The following paper is divided into two parts, The first part contains an account of observations on the development of the Wolffian duct and anterior Wolffian tubules in the chick, being supplementary to my paper on the “Kidney of the Chick”* 1 The second part is devoted to a discussion of the vertebrate excretory system in general.

The first trace of the Wolffian duct is visible in an embryo with eight protovertebræ as a slight projection from the intermediate cell mass towards the-epiblast in the region of the 7 th and 8th protovertebrae. The projection also extends back behind the region of the protovertebrae for a short distance. In a chick with nine or ten protovertebræ a similar condition is found, i.e. a projection from the intermediate cell mass towards the epiblast in the region of the 7th, 8th, 9th, and 10th protovertebrae, and for a short distance behind the region of the protovertebræ.

In a chick with ten protovertebræ the projection is beginning to show signs of separation from the intermediate cell mass at certain points. The appearance presented by the rudiment of the Wolffian duct in the 10th segment of a chick with ten segments is shown in fig. 1.

In a chick with eleven protovertebræ the rudiment of the Wolffian duct is still present as a projection from the intermediate cell mass in the region of the 7th, 8th, 9th, 10th, and 11th protovertebrae; but behind the region of the protovertebræ it has grown back for a short distance between the epiblast and mesoblast as an irregular cord of cells not connected to the peritoneal epithelium. A partial separation of the Wolffian duct from the intermediate cell mass is now effected in the region of the 7th to the 10th protovertebrae. This separation is not, however, complete; but the Wolffian duct remains connected io the peritoneal epithelium at certain intervals by short cords of cells.

In a chick with twelve protovertebræ the separation of the Wolffian duct from the intermediate cell mass in the region of the 7th to the 11th protovertebræ inclusive is as complete as it ever will be, i.e. it has separated for the greater part of its length, but remains attached to the peritoneal epithelium at certain points, by cords of cells (fig. 2) derived from the cells of the intermediate cell mass connecting the rudiment of the Wolffian duct with the peritoneal epithelium. These cords of cells are the commencing Wolffian tubules of the anterior part of the Wolffian body, and are more numerous than the segments in which they are placed. Behind the region of the protovertebræ in a chick of this age (twelve protovertebrse), the Wolffian duct has grown back as an irregular cord of cells (fig. 6), independent of the intermediate cell mass, for a short distance, thus repeating the feature of the last and succeeding stages in this particular. In the region of the last (12th) protovertebra, however, the cord of cells constituting the Wolffian duct at this stage is now continuous with the intermediate cell mass at certain intervals. Comparing the sections through the 12th segment of this stage with those just behind the 11th protovertebra of the previous stage, it is seen that the Wolffian duct has enlarged, and by a downgrowth of cells from it, with which probably is connected an upgrowth from the intermediate cell mass, has become in certain places connected with the intermediate cell mass. These secondary connections constitute the commencing tubules of this part of the Wolffian body.

In a chick with thirteen protovertebrse an advance precisely similar to that characterising the previous stage has taken place, i.e. the Wolffian duct has become connected with the intermediate cell mass in the 13th segment (fig. 7), and behind this point is free from adjacent structures.

In a chick with fourteen or fifteen protovertebræ the process of development remains the same. So that in a chick with fifteen segments the following is the condition of the Wolffian duct:—It extends from the 7th to the 15 th segment as a solid cord of cells, connected at intervals with the peritoneal epithelium by the commencing Wolffian tubules; behind the 15th segment it extends for a short distance as a free cord. The further development differs from that just recorded in this important particular; the duct does not become connected with the intermediate cell mass of the newly-formed last segment, but remains separate for a considerable interval of time (till towards the end of the third day) from it. In other words, the formation of the Wolffian tubules and their connection with the Wolffian duct is deferred behind the 15th segment.

To sum up the developmental changes above recorded, the Wolffian duct arises as a continuous ridge of cells projecting from the intermediate cell mass towards the epiblast in the region of the 7th to 11th protovertebræ inclusive. This ridge’ separates from the intermediate cell mass from before backwards, remaining, however, connected with it at intervals by the rudimentary Wolffian tubules. Meanwhile, from the hind end of it there grows back a cord of cells independent at first of the adjacent structures, but immediately on the formation of the hinder segments becoming connected with the intermediate cell mass of each segment in turn. This happens as far back as the 15th segment; behind this point it grows back as a solid cord, which does not become connected with the intermediate cell mass until the tubules of the Wolffian body have made considerable advance in tbeir development.

Figs. 1—7 are meant to illustrate the above method of development. Figs. 1—5 are from the 10th segment of chicks, with ten, twelve, thirteen, and fourteen protovertebrse respectively. They are all taken through points where the Wolffian duct remains attached to the peritoneal epithelium, i.e. through a rudimentary tubule, excepting fig. 4, which is from a section close to fig. 3, and shows the condition of things in one of the intervals between the points of continuity.

Fig. 6 is taken from a section just behind the last segment of a chick with twelve segments, and shows the complete independence of the Wolffian duct.

Fig. 7 is from the 13th segment of a chick with thirteen segments, i.e. from the same region as fig. 6, and it shows the connection which has become established between the Wolffian duct and the intermediate cell mass by a mutual growth of these structures.

Fig. 8 is from the 16th segment of a chick with twenty-two protovertebrse, and is illustrative of the fact derived from an inspection of all the sections of the segment, that the Wolffian duct is independent of the peritoneal epithelium. From the 15th segment the Wolffian duct grows back independently to the cloaca, into which it eventually opens, and a lumen appears in it from before backwards.

In fig. 11,’taken from a chick at the end of the third day, it is still distinct from the now considerably developed Wolffian tubule (w.t.).

For purposes of description I shall divide the Wolffian body into three regions—(1) The part found within the limits of the 7th—11th segments inclusive; (2) the part found within the 12th—15th segments inclusive; (3) that found behind the 15th segment.

In a previous paper I have described at some length the early development of the Wolffian body behind the 16th segment, and I have there shown that that part may be divided into two parts, each characterised by a peculiarity in the early development. In this paper I shall make but little reference to the development of the Wolffian body in this region, confining myself almost entirely to that part lying within the area of the 7th to the 15th segments inclusive.

The Wolffian tubules and Wolffian duct in this region attain but a slight development. They may almost be said to have reached their highest point at the stage with fourteen provertebrse, the only difference in later stages being the development of a lumen in them. The lumen in the tubule may acquire an opening into the Wolffian duct in some cases. In this case the string of cells seen in fig. 5 becomes very short, and the Wolffian duct appears as a narrow groove in the peritoneal epithelium. This state of things is usually found in chicks with from nineteen to thirty-two protovertebrse.

The Wolffian duct in this region exhibits great variations in calibre, and in later stages parts of it appear to atrophy, and isolated portions are found connected with rudimentary tubules. An enlarged section of the Wolffian duct in front is nearly always found as Gasser 2 has described. The duct and tubules in this region appear entirely to atrophy in chicks with more than thirty-five protovertebræ.

I have not thought it worth while to preserve figures of the duct and tubules in this region of the Wolffian body after their first appearance, as the arrangement just described may be easily observed in sections of an embryo chick of the third day.

The interest in the development of this region lies in the fact of the continuity of development of the Wolffian tubules and Wolffian duct. It has always appeared to me astonishing that the Wolffian duct developed as a continuous ridge from the intermediate cell mass, which, from our knowledge of Elasmobranch development, may be called the peritoneal epithelium, should entirely separate from it and then secondarily become connected with it by the tubules of the Wolffian body. My investigations, which have been made with some care on a large number of chicks of all ages from nine to thirty proto vertebrae, have entirely convinced me that the usual statements on this point are not true, and show to my mind most conclusively that the duct and tubules of the Wolffian body in the region in question do develop in continuity, precisely as do the duct and peritoneal openings of the head-kidney in most Ichthyopsidan types.

The number of rudimentary tubules in each segment of this region I have not determined precisely. They occur as often as not between the segments, and there seems to be about two for each segment. In the seventh segment I have never seen more than one.

Before proceeding to give an account of the further development in the next region, I will briefly refer to the points in which my observations differ from those of previous observers on the development of the Wolffian duct.

Gasser’s account of the development of the Wolffian duct is the most recent and exact. In his valuable paper will be found a complete account of the literature of the subject, to which I need not further refer.

“The first trace of it which he finds is visible in an embryo with eight protovertebræ as a slight projection from the intermediate cell mass towards the epiblast in the region of the three hindermost protovertebræ. In the next stage with eleven protovertebræ, the solid rudiment of the duct extends from the 5th to the 11th protovertebræ; from the 8th to the 11th protovertebræ it lies between the mesoblast and epiblast, and is quite distinct from both, and Dr. Gasser distinctly states that in its growth backwards from the 8th protovertebræ the Wolffian duct never comes into continuity with the adjacent layers. In the region of the 5th protovertebræ, where the duct, &c., was originally continuous with the mesoblast, it has now become free, but is still attached in the region of the 6th to the 8th. In an embryo with fourteen protovertebræ the duct extends from the 4th to the 14th, and is now free between epiblast and mesoblast for its whole extent.”

