The Development of the Amphibian Kidney: PART I: THE DEVELOPMENT OF THE MESONEPHROS OF RANA TEMPORARIA

An extraordinary amount of confusion seems to exist at present as regards the development of the mesonephric units of Anura. Almost every worker who has worked upon the subject has not only put forward a new theory, but has, in most cases, boldly stated this theory to apply to the whole group. This would be the more easy to understand if each author had worked upon many genera; but in the majority of cases only a single genus has been investigated and there has been little or no attempt to bring the new facts observed into line with the statements made by previous authors. Though there should obviously be a general similarity in the development of the mesonephros throughout the whole group Anura, I do not suggest that the very detailed account of Rana temporaria which I am about to put forward is directly applicable to any other genus.

Only one previous paper, that of Marshall and Bles (1890), has been exclusively devoted to the mesonephros of R. temporaria; indeed, the whole genus Rana has been surprisingly neglected, since the two most recent papers are those of Hall (1904), who worked upon R. sylvatica, and Filatow (1904), who investigated R. esculenta. The work of Hall has been so widely quoted, and appears to have been so generally accepted, that it is proposed to use it as the base of this summary.

Hall, then, states that ‘in the region of the mesonephros the mesomers detach themselves from the somite and fuse to form a continuous mesonephros blastema in which swellings are seen. These are the mesonephric blastulae—the fundaments of the mesonephric units’. This view, first advanced by Clarke (1881), was in direct contradiction to the accounts usually accepted at this time. Fürbringer (1887) had stated that the mesonephric units arose as solid ingrowths from the peritoneal epithelium, which were joined by scattered mesenchyme cells before severing their connexion with the peritoneum and becoming differentiated into tubules. This account agreed in the main with those of Spengel (1876) and Goette (1875), save that the latter supposed the original connexion with the peritoneum to he maintained through a ciliated funnel. The accounts given by Filatow agree in the main with those of Hall (with whose work he was obviously unacquainted) and do not merit a separate description.

Hall also confirmed the fact, already stated by Spengel (1876), Nussbaum (1880), Hoffmann (1886), Marshall and Bles (1890), and Farrington (1892), that the peritoneal funnels¹  of the adult open into the blood-system; yet even if we take this fact for granted, we find that much confusion exists as to the manner in which this connexion is formed. The account usually accepted to-day is that put forward by Hall, who supposed the developing funnel and the malpighian capsule to form the two short arms of a Y, whose base opened into the archinephric duct; the attached end of the funnel then broke away and opened into a blood-vessel. He was uncertain as to whether the lumen of the funnel was at any time in continuity with the lumen of the capsule, but inclined to the view that it was not. Spengel (who was the original discoverer of these funnels) supposed them to have had a primary connexion with the malpighian capsule, but found himself unable to trace this connexion at any stage. Nussbaum and Hoffmann merely confirmed the results of Spengel and stated themselves to be in agreement with his conclusions; the latter also pointed out (as is remarked in passing by Farrington) that the capsulo-coelomic connexion through the peritoneal funnel was maintained throughout life by the Urodela.

Marshall and Bles, working upon R. temporaria, pointed out that the funnels opened into the veins at the earliest stage investigated by them (20 mm.); they presumed that there must have been a connexion with the capsule at some stage of the lifehistory, but stated themselves unable to find this connexion. With regard to the development of the malpighian capsules, they stated that ‘in tadpoles of 10 to 12 mm. length the Wolffian tubules arise as little masses of cells in the mesoblast between the aorta and the archinephric duct, a little distance from the peritoneum and quite independent of it. These masses of cells are at first segmentally arranged; they are ill-defined groups of spherical, or slightly branched cells, which rapidly acquire a more definite rod-like shape, then become tubular and, growing outwards, meet and open into the archinephric duct, while at their opposite ends malpighian bodies are formed at a slightly later date.’

We have, therefore, two distinct schools of thought as regards the original formation of the malpighian units. The first school supposes them to be derived from the peritoneum (Goette and Spengel); the second school (Clarke, Hall, and Marshall and Bles) supposes them to arise from mesodermal cells of doubtful origin. These two schools are more or less united in the work of Ffirbringer, in that he regards the proliferated peritoneal cells as being joined by mesenchyme cells; both schools unite in presuming the peritoneal to be derived from some portion of the malpighian unit, from which it severs its connexion in order to open into a vein.

It has long been realized that the kidney of Amphibia is composed of two types of unit whose origins differ; the brief outline which we have just given applies only to the earliest set of units, termed variously the ‘early units’ and the ‘primary units’. The development of the later units has also given rise to two distinct schools of thought. The school which regarded the first of these sets as originating in the peritoneum, supposed the secondary units (as they termed them) to be derived from buds upon the primary units and in this view they were joined by all other workers except Hall. He, however, still further divides the later units into secondary and tertiary derivates of the blastema together with a further set of ‘outer units’. The secondary and tertiary units were derived exactly as were the primary, i.e. as swellings in the solid cord of cells which he termed the blastema, but he frankly states that he is not satisfied as to the origin of the ‘outer units’, but suggests that they are formed by the splitting of already formed units. He makes special mention of the peritoneal funnels, suggesting that in addition to some formed by splitting of primary funnels, others might be produced from cells already in situ; from whence these hypothetical cells could have come he does not say.

The only other worker who seems to have regarded the adult funnels as presenting a special problem is their discoverer Spengel. He endeavoured to show that these funnels, about whose origin he is doubtful, maintained throughout life a connexion with some portion of a malpighian unit. He found himself, however, quite unable to prove his point, though in one case he has recorded a long ciliated tube, the free end of which was applied to the peritoneal epithelium and whose dorsal end appeared to lead to the collecting trunk. (In common with other workers he termed the ‘collecting trunk’, that portion of the primary tubule from which had been given off the secondary tubules.) The presence of this long ciliated tubule is of particular importance in view of the confirmation it gives to our own account of the development of these later peritoneal funnels (vide subter).

The tadpoles and newly metamorphosed frogs used in this work were all bred in the laboratory from R. temporaria obtained in the London district; post-metamorphic stages were captured in the same district. The tadpoles were graded into stages according to the length from the tip of the snout to the tip of the tail, and it was found, under the artificial conditions employed, that the following lengths, ages, and conditions corresponded.

These stages are singled out for special mention since they are those stages to which reference is made in the text and in the plates; a much more complete series was, of course, investigated.

The material was fixed for about 12 hours in Bonin’s picroacetic-formol and then washed in 70 per cent, alcohol till all the yellow colour was removed. Embedding, after dehydration in 90 per cent., 95 per cent., and 99 per cent, alcohols and clearing in oil of cedarwood, was carried out in 56 degrees paraffin wax and 10 μ. sections cut in transverse, frontal, and sagittal planes. The best stain for the sections was found to be Delafield’s haematoxylin, with subsequent ‘blueing’ in ammonia vapour. In the differentiation of the different parts of the more advanced kidneys Pacini’s triple stain was found useful, but it tends to obscure histological detail. The sections were without exception mounted in Gurr’s neutral balsam.

