From host plants to host ants: Evolution of phyto-predation in butterflies

The aim of this project has been to study the evolution of ant association, and especially parasitic associations, in butterflies using a phylogenetic approach. Many butterflies (especially in the families Lycaenidae and Riodinidae) are associated with ants in some way. These associations are mostly mutualistic where the caterpillars produce a nutrient rich secretion from special organs on their back, which ants like to eat, and the ants protect the caterpillars against parasites and predators. In a few cases, mostly in single species within otherwise fully plant feeding genera, this mutualistic association has turned into parasitism. In these species the caterpillars either eat the ant brood directly or are being fed mouth to mouth by the ants. The main group of study for this project is one of the few large radiations of ant eating butterflies. The genus Lepidochrysops occurs in Africa and many species are very rare since they depend on both a specific plant and ant species for larval food, and can only survive in places where both of these are found. I first wanted to produce a molecular phylogeny of the genus and closest relatives to get information on the relationships between species and the age of the species radiation. Obtaining information about species limits and relationships between the species is important for their conservation. If you don't know what species occurs where it is difficult to make conservation plans for the species. To generate this phylogeny I used a phylogenomic approach to sequence around 600 genes for 220 specimens. I optimized the method to be able to use butterfly specimens in museum collections in addition to fresh material. I spent 10 months in the field collecting specimens in South Africa, Zambia, Malawi, Kenya and Cameroon and utilized museum material from multiple museums and private collections around the world. Sequencing will happen shortly and the two main papers should be published towards the end of 2015. In part two of the project I inferred the first higher-level molecular phylogeny of the butterfly family Riodinidae, where ant associations are present in at least tree clades. The classification of the family has been in a state of flux for decades and many genera are still not placed in a tribe.I used a traditional Sanger sequencing approach and sequenced 6 genes for 200 specimens covering all subfamilies, tribes and subtribes of the family to stabilize the classification and study the evolution of ant association and the biogeography of the family. The paper will be submitted by the end of this year. The third group of study was the mainly African subfamily Aphnaeinae within the family Lycaenidae. The caterpillars in this group are mainly plant feeding and mutualistically associated with ants, but in nine of the 17 genera one species has turned parasitic. We inferred a phylogeny of the group and studied the host ant and host plant associations in the genus. The paper has been published (online early) in the journal Systematic Entomology. In the last project I look at ant association on the gene level and we reared the mutualsitic butterfly Jalmenus evagoras (Theclinae) with and without ants in the laboratory and will look for candidate genes involved in ant association. Currently extraction of RNA and RNA library prep are being done. One paper will be published probably next year. My work was presented at multiple international and national conferences, and I was invited to give seminars in Australia, Sweden, USA and Germany. Additionally I spent time in the field and in the class room with school children in Africa and the US to teach them about butterflies, ants, plants and the environment.

This research will study life history evolution of African butterflies in the genus Lepidochrysops. Unlike 99% of Lepidoptera (moths and butterflies), the species in this genus are phyto-predaceous: caterpillars initially feed on flowers, but eventually d rop to the ground where they are carried by ants into the brood chamber of the nest. Here they either eat the brood directly, or are fed mouth-to-mouth like cuckoos for the remainder of their development. The some 120 species of Lepidochrysops are all tho ught to be phyto-predaceous. Preliminary results using stable isotope analyses to measure trophic status suggest that this may not be true, and we propose to expand our sampling to analyze feeding behavior throughout the genus. We will investigate the evo lution of phyto-predation in a phylogenetic framework by inferring the evolutionary history of all the main lineages using molecular characters, and then applying comparative methods to assess life history evolution and diversification patterns in the gro up. In making collections, we will also assess distribution patterns of different taxa, and use this information to examine the effects of anthropogenic and climate change. The life histories of African Lepidochrysops are very similar to those of Palearct ic Maculinea species, many of which are the focus of considerable conservation attention. Probably as a consequence of their complex life histories, the butterflies are amongst the first to be threatened when habitat is destroyed as a result of human inte rference; one species of Lepidochrysops is already recorded as extinct. However, Lepidochrysops species parasitize ants in the subfamily Formicinae, whereas Maculinea species parasitize Myrmicinae, and this difference is likely to shape their population d ynamics. Our results will therefore be of great interest in a comparative framework, and will assist in establishing guidelines for conservation and management of these specialized insects, their hosts and habitats.