Through the kindness of Professor Minchin and Mr. Pocock I have been able to examine embryos and larvæ of a South African species of Archispirostreptus from Port Elizabeth. The adult specimens were obtained by Mr. Pocock for the Zoological Society, and the eggs were laid in the insect house at the Society’s gardens.
The work was undertaken with a view to finding out, from embryological evidence, the number of segments and appendages concerned in the formation of the Diplopod gnathochilarium. Unfortunately none of the larvæ from the two batches of eggs that were laid lived long enough to allow the development of the gnathochilarium to be observed. The examination of some of the embryos has, however, led to one or two interesting results, and incidentally to the solution of the gnathochilarium problem itself.
Many thanks are due to Mr. Pocock, who spent much of his valuable time in preserving the earlier stages daily.
In an account of the developmental stages of several Chilognatha, Metschnikoff (1874) states that in them the mouth parts are only two in number ; namely, mandibles and one pair of maxillæ. Heathcote’s (1888) observations confirm this statement as also do those of vom Rath (1886), whose dissertation unfortunately I have been unable to see. While Heathcote merely mentions the fact of there being but one pair of maxillæ, Metschnikoff insists upon it as being important, and includes it among other points of resemblance between the Poduridæ and the Chilognatha.
More lately, zoologists have been quite awake to the significance of there being only two pairs of mouth parts in the Chilognatha, though but few observations have been made on the development of these myriapods. Folsom (1900) quotes Packard (1883) as saying that there can only be two pairs of oral appendages; but adds that he (Folsom) would suggest that further embryological studies on Diplopods might show more. In his summary he homologises the Diplopod gnatho-chilarium with three Hexapod somites, namely, those of the superlingnæ, maxillæ, and labium.
Korschelt and Heider (1898) also, in their text-book quote Metschnikoff and vom Rath, and allude to the necessity for further investigation into this matter.
Heymons (1897), in an account of the embryo of Glomeris, States very decidedly that its gnathochilarium is formed of only one pair of appendages and the hypopharynx (hypostome), but ventures no surmises as to the origin of the hypostome.
Silvestri (1898), in describing the larva of Pachyulus communis says that though there is only one pair of appendages in the’ gnathochilarium yet two sterna are also concerned in its formation ; namely, the sternum of the maxillary segment and that of the post-maxillary segment, and that it is this latter sternum which corresponds with the hypostome of Latzel. This implies the existence of two maxillary segments, though one of them is without appendages.
The fact that only two jaw-bearing segments were known in Diplopods was one of several reasons given by Professor’ Lankester (1903), for placing the Diplopods (Dignatha monoprosthomera) far a Way from the Chilopods, Hexapods, and Crustacea (Pantagnatlia triprosthomera), in his recent classification of the Arthropoda.
The latest writer on the subject, Carpenter (1905), has found what he interprets as two pairs of maxillæ in the gnathochilarium of Polyxenus (adult). He also expresses the confident anticipation that the median parts of the gnathochilarium will be found to belong to the post-maxillary segment. This would imply the existence of three maxillary segments in Diplopoda.
Professor Lankester, considering the head as monopros-thomerous, places the Diplopoda far away from the Chilopoda, Hexapoda, and Crustacea.
Heymons considers the differences between the Diplopod head and those of the other groups to be not radical, for he admits the existence of a tritocerebrum in the adult Diplopod (St. Rémy, 1890), and also because he would include the post-maxillary segment in the head, considering it as a rudimentary maxillary segment, thus giving the head six segments as in the other tracheates.
Carpenter too is unable to consider the Diplopod head as monoprosthomerous, partly because he believes in the existence of an ocular segment, and also because he is in hopes of a tritocerebral segment being demonstrated in an embryo millipede.
The eggs were fixed for the most part in Perenyi’s fluid, but for some corrosive sublimate and acetic acid were used. Perenyi is excellent for hardening the yolk, but the second method gives far better histological results.
I can fully endorse Folsom’s (1900) statement that the best and easiest way of observing the minute embryonic appendages is to study the embryos whole with simply their shells removed. When this has been done sections can be cut, and they are of great use in giving complementary and explanatory evidence. The whole embryos were stained with Delafield’s hæmatoxylin, the sections with Kleinenberg’s haematoxylin and orange.
Description of the Stages
My observations were made chiefly on the two embryonic stages which are described below.
Stage A (fig. 1)
At this stage one can see on the ventral surface of the blastoderm :
The mouth which has the appearance of a semicircular pore.
