PSYOHE.

IMAGINAL DISCS IN INSECTS.

BY HENRY S. PRATT PH. D. HAVERFORD COLLEGE, PA.

[Annual address of the retiring president of the Cambridge Entomological Club, January, I897].

Among nsects which have a completemetamorphosis the organization of thelarva is in general similar to that of theadult or imago, but the larva alwayslacks certain organs which the imagopossesses. The larval beetle or butter-fly lacks wings, the larval bee hasneither wings nor legs which is also trueof the larval gnat, while the larval flyhas neither legs, wings nor head; andin each case the imago possesses theseorgans. As we ascend the scale ofdevelopment from the less highly to themore highly specialized insects havingcomplete metamorphosis, we find aconstantly increasing sum of differencesbetween the larval and the imaginalforms, and a correspondingly increas-ing number of organs which are pos-sessed by the imago and not by thelarva. This drawing apart is due, onthe one hand, to the higher specializationof the imago and its consequent furtherdeparture from the ancestral stage of itsontogeny represented or suggested bythe larva, but also largely to the retro-gressive development of the larva itself.In the highest insects, where the imagois a highly specialized animal capableof living only in a certain restrictedenvironment, the larva is perhaps as

highly specialized as is the imago:its environment is as sharply restrictedand its structure departs as far fromthe phyletic type or stage it representsas is the case with the imago.

In the coleoptera, to consider firstone of the less highly specialized groupsof" holometabolic insects, the environ-ments of the larva and of the imago areusually quite similar, or perhaps theyare exactly the same the organs of thetwo forms are correspondingly similar,and the transformations which must beaccomplished on the body of the larvato produce the imago are but slight.The imago differs from the larva prin-cipally in that it has acquired wings,elytra, compound eyes, and external re-productive organs, but all the larvalorgans with the exception of the mid-gut become imaginal ones without greatchange. The midgutin all hoiometabo-lic insects undergoes a complete trans-.formation during the metamorphosis,*

In the lepidoptera, to come to asomewhat more highly specializedgroup, the larval and imaginal environ-ments are apparently widely different

* Kowalevsky, A. Beitrige Kenntnis der nachem-bryonalen Entwicklung der Muscden. Zeit. f. wiss. Zool.45; 3, x887. p. 565

16 PS:/’CI.E. [February ,897.

from each other. But, after all, bothlarva and imago live on vegetable food,and more than this, in the higher mem-bers ofthe group at least, on very nearlythe same kind of vegetable food, theleaves and flowers of phanerogamousplant,s. The larva eats and lives amongthe green leaves, while the imago findsits nourishment in the flowers which aremodified leaves. So that we shall notbe surprised if we find that the transfor-mations which result in the productionof the butterfly from the caterpillar arenot as great as the apparently great dif-ference between the two forms mightsuggest. /ks is the case with thecoleoptera, the imago acquires wings,compound eyes, and external reproduc-tive organs, and all the larval organswith the exception of the midgut passdirectly into the imaginal organs, al-though some of them are highly modi-fied in the process;but there is nocomplete making over, no general his-tolysis.

In the hymenoptera the larva andinago live in general on the same kindof food. But the conditions of colon-ial and family life which prevail amongthe higher members of the order haveresulted in a marked retrogressive de-velopment on the part of the larva, sothat it is very different structurally, fromthe imago. Not having to find its ownfood, to protect itself, or to escape fromenemies, it has lost its extremities. Inthe nematoceran diptera other conditionshave produced similar results, and wealso find apodous larvae. The transfor-mations, now, which the pupae of these

insects must undergo to become imagin-es are much greater than is the case inthe coleoptera or lepidoptera. Not onlymust wings be acquired during themetamorphosis but legs as well, and thelarval organs require a much greatermodification before they can serve asimaginal organs. But yet no-completemaking over, no general histolysis *takes place.When we come to the brachyceran

diptera, the most highly specialized in-sects in my opinion, we find the great-est structural differences between thelarva and the imago to be met amonginsects. The environment of the larwis as a rule totally dif[erent fi’om that ofthe imago, and the larval structurecorrespondingly different fl’om the imag-inal. Thelarva, too, has undergone anextensive retrogressive development.In the case of the nematocera, as I havejust said, the larva is without legs, butin the brachycera the retrogression hasgone much further and the larva hasneither legs nor head. There are alsogreat internal differences. In the meta-morphosis, consequently, a very differ-ent animal must come out of the pupa-rium than went into it. The imagomust acquire not only external repro-ductive organs, compound eyes, wings,legs, and a head, none of which thelarva possessed, but also internal organsvery different fi’om those of the larva.A complete making over accompaniedby general histolysis is the result.

* Ganin, M. The post embryonic development of insects.(Russian) Warsaw x875. Reviewed in Zeit. f. Wiss. Zool.a8Bd. 877; P. 386.

