curred on 8 and 22 June, 6 and 20 July, and 3 .... spider mites (Table 3), suggesting a potential early ..... versity of California Systemwide Integrated Pest Man-.
This article is the copyright property of the Entomological Society of America and may not be used for any commercial or other private purpose without specific written permission of the Entomological Society of America.
Natural Enemies of Spider Mites (Acari: Tetranychidae) on Cotton: Density Regulation or Casual Association? L. T. WILSON,' P. J. TRICHILO,' AND D. GONZALEZ2 Department of Entomology, Texas A&M University, College Station, Texas 77843
Environ. Entomol. 20(3):849-856 (1991)
This study addresses the potential impact of natural enemies on the abundance of spider mites, Tetranychus spy., on cotton in the San Joaquin Valley of California. These natural enemies are omnivorous predators, and include the big-eyed bug, Geocoris pallens StAl and G. punctipes (Say),the minute pirate bug, Orius trfsticolor (White),and the western flower thrips, Frankliniella occidentalis (Pergande). Simple linear regression suggested that omnivorous predators were potentially effective in delaying the buildup of spider mites, with the highest rz (0.65) recorded for adult F. occidentalis. Geocoris showed the potential to suppress the rate of spider mite population increase (rl = 0.73). All three tested predator species exhibited the capacity to suppress early season spider mite abundance, with the highest r2 (0.62) recorded for Ceocoris and Orius. Predators were also potentially able to suppress mid- to late-season spider mite populations. Multiple regression analysis indicated a significant negative correlation between mid- to late-season spider mite abundance and early season predators. Results from a second year were less conclusive, suggesting that the reduced range of spider mite abundance limited our ability to discern potentially significant interactions during that year. ABSTRACT
KEY WORDS
.
Insecta, Cossypium, Tetranychus, predators
SPIDERMITES,Tetranychus spp. can cause serious economic injury to cotton, Gossypium hirsutum L., in the San Joaquin Valley of California (Leigh 1963; Leigh & Burton 1976; Wilson et al. 1983, 1985). Barring insecticidal intervention, natural enemies of spider mites, including omnivorous arthropod predators, are believed to maintain spider mites below economic injury levels (van den Bosch & Hagen 1966, Gonzalez & Wilson 1982, Gonzalez et al. 1982). Therefore, methods of pest management that preserve and enhance the natural enemy component should be considered desirable. During 1982 and 1983, we engaged in a field study to determine the impact of spider mites on cotton yield, which indicated that economic damage from spider mites occurs only when the rate of population increase exceeds a critical Infestation rate (Wilson et al. 1985). At rates progressively greater than the critical rate, there is a precipitous loss of yield. Infestations that exceeded the critical rate occurred only when permethrin, methyl parathion, or both were applied. Thus, consistent with other reports (Bower & Kaldor 1980, McWhorter 1982, Botha et al. 1986, Bentley et al. 1987), economically damaging levels of spider mites were closely associated with the application of insecticide. I Department of Entomology. -. University of California, Davis. Calif. 65616. Division of Biological Control. University of California. Riv-
erside, Calif. 92521.
Several mechanisms have been suggested to explain the spider mite-insecticide phenomenon. One explanation is that insecticides decimate large numbers of natural enemies, thereby reducing predation pressure and allowing the herbivore to increase rapidly in abundance (Huffaker et al. 1970, DeBach et al. 1976, Stoltz & Stern 1978, Trichilo & Leigh 1986b). Another view is that insecticides cause increased spider mite reproductive rates, either directly, or indirectly through the plant (Bartlett 1968, van de Vrie et al. 1972, Iftner & Hall 1984). A third view is that synthetic pyrethroids such as permethrin have a repellancy effect on spider mites, increasing the rate of dispersal to areas which have lower spider mite densities and higher food quality (Hoyt et al. 1978, Hall 1979, Penman & Chapman 1988). Lower densities and improved food quality generally lead to higher spider mite reproductive rates (Wrensch & Young 1978, Iftner & Hall 1983). While many investigators have long believed that arthropod predators exert significant regulation of herbivore densities, it has been difficult to demonstrate quantitatively this effect for spider mites on a field crop such as cotton. Unlike parasitoids, predators of spider mites leave no easily recognized evidence of their activity. Thus, assessment of predator effectiveness in the field on a large scale is relegated to inducing and observing changes in predator frequencies and any associated fluctuation in populations of their prey.
