1997. The Journal ofArachnology 25:352-360 EFFECTS OF PREY SUPPLEMENTATION ON SURVIVAL AND WEB SITE TENACITY OF ARGIOPE TRIFASCIATA (ARANEAE, ARANEIDAE): A FIELD EXPERIMENT Bonnie Jean McNett and Ann L. Rypstra^: Department of Zoology, Miami University, Oxford, Ohio 45056 and Hamilton, Ohio 45011 USA ABSTRACT. The effects of prey capture on web site tenacity and survivorship ofArgiope trifasciata (Araneae, Araneidae) were studied in two old field habitats in southwestern Ohio. Adult females were studied in habitats dominated by grass or thistle plants. In manipulation plots, we added two crickets to the webs of approximately half the spiders. We were able to quantify differences in prey intake using morphological measurements that changed with food consumption. The spiders that did not receive sup- plemental food were similar in size to unmanipulated spiders in other areas that we censused. No differ- ences were observed in survivorship or web site tenacity of spiders in grass vs. thistle habitats. No difference in survivorship was observed between fed spiders and those left to natural prey capture. How- ever, spiders receiving supplemental prey relocated their webs less frequently than those spiders that were unsupplemented. The selection of a site in which to live and with conspecifics, the spider’s age, and/or the forage can be a critical decision for a spider action of predators, can influence a spider’s since food intake can have a substantial effect decision to leave orremain in a given location on the spider’s ability to survive, grow, and (Eberhard 1971; Enders 1975, 1976, 1977; & ultimately reproduce (Riechert Gillespie Wise 1975; Pasquet 1984; Spiller 1984; Voll- 1986; Vollrath 1987). The webs that spiders rath & Houston 1986; Gillespie & Caraco use as foraging tools are energetically costly 1987; Craig 1987; Smallwood 1993). Clearly, (Prestwich 1977; Peakall & Witt 1976), and it if a particular population is not food limited, is not possible for web-spiders to sample their prey capture should not have an effect on web habitat extensively before settling to forage in site tenacity (Eberhard 1971; Enders 1976; a particular place (Janetos 1986; Vollrath Wise 1993). However, Olive (1981) argues 1985, 1987). As a result, the initial selection that the phenologies of orb-weaving spiders, of a site must be based on habitat features and particularly those in the genus Argiope Au- the appropriateness of web attachment sites douin 1827, are tied to the seasonality of in- (Pasquet 1984; Hodge 1987a; Bradley 1993). sects in their environment and that they Once the initial web is constructed, the spider evolved under the constraints of food limita- acquires additional information on prey cap- tion. In enclosure experiments with Argiope ture which can influence whether it stays or trifasciata (ForskM 1775), he found that they leaves. A number of studies on a variety ofspecies abandon sites with lower rates ofprey capture and aggregate in areas where he supplied prey suggest that web-spiders use recent informa- A at a higher rate (Olive 1982). field study tion on prey capture in deciding whether to withA. keyserlingi Karsch 1881, revealed that stay or leave a particular site (Turnbull 1964; food supplementation, even over a few days, Janetos 1982; Olive 1982; Pasquet 1984; Voll- decreased the tendency of individuals to re- rath 1985; Gillespie 1987; Rubenstein 1987; Hodge 1987b; Provencher & Riechert 1991; locate their webs (Bradley 1993). However, in A experiments with A. aurantia Lucas 1833, Bradley 1993). variety of other factors un- related to prey capture, such as the frequency prey supplementation had no effect on web of web destruction or damage, interactions site tenacity; and the likelihood of wind dam- age appeared to be more critical to the web 'To whom correspondence should be addressed. relocation decision (Enders 1975, 1976). 