The clypeus lying in front of the mouth, and already showing division into two lobes.
A rudimentary brain consisting of three lobes on either side.
An incomplete nerve ring on which lies one pair of post-oral ganglia.
Two other pairs of post-oral ganglia.
A large pair of procephalic lobes.
Four pairs of rudimentary post-oral appendages.
Markings or grooves showing divisions between the segments in which the appendages lie.
The nerve-ring, with its thickenings, is here very minute. I have represented it (fig. 1) and them as being slightly raised above the rest of the blastoderm. This is done to emphasise their appearance.
Heymons (1901), in his account of the development of Scolopendra, describes the primitive brain as consisting of three lobes on either side. In his drawing, which, like mine, is somewhat diagrammatic, the lobes lie moré in series than do those shown here (Pl. 37, fig. 1). Nevertheless, there is a great similarity between the two brains.
It seems that we have in the Archispirostreptus embryo an archicerebrum, in Professor Lankester’s (1881) sense of the word (Heymons uses the words “archicerebrum “and “syncerebrum” with a different meaning), and one cannot help being reminded of the fact that in the Chætopod brain, also an archicerebrum, there are three areas (Pruvot, 1885, and Racovitza, 1896).
The first post-oral ganglion, which is here very small, consisting of a just visible expansion of the thickening, is that belonging to the first post-oral segment which bears the antenna. I can find no trace of a pre-antennary segment in this embryo.
The segment which lies immediately behind the imperfectly closed nerve-ring is apparently the tritocerebral segment (intercalary segment of Heymons and others). It has ho appendages, but shows distinct rudiments of a pair of ganglia, and is definitely cut off by grooves from the antennary segment in front and the mandibular segment behind.
The “intersegmental furrows” (Heymons) seem here to appear first in the median region of the blastoderm, while in Scolopendra, according to Heymons (1901) they make their first appearance laterally.
The rudimentary nervous system is here extremely small and difficult to make out. From the deep way in which it stains with Delafield’s hæmatoxylin, I believe it to be as yet merely a thickening of the ectoderm.
Behind the mouth, a very little way in front of the first “intersegmental furrow,” there is a very slight expansion of the thickening on both sides of the nerve-ring. This expansion I take to be the first indication of the antennary ganglion. The ring is not completely closed, but the two cords lie very near each other behind the imperfect closure. These cords Lave a rather wavy outline all through their length, but after much careful observation I was able to make out in each cord a dip or incurving, surrounded by a semicircular thickening, of the cord. This incurving lies about half-way between the imperfect closing of the nerve-ring and the second “intersegmental furrow”—i. e., in the middle of the second postoral segment. This appearance is repeated in the mandibular segment. These semicircular, thickened incurvings of the cords must be the rudimentary tritocerebral and mandibular ganglia.
One cannot help being struck by the well-marked doubleness of the nervous system shown here. In Scolopendra, according to Heymons, the median unpaired ectodermal thickening (which he calls the archicerebrum) is the part of the nervous system which appears first. Later on it forms a junction between the two halves of his syncerebrum, ultimately becoming fused with them. These two halves, therefore, are never separated from each other by non-nervous tissue. But Heathcote (1886), if I understand him rightly, records a double origin for the nervous system of Julus, and Metschnikofi (1874), with more clearness, speaks of the first rudiment of the brain in Strongy loso ma as consisting of two “Scheitelplatten.”
I do not believe that the tritocerebral segment has, up till now, been observed in a Diplopod embryo. It is, however, not surprising to find it, since St. Rémy (1890) has described a well-marked tritocerebrum in the brains of Julus and Glomeris (adult).
Before leaving the future head, a word or two must be said about the procephalic lobes. They resemble those found in embryonic insects, arachnids, etc., and those figured by Heymons (1897) in the embryo of Glomeris. Their size would seem to me to almost preclude the possibility of finding a separate præ-antennary segment here.
Behind the tritocerebral segment lies that of the mandibles, and following that two segments bearing maxillæ. I have not been able to demonstrate the ganglia belonging to these last two segments. They have probably not yet made their appearance.
The antenuæ and mandibles are slightly raised above the rest of the blastoderm, and each of these appendages shows, a slight depression in its centre. The anterior pair of maxillæ are less rounded and smaller than the second pair, and they also lie rather nearer to the middle line. They are most likely homologous with the maxillulæ or superlinguæ described by Hansen (1893) and Folsom (1900) in some of the Thysanura and Qrthoptera. They are also the homologues of the first maxillæ in Cbilopoda and the first maxillæ in Crustacea.
