell, it extends down the vertical sides also (Pl. 40, fig. 136). This process of separation between the cells has an obvious significance in connection with the process of flattening which they undergo when the stomach is dilated after feeding; it can be regarded as an adaptation to the blood-sucking habit. When the flea gorges itself each cell is so stretched that its tallest part in the vertical direction is scarcely thicker than the nucleus, which bulges out the middle part of the cell in an even curve towards the lumen of the stomach, while towards the periphery the vertióle height of the cell diminishes to the isthmus connecting it with its neighbours (Text-fig. 2 c). As the cell resumes the columnar form the nucleus remains at or near the base, as a rule, and the cytoplasm of the cell is heaped up over it. In the extreme columnar form the apex of the cell is generally slightly expanded, the middle region more narrowed, so that spaces are left between adjacent cells, into which a considerable quantity of blood-débris penetrates (Pl. 39, fig. 126). The nucleus is usually situated at the base of the cell, but occasionally towards the apex (Pl. 40, fig. 136). The border clothes the whole free surface of the cell, whether flattened or columnar, and is of considerable thickness over the apex and the sides, becoming thinner as it approaches the isthmus, but in the columnar form of the cell, when its apical region is expanded, the border may be thinner, as if stretched, at the apex of the cell (Pl. 40, fig. 144).

The blood-debris has a great tendency to adhere closely to the border, so much so that the border is often more sharply marked off from the cell-contents within than from|the blood-d c b ri s without, in the sections, but places can be found occasionally where the blood-débris has split away from the epithelium, leaving the border distinct and sharp. The border appears usually homogeneous and réfringent, though in some preparations indications are seen of a vertical striation, as if it were composed of little darkly-stained rods, placed at right angles to its two limiting surfaces, and separated by intervening substance of lighter colour (PL 39, fig. 129, and Pl. 40, fig. 142). After sublimate-fixation the border is colourless, but when stained with iron-hæmatoxylin the blood-débris adhering to it often hinders the extraction of the stain and at these spots it remains black; when the haematoxylin is extracted it tends to take up the green from the Lichtgriin-picric mixture (Pl. 40, fig. 147). With Giemsa after sublimate-fixation it stains a pinkishyellow. After Flemming-fixation the border is yellowish, as if tinged by the chromic acid in the mixture, and when this fixation is followed by the Giemsa-stain the border is coloured green (Pl. 38, figs. 99-103), There is no “bordure en brosse,” or palisade of stiff rod-like cilia, external to the border, as in many insects. The condition in the flea more resembles that figured by Léger and Duboscq for Scolopendra (1. c., pl. vi).

The border is evidently a fairly tough structure since in teased up stomachs examined fresh, the borders of cells are often seen quite empty, but retaining their shape, like shells.

The nucleus of the epithelial cell calls for no special comment: as can be seen in our figures, it is rounded or oval, with the typical structure seen in tissue-cells, namely, a distinct membrane, a reticulum containing chromatin-grains of various sizes, and one or more nucleoli which stain black, like the chromatin, after-iron-hæmatoxylin. Mitoses of the usual type are found commonly in the crypts of regeneration, but we have never seen the slightest evidence of nuclear division in cells forming part of the general epithelium outside the crypts.

The cytoplasm of the epithelial cell varies at different ages. In the youngest cells bordering the crypts the cytoplasm appears more or less homogeneous and finely granular in aH parts of the cell; it stains light purplish-grey or grey-black after iron-hæmatoxylin, bluish-purple after Giemsa, and no coarse granulations are to be seen. In the fully developed cell the cytoplasm has undergone local differentiation; round the nucleus, in the basal half of the cell, it has a denser texture, but above the nucleus, in the apical region, it has become of looser consistence, more spongy, so to speak, in appearance, with irregular spaces (Pl. 39, figs. 126,127, Pl. 40, figs. 136,144), containing fluid in the living condition, and transversed by strands of protoplasm disposed irregularly. The more the apical part of the cells is expanded the more watery its contents appear. Sometimes the apical region appears almost empty in the sections, with only a few traces of cytoplasm close under the border and at the sides. It is in this region in which the stages of the trypanosomes are most often found, and into which the parasites first penetrate.