0046-225X/91/0849-0856$02.00/0@ 1991 Entomological Society of America
850
ENVIRONMENTAL ENTOMOLOGY
Vol. 20, no. 3
curred on 8 and 22 June, 6 and 20 July, and 3 August. Experimental differences between blocks were based solely on release dates of T. turkestani, and accompanying applications of permethrin. A sixth block, representing the modified part of the design, was subdivided into six plots, of which five received inoculative release of T. turkestani (but no permethrin), corresponding to the dates of the first five blocks. The sixth plot was untreated and allowed for testing of spatial homogeneity (see Trichilo et al. 1990). This design translated into a five block by five treatment factorial, consisting of the following five treatments: (1) no mites released, and no pesticides, (2) mites released, permethrin, and dicofol at 0.2 P(I), (3) mites released, permethrin, and dicofol at 0.8 P(Z), (4) mites released, permethrin, but no dicofol, and (5) mites released, but no pesticides. Timing, application rates, and experimental designs for 1982 and 1983 are described in detail by Trichilo et al. (1990). Sampling. Spider mite (immature and adult) and predator (egg through adult) seasonal abundance was estimated using the proportion of sample units infested, P ( I ) , with one or more individuals of a given species (Wilson et al. 1983). Plots were sampled weekly, beginning on 26 May and ending on 14 September in 1982, which yielded 17 samples. In 1983, weekly sampling began on 25 May and ended on 8 October, which yielded 20 samples. Physiological time (degree-days (DD) after planting) was calculated, based on an approximate minimum developmental temperature of Materials and Methods 12OC for spider mites, as extrapolated from develExperimental Design. This study represents the opmental data from Carey (1980). In each year, a sample consisted of 40 mainstem analysis of data collected on cotton during 1982 and 1983 at the University of California, Westside leaves, one leaf per plant; each plant was chosen Field Station. The 1982 study was a factorial ex- at random from a given plot. The leaf chosen repperiment based on a split-plot design of 6.7 ha, in resented the main stem node leaf most likely to which each of four blocks representing four levels have spider mites or predators (see Wilson et al. of permethrin was divided into six subplots or treat- 1983). Insufficient levels of Geocoris adults and ments. Permethrin was used as an insecticide to nymphs were detected to warrant including these suppress natural enemies, while the acaricide di- stages in the statistical analysis. Statistical Analysis. Simple linear regression was cofol was used to suppress spider mites. Methyl parathion was used as a pesticide to contrast di- used to test the influence of major predator species cofol. The six treatments included: (1) no pesti- on spider mite population dynamics. Analyses were cides, (2) dicofol applied (at half label rate) when also performed using multiple linear regression, to the proportion of infested leaves,.P(Z),reached 0.2, assess the effects of early season predator popula(3) dicofol applied (at full rate) at 0.2 P(Z), (4) tions on early and mid-to-late season spider mites. dicofol applied (at full rate) at 0.8 P(Z), (5)methyl For all regression analyses, significant differences parathion applied at half rate, and (6) methyl para- were assumed to exist at P < 0.05. Spider mite infestation rate (dp = change in thion applied at full rate. To guarantee a minimum level of spider mite infestation, a small number of proportion [infested leaves]/100 DD), was used as cotton leaves from another field, with damage a dependent variable. A second dependent varisymptoms resembling that of Tetranychus turke- able, initiation of rapid spider mite increase, was stani Ugarov and Nikolski, was distributed to all represented by the sample date (in degree-days) immediately preceding the phase in which spider plots on 4, 9, 15, and 22 June. The 1983 study was based on a modified factorial mites most rapidly increased in proportion of indesign, also in the form of a split-plot, in which fested leaves. The third and fourth dependent varieach of five blocks was subdivided into four treat- ables were generated by computing the area under ments. Each block represented five separate dates the spider mite P ( I ) curve, as a function of physof permethrin application, followed by inoculative iological time for a given treatment, from sample 0 > 12OC from planting) through samrelease of T. turkestani. Spider mite releases oc- 1 ( ~ 3 0 DD
