Prey Selection and Predatory Importance of Orb-Weaving Spiders (Araneae: Araneidaeo Uloboridae) in Texas Cotton

M. NYFFELER, D. A. DEAN, NNO W. L. STERLING

Department of Entomology, Texas A&M University, College Station, Texas 77843

Environ. Entomol. f8(3): 373-380 (f989) ABSTRACT In an unsprayed cotton field in east Texas, orb weavers were one of the numerically dominant groups of spiders, constituting 1O% of all spiders collected by D-vac during the summer of 1985 (range, 0.04 individuals/m'9 in June to 0.72 individuals/m'z in Augusl). Direct counts, conducted during peak orb-weaver density in August, showed that 0.86 individuals/m" were found. More than two-thirds of all orb weavers collected by D-vac in cotton consisted of the five species Acanthepeira stellata (Walckenaer), Neoscona arabesca (Walckenaer), Cea heptagon (Hentz), Tetragnatha laboriosa Hentz, and Uloborus glomosus (Walckenaer). Their prey consisted of insects (>99%i") and spiders (<1%). Aphids, which occasionally reach pest status in Texas cotton, were the most abundant prey of all five spiders (34.6-90%). Other important prey included small dipterans, cicadellids, and hyme- nopterans. Furthermore, coleopterans were an important component in the prey of A. stellata and, N. arabesca. Together, these ffve insect groups made up >90Vo oI the prey of the orb- weaving spiders, which are characterized as generalist predators. Differences among the live spider species indicate that prey selection was occurring; this seems to be determined by web location, web inclination, and web strength. Of the orb weavers occurring in cotton, 99% were small-sized spiders (primarily C. heptagon) that intercept small prey with their delicate (about 4 cm diameter) webs. These orb weavers are predators primarily of small- sized pests such as the cotton aphid, Aphis gossApü Glover, and the cotton fleahopper, Pseudatomoscelis seriatus (Reuter).

KEY WORDS Arachnida, orb-weaving spiders, cotton, predation

Most spncrns of Araneidae and Uloboridae spin bordered by meadows composed of various grasses spiraling orb webs. Although the general biology and low growing annual Dicotyledonae that were of orb-weaving spiders is well known (reviews in mown once during this study. The cotton field had Witt et al. 1968, Levi 1978), the significance of an area of 6.5 ha with I m between rows and about these predators in the natural control of pest insects l0 cotton plants per meter of row. Cotton (variety is poorly understood, although in some agroeco- CAMD-E) was planted on 27 May and emerged systems, orb weavers constitute the most abundant in the first week of June. Parts of this field were spiders. Prey analyses have been conducted in soy- heavily infested with weeds (johnsongrass). The bean fields in Illinois (LeSar & Unzicker 1978) and field was cultivated on 10 and 29 June. We finished Kentucky (Culin & Yeargan 1982), as well as in our investigation on 16 September, at which time Polish meadows (Kajak 1965) and Swiss cereal fields the cotton had not been harvested. (Nyffeler & Benz 1979). The goal of this paper is Evaluation of Numbers of Orb-Weaving Spi- to give insight into the effect of orb weavers in an ders in Cotton. Numbers of orb weavers per square east Texas cotton field that can be used to adjust meter were estimated in two ways: through direct the species-specific indices of efficiency used in the counts in the field, and with a D-vac suction ma- tritrophic cotton insect TEXCIM model (Hartstack chine (D-vac, Riverside, Calif.). & Sterling 1988). Studies on the ecology of other Direct Counf. Numbers of small diurnal orb spiders occurring in this agroecosystem are pre- weavers can easily be assessed by counting webs sented elsewhere (Nyffeler et al. 1986, 1987a,b,c, per square meter during daylight hours. On 7 Au- 1988a,b; Dean et al. 1987). gust, small orb weavers were counted in 50 ran- domly selected l-m'samples by searching the cot- ton foliage for webs. Adults of large nocturnal Methods Materials and orb-weaving species may be overlooked because Study Area. Investigations were conducted dur- many of these spiders remove their webs during ing the summer of 1985 (June to mid-September) the daylight hours and construct retreats under in a cotton field that received no pesticide appli- cotton foliage. To accurately estimate their num- cations. This field was located 8 km west of Aus- bers, additional counts were made after dark with tonio, Tex., near Crockett in Houston Co., and was a headlamp on 14 and 19 August by walking along