I am guided to the making of this homology mainly by Hansen’s (1893) observation that the structure of the first maxillæ in Machilis agrees with that of the second maxilla in the Eumalacostraca, though I am well aware that serial position is of more importance than structure in such a case as this.
Folsom’s (1900) account of the development of the maxillulæ or superlinguæ in Anurida is somewhat unsatisfactory, for he sa.ys that they do not make their appearance until some little time after the maxillæ have become evident. This statement may, however, be due to an error in observation, and the ultimate position of these small appendages is between the mandibles and the first maxillæ.
Another piece of evidence which, though structural, seems to me to count for a good deal is this :—My observations on the developing Archispirostreptus have enabled me to giveample confirmation to Heathcote’s (1888) statemeht as to the mesodermal origin of the salivary gland in Diplopoda. This gland, then, must be a coelomoduct, and is most probably homologous not only with the salivary gland in P eripatus, but also with maxillary gland (in the second maxilla) in Crustacea.
Hansen has recently (1903) found maxillulæ in Scolopendrella, and this gives another reason in favour of the above-mentioned homology ; for if the Symphyla be not true Diplopods they are, at any rate, very nearly related to them, as they are also to the Thysanura.
I would, therefore, arrange the mouth parts in the four groups thus :
In stage B the size of the first maxillæ as compared with that of the other appendages is much reduced, and no trace of them can be found in the mouth parts of older larvæ. My material was very scanty, and only three series of longitudinal sections through stage B were cut. In each of these series, however, one could see the first maxillæ apparently in process of fusion with the mandible. Perhaps it would be better to say that one could see it being absorbed by the mandible (fig. 4). And just as one could see the apparent absorption of this maxillula by the mandible, so also could one see the absorption of its ganglion by the mandibular ganglion (6g. 5). Provided that any fusion of ganglia or appendages takes place the above is the manner in which one would expect it to do so. For the tritocerebral rudiments undoubtedly join the deuterocerebrum which lies in front of them, and not the mandibular ganglion which lies behind them, and the general tendency of fusion in the Arthropoda is from behind forwards.
If, as I believe, these maxillulæ fuse with or become part of the mandibles, they can hardly be homologous with the maxillulæ found by Carpenter (1905) in the adult Polyxenus, for the inferences which he draws from his observations on the mouth parts of that Diplopod would lead one to suppose that the maxillulæ there had fused with the maxillæ during the growth of the gnathochilarium.
This matter certainly requires further investigation, though it seems, for reasons which are given below, extremely improbable that more than one pair of appendages are concerned in the formation of the gnathochilarium.
Stage B (fig. 2).
This is the stage which immediately precedes the breaking and final casting off of the egg-shell. In an external view of it there can be seen :
The three-lobed archicerebrum.
The antenuary ganglia, to which the tritocerebral rudiments are now joined, and which still lie behind the mouth.
The ganglia of the mandibles.
The ganglia of the maxillulæ, much reduced in comparative size.
The ganglia of the maxillary segment.
The ganglia of a post-maxillary segment which is without appendages.
The ganglia of the three segments bearing rudimentary legs.
There are also to be observed the mouth and the clypeus, as well as a pair of appendages for each pair of ganglia, excepting that in the post-maxillary segment.
In each of the three lobes of the archicerebrum and in every other ganglion there can now be seen a distinct depression, a little pit in fact, the opening of which lies in about the middle of the ventral surface of the ganglion. Heathcote (1888) mentions these depressions as occurring in the ganglia of the embryonic Julus, but states that they do not appear until the ganglia have left the ectoderm. He also says that he does not consider them to have anything to do with the “cerebral grooves.” These “cerebral grooves” are similar cavities which he found in the ganglia of the embryonic brain.
In my specimens I can find no difference between the depressions in the ganglia forming the archicerebrum and those in the ganglia of the cords ; and further, these depressions seem to me to appear before the ganglia have com-pletely left the ectoderm (see fig. 5), which represents a longitudinal section through part of one of the nerve cords in stage B). In this section there can also be seen the small tritocerebral rudiment as well as the rudimentary ganglion of the maxillulæ. This last is being absorbed by the mandibular ganglion.
The tritocerebral rudiment here shows no depression. It is simply a mass of nervous tissue, lying between the antennary ganglion and that of the mandible. The ganglion of the maxillula on the other hand, does show traces of a depression, although it is fusing with, or rather becoming part of, the mandibular ganglion.
