2001. The Journal of Arachnology 29:64±71

SEXUAL SIZE DIMORPHISM AND JUVENILE GROWTH RATE IN TRIANGULARIS (, ARANEAE)

Gary H.P. LaÊng: Department of Zoology, GoÈteborg University, Box 463, SEÐ405 30 GoÈteborg,

ABSTRACT. On three separate occasions during the growth season four populations of the sheet web were sampled, twice as immatures and once as adults. For the immature specimens, ®ve linear size characteristics (length and width of the cephalothorax, length of tibia of the ®rst leg, and length and height of the abdomen) were measured in the laboratory and compared with fresh weight. The best predictor of weight was abdomen length, closely followed by cephalothorax width. Cephalothorax width was used to compare the size of immatures with the adult size at time of maturity because the abdomen shrinks in the non-foraging adult males. Mean cephalothorax width was larger for males than for females in both immature and adult specimens. The difference increased from the earliest immature population samples to the adult samples. The relationship between cephalothorax width and abdomen length was linear and equal between the sexes over all immature samples. This means that there was no difference in the allocation of resources to body parts important to female fecundity (the abdomen) vs. body parts important to male ®ghting ability (the cephalothorax) between males and females. Selection for large male size thus seems to be greater than selection for large female size in this web-building spider, resulting in an overall faster growth rate in males. Males grow Ͼ10% larger than females despite the distinct protandry in this species. Keywords: Sexual dimorphism, Linyphia triangularis, growth rate, spider

Web-building have normally a sex- mortality risk as females during juvenile stag- ual size dimorphism (SSD) with the female as es, but after maturation they leave their pro- the larger sex (Vollrath & Parker 1992; Head tective web and move around in the vegetation 1995; Foelix 1996; Coddington et al. 1997). searching for females. This behavior is prob- This has been attributed to (1) the develop- ably associated with a high risk of mortality ment of dwarf males, favored by a high se- (Vollrath 1980), as moving around makes the lective adult mortality on males that relaxes male more vulnerable to predators and reduc- sexual selection for large size (Vollrath & es energy reserves. Small size could give an Parker 1992), and to (2) the development of advantage in avoiding visual predators (Gun- giant females (Coddington et al. 1997) in re- narsson 1998); and, because metabolism is sponse to fecundity selection for large females lower for small spiders, more time can be used as suggested already by Darwin (Andersson for reproductive activities (Reiss 1989; 1994). Blanckenhorn et al. 1995). Both these views support the idea that nat- Linyphiid spiders exhibit all of the charac- ural selection favors larger size in female spi- ters of other web-builders, except in regard to ders because fecundity is positively correlated size dimorphism, where both sexes are of al- with female size in spiders, both intraspeci®- most equal size (Vollrath & Parker 1992; cally (Rubenstein 1987; Suter 1990; Beck & Head 1995; Prenter et al. 1997). This probably Connor 1992; Higgins 1992) and interspecif- depends on a more intense sexual selection for ically (Eberhard 1979; Marshall & Gittleman large male size through male-male competi- 1994; Simpson 1995). Females, which expe- tion over mating opportunities. Protandric rience a constant and relatively low mortality mating systems seem to be the rule in liny- risk throughout their lives, should grow large phiid spiders (Toft 1989; Gunnarsson & to maximize fecundity and reproductive suc- Johnsson 1990; Watson 1990), probably be- cess (Marshall & Gittleman 1994; Head cause the ®rst male has precedence in fertil- 1995). ization success (Austad 1982; Watson 1991). Males of web-building spiders run a similar Males that wait too long for their ®nal molt

