PTERALIA OF THE PALEOZOIC INSECT ORDERS PALAEODICTYOPTERA , MEGASECOPTERA AND DIAPHANOPTERODEA ( PALEOPTERA ) * BY

For an understanding of insect evolution the structure of the wing base is of major significance. However, the fossil record o pteralia in extinct orders is extremely scanty. This paper is concerned with the wing bases of certain Paleozoic Paleoptera, namely, Palaeodictyoptera, Megasecoptera and Diaphanopterodea rom the Upper Carboniferous (Namurian) of Czechoslovakia, the Upper Carboniferous (Stephanian) of France, and the Lower Permian of Czechoslovakia and Kansas. Independently of the Neoptera, the Diaphanopterodea acquired the ability to flex the wings backwards over the abdomen. In this respect, the order is of special interest, and an attempt is made here to compare the pteralia o the Diaphanopterodea with those of extant Ephemeroptera. Our present knowledge of the wing base in Paleozoic Paleoptera is restricted to the axillary plate o several palaeodictyopteran adults (Kukalova, 96o, 969-7o), and palaeodictyopteran nymphs (Wootton, 972; Sharov, 97I) [See figures , 2 and 4]. Recently, the wing base has been described in the Diaphanopterodea, Family E1moidae (Kukalova-Peck, 974) (/]g. 8). In the present paper, the axillary plates in Martynoviidae and Asthenohymenidae o the Order Diaphanopterodea are included, and for the first time the axillary sclerites in Megasecoptera are described. The interpretation and terminology of the pteralia in extinct Paleoptera are necessarily dependent upon the detailed functional morphology of extant Ephemeroptera and Odonata. At the same time, the wing base structures found in extinct orders provide an evolutionary view and might be helpful in unraveling the enigmatic archi-

the wing bases of certain Paleozoic Paleoptera, namely, Palaeodictyoptera, Megasecoptera and Diaphanopterodea rom the Upper Carboniferous (Namurian) of Czechoslovakia, the Upper Carboniferous (Stephanian) of France, and the Lower Permian of Czechoslovakia and Kansas. Independently of the Neoptera, the Diaphanopterodea acquired the ability to flex the wings backwards over the abdomen.
In this respect, the order is of special interest, and an attempt is made here to compare the pteralia o the Diaphanopterodea with those of extant Ephemeroptera.
Our present knowledge of the wing base in Paleozoic Paleoptera is restricted to the axillary plate o several palaeodictyopteran adults (Kukalova,96o,, and palaeodictyopteran nymphs (Wootton,972;Sharov,97I) [See figures , 2 and 4]. Recently, the wing base has been described in the Diaphanopterodea, Family E1moidae 974) (/]g. 8). In the present paper, the axillary plates in Martynoviidae and Asthenohymenidae o the Order Diaphanopterodea are included, and for the first time the axillary sclerites in Megasecoptera are described.
The interpretation and terminology of the pteralia in extinct Paleoptera are necessarily dependent upon the detailed functional morphology of extant Ephemeroptera and Odonata. At the same time, the wing base structures found in extinct orders provide an evolutionary view and might be helpful in unraveling the enigmatic archi- 1974] Kukalova-Peck Pteralia 417 S C R P S R} MPI M C U p I/,"k-,: CU dictyoptera, Homoiopteridae). Enlarged base of fore wing. Redrawn from Kukalova, 1960. AP1--anal (966), but interpretation of these structures is stilI subject to debate. The axillary plate is considered either as the fused radio-anal plate, or as a compound structure, with the radio-anaI plate, axillary sclerites, and subalare incorporated. However, the wings of the Odonata are highly specialized and possess many unique 'features and therefore are less suited for comparison with Paleozoic orders than the more primitive wings of the Ephemeroptera.

