Social evolution in a new environment: The case of introduced fire ants (queen number/relatedness/mating system/male diploidy/solenopsis invicta)

Transcription

1 Prc. Natl. Acad. Sci. USA Vl. 93, pp , April 1996 Evlutin Scial evlutin in a new envirnment: The case f intrduced fire ants (queen number/relatedness/mating system/male diplidy/slenpsis invicta) KENNETH G. Rss*t, EDWARD L. VARGOt, AND LAURENT KELLER *Department f Entmlgy, University f Gergia, Athens, GA ; tbrackenridge Field Labratry and Department f Zlgy, University f Texas, Austin, TX 78712; and Institut de Zlgie et d'eclgie Animale, Universit6 de Lausanne, Batiment de Bilgie, 115 Lausanne, and Zlgisches Institut, Universitat Bern, Ethlgische Statin Hasli, Whlenstrasse 5a, CH-332 Hinterkappelen, Switzerland Cmmunicated by Charles D. Michener, The University f Kansas Natural Histry Museum, Lawrence, KS, December 26, 1995 (received fr review September 15, 1995) ABSTRACT The inadvertent intrductin f the fire ant Slenpsis invicta t the United States frm Suth America prvides the pprtunity t study recent scial evlutin by cmparing scial rganizatin in native and intrduced ppulatins. We reprt that several imprtant elements f scial rganizatin in multiple-queen nests differ cnsistently and dramatically between ants in Argentina and the United States. Clnies in Argentina cntain relatively few queens and they are clse relatives, whereas clnies in the United States cntain high numbers f unrelated queens. A crllary f these differences is that wrkers in the native ppulatins are significantly related t the new queens that they rear in cntrast t the zer relatedness between wrkers and new queens in the intrduced ppulatins. The bserved differences in queen number and relatedness signal a shift in the breeding bilgy f the intrduced ants that is predicted n the basis f the high ppulatin densities in the new range. An additinal difference in scial rganizatin that we bserved, greater prprtins f permanently unmated queens in intrduced than in native ppulatins, is predicted frm the lss f alleles at the sex-determining lcus and cnsequent skewing f peratinal sex ratis in the clnizing ants. Thus, significant recent scial evlutin in fire ants is cnsistent with theretical expectatins based n the altered eclgy and ppulatin genetics f the intrduced ppulatins. Scial rganizatin-the number, relatinships, and reprductive utput f members f a grup-cnstitutes the defining features f scial life (1). The eclgical and genetic factrs causing evlutin f scial rganizatin in ants are f great interest t evlutinary bilgists, because knwledge f these factrs can illuminate the mechanisms by which diverse frms f cperative breeding riginate and are elabrated (1-7). In euscial insects, eclgical factrs are widely believed t be imprtant in shaping such basic features f scial rganizatin as clny queen number and queen reprductive rles (4-1), but demnstratin f the effects f such factrs remains elusive (5, 11). Specific genetic factrs influencing reprductive rles and the frm f scial rganizatin have seldm been dcumented (1, 12-16). The inadvertent intrductin f the fire ant Slenpsis invicta t the United States frm Suth America prvides a unique pprtunity t study recent scial evlutin by cmparing scial rganizatin in the native and newly clnized ranges. Such cmparative studies are especially pertinent because f imprtant advances in ur knwledge f the eclgy and genetics f the ants in bth areas. Recent eclgical studies have revealed that intrduced fire ants attain clny sizes and densities far in excess f thse that ccur in The publicatin csts f this article were defrayed in part by page charge payment. This article must therefre be hereby marked "advertisement" in accrdance with 18 U.S.C slely t indicate this fact. 321 the native range (17), presumably due t eclgical release frm the effects f natural enemies that nrmally limit the survival and grwth f established nests. High ppulatin densities in the USA are likely t reduce the availability f new nesting sites as the pen, disturbed habitats required by this species fr nesting (18) becme increasingly saturated. Increased habitat saturatin is hypthesized t influence clny queen number and related elements f scial rganizatin in ants by favring the adptin f new queens int existing nests and disfavring slitary nest fundatin by dispersing queens (1, 5, 7, 11, 19, 2). Mrever, habitat saturatin may prmte incidental mixing f nnnestmates, ptentially erding the discriminatin abilities f wrkers invlved in the recruitment f new queens and the regulatin f clny queen number (21-24). Bth mechanisms predict a psitive assciatin between degree f habitat saturatin and clny queen number. Recent genetic studies have revealed that genetic variatin has been lst in intrduced fire ants because f a ppulatin bttleneck