352 — McNETT & RYPSTRA—PREY SUPPLEMENTATION EFFECTS 353 Olive’s (1981, 1982) results influenced us in thistle and three 5 X 5 m census plots in m to further investigate the relationship between grass, located at least 100 away from the prey capture, survivorship, and web site te- manipulation plots, were used as control areas nacity in A. trifasciata. Since the abdomen of in which the spiders were counted and mea- a spider expands as it feeds (Anderson 1974; sured but not fed. — Jakob et al. 1996), we were able to quantify Prey availability. Background prey differences between spiders that receive sup- availability was assessed in both the thistle plemental prey and those left to natural prey and grass habitats using sticky traps. Eachtrap capture withoutdisturbing themontheirwebs. consisted of a 20 X 20 cm sheet of plastic to In this way, we were able to verify that the which a thin layer of Tangle Trap® (Tangle spiders we fed consumed the prey we provid- Foot, Grand Rapids, Michigan) was applied. ed and experienced a change in their overall Traps were suspended with string thatwas tied body condition in some significant way as a either to natural vegetation or, when neces- result ofprey consumption. We then testedthe sary, to metal reinforcing rods (3.5 m in hypothesis that prey supplementation would height). Trap height was randomly determined increase the survivorship and web site resi- within the range of 15-92 cm. These values dence time ofadult females in two structurally were selected because they corresponded to distinct habitats. the range ofheights at which webs were found in 1993 (McNett 1995). Trap orientation was METHODS determined by randomly selecting a compass Study species. Argiope trifasciata is a direction. A total of nine 400 cm^ traps were conspicuous orb-weaving spider found in gar- hung in each of the six census plots in the dens, tall weeds, and grasses in the eastern early morning of 6 October 1994 and left for United States (Kaston 1948). Spiders emerge 24 h. The arthropods collected were identified fromegg sacs in May andJune (Kaston 1948). to order, counted and measured to the nearest Females mature in September when they are 0.1 mm. mm 15-25 in length, lay eggs in October and Mor—phological changes in the labora- November, and die with the onset of winter tory. Twenty-two adult femaleA. trifasciata (Scheffer 1905; Tolbert 1976). We selected were collected and allowed to establish webs this species for our investigation ofthe effects in acrylic plastic (Plexiglas®) cages measur- ofprey capture on web relocationbecause: (1) ing 45 X 45 X 7.5 cm in the laboratory. The their large size makes them easy to monitor total body length and abdomen width of all in the field; (2) in 1993, the year before this the spiders were measured afterweb construc- study was conducted, we found them to be tion. We selected these measures because it very abundant in old field habitats with den- was possible to take them without disturbing sities as high as 0.82 spiders per m^ (McNett the spider in its web. After measurement, nine 1995); and (3) although they rebuild the cap- spiders were fed one cricket (Acheta domes- ture spiral of their web each day, they reuse tica; approximately 150 mg in weight). All of the framework which attaches the web to the the spiders were then left for 24 h during vegetation and, as a result, web relocation is which time each individual replaced its cap- costly in comparison to remaining at the same ture spiral once. At that time, all ofthe spiders site (Enders 1976; Olive 1981). were measured again to determine ifmorpho- Study site.—The study population inhab- logical differences as a result of feeding ited old fields ofthe Miami University’s Ecol- would be detectable. ogy Research Center, three miles north ofOx- During the course of two years of study of ford, Butler County, Ohio, USA. Two this species we were able to obtain morpho- manipulation plots (25 X 20 m) were estab- logical measurements in the laboratory of six lishedforprey experiments. Onemanipulation females before and after eggsac deposition. plot was set up in an areadominatedby thistle Spiders were measured and left for 24 h. At (Cirsium arvense) and the second in an area that time the egg sac was noted and the spider dominatedby grasses (Elymus sp., Fustuca sp. remeasured. None of these spiders were fed andPhleum sp.). These two habitattypes were between measurements.