Only one difficulty was encountered—the tendency of tadpoles from 10 to 30 mm. long to take grit into the alimentary canal; in order to obtain good sections of these stages it was found to be absolutely essential to remove the entire gut from the larva before embedding.

The reconstructions were carried out by ordinary graphic methods and subsequently shaded, with the aid of the actual sections, to simulate relief.

If we examine the mesonephric region of a 10 mm. tadpole, we notice that there is an irregular mass of cells occupying a tract of the retro-peritoneal tissue along the dorso-medial wall of the archinephric duct; this tract of tissue is the ‘blastema’ of Hall (vide p. 508). It seems to me to be highly improbable that it should, as he states, have been cut off as a solid mass of cells at the time of the division of the mesoderm into dorsal and ventral portions, and I am far more inclined to Ffirbringer’s view that it is an agglomeration of specialized mesenchyme cells; that these mesenchyme cells represent that portion of the mesoderm known as the intermediate cell-mass is, of course, obvious, since in every known vertebrate it is from this portion of the mesoderm that the excretory system arises.

The development of the early units from this blastema is essentially as described by Marshall and Bles; but since their account occupied less than a hundred words, very little excuse seems to be necessary for presenting a more detailed description.

A remarkable fact, for which I can find no previous mention, is that the early units develop asymmetrically on each side of the animal. In every case which I have examined, the mesonephric units of the right side are far less developed than those of the left side. It is true that all the larvae examined by me are from the London district, but I do not consider it possible that such a condition should be merely a local abnormality, and I am at a loss to explain why the fact has not been recorded by earlier workers. This asymmetry is of the utmost value to the investigator since it permits of the tracing of the whole history of the unit in a single tadpole. Text-fig. 1, for example, is a complete reconstruction of both mesonephroi from a 17 mm. tadpole; every stage in the development, from the first nephroblast vesicle (vide s u b t e r for the explanation of this term) to the much-coiled unit, is there shown, but in order to get a better idea of the early development it is necessary to discuss, first of all, the condition in a rather younger tadpole.

The examination of a series of sections through the mesonephric region of a 13 mm. tadpole shows that a series of condensations have taken place in the mesonephric blastema—for reasons which will later become apparent I propose to term these condensations the ‘nephroblast vesicles’-¹  Each consists, on the average, of about seven or eight cells, loosely pressed together and whose position bears no constant relation to the segmentation of the animal, since there are some six or seven such vesicles extending over the space of five or six segments. The vesicles are in no way joined one to the other, but are quite distinct condensations within the main irregular mass of the blastema; it must further be clearly understood that the number of cells entering into these condensations form only a quite small proportion of the total number of cells in the blastema, and the nephroblast vesicles of R. temporaria are not, therefore, identical with the ‘hlastulae’ of Hall which were stated to be regularly placed swollen nodes connected by a smooth cord of cells.

All the nephroblast vesicles arise about the same time but develop rather more rapidly in the posterior than in the anterior region; it has frequently been stated, in this connexion, that the mesonephros forms in a regular manner from the posterior to the anterior. This is a misuse of the word ‘form’: the fundaments of all the vesicles are formed at the same time, but they develop rather more rapidly in the posterior region. This statement is entirely relative, for a glance at Text-fig. 1 shows that there is no uniform gradation in stage from the most anterior to the most posterior unit. The right mesonephros (shown on the left-hand side of the figure) has the eighth unit (R 8) most highly developed, while in the left hand the seventh, eighth, and ninth units (L 7, L 8, and L 9) are at an almost equally advanced stage of development; again, the fifth unit on the left kidney (L 5) is far in advance of the more posteriorly placed sixth unit (L 6).

When the nephroblast vesicles are first formed they are roughly spherical in shape (L 1, L 2, R 1, and R 10 on Textfig. 1, for example), but soon become oval (R 2) owing to the elongation of the cells of which they are formed. This oval vesicle then develops a lumen and commences to grow rapidly at each end. The end nearest to the archinephric duct forces its way into the wall of the latter so that the lumen of the vesicle becomes continuous with that of the duct; this is diagrammatically represented in Text-figs. 2,A and 2 B. While this connexion is being formed the other end of the oval vesicle is growing downwards round the wall of the archinephric duct; the proliferation of cells at this growing tip, moreover, has not only caused the lengthening of the vesicle but has also given rise to a slight dilatation of the tip, so that at this stage the developing unit presents the subclavate shown at R 2,Text-fig. 1. A section which is transverse to the archinephric duct will, at this stage, cut the tip of the developing unit obliquely; such a section is shown at fig. 1, Pl. 27, and it will be noticed that the nephroblast vesicle (n.v.) is still surrounded by blastema cells (b) and does not, as stated by both Ffirbringer and Marshall and Bles, lie freely in the connective tissue.

So far only the cells comprising the two ends of the developing vesicle have increased in number, but after the connexion with the archinephric duct has been formed, the cells of this end remain quiescent while those of the central and terminal portions continue to divide rapidly, so that the vesicle becomes elongated into an embryo tubule; but the cells of the central portion do not divide equally on both sides of the duct and the terminal portion becomes bent back upon the remainder into the form shown at L 4,Text-fig. 1. The cells of the tip, however, continue to divide evenly and rapidly so that a mass of cells is formed which, upon the inner side, soon commences to push inwards and obscure the lumen; this mass is the rudiment of the malpighian glomerulus and is represented at FG,Text-fig. 2 c. This ingrowing portion continues to increase rapidly in size (Text-fig. 2 D), its cells become reorientated and it finally breaks away from the capsule as the completed glomerulus (Text-fig. 2 B). This breaking away is not entirely complete, but a narrow strand, representing the blood-connexion, remains in connexion with the capsule; finally, the increase in size of the glomerulus stretches the walls of the capsule until they assume the membranous condition shown in figs. 4 and 5 on Pl. 27.

While this differentiation has been going on within the capsule, the increase in length of the tubule has carried the malpighian capsule at its tip into a position alongside the peritoneal wall of the kidney; the continued division of the tubule cells can therefore no longer give rise to an increase in length, but must result in coiling. This coiling, which is perfectly regular, takes place in the following manner.