during funding f the USA ppulatins (25). Lss f variatin at the majr sex-determining lcus is especially evident, as it is assciated with a large increase in the frequency f sterile (diplid) males in intrduced ppulatins (25). Such an increase in male sterility may affect anther basic feature f scial rganizatin, the frequency f egg-laying queens that remain permanently unmated, by altering peratinal sex ratis and disrupting the mating system (26). Thus, clnizatin f the USA by S. invicta has been accmpanied by well-dcumented eclgical and genetic changes predicted t have significant cnsequences fr fundamental prperties f scial rganizatin such as clny queen number, nestmate relatedness, and queen mating success. In this reprt we examine differences in these prperties in multiple-queen (plygyne) clnies f S. invicta frm the native and intrduced ranges. The results indicate great similarity in scial rganizatin between lcalities within each range but prnunced differences between the native and intrduced ants. The significant evlutin f scial rganizatin inferred t have ccurred in intrduced ppulatins f the plygyne frm is cnsistent with changes predicted n the basis f the altered eclgy and ppulatin genetics in the new range. Of special interest is the transitin frm nests with few clsely related queens t nests with many unrelated queens, because the cnsequent rearing f unrelated sexuals by wrkers in the intrduced ppulatins pses an apparent paradx fr inclusive fitness thery (7, 27). MATERIALS AND METHODS The fire ant S. invicta was intrduced int the USA frm Suth America in the 193s (28) and has since expanded its new range rapidly. Tw scial frms f S. invicta exist in the native and tt whm reprint requests shuld be addressed.

2 322 Euin se l intrduced ranges, ne in which nests cntain a single egglaying queen (mngyne frm) and ne in which nests cntain multiple queens (plygyne frm). In the mngyne frm, clny funding is independent: newly mated queens start clnies withut wrkers after dispersing sme distance frm the parent clny (29). In the plygyne frm, the subject f this study, clny funding usually is dependent: queens and wrkers frm an existing clny emigrate as a grup t frm a new clny near the parent clny (3). Clny funding in the plygyne frm may ccur independently as well in sme cases, suggesting plymrphism in queen reprductive strategies. Queens f bth frms mate with a single male (25, 31), presumably nly (mngyne frm) r mainly (plygyne frm) during extensive swarming flights. The mngyne frm is presumed t be ancestral in the intrduced range (1), with lcal enclaves f plygyny arising secndarily in established mngyne ppulatins. Ants were cllected frm plygyne nests excavated at tw lcalities in Argentina (Crrientes, Frmsa) and tw in the USA (Gergia, Texas). The Argentine lcalities are separated by a majr bigegraphic bundary (the Ri Parana) and are genetically distinct (32); the USA lcalities are at ppsite ends f the intrduced range. All excavated nests were recgnized as cntaining multiple egg-laying queens because nestmate wrkers pssessed gentype arrays at several electrphretic lci incmpatible with their being the ffspring f a single queen (31, 33). The average numbers f mated egg-laying queens per nest were estimated indirectly frm the genetic variatin within nests and directly frm the cunts f cllected queens. The indirect estimate is the effective queen number, estimated frm the genetic relatedness f daughters f single queens (rs), the relatedness f nestmate wrkers (rw), and the relatedness f mated nestmate queens (rq) (33, 34). The value f rs is fixed at.75 because S. invicta queens mate nce (25, 31). Values f rw and rq were estimated frm electrphretic gentypic data frm a large number f plymrphic lci in each ppulatin (see legend t Table 1). Only lci exhibiting mendelian patterns f inheritance and shwing n evidence f being under selectin (15, 25, 33, 35) were used fr estimating relatedness (36). Direct estimates f the numbers f egg-laying queens are the arithmetic and harmnic means f the cunts made fllwing nest excavatins. The harmnic mean queen number is f interest because it is expected t equal the effective number (indirect estimate) under simple assumptins (33, 34). Variances fr the relatedness estimates and direct estimates f queen number, btained by jackknifing ver nests, were used t generate 95% cnfidence intervals by assuming the t distributin. Mating status was determined by dissecting queens and examining the spermatheca (sperm strage rgan). Prprtins f mated queens were estimated by using a resampling prcedure t avid clny-level effects. A single queen was drawn at randm frm each nest 1 times (with replacement), with the arithmetic mean f the 1 resampled ppulatin prprtins taken as the ppulatin estimate. The 95% cnfidence intervals abut these