— those that the spiders preferred in 1993 Prey manipulation. On 26-27 Septem- (McNett 1995). Three 5 X 5 m census plots ber 1994, 80 adult female A. trifasciata were 354 THE JOURNAL OF ARACHNOLOGY m collected from areas at least 100 outside the likelihood that a spider would be falsely as- plots and individually marked on their abdo- sumed dead. If we found the spider, we re- mens with non-toxic paint. Spiders were held corded its new location but, ifwe were unable in the laboratory at 15 °C in vials 1 cm in to find it, we assumed it was dead. diameter which were not large enough to al- We measured abdomen width and total low web construction andthereforeminimized body length of all spiders in the manipulation mm any changes in theircondition orhungerlevel. plots to the nearest 0.1 on 4, 8 and 16 On 28-29 September 1994, all naturally oc- census days after the prey supplementation curring A. trifasciata from each ofthe two 25 commenced. On those same dates, we counted X 20 m manipulation plots were removed. In and measured all ofthe spiders in our six cen- the early morning of 30 September 1994, we sus plots. introduced 40 randomly-chosen marked spi- Statistical analysis.—The number and size ders to each plot by placing them on vegeta- ofinsects captured on sticky traps in grass and tion approximately 2 m away from other in- thistle were compared with a one-way ANO- dividuals. The next day, we searched the plots VA. The change in body size of laboratory and marked the location of each spider’s web spiders was compared using the t-test. The by tying flagging to the vegetation near the number of spiders in grass and thistle census web. There was a low establishment rate of plots over the course of the study were com- marked spiders, so all unmarked spiders that pared using arepeated measures ANOVA. We moved into the plots after that date were as- compared the abdomen width andbody length signed to a treatment group and included in of field measured spiders in three treatments further data collection. In comprehensive sur- (supplemented, unsupplemented and cen- veys conducted in 1993, we discovered that sused) in two habitat types (grass and thistle) individuals never moved more than two me- using an two-way factorial ANOVA and then ters (McNett 1995), so we were confident in differences among the specific treatments our ability to follow and monitor web site were compared using Fisher Pairwise Com- changes of unmarked spiders that moved into parisons. These three groups were compared our manipulation plots. 4, 8, and 16 census days after the prey sup- Introduced as well as unmarked individuals plementation was begun. Fed spiders on day that established in the manipulation plots, 4 would have received prey twice (a total of were assigned randomly to one of two treat- four crickets), on day 8 they would have re- ments: one group received supplemental prey ceived prey four times (eight crickets), and on and the other group was left to natural prey day 16 they would have received prey eight capture. Supplemented spiders received two times (16 crickets). adult crickets {Acheata domestica; approxi- In order to determine the impact of supple- mately 300 mg) every other day in addition mental prey on survivorship, the total number to the prey they captured naturally. Spiders ofdays over which we were able to locate fed received supplemental prey until they could and unfed spiders in the two habitats was no longer be located, at which time they were compared using a two-way factorial ANOVA. presumed dead. A total of 26 spiders was In order to determine the impact ofprey sup- monitored in the thistle plot, 12 ofwhich were plementation on web relocations, we also used ANOVA fed crickets, and 30 spiders were monitored in the two-way factorial to compare the the grass plot, 14 ofwhich wereprovidedwith movement frequency offed and unfed spiders crickets. in the grass and thistle habitats. Spider location was monitored daily from RESULTS 1 October until no spiders could be found on 27 October 1994. If a spider was not found Spider abundances.—There was no differ- where it had been the previous day, the sur- ence between the numberofspiders inhabiting rounding 60 m^ area was visually searched. thistle or grass in the census plots (Repeated We were able to identify unmarked individu- measure ANOVA, F = 1.2, F = 0.3) (Table als by a combination of web location and ab- 1). Census plots had more spiders than we dominal patterns. Since we never observed a were able to establish in our manipulation spider move more than 2 m from a previous plots {F = 18.96, P = 0.032) (Table 1). Since web site, this large search area eliminated the densities in manipulation plots were low com- McNETT & RYPSTRA—PREY SUPPLEMENTATION EFFECTS 355 — Table 1. Number of spiders (mean ± standard duced both spider abdomen width (t = 8.2, F error) per square meter in old field habitats domi- = 0.0004) and total body length (t = 7.9; F nated by grass or thistle. In a two-way ANOVA, = 0.0005) in the laboratory (Table 2). In ad- there werenodifferencesbetweendensitiesingrass dition, the spider’s abdomen appeared shrunk- ianndmatnhiisptulleahtaibointaptlsot(sFw=er1e.2,siFgni=fi0c.an3t)lbyultodweenrsitthiaens en and wrinkled after e—ggs were laid. densities in census plots (F = 18.9, F = 0.03). Prey manipulation. Although the spiders increased in size over the course of the ex- Plot type Grass Thistle periment, there were no significantdifferences observed between spiders inhabiting grass or Census 0.52 ± 0.25 0.39 ± 0.12 Manipulation 0.06 ± 0.02 0.05 ± 0.02 tlheinsgttleh (iTnwtoh-ewaaymofaucnttoritahlatANeiOtVheAr,toFta=l b1.o3d4y, F = 0.26) or abdomen width (F = 0.24, F = 0.63) changed during the course of the study pared to natural densities, we believe we suc- (Table 3). Unsupplemented spiders within our cessfully eliminated density as a potentially manipulation plots were not different in either confounding factor in our study of the effects measure ofsize from the control spiders in the of habitat type and prey capture on web re- census plots 4, 8 and 16 days after the food location. supplementation was begun (Fisher pairwise capPtrueryeda6b0u.0nd±anc10e..1—iSntsieccktys itnrap2s4 ihnwghriacshs comparisons, F > 0.05) (Fig. 1). However, was not significantly different from 80.7 ± spiders that received supplemental prey had wider abdomens than spiders in the other two 20.2 insects captured by these traps in thistle (One-way ANOVA, F = 0.21, F = 0.4). The groups (unsupplemented and censused) on all mean size of the insects captured by sticky three dates tested (Fisher pairwise compari- traps in grass (2.39 ± 0.19 mm) was also very sons, F < 0.05) (Fig. 1). Likewise, spiders similar to the mean size captured in thistle receiving additional prey were longerthan un- (2.46 ± 0.22) (F = 0.07, F = 0.7). The insect supplemented spiders in manipulationplots on orders Diptera and Hymenoptera made up all of those same dates and were longer than more than 90% of the captures in both habi- undisturbed spiders in the census plots on day eight. (Fisher pairwise comparisons, F < tats. — Morphological changes. In the labora- 0.05) (Fig. 1). tory, the consumption of one cricket was We were able to find fed spiders for 13.6 ± enough to increase total body length by 0.42 1.4 days which was not significantly different ±0.16 mm, whereas unfed individuals shrank from the survival of 11.5 ± 1.4 days we ob- by 0.21 ± 0.16 mm in 24 h (r = 7.35, F = served for unfed individuals (Two-way fac- 0.0005) (Table 2). Likewise, the abdomen torial ANOVA, F = 0.99, F = 0.33) (Table width offed spiders increased by 1.34 ± 0.11 3), We were also able to locate spiders in the mm while the abdomen width of unfed indi- thistle 14.6 ±1.5 days and in the grass 10.7 viduals decreased by 0.22 ± 0.15 mm in 24 ± 1.2 days, but this difference was not sig- h (t - 60.70, F < 0.0001) (Table 2). These nificant at the 0.05 level (F = 3.68, F = 0.06) differences verify that these measurements are (Table 3). an indicator ofrecent feeding history and spi- Fed individuals remained at web sites an der condition. The deposition ofan eggsac re- average of 12.5 ±1.4 days which was signif- — Table2. Measurements (mm)ofthetotalbodylengthandabdomenwidthoffemaleArgiopetrifasciata in the laboratory (mean ± standard error). Fed individuals received one cricket (150 mg) whereas unfed individuals and those that produced eggsacs received no food. First measurement After 24 hours Difference Status n Length Width Length Width Length Width Fed 9 14.1 ± 0.8 7.1 ± 0.5 14.5 ± 0.7 8.5 ± 0.5 +0.4 ± 0.2 +1.3 ± 0.1 Unfed 13 15.2 ± 0.5 7.8 ± 0.4 15.0 ± 0.6 7.6 ± 0.4 -0.2 ± 0.2 -0.2 ± 0.2 Produced eggsac 6 14.1 ± 0.8 7.4 ± 0.4 12.5 ± 0.4 5.8 ±0.4 -1.6 ± 0.2 -1.5 ± 0.2 356 THE JOURNAL OF ARACHNOLOGY — Table 3. Results of prey manipulation experiment in which fed spiders in thistle and grass were compared to spiders left to natural prey capture. Morphological measurements (abdomen width and total length) represent the difference between the fourth day afterprey supplementation began andthe sixteenth day after supplementation began. Data are expressed as mean ± standard error. Fed Unfed Treatment Vegetation Interaction Change in {n = 10) {n - 