are three extant (living) suborders: Zygoptera (damselflies), Anisoptera
(Dragonflies), and Anisozygoptera.
Features that occur in only one suborder (though not in every family) include:
Anisozygoptera (one family, two species: Japan and Himalayas) have an anisopteran-like body but zygopteran-like wings.
Note that two common rules of thumb do not work for the entire World fauna:
· dragonflies perch with wings open, damselflies with wings closed
These rules, however, do work for the majority of European and North American species.
Identification of Odonata to superfamily mostly is straightforward. However, for a minority of specimens such identification can be difficult, either because the required characters are confined to one sex or the species is in some way anomalous. Unfortunately, many odonate taxa grade into one another and some boundaries are not as well defined as they should be.
This key is designed to place any specimen to its currently accepted superfamily. In rare cases this process may be slow, but all the common forms should fall out quite readily.
The two most difficult superfamilies to separate are the narrow-winged, zygopteran superfamilies Coenagrionoidea and Lestoidea. In a majority of Coenagrionoidea the discoidal cell is sharply pointed and vein IRiii is at, while vein Riv is very near, the subnodus. All sorts of other combinations occur in Lestoidea.
Unfortunately, some South American Lestoidea in family Megapodagrionidae (subfamily Megapodagrioninae) show this same combination of characters. The taxonomist's distinction between these superfamilies depends on features of the male secondary genitalia, but this is of no use for identification unless the specimen can be killed and dissected, and is male.
To positively separate the common and widespread Coenagrionoidea from this subfamily of South American Lestoidea, check for short, supplementary longitudinal veins in the outer part of the wing. Coenagrionoidea never have these but many Lestoidea do, and in Megapodagrioninae they are very well developed. The relevant character has been placed at the end of the key.
There are six families in the Calopterygoidea. Most Calopterygoids are readily identifiable to family and subfamily level and most identifications using this key should take no more than 3-4 steps.
However, in order to cover every genus worldwide, allowance has been made for groups which do not fit the 'typical' pattern for their family. Because of this, some combinations of character states do not eliminate taxa which a user might expect would be eliminated.
Choices which always will work include the following:
Amphipterygidae, and no other Calopterygoidea (as the term is defined for this key, but see Lestoidea for similar taxa), always have two primary antenodals plus a few incomplete secondaries (character 3).
There are six families in the Coenagrionoidea, with a number of subfamilies. Identification to subfamily can be tricky. Here is a general guide to identification, with additional notes to the six subfamilies of the Coenagrionidae.
Three families have a narrowed wing with reduced venation in the cubital and anal fields. These are Isostictida, Protoneuridae, and Platystictidae.
Isostictidae: Extremely narrow wings, the cubital vein reaching at most 1-2 cells beyond the level of the subnodus, the anal vein not developed at all.
A fourth family, Pseudostigmatidae (the Forest Giants) can be identified by very large size and the pterostigma either absent or replaced by a 'false pterostigma' with veins running through it.
The two remaining families, Platycnemidae and Coenagrionidae, have a non-reduced venation.
Platycnemidae: Dc is rather blunt. In most species the tibiae (2nd-last leg segment) of the males are distended, hence the common name "Featherlegs".
Subfamilies of Coenagrionidae
The differences among the subfamilies are small and some characters are rather easily mistaken. As a guide:
Agriocneminae are very small damselflies in which the arculus is well distal of the second antenodal.
Superfamily Lestoidea contains five families: Synlestidae, Lestidae, Lestoideidae, Megapodagrionidae and Perilestidae.
Although this key notionally runs to subfamily level, many subfamilies of Lestoidea contain only one genus, sometimes just a single species. As well, the geographical distributions of most lestoid species are restricted and this can be an additional clue to identification. In effect, the key may often allow identification to below subfamily level.
Terminal taxa are subfamilies, except for Lestoideidae (which is keyed as three separate genera because they are so different from one another) and the megapodagrionid subfamily Rhipidolestinae (for which the anomalous genus Pseudolestes gets a separate entry). The final family, Perilestidae, is not divided into subfamilies.
Some Lestoidea are very readily identifiable to subfamily but others may be difficult. This is particularly so for members of the larger, more heterogeneous families. Some taxa are defined more by features they don't have than by those they have, and taxonomic work is needed to clarify the relationships both among and within the families. Here is a broad guide to families and subfamilies:
Synlestidae: Cu arches forward as it leaves Dc (see also Perilestidae). There are three subfamilies. In Chorismagrioninae the forewing Dc is open at its base. In Synlestinae, IRiii/Riv arises near the nodus and the anal vein leaves the wing border well after Ac. In Megalestinae, Riii/Riv arises near the arculus and A leaves the wing border at or proximal to Ac.Distributions are: Chorismagrioninae, NE Queensland (Australia), Synlestinae (=Chlorolestinae), southern Africa, southern Asia, Australia, Haiti. Megalestinae, India through China.
Lestoideidae: There are no obvious characters that unite this family but the three genera are distinctive.
Lestoidea: Moderately small size, Dc is short and rectangular. Only two antenodals. Cu is short and the anal vein is minute.
Perilestidae: Thin damselflies with very long bodies for the size of their wings. Dc touches or almost touches the wing border, the anal field being completely reduced at this point. Cu arches forward as in Synlestidae. Wings with brown or black bands. Distribution: South America.
Lestidae: The distal angle of Dc is pointed. IRii/Riv arises closer to the arculus than to the nodus. There is a small but distinct oblique vein between Riii and IRiii. None of these features is unique to this family but Lestidae can be distinguished from others by not having the features of Synlestidae or Perilestidae.
The two subfamilies of Lestidae are speciose and widely distributed:
Lestinae have Dc equal in fore- and hindwings and typically rest with the wings half open.
Megapodagrionidae: This family was formed to hold lestid-like genera in which IRiii/Riv arises near the nodus and the male genital hamules are quadrate. "Megapod" suggests unusually large legs or feet but this is true only of the largely South American Megapodagrioninae. Members of most subfamilies habitually rest with wings outstretched.
Although this family is split into seven subfamilies, three contain only a single species, one contains two species and one 11 species. The two large subfamilies are Argiolestinae (28 genera) and Megapodagrioninae (4 widely disjunct genera).
Argiolestinae: Small to moderately large megapodagrionid damselflies with not particularly long legs, Dc less pointed than in Megapodagrioninae. Anal vein leaves the wing border distal of Ac, often level with the proximal end of Dc. Widespread.
Superfamily Aeshnoidea contains four families: Aeshnidae Gomphidae, Neopetaliidae, and Petaluridae.
Superfamily Cordulegastroidea contains one family, Cordulegastridae, which has many features in common with Aeshnidae and is included for convenience in this key.
These five families are fairly easy to separate provided, in the case of Cordulegastridae, that specimens of the appropriate sex (female) are to hand. Male cordulegastrids can look a lot like some aeshnids.
In Aeshnidae the eyes touch or fuse completely along the midline. Most are large, powerful insects. Some have bands of brown or other colour on the wings, and coloured markings on the body are commonplace. This family is widespread and ubiquitous.
The simplest description of Neopetaliidae is "dragonflies of aeshnid appearance but with a series of red spots along the leading edge of each wing". At times, Neopetaliidae has been treated as a subfamily of Aeshnidae. This small family of nine species is confined to South America west of the Andes and to south-eastern Australia.
Cordulegastridae (superfamily Cordulegastroidea) are yellow and black (or brown) aeshnid-like dragonflies. The females carry a unique, elongated, secondary ovipositor.
In both Gomphidae and Petaluridae the eyes are well separated, not touching at all on the midline. (The eyes of Cordulegasteridae females also fail to touch while those of the males just touch.)
Gomphidae have a reduced ovipositor in which the terebra, even if well developed (usually they are vestigial) do not enclose the usual complement of valves. The family is divided into four subfamilies, although the divisions are not very clear and it is not certain that the current classification follows the true relationships. Subfamilies are separated on relatively minor venational characters, mainly of the triangles and the anal loop.
Petaluridae males have distinctive, leaf-like superior appendages. The females have a fully functional ovipositor. This small family of 11 species is distributed in Australia, New Zealand, Japan, and North and South America.
