of Descen$ of Insectsdownloads.hindawi.com/journals/psyche/1938/018939.pdfinsects, since they are much more specialized than the Machilide. In fact, the Dicellura could readily be

1938] Lines of Descen$ of Insects 165

THE INTERRELATIONSHIPS AND LINES OFDESCENT OF LIVING INSECTS

BY G. C. CRAMPTON,

Massachusetts State College, Amherst, Mass.

The lines of descent shown in the accompanying phylo-genetic tree should be interpreted as though the figure werea three dimensional one (as is indicated by the basal attach-ments of the branches of the tree), since the usual methodof portraying the lines of descent in the form of a dicho-tomously branching tree, drawn in one plane, does not bringout the fact that several lines of descent may converge upona common ancestry, and does not indicate the complicatedinterrelationships of these lines of descent at all accurately.In fact, sections of cones made up of converging lines wouldbetter illustrate the fact that some lines of descent inter-grade "horizontally" as well as "vertically," but the methodof illustrating the interrelationships of the lines of descentshown in the accompanying figure will serve well enoughfor all practical purposes, if the figure is interpreted as athree dimensional one.The hypothetical "Protomalacostraca" shown at the base

of the phylogenetic tree, represent the extinct common an-cestors of the higher Crustacea (such as the Tanaidacea,Mysidacea, Anaspidacea, etc.) insects and "myriopods".The character of the head, with its sessile eyes, the mono-condylar mandibles, with their differentiated incisor andprojecting molar regions, the large paragnaths, the slendermultiarticulate, cerci-like uropods, and other feature ofthe Tanaidacea (such as Tanais, Apseudes, Leptochelia,etc.) are strikingly suggestive of the precursors of similarstructures in the Machiloid insects, and the ancestors ofthe Tanaidacea (represented by the hypothetical "Proto-tanaidacea in the diagram) must have been extremely closelyrelated to the more dir.ect ancestors of the Hexapoda (rep-resented in the diagram by the hypothetical "Protohexa-

166 Psyche [December

poda") and those of the ’"Myriopoda" (represented in thediagram by the hypothetical "Protosymphyla"whichwould include the direct ancestors of the Chilopoda, Pauro-poda, etc., as well as those of the Symphyla themselves).The hypothetical "Protapterygota," or common ancestors

of the Apterygota (and consequently the ultimate ancestorsof the Pterygota also), are best represented by the Machilis-like Apterygota, although some investigators insist that theDicellura (such as Campodea, Ana]apyx, Japyx, etc.), orthe Protura, are the most primitive or most "ancestral" in-sects, despite the fact that the Machilis-like Apterygota aremorphologically the most primitive (and hence the most"ancestral") of all insects. Some of the features whichindicate that Machilis (or the family Machilidm), ratherthan the Dicellura or other forms, represents the ancestraltype better than any other living insect, is indicated by thefollowing facts. (1) Machilis has more abdominal limbs(represented by eight pairs of distinct, styli-bearing cox-ites) than any other insect, including Campodea and otherDicellura. (2) The abdominal limbs, or coxites, of Machilisproject free and distinct from the sternites (and the coxitesof the ninth segment are hugely developed), while theabdominal limbs of the Dicellura, for example, are reducedto mere styli-bearing areas scarcely distinguishable fromthe abdominal sternites with which they have merged andno other Apterygota have more than three pairs of abdomi-nal limbs. (3) The cerci, or limbs of the eleventh abdominalsegment, are larger and better developed than those of anyother insect. (4) The terminal abdominal segments ofMachilis are more distinct, or less iused, than those of theDicellura, for example. (5) The thoracic terga are betterdeveloped and overlap the lateral regions in a more primi-tive, or Crustaceoid fashion, in Machilis (and Lepisma),while in other Apterygota the thoracic terga are not of thistype. (6) The thoracic limbs are best developed in Machilis,and bear styli (epipodites?) in the Machilidae alone.(7) Machilis has primitive compound eyes structurally simi-lar to those of Crustacea, and also has well developed ocelli(which likewise occur in certain trilobites, Anaspidacea,etc.), while the Dicellura are eyeless as is also the casein the Protura (which likewise lack antennae). (8) Machi-