In addition to these changes in the cytoplasm, numerous grains and enclosures of various kinds make their appearance in it. A detailed study of these granulations would require a lengthy investigation, an expenditure of time and trouble, that would go beyond the scope and objects of this work. We must confine ourselves to a brief summary of the appearances seen in our sections, without attempting to give physiological explanations of the various conditions seen. It is obvious that the bare observation that a granule is stained black by iron-hæmatoxylin or red by Giemsa’s stain does not permit very far-reaching conclusions as to its nature or function in the cell; bodies of most diverse properties might agree to this extent in their reactions.

The first granulations to appear are minute grains which, whatever the fixation, Flemming or sublimate, stain black after iron-hæmatoxylin and red after Giemsa. They are seen at first chiefly at the sides of the nucleus, between it and the cell-wall and extend up the sides of the cell close under the border. Scanty at first in the apical spongy part of the cell, they are soon deposited in this region also, appearing often in considerable numbers and varying in size from small granules to conspicuous grains, and even large masses (Pl. 40, figs. 136,137.) The larger grains are seldom homogeneous, but appear as rings, black or red, as the case may be, with clear centre, apparently hollow (Pl. 38, fig. 99). Those of still larger size show, especially after Giemsa, darker and lighter parts disposed in various ways; inside the peripheral deeply-stained shell there may be darker grains or patches. After iron-hæmatoxylin, however, the whole mass may be opaque black, but usually shows lighter inner portions. The extent to which these granulations are developed varies in different stomachs, doubtless in relation to their secretive or absorptive activity at the moment of preservation. When a number of stomachs are cut in the same block, one stomach all through the series may show the cells clear and very free from granulations, while another stomach shows nearly every epithelial cell loaded with coarse grains in its apical region.

The red grains, as they may be termed from their distinctive reaction to Giemsa’s stain, appear to be always present in greater or less quantity in the fully-developed cells of every stomach. In addition there are often found, lodged in the apical spongy region of the cell, masses of relatively large size which do not retain the iron-hæmatoxylin stain firmly throughout their substance, and consequently appeal for the most part light grey in colour after this stain (Pl. 40, figs. 145, 146); after Giemsa they are either scarcely stained at all, appearing a sort of neutral tint, or they are coloured bluish-purple in various shades, sometimes very deeply, with streaks and blotches more reddish in tint (Pl. 38, fig. 102). These masses vary considerably in size and contour, and show differentiation of their substance into lighter and darker parts. With superficial examination they often simulate the intracellular stages of the trypanosomes to a remarkable degree, especially in the living condition, when they are often very conspicuous; for a long time we confused them with the spheres in the freshly teased-up stomachs, and spent much time watching them in the expectation, never of course fulfilled, of seeing them perform the characteristic movements. After we had made smears of stomachs in which these bodies were abundant, without finding any intracellular stages of the trypanosomes in such preparations, we came to the conclusion that these motionless spheres (as they appeared to be) were merely cell-products, and referred to them in our notes as “pseudospheres.” Even in sections the pseudospheres often mimic the true spheres and might be confused with them at first sight, but only by an inexperienced observer who had never seen the actual intracellular stages of the trypanosome in the epithelium. The idea occurred to us at one time that some of the pseudospheres might possibly be degenerated stages of the trypanosomes, destroyed, and in process of absorption, within the epithelial cells into which they had penetrated: but we have found no decisive evidence for this. It is most probable that the pseudospheres represent secretion-masses formed by the cell itself.

