Using geometric wing morphometrics to distinguish Aedes japonicus japonicus and Aedes koreicus

Background: Aedes japonicus japonicus (Theobald, 1901) and Aedes koreicus
(Edwards, 1917) have rapidly spread in Europe over the last decades. Both
species are very closely related and occur in sympatry. Females are
difficult to distinguish, and no distinctive morphological characters are
known for males. However, accurate species discrimination is important as
both species may differ in their vectorial capacity and spreading
behaviour. In this study, we assessed the potential of geometric wing
morphometrics as an alternative to distinguish the two species. 
Methods: A total of 147 Ae. japonicus specimens (77 females and 70 males)
and 124 Ae. koreicus specimens (67 females and 57 males) were collected in
South-West Germany. The left wing of each specimen was removed, mounted
and photographed. The coordinates of 18 landmarks on the vein crosses were
digitalised by a single observer. The resulting two-dimensional dataset
was used to analyse the differences in the wing size (i.e., centroid size)
and wing shape between Ae. japonicus and Ae. koreicus by means of
geometric morphometrics. To analyse the reproducibility of the analysis,
the landmark collection was repeated for 20 specimens per sex and species
by two additional observers. Results: The wing size in female Ae. koreicus
was significantly greater than in Ae. japonicus but did not differ
significantly for males. However, the strong overlap in wing size for the
females would not allow for discriminating the two species. In contrast,
the wing shape clustered species-specific and a leave-one-out validation
resulted in a reclassification accuracy of 95% for the females and 91% for
the males. The data collected by different observers resulted in a similar
accuracy, indicating a low observer bias for the landmark
collection.  Conclusions: Geometric wing morphometrics provide a
reliable and robust tool to distinguish female and male specimens of Ae.
japonicus and Ae. koreicus.