DRAGON-FLY (Ger. Wasserjungfer; Swed. trollslända; Dan. guldsmed; Dutch, scherpstekendevlieg; Fr. demoiselle), the popular English name applied to the members of a remarkable group of insects which formed the genus Libellula of Linnaeus and the ancient authors. In some parts of the United States they appear to be known as "devil's darning needles," and in many parts of England are termed "horse-stingers." It is almost needless to say that (excepting to other insects upon which they prey) they are perfectly innocuous, though some of the larger species can inflict a momentarily painful bite with their powerful jaws. Their true systematic position is still contested and somewhat uncertain. By most of the older systematists they were placed as forming part of the heterogeneous order Neuroptera. J. C. Fabricius, however, elevated them to the rank of a distinct order, which he termed Odonata; and whatever may be the difference of opinion amongst authors at the present day, that term is almost universally employed for the group. W. F. Erichson transferred all the groups of so-called Neuroptera with incomplete metamorphoses, hence including the dragon-flies, as a division of Orthoptera, which he termed Pseudo-Neuroptera. K. E. A. Gerstäcker more recently also retains them in the Orthoptera, terming those groups in which the earlier states are subaquatic Orthoptera amphibotica. All entomologists are agreed in maintaining the insects as forming a group marked by characters at once extraordinary and isolated in their nature, and in most modern classifications they are treated as a distinct order.
The group Odonata is divided into three families, and each of these again into two subfamilies. The families are the Agrionidae, Aeschnidae and Libellulidae - the first including the subfamilies Calopterygina and Agrionina, the second Gomphina and Aeschnina, and the third Cordulina and Libellulina.
Anatomy. - The structure of a dragon-fly being so very remarkable, it is necessary to enter somewhat extensively into details. The head is comparatively small, and excavated posteriorly, connected very slightly with the prothorax, on which it turns almost as on a pivot. The eyes are, as a rule, enormous, often contiguous, and occupying nearly the whole of the upper surface of the head, but sometimes (Agrionidae and Gomphina) widely distant; occupied by innumerable facets, which are often larger on the upper portion. The antennae, which are smaller in proportion than in almost any other insects, consist only of two short swollen basal joints and a 5 or 6-jointed bristle-like thread. The large labrum conceals the jaws and inner mouth parts. The lower lip, or labium (formed by the conjoined second maxillae), is attached to a very small chin piece (or mentum), and is generally very large, often (Agrionidae) divided almost to its base into two portions, or more frequently entire or nearly so; on each side of it are two usually enormous hypertrophied pieces, which form the "palpi," and which are often furnished at the tips with an articulated spine (or terminal joint), the whole structure serving to retain the prey. Considerable diversity of opinion exists with respect to the composition of the mouth parts, and by some authors the "palpi" have been termed the side pieces of the lower lip. The prothorax is extremely small, consisting of only a narrow ring. The rest of the thorax is very large, and consolidated into a single piece with oblique sutures on the sides beneath the wings.
The abdomen varies excessively in form, the two extremes being the filiform structure observable in most Agrionidae, and the very broad and depressed formation seen in the familiar British Libellula depressa. It consists of ten distinct segments, whereof the basal two and those at the apex are short, the others elongate, the first being excessively short. In a slit on the under side of the second in the male, accompanied by external protuberances, are concealed the genital organs: on the under side of the eighth in the female is a scale-like formation, indicating the entrance to the oviduct. The tenth is always provided in both sexes with prominent appendages, differing greatly in form, and often furnishing the best specific (and even generic) characters.
The legs vary in length and stoutness, but may, as a rule, be termed long and slender. The anterior pair probably assist in capturing and holding insect prey, but the greatest service all the legs render is possibly in enabling the creature to rest lightly, so that it can quit a position of repose in chase of passing prey in the quickest possible manner. The coxa is short and stout, followed by a still shorter trochanter; the femora and tibiae long and slender, almost invariably furnished on their under surface with two series of strong spines, as also are the tarsi, which consist of three slender joints, the last having two long and slender claws.
The wings are always elongate, and furnished with strong longitudinal neuration and dense transverse nervules strengthening the already strong (although typically transparent) membrane. In the Agrionidae both pairs are nearly equal, and are carried vertically and longitudinally in repose, and the neuration and membrane are less strong; hence the species of this family are not so powerful on the wing as are those of the other groups in which the wings are horizontally extended in a position ready for instant service. The neuration is peculiar, and in many respects without precise analogy in other groups of insects, but it is not necessary here to enter into more than some special points. The arrangement of the nervures at the base of the wing is very singular, and slight differences in it form useful aids to classification. In the Aeschnidae and Libellulidae this arrangement results in the formation of a triangular space (known as the "triangle"), which is either open or traversed by nervules; but in many Agrionidae this space, instead of being triangular, is oblong or elongately quadrate, or with its upper edge partly straight and partly oblique. This fixitude of type in neuration is not one of the least important of the many peculiarities exhibited in these insects.