N46-225x/ 89 / 0373-0330S02.00/0 @ 1989 Entomological Sciety of America This article is the copyright property of the Entomological Society of America and may not be used for any commercla! gr othel private purpose without specific written permissisn ol the Entomological Society of America 374 ENvlnoNNaeNrel ENrouot ocy Vol. 18, no.3

Table l. Web area versus ground area spun by orb Evaluation of rhe Spiders' Prey. Evidence of weavers in eastern Texas cotton during their peak num- predation by orb weavers in cotton was obtained bers, August 1985 by removing arthropod remains from webs. All No. spiders/mz Estimated cm2 arthropods found dead in spider webs were con- determined by web areaf m2 sidered as prey. Webs were searched for prey dur- Spider group visual counts (Z) ground areaa ing the day and night, and prey items were re- Mean (% total) Mean (70 total) moved from webs with forceps and preserved in 70% ethyl alcohol. They were later identified and Small orb-weaversü counted under the microscope. Some of the insect Cea heptagm 0.68r (78.9) e.66 (57.2) Othersc 0.r76 (20.4) 2.so (r4.8) carcasses are discarded by the spiders from the web after the meal (Turnbull 1973); these dropped car- Large orb-weaversd casses are usually removed by the scavenging ac- Acanthepeira stellata o.oo3 (0.3) 2.u (t2.3) Neoscona arabesca 0.003 (0.3) 2.66 (r5.7) tivities of Solenopsis inoicta Buren worker ants. Total 0.8ffi (ro0) 16.89 (roo) Because of different handling times exhibited by the spiders for differing prey taxa, our data may o Orb-web areas calculated as approximation to a circular area: be biased. (r)(D2)(Z)/5, where D is average web diameter in centimeters. ö Immatures and adults of small-sized species and immatures of Sratisrical Procedures for Testing Interspeciffc large-sized species. Difrerences of Prey Selection. Interspecific dif- c T.laboriosa, U. glomosus, and others. ferences in the selection o{ "type" of prey by the d Adults of large-sized species. investigated orb weavers were tested by computing the proportion of prey items in the four "types" of prey categories-"flying insects;" "jumping in- cotton rows and recording the numbers of large sects;" "wingless, mobile arthropods;" and "wing- orb webs hanging across the free space between less, immobile arthropods"-for each orb-weaver adjacent rows or in gaps within a row. Each night, species. Interspecific differences of proportions spiders were counted along a distance of 500 m within a type of prey category were tested by com- (walking speed about 25O rn/h), and data were paring the 0.95 conffdence intervals (CI) for pro- later converted into average number of spiders per portions using tables in Documenta Geigy (1968); square meter. Numbers per square meter of small nonoverlapping 0.95 CIs indicate signiffcant dif- orb weavers (assessed by day) and of large orb ferences at the 5% level. However, because confi- weavers (assessed by night) were combined (Table dence intervals do not constitute a rigorous statis- l) to provide an estimate of the total number of tical test, those signiffcant differences of proportions orb weavers in this cotton ffeld in August. discussed in the text were doublechecked using the D-oac Method. Twenty-ffve D-vac suction sam- 12 test for the comparison of proportions. ples (Dietrick f96f), each of I m of row, were taken weekly over a l4-wk period during the sum- mer of 1985 to obtain estimates of numbers of Results spiders and of potential prey composition. Details In this siudy, orb-weaving spiders were a nu- are described in Nyffeler et al. (1987b). Number merically dominant spider group constituting I0% of spiders per meter of row represents the number of all spiders collected during the summer by per square meter, because the lateral distance be- D-vac (total n:923; monthly means: Jrne, 15.6%; tween rows was I m. Jdy,14.6%; August, 9.3%; September, 3.5%). The Assessment of Web Size and Catching Area. taxonomic composition of orb weavers in Texas Because most orb-weaving spiders spin slightly cotton ftelds is presented in Dean et al. (1982) and asymmetrical orb webs (Nentwig f985), it follows Dean & Sterling (1987). Five species, Acanthepeira that often horizontal diameter is not equal to ver- stellata (Walckenaer), Neoscona arobesca tical diameter. In this study, we measured hori- (Walckenaer), Gea hept agon (Hentz), T etra gna- zontal diameter and vertical diameter of an orb tha laboriosa }Jentz, and Uloborus glomosus web with a meter stick, and from these two values (Walckenaer), constituted more than 80% of the we calculated the arithmetic mean used as an es- 88 orb weavers sampled by D-vac in 1985. In early timate of parameter D (see below). The catching June, when the cotton plants emerged, very