Btor,ocrcer, CoNrnor- Impact of the Striped Lynx (Araneae: Oxyopidae) and Other Natural Enemies on the Cotton Fleahopper (Hemiptera: Miridae) in Texas Cotton

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

Deparhnent of Entomology, Texas A&M University, College Station, TX 77843

Environ. Entomol. 2f(5): ff78-ff88 (1992) ABSTRACT Natural predation on nymphs and adults of the cotton feahopper, Pseud,ato- moscelis seriatus (Reuter), was assessed during 108 h of visual observation in an insecticide-free cotton ffeld in cenbal Texas. Predaceous of 13 species (from nine families) were observed to forage on the feahopper. More than 807o of üre predation events observed were afüibutable to . The striped lynx spider, Oxgopes salticus Hentz, was dominant among the predators observed eating fleahoppers (15 records of feeding in action). Cotton fleahoppers composed, -25Vo of the total prey of O. salticus during June and July. It was estimated during midseason that once every 4 d, one O. salticus would kill one cotton fleahopper. The assessment of the killing power of O. salticus, based on the predation rate and the predator-to-prey ratio (i.e., number of O. salticus individuals per feahopper), suggests that these spiders are important mortality agents of the cotton fleahopper (>l1Vo prey mortality per day in the middle of the growing season). Additional feahopper mortality is attributable to other predaceous arthropods such as oiridans (Hentz) (Oxyopidae), jumping spiders (Salticidae), crab spiders (), web-building spiders (Araneidae, , Theridiidae), damsel bugs (Nabidae), and ants (Formicidae).

KEY WORDS Pseudatomoscelis, Oxgopes, colton

THn cotroN FLEAHorer,r., Pseudatomoscelis fleahopper was monitored (three prey records seröatus (Reuter), is a major pest of cotton in over an 85-h observation period, or 0.03 record Texas (Adkisson 1973, Sterling et al. 1992b). per hour) (Table 6). Numbers of cotton fleahop- Fleahoppers are eaten by various polyphagous pers counted in that cotton field in 1985 was predators as has been detected by vi- 0.04-1.3 individuals per meter of row (early sea- sual observation (Whitcomb et al. 1963, Dean et son until bloom). This is below the economic al. 1987, Lockley & Young 1987) and by ""P- threshold of 15-35 fleahoppers per 100 plants labeling (Breene & Sterling 1988). These poly- (-LL3.5 individuals per meter of row in the phagous predators are numerous in some cotton Austonio field) recommended by the Texas Ag- fields (Whitcomb & Bell 1964, van den Bosch & ricultural Extension Service. The low predation Hagen 1966, Johnson et al. 1986, Dean & Ster- rates on fleahopper prey apparently reflected the ling 1987), suggesting that they may contribute reduced fleahopper numbers on cotton (Nyffeler to the natural mortality of the cotton fleahopper. et al. 1987a). Possibly the fleahoppers were kept An observational study (>85 h) was conducted in check by the numerous predators on the wild in an east Texas cotton agroecosystem during the host plants in the "reseryoir habitats" before summer of 1985 to evaluate quantitatively the they migrated into cotton (unpublished data). effect of arthropod predation on the population Nyffeler et al. (1987a, b) sbessed the need to dynamics of the cotton fleahopper (Nyffeler et al. repeat a similar visual observation project in an- 1986; 1987a, b, ci 1988a, b; 1989; Dean et al. other cotton agroecosystem where cotton flea- 1987). The study site was an insecticide-free cot- hoppers were more abundant. ton agroecosystem surrounded by extensive During the summer of 1988, the effect of ar- tracts of minimally disturbed noncrop habitats thropod predators on fleahopper numbers was composed of various wild plants and grasses. evaluated quantitatively in a cotton field in cen- From these "reservoir habitats," large numbers tral Texas, where cotton fleahoppers occurred in of predators (primarily spiders and ffre ants) mi- fairly high numbers (two per meter of row in grated into the cotton agroecosystem. Large midseason). Predation activities of insectivores numbers of predators were observed on cotton, on the various instars of the cotton fleahopper but a very low frequency of predation on the were observed, and the killing power of the nu-