In some of the stomachs, especially in those preserved about twenty-four hours after feeding, there are to be seen dense and very conspicuous accumulations of coarse grains in the epithelial cells immediately below the border (Pl. 38, fig. 98, Pl. 40, fig. 147). The grains in question are especially distinct after fixation with Maier’s fluid; they are more difficult to make out in the stomachs fixed with Flemming. The grains resemble very closely those of the blood-débris adherent to the border external to the cell, so much so that the first impression gained is that the débris has been absorbed into the cell through the border. It is easy to imagine this after iron-hæmatoxylin, which stains both these granules and the débris very black after sublimate-fixation (fig. 147); but the Giemsa-stain colours the grains within the cell differently from the debris (fig. 98), and when the digestion of the blood has gone beyond a certain point the grains inside the cell may be stained much darker with iron-hæmatoxylin than the grains in the blood-débris. It is improbable that the coarse grains of the débris would pass bodily through the border, which is to all appearances a dense, tough structure; but it is probable that these grains are formed in the cell in direct relationship with the process of absorption of nutriment from the blood.

Amongst the enclosures of the epithelial cell must be mentioned finally peculiar yellow grains which occur with great frequency in some stomachs, not at all in others. Their presence or absence is in no way connected with that of the trypanosomes, and they occur both in normal as well as in degenerating cells, though perhaps more abundantly in the latter. In the Flemming-iron-hæmatoxylin sections these grains have a brownish-yellow tint, often with a darker shell (Pl. 40, fig. 141). They vary in size from small granules up to the large grains reaching as much as 13μ in diameter (fig. 142). Their tint also varies in depth, being usually much lighter in the larger grains. With Giemsa, after Flemming, they are stained bright green (Pl. 38, fig. 103), probably as the result of a blue stain (azure) imposed upon their original yellow tint.¹ These yellow bodies are very similar to, probaby identical in nature with, the enclosures characteristic of the pericardial cells. As one of us has described elsewhere (E.A.M., 1910), the pericardial cells of the flea may beso crammed with yellowish-brown grains and spheres that the cell becomes visible with the naked eye as an opaque black spot through the integument of the living flea. In some of our stomachsections there are also casual sections of pericardial cells which have been pulled out of the flea together with the stomach, so that we have had the opportunity of making a direct comparison between the yellow grains in the epithelial and pericardial cells. The yellow grains are probably an excretory product, eliminated by the flea under certain physiological but apparently normal conditions, and elaborated either in the epithelial cells, to be cast out into the lumen of the stomach, or in the body-cavity, to be taken up by the pericardial cells.

We come now to the process of cell-degeneration which occurs in the effete, senile epithelial cells. This process is very different in the flea’s stomach from that described by Léger and Duboscq in various insects, none of them of blood-sucking habit. It is described by these authors as a “Dégénérescence mucoide,” an infiltration of the cells with mucoid substance. In the flea’s stomach the process appears to be more of the nature of a fatty degeneration, combined perhaps with a mucoid infiltration.

In our sections fixed with Flemming’s fluid and stained with iron-hæmatoxylin the intensely black, often perfectly opaque, degenerated cells, which are seen frequently detached completely or in process of detachment from the epithelium, are very distinct from the clear, lightly-stained cells originating from the crypts of regeneration and taking the place of the degenerated cells (Pl. 44, fig. 316). In some of our sections the stain has been over-extracted: the trypanosomes have become ghosts, faintly visible only to the practised eye, the nuclei of the epithelial cells are pale, and even the blood-débris has had its usually intense black stain reduced to a shade of brown; but the black grains and masses in the epithelial cells remain as black as ever, showing that they do not owe their colour to the stain but to the fixation, that is to say, to the osmic acid in the Flemming’s fluid. Such preparations, spoilt for other purposes, are very useful for showing the gradual process of deposition of the blackened grains. First they appear as fine granules round the nucleus, near the base of the cell (PL 40, figs. 143, 144). Next, other, and for the most part larger masses, are deposited in the cytoplasm above the nucleus. The cell then becomes gradually filled up with black grains from below towards the apex; often an empty space is seen at the apex, immediately below the border (Pl. 40, fig. 138), but finally this, too, is filled up and the whole cell becomes an opaque black mass (fig. 139).