few area per spider was calculated as an approximately orb-weaving spiders were in the field (0.04 + 0.04 circular area; the average square centimeters of individuals/m' [- + SE] in D-vac samples), but a web area per square meter of ground area was large number already existid in the adjacent mead- estimated as follows (see Table l): ows (assessed by direct observations and sweep sampling; D.A.D., unpublished data). The same species were found in the meadow as in cotton, cm2web areaf m2ground area:@+@ with G. heptagon constituting 75Vo of all orb weav- ers sampled in the meadow by D-vac (D.A.D., un- where D is _the average web diameter in centi- published data). Those found in cotton and the meters and Z is the average number of spiders per adjacent meadow move readily through the air by square meter (see above). ballooning (Dean & Sterling 1985), and it is likely June 1989 Nvppelrn ET AL.: Pneo,tronv Il,rponra,Ncn oF ORB Splorns 375 that meadows function as reservoirs for the colo- nization of cotton fields by orb weavers. Numbers in cotton increased to a maximum in August (0.72 € + 0.18 individuals/m'? + SE] in D-vac samples). [f z The 0.86 individuals/m' (Table 1) that were vi- sually counted is about 2O% more than the average c numbers assessed by D-vac sampling. In August, o these spiders spun an estimated average web area r of about 17 crnzfrn'ground area (Table l). 9 Of these orb weavers, about 99% were small- I si,zed-spiders (5 mm long, with G. heptagon clear- Iy dominating (Table 1). Large orb weavers, such as adults of A. stellata and N. arabesca )7 mm Iong, constituted 90% of the orb weavers prey. the plants, and T. laboriosa spins small to medium- One cotton fleahopper, Pseudatomoscelis seriatus sized webs (about 1l cm diameter) oriented at var- (Reuter), and one adult bollworm moth, Heliothis ious angles in the top half of the plant. The large, zea (Boddie), were observed as prey of orb weavers almost vertically oriented webs (about 30 cm di- (Table 2), indicating that orb weavers occasionally ameter) of adult A. stellata and N. arabesca are capture these cotton pests. In the cotton field, evi- normally hung across the free space between ad- dence of prey selection between the ftve orb-weav- jacent cotton rows and sometimes in gaps within a er species was found. row. A statistically significant difference was found A high proportion (>8O%) of flying insects was between average web diameter of large orb weav- captured by the two large orb weavers A. stelLttta ers (A. stellata, N. arabesco) and smaller orb weav- and N. arabesca, which spin their strong webs across ers (G. heptagon, T. Iaboriosa) (P < 0.00f , Mann- the free spaces between adjacent cotton rows. The Whitney U test, two-tailed). percentages of flying insects in the prey of these Cea heptagon and lJ. glomosus remain on the two species did not differ significantly (P > 0.05, hub of their webs day and night, whereas the adults 1'). The three smaller orb-weaver species, which of A. stellata and N. arabesca are nocturnal for- spin their orbs within and on the cotton plants, agers. T. laboriosa were found on the hub of their captured a significantly (P < 0.05, 1'z) lower pro- webs during the night and sometimes during the portion (<\OVa) of flying insects (Table 3). daylight hours. A. stellata, N. arabesca, and T. Jumping insects and wingless, mobile arthropods laboriosa were observed constructing webs shortly each constituted. <13% in all five orb weavers' prey before sundown. C. heptagon was observed spin- (Table 3). Seventeen workers of the red imported ning webs in the laboratory in the evening or early fire ant, S. inaicta, were trapped in webs of C. morning (Sabath 1969). No observations were made heptagon. Because this spider builds its small orbs on the time of day of web spinning of IJ. glomosus, attached to cotton leaves close to the ground, ant but another species oI Uloborus was observed workers occasionally get stuck in the sticky threads building its web I or 2 h before dawn (Eberhard and are wrapped into silk by the spider. In orbs of l97l). Feeding spiders were found in only 7Vo of other spider species positioned on higher levels the webs of C. heptagon as assessed in the first half above ground, S. inuicta workers were rarely cap- of August during daylight hours. tured; on one occasion a T. laboriosa was observed The prey of the five orb-weaving spiders in cot- consuming an S. inoicta worker. ton consisted of insects (>99Vo) and spiders (25%) in the smaller orb weavers' prey abundant potential prey on the cotton foliage. Oth- (Table 3); this difference is statistically signiftcant er important prey were small dipterans, cicadellids, (P < 0.05, x'?). 376 ENvrnoNrrpNtal ENtovol,ocv Vol. 18, no. 3