The first bend, as already stated, is such as to cause the whole unit to become recoiled upon itself. The second is similar, but in a different direction, to the first, so that an ‘S’-shaped bend is produced; this bend is, of course, the well-known ‘Henle’s loop’, and is extremely well illustrated by the condition of the units R 8 and L 10 in Text-fig. 1. The distal (that is to say, farther from the archinephric duct) curve of the S bend now sends out a loop parallel to the plane of the archinephric duct; the resultant form, which is far easier to figure than to describe, is shown by the unit L 6. The section in fig. 3, Pl. 27, is across the region indicated on L 6,Text-fig. 1, and shows the rapid cellular division from which this bend takes its form. The proximal curve of the original S now pushes rapidly towards the median plane of the animal, and since there is little, or nothing, to impede its growth, it pushes farther than the bend last described so that this medially directed loop becomes rather more drawn out, and in consequence rather thinner, than the other portions of the tubule; this thin loop is well shown in fig. 6, Pl. 27, and in the unit L 7 in Text-fig. 1. The remaining central portion of the original 8 now elongates and becomes bent into the form of the Greek letter Σ; these bends are shown in the units L 7 and L 9,Text-fig. 1. Beyond this point the coiling becomes quite irregular, since the units have now attained such proportions that they tend to press one against the other with a resultant modification of the position of their coils.

It will be noticed, on reference to Text-fig. 1, that the anterior vesicles are shown as lying against the dorso-medial wall of the archinephric duct while the more developed posterior units are shown as inserted ventrally; this is accounted for by the ‘rotation’ of the archinephric duct. We place the word rotation between quotation marks since this word is far more applicable to the result than to the means by which it is produced; actually the cells of the dorsal wall of the duct divide so that the area occupied by this wall is increased and, without any actual movement of the duct, the apparent orientation of its walls is thus altered, what was originally dorsomedial being pushed round until it becomes ventral. This apparent rotation of the duct continues until the point of insertion of the original units comes to occupy the lateral wall; the significance of this change becomes apparent when we deal with the development of the later units.

Now let us turn our attention to the development of the peritoneal funnels of the early units; these are not shown in the reconstruction since they would merely serve to obscure the outlines of the units.

When the growing tip of the unit reaches the peritoneum there is found to be a small roundish mass of cells lying between the two; these cells are the rudiment of a funnel. Whether these cells are blastema tissue which has been pushed down along with the tip of the unit, or wether they are a product of this latter, it is impossible to say. I am inclined to think that the former is true and that the funnel is therefore differentiated from the rest of the units at the very commencement of its lifehistory; at any rate, it may be quite definitely stated that this mass of tissue lies at the tip of the developing malpighian capsule and never, in R. temporaria, occupies the position indicated by Hall for R. sylvatica. This mass, which is shown diagrammatically at FN,Text-fig. 3 A, then becomes elongated into a spindle-shaped tubule, closed at both ends; this is represented in Text-fig. 3 B. The plane of this spindle-shaped tubule is parallel to that of the peritoneum—that is to say, transverse to the plane of the long axis of the malpighian capsule; we speak in terms of planes since the actual orientation within this plane appears to be a matter of chance. Both Hall and Marshall and Bles lamented the fact that, in the words of the latter, ‘transverse sections do not cut the nephrostomes in a favourable plane for observation’. This is perfectly true—but it would be equally true were the statement made of frontal or sagittal sections. There is absolutely no means of foretelling in what plane any particular funnel will lie; it is largely this fact which causes me to believe that the funnel rudiment is early differentiated from the malpighian capsule; for if it were the product of the latter, one would expect to find some form of constant relation between the two. A transverse section of the spindle condition is shown at f.n.,fig. 2, Pl. 27, and it will be noticed that the developing funnel is only very loosely applied against the wall of the malpighian capsule.

The lumen of the spindle now increases in size, as do the cells of its wall, these latter acquiring cilia. The lumen then forms a connexion with the coelom and with an outgrowth from a neighbouring vein. Fig. 4, Pl. 27, shows the connexion (o.n.) between the coelom and the funnel, while fig. 5 on the same plate shows the connexion of a funnel with a blood-vessel (b.v.); both these connexions appear to be formed about the same time. There is no doubt whatever that the lumen of the funnel is never in continuity with that of the malpighian capsule; for, as we have already stated, the lumen of the developing funnel lies in a plane at right angles to that of the developing malpighian capsule, so that this latter only comes in contact with the central portion of the funnel wall.

Now so far we have only traced the development of the vesicle; the cells of the blastema develop as rapidly as do those of the vesicle. Thus in fig. 1, Pl. 27, the blastema (b) consists of a few cells lying around the nephroblast vesicle. A section inter mediate between two vesicles would show a solid blastema mass occupying the same area as is occupied, in the section mentioned, by both the vesicle and the blastema put together; in short, as has already been pointed out, but cannot be too heavily em-phasized, they are not condensations of the blastema, but in the blastema.

As the units develop, therefore, they are always surrounded by the growing blastema; but the increase in the space occupied by the growing units tends to separate the main mass of blastema from the region of the archinephric duct, as is shown in the series of sections, figs. 1 to 6, PL 27. In figs. 1, 2, 3, and 4, Pl. 27, the archinephric duct (a.d.) still retains a connexion with the blastema, but in 5 a tubule has forced its way up between the two, and in 6 two tubules occupy this position. The blastema does not become separated from the archinephric duct along its entire length, but remains connected to it by five or six ‘straight tubules’ whose origin is as follows.

An examination of fig. 4, Pl. 27, shows, at the dorso-medial angle of the archinephric duct, a small mass of cells (f.st.) which at first sight appears to form a nephroblast vesicle; this mass of cells is the rudiment of one of the straight tubules. Each develops as if it were a nephroblast vesicle up to the point at which the connexion with the archinephric duct is formed. At this stage, it will be remembered, the developing unit has the form of a short tubule whose end is swollen in a manner suggestive of a malpighian capsule; but in the case of the straight tubule this malpighian capsule never becomes perfected, neither does the tubule ever become coiled. It remains as a straight connexion between the archinephric duct and the blastema mass, increasing in length as the distance between the two increases. The development of this straight tubule is diagrammatically shown in Text-figs. 6 A to 6 c, and will be given in more detail when we come to discuss the development of the later units.

To sum up, then, the development of the kidney from 12 to 17 mm., we may say that about (varying in different individuals) ten units are formed upon each side, each unit being derived in its entirety from a nephroblast vesicle, which is itself a condensation in the blastema. At a slightly later date five or six vesicles arise which do not develop into perfect units, but remain as straight tubules running from the dorso-medial blastema to the dorso-lateral archinephric duct. Further, the peritoneal funnels, which never have any connexion with the lumen of a malpighian capsule, develop from masses of tissue lying between the malpighian capsule and the peritoneum and form a clear connexion between the coelom and the blood-system.

i. General.—

The development of what I propose to term ‘later units’ (secondary, tertiary, and outer units of Hall) commences at about 18 mm. and continues up to about two years. The processes which are about to be described have been worked out from sections of larvae varying between these two ages, but the actual sections figured are for the most part from 20 mm. and newly metamorphosed tadpoles. These two stages have been selected solely because the development of the units appears to be carried on most actively at these times; almost any stage in the development of either a malpighian unit or a peritoneal funnel may be found in any specimen between the two ages mentioned.

