Digitized by the Internet Archive

in 2012 with funding from

Royal Ontario Museum

http://archive.org/details/variationinbatsoOOpete

%u

5/ty/

!n «.

*CT<

P&

gj-^-

C9

ROM

Life Sciences Occasional Papers

Royal Ontario Museum August 5, 1970

No. 17

ROYAL ONTARIO MUSEUM LIBRARIES

3 1761 05162568 9

Variation in the Bats of the Genus Harpyionycteris, with the Description of a New Race

by R. L. Peterson1 and M. Brock Fenton2

The genus Harpyionycteris, established by Thomas (1896) with H. whiteheadi as the type species, was based on a single specimen of unknown sex (BMNH 97.5.2.7) from Mindoro Island in the Philippine Islands. The holotype was collected by J. Whitehead at 5,000 feet altitude in December, 1895. Thomas (1898) later provided an illustra- tion, in colour, of the type and figured the skull. Miller (1907) was unable to establish any close relationship of this genus, with its multicuspid dentition, with any other known taxa in the family Pteropodidae and pro- posed a distinct subfamily, Harpyionycteri- nae, to contain this most aberrant genus of the family. Andersen (1912) provided an illustration of the holotype skull (Fig. 78, p. 800) as well as excellent details of the dentition (Fig. 79, p. 802).

The holotype remained the only known specimen of the genus until Miller and Hollister (1921) described H. celebensis from a single adult female (USNM 219349) collected by H. C. Raven on 23 August 1917 from Gimpoe, middle Celebes. This new taxon was said to be ... "like Harpyio- nycteris whiteheadi Thomas, of Mindoro, but molars with crowns lower and cusps relatively higher, and PM:! with a conspicu- ous secondary cusp on each side of main

1. Department of Mammalogy, Royal Ontario Museum

2. Department of Biology, Carleton University, Ottawa, and Research Associate, Department of Mammalogy, Royal Ontario Museum

outer cusp" (Miller and Hollister, op. cit., p. 99).

Tate (1951) reviewed the genus and pro- vided data on additional specimens, six from the Buitenzorg Museum collection from Roeroekan, north Celebes, collected at 1,000 metres (3,300 ft.) in January, 1931, by G. Heinrich (Buitenzorg Mus. 2828-33), a male from Tanko Salokko, Mengkoka Mountains, south Celebes (AMNH 153590; examined by us) collected by G. Heinrich in 1932 at 1,500 metres (4,950 ft.), and a female from Negros Island, Philippines (FMNH 66302?). The latter specimen from Mambaho Cave, Pagyabunan, Bais, Negros Island, collected by D. S. Rabor on 13 May 1949, was also listed by Sanborn (1952), who provided detailed measure- ments and a sketch of the palate. The alti- tude of this locality was not stated by the above authors and is not included on the data accompanying the specimen (examined by us). However, judging by maps available to us, the general region appears to be well above 1,000 feet in elevation. A specimen in the Zoologisk Museum, Copenhagen, taken at Kaatoan, Katanglad Volcano at an altitude of about 4,000 feet, Bukidnon Province, Mindanao, and previously re- ported by Sanborn (1953), was examined and the skull restored sufficiently to obtain several measurements. The forearm, hind foot and tibia lengths were ascertained by radiograph.

The collections of the ROM now contain

eight additional specimens from the Philip- pine Islands, including one ( S skin and skull) collected by D. Empeso on 26 May 1967 at Caterman, Mount Mamajao, Lan- goangon, Camiquin Island (off Mindanao, elevation between 2,500 and 3,000 feet), and seven taken by the same collector on 7 and 8 January 1968 between 500 and 1,000 feet at Balanan, Saiton, Negros Island. The latter series, together with FMNH 66302 9 , represents an age spectrum from subadult to middle age and provides an opportunity to detail certain changes that take place in the skull during the aging process of H. white- headi.

Cranial Variation with Age — This series from Negros Island, represented by three females and five males, is not sufficiently large to detect any consistent cranial charac- teristics to distinguish the sexes. An attempt was made to arrange the series in an age- developmental sequence from youngest to oldest, and certain characteristics are sum- marized in Table I. From this series, six have been selected to show some of the more apparent aging characteristics illus- trated in Fig. 1. The change in the angle of deflection of the occiput, from a strong de- flection as shown in the two youngest indi- viduals, to a more moderate one in the older specimens, is particularly noteworthy, as this angle of deflection has been used as a taxonomic character in several pteropodid genera.