February 8c7. 17

A dissection of an old larva or a youngpupa of either of the insect ordersabove mentioned would show that thesenew organs which the insect is to ac-quire during its metamorphosis arereally already present, not as fillyformed organs, however, but in the formof rudiments or anlagen. In the bodycavity of the caterpillar, for instance,buried beneath the dorsal meso- andmetathoracic integument are two pairsof small disc-like islands of cells. Theseremain unfunctional and inactive dur-ing the caterpillar’s lifetime althoughgrowing constantly, but during its meta-morphosis they develop into the two pairsof wings of the butterfly.* Similar cell-islands are present in the larval coleop-teron." The larval hymenopteron alsopossesses theme; while beneath itsventral thoracic integument are threeother pairs of cell-islands whose fate it isto furnish the imaginal legs. In thenematocera, also, ventral and dorsalpairs of cell-islands are present in thelarva, as the observations of Weismannon Corethra first showed. This classicinvestigation demonstrated the presenceof three dosal pairs of cell-islands aswell as three ventral pairs. They are sit-uated in the body cavity of the larva justbeneath the integument, a dorsal and aventral pair in the prothorax, destinedto form the pupal spiracles and the

*Swammerdam. Bibel der Natur. Leipzig 75, p. 3. Pan-crifius. Beitr. Kennt. d. Fliigelentw. b. d. Insecten.KSnigsberg, 884.

Ganin. 1. p. 387.

Weismann. Die Metamorphose Corethra plumicor-his. Zeit. f. wiss. Zool. x6 Bd. x866. p. 45.

imaginal prothoracic legs, respectively;a dorsal and a ventral pair in the meso-thorax, destined to tbrm the wings andthe mesothoracic legs; and a dorsal anda ventral pair in the metathorax, des-tined to form the balancers and themetathoracic legs.

In the brachycera Weismann wasagain the first to prove the existence ofthese cell-islands. It was in I864 thathe published in his account of the post-embryonic development of the muscidsthe first correct and extended observa-tions on these peculiar cell-islands in anyinsect. He called them imaginaldiscs.He showed that in Musca six pairs arepresent in the larval thorax, not nearthe surface as in Corethra, but in the cen-ter of the larva, and that their fate isexactly the same as in Corethra. Inaddition to these thoracic imaginal discshe described two large cephalic discssituated in the forward portion of thethorax and connected with the larvalpharynx, the fate of which is to formthe imaginal head. Weismann alsoshowed that but a small portion of thelarval body passes directly into the im-aginal body, but that most of it under-goes disintegration so that the differenttissues entirely lose their identity, afterwhich the imaginal body is built up fl’omthe imaginal discs. To this process, theentire significance of which was not,however, understood until later, he gavethe name histolysis.

Wesmann, A. Die nachcmbryonale Entwick. d. Musci-den nach Beob. Musca vom. Sarcaphoga carnaria.Zeit. f. wiss. Zool. x4 Bd. Heft 3. 854. p.

1. p. 322 and 325.

18 PST’CIq.E. [February x897.

These observations of Weismannopened a new field .of investigation inthe development of insects. This hasbeen entered by several eminent inves-tigators ofwhom Ganin, Ktlnckel d’Her-culais, Viallanes, Kowalevsky, and VanRees have been the most successful.Ganin * in 1876 demonstrated the pres-ence of many other imaginal discs thanthose which Weismann described.Weismann supposed that the hypoder-mis of the larval abdomen went di-rectly with some modifications to formthat of the imago. Ganin showed, now,that in the hypodermis of ech of theeight abdominal segments are four cell-islands, two dorsal atd two ventral,which resemble in every respect thetissue of the imaginal discs that theyare in fact imaginal discs and are des-tined to form the starting points for thegrowth of the imaginal abdominal hyp-odermis. Ganin also discovered similardiscs in the epithelium of the midgutwhose fate it is to form, in the sameway, the imaginal midgut, and alsothe important fact that each imaginaldisc in the larva is made up of twokinds of embryonic tissue, ectodermand mesodenn or mesenchyrn. InI875 Kiinckel d’Herculais found inthe last abdominal segment two pairsof imaginal discs of the external geni-tal organs. In 883 Meschnikoff ++

* 1. p. 386.Ktinckel d’Herculais. Recherches l’organisation et

le d6veloppement des olucelles. Paris, x875, p. x49.

Metschnikoff. Untersuchung tiber intraccellulire Ver-dauung bei Wirbellosentieren. Arb. d. Zool. Inst.Wien. Bd. x883, p. 4.

published tile first of his epoch-mak-ing observations on the destruction ,oftissues in certain invertebrates by leu-cocytes or as he called them phagocytes.He discussed Ganin’s paper and espe-cially his statement that during thehistolysis of the pupal muscid the larvalorgans are destroyed by amoeboid mes-oderm cells. These cells he suggestsare none other than phagocytes. In884 Van Rees and in 1885 Kowalev-sky proved the correctness of this posi-tion; they showed that the process ofhistolysis is the tearing down and diges-tion of the functional larval tissues byphagocytes and the building up of imag-inal tissues fi’om imaginal discs.

In 1888 Van Rees published his ex-tensive paper on the post-embryonic de-velopment of muscids, and completedour knowledge of this phenomenon.He showed that when tile muscidianlarva has entered upon the pupal stage,histolysis is inaugurated by the destruc-tion of the larval muscles, they becom-ing unfunctional directly after pupationand a natural prey to the phagocytes.Soon the thoracic hypodermis and theinner organs are attacked, and at thesame time the imaginal discs begin togrow and widen out to supply the placeof tile tissues which are being destroyed.The cont.inuity of the hypodermis and

Van Rees. Over intra-cellulaire spijsverteering overde beteckenisder witte bloedllchampjes. MaandbladNaturwetenschappen. Jaarg. x, x884.

