SMITHSONIAN MISCELLANEOUS COLLECTIONSVOLUME 81, NUMBER 3
MORPHOLOGY AND EVOLUTION OF THEINSECT HEAD AND ITS APPENDAGES
BYR. E. SNODGRASSBureau of Entomology
(Publication 2971)
CITY OF WASHINGTONPUBLISHED BY THE SMITHSONIAN INSTITUTIONNOVEMBER 20, 1928

SMITHSONIAN MISCELLANEOUS COLLECTIONSVOLUME 81, NUMBER 3
MORPHOLOGY AND EVOLUTION OF THEINSECT HEAD AND ITS APPENDAGES
BYR. E. SNODGRASSBureau of Entomology
S*^li(^,5"^'[is:w.'o^P^ik7l Zl
(Publication 2971)
CITY OF WASHINGTONPUBLISHED BY THE SMITHSONIAN INSTITUTIONNOVEMBER 20, 1928
BALTIMOriE, MD., V. S. A.
MORPHOLOGY AND EVOLUTION OF THE INSECT HEADAND ITS APPENDAGESBy R. E. SNODGRASSBureau of EntomologyCONTEXTS PACKIntroduction 2I. Evolution of tlie arthropod head 2Cephalization 3Development of the body iit segmented animals 12The protocephalon 19The definitive arthropod head 27II. General structure of the insect head 33The head capsule 34The labrum and epipharynx 41The stomodeum 42The hypopharynx 45The tentorium 50III. The head appendages 56The antennae 57The postantennal appendages 59The gnathal appendages 60The mandibles 62The first maxillae 74The second maxillae 77Morphology of the gnathal appendages 79IV. Summary of important points 9°V. The head of a grasshopper 94Structure of the cranium 95The antennae 99The mandibles ^0°The maxillae 102The labium 106The preoral cavity and the hypopharynx 107The stomodeum 112The mechanism for moving the head 118VI. Special modifications in the structure of the head 120Modifications in the fronto-clypeal region 120Modifications in the posterior ventral region of the head 125VII. The head of a caterpillar 131Structure of the head capsule • ^3~The antennae ^37The mandibles ^38The maxillae and lal)ium I39The stomodeum ^45The musculature of back of head, and nature of insect neck 150Abbreviations used on the figures 153References ^55Smithsonian Miscellaneous Collections, Vol. 81, No. 3
2 SMITHSOMAX MISCKLLANEOLS CULLECTlUXS VOL. 8lINTRODUCTIONTt is regrettable that we must arrive at an understanding of thingsby way of the human mind. Lacking organs of visual retrospection,for example, we can only hold opinions or build theories as to thecourse of events that have preceded us upon the earth. Knowledgeadvances by what biologists call the method of trial and error, but themind can not rest without conclusions. Most conclusions, therefore,are premature and consequently either wrong or partly wrong, and,once in every generation, or sometimes twice, reason back tracks andtakes a new start at a different angle, which eventually leads to a newerror. By a zigzag course, however, progress is slowly achieved. Error,then, is a byproduct of mental growth. It is not a misdemeanor inscientific research unless the erring one clings to his position when heshould see its weakness. It is better to write beneath our most positivecontentions that we reserve the right to change of opinion withoutnotice. The reader, therefore, should not take it amiss if he findscertain conclusions drawn in this paper that do not fit with formerstatements by the writer, for no apology will be ofifered.
I. EVOLUTION OF THE ARTHROPOD HEADThe head, as a dififerentiated region of an animal, is a more ancientstructure than is any other specialized part of the body, and a properunderstanding of the head structure involves an examination of theevidence of cephalic evolution from the very earliest period whenevidence of head development can be found. Most of the Arthropodahave well developed heads, and that the arthropod head is a specializedbody region, just as is the thorax or the abdomen in forms where thesebody regions are difi:'erentiated, is shown by the fact that in the embryoit consists of a series of body segments. In most cases, and particularlyin insects, however, the head differs from the other body regions inthat its component segments become so thoroughly consolidated inthe adult as to leave little evidence of the primitive elements in thehead structure. Even in the ontogenetic record the true history of thehead development is so oljscure in many respects, and so much deletedin the early passages, that, though all the facts of embryology wereknown, it is probable that the assembled information would still givebut an incomplete account of the phylogenetic evolution of the head.It is only by a comparative study of the head structure and its develop-ment in the various arthropod groups, and by an effort to correlatethe known facts of arthropod organization with what is known inother animals successively lower in the scale of evolution, that we
NO. 3 INSECT HEAD SNODGRASS
may arrive at a satisfactory conclusion as to the steps by which thecomplex head of an insect has been evolved—and even then we mustallow much for errors of judgment.CEPHALIZATION
It has been but little questioned that the numerous groups of meta-zoic animals are derived from a creature resemliling the 1)lastula ofembryonic development (fig. i A). The embryonic blastula is exem-plified, among living animals, in the early stage of the free-swimminglarval'planula of the Coelenterata (fig. 2 A). The planula developsBid
Fig. I.—Typical early stages in general embryonic development.A hlastula stage diagrammatic, consisting of a blastoderm (SW) surround-ing a blSocoele' cavity" (BIc). B, C, D,. stages in development of a chUonCfrotn Kowalevskv i88^) : B, d fferentiation of cells m blastula, L, gastruia
^,on 7oS gast icoele cavit; (GO. lined with endoderm (tud), and opemngthrough blaftfpore (Bp) ; D, later stage, showing ongm of mesoderm layers(Msd) just within lips of blastopore.directly from the coelenterate egg, and has the form of a hollow massof cells the outer surface of which is covered with vibratde cilia.The uniform motion of the cilia propels the animal through the waterin the direction of one axis of the body (fig. 3), and thereby one endis distinguished as anterior and the opposite as posterior. The creaturethus becomes uniaxial and bipolar, though as yet there may be nodifferentiation of body structure. The functional differences at thetwo poles of the body, however, determine the course of the subse-quent development of physical characters. Structural dififerentiationof the end of the body that is forward in usual progression is calledcephalization, a term meaning the process of evolving a head.
SMITHSONIAN MISCELLANEOUS COLLECTKJNS VOL.The body of the planula is usually larger at the anterior end(figs. 2,3), and only in this does the planula attain cephalization inthe Strict sense. Its principal structural dififerentiation occurs at theposterior pole, where there takes place an ingrowth of cells(fig. 2 B-D) that soon fills the hollow of the body, and finally, by theappearance of a cavity within its mass, becomes the wall of the stomachof the mature animal. The process of forming a primitive stomach, orarchenteroii. as it takes place in the planula, is typified by that ofgastrulation in ordinary embryonic development (fig. i A-D). Theplanula, of course, is a specialized larval form, and its manner ofcndoderm formation can not be taken as showing how the archenteronwas evolved, but the free-swimming planula does show that the primi-tive mouth, or hlasfoporc (fig. jC,D,Bp), was formed at the
A B C DFig. 2.—Formation of the endoderm in a coeleiiterate planula larva by pro-liferation of cells from posterior pole. (From Hatschek, 1888, after Claus.)Blc, blastocoele ; Pld, blastoderm ; Bed, ectoderm ; End, endoderm.posterior pole of the body, and not at the anterior pole. It is interest-ing to note, therefore, that the position of the mouth opening wasnot necessarily a primary determining factor of cephalization ; thepractical site for a mouth in a free-swimming, ciliate animal wasdetermined by the direction of the animal's movement. Korscheltand Heider (1895) have stated, if a monaxial, heteropolar planula isallowed to swim through water containing particles of carmine, itcan be seen that the particles are rej^ulsed at the anterior and lateralparts of the body, but that they accumulate at the posterior pole.
'' Here accordingly," say Korschelt and Heider, " was a favorableplace for the reception of particles of food, and by a flattening orshallow invagination of the posterior pole these favorable conditionswere increased. The archenteron, therefore, in its earliest beginningswas a pit in which to catch particles of food."
NO. 3 INSECT HEAD—SNODGKASS 5This is a satisfactory explanation of the origin of the gastrula ifnot questioned too closely; but Bidder (1927) rather disturbs theidea with his statement that " the laws of viscous matter make it clearthat the free-swimming gastrulae we observe as larvae could neverearn their own living, since the stream-lines would carry every particleof food outside the cone of dead water which is dragged behind thegastrula mouth." On the other hand, Bidder admits, '" creepingplanulae or gastrulae might pick things up." A creeping animal,however, would never in the first place develop a mouth at the rearend of the body. What we want is an explanation of the originalposterior position of the blastopore, and if none offered will sufifice.we must be content with the fact.The further history of the coelenterate larva has no bearing on theevolution of insects, for the creature soon becomes attached by itshead end, and. probably as a result of the sedentary, plant-like habits
Fig. 3.—Free-swimming plaiuila larva of a coelenterate, Sympodiuin corral-1aides, with ciliated ectoderm, and completeb'-formed endoderm. (From Ko-walevsky and Marion, 1883.)
of its immediate ancestors, develops into a polype or jellyfish havinga radiate, flower-like type of structure. Some writers have suggestedthat the worms and the arthropods may have been evolved from anelongate medusa, but it seems more probable that the Coelenterata,the Annelida, and the Arthropoda are all to be traced back to a free-swimming gastrula ancestor. The mature planula is a specializedgastrula, but it is of general interest in that it gives us a ])assingglimpse of a free-living animal in the blastula and gastrula stages ata time when cephalization was first established in the Metazoa.The structure and development of the arthropods suggest thatthese creatures were developed from forms adapted to a creepingrather than a swimming mode of progression. Some planula larvaelack cilia and have creeping habits, and such forms, though they havenothing to do with the arthropod ancestors, show that a free-livingcreature in the blastula or gastrula stage may change its mode of])rogression. The creeping habit as an habitual mode of progressionentails some fundamental structural adaptations. An animal that crawls
6 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
must keep one surface against the support, and thus it estabHshes afunctional distinction between its upper surface and its lower surface,which has led to the structural differentiation of dorsivcntrality ; andfrom this, in combination with movement in one direction, finally,bilateral symmetry of organization necessarily follows.Progression by crawling instead of by swimming alters the wholestatus of the relation between the animal and the environment. Amouth at the posterior end of the body now becomes quite impracti-cable, and embryonic history shows that crawling animals proceeded torectify the defect, supposedly inherited from their free-swimmingciliate ancestors, by lengthening the mouth, or blastopore, in a for-ward direction on the under side of the body. In the young Peripatusembryo, for example (fig. 4 A, B), the l)lastopore is a long slit on the
Fig. 4.—Early stages in the development of Peripatus capensis. (From Balfour,1883.)The blastopore {Bp) elongates on ventral surface of embryo, and thencloses except at the two ends (C) where open extremities become mouth andanus. Segmentation appears as series of coelomic sacs in mesoderm (see fig. 6,Msd).
ventral surface of the blastoderm. Later, the edges of the slit cometogether (C) and unite except at the two ends, where openings re-main into the archenteron that become the mouth and anus of themature animal. In insects and other arthropods, the process of gas-trulation in the embryo (fig. 5 A) is clearly a modification of that inPeripatus, by which many of the details have been omitted and thewhole procedure greatly altered. In most insects (fig. 5B), gastru-lation resembles that of the planula (fig. 2) in so far as it takes placeby an internal proliferation of cells from the blastoderm, but mostof the gastrulation area gives rise to mesoderm, the true endodermbeing formed only at the two extremities of the inner layer (fig. 5 C,AMR).
NO. 3 INSECT HEAD SNODGRASSThe mesoderm, and the associated mesenchyme, play an importantpart in the organization of all the higher Metazoa, since they form theinternal organs that lie hetween the ectodermal covering of the bodyand the endodermal epithelium of the alimentary canal. The meso-derm is of particular importance in segmental animals because it isin this layer that metamerism originates. Mesoblastic tissue is pro-duced in a gastrulated embryo in two ways : First, in the form ofscattered cells proliferated from the inner surface of the invaginatedendoderm ; and second, in the form of cell layers. The scattered cellsPre
B r/ AsdBp EcdAMRMsd
Ecd
Fig. 5.—Gastrulation in insects.A, embryo of Lcptiuotarsa deciml'mcata with long gastrulation groove, orblastopore {Bp), on ventral surface. (From Wheeler, 1889.)B, cross section through blastopore of embryo of Forficula, showing mesoderm(Msd) formed by invagination of middle plate. (From Heymons, 1895.)C, anterior mesenteron, or endodermic, rudiment (AMR) formed at anteriorend of mesoderm (Msd) in embryo of honeybee. (From Nelson, 1915.)form a loosely coherent mesenchyme ; the cell layers constitute thetrue mesoderm. The primitive mesoderm cells are given ofif from theendoderm near where the latter joins the ectoderm, that is, just withinthe lips of the blastopore (figs, i D,6, Msd). In the young annelidlarva, the mesoderm cells first form two lateral bands of tissue at theposterior end of the body (fig. 7 D, Msd). Later, the extended meso-derm tracts become excavated by a series of cavities, the coelomicsacs, which mark the beginning of segmentation. In Pcripatiis (fig. 4),likewise, two rows of coelomic sacs {Msd) are formed as pairedcavities in the mesoderm, which extends laterally between the ecto-derm and the endoderm along the line of junction between these two
8 SMITHSONIAN AnSCELLANEOUS COLLECTIONS vor 8
1
layers (fig. 6,Msd). In the annelids, the coelomic sacs form theentire segmented body cavity ; in Pcripatus and most of the arthropods,the greater part of the definitive body cavity is derived from a spacebetween the ectoderm and the endoderm lined with mesenchymaticcells.It is most important to bear in mind the intimate relation that existsbetween the mesoderm and the endoderm. In the arthropods, especi-ally in insects, the process of gastrulation, as above noted, is greatlymodified, and mesoderm tissue alone is proliferated along the greaterpart of the length of the blastopore area, which in only a few general-ized forms appears as a true opening. At each end of the mesoderm,however, endodermal tissue is formed (fig. 5 C, AMR), and the two
~-M5d-
FiG. 6.—Formation of mesoderm in Peripatiis cafciisis. (From Balfour, 1883.)Cross sections of embryos through blastopore, showing formation of meso-dermic coelomic sacs (Msd) from endoderm (End) just within lips of blasto-pore (Bp).
endoderm rudiments mark the anterior and the posterior limits ofthe mesoderm—consequently, they define the area of segmentation.It is unnecessary to speculate as to the phylogenetic steps that mayhave led from the early creeping gastrula form of animal to the worm-like ancestor of the arthropods, but we must note the importantadvance in cephalization, and the possibilities of further head develop-ment that came with the establishment of a mouth at the anterior endof the body. Food, whether living or inert, had now to be recognizedand seized on contact. Consequently, it became highly important tothe animal to be able to determine its course according to favorableor unfavorable conditions of the surroundings. The ectoderm of theanterior end of the body developed a special sensitiveness to environ-mental changes, and, probably by means of ectodermal processes ex-tending into the body, communicated the stimuli received from with-out to the internal tissues. Certain groups of the sensitive cells then
NO. 3 INSECT lIEAll SXODCRASS 9
were withdrawn into the hody where they became the rudiments of acentral nervous system. Other sensory cells, remaining at the surfacebut sending processes inward to the buried cells, formed the peripheralsensory system. This anterior differentiation of sensory and con-ductive tissues opened still other possibilities of cephalization, whichhave led to the development of the brain and all the com])lex senseorgans located on the head in higher animals.It is difficult to establish, by concrete example, the contention thatthe change in the position of the mouth resulted from a change in themanner of locomotion, but it is indisputable that the ancestors of theworms and the arthropods must have assumed the crawling habit atsome stage in their evolution. The chaetopod annelids, in their em-bryonic development, arrive at a first larval stage known as a trocho-phore (fig. 7 D), which is a free-swimming creature with well differen-tiated anterior and posterior poles, and a dorsal and a ventral surface,with the mouth situated anteriorly in the latter. If dorsiventrality isto be attributed to a creeping mode of locomotion, then there mustbe some stage omitted between that represented by the free-swimmingplanula, and that of the free-swimming trochophore, because there isno evident reason, otherwise, why two forms having the same modeof life should have an organization so different. The trochophore iswithout doubt a specialized larval form modified secondarily for aswimming habit. It can not, therefore, be taken to represent an an-cestral form of the worms ; but it is the only free-living creature thatshows us the beginning of the worm organization, and its structurecan certainly be traced into that of the arthropods.The annelid trochophore is typically ovate in shape with the largerend forward (fig. /D), or rather, upward, since the creature floatsupright in the water, but the side in which the mouth (Mth) is locatedis called the ventral surface because it becomes the under surface ofthe mature worm. The mouth lies a little below the middle of thebody, and the anus (Aji) is situated at the posterior pole. The bodyis surrounded by several bands of vibratile cilia. The principal band(b), comprising usually two rows of cilia, is situated on the widestpart of the body and just before the mouth. It divides the animal intoa preoral, or prostornial, region (Pst), and into a postoral, or iiictas-tomial, region (Mst). A second band of cilia (c) is generally presenta short distance behind the mouth, and sometimes there is a third,preanal band (G, d) near the posterior end, which sets off a terminalcircumanal region, or pcriproct (Ppt). At the anterior end of thebody there is a central tuft of tactile hairs (G, a), a pair of smalllateral tentacles (Tl), and one or more simple eye spots (0).
10 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lThe mouth of the trochophore (fig. 7 D, Mth) opens into an ecto-dermal stomodeum (Stoiii), which leads into a large endodermalstomach, or mesenteron (Ment), of two parts, the second of whichcommunicates with the exterior through a short ectodermal procto-M5b
Mth
^M5t
VMcl VNvFig. 7.—Structure and development of an annelid trochophore larva. (FromHatschek, 1888-91, but figures A to F turned in position of adult worm withmouth downward.)A, blastula stage with endoderm cells (End) differentiated at posterior pole.B, gastrulation, showing a primitive mesoblast cell (Msb) of one side. C,gastrulation completed, mesenteron {Mcnt) detached from ectoderm at posteriorend, its anterior end opening through stomodeum (Stoni) and mouth. D, full-grown trochophore larva of Polygordhis. E, diagrammatic view of the musclesystem. F, the nervous system. G, ventral surface of a trochophore.a, apical tuft of cilia; An, anus; AP, apical plate; b, preoral band of cilia;BIc, blastocoele ; Bp, blastopore : c, postoral band of cilia ; d, circumanal band ofcilia; DMcl, dorsal muscle; DNv, dorsal nerve; End, endoderm; Ment, mesen-teron; Msb, primary mesoblast cells; Msd, mesoderm; Mst, metastomium ; Mth,mouth ; Nph, nephridium ; O, eye spot ; Ppt, periproct ; Proc, proctodeum ; Pst,prostomium ; Stoni, stomodeum ; TI, tentacle ; VMcl, ventral muscle ; ]''Nv.ventral nerve.demn {Proc) . In development, the endoderm is formed by invagina-tion at the posterior end of the body (fig. 7A, B, iiwrf), but theblastopore {Bp) soon shifts to the ventral surface (C) and elongatesforward. The posterior part of the blastopore is then closed ; the
NO. 3 INSECT HEAD SNODGKASS II
anterior open extremity is carried inward by an ectodermal invagina-tion which becomes the stomodeum (C, Stom), the external openingof which is the definitive mouth (Mth). The proctodeum and theanus are formed later by a posterior invagination of the ectoderm,and the proctodeum secondarily opens into the posterior end of thestomach. At the anterior end of the preoral region of the body, orprostomium, the ectoderm is thickened to form a sensory apical plate(D, G, AP) beneath the sensory organs here located, and from itectodermal nerve tracts extend posteriorly in the body wall (F).Typically, there is a pair of dorso-lateral longitudinal nerves (DNv),and a pair of ventro-lateral nerves (VNv). The simple musculatureof the trochophore (E) is developed from mesenchyme tissue; theepithelial mesoderm forms only the pair of mesoderm bands (D, Msd)and a pair of nephridia (Nph) in the posterior part of the body.This description of the trochophore is based mostly on that of Hats-chek (1888-1891), from whose work the illustrations of figure 7 aretaken.The trochophore develops into the worm form of its parents by ametamorphosis involving an elongation of its posterior end (fig. 9),accompanied by a reduction of the cephalic swelling, until finally, inthe adult, the only dififerentiation in the head region is an anterior,median prostomial lobe overhanging the mouth (fig. 10, Pst). Theprostomium usually bears the principal sensory areas or organs ofthe worm, and a ganglionic nerve mass is differentiated from theinner surface of its ectoderm, which becomes the supraoesophagealganglion, or brain, of the annelid. In the Polychaeta, the prostomiummay bear one or more pairs of eyes, and several pairs of sensory ten-tacles (fig. 10). As the body of the young worm elongates, it be-comes transversely segmented, the somites increasing in numberposteriorly as the segmented area lengthens.The young arthropod embryo, in its first definite form (fig. 8 A),consists of a large head region, the so-called cephalic lobes (Pre), andof a slender body (Bdy). The mouth (B, Stom) is situated on theventral surface of the cephalic enlargement. The proctodeal invag-ination and the anus are formed, usually in a later stage, at the pos-terior end of the body.The large-headed stage of the young arthropod embryo has a cer-tain resemblance to the trochophore stage of the annelid larva ; but itis probable that the similarity between the two forms has no geneticsignificance, and that the size of the cephalic lobes in the arthropodembryo is to be explained as an acceleration of development. Yet,it is evident that the cephalic region of the arthropod embryo cor-
12 SM iTHSOXlAX M ISCKLLANEOUS COI.LECTIONS NUE. 8l
responds with the prostomial and metastomial regions of the trocho-phore, and includes also the next following somite, for the first an-tennae, which are the appendages of the second somite of the arthro-pods, are developed on the cephalic lobes of the embryo (fig. 8 B, C.
'D.Avt). In the insect embryo, furthermore, the region of the rudi-mentary second antennal appendages, or the tritocerebral segment(fig. 8 C, ///), is often incorporated into the cephalic lobes. It isprobable, therefore, that the very early insect embryo represents ahigher stage of cephalic evolution than does the annelid trochophorePre .Pre
-Ant PrePnt
CFig. 8.—Young stages of insect embryos, showing cephalic lobes, beginningof segmentation, and formation of appendages.A, germ band of Blatclla gcnnanica on seventh day, with cephalic lobes {Pre)indicated. (From Riley, 1904.)B, embryo of same, about nine days old, with cephalic lobes developed intoa distinct protocephalon (Pre), antennae (Ant) appearing, stomodeum iStotn)indicated as thickening of ectoderm. (From Riley, 1904.)C, young embryo of Lcpisiiia, with well-developed protocephalon (Pre),bearing stomodeum and rudiments of antennae, with tritocerebral segment(///) closely associated with its base. (From Heymons, 1897.)D, embryo of BhilcUa late in tenth day, with labrum (Lm) , mouth (Mth),and antennae (Ant) on protocephalon {Pre), followed by rudiments of post-antennal appendages (Put), mandibles (Md), first maxillje (iMx), secondmaxillje (jAIx), and legs (Li). (From Riley, 1904.)larva, in as much as it has already progressed to a point where thehead includes two or three of the body segments.The definitive head of the arthro])od may contain as many as six orseven of the body segments. Before going farther in the study of]irogressive cephalization. then, it will be necessary to understandsomething of the development and general organization of the bodyin segmented animals.
r>i:\EL01^MENT OF THE BODY IN SEGMENTED ANIMAL.SIn the Annelida, the worm form is developed from that of thetrochophore by an elongation of the posterior part of the larval body
NO. 3 INSECT HEAD—SNODC.KASS 13
(fig. 9), and by a decrease in the relative size of the cephalic enlarge-ment. The young worm is a cylindrical creature with only a com-paratively small prostomial lobe projecting before the mouth. Withthe elongation of the body, the alimentary canal and the mesodermbands are correspondingly lengthened, and the trochophore musclesand nerves are continued into the new region. The external surfaceof the body of the trochophore is marked ofif into several areas bycircular bands of cilia ; the worm body, on the other hand, is con-stricted by transverse grooves into a series of segments, or somites.The segmentation of the adult worm originates in tJie mesodermbands by the development in the latter of a series of paired coelomicsacs. Secondarily, the mesodermic divisions become impressed u])on
-MentAFig. 9.—Diagrams of the development of an annelid trochophore larva, andearly stage in its metamorphosis into a segmented worm. (F"rom Hatschek,i888-'9i.)A, early larval stage, showing a primary mesohlast cell {Msh) of oneside. B, later stage in which the mesoblast has formed scattered mesenchymecells (Msc), and a ventro-lateral band of mesoderm (Msd) in each side of thebody. C, early stage of metamorphosis in wliich each mesoderm band hasdivided into a number of primary segments.the body wall, and the segmentation expressed externally by a seriesof transverse, circular grooves on the intersegmental lines. In theworms, the segments increase in number from before backward bythe differentiation of new segments between the last one formed andthe periproct. The latter remains as an undifferentiated terminal pieceof the body bearing the anus. The prostomial region of the trochophorebecomes the prostomium of the adult worm; the metastomial regionin the Archiannelida constitutes the first body segment, or that im-mediately behind the mouth ; but in the Polychaeta and Oligochactathe metastomium is said to unite with the next somite to form a com-pound peristomial segment.In the adult annelid (fig. 10), the body, as distinguished from thehead, is all that part of the worm that lies posterior to the mouth(A, MfJi), and the only differentiated head region is the prostomium
14 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
(Pst), though the peristomium (Prst) is sometimes called a part ofthe " head ". The prostomium varies in size from a small lobe over-hanging the mouth, to a large area extended posteriorly into thedorsal region of the peristomium (B). It bears the principal sensoryorgans, eyes and tentacles (O, TI), of the anterior part of the worm.The alimentary canal extends entirely through the body from themouth at the anterior ventral edge of the peristomium to the anusat the end of the periproct.The nervous system of the adult annelid consists of a brain locatedbefore or above the oesophagus, derived from the ectodermal apicalplate of the prostomium of the trochophore (fig. 7 F, AP), and ofa ventral nerve cord of segmental ganglia and intervening connec-tives, formed from two nerve strands developed from the ectoderm
Fig. 10.—Anterior end of an adult polychaete annelid worm, Nereis virens.A, lateral view. B, dorsal view.Mth. mouth; 0, ocellus; Pip, prostomial palpus; Ppd, parapodium ; Prst.peristomium ; Pst, prostomium ; Tl, prostomial tentacles.
of the ventral body wall and prolonged from the ventro-lateral nervestrands of the trochophore. The ventral cord, therefore, is connectedwith the brain by a pair of connective nerve cords passing to the sidesof the oesophagus. Usually, the two ventral nerve strands are unitedalong the midline in adult worms, but in certain polychaete forms(Serpulidae) the two cords are said to remain separate, though con-nected by transverse, interganglionic cominissures in each segment.In the Arthropoda, body segmentation begins during an early em-bryonic stage, and the somites are added, in general, as in the annelids,from before backward by the differentiation of new segments behindthe last one formed. In some of the arthropod groups, segmentationis completed in a postembryonic stage; in insects (except Protura),however, the somites are all defined before the creature leaves theQgg, and the typical sequence of segmentation is not always followed.
NO. 3 INSECT HEAD SNODGRASS I 5The component segments of the cephalic loljes, or head of the arthro-pod embryo, are never distinct, but the subsequent development ofthe anterior nerve centers shows that the lobes comprise two segmentsat least, in addition to a prostomial region, and that usually a thirdsegment is more or less included in their posterior part.The way in which metamerism arose in the phylogenetic history ofsegmented animals is not known, and it is not necessary to believethat the method of segment formation in either the annelid larva orthe arthropod embryo gives a picture of primitive segmentation inthe course of evolution. The development of the trochophore into theworm is clearly a process of metamorphosis, that is, it is the returnof a specialized, aberrant larva to the ancestral form represented morenearly in that of the adult ; and it is well known that embryos do notkeep closely to the phylogenetic path in the details of their development.Since so many other essential features in the body structure of animalsare connected with the mode of locomotion, the writer holds as mostprobable the idea that segmentation also had its beginning as anadaptation to a specific kind of movement. The creeping, worm-likeancestors of the annelids and arthropods certainly at an early periodmust have developed a contractile tissue in their mesoderm bands—
-
that they did so is attested by the early development of a centralnervous system consisting of motor neurons, following the lines of thelater established ventral longitudinal muscle bands. It is, then, clearthat a breaking up of the contractile tissue into short lengths wouldgive a greater efficiency of movement, with the possibility of morevariety of action, and that, wdth the differentiation of true musclefibers, the attachment of the ends of the fibers to the ectoderm wouldcarry the metamerism into the body wall. The fact that embryonicsegmentation begins anteriorly and progresses backward, in itselfsuggests that metamerism originated in a creeping animal ; in a free-swimming form, the progress of segmentation should be the reverse,for the motile region of the animal would be the tail end. Organs de-veloped at the time of metamerism or subsequent to it, such as ne-phridia, tracheae, and external appendages, are repeated in each seg-ment, those antedating segmentation either remain unsegmented, asthe alimentary canal, or take on a secondary segmentation, as do thebody wall and the nervous system.There are other theories of metamerism: Hatschek (i 888-1 891)enumerates five views that have been proposed to explain the originof body segmentation, but none of them is based on the simple factthat in embryonic development, metamerism begins in the mesoderm
i6 SMITHSONIAN M iSCKLLANEOUS COLLECTIONS VOL. 8l
and secondarily spreads to other tissues. The older locomotion theorywas defective in that it attributed the formation of segments to themechanical stress of movement.At the completion of metamerism, a segmented animal has attainela generalized structural stage in which it consists of a segmented bodypart coextensive with the length of the alimentary canal (fig. ii),and of a prostomial region (Pst) anterior to the mouth (Mth). Sincethe mouth in annelids and arthropods marks the site of the originalanterior extremity of the blastopore on the ventral surface of thebody (figs. 4 B, 5 A, Bp), it is evident that nicsodcniuil segments cannot be formed morphologically anterior to the mouth, and therefore,that the preoral region is never truly segmented. The common idea,then, that the arthropod mouth lies behind the first head segment, or,as proposed by some writers, behind the second or even the third seg-Stom Ment Proc
VNC
-Diagram of the structure of a theoretically generalized segmentedFig. II.—animal.An, anus; Arc, archicerebrum : Mcut, mesenteron; Mth, mouth; Ppt, peri-proct; Proc, proctodeum; Pst, prostomium ; Stoiii, stomcdeum; VNC, ventralnerve cord.
ment, disregards the fundamental relation between the endodermal andmesodermal layers. Segmentation can not exceed the extent of themesoderm, and the primitive extent of this layer in the annelids andarthropods is defined by the positions of the mouth and the anus. Theblastopore never extends quite to the true cephalic extremity. Thestomodeal invagination, which gives rise to the definitive mouth, isthus preceded by an unsegmented prostomium. The closed posteriorend of the blastopore, however, is at the posterior extremity of thebody, where the blastopore and endoderm originated, and the laterformed ]n-octodeum, therefore, opens terminally in the periproct. Insome arthropods a median lobe, or suranal plate, grows out over theanus from the periproct, and simulates the prostomial lobe at theanterior end of the body. Likewise, there may be lateral and subanallobes of the periproct.In as much as the most important evidence of the segmentation ofthe arthropod head is derived from a study of the cephalic nerve
NO. 3 INSECT HEAD SNOUGRASS 17
masses, it will be necessary to understand next the essential featuresin the evolution of the central nervous system in segmented animals.The annelids, as already noted, have a ganglionic nerve mass lyingin the anterior part of the body, before or above the stomodeum,which takes its origin from the ectodermal apical plate of the pro-stomium (fig. 7 F, Ap) . This, the most primitive brain of the annelid-Arc
a,.NCCom
Fig. 12.—Diagrams suggesting the evolution of a central nervous S3'stem ofannelid-arthropod type of structure.A, theoretical structure of nervous system in an unsegmented pre-annelidform, consisting of a prostomial archicerebrum (Arc), and of two ventro-lateral nerve cords (NC), connected medially by transverse nerves, and givingoff nerves laterally to body wall and internal organs. Nerve cells dififused alongthe cords.B, simple nervous system of the ladder type in a segmented animal. Thenerve cells aggregated into segmental groups, or ganglia (Gng), along the cords ;the intervening parts of cords converted into connectives (Con), and thetransverse ventral nerves forming commissures (Com) between the ganglia.C, the segmental pairs of ganglia united into compound ganglia of a medianventral nerve cord (J'NC), in which the first, or suboesophageal, ganglion(SocGng) is postoral, and connected with archicerebrum of prostomium (Arc)by connectives (CocCon^ encircling the mouth (Mth).
arthropod series (fig. 12 A, Arc), has been named by Lankester(1881) the archicerebrum (a happy, though mismated union of lin-guistic elements). In the trochophore, a pair of dorsal and a pair ofventral nerves (fig. 7 F, DNv, VNv) extend backward from theapical plate, but in the adult worm and in arthropods only the nervesof the ventral pair are retained. In the more primitive condition, thetwo ventral nerve strands have a latero-ventral position (fig. 12 A,
./
l8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lNC), and it seems reasonable to attribute the special development ofthese nerves in a creeping animal to the special development of sensi-tivity in the ectoderm along the edges of the body in contact with thesupporting surface.That the nerve strands were primarily unsegmented is shown by thefact that they are not ganglionated in the Archiannelida, and by theircondition in Pcripatiis where the nerve cells are still distributed alongthe length of the cords, and segmental grouping of the cells is butslight. A concentration of the nerve cells of the cords in each seg-ment is, then, only a simple adaptation to efficiency where metamerisml)ecomes the established body structure. After the segregation of thenerve cells into pairs of segmental ganglia, the intervening fibroustracts of the cords remain as connectives between the successive gangliain each chain, while transverse ventral nerves, originally going fromone cord to the other, become commissures uniting the ganglia ofeach segmental pair. In this way, apparently, a simple nervous system,formed primarily as two parallel strands of nerve tissue, became asegmented system of the ladder type (fig. 12 B) . In the further courseof evolution, the ganglia of each segment come together medially andcombine into a single ganglionic mass, or segmental ganglion (C),which, in some arthropod groups, acquires an addition from a sec-ondary median cord of nerve tissue developed from the ventralectoderm along the midline of the body. The transverse commissuresare now internal fibrous tracts of each double ganglion, but the length-wise cords persist usually as paired interganglionic connectives. Eachdefinite body ganglion, or pair of ganglia, innervates, in general, onlythe parts and organs of its own segment, but all the ganglia show atendency to migrate along the cords, especially in a cephalic direction,and to unite with other ganglia to form composite ganglionic masses.Whatever may be the final position of any pair of ganglia, however,its nerves in most cases still go to the segment in which the gangliaoriginated. The nervous system, thus, often gives a key to the bodysegmentation where the latter is obscure or obliterated.The next important stage of development is that, characteristic of thearthropods, in which are formed the external segmental appendages.The organs designated " appendages " in the limited sense are hollow,ventro-lateral outgrowths of the body wall (figs. 13, 14, 22), whichbecome movable by muscles inserted on their bases, and flexible by aseries of joints in their walls, also provided with muscles. Here again,we connect structural evolution with movement, for undoubtedly thesegmental appendages in the first place were all organs of locomotion,giving a new power of movement supplanting the wriggling and
NO. 3 INSECT HEAD SNODGRASS 19
creeping of earlier ancestral forms. The question of whether theappendages were first used for propulsion through the water, or forprogression on a solid support will not be discussed here, but, in thecourse of their evolution, the appendages have become specialized toserve a great variety of functions. Moreover, by the functionalgrouping of the appendages, the corresponding body segments havethemselves become differentiated into groups forming often quitedistinct body regions (fig. 13 B), of which the head of an insect is oneof the most highly evolved.