10) F = 3.61 F = 0.24 F = 1.04 abdomen width P = 0.08 P = 0.63 P = 0.32 (mm) in 12 days Grass {n = 7) 1.50 ± 0.03 1.15 ± 0.55 Thistle (« = 13) 2.18 ± 0.46 0.94 ± 0.20 Both habitats 1.80 ± 0.30 0.98 ± 0.18 Change in body F = 1.22 F = 1.91 F = 0.08 length (mm) in {n = 10) {n = 10) P = 0.28 P = 0.19 P = 0.78 12 days Grass {n = 7) 0.64 ± 0.73 0.10 ± 0.50 Thistle (« = 13) 2.06 ± 0.54 0.83 ± 0.45 Both habitats 1.19 ± 0.45 0.68 ± 0.38 Total days located F = 0.99 F = 3.68 F = 1.71 {n = 26) {n = 30) P = 0.325 P = 0.061 P = 0.187 Grass {n = 30) 11.0 ± 1.8 6.8 ± 1.2 Thistle {n = 26) 14.3 ± 2.3 10.8 ± 2.2 Both habitats 13.1 ± 1.8 8.6 ± 1.6 Web relocations {n = 26) {n = 30) F = 5.60 F = 0.12 F = 2.98 per spider P = 0.022 P = 0.750 P = 0.09 Grass {n - 30) 0.3 ± 0.1 0.4 ± 0.2 Thistle {n = 26) 0.0 ± 0.0 0.6 ± 0.2 Both habitats 0.2 ± 0.1 0.5 ± 0.1 icantly longer than the unfed spiders which behavior we observed. The fact that we could remained only 8.7 ± 0.2 days. This difference take these hunger measurements in the field verifies that fed spiders had significantly few- without disturbing the spider is a desirable er web relocations (Two-way factorial ANO- feature of this system. Since we were able to VA, F = 6.99, P = 0.011) (Table 3). Spiders demonstrate that there was no impact of hab- in the thistle relocated their webs with the itat or manipulation on these measures, only same frequency as the spiders located in the the supplemental prey that we provided can grass {F = 0.0001, P = 0.95) (Table 3). account for the differences we observed. Nu- merous studies have associated prey capture DISCUSSION with web site tenacity (Turnbull 1964; Janetos An increase in prey capture by adult female 1982; Olive 1982; Riechert & Gillespie 1986; Argiope trifasciata influences the decision to Gillespie 1987; Vollrath 1987; Rubenstein relocate or continue foraging in the same web 1987; Bradley 1993 and references therein) site. These data are consistent with the results but the quantification of prey capture in the of Olive’s (1982) enclosure experiments in past has always been prey in the web rather which A. trifasciata individuals tended to than some measure ofactual intake by the spi- leave areas in enclosures where food was not der as we were able to accomplish. provided and aggregate in regions where food One possible confounding factor that might was supplemented. The fact that we were able affect ourmorphological measurements would to quantify an increase in spider condition via be the production of an eggsac which sub- morphological measurements verifies that the stantially reduces the spider’s abdomen size food we were providing was sufficient to af- and changes its appearance. However, we did fect the spiders and provides a close link to not observe the same kind ofemaciation after food as the mechanism causing the changes in egg laying in individuals we were monitoring McNETT & RYPSTRA—PREY SUPPLEMENTATION EFFECTS 357 in the field that we saw in laboratory spiders. Since eggsacs are deposited very late in the season and since this species produces only one clutch per year (Tolbert 1976), it is likely that the spiders were dying shortly after the production of their egg sacs in the field, per- haps due to an increase vulnerability to pre- dation or other environmental stressors. In any case, we would predict that the spiders receiv- ing food supplements would be most likely to produce eggs since food intake is positively correlated with egg production in many spe- cies of spiders (Wise 1993 and reference therein). Therefore, ifegg sac production were confounding our results, it would have re- duced the likelihood of seeing the significant differences in body size we observed between the food supplemented and unsupplemented spiders in this study. Spiders are frequently categorized as food limited in nature because they can survive long periods of starvation (Anderson 1970, 1974), have low metabolic rates (Anderson 1970; Carrel & Heathcote 1976; Nakamura 1987), and the fact that they tend to aggregate in highprey areas (Olive 1982; Rypstra 1989). It has been suggested that the plasticity ofthe abdomen in spiders is an adaptation to prey shortages because it enables spiders to con- sume large amounts of prey when it is abun- dant and store it for subsequent lean periods (Wilson 1971; Anderson 1974). Since the ability of a spider’s abdomen to expand with