0/Xü22l5XlS2llL78-f f88$02.00/0 @ 1992 Entomological Society of America This article is the copyright property of the Entomological Society of America and may not be used for any commercial !r othq p{ygte pqrpose without specific written permission of thg Entomological Society of Amedca October 1992 Nvppsrsn ET ÄL.: Spropn PnnoerroN oN CorroN Flnenoppnn I179 merically dominant predator species was esti- remains of dead fleahoppers from the spider mated. webs ("durable evidence" lDEl sensu Sterling [1989]). "Total evidence" is deffned as the com- bined data of"observational evidence" plus "du- Materials and Methods rable evidence" (OE + DE). Sfudy Area. The study site was a weed-free Predators in possession of prey were killed, cotton agroecosystem untreated with insecti- preserved (along with their prey) in TOVo ethyl cides in central Texas (Burleson County), -!Q alcohol, and later identified in the laboratory un- km southwest of College Station. This cotton der a dissecting microscope. At the same time, ffeld (f3.6 ha) was surrounded by grassland the age (instar) of each fleahopper prey was de- (grazed pastures), with wild plants growing on termined and recorded. For methodological de- the ffeld borders and in adjacent grasslands. Cot- tails see Nyfeler et al. (1987a, b, cj I988a, b; ton, sorghum, and corn were grown in nearby rese). ffelds. The cotton ('Paymaster 145') was planted Estimate of Predation Rate of O*gropes salticus on 8 April 1988 with a distance between rows of Hentz According to Edgar (1970) and Kiritani et 1 m. The cotton yield was 950 kg/ha (I.7 bales/ al. (1972), the predation rate of nonweb-building acre). spiders can be estimated based on the average Evidence of Predation. Field observations proportion of prey-carrying spiders observed in wk, from mid- were conducted for 9 consecutive the ffeld. It is necessary to know the average time June to mid-August 1988, during the daylight required to handle an individual prey (handling hours (the majority between 1200 and 1800 hours time) and the hunting time (hours per day), so CST). & Young (1987) noted that pred- Lockley that the data obtained in the field can be con- ator activity was higher in the morning hours verted into the number of prey eaten per day compared with the afternoon hours in Missis- (Edgar 1970, Kiritani et al. 1972). The predation sippi. In a previous sfudy conducted in Texas was in middle cotton, the feeding activity of the numerically rate in this study estimated the of dominant predators did not differ signiffcantly the growing season, when the O. salticus popu- between the morning and afternoon hours lation had a nearly uniform age-size class struc- stages (late instars- (Nyffeler et al. 1987a, b), although we cannot rule ture dominated by larger out that the unknown feeding activities of some adults, sensu Whitcomb & Eason [f967]). Young (1986) less abundant species may peak in the morning & Lockley conducted laboratory experi- small (see also Culin & Yeargan 1982). In total, 108 ments with O. salticus and found that (0.58 * person-hours ofvisual observation were spent in spiders 0.04 mm carapace width) killed signiffcantly less prey than medium-sized spi- the ffeld; 34 h in h in July, and 24 h in June,50 ders (0.81 -+ 0.07 mm carapace width) or large August, with an average of 3 h/d. The numbers of -t- predators were monitored by counting them on spiders (f .34 0.29 mm carapace width), whereas the two larger plants during l-h periods (walking speed -0.8 the difference between km/h along the field rows). During each observa- size categories was not statistically signiffcant. predation tion period, the following data were recorded: Evidently the difference of the rate (l) (2) (3) predators between the larger O. salticus size classes is Date, time of day, numbers of justifies without prey per observation hour, (4) numbers rather small, which the assessment of a group of predators with fleahopper prey per observa- single predation rate for the entire of growing tion hour, (5) numbers of predators with alternate larger O. salticus in the middle of the prey per observation hour, and (6) numbers of season. (Pd", potential fleahopper prey per observation hour. The daily rate of predation on all prey per Versatile predators (nonweb-building spiders number of prey organisms killed per spider insects) prey their - day) of O. salticus was assessed with equation I and with in (Edgar mandibulae were capfured by hand with a trans- 1970, Nyffeler et al. 1987a): parent cup (7.5 cm upper diameter, l0 cm depth). This method monitored "observational evidence Pd.: (T1x 60 x F)l(Th x 100), (l) of predation in action" (OE values [see Sterling 19891). One ffre ant worker transporting a wig- where 60 is minutes and 100 is used to convert to gling fleahopper was listed in the category of percentage, Tp is the hunting time (hours per "predators feeding," although the ant was not day) available'for prey capture and feeding in the actually seen eating; however, subsequent feed- field, 77" is the average time (minutes) required ing by the colony could be expected (Breene et to handle an individual prey, and F, is the aver- al. 1989b). age feeding frequency at a given time (mean For sedentary web-building spiders, evidence percentage prey-carrying spiders observed in the of predation was obtained in two ways: (l) by ffeld [see Edgar 1970J). The proportion ofprey- capfuring spiders with prey in their chelicerae carrying O. saltöcüs within the population was (observational evidenee of predation in action recorded on 4 consecutive d (20-23 July, with [OE], see above), and (2) by collecting the 3-h observations per day) and the mean (*Sn) of ll80 ENvlnoNNrsNrer, ENtouolocY Vol. 21, no. 5 the four observation periods used as an estimate principal groups of foragers (web-building spi- for the F, value. Hunting (finding) time (T,") and ders versus nonweb-building spiders), because handling time (Tr,) of O. salticus in Texas öotton the number of observed predation events was too had previously been determined in a field study low for a meaningful between-species statistical by Nyffeler et al. (1987a). As a polyphagous pred- analysis. A f test of independence was used to ator, O. saltöcus feeds on multiple prey species examine whether the immature/adult feahopper (Nyffeler et al. 1987a, 1992), and the Pdo value ratio in the prey differed significantly between expresses the rate ofpredation on all prey (flea- the two forager groups. hopper prey plus alternate prey). The feahopper The same statistical test was used to determine prey/all prey ratio was estimated based on field whether the ratio of fleahoppers/alternate prey observations (Table 3), and used as a correction consumed by the predator complex (monthly factor to convert the rate ofpredation on all prey pooled data for combined predator species) dif- (Pd" vahrc) to the rate of predation on fleahop- fered signiffcantly between months, indicating pers (Pd" value). possible temporal shifts of the pred_ator activities Estimate of Fleahopper Mortality Caused by (see also Breene et al. 1989a). A f test of inde- O, salticus. Based on the predation rate by O. pendence was also applied to compare feeding saltöcus upon fleahopper prey and on the preda- frequencies on fleahopper prey (predators eating tor/prey ratio (i.e., number of O. saltöcu.s individ- a fleahopper per total predators, monthly pooled uals per fleahopper), the daily percentage mor- data) between months, which provides informa- tality (M) of the cotton feahopper caused by O. tion on the seasonal dynamics of the predator saltöcus was estimated. The M value, in the mid- activities. f values were computed by means of dle of the growing season, was estimated with the uncorrected formula (without Yates' correc- the following equation: tion) (Sokal & Rohlf 1969). M: 100 x Pdcx R, (2) Results and Discussion where Pd" is the predation rate on cotton flea- (number per spider Predator Determination and Efficiency. Over- hoppers of fleahoppers killed pre- the predator/prey ratio all, 3,981 spiders (and numerous uncounted per day), and R is by visual ob- (number of O. salticus individuals per fleahop- daceous insects) were encountered August in cotton (Table f per). In this study, the R values were based on servation from June to ). assemblage (Table represents a estimates (relative and ab- The spider l) two different density cotton ffelds, Relative popula- species complex typical for Texas solute densities, respectively). (Oxyopidae) predominating densities (individuals observed per hour) with lynx spiders tion (compare Dean et al. 1982; Dean & Sterling were counted from 1300 to 1400 hours on 14 and 1987; Nyffeler et al. 1987a, b). Spider numbers 25 Absolute population densities (individ- July. increased with time (Fig. 1). The phenology of uals per meter of row) were assessed by whole- predators correlated with the fruiting rate of plant examination; twenty-ffve random samples is the cotton plant (Dean & Sterling 1992). In the each of I m of row were obtained between 1100 middle of the growing season, the spider density 1200 hours on l9 and July. at 2.84 -F 0.39 individuals per The predator/prey ratio was assessed as fol- was estimated meter (mean + SE, whole-plant examina- lows for the relative estimate: square tion on 19 July). Rr: Sr/Cr, (3) During this study, a total of 97 cases of arthro- pod predation upon the cotton fleahopper was where S. is the average number of striped lynx documented (total evidence, Table l). The age spiders recorded per hour, and C, is the average structure of fleahoppers killed by predators (Ta- number ofcotton fleahoppers recorded per hour. ble 1) was: third instar (2 records l2Vol), frfth The predator/prey ratio based on absolute den- instar (3 records [37o]), unidentified instar (l sities was computed as follows: record [17o]), and adults (91 records [9a%o]). However, these observations are biased by the Ro: SolCo, @) fact that small feahoppers are not easily ob- where So is the average number of striped lynx served on the plant. If caphrred, they are likely spiders per meter of row, and Co is the average consumed rapidly (low handling time) so are less number of cotton fleahoppers per meter of row. likely to be observed as prey (see Edgar [f970] Statistical Tests. A statistical comparison of the for an analysis of handling time as a function of ratio of immature/adult fleahopper prey con- prey size). Because they do not fy, immatures sumed by the various predator species provides ire less likely to be observed in spider webs (see information that can be used to adjust the pred- below). Therefore, other experimental methods ator group-speciffc indices of efficiency used in may be needed to assess the predation rates of the tritrophic cotton insect TEXCIM model (see predators on small immafure fleahoppers accu- Breene et al. 1989a, Nyffeler et al. 1989, Sterling iately. Web-building spiders, which are "sit-and- et al. 1992b). The data were pooled into two wait" foragers, intercepted with their webs pre- October 1992 Nvrrrr-rn ET AL.: Sproen Pnroatrort oN Cor"roN Flpnnoppsn 118l

