larndq H.q4d4\ ed

Biology and Commelcial Utilization of Acrochordid Snakes, withSpecial Reference to KafiJr.g (Acrochordus javanicus\

RICHARD SHNE] PrrrR HARLow,r J. Scorr KEocH, AND BoEADI'

'Srhaol al BiolaSicol Scletes AA8, Th. Uni!.tsity of Svdne!, NSW 20a6, AtsttoLk, and1c."tft lor R5edrch in Eialqy, Muvu of Zaotalr, LlPl, BDSot 16122, thdon.sid

AssrRAcr. - we measur€d an<l di$ected fil€snak€s (A.tu.hordus javani.ds, colle.red foi th€ comm€r.ialskin rrad€ in southern Sumatra, to pfovide information on morphologt and r€producriv€ biology, Ircombinarion with publiehed and ori8inal d.ta on theothertwolivingacbchordid sp€cies,thbinrormahonenablesus to comp.re lhes€ ihr€e l.xa.nd to examine wherhei or not tlte existing commercial narvesl ofA. ri,v,ni.6 is litely ro b€ sust.inable in rhe long term , a.rochordus javanicts is supe i.ially similar rothe A6ir.lian ,, ,.!lu.a€, but is nore heavy-bodied (almost twice lhe ma$ at thf 5ame SvL), wirn alaraer h€ad.nd a longer tail. all thr.e acro.hordid spe.ies show siAnificant s€xuel dimorphism in bodilyproportions(mas/Svl-,head l€ngrh/SVl,t.il length/SVL),butsexualsiz€ difrorphism is l€ss pronouncedin ,{, g/,rular,s ihan in th€ rwo larAer sp€.ies. Reprodu.tion is seasonal in a1l thr€e ado.hofdids, withovul.tion aroundJuly and parturition five orsir monthslater. L.raer female a. jrv,,i.us produ.e larAerlitters, and about two-thirds of r[. adult females in our sampl€ w€re reproductiv€ in lhe tear th€y w€re.olle.ted. rnr€r mass€s Elative to mat€nal body nass (=RCMS) are higher in A, rrtttae aad a. javani.usthan in A, Annularus The sihilarity in RCM in,1, ai,f{n. aad A. javani.us, despi,re a rwofold differ€n.ein nean narernal oa3s, resulb primarily from rhe mu.h laraer li$er siz€ oI A, i,v,ni.us(29.3, y€6us 16,9in c. ara&rr€). Our dara otr the biolo8y of -4, ji:r,ni.!6, and on the ways ir wnich rhis species is.oll€ct€dfor thc commer.ial slin indostry, suggesr thai the current harvest is unlik€ly to seriously reduc€ wildpoprlaiions. Th€ r€l.lively .*asonal prccip ation r€gine in this area, the e{ensive (and largely ina.-(essiSie) habitat, rhe lack of sp€cific and ef6.ient techniques to capture 3nakes, the nigh reprodu.riv.output of rhe snales lnd ihe low ..onomi. value or their skins, are som€ of ihe factors .ontribuling lothis.ontinu€d susr.irabiliry. rn (ortrasi, the bioloAy of Astr.li.n A, ,.afu.ae sugg€sts that this spe.ietis poorty suiled to comm€rcial harv€sting.

AGrRAcr (Bahasa Indon.sia).-Kami telah mengukur d.n memb€dah uld-ular latt^g A.ro.hordu.ji,v,.i.u, yana k.mi kumpulkan dai kulickulit yang diperdaganglan di slmrter. selatan guna men-dapatkan informdi tentan8 morlologi dan k€hid.pan p.rkembangbiakannya, DenAan nenga.u pada datayang telah ditulis dan d e.bitk.n s€rla menAaabungkanny. den8an data prim€r menAenai dua rumpuna.rochordid iainnya yrng nasih hidup, inlormasi ini telah h€hungkinkan kami untuk mengadakanperbrndingan pada b.ngkai (take) ke tiga empln rld tersebut, serta menyelidi*i apatah ,4../,rvfii.,sini dap.l diruai s€cara komeGial untuk iangla waktu yana panjang. Daii llanya A. javtni.lslni, mnipdengan 4,,laf,laeAuslralia, tapi l€bih b€rat (isi tubuhnya h.mpir du. kali lipat dengan SVI- yana s.ma),d€nAan kepala lebih besar d.n elor lebit paniang. Ke tiga rufrpun acrochoidid nehperlihatkan dimorf-isme (dua b€ntuk d€ng.n strukl!r yang berbeda) kelamin d€n8an proposi lub!h (isi /Svr, panianC k€pala/SvL, paniang €ko4svu, namutr pada a, gnn,l:t,r dimorfismeny. ridakseielas ke du. tun?un lainnya,Reproduksi p.d. le tiga !cochordid ini msim.n silatny., dengan ovllasi s€kitar bdlan Jrli dan p€nerasanlima ara! enam bllan lemrdian. B€tina ,4, i,ra"icue yang bes.r, m€lahirkan ular-ul.r yang berukuranlebih b6.r, d.n s€kitar dla pertiga b€tina il€wasa dati smp€i kami reproduktif dal.n set.hun masapenangkapannta. B€nr keiurunan ut.r yan8 r€latif d.n8an isi tubuh indultnya (:RCM) lebih tinggijumrahnya pada rufrpun A. aralurae dan A. javani.ls daripada rumpun A, Aftnd|,.,.. ?e6am.an RCMpada rumpun c, atfd&e dan A, rirvanrtur terutam. akibat rn.k-anak A,,;r,,rtus yang lebih beser (?9,3berbandina16.9 p.da c,,/,lurae, nes&ipun berar A,,ralurae.lua tali lipat da.i A,.,:'v,nrtur, D.la kamimeng€nai biologi A, i,vari.u6 dan .ara-.ara tumpun ul.f teGebur dikumprlkan dari industri lulit ko-me6ial, menunjukkan besar kemunAkinan b.hwa luaian saal ini tidak ,k.n m€ngurangi popllasi ularliai Jutrh hlian yang kuranA m€n€ntu, habit.t yanA amar luas {dan kebanyakan tidak deprt dicapai),kur.nany. tekniklh$us yangefisien untuk penanSk.pan ular p€rkembangbi.kan ul.r yan8 amat subur,s€rta nilai ekonomi6 kulitny, yang.endah, fre.upatan beb€rala faktor yang nendukunA k€lesrariannya.Sebaliknya, biologi ,4, ,r,1d€e memp€rliharkan brhwa rumpun ul..:ni sangar tidak cocok untuk tuaian

