2005. The Journal ofArachnology 33:703-710 MALE RESIDENCY AND MATING PATTERNS IN A SUBSOCIAL SPIDER Barrett A. Klein^ Todd C. Bukowski^, and Leticia Aviles^^: ^Department of Entomology, Forbes Building, Room 410, University of Arizona, Tucson, AZ 85721 USA. E-mail: [email protected];; ^Department of Ecology and Evolutionary Biology, Biological Science West, University of Arizona, Tucson, Arizona 85721 USA; ^Department of Zoology, University of British Columbia, Vancouver, BC, V6T 1Z4 Canada ABSTRACT. Male mating strategies are often deployed with regard to female maturity and receptivity, possibly in response to sperm utilization patterns on the part of the female. We examined the pattern of male residency with females during the mating period of the subsocial spider Anelosimus cf. jucundus (Araneae, Theridiidae). We first examined patterns of male cohabitation with naturally occurring penul- timate instar and adult females in the field. Males were significantly more likely to be found in association with adult females, ratherthan with penultimate instarfemales. Penultimate instarand virgin adultfemales ofknown age were then placed into the field and monitored forresidency by subsequently marked males. Males were, again, significantly more likely to be found in association with adult females, rather than with penultimate-instar females, although we were unable to determine if this pattern was due to differ- ential arrival or to differential retention of males at adult female web sites. Aspects ofA. cf, jucundus natural history, including duration of male residency and frequency of mating in the field, are provided for the first time. We discuss the patterns of male residency in relation to predictions based on sperm utilization patterns by female A, cf.jucundus spiders. Keywords: Anelosimus, female maturity, male cohabitation, residency, sperm utilization Male reproductive success is largely deter- out already mature females when paternity is mined by the number of mates males are able not biased with respect to male mating order to access (Bateman 1948; Jones, et al. 2000). (Eberhard et al. 1993; Schneider 1997) or In spiders, where males tend to move around when paternity is biased towards the last male in search of females, a male’s mating success to mate (Uhl 1998; West & Toft 1999). In the will depend on his ability to locate females of latter case, post-copulatory guarding of fe- the appropriate age and reproductive status males is expected. We examined male mate- and to assess potential paternity success once finding and residency patterns in relation to afemale has been located. Whenfemales mate female maturity inAnelosimus cf.jucundus, a multiple times, sperm priority patterns should subsociai spider species in which, for reasons influence male reproductive strategies (Austad we discuss below, we suspected sperm utili- 1984; Eberhard et al. 1993). When paternity zation patterns to be unbiased with respect to is biased towards the first male to mate, males male mating order. should seek out and guard females who are Anelosimus cf.jucundus, a species to be de- approaching the final molt (Jackson 1980; scribed shortly (I. Agnarsson in press), is rel- Christenson & Goist 1979; Austad 1982; Toft atively common in riparian regions of south- 1989; Watson 1990; Dodson & Beck 1993; ern Arizona. Following a period of maternal Eberhard et al. 1993; Bukowski & Christen- care, A. cf. jucundus siblings remain together son 1997 Bukowski et al. 2001 but see Mas- in their natal nest until close to sexual matu- ; ; umoto 1991). In contrast, males should seek ration, communally capturing and feeding on prey. All clutchmates eventually disperse and Current address: Dept, ofEcology, Evolution and establish individual webs at relatively short Behavior, Section of Integrative Biology, Univ. of distances from the natal nest (5 cm“-5 m, me- Texas at Austin, Austin, TX 78712. dian ^ 46 cm; Powers & Aviles 2003). Dis- 703 704 THE JOURNAL OF ARACHNOLOGY persal typically occurs during the ante-penuL experimental phase, in which we examined & timate and penultimate stadia (Aviles male residency patterns in artificially-estab- Gelsey 1998). Following dispersal, males and lished subadult and adult females’ webs. females mature in their individual webs. AL Ear—ly census of naturally-occurring though both sexes mature during the same sta- webs. On 14 and 24 June 2000, we exam- dium (Aviles & Gelsey 