Table l. Obeerations of arthropod predation on the cotton fleahopper in r cotton field near College Station, TX, during aummer 1988 (lOB h total obgenarion time)

No. feahopper prey' Total no. Predator Predator species predators Predator Prey in Total stage" observed feeding web evidence (oE) (DE) (OE + DE) Araneae Oxyopidae Orgopes salticus lJenu Imrn, ad 2,402 15 r'f Peucetia oirtdans (Henu) Imm 626 4 4e Salticidae Phidöppus aud.ax (IIenE) Imm I ls Metaphidippus galathea (Walckenaer) I lh Unidentiffed 1- 2r3. 0 0 Thomisidae Misumenops spp, Imm 228 I 18 Araneidae C g clo s a turbinat a (Walckenaer) Imm, ad r45 I ll lze Argäope aurantiaLtc s Imm 3l 0 I ls N eo s co na arabe s c a (Walckenaer) 206c 0 2 28 Unidentiffed l' 0 I ls Dictynidae Dictgna segregata Gertsch & Mulaik Imm 90 I 53 548 Theridiidae küro dc ctu s mact an s (F.) Ad 2 0 I Le Unidentiffed spiders 38 0 I li Hemiptera Nabidae Reduoiolus alte rnatus (Parshley) Ad _d Is Unidentiffed (Reduviidae ?) Ad _d Ie Hymenoptera Fomicidae S ole nop sis inrsict a Bur en _d I 18 Total spiders I 3,981 2/t 70 94 Total insects _d J

"Only predator individuals in possession offleahopper prey. Imm, immature; ad, adult. b OE, obseruational evidence ofpredation in action; DE, durable evidence (predator not feeding); OE + DE, total evidence (observational plus durable evidence). " All Salticidae; N. arabesca and unidentiffed Araneidae. d not counted. " These-, predators do not make a web; therefore, no prey can be found in webs (indicated by a dash). tThird instar, 2 fffth instar, unidentiffed instar, 1l adults. s Adults only. h Third instar. i Fifth instar. ponderantly mobile winged adults of the 1987; Nytreler et al. 1987c, 1989; Breene et al. fleahopper (71 adults versus I nymph) (Table 1). 1988, r989b). A significantlv (f : 12.84, df : l, P < 0.001) Total evidence (OE + DE) presented in Table lower proportion of adult fleahopper prey (17 I shows a predominance of web-building spider adults versus 5 nymphs) (Table 1) was captured prey records (mostly Döctgna segregata Gertsch by the nonweb-building spiders which actively & Mulaik and Caclosa turbinata (Walckenaer)), search the plant surface for nymphs and adults of which is deceptive because web-building spi- the feahopper (exception: crab spiders tend ders tend to store prey in their webs for longer toward a "sit-and-wait" foraging strategy). Evi- time periods (up to several days, "durable evi- dence ofpredation on fleahoppers was obtained dence" (DE) sensu Sterling [f989]), whereas a on spiders of l0 different species (six families) prey organism remains in possession of a non- and 3 insect species (three families) (Table 1). web-building spider for only a short time period Among the spider predators ranging from 1.2 to (To <1 h, O. saltöctrs) (Nyffeler et al. 1987a; 7.4mm in length were 5 species each of nonweb- M.N., unpublished data); whereupon the evi- building spiders (Oxyopidae, Salticidae, and dence is destroyed. Total evidence (OE + DE) Thomisidae) and web-building spiders (Aranei- for web-building spiders versus nonweb- dae, Dictynidae, and Theridiidae) (Table 1). building spiders (long versus short retention With the exception of the black widow spider, time), therefore, cannot be compared quantita- Latrod.ectus nxact&ns (F.), all arthropod preda- tively, Feeding times, however, can be com- tors listed in Table I have been reported to be pared because it takes web-building spiders, predaceous on the cotton fleahopper (Dean et al. nonweb-building spiders, and predaceous in- 1182 ENvrnoNunNter- ENrolrolocY Vol. 21, no. 5