The fusion of sutures follows a fairly consistent pattern if our sequence has been correctly arranged by age (see Table I). There is a general increase in size with age (as expressed by forearm length, condylo- basal length, zygomatic and mastoid breadths), although, as might be expected, these increases in size are not entirely con- cordant with the sequence of fusion of sutures.

An inverse ratio is apparent in the width of the postorbital constriction. There is an obvious decrease in width with age, from 6.8 mm in the subadult to 5.3 mm in the mature adult. This increased constriction with age is also coincident with changes in the fronto-parietal region and the develop- ment of a sagittal crest (Fig. 1).

Specimens FMNH 66302 9 (not illus-

trated) and ROM 46146 9 agreed well in age characteristics. Specimen ROM 46148 <5 (not illustrated in Fig. 1), similar in age to ROM 461509, was the most narrow in zygomatic breadth and equal to the most narrow in mastoid breadth of the series. It had an aberrant occiput with a malformed occipital condyle, and this deformity prob- ably accounts for the aberrant width meas- urements in the affected areas of the skull. It is likely that the age span in our younger series (all but specimen ROM 46149 <5 ) represents a fairly short period of time (perhaps only a few weeks) beyond the juvenile stage, by which time they have become volant and independent but have not yet reached full maturity. The three youngest specimens could be classified as subadults and are characterized as follows: occiput strongly deflected, fronto-parietal and basisphenoid-basioccipital sutures un- fused, and metacarpo-phalangeal joints no- ticeably swollen and cartilaginous. The absence of any appreciable wear on the dentition of specimen ROM 46149 $ sug- gests that even this individual was no more than middle age (possibly no more than two or three years old). Of other known specimens, the holotype of H. whiteheadi appears to be comparable to ROM 46149 $ in age characteristics, perhaps even slightly younger. Specimen ROM 43669 S from Mindanao exhibits the same suture-fusion characteristics as 46149, except that the jugal-temporal suture is only partially fused, and the frontal-parietal ridges are not fused into the sagittal crest as far anteriorly. This specimen was judged on cranial characters to be slightly younger, and when the skin was subsequently examined by radiograph the metacarpo-phalangeal joints clearly had cartilaginous pads not yet completely fused to the bone ends as in 46149 c5 . The speci- men from Mindanao in the Zoologisk Mu- seum, Copenhagen, ZM 23719, was an old individual with well worn teeth (particu- larly the molars) and a well developed sagittal crest. The appearance of the mam- mae suggests that it may have been lactating at the time of collection. The articulating ends of the metacarpo-phalangeal joint are completely ossified (verified by X-ray). The specimen AMNH 153590 c? from the Celebes is the oldest specimen seen by us and

- -

3

e

■3

-a

I

5

cu M

cs

8

"3

M

O

o

f

o E

M e cs

3 ^ ov

< Pi VO

3 MO

o

3 Pi VO o "i"

"b

3

MO 2; 3 Ph vo

o ■*

O "*

3_ ^b

T3 ^ T-H

cS O "^

"§QiS C/3 ^

3 EG

o

OS

r-^o-Hooroo — oo

Ph Ph

^ Tf rt (N *-. — i -h

o

£ft

t}- tj- cj — — < —

Tfnt^M-rtTtMOO

fe fc £ O O O O O d o\ ri Th vd d vo «o

■* ro <N ^h -h — '

— i <n o o m \o n on

fe ^ -8 & & & O O O O O - oVi vi vo" d vo vd

HH HH HH MH . ^ ^ ^

P_ PL| O Ph Ph

vo-^oor~-oom — ro

mhhHqI-hm T+mrj— -h -h

Ph Uh

inr^<-^mvo>oow->

£ £ ° O O O O O O O O - d <n Tt vd © vd vd

^ ^ £2 •* tJ- <N — _h —

oooo^h\d>tivooooo

fco gOOOOOOOO— 'd-n'^vo'd'ovd

g1 m^ON^fOOVlN'H

ft O o O O O O O O O O ©" oo r-i Tf vd d vo vd

& <*> m

3 3

o

O cS

is *->

o> 70

o o

.2 <»

en O)

