The Geroptera comprise a single family of primitive odonatopteroids, Eugeropteridae, from the early Late Carboniferous (Early Bashkirian = Namurian) of Argentina (Riek and Kukalova-Peck, 1984). While a quick glance at the primitive wings of geropterans reveals little affinity to anything one might identify as a dragonfly or damselfly, finer study indicates a shared, albeit distant, ancestry between the groups

(as noted earlier). Superficially, the order more closely resembles the Palaeodictyopterida and, like some palaeodicty-opterids, geropterans had pronotal lobes but lacked an archedictyon (Figure 6.24). Very little is known of Geroptera and the monophyly of the few species in the order is entirely speculative. Within the Odonatoptera the pronotal lobes are derived, and so these may serve as a defining feature of the Geroptera, though they are convergent with similar structures in unrelated Paleozoic orders and apparently occurred in Erasipteridae (a primitive protodonate family that might be best included in Geroptera). Alternatively, paranotal lobes may be the groundplan design of all pterygotes, retained in basal members of various orders but lost independently in most species in each lineage.

It is entirely unknown whether or not the Geroptera, like Odonata, had aquatic nymphs. The ancestor of Odonata certainly had an aquatic nymph, and some protodonatans may have had as well, but it is not known whether this mode of life evolved in the common ancestor of Protodonata + Odonata, or was even a feature of all Odonatoptera. It would further be very significant to know if Geroptera had terrestrial immatures because it would cast light on whether aquatic living in Odonatoptera and Ephemeroptera were independently derived, as current evidence suggests. Other basal metaptery-gotans, the Palaeodictyopterida, were certainly terrestrial and tentatively suggest that Odonatoptera evolved aquatic nymphs independently of Ephemeroptera.


The Holodonata includes two orders that better approximate what most people know as dragonflies and damselflies. The wings had a characteristically long, slender, "odonatoid" appearance. In fact, most of the defining features of this

6.24. Reconstruction of Eugeropteron (Eugeropteridae). Although superficially resembling a palaeodicty-opteran, eugeropterids were early odonatopterans and perhaps stem-group relatives of odonates.

6.25. Wing of Megatypus schucherti (Meganeuridae: Protodonata) from the Early Permian of Elmo, Kansas, shown here at life size. The mega-neurid "griffenflies" were early relatives of modern odonates and included the largest known insect, which was Meganeuropsis permiana with a wingspan of approximately 640 mm (26 in.). YPM 1021; length 160 mm.

6.25. Wing of Megatypus schucherti (Meganeuridae: Protodonata) from the Early Permian of Elmo, Kansas, shown here at life size. The mega-neurid "griffenflies" were early relatives of modern odonates and included the largest known insect, which was Meganeuropsis permiana with a wingspan of approximately 640 mm (26 in.). YPM 1021; length 160 mm.

group occur in the structure of the wings. Unlike the Geroptera, holodonotans have a large, proximal, hornlike sclerite on the posterior articular plate of the wing articulation, and there is fusion of various veins. For example, MA is fused with RP while the stem of M at the wing base is vestigial (in Protodonata and some basal Odonata), or it is fused entirely with the Cu stem (in most Odonata). Vein MP originates from the combined stem of veins M + Cu, rather than as part of a stem of M as in Geroptera, and the area between veins MA and MP is expanded and filled with intercalary, longitudinal veins. An interesting feature of the Holodonata is that the thorax is slanted posteriad. The slanting is weak in the few Permian Protodonata that had some pleural sclerites preserved. In Odonata the slanting is much more dramatic, particularly in damselflies where the wings extend over the abdomen when held together at rest, similar to the neopter-ous condition but where the wings are actually folded flat over the abdomen.


Although frequently called giant dragonflies, the Protodonata cannot truly be considered dragonflies. This is a Paleozoic stem group to the true Odonata, the dragonflies and damselflies. The name "griffenflies" more aptly highlights this distant relationship, rather than the name "giant dragonflies," which implies a much closer affinity. As the cladogram shows, dragonflies (Epiprocta) are distantly removed from the Protodonata and radiated after most protodonatan lineages became extinct.