60 dation frequency. The predator groups that dem- onstrated a higher predation frequency were the 50 jumping spiders (Salticidae, I.37o feeding indi- viduals) and lynx spiders (Oxyopidae, 0.9Vo). o!+o E Predator groups with lower predation frequency E30 (Thomisidae) o were the crab spiders and web- p building spiders (Araneidae, Dictynidae, and ,h4 20 Theridiidae, combined 0.5Vo). L very low preda- tion frequency is attributable to the red imported 10 ffre ant (Solenopsös inoöcta Buren, <0.1%). No 0 predation on fleahoppers (UVo) was observed for JUNE JULY AUGUST big-eyed bugs (Ceocoris punctöpes (Say)), plant (Lgg1.ts (Hippodamia DATE bugs spp.), lady beetles conoergens (Guerin-Meneville) and other Coc- Fig. l. Seasonal variation in the numbers of spiders cinellidae), or lacewings (Chrysopidae) (for a de- (visual encountered per hour observations) in a cotton scription ofthese predator groups see Sterling et ffeld near College Station, TX. H, O. salticus; U, other case nonweb-building spiders (including Pezcetia oiridans, al. f992b). In the of the damselbtgs,Redu- Salticidae, Thomisidae, Lycosidae, Philodromidae, oöolus alternatns (Parshley) and other preda- and others); l, web-building spiders (including Ara- ceous Hemiptera, no value could be computed. neidae, Dictynidae, Theridiidae, Tehagnathidae, Ul- These values (100 x OEIA) were converted oboridae, and others). Monthly pooled data collected into a standardized value ("efficiency index" during daylight hours, June-August 1988. [100 x OE]1U.3 x Al), ranging between 0.0 and 1.0. The predator group with the highest preda- sects each a short time to consume small-sized tion frequency (jumping spiders), considered to prey such as fleahoppers. Observational evi- be the most efficient natural enemy, is weighted dence of predation in action (based on feeding with an efficiency index of l. Other predators are records, OE-values from Table 1) provides a less compared in efficiency with the jumping spiders biased quantitative comparison of the various (concept according to Sterling et al. [1989]). The predators (Table 5). standardized values obtained in our study (Table Based on the total numbers of observed pred- 2) agree fairly well with efficiency indices previ- ators (A) and the number of predators found feed- ously used by the Texas Cotton Insect Model ing on fleahoppers (OE), the percentage individ- (TEXCIM; see Breene et al. 1989a). uals within a particular predator group feeding Percentage Fleahoppers in Predators' Diet. on fleahoppers (100 x OEIA) was computed, and Approximately 20Vo (n : f3a) of the overall diet the values for the various predator groups were of combined predators was composed of fleahop- compared (Table 2, pooled data for June and pers (Table 3). The proportion of fleahoppers in July). These values provide an estimate for pre- the diet of combined predators did not differ

Table 2. Frequency of predation on cotton fleahoppen obeered in a cotton field near College Station, TX (data pooled forJune and July 1988)

No' p,redators 7o predators Total no. standardized value leedrnq on leeorng on Predator taxon Dredators "f,l1l,X:ilor '.läEi"^ ,l;YU,:, ,lo6T"ää;iiäu:i]' Striped lynx (O. saltöcus) 1,645 l5 0.9 o.7 Green lynx (P. oiridans) 460 4 0.9 o.7 Jumping spiders (Salticidae) r57 2 1.3 1.0 Crab spiders (Thomisidae) 201 I 0.5 0.4 Web-building spiders (Araneidae, Dictynidae, Theridiidae) 398 2 0.5 0.4 Fire ants (5. inoicta) I <0.1 <0.1 Damsel bugs (R. alternatus) and unidentiffed bugs (Reduviidae)b 2 (0.8)d (0.6)d Big-eyed bugs (G. punctöpes)b 0 0.0 0.0 Plant bugs (Lygas spp.)b 0 0.0 0.0 Lady beetles (Coccinellidae)b 0 0.0 0.0 Lacewings (Chrysopidae)b 0 0.0 0.0

" Standardized value as an estimate of "predator efficiency", ranging between 0.0 and 1.0. The predator group with highest predation frequency (umping spiders), considered to be the most efficient natural enemy ofthe cotton fleahopper, is weighted with an efficiency index of l. Other predators are compared with jumping spiders. a For a description ofthese predator groups, see Sterling et al. (1992b), c not counted. d -,Empirical estimate. October 1992 Nvppnr-rn ET AL.: Sprprn PnnolrroN oN CorroN Fr,rlgopprn 1183

Table 3. Cotton fleahoppers in diet ofpolyphagoue an that time composed of individuals with similar thropod predatom, obeered in a cotton field near College energy requirements, which justifies the evalua- Station, TX, during eumer 1988 tion of a single predation rate for this entire No. No. group of larger spiders (see Methods). predators predators The predation rate (Pd") of O. saltöcus was Predator taxon Month observed feeding on estimated with equation l, using the following Geding" feahopper values: Ft : 3.0 + 0.61 (f + SE of four samples, O. salticus fune 24 6 n :74, n :77, n :'77, n : IO2 observed spi- July 3l I ders), Tr: 24 (based on Nyffeler et al. [1987a]), Aog. 8 0 and T6 : 49 (mean value for penultimate-adult P. oirid,ans June 13 Joly 15 I O. salticus [Nyffeler et al. 1987a]). Because the Aug. 3 0 handling time is a function of the spiders' prey Salticidae June 6 I size (Edgar f970), a low SE of mean prey length July 7 I (2.72 t 0.36 mm, : 10) observed during mid- Aug. 5 0 n Thomisidae June 2 0 season justifies the use of a single average T6 Jrly 8 I value for the entire group of larger O. salticus. Aug. 0 0 o On this basis, we estimated that an O. salticus Web-building spiders June 6 spider captured, in the middle of the cotton- Jrly I 0 Arg. 0 0 growing season, an average of approximately one Predaceous insects June 3 2 prey daily (Pd":0.9). This estimate is slightly July 2 I lower than the daily predation rate of O. saltöcus Arg. 0 0 assessed in another Texas cotton ffeld (Nyffeler Combined total June 54 14 (25.9Vo)ab Jrly U t3 (2o.3/o)b et al. I987a) and in laboratory feeding tests Aug. 16 0 (0.0%)ab (Guillebeau & All 1989), where the larger stages Total 134 27 (2O.lVo) of this spider captured on the average a little more than one prey per day. " Fleahopper prey plus altemate prey. b Percent fleahoppers in diet of combined predators. Values Because O. salticus is a polyphagous feeder followed by the same letter are signiffcantly different (P < 0.05, (Nyffeler et al. 1987a), the obtained Pdo value f test of independence). expresses the rate of predation on multiple prey species (see above). About every fourth prey or- ganism captured by O. salticus was a cotton flea- signiffcantly (f : 0.52, df : 1, P > 0.05) be- hopper (see above), which suggests that one flea- tween June and July (26 versus 2OVo) (Table 3) hopper may have been killed per spider about but declined significantly (f : 3.88, df : l, P < every 4 d in the middle of the growing season 0.05) from July to August (20 versus IVo) (Table (Pd. : 0.25). This is a rough estimate that ap- 3). pears to be rather conservative compared with In June, we found 24 prey-carrying O. salticus the Pd." values for O. saltöcus evaluated in other spiders with a fleahopper prey/all prey ratio of sfudies. Ten to thirteen times higher mean Pd. t:4 (Table 3). In July, we collected 31 prey- values estimated for O. salticus were reported by carrying O. salticus spiders, with a correspond- Breene et al. (1989a, 1990). ing ratio of approximately l:3.5 (Table 3). This The average population density of O. salticus, implies that overall, approximately one in four in the middle of the growing season, was 1.48 + prey captured by O. salticus was a fleahopper 0.24 individuals per square meter (So value for during June and July. 19 July) in the investigated cotton field, which Mussett et al. (1979) obtained a correlation (r : implies that one fleahopper may have been 0.62) between the abundance of combined pred- killed per square meter about every 3 d (Pd" x S ators and cotton fleahoppers. Whitcomb & Bell : o.25 x 1.48 : 0.37). " (1964) and Mussett et al. (1979) suggested that Killing Power of Dominant Predator, O. saltö- fleahoppers are among the cotton arthropods cus. The killing power of the dominant predator serving as a food source which help maintain the species, O, salticus,was evaluated quantitatively abundance of polyphagous predators. The high based on estimates of predation rate, spider den- percentage offleahoppers in the diet ofpolyph- sity, and fleahopper density. agous predators observed in this study (Table 3) The daily percentage mortality (M) of the cot- supports Whitcomb & Bell's hypothesis. ton fleahopper caused by the most abundant spi- Estimate of Predation Rate of O. salüdcus. This der species (O. saltöcus) in the middle of the estimate was conducted in the middle of the cot- growing season was assessed with equation 2. In ton-growing season when the O. saltöcus popu- the relative estimate, values used were S, : lation had a nearly uniform age-size class struc- 29.50 and C, : 48.50 (mean value of 14 and 25 ture dominated by larger stages (mean body July) (Table 4), which resulted in R, : 0.61. In length 4.34 'r 0.23 mm, n : 18, on 20 July); the absolute estimate, values used were So : low-SEM carapace width (1.55 -F 0.08 mm, n = 1.48 + 0.24 and. C" : 2.04 + 0.36 (data for 19 18) implies that the O. saltöcus population was at July), which resulted in Ro : 0.72. Assumingthat Il84 ENvrnoNunNTAL ENToMoLocY Vol.2l, no. 5