PALAEODICTYOPTERA AND EPHEMEROPTERA
In the most primitive, yet unknown, Paleoptera all main veins probably originated at the wing base rom the separate basal plates.
In the process of evolution, the basal plates gradually became fused.
The intermediate stage of partial usion is documented in the primitive palaeodictyopteron, Boltopruvostia nigra (Namurian C of Czechoslovakia, fig. ). In this species the subcostal and radial plates are fully grown together (forming the subcosto-radial plate), the median plate is separate, and the cubital plate is fused with the small anal plate (forming the cubito-anal plate). The partially separated basal plates are also indicated in the wing pad of the nymph Idoptilus onisciformis (Wootton 972, Westphanian of England) (i]g. 2).
Within the evolutionary history of Palaeodictyoptera, full fusion of the plates was acquired in the more specialized forms. Basal plates used together into a single subcosto-anal plate but still retaining an indication of individual outlines, have been found in Moravia convergens (Lower Permian of Czechoslovakia, ig. 3). In Palaeodictyoptera rom the Upper Carboniferous o France (tqg. 4), the subcosto-anal basal plate is fully used and its subcostal part becomes reduced in size.
In the primitive Ephemeroptera (Siphlonuridae, (ig. 6), the costal brace in the ore wing starts at the anteroproximal angle of the axillary plate; in the hind wing, it sometimes starts rom a transverse elevation on the axillary plate, forming a convex hook directed against the stem of M. The costal brace is, in all probability, homologous to the postcostal vein of Palaeodictyoptera (ig. 4B, pv), plus the cross vein between Sc and R. The base o ephemeropteran wings is weakly sclerotized and the stems of the main veins adjacent to the axillary plate are mostly not discernible. However, i the vein stems 1974] Kukalova-Peck Pteralia 419 in primitive mayflies are thoroughly examined, their connection with the axillary plate can be traced. This is true in many Siphlonuridae, especially in the hind wings ( fig. 6B). In the ore wing o 8i#hlonurus mirus ( fig. 6A), the axillary plate is domed in the middle. At the anteroproximal angle, there is a short groove, which is also present in Palaeodictyoptera (tg. I, 3). Expanded stems o Sc and R (which might in this Sithlonurus be represented by the used R + M + Cu stems) meet at a suture and clearly penetrate the axillary plate. In the hind wing of the same species, the groove at the anteroproximal angle is reduced to a small pit. R is located at the anterior part o the dome and is closely ollowed by the short stem ot M, rom which MA and MP diverge almost immediately.  4).
The main difference between the axillary region o1: the Palaeodictyoptera and that o1: the Ephemeroptera is in the attachment o1: the anal veins. In the Palaeodictyoptera, the anal veins probably always originated directly 1:rom the subcosto-anal plate ((igs. 3, 4), or the anal plate was closely attached to the cubital plate (tqg. ). The anal area in Palaeodictyoptera is o1:ten crossed by diverse convex ridges, which are sometimes V-shaped (Homaloneura, Kukalova 969), and by transverse grooves (figs. 3, 4A), or the anal veins start 1:rom a cuticular thickening (Dunbaria, Kukalova-Peck 97). _As pointed out by Wootton (974, personal communication), these structures are in the place o1: the anal brace o1: Ephemeroptera. In the primitive mayflies, the anal veins start at the anal brace, which is basally attached to a small plate, separated 1:rom the large axillary plate. It seems possible that while the axillary plate originated by (usion o1: the subcosto-cubital plates, the anal plate stayed detached. This condition is partially reminiscent o1: Diaphanopterodea and Megasecoptera, as will be shown below. The anal brace in Ephemeroptera is 1:ollowed, both anteriorly and posteriorly, by two concave t:urrows. By location and, very probably, by (unction for flight, the 1:urrows may be compared with the concave transverse 1:urrow, crossing the anal area in almost all Palaeodictyoptera ((igs. 3, 4A).
The weak sclerotization o1: the axillary plate in Ephemeroptera and the vanishing o1: the adjacent stems are probably the changes which p,aC  Kukalova, 1970. pa--precostal

1974]
Kukalova-Peck Pteralia 42 occurred later in evolution. This conclusion is supported by the morphology of primitive Permian mayfly families, the Protereismatidae and the Misthodotidae. In both, the stems of main veins were strong, reaching the wing base, and the anal brace was less pronounced than in extant mayflies ( fig. 7). It should be noted that the adult life span in Permian Ephemeroptera was probably longer than in Recent species; the presence of large and apparently fully unctional mouthparts in the adult Permian Misthodotidae (Tschernova, 965) is very suggestive that these particular mayflies, at least, were active feeders. Flight was presumably important in obtaining food, as well as in courtship activities.
According to Brodskyi (97o), the second axillary sclerite in Recent Ephemeroptera is firmly fused with the axillary plate, although this opinion is not shared by some other specialists; the schematic figure of the axillary region in extant mayflies (Leptophlebiidae) in the interpretation of Tsui and Peters (972)