Fig. 13.—Young insect embryos at a stage when the thorax is alreadydifferentiated, but in which the gnathal segments are not yet added to theprotocephalon to form the definitive head.A, embryo of Lepisma saccharina (from Heymons, 1897). B, embryo ofRanatra fusca (from Hussey, 1926).Ab, abdomen; Ccr, cercus; Gn, gnathal segments; ///, tritocerebral segment:Li, first leg; Lm, labrum ; Md, mandible; iMx, first maxilla; sMx, secondmaxilla ; Pp, " pleuropodium " ; Pre, protocephalon ; Th, thorax.THE PROTOCEPHALONThe arthropods differ from the annelids in the possession of a com-posite head, or syncephalon, formed by the union of several of theanterior segments with the prostomium.In the embryonic development of most Arthropoda the head isfirst differentiated as a swelling of the anterior end of the body, form-ing the so-called cephalic lobes (fig. 8 A, B, Pre). On this region aredeveloped the labrum (D, Lm), the eyes, the stomodeal invagination(B, Stom), or mouth (D, Mth), the antennae (Ant), and in somecases the postantennal appendages, when the last are present (fig.22 A, 2Ant). The cephalic lobes soon become a very definite em-bryonic head (fig. 13A, B, Pre), which either remains as the entire
20 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
adult head (certain crustaceans), or constitutes the basic structure towhich other body segments are added later to form the definitive head.It is impossible, therefore, to escape the conclusion that the primaryembryonic head represents an early phylog-enetic stage of cephaliza-tion, which was characteristic of the ancestors of all the arthropods.This first arthropod head may be termed the protoccphalon (proceph-alon. Patten, 1912) to distinguish it from the prostomial head of theannelids, which might fittingly be designated an archieephalon, thoughCram]iton (1928a) has proposed this term to denote a later formedcephalic region composed of the protocephalon and the mandilnilarsegment.There has been some uncertainty as to the number of segments in-volved in the protocephalon, for the segmentation of the cephalic lobesis never clearly marked in the eml)ryo, and the existence of primaryhead segments is usually indicated rather by the presence of the headappendages, and by the divisions of the cephalic nerve mass, than bythe appearance of metamerism in the head itself. It appears most prob-able, however, for reasons to be given presently, that the protocephaloncomprises a prostomial region and two or three primitive somaticsegments. The adult arthropod brain is a syncerebrum, consistingalways of two parts, the protocerebral and deutocerebral lobes, towhich in most cases are added the ganglia of a following segment,which constitute then the tritocerebral brain lobes. The protocerebrallobes are the most complex part of the brain, and they are probablyformed of elements derived from a primitive prostomial region andfrom the ganglia of a preantennal segment. The deutocerebral lobesare simple developments of the ganglia of the antennal segment. Thepostantennal ganglia do not always enter into the composition of thedefinitive brain, and their segment is often not a part of the proto-cephalic head of the embryo, as indicated by the position of its appen-dages (fig. 8 D, Put, fig. 22 B, C, Ch ; D, Pnt).The segmental position of the mouth has been the subject of muchdifference of opinion. Most writers hold that the stomodeal invagina-tion is situated in or Ijefore the first true head segment ; others claimthat it lies l)eliind the second, or even the third segment (Comstockand Kochi, 1902; Holmgren, 1909, 1916; Henriksen, 1926). It waslong ago pointed out by Lankester (1881) and by Goodrich (1898),however, that only on the assumption that all the true head segmentsof arthropods are pastoral in position can the arthropod head seg-mentation be homologized with the anterior body segmentation of theannelids. Whatever part of the head is truly preoral, according tothis view, belongs to the prostomium. Moreover, Lankester argued,
NO. 3 INSECT HEAD SNODGRASS 21
the arthropod brain must contain a median anterior rudiment derivedfrom the prostomial ganglionic mass, or archicerebrum, in addition tothe ganglia of the component segments. " In the Chaetopoda," Lan-kester says, " the prae-oesophageal ganglion appears always to remaina pure archicerebrum. But in the Crustacea (and possil)ly all otherArthropoda * * * ) the prae-oesophageal ganglion is a syn-cerebrumconsisting of the archicerebrum and of the ganglion masses appropri-ate to the first and second pair of appendages which were originallypostoral, but which have assumed a praeoral position whilst carryingtheir ganglionic masses up to the archicerebrum to fuse with it."According to Lankester's view, then, the arthropod head shouldcomprise a prostomial region and several postoral segments, and thebrain correspondingly should include the prostomial archicerebrumClp ^ PrntLm
IIIGr Ant
~MdA ^-^ B iMxFu;. 14.—Young embryos of a chilopod and an insect showing rudiments ofpreantennal appendages.A, anterior end of embryo of Scolopcmira (from Heymons, 1901). B, sameof a phasmid, Carausiiis niorostis (from W'iesmann, 192(1).Ant. antenna; Clp, clypeus ; Hphy, hypopharynx ; IIIGng, tritocerebral gan-glion ; Lm, labrum ; Md, mandible ; iMx, first maxilla ; 2Mx, second maxilla
;
Pnit, preantenna.and the paired postoral ganglia of the first two segments, with theganglia of the third segment added in most cases. This idea, expressedtheoretically by Lankester and by Goodrich, has been given substantialsupport by Heymons in his study of the development of Scolopendra,and more recently by Wiesmann from a study of the embryo ofCarausius.The head of Scolopendra, Heymons (1901) says, is formed duringembryonic development by the union of an unsegmented preoralregion and six postoral segments. The preoral part Heymons callsthe " acron," taking this term from Janet ( 1899) in a slightly alteredsense ; it is the primary " Kopfstiick," which clearly is the pro-stomium. The first true metamere, or postoral segment, is marked bya pair of small coelomic sacs in the mesoderm, and bears a pair ofevanescent preantennal appendages (fig. 14, Prnt), which at an early
22 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
stage lie on a line posterior to the mouth. Later, this segment losesits identity, and it can not be traced in the composition of the adulthead. The second metamere is the antennal segment, bearing theantennae of chilopods (fig. 14 A, Ant) and insects (fig. 13 A, Ant),or the corresponding first antennae (antennules) of Crustacea (fig.22 A, IAnt) . The third metamere is the so-called intercalary segment,marked by a pair of coelomic sacs and corresponding ganglia inScolopcndrd (fig. 14 A, IIIGng), bearing the highly developed secondantennae of Crustacea (fig. 22 A, 2Ant) , the chelicerae of Arachnida(fig. 22, B, C, C/i), or the rudimentary post-antennal appendages ofinsects (fig. 22 'D,Pnt). The fourth, fifth, and sixth metameres ofthe definitive chilopod head are the segments of the gnathal appendages(fig. 14 A, Md, iMx, 2Mx), which have united with the proto-cephalon.The adult brain of Scolopendra, Heymons finds, is a composite ofpreoral and postoral ganglionic elements. The preoral parts are de-rived from the ectoderm of the prostomial region, the postoral partsare the paired ganglia of the first three head metameres. The pro-stomial elements include a median archicerebral rudiment that becomesthe anterior part of the supraoesophageal commissure, and pairedlateral rudiments, which form the dorsal cortical plate, the frontallobes, and the optic lobes of the definitive brain. The ganglia of thefirst metamere, or preantennal segment, are a pair of small nervemasses which unite with the prostomial rudiments to form the proto-cerebral lobes of the adult brain. The ganglia of the antennal segmentconstitute the deutocerebrum ; those of the postantennal segment be-come the tritocerebral lobes. The definitive location of the preantennaland antennal ganglia anterior to the mouth is a secondary one, andtheir union before or above the stomodeum, Heymons explains, comesabout ontogenetically through the late development of the transversecommissures, which are not formed until the respective ganglia ha^'eacquired a preoral position. Wheeler (1893) had sviggested that " thearthropod protocerebrum probably represents the annelid supraoesoph-ageal ganglion, while the deuto- and tritocerebral segments, orig-inally postoral, have moved forward to join the primitive brain."This essentially is also Heymon's earlier view ( 1895) , but the existenceof a separate pair of preantennal segmental ganglia was not suspectedat that tim'e.For many years Heymons' observations on the development of thehead of Scolopendra have remained unverified. It is, therefore, ofparticular interest to find essentially the same structure now describedfor an insect. Wiesmann (1926), studying the development of aphasmid. Caransius morosns, reports that the head is composed of
NO. 3 INSECT HEAD SNODGRASS 23
a prostomial region and of six postoral metameres with paired coelomicsacs, of which the first metamere bears a pair of small, evanescentpreantennal appendages (fig. 14 B, Pnit). Wiesmann, however, claimsthat the prostomium is a segment, because he finds in its mesoblastictissue a pair of small cavities at the base of the paired rudiments of thelabrum. The prostomial region of the adult arthropod contains a partof the body lumen, but from this it does not necessarily follow that itsprimitive mesoblastic cavities are homologous with the coelomic sacsof the true mesoderm, the extent of which should be limited by thelength of the blastopore (see page 16). More likely, the mesoblast ofthe prostomium is a mesenchyme. In any case, however, it is only amatter of definition as to what we shall call a " segment."The assumption of the presence of one or more preoral segmentsin addition to the prostomium disregards the fundamental relationbetween the embryonic germ layers. As already pointed out, theposition of the mouth, or of the stomodeal invagination, marks theanterior end of the blastopore ; the extent of the endoderm, except asit expands within the body, is determined by the length of the blas-topore; the mesoderm is derived from the endoderm, and in themesoderm metamerism originates. Therefore, in a bilateral animal,it seems clear, true segments can not lie morphologically anterior tothe mouth. In the insect embryo, the anterior mesenteron rudimentactually defines the anterior limit of the mesoderm. Later formedsegmental regions or appendages that appear to be preoral must, then,have acquired this position secondarily. In the figure of a Peripatusembryo (fig. 4 D) it is clearly seen how the anterior coelomic sacsmay extend laterally before the mouth, and how correspondingappendages might come to have a preoral location topographically,though being morphologically postoral.In the insect brain, there has never been noted a distinction betweenganglionic rudiments of a preantennal segment and prostomial ele-ments in the composition of the definitive protocerebral lobes, andthe optic lobes are commonly referred to the first segment, thoughtheir independent origin is recognized. In the Crustacea, however, pre-antennal ganglia have been recorded, and Daiber (1921) says, " sinceontogeny appears to give support to the view that the optic lobes aresecondary structures, we must suppose that the segmental gangha ofthe preantennal segment have been mostly suppressed, and that remainsof them are represented in the ganglion cells of the roots of the oculo-motor nerves. The ganglion pair found in the embryo of Astacus andlacra between the ganglionic fundaments of the optic lobes and those
24 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
of the antennal ganglia, and which later fuse with the brain ganglia,are probably to be explained as the true segmental ganglia of thepreantennal appendages."It may be questioned if there is ever a true segmental separationbetween the ocular and antennal region of the insect head, since what-ever division does occur between the two parts appears relatively latein development, and is, therefore, probably of a secondary nature.Holmgren (1916), from a comparative study of the histology of thebrain of annelids and arthropods, concluded that the protocerebraland deutocerebral parts of the definitive arthropod brain are secondarysubdivisions of one primitive nerve mass, which, moreover, Holmgrenwould identify with the archicerebrum of the annelids. This con-clusion is scarcely tenable, because, interpreted literally in terms ofannelid structure, it would assign the antennae to the prostomium.and because it disregards the evidence of the postoral position of boththe preantennal and antennal rudiments in the embryo.It is usually assumed that the compound eyes of crustaceans andinsects belong to the preantennal segment, which, on this assumption,is designated the "ocular" segment. Heymons (1921), however,claims that in Scolopendva the eyes and the optic lobes are derivedfrom the ectoderm of the prostomial region. It is perhaps not neces-sary to believe that the grouped ocelli of the Chilopoda, even thecomposite " pseudo-compound " eyes of Scutigera, are related to thetrue compound eyes of crustaceans and insects, since the details ofstructure in the two cases are quite dififerent ; but it would seem lessprobable that the optic lobes of the brain should have a separateorigin in the different arthropod groups. In many of the Crustacea,the compound eyes are pedunculate, being situated on segmentedstalks having an ample musculature innervated from the protocere-brum, and this fact gives strong support to the idea that the eye-stalks are appendages of the preantennal segment. Experiments haveshown that if an eye-stalk is amputated, an antenna-like organ isoften regenerated from the stump, on which an eye is not developed.These results recall the experiments of Schmitt-Jensen (1913, 1915)who cut ofif the antennae of a phasmid (Caraiisius inorosiis) and foundthat the appendages were regenerated in a form closely resemblingthe tarsi of the thoracic legs, each, in some cases, with a pair of ter-minal claws and a pulvillus.It is difficult to evaluate these regeneration phenomena, for it seemshighly improbable that the insect antenna ever had the specializedstructure of the thoracic appendages of modern adult insects. Many
NO. 3 INSECT HEAI -SNODGRASS
writers hold that the crustacean eye-stalks are secondary outgrowths
;
and, as for their innervation from the protocerehral lobes, it mightbe claimed that the roots of the oculo-motor nerves come from a partof the protocerebrum derived from the prostomial archicerebrum. Adefinite opinion on these matters must await the results of further re-search. Since, however, in the Annelida, the prostomium is the seatof primary sensory development, and of the principal sense organs
AntNv^- MdNvLbNv
Fir,. 15.—Evolution of the insect brain as it must be conceived if it includesan archicerebral rudiment, and ;/ the compound eyes and the optic lobes arederived from the prostomial region, as claimed by Heymons.A, theoretical generalized condition in which the ganglia of the prostoinium(Pst), preantennal segment (/), antennal segment (//), and postantennalsegment (///) are yet distinct, and in which the prostomial archicerebrum (Arc)is the brain.B, the prostomium and the first three postoral segments united into a proto-cephalon ; the brain composed of protocerehral lobes (iBr) formed of thearchicerebrum (Air) and ganglia of preantennal segment (/), and of deuto-cerebral lobes (jBr) representing ganglia of antennal segment ( // ) ; ganglia ofpostantennal segment (///) distinct and connected by postoral commissure. Thiscondition retained in some lower crustaceans.C, the definitive condition in all insects : the ganglia of postantennal segment(III) are added to the brain to form the tritocerebral lobes (sBr) of thelatter; the ganglia of the gnathal segments (/F, V, VI) united in a compoundsuboesophageal ganglion (SocGnc;).Aiit, antenna; AntNv, antennal nerve; Arc, archicerebrum; iBr, protocere-brum; 2Br, deutocerebrum ; sBr. tritocerebrum ; 3C0111. tritocerebral com-missure ; E, compound eye ; LbNv, labial nerve ; Md. mandible ; MdNv, mandib-ular nerve ; iMx, first maxilla ; 2Mx, second maxilla ; MxNv, maxillary nerve
;
O, ocellus, OpL, optic lobe; Put, postantennal appendage; Pre, protocephalonPst, prostomium; SocGng. suboesophageal ganglion; Stom. stomodeum.(fig. 10), it is at least in harmony with the assttmed annelid ancestryof the Arthropoda to suppose that the arthropod eyes had their originon the prostomial region of the head, and that their definitive posterior,dorsal location has resulted from the backward revolution of theanterior part of the head, a transformation that actually takes placein the growth of the embryo.We may conclude, without going farther into matters of contro-versy, that the immediate ancestors of the arthropods possessed a
26 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
long, segmented body, at the anterior end of which was a specializedcephalic region, differing from the annelid head in that it comprisedboth the prostomium and the first two or three primitive body seg-ments. In this early arthropod head, or protocephalon, the prostomiumwas still an important element ; it perhaps carried the ocular organs,though tentacles were probably lacking, and it was extended dorsallyon the facial aspect of the head between the bases of the antennae
;
on its ventral part, just before the mouth, there was a median lobe,the labrum. The first true head segment was much reduced, and itsappendages were vestigial, or absent, unless they are represented in theeye-stalks of modern Crustacea. The second head segment bore theantennae, simple, jointed appendages, which acquired a preoral positionon the sides or front of the head by a secondary forward migrationof their bases. These two segments and the prostomium became in-timately fused, and in the ontogenetic development of present-dayarthropods they appear as a unified, bilobed cephalic enlargement ofthe young embryo (figs. 8, 13, 16 A, B, C, 22 C, D, Pre). The brainat this stage was a syncerebrum, consisting of the archicerebrum andoptic lobes fused with the ganglia of the preantennal and antennalsegments, the two lateral nerve masses being united above the stomo-deum (fig. 15 B). The third postoral segment was probably moreor less closely associated with the second, but, judging from embryonicevidence (fig. 8D), it did not at first form an integral part of theprotocephalon. Its ganglia (later the tritocerebral lobes of the brain)at this stage constituted the first ganglia of the ventral nerve cord.There can be no question that the arthropods are to be divided intotwo principal groups, one represented by the modern mandibulateforms, the other by those in which the appendages of the fourth seg-ment retained the more generalized structure of the pedipalps ofmodern arachnids and xiphosurans. The separation of the two groupsmust have taken place in the protocephalon stage, for, as will laterbe shown, the unity of structure in the mandibles of all the mandibu-late forms is such as to leave no doubt that the mandible is a commoninheritance from a primitive mandibulate ancestor. But, before thedefinitive gnathal segments were added to the head, it would seemthat the postantennal, or tritocerebral, appendages must have as-sumed the principal gnathal function by means of basal endites thatserved as masticatory lobes. In the xiphosurans and arachnids, theseappendages have become the chelicerae, if modern embryology isrightly interpreted ; in the crustaceans they lost their gnathal functionand were developed into the second antennae ; in the land-inhabiting
NO. 3 INSECT HEAD—SNODGRASS 2J
niyriapods and insects they have become reduced to rudiments, orto embryonic vestiges.Insects were thoroughly modern in the later part of the Carbon-iferous period, when their remains are first known from the geologicalrecords. They must have been in the course of evolution during allthe preceding extent of the Paleozoic era. Scorpions are found inthe Silurian rocks, eurypterids in the Ordovician. Crustaceans, asrepresented by trilobites and other forms, were well developed in theCambrian. The common arthropod ancestors in the protocephalicstage, long antedating the divergence of the several modern groups,must have existed, therefore, in remote ages of Pre-Cambrian time.THE DEFINITIVE ARTHROPOD HEADIn all modern arthropods, at least one pair, and usually severalpairs of the segmental appendages following the protocephalon aremodified to form organs of feeding, and they are crowded forwardtoward the mouth, those of the first pair coming to lie at the sidesof the mouth opening. These appendages become the " mouth parts "of insects, and in general they may be termed the gnathal appendages.As a consequence of the forward transposition of the gnathal appen-dages, the postoral, sternal parts of the protocephalic segments arereduced and in most cases practically obliterated, their places beingtaken by the sterna of the gnathal segments. Early in the course ofevolution, therefore, the gnathal segments themselves must have hada tendency to fuse with the protocephalon to form an enlarged headregion ; and nearly all the arthropods show in some degree the resultsof this tendency toward a more extensive cephalization of the anteriorsegments in the formation of a composite definitive head.The condensation of the anterior segments has resulted in the for-mation of a definite cephalic structure in many of the arthropodgroups. Among the Crustacea, however, there is much variation inthe composition of the head. In the decapods, the protocephalon aloneforms a distinct though immovable head piece—it is that part attachedwithin the anterior end of the carapace, overhung by the rostrum, thatbears the eyes, the antennules, the antennae, and the labrum, andwhich may be easily detached from the region covered by the cara-pace (fig. i/B). The segments of the mandibles, the maxillae, themaxillipeds, and the legs are united dorsally in the wall of the cara-pace. The jaws of the decapods, therefore, are not attached to theprimitive head, and though the protocephalon and carapace may besaid to constitute a " cephalothorax," there appears to be no reason
28 SMITHSONIAN MlSCliLLANEOUS COLLECTIONS \OL. 81
for regarding the region of the carapace formed of the gnathal seg-ments as a part of the head, since there is no evidence that the decapodhead ever included more than the protocephalon.The generaUzed malacostracan crustacean, Anaspides, also retainsthe protocephalon as an independent head piece attached within theprojecting anterior rim of the mandibular segment. The large mandib-ular segment is likewise free from the following maxillary segment,but the two maxillary segments and the first maxilliped segment are
Fig. 16.—Four stages in the development of Forficula. (From Heymons, 1895.)A, embryo differentiated into a protocephalic head, and a body. B, appendagesof gnathal segments (Md, iMx, 2M.r) well developed. C, the gnathocephalicregion (Cnc) compact, but still distinct from protocephalic region. D, proto-cephalic and gnathocephalic' regions united in the definitive head (H).Ab, abdomen; Ajii, amnion; Ant, antenna; Clio, chorion; Gnc, gnatho-cephalon ; H, definitive head ; Li, first leg ; Lin, labrum ; Md, mandible ; iMx, firstmaxilla; J>71/.r, second maxilla; Pre, protocephalon; Set, serosa; Th, thorax.fused into a composite region bearing the maxillae and the firstmaxillipeds.In most of the other Crustacea, the head either is a unified cephalicstructure consisting of the protocephalon and the three gnathal seg-ments, in some forms with one or two of the maxilliped segmentsadded, or it exhibits varying stages in the condensation of the gnathaland maxilliped segments with the protocephalon. A relatively primi-tive condition is shown by Eiihranchipus (Anostraca), in which theprotocephalon itself is a distinct and well-developed head capsule (fig.17 A, Pre) carrying the first and second antennae (lAiif, 2Ant), the
NO. 3 INSECT IIKAD SNODGRASS 29
eyes (£). and the laliruni (Lin) ; but to it is attached the tergum ofthe mandibular segment (I\ ) bearing the large, jaw-like mandibles(Md). Following the mandibular segment, comes the region of thetwo maxillar}' segments (V + VI) with the rudimentary first andsecond maxillae on its under surface. Euhranchiptis thus re|)resents astage in the evolution of the head almost equivalent to that in the em-bryonic development of insects shown in figure i6 C where the gnathalsegments {Gnc), in process of being united with the protocephalon(Pre), still constitute a distinct body region. In Limnaclia (Choncos-traca), the structure of the head is essentially as in Euhranchipus, butthe gnathal segments are more intimately united with the proto-cephalon, and the second antennae are typical biramous appendages.In Apus (Notostraca) the head is more highly evolved (fig. 17 D.E), and its lateral and posterior margins are produced into a largecephalic carapace (Cp). The protocephalon and the gnathal segmentsare imited, but their respective areas are well defined dorsally (D).The protocephalon (Pre) is set ofif from the mandibular tergum (IV)by a sinuous transverse groove (x) ; on its upj^er surface it bearsthe group of head sense organs, including the compound eyes (E)
,
and, on its lower surface (E), the antennae (Ant) and the labrum(Lin). The tergal region of the mandibular segment (D, /F^) is dis-tinctly limited posteriorly by a second suture ( \' ) on the dorsal sur-face of the carapace, behind which is a narrow area representing thedorsal wall of the two maxillary segments (['" + VI), from the pos-terior edge of which is reflected the median part of the carapace.Back of the head, and partly covered by the carapace, is the long,flexible body of forty or more segments. Here is a condition f|uitedifferent, therefore, from that of the decapods (fig. 17B. C), inwhich latter the protocephalon has retained its individuality, whilethe gnathal segments have united with those of the maxillipeds andthe ambulatory limbs to form the region of the carapace (C, Cp).In the Amphipoda and the Isopoda, the head consists of the pro-tocephalon, the three gnathal segments, and one or two of the maxilli-ped segments. In these groups, however, the head segments are fusedinto a cranium-like capsule (figs. 17 F, H, 28 A), in which little nr notrace of the original head segmentation is to be discovered. In formand general appearance, the amphipod head (fig. 17 H) often cu-riously suggests the head of an insect, but both the amphipod and theisopod cranium appears to contain at least one more segment than isknown to be included in either the insect or the myriapod head.The head in the Chilopoda (fig. 17 G), Diplopoda (K). and Hexa-poda (I), is a highly evolved cranial capsule composed of the protoce-
30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lphalon and the gnathal segments, but so thoroughly fused are all thecephalic elements that the segmental composition of the head is nolonger discernible in the head wall. The insect head is a well-standard-ized structure, which, though varying greatly with regard to form,is the same in fundamental construction throughout all the insectorders. The myriapod head, likewise, exhibits no modifications in itsbasic structure, and, from a study of the head alone, it is impossibleto judge whether the cephalic structures of the myriapods and of theinsects has had a common origin, or whether in each group the headhas been evolved along a separate line of development. Consideringthe dififerences in the head appendages, and especially in the mandibles,as will be shown later, it appears probable, however, that the myria-pods and insects are not as closely related as the form of the headmight otherwise suggest. The insect head resembles also the head ofthe amphipods and isopods, as already pointed out, but it can beshown that the evolution of the head appendages has run parallel inthe insects and the crustaceans, and here again, therefore, we mustconclude that the similarities in the head structure are only equalresults of the primary tendency toward a condensation of the gnathalsegments with the protocephalon, in consequence of the drafting ofthe appendages of these segments into the service of the mouth.Considering all the evidence, especially that which will be adducedfrom a study of the mandibles, it seems most probable that the severalprincipal arthropod groups represent independent lines of descentfrom ancestors differentiated at an early stage in the evolution of thecomposite head structure. The early development of the thorax inthe insect embryo, before the gnathal segments are added to the head(fig. 13), is evidence that the insects formed a distinct arthropodgroup long before the completion of the definitive head, unless thedifferentiation of the thorax in the young insect embryo is to be re-garded as a precocious embryonic development, comparable with theearly development of the head in the vertebrate embryo. It is scarcelynecessary, however, to postulate, as suggested by Walton (1927), aseparate origin of the insects from annelids.In the Arachnida, the protocephalon constitutes a distinct head atan early embryonic period, but, as shown in Balfour's illustration (fig.22 C, Pre), it does not include at this stage the tritocerebral segment(///) in its composition. At a later stage, however, the tritocerebralsegment and the five following segments are usually added to theprotocephalon to form a cephalothorax (fig. 17 J, CtJi). The appen-dages of the cephalothorax of an adult arachnid are the chelicerae
Pre X r/ VA'I
Fig. 17.—Head or head region of various arthropods.A. head and anterior body region of Eubranchipus vernalis (Phyllopoda,Anostraca), with gnathal segments (/F, V, VI) distinct from protocephalon.B, protocephalic head piece of Spironfocaris polaris (Decapoda) separated fromthe carapace. C, carapace of Spironfocaris polaris from which the proto-cephalon (B) has been detached. D, head and head carapace of Apus louf/i-candata (Phyllopoda, Concostraca ) , dorsal view, showing segments IV , V , VIadded to protocephalon {Pre) and forming carapace. E, ventral view of
.same. F, head of PorcclUo (Isopoda) with maxillae removed. G, head ofScntiqcra forceps (Chilopoda). H, head of Orchcstoidea californica (Amphip-oda)." I, head of Machilis (apterygote insect). J, cephalothorax and anterioralxiominal segments of a scorpion (Arachnida). K, head of Eurynrus cryfhro-pygus ( Diplopoda )
.
a, dorsal (or posterior) articulation of mandible; Ant. antenna; lAiit, firstantenna; ^Ant, second antenna; c, anterior articulation of mandible; Ch,chelicera; Cp, carapace; Cth, cephalothorax ; E, compound eye; Gch. gnatho-chilarium; II', mandibular segment; iL, first leg; Lm, labrum; Md. mandible;tMx. first maxilla; jM.v, second maxilla; iMxp, first maxilliped ; MxPlp,maxillary palpus; Pdp. pedipalp ; Pre. protocephalon; iT , first tcrgum ; V, firstmaxillary segment; /7, second maxillary segment; x, suture between proto-cephalon and mandibular segment; y, suture between mandibular and firstmaxillary segments.3 31
32 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l(Ch), or tritocerebral appendages, the pedipalps (Pdp), which are themandibular appendages of other groups, and four pairs of legs (L),which are represented by the two maxillae and the first two pairs ofmaxillipeds in the Crustacea. Antennae are lacking in all adult arach-nids, but some writers (Croneberg, 1880, Jaworowski, 1891) have re-ported the presence of antennal rudiments in the embryos of certainspecies (fig. 22 B, Ant). The comparative lack of specialization in thearachnid limbs suggests that the Arachnida are an ancient group ofarthropods having little direct relationship to other forms, except to theXiphosura and possibly to the extinct eurypterids. In the Solpugida.the cephalothoracic region is divided into an anterior cephalic partcarrying the eyes, the mouth, the chelicerae, the pedipalps, and thefirst pair of legs, and into a posterior thoracic |)art carrying the second,third, and fourth pairs of legs. The division between these two bodyparts, as compared with insects, falls between the first and secondmaxillary segments, and the parts, therefore, are in no way compar-able with the insect head and thorax. In the ticks (Ixodoidea), thehead-like structure known as the capitulum is said to bear only thechelicerae and the pedipalps. In its composition it is thus equivalentto the protocephalon with only the first gnathal segment added.Cephalization in the Arthropoda, then, apparently has progressedfrom the prostomial stage (archicephalon) to the formation of aprotocephalon, from a protocephalon to the usual definitive head, ortelocephalon, and finally to the union of head and body regions in acephalothorax. The archicephalic stage is to be inferred from theevident derivation of the arthropods from an annelid-like ancestorhaving the prostomium as the only defined head. The protocephalicstage is shown in the development of all arthropod embryos, and isretained in the decapods and related crustaceans, where the carapaceis a gnatho-thoracic structure. The telocephalic stage exhibits aprogressive evolution in phyllopods, amphipods, and isopods by theaddition of one, two and three, four or five segments to the protoce-phalon ; in insects and myriapods it has reached a standardized con-dition in which the head is composed of six segments and the pro-stomium. The cephalothoracic stage is characteristic of the Xiphosuraand Arachnida. in which the segments of all the fully developedappendages are united, and combined with the prostomium.A study of the head alone does not furnish a sufficient basis for adiscussion of the inter-relationships of the various arthropod groui)s.but it must be recognized that the facts here given, and others to bedescril)ed in this paper have an important bearing on the subject.
NO. 3 INSECT HEAD ^SNODGRASS 33
and that their signiticance has not been fully taken into account bythose who have formulated theories of arthropod relationships anddescent.
II. GENERAL STRUCTURE OF THE INSECT HEADThe almost complete suppression of the primitive intersegmentallines in the insect head makes a study of the head segmentation ininsects a difficult matter, and investigators differ widely in theirviews as to the parts of the adult head that have been derived fromthe several head segments. Since the prostomial region and the threesegments of the protocephalon are never distinct, even in the earliestembryonic stages, it seems fruitless to speculate as to what areas ofthe adult cranium are to be attributed to them individually, but thegeneral protocephalic region must be at least the region of the clypeusand frons, the compound eyes, and the antennae. In as much as themuscles of the three pairs of gnathal appendages have their origins inthe posterior parts of the head, it is reasonable to assume that theareas upon which these muscles arise represent the walls of the gnathalsegments that have been added to the protocephalon.According to Heymons (1895), who bases his conclusions on astudy of the embryonic development of the head in Periplaneta andAnisolahis, the entire cranium except the frons and the region ofthe compound eyes and the antennae is formed from the walls of themandibular, maxillary, and labial segments. Janet ( 1899) , taking theattachments of the muscles of the appendages on the head walls ascriteria of the respective segmental limits, maps the cranium intoareas that closely correspond with the segmental regions claimed byHeymons. From Riley (1904), on the other hand, we get a quitedifferent conception of the definitive head structure. According toRiley's account of the development of the head of Blafta, the greatcephalic lobes of the embryo form most of the adult head capsule.The dorsal and lateral walls of the gnathal segments, Riley says, areso reduced by the posterior growth of the cephalic lobes that littleremains of them in the adult head—only the extreme posterior andpostero-lateral parts of the cranial walls, and the postoral ventral regionbeing refera1)le to them. This view must assume that the muscles ofthe gnathal segments have moved forward to the protocephalic regionas their own segments became reduced, and it would nullify the evi-dence of head segmentation based on muscle attachments. The writeris inclined to agree with Heymons and Janet that the muscle attach-ments on the lateral and dorsal walls of the cranium should be prettycloselv indicative of the limits of the gnathal terga in the composition
34 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
of the head, but it must be admitted that muscle bases can undergorather extensive migrations. That the gnathal segments contribute aconsideral)le part to the cranial walls of the definitive insect head isclearly suggested by Heymons' figures of the development of Forficula(fig. i6), and, as already shown, there can be no doubt that thesesegments enter bodily into the head composition of Crustacea that havea well-defined composite head. With insects, it is a question of thedegree of reduction that the gnathal segments have suffered after theirunion with the protocephalon.By whatever phylogenetic course the cephalic region of the insectbody has arrived at its definitive state, it acquired long ago a cranium-like form, and a definite structure that has since been modified only insuperficial characters, adaptive to diflferent modes of living and todififerent ways of feeding in the various groups of modern insects.THE HEAD CAPSULEThe chitinous walls of the definitive head capsule constitute theepicranitiiii. In an adult insect head preserving the typical embryonicposition, with the facial aspect directed forward (fig. i8B), themouth parts are suspended from the ventro-lateral edges of the epi-cranium. A pair of compound eyes (E) typically have a lateral ordorso-lateral position, and three ocelli (O) occur between them onthe dorsal or facial area of the head (A). The antennae (Ant) varyin their location from positions just alcove the bases of the mandibles(fig. 50 A, Aut) to a more median site on the dorsal part of the face(fig. 18 A. B). The top of the head, or vertex (fig. 18 A, B, J\\-).is marked by a median coronal suture (A, cs) that turns downwardon the face and divides into the frontal sutures (fs). which divergeventrally to the anterior articulations of the mandibles (c). Thecoronal suture and the frontal sutures together constitute the epi-cranial suture. The lines of these sutures are marked internally byridges, and the coronal ridge is sometimes developed into a platesupporting muscle attachments. The median facial region betweenand below the frontal sutures is the frons (Fr). ventral to which isthe clvpeus (Clp). with the iabruni (Liu) suspended from the lowermargin of the latter.The posterior surface of the epicranium (fig. 18 C) is occupiedby the opening (For) from the head cavity into the neck, usually alarge aperture, properly termed the foramen magnum by analogy withvertebrate anatomy, but commonly called the " occipital foramen "by entomologists. The surface of the head surrounding the foramen
NO. 3 INSECT HEAD SNODGRASS 35dorsally and laterally is the occipital area. Its anterior limit is de-fined in orthopteroid insects by the occipital suture (ocs). The occip-ital area is subdivided by a suture lying close to its posterior margin,
Fig. i8.—Generalized structure of the head of an adult pterj-gote insect,diagrammatic.A, anterior view. B, lateral. C, posterior. D, ventral.a, posterior articulation of mandible ; Ant, antenna ; as, antennal suture
;
at, anterior tentorial pit ; c, anterior articulation of mandible with craniumCIp, clypeus ; cs, coronal suture; Cv, neck (cervix); cv, cervical sclerite;E, compound eye ; c, articulation of maxilla with postgenal margin of cranium/, articulation of labium with postoccipital rim (Poc) of epicranium; For,foramen magnum ; Fr, frons ; fs, frontal suture ; g, postoccipital condyle forarticulation of first cervical sclerite with head ; Ge. gena ; Hphy, hypopharynxHS, suspensorium of hypopharynx; Lb, labium; Lin, labrum ; Md, mandible;MdC, opening in head wall where mandible removed; Mth, mouth; Mx, maxilla;MxC, opening in head wall where maxilla removed ; O, ocelli ; Oc, occiput ; os,ocular suture ; ocs, occipital suture ; Pyc, postgena ; Poc, postocciput ; pos,postoccipital suture ; pt, posterior tentorial pit ; sgs, subgenal suture ; SIO,orifice of salivary duct; Vx, vertex.here named the postoccipital suture (fig. i8 B, C, pos), which sets ofif anarrow marginal rim of the cranium, or postocciput {Poc), to whichthe neck membrane is directly attached. The postoccipital suture,though sometimes inconspicuous by reason of the reduction of the
Mt,
36 SMITHS()NIAi\ MJSCELLANEOUS COI.I.ECTIONS VOl,. 81
postoccipital rim, is the iiKjst constant suture of the cranium. Thedorsal part of the occipital area before it is termed the occiput (Oc),and the lateral ventral parts the postgenae {Pge). Rarely the occiputand the postgenae are separated, as in Mclanoplus, by a short sutureon each side.The lateral areas of the cranium, between the occipital suture andthe frontal sutures, and separated dorsally by the coronal suture, havebeen appropriately termed by Crampton (1921) the parietals. Theparietal area behind and below the compound eye is the gena (fig.18 B, Ge), that between the eyes is the vertex. The lower marginalarea of each lateral wall of the head is commonly marked by a sub-marginal suture (fig. 18 A, B, j-f/.?), which forms an internal ridgestrengthening the ventral lateral edge of the cranium (fig. 39 A,SgR). The suture has been termed the '' mando-genal '* suture(Yuasa, 1920, MacGillivray, 1923), but, for grammatical reasons,the writer would substitute the term subgcnal suture, and call thecorresponding ridge the siibgenal ridge. The ridge is sometimes knownas the " pleurostoma." When an epistomal ridge separates the clypeusfrom the frons, it unites the anterior ends of the subgenal ridges.The true ventral wall of the head is the region between the bases ofthe mouth parts (fig. 18 D), the median area of which is producedinto the variously modified lobe known as the liypopharynx (Hphy).Anterior to the base of the hypopharynx, and immediately behind theposterior, or epipharyngeal, surface of the lal:)rum and clypeus is themoutli {Mth). The space inclosed by the labrum and the mouthparts is often called the " mouth cavity," but, since it lies entirelyoutside the body, it is more properly a preoral cavitv.The frons, clypeus, and labrum belong to the prostomial region ofthe head. The frons and clypeus are not always distinct, but when theyare separated, the dividing fronto-clypeal groove, or epistomal suture(fig. 18 A, B, c.?), extends typically between the bases of the man-dibles. That the more primitive division of the prostomium, however,is that between the labrum and the clypeal area is evidenced by thefact that the labral retractor muscles always extend from the base ofthe laljrum to the frontal area (fig. 19, 2, 5). The clypeus, on theother hand, can not be regarded as a mere articular region betweenthe labrum and the frons, secondarily developed into a chitinous plate,as some writers have suggested, because the most anterior of thedilator muscles of the stomodeum have their origins upon its innersurface (fig. 41, jj, 34). The external suture separating the clypeusfrom the frons appears to be incidental to the development of aninternal epistomal ridge (fig. 39 A, B, C, iii?) forming a brace be-
NO. 3 INSECT HEAD- SNODGUASS 37tweeu the anterior articulations of the mandibles. The typical positionof the fronto-clypeal suture is on a line l)et\veen the mandibular basespassing- through the roots of the anterior arms of the tentorium; butthe suture and its ridge are often arched upward, as in the Hymen-optera, Psocidae, and Homoi^tera (fig. 46 E, F, G, H), or bentdorsally in an acute angle, as in the caterpillars (fig. 50 A). Thefronto-clypeal suture is to be identified by the origin of the anteriorarms of the tentorium from its internal ridge ; the frontal regionabove it is marked by the attachments of the labral retractor muscles,and the clypeal region below is distinguished by the origins of the firstanterior stomodeal muscles on its inner surface. The value of thesecharacters wall be illustrated in succeeding parts of this paper. Theclypeus may be secondarily divided into an anteclypcus and a postcly-peus, the latter sometimes attaining a special development, as inHomoptera.If the prostomial region of the adult head embraces only the labrum,clypeus, and frons, the frontal sutures must separate the prostomialarea from the area derived from the segmental elements of the head,as maintained by Riley (1904) ; but, if the compound eyes and theoptic lobes of the brain had also a prostomial origin, as claimed byHeymons (1895, 1901), then an area between and including the com-pound eyes must be regarded as a part of the general prostomialregion. Following Heymons' interpretation, Berlese (1909) recog-nizes a " postfrons " embracing the ocular region, and a " pref rons,"which is the ordinary frontal sclerite. Whatever the facts of the casemay be, it will be most convenient to retain the name " frons " forthe latter sclerite. In general, the frontal sutures mark the lines ofcleavage in the facial cuticula at the time of a molt, but there areexceptions to this rule, for the cuticular splits, when extended fromthe end of the coronal suture, may diverge to the sides of the frons,and may even extend laterad of the bases of the antennae, as inOdonata (fig. 46 I).The frontal sutures are often obscured or are lacking, and the fronsthen becomes confluent with the lateral epicranial walls. The anteriormedian ocellus, when present, is located upon the frons, or on thefrontal region ; the paired ocelli usually lie above or posterior to theupper ends of the frontal sutures, though in some cases they appear tobe in the sutures. The antennae are usually situated on the facialaspect of the head, but they never truly arise upon the frons. Inpost-embryonic stages, the antennae occupy positions varying frompoints just above the mandibles, as in caterpillars, to points laterad ofthe upper end of the frons ; they sometimes lie against the frontal
JI
38 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81
sutures, and by an approximation of their bases, they may constrictthe frons between them. The reversed relative position of the antennaeand the compound eyes, as between embryonic and adult stages, comesabout through the posterior revolution of the ocular region and theforward migration of the antennae. The antennal socket is generallystrengthened by an internal circular ridge on the cranial wall sur-rounding it (fig. 39 A, AR), and the compound eye is likewise en-circled by an inflection of the cuticula close to its base (OR). Theseridges and their external sutures set off the so-called ocular andantennal sclerites (fig. iSA, B).The posterior, or occipital, surface of the epicranium (fig. i8 C)is usually but a narrow area surrounding the foramen magnum (For)dorsally and laterally, the foramen Iieing normally completed ven-trally by the base of the labium (Lb), or by the neck membrane inwhich the labium is suspended. When the foramen is small, however,the occipital area often becomes a wide transverse surface on the backof the head, and its ventral, or postgenal, parts may form medianprocesses that sometimes unite into a bridge beneath the foramen, inwhich case the latter becomes entirely surrounded by chitinous walls(fig. 48 B, C). The occipital suture (fig. i8 B, C, ocs), when present,is generally located a])out where the dorsal and lateral areas of thehead wall are reflected upon the posterior surface. It does not seemprobable that the occipital suture is a primitive intersegmental line ofthe head, for, though it lies approximately between the mandibularand maxillary regions, it does not consistently separate the bases ofthe mandibular and maxillary muscles, and the posterior articulationof the mandible is with the postgena posterior to the lower end of thesuture (fig. i8 B, a). As is the case with most of the skeletal grooves,it is pro])able that the occipital suture has no significance in itself, andthat it is merely incidental to its corresponding internal ridge, whichstrengthens the posterior part of the cranium along the line wherethe dorsal and lateral areas are reflected into the posterior surface.In the Machilidae the posterior part of the epicranium is crossed bya prominent suture lying close behind the eyes dorsally (fig. 17 I, 3')and extending" downward on each side of the head to a point onthe lateral margin of the cranium between the base of the mandible(Md) and the base of the maxilla (Mx). This suture, therefore, ap-pears to separate the region of the mandibular segment from that ofthe maxillary segment in the cranial wall, and if it does so, it may bethe homologue of the mandibulo-maxillary suture in the phyllopodcrustaceans (fig. 17 A, D, y), and of the corresppnding suture in themore generalized malacostracan forms, such as Anaspides. Crampton
NO. 3 INSECT HEAD SNODGRASS 39(1928a) has called the mandibulo-maxillary suture the " archice-phalic " suture, since he calls the region before it the " archicephalon,"but the term thus applied denotes too much antiquity for a stage thatis clearly subsequent to several others in the head evolution. A simi-larly-placed suture is present in the head of Japyx (fig. 30 B, PcR),but the relation of the suture here to the bases of the head appendagescan not be determined. The occipital suture of the pterygote insecthead, ending laterally before the posterior mandibular articulations,therefore, is probably not the mandibulo-maxillary suture of thesimpler crustaceans, or the homologue of the posterior suture in thehead of Machilis.The postoccipital suture (fig. i8B,C, pos) is a most importantlandmark of the head because it is invariably present, and because ofits constant anatomical relations to other parts. The posterior tentorialpits (pt) are always located in its lower ends, and if the pits migratein position, as in some of the Coleoptera and other insects, the lowerends of the suture are correspondingly lengthened (fig. 49 C, pt, pt)
.
Frequently the suture is inconspicuous by reason of its closeness tothe margin of the cranium, and for this reason, probably, it has notbeen given sufficient attention by entomologists. Comstock and Kochi(1902) believed that the suture is the groove between the pleuritesof the maxillary segment; but Riley (1904) claimed, from a study ofthe developing head of Blatta, that the suture is the intersegmentalgroove between the maxillary and the labial segments, and that thepostoccipital sclerite is a remnant of the wall of the labial segment,which segment is otherwise obliterated or represented in the anteriorpart of the neck membrane. This view is at least in harmony withcertain anatomical relations in the adult head, and is tentativelyadopted in this paper.Internally, the postoccipital suture forms a postoccipital ridge (fig.39 A, PoR) just within the foramen magnum, and upon this ridge areattached the anterior ends of the dorsal muscles of the prothorax(figs. 45 A, 57A, B, C). The ridge, therefore, must be a primaryintersegmental fold corresponding with the ridges or phragmata sup-porting the longitudinal muscles in the thorax and abdomen. If itdoes not represent the fold between the maxillary and labial segments,it should be that between the labial segment and the prothorax. Ifthe first possibility is true, as claimed by Riley, there is an interseg-mental line lost somewhere in the neck, and the muscles going fromthe first phragma of the thorax to the postoccipital ridge of the headmust be regarded as extending through the region of two primarysegments. If, on the other hand, the posterior ridge of the head is the
40 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lintersegmental fold between the labial and the prothoracic segments,the muscles in the neck are all muscles of the prothorax, and the neckitself is prothoracic. It is evident that nuicli mori)hological significancehinges on this problem. The neck sclerites, for example, in the firstcase, might belong either to the labial segment or to the prothorax,or to both ; in the second case, they could pertain only to the prothorax.The relation of the posterior arms of the tentorium to the postoccipitalsuture and ridge will be noted under the special description of thetentorium (page 50).The labrum, the appendages of the gnathal segments, and thehypopharynx constitute the mouth parts of insects. The gnathalappendages are the mandibles, the first maxillae, and the secondmaxillae, which last are united in insects to form a labium. Themorphology of these appendages will be discussed in a later section(pages 79-90), but it is important here to understand their relationsto the cranial wall. The mandible in biting pterygote insects is typi-cally suspended from the lower edge of the gena and postgena, andswings outward and inward on a longitudinal axis between anteriorand posterior articulations with the head wall. The anterior articu-lation is with a condyle at the contingent angles of the gena andclypeus (fig. 18 A, B, D, c), the posterior with a shallow facet on thelower margin of the postgena (B, D, a).The maxilla hangs from the lower edge of the postgena, uponwhich it is freely movable by a single articular point just before thelower end of the postoccipital suture (fig. 18B, D, g). The labium,in generalized insects, is suspended from the neck membrane, but eachlateral angle of its transverse base is closely attached to the postoc-cipital rim of the head (B, C, D, /). The positions of the maxillaryand labial articulations relative to the postoccipital suture (pos) arein harmony with the idea that this suture is the intersegmental groovebetween the maxillary and the labial segments. In some insects, thelabium is shifted forward between the ventral edges of the postgenae,and thus becomes removed from its primitive position. In such cases,as in caterpillars (fig. 53 A) and adult Hymenoptera (fig. 48 B, C),the ventral angles of the postgenae may approach each other medially,or even unite into a ventral bridge (hypostoma) behind the labium.In other insects, in which the posterior part of the head is lengthened,the base of the labium is elongated between the postgenae, forming the]>late known as the gula. These modifications, however, will be dis-cussed more fully in section VI of this paper.The head is attached to the thorax by a cylindrical, membranousneck, or cervix (fig. 18 B. Cv). In each lateral wall of the neck there
NO. 3 INSECT HEM) SNODGKASS 4I
is typically a pair of lateral neck plates, or cervical sclerites, hingedto each other. The first is articulated anteriorly to a small process,the odontoidea (Yuasa, 1920), or the occipital condyle (Crampton,1921), on the rear margin of the postoccipital rim of the head(B, C, g) just above the base of the labium. The posterior neck platearticulates with the anterior margin of the prothoracic episternum.Other cervical sclerites of less constant form are sometimes present inthe ventral wall of the neck, and occasionally there are chitinizationsalso in the dorsal wall. The lateral neck sclerites are important ele-ments in the mechanism for moving the head on the thorax. Uponthem are inserted muscles from the postoccipital ridge of the head,and from the inner surface of the prothoracic tergum (fig. 45 A, B).The uncertainty of the morphology of the insect neck, and con-sequently of the neck skeleton, furnishes a problem still to be solved.As already pointed out, the status of the neck and of its scleriteswill depend upon that of the postoccipital rim of the head : if thelatter is an anterior remnant of the labial segment, the neck scleritesmay belong to the labial segment, or also to the prothorax ; if , however,the postoccipital ridge of the head, upon which the anterior ends of thedorsal prothoracic muscles are attached, is the infolding between thehead and the prothorax, then the neck can only be a part of the pro-thorax. The second assumption looks improbable in view of theposition of the labial articulations in generalized insects (fig. 18 B/).THE LABRUM AND EPIPHARYNXThe labrum is a characteristic feature of the arthropod head, andprobably corresponds with the tip of the annelid prostomium. In theembryo (figs. 8 D, 13, 22 A, D, Lrii), it appears at an early stage asa median ventral lobe of the prostomial region, lying just before thepoint where the stomodeal invagination will be formed. In the maturehead the mouth opening (figs. 18 D, ig, Mth) is immediately behindthe base of the labrum (Lm), and the posterior, or epipharyngeal,surface of the latter is continued directly into the dorsal wall of thepharynx (fig. 19, Phy). The adult labrum takes on various formsin different insects, but it is typically a broad flap freely suspendedfrom the lower edge of the clypeus (fig. 18 A, Lin). When movable,the labrum is provided with muscles inserted on its base, having theirorigin on the inner surface of the frons. Typically, there are twopairs of these muscles, one pair (fig. 19, 2) inserted anteriorly onthe labral base, the other (j) posteriorly on the chitinous bars of theinner face of the labrum known as the tormae (figs. 37 B, 42 A).