Table 2. Prey (7o by number) of five orb-weaving species and their estimated potential prey in a cotton agroecosystem in eastern Texas, summer 1985

Spider

Prey A. stellata N. arabesca C.heptagon T.laboriosa U. glomosus Estimate Zo n: lO4, n:22, n:).47, n:41, n:5O, potential prel z: 44 z: 15 z: 111 z:23 z: 16 n :58.528

Homoptera Aphididae, wingless 2.9 0 25.2 43.9 68.0 72.2 Aphididae, winged 31.7 45.5 27.2 34.1 22.O Cicadellidae 8.7 4.5 ro.2 0 4.5 Diptera, small 30.8 4.5 r5.6 t2.2 4.O 2.t Hymenoptera S. inücta, worker 1.0 0 11.6 2.4 6.0 12.6 Other, small 2.9 4.5 4.1 0 0 1.9 Coleoptera Curculionidae 00 0 0 0 0 o.l Scarabaeidae 4.8 22.7 o 0 0 0 Other 6.7 9.t o.7 0 0 0.3 Heteroptera 1.9 4.5 t.4b 0 0 1.2 Lepidoptera, adults 2.V 4.5 0 0 0 _d Orthoptera 0 0 2.t 0 0 _d Araneae, orb weavers 1.0 0 0 0 0 t.6 Other 1.9 0 t.4 0 0 o Total 100 r00 100 loo lfi) r00

n : no. prey items; a : no. spider webs. d Based on D-vac sample data; those samples have bias toward some prey types (Pieters & Sterling f973) but are at least an estimate of potential prey composition. b Includes one cotton fleahopper. c Includes one H. zea adrlt. / Mising in samples because D-vac method not suitable for collrcting Lepidoptera or Orthoptera.