It can be noticed that all the appendages (first maxillæ or maxillulæ excepted) have grown in size, as also has the clypeus. The mouth has moved a little further back, but still lies in front of the antenna and its ganglion. The first maxillæ appear very much flattened between the mandibles and the second maxillæ.
Behind the second maxillary segment there is a segment definitely marked off from those in front of and behind it, and showing distinctly a pair of ganglia. This is the postmaxillary segment. Behind it there follow three segments, each of which bears a pair of rudimentary legs. These appendages spring from the hindmost region of the segments in which they occur, so that they have the appearance of belonging to the segment behind their own. Heymons (1897) has observed a similar appearance in the embryo of Glomeris, and, what is far more important, he has noticed in the same embryo a post-maxillary segment which bears no appendages. While being firmly of the opinion that there go to the making of the gnathochilarium only one pair of appendages and the hypopharynx, he homologises the postmaxillary segment with the labium-bearing segment in Hexapods, and the second maxillary segments in Chilopods. The knowledge of the existence of a pair of maxillæ in front of the gnathochilarium compels one now to homologise these segments differently. (See Table given above.)
Silvestri (1898) found a post-maxillary, legless segment in the larva of Pachyulus communis. He expresses the opinion that the sternum of this (post-maxillary) segment corresponds with the hypostome of Latzel, and that therefore the gnathochilarium consists of the sternum and appendages of the maxillary segment together with the sternum of the labial (post-maxillary) segment.
Further, he says that the dorsal part of this labial (postmaxillary) segment does not join the head at all, but forms a neck joined to the first segment of the body. This last statement is fully confirmed by my observations in S pi rest reptus. My reasons for thinking that the sternum of this segment does not form the hypostome are given further on.
Unfortunately none of the larvæ hatched in the Zoological Gardens lived long enough to enable the actual formation of the gnathochilarium to be worked out. But that is now a matter of comparatively small importance since the real question at issue was not the formation of the lower lip, but the number of maxillary segments present in the Diplopoda.
When the larval Archispirostreptus leaves the egg the mandibles and the second pair of maxillæ are clearly laid down, as indeed can be seen from Pl. 37, fig. 2, which represents an embryo still within the egg-shell. The first maxillæ have disappeared, being by this time completely absorbed by the mandibles. But the gnathochilarium is far from being fully formed, and in this my observations do not agree with those of vom Rath (1891), who found that the young larvæ of the Julidæ and Glomeridæ left the egg with mouth parts like those of the adult.
As the larvæ grew in age and size the maxillæ gradually approached each other, and in that which lived longest they touched each other -, but as yet they were not fused together, nor could I see that they had fused with any other part of the head, or body (see text-figures 1—4).
There were no adult Archispirostreptus at my disposal, I have therefore made use of some adult Spirostreptus collected by Professor Minchin in Uganda.
In the adult Spirostreptus there is behind the hypostome a piece of chitin, which, though larger than the hypostome, resembles it in shape (text-fig. 5). The simplest way of accounting for this would be to consider it as representing the sternum of the post-maxillary segment. The tergum of this segment is exceptionally large, and seems to represent two terga, namely, that of the post-maxillary segment and that of the first leg-bearing segment. The segment following that which bears the third pair of legs (fifth body segment) is apodous in Spirostreptus.
Heathcote (1888) in his account of the development of Juins says that he believes the first body segment, i.e. the post-maxillary segment, to be the apodous one. He states that he believes it to be equivalent to the second maxillary segment in the Insecta, and also asserts that its ganglion fuses early with that of the segment in front of it. I can find no trace.of such a fusion in either the embryos or larvæ of Archispirostreptus. Moreover, I have cut longitudinal sections through the subœsophageal ganglion in several adult Spirostreptus, and have found it to consist of only two pairs of ganglia (Pl. 37, fig. 6), namely, those of the mandibles and maxillæ.
I am well aware that the evidence is incomplete and that what is really-needed is an.examination of larvæ intermediate in age between, the oldest which I have had and the adult. But such evidence as we have at present seems to show that only one pair of appendages (the second pair of maxillæ) are concerned in the making of the gnathochilarium, though the sternum of the segment which bears them most probably also takes part in it. The fact that the subœsophageal ganglion consists of only two pairs of ganglia seems to point strongly to this conclusion.