consumption should decrease as it reaches its maximum, the substantial morphological changes we observed suggested that the spi- ders in our population were not close to sati- ation. Likewise the fact that manipulating the prey they consumed altered their web site te- nacity provides further evidence that food is a limiting resource for this web-building spi- WIDTH (MM) der (Wise 1993). Food supplementation had more consistent — Figure 1. Totalbody length and abdomenwidth effects on the spider’s abdomen width than on (mean ± SE) of spiders measured 4, 8 and 16 days total body length (Fig. 1). The abdomen is after prey supplementation was begun. On Day 4, flexible and therefore changes size with feed- abdomen width was significantly different among ing, whereas the cephalothorax is fixed in size glaerbnodguotphmsew(naFsw=inodtt4h,(4F3(,F==P=2.=3,3.0P9.50,>03P0).0=w5h).e0r.Oe0na0s7D)atoaytnal8d,bbtoootdtayhl dfoormaengiwviednthinisstaar.moOrfeoduirremcetamseuarseumreentmse,ntabo-f body length (F = 2.6, P = 0.048) were signficantly the changes in condition the spider experi- different among the treatments. On Day 16, abdo- enced since any abdominal changes reflected men width was signficantly different (F = 4.25, P in total body length are damped by the ceph- = 0.0085) but total body length was not (F = alothorax size, which cannot change. As a re- 1.389, F > 0.05). sult, we saw less consistent differences among 358 THE JOURNAL OF ARACHNOLOGY treatments over the course of the experiment we never observed individuals moving more m in body length than in abdomen width. In ret- than two in a web relocation event (McNett rospect, a more accurate assessment of spider 1995), suggest that intraspecific interactions condition would have been obtained ifwe had were not very important in our these experi- taken measurements of the cephalothorax ments. Additionally, it may be that the low alone or some other body part that we knew densities with which we were working and the did not change with feeding. Then we could elimination of spider-spider interactions as a have scaled body condition on absolute body disturbance, accounts for the relatively long size as reconamended by Jakob et al. (1996). residence times that we observed compared to Optimality theory predicts that the amount other orb-weaving spiders. of time an organism remains at a site should The size and web relocation behavior ofA. be related to some combination of prey cap- trifasciata in the grass and thistle habitats we ture at that site and their investment in that compared were surprisingly similar. Prey cap- site (Pyke et al. 1977). If this is true then, in ture of spiders in thistle must have been sim- a given habitat, spiders with more energeti- ilar to that in the grass because we uncovered cally costly webs should have longer web res- no morphological differences in the spiders idence times since it should take them longer inhabiting the two habitats (Table 2). This re- to recoup the investment in the web itself (Ja- sult is supported by our captures in insect netos 1986; Riechert & Gillespie 1986). The traps which failed to reveal any differences residence times that we recorded for unsup- between these two habitats in prey activity at plementedA. trifasciata were around 8.5 days this time in the season. Since Enders (1975, which is substantially longer than the time re- 1976) related web relocation to destruction by ported (3 days) for a wide variety ofotherorb- wind, we expected to see more relocation weaving spiders (Janetos 1982; Olive 1982; events by spiders living in grass since it offers Riechert & Gillespie 1986; Smallwood 1993). a less sturdy web support than thistle. Perhaps, Even the linyphiids with semi-permanent at least in the season of this study, wind was webs that Janetos (1982) studied had resi- not sufficiently damaging to affect the spider’s dence times around 5 days. In contrast, resi- behavior. dence times of the linyphiid with a semi-per- Although not significant at the 0.05 level, manent web, Neriene radiata (Walckenaer it is tempting to speculate on the nearly sig- 1844), were about 10 days; a value much clos- nificant difference in survival between ani- er to those we observed inA. trifasciata (Mar- mals in the thistle and those in the grass (P = tyniuk 1983). Since A. trifasciata has a large 0.06, Table 2). Indeed, since we were moni- web and