Within the cotton foliage, wingless aphids con- by orb-weaving spiders (LeSar & Unzicker 1978; stituted a significantly (P < 0.05, 1'?) higher pro- Culin & Yeargan 1982; M.N., unpublished data). portion in the prey spectrum of U. glomosus Thus, we assume that most species of insects found (>60Vo), which builds horizontal webs, than in that in webs in this study were used as food. According of G. heptagon (<3O%), which builds vertical webs. to a field study by Turnbull (1960), a web-building The percentage of wingless, relatively immobile spider accepted 98% of 153 species of insects insects in the prey of T.Iaboriosa (44%, Table 3), trapped in the web, demonstrating the low rejec- whose webs are oriented at various angles, was tion rate of such spiders. Intuitively, many defenses between the values of G. heptagon (25%) and U. of prey may be largely immobilized by capture in glomosus (68%) and differed from them signifi- a web. cantly (P < 0.05, 1'). If we compare the ftve spiders of this study and their selection of type of prey in relation to the different web positions, web inclinations, and other Discussion web characteristics (Fig. l; Table 3), the following The investigated orb weavers captured a variety pattern emerges. of prey taxa characterizing them as generalist pred- Web position was found to be a determinant of ators (Table 2), which is in agreement with previous prey selection by comparing small orb webs on prey analyses of orb-weaving spiders (LeSar & Un- cotton foliage with the large orb webs spun be- zicker 1978; Nyffeler & Benz 1978; Culin & Year- tween cotton rows. We found that orb webs be- gan 1982; Nyffeler 1982; Nyffeler et al. 1986, tween the rows captured a higher proportion of 1987c). One could question if the carcasses of ar- flying and a lower proportion of wingless, immobile thropods found in spider webs actually represent prey. the diet or if some of them may have been caught Orb webs may function as protective barriers in the webs accidentally. From the point of view (sensu Turnbuil f973) against S. inoicta, which are of natural pest control, it matters little whether the aggressive predaceous ants occasionally biting into spider actually feeds on insects caught in webs if the legs of spiders (M.N., unpublished data). Web the "prey" dies anyway. However, the insect groups position also was found to be a determinant of prey which are considered in this paper as prey of orb selection in other studies (Nyffeler & Benz 1978, weavers (such as aphids, cicadellids, dipterans, ants, Olive 1980, Pasquet 1984). coleopterans, heteropterans, lepidopterans, and Web orientation may be another factor deter- others) have been observed in the field being eaten mining prey selection, evidenced by the observa- June 1989 NvppsI-nn ET AL.: Pnnperonv IuponteNcr on Ons Spronns ö/ /

6 o x O O F 8 nN o d c o o O ! o 6 - ='E; - -- o oo :- 9;- +ri6 41 : i;6: I I I I \ dJ O A ! 6Oi6 o cE c a ivv6 bo '=tr öN-,. -L O a ^ o o i F3 E E ;OöJ O 4 o > L6V- o öo i ü o ü 0 O o E ts o -9 o +_9 + F e eboa Y ^^^ tr S o o FEE: EqTA I ( u! l;:l c -iil t q.ix;: -- tr e 6v:= Y:) '= d= o Q 66so b0 > ! 8 ä @r6i6 o qt =}i.;o> * $o 6q) ' u o o

öo o I ! = o ;* L o : d X! 666c4 ä .i€s o O A A ! +@r-fiiJ o i? 'i.Y=' o O ör^6N odi .: F ! ö d oNi6 o ts 6 >-e tri> niid O >-69? - :f, e eeä o.; O a >l o ü- ' b0 (, ud € .id e- Ju ö H OO 6 oo o 0 3 o d! ä Y o^^^ r ugi oÖnn oi eP B : tr -!l - idii =o ä --z =! a 1? q! . tr= 4 U -v o UF ln !:;9"^9 ! r.: d* nsoo"qn o OU o\ cq>> o o o 9-e

