Records afthe Westem Australian Museum 17: 125""142 (1995).

Systematic review of Nyctimene cepJzalotes and N. albiventer (Chiroptera:Pteropodidae) in the Maluku and Sulawesi regions, Indonesia

D.J. Kitchener l, W.e. Packerl and A. Suyanto2

1 Western Australian Museum, Francis Street, Perth, Western Australia 6000

2 Museum Zoologicum Bogoriense, L1PI, Jalan Ir. H. Juanda 9, Bogor, Indonesia, 16122

Abstract - A univariate and multivariate statistical study of the morphologyof island populations of Nyctimene in the Maluku region, Indonesia,distinguishes Nyctimene keasti Kitchener, 1993, from the Tanimbar and KaiIslands, as a species. The form from the Tanimbar islands is described as anew subspecies of N. keasti; it is not associated with N. cepha/otes aspreviously considered in lilt. Nyctimene cepha/otes is restricted to Sulawesiand the Maluku region. The Sulawesi population of N. cepha/otes is describedas a distinct new subspecies. The Aru population of N. a/biventer issomewhat differentiated from the Papua New Guinea population of N. a.papuanus; its subspecific status is not determined.

Close island populations of N. c. cepha/otes, N. k. keasti and N. k. subsp.novo show considerable morphological differentiation.

INTRODUCTIONKitchener et al. (1993) examined morphological

variation among populations of Nyctimenealbiventer (Gray, 1863) in the Maluku region. Theyshowed that the population on Am wasmorphologically similar to the Papua New GuineaN. a. papuanus K. Andersen, 1910; while those onthe Kai and Banda Neira Islands (N. a. keastiKitchener, 1993) were distinct from both the Amform and the nominate subspecies on Halmaheraand Ternate Islands.

The discovery of N. a. keasti confuses somewhatthe previous diagnostic distinction between N.albiventer and the closely allied N. cephalotes (pallas,1767). This distinction had been based on the largersize of N. cephalotes and the fact that the femaleshad a dorsum that was much paler than the males,whereas in N. albiventer there was no such colourdistinction between the sexes (Anderson 1912). Anumber of Nyctimene a. keasti specimens hadforearm lengths overlapping those of both N. a.albiventer and N. cephalotes. Further, female N. a.keasti were much paler dorsally than the males.

In the study region, N. cephalotes is reported fromthe following islands: Ambon (fixed as the typelocality by Andersen 1912); Seram; Bum, LaratIsland (Tanimbar group); Timor and Sulawesi(Andersen 1912; Hill in Corbet and Hill 1992). Therecords from Larat Island and Timor Island appearto devolve on single specimens collected late lastcentury (Andersen 1912, Goodwin 1979).

Between 1987 and 1993 expeditions by staff fromthe Western Australian Museum and MuseumZoologicum Bogoriense carried out a survey of the

terrestrial vertebrate fauna of South Sulawesi,Maluku and Nusa Tenggara, Indonesia. On theseexpeditions, series of Nyctimene that would beattributed to N. cephalotes by earlier authors, werecollected from the following islands: Seram;Ambon; Bum; Yamdena and Selam, Tanimbargroup; and South Sulawesi. This paper reports onan examination of these extensive moderncollections which allow both a re-appraisal ofmorphological variation among island populationsof N. cephalotes and N. albiventer and areassessment of their taxonomy in the studyregion.

MATERIALS AND METHODS

A total of 142 adult specimens (listed inspecimens examined section) was examined. Thesewere from Banda Neira Neira I. (2); Dullah I., KaiKecil (7); Kai Besar I. (5); Wokam I., Am islands(27); the Halmahera group (6); Papua New Guinea(13); Ambon (10); Seram (16); Bum (2); Sulawesi(9); Selaru (13) and Yamdena (32). The localities ofthese specimens are shown in Figure 1. Apart fromthe Halmahera group specimens (AustralianMuseum, Sydney), all specimens are currentlylodged in the Western Australian Museum.

Seventeen measurements of skull, dentary anddental characters and six of external bodycharacters (all in mm) were recorded from adultspecimens.

The measurements recorded were (allmeasurement involving teeth were to alveoli): GSL,

c:...·\:)6

+

D.}. Kitchener, W.e. Packer, A. Suyanto

+

+o

140

•

•~_J,~~. ~·t·- ...

O' ~

~~'~'G·t·.. . \t, ~. ,•• Jo~l

o .~.....,~& .

.. • ~ AA

+

J - ~

~ D'~ .' ....~

• X •" .. .'\. P '.~

~!J~/7~.,f/

126

Figure 1 Locality of Nyctimene in this study; N. albiventer albiventer, .; N. a. subsp. indet., G; N. a. papuanus, A.; N.keasti keasti, 0; N. k. tozeri subsp. nov.,~;N. cephalotes cephalotes, .; N. c. aplini subsp. nov., X.

greatest skull length; CBL, condylobasal length;PIF, minimum length from posterior margin ofincisive foramen to palate posterior margin; RL,rostrum length, from anteriomost internal marginof orbit to nares; RH, rostrum height, from uppercanine alveoli to level of dorsal surface of nasals;BB, braincase breadth above zygoma; ZW,zygomatic width; MlMl and ClCl, width acrossMIMl and ClCl respectively, from the labial side;ClMl and ClM2

upper and lower canine to lastmolar length; lOB, minimum interorbital breadth;POB, minimum postorbital breadth; MFB,mesopterygoid fossa breadth, at the widest pointof the palatal flange; p4p4, palatal width betweenthe lingual aspect of p4p4; ML, mandible length,from condyle to anteriomost point of dentary; CH,dentary coronoid height; TV, tail to anus length;EAR, basal notch to apex length; FA, forearmlength; and MC3-5, metacarpal 3,4 and 5 length.

The skull, dentary and dental characters weremeasured to an accuracy of 0.01 mm, while theexternal body characters were measured to 0.1mm. Terminology used in the description of skull,dentary, dental and external body charactersfollows Hill and Smith (1984). Pelage descriptionsfollow the colour terminology of Smithe (1975).

Adults were diagnosed as those specimens withthe following sutures fused: basioccipital ­basisphenoid, basisphenoid - presphenoid andpalatine - maxillary (these sutures are illustrated inHeaney and Peterson 1984: Figure 4).

The effect of sex and island on all characters wasexamined by multiple regressions for the islands:Yamdena, Selaru, Kai Kecil, Kai Besar, Aru,Ambon, Seram, Sulawesi and Papua New Guinea.Examination of the residuals from regressionanalyses gave no indication of heteroscedasticity.

Stepwise canonical variate (discriminantfunction) analyses (OFA) were run for skull,dentary and dental characters and externalcharacters using all 23 characters for males andfemales combined, after first testing for sexualdimorphism. A reduced set of five of thesecharacters was used in all presented OFA becausein all instances they provided similar OF plots tothe full set of 23 characters. This reduced set ofcharacters was selected in all these analysesbecause the sample size of the smallest a priorigroup selected was always less than the totalnumber of characters. This reduced set ofcharacters was selected because they providedvalues that minimise Wilk's Lambda.

Systematic review of Nyctimene ceplla/otes and N. a/biventer 127

STATISTICS: RESULTS AND DISCUSSION

Univariate statisticsMean, standard deviation, mInimUm and

maximum values and sample size for each islandare presented in Table 1, this paper, and Kitcheneret al. (1993: Table 1) for all characters examined.