The points in which the preceding account differs from that of Dr. Gasser’s briefly are :

1. The position of the continuous ridge of the Wolffian duct.

2. The subsequent complete isolation of the duct in the region of the ridge.

3. The independence of the backward growth of the duct in the 12th to the 15th segment.

I have never seen any trace of the Wolffian duct in front of the 7th segment, and in all the chicks I have examined I find that the continuous ridge extends from the 7 th to the 11th segments.

With regard to Gasser’s statement of the complete isolation of the duct in the anterior region from the intermediate cell mass, I can only say that my observations point to an entirely different conclusion.

Thirdly, I differ with him in his statement that the duct in the growth back from the attached extremity does not come into relation with adjacent structures.

As stated above, it seems to me that for the space of four segments the small cord of cells which grows back from the hind end of the ridge, does almost immediately become connected with-the intermediate cell mass.

I now pass to the most interesting point which has turned up in my investigations on the excretory system of the chick.

In a paper by Mr. Balfour and myself in the ‘Quart. Journ. of Mier. Science,’ vol. xix, describing the development of what we believed to be a rudimentary head-kidney in the chick, we drew attention to a structure which so closely resembled the glomerulus1 of the head-kidney of the Ichthyopsida that we identified it as an homologous structure.

Gasser2 has also independently discovered and similarly identified this structure.

In the paper just referred to no attempt was made to trace the development of this glomerulus, but it was merely described as it appeared at the time of its greatest development.

The following description is taken from that paper :

“In the chick the glomerulus is paired, and consists of a vascular outgrowth or ridge projecting into the body cavity on each side at the root of the mesentery. It extends from the anterior end of the Wolffian body to the point where the foremost opening of the head-kidney commences. We have found it at a period slightly earlier than that of the first development of the head-kidney….In the interior of this body is seen a stroma with numerous vascular channels and blood-corpuscles, and a vascular connection is apparently becoming established, if it is not so already, between the glomerulus and the aorta. The stalk connecting the glomerulus with the attachment of the mesentery varies in thickness in different sections, but we believe that the glomerulus is continued unbroken throughout the very considerable region through which it extends. This point is, however, difficult to make sure of, owing to the facility with which the glomerulus breaks away. At the stage we are describing no true Malpighian bodies are present in the part of the Wolffian body on the same level with the anterior end of the glomerulus, but the Wolffian body merely consists of theWolffian duct. At the level of the posterior part of the glomerulus this is no longer the case, but here a regular series of primary Malpighian bodies is present, and the glomerulus of the head-kidney may frequently be seen in the same section as a Malpighian body. In most sections the two bodies appear quite disconnected, but in those sections in which the glomerulus of the Malpighian body comes into view it is seen to be derived from the same formation as the glomerulus of the head-kidney.’”

The point which is left in doubt in the above description, viz. as to whether the glomerulus constitutes a continuous structure, is at once decided by a study of its development.

I may here state that it is not a continuous structure, but consists of a series of external glomeruli, each of which corresponds and is continuous with the glomeruli of the Malpighian bodies found in this part of the trunk.

The first development of the Wolffian tubules in the region under consideration has already been described. They appear as outgrowths from the Wolffian duct meeting outgrowths from the intermediate cell mass immediately on the formation of the segment in which they are placed ; so that in a chick with fifteen protovertebrse the Wolffian duct is connected with the intermediate cell mass by a certain number of cell cords in the 12th, 13th, 14th, and 15th segments.

The duct and cords, which have at first rather an irregular outline, soon become well-defined compact structures.

Fig. 12, taken from the 12th segment of an embryo with twenty-two segments, represents the condition of things at this age-….

The Wolffian tubules in this region are derived from two distinct structures—(1) the outgrowth from the Wolffian duct; (2) part of the intermediate cell mass.

The intermediate cell mass is at first continuous with the peritoneal epithelium in every section; but, as described in a previous paper, this connection soon becomes lost at certain points (fig. 9), and maintained at others (fig. 10). Figs. 9 and 10 are contiguous sections through the 15th segment of a chick with twenty-two segments, showing this point. At these points, where the continuity is retained, a peritoneal funnel is subsequently formed by the development of a lumen extending from the body cavity into the intermediate cell mass.

The features of the stage of development now reached are well known; it is that of the S-shaped cords of cells which have been so often described. In the adjoining woodcut is represented part of one of these S-shaped strings, showing clearly the above features of a tubule, &c., viz.—(1) the Wolffian duct, in which a lumen has appeared ; (2) the outgrowth from it to the intermediate cell mass forming the upper limb of the S ; (3) the intermediate cell mass with the commencing lumen from the body cavity.

In the next section the intermediate cell mass is not connected to the peritoneal epithelium.

In chicks of gradually increasing number of protovertebrse this cavity in the intermediate cell mass gradually becomes more marked (figs. 13, 1.4), and extends into that part of it immediately behind the peritoneal connection (fig. 15).

Figs. 13, 14, and 15 are three successive sections through the 13th segment of a chick with about thirty segments, showing the features of a tubule at this stage.

The Wolffian duct is connected with the lower end of the intermediate cell mass in all the three sections. A distinct lumen has appeared in the intermediate cell mass which opens into the body cavity in front (figs. 13 and 14), but is separate from the body cavity in the hindermost section (fig. 15).

Comparing these figures with figs. 9 and 10 it is seen that fig. 13 or 14 corresponds to fig. 9 in the fact of the continuity between the intermediate cell mass and peritoneal epithelium; while fig. 15 corresponds to fig. 10, in both the continuity having been lost. The difference between them consists in the presence of a distinct lumen in the older series, opening into the body cavity, and continued behind into the part of the intermediate cell mass which has separated from the peritoneal epithelium. This part, marked i. c. m. in fig. 15, will in the next stage become converted into; that part of the tubule in which a Malpighian body is developed, while the anterior part, which is open to the body cavity, will widen out considerably, and give rise to a wide peritoneal funnel.

In fig. 11 is represented a section through a developing Wolffian tubule in the hinder part of the Wolffian body. The tubule (w. t1.) in this section precisely resembles the part of the tubule (i. c. m.) represented in fig. 15. Supposing the anterior part of w. t1. were open to the body cavity it would almost be a repetition of the anterior tubule, save in the fact that it is not yet united to the Wolffian duct. But the hinder tubule (fig. 11) does not develop until after the intermediate cell mass has separated from the peritoneal epithelium, i. e. subsequent to the obliteration of the rudiment of the peritoneal funnel.

Not only do the Wolffian tubules in the region of the 12th to 15th segments develop a lumen while still continuous with the peritoneal epithelium, but further, a glomerulus appears in them while still open to the body cavity; and this glomerulus not only appears in the hinder part of the tubule (fig. 15) which has separated from the peritoneal epithelium, but also in the anterior part (figs. 13 and 14) where it is open to the body cavity. This is at once clear on inspection of figs. 16, 17, 18. These figures are taken from the 13th segment of a chick with thirty-four protovertebra;. There was a section not figured between fig. 17 and 18, otherwise the sections are successive, fig. 16 being the anterior.

In fig. 16 is seen the commencement of the peritoneal funnel as a bay lying between the Wolffian duct and mesentery.

In fig. 17, a glomerulus (gl) has appeared projecting into this bay. In the next section, not figured, the bay was almost closed up by an approximation of its edges, while in fig. 18 the bay is completely shut off from the body cavity, and we have a section of a true Malpighian body with its contained glomerulus.

Fig. 18 clearly corresponds to fig. 15 of the previous stage, while fig. 17 corresponds to fig. 14, the difference being that a distinct cellular projection (gl.) has appeared at the point where the projection of cells from the Wolffian duct joins the intermediate cell mass.

I have given a diagram (fig. 22) representing an ideal longitudinal dorso-ventral section through, two of these Wolffian tubules at this stage. This diagram has been made from a study of many embryos showing the development of the external glomerulus. ;

The open peritoneal funnel is represented, at p,f., the arrow pointing into it. Through it is projecting the anterior part of the glomerulus (gl.), that part which I shall call the external glomerulus. A transverse section through this part would give the appearance represented in fig. 17,..

Into the closed hinder part of the tubule (mb) is projecting the hinder part of the glomerulus (i. gl..), which I shall call the internal glomerulus. It was not possible to represent satisfactorily in this diagram the Wolffian duct, which, obviously from its position in transverse section, would not be seen in a longitudinal section passing through the attachment of the glomerulus.

In fig. 23 is represented somewhat diagrammatically a transverse section through a chick with thirty-three protovertebrae, i. e. from a slightly younger embyro than that from which figs. 16—18 were taken, in which the cord of cells connecting the Wolffian duct with the cavity of the glomerulus had acquired a distinct lumen, the cavity of the Wolffian duct being here distinctly continuous with that of the bay in which is placed the rudimentary external glomerulus, and so with the body cavity. At subsequent stages this part of the tubule appears to persist, but only in a rudimentary fashion.