64 LAÊ NGÐSEXUAL SIZE DIMORPHISM 65 risk losing valuable mating opportunities be- and SE site (VaÈrnamo) were inland sites. The cause females reaching sexual maturity mate SW site was a sparse pine forest, Pinus syl- with the male present in their web. This means vestris, with mostly Vaccinium myrtillus as that the male growth period is shorter and, ground cover. The NW site was dominated by assuming an equal growth rate, adult males heather, Calluna vulgaris, with some junipers, are expected to be smaller than adult females Juniperus communis, as the only higher veg- at maturity. etation. The SE site is a pine forest, Pinus One linyphiid spider that lives on spruce sylvestris, with mixed patches of Vaccinium branches and has a bi-annual life cycle, Pi- uliginosum and V. myrtillus as ground cover. tyohyphantes phrygianus (C.L. Koch 1836) The NW site is on the edge of a mire, with has been shown to have a male-biased SSD scattered small (Ͻ 5 m) birches, Betula sp., (Gunnarsson 1988). Gunnarsson suggests that and pines, Pinus sylvestris, as the only higher this depends on behavioral differences be- vegetation, and a ground cover of Calluna tween sexes during winter when males, de- vulgaris, Erica tetralix and Empetrum nigrum. spite high risk of predation, forage more ac- The shortest distance between two sites was tively than females (Gunnarsson 1998). about 100 km (SW-SE) and the longest dis- The species used in this study, Linyphia tri- tance was about 200 km (SW-NE). angularis (Clerck 1757), is univoltine and in- The spiders were collected on three separate habits low shrubs and bushes in many types occasions from each site between 2 June±5 of habitats in the southern part of Sweden. September 1996. The ®rst sampling was Linyphia triangularis overwinters as eggs, planned to occur about four weeks before the hatches in late March/early April, and grows expected maturation but, due to bad weather to adult stage in about four months. Males conditions, the sampling was delayed at two molt to maturity about a week before females locations. The second sampling was done (Toft 1989; Stumpf 1990), and in southwest- close to the expected maturation date, and the ern Sweden the ®rst adults are seen at the end last sampling was made when the reproduc- of July. The adult males guard subadult fe- tive period was over. To ensure that all spiders males, ®ghting with other males over access had put up new webs, sampling was done only to the female, and mate with the female im- after at least 24 hours of relatively still, clear mediately after maturation (Rovner 1968; weather. On each occasion, at least 30 speci- Nielsen & Toft 1990), leaving mating plugs in mens were collected by ``hand-to-jar'' sam- the epigynum that impedes further copulation pling at random co-ordinates in a selected area and/or reduce female receptivity (Stumpf 2 1990; Stumpf & Linsenmair 1995). These of about 10,000 m with relative homoge- studies say nothing about a size dimorphism neous vegetation. The spiders were brought to in either direction, but emphasize the ®erce the laboratory and placed in darkness at 4 ЊC male ®ghts that occur on the web of virgin overnight before measuring. females, presumably creating a strong sexual On all non-adults I measured the length of selection for large male size. This study was the tibia on the ®rst leg, the length (from clyp- initiated to test whether sexual selection for eus to pedicel) and maximum width of the large male size could overcome fecundity se- cephalothorax, and the length and height of lection for large female size in a linyphiid spi- the abdomen to the nearest 0.02 mm with an der with a continuous growth period. I used ocular eye piece on a Wild stereo-microscope. four separate populations to control for local The fresh weight of the spiders was measured variation in size or size dimorphism between using a Sartorius electronic balance to the populations. nearest 0.1 mg. To control the sex of each individual, the spiders were placed in 250 ml METHODS plastic jars and reared to maturity. Spiders that Four populations of Linyphia triangularis could not be unambiguously sexed before they in southwestern Sweden were sampled to con- died were excluded. In adults, most males had trol for possible geographic or habitat varia- shrunken and severely distorted abdomens so tion. The NW site (GoÈteborg) and SW site only the cephalothoracic measurements were (Halmstad) were within 5 km from the Swed- noted in these specimens. Statistical analyses ish west coast, whereas the NE site (SkoÈvde) were performed with StatView v.5.0 (SAS In- 66 THE JOURNAL OF ARACHNOLOGY

Table 1.ÐRegression statistics of log10 weight against the different log10 linear measurements of body size in juvenile stages. Data from all sites are pooled.