The internal structure is comparatively simple. The existence of salivary glands, denied by L. Duprix, has been asserted by O. Poletajewa. The rest of the digestive apparatus consists of an elongate canal extending from mouth to anus, comprising the oesophagus, stomach and intestine, with certain dilatations and constrictions; the characteristic Malpighian vessels are stated to number about forty, placed round the posterior extremity of the stomach. Dragon-flies eat their prey completely, and do not content themselves by merely sucking its juices; the harder portions are rejected as elongate, nearly dry, pellets of excrement.
Fig. 1 - The anterior portion of the body of Aeschna cynea freed from the nymph-cuticle. Fig. 2. - The tail being extricated.
Fig. 3. - The whole body extricated.
Pairing. - But the most extraordinary feature in the economy - one which has attracted the attention of naturalists from remote times - is the position of the genital organs, and the corresponding anomalous manner in which the pairing of the sexes and impregnation is effected. In the male the intromittent organ is situated in a slit on the under surface of the second abdominal segment; it is usually very crooked or sinuous in form, and is accompanied by sheaths, and by external hooks or secondary appendages, and also by seminal vessels. But the ducts of the vessels connected with the testes unite and open on the under surface of the ninth segment; hence, before copulation can take place, it is necessary that the vessels in the second segment be charged from this opening, and in the majority of cases this is done by the male previously to seeking the female. In the latter sex the entrance to the oviduct and genital organs is on the under surface of the eighth abdominal segment. The act of pairing may be briefly stated as follows. The male, when flying, seizes the prothorax of the female with the strong appendages at the extremity of the abdomen, and the abdomen of this latter sex is then curved upward so as to bring the under side of the eighth segment into contact with the organs of the second segment of the male. In the more powerful Libellulidae, etc., the act is of short duration, and it is probable that polygamy and polyandry exist, for it possibly requires more than one almost momentary act to fertilize all the eggs in the ovaries of a female. But in many Agrionidae, and in some others, the male keeps his hold of the prothorax of the female for a lengthened period, retaining himself in flight in an almost perpendicular manner, and it may be that the deposition of eggs and pairing goes on alternately. There is, however, much yet to be learned on these points. The gravid female usually lays her eggs in masses (but perhaps sometimes singly), and the operation may be witnessed by any one in localities frequented by these insects. She hovers for a considerable time over nearly the same spot, rapidly dipping the apex of her abdomen into the water, or at any rate touching it, and often in places where there are no water-weeds, so that in all probability the eggs fall at once to the bottom. But in some of the Agrionidae the female has been often noticed by trustworthy observers to creep down the stems of aquatic plants several inches below the surface, emerging after the act of oviposition has been effected; and in the case of Lestes sponsa, K. T. E. von Siebold saw the male descend with the female. The same exact observer noticed also in this species that the female makes slight incisions in the stems or leaves of water plants with the double serrated apparatus (vulva) forming a prolongation of the ninth segment beneath, depositing an egg in each incision. He has seen two pairs thus occupied beneath the surface on one and the same stem.
Fig. 4. - The perfect insect (the wings having acquired their full dimensions) resting to dry itself, preparatory to the wings being horizontally extended.
Larva and Nymph. - The duration of the subaquatic life of a dragon-fly is no doubt variable, according to the species. In the smaller forms it is probably less than a year, but precise evidence is wanting as to the occurrence of two broods in one year. On the other hand, it is certain that often a longer period is requisite to enable the creature to attain its full growth, and three years have been stated to be necessary for this in the large and powerful Anax formosus. Like all insects with incomplete metamorphoses, there is no quiescent pupal condition, no sharp line of demarcation between the larval and so-called "nymph" or penultimate stage. The creature goes on eating and increasing in size from the moment it emerges from the egg to the time when it leaves the water to be transformed into the aerial perfect insect. The number of moults is uncertain, but they are without doubt numerous. At probably about the antepenultimate of these operations, the rudimentary wings begin to appear as thoracic buddings, and in the full-grown nymph these wings overlap about one-half of the dorsal surface of the abdomen. In structure there is a certain amount of resemblance to the perfect insect, but the body is always much stouter and shorter, in some cases most disproportionately so, and the eyes are always separated; even in those genera (e.g. Aeschna) in which the eyes of the imago are absolutely contiguous, the most that can be seen in the larva is a prolongation towards each other, and there are no ocelli. The legs are shorter and more fitted for crawling about water plants and on the bottom. In the mouth parts the mandibles and maxillae are similar in form to those of the adult, but there is an extraordinary and unique modification of the lower