Kowalevsky. Beitrtge nachembryonale Entwieklungd. Musciden. Zool. Anz. Bd. 885, p. 98.

Van Rees. Beitrtge Kenntniss d. inneren M.eta-morphose Musca vomitoria. Zool. Jahrb. Abth. f. Anat.

Ontog, 3. Bd. x888, p.

February 897.] -SYC/.E. 19

of most of the internal organs is thusnot at any time broken. As these pro-cesses go on, the two large head imag-inal di,scs, which form two irregularlyshaped sacks extending as diverticulafrom the dorsal wall of the pharynxback to the brain, begin to pass forwarddragging the brain with them. Theiranterior ends bend and pass ventradembracing the pharynx between them.At the same time their communicationswith the pharynx enlarge and theirlumena fuse more completely with thepharyngial lumen. These communica-tlons now continue to enlarge theyfuse into one single median openingwhich, ever increasing in ize, travelsfrom the anterior end of the pharynxposteriad, obliterating the dorsal phar-yngial wall in its oourse. Finally thelumena of the discs and that of’ thepharynx become completely mergedand form together a single continuousspace, and the walls of the discs and ofthe pharynx form a single continuousvesicle. This is the head-vesicle ofWeismann * and Van Rees "f wDich isdestined to become the imaginal head.On its anterior ventral surface epithelialthickenings appear which are destined toform the imaginal antennae and mouthparts, while at its posterior end arethickenings which are to become thecompound eyes and which are still incontact with the brain. The head-vesicle remains buried within the pupalthorax until near the end of the pupal

period when it evaginates and forms theimaginal head. This evagination hasbeen observed by Weismann $ to be theconsequence of the pressure of bloodwhich at the right moment rushes fromthe abdomen into the thorax and pushesthe head-vesicle forward.The metamorphosis ofthe thorax goes

on simultaneously with the formation ofthe head-vesicle. In proportion as thelarval hypodermis disappears under theattacks of the phagocytes, as I havealready mentioned, the edges of theimaginal discs grow and take its place,forming the imaginal hypodermis._As we have seen, there are six pairs ofthese discs, three dorsal and threeventral, nd they are in the center ofthe larva. Each disc is, however, con-nected with that portion of the hypo-dermis of the segment to which itgenetically belongs and where it is des-tined to appear as an extremity, by avery fine, hollow chord. This chord,now, begins to shorten and its lumen toenlarge. The disc is thus broughtnearer the surface and, as it advances, itincreases in size. The lumen of thechord then opens through the hypoder-mis to the outside, and finally becomesso wide and the chord itself so shortthat the disc is brought through thehypodermis to the outside. The hol-low chord has of course been obliteratedby this process and the edges of theproximal end of the disc brought intodirect connection with the hypodermis.The disc has by this time assumed its

$1. p. 259.

ADS’C.H.’. [February 897.

position as an extremity. It is an ap-pendage ofthe body wall it has becomeirlegularly cylindrical in shape and pos-sesses a number of constrictions andfolds, which in the case of the ventraldiscs are equivalent to the joints of thefuture leg. The proximal edges of thediscs, those in contact with the larvalhypodermis, grow and extend them-selves and take the place of larval hy-podermis in proportion as this is de-stroyed by phagocytes.The metamorphosis of the abdomen

is retarded and does not begin until thatof the head and thorax is well advanced.Then in each abdominal segment thetwo ventral and four dorsal discs (VanRees found two additional dorsal discsin each segment) begin to grow and takethe place of the disappearing larval hy-podermis.Kowalevsky * discovered that the discs

of the last segment do not take part inthe formation of hypodermis, but of theendgut with the rectal glands, and thatthey are situated in the vicinity of thelarval anus. The metamorphosis of thelarval internal organs was correctly re-ported first by Kowalevsky " in the yearpreceding the publication of Van Rees’paper. All of these organs are de-stroyed by phagocytes except thecentral nervous system, the heart, thereproductive organs, and three pairs ofthoracic muscles. These with the excep-tion of the reproductive organs remain

Kowalevsky. Beitrtge Kenntniss d. nachembry-onalen Entwicklung d. Musciden. Theil. Zeit. f. Wiss.Zool. 45 Bd. 887, p. 557.

1. p. 542.

active, functional organs during thepupal period and are not attacked byphagocytes, but pass directly into theimago without great change and be-come imaginal organs. The organsdestroyed are reconstructed from imag-inal discs in a way similar to thatalready described.The only paper dealing with imagi-

nal discs which has appeared sinceVan Rees’ is one of my own publishedin 893.$ It contains a description ofthe larva of Melophagus ovinus, a pupi-parous dipter.The pupipars are cyclorrhaphic bra-

chycera and very closely allied to themuscids, so that we may expect to findthe same imaginal discs in their larvaeas in the muscids. _And we do, in fact,find in general similar conditions, butthere are several interesting differences.The larva is apodous and acephalouslike the muscidian, but in many respectsit is much less highly specialized;it seems, in fitct, as if it mightrepresent the ancestral stage in dip-teran phylogeny at which the muscidsare beginning to draw away fromtheir relatives, to occupy a position be-tween Corethra and Musca. In theposition of the thoracic discs, for in-stance, it closely resembles Corethra.We find these discs just beneath theintegument in two very regular rowsand not in the center of the larva as inMusca. The accompanying wood cutrepresents dorsal and ventral frontal

Pratt. Beitrlige Kenntnis d. Pupiparen. Die Larveyon Melophagus ovinus. Archiv f. Naturgeschichte. Bd.