Measurement (x) Least squares regression on weight (y) r2 Cephalothorax length y ϭ 0.095 ϩ 2.833 x 0.861 Cephalothorax width y ϭ 0.464 ϩ 3.307 x 0.928 Tibia 1 length y ϭ 0.080 ϩ 1.963 x 0.714 Abdominal length y ϭϪ0.212 ϩ 2.689 x 0.958 Abdominal height y ϭ 0.349 ϩ 1.939 x 0.738 stitute 1998) and SuperAnova v.1.11 for Mac- adult male specimens were more-or-less intosh (Abacus Concepts 1989). starved and had severely shrunken abdomens. The correlation was high for all the measures RESULTS in juveniles (Table 1), but the best predictor I considered weight as the measure best re- of weight in the juvenile specimens was ab- ¯ecting the general size of the . How- domen length, closely followed by cephalo- ever, the contrasting adult lifestyles of males thorax width. I chose cephalothorax width as and females make the use of adult weight dif- indicator of general size instead of weight or ®cult to accurately assess the actual size at other size measurements including the abdo- maturity. In order to establish the linear mea- men because this study focuses on the com- surement that provided the best estimate of parison between SSD in juveniles and adults. general size in L. triangularis, I made a log- The deterioration of the males' abdomen dur- arithmic regression of ®ve linear measure- ing their non-feeding, mate-searching adult ments against the weight of immature spiders. life inhibits the use of weight or abdomen The adults were not used for this test as the length for this purpose. In order to examine if there was any dif- ference between males and females in allo- cation of resources to different body parts, i.e., if the relative size of the cephalothorax vs. the abdomen changed during growth, I performed a one factorial ANCOVA on abdomen length with sex as factor and cephalothorax width as covariate. Males and females had very similar expected abdomen length for any given ceph- alothorax width and the only signi®cant factor in the analysis is the covariate, cephalothorax width (F ϭ 934, 3; P Ͻ 0.0001). Neither the interaction term, sex∗cephalothorax width (F ϭ 0, 029; P ϭ 0, 86), nor the singular factor, sex (F ϭ 0, 007; P ϭ 0, 93), had any effect on abdomen length. The regression lines de- scribing the relationship between maximum cephalothorax width and abdomen length are virtually the same for males and females (Fig. 1). This means that there is no difference be- tween males and females in the relative growth rate of different body parts in juvenile Figure 1.ÐThe relationship between cephalotho- stages. rax width and abdomen length in immature Liny- Males were, on average, larger than females phia triangularis. ⅙ ϭ females, ᭝ ϭ males. Re- gressions are: female abdomen length ϭϪ0.37 ϩ in maximum cephalothorax width in all sam- 2.18 ϫ cephalothorax width, r2 ϭ 0.84; male ab- ples except in the second sample at site SE. domen length ϭϪ0.36 ϩ 2.16 ϫ cephalothorax In the ®rst sampling at the four sites, males width, r2 ϭ 0.80. were on average 9±14% larger than females. LAÊ NGÐSEXUAL SIZE DIMORPHISM 67

Figure 2.ÐDevelopment of mean cephalothorax width Ϯ S.E.M. in four populations of Linyphia tri- angularis in SW Sweden. The ®rst sampling yielded only immatures. In the second sampling adults found in the populations were excluded and in the third sampling only adults were found. NW, NE, SW, SE ϭ population labels after relative geographic position. Day 1 ϭ July 1. ⅜ ϭ females, ᭝ ϭ males.

Mean male cephalothorax width ranged from site, while female cephalothorax width ranged 0.91 mm in the NE site to 1.17 mm in the SE from 1.16 mm in the NE site to 1.26 mm in site, while female mean cephalothorax width the SE site (Fig. 2). The abdomen length var- ranged from 0.81 mm in the NW site to 1.07 ied in a similar way, the males being 7±13% mm in the SE site (Fig. 2). Male abdomen larger than females at three sites, but 28% length was on average 12±17% larger than fe- smaller at the SE site. male abdomen length in the four sites. The third sampling was made about three In the second sampling, made just before weeks after the normal time for maturation. the expected molt to maturity, the mean ceph- The average cephalothorax width of adults alothorax width of males was on average 5± was larger for males than for females in all 9% larger than that of females in three sites. areas (8±22%). The absolute difference of the In the fourth site (SE), few subadult males average maximum cephalothorax between were found (n ϭ 2); and they were much males and females width was larger in all pop- smaller than the subadult females. Male ceph- ulations compared to the initial collection. alothorax width at this occasion ranged from Male adult mean cephalothorax width ranged 1.07 mm in the SE site to 1.30 mm in the SW from 1.54 mm in population SE to 1.81 mm 68 THE JOURNAL OF ARACHNOLOGY

Table 2.ÐANCOVA on cephalothorax width with site and sex as factors and date of sampling as a covariate.