While "dragonflies" is a misnomer, "giant" is not. Among the Protodonata insects attained grandiose proportions (Figures 6.25, 6.26, 6.27). The largest insect to have ever lived was Meganeuropsis permiana, from the Early Permian of Elmo, Kansas, and Midco, Oklahoma (it is also known by the synonymic name M. americana) (Carpenter, 1939, 1947). This magnificent griffenfly attained wingspans of approximately 710 mm (28 inches), which dwarfs the largest odonates found today. Protodonatans were almost certainly predaceous, as all nymphal and adult odonates are today. Most fossils of

6.26. Wing of Tupus gracilis (Meganeuridae: Protodonata) from the Early Permian of Midco, Oklahoma. MCZ 4818; length 145 mm.
6.27. Wing of Arctotypus sinuatus (Meganeuridae: Protodonata) from the Permian of Russia. PIN 3353/87; length 120 mm.

these insects consist only of wings, but among the few preserved body parts are large, toothed mandibles, enormous compound eyes, and stout legs with spines, thrust forward in a similar manner to Odonata - all indicative of their being aerial predators. It is intriguing to imagine how these insects flew, perhaps streaking through Paleozoic swamps and forests, landing on unsuspecting animals like a bird of prey. At their prodigious size, they must have preyed on virtually all other insects and even small vertebrates.

Although immature griffenflies are yet unknown, the close relationship of the group to Odonata might suggest that their naiads were aquatic, but there is no direct evidence. It would be particularly fascinating to know if griffenfly naiads possessed the labial "mask" characteristic of odonate naiads (Figure 6.33), which is the prey capture device. Some of the larger naiads of Recent Odonata can even capture small vertebrates such as fish or tadpoles, so given the size of Mega-neuropsis (its naiad must have been up to 18 inches in length), it must have been a formidable predator. It would also have been impressive to see such a naiad eclose into an adult, its huge, soft gossamer wings gradually expanding before taking its first flight. Evidence for aquatic protodonate naiads, however, like those of Paleozoic Ephemeroptera, is equivocal (Wootton, 1988).

The famous French paleontologist Charles Brongniart (1893) brought us our first image of these giants and coined their scientific name. Brongniart's dissertation was a study of the insects from the Carboniferous coal measures of Com-mentry, France, and among the fossils he discovered was the first griffenfly, Meganura monyi, which was the largest insect known until the discovery of Meganeuropsis by Carpenter approximately 50 years later. He even published in his dissertation a life-sized, fold-out reconstruction of the insect, although at the time only the wings were known. Unfortunately, even today little is known of Protodonata, much remains educated speculation, and even some is pure myth, such as the existence of extinct dragonflies with 2 m (6 ft) wingspans. Most specimens are preserved as wing fragments only, a few as virtually complete wings, and even fewer with some body structures (Figure 6.28).

The wings of protodonatans, unlike those of the Odonata, lacked the distinctive pterostigma or a nodus (formed by the abrupt termination of Sc into a transverse nodal crossvein near wing midpoint), among other typical odonate features (Figure 6.29). Although frequently believed to be an unnatural (paraphyletic) group, the exclusion of Geroptera from Protodonata, as well as a few basal odonate suborders, makes the traditional families Meganeuridae and Paralogidae apparently monophyletic. In a recent study of odonatoid relationships (Rehn, 2003) Protodonata were redefined in a more restricted sense, based on a small lobe on the outside edge of the costal axalare (a portion of the anterior axillary plate at the wing base) and intercalary longitudinal veins between IR1 and RP2. The evidence is not substantial, so it is still possible that griffenflies are paraphyletic, stem-group odonatoids.

The Protodonata ruled the Paleozoic skies from the Late Carboniferous until the Late Permian, disappearing from the

6.28. The early odonatopteran, Erasipteroides valentini from the Late Carboniferous of Hagen-Vorhalle in Germany. Although odonates today do not have ovipositors, stem-group taxa such as Erasipteroides had well-developed, primitive ovipositors. Redrawn from Bechly et al. (2001).