In taking out the subœsophageal ganglia from several adult Spirostreptus I naturally looked at their gnatho-chilaria. In two of these I found the hypostome to be distinctly double (text-fig. 6). This would seem to show that it had its origin rather in two appendages than in one sternum.
If, as above suggested, the gnathochilarium consists of but one pair of appendages we have another reason in favour of the homology proposed in the first part of this paper.
Of making many homologies for the segments in different arthropods there is no end, but most authorities are agreed in assigning six segments to the head in Crustaceans, Hexapods and Chilopods. Now the head of the adult Diplopod is very distinctly marked off from the body, being separated from it by a narrow neck, which forms, as it were, a joint. This is more apparent in the recently hatched larva before the chitin has developed than in the adult (fig. 3). Even in so young an embryo as that shown in fig. 2 the ganglia of the head have a different appearance from those of the body. The Diplopod head can now be shown to consist of six segments thus :
A segment which, together with the acron (Heymons), forms the procephalic lobes.
The antennary segment.
The tritocerebral segment, representing the second antennary segment of Crustacea, and tritocerebral rudiments in Hexapoda and Chilopoda.
First maxillary (rudimentary).
Second maxillary segment, the appendages of which are fused to form the gnathochilarium.
As the head of Spirostreptus is distinct from the body so also, notwithstanding the forward movement of the ganglia, is the subœsophageal ganglion distinct from that of the post-maxillary segment. I mention this as further evidence to show that the post-maxillary segment belongs to the body and not to the head.
Summary of Results
There are, in the embryo of Archispirostreptus, two segments, the possession of which would seem to give the Diplopoda a place in the Arthropod system nearer to the Chilopoda and Hexapoda than that which has of late been assigned to them.
These additional segments are :
A tritocerebral segment representing the tritocerebral rudiments found by Wheeler (1893) and others in Hexapoda, and by Heymons in Scolopendra, and also the tritocerebral segment in Crustacea.
A pair of maxillæ (rudimentary) lying in front of the pair which forms the gnathochilarium in the adult. These are most likely homologous with the first maxillæ in Chilopoda and Crustacea, and with the superlinguæ (Folsom) of Hexapoda.
With regard to the development of the gnathochilarium I have unfortunately not been able to add much to our previous knowledge. But the importance of this matter has now become comparatively small, since the existence of a first pair of maxillæ has been demonstrated.
The evidence that I have is incomplete, but it certainly goes to show that the gnathochilarium is part of the head, being formed by the second pair of maxillæ which are the only appendages in it ; and that the post-maxillary segment of Heymons and Silvestri takes no part in the formation of this gnathochilarium, but is purely a body segment.
REFERENCES TO LITERATURE
Ag. Antennary ganglion. Aul. Antenna. Arc. Arcliicerebrum. Cig. Clypeus. Ec. Ectoderm. M. Mouth. Md. Mandible. Mdg. Ganglion of mandibular segment. Mes. Mesoderm. MA. First maxilla. MAg. Ganglion of first maxilla segment. MA’. Second maxilla. MA’g. Ganglion of second maxilla segment. Pmx. Post-maxillary segment. Pmxg. Ganglion of postmaxillary segment. Pmxt. Post-maxillary tergum. Pre. Procephalic lobe. Isf. Intersegmental furrow. Trc. Trito-cerebral segment. Trcg. Tritocerebral ganglion.
FIG. 1.—Embryo of Archispirostreptus (sixteen days old) about four days before leaving the shell.
FIG. 2.—Embryo of Archispirostreptus one day before leaving the shell
FIG. 3.—Larva of Archispirostreptus five days after hatching.
FIG. 4.—Longitudinal section through appendages of embryo shown in fig. 2.
FIG. 5.—Longitudinal section through ganglia of embryo shown in fig. 2.
FIG. 6.—Longitudinal section through the suboesophageal and post-maxillary ganglia in the adult Spirostreptus.
Diagrammatic plan of the mouth-parts of the larva of Archispirostreptus five days after it leaves the egg-shell.
Mouth-parts of larva fifteen days after leaving the shell
Mouth-parts of larva twenty days after leaving the shell. Ant. Antenna. Cly. Clypeus. W. Mandible. Mx.” Second maxilla.
Gnathochilarium from adult Spirostreptus, with hypostome as usually present.
Gnathochilarium from adult Spirostreptus,showing double hypostome.
Gn. Gnathochilarium. Hy. Hypostome. St. Sternum of postmaxillary segment. Gnathocbilarium. By. Hypostome. St. Sternum of postmaxillary segment.