Tabte 4. Numben of cotton fleahoppem and striped 0.5 lynx epidere counted per hour on cotton foliage in a field 1988 near College Station, TX, during sumer 0.4 o: Predator/ No. individuals/h" prey ratio ä 0.3 n.: (s"y o (c,) (s,) (c.) o.z Et 6 4 July 92.9b 32.6 0.35 o 7 84.5b 31.0 0.37 tr July 0.1 I July 76.0. 30.4 0.40 14 July 55.0" 27.8 0.51 25 July 42.O. 31.2 o.74 3.60 26 July 9.0d 32.0 JULY AUGUST 4 Aug. 5.3d 34.0 6.42 DATE Records made between 1200 and 1500 hours CST. Seasonal variation in the frequency ofpre- " C., cotton fleahoppers; S,, striped lynx spiders, Abundance Fig.2. of striped lynx spider apparendy not changing with time. dation on the cotton feahopper (measured as number Based on larger data (June to August), however, a visible of feahopper prey records per observation hour) in a change in the abundance of O. salticus with the progressing cotton ffeld near College Station, TX. H, O. salticus; Z, season was found (Fig. l). other nonweb-building spiders (including Peucetia b All adult. oirid.ans, Salticidae, and Thomisidae); l, web-building " >90% adult. spiders (including Araneidae, Dictynidae, and Theridi- d <90vo adt:Jt. idae); ffi, insects (including Solenopsis inoicta and pre- daceous Hemiptera). Monthly pooled data collected during daylight hours, JuneAugust 1988. Pd,": 9.25 is an accurate predation estimate (see above), the daily mortality was computed to be M. : lSVo per day (relative estimate) and Mo : further declined significantly (f : 6.6f, df : 1, lSVo per day (absolute estimate), respectively. P < 0.01) from July to August (0 in 1,094). The two estimated values are of similar mag- Relative Importance of Various Predator nitude (only SVo difference). The other preda- Groups (Observational Versus Experimental Ev- ceous arthropods such as t}re green lynx spider idence). In another Texas cotton agroecosystem, Peucetöa oirödans (Hentz), jumping spiders, Breene et al. (1989b) conducted a ffeld experi- crab spiders, web-building spiders, predaceous ment by releasing :30,000 fleahopper nymphs Hemiptera, and red imported ffre ants contrib- labeled with 32P and thereafter recovered radio- uted additional mortality (Tables l-3) (see also active predators. Although this experimental de- Breene et al. 1989a, b). sign has the limitation that it cannot distinguish Because data for predators and prey were lim- primary from secondary predation (Breene & ited, fleahopper mortality could not be quantita- Sterling 1988), it has the advantage that evidence tively assessed except for the middle part of the is based on much larger samples compared with growing season. A comparison of predation rec- the very time-consuming visual observation per (number feahopper prey method (n = 282 versus n:24 for spiders) (Ta- ords hour of 32P counted per hour, monthly pooled data) in dif- ble 5). In addition to this, the method is ferent months (Fig. 2) suggests a declining trend advantageous by measuring the combined activ- offleahopper predation by the predator complex ity of diurnal and nocturnal predation. The ob- with the progressing season (decrease of -4OVo servational data of our project (OE values in Ta- from June [-0.5 record per hour] to July [-Q.3 ble 1) and Breene's data are comparable because per hourl, down to zero in August) (Fig. 2). Ob- both studies were conducted in insecticide-free served predation on fleahoppers by O. salticus cotton fields in the same geographic area (near alone, however, did not differ visibly between College Station, TX). Table 5 compares the rel- June and July (-9.2 record per hour). A low pre- ative importance of various predators between dation rate of O. saltöcus on fleahoppers was the two studies. The comparison reveals that the monitored in August (Table 3; Fig. 2) after the results of the present observational sfudy are decline of fleahopper numbers in late July (Ta- strongly supported by Breene's experimental ble 4). work. Another way of examining the seasonal In our study, 897o of the predators found feed- dynamics of predation on the cotton fleahopper ing on fleahopper prey (OE) were spiders, and is given by comparing feeding frequencies 97Vo of all predation events recorded (total evi- (predators eating a fleahopper per total preda- dence, OE + DE) were attributable to spiders tors, monthly pooled data) between months. The (Table l). This is basically conffrmed by the feeding frequency on feahopper prey by the spi- work of Breene et al. (1989b) and the observa- der complei decreased significantly (f = 4.12, tions of Reinhard (1926), who also concluded df: 1, P < 0.05) from June (12 in 893 spiders that spiders are superiör as predators compared eating a feahopper) to July (12 in 1,994) and with the predaceous insects. In our study, the October 1992 Nvrrnr-nn ET AL.: Sprorn PnrlerroN oN CorroN Fr-negoppnn rl85