6.29. A reconstruction of one of the large, extinct odonatopterans, Namurotypus sippeli, from the Carboniferous of Hagen-Vorhalle. Based on Bechly et al. (2001).

record after the End Permian Event (ca. 247 mya). While the Protodonata vanished, the Odonatoptera as a whole persisted through this catastrophe, including true Odonata from the Permian. The flourishing of Odonata in the Mesozoic may be a result of the demise of the protodonatans.


Griffenflies were not the only giants during the Paleozoic. Enormous mayflies, myriapods, scorpions, palaeodicty-opterids, and others were all contemporaries of these aerial juggernauts (Kraus, 1974; Hunicken, 1980; Briggs, 1985; Shear and Kukalova-Peck, 1990; Kraus and Braukmann, 2003). But, it is a very common misconception that all Paleozoic insects were giants when in fact most species were only a few centimeters or less in size, not unlike the situation today. Also, gigantism occurred in primitive amphibians during the Carboniferous (Carroll, 1988), and of course in some lineages of nonavian dinosaurs. The development of gigantism and its disappearance is an intriguing evolutionary and mechanistic question. The repeated evolution of unusually large size can be a feature of the lineage (e.g., sauropod dinosaurs), but in general there must also be some environmental factors conducive to gigantism, such as defense against predators (Vermeij, 1987; Shear and Kukalova-Peck, 1990). Another explanation concerns changes in the atmospheric concentration of gases (specifically oxygen) during the Late Paleozoic and Early Mesozoic (Graham et al., 1995), which has been discussed mostly in terms of insects because of their manner of respiration.

Insects breathe through a tubular system of tracheae, which are connected to the outside of the animal by minute, valved openings (spiracles). Air moves through the insect's tracheae and it is the passive diffusion of oxygen that allows the insect to respire. Insects can enhance the movement of airflow by contracting small "bellows" located at various points in the tracheal system or by expanding and contracting their abdomen, but overall it is the simple physics of diffusion that allows them to breathe. This action immediately imposes constraints on the body size of an insect because it becomes increasingly difficult to get oxygen to the interior of a larger animal: the greater the mass, the disproportionately more tracheae are required to reach the deepest muscles, and respiration becomes very inefficient. Increased partial pressures of oxygen in the atmosphere have the effect of allowing the gas to diffuse further through the network of fine tubes. Thus, as the atmosphere becomes hyperoxic the upper limits of arthropod size may have increased (Graham et al., 1995). Indeed, several authors have hypothesized that giant insects would have occurred during episodes of increased oxygen concentration (Rutten, 1966; Schidlowski, 1971; Tappan, 1974; Budyko et al., 1987; Dudley, 1998). Interestingly, the Late Paleozoic, when these giants existed, was a period of high oxygen concentrations.

During the Devonian, plants invaded land and rapidly proliferated. This expanded flora produced large volumes of oxygen as a photosynthetic byproduct, and concentrations continued to increase until reaching a peak during the Late Carboniferous (Berner and Canfield, 1989; Graham et al., 1995; Dudley, 1998, 2000). Although the peak was in the Carboniferous, what might be considered a hyperoxic atmosphere first came about during the mid- to Late Devonian when oxygen concentrations began to exceed today's levels. During the very end of the Paleozoic, oxygen concentrations began to decline, and indeed concentrations went steadily from their Carboniferous peak to well below today's level across the Permian. Thus, the decline of giant insects may not have been a result of the fateful End Permian Event but instead a factor of physics. This is highlighted by the brief reappearance of giant mayflies in the Hexagenitidae in the Cretaceous, when hyperoxic conditions were reached once again. It must also be emphasized that the Paleozoic giants were probably actively flying insects. The structure of mega-neurid wings and wing veins indicate that these insects had a maneuvered flight, which would have been too metabolically demanding without high oxygen levels.

Caution is required, though, when interpreting the gigantic sizes of some extinct insects. First, our basis of comparison is only the Recent - a geologically instant slice of time. The 400-myo fossil record of insects draws from a collective insect fauna that was many orders of magnitude more diverse than what exists today, so we may just be more likely to encounter rare giants by surveying the fossil record. This is especially true given the preservational bias toward larger insects as compressions in rocks. Also, while we tend to think that the giant insects have vanished, some still persist. Damselflies of the family Pseudostigmatidae can reach wingspans similar to those of Megatypus griffenflies (Protodonata: Meganeuridae), and there exist beetles today that are 4-6 inches long, walking sticks about 12 inches long, katydids (Tettigoniidae) with 8-inch wingspans, and some lepidopterans (e.g., the "white witch," Thysania agripinna [Noctuidae]) can have a wingspan up to 10 inches. No insects, though, ever matched the size of some of the meganeurid griffenflies.


