CONTRIBUTIONS TO THE BIOLOGY OF THE QUEENLESS PONERINE ANT DIACAMMA CEYLONENSE EMERY (FORMICIDAE) (Withfive text-figures) Vedham Karpakakunjaram1 Padmini Nair1 Thresiamma Varghese1 George Royappa1 , , , , Milind Kolatkar2 and Raghavendra GADAGKAR1’2’3 Key words: Diacamma ceylonense queenless ant, Ponerinae, Formicidae, Hymenoptera , All ants are generally classified as highly eusocial. However, there are some 100 species ofants belonging to the morphologically primitive subfamily Ponerinae, which lack a morphologically differentiatedqueen. One orasmall numberofmated workers (gamergates) function as queens, i.e. they produce male as well as female offspring. Such ant species are ofgreat interest as they provide unique opportunities to understand the causes and consequences ofqueenlessness. This is the first report ofa long-term field study we have initiated on a large population ofthe queenless ponerine ant, Diacamma ceylonense Emery, on the campus ofthe Indian Institute of Science, Bangalore. Data are provided on the numbers ofcolonies present during a 213-week period, from which the probable time ofcolonisation ofthe site by D. ceylonense is suggested. Twenty-six entire colonies were excavated to study the adult and brood composition. While some colonies did not appear to have a gamergate at the time ofexcavation, others had only one gamergate per colony. Only the gamergate was mated and had well developed ovaries, while all the workers were unmated and had undeveloped ovaries. Excavated colonies had an average of230 adult females, 2 males and 80 items of brood. The gamergates were indistinguishable from their workers in body size but there was significant variation in the size ofworkers between colonies. On an average, colonies employed 24% oftheir workers outside the nest for foraging and other duties. The relative constancy ofthis proportion permits estimation oftotal colony size by merely estimating the number ofextranidal workers and thus without having to excavate the nests. In the five colonies studied, foragers spent an average of 9 to 23 days oftheir lives in performing foraging duties. Most foragers remained faithful to one oranarrow range ofcompass directions, although such directional preferences ofall ofthem put together permitted colonies to exploit resources in all directions. Introduction In this scale of social evolution, all ants are generally classified as highly eusocial (Holldobler Eusocial insects are defined as those and Wilson 1990, Bourke and Franks 1995). insects whose colonymembers exhibit overlap of However, some 100 species ofants, belonging to generations, cooperative brood care and the morphologically primitive subfamily reproductive caste differentiation into a queen Ponerinae, lack amorphologically distinguishable caste and a worker caste. Highly eusocial species queen. Their colonies consist only ofworkers (in are those in which the queen and worker castes additionto males ofcourse), one or a few ofwhom are morphologically differentiated (Wilson 1971). are mated and take on the function ofqueens, i.e. production of male and female offspring. These 'Centre for Ecological Sciences, mated, reproducingworkersaretermed gamergates Indian Institute of Science, Bangalore 560 012, Karnataka, India. (Peeters 1991). Queenless ants are ofgreat interest Evolutionary and Organismal Biology Unit, for several reasons: (1) They provide an Jawaharlal Nehru Centre for Advanced Scientific opportunity to understand the conditions under Research, Jakkur, Bangalore 560 064, Karnataka, India. which the queen caste may be lost. (2) In the Email: [email protected] absence ofwinged queens, new colonies have to JOURNAL, BOMBAY NATURAL HISTORY SOCIETY, 100(2&3), AUG.