The next stage which I propose to describe was found in a chick in which thirty-six protovertebrse could be counted, but possibly there were more.

The glomerulus has grown immensely (figs. 19, 20, 21), and has now acquired the peculiar histological features which characterise it at the time of its greatest development, and which have already been described in a former paper.

Anteriorly the bay has widened out considerably (fig, 19), and the glomerulus (e. gl.) projects directly into the body cavity. Posteriorly the bay remains deep (figs. 20, 21), and the glomerulus almost completely fills it and projects beyond it into the body cavity. In sections behind fig. 21 there was seen a fairly well-developed internal glomerulus.

The edges of the bay are gathering round the glomerulus preparatory to fusing with it, and so closing up the peritoneal funnel and dividing the glomerulus completely into two parts, the internal vascular tissues of which, however, are continuous.

In this stage the epithelial covering of the external glomerulus (e. gl.) was distinctly, as in the previous stage, continued behind directly into that covering the posterior internal glomerulus.

When, however, the peritoneal funnel closes by the completion of the process commencing in figs. 20 and 21, this epithelial continuity is lost, and we have the final stage of the glomerulus, the last which I have observed, in which the separation above described is complete, so that in this stage, which is that of the greatest development of the external glomerulus, and corresponds with the commencing formation of the head-kidney, the glomerulus belonging to one tubule is divided into three parts.

(1) An anterior1 part projecting into the body cavity. This corresponds to a further development of fig. 19.

(2) A middle part, continuous with (1), also projecting freely into the body cavity, but also connected by vascular structures with an internal glomerulus. This part is figured in fig. 26, and corresponds to a further development of the part from which fig. 20 and 21 were taken.

(3) A posterior part, in which there is no external glomerulus, but merely an internal one belonging to a true Malpighian body of the mesonephros, which I have not thought it necessary to figure in this or the previous stage. It is a further development of fig. 18. This stage, which may be observed about the middle of the fourth day of incubation, brings to a close my observations on this extraordinary structure. It appears that in the chick the stage just described is that of the greatest development of the external glomerulus. In the duck, however, I have often met with it even larger and more developed, and it appears to me after its separation from the internal glomerulus to get an independent growth, and while the latter is undergoing atrophy to become larger and extend itself posteriorly, so as almost to overlap the external glomerulus of the next tubule.

With regard to the number of the external glomeruli in the chick and the exact limits of their occurrence, the following is briefly what I have been able to make out in a chick with thirty protovertebrae:

In the 11th segment there are two rudimentary tubules running from the Wolffian duct to the peritoneal epithelium. At the point of attachment of these there is a small rudiment of the external glomerulus, visible for only one section in each case.

In the 12th segment there is at the beginning a Wolffian tubule and a well-marked external glomerulus extending through three sections. At the hind end of the 12th segment and beginning of the 13th there is an external glomerulus for three sections continued into part of the segmental tube behind, in which an internal glomerulus will subsequently be developed.

In the 13th segment there is an external glomerulus for three sections.

In the 14th segment there are two segmental tubes with developing external glomeruli.

In the 15th segment no external glomeruli appear to be developed, the segmental tubes being already separated from the peritoneal epithelium.

In later stages only the three or four hindermost of the above external glomeruli appear to develop further. The anterior glomeruli soon atrophy with the adjoining tubules and duct.

In the duck a much greater number become developed, and they may be seen in the anterior segments after their respective tubules have entirely atrophied.

The bearing of the developmental processes above recorded on any hypothesis as to the phylogenetic history of the vertebrate excretory system I propose to examine in the second part of this paper (pp. 460—462; 464).

The most peculiar feature of the excretory system of the vertebrata is the presence of three more or less distinct parts, the pronephros, the mesonephros, and the metanephros or kidney proper. In the following pages my object will be to explain the relation of these parts, more especially those of the pronephros and mesonephros, and to show that they have arisen as differentiations of a primitively uniform structure.

For this purpose it is necessary briefly to recapitulate the more important features in the development which have a bearing on my argument.

The first part of the excretory system to make its appearance is always a duct. This duct has received various names, but its homology in different forms is undisputed. I shall call it the segmental duct.

In the chick the segmental duct is commonly known as the Wolffian duct.

All the Ichthyopsida whose development is known, with the exception of Elasmobranchs, possess a structure called the headkidney or pronephros. The pronephros when present always develops in continuity with the anterior end of the segmental duct.

In the Amphibian the segmental duct arises as a groove of the parietal peritoneum, just ventral to the place where the body cavity is connected with the cavities of the muscle plates. This groove, which arises first of all anteriorly just behind the branchial region, is continued for a certain distance backward. It soon, however, becomes constricted into a canal lying between the ectoderm and parietal peritoneum. This constriction has been described as taking place in the following manner:—It first appears in the middle region of the groove, giving rise to a canal opening into the body cavity in front and behind. It then is continued backwards until the groove is completely converted into a canal behind, which soon acquires an opening into the cloaca. Anteriorly the wide opening meanwhile is divided up into two,1 three,2 or four3 openings, according to the species.

The canal immediately behind the last of these openings becomes coiled and placed on the same level but ventral to the openings. The part of the body cavity into which the openings of the segmental duct pass widens out, a vascular projection— the glomerulus—from the dorsal inner wall is formed, extending uninterruptedly from opposite the anterior opening of the segmental duct to as far back as the posterior. The dilated section of the body cavity in which the glomerulus lies, and into which the segmental duct opens, is partially separated from the rest of the body cavity. The whole structure, including openings of duct, ventral coiled part of duct, glomerulus, and dilated part of body cavity, is known as the pronephros. The number of openings from the segmental duct into the body cavity corresponds with the number of segments through which the pronephros extends.1

With its excretory system in this condition the young Amphibian is hatched. Fundamentally the head-kidney retains the above structure, increasing only in size until it begins to atrophy, an occurrence which takes place on the development of the mesonephros.

This method of development of the segmental duct and pronephros is fundamentally repeated in other animals which possess a pronephros.

About the marsipobranch development very little is known. Fürbringer (loc. cit.), quoting W. Müller and his own observations, makes the following statements for Petromyzon :—In the earliest stage which has been observed there was present at about the level of the heart a groove in the parietal peritoneum, which leads behind into a duct, which eventually, by a backward growth, reaches the cloaca and opens into it. The anterior groove or opening of the duct soon becomes divided up into four openings.

In the young Ammocætes there is present a pronephros made up of a complicated coiled duct and four or five openings into the body cavity, opposite which is placed a vascular glomerulus ; the whole structure extends over four or five segments.2The pronephros atrophies in the adult.

In Myxine nothing is known of the development, but in the adult a pronephros has been described, which, however, is not functional in old individuals (adult ?), as in them it has lost its connection with the backward continuation of the segmental duct.

It3 consists of the segmental duct, which gives off dorsally a number of diverticula, in which are found glomeruli, and ventrally a number of coiled canals, which open apparently into the pericardial cavity.

The fully-formed pronephros of Petromyzon then resembles in structure very closely that of Amphibia, while the pronephros of Myxine differs in certain important points.

The Teleostei possess a pronephros, which persists as a large organ in the adult. It develops in connection with the segmental duct precisely as does the pronephros in Amphibia. The only difference between the two is that in Teleostei the segmental duct has never more than one anterior opening, and the part of the body cavity into which it opens, and in which the glomerulus lies, is completely constricted off from the rest of the body cavity, and comes to resemble exactly an enormous Malpighian body.1

I may here sum up the common features characterising the ontogeny of the pronephros and its duct (segmental duct) in all the forms of the Ichthyopsida in which the development is at all known:

1. The segmental duct arises first as a ridge from the parietal peritoneum. This ridge usually contains a diverticulum from the body cavity, and is continuously constricted off to form a duct.2

2. Except anteriorly, where the constriction only takes place at intervals, leaving the openings of the pronephros (except in Teleostei, where there is only one opening).

3. These openings correspond in number with the segments which the pronephros occupies.

4. A vascular structure, called glomerulus, is formed, projecting on each side of the aorta into a specialised dilatation of the anterior part of the body cavity. Myxine forms a peculiar exception to this otherwise universal fact.

5. This dilated part of the body cavity may become partially or completely separated off to form a capsule, into which the glomerulus projects and the anterior end of the segmental duct opens.

6. The pronephros in all those Ichthyopsida in which it is found attains a functional development, but is usually only active during a period intervening between the hatching and the attainment of full maturity, i. e. it only functions in the larva.

In Elasmobranchs, which do not, so far as is known, possess a pronephros, the segmental duct arises as a solid ridge from the somatic layer of the intermediate cell mass in the anterior region of the trunk. From this ridge there grows back a column of cells to the cloaca. On the development of a lumen the segmental duct, with its peritoneal opening, is established. The duct develops quite independently of adjacent structure behind the poifit of its original attachment, and does not unite with the segmental tubes till considerably after its first development.

