Taxonomy of Rhinolophus simplex 
13 
Table 4 Canonical variate function coefficients for the five taxa: Rhinolophus megaphyllus, R. simplex simplex, R. s 
parvus, R. s. subsp. nov. and R. borneensis parvus. R. s. keyensis unallocated. Standardised values, followed by 
(in brackets) unstandardised values, (a) skull and dental; (b) external characters. 
Table 4a 
Character 
Function 1 
Function 2 
Function 3 
MW 
0.3517 
(1.7477) 
0.4891 
(2.4302) 
0.6221 
(3.0911) 
NIL 
0.2224 
(0.9304) 
0.7231 
(3.0253) 
-0.7308 
(-3.0575) 
M'L 
0.5114 
(8.5400) 
-0.4025 
(-6.7206) 
-0.1993 
(-3.3275) 
ZW 
0.3518 
(1.3538) 
-0.8787 
(-3.3811) 
-0.4380 
(-1.6854) 
SFB 
0.1374 
(0.9804) 
0.4699 
(3.3522) 
0.7509 
(5.3568) 
CONSTANT 
-45.9212 
-6.4634 
-6.3208 
VARIATION 
86.9 
10.2 
2.8 
EXPLAINED (%) 
Table 4b 
Character 
Function 1 
Function 2 
Function 3 
FA 
0.6108 
(0.5706) 
-0.1597 
(-0.1492) 
-0.1565 
(-0.1462) 
PES 
0.2590 
(0.7415) 
1.0000 
(2.8640) 
-0.4652 
(-1.3324) 
VSH 
0.3455 
(1.2204) 
-0.2507 
(-0.8854) 
0.2965 
(1.0471) 
D4P2 
-0.1158 
(-0.1892) 
0.2766 
(0.4519) 
1.1051 
(1.8054) 
D4M 
0.3276 
(0.3474) 
-0.6520 
(-0.6915) 
-0.1803 
(-0.1912) 
CONSTANT 
-42.6952 
3.4166 
-2.7095 
VARIATION 
77.3 
16.6 
4.1 
EXPLAINED (%) 
based on this reduced set of five characters are 
presented and discussed below. 
The DFA produced four significant canonical 
functions. These functions combined explained 100 
percent of the variance (Table 4b) with function 1, 
77.3 percent; function 2, 16.6 percent; function 3, 
4.1 percent and function 4, 2.0 percent. A total of 
95.6 percent of individuals was correctly classified 
to their appropriate taxon. Four specimens were 
incorrectly classified. One R.s. subsp. nov. was 
classified as R.s. parvus, one R.s. parvus was 
classified as R.s. subsp. nov. and two R.s. simplex 
were classified as R.s. parvus. 
The plot of functions 1 and 2 and functions 1 and 
3 most clearly separates the taxa (Figure 3b) and 
plots of other combinations of functions 1-4 does 
not further clarify graphically this separation. From 
Figure 3b, R. bomeetisis, R. megaphyllus, R.s. simplex 
and R.s. subsp. nov. cluster separately on function 
1 with R.s. parvus also clustering separately from R. 
megaphyllus and R. bomeetisis on this function. 
Additionally R. megaphyllus clusters separately 
from both R.s. parvus and R. borneensis on function 
2. The unallocated R. simplex keyensis specimens 
cluster close to R. s. simplex. 
The character loading most heavily (>0.6) on 
function 1, which is presumed to be a most 
important discriminant between most of these 
taxon, was forearm length (Table 4b). the 
characters loading heavily (>0.6) on function 2, and 
presumed important discriminants between R. 
megaphyllus and both R.s. parvus and R. borneensis 
were pes length, PES, and digit 4 metacarpal 
length, D4M (Table 4b). 
R. simplex analysis 
A DFA was carried out on the three subspecies 
of R. simplex (R.s. simplex, R.s. parvus, and R. s. 
subsp. nov.) after combining both sex and age 
groups. This analysis was carried out for skulls, 
dentary and dental characters only after deleting 
three skull characters shown by the multiple 
regression analysis to be significantly (P<0.01) 
influenced by sex or age (BB, P'W, and M 2 W, see 
Table 3a). Also, MSF and C'C'W were deleted 
because too many individuals had these values 
missing. The skull analyses was then run with the 
reduced set of 25 characters. 
Similarly the DFA of external characters was run 
after deletion of ear length from the analysis 
because there was a significant (P<0.01) interaction 
between sex and island for this character (Table 
3b). 
Skulls. The DFA for the R. simplex subspecies was 
first run using the reduced set of 25 characters and 
using islands as the a priori groupings. When these 
islands were then grouped to represent the three 
taxa R. s. simplex, R. s. parvus and R. s. subsp. nov. 
(R. s. keyensis unallocated), the configuration of the 
taxon clusters in discriminant function space was 
similar to that produced above. However, as the 
number of characters was considerably larger than 
the number of individuals in one taxon group (R. 
simplex parvus, 12) a reduced set of five characters 
