MEGASECOPTERA FROM THE LOWER PERMIAN OF MORAVIA

The M,egaseco.ptera are separated rom the closely related Paleozoic Order Palaeodictyoptera by having nearly homonomous, elongate, petiolate (or basally narrowed) wings; usually a simple venation and Sc and R crowded towards the costal margin. The body structures in both orders are virtually identical, so a.r as known. The. r.elationship o] .these groups gives the impression o.t being like that between dragonflies and damselies within the Order Odonata, rather than that between two dierent orders (Kukalova-Peck, 973). Only recently has the wing attachment to the body in both groups been studied (Kukalova-Peck, 974), revealing evidence which justites the recogniti.on ot Megasecoptera. as a distinct order. The Palaeodictyoptera have been tound to retain many eatures o,t the primitive paleopterous wing base, such as the ully developed subcostoanal axillary plate. In the Megasecoptera, however, the subcostoanal axillary plate is missing and the axillary sclerites are highly specialized and largely used. The morphology o the axillary region in Megasecopter is unique within the Paleoptera and shows once again the potential or extraordinary diversity in the paleopterous orders. The Lower Permian locality o. Obora in Moravia (Czechoslovakia) has recently yielded twelve specimens o.t Megasecoptera, reterred here to six new tamilies: Caulopteridae, Hanidae, Arcioneuri-

The Upper Carboniferous (Westphalian) families Frankenholzii- dae (Germany), VIecynopteridae (Belgium) and the new Camily Dictyoneurellidae (France) (described below), all of which were previously considered palaeodictyopterous, are apparently related to. the newly described families mentioned above and to the primitive 'family Anchineuridae (Upper Carboniferous, Spain) of the Order VIegasecoptera.They are herein referred to that order.
Besides the adults, three nymphal forms are described: Caulotera colorata, d rcioneura juveniles, and Alectoneura eurolaea.The nymphal wings are easily distinguished from the adult wings, which are held at right angles to the body, by the characteristic bend in the proximal third of the wing length.This bend determined the oblique-lateral position of the wing pads in the living nymphs, as described by Carpenter and Richardson (968) in the completely preserved early stage of the megasecopteran nymph o Mischoltera douglassi.Ho.wever, in all isolated nymphs from Obora the angle of the bend is more obtuse than in Mischotera douglassi.This fact together with the larger size and good sclerotization of the venation (the sclerotization being weak in douglassi) implies that the nymphal wings from Obora belonged to late, possibly the ultimate, instars.There is no evidence as to whether or not a subimaginal stage was present in Megasecoptera.In the Palaeodictyoptera, parts of a shed cuticle have been found attached to a emale of Lycocercus golden- bergi (Kukalova, 969, p. 449, fig.32), including a smaller and less sclerotized ovipositor than that of the adult.It is therefore conceiv- able that the related Megasecoptera, which clearly descended rom the Palaeodictyoptera, had adult molts.Nymphal wing: petiolate, broadened beyond mid-wing, narrowed in the apical third; posterior margin undulated.Sc long; stems o.f R and M not coalesced in the basal third; Rs diverging from R1 before mid-wing; MA and MP branched; MP anastomosed with CuA for a short distance.;CuA simple, CuP sending off several branches; A with a series of branches.Cross veins numerous, irregular, forming in'tercalated sectors.
This family differs from the Anchineuridae in the petiolation of the wing, the branched MA and the coalesced M P and CuA.
Although the shape of the wing in the family Caulopteridae and in the following three families is specialized and typical for Megasecoptera, the venation is reminiscent of that in the Pala.eodictyoptera.If the proximal, narrow part of the wing has not been preserved, the megaseco.pterouscharacter of such a fragment would be obscure.This seems to be true for several incompletely preserved wings pre- viously considered palaeodictyopterous, discussed below.
At present this family includes only the following genus: Cauloptera, new genus Nymphal wing: moderately slender, with a broad lobe in the third quarter of the wing length; posterior margin undulated beyond mid- wing only within the mp area.Sc, R1 and Rs equally distant from each other, close to., but not crowded towards the costal margin; Rs with 2 long branches; MA with several short branches.
The generic name is derived from the Greek words kaulos and ltera (petiol.e + wings) and is considered neuter plural.

Figure
This species is based on a detached, well-preserved wing pad of a late nymph; the base is narrowed, perhaps petiolate; the basal third is missing.
Nymphal wing: length (estimated) about 4 ram, width 9 ram.Broadened part of the wing forming a smooth, broad lobe; undula- tions in the mp area shallow; apex comparatively broad and rounded.

