Corduliidae

Fleck, Günther & Neiss, Ulisses G., 2012, The larva of the genus Paracordulia Martin, 1907 (Odonata: Corduliidae s. s.) and a generic key to the larvae of Corduliidae s. l. occuring in South America, Zootaxa 3412 (1), pp. 62-68 : 66-68

Key to larvae of Corduliidae s. l. genera occurring in South America (late stadia)

We follow Garrison et al. (2006), Carvalho et al. (2008) and Fleck (in press) who considered Schizocordulia , erected by Machado (2005) for the monotypic Aeschnosoma rustica Hagen in Selys, 1871, as a synonym of

Aeschnosoma . Machado (2012) has revalidated the genus Schizocordulia on the base of six characters. We think it is unnecessary to revalidate this genus because of the following reasons:

- Machado’s revalidation is mainly based on male characters; as the male of A. auripennis Geijskes is still undescribed, it is premature to consider characters exclusive to Schizocordulia prior to knowing their states in all known Aeschnosoma species ;

- Machado mingles practical characters used in a key and generic diagnoses; he also ignores non-neotropical Corduliidae and does not explain why, for example, the closely related genus Pentathemis Karsch , sharing characters with Schizocordulia and Aeschnosoma , is not considered;

- the larva of Schizocordulia has been recently discovered (Pinto, unpublished) and is a “typical” Aeschnosoma , morphologically strongly similar to the other larvae of the genus (Pinto, pers. com.);

- the membranule of A. hamadae Fleck & Neiss ends slightly distal to the anal triangle, and the membranule of a new species of Aeschnosoma (Fleck, in press) ends distinctly distal to anal triangle; Machado’s character (6) is therefore invalidated;

- the genital lobes of A. auripennis (pers. obs.) and those of a new species of Aeschnosoma (Fleck, in press) are distinctly much more reduced than those of remaining Aeschnosoma ; Machado’s character (5) loses its consistency as size of genital lobes can be considered greatly variable within the genus Aeschnosoma ;

- the shape and length of vulvar lamina are generally variable among species of a given gomphid or libelluloid genus; they can even appear extremely variable in many genera ( Somatochlora [ albicincta vs calverti ], Leucorrhinia [ caudalis vs rubicunda ], Sympetrum [ corruptum vs janeae ], etc.); the shape of the vulvar lamina of a new species (Fleck, in press), and to a lesser extent that of the related A. auripennis , also strongly differs from that of remaining Aeschnosoma , and like Schizocordulia presents two tapering subtriangular processes (even if distinctly shorter); therefore it is not a unique character of Schizocordulia , and proves at least the great variability of the shape of the vulvar lamina within the genus Aeschnosoma ; Machado’s character (4) is invalidated;

- a tenuous S7 pilosity visible on A. forcipula is considered homologous to that of Schizocordulia (Fleck, pers. obs.; see also Geijskes 1970: Fig. 12); furthermore, as marked pilosity on S8 can also be present in some Aeschnosoma and absent in others, the male ventral pilosity of the genus Aeschnosoma (and related pilose plates) can be considered variable, and Machado’s character (3) regarded as little convincing;

- the posterior hamulus of A. auripennis (pers. obs.) and that of a new species of Aeschnosoma (Fleck, in press) exhibit a well developed internal branch, thus Machado’s character (2) is invalidated;

- a bifid male epiproct is not a strict generic character, and several libelluloid genera include species with “irregular” bifid epiproct ( Somatochlora [ cingulata vs provocans ], Leucorrhinia [ frigida vs intacta ], Tetrathemis [ bifida vs polleni ], etc.); the monotypic genus Corduliochlora Marinov & Seidenbush has a moderately bifid epiproct but should be considered as a synonym of Somatochlora (see Fleck et al., 2007); as the shape of the epiproct of a new species of Aeschnosoma (Fleck, in press) is very different to that of A. elegans Selys (even if not bifid), the genus Aeschnosoma should be considered to have a great variability in size and shape of the male epiproct, and Machado’s character (1) as just another different model compatible with this great intrageneric variability.

On the light of these arguments, the genus Schizocordulia should be regarded as a junior synonym of Aeschnosoma . As already stated by Fleck (in press), the genus Aeschnosoma must be considered as greatly polymorphic, and prior to evaluating generic or intra-generic taxonomy, a revision of this genus, integrating sister genera and a phylogenetic discussion, is needed.

The larva of Paracordulia differs significantly from all other known South American larvae, and in the current state of knowledge, misidentifications are very unlikely. From larval characters, it is also apparently not related to any known Neotropical larva and could represent an old stock of the Amazonian forest. This genus completes the knowledge of larvae at the generic level for South America with eight known genera, and even for the New World with 15 genera ( Garrison et al. 2006). Nonetheless it must be kept in mind that larvae of only a few species are described from South America (see Carvalho et al., 2008). Therefore the following key treating larvae of corduliid genera occurring in South America, partly modified from that of Carvalho et al. (2008), is susceptible to change with the discovery of new larvae. 1 Indentations of the distal margin of labial palps large and few in number (<6) and totally deprived of raptorial setae; abdominal lateral spines absent.................................................................... Gomphomacromia

- Indentations of the distal margin of labial palps more numerous (>5) and with raptorial setae; abdominal lateral spines present ................................................................................................... 2

2 Abdominal mid-dorsal hooks (spines or knobs) present........................................................ 3

- Abdominal mid-dorsal hooks absent....................................................................... 6

3 Abdominal lateral spines present anterior to S8 (those of S5–S7 generally very distinct)............................ 4

- Abdominal lateral spines anterior to S8 absent.............................................................. 5

4 No conical protuberances on dorsal surface of occiput; frontal plate prominent and furnished with distinct spinelike and scalelike setae........................................... Lauromacromia [apparently absent from Chile and Argentina]

- A pair of prominent conical protuberances present on dorsal surface of occiput, frontal plate more reduced and covered with microscopic prickles..................... Rialla [reported only from South Chile and Argentinean boundary with Chile]

5 Lateral abdominal spines on S8 and S9 strongly developed, with those of S9 largely overtaking the apex of the anal pyramid, distinctly divergent, and with their apices upturned................................................. Paracordulia

- Lateral abdominal spines on S8 and S9 not greatly developed, with those of S9 not overtaking the apex of the anal pyramid, not divergent, and with their apices not upturned...................................................... Navicordulia

6 Lateral abdominal spines on S9 very strongly developed and largely overtaking the apex of the anal pyramid; posterior margin and lateral margins of the occiput separated dorsally by a low tubercle made by a verticil of strong spinelike setae................................................................................................... Aeschnosoma

- Lateral abdominal spines on S9 poorly developed and never overtaking the apex of the anal pyramid; posterior margin and lateral margins of the occiput not separated dorsally by a low tubercle made by a verticil of strong spinelike setae........... 7

7 Antennae short, shorter than middorsal length of head; frontal plate prominent and with scalelike setae......... Neocordulia

- Antennae long, longer than middorsal length of head; frontal plate not prominent and without scalelike setae. Cordulisantosia