IILESNAKE BIOLOGY 353

The three living species ofaquatic fiLesnakesthat compiis€ the family Acrochoididae mustsurely q;alfy as among the most remarlablesnakestu the world. These troPi.al animalspossess a bizarre suit. of chara.teristics noi knownin any other snaket, in terms of theii mor-phology (e.9., rugose skin, convoluted intesti-nal tract), physiology (e.g.,1ow metabolic rate,high blood volume), behavbi (e.8., tongue_flicking ra te, prey capture techniqu€t and ecology (e.9., pronounced sexual dimorPhism,lowrates of feeding and reproducin8) (see reviewsby Lillywhite, 1991; Shin€ and Houston, 1993).Extensive infornation on reProdu.tive biologyand natural histoiy a.e available for ihe laiSefreshwatei AustElian spe.ies Acrochotdus ata'

furd. (Shine, 1986a, b, c, Shine and Lambeck,1985; Houston and Shine, 1993, 1994) and thediminuitive madne A. 3rdrrlrtrs from lndonesia (Rersnan, 1958), Malaysia (Voris and Glo-dek,1980; Lemen and Vor is ,1981)and the Phi l -ippines (Dunson and Minton, 1978; Corman eial- 1981). Ho*evei, the remaininS acrochoididspccies,,4. /dtarrrs f.om freshwater and biack-ish-water habitats in southeast Asia, has attract-ed less scientific attention. Most Published information on the ecology of the Iavan Filesnakeis anecdotal in natule (e.8., Dowlin8,1960; Boo-Liat, 1964), with the only extensive data setcoming from Bergman's (195E) studies on ]avan

Quantitative information on the morPhologyand ecology of,4.ro.lordlsi4trrnrs is of inteFest in two main .espe.ts. Iirstly, it Provi.tes abasis fiom which to cairy out an ov€rall coh-parison of the thie€ living a.ro.holdid tdaGiven the .onsiderable norPhological (Mc-Dowell, 1979) andecoloEi.al (Shine, 1986a) di-vergence between ,a. ,/aFtoe a\d A. gtnnuLat$,the characteristics of ,'{- i4?tti.rs are ol Partic_ulal interest. Secondly,A. jotaii.us, Iike A Srdnrlalrs, supports a significant .ommercial tra.lein snake'skins (Litlywhite, 1991). Biologi.al dataon a.rochordid snakes, and infoimation on theways in which the snaNes are .ollectcd, can thusbe oI value in €valuating whether oi not thisharvest is likely to be sustainable in the lonStern (IritzSerald et al., r993). The plesentPaPeris based primarily on our measurements anddissections of a sample of ActDchotdus iarani.uscollected for ihe commercial skin tra.le insouthern Sumatra, but we alsoProvide oiiSina]data from measurements of Preserved A. 8/4,

M^rERrALs AND MErHoDs

A , r J . h o r d u \ F ! a n r . u s - W F v i . i L ' d t h L. o t r m p h i " l . n r l e s \ i n n r n g u P p r d r , o n o ' n r rl e r b J n S . l h e . a p r l r l c . t v o f . o u r h P r n 5 u n a l r ul roT 7q tu ) to o Augu. t loq l 'nd f rom 4 lo

21 O.tober 1994. we interviewed the ownersand staff of these enteiprises about the ways inwhich the snakes are .ollected and transPortedto the skinning fa.tories. we also visited a vil-lase (Tanjung Laut, 52 km w of Palenbang)whefe many filesnakes are collected by lo.alfishermen for sale to the skinners, and we weretaken out into the extensive iidal swamPs fromwhence the snakes originate. The rest of ourtime was spent in the skinning factodes in Palembang, where we measured andweighed thesnakes pdor iolkinning (immediately aftertheanimals had been killed), and then examinedthe Sut.ontents and gonads of the slinned car.asses. wc recorded snout vent lenSih (SVL'head lenAth (straight-line distance along thelower jaw, from the snout to the posterior angleof ihe qladrate-mandibular arti.ulation), bodymass, gender, presence of Prey items in thestoma.h, and reproductive condition (maturityor immaturity), We did not measure truncatedtails, or heads that hadbeen substantially dam-aged when the snake was killed. Males werescoredas matule if the testes were turgid ortheelJerent ducts were enlarged and containe.topaque white nuid (presumably, sPern). Fe-males were .lassed as maiuie if they wele Srav_id, if their ovalies coniained vitelloSenic fol-licles (>5 mn dianeter) or colPora albicantiafrom previous ovulations (these were obviousas dark yeLlow specls on the ovary), and if theovidu.ts wer€ thickened and muscular. Ovidu.tal embryos were counted and measur€d,and the oviducts and their contents were re-moved and weighed separately.

Aco.hordus granulaius. The senior authorexamined and measured 64 prese ed sPecimens of marine Iilesnakes in the colle.tion ofihe Field Museum ofNatunl History (Chicago,USA). These snakes were collectedby D!- H. K.Vois and his .olleagues in Malaysia, and haveserved as the basis of earlier Publications onotheiaspccts of the biology ofthis iaxon (Vorisand Glodek, 1980; Lemen and Voris, 1981).