Psyche
[March Sc terminating on R1; Rs originating before mid-wing; MA sending off 3 branches; MP with 2 branches; CuA simple, CuP with 2 branches; 1A with a series of about three branches; intercalated sectors mostly short.Wings dark-colored, with irregular light bands becoming crescent in the apical area.
Family Hanidae, new amily Wing: petiolate, broadened at about mid-wing, narrowed in cua area; posterior margin deeply undulated.Venation simplified; Sc terminating shortly beyond mid-wing; stems of R and M not co- alesced in the basal third; Rs diverging from R1 about mid-wing; MA and CuA simple, MP and CuP with few branches; MP anastomosed with CuA for a short distance; IA simple.Archedictyon and irregular cross veins present.Intercalated sectors absent.
The family Hanidae differs from the related family Caulopteridae in the more deeply undulated posterior margin, in the short Sc, the simplified venation, the presence of the archedictyon and in the ab- sence of the intercalated sectors.
The large, elongate, extremely narrow wings of Hanidae, with an intricate outline, are among the most unusual wings within the Paleopt.era.The preservation of a primitive archedictyon, which in other families becomes reduced to cross veins by Westphalian time, is a remarkable example of mosaic evoluation in Paleozoic Paleoptera.

Hana, new genus
Wing: very long and narrow; posterior margin forming several prominent lobes in the median and cubital areas.Sc distant from anterior margin in the basal half of the wing; Rs with about 3 short branches; IA following closely the posterior margin.Wing mem- brane with a dense covering of tubercles.This genus is represented by two species: H. filia, ba.sed on the distal half of a wing, and H. lineata, n.sp., based on the proximal half of a wing.The genus is named for my daughter Hana, who has been very helpful to me in the field work.

Figures 3-6a
This species is based upon beautifully preserved distal halves of the fore and hind wings, in normal positions (see reconstruction of the outlines, fig.6a).The hind wing is somewhat longer and broader at the end of the anal area than the ore wing.The venation is almost identical in the wings, though the first branch of Rs originates more markedly distally in the hind wing.In both wings, the costal margin is proximally broken away; in the fore wing, the posterior part of the wing is superimposed over R.Both wings are restored to the original widths in figure 3- Besides the holotype, a small fore wing fragment (specimen no. 26/1974, fig. 4),and a hind wing fragment (specimen no 27/1974, ig. 5) were found, showing a coloration and a lobation o.f the pos- terior margin remarkably similar to that of the holotype.
Wing: length about 96 mm, width 8.3 ram.Posterior margin convexly bent at the end of all branches of CuP; Rs originating beyond mid-wing; first branch of MP long, starting before mid- wing; CuP with a series o 3 short branches; archedictyon and cross veins similar to those o.f filia, but finer.Wing membrane densely covered with small tubercles.A few dark spots occur in the sub- costal area.
This species differs from filia in its narrower wings, finer mem- brane and its series o.f 3 .CuP branches.
Holotype: no.28/I974 (wing fragment, length 58 mm, width 8.3 ram; reverse).Paleontological Institute of Charles University, Prague.Family Arcioneuridae, new family Nymphal wing: petiolate, narrow, broadened at about mid-wing; posterior margin undulated.No coalescing veins; Sc short; Rs di- verging from R1 we'll before mid-wing; MA and MP with several branches; CuA and CuP with few short branches.Cross veins numerous, forming intercalated sectors.This t:amily differs from the Caulopteridae in the shorter Sc and in lacking the coalesced MP and CuA.Type genus: lrcioneura, new genus.
Arcioneura, new genus 'Nymphal wings: very long and narrow.R1 close to the costal margin in the distal half of the wing; Rs with about 2 long branches; MA with a series of short branches; MP orming a long branched ork.Cross veins almost regular.
The generic name is derived from the Greek words arkios and neura (certain + veins) and is considered neuter plural.
Arcioneura juveniles, n.sp. Figure 8 This species is based upon a detached wing of a nymph, bent typically into oblique-lateral position in the basal third.The slight angle of the bend and the advanced sclerotization of veins and cross veins indicate that the nymph was .close to the adult stage.
Nymphal wing" length about 9 ram, width 2.6 ram.Sc termi- nating shortly beyond mid-wing; Sc much closer to.R and R1 than to the costal margin; Rs detaching from R1 at the end o.f the basal third, the /]rst branch of Rs starting shortly beyond mid-wing; MA sending off 4 branches, MP probably 3 times branched; CuP with branch; A probably simple.Type species" z/nconeura havlatai, n.sp.
Figure 9 This species is monotypic, based upon an adult wing with a well preserved, long, petiolate base.. Distally from mid-wing, the wing is 'broken into several dislocated pieces.In figure 8, the venation is shown restored to its normal pattern.
Wing: length about 41 mm, .width5-7 mm.Wing broadest shortly beyond mid-wing; posterior margin with 4 undulations; at its proximal end the petiole broadens for its attachment to the thorax.
Rs originating at about the end of the basal third, the first branch of Rs starting before mid-wing; MA sending o.ff 3 branches; 3alP probably branched 4 times; CuA probably branched 5 times; 'CuP with branch.A close and parallel with the posterior margin, with one branch.
The species is named for Karel Havlata of Prague, who has col- lected hundreds of specimens of insects in the Obora beds.
Family Anco.pteridae, new family Wings" elongate; almo.stequally broad beyond mid-wing; posterior margin undulated.Sc somewhat shortened; stems o.f R and M not coalesced in the basal third; Rs diverging from R1 much before mid- wing; MA and CuA simple, MP and CuP branched.Cross veins numerous, forming conspicuous intercalated sectors.
The family Ancopteridae differs from the related family Dictyo- neurellidae (Westphalian, France), and Anchineuridae (U..Carboniferous, Spain) in the presence of the undulated posterior margin,