Very instructive is one of oui series preserved in Flemming, in which there is one stomach in which nearly all the epithelium is degenerate. The sections of this stomach are spread over seven slides, six of which were stained with iron-hæmatoxylin, while the seventh, on which are sections through the hindmost region of the stomach, was stained with Giemsa. On this slide the degeneration is not so far advanced as in the more anterior region of the stomach, and in different parts even of the same section the following conditions are to be found: (1) Cells of normal type, with clear cytoplasm containing a few red granules (Pl. 38, fig. 99); (2) cells with cytoplasm of a darker bluish-purple tint, with many more red granules and amongst them a few coarser grains intensely black in colour (Pl. 38, fig. 100); (3) cells in which both the red and the black grains, but especially the latter, are greatly increased in number, leading up to (4) opaque black cells in which nothing can be focussed clearly. The black grains, it is obvious, can only owe their colour to the action of the osmic acid in the fixation, and must therefore be of a fatty nature. On the other hand there is also a marked increase of the red grains in the degenerating cells, indicating, perhaps, that in addition to deposition of fat, there is also a tendency to mucoid infiltration, as described by Léger and Duboscq. The darker tint of the cytoplasm, in so far as this is not an optical effect due to crowding of the grains, indicates that it becomes impregnated with the substances produced in the process of degeneration.

The deposition of the fat round the nucleus in the first instance indicates that the nucleus takesan active share in the process, and this is borne out by the fact that the nuclei themselves become very dark in the degenerating cells and are sometimes quite opaque.¹

In sections of stomachs fixed with sublimate mixtures the blackening of the degenerating cells seen in the Flemming-fixed sections is conspicuously absent, so that at the first glance it is difficult to pick out the senile portions of the epithelium. More careful study of the sublimate sections shows that here the degenerated epithelium is distinguished from the regenerated by its pale, empty appearance, owing to the fat-grains having entirely disappeared, leaving empty spaces to mark their former position. This is best seen in stomachs fixed in sublimate-acetic, since, after sublimate-alcohol mixtures (Maier’s and Schaudinn’s) the cells are often much deformed and shrunk. In a favourable spot it is seen that the young cells, freshly produced from the crypts, have denser cytoplasm filling the cell throughout, except in the apical expanded portion of the cell; the cytoplasm stains deep grey or neutral tint after iron-hæmatoxylin and shows relatively few enclosures. The senile cells, on the contrary, are full of cavities, so that the cytoplasm has a spongy appearance throughout the cell and not merely in its apical region; and scattered through the spongy cytoplasm are grains, fine or but moderately coarse, which are stained black after iron-hæmatoxylin, red aftei Giemsa.

The difference between the senile cells after the two methods of fixation is easily explained if the grains deposited in them are principally fat. In all the sections alike the fat has been dissolved away during the process of imbedding in paraffin. In the Flemming-fixed sections, however, each fat-grain has reduced the OsO₄ to metallic osmium, and consequently is represented in the sections by a black mass, a model of the fat-globule in metallic osmium. In the sublimate-fixed sections no such reduction takes place, and the fat-globule is represented by an empty space; only the mucoid grains (if we are right in calling them so) remain in the cytoplasm, stained red or black according to the stain used.

It should be mentioned finally that after sublimate-fixations the blood-débris is stained very much blacker by iron-hæmatoxylin, and holds the stain much more tenaciously than after Flemming-fixation. This is especially true of that part of the débris which penetrates down between adjacent epithelial cells, and which often remains jet-black after all the rest of the débris has become pale in tint. In consequence the cells of the columnar epithelium in sublimate-fixed sections are often seen to be separated by black masses, which careless observation might confuse with the black stain of the degenerated cells after Flemming-fixation, especially when, as often happens in such sections, the main mass of the débris has shrunk away from the epithelium into the centre of the stomach-lumen. Such a mistake could only be made, however, with powers too low to discern that the black masses are between the cells and not in them.