\\

\

Io

-- ./o'

o

oo

•

o -0

00600S°"-

rt0Gl. (] "0,o ... 0 \. \

o \. \. \

•'V'

/

o

!•!If o.f

//

/

-4- 8 +-~,--~,....--r~-r~-,--~-,-~,--~,....--r~-

- 5

c:Q­uc:::>...

Sex

Metacarpal 3-5 lengths and postorbital breadthwere significantly influenced by sex (F196 = 6.11, P= 0.015; F1,96 = 4.67, P = 0.033; F1,96 =' 4.653, P0.033; and F1.96 =4.639, P =0.034, respectively.

IslandOnly postorbital breadth was not highly «.01)

significantly related to island. Clearly there wasconsiderable difference in morphology betweenthese island populations.

Multiple regressionsMultiple regressions were run for skull, dentary,

dental and external characters on sex and nineisland populations. The islands of Buru, BandaNeira, Halmahera, Ternate and Moratai wereexcluded because of their small sample size and/or absence of either males or females (Table 2).

InteractionsThere were no significant interactions.

Multivariate analysesBecause the multiple regression analysis

involved testing a large number of interactions,some of these tests will be significant at0.05>P>0.01 by chance alone. For this reason, noneof the characters measured were considered to bemarkedly influenced by sex. As a result both malesand females are combined for all 23 charactersexamined in the following DFA.

Function 2

Figure 2 Canonical variate analysis on the followingisland populations: Aru, [;] ; Kai Besar, 0; KaiKecil, 0; the Halmahera group (Halmahera,Ternate and Morotai), -; Papua New Guinea,...; Selaru 0; and Yamdena, II (Tanimbargroup); Seram, +; Ambon, e; and Sulawesi,X; Banda Beira +, and Buru *, wereungrouped. The DFA plots of functions 1 and2 were based on a selection of five characters(skull, dentary, and external body), withmales and females combined.

All populationsThe DFA was first run using all 23 skull, dental,

dentary and external body characters and theislands Aru, Kai Besar, Kai Kecil, the Halmaheragroup (Halmahera, Temate and Morotai) PapuaNew Guinea, Selaru and Yamdena (TanimbarIslands), Seram, Ambon and Sulawesi. BandaNeira and Buru were ungrouped because therewere only two individuals from each of theseislands. The above analysis was then repeatedusing only five characters (forearm length, FA; earlength EAR; C1M 2

length, C1M2; width across M1M 1

from labial side, MIM1; and rostra I length, RL).This analysis extracted four significant functionswhich together explained 99.6'X, of the variation,with 68.0'j!o of individuals classified to their correctisland group.

From Figure 2, three broad groups were defined:

the Ambon group (Ambon-Seram-Buru-Sulawesi);the Kai group (Kai and Tanimbar islands); and theHalmahera group (Halmahera, Aru and PapuaNew Guinea). When a DFA was again run usingthe reduced set of five characters and these threeisland groups, the analysis extracted twosignificant functions (Figure 3). Function 1, whichexplained 78.9% of the variance, separated theAmbon and Kai groups from the Halmaheragroup. The character loading heavily on thestandardised canonical variate coefficients (>0.6)on this function was C

jM

2length (Table 3).

Function 2, which separated the Ambon groupfrom the Kai group, explained 21.1% of thevariation. The characters loading heaviest (>0.6) onFunction 2 were forearm and ear length (Table 3).A total of 99.2% of individuals was correctlyclassified to their appropriate island group. Onespecimen of the Ambon group classified to the Kai

128 D.J. Kitchener, W.e. Packer, A. Suyanto

Table 1 Measurements, in mm, for skull, dentary, dental and external body characters (see Material and Methodssection for explanation of charas!er codes) of adult Nyctimene cephalotes aplini, N. c. cephalotes, N. keasti keastiand N. k. tozeri. N, sample size; X, mean; SO, standard deviation; MIN, minimum; MAX, maximum.

GSL CBL PIF RL RH BB ZW

Nyctimene cephalotes aplini(Sulawesi)

N. cep1ullotes cephalotes(Seram, Ambon, Buru)

N. cep1ullotes(Sulawesi, Seram,Ambon, Buru)

N. keasti keasti(Kai Besar, Kai Kecil)

N. keasti tozeri(Selaru, Yamdena)

N. keasti(Kai Is, Tanimbar Is)

NXSOMINMAX

NXSOMINMAX

NXSDMINMAX

~XSOMINMAX

NXSDMINMAX

~XSDMINMAX

931.740.47

31.0532.42

2530.130.68

28.3231.05

3430.560.95

28.3232.42

1228.770.50

28.0429.38

4429.500.56

28.4530.78

5629.350.62

28.0430.78

930.220.33

29.7130.67

2528.660.64

26.9829.69

3429.070.90

26.9830.67

1227.280.52

26.0227.88

4427.740.48

26.6428.74

5627.650.52

26.0228.74

912.771.48

11.7613.54

2612.230.48

11.4513.70

3512.370.53

11.4513.70

1212.060.26

11.5812.47

4411.850.43

10.9212.81

5611.910.41

10.9212.81

96.390.39

5.727.00

265.940.275.276.39

356.060.365.277.00

125.520.344.966.00

445.730.275.076.30

565.680.304.966.30

97.250.276.647.59

267.050.346.167.79

357.100.336.167.79

126.360.375.756.96

446.720.286.227.30

566.640.335.757.30

913.400.27

12.9513.81

2612.760.37

11.5913.39

3512.930.44

11.5913.81

1212.340.31

11.6412.68

4412.720.28

12.2013.94

5612.640.33

11.6413.94

921.370.48

20.7422.24

2619.800.72

17.8120.90

3520.210.96

17.8122.24

1219.060.33

18.6319.71

4419.340.45

18.1120.40

5619.280.44

18.1120.40

99.560.199.319.92

268.880.327.819.42

359.060.427.819.92

128.160.247.668.45

448.700.238.099.09

568.580.327.669.09

95.800.275.436.17

265.570.245.046.02

355.630.265.046.17

125.460.245.185.80

445.570.185.245.96

565.550.205.185.96

group. Of the two "ungrouped" Banda Neiraspecimens, both classified to the Kai group.

These three island groups represent threedescribed taxa as will be discussed below. Thesetaxa are Nyctimene cephalotes (Ambon, Seram, Buru,Sulawesi); N. keasti (Kai and Tanimbar groups,with Banda Neira specimens referred to thisgroup); and N. albiventer (Halmahera, Aru andPapua New Guinea).

Nyctimene cephalotes populationsA OFA was run for all 23 characters for the

islands Ambon, Seram and Sulawesi, with Buruunallocated. A reduced set of five characters wasselected from this above analysis. These were:condylobasal length, CBL; zygomatic width, ZW;mesopterygoid fossa breadth, MFB; mandiblelength, ML; and upper canine to first molar cusplength, CIMI. The selection of only skull andmandible characters in this subset of five charactersenabled the inclusion of an additional sample ofSulawesi individuals (which were missing some

external measurements) to be included in this OFA.This latter OFA extracted two significant functionswhich explained 100% of the variation (Figure 4).A total of 73.5% of individuals was correctlyclassified to their correct island population, withmisclassification occurring between the Ambonand Seram populations. One of the Buru specimenswas allocated by OFA to the Seram population andthe other to the Ambon population.