Odonata are "bird-watcher's" insects (Figure 6.30). The aerial displays and complex behaviors of the order invite even casual observers, and it is little wonder that ornithologists make excellent odonatologists. Species are almost entirely diurnal and have acute vision and an active, powerful, and maneuvered flight. Because of their popularity and diversity (approximately 6,000 species), dragonflies and damselflies have received a great deal of attention from professionals and laymen alike. Overviews of odonate biology and taxonomy are by Allen et al. (1984, 1985), Bechly (1996), Steinmann (1997a,b),

6.30. A dragonfly rests before taking flight in White Rock, Canada. Photo: R. Swanson.

with major regional treatments by Westfall and May (1996) and Needham et al. (2000) for North America, Askew (1988) for parts of Europe, Needham (1930) for parts ofAsia, Watson et al. (1991) for Australia, and Pinhey (1951, 1961) for Africa. Classic references to the Odonata are by Tillyard (1917a), Corbet (1999), and Silsby (2001). The order has undergone numerous recent classificatory rearrangements based on phylogenetic studies (e.g., Fraser, 1954, 1957; Pfau, 1971, 1986, 1991; Carle, 1982; Trueman, 1996; Lohmann, 1996; Bechly, 1996), but the cladogram presented herein is based on the excellent study of adult and immature morphology by Rehn (2003) (Figure 6.31).

The order is defined by numerous traits discussed by Rehn (2003) as well as others. They include the development of a pterostigma, the formation of a nodus (albeit somewhat weak in some of the extinct, basal suborders), a complete absence of vein CuP beyond its attachment to CuA at the wing base, the presence of an arculus (Figure 6.32), the reduction of the thoracic terga, a mespisternum nearly touching the wings, direct flight muscles that power the wings out of phase with each other, a prehensile labial mask in naiads (Figure 6.33), a bristle-like antennal flagellum (convergent with Ephemeroptera), and pronounced skewness of the thorax. The degree of thoracic skew is variable within the order; in fact, the extremely oblique thorax of damselflies is a defining feature for that suborder (Needham and Anthony, 1903; Rehn, 2003). Although entirely paleopterous (i.e., unable to flex the wings posteriorly so that they fold over the abdomen during rest, usually flatly), damselflies bring their wings together during rest but over the abdomen (Figure 6.35). In damselflies the dorsal surface of the thorax nearly faces to the rear such that the wing apices are directed to the tail end.

Perhaps the most remarkable trait for the order is, however, the suite of modified male copulatory structures (Figure 6.34). The male terminalia have evolved into grasping appendages, while the actual copulatory organs are distant, on the ventral surface of abdominal segments 2 and 3. Males still produce sperm and emit sperm from a gonopore on the ninth abdominal segment at the tip of the abdomen, but the sperm must be transferred to the secondary genitalia before copulation is initiated. As such, reproduction in Odonata is far from simple. It begins when a female enters the territory of a male. Territorial males occur in two types: perchers and fliers. Fliers incessantly patrol a particular habitat, whereas perchers make regular, short exploratory flights from a fixed, local position, defending their territories from conspecifics and at times from males of other species. Con-specific males and females first recognize one another by their flight behavior, followed by coloration and overall body shape. Sometime during this watch a male will transfer sperm from his gonopore at the tip of the abdomen to his secondary genitalia underneath abdominal segment 2. Males will then grasp females as quickly as a positive identification can be made, generally in flight. The male will then attempt to align himself in tandem (tandem-linkage), grasping the female behind her head (in Epiprocta) or on the prothorax (in Zygoptera) with his terminalic claspers. The male may hold the female in this manner for several minutes to several hours, until she initiates copulation. Copulation takes place when the female extends her abdomen underneath her and forward, and her genitalia interlock with the male secondary genitalia, thereby transferring the sperm. Such a mating couple forms a characteristic copulation wheel (Figure 6.35).

