Why are There so Many Kinds of Passerine Birds? Because They Are Small. A Reply to Raikow

Abstract

In a recent note Raikow (1986) addressed cause of passerine diversity from a cladistic perspective. He noted that 5,274 of 9,021 extant birds are placed in Order Passeriformes, whereas remaining 3,747 species are placed in remaining nearly two dozen orders. The aim of Raikow's paper was to consider whether diversity of Passeriformes can be attributed to existence of adaptations, uniquely possessed by Passeriformes as a monophyletic group. After describing few synapomorphic characters by which Passeriformes are distinguished from other birds (the unusual insertion of a forelimb muscle, spermatozoan morphology, arrangement of bellies of a thigh muscle, structure of foot, and aegithognathous condition of palate), he states that the synapomorphies of Passeriformes, whether considered individually or in a broader functional context, fail to provide any convincing key innovations that might explain evolutionary success of group (p. 257). He concludes that notion of adaptation cannot be used to explain diversity of passerines, and that in fact it may be impossible to ever identify key adaptations in a rigorous, scientific fashion. Raikow briefly alluded to potential significance of body size in passerine evolution by pointing out that their small size makes them well adapted to feeding on arthropods and flowering plants; however, he did not explicitly consider potential causal relationship between size and diversity. Although relative sizes of organisms are not traditionally used in cladistic analysis, size is nonetheless an important factor to consider in analyses of taxonomic diversity. Van Valen (1973) found that taxa comprised of small organisms tend to have more species than taxa comprised of large organisms. He documented a negative relationship between mean organism size and species diversity for angiosperms, birds, and mammals. The pattern in mammals is illustrative: of 4,004 mammal species of world listed by Corbet and Hill (1980), 1,591 (40%) are in Order Rodentia and 950 (21%) in Order Chiroptera; species in these orders are generally among smallest of mammals. This negative relationship has been repeatedly observed in many taxonomic groups (Hutchinson and MacArthur, 1959; Lindsey, 1966; Fleming, 1973; Stanley, 1973; May, 1986). Because Van Valen did not statistically evaluate this relationship within birds, we recorded coordinates of each observation from Van Valen's bird scatter plot (his fig. 6) for analysis. In Van Valen's data set, values of mean weight and numbers of species were log transformed. A regression of these data showed a significant negative relationship between numbers of species in bird families and mean body weight of species within these families (n = 102, m = -0.23, r = 0.10, P = 0.001). Such analyses are problematic because of non-independence of data points drawn from a hierarchically structured system (Felsenstein, 1985). In addition, regression of these data is not entirely valid statistically, because mean family weights and numbers of species per family are bimodally distributed (n = 103: body weights: D = 0.13, P < 0.01); number of species: D = 0.10, P = 0.014; D = Kolmogorov statistic for normality). This bimodal distribution is due largely to significantly lower mean weights of passerine families (t = 9.54, P < 0.0001, df = 93.5),