Although a variety of arthropod predators, such as nabids, lacewings, predatory mites, and predaceous thrips, are found on cotton, they appear to be insignificant in controlling spider mite populations during critical infestation periods in the San Joaquin Valley due to low numbers and normally late season occurrence. The major natural enemies of spider mites on cotton in this region are generalist, omnivorous predators, and include the western flower thrips, Frankliniella occidentalis (Pergande), the big-eyed bugs, Geocoris pallens StAl and G . punctipes (Say), and the minute pirate bug, Orius tristicolor (White) (van den Bosch & Hagen 1966, Gonzalez & Wilson 1982, Gonzalez et al. 1982, Trichilo & Leigh 1986a,b). If natural enemies suppress spider mite densities, we would expect to observe a negative association between snider mite abundance and that of their predators over a given block of time, thus, at least partially explaining why spider mites dramatically increase in abundance after insecticides are used. When predators are at normal (i.e., maximum) levels throughout the season, overall levels of spider mites should be relatively low, and vice versa. However, there is little quantitative evidence that effectively demonstrates the impact of natural enemies on populations of spider mites on cotton fields. Our objective was to assess the relationship between spider mites and the natural enemy complex during the cotton season.
WILSONET AL.: NATURAL
June 1991
0.2
-
.
EARLY SEASON
0.0 1.0
-B
0.8
-
,
I
;! ;!
I SEASON
I----
-.- -
,
m ~ a r v a e
I . .-.WFT Adults I------Orius Nympha I -1 Geocorii Eggs 1
PHYSIOLOGICAL TIME (DD>12OC From Planting) Fig. 1. Seasonal infestation levels, P(,I),for (A) spider mites, Tetranychus spp., and (B) three major arthropod predators in the cotton agroecosystem.
PHYSIOLOGICAL TIME (DD>lZ° From Planting) Fig. 2. Estimated seasonal density of (A) spider mites, Tetranychus spp., and (B) three major arthropod predators in the cotton agroecosystem.
pie 8 (ss850 DD) and from sample 9 ( ~ 8 5 DD) 1 through sample 16 (a1622 DD). These variables were designated early season and mid-to-late season spider mite proportion infested-degree-days (P(I)DD), respectively. The fifth and sixth dependent variables were derived by converting proportion infested leaves into estimated absolute densities (see Wilson et al. 1983), and then calculating spider mite degree-days (DD) for early and mid-to-late season. Predator independent variables, P(1)-DD and predator-DD, were derived similarly to spider mite variables for the early and mid-to-late season periods. Predator impact was also analyzed in terms of the population immediately following applications of permethrin in 1982. These predator variables comprised the areas under the P(1) and density curves, between sample dates 4 through 7, designated early season post=insecticide predators. Because applications of permethrin were staggered throughout the season in 1983, the post-insecticide variable was not derived for that year. Treatments from each year that involved applications of dicofol were omitted from the analyses, because spider mite abundance would be artificially low in relation to the frequency of predators. Because all treatments in 1983 involved inoculative releases of T. turkestani, except the control (no mites released, no pesticides), control treatments were omitted from the analysis.