REsuLrs

Aclochofdus javanicus. We examin€d 101lre,n y- f l led A. ;a on d o( whi , \ 21 serc" d u r l m d e - o T h p ' e a o r l l l P m d l e ' a n d t h e t c {sr-e iuven. .e. Adul l lPmd e. were con' idpFabl t .Jrse, ,han JJur t ra le. In rern bootlensrh. ( . -e lJb.e l : 1 . " b.07. r 0 .0001)mein head lengths (t,s: 12.9, P < 0 0001), andmean body masses (t.0 = 8.37, P < 0.0001), burmean tail lensths did not ditrei between adulr. p e . i r . n " o r i h e w o . e . " . r t " I 0 1 . f ' 0 . l l r

M o r e d e r d , t " d d f , ) . i , . h o h e o r h d ' r h e s e \ d i fferen.es in head siz€s and body masses werenot a simple consequence of the larAer m€anbody lent lh . o l l .m,re ' : : l ,hc 'dnc sVL fe-

354 R. SHINE ET AL.

E

a0

E

:

400

300

200

100800 1000 1200 1400

800 1000 1200 1400

' e n d l . - h d v e . \ o r t e r t d i l . J l J n ) 6 i ! e l s v t . Ahe leroScne i l ) u f s lope- te , r .nowpd no drver8en.e be lh .en rhe "e \e , , . td i l lpnSth re rd t i v .t o - V l ( F , s . 0 . 0 7 . P + 0 . 7 o r o u r r n . i 1 . . o rcovaian.e with sex as the factorandSVL as the.ovariate showed a highly signilicant differencebptM-e1 na le , a rd lemJ ip" i i tne e te t r t ion ort \ e r e g r e s \ ' o n l , n e , r l . 4 l o l . 1 0 . 0 0 0 t ,see Fi8. 1). Analysis of head len8ths letative toSVL gave ihe same result (heterogeneiry of. l o p e . ( e s l f . 2 2 8 . P 0 . t 1 . A \ L O V A t ,' | 12 .1 . f 0 .0001, . Bod) rd$ .ncrpd \ed nore.apidly with in.reasing SVL in femates than inhd le - rF ig I h . re rogrner ly o t . t , rpe \ rp - t . | ": 6 . 3 r . t - 0 0 2 )

Mpi.urement . . f fur l tprh "mb-yo" f rJn rsofemdlps. . l lecrpo In Oclooer "e\edlea thaL,e^-ual dihorphism is present plior to birrh. In both' i r reh. temale. dvFrdgcd.hor ter thdn mate" Ls V I / i t l e r | - n e c n , o t 2 ] b . 4 r - . 2 4 o 4 m t r . , -: 3 .63, P < 0 001, l i r ter 2-259 9 vs. 28b 5 mmtr - 5.52, P < 0.0001) and tail lenSrh (litiei 1-54.8 vs. 58.2 mh, t," : 2.84, P < 0.01; tiiter 2-<8 4 vs b4.o mm.l . 4 2t . f 0 .00011 L mdtFnponrte. J l .o hdd ldrSer hcrd. . re. r r iv" ro s \ Ithan did iheir b.others (litter 1-heteroseneih-o f . l o p p c l . , 0 . t 1 f 0 . 7 2 . c \ C o v i r , -17.c4 , - . 00005. t t r r 2_hprerr8erer I o l. lope. F. . . 0 15 | 0 . / t AN( O\ A I , ,9 .9r , P < 0.005).

ThA ld.8e.r l rmdrure nate,A. puon\u. wa\l0c0 mh 5\ L but rhree .ndt ter da es l the.r r ' lesr wds oo0 mm SVL) were mdlurp. Thu.nate 4. iauantu- in .outhp.n \Lmrr-d mdtur(dtdround ,000 md 5\ L h e d l .o nored coned.erable overlap in body sizes of matuie and im-hature animars in female A_ janfticrs,btr at alarger size than in conspecific hates Gmallesrmatde : 1140 mm, largest immature = 1300mm SVL). All of the adult males had efferentducts packedwithsperm,and theirtestes.ansedfrom fully turEid to alhost flaccid. Forty th;eeof the 6zaduitfemales (64%)w€rerep.odu.tive,with all of the others showins snait (=6 nn)ovadan fouicles and thin oviducts. The Dres-en.e of coropora albicantia in the ovaii;s ofthese animals indicated that thev had repro,duced previously, and hence rheiwere cdsedas mature. Some reprodu.tive females ihat hadbeen mdrnta ined In capnvi tv p lor to k i l t ehad larSe resorb,ng fo i t ic tes o; ovrductat en:bryos. Single femates in both of our tlips (Julyand October) weie imnediarety pre-oviilator;.All of the other reproductivi females weieg.avid. Apart from one pre ovulatory fehale(se€ above), the .eploductive adutt femates ex-amined in O. tober aU conta ined tarSe prghented cnbryos. and rwo were ctose to fullrerm. Of the snaLes e\am,ned in Jub Auausr ,mo+ had ovujat.d nor tong before ihei were

30

1600

16007000

6000

5000

40QO

3000

20aa

1000

01000 1200 1400 1600

snout-vent length (mm)IIc.1. Sexual dimorphism in l.var htesmkcs, A.

rcchottl,s ja"onieus, lr.^ sorthern Sunat.a. Relativeto snout-vent length, mJe filesnares (op€n .ncles)h.ve rongei t ils (u!!er g ph) and shorter heads(middle grnph), and weigh huch les (lower 8.aph)than .1o conspecifi. females. Sec terr for srarjsticalanalyses of these differcnc€s.

males had larger heads than did males, as we]las beinA much heavier (FiE. l). Atl of rhese sexdiffeiences weie statisricaly siSni6cant. The lackof a significani sex differen.e in mean aduti taitlenSth,despite the larSedisparity in mean bodysues between thesexes,was due to the fa.t that

400

killed. One of these females was immediatelyposr-ovulatory, with the corpora lutea stiU inthe forn of large collapsed follicles 25 to 30 nnin diameter. ln all of the other gravid femalesfrom ldy August, the colpora lutea had re-gfessed to a diameter of 7 to 10 mm, with em-bryos tiny and unpigmented in all except fourof the females. The most fully developed embryor seen in July-August wefe 135 mm SVL,probably less than one third throu8h gestation.One female with enlarged ovarian follicles con-tained thfee fullterm, piSmented embryos froma previos litter, that had died and become en'capsulated in a damaEed ovidu.t.