The degenerated cells are thrown off bodily into the lumen of the stomach, which often contains great numbers of them in the blood-débris. There they are doubtless digested and absorbed along with the other contents of the stomach. Légei’ and Duboscq described a process of mucoid degeneration in which the entire cell, having a remarkable and deceptive resemblance to a gregarine, is engulphed by a basal cell; ultimately the latter also degenerates, and is thrown off with the cell it has taken in (1. c., p. 451). We have seen nothing of this sort in the flea, in which basal cells do not occur in the epithelium of the stomach.

We have already described above our methods of dissecting the flea and extracting from it the organs which it is required to examine for the presence of stages of T. lewisi. Here we propose to describe the methods by which the trypanosomes, when found, were preserved as permanent preparations for microscopic study.

The organs of the flea, extracted in the manner described above, are at once examined carefully under the microscope for the presence of trypanosomes in their various phases of development. When trypanosomes were found in any of the internal organs, after note had been taken, or sketches made, of their forms, position, and other points of interest, we proceeded to make permanent preparations of them. For this purpose the coverslip is carefully raised up, by means of the pair of fine needles that were used in the dissection of the flea, lifted off, and dropped at once with wet surface downwards into a suitable fixative. The slide is then handed to the collaborator or to an assistant, who places it bodily into a tube containing a small quantity of four per cent, solution of osmio acid. In the tube the slide remains about ten to fifteen seconds, tightly corked up, in order to fix the trypanosomes with the vapoui’ of osmic acid. Subsequently the slide is fixed with absolute alcohol for about fifteen minutes and stained with Giemsa’s stain in the usual manner.

For the fixation of the coverslip-films we used, in the earlier periods of our investigation, either Schaudinn’s fluid (corrosive sublimate, saturated solution in distilled water, 100 c.c.; absolute alcohol, 50 c.c.; glacial acetic, a few drops) or sublimate-acetic (Hgd₃ saturated in H₂O, 95 volumes; glacial acetic, 5 volumes). Both these fixatives gave results about equally good; it is difficult to choose between them. Latterly, however, we used only Maier’s modification of Schaudinn’s fluid (distilled water, 200 c.c.; absolute alcohol, 100 c.c.; sodium chloride, 1·2 grm.; HgCl₂, 10 grm.), since this appeared to us to give better preservation, and, in particular, less shrinkage of the bodies of the trypanosomes, than the others. The fluid being put into a large watch-glass, the coverslip is dropped into it with the film downwards. The coverslip usually sinks in the fluid and ‘then rests on its corners on the rounded bottom of the watch-glass, so that the film itself escapes any friction or injury. The coverslips are left in the fixative from ten minutes to half-an-hour or longer (the exact time appears to be immaterial), and are then passed through 50 and 70 into 90 per cent, alcohol, where they can be kept until it is convenient to stain them.

The coverslip films were stained almost invariably with Heidenhain’s iron-hæmatoxylin, using 3 per cent, iron-alum solution and per cent, hæmatoxylin-solution, both in distilled water. The filon, after having been brought down through graded strengths of alcohol (80, 70, … 10 per cent.) to water was left about twenty-four hours in the iron-alum, then as long in the haamatoxylin. Immediately before using the hæmatoxylin-solution a few drops of a saturated watery solution of lithium carbonate was added to it, drop by drop, until the solution, when shaken up, was a bright claret-red colour. After the film had been twenty-four hours in the hæmatoxylin-solution the differentiation of the stain was carried out under control by the microscope in a weak (light brown) watery solution of iron-alum. When differentiation was complete the film was washed in a current of tap-water for at least twenty minutes, then rinsed in distilled water and brought up through graded strengths of alcohol to absolute alcohol. At this stage the coverslip was usually dipped for a moment into Lichtgriin-picric solution (Lichtgrün, 1 grm.; picric acid, grm.; absolute alcohol, 100 c.c.), then washed again in absolute alcohol, passed through xylol, and mounted in pure xylol-balsam on a slide. The Lichtgrün stain must be used very rapidly, as it stains intensely.