In order clearly to understand what follows, it is essential to realize the general structure of the mesonephros of a young tadpole; Pl. 28 shows photographs of transverse (fig. 7, Pl. 28) and obliquely parasaggital (fig. 8, Pl. 28) sections of a 20 mm. stage. The asymmetric condition of the mesonephros is very clearly shown. The blastema (b.) occupies (in fig. 7, Pl. 28) the dorso-medial angle of both mesonephroi; running down from this in a ventral direction is a ‘string’ of three malpighian glomeruli, of which the dorsal is the least developed, the middle one (m.c.) clearly recognizable, and the lowest one in an almost perfect condition. Lying against the side of this last is a peritoneal funnel whose outer opening (o.n.) connects with the coelom, and whose inner opening (i.n.) is clearly shown as leading into the large blood-vessel (b.v.). Following this latter round in a ventro-medial direction to the ventral margin of the kidney, we find two tubules (nst.t.) actually lying within the bloodvessel. Just to the left of these two tubules there is a further malpighian capsule whose glomerulus is very small and has no connexion (either in this or in any other of the neighbouring sections) with the blood-supply; from this, and from the fact that the walls of the capsule are being crushed in by the surrounding tubules, we may presume that the glomerulus is degenerating. Returning once more to the dorsal side of the mesonephros, we see that the archinephric duct (a.d.) has now become pushed a considerable distance from its original position, towards the lateral margin of the kidney; the duct bears upon its ventral surface a small conical mass of cells which, when seen under a higher magnification (Text-fig. 4), shows a glassy protoplasm and irregular nuclei—which is, in fact, a typical mass of degenerating tissue. The remainder of the area of the kidney is occupied by large tubules whose lumina appear to be partially obscured by a fine reticulum.

Turning, for a moment, to the longitudinal section (fig. 8, Pl. 28), we see that the blastema (b.) (in that portion of the section in which it is cut) lies along the upper margin of the kidney as an irregular band of darkly staining cells, from which hangs down, at one point, a string of two malpighian capsules with their glomeruli. The more ventral (and the more highly developed) of these two glomeruli is associated, at its tip, with a funnel (n.). At about the centre of the dorsal margin of the section there is a reorientation of the blastema cells into a form strongly suggesting a downwardly growing tubule (nst.v.). Owing to the plane in which the section lies, the archinephric duct is not cut at any point, but at s.t. there is a darkly staining duct (cut in transverse section) which may be easily identified as one of the ‘straight tubules’ to which some reference has already been made; growing down diagonally from the straight tubule, there is a further mass of cells (o.s.t.).

Thus it will be seen that the kidney at 20 mm. differs sharply from that at 15 mm. in many particulars. The malpighian capsules at this latter stage are arranged at approximately equal intervals in a longitudinal direction; at 20 mm. they appear to be arranged in a series of dorso-ventrally hanging ‘strings’. Further, tubules are now present which actually lie in a bloodvessel. There also remain to be explained the lobes of degenerating cells on the archinephric duct and the apparently degenerating malpighian glomerulus.

All these differences have been remarked from the consideration of two isolated sections; let us now further investigate these differences by means of a reconstruction (fig. 9, Pl. 29). In this plate the general mass of black represents the outline of a portion of the mesonephros of a 20 mm. tadpole; on this have been superimposed the peripheral blood network (red), a malpighian capsule (yellow), the archinephric duct, and ‘straight tubule’ (blue), a peritoneal funnel (green), and the tubule (also green) which in the section was seen to be inside a blood-vessel. It has been thought desirable only to include one nephrostome and one malpighian capsule within the reconstruction in order to avoid the confusion which would inevitably arise were all the malpighian capsules with their intercoiling tubules shown.

It now becomes evident that what we have so far referred to as the straight tubule is the ‘collecting ‘trunk’ of previous writers. Its development, which will be later given in more detail, has already been briefly outlined, and in the example reconstructed there are no features of particular interest attaching to it. The peripheral blood network is, so far as we know, of a form common to all immature mesonephroi and presents no feature of special interest.

The malpighian unit, on the contrary, is altogether extraordinary, since a reconstruction of its tubule (also yellow) shows this latter to end blindly. I would at this point categorically state that I have never found any tubule, other than the straight tubule or collecting trunk, opening into the archinephric duct in any tadpole of more than 20 mm. length. What, then, has become of the early units which we left in an apparently functional condition at 17 mm.? This question is, in my opinion, fully answered by the little masses of degenerating tissue which, at 20 mm., we found adhering to the archinephric duct; that is to say, I regard these masses of tissue as the points at which the early units have severed their connexion with the archinephric duct. I have been fortunate enough, in one instance, to find an abnormal case (in a metamorphosing frog) in which the primary tubule had not severed its connexion. Text-fig. 5 is a photograph of this case as seen in transverse section and it will be noticed that the rapidly degenerating connexion (DC) is upon the opposite side of the archinephric duct (AD) to the straight tubule (ST). Now it will be remembered that the archinephric duct ‘rotates’ upon its axis (vide p. 518), and this one section therefore bears out the two new points which I have so far introduced into this account, viz.:

i. That the straight tubule or collecting trunk has no relation at all to the early units; since if it had it would have to be inserted upon the same side of the duct as are these units.

ii. That the early units sever their connexion with the archinephric duct; since it is well known that this duct in the adult condition runs along the outer edge of the mesonephros and the position where were inserted the early units would therefore come to lie along that wall of the duct which is farthest from the main mass of the kidney—a reductio ad absurdam.

There remains, in the reconstruction which we are at present describing and in the two sections which we have already described, the ‘tubules lying within the blood-vessel A reconstruction of this tubule (green in Pl. 29) shows it to have one extremity ending blindly in the dorso-medial blastema mass and the other end, which is internally ciliated, ending blindly just outside a blood-vessel; the coils of the tubule between these two extremities lie actually in the course of a bloodvessel. Such parts of the tubule as do not actually lie within the vessel are, in the tadpole, embedded in the general blastema which still (20 mm.) surrounds the developing tubules and therefore forms the ‘interstitial tissue’ of the mesonephros. The presence of this ‘interstitial tissue’ will be the more easily understood if we emphasize once again that the original nephroblast vesicle was surrounded by a mass of blastema and that this blastema has grown proportionally with the tubules.

To sum up, then, what we have learned from both sections and reconstructions of the mesonephros of a 20 mm. tadpole, we find:

i. That the archinephric duct is furnished at intervals with collecting trunks.

ii. That none of the malpighian units appear to have any direct connexion with the archinephric duct.

iii. That the greater part of these units appear to hang down in dorso-ventrally directed strings.

iv. That there is a curious set of tubules, ending blindly at both ends, which more or less closely follow the interior of the blood-system.

