PertanikaJ. Trap. Agric. Sci. 21(1): 23 - 28(1998) ISS T: 1511-3701© Uniyersiti PlItra l\Ialaysia Press

Plantain Squirrel Callosciurus notatus In a Plantation Habitat

M.N. HAFIDZI

Department of Plant Protection

Faculty oj Agriculture

Universiti Putra Malaysia

43400 UPM, Serdang, Selangor~ Malaysia

Keywords: actogt"am, activity pattern, Callosciurus notatus, oil palm plantation

ABSTRAK

Corak aktiviti /wrian atau aktogram Callosciurus notatus dalam ladaJlg I?elapa sawit daJl /<0/(0 di bawalikelapa ditentu/wll dengan mengguna!wn radiotelemetli . Hasil meJ11l'J1ju!dwJl C. notatus mempamerkaJl comk

al?tiviti harian dari jam 0630 hingga 1730 del/gan beberapa pU'llcak aktiviti. Dua Plillcak lltama di/wwl pastiiaitu pada awaL pagi (jam 0730) dan sebeLu1n tengah hari (jam 1130). Ini dii/wti dengall fasa relwt ),ang relatifpada waktu tengah hori (1430 hingga 1630). puncal? yang lebih kecil dicemp pada waMu pelang UalJl. 1730).KepeLbagaiall IJ/,ung/<in wujud I?erana satu individ'll mellllnjllkkan elllpal pUJlcak yang diselallg seli dengall

beberapa fasa rehat yang seng/wt. Puncak aktiviti yang kerap lII'Ll1lgkin lIIeJlggalllbar!wn strategi pengge/inlaranuntuk m,emeuuhi keperLuan maharwn yang terlent'll. KajiaJl isi perut 'Inelllllljll!dwn serallgga lIIerujJal?all

lIIe/wnan utama pada awaL pagi tetapi menunjul?/wu lIIa/wnan yang pe/bagai lIIeliplili seraJlgga daJl bahall­

bahan tumbuhan selepasnJa.

ABSTRACT

The dail)' activit)' pattern or actogram offour Callosciurus notatus in a mixed plaJllalioJl ClDP of oil j){tlm andcacao under coconut in Peninsula-r Malaysia was deterlllined by using radiotelelllelr)'. Resulls suggest that C.notatus exhibits a daiLy activity pattern from 0630 to 1730 with several pea/<s. Two major pea!<s are identified,one at daybreak (abollt 0730) and another before midday (about 1130). 77le two peaks were followed by a relative

ressation of activity in the aflemoon (1430 - 1630). a third smaller peak was recorded in the eveniJlg (1730).Individual valiation exists as one individual showed four small peaks interspersed by short periods of resl. The

multiple fJeaks may -reflect the foraging strategy in meeting specific dietm)' needs. A slllvey of stomach contentssuggests that insects constitute the preferred item in the first couple of hours and a mixed diet of illsects and plant

matter later during the day.

INTRODUCTION

Rats (Rattus spP). and the plantain squirrel(CallosciurLls notatus) are important vertebratepests of many plantation crops in Malaysia (forpests of cacao, see Kamarudin and Lee 1981;Hafldzi 1992). To gain a better understandingof the foraging behaviour of plantain squirrels,the present study investigated their activitypattern. Such a study may contribute to moreeffective control programmes.

Several techniques have been used instudying the activity pattern of small mammals.Among these are unaided sighting (Shorten

1962), time-scheduled trapping (Kamarudin1982) and radio telemetry (Tonkin 1983). \i\Thileunaided sighting causes the least disturbance, itmay not be suitable for less conspicuous animalslike especially in a habitat characterized by densefoliage. Live-trapping techniques are suitable forstudying population dynamics but will reveallittle of the ranging behaviour as animals areimmobilized at points of capture. Radiotelemetrywas employed in this study as one clear advantageof this technique is that it allows continuousmonitoring of animals with minimal disruptionof their activities in their natural environment.