1998), males do so on ined naturally occurring, active, post-natal dis- ~ average nine days earlier than their sibling fe- persal webs {n 293) for the presence ofsub- males (Bukowski & Aviles 2002). After mat- adult and adult males and females. We uration, females typically remain in the webs identified new, active webs containing dis- where they matured while males set out in persed individuals by the relative lack of de- search offemales. While the sex ratio in nests bris, smaller size, and presence of recently prior to dispersal is even, postdispersal sex ra- maintained capture threads. We recorded the & tios are significantly female-biased (Aviles instar (immature versus adult) and sexes ofall Gelsey 1998). animals in each web. — The patterns of sexual receptivity in A. cf. Artificially-established webs. On 8 and jucundus differ for males and females. Males 13 July 2000, we returned to their collection become sexually active within approximately site 27 penultimate-instar females and 54 two days following their final molt, while fe- adult females that had been individually raised males become sexually receptive an average in the laboratory. These spiders had been col- of ten days following their final molt (Bu- lected as penultimate-instar females one to kowski & Aviles 2002). The probability of a two weeks earlier, held individually in 125 ml male courting a female appears to increase as or 30 ml plastic containers, and fed ad libitum the female gets older (Bukowski & Aviles on house flies {Musca domestica), walnut flies 2002). (Rhagoletis juglandis), and fruit flies {Dro- Females readily remate under laboratory sophila melanogaster). conditions and males do not release signifi- The spiders were returned to a large patch cantly different numbers of sperm to virgin of naturally occurring A. cf.jucundus webs to and non-virgin females (Bukowski & Aviles, ensure the presence of naturally occurring unpub. data, using methods ofBukowski et al. males. We placed individual females in open 2001 to quantify sperm). Given that paternity 125 ml containers, which we attached to patterns in spiders largely reflect the numbers branches of juniper trees. We covered each of sperm released (Christenson 1990; Bu- branch with a fine nylon mesh net to encour- kowski & Christenson 1997; Schneider et al. age web-construction at the selected site and 2000; Elgar et al. 2000; Bukowski et al. prevent males or predators from visiting until 2001), we predict that A. cf. jucundus will initiation of the observation period (Fig. 2). have a paternity pattern that is unbiased with When the nets were removed 48 hours later, respect to male mating order. In such a case, females had usually expanded their webs from males should preferentially seek out adult the containers to the surrounding vegetation, rather than subadult females. Here we exper- Web sites (defined here as the area within imentally examine this prediction and present approximately ten centimeters of the female’s the first natural history data on mating fre- web) were censused every 1~2 hours over a quency under field conditions in this spider 24 hour period, every other day. Females re- species. turned to the field on 8 July were censused over a period of seven days, and females re- METHODS turned to the field on 13 July were censused We conducted our studies in Garden Can- over a period of three days. During each cen- yon, a riparian area in the Huachuca Moun- sus, females were recorded as present or ab- tains of southeastern Arizona (31.51°N, sent and the occurrence of mating and male- 110.31°W; 1600-2000 m). Anelosimus cf. ju- male physical contact was monitored. cundus primarily inhabit juniper trees along- We individually marked all males that ap- side permanent streams in this area (Fig. 1). peared at a female’s web site {n = 76) with Our study involved an early phase, in which water-based acrylic paints so that we could we censused naturally occurring webs for pat- determine their duration of residency, occur- terns of male/female cohabitation, and a later. rences of mating and relocation distances. KLEIN ET AL.