Litter sizes of our dissected A. jardflicts nnredftom 13 to 52otrspring(mean = 29.3,5D:8.74,N = 43) and were significantly correlated withmaternal snout-vent length (Fig- 2; N : 43, r: 0.52, P < 0.0003). A relatively high pfopor-tion of ovidu.tal ova were infertile (mean =267,, SD : 31, iange - 0-1007', N = 27). TheRelative clutch mass (RCM, : combined massof oviducts and their contents, divided by ma-ternal .ar.ass mast ranged froh 0.15 to 0.47(mean = 0.29, SD = 0.07, N = 36). Howevetthis estimate of RCM will be mu.h lower thanthe RcM at parturition, be.ause of the increasein embryoni. mass throu8h Bestation (Shine,1986a). The RCM at parturitioncan be estimaiedfrom our data on litter sizes and matemal car-cass masses, combined with information on av-eEAe neonatal mass (:42 g from Beigman,1958,linear dim€nsions of his neonates were simila!to those of the fullterm embryos we found). Theresulting estimate of RCM (nean : 0.a3, SD :0.08, N : 36) should offer a reasonable approximation to the a.tual value-

Only thiee of the disse.ted a.rochordids .on-tained identifiabl€ remains of prey. Two .on-tained fishes and one contained a lar8e aquaticsnaiL (Pt4spp.)that may havebeen a secondaryprey item. Most of the snakes also containedfish-hooks embedded in their stoma.h walls.Our inteFiews with villagere at TanjunS Lautconfirmed that the filesnakes (called "u1ai ka'rung" or "ular berot" in Baha$a Indonesia) aiecauSht at niShi on untended fishing linesusingsmall fish as bait. There is a diverse fish faunain the reSion, with the most abundant largespecies beinS eleotrids (e.9., Otyeleottus kta|h old t us, O p h idac c p h a I u s spp.). These are sluSSishbottom-dwelling fishes that attain maiimumsizes of >3 kg. The villageis told us that th€price paid by the skinners for filesnakes is solow that many of the snakes are simply dis-carded. In keeping with this story, we foundone large male filesnake floating, dead, in theswamps nearthevillage, alength of fishinS lineprotruded from its mouth. The fishermen catchthe f i lesnakes in the huge brackish-water

snout-vent length (mm)Frc. 2. Reproductive output of Javan filesnakes,

Aorhordrs jraonnus, fron iouthern Sumaba. Largcrfenale filesna(es produc€ larger littes; 6ee rert forstatisticat analysls of this resutt.

swamps thatsuround the village. There is littleopen wate! in the area, with only small boatchannels among dense mats of floating glass,the exoti. wateMeed S4i.inia ftokna, and, nypapalns (NX/r tllt.art.

Inteniewswiih the skin'traders indicate thatthe fiiesnakes processed in Palembang are .olle.ted flom a wide area in south€rn Sumatra.The snakes are taken as a by-catch, and are nottalgett€d specifically by fishermen. The majorskin-dealers have networks of lepresentativesin larger rural villages, and ihese people pur-chase the snakes from fishermenand send themon to PalembanS. Filesnake skins are trot asvaluable as skins of sorne of the othel reptilespecies ha*ested in the area, notably PyllD,teti.ulntus P. curtus, and Vdlrras salraiol (whichare allproc€ssed in the same skinning factoriesas the filesnakes). However, the healTsetbuildof A. jowflicus and its lack of enlarged beuy-shjelds mean that the skins of this sp€.i€s arevery wide, and henc€ well-suited to the man-ufacture of a considerable range of leather items.Ma.Lei forces control the volumes of filesnakeskins processed and exported, be.ause even asmall drop in pd.es paid for the skins can re-move the economi. in.entive for village$ tocollect th€ animals and transpo* them to placeswhere they can be sold.

Aciochordus granulatus. In olde. to fa.ili-tate .omparisons amonS the three aclochordidspecies, we report heie some original data onihe norphology of preseived specimens of ,4.grrrrlatrs. These data.larify the degree of sex-ual divergen.e in size and morpholo8y in thenarin€ filesnake,asubjectuponwhich previousreports have been contradictory. Our measure-ments of 24 male and 25 fenale A. BranuLdtus

IILESNAKE BIOLOGY 355

1200 1400 1600 1800

,i +o

E 3 0

1 0

356 R. SHINE ET AL,

4AA 400

300

300 400 500 600 700 800

snoutvenl length (mm)r rc 1 5. .ur l d norp\Fn rn n dnn" 6 lmn. ( . '

A. a. \ord, . ? .dnLlatL f r 'T Vd,d) jc . Re.d l r \e ro--4ur \€n le-g h. m" lc 6 l " "n" l " " rop n. n " . , .v",l'onFr hpdd- rupp.' B dphj drd w"rgh ru, h e,'lo$er 8tupl - ) rL in do ro- .p- :1. rerc le- . ( " " re 'ror statistical analyses of these dlfierences