6.31. Phylogeny of the Odonatoptera (living odonates and their extinct relatives), with significant characters indicated (Table 6.2). Modified after Rehn (2003).

Interestingly, the male copulatory organ is designed not only to insert sperm but also to remove it. Females are not monogamous and sperm competition between competing suitors can be intense. The penal structures of males are modified in different lineages to scoop any competitors' sperm out of the female's bursa copulatrix before depositing his own. Alternatively, some males will simply pack competitors' sperm tightly into the female such that his own will be accessible during the fertilization of eggs prior to oviposition.

This strategy is quite successful because sperm from the last male are more likely to encounter the egg as it passes the fertilization pore during oviposition. Males of some species further ensure the success of their progeny by forcibly guarding females after copulation and until the female has deposited her eggs. This guarding behavior can even take the form of essentially holding onto the female until she has oviposited, or even forcing eggs out (e.g., Libellulidae) by using his abdomen to thrust the female's abdomen into the water. It is

TABLE 6.2. Significant Characters in Odonatoptera Phylogeny0

1. Proximal process on posterior articular plate

2. MA fused with RP, base of M vestigial or fused to stem of Cu

3. Thorax slanting (slightly in Protodonata, more so in most Odonata)

4. Formation of nodus (albeit incipient in basal suborders)

5. CuP absent beyond attachment to CuA

6. Presence of arculus

7. Reduction of thoracic terga

8. Mesepisterna nearly touching dorsally

9. Prehensile labial mask in naiad

10. Antennal flagellum reduced ("aristate")

11. More completely formed nodus

12. Formation of true pterostigma

13. Loss of secondary branches in anal vein system

14. Fusion of posterior anal vein with posterior wing margin

15. Formation of posterior arculus at RP-MA separation

16. Subdiscoidal vein present

17. Frons bulbous

18. Epiprocts enlarged to form appendage of male claspers

19. Head transversely elongate

20. Absence of epiprocts in adults

21. Thorax strongly oblique

22. Three caudal gills in naiads a Numbers correspond to those on phylogeny, Figure 6.31.

6.32. Representative odonate fore wing.

unclear at what point in the evolution of Odonata (or Odonatoptera) this remarkable system of mating first came about.

Females deposit eggs directly into the water or in vegetation near fresh or brackish waters; some species prefer phy-totelmata (e.g., water captured in plants, such as in the tanks of bromeliads). Development proceeds quickly through the earliest instars (total number ranging from 11 to 13 instars), which rely mostly on nutrition from stored yolk while the young refine their predatory behavior. After the second or third instar, the naiads become more aggressive and adept at prey capture. Odonates use a prehensile mask, or modified labium, for prey capture (Figure 6.33). Prey are generally detected visually and captured by a rapid extension of the

6.33. Odonate naiads have a prehensile labial mask for capturing prey. The labium unfolds, spine-like palps impale the prey, and the labium quickly folds back to the mouth with the prey. Scanning electron micrographs. Photos: W. Wichard.

mask. Spine-like palpi at the apex of the labial mask impale the prey, and the folding labium draws the prey into the mandibles and mouth. Depending on the size of the species, prey ranges from small invertebrates (typically arthropods) to larval fish and amphibians. Aquatic respiration is achieved through the integument, supplemented in Epiprocta by a rectal chamber lined with gill pads or by long, external anal gills in Zygoptera. Eventually the naiad crawls from the water onto nearby vegetation to molt to the adult (Figure 6.36).