59, 1893, p. 151.

February I897. PSYCHE. 21

sections through the anterior end of anold larva showing the position of thethoracic discs. The dorsal prothoracic

Fig. A, ventral frontal section; /, dorsal frontaltion; a, anterior-pole; -Pr, W[es. and 2t/et, pro-, meso-, andmetathoracic imaginal discs b, body wall.

pair arise during the larval period andare not embryonic organs as are all theothers (Van Rees noticed the samefact about the corresponding muscidiandiscs). They are invaginations of thelarval hypodermis and have externalopenings which do not close, and theirinner surface is lined with .a delicatecuticula which is a continuation of theexternal cuticula of the larva. Thesediscs are rudimentary organs and donot develop into any pupal or imaginalorgans. In structure the meso-andmetathoracic discs stand exactly mid-way between the same discs in Corethraand in Mu’sea. In Corethra* all thethoracic discs are of larval origin, aris-

Fig. D, imaginal disc Fig. 3. D, imaginal discin Corethra; /-/,.hypoder- in Melophagus; H, hy-mis C, cuticula, podermis C, cuticula P,

parapodial membrane.

* Weismann, Corethra. 1. p. 78.

ing as they do after the last larval moult,and each one is a double fold of thehypodermis, of which it remains a partas Fig. z shows. In Melophagus, onthe other hand, each of these discsarises in the embryo, as is also the casein Musca it is also a double fold ofthehypodermis but becomes constricted 0fffrom it as is shown in Fig. 3. VanRees’? has called that portion of theinvaginated hypodermis which enclosesthe disc proper, the parapodial mem-brane (P. Figs. 3 and 4), and the spaceit encloses, the parapodial space. InMuscat, now, the disc not only becomesconstricted off fl’om the hypodermis butsuffers removal to the center of the larvaas is shown diagramatically in Fig.4, andthe parapodial membrane lengthens toform the hollow chord which connects it

Fig. 4. D, imaginal disc in Muca .r/, hypodermis C,cuticula P, parapodial membrane; CI,, chord.

with its old position at tlxe hypodermis.The fate of the thoracic discs in Melo-phagus is exactly the same as in Corethraor in Musca.

In the cephalic discs, ow, we ,findthe conditions similar to those in M,usca,but even more complicated. Instead ofa single pair of head-discs we find twopairs, one dorsal and one ventral. Thedorsal pair corresponds to the muscidiaa

"tl. p. 24.,Van Rees. 1. p. 23.

lSY"CttE [February x897.

head-discs in every respect; they aredestined to form the dorsal and lateralportions of the imaginal head togetherwith the compound eyes. The ventralhead-discs have no counterpart inMusca. In the embryo, as we shall see ina moment, they appear as a single medianthickening, but in the young larwt theyhave become paired dive,’ticula of theventral pharyngial wall; and at the bot-tom of each diverticulum there arises aprojection. These diverticula soon fusein the median line and the wall betweenthem gradually disappears. In the oldlarva, consequently, the ventral discsappear as a single ventral dive,’ticulumof the pharynx at the bottom of whicha pair of long projections extend towardhe wide opening. The fate of thesediscs is to form the ventral portion ofthe head, the paired projections formingthe anlage of the proboscis.The formation of the head-vesicle pro-

ceeds in a way similar to that in Musca.The ventral disc fuses early at its lateraledges with the dorsal pair. The com-municatiots between both ventral anddorsal discs and the pharynx rapidlywiden (in the old larva they have alreadybecome very large), and soon thediscs and pharynx form together asingle vesicle, which is the head-vesicle.The imaginal discs of the abdomen

in Melophagus are extctly the same asthe corresponding otes in Musca.

I shall now speak of the origin andearly development of imaginal discs.But little has been published on thissubject and nothing, so far as I know,on the origin of muscidian discs.

Weismann* showed in I866 that inCorethra, that is, in the nematocera, thethoracic discs arise as ectodermic infold-ings after the last larval moult, and justbefore pupation (Fig. 2). This israther surprising, as we should expectthese discs, as they represent extremi-ties, to appear in an early embryonicperiod when the extremities usuallyarise in insects. In Musca, Weismann"did in fact find that the imaginal discsarise in the embryo, but their genesiswas not observed by him, and, as hefound them invariably attached to a tra-chea or a nerve, he made the mistake ofsupposing that they take their origin inthe epithelial coverings of these organs.This method of growth was, however,early discredited by Ktinckel d’Hercu-lais: who, in 875 found the chord con-necting the thoracic discs with the hypo-dermis, and rightly concluded that theyhave an ectodermic origin. Balfouralso, in his text book, declared that thethoracic and cephalic discs must be de-rivatives of the ectoderm in Musca asthey had been proved to bd n Corethra.Dewitz[[ in 878 confirmed d’Herculais’observation of the chord connecting thedisc with the hypodermis, and VanRees some years later in his paper al-ready quoted showed that this chord is

Weismann. Die Metamorphose Corethra etc. p. 78.Weismann. Die nachembry. Entw. v.d. Musciden, etc.,

p. 23.