The OFA was run again using 23 characters forthe two a priori groups: Sulawesi and the groupAmbon, Seram and Buru. A subset of fivecharacters (condylobasallength; CIMI length; MIMIbreadth; rostrum height; incisive foramen toposterior palatal length) was selected from theabove analysis. This latter analysis extracted a verysignificant function and all individuals werecorrectly classified to their appropriate group(Figure 5). The characters loading most heavily onthis function were CBL (1.9), CIMI (-1.4) and MIMI(0.6) (Table 4).

The type locality of the nominate subspecies of

Systematic review of Nyctimene ceplla/otes and N. a/biventer

Table 1 (continued)

129

9 91062 6.440.28 0.40

10.34 6.0311.11 7.17

26 2610.16 5.650.43 0.359.02 4.94

10.82 6.13

35 3510.28 5.850.44 0.509.02 4.94

11.11 7.17

12 129.57 5.690.24 0.248.98. 5.299.80 6.12

44 4410.01 5.810.33 0.288.39 5.30

1050 6.49

56 569.91 5.780.36 0.278.39 5.29

10.50 6.49

POB

96.180.605.517.39

265.770.445.176.94

355.880.515.177.39

126.020.355.216.44

445.660.415.036.65

565.740.425.036.65

MFB

95180.134.965.37

255.160.204.795.61

345.160.194.795.61

124.760174.505.05

445.000.174.565.39

564.950.204.505.39

96.020.235.476.26

265.580.234.895.94

355.690.304.896.26

125.190.105.035.38

445.600.195.245.97

565.510.245.035.97

ML

92379

0.4123.2024.37

2622.55

0.67210223.69

3522.870.82

21.0224.37

122149

0.5120.5822.11

4422.000.49

21.0423.10

5621890.53

20.5823.10

CH

913920.31

13.4114.47

2612.99072

11.4714.23

3513.230.76

11.4714.47

1212.860.31

12.3313.32

4413.380.47

12.3414.46

5413.260.49

12.3314.46

91196025

11.6612.41

2611.460.44

102112.09

3511.590.46

10.2112.41

1210.840.26

10.2711.18

4411.280.29

10.7612.10

5611.180.34

10.2712.10

FA

5686

0.668.169.5

2463.4

2159.167.6

2964.3

2.759.169.5

1258.213

55.159.7

4460.1

1.258.063.0

5659.7

1.455.163.0

MC3

549119

47.4518

2445116

41448.0

2945.8

2.241.4518

1242.2

1.140.744.5

4443.510

41.645.7

5643.2

1.240.745.7

MC4

544.822

423477

24418

1538.644.1

2942.3

2.038.647.7

1238.5

0.737.539.6

4439.6

1.137.543.1

5639.4

1.137.543.1

MC5

.s486

2.7452520

2444.6

1.740.247.2

2945.3

2.440.252.0

1240.610

38.742.7

4442.110

40.044.2

5641812

38.744.2

TV

5251

1.0237262

2422.413

20624.8

2922.916

206262

12207

1.518.624.1

4421.3

1.618.025.4

5621216

18.025.4

EAR

516203158165

2414.20612.915.2

2914.510

12.916.5

1215.00.5

14215.6

4415.40.614.316.9

5615.30.614.216.9

N. cephalotes was fixed as Ambon by Andersen(1912). The above documentation of themorphological separation of the Sulawesipopulation from N. c. cephalotes warrants itsdescription as a new subspecies of N. cephalotes.

Nyctimene keasti populationsOFA was run for all 23 characters for the islands

Kai Besar, Kai Keci!, Selaru and Yamdena, withBanda Neira unallocated. A reduced set of fivecharacters was selected from this above analysis.These were palatal width between the lingualaspect of P41>l; ear length; forearm length; braincasebreadth; minimum length from posterior margin ofincisive foramen to palate posterior margin. Thislatter DFA produced two significant functions,which combined explained 98.4'10 of the variation(Figure 6). A total of 81.5% of individuals werecorrectly classified to their appropriate island, withmisclassification occurring between the twoTanimbar islands. One of the unallocated BandaNeira specimens classified with the Selaru

population and the other with the Yamdenapopulation.

The OFA run with 23 characters and only two apriori groups: Tanimbar group (Selaru andYamdena) and Kai group (Kai Kecil and Kai Besar)produced a similar discriminant histogram to thatproduced from a subset of the five characters thatwere selected from the above analysis (palatebreadth at p4p4; incisive foramen to posterior palatelength, PIF; braincase breadth, BB; postorbitalbreadth, POB; and forearm length, FA). Thefunction extracted from this latter analysis washighly significant and resulted in all individualsbeing correctly allocated to their appropriate group(Figure 7). The characters loading heavily (>0.6) onthis function were p4p4, FA and BB (Table 5).

The type locality of the nominate subspecies ofN. keasti is Pulau Oullah, Kai Kecil. The aboveanalysis documents clearly the morphologicalseparation of the Tanimbar form from N. k. keasti.It indicates that the Tanimbar form warrantsdescription as a new taxon. It also shows that some

130 D.J. Kitchener, W.e. Packer, A. Suyanto

Table 2 Multiple regressions on sex and island populations of Nyctimene a/biventer, N. cepha/otes and N. keasti forskull, dentary, dental and external body characters. F values are presented for the main effects and theirinteractions. For explanation of character codes see Material and Methods section. Significance levels are *,0.05>p>0.01; ** O.01>p>O.OOI; ***, p<O.OO1.

MAIN EFFECT INTERACTIONCHARACTER SEX ISLAND SEX. ISLAND

GSL 0.354 54.340*** 0.774CBL 0.050 45.776*** 0.968PIF 0.646 27.866*** 0.831RL 0.530 41.475*** 0.911RH 0.015 10.427*** 0.795BB 0.335 17.716*** 0.817ZW 0.618 26.450*** 0.877MIMI 0.367 17.584*** 0.774CICI 1.770 5.361*** 1.378CIMI 0.083 32.995*** 0.430lOB 0.110 15.210*** 1.012POB 4.631* 0.959 0.875MFB 0.000 16.366*** 0.376p4p4 2.096 6.823*** 0.367ML 0.063 41.255*** 0.425CH 0.010 28.720*** 1.173CIM2

0.084 45.256*** 0.695FA 1.887 62.079*** 0.725MC3 6.110* 42.895*** 0.914MC4 4.667* 46.285*** 1.117MC5 4.653* 41.120*** 1.070TV 2.469 3.331** 0.818EAR 0.422 30.529*** 0.571dJ. 1,96 8,96 8,96

morphological differentation has occurred betweenthe two Kai Island populations.

Nyctimene albiventer populationsKitchener et al. (1993) examined the

morphological differences between the Halmaheragroup (N. a. albiventer), Aru and Papua NewGuinea populations (attributed to N. a. papuanus)

Table 3 Canonical variate function coefficients fromDFA between the Nyctimene species (N.albiventer, N. cephalotes and N. keasti).Standardised values, followed by (in brackets)unstandardised values, for skull, dental andexternal body characters. For explanation ofcharacter codes see Materials and Methodssection.

CHARACTER FUNCTION 1 FUNCTION 2

C1M20.6941 (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)M1Ml -0.4455 (-1.1545) -0.4087 (-1.0591)Constant -30.0241 2.4944Explainedvariation 78.9 21.1

and Kai group (N. a. keasti) populations. Theremoval of the form keasti from association with N.albiventer in this study may affect the perceivedmorphological association between theseremaining populations of N. albiventer using DFA.