Hunting is also well developed in adults that, unlike mayflies, continue to feed and are, in fact, voracious aerial predators. Hunting behavior is slightly different between major lineages and once again breaks down into the fliers versus

6.34. The primary and secondary male genitalia of a damselfly (Zygoptera). Before mating the male will transfer sperm from the primary genitalia to his secondary genitalia. Photo- and scanning electron micrographs.

the perchers. The slanted thorax proves to be advantagous for prey capture. The oblique thorax, while pushing the wing bases backward, thrusts the legs forward. The legs are elongate and beset with numerous spines that together can form a basket of sorts. This allows individuals to grasp and control a victim effectively while aloft and to snag midges and other small insects in flight. Fliers, not surprisingly, capture and consume their prey while in flight. Perchers, which include most damselflies as well as most Gomphidae, Petaluridae, and Libellulidae, capture prey and then return to their roost to feast. Some damselflies pluck small arthropods off stems, and the long, tropical pseudostigmatines hover quite nicely while plucking insects out of spider webs. All types have visual acuity that is remarkable for arthropods. In fact, Odonata have

6.35. A damselfly mating wheel. The male (above) uses his primary genitalic claspers to grasp the female (below) by the neck and the female becomes impregnated by coupling with his secondary genitalia. Photo: S. Marshall.
6.36. The exuvium of a dragonfly naiad. Before emergence the naiad climbs up on a stem; even the chitinous tracheae (the light filaments here) are shed. Photo: V. Giles.

the largest compound eyes (these occupy nearly the entire head), which also possess the most facets of all insect eyes.

Subordinal Relationships and Early History

While the odonatopteran lineage is old, dragonflies (Epiprocta) and damselflies (Zygoptera) in the strict sense are not all that more ancient than many other insect lineages (see discussion that follows). Odonata in its broader sense stems well into the Paleozoic, based on suborders that are basal to a monophyletic lineage consisting of the Zygoptera and Epiprocta. The Odonata as a whole have had a remarkably successful geological history, and although still quite diverse today, they were equally diverse, if not slightly more so, in the past.

Suborder Protanisoptera. This group includes the extinct families Ditaxineuridae and Permaeschnidae (Carpenter, 1931, 1992) from the Early to Late Permian, which disappeared by the End Permian Event. Species have been recovered from deposits ranging from the central United States to Russia and Australia, and the suborder was likely cosmopolitan in distribution during its day. Unfortunately, protanisopter-ans are known only from their wings; the bodies and naiads remain to be discovered. They had relatively broad wings, nonpetiolate wings; the forewing being more slender than the hind wing with its broader base. The distinctive odonatan nodus was present but poorly formed, and the nodal crossvein was not completely developed (though the wing margin possessed a notch where such a nodal point would later appear). Similarly, the arculus was not entirely formed yet, and the "pterostigma" was merely a diffuse precursor crossed by the radial vein. Protanisoptera is currently held as the basalmost group of Odonata, all other members being united by a more complete formation of the nodus (albeit still somewhat incipient in Archizygoptera); a formation of a true pterostigma; a loss of secondary branches in the anal vein system, leaving just the anal brace; and a fusion of the posterior anal vein with the posterior wing margin (see also Rehn, 2003).

Suborder Archizygoptera. Basal relationships among early odonate families are contentious and continue to fluctuate in studies. We have therefore taken a conservative position herein and included the Protozygoptera within the Archizy-goptera (as was done by Carpenter, 1992). The Archizygoptera are the first suborder to truly take on a definite odonate appearance. The group is perhaps paraphyletic with respect to all other Odonata, and resurrection of Protozy-goptera and other putative suborders may be warranted as future studies continue to resolve relationships among families, some of the best know of which are the extinct families Kennedyidae, Permagrionidae, Protomyrmeleontidae (not to be confused with antlions of the Neuroptera!), Permolesti-dae, Permepallagidae, and Batkeniidae. Archizygopterans occurred throughout Europe, today's northern and central Asia, North America, Australia, and the Falkland Islands from the Permian through the Jurassic. Overall most species were relatively small and had petiolate wings resembling those of damselflies (hence the name Protozygoptera). Unlike the

Protanisoptera, Protodonata, and Geroptera, the Archizygoptera lacked a precostal area (similar to other Odonata); however, they had an incipient nodus (more closely approximating that of other Odonata) and a completely developed pterostigma, although the arculus remained undeveloped in most taxa.