(in.luding jlveniles as well as adults) revealedno sex ditrerence in mean SVL (nales-nean: 569, SD : 75 mfr ; females-mean:581, SD: 86 mm, unpaiied two'iailed t, = 0.50, P :0.60). Females tended to be heavier than males,but the difference fell short of statistical si8-nificance (males mean : 101.0, SD : 32.5 g;fetuales-mean = 126.r, SD : 56.5 g, unpairedtwo tailed tr : 1.84, P : 0.07). Females hadsigni6cant ly large! heads than d id males(maies mean: 16.7, SD = 1.6 mm; fenales-mean = 19-0,SD:2.3 mm; unpaired two-tailedta,:4.03,1 < 0.002). The relationship betweenSVL and head len8th also differed beiween thesexes: the slopes were similar (hetelogeneity ofslopes test, with SVL as covariate and headlength as dependent variabler F,4 : 2.40, P :0.13) but the elevations ditrered (ANCOVAwithSVL as the covaliate: I,.." : 62.0, P < 0.0001,Fig. 3)- Mass increased more rapidly with SVLin {emales than in males (hetelogeneity of slopes

l e5 ' . w i 'h cVL ds (ovdr . r re dnd mas. d , depen-dent vadable, fL11 : 4.27, P < 0.05; Fig. 3).

DrscussroNOur data on a.rochordids are g.neally .on

sistent with earlie. work on the same spe.ies.ln particular, our information on sumatran,4.iarauc,raccords welt with Bergnan's (19s8) re-sults based on ]avan specimens. Iof exahple,orf stu.lies agree on the patterns of sexual di-morphism in size and proportions, on sizes al. e \ u d 1 m " t u r J l r o n . o n ' h " , e d . o n d l r i m . n 8 o lovuLation,litter sizes, and the hiah proportio!of infertile oocytes (and see de Rooij, 19171Smith,1943). Our data on preserv€d specimensot A. gtanulatrs also .ontifm and extend the re-sults of earlier wolkers, in.luding significanlserual differences in relative tail lengih and thevirtual absence of sexual dimorphism in meanSVL (Bergman, 1958r Gorman et at., 1981). Ouranalyses show that female A. grawlatus arc sig'nificantty more heavy-bodied than aie malesolthe same SvL (FiB. 3), as tentativeiy r,ugSestedby Bergnan (19s8) but rejected by corman etal. (1981). We also found clear sex divelgencein relative head lengths (Fig. 3), a topic notinvestigated in previous analyses of ,4. grara,

The encapsulatedoviductal embryos we foundin ,4. /araricrs, like those that one of us (RS) hasseen previously in ,4. afatra., remain rcma.k-ably well-preserved for long periods of time.Civen that parturition in Sumatran ,,1. lara"icrsusually oc.urs around December (f!om Berg-man's 1958 data and o own), the three em-bryos we eramined had probably been dead forat least 6 months, and perhaps much longer.This virtual mummification inside the oviductsmay plovide an exptanation for a puzzlinS ob-servaiion made by Magnusson (1929), whofound a dead and paitly de.onposed enbryoinsidea fenale,,1. a/afula. that had been isolatedfo! seven years. Rather than sperm storage orparthenogenesis (as suggested by Masnusson)/this case may be explained by the abiliiy offehale a.ro.hordids to maintain dead ovidu.taloffspring in remaikably Sood condition for long

Cantparisoks olnohg A.\Ehotdtd Speci.s. -'l ablcI summarizes inforhation on the morphologyand leProductive biology of the three acro'chordid spe.ie$, based on a compilation of published data plus inforhation Sained durinS thepresent study. This .ompadson reveals manysimilarities among the three tda, bui also somestriking differences. Pojnts of similarity amongA.gla ttatus, A. dtdfu/de and ,4.la,a,i.!s includege^eral norphology, aquati. habiis, viviparityand piscivory. The apparentimportance ofslow-moving eleotdd fishes in the diets of ,.1. Srdt

24

^ 2 2

1 2300 700

E 1 0 0

FILESNAKE BIOLOCY

,laias (voris and Clodek, L98r), A. Infune(shi^e, 1986a, Houston and Shine,1994)and^.lrrdricls (based on the abundan.e of these fishes in the areas where the snakes are common)suSgests that the availability of these 6shesmayinfluence the distribution and abundance ofac-rochordid snales. A1l three a.rochordid speciesappear to have seasonal rePfoductive cycles, de-spite living in tropi.al habitats which Permityea!-lound activiiy (Gorhan et a1.,198r; Shine,1986a) and in all ihr€e taxa, gestation b€8insaround mid-year and finishes near the end ofthe year (Beignan, 1958; Goiman et a1., 1981tShine, 1986a). The only exception to this sea'sonality is Voris and Clodek's rePort ot year-round repfoduction in Malaydan A. Sto ulatus,but Gorman et al. (1981) hav€ arSued againstthis interpretation. All three acrochordid sPe-.ies also show less than-annual reProdu.tivefrequency in adult females, andall three sPeciesshow statisti.ally siSnificant sexual dimoi'phism in bodily proportions. Iemale acrochor'dids have larSer heads, shorter tails and areheavier-bodied than are conspecifi. males ol

Despite ihese similarities, many differencesare also appar€nt among the three a.iochordidspecies, as mi8ht be expecied in a Senus thatspans the entile range from ffeshwater (n. aral,fae) throush brackish (,1. la,aflt."s) to saltwater (,4. j/arrlrtrs), and whose constituentspecies "ditrer from one another anatomi.allyas mu.has do geneia of Boidae and Colubidae"(McDowell,1979, p.61). The most obvious dif'felence in external morPholosy is in size; onaveESe, adult female A.8/arrldirs weigh only6% as much as adult female A. jaranic$. Thedegree ofsexual size dinorphism is also Sreatlyreduced in A.8/ar!lat!s (Table 1). The conParison between ,4. dtalutae a\d, A. iauaticus is aparticularly interestinS one, because the twospecies are very similar in SVLS at birth, at mat-uration, and at mean adult body siz€ in eachsex. Howeve!, ,a. iu.dti.ls weighabout iwice amuch as ,4. dnlrl4e of the same SvL, have muchlarEer heads, and much longer tails (Table 1).Th€ difference in body mds between,4. afdtrdeand,4.la?rri.as is not an artifact of the hjgheireproduciive frequen.y of ,4. /r?a,ic,s (whi.hmeans that litter masses are included in bodynasses of nany A.la"ad.usbut rew A. a furae).Even aft€r subtraciion of littei masses, the car-.asses of our gravid fenal e A. jaodnic$ a! etaged2.5 kg, still mu.h heavier than average adultfemate A. dnfune \=).4k8). rhe adaptive si8-nifi.an.e (ifany) of the longer tails of,4. j4"ar;.!s remains elusive, but may be related to themuch largerandhore spinosehemipenisof thisspecies thar. ot A. anfutoe (McDoweil, 1979).

