Taxonomy of Rhinolophus simplex 
11 
Table 3b 
Character 
Sex 
Main Effects 
Age 
Island 
Sex. Age 
Interactions 
Sex. Age. 
Island Island 
Sex. Age. 
Island 
SV 
0.176 
0.069 
28.275 
0.150 
0.520 
0.694 
2.084 
TV 
0.284 
2.670 
3.109 
0.413 
1.036 
1.862 
0.420 
EL 
0.098 
0.063 
1.798 
0.004 
5.505 
2.437 
4.712 
TIB 
0.005 
0.083 
41.150 
0.926 
0.281 
0.365 
1.472 
PES 
0.116 
0.040 
12.590 
0.904 
0.917 
0.912 
0.818 
FA 
0.543 
1.035 
24.474 
1.183 
0.838 
0.931 
3.275 
D2M 
0.532 
0.000 
17.119 
1.075 
0.543 
0.770 
1.490 
D3M 
0.072 
1.232 
14.872 
1.342 
0.547 
0.679 
0.249 
D3P1 
1.012 
0.124 
15.439 
0.000 
1.498 
0.608 
0.209 
D3P2 
0.799 
1.046 
17.561 
0.199 
0.880 
0.426 
0.002 
D4M 
0.402 
3.601 
33.937 
4.652 
2.621 
1.614 
1.607 
D4P1 
4.457 
0.012 
13.284 
1.016 
2.557 
0.866 
0.019 
D4P2 
0.041 
0.157 
17.845 
2.890 
0.379 
0.922 
5.557 
ALB 
0.000 
0.583 
2.459 
0.055 
1.251 
1.045 
0.018 
BSL 
0.094 
0.165 
4.530 
1.108 
0.073 
1.021 
1.099 
BSB 
0.932 
0.007 
0.699 
1.916 
0.533 
1.302 
5.935 
VSH 
0.115 
0.490 
8.392 
3.426 
1.103 
0.692 
2510 
VSB 
0.580 
5.218 
2.924 
0.205 
0.294 
1.351 
0.381 
DEGREES OF 
FREEDOM 
1,23 
1,23 
5,23 
1,23 
3,23 
4,23 
1,23 
incorrectly classified: Four R. simplex sp. nov. were 
classified as R. s. parvus and one R. s. simplex was 
classified as R. s. parvus. 
The plot of functions 1 and 2 most clearly 
separates the taxa (Figure 3a) and plots of other 
combinations of functions 1-3 do not further clarify 
graphically this separation. From Figure 3a, the R. 
borneensis and R. megaphyllus clusters are clearly 
separate on function 2 and these two species clearly 
cluster separately from the R. simplex subspecies 
on function 1. Further, the R. simplex subspecies 
clusters, which partially overlap, separate on 
function 1. The unallocated Rhinolophus s. keyensis 
grouped closely with R. s. simplex. 
The character loading most heavily (>0.5) on 
function 1 and which is presumed an important 
discriminant between R. borneensis, R. megaphyllus 
and R. simplex, and among the R. simplex 
subspecies, was M 1 length (Table 4a). The 
characters loading heavily (>0.5) on function 2 and 
presumed important in discriminating between R. 
borneensis and R. megaphyllus were supraorbital 
length, NIL, and zygomatic width, ZW (Table 4a). 
Externals. The DFA for the five taxa was first run 
using the reduced set of 17 characters and using 
island as the a priori grouping. When these islands 
were grouped to represent the five above taxa the 
configuration of the taxon clusters in discriminant 
function space was very similar to that produced 
above. However, because the number of 
individuals in some taxa was less than the number 
of characters measured (e.g. R. borneensis, 9; R. 
megaphyllus, 12) fewer characters were used in the 
analysis. Five the 17 characters used in the DFA as 
the a priori groups were selected (forearm length, 
FA; pes length, PES; vertical sella height VSH; digit 
4, phalanx 1 length, D4P2) for analysis because they 
provided values that minimise Wilk's Lambda. 
These five characters produced similar DFA plots 
to those from the 17 characters. Only the DFA 