M.N. HAFIDZI

MATERIALS AND METHODS

Radio-tracking Equipment

Basic radio-tracking equipment was used;consisting of radio collars, a portable radioreceiver (Model M57, Mariner Radio, UK) and athree-element Vagi antenna. Each radio-collarwas made up of a SS-1 transmitter (Biotrack UKLtd.), a 1.5-V battery and an external antennamounted on a cable tie collar with self-lockingratchet. The transmi tter and battery were bothcoated in epoxy for protection against gnawingand water-proofing. The whole package weighedapproximately 10 g. All animals fitted with radiosweighed more than 200 g so that the transmitter­animal weight percentage did not exceed 5%. Agreater weight probably affects behaviour andmovement of animals (Wolton and Trowbridge1985; Pouliquen et al. 1990)

Study Site and Radio-tracking TechniquesThe study area was a mixed plantation of matureoil palm and cacao planted under coconut. Thesquirrels were trapped using ordinary trap cagesset in a row on every fourth cacao treeapproximately 10m apart. Trapping was carriedout at the edge of the cacao near the cacao-oilpalm boundaries. A previous trapping exerciseshowed most captures were recorded at intercropboundaries. Squirrels were identified individuallyby radio transmitters operated at uniquefrequencies. The sex, weight, breeding conditionand capture locations were recorded for eachsquirrel (Table 1). Radio collars were attachedwhile the animals were lightly anaesthetized forapproximately 60 seconds with either diethylether or chloroform. The radio-tagged squirrelswere kept in a cage overnight, allowed to recoverfull locomotor activity, and released at theirrespective site of capture. This measure wasnecessary to ensure that the radio collars werefully secured and did not physically harm theanimal.

A total of six squirrels (three males andthree females) were radio-tracked for up to sevendays. The tracking period for each animal isshown in Table 1. Location was determinedevery hour from 0530 to 1730 h eitherthroughout or for part of the period. Trackingwas not done on rainy days when animals seemedless active. Radio signals were detectable from atleast 100 m awayand were pin-pointed by walkingalong the path of the strongest signal. Apparentanimal location was determined to be accuratewithin 1 - 2 metres as confirmed by actualsighting. The presence of coconut and oil palmsgrown at 15-m intenrals across the study areaconveniently senred as grid markers to' markanimal locations on the map. When radiolocation was uncertain, triangulation (Kenward1987) was employed; this involves taking bearingsfro~n at least three different points. The point atwhIch the bearings intersect was designated asthe animal location. The distance betweensuccessive radio locations was taken to representthe level of activity during a I-hour period. Ifthe animal did not move from its previouslocation over a period of time, it was assumedfor practical reasons that it had remainedin~ctive. To further qualify this assumption,a111 mal location was checked every 10 - 15minutes. The mean distance moved was used to

plot the daily activity pattern for each animal.Palomares and Delibes (199 I) showed thatdistance travelled gave similar results to net~ctivity time in estimating daily activity patterns111 the Egyptian mongoose (Herpestes ichneumon).

SunJf!)' of Stomach Contents

A I-day shooting exercise was organized in thestudy area to determine the feeding habits of C.nota/us based on the stomach contents. Shootingstarted at daybreak and ended before noon tocover the time of the day when animals wereobserved to be most active. The exact time of

TABLE 1Summary of breeding condition, weight and tracking dates of four Callosciurus notatus

Male AMale BFemale AFemale B

Breedingcondition

BreedingBreedingNon-breedingBreeding

Wt(S)

246232201226

Tracking dates

June 5 -8, 10 - 12, 1990Jul 6 - 7,9, 12 - 16, 1990July 2 - 5, 7 - 8, 1990July 20 - 28, 1990

PERTANlKA.J. TROP. AGRIC. SCI. VOL. 21 NO.1, 1998

PLANTAIN SQUIRREL CAfLW)(;f{/fWS NO'{ATUS IN A PLANTATION HABITAT

the shooting of each individual was recorded.The gut was opened and contents broadly iden­tified as oil palm fruit, cacao mucilage, uniden­tified plant matter and insects. The proportionof each item was subjectively quantified.