:l:l. p. i46.Balfour: Comparative Embryology, vol. I, p. 423IIDewitz. Beitrg.ge Kenntnis d. postembry. Glied-

b. d. Insecten. Zeit. f, wlss. Zool. Bd. 30.Sup. p. 95.1. p. 25

February 897. PSYCI-t.E. 28

hollow, that its lumen is a continuationofthe parapodial space, as shown in Fig.4, and that both lumen and parapodialcavity are lined by a fine cuticula. Heasserts that this discovery is anatomi-cal proof that the chord, the parapodialmembrane, and the disc itself are all partsof a single invagination of the embryonicectoderm. Graber,* however, in x889,supported by certain observations onCalliphora, a muscid, asserts that thebeginnings of both the thoracic andcephalic imaginal discs are not ec-todermic invaginations, but ectodermicthickenings towards the inside, fol-lowed by delaminations.There exists in fact at the present

time no embryological evidence on theorigin and first stages of development.of the thoracic and cephalic imaginaldiscs in the brachyceran diptera, al-though all the later writers exceptGraber, supported by the fact of theirundoubted ectodermic origin in thelarva of Corethra, and also by the ana-tomical evidence adduced by Van Rees,"are of the opinion that they arise as in-vaginations of the embryonic ectoderm.

I have been for some time studyingthe embryonic development of Melo-phagus ovinus, and, although the study isnot yet completed, I am able to give adetailed account of the origin and earlyhistory of the imaginal discs in this in-sect. Melophagus is a cyclorrhaphicbrachyceran. Leuckart? early showed

Graber. Vergl. Stud. fiber, d. Embry. d. Insecten.Denks. Acad. Wiss. Wien. Bd. 56. p. 3o6.

Leuckart. Die Fortpflanzung Entwick. d. Pupiparen,1858.

the striking similarity of its larvaland pupal forms, and of the pupiparsin general, to those of the muscids; inan earlier portion of thispaper, I haveemphasized the same fact Mtlggenburg;thas described the mouthparts of almostall the pupipars and homologisedthem with the parts ofthe fly’s proboscisBrauer considers the pupipars de-generate flies. The relationship betweenthe pupipars and the muscids isundoubtedly a very close one, and obser-vations on the former must be conclusivein determining the method of origin ofimaginal discs in the latter, as well as inall the higher dipters.

In Melophagus the cephalic and thor-acic imaginal discs first appear as localthickenings followed by invagination ofthe embryonic ectoderm. The cephalicdiscs make their appearance first, andvery early in the ontogeny of the insect.The ventral plate, as in other dipters,is not confined, to.the ventral side ofthe embryo, but it encircles the anteriorend and covers the anterior third ofthe dorsal side, as is shown in Fig.. 5"The stomatodeum appears as an ecto-dermic depression on the dorsal sideat some distance from the anterior end(S, Fig. 5), the proctodeum a.s a cor-responding depression on the dorsalside, slightly distant from the poste-rior end (P, Fig. 5)" These are not,however, the permanent positions of

Mfiggenburg. Der Riissel d. Diptera pupipara.Archiv f. Naturgeschicite. Jahrg. 58, I892 p. 9-87.

Brauer. Systemat.-zool. Studien. 9x Bd. d. Sitzungsbr.d. Kais. Acad, d. Wiss. ,st Abt. x885, p. 405.

t’SYCHE. [February x87.

these organs. They gradually migratetowards the two poles of the egg, andfinally attain positions slightly ventral

Fig. 5. Embryo Melophagus, side view a, anterior end//, ventral side; S, stomatodeum; P, proctodeum; //P,ventral plate T, tracheal invaginations H, head,fold.

to the anterior and posterior polesrespectively. Long before the stom-atodeum has attained its definitiveposition, however, there appear beforeand behind it crescentic, ectodermalthickenings which partly encircle it asshown in Fig. 6. These are the begin-

Fig. 6. Anterior end of embryo, dorsal view; H, head-fold; S, stomatodeum; //t9, ventral disc; DD, dorsaldiscs.

nings of the cephalic imaginal discs,which, it will be seen, appear veryearly in the ontogeny of the insect,

while, in fact, the intestinal tract, thetracheal, and nervous systems are intheir firstanlagen. The crescentic thick-enings, now, are three in number, apair just behind (D D, Fig. 6) and asingle median one just before(VD, Fig.6) the stomatodeum. The latter thick-ening is destined to form the ventralcephalic disc; the fate of the pair, onthe other hand, is to form the dorsaldiscs, those homologous with the ceph-alic discs of Musca. At the posterior sideofeach ofthe latter thickenings, now, ainvagination begins to appear whiclafinally becomes a deep pocket. Eachpocket early shows an intimate connec-tion with a supra-oesophageal ganglion,’the latter abutting the posterior sideof the pocket, but having no structuralunion with it. During the formationof these pockets, however, the stoma-todeum, together with the dorsal andventral discs, has been continuing itsmigration towards the anterior end ofthe embryo, and the fornaation of the"intestine and closure of the back ofthe embryo have been going rapidlyforward. Finally, when the mouth hasarrived at its definitive position at theanterior pole .of the .embryo, the dorsalpockets have come to occupy a positionon the dorsal side of the embryo justabove it, and the back is entirely closed.The openings of the pockets have by

this time moved to the mid-dorsal .lineand merged into a single median open-ing. In Fig. 7, DD represents thiscommon opening. The proximal orupper portions of the pockets have alsoflsed, but their distal ends are still fi’ee