DFA was run on all 23 characters for the islands:Halmahera, Aru and PNG. A reduced set of fivecharacters was selected from the above analysis.Those were palatal breadth at ptpt, ptpt; rostrumheight, RH; CIMI length; mesopterygoid fossabreadth, MFB; and greatest skull length, GSL. TheDFA plots based on these five characters producedtwo significant functions (Figure 8). A total of90.5% of individuals was correctly classified totheir appropriate island. Three individuals fromAru were misclassified to PNG, and one individualfrom PNG was misclassified to Aru. Function 1,which explains 74.5% of the variation, separatesthe Halmahera group from both the Aru Islandand PNG populations. The characters loadingheavily (>6) on Function 1 were p4p4 and GSL(Table 6). Function 2, which explains 25.5% of thevariation, separates the Halmahera group from thePNG population. The characters loading heavily(>0.6) on Function 2 were CIMI and RH (Table 6).

When the DFA was run for the Aru and PNGpopulations and for a subset of five characters (ear

Systematic review of Nyctimelle ceplla/otes and N. a/biventer 131

x

x"

... . .

length, EAR; rostrum height, RH; rostrum length,RI,; ClM' length and interorbital breadth, JOB). thefunction extracted was highly significant with94.9'1,) of individuals correctlv classified to theirappropriate population (Figure 9). Only two Anlindividuals were misclassified to the PNG group.The characters loading heaviest (>0.6) on thisfunction were C'M' and RH (Table 7) .

Summary of multivariate analysesThe above DFA supports the view that the

recently described taxon keasti is equally distinctfrom both N. a/biventer and N. ceplza/otes andshould be treated as a species. Further, it showsthat the Tanimbar populations, rather than beingreferred to N. ceplza/otes as indicated by Andersen(1912) and Hill in Corbet and Hill (1992), are moreclosely allied to N. keasti.

The Sulawesi populations referred to N.cepha/otes are distinct from Maluku population ofthe species and are identified as a subspecies of N.cep/Ia/otes .

The species N. cepha/otes, N. keasti and N.albiventer are rediagnosed for the study region andthe Sulawesi and Tanimbar forms described as new

•.•

o 0

oo •..o

00

li o•

• 0

o

o/

o

-4

•

c 0.~-::;C:::l.....

- 2

·4 ·3 o

Function 2

Figure 3 Canonical variate analysis on three groupingsof islands, (i), Aru, Halmahera and PapuaNew Guinea; (ii) Kai Besar, Kai Kecil,Yamdena, and Selaru; and (iii) Sera m,Ambon, Buru and Sulawesi: Banda Neirawas ungrouped. The DFA plots of functions1 and 2 were based on a selection of the samefive characters as in Figure 2, with males andfemales combined. Symbols are as for Figure2.

cCl

'"C:::l.....

o

xx

Table 4 Canonical variate function coefficients fromDFA between the population of Nyctimenecephalotes apfini from Sulawesi and thecombined populations of N. c. cephalotes fromSeram, Ambon and Buru. Standardisedvalues, followed by (in brackets)unstandardised values, for skull and dentalcharacters. For explanation of character codessee Materials and Methods section.

CHARACTER FUNCTION 1Function 2

CBLC'M'M'M'RlI('IFConstant

19328 ( 33396)-1.3704 (-3.4858)0.6323 (2.1542)

-D.4987 (-1.5196)-0.4524 (-0.9386)

-58.3252

Figure 4 Canonical variate analysis among threeisland populations of Nyctimelle cepha/otes(Seram, Ambon and Sulawesi) with the Burupopulation ungrouped. The DFA plots offunctions 1 and 2 were based on a selectionof five skull characters with males andfemales combined. Symbols as for Figure 2.

132 D.J. Kitchener, W.e. Packer, A. Suyanto

Figure 5 Canonical variate analysis between theSulawesi population (Nyctimene cephalotesaplini subsp. nov.) (.) and those of the groupof Seram, Ambon and Buru (N. c. cepitalotes)(0). The histogram for function 1 was basedon a selection of five skull characters, withmales and females combined.

Function 1

6

4

f

2

-4 -2 o 2 4

Table 6 Canonical variate function coefficients fromDFA between populations of Nyctimenealbiventer, N. a albiventer, N. a. papuanus andN. a. subsp. indet. from Halmahera group,and Papua New Guinea and Aru,respectively. Standardised values, followed by(in brackets) unstandardised values, for skulland dental characters. For explanation ofcharacter codes see Materials and Methodssection.

CHARACTER FUNCTION 1 FUNCTION 2

p4p4 0.7335 (3.0491) 0.4262 (1.7717)CIMl 0.1907 (0.7626) -0.9287 (-3.7137)RH -0.2628 (-0.8189) 0.8977 (-2.7979)MFW -0.5806 (-2.3344) 0.1779 (0.7151)GSL 0.6701 (1.0692) 0.1438 (0.2295)Constant -35.8486 -3.3963Variationexplained (%) 74.5 25.5

4,.------------------..,

o 1

Function 2-1-2- 3

-4

Nyctimene cephalotes (Pallas, 1767)

SYSTEMATICS

DiagnosisNyctimene cephalotes differs from Nyctimene

subspecies below.The Aru form, is still phenetically close to the

Papua New Guinea form, but is moredifferentiated morphologically from the PapuaNew Guinea form than was thought to be the casein Kitchener et al. (1993). Its subspecific status isnot determined. Within the island groups ofTanimbar and Kai there has been considerablemorphological differentiation between populationsof subspecies of N. keasti inhabiting differentislands. This has occurred to a lesser extent alsobetween the Ambon and Seram populations of N.c. cephalotes.

Table 5 Canonical variate function coefficients fromDFA between populations of Nyctimene keastikeasti from the Kai Islands and N. k. tozerifrom the Tanimbar Islands. Standardisedvalues, followed by (in brackets)unstandardised values, for skull, dental andexternal body characters. For explanation ofcharacter codes see Materials and Methodssection.

CHARACTER

p4p4FABBPOBPIFConstant

FUNCTION 1

0.6922 (3.9545)-0.7606 (-1.9787)0.7532 (2.5839)

-0.4283 (-1.0462)0.3125 (0.2625)

-40.5598

Figure 6 Canonical variate analysis among four islandpopulations of Nyctimene keasti: Kai Besar,Kai Kedl (N. k. keasti), Selaru and Yamdena(N. k. tozeri subsp. nov.). The Banda Neirapopulation was ungrouped. The DFA plotsof functions 1 and 2 were based on a selectionof five skull, dentary and external bodycharacters, with males and femalescombined. Symbols as for Figure 2.

Systematic review of Nyctimene cepll1llotes and N. albiventer 133

Canonical variate analysis between Nyctimenekeasti populations from the Kai Island group(Kai Besar and Kai Kecil) (N. k. keasti) (.)and the Tanimbar Island group (Selaru andYamdena) (N. k. tozeri subsp. nov.) (0). TheBanda Neira population was ungrouped (~).

The histogram for Function 1 was based on aselection of five characters (skull, dentaryand external body), with males and femalescombined.

\

Table 7 Canonical variate function coefficients fromDFA between populations of Nyctilllellealbiventer from Aru islands and Papua NewGuinea. Standardised values, followed by (inbrackets) unstandardised values for skull,dental and external body characters. Forexplanation of character codes see Materialsand Methods section.