.= o -.v 3 E 8 ! u i!!o-=t s b0 E EÄ Px *- 9 !t ä oa h b0 E d: Y d oo l: >*: {) .1 : d '?,.4 ö I E;rl a ! uY+- ^^iN^^ E E * ö F€- qi\o9 : : ö c (osN@ I - Tn- - o o d tr-6 i o ;;8r'i 3 \\qe ! ! q>>!.: d $odN o l.E{hi AE I€ .N p-A--9-q € o4 .9 >'3 y äg {'ä !o ä f,E= ää;e :6-! ga i ,' o ! ; .,! ä { .Fo q .9; >n - &e p-B E _ _*i .: H:Pö€ E':.! (,äOd E ä :..?äE aaa?ä-o.P.ä € E;.EEE.E ü? X;,95O% of the moths encountering the webs of a winter wheat fields that wingless aphid morphs fell large orb weaver were able to escape; on the other from foliage to the ground at a rate of up to >100 hand, Whitcomb et al. (1963) found many noctuid individuals/m'/d. Many of these falling aphids were moths captured in large orb webs in Arkansas cot- intercepted in horizontally oriented spider webs. ton fields. Chacon & Eberhard (1980) suggested that hori- Feeding spiders were found in 5O% in total prey of all five in August has on the average a 50 times smaller orb-weaver species). No correlation between mesh catching area than an adult large-sized orb weaver size and prey size was found by Nentwig (1983) in (".g., A. stellata). However, G. heptagon, having experiments with artificial webs. reached maturity in fall, build webs of 10-12 cm Web strength, which is a function of web size diameter (Sabath I969). (according to Craig [1987], high-energy-absorbing In east Texas where small orb weavers predom- webs tend to be large), was found to be another inate, these spiders are primarily predators of small- determinant of prey selection by orb weavers. Small, sized pests such as the cotton aphid, Aphis gosswü delicate orb webs (e.g., those of G. heptagon, U. Glover, and the cotton fleahopper. Orb-weaver glomosus, and T. laboriosa) are designed for the predation on these two insects was reported by interception of small insects and are not suited to Kagan (1943) in Texas cotton. Predation of the boll capture larger prey such as Lepidoptera and Co- weevil and Hellothls spp. by orb-weaving spiders leoptera (LeSar & Unzicker 1978, Culin & Yeargan appears to be insigniffcant because of the apparent f982). On the contrary, the large, strong orb webs inability of small orb weavers to overcome larger of adult A. stellata and N. arabesca are designed pests and because of the low numbers of large- for the interception of small and large prey. This sized orb weavers in the cotton fields (see above). is consistent with observations of Castillo & Eber- In 1985, all key pests (sensu Bohmfalk et al. 1983) hard (i983), who reported that larger orb weavers combined constituted only about 1% of the prey captured larger prey and a greater variety of prey of G. heptagon and A. stellata and were missing sizes than smaller spiders. in the prey spectra of the other orb-weaver species Coleoptera are excluded as prey of small orb (Table 2). The low frequency of key cotton pests weavers, as observed in T. laborioso, which was is in part because of the low densities of those key seen eliminating entangled beetles from the web pests in that area during the summer of 1985 by violently shaking the web until the beetle fell, (D.A.D., unpublished data). Conversely, occasional by ignoring the beetles until they worked them- pests (sensu Bohmfalk et al. 1983) such as aphids selves free and could escape, and by cutting the constituted >3O% of the spiders' prey. Predaceous web around an adult beetle and allowing it to drop arthropods were rarer (<7%) in the prey of orb from the web (LeSar & Unzicker 1978, Culin & weavers except for occasional capturing of S. in- Yeargan 1982). oicta by G. heptagon. Adult Coleoptera constituted >lOVo of the prey of the large orb weavers in this study and almost Acknowledgrnent SOVo of the prey of N. arabesco in Kentucky soy- bean fields (Culin & Yeargan 1982). These authors We thank L. N. Brown, farmer in Austonio, for per- assume that the high proportion of Coleoptera in mission to carry out this project in his cotton ffelds and June 1989 Nvrrulnn ET AL.; Pnnoaronv InpontaNcp oF ORB Sprlsns 379