1974]
Kukalova-Peck Pteralia 423 1963). In comparison with the average size o Palaeodictyoptera, they are much smaller and they developed the ability to flex the wings backwards over the abdomen. The axillary plate o the wing base and the basal old were recently described by Kukalova-Peck (1974) in Permodiapha carpenteri of the family Elmoidae ( fig. 8A) 8B).
In both forms, the general patterns of the wing base structure are similar. The larger part of the axillary plate in Martynovia is formed by the fused medio-cubital plate, or perhaps by the cubital plate itself. Different is the development of the anal plate, which is distinctly outlined and two-lobed, located anteriorly to the basal fold. The anal plate is firmly connected by the anterior margin with the axillary plate. The basal fold encircles the axillary plate anteriorly and distally and separates it from the sclerotized region, extending between CuP and the posterior wing margin ( fig. 8B, S) with the proximal margin of the axillary plate and with R. It is in the position of 2 Ax in Recent Ephemeroptera (sensu Tsui and Peters 972, fig. 5).
In Asthenohymenidae, the incompletely preserved wing base, including only the basal fold and part 05 the sclerotized region, is known. The structure of the wing base, as preserved in Asthenohymen dunbari (specimen No. 3835, Museum of Comparative Zoology, fig. 9) seems to be basically similar to that of Martynovia and Permodiapha.
The fused axillary plate of Diaphanopterodea is homologous only to the posterior part of the subcosto-anal plate of Palaeodictyoptera.
The subcostal, radial, and possibly also the median basal plates became completely reduced, perhaps in connection with narrowing of the wing base, coalescence of veins and with acquisition of wing folding ability. major difference in the pteralia, not only with respect to Palaeodictyoptera, but also to all other paleopterous orders.
The well preserved wing bases with axillary sclerites were found in specimens of two comparatively small species: Protohymen permi- Towards the base, the wing narrows markedly in the posterior part, while the anterior margin remains almost straight. Shortly before the base, the posterior margin curwes posteriorly to form a prolonged axillary lobe parallel with the tergal margin. The serrated costa is attached at the very base to a long narrow plate on the |at.eral margin of the tergum. This plate is connected anteriorly with a prominent, darkly colored lobe. The subcosta is convexly bent near the base. Both costa and subcosta terminate on a triangular, highly convex basal plate, which is posteriorly connected to R + VI and proximally to the small, convex axillary sclerite. This is associated with the tergal margin, the basal plate, the costa, the narrow plate on the tergum, and the anterior arm of the large elongated sclerite.
The cubitus is concavely bent towards the very end of R shortly before the base, so that it creates the mirror image of the subcosta. The anal vein originates from a small anal plate, which is divided by a perpendicular suture into two small equal sections. The anal plate is directed obliquely and posteriorly against the posterior arm of a large, anteropo.steriorly elongated sclerite. This closely contacts the tergum along almost its entire proximal margin. The elongated sclerite is highly convex posteriorly from the level of the end of the anal plate. In the anterior, flatter part, the axillary sclerite divides into two arms. The proximal arm is longer and slightly bent to attach to R + M and to the small axillary sclerite; the distal alan is shorter and directed to connect the curved end of the anal plate. This sclerite is markedly larger in the hind wing.
The megasecopteran pteralia, as described above (figs. o, ), are so highly specialized that homology at our present level of knowledge is nearly impossible. The narrow plate laterally on tergum (np) and the elongate sclerite (Axe) as well seem to be compound structures in which several elements might be incorporated. The small convex sclerite (Axc) is approximately in the position of Ax of Ephemeroptera (se'nsu Tsui and  Kukalova-Peck Pteralia 429 SUMMARY In the several paleopterous orders, the "axillary plates" are by no means to be considered as fully homologous structures. They are formed by fusion of various basal plates, which can eventually vary in the degree of fusion and in the individual size within a single order (for instance in Palaeodictyoptera, and in Diaphanopterodea).
There is little doubt that in the hypothetical primitive ancestor of Paleoptera the main veins originated from the separate basal plates.
From this condition, the axillary plate was derived by fusion of the following plates: in Palaeodictyoptera, of subcosto-anal plates; in Ephemeroptera, of subcosto-cubital plates; in Odonata, of radioanal plates; in Diaphanopterodea, of medio-cubital plates; in Megasecoptera, only small subcostal plate remained.
In spite of the variability mentioned above, the general plan of the axillary region and adjacent part of the wings is shared by all paleopterous orders, and certain structures, functionally significant for the flight, repea: many times throughout Paleoptera. For instance, separated or semi-separated anal plates occur in Magasecoptera, Diaphanopterodea, and Ephemeroptera. A deeply concave furrow crossing transversely the anal area is found in almost all Palaeodictyoptera and in the Ephemeroptera. Transverse reinforcement of the proximal half of the wing formed by coalesced veins and supporting cross veins, cuticular thickenings and pigmented stripes is present in Odonata and in "odonatoid" (dragonfly-convergent) Palaeodictyoptera, etc. These structures will not be understood until more study of the functional morphology has been done, particularly on extant mayflies, dragonflies, and on wings in general.