RESULTS

Of the six squirrels, sufficient data for analysiswas collected from only four. Signals were notpicked up from the other two and their trans­mitters were assumed to be lost. Daily activityprofiles were plotted for each squirrel (Fig. 1).

In terms of habitat utilization, Male A wastracked in both cacao and oil palm, Female A incacao only and Male B and Female B in oil palmonly. The small sample size precludes compari­son of activity patterns in the two habitats. Indi­vidual daily activity patterns (Fig. 1a-cl) are quiteconsistent. Differences in the mean distancetravelled during the tracking period were notsignificant between Male A and B (Z=1.345,P > 0.1) and between Female A and B (z=0.275,P > 0.1). All squirrels started movement at first

light i.e. around 0600 h and maintained a highlevel of activity until 0930 h. Since trackingstarted 30 min earlier i.e. at 0530 h, it is as­sumed that squirrels stayed at their nest sitewithin that period. In 12 out of 14 instanceswhere tracking was carried out on consecutivedays, squirrels were located at their last position(1730 h the previous day), suggesting that theyremained inactive throughout the night. Twomorning peaks were identified, the first at 0730h and the second from 1030 - 1130 h. The firstpeak was higher than the second except forMale A, where the level of activity was similar atboth peaks. The afternoon period (1230 - 1630h) was marked by a period of ;:erestJE except forFemale A, which showed a third peak about1530 h. All squirrels showed increased activityfrom 1730 h to just before nightfall. In terms ofthe daily activity profile, they generally showed aprogressively decreasing level of peak activity,except for Female A where the four peaks weresimilar. Female A, the smallest (201 g) was

'>0 50

40 40

- :10 - 30

- -

20 20

~

C C10 10

0700 O!lOO 0900 1000 1100 1200 1300 1400 1500 1600 1700 1800 0700 0800 1J900 1000 1100 1200 I:~OO 1100 1500 1600 1700 I!lOOActual time Actllallilllt:'

En or bar: YY"'c Confidence IiIllil EnOl bar: 95~ Confidence limit

Fig. 1a. Daily activity pattern offemale A (radio-tracked incacao only)

Fig. 1b. Daily activity pattern offemale B (radio-tracked inoil palm only)

07000800 0900 1000 1100 1200 1300 1400 1500 1600 1700 1800

Error bar: 95% Confidence limitAClUal lIlIle

lO 40

- :10 --c -c :10

20 20

r ~::S is

10

0700 O!lOO 0900 1000 1100 1200 1300 1400 1500 1600 1700 1800

Actual time

'>O~----------------....,

EIIOI bar: 9:)!1 <;(Hllirlence limit

Fig. 1c. Daily activity pattern of male A (radio-tracked incacao and oil Im/m)

Fig. 1d. Daily activity pattern of male B (radio-tracked in oilpalm only)

PERT IKA]. TROP. AGRIC. SCI. VOL. 21 O. 1, 1998 25

Yl.N. HAFIDZI

- - __ Female B

• - - - - • • Male A

-- - l..,lale B

50

40

I30

]~g 20

"uc:~i:S 10

,<

I

"I • -Female A

Period squirrelsmost aClive

0700 0800 0900 1000 I 100 1200 1300 1400 1500 1600 1700 1800

Actual time

Fig. 2. Comparison of the duration of the active phases of the four radio-trackedCallosciurus notatus

• - pooled from 1000 due to few squirrels

TABLE 2Number of squirrels shot every hour

from 0700 - 1200

therefore active for a relatively longer period. Acomparison of the duration of the active phasesis shown in Fig. 2.

A total of 18 shot squirrels were recoveredfor feeding analysis (Fig. 3). The first two squir­rels, recovered at 0700 h, both had empty stom­achs. The first squirrel with food in its stomachwas recovered at 0735 h. Insects were found inall individuals in varying amounts but were par­ticularly substantial before 0900 h. Oil palmfruits constituted a substantial amount from in­dividuals recovered from 0955 h onwards. Cacaomucilage was only identifiable in two individu­als.