CS H

03 2

•a 3

E E,

■a a

c

0

O

CS C

C 3

550C

O —

a 2 2 E

— - 'Lo o

CS cs

CS CS

1- 4) 3

CS I S

'* "3 S

« 2? £

0) In 3 in CS	O c 0 3	3 oa	6b	X!
& CO U — P-	0 a	c 0		■a CS u
'3 CS E	T3 '5 3	CO		U CS E	-0 0
0		1)	a	0 M	CO
p 0	'3	OunS
ft, PQ

5 _ 'C X! ^3

o a -

'^1

M

C

IN

o

«n £ I

£s

CS

u

3

z

II

x z

Ph

O

oi

o Pi

3 C/3

&- 2

■a 3 ^^

a I

cs .3

Ph ^

11 _y

Uh-3

P. C

o o

"cS co .3

£.£■

3 CJ

3 8

CO

a S

SMSff^^^JSESCr " f"°m ~" ~

is characterized by excessive wear on all teeth (Fig. 4). In other respects it exhibits the same age characteristics as specimen ROM 46149 6 from Negros Island. Tate (1951) did not see this skull, as he listed the specimen as a skin only. Buitenzorg Museum specimen 2833, illustrated by Tate (1951), is obviously a subadult. Judging by the measurements of the northern Celebes specimens given by him, one was a juvenile, and the other five were all probably sub- adults. Almost certainly none were older than young adult.

Wing Variation with Age — Measurements of the wing elements of the Negros series are summarized in Table II. The metacarpo- phalangeal joints are normally quite swollen in appearance in many pteropodid bats, but the joints of three youngest alcoholic speci- mens (ROM 461469, 66302 9 and 46151 S ) are obviously less ossified than those of older individuals when examined externally. Radiographs were taken on the entire Negros series, and the swollen carti- laginous pads are clearly evident in the three subadults, somewhat less swollen but clearly distinct in all four young adults and absent in the adult. The length of the fore- arm varies from 82 to 86.5 mm in the sub- adult-young adults, as compared to 89.5 mm in the mature adult. The length of the hu- merus, as determined by radiographs, is fairly consistent in the Negros series. The subadults and young adults vary from 51.2 to 55.0 mm, whereas the adult is 56.0 mm. It is apparent that the thumb with claw (pollex) and second digit elements (with claw) reach maximum length early in life, and no further growth is obvious in this series. Moreover, there is little change in the length of the elements of the third, fourth and fifth digits in the younger series. The mature adult, however, has all of these ele- ments consistently longer. Of the six north Celebes specimens measured by Tate (1951), one was a juvenile, and the other five, thought by us to be subadults, had fore- arms varying from 79 to 84 mm in length, whereas the holotype of H. celebensis was listed as 90 mm and the old adult from south Celebes as 91 mm (92.5 by us). These measurements are remarkably consistent with those of our Negros series. The humerus

of the south Celebes specimen (AMNH 153590 $ ) is 58.5 mm long compared with 56.0 in the adult from Negros (ROM 46 149 6; both measured by radiograph).

Variation in Hind Limbs — The genus Har- pyionycteris is characterized by a short tibia. Unfortunately, precise measurements are difficult to obtain by ordinary methods in either study skins or preserved specimens. Radiographs taken with the feet and tibias positioned as near horizontal as possible yield much greater accuracy. The Negros series is notably consistent in lengths of both hind feet and tibias, and both are virtually identical in length, varying from 22.0 to 23.5 mm in both cases. It is obvious that these members, like the thumb or first digit of the wing, reach maximum size early in life with no apparent increase beyond the subadult age.

The femur is only slightly shorter than the tibia in five of the Negros series (properly exposed for measuring), varying from 21.0 to 22.0 mm. The Camiquin specimen (ROM 43669 6 ) measures 21.6 mm and the Mindanao specimen (ZM 2371 9 ) measures 22.0 mm. By contrast the Celebes specimen (AMNH 153590 & ) has a femur length of 25.0 and a tibia length of 26.0 mm, both above the upper limits observed in Philippine specimens. The holotype of H. celebensis is said to have a tibia length of 30 and hindfoot length of 29 mm. Tate (1951) shows measurements of AMNH 153590 <$ from south Celebes as tibia, 31, and hindfoot, 22 mm, whereas by X-ray technique these prove to be 26 mm in both cases. It seems likely that, by using our X-ray technique, these measurements for the holotype will prove to be somewhat less and to be closer to those of the south Celebes specimen.

