Acrocera orbiculus

Kehlmaier, Christian & Almeida, Jorge Mota, 2014, New host records for European Acroceridae (Diptera), with discussion of species limits of Acrocera orbiculus (Fabricius) based on DNA-barcoding, Zootaxa 3780 (1), pp. 135-152 : 140-144

Acrocera orbiculus —a species complex?

Both molecular analyses of the COI-dataset yielded similar results, though with differing degrees of resolution. The maximum likelihood analysis reveals two clades within A. orbiculus (fig. 4). Clade A (six specimens/ localities) reaches from Spain to Iran and has good statistical support values (BT: 99 %) (fig. 6, Appendix). Clade B (nine specimens/localities) is confined to mountainous regions in northern Italy, southern France and central Portugal and is weakly supported (BT: 66%). Branch support can be increased if CK684 from Portugal is excluded and assigned to a separate clade C (BT: 81%). The Bayesian analysis (fig. 5) shows six evolutionary lineages within Acrocera placed in a basal polytomy, four of those comprising A. orbiculus . The only Bayesian clade of A. orbiculus with reasonable support (PP: 84) corresponds to clade A from the ML analysis. However, clade B is not supported and broken up into three lineages.

Uncorrected genetic pairwise distances are summarised in table 1. The maximum intraspecific genetic variability of all A. orbiculus lies at 6.5% (CK565 and CK641), which is conspicuously large (see below). Within members of clade A, the maximum genetic distance is 4.0% (CK598 and CK641), and within members of clade B it is 4.4% (CK565 and CK684), respectively. The latter is reduced to 3.1% if CK684 is excluded. The minimum genetic distance between members of both clades lies at 3.6% (CK572 and CK609). Hence, there are at least members of clade A that are genetically closer to members of clade B than to other members of their own clade. When mapping and comparing the genetic variability within clade B (fig. 7), specimens sharing an almost identical barcode, (i.e., maximum genetic distance of 0.6%) and located in geographic proximity, are considered to belong to the same population. Thus, the nine specimens sampled represent five populations that are separated by a 2% or greater p-distance, pointing towards a low degree of gene flow amongst them.

From a morphological perspective, no clear differences could be observed between the two clades, other than males and females of clade B tend to have slightly darker leg coloration compared to members of clade A (table 2), the main feature used by Chvála (1980) to distinguish between the two forms. Therefore, it is assumed that clade B corresponds to “sp. (globula PANZ.)” of Chvála (1980), with the exception of specimen CK599 of clade B which is slightly paler than CK641 of clade A. No structural differences were observed in the male genitalia. Nartshuk (1982) highlighted the high variability of A. orbiculus in terms of body size and coloration, by distinguishing ten different colour forms of abdominal tergites 2–4 (4 in males and 6 in females) based on 22 specimens (8♂ 14♀) from the Saint Petersburg area. Our findings include a series of 12 ♀ from the Ore Mountains (south-eastern Germany) and 79 ♀ from the Parc Naturel Régional des Pyrénées Ariégeoises (southern France) (see table 1 for label and deposition data) that vary considerably in body size and abdominal coloration. This high degree of phenotypic plasticity is illustrated in fig. 8 for voucher specimens CK680 (2.5 mm body length), showing small yellow markings on tergite 3 and 4 only, and CK681 (4.2 mm body length), with small yellow markings on tergite 2 and predominantly yellow tergites 3 & 4. Both specimens were collected at the same locality and share an identical DNA-barcode.

Table 1. Uncorrected genetic distances of COI dataset.

Clade A Clade B

Coloration of legs Yellow except distitarsi dark brown. [n=3] Yellow except fore and mid coxa on anterior surface

in males and fore femora especially in basal half light brown;

distitarsi dark brown. [n=1]

Coloration of legs Yellow except coxae light brown on anterior Yellow except coxae light to mid brown on anterior

in females surface; femora yellow to light brown except at base surface; femora light to mid brown except at base and apex; tibiae yellow to light brown medially; and apex; tibiae yellow to mid brown medially; tarsal segments yellow to light brown; distitarsi tarsal segments yellow to mid brown; distitarsi dark dark brown. [n=4] brown. [n=5]

Based on the molecular and morphological results presented for A. orbiculus , Chvála’s (1980) assumption of an unnamed taxon “sp. (globula PANZ.)” cannot be supported. Even though the maximum likelihood analysis shows two clades, the low statistical support and the lack of resolution in the Bayesian reconstruction do not support the presence of multiple cryptic species within A. orbiculus . Furthermore, the absence of a distinct “barcoding gap” between the two clades points towards a genetic continuum between clade A and B. On the other hand, such an overlap has repeatedly been observed (e.g. Kehlmaier & Assmann 2008; Smith et al. 2006) to the point that morphologically distinct species can share an identical DNA-barcode (e.g. Skevington et al. 2007). In the case of A. orbiculus , no consistent morphological differences between the two clades were detected in either genital morphology or body coloration. Whereas body size is positively correlated with host size (i.e., food supply) in at least some endoparasitic Diptera like Pipunculidae (Kehlmaier, pers. observ.), body coloration might be influenced by thermal conditions during pupal development, as documented for some Syrphidae ( Dušek & Láska 1974, Ottenheim et al. 1995). Also, quantity and quality of food might vary to such a degree that it directly effects the forming of body pigmentation in the cuticle, assuming that the development of pale markings requires more energy than dark markings (see Gilbert 2005 for a summary of this issue). The pattern of abdominal variability according to body size in A. orbiculus has also been observed in the syrphid Pelecocera tricincta Meigen whose larval feeding mode is still unknown ( Kehlmaier 2002).

In recent years, there has been some dispute as to whether a fixed threshold of genetic divergence of the COI gene can be applied for species delimitation, i.e., Hebert et al. (2003) suggesting 3% for Lepidoptera and Hebert et al. (2004) postulating a ‘10× divergence criterion’. If applied here, a 3% threshold would lead to an unjustifiable/ unnecessary splitting of A. orbiculus , considering that the specimens barcoded have an average p-distance of 4.0%. The ‘10× divergence criterion’ would ultimately lead to the lumping of probably the entire family, as A. orbiculus has a maximum intraspecific genetic variability of 6.5%, resulting in a species threshold of 65%. The development of the observed divergent evolutionary lineages within clade B, being separated by a genetic variability of at least 2%, might be explained by the limited ability of A. orbiculus for active dispersal, resulting in a low exchange of genetic material between populations. Acroceridae are generally regarded as poor flyers, being active only on warm sunny days, and resting on vegetation in morning and evening hours and on colder and cloudy days ( Nartshuk 1997). Taking into consideration that the life-span of adult Acroceridae is confined to 3–30 days ( Schlinger 1987), most dispersal probably takes place during the larval development within the araneomorph host, which lasts 6–10 months ( Schlinger 1987).

Adding up the evidence at hand (mtDNA and morphology), it is our understanding that A. orbiculus represents a single biological species with a slight degree of sexual dimorphism and a high degree of phenotypic plasticity regarding body size and coloration, instead of many separate, closely allied taxa that cannot be reliably differentiated on a morphological basis. An expanded sampling covering the entire Holarctic distribution and a more detailed morphological and molecular approach would be desirable to further investigate this question.