mean values were btained by drpping the 25 highest and the 25 lwest f the 1 resampled prprtins. Sample sizes fr estimating clny queen number, relatedness, and prprtins f mated queens are given in the legends t Tables 1 and 2. Lcal ppulatin genetic structure was assessed in Argentina by calculating Wright's inbreeding cefficient (Fis) as well as the relatedness between nnnestmates in neighbring nests. Estimates f FIs were btained frm allele and gentype frequencies generated by resampling single (independent) gentypes 1 times (with replacement) frm each nest included within a 5-km diameter deme (33, 37). Estimates were based n eight lci (15 alleles) surveyed in 516 ants frm five Prc. Natl. Acad. Sci. USA 93 (1996) demes in Crrientes and n eight lci (14 alleles) surveyed in 48 ants frm five demes in Frmsa. Estimates f relatedness between nests lcated within clusters f diameter < 1 m were based n eight lci surveyed in 276 ants frm 1 clusters in Crrientes and n seven lci surveyed in 162 ants frm 7 clusters in Frmsa. RESULTS Intrduced ppulatins f plygyne S. invicta differ dramatically frm native ppulatins f this frm in the number f egg-laying queens per nest and the relatedness f these queens (Table 1). The number f mated queens per nest estimated indirectly frm the genetic data is ver twice as high in intrduced ppulatins as native ppulatins, a disparity reflected als in the direct cunts f queens cllected frm the excavated nests. The increase in queen number in intrduced ppulatins is accmpanied by a prnunced drp in the average relatedness f nestmate queens. Relatedness values in Argentina are thse expected if queens ften are sisters r mther-ffspring pairs, whereas values in the USA are statistically indistinguishable frm zer. As expected, the increased number and decreased relatedness f egg-laying queens has led t a significant drp in relatedness between wrkers and the new queens they rear in the intrduced ants cmpared t the native ants (Table 1); indeed, wrkers in the USA are n average n mre clsely related t the yung winged (nnegg-laying) queens frm their wn nest than they are t thse frm ther nests in the same ppulatin. These differences in queen number and relatedness signal a significant shift in the breeding system f the plygyne frm in the intrduced range. Nests in Argentina must recruit new egg layers almst entirely frm queens that are reared in (and mate in r near) the same nest, because ther mechanisms that cnceivably culd generate the bserved high levels f queen relatedness can be discunted. Fr instance, high queen relatedness des nt result frm recruitment f queens frm highly inbred lcal demes, because inbreeding is insignificant (FIs + SE =.11 ±.14 in Crrientes and.6 ±.27 in Frmsa) and neighbring nests are nt clsely related (between-nest r ± SE = in Crrientes and in Frmsa). The lw relatedness f nestmate queens in the USA is cnsistent with data frm mark-recapture studies (38, 39), indicating that at least sme (and perhaps mst) new queens are recruited frm freign nests after having participated in mating flights. The native and intrduced ppulatins differ als in the prprtins f egg-laying queens that remain permanently unmated, with such queens ccurring at least twice as frequently in intrduced as in native ppulatins (Table 2). This difference means that the disparity in number f egg-laying queens per nest between native and intrduced ppulatins is even greater than indicated by the data in Table 1, because bth the indirect and direct estimates cncern nly mated queens. DISCUSSION The differences in clny queen number and relatedness bserved between native and intrduced S. invicta f the plygyne frm are cnsistent with predictins based n the different ppulatin densities in the tw ranges. Cnstraints n independent funding in the intrduced range are expected t generate increased selectin n queens t seek adptin int existing clnies and n members f established nests t accept such queens (5, 11, 2, 4). High ppulatin densities in the USA suggest that apprpriate nesting habitats are mre ften saturated, and pprtunities fr independent funding by dispersing queens are crrespndingly mre limited, than in the native range (19). Thus, queens attempting either t