200 DD before spider mites (Fig. 1 and 2). Adult and larval F.occidentalis usually exhibited a sharp increase in abundance, immediately following their appearance on cotton, followed by an abrupt drop 150-200 DD later. Beyond this point, their populations fluctuated in a random and unpredictable manner during each year. Geocoris eggs and Onus nymphs were first observed around 550 DD, often in conjunction with the initiation of spider mite infestations (Fig. 1 and 2). However, occurrence of eggs and juveniles implies that Geocoris and Orius adults had entered cotton fields at an earlier date. Simple Associations. In 1982, the greater the abundance of early season F . occidentalis adults, Geocoris eggs, and Onus nymphs, the later the estimated time when the spider mites initiated the period of rapid population increase (Table I), suggesting that predators can delay the onset of spider mite infestation. The subsequent rate at which the spider mite population increased was negatively correlated only with early season Geocoris eggs. Early season spider mite P(1)-DD was negatively correlated with F . occidentalis adults, Geocoris eggs, and Orius nymphs and adults (Table 2). Mid-to-late season spider mite P(1)-DD was negatively correlated with early season Geocoris eggs, early season Orius nymphs and adults, and midto-late season F. occidentalis larvae and adults (Table 2). There was also significant negative correlation between mid-to-late season spider mite-DD (i.e., absolute numbers) and early season Geocoris eggs, early season Onus nymphs, and mid-to-late season F. occidentalis adults. Regression on post-insecticide F. occidentalis
Results Population Phenology. Frankliniella occidentalis was usually observed around 350 DD, 150-
Vol. 20, no. 3 Table 1. Coefficients of determination (r2), for spider mite infestation rate and initiation of rapid spider mite increase, regressed on early season predator abundance for 1982 -
Early season predator P(7)-DD WFT* larvae WFT adults Geocoris eggs Orius nymphs Orius adults a
Spider mite abundance variables" Initiation of rapid ppulation inCree Infestation rate (+) 0.02 NS (+) 0.65** (+) 0.52** (+) 0.45* (+) 0.11 NS
(-) (-) (-) (-) (-)
0.01 NS 0.17 NS 0.73*** 0.17 NS 0.01 NS
Parentheses enclose sign of the regression slope; *, P < 0.05; < 0.01;***. P < O.Ool;NS, P > 0.05;n = 12. WFT, western flower thrips, F. occfdentalfs.
**, P
*
larval populations (samples 4-7) indicated a slightly significant negative association between early season F . occidentalis larvae and mid-to-late season spider mites (Table 3), suggesting a potential early season impact by F. occidentalis larvae on later season spider mites. Using the post-insecticide time period also supported earlier findings on other predators as well. In 1983, there was a significant negative correlation between mid-to-late season spider mite P(Z)DD and early season F . ocddentalis adults and larvae (Table 4). This effect was also true for estimated numbers of snider mites. There. were no .-significant ciorrelations between snider mite abu dance and Ceocoris eggs, but there was a significant positive correlation with Orius nymphs (Table 4). Multiple Associations. Multiple regression analyses were conducted for all possible combinations of spider mite and predator variables. However, many were not statistically significant, and therefore, only those analyses which exhibited significant individual and overall effects are presented here. Regressing early season spider mite P(Z)-DD against early season predators in 1982 produced
the first relationship in Table 5. All coefficients were negative, suggesting that all of these predators had a negative impact on early season spider mites. The potential impact of early season predators on mid-to-late season spider mites is depicted in the second function of Table 5. All coefficients, except for F. occidentalis adults, were negative, suggesting that early season activity of natural enemies can negatively affect spider mite abundance later in the season. Removing the effect of F. occidentalis adults from the relationship resulted in the third function of Table 5. The fourth function was produced using estimated spider mite and predator densities. Coefficients from this function suggest that all early season predators exhibited some potential for impact on densities of mid-to-late season spider mites. In general, results from 1983 show little or no early season potential effectiveness by Geocoris or Onus. There were no significant multiple associations between predators and early season spider mites, and only a few significant or nearly significant associations between early season predators :ind mid-to-late season spider mites (Table 5). Fhe iabsence of early season activity by these predaitors i1s in sharp contrast to the results of 1982, wl-~ich *h/^*È-^th^'m tn hft..,., ff-m."" ../^t^yitlnI ^>/È-l. SJIJIV 1y 3c;aSOn impact on spider mite populations. Mean spider mite P(Z)-DD was lower in 1983 (596) than in 1982 (651), which constituted an 8% drop. However, estimated spider mite-DD were on average 34% lower in 1983. The range between high and low values of spider mite P(Z)-DD was also much less in 1983 (196) than in 1982 (385), representing a drop of 49%. Similarly, the range in estimated spider mite-DD was 70% less in 1983 than in 1982.