Profound ditrerences among the three acro-

.hordid species are also evidentin leproductivebiology. The madne A- Sfarrlalls produces asnall litter (mean - 4.8 offsprinA) as might beexpe.ted from its small body size. However,RCM also appea.s to be lowel in this taxon thanin its lar8er ielatives. The.ompalison betweenA. olilurae and A. jrrdnkas in terms of rePro-duciive €rpenditure is especially interestinS,be.ause these two congeneric spe.iesditrer sub-stantiallyinbodily proportions(andths, mass)buinot in SVL. The sihitarity in RCMS betweenthese two taxa is due to a slight intersPedfi.difference in ofisprinSmass, andan almost two-fold difieren.e in littersize (Table i). This.om-parison offers suppo for the hypothesis thatbody volume constnins reproductive outPut insnakes (Shine, 1992).

The other obvious interspecific difference inreproductive output among a.ro.holdids is inthe propoftion of adult females that are repto-ductive in the year that they are couected. Thedata suggest that adult female ,4. 8/arrlniizs and,4.la,atc,s may reploduce, on average, at leaston.e every two years. The remalkably low pro-portions of reproductive fenale A. arafurae(<107o) initially reported by shine (r986) havebeen confirmed by subsequent more extensivestudies (Houston, 1992). Reprodu.tive ftequen-.ies in,4. a/aFla. tend io be hiShest followingyears of unusually heavy rainfall (Shine et al.,unpubl. data), suggestinS that the greater arid_ity and annual .limatic vaiability of Australianhabitats may be responsible foi the low repro-ductive ftequency in female A. drat'wae. Themaline habitats occupied by A. gfanttdtus, andthe bra.kish swamps inhabited by A- idraflt.rs,show fai less seasonal and annual variation inwater levels than do ihe Australian biliabongsin which A. a/atr"re has b€en studied (Shine,1986a). This hypothesis that low water levelsconsfain reprodu.tive frequency in,4. 4f4l!f4efits well with the obsewation that this speciesfeeds at much higher rates duing the annualfloods than ai other times of the year (Houstonand Shine, 1994). Given that food availabilityis likely to be thc mosi important pioximateinfluen.e on leproductive output in snakes (e.9.,Sei8el and Ford, 1987), the viftually year-roundfeeding opportunities avallable to A. granuLdtusand A. iotanitus may allow femaLes to a.cu-mulate energy ieserves for reProduction morequi.kly than is possiblein the severely seasonalhabitat occupied by ,4. a/aFl4e.

CorLnerclal UtiLizdtioh.-Civen that two of thethree a.ro.hordid species are heavily haflested,it is worth trying to evaluate whether oi notsuch a harvest is likely to be sstainable in thelong term. Such an evaluation is comPli.aiedby the diff,culty of obtaining acclrate PoPula-tion .ounts of any snales (Turne., 1977), let

358 R, SH]NE ET AL.

TaFLE 1. A comparison ofmo|Phology and.eproductiv€ biology in the three livingspc.ies ofacrocnordidsnakes. Table gives ranges or mean values, with standard deviatior in brackets. Data ror.1 */,,,rat,r.omelrinadly from Bergnan ( 1958) and our neduren€nts of preaerved specinens in the Chicago lield Musern,wjth RCM data from Lenen and voris (1981). Data oh a. //,l,rne lroh Snine (1986a D) and subsequentunpublish€dstudies. Daraon/{. j,.,,i.r! lron the!.€sent study Gample sizes in terr), exc.F tnat inrornarionon size at birth cones from lergman (19s3).

snouFvent tength (mn)

At temale maturatior

At female maturation

At lenale flatffafiDn

Mas (s)

At fehale maturation

fenales r€lrodu.riveRelative clutch mass

280 322574540668 (4.7J678 (11 2)

3972

8.'75

1 01 6l 9

100

lss (3.7)r81(9.1)

4.3 (2.2)

0.24

320-343820-900

1150 13001067(82)t361\122)

52 78150170189.8 (22.3)212.4 (21.1)

12.5-r9.528

31.5 (2 0)5r .3 (5.3)

24.5-36.6380

1000473 (110.2)

15,15(6301)

16.914.7)

0.53

280 36010001140-1300118s(108)1351(120)

30210220756.7123,2)263.7129.9)

19365141.5 (3.s)s7.9 (5.8)

32 58900

20001397(539.7)3382(1105 6)

29 .3(3 .7)

0.64

alone aquati. snakes living in extensive tropicalswamps. In pn.ti.e, logisticat problems maleit virtually impossible to monitor densiiies offree livin8 populations ofany of the hanestedAsian ieptile taxa. Analyses based on the re-cords of .ommer.ial operations offer an alter-native source of infoimation from which thesustainability of the haFest may be evaluated(e.9., LurhooE and Cloombrid8e, 1989), butthe actral volume of traffic in skins of a partic-ularspecies may be disguisedby iLleSal tradinSand intentional or a.cid€ntat mis,labeuins ofskins. Information on the basi. biology of thespecies may also provide a basis for evaluatingsustainability, however, and the daia oD,4. a/-dl"roe and A. jatanicus are sufficient for us toattempt this exercise.