Tsble 5. Relative importance of varioua spider tm aa predaton of cotton fleahopper; comparison of two methode

Observational evidence Spiders feeding on Diflerence fleahoppef (A-B) No. events lo Total No, events 4o Total recorded (Ä) recorded (B) O. salticus l5 62.5 163 57.8 4.7 P. airid,ans 4 16.7 5 1.8 14.9 salticidae 2 8.3 81 28.7 -20.4 Thomisidae I 4.2 6 2,7 2,7 Other nonweb-building spiders 0 0.0 8 2.8 -2.8 Web-building spiders 2 8.3 l9 6.7 1.6 Total 2,4, lo0 282 100

Both studies conducted in insecticide-free cotton ffelds near College Station. " Numbers ofspiders feeding on fleahoppers observed in a cotton field during summer 1988 (OE-values from Table l). ö Numbers of radiolabeled spiders recovered with a D-Vac following release of radiolabeled fleahoppers in a cotton ffeld (summers 1986-1987, data from Breene et al. 1989b). Predation evidence based on assumption that predators became radioactive while feeding on radiolabeled fleahoppers (Breene & Sterling 1988). only insects predaceous on the fleahopper were Breene indicate that under the conditions of two individuals of Hemiptera and one individual these studies, web-building spiders are of less red imported ffre ant (Table l). Breene et al. importance than Oxgopes as predators of the flea- (1988, 1989b) provided experimental evidence hopper. for red imported ffre ants feeding on the fleahop- In our study (OE values) and in the experimen- per. They pointed out that the rate and extent of tal work by Breene et al. (1989b), -907o of the such ant predation cannot be reliably estimated monitored spider predation on fleahoppers was using 32P. In the course of this project, during attributable to the nonweb-building spiders (Ta- 108 h of direct observation in the field, we wit- ble 5). A higher relative frequency ofpredation nessed only one case of a red imported ffre ant events attributable to jumping spiders was mon- (minor worker) carrying a wiggling fleahopper itored in Breene's study compared with the (Table l). In other cases (not listed in Table 1), present project. The apparent difference ob- red imported ffre ant workers were transporting served inP. oöridans is based on a low number of dried-out (unnatural coloration) fleahopper car- predation records in both studies. O. salticus was casses, which suggests a scavenging foraging be- the dominant predator of the fleahopper in both havior. Thus, ant feeding traced by Breene et al. studies (more than half of the predation events (1988, 1989b) using 32P may consist ofboth scav- recorded in Table 5). As in our study, high flea- enging and predation. hopper mortality caused by lynx spider preda- Among the spider predators found feeding on tion was also monitored by Breene et al. (1989b), fleahoppers, llUVo were web-building spiders indicating a high killing power of these preda- (two feeding records, OE-values in Tables 1 and tors. 5), which agrees with Breene's results where, The present observational project is based on likewise, a low proportion (<10%) of monitored the data of I yr (1988) only. Breene's project predation activity was attributable to web- (f986-f987), however, was conducted in the building spiders (Table 5). The flight paths of the same geographic area in an insecticide-free ffeld; fl eahoppers spatially-temporarily coincide with hence, the two projects complement each other, the web positions of the spiders (M. N. & W.L.S., providing combined data over a continuous 3-yr unpublished data). Based on the fairly large period (f986-f988). The similarity of the preda- numbers of winged fleahoppers observed in the tion patterns observed in the two projects (Table ffeld during the first half of the growing season 5) provides shong mutual support for their accu- (Table 4), one may expect frequent capture of racy. these insects in spider webs. However,

Table 6. Prey records obtained per hour; comparieon between two ineecticide-free cotton agroecosyttema in Texae baeed on total evidence data (predatore obsened feeding plue fleahopper carcaaaes found in webs)

Central Texas (summer 1988)" East Texas (summer 1985)b Predator taxon Total no. Total no. No. records/ No. recordsÄr" fleahopper prey fleahopper prey hd Nonweb-building spiders oo 0.20 0 0.00 Web-building spiders 0.67 0.03 Predaceous insects ü 0.03 0 0.00 Combined total 97 0.90 0.03