Individual Variation in Dentition — The cusp and dental patterns of mammal teeth tend to be relatively stable elements and have long been relied on for species discrimina- tion, especially in fossils. The only basic distinguishing character ascribed to H. cele- bensis by the original authors was the de- velopment of an extra cusp on P3 and the greater height development of the cusps in general. The series of eight specimens from

w

CO

c

o Z

E

o

5'

•2

E

—

c S

C

o

^ OS

IS

o-

2-°

£g

lOOOOO'^l'^OO'OOOvOTl-iy-iO

(N <N <

O mir>ir>0"~>'/~iOOO''~>w">©'/i©©©

so l>/-)©Or«->©rit--oo</->-'3-0>0\0\<Nror'> 00 (mt«-\0^«lflmmin(NMN(NM

inOO'OOOOOO'^iOi^iOTTO

o o

\0ir>>/-)0— '<-<"> — Tft--t--sOi/-)O\0O<N

o o

T]--r)-Tj-oo©(N©Tl->ot--i/-)>/-iooCN©

©r4©0©©'/l©>/->i/->Ouiir>'/'>w"}rl©©

OOOmO\>oOOO

CA

&M

CO to r-

"O 3 „ 13 u „ T3 D « 'O S „ 13 S *Z •— rN «=S — OJ »S ^* <N «S — <N

■SSllSl-s

C3 03 r^ ra C3 r-

' -^ 0. ^Q. 03 Ph

c 2 eu

is "3 rt

03 CT3 j—

2p-

E ?

il-a

0 M -^ c

'5>

'q

on 5

'oo S

INN C

3 O

O — nr~o<Noorn-'3-voi/->!/-)r^ooo — n n oo oo in r»i(n- — * *n -*t w~t v~i en m u~i c\ m <^i (H c<\ c

O

-^a\<N"~>©>'->i/->oooo>/->>oi/-)0>/")00

«1 — \000<Nr-4as-^tTj-r-\0>00\000\r|r^ro

"O

3 1/3

000000003 T3

3

* °

R £ S Si

<d r= -j= M

U* H £ <

2 8?

a

o o I

X

46146

66302

46151

46152

46147

46150

46148

46149

Figure 2

Variation in the cusp pattern of M1 and M- of Harpyionycteris whiteheadi from Negros Island, Philippine Islands. No. 66302 is from Mambaho Cave, Pagyabunan, Bais (FMNH); all others are from Balanan, Saiton (ROM ). See Table I for sex and age development of each specimen.

Negros affords an opportunity to assess the range of individual variation in dental pat- tern. A careful examination of M1 and M- shows the cusp pattern to be extremely variable (see Fig. 2). In fact, no two of the eight specimens can be said to be even close to identical. The basic pattern of six cusps as defined by Andersen (1912, pp. 801-3) is altered in this genus by the addition of accessory cusps or by the division of primary cusps into two. Not only do the cusps vary in size, position and number in each indi- vidual, but there is also a comparable amount of difference in the left and right teeth of individuals. Minor variations are apparent in the other teeth as well. None show even a trace of the secondary cusp on the buccal side of P3 seen in H. celehensis. Tate (1951) was in error when he stated that the Negros Island specimen (FMNH 66302 9 ) had the supplemental cuspule on P2 [= P3] as in celehensis. The known ex- amples of that taxon, as well as the holotype

of H. whiteheadi and ROM 43669 & from off Mindanao, have a well-developed an- terior supplemental cuspule on P:!. In the Negros series this cuspule is absent from three specimens and weakly developed in five, none being as distinct as in the above examples. The single pair of lower incisors are quite small and probably non-functional. Of the specimens seen by us, the holotype of H. whiteheadi, FMNH 66302 9 and ROM 46151 £ have only one incisor (on one side only) present, and all others have none.

Although individual variation in dental patterns is normal and to be expected, as in many species of mammals, the variability in the cusp patterns exhibited by Harpyio- nycteris is unique among bats and must certainly rank high among all mammals.