Overau, oui data on the iife-historyand 8en-eral bioloBy ofA.jdrnric!s sugAest that thisspe-cies should be capable of withstanding consid-

erabl€ harvesting pressuresi probably, pressures hi8her than those currently in operation.We base this conclusion on the foliowing chaFa.teristics of ,,1. /a?aricrs:

(i) It is an aq@ti. spe.ies living in extensivebraclish swanps, nany parts of whi.h are vii-tualty jmpenetFble to humans. Human popu-lations are currenily at reLatively low densitiesovei mu.h of this range in Sumatra and Kali-mantan (especially, in contrast to the very highpopulation densiti€s of humans in Java)- Thus,a high proportion ofthe habitat occupied by,4.iaran'crs is not currently haryested, no. is ita.cessible to harvesting unless there is a targeincrease in human poputation in this area.

(ii) Because water levels remain high in theseswahps year'round, ihe snakes appear not tobe.oncentrated (and poientially vulnerable toovefexploitation) at any time of the yeaf. Also,the continuous natuie of the available habitat

FILESNAKE BIOLOGY 359

over larSe areas fa.ilitates migration ofsnakes,reducins any impacts of lo.al over-exPloiiation

(iii) The snakes are coUected serendiPitously,as a by .atch of fishing activities using baitedhooks. Unlike other methods su.h as nets (asused for ,4. lltrdrrlarr -Voris and Glodel, 1980)or hand .ollectinS when populations are .on-centrated by seasonal aridity (as used foi,4. 4/dlari. Shine, 1986.), the Potential for a largepioporiion of any lo.al population of snakes iobe collected in a short period of time is verylow. Thus, reasonable PoPulations of thesesnakes persist even in the vi.inity of villages

(iv) The conmercial value of filesnake skinsis low, so that there is little economi. incentiveto exploit the resource ai a hiSher !ate. Thus,even were a more effective method of cat.hin8filesnakestobe developed, thele would be littlepressure to adopt ]t._

(v) Iemale filesnaNes reProdu.e relativelyfrequently (approximately two_thirds of ouladult females were reProdu.tive when colle.t-ed) and have a larEe littei size (nean : 29.3offspring per litter). These attributes should fa-cilitate recovery from shorrterm overexploitation of local populations-

(vi) Because both sexes of A. jovdii.us attainlarge sizesandare relatively heavy bodied,bothprov ide 5ur table s l rns for the.ommercia l t radeThus, boih se\es i re l ike ly to be harvested ins i g r t i . r n r n u m b e h { 4 . h a . l h e . a s e ' n o u r " d " l.dm plp, . rdrhe- rhdr e\p lo. t . l ion (o1! en, rar inBun lemJ.e. . rh. dpmosrdPhicdl t more.mPoFtantsex (asoc.ursinA. ardFfd. Shine, 1986c)

I h- .F.hJr .c te- i ' r i ( - u , 4. laJ l r i , wh, .h -u. ritwell to commercial erploitation ale not sharedby A. ot a l a at Al lpdst .n lne AL. l€ l ,dn hdbi !dt5in which it has been studied to Cate, A. alafwdPseems poorly suited to withstand any signiiiLdn ' .ommphrdl hdrvesr ,nS. l ls hdbi lars dretudErenrpd r iso ldr .d b i l ldbonS:) dnd gencrJhd<cesrble to h i r !est ing. i t i5 seasonat ly vu lnerable to hand .olleciingin large numbers because of the .oncentration of PoPuiations during,er-on. l orouEl ' r , . , ' re ldt ' \e l \ rh,n_bod_iad ( -o l_dr 'hc - l .n . e-pp. id l l ! rho 'e ot ndie. .are e. . \a luable lhdn dre l l 'o"e o i ,4. l4r ,4n, r .dno ferd le, r .produle o l ly o. .drondl lv . $ . rhgFat .limatically-indu.ed stochastic variation,r reprodJct ive ou 'Pu' r rd PoPuld l 'on ! l ru(_rLret \h inF ld8bd. Houslon. loo2: Hou"ton dnd<hi le . lqo4r . Al -o. Erdvid fera le ' dre Pd-t i ,ul " r l y v u l n e t u b l e l o \ r n d . o l l e i t ' 1 8 ( S h . n etd8bb). Al l o l the\e.hdidc(er is t iL ' 'u6ge- l lha l4. 4 j4 ld,4 ,n Au.r rd l .J would bF poor ly "u, (edto (ormcr. i r l e \P.o i tdt ion f .onomic cor5 iderations reinforce this con.lusion, since skinsfron,1. afaFla. would have to compete on th€' n r e r l d t o n J l n d r l e r i s i l h , l . n s f r o r l h e 1 J 1 8 -e t 4 t o t n ' , u - L o b o r , o s l . d t . . o 1 - r d c r d b l v

higher in Australia than in lndonesia. Thus, ourdata suSSest that any lonS+elm comme..ialex-ploitation of Austmlian poPulations of ,4. a/a-lrlre is unlikely to be either e.ologically suntainable or commer.iaily liabIe. Actochardus jo,4fl t.rsappears tobe better-suiied io harvesting,but further data are needed before any recom'mendations can be made about sustainable lev'

A ckioul. *enekts. -w e rhank esPecially Mr'Hasan and his staff at Budiman, who gave sthe opportunity to cary out this study, and werevery helpful in all stages of the work. I. BYuwo^o assisted with logiEtics and data colle.tion. The work would have been imPossiblewithout the support of LIPI (especially, Dr. Soe-tikno Wiioatmodjo) md Perennel of the PHPA.We thank R. W. G.lenkins and G. J. W. Webbfor insights into the .ommercial utilization oIreptiles, and H. K. Voris for allowin8 us to er-a;ine spe.inens cotlected durinS his Malay-sian studies. The Abstract was translated by LGuldberg and R. Kumudawati. Financial sup_port for the ttudy was plovided by the Austra_lian Reseal.h Committee.

