The Development of the larval and adult mid-gut of Calandra oryzae (Linn.): The Rice Weevil

IN November 1925, while examining a series of sections through the larva of the rice weevil (Calandra oryzae, Linn.), I was puzzled by the presence of a comparatively large mass of cells between the nervous system and the alimentary canal in the region where the fore-gut passes into the mid-gut. Later on, by examining sections through metamorphosing larvae, it was noticed that this mass takes part in the formation of the adult mid-gut. This suggested that a detailed study of the origin and behaviour of this mass, together with a study of the development of the larval mid-gut and that of the adult, might be of some interest to the embryologist and probably to the systematist also.

The development of the mid-gut epithelium of insects has for the last fifty years been the subject of numerous investigations, not only from the point of view of comparative insect embryology, but also from the point of view of its relation to the germ-layer theory. According to this theory the material from which the mid-gut epithelium is built should be derived from the innermost of the germinal layers, i. e. the endoderm.

It is not the intention here to enumerate the results obtained by the various investigators, nor to describe in detail the endeavours made to apply the germ-layer theory to the case of insects, since the whole problem has often been reviewed at length in the text-books of embryology, but only to give a brief account of the main views which still divide the more recent insect embryologists into two opposing groups.

1. Mid-gut epithelium derived from an anterior and a posterior endodermal rudiment and therefore of an endodermal origin.

Grassi (1884) showed that in Apis the mid-gut epithelium was derived from the anterior and posterior ends of the inner layer which he considered as a combination of the mesodermal and endodermal layers.

Kowalevsky (1886) observed the same phenomenon in Musca, and was led to regard the insect egg at the time of the formation of the germinal layers as comparable to a gastrula with an extremely elongated furrow-like invagination. Through this lengthening the endodermal layer was cut into halves, which were dragged apart to the anterior and posterior ends of the inner layer. From these two halves the larval mid-gut developed.

The view of Grassi as elaborated by Kowalevsky was accepted by Heider (1889), Wheeler (1889), Escherich (1900), Nussbaum and Fulinski (1906 and 1909), and in a modified form by Strindberg (1913).

Hirschler (1909) confirmed Grassi’s observations, and in 1912 gave an interpretation which is totally different from that of Kowalevsky. According to Hirschler there are two gastrulation phases in the development of pterygote insects. The first phase involves the segregation of the primary yolk-cells— which this investigator called primary endoderm—from the blastodermal skin which was referred to as the ectoderm. The second phase lies in the differentiation of the mid-gut epithelium rudiments from the inner layer. This observer also maintained these views in Schroder’s ‘Handbuch der Entomologie ‘(1924).

While all the above-mentioned investigators agreed that the rudiments of the mid-gut epithelium are differentiated from the inner layer, Carrière (1890), Carrière and Biirger (1897), and Noak (1901) claim that, at least in the species they investigated, the mid-gut epithelium rudiments arise quite independently of the inner layer, from proliferating areas of the blastoderm, one at each end of the germ-band corresponding to the future location of the stomo- and proctodaeum respectively. According to these investigators the proliferated cells represent the endoderm while the inner layer gives rise exclusively to the mesoderm. This view received partial support from the work of Nelson (1915).

1. Mid-gut epithelium derived from proliferations of the inner ends of the stomo- and proctodaeum and therefore of an ectodermal origin.

Ganin (1874) observed that in certain insects the mid-gut epithelium was derived from the inner ends of the stomo- and proctodaeum. He concluded that owing to the fact that these organs are developed from the ectodermal layer, the mid-gut epithelium which is derived from them is of an ectodermal origin.

Voeltzkow (1888, 1889, and 1889 a), working on Musca and Melolontha, while confirming Ganin’s observations, came to an entirely different conclusion. According to this observer, the stomodaeal and proctodaeal invaginations originated from the base of the gastrula-furrow and therefore belonged to the inner layer, a conclusion which meant that the stomodaeum and proctodaeum of insects were not of ectodermal origin. To the adherents of the germ-layer theory neither Ganin’s nor Voeltzkow’s conclusions were acceptable.

Heymons (1895), who devoted special attention to the development of the mid-gut in Ortboptera and Dermaptera, fully confirmed Ganin’s and Voeltzkow’s observations, and concluded that all the larval and imaginai structures of insects were derived from two germinal layers only, viz. the ectoderm and the mesoderm. This conclusion led Heymons to believe that the germ-layer theory had been shaken, and he endeavoured to explain the case of insects by assuming that the yolk-cells originally gave rise to the mid-gut epithelium. He also suggested that, owing to the gradual displacement of the original mid-gut epithelium by epithelial growths from the ends of the fore- and hind-guts, the assimilating action of the yolk-cells had become restricted to the embryonic period. This hypothesis received great support from Heymon’s discovery (1897) that in Lepisma, an Apterygote insect, the mid-gut epithelium is actually derived from the yolk-cells.

In 1903 Heymon’s hypothesis received further support by Madame Tschuproff-Heymons’s discovery that in some representatives of the O donata (Libellulidae) the middle portion of the mid-gut epithelium is derived from the yolk-cells while the anterior and posterior portions are formed from the stomo- and proctodaeum respectively. This worker considered this condition as constituting a transitional stage between the Lepisma type and that of the more highly developed pterygote insects as described by Heymons.

Lécaillon (1898), studying the embryology of certain Chrysomelid beetles, again confirmed Ganin’s and Heymons’s observations. This investigator concluded that, owing to the fact that insects are the most highly developed group among Arthropods, and also owing to the presence of a large mass of yolk in the developing egg, insects follow an atypical form of development, and their case can in no way affect the germ-layer theory which holds good for all the animals whose developmental processes have not been interfered with.

In considering the theoretical problems involved, Lécaillon accepted Heymons’s theory and proceeded to homologize the very early developmental stages of insects with those of other animals. According to this investigator there is no blastula stage in the more highly developed pterygote insects, and the period when the primary yolk-cells are segregated from the other cleavage cells corresponds to the gastrulation period of other animals.

The observations of the supporters of the ectodermal origin have been confirmed by Deegener (1900), Czerski (1904), Friederichs (1906), and Baling (1907) in the Coleóptera; Schwartze (1899) and Toyama (1902) in the Lepidoptera; Habito (1898) in the Orthoptera; and Pratt (1900) in the Díptera.

Kowalevsky (1887) described the adult mid-gut epithelium of Musca as derived from scattered groups of cells between the larval digestive epithelium cells. During metamorphosis these groups of cells, through active proliferation, form a continuous layer which eventually becomes the adult digestive epithelium, while the larval one is sloughed off into the lumen, awaiting rejection to the exterior when the adult emerges. With only one exception, such has been found to be the case in all holo-metabolic insects that have been studied. The exception we owe to Poyarkoff (1910), who found that in Galerucella ulmi, a Chrysomelid beetle, the adult mid-gut epithelium is derived from some of the cells proliferated from the posterior side of the oesophageal valvule.

To my knowledge no account of the embryonic or post-embryonic development of the, mid-gut epithelium of Calandra oryzae has been published, but Tschomiroff (1890) gave an extremely short account of the embryology of Calandra granaría, a closely-related species to the weevil under consideration. The larval mid-gut here was described as formed from the yolk-cells.

This weevil breeds without any obvious difficulty in an incubator at about 30° C.

The eggs are laid singly inside the grain.

The fertilized female when ready to oviposit eats out a hole large enough to receive its egg. Turning round, it brings the apex of the abdomen just above the hole, inserts its ovipositor, and lays the egg. The egg is elliptical in form, about 0·7 mm. in length and 0·3 mm. in breadth. It occupies about the inner four-fifths of the hole, which is filled up with a mucilaginous secretion flowing from the ovipositor. This secretion on drying up acquires the colour of the testa of the grain, thus making it extremely difficult to determine the spot where the egg has been laid.

Hatching occurs in about 4–5 days.

The larva is apodous and lives inside the grain.

There are four larval instars. The first and second require 1–2 days each, the third 5–6 days, and the last 7–9 days.

About 3 days before pupation the larva ceases to feed and, by sticking together the loose debris which has accumulated inside the grain, it constructs a neat cell, inside which it undergoes the whole process of metamorphosis. In this cell the metamorphosing larva contracts rhythmically, peristaltic movements being recognizable from the thorax backwards for about a day and in the meantime undergoes marked change in form. After being nearly hemispherical with the head capsule partially sunk in the thoracic region, and with no marked difference between the latter and the abdomen, the larva becomes very much reduced in girth with the head capsule well marked from the rest of the body. This stage, which is only a transitional one, is the ‘prepupal ‘stage.

After about a day the pupal moult sets in, and in the cell appears a pupa libera and the old larval skin.

The pupal stage lasts 6–7 days. Towards the end of this period the pigmentation of the adult begins.

When the adult emerges it has a brownish colour which gradually gets darker. After about 3 days inside, the adult bores its way to the open and the cycle starts again.

Besides following the development of the mid-gut in all the above-mentioned stages of Calandra oryzae, stages in the life-cycle of the following insects were examined for checking the results and also for comparative purposes.

The fixative that proved most suitable for this work was Carl’s, which is a mixture of 15 pts. absolute alcohol, 6 pts. formalin, 2 pts. glacial acetic acid, and 30 pts. distilled water. This gave the best results when used hot at about 70° 0. for from 3 to 8 minutes, according to the size of the object. It was then followed by 70 per cent, alcohol for at least 24 hours. After dehydrating, the material was then cleared in cedar-wood oil and embedded in paraffin wax in the usual way. Sections were cut 8μ thick and stained in Delafield’s haematoxylin.

From what has been mentioned about the life-history of Calandra oryzae, it is to be gathered that the duration of the embryonic, larval and pupal stages is not fixed. For this reason it has been thought advisable not to consider the age as a definite criterion of the progress of development, but to refer to the various stages according to certain developmental processes.

For the purpose of this work the earliest embryonic stages have been omitted, and the fully formed blastoderm is the starting-point of the developmental processes dealt with.

In the developing egg at the fully formed blastoderm stage, the following can be easily recognized:

1. A peripheral unicellular layer.

2. Yolk material.

3. Cells in the yolk.

4. A mass of cells at one end, between the peripheral cellular

layer and the yolk.

The peripheral unicellular layer consists of uniform cubical cells closely packed together round the yolk.

The yolk in fixed preparations is in the form of globules and, scattered among these globules, are the few ‘cleavage cells ‘that did not take part in the formation of the peripheral cellular layer. Similar cells in other insects have been described by various authors and are the ‘primary yolk-cells

The mass of cells at one end between the yolk and the peripheral unicellular layer is the genital rudiment, which now marks the posterior end of the developing egg.

In the advanced blastoderm stage the peripheral unicellular layer is no longer of a uniform thickness, there being a longitudinal band of columnar cells along one side and a similar band of distinctly flattened cells along the other, the two bands passing insensibly one into the other. The columnar band indicates the ventral side of the future embryo.

The appearance of a mid-ventral groove in the columnar band denotes the beginning of the formation of the germinal layers (fig. 1, Pl. 29). This groove first appears in the posterior region, but later on extends the whole length of the egg. During the formation of this groove two lateral furrows appear along the edges of the columnar cell-layer, and these mark off between them the definitive embryonic band of the developing egg (fig. 1, Pl. 29). This band is divided into three regions, viz. one median, which constitutes the walls of the mid-ventral groove, and two laterals.

The cells of the walls of the ventral groove multiply rapidly, and on account of this activity and also owing to the approach of the edges, the cavity of the groove is gradually reduced and finally disappears (fig. 2, Pl. 29).

The reduction and disappearance of the ventral groove causes the two lateral bands to unite, and this broad band of tissue has been regarded by all previous workers as constituting the ectoderm.

Some of the cells proliferated from the walls of the ventral groove retain their extra-vitelline situation while others wander into the yolk (fig. 2, Pl. 29). These two sets of cells are totally different from one another. The cells entering the yolk are very poorly defined and even form a syncytium with deeply staining constituents. The extra-vitelline cells, on the other hand, are irregular in form but with clearly marked cell walls (fig. 2, Pl. 29), and possess very feebly staining cytoplasm.

Soon after the disappearance of the ventral groove, the extravitelline products of its walls, probably through amoeboid movements, spread themselves between the yolk and the newly formed ventral columnar layer, i. e. the ectoderm (fig. 3, Pl. 29). In their new positions these cells are quite separate from one another and form a non-continuous layer.

By this time the entrance of the other cells into the yolk is completed. No evidence of them can be seen anywhere near the non-continuous layer. Thus we are now left with only two layers outside the yolk: the ectoderm, the origin of which has already been explained, and the non-continuous layer within, so that there are no anterior and posterior rudiments such as have been assumed to exist and have been described as endoderm.

In most of the details given the mode of formation of the germinal layers of Calandra oryzae is similar to that described in other insects, and the chief points of importance seem to be:

1. The absence of the so-called endodermal rudiments, and

2. The determination of the origin of the cells which disappear into the yolk, a point apparently missed by most other observers (cf. Graber, 1888; Voeltzkow, 1888; Wheeler, 1889 and 1893; Heymons, 1895; Schwartze, 1899; Noak, 1901; Petrunkewitsch, 1901; Schwangart, 1904; Hammerschmidt, 1910; Nelson, 1915; and Leuzinger, 1926..

Students of embryology recognize the ¹ gastrula ‘as the stage produced by an invagination of part of the ‘blastula ‘wall, the invaginated cells forming the lining of the primary gut, this layer being the ‘endoderm

In order to trace the origin of the endoderm in the pterygote insects, where the details of development are not so simple as just described, the first point to be decided is whether the insect blastoderm corresponds to the ‘blastula ‘of other types, and various views have been expressed upon this point.

Kowalevsky (1887) considered the blastoderm stage of the insect embryo as corresponding to the blastula stage of other animals. To this worker the blastoderm of insects was a blastula full of yolk. Kowalevsky also assumed that the preparation of this yolk for assimilation by the developing embryo was carried out by cleavage cells left in the yolk during the formation of the blastoderm, i. e. the ¹ primary yolk-cells ‘, and to this observer these cells had no phylogenetic significance.

As to the ventral groove, Kowalevsky homologized it with an elongated shallow gastrula invagination, and derived the endodermal and mesodermal layers from its walls. In the differentiation of these germinal layers, he assumed that the endodermal layer, through the lengthening of the gastrula invagination, was dragged apart to opposite ends of the developing embryo.

Strindberg (1913) differed from Kowalevsky in regarding the endodermal layer as being formed along the whole stretch of the ventral groove, and not at the anterior and posterior ends only. This observer also took into consideration the cells entering the yolk after the formation of the blastoderm, and concluded that they were undifferentiated endodermal cells, i. e. mesoendodermal cells.

On the other hand Lécaillon (1898) considered that in the development of the more highly organized pterygote insects there is no true blastula stage, and that the period towards the end of the cleavage, when some of the cells rise to the surface of the developing egg to form the blastoderm, corresponds to the gastrulation period. According to this worker the primary yolk-cells represent the endodermal layer, while the blastodermal wall which he called ‘primitive ectoderm ‘contains the rudiments of the mesodermal and ectodermal layers only. The groove itself Lécaillon considered as a coenogenetic feature of the developing insect egg, and referred to it as the ¹ mesodermal groove ‘.

The third and last view to be considered here we owe to Hirschler (1912). According to this investigator there are two gastrulation phases in the development of the pterygote insects. The first phase occurs towards the end of the segmentation period when the wall of the blastoderm is being formed and a few cells are left behind in the yolk. The peripheral layer Hirschler called ectoderm, while the cells left in the yolk, i. e. the primary yolk-cells, he called the ‘primary endoderm ‘. The second gastrulation phase is marked by the appearance of the ventral groove, whose cells later on differentiate into mesodermal and secondary endodermal layers, both of which remain outside the yolk. The cells entering the yolk after the blastoderm stage Hirschler considered of secondary endodermal origin and as representing a secondary phenomenon. Neither the blastoderm stage nor the ventral groove, according to this view, has a phylogenetic significance.

Now it seems to me that whether or not the blastoderm stage corresponds to the ‘blastula ‘stage depends entirely on the significance of the ventral groove and whether or not its walls contain the endodermal elements.

A consideration of the behaviour of the endodermal layer among Arthropods other than pterygote insects, and possessing eggs rich in yolk, will assist us in the solution of this problem. Among Crustacea (vide MacBride’s ¹ Text-book of Embryology’), as shown in the case of Homar us, Palaemon, Mysis, and representatives of Amphipods and Isopods; among Arachnids, e.g. Agelena, Euscorpius and Limulus; and in Myriapods as described for Scolpendra, the endodermal layer is proliferated either from the walls of an invagination or from a ridge or thickening along the blastoderm. Soon after proliferation the constituent cells enter the yolk and spread themselves there.

Again, among apterygote insects it is definitely established that the endodermal cells reside in the yolk. In Anurida mari tima, which possesses an egg comparatively poor in yolk, Claypole (1898) observed that the endodermal cells are differentiated at an early stage of development, probably by a process of gastrulation, and that these cells remain in the yolk quite distinct from the vitellophags. According to Heymons (1897) and Uzel (1897), in Lepisma and other apterygote insects, the endodermal cells are budded off into the yolk from a thickened area of the blastoderm. The budded-off cells soon become unrecognizable from the cells left in the yolk during the formation of the blastoderm.

Similarity between the cells entering the yolk in the abovementioned Crustacea, Arachnida, Myriapoda and apterygote insects, and those budded off into the yolk in the case of Calandra oryzae and the other pterygote insects, is at once suggested.

It is therefore not unjustifiable to conclude that in the pterygote insects the cells budded off into the yolk from the wall of the ventral groove are probably homologous with the endodermal cells of other Arthropods and apterygote insects, and thus represent the true endodermal cells: hence the wall of the ventral groove here contains endodermal elements, and in accordance with the embryological definition of the ‘gastrula ‘the ventral groove will represent an elongated gastrula invagination and the stage prior to the appearance of the ventral groove must be the ‘blastula ‘and not a post-gastrula stage, as Lécaillon and Hirschler had concluded.

Kowalevsky’s view, so far as it concerns the significance of the blastoderm, the meaning of the primary yolk-cells and the significance of the ventral groove, is fully supported by the writer’s observations on Calandra oryzae, but as to the mode of origin and the behaviour of the endodermal cells it is unsound, as will be shown now and later on.

Kowalevsky assumed that originally the endodermal layer was dragged apart to opposite ends of the embryo, and thus it was represented by an anterior and a posterior rudiment. My observations on Calandra oryzae fail to show the presence of such rudiments, but demonstrate very clearly the entrance of the endodermal cells into the yolk; a fact which is in full harmony with what is known to occur in other Arthropods including the apterygote insects.

A table, showing the views expressed by different authors, is here inserted (p. 327) as a summary of the first part of this paper.

During the formation of the germinal layers, the anterior and posterior ends of the embryonic band grow very actively. As a result of this growth, which is more marked posteriorly, two folds appear one at either end of the egg. These folds are the anterior and posterior amniotic folds.

The growth of the anterior end is more marked laterally where it forms the cephalic lobes.

On the other hand, the growth of the posterior end is in a longitudinal direction only, and the germ-band extends round the posterior end of the egg to the dorsal side where it continues to grow until it nearly reaches the anterior extremity. At this stage, therefore, the embryonic band is nearly twice the length of the egg.

The growth in length involves the three longitudinal strips of the embryonic band, i. e. the median or walls of the ventral groove and the two laterals. The newly formed material is being rapidly differentiated into endodermal cells which wander into the yolk, mesodermal and ectodermal cells which behave as described above.

As this growth proceeds the genital rudiment correspondingly shifts so that it is all the time situated in a morphologically fixed position, i. e. between the posterior end of the embryonic band and the yolk.

After the formation of the germinal layers the stomodaeum appears as an ectodermal invagination in the anterior part of the embryonic band, in a mid-ventral position (fig. 4, Pl. 29).

It is to be emphasized here again, that at this stage of development the germinal layers represented outside the yolk are the columnar ectodermal layer and the non-continuous mesodermal one together with the genital rudiment which is situated at the posterior end of the embryonic band.

The stomodaeum pushes its way dorsally between the scattered mesodermal cells, and for some time remains as a well-defined pit with uniform walls (fig. 4, Pl. 29; fig. 6, Pl. 30).

While the stomodaeum is appearing, peculiar structures, which probably are disintegrating cells, are to be seen just inside the yolk in the stomodaeal region (deg.c., fig. 4, Pl. 29; fig. 6, Pl. 30). Probably these structures are of mesodermal origin.

The proctodaeum originates in a similar manner from a median ectodermal invagination at the posterior extremity of the embryo. The pit soon becomes deep and at first is at right angles to the embryonic band (fig. 5, Pl. 29), but later on becomes parallel to it and extends forwards along it (fig. 11, Pl. 30). The wall of the proctodaeal tube at this stage is not of uniform thickness, since the morphologically dorsal side thins out gradually as it passes into the amnion (fig. 11, Pl. 30).

The posterior wall of the stomodaeal pit, which is rounded evenly (fig. 6, Pl. 30), soon becomes the seat of active cellular division which leads to the formation of two processes one on either side of the pit (fig. 7, Pl. 30). In fig. 8, Pl. 30, one of these processes is shown in an oblique section.

At first the component cells of these processes are packed together and thus exhibit a polygonal form. Their cytoplasm is dense, and in all respects is exactly similar to that of the cells of the stomodaeal wall. Gradually these cells become irregular in form and show light and vacuolated cytoplasm, but nevertheless they continue to divide actively and, insinuating themselves posteriorly between the yolk and the mesodermal cells, they form two lateral discontinuous bands one on either side of the mid-ventral line (fig. 9, PL 30). These bands, meet similar but less evident structures advancing anteriorly in a similar manner from the wall of the proctodaeum after the latter has given rise to the rudiments of the Malpighian tubules (fig. 11, PL 30).

The lateral bands, which are definitely derived from the posterior and anterior walls of the stomodaeum and proctodaeum respectively, contain the rudiments of the future larval mid-gut epithelium.

Towards the end of the proliferation of these bands numerous cells are to be seen migrating into the yolk from the blind end of the stomodaeum (fig. 10, PL 30). At first these cells are recognizable by the deep stain of their nuclei, but soon they lose this character and become indistinguishable from the ones already present in some of the spaces in the yolk. Nothing definite can be said about the fate of these cells. It is only certain that they do not take part in the formation of any definite structure. It is possible that they are destined to help in the preparation of’ the yolk for assimilation.

The derivation of the larval mid-gut epithelium from proliferations of the blind ends of the stomodaeum and proctodaeum was first claimed by Ganin (1874). Since then this claim has been supported by Heymons (1895), Korotneff (1894), and Habito (1898) in the Orthoptera; Witlaczil (1884) and Heymons (1897) in the Rhynchota; Schwartze (1899) and Toyama (1902) in the Lepidoptera; Lécaillon (1898), Deegener (1900), Czerski (1904), Friederichs (1906) and Baling (1907) in the Coleóptera; Pratt (1900) in the Diptera; and Heymons (1895) in the Dermaptera.

All the above-mentioned investigators concluded that owing to the fact that the stomo- and proctodaeum are ectodermal structures, the mid-gut epithelium derived from them is ectodermal in origin. This conclusion is obviously at variance with the germ-layer theory and, in spite of the numerous confirmations Ganin’s view had received, the derivation of the mid-gut epithelium from ectodermal structures remains unaccepted by practically all the embryologists.

The majority of those who contested Ganin’s view originally accepted Kowalevsky’s theory (see p. 323), and derived the mid-gut epithelium from anterior and posterior endodermal rudiments.

It has been mentioned (see p. 326) that the representation of the endodermal layer by two cellular masses one at either end of the embryonic band is quite contrary to what occurs among other Arthropods with eggs rich in yolk and also in Calandra oryzae.

Now, if the mid-gut epithelium of insects has an endodermal origin, one would expect it to be formed from the endodermal cells spread in the yolk. That this should be the case is evident from the study of the development of the apterygote insects. According to Heymons (1897), Uzel (1898), and Claypole (1898), the mid-gut epithelium of various representatives of this group is formed from the endodermal cells which were spread in the yolk in a way similar to that described for various Arthropods. That such probably was the case among the ancestors of the pterygote insects is an hypothesis which was originally suggested by Heymons (1895) and has since received support from various authors. Carrière and Bürger (1897) recorded that ‘at the time of the appearance of the endoderm bands (i. e. the rudiments of the mid-gut epithelium which these observers derive from an anterior and a posterior endodermal mass), the yolk-cells at the periphery, regularly distributed and connected together, form a complete sac, or if you will a primary mid-intestinal epithelium ‘(p. 358). A similar phenomenon has been described by Wheeler (1889), Heymons (1895), Hammerschmidt (1910), and Leuzinger (1926). One is therefore led to conclude that the anterior and posterior endodermal rudiments of the various authors are not homologous with the true endoderm of the apterygote insects, and if such rudiments do really exist they will be merely analogous in as far as they contribute to the formation of the mid-gut epithelium. This conclusion receives great support from a close examination of the works of Wheeler, Nusbaum and Fulinski, and Hirschler, all of whom endeavoured to demonstrate the presence of the anterior and posterior endodermal rudiments.

Wheeler (1889) described an anterior and a posterior thickening in the inner layer of Dory pho r a at the moment the stomodaeal invagination is formed. He claimed that from these thickenings the endodermal rudiments are differentiated. Wheeler’s fig. 88 which is intended to illustrate this differentiation in the posterior mass fails to do so. The liné of demarcation between what this author believed to be endoderm and the rest of the mass looks very artificial and probably was due to an artifact. As to the anterior mass this writer did not say much. From a comparison between Wheeler’s fig. 82 and a corresponding stage in Calandra it is perhaps not unreasonable to conclude that the posterior undifferentiated mass of this author is probably the genital rudiment. A clear idea of this can be gained by comparing the posterior part of Wheeler’s fig. 82 with the writer’s fig. 5, Pl. 29.

Figs. 3 and 4 of Nusbaum and Fulinski (1906) show at once that the supposed anterior endodermal rudiment is an outgrowth from the posterior wall of the stomodaeum. These workers (1909), while admitting that at certain developmental stages the rudiments of the mid-gut epithelium look like outgrowths of the stomodaeum and proctodaeum, assumed that in such cases the differentiation of the anterior and posterior endodermal rudiments is slightly retarded while the appearance of the stomodaeal and proctodaeal invaginations is somewhat earlier, so that each of the undifferentiated regions of the germband at the most anterior and posterior ends of the gastrula furrow correspond to the stomodaeal and proctodaeal invaginations respectively.

Nusbaum and Fulinski’s assumption is based on Kowalevsky’s theory, which is at variance with what occurs in the development of the apterygote insects and the other Arthropods with yolky eggs, and therefore it cannot be accepted as an explanation of the origin of the mid-gut epithelium.

Hirschler (1909), on the other hand, described the appearance of blade-like processes from the posterior stomodaeal wall of D o n a c i a, and stated that similar structures occur in certain other insects. His figs. 33 and 36 are intended to illustrate these processes. He, however, believed that these outgrowths gradually pass inwards into the stomodaeal wall, and probably form the separating lamellae between the stomodaeum and the mid-gut. In deriving this latter organ, Hirschler assumed that the median cells of the inner layer at the anterior end of the embryonic band are endodermal and derived the mid-gut epithelium from caps of flattened cells between the stomodaeal and proctodaeal invaginations and the yolk. However, Hirschler did not explain why these cells were to be regarded as endodermal.

Three years before Hirschler’s work the development of Dona ci a was studied by Friederichs (1906), who agreed with Hirschler in the description of events but described the processes as ectodermal, and said that they give rise to the mid-gut epithelium.

One is therefore greatly tempted to believe that the caps which Hirschler claimed to be the rudiments of the mid-gut epithelium probably correspond to the few mesodermal cells dorsal to the stomo- and proctodaeal invaginations, while the blade-like ectodermal outgrowths to which this author referred correspond to the stomodaeal outgrowths which give rise to the greater bulk of the mid-gut epithelium rudiments in Calandra.

It is now obvious that Wheeler, Nussbaum and Fulinski, and Hirschler have not been referring to identical structures in their description of ‘endodermal rudiments ‘, and that in no case do the structures they refer to exhibit any similarity of composition or origin and thus cannot be called ¹ endodermal ‘in the proper embryological sense of the word. In fact no clear exposition of the differentiation of the so-called ‘endodermal rudiments ‘has yet been given by any author, and my observations in the case of Calandra oryzae confirm the view that no such rudiments do differentiate in the regions where the mid-gut epithelium originates and that therefore the larval mid-gut is indeed ectodermal in origin.

Therefore we may conclude that the origin of the larval mid-gut epithelium in Calandra oryzae, and probably all pterygote insects with the exception of the Libellulidae (vide Heymons, 1896, and Tschuproff-Heymons, 1903), differs entirely from that in the apterygote insects and other Arthropods with yolk-bearing eggs where the true endodermal cells which at first wander into the yolk become the mid-gut epithelium. In most of the pterygote insects these cells disappear, and their function of producing the larval mid-gut epithelium is taken on by ectodermal cells proliferated from the stomodaeum and proctodaeum.

(e) Formation of the Larval Mid-gut and the Appearance of the ‘Accessory Cell-mass’.

The rudiments of the larval mid-gut epithelium which now lie beneath the yolk become closely associated with the inner walls of some of the abdominal mesodermal somites (fig. 12, Pl. 31). These portions of the somites are destined to form the muscular wall of the mid-gut.

The cells of the rudiments of the mid-gut epithelium divide actively and, together with their mesodermal associates, spread round the yolk in a dorso-ventral direction (fig. 13, Pl. 31).

During the spreading a very curious phenomenon occurs. Some of the cells proliferated from the mid-gut rudiments enter the yolk instead of spreading round it (fig. 13, Pl. 31). During their passage into the interior these cells undergo remarkable cytological changes. Their nuclei become much lighter than the original ones, while their cytoplasm becomes denser and acquires a very fine reticulate texture. These transformed cells agglomerate in the middle of the yolk as a comparatively large mass (figs. 13 and 14, Pl. 31). This group of cells has a peculiar history, as will be seen during the course of this paper, but as at present it is only certain that its cells are accessory alimentary cells with an unknown function, they will be referred to merely as ¹ accessory cells ‘.

Meanwhile the stomodaeal tube has been growing posteriorly into the yolk, and eventually it abuts on the anterior end of the accessory cell-mass (fig. 14, Pl. 31).

After the epithelial rudiments have budded off the accessory mass they continue to spread themselves round the yolk together with their mesodermal associates, so that by the time the formation of the dorsal wall of the embryo has been completed, the yolk is to be seen surrounded by a rather loose layer of irregular cells derived from the rudiments of the mid-gut epithelium, and a layer of elongated cells with long nuclei, derived from the inner walls of the mesodermal somites. The former layer will form the future larval mid-gut epithelium, while the latter will give rise to the mid-intestinal musculature.

During the growth period of the epithelial and muscular rudiments, the accessory cell-mass has been undergoing slight changes in position. Probably partly due to the activities of its own cells, and partly on account of the elongation of the stomodaeal tube, this mass is to be seen in a more anterior situation. Again, probably because of the development of the supraoesophageal ganglion, it is also pushed downwards to a more ventral position. As a result of these processes a portion of this mass now separates the ventral stomodaeal wall from the ventral rudimentary layers of the mid-gut. As development proceeds the accessory cell-mass slowly moves from the yolk into a sub-stomodaeal situation (fig. 15, Pl. 31). Eventually it entirely leaves the yolk (fig. 16, Pl. 31), and is to be found dorsal to the nervous chain and closely applied to the wall of the alimentary canal at the point where the stomodaeum opens into the mid-gut.

In its sub-stomodaeal position the accessory mass remains during the latter part of the embryonic life, and the whole period of larval growth apparently undergoing no visible change apart from increasing in size. Its fate and behaviour during metamorphosis will be discussed later on.

As soon as the accessory mass is extruded from the yolk, the cells of the rudimentary epithelial layer of the larval mid-gut become closely packed together to form a sac whose ends join the stomodaeal and proctodaeal tubes. The wall of this newly formed sac is the mid-gut epithelium of the larva.

At this stage the epithelial cells of the mid-gut are large with weakly staining cytoplasm in contrast with those of the stomodaeum and proctodaeum, which are much smaller in size and possess very dense cytoplasm (fig. 16, Pl. 31).

The rudimentary muscular layer gives rise to the circular and longitudinal muscles of the mid-gut which may be very loosely joined to those of the stomodaeum, but in all sections examined there is a gap between the two.

By the time the mid-gut is fully formed the other embryonic processes are also completed and hatching soon follows.

In the feeding larva the cells of the mid-gut epithelium, although retaining their large size, are quite different in character from the ones described for the fully formed embryo. The most marked change lies in the cytoplasm, which is now vacuolated and stains very deeply. Dispersed between these cells and situated at their bases are the replacement cells (figs. 17 and 18, Pl. 32).

Apart from these cytological differences, the presence of a striated fringe and a chitinous intima on the inner side of the mid-gut, and stomodaeum and proctodaeum respectively, marks the three regions of the alimentary canal.

The anterior portion of the mid-gut is dilated and ends posteriorly in a narrow long tube which is coiled twice and passes into the proctodaeum. Mesenteric caeca are only present on this narrow portion. They are papillae-form, consisting of about ten cells each.

The accessory mass must not be confused with the suboesophageal body described by Wheeler (1893), Nussbaum and Fulinski (1906 and 1909), and Hirschler (1909). Such a body is present in Calandra oryzae (fig. 14, Pl. 31) but it takes no part in the formation of the mid-gut as Nussbaum and Fulinski and Hirschler have suggested, and in the insects these authors investigated no accessory mass has been described.

In Calandra the cells of the sub-oesophageal body are mostly binucleate, suggesting that this body is nephrocytic. (Compare the ventral nephrocytes of some Dipterous larvae (Keilin, 1917), the peri-oesophageal nephrocytes of Pediculus hum an us (Nuttall and Keilin, 1921), and similar structures of Cimex lectularius (Puri, 1924).) This body is to be seen in Calandra before the migration thither of the accessory mass (fig. 14, Pl. 31), and is actually pushed away from the wall of the gut by the advancing accessory cells (fig. 16, Pl. 31).

Examination of eight Curculionid species, three Scolytids, two Anobiids, two Chrysomelids, and one Lamellicorn (see p. 320), revealed the fact that an organ similar in all respects to the accessory mass is present in Calandra granaría, Odoiporus glabricollis, and Hylobius abietis, while a similar organ but in a totally different situation is present in Rhyncolus lignarius. In this last-mentioned species it occupies a dorsal situation somewhere near the junction of the mid-gut and the proctodaeum.

While one can perhaps assume that in Calandra granaría, Hylobius abietis and Odoiporus glabricollis, the mode ôf origin and behaviour of the accessory mass is similar to that described in Calandra oryzae, one is hardly justified in drawing a similar conclusion for Rhyncolus 1 i g n a r i u s, and it is hoped to carry out further research upon this species as soon as the necessary material has been procured.

The processes involved in the metamorphosis of the mid-gut begin towards the end of the last larval instar about three days before the pupal moult sets in. The habits of the larva and the external changes in form during this period have been described on p. 319.

The internal changes are most peculiar and are begun by the collapse of the ventricular portion of the gut to a narrow tube with an irregular wall. The posterior portion also undergoes diminution in lumen. Not only does the larval mid-gut diminish considerably in width, but also in the same time it is markedly reduced in length. Its coils become less evident, and it is now transformed to a narrow and almost straight tube which is blocked anteriorly with numerous degenerating cells and is totally cut off from the posterior end of the fore-gut (fig. 19, Pl. 32).

During the collapse, the replacement cells between the bases of those of the epithelium grow to form a more or less continuous flattened layer, which envelops the degenerating epithelial wall and separates it from the muscular one (fig. 19, Pl. 32).

As a result of these processes the larval mid-gut epithelium with its replacement cells, except that it remains loosely joined to the proctodaeum, now lies loose within the muscular wall (fig. 19, Pl. 32). The latter is undergoing regeneration, and will become the muscular wall of the adult mid-gut-

At the same time the cells at the posterior end of the fore-gut become elongate, multiply rapidly, and obhterate the opening into the larval mid-gut. The stomodaeum then becomes cut off from the latter and grows backwards into a blind tube inside the transforming muscular layer of the mid-gut (fig. 20, Pl. 32), pushing before it the degenerating larval epithelium until the new tube reaches the proctodaeum. By the time this happens only a few irregular cells and scattered pieces of chromatin represent the last of the larval mid-gut (fig. 21, Pl. 32). The walls of this newly formed tube will later on yield the definitive epithelium of the adult mid-gut.

As soon as the collapse of the larval mid-gut begins, the accessory mass awakens from its larval quiescence. After being closely applied to the ventral side of the alimentary canal (fig. 17, Pl. 32), it fuses with the wall of the metamorphosing fore-gut (fig. 19, Pl. 32). This fusion takes place in the region between the muscular walls of the fore- and mid-guts of the larva (fig. 18, Pl. 32). The anterior margin of the fusion denotes the anterior boundary of the mid-gut tube.

The fusion now extends posteriorly (fig. 20, Pl. 32), probably partly due to the intercalary growth of the portion of the midgut tube fused with the accessory mass and partly to the activities of this mass itself. Extension also occurs laterally, and eventually the mid-gut tube is surrounded by the accessory cells (fig. 20, Pl. 32). Posterior extension continues, and by the time the mid-gut tube reaches the proctodaeum only about one-third of it remains free from these cells (fig. 21, Pl. 32).

Up to the stage when the mid-gut tube reaches the proctodaeum its lumen is extremely narrow (fig. 21, Pl. 32), but it then becomes enlarged anteriorly (fig. 22, Pl. 33) owing to the appearance of a homogeneous substance which is probably secreted by the cells of the walls of this tube. This substance is the yellow body; a term which was originally given by Weismann (1864) to the contents of the gut of Musca at a similar stage of development.

During the secretion of the yellow body the pupal moult sets in and the development of the adult mid-gut continues without any break during the pupal stage.

(d) Appearance of the Mesenteric Caeca and the Fate of the Accessory Mass.

The development of the mesenteric caeca, which begin to appear all along the mid-gut, is more easily seen in the posterior region where the accessory cells are lacking. Here they appear as groups of cells derived from the wall of the mid-gut tube. These groups of cells are numerous and are scattered all over the outer surface of the wall.

The individual groups, through active division of their cells, form small tubes closed at their outer ends and opening into the lumen of the gut (fig. 25, Pl. 33).

In the anterior region, although development commences, the caeca do not grow into tubes, probably because of the pressure of the accessory cells which collect round the budding caeca, squeezing them, so that the lumens are obliterated (fig. 24, Pl. 33).

By the time the formation of the mesenteric caeca is completed, the accessory cells, which once formed a more or less regular layer round most of the mid-gut tube, have become segregated into small groups, forming the outer walls of most of the caeca.

(e) Formation of the Definitive Mid-gut Epithelium of the Adult.

During the formation of the mesenteric caeca the remainder of the cells of the mid-gut tube have been undergoing active division and reorganization. After being extremely irregular and distinctly flattened, where the accessory cells exist (fig. 22, Pl. 33) they become more or less cubical, and by the time the accessory cells have become incorporated in the walls of the numerous mesenteric caeca these cubical cells form a columnar layer which is the definitive mid-gut epithelium of the adult.

From the beginning of the development of the adult mid-gut its lumen has been in direct communication with that of the metamorphosing fore-gut. The hind-gut communication, however, is not effected until after the definitive mid-gut epithelium has been fully formed.

By this time the other metamorphosis processes have been completed and the adult is almost ready to emerge.

The post-embryonic processes detailed above can be summed up as follows:

1. The degeneration of the larval mid-gut and the formation of that of the adult start about three days before the pupal moult sets in.

2. The larval mid-gut epithelium with the replacement cells collapses, degenerates and totally disappears before the emergence of the imago.

3. The adult mid-gut epithelium is derived entirely from the posterior end of the metamorphosing fore-gut.

4. The accessory alimentary cells envelop the developing mid-gut epithelium and later on become localized in the walls of most of the mesenteric caeca.

It was pointed out on p. 336 that in certain other Rhynchophora the accessory mass had been observed, and in the forms studied the type of adult mid-gut formation is similar to that described for Calandra oryzae.

In Scolytus destructor, Anthonomus pomorum, and Anthonomus grandis no accessory cell-mass is present, but in all these cases examined the adult mid-gut epithelium arises exactly as in Calandra oryzae.

This shows that the accessory mass has nothing to do with the type of development of the adult mid-gut epithelium described above. To the embryologist the significance of this type lies in the fact that the adult mid-gut epithelium is derived from a structure which has a definite ectodermal origin.

This recalls and supports what has been stated earlier in this paper about the origin of the larval mid-gut epithelium of most pterygote insects.

The activities of the accessory cells during imaginai life have not been so far studied. From a brief examination it can be said that the number of these cells becomes markedly reduced. After being densely aggregated in the walls of most of the mesenteric caeca (fig. 24, Pl. 33), they apparently disappear gradually during the imaginai life of the individual until in aged adults only a few cells are to be found near the distal ends of these caeca (fig. 26, Pl. 33).

In considering the possible function of these cells a digestive one is at once suggested on account of their close intimacy with the digestive tract, but there is at present no experimental evidence in support of this. If this should prove to be correct, it suggests that probably the adult extracts something from the food material which was unobtainable by the larva.

The type of formation of the adult mid-gut epithelium described here for Calandra oryzae and other Rhynchophorous species is totally different from that described by Regnel (1897) in Tenebrio molitor, Karawaiew (1898) in Anobium paniceum, Deegener (1904) in Cybister roseli, and Korschelt (1924) in Dy tiscus marginalis. In all these various Coleopterous insects the adult mid-gut epithelium was found to develop from islands of replacement cells scattered at the bases of the larval mid-gut epithelium cells, and my own observations on Ptinus’ tectus agree with those of the authors mentioned. This ‘Ptinus’ type can easily be differentiated from the ¹ Calandra ‘during the pupal stage. At this stage in insects with the ‘Ptinus ‘type the ‘yellow body ‘is to be found surrounding a large mass of cellular debris which is the sloughed-off larval mid-gut epithelium (Text-fig, 1, B, relep). In those insects of the ‘Calandra’ type the ‘yellow body ‘contains no remains of the larval mid-gut epithelium (Text-fig. 1, A, and fig. 22, Pl. 33), since such remains have been continually pushed backwards in front of the advancing midgut tube (Text-fig. 1, A, relep, and figs. 20 and 21, Pl. 32).

On the question as to which of these two types is to be regarded as the more primitive, we might agree with Pérez (1902) that the formation of the adult mid-gut epithelium from scattered replacement cells is comparable to the ordinary processes of regeneration occurring in the functional larval and nymphal mid-gut of the various insects.

Now the replacement cells are represented in the larva of Calandra oryzae (figs.17andl8, Pl.32). During metamorphosis they form a more or less continuous layer (fig. 19, Pl. 32), but this layer undergoes the same fate as the cast-off larval midgut epithelium. Possibly, therefore, similar replacement cells originally gave rise to the mid-gut epithelium of the ancestral adult.

The failure of the replacement cells to form the epithelium of the adult mid-gut in the Rhynchophorus species studied recalls the reference earlier in this paper to the failure of the endodermal cells to form the larval mid-gut—two points, the significance of which must for the present remain obscure.

The type of development of the adult mid-gut epithelium described for Calandra oryzae, on the other hand, exhibits a distinct relationship to that described by Poyarkoff (1910) in Galerucella ulmi, a phytophagous beetle. In this insect Poyarkoff found that during metamorphosis the larval epithelium is stripped off with its replacement cells from the muscular wall and is passed into the hind-gut. According to this observer, the cells of the posterior side of the oesophageal valvule multiply rapidly and extend backwards in the form of a solid plug, totally blocking the space once occupied by the mid-gut epithelial sac. These proliferated cells, which are fusiform, are of various sizes. Some of the smallest become fixed to the metamorphosing muscular wall and are destined to form the definitive adult mid-gut epithelium, while few larger ones form a provisional epithelium which lasts only during the pupal stage. The remaining fusiform cells perish and take no part in the formation of the adult organs.

The important point in Poyarkoff’s account is the formation of the adult mid-gut epithelium from some of the cells proliferated from the posterior end of the fore-gut.

An incomplete examination of Melasoma populi, another phytophagous beetle, suggests that in this species the development is similar to that described by Poyarkoff.

This resemblance between examples of Phytophaga and Rhynchophora, if it turns out to be characteristic of the groups, is interesting because of the view of recent systematists as to the close relationship between these two series which Lameere (1903), on certain anatomical characters, actually united into one group, Phytophaga.

With reference to the assumed relations of the Phytophaga and Rhynchophora, it is worth mentioning that the Lamelli-cornia agree with these two series in the possession of compound pedicellate testis which are only present in these three groups among the Coleóptera. An examination of Dorcus paral-lelopipedus, however, revealed the fact that the mode of adult mid-gut formation is of the ‘P tinus ‘type. If the same is true of other Lamellicorns, it may indicate either that this test is of no value or that the Lamellicornia are not very closely related to these series.

With regard to the question of origin of the Phytophaga and Rhynchophora, Kolbe (1908) thought that they have been derived from primitive Clavicornia. Ganglbauer (1903), on the other hand, suggested that these series might have had a common descent with the Malacodermata. Possibly the systematic study of the mode of formation of the adult mid-gut, which I am undertaking, will throw light on this point.

1. The blastodermal stage of insects corresponds to the blastula stage of other animals.

2. The primary yolk-cells are a secondary feature of the developing insect egg.

3. The ventral groove corresponds to the invagination of the gastrula.

4. The endodermal cells of insects are budded off into the yolk from the walls of the ventral groove or from thickenings in the blastodermal wall.

5. The endodermal cells inC. oryzae and most Pterygota do not give rise to any larval or imaginai structures.

6. The larval mid-gut epithelium in C. oryzae and most pterygote insects is derived from the inner ends of the stomodaeum and proctodaeum and is therefore of ectodermal origin.

7. The accessory alimentary cell-mass present in the larva of Calandra oryzae is developed from the rudiments of the larval mid-gut epithelium.

1. The metamorphosis of the mid-gut in C. oryzae begins about three days before the pupal moult occurs.

2. The larval mid-gut epithelium collapses and degenerates.

3. The replacement cells take no part in the formation of the mid-gut of the adult.

4. The adult mid-gut epithelium in Calandra oryzae and five other Rhynchophorus species is derived from the posterior end of the metamorphosing fore-gut, and is therefore of definite ectodermal origin.