Adaptive Significance of Dentition — There has been some speculation as to the affinities of Harpyionycteris with other bats. Tate

(1951) summarized this question by sug- gesting that it ..."undoubtedly branched off from the pteropid stem in very ancient times. ... and that the multicuspid state of the molars, and indeed of the whole dentition, of Harpyionycteris must be regarded as secondary rather than as a surviving example representing a formerly widespread con- dition in the Megachiroptera." If Harpyio- nycteris has as old a history as has been suggested by Tate, why the cusp pattern has not become genetically fixed within much narrower limits, as we have seen in most mammals, poses a most interesting question. The adaptive significance of the multicuspid condition in the genus remains a mystery. To our knowledge its food habits are still unknown. An examination of the digestive tracts of the ROM specimens showed them all to be empty of any diagnostic food items. The meagre evidence from our examination suggests animal matter is not a prime food because of the absence of insect chitin. bone, scales or hair. We can only speculate that, like most of its allies, it is a fruit eater, perhaps adapted for a particular type of tough-textured fruit for which the multicusp teeth would be advantageous in extracting the juice. Possibly the fruit fibres are dis- carded and not normally ingested - this might explain the absence of diagnostic material in the digestive tracts! This sug- gests that Harpyionycteris may represent yet another line of highly specialized feeders - fruit juice-feeding bats.

Geographic Variation — The known range of the genus Harpyionycteris, based on 18 specimens, extends from Mindoro ( 1 speci- men) in the north, southward through Negros Island (8 specimens), Camiquin Island, offshore from Mindanao ( 1 speci- men), Mindanao Island ( 1 specimen) to the Celebes, including the north (6 specimens), middle ( 1 specimen) and south portions of the island (1 specimen) (Fig. 3). Of this total, only six are mature adults that can

be critically compared, and these have been taken from five separate areas (Table III; Fig. 4).

Mature specimens from the five localities are approximately similar in head and body length. The two Celebes adult specimens have the longest tibia, forearm and thumb, while one has the maximum lengths of femur and humerus. Of the other wing elements, all seem to be of approximately the same mag- nitude, although the Mindoro, Camiquin and Mindanao specimens tend to be slightly smaller than the others. The Negros speci- men has the longest metacarpals on the third, fourth and fifth digits.

In greatest skull length, condylobasal length and palatal length, the six adults are remarkably similar. In the measurements of breadth, the Negros Island specimen has the narrowest zygomatic, mastoid, brain- case, M2-M2, P3-P3, interorbital, postorbi- tal and upper C-C measurements. The Mindoro, Camiquin and Mindanao speci- mens are strikingly similar to each other in most skull measurements and tend to be either equal to or smaller in breadth mea- surements than the Celebes specimens. Maximum size is found in the latter in the mastoid, braincase, P3-P:\ interorbital and the upper C-C breadth measurements.

In the length of the lower jaw, all adults are similar. However, the Negros specimen has the lowest height of ascending ramus and is lighter in build, having the least depth between P4 and Mi as well as behind M3. In the dentition, the length of C-M- and C-M.i tend to be similar in all adults, although there are a number of variations in the conformation of the individual teeth.

The Camiquin specimen (ROM 43669 $ ) has been compared with the holotype of H. whitehead i and agrees remarkably well in all essential details of both the skull and the skin (including colour of pelage). Even in the highly variable M- the cusp pattern agrees well, with only a better development of the posterior buccal cusp (pc, of Ander-

Figure 3

Map of the Philippines-Celebes area showing the distribution of the known specimens of Harpyionycteris. 1. Mindoro Island (type locality of H. whiteheadi whiteheadi). 2. Mambaho Cave, Pagyabunan, Bais, Negros Island. 3. Balanan, Saiton, Negros Island (type locality of H. whiteheadi negrosensis) . 4. Cater- man, Mount Mamajao, Langoangon, Camiquin Island. 5. Kaatoan, Katanglad Volcano, Bakidnon, Min- danao. 6. Gimpoe, middle Celebes (type locality of//, celebensis). 7. Tanko Salokko, Mengkoka Moun- tains, south Celebes. 8. Roeroekan. north Celebes.

TABLE III

Measurements (in mm) of the known full adult specimens of the genus Harpyionycteris: 1. H. whiteheadi

whitehead! 2. H. w. negrosensis 3. H. celebensis. AMNH = American Museum of Natural History;

BMNH = British Museum (Natural History); ROM = Royal Ontario Museum; USNM = United States

National Museum; ZM = Zoologisk Museum, Copenhagen.