The difference in the development of the segmental duct in the forms possessing a pronephros and in Elasmobranchs is only one of degree.

In both cases it at first arises as a projection, either solid or containing a diverticulum from the body cavity, from the parietal peritoneum just ventral to the muscle plates; but in the one case this groove has a greater longitudinal extension than in the other. In all probability the hinder part of the segmental duct is in all cases formed by an independent growth from the hind end of this groove.

Amongst the Amniota the chick is the type in which the development of the segmental duct has been most carefully examined.

In the chick it arises as in Amphibia as a projection (solid in the chick) from the parietal mesoderm just ventral to the muscle plates; and the extent of the ridge is the space occupied by five segments.

This ridge is constricted off at intervals from the intermediate cell mass, but remains attached at certain points. The hind end of the duct is formed by a growth back from the hind end of this ridge, which takes place independently of adjacent structures.

The question now presents itself: are these structures at the anterior end of the segmental duct in the chick, which so closely resemble in development the openings of the Ichthyopsidan head-kidney, homologous with that head-kidney ?

To a consideration of this question I shall return.

The mesonephros obtains a large development in all the groups of the Vertebrata; but it does not persist as an excretory organ in the adult of the Amniota.

It develops in three very markedly distinct ways.

The first of these characterises the Elasmobranchii.

The second the Amphibia, Teleostei, Ganoidei, Marsipo-branchii.

The third the Amniota.

The segmental tubes of Elasmobranchii were originally described by Balfour as arising as solid diverticula of the peritoneal epithelium. An examination of Balfour’s specimens led me, however, to conclude that they originated as specialised parts of the body cavity, viz. from the canals in the intermediate cell mass which connect the muscle plate cavities with the general body cavity; and Balfour has now given his adherence to this view (‘Comp. Embryology/vol. 2, p. 570).

These canals having lost their connection with the body cavity of the muscle plates acquire an opening into the segmental duct, and differentiate1 into the typical Wolffian tubules. The connection with the general body cavity may or may not be retained in the adult. The secondary tubules develop as outgrowths from that part of the primary tubules, which will give rise to a Malpighian capsule. These outgrowths grow forward and eventually acquire an opening into the terminal portion of the tubule of the segment in front. Later they loose their connection with the Malpighian capsules, though a rudiment of this is sometimes retained as a solid cord of cells.

The method of development of the secondary, tertiary, &c., tubules has not been followed.

The primary tubules open into the segmental duct very shortly after the latter has acquired an opening into the cloaca.

The formation of the Malpighian bodies and the outgrowths from them to form secondary tubules occur later.

For a full account of the development of the mesonephros in Elasmobranchs I must refer to the works of Balfour and Semper, to whom we owe the whole of our knowledge.

As a type of this development I will take an Amphibian, Salamandra, in which animal it has been more completely elucidated by Fürbringer than in any other.2

Fürbringer describes the formation of the mesonephros as taking place entirely during larval life; no trace of the gland being seen in the newly hatched larva. It arises as a series of ingrowths of the peritoneal epithelium, which soon become separate from the latter. The primary tubules are hollowed out in the cell masses so formed independently both of the body cavity and segmental duct (Wolffian duct), but subsequently they acquire an opening into both.

The secondary tubules arise in a blastema, the origin of which is not clear, but is apparently derived from the just mentioned serial ingrowths. They acquire an opening into the collecting part of the primary tubule and into the body cavity. The remaining dorsal tubules have an equally obscure origin.

As the mesonephros becomes more developed the pronephros retrogrades, and is eventually entirely, as far as its function is concerned, replaced by the former.

The development of the mesonephros in Teleostei, Marsi-pobranchii, Ganoidei, is similary described as taking place in the free young (larva) from strings of cells derived from the peritoneal epithelium. In Marsipobranchii as in Amphibia the young are hatched with a functional pronephros, and no trace of the mesonephros; but the former is, in the further growth of the young animal, gradually replaced functionally by the latter, and more or less retrogrades. In the Teleostei, however, and Ganoidei, it persists with the mesonephros as an important functional organ in the adult. In some Teleostei the pronephros is the only functional adult kidney, the mesonephros not being developed.

I have made some observations on the development of the mesonephros in the Frog (Rana, temporaria), Salmon and Sturgeon, and my observations lead me very strongly to doubt whether Fürbringer and other observers are right in describing the origin of the cells which give rise to the mesonephros as actual ingrowths from the peritoneal epithelium.

In the case of the Frog this is certainly not the case. In fig. 25 is represented a section through a Tadpole of 11 mm., showing the first trace of the cells (K B) from which the Wolffian tubules arise. At their first appearance they are independent of the peritoneum, and only secondarily become connected with it. Fürbringer figures from the Salamander a section in support of his statement; I have also seen such appearances in the Tadpole, but in this animal these strings are only found in that part of the animal in which, I am confidently able to state, no Wolffian tubules are ever developed. I have examined and compared segment with segment of Tadpoles of various ages, and have never found these strings of cells developing into Wolffian tubules. The cell strings appear to me to arise from a blastema of cells developed in situ becoming connected with the peritoneal epithelium, and they are, no doubt, rudimentary tubules.

Fürbringer in his paper gives no evidence of the origin of these cells from the peritoneal epithelium, except a drawing of a stage in which the blastema is connected with the peritoneal epithelium. 1 I have also seen this stage, as mentioned above, in my sections of the Frog, but have completely failed to find the earlier stages of this ingrowth. One would expect to see it preceded by a thickening of the very flat cells lining the body cavity at this point; one would hardly expect the flat cells so specialised to form the lining of the body cavity of the young larva suddenly, and without showing any change to begin to grow inward. Further, if the cell cords described by Fürbringer in the Salamander are really only rudimentary structures belonging to the anterior part of the mesonephros, as is certainly the case in the Frog; and if the process which Fürbringer describes for the posterior part of the mesonephros of the Salamander takes place for all fully-developed parts of the mesonephros, as is the case in the Frog, then part of the difficulty caused by the peculiar secondary development of the peritoneal funnels disappears. In other words, I believe Fiir- bringer has made a mistake, precisely similar to that which was made about the development of the Avian Wolffian body. He has seen in the anterior part of a young larva the cell cords mentioned above; which were present at a time when there was no trace of the posterior part of the mesonephros. He has also seen in the hinder part of older larvae the blastema of cells separate from the peritoneal epithelium from which the Wolffian tubules arise. Finally, he has connected these two conditions, which are, as I believe, found in different regions of the trunk, and has concluded that the cell strings of the anterior part have separated from the peritoneal epithelium and given rise to the cell masses of the posterior part which really develop independently of the peritoneal epithelium, and eventually, give rise to the Wolffian tubules.

My observations on Teleostei lead me, for similar reasons, to assert an origin, in situ, of a continuous blastema, which later, breaking up, will give rise to the Wolffian tubules.

On the other hand, the older observers, including Vogt and Rosenberg for Teleostei, Rathke, Johan. Müller, Reichert, Vogt, for Amphibia,2 are quoted by Fürbringer as asserting an origin of the tubules as a series of excavations in a blastema of cells lying just internal to the segmental (Wolffian) duct. And it seems to me that the older observers were,3 as in their statements concerning the development of the mesonephros in the chick, not far from the truth. In the Sturgeon my observations point to a similar conclusion; in the just-hatched young a few mesoblast cells are seen lying internal to the segmental duct. These, at a later stage, are replaced by a more compact mass of cells, occupying the position of which, in a still older animal, Wolffian tubules are seen1.

The point I wish to insist upon is that sufficient proof of an actual ingrowth of cell from the peritoneal epithelium has not been given; but that it is much more probable that the kidney blastema arose in situ, in some cases perhaps in continuity with the peritoneal lining, and in other cases independently of it, but soon becoming united with it to form the nephrostomata.

The development of the mesonephros in the Amniota has been most fully elucidated in the chick. 2

In a recent paper I have described the development of the posterior Wolffian tubules from a continuous blastema of cells derived from the intermediate cell mass; and in the first part of this paper that of the anterior tubules from the cell cords left connecting the Wolffian duct and intermediate cell mass.

Further, in the chick there is a kind of intermediate method of development of the tubules of the 12th—15th segments (see above).

The question here again recurs which was asked before: Are these tubules of the anterior part of the Avian Wolffian body really tubules of the Wolffian body, or have they something to do with the head-kidney ? For a discussion of this question I must refer below to p. 460.

In a recent paper3 I have attempted to show that the metanephros, which is found only in the Amniota, is developed from a blastema of cells which arises continuously with but behind the blastema from which the Wolffian tubules develop.

Although the blastema which will give rise to the greater part of the metanephros arises at a comparatively early stage in development, still it is not till a much later stage that it shifts its position, and begins to show signs of developing into the Wolffian tubules. This late development of the kidney, which in this point to a certain extent resembles the Amphibian mesonephros, is a very remarkable fact. I shall return to it again.