3 Evlutin: Rss et al. Prc. Natl. Acad. Sci. USA 93 (1996) 323 Table 1. Number f mated egg-laying queens and nestmate relatedness in plygyne nests in native and intrduced ppulatins f S. invicta Native ppulatins Intrduced ppulatins Crrientes Frmsa Gergia Texas Number f mated egg-laying queens per nest Indirect estimate Direct estimates Harmnic mean* ( ) ( ) ( ) ( ) ( ) ( ) Arithmetic mean ( ) ( ) ( ) ( ) ( ) ( ) Relatedness f mated egg-laying queens (.3-.6) ( ) (-.3-.6) (.-.1) (-.3-.7) 199 Relatedness between wrkers and yung winged (nn-egg-laying) queens ( ) (.8-.23) (-.3-.3) The 95% cnfidence intervals fr direct estimates f queen number and relatedness are in parentheses. Values fr relatedness f mated queens (rq) were based n fur lci in Crrientes, seven in Frmsa, tw in Gergia in 1984, six in Gergia in 199, and six in Texas. Relatedness between wrkers and yung winged (nn-egg-laying) queens was estimated frm five lci in Crrientes, six in Frmsa, and seven in Gergia. Values fr the relatedness f nestmate wrkers (rw), used t btain indirect estimates f queen number, were based n eight lci in Crrientes, seven in Frmsa, tw in Gergia in 1984, and eight in Gergia in 199. All relatedness estimates fr Frmsa were crrected fr the lw but significant relatedness bserved between neighbring nests. Sample sizes fr estimating relatedness are as fllws: Crrientes, 528 wrkers frm 44 nests, 181 queens frm 26 nests, 17 winged queens frm 2 nests; Frmsa, 399 wrkers frm 34 nests, 18 queens frm 2 nests, 165 winged queens frm 22 nests; Gergia (1984), 826 wrkers frm 2 nests, 1261 queens frm 26 nests; Gergia (199), 895 wrkers frm 27 nests, 765 queens frm 27 nests, 1662 winged queens frm 27 nests; Texas (1995), 31 queens frm 3 nests. Sample sizes fr direct estimates f queen number are as reprted abve fr Crrientes, Frmsa, and Gergia (199). Sizes f the remaining samples are as fllws: Gergia (1985), 252 queens frm 74 nests; Texas (1987), 66 queens frm 33 nests; Texas (1995), 78 queens frm 2 nests. *Harmnic means f the direct cunts f queens are expected t equal the effective queen numbers btained indirectly (33, 34); the fact that they cnsistently are 6-7% lwer suggests that similar prprtins f egg-laying queens eluded capture during the excavatins in each ppulatin. remain in their natal nest r t enter anther nest t becme egg layers may have higher average reprductive success than queens attempting t fund nests independently, even thugh queens using the frmer strategies have reduced fecundity cmpared t independently funding queens (41). Wrkers initially are expected t accept nly new queens riginating frm their wn nest, but their nestmate discriminatin abilities may weaken as queen number increases (21-24), pssibly leading t a runaway prcess f acceptance f bth nestmate and freign queens (7, 42). Incidental mixing f nnnestmates in dense ppulatins may accelerate this breakdwn in discriminatin abilities. A majr cnsequence f this increase in queen number is a decrease in relatedness amng clny members, with the effect that wrkers in the USA typically rear unrelated sexuals. Plygyne ants generally can be characterized by the alternative syndrmes f few clsely related queens r many distantly related queens per nest (7, 2), suggesting that the transitin frm ne syndrme t the ther in S. invicta reflects a fundamental eclgically driven shift in scial rganizatin in ants. The difference in prprtins f unmated queens bserved between Argentina and the USA supprts the prpsal that frequent male sterility caused by a lss f genetic variatin at the sex-determining lcus has led t the elevated abundance f such queens in the intrduced range (26). Lss f alleles, dcumented in intrduced S. invicta fr a large number f genes (25), is the expected cnsequence f a ppulatin bttleneck such as that which presumably accmpanied initial establishment f the ants in the USA (43). Such a lss is likely t have a prnunced effect n sex determinatin and sex ratis in scial Hymenptera such as fire ants because f their unique genetic mde f sex determinatin. Sex is determined via a system f genetic cmplementarity whereby diplids heterzygus at the sex-determining lcus becme females, and haplids (hemizygtes) r hmzygus diplids becme males (25, 44). A reductin in sex-allele diversity leads t increased hmzygsity at this lcus and cnsequent elevated frequencies f males that are diplid; thus, such males are cmmn in the USA but uncmmn in Argentina (25). Because diplid males are sterile (45) and appear t be prduced at the expense f haplid males (1), the frequent ccurrence f diplid males in the USA creates a highly female-biased peratinal sex rati (estimated at 6.2 queens per fertile male in the Gergia plygyne ppulatin; ref. 46). The apparent cnsequence f this female-biased sex rati in an rganism in which bth sexes mate singly (31, 33) is that a Table 2. Prprtin f egg-laying queens that are permanently unmated in plygyne nests in native and intrduced ppulatins f S. invicta Intrduced ppulatins Native ppulatins Gergia Texas Crrientes Frmsa (.-.7) (.8-.19) ( ) ( ) ( ) (.16-.4) The 95% cnfidence intervals abut the values are shwn in parentheses. Sample sizes fr estimating the prprtins are as fllws: Crrientes, 185 queens frm 28 nests; Frmsa, 121 queens frm 26 nests; Gergia (1985), 3595 queens frm 74 nests; Gergia (1993), 1156 queens frm 45 nests; Texas (1987), 942 queens frm 33 nests; Texas (1995), 144 queens frm 5 nests.

