Variation in nocturnal distributions and relative abundances
Yaeyama fruit bats are solitary or form mostly very small groups while foraging, but at a larger scale, their nocturnal spatial distribution is slightly clumped. Among habitats, the mean abundance and density were lower in cultivated areas than in villages and inland forests. This explains the across-island variation in bat distribution, where higher abundances were found in western than in eastern village sites, and supported our predictions. Villagers appeared tolerant to flying foxes, including their excretion and sounds (Y.F. Lee, unpubl. data; M. Matsumoto, pers. comm.). On average, however, eastern villages devoted a four-fold larger area of land to agriculture/husbandry activities, areas presumably with a higher rate of disturbance and intensity of human activities. Eastern sites also had lower fruit-tree density and heterogeneity range in tree composition. On Daito Island, where a high intensity of cultivation has made mangrove forests one of the few remaining undisturbed habitats, fruit bats are consistently found roosting in the border of mangroves (Izawa et al., 2001).
The distribution of foraging bats was correlated with land use patterns, habitats, as well as with food supplies, which is consistent with that reported for grey-headed flying foxes P. poliocephalus in Australia (McDonald-Madden et al., 2005). Bat density was correlated moderately with heterogeneity of tree composition, strongly with the density of major fruiting trees, and supported our predictions. Shifts in foraging sites by flying foxes for better food availability have been frequently documented and may occur between types of forests in an area (Banack, 2002), across various distances within a short period of time (Tidemann and Nelson, 2004; McConkey and Drake, 2007), or during migration (Hall and Richards, 2000). Nakamoto et al. (2007a) reported Orii’s fruit bats using certain fruiting trees intensively in an urban area. On Iriomote, bats in larger groups were observed frequenting some village sites, such as Hoshidate and Sonai, where large fruiting trees and ample food supplies occurred.
We observed variation between habitats in the height and DFT of perching bats. Heights of bats were lower in cultivated areas, and were in correspondence with lower canopy and shrub heights, and less canopy coverage, in this habitat. The DFT of perching bats was probably, at least in part, due to environmental features that restrict accessibility, such as tidal patterns in mangroves, and paths tended to be narrower and tree boundaries closer to trails in inland and coastal forests. An alternative but not mutually exclusive explanation is that the height and DFT may reflect a perceived or actual risk of predation. Fruit bats may fall prey to snakes, raptors, and other mammals (Pierson and Rainey, 1992; Klose et al., 2009). On Iriomote these at least include Beauty Rat Snakes Elaphe taeniura (Cope, 1861), Japanese Lesser Sparrowhawk Accipiter gularis (Temminck and Schlegel, 1844), feral cats Felis catus Linnaeus, 1758, and Iriomote Cats Prionailurus iriomotensis (Imaizumi, 1967) (Watanabe et al., 2003). Lower perching heights may reduce the risk from aerial raptors, particularly at trees of less canopy coverage, and in more open cultivated areas; lower shrub heights may make terrestrial predators more difficult to conceal and less possible to approach a perching tree from surroundings. The latter is also supported by the negative correlation between shrub height and bat density. At greater distances from paths and often depending on tides, bats in mangroves may be free of predation pressure from most terrestrial predators. This is supported by the fact that with a generally higher mean upper-canopy layer, perching heights of bats in mangroves were among the lowest, where terrestrial predators are most restricted.