" Present paper. b Based on Nyffeler et al. (1986; 1987a, b, c; 1988a, b; f989), Dean et al. (f987). " Total number of fleahopper prey divided by 108 h. d Total number of fleahopper prey divided by 85 h. 'Fleahopper carcasses found in webs, but spiders not obserued feeding (durable evidence) based on Nyffeler et al. (1987c, le89). insects of a neutral economic stafus, and preda- nalism reported by Nyffeler et al. [1987a]). ceous arthropods as well. High levels of "intra- Thus, this spider is a "time generalist," guild predation" (sensu Polis et al. f989) by lynx which increases the probability of encoun- spiders were recorded in Texas cotton ffelds tering fleahopper prey. (Nyfieler et al. 1987a, b; 1992); the overall eco- (5) These spiders readily feed on the various logical and economic implications of this phe- stages of the fleahopper (Table l) and exhibit nomenon, however, are not yet known. a sigmoid functional response to fleahopper Orgopes saltöcus is predaceous on various cot- availability (Breene et al. 1990). ton insect pests (Young & Lockley 1985, Nyffeler et al. 1990). Although no experimental evidence The high values of fleahopper mortality esti- "irreplaceable (sensu for mortality" Sterling et mated in our sfudy and in that of Breene et al. fleahoppers caused salticus al. 1989) of by O. (1989b) provide evidence that these spiders con- exists currently, these spiders show several char- tribute to fleahopper mortality in Texas cotton. acteristics suggesting that they are major preda- these spiders as mortality tors of fleahoppers in the Texas cotton agroeco- The contribution of system: agents, however, varies between the different fields and within different years because of the (f) They have good dispersal capabilities (Dean spatial and temporal fluctuations of the abun- & Sterling f985, 1990) and appear to be ex- dance patterns ofspiders and fleahoppers (Dean cellent colonizers well adapted for survival & Sterling 1987, Breene et al. 1989a). We re- (foraging and reproducing) in the cotton corded times higher frequency of predation (Dean 1987, -30 agroecosystem & Sterling on fleahoppers compared with another Texas cot- Nyffeler et al. 1987a). Therefore, they colo- ton ffeld (0.90 versus 0.03 prey record per hour) nize cotton ffelds in high abundance relative (Table 6). Consequently, the economic benefit other predators (Table 1) (Johnson et al. to attributable to these predators varies in different 1986, Dean & Sterling 1987, Nyffeler et al. sifuations. f987a). Because these spiders can build up TEXCIMSO model (Sterling et al. large numbers, they may sometimes become With the lynx spiders, other more abundant than their feahopper prey 1992b), the economic value of predaceous the (Table 4) (Breene et al. 1989a). Because of spiders, ffre ants, and bugs in their polyphagous feeding behavior, these control of cotton fleahoppers can be forecast for pred- spiders can survive in a ffeld with low flea- each ffeld. The value ofspiders and other preda- hopper numbers (Nyffeler et al. 1987a). ators depends on many variables such as (2) They are among the first predators arriving in tor density, cotton fleahopper density, weather, spring in the cotton fields (Nyffeler et al. insecticides, crop value, other herbivores, crop 1987a). Even the smaller immature O. saltö- growth, etc. TEXCIMS0 takes these and many cus (<3 mm long) are already capable of other factors into consideration in forecasting the overpowering fleahoppers (Nyffeler et al. value of spiders. During a S-yr study, the value of ree2). all predators of cotton feahoppers ranged from (3) They forage for prey throughout the entire $2.12 to $38.30 per ha (Sterling et al. 1992a). cotton plant, from the top to the ground and Few quantitative evaluations ofthe predation even under leaves, which enables them to effect of spiders have been published (review in detect fleahoppers hiding in refuges (Whit- Nyffeler & Benz 1987, f989). The mortality esti- comb et al. 1963; Dean et al. 1982; M. N., mates presented here suggest that nonweb- unpublished data). building spiders can exert predation pressure on (4) Thev forage for prey day and night (noctur- herbivores, which agrees with the quantitative October 1992 Nyrrpr-rn ET AL.: Sprorn Pnroerrou oN CorroN Fr-n.a.nopppn 1187 evaluations by Van Hook (1971) and Kiritani et Guillebeau, L. P. & J. N. All. 1989. Geocoris spp. al. (1972) in other habitats. (Hemiptera: Lygaeidae) and the striped lynx spider (Araneae: Oxyopidae): cross predation and prey preferences. J. Econ. Entomol. S2: 1106-1110. Acknowledgments Johnson, S. J., H. N. Pitre, ]. E. Powell & W. L. Ster- ling. 1986. Control of Heliothis spp. by conser- We thank farmer Vajdak for permission to carry out vation and importation of natural enemies, pp. 132- this project in his cotton ffeld, and |. D. Lopez, Jr., and 154. In. S. E. G. King & R. Bradley, (USDA, J. Johnson, f. Jr. M. A. Latheef College Station) for information [eds.], Theory and tactics of Heliothis population about the ffeld. We also thank R. G. Breene and T. C. management: cultural and biological conhol. So. Lockley for their review of this paper. Support has Coop. Ser. Bull.316. been provided by the Expanded Research Project Kiritani, K., S. Kawahara, T. Sasaba & F. Nakasuji. H-6903-2100 of the Texas Agricultural Experiment Sta- 1972, evaluation ofpredation spi- publication Quantitative by tion. Approved for as TA 30268 by the ders on the green rice leafhopper, Nephotettir Director, Texas Agricultural Experiment Station. cincticeps Uhler, by a sight-count method. Res. Popul. Ecol. 13: 187-200. Lockley, T. C. & O. P. Young. 1987. Prey of the References Cited striped lynx spider Oxgopes salticus (Araneae, Adkisson, P. L. 1973. The integrated control ofthe Oxyopidae), on cotton in the Delta area of Missis- insect pests of cotton. Proc. Tall Timbers Conf. sippi. I. Arachnol. 14: 395-397. Ecol. Anim. Control Habitat Manage. 4: l7F-188. Mussett, K. S,, J. H. Young, R. G. Price & R. D. Mor- Breene, R. G. & W. L. Sterling. 1988. Quantitative rison. 1979. Predatory arthropods and their rela- phosphorus-32 labeling method for analysis of tionship to fleahoppers on llaliofhis-resistant cotton predators of the cotton feahopper (Hemiptera: varieties in southwestern Oklahoma. Southwest. Miridae). J. Econ. Entomol. 8l: 1494-1498. Entomol. 4: 35-39. Breene, R. G., W. L, Sterling & D. A. Dean. f988, Nyfreler, M. & G. Benz. 1981. Some observations Spider and ant predators ofthe cotton feahopper on on the escape of adult Lepidoptera from webs of woolly croton. Southwest. Entomol. 13: 177-183. orb-weaving spiders. Anz. Schaedlingskd. Pflan- Breene, R. G., A. W. Hartstack, W. L. Sterling & M. zenschutz Umweltschutz 54: ll&-ll4 (in German). Nyffeler. 1989a, Natural conhol of the cotton 1987. Spiders in natural pest control: a review. J. fleahopper, Pseudatomoscelis seriatus (Reuter) Appl. Entomol. 103: 321-339. (Hemiptera, Miridae), in Texas. J. Appl. Entomol. 1989. Foraging ecology and predatory importance 108:298-305. of a guild of orb-weaving spiders in a grassland Breene, R. G., W. L. Sterling & D. A. Dean. 1989b. habitat. J. Appl. Entomol. 107: 166-184. Predators ofthe cotton fleahopper on cotton. South- Nyffeler, M., D. A. Dean & W. L, Sterling. f986. west. Entomol. 14: 159-166. Feeding habits of the spiders Cgclosa turbinata Breene, R. G., W. L. Sterling & M. Nyfreler. 1990. (Walckenaer) and Lgcosa rabida Walckenaer. Efficacy of spider and ant predators of the cotton Southwest. Entomol. 1l: l9A-201. fleahopper [Hemiptera: Miridae]. Entomophaga 35: 1987a. Evaluation of the importance of the striped 39L40t. lynx spider, Oxgopes salticus (Araneae: Oxy- Culin, J. D. & K. V. Yeargan. 1982. Feeding opidae), as a predator in Texas cotton. Environ. behavior and prey ofNeoscona arabesca (Araneae: Entomol. 16: 1114-1123. Araneidae) and Tetragnatha laboriosa (Araneae: f987b. Predation by the green lynx spider, Peucetia Tetragnathidae) in soybean ffelds. Entomophaga airidans (Hentz) (Araneae: Oxyopidae), inhabiting 27:417-424. cotton and woolly croton plants in east Texas. Dean, D. A. & W. L. Sterling. f985. Size and phe- Environ. Entomol. 16: 35L359. nology of ballooning spiders at two locations in 1987c. Feeding ecology of the orb-weaving spider eastern Texas. J. Arachnol. 13: lll-120. Argiope aurantia (Araneae: Araneidae) in a cotton f987. Distribution and abundance patterns of spi- agroecosystem. Entomophag a 32: 367 -37 5. ders inhabiting cotton in Texas. Tex. Agric. Exp. 1988a. The southern black widow spider, Latrodec- stn. Bull. 8-1566. tus nactans (Araneae: Theridiidae), as a predator of 1990. Seasonal patterns of spiders captured in suc- the red imported ffre ant, Solenopsis inoicta (Hy- tion baps in eastern Texas. Southwest. Entomol. 15: menoptera: Formicidae), in Texas cotton ffelds. J. 399-4r2. Appl. Entomol. 106: 5257. 1992. Comparison of sampling methods to predict f988b. Prey records of the web-building spiders phenology of predaceous arthropods in a cotton Dictgna segregata (Dictynidae), Theridion australe agroecosystem. Tex. Agric. Exp. Stn. Misc, Publ. (Theridiidae), Tidarren haemorrhoidale (Theridi- 1731. idae), and Frontinella pgramitela (Linyphiidae) in Dean, D. 4., W. L. Sterling & N. V. Homer. 1982. a cotton agroecosystem. Southwest. Nat. 33: 215- Spiders in eastern Texas cotton ffelds. J. Arachnol. 218. l0: 251-260. f989. Prey selection and predatory importance of Dean, D. 4., W. L. Sterling, M. Nyfreler & R. G. orb-weaving spiders (Araneae: Araneidae, Ul- Breene. f987. Foraging by selected spider pred- oboridae) in Texas cotton. Environ. Entomol. 18: ators on the cotton fleahopper and other prey. 373-380. Southwest. Entomol. 12: 263-270. Nyffeler, M., R. G. Breene, D. A. Dean & W. L. Ster- Edgar, W. D. 1970. Prey and feeding behaviour of ling. 1990. Spiders as predators of arthropod adult females of the wolf spider Pardosa anlentata eggs. J. Appl. Entomol. 109: 490-501. (Clerck). Neth. J. Zool.2O: 487-49L. Nyffeler, M, D. A. Dean & W. L. Sterling. 1992. 1188 ENvrnoNunxter. ENrolroLoGY Vol.2l, no.5