-
TT
uu
J L A W A A ~~ V L G J U L ~ G ~ a J l~
U A W ~ S . ~ I"A ~ C T V -
"-A
Discussion Population Phenology. The time of initiation of rapid spider mite increase occurs relatively early
Table 2. Coefficients of determination (r2), for spider mite abundance, regressed on early season and mid-late season predator abundance for 1982 Spider mite abundance variables"'* Predator
WFT" larvae
WFT adults
Ceocoris eggs
Ortus nymphs Orius adults
Spider mite P(1)-DD
Time
early mid-late early mid-late early mid-late early mid-late early mid-late
Spider mite DD
Early
Mid-late
Early
Mid-late
(-) 0.00 NS
(-) 0.00 NS (-) 0.67**
(-) 0.01 NS
(-) 0.02 NS (-) 0.03 NS
- ) 0.57** (-) 0.62**
- ) 0.25 NS (-) 0.56** (-) 0.60** (-) 0.01 NS (-) 0.76*** (-) 0.01 NS (-) 0.49* (+) 0.24 NS
(-) 0.20 NS
(-) 0.28 NS (-) 0.77(-) 0.65** (-) 0.15 NS (-) 0.60** (+) 0.11 NS (-) 0.22 NS (+) 0.15 NS
-
-
(-) 0.62**
-
(-) 0.37*
-
-
-
(-) 0.44*
(-) 0.29 NS -
(-) 0.45*
Parentheses enclose sign of the regression slope; *, P < 0.05; **, P < 0.01; ***, P < 0.001; NS, P > 0.05; n = 12. *Spider mite P(1)-DD regressed on Predator P(1)-DD, and Spider mite DD on Predator DD. WFT, western flower thrips, F. occfdentalfs.
a
June 1991
WILSONET
AL,:
Table 3. Coefficients of determination 1982
reda at or abundance for
NATURAL ENEMIES OF SPIDER MITES
853
(s), for spider mite abundance, regressed on early season post-insecticide Spider mite abundance variables"'^
Predator
Spider mite P(1)-DD Early (-) (-) (-) (-) (-)
WFTc larvae WFT adults Ceocoris eggs Orius nymphs Orius adults a
Spider mite DD
Mid-late (-) (-) (-) (-) (-)
0.11 NS 0.22 NS 0.58** 0.510.33 NS
Early
0.36* 0.11 NS 0.46* 0.38* 0.49*
(-) (-) (-) (-) (-)
Mid-late
0.00 NS 0.01 NS 0.55;0.35* 0.26 NS
(-) 0.03 NS (-) 0.01 NS (-) 0.66** (-1 0.4V (-) 0.22 NS
Parentheses enclose sign of the regression slope; *, P < 0.05; **, P < 0.01; *#, P < 0.001; NS, P > 0.05; n = 12. Spider mite P(1)-DD regressed on Predator P(1)-DD, and Spider mite DD on Predator DD. WFT, western flower thrips, F. occidentalis.
in the season, and in heavy mite infested treatments, P ( I ) can reach 1.0 by midseason. During this period, spider mites become established on cotton plants and expand their distribution (i.e., disperse). Frankliniella occidentalis is the earliest arthropod to colonize cotton and has the potential to be an effective regulating agent .by restricting spider mite infestations, delaying them, or both. As later arrivals, Geocoris and Orius would seem to be less well positioned in time than F. occidentalis to affect early spider mite infestations. However, our 1982 data indicate the potential for early season impact by both Geocoris and Orius. Being polyphagous, Orius and Geocoris prey on F. occidentalis as well as spider mites (van den Bosch & Hagen 1966, Huffaker et al. 1970, Gonzalez & Wilson 1982, Letourneau & Altieri 1983, L.T.W., unpublished data). Early season F. occidentalis can serve as an alternative food source for these other predators until spider mites become established on cotton. As an herbivore, a food source for Geocoris and Orius, and a predator of spider mites, F . occidentalis is of considerable importance in the cotton-arthropod food web.