42 SMITHSONIAN MISCELLANEOUS COLLECTIONS \0L. 8
1
The points of origin of the lahral muscles serve to identify the frontalsclerite, or the true frontal region when the frontal sutures are lack-ing. Frequently there is only one pair of lahral muscles (fig. 50 E,G), and when the labrum is immovable on the clypeus, both pairsare lacking. The labro-frontal muscles are to be regarded as medianmuscles of the prostomium. On the posterior surface of the labrumthere is often a median lobe, the cpipharynx (fig. 19 EpJiy), that fitsbetween the bases of the closed mandibles, and obstructs the entranceto the mouth (Mth) when the labrum is closed upon the hypopharynx.THE STOMODEUMIn the embryonic development of arthropods, the endodermal partof the alimentary canal, which becomes the true stomach, is formedwithin the body and has at first no opening to the exterior. The an-terior and the posterior ectodermal parts, or stomodeiim and procto-deum, of the definitive alimentary tube are ingrowths of the ectodermat the two extremities of the blastopore. Their inner ends abutagainst the ends of the endodermal sac, and their final union withthe latter takes place by an absorption of the adjacent walls. In someinsects the proctodeum does not open into the ventriculus until theend of the larval stage.If the ontogenetic development of the alimentary canal is to betranslated literally into phylogenetic evolution, we should have tobelieve that the arthropod stomach was once a closed sac, and that thestomodeum and proctodeum are secondary means of communicationwith it. But, if insects have had a continuous line of free-livingancestors, this seems unlikely, and it is more probable that, in theiractual history, the stomodeum and the proctodeum have been formedas open invagination of the primitive circumoral and circumanalregions, and that the discontinuous development of the three parts ofthe alimentary canal in ontogeny is an adaptation to embryonic orlarval conditions.It has been proposed by Janet (1899, 1911) that the stomodeumconsists of the walls of three primitive segments that once formedthe true anterior end of the body, but which have been inverted, as theprimitive mouth, now the orifice from the stomodeum into the stomach,was retracted. This theory would give a plausible explanation of thepresence of the stomodeal ganglia, but it must assume that theseganglia have been formed from paired ventral rudiments which havemigrated dorsally and fused on the upper surface of the stomodeum.The known origin of these ganglia from the epithelium of the dorsal
NO. 3 INSECT HEAD—SNODGRASS 43
wall of the stomodeum, however, is direct evidence that they do notbelong to the system of the ventral nerve cord.The stomodeum (fig. 19) is usually differentiated into several partsin the mature insect, which may include a buccal cavity (BuC), apharynx (Pliy), an oesophagus (OE), a crop {Cr), and a proven-trkulus {Pvent). The entire length of the tube, except the extremeanterior end, is surrounded by circular and longitudinal muscles. Ingeneral the circular muscles form an external layer, the longitudinalsan internal layer, but the arrangement and relative development ofVx Oc Poc Cv ,T, Ao Cr Pvent \ent-FrGne -
PrC Epby HphyFig. 19.—The stomodeum of an insect, and its relation to associated organsin the head, diagrammatic.Ao, aorta; Br, brain; BnC, buccal cavity; Clp, clypeus ; Cr, crop; Ephy,epipharynx ; es, epistomal suture; Fr, frons; FrGng, frontal ganglion; Hphy,hypopharynx; Lb, labium; Liii, labrum ; LNv, lateral stomodeal nerve; Mth,mouth ; Oc, occiput ; CE, oesophagus ; Pliy, pharynx ; Poc, Postocciput ; PoR,postoccipital ridge ; PrC, preoral cavity ; Pvent, proventriculus ; SID, salivaryduct ; SIO, orifice of salivary duct ; Ti, tergum of prothorax ; Tent, body oftentorium ; Vent, ventriculus ; J '.1% vertex ; 2, anterior labral muscle ; j, posteriorlabral muscle ; 3S, retractor muscle of the mouth angle.the two layers varies much in different insects, as will be illustratedin the grasshopper and the caterpillar (pages i [5 and 145). The buccalcavity, the pharynx, the oesophagus, and the crop are provided withdilator, or '' suspensory " muscles arising on the walls of the head,on the tentorium, and on the walls of the thorax (figs. 41, 44, 55)-The parts of the stomodeum can not be concisely defined, becausethey are functional adaptations of structure varying in different in-sects, rather than strictly morphological regions of the stomodeal tube.The buccal cavity is the anterior, or ventral, end of the stomodeum, in-cluding the region of the mouth opening (fig. 19, BuC). The dilatormuscles of the stomodeum that have their insertion on the dorsal wall
44 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
of the buccal cavity arise upon the clypeus, and this relation betweenthe region of the buccal cavity and the clypeus appears to be a constantone. In the cicada, the sucking ]5ump is a mouth structure quitedistinct from the true pharynx, and the origin of its dilator musclesupon the large striated facial sclerite of the head wall helps to identifythis plate as the clypeus (fig. 46 H, Clp). In many insects, however,there is no structural distinction between the region of the buccalcavity and that of the pharynx. The retractor muscles of the mouthangles (fig. 19, ^8) have their origin on the inner surface of the frons,and their points of attachment give another character, in addition tothat furnished by the labral muscles, for the determination of thefrons when the limits of this sclerite are obscured, or the identity ofthe plate otherwise doubtful. The mouth retractors are inserted uponchitinous processes that extend into the stomodeal walls at the mouthangles from the suspensorial rods of the hypopharynx (fig. 42 B, y).Usually these processes are short and inconspicuous, but in the beesthey form long arms united at their bases in a chitinous plate in thefloor of the buccal cavity.The region of the pharynx is usually marked by a dilation of thestomodeum, and sometimes it forms an abrupt enlargement of thetube. The frontal ganglion is situated on its dorsal wall (fig. 19,FrGiig), and the circumoesophageal connectives lie at its sides. Thedorsal dilator muscles of the pharynx have their origin on the frons, onthe parietals, on the dorsal arms of the tentorium, and rarely one or twol^airs may encroach on the area of the clypeus (caterpillars). Thepharynx of the Orthoptera is divided into an " anterior pharynx " anda "posterior pharynx" (Eidmann, 1925), but the part called theposterior pharynx, the dorsal dilator muscles of which arise on theposterior dorsal walls of the head, appears to correspond with theoesophageal region in some other insects.The oesophagus, when there is a distinct oesophageal region, is anarrow tubular part of the stomodeum following the pharynx (fig.19, OE). and varies much in length in dififerent insects. Its posteriorend enlarges into the crop (Cr), or the crop is sometimes a lateraldiverticuhun. The terminal part of the stomodeum in biting insectsis usually a well-defined proventriculus (Pveiit). The chitinousintima of all parts of the stomodeum may be provided with short hairs,spicules, or chitinous nodules, but the inner cuticular structures arebest developed in the proventriculus. where they generally have theform of longitudinal ridges or ])lates, with deep grooves betweenthem.
NO. 3 INSECT HEAD SNODGRASS 45According to the views of the earlier students of the digestive organsof insects, the proventriculus constituted a gizzard ; its inner chitinousfold, and its sheath of strong muscle fibers, it was pointed out, mustserve to break up the larger or harder pieces of the food material notsufficiently crushed by the jaws. Experimental evidence of thisfunction, however, is lacking, and Plateau (1874, 1876) argued thatthe proventriculus is merely an apparatus for passing the food fromthe crop into the stomach. More recently, Ramme (1913) has shownthat the proventriculus, in Orthoptera and Coleoptera at least, hasanother important function in that the furrows between its chitinousridges serve to conduct the digestive secretions of the ventriculus intothe crop, where they attack the food material in advance of its en-trance into the stomach. The channels between the proventricularfolds, then, rather than the folds themselves, are to be regarded ashaving the primary functional importance. Otherwise, the proven-triculus serves to conduct the food mass into the ventriculus. InDytiscus, according to Ramme, the armature of the proventriculusretains the indigestible ]iarts of the food, which are later ejectedfrom the mouth ; but in Orthoptera all the food matter passes throughthe alimentary canal. THE HYPOPHARYNXWhen the gnathal segments are added to the protoce])halun duringembryonic growth, their sternal parts lose their identities in the gen-eral postoral ventral wall of the definitive head. On this region thereis developed a median lobe between the bases of the mouth partsknown as the hypopharyux (fig. 18 D, Hphy). The name is poorlychosen, because the organ in question lies on an exterior surface of thehead entirely outside the pharynx, but it is a heritage of earlier daysin entomology and is now w^ell established in entomological ter-minology.There is a diilerence of opinion among embryologists as to howmany of the gnathal sterna contribute to the formation of the hypo-pharynx. According to Heymons (1901), the hypopharynx is formedin insects on the sternal region of the mandibular and first maxillarysegments, but in the chilopods it arises on the mandibular segmentalone. The fusion of the bases of the second maxillae in insects, andthe similar union of both pairs of maxillary appendages in the chilo-pods gives reasons for this view, but, as will be shown ]M-esently, theprimitive adductor muscles of all the gnathal appendages have theirorigin on the hypopharyngeal region in the chilopods and in theaptervgote insects—a condition which indicates that at least some part
46 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
of each gnathal sternum enters into the hypopharyngeal region. Riley(1904) descrilies the hypopharynx of Blatclla germanica as formedin the embryo from the sterna of the mandibular, first maxillary, andsecond maxillary segments.In the more generalized pterygote insects, the hypopharynx hangslike a tongue in the preoral cavity (fig. 19, Hphy) behind the mouth(Mtli), shut in anteriorly by the labrum, laterally by the mandiblesand maxillae, and posteriorly by the labium. Its base generally extendsposteriorly to the labium (figs. 18 D, 19), and in the groove betweenMth
-4
Lin Slin Fig. 20.—The hypopharynx.A, three-lobed hypopharynx of an ephemerid nymph, with ventral adductormuscles of mandibles (KLh) attached to its base. B, head of embryo of Amiridamaritima (from Folsom, 1900), ventral view, showing median lingua (Lin)and paired superlinguae {Sliii) that combine to form hypopharynx of adult.C, transverse section through mandibles of embryo of Tomocerus plmnbcns(from Hoffman, 1911), showing origin of superlinguae (Slin) from innerangles of mandibles. D, hypopharynx of Microccntrum rhoinbifolmiii, ventral,with rudiments of suspensorial arms (HS) on which ventral mandibular ad-ductors (KLh) are attached.the two organs is situated the orifice of the salivary duct [SLO) . Ingeneral, therefore, the salivary orifice serves as a landmark for sep-arating the hypopharynx from the labium, or for determining thehypopharyngeal region when a specific hypopharyngeal lobe is lacking,as in the honeybee ; but the opening of the salivary duct may be at theapex of the hypopharynx, as in Homoptera, or, when the hypopharynxand labium are united, as in many insect larvae (fig. 54 A, D), it maylie at the tip of the combined labio-hypopharyngeal structure.In some insects the hypopharynx consists of a median part and oftwo lateral lobes. In such cases it usually projects forward like a lower
M
NO. 3 INSECT IIEAID SNODGRASS 47
lip beneath the mouth opening. The lateral lobes are best developedin the more generalized insects, both apterygote (fig. 21 D. Ilphy)and pterygote (fig. 20 A), and in coleopteran larvae, but possibletraces of them are to be found in many of the higher orders. The oc-currence of the hypopharyngeal lobes has been w^ell reviewed l)yCrampton (1921a) and by Evans (i92i),and those of lepidopteranlarvae have been described by de Gryse (1915). The median lobe ofthe hypopharynx is best distinguished as the lingua, though somewriters call it the " glossa " ; the lateral lobes have been termed " para-glossae " and " maxillulae," but Folsom (1900) has given them themore distinctive name of superlinguae, because the lateral lobes of thelabium are commonly known as the paraglossae.The nature of the superlingual lobes of the hypopharynx has beenmuch discussed. Hansen (1893) proposed that they represent thefirst maxillae, or maxillulae, of Crustacea, and Folsom (1900) be-lieved that their identity as such was established in the discovery ofwhat he regarded as a corresponding pair of ganglia in the embryonichead of Anurida. Crampton (1921a), on the other hand, argued thatthe superlinguae of insects must be the homologues of the paragnathaof Crustacea, and it will be shown later in this paper that the identityin the relations of each of these organs to other structures of the headcan leave little doubt of the truth of Crampton's contention. The super-linguae, then, are not the first maxillae of Crustacea ; but if thesuperlinguae represent a segment in the insect head, the paragnathahave a like significance in the crustacean head. It now appears prob-able, however, that neither of these organs has a segmental value,since Folsom's claim of the presence of a pair of superlingual gangliahas not been verified by subsequent research, and Hofifmann (1911)appears to have demonstrated that in the collembolan, Tomocerusplumheus, the superlinguae are derived during embryonic develop-ment from the inner basal angles of the mandibles (fig. 20 C, Slin)
.
In the Chilopoda and Diplopoda there is a single median hypo-pharyngeal lobe forming a projecting lip below the mouth opening(fig. 21 A, B, C, Hphy). In the Crustacea, the paragnaths usuallylie to each side of the median line, and are associated with the firstmaxillae, but in some forms, as in Gammarus, they are united on acommon median base, forming a bilobed structure very similar to thehypopharynx of the apterygote insect Japyx (fig. 21 D).The base of the hypopharynx is supported anteriorly, in generalizedinsects, by a pair of chitinous plates or bars that extend laterally ateach side of the mouth, and form a suspensorial apparatus for the
48 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
hypopharynx (fig. i8 D, HS). The plates appear to be chitinousremnants of the mandibular sternum. They are best developed in themyriapods. In Lifliobiiis (fig. 21 A). Scolopcndra (B). and Scutigcra(C), each plate is a large, irregular sclerite ( HS) attached laterally tothe lower margin of the head wall at a point (d) before the base of themandible, and ending mesally in the side of the hypopharynx. In somechilopods a process on the anterior free part of the mandible articu-lates against the hypopharyngeal plate of the same side.Attems (1926) describes the suspensorial plates of the hypo-pharynx in the chilopods as a mandibular support (" kommandibularGeriist "), but the homologous sclerites and their apodemal processesin the diplopods he calls the " tentorium." The writer has not ob-served corresponding structures in the Crustacea. In insects thehypopharyngeal supports are variously developed, but are usuallyreduced, and often rudimentary. In Machilis (fig. 21 E, HS) theirouter ends are broadly fused with the basal angles of the clypeusiCIp) ; in Japyx ( D) the plates are reduced and united in a W-shapedsclerite in the base of the hypopharynx ; in Dissostcira (fig. 42 B, C,HS) they are slender bars extending outward to the bases of theadductor apodemes of the mandibles ; in Microcentrum (fig. 20 D,HS) they are rudimentary prongs diverging from the base of thehypopharynx. In many cases a process extends from each hypo-pharyngeal bar into the lateral walls of the mouth, where it supportsthe insertion of the retractor muscle of the mouth angle (figs. 42 B,44,^8), and may give rise to an extensive pharyngeal skeleton. Inthe bees these processes form the long rods bearing the protractormuscles of the pharynx, though the hypopharyngeal bars themselvesare lacking.In the chilopods and in the apterygote insects, an apodemal processarises from the inner end of each suspensorial plate of the hypo-pharynx, and extends posteriorly below the sides of the pharynx (fig.21 A, C, D, HA). Upon these apodemes arise the retractor musclesof the hypopharynx, the ventral dilators of the pharynx, and ventraladductors of the mandibles, the first maxillae, and the second maxillae.These muscles are all properly sternal muscles, and their origin in theChilopoda and Apterygota on the hypopharyngeal apodemes, whichare sternal apophyses of the head, attests a primitive relation in thesegroups between the muscles of the gnathal appendages and the sternalparts of their segments. In some of the Crustacea, the correspondingmuscles have their origins on a central endoskeletal structure thatarises on the sternal region of the gnathal segments behind the mouthIn many Crustacea, however, and in the Diplopoda, the ventral
NO. 3 INSECT IIEAI -SNODGRASS 49
muscles of the gnathal appendages, especially those of the mandibles,show a highly specialized condition in that they are mostly separatedfrom their sternal connections and united upon a common transverse
,-DT
Fig. 21.—The hypopharyngeal apophyses and the tentorium.A, under surface of head of Lithobius, mandibles and maxillae removed,showing suspensorial plates (HS) of hypopharynx suspended from points (d)on margins of head, and hypopharyngeal apophyses (HA) invaginated fromtheir inner ends and connected by ligamentous bridge (/) beneath pharynx.B, Head of Scolopcndra, ventral, maxillae and right half of cranium re-moved, showing attachment of mandibular adductors (KL, KL) on ligamentuniting the hypopharyngeal apodemes.C, Sciitigera forceps, ventral view of hypopharynx {Hphy), suspensorialplates (HS), their apodemes (HA) and vmiting ligament (/).D, Heterojapyx gaUardi, ventral view of right maxilla, hypopharynx (Hphy),and hypopharyngeal apodemes (HA) upon which arise muscles of the maxilla(admx), the labium (Ibmcl), and the mandibles (not shown).E, Ncsoniachilis maoriciis, posterior view of unconnected anterior and pos-terior arms of tentorium (HA, PT), part of the head wall with clypeus (CIp)and labrum (Lin), base of maxilla (Mx), and mandible (Md).F, Ephemerid nymph, ventral view of tentorium and part of left side of head,showing anterior tentorial arms (AT) arising from ventral margin of gena(Ge).ligament. In the pterygote insects the hypopharyngeal muscles, theventral dilators of the pharynx, and most of the fibers of the ventraladductors of the mouth part appendages arise on the endoskeletalstructure of the head known as the tentoriion. Evidentlv, then, the
50 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
tentorium must have some relationship with the hypopharyngealapophyses of the Apterygota and the Chilopoda, and with the sternalapodemes of the gnathal segments in the Crustacea. The nature ofthis relationship will be shown following the anatomical description ofthe tentorium. THE TENTORIUMThe tentorium of orthopteroid insects is a horizontal, X-shapedbrace between the lower edges of the cranial walls (fig. 39 B, Tut).It consists of a central body with a pair of divergent anterior arms(AT) and a pair of divergent posterior arms (PT) . The arms arehollow invaginations of the head wall. The roots of the anterior armsappear as external pits, in most insects lying just before the anteriorarticulations of the mandibles (fig. 18 A, B, at) in the epistomal suture,when the latter is present ; the roots of the posterior arms form de-pressions in the lower ends of the postoccipital suture (B, C, /^O-Usually there is a pair of internal processes, or dorsal arms of thetentorium (fig. 39 A, C, DT), arising centrally at the junction of theanterior arms with the body, and extending dorsally and anteriorlyto the facial wall of the head near the bases of the antennae. Some-times these arms are fused with the cuticula of the cranial wall, butgenerally they are attached only to the hypodermis, and often theirouter ends are weak and tendinous. Riley (1904) says that the dorsalarms of the tentorium of Blatta arise in the embryo as processes fromthe inner ends of the anterior arms. The tentorium undergoes manymodifications of form in different insects, according as certain partsbecome more highly developed and others reduced, but its typicalstructure is seldom obscured.In its typical form, the tentorium is a simple " tent," as its nameimplies, composed of the central plate, or body, suspended by the fourstays, or arms, from the four ventral angles of the head. Yet, mor-phologists have always been suspicious of accepting the tentorialstructure at its apparent face value. Some writers would homologizethe arms with the apophyses of the thoracic pleura, others with theapophyses of the thoracic sterna. Either disposition suggests, then,that there should be a pair of such processes for each of the headsegments. Wheeler (1889) thought that he found in the embryo ofLcptinotarsa (Doryphora) five pairs of tentorial invaginations, repre-senting each head segment but the last. Other investigators have notverified this, and most students of the development of the insect headreport the presence of only the two pairs of invaginations that formthe anterior and the posterior arms of the definitive structure.
NO. 3 INSECT HEAD SNOIX^RASS 5E
Besides bracing the walls of the cranium, the tentorium gives attach-ment to muscles of the hypopharynx, of the mandibles (in some in-sects), of the maxillae, of the labium, of the pharynx, and, whendorsal arms are present, to muscles of the antennae. Such a com-prehensive relation to the musculature of the head appendages, there-fore, furnishes ample ground for the suspicion that the tentoriumincludes in its composition more than is evident in its adult structure.Janet (1899), after making a careful analysis of the muscles arisingupon the tentorium in the head of an ant, concluded that the tentoriummust be composed at least of three pairs of processes correspondingwith the antennal, the maxillary, and the labial segments. The antennalprocesses, according to Janet's homolog}^ are the anterior arms, thelabial processes are the posterior arms ; the maxillary processes areassumed to have lost their connection with the head wall, after theirinner ends had united with those of the other processes in the formationof the central tentorial body. Janet's scheme, however, is not completewithout the assumption of mandibular elements in the tentorium, for,in some of the lower insects, certain muscles of the mandibles areattached upon the tentorium. Since these muscles were not thenknown, Janet suggested that the homotypes of the mandibular ten-torial processes are represented on the mandibular segment by thepoints where the corpora allata have their origin in the hypodermis.All the tentorial processes, both real and hypothetical, Janet regardedas homologous with the fureal invaginations of the thoracic sterna,because the tentorium of the adult insect supports the adductormuscles of the head appendages. This is sound reasoning, and theconclusion probably comes as close to the truth in the matter as thetruth may be approached by induction from the facts presented bythe higher insects ; but a study of the Apterygota, the Myriapoda, andthe Crustacea throws an entirely new light on the origin and evolutionof the tentorium, and dispels the obscurity which has led to so manytheories concerning the nature of this head structure.The morphology of the tentorium, briefly summarized from factslater to be described, is as follows : The anterior arms and the part ofthe body of the tentorium on which the ventral adductor muscles of themandibles, the maxillae, and the labium have their origin are identicalwith the hypopharyngeal apophyses of the Myriapoda and Apterygota,and have their prototypes in the ventral apodemes of the gnathal seg-ments in Crustacea. From their positions just laterad of the hypo-pharynx, the bases of the apophyses have moved outward in the ventralwall of the head before the leases of the mandibles to the lateral ventraledges of the cranium, where thev come to lie in the subgenal sutures.
52 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lThen, proceeding forward, they have migrated to the fronto-clypealsuture on the facial aspect of the head. The primitive condition isfound in Chilopoda, Diplopoda, and Apterygota; intermediate con-ditions occur in the Ephemerida and Odonata ; the final condition ischaracteristic of all Pterygota, except the Ephemerida and Odonata.The posterior tentorial arms are invaginations in the lower ends ofthe postoccipital suture of the cranium, which is probably the inter-segmental groove between the first and second maxillary segments.These arms are absent in the Myriapoda and most Apterygota ; theyare present in Machilis and in some Crustacea, where their inner endsare united to form a transverse bar through the back of the head ;they are present in all Pterygota, where the anterior arms are unitedwith them to form the typical four-branched tentorium. The dorsaltentorial arms are processes of the anterior arms and may secondarilybecome attached to the dorsal or facial wall of the cranium.The muscles of the tentorium, with the exception of the antennalmuscles usually arising on the dorsal arms in pterygote insects, areall muscles that primitively have their origin on the sterna of thegnathal segments. They include two sets of median longitudinalventral muscles, one set going anteriorly to the hypopharynx, andthe other posteriorly to the sternum or sternal processes of the pro-thorax ; they include also the transverse ventral adductors of themandibles, the first maxillae, and the second maxillae, and the ventraldilators of the phaiynx. In the Chilopoda and Apterygota, all thesemuscles arise from the hypopharyngeal apodemes, except some of themandibular muscles which may become detached from the apophyses,or retain a direct connection with the base of the hypopharynx. Thehypopharyngeal apodemes are, therefore, paired apophyses of theregion of the gnathal sterna. There is no evidence that they are com-posite structures ; each appears to be a single process invaginatedfrom a chitinous remnant of the mandibular sternum (the suspensorialplate of the hypopharynx), but since it bears the sternal muscles ofthe three gnathal appendages, either the bases of these muscles havemigrated forward, or each apophysis is a process of the three unitedsterna. When the two apophyses move to the positions on the frontwall of the head characteristic of the orthopteroid branch of thePterygota, they retain the muscle attachments, and when they unitewith the posterior arms to form the typical tentorium, the head pre-sents the aspect of having none of the ordinary sternal muscles of theappendages attached on its sternal region, except for the small mandib-ular adductors present in some of the lower Pterygota that haveretained their origin directly on the base of the hypopharynx.
NO. 3 INSECT HEAD SNODGRASS 53The antennal muscles that take their origin on the dorsal arms ofthe tentorium in most adult pterygote insects have evidently migratedsecondaril}^ to this position after the attachment of these arms to thedorsal wall of the cranium. In the crustaceans, myriapods, and manyinsect larvae, the antennal muscles have the primitive attachment onthe walls of the head capsule (figs. 23 B, 50 B, C, E, F, 5).Evidence fully supporting the above statements is easily adducedfrom a comparative study of the head structure and the gnathal mus-culature in the Myriapoda, Apterygota, Ephemerida, Odonata, andorthopteroid Pterygota. Many of the facts have been described byother writers, but their significance appears to have been unrecognized.In the Chilopoda, the hypopharyngeal apodemes are large chitinousprocesses (fig. 21 A, B, HA) that arise from the inner ends of thesuspensorial plates of the hypopharynx (HS) close to the base ofthe hypopharynx (Hphy). Each projects posteriorly at the side ofthe pharynx, and the two are bridged below the pharynx by a sheetof ligamentous tissue (A, C, ;'). Upon the arms, or on processes ofthe arms, and on the uniting ligament arise the ventral adductormuscles of the mandibles (B,KL), and of the first and secondmaxillae. The relations here are the same as in the thorax of an insectwhere the ventral leg muscles arise from a pair of sternal apophyses.In the Diplopoda, a highly specialized condition has arisen throughthe separation of the inner ends of the muscles from the apodemesand their union across the median line by a tough transverse ligament(fig. 26 A, A-). The large mandibular adductors (KL, KL) herepull against each other from the two ends of the ligament. The liga-mentous bridge suggests, in a way, the body of a tentorium, but aswill be seen it has no relation to the insect tentorium. A similar con-dition of the mandibular adductors exists in many of the Crustacea(fig. 2y A, B, KL), and in some of the fibers of these muscles in theapterygote insects (C, D, KLk), as will be descril)e(l later in con-nection with the mandibles (page 62).In the Apterygota, the hypopharyngeal apodemes are well developedand extend far back in the head. Those of Japyx (fig. 21 D, HA) areslender rods running parallel beneath the sides of the pharynx andthen diverging outward and posteriorly to the head wall behind thecardines of the maxillae (Cd), but their ends appear to be free andnot attached to the cuticula of the cranium. Upon these arms arisethe hypopharyngeal retractor muscles, a set of mandibular adductors(fig. 2'/ C, KLt), the adductors of the maxillary stipes and cardo(fig. 21 D, admx), and muscles of the labium (Ibmcl). The hypo-pharyngeal skeleton of Japyx was described first by Meinert (1867),
54 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
and later by von Stumnier-Traunfels (1891). The latter writer calledit the " Stiitzegerust,'' or supporting framework of the hypopharynx
;
he figured it in Tetrodoiitopliura gigas and in Campodca staphylinus,and he says it has essentially the same structure in Japyx, Campodea,and Collembola. Folsom (1899) described the hypopharyngeal skele-ton of the collembolan, Orchcsella cincta, as consisting of a thinmedian plate with paired anterior, dorsal, and posterior arms. Theanterior arms, he says, are united with the lateral lobes of the hypo-pharynx, the others are attached to the cranial walls by fibrous strands.This structure of the collembolan head, upon which arise musclesof the pharynx, the mandibles, and the maxillae, Folsom points outis the true tentorium, homologous with that of the Orthoptera andother mandibulate insects. The failure to recognize this fact, hesays, " has led students to assign an altogether undue importance tothe * Stutzapparat ' of the ligula (hypopharynx), which has errone-ously been regarded as a sort of substitute for a tentorium." " Partlyas a result of this error," he adds, " systematists have acquired anexaggerated opinion of the dififerences which separate Collembolaand Thysanura from insects of other orders."The tentorium of the Protura has been described by Berlese (1910)and by Prell (1913). The anterior arms of the structure are unitedin a median bar, but each arm itself is forked anteriorly, and thetwo forks are said by Prell to make connections with the base of thehypopharynx and with the fronto-clypeal ridge of the head. BothBerlese and Prell call this endoskeletal structure of the proturan headthe " tentorium," but Prell observes that it has a close resemblanceto the " Zungenapparat " of the Collembola and suggests a homologywith this structure. It is now to be seen that the two structures are,indeed, identical, and that the hypopharyngeal apophyses of theApterygota are the primary elements of the pterygote tentorium.In Machilis (figs. 21 E, 27 D), the hypopharyngeal apodemes(HA) arise from suspensorial plates (fig. 21 E, HS) connected later-ally with the cranial walls as in the chilopods, but their points of originfrom these plates are at the basal angles of the clypeus (Clp). Thereis in Machilis also a well-developed posterior tentorial bar (PT) ex-tending transversely through the back of the head from pits (pt)in the lower ends of the postoccipital suture. The maxillary cardines(Cd) are attached to the margin of the cranium just anterior to theseposterior tentorial depressions. The inner ends of the hypopharyn-geal apodemes (HA), or anterior tentorial arms, of Machilis becomeweak and fibrous, and in specimens cleaned in caustic they do notconnect with the posterior tentorial bar. The tentorium of Machilis.
NO. 3 INSECT HEAD SNOUGRASS 55therefore, appears to be in an intermediate stage of development inwhich the anterior and posterior elements are still independent ofeach other. A two-branded fiber (q) extends downward in the headfrom the middle of the posterior bar. A similar tentorial bar isstrongly developed in the crustacean, Gaimnarus (fig. 28B, PT).In all the pterygote insects the anterior and the posterior arms ofthe tentorium are united with each other, and typically the lateralelements are fused across the median line to form the central plate-like body of the tentorium (figs. 21 F, 39 B, Tut). The median plate,however, is not developed in all cases ; in the caterpillars the posteriorarms form only a slender bar through the back of the head, to whichthe anterior arms are attached on each side (fig. 53 D, Tnt), and asimilar condition exists in adults of the higher Hymenoptera, wherethe posterior bar appears as a slender yoke between the posterior endsof the large anterior arms. In all insects of the orthopteroid branchof the Pterygota, the roots of the anterior tentorial arms lie in thefronto-clypeal suture (figs. 18 A, B, 36 A, B, 46 A, B, D, F, at, at).So constantly do they have this position that they become diagnosticmarks of the suture, or of the fronto-clypeal line when a suture isabsent. In the Ephemerida (nymphs), however, the roots of thebroad anterior arms (fig. 21 F, at) lie at the edges of the inflectedventral areas of the genae (Ge), before the bases of the mandibles.Here, clearly, is a more primitive condition, dififering from that ofthe Myriapoda and Apterygota only in that the bases of the hypo-pharyngeal apodemes have moved outward from the hypopharynxto the lateral walls of the cranium. In the Odonata the roots of theanterior tentorial arms lie on the sides of the head, above the bases ofthe mandibles, in the subgenal sutures. This is a second step towardthe orthopteroid condition, in which finally the tentorial roots havemigrated anteriorly into the fronto-clypeal suture on the facial aspectof the head.The writer believes that the facts presented in the foregoing de-scriptions solve the riddle of the insect tentorium, and explain theseeming anomalies of the gnathal musculature, though he has notshown the mode of union between the anterior and the posterior armsin forming the characteristic tentorium of pterygote insects, andthough the method of the change in the connections of the anteriorarms from the base of the hypopharynx to the facial aspect of thehead may still be held as not exactly determined. The origin of theanterior tentorial arms as apophyses of the sternal region of thegnathal segments, however, shows that the adductor muscles of thegnathal appendages, which arise on the tentorium in pterygote insects.
56 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL.
are the true sternal muscles of the head appendages, and this relationbrings the musculature of these appendages directly in line with thatof the thoracic legs, which are moved by sets of muscles arising onthe tergum and the sternum in each segment. In the Pterygota, itwill further be shown, the mandibles lose their primitive sternal ad-ductors, and, by a change in the nature of the mandibular articulationwith the head, the primitive tergal promotor and remoter muscles ofthe jaw become the functional abductors and adductors.
III. THE HEz\D APPENDAGESThe segmental appendages of the head in an adult insect are theantennae, the mandibles, the maxillae, and the labium. The antennaePi
-" Lm
-- ,^ ' ^ , lAnti^-^ vi^ -2AntMthAn ALm Ant
.Ch
Fig. 22.—Arthropod embryos showing relative development of the trito-cerebral appendages.A, embryo of a crayfish, Aslacus (Potatiwhius) asiactts (from Reichen-bach, 1877). B, embryo of a spider, Trochosa singoricnsis (from Jaworowski,1891). C, embryo of a spider, Angeleua labyrinthca (from Balfour, 1880).D, embryo of an apterygote insect, Auurida maritiina (from Wheeler, 1893).An, anus; Ant, antenna; lAiit, first antenna; 2Ant, second antenna; Ch,chelicera ; ///, tritocerebral segment ; /L, first leg ; Li, prothoracic leg ; Lm.labrum ; Md, mandible ; Mth, mouth ; Pdp, pedipalp ; Pi. pit on head region
;
Pnt, postantennal appendage ; Pre. protocephalon.belong to the second, or deutocerebral, segment of the protocephalon.the other appendages to the gnathal segments. In many insect em-bryos there is present a pair of small lol^es on the third protocephalicsegment, which lobes are unquestionably rudiments of the tritocere-bral appendages. Preantennal a)3pendages have been reported in Scolo-pendra and in the phasmid insect, Caransius (fig. 14 A, B Prnt) . Asalready pointed out, there is some reason for regarding the crustaceaneye stalks as being the appendages of the preantennal segment, thoughthe true status of these organs has not yet been demonstrated.
NO. 3 INSECT HEAD SNODGKASS 57The eye stalks of the decapod crustaceans arise from the ends of atransverse ridge on the top of the protocephalon, and project later-ally from mider the base of the rostrum, the latter being a process ofthe anterior edge of the carapace, and, therefore, from the tergumof the mandibular segment. Each eye stalk (fig. 17 B) consists of twomovable segments, a narrow basal one forming a short peduncle, anda large terminal one capped by the hemispherical compound eye.Schmidt (1915) enumerates ten individual muscles for each eye stalkin the crayfish, the basal segment being provided with muscles arisingon the head walls that move the appendage as a whole, while musclesfrom the basal segment move the terminal eye-bearing segment. Theeye muscles are innervated by an oculo-motor nerve arising from thebrain near the base of the sensory optic nerve.THE ANTENNAEThe insect antenna is typically a many-jointed filament. Usuallythe first two basal segments are dififerentiated from the rest of the
BFig. 23.—The antenna.A, diagram of typical segmentation and articulation of an insect antenna.B, head of a chilopod, Scutigcra forceps, dorsal, showing dorsal articulation ofantennae, and origin of antennal muscles on walls of cranium.Ant, antenna; as, antennal suture; E, eye; 11. articular pivot of antenna;Pdc, pedicel ; Sep, scape ; FI, flagellum.
shaft (fig. 23 A). The first segment serves to attach the antenna tothe head, and, being often, thicker and longer than the others, forms abasal stalk, or scape (Sep), of the appendage. The second segment,or pedieel (Pde), is short, and in nearly all insects contains a specialsensory apparatus known as the organ of Johnston. The part of theantenna beyond the pedicel is termed the flagelliiiii or clavola (Fl).The flagellum may be long and tapering and made up of many smallsegments, or it may be abbreviated, and reduced even to a single seg-ment. The scape is set upon a small membranous area of the headwall, sometimes depressed to form a cavity, or autcnual socket.
58 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81The head wall surrounding the antenna! base is strengthened by anniternal ridge, the line of which is marked externally by a suture(fig. 23, as), setting- off a circular, marginal rim known as the antennalsclerite. Usually a pivot-like process (;/) from the rim of the scleriteforms a special support and articular point for the base of the scape,and allows the antenna a free motion in all directions. In its singlepoint of articulation with the head wall, the antenna resembles themaxilla, or the mandible of those apterygote insects in which the jawdoes not have a double hinge with the cranium. In most pterygoteinsects the antennal pivot is ventral or postero-ventral in position,relative to the base of the antenna (fig. 23 A), while the single mandib-ular or maxillary articulations are dorsal. The ventral position ofthe antennal articulation might be supposed to have shifted during theforward and upward migration of the appendage from its primitiveventral and postoral situation ; but in Japyx the antennal pivot isdorsal, as it is also in the Chilopoda (fig. 23 B, n).Each antenna is moved by muscles inserted upon the base of thescape. The origin of the antennal muscles in adult pterygote insectsis commonly on the dorsal, or dorsal and anterior arms of the ten-torium (fig. 38 D, DT, AT), but in the caterpillars (fig. 50 B, C, E, F)and in some coleopteran larvae, the antennal muscles arise uponthe walls of the epicranium. The cranial origin of the muscles is prob-ably the primitive condition, for, as already shown, the tentoriumbelongs to the gnathal segments only. The attachment of the anten-nal muscles on the tentorium, therefore, appears to be a secondarycondition that has resulted from the migration of the muscle bases tothe dorsal tentorial arms when the latter make contact with the dorsalwall of the head. In Crustacea and Chilopoda the antennal muscleshave their origin on the head wall. In Sciitigera (fig. 23 B) a dorsalset to each antenna arises on the dorsal wall of the cranium mesadand posterior to the antennal base, and a ventral set arises on the lat-eral walls below the antenna, and below the eyes. The insertion pointsof these muscles, distributed on three sides of the articular pivot {n),allow the muscles to act as levators, depressors, and rotators of theappendage. The part of the insect antenna distal to the scape is movedby muscles arising within the scape and inserted on the base of thepedicel (fig. 23 A). The segments of the flagellum in insects, however,so far as known to the writer, are never provided with muscles, andtheir lack of muscles suggests that the flagellum is a single segmentsecondarily subsegmented, corresponding with the flagellum of acrustacean antenna (fig. 24 B, Ft), which is a many-jointed dacty-lopodite. In the Myriapoda, however, all the antennal segments may
j^O. 3 INSECT HEAD SNODGRASS 59be individually provided with muscles {Scolopendra, Spiroholus).The first antenna, or antennule, of the crayfish, according to Schmidt(1915), has paired antagonistic muscles for each of its first threeproximal segments, and the third segment contains a single reductorinserted on the base of the dorsal branch of the flagelluni. but other-wise none of the flagellar segments is provided with muscles.The Arachnida and Xiphosura lack antennal appendages in theadult stage. Croneberg (1880) describes a pair of head lobes in thearachnid embryo, which he says fuse into a median rostrum in themites and in the higher arachnids, and which he believes representthe antennal appendages. Jaworoski (1891) likewise describes inthe embryo of a spider, Trochosa singoriensis, a pair of lobes situatedbefore the chelicerae, which he claims are rudiments of the antennae(fig. 22 B, Ant), but he says the lobes disappear during later develop-ment. THE POSTANTENNAL APPENDAGESThe pair of postantennal appendages on the tritocerebral segmentof the head, known also as the antennae (Crustacea) , second antennae,premandibular appendages, and intercalary appendages, are at bestrudimentary in all insects. According to Uzel (1897), two smalllobes in the adult head of Campodea, lying between the labrum andthe maxillae, in the space left free by the retracted mandibles, arethe tritocerebral appendages ; the writer has found a pair of smallpapillae in Dissosteira between the bases of the mandibles and theangles of the mouth (fig. 42 B, Put) that might be vestiges of theseorgans Otherwise tritocerebral appendages are known m msects onlyas evanescent rudiments in the embryo (fig. 22 D, Put) . In the Myna-poda likewise, the postantennal appendages are lackmg, or possiblyare present as temporary premandibular lobes on the head of theembryo ("rudiments of lower lip" in GcophUus, Zograf, 1883). Inthe Crustacea, on the other hand, the appendages of the tritocerebralsegment, though sometimes reduced or lacking, are commonly highlydeveloped, biramous organs, the second antennae, or " the antennaeaccording to the terms of carcinology. In the decapods each appen-dage consists of a two-segmented base (fig. 24 B. Prtp), of a large,one-segmented exopodite (Exp), and of a long, slender endopodite(Endp) of which the terminal segment is the many-jointed flagellum\fI). The exopodite is independently movable by abductor and ad-ductor muscles arising in the second segment of the baseIn Xiphosura and Arachnida. the chelicerae (fig. 24 A) ^^'"^ f""erally regarded as the appendages of the tritocerebral segment. Their
60 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
rudiments in the embryo of a spider (fig. 22 C, Ch) bear a relation-ship to the head so similar to that of the tritocerebral rudiments inthe insect embryo (D, Put), that the identity of the two sets of organscan scarcely be questioned. Holmgren (1916), furthermore, claimsthat the histology of the arachnid brain shows that the chelicerae areinnervated from the tritocerebral region of the brain. If this homologyis correct, there is no reason for calling the tritocerebral appendages
" second antennae " except in the Crustacea. The arachnid chelicerais a uniramous organ, that of a scorpion (fig. 24 A) having threewell-developed segments.