—MALE RESIDENCY AND MAl'lNG IN ANELOSJMUS 705 — Figure 1. Juniper trees in the Huachuca Mountains of Arizona, where we artificially established webs of female A. cf jucundus spiders in a community of naturally-occurring conspecifics. — . Figure 2. Artificially-established female web site, temporarily surrounded by netting to deterpredation or escape as she expanded her web beyond the cup (labeled “2” and attached to ajuniper branch). — Figure 3. Copulation (male on left, female on right). — Figure 4. Following an extended bout of male-male aggression, this male had tumbled below its combatant, who proceeded to court and mate with the resident female. First, each unmarked male was removed from mating. Matings typically last approximately a female’s web site immediately upon detec- 135 minutes per pedipalp, with an interim of tion or following copulation. Each male was 35 minutes between pedipalps (Bukowski & uniquely marked by gently guiding him into Aviles unpub. data). Matings, in light of their a piece of mesh netting and dabbing acrylic lengthy durations, were unlikely to have been paint onto his opisthosoma. Following mark- undetected during a day of censusing with 1- ing, the male was returned to his place of re- 2 hour inter-census period—s. moval. Most males remained without obvious Statistical analyses. For all maturity behavioral long term effects, although six (of analyses, each female was classified as pen- 82) males were dropped and lost, and two ultimate instar or adult. If seen at a web site males were being consumed by female resi- during consecutive census periods, a spider dents during the census following each male’s was assumed to have remained for the dura- marking. tion between observations. If observed only Matings were defined as pairs in copula during a single census period, a spider was with at least one male pedipalp inserted in the considered to have remained for one hour (for female. Anelosimus cf. juciindus males typi- duration study purposes). Each adult female cally have one insertion with each palp during web site {n ~ 54) was checked an average of 706 THE JOURNAL OF ARACHNOLOGY habitation with immature females was rarely observed in naturally-occurring A. cf. jucun- dus webs during our early census. The webs of adult females were far more likely to con- tain an adult male (42 of 107 females, 39.3%) than were the webs of penultimate-instar fe- males (5 of 93 females, 5.4%; - 35.79, P < 0.0001). Of those webs that contained males, most adult females {n = 39) and all five penultimate-instar females each contained a single male. Two of the adult females each cohabited with two males, and the web ofone Femalematurity adult female contained three male visitors — (Fig. 5). Since we were interested in male res- Figure 5. Male residency in relation to female idency with females and female-female resi- maturity in naturally-occurring webs. Number of dency was rare, four webs that each contained males residing in each web (zero, one, two, orthree males) is plotted against the number of naturally- two adult females and two webs that each occurring female webs with respect to female ma- contained two penultimate-instar females turity. These data are the result of an early field were excluded from the previous analysis. survey. All penultimate-instar males were found as solitary individuals {n = 48). In contrast, at least as many adult males were found with a 30 times and each penultimate-instar female = — web site {n — 21) was checked an average of female {n 46) as without {n 38). In one additional case, three adult males were found 27 times. The proportion of observations that a female’s web site was visited by a male and together in one web without a female. the number of different males involved was Male residency in relation to fem—ale ma- turity: artificially-established webs. Ofthe recorded for each female. Each female then females placed into the field, many (47 of 81, served as a single observational unit for the 58.0%) disappeared from their web sites be- puipose of analyses. All analyses used data fore the study ended. The web sites of adult from the entire study period, except for the females were, again, far more likely to contain duration of male-female encounters and mat- an adult male than were the web sites of pen- ing analyses, which used data collected after ultimate-instar females (27.9 ± 4.3% versus the first day (when male marking began). Ex- 2.4 ± 6.1% of the observations per female; n cseenpDttatawanhdefrlreiovmenosttpheiedd,etrowsnolwyesredtesataaonfacloaynnzcieedm.ranlisng(thporse-e P= =360a.n0d01138; fFeimga.le6s)., Aredsupletctfiveemlayl;est^2al~so3.h4a1d, a greater number ofmale cohabitants perhour placed into the field on 8 and 13 July) were