Let us examine these points in the order in which we have given them.

ii. Collecting trunk.—

Though we have already, in order to link up the development of the early with the later units, given a brief outline of the history of the trunk, it would be as well to describe this in more detail.

All workers, from Fürbringer onwards, have noted the existence of the collecting trunk, but all have, without exception, supposed it to be derived from one of the early units. In this connexion it is significant to note that all these workers, save only Marshall and Bles (who regarded the collecting trunk as an accomplished fact), have worked upon other forms than R. temporaria, and I am perfectly prepared to admit that the description which follows may be but another instance of the highly modified developmental history exhibited by this form.

In R. temporaria, then, the first origin of the collecting trunk is a small spherical condensation in the dorso-medial blastema mass and is, at first, indistinguishable from those from which the early units are derived. These condensations usually appear at about the time when the first of the early units has attained its highest development—that is to say, at about the stage reached in the reconstruction, Text-fig. 1. There are five or six such vesicles and there appears to be no definite segmental arrangement nor yet any relation in position to the early units. One of these vesicles, as already noted, is shown at f.st, fig. 4, Pl. 27. This vesicle elongates as the archinephric duct is pushed away from the blastema and at the same time develops a swelling at its tip; fig. 15, Pl. 31, shows a section of this stage, f.st. being the fundament of the straight tubule. I have no doubt whatever that this straight tubule is a modified unit—modified, that is, long before it becomes in any way comparable to a unit—and we may therefore safely regard the swollen tip as the vestige of what was once a malpighian capsule. The plane of this section shows the abortive malpighian capsule as separate from the fundament of the straight tubule, as there is, in this section, a slight upward bend of the tubule at this point.

Very shortly after the stage figured has been reached the elongated vesicle acquires a lumen and forms a connexion with the archinephric duct. As this latter is pushed farther and farther from the main mass of the blastema so does the straight tubule increase in length until (fig. 16, Pl. 31) it extends across the whole width of the dorsal surface of the mesonephros. The lumen of the tubule never penetrates the abortive malpighian capsule, which retains, at any rate up to two years, the appearance of a solid ball of cells.

So much for the actual tubule; we have yet to explain the mass of downgrowing cells (o.st.) shown, in fig. 8, Pl. 28, as attached to it. These outgrowths appear at irregular intervals along the length of each straight tubule arising as slight swellings upon the wall of the latter. These thickenings grow outwards as buds which continue to elongate and, when they have attained a length of about twice the diameter of the straight tubule, acquire lumina which become contiguous with that of the straight tubule. This condition is very clearly shown in fig. 17, Pl. 31, where st. is the straight tubule and o.st. the bud; the abortive malpighian capsule is also shown (a.m.c.). This whole course of development is shown in Text-fig. 6.

We are left, therefore, with a straight tubule, one end of which ends blindly in a mass of cells representing an ancestral malpighian capsule and the other end of which opens into the archinephric duct; at irregular intervals along the straight tubule there are buds. In order to understand the function of these buds we must first of all discuss the second point given in the summary above—the development of the malpighian capsules and glomeruli of the later units.

iii. Malpighian capsules and glomeruli.—

The main blastema mass of a 20 mm. tadpole consists, as we have already stated, of a continuous mass of tissue along the dorsomedial angles of both mesonephroi. In this blastema there have already arisen the nephroblast vesicles, giving rise to the early units, and a set of some six further vesicles which give rise to the straight tubules.

There now arise, intermediate between, but not segmentally related to, the early units, another set of vesicles which we propose to term ‘capsuloblast vesicles’;¹  the explanation of this name is afforded by the fact that these vesicles give rise to the later malpighian capsules and glomeruli. Each capsuloblast vesicle, when it first arises, is identical with one of the early nephroblast vesicles—i.e. it is a spherical condensation of some seven or eight blastema cells. The blastema cells immediately surrounding this condensation then withdraw slightly so as to leave a small vacuole about the lower hemisphere of the vesicle. This stage is shown in fig. 10, Pl. 30, where b. is the blastema and cp.v. the capsuloblast vesicle.

The cells comprising this latter then multiply very rapidly till a solid, compact sphere of cells has been formed which may now be clearly distinguished, both by their smaller size and by their darker staining reactions, from the surrounding blastema. Such a condition is shown in fig. 11, Pl. 30, where it will be noticed that one of the central cells of the capsuloblast vesicle is actually in a condition of mitotic activity. While this increase in the number of the vesicle cells has been going on, the surrounding blastema cells have still farther withdrawn so that the condensation is now entirely cut off from the blastema save by a narrow neck of cells. This is the first stage at which the condensation may be clearly recognized as an embryo malpighian unit. The cells of the vesicle now become reorientated into a series of concentric spheres surrounding a solid inner mass. These concentric spheres are the fundament of the glomerulus; why this concentric sphere arrangement should take place I do not know. There is no suggestion of this formation in the perfect glomerulus, and indeed this form is soon lost by the continued rapid multiplication of the cells. It is at this stage (fig. 12, Pl. 30) that one begins to find traces of the formation of the tubule, in the blastema cells which form the boundary of the vacuole in which lies the capsuloblast vesicle. There is a tendency on the part of these cells to become elongated and reorientated into the form of an investing sheath (f.k., fig. 12, Pl. 30) which becomes more thickened and clearly marked out as development proceeds. One portion of this sheath (that actually marked by the line f.k.) becomes especially thickened and later grows out as the actual tubule.

The next change observed is in the formation of the bloodsupply to the glomerulus. This is formed by an outgrowth (&.s., fig. 13, Pl. 30) from the cells originally comprising the capsuloblast vesicle, which connects with a blood-vessel. At the same time as this is occurring the fundament of the tubule increases rapidly in size, acquires a lumen, becomes elongated into a regular tubular form, and commences to coil. The vacuole (which is now, of course, the cavity of the malpighian capsule) increases in size and the glomerulus assumes a typical form. The completed malpighian unit has been too well described to need re-description here, but in order to render the series of illustrations complete it is shown in figure 14, Pl. 30.

So far we have only dealt in detail with the capsule and the glomerulus; there remains the tubule. If this description has been followed, it will have been realized that the capsule and glomerulus are formed in a portion of the kidney which is now widely separated from the archinephric duct. How, then, can the connexion between the two he acquired ? The answer is that as the tubule grows out it turns in an upward direction and fuses with one of the buds which has itself grown out of the straight tubule. Thus a mental picture of the malpighian units at this stage shows us a dorsally running transverse straight tubule from which hangs by the malpighian tubule, as fruit hangs from a branch, a malpighian capsule.

