Smart Engineering in the Mid-Carboniferous: How Well Could Palaeozoic Dragonflies Fly?

See allHide authors and affiliations

Science  23 Oct 1998:
Vol. 282, Issue 5389, pp. 749-751
DOI: 10.1126/science.282.5389.749


The wings of archaic Odonatoidea from the mid-Carboniferous of Argentina show features analogous to “smart” mechanisms in modern dragonflies that are associated with the agile, versatile flight necessary to catch prey in flight. These mechanisms act automatically in flight to depress the trailing edge and to facilitate wing twisting, in response to aerodynamic loading. The presence of similar features suggests that the earliest known odonatoids were already becoming adapted for high-performance flight in association with a predatory habit.

Insect wings are the principal adult locomotory structures of the largest group of animals. They are proving to be spectacular examples of microengineering. They lack internal musculature, and their three-dimensional shape during the flapping cycle is largely determined by their elastic response to aerodynamic and inertial forces, moderated by thoracic muscles inserted at or near the base (1).

The dragonflies (order Odonata) are supremely versatile, maneuverable fliers, and this is reflected in their wing morphology. High-speed cinefilm, videotape, and still photographs show that the wings twist extensively along their span, allowing the insects to develop weight support on both up- and downstrokes and giving fine instantaneous control of aerodynamic forces over a wide speed range (2–5). This twisting is facilitated by a flexible, relatively unsupported posterior wing margin. Useless flaglike fluttering is prevented by a series of automatic devices within the wings (2, 4, 5); the latter are in effect “smart” aerofoils, passively maintaining appropriate profiles and attitudes throughout the stroke in direct response to the external forces that they experience.

Odonata show several such devices. Two examples are the “nodus,” and the basal complex comprising the “supratriangle” and “triangle” of entomologists, characteristic of the suborder Anisoptera (Fig. 1A).

Figure 1

(A to C) Hind wing ofAeshna cyanea (Müller) (Odonata, suborder Anisoptera), showing the main longitudinal veins and the detailed structure of the nodus and the triangle-supratriangle complex. (A) The whole wing, with the leading-edge spar shaded; (B) the nodus; (C) the basal complex. MA, anterior median vein; ScP, posterior subcostal vein. (D) Diagrammatic representation of the operation of the triangle-supratriangle complex of anisopterous Odonata in lowering the trailing edge in response to an upward force applied in the distal median area of the wing.

The nodus (Fig. 1B) is a combined brace and shock absorber, at the junction of two regions of the leading edge spar with very different mechanical properties. The proximal region consists of three veins linked by angle bracket–like cross-veins into a rigid girder. At the nodus the concave subcostal vein (ScP in Fig. 1B) ends, and the spar continues distally as a shallower girder with weaker cross-veins and an inverted V-shaped section. This latter conformation typically allows supinatory twisting while restricting pronatory twisting (5, 6) and permits the passive upstroke torsion described above. This differentiation of the supporting spar, basally rigid and distally compliant, is hence an adaptation for versatile, maneuverable, multispeed flight.

The basal complex of Anisoptera (Fig. 1C) is an angular, strongly three-dimensional conformation of veins. A force manually applied more distally to the underside of the wing of a living insect, so simulating the lift, levers the trailing edge down about this basal complex, cambering the wing and increasing its angle of pitch (2,5). Figure 1D shows diagrammatically the operation of this mechanism, which is easily modeled in thin card. A force applied at W, representing the anterior median vein (MA in Fig. 1, A and C), lowers the regions anterior and posterior to the modeled basal complex XYZ. In the actual wing the anterior area is relatively rigid, so that downward deflection is concentrated posteriorly. In flight this holds down the trailing edge, improving the wing's camber and attitude in direct response to aerodynamic loading (2,5).