T. L. Payne for the use of his trailer as a field lab. We 1985. Top-bottom asymmetry in vertical orbwebs: a also thank C. W. Agnew and R. G. Breene for their functional explanation and attendant complications. review of this paper. This research was made possible Oecologia 67: lll-112. by a postdoctoral fellowship of the Swiss National Science Nyfreler, M. 1982. Field studies on the ecological role Foundation (Research Commission of the Swiss Acade- of the spiders as predators of insects in agroecosys- my of Science) granted to M.N. and by the Expanded tems. Ph.D. dissertation, Swiss Federal Institute of Research Project H-2591-2100 of the Texas Agricultural Technology, Zurich. Experiment Station. Approved for publication as TA Nyfreler, M. & G. B,enz. 1978. Prey selection by the 22525. web spiders Argiope bruennichi (Scop.), Araneus quadratus Cl., and Agelena labgrinthica (Cl.) on fal- References Cited low land near Zurich, Switzerland. Rev. Suisse Zool. 85: 747-757. Bohmfalko G. T., R. E. Frisbieo W. L. Srerling, R. B. 1979. Studies on the ecological importance of spider Metzer & A. E. Knutson. 1983. Identification, bi- populations for the vegetation of cereal and rape ffelds ology and sampling of cotton insects. Texas Agricul- J. Appl. Entomol. 87: 348-376. tural Extension Service Bulletin B-933. Nyffeler, M., D, A. Dean & V. L. Srerling. 1986. Castillo, J. Ä. & W. G. Eberhard. 1983. Use of ar- Feeding habits of the spiders Cgclosa turbinata tificial webs to determine prey available to orb weav- (Walckenaer) and Lgcosa rabida W alckenaer. South- ing spiders. Ecology 64: 1655-1658. west. Entomol. ll: I95-20I. Chacon,P. &W.G. Eberhard. l9BO. Factorsaffecting l987a. Predation by the green lynx spider, Peucetia numbers and kinds of prey caught in artificial spider uiridans (Hentz) (Araneae: Oxyopidae), inhabiting webs, with consideration of how orb webs trap prey. cotton and woolly croton plants in east Texas. Envi- Bull. Br. Arachnol. Soc.5:29-38. ron. Entomol. 16: 355-359. Craig, C. L, 1987. The ecological and evolutionary f9B7b. Evaluation of the importance of the striped interdependence between web architecture and web lynx spider, Oxgopes salticus (Araneae: Oxyopidae), silks spun by orb web weaving spiders. Biol. J. Linn. as a predator in Texas cotton. Environ. Entomol. 16: Soc.30: 135-162. 1 I l4-l 123. Culin, J. D.&K.V.Yeargan. 1982. Feedingbehavior l987c. Feeding ecology of the orb-weaving spider and prey of Neoscona arabesca (Araneae: Araneidae) Argiope aurantia (Araneae: Araneidae) in a cotton and Tetragnatha Laboriosa (Araneae: Tetragnathi- agroecosystem. Entomophaga 32: 367 -375. dae) in soybean fields. Entomophaga 27:417-424. l9BBa. Prey records of the web-building spiders Dic- Dean, D. A. & W. L, Sterling. 1985, Size and phe- tAna Eegregata (Dictynidae), Theridion australe nology of ballooning spiders at two locations in east- (Theridiidae), Tidarren haemorrhoidala (Theridi- ern Texas. J. Arachnol. l3: 111-120. idae), and Frontinella pEramitela (Linyphiidae) in a 1987. Distribution and abundance patterns of spiders cotton agroecosystem. Southwest. Nat. 33: 215-218. inhabiting cotton in Texas. Texas Agricultural Ex- lgBBb. The southern black widow spider, Latrodectus periment Station Bulletin 8-1566. mactans (Araneae' Theridiidae), as a predator of the Dean, D. A., W. L. Sterling & N. V. Horner. 