Now all this description has dealt with the formation of a single unit and tubule; we have not yet furnished any explanation of the phenomenon previously remarked—that the capsules and their contained glomeruli appear to hang down in ‘strings’ from the dorsal blastema. The explanation of this phenomenon is that the capsuloblast vesicles do not arise irregularly throughout the whole length of the blastema but only in certain definite tracts, each tract being apparently capable of giving rise to an indefinite number of units. As each capsuloblast vesicle develops, the growth of the blastema mass above it causes it to be pushed ventralwards; thus a clear tract of blastema is formed immediately dorsal to the vesicle and in this tract a further vesicle arises, which is in turn pushed downwards. This second vesicle is now pushed downwards to make room for a third, the third pushed down to make room for a fourth, and so on. The development of the individual vesicle appears to take place at such a rate that the unit assumes its final form at the time when it has reached the ventral wall of the mesonephros; this rate of development is therefore variable, since the older the mesonephros, the farther will the vesicle have to be pushed before it reaches the ventral wall. Thus in Pl. 28 the mesonephros shown in transverse section is rather more advanced than that shown in longitudinal section; in both kidneys the ventral capsule is at the same advanced stage of development, but in the younger example the string is only composed of two capsules while in the elder it is composed of three.

The course of development which we have just described is summed up diagrammatically in Text-fig. 7. The capsuloblast vesicles (CF, 7 A) arise in the blastema (B, 7 A), each vesicle descending as it develops and further vesicles arising above it. An outgrowth (OT, 7 B) appears on the straight tubule (ST, 7 B), and the developing capsule sends out a tubule (FT, 7 c) which meets and fuses (7 D) with this outgrowth.

It will be noticed in 7 D that the lowest malpighian vesicle does not actually lie in contact with the ventral wall of the kidney (WK, 7 D), but is separated from it by a small mass of blastema (FN, 7 D); this mass of blastema is the fundament of one of the later peritoneal funnels, to whose development we will now proceed.

iv. Peritoneal funnels.—

There are two quite distinct methods of formation of the later funnels. The first method, which is only found in quite young (20 to 22 mm.) tadpoles, is analogous with the formation of the funnels of the early units. Each string of malpighian capsules bears at its tip of blastema cells which, when they reach the wall of the kidney, reorientate themselves into a spindle which acquires a lumen and later opens both from the coelom and to a blood-vessel. Text-fig. 3, in which is diagrammatically shown the development of the peritoneal funnel of an early unit, displays with equal accuracy the changes which take place in the course of the formation of one of these later funnels.

Now it is obvious that the funnels produced in this manner will be numerically fewer than the malpighian units; but it has been frequently recorded that the funnels in an adult kidney outnumber the malpighian units by three or four to one. Two suggestions have so far been advanced to account for these later funnels: (i) that they are formed in situ from the wall of the peritoneum, or (ii) that they are produced by the division of already existing funnels. I have observed nothing in any of my sections which in any way bears out either of these suggestions; in my opinion the peritoneal funnels are produced in the following manner.

If we revert, for the moment, to our examination of the blastema of a 20 mm. tadpole, it will be remembered that not only are there capsuloblast and straight tubule-forming vesicles produced, but also a further set of condensations (of which one is shown at nst.v.,fig. 8, Pl. 28) for which we have so far recorded no function. This vesicle (which for reasons later becoming apparent will in future be referred to as the vesicle of a ‘funnelforming tubule’)¹  develops in a manner utterly different from that of any of the other vesicles.

The original spherical shape of this vesicle soon becomes modified to that of a hemisphere by the excessive growth of one side of the vesicle and the consequent crushing of the other side against the upper margin of the kidney; this hemispherical condition is apparent in the section already quoted. The rapid growth of the lower side then continues so that the vesicle assumes the form of a blindly ending tubule; this slightly later stage is represented at nst.v., fig. 10, Pl. 30. The tubule grows rapidly in length, pushing downwards towards the lower margin of the kidney; shortly before it reaches this margin, the tip of the tubule enters the peripheral blood-system by forcing its way through the inner wall. We have here, therefore, the explanation of the tubules which have already been noted as lying within this blood-system. The sole function of these tubules is to give rise to peritoneal funnels; since both the existence of these specialized tubules and the function subserved by them is here noted for the first time, it is obvious that a name must be found to describe them and we propose to employ the term ‘funnel-forming tubules’¹  The actual method of funnel production from these tubules is as follows.

When the tip of the tubule has accomplished about half of its journey towards the peripheral blood-system, it becomes internally ciliated; by the time that the tip has penetrated the blood-vessel the whole lower third of the tubule has acquired these internal cilia. Shortly after its entry into the bloodsystem the tip of the tubule lays itself parallel to the outer wall of the blood-vessel—that is, to the outer-wall of the kidney. This orientation of the tip is shown in fig. 18, Pl. 31, where nst.t. is the ciliated tip. That wall of the tubule which lies against the periphery of the blood-vessel then breaks down (at the point bn., in fig. 19, Pl. 31), and in this is soon followed by the inner wall of the tubule; in short, the whole tip detaches itself from the tubule and lies against the outerwall of the kidney. This wall is not, as it is usually figured, composed of a thin sheet of squamous epithelium, but is strongly reinforced in many places by agglomerations of the blastema cells; it is to these blastema cells (b., in the two figures already quoted) that the now severed tip becomes attached. This severed tip, which, it will be remembered, is internally ciliated, may now be termed a ciliated funnel.

We have then, at this stage, a funnel which opens internally to a blood-vessel but whose outer extremity ends blindly in a mass of blastema cells. This blindly ending tip then forces its way to the surface of the kidney where, as in the case of the funnels produced by other methods, a coelomic connexion is formed.

Now let us return to the funnel-forming tubule. The end, from which has been separated the funnel, soon heals, but even before this healing has taken place there is an exceedingly rapid growth of the cells comprising the wall of the tubule which lies farthest from the periphery of the kidney; this rapid growth causes the tubule to turn sharply back upon itself, as is shown in the tubule reconstructed on Pl. 29 where that portion of the tubule most nearly approximated to the funnel presents a ‘V’shaped formation. This ‘V’ shape is very typical, being present in every funnel-forming tubule which I have reconstructed and forms, in my opinion, one of the most convincing arguments in favour of the method of funnel production just described. For there appears to me to be no good reason to account for a sudden directional change in an otherwise smoothly coiled tubule, other than that the tubule has been damaged at the point at which this directional change occurs; such damage is more than sufficiently accounted for by the detaching of the entire tip.

The typical appearance presented in section by a newly formed funnel and its attendant funnel-forming tubule is shown in the series of sections (figs. 20 to 22) on Pl. 31, which are taken at about 25 μ intervals from a uniform series. Fig. 20, Pl. 31, shows the internal opening of the funnel to the blood-vessel; fig. 21, Pl. 31, shows this funnel running through the centre of a mass of blastema (b.) from whose edge the funnel-forming tubule (nst.t.) has not yet become entirely separated; fig. 22, Pl. 31, shows the external opening of the funnel from the coelom as well as the apparently double funnel-forming tubule, this latter having been cut just behind the tip of the ‘V ‘shaped bend.