Time (I-hour period)

0800090010001200'Total

No. of squirrelsanalysed

8433

18

26

100

c80

:~00-E0 60u~

Cv~v 400-

20

0

Fig. 3.

58.3o cocoa

70insects

24.14 o oil palm

0800 0900 1000 1200hourly period

The percentage composition of oil palm, insects and cacao from stomach contents ofCallosciurus notatus shot in the study area

PERTANlKAJ. TROP. ACRIe. SCI. VOL. 21 NO.1, 1998

PLANTAIN SQUIRREL CAU-O.liCIUIWS NO'j'!l'j'US IN A PLANTATION HABITAT

Due to the small sample size, squirrelscollected within a one hour period beginningfrom 0700 were grouped together. The respectivefood items from each stomach were pooled andthe mean proportion estimated for each group.The percentage composition of each food itemfor every I-hour period is shown in Fig. 3.

DISCUSSION

The present study suggests that C. notat'LlS exhibitsa daily activity pattern with at least three distinctpeaks; two in the morning and one in the lateafternoon. However there is individual variation,as exemplified by one individual, which hadfour peaks. Three or more activity peaks per dayconstitute a variation of the typical biphasic orbimodal pattern that has been observed in manyanimals (Aschoff 1966). Tonkin (1983) suggestedthat this type of behaviour is either amanifestation of the physiological need for restor that a break following the morning feedingperiod is needed when animals have eaten theirfill. Whereas this might be reflected in thepresent study, the existence of double morningpeaks punctuated by a short break and variabilityamong individuals raised some interestingquestions.

The double peaks may also be associatedwith a particular resource or foraging strategy tooptimize food searching effort, hunting successand energy return. The difference in the intensityand duration of the active phases may reflect thenutrient value of the associated food source.This is supported by evidence from stomachcontents of squirrels shot in the study area,which shmved that insects formed the primaryfood item during the early morning hours (0730- 0930 h). Oil palm was only identified insubstantial amOUI1lS from individuals recoveredfrom about 1000 h onwards. This suggest thatthe two active phases in the morning may accountfor two different kinds of food. The differencein the level and duration of the two peaks isagain indicative of such food preference. Thefirst peak may account for an insect diet. Insectsare highly nutritive but are relatively scarce andhighly mobile, thus less predictable in terms ofdistribution. Insects may also be more activeduring the early morning. Therefore, animalshave to intensify their search to take advantageof the temporarily abundant insect food resource.The short rest before resuming the second active

phase may indicate the point whereby foragingfor insects is no longer economic in terms ofenergetics. Energetic cost is an important factorin the foraging decisions in animals (Krebs et al.1983). Oil palm fruits and cacao pods, on theother hand, are more evenly distributed, highlypredictable, conspicuous and immobile, and thusrequire less effort to secure. The second activephase could be associated with a primarily oilpalm diet. A longer phase may suggest a greatervolume consumed to meet dietary and energeticsrequirements. In general, foraging time andintensity are influenced by availability. nutrientvalue, predictability and handling time of foodsource (Lewis 1980).

There are other possible explanations forthe double morning peak. The first of the twopeaks may account for the daily re-establishmentof exclusive feeding ranges. C. l/otat'LlS has beensuggested to exhibit some form of territorialbehaviour (Duckett 1982). Such behaviour wasalso observed in this study. C. I/olalus is mostconspicuous at daybreak, when it can be seen inhot pursuit of other individuals apparentlydriving away intruders from a dray or a favouritefeeding site.

