The biodiversity of Diptera in Old World rain forest surveys: a comparative faunistic analysis

Aim Identify the taxonomic patterns and the relative importance of particular families of Diptera sampled in comparative biodiversity surveys carried out at seven rain forest locations. We test and quantify the contention that different trapping methods routinely target different families. We identify the south-north (and upland/lowland) patterns and generate a set of hypotheses concerning mechanisms underlying these patterns. Location Australia and Papua New Guinea. Methods A total of 28,647 Diptera collected using canopy knockdown, yellow pan (water) traps and Malaise traps have been sorted to 56 families following these surveys. Comparative analyses across sites from Lamington National Park in south-east Queensland, Australia to the Kau Wildlife area in Madang Province, Papua New Guinea, of the dipteran assemblages, and separately, of the 14 families which collectively made up 95.8% of the sample, are presented. Results Ordination by multi-dimensional scaling and analyses of variances showed that the three methods complemented each other in terms of target families and, together, sampled a large proportion of the expected fauna of these sites. Ordinations on a method-by-method basis permitted the identification of groups of sites and analyses of variance indicated which taxa differed significantly across these groups. Main conclusions Recurrent patterns and associated hypotheses about their generation emerge from the data. These mirror floristic differences and reflect the biogeographic history of the sites since the Miocene. Clear linkages between the lowland faunas of Papua New Guinea and northern Australia are evident and are reflected in the abundances of the Dolichopodidae, Empididae, Muscidae and Tipulidae (other groupings underlined the essential difference of the New Guinean fauna which had characteristic proportions of Cecidomyiidae, Chironomidae, Dolichopodidae, Phoridae and Psychodidae). A subtropical grouping of families was evident comprising, inter alia, Chloropidae, Mycetophilidae, Drosophilidae and Phoridae which was frequently linked with the higher elevation tropical fauna at Robson's Creek, Atherton Tablelands. The long isolated, high elevation, rain-forested massif at Eungella, central Queensland often emerged as a unique entity in the analyses, characterized by the high numbers of and proportions of Chironomidae, Psychodidae, Tipulidae and Empididae. This study supports the case for the wider use of Diptera in biodiversity analyses, complementing extensive earlier analyses which have used, predominantly, large coleopteran assemblages. The results indicate the potential power of family-level analyses at large geographical scales and contribute to the ongoing debate on 'taxonomic sufficiency'.