Suborder Triadophlebiomorpha. This group derives from the Triassic of central Asia and has at times been confused as a post-Paleozoic member of the Protodonata, owing to the loss of the pterostigma (e.g., Grauvogel and Laurentiaux, 1952). Moreover, based on incomplete specimens, it was believed that the nodus was similarly absent, once again a primitive feature of Protodonata. However, the nodus was actually present in these insects, and an arculus was similarly developed. The wings, like Archizygoptera, tended to be petiolate although with crowding and fusion of longitudinal veins at the wing base. A complete formation of the nodus, including aligned nodal crossveins, and the formation of the discoidal cell unite Triadophlebiomorpha with all other Odonata. The group contains the Triassic families Triadophlebiidae, Triadotypidae, Mitophlebiidae, Zygophlebiidae, and Xameno-phlebiidae.

Suborder Tarsophlebioptera. The extinct family Tarsophlebi-idae has the distinction of being the sister group to the remainder of the modern Odonata. Although presently known only from the Jurassic (Figure 6.37), the tarsophlebiids likely occurred as early as the Permian owing to their phylo-genetic position. Tarsophlebia and its relatives appear to be the sister group to Zygoptera + Epiprocta based on a costal triangle in the wing and vein Sc turning sharply perpendicular to the long axis of the wing before meeting the costa (i.e., the completed formation of the nodus), among other traits discussed by Bechly (1996) and Rehn (2003).

The remaining two suborders are what most students of entomology truly think of as Odonata. The damselflies (Zygoptera) and dragonflies (Epiprocta) are supported as monophyletic in almost every study of odonate relationships (see earlier references). Defining features of the group include the development of the posterior arculus at the point of separation between veins RP and MA and the presence of a subdiscoidal crossvein.

Suborder Zygoptera. The damselflies are monophyletic and defined by the broad head, with the compound eyes separated by more than their own width; an absence of epiprocts in adults; a strongly oblique thorax; and the presence of three caudal gills in naiads. Damselflies are also characterized by similar, petiolate fore- and hind wings (Figure 6.38). The group occurs throughout the world, with species in the tropics reaching particularly large sizes (e.g., up to 6 inches in length and 8-inch wingspans in some pseudostigmatines).

6.37. A primitive odonate, Turanophlebia, from the famous Jurassic limestone of Solnhofen, Germany. Tarsophlebioptera were primitive odonates, and although they looked like dragonflies (Epiprocta), they were actually relatives of both Epiprocta-Zygoptera together. NHM In. 46336; forewing length 38 mm.

The earliest definitive Zygoptera are of the family Triassolesti-dae from the Triassic of South America, Australia, and Central Asia (Pritykina, 1981; Carpenter, 1992). Although called dam-selflies, the Late Permian Saxonagrionidae were actually more basal than the Zygoptera + Epiprocta lineage (Nel et al., 1999). Damselflies are abundant in Cretaceous and Tertiary deposits (Figure 6.39), although those in fossiliferous resins are rare (Figure 6.40).

Suborder Epiprocta. The suborder Epiprocta was established by Lohmann (1996) for the former suborders Anisoptera and "Anisozygoptera." Epiproctans are what most individuals consider true dragonflies (Figures 6.41, 6.42). These insects are robust with huge compound eyes that touch (or nearly so) and nearly occupy the entire head, and their wings are held out to the sides at rest. Defining features of the suborder include the large, bulbous frons; the enlargement of the epiprocts to form an inferior appendage as part of male ter-minalic grasping organ; and the development of a rectal chamber in the naiad lined with rectal pads for respiration. Additional features based on wing venation are discussed by Rehn (2003).

infraorder epiophlebioptera. The infraorder Epio-phlebioptera of Epiprocta contains only the family Epio-

Labium Odonata
6.38. Representative Recent damselflies. To the same scale.

6.39. An Early Cretaceous damselfly, Eoprotoneura hyperstigma, from the Santana Formation in Brazil. AMNH 44203; forewing length 18 mm.

phlebiidae. The family consists of only two species, Epiophlebia superstes from Japan and E. laidlawi from the eastern Himalayan region (Nepal). Epiophlebiids are "living fossils," which combine physical attributes of both Zygoptera and Anisoptera (Asahina, 1954). Overall the body of Epiophlebia resembles that of any other dragonfly. However, like the damselflies, the wings are petiolate, and the fore- and hind wings are relatively similar in overall shape. Both species occur in mountainous areas and breed in fast-flowing waters (possibly even in high-altitude waterfalls in the case of E. laidlawi). Epiophlebiids, unlike other Epiprocta, have functional ovipositors (similar to Zygoptera and more basal extinct lineages), and even though the naiads also have a rectal chamber lined with gills, they are apparently not capable of the "jet" propulsion so characteristic of Anisoptera. Epiophlebiids are not known from the fossil record, but their position as the living sister group to Anisoptera suggests that the lineage is at least as old as the Triassic (based on early "anisozygopteran" fossils).