1938] Lines of Descent of Insects 167

lis has preserved the primitive archicephalic or supraman-dibular suture, characteristic of such Crustacea as Branchip-pus, Anaspides, etc. (9) The huge flagelliform antennm ofMachilis are the most like those of trilobites and other primi-tive arthropods. (10) The huge monocondylar mandiblesof Machilis are better developed and are more Crustaceoid(with separate incisor and elongated molar regions) thanthe reduced and highly specialized mandibles of the Dicelluraand Protura. (11) The well developed paragnaths (super-linguee) of Machilis are larger and more Crustaceoid thanthose of other Apterygota. (12) The well developed maxillmof Ma.chilis, with their huge limb-like palpi, are far moreprimitive than those of any other insects. (13) Thelacinial fringes of Machilis are more primitively Crustaceoidthan those of other insects. (14) The labial palpi etc., ofMachilis are better developed, and are of a more primitivecharacter than those of other Apterygotan insects. (15)The traces of the second maxillm forming the under lip aremore distinct in Machilis than in other Apterygota. (16)The head of Ma.chilis is of the ectognathous type, and ismuch more primitive than the entognathous type (withovergrown mouthparts) found in the Dicellura and Pro-tura, etc.

There are many more features which might be cited toprove that Machilis is more primitive or "ancestral" thanany other insect, but the facts cited above should be suffi-cient to convince any impartial investigator that Mac.hilisis the most ancestral of all insects. Furthermore, its numer-ous unmistakably Crustaceoid features clearly prove thatMachilis, and the other primitive insects descended fromsimilar ancestors, were derived from Crustacea resemblingthe "Prototanaidacea" in many respects. The well knownfact that many Apterygota exhibit striking similarities tovarious types of "myriopods," merely indicates that bothinsects and "myriopods" were ultimately descended fromthe same (Crustacean) ancestry and the Crustacea arethe only intermediate forms serving to connect the insectsand "myriopods" with the trilobites and other primitiveforms at the base of the common arthropodan stem.

If the Machilidm are the most primitive or "ancestral"insects, it is obviously misleading to insist that the Dicellura


(such as Campodea, Anajapyx, .etc.) are the most ancestralinsects, since they are much more specialized than theMachilide. In fact, the Dicellura could readily be derivedfrom a slender blind Nicoletia type of Lepismatid Tysanu-roid insect (descended from Machilis like forebears) butby no stretch of the imagination could the primitive Mac’hilistype of insect be derived from any Dicelluran type whichshould be the case if the Dicellura represent the ancestralinsects (leading back to some type of "myriopod").

Because they have mouthparts of the concealed type, theDicellura are sometimes grouped with the Collembola (andProtura) in the division Endognatha, in contradistinctionto the Ectognatha, or Thysanuroid forms with mouthpartsof the exposed type. The modifications of the mouthpartsof the Dicellura, however, are not very similar to the modi-fications exhibited by the Collembola and Protura; and theDicellura are only very distantly related to the other"endognathous" Apterygota, while the occurrence of cerci,and styli-bearing coxites, in the Dicellura, allies them moreclosely with the Thysanuroid Apterygota. It is thereforepreferable to unite the Dicellura with the Lepismatidm(Thysanura) and Machilidae (Protothysanura) in the sec-tion Styligera, characterized by the occurrence of styli-bearing coxites, and cerci, as opposed to the Astyligera(Collembola and Protura), in which these structures arelacking. The Dicellura were possibly derived from someNicele$ia-like Lepismatid insect, and the Lepismatidm them-selves were apparently derived from ancestors closely alliedto the Machilide. The Lepismatidm, in turn, serve to con-nect the Ma.chilis-like ancestors of the Apterygota with theancestors of the lterygota, which are best represented by"larval" Ephemerida (which have three caudal filamentslike those of Thysanuroid Apterygota), so that the Lepis-matid line of development is an extremely important oneor the study of the evolution of the higher insects.The Protura are the most primitive representatives of

the section Astyligera, although they have lost the eyes andantennae, and their mouthpar.ts are rather highly specialized.They differ from the rest of insects by the fact that theyexhibit a postembryonic increase in the number of segments(they have nine abdominal segments as "larvae", and eleven