0.7909 (3.0302)-D.6888 (-2.1155)0.5633 (0.6949)04382 (1.6215)

-04345 (-1.4417)-23.1370

FUNCTION 1

ClMl

RHEARRLlOBConstant

CHARACTER

a/biventer by averaging larger in all measurements(Table 1). Incisive foramen to posterior palatallength shorter relative to postorbital breadth(Figure 10); C1M2

length generally longer relativeto M1Ml length (Figure 11); and forearm longerrelative to ear length (Figure 12). Female dorsumpaler than that of males.

It differs from Nyctimene keasti by averaginglarger in all measurements except for cranial height(Table 1). Condylobasal length generally longerrelative to both rostrum height, interorbital breadthand cranial height (Figure 13A,B,C, respectively);and forearm longer relative to ear length (Figure12).

42

j IL

/ L)\j IJ

o

Function I

-2-4

4

3

Figure 7

Function 2

a

c: 0.::-'"'c:

'"... .,

•I-~

\

a

/ .a a j ~./\

"\\ ./\\ .-, ~ \\ .\

\ .\\

\\ \\ \

\\.

o

Nyctimene cephalotes aplini Kitchener,subsp. novo

HolotypeMuseum Zoologicum Bogoriense Number MZB

15903; adult male; weight 46 g; carcase fixed in10% formalin and preserved in 75% ethanol; skulland mandibles separate; collected by DrsAugustine Suyanto on 28 July 1990.

3

2

f

2Figure 8 Canonical variate analysis between Nlfctimenea/biventer population; from the Haimaheraisland group (Halmahera, Morotai andTernate) (N. a. a/biventer), Aru (N. a. subsp.indet.) and Papua New Guinea (N. a.papllanlls). The DFA plots of functions 1 and2 were based on a selection of 5 characters(skull and dental), with males and femalescombined. Symbols as for Figure 2.

-4

Figure 9

-2 0

Function 1Canonical variate analysis between twopopulations of Nyctimene a/biventer: Aru (N.a. arllensis subsp. nov.) (.) and Papua NewGuinea (N. a. papllanlls) (0). The histogramfor function 1 was based on a selection offive characters (skull, dental and externalbody), with males and females combined.

134 D.J. Kitchener, W.e. Packer, A. Suyanto

14,------------------------, Description

\Skull, dentary and dentition

Large skull with greatest length 31.7 ± 0.47, broadzygomatic width 21.4 ± 0.48, palate long and broadboth anteriorly, with ClCl breadth 5.8 ± 0.27, andposteriorly, with M1M1 breadth 9.6 ± 0.19; rostrumshort, nasal dorsal surface almost horizontal,occasionally with dorsal distal inclination; rostrumhigh; premaxillary part of nares terminates directlyabove incisors or projects slightly anterior to them;braincase typical of N. cephalotes with craniuminflated to maximum height at junction of strongsagittal and weak postorbital crests; postorbitalbreadth large; usually deeply basined and opensanteriorly into deep frontal sulcus; frontalsconverging slightly posteriorly; sagittal andlambdoidal crests moderately high; incisiveforamen subcircular, reach posteriorly almost toline joining p2p2 anterior face; posterior palateextends well beyond M1, its margin a broad Ushape; basi- and presphenoid median ridgeprominent; pterygoid process low and slightlycurved ventrally and laterally.

Upper tooth row slightly curved; upper incisorsin contact, sited close to line joining C1C1

7,56.55.59+---~-___,_--~--__,_--~-____;

4.5

12

10

11

13

­..Cl~,c:ell

Ec"­Cl"­ell::.';;'uc:

Postorbital Breadth 13,----------------------,

Figure 10 Plot of incisive foramen to posterior palatelength versus postorbital breadth for allisland populations of Nyctimene studied.Males and females combined. Symbols as forFigure 2.

ParatypesTornado, Lake Lindu area Central Sulawesi,

WAM M27659 AS!, skin and skull; Soroako,Sulawesi Selatan, WAM M27676 AS!, skin andskull; WAM M33145, near Bontobonto, SulawesiSelatan (c. 5°07'S, 119°39'E), AS!, spirit carcase andskull; Maros, Sulawesi Selatan (c. 4°59'S, 119°35'E),WAM (M33146, M33207), 200 skin and skull,WAM (M33208, M3321D-12), 400, spirit carcaseand skull.

Type localityMaros, Sulawesi Selatan (= S. Sulawesi) (c. 4°59',

119°35'E), altitude c. 50 m; mist-netted.

DiagnosisNyctimene cephalotes aplini differs from N. c.

cephalotes in averaging larger in all skull, mandibleand external measurements; in some it is absolutelylarger (Table 1). Condylobasal length greaterrelative to both mesopterygoid fossa breadth, MFB,and CMl length (Figures 14A,B, respectively); andzygomatic width greater relative to MFB (Figure15).

12

::11en

c:ell~

N1:..,

10

10

Outside M I M 1 Breadth

Figure 11 Plot of C1M

2length versus outside MIM1

breadth for all island population ofNyctimene studied. Males and femalescombined. Symbols as for Figure 2,

Systematic review of Nyctimelle cepllalotes and N. allJivellter 135

69

Ear Length

//

DistributionCentral and South Sulawesi, probably also North

Sulawesi.

Nyctimene cepllalotes cepllalotes Wallas, 1767)

Vespertilio cephalotes Pallas, 1767: 10, pis 1,2"Moluccas" .

Vespertilio cephalotes melinus Kerr, 1792: 98. Typelocality restricted to Amboina by Andersen(1912).

Cephalotes pallasi Geoffroy, 1810: 107 (renaming ofcephalotes Pallas).

EtymologyNamed after our colleague, Or Ken Aplin,

Western Australian Museum, who participated inthe expedition to Timor Island, Nusa TenggaraTimur, and who frequently has assisted usinterpret the lndo-Malay mammal fauna.

Externals (measurements of ethanol preservedspecimens followed by, in square brackets,'cabinet' skin specimens).

Large body and ears. For example, forearm long68.6 (68.1-69.5) 5 and [70.9 (69.0-73.0) 4]; ear long16.2 (15.8-16.5) 5 and [16.6 (15.1-18.3) 4] and taillength 25.1 (23.7-26.2) 5 and [23.3 (19.9-27.2) 4].Wing membrane from phalanx 1 of digit 2 in allnine specimens examined.

Males and females with similar coloured pelage.Slight colour patterning on dorsum and venter:neck to lower scapular region a broad band ofpaler Smoke Gray to Buff Yellow, occasionallytinged with Clay Color on shoulders and behindears; remainder of back a darker Drab with hairs inmid dorsal region c. 9 long, merges into FawnColor to Clay Color on flanks; median centraldorsal stripe from head to tail, 2-6.5 wide andOlive Brown. Basal one-third to half of dorsumhairs pale-dark Olive Brown. Forehead and faceDrab to Cinnamon; chest and abdomen StrawYellow to Buff Yellow with abdomen occasionallySmoke Gray merging to darker Grayish Olive atflanks. Uropatagia Fuscous. Spectrum Yellow spotson patagia, ears, and skin covering all digits,particularly noticeable on forearm where spotsoften merge to form a broad line of colour.

approximately in line of tooth row, shorter labialcusp distinct; P4 considerably shorter than P" labialcusp taller than lingual cusp but less so than in P"these cusps connected by commissure that arcsaround anterior face of tooth; M

jlonger and

narrower than P4 but with both labial and lingualcusps much reduced on those of P4; M

2small, about

half occlusal area of Ml. M l posterior shelf slightlylarger than that of p) and P4'

17161514131249 +-~--.----,-~-,--~-,----,--~--I

11

Figure 12 Plot of forearm length versus ear length forall island populations of Nyctimene studied.Males and females combined. Symbols as forFigure 2.