Psyche
[March in the longer Rs and in the branching of the main veins, which gen- erally arise much more basally.From the Arcioneuridae it differs.
in the longer Sc, in the simple MA and CuA; rom the Caulopt.eri-dae in the simple MA and the absence of coalescence between MP and CuA.Type genus: Ancoptera, new genus.
Ancoptera, new genus Wings: long but only slightly narrow, with broadly rounded apex; posterior margins with only shallow undulations.C broad, flattened; Sc and R1 crowded towards the costal margin in the distal third, but well remote rom it proximally; Rs with 2 long branches; stem of R, stem of V[ and CuA touching each other in the proximal third of the wing; mp area large.
The generic name is derived from the Greek words ankos and ptera (curve + wings) and is .consideredneuter plural.
Figure 2 This species is based upon a well preserved adult wing, with the narrowed basal part missing.The long anterior branch of R1 is probably only an individual variation.Besides the holotype, another specimen (no. 3 r/1974, fig.Io) was found, this being the distal part of a wing.In this wing fragment, the apical part is more pointed and the Rs branches are much shorter ,than in the holotype.The veins and cross veins are surrounded by dark-colored, narrow bands, as in Dictyoneurella.Specimen no.3I/I974 probably belongs t.o a separate species, but because of the small size of the fragment it is referred here to ACncoptera sp.
Wing: length about 43 mm, width, 9.8 ram.Posterior margin with 4 undulations in the distal part of the wing, Rs diverging from R1 at about the first fifth o.f the wing length; first branch of Rs originating much before mid-wing; MP with a series of 3 branches; CuP forming a wide fork, mp-cua area and cup-anal area are.wide.Intercalated sectors long and regular.
Family Engisopteridae, new family Wings: narrow, but not petiolate, almost uniformly broad in distal halt:; posterior margin not undulated.Postcostal vein present; Sc terminating shortly before mid-wing, subcostal area relatively broad; Rs short; ma area large and rich in long branches, mp area narrow; CuA simple, CuP with long branches.Cross veins numerous, inter- calated sectors present.
This family differs rom the Arcioneuridae in the broader wings, the presence of the postcostal vein, the shorter Rs, and larger ma and cup areas.From the primitive family Archineuridae it differs in its relatively broad subcostal area, short Sc and Rs and in the more nearly parallel arrangement .ofmedian and cubital branches.
The family Engisopteridae has several characteristics o venation which are among the most primitive known in the Megasecopterathe relatively broad subcostal area with a series of Sc branches, the postcostal vein, and the generalized, richly branched venation (except t:or a simple CuA).These particular features are palaeodictyopterous in character, and they are more primitive than their structure in Anchineuridae.
The postcostal vein, which I consider homologous with the "costal brace" of Ephemeroptera, is present in almost all Palaeodictyoptera.In addition to the Engisopteridae, the vein has been ound in the Mischopteridae (Carpenter & Richardson I968,fig., 2), in both adult and nymphal wings.It is probable that this inconspicuous vein has been overlooked in several other families.The postcostaI veins are completely reduced in families with a convexly bent Sc at the very base, as in Protohymenidae or Alectoneuridae.