as adults) and are sometimes called Anamerentoma, in con-tradistinction to the rest of insects (called Holomerentoma)which exhibit no such postembryonic increase in segmenta-tion. Their line of development evidently branched off atthe base of the Apterygotan stem, and ends blindly, unlessit leads to the Collembola.The line of development of the Collembola is a rather

isolated one, but the Collembola resemble the Protura in thatthey have a ventral head-groove, a postantennal organ, andsimilarly modified endognathous mouthparts, etc. TheCollembolan line of development may have branched offfrom that of the Protura, although it is also possible thatthe Collembola represent degenerate offshoots of the primi-tive Machiloid ancestors of the Apterygota, since the lacinialfringes of such Collembola as Tetrodontophora are verylike those of certain Machilids. In any case, the line ofdevelopment of the Collembola is a very isolated one, andhas no significance for tracing the lines of descent of otherinsects.As was mentioned before, the "larvm" (naiads) of the

Ephemerida, with their Lepisma-like terminal filaments (apair of cerci and an unpaired median terminal filament),and their large paragnaths (superlingum), and primitivetype of mandibles, provided with a lacinia mobilis like themandibles of higher Crustacea, etc., are the most archaicrepresentatives of the Pterygota, and suggest that wingedinsects arose irom Lepisma-like orebears, which lead backto the Crustaceoid Apterygota such as Machilis. "Larval"Ephemerida and Odonata agree in having the lacinia andgalea united to orm a single lobe in the maxilla, and "larval"Ephemerida and Zygoptera agree in having an unpairedmedian terminal structure, represented by a gill plate inthe Zygopteran naiad, and by a terminal filament in theEphemerid naiad.The Ephemerida, Odonata, Megasecoptera and Palmo-

dictyoptera, etc., comprise the division of Pterygota calledthe Paltvopterygoa, characterized by their inability to laythe wings back along the body in repose. They consequentlydo not develop a basal fold of the wing, and do not havemore than two or three axillary sclerites. Their wings areprimitively homonomous, and a neala is not developed in


them. The rest of the Fterygota comprise the divisionNeopterygota, characterized by their ability to lay thewings back along the body in repose. They consequentlydevelop a basal fold in the wing, and have more than twoor three axillary sclerites. Their wings were originallyheteronomous, with an anal fan in the hind wing (thoughthis is lost in many of their descendents), and a neala isdeveloped in the postero-basal region of the wings. Thisdivision of the Pterygota is a much more fundamental onethan the usual division into Exopterygota and Endoptery-gota (on the basis of the external or internal developmentof the wings) which makes an unnatural separation ofclosely related forms, and lumps together others which arenot at all closely related.The common ancestors of all of these forms are repre-

sented in the diagram by the hypothetical "Protopalmo-dictyoptera". The Palmodictyoptera are the nearest knownrepresentatives of these common ancestors, and representtheir direct descendents. The Ephemerida were probablyderived from the common ancestors of the group by way ofthe Protephemerida, while the Odonata were derived fromthem by way of some unknown, extinct forms, to which theProtodonata are very closely related; and the line of descentof the Megasecoptera apparently branched off from thecommon Palmodictyopteriod stock near the origin of theancestors of the Odonata. The Odonata are the most"Orthopteroid" of the above-mentioned insects, while theEphemerida are the most primitive living representativesof the group. The Ephemerida and Protephemerida mightbe grouped into a superorder called the Panephemeroptera(or Ephemeropteria), characterized by the occurrence ofthree caudal filaments in many members of the superorder,while the Odonata and Protodonata might be grouped in asecond superorder, the Pantyloptera (or Tylopteria) char-acterized by the skewness of the thorax etc., but too little isknown of the morphological details of the fossil forms toenable us to group them correctly at this time.The Neopterygota may be grouped into three divisions

called the Orthopteroid insects (Paurometabola or Orthop-teradelphia), the Hemipteroid insects (Parametabola orHemipteradelphia), and the Neuropteroid insects (Holo-


metabola or Neuropteradelphia); and the Orthopteroidinsects represent the ancestral types of the group as nearlyas any known forms. The Orthopteriod insects were ap-parently not derived directly from the Paleeodictyoptera,but were probably derived from the Palmodictyopteranstock by way of Synarmoge (or Syna.rmogoge as it is some-times spelled). Since only a fragment of one wing ofSynarmogoge is known, however, all that can be said con-cerning it is that it exhibits certain characters intermediatebetween the Palaeodictyoptera and the Protorthoptera.The Protorthoptera, shown at the base of the lines of de-