dciofmfbeirneencdesw.hePnertcheenytaegxheibdiattead wnoerseiganirfcisciannet m(0a.l3e7s ±(0.00.306)± t0h.a0n8;di^5d2 p=enu3l.2t8i,maPte-=ins0t.a0r01f9e;- square root transformed prior to analysis. Du- n = 36 and 18 females, respectively; Fig. 7). All but five (of 54) male visitations were by ration data, which were non-normally distrib- new, unmarked males. The five exceptions in- Sutuemd,mawerryesatnatailsytizcesdoufscionngtnionnupoaursavmaertiraibcletsesatrse. cluded one male who returned to the same fe- reported as X ± standard error (SE). Alpha male after a 37 hour absence, two males who wtwaos-tasielteda.t 0D.a0t5afwoerreallantaesltyszeadndwiatlhl ttehsetsJMarPe emaaclhe wtrhaovelterdavetloedatosetchoreneddfieffmearleen,t faenmdaleosn.e IN (version 4.0.3; SAS Institute Inc. 2001) Although males appeared to arrive to adult female web sites at twice the rate than to sub- computer package, or, in the case of rates of male arrival, with Systat (Systat Software, adult female web sites (to 33 out of 118 avail- able adult females, or 28%, versus 6 out of43 Inc.). available subadult females, or 14%), this dif- RESULTS ference was not statistically significant with Male residency in relation to f—emale ma- our sample size (Mantel-Haenzel — 2.37, P turity: naturally-occurring webs. Male co- — 0.12, for the comparison ofnumbers ofnew KLEIN ET AL.—MALE RESIDENCY AND MATING IN ANELOSIMUS 707 period either before or after the start of a cen- sus day, as appropriate to the case, gives us a maximum possible residence time. Given these considerations, the median male resi- dence time we observed was bracketed be- tween a minimum of five and a maximum of 11 hours. Based on our minimun male residence es- timates (period actually observed), males spent more time per visit with adult females with whom they mated than with any other females (median = 18.5 hours, versus 5 hours for adults not observed to mate, and 3 hours 6 = for penultimate-instar females; n 10, 37, and 6 visits, respectively; ^ 12.1, 2 df, P = 0.002; Fig. 8). Post-copulatory periods = ranged from three to 33 hours (ji 10), with seven cases lasting seven hours or less. If three cases in which the period had not yet concluded when observations ended are in- cluded, a floor for the median post-copulatory period is estimated at six hours. In a few instances, more than one male could be observed at a female’s web site (Ta- ble 1). Cohabiting males engaged in agonistic penultimate-instar adult interactions, including foreleg tapping and Female maturity locking ofchelicerae and legs {n = three pairs of males in the presence of three different fe- — Figure 6. Male residency in relation to female males). One battle sent a male tumbling and maturity in artificially=established webs: percent of appearing temporarily dead (Fig. 4), while his field observations in which at least one male was combatant copulated with the resident female present (± SE). (Fig. 3). Male-induced coitus interruptus, re- — Figure 7. Male residency in relation to female sulting in no resumption of copulation, was maturity in artificially-establishedwebs: meannum- also observed in one case where two males ber of male residents per hour (± SE). were simultaneously pr—esent with a female. Mating frequency. Nineteen of the 36 adult females (52.7%) were observed mating males arriving per total females available at with at least one male over the study period. each of 18 different census periods). Males Eleven of these females (57.9%) were each also tended to stay longer at adult female web observed mating with a single male. The re- sites (see next section), but this effect ap- maining eight of these females (42.1%) were peared to reflect whether copulation occurred each observed mating with two males. The av- or not, rather than female age pe—r se. erage number of males a female copulated Duration of male residency. Because of with over the active observation period (an the periods between census days during which average of38.8 observation-hours per female) the nests were not monitored, we can only was therefore 0.75, which corresponds to 0.47 provide estimates of the minimum and maxi- males per female per day, ifwe assume a sim- mum possible male residence times. In cases ilar mating rate during non-observed periods, where male residence periods had either al- or, more conservatively, 0.22 males per fe- ready been initiated when observations had male per day if we