Each funnel-forming tubule gives rise to a large number of funnels; for obvious reasons it is impossible to say exactly how many. It is also, unfortunately, impossible to say for how long these tubules continue to exist. They first appear about 20 mm. and are both present and extremely active at metamorphosis; beyond this stage it becomes almost impossible to trace any individual tubule through a number of sections, and though I can quite definitely state that I have found no trace of these tubules in a frog about 18 months old, I would not care to be held responsible for the statement that they are not then present.

The method of funnel production which has just been described is summed up diagrammatically in Text-fig. 8. 8 A shows the vesicle of the funnel-forming tubule (FNT) lying in the blastema (B) and already commencing to grow downwards towards the lower wall (W) of the kidney, against which lies the blood-vessel (V). 8 B shows the ciliated tip (TNT) of the tubule (NT), while in 8 c this tip has penetrated the blood-vessel and arranged itself parallel to the outer wall. At 8 D the tip has become detached as a funnel (N), which latter is shown in its final condition in 8 E.

To sum up, then, all that we have learned of the formation of the later funnels, we may say that:

i. There are two methods for the production of such funnels;

ii. The first set arise one at the end of each string of malpighian units;

iii. The second set are produced from tubules modified to subserve this function;

iv. No peritoneal funnel ever opens into a malpighian capsule.

It would seem at first sight that the results here published differ widely from those of previous writers; but if a more thorough comparison be made, it will be found that such new facts as have been introduced into this description serve to fill some of the more obvious gaps left in former works. Let us, then, take each point in the order in which it has been made and contrast it with other accounts.

We come, first of all, to the early units. These develop in a manner common to many vertebrates and which has already been roughly outlined for R. temporaria in the work of Marshall and Bles; that is to say, cells which represent the intermediate mesoderm form an agglomeration which acquires tubular form, becomes connected to the archinephric duct, coils in a regular manner, and acquires a malpighian glomerulus.

With relation to the coiling, it is of great interest to compare some of my reconstructions with those published by various authors for various forms of fish. It will be found that the stage figured at L 6,Text-fig. 1, is common to both fish and Rana; whether any significance can be attached to this point will be discussed later. Our account, therefore, of the early units contains only one entirely new point—the method of formation of the peritoneal funnels. The remainder of the description only serves to give in detail what has already been roughly sketched by Marshall and Bles.

The development of the later malpighian units, however, is not clearly analogous to the same process in any other vertebrate; hut it is, none the less, quite easy to see how this development has been evolved. The entire crux of the matter, in my opinion, lies in the early separation of the blastema from the archinephric duct and the need for great speed in production due to the very short larval life.

We may, I think, take it as axiomatic that a vertebrate malpighian unit arises, in some way or another, from the intermediate mesoderm; when this mesoderm lies closely applied to the archinephric duct, the most economical—but at the same time the slowest—method of production is that shown by the early units. When the intermediate mesoderm, however, has become separated from the duct it would entail a great waste of time were each unit to send out a connexion to the archinephric duct across the whole breadth of the kidney; hence the early specialization of a unit to the function of collecting trunk.

It is stated (and with a great deal of confirmatory evidence) that in Urodeles this trunk is in itself a functional unit from which the later units are derived as buds. If this is so, then our description merely bears out what has already been frequently remarked—the great shortening of developmental processes shown by R a n a; for if our description has been followed it will be quite obvious that the straight tubule is, as shown by the abortive malpighian capsule at its tip, merely a unit which becomes modified at an early, instead of at a late, stage of its existence. That the malpighian capsules and glomeruli should be formed from a separate piece of blastema and not in any way from an existing tubule is a modified condition which finds no exact parallel in our present accounts of other Amphibia.

I consider, therefore, that the later units of Rana are directly comparable to what has been described as the whole history of the mesonephros in Urodeles. How, then, can we account for the early units of Rana? It seems a perfectly fair assumption that they represent an ancestral condition—a condition where a longer larval life had not rendered speed an essential factor in production. This theory is rather borne out by the great similarity shown in the method of coiling of these tubules and of fish kidney tubules. It is a well-known fact that all vertebrate kidney tubules show the familiar ‘S’ bend (Henle’s loop); it is difficult to see why this form should have been adopted by so many animals unless we assume it to have been the form adopted by the very earliest ancestors of the group. Now if we accept the view that one type of bend has been retained from what may well be termed a primordial ancestor, we should have very little trouble in believing that other regular bends, common to the fish and to an early amphibian kidney, form evidence of a relationship between the two.

There is one point arising out of this argument which is a little difficult to understand—the apparent absence of this early set of units from the Urodelan kidney. The word apparent is employed since it seems very probable that a thorough reinvestigation of this type may show the early units to be present. If this proves to be the case, then the whole evolution of the kidney of R. temporaria apparently took place along the following lines:

An ancestral form, with a long larval life and consequently no need to produce an adult kidney in the quickest possible manner, has given rise firstly to the Urodelan kidney with its fairly rapid method of production (the budding off of later units from a functional early unit); and secondly, the kidney just described, in which the period occupied by Urodeles in the formation of a nephrically functional collecting trunk has been suppressed, and further, in which a method of producing a number of malpighian units in quick succession has been evolved.

So much for the malpighian units; now let us examine the question of the peritoneal funnels. In Urodeles the funnels lead from the coelom into a short tubule which opens directly into the cavity of a malpighian capsule—that is to say, the coelomic fluid is in osmotic connexion with the blood-supply. For some reason—the exact reason I leave to the physiologist—R ana has found it necessary to have a direct connexion between the coelomic fluid and the blood-supply at the earliest possible moment. This need has been met by the early units, in which those cells which would have formed the tube leading from the funnel to the capsule remain quiescent and the funnel itself is therefore forced to open into the blood-supply. Now, not only does it appear to he necessary that the funnels should be formed as early as possible, but also that they should continue to be formed as rapidly as possible; but the method adopted for the early production of malpighian capsules definitely cut down the rate at which funnels could he formed, for each malpighian capsule takes some time to reach the periphery of the kidney. It therefore became impossible for the production of the peritoneal funnels and malpighian capsules to remain in any way correlated with each other—hence the evolution of the funnelforming tubule which carried the developing funnel directly to the point at which it was most required.

It would be as well to seize this moment to digress slightly and to demonstrate that no research is ever altogether new. Surprising though it may sound, this funnel-forming tubule was observed by Spengel, the earliest of all workers upon the Anuran kidney. He started with the assumption that the funnel must open into a capsule, but failed to find any proof of this; he did, however, find a single long tubule, internally ciliated at its tip, which he thought that he had succeeded in tracing into the collecting trunk. This error on his part is the more easy to understand when we remember that both the collecting trunk and the funnel-forming tubule end blindly in the upper blastema mass; it obviously requires only a very slight mistake to connect two tubules whose blind ends are separated at the most by a millimetre of irregular tissue.