February I897.] P T"CI-I.E 25

from each other and each still abutsa supra-oesophageal ganglion. /ks tothe median ectodermic thickening inFig. 6 (V D) which is destined to form

.VD IDD

Fig. 7. SagittM section of anterior end of embryo iustbefore involution of head; mouth;f, intestine;ventrM disc; 2)D, dorsal discs; fs, muscle proiecfionU, ventral side , dorsal side.the ventral cephalic disc, it migrates for-ward with the stomatodeum, suffering atthe same time a slight invagination,and finally takes a position immediate-ly below the mouth (V D, Fig. 7).The relation of the other organs ofthe anterior portion of the embryo tothe discs, is well shown in Fig. 7"Just dorsad of the mouth and betweenit and the median portion of the dorsaldiscs is seen a muscular projection(Mus.) whose later history will be foundto be interesting.A very important change now takes

place in the development ofthe animal,namely, the involution .of the embry-onic head. /kn ectodermic fold startsjust posteriad of the discs, both dorsaland ventral, and grows rapidly forwardtowards and over the mouth. Themouth, with the ventral disc just ventradof it and the muscular projection anddorsal discs just dorsad of it, is rolled in

by this process. A new mouth is formed(M, Fig. 8) and just back of itis a new portion of the intestinal tract(P, Fig. 8). This is the so-called phar-ynx of Weismann and Van Rees, de-scribed by them in the larva of Musca.Just back of the pharynx is the oesopha-gus and the old mouth, dorsad ofwhichis the muscular projection (Mus. Fig.

Fig. 8. Sagittal section of embryo just after involution ofhead 2, mouth _P. pharynx DD, dorsal discs V.D,ventral disc I, intestine 3Ius, muscle projection V,tral side; D, dorsal side.

8). This projection becomes in thelarva the most active organ in the ani-mal. It is a. sucking tongue and by itsregular pulse-like contractions causes aflow of the milk-like secretion, whichis present in the uterus and formsthe food of the larva, into the mouth.*Connected with the dorsal wall of thepharynx is the fused median portionof the dorsal discs (DD, Fig. 8) whichin Fig. 7 opened to the outside of theembryo. The distal portions ofthe discsrepresent diverticula which extend backto the supra-oesophageal ganglia. Onthe ventral side of the pharynx is theventral disc (VD, Fig. 8) which showsa slight invagination and Which is des-tined to undergo considerable changes

Pratt. 1. p. I70.

.PST’CHE. February I897.]

before it reaches its final form.* Theforward end of the animal does notchange materially, now, from the con-dition we find represented in Fig.. 8during the remainder of the embryonicand the entire larval period. Early inthe pupal period, however, the dorsaland ventral discs unite to form the headvesicle, as explained in a previousportion of this paper, which becomesthe imaginal head.So much for the origin and embryonic

history of the cephalic imaginal discs.The thoracic discs (with the.exception ofthe dorsal prothoracic which arise inthe larva) do not make their appearanceuntil the time of the involution of thehead. This is late in the embryonic lifeand at. a period when its organizationand the formation of the head discs arecompleted. The thoracic discs arise,like the cephalic, as ectodermic thicken-ings. A dorso-ventral section to oneside of the median line of the same em-bryo of which Fig. 8 represents a sagit-tal section, is shown in Fig. 9" The

Fig. 9. Dorso-ventral section of anterior end of embryoPro, Wges, and /P/el, ventral pro-, meso-, and metathoracicdiscs; 21/Ies. D, dorsal mesothoracic disc; /4", heac-disc; I,ntestine D, dorsal side V, ventral side.

Pratt. 1. p. I88.

three ventral ectodermal thickeningsare the beginnings of the ventral pro-meso-and metathoracic imaglnal discson one side of the embryo. The dorsalthickening is the beginning of the dorsalmesothoracic disc. The dorsal metatho-racic disc does not appear in this section.All these, discs begin very soon to invag-inate, the ventral prothoracic begin-ning first, then the mesothoracic, andfinally the metathoracic taking its turn.The invagination begins in each disc atits posterior border as is shown in theventral metathoracic disc i,n Fig. O;then the anterior border sinks in as isshown in the mesothoracic disc in Fig.Io finally the entire disc sinks beneaththe surface as is shown in the prothor-acic disc in Fig. o; then the ecto-

Fig. Pro, Mes, and Met, Fig. Disc with invag-ventral pro-, meso-, and meta- ination completed.thoracic discs, showing begin-ning of invagination.

derm closes over it as is shown inFig. II, at which stage it is com-parable to the thoracic discs in Corethraas represented in Fig. 2. Thedisc now becomes separated fromthe ectoderm, the thickened middleportion sinks in and forms the disc

February 897. PSUCItE. 27

proper, the real anlage of the extremity,while the portion at the side becomesthe parapodial membrane, and the discis formed as we find it throughout thelarval period (Fig. 3).As to the discs of tlae internal organs

and of the abdominal hypodermis, Ihave not observed them in the embryoand think it is probable they appear firstin the larva.