Predator-Spider Mite Associations. All predators exhibited the potential to suppress some aspect of spider mite population growth. The early season potential of Geocoris, Onus. and F . occidentalis to affect spider mite abundance was supported by 1982 results, but less so by 1983 results. In 1982, the greater the early season abundance of each of these predators, the later the time at which the spider mites initiated theil phase of rapid population increase, and lower the abundance of midto-late season spider mites. :Fhis effect was obscured * by the reduced abundance of spider mites in 1983, and possibly by the staggered release of T. turkestunt, and accompanying applications of permethrin. Frankliniella occidentalis has been characterized as an opportunist and a passive predator (Trichilo & Leigh 1986a). This omnivore was observed to feed only on newly laid eggs of spider mites, and has difficulty maneuvering in spider mite webbing. Moreover, this species is primarily herbivorous and only becomes noticeably predaceous at high egg densities. Flower thrips do not actively seek spider mite eggs, but will feed on them if encountered on the leaf surface (P.J.T., unpub-
-
.. . . .
.
Table 4. Coefficients of determination (r2), for spider mite abundance, regressed on early season and mid-late season predator abundance for 1983 Spider mite abundance variables"^ Predator
Time
Solder mite P(I)-DD
Snider mite DD -
Early WFTc larvae WFT adults
Geocorfs eggs
Ortus nymphs
Ortus adults^ a
^
early mid-late early mid-late early mid-late early mid-late early mid-late
(-) 0.13 NS
-
(-) 0.17 NS
-
(+) 0.05 NS
-
(+) 0.42*
-
Mid-late (-) (-) (-) (-)
0.59** 0.10 NS 0.69;. 0.33 NS (-) 0.10 NS (+) 0.01 NS (+) 0.01 NS (+) 0.52*
(+) 0.02 NS
Early (-) 0.00 NS
-
(-) 0.15 NS
-
(+) 0.02 NS (+) 0.52*
-
--
Mid-late (-) 0.66**
(-) 0.28 NS (-) 0.69** (-) 0.61** (-) 0.04 NS (-) 0.19 NS (+) 0.00 NS (+) 0.30 NS
-
(+) 0.18 NS
Parentheses enclose sign of the regression slope; * P < 0.05; **, P < 0.01; ***, P < 0.001; NS, P > 0.05; n = 10. Spider mite P(1)-DD regressed on Predator P(1)-DD, and Spider mite DD on Predator DD. WFT, western flower thrips, F. occidentalis. No Orius adults observed in early season samples.
ENVIRONMENTAL ENTOMOLOGY
854
Vol. 20, no. 3
Table 5. Coefficients and corresponding statistics, resulting from multiple linear regression of spider mite abundance, on early season predator abundance for 1982 and 1983
Dependent variable SMmP(1)-DD SMmhP(1)-DD SMmk P(1)-DD SMmkDD
TLes
189.622 742.507 946.229 7.143.777.135
statistics
Independent variables
Y-intercept
-0.094 -1.168 -0.602 -7.069
TAes
-0.415 +1.905
-
-31.498
+
GEes
0%
1982 -2.608 -13.691 -10.512 -1.002.866
-1.102 -9.671 -10.454 -366.156
+
+
0'4-
-4.020 -27.185 -20.318 -1.155.033
r2
0.87 0.94 0.91 0.89
F
P
8.336 19.911 16.675 9.894
0.0113 0.0011 0.0011 0.0073
+
Regression equations of the form: SM = fSy + ffi TL flv TA fti GE b4ON Bs OA. es, early season; mls, mid-late season;SM, spider mite; TL, flower thrips, F.occidentalis, larva, TA, flower thrips adult; GE, Ceocoris egg; ON, Orius nymph; OA, Orius adult. Spider mite P(1)-DD.
regressed on Predator P(1)-DD,
and Spider mite DD on Predator DD.