lip. This is attached to an elongate and slender mentum articulated to the posterior portion of the lower surface of the head, slightly widened at its extremity, to which is again articulated the labium proper, which is very large, flattened, and gradually dilated to its extremity; but its form differs according to group as in the perfect insect. Thus in the Agrionidae it is deeply cleft, and with comparatively slender side-pieces (or palpi), and strongly developed articulated spines; in the Aeschnidae it is at the most notched, with narrow side-pieces and very strong spines; in the Libellulidae it is entire, often triangular at its apex, and with enormously developed palpi without spines, but having the opposing inner edges furnished with interlocking serrations. The whole of this apparatus is commonly termed the mask. In a state of repose it is applied closely against the face, the elongated mentum directed backward and lying between the anterior pair of legs; but when an approaching victim is seen the whole apparatus is suddenly projected, and the prey caught by the raptorial palpi; in some large species it is capable of being projected fully half an inch in front of the head. The prey, once caught and held by this apparatus, is devoured in the usual manner. There are two pairs of thoracic spiracles, through which the nymph breathes during its later life by thrusting the anterior end of the body into the air; but respiration is mostly effected by a peculiar apparatus at the tail end, and there are two different methods. In the Agrionidae there are three elongate flattened plates, or false gills, full of tracheal ramifications, which extract the air from the water, and convey it to the internal tracheae (in Calopteryx these plates are excessively long, nearly equalling the abdomen), the plates also serving as means of locomotion. But in the other groups these external false gills are absent, and in their place are five valves, which by their sudden opening and closing force in the water to the rectum, the walls of which are furnished with branchial lamellae. The alternate opening and closing of these valves enables the creature to make quick jerks or rushes (incorrectly termed "leaps") through the water,  and, in conjunction with its mouth parts, to make sudden attacks upon prey from a considerable distance. Well-developed Aeschnid larvae have been observed to take atmospheric air into the rectum. The lateral angles of the terminal abdominal segments are sometimes produced into long curved spines. In colour these larvae are generally muddy, and they frequently have a coating of muddy particles, and hence are less likely to be observed by their victims. If among insects the perfect dragon-fly may be termed the tyrant of the air, so may its larva be styled that of the water. Aquatic insects and larvae form the principal food, but there can be no doubt that worms, the fry of fish, and even younger larvae of their own species, form part of the bill of fare. The "nymph" when arrived at its full growth sallies forth from the water, and often crawls a considerable distance (frequently many feet up the trunks of trees) before it fixes itself for the final change, which is effected by the thorax splitting longitudinally down the back, through which fissure the perfect insect gradually drags itself. The figures indicate this process as observed in Aeschna cyanea.
The Complete Insect.-For a considerable time after its emergence a dragon-fly is without any of its characteristic colours, and is flaccid and weak, the wings (even in those groups in which they are afterwards horizontally extended) being held vertically in a line with the abdomen. By degrees the parts harden, and the insect essays its first flight, but even then the wings have little power and are semi-opaque in appearance, as if dipped in mucilage. In most species of Calopterygina, and in some others, the prevailing colour of the body is a brilliant bronzy green, blue or black, but the colours in the other groups vary much, and often differ in the sexes. Thus in Libellula depressa the abdomen of the fully adult male is covered with a bluish bloom, whereas that of the female is yellow; but several days elapse before this pulverulent appearance is attained, and a comparatively young male is yellow like the female. The wings are typically hyaline and colourless, but in many species (especially Calopterygina and Libellulina) they may be wholly or in part opaque and often black, due apparently to gradual oxydization of a pigment between the two membranes of which the wings are composed; the brilliant iridescence, or metallic lustre, so frequently found is no doubt due to interference - the effect of minute irregularities of the surface - and not produced by a pigment. A beautiful little genus (Chalcopteryx) of Calopterygina from the Amazon is a gem in the world of insects, the posterior wings being of the most brilliant fiery metallic colour, whereas the anterior remain hyaline.
These insects are pre-eminently lovers of the hottest sunshine (a few are somewhat crepuscular), and the most powerful and daring on the wing in fine weather become inert and comparatively lifeless when at rest in dull weather, allowing themselves to be captured by the fingers without making any effort to escape. Many of the larger species (Aeschna, etc.) have a habit of affecting a particular twig or other resting place like a fly-catcher among birds, darting off after prey and making long excursions, but returning to the chosen spot. A. R. Wallace, in his Malay Archipelago, states that the inhabitants of Lombok use the large species for food, and catch them by means of limed twigs.