		1.		2.		3.
					South	Middle
	Mindoro IV	indanao	Camiquin	Negros	Celebes	Celebes
	BMNH	ZM	ROM	ROM	AMNH	USNM
	97.5.2.7	2371 9	43669 d*	46149 d"	153590 &	219349 9
	Holotype			Holotype		Holotype
Total length	140		145	152	143	* 153
Hindfoot	23	24t	23. 5t	23. 5t	26t	29(?)
Tibia	24.5	24t	23. 5t	23. 5t	26t	30(?)
Forearm	82.5	84	85	89.5	92.5	90
Thumb with claw	33.0	31±	33.7	35.0	36.1	39.0
Second Digit: Metacarpal	36.5	38.2	40.0	41.0	43.7	—
1st Phalanx	11.5	8.0	10.0	10.0	11.7	—
2nd Phalanx	13.0	11.5	11.3	12.5	12.0	—
Third Digit: Metacarpal	59.0	60 ±	60.0	64.5	64.4	63.0
1st Phalanx	43.5	44.0	43.1	46.0	50.2	—
2nd Phalanx	54.0	59.5	55.0	62.0	58.0	—
Fourth Digit: Metacarpal	54.5	56 ±	57.0	62.5	60.7	—
1st Phalanx	35.5	36.6	36.0	36.8	42.5	—
2nd Phalanx	36.0	36.7	36.5	39.6	37.3	—
Fifth Digit: Metacarpal	55.5	54 ±	59.0	64.4	61.0	—
1st Phalanx	29.0	28.5	27.9	30.5	32.6	—
2nd Phalanx	29.0	30.8	30.5	34.0	35.4	—
Skull: Greatest length	43. 0±	43 ±	43.1	43.7	43. 5±	43.0
Condylobasal length	40. 0±	41 ±	41.7	42.6	42. 0±	41.6
Palatal length	23.0	23 ±	22.5	22.3	23. 5±	—
Zygomatic breadth	24.0	—	24.3	23.1	24.3	24.0
Mastoid breadth	15.6	15. 5±	15.7	14.8	16.0	—
Braincase breadth	16.7	16.5 ±	16.5	15.7	16.9	—
M2 — M2 breadth	11.6	—	11.2	11.0	11.6	—
P3 _ p3 breadth	9.2	—	9.3	8.9	9.5	—
Interorbital breadth	6.5	6.4±	6.9	6.1	7.0	7.0
Postorbital constriction	6.0	5.9±	6.4	5.3	5.6	6.2
Canine — Canine (upper)	7.8	—	7.8	7.4	7.9	—
Height of braincase	13.5	14.6	14.0	13.4	13.6	—
C — M2 length	17.8	17.5	17.6	17.8	17.5	16.6
Length of mandible	35.0	34. 8±	33.1	34.0	34.0	35.0
Height of ascending ramus	16.5	16.8	16.3	15.8	16.1	—
Depth of mandible between
P4 and Mi	4.4	4.4	4.4	3.7	4.1	—
Depth of mandible behind M3	5.6	5.6	6.1	5.0	5.4	—
C — M3 length	19.2	19.3 ±	18.8	19.4	18.9	17.8

*vide Miller and Hollister (1921) and Tate (1951)

fdetermined by X-ray

10

Figure 4

Dorsal and ventral views of skulls of Harpy ionycteris.

Left: H. whiteheadi whitehead! ROM 43669 <£ from Camiquin Island, off Mindanao Island.

Centre: H. whiteheadi negrosensis subsp. nov. Holotype ROM 46149 c? from Balanan, Saiton,

Negros Island, Philippine Islands.

Right: H. celebensis AMNH 15390 <S from Tanko Salokko. Mengkoka Mountains, South Celebes.

11

sen, 1912). The palatal ridges of the Cami- quin specimen (Fig. 4), however, do not conform with the illustrations of the Negros Island specimen given by Sanborn (1952). Even though there are minor individual dif- ferences among our specimens, all, including our Negros series, are similar in having five undivided anterior ridges and three pairs of divided posterior ridges, the latter exhibiting the greatest variation in length, shape and position. There is also an undivided ridge near the posterior edge of the palate in all the specimens.

Figure 5

Details of the face of H. w. negrosensis ROM 46150, young adult female, prior to extraction of skull. Note lateral deflection of the nostrils.

The Negros Island population appears to have been isolated a sufficient length of time to have developed distinctive characteristics consistent with subspecific variation and may be known as

Harpyionycteris whiteheadi negrosensis, subsp. nov.