scent of the higher insects in the accompanying phylogenetictree, include the Protoblattids (which have a demarkedclaval region in the fore wings) as well as the Protorthopterain the narrower sense (which have no demarked clavalarea), since the Protoblattids and Protorthoptera merge soindistinguishably that they may be combined into a singleancestral group from which all of the higher insects wereultimately derived. The Protoblattids are the most primi-rive representatives of the group, and are more like thedirect ancestors of the Blattids, Mantids and Isoptera, whilethe other members of the Protorthoptera are somewhatcloser to the direct ancestors of the Orthoptera.The Orthopteroid insects are characterized by the fact

that the cerci are well developed and the parapodial plates(paraprocts) are distinct in all of the members of the group.An anal fan is developed in the hind wings of most of theOrthopteroid insects, but in some of them, such as theEmbiida and Isoptera (excepting Mastotermes) the wingsare secondarily homonomous. The Orthopteroid insectshave been grouped into three superorders called thePanorthoptera (or Orthopteria), the Panplecoptera (orPlecopteria), and the Panisoptera (or Isopteria).The superorder Panisoptera (Isopteria) includes the

Paleeoptera (Blattids and Mantids) and the Isoptera, andis characterized by the fact that the lateral cervical scleritesare contiguous in the midventral line, the mesothoracictrochantins do not unite basally with the episternum; aclaval area is usually demarked in the fore wings, and theseventh abdominal sternite projects below the ovipositor inthe members of this superorder. The Blattids have pre-


served the mo.st primitive venation of any living membersof the superorder, while the Isoptera have preserved thevarious features of the body in as primitive a condition asany members of the superorder. These insects are themost primitive of the Orthopteroid insects, and are prac-tically the direct descendents of the Protoblattid type ofProtorthoptera.The Embiids and Plecoptera are included in a superorder

called the Panplecoptera (or Plecopteria), characterized bythe fact that the postscutellum of the mesothorax is welldeveloped, and the trochantin of the mesothorax unitesbasally with the episternum above it, in the members of thissuperorder. The mesothoracic coxm tend to become ring-like rather than conical, and the tarsi are trimerous. Theeighth and ninth abdominal segments are not greatly nar-rowed in the females of these insects, which are ovipositor-less. The Plecoptera are usually grouped with the Odonataand Ephemerida (rather than with the Embiids among theOrthopteroid insects) but the character of their thoracicsclerites is so strikingly similar, and the venation of thefossil forms intergrades so markedly, that there can be nodoubt that the Embiids and Ilecoptera are extremely closelyrelated, and were descended from a common Protorthopteranancestry. The fossil Protoperlaria are rather specializedPlecopteroid insects which branched off at the base of thePlecopteran stem, and the fossil Protembiids apparentlybranched off at the base of the Embiid stem, but the actualProtorthopteran ancestors from which all of these insectswere ultimately derived have not as yet been discovered.The Orthoptera (including the Grylloblattidm) and the

Cheleutoptera, or Phasmida, and possibly the Dermapteraalso (although the closest affinities of the Dermaptera maybe with the Blattoid insects comprising the superorderPanisoptera), are included in the superorder Panorthoptera(or Orthopteria), characterized by the huge development ofthe anal fan, and the consequent reduction of the preanalregion of the hind wings. The ovipositor is well developedin most of them, and is overlapped basally by the eighthabdominal sternite (excepting the Dermaptera, which maynot belong in this superorder). The male genitalia areusually rather symmetrically developed and the cerci fre-


quently bear mesal prongs, etc., in the members of thissuperorder.