-DEC. 2003 533 BIOLOGY OF THE QUEENLESS P0NER1NE ANT be necessarily founded by walking gamergates, Diacamma ceylonense Emery, which is which must result in a rather unusual population abundantly distributed in and around Bangalore. genetic structure. (3) They can be used as model There is a large population of D. ceylonense in systems to understand how one ora small number the so-called Jubilee Garden in the campus ofthe of individuals can establish and maintain Indian Institute ofScience, Bangalore (13° 00' N reproductive monopoly among a group of 77° 32' E). Such an easily accessible population identical, ornearly identical, individuals. This is a facilitates observation ofcolonies in their natural problem that has been extensively investigated habitats. Most previous work on Diacamma and , with bees and wasps (see Gadagkar 2001 and indeed on most ponerine ants, has been largely references therein) but because queenlessness restricted to observation in artificial nests in the in ponerine ants is evolutionarily derived from laboratory, ofants removed far from their natural the queenright condition (Baroni Urbani et al. habitat. 1992), a new perspective is expected from a study For all these reasons, we have initiated a of these ants. long-term study of the population of In this context, the genus Diacamma D. ceylonense in the Jubilee Garden ofthe Indian represents a most fascinating example. In Institute of Science, Bangalore. This is the first Diacamma all females eclose with a pair ofclub- report of this ongoing study. , like thoracic appendages called gemmae. The gamergate mutilates all workers who eclose after Material and Methods her, by clipping their gemmae. Such mutilation results in poorly understood neurological changes This study was conducted in the Jubilee in the victims, making them incapable of sexual Garden, Indian Institute of Science, Bangalore. calling and mating. This unique mechanism by The site is a plantation ofAcacia auriculiformis , which the gamergate maintains her reproductive with a few individuals of Polyalthia longifolia, monopoly seems to have a built-in mechanism to Tamarindus indica and Ficus sp. The area (about ensure that a colony does not necessarily die if two hectares) was divided into 60 line transects. its gamergate dies. Mutilated workers do not Initially, an extensive search was conducted to mutilate others so that, upon the death of a locate the nests ofD. ceylonense. All nests were gamergate, the next individual to eclose retains marked with sequentially numbered aluminium her gemmae, mutilates all those that eclose after tags and the tree nearest to each nest was marked her and takes over as the next gamergate ofthe with red paint. Every Monday, the nests were colony (Fukumoto et al. 1989, Peeters and monitored for activity and for the occurrence of Higashi 1989). While this description fits all other males. Nests that appeared to be abandoned were species ofDiacamma examined so far, there does checked by inserting a stick inside the nest appear to be an exception. There is at least one entrance and waiting for the appearance of species in which the gamergate seems to be workers. In addition, such nests were also capable of maintaining reproductive monopoly monitored for activity during the subsequent 4-5 without mutilating her workers. This species, weeks, before confirmingthem as abandoned. On which was first found at the foot of the Nilgiri the first Monday and Tuesday of every month, hills, and mistakenly identified as Diacamma the entire study site was searched systematically vagans Smith (Peeters et al. 1992), is yet to be to record any newly initiated nests. Data thus described. We therefore refer to it here as obtained, during 213 weeks, on the numbers of Diacamma sp. from Nilgiri. Diacamma sp. from nests, numbers of newly initiated nests, Nilgiri is morphologically very similar to abandoned nests and number of males, were 534 JOURNAL, BOMBAY NATURAL HISTORY SOCIETY, I00(2&3), AUG.-DEC. 2003 BIOLOGY OF THE QUEENLESS PONERINE ANT analysed in relation to abiotic factors such as observations and continuing throughout the maximum and minimum temperatures and total period ofobservation. All extranidal behaviours rainfall usingthe Kendall’s correlation coefficient. visible to the observer, positioned near the nest Climatic data were obtained from the India mound, were recorded. The 852 hours of data Meteorological Station, Bangalore. As there was a thus collected, were used to compute the number monotonic increase in the number ofnests during the of extranidal ants active per day, the number of first 120weeksofstudy, alinearregressionmodelwas days for which different ants were active, and fitted to these datawhich was then used to predictthe the rate per hour at which they performed probable time of colonisation of Jubilee Garden by extranidal activities. When the ants left the Diacamma ceylonense. vicinity ofthe nest mound, the compass direction Twenty-six colonies with all individuals and in which they departed was recorded in eight brood were collected from within the Indian discrete directions namelyNorth, Northeast, East, Institute of Science campus, but outside the Southeast, South, Southwest, West and Jubilee Garden. Nest characteristics such as height Northwest. When the foragers returned, what ofthe mound, number ofbrood chambers, depth they brought back (food, building material or of the nest and commensals encountered, were nothing) was noted. The proportion ofsuccessful recorded for all these nests. Head width and and unsuccessful trips made by foragers who alitrunk length ofall individuals were measured in had made at least three trips were then computed 21 of these colonies. Data on head width and and compared. A successful trip was defined as alitrunk length were subjected to one-way analysis the act ofreturning to the nest with either food or of variance to partition variance within and building material. An unsuccessful foraging trip between colonies. In the nine colonies that had a was one in which the ants returned with neither gamergate, one-sample t-tests were performed to food nor building material. Data on foraging compare each gamergate with her workers. In 15 directions of all active foragers were subjected ofthese 2 colonies (including the nine that had a to Rayleigh’s test (Batschelet 1981) of 1 gamergate), the numbers of mature eggs, width randomness to determine the degree of and length ofproximal oocytes and the presence randomness in the foraging paths, both for (or absence) of sperm in the spermatheca, were individual ants as well as for colonies as a whole. recorded for all individuals by dissecting their Another eleven colonies were excavated in gaster. an attempt to predict the colony size from the Five other colonies were chosen for numbers ofants engaged in extranidal tasks. Over observation of the extranidal behaviours. Four a period of three days, all ants seen outside the ofthese colonies were observed for four weeks nest were marked. At least two persons spent 3 each while one was observed for 55 weeks. Each hours and 30 minutes each, per day (30 minutes at observation session was of4-hour duration and the beginning ofevery hour from 0600 hrs-1800 was conducted either in the morning (0600 to hrs) for three consecutive days, marking the ants. 1000 hrs), midday (1000 to 1400 hrs) or afternoon Pilot experiments had indicated that this effort (1400 to 1800 hrs). One each of the morning, was morethan adequate to markall extranidal ants. midday and afternoon sessions were completed On the fourth day the nest was excavated to in each week, thus yielding 12 hours ofdata per determine the colony size (these nests were then week, per nest. All new workers seen outside the used for other experiments not reported here) so nest were uniquely marked with quick-drying that the relationship between the number of coloured paint on the thorax or abdomen, starting extranidal ants and the total colony size could be two weeks before the commencement of ascertained. JOURNAL BOMBAY NATURAL HISTORY SOCIETY, I00(2&3), AUG.-DEC. 2003 535 BIOLOGY OF THE QUEENLESS P0NER1NE ANT Results and Discussion Table 2: Co-inhabitants observed in Diacamma ceylonense nests Diacamma ceylonense nests were found Commensal No. of nests in which to occur quite commonly in the study area. The observed nests are subterranean, with a distinct mound at Cockroaches 12 the entry of the nest. The mounds are often Crickets 11 decoratedwith drytwigs, leaves, insect and spider Pseudoscorpions 9 Bagworms 3 exuviae, and occasionally, bird feathers. Moffett Grubs 1 (1985) observed that these decorations are used Earthworms 1 to harvest dewdrops, which the ants drink in the Isopods 14 mornings. The excavated nests contained 230.5 Other ant species 1 ±108.6 adultfemaleants, 2.3 ±5.3 males, 18.5 ±19.4 eggs, 22.6±18.7 larvae and40.0 ±36.6 pupae. The yet complete. Of the 15 colonies used for nests were 63.9 ± 20.6 cm deep under the ground dissection, 9 had one gamergate each, while the with 4.1 ±1.2 brood chambers and the mounds remaining 6 colonies did not have a gamergate. were 3.2 ±4.1 cm above the ground (Table 1). The number of mature eggs in the gamergates Males were usually found in the first chamber, ranged from 0-24 with amean of8.78 ±10.28 eggs. closest to the entrance. The next 2-3 chambers None ofthe workers had developed ovaries and had brood and the last chamber at the bottom none ofthem were mated in any ofthe colonies, contained refuse. We often found cockroaches, including those without a gamergate (Table 1). isopods, crickets, pseudoscorpions, bagworms, Thus D. ceylonense colonies are monogynous grubs, earthworms and a few other ant species in and recent molecular studies suggest that they the nests; we presume that these were commensals are also monoandrous but with some serial (Table 2). Nine out of the twenty-one colonies polygyny (gamergate turnover) (Gopinath 2002). excavated, seemed to lack a gamergate; the For both head width and alitrunk length, possibility that the gamergate was lost during there was a significant added variance between excavation seems unlikely but cannot, ofcourse, colonies. In the case of head width, this added be entirely ruled out. The remaining 12 colonies variance between colonies was significantly invariably had only one gamergate each. In two higher than the variance within colonies, while in nests, one or two individuals other than the the case of alitrunk length, this variance was gamergate also had gemmae, but only on one side significantly less than the variance within oftheir thorax; perhaps their mutilation was not colonies (Table 3). Variance between colonies is Table 3: One-way analysis of variance of head width and alitrunk length of individuals of Diacamma ceylonense from 21 colonies Source DF SS MS F P Head width3 Among colonies 20 1811.07 90.55 791.42 0.001 Within colonies 4731 541.31 0.11 Total 4751 2352.38 Alitrunk lengthb Among colonies 20 1151.20 57.96 94.61 0.001 Within Colonies 4683 2849.00 0.61 Total 4703 4000.19 3Variance between colonies significantly higher than variance within colonies bVariance between colonies significantly lesser than variance within colonies 536 JOURNAL BOMBAY NATURAL HISTORY SOCIETY, I00(2&3), AUG.-DEC. 2003 — —— —— — —— — —— —— — —— — - BIOLOGY OF THE QUEENLESS PONERINE ANT MO— CD CO CD CD on n M oM" to—c• 0^•> CO CO CM CM CM _s*oc(-zJD>• §2EQ.o°O*ESc 11 11 1I 1I 1I 11 11 I1 1I odCd-oDH oCd+O1 Cddn-DH Cdhd-~OH 1i 1I Cd-OH d-H i1 Co4+Mi C4Cd+OO1 dCdn+M1 1I 1| 1j 11 11 -0vC0^C_D3 TiO3oOs a !E a> oCD CO O OM" CoO CoO CoO o00 $ 0 snz proximal d d d o d d d d d 0 5 oocyte (mm) 11 !i 1i 11 11 1i !1 11 I1 +1 +i H +i 11 11 +i -H |1 +! +i +i 11 |I 11 11 I1 ° 0 4CdO cor^o- CMO" CCdMD 0Cd0O CdMD- CdMO- dM; CCdOD S*-Z• *o0-> d m1 HO— 0t3— XC)/» o co o o co CD •m- co o 0o0d CdO CCuDD .~m505 ceylonense Zd TEO OOcpJ) 11 1I 1I 1I 1i 1[ 1| 11 1| 4 1I I1 4— I1 CM CM 11 11 11 1| II 4-H 03 005 M" o CO in m CD o CD C—O CD n CD M- CO co M- r^- 'vT 4— ETO ^E> O d d d d d CdM 4o 4o d d 4o d d d 4o 4d CdM d d d CdM c 03 . oCO TO Diacamma 3L. |S4o—Z>« £E 11 C-DH i+ni i-nH -H o+i -H | j1 Cf+-Di- To+~1 oi+n1- OC+Di CoHMH Cn-MH CC-MDH CC-MOH 0+0i n+1 CC+MDi CC+MOi 11 C4+M—i Co+Moi |1 CC+OMt C+O! hw-I1 0Co</) =<D) cd •M- O' cd M- M" M- M" •M- m- -M- M- M- M- M" M" CII o 3• - O TO in T> 0-3TO ^ iodn CodO CodD COdD CdOD dion rod-4- 0Od0 COdD nod Ooh- odD- CCMD od 0od0 0od0 0Od0 0od0 odn C0M0 0od0 Ohd~ T~ 0Eo EaE)? individuals X0CO 11 CC-DDH CC+MDi CC+MDi OC-MH Co+M1 oO-H 1| 11 oC-OH OC-DH 4o-—H oo-H on+i oT-—H or+s~1- OM+-i qC+M! oC+Oi OC+Di oC-DH 1| oo+i CO+D1 11 OC-DH OM+"i I>D3 CCD tCoD 4—1 CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CM CII (/) ° T0O stzz <1/—) Q) the JQ 4— 3 >4 of >ZOd4— •-°oEoQ _tomrcoo M- CO m CO CM M" in "fr co CD M- CO n CM n n h- CO M- n n M' M- n CM ±1.2 3t^f) oCOO P fc 03 ? measurements Heightof mound (cm) o CO o o O O o in O r^- o O o o O n o o o O o 10 00 n 00 10 CcMd —c0z c3 +i O 4— ZcS. o body JLn•C04ZJ—L» *v0COt>r —£O oin oin Toi_n hm~ om mCD imn CmD CinD mM" on oD- CoD CoD oCD on 0n0 CoO CoO CnD CoD on o nCD oh- nCD CcCDdD CdC+DM1 T^d33 ColD CZ CD 4- 0) o TO LLJ and q CO m m o o m m O O o n n o O d CD SZ 13 Zri 3QQ.. T_ T“ 0b0- - CO T_ oCO iO'n CO co CD o4o TC_M n CCOM c0o0 h- C4O T4_— CMD- CnO CM-D Cc+Doi "5O § 0 T3 X£3 composition HZod- CT£OO "T—" cTo_ CCMO 4m— oCM CCDM imn Ti—n CO C4—O CCDO CO O o o 4— ccoo oT— 4<—M CCDM 4— CD cno CMO" CCDO o qCCMM oh-d cC0oO 3"cS +i Colony HZoo— Cy0O)D) T3 o Oa oCoM 00 CCOD CMM- 4 CM-D CCDD o CCOO o o P0 cO’o) CnO CMM" M•M"- o O CD 44 44 o n4od 4CM+—-D1 *T04O3C*/3“3 0TEET30OO 1: *o4— c0o CO CO 0 T0O Zd CfE—O o o o CO O oo o O o O T_ o o o o o 4CO o 4— O4— CCMM o O o o 4— CM n+1 0$ E>4 Table c0/> i5„ "oO oTO >o4— n q T3O oco 0 o «T£04O— oCCCMDM 4COT’O* <CCOuD'D C4C—DO CoOD' CiTO_n <oCMDM- oOCCMM o0T~0 4nC—M C0CD0M 4C—D C44—“D 00CT30O_ Q04)—0 onCM cCCMoD 0o0T)0_ C4T—“O CChM-D CCCDDM orcC^oM- 0OoC)M 0Cc43Mo oCCCMDM 0n43J- dcCoM ood “•Ei>Ti3 £^•O- STE—ZO +i 10 O>coo4 T0OOO Q v_ CM CO O' in CD D- CQM 00 CD OT_ T_ C4M— CT_O 4M—' n4~ C’rD" h,p-" 0v~0 CT“D D-3 OCM CM QM-i CCMM '^O0 E0 -CC0T/OL) 3 <Q«oCC+*—D-O* "ocn0oo CoiC4O>—nM C4oiC—DnM CCOCMMDM-M" Coc4cD—do CCrOD-M SCi^orMn CCC0ODMD0 COC•0MDM0- C0CMcoD"0Md C0OOM4"D0O MO0CCC"DDM0 COTCM"D—M CO4I4f'D—) COCC4DMD— C4no4c—Dd CCiqoDMn CocnoqDdd CnoioMDrn Cnno4D— CCnCo4DDD— Cn4hcqD~d CnChnoD-M qoqCncoMd CCnoc4DD—d CCno4nDD— adCn4—>M Z<C O4oC32JQN°D. Sao3%Z (EVC<Q0//—).) JOURNAL BOMBAY NATURAL HISTORY SOCIETY, 100(2&3), AUG.-DEC. 2003 537 , 1 BIOLOGY OF THE OUEENLESS PONERINE ANT 200 likely to have a genetic basis and variance within colonies, a nutritional basis. As expected from the factthatthe gamergate is simplyan unmutilated worker, in 17 out of18 tests (9 colonies, eachtested for headwidth and alitrunk length), the gamergate was not significantly different in body size as comparedtothe workers in the colony (one sample t-test; Table 3). The number of active nests steadily increased from the commencement ofthis study (December 1993) up until 120 weeks. After that the number ofactive nests decreased to an extent (Fig. 1a). In our213 weeks ofstudy, we recorded a total of 1 17 nests, active at one time or another. Although new nests were initiated and abandoned throughout the study period, initiations usually outnumbered abandonings during the first 175 weeks, while abandonings outnumbered initiations during the next 38 weeks (Fig. lb,c). Twenty-one out of 117 nests were active throughout the 2 3 weeks ofobservations. Given 1 that they were already present at the beginning of the study, they must have been older than 213 weeks. Males were few in number but were recorded more or less throughout the year (Fig. Id). Climatic data are depicted in Fig. 1e. There was a positive correlation between minimum temperature and number of nests initiated (x = 0.26 at p<0.008) and abandoned 150 200 Week (i = 0.38, at/?<0.000 ). There was also a positive 40 450 1 correlation between rainfall and number ofnests initiated (x = 0.43 at /?«0.01) and abandoned O 300 o (t = 0.39, at p<0.0001) and number of males A~\A CL 20 JvT Rainfall (x = 0.29, at p<0.005). Maximum temperature £ - 150 showed no such correlation. j UL uliJl Since the number ofnests increased steadily , duringthe first 120 weeks, a linearregression model Dec Dec Dec Dec 1093 1994 1995 1996 1997 was fitted to the data (Fig. 2). The model is highly Month significant. Extrapolatingbackwards in time, using the slope and its 95% confidence interval, we may Fig. 1 : Thenumberof(a) active nests, (b) nests initiated, (c) nest abandoned, (d) males and (e) mean monthly hypothesize that the Jubilee Garden was colonised rainfall (bars), mean monthly maximum temperature by D. ceylonense some 85 to 1 13 weeks before the (upper line) and mean monthly minimum temperature beginningofthe study, i.e. between September 199 (lower line), during the 213 weeks of census of the and May 1992 (Fig. 2). Jubilee Garden population ofDiacamma ceylonense 538 JOURNAL, BOMBAY NATURAL HISTORY SOCIETY, 100(2&3), AUG.-DEC. 2003 BIOLOGY OF THE QUEENLESS P0NER1NE ANT Fig. 2: Number ofactive nests from the 120-week data was used to fit a linear regression model which was extrapolated backwards using the slope and its 95% confidence intervals, to estimate the probable time of colonisation ofJubilee Garden by Diacamma ceylonense Diacamma ceylonense workers are extranidal tasks like nest cleaning and shifting of individual, diurnal foragerswith directional fidelity materials around the nest increased soon after within a narrow range ofdirections (as seen from rainfall. On finding an item of prey, the forager our observations given below). Their major food paralysed it by stinging and returned with it to consisted of termites, although bugs, the nest quickly and in a short, more linear path cockroaches, grasshoppers, crickets, spiders and as compared to the often long-winding path it other species of ants were also observed being had followed during its outwardjourney. When a brought to the nest on some occasions. Prey, termite nest or trail was discovered, the successful however large, was never retrieved in groups, but forager made repeated trips to the same source. there was some cooperation among the workers The mean number ofactive individuals seen near the nest to manoeuvre it through the nest outside the nest during any 4-hour observation entrance. Foraging continued for most ofthe day session varied from about 6 to 32 and the mean on colder days, but was restricted to the mornings foraging life span ofworkers ranged from 9 to 23 and evenings only, on warmer days. Foraging was days (Table 4). Foraging was the most frequently suspended during heavy rains although some performed extranidal activity. In four out offive Table 4: Number of active animals per observation session, and foraging life span in five colonies Nests No. of sessions & No. of active animals/session3 Foraging life span in days No. of days observed Mean ±SD (Range) Mean ±SD (Range) 071 165 sessions; 385 days 6.7 ±3.6 (1-17) 23.3 ±22.2 (1-154) 025 12 sessions; 26 days 17.2 ±2.2 (11-21) 13.2 ±10.2 (2-26) 056 12 sessions; 27 days 32.3 ±8.0 (22-51) 9.0 ±7.6 (2-27) 0117 12 sessions; 26 days 12.9 ±3.7 (5-18) 12.9 ±8.7 (2-26) 0134 12 sessions; 23 days 16.2 ±5.5 (4-22) 15.5 ±6.1 (2-23) a session = 4 hours 1 JOURNAL BOMBAY NATURAL HISTORY SOCIETY, 100(2&3), AUG.-DEC. 2003 539 BIOLOGY OF THE QUEENLESS PONERINE ANT Fig. 3: Proportions ofsuccessful (shaded bars) and unsuccessful (blank bars) trips made by foragers in five colonies (numbers above bars are the numbers ofsuccessful/unsuccessful trips) colonies (Fig. 3; Mann Whitney U test; Colony individual foragers were non-random with respect 71, 56and 134atp<0.001,andColony 1 17atp<0.05), to the directions in which they made foragingtrips, the number of successful foraging trips was butthat the colony as a whole (with all its foragers significantly less than the number ofunsuccessful put together) was random in this regard. In each trips. In one colony (Fig. 3, Colony 25), foragers colony, even the most random forager was less were somewhatmore successful in retrieving prey so than the colony as a whole (according to items, but the number of successful and Rayleigh’s test of randomness, p<0.05, see Fig. unsuccessful trips did not differ significantly from 5). The mechanism by which the foragers achieve each other (Mann Whitney U test, p>0.05). such collective randomness, in spite ofindividual Foragers who had made at least three trips were non-randomness, remains to be investigated. used for this analysis. The proportions of Experiments in which all the ants seen successful and unsuccessful trips were computed outside the nests were marked before nest for each worker and then averaged across workers excavation demonstrate that the number of ants in each colony. seen outside the nest is a reasonably constant Many foragers seemed to be faithful to proportion ofthe total colony (24.39% ±9.63%) certain foraging directions (Fig. 4), although the (Table 5). This information can be used to colony itself seemed to have foragers covering approximately infer colony size without all directions. Rayleigh’s test of randomness necessarily having to excavate nests. Besides confirms that in all five colonies studied, helping to avoid unnecessary excavation ofnests, 540 JOURNAL BOMBAY NATURAL HISTORY SOCIETY, 100(2&3), AUG.-DEC. 2003 BIOLOGY OF THE QUEENLESS PONERINE ANT Table 5: Colony size and percentage of extranidal ants Colony No. of Colony Percentage code extranidal size of foragers Colony 71 ants marked n=96; T=1830; DC28 69 457 15.10 DC29 89 376 23.67 100n DC30 135 593 22.77 80- DC31 23 192 11.98 DC42 23 121 19.00 60 - Colony 25 DC47 47 145 32.41 n=28; T=265; DC52 47 271 17.34 DC53 24 137 17.51 DC64 60 145 41.38 DC65 39 103 37.86 DC66 12 41 29.26 Colony 56 Mean ±SD 51.6 ±35.9 234.6 ±171.6 24.39 ±9.63 n=65; 1=231; the proportion of foragers is in itself a useful parameter in many kinds of studies (see for e.g. Andre et al. 200 Gopinath 2002). Investigations 1 , to determine intra-colony genetic relatedness using microsatellite markers, population genetic Colony 117 structure using both nuclear and mitochondrial n=23; T=94; markers and the mechanism of mutilation of the gemmae are in progress and would form the subjects of future publications. Acknowledgements We thank the Ministry ofEnvironment and Colony 134 Forests, Department ofScience and Technology, n=19; 1=66; Council for Scientific and Industrial Research, Department of Biotechnology, Government of India and the Jawaharlal Nehru Centre for 4 5 6 7 8 Advanced Scientific Research, Bangalore, for Number of directions foraged financial assistance, and several student trainees n = number of foragers observed and visitors, including Karthik, Nutan, Sumana, T = total number of trips observed Vidya, Yuvana, Andre, J.-B. and Ravi Kumar for help with marking ants in the field, for various Fig. 4: Frequency distribution ofnumbers offoragers experiments. We also thank Christian Peeters for who foraged in different numbers ofdirection helpful comments on an early draft ofthis paper. in five colonies JOURNAL BOMBAY NATURAL HISTORY SOCIETY, 100(2&3), AUG.-DEC. 2003 541 , BIOLOGY OF THE QUEENLESS PONERINE ANT N (a) (b) (c) T 239 15 W 244 „ . 268 h>- c 332 o % W^ 104 166 206 (n=96; r=0.13) lO 11 CM 11 c o 21 o O 25 (n=28; r=0.59) 25 CO to > c o o O 53 (n=65; p=0.11) (r=0.36) n = no. of foragers for the colony as a whole; r = randomness Bars are proportions of number of trips made by forager(s) in different directions. Values beside each bar indicate the number of trips made by one or more foragers (in a) and one forager (in b and c). Fig. 5: Foraging direction for (a) colony as a whole, (b) the most non-random forager and (c) for the relatively most random forager, in the five colonies studied. 542 JOURNAL, BOMBAY NATURAL HISTORY SOCIETY, I00(2&3), AUG.-DEC. 2003