Holotype— ROM No. 46149 adult male preserved in alcohol with skull extracted; collected on 7 January 1968 between 500 and 1 ,000 feet elevation at Balanan, Siaton, Negros Island, Philippines, by D. E. Em- peso (see Figs. 1,2,4, and 6).

Range — Known only from Negros Island, Philippine Islands.

12

Diagnosis — Allied with H. whiteheadi whiteheadi and of the same general size but with narrower skull proportions, including zygomatic breadth, mastoid breadth, rostrum breadth (C-C and P3-P3), interorbital breadth, and breadth of braincase. Nasal aperture, viewed from the front, noticeably higher than deep rather than as wide or wider than deep. Mandibles are lighter in build, shallower at the junction of P4 and M: and behind M3 (see Table III). The last upper molars (M2) lie much closer to the shelf of the zygomatic plate.

Measurements — See Tables I, II and III.

Comparisons — The pelage coloration is un- known from fresh skins, but preserved speci- mens, including the holotype, ROM 46 1 50 2 , and one skin prepared from a preserved specimen (ROM 46148 c?) are darker than H. w. whiteheadi, with the underparts approaching Bister (Ridgway, 1912) rather than Buffy Brown or Dresden Brown. The dorsum is more rusty coloured, ap- proaching Angus Brown or Brussels Brown rather than Sepia or Proufs Brown. In general the fur appears to be somewhat longer and thicker, both above and below. Light spots are also present on the wings but are more pinkish than yellowish in colour; these as well as pelage characters may be of limited taxonomic significance.

The darker pelage coloration is closer to H. celebensis as represented by AMNH 153590 c! , but the length of the fur, both above and below, is noticeably shorter and is more nearly unicolour on the middle of the back, rather than bicolour, with a much paler base.

The ears of H. w. negrosensis are nar- rower and more pointed than either H. w whiteheadi or H. celebensis.

The great variability in the dental cusps, as illustrated in the Negros Island series suggests that caution should be exercised in using dental characters in this genus for taxonomic discrimination. The well-de veloped extra cusp on P3 found in H. cele bensis has not been observed in any Philip pine specimen and may prove to be a ge netically fixed character of taxonomic value Examination of AMNH 153590^ and th illustrations provided by Tate (1951) sug

gest the possibility that the distinctive shape and angle of projection of the upper canine may be equally significant, if not more so, to distinguish H. celebensis from H. white- headi (see Fig. 6). In H. celebensis the upper canine is relatively shorter and broader with a less anterior angle of for- ward projection and has a much wider posterior cusp (viewed from the side).

Systematic Discussion — The past paucity of specimens of the genus Harpyionycteris has precluded any meaningful analysis of taxo- nomic characters and evolutionary trends. The addition of eight new specimens helps in an assessment of some trends but still falls far short of an adequate appraisal. If our assignment is correct, there are still only three examples of the nominate taxon now known, the holotype from Mindoro, the specimen from Mindanao (UZ 2371 9 ) and our specimen from Camiquin. These repre- sent two widely separated localities lying both north and south of Negros Island but apparently separated by smaller water gaps from each other than from Negros Island. This presupposes that ultimately this species will be found in suitable habitat, should it exist, on the islands of Leyte, Samar and southern Luzon. If our suspicion concerning its food habits is confirmed, then its distri- bution may well be correlated with a specific type of fruiting plant, probably occurring primarily at higher elevations (about 1,000 feet or more).

The systematic relationships of the known populations of Harpyionycteris pose interest- ing problems common to other insular taxa. Tate (1951) suggested that the Celebes population probably represents only a sub- species of H. whiteheadi, a view also ac- cepted by Laurie and Hill (1954). Such a conclusion would presuppose that there is, or has been in the recent past, some gene flow between these insular populations. If it is assumed that the species exists throughout Mindanao, the distance between the south- ernmost point of land and northernmost Celebes would be about 250 miles. Although there are small islands between that might serve as "stepping stones," there are still nearly 100 miles of open water to be trav- ersed— a distance usually more than ade- quate to serve as an effective barrier with

most species of bats. If the diet of these bats proves to be a specialized one requiring high altitude species of fruit, the smaller islands would be of limited value as "step-

Figure 6

Diagram comparing the shapes of the rostrums and dentitions of H. w. whiteheadi (ROM 43669c? ), top; H. w. negrosensis, Holotype, centre; H. celebensis (AMNH 153590 c? ), bottom.