These insects are the more or less direct descendents ofthe Protorthoptera in the restricted sense (i.e., the Pro-torthopteran forms other than the Protoblattids), and theirmost primitive representatives are the Grylloblattids, whichare practically living Protorthoptera, closely related to theStenopelmatoid Orthoptera (including the Gryllacris-types).The latter are connected with the Grylloid Orthoptera bythe Prophalangopsidm (and Tridactyloid types), and areconnected with the Acridoidea by the Tettigoniide (unlessthe Tridactyloidea furnish the intermediate forms leadingto the Acridoidea).The Hemipteroid or Psocoid insects comprise the division

Hemipteradelphia, or Parametabola, characterized by thedevelopment of a mesal detached lacinial structure forminga setiform, or a chisel-like portion of the maxilla (exceptingthe Zoraptera, which have a normal type of maxilla). Theinsects belonging to this division may be grouped into twosuperorders, namely the Panpsocoptera (Psocopteria) in-cluding the Psocoptera, Mallophaga and Anoplura, and thePanhemiptera (Hemipteria) including the Hemiptera andThysanoptera. The lacinial structures are usually chisel-like in the members of the superorder Panpsocoptera,and are setiform in the members of the superorderPanhemiptera.The Zoraptera are the most primitive representatives of

the Hemipteroid insects, and exhibit so many characterssuggestive of a close relationship to the Isoptera, that thismight be taken to indicate that they and the Hemipteroidinsects in general were derived from the same Protorthop-teroid ancestry as the Isoptera were. On the other hand,the venation of the Zoraptera shows that they are membersof the order Psocoptera, closely related to such Psocids asArchipsocus and Embidopsocus which exhibit some char-acters suggestive of a derivation from Embiid-like ances-tors; and the Psocoptera in general were probably derivedfrom the common Protorthopteroid ancestors of the Isopteraand Embioptera. The Mallophaga are undoubtedly de-scended from ancestors extremely closely related to thePsocoptera, and may represent merely degenerate wingless

174 Psyche Decembe

Psocoptera. The Anoplura were probably derived fromancestors closely allied to the Mallophaga; and all of theseinsects are sometimes grouped together as Corrodentia,although they are apparently worthy of ordinal rank.The mouthparts of the Thysanoptera (which have distinct

maxillary and labial palpi) are much more primitive thanthose of the Hemiptera, and it is possible that the Thysanop-tera were descended rom a slightly more primitive typethan the Hemiptera were. At any rate, the Thysanopterawere apparently descended from Protorthopteroid ancestorsvery closely related to those from which the Hemiptera werederived, and these ancestors evidently resembled the Psocop-tera very closely. The Hemiptera were evidently descendedfrom the same Protorthopteroid ancestors from which thePsocoptera were derived, and the Psocopt.era have departedthe least of any living insects from the types ancestral tothe Hemiptera. Not only does the venation of livingPsocoptera parallel that of certain Hemiptera-Homopterastrikingly closely, but the venation of certain fossil Hemip-tera merges with that of certain fossil Psocoptera so in-timately that there can be no doubt that the two groups hada common Protorthopteroid ancestry; and the iact thatmany Hemiptera have a claval area demarked in their orewings may possibly indicate that their Protorthopteroidancestors resembled Protoblattids (in which the claval areais also demarked) in some respects. At any rate, theHemiptera could not possibly have been derived from suchPalmodictyopteroid insects as Eugereon (mistakenly called"Protohemiptera"), since Eugereon belongs in the sectionPalmopterygota, whose members are incapable of laying thewings along the body in repose, while the Hemiptera wereevidently descended from ancestors capable of laying thewings back along the body in repose, and the venation oprimitive Hemiptera does not bear the slightest resemblanceto that of Eugereon.Some Parametabola (Hemipteroid insects) parallel the

Holometabola remarkably closely in their method of devel-opment, and indicate very clearly that complete meta-morposis arose through an increasing divergence betweenthe immature and mature forms (rather than through theprecocious emergence of "i’ree-living embryos", as certain


investigators insist is the case) as the result of mutationalchanges in the genetic or hereditary material of theseinsects. Thus in larval Aleurodidm, for example, the wingsarise internally 2rom wing buds, and become external in aquiescent pupal stage, as they do in the Holometabola;and the development of certain Thysanoptera and otherHemipteroid insects clearly suggests the beginning oHolometabolism.

