| Evolutionary ecology of hoverflies | Back : Forward |
My research programme in this field mostly concerns the evolutionary biology of the hoverflies (Diptera, Syrphidae), a very large group of flower-visiting flies whose diverse larval feeding strategies are particularly interesting from an evolutionary standpoint. Using comparative and experimental approaches, I concentrate on morphology, life-histories, feeding specialization and mimetic communities. Funded initially by a Leverhulme fellowship, I am still struggling to complete a monograph on the evolutionary biology of the Syrphidae. This summarises and reinterprets information from more than 10000 published works on the family, most of which I have scanned as pdfs. The databases that abstract the information from the literature will also be made available free to download. The basic information for all comparative approaches is a phylogeny of the group. With my colleague Dr Graham Rotheray of the National Museums of Scotland, some years ago we completed a very large phylogeny of most of the Palaearctic genera of Syrphidae using cladistic methods on larval characters (Rotheray & Gilbert 1989, 1999, 2005). The large scale of this work makes it possible to use the phylogeny to test many hypotheses about the evolution of morphological and ecological attributes: for example, whether particular life-history attributes promote the rate of speciation (Kazourakis et al 2001) or the multiple evolution of wasp mimicry (Leavey et al 2021). Molecular phylogenies of parts of the syrphid tree are now available, and we await the huge efforts of a worldwide consortium in sequencing multiple genes for nearly all 268 genera - the preliminary results of the estimated phylogeny (Pauli et al 2018) look very similar to our tree based on larval characters. I am interested in the evolution of feeding modes and in specialization. The Syrphidae are an excellent model system to use for this because of their diverse larval feeding ecologies: substantial numbers of species are phytophages, saprophages and predators, including economically important aphid predators. Research on the evolution of feeding specialization has found a relationship between prey preference by ovipositing females and the larval performance feeding on those prey, even in an extreme generalist (Sadeghi & Gilbert 1999, 2000a-c). Specialization may also evolve in response to mortality by natural enemies, the idea of ‘enemy-free space’. The phylogenies of syrphid hosts and diplazontine parasitoids match, and reveal the parallel evolution of feeding specialization in both taxa. A review of the interactions known to occur in the communities where aphidophagous syrphids live (Gilbert 2005a) shows that the community context is an important determinant of feeding specialisation. I am particularly interested in the evolution of the mimetic colour patterns of hoverflies, partly because aspects of hoverfly mimetic communities do not match predictions from mimicry theory: mimics are too common relative to their models, and there are also many imperfect mimics. We developed new approaches using the hoverfly phylogeny, image analysis, neural networks and operant conditioning methods, in collaboration with various researchers and students including Winand Dittrich, Peter McGregor, Patrick Green, Graham Holloway, Tom Reader, Dave Grewcock, Salma Azmeh, Chris Taylor and Tom Sherratt. These have shown that: a) birds categorise hoverfly mimetic patterns in a very similar way to humans (Dittrich et al 1993); b) image analysis allows a quantitative assessment of mimicry (Taylor et al 2013); c) current relative abundances are far from their natural values in the absence of human habitat interference (Azmeh et al 1998); d) ‘mimics’ may actually be aposematic rather than mimetic, advertising their unprofitability (via flight agility or possibly unpalatability); e) pattern variability implies weaker selection on better mimics (Holloway et al 2002). In a review I summarized possible explanations of imperfect mimicry in syrphids (Gilbert 2005b). Using a neural network approach, it is possible to dissect which components of the pattern are used by birds to classify mimics and nonmimics (see Bain et al 2007). Now with my colleague Tom Reader and student Chris Taylor (PhD 2011-15) we are testing the multiple-model hypothesis and taking a new look at the accuracy of mimetic resemblence via image comparisons (eg Taylor et al 2013). We have recently completed an assessment of the multiple evolution of wasp mimicry (Leavey et al 2021). I collaborate with a number of workers around the world on hoverfly biology, and organize an email discussion group for promoting international collaboration. With Ante Vujic I am coordinating the Red Listing of European syrphids in an IUCN project funded by the EU. See my related site: Selected publications: |