The rapid increase of natural ecosystems destruction has leaded to the extinction of a great number of species. Because most species are still undescribed, cataloging and explaining current diversity is the focus of much active research. One of the immediate consequences of these efforts has been an increase in the research on cryptic species over the past two decades, mainly due to the increasing availability of DNA sequences. The identification and description of cryptic species provides opportunities to study important mechanisms of speciation, mate recognition and natural resource protection and management (Bickford et al. 2007).
The biological species concept, defined and elaborated mainly by Enrst Mayr (1996) defines species as ‘groups of interbreeding natural populations that are reproductively isolated from other such groups’. Under this concept, new species are formed when they are reproductively isolated. However, as mechanisms of reproductive isolation differ among taxa, no universal criterion to delimit species exists yet. DNA barcoding (i.e. the generation of DNA sequences for all named species on the planet) has recently emerged as a powerful tool for taxonomy, as it allows us to distinguish species and to discover new ones, using genetic (or molecular) information. However, a current standard is that DNA alone should not be the unique source of information for species designations and that diagnostic characters from classical taxonomic approaches (morphology, ecology) should be used to infer the rank of populations (DeSalle et al. 2005).
The work of Damm et al. (2010) is based on the ‘taxonomic circle’ proposed by DeSalle et al. (2005), a term which designs a methodology for new species discovering that combines traditional taxonomic work with the power of modern molecular tools. The main idea within the concept is to ‘break’ the circle of tautological reasoning, i.e. taxonomists should use information based on geography (or morphology, or other relevant discipline) to generate a null hypothesis. Then, this hypothesis is tested with relevant information on the geographical distribution, morphology, ecology, reproduction and behaviour of the putative taxa, integrated in the circle. Only when at least two lines of evidence support the presence of several taxa, are these entities recognized as species and the circle is broken [see DeSalle et al. (2005) for more details].
Damm et al. (2010) initially started a population genetic study in the African libellulid dragonfly Trithemis stictica (for Trithemis spec. nov. see Fig. 1) but, surprisingly, three genetically distinct clusters were found, based on genetic evidence from NDI (NADH dehydrogenase 1) and COI (cytochrome c oxidase subunit I). Genetic distances between the clades were surprisingly high (5–9%), whereas distance within clades was low (0–1%). Individuals of two of the clades occurred sympatrically in one site in Namibia (Popa Falls, Fig. 2).
This finding was the basis for the hypothesis of three cryptic species in T. stictica, which was tested by using the ‘taxonomic circle’ of DeSalle et al. (2005), including information on genetics, morphology and ecology of the different clades found.
Phylogenetic analyses corroborated the results of the distance analyses and separated individuals into three clades: clade 1 grouped together with other Trithemis species according to the classical taxonomy of T. stictica and thus it was identified as the originally described T. stictica. Clades 2 and 3 appeared as sister species, with high phylogenetic support. Furthermore, the three clades are distinguishable by unambiguous diagnostic characters (or barcodes).
A re-examination of 43 male specimens, previously genetically characterized, showed consistent differences in morphology between the clades. The ‘true’T. stictica can be distinguished from clades 2 and 3 by eye and base wing coloration, as well as by differences in penis structure. Clades 2 and 3 are distinguished only by size.
Finally, an analysis of ecological patterns revealed differences in habitat preferences among the three genetic clades, suggesting again that several taxa were involved in this system.
In summary, significant genetic isolation, differences in habitat preferences, fixed size differences and reproductive isolation together provide evidence to support the hypothesis of two new sympatric Trithemis species, which shows the importance of combining information from different disciplines to define species limits (Damm et al. 2010).
Odonates have been used for decades as model systems for many evolutionary and ecological studies (Córdoba-Aguilar 2008). Their elaborate mating behaviour, which includes in some cases a complex courtship, based on visual displays and their powerful flight (Corbet 1999) a priori suggest that cryptic speciation is unlikely to occur in this insect group. This work by Damm et al. (2010) demonstrates that these prejudices are no longer justified.
African odonates have generated several surprises on what a species is and how to delimit them. Very recently, close examination of one of the commonest African dragonflies, Brachytemis leucosticta, showed that it included two morphological types. In this case males are distinguished by coloration and morphology, but females are not (Dijkstra & Matushkina 2009). A different example where two species were hidden under one name was shown in Palpopleura lucia and P. portia. Here two male morphs were considered to be ‘male polymorphism’ for centuries, until a difference in DNA sequences shed light on the erroneous species delimitations (Mitchell et al. 2006).
The study of Damm et al. (2010) shows the importance of a detailed analysis of species diversity, especially in tropical countries, where threats to conservation make this task more urgent.
This study also exemplifies the methodology proposed by DeSalle et al. (2005) and provides a good guidance when perplexing and unexpected results are found: rather than trying to find ad-hoc explanations for new, unforeseen results, use them to inspire new ideas. Then, test these ideas with new, independent evidence (Johnson 2002). If the results were spurious, the new evidence will show that. If not, a little step in the right direction has been made.