The Odonatoidea have a long and excellent fossil record. Apparent automatic mechanisms familiar in modern forms can be identified in many Mesozoic fossils, and their evolution would reward detailed study. They are less evident in most Palaeozoic members of the odonatoid stem group. The familiar giant meganeurid protodonates of the Carboniferous and Lower Permian, though strikingly convergent in wing planform with modern Anisoptera, lack both nodus and basal complex. However, the most archaic known odonatoids are the far smaller Eugeropteridae from the mid-Carboniferous of La Rioja, Argentina (7). They had wingspans between 80 and 100 mm, well within the size range of modern dragonflies. Until recently only wings were known, but new material, currently being studied, shows Eugeropteridae to have been proportioned rather like modern Libellulidae, but with relatively smaller bodies, and with prothoracic winglets—an archaic feature familiar in some other Carboniferous insects, but previously unknown in odonatoids (Fig. 2A).

Figure 2

(A) Undescribed eugeropterid from la Rioja, Argentina (Carboniferous, Namurian/Westphalian). Photograph of a latex cast. (B) The hind-wing base of Eugeropteron lunatum Riek. (C) Diagrammatic representation of the operation of the basal mechanism inEugeropteron, as here interpreted. CuA, anterior cubital vein; MA, anterior median vein.

Despite their many plesiomorphic characters, these insects show specializations that parallel those just described in modern dragonflies and provide good evidence for early development of versatile flight techniques. There is no nodus, but the stiffening posterior subcostal vein meets the fore-margin well before the wing tip, so that the distal part of the wing would have been relatively compliant to supinatory twisting.

Both fore and hind wings show a basal, three-dimensional vein complex that superficially resembles that of Anisoptera, and when modeled in card responds to manual loading in a similar manner. These are certainly analogous, not homologous, adaptations, because different veins are involved in the two groups, and the Eugeropteridae are unlikely to be directly ancestral to modern Odonata. It seems clear that smart, trailing edge–lowering mechanisms arose independently in these early forms, associated as now with a flexible posterior margin in torsionally compliant wings.

Figure 2B shows the form of the basal part of the wing, and Fig. 2C a schematized version of the mechanism as we interpret it, with the complex shown as a double, three-dimensional parallelogram. An upward force applied at W—here representing the anterior cubital vein (CuA)—flexes the parallelogram about its diagonal and applies depressing torques to the wing both anteriorly and posteriorly. In the actual wing the diagonal flexion of the complex would slightly bend three veins, and their elastic recovery may have aided the reversal of the mechanism when the aerodynamic load was removed. The mechanism in Eugeropteron parallels that in modern Anisoptera in that both involve raising the apex of an L-shaped vein formation—XYZ in both Figs. 1D and 2C—so tending to twist downward the regions in front and behind. Again, the greater rigidity of the anterior part of the wing would resist this torsion, and to compensate, the posterior area would be depressed further.

How, and how well, did these insects fly? Certainly not as skillfully as modern Anisoptera, which coevolved in the Mesozoic and Tertiary with the many groups of agile insects on which they feed. Carboniferous odonatoids were already predatory (8), but no contemporary insects would have approached the maneuverability of many extant species. The relatively smaller, less sturdy bodies and consequent lower wing-loadings of Eugeropteridae indicate lower maximum speeds than in Anisoptera, and the absence of a nodus suggests a poorer capability for wing twisting, truncating the lower end of their speed range: It is unlikely that they could hover like modern dragonflies. Nonetheless, the flexible trailing edge and the shortened subcostal vein indicate that some supinatory twisting was possible, and the group appears to be following a trend, parallelled in many other insect groups, toward improving flight versatility by recruiting upstroke forces to supplement those of the far more effective downstroke, and varying their magnitude and direction at need. The “smart” wing-base mechanism is best interpreted as an elegant means of maintaining downstroke efficiency in the presence of these adaptations to improve upstroke usefulness.

  • * To whom correspondence should be addressed. E-mail: r.j.wootton{at}


View Abstract

Navigate This Article