180 
D.J. Kitchener, Y. Konishi, A. Suyanto 
dimensions. H. bicolor is also reported from nearby 
Flores island by Hill and Rozendaal (1989). 
More recently, collections of H. bicolor have been 
made from the following islands in the Lesser 
Sunda group: Selaru (Tanimbar Is), Timor, Roti, 
Savu, Sumba - see Figure 1 map. These specimens 
exhibit some morphological differences from the 
described forms of H. bicolor. This paper reports on 
an examination of this morphological variation, 
particularly that of taxonomic significance. 
MATERIALS AND METHODS 
76 adult specimens from six islands in the Lesser 
Sunda group and from Java and Peninsula 
Malaysia were measured. 
Adult condition was judged by an absence of 
swelling of the epiphyseal joints of the wing 
metacarpals and phalanges. 
The description of the colour of the pelage and 
patagia, where they follow Smithe (1975), are 
capitalised. 
Measurements were as follows: GSL, greatest 
skull length; CB, braincase breadth; ZW, zygomatic 
width; MW, mastoid width; RL, rostrum length; 
LIB, least interorbital breadth; TBL, tympanic bulla 
length; ILD, interlachrymal distance; CL, cochlea 
length; CW, intercochlear distance; C'B, upper 
canine breadth at base; P 3 B, second upper premolar 
cusp breadth; M 3 B, last upper molar cusp breadth; 
M'M 3 L, upper molar row cusp length; C‘M 3 L, 
upper canine to last upper molar cusp length; 
M 3 M 3 W, width across M 3 to M 3 from outer alveoli 
edge; IjM 3 L, lower tooth row length; DL, dentary 
length; SVL, snout to vent length; TVL, tail to vent 
length; EL, ear length; TIL, tibia length; FA, 
forearm length; D2-5 MC, digits 2 to 5 metacarpal 
length; D3-5 PI, digits 3 to 5 phalanx 1 lengths; D3- 
5 P2, digits 3 to 5 phalanx 2 lengths; ANB, anterior 
noseleaf maximum breadth; PNB, posterior 
noseleaf maximum breadth. 
Sexual dimorphism of all cranial, dentary, dental 
and body characters that were measured was 
investigated by standard multiple regression 
analyses (where all effects were assessed 
simultaneously of each character on sex and 
island). Graphical inspection of raw data using 
plots from regression analyses gave no indication 
of heteroscedasticity. 
Stepwise Canonical Variate (Discriminant 
Function) analyses were run for all cranial, 
dentary, dental and external body characters 
measured for both sexes combined. 
RESULTS 
Multiple regressions 
Localities included in these analyses were from 
Peninsula Malaysia, Java, Selaru, Timor, Roti, Savu 
and Sumba islands. 
The analyses showed that there were no 
significant interactions between locality and sex for 