Engisoptera, new genus
Wings: more or less elongate-oval in shape., narrowed across the anal area.C, Sc and RI equidistant from each other in the distal half of the wing; Sc terminating on C; Rs very remot.e from R1, sending off 2 short branches; MA, MP, and CuP branches o.rigi- nating very anteriorly.Cross veins almost regular.
The most conspicuous character of this new genus is the branches of the main veins originating very anterio.rly in the wing.A similar branching pattern is present in the hind wings of some Pa.laeodicty- optera.
The generic name is derived from the Greek words engys and ptera (near + wings) and is considered neuter plural.
Type species" Engisoptera simplices, n.sp.   Figure 11 This species is based upon .adetached, very delicate wing with the anal area and the apical area missing.
Wing: length about 20 mm, width 4 mm.Rs div,erging ro.m R1 shortly beyond mid-wing, rs area small, triangular; MA with 2 long branches, the distal one twice branched; MP with a single, long parallel branch; CuP with a series o.f several relatively long branches.
Family Aleetoneuridae, new family Wings: very narrow in the proximal third but not petiolate, broad- est beyond mid-wing.Venation simplified; Sc terminating before the apical third; stems of R and M not coalesced in the basal third; Rs diverging from R1 before mid-wing; MA adjoining or .co.alescing with Rs for a short distance.IA with a series of long branches.

Cross veins few.
The venation of Alectoneuridae is very close to.that of Mischop- teridae.However, the shape of the wing is less specialized in lacking the petiole, and the anal area has a more generalized venation.The basal part of the wing, including the convex basal bend of the subcosta., is very reminiscent o.f the Protohymenidae I974).The Alectoneuridae combines morphological features of two.groups of families, which originated from common ancestors" Mischopteri- dae, Sphecopteridae, Carbonopteridae, Ischnoptilidae, Corydaloididae, Foririidae on one side; and Aspidothoracidae, Scytohymenidae, Aspi- dohymenidae, Vorkutiidae, Bardohymenidae, and Pro.tohymenidae on the other side.The amily Alectoneuridae probably descended from the ancestors of both groups and has a morphologically intermediate position similar to that of the Moravohymenidae (Kukalova-Peck I972, p. 256).
Aleetoneura, new genus Wing: C and Sc convexly bent before the base; subcostal area relatively broad in the proximal half of the wing, with a series of Sc branches; M dividing into MA and MP at the level of the origin of Rs; anal area relatively broad and long, reaching beyond mid-wing.
In its venation, zllectoneura recalls Psilothorax (Mischopteridae) and in the structure of the basal part of the.wing, Protohymen (Protohymenidae).The broad subcostal area, with a series of numerous subcostal branches and the relatively large anal area with richly branched IA, are the characters typical for primitive.Megasecoptera with generalized venation.
The generic name is derived from the.Greek words alecto and neura (indescribable + veins) and is considered neuter plural.
Type species: d lectone.uraeuropaea, n.sp.This species is based upon a single wing, which is slightly bent at the end of its basal quarter.The wing might have belonged to a subimago of a species like Sphecoptera elegantissima, since all other wings in all related families are straight in the basal part.In this case, the deep concavity of the posterior margin at the 3rd branch of IA would probably become less pronounced in the adult wing.In the holotype, the very fine wing membrane has split along the veins and become slightly superimposed posteriorly over other parts ot: the membrane.In figure 12, the venation is restored back into normal position.
Wing: length of the fragment 12.8 mm, width 3.3 mm.Anterior margin broadly bent in the basal quarter; posterior margin slightly undulated at about the middle of the anal area; Sc close to R along its whole course; Rs and MA adjoining for a short distance; first branching of Rs broadly dichotomous; M and Cu branches almost parallel and probably simple; CuA soon after its origin connected with M by a short cross vein; IA with 4 long branches; 2A short, simple.Cross veins near the posterior margin sigmoidal, almost parallel with the longitudinal axis.