Fig. 24.—Postantennal appendage of adult arthropods.A. chelicera of a scorpion, left, ventral view, showing uniramous structureand three segments. B, second antenna of a decapod crustacean (Spirontocaris(jrociilaiidiciis), left, ventral view, showing biramous structure, consisting oftwo-segmented base (Prtp) bearing an exopodite (Exp) and an endopodite(Endp). THE GNATHAL APPENDAGESThere can be no doubt that the gnathal organs—the mandibles, thefirst maxillae, and the second maxillae—constitute a distinct groupof appendages in the eugnathate arthropods. The mandibles are themost highly modified of the gnathal appendages, and, in most cases,their structure has lost all resemblance to that of the more generalizedinsect maxillae. A maxillary appendage, therefore, should be studiedfirst as affording a better example of the basic structure of the gnathalorgans, and, in insects, the first maxilla preserves most nearly theprimitive structure, since the second maxillary appendages are unitedto form the labium.The first maxilla of an insect with typical biting mouth parts, ofwhich the roach oft'ers a good example (fig. 25 A), consists of a basalstalk, two terminal lobes, and a palpus. The base is divided into aproximal cardo (Cd) , suspended from the head by a single point ofarticulation (c), and a distal stipes (St). The cardo and stipes arefreely flexible on each other by a broad hinge line, and their planesmay form an abrupt angle at the union, but neither has an inner wall,
NO. 3 INSECT HEAD SNODGRASS 6i
the two being merely strongly convex sclerites set upon the mem-branous lateral wall of the head, and their cavities are a part of thegeneral head cavity. Only the terminal maxillary lobes and the palpusare free parts of the appendage. The lobes arise from the distal endof the stipes, one, the lacinia (Lc), being internal, the other, the galea(Ga) external. The galea is also anterior to the lacinia (or dorsalto it in insects with the head flattened and held horizontal). The
KLcd
Fig. 25.—Maxilla of Periplaneta.A, left maxilla, posterior (ventral) surface. B, internal surface of cardo.C, right maxilla, anterior (dorsal) view, showing muscles.Cd, cardo; e, articulation of cardo with cranium; fga, flexor of galea;Ucc, cranial flexor of lacinia ; Acs, stipital flexor of lacinia ; ft, femoro-tibialjoint of palpus; Ga, galea; /, promotor of cardo; KLcd, adductor of cardo(origin on tentorium) ; KLst. adductor of stipes (origin on tentorium) ; Lc,lacinia ; O, levator of palpus ; Plf, palpus ; ipip, first segment of palpus ; Q,depressor of palpus; q, submarginal suture (and internal ridge) near innermargin of stipes ; r, internal ridge of cardo ; St, stipes ; T, depressor of fourthsegment (tibia) of palpus; V, depressor of fifth segment (tarsus) of palpus.galea is usually a soft lobe ; the lacinia is more strongly chitinized,and ends in a strong incisor point provided with one or more apicalteeth curved inward. Both lobes are movable on the end of the stipes ;the galea can be deflexed, and the lacinia can be flexed inward. Thepalpus (Pip) arises from the lateral surface of the stipes, a shortdistance proximal to the base of the galea. The palpus of the roach isfive-segmented.The musculature of the maxilla (tig. 25 C) comprises muscles thatmove the appendage as a whole, and muscles that move the terminal
62 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
lobes and the palpus. The first group includes a tergal muscle (/)arising on the posterior dorsal wall of the head, and two sets of sternalmuscles {KLcd, KLst) arising on the tentorium in most insects, oron the homologous hypopharyngeal apodemes in some apterygoteinsects (fig. 30 B, HA). The single tergal muscle (fig. 25 C, /) isinserted on the proximal end of the cardo just before the articulationof the latter with the head (c) ; it is probably a promotor, serving toswing the appendage forward. The sternal muscles (i. e., the tentorialor hypopharyngeal muscles) consist of two large flat bundles oflibers, one group {KLcd) inserted on the inner face of the cardo, theother {KLst) on an internal ridge of the stipes near the mesal borderof the posterior face of the latter (A, q). These muscles are the ad-ductors of the maxilla ; the fibers of the cardo muscle arise anterior(or dorsal) to those of the stipes muscle and cross them obliquely.The muscles of the movable parts of the maxilla include musclesof the galea, the lacinia, and the palpus. The galea has a single muscle(fig. 25 C, fga) arising on the posterior wall of the stipes and insertedon the posterior rim of the base of the galea ; it is a reductor in asmuch as it serves to flex the galea posteriorly ( or ventrally) . Thelacinia has a large flexor {Hcs) arising in the base of the stipes, and asecond muscle {flee) arising on the posterior dorsal wall of thecranium. In the roach these two muscles are inserted by a commonbroad tendinous base on the inner proximal angle of the lacinia ; inother insects they usually have separate insertions (fig. 30 B, Hes andflee, fig. 40 B, 14, 75). The palpus is provided with two muscles(fig. 25 C,0,Q), both of which arise within the stipes and are in-serted on the base of the first segment of the palpus (A, iplp). Thetwo palpus muscles are more distinct in most other insects than in theroach (fig. 31 A, B, C, E), and since one is dorsal and the other ven-tral, relative to the morphologically vertical axis of the maxilla, theyare clearly a levator and a depressor, or abductor (O) and adductor(()), of the palpus. The muscles within tbe ]ial])us vary somewhat \v.dififerent insects. In the palpus of the roach, a levator of the secondsegment arises in the first, where also a long depressor of the fourthseginent (T) has its origin. A depressor of the terminal segment {V)arises ventrally in the penultimate segment.THE MANDIBLESThe most generalized manciil)u!ar api»eu(lage in the arthropods.i. e., one corresponding most closely in structure and musculature witha typical maxilla, is to be found, not in the insects or crustaceans, butin the myriapods, and best developed in the Diplopoda.
NO. 3 INSECT HEAD SNODGRASS 63The diplopod mandible consists of a large basal plate, which appearsto form an extensive part of the lateral head wall (fig. 17 K, Add),and of a movable terminal lobe mostly concealed in the normal con-dition by the gnathochilarium (Gch). The basal plate is subdividedinto several regions, but particularly there is a proximal piece (fig.26 A, Cd) and a distal piece (St), separated by a line of flexibility.The proximal piece is loosely articulated to the head wall by a singlepoint on its dorsal posterior angle (a). The entire mandibular baseis slightly movable by its membranous union with the head, but it isnot of the nature of a free appendicular structure, since it has no innerwall—it is merely a convex plate in the lateral wall of the head, butKL I flee flee a
.^^:^m^^ /KLij' . " y^ KL ••=*•^- fTxair-ii/n-^i-f..
flcs A B Le-Fig. 26.—Mandibles of Myriapoda.A, right mandible of a diplopod, Thyropygus (Spirostreptus), dorsal, showinglarge dumb-bell adductors (KL, KL) from opposite mandibles, united by mediantendon (k). B, left mandible of a chilopod, Scutlgera forceps, lateral view.C, right mandible of Sciitigera, dorsal, somewhat diagrammatic.a, articulation of mandible with cranium; BP, basal plate of inaudible; Cd," cardo " of mandible ; flee, cranial flexor of lacinia ; flcs, stipital flexor oflacinia ; /, promoter of mandible ; /, remotor of mandible ; k, median tendon ofmandibular adductors ; KL, mandibular adductors, united by median tendon indiplopod (A, k) to form dumb-bell muscle; Lc, lacinia; St, " stipes" of mandible.
separated from the cranium by a membranous suture. The free ter-minal lobe of the mandible is a strongly chitinized, jaw-like structurewith a proximal molar area and terminal incisor point (fig. 26 A, Lc )
.
It is hinged by a dorsal articulation at its base with the end of the basalplate.So closely do the parts of the diplopod mandible (fig. 26 A) re-semble the cardo, the stipes, and the lacinia of an insect maxilla(fig. 25 A), that the imagination at once sees in the diplopod ja\van appendage similar to the maxilla, lacking only a galea and a palpus.That the fancied resemblance is real is easily demonstrated by a studyof the musculature.The musculature of the diplopod mandible consists of muscles thatmove the appendage as a whole, and of muscles that move the lacinial5
64 SMITHSUMAN MISCELLANEOUS COLLECTIONS VOL. 8l
lobe. As in the insect maxilla, the muscles that move the entire organinclude a tergal promoter and a group of ventral adductors. Thepromoter (fig. 26 A. /) arises on the wall of the cranium dorsal andposterior to the articulation of the basal plate with the head. It isinserted on the dorsal (anterior) margin of the distal division {St)of the basal plate, and in its point of insertion alone does this musclediffer from the promoter of the insect maxilla, which is inserted enthe edge of the carde (fig. 25 C, /). Functionally, however, the twomuscles are the same, and a shift in the point of attachment is not amorphological difference.The adductor muscles of the diplopod mandible consist principallyof a great mass of fibers (fig. 26 A, KL) filling the cavity of both divi-sions of the basal plate (Cd and St). These muscles are clearly thehomelogues of the adductors of the cardo and the stipes in the insectmaxilla (fig. 25. KLcd, KLst), which have their origins on the ten-torium, or on the hypepharyngeal apodemes. In the diplopod mandi-bles, however, the fibers of the adductor muscles converge mediallyfrom each jaw upon a large, tough, transverse ligament (fig. 26 A, k),and the two conical fiber masses, together with the connecting liga-ment, form a great dumb-liell-shaped muscle uniting the two man-dibles. The two sets of fibers pull against each other to close the jaws.Clearly, the inner ends of these muscles have become detached fromthe hypepharyngeal apodemes. and the fibers from opposite sides havebeen united across the middle of the head by means of a transverseligament. There is also, however, a small group of adductor fibers toeach mandible (not seen in the figure) that still retains a connectionwith the corresponding apodeme of the hypepharynx. Besides themandibular muscles, other muscles have their origin on the transverseligament, including muscles to the gnathochilarium, which is either theunited second maxillae, or the combined first and second maxillaryappendages. In the Diplopoda, therefore, the ventral adductors of allthe gnathal ap]')endages have lost their sternal connections by their de-tachment from the hypepharyngeal apodemes. This is a specializedcondition, and the ligamentous bridge en which the muscles arise hasno relation to the insect tentorium.The muscles of the free terminal lobe of the diplopod mandible(fig. 26 A./,r) include a muscle inserted directl}- on the base of thelobe (fics) arising within the stipes (St), and a large cranial muscle(Hcc) arising on the dorsal wall of the head and inserted by a strong,chitineus apodeme on the inner basal angle of the lobe. These musclescorresijond exactly with the lacinial flexors of the insect maxilla, oneof which (fig. 25 C. tJcs) arises wnthin the stipes, the ether (tlcc) en the
NO. 3 INSECT HEAL) SNODGKASS 65dorsal wall of the cranium. In most insects the second muscle is in-serted, as in the diplopod, on a chitinous apodeme from the inner angleof the lacinia (fig. 30 B, flee) . There can be little question, therefore,that the single lobe (Le) of the diplopod mandible is the lacinia, andthat the jaw of the Diplopoda has a structure identical with that ofthe insect maxilla, except for the lack of a galea and a palpus.The mandible of the Chilopoda is more specialized in structure thanis that of the diplopods, but in its musculature it is in some respectsmore generalized. In Scolopendra, Lithohius, Scutigera, the jaw isslender and greatly elongate. In Scutigera (fig. 17 G, Md) its taper-ing base is exposed on the side of the head where it is articulated tothe cranial margin (a), but in Seolopcndra (fig. 21 B) the end of themandible is buried in a pocket of the head wall lying mesad of thebase of the maxilla {MxC). The long basal plate of the chilopod jawis undivided (fig. 26 B, BP), and is articulated to the head wall by itsapical point (a). In some chilopods there is an anterior articulationbetween the mandible and the suspensorial plate of the hypopharynx,but this articulation is a mere contact between external surfaces.As in the diplopods, the basal plate has no inner wall. The distalpart of the mandible is a free lobe {Lc) movable on the base, but notso definitely hinged to the latter as is that of the diplopod mandible.The musculature of the chilopod mandible is practically alike inboth the Pleurostigma and the Notostigma, and is essentially the sameas in the diplopods. though the muscles differ in relative size. Thebasal plate is provided with both tergal and sternal muscles. Of theformer, there are two sets of fibers, one inserted on the dorsal (an-terior) edge of the proximal part of the plate (fig. 26B,C, /), theother (/) on the ventral (posterior) edge; both have their originson the dorsal wall of the cranium. These muscles apparently serveto rotate the mandible on its long axis, and they probably act as pro-tractors where the mandible is capable of a longitudinal movement
;
but clearly the first would be a promotor, and the second a remotorin an appendage with primitive relations to the head. The sternalmuscles of the mandible consist of a conical mass of adductor fibers(fig. 26B,C, A'L) spreading upon the inner wall of the basal platefrom their median origin (fig. 21 ?>, KL), which is on the ligamentousl)ridge uniting the two apodemes of the hypopharynx (fig. 21 A, C, /).The adductors of the chilopod mandibles are unquestionably homo-logues of the dumb-bell muscle of the diplopods. The condition ofthe mandibular adductors, therefore, is more primitive in the Chilo-poda, for here the muscles retain their connections with the sternal,hypopharyngeal apodemes.
66 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lThe movable terminal lobe of the chilopod mandible (fig. 26 B,C, Lc) is provided with the same muscles as is the corresponding lobeof the diplopod mandible (A. Lc) and the lacinia of the insect maxilla(figs. 25 C, 30 15, Lc) . The muscle from the lobe to the basal plate inthe chilopod jaw is very large (fig. 26 B, C, flcs), suggesting that ofJapyx (fig. 30 B. -flcs), and is composed of two groups of fibers. Thecranial muscle {ficc) arises by a broad base on the dorsal wall of thehead, and is inserted on a slender apodeme from the inner angle of thelobe. In the chilopod mandible, therefore, there is a basal plate (fig.26 B, BP) corresponding with the cardo and stipes of the insectmaxilla, but not divided as in the diplopods, and a free terminal lobe{Lc) which represents the lacinia. In retaining the connection of theadductor muscles with the hypopharyngeal apodemes, the chilopodmandible preserves the primitive condition shown by the maxilla ofJapyx {fig. 2>o^).In the Crustacea and Hexapoda, the mandible, or the jaw part ofthe mandibular appendage, which may bear a palpus, consists of asingle piece. Whatever may be the primitive elements that haveentered into its composition, these elements are fused into a solidgnathal organ. There are, hence, never muscles entirely within ^themandible, except those that pertain to the palpus, when a palpus ispresent. The mandibular musculature consists exclusively of themuscles that move the appendage as a whole, and these musdles cor-respond with the muscles of the basal plate of the myriapod mandible,or with those of the cardo and stipes of the insect maxilla.In the phyllopod crustacean Apits, the large mandibles (fig. 27 A,Md) hang vertically from the wall of the mandibular segment (IV).Each is a strongly convex, elongate oval structure, attached to thelateral membranous wall of the head by most of its inner margins,leaving only a ventral masticatory part projecting below as a free lobe.A single, dorsal point of suspension (a) allows the base of the man-dible to turn on its vertical axis, or to swing inward and outward asfar as the membranous lateral head wall will permit. The musculatureis correspondingly simple: two dorsal muscles from the tergum ofthe mandibular segment (IV) are inserted on the! base of the jaw, one(/) on the anterior margin, the other (/) on the posterior margin
;
the hollow of the mandible is filled with a great mass of fibers (KL)which converge upon a median transverse ligament (k) that receiveslikewise the muscles from the opposite jaw. Here, then, is a ventraldumb-bell adductor, as in the diplopods, and two dorsal muscles, whichmay function either as productors and reductors, or as anterior andposterior rotators. It is not clear as to what constitutes the mechanism
I
NO. 3 INSECT HEAD SXODGRASS 67
of abduction in appendages with this type of articulation and mus-culature.The Apus type of mandible is probably characteristic of most ofthe more generalized Crustacea; it is present also in some of thePcR
Md
Fig. 27.—Mandibles of Crustacea and Apterygota.A, mandibles of Apjis longicaudata (phyllopod), anterior. B, mandibles ofSpirontocaris grocnlandicus (decapod), anterior. C, mandibles of Hetcrojapyxgallardi (apterygote insect), anterior (dorsal). D, mandibles of Xcsoiiiachilismaoricns (apterygote insect), posterior.a. articulation of mandible with cranium, or with wall of mandibular seg-ment (IV); HA, hypopharyngeal apophysis; /, promoter of mandible; /,remoter of mandible; k, median tendon of mandibular adductors of dumb-bellmuscle {KL or KLk) ; KLk, fibers of mandibular adductors united by tendon{k) ; KLf, fibers of mandibular adductors retaining origin on hypopharyngealapophyses (D, HA); m, suspensory tendons of mandibular adductors; Md,mandible ; PcR, posterior cranial ridge ; f, branch of labral muscle attached onmandible.decapods (Virbiiis, Spirontocaris). In Spirontocaris (fig. 27 B), themedian ligament (k) of the dimib-bell adductors (KL) is connectedwith the hypodermis of the dorsal wall of the body by a branched arm(m) on each side. As before pointed out, however, the adductor liga-
/68 SMITHSONIAN MISCELLANEOUS COLLKCTIOXS VOL. 8l
meat is in no sense to be homologized with the tentorium as developedin some of the higher Crustacea and in the pterygote insects. Eachmandible of Spirontocaris is provided with two dorsal productormuscles (/), but a reductor was not observed. Spironiocavis preservesthe primitive single dorsal point of articulation of the mandible (a)with the wall of its segment. In higher decapods, the am])hipods. andthe isopods, where the mandible may have a double hinge with thewall of the head, the musculature of the organ is modified in a mannerto be described later.The simple mechanism of the mandible of the higher pterygote in-sects is well understood ; the complicated musculature of the mandiblein Apterygota has l)een given scant attention, and the derivation ofthe pterygote jaw mechanism from that of the Apterygota has beenalmost ignored. Borner ( 1909) has given the first comparative accountof the mandibular musculature in the more generalized insects, andhas pointed out certain points of similarity with the musculature ofhigher crustaceans. He did not, however, carry his comparisons tothe myriapods, and thereby missed some fundamental relations.The mandibles of the Machilidae will serve best as an example ofthe more generalized apterygote jaw. The mandible of MacJiilis orof NcsoinachUis (fig. 27 D, Md) is surprisingly similar in form tothat of the crustacean Apus {A), except that it has a long incisorpoint in addition to a broad molar lobe. In this latter character themachilid jaw resembles the mandibles of some of the decapod crus-taceans, such as Spi'irotocar'is and Virhhis, as has l)een pointed out byCrampton (1921b). The mandible of Machilis is suspended by asingle dorsal point of articulation (a) against the lateral wall of thehead. The cavity of the elongate base of each organ is filled by a massof muscle fibers (KLk). and these fibers from the two mandibles con-verge upon the ends of a common transverse tendon (k) that passesthrough the base of the hypopharynx. Here, in an insect, therefore.we find the same type of dumb-bell adductor uniting the two mandiblesas occurs in the Diplopoda and in lower Crustacea. In Madiilis.however, there is a second and larger set of adductor fil)ers (KlJ)which has its origin on the hypopharyngeal apodemes (HA). Machi-lis, therefore, in the possession of two dififerentiated sets of mandibularadductor fibers, combines the primitive condition of the chilopodswith the specialized condition of the diplopods and lower crustaceans.The tergal musculature of the mandible in Machilis is simple, con-sisting of an anterior promotor (/) and a posterior remotor (/). Thetwo muscles are disposed exactly as in Apus (A), and are in entire
NO. 3 INSECT HEAD-—SNOlKiRASS 69
conformity with the tergal musculature of the basal plate of the jawof Scutigera (fig. 26 B, C, /, /) and other chilopods.The machilid type of mandibular musculature appears to be char-acteristic of most apterygote insects except the Lepismatidae. InJapyx and Campodea, the bases of the elongate mandibles and maxillaeare deeply retracted into the head above the labium, and the edgesof the labium are fused to the postgenal margins of the head, so thatthe distal tdgQ of the labium appears as the ventral lip of a pouchcontaining the other gnathal appendages and the hypopharynx.The mandibles of Heterojapyx (fig. 27 C, Md) are simple, slenderorgans, each consisting of a long, hollow basal piece, and of a morestrongly chitinized free terminal lobe with a toothed incisor edge.The proximal tapering end of each jaw is set ofif from the rest bya thick internal ridge, superficially suggesting the division of themaxillary base into cardo and stipes ; but the " division " in the Japyxmandible gives rigidity instead of flexibility. The two mandibles ofHeterojapyx are connected by a large dumb-bell adductor muscle(KLk), the spreading fibers of which fill the basal cavities of theorgans. Besides this muscle there are also sets of ventral fibers (KLt)to the mandible that arises on the hypopharyngeal apophyses. Thetergal muscles of the mandibles are large : they include for each jawan anterior muscle (/) arising against a dorsal cranial ridge (PcR),and a wide fan of posterior fibers (/) arising along a median coronalridge. Because of the retraction of the mouth appendages, the hypo-pharyngeal muscles of the mandibles (KLt) would appear to functionas protractors, and the tergal muscles as retractors ; but the former areclearly the hypopharyngeal adductors of Machilis (D, KLt), and thelatter the tergal promotors (7) and remotors (/). A peculiaritynoted in Heterojapyx, if the writer observed correctly, is the attach-ment of a branch of the retractor of the labrum (t) on the base ofthe mandible.In the Collembola, which also have retracted mandibles and max-illae, the mandibular musculature would appear, from Folsom's
( 1899) account of OrchcscUa c'mcta, to be of the same essentialnature as that of Japyx. Folsom enumerates ten muscles for eachmandii^le of OrchcscIIa. but they all fall into three groups accordingto their origins, namely, muscles arising on the walls of the head,muscles arising on the "tentorium" (hypopharyngeal apodemes),and fibers from one mandible to the other. The second and thirdgroups constitute the adductors of the jaw ; their fibers are inserted.Folsom says, on the inside of the lateral wall of the mandible, andmost of them have their origin on the " tentorium." but a few of the
70 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
fibers, he adds, '' pass under the tentorium and become continuouswith similar fibers from the opposite mandible." Folsom, it will benoticed, says the adductor fibers connecting the mandibles pass be-neath the tentorial arms. In Japyx the tendon of the dumb-bell muscledistinctly lies dorsal to the hypopharyngeal apodemes. In Machilisthe apodemes are so loosely connected with the base of the hypo-pharynx and so strongly united with the lateral inflections of the headwall, that in dissections their hypopharyngeal connections are easilylost, and the impression is given that the tendon of the dumb-belllAnt J-~.
iMxpAFig. 28.—Head of Gammarus Jocusta (amphipod crustacean).A, lateral view of head, showing tergal abductors (/) and adductors(/) of left mandible, and base of ventral adductor (KL). B, postero-ventralview of back of head, showing origin of ventral adductors {KL) on posteriortentorial bar (PT).IAnt, first antenna ; 2Ant, second antenna ; /, abductor of mandible ; /, dorsaladductor of mandible ; KL, ventral adductor of mandible ; Lm, labrum ; Md,mandible; MdPIp, mandibular palpus; lALr, first maxilla; pMx, second maxilla;iMxp, first maxilliped ; pt, posterior tentorial pit ; PT, transverse posteriortentorial bar.adductor lies ventral to the apodemes. It does lie ventral to the sus-pensorial plates uniting the apodemes with the lateral walls of thehead, but it passes anterior, i. e., dorsal, to the ends of the apodemesthemselves. Folsom's statement, therefore, should be verified, fora discrepancy in the relations of the parts in question seems hardlypermissible if we are dealing with homologous structures.The mandibles of the Protura, as described by Berlese (1909), areprovided each with retractors and protractors that have their originson the head wall, and with a protractor arising on the tentorialapodeme. Berlese, however, does not mention a muscle continuousbetween the two mandibles. The muscles present clearly representthe usual tergal muscles, and the hypopharyngeal adductor.
NO. 3 INSECT HEAD SNODGKASS 7
1
In all the apterygote forms thus far described, the mandible hasa free attachment to the head, being implanted by most of its lengthin the ventro-lateral membranous part of the head wall, and articulatedto the margin of the chitinous cranium by only a single dorsal point ofcontact. In the Lepismatidae, a new condition is established in themandible through the elongation of its dorsal loase line forward andventrally to the anterior end of the lower genal margin of the epicra-nium. The jaw thus becomes hinged to the head on a long basal axisextending from the primitive dorsal articulation, which is now pos-terior, to the angle between the genal margin of the head and theclypeus. At the latter point a secondary, anterior articulation isestablished between the mandible and the cranium. Bonier (1909)describes the articulation of the mandible of Lepisma, but he doesnot observe that its type of structure is characteristic of the Lepis-matidae only, not of the Apterygota in general. The alteration in themandibular articulation involves a change in the entire mechanismof the jaw, and initiates the series of modifications that have ledto the evolution of the pterygote type of mandibular musculature fromthat of Machilis, Japyx, and the Collembola.The musculature of the mandible of Lepisma, as described byBorner (1909), is apparently almost the same as that of Machilis.The adductor muscles inserted within the body of the mandible consistof a large dorsal set of fibers (fig. 29 B, KLt) from the tentoriumrepresenting the fibers that arise on the hypopharyngeal apodeme ofMachilis (figs. 27 D, 29 A, KLt), and of a small ventral set (KLh)arising directly from the hypopharynx. The tergal muscles comprisea pair of abductors (/) inserted on the outer margin of the mandibularbase between the two articular points (a, c), and a large dorsal ad-ductor (/) inserted on the inner margin mesad of the posterior artic-ulation. The tergal abductors and adductor, however, are clearlythe promotor and the remotor of the mandible of Machilis (fig. 29 A,/, /) and of all other generalized forms, which have assumed a newfunction by reason of the change in the nature of the mandibulararticulation.The structure and musculature of the mandible in nymphs of Ephe-merida is essentially the same as in Lepisma. Borner describes andfigures the nymph of Cloeon dipteruni, showing the presence of alarge tentorial adductor and a very small hypopharyngeal adductor,in addition to the dorsal abductors and adductors ; the writer hasverified the existence of all these muscles in another ephemeridspecies. In a dragonfly nymph, Aeschna, a small hypopharyngealadductor was found, but no tentorial fibers were observed. In the
72 SMITIISOXIAX ^[ISCKLLANEOLS COLI-ECTJOXS VOL. 8l
orthopteron, Locusta, Bonier shows two small tentorial adductors ofthe mandible (fig. 29 C, KLt) , and a small hypopharyngeal adductor(KLIi). The same muscles the writer has found in Microccntrum,the hypopharyngeal fibers being attached medially on the tips of therudimentary suspensorial arms of the hypopharynx (fig. 20 D, KLJi) ;but no trace of either set could be discovered in the acridid, Dissos-teira. Mangan ( 1908) described in the roach, Pcriplancta auslrakisiae,both a tentorial adductor and a hypopharyngeal adductor. The first
'--KLt
KLk' AFig. 29.—Three sta.sfes in the evolution of the mandibular mechanism inbiting insects.A, mandible of Machilis. outer surface, with single dorsal point of articulation(a) with cranium; the jaw moved by tergal promoter (/) and remotor (/),and by ventral adductors (KLk, KLt, see fig. 27 D).J>, mandible of Lcpisma (from Borner, 1909), articulated with cranium onlong basal hinge inclined downward anteriorly from dori;al articulation (a) toanterior articulation (r) ; the promoters (/) here become abductors, and theremotor (/) becomes a tergal adductor; ventral adductor (KLh, KLt) retained.C, head of Locusta (from Borner, 1909), showing common type of mandibulararticulation in pterygote insects, w-ith hinge line inclined downward posteriorlyfrom anterior articulation (c) to posterior articulation (a) ; tergal abductorsand adductors (/, /) highly developed, ventral adductors {KLh, KLt) rudi-mentary. In In'gher Pterygota the ventral adductors disappear.mention of either of these muscles is liy Basch (1865), who foundthe tentorial adductor in the mandible of Tcrines Havipes.The adductor fibers arising directly from the base of the hypo-pharynx are evidently remnants of the primitive sternal adductors thathave retained their original connections. /;/ the insects, tlicreforc,the primary, sternal adductor muscles {KL) of the mandibles havebecome differentiated into three groups of fibers, the fibers of onegroup {KLh) retaining the primitive sternal connection, those of thesecond {KLt) being carried inzvard upon the sternal (hypopharyngeal)apophyses, those of the third (KLk), after having united medially
INO. 3 INSECT HRAl) SNODGKASS 73
z^'ith the corresponding set from the opposite mandible, having beendetached from all connections except their points of insertion on themandibles. With the change in the mandibular articulation from asingle dorsal suspensory point to a long basal hinge, the primary ad-ductors have lost their importance, and the function of adductionhas been secondarily taken over by the primary tergal remotor, whilethe original tergal jjroniutor becomes the alxluctor. Remnants of theprimary adductors in insects having a hinged mandi])le ])ersist inthe Lepismatidae, Ephemerida, r)rth()ptera, and Isoptera, but in thehigher orders they have disappeared.A still further evolution in the mandibular base has reversed thetilt of the hinge line. Instead of sloping from the posterior articulationdownward and forward, as it does in Lepisma and in some ephemeridnymphs, the base of the jaw in all higher insects is inclined from theanterior articulation downward and posteriorly (fig. 29 C). Thischange in the slope of the axis of the hinge causes the apex of thejaw to swing inward and posteriorly during adduction, instead of in-ward and anteriorly as in the first condition.In the higher decapod crustaceans, and in the amphipods and iso-pods, the mandible has undergone an evolution parallel to that whichhas taken place in insects. Bonier (1909) has descril)ed the mandibleand mandibular musculature of Gammarus, an amphipod, and hasshown the structural similarity with the mandible of Lepisma. InGammarus locusta (fig. 28 A) the mandible is hinged to the craniuml)y its long base, which slopes downward and forward from the pos-terior point of articulation. The primitive tergal ])roniotor muscle(7) has then become an al)ductor, and the remotor (/) has becomea dorsal adductor. The primitive ventral adductor (KL) has itsorigin on a well-developed transverse tentorial bar (B, PT) passingthrough the back of the head ; a hypopharyngeal branch of theadductor is lacking. In the crayfish (Astacus), Schmidt (1915)describes an anterior ventral adductor of the mandible arising on theanterior end of the ventral head apodeme. In the isopods the mandibleattains a stage almost exactly comparable with that of the higherpterygote insects (fig. 17 F)—the basal hinge line of the jaw^ slopesposteriorly and downward, and the only muscles present, so far asthe writer could find, are the tergal abductors and adductors.The homologues of the mandibles in Xiphosura and Arachnida,the so-called pedipalps (fig. i/],Pdp), scarcely need considerationhere. The pedipalps never attain a jaw-like form, but retain alwaysthe structure of a jointed limb, though the basal segment may developa strong gnathal lobe.
74 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lTHE FIRST MAXILLAEThe leading features of the first maxilla have been sufficientlynoted in the description of a generalized gnathal appendage (page60) based on the maxilla of Penplancta (fig. 25). In none of theother arthropods are the maxillary appendages so highly developed asin the insects, but, in all the arthropods, it appears that the mandiblehas been evolved from an appendage that was originally very similarCd e PcR'
GaA BFig. 30.—Maxilla of Heterojapyx gallardi.A, left maxilla, posterior (ventral) surface. B, right maxilla and muscles,anterior (dorsal) view.Cd. cardo ; c. articulation of cardo with cranium; ftia, flexor of galea; flee.cranial flexor of lacinia ; flcs. stipital flexor of lacinia ; Ga, galea ; HA, liypo-pharyngeal apophysis ; HS, rudiment of suspensorial arm of hypopharynx ; /,promotor of cardo ; KLcd, adductors of cardo ; KLst, adductors of stipes ; Lc,lacinia; OQ, muscle of base of palpus; p. muscle of terminal segment of palpus;PcR, posterior cranial ridge; Pip, palpus; iplp. first segment of palpus; St.stipes ; u, line of internal ridge of stipes.
to the generalized insect maxilla. In many of the higher insects themaxillae, too, have become specialized, always in adaptation to specialmodes of feeding, but a description of the modifications involved isbeyond the scope of the present paper. The musculature of the organis essentially the same in all groups of biting insects, except as itsuffers a reduction where the appendages become reduced or unitedwith the labium.The maxilla of Japyx { fig. 30) presents a more generalized con-dition in its relation to the head than does the maxilla of the roach,
NO. 3 INSECT JIEAD SNOUGKASS 75in that the head apophyses (B, HA) upon which the adductor musclesof the appendages arise are still connected with the hypopharynx,whereas in Periplaneta the corresponding endoskeletal arms have losttheir primitive sternal connections and have become a part of thetentorium. The adductors of the cardo in Heterojapyx (fig. 30 B,KLcd) are well differentiated from those of the stipes (KLst), andcross obliquely the inner ends of the latter. The promotor of the cardo(/) arises against a median ridge of the dorsal wall of the cranium.The lacinia (Lc), which is mostly covered dorsally by the galea, has abroad flexor arising within the stipes (Acs), and a large cranial muscle{flee) arising against the dorsal cranial ridge (PeR) on the top of thehead, and going dorsal to the other muscles of the appendage to beinserted on a slender apodeme from the inner angle of the lacinialbase. The galea (Ga) is provided with a single long flexor (figs. 30 B,31 D, fga) arising within the stipes, which splits into two bundles offibers toward its insertion on the ventral wall of the base of the galea.The palpus (Pip) is reduced and otherwise modified as compared withthat of the roach (fig. 25), consisting of only three segments, ofwhich the basal one (figs. 30 A. 31 D, iplp) is much elongate and isunited with the outer wall of the base of the galea (Ga). Theremight be some question as to the homology of this basal region of thepalpus of Japyx, but the insertion upon its base of the muscle (OQ)from the stipes, evidently representing the usual pair of palpal muscles,and the origin within it of a muscle (p) going to the distal segment ofthe palpus identify the part in question as the true basal segment ofthe palpus.The cardo and the stipes of many insects appear externally to bedivided into sub-sclerites, but in most such cases it is found that theso-called " sutures " are but the external lines of inflections that haveformed internal ridges, the ridges being developed either for givingstrength to the sclerite, or to furnish special surfaces for muscleattachment. The cardo of Periplaneta, for example (fig. 2^A,Cd),has a " divided " appearance externally, but when examined fromwithin (B) it is seen that the regions apparent on the surface resultfrom the presence of a strong Y-shaped ridge (r) on the inner wall,which extends distally from the base to reinforce with its divergingarms the extremities of the hinge line with the stipes. This structureof the cardo is characteristic of other orthopteroid insects. Cranipton(1916) distinguishes the area of the cardo between the arms of theridge as the " veracardo," and calls the rest of the sclerite the " juxta-cardo." The terms may have a descrijjtive convenience, but theyare misleading if taken to signify a division of the cardo into two parts.
76 SAllTHSUNJAN MISCELLANEOUS COl-f.ECTIONS VOL. 8
1
The stipes is usually marked by a prominent groove parallel to itsinner edge (fig. 25 A, q), setting off a narrow marginal strip. Thegroove is here likewrise but the external line of an internal ridge orplate upon vi'hich are inserted the adductor muscles of the stipes( B, KLst) . Crampton designates the area of the stipes external tothe ridge as the " verastipes," and that mesad to it as the " juxta-stipes." In Hctcrojapyx the basal part of the stipes is .strengthenedl)y an internal ridge (fig. 30 A, u) that forks proximally to the endsof the hinge line with the cardo.
Fjg. 31.—Maxillae of insects and of a chilopod.A, maxilla of Ncsoiiiachilis. B, maxilla of Tlicniiobia (Lepismatidae). C,maxilla of larva of Sialis. D, terminal part of maxilla of Hctcrojapyx. V.,maxilla of an adnlt stonefly (Pteroiiarcys). F, first maxillae of a chilopod(Litliobiiis).Base of palpus to be identified by insertions of levato'" and depressor mnscles(O, Q) ; the palpifer (Plf) has no muscles, and appears as a mere subdivisionof stipes; in Sialis larva (C). lobe is not the galea, but an endite of firstsegment of palpus {iplp), the latter identified by its muscles (O, (J).The ventral, or distal, end of the stipes bears the lacinia and galea,and to its lateral surface is attached the palpus. The lacinia and galeaare movable lobes, each being jirovided with muscles having theirorigin in the stipes, by which they can be fiexed posteriorly (or ven-trally if the mouth appendages are horizontal). The lacinia, in ad-dition, has a muscle from the cranial wall inserted on the iinier angleof its base, which gives it a mesal fiection, or adduction. The base ofthe galea commonly overlaps anteriorly the base of the lacinia.The maxillary palpus arises from the outer wall of the stipes,usually only a short distance proximal to the base of the galea. The
«NO. 3 INSECT HEAD SNODGRASS "JJ
area supporting the palpus is frequently differentiated from the restof the stipes, and is then distinguished as the palpifer (fig. 31 A,F//). When the delimiting suture of the palpifer region extends tothe galea, the palpifer appears to support both the galea and thepalpus. That the palpifer is not a segment of the appendage is shownby the fact that inuscles neither arise tinthin it nor are inserted uponit. The muscles that move the palpus as a whole have their originswithin the main part of the stipes, and always pass through the pal-pifer, if the latter is present, to 1)e inserted on the ])roxinial segmentof the palpus (figs. 25 C, 31 A-E, 0,0). The pal])us muscles, then,may be taken as identification marks of the true basal segment of thepalpus. Since they are typically inserted one dorsally and the otherventrally, relative to the vertical axis of the appendage, they areevidently a levator ( O ) and a depressor ( ) of the palpus. Thenumber of segments in the maxillary palpus varies much in differentinsects. Machilis perhaps presents the maximum number of seven(fig. 31 A) : the palpus of the roach with five segments is more typical(fig. 25). Evidence will later be given indicating that the palpus isthe telopodite of the maxillary appendage, and that its basal articula-tion with the stipes, or palpifer, is the coxo-trochanteral joint of amore generalized limb (fig. 35 A, B, C, ct) . A joint near the middleof the palpus (figs. 25 C, 35 A, B. ft) often suggests the femero-tibial fiexvn-e. THE SECOND MAXILLAEThe second maxillae of insects are unquestionably united in thelabium. The correspondence in external relations between the parts ofeach half of a typical labial appendage and those of an entire maxillais so close that most entomologists have not hesitated to assume anhomology of the submentum (figs. 32 A, 40 D, Suit) with the cardines,of the mentum {Mt) with the stipites, of the glossae {Gl) with thelaciniae, and of the paraglossal {Pgl) with the galeae. Some writers,however, have contended that the sul)mentum, or both the submentumand the mentum represent the sternum of the labial segment. Thus,Crampton in a recent paper (1928) adopts the idea of Holmgren( 1909) that the submentum and mentum arc derived from the sternumof the labial segment.In an orthopteroid labium (fig. 40 D), the muscles of thepalpi {28, jq), and the muscles of the terminal lobes (-^5) arise in thementum {Mt), and this relation, together with the presence of musclesfrom the mentum to the tentorium (i-J, 24). must certainl\- identifythe region of the mentum in the la1)ium with that of the stipes in a
78 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
first maxilla (fig. 40 B, St). The wall of the mental region, however,may not be entirely or continuously chitinized (fig. 32 A), and, hence,a distinction must be drawn between the entire region of the mentum,and the area occupied by one or more mental sclerites. The labiummay contain muscles not represented in the maxillae, such as themuscles associated with the orifice of the salivary duct in the grass-hopper (fig. 40 D, 26, 2/), or with the silk press in the caterpillar(figs. 53 C, D, 54, A, B, C, ly, 18, 19).The submentum corresponds functionally at least with the cardinesof the maxillae, since it serves to attach the labial appendage to thewalls of the head. The lateral articulation of its basal angles to the
BFig. 32.—Second maxillae.A, typical second maxillae of an insect (Periplaneta) united to form thelabium. B, second maxillae of a chilopod (Lithobius) united by inner anglesof coxae.ct, coxo-trochanteral joint; Cx, coxa; Gl, glossa; Mt, mentum; O, levatormuscle of telopodite (palpus) ; Pgl, paraglossa; Pig, pelpiger; Pip, palpus; Q,depressor muscle of telopodite (palpus) ; Smt, submentum.
margins of the cranium in orthopteroid insects (figs. 18 B, C, 36 C, /)suggests, moreover, that the points of attachment are the truebasal articulations of the second maxillae with the cranium, corre-sponding with the articulations of the cardines (e) in the first maxillae.It is possible, of course, that a median part of the labial sternum hasbeen incorporated into the submentum. To accept the proposal, how-ever, that the entire submentum is the sternum of the labial segment,is to assume that the sternum itself has become articulated laterallyto the tergum of its segment, and that it alone bears the segmentalappendages. Such assumed relations violate the basic principles ofsegmental inorphology, and thus throw suspicion on the evidencegiven in their support.
NO. 3 INSECT HEAL) SNODGRASS 79
It will be shown in the next section of this paper that the cardinesof the maxillae are not true proximal segments of the maxillaryappendages, but are secondary subdivisions of the bases of theseappendages. It appears probable, therefore, that the submentumrepresents likewise proximal subdivisions of the bases of the secondmaxillae, retaining the lateral articulations with the margins of thecranium in generalized insects (fig. 36C, /), and perhaps includingbetween them a median part of the labial sternum.If the insect labium (figs.< 32 A, 40 D) is compared with the secondmaxillae of a chilopod (fig. 32 B), it will be seen that the united basalsegments of the latter {Cx), containing the origins of the palpalmuscles {O, Q) , correspond at least with the mentum of the labium.The large proximal segments of the chilopod maxillae are clearly thebases of a generalized limb, the coxae, according to Heymons (1901),and the limb base, or a subcoxal division of it, bears the primitivedorsal articulation of the appendage with the body. The mentum,and at least the lateral parts of the submentum, therefore, appear tobe subdivisions of the primary bases of the second maxillary ap-pendages, corresponding with the stipites and cardines of the firstmaxillae in insects, and with the similar subdivisions of the bases ofthe mandibles in the diplopods (fig. 26 A, Cd, St).The median, terminal duct of the labial, or " salivary," glands opensanterior to the labium, and, in typical forms, at the base of the mentum(figs. 18 D, 19, SIO). The position of the orifice, anterior to the sub-mentum, however, does not argue that the latter is entirely the sternumof the labial segment, but rather the reverse, for it is likely that theorifice of the duct has not left the sternal region of its segment, andthat it has been crowded forward in the latter by the median approachof the labial cardines. The common duct of the lal)ial glands resultsduring embryonic development from the union of the two primaryducts of paired lateral glands of the labial segment.