Source df Sum of squares Mean square F-value P-value Site 3 1.142 0.381 20.586 Յ0.0001 Sex 1 0.94 0.094 5.060 0.025 Date of sampling 1 12.262 12.262 663 Յ0.0001 Site ϫ Sex 3 0.009 0.003 0.154 0.93 Site ϫ Date 3 0.504 0.168 9.085 Յ0.0001 Sex ϫ Date 1 0.053 0.053 2.865 0.091 Site ϫ Sex ϫ Date 3 0.042 0.014 0.751 0.52 Residual 323 5.971 0.018

in the NW population, while female adult males have a higher juvenile growth rate than mean cephalothorax width ranged from 1.43 females, as a solution to two apparently op- mm in the SE population to 1.52 mm in the posing selection pressures on male size in this NE population. All population means of fe- species. First, males need to mature earlier male cephalothorax width were lower than than females. Female L. triangularis mate those of male cephalothorax width (Fig. 2). with the male present immediately after molt- These results were tested for differences in ing to maturity (Toft 1989) and ®rst male to cephalothorax width with a full interaction, mate with the female sires most of the off- two factor ANCOVA with collection day as a spring (Stumpf 1995). This gives males a covariate, to control for the time differences shorter period of time for growth. Second, be- between dates of collection, and with sex and cause of intense male ®ghts for access to un- site as singular factors (Table 2). There was mated females, males also need to become as one signi®cant interaction term in the AN- large as possible to be able to defend or take COVA between site and date of collection. over subadult females from other males. The date of collection, as well as the site of The decrease of the SSD in the second sam- origin, also was highly signi®cant as singular pling (compared to the ®rst) in some of the factors; but as the interaction term between populations is probably due to the low num- these two factors was also signi®cant, these bers of subadult males found in the sample, factors cannot be considered independently and that these males were probably smaller (Sokal & Rohlf 1995). However, the singular than the rest of the cohort. This could depend factor sex was not involved in any signi®cant on (1) reproductive success that is increasing interaction terms and had a signi®cant effect with size but decreasing as time of maturity on cephalothorax width (P ϭ 0.0252). That is delayed and (2) the decision to molt to ma- shows that males on average have a larger turity that will be a compromise between op- cephalothorax width than females across all timal size and age. The result of this would sites and times examined. be that larger males mature earlier than small DISCUSSION males (Stearns 1992) and that a late sample The results in this study show that sexual would then necessarily underestimate the av- size dimorphism in L. triangularis is generally erage size of subadults. male-biased in all stages and populations sur- Sex-related variation in growth rate was veyed (Fig. 2). Males are larger than females previously described by Wiklund et al. (1991) both as juveniles (®rst sampling, mostly pre- for the butter¯y Pieris napi L. 1758 in south- subadult stages) and as adults despite their ern Sweden. This butter¯y is partially bivol- earlier maturation. The size difference, mea- tine with discreet generations and must dia- sured as cephalothorax width, increases from pause as pupae during winter. Males are the juvenile stages sampled in early July to generally larger than females; but the differ- the adult stages sampled in August. This sug- ence varies depending on the season, as does gests that males reach a greater weight and the mechanism by which the size dimorphism overall size at maturity than females do. If so, is achieved. Overwintering pupae produce a LAÊ NGÐSEXUAL SIZE DIMORPHISM 69