19A2. red imported fire ant, Solenopsis inoicta (Hyrrrenop- Spiders in eastern Texas cotton fields. J. Arachnol. 10: tera: Formicidae), in Texas cotton fields. J. Appl. 25r-260. Entomol. 106: 52-57. Dean, D. Ä., V. L. Sterling, M. Nyfreler & R. G. Breene. Olive, C. W. 198O. Foraging specializations in orb- 1987. Foraging by selected spider predators on the weaving spiders. Ecology 6l: 1133-II44. cotton fleahopper and other prey. Southwest. Ento- Pasquet, Ä. 199.4. Prcy and predatory strategies of mol. 12: 263-270. two orb-weaving spiders: Argiope bruenniclvi and Diericko E. J. 1961. An improved backpack motor Araneus marfiroreus. Entomol. Exp. Appl. 36: I77- fan for suction sampling of insect populations. J. Econ. 184. Entomol. 54: 394-395. Pieters, E. P. & V. L. Srerling. 1973. Comparison Documenta Geigy. 1968. Pp. 85-fffi, 109-123. In of sampling techniques for cotton arthropods. Texas Scientific tables, 7th ed. Basel (in German). Agricultural Experiment Station Miscellaneous Pub- Eberhard, V. G. 1971. The ecology of the web of lication 1.1.20. U loborus diaersus (Araneae, Uloboridae). Oecologia Risch, P. l9?7. Quantitative analysis of orb web pat- 6:328-342. terns in four species of spiders. Behav. Genet. 7: I99- Eisner, T., R. Alsop & P. Euershank. 1964. Adhe- 238. siveness of spider silk. Science 146: 1058-1061. Robinsono M. H. & B. Robinson. 197O. Prey caught Hartstack, A. W. & V. L. Sterling. f 988. The Texas by a sample population of the spider Argiope argen- cotton-insect model-TEXCIM. Texas Agricultural tata (Araneae: Araneidae) in Panama: a year's census Experiment Station Miscellaneous Publication 1646. data. J. Linn. Soc. Zool. 49: 345-358. Kagan, M. 1943. The Araneida found on cotton in Sabath, L. E. 1969. Color change and life history central Texas. Ann. Entomol. Soc. Am. 36:257-258. observations of the spider Cea heptagon (Araneae: Kajak, A. 1965. An analysis of food relations between Araneidae). Psyche 76: 367-374. the spiders Araneus cornutus and A. quadratus and Sunderland, K. D., A. M. Fraser & A. F. G. Dixon. their prey in meadows. Ekol. Pol. A 13: 717-768. f 986. Field and laboratory studies on money spiders LeSar, C. D. & J. D. Unzicker. 1978. Life history, (Linyphiidae) as predators of cereal aphids. J. Appl. habits, and prey preferences of Tetragnatha labo- Ecol.23: 433-447. riosa (Araneae: Tetragnathidae). Environ. Entomol. Turnbull, A. L. 1960. The prey of the spider Lingphia 7: 879-884. triangularis (Clerck) (Araneae, Linyphiidae). Can. J, Levio H. W. 1978. Orbweaving spiders and their webs. Zool. 38: 859-873. Am. Sci. 66:734-742. 1973. Ecology of the true spiders (Araneomorphae). Nentwigo W. 1983. The non-fflter function of orb webs Annu. Rev. Entomol. l8:305-348. in spiders. Oecologia 58:418-420. Uetz. G. W.. A. D. Johnson & D. W. Schemske. 1978. ENVIRONMENTAL ENTOMOLOGY Vol. 18, no. 3

Reed D. B. Peakall' 1968' A Web placement, web structure, and prey- capture in Witt, P. N.' C. F. & problems in regulatory biology Spring- orb-wäaving spiders. Bull. Br. Arachnol Soc' 4: l4l- spider's we-b: er, York. 148. New Vhitcomb' V. H.o H. Exline & R' C' Hunrer' 1963' -Spid".t'of 22 Decembet 1987; accepted the Arkansas cotton field' Ann Entomol' Receioed for publication 1988. Soc. Am. 56:653-660. 29 Juls