4 324 Euin se l a) c. 1 E z ' I I AIA 'I. C.8- C:. 13, B -,-B 9 c,6 IIZ J,S 111 Prc. Natl. Acad. Sci. USA 93 (1996) c E c E C.2- CD ) a) c.1- C. va ~ O- I #Wm FIG. 1. Number f alleles, prprtin f males diplid, and prprtin f queens remaining permanently unmated in native (Argentina) and intrduced (USA) ppulatins f the plygyne frm f S. invicta. The lss f alleles at prtein markers in the USA (A) is paralleled by a lss f alleles at the sex-determining lcus, which leads t increased hmzygsity at this lcus and elevated prprtins f males that are sterile diplids (B). This increased frequency f male sterility creates strngly female-biased peratinal sex ratis respnsible fr increased prprtins f queens failing t acquire mates in the USA (C; als Table 2). The relatinship amng the three depicted parameters predicted by this scenari hlds nt nly between the native and intrduced ants but als between the tw native ppulatins. Prtein alleles are thse present at 76 electrphretic lci in samples f 3-36 nests frm each ppulatin (25). Prprtins f diplid males are maximum-likelihd estimates based n individual banding phentypes at plymrphic electrphretic lci [Crrientes, 161 males frm 17 nests (9 lci); Frmsa, 326 males frm 26 nests (1 lci); Gergia, 1182 males frm 25 nests (tw lci)] (25). Data fr number f alleles and diplid males are nt available fr the intrduced ppulatin frm Texas. substantial prprtin f queens are unable t btain mates despite being reprductively active (1, 47). In Argentina, where diplid males are rare, significantly lwer prprtins f unmated queens are expected than in the USA, as we have fund. Parallel patterns in abundance f marker alleles, frequency f diplid males, and frequency f unmated queens exist acrss several native and intrduced ppulatins (Fig. 1), prviding strng supprt fr this hypthesized effect f lss f genetic variatin n mating success f plygyne queens. This study dcuments a unique histrical example f significant scial evlutin that has ccurred in assciatin with well defined eclgical and genetic changes. An imprtant cnclusin is that high ppulatin density and intense intraspecific cmpetitin can prmte the evlutin f a scial rganizatin in which wrkers rear unrelated sexuals, implicating such eclgical factrs in the evlutin f ther ant scial systems nt based n clse kinship. Als imprtant is the cnclusin that lss f genetic variatin can mediate evlutin f scial rganizatin by means f the effects f reduced sex-allele diversity n sex ratis and mating systems. We thank P. Gwaty, D. Haig, C. Michener, L. Partridge, and W. Wcisl fr cmments n an earlier versin f the manuscript. Supprt fr this research was prvided by the Gergia Agricultural Experiment Statins f the University f Gergia, the Texas Advanced Technlgy Prgram, and grants frm the Natinal Gegraphic Sciety and the Swiss Natinal Science Fundatin. 1. Rss, K. G. & Keller, L. (1995) Annu. Rev. Ecl. Syst. 26, Burke, A. F. G. & Franks, N. R. (1991) Bil. J. Linn. Sc. 43, Keller, L., ed. (1993) Queen Number and Sciality in Insects (Oxfrd Univ. Press, New Yrk). 4. Herbers, J. M. (1993) in Queen Number and Sciality in Insects, ed. Keller, L. (Oxfrd Univ. Press, New Yrk), pp Burke, A. F. G. & Heinze, J. (1994) Phils. Trans. R. Sc. Lndn B 345, Keller, L. & Reeve, H. K. (1994) Trends Ecl. Evl. 9, Burke, A. F. G. & Franks, N. R. (1995) Scial Evlutin in Ants (Princetn Univ. Press, Princetn, NJ). 8. Strassmann, J. E. & Queller, D. C. (1989) in The Genetics f Scial Evlutin, eds. Breed, M. D. & Page, R. E. (Westview, Bulder, CO), pp d 4q,"I, I c' c ( ^l #,11 9. Reeve, H. K. & Ratnieks, F. L. W. (1993) in Queen Number and Sciality in Insects, ed. Keller, L. (Oxfrd Univ. Press, New Yrk), pp Heinze, J., Lipski, N., Hlldbler, B. & Burke, A. F. G. (1995) Zlgy 98, Herbers, J. M. (1986) Behav. Ecl. Scibil. 19, Mritz, R. F. A. & Hillesheim, E. (1985) Behav. Ecl. Scibil. 17, Heinze, J. & Buschinger, A. (1989) Insectes Sc. 36, Buschinger, A. & Heinze, J. (1992) in Bilgy and Evlutin f Scial Insects, ed. Billen, J. (Leuven Univ. Press, Leuven, Belgium), pp Rss, K. G. (1992) Nature (Lndn) 355, Page, R. E. & Rbinsn, G. E. (1994) Behav. Ecl. Scibil. 35, Prter, S. D., Fwler, H. G. & MacKay, W. P. (1992) J. Ecn. Entml. 85, Tschinkel, W. R. (1988) Ann. Entml. Sc. Am. 81, Nnacs, P. (1993) in Queen Number and Sciality in Insects, ed. Keller, L. (Oxfrd Univ. Press, New Yrk), pp Keller, L. (1995) Trends Ecl. Evl. 1, Hlldbler, B. & Wilsn, E.. (1977) Naturwissenschaften 64, Mrel, L., Vander Meer, R. K. & Lfgren, C. S. (199) Ann. Entml. Sc. Am. 83, Obin, M. S., Mrel, L. & Vander Meer, R. K. (1993) J. Insect Behav. 6, Hlldbler, B. & Michener, C. D. (198) in Evlutin f Scial Behavir: Hyptheses and Empirical Tests, ed. Markl, H. (Verlag Chemie, Weinheim, Germany), pp Rss, K. G., Varg, E. L., Keller, L. & Trager, J. C. (1993) Genetics 135, Rss, K. G. (1989) in The Genetics fscial Evlutin, eds. Breed, M. D. & Page, R. E. (Westview, Bulder, CO), pp Crzier, R. H. (1979) in Scial Insects, ed. Hermann, H. R. (Academic, New Yrk), Vl. 1, pp Lfgren, C. S. (1986) in Fire Ants and Leaf-CuttingAnts: Bilgy and Management, eds. Lfgren, C. S. & Vander Meer, R. K. (Westview, Bulder, CO), pp Markin, G. P., Cllins, H. L. & Dillier, J. H. (1972) Ann. Entml. Sc. Am. 65, Varg, E. L. & Prter, S. D. (1989) Ann. Entml. Sc. Am. 82, Rss, K. G. & Fletcher, D. J. C. (1985) Behav. Ecl. Scibil. 17, Rss, K. G., Varg, E. L. & Keller, L. (1996) Evlutin, in press. 33. Rss, K. G. (1993) Am. Nat. 141,

5 Evlutin: Rss et al. 34. Queller, D. C. (1993) in Queen Number and Sciality in Insects, ed. Keller, L. (Oxfrd Univ. Press, New Yrk), pp Shemaker, D. D., Csta, J. T. & Rss, K. G. (1992) Heredity 69, Queller, D. C. & Gdnight, K. F. (1989) Evlutin 43, Weir, B. S. & Cckerham, C. C. (1984) Evlutin 38, Glancey, B. M. & Lfgren, C. S. (1988) Fla. Entml. 71, Prter, S. D. (1991) J. Entml. Sci. 26, Seger, J. (1993) in Queen Number and Sciality in Insects, ed. Keller, L. (Oxfrd Univ. Press, New Yrk), pp Prc. Natl. Acad. Sci. USA 93 (1996) Varg, E. L. & Fletcher, D. J. C. (1989) Physil. Entml. 14, Pamil, P. (1991) Am. Nat. 138, Nei, M., Maruyama, T. & Chakrabrty, R. (1975) Evlutin 29, Ck, J. M. & Crzier, R. H. (1995) Trends Ecl. Evl. 1, Hung, A. C. F., Vinsn, S. B. & Summerlin, J. W. (1976) Ann. Entml. Sc. Am. 67, Varg, E. L. (1996) J. Evl. Bil., in press. 47. Rss, K. G. & Keller, L. (1995) Am. Nat. 146,