Reliable identification of Aphidiinae species (Braconidae) is a prerequisite for conducting studies on aphid-parasitoid interactions at the community level. However, morphological identification of Aphidiinae species remains problematic even for specialists and is almost impossible with larval stages. Here, we compared the efficiency of two molecular markers [mitochondrial cytochrome c oxydase I (COI) and nuclear long wavelength rhodopsin (LWRh)] that could be used to accurately identify about 50 species of Aphidiinae that commonly occur in aphid-parasitoid networks in northwestern Europe. We first identified species on a morphological basis and then assessed the consistency of genetic and morphological data. Probably because of mitochondrial introgression, Aphidius ervi and A. microlophii were indistinguishable on the basis of their COI sequences, whereas LWRh sequences discriminated these species. Conversely, because of its lower variability, LWRh failed to discriminate two pairs of species (Aphidius aquilus, Aphidius salicis, Lysiphlebus confusus and Lysiphlebus fabarum). Our study showed that no unique locus but a combination of two genes should be used to accurately identify members of Aphidiinae.

Molecular methods are increasingly used to detect and identify parasites in their hosts. However, existing methods are generally not appropriate for studying complex host–parasite interactions because they require prior knowledge of species composition. DNA barcoding is a molecular method that allows identifying species using DNA sequences as an identification key. We used DNA amplification with primers common to aphid parasitoids and sequencing of the amplified fragment to detect and identify parasitoids in their hosts, without prior knowledge on the species potentially present. To implement this approach, we developed a method based on 16S rRNA mitochondrial gene and LWRh nuclear gene. First, we designed two primer pairs specific to Aphidiinae (Hymenoptera), the main group of aphid parasitoids. Second, we tested whether the amplified regions could correctly identify Aphidiinae species and found that 61 species were accurately identified of 75 tested. We then determined the ability of each primer pair to detect immature parasitoids inside their aphid host. Detection was earlier for 16S than for LWRh, with parasitoids detected, respectively, 24 and 48 h after egg injection. Finally, we applied this method to assess parasitism rate in field populations of several aphid species. The interest of this tool for analysing aphid‐parasitoid food webs is discussed.

No abstract