Determinants of Butterfly Species Diversity : Plant Diversity, Foliage Height Diversity and Resource Richness Across Vegetation Types

Abstract

MacArthur and MacArthur's (1961) theory of foliage height diversity, which was originally discussed as a determinant of bird species diversity, has been a strong concept describing the possible role of vegetation structure in deciding resource use, habitat selection and diversity of organisms. They measured foliage profile, and calculated foliage height diversity using information theory. Their results showed increase in foliage height diversity from simple to complex vegetation types, and bird species diversity was positively correlated with it. They then proposed an evolutionary mechanism to explain this correlation. A continuous flow of studies on vegetation structure and species diversity or habitat use of organisms followed after MacArthur and MacArthur (1961), but mostly they focused on birds. Subsequently, vegetation parameters other than foliage height diversity, such as horizontal vegetation heterogeneity or patchiness, also proved equally useful or better predictors of bird species diversity. Moreover, some studies revealed that foliage height diversity and bird species diversity were not correlated in a few cases. This brought forth a question whether or not foliage height diversity is a strong determinant of bird species diversity, or a frequent correlate of some other factor that has a strong influence on species diversity. To find out why in some cases MacArthur and MacArthur's (1961) theory did notexplain species diversity, I identified following assumptions in their theory, which were unstated in their paper: i)resources for a species are evenly dispersed within a horizontal vegetation layer, ii)resources are unevenly dispersed across the horizontal layers, iii)even for generalist species, the cost of switching over between layers is high, which would preclude switching over between layers, and iv)if the first three assumptions are true, then each horizontal layer of vegetation would have a different set of species, each layer adding a similar magnitude of diversity in more complex habitats. If these assumptions do not hold true for real biological communities, one would expect that foliage height diversity would not be a good predictor of species diversity in all situations. To test MacArthurs' assumptions and explore patterns of butterfly diversity across vegetation types, I studied butterflies at the Anamalais, southern Western Ghats, in southern India. I sampled foliage height diversity in vegetation plots and butterflies on count lines as well as all-aut-walks (opportunistic sampling), recorded body measurements of the butterflies, floral parameters of the plants on which the butterflies fed, and compiled information on larval and adult feeding plant resources. Using these data, I tested three possible degeminates of butterfly species diversity namely, foliage height diversity, 2.plant species diversity, and 3.resource richness in different vegetation types. This was probably the first attempt to analyse growth form-wise resource richness and utilisation by butterflies across vegetation types, in the light of foliage height diversity theory. Although the correlations between butterfly species diversity, plant species diversity and foliage height diversity were statistically significant, the butterfly species diversity was not perfectly correlated with these two determinants. Butterfly species diversity increased from grassland to deciduous forest, through shrub savannah and teak plantation, but was less in the mid-elevation evergreen forest as compared to the deciduous forest. Plant species diversity and foliage height diversity, on the other hand, increased from the grassland to the evergreen forest, with the deciduous forest falling before the evergreen forest. The resources for butterflies were not distributed in the vegetation types as assumed by MacArthur, and butterflies were not observed following the pattern of habitat or resource utilisation as predicted by the theory. As a result, foliage height and plant species diversity did not satisfactorily explain the butterfly species diversity. However, larval host plant and total plant resource richness correlated significantly and appropriately with the butterfly species diversity. Therefore resource richness seems to be an appropriate and better predictor, or at least a correlate, of butterfly species diversity in situations when other predictors may not be correlated with it. This study suggests that butterfly species diversity is dependent more on the resource diversity. However, foliage height diversity and plant species diversity need not be satisfactory surrogates for resource diversity in all situations. Therefore butterfly species diversity may vary independently of plant species and foliage height diversity. Another important result of this study was that even for butterflies, which are directly dependent on plants, the resource richness is not necessarily correlated with the plant species diversity. It means that butterflies selectively use certain plant resources more than other plant resources. I speculate that it would be underlying utilizable chemical diversity and "sociability" of plant resources used that would ultimately decide butterfly species diversity within a vegetation type. I define a "sociable host plant" as the one that supports many insect "guests". The butterfly assemblages using plant resources with maximum utilizable chemical diversity and sociability would be most diverse in the vegetation type supporting these sociable host plants. However, sociability of the utilizable plant species, and resultant butterfly diversity, may be independent of total plant species diversity or foliage height diversity in the vegetation type.