lished data). Consequently, their effectiveness as predators should increase as they become more abundant. Significant negative associations between spider mites and mid-to-late season F. occidentalis in 1982 suggest that they have the potential to affect spider mite abundance throughout the season. Frankliniella occidentalis had potential impact during mid-season, when spider mites were typically increasing " in density. If all 1iredators were equally abundant, we woilid not expt;ct F. occidentalis to have as dramatic an effect as Geocoris or Onus. By virtue of their larg;er size, Geocoris and u r w s could be expected to take more prey per unit time than F . occidentalis. Moreover, these two larger predators appear to be more aggressive, and probably search more acprey than does F. occidentalis. Data fr()m tivel!y for 1982 support this prediction, with smaller sim]?le and multiple linear regression coefficients for F. . . . * ,. -occidentalis than tor the other two predators. However, early season F. occidentalis often outnumbered early season Geocoris by more than 20:1, and therefore could have greater overall impact. Frankliniella occidentalis was difficult to suppress with insecticide for extended periods. Adult F . occidentalis primarily inhabit cotton flowers during peak blooming (Pickett et al. 1988, P.J.T., unpublished data), which could afford some escape from insecticide exposure during this period. Adults are highly mobile and can invade insecticide-treated plots relatively soon after the chemical has dissipated. With a reproductive rate similar to spider mites (Trichilo & Leigh 1985, 1988), larvae can quickly increase in numbers. Rapid replacement of larval F. occidentalis undoubtedly interfered with our ability to clearly quantify relationships between these predators and spider mite abundance. Results of our current study emphasize the difficulty in consistently observing significant associations among populations within a given year and from one year to another. The ability to demonstrate significant associations is dependent upon producing sufficient variation in both spider mite ,
.
A
.
. .
...
and predator frequencies to allow the impact of predators to be expressed. Lower average spider mite abundance, and especially the greatly reduced range of variation in the numbers of spider mites, undoubtedly contributed to the difficulty in demonstrating significant correlations between spider mites and their natural enemies in 1983. Greater spider mite abundance in permethrin treated plots might be the result of increased spider mite dispersal (Iftner & Hall 1983, Penman & Chapman 1988). Increased dispersal should be observed as increased infestation rates (i.e., increase in proportion of infested leaves), as spider mites move away from heavily or moderately infested leaves to uninfested leaves. Reduced competition and higher food quality associated with less crowded leaves is generally believed to stimulate higher spider mite reproductive rates (Wrensch & Young 1978). Negative correlations between natural enemies and their prey do not demonstrate cause and effect. Absence of consistent correlations from one year to another could suggest that the fluctuation in spider mite densities was coincidental and not necessarily the result of predator activity. However, it is more likely that the extremely low abundance of spider mites during the second year of the study was the major reason for the loss of significant effects that year. Flower thrips and spider mites share the same photosynthetic resource, although F. occidentalis adults acquire a major portion of their nutrition from pollen (Trichilo & Leigh 1988). A common food source and difficulty with spider mite webbing could constitute reasons for avoidance of spider mites by F. occidentalis. This option becomes less likely later in the season when spider mites are increasing in abundance and the resource of fresh new cotton leaves is diminishing. The predaceous nature of all three of these natural enemies has been documented (York 1944, Iglinski & Rainwater 1950, van den Bosch & Hagen 1966, Huffaker et al. 1970, Trichilo & Leigh 1986a), and there is no evidence to support avoidance of spider mites by Geocoris or Onus. Thus, there is
-
June 1991
WILSONET AL.: NATURAL ENEMIES OF SPIDERMITES
adequate reason t o suggest that much of the observed differences in spider mite abundance between experimental units resulted from variation in predation pressure.
and permethrin on Tetranychus urticae Koch (Acari: Tetranychidae) dispersal behavior. Environ. Entomol. 12: 1782-1786. 1984. The effects of fenvalerate and permethrin residues on Tetranychus urticae Koch fecundity and rate of development. J. Agric. Entomol. 1: 191-200. lglinski, W., Jr., & C. F. Rainwater. 1950. Onus Acknowledgment insidiosus, an enemy of spider mites on cotton. J. This research was supported in part by a grant to Lo Econ. Entomol. 43: 567-568. T. Wilson from the EