The Later Embryology of Rhodnius prolixus

The present paper deals with the embryology of the bloodsucking bug, Rhodnius prolixus Stäl, from the end of the gastrulation stage until the embryo hatches from the egg. The earlier development has already been described (Mellanby, H., 1935). The technique used for the later stages is similar to that which was employed for the earlier ones. Celloidin embedding was continued, as the results obtained were better than those from ordinary wax embedding. Whole mounts were sometimes stained over night with dilute paracarmine (Meyer’s); this was done if the pigment in the embryos was not well developed. Older embryos (from 12 days onwards) were also dissected to obtain the exact relationships of the internal organs.

The adult bugs were kept at 21° C. and the eggs were incubated at 21° C. and 90 per cent, relative humidity. Sections and whole mounts were made of eggs and embryos from 5 days old to 29 days old (the day of hatching under the above conditions) with daily intervals between each age of the eggs. As previously stated (Mellanby, H., 1935), the temperature of 21° C. is one at which development proceeds very slowly, and this justifies the daily interval between successive stages used for this work.

At 5 days the embryo shows cephalic, thoracic, and abdominal regions, segmentation being absent in the latter. Appendages have begun to appear on the head and thorax, and the general relation of the embryo and the egg is similar to that figured for a slightly later stage (Text-fig. 2). The position of the embryo is reversed with regard to the egg axis, because of the early invagination and blastokinesis which bring it into a superficial dorsal position with the head end of the embryo at the posterior pole of the egg. The head of the embryo is flexed round the posterior pole of the egg so that its extreme anterior portion is ventral in position. This position of the embryo is similar to that described for other Hemiptera which have been studied; for instance, Aphids (Witlaczil, 1884), Cimex (Heymons, 1899), Notonecta (Heymons, 1899), Pyrrhocoris ap- terus (Seidel, 1926), Belostoma flumineum (Hussey, 1926). In Cimex and Pyrrhocoris the embryo is flexed both at the head and abdominal ends, and the median portion of the embryo lies ventrally in the yolk, whereas the embryo of Rhodnius always lies superficially on the dorsal side of the yolk.

The head of the embryo in front view presents the appearance of Text-fig. 1. It will be seen that the cephalic lobes are well developed. The labrum appears to have a paired origin somewhat similar to that of Cimex and Pyrrhocoris (Heymons, 1899) and Pieris (Eastham, 1930). From the posterior margins of the cephalic lobes the antennae grow out as blunt rounded appendages. At their base, on their inner side, there appears another pair of small appendages, which are only rounded knobs; they are resorbed and they are not visible in embryos older than 6 days. They would appear to be the premandibular appendages. In 5-day embryos the mouth appendages and the thoracic appendages are all simple outpushings of the body-wall, but those of the future legs are slightly larger than the others. The segmentation of the thorax is not yet very distinct, and whereas the head and thoracic parts of the embryo lie very superficially on top of the yolk, the abdomen is slightly sunk in the yolk.

At 6 days the embryo has reached its maximum length. Not only does the extreme posterior part of the abdomen almost touch the egg operculum when the chorion is intact, but the abdomen is flexed downwards into the yolk because it is now too long to lie straight out (Text-fig. 2). The appendages of the head and thorax have elongated. Those of the thorax have turned inwards from the former lateral position as they increased in length. The bifid labrum has grown downwards over the stomodaeum, which is just being formed at this time. A bright red pigment has begun to appear in the ectodermal parts of the head round the cephalic lobes. This pigment is present in the nymphal and adult stages of the bug, and it has been suggested by Wigglesworth (1933) that it may be an excretory by-product of the cells. The quantity of pigment present seems to be subject to much individual variation in embryos. It also appears to be affected by temperature, embryos developing at temperatures above 80° C. having very much more pigment and being very red (Mellanby, H., unpublished work).

Segmentation of the thorax is now very obvious, and the first four segments of the abdomen are also clearly distinguishable (Text-fig. 2). Along the mid-ventral line of the embryo there runs a distinct furrow extending from the mandibular segment to about the middle of the abdomen. This is the neural groove, which is also very conspicuous in sections of this and slightly later stages (Text-fig. 9). The somites appear as distinct square blocks of tissue in whole mount embryos. The embryo is covered by the amnion and the serosa, which can easily be removed with dissecting needles.

At a slightly later stage the total length of the embryo has decreased, and the embryo, particularly in the abdominal region, has become broader. This decrease in length has resulted in the abdominal portion of the embryo becoming straightened out, since it can now be accommodated in the egg without flexure. When the first six abdominal segments have differentiated, limbs appear on the first and second abdominal segments. They are small lateral outgrowths (Text-fig. 2). They soon become covered up by the downward growth of the third thoracic limbs, and can only be seen properly after their removal. Appendages belonging to the first abdominal segment have been noted in most of the other Hemiptera whose development has been observed. They also occur in other insect orders such as the Orthoptera, some Coleoptera, and Eastham describes them in Pieris. Hussey (1926) gives a detailed description of their development in Belostoma flumineum and Ranatra, f u s c a; the stage at which they arise in these two bugs agrees exactly with their time of origin in Rhodnius. It may be worth noting that their position in Rhodnius is lateral (Text-fig. 2), whereas those of some other bugs, namely Cimex and Pyrrhocoris, grow out ventrally from the body-wall. In the Polyctenid, Hesperoctenes fumarins, Hagan (1931) shows that these abdominal appendages are extremely well developed and that they occupy a very large proportion of the embryo. He ascribes to them a nutritive function.

By the eighth day of incubation the chief advance in development is seen in the further shortening and broadening of the embryo and in the differentiation of the appendages. The labrum still shows its paired origin. The antennae are very prominent (Text-fig. 3), and the first and second maxillae have elongated somewhat. The mouth appendages show an increase in size as we pass backwards, the mandibles being the smallest, the second maxillae the largest. The thoracic appendages, on the other hand, are all of equal length and are considerably longer than the largest pair of mouth appendages. The abdomen now possesses eight segments and a terminal portion. Though abdominal limbs are developed on all the abdominal segments, they are only well developed on the first two segments. On all but the latter segments, appendages are not readily visible in whole mounts; they are much more obvious in sections.

During the next 3 days the embryo becomes still shorter, and its posterior extremity comes to lie at about one-third the length of the egg from its anterior end. The features of the embryo at this stage are shown in Text-fig. 4. The head of the embryo remains at the posterior end of the egg, and the cephalic flexure is still very evident. The first three head-segments are nearly at right angles to the rest of the body, and the cephalic lobes curve back on to the ventral surface of the yolk. Except for the head, the embryo is not visible on the surface of the egg because the thorax and abdomen are slightly embedded in the yolk, and to see the embryo properly the latter has to be removed (Text-fig. 4). It will then be seen that the appendages have all grown in length. The labral lobes now appear externally as a single median appendage, though internally they still show evidence of their paired origin. The first maxillae have become bi-segmented. This early division into two seems to occur in the maxillae of all Hemiptera. The basal joint gives rise to the maxillary plate of the nymph and adult, while the distal joint becomes the maxillary stylet (Imms, 1925). The abdomen now shows ten segments, and a small terminal unsegmented portion. The stomodaeum and proctodaeum are visible because of the extension inwards of the ectodermal red pigment which first appeared in the head (see p. 3). By 11 days the pigment is to be found in bands along the intersegmental membranes of the thorax and abdomen, and also there is one patch on each thoracic leg (Text-fig. 4). The pigment is particularly in evidence on the lateral portions of the head, where the eyes later develop. The first pair of abdominal appendages are no longer present as outgrowths of the first abdominal segment. They have become sunk into the ectoderm of that segment, and form a pair of glands (see p. 12). The remaining abdominal limbs have been resorbed.

After 12 days of incubation at 21° C. the second stage of blastokinesis occurs. The embryo, which has been developing with its head on the ventral side of the yolk while the rest of its body is on the dorsal side, now moves its thorax and abdomen through 180°, while its head moves forward to the anterior end of the egg (see Text-fig. 4 A). At the same time as this movement occurs, the embryo becomes much broader, with the result that it now covers the ventral side of the yolk (except at the anterior end) (Text-fig. 6). During the next 3 days all the appendages elongate very considerably. The labrum grows downwards in the mid-line. The mandibles of each side become invaginated into ectodermal pockets at their base along with the maxillae; the second maxillae have met in the mid-line to form the labium, and they form a groove which is open dorsally. The thoracic legs soon reach the posterior end of the abdomen and then become bent round so as to lie against the opposite sides of the embryo. The embryo moves right up to the anterior end of the egg, and its body-wall at the sides begins to grow up dorsally round the yolk so as finally to enclose the latter within the embryo. The two sides of the embryo meet along the middorsal line of the egg, first at the extreme anterior and posterior ends, and rather later in the middle. The yolk is completely enclosed after 16 days of incubation (Text-fig. 5). At this stage a thin iridescent cuticular membrane, which can easily be removed, surrounds the embryo.

From 16 days until 29 days there is very little difference in the external appearance of the embryo, though the internal changes are great. The eyes, which up till now have been patches of red pigment, gradually darken, and they are quite black 3 days before hatching. The openings of the respiratory system, the spiracles, become obvious little pink spots at about 20 days.

When the embryo is about to hatch, the pressure of blood in the head dislodges the operculum, and the insect slowly emerges with the operculum resting on the top of its head. When about two-thirds of the way out, a split occurs in the dorsal thoracic region of the cuticle. The insect moults as it proceeds with the end of its hatching. The shed cuticle usually remains attached to the inside of the chorion by the antennal covering, but not infrequently the insect hatches with its larval skin still attached to its third pair of legs. This ‘embryonic moult’ is a true moult (Gailliard, 1934), the cuticle being spiny and enveloping each appendage separately. Such moults are known to occur in other insect orders such as some Orthoptera (Williams and Buxton, 1916).

The formation of the amnion and the serosa at the time of the first invagination of the germ-band in the yolk has already been described in my previous paper. It remains to trace the fate of these membranes at later stages.

After the embryo has become invaginated into the yolk, it remains covered by the amnion until the second stage in blastokinesis (see p. 8). The serosa forms a very thin layer over the yolk, except at the extreme anterior end, where it is slightly thickened (Text-fig. 14). The amniotic cavity, which is at first very small, becomes enlarged as the appendages grow. They extend into the cavity and cause the amnion to be raised up from the body-wall of the embryo. The layer of yolk between the amnion and serosa, which was at first quite considerable, becomes gradually reduced until amnion and serosa are no longer separated at the head end of the embryo, and only slightly separated in the thoracic and abdominal regions (Diagram, p. 8). Growth of the head region is probably responsible for the rupture of the amnion and serosa. After rupture the thorax and abdomen of the embryo move on to the ventral side of the yolk. The amnion, which is now continuous with the embryo and with the serosa (Text-fig. 6), encloses the yolk dorsally, and the serosa becomes very much thickened and restricted to the anterior end of the egg. The thickened serosa consists of cubical cells with large nuclei.

Both Slifer (1932) and Roonwal (1935) find that blastokinesis in the grasshopper and in Locusta is due to movements of the living embryo, this being the first activity shown by the embryo. I have been unable to confirm this in B, hodnius owing to the opaqueness of the chorion. Previously it was thought that the sudden thickening of the serosa together with the growth of the embryo were responsible for the movements of the embryo.

While the embryo is growing to enclose the yolk, the area of the amnion becomes more and more restricted. Finally, amnion and serosa contract to form a compact cup-shaped invagination of cells in the dorsal thoracic region of the embryo. This ‘dorsal organ ‘is made of very regular cells which contain many vacuoles (Text-fig. 7). It becomes overgrown by the ectoderm of the body-wall, and the cells disintegrate in the anterior part of the embryo’s mid-gut. Hagan (1931) describes a similar fate for the embryonic membranes of a viviparous Polyctenid, only in his case there was only a mass of cells in place of the neatly formed ‘dorsal organ’.

During the process of blastokinesis the embryo and the membranes covering the yolk become surrounded by a layer of pale non-cellular material which does not stain. A little later the outer part of this material has very much the appearance of a cuticular layer. At a stage when the embryo has just enclosed the yolk, the whole embryo is surrounded by this thin iridescent membrane, which can be easily removed. This layer of cuticle does not appear to be made by the hypodermal cells, since there is no appearance of the characteristic changes in the various ectodermal structures which should precede cuticle formation (Wigglesworth, 1933). Also this layer does not envelop the appendages individually; it simply surrounds the embryo like a sac. The appendages of the first abdominal segment become sunk into the ectoderm and form conspicuous glandular structures by the time the embryo is 11 days old (Text-fig. 8). These glandular structures or ‘pleuropodia ‘produce long extensions of their cells, which become very conspicuous from 11 days onwards until 2 days after the finish of blastokinesis. It seems reasonable to assume that the formation of this third embryonic membrane is due to the activity of these cell extensions, and that the cell extensions make the cuticular layer. Wheeler (1890), who first described these curious first abdominal appendages in insects, decided from the evidence that he collected, that they had a glandular function. The general opinion seems to be that the function of this gland is to secrete material which helps to rupture the embryonic membranes. In the Polyctenid described by Hagan (1931), the relatively enormous pleuropodia surround the embryo with their cell extensions which then secrete a cuticular layer. Hagan claims that this layer protects the embryo (which develops in the oviduct of the mother) and that it absorbs food from the oviduct. Rhodnius would appear to have a similar arrangement, and here also this may be capable of having a nutritive function. Before blastokinesis there is a quantity of yolk between the amnion and serosa (Text-fig. 16). After evolution this yolk is left outside the embryo and the amnion (Text-fig. 6), but it becomes enclosed by the ‘third embryonic membrane’ and as the yolk disappears it is presumably absorbed through the ectoderm of the embryo.

When at 18 days the embryo has begun to lay down its own cuticle round the ectoderm, the cell extensions of the pleuropodia have become much reduced, and only a very thin cuticular layer remains.

The nervous system is formed from neuroblast cells which are differentiated early in development before segmentation is tvell advanced, i.e. at about 5 days. The first neuroblasts appear at the same time as the neural groove on the ventral side of the embryo. They are ectoderm cells whose nuclei become specially large and pale- staining. They lie internally to the other ectoderm cells, and are situated on either side of the neural groove. At first there is only one chain of neuroblasts on either side (see previous paper), but soon there are three. They run the whole length of the embryo. By 7 days the neuroblasts have begun to separate from the outer ectoderm cells or dermatoblasts, and a space appears between the two which is the beginning of the subneural sinus (Text-fig. 14). The detailed formation of the ventral chain has been described for Pieris by Eastham (1930); for other references see his paper. In Bhodnius the ventral chain appears to arise in an exactly similar fashion. It is at first unsegmented, but later becomes segmented when the germband segments. Protoplasmic extensions of the ganglion cells give rise to the nerve-fibres. By the time that the embryo is 11 days old, there are three thoracic and eight abdominal ganglia.

The brain arises from neuroblasts situated at the anterior end of the embryo. The neuroblast cells in the protocerebral segment cover a very large area, and they give rise to the large flattened protocerebral region of the brain. The deutocerebrum and tritocerebrum develop from neuroblasts in the antennary and premandibular segments. The developing stomodaeum pushes its way between, and rather ventral, to the tritocerebral ganglia (Text-fig. 10). Sections of 11-day embryos show the three ganglia forming the future supra-oesophageal part of the brain, and the three ganglia of the mandibular, first and second maxillary segments forming the sub-oesophageal nerve-mass. In dissections of older embryos (14—20 days) the brain shows quite definitely that it is composed of three supra-oesophageal parts, and that there are four more large ganglia in the rest of the head and thorax.

At 13 days there are still separate abdominal ganglia, but between then and 20 days concentration of abdominal and thoracic ganglia occurs. The abdominal part of the ventral nerve-cord appears to be pulled up into the thorax. This concentration occurs as soon as blastokinesis is finished (Text-fig. 6). In 18-day-old embryos, there is a supra-oesophageal nerve-mass, a fairly large sub-oesophageal mass, and three thoracic ganglia.

The last thoracic ganglion is very much elongated and represents the abdominal ganglia fused with it (Text-fig. 11). There is little difference between this stage and the one at which the insect hatches. The third thoracic ganglion becomes more compact, but remains much the largest of the three (Text-fig. 22).

The sympathetic nervous system arises from the dorsal wall of the stomodaeum. Cells in the mid-dorsal part of the stomodaeum become palestaining in 11-day-old embryos. These cells enlarge and become more rounded in outline than those of the rest of the stomodaeal wall. Transverse sections through 13-day embryos, in the posterior region of the trito-cerebral ganglion, show a rounded mass of pale cells which are continuous ventrally with the ectoderm of the stomodaeum (Text-fig. 10). Most of this mass of cells develops into the sympathetic ganglion of the adult, but from the anterior part, the frontal ganglion is formed. At first the frontal ganglion is very close to the sympathetic ganglion. Later the frontal ganglion moves forward to a position in front of the brain, but remains connected by a small string of nervecells to the sympathetic ganglion. This string of cells forms the recurrent nerve and runs along the dorsal stomodaeal wall. Anterior to the frontal ganglion a frontal nerve is differentiated from the mid-dorsal line of the stomodaeal wall. In 15-day embryos this differentiation may be seen in progress. In embryos younger than 15 days the sympathetic nervous system consists of pale rounded cells only. At 15 days nerve-fibres begin to be formed in the ventral part of the sympathetic ganglion. These fibres enter the posterior part of the tritocerebral ganglion on each side and thus become continuous with the fibres in the posterior part of the brain. The fibres form an arch over the stomodaeum.

All the parts of the sympathetic nervous system of the adult Rhodnius (see Wigglesworth, 1934) are now accounted for. There remains the corpus allatum, which in the adult is closely associated with the sympathetic ganglion and receives its nerve- supply from it. This structure is paired in most insects, but in the nymphal and adult Rhodnius it is a single structure lying dorsally to the oesophagus immediately behind the sympathetic ganglion. It originates as a pair of ectodermal cellmasses which appear to invaginate at the base of the first maxillae. They start to be formed in 12-day embryos. They soon lose connexion with the ectoderm and are to be seen in 15-day embryos as a pair of solid cell-masses lying against the sides of the stomodaeum near its blind end. By 18 days these masses have become continuous dorsally. The stomodaeum is now much longer, so they are to be found some distance in front of its blind end, and are situated close behind the sympathetic ganglion. A nerve grows backwards from the sympathetic ganglion into the corpus allatum.

The corpora allata of insects usually originate as paired invaginations of the mandibular segment as in Apis (Nelson, 1915), Pieris (Eastham, 1930). Heymons (1895) found that they originated from the maxillary segment in Dermaptera, and Carriere and Bourger (1898) figure them as a pair of cell-masses on either side of the fore-gut. In Bhodnius they appear to arise from the maxillary segment, but it is difficult to be certain of this fact, since the mandibles and maxillae have already become pushed into the head and the nervous system has begun to become concentrated by the time that the corpora allata are first visible.

Insect embryos develop paired ectodermal invaginations on most of their segments. In the head these invaginations form the internal skeleton and various glands, while in the thorax and abdomen they give rise to the tracheal system. Eastham (1930) gives a detailed account of this process in the head of Pieris, and he maintains that they help to solve the problem of head segmentation.

In a Rhodnius embryo of 11 days the various ectodermal invaginations are beginning to be easily recognizable. These are paired shallow pits, whose cells in most cases contain red pigment. They occur behind the antennae, in front of the mandibles, behind the mandibles, behind the first and second maxillae. The three thoracic, and the first eight abdominal segments each have a pair.

The Tentorium and Associated Glands.—The anterior arms of the tentorium are made from the tubular ingrowths behind the antennae. They elongate dorsally, then bend towards the posterior end of the embryo and grow backwards at the sides of the brain. In embryos more than 20 days old they appear bright red. In Text-fig. 12 (a) they are shown in transverse section. They finally meet the posterior tentorial arms, and where they join a cross-piece, the body of the tentorium, is developed below the oesophagus.

The invaginations just in front of the mandibles, which may belong to the premandibular segment, have from the first a very narrow lumen. They extend inwards towards the mid-line and later join the anterior tentorial arms near their point of fusion with the posterior arms. These are the dorsal tentorial arms.

The invaginations of the mandibular segment are large, and at 12 days they form a conspicuous gland-like structure, situated laterally.

The opening of the first maxillary invagination is carried into the ectodermal pocket in which the bases of the mandibles and maxillae are to be found from 12 days onwards. These invaginations grow outwards and upwards to join the anterior tentorial arms, and they form the posterior arms of the tentorium.

The labial invaginations form the salivary glands. They are situated laterally, and they grow inwards and backwards, coming to lie on either side of the oesophagus. Their glandular portions become very prominent bilobed structures on either side of the thorax. Each gland connects with the stomodaeum by means of a long straight duct (Text-figs. 12 (a), 18 (b)).

While all these head invaginations are developing, the mouthparts become concentrated so that it is difficult to determine the exact extent of the head segments. The head itself elongates greatly and becomes flexed ventrally against the thorax. The head of the nymphal and adult Rhodnius is a very compact structure (Muir and Kershaw, 1911), and for these reasons it is difficult to follow the fate of the invaginations.

Three pairs of thoracic and eight pairs of abdominal spiracles develop in the first instance. All these persist except the eighth abdominal, which closes. The main tracheae develop as ectodermal tubes from the bases of the spiracles. In each segment there is a transverse as well as a longitudinal trachea developed. All the main tracheae contain the red pigment and are therefore very easily seen in sections (Text-figs. 12 (b), 16).

The oenocytes are very large cells which are derived from ordinary ectoderm cells. They begin to appear in 12-day embryos, and by 13 days there are several in each segment near the spiracular invaginations. They appear to have a segmental arrangement, and they are lateral in position lying immediately beneath the epidermal cells. These first oenocytes are very large (Text-fig. 12 (b)). Later they become more numerous, and smaller. The oenocytes in insects are probably concerned with synthesizing some of the non-chitinous constituents of the cuticle (Wigglesworth, 1933). During the moulting of the Rhodnius nymphs, the oenocytes go through a definite cycle. They reach a maximum size before the new cuticle is formed and then they decrease. In the embryo the cuticle of the embryonic moult is laid down from 18 days, and it is well formed at 21 days. The oenocytes become more numerous but noticeably smaller after 21 days. So the embryonic moult would appear to be preceded by the same oenocyte cycle as occurs before the nymphal moults.

The ectodermal invagination to form the stomodaeum is the beginning of fore-gut formation. This starts to take place just later than 5 days. The invagination begins anterior to the antennae (Text-fig. 13), but the latter soon take up a pre-oral position. At first the stomodaeum is only a shallow pit, but it becomes deeper and its lumen narrows. The anterior endoderm proliferation is attached to the blind end of the stomodaeum, and these endoderm cells are carried inwards. By 11 days the fore-gut forms a definite channel, which is growing upwards and backwards between the neuroblast cells which will form the tritocerebrum, and those which will form the nerve ganglia of the mandibular segment. The cells lining the stomodaeum contain the red pigment already mentioned (see p. 3), and as a result of this the structure is easily seen in whole mounts. The fore-gut remains short until blastokinesis is finished. After that it lengthens rapidly, grows in an anterior direction and then curves round dorsally and extends backwards through the thorax. When the cuticular layer is being laid down on the outside of the embryo, a very distinct cuticular layer is developed surrounding the lumen of the fore-gut; the cavity is almost obliterated by it (Text-figs. 11 and 22). By 19 days the stomodaeum has reached its maximum length; the walls of the tube become much thinner as its length increases. In the region of the metathorax the blind end of the fore-gut comes in contact with the first part of the mid-gut. When the embryo is ready to hatch there is still no through connexion between the fore- and mid-gut, though they are in contact with one another. Presumably the stomodaeum becomes continuous with the mid-gut when the yolk has all been used up some days after hatching. It is well known that the first stage nymphs will not feed for several days after hatching (Buxton, 1930).

The hind-gut starts as an ectodermal invagination just later than the stomodaeum. At 8 days it is a straight, rather narrow tube ending blindly. The posterior endoderm rudiment is continuous with the cells at its blind end (Text-fig. 14). By 11 days the Malpighian tubules have begun to develop. They are outgrowths from the sides of the proctodaeum close to its blind end (Text-fig. 15), and they appear to be ectodermal structures. The Malpighian tubules have been claimed to have an endodermal origin in Pieris (Henson, 1932), but I am unable to support this origin for Rhodnius. The Malpighian tubules early become surrounded by mesoderm cells (Text-fig. 15). They are four in number, and between 11 and 15 days they lengthen considerably and become coiled, so that they may be cut several times in one section. The proctodaeum remains a relatively short straight tube, the form which it takes in the adult (Wigglesworth, 1931). The Malpighian tubes communicate with the anterior end of the proctodaeum.

During the formation of the germ-layers in Rhodnius, there appears an anterior and a posterior region of proliferating cells. These masses of proliferating cells are near the position of the future stomodaeal and proctodaeal invaginations of the embryo. The anterior mass is very like that found by Eastham (1927) in the embryo of Pieris. A large number of these cells are given off into the yolk in company with lower-layer cells along the whole length of the germ-band. At a slightly later stage these masses appear to be considerably smaller than they were at first, and they become attached to the blind ends of the invaginating stomodaeum and proctodaeum (Text-fig. 14). They next begin to proliferate very actively, especially at the sides of the stomodaeum. They form a very close and compact cellmass. Longitudinal sections of 11-day-old embryos show the endoderm cells in the region of the proctodaeum, and it appears here that the ectodermal cells of the blind end of the proctodaeum pass indistinguishably into the more compact rounded cells of this posterior cell-mass (Text-fig. 15). Both anteriorly and posteriorly, a strand of cells passes out from these proliferations, and forms a thin layer of cells on the yolk side of the lower layer or mesoderm (Text-fig. 21 (b)). In 11-day embryos the strands from the stomodaeum and proctodaeum have nearly met to form a continuous sheet of cells. From this sheet of cells the median portion of the adult mid-gut is formed.

The mid-gut of insects is claimed by some workers to be of ectodermal origin, and by others to be endodermal (see Eastham’s review, 1930). Mansour (1934) finds the adult mid-gut of some beetles to be made from larval ectoderm, and this he claims supports the view that the mid-gut of the embryo in the rice weevil, Calandra oryzae (Mansour, 1927) is also of ectodermal origin.

Richards (1932) attempts to rationalize these opposing views on mid-gut derivation by regarding the determination of the mid-gut as ‘a function within the whole’ (Driesch). The cells which are in the correct position when endoderm is due to be formed will be ‘determined ‘to form mid-gut.

Henson (1932) homologizes the anterior and posterior proliferations, found at either end of the germ-band in many insects, with the blastopore of Peripatus (Sedgwick, 1885). In his work on the embryology of Pieris (1932), in which reference is made to that of Eastham (1927) on the same insect, he refers to the ‘mesendoderm ‘rudiments as anterior and posterior blastoporic areas. The blind ends of the stomodaeum and proctodaeum are then not ectodermal; they are blastopore lips, and as such are capable of proliferating ectoderm, endoderm, and mesoderm. This appears to be a reasonable solution of the problem, and it holds good for the various methods of mid-gut formation described for insects. In Rhodnius the cells at the blind end of the stomodaeum and proctodaeum do look like blastopore cells. Here ectoderm and endoderm pass indistin-guishably into one another (Text-fig. 15).

The method of endoderm formation in Rhodnius agrees with the findings of Seidel (1926) for the bug Pyrrhocoris apterus, and is similar to that found for Pieris (Eastham, 1930) and for the beetle Europa terminalis (Paterson, 1932).

The nymphal and adult Rhodnius has a mid-gut made of three distinct parts. The oesophagus communicates with a short narrow anterior portion of mid-gut which leads into a very wide crop. From the crop a narrow convoluted tube communicates with the hind-gut.

In the embryo the most anterior part of the mid-gut develops from the endoderm cells which do not migrate to form a layer round the yolk. It becomes covered on the outside by mesoderm from which the muscle-layer is derived. This muscle-layer is continuous with that which surrounds the next part of the mid-gut. This second part is very large, and by 18 days the yolk has become restricted to its interior, the whole forming a ‘yolk-sac’ (Text-fig. 11). The hindmost part of the mid-gut is made by a cavity appearing in the cells of the posterior endoderm proliferation. It is considerably longer than the first part, and it is convoluted. The cells in all parts of the mid-gut are large, pale-staining, and vacuolated from 14 days onwards.

After the germ-band has become sunk into the yolk, a lower layer of cells is formed along the whole length of the embryo. A gastral groove first develops, and when this has been overgrown, the lower layer is complete. At the anterior and posterior extremities, the lower layer gives rise to the endodermal proliferations. The cells in between constitute the mesoderm. The mesoderm is at first a solid band of cells lying on the yolk side of the ectoderm. After the overlying ectoderm becomes segmented the mesoderm becomes constricted segmentally. At 6 days the mesoderm has become spread out over the ectoderm and arranged in two longitudinal bands which are connected in the mid-line by a very thin layer of mesoderm. The cells in the longitudinal bands become neatly arranged to form somites. A small cavity appears in the middle of each somite. These are the coelomic cavities. After the ectoderm has become pushed out to form the appendages, mesoderm migrates into these hollow ectodermal outpushings and this enlarges the coeloms (Text-fig. 14). The neat arrangement of the somite is upset, and the coelomic cavities extend down into the appendages. In the head and thorax the somites are only present for a short time, as it is here that the appendages first develop. After the derangement of the somites the coelomic cavities also open towards the yolk in the mid-line, and in the thorax the mesoderm is very obviously continuous longitudinally (Text-fig. 14).

By 11 days the mesoderm, in a typical segment shows the beginnings of the organs and tissues which it is going to form (Text-fig. 16). The thin median portion overlying the nervecord begins to break up into large cells which break loose, and are now to be found occupying the space between the nervecord and the yolk. These cells are the blood-cells, and they begin to migrate at 8 days. By 11 days the median part of the mesoderm has almost disappeared. The yolk recedes from the embryo in the mid-line and the space formed, and in which the blood-cells are found, is the epineural sinus. The sinus is continuous with the spaces between the somites when these open towards the mid-line of the embryo. In this way the development of the future haemococle is begun. The mesoderm, which originally was in the form of somites, gives rise to the muscles of the appendages, the muscles of the body, the gut musculature, the heart, and the fat-body (Text-fig. 16).

The mesoderm which migrates into the appendages becomes divided into two (Text-fig. 16). Both parts become attached to the ectoderm, and at their points of attachment the ectoderm becomes thickened. The more internal part becomes in the thoracic segments the intrinsic leg muscle, the outer part the extrinsic leg muscle. The mesoderm dorsal to the extrinsic leg muscle develops into the longitudinal sternal muscle. Mesoderm cells between the intrinsic leg muscle and the ventral nervecord form the ventral fat-body. The remainder of the mesoderm occupies at 11 days a position at the sides of the embryo on a level with the spiracle invaginations. That part of it which is closely applied to the ectoderm gives rise to dorso-lateral and dorso-longitudinal body muscles. On its inner edge this part is continuous with the dorsal fat-body. The small piece of mesoderm lying dorsal to the last-mentioned portion forms the heart and the gut muscles. This latter becomes covered on the outside with a layer of endoderm cells, which are backward extensions down either side of the body from the anterior endodermal mass surrounding the base of the stomodaeum (Text-figs. 16 and 21).

It is not until the yolk has been enclosed by upward growth of the body-wall, that differentiation of the mesoderm into muscle-fibres becomes obvious. Two kinds of somatic muscles are formed. In the appendages the myoblast cells forming muscles become arranged end to end, their nuclei remaining central. Each row of cells becomes a muscle-fibre with its nuclei placed centrally, and its contractile material arranged round the outside (Text-fig. 17 c). The rest of the somatic muscles are formed in a totally different way. The nuclei of the cells forming the fibre remain on the periphery, while the centre of the fibre is composed of contractile material. These muscle-fibres show a conspicuous sarcolema (Text-fig. 17 b). Muscles such as the cervical muscles, longitudinal tergal, and dorso-ventral muscles show this arrangement of their fibres. For further information on this subject in adult insects see Weber (1933).

A very thin layer of mesoderm comes to surround the fore-, mid-, and hind-gut. The cells are elongated and flat, and they become differentiated into the layer of muscle surrounding the alimentary canal.

The fat-body, which arises from the somites in dorsal and ventral portions (Text-fig. 16), becomes particularly well developed in the head and thorax, where it forms a packing tissue round all the organs. In the abdomen it is restricted to a thin layer of cells lying beneath the epidermis and showing a segmental arrangement. This layer is much interrupted by the abdominal muscles. The fat-body cells in the abdomen often have the typical ‘signet ring ‘appearance, while those of the thorax appear to have less cytoplasm.

The heart is developed from the most lateral part of the mesoderm, which also gives rise to the splanchnic muscles (Text-fig. 16). The cardioblasts separate from the splanchnic muscle mesoderm in 13-day embryos. They become attached to the ectoderm at the extreme edge of the body-wall (Text-fig. 18 (a)). As the body-wall grows up dorsally to enclose the yolk, the cardioblasts are carried with it until they meet in the mid-dorsal line. On meeting they form the tube-like heart and the dorsal diaphragm (Text-fig. 18 (b)).

The aorta comes into existence before the heart. It is formed above the stomodaeum from two small tubular portions of mesoderm which are presumably the antennary somites, though this cannot be ascertained with certainty. These meet and form one tube (Text-fig. 18 c). This takes place at 13 days, and the aorta later becomes continuous with the heart. The aorta ends anteriorly in a sinus which is continued forwards through the brain and backwards beneath the sympathetic ganglion.

The germ-cells first appear as a group of cells at the posterior pole of the egg at a time when the germ-band has become differentiated, but before its involution (Mellanby, H., 1935). With the invagination of the germ-band the germ-cells are carried forward through the yolk, and they come to lie at the extreme posterior end of the developing embryo. They appear to migrate through the lower-layer cells (mesoderm), and in 8-day-old embryos they are found round the small coeloms of abdominal segments 8-10 (Text-fig. 14). By 11 days they have moved more anteriorly, and form two continuous strands of cells running through abdominal segments 6-8. The strands are closely applied to the yolk side of the mesoderm (Text-fig. 15) and are lateral in position, at the same level as the developing spiracular invaginations. After evolution has occurred, and the yolk becomes enclosed by the growth of the lateral parts of the embryo, the germ-cells take up a dorso-lateral position. They are placed very close to the lateral body-wall of the embryo because the whole of the middle abdominal region of the embryo is filled with the enclosed yolk. In 19-day embryos the germcells have developed to form a definite organ. The main longitudinal strand has become divided up at its anterior end into about 8 follicles giving the appearance of a typical insect gonad (Text-fig. 19). In some embryos the length of the follicles was much greater than in others of the same age. Possibly the longer ones are ovaries, and the short ones testes. The gonads remain in approximately this condition until the insects become adult.

The head of Rhodnius appears to consist of an anterior non-segmental portion or acron, followed by six segments. In longitudinal sections of a 6-day embryo (Text-fig. 20) it will be seen that there are three segments in front of the mandibular segment. From the first segment the bifid labrum grows out. The second has the antennae as appendages. The third has the very short-lived premandibles for its appendages. In Text-fig. 20 the very small coelom of the premandibular segment is shown. These first three segments are followed by three more, those of the mandibular, first and second maxillary segments. The antennae are at first post-oral, and longitudinal sections show the stomodaeal invagination anterior to the antennae.

Later the antennae become pre-oral. The labrum of Rhodnius appears to arise as a very definitely paired structure (Text-fig. 1) and the mesoderm has an obvious coelom (Text-fig. 20). Later the mesoderm in transverse sections is paired, though at this stage, 11 days, there is no evidence of paired coelomic sacs. Eastham (1930) suggested from his observations on the bifid labrum of Pieris that the insect labrum may represent a pair of fused appendages. This idea receives support from the recent discovery of Roonwal (1935) that the labral segment in Locusta has a pair of prominent coelomic sacs. This would seem to show that the insect labrum does represent a pair of appendages. However, it must be remembered that the labrum appears to originate as a single structure in some insects. The evidence from the embryology of Rhodnius is in favour of the labrum being a pair of appendages.

From dissections of embryos aged 11-16 days, there appear to be three neuromeres involved in the supra-oesophageal nervemass, and from longitudinal sections of 11-day embryos three in the sub-oesophageal ganglion. This is evidence in support of the head consisting of six segments, and if ‘procephalic ‘be substituted for ‘labral segment’ then this agrees with the interpretation given by Goodrich (1897) for the segmentation of the insect head. The following table would seem to represent the plan of head segmentation in Rhodnius.

It will be observed that this table agrees with that drawn up by Eastham for Pieris (1930). It does not agree with the findings of Wiesmann (1926) on the stick insect, Carausius morosus. Wiesmann includes a pre-antennary as well as a labral segment. I can find no evidence of two segments anterior to the antennal one in Rhodnius.

The very specialized mouth parts of nymphal and adult bugs were shown to be developed from ordinary embryonic outgrowths by Heymons (1899) and Muir and Kershaw (1911). Until the embryo is 11 days old, the mouth appendages of Rhodnius remain as simple outgrowths except for the bisegmentation of the first maxillae.

After the embryo has begun the second stage in blastokinesis, the mouth parts begin to differentiate rapidly. The labrum becomes longer and more pointed at the distal end. The mandibles and the first maxillae become much longer and more like stylets. At the same time pocket-like invaginations are formed at their bases. Both mandibles and first maxillae of each side grow backwards into a common pocket, and then each has its own individual pocket leading out of the common one (Text-fig. 21 a and b). These pockets are invaginations which grow deep into the head, and eventually form a coiled pocket into which the mandibles and maxillae can be partially withdrawn. The second maxillae grow towards each other in the mid-ventral line. They become joined distally, and at the same time they segment into three portions. The joined portion of the second maxillae forms a groove which is open dorsally. The mandibles are just dorsal to this groove by the time the embryo is 13 days old. The maxillae remain outside for some little time longer, and they are very prominent structures in whole mounts.

With the exception of the first abdominal appendages, the abdominal appendages are all very evanescent structures. They appear progressively from before backwards as the abdomen segments. They begin to appear in 7-day embryos, and by 11 days all except the first pair have begun to be resorbed. The first pair form the pleuropodia already mentioned (see p. 13). In Rhodnius the pleuropodia are well developed and form a conspicuous gland-like structure (Text-fig. 8). Extensions from the gland-cells pass out into the amniotic cavity through the narrow orifice of the gland. They are still quite prominent structures at 18 days, but from then onwards their cells gradually break down and are absorbed.

In conclusion I should like to thank Professor Watson, in whose department this work was done, and also Professor Eastham for reading the typescript and giving me many helpful suggestions.

Eggs incubated at 21° C. and 90 per cent, relative humidity.

Age. Stage of Development.

5 days. Germ-layers completed. Head appendages just beginning to develop. Segmentation begun in head and thorax. Neuroblast differentiation.

6 days. Premandibular appendages visible. Somites distinct in segmented part. Coeloms developed in head segments. First appearance of stomodaeal invagination. Serosa thickened anteriorly. Neural groove obvious. Abdomen four to five segments. First and second abdominal appendages.

7 days. Stomodaeum farther advanced. Premandibular appendages resorbed. Abdomen six to seven segments and terminal unsegmented part. Proctodaeum invagination begun. Short abdominal appendages developed on segmented part of abdomen.

8 days. Head and thoracic appendages all much lengthened. Abdomen seven segments with appendages. Anterior and posterior endoderm rudiments attached to blind ends of stomodaeum and proctodaeum. Germ-cells round posterior abdominal coeloms. Break down of somites begun.

9 days. Abdominal appendages have reached maximum development. Abdomen consists of nine segments.

10 days. Embryo sunk in yolk and ventral part of embryo completely concealed by yolk. Abdomen ten-segmented. Ectodermal invagination to form tentorium and tracheal system just beginning. Nerve-fibres visible in brain and ventral nervecord. Malpighian tubules just appearing.

11 days. Ectodermal invaginations well advanced. Abdominal appendages other than the first resorbed. First abdominal appendages have begun to take on their glandular structure. First maxillae bi-segmented. Nervous system with six head ganglia, three thoracic, and eight abdominal. Formation of sympathetic nervous system. Somites broken down to form beginnings of muscles and fat-body. Epineural sinus present containing blood-cells. Endodermal layer completed on yolk side of embryo. Germ-cells in two rudiments in abdominal segments 6-8.

12 days. Embryo moves on to ventral surface of egg and head is once more anterior. Embryo broadens. Serosa contracted to anterior end. Pleuropodia secrete a third embryonic membrane. Mandibles and maxillae start to be invaginated into the head at their base. Corpora allata arise. Sub- oesophageal nerve-mass formed by fusion of the three neuromeres belonging to the mandibular, first and second maxillary segments.

13 days. Embryo nearly reaches anterior end of egg. Growth of embryo round yolk begun. Second maxillae grow together, forming labium. Cells forming muscles (myoblasts) begin to show a definite end to end arrangement. First appearance of oenocytes. Contraction of abdominal nerve ganglia towards thorax. Aorta developed.

14 days. Head of embryo reaches anterior end of egg. Amnion becomes retracted with serosa to dorsal thoracic region of embryo forming ‘dorsal organ’. Yolk continues to be enclosed by growth of embryo. Muscle development proceeds. Corpus allatum still paired.

15 days. Yolk enclosed by embryo except in region just behind ‘dorsal organ’. Tracheal tubes begun. Aorta completed.

16 days. Yolk completely enclosed. ‘Dorsal organ’ disintegrates. Sympathetic nervous system complete. Ventral nerve-cord further contracted. Stomodaeum much lengthened. Formation of heart.

17-18 days. Abdominal part of ventral nerve-cord fused with third thoracic ganglion. Yolk restricted to ‘yolk sac’. Posterior part of mid-gut tubular. Longitudinal body muscles differentiated. Ectoderm cells covered by freshly secreted cuticle.

19-21 days. Musculature differentiated all over body. Anterior part of mid-gut formed. Germ-cells arranged to form gonads. Cuticular layer over ectoderm complete. Third embryonic membrane has become very thin.

22-6 days. Mandibles and maxillae become much lengthened and styletlike. Decrease in size of metathoracic ganglion. Fat-body in thorax becomes well developed. Great lengthening of Malpighian tubules. Anterior part of mid-gut takes a tubular form. Disappearance of yolk nuclei. Tracheae well developed. Salivary gland well developed at sides of stomodaeum in thorax.

26-9 days. Cuticle becomes very ‘frilled’ in appearance. Outer layer becomes detached and this will be moulted during the embryonic moult. Anterior part of mid-gut touches blind end of stomodaeum, but there is no through communication. Quantity of yolk in yolk-sac considerably diminished.