S4

.N \+ •

\

•" .64

anteriormost face, not in contact with Cl; Clprominent cusp with gentle lingual ridgeconnecting to lingual and posterior cingulum,lingual cingulum occasionally with posterolingualcusplet; Cl with moderate secondary labial cusp onposterolabial ridge, this cusp not apparent inspecimens with worn teeth; p2 subcircular, buccalarea half to three-quarters that of incisors, widelyspaced between Cl and p3; p3 occlusal viewsuboval, with prominent labial cusp and muchlower separate lingual cusp; posterior basal shelfwell defined and forms slight basin, occlusal areaslightly larger than p4; p4 occlusal viewsubrectangular with posterolingual salient frombasal shelf; labial cusp lower than that of pI, lingualcusp less clearly defined than that of pI, withlonger posterior ridge that reaches almost toposterolingual edge of posterior basal shelf, lingualcusp closer in height to labial cusp than in pI; Mlsubrectangular, labial cusp low and only slightlytaller than lingual cusp, posterior basal shelfdefinition similar to that of p4; Cl not in contactwith each other or with P,; P, suboval, slightlytaller than Cl cingulum; P; with tall labial cuspwith gently sloping anterior and posterior flanges

E..'"'"..Q...

136 D.J. Kitchener, W.e. Packer, A. Suyanto

31 ,---------------------, 31,----------------------,

30

27

28

29

26

:SClCQl....'0..III.clCl

'El1:1CClU

29

30

28

27

26

6 7.06.025 +---~---_,__--~---,__--~

5.0

25 +---~-.........,r__----__r--~--_i5

A Rostrum Height B Interorbital Breadth

31,-------------------,

30

29

:SClCQl....'0 28..Cl.clCl'El1:1CClu

27

26

Figure 13 Plot of condylobasal length versus (A)rostrum height, (B) interorbital breadth and(C) cranium height for individual Nyctimenekeasti and N. cephalotes. Symbols as forFigure 2.

HolotypeAndersen (1912: 707) states that "The species was

described by Pallas from two female specimensfrom the 'Moluccas', at least one of which seems tohave been in Amsterdam (in the possession of J.A.Schlosser). Whether any of the types are still inexistence is unknown".

Type localityType locality fixed as Amboina by Andersen

(1912: 707).

Specimens examined

INDONESIA

Ambon Island: Ambon, 3°41'5, 112°lO'E, WAMM(38762,38778,38821-22,38861,42351,42364)(4gg,30o).

Desa Amahusa, 3°41'5, 128°lO'E, WAM M(43128,43921) (1 g, 1 sex unknown)

1412 13

Cranium Height

25 +--~-_,_-~-__r----_,__-~--j11

c

Systematic review of Nyctimene cepllalotes and N. albiventer

'x

137

x xN. c subsp nov

]0.5N. c. ,lp/m! subsp_ nov

\x

\x

295

..29

.M M J::J:: .- • • .

er> er>M . .

C C

'" '" •.... ....0; 0;", N, c. cephalotes ",., .,,Q ,Q 28.5

.:! .:!

"" 28 ""'C 'CC CQ Q

'-' '-'

N c. cephafofes

27.527

55

A

26 +------~-----,._-------I45

Mesopterygoid fossa Breadth B

26.5+---,.---,._--~--,---~--8.5 95 105 11.:,

elMI Length

Figure 14 Plot of condylobasal length versus (A) mesopterygoid fossa breadth and (B) e1M1 length for individualNyctimene cephalotes. Symbols as for Figure 2.

Bum Island: S.W. Bum, c. 3°40'S, 126°30'E, WAMM(43883-84) (2 <jl <jl).

Seram Island:Kanikeh 3°06'S, 129°28'E, WAMM(34560-3, 34565) (1 <jl, 400). Solea, 20 53'S,129°32'E, WAM M(34548, 34550-6, 34558-9 (3 <jl <jl ,

700).

DiagnosisNyctimene c. cephalotes differs from N. c. aplini as

described in the earlier diagnosis of thatsubspecies.

DescriptionSkull moderately long 30.1 ± 0.68, narrow with

braincase breadth 12.8 ± 0.37 and zygomatic width19.8 ± 0.72; palate posterior margin V or sharp Ushape; mesopterygoid fossa moderately wide 5.2 ±

0.20; CIMI tooth row moderately long 10.2 ± 0.43; Pgenerally closely opposed to Cl, sometimes incontact with Cl; incisive foramen subcircular,small, generally not reaching posteriorly to a linejoining p2_p2 anterior face; frontal betweenpostorbital process flat or only very slightlybasined.

Apart from the above characteristics andgenerally smaller size, the shape of the skull,

dentary and dentition is very similar to N.cephalotes aplini.

ExternalsForearm moderately long 63.4 ± 2.07; ear

moderately long 14.2 ± 0.61. Wing membranegenerally connected between phalanx 1 of digits 2and 3 (46%, N = 13) but also commonly fromphalanx 1 of digit 3 (29'10, N = 8) and commonlyfrom phalanx 1 of digit 2 (25%, N = 7).

Pe/ageMales and females with similar coloured dorsum

but females with paler head, face and ventralpelage and paler patagia. The dorsal surface ofmales and females predominantly Drab but tippedwith Olive Brown on shoulders and patchily withFawn Color on lumbar and lower back regions;median dorsal stripe from crown of head to base oftail Olive Brown varying in breadth from 2-5.Hairs in mid dorsal region c. 11 long, basal one­third Drab. On head and behind ears to shoulders,Clay colour in males; in females a paler StrawYellow. On face dark Smoke Gray in males; infemales a paler Smoke Gray. Ventral surface inmales Cream Color on chest and abdomen withneck and flanks Buff Yellow to Buff; in females

23,------------------,

138 D.J. Kitchener, W.e. Packer, A. Suyanto

Nyctimene keasti tozeri Kitchener, subsp. novo

Type localityLorulun (Figure 1) c. 20 km N Saumlaki,

Yamdena Island, Tanimbar group; MalukuTenggara, Indonesia (7°52'S, 131°25'E); altitude 200m. Mist netted at c. 2m height in disturbed vineforest with the following genera of trees prominent:Ficus, Erythrina, Podocarpus, Albizia as well as'Kenari' trees. The understory was.mostly Lantana.There were numerous vines and creepers (seeKitchener and Maryanto 1994: Figure 5).

HolotypeMuseum Zoologicum Bogoriense Number MZB

15904, adult male, weight 29.5 gm, carcase fixed in10% formalin and preserved in 75% ethanol; skulland mandibles separate, liver preserved inultrafreeze at WAM, collected by DJ Kitchenerand R.A. How on 27 Apri11993.

DiagnosisNyctimene keasti tozeri differs from N. k. keasti in

averaging larger in all skull, mandible and externalmeasurements, except for incisive foramen toposterior palate length (PIF) and postorbitalbreadth (Table 1). Palatal length between p4p4lingual surface broader relative to both postorbitalbreadth and PIF (Figures 16 A,B, respectively);braincase breadth generally larger relative topostorbital breadth (Figure 17). It differs also fromN. k. keasti in females having pelage of dorsum asimilar dark Drab colour to that of males ratherthan females with a dorsum of Buff Yellow whichis paler than the male Fawn color dorsum.