In this way two preparations were obtained of the contents of each organ—one on the coverslip, the other on the slide— and as a rule trypanosomes were found more or less abundantly on both of them, so that it was possible to compare corresponding phases of the development prepared by distinct methods of technique. It is very important, however, that the operation of removing the coverslip and fixing the films should be performed very rapidly and expeditiously, in order to avoid any drying taking place. The coverslip is particularly liable to dry, since the film of liquid that adheres to it is very thin; the slide, on the contrary, does not dry so quickly. A coverslip that has dried before fixation is quite useless for staining by the iron-hæmatoxylin method; the trypanosomes acquire a characteristic shiny appearance, as if they had been glazed, and when the stain is extracted in order to differentiate the preparation, it does not come out of the cytoplasm evenly, but gives a blotchy appearance, with no sharp differentiation of the nucleus or flagellum. It sometimes happens that a coverslipfilm may be otherwise satisfactory, but may have dried slightly at or near the edges, thus affording opportunities for comparing the effects of desiccation on the trypanosomes with the condition of others that have never been dried. It is then seen that, in addition to the defective staining already described, the trypanosomes are flattened and distorted in various ways.

The fragments of tissue in the dissection adhere, for the most part, to the coverslip; it is not possible, however, to make out anything of trypanosomes which remain within the organs in film-preparations, and it is therefore necessary to tease up the organs well, after dissecting them out, in order to set free the trypanosomes. In the case of those phases which are attached to the gut-wall many remain so attached even when the wall is teased up, but a certain number are usually set free. When such forms are seen in the fresh film they should be dislodged, as far as possible, by tapping gently on the coverslip with a needle.

In some cases a coverslip-film which had been stained with iron-hæmatoxylin was unmounted by dissolving the Canada balsam in xylol, after the trypanosomes on it had been studied and drawn, and the haematoxylin-stain completely extracted by placing it for twenty-four hours in a 3 per cent, solution of iron-alum. The coverslip was then washed for an hour in a current of tap-water, and could then be restained by some other method—for example, Twort’s stain. Trypanosomes that had been already drawn after the haematoxylin-staining could then be drawn again after being stained in a different manner. This double staining did not seem to injure the trypanosomes in any way, but it is noteworthy that after re-staining with Twort’s stain they always came out a little smaller, when re-drawn with the camera lucida, than they had done previously after the hæmatoxylin-stain (compare figs. 260-63, Pl. 42, with figs. 260a-263a, Pl. 38).

When, as sometimes happened, the trypanosomes were so scanty on the coverslip as to require prolonged searching to find them, it was often very difficult to judge the right amount of extraction of the haematoxylin in the process of differentiation by means of iron-alum. Morever, a degree of differentiation which is sufficient for trypanosomes in the thinner parts of the film is insufficient for the thicker parts. Hence it was often necessary to unmount the preparations and differentiate them further, perhaps two or three times, before the right degree was attained. It is difficult to judge of the required differentiation by the fragments of tissue in the films, since the minute bodies of the flagellates give up the stain much more quickly than the relatively thick tissue-cells, and in a preparation iu which the latter are satisfactorily differentiated the trypanosomes become mere ghosts, requiring to be re-stained altogether. The counter-stain with the Lichtgrüu-picric mixture was found to show up the cytoplasm and flagellum of the trypanosomes more clearly.

However carefully the preparations have been made, it is often difficult to make out clearly and with certainty the structural details of some of the minuter phases of the lifecycle, and for this purpose the best optical apparatus was required, both as regards the objectives and the illumination used. All trypanosomes in the permanent preparations were drawn by Miss Rhodes, under our supervision, with the camera lucida at a constant scale of magnification which was as nearly as possible 3000 diameters in the case of the filmpreparations, 2000 diameters in the case of sections.

Our study of the development of T. lewisi in the flea was based principally upon the examination of films, made as described above, but it was found necessary also to cut sections both of the stomach and rectum of the flea. The following is an account of the technique employed by us in preparing sections of the stomach; the same applies to sections of the rectum, the only difference being that the stomachs were cut transversely, the recta longitudinally.