assume that all matings started or had not yet concluded when obser- were observed during the recorded period of vations ended, we have taken the period ac- each female. tually observed as the minimum male resi- Some marked males (10 of76, 13.2%) were dence time. This period plus the unobserved observed at the web sites of two {n == 9) or 708 THE JOURNAL OF ARACHNOLOGY The mechanism responsible for these di- vergent residency patterns remains unclear. Males could preferentially arrive at the webs ofadult females or arrive equally at both adult andjuvenile female webs, but be preferential- ly retained by adult females. Our data show a nonsignificant trend towards differential arriv- al at adult webs and significant retention when copulation occurs. Although a greater sample size will be needed to definitely address this issue, the trend towards preferential arrival at adult female web sites suggests that females Female maturity may be producing a distance-acting signal or — Figure 8. Estimates of the range (outermost cue guiding males to their webs. Distance-act- lines in each graph), 25th and 75th quantiles (edge ing pheromones released by females have of boxes), and medians (lines inside the boxes) of been demonstrated to attract males in both the minimum duration of male residency (based on Pardosa milvina wolf spiders (Searcy et al. period actually observed) with respect to female 1999) and in Agelenopsis aperta desert spi- avigseitanwdithmataidunlgtstfaetmuas.leMsalweisthspwehntommorteheytimmeatpeedr ders (Papke et al. 2001). If males arrive at the (median = 18.5 hours; n — 10 visits), than with webs ofadult andjuveniles equally, then some adults with whom they were not observed to mate process associated with interaction with the (median = 5 hours; n = 37 visits), or with penul- female must influence duration of male resi- timate-instar females (median = 3 hours; n = 6 dency. visits). Nearly halfof the females observed mating copulated with more than one male, suggest- ing that multiple mating by females is com- three (n = females. However, because mon in this species. Matings were not pre- 1) males could not be tracked as reliably as fe- dictably followed by continued male residence males (and several ofthe females were absent at the females’ web sites (in 60% ofthe cases or dead at the time of male visitation), only the male departed in six hours or less), so pro- one of these males was actually observed to longed post-copulatory guarding appears to be mate more than once. Overall, ten of 48 absent in these spiders. Male residency sub- (20.8%) visits by marked males to adult fe- sequent to copulation could simply involve male web sites were observed to result in cop- time spent by males inducting sperm into their ulation (Fig. 8). pedipalps for subsequent matings or their fac- Measurements of distances between female ultative use of females’ webs for food and web sites that males successively visited shelter, rather than a means of exploiting the showed an average travel distance of 2.0 ± resident females’ reproductive biology. Given 0.4 m (range: < 1-4 m, /? = 11 males) over that males differentially reside with adult fe- an average of 21.5 ± 4.5 hours (range: 1.5- males, females multiply mate, and males do 1 51.0 hours, n = 11). Three of the males each not exhibit post-copulatory guarding, male traveled four meters in an average 32.3 ± 1 1.4 mating order may not be an important deter- hours. minant of paternity in A. cf. jucundus. At times, more than one male entered the DISCUSSION same adult female site. This could result in Anelosimus cf. jucundiis males were much multiple matings by females with different more likely to be found on the webs of adult resident males, or male-male aggressive inter- females than on those ofpenultimate-instarfe- actions, as described earlier. Because of the males, both in a survey offree-ranging spiders scarcity of extended multiple male residen- and when females of known age and repro- cies, aggressive interactions may drive some ductive history were placed into the field. males to search for different, unattended fe- Adult females also had a greater number of males. male residents per hour than did penultimate- Many adult females were observed mating instar females. during the relatively short study period. Of KLEIN ET AL.