5. The accessory alimentary cells are incorporated in the walls of most of the mesenteric caeca.

6. The presence of the accessory cell-mass has nothing to do with the ‘Calandra’ type of formation of adult mid-gut epithelium.

7. The number of the accessory cells is reduced during imaginai life.

8. In insects with the ‘Calandra’ type the ‘yellow body ‘contains no remains of the larval mid-gut epithelium. Such remains characterize the ‘Ptinus ‘type, and differentiate it from that of Calandra.

9. The ‘Ptinus ‘type probably is the more primitive.

This work has been carried out under the supervision of Professor F. Balfour-Browne, without whose constant encouragement and stimulating advice it would not have been accomplished. The writer also has to thank Professor E. W. MacBride, Professor H. Graham Cannon, and Dr. J. W. Munro for the interest they have taken in this work and the help they gave.

All figures are from Camera lucida drawings.

acc c, accessory cell; acc m, accessory cell-mass; am, amnion; ap, appendage; be me ca, beginning of a mesenteric caecum; be mg r, beginning of midgut rudiment; be mg t, beginning of mid-gut tube; br, brain; ch i, chitinous intima; ci mus, circular muscles; co me ca, core of mesenteric caecum; deg c, degenèrating cell; deg I ep, degenerating larval mid-gut epithelium;

ect, ectoderm; end, endoderm; fg, fore-gut; ge r, genital rudiment; i w so, inner wall of mesodermal somite; If, lateral furrow; Ig mus, longitudinal muscles; Ip, lateral plate; Mal.t, Malpighian tubule; mdp, middle plate; mes, mesoderm; mg ep, mid-gut epithelium; mg ep c, mid-gut epithelium cell; mg r, mid-gut rudiment; mg t, mid-gut tube; mig c, migrating cells; mus c, muscle cell; mus w, muscular wall; neph, nephrocyte; nrb, neuroblast; phag, phagocyte; proof, proctodaeum; pr ye, primary yolk-cell; p w stom, posterior wall of stomodaeum; re I mg, remains of larval mid-gut epithelium; re y, remains of yolk; r Mai, rudiment of Malpighian tubule; ry mg w, rudimentary wall of mid-gut; ry mus w, rudimentary muscular wall; rp c, replacement cells; so b, sub-oesophageal body; stom, stomodaeum; stom mus,stomodaeal muscles; stomw, stomodaeal wall; strfr, striated fringe; vg, ventral groove or gastrula furrow; vnch, ventral nerve-chain; y, yolk; y b and yeb, yellow body.

All Figures are from Sections of Eggs.

Fig. 1.—Ventral portion of a transverse section through an advanced blastoderm showing the gastrula furrow (vg) and the lateral furrows (If). ×380.

Fig. 2.—Ventral portion of a transverse section to illustrate the entrance of the endodermal cells (end) into the yolk. × 380.

Fig. 3.—Portion of transverse section showing the mesoderm (mes) and the ectoderm (ect) after the entrance of the endodermal elements into the yolk, × 380.

Fig. 4.—Portion of transverse section through the early stomodaeal pit. × 380.

Fig. 5.—Portion of a longitudinal section through the early proctodaeal pit. × 400.

All Figures are from Sections of Eggs.

Fig. 6.—Portion of a transverse section through the posterior wall of the early stomodaeal pit. × 380.

Fig. 7.—Portion of a transverse section through the posterior wall of the stomodaeum to show the beginning of the formation of the rudiments of the mid-gut epithelium of the larva, × 380.

Fig. 8.—Portion of an oblique section through the posterior wall of the stomodaeum and one of the rudiments of the mid-gut epithelium of the larva (mgr). × 380.

Fig. 9.—Portion of a transverse section through the thoracic region of the embryo to show the presence of the rudiments of the mid-gut epithelium (mgr) on either side of the mid-ventral line, × 380.

Fig. 10.—Portion of sagittal section illustrating the entrance of cells from the blind end of the stomodaeum into the yolk. × 400.

Fig. 11.—The posterior portion of a sagittal section showing the beginning of one of the rudiments of the larval mid-gut epithelium derived from the proctodaeum (mg r). × 400.

All Figures are from Sections of Eggs.

Fig. 12.—Lateral portion of a transverse section showing the association of one of the rudiments of the larval mid-gut epithelium (mg r) with the inner wall of an abdominal mesodermal somite (i w so). × 380.

Fig. 13.—The inner portion of a transverse section showing the budding off of the cells of the accessory mass (acc m) from the rudiments of the larval mid-gut epithelium into the yolk. ×300.

Fig. 14.—The anterior portion of*a sagittal section showing the stomodaeum abutting on the accessory mass (acc m). This figure also shows the sub-oesophageal body (so b). ×300.

Fig. 15.—Portion of a sagittal section showing the accessory mass migrating from the yolk into a sub-stomodacal situation. × 400.

Fig. 16.—Portion of a sagittal section through a fully formed embryo showing the accessory mass (acc m) outside the lumen of the alimentary canal. This figure also shows the binucleate cells (neph) derived from the sub-oesophageal body. × 400.

Figs. 17 and 18 are from Sections of a third Instar Larva.

Figs. 19 and 20 are from Sections of Prepupae.

Fig. 17.—Portion of a sagittal section showing relative position of accessory mass in the growing larva. × 80.

Fig. 18.—Portion of a sagittal section illustrating the difference between the cells of the stomodaeum and those of the mid-gut. This figure also shows the gap between the muscles of the stomodaeum and those of the mid-gut. × 600.

Fig. 19.—Portion of a sagittal section through a metamorphosing larva, showing the collapse and the degeneration of the larval mid-gut and the fusion of the accessory mass (accm) to the ventral side of the stomodaeum. × 80.

Fig. 20.—The same as the previous one but more advanced, showing the increased fusion of the accessory mass and the growth of the stomodaeum to form the mid-gut tube (mg t). × 80.

Fig. 21.—The same as the previous sections but more advanced, showing the remains of the larval mid-gut epithelium (re I mg) and the mid-gut tube nearly reached the proctodaeum. × 80.

Figs. 22–5 are from Sections of Pupae.

Fig. 26 is from a Section of the Gut of an aged Adult.

Fig. 22.—Portion of a longitudinal section showing the lumen of the midgut tube much enlarged. × 80.

Fig. 23.—Portion of a longitudinal section of the wall of the developing mid-gut showing the beginning of the formation of a mesenteric caecum where the accessory cells are present. ×380.

Fig. 24.—Section through one of the anterior mesenteric caeca in an advanced pupa showing the accessory cells (acc c) arranged round a core of cells continuous with the cells of the mid-gut epithelium, × 220.

Fig. 25.—Section through one of the posterior mesenteric caeca. × 220.

Fig. 26.—Section through one of the most anterior mesenteric caeca in an aged adult showing the reduction of the number of the accessory cells. × 220.