I have thus run over very rapidly the most salient features in the development of the various parts of the Vertebrate excretory system, so far as it is at present known to us. I now turn to a consideration of the bearing which these facts have upon any hypothesis as to the phylogenetic connection of these various organs.

But, before so doing, it will be well to consider the nature of the problem which presents itself. It is universally admitted that the Craniata have had a common ancestor. The problem to be solved is contained in these questions: What was the structure and development of the excretory system of that ancestor? How has it been modified to produce the excretory organs which we see in Vertebrates now living?

I am but too well aware how complicated and difficult the problem is, and how insufficient are the data we at present possess to enable us to solve it. Of the two sources (geology and embryology) from which we can hope to obtain these data, palaeontology can throw no light whatever upon the primitive Vertebrate or its ancestors, for the Vertebrates have apparently an antiquity greater than that of the oldest fossil-bearing rocks; and even if there are in existence fos-siliferous rocks bearing the remains of the ancestor of Vertebrates (excluding Amphioxus), we can hardly hope, when they are found, to obtain any knowledge of the ontogenetic development or structure of soft parts, and the light which palaeontology throws upon the later history is at present difficult to use in settling questions of this kind, so that we are thrown almost entirely upon embryology for the facts; but the facts which embryology at present supplies us with are quite inadequate to enable us, even approximately, to solve the problem. But still, such as they are, it seems worth while to put them together, and to discuss the conclusions to which they seem to point.

Mr. Balfour1 has compared the embryonic record to an ancient manuscript in which many leaves are missing, many moved out of their proper order, and many spurious ones interpolated by later hands. It is the duty of an embryologist to try to reconstruct the manuscript and see exactly what it contained when it was first written. In doing this he is aided by the fact that he has access to many copies of the manuscript, which have each been used and altered by very different people. He is thus able, by comparing the different copies, and by studying the characters, &c., of the people by whom they have been possessed, to arrive at a more correct idea as to what the original was like than if he had only one copy.

In studying the various embryonic records we have we can pick out certain features common to all, and which may be assumed to have had their counterpart in the phylogenetic history. But the majority of features have been so altered that it is only possible to arrive at anything like a conclusion by taking into account the complicated conditions in which the animals have lived.

While the pronephros is characterised by a very similar structure and development in all the animals in which it occurs, the mesonephros, though possessing in all animals a fairly similar adult structure, presents most remarkable differences in development in the different groups. While the mesonephros is universally (few Teleostei excepted) present, the pronephros is only present in certain forms. Considering first the Ichthyopsida, it is at once seen that the presence or absence of a pronephros is correlated with another peculiarity. When the pronephros is present the egg contains a relatively small amount of food yolk, and the young undergo a considerable part of their development after leaving the egg; while, when the pronephros is absent, the egg contains a very bulky food yolk, and the young undergo far the greater part of their development within the egg (Elasmobranchii).

Further, again considering the Ichthyopsida, we find that one method of development of the mesonephros is found in those animals with a pronephros, while the other method is found in those animals without a pronephros. Of the two methods of development of the mesonephros, while one (that found in Elasmobranchii) may be considered as in some respects primitive, the other must be regarded as very much modified.

Whatever may have been the phylogenetic origin of the Wolffian tubules, the ontogenetic origin, as seen in Amphibia, Teleostei, Ganoids, Marsipobranchii, cannot possibly be regarded as in any way approaching the former. We cannot suppose that a definite serial organ like the mesonephros developed in phylogeny as a series of independent cavities in a mass of mesoblastic cells. At any rate, I think I am justified, in the present state of our knowledge, in making this statement. It is completely opposed to our ideas, and can only be accepted when all other hypothesis as to the origin of the mesonephros in phylogeny, based on the facts of embryology, have been shown to be untenable.

The tubules of the mesonephros in Elasmobranchii, however, in which group they arise from parts of an organ previously developed, present a method of development which is not at all at variance with our a priori views as to their phylogenetic origin. From considerations of this kind it seems to me a fair assumption that the development of the tubules in Elasmobranchs from parts of the body cavity more nearly resembles the method by which the organ arose in phylogeny than does that of the Wolffian tubules of the remaining Ichthyopsida.

In Elasmobranchs the Wolffian tubules have a segmental arrangement; one is found in each segment. In all probability this also is a primitive condition.

The arrangement of the tubules in the other vertebrata, although it does not actually afford support to this view, still it does not disprove it. It is a well-known fact that the segmental tubes have very rarely a segmental arrangement in the adult or even in the embryo. But in this connection it must be remembered that the tendency of development always seems to be to render that part of the mesonephros, which is going to function in the adult as an excretory organ, more compact, i.e. to bring its constituent parts closer together. I need only refer to the kidneys of the Urodele Amphibia. Here the posterior part of the mesonephros, which is going to function in the adult as kidney, becomes distinguished by its size and the course of its ducts from the anterior part, and in the female by its size only from the anterior part. And Fürbringer has shown, in Salamandrra maculata, that in correspondence with the increasing size of the posterior region there is found an increased number of primary tubules in a segment, as well as of dorsal secondary tubules. 1

Spengel has also shown that even in different species of one genus the number of primary tubules in a segment differs, e.g. in Spelerpes variegatus there is one primary tubule in a segment, in Spelerpes fuscus there are two.

Further, Fürbringer states that in the species investigated by him the number of primary tubules in a segment increases with the age of the animal.

“Die Anlagen sind in ihren früheren Entwickelungsstadien leicht zu scheiden; später hingegen lagern sie sich so innig an einander, dass eine Abgrenzung unmöglich wird.”1

Finally, there seems to be a distinct relation between the closeness of aggregation of the tubules with regard to the body segments and the number of segments found between the mouth and the anus.

In the Anourous Amphibia, where there are very few segments in the adult in this region, we find a very compact and complex kidney..

In the Urodeles, in which the number of segments is greater, the kidney occupies a greater number of segments, and is not nearly so compact, while in Coecilia, in which the anus is almost terminal, very few segments being placed behind (tail undifferentiated), we find that the kidney is segmental, i. e. one primary tubule is found for each segment, and it occupies in the adult as many as sixty segments.2

Turning to the Amniota, we find that in Lacertilia3 the mesonephros has at first a segmental arrangement, one primary tubule for each segment, and although it has not been shown that the fully developed mesonephros of lizards has lost this feature, still there can be little doubt, considering its resemblance to that of Aves, that it has; while in the case of the chick4 the number of primary tubules in a segment increases with the age of the embryo.

These three facts, viz.—-(1) The variability of the number of primary tubules in a segment in closely allied forms, (2) the increased5 number in a segment as development proceeds, (3) the relation between the compactness of the kidney and the number of segments over which it extends, all point in the same direction. They seem to indicate that the tubules of the Wolffian body are capable of shifting their position according to the wants of the particular species.

We know very well other organs can do this, and I need only mention the anus placed so near the head in frogs, and so far off in Coecilia, and it seems only probable that an important gland like the kidney should be capable of acquiring a position and arrangement of its constituent parts different from the position of their development, if it is advantageous for the performance of the function of the organ.

The evidence which at the first look appeared so strong against the primitiveness of the Elasmobranch arrangement of one primary tubule to each segment proves on examination to lose a great part of its force.

I now come to a difficulty which apparently at present presents an insuperable obstacle to a successful solution of the question under consideration, viz. What was the structure and development of the excretory system of the ancestral Vertebrate?

Assuming that the development of the Elasmobranch mesonephros presents primitive features in the two details already considered, its development in a third particular can by no means be assumed to be primitive. The fact that the segmental duct develops independently of the tubules cannot, in the present state of our knowledge, be regarded as primitive. Objections of precisely the same kind as those used in arguing against the development of the tubules in Amphibia, &c., being primitive present themselves here.

Any phlyogenetic hypothesis which presents difficulties from a physiological standpoint must be regarded as very provisional indeed. The physiological difficulty present in the conception that in the evolution the mesonephros has arisen by the fusion of two distinct parts, viz. the duct and tubule, is so great that until facts are brought forward to show a different origin we must consent to admit our total ignorance on this point. I think that the observations recorded in the first part of this paper on the development of the Avian Wolffian duct and anterior tubules are of great interest in this relation. Here we have the Wolffian duct and tubules developing in continuity in the anterior part of the excretory system, which has been always admitted to present the most primitive development. But this point I must again keep for later consideration.

So far, then, the following conclusions have been reached—the development of the mesonephros of Elasmobranchii is in part primitive (tubules), and in part very much modified, while the development of the mesonephros of Amphibia, Teleostei, &c., is in all respects modified.

Turning to the development of the segmental duct, we find ourselves obliged, for precisely similar reasons to those already given in the case of the mesonephros, to suppose that that ontogeny is in this respect more primitive in which the duct arises as a continuous groove constricted off from the body cavity than that in which it arises as a solid knob (modified groove) for only a very small part of its course, and undergoing the major part of its early growth quite independently of surrounding structure.

In Elasmobranchii that part which develops as a groove persists as a groove throughout life (abdominal opening of Mullerian duct).