13

ping stones" between the widely separated populations of Harpyionycteris. Celebes lies east of Wallace's Line and its fauna normally has been associated with the Australian zoogeographic region, whereas the Phillip- pines lie west (and north) of Wallace's Line and its fauna is more related to that of Borneo as an insular extension of the Orien- tal zoogeographic region. It is of interest that no Harpyionycteris has been recorded from Borneo, which is separated from nearest contact with Celebes by only about 75 miles. The apparent isolation of the population on Negros Island, with only 25 or 30 miles of water separating them from possible land contact with other populations, suggests to us that any normal contact between the Celebes and Philippine populations is un- likely. Although dental characters are highly variable in this genus, the distinctive shapes of the upper canines and P3 repre- sent morphological differences that would normally be regarded as greater than a sub- specific level of variation in any other species of bat. In the absence of any specific evi- dence to the contrary, we therefore take the view that the Celebes population probably has been isolated from the Philippine ones for a considerable period of time and is suf- ficiently distinct to be regarded as a full species.

Acknowledgments — We wish to acknow- ledge the continued co-operation of Mr.

Domingo Empeso for conducting systematic collection of bats in the Philippines which provided the ROM specimens here reported on. Co-operation and assistance was pro- vided by the following institutions and indi- viduals for loan of specimens or study of material in their care: British Museum (Natural History), G. B. Corbet and John Edward Hill; American Museum of Natural History, Richard G. Van Gelder and Karl F. Koopman; Field Museum of Natural His- tory, Joseph C. Moore; Zoologisk Museum, Copenhagen, F. W. Braestrup. Line draw- ings were prepared by Paul Geraghty and Mrs. Sophia Poray. The photograph of the face (Fig. 5) was made by Lee Warren. The entire staff of the Department of Mammal- ogy of the ROM deserve our special thanks for continued assistance and co-operation, and especially J. R. Tamsitt for editorial council. The radiographs used in this study were provided by the Laboratory of Sys- tematic and Evolutionary Zoology of the University of Toronto established in the Royal Ontario Museum with National Re- search Council of Canada Negotiated Development Grant funds. This study, aris- ing out of our bat research program, has had the financial assistance of the Canadian National Sportsmen's Show, the Pure and Applied Sciences Committee of the Uni- versity of Toronto through the Department of Zoology, and the National Research Council of Canada, as well as the Royal Ontario Museum.

Literature Cited

Andersen. Knud

1912 Catalogue of the Chiroptera in the collection of the British Museum. 2d. ed. Vol. 1: Megachiroptera. London, Brit. Mus. (Nat. Hist.), pp. 1-854. Laurie, Eleanor M. O. and J. E. Hill

1954 List of land mammals of New Guinea, Celebes and adjacent islands, 1758-1952 London, Brit. Mus. (Nat. Hist.), pp. 1-1 75. Miller, GerritS., Jr.

1907 The families and genera of bats. Bull. U. S. Natn. Mus.. no. 57, pp. 1-282. Miller, Gerrit S., Jr. and N. Hollister

1 92 1 Twenty new mammals collected by H. C. Raven in Celebes. Proc. Biol. Soc. Wash. vol. 34, pp. 93-104. Ridgway, Robert

1912 Color standards and color nomenclature. Washington, D.C., Published by the author pp. 1-44, 53 pis.

14

Sanborn, Colin C.

1952 Philippine Zoological Expedition, 1946-1947. Mammals. Fieldiana: Zool., vol. 33, no. 2, pp. 87-158.

1953 Mammals from Mindanao, Philippine Islands, collected by the Danish Philippine Expedition, 1951-1952. Vidensk. Meddr. Dansk Naturh. Foren., vol. 115, pp. 283-288.

Tate, G. H. H.

1951 Harpyionycteris, a genus of rare fruit bats. Am. Mus. Novit. no. 1522, pp. 1-9. Thomas, Oldfield

1 896 On the mammals from Celebes, Borneo, and the Philippines, recently received at the

British Museum. Ann. Mag. Nat. Hist., vol. 18, ser. 6, pp. 241-250. 1 898 On the mammals obtained by Mr. John Whitehead during his recent expedition to the Philippines, with field-notes by the collector. Trans. Zool. Soc. Lond., vol. 14, pt. 6, pp. 377-412.

15

*">3S,

TllAi

OlijV

J.M0

^^8/7