130 
D.J. Kitchener, W.C. Packer, A. Suyanto 
Table 2 Multiple regressions on sex and island populations of Nyctimene albiventer, N. cephalotes and N. keasti for 
skull, dentary, dental and external body characters. F values are presented for the main effects and their 
interactions. For explanation of character codes see Material and Methods section. Significance levels are *, 
0.05>p>0.01; ** 0.01>p>0.001; ***, pcO.001. 
CHARACTER 
MAIN EFFECT 
SEX ISLAND 
INTERACTION 
SEX. ISLAND 
GSL 
0.354 
54.340*** 
0.774 
CBL 
0.050 
45.776*** 
0.968 
PIF 
0.646 
27.866*** 
0.831 
RL 
0.530 
41.475*** 
0.911 
RH 
0.015 
10.427*** 
0.795 
BB 
0.335 
17.716*** 
0.817 
ZW 
0.618 
26.450*** 
0.877 
M'M 1 
0.367 
17.584*** 
0.774 
C'C 1 
1.770 
5.361*** 
1.378 
C’M 1 
0.083 
32.995*** 
0.430 
IOB 
0.110 
15.210*** 
1.012 
POB 
4.631* 
0.959 
0.875 
MFB 
0.000 
16.366*** 
0.376 
p4p4 
2.096 
6.823*** 
0.367 
ML 
0.063 
41.255*** 
0.425 
CH 
0.010 
28.720*** 
1.173 
c,m 2 
0.084 
45.256*** 
0.695 
FA 
1.887 
62.079*** 
0.725 
MC3 
6.110* 
42.895*** 
0.914 
MC4 
4.667* 
46.285*** 
1.117 
MC5 
4.653* 
41.120*** 
1.070 
TV 
2.469 
3.331** 
0.818 
EAR 
0.422 
30.529*** 
0.571 
d.f. 
1,96 
8,96 
8,96 
morphological differentation has occurred between 
the two Kai Island populations. 
Nyctimene albiventer populations 
Kitchener et al. (1993) examined the 
morphological differences between the Halmahera 
group (N. a. albiventer), Aru and Papua New 
Guinea populations (attributed to N. a. papuanus) 
Table 3 Canonical variate function coefficients from 
DFA between the Nyctimene species (N. 
albiventer, N. cephalotes and N. keasti). 
Standardised values, followed by (in brackets) 
unstandardised values, for skull, dental and 
external body characters. For explanation of 
character codes see Materials and Methods 
section. 
CHARACTER 
FUNCTION 1 
FUNCTION 2 
C,M 2 
0.6941 (1.7737) 
0.5583 ( 1.4266) 
FA 
0.4366 (0.2167) 
-0.8945 (-0.4438) 
EAR 
-0.0297 (-0.0373) 
0.9784 ( 1.2309) 
RL 
0.4542 (1.5569) 
-0.0234 (-0.0802) 
M'M 1 
-0.4455 (-1.1545) 
-0.4087 (-1.0591) 
Constant 
-30.0241 
2.4944 
Explained 
variation (%) 
78.9 
21.1 
and Kai group (N. a. keasti) populations. The 
removal of the form keasti from association with N. 
albiventer in this study may affect the perceived 
morphological association between these 
remaining populations of N. albivetiter using DFA. 
DFA was run on all 23 characters for the islands: 
Halmahera, Aru and PNG. A reduced set of five 
characters was selected from the above analysis. 
Those were palatal breadth at P 4 P, P 4 P; rostrum 
height, RH; C'M 1 length; mesopterygoid fossa 
breadth, MFB; and greatest skull length, GSL. The 
DFA plots based on these five characters produced 
two significant functions (Figure 8). A total of 
90.5% of individuals was correctly classified to 
their appropriate island. Three individuals from 
Aru were misclassified to PNG, and one individual 
from PNG was misclassified to Aru. Function 1, 
which explains 74.5% of the variation, separates 
the Halmahera group from both the Aru Island 
and PNG populations. The characters loading 
heavily (>6) on Function 1 were P'P 1 and GSL 
(Table 6). Function 2, which explains 25.5% of the 
variation, separates the Halmahera group from the 
PNG population. The characters loading heavily 
(>0.6) on Function 2 were C'M 1 and RH (Table 6). 
When the DFA was run for the Aru and PNG 
populations and for a subset of five characters (ear 