infraorder anisoptera. The infraorder Anisoptera is used here in an expanded sense to include the large part of the former "Anisozygoptera." Anisozygoptera is a paraphyletic stem group to Anisoptera, and, despite some recent contentions (e.g., Carpenter, 1992), it does not include the living, relict family Epiophlebiidae. Instead, the numerous fossil Anisozygoptera are a stem-group assemblage leading to the Anisoptera (Rehn, 2003), which is now considered an infraorder of Epiprocta (Lohmann, 1996). The Anisozygoptera, excluding Epiophlebiidae, is linked with Anisoptera based on a hind wing that is broader than that of the forewing and which has a different venation, as well as a costal nodal "kink" (a small extension of the costal vein along the nodal crossvein).

Defining features of the Anisoptera in its traditional sense (here called Eteoanisoptera) are the following: a vestigial ovipositor; the ability of the naiad to propel itself using water pressure from its rectal chamber ("jet" propulsion); a pterostigmal brace vein (a supporting vein under the inner edge of the pterostigma); a distinctive anal loop (a series of distinctive cells demarcated in the anal region of the wing); a small, frequently darkened area of veinless membrane at the base of the wings near the anal region (often homologized with the jugum, though it appears to be secondarily evolved); a secondary "CuP" vein that is associated with an expanded

6.41 (left). Representative Recent dragonflies. From Genera Insectorum.

6.42 (right). Representative Recent dragon-flies. From Genera Insectorum.

6.43. A stunning Jurassic dragonfly, Libellulium longialata, from Solnhofen, Germany. NHM In. 28201; wingspan 140 mm.

anal region in the hind wing; and division of the discoidal cell into two triangular cells in the fore and hind wings. The Anisoptera includes the eteoansiopteran superfamilies Aesh-noidea, Libelluloidea, Petaluroidea, Gomphoidea, and Cord-ulegastroidea; extinct anisozygopteran lineages include families such as Archithemistidae, Asiopteridae, Euthemistidae, Heterophlebiidae, Isophlebiidae, Karatawiidae, Liassophlebi-idae, Oreopteridae, Progonophlebiidae, and Turanothemisti-dae. Relationships among these extinct families is uncertain, although Lohmann (1996), Bechly (1996), and Rehn (2003) all indicate them to be paraphyletic to Eteoanisoptera. The earliest fossils are from the Triassic, with definitive Eteoanisoptera in the Early Jurassic (e.g., Liassogomphidae of western Europe). Fossils of dragonflies are abundant (e.g., Figures 6.43, 6.44), including those of the aquatic naiads (e.g., Figures 6.45, 6.46), and are rich in characters that allow deciphering their relationships to modern odonate lineages.

6.45. An impressive naiad of the dragonfly Pseudomacromia sensibilis (Macromiidae) from the Early Cretaceous of Brazil. The long legs of this naiad suggest that it climbed amongst submerged vegetation to ambush prey. The long antennae may have served to detect prey in dense growth. AMNH 44205; length 21.3 mm.

6.45. An impressive naiad of the dragonfly Pseudomacromia sensibilis (Macromiidae) from the Early Cretaceous of Brazil. The long legs of this naiad suggest that it climbed amongst submerged vegetation to ambush prey. The long antennae may have served to detect prey in dense growth. AMNH 44205; length 21.3 mm.

6.44. Not all extinct odonatoids were giants; the smallest odonate that ever lived was Parahemiphlebia mickoleiti from the Early Cretaceous of Brazil. It is shown at approximately life size in the box. AMNH; wingspan 18.5 mm.
6.46. A dragonfly naiad in Early Cretaceous limestone from the San-tana Formation in Brazil. AMNH; length 21 mm.
+1 0

Post a comment