So far we have offered no explanation as to the method by which these funnel-forming tubules might have been evolved; for our suggestions let us turn once more to the Urodeles. Ftirbringer, Hoffmann, and Farrington have all recorded that both the funnels and the connexion of the funnels to the capsules split in Urodeles so that two funnels may lead into a single capsule. If we postulate that this splitting process slowly receded farther and farther along the length of the unit, we obtain a condition in which the funnel and malpighian capsule form the two short arms of a ‘Y’; this condition would be the less easy to visualize were it not described by Hall as a stage in the development of a unit of Rana sylvatica. From this stage to a ‘V’—one of whose arms bears a developing malpighian capsule and the other a developing funnel—is but a short step, and from this to the condition in R. temporaria in which the two arms of the ‘V’ develop entirely independently of each other, a still shorter step. I do not, of course, suggest that the funnel ever opened into the unit behind the malpighian capsule—but I do suggest that the funnel-forming tubule may have done so.

There is one further point in the earlier literature which bears out both the existence and the function of these funnel-forming tubules; this point is that F ürbringer, Spengel, and Farrington have all commented upon the extraordinary manner in which the peritoneal funnels of the adult are grouped along the course of the blood-vessels. This ceases to be extraordinary when we remember that this is the very course followed by the funnelforming tubule.

I trust, therefore, that I have succeeded in showing what I stated at the beginning of this discussion would be shown—that however extraordinary the course of the development of the mesonephros of R. temporaria may appear, it is possible to show both how such a course of development may have been derived from an earlier form and to find corroborative evidence for it in the accounts of other workers.

In conclusion, I would like to record my great gratitude to Professor E. W. MacBride, F.R.S., who not only suggested that I should work upon this subject, but has also given me the greatest help and encouragement throughout the whole course of the research. Mr. H. R. Hewer, M.Sc., has also provided me with many helpful suggestions and has freely given me his assistance with the technical problems encountered.

I would also like to express my appreciation of the many constructive criticisms which have been offered by my friends and colleagues in the Zoological Research Laboratory of the Imperial College of Science, amongst whom Mr. H. K. Mookerjee and Miss D. E. Sladden have been especially helpful; indeed, it is to her skill in the rearing of animals under artificial conditions that I am indebted for the greater part of my material.

My friend Miss Kitty Edridge also afforded me considerable assistance in the somewhat tedious work of transferring the reconstruction in Text-fig. 1 from squared paper to its present form.

1. The entire mesonephros of R. temporaria is derived from the mesonephric blastema, a mass of cells originally occupying a position along the dorso-medial wall of the archinephric duct and later along the dorso-medial angle of the kidney.

2. The mesonephros arises as two perfectly distinct sets of units which are termed the ‘early units’ and the ‘later units

3. The early malpighian units are derived as from small spherical condensations (termed ‘nephroblast vesicles’) in the blastema.

4. Each of these nephroblast vesicles elongates into a tubule which forms a connexion with the archinephric duct at one end and develops a malpighian capsule at the other.

5. These early units later sever their connexion with the archinephric duct and degenerate.

6. The early peritoneal funnels are derived from masses of blastema lying between the early malpighian capsules and the periphery of the mesonephros.

7. The lumina of these funnels never form any connexion with the lumina of the malpighian capsules, but form a direct connexion between the coelom and the blood-system.

8. The growth and coiling of the early units forces the archinephric duct away from the blastema into the dorso-lateral angle of the kidney.

9. During the course of this separation a special set of condensations arise.

10. These condensations elongate into straight tubules which maintain a connexion between the archinephric duct (into which they open) and the blastema.

11. That end of the straight tubule which lies in the blastema develops an abortive malpighian capsule at its tip.

12. The later malpighian units arise as condensations (‘capsuloblast vesicles’) in the blastema when this latter has become separated from the archinephric duct.

13. The capsuloblast vesicles are not formed singly but in vertically hanging strings; this is due to the fact that as each vesicle develops it is pushed downwards by the growth of the blastema above it, while a further vesicle condenses in the clear patch of blastema so left.

14. The capsuloblast vesicle differentiates into capsule and glomerulus, from the former of which a tubule grows out.

15. This tubule forms a connexion with a bud which has grown out from one of the straight tubules.

16. When the most ventral capsule of each string approaches the peritoneal wall of the kidney, it is seen to be separated from the latter by a small mass of blastema.

17. This mass of blastema develops into a peritoneal funnel exactly as do the blastema masses lying between the early malpighian capsules and the peritoneum.

18. A further set of condensations (the vesicles of the ‘funnelforming tubule’) now arise in the blastema.

19. Each of these vesicles elongates into a tubule which runs down towards the peritoneal wall of the kidney.

20. The tubule so formed becomes internally ciliated along the lower third of its length.

21. The tip of one of these tubules enters one of the peripheral blood-vessels and places itself parallel to the outer wall of the mesonephros.

22. The whole tip of the tubule now breaks off and, by acquiring a connexion with the coelom, becomes a perfect peritoneal funnel.

23. It is suggested that the ‘early units’ represent an ancestral mesonephros and that the later units are homologous with the whole of the Urodelan mesonephros as it is at present known.

In view of the fact that several new terms have been coined to describe structures whose existence or function has hitherto been unknown, it has been thought as well to include a glossary of these new terms.

Capsuloblast vesicle.—

A condensation in the blastema which gives rise to a later malpighian capsule and glomerulus.

Nephroblast vesicle.—

A condensation in the blastema which gives rise to an entire early malpighian unit.

Funnel-forming tubule.—

A tubule whose sole function is the production of later peritoneal funnels.

Straight tubule.—

Employed as synonymous with, but more descriptive than, the earlier term ‘collecting trunk’.

a.d., archinephric duct; a.m.c., abortive malpighian capsule; blastema; b.m., point of separation of funnel from funnel-forming tubule; b.v., blood-vessel; cp.v., capsuloblast vesicle; d.m.c., degenerating malpighian capsule; f.k., rudiment of malpighian tubule; f.n., rudiment of peritoneal funnel; f.st., rudiment of straight tubule; i.n., opening of peritoneal funnel to blood-vessel; k., malpighian tubule; m., myotome; m.c., malpighian capsule; n., peritoneal funnel; nst.t., funnel-forming tubule; nst.v., vesicular rudiment of funnel-forming tubule; o.n., opening of peritoneal funnel from coelom; o.st., outgrowth from straight tubule; r.p.t., retro-peritoneal connective tissue; st., straight tubule.

PLATE 27.

Figs. 1 to 6.—Transverse sections across the nephric region of a 17 mm. tadpole, in the regions indicated in Text-fig. 1.

PLATE 28.

Figs. 7 and 8.—Microphotographs of sections of the mesonephros of a 20 mm. tadpole.

Fig. 7. Transverse section.

Fig. 8. Obliquely parasagittal section.