They are distributed over the whole world excepting the polar regions, but are especially insects of the tropics. At the present day about 2200 species are known, dispersed unequally among the several subfamilies as follows: Agrionina, 700 species; Calopterygina, 280; Gomphina, 320; Aeschnina, 170; Corduliina, 130; Libellulina, 600. In Europe proper only 100 species have been observed, and about 46 of these occur in the British islands. New Zealand is excessively poor, and can only number 8 species, whereas they are very numerous in Australia. Some species are often seen at sea, far from land, in calm weather, in troops which are no doubt migratory; the common Libellula quadrimaculata, which inhabits the cold and temperate regions of the northern hemisphere, has been frequently seen in immense migratory swarms. One species (Pantala flavescens) has about the widest range of any insect, occurring in the Old World from Kamtchatka to Australia, and in the New from the Southern States to Chili, also all over Africa and the Pacific islands, but is not found in Europe. The largest species occur in the Aeschnina and Agrionina; a member of the former subfamily from Borneo expands to nearly 6 in., and with a moderately strong body and powerful form; in the latter the Central American and Brazilian Megaloprepus caerulatus and species of Mecistogaster are very large, the former expanding to nearly 7 in., and the latter to nearly as much, but the abdomen is not thicker than an ordinary grass-stem and of extreme length (fully 5 in. in Mecistogaster).
Fossils. - Among fossil insects dragon-flies hold a conspicuous position. Not only do they belong to what appears to have been a very ancient type, but in addition, the large wings and strong dense reticulation are extremely favourable for preservation in a fossil condition, and in many cases all the intricate details can be as readily followed as in a recent example. From the Carboniferous strata of Commentry, France, C. Brongniart has described several genera of gigantic insects allied to dragon-flies, but with less specialized thoracic segments and simpler wing-neuration. These form a special group - the Protodonata. True Odonata referable to the existing families are plentiful in Mesozoic formations; in England they have been found more especially in the Purbeck beds of Swanage, and the vales of Wardour and Aylesbury, in the Stonesfield Slate series, and in the Lias and Rhaetic series of the west of England. But the richest strata appear to be those of the Upper Miocene at Oeningen, near Schaffhausen in the Rhine valley; the Middle Miocene at Radaboj, near Krapina in Croatia; the Eocene of Aix, in Provence; and more especially the celebrated Secondary rocks furnishing the lithographic stone of Solenhofen, in Bavaria. This latter deposit would appear to have been of marine origin, and it is significant that, although the remains of gigantic dragon-flies discovered in it are very numerous and perfect, no traces of their subaquatic conditions have been found, although these as a rule are numerous in most of the other strata, hence the insects may be regarded as having been drowned in the sea and washed on shore. Many of these Solenhofen species differ considerably in form from those now existing, so that Dr H. A. L. Hagen, who has especially studied them, says that for nearly all it is necessary to make new genera. It is of great interest, however, to find that a living Malayan genus (Euphaea) and another living genus Uropetala, now confined to New Zealand, are represented in the Solenhofen deposits, while a species of Megapodagrion now entirely Neotropical, occurs in the Eocene beds of Wyoming.
A notice of fossil forms should not be concluded without the remark that indications of at least two species have been found in amber, a number disproportionately small if compared with other insects entombed therein; but it must be remembered that a dragon-fly is, as a rule, an insect of great power, and in all probability those then existing were able to extricate themselves if accidentally entangled in the resin.
See E. de Selys-Longchamps, Monographie des Libellulidées d'Europe (Brussels, 1840); Synopses des Agrionines, Caloptérygines, Gomphines, et Cordulines, with Supplements (Brussels, from 1853 to 1877); E. de Selys-Longchamps and H. A. L. Hagen, Revue des Odonates d'Europe (Brussels, 1850); Monographie des Caloptérygines et des Gomphines (Brussels, 1854 and 1858); Charpentier, Libellulinae europeae (Leipzig, 1840). For modern systematic work see various papers by R. M'Lachlan, P. P. Calvert, J. G. Needham, R. Martin, E. B. Williamson, F. Karsch, etc.; also H. Tumpel, Die Geradflugler Mitteleuropas (Eisenach, 1900); and W. F. Kirby, Catalogue of Neuroptera Odonata (London, 1890). For habits and details of transformation and larval life, see L. C. Miall, Natural History of Aquatic Insects (London, 1895); H. Dewitz, Zool. Anz. xiii. (1891); and J. G. Needham, Bull. New York Museum, lxviii. (1903). For geographical distribution, G. H. Carpenter, Sci. Proc. R. Dublin Soc. viii. (1897). For British species, W. J. Lucas, Handbook of British Dragonflies (London, 1899). For wings and mechanism of flight, R. von Lendenfeld, S.B. Akad. Wien, lxxxiii. (1881), and J. G. Needham, Proc. U.S. Nat. Mus. xxvi. (1903). For general morphology, R. Heymons, Abhandl. k. preuss. Akad. (1896), and Ann. Hofmus. Wein, xix. (1904).
(R. M'L.; G. H. C.)
 A similar contrivance was suggested and (if the writer mistakes not) actually tried as a means of propelling steamships.
Note - this article incorporates content from Encyclopaedia Britannica, Eleventh Edition, (1910-1911)