Before closing I wish to dwell for amoment on one or two theoretical ques-tions which naturally present them-selves. In the first place, why is it thatin the brachyceran dipters the phago-cytes, that is the blood corpuscles,during the metamorphosis do not attackall the larval tissues indescriminatelyinstead of being selective in their oper-ations. We find that in the histolysiscertain organs only, such as the larvalhypodermis, intestine, muscles, etc., areattacked and destroyed while others,such as the imaginal discs, the heart,central nervous system, reproductiveglands, and even some of the muscles areleft intact. Kowalevsk)* seeks to an-swer this question. He says the reason isthe same as that which accounts for thefact, first observed by Metschnikoff,-that certain virulent bcteria, as theform vhich is the cause of Anthrax, arenot attacked by leucocytes, while thesame form in Pasteur’s vaccine for An-thrax, which las been weakened androbbed to a certain extent of its virulentpower, is attacked and consumed. With

1. c. p. 555.

Metschriikoff. Archiv f. pathol. Anat. Bd. 97, P. 52.

the beginning of the pupal stage inMusca the larval tissues, being no longe:active, functional tissues, become weakand degenerate and are attacked and de-stroyed by the amoeboid blood cor-puscles of the pupa, as are the non-viru-lent bacteria of the vaccine by the leu-cocytes. The imaginal discs, on theother hand, which are a fresh, youngtissue, wlaose cells are rapidly proliferat-ing, are immune, as are also the viru-lent bacteria. Kowalevsky also saysthat perhaps the tissue of the imaginaldiscs also secretes some substance whichrenders it secure against the attacks ofphagoc)tes as do the virulent Bacteria.Van Rees says that the phagocytes do,as a matter of fact, attack all the larvalorgans indiscriminately, but that theactive metabolism of the cells of theimaglnal discs preserves them fromthese attacks. He thinksKowalevsky isprobably right in supposing that thediscs render themselves immune bysome poisonous secretion.

It seems to me that the supposition ofa protecting secretion is hardly neces-sary to account for the phenomenon,and it certainly would not account forthe preservation of those functional lar-val tissues which are not destroyed.As soon as the larval life is over, in thoseorgans and tissues which then becomefunctionless and inactive, a healthymetabolism would cease, and they wouldbegin to exhibit signs of weakness anddegeneracy. ]n this condition theyw.ould be an easy prey to phagocytes ordisintegrating influences of any sort.

.$1. p.

28 [February x897.

Weak or functionless organs in any ani-mal are peculiarly susceptable to dis-ease. Healthy organs may be exposedto the same influences without danger.In the same way the imaginal discs, inwhich there is an exceedingly activemetabolism, and all the larval organswhich remain functional during themetamorphosis or during a greater partof it are immune from the attacks of thephagocytes. The heart in tle muscidscontinues to beat, as Kttnckel d’Hercu-lais has observed, during the entire per-iod of the metamorphosis with the ex-ception of a day or two in the latterhalf of it. The nervous system mustcontinue functional during the entiretime. The three pairs of thoracicmuscles which pass intact from the lar-va to the imago are probably employedin respiration during the metamorphosis.The reproductive glands are, like theimaginal discs, rapidly growing organs.It would be interesting to know exactlyto what extent the blood corpuscles inthe other orders of insects besides thebrachyceran dipters become phagocytesduring the metamorphosis, and whyhistolysis is so much more extensivehere than elsewhere. The answerwould undoubtedly be in accord withwhat I have just said above. The bra-chycera are the most highly specializedinsects; the structural differences be-tween the larva and imago are the mostprofound; their metamorphosis is thethe most complete; but two or threeof the functional, larval organs remainfinctional during the metamorphosisamoeboid blood corpuscles in tlae form

of phagocytes consume the rest, sparing,however, also the imaginal discs fromwhich the imaginal organs are built up.Among the other holometabolic insects,on the other hand, many or most of thelarval organs remain functional duringthe metamorphosis, hence there is butlittle histolysis. But the larval intes-tine wotld always necessarily becomeunfunctional, and, as we have seen,Kowalevsky is of the opinion that thelarval midgut in all holometabolic insectscontains imaginal discs and undergoesdegeneration during the metamorphosis.

It is interesting to note that each ofthe three thoracic and eight abdomin.alsegments which make up the larvalbody in the brachyceran dipters hastwo pairs of imaginal discs, a dorsaland a ventral pair. Thus there is adouble rov of discs extending the lengthof the body on the dorsal side and an-other on the ventral side of the larva.This is seen especially well in Melo-phagus where the thoracic discs are nearthe surface and not in the center of thelarva as they are in Musca the thoracicdiscs are here in straight rows with theabdominal ones. The question nowpresents itself,are the thoracic andabdominal discs homodynamous organs.The different discs furnish very differ-ent imaginal organs; legs, wings, pupalspiracles, balancers, and hypodermisgrow fl’om the thoracic, and hypoder-mis, rectal glands, and perhaps externalsexual organs from the abdominal discsbut yet they are all similar to each otherin several very important respects.In the first place the method of ori-

February I897.] PSYCHE. 29

gin is exactly the same in all. Theyall arise as ectodermal thickeningseither in the embryo or in the larva;and further, the position of each pairwhether dorsal or ventral, in each seg-ment whether thoracic or abdominal,is the same as that of every other pair.The thoracic thickenings, it is true,develop further and become folds orpockets of the ectoderm because theymust furnish the legs and wings of theimago, while the abdominal discs do notdevelop into pockets, with the exceptionof those of the external sexual organs*.Then again the thoracic and abdominaldiscs are alike in that they all help toform the imaginal hypodermis.