Libelluloidea is a huge superfamily with very many species. At a local level, most libelluloid dragonflies are quite easily recognised, often to genus or species. However, identification to the higher taxonomic categories can be far from easy. This is because the categories themselves are quite poorly defined. In some instances they grade into one another. The true evolutionary relationships within Libelluloidea are poorly known and many of the named taxa appear to be quite arbitrary.
There are five families: Chlorogomphidae, Macromiidae, Synthemistidae, Corduliidae, and Libellulidae. The first three of these each contain relatively few species and have a distinctive wing venation. Two of them are geographically quite restricted. These families are easy to recognise.
The other two families, the very speciose Corduliidae and the even more speciose Libellulidae, are not so easy. There is no single character that separates Corduliidae from Libellulidae. Genera are assigned to one or other family on the basis of combinations of traits.
The same applies at subfamily level. Each of these two families is too unwieldy to be treated as a single unit and has been divided into several subfamilies. Whether we take the scheme proposed by F. Ris (Monograph Libellulines, 1910-1919), the one by F. C. Fraser (A Reclassification of the order Odonata , 1957), or any of several later modifications, the subfamilies remain ill-defined. Here, we recognise six subfamilies in Corduliidae and 11 in Libellulidae: broadly following Davies and Tobin (1985) and Bridges (1994) who in turn follow Fraser's basic plan. However, these almost certainly do not arrange the genera in any evolutionary order.
A difficulty for any interactive key is that when all character states appear in all taxa, albeit in different combinations, the selection of a state will fail to eliminate any taxa. Identification then becomes impossible. The alternatives are (i) break the taxa into smaller parts (e.g., treat individual genera separately), (ii) invent complex characters (e.g., a character such as "median space free and discoidal field starting with one row of cells, versus median space crossed and pterostigma small", might work), or (iii) include computer instructions "in the background" to disallow combinations of states that do not occur.
We have used all three strategies, but mainly (i) and (iii), preferring to keep the list of characters as simple as possible. Even so, it is by no means certain that every species will key out completely using this key. The most difficult groups to separate at-the margin (i.e., when considering those odd genera that could be in one subfamily but could be in another) can be expected within the series of libellulid subfamilies from Brachydiplacinae to Urothemistinae.
Brachdiplacinae (Dwarves) are similar to but smaller than Sympetrinae (Darters). These in turn can be rather similar to some Libellulinae (Chasers). The Leucorrhininae (White Faces) are rather like Darters and rather like Chasers, and not all of their faces are white. The Trithemistinae (Dropwings) grade somewhat imperceptibly toward Sympetrinae both in morphology (wing venation) and behaviour (resting with the body pointing up and the wings pointing down).
The Onchothemistinae have a coppery metallic thorax which readily separates them from other Libellulidae (except Zygonychinae) despite a complete lack of defining wing vein characters. Palpopleurinae (Widows) would be well distinguished by the unique way in which the costal margin of the forewing bulges forward, except that this does not occur in the genus Perithemis which gets placed here largely because the markings on its wings are similar to those in some Palpopleura.
The Trameinae (Gliders) could by no means be mistaken for Sympetrinae, yet in their short bodies and long wings they sometimes are reminiscent of Libellulinae genera such as Libellula. The base of the wing is more developed, but just how much more it takes to make a Trameinae is quite difficult to quantify. The similarities are greatest in the tribe Trameini. Tribe Rhyothemistini can be identified purely on (lack of) body length, and tribe Zygommatini on the shape of the anal loop (which is open at the wing border).
The Urothemistinae (Baskers) return us to a sympetrine- or brachydiplactine- like form, albeit with fewer cells in the wings, the sectors of the arculus are not fused and the two primary antenodals are just distinguishable, as in Corduliidae. This subfamily sometimes is regarded as a separate family (not as Urothemistidae, but Macrodiplactidae).
Finally, the Zygonychinae (Cascaders) recall the family Corduliidae rather than any other subfamily of Libellulidae. Zygonychinae are emerald green and black, metallic insects with corduliine looks and corduliine behaviour (for example, they rest with the body hanging downward). Yet, venationally, they have all the characteristics of "advanced" libellulids such as Onychothemistinae or Trameinae.