56 v. 26.98 (24.67-28.45) 45; braincase breadth 12.64(11.64-13.94) 56 v. 11.93 (11.10-13.06) 45; zygomaticwidth 19.28 (18.11-20.40) 56 v. 18.07 (16.55-19.69)45; CIMl length 9.91 (8.39-10.50) 56 v. 8.82 (8.20­9.51) 45; forearm length 59.7 (55.1-63.0) 56 v. 54.9(49.9-59.2) 48; and ear length 15.3 (14.2-16.9) 56 v.13.2 (11.7-14.7) 48; forearm length generally longerrelative to ear length (Figure 12); incisive foramento posterior palate length, PIF, relatively greaterrelative to postorbital breadth (Figure 10); ClM2

length greater relative to MIMl width (Figure 11).

ParatypesYamdena Island: Desa Sifnana Omele, 7°56'S,

131°18'E, WAM 44398-401 (3 <;> <;>,10); Desa Lorulun(7°52'S, 131°25'E), WAM M(43776, 43782-88, 4394­95,43798,43801-06,43809) (10<;> <;>,8(0); Saumlaki,7°59'S, 131°22'E, WAM M(43511-12, 43555-58,43564-66, 43600) (8 <;> <;>, 10). Selaru Island: DesaAdaut, 8°09'S, 131°08'E, WAM M(44201-Q5, M4408,M44300-Q3, M4431O-11, M44316) (11 <;> <;>,2(0).

5.85.65.45.25.04.8

Plot of zygomatic width versusmesopterygoid fossa breadth for individualNyctimene cephalotes. Symbols as for Figure2.

Mesopterygoid Fossa Breadth

17 +-~-.__~__r-~_,_~-_,__~___,.__~---14.6

Figure 15

18

22

21

J::...~

Si 20;;CleClCl):::JlN

19

DistributionSeram, Ambon and Buru Islands, Maluku

Tenggara, Indonesia. It is unlikely that theNyctimene from Timor Island attributed to N.cephalotes by Anderson (1912), Goodwin (1979) andHill in Corbet and Hill (1992) is of that species. It ismore likely to be associated with N. keasti.

Nyctimene keasti Kitchener, 1993

Nyctimene albiventer keasti Kitchener et al., 1993:408-412

ventral surface a uniform Cream Color. Patagia inmales Olive Brown; in females a paler Fawn Color.Spots on patagia spectrum Yellow and distributedsimilarly to N. cephalotes aplini.

DiagnosisNyctimene keasti differs from N. cephalotes as

described in the earlier diagnosis of N. cephalotes. Itdiffers from N. albiventer in averaging larger in allskull, mandible and external measurements. Forexample, greatest skull length 29.34 (28.04-30.78)

Systematic review of Nyctimene cephalotes and N. albiventer 139

N. k. tozen 5ubsp_ rlOV

S.B

AA

5.8A

AA

.A A

cc.c-'ClI:lQ)~

coQ)-lI:l;;Q,.

5.6

54

5.2

•

A

A +

A

o

'1'Q,.'1'"-cc.c-'ClI:lQ)~

co

'";;;;Q,.

5.6

t A

N. k. tozen 5ubsp_ nov.1 It.

I

54

52

A

N. k. keasti N. k. keastl

12511.55.0+---~---~--~---~------..j

10.550+----~-----r--------_____1

5

A Postorbital Breadth B Incisiue Foramen-Post Palate length

Figure 16 Plot of palate breadth at p4p4 versus (A) postorbital breadth and (8) incisive foramen to posterior palatelength for individual Nyctimene keasti. Symbols as for Figure 2.

Description

Skull, dentary and dentitionIntermediate sized skull with greatest length

29.5±0.56; zygomatic width 19.4±0.45; braincasebreadth 12.7±0.29 and elMl length 10.0±0.34 (Table1). Very similar in shape to Nyctimene cephalotesaplini except that the junction of the postorbitalridges and sagittal crest closer to postorbitalprocess base and frontal area immediately anteriorto this junction is less basined than is the case in N.c. aplini.

ExternalsForearm moderately long 60.1±1.12; ear

moderately long 15.4±0.60. wing membraneusually attached to pes between phalanx 1 of digits2 and 3 (56'Yo, N=25); also commonly attached tophalanx 1 of digit 2 (16%, N=7), and phalanx 1 ofdigit 3 (27'Yo, N=12) and rarely to between digits 3and 4 at level of phalanx 1 (2'X" N=1).

Pelage and patagia colouration very similar to N.c. cephalotes. Namely male and female dorsalsurface of similar colour but with females having apaler head, face and ventral surface.

DistributionYamdena, Salaru and Larat islands, Tanimbar

group, Maluku Tenggara, Indonesia.

EtymologyNamed after Mr Tom Tozer, a distinguished

retired gentleman, who for many years providedhonorary assistance in the mammal section,Western Australian Museum, primarily with thecuration of the Indonesian collection.

Referred specimensThe two specimens from Banda Neira Island,

Band Group were included in the paratype seriesof Nyctimene albiventer keasti Kitchener, 1993 inKitchener et al. (1993). These two specimens aretreated as unallocated in the preceedingmultivariate analysis because they are not readilyattributable to either N. cephalotes or N. keasti. Theymay represent a distinct form of Nyctimene but ifreferrable to a recognisable form of keasti, they areclearly associated with N. k. tozeri. The taxonomicrelationships of the Banda Island form will beclarified when more specimens become availablefor study.

The specimen WAM M34549, adult male, alcoholand skull specimen, collected from Solea, Seram,was included in all the preceeding statisticalanalysis as an unallocated specimen. It has a shortforearm (58.0) and skull and externalmeasurements similar to N. keasti tozeri. It wasclearly allocated to this taxon in the DFA.

140 O.J. Kitchener, W.e. Packer, A. Suyanto

14.,.-----------------,~A

length (see Kitchener et al. 1993: Table 1). p4p4breadth also smaller relative to greatest skull length(see Kitchener et al. 1993: Figure 3a); and ear lengthlonger relative to forearm length (see Kitchener etal. 1993: Figure 4).

13

12

DescriptionA small subspecies of N. albiventer with greatest

skull length 25.68 ± 0.60 and forearm length 51.2 ±1.2 (see Kitchener et al. 1993, Table 1). The overallshape of skull, dentary, and dentition similar to N.cephalotes aplini. Its pelage and colour are similar tothat described earlier for N. c. cephalotes. Wingmembrane attached to pes on phalanx 1 of digit 3(Andersen 1912: 701).

DistributionMorotai, Gilolo, Ternate and Goal area, N.

Halmahera Island.

Nyctimene albiventer papuanus K. Andersen, 1910

Nyctimene papuanus K. Andersen, 1910: 621.

11+----~---___r----~---_i HblotypeThe Natural History Museum, London Number

Postorbital Breadth

10

elMI length

Plot of rostrum height versus e1M1 length forindividual Nyctimene a/biventer subsp. indet.and N. a papuanus. Symbols as for Figure 2.

5.0+----~---__,_-------__1

8

Figure 18

7.0

~

'"'Qj::E:::l.=; 6.0Cl:

Specimens examinedListed in Kitchener et al. (1993)

Figure 17 Plot of braincase breadth versus postorbitalbreadth for individual Nyctimene keasti.Symbols as for Figure 2.