In this connection, it may be noted that the internal de-velopment o the wings in the Aleurodidm, and the .externaldevelopment of the wings in the closely related Psyllidm andother Hemiptera-Homoptera, clearly indicates that the usualdivision of winged insects into Exopterygota and Endoptery-gota, on the basis of the external or internal development ofthe wings, is utterly meaningless irom the standpoint ofphylogeny. Such a division would group together theremotely related Palmodictyopteroid, Orthopteroid andHemipteroid insects, and would separate the Holometabolarom their Orthopteroid relatives, etc. and this division ofwinged insects should be abandoned in avor of the morenatural and 2undamental grouping of winged insects intoPalmopterygota and Neopterygota (on the basis of themethod of olding the wings in repose), since this groupingdoes not separate the Holometabola rom their Hemipteroidand Orthopteroid relatives.The section Holometabola (or Neuropteradelphia) in-

cludes all insects with complete metamorphosis, and thegroup is evidently a monophyletic one since their larvmintergrade so intimately that all of the Holometabola musthave had a common ancestry. The fact that some Holo-metabola are somewhat lsocid-like, and the fact that somePsocoid (or Hemipteroid) insects, such as the Aleurodidm,etc., clearly oreshadow Holometabolism, would seem to in-dicate that both Parametabola (Hemipteroids) and Holo-metabola were descended from closely allied Protorthopterawhich may have resembled the Protoblattids venationally,although their bodies probably exhibited features occurringin the ancestral Isoptera, Embiids and Grylloblattids.The most primitive representatives of the Holometabola

are the Neuroptera, Hymenoptera and Coieoptera, and their

A quiescent pupal stage is also ioreshadowed in the Isoptera.


complicated interrelationships (both "vertical" and "hori-zontal" in the phylogenetic groupings) make it very difficultto decide how to distribute them in the superorders of theHolometabola. There are at least two superorders ofHolometabola, in one of which, the Pancoleoptera orColeopteria (including the Coleoptera and Strepsiptera),the mesothoracic coxae are not divided into eucoxa andmeron, the mesothoracic postscutellum is not well developed,and the cerci are usually not developed, while in the othersuperorder called the Panneuroptera or Neuropteria (in-cluding the Neuroptera, Mecoptera, Diptera, Trichoptera,Lepidoptera and Siphonaptera) the mesothoracic coxae aredivided into a eucoxa and meron, the mesothoracic post-scutellum is usually well developed, and cerci are requentlypresent.The Coleoptera are the most Orthopteroid representatives

of the Holometabola, and are strikingly similar to theDermaptera in numerous features of the body. It is pos-sible that the Coleoptera and Dermaptera were derivedfrom Protorthopteroid ancestors which had bodies some-what like those of the primitive Isoptera, while the wings oftheir ancestors may have been like those of certain Proto-blattid Protorthoptera. At any rate, the wings of the so-called Protocoleoptera (such as Protocoleus) are veryOrthopteroid, and may have been derived from a Protor-thopteroid type related to the Protoblattids.The most primitive Coleoptera are the Cantharoid

(Lampyroid) beetles, and the Strepsiptera may have beenderived from ancestors resembling Cantharoid beetles insome respects. The larvm of the Strepsiptera are veryMeloid in appearance, while the adult Strepsiptera resembleRhipiphorid beetles in some respects, so that it is very prob-able that the Strepsiptera were derived from a Coleopteroidstock, although the exact character of their ancestors hasnot been determined.The Neuroptera have retained the most primitive type of

venation occurring in any Holometabola (with the possibleexception of the Mecoptera), and the structures of the bodyof the primitive Sialid Neuroptera suggest that their an-cestors had bodies resembling those of primitive Isopteraand Embiids in many respects, while the venation of the


Corydalid Neuroptera is rather suggestive of that of certainProtoblattid Protorthoptera. The venation of the Protor-thopteroid insect Metropator (which is regarded as a"Palmodictyopteroid" insect by Handlirsch) exhibits certaineatures suggestive o the precursors of the Sialid type ofvenation, but Metropator is an oligoneurous form (with fewveins) and it is hardly probable that the more richly veinecl(polyneurous) types of Neuroptera, especially the ossilorms, were derived from the oligoneurous Metropator typeof Protorthopteroid insect. In act, it is very probable thatthe Protorthopteroid ancestors of the Neuroptera were botholigoneurous and polyneurous, and both tendencies wouldnaturally reappear in their Neuropterous descendents ifboth tendencies occurred in the ancestral stock. Some in-vestigators consider that the venation of the Neuropteraindicates that they were derived 2rom Palmodictyoptera, butthe Neuroptera are clearly Neopterygota capable of layingthe wings back along the body in repose, and their larvaeintergrade extremely closely with those of the Coleoptera,clearly indicating that the Neuroptera were derived romthe same Protorthopteroid ancestors as the Coleoptera, andthe venation o the primitive Neuroptera might readily bederived rom the type exhibited by certain Protoblattidtrotorthoptera.The Mecoptera are extremely closely relatecl to the