—MALE RESIDENCY AND MATING IN ANELOSIMUS 709 — Table 1, Number and proportion of census ob- males first become sexually receptive (Austad servations with zero, one, two, or three males co- 1982; Christenson & Cohn 1988; Watson habiting with adult females in their artificially-es- 1990; Bukowski & Christenson 1997; Bu- tablished webs. Multiple male residency at a given kowski et al. 2001). When the fertilization female’s web was uncommon. Calculations are pattern is biased towards last males, males pooled across females. should preferentially seek out and guard adult # males # observations % observations females (Uhl 1998; West & Toft 1999). When the sperm of two or more males mix equally 0 355 57.1 within the female, males should seek out adult 1 211 > 33.9 females regardless offemale age (Eberhard et 2 53 8.5 ah 1993). Our data suggest that A. cf. jucun= 3 3 0.5 Total 622 100 dus exhibit either last male precedence or sperm mixing. Other data on sperm release patterns in this species, along with no evi- dence of mate guarding, provide support for those females, about half were observed to sperm mixing, since the first and second males mate with at least two males. Given that fe- to mate with a female were found to transfer males become sexually receptive an average equal numbers of sperm (Bukowski & Aviles, of ten days after the final molt and cease sex- unpub. data). Several studies have shown that & ual receptivity after oviposition (Bukowski when two males mating with a female transfer Aviles 2002), a total of 20 days spans the av- equal numbers of sperm, the two males typi- erage period of sexual receptivity for females cally sire equal numbers of offspring (Bu- in this species. Assuming that a female’s pro- kowski & Christenson 1997; Schneider et al. pensity to mate does not alter dramatically 2000; Elgar et al. 2000). If a male were to over the course of these 20 days and that all reside with a penultimate instar female until matings are considered to have been observed she became sexually receptive, he would like- throughout the period of censusieg, a female ly visit and mate with fewer females, siring may be calculated to mate with an average of fewer offspring than a male that exclusively 4.4 males during her lifetime (0.22 males per visits adult females. female per day X 20 days). This figure may Understanding the role of female maturity, underestimate the numberofpotential matings mating receptivity and subsequent sperm uti- if we assume that additional matings occurred lization is contingent upon learning more during the unobserved periods. Alternatively, this figure may overestimate the number of about the natural history of A. cf. jucundus spiders. Precisely determining male arrival potential matings if female receptivity or the rates and residency durations relative to fe- frequency of male visitation to females’ web male maturity could serve as the next step in sites diminished as the 20-day female active period progressed, although we have no evi- understanding the mechanisms driving their sexual interactions. dence negating or supporting either a dimin- ished female receptivity or reduced male vis- ACKNOWLEDGMENTS itation over time. All matings within census days were likely to have been observed and We thank Sheridan Stone and the Wildlife recorded because times between censuses Management Office for access to the spiders were shorter than the average copulation du- ofGarden Canyon, and Natalie Doerr and Ter- ration. Matings occurring during the much ry A. Bukowski for helping with field data longer period between census days, on the collection. Asher Cutter, Greta Binford, Jeff other hand, could have been missed. Smith, Eileen Hebets, Kim Powers, two anon- The male residency and female mating pat- ymous reviewers and the editors offered fruit- terns exhibited by A. cf. jucundus have im- ful comments on drafts of this paper. Voucher portant implications for the pattern of sperm specimens ofboth sexes reside within the Na- utilization at fertilization. When the fertiliza- tional Museum ofNatural History Spider Col- tion pattern is biased towards first males, lection, Washington D.C., USA. This research males differentially cohabit with juvenile fe- was supported by NSF grant DEB-9815938 to males approaching the final molt when the fe- Leticia Aviles. 