®rst generation of adults in the spring. The pect a signi®cant interaction term in the AN- eggs laid by these adults can either develop COVA on abdomen length and the slope of directly, producing a second generation of the correlation between cephalothorax width adults the same year, or they can develop and abdomen length (Fig. 2) would be steeper slower into diapausing pupae. Wiklund et al. for females than for males. Instead, the slopes (1991) shows that the growth rate of the time- between size measurements are virtually iden- stressed larvae that develop directly to second tical for males and females. The ANCOVA generation adults are higher than for those that reveals no difference between males and fe- develop to diapausing pupae. In the directly males in growth of different body parts. This developing larvae, males achieve a greater suggests that size dimorphism in L. triangu- size than females through a higher growth laris depends on an overall higher growth rate rate. In the diapausing individuals males grow in males, as opposed to a sex-related alloca- larger than females because they have a longer tion of resources between body parts. development time. The authors suggest that In Pityohyphantes phrygianus, Gunnarsson growth rate must be considered as a life his- (1988) showed that subadult males were sig- tory trait in its own right amenable to evolu- ni®cantly larger than subadult females in both tionary change and not as a parameter con- cephalothorax width and abdomen height both trolled passively by temperature and food before and after winter. A re-analysis of the availability in the environment. data shows that the relationship between the The male-biased SSD described in this two measurements varies considerably be- study is unusual among web-building spiders, tween years and time of season, but the chang- and has not been described in other members es are similar for males and females within of the family. Earlier works of SSD in spiders each sample. This suggests that overall growth have often used only the total body length as rate is higher for males in this species as well measure of general size, and this may in part and that allocation of resources does not differ explain the difference from the pattern found between sexes. Pityohyphantes phrygianus is here. Male-biased SSD in cephalothorax size biannual and overwinters twice before matur- has previously been described in a study on ing in the second spring. In spite of the fe- the orbweaver Metellina segmentata (Clerck male-biased primary sex ratio (Gunnarsson & 1757) by Prenter et al. (1995). They suggested Andersson 1992), it seems to be strong selec- that the size dimorphism found in the adults tion on large male size. Males grow larger of this species depends on sex-speci®c allo- probably because they forage more actively cation of resources. This hypothesis states during winter than females. But foraging ac- that, because eggs are more costly to produce tively increases risk of predation, and the re- than sperm, the females make a larger invest- sult is a more female-biased sex ratio after ment in reproductive tissues (i.e., the abdo- winter (Gunnarsson 1998). men). Males can therefore transfer more of For L. triangularis, which is a species with their energy to other body parts that assist a continuous growth period from egg to adult, them in ®nding and guarding females (such as such an explanation is not possible. Neverthe- longer walking legs and a larger, more pow- less, an increased growth rate during juvenile erful cephalothorax). However, the large re- stages would also bene®t the female fecundi- productive costs for females are the eggs and ty, unless there is some cost invoked by a high the yolk associated with the eggs. Even growth rate. Costs could be developmental, though oocytes are to some extent present at for example, if a too-large mass increment be- sexual maturity, the largest part of the egg, the tween molts would make molting dif®cult, re- yolk, is not added to the oocytes until after sulting in loss of one or more limbsÐor death. copulation (Seitz 1971). Therefore, the costs Costs could also arise from a risk-prone feed- of reproductive structures that develop before ing behavior, such as responding indiscrimi- maturity are likely to be similar in males and nately to any vibration in the web as if it were females and should not have an effect on prey. This would increase the chance that prey overall size dimorphism at maturity. falling onto the web is caught, but it also in- Also, if there was a difference in resource creases the risk to become prey to a larger allocation prior to maturity between females predator. All of these costs should result in a and males of L. triangularis, one would ex- differential mortality of males, affecting the 70 THE JOURNAL OF ARACHNOLOGY sex ratio of the adult population, and possibly Gustavo Hormiga, Jonathan Coddington, Pe- also increase the variance of male sizes com- tra Sierwald and Jim Berry for valuable com- pared to female size variance. From the data ments. Birgit Lundell helped with the food for in this study, it is not possible to conclude if reared spiders. This study was supported by there are sexual differences in mortality re- grants from W.