THE PROBLEM OF RECOGNIZING GENERALIZED MEGASECOPTERA
Before this new material from Y[o.ravia was made known, mega- secopteran wings were readily separated 'from those ot the Palaeo- dictyoptera by the crowding of Sc and R towards .thecostal margin, the basally narrowed wings, and the simplified venation (Carpenter, 1963).However, with the presence o.t a rather broad subcostal area in Ancopteridae, Alectoneuridae, Engisopteridae and Hanidae, and the richly branched venation in Caulopteridae, Arcioneuridae and Engisopteridae, the recognition of primitive Megasecoptera becomes more difficult.This is true especially for ragmentary wings, in which the proximal, narrow part may not be preserved.
For identifying Megasecoptera with a generalized venation, the most reliable crit.eria are as follows: long, nearl.yhomonomous wings; a narrow anal area; at least a partial crowding o.t Sc and R owards the costal margin; and the wing conspicuously broad at about mid- wing but narrowed both proximally and distally.An irregular and not very dense cross venation, coarse archedictyon, and incomplete intercalated sectors are often present.However, each single feature mentioned above occurs randomly in the Palaeodictyoptera and should be carefully considered if found separately.
Keeping in mind the similar eatures of primitive megasecopterous and palaeodictyopterous venation, I believe that the following species, based upon fragrnents and previously considered to belong to. the Palaeodictyoptera, need to be reconsidered or possible inclusion within the Megasecoptera.
Dictyoneurella #er[ecta Laurentiaux, 1949  (Westphalian, northern   France) is probably megasecopterous because of the shape o the wing, which is broadest at the middle, and narowed both proximally and distally; and because o the general resemblance to.&nchineuridae and Anconopteridae.This species is herein re(erred to a new family, Dietyoneurellidae, characterized by having a simple lobe on the posterior wing margin, and a primitive venation that lacks coalescence of veins, a relatively broad subcostal area, simple MA and CuA, richly branched MP and CuP, and irregular cross veins forming intercalary sectors. The amily Frankenholziidae, based on Frankenholzia culmani Guth6rl, 1962 (Lower W'estphalian D, Germany) is in all prob- ability megasecopterous, since the wing is broadest at mid-wing and narrows di.stally and proximally.It differs from the Arcioneuridae by having broader wings, a broad lobe on the posterior margin and by less abundant cross veins [The vein marked by Guth6rl, 1962,   fig.3, as Sc is in reality R1, R is Rs, and A is CUP].
The i:amily Mecynopteridae, based on Mecyno#tera sldendida Handlirsch, I9O4 (W'estphalian C, Belgium), is almost certainly megasecopt.erous, because Sc and R1 are crowded towards the costal margin and the wing is elongate.It is related to the Upper Carboni- ferous Anchineuridae and Dictyoneurellidae and to. the Lower Per- mian Ancopteridae, all char.act.erized by the presence of irregular cross veins and intercalated sectors.Mecynotera tuberculata Bolton, 9:, is a wing fragment too small to .bereferred to.any particular genus; and 3Jerle'nbachia grimaldi Waterlot, 1934, classified by its author in the M.ecynopteridae, is a wing fragment of dubious sys- tematic position, and may even be neopterous.Saarodictyum abnormis Cmth6rl, 1939 (Westphalian D, Germany)   is very probably megasecopterous, because of the typical narrowing of the wing towards the base, almost into a petiole.Very likely it represents a separate tamily with primitive venation and a dense archedictyon, and rather broad subcostal area.Unfortunately, the Psyche [March venation o the holotype is misinterpreted in the original figure ('Guth6rl, 939, fig.), and the type specimen, which was depo.sited in a private collection, is not accessible for study.Hence, this species is referred to the Megasecopte.ra,family uncertain.
Orthocosta lithomantidoides Waterlot, I934 (Westphalian C, Belgium) is based upon a small wing fragment.Because o the shape ,of the posterior margin and the irregular cross venation, with incompletely developed intercalated sectors.,I believe it belongs to the Megasecopte.ra,though of uncertain f,amily position.Orthocosta splendens Bolton, 912, on the other hand has a loose "palaeodicty- opterous" type of reticulation and is referred here to the Palaeodicty- optera, uncertain family position.
Since it seems clear that the most primitive Megasecoptera merge fully in the wing venation pattern with the Pala.eo.dictyoptera, we must accept the fact that some little-known species cannot be identi- fied to either .orderwith certainty.
The following is a summary of my present concept of the relation- ships of the species mentioned above: Order Mega'secoptera Family Dictyoneurellidae Kukalo,va-Peck (herein described) Dictyoneurella erfecta Laurentiaux, I949

Figure 12
Figure 12 ram, width lO.5 ram; length of the hind wing about 88 ram, width 12 ram.Posterior margin deeply incised at the end of all branches of 'CuP and at MP2; Rs origina'ting before mid-wing, Rs branches originating in the distal quarter of the wing; MA and CuA simple; MP with 2 branches, CuP probably with short branch; archedictyon and cross veins strong, conspicuous.Wing membrane covered with numerous prominent tubercles and with extensive, irregularly outlined dark areas; hind 'wings darker than fore wings.
Wings: very long and narrow; posterior margin undulated beyond mid-wing.R1 close to the costal margin; Rs with about 4 long branches, sometimes forked; MA with a series of short branches; MP with a long, branched fork; CuA forked.Cross veins irregular.ztncone'uradiffersfrom Arcioneura by its larger wings, which have more numerous branches on Rs and CuA.The generic name is derived from the Greek words ankos and neura (curve + veins)and is considered neuter plural.