Diets, feeding specialization, and predatory role of TEXCIMS0: The Texas cotton insect model. Tex. two lynx spiders, Oxgopes salticus and, Peucetia Agric. Exp. Stn. Misc. Publ. MP-1646 (revised). oiridans (Araneae: Oxyopidae), in a Texas cotton van den Bosch, R. & K. S. Hagen. 1966. Predaceous agroecosystem. Environ. Entomol. (in press). and parasitic arthropods in California cotton ffelds. Polis, G. A., C. A. Myers & R. D. Holt. 1989. The Calif. Agric. Exp. Stn. Bull. 820. ecology and evolution of intraguild predation: po- Van Hook, R. I. 1971. Energy and nuhient dynam- tential competitors that eat each other. Annu. Rev. ics of spider and orthopteran populations in a grass- Ecol. Syst. 20: 297-330. land ecosystem. Ecol. Monogr.4l: l-26. Reinhard, H. J. 1926. The cotton fleahopper. Tex. Whitcomb, W. H. & K. Bell. f964. Predaceous in- Agric. Exp. Stn. Bull.339. sects, spiders and mites of Arkansas cotton ffelds. Riechert, S. E. T. Spiders as bio- & Lockley. 1984. Ark. Agric. Exp. Stn. Bull. 690. Iogical control agents. Entomol. 29: Annu. Rev. Whitcomb, W. R. Eason. 1967. history 299-320. H. & Life and predatory importance of the striped lynx spider Sokal, R. R. & F. Rohlf. 1969. Biomeky: the prin- J. (Araneida: Oxyopidae). Ark. Acad. Sci. Proc. ciples and practice of statistics in biological re- 2l:54- 58. search. Freeman, San Francisco. Whitcomb, Sterling, W. L. 1989. Estimating the abundance and W. H., H. Exline & R. C. Hunter. fgffi. impact of predators and parasites on Heliothis pop- Spiders of the Arkansas cotton ffeld. Ann. Entomol. Soc. Am.56: 653-660. ulations, pp. 37-56. In E. G. King & R. D. Jackson Ieds.], Proceedings ofa workshop on biological con- Young, O. P. & G. B. Edwards. 1990. Spiders in hol of Heliothis: increasing the efiectiveness ofnat- United States ffeld crops and their potential efiect ural enemies, ll-15 November 1985, Far Eastern on crop pests. f. Arachnol. 18: I-27. Regional Research Office, USDA, New Delhi, In- Young, O. P. & T. C. Lockley. 1985. The striped dia. lynx spider, Orgopes salticus (Araneae: Oxy- Sterling, W. L., K. M. El-Zik & L. T. Wilson. f989. opidae), in agroecosystems. Entomophaga 30: 329- Biological control ofpest populations, pp. f55-f89. 346. In R. Frisbie, K. El-Zik & L. T. Wilson [eds.], Inte- 1986. Predation ofstriped lynx spider, Orgopes sal- grated pest management systems and cotton pro- ticus (Araneae: Oxyopidae), on tarnished plantbug, duction. Wiley, New York. Lggus lineolards (Heteroptera: Miridae): a labora- Sterling, W. L., A. Dean & N. M. Abd El-Salam. tory evaluation. Ann. Entomol. Soc. Am. 79: 879- 1992a. Economic beneffts ofspider (Araneae) and 883. insect (Hemiptera: Miridae) predators of cotton fleahoppers. J. Econ. Entomol. 85:52-57. Receüsed for publication 27 Noaember 7997; ac- Sterling, W. L., A. W. Hartstack & D. A. Dean. 1992b. cepted 26 Mag 1992.