Nyctimene albiventer albiventer (Gray, 1863)

Cynopterus albiventer Gray, 1863: 261-63

Type localityMorotai Island, Maluku Utara, Indonesia (c.

2°30'N, 128°30'E)

HolotypeThe Natural History Museum, London Number

BMNH 62.10.21.5; adult, 'cabinet' skin and skull,collected by Dr. A.R. Wallace in 1861.

DiagnosisNyctimene albiventer albiventer differs from N. a.

papuanus in averaging smaller in all measurementsexcept rostrum length (see Kitchener et al. 1993:Table 1). Forearm length generally smaller relativeto ear length (Kitchener et al. 1993: Fig 4) andpalatal breadth at P"P" generally smaller relative torostrum length (Kitchener et al. 1993: Figure 3b).

If differs from N. albiventer subsp. indet. fromAru Island in averaging smaller in allmeasurements, except rostrum length,mesopterygoid fossa breadth and snout to vent

Systematic review of Nyctimene cephalotes and N. albiventer 141

BMNH 99.12.3.2, adult male, 'cabinet' skin, skulland mandibles separate, collected by AS. Meek on19 March 1899.

Type localityMilne Bay, Papua New Guinea, (ca. 10°20'5,

1500 30'E)

Species examinedListed in Kitchener et al. (1993)

DiagnosisNyctimene albiventer papuanus differs from N. a.

albiventer as described in the earlier diagnosis ofthat species.

It differs from N. albiventer subsp. indet. fromAru in having skull, tooth row and externalmeasurements that generally average smaller.Rostrum height generally less relative to CIMIlength (Figure 18).

DescriptionIntermediate in overall skull, dentary, dental and

external characters between N. a. keasti and N. a.albiventer (for measurements see Kitchener et al.1993: Table 1).

Pelage colour similar to earlier N. c. cephalotesdescription. Wing membrane attached equallycommonly to phalanx 1 of digit 2 (47%, N = 7) or atthis level between digits 2 and 3 (53%, N = 8).Andersen (1912) stated that in the 14 specimens ofpapuanus he examined, this membrane was usuallyattached to phalanx 1 of digit 2 (79%, N = 11), andoccasionally between digits 2 and 3 at level ofphalanx 1 (14%, N=2) and on phalanx 1 of digit 3(7%, N = 1).

DistributionIrian Jaya: Andai, Skiu. Papua New Guinea:

Eaga, Stephansort, S. Coast, Milne Bay; Olsbip,Upper Fly River; near Wewak, E. Sepik; near Baku,Gogol Valley, Madang; Lobota Cave, and Buso,Morobe. Admiralty Island and BismarkArchipelago (Andersen 1912, Hill 1983).

Nyctimene albiventer subsp. indet.

DiagnosisThis is a large form of N. albiventer which has the

overall shape of the skull, dentary, dentition andcolour of pelage of N. a. papuanus (formeasurements see Kitchener et al. 1993: Table 1). Itmay be diagnosed against N. a. albiventer and N. a.papuanus as described in the earlier diagnoses ofthose subspecies.

The wing membrane attaches to the pes usually

on phalanx 1 of digit 2 (62'X), N 10) but alsocommonly at level on phalanx between digits 2 and3 (25%, N = 4), or on phalanx 1 of digit 3 (13%, N2).

DistributionWokam Island, but probably all other islands in

the Aru group.

Specimens examinedListed in Kitchener et al. (1993)

ACKNOWLEDGEMENTS

We gratefully acknowledge the support of Mr ].Bannister, previous Director, Western AustralianMuseum and the current Director Mr A Reeves;Drs M. Amir, Director, Balitbang Zoologi (LIPI);and the Director of the Indonesian Departmentresponsible for the conservation of wildlife inMaluku Tenggara, Ir. ]. Rustandi, who provided uswith assistance in the field, as did a number of hisstaff. To our colleagues Or Ric How and Mr RonJohnstone, Western Australian Museum, and IrIbnu Maryanto, Museum Zoologicum Bogoriense,who participated in the exhausting field work, weextend our thanks for both their effort andcompanionship. Or Ken Aplin and Or R.A How,Western Australian Museum, kindly reviewed themanuscript. We are gratefUl to Or Tim Flannery,Australian Museum for the loan of specimens ofNyctimene from Ternate, Morotai and Halmaheraislands. Expedition costs were defrayed by grantsto D. Kitchener from: National Geographic Society,Washington and Australian National Parks andWildlife Service, Canberra. Some expenses wereborne personally. Garuda Indonesia kindlydefrayed freight costs of the expeditions. Mr C.Keast and Mrs N. Cooper, Western AustralianMuseum removed and cleaned the skulls. Mrs N.Cooper, Western Australian Museum, ran thecomputer analyses. Mrs S. Dalton, WesternAustralian Museum, typed the manuscript.

REFERENCES

Andersen, K. (1910). Ten new fruit bats of the generaNyctimene, Cynoptems and Eonycteris. Annals of theMagazine of Natural History 6: 621-625.

Andersen, K. (1912). Catalogue of the Chiroptera in thecollection of the British Museum. British Museum(Natural History), London. 2nd Ed.

Corbet, G.B. and Hill. ].E. (1992). The mammals of theIndo-Malayan region; a systematic review. NaturalHistory Museum Publications, Oxford UniversityPress.

Geoffroy, E. (1810). Description des rousettes et descephalotes, deux nouveaux genres de la famille des

142

chauve - souris. Annals du Museum Nationald'Histoire Naturelle, Paris 15: 86-108.

Goodwin, RE. (1979). The bats of Timor: systematicsand ecology. Bulletin of tile American Museum ofNatural History 163: 73-122.

Gray, J.E. (1863). Descriptions of some new species ofmammalia. Proceedings of tile Zoological Society ofLondon (1862), 261-63.

Heaney, L.R and Peterson, RL. (1984). A new species ofTube-nosed Fruit Bat (Nyctimene) from NegrosIsland, Philippines (Mammalia: Pteropodidae).Occasional Papers of tile Museum of Zoology, UniversityofMiclligan 708: 1-16.

Hill, J.E. (1983). Bats (Mammalia: Chiroptera) from Indo­Australia. Bulletin of tile Britisll Museum of NaturalHistory (Zoology) 45: 103-208.

Hill, J.E. and Smith, J.D. (1984). Bats, a natural lIistory.

D.J. Kitchener, W.e. Packer, A. Suyanto

British Museum (Natural History), London.Kerr, R (1792). Tile Animal Kingdom; Class 1. Mammalia

1: xviii, 98.Kichener, D.J., Packer, W.e. and Maryanto, I. (1993).

Taxonomic status of Nyctimene (Chiroptera:Pteropodidae) from the Banda, Kai and Aru Islands,Maluku, Indonesia - implications for biogeography.Records of tile Western Australian Museum 16: 399-417.

Kitchener, DJ and Maryanto, I. (1994). A new species ofMelomys (Rodentia, Muridae) from Yamdena Island,Tanimbar group, Eastern Indonesia. Records of tileWestern Australian Museum 17: 43-50.

Pallas, P.5. (1767). Spicilegia zoologica ete. Berlin.Smithe, F.B. (1975). Naturalists' calor guide. American

Museum of Natural History, New York.

Manuscript received 10 January 1994; accepted 15 July 1994.