The stomachs of which sections were cut were taken from fleas fed eighteen, twenty-four, or thirty-six hours previously on an infected rat; the fleas themselves had been collected from the non-infected breeding-cage and kept hungry for about three days before being put on the infected rat. The stomach in each case was carefully dissected out from the flea, if possible without puncturing or injuring the stomach in a drop of salt-solution on a slide, and then plunged into the fixative by inverting the slide in such a way that the stomach alone, all other parts of the flea having been removed, was in a hanging drop. If the stomach was ruptured or punctured in the process of extraction it was not, as a rule, preserved, except perhaps as a smear after teasing it up.

A number of different fixatives were tried, but the best results were obtained with Flemming’s fluid¹ and Maier’s modification of Schaudinn’s fluid, and especially with the former. After Flemming the histology of the stomach is extremely good in all details; the blood fills the whole section and is not shrunk away from the wall, and the trypanosomes, both free and intracellular, are well-preserved both in structure and form, and they stain well either with iron-hæmatoxylin or Giemsa, especially the former. After Maier’s fluid the histology of the stomach-tissue is not so good; the cells are shrunk and the minute structure of the nuclei is de Formed. It is evident from a careful study of the preparations that the defects of Maier’s fluid are due to unequal or differential penetration of its constituents; the alcohol evidently diffuses into the tissues first and produces the shrinkage and deformation of the nuclei; the sublimate does not get to the various tissue-elements until they have already been fixed in a defective manner by the alcohol. The blood-débris is also much shrunk after the Maier; while the greater part, sometimes the whole of it, contracts to form a central mass in the section, a certain amount remains usually adherent to the epithelium at the periphery, leaving an irregular empty ringshaped space between the central and peripheral zones of the blood-débris. But to compensate for these disadvantages, the trypanosomes are extremely well-preserved and stain admirably with Giemsa’s stain; some of our stomach-sections prepared in this way are as clear and demonstrative, so far as the trypanosomes are concerned, as any smear or filmpreparation; in fact more so in the case of the large “spheres,” which do not suffer so much from the tendency to opacity which is so disagreeable a feature in the smears. One is here confronted with the extraordinary difference, familiar to everyone who has worked at trypanosomes, between the reaction of these parasites, and that of tissuecells, to the ordinary fixatives and stains used in cytological technique.

Whatever the fixative used, it was allowed to act for about an hour. The stomachs preserved in Flemming were well washed in tap-water and then brought up through a series of alcohols of gradually increasing strength; those preserved in Maier were transferred from it direct to 50 per cent, alcohol. In either case the objects were brought up to 90 per cent. alcohol and there fixed on liver preparatory to being imbedded for section-cutting. Amyloid human liver was used. A moderately thin slice of a block of liver preserved in alcohol was cut by hand with a razor wetted with alcohol, and floated into a shallow glass vessel with a flat bottom, placed on the stage of the dissecting-microscope, and containing 90 per cent, alcohol to the depth of about a centimeter. The stomachs, taken up in a pipette of suitably coarse calibre, were placed on the slice of liver and carefully arranged side-by-side, th eh’ axes parallel to one another and similarly orientated, with their proventriculi all at the same level and all pointing in one direction, their pylori in the opposite direction. Then a tiny drop of glycerine and albumin solution, such as is used commonly for sticking sections on slides, was taken up on the point of a needle and caused to touch the surface of the alcohol immediately above the stomachs. The dense albumin-solution falls at once through the alcohol and spreads out over the stomachs on the liver; at the same time the glycerine is extracted and the albumin coagulated by the alcohol, with the result that the stomachs are stuck to the slice of liver. From six to nine stomachs were thus attached side-by-side on a slice of liver. As the stomachs, before being stuck on, are very liable to roll about or become shifted in position with the slightest disturbance or touch of the microscope, it was found best in practice to put them on not more than three at a time; that is to say, three stomachs having been arranged and fixed upon the liver, three more are then put on beside them. When the required number of stomachs have been stuck on, the slice of liver is trimmed with a scalpel into a rectangular form, in such a way that the longitudinal axes of the stomachs are parallel to the shorter sides of the rectangle; so that by cutting sections of the liver parallel to the longer sides of the rectangle the stomachs are all cut transversely at the same time.