710 THE JOURNAL OF ARACHNOLOGY LITERATURE CITED cannibalistic spider Argiope keyserlingi. Pro- Agnarsson, 1. In press. Arevision oftheNewWorld Bc-eBeidoilnoggsicoaflthSeciReonycaels S2o6c7i:e2t4y3o9f-2L4o4n3d.on, Series eximiiis lineage of Anelosimus (Araneae, Theri- Jackson, R.R. 1980. The mating strategy ofPhidip- diidae) and a phylogentic analysis using world- wide exemplars. Zoological Journal of the Lin- pus johnsoni (Araneae, Salticidae): II, Sperm competition and the function ofcopulation. Jour- nean Society. nal of Arachnology 8:217-240. Austad, S.N. 1982. First male sperm priority in the bowl and doily spider Frontinella pyramitela Jones, A.G., G. Rosenqvist, A. Berglund, S.J. Ar- nold, and J.C. Avise. 2000. The Bateman gradi- (Walckenaer). Evolution 36:777-785. Austad, S.N. 1984. A classification of alternative ent and the cause of sexual selection in a sex- role reversed pipefish. Proceedings of the Royal reproductive behaviors and methods for field- testing ESS models. American Zoologist 24: Society of London Series B-Biological Sciences 267:677-680. 309-319. Aviles, L. & G. Gelsey. 1998. Natal dispersal and Masumoto, T. 1991. Males’ visits to females’ webs demography of a subsocial Anelosimus species and female mating receptivity in the spider,Age- lena limbata (Araneae, Agelenidae). Journal of and its implications for the evolution ofsociality Ethology 9:1-7. i2n137sp-i2d1e4rs7.. Canadian Journal of Zoology 76: Papke, M.D., S.E. Riechert, & S. Schulz. 2001. An airborne female pheromone associated with male Bateman, A.J. 1948. Intra-sexual selection in Dro- attraction and courtship in a desert spider. Ani- sophila. Heredity 2:349-368. Bukowski, T.C. & L. Aviles. 2002. Asynchronous mal Behaviour 61:877-886. maturation ofthe sexes may limit close inbreed- Powers, K.S. and L. Aviles. 2003. Natal dispersal ing in a subsocial spider. Canadian Journal of patterns of a subsocial spider Anelosimus cf.ju- cundus (Theridiidae). Ethology 109:1-13. Zoology 80:193-198. Bukowski, T.C. & T.E. Christenson. 1997. Deter- Schneider, J.M. 1997. Timing ofmaturation and the mating system of the spider, Stegodyphus linea- omirnbawnetasvionfgspspeirdmer.relAenaismeaalndBeshtaorvaigoeurin5a3:s3p8i1n-y tus (Eresidae): how important is body size? Bi- ological Journal of the Linnean Society 60:517- 395. Bukowski, T.C., C.D. Linn & T.E. Christenson. 525. Schneider, J.M., M.E. Herberstein, EC. DeCrespig- 2001. Copulation and sperm release in Gasterci- ny, S. Ramamurthy, & M. Elgar. 2000. Sperm cantha cancriformis (Araneae: Araneidae): dif- competition and small size advantage for males ferential male behaviour based on female mating of the golden orb-web spider Nephila edulis. history. Animal Behaviour 62:887-895. Journal of Evolutionary Biology 13:939-946. Christenson, T.E. 1990. Natural selection and repro- Searcy, L.E., A.L. Rypstra, & M.H. Persons. 1999. duction: a study of the golden orb- weaving spi- der, Nephila clavipes. Pp. 149-74. In Contem- Airborne chemical communication in the wolf spider Pardosa milvina. Journal of Chemical porary Issues in ComparativePsychology. Edited by D.A. Dewsbury. Sinauer, Sunderland, MA. Ecology 25:2527-2533. Christenson, T.E. & J. Cohn. 1988. Male advantage Toft, S. 1989. Mate guarding in two Linyphia spe- cies (Araneae: Linyphiidae). Bulletin ofthe Brit- for egg fertilization in the golden orb-weaving spider {Nephila clavipes). Journal of Compara- ish Arachnological Society 8:33-37. tive Psychology 102:312-18. Uhl, G. 1998. Mating behaviour in the cellar spider, Christenson, T.E. & K.C. Goist. 1979. Costs and Pholcus phalangioides, indicates sperm mixing. Animal Behaviour 56:1 155-1159. benefits of male-male competition in the orb- weaving spider Nephila clavipes. Behavioral Watson, PJ. 1990. Female-enhanced male compe- Ecology and Sociobiology 5:87-92. tition determines the first mate and principle sire Dodson, G.N. & M.W. Beck. 1993. Pre-copulatory in the spider Linyphia litigiosa (Linyphiidae). guarding of penultimate females by male crab Behavioral Ecology and Sociobiology 26:77-90. spiders, Misumenoides formosipes. Animal Be- West, H.R & S. Toft. 1999. Last-male sperm pri- haviour 46:951-959. ority and the mating system of the haplogyne Eberhard, W.G., S. Guzman-Gomez & K. Catley. spider Tetragnatha extensa (Araneae: Tetrag- 1993. Correlation between spermathecal mor- nathidae). Journal of Insect Behavior 12:433- phology and mating systems in spiders. Biolog- 450. ical Journal of the Linnean Society 50:197-209. Elgar, M.A., J.M. Schneider & M.E. Herberstein. Manuscript received 7 October 2003, revised 30 2000. Female control ofpaternity in the sexually June 2004.