In Amphibia, &c., that part which develops as a groove becomes constricted off first in the middle, and then backwards and forwards, but in front it is constricted in a manner, according to Fürbringer not understood, so as to leave the variable numbers of openings of the pronephros.

However this may be, apparently the openings of the pronephros develop as unclosed portions of the anterior end of the groove from which the duct arose, and they open into a space placed at the root of the mesentery close to the notochord and close to the point where in a previous stage the body cavity communicated with the muscle plates.

In the Amphibian, and apparently in the Teleostean, there is no marked structure corresponding to the intermediate cell mass of Elasmobranchii. The muscle-plate cavity is, after its separation from the general body cavity, only separated from the latter by a double layer of cells, forming its ventral wall and the wall of the body cavity; i. e. there is no portion of the body cavity at first continuous, but subsequently divided up by the coming together of its walls into a series of canals connecting the general body cavity with the muscle plates.

Now the glomerulus of the pronephros develops in a part of the body cavity anatomically corresponding to the intermediate cell mass of Elasmobranchii, only in Amphibia it does not, in this region, become divided up into chambers corresponding to the segments.

With this part of the body cavity, from the somatic walls of which the original groove arose, the openings of the head-kidney communicate. The number of these openings corresponds with the number of segments occupied by the pronephros in all those animals in which they exceed one, except Myxine; but the development of the pronephros in Myxine is not at all known, and its adult structure is, on the whole, obscure.

Turning again to Elasmobranchs, we find that the anterior knob of the segmental duct arises from the intermediate cell mass, i. e. from a part of the body cavity corresponding serially with that with which in the succeeding segments it later unites when the young segmental tubes acquire a communication with the segmental duct.

In Amphibia the segmental duct, when larval life is tolerably advanced, opens into a Wolffian tubule, which arises from a mass of cells, the origin of which is obscure, but which apparently does not appear till after the larva has left the egg. Now the Wolffian tubule of an Amphibian is homologous with that of an Elasmobranch; it is similarly constructed, and opens into the body cavity at a corresponding point. Hence we are driven to the conclusion that the cells from which the Wolffian tubule in an Amphibian arise are homologous with the intermediate cell mass of an Elasmobranch.

But in Amphibia these cells are not developed where, if Elasmobranch development is primitive, they should be; and appear later in a way which gives no clue to their relationship to the intermediate cell mass in Elasmobranchii.

What is the meaning of this extraordinary method of development ?

In Elasmobranchs the development of the segmental duct is modified, while the development of the mesonephros is primitive in its segmental arrangement and origin as a specialised part of an organ present at an earlier stage.

In Amphibia the development of the segmental duct is more primitive, but that of the mesonephros very modified, and this very latter fact always goes hand in hand with the presence of a pronephros. Turning to the pronephros, it is found to develop in continuity with the segmental duct. It is found to possess, with regard to its openings into the body cavity, a segmented structure. It is also found to possess a structure, the glomerulus, resembling extraordinarily closely the glomerulus of an ordinary Malpighian body of the mesonephros. This glomerulus lies in a special part of the body cavity, just as a glomerulus of a Malpighian body in the mesonephros of an Elasmobranch lies in what from its origin may be called a specialised part of the body cavity; and both these specialised sections in their anatomical position precisely correspond (see above, p. 457).

With all these similarities can the inference be avoided that the head-kidney is descended from the same primitive excretory system as the mesonephros, which has appeared early in development to supply the larva with an excretory organ, and has been able to retain a more primitive development ? The larva, having this, has not wanted the hinder part, and in consequence, having all its energy occupied while within the egg in developing those organs which it will really require as a larva, it leaves over the development of the organs not so required until after it is hatched; and in order that it may not be burdened by useless organs, the cells from which the tubules after appear and which should appear, if keeping the phylogenetic order, quite early iu embryonic life, in a way already indicated, are reduced so as hitherto to have escaped observation.

It is perfectly true that the pronephros does present peculiarities of structure not presented by the mesonephros, such as the unsegmented nature of the glomerulus, and in the fact that the tube connecting the cavity in which the glomerulus lies with the segmental duct not being coiled. But in the fundamental structure, i.e. in the possession of a glomerulus placed close to the main vascular channel (aorta), in the segmental arrangement of the openings of the segmental duct into the cavity (anatomically corresponding in both cases) containing the glomerulus, in the cavity containing the glomerulus being a specialised part of the body cavity; in all these points the pronephros and mesonephros resemble each other.

Assuming for the moment the truth of this suggestion, we find the pronephros to present that method of development which à priori we are bound to assume would be if it were not for disturbing causes, the development of the mesonephros, because it represents the most probable method by which the mesonephros and its duct can have arisen in phylogeny.

The question now arises, What are the disturbing causes winch in Amphibia have so changed the phylogenetic development ? The answer has already been given, but I will repeat it here. It has been brought about by the action of natural selection on the innumerable larvae produced, so that only those animals reached the adult state which in their prelarval and larval development conformed to the type of development we have before us.

Admitting the possibility of both prelarval as well as larval development varying at any particular stage, the tendency has been to produce a dissimilarity in the early structure of the excretory organs of Elasmobranchii and Amphibia greater than that which exists in the adult state, a result entirely in opposition to what we should expect from the application of that principle which has been laid down as regulating embryonic development, viz. that embryos of different animals, starting as fairly similar, become more and more dissimilar as their development proceeds.

To get any actual proof from embryonic development in favour of the above hypothesis must, from the nature of the case, be very difficult. For the very reason of the existence of the pronephros as an anterior part of the excretory system well marked off from the posterior makes it improbable that anything more than a trace of the hinder part should appear simultaneously in embryonic development with the anterior part. If the rest of the mesonephros developed continuously with the duct and simultaneously with the pronephros, then, on the above hypothesis, we should not be able to distinguish a pronephros from the hinder part; and it is opposed to all our ideas of economy to suppose that a rudiment of the mesonephros should appear at what phylogenetically would be the proper time, remaining over as a rudiment in the larva, i. e. as a useless organ forming merely a burden until it was wanted.

It seems to me that we can only expect, at the very utmost, to find a very small trace of the mesonephros in embryonic development at what phylogenetically we should consider, on the above hypothesis, to be the proper moment relative to the pronephros.

I have been examining the development of the segmental duct in an Amphibian, the frog, to see if at the time of closure of the groove of the segmental duct any trace of a discontinuous closure such as we find in the head-kidney existed. If the pronephros is merely the anterior part of a segmental organ of which the mesonephros is the posterior part, and if phylogeny is in any way repeated in the development of the pronephros, we should expect to find that the discontinuous (segmented, see above) closure of the pronephros would be repeated behind, showing some traces at least of the openings of the segmental duct and of the specialised part of the body cavity which later forms the Wolffian tubule and contains the glomerulus. So far it cannot be said that my search has been from my point of view successful. To get any evidence of what I was searching for requires a very complete series of sections in a state of preservation favorable for observation. The difficulties presented by the embryonic Amphibia in their early stages to such a successful result are very great. In the first place they are very brittle, and comparatively very few of the sections, even if thick, can be mounted uninjured. Of these, very few, indeed, can be obtained perfect, and those so obtained are apparently more difficult to see anything in than the thick ones. The cells are full of yolk granules which seem to escape and obliterate the outlines of the cells from the sight.

While my results have not been such as to unable me to speak with any confidence either one way or the other, yet on the whole they have convinced me that a re-examination with a new method of the development of the segmental duct in Amphibia, &c., would repay the trouble.

In the chick, on the other hand, the anterior part of the segmental duct, for the space of five segments, develops exactly in the manner of the segmental duct and head-kidney of the lehthyopsida. Are the cell cords connecting the duct and peritoneal epithelium in these segments rudimentary Wolffian tubules, or are they rudiments of a head-kidney ? In the absence of a continuous glomerulus opposite them they differ from the openings of the pronephros. In their development they resemble the latter. If they are Wolffian tubules they develop quite differently from all other Wolffian tubules. If they are rudimentary pronephric funnels, then the chick possesses a rudiment of a pronephros which resembles exactly the hinder developing Wolffian tubules.

It seems to me that these structures, under the light of the above hypothesis, present no difficulty, and I cannot help thinking that the discovery of their method of development is striking evidence in its favour. They belong, on that hypothesis, to the anterior part of the excretory organ, which has retained the primitive method of development originally characterising the whole organ. They, in some Avian ancestor, have constituted the first developed part of the excretory system, which has been utilised by the larva as its excretory organ. Supposing that Avian ancestor existed now, we should find that its larva possessed an organ which we should call pronephros, having a structure less modified probably from the hinder part of the excretory system than in the case of the lehthyopsida, i.e. an organ the serial homology of which, with the mesonephros, would no more be disputed than is that of the metanephros with the mesonephros.