We have thought it worth while to describe the method of fixing the stomachs on liver, although no novelty is claimed, for it,¹ in some detail, as it may not be familiar to some investigators working on similar objects, and because it is a procedure which saves much time and trouble. In the first place, it is much easier to imbed a relatively large block of tissue than a number of separate tiny little stomachs, and the orientation of the objects can be made much more accurate. In the second place, a great economy of labour in the section-cutting and of space in the slides and preparations is effected. To have a number of stomachs cut in the same section diminishes the labour of looking through the preparations under the microscope, and the presence in the section of the slice of liver makes it much easier to go from one section to the next under the high power. Thirdly, with a little experience the liver itself furnishes useful guidance in staining the sections, especially by the iron-hæmatoxylin method; one soon learns what degree of extraction of the stain from the liver-cells gives the best results for the trypanosomes, so that the process of differentiation can be carried out under low powers of the microscope—a great advantage. And finally, since it may be assumed that all the stomach-sections contained in one and the same microtome-section have received exactly the same treatment, it is legitimate to ascribe the very considerable differences seen in different stomachs in the same section to constitutional or functional differences in the stomachs themselves and not to varying local effects of the stain.

The stomachs, after being fixed to the liver in 90 per cent, alcohol, were imbedded in the usual way in paraffin, with a melting point of about 54° C. Methods of celloidin-imbedding were tried, but yielded no advantages to compensate for the extra trouble, especially that of extracting the celloidin from the sections—an indispensable preliminary to staining them. The best thickness for the sections of stomachs was found to be 6μ; with less than that the trypanosomes are too fragmentary. The recta may with advantage be cut thinner than 6 μ, since the crithidial forms are very minute.

Various methods of staining were tried on the sections, but the results of the trials were that wè kept finally in practice to two methods only, namely, iron-hæmatoxylin (Heidenhain), followed by Lichtgriin-picric in absolute alcohol as a counter-stain, and Giemsa’s method. For the iron-hæmatoxylin method the sections were treated first as has been described above for the coverslip-films. The Lichtgriin-picric, which stains very rapidly, was merely washed over the sections for a moment and then washed off again with absolute alcohol. Giemsa’s stain was used, according to the published prescription, as follows: The sections have their paraffin removed, and are brought down to water in the ordinary way. They are then washed in tap-water and put into dilute Lugol’s solution (1 c.c. of Lugol to 25 c.c. of distilled water) for ten minutes. After this they are rinsed quickly in tap-water and put into a 05 per cent, watery solution of hyposulphite of soda for ten minutes. Next they are washed in a current of tap-water for five minutes or longer, and then put into the stain. The distilled water used to dilute the Giemsa-stain has to be neutralised in the way prescribed by Giemsa.¹ The sections were first placed iu fairly strong stain—say, 1 drop of Giemsa to 1 c.c. of neutralised distilled-water—for about an hour, and then were left overnight in a weaker stain—1 drop of Giemsa to 4 or 5 c.c. of neutral distilled-water. The excess of stain is removed by rinsing in water, and after the excess of water has been drained off differentiation of the stain is carried out with different strengths of acetone mixed with xylol, beginning with 95 per cent, acetone used for a very short time, in order to dehydrate the sections and extract the stain, and ending with pure xylol, after which the slides are mounted in dammar or Canada-balsam.

Of the two staining methods principally used, iron-hæmatoxylin gave admirable results after Flemming, especially for the intracellular stages; for the extracellular trypanosomes this stain is not so satisfactory, owing to the fact that the blood-debris, especially in the earlier stages of digesti