Research determining the limits and history of plant-herbivore associations can help us understand why phytophagous insects are so diverse. Many studies have tested for “host associated differentiation” of insect herbivores that feed on two or more host plants. However, few such studies have focused on apparent broad generalists that feed on many different plant species. Such apparent generalists might actually include genetically distinct clades feeding on specific host plant species or in different habitats. Historically, many leaf mining flies have been considered (on morphological grounds) to be generalists. One such leaf miner is the agromyzid fly Nemorimyza posticata, which has been reported from at least 20 species spanning the herbaceous plant tribe Astereae (Astereaceae). We used nuclear and mitochondrial genes to test for cryptic genetic differentiation in this species to ask whether any such differentiation is associated with host species, habitat type, or geography. We collected 867 larvae of N. posticata feeding on 15 host plant species in three very different habitat types (salt marshes, old fields and forests) of New Brunswick and Quebec. We sequenced ~1200 bp of CO1 and ~800bp of CAD, and estimated relationships among sequences and divergence among clades building a maximum likelihood tree.
Results/Conclusions
There is much molecular variation in the apparent generalist Nemorimyza posticata, suggesting a potential cryptic species complex, with evidence for both host- and habitat-associated differentiation. Average genetic distance is 5.3% within this species complex. There are three main groups initially separated by habitat. The genetic distance between the groups varies between 4.3 and 6.4%. One group is a true host generalist feeding on several host plants, but is specialized in habitat to open fields and forest edges. A second group is a forest specialist, which has further host associated genetic structuring. The third group is a salt marsh specialist. Nemorimyza posticata, once believed a single host and habitat generalist, is instead a clade of closely related, morphologically indistinct, but habitat- and host- associated flies. Our results highlight the importance of testing whether apparently generalist species are really generalists, and, if they prove to be complexes of specialists instead, determining the ecological axes along which differentiation has occurred. Neither a “generalist” nor a “one specialist per host” model fits N. posticata’s ecological genetics well; instead, this species shows surprising complexity in its history of diversification.