Neuroptera, and were evidently derived from the sameProtorthopteroid ancestors rom which the Neuropterawere descended. The venation of the primitive Mecopterasuggests that their ancestors were similar to the ProtoblattidProtorthoptera in certain respects, although Tillyard con-siders that the Mecoptera were derived rom ancestors ofthe Metro.pator type, and Tillyard likewise considers thatthe Mecoptera are more primitive than the Neuroptera.The head and mouthparts of the Mecoptera are more spe-cialized than those of the Sialid Neuroptera, however, andthe thoracic sclerites of the Neuroptera are of a much moreprimitive type than those of the Mecoptera, while someNeuroptera such as Raphidia have retained an Orthopteroidovipositor which is lost in typical Mecoptera, so that theNeuroptera are more primitive, in general, than the Mecop-tera are, and have departed less than the Mecoptera have,


from the ancestral Holometabolous stock, which was prob-ably a Protorthopteran type with body structures like thoseof primitive Isoptera and Embiids, and with a venation re-sembling that of certain Protoblattids. Handlirsch wouldderive the Mecoptera from M.egasecoptera, but the Megase-coptera belong in the section Paleopt.erygota, whose mem-bers are incapable of laying the wings back along the bodyin repose, while the Mecoptera are clearly Neopterygotacapable of laying the wings back along the body in repose,and the resemblance between the venation of certainMecoptera and Megasecoptera is apparently the result ofconvergence.The Diptera were undoubtedly descended from Mecoptera-

like forebears, as is evidenced by all of their structuralfeatures; and fossils such as Aristopsyche (called Paratri-choptera or Protodiptera), which have a venation strikinglysuggestive of the ancestors of the Diptera, merge so indis-tinguishably with the Mecoptera, that it is very doubtfulthat they are worthy of ordinal rank, and it is preferable togroup them with the Mecoptera as a suborder of Mecoptera.Among the living Mecoptera, such forms as Nannochoristahave preserved numerous features suggestive of the ances-tors of the Diptera, and clearly indicate that the labella ofDiptera, for example, are merely modified segments of thelabial palpi.The Trichoptera have likewise preserved a great number

of characters strikingly suggestive of the precursors of theDiptera, but the Trichoptera are more closely allied tothe Lepidoptera than any other insects. In fact, theRhyacophilid Trichoptera merge so indistinguishably withthe Micropterygid Lepidoptera that it is very doubtful ifthe two groups are worthy of ordinal rank, although it ispreferable to treat them as distinct orders for the sake ofconvenience. Fossils such as Belmontia (usually placed ina distinct order the Paramecoptera or Prototrichoptera)are strikingly suggestive of the common ancestors of theTrichoptera and Lepidoptera and lead back to ancestors2The Tanyderidae and Trichoceridae are the most primitive living

Diptera, and the Anisopodidae are very like the ancestors of theBrachycera, whose most primitive representatives are the Therevidaeand Rhagionidae (Leptidae), while the Syrphidae have departed butlittle from the ancestors of the Cyclorrhapha.

19381 Lines of Descent of Insects 179

resembling the fossil Mecoptera. The venation of Belmontiais so like that of certain Rhyacophilid Trichoptera that thedifferences hardly seem to be of ordinal value. At any rate,the Trichoptera have departed the least of any living insectsfrom the ancestors of the Lepidoptera, and the Mecopteraare the nearest living representatives of the forms ancestralto the Trichoptera.The origin and closest affi.nities of the Siphonaptera, or

fleas, is still a subject of much dispute, and it is impossibleto decide the question in the present state of our knowledgeof the group. It is quite evident, however, that the Sipho-naptera resemble both Diptera and Trichoptera in numerousfeatures of their larval and adult anatomy, and this prob-ably indicates that the fleas were descended from thecommon ancestors of the Diptera and Trichoptera. Sincethe Mecoptera are the nearest living representatives of thecommon ancestors of the Diptera and Trichoptera, the fleaswere doubtless derived from Mecopteroid ancestors as yetunknown.The Hymenoptera combine in themselves so many char-