From direct observation, most of theafternoon was normally spent dozing, groomingand other less energy expensive activities.Continuous radio surveillance indicatcd thatanimals sometimes staycd in the samc locationfor up to four hours. The final activc phase inthe evcning can be ascribed to a final feedinground or nest run. Results also suggest individualvariation in thc observed daily activity pattern.Tonkin (1983) observed a similar phenomenonamong a population of the red squirrel (Sciurusniger). He suggestcd that newly independentyoung adults tend to be active during periodswhen fewer adults were active to avoid directcompetition for food and space. Similarly, thiscould be viewed as a form of a strategy adoptedby the small, non-breeding Fcmale A (Table 1)to reducc competition. However, it could alsobe random variation since observations werebased on one individual from a sample of four.Daily activity pattern can also be a function ofthe physiological needs of the individual. Maier(1992) showed in pipistrellc hats Pipistrelluspipistrelhls that variation cxisted betwecn fcmaleswhich exhibit unimodal activity pattern duringpregnancy and bimodal during lactation.

PERTANIKAJ TROP. AGRIe. SCI. VOL. 21 NO. 1. 1998 27

M.N. HAFIDZI

CONCLUSION

The results of this study show that C. no/atus ismore active during early morning and just beforenoon. Therefore, shooting exercises to controlsquirrel infestation in plantations, particularlyoil palm, should be carried out during those twoperiods i.e. when squirrels are more active andthus more conspicuous.

ACKNOWLEDGEMENTS

The author wishes to extend his thanks to Dr.G.vV.H. Davison for his supenrision of the work,the Department of Zoology at UniversitiKebangsaan Malaysia for funding the research,Ladang Prang Besar, Kajang for permission towork in their plantation and the many individualswho have helped in the study.

REFERENCES

Al.iCHOFF, J. 1966. Circadian activity pattern with twopeaks. Ecolog)' 47: 657-662.

DlICKETT, J.E. 1982. The plantain squirrel in oilpalm plantations. Mala)'. Nat. J. 36: 87-89.

H-\FI1)ZI M.N. 1992. Some notes on bark strippingbehaviour of Calloseiurus notatlls. The Planter.68: 501-505.

K-\~IARUDIN, K.A. 1982. The ecology, pest status andcontrol of the Malaysian wood rat Rattus/iolJlanicus (Miller) in a cocoa-coconutplantation. Ph.D. Thesis, Michigan State Uni­versity.

K-\~IARUDI , K.A. and C.H. LEE. 1981. Modes ofcocoa pod depredation by three smallmammals. iVIARDI Res. Bull. 9(1): 42-48.

KENWARD, R.E. 1987. Wildlife Radio Tagging:EqllijJlllent, Field Technique and Do/a Analysis.London: Academic Press.

KREBS, J.R, D.W. STEPHENS and WJ SUTHERlAND.1983. Perspectives in optimal foraging. InPersjJeetives in Ornithology, ed. A.H. Brush andG.A.R Clark. Cambridge: Cambridge UniversityPress.

LEWIS, A.R 1980. Patch use by gray squirrels andoptimal foraging. Ecology 61: 1371-1379.

MAIER, C. 1992. Activity patterns of pipistrelle bats(PijJistrelllls jJijJistrelllls) in Oxfordshire. J. Zool.Land. 228(1): 69 - 80.

PALO~IARES, F. and M. DELll'.ES. 1991. Assessing threemethods to estimate daily activity patterns inradio-tagged mongooses. J. lVildl. l"fgl1l/. 55(4):698-700.

POULIQUEN, 0., M. LElSH~IAN and T.O. REDHEAD.1990. Effects of radio collars on wild mice iVIllsdomes/iellS. Can. J. Zool. 68: 1607-1609.

SI!ORTEN, M. 1962. Squirrels: Their Biology andControl. Bulletin of the LVlinistl)' of Ag!icllltllre,Fisheries and Food, 184. London.

TONKI I, J.M. 1983. Activity patterns of the redsquirrel (Sciurus vulgaris). 1V!aJlIIII. Rev. 13: 99­111.

WOUO ,RJ and BJ. TRO\l·BRI1)(;E. 1885. The effectsof radio collars on wood mice AjJodelJluss)'lvatie1ts. J. Zool. Lond. 206: 222-224.

(Received: 25 AjJlil 1995)(AccefJted: 26 l"fo)' 1998)

28 PERTANIKA.J. TRap. AGRIe. SCI. VOL. 21 NO.1, 1998