It seems to me that the ventral thor-acic and abdominal discs at least, arehomodynamous organs. There can beno doubt that the ventral discs of thedifferent thor’acic segments are homo-dynamous among themselves, likewisethe ventral discs of the different abdom-inal segments among themselves. Theventral thoracic discs, too, are undoubt-edly homologous to the thoracic extrem-ities of the other insects; and I thinkthere can be no doubt that the ventralabdominal discs are homologous to therudimentary, extremities which appearin the embryos of all other insects, butnot in the brachfceran dibblers. Butthe thoracic and the embryonic abdom-inal exfi’emities in other insects areundoubtedly homodynarnous organs,therefore, the ventral thoracic and ab-dominal imaginal discs in the brachycera

Pratt. 1. c. p. x97.

are also homodynamous organs, asthings which are equal to the samething are equal to each other. In thethorax these organs furnish the legs andthe ventral half of the imaginal hypo-dermis, in the abdomen they furnishonly the hypodermis, there being nolegs. The two pairs of discs whichfurnish the external sexual organs are,I think, the ventral discs of the absentninth and tenth abdominal segments.When we consider the dorsal discs

we find the matter is much. more diffi-cult. We cannot prove that the dorsalthoracic and the dorsal abdominal discsare homodynamous in the same way,because the dorsal abdominal surfaceof the insect has no extremities and norudiments of any at any time. But Ithink, although reasoning from analogyis very unsafe in such matters, it is atleast very probable that the same homo-dynamy exists on the dorsal side as onthe ventral side of the insect. The dor-sal discs have exactly the same appear-ance on the thortx and abdomen as theventral discs, and the same method oforigin, and if these facts go for anythingthere can be no doubt of the homody-namy.

If, now, this is really the case, whatis its significance? The ventral discs,thoracic and abdominal, are homolog-ous to extremities. The dorsal thoracicdiscs are homologous to wings. If theyand the dorsal abdominal discs arehomodynamous organs, are the latterhomologous to wings, too? Such anassumption is of course impossible,but it is not impossible that there

0 _PS/’Cf/.E. [February 897.

existed in some previous phyletic late, in the larval development. Forstage, paired rows t)f external, seg- instance, in the lower orders ofmental organs running down the back holometabolic insects, as in those havingof the insect from one end to the incomplete metamorphosis, theanlagenother, just as the legs at one phyletic .of the extremities appear very early inperiod extended the entire length of theventral surface, and still do in the verylowest insects, and firther, that on thethorax these organs developed finally, inthe evolution of insects, into wings.Tracheal gills might represent such or-gans. _And the fact that the dorsalprothoracic discs in Musca and thenematocera develop into the pupalspiracles lends great weight to thisnotion as these, like the tracheal gills,are respiratory organs. The wellknown theory of Gegenbaur and Lub-bock, tracing the origin of wings ininsects to tracheal gills, seems thus toobtain a new support.Another matter which seems worth

mentioning is that in different holomet-abolic insects, the extremities or thethoracic and abdominal imaginal discs(when such are present) may appear atvery different times in the ontogeny.In some insects these appear early, andin some late, in the embryonic develop-ment, in some early, and in some

the embryo. In Melophagus the tho-racic discs, homologous organs, appearrather late in the embryo, while theabdominal discs appear probably earlyin the larval period. In Corethra theimaginal discs, also homologous toextremities, delay their appearance un-til just before pupation. Thus theepigenetic period in insects, when neworgans are tbrming, does not end withthe birth of the larva from the egg, butextends over the larval and even overthe pupal period. The eabryonicdevelopment of the insect really does notend until the imago bursts from thepuparium, the embryonic, larval, andpupal periods being essentially identical.The principal significance of the pupalperiod and the metamorphosis is thatit is the time when the larval characterswhich were adopted for use during aperiod of free life in the midst of thedevelopment, and which would bevalueless to the imago, are correctedor abandoned.

DIAPHEROMERA FEMORATA.

I FIND among my notes the following ob-servations on this insect in captivity.The general color of the female is brown,

marked by streaks and dots of a lighterbrown or shaded darker at the sides of thebody and at each joint. The face is orange,the antennae and palpi brown. The legshave a greyish green tinge and are lighter

than the body, but darker at the ends of thejoints. The fore legs are always different incolor from the others being brown above anddull yellow below and when stretched for-ward beside the appressed antennae (whichjust surpass them), as is always the caseat rest, they make the insect appear a thirdlonger than it is. They eat the edge of aleaf, usually straddling it with their legs andin an hour will devour a piece an inch long

Submit your manuscripts athttp://www.hindawi.com

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Anatomy Research International

PeptidesInternational Journal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporation http://www.hindawi.com

International Journal of

Volume 2014

Zoology

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Molecular Biology International

GenomicsInternational Journal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

The Scientific World JournalHindawi Publishing Corporation http://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

BioinformaticsAdvances in

Marine BiologyJournal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Signal TransductionJournal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

BioMed Research International

Evolutionary BiologyInternational Journal of

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Biochemistry Research International

ArchaeaHindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Genetics Research International

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Advances in

Virolog y

Hindawi Publishing Corporationhttp://www.hindawi.com

Nucleic AcidsJournal of

Volume 2014

Stem CellsInternational

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

Enzyme Research

Hindawi Publishing Corporationhttp://www.hindawi.com Volume 2014

International Journal of

Microbiology