It may be objected to this view of the anterior part of the Avian excretory system, that it differs in certain marked features from the pronephros of other forms. Of these differences the most important is, perhaps, the fact that there is always found an interval unoccupied by segmental tubes between it and the mesonephros. But in Amphibia Salamandra Fürbringer1 distinctly states that rudiments, as masses of cells, occupying the same relative position to the segmental duct as do segmental tubes, are found intervening between the two. If these rudimentary tubules underwent full development there would be no such gap as that we now find between the pro- and mesonephros of Amphibia.

But this difficulty is merely part of another difficulty which it seems to me must exist whatever view be taken of the nature of the pronephros, namely, why does this organ, so well developed in the larva and apparently perfectly well performing the functions of an excretory organ, atrophy in the adult ? And this difficulty only seems capable of the unsatisfactory explanation, that though perfectly well suiting the requirements of the larva, its position is unsuitable for the satisfactory performance of its functions in the adult. Balfour has suggested1 that the atrophy of the pronephros is due to its position in that part of the body cavity which eventually becomes the pericardium ; and has pointed out, as a confirmation of this view, that it only persists in the adult of those animals in which it is completely shut off from the body cavity, e.g. Teleostei.

(The enormous size which the pronephros attains in adult Teleostei is peculiar, but, coupled with the remarkably feebly developed mesonephros in the adult, is not astonishing. The pronephros seems capable of carrying on all the excretory work in some adult Teleostei, in which the mesonephros is not present. The absence of the mesonephros in these cases is probably purely secondary, and, no doubt, traces of it would be found if a close examination were made. The survival of a larval character into the adult state is paralleled by the Axolotl’s gills.)

A second feature of difference between this anterior part of the Avian excretory system and the Amphibian pronephros, is the absence in the former of a continuous glomerulus. This may be abortion from disuse, and does not really present a serious difficulty.

A third feature of difference is that the Avian pronephros extends over a much greater area than that of the Ichthyopsida, but when I draw attention to the fact that this difference is found amongst the various members of the Ichthyopsida themselves, I think it can hardly be looked upon as a difficulty. In Teleostei the head-kidney is distinguished by one peritoneal opening and a correspondingly short glomerulus. From this we have all stages to the five peritoneal openings of Petromyzon.

Finally, even if the Avian pronephros did differ in certain features from the Ichthyopsidan pronephros, this can hardly be regarded as a serious difficulty.

The pronephros of Teleostei with its Malpighian capsule containing the isolated glomerulus, and with its one peritoneal opening, surely differs considerably from the pronephros of the frog with its three peritoneal openings and its glomerulus lying free in the body cavity.

Again, without laying too much stress upon it, I point to the pronephros of Myxine, which differs still more remarkably from that of other types.

The difficulty presented by the Elasmobranchii, in which the tubules, though retaining certain primitive features of development, do not develop in continuity with the duct, is very great, and in the present state of our knowledge no satisfactory explanation, founded on facts of development, can be given of it. I will suggest a possible, but entirely rough and hypothetical, solution on the lines so far followed.

Before the Elasmobranchii produced eggs with the large food yolk they at present possess, they may have undergone a large part of their development in the surrounding medium as free larvm. These larvae must have left the egg at a time when the cavities of the muscle plates were still open to the body cavity, and when the segmental duct had only just commenced to be formed in front, and before the development of the vascular system, and therefore before the glomerulus, the functions of which were probably carried on by the walls of the body cavity. The segmental duct was quickly developed from a groove into a duct, the larvae thus precociously developing a recently acquired adult structure. With this constitution the larva of the ancestral Elasmobranch quickly developed the rest of its excretory system. In consequence of the larva having been hatched at a very primitive stage, before the muscle plates were separated from the body cavity, certain primitive characters in the development of the segmental tubes were retained. These characters have been more or less transmitted to the present day, this having been rendered possible by the acquisition of food yolk and abolition of the larval state.

However this may be, and it is useless now to make hypotheses of this kind, we can only wait till a more close study of Elasmobranch development has been made to see if any traces can be found of the disturbing cause which has produced the modification in the development of the excretory system assumed on the above hypothesis, and very possibly in the search along the lines which this hypothesis indicates quite a different view as to the phylogeny of the vertebrate excretory system may present itself.

Before concluding I will briefly state what I think to have been the structure of the primitive excretory system in the ancestral Vertebrate.

There was a duct occupying the position of the segmental duct, i. e. at the ‘dorsal outer angle of the body cavity, at the point where the latter becomes separated from the cavities of the muscle plates. This duct opened in each segment into the dorsal part of the body cavity. On the inner wall of the latter projected on each side a vascular ridge formed by the aorta. Behind, the segmental duct opened into the cloaca.

As differentiation proceeded the vascular aortic ridge became more especially developed opposite each opening of the segmental duct, and parts of each of these enlargements became successively enclosed in a special part of the body cavity, giving rise to the commencement of the secondary glomeruli. With this division of the glomerulus segmentally, and of each segment of it into further secondary glomeruli, each lying in a specialised part of the body cavity, the openings of the segmental duct began to fold and divide, incompletely at first, into special openings, one for each secondary glomerulus. “Finally, this division was completed, and the segmental duct communicated by a number of openings in each segment with specialised parts of the body cavity containing a portion of the original aortic ridge. The specialised parts containing these glomeruli being still open to the body cavity, and the glomeruli being still all distinctly attached by a common stalk to the walls of the body cavity, and the intermediate parts of the original continuous ridge having completely vanished, now the capsules enclosing the glomeruli became more and more completely marked off from the body cavity. The openings putting them in communication with the segmental duct elongated into tubules which became coiled, and the glomeruli themselves gained a greater independence of each other by a development of intermediate tissue.

A trace of the original state of things has descended to the present time in the pronephros, with its continuous glomerulus opposite the opening of the segmental duct, and placed in a specialised part of the body cavity. Differences in structure from the supposed primitive state of things have of course arisen, in consequence of the specialisation of the pronephros as the larval excretory organ.

In the same way a trace of the division of the primary glomeruli into primary, secondary, &c., glomeruli, is left in the curious development of the external glomeruli of the anterior part of the Avian mesonephros. Only in this case no cause can apparently be given for the retention of this primitive feature of development.

An examination of an early stage in the development of the Avian Wolffian tubules, when the primary and secondary tubules are both fairly well established, but not very complicated in structure, points very distinctly to the fact that the glomeruli of the two tubules are parts of one primitive glomerulus. They appear to be continuous, and while one looks ventrally, i. e. the so-called primary glomerulus, the other looks dorsally. A glance at the accompanying woodcut will make this clear.

If this drawing of a section through the Wolffian body of a chickin a part with primary and secondary tubules, be compared with fig. 24, which is from the anterior part of the same chick where there are no secondary tubules, it will be seen that the step between them is not great.1 It is merely necessary to suppose the division of the glomerulus (in fig. 24) into two parts, and a simultaneous development of certain folds from the Wolffian duct to form the tubules, and the original single tubule would have been transformed into a ventral primary and a dorsal secondary tubule.

Further, as I have pointed out in another paper,2 the secondary tubule always arises in close proximity, apparently from a blastema continuous with a part of that from which the primary tubule arose.

A modification of development is to be expected, because in those animals in which the mesonephros develops after hatching, it clearly comes gradually into use. The whole is not wanted at once, but with the increasing size of the larva, more tubules are wanted. The first developed (primary) in Salamandra acquire a structure with which they can apparently perform their function when there is hardly a trace of the secondary tubules (Fürbringer, loc. cit., fig. 26).

A cause of abbreviation is so clear in this case that I need not waste time in stating it.

But the whole details of the development of the secondary, &c., dorsal tubules needs reworking, for, with the exception of the observation of Mr. Balfour’s for Elasmobranchs, we have no real knowledge of their exact method of development. The result of such an investigation cannot but be exceedingly interesting from a phylogenetic standpoint.

I cannot help thinking, as before stated, that the development of the external glomeruli in the chick may have some interest in this relation.

The modification of the mesonephros of the Amniota is, on the above hypothesis, due to the fact that some Avian ancestor possessed a larva in which the anterior part of the excretory system was early developed, the development of the hinder part being deferred, and consequently modified, just as we see to be the case now in the Ichthyopsida.

The still greater modification und retardation of the development of the metanephros or true kidney of the Amniota, and the great size which the Wolffian body reaches in the embryo, are striking facts which demand consideration in any discussion of the Vertebrate excretory system.

In my paper on the “Development of the Kidney” I have stated my views on the relation of the Amniote kidney to the mesonephros. But one point in that paper is left untouched.

Why does the kidney appear so late ? and also why does the Wolffian body become so large and complex—so much larger than the small-sized chicks, in which it is fully developed, can need?

And, further, why should this organ, apparently so well adapted to serve as the excretory organ of the adult chick, atrophy ?

It may be said, in answer to the latter question, that only those tubules of the mesonephros which open into the cloaca independently of the Wolffian duct can function in the adult, as those which have not so changed their course would interfere with the function which the Wolffian duct later acquires—the carriage of semen.

