Taxonomy of Rhinolophus simplex 
9 
Table 2b 
Character 
Sex 
Main Effects 
Age 
Taxon 
Sex. 
Age 
Interactions 
Sex. Age. 
Taxon Taxon 
Sex. Age. 
Taxon 
SV 
15.293 
7.898 
40.602 
7.900 
5.050 
2.188 
2.307 
TV 
0.257 
2.221 
5.135 
0.019 
0.514 
1.982 
0.624 
EL 
0.083 
0.003 
14.580 
0.001 
0.309 
0.412 
0.067 
TIB 
0.662 
0.290 
48.605 
0.269 
1.402 
1.398 
1.326 
PES 
0.847 
0.440 
*** 
59.904 
0.005 
0.505 
0.573 
1.163 
FA 
0.206 
0.173 
77.969 
3.968 
1.403 
0.609 
0.960 
D2M 
0.381 
0.137 
47.522 
0.290 
1.241 
0.486 
1.121 
D3M 
0.002 
0.084 
55.055 
1.323 
1.162 
0.927 
1.142 
D3P1 
0.348 
3.552 
41.778 
0.884 
1.388 
1.112 
1.485 
D3P2 
0.468 
0.733 
34.810 
0.001 
1.831 
1.926 
0.093 
D4M 
0.751 
0.737 
78.263 
1.460 
2.230 
1.571 
1.380 
D4P1 
1.229 
0.055 
52.247 
1.843 
2.900 
0.892 
2.660 
D4P2 
2.259 
1.882 
27.121 
0.711 
1.768 
1.225 
1.513 
ALB 
2.217 
2.987 
15.412 
0.134 
0.346 
1.688 
0.112 
BSL 
0.601 
0.567 
2.120 
0.057 
0.487 
0.562 
0.607 
BSB 
2.481 
0.165 
11.389 
0.142 
1.086 
1.312 
0.834 
VSH 
1.135 
0.002 
17.864 
2.984 
2.412 
1.344 
1.169 
VSB 
0.062 
0.311 
5.624 
1.748 
0.363 
0.319 
0.282 
degrees of 
FREEDOM 
1,59 
1,59 
4,59 
1,59 
4,59 
4,59 
2.59 
borneensis importunus, R. simplex simplex, R. s. parvus 
and R. simplex subsp. nov.). Rhinolophus s. keyensis 
was included in this analyses but unallocated 
because of its small sample size (N = 2). Both sex 
and age groups were combined. However, this 
analysis was carried out only after deleting nine 
skull characters shown in the multiple regression 
analysis to be significantly (P<0.01) influenced by 
sex or age (GSL, BB, RH, RL, CCW, M'M’W, 
C'M’L, I‘M~’L, DL, see Table 2a). Also MSF, and 
C'C'B were deleted because too many individuals 
had these values missing. The skull analyses used 
was with this reduced set of 21 characters. 
Similarly, the DFA of external characters was run 
following deletion of snout to vent length, SVL, 
because there were significant (PcO.Ol) interactions 
between sex and age for this character (Table 3b). 
Skulls. The DFA for the five taxa was first run 
using the reduced set of 21 characters and using 
islands as the a priori grouping. When these islands 
were grouped to represent the five above taxa the 
configuration of the taxon clusters in discriminate 
function space was very similar. However, because 
the number of individuals in some taxa was less 
than the number of characters measured (R. 
borneensis, 8; R. megaphyllus, 15) a reduced set of 
characters was used in the analysis. Of the 21 
characters used in the DFA of the five taxa as the a 
priori groups, five were chosen because they 
provided values that minimise Wilk's Lambda. The 
plots of the discriminant functions 1 to 3, based on 
the reduced set of five characters (mastoid width, 
MW, supraorbital length, NIL; M 1 length, M’L; 
zygomatic width, ZW; and sphenorbital fissure 
width, SFB) produced very similar plots to the 
above analyses, and so only these are presented 
and discussed below. 
The DFA produced three significant canonical 
functions. These three functions combined 
explained 99.9 percent of the variance (Table 4a) 
with function 1, 86.9 percent; function 2, 10.2 
percent and function 3, 2.8 percent. A total of 95.3 
percent of individuals were correctly classified to 
their appropriate taxon. Only five individuals were 