acters occurring in both the Coleopteroid and the Mecop-teroid insects that it is very difficult to determine theirclosest affinities, although they have a greater number ofcharacters in common with the Mecopteroid insects thanwith any other group, and their larvae (particularly thoseof the sawflies) are strikingly similar to larval Mecoptera,Lepidoptera, etc. These facts may be interpreted as indi-caring that the Hymenoptera should be grouped with theMecopteroid insects in the superorder Panneuroptera(Neuropteria), but since the Hymenoptera are in manyrespects intermeliate between the Coleopteroid and Mecop-teroid insects their annectant character may be betterindicated by placing them in a distinct superorder, hePanhymenoptera (or Hymenopteria), occupying a positionintermediate between the Mecopteroid insects (Panneurop-tera) and Coleopteroid insects (Pancoleoptera), and char-acterized by the occurrence of an ovipositor (represented bya saw or a sting), forcipate male genitalia, cerci, anOrthopteroid head, etc., and by the absence of the meron inthe mesothoracic coxm.

Tillyard considers that the Hymenoptera were derived


from the so-called "Protohymenoptera", which have beenshown by Carpenter to be merely modified Megasecoptera,having nothing to do with the true ancestors of the Hymen-optera. The Megasecoptera belong in the division Palmop-terygota, whose members are incapable of laying the wingsback along the body in repose, while the Hymenoptera wereevidently descended from ancestors which were capable oflaying the wings back along the abdomen in repose. Theirancestors apparently were Protorthoptera with bodies likethose of primitive Isoptera and Embiids, while the venationwas probably like that of certain Protoblattids. In otherwords, the Hymenoptera were descended from a commonancestry with the Sialid Neuroptera (and the LampyroidColeoptera), and their line of descent also merges with thatof the M.ecoptera and Trichoptera, or branches off from thecommon stem near the point of origin of the lines of descentof the Mecopteroid insects. The most primitive living rep-resentatives of the Hymenoptera are the Xyelidm, and theCephidm are the nearest representatives of the ancestorsof the higher Hymenoptera (or Clistogastra), whose mostprimitive representatives are the Trigonalide.The views briefly summarized above, have already been

presented in a series of papers dealing with the comparativemorphology of recent insects in the light of what is knownof the fossil forms, but since these views differ very radicallyfrom those commonly accepted by recent writers, or by thosewho have reviewed the recent progress in insect phylogeny,they have not been taken into consideration by recentwriters. When a more extensive study of the availableevidence is made, however, it becomes readily apparent thatmany of the currently accepted views are quite untenable;and the foregoing brief summary of certain alternativeviews on the subject has been made in order to call attentionto the fact that the currently accepted views concerning theorigin and interrelationships of the insectan orders are notthe only possible ones, or necessarily the correct ones, andsome consideration should also be given to these alternativeviews if they are evidently more nearly in accord with theavailable evidence on the subject.


PSYCHE, VOL. XLV PLATE 17

PROTO/

,PALAEODICTYOPTERA

PROTAPTp!’Y’GOTAPROTOt’IE,XAPODA

I:tROTOTANAIDACEA tROTOSYMpHYLAPROTOM COSTRAGA

CRAMPTON- DESCENT OF INSECTS

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

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

Anatomy Research International

PeptidesInternational Journal of


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

International Journal of

Volume 2014

Zoology


Molecular Biology International

GenomicsInternational Journal of


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


BioinformaticsAdvances in

Marine BiologyJournal of



Signal TransductionJournal of


BioMed Research International

Evolutionary BiologyInternational Journal of



Biochemistry Research International

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


Genetics Research International


Advances in

Virolog y

Hindawi Publishing Corporationhttp://www.hindawi.com

Nucleic AcidsJournal of

Volume 2014

Stem CellsInternational



Enzyme Research


International Journal of

Microbiology

Documents

of Descen$ of Insectsdownloads.hindawi.com/journals/psyche/1938/018939.pdfinsects, since they are much more specialized than the Machilide. In fact, the Dicellura could readily be