Geographic variation in the phenotype of an African horseshoe bat species, Rhinolophus damarensis, (Chiroptera: Rhinolophidae)

Abstract

Studies involving geographic variation in the phenotypes of bats help scientists to explain why these mammals are the most species rich mammalian order second only to rodents, with well more than 1 300 species occurring worldwide. Such species richness or high diversity is the manifestation of the generation of biodiversity through the splitting of lineages within bat species. A lineage of bat species can diversify into several lineages which then differentiate from each other in allopatry. Thus, the spatial separation of a lineage into several lineages could be attributed to geographical, ecological and environmental factors across the distributional range of the species. Similarly, vicariant events may also play a role in separating lineages within species. The Damara horseshoe bat species, Rhinolophus damarensis, is widely distributed but restricted to the western half of southern Africa, where it occurs across several major biomes. Formerly regarded as the subspecies, R. darlingi damarensis, it was elevated to full species status on the basis of genetic and phenotypic differences between it and R. darlingi darlingi. Rhinolophus damarensis is itself made up of two ecologically separated genetic lineages. A total of 106 individuals of R. damarensis were sampled from seven localities across its distributional range, with a view to determining and documenting the extent of geographic variation in body size, echolocation parameters, wing parameters, cranial shape and postcranial morphology of individuals from populations of R. damarensis across the distributional range of the species. Firstly, an investigation into geographic variation in resting echolocation frequency (RF) of the horseshoe bat species, R. damarensis was carried out in the western half of southern Africa (Chapter 2). Three hypotheses were tested. The first one, James’Rule (JR), states that individuals occurring in hot humid environments generally have smaller body sizes than conspecifics that occur in cooler, dryer environments, and the largest are expected to occur in cool, dry areas. On this basis and because of the known relationship between body size and RF, it was predicted that there should be a correlation between body size and climatic factors and between body size and RF. The second hypothesis was Isolation by Environment (IbE) mediated through sensory drive, which proposes that diversification of lineage may be driven by environmentally-mediated differences in sensory systems. Under this hypothesis, it was predicted that call frequency variation should be correlated with climatic variables. The third hypothesis was that Isolation by Distance (IbD) can influence phenotypic trait variation by restricting gene flow between populations. Under the Isolation by Distance (IbD) Hypothesis, it was predicted that call frequency variation should be partitioned in accordance with geographic distance between populations. To investigate the probability of the JR, IbE and IbD, the Akaike’s information criterion AICc candidate models were evaluated with different combinations of environmental (annual mean temperature and relative humidity), spatial (latitude and region) and biological (forearm as a proxy for body size) predictor variables to determine their influence on resting frequency (RF) across the distributional range of R. damarensis. Linear mixed effects models (LMEs) were employed to analyse the relationship between the response variable (RF) and the environmental, spatial and biological predictor variables. The influence of prey detection range and atmospheric attenuation was also investigated. The results showed no evidence for JR or for random genetic drift. Body size was neither correlated with RF nor environmental variables, suggesting that variation in RF was not the result of concomitant variation in body size as proposed by JR. Similarly, the Mantel test showed no IbD effect and there was therefore no evidence that genetic drift was responsible for the variation in RFs. In contrast, the LMEs showed that there was support for IbE in the form of an association between RF and region (in the form of the variable “Reg”) which was based on the two geographically separated genetic lineages. Furthermore, RF variation was also associated with the climatic variable AMT. The taxonomic status of R. damarensis was investigated using ecological traits and phenotypic characters including skulls, wings and echolocation (Chapter 3) and three dimensional (3D) scanned skulls and mandibles (Chapter 4). The main objective (Chapter 3 and Chapter 4) was to test whether previously reported genetic divergence between the two R. damarensis lineages was associated with phenotypic divergence. Morphometric and echolocation measurements were taken from hand held individual bats in the field, and skull measurements were taken from field collected voucher specimens as well as museum specimens. Discriminant Function Analyses (DFA) revealed that there was geographic variation among populations and lineages of R. damarensis. Multivariate Linear Regressions (MLV) and Linear models (LM) on the basal parts of bacula revealed significant differences between the southern and northern lineages of R. damarenis. The bacula of the two lineages of R. damarensis appear to have different shapes. Diversification through shape analyses (Chapter 4) was investigated using three dimensional (3D) geometric morphometric analyses based on X-ray microcomputed tomography (µCT) scanning of dried skulls and mandibles of R. damarensis. Procrustes Anova results of both mandibles and skulls indicated that there were no significant differences between sexes but that the shape of skulls and mandibles varied across different localities (Chapter 4). Canonical Variate Analysis (CVA) suggested that geographic variation in R. damarensis mandibles was based on the shape and thickness of the alveolar bone. Geographic variation in the skulls was based on changes in the rostrum, anterior medial swelling and brain case. Some populations had slightly deeper rostra, slightly larger anterior medial swellings and smaller braincases, whilst others had slightly shallower rostra, slightly smaller anterior medial swellings and larger braincases. The northern lineage was found to be separated from the southern lineage based on the changes in skull and mandible shape. Therefore, separation of lineages within R. damarensis (Chapter 4) could be associated with the foraging and feeding behaviour of the species under different ecological conditions due to ecological opportunity. Overall, differences in the RF were found to be associated with Isolation by Environment mediated through sensory drive and this has led to the formation of two regional (northern and southern) groupings in RF (Chapter 2). The two lineages were supported by both the phenotypic divergence (Chapter 3) and shape variation within R. damarensis skulls and mandibles (Chapter 4). Thus, phenotypic differences corresponded to genetic differences between the two lineages and provide support for IbE.