MORPHOLOGY OF THE GNATHAL APPENDAGES
It has often been assumed that the segmental appendages of allarthropods are derived from a primitive limb having a biramous typeof structure. A two-branched limb, however, occurs actually only inthe Crustacea, and there is no certain evidence of a biramous limbstructure ever having prevailed in other arthropod groups. In allforms, including the Crustacea, the segmental appendages first appearin the embryo as simple protuberances of the body wall, and somezoologists now believe that the exopoditc branch, when present, is
8o SMll'HSONIAX M1SCI:LLANE0US COLI.iaTlONS vol.. 8
1
merely a specially developed exite lobe of a single shaft. Borradaile(191 7) expresses the opinion that " probably the primitive crustaceanappendage resembled that of the Branchiopoda in being uniramous."Movable lobes individually provided with muscles, however, may bedeveloped along both the outer and the inner margin of the Umb, andan excessive development of one of the outer lobes might give rise to
L.:
Fig. 33.—Parapodium and parapodial musculature of an annelid worm(Nereis virens).A, B, first and third parapodia, left, anterior surfaces.C, cross section of left half of a segment from middle of body, cut anteriorto base of parapodium. showing muscles of setae inserted on end of setalpouch (a), and ventral promotor (A.') and remotor (L) muscles of parapodium.DMcl, VMcl, dorsal and ventral bands of longitudinal body muscles.D, musculature of third parapodium, right, inner view, showing tergal pro-motor (/) and remotor (7), and sternal promotor (K) and remotor (L).E, musculature of right side of a segment from middle of body, internalview, lateral oblique muscles and setal muscles removed : b, c, anterior and pos-terior pleuro-sternal muscles ; DMcl, part of dorsal longitudinal muscles ; /, tergalpromotor of parapodium ; /, tergal remotor ; /, accessory remotor arisinganteriorly from intersegmental fold ; K, sternal promotor ; L, sternal remotor
;
.SV^ bases of setae.
a secondary biramous structure of the appendage. Hansen (1925)recognizes the definitive two-branched structure of the typical crus-tacean appendage, but he says it seems " impossible to deny the possi-bility that the exopod may be analogous with the epipod, and if sothe primitive appendage is uniramous."The segmental appendages, or parapodia, of the polychaete annelidsare in some cases simple lobes ; in others they are of a two-branched
NO. 3 INSECT HEAD SNODGRASS 8l
structure owing to the presence of two groups of setae on each (fig.33 B,C). In Nereis vircns, though most of the parapodia are dis-tinctly cleft, those of the first and second segments do not have thedouble structure (fig. 33 A). Whatever relations, however, may betraced, or assumed to exist, between the annelids and the arthropods,the relationship must be presumed to have come through a remoteworm-like ancestor common to both groups, for none of the highlyorganized modern annelids can be taken to represent the ancestralform of the arthropods.A comparative study of the legs of mandibulate arthropods willshow that in each group there is a maximum of seven limb segments,l)eyond a subcoxal base, that are individually provided with muscles.The relative size and form of the segments, the character of the articu-lations, and the nature of the musculature present many variations,and it is not to be assumed that segments are to be homologized inall cases by their numerical order beyond the base of the limb.The gnathal appendages undoubtedly constitute a group of organsthat are individually homologous in arthropod groups, whether theirsegments are united with the protocephalon to form a larger head,or with the body segments following. The similarity of the structureof the mandible in all the eugnathate arthropods, and the commonplan of its musculature, allowing for modifications of which theevolution can easily be followed, leave no doubt concerning the identityof the jaw in the various groups, or that the jaw attained its basicstructure in some very remote common ancestor. The primitive struc-ture of the mandible is not entirely preserved in any arthropod : inthe Diplopoda and Chilopoda the movable lacinia is retained, but thepalpus has been lost ; in the Crustacea and Hexapoda, the lacinia haslost its independent mobility and has become solidly fused with thebase of the appendage, but in many crustaceans a mandibular palpuspersists.The first maxilla of the Hexapoda has the structure of a generalizedmandible, i. <?., it consists of a base supporting a palpus and at leastone movaTjIe lobe, the lacinia, though generally there is present asecond lobe, the galea. The insect labium consists of a pair of ap-pendages that probably once had the structure of the first maxillae.In the Chilopoda the maxillary appendages appear to have under-gone but little modification of structure, and those of the secondpair still retain a form similar to that of the body appendages. Thecorresponding appendages of the Diplopoda are now so highly spe-cialized that it is useless to speculate as to their earlier form. In theCrustacea both pairs of maxillae have been reduced in size and modified
82 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Si
in structure to serve as organs accessory to the mandibles, but theyhave not attained the highly specialized form of the correspondingappendages of insects.We may conclude, therefore, that in the common ancestor of theseveral groups of modern eugnathate arthropods, the mandible alonehad attained a gnathal function, and that in form and structure it re-sembled the maxilla of a present day insect, though perhaps lackinga galea, or outer endite lobe of the base. The two maxillae at thisperiod were more or less modified to serve as organs accessory to themandibles.When the modem groups of arthropods were differentiated, themandible, in the Diplopoda and Chilopoda, retained the movablelacinia, but lost the palpus ; in the Crustacea and Hexapoda, the laciniafused with the base of the appendage to form a solid jaw, while thepalpus was preserved by the crustaceans, and lost by the insects. Thetwo maxillary appendages retained the leg-like form in the Chilopoda
;
in the Diplopoda they became highly specialized in a manner peculiarto the diplopod group ; in the Crustacea and Hexapoda they weremodified for an accessory gnathal function, but in the insects they ac-quired a form almost identical with that of a primitive mandible.Finally, in the insects, the second maxillae united basally to form thelabium. While the insect maxilla appears to be a highly specializedappendage, it will be shown that its basic structure is not far removedfrom that of a thoracic leg.While the status of the gnathal appendages relative to one anotherin the various groups of the eugnathate arthropods seems fairly clear,it is a more difficult matter to homologize their parts with the segmentsof the ambulatory appendages. The structure of the first and secondmaxillae of a chilopod, or of the first maxilla of an insect, suggeststhat the gnathal appendages have been derived from an appendageof the ambulatory type—the insect maxilla is certainly more like theleg of an insect, a chilopod, or a decapod than it is like one of thebody appendages of Apits (fig. 35 C), or of any other of the lowercrustaceans in which the appendages are used for swimming. Thiscondition suggests, therefore, that the ambulatory leg more nearlyrepresents the primitive type of arthropod limb than does an appen-dage, such as that of Apiis, clearly modified for purposes of purelyaquatic locomotion. If we consider, furthermore, that the appendagesof the chelicerate arthropods (Xiphosura and Arachnida) are also ofthe ambulatory type, the evidence becomes all the more convincingthat the primitive arthropod limb was a walking leg and not a swim-ming organ. If this deduction is acceptable, we must conclude that
i
NO. 3 INSECT HEAD SNODGRASS 83
the Crustacea represent a group of arthropods that have secondarilyadopted an aquatic life, and that, while certain forms have becomethoroughly adapted to a free life in the water, others have retained,with but little modification, some of the organs that were developedprimarily for terrestrial locomotion. This view, perhaps, is contraryto generally accepted ideas concerning the evolution of the arthropods,but it is clearly futile to attempt to derive the appendages of arth-ropods in general from the swimming appendages of Crustacea.If the ambulatory limb be taken as more nearly representative ofthe basic structure of an arthropod appendage than is the natatory
Fig. 34.-malic. -Generalized segmentation and musculature of an insect leg, diagram-A, theoretical segmentation and musculature of a primitive arthropod leg,anterior view: the appendage, consisting of a limb base (LB), and a telopodite{Tip) of two segments, moved forward and backward on vertical basal axis{a-b) bv tergal and sternal promoters (/. K) , and tergal and sternal remotorsU,L).'D, definitive segmentation of an insect leg by division of limb base (A, LB)into subcoxa {Sex) and coxa {Cx), and by subsegmentation of first part oitelopodite into trochanter and femur, and of second part into tibia, tarsus, andpraetarsus.a-b, basal axis of limb base ; ct, coxo-trochanteral joint ; Cx, coxa ; F, femur
;
ft, femoro-tibial joint; /, tergal promotor; /, tergal remotor; K, sternal pro-motor ; L, sternal remotor ; O, levator of telopodite ; P, tergal depressor of telo-podite (characteristic of insects) ; Ptar, pretarsus; Q, depressor of telopodite;Sex, subcoxa ; T, depressor of distal segment of telopodite ; Tar, tarsus ; Tb. tibia.limb, we have only to inquire as to what was its probable form in theancestors of terrestrial arthropods. The primitive appendage un-doubtedly turned forward and backward on a vertical axis through itsbase (fig. 34 A, fl, &), as does the parapodium of a modern polychaeteannelid (fig. 33 A, B, C). For walking purposes, however, the limbmust have acquired joints, and, as Bonier (1921) has shown, thesimplest practical condition would demand at least tzvo joints withvertical movements (fig. 34 A), one near the union of the leg with thebody {ct), dividing the limb into a basal piece {LB) and a telopodite
84 SMITHSONIAX MISCELLANEOUS COLLECTIONS VOL. 8l(Tip), and one (ft) near the middle of the telopodite. These primaryjoints persist, evidently, as the coxo-trochanteral joint (B, ct) and thefemero-tibial joint (ft) of the leg {Hiiftgelenk and Kniegclenk, ac-cording to Borner). The type of leg-segmentation resulting fromtwo joints so placed applies at least to the Chilopoda, Diplopoda,Hexapoda, and Crustacea ; in the Xiphosura and Arachnida, however,it is possible that the mechanism of the leg is given by three primaryjoints, the second and third setting off a horizontal middle sectionof the leg (patella).The further segmentation of the limb has been produced by thesubdivision of the principal parts of the telopodite. In the mandibu-late arthropods (fig. 34 B), one or two small segments cut oft' fromthe basal end of the j^roximal piece of the telopodite form the tro-chanters (Tr), while the rest of this part becomes the femur (F) ;the distal section beyond the knee joint {ft) subsegments into thetil)ia (Tb), tarsus (Tar), and praetarsus (Ptar). This type of seg-mentation is clearly shown also in the maxillipeds or in any of theanterior body appendages of Apus. In the third maxilliped (fig. 35C) there are two principal flexures, one {ct) between the limb base{LB) and the telopodite {Tip), the other {ft) beyond the middle ofthe latter. The part between the two points of flexure is the femur(F) with two indistinctl}' separated trochanters {Tr) ; that beyondconsists of two shortened segments, and the terminal praetarsus, ordactylopodite. The limb base of Apus is entire, but in some arthropodsthe basis appears to have become subdivided into a coxa (fig. 34 B,Cx) and a subcoxa {Sex). The coxa may then become the freefimctional base of the appendage, since the subcoxa usually forms achitinization in the pleural or the sternal wall of the segment.The primitive musculature of the limb base was such as to swingthe appendage forward and backward ; it must have comprised, there-fore, promotor and remotor muscles. Probably there was a tergalpromotor (fig. 34 A, /) and a tergal remotor (/), and a sternal pro-motor {K) and a sternal remotor (L). In a thoracic leg of an insect,the base of the telopodite is provided with a depressor muscle (F)having its origin on the tergum of the segment, which greatly increasesthe lifting power of the appendage, but this muscle is not to be con-sidered as a primitive element of the limb musculature. The usuallevator and depressor muscles of the telopodite {O, Q) have theirorigin within the limb base.The simple type of musculature shown in figure 34 A, and hereassumed to be the primitive musculature of an arthropod limb base,is actuallv present in typical form in the simpler anterior parapodia
NO. 3 INSECT HEAD SNODGRASS 85
of the annelid, Nereis virens (fig. 33 D). Here a dorsal proniotorand a remoter (/, /) arise on the tergal wall of the segment, and aventral promoter and a remotor {K, L) on the midline of the sternalwall. The ventral muscles are repeated regularly in all the segmentsof the worm (C, E, K, L), but in the more posterior segments thedorsal muscles, though present (E, I, J), are less symmetrical in ar-rangement, and the primary remotor (/) is subordinated to a largeoblique remotor (/) that arises on the anterior margin of the seg-ment. This last muscle is described by Borner (1921) as being thetypical dorsal remotor of the parapodium, but by comparison with thesimpler anterior appendages (D) it appears to be a secondary acquisi-tion, for the muscle (E, /) dipping beneath it has the same insertionon the parapodial base as that of the tergal promoter of tlic anteriorparapodium (D, /). Borner's claim, however, that this simple type oflimb musculature presented by Nereis must represent the primitivemotor mechanism of an appendage turning forward and backward ona vertical axis through its base is scarcely to be questioned.The basal muscles of an appendage do not necessarily retain theiroriginal functions, nor their primitive simplicity, for an alteration inthe basal articulation of the appendage may change the fundamentalmovements of the limb, and thereby give quite a different action tothe muscles, which, in turn, may shift in position, or become split upinto segregated groups of fibers, thus multiplying the number of in-dividual muscles actually present.Returning now to a further consideration of the muscles of thegnathal appendages of the arthropods, it is not difficult to draw aparallel between the musculature of a mandible, or of an insectmaxilla, and that of the annelid parapodium (fig. 33D, E), or withthe hypopthetically primitive musculature of an arthropod limb baseas expressed in figure 34 A. In the mandible of Scutigera (fig. 26 B,C), Apus (fig. 27 A), Hetcrojapyx (fig. 27 C), Ncsomachilis (fig.27 D), a tergal promoter (/) and a remoter (/) have the typical re-lation to the appendage. In some forms the tergal remotor appearsto be lacking (Diplepeda, fig. 26 A; Spiroiitocaris, fig. 27 B). Thecranial flexor of the mandibular lacinia m diplopods and chilopods(fig. 26 A, B, C, iJcc) is probably derived from the tergal promoter,since it arises en the dorsal wall of the head and goes dorsal (an-terior) to the ventral muscles. The sternal promoter and remotor,which are distinct muscles in the annelid parapodium (fig. 33 C, D, E,K L), are united in the gnathal appendages of the arthropods, wherethey become ventral adductors (A'L) as a result of the free movement
86 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8
1
of the base of the appendage on a single dorsal point of articula-tion (a).The adductor fibers of the mandible may all retain their connectionwith the sternal, or hypopharyngeal, apophyses (Chilopoda, fig. 21 B),or they may become detached from the apophyses and united with thefibers from the opposite jaw to form a transverse dumb-bell muscle(Diplopoda, fig. 26, A, KL; some Crustacea, fig. 27 A, B ; mostApterygota, fig, 27 C, D, KLk), though at the same time some ofthe fibers may retain their connections with the apophyses, or withthe tentorium (Apterygota, fig. 27 C, D, KLt; Orthoptera, fig. 29 C,KLt) or with the base of the hypopharynx (Lepisma, fig. 29 B, KLh ;Locusta, fig. 29 C, KLh ; Microcentrum, fig. 20 D, KLh ; ephemeridnymph, fig. 20 A, KLh). The evolution of the mandibular muscles inthe higher insects has been detailed in an earlier paragraph, wherein itwas shown that the ventral adductors are reduced and finally obliteratedafter the jaw has acquired a double hinge with the edge of the cranium,and that the tergal promotor and remotor muscles then become respec-tively the functional abductors and adductors.The basal musculature of the insect maxilla, as already shown,coincides almost exactly with the basal musculature of the mandibleof a diplopod or a chilopod, and may be derived from the simple planof the musculature of the annelid parapodium. The tergal promotoris evidently separated into two groups of fibers inserted on the dorsaland ventral extremities of the anterior rim of the appendage base,the upper set being the muscle of the cardo (figs. 25 C, 30 B, 7), thelower set the cranial flexor of the lacinia {Hcc). A tergal remotor islacking in the insect maxilla, but so it appears to be also in the diplopodmandible (fig. 26 A). The sternal promotor and remotor muscles (figs.33 D, E, 35 B, i<^, L) are united, as in the mandible, to form a ventraladductor (KL), the fibers of which almost always retain their originon the hypopharyngeal apophyses, or on the corresponding part ofthe tentorium, and are distributed to both the cardo and the stipes(figs. 25 C, 30 B, KLcd, KLst) . In Machilis the maxillary adductorsfrom opposite appendages are united with each other medially, andappear to be detached from the ventral apophyses.The margin of the basal cavity of the maxilla (fig. 35 A) includesthe region of the cardo, the stipes, and the lacinia ; and the tergaland sternal muscles (/, /, KL) of the appendage are distributed tothese three parts. The entire base of the maxilla, therefore, has thefundamental character of a single segment, and there can be no doubtthat this segment is the true primitive base of the appendage (fig.34 A, LB) . The base of a leg appendage may be divided into a coxa
NO. 3 INSECT HEAD SNODGRASS 87
and a subcoxa (fig. 34 B, Cx, Sex), and Borner (1909, 1921) wouldidentify the cardo of the maxilla with the subcoxa of a leg. The sutureseparating the cardo from the stipes, however, terminates at bothends in the marginal rim of the maxillary base (fig. 35 A) insteadof running parallel with it; the cardo, therefore, does not have therelation of a true segment to the rest of the appendage. It is perhaps,
Fig. 35.—The relation of a maxilla to a generalized limb.A, theoretical generalized structure of a gnathal appendage, consisting of alimb base (LB), bearing a divided basendite (Lc, Ga), and a six-segmentedtelopodite, or palpus (Pip), with the principal downward flexure at the femoro-tibial joint (ft). The limb base provided with tergal promotor and remotormuscles (/, /), and sternal adductors (KL).B, showing basal musculature of a gnathal appendage analysed into thefunctional elements of the musculature of an annelid parapodium (fig. 33 D).C, third maxilliped of Apus longicaxidata, left, anterior surface, showingdivision into a limb base (LB) and a telopodite (Tip) ; the base movable on atransverse axis (a-b), the telopodite with a principal flexure (ft) between itsthird and fourth segments.a-b, basal axis of limb base; Be, basendite; ct, coxo-trochanteral joint; F,femur ; fga, flexor of galea ; flee, cranial flexor of lacinia ; flcs, stipital flexor oflacinia; ft, femoro-tibial joint; Ga, galea; /, tergal promotor; /, tergalremotor; K, sternal promotor; KL, ventral adductors (K and L united); L,sternal remotor ; LB, limb basis ; Le, lacinia ; O, levator of telopodite ; Pip,palpus (telopodite) ; Q, depressor of telopodite; Tip, telopodite; Tr, trochanters.
though, to be questioned if the subcoxa! chitinization, the pleuron, atthe base of a thoracic leg is not also a mere subdivision of the basallimb segment similar to the cardo, rather than the remains of a trueindependent segment. If the cardo does, in any sense, represent thethoracic subcoxa, it is to be noted that the hinge line between it andthe stipes has a horizontal position with reference to the axis of theappendage, and this, the writer has argued (1927), must have beenthe primitive position of the subcoxo-coxal hinge in a thoracic leg.
88 SMITHSONIAN MISCFXLANEOUS COLLECTIONS VOL. 8lThough the homology of the cardo must be left in question, thewriter would agree with Borner (1909, 1921) that the part of themaxilla bearing the lacinia, galea, and palpus represents the coxal re-gion of the leg base, and that the basal segment of the palpus is a tro-chanter (fig. 35 A, Tr). The lacinia and galea, then, are coxal endites,and, as Borner proposes, the stipes and palpifcr arc corresponding sec-ondary subdivisions of tJic coxa, or of the coxal region of the maxillarybase. In the maxillae and maxillipeds of the Crustacea, Bornerclaims, the segments bearing the lobes homologous with the insectlacinia and galea are also subdivisions of the coxa. By this interpre-tation, the palpus (fig. 35 A, Pip) becomes the telopodite of the maxil-lary appendage (B) ; its basal union with the stipes or palpifer is thecoxo-trochanteral joint (ct), and its principal distal articulation hav-ing a ventral flexure is the femoro-tibial joint (ft).Other writers have held somewhat divergent views concerning thehomologies of the maxillary segments. Goldi (1913) interpreted thecardo as the coxopodite, the stipes as the basipodite, to which heassigned the lacinia and galea as endite lobes, and the basal segmentof the palpus as the ischiopodite. Crampton (1922) gave a modifica-tion of this view in that he proposed that the palpifer represents asegment, the ischiopodite. and that the galea is an endite of this seg-ment. Uzel (1897) appears to give confirmation to this view in hisdescription of the development of the maxilla of Campodea ; the maxil-lary rudiment, he says, is first divided into an outer and an inner lobe,and then the outer lobe splits into two parts, one of which becomesthe palpus, the other the galea. If the maxilla of Campodea resemblesthat of Japyx (fig. 30 A), however, it is easy to believe that thestructure in the embryo might be misinterpreted, in as much as theadult structure is misleading until the muscle relations are taken intoconsideration (fig. 31 D) ; then it is seen that the basal region of thepalpus, which is united with the base of the galea, is the true basalsegment of the palp, and not the palpifer—clearly a secondary modi-fication.It has already been pointed out that the entire lack of muscle con-nections in the palpifer is a condition that disavows the segmentalnature of the palpifer region. Crampton's best example among in-sects of a structure corresponding- with his idea of the segmentationof a maxillary appendage is the maxilla of the larva of Sialis (fig. 31C), in which there is a small lobe (o) borne on the apparent first seg-ment of the palpus. This lobe Crampton would identify as the galea,making the supporting segment the palpifer. The muscles (O, Q) in-serted on the base of this segment, however, clearlv demonstrate that it
NO. 3 INSECT HEAD SNODGRASS 89
is the true base of the palpus {iplp)—therefore, not the palpifer.which lacks muscles—and that the lobe in question is not the galea,as also the absence of a muscle to it would indicate. The maxilla ofthe Sialis larva, then, is not a generalized appendage in the sense thatCrampton would infer, since it lacks a true galea and is provided withan accessory lobe on the first segment of the palpus. Similar lobesof the palpus segments occur in other insects, particularly in larvaeof Coleoptera ; they have a suggestion of the endite lobes of the telo-podite in such crustaceans as Apus (fig. 35 C).In the thoracic legs the limb is always flexible at the union betweenthe basis and the telopodite, i. e., at the coxo-trochanteral joint, andin no appendage, where the facts can be clearly demonstrated, isthere a union between the coxa and the trochanter. It does not seemreasonable, therefore, to suppose that the proximal segment of thetelopodite (the trochanter) should have been incorporated into thelimb base in the case of a maxillary appendage. Especially is sucha supposition unreasonable in the face of nnich specific evidence tothe contrary. The whole body of evidence l)earing on the limb mech-anism points to a primitive uniformity of flexure in all the appen-dages, whereby the limb is divided into a basis and a telopodite. andindicates that the articulation between these two parts is preserved inthe entire series of appendages, except, of course, where the telo-podite is lost.The maxillipeds and the anterior body appendages of Apus beareach five endite lobes. The first lobe (fig. 35 C, Be) is a basendite,the second is carried by the proximal division of the trochanter, thethird by the femur, the fourth by the tibia, and the fifth by the tarsus.Each endite is independently movable by muscles inserted upon orwithin its base. The maxillae of Apus are reduced to single lobes, butthe first maxilla appears to represent the rudimentary limb base withthe large basendite, since it falls exactly in line with the series ofbasal endites on the following appendages. The basal endites ofarthropod limbs in general, including the " gnathobases " of the trilo-liite appendages, the gnathal lol)es of the pedipalps in Xiphosura andArachnida, and the " styli " of the legs of Scolopendrella, are almostcertainly analogous lobes in all cases, and they must be representedby the laciniae (at least) of the insect maxillae, by the laciniae of themandibles of diplopods and chilopods, and by the incisor and molarlobes of the mandibles in all arthropods. It is a question, therefore,whether the galea of the insect maxilla is an accessory lobe of thelimb base, or a subdivision of the primary basendite (fig. 35 B, C, Be)The latter seems probable, since, in the more generalized insects, the
90 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
lacinia and galea overlap each other basally, and both are flexed bythe muscles inserted upon their bases that have their origin in thestipes.It is impossible at present to arrive at final conclusions on the manyproblems connected with the morphology of arthropod appendages,and the most that the writer would claim for the present attempt atadvancing the subject is that the material here presented gives atleast a substantial enlargement to the foundation of known facts fromwhich future work must proceed. There is no question that studentsof arthropods have given far too little attention to the relationshipbetween skeletal structure and musculature. The more the svibject islooked into, the more it will be seen that the characters of the arthropodskeleton are in large part adaptations to the strain of muscle tension,and that they are to be correctly interpreted only through an under-standing of the entire mechanism of which they are a part. Thesclerites of the insect cuticula, in particular, have been studied as ifthey were skeletal elements deftly fitted together in such a manner asto cover the outside of the animal, and we entomologists have playedwith them, as we might with the sections of a picture puzzle, withoutlooking for their significance in the mechanics of the insect. Thearthropod skeleton, it is true, has been formed from a few majorcenters of increased chitinization, but the minor " divisions " are inalmost all cases adaptations to flexion, or the opposite, namely, thestrengthening of the skeleton by the development of internal ridges.The scientific study of the comparative anatomy of insects must lookfor its advance in the future to a wider knowledge of muscles andmechanism. IV. SUMMARY OF IMPORTANT POINTS
I. The arthropods have been derived from creeping animals, notfrom forms specially modified for swimming ; their immediate pro-genitors were annelid-like in structure.
^2. The stomodeum marks the anterior end of the blastopore. Thereare, therefore, no true mesodermal segments anterior to the mouth.The unsegmented preoral part of the animal is the prostomium, andconstitutes the most primitive head, or archkephalon, of segmentedanimals, since it contains the first nerve center, or " archicerebrum,"and bears the primitive sense organs.3. The first stage in the development of a composite head in thearthropods, as represented in the embryo, comprises the prostomiumand the first two or three postoral segments. The head in this stagemay be termed the protocephalon; it is represented by the cephaliclobes of the embryo, which may or may not include the third segment.
NO. 3 INSECT HEAD SNODGRASS QI
4. The protocephalon carried the labrum, the mouth, the eyes, thepreantennae, and the antennae, also the postantennae when it includedthe third body segment.5. During the protocephalic stage of insects, as shown by the em-bryo, the thorax was differentiated as a locomotor center of the body,and the region between the head and the thorax, consisting of thefourth, fifth, and sixth body segments, became a distinct gnathalregion.6. The gnathal region was eventually added to the protocephalon toform the definitve head, or telocephalon. In the Crustacea, in whichthere was no thoracic region corresponding with that of the insects,the gnathal region was not definitely limited posteriorly, and thedefinitive head in this group may include as many as five segmentsfollowing the protocephalon. In some of the crustaceans the gnathalsegments have united with segments following to form a gnatho-thorax, leaving the protocephalon as a separate anterior head piece.In the Arachnida the protocephalon included the prostomium and twopostoral metameres, and it has combined with the following six seg-ments to form the cephalothorax.7. In the definitive insect head, the prostomium, according to someembryologists, contril)utes the clypeus and frons and the region ofthe compound eyes ; according to others it forms the clypeus andfrons only. The labrum is a median preoral lobe of the prostomium.8. The arthropod brain probably always includes the median pro-stomial ganglion, combined with the ganglia of the preantennal segmentto form the protocerebral lobes. It may still be questioned whetherthe optic lobes are derived from the prostomium or from the prean-tennal segment. The ganglia of the antennal segment form the deuto-cerebrum. The commissures of the protocerebrum and the deutocere-Ijrum are formed above the stomodeum, and unite with the archicere-bral rudiment to form the median part of the brain. The ganglia ofthe postantennal segment, when united to the preceding ganglia,become the tritocerebral brain lobes, but they, remain separate fromthe brain in some crustaceans, and their uniting commissure alwayspreserves its sub-stomodeal position.9. The prostomial region of the adult arthropod is innervated fromthe postantennal ganglia, but this is probably a secondary conditionowing to the loss of the true prostomial nerves.10. The appendages of the definitive insect and myriapod head arethe preantennae, the antennae, the postantennae, the mandibles, thefirst maxillae, and the second maxillae. Rudimentary, evanescentpreantennae have been reported only in the embryo of Scolopcndra
92 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l(Heymons) and in the embryo of Carausius (Wiesmann). Postan-tennae are commonly present in insect embryos, but their rudimentspersist in only one or two doubtful cases in the adult. The postantennalappendages are the second antennae of Crustacea, and probably thechelicerae of Arachnida and Xiphosura. Endites of their bases mayhave been the functional jaws of the insectan and myriapodan an-cestors in the protocephalic stage.11. The gnathal appendages have been derived from organs havingthe structure of uniramous ambulatory legs. All the primitive ar-thropod appendages were probably uniramous ambulatory limbs. Bira-mous and natatory appendages are characteristic of the Crustacea only,and are probably secondary adaptations to an aquatic life.12. The mandible is a common inheritance from an early ancestorof the eugnathate group of arthropods. Its primitive structure re-sembled that of the first maxilla of modern insects, and is best pre-served in the Myriapoda.13. The diplopod mandible consists of a base subdivided into cardoand stipes, bearing a large movable lacinia. but lacking a galea and apalpus. In the typical chilopod mandible, the division between cardoand stipes has been lost, and the lacinia is less free. The musculatureof the chilopod mandible is more primitive than that of the diplopodmandible. In the crustaceans and insects the mandibular lacinia iseither lost, or is fused with the base to form a solid jaw. The mandib-ular palpus is retained in many Crustacea. The mandible is repre-sented by the pedipalp in Arachnida.14. The first maxillary appendage is best developed in the insects,and probably here preserves the primitive structure of the mandible.Its musculature is exactly duplicated in the musculature of the man-dibles in the diplopods and chilopods. Neither the first nor the secondmaxillae of the chilopods gives any evidence of ever having at-tained the special structure of the primitive mandibles and the insectmaxillae.15. A primitive gnathal appendage had the structure of a gener-alized ambulatory appendage, consisting of a liinh basis and a fclo-poditc. The basis represents the coxa and subcoxa of a thoracic leg,but its division into cardo and stipes is not a true segmentation. Thegalea and lacinia are movable endites of the basis, with the originof their muscles in the stipital region of the latter. The telopoditebecomes the palpus of the gnathal appendage, and its basal articula-tion is the homologue of the coxo-trochanteral joint in the leg. Thepalpifer is not a segment of the limb, but a subdivision of the stipes
NO. 3 INSECT HEAD SNODGRASS ()3bearing the palpus and the galea (as claimed by Borner) ; the musclesof the palpus and the galea pass through the palpifer, but never arisewithin it.i6. The primitive appendage was implanted in the soft lateral wallof its segment, and turned forward and backward on a vertical axisthrough its base, as does an annelid parapodium. The first joint setoff the telopodite, and gave the latter a mobility in a vertical ])lane.17. The primitive muscles inserted on the base of a generalized limb,as on an annelid parapodium, consisted of a dorsal promotor and a re-motor, and of a ventral promotor and a remoter. When tergal plateswere developed, the gnathal appendages of the arthropods becameattached to their lateral margins, each by single point of articulation.The ventral muscles of the appendages then became sternal adductors.18. The points of origin of the ventral adductors of the gnathalappendages in myriapods and insects were probably crowded togetherwhen the gnathal segments were added to the protocephalon. Theyhave since become supported on a pair of apophyses arising at firstfrom the base of the hypopharynx. In the myriapods and in most ofthe apterygote insects, the apophyses still maintain their hypopharyn-geal connections : but in the pterygote insects their l)ases have migratedlaterally to the margins of the cranium, and in all but some of thelower forms have finally moved to a facial position in the epistomalsuture. Their posterior ends have united with the transverse ten-torial bar developed in the back part of the head. The hypopharyngealapophyses of the Myriapoda and Apterygota have thus come to bethe anterior arms of the pterygote tentorium.19. The adductor muscles of the insect maxillae, arising on thetentorium, are the sternal adductors of the appendages, correspondingwith the sternal adductors or rotators of the thoracic legs, and arederived from the primitive ventral promoters and remotors of thelimb.20. The ventral adductors of the mandibles in the Chilopoda re-tain their connections with the sternal, or hypopharyngeal, apophyses.In the Diplopoda, Crustacea, and Apterygota, groups of the adductorlil)ers from the mandibles have lost their sternal connections and haveunited with each other by a median ligament to form a dumb-bellmuscle between the two jaws. Other groups of fibers may retain theirconnections with the apophyses, or direct with the base of the hypo-I)harynx. In the Pterygota, the ventral adductors of the mandibles havebeen lost, except for a few rudiments in some of the lower orders.21. The mandible of Lcpisma and of pterygote insects is hingedto the head on a long base line with anterior and posterior articulations.
94 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lThe posterior articulation is the primitive one, the anterior a secon-dary one. By this change in the articulation and movement of thejaw, the primitive tergal promotor muscle becomes an abductor, andthe primitive tergal remotor becomes an adductor. The base line ofthe mandible slopes downward and forward in Lepisina and in a fewof the lower pterygotes ; in all others its slope is reversed, allowingthe tip of the jaw to swing inward and posteriorly during adduction.A similar evolution of the mandible has taken place in the Crustacea.22. The ridge of the base of the insect cranium, on which the pro-thoracic and neck muscles are inserted, is probably a chitinization ofthe intersegmental fold between the maxillary and labial segments.The posterior tentorial arms arise from its ventro-lateral ends by in-vaginations in the external suture. The neck of the insect, therefore,may be unchitinized parts of both the labial and the prothoracicsegments. V. THE HEAD OF A GRASSHOPPERAfter laying down the general principles worked out in the pre-ceding sections, it will be well to test them with a few specific examples.The head of a grasshopper is a good subject for an elementary studyof the structure of the pterygote insect head, because it preserves thegeneralized orientation in having the face directed forward and themouth appendages hanging downward. Terms of direction, therefore,do not have to be qualified
—
ventral is downward, dorsal is upward,and anterior is forward. The descriptions here given are based onthe Carolina locust (Dissosteira Carolina), a fairly large grasshopperto be obtained in almost any part of the United States.The muscles are designated numerically for convenience of refer-ence only, and the same numbers do not refer to corresponding musclesin the grasshopper and in the caterpillar (Section VII) . The myologyof insects is as yet too little advanced to furnish a satisfactory generalnomenclature for insect muscles, and no attempt is made here to usea set of names for the muscles of the grasshopper that could in allcases be applied to the muscles of other insects. The usual method ofnaming muscles according to their function, or their supposed func-tion, gives terms fitting for the species described; but in many cases,by a change in the articulation between the skeletal parts involved,muscles that are clearly homologous have their functions completelyaltered. Again, it is impossible to name muscles consistently accordingto their points of origin and insertion, for either end of a muscle mayshift and may migrate into a territory quite foreign to its original
NO. 3 INSECT HEAD SNODGRASS 95
connections. A third feature disturbing to a uniform muscle nomen-clature is the fact that any muscle may break up into groups of fibers,or, at least, a single muscle in one species may be represented func-tionally by several muscles in another. Finally, there are muscles thatare evidently new acquisitions developed in connection with specialmechanisms. The importance of the study of musculature for theunderstanding of the insect skeleton, however, can not much longerbe ignored. STRUCTURE OF THE CRANIUMThe walls of the head in the grasshopper are continuously chitinizedon the anterior, dorsal, and lateral surfaces (fig. 36 A, B), and theAnt Vx ocs Oc
Fig. 36.—Head of a grasshopper, Dissosteira Carolina.A, lateral. B, anterior. C, posterior.a, posterior articulation of mandible; Ant, antenna; at, anterior tentorial pit;r, anterior articulation of mandible; Clp, clypeus ; cs, coronal suture; cv, cervicalsclerites ; E, compound eye ; c. articulation of maxilla with cranium ; cs, epistomalsuture; /, articulation of labium with cranium; For, foramen magnum ;_ /.y,frontal suture; g. condyle of postocciput for articulation with cervical sclerite;Gc, gena ; h, subocular ridge ; /, frontal carina ; ;, subantennal suture ; k, flexiblearea between lower edge of gena and base of mandible ; Lb, labium ; Lm, labrum
;
Md, mandible; Mx, maxilla; O, ocellus; Oc, occiput; ocs, occipital suture;Pgc, postgena ; Poc, postocciput ; PoR, postoccipital ridge ; pos, postoccipitalsuture ; PT, posterior arm of tentorium ; pt, posterior tentorial pit ; sgs, subgenalsuture ; Vx, vertex.dorsal and lateral walls are reflected upon the posterior surface (C)to form a narrow occipito-postgenal area surrounding the foramenmagnum (For). The foramen is closed below by the neck membrane,in which is suspended the base of the labium (Lb). The ventral wallof the head, between the bases of the gnathal appendages, is occupiedinostly by the large median hypopharynx, it being otherwise reducedto the narrow membranous areas between the lateral margins of thehypopharynx and the bases of the mandibles and maxillae.
96 SMITHSONIAN MTSCl'lLLANEOUS CULI.ECTIONS VOL. 8[The facial aspect of the cranium is distinctly separated by theepistomal suture (fig. 36A, B, r^-) from the clypeus, but there is nodemarked frontal sclerite. The apex of the f rons, however, is definedin Dissostcira by two short remnants of the frontal sutures (B, fs)diverging- from the end of the coronal suture (cs). The facial areaof the head is limited on each side by an impressed line (/{) extendingfrom the lower angle of the eye to the anterior articulation of themandible. The median part of this area forms a broad frontal cosfa,margined laterally by a pair of sinuous carinae (/') reaching from thetop of the head to the lower part of the face. A short, transversesuhantcnnal suture (j) lies on each side of the frontal costa justbelow the level of the median ocellus. The inner ridges of thesesubantennal sutures have a close relation to the attachments of themore important muscles of the frons (fig. 38 D, ;'). The true frontalarea of the grasshopper, therefore, must include the region of thesesutures and extend dorsal to them between the bases of the antennaeinto the angle between the short remnants of the frontal sutures.The lateral areas of the head (fig. 36 A) have no special character-istics. The subgenal suture (sgs) on each side is continuous anteriorlywith the epistomal suture (es). The compound eye is surrounded bya distinct suture forming a high ridge internally (fig. 39 A, OR), andsetting oft" a narrow rim, or ocular sclerite, around the base of theeye (fig. 36, A, B, C).On the posterior surface of the head (fig. 36 C), the occipito-postgenal area (Oc, Pge) is included between the well-marked occip-ital suture {ocs) and the postoccipital suture {pos). In Dissosteirathe occiput and postgenae are continuous ; in Mclanoplus the occipitalarch is separated from the postgenae by a short groove on each sideon a level with the lower angles of the compound eyes. Posteriorto the postoccipital suture is the postoccipital rim of the head (Poc),widened above and below on each side, to which the membrane of theneck is attached. The postoccipital suture forms internally the ridgeon which the muscles of the neck and prothorax that move the headare inserted (fig. 45 A, PoR). Laterally the postoccipital ridge iselevated as a high plate (fig. 36 C,Fo/?), from the ventral ends ofwhich the posterior arms of the tentorium (PT) proceed inward. Theroots of the tentorial arms appear externally as long open slits in thelower ends of the postoccipital suture (pt, pt).The clypeus and labrum form together a broad free flap (fig. 36 A,B, Clp, Lni) hanging before the mandibles from the lower edge of thefrontal region. The fronto-clypeal, or epistomal, suture (es) is adeep groove forming internally a strong epistomal ridge (fig. 39 A, kMl
NO. 3 INSECT IIEAl -SNODCKASS 97B, C, ER), from the lateral parts of which arise the anterior tentorialarms {AT). The roots of these arms appear externally as lateralslits in the epistomal sutm-e (figs. 36 B, 37 A, at, at), just mesadto the anterior articulations of the mandibles {c, c). The clypeus ofDissostcira is partially divided by transverse lateral grooves intoanteclypeal and postclypeal areas. The labrum is a 1)road oval plate,notched at the middle of its ventral margin, freely movable on thelower edge of the clypeus. A median quadrate area on its basal halfis limited below by a sinuous transverse groove (fig. 37 A) that formsa low ridge on the inner surface of the anterior wall (B, /.) On the
AT
Clp—
-Lm— TorA BFig. 2)7-—Clypeus and labrum of Dissostcira Carolina.A, anterior surface. B, posterior surface of anterior wall, showing muscleattachments, and bases of anterior tentorial arms.AT, anterior arm of tentorium; at, anterior tentorial pit; c", anterior articula-tion of mandible ; Clp, clypeus ; es, epistomal suture ; /, ridge of anterior labralwall; Lm, labrum; Tor, torma ; /. labral compressors; ^, anterior retractors oflabrum; 3, posterior retractors of labrum; 34, 35, points of origin of anteriordilators of buccal cavity (figs. 41, 44).posterior surface of the clypeo-labral lobe, the area of the clypeus isseparated from that of the labrum by two small chitinous bars, thetonme (figs. 37 B, 42 A, Tor). A median Y-shaped thickening (fig.42 A, 7/0 of the cuticula of the labrum makes a ridge on the innersurface of the posterior labral wall.The clypeus has no muscles for its own movement, but the firsttwo pairs of anterior dilators of the buccal cavity (figs. 41, 44, 33^ Si)are inserted on the inner surface of its anterior wall (fig. 37 B, jj,34) . The labrum is provided with three sets of muscles, as follows
:
I.—Compressors of the labrum (fig. 37 B).—A pair of short mus-cles arising medially on transverse ridge of anterior labral wall (/) ;diverging to arms of Y-shaped ridge in posterior wall (fig. 42 A. m).
98 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81
2.—Anterior retractors of the labrum (fig. 37 B).—A pair of longmuscles arising on subantennal ridges of frons (fig. 38 D, ;") ; con-verging downward to insertions on base of anterior wall of labrum(%-37B).
_J.
—
Posterior retractors of the labrum (fig. 37 B).—A pair of longmuscles arising on subantennal ridges of frons, each laterad of 2 (fig.38 D) ; inserted on dorsal processes of tormae at base of posteriorwall of labrum (figs. 37 B, 42 A).The articulations of the gnathal appendages occupy typical positionsalong the lower lateral margins of the cranium. The mandible articu-lates anteriorly with a condyle (fig. 39 A, c) supported at the junctionof the epistomal and subgenal ridges {ER, SgR), but projecting onthe external surface of the head. Posteriorly the jaw articulates witha facet on the ventral edge of the postgena (figs. 36 C, 39 A, C, a).This articulation as the first is outside the membranous connectionof the mandible with the head. The axis of the mandible slopesstrongly downward and posteriorly between the articular points. Thelateral edge of the mandibular base is separated from the margin ofthe gena by a narrow, flexible strip of weakly chitinized articularmembrane (fig. 36 A, k), at the ventral margin of which arises theabductor apodeme of the mandible (fig. 39 D, 8 Ap).The maxilla articulates by a single point on the base of the cardowith a shallow facet on the edge of the postgena (figs. 36 C, 40 C, c)almost directly below the posterior tentorial pit {pt). The maxillaryarticulation is thus crowded unusually far posteriorly in the grass-hopper. In most generalized insects it lies well before the line of thepostoccipital suture, as in a roach or a termite, and is often muclifarther forward.The labium is loosely articulated by the elongate basal angles of tbesubmentum with the posterior margin of the postocciput at points ashort distance above the posterior lower angles of the latter (figs. 36C, 40 C, /).The tentorium of the grasshopper has the form of an X-shapedbrace between the lower angles of the cranial wall (fig. 39 B). Theanterior arms {AT) arise from the lateral parts of the epistomal ridge(ER), their broad bases extending from points above the mandibulararticulations half way to the median line of the face. In this respectDissosteira shows an advance over Periplaneta. in which the bases ofthe anterior tentorial arms arise from the subgenal ridges and extendonly a short distance mesad of the mandibular articulations. 'V\vposterior tentorial arms of Dissosteira (fig. 39 B, PT) arise from thelower ends of the postoccipital ridge (fig. 45 A. PoR). The median
NO. 3 INSECT HEA SNODGRASS 99body of the tentorium is concave below (fig. 39 B, C, Tnt). A thin,flat dorsal arm of the tentorium (fig. 39 C, DT) arises from the baseof the inner end of each anterior arm and extends upward and anter-iorly to the wall of the cranium just before the lower angle of thecompound eye. The dorsal tentorial arms are attached to the hypo-dermis of the head wall, but make no connection with the cuticulain Dissosteira. THE ANTENNAEEach antenna consists of two larger basal segments, and of a longslender flagellum broken up into about 24 small subsegments. In
4a
ii\ ^-tb ER DFig. 38.—Antenna of Dissosteira Carolina and of Periplaneta.A, base of left antenna of Dissosteira, ventral surface. B, the _ same _ ofPeriplaneta. C, base of right antenna and antennal muscles of Dissosteira,dorsal view. D, base of right antenna, antennal muscles, anterior tentorial arm,muscles arising on frons, and associated structures of Dissosteira, interior view.SAp, apodeme of depressor muscles of antenna; AR, antennal ridge; At,anterior arm of tentorium ; c, anterior articulation of mandible ; DT, dorsalarm of tentorium; ER, epistomal ridge; es, epistomal suture; ;, subantennalridge; n, pivot of antenna; OR, ocular ridge; Pdc, pedicel; Sep, scape.Dissosteira the antenna of the male is a little longer than that ofthe female. Of the two basal segments, the proximal one, or scape(fig. 38 A, Sep), is the larger. It is articulated to the rim of theantennal socket by a small process on the lateral ventral angle of itsbase that touches upon the margin of the socket {n). The motion ofthe scape on the head, however, is that of a hinge joint moving in avertical plane on a transverse axis. The base of the scape is provided
]00 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
with levator and depressor muscles (C, D). In other insects theantenna is more commonly pivoted on a ventral point of articulationwith the rim of the socket, as in Periplaneta (fig. 38 B, n), and thus hasgreater freedom of movement. As already noted, how^ever, the artic-ular point may be dorsal, as in Japyx, and in the Chilopoda (fig. 23B, n). The thickened rim of the antennal socket (fig. 38 D, AR) isbraced by a short arm against the anterior margin of the heavy cir-cumocular ridge {OR). A crescentic area of the head v^^all just aboveand mesad to the antennal socket is depressed externally, and theinflection tilts the place of the antennal socket somewhat dorsally,giving the antenna a more upward play than it otherwise would have.The second basal segment of the antenna, the pedicel (fig. 38 A,B, Pdc), is movable in a horizontal plane on the end of the scape bymeans of muscles arising within the scape. The other segments of theantenna are flexible but have no muscles.The muscles of the antenna comprise muscles inserted on the baseof the scape that move the antenna as a whole, and the muscles of thepedicel that move the pedicel and flagellum. They are as follows
:
4.—Levators of the antenna (fig. 38 C, D).—Two muscles arisingon tentorium, one (D, 4a) on dorsal arm, the other {4h) on anteriorarm ; both inserted by a short tendon on a lobe of dorsal side of baseof scape (C, D).5.
—
Depressors of the antenna (fig. 38C, D).—Two muscles aris-ing on dorsal arm of tentorium (D, ^a) and on anterior arm {^b) ;both inserted on a long slender tendon arising near ventral margin ofscape (A, sAp) in articular membrane of antenna.6.—Extensor of the flagellum (fig. 38 C).—Arises dorsally andmedially in base of scape ; inserted medially on base of pedicel.7. Flexor of the flagellum (fig. 38 C).—Arises dorsally and later-ally in base of scape ; inserted laterally on base of pedicel.THE MANDIBLESThe mandible of the grasshopper is a strongly chitinous jaw—
a
short, hollow appendage with triangular base, thinning down to thecutting margin. The anterior and the posterior angles of the lateralbase line carry the articular points with the head, and the apodemeof the adductor muscles arise at the median angle.The distal edge of each mandible presents an incisor and a molararea. The first (fig. 39 D, 0) forms the compressed and toothedapical part of the jaw, the second (/>) forms a broad grinding surfaceon the anterior median face closer to the base of the mandible. The in-cisor and yiolar areas are not exactly alike on the two jaws, each being
( NO. 3 I XSKCT ]T RAD SNULHiKASSbest developed on the right. The molar area of the right mandibleconsists of strong, heavy ridges forming a projecting surface ; theridges of the left jaw are low and their area does not project. Thetwo molar surfaces, therefore, fit one upon the other without interfer-ence when the jaws are closed. The incisor lobes of the mandiblesclose upon the ventral end of the hypopharynx, the molar surfacesover its base, and the anterior contour of the hypopharynx is modeled
AR
Fig. 39.—Internal structure of the head of Dissosfcirn caroJ'uia, and themandible and its muscles.A, inner surface of right half of epicranium. B, tentorium and lower marginof epicranium, ventral view. C, iimer view of right half of head, with rightmandible and its muscles in place. D, right mandible, postero-mesal view.a, posterior articulation of mandible ; SAp, abductor apodeme of mandible
;
gAp, adductor apodeme of mandible; AR, antennal ridge; AT, anterior tentorialarm ; at, anterior tentorial pit ; c, anterior articulation of mandible ; Clp, clypeuscv, cervical sclerite; DT, dorsal tentorial arm; E, compound eye; ER, epistomalridge ; es, epistomal suture ; Fr, frons ; g, condyle of articulation of cervicalsclerite ; /;, subocular ridge ; /, subantennal ridge ; Lm, labrum ; Md, mandibleO, ocellus ; 0, incisor lobe of mandible ; p, molar area of mandible ; Poc, post-occiput ; PoR, postoccipital ridge; PT, posterior tentorial arm; pt, posteriortentorial pit; SgR, subgenal ridge; Smt, submentum ; Int, body of tentorium.
according to the irregularities of the mandibular surfaces. The pos-terior slope of the mandibular hinge lines cause the points of thejaws to turn inward, upward, and posteriorly during adduction. Atthe base of each molar area of the mandibles a flat brush of hairs (fig.39 C, D) projects inward, and the two brushes come together anteriorto the mouth opening when the mandibles are closed, serving thusevidently to prevent the escape of masticated food material from be-tween the jaws. The anterior surfaces of the mandibles are overlappedby the epipharyngeal surface of the clypeus and labrum, and the
I02 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
asymmetry of the mandibular surfaces and contours is reflected inthat of the epipharyngeal surface (fig. 42 A).The mandibles of Dissosfcira are moved, so far as the writer coulddiscover, only by tergal abductor and adductor muscles, which, as al-ready explained, are the primitive tergal promotors and remotorstransformed in function by the change from a monocondylic to adicondylic articulation in the mandible (fig. 29 A, B, C). Smallventral adductors of the mandibles arising on the hypopharynx andon the tentorium persist in some of the Tettigoniidae ( figs. 20 D,KLh, 2gC,KL}i, KLt), but these muscles appear to be lost in theAcrididae, as they are in all higher pterygote insects. The fibers of thefunctional abductors and adductors arise on the walls of the craniumand are inserted on flat apodemal plates of the jaws. The abductorapodeme is a small plate (fig. 39 D, 8Ap) arising from the articularmembrane close to the outer margin of the mandibular base and nearthe posterior articulation (a). The adductor apodeme (pAp) con-sists of two large thin plates borne upon a common stalk, whicharises from the articular membrane at the inner angle of the mandib-ular base, and lies in the lateral angle between the anterior and pos-terior arms of the tentorium (C). One plate extends dorsally in alongitudinal plane, the other, which is smaller, lies in a transverseplane. Each mandibular apodeme is a chitinous invagination from thearticular membrane close to the base of the jaw. The muscles of themandible correspond with the apodemes. They are as follows
:
8.—Abductor of the mandible.—A small fan of fibers, arising onventral part of postgena and on extreme posterior part of ventral halfof gena ; inserted on abductor apodeme of the mandible.p.
—
Adductors of the mandible (fig. 39 C).—Two sets of fiberscorresponding with the two divisions of the adductor apodeme. Thefibers of one set (pa) arise on dorsal wall of cranium, from a pointbetween compound eyes to occiput, with one bundle attached on post-occiput (Poc) ; inserted on both sides and on posterior margin of themedian apodemal plate. Those of the other set (pb), inserted on thetransverse plate of the apodeme, arise on lateral walls of craniumfrom subocular ridge (h) to postgena, and some of the posteriorfibers encroach upon outer end of posterior tentorial arm.THE MAXILLAEThe maxilla of the grasshopper (fig. 40 A) is so similar to that ofthe roach (fig. 25 A), already described, that its major features willneed no special description. It consists of a triangular cardo (fig.40 A, Cd). a quadrate stipes (St), with a well-developed ])alpifer
NO. INSECT HEAD SNODGRASS 103
Fig. 40.—Maxilla and labium of Dissosteira Carolina.A right maxilla, posterior surface. B, left maxilla, anterior view, exposingmuscles of cardo and stipes. C, posterior region of cranium, with cervicalsclerites and maxilla, left side. D, labium and its muscles, posterior view. L,-stipes and palpifer with bases of palpus, galea and lacmia, lacmial muscles re-moved, anterior view. . , / ^ r „^,i^.a, posterior articulation of mandible; loAp, apodeme of promotor of _ cardoCd cardo ; cv, cervical sclerite ; e, articulation of maxilla with cranium ; t,articulation of labium with cranium; Ga, galea; Gl, glossa; Lc, lacmia; Mtmentum; Mx, maxilla; Oc, occiput; ocs, occipital suture; P^, postgena ; /^//,palpifer; Pig, palpiger ; Pip, palpus; Por, postocciput; pos, postoccipital suture;^^ posterior tentorial pit; q, suture and internal ridge near inner margin otstipes; r, internal ridge of cardo; s, apophysis of cardo for muscle inseition,SID, salivary duct; Suit, submentum ; St, stipes; t, suture and internal ridgeseparating p'alpifer from stipes ; Tut, body of tentorium : u, inner ridge at baseof posterior wall of galea ; v, keel of salivary cup.
104 SMITHSONIAN JtllSCIiLLANEOUS COLLECTIONS VOL. 8l{Plf), and two terminal lobes, lacinia (Lc) and galea (Ga), and afive-segmented palpus (Pip).The cardo presents an irregular topography on its external surface,and is marked into several areas by the lines of a strong branchingridge on its internal surface (fig. 4oB,r). Crampton (1916) callsthe part proximal and posterior to the ridge the juxtacardo and therest of the sclerite the vcracardo, but the inference that these areasare " divisions " of the cardo is scarcely warranted, since the ridge isclearly a mere strengthening device. The articular point {c^ of thecardo with the cranial margin is a knob on the posterior angle of itsbase, anterior to which is a long arm to which is attached the apodeme( loAp) of the promotor muscle (C, 10) . A pit in the distal part of theexternal surface of the cardo (A, s) marks the site of an internalprocess on which one of the adductor muscles is inserted (B, iia).The distal margin of the cardo is articulated by a long, flexible hingeline with the base of the stipes, but there are no muscles extendingbetween the cardo and stipes.The quadrate stipes (fig. 40 A, St) has a strong plate-like ridgeon its internal surface near the inner margin (g), on which is in-serted one of the adductor muscles (E, 12). Crampton distinguishesthe body of the stipes as the vcrastipcs, and the flange mesad of themuscle-bearing ridge as the jnxtastipcs. The region of the palpifer(A, Plf) is well separated from that of the stipes by an internal ridge(E, /), but the muscles of the palpus ( //, 18), as well as the muscleof the galea {16). have their origin in the stipes, suggesting that thepalpifer is a subdivision of the stipes, and not a basal segment of thepalpus.The lacinia (fig. 40 A, B, Lc) is borne by the distal end of the stipes,and is capable of flexion anteriorly and posteriorly on an oblique axiswith the latter. Distally it tapers and ends in two claws turned in-ward. The lacinia is flexed by a pair of strong muscles arising withinthe stipes (B, i^^a, 13b), and by a slender muscle (/./) having itsorigin on the wall of the craniuuLThe galea (fig. 40 A, Ga) is carried by a distal subdivision of thepalpifer, which Crampton (1916) calls the basigalea. In form, thegalea (A, B, C, Ga) is an oval, flattened lobe ; its walls are but weaklychitinized. Its inner margin lies against the lacinia, and its outersurface is modeled to fit the outer part of the posterior surface ofthe mandible, against which it can be tightly closed. The base of thegalea is marked on the posterior wall by an internal ridge (E, m), uponwhich is inserted its single flexor muscle {j6).
f so. 3 INSECT llRAl) SNODGKASS I05The maxillary palpus consists of five segments (lig. 40 A, B, C.Pip). The basal segment is provided with levator and depressor mus-cles (B, E, I/, 18) arising within the stipes; each of the other seg-ments has a single muscle arising in the first or second segment proxi-mal to it.The muscles of a maxilla are as follows
:
10.—Promotor of the cardo (fig. 40 C).—A small fan of fibersarising on lower posterior part of postgena, external and anterior tothe mandibular abductor ; inserted on slender apodeme of basal armof cardo.//.
—
Adductors of the cardo (fig. 40 B).—Two muscles arising onposterior end of anterior arm of tentorium, extending ventrally, pos-teriorly, and outward; one {iia) inserted on process {s) of innerface of cardo, the other (iih) mesad to distal end of ridge (r) ofcardo.12.—Proximal adductor of the stipes (fig. 40 E).—Arising on ex-treme posterior end of anterior arm of tentorium; inserted on ridgeof inner margin of stipes.7j. Distal adductors of the stipes (fig. 40 E).— Two musclesarising on tentorium, the first (13a) a slender muscle arising, alongwith iia, lib, and 12, on posterior end of anterior arm of tentorium,the second (isb) a large, thick, digastric muscle arising laterally onconcave ventral surface of body of tentorium ; both muscles insertedon a slender apodeme attached to inner distal angle of stipes.Muscles II, 12, and 13 correspond with the adductors of the cardoand stipes that in Apterygota arise on the hypopharyngeal apodemes(fig. 30 B, KLcd, KLst), representing the sternal adductors, or ster-nal promotor and remotor, of a primitive appendage (fig. 35 B, A', L).14.—Cranial flexor of the lacinia (fig. 40 B).—Arises on gena justbefore upper end of promotor of cardo (C, 10) ; inserted on innerangle of base of lacinia. This muscle is the homologue of the cranialflexor of the maxillary lacinia in Apterygota (fig. 30 B, /?cc), and ofthe corresponding flexor of the mandibular lacinia in Myriapoda (fig.26 A, B, Cflcc).75. Stipital flexor of the lacinia (fig. 40 B).—A large two-branched muscle arising in base of stipes, one branch {13a) medially,the other (ifib) in outer basal angle ; both inserted on anterior marginof base of lacinia. These muscles flex the lacinia forward. Berlese
( 1909) describes the posterior branch of this muscle in Acridium asattached to the posterior wall of the lacinia, and as being an antagonistto the anterior branch, but in no insect has the writer observed anantagonist to the lacinial flexor.
I06 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
i6.—Flexor of the galea (fig. 40 B, E).—A large muscle arisingmesally in base of stipes, external to lacinial muscles and depressor ofpalpus; inserted posteriorly on ridge (E, u) at base of galea. Thismuscle probably flexes the galea forward and inward, the point of flex-ion being at the base of the subgalea.//.
—
Levator of the maxillary palpus (fig. 40 B, E).—Origin inmedian basal part of stipes ; insertion on dorsal margin of basal seg-ment of palpus.18.—Depressor of the maxillary palpus (fig. 40 B, E).—Origin oninner edge of stipes; crosses anterior to muscle of galea {16) to in-sertion on ventral margin of basal segment of palpus.If the basal segment of the palpus (fig. 35 A) corresponds withthe trochanter of the leg (fig. 34 B, Tr), then muscles ij and iSrepresent the levator and depressor of the telopodite (fig. 34 A, 0,Q) arising in the coxal region of the leg base (LB).ip, 20, 21, 22.—Muscles of the maxillary palpus (fig. 40 B).—
A
single muscle for each segment, the first (ip) a levator of second seg-ment, the second {20) a productor of third segment, the third (21)a depressor (adductor) of fourth segment, the fourth {22) a reductorof terminal segment.The joint between the third and fourth segments of the palpusapparently represents the femero-tibial flexure of a leg (figs. 34, 35A, ft), the two small basal segments of the palpus being trochanters.THE LABIUMThe labium of the grasshopper (fig. 40 D) is simple in construction,and typical of the labium of biting insects, except in the reduction ofthe glossae. It consists of a largei submentum (Smt) with the elongatebasal angles loosely attached to the posterior margin of the craniumbehind the roots of the posterior tentorial arms (C, /). The mentum(D, Mt) is broad, with imperfectly differentiated palpus-bearinglobes, or palpigers (Pig), at the sides of its base. On its ventralmargin the mentum bears a pair of large flat lobes, the paraglossae(Pgl), with a pair of rudimentary glossae (Gl) between them. Eachpalpus is three-jointed.At the base of the anterior surface of the mentum, where the wallof the mentum is reflected into that of the hypopharynx (fig. 41),there is a small, median, oval, cup-shaped depression into which opensthe duct from the salivary glands (SID). A small prominence on thebase of the hypopharynx fits into the salivary cup and apparentlycloses the latter when the labium is pressed against the hypopharynx.
NO. 3 INSECT HEAD SNODGRASS loyThe walls of the salivary cup are chitinous, and its posterior innersurface bears a strong chitinous keel (figs. 40 D, 41 z') projectinginto the interior of the labium in the base of the mentum. Two pairsof muscles (figs. 40 D, 26, 2y) are attached upon the keel and thewalls of the salivary cup.The musculature of the labium is in general similar to that of themaxillae. It includes the following muscles
:
i'J.
—
Proximal retractors of the mentum (fig. 40 D).—A pair ofmuscles arising on ventral surfaces of posterior tentorial arms ; in-serted on lateral basal angles of mentum.24.—Distal retractors of the menttim (fig. 40 D).^—A pair of mus-cles arising on posterior surfaces of posterior tentorial arms ; extend-ing through submentum and mentum to be inserted on anterior wallof labium at inner basal angles of the glossae. The distal parts ofthese muscles are not seen in figure 40 D, being covered posteriorlyby muscles 2^ and 2j. The labial muscles 2^ and 24 evidently cor-respond with the tentorial adductors of the maxillae (E. 12, jj).2J.—Flexors of the paraglossae (fig. 40 D).—A pair of large mus-cles arising in lateral basal angles of mentum ; inserted on bases ofparaglossae, to posterior walls, near inner ends. Each of these musclescorresponds with the flexor of the galea in the maxilla (E, 16).The small labial glossae of Dissosteira have no muscles.26, 2/.—Muscles of the salivary cup (fig. 40 D).—Two pairs ofmuscles: one pair (26) arising on basal angles of mentum, converg-ing to insertions on keel of salivary cup ; the other pair (<?/) arisingon posterior wall of mentum near bases of palpi, converging proxi-mally to insertions on sides of salivary cup. These muscles apparentlyhave no homologues in the maxillae ; perhaps they are special labialmuscles having something to do with the regulation of the flow ofsaliva from the salivary duct.28.—Levator of the tahial palpus (fig. 40 D).—Origin in lateralbasal angle of mentum ; insertion on dorsal rim of base of palpus.2Q.—Depressor of the labial palpus (fig. 40 D).—Origin in distalmedian angle of mentum ; insertion on ventral rim of base of palpus.30, 5/.
—
Muscles in the labial palpus (fig. 40 D).—The first (30)a levator of second segment; second (5/) a depressor (adductor) ofthird segment.THE PREORAL CAVITY AND THE HYPOPIIARYNXThe intergnathal space, or preoral cavity, of the grasshopper (fig.41, PrC) is of large size, but it is mostly filled by the thick, tongue-like hypopharynx suspended from its roof (Hphy). Its anterior wall
io8 SMITHSONIAN M JSCi;i.LANEOUS COl.r.KCTli )\S VOi.. 81
is the posterior surface of the clypeiis and labruni {Lip, Lin ). whichin the grasshopper is not produced into a specially developed lohe,or epipharynx. The lateral walls are the inner faces of the mandiblesand maxillae ; the posterior wall is the anterior surface of the labium{Lb). The dorsal wall of the cavity represents the true sternal regionof the head, sloping downward and posteriorly from the mouth open-ing to the base of the labium. It is mostly produced into the large,
HphyFig. 41.—Stomodeum, and its dilator muscles in right half of head ofDissostcira Carolina.Clp, clypeus ; Cr, crop ; cs. epistonial suture, Fr, frons ; Hphy, hypopharynx,Lb, labium; Lm, labrum ; Mth, mouth; Fhy, pharynx; Pre, preoral cavity; SID,salivary duct ; Tnt, tentorium ; v, salivary cup.
median hypopharynx {Hphy), leaving otherwise only a narrow mem-branous area on each side between the base of the hypopharynx andthe bases of the mandible and maxilla.The anterior or epipharyngeal wall of the preoral cavity presentsa number of features of special interest (fig. 42 A). The lateralparts of the labral region of this wall are concave and fit closely overthe smooth, rounded, anterior surfaces of the mandibles. The bandsof hairs directed inward on the labral surface guard the exits from
NO. 3 INSECT HKAL) SxXOlXIKASS IO9
l)et\veen the mandibles, and the asymmetrical forms of the hair-covered areas here correspond with the different shapes of the twomandibles. Minute sense organs are scattered over this labral surface,especially on the bare lateral regions. A special group of similar butsomewhat larger sense organs lies at each side of the notch in theventral border of the labrum. The median area of the basal half ofthe labral surface forms a low elevation, the sides of which are thicklycovered with long spine-like hairs curved inward and upward. Thiselevation projects between the inner edges of the closed mandibles,and its irregular contours fit with the lines of the opposing jaws. Itsmedian surface is depressed and embraces the region of the internalY-shaped ridge ( /// ) . The elevation is continued upward on the clyp-cal region, above the spreading arms of the Y-shaped ridge, and be-tween the inner recurved ends of the tormae (Tor), and then intothe mouth (Mth) and upon the anterior wall of the buccal cavity. Asinuous groove begins upon the elevation ventrally between the tormae,which extends dorsally and enlarges into a deep, median channel con-tinued into the anterior wall of the mouth and pharynx. At the sidesof the lower end of the channel, between the slender arms of thetormae, are four asymmetrically placed, oval groups of small peg-likesense organs with large circular bases, partly covered from the sidesby fringes of long recumbent hairs.The hypopharynx is a large median lobe suspended, as already noted,from the ventral wall of the head between the mouth and the baseof the labium (fig. 41, HpJiy). Its posterior end is closely coveredby the paraglossal lobes, and its sides are concealed by the mandiblesand maxillae. In form, as seen from below (fig. 42 C), the hypo-])harynx is somewhat ovate, with the smaller end anterior, but its pos-terior end is set off as a narrowed lobe by lateral constrictions. Thelateral surfaces of the anterior division fit into the posterior concavi-ties of the mandibles, those of the posterior lobe are embraced by theconcave inner faces of the laciniae. The posterior, basal extremity ofthe hypopharynx projects as a small median process into the salivarycup on the base of the labium (tig. 41 ). The lateral line of the hypo-pharyngeal base is marked bv a slender, sinuous, chitinous bar on eachside (zv). The arrangement of the hairs clothing the hypopharynx issufficiently shown in the figures (figs. 41. 42 C). On its sides andat the posterior end near the salivary cup are a few small sense organssimilar to those of the labrum.Dorsal to the anterior end of the hypopharynx is an area that leadsdirectly upward into the floor of the mouth. It possibly represents thesternal region of the protocephalic segments of the head. On itsk
no SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
median surface (fig. 42 B, C) is a ridge, bordered by long hairs di-rected inward and upward, that continues dorsally from the narrowedend of the hypopharynx, and which is excavated by a median channelwhere it enters the mouth. At each side of this channel is an ovalgroup of sense organs. Flanking the ridge are two chitinous bars(HS), the ventral ends of which articulate with the anterior extrem-ities of the lateral basal rods of the hypopharynx (w). Dorsallyeach bar forks into two arms, of which one (x) goes posteriorly to
Fig. 42.—The epipharyngeal surface, and the hypopharynx of DissosteiraCarolina.A, epipharyngeal surface of clypeus and labrum, and ventral extremity ofpharynx. B, lateral view of mouth opening, and of suspensorial apparatus ofhypopharynx. C, antero-ventral view of hypopharynx and its suspensory rods.9Ap, apodeme of adductor muscle of mandible ; Clp, clypeus ; Hphy, hypo-pharynx ; HS, suspensorial bar of hypopharynx ; Lm, labrum ; in, Y-shaped ridgein epipharyngeal wall of labrum ; Mth, mouth ; Phy, pharynx ; Pnt, small lobebehind angle of mouth, possibly rudiment of tritocerebral appendage; Tor,torma ; iv, lateral basal bar of hypopharynx ; x, mandibular branch of sus-pensorial bar (HS) of hypopharynx; y, oral branch of same.the base of the adductor apodeme of the mandible (pAp), and theother (y) goes anteriorly, laterally, and dorsally into the angle of themouth, where it forms a support for the insertion of the retractormuscle of the mouth angle (A, B, ^8).The two lateral bars (fig. 42 B, HS) in the space between the hypo-pharynx and the mouth, with their posterior dorsal arms (x) bracedagainst the bases of the mandibular apodemes. and their ventralends articulated with the basal rods (zv) of the hypopharynx, consti-tute a movable suspensorial apparatus of the hypopharynx. It isevident that a contraction of the mouth angle muscles (38) inserted
NO. 3 INSECT HEAD SNODGRASS III
on the anterior dorsal arms (y) of the bars must effect a movementof the hypopharynx, and that the latter would be lifted and swungforward beneath the mouth opening. The pull of the mouth muscles,however, also retracts the mouth angles, and there is probably thusaccomplished a closing of the mouth upon the food mass accumulatedin the preoral space above the anterior end of the hypopharynx. Inthe grasshopper, the mouth is closed also by the opening of the jaws,but, so far as can be observed in a dead specimen, the closing of themouth in this case results mechanically from the transverse stretch-ing of the oral aperture between the separating bases of the adductorapodemes of the mandibles.Posteriorly the hypopharynx is fixed to the base of the labium, whereits wall is reflected into that of the latter (fig. 41). The hypopharynx,therefore, can swing forward only in unison with the labium, but other-wise it is free to move to the extent permitted by the membranousareas laterad of its base. The only muscles properly belonging to thehypopharynx are the following:32.—Retractors of the hypopharynx (figs. 40D, 41)—A pair ofmuscles arising posteriorly on extreme lateral ends of anterior armsof tentorium (fig. 40 D) ; inserted on posterior parts of basal rods ofhypopharynx (fig. 41).The contraction of these muscles probably retracts the hypopharynx,and pulls the hypopharynx and labium posteriorly. The mouth aper-ture is opened by the contraction of the dilator muscles inserted onits anterior and posterior walls (figs. 41, 44, JJ, 34, 41).The rods {HS) of the suspensory apparatus of the hypopharynxin the grasshopper are evidently remnants of the much larger sus-pensory plates of the hypopharynx in Apterygota and Myriapoda(fig. 21 A, B, C, E, HS). In Microcentrum, as already shown (fig.20 D), a small hypopharyngeal adductor muscle of the mandible isattached to the end of each rod. In the roach (Periplancta) the chi-tinous parts of the hypopharyngeal suspensorium are more stronglydeveloped than in the grasshopper, and their action can be more clearlydemonstrated. In the bees, though the hypopharynx itself may belacking, the oral arms of the suspensory bars are prolonged as slenderrods into the lateral walls of the pharynx, and their basal ends arebridged by a wide plate on the pharyngeal floor.In Dissosteira there is at each side of the mouth, in the angle betweenthe dorsal arms of the suspensorial bar of the hypopharynx, a verysmall but distinct membranous lobe of a definite form (fig. 42 B, Pnt),but having no apparent function, and bearing neither hairs nor senseIk
112 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
organs. The position of these lobes between the mouth and the ad-ductor apodemes of the mandibles strongly suggests that they arerudiments of the postantennal appendages of the tritocerebral seg-ment, which have otherwise not been observed in the adult of anypterygote insect. THE STOMODEUMAt the upper end of the preoral cavity (fig. 41, PrC), anterior tothe base of the hypopharynx, and immediately behind the base of theclypeus is the true mouth (Mth), or external opening of the stomo-deum. The mouth of Dissostcira is a transverse aperture havingacute lateral angles, I)Ut without definite " lips," for the epipharyngeal
Fig. 43.—Pharynx, crop, anterior gastric caeca, and associated organs ofDissosteirn Carolina.Ao, aorta; CA, corpus allatum (?) ; Cr, crop; FrGng, frontal ganglion; GC,gastric caecum ; LNv, lateral stomodeal nerve ; OeGiiy, posterior median oroesophageal ganglion ; Phy, pharynx ; PLGng, posterior lateral stomodeal gan-glion ; Pvcnt, proventriculus.and the supra-hypopharyngeal walls are directly continued into theanterior and posterior walls of the buccal cavity and pharynx.The stomodeum of the grasshopper extends from the mouth upwardin the anterior part of the head (fig. 41). then turns posteriorly abovethe tentorium, and continues rearward through the head and thoraxinto the base of the abdomen (fig. 43). By differences in its diameterand in the character of its walls, the stomodeum is differentiated intoseveral parts, but only three parts are well defined in the grasshopper
;
these are the pharynx, the crop, and the proventriculus.The pharynx, or first division of the stomodeum. is a narrow, mus-cular-walled tube bent downward to the mouth between the anteriorarms of the tentorium (figs. 41, 43, 44, PJiy). The region of themouth, including the upper end of the preoral cavity (fig. 41, PrC)and the part of the stomodeum just within the oral aperture, may bedistinguished as the buccal covitv because the muscles inserted on it
NO. 3 INSECT HEAD SNODGRASS II3(figs. 41, 44, J5, S4' 3^^ 41) function in connection with the mouth.The dorsal dilators {33, 34) arise upon the clypeus (fig. 41, Clp).The true pharyngeal region of the stomodeum of the grasshopperis differentiated into an anterior pharynx and a posterior pharynx,the two parts being thus named by Eidmann (1925) in the roach.The principal differences between the two parts of the pharynx, how-ever, are in the conformations of the cuticular lining, though theposterior end of the anterior pharynx is marked externally by a slightbulging of the lateral walls. The circumoesophageal connectives (fig.44, CoeCon) lie approximately between the two pharyngeal sections.The crop (fig. 43, Cr) is a large, rather stiff-walled sack, represent-ing probably both oesophagus and crop in insects with a long oesoph-ageal tube, though the posterior section of the pharynx in the grass-hopper appears to be the oesophageal region in the caterpillar (fig.55). The anterior end of the crop in Dissosteira lies in the back ofthe head where it rests upon the bridge of the tentorium (fig. 41) :the ventral surface of the thoracic part of the organ is supported bythe spreading apophyses of the thoracic sterna. The anterior third ofthe crop (fig. 43) is somewhat set off from the rest by a slight nar-rowing of the walls ; the posterior part tapers between the largeanterior caecal pouches of the ventriculus, and ends in the proventric-ulus (Pvent). The proventriculus is a small, cup-shaped enlarge-ment of the posterior end of the stomodeum, mostly concealed betweenthe bases of the ventricular pouches (GC).The frontal ganglion of the stomodeal (stomatogastric) nervoussystem (fig. 43, FrGng) rests against the anterior wall of the pharynx,and the posterior median oesophageal ganglion (OeGng) lies overthe posterior end between the spreading bases of the last pair of dorsaldilator muscles of the pharynx (j/). From this second median gan-glion a long lateral nerve {LNv) goes posteriorly on each side of thecrop, ending on the rear part of the latter in a posterior lateral gan-glion (PLGng). A pair of short anterior lateral nerves from theoesophageal ganglion go laterally to a pair of globular bodies, possiblythe corpora allata (CA) , lying at the sides of the posterior pharynx.The anterior dilated end of the aorta (Ao) rests upon the oesophagealganglion, and its open, trough-like lower lip is extended forward be-neath the brain.The Inner Wall of the Stomodeum.—The surface of the intima, orcuticular lining, of the pharynx, crop, and proventriculus is diversifiedby various folds and ridges, most of which are clothed with hairs orare armed with small chitinous teeth.
114 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lThe channels on the walls of the preoral cavity that lead into themouth are continued upon the inner w^alls of the anterior pharynx.The median epipharyngeal groove proceeds upward on the anteriorpharyngeal wall between two converging ridges, but it soon endsin a thick median fold which follows the midline of the roof of theposterior pharynx to the end of the latter. Likewise, the medianchannel leading upward from the base of the hypopharynx is con-tinued on the rear wall of the anterior pharynx, between two con-verging ridges, and ends in a median ventral fold on the floor of theposterior pharynx. From the lateral angles of the mouth, wide chan-nels go dorsally in the side walls of the anterior pharynx, but theseagain end each in a lateral fold of the posterior pharynx. Thus therelative positions of the principal ridges and grooves in the walls ofthe two parts of the pharynx are reversed. In the posterior pharynxthere is a slenderer intermediate fold between each two of the majordorsal, lateral, and ventral folds. These eight folds of the posteriorpharynx end at the entrance of the crop, giving the aperture a stellateappearance when seen from the lumen of the crop. All the pharyngealfolds, except the midventral fold of the posterior pharynx, are clothedwith hairs directed backward.In the crop, a wide dorsal channel proceeds from the pharyngealopening posteriorly on the anterior third of the upper wall betweenconverging folds of the intima. A narrower ventral channel followsthe midline of the floor between a pair of folds that diverge posteriorlyand are lost beyond the middle of the organ. The lateral walls ofthe anterior half of the crop are closely corrugated by obliquely trans-verse ridges, which bear rows of small, slightly curved, sharp-pointed,chitinous teeth projecting backward. The anterior three or fourtransverse ridges on each side are particularly conspicuous by reasonof their greater width, and because they are thickly beset with similarbut slightly larger teeth than those of the other ridges. In the posterior,narrowed part of the crop the transverse ridges are replaced by fine,parallel, lengthwise folds, following the lines of the longitudinalmuscle fibers. Numerous teeth are present here also, but they aresmaller and blunter than those of the anterior region, and are mostlyarranged in small groups, usually two or three together, on elevationsof the intima along the folds. The interior characters of the crop arebetter developed and the teeth are more numerous in the larger organof the female grasshopper than in that of the male. They can bestudied best on pieces of the intima stripped from the tough muscularsheath of the crop.
NO. 3 INSECT HEAI3 SNODGRASS II5The walls of the short proventriculus are produced into six flat,triangular elevations having their bases contingent anteriorly, and theirapices directed backward, where they all end on the rim of the wide,round orifice into the ventriculus. The proventricular ridges are notmere folds of the intima, for each is formed by a thick mass of theunderlying epithelial cells. The surface of the intima in the pro-ventriculus is smooth, except for a few very small teeth on the edgesof the triangular ridges, and areas of minute granulations on the distalhalves of the latter. The posterior margin of the proventricular wallis reflected outward upon itself to form a short circular fold project-ing into the anterior end of the ventriculus, reaching just past theopenings of gastric caeca. The intima covers the outer surface of thefold, but terminates at the base of this surface. The line of the latter,therefore, marks the end of the stomodeal or anterior ectodermalsection of the alimentary canal.The Muscular Sheath of the Stomodeum.—The stomodeal wallsare everywhere covered with flat bands of muscles, which in generaltake a transverse and a longitudinal direction, the transverse bandsl)eing external and the longitudinal internal; but the distribution ofthe two sets is not such as to form a regular net-pattern on all partsof the stomodeum. On the posterior two-thirds of the crop, the ex-ternal transverse fibers have the form of continuous rings encirclingthe organ, and the longitudinals run with its length. On the anteriorthird, however, the ring muscles are interrupted laterally and dorsally.and their layer is continued only on the ventral surface as a series ofventral arcs ; but the fibers of a latero-ventral tract of the posteriorlongitudinal muscles on each side curve upward on the lateral wall ofthe crop where the circular bands are interrupted, and are continuouswi'th those from the opposite side over the dorsal surface as an externallayer of obliquely transverse fibers reaching to the base of the pharynx.On the pharyngeal tube the muscles again take the pattern of regularlyarranged external circular and internal longitudinal fibers. The cir-cular fibers of the pharynx may belong to the interrupted set of circularfibers of the crop, but the longitudinal fibers are continued irregularlyinto the walls of the crop on the inner surface of the anterior circularfibers of the latter, and they do not, therefore, belong to the samelayer as the posterior longitudinal crop muscles. A close study ofthe stomodeal musculature of the grasshopper would show some com-plexity of detail in the arrangement and relationship of the musclefibers, but nothing approaching the intricacy of the fiber connectionsin the muscular layers on the pharynx and crop of the caterpillar, tobe described later.
ii6 SMITHSONIAN MISCF.LLANEOrS COLLECTIONS VOL. 8lThe Dilator Muscles of tJic Stomodeuni.—Th\rttQ.n paired sets ofmuscle fibers and one median unpaired muscle arising on the skeletalparts of the head or thorax are inserted on the stomodeal walls inDissosteira (figs. 41, 43, 44). These muscles may be classed as dor-sal, lateral, and ventral according to their insertions, though becauseof the downward flexure of the pharynx, the first " dorsal " and
" ventral " muscles are anterior and posterior. The dilator musclesof the stomodeum, sometimes called also suspensory muscles, enumer-ated from 55 to 46 inclusive, are as follows
:
Fig. 44.—Dilator muscles of the buccal region, pharynx, and crop ofDissosteira Carolina.CocCon, circumoesophageal connective: Cr, crop; Mth. mouth; Tut. tentorium.
55.
—
First anterior dilators of the buccal cavity (figs. 41, 44).
—
A pair of fan-shaped muscles arising on inner wall of clypeus (fig.37 B), the fibers spreading to their insertion on anterior wall of buccalcavity (fig. 44) mostly distal to oral aperture.
^4.—Second anterior dilators of the buccal cavity (figs. 41, 44).A pair of fan-shaped muscles similar to j?, arising on clypeus nearepistomal ridge (fig. 3/ B) ; inserted laterad of jj and mostly proxi-mal to oral aperture.J5. First dorsal dilators of the pharynx (figs. 41, 44).—A pair ofslender muscles arising on frontal area of head wall, each attachedbetween labral retractors of same side (fig. 38 D) ; inserted on anteriorwall of pharynx.
t NO. 3 INSECT HEAD SNODGKASS II736.—Second dorsal dilators of the pharynx (figs. 41. 44).—Eacharises by a slender stalk on sulmntennal ridge of frons (fig. 41) ; in-serted by spreading base on upper end of anterior pharynx.37.—Third dorsal dilators of the pharynx (figs. 41, 43. 44).—Eacharises by slender stalk on vertex near inner rim of compound eyejust anterior to first dorsal fibers of mandibular adductor ; insertedby widely spreading base on dorsal wall of posterior pharynx.38.—Retractors of the mouth angles (figs. 41, 44).—These, thelargest muscles of the stomodeum, and the first of the lateral series,arise on the subantennal ridges of the frons (fig. 38 D), and extenddownward and posteriorly to their insertions on the oral arms of thehypopharyngeal suspensorial rods (figs. 42 A, B, 44) in the lateralangles of the mouth.jp.
—
First lateral dilators of the pharynx (fig. 44).—A pair ofslender muscles arising laterally on frontal region ; inserted on sidesof anterior pharynx.40.-—Second lateral dilators of the pharynx (figs. 41, 44).—A slen-der muscle on each side, arising on posterior face of distal end of dor-sal arm of tentorium (fig. 41) ; inserted by spreading base on upperend of anterior pharynx (fig. 44).41.—Ventral dilator of the buccal cavity (figs. 41, 44).—A median,unpaired, strap-like muscle arising on ventral face of body of tentor-ium ; inserted on median groove of posterior wall of mouth.42.—First ventral dilators of the pharynx (figs. 41, 44).—A pairof fibers arising on ventral surface of tentorium ; going anteriorly toinsertions medially on lower end of posterior wall of anterior pharynx.43.—Second ventral dilators of the pharynx (figs. 41, 44).—
A
group of diverging fibers on each side, arising on anterior edge oftentorium ; inserted latero-ventrally on anterior pharynx.44.—Third ventral dilators of the pharynx (figs. 41, 43, 44).—large fan of fibers on each side, arising on dorsal edge of posteriorarm of tentorium ; the spreading fibers inserted ventro-laterally alongentire length of posterior pharynx.45.—Anterior dilators or protractors of the crop (figs. 41, 43, 44).
—A large group of fibers arising on each posterior tentorial arm.behind origin of 44 ; spreading posteriorly to insertions ventro-later-ally along anterior third of crop. These are the last of the stomodealmuscles that have their origin in the head.46.—Posterior protractors of the crop and gastric caeca (fig. 43).
—
A pair of long, branched muscles, each arising by a slender stalk oninner surface of prothoracic tergum, just anterior to base of troch-antinal muscle ; branching downward and posteriorly, one branch in-
Il8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
serted on lateral wall of crop just above posterior lateral stomodealganglion {PLGiig), the others on tii)s of the gastric caeca (GC) ofsame side. THE MECHANISM FOR MOVING THE HEADThe head of the grasshopper is freely attached to the prothorax bya membranous neck, but its movements are somewhat limited by theoverlapping anterior edges of the protergum, and by the pair ofcervical sclerites on each side (fig. 45 B, icv, 2cv) which link the headwith the concealed episternal plate of the prothorax (Eps^).The cervical sclerites, however, constitute an important part of themechanism for moving the head The two plates of each pair arearticulated end to end, and ordinarily they are bent downward at anangle to each other (fig. 36 A, cv). The first is articulated anteriorlyto the posterior margin of the postoccipital rim of the head (fig. 45,g), the second posteriorly to the anterior edge of the prothoracicepisternum (EpSi). The neck plates thus constitute a fulcrum on eachside between the head and the thorax, giving a leverage to the dorsaland ventral muscles extending from the postoccipital ridge and ten-torium to the prothorax and the first thoracic phragma. Moreover,upon each plate are inserted strong levator muscles (fig. 45 B) arisingon the back of the head and on the prothoracic tergum, and the con-traction of these muscles, with the consequent straightening of theangle between the two plates of each pair, must cause the protractionof the head. From each anterior plate a horizontal muscle extends tothe prosternal apophysis of the opposite side (fig. 45 A, B, 5^). Be-sides the muscles that connect the skeletal parts of the head, neck,and prothorax, there are two muscles on each side inserted directlyupon the neck membrane (A, 5<5, 57).It is difficult to give names signifying function to the neck muscles,for it is evident that the function will depend on whether the twomuscles of any pair act in unison, or as antagonists. The neck musclesof Dissosteira are as follows, on each side
:
4/.—First protergal muscle of the head (fig. 45 A).—A slendermuscle arising dorsally on prothoracic tergum ; inserted dorso-later-ally on postoccipital ridge of head (PoR).48.—Second protergal muscle of the head (fig. 45 A).—A largermuscle arising on principal ridge of protergum (e) ; inserted with 4/on postoccipital ridge of head.4p.—Longitudinal dorsal muscle of the prothorax (fig. 45 A).—
•
Extends from first thoracic ])hragma ( rPh) to postoccipital ridge(PoR) just l^elow 48.
NO. 3 INSECT HEAD SNODGRASS 11950> 5^-—Cephalic muscles of the cervical plates (fig. 45 A, B).Origin on postoccipital ridge below 4p; both extend ventrally andposteriorly, the first (30) inserted on first cervical plate, the second(57) on second plate.5^. 53-—Protergal muscles of the cervical plates (fig. 45 B).—Origin dorso-laterally on prothoracic tergum; both extend ventrallyand anteriorly, crossing internal to 50 and 51, to be inserted on first
Fig. 45.—Muscles of the neck of Dissosfeira Carolina, right side, internalview.A, muscles extending between head and prothorax, omitting 52, 5-3 and 54,inserted on cervical sclerites (B). B, head and prothoracic muscles of cervicalsclerites.Bsi, basisternum of prothorax; c, first ridge of protergum; icv, first cervicalplate ; ^cv, second cervical plate ; d, second ridge of protergum ; e, third ridgeof protergum; Epsi, episternum of prothorax; EpS2, episternum of mesothorax
;
g, process of head articulating with first cervical sclerite; H, head; iPIi, firstthoracic phragma; PoR, postoccipital ridge; FT, base of posterior arm oftentorium ; Rd, posterior fold of protergum ; SA, apophysis of prothoracicsternum ; Spn, spina ; Ti, tergum of prothorax.
cervical plate, the first {32b) with a branch {32a) to articular process(g) oi postoccipital ridge.34.—Prosternal muscle of the first cervical plate (fig. 45 A, B).—
A
diagonal, horizontal muscle arising on apophysis of prothoracic ster-num (A, SA), crossing its fellow to insertion on inner edge of firstcervical plate of opposite side (B).33.—Longitudinal ventral muscle of the prothorax (fig. 45 A).—broad, flat muscle from prosternal apophysis (SA) to base of poste-rior arm of tentorium (PT).
'Wl
120 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL.
56.
—
Dorsal lateral neck nittsclc (tig. 45 A).—A band of slenderfibers from first phragma ( iPli), inserted on base of neck membrane.57. J^cntral lateral neck muscle (fig. 45 A, B).—A short, flatmuscle from anterior edge of prothoracic episternum {EpSi), insertedon base of neck mem]:)rane.
VI. SPECIAL MODIFICATIONS IN THE STRUCTUREOF THE HEADThe important structural variations in the head of biting insectsafifect principally the fronto-clypeal area, and the posterior lateral andventral regions. Modifications of the facial plates are often to becorrelated with variations in the relative size of the buccal and pharyn-geal parts of the stomodeum, or with a special development of themouth cavity. Modifications in the posterior ventral parts of the headare correlated with a flattening and elongation of the cranial capsule,usually resulting from an upward tilting of the head on the neckby which the mouth parts become directed forward, and, in certainorders, are accompanied by an elongation of the submentum an-teriorly, with a dififerentiation of this plate into a posterior gularsclerite and a secondary anterior submental sclerite.MODIFICATIONS IN THE FRONTO-CLYPEAL REGIONThe prostomial part of the insect head includes the frons, the clyp-eus, and the labrum. Whether or not it comprises also the regionof the compound eyes may be regarded as an open question, and onefor the embryologists to settle. If the compound eyes belong to thefirst true segment of the head, it is probable that the frontal suturesdefine the posterior limit of the prostomium ; otherwise the suturesmust be secondary formations within the area of the prostomium. Thefrontal sutures do not always mark the lines of cleavage in the headcuticula at the time of a molt. In an odonate nymph, for example(fig. 46 I), the facial clefts (t) of the molting cuticula extend fromthe coronal suture outward and downward on each side between theeyes and the bases of the antennae, far outside the possible limits ofthe frons (Fr).The part of the postembryonic head that may be defined as thefrons is the area included between the frontal sutures, where thesesutures are fully developed (fig. 46 B, Fr). The frontal sutures (fs)extend typically from the coronal suture (cs) to the neighborhoodof the anterior articulations of the mandibles (c, c). The true frontalregion, therefore, can not include the bases of the antennae, which
NO. 3 INSECT IIEAl -SNODGKASSin-gans belong to the second head segment behind the prostomiuni.and acquire their facial positions secondarily by a forward and upwardmigration. \^entrally the frons is limited, and separated from the
'-at
122 SMITHSONIAN MISCELLANEOUS COLLECTIONS' VOL. bl
these characters, especially the position of the median ocellus and theorigin of the labral muscles, the true frontal region is to be identifiedwhen the frontal sutures are imperfect or obsolete (fig. 46 E, F, Fr).As was shown in the study of the grasshopper (fig. 36 B), thefrontal region of the face may present a number of secondary linesformed by ridges of the inner surface. In the Dermaptera two sutures(fig. 46 A, s) diverge widely from the end of the coronal suture {cs)and extend outward to the compound eyes. It appears doubtful thatthese are the frontal sutures, for the true frontal region should bethe smaller triangular area indistinctly defined on the median part ofthe face.The clypeus (fig. 46 B, CIp) is a distinct area of the prostomialregion, and is to be identified by the origin of the dilator musclesof the mouth and buccal cavity on its inner wall. It is almost alwaysin biting insects separated from the labrum by a flexible suture, and itis demarked from the frons whenever the epistomal suture is present.The clypeus is sometimes divided into an anteclypeus and a postclyp-eus by a partial or complete transverse suture ; but often the term
" anteclypeus " is given to a more or less membi'anous area betweenthe clypeus and the labrum (fig. 46 G, Aclp), and it is likely thatregions named " anteclypeus " are not equivalent in all cases.The labrum (fig. 46 B, Lm) hangs as a free flap before the mouth.It is a preoral lobe of the prostomium characteristic of insects, myria-pods, and crustaceans. The insect labrum is usually movable, and isprovided with one or two pairs of muscles (though both may be ab-sent), which, as above noted, have their origin on the frons. Thelabral muscles, therefore, are strictly muscles of the prostomium.The principal departure from the typical structure in the pro-stomial sclerites arises from variations in the development or in theposition of the epistomal suture, and from a partial or complete sup-pression of the frontal sutures.The epistomal suture is the external groove formed incidentally tothe development of an internal transverse ridge across the prostomialarea. Since this ridge in generalized insects lies approximately be-tween the anterior articulations of the mandibles, its primitive positionsuggests that it was developed to strengthen the lower edge of the facebetween the mandibular bases. The epistomal ridge itself is a con-tinuation of the subgenal ridges, and the epistomal suture is, there-fore, continuous with the subgenal sutures. In the Ephemerida andOdonata, as we have seen, the anterior arms of the tentorium arisein the subgenal sutures laterad of the bases of the mandibles. Insome of the Orthoptera, as in the roach, and in larvae of Coleoptera.
NO. 3 INSECT HEAD SNODGRASS 123
the tentorial arms have moved forward to a position above the mandib-ular articulations, and their external openings, the anterior ten-torial pits, appear in these positions (fig. 46 B, at, at).In some of the more generalized insects, the epistomal ridge andits suture are lacking, as in the roach, and there is then present onlya single fronto-clypeal sclerite (fig. 47 A, Fr-Clp). In such cases,the tentorial pits (at) lie in the anterior extremities of the subgenalsutures (sgs), above the anterior articulations (c) of the mandibles.Where an epistomal ridge unites the subgenal ridges across the face,separating the clypeus from the frons, the tentorial pits may retain
LmMcls /A ^ Fr-Clp
A
Fig, 47.—Diagrams showing variations in the position of the epistomalsuture (es), and the relations of the frons and the clypeus.Aclp, anteclypeus: at, anterior tentorial pit; c, anterior articulation ofmandible ; CIp, clypeus ; es, epistomal suture ; Fr, frons ; //-, " adfrontal ; Fr-Clp.fronto-clypeus ; fs. frontal suture ; h, line of secondary ridge across lower partof clypeus ; Lin, labrum ; LmMcls, labral muscles, with origm always on frons
;
O, median ocellus.
their positions above the mandibular articulations (fig. 46 A, B, at,at) ; but more commonly they move into the epistomal suture (fig.47 B). In any case, the tentorial pits identify the epistomal suture,when this suture is present. The mandibular articulations (c, c) arecarried by the ventral margin of the epicranium and are not true land-marks of the epistomal suture, as has been pointed out by Yuasa(1920), and by Crampton (1925).As long as the epistomal suture maintains its direct course acrossthe face, no complications arise ; but the suture is frequently archedupward, and this shift in the position of the suture extends the clyp-eus into the facial region above the bases of the mandibles, and re-duces the area of the frons (fig. 47 C). A modification of this kind
ft
124 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8[has taken place in the Hymenoptera. In the larval head of Vespa(fig. 46 D) the clypeus has clearly encroached upon the area of thefrons by a dorsal arching of the epistomal suture {cs). In an adulttenthredinid (E), the same condition is observed, but the lower partsof the frontal sutures {fs) are lost, and the bases of the antennaehave approached each other mesally, and have constricted the frontalarea between them. In the adult of Apis (F) the condition is moreexaggerated—the epistomal suture {es), identified by the tentorialpits {at, at), is arched upward almost to the bases of the antennae,and the frontal sutures are obsolete. The frontal area (Fr), however,is to be identified by the position of the median ocellus, and the pointsof origin of the labral muscles between and just aljove the antennalbases. The head of a larval tenthredinid (fig. 46 C) presents aspecialized condition, for the single large facial plate is here clearlya fronto-clypeus, as shown by the origin of the labral muscles onits upper parts, and by the origin of the tentorial arms {AT) fromthe ridges at its sides. Evidently, the median part of the epistomalridge and its suture has been suppressed. A similar condition is tobe observed in some trichopteran larvae.A still greater degree in the upward extension of the clypeus isshown on the face of a psocid (fig. 46 G). Here the epistomal suture{es) is arched high above the tentorial pits {at, at), and the clypeus{Clp) becomes the large, prominent, shield-shaped plate of the facebetween the bases of the antennae. The frontal sutures are lacking,but the frontal area (Fr) is that between the bifid end of the coronalsuture and the clypeus, on which is located the median ocellus. Aweakly chitinized area below the clypeus is sometimes called theanteclypeus {Aclp), but it appears to be only a chitinization of theconnecting membrane between the clypeus and the labrum.The clypeus, finally, attains its greatest development at the expenseof the frons in the Homoptera (fig. 47 D). In the cicada (fig. 46 H),the clypeus is the great bulging, striated plate of the face upon whicharise the dilator muscles of the mouth pump. The dorsal arch of theepistomal suture {cs) lies on a level with the antennal bases, and theanterior tentorial pits {at, at) are in its upper lateral parts, just abovethe dorsal extremities {c, c) of the mandibular plates (Md). Thefrons is a small, indistinctly defined triangular area {Fr) bearingthe median ocellus in the adult. It is more strongly marked in thenymph, and is cut out by the opening of the frontal sutures at thetime of the molt. The plate below the principal clypeal sclerite isprobably an anteclypeus {Aclp), because in some Hemiptera it is notdistinctly separated from the area above it, but it is questionable if
NO. 3 INSECT HEAD SNODGRASS 1 25
it is homologous with the preclypeal area of the psocid (fig. 46 G,Adp). The terminal piece in the cicada (H, Lm) that closes thegroove in the upper part of the labium would appear to be the labrumby comparison with Heteroptera. The "mandibular plates" (Md)on the sides of the head must be the true bases of the mandibles. Theirupper ends (c, c) have the same relations to the surrounding partsthat the anterior mandibular articulations have in biting insects. Themandibular bristles are chitinous outgrowths from the ventral pos-terior angles of the plates, and the protractor apparatus of each bristlein the adult is differentiated from the posterior margin of the mandib-ular plate, as the writer has elsewhere shown (1927).In the larvae of Lepidoptera, a somewhat different type of modi-fication has produced an unusual distortion in the relation between thefrons and the clypeus. The caterpillar head shows no essential varia-tion within the order, but the homologies of the facial structures areclear if interpreted by the characters w^hich serve as identificationmarks in the other orders. The triangular facial plate (fig. 50 A)thus becomes the clypeus, because the suture (cs) bounding it is identi-fied as the epistomal suture by the origin of the anterior tentorial armsfrom its lateral parts (fig. 50 I, AT). Upon this plate arise themuscles of the buccal region of the stomodeum. The median part of thefrons is invaginated and forms the thick internal ridge (Fr) dorsalto the apex of the clypeus, which is to be identified as the frons bythe origin of the labral muscles upon it. The so-called " adfrontals "(A, fr) are probably lateral remnants of the frons at the sides of theclypeus, and the " adfrontal " sutures are the true frontal sutures(fs). That the relations of the plates of the caterpillar's head, asthus established, are identical with those in other insects is made clearin the diagram given at E of figure 47. The clypeus (Clp) has simplyextended into the area of the frons, and the median part of the latterplate (Fr), bearing the origins of the labral muscles, has been in-flected, while its distal parts, the so-called " adfrontals " (fr), maintainthe original lateral ground of the primitive frontal area. The lower]:»art of the clypeus is sometimes strengthened between the ])ases ofthe jaws by a secondary thickening forming a submarginal ridge (h)on its inner surface.MODIFICATIONS IN THE POSTERIOR VENTRAL REGION OF THE HEADThe structural changes in the posterior parts of the head describedhere are associated with an elongation of the postgenal regions, re-sulting in the production of a long interval between the foramen
126 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. Simagnum and the posterior articulations of the mandibles. Two dif-ferent types of structure follow from this style of modification, oneshown in adult Hymenoptera and in the larvae of Lepidoptera, theother in those orders in which a gular plate is developed.The morphology of the posterior surface of the hymenopteran headis comparatively easy to understand, for, in the larval stages, therear aspect of the head presents the same structure as does that ofan adult orthopteron (fig. 36 C). In the head of the larva of Vespa,for example (fig. 48 A), the details of the structure are exactly asin the grasshopper. There is a distinct postoccipital suture {pos)ending below in the invaginations of the posterior arms of the ten-torium {pt, pt). The postocciput {Poc) is very narrow, but it formsthe marginal lip of the head capsule behind the postoccipital suture.
Cd
Fig. 48.—Development of the posterior head region in Hymenoptera.A, posterior surface of head of larva of Vespa maciilata. C, same of theadult. D, corresponding view of head of adult Apis mellifica.Cd, cardo ; Lb, labium ; Oc, occiput ; Pge, postgena ; Poc, postocciput ; pos,postoccipital suture ; pt, posterior tentorial pit ; St, stipes.The labium (Lb) is suspended from the ventral neck membrane,and the cardines of the maxillae (Cd) are articulated to the ventralcranial margins just anterior to the tentorial pits.In the adult wasp (fig. 48 B) the back of the head presents a quitedifferent appearance from that of the larva. The foramen magnumis greatly contracted and is reduced to a small aperture in the centerof a broad occipito-postgenal field. It is surrounded by a wide post-occipital collar (Poc) set off by the postoccipital suture (pos), inwhich suture are located the posterior tentorial pits (pt, pt). Thelabium (Lb) is detached from the neck and displaced anteriorly(ventrally), and the space between its base and the neck is closed bymesal extensions of the inner angles of the postgenae (Pge, Pge).The articulations of the cardines (Cd) are also far removed from thetentorial pits (pt, pt), and are separated from them by the interven-ing bridge of the postgenae. In the wasp the postgenal bridge pre-
NO. 3 INSECT HEAD SNODGRASS I27
serves a median suture, but in the honeybee (C) the line of unionbetween the postgenal lobes is obliterated, and the bridge presents acontinuous surface in the space between the foramen magnum andthe fossa containing the bases of the labium and maxillae. In anadult tenthredinid {Ptcronidca) , on the other hand, the foramenmagnum, though greatly reduced in size by the development of awide occipito-postgenal area, is still " open " below, that is, it is closedby a narrow remnant of the neck membrane between the approximatedangles of the postgenae. The labium, however, is displaced ventrallyand united with the bases of the maxillae.In the Hymenoptera, then, there can be little question as to theline of evolution that has produced the structure of the back of thehead in the higher forms. The resulting condition has Ijeen correctlyobserved by Stickney (1923), who says: "In many Hymenopterathe mesal margins of the postgenae are fused between the occipitalforamen and the articulation of the labium." A very similar modi-fication of the head has taken place in the caterpillars, as will be shownlater, in which the parts constituting the " hypostoma " (fig. 51 A,Hst) correspond with the postgenal bridge of adult Hymenoptera.In either case, an unusual thing has happened in that the labium, afterbeing moved forward to unite with the maxillae, has been separatedfrom its own segment by the intervention of parts of the first maxil-lary segment. If the postgenae are lateral tergal elements of thehead wall, their ventral union finds a parallel in the prothorax of thehoneybee, which is completely encircled behind the bases of the legsby the prothoracic tergum.The modifications in the posterior ventral parts of the head in thoseorders in which a " gula " is developed are difficult to explain if studiedonly in the higher phases of their evolution, but they can be understoodif traced from forms that show the simpler earlier stages of departurefrom the normal.In the Blattidae. the cranium is much flattened, but the essentialhead structure has not been altered, its posterior parts retaining thesame form as in the less movable head of the grasshoppers. In manyinsects, especially in the Neuroptera and Coleoptera, however, theflattened head is not only turned upward on the neck, causing the trueanterior surface to become dorsal and the mouth parts to be directedforward, but the ventral surface of the head has been elongated topreserve the vertical plane of the foramen magnum. In such insectsthe bases of the mouth parts become separated from the foramenmagnum by a wide space, and in this space there appears a median9
128 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
plate called the " gula." The nature of the gula has long been apuzzle to entomologists, but Crampton (1921, 1928) has given reasonsfor believing that it is a differentiation of the base of the labium,and a few examples taken from the Coleoptera will amply substanti-ate this view.In a scolytid or scarabaeid beetle larva the structure of the headdoes not differ essentially from that of the grasshopper. The faceis directed forward, the mouth parts hang downward, and the undersurface of the head is short. In the scarabaeid larva (fig. 49 A) theoccipital and postgenal regions terminate in a postoccipital sutureipos), in the ventral ends of which are situated the large invagina-tions (pf, pt) of the posterior arms of the tentorium. Beyond thesuture is a narrow postoccipital rim of the cranium (Poc), best de-veloped ventrally, where the lateral cervical sclerites (ci') are articu-lated to it. The postoccipital ridge is developed on each side of theforamen magnum into a broad apodemal plate (PoR), the two platesconstricting the foramen laterally, and uniting ventrally in the broadtentorial bridge, which is concealed in the figure by the ventral partof the neck membrane (NMb). The labium, the maxillae, and themandibles of the scarabaeid larva are suspended from the ventraledges of the cranium exactly as in the grasshopper (fig. 36 C), butthe two forms differ by the elongation in the beetle (fig. 49 A) of thepostgenal margins of the head between the articulations of the car-dines (e) and the posterior articulations of the mandibles (a).The basal part of the submental region of the labium in the scara-baeid larva, Popillia japonica (fig. 49 A), is chitinized to form atriangular plate (Smt). This plate is attached to the mesal points ofthe postgenae (Pge), and has its extreme basal angles prolonged be-hind the tentorial pits to points (/, /) corresponding with the basalarticulations of the submentum with the postocciput in an orthopteron(fig. 36 C, /). There can be no doubt that this; plate in the beetle headis the submentum, or a chitinized basal part of the submentum. It ismarked by a transverse groove between the tentorial pits ( pt, pf).In a silphid larva (fig. 49 B) the general structure of the head issimilar to that in the scarabaeid larva, but the ventral postgenal mar-gins between the articulations of the cardines (c, c) and the mandibles(a) are much longer, and the posterior tentorial pits (pt, pt) areapproximated in the mesally prolonged basal angles of the postgenae.The submentum (Smt) is large; its base is narrowly constricted be-tween the tentorial pits, which here almost cut off a small but distinctproximal area (Gu). The lateral angles of this extreme basal area ofthe submentum are prolonged behind the tentorial pits and become con-
NO. 3 INSECT HEAD—SNODGRASS 129
tinuoLis with the postocciptal rim of the cranium (Poc), which is set
oft' by the postoccipital suture (pos) ending ventrally in the tentorialpits ipt, pt).NMb
B
Poc
Fig. 49.—Evolution of the " gula " in Coleoptera.A, Postero-ventral view of the head of a scarabaeid larva, Popillia japonica.B, same of a silphid larva, Silpha ohscura. C, ventral surface of an adultmeloid, Epicauta pcunsylvanica. D, same of a carabid larva, Scaritcs.a. posterior articulation of mandible; Cd, cardo; cv, cervical sclerite; e, an-terior articulation of mandible ; Gu, gula ; Pge, postgena ; Poc, postocciput
;
PoR, postoccipital ridge; pos, postoccipital suture; pt, posterior tentorial pit;Sml, submentum.The characteristic structure of an adult coleopteran head is wellillustrated in the head of a meloid beetle (fig. 49 C). The form of thecranial capsule here differs principally from that of the scarabaeidor silphid larva in the lengthening of the postgenal regions betweenthe foramen magnum and the articulations of the cardines {e, e).
T30 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lThe extension of the ventral surface of the cranial wall accommodatesthe head to its horizontal position, and has involved a great elongationin that part of the submentum which lies between the posterior ten-torial pits {pt, pt) and extends forward to the articulations of thecardines {e, c). This region of the submentum is known as the gula.In Epicauta (fig. 49 C) the tentorial pits lie at about the middle of thelateral margins of the gula, and the ventral ends of the postoccipitalsuture {pos) are, consequently, turned anteriorly and lengthened inthe same direction behind the pits. The ventral parts of the post-occipital suture, terminating in the tentorial pits, now become the so-called " gular sutures." It is evident that the large gular region in theadult meloid head (fig. 49 C) lying posterior to the tentorial pitsand continuous basally with the postoccipital rim of the cranium(Poc) is produced from the small but corresponding area in the larvalsilphid head (B, Gu) , and that this area, in turn, is merely the basalstrip of the submentum in the scarabaeid larva (A, Sint), attached tothe postocciput by its lateral extremities (/, /).In adult Coleoptera the distal end of the gula may be dififerentiatedas a " pregula " or " gular bar " (C, Pgu). It supports the terminalpart of the original submental plate (Suit), which lies between thebases of the maxillae, and which, in a restricted sense, is usually called
" the submentum " by coleopterists. The pregular region may fuselaterally with the " hypostomal " regions of the postgenae, and inother ways the more primitive structure may become so obscured thatthe relations of the parts are difficult to determine except by studyingthem in a gradient series of simpler forms. The comparative studiesmade by Crampton (1921, 1928) on the gula in various orders showfully its numerous variations, and demonstrate its origin from theproximal part of the primitive submental plate. Stickney (1923) alsohas well illustrated the structure of the gula and associated parts ina large number of coleopteran forms. Stickney fails to recognize,however, that the " gular sutures " are direct continuations of theventral ends of the postoccipital suture, and that, therefore, the gularplate between them must be the basal part of the submentum. Hewould explain the gular bridge in the Coleoptera as a product of theventral fusion of the edges of the postgenae, and the gular scleriteas a plate cut out of this newly-formed region by the anterior exten-sion of the " gular sutures." As we have seen, the ventral bridge ofthe cranial walls is formed in this manner in the Hymenoptera (fig.48), as Stickney has pointed out, but in the Hymenoptera the ten-torial pits have remained at the sides of the foramen magnum, andthe labium has lost its original connection with the postoccipital region.
NO. 3 INSECT HEAD SNODGRASS I3IThe facts are quite otherwise in the Coleoptera, for here the labiumretains its postoccipital connections, and its base has been drawn outbetween the lengthened postgenal margins to form the gula.In certain Coleoptera the postgenal margins do become closely ap-proximated (fig. 49 D), but, in such cases, the gula is compressed be-tween the postgenae, and sometimes almost obliterated. The gularsutures may then be partially or wholly united into a median gularsuture, with which are closely associated the two tentorial pits (pt,pt). Intermediate stages of this condition are well shown in someof the Rhyncophora, in which the head is drawn out into a " snout."In the Neuroptera, both larvae and adults, and in larval Trichoptera,a gular plate is developed showing essentially the same structure andvariations of form as in the Coleoptera. The gular structure has beendescribed in various members of these orders and others in additionto the Coleoptera by Crampton (1921, 1928). In the Termitidae,Crampton shows, the gular region of the submentum may be verymuch elongated, and in the soldier of Termopsis its margins becomeunited with the lengthened edges of the postgenae to fomi a typicalgular plate.The question of the derivation of the gula, the answer to which is,that the gula is a part of the submental region of the labium, is notto be confused with the question as to the origin of the submentumitself. The various views concerning the nature of the submentumhave been already discussed in an earlier section of this paper (pageyy), and the writer will reiterate here only his own personal opinionthat, since the submentum in generalized insects is attached laterallyto the postoccipital tergal region of the head, it comprises the basalparts of the second maxillary appendages, to which, however, theremay be added a median field of the sternum of the corresponding seg-ment. If the submentum is regarded as entirely the labial sternum,then the sternum becomes suspended directly from the tergum of itssegment, and bears the appendages
—
Sl condition so at variance withordinary morphological relations as to discredit the premises fromwhich it is deduced.VII. THE HEAD OF A CATERPILLARThe caterpillars are remarkable for their standardization of struc-ture. In none of the other larger groups of insects is there such uni-formity in fundamental organization as in the larvae of the Lepidop-tera. Some species are superficially specialized, but apparently there isno " generalized " caterpillar. Ontogenetically, the caterpillars prob-ably represent a stage below that of the larvae of Neuroptera, and of
132 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
the larvae of the more generaHzed adult Coleoptera (Adephaga),since the young of these insects are closer in form to that of a typicaladult insect. The caterpillars show primitive conditions in the originof the antennal muscles on the walls of the cranium, in the musculatureof the thoracic legs, in the monocondylic leg joints, in the dactylopo-dite-like end segments of the legs, and in the retention of the abdominal
" legs," if these organs are remnants of true abdominal appendages,as they appear to be. The general form of the alimentary canal, ofthe tracheal system, and of the nervous system are fairly generalized,though the brain is specialized by an extreme condensation of itsganglia. On the other hand, the head, the maxillary appendages, themuscle sheath of the alimentary canal, and the body musculature areall highly specialized. While the form of the caterpillar's body isworm-like, it is not to be supposed that it represents a worm stageor even a primitive stage in the insect ancestry, for the structure ofthe head shows that the caterpillar belongs to the highly evolved stageof the pterygote insects. The caterpillar's form is merely one thatadapts the insect to a wide feeding environment. The extremely com-plicated body musculature must be regarded as acquired through anexcessive multiplication of the segmental muscles to give unlimitedmobility to a soft-bodied animal. The fly maggot likewise has anintricate body musculature, but of quite a different pattern from thatof the caterpillar.STRUCTURE OF THE HEAD CAPSULEThe caterpillar head is an example of the type of head structure inwhich the lower genal and postgenal regions of the cranium (fig. 51 E)are lengthened to give a long ventro-lateral area on each side betweenthe foramen magnum and the posterior articulation of the mandible.The facial aspect of the head (fig. 50 A) is characterized by the ex-tension of the clypeus into the area of the frons, and by the invagina-tion of the median part of the frons dorsal to the clypeus.The prominent triangular plate so characteristic of the facial aspectof a caterpillar's head is unquestionably the clypeus (fig. 50 A, B, C,F, H, Clp), though it has usually been called the " frons." Its marginsare defined internally by a strong V-shaped ridge (E, I, ER), the in-verted apex of which is continued into a thick median ridge of thedorsal wall of the cranium. From the arms of the V-ridge arise theanterior tentorial apophyses {AT), and the latter identify the V-ridgeas the epistomal ridge {ER). The space between the diverging arms,therefore, is the true clypeus {Clp). It has already been shown thatthe clypeus in other orders of insects may be extended into the facialregion dorsal to the mandibular articulations (figs. 46 D, F, G, 47 C).
NO. 3 INSECT HEAD SNODGRASS 133Further evidence that the area thus designated the clypeus in thelepidopteran larva is the true clypeal area, and not the frons, is given
Fig. 50.—Head structure of caterpillars : anterior cranial wall, labrum,antenna, and mandibular muscles.A, anterior surface of head of Lycophotia {Peridroma) margaritosa. _B,same of Thrydoptervx cphemeraefonnis. C, same of Sibeiic stimulca. showingareas of origin of "mandibular muscles (4, 5a). D, antenna of Malacosomaamcricana, left, anterior view. E, interior view of anterior wall of head ofPrionoxyshis robiniae (Cossidae), with labral muscles and adductor of leftmandible in place. F, anterior surface of head of Mnemonica aurocyanea. G.labrum of Lycophotia maryaritosa, anterior view, showing muscle insertions.H, fronto-clypeal area of same. I, inner view of same.AT, anterior arm of tentorium; c, anterior articulation of mandible; Clp.clypeus ; ER, epistomal ridge ; cs, epistomal suture ; Fr, frons ; jr, " adfrontal " ;fs, frontal suture; h, submarginal thickening of clypeus; Lm, labrum; Md,mandible; Nv, antennal nerve; Tra, antennal trachea.by the origin of the clypeal dilator muscles of the stomodeum upon it(fig- 55. 20, 21). Finally, it is to be observed, the muscles of the la-brum, which, in all cases where the identity of the facial plates is clear,
134 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
arise on the frons, are never attached to the triangular plate of thecaterpillar face, but take their origin from the median ridge dorsalto it (fig. 50 B, E, /). In many caterpillars the lower part of the clyp-eus is strengthened by an internal sul)marginal thickening (E, I, h)forming a bracing ridge between the articulations of the mandibles(c, c).The frontal area of the head, as has been shown, is to be identifiedby the origin of the labral retractor muscles upon its inner surface(fig. 47 B, C). In the caterpillar the labral muscles arise either uponthe median internal ridge of the cranium that extends between the apexof the posterior emargination of the vertex and the apex of the clyp-eus, or upon the dorsal bifurcations of this ridge that are continuedinto the margins of the vertical emargination (fig. 50 B, E, 53 E, /).This ridge, then, is at least a part of the frons. It is formed by a deepinflection of the median line of the cranium dorsal to the apex of theclypeus, which appears externally as a median suture (fig. 50 A, B, C,H, Fr). In a softened specimen this frontal invagination can often bewidely opened, when it is seen that its inflected surfaces are continuouswith the so-called " adfrontal " strips lying laterad of the clypeus andextending ventrally to the bases of the mandibles. The sutures, 01membranous lines, along the outer margins of the " adfrontals " thusbecome the true frontal sutures (fig. 50 A, H, I, fs).The frontal region of the caterpillar, therefore, includes the invag-inated frontal groove (fig. 50 A, E, Fr), the " adfrontals " (fr), andperhaps the apical margins of the vertical emargination. When themature caterpillar sheds its skin at the pupal molt, the head cuticulasplits along two lines, which, beginning at the notch of the vertex,follow the external lips of the median frontal invagination and thendiverge along the " adfrontal " sutures to the bases of the mandibles.An elongate piece is thus cut out which includes the median frontalinflection, the " adfrontals " and the clypeus. In some caterpillars themolting cleft follows only one of the adfrontal sutures, the other re-maining closed.The median part of the vertex in the caterpillar's head is obliterated
•by the dorsal emargination, and the angle of the emargination usuallyextends into the frontal invagination (fig. 50 I) ; in some cases thenotch is so deep that the latter is reduced to a very small area dorsalto the apex of the clypeus (F).The labrum of the caterpillar (fig. 50 A, B, Lin) is commonly sep-arated from the lower edge of the clypeus by a wide, flexible membran-ous area. Some writers, having mistakenly identified the true clypeusas the frons, have regarded this membranous area as the clypeus,
NO. 3 INSECT HEAD SNODGRASS I35but the error of this interpretation is shown by the fact that noneof the stomodeal muscles arise upon the membrane, the clypeal dila-tors having their origin on the triangular plate above. The caterpillarlabrum has but a single pair of muscles
:
/.
—
Retractor muscles! of the lahrmn (figs. 50 E, G, 53 E).
—
A pairof long slender muscles arising on the inflected frons (figs. 50 E, 53 E,Fr) ; inserted by long tendons on bases of tormae (figs. 50 G, 53 E)The ventral surface of a caterpillar's head presents a number ofsecondary modifications that, at first sight, somewhat obscure the basicstructure ; but, when the general head " landmarks " are once recog-nized, it is not difficult to see that the fundamental structure is nodifl:'erent from that in an orthopteroid head.As we have noted, the caterpillar head is characterized by an elon-gation of the postgenal regions between the foramen magnum, or theend of the neck membrane (fig. 51 E, NMh) , and the posterior articu-lations of the mandibles (a). On each side, a posterior median partof the postgena (A, E, Hst) is separated from the more lateral post-genal region (Pge) by a suture (/).The median area thus set off is called the hypostoma {Hst), and theinner angles of the two hypostomal areas are approximated andsometimes united on the median line behind the base of the labium,which is thus separated from its usual basal connection with the neckmembrane, or with the postoccipital rim of the cranium. In thismanner a condition has been evolved which is almost a replica of thatin the head of adult Hymenoptera (fig. 48 B, C), except that in thelatter the hypostomal areas are not separated from the rest of thepostgenal regions.In some caterpillars a well-developed subgenal ridge (fig. 51 D,SgR) follows the outer margin of the membranous area of the an-tennal base from the anterior articulation of the mandible (c) to theposterior (a), and is then continued along the anterior mesal marginof the hypostoma (Hst). Some entomologists distinguish the partof the subgenal ridge that skirts the mandibular area as the " pleuro-stomal ridge," or " pleurostoma," and that part which follows thehypostomal margin as the " hypostomal ridge." The external suturethat defines the hypostomal area on each side (E, ;) forms internallya strong ridge (D, ;) extending from the subgenal ridge at the pos-terior mandibular articulation (a) to the postoccipital ridge (PoR).The subgenal ridge, especially its hypostomal part, is lacking or butweakly developed in some caterpillars (C), but the ridge of thehypostomal suture (;) is always well developed, and apparently servesto brace the genal area between the mandible and the posterior rim
136 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81
of the head. The maxillae are suspended in the usual manner by thearticulations of the cardines against the margins of the hypostomalareas of the postgenae (C, Cd, E, e).
NMb
(2 Sg-R-^ Hst t=^ DFig. 51.—Structure of the posterior and ventral parts of the head of acaterpillar.A, postero-ventral view of head of a noctuid (Lycaphotia iimrgaritosa). B,dorsal view of»same. C, interior view of postgenal and hypostomal regions,showing posterior arm of tentorium (PT), and articulation of cardo (Cd). D,inner face of same region in Malacosoma americana. E, ventral view of righthalf of cranium, with mandible and antenna, of Estigmene acraea.a, posterior articulation of mandible; Ant, antenna; 4AP, base of adductorapodeme of mandible; AT, anterior arm of tentorium; c, anterior articulationof mandible ; Cd, cardo ; dap, dorsal apodemal plate of postoccipital ridge ; e,articulation of cardo to cranium; FrR, frontal ridge; Hst, hypostoma ; /, lineof base of neck membrane; j, hypostomal suture, hypostomal ridge ; Lb. labrum
:
Md, mandible; Mx, maxilla; NMb, neck membrane; Pge, postgena ; PoR, post-occipital ridge; PT, posterior arm of tentorium; Tnt, transverse bar of ten-torium ; vap, ventral apodemal plate of postoccipital ridge.The foramen magnum is extraordinarily large in the caterpillar,being almost as wide as the cranium, and is extended forward dorsallyin the median notch of the vertex (fig. 51 A). The postoccipital ridge{PoR) is inflected from the rear margin of the cranial walls, there
NO. 3 INSECT HEAD SNODGRASS I37being no perceptible chitinization beyond it to form a postoccipitalrim in the neck region. The postoccipital ridge gives origin to plate-like apodemes that constrict the actual opening of the head cavity intothat of the neck. Usually there is a pair of dorsal apodemes (A, B.dap) in the notch of the vertex, and a pair of larger ventral apodemes(A, D, E, vap) arising from the postgenal and hypostomal parts ofthe postoccipital ridge. The apodemes vary much in size and shape indifferent species, but those of the ventral pair are usually the largerand the more constantly developed. The apodemes furnish surfacesof attachment for the anterior ends of prothoracic muscles insertedon the back of the head (fig. 57 A, C). In the caterpillars the foramenmagnum is crossed laterally by oblique foraminal muscles, which arethe following
:
2.—Muscles of the foramen magnum (figs. 51 E, 57 A).—Attachedbelow on each side to ventral postoccipital apodeme (fig. 51 E, vap)laterad of posterior root of tentorium ; spreading dorsally and laterally,sometimes as a broad fan (fig. 57 A), to the dorso-lateral parts ofpostoccipital ridge. The foraminal muscles are of the nature of thetransverse muscles of the intersegmental folds in the body of thecaterpillar. From their position it would appear that they must pro-duce a tension on the hypostomal regions of the head wall. Foraminalmuscles are not present in insects generally.The tentorium of the caterpillar is a simple structure consisting oftwo slender longitudinal bars, and of a delicate transverse posteriorbridge. The longitudinal bars, which represent the anterior arms ofthe tentorium (fig. 53 D, E, AT), arise from the lateral parts of theepistomal ridge at the sides of the clypeus (fig. 50 E, I, AT). Theyextend horizontally through the head (fig. 53 E), and are unitedposteriorly with the ends of the posterior bridge (figs. 51 A, C, E,53 D, Tnt). The bridge represents the united median parts of theposterior tentorial arms (fig. 51 A, C, PT), the origins of which (E,pt) are at the posterior angles of the hypostomal plates in the deepinflections that form the inner ends of the ventral postoccipital apo-demes {vap). The positions of all the tentorial roots in the caterpillar,thus, are identical with those of the tentorial roots in an orthopteroidhead, notwithstanding the considerable alterations which the surround-ing parts have suffered. THE ANTENNAEThe antennae are much reduced in all caterpillars, being so smallby comparison with the adult organs that the latter are forced to de-velop by recession, and during the propupal stage their tips only liewithin the antennae of the larva. The antennae of the caterpillar are
138 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8l
situated on membranous areas just laterad of the bases of the man-dibles, while the antennae of the adult arise from the facial regionabove the compound eyes. The ventro-lateral position of the larvalantennae, therefore, appears to be a primitive character in the cater-pillars.Each antenna of the caterpillar consists of three segments, of whichthe middle one is usually the largest, the proximal segment beingoften reduced to a mere basal ring (fig. 51 E, Ant), and the terminalone appearing as a minute apical papilla of the second. The mem-lirane of the antennal base may form a large mound with the antennaretractile into it, or sometimes a long cylindrical projection simulat-ing a basal segment (fig. 50 C). A hypodermal fold projects inwardfrom the base of the antenna (fig. 50 D) which receives the antennalnerve and trachea. Each antenna is moved by a single set of musclefibers, which are
:
J.
—
TJie retractor muscles of the antenna (fig. 50 B—F).—-A groupof slender fibers arising on the parietal walls of the cranium lateradof adfrontal area ; inserted on anterior inner angle of base of proximalantennal segment. Extension of the antennae is probably effected byblood pressure from within the head.THE MANDIBLESThe mandibles of the caterpillar are typical insect jaws suspendedfrom the lower margins of the cranium by a hinge line sloping down-ward posteriorly, with well-developed anterior and posterior articula-tions. The anterior articulation of each mandible consists of a condyleon the cranial margin placed just laterad of the clypeus (fig. 52 A, c),received into a socket on the base of the jaw ; the posterior articulation(a) is the reverse, consisting of a socket on the cranial margin receiv-ing a condyle of the mandible. As in all insects, the articular pointsof the jaw lie outside the membrane that connects the base of themandible with the head. A line between the two articulations dividesthe base of the jaw unequally (fig. 52 B), the larger part being mesadto the axis.The muscles of the mandibles are inserted on large but weaklychitinized apodemal inflections arising at the outer and inner marginsof each jaw. The muscles take their origin on the walls of the craniumand on the ventral apodemes of the postoccipital ridge. Their fibersoccupy most of the cavity of the head, and the cranial hemispheresappear to model their form on that of the bases of the great adductormuscles of the jaws.
NO. 3 INSECT HEAD SNODGRASS 139
4.—The abductor muscles of the mandible (figs. 50 C, 52 B).—
A
group of fillers, small by comparison with the adductor group, arisingon lower lateral and posterior walls of cranium, and on ventral apo-deme of postoccipital ridge laterad of posterior root of tentorium ;fibers converging ventrally, anteriorly, and mesally to insertion onabductor apodeme of mandible.5.
—
The adductor muscles of the mandible (figs. 50 C, E, 52 B, 53E).—An enormous mass of fibers disposed in two sets (figs. 52 B,53 E, 5a, 3b). The fibers of one group arise from almost entire dorsal,anterior, lateral, and posterior walls of corresponding half of epi-cranium above the ocelli (figs. 50 C, E, 53 E, 3a) ; they converge down-ward upon both surfaces of the broad, adductor apodeme of mandible.The fibers of the other group (figs. 52 B, 53 E, 5/;) arise on ventralapodeme of postoccipital ridge (fig. 53 E, vap) mesad of bases ofClp
Fig. 52.—Mandibles of a caterpillar.A, mandibles and antennae of Estigmenc acraea, ventral view. B, left mandibleof a noctuid, with bases of muscles, dorsal view.a, posterior articulation of mandible; Ant, antenna; c, anterior articulation ofmandible ; Clp, edge of clypeus ; Mds, mandibles ; 4, abductor muscle of mandible
;
^a, fibers of adductor arising on wall of cranium; 5/', adductor muscles arisingon ventral apodeme of postoccipital ridge (see fig. 53 E).abductor fibers, and extend horizontally to posterior edge of adductorapodeme of mandible.The obliquity of the mandibular axes causes the points of the jawsto turn upward and somewhat posteriorly during adduction. Whenthe mandibles are closed, the teeth on the cutting edges of the twojaws are opposed to each other (fig. 52 A), not interlocked ; but usuallyone mandible closes first and its toothed edge passes inside that of theother. Live caterpillars examined by the writer always closed the rightmandible over the left, and species of several families preserved inalcohol were found to have the jaws in the same position.THE MAXILLAE AND LABIUMThe basal parts of the maxillae and labium are united, and theirchitinous areas are reduced or variously broken up into small plates(figs. 51 A, 53 A), which may dififer much in dififerent species. With
140 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81
the anterior wall of the labium, apparently, is united also the hypo-pharynx (fig. 54 D, Hphy), and the duct of the silk gland opensthrough a hollow spine, the spinneret, at the tip of the labium.Each maxilla includes a cardinal area (fig. 53 A, Cd), a stipitalarea {St), both united with the basal part of the labium, and a freeterminal lobe {Lc) , which appears to be the lacinia. A maxillary palpusis lacking. The area of the cardo- includes one principal sclerite (fig.53 A, B, E, F, Cd), and generally one or two accessory plates (A, E,F, k, k). The principal sclerite is always articulated to the hypostomalmargin at a point (c) corresponding with the articulation of the cardoto the cranium in orthopteroid insects. The area of the stipes (St)is variously chitinized, or unchitinized, but it always preserves theridge (q) of its inner margin, upon which are attached all the stipitalmuscles. The homology of the terminal lobe of the maxilla is difficultto determine.The musculature of the maxilla of a caterpillar comprises musclespertaining to its three parts, most of which are comparable to themaxillary muscles of the grasshopper or other generalized insects,though there is little similarity in the general appearance of the struc-ture in the two cases. The cardo, in the caterpillar, is provided withtwo or three muscles (fig. 53 B, E, F. 6, 7, 8), all of which arise onthe anterior arm of the tentorium (D, E), and, therefore, representthe tentorial adductors of the cardo in orthopteroid insects. The usualcranial muscle of the cardo (fig. 25, /, fig. 40 C, 10) is lacking in thecaterpillar. The stipes is provided likewise with tentorial adductors(fig. 53 B, D, E, F, g, 10, 11) inserted on its mesal chitinous ridge {q).The terminal maxillary lobe is moved by muscles that arise within thestipes (B, F, 12, jj), and also by a long muscle (B, 14) having itsorigin in the posterior angle of the hypostomal plate (Hst) of theepicranium. These three muscles are inserted upon a basal scleritein the ventral wall of the maxillary lobe (A, B, /). The first twosuggest the ordinary stipital muscles of the lacinia, but the third (14)appears to have no homologue in more generalized insects, since theusual cranial flexor of the lacinia (fig. 30 B, ficc) is inserted on themedian angle of the latter. The insertion of the three muscles on asingle sclerite in the base of the maxillary lobe leaves no evidence toindicate the presence of a galea, and suggests that the lobe is the laciniaalone, complicated in form by the development of large sensorypapillae. Certainly, the musculature of the lobe shows that none ofthe papillae can be a palpal rudiment.
NO. 3 INSECT HEAD SNODGRASS 141
Fig. 53.—Maxilla, labium, and silk press of a caterpillar.A, Estiijmenc acraca, maxillae and labium, with hypostomal plates of head,posterior (ventral) view. B, internal view of left maxilla and hypostomalregion of same, showing muscles. C, Malacosoma americana, distal part oflabium and hypopharynx, lateral view, showing silk press and muscles. D.Lycophotia margaril osa, muscles of maxillae, labium, and hypopharynx, internal(dorsal) view. E, the same, right side of head, internal view, showing musclesof labrum, mandible, maxilla, and labium. F, cardo and lateral parts of stipesof right maxilla, showing bases of muscles, dorsal (anterior) view. (Comparewith E.)a, anterior articulation of mandible; AT, anterior arm of tentorium; Cd,cardo ; Clp, clypeus ; dap, dorsal apodeme of postoccipital ridge ; c, articulationof cardo with cranium; Fr. frons ; /;, submarginal ridge ot clyjeus : Hpliy.hypopharynx ; Hst, hypostoma ; /, hypostomal suture ; k, accessory plates ofcardo; l, basal sclerite of lacinia ; Lc, lacinia ; Lm, labrum; Md, mandible; Mt,mentum; NMb, neck membrane; PT, posterior tentorial arm; pt, posteriortentorial pit ; q, ridge on inner edge of stipes ; r, articular nodule betweenend of stipital ridge (q) and mentum; SID. silk gland ducts; Suit, submentum ;Spf, spinneret; St, stipes; Tut. transverse bar of tentorium; vap, ventral apodemeof postoccipital ridge.
W'
142 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81The muscles of the maxilla may be enumerated as follows, andthey will probably be found to dififer but little in different species ofcaterpillars
:
6.—First adductor of the cardo (fig. 53 B, D, E, F).—Origin onposterior end of anterior arm of tentorium (AT) ; goes ventrallyto insertion on base of cardo.7.
—
Second adductor of the cardo (fig. 53 B, D, E, F).—Originanteriorly on tentorial arm (D, E) ; insertion on distal end of cardo.8.—Third adductor of the cardo (fig. 53 D, E, F).—This musclefound in noctuid larvae, perhaps a subdivision of 7. Origin anteriorto 7 on tentorial arm (D, E) ; insertion on accessory plate (E, F, k)mesad to the articulating sclerite of cardo {Cd).9. First adductor of the stipes (fig. 53 B, D, E, F).—Arises nearanterior end of anterior tentorial arm (D. E) ; goes obliquely ventrallyand posteriorly to insertion on marginal ridge (B, D, E, F, q) ofstipes.70.
—
Second adductor of the stipes .(^g- 53 B, D, E, F).—Originat anterior end of tentorial arm, just before p (D, E) ; insertion onstipital ridge (D, E, F, q) anterior to p.II.—Third adductor of the stipes (fig. 53 B, D, E, F).—Arises pos-teriorly on anterior tentorial arm, just before first adductor of cardo(6) ; goes obliquely ventrally and anteriorly (D, E), internal to 7, 8,p, and 10, to insertion on anterior end of stipital ridge (B, D, E, F, q)12.—External retractor of the lobe (fig. 53 B, F).—Origin onbase of stipital ridge {q) ; insertion laterally on basal plate (A, B, /)of terminal lobe of maxilla./?. Internal retractor of the lobe (fig. 53 B. F).—Origin onbase of stipital ridge (q) ; insertion mesally on basal plate (A, B, /)of terminal lobe of maxilla.14.-—Cranial abductor of the lobe (fig. 53 B).—Origin in basalangle of hypostomal plate of epicranium (Hst) ; insertion on outerend 'of basal plate (/) of terminal lobe of maxilla. A correspondingmuscle is not present in orthopteroid insects.The labium of the caterpillar (fig. 53 A) lies between the maxillae.The broad membranous surface of its large submental region is unitedon each side with the marginal ridges (q) of the stipites, and itsbasal part is continuous laterally with the membrane of the cardinalareas. Proximally the labium may be continuous with the neck mem-brane (NMb) between the approximated ends of the hypostomalplates (Hst), but, when the latter are united, the labium becomes
NO. 3 INSECT HEAD SNODGRASS 143
separated from the neck. A large suljniental plate occupies the medianbasal part of the submental region in some species (A, Smt).The distal, free lobe of the labium probably represents the mentumand ligula of other biting insects, combined with the hypopharynx,which forms its anterior surface (fig. 54 A). Evidence of this in-terpretation is found in the fact that the labial and hypopharyngeal
Hphy,
SID
--Pr
Hphy
Fig. 54.—Distal part of labium, hypopharynx, and silk press of a noctuidcaterpillar.A, mentum and hypopharynx, with silk press partly exposed, lateral view.B, the same, dorsal view. C, the same, posterior view, showing support on armsof stipites iq, q). D, lateral view, showing muscle attachments.Hphy, hypopharynx ; Mt, mentum ; Pr, silk press ; q, q, ridges of stipes ; r, r,articular nodules between stipital arms and mentum ; SID, silk duct ; Spt,spinneret.
muscles are inserted on the base of the lobe (figs. 53 C, D, 54 C, D,15, 16), and in the position of the spinneret (fig. 54 A, D, Spt), whichcontains the opening of the silk duct (salivary duct), the latter beingnormally situated between the labium and the hypopharynx (fig. 18D, SIO).The mental region of the mento-hypopharyngeal lobe appears tobe that occupied by the large proximal plate (fig. 53 A, Mt) that em-10
144 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lbraces the base of the lobe ventrally and laterally, but which is notcontinued across the hypopharyngeal surface (figs. 53 C, 54 A, C, D,Mt). This plate is supported upon the distal ends of the ridges ofthe stipites (fig. 54 C, D, q, q), which are turned forward and artic-ulated with the dorsal arms of the mentum (Mt) by small, chitinousnodules (r, r). By this mechanism, the mentum-hypopharynx, whichcarries the spinning apparatus, is freely movable on a transverse axisbetween the ends of the supporting stipital ridges. The motion in avertical plane is the only movement that can be given to the spinningapparatus, except by the action of the entire head ; but the head ofthe caterpillar is highly mobile by reason of the great number of mus-cles inserted upon its posterior margin (fig. 57). The musculatureof the mentum-hypopharynx, or spinning organ, is as simple as itsmechanism, consisting of two pairs of muscles, as follows
:
75.
—
Reductors of the spinning organ (figs. 53 C, D, E, 54 C, D)—A pair of double muscles arising at posterior ends of tentorial arms(fig. 53 D, E) ; converging ventrally and anteriorly to insertions onventral edge of mentum (figs. 53 C, E, 54 C, D, Mt). These musclesprobably represent the mento-tentorial muscles of orthopteroid insects(fig. 40 D, ^5), which are primitive adductors of the second maxillae.i6.—Productors of the spinning organ (figs. 53 C, D, 54 C, D).
—
A pair of broad muscles arising medially on transverse bridge of ten-torium (fig. 53 D, Tnt), diverging ventrally and anteriorly to base ofhypopharynx (figs. 53 C, D, 54 C, D, Hphy). These muscles are prob-ably the retractors of the hypopharynx in orthopteroid insects (fig.41, J'?)-The silk press of the caterpillar is a special development of thecommon duct of the labial glands (here, the silk glands). The deeplyinvaginated dorsal wall of the organ exerts a pressure on the silk ma-terial, which is regulated by two sets of opposing muscles that, prob-ably acting together, effect a dilation of the lumen of the press byelevating the invaginated roof. The muscles of the press arise withinthe mentum, and the two sets may be distinguished as follows
:
//, 18.—Dorsal muscles of the silk press (fig. 54 A, B, C).—Twolateral series of muscles, the number on each side varying in differentspecies of caterpillars, arising on dorsal arms of mentum ; converg-ing to insertions on chitinous raphe in dorsal (anterior) wall of press.ip.—Ventral muscles of the silk press (fig. 54 A, B, C).—Originin ventrolateral parts of mentum ; insertion on dorso-lateral edges ofsilk press. These muscles are antagonists to the dorsal muscles, sincethe fibers of the two sets oppose each other in the crossed lines of an X
NO. 3 INSECT HEAD SNODGRASS 145
(fig. 54 C) ; but in function the ventral muscles are probably accessoryto the dorsals by counteracting the pull of the latter on the press.It is difficult to discover a parallelism between the muscles of thesilk press in the caterpillar and muscles of the labium in other insects.However, it may be possible that the two sets of muscles in the labiumof the grasshopper (fig. 40 D, 26, 2y) inserted on the salivary cup {v)are the prototypes of the silk press muscles, though their insertionpoints are ventral instead of dorsal.THE STOMODEUMThe stomodeum of the caterpillar (fig. 55) is diflferentiated intofour parts. The first part is a bucco-pharyngeal region {BuC, Phy) ;Cr
BuC- ^./
Fig. 55.—Anterior part of the stomodeum of a noctuid caterpillar, showingmuscles of the stomodeal wall, and the dilator muscles arising in the head.a-m, muscles of stomodeal wall ; BnC, buccal cavity ; Cr, crop ; OE, oesoph-agus ; Phy, pharynx ; 20-23, muscles of buccal region, arising on _ clypeus
;
24-27, dorsal dilators of anterior pharyngeal region; 28-30, dorsal dilators ofoesophagus (posterior pharyngeal region) ; 31-36, ventral dilators.the second, a cylindrical tube with strong transverse muscle rings,constitutes an oesophagus (OE) in the caterpillars, but it evidentlycorresponds with the posterior section of the pharynx in Orthoptera
;
the third part is the large sack-like crop (Cr) ; the fourth is the con-stricted posterior region of the stomodeum (fig. 56 F, Pvent), whichmay be termed the proventriculus, though it has no special develop-ment of the lining intima, such as usually distinguishes the proven-tricular region in other insects.The muscular sheath of the entire alimentary canal of the caterpillaris strongly developed, and in some parts becomes highly complicatedin structure. The alimentary muscles are particularly strong in thenoctuids, and the following descriptions are based mostly on Lyco-photia margaritosa.
146 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81The lateral walls of the bucco-pharyngeal region are marked on eachside by an oblique ridge (fig. 55), formed by a specially chitinizedgroove of the intima, which gives a firm line of insertion for the ex-ternal muscles. The latter consist of thick, broad bands of stronglyfibrillated muscle tissue, for the most part lying in one plane, thoughvarying in position from transverse to longitudinal. The anterior-most muscles consist of two dorsal arcs (a, /?), and of a correspondingwide ventral arc (d), their ends inserted laterally on the obliqueridges. This part of the stomodeum may be defined as the buccalregion because its dilator muscles (20-2^) have their origins on theclypeus. The anterior end of the pharyngeal region following is cov-ered dorsally by a broad transverse muscle (c) attached laterally onthe oblique ridges. The frontal ganglion lies over the posterior borderof this muscle. Each side of the pharynx presents two muscle plaques(c, f) attached to the ventral margins of the upper half of the obliqueridge, but extending posteriorly to the oesophagus. The posteriordorsal wall of the pharynx is covered with several longitudinal mus-cles, the most prominent of which is a wide, median, external bandof fibrils {g) deflected from the posterior part of the broad anteriortransverse muscle (c) . Concealed by this muscle are two longitudinalsof a deeper set, arising anteriorly on the buccal region beneath thefirst transverse muscle (a) and extending posteriorly to the anteriorend of the oesophagus. Several superficial longitudinal fibers lie morelaterally.The buccal region of the stomodeum is thus distinguished by itsstrong circular musculature, which evidently gives it a powerfulconstrictor action. The pharynx is provided principally with longi-tudinal muscles, and its action, except for that produced by the an-terior dorsal transverse muscle, must be one of lengthwise contrac-tion.The entire length of the oesophageal tube is sheathed in a closeseries of strong circular fibers {i) which are complete rings, excepta few of the most posterior interrupted dorsally at the anterior end ofthe crop.The inner walls of the pharynx and oesophagus form four longi-tudinal folds—one dorsal, one ventral, and two lateral. The dorsalfold is broad, flat, and straight-edged. It arises at the base of thelabrum, where its margins begin at the tormae, and continues to theposterior end of the oesophagus, where it is lost with the suddenwidening of the stomodeal tube in the crop. Between the pharynx andthe oesophagus, the continuity of the dorsal fold is interrupted by atransverse fold. The ventral and lateral folds are less definite, rounded
I NO. 3 INSECT HEAD SNODGRASS I47inflections of the stomodeal wall, continuous from the pharynx intothe oesophagus. In Lycophotia margaritosa each of these folds endsat the opening of the crop in a prominent fleshy papilla covered withsmall chitinous points. Between the folds are four deep channels ex-tending from the mouth to the crop, two dorso-lateral, and two latero-ventral. Possibly it is through these channels that the alimentaryliquid, which caterpillars frequently eject from the mouth whenirritated, is conveyed forward from the crop.The muscles of the crop (fig. 55, Cr) are arranged longitudinallyand circularly. The circular muscles (/), except for a few closelyplaced anterior bands (k), are widely spaced, external circular fibers.They all completely surround the crop like the hoops of a barrel. Atthe junction of the crop with the oesophagus, there are several shorttransverse fibers (/) confined to the dorsal surface. All the musclesof the crop are strongly fibrillated (fig. 56 A, B, C, D). The circularbands have distinct nuclei, but nuclei were not observed in the longi-tudinal muscles of noctuid species examined.The longitudinal muscles of the crop (fig. 55, vi) have their originin single fibrillae (fig. 56 A) or small bundles of fibrillae (B) givenoff from the posterior margins of the circular fibers. They are, there-fore, of the nature of branches of the circular fibers, and this factmay account for their lack of nuclei. Moreover, the longitudinalmuscles are not continuous, individual bands, but are everywherebranched and intimately united by intercrossing bundles of fibrillae insuch a manner that the entire layer becomes a plexus of muscle tissue(fig. 56 C) . Most of the fibrillae of this layer spring from the anteriorcircular fibers, but probably all the circular fibers contribute at leasta few elements to the longitudinal plexus. On the anterior end of thecrop, the longitudinal fibrillae appear as simple connectives betweenthe transverse fibers (fig. 55, /). On the posterior end of the crop(fig. 56 F), the longitudinal muscles again break up into smaller fibrilbundles, and at last into fine strands that reunite with the externalcircular fibers of the crop or the proventriculus.The proventricular region (fig. 56 F, Pvcnt) resembles the oesoph-agus in being surrounded by a close series of strong circular musclefibers (w). There is no distinct inner muscular sheath here, but thecircular fibers are all connected by small bundles of fibrillae going fromone to another (G), some to the first neighboring fibers, others tothe second, third, or fourth removed in either direction. The proven-triculus has a special feature in the presence of an external layer offine, widely-spaced, longitudinal muscles, stretched freely between itstwo ends (fig. 56 F, o). These threadlike strands arise anteriorly
148 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81from branches that spring from the posterior ends of the longitudinalcrop muscles, and from the anterior circular fibers of the proventric-ulus. Posteriorly they again break up into branches that are lost ina plexus of fibers at the junction of the proventriculus with the ven-triculus (Vent).A study of the stomodeal muscle sheath of the caterpillar thusshows that the usual brief statement that the insect stomodeum issurrounded by an external layer of circular fibers and an internal layerof longitudinal fibers must be considerably modified and amplified tofit conditions in the caterpillar. The proctodeal muscles of the cater-pillar are even more complicated than are those of the stomodeum.The high degree of development in the alimentary musculature of thecaterpillars accords with the general specialization of the caterpillaras an animal most efficient in feeding, and the extreme developmentof the somatic musculature is only another adaptation to the same end.The dilator muscles of the stomodeum are inserted dorsally andventrally on the stomodeal walls. The dorsal muscles are grouped intothree sets corresponding with the buccal, pharyngeal, and oesophagealregions of the stomodeum. The dilator muscles of the dorsal andcent;:al series, enumerated according to the order of their insertions,are as follows
:
20.—First dorsal dilators of the buccal cavity (fig. 55).—A pairof slender muscles arising on submarginal ridge of clypeus (fig. 50I, h) ; extending posteriorly to insertions laterally on roof of mouthcavity just before first band of circular stomodeal muscles.21.—Second dorsal dilators of the buccal cavity (fig. 55).—Originson clypeus, above middle and close to lateral margins ; insertionsmedially on dorsal wall of mouth cavity between insertions of 20.22, 2^.—Third and fourth dorsal dilators of the buccal cavity (fig.55).—Two pairs of slender muscles: those of each side arising to-gether in ventral angles of clypeal triangle just above ends of sub-marginal ridge ; inserted dorso-laterally on buccal region, 22 beforesecond band of transverse muscles (b), <?j behind it.A wide space occupied by the third transverse muscle band (c)intervenes between the dilators of the buccal region and those of thetrue pharyngeal region.24.—First dorsal dilators of the pharynx (fig. 55).—Origin onupper part of clypeus just internal to epistomal ridge; insertionmedially on dorsal wall of pharynx laterad of frontal ganglion. Theseare clearly true pharyngeal muscles ; their points of origin have evi-dently crossed the epistomal ridges to the clypeus.
r NO. 3 INSECT HEAD—SNODGRASS 149
r-; [ 'j~>-^r'/V-mmm^ A Bllim ^mmM 1: Bm
lllllll'Hiiiilifii Ec
,.»;V^fr'?S'(?''i^'«-jCr
^^:
.cPvent lilit^gEl^--vent;U§flipf,^ p
^' '-•
-miJ f l^ 1 ^fFig. 56.—Muscles of the stomodeum of a noctuid caterpillar.A, B, origin of longitudinal muscles (in), of crop (see fig. 55) from fibrilsdeflected from the anterior circular muscles (/, k). C, plexus of longitudinalmuscles, anterior part of crop. D, piece of circular fiber from anterior partof crop. E, a connecting fiber between circular and longitudinal muscles. F,posterior end of crop (Cr), proventriculus (Pvcnt), and anterior end of ven-triculus (Vent): I, m, circular and longitudinal muscles of crop; n, circularmuscles of proventriculus ; o, external longitudinal fibers of proventriculus ; p,first suspensory muscles of ventriculus. G, parts of seven consecutive circularfibers of proventriculus, showing bundles of uniting fibrils, external.
150 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81
^5.
—
Second dorsal dilators of the pharynx (fig. 55).—Origins onepistomal ridges near union with frontal ridge ; insertions dorso-laterally on pharynx.26.—Third dorsal dilators of the pharynx ( fig. 55) .—Each arises oncranial wall laterad of origins of antennal muscles ; extends medially,posteriorly, and downward to insertion on pharynx just laterad of 2-,.The insertions of muscles 24, 2§, and 26 all lie posterior to thefrontal ganglion connective.2/.—Fourth dorsal dilators of the pharynx (fig. 55).—A group offibers on each side, arising on outer surface of lower end of frontalridge ; converging to one or two stalks inserted on dorsal wall ofpharynx just before brain.The following dorsal muscles are inserted behind the brain and onthe region of the stomodeum that may be distinguished in the cater-pillar as the oesophagus, but which is the so-called posterior pharynxin Orthoptera.28, 2p, 50.
—
Dorsal dilators of the oesophagus (fig. 55).—Threefans of muscles arising on posterior margin of cranial walls on eachside of vertical emargination ; the spreading fibers inserted dorso-laterally on oesophagus from brain to crop.J/. First ventral dilators of the pJuirynx (fig. 55).—A pair oflong slender muscles arising on transverse bar of tentorium (fig. 53D, Tnt), converging to ventral wall of pharynx where inserted justbehind first ventral transverse muscle {d)
.
^2, 55.
—
Second and third ventral dilators of the pharynx (fig. 55).
—A pair of small muscles on each side arising on extreme outer endsof transverse tentorial bar ; fibers spreading at insertion ventro-later-ally on pharynx just before anterior circular muscles of oesophagus.34' 35> 3^-—Ventral dilators of the oesophagus (fig. 55).—Threelarge fans of fibers arising on postoccipital apodemes on each sidelaterad of posterior roots of tentorium ; the spreading fibers insertedventro-laterally on oesophagus from circum-oesophageal nerve con-nective to crop.THE MUSCULATURE OF BACK OF HEAD, AND NATUREOF THE INSECT NECKThe head of the caterpillar is remarkably mobile. It is providedwith a wonderful system of muscles, the fibers of which arise mostlyin the prothorax and are distributed at their insertions upon the post-occipital ridge of the head in such a manner as to enable the caterpillarto make all possible head movements of which it conceivably mighthave need (fig. 57 A, B, C).
NO. 3 INSECT HEAD SNODGRASS 151The muscles of the prothorax of the American tent caterpillar,Malacosoma americana, are illustrated in figure 57. At A are shownthe lateral and ventral muscles as seen from a posterior dorsal view,with the head turned somewhat downward on the neck ; B shows thedorsal muscles as seen from below; C presents an inner view of all
Fig. 57.—Muscles of the prothorax of a caterpillar, Malacosoma americana.A, prothoracic muscles inserted on lateral and ventral parts of back of head,and ventral muscles of mesothorax, posterior view. B, dorsal muscles of pro-thorax and mesothorax inserted on dorsal half of back of head, seen frombelow. C, innermost layers of muscles of right half of prothorax, internalview. D, external muscles of right half of prothorax.Isg, intersegmental fold ; Li, base of prothoracic leg ; PoR, postoccipitalridge of head ; s, edge of tergal plate of prothorax ; iSp, first spiracle ; Tnt,transverse bar of tentorium ; vap, ventral apodemal plate of postoccipitalridge.the muscles in the right half of the prothorax inserted on the head
;
and D gives the muscles of the same side that lie external to thoseshown in C, except the single fiber arising just dorsal to the spiracle,which is shown in both figures.
152 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81The various fibers of the head muscles are mostly arranged ingroups, and it is easiest to trace them from their points of insertionon the back of the head. Inserted in the median notch of the vertexthere is a dorsalmost group of long fibers that diverge posteriorly tothe dorsal wall of the prothorax (B, C), the middle fibers of eachgroup going to the posterior margin of the segment. External tothese muscles, a group of short fibers, inserted serially on each side,extends posteriorly and dorsally to the tergal plate of the prothorax.Laterally there are inserted on the postoccipital ridge several fibersthat spread to their origins on the tergal plate, and a group of fourlong fibers going dorsally and medially to the intersegmental fold(Isg), with the two median fibers crossing the latter to the dorsum ofthe mesothorax. Three lateral groups of fibers ( A, C) go ventrally andposteriorly from their head insertions, one to the sternal interseg-mental fold, another to the region just before the base of the protho-racic leg, and the third to the median longitudinal fold between thelegs. Ventrally there are inserted on the ventral apodeme of thehypostomal region (C, z'ap) the anterior ends of the ventral longitudi-nal muscles of the prothorax (A, C), and a group of four long fiberson each side that arise on the region above the spiracle.It is of particular interest to observe that, in the caterpillar, theventral longitudinal muscles of the prothorax are not inserted on thetentorium (fig. 57 A, C) as they are in orthopteroid insects, and fur-thermore, that all the principal longitudinal ventral muscles of thethorax have their origin on the intersegmental folds, and not on in-trasegmental apophyses. The primitive anterior insertion of thesemuscles in the prothorax, therefore, should be on a ventral interseg-mental fold between the prothorax and the last head segment. Wehave already seen that there is evidence of the loss of the true labio-prothoracic intersegmental fold, since the postoccipital ridge, whichbears the anterior attachments of the prothoracic muscles in all knowninsects, appears to be the fold between the maxillary and the labialsegments. If so, the original attachments have been lost and themuscles now extend through the length of two primary segments.Furthermore, the attachment of the ventral muscles of the cater-pillar on the hypostomal regions of the head must signify a migrationof the muscles from their primitive sternal insertions, for the hypo-stomal lobes clearly belong to the postgenae, and are, therefore, ventralextensions of the tergal area of the head wall. In any case, an attach-ment of the ventral muscles on the bridge of the tentorium certainlyrepresents a farther displacement of the muscle insertions by a final
I NO. 3 INSECT HEAD SNODGRASS 153
migration from the tergal postoccipital ridge to the posterior ten-torial apophyses.The question of the morphology of the cervical region of the insectmust yet remain a puzzle ; but the musculature gives no evidence of theexistence of a neck segment. On the other hand, the fold in theintegument of the caterpillar between the neck (fig. 57 D, Cv) andthe prothoracic tergum {To) is suggestive of being the true interseg-mental line between the labial segment and the prothoracic segment,and several muscles of the prothorax have their anterior attachmentsupon it (D). If the primitive insect is conceived as a continuouslysegmented, vermiform animal, the neck, or any other secondary inter-segmental area, must be a part of a primary segmental region. Fromthe evidence at hand it seems more probable that the region of theinsect neck belongs to the labial segment, than to an anterior partof the prothoracic segment.ABBREVIATIONS USED ON THE FIGURESAb, abdomen.abplp, abductor of palpus.adplp, adductor of palpus.Am, amnion.AMR, anterior mesenteron rudiment.An, anus.Ant, antenna.AntNv, antennal nerve.Ao, aorta.AP, apical plate.AR, antennal ridge.Arc, archicerebrum.as, antennal suture.AT, anterior arm of tentorium.at, anterior tentorial pit.BC, body cavity.Bdy, body.Blc, blastocoele.Bid, blastoderm.Bp, blastopore.iBr, protocerebrum.2Br, deutocerebrum.SBr, tritocerebrum.Bs, basisternum.BxiC, buccal cavity.CA, corpus allatum.Cd, cardo.Cer, cercus.
Ch, chelicera.Cho, chorion.Clp, clypeus.CoeCon, circumoesophageal connective.Com, commissure.SConi, commissure of tritocerebrallobes.Con, connective.cs, coronal suture.ct, coxo-trochanteral joint.Cth, cephalothorax.Cv, neck, cervix.cv, cervical sclerite.Cx, coxa.dap, dorsal apodemal plate of postoc-cipital ridge.DMcl, dorsal longitudinal body muscle.DNv, dorsal longitudinal nerve.DT, dorsal arm of tentorium.dt, attachment of dorsal tentorial armto wall of cranium.E, compound eye.Ecd, ectoderm.End, endoderm.Endp, endopodite.Ephy, epipharynx, epipharyngeal sur-face.Eps, episternum.
154 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 8lER, epistomal ridge.es, epistomal suture.Exp, exopodite. KLt, ventral adductors arising on ten-torium, or hypopharyngeal apo-demes.
F, femur.fga, flexor of galea.Fl, flagellum.ftc, flexor of lacinia.flee, cranial flexor of lacinia.fles, stipital flexor of lacinia.For, foramen magnum, or " occipital "foramen.Fr, frons./;-, " adfrontal "FrGng, frontal ganglion.fs, frontal suture.ft, femoro-tibial joint.
Ga, galea.GC, gastric caecum.Gc, gastrocoele, archenteron.Gch, gnathochilarium.Ge, gena.Gl, glossa.Gn, gnathal segments.Gnc, gnathocephalon.Gng, ganglion.Gil, gula.H. head.Hphy, hypopharynx.Hst, hypostoma.
/, tergal promoter muscle of an appen-dage.I-VI, segments of the head.Isg, intersegmental fold.
/, tergal remoter muscle of an appen-dage.K, sternal promotor muscle of an ap-pendage.KL, ventral adductor muscles.KLh, ventral adductors arising on hy-popharynx.KLk, ventral adductors united by liga-ment {k) forming "dumb-bellmuscle."
L, leg. iL, first leg. L^, prothoracicleg.sternal remoter muscle of an appen-dage.LB. primitive limb base (coxa andsubcoxa).Lb, labium.LbNv, labial nerve.Ibmcl, labial muscles.Lc, lacinia.Lni, labrum.LNv, lateral stomodeal nerve.Md, mabdible.MdC, mandible cavity.MdNv, mandibular nerve.Ment, mesenteron.Mps, mouth parts.Msb, primary mesoblast.Mse, mesenchyme.Msd, mesoderm.Mst, metastomium.Mt, mentum.Mth, mouth.Mx, maxilla.iMx, first maxilla.2Mx, second maxilla.MxC, maxilla cavity.MxNv, maxillary nerve.NC, nerve cord.NMb, neck membrane.NpJi, nephridium.O, ocellus.levator muscle of palpus, or of tro-chanter.Oc, occiput.OcR, occipital ridge.ocs, occipital suture.OE, oesophagus.OcGng, oesophageal, or posterior me-dian stomodeal ganglion.OpL, optic lobe.OR, ocular ridge.OS, ocular suture.
NO. 3 INSECT HEAD—SNODGRASS 155
P, thoracic depressor muscle of tro-chanter.PcR, posterior cranial ridge.Pdc, pedicel.Pdp, pedipalp.Pge, postgena.Pgl, paraglossa.Ph, phragma.Phy, pharynx.PLGng, posterior lateral stomodealganglion.Pip, palpusPnt, postantennal appendage.Poc, postocciput, postoccipital rim offoramen magnum.PoR, postoccipital ridge.pos, postoccipital suture.Pp, " pleuropodium," specialized ap-pendage of first abdominal seg-ment.Ppd, parapodium.Ppt, periproct.PrC, preoral cavity.Pre, protocephalon.Priit, preantennal appendage.Proc, proctodeum.Prst, peristomium.Prtp, protopodite.Pst, prostomium.PT, posterior arm of tentorium.pt, posterior tentorial pit.Ptar, praetarsus.Q, depressor muscle of palpus, or oftrochanter.Rd, posterior fold of tergum.
rh, retractor of hypopharynx.RNz>, recurrent nerve.
SA, sternal apophysis.Sep, scape.Sex, subcoxa.Ser, serosa.Set, seta, setae.SgR, subgenal ridge.sgs, subgenal suture.SID, salivary duct, silk gland duct.SJO, opening of salivary duct.Sint, submentum.SeoGng, suboesophageal ganglion.Sp, spiracle, iSp, first thoracic spir-acle.Spn, spina.Spt, spinneret.St, stipes.Stoni, stomodeum.
T, tergum.depressor muscle of tibia.Tar, tarsus.Tb, tibia.Th, thorax.77, tentacle.Tip, telopodite.Tnt, tentorium.Tor, torma.Tr, trochanter.
J\ fifth head segment.vap, ventral apodemal plate of postoc-cipital ridge.VI, sixth head segment.VMcI, ventral longitudinal body mus-cle.VNC, ventral nerve cord.VNv, ventral longitudinal nerve.Vx, vertex.
REFERENCESAttems, C. (1926). Myriapoda. Krumbach's Handbneh dcr Zoologie, 4. Ber-lin, Leipzig.Balfour, F. M. (1880). Notes on the development oT the Araneina. Quart.Jonrn. Micr. Set., n. s., 20: 1-23, pis. 19-21.(1883). The anatomy and development of Peripatus capensis. Quart.Journ. Micr. Sci., n. s., 23:213-259, pis. 13-20.Basch, S. (1865). Untersuchungen iiber das Skelet und die Muskcln desKopfes von Termes flavipes (Kollar). Zeit. luiss. Zool, 15:56-75, pl- 5-^
156 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81
Berlese, a. (1909). Gli Insetti, i. Milan.(1910). Monografia dei Myrientomata. Redia, 6:1-182, pis. 1-17.Bidder, G. P. (1927). The ancient history of sponges and animals. Rep. BritishAssoc. Adv. Sci., 1927:58-71.BoRNER, C. (1909). Neue Homologien zwischcn Crustaceen and Hexopoden.Die Beissmandibel der Isekten und ihre phylogenetische Bedeutung. Zool.Anc, 34: 100-125, 9 figs.(1921). Die Gleidmassen der Arthropoden. Lang's Handhuch dcr zvir-bellosen Tiere, 4:649-694, 57 figs. Jena.BoRRADAiLE, L. A. (1917). On the structure and functions of the mouth-partsof the palaemonid prawns. Proc. Zool. Soc. London, 1917:37-71, 51 figs.CoMSTOCK, J. H., and Kochi, C. (1902). The skeleton of the head of insects.Amer. Nat., 36: 13-45, 29 figs.Crampton, G. C. (1916). A comparative study of the maxillae of the Acridiidae(Oedipodinae and Tettiginae), Phasmidae and Phylliidae. Psyche, 23: 83-87,pi. II.(1921). The sclerites of the head, and the mouth parts of certain im-mature and adult insects. Ann. Ent. Soc. Amer., 14:65-103, pis. 2-8.(1921a). The origin and homologies of the so-called " superlinguae "or " paraglossae " (paragnaths) of insects and related arthropods. Psyche.28:84-92, pi. 5.(1921b). The phylogenetic origin of the mandibles of insects and theirarthropodan relatives. A contribution to the study of the evolution ofArthropoda. Journ. Nczu York Ent. Soc, 29:63-100, pis. 6-8.(1922). A comparison of the first maxillae of apterygotan insects andCrustacea from the standpoint of phylogeny. Proc. Ent. Soc. Washington,24: 65-82, pis. 8, 9.(1925). The external anatomy of the head and abdomen of the roach,Periplaneta americana. Psyche, 32: 195-220, pis. 5-7.(1928). The eulabium, mentum, submentum and gular region of in-sects. Journ. Ent. and Zool., 21: 1-15, 3 pis.(1928a). The evolution of insects, chilopods, diplopods, Crustacea andother arthropods indicated by a study of the head capsule. Canad. Ent., 60:129-141, pis. 8-12.Croneberg, a. (1880). Ueber die Mundtheile der Arachniden. Arch. Natiir-gesch., 46 (i) 1285-300, pis. 14-16.Daiber. Marie (1921). Crustacea. Lang's Handhuch der Morphologic derzvirbcllosen Tiere, 4:9-252. Jena.EiDMANN, H. (1925). Vergleichend-anatomische Studien iiber die Pharynxmus-culatur der Insekten. Zool. An::., 62:49-64.Evans, A. M. (1921). On the structure and occurrence of maxillulae in theorders of insects. Journ. Linn. Soc. Zool., 34:429-456, pi. 31.FoLSOM, J. W. (1899). The anatomy and physiology of the mouth-parts ofthe collembolan, Orchesella cincta. Bull. Mus. Comp. Zool., 35:7-39,pis. 1-4.
• (1900). The development of the mouth-parts of Anurida maritimaGuer. Bidl. Mus. Comp. Zool., 36:87-157, pis. 1-8.
I NO. 3 INSECT HEAD SNODGRASS I57
GoLDi, E. A. (1913)- Die bisherige Anschauung bezuglich der Homologie derInsekten-Mundteile mit der Derivaten des Spaltfusses der Crustaceen undeine notwendig gewordene Modification. Mitt. Schweiz. Entom. Gcsells.,12: 146-151, I fig.Goodrich, E. S. (1898). On the relation of the arthropod head to the annelidprostomium. Quart. Journ. Micr. Sci., n. s., 40:247-268, 12 figs.DE Gryse, J. J. (1915). Some modifications of the hypopharynx in lepidopter-ous larvae. Proc. Ent. Soc. Washington, 17: 173-178, pis. 17-19.Hansen, H. J. (1893). Zur Morphologic der Gliedmassen und Mundtheile beiCrustaceen und Insekten. Zool. Anz., 16: 193-198, 201-212.Hatschek, B. (i888-'9i). Lehrbuch der Zoologie (first to third parts). Jena.Henriksen, K. L. (1926). The segmentation of the trilobite's head. Saert.Medd. Dansk. gcolog. For., 7: 1-32, 27 figs.Heymons, R. (1895). Die Embryonalentwickelung von Dermapteren undOrthopteren. 136 pp., 12 pis. Jena.(1897). Entwicklungsgeschichtliche Untersuchungen an Lcpisma sac-charina L. Zeit, zviss. Zool., 62:581-631, pis. 29, 30.(1901). Die Entwicklungsgeschichte der Scolopender. Zoologica. Orig.-Abhandl. Gesammt. Zool., S3, 244 pp., 8 pis.Hoffmann, R. W. (1911). Zur Kenntnis der Entwicklungsgeschichte derCollembolen. (Die Entwicklung der Mundwerkzeuge von Tomocerusplumbeus L.) Zool. Anz., ST. 2)S3-2>77, 7 figs-Holmgren, N. (1909). Termitenstudien. i. Anatomische Untersuchungen.Kuiigl. Svenska Vetenskapsakad. Handl., 44, No. 3, 215 pp., 3 pis.(1916). Zur vergleichenden Anatomic des Gehirns von Polychaeten,Onycophoren, Ziphosuren, Arachniden, Crustaceen, Myriopoden, undInsekten. Kungl. Svenska Vetenskapsakad. Handl., 56, No. i, 303 pp.,12 pis.Hu.ssey, Priscilla B. (1926). Studies on the pleuropodia of Belostomaflumineum Say and Ranatra fusca Palisot de Beauvais. EntomologicaAmericana, 7: 1-59, pis. i-ii.Janet, C. (1899). Essai sur la constitution morphologique de la tete de I'insecte.74 pp., 7 pis. Paris.(1911). Constitution morphologique de la bouche de I'insecte. 35 pp.,2 pis. Limoges.Jaworowski, a. (1891). Ueber die Extremitaten bei den Embryonen derArachniden und Insecten. Zool. Anz., 14: 164-169, 173-176, 4 figs.KoRSCHELT, E. and Heider, K. (1895). Text-book of embryology of inverte-brates. English transl., Vol. i.Kovvalevsky, a. (1883). Embryogenie du Chiton polii (Philippi). .Inn. Mus.Hist. Nat. Marseille, Zool., i, 46 pp., 8 pis.Kowalevsky, a. and Marion, A. F. (1883). Documents pour I'histoire em-bryogenique des Alcyonaires. Ann. Mus. Hist. Nat. Marseille, Zool., i,50 pp., 5 pis.Lankester, E. R. (1881). Observations and reflections on the appendages andon the nervous system of Apus cancriformis. Quart. Journ. Micr. Sci.,n. s., 21:343-376, 3 figs.MacGillivray, a. D. (1923). External insect-anatomy. Urbana, 111.Mangan, J. (1908). On the mouth-parts of some Blattidae. Proc. R. IrishAcad., 27, B: i-io, pis. 1-3.
158 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 81Meinert, F. (1867). On the Campodeae, a family of Thysanura. Englishtransl., Ann. Mag. Nat. Hist., ser. 3, 20: 361-378, 3 figs.Patten, W. (1912). The evolution of the vertebrates and their kin. Phila-delphia.Prell, H. (1913). Das Chitinskelett von Eosentomon. Zoologica Orig.-Abhandl. Gesammt. Zool., 25, 58 pp., 6 pis.Ramme, W. (1913). Die Bedeutung des Proventriculus bei Coleopteren undOrthopteren. Zool. Jahrh., Anat., 35:419-456, pis. 22-24.Reichenbach, H. (1877). Die Embryonalanlage und erste Entwicklung desFlusskrebses. Zeit. miss. Zool., 29: 123-196, pis. 10-12.Riley, W. A. (1904). The embryological development of the skeleton of thehead of Blatta. Amer. Nat., 38:777-810, 12 figs.Schmit-Jensen, H. O. (1913). Homeotisk Regeneration af Antennen hos enPhasmide, Carausius (Dixippus) morosus. Vidensk. Med. Dansk. naturh.Foren., 65: 113-134. English transl., Ann. Rep. Smithsonian Inst, for1914: 523-536, 7 figs.Schmidt, W. (1915). Die Muskulature von Astacus fluviatilis (Potamobiusastacus L.) Zeit. zuiss. Zool., 113: 165-251, 26 figs.Snodgrass, R. E. (1927). The head and mouth parts of the cicada. Proc. Ent.Soc. Wash., 29: 1-16, pi. I.(1927a). Morphology and mechanism of the insect thorax. Smith-sonian Misc. Coll., 80, No. I, 108 pp., 44 figs.Stickney, F. S. (1923). The head-capsule of Coleoptera. Illinois Biol. Mono-graphs, 8, No. I. 51 pp., 2() pis.voN Stummer-Traunfels, R. (1891). Vergleichende Untersuchungen iiberdie Mundwerkzeuge der Thysanuren und Collembolen. Sitz. Kaiscrl. Akad.Wiss. Wien., math.-natur., 100, Abth. 1:216-235, 2 pis.UzEL, H. (1897). Vorlaufige Mitteilung iiber die Entwicklung der Thysanuren.Zool. Anz., 20: 125-132.(1897a). Beitrage zur Entwicklungsgeschichte von Campodea staphy-linus Westw. Zool. Ans., 20:232-237.Walton, L. B. (1927). The polychaete ancestry of the insects. I. The exter-nal structure. Amer. Nat., 61:226-250, 11 figs.Wheeler, W. M. (1889). The embryology of Blatta germanica and Dory-phora decimlineata. Journ. Morph., 3:291-386, pis. 15-21.
' (1893). A contribution to insect embryology. Journ. Morph., 8:1-160, pis. 1-6.Wiesmann, R. (1926). Zur Kenntnis der Anatomic und Entwicklungsge-schichte der Stabheuschrecke Carausius morosus Br. III. Entwicklung undOrganogenese der Colomblasen, pp. 123-328, 86 figs. Published by Zool.-vergl. anat. Inst. Univ. Zurich.Yuasa, H. (1920). The anatomy of the head and mouth-parts of Orthopteraand Euplexoptera. Journ. Morph., 33:251-290, 9 pis.Zograf, N. (1883). Contributions to a knowledge of the embryological de-velopment of Geophilus ferrugineus L. K. and Geophilus proximus L. K.(in Russian). Studies Lab. Zool. Miis. Moscoiv Univ., 2, pt. i, 77 pp.,108 figs.