It seems to me that the only answer which can be given to the first of these questions is this:

The kidney is thrown back in development for the same reason that the mesonephros of the Amphibia is, viz. because the ancestor of the chick underwent part of its development out of the egg, at which stage of development the testis, not being developed, did not interfere with the excretory functions of the Wolffian tubules, or vice versa. The large size of the mesonephros, then, is to be explained on the supposition that the larva of the chick’s ancestor used it for a considerable period of its early life as an excretory organ, so that it may be said that the pronephros holds the same general relation to the mesonephros in the Ichthyopsida as does the mesonephros to the metanephros in the Amniota.

I do not mean to affirm that the above explanation of the lateness of the development of the metanephros is absolutely valid, for I think that a careful consideration of the development of the hind part of the mesonephros in Amphibia and Elas-mobranchii might necessitate a slightly different explanation.

But an explanation of that kind must be sought to explain the remarkably late development in the chick of an organ which phylogenetically must be assumed to have had an origin simultaneous with that of the mesonephros.

With regard to the relation which the testes enters into with the mesonephros, it is interesting to notice the modified development which always characterises this connection.

Here it can be definitely affirmed that the lateness and consequent modification of the process is due to the fact that the apparatus has not been required in the larvae of the Ichthyopsida and of the Amniote ancestors, and consequently has been put off and modified in development. The explanation is exactly similar to that given for the modification in development of the Amphibian mesonephros, except that here we are supposed to be able to assert with greater reason that the putting off and consequent modification is due to the fact that the connection between the testes and mesonephros was not wanted sooner, and so was not developed.

The whole of the Vertebrate excretory system, including pronephros, mesonephros, and metanephros, are derived from a primitive organ possessed by the ancestral Vertebrate. This organ had a segmental character, and consisted of a duct, the segmental duct opening in every segment into the body cavity, close to a continuous structure, known now as the glomerulus, which was placed close to the main vascular channels and acted as an excretory organ.

The anterior end of this organ was used by the larva, and developing more or less with regard to other structures at the normal time, retained many primitive features of development originally characterising the whole organ, and is known to us as the pronephros. The posterior part of the organ had its development delayed with regard to other structures, particularly those in connection with which it primitively developed ; the development was consequently modified. This part is known to us as the mesonephros.

The same hypothesis was applied to account for the retardation and modification of the development of the metanephros with regard to the mesonephros in the Amniota.

The main facts in favour of the hypothesis are—

1. The development of the segmental tubes in Elasmobranchii and of the pronephros and segmental duct of the Ichthyopsida as parts of the body cavity.

2. The obvious modification in development of the mesonephros, accompanying also the presence of a pronephros in most of the Ichthyopsida.

3. The resemblance in structure between the pronephros and mesonephros, particular stress being laid on the fact that the glomerulus in both glands is developed in anatomically corresponding, i. e. homologous, parts of the body cavity

I may point out before leaving the subject that other views concerning thenature of the pronephros have been expressed by Gegen-baur, Fürbringer,1 and Balfour.2 The two former authors look upon the pronephros as having an antiquity greater than that of Vertebrates, greater even than that of the segmented ancestors of Vertebrates. They regard it as being descended from the primitive excretory system possessed by the unsegmented ancestor, which has been retained in such forms as Turbellaria and Rotifera, the segmented posterior part having been added when the segmented state was reached.

Balfour’s views as to the phylogeny of the Müllerian duct and its homology throughout the Vertebrata are well known. He supposes it is one or, in the chick, more of the head-kidney openings which have become modified for generative purposes.

I still adhere to the view expressed in the paper on the “Rudimentary Head-Kidney of the Chick” as to the meaning of the peculiar structures at the anterior end of the Müllerian duct, and I think that there are grounds, which it is not necessary to enter into here, for supposing that the abdominal opening or openings of the Mullerian duct have been derived from the anterior part of the excretory system after its modification to form the pronephros. But I quite admit that a fuller knowledge of the early development of the Elasmobranch segmental duct may necessitate an alteration in this view.

1

“Development of the Kidney in its relation to the Wolffian Body in the Chick,” ‘Quart. Journ. Mie. Sci.,’ vol. xx.

1

‘Arcli. Für Mic. Anat.,’ vol. xiv.

1

I have already given a preliminary account of the development of this structure in the ‘Proc. Cambridge Phil. Soc.,’ May 3, 1880.

5

‘Sitzungsberichte der Gesellschaft zur Bedford d. gesam. Naturwiss.,’ No. 5,1879.

Transverse Section through the Trunk of a Duck Embryo with about twenty four Mesoblasiic Somites am. amnion ; so. somatopleure ; sp. splanchnopleure ; wd. Wolffian duct ; st. segmental tube ; ea.v. cardinal vein ; ms. muscle-plate ; spy. spinal ganglion ; sp.c. spinal cord ; ch. notochord ; ao. aorta ; kg. hypoblast.

1

Fig. E, Pl. II, in the paper on the “Head-Kidney of the Chick,” ‘Quart. Joum. Mic. Sci.,’ vol. xix.

1

Urodela. 2 Anura. 3 Cascilia.

1

Fürbringer,‘Morph. Jahrbuch,’ Bd. 3, p. 5.

2

Scott, in a recent paper (‘Morph. Jahrbuch, ’ vol. viii), states that the segmental duct in Petromyzon, develops as a solid cord of cells from the somatic mesoblast, which subsequently becomes hollow. The peritoneal openings of the head-kidney are developed as outgrowths from the anterior end of this duct to the body cavity.

3

‘Jenaische Zeitschrift, ’ vol. vii, 1873.

1

There is a functional head-kidney in adult Ganoids. It appears to be formed on the Teleostean type (vide Balfour, ‘Comp., Embryology, vol. 2, p. 51).

1

In Petromyzon, Scott (see note, p. 445) states that this duct arises as2a solid rod of cells, which secondarily becomes connected with the bodycavity epithelium, to form the pronephric funnels. This account, in my opinion, needs confirmation.

3

Fürb.,’ p. 5, p. 42.

1

‘Elasmobranch Fishes,’ p. 260 el. seq.

5

Loc. cit.

1

Götte also, in his latest writings on the subject, agrees with Fürbringer as to the origin of the cells which give rise to the mesonephros. But 1 may draw attention to the fact that Gotte has held three views on this point, the last of which did not appear (see Fürbringer, loc. cit.) till 1875, i.e. after the publication of Balfour and Scmper’ s works on ‘Elasmobranchi.’

1

‘Fürbringer,’ loc. cit., p. 46.

2

‘Ibid., loc. eit., p. 12.

3

Self, ‘Quart. Journ. Mic. Sci.,’ April, 1880.

1

Balfour has recently described the existence of solid cords of cells, connected with the peritoneal epithelium, in the anterior part of the mesonephros of the sturgeon (‘Comp. Embryology,’ vol. ii, p. 581). The origin of these cords is not clear, neither is it certain that they undergo full development.

2

Loc. cit.

3

Loc. cit.

1

In making out the phytogeny of organs which have had an early origin, it seems to me that geology can help us ill this way (amongst others). Those forms which are found in the oldest rocks, and which have existed as small isolated groups, very little changed apparently in structure, to the present day, probably retain the same method of development now as then. By examining the embryology of such living forms we might expect to find the development of certain organs different to that in other animals belonging to larger living groups. Turning to the Brachiopoda, a group of great antiquity, we find a development of the body cavity which is shared by but few animals, and which à priori we regard as the most primitive method of development of that organ known. Now, of the animals which resemble the Brachiopoda in this respect, Balanoglossus, Amphioxus, and Sagitta are soft bodied, and so not found as fossils; but their very isolation at the present day, with regard to their relations to other groups, suggests that they are survivals of some larger groups, the other members of which have undergone so much evolution that their relationship is unrecoguisable. The other group, Echinodermata, which presents this method of development, is found at its greatest development. In Palæozoic rocks, and has not undergone any very marked changes since that time. It seems to me that, by following this line, some very important help might be obtained in helping us to decide questions of organ phylogeny.

1

‘Comp. Embryology.’

1

Spengel however asserts, that in the female of those Amphibia he has investigated, the kidney (mesonephros) contains an uniform number of segmental tubules in each segment over its whole area; while in the male, he finds that they increase in number behind.

1

Loc. cit., p. 19.

2

Spengel.

3

Braun.

4

Self, ‘Quart. Journ. Mier. Sei.,’ April, 1880.

5

There is no evidence that this is effected by intercalation in the chick at any rate.

1

Loe. cit.

1

‘Comp. Embryology,’ vol. ii.

1

It will be observed that iu this figure the tubule connecting the Wolffian duct and capsule is hardly developed. In all probability, this was on the analogy of the pronephros, the primitive state of things, the tubule, being a secondary differention of the duct near each glomerulus.

2

‘Quart. Journ. Mier. Sci.,’ April, 1880.

1

Loc. cit.

2

Balfour looks upon it as the most primitive part of the excretory system which has been retained by the larva, as so many ancestral organs are, long after they have been lost by the adult. ‘Comparative Embryology,’ vol. ii.