&M. Lundgrens Foundation, sulting in a skewed sex ratio in the adult pop- the Royal and Hvitfeldtska Foundation, ulation. Toft (1989) noted that there is an even A.&G. Vidfeldts Foundation, Collianders sex ratio in L. triangularis just before maturity Foundation, and various funds from the Swed- and concluded that the operational adult sex ish Royal Academy of Sciences (to the au- ratio therefore is male-biased. This suggests thor). The study also bene®ted from grants that such costs are negligible. from the Swedish Natural Science Research Another explanation for this male-biased Council (to Bengt Gunnarsson). size dimorphism in both juvenile and adult populations could be that males get a headstart LITERATURE CITED in life. This could occur either because male Abacus Concepts. 1989. Super Anova. Abacus eggs hatch earlier than female eggs or because Concepts, Inc., Berkeley, California. male eggs are larger than female eggs. Some Andersson, M. 1994. Sexual Selection. Princeton egg sacs produced by females in the labora- Univ. Press. Princeton, New Jersey. 599 pp. tory were allowed to hatch, but there was no Austad, S.N. 1982. First male sperm priority in the apparent differentiation in hatching date with- bowl and doily spider, Frontinella pyramitela in egg sacs. Other egg sacs were opened and (Walckenaer). Evolution 36(4):777±785. the diameter of the eggs within each egg sac Beck, M.W. & E.F. Connor. 1992. Factors affecting was measured. There was no deviation from the reproductive success of the crab spider Mis- an expected normal distribution of egg size umenoides formosipes: The covariance between juvenile and adult traits. Oecologia (Berlin) 92: within these egg sacs (unpubl. data). No study 287±295. on spiders that I am aware of has reported a Blanckenhorn, W.U., R.F. Preziosi & D.J. Fairbairn. differential investment in male and female 1995. Time and energy constraints and the evo- eggs nor a differential hatching time between lution of sexual size dimorphism - to eat or to eggs. Differential investment might also be mate? Evolutionary Ecology 9:369±381. very unlikely as the fertilization occurs to- Coddington, J.A., G. Hormiga & N. Scharff. 1997. gether with oviposition when the yolk of the Giant female or dwarf male spiders. Nature 385: egg already has been supplied (Foelix 1996). 687±688. In conclusion, this work shows that males are Eberhard, W.G. 1979. Rates of egg production by larger than females at maturity in L. triangu- tropical spiders in the ®eld. Biotropica 11(4): laris and that this depends on differences in 292±300. Foelix, R.F. 1996. Biology of Spiders (2nd ed.). Ox- behavior or physiology that affect the growth ford Univ. Press, Oxford. 330 pp. rate during juvenile stages. This result sug- Gunnarsson, B. 1988. Body size and survival: Im- gests that, because of competition for mating plications for an overwintering spider. Oikos 52: opportunities, selection for large size in males 274±282. is stronger than is selection for large size in Gunnarsson, B. 1998. Bird predation as a sex- and females. Since there is no evidence on costs size-selective agent of the arboreal spider Pityoh- of rapid growth in males, this could mean that yphantes phrygianus. Functional Ecology 12: there is intraspeci®c competition that affects 453±458. males and females asymmetrically. It is also Gunnarsson, B. & A. Andersson. 1992. Chromo- possible that female lifetime reproductive suc- some variation in embryos of a solitary spider, cess does not increase monotonically with Pityohyphantes phrygianus, with skewed sex ra- size, hence the optimal body size for a female tio. Hereditas 117:85±91. Gunnarsson, B. & J. Johnsson. 1990. Protandry is lower than the maximum attainable size. and moulting to maturity in the spider Pityohy- ACKNOWLEDGMENTS phantes phrygianus. Oikos 59:205±212. Head, G. 1995. Selection on fecundity and varia- I am grateful to Donald Blomquist for help tion in the degree of sexual size dimorphism with statistics. Bengt Gunnarsson, Malte An- among spider species (Class Araneae). Evolution dersson and Ingela Danielsson commented on 49(4):776±781. early versions of the manuscript. I also thank Higgins, L.E. 1992. Developmental plasticity and LAÊ NGÐSEXUAL SIZE DIMORPHISM 71

fecundity in the orb-weaving spider Nephila cla- Sokal, R.R. & F.J. Rohlf. 1995. Biometry (3rd ed.). vipes. Journal of Arachnology 20:94±106. W.H. Freeman & Co., New York. 887 pp. Marshall, S.D. & J.L. Gittleman. 1994. Clutch size Stearns, S.C. 1992. The Evolution of Life Histo- in spiders: Is more better? Functional Ecology 8: ries. Oxford Univ. Press, Oxford. 249 pp. 118±124. Stumpf, H. 1990. Observations on the copulation Nielsen, N. & S. Toft. 1990. Alternative male mat- behaviour of the sheet-web spiders Linyphia hor- ing strategies in Linyphia triangularis (Araneae, tensis Sundevall and Linyphia triangularis Linyphiidae). Acta Zoologica Fennica 190:293± (Clerck) (Araneae: Linyphiidae). Bulletin of the 297. Society of European Arachnology 1:340±345. Prenter, J., W.I. Montgomery & R.W. Elwood. Stumpf, H. & K.E. Linsenmair. 1995. Observations 1995. Multivariate morphometrics and sexual di- on the mating systems of two spiders, Linyphia morphism in the orb-web spider Metellina seg- hortensis Sund. and L. triangularis (Cl.) (Liny- mentata (Clerck, 1757) (Araneae, Metidae). Bi- phiidae: Araneae). Revue Suisse de Zoologie ological Journal of the Linnean Society 55:345± (vol. hors seÁrie) 627±634. 354. Suter, R.B. 1990. Determinants of fecundity in Prenter, J., W.I. Montgomery & R.W. Elwood. Frontinella pyramitela (Araneae, Linyphiidae). 1997. Sexual dimorphism in northern temperate Journal of Arachnology 18:263±269. spiders: Implications for the differential mortal- Toft, S. 1989. Mate guarding in two Linyphia spe- ity model. Journal of Zoology, London 243:341± cies (Araneae: Linyphiidae). Bulletin of the Brit- 349. ish Arachnological Society 8(2):33±37. Reiss, M.J. 1989. The Allometry of Growth and Vollrath, F. 1980. Male body size and ®tness in the Reproduction. Cambridge Univ. Press, Cam- web-building spider Nephila clavipes. Zeitschrift bridge. 182 pp. fuer Tierpsychologie 53:61±78. Rovner, J.S. 1968. Territoriality in the sheet-web Vollrath, F. & G.A. Parker. 1992. Sexual dimor- spider Linyphia triangularis (Clerck) (Araneae, phism and distorted sex ratios in spiders. Nature Linyphiidae). Zeitschrift fuer Tierpsychologie 360(6400):156±159. 25:232±242. Watson, P.J. 1990. Female-enhanced male compe- Rubenstein, D.I. 1987. Alternative reproductive tition determines the ®rst mate and principal sire tactics in the spider Meta segmentata. Behavioral in the spider Linyphia litigiosa (Linyphiidae). Ecology and Sociobiology 20:229±237. Behavioral Ecology and Sociobiology 26:77±90. SAS Institute. 1998. StatView, 2nd ed. SAS Insti- Watson, P.J. 1991. Multiple paternity and ®rst mate tute Inc., Cary, North Carolina. sperm precedence in the sierra dome spider, Lin- Seitz, K.-A. 1971. Licht-und elektronenmikrosko- yphia litigiosa Keyserling (Linyphiidae). pische Untersuchungen zur Ovarienentwicklung Behaviour 41(1):135±148. Wiklund, C., S. Nylin & J. Forsberg. 1991. Sex- und Oogenese bei Cupiennus salei Keys. (Ara- related variation in growth rate as a result of se- neae, Ctenidae). Zeitschrift fuer Morphologie der lection for large size and protandry in a bivoltine Tiere 69:283±317. butter¯y, Pieris napi. Oikos 60:241±250. Simpson, M.R. 1995. Covariation of spider egg and clutch size: The in¯uence of foraging and Manuscript received 27 July 1999, revised 1 June parental care. Ecology 76(3):795±800. 2000.