LMUreRE C'fD

BBCMAN, R. A M. 1958. The anatony oi the A._ro .nordrnae. f o , a ldJ . l tP lPn-h ,5 re 'c ,6 l145 184.

B @ L A , I r o o 4 . \ o r e . o - r h p t l c p h d n ' . 1 ' J n \5ndtp dnd rne tJh- td .ed r d tp r \ - r le n | iud dLunpur. Malayan Nature ]ournal 18:179-r&

DE Root, N. 1917. The Repriles oi the Indo Austra.lian Archip€laAo. E.I. BriU, Leiden.

Do* Nc ! l C .obo. lhpru ' J - - f .ed n8 l 'db iFothe Java wart Snake. Aninal KinSdon 6313 15.

D . N . o N l ^ . A . s D \ A V r ! r o \ . . 4 7 8 . D ' \ e r q i l ) .J i " rnbur .on d-o .oLo8) o f lh l !P ' r ' d r n "srakes (Reptilia, Serpent€s). I. l{crPetoL. 12:281286.

f ' r l . rMrD L . q . f B (cu l . $DL fuonr laq j .The reproductive .y.le and lhe size at naturity oITuti,a,tbk tula{ens lqawiat Ttiidae) in lhe dryChaco or Argertina J. Herpetol. 27:70 78.

uoRlA u ' I l r Hr . $o I ML(o uv . lo8 lAn- -11 rpprodu f l \e pd f { rn \ n lh rPe.PP! i ! ' o lr ' n r p n d l e . l r o n r h ( ' e l r " f h r . r p p n e l H P rpetol. r5335 354.

Hor - roN D I loo2 t .o log) v l rP 6 lP .no lP /4,o ,aJ , , - , d r i . , . . rh .D.D 's i r q \dne) .Ncw.

Hou. ro \ n . A \D R. qHNr lao l 5 . \Ld d imo 'ph -m 'nd n . i 'edr \e rge- re : l "Pd ' r8hdb 'For h .Ar.tura fitesnake.I Anifu E ol.62.737-749.

Idq4. I o { d row ' l ' r " tP \ dndd" d ) .a nd l - rduon n A , lu r i 6 le .ndre rs ' rpen le \ . Acro ' horq id re \ 'n ' -oPrd l Au ' r - l d ( ope!,a 1994:726-73r.

LEWN, C. A, AND H. K. Vo{s. 1981. A conParisono l 'eP odu. l . \ " { rd re8 'e ' r fonS n dr ine i .1 reJ. Anim. tjcol.50:89 101.

360 R. SHINE ET AL.

I-rLLWHE, H. B 1991. Tne biology and consena-tion of acrochordid snrkes. Hama.lryad 16:1-9.

LuxMooRt R., ̂ ND B. GR@MBSDGE. 1989. Asianmonitor lizards. A revi€w ol distribution, status,€rpLoitationand trade in four scl..icd rp..ics. DraftRcport to th. CITES Secietariat, World Conser-vation Monitoiing Centre, CambridAe, U.K.

MacNUssoN, W. t. 1979. Irodudnn of an .nbryobt d n ,4.ro.lord, ! iar,r,./s iso lated for seven yeaE.Cope\r 1979:744-745.

MCDoMLL, S. B. 1979. A.alalogueofrhesnakesolNew Cuin€a and ihe SoLomons, with rPecial ref-erence t. those in thc acrnice P. Bisho! Mseum.Part III Boinne nnd A.rochordoidea. J. Herpetol.13:I-92

SECEL, R. A., AND N B FoRD 1937. Reprodu.tiveecologl. l, R ,\. s€i8el, J. r' Collins, and S. 5.Noval (eds.), Snakes: EcoloSy and EvolutionaryBiology, pp. 210 252. Ma.millan Pub. co., New

SIrNE, R. 1986a. tcoLosy oIa low-cncrgy spccialist:lood hlbits and rcprodu.tiv. biology of the Ar.-fuia lilesnale (Acrochordidae) Copeia r986:424437.

-. 1986b. Pr€dation upon filesnakca(,4.'o.rofd6,irrl,/ar) by aboriginat nunters: *lectivily withreslect to size, s€x and reproductive condition.Copeia 1986:238 239

. 1986c. SeruaL difier€nces in morphology and

niche utilization in an aquatic snake,,4.ro./,ord!sar,rtrac. oecologia (Berlin) 69:260 267.

-. 1992. Relative.lutch nass an.t body shapcin liards and snakes: is re!.oductive investmentconstrnine.t or optimizcd? Evolution 46:828 833.

SENE, R.. ND D. HousroN. 1993 Acrochoididae.L D. Walton (eds.), Fauna and Flora ofAustralia,Volnmc 2, pp.322-324. Aust Govt. Publ. Servi.e,

-/ ,AM R. LNBE K. 1985. ?\ radiorelemctri.study of novemenis, thermoregulation and hab-itat utilizriior of Aralura hlesnales (Serpentes,Acrochoi.lldne). Herpetoldgi.a 41i351 361.

SMnn, M. A 1943. The fauna or Biitish India, Ccv-lon and Blrna including th€ whol€ oi the rndo-Chinese sub-region. R.ptilia and Amphibia. VoLume III serpentes. Taylor and Fran.is, London.

TURNER, L B 1977. The dyna mics ol populatiors ofsquamates, .ro.odilians nnd rhy n.occp h! I i a n s, !p.157 26,1. ln C. Gans and D. w. Tin(le (Ids.), Ri-olo8t of tne Reltilia, Volune 7. Academlc Press,

Votus, H K..^NDC.5 CLoDE( 1980. Habirat,dret,ard reproduction of the 6le sna(e, A./d.fi./r,s8rd,,lrhr in the Straits ol Mala.ca. J. Herpetol14 :103-111.

A...picd:2 April 1995.