reuses some portion of the support toring such a short period in the end of the infrastructure, the construction of an entirely spider’s life, it is surprising that there is any new web in a new location may be more cost- suggestion of a difference by habitat in the ly than the other orb-weavers investigated. timing of their death at the onset of winter. The large body size of this spider at late in- Horton (1980) found that Argiope in North stars prevents from moving by ballooning and American old field habitats are subject to sub- it appears to walk awkwardly off of the web. stantial bird predation and that the zig-zag sta- As a result, exploring for new web sites is a bilamentum offers them some protection from risky and energetically costly endeavor forA. birds. For those of us who have monitored trifasciata. spiders in thistle habitats, it is not difficult to When spiders reach high densities then in- believe that the irritating leaves of this plant teractions with one another can influence web could provide the spiders some protection site tenacity (Hoffmaster 1986; Rypstra 1985; from a variety of vertebrate predators which Smallwood 1993). It seems unlikely that web may have contributed to the near significant take-overs or spider interactions on the webs difference in survival we observed. were factors in this study. In experimental In summary, these data demonstrate that plots the spider density was only 0.05 indi- change in prey intake is a major factor influ- viduals per m^ in the thistle and 0.06 individ- encing web site tenacity of these large orb- uals per m^ in the grass and the spacing was weaving spiders. The difference in body con- fairly uniform across the plots. The low den- dition between spiders that received supple- sities in these experiments and the fact that mental prey and those that did not was the McNETT & RYPSTRA—PREY SUPPLEMENTATION EFFECTS 359 overriding difference between the spiders Gillespie, R.J. 1987. The mechanism ofhabitat se- studied here even though we also compared lectionin the long-jawedorb-weaving spider, Te- spiders in two structurally different old field tragnatha elongata (Araneae, Tetragnathidae). J. habitats. The ease with which we could verify Arachnol., 15&:81-90. changes in body condition make detailed anal- Gillespie, R.J. T. Caraco. 1987. Risk sensitive foraging strategies of two spider populations. ysis of the impact of food intake on the ecol- Ecology, 68:887-899. ogy and behavior ofA. trifasciata in a natural Hodge, M.A. 1987a. Factors influencing web site setting possible. residence ofthe orb wea—ving spider, Micrathena ACKNOWLEDGMENTS gracilis. Psyche, 94:363 371. Hodge, M.A. 1987b. Macrohabitat selectionbythe We especially thank Cameron Eicherforhis orb-weaving spider,Micrathenagracilis. Psyche, help and support at critical times during this 94:347-361. study. C. Ball, L. Barghusen, C. Brandt, G. Hoffmaster, D.K. 1986. Aggressionintropicalorb- Cochran. J. Dobyns, & S. Modica also pro- weaving spiders a quest for food? Ethology, 72: vided valuable assistance with various aspects 929-945. of this research project. Early drafts of this Horton, C.C. 1980. A defensive function for the paper benefited from input from D. Claussen, stabilimenta of two orb weavin—g spiders (Ara- neae, Araneidae). Psyche, 87:13 20. Rsh.alLl.eeW,eO.areLogurcatkesf,ulBt.oSRt.eiSnclyh,aeafenrdfSo.r eMxatrr-a Jaktoibm,atEi.Jn.g,fSi.tDne.ssM:arAshcaolm—lp,a&riGso.nW.ofUebtozd.y1c9o9n6d.itEiso-n patience, advice, and assistance with the sta- indices. Oikos, 77:61 67. tistics. Funding for this project was provided Janetos, A.C. 1982. Foraging tactics oftwo guilds by Sigma Xi, the Scientific Research Society, ofweb-spinning spiders. Behav. Ecol. Sociobiol., the Department of Zoology and the Hamilton 10:19-27. Campus of Miami University. Voucher spec- Janetos, A.C. 1986. Web-site selection: Are we imens are available in the Hefner Zoology asking the right questions? Pp. 9-22. In Spiders, Museum, Miami University, Oxford, Ohio Webs, Behavior and Evolution. (WA. Shear, 45056 USA. ed.). Stanford Univ. Press, Stanford, California. Kaston, B.J. 1948. Spiders ofConnecticut. Bulletin LITERATURE CITED 70, State Geol. Nat. Hist. Surv., Hartford, Con- Anderson, J.E 1970. Metabolic rates of spiders. necticut. Comp. Biochem. Physiol., 33:51-72. Martyniuk, J. 1983. Aspects ofhabitat choice and Anderson, J.E 1974. Responses to starvationinthe fitness in Prolinyphia marginata (Araneae: Lin- spiders Lycosa lenta Hentz and Filistata hiber- yphiidae): Web-site selection, foraging dynam- nalis (Hentz). Ecology, 55:576-585. ics, spermcompetitionandoverwinteringsurviv- Bradley, R.A. 1993. The influence of prey avail- al. Ph.D. dissertation, SUNY, Binghamton, New ability and habitat on activity patterns and abun- York. dance of Argiope keyserlingi (Araneae: Aranei- McNett, B.J. 1995. Mechanismofhabitat selection dae). J. Arachnol., 21:91-106. in the silver garden spider {Argiope trifasciata). Carrel, J.E. & R.D. Heathcote. 1976. Heart rate in M.S. Thesis, Miami Univ., Oxford, Ohio. spiders: Influence ofbody size and foraging en- Nakamura, K. 1987. Hunger and starvation. Pp. ergetics. Science, 193:148-150. 287-295. In Ecophysiology ofSpiders. (W. Nen- Craig, C.L. 1989. Alternative foraging modes of twig, ed.). Springer-Verlag, Berlin. orb web weaving spiders. Biotropica, 21:257- Olive, C.W 1981. Optimal phenology and body- 264. size of orb-weaving spiders: Foraging con- Eberhard, W.G. 1971. The ecology of the web of straints. Oecologia, 49:83-87. Uloborus diversus (Araneae: Uloboridae). Oec- Olive, C.W. 1982. Behavioral response of a sit- ologia, 6:328-342. and-wait predator to sp—atial variation in foraging Enders, E 1975. Change in web site in Argiope gain. Ecology, 63:912 920. spiders (Araneidae). American Mid. Nat., 94: Pasquet, A. 1984. Predatory-site selection and ad- 484-490. aptation ofthe trap in four species oforb-weav- Enders, E 1976. Effects of prey capture, web de- ing spiders. Biol. Behav., 9:3-19. struction, and habitat physiognomy on web-site Peakall, D.B. & P.N. Witt. 1976. The energy bud- tenacity of Argiope spiders (Araneidae). J. Ar- get of an orb-weaving spider. Comp. Biochem. achnol., 3:75-82. Physiol., 54:257-264. Enders, E 1977. Web siteselectionbyorb-webspi- Prestwich, K.N. 1977. The energetics of web ders, particularlyArgiope aurantia Lucas. Anim. building in spiders. Comp. Biochem. Physiol., Behav. 25:694-712. 57:321-326. 360 THE JOURNAL OF ARACHNOLOGY Provencher, L. & S.E. Riechert. 1991. Short-term Tolbert, W.W 1976. Population dynamics of the effects of hunger conditioning on spider behav- orb-weaving spiders,Argiope trifasciata andAr- ior, predation, and gain of weight, Oikos, 62: giope aurantia (Araneae: Araneidae). Ph.D. dis- 160-166. sertation. Univ. Tennessee, Knoxville, USA. Pyke, G.H., H.R. Pulliam & E.L. Chamov. 1977. Turnbull, A.L. 1964. The search for prey by the Optimal foraging: A selective review of theory web-building spider Achaearanea tepidariorum and tests. Q. Rev. Biol., 52:137-154. (C.L. Koch) (Araneae: Theridiidae). Canadian & Riechert, S.E, R.G. Gillespie. 1986. Habitat Entomol., 96:568-579. choice and utilization in web-building spiders. Vollrath, F. 1985. Web spider’s dilemma: A risky Pp. 23-48, In Spiders, Webs, Behavior andEvo- move or site dependent growth. Oecologia, 68: lution (W.A. Shear, ed.). Stanford Univ. Press, 69-72. Stanford, California. Vollrath, F. 1987. Growth, foraging and reproduc- Rubenstein, D.I. 1987. Alternative reproductive tive success. Pp. 331-347, In Ecophysiology of tactics in the spider Meta segmentata. Behav. Spiders. (W. Nentwig, ed.). Springer-Verlag,Ber- Ecol. Sociobiol., 20:229-237. lin. Rypstra, A.L. 1985. Aggregations ofNephila cla- Vollrath, F. & A.L Houston. 1986. Previous expe- vipes L. (Araneae: Araneidae) in relation to prey rience and site tenacity in the orb spider,Nephila availability. J. Arachnol., 13:71-78. (Araneae, Araneidae). Oecologia, 70:305-308. Rypstra, A.L. 1989. Foraging success of solitary Wilson, E.O. 1971. The Insect Societies. Belknap and aggregated spiders: Insights into flock for- Press, Cambridge, Massachusetts. mation. Anim. Behav., 37:274-281. Wise, D.H. 1975. Foodlimitationofthe spiderLm- Scheffer, T.H. 1905. The cocooning habits of spi- yphia marginata: Experimental field studies. ders. Kansas Univ. Sci. Bull., 111:83-114. Ecology, 56:637-646. Smallwood, P.D. 1993. Web-site tenure in the long Wise, D.H. 1993. Spiders in Ecological Webs. jawed spider: Is it risk sensitive foraging orcon- Cambridge Univ. Press, Cambridge, England. specific interactions? Ecology, 74:1826-1835. Spiller, D.A. 1984. Competition between two spi- der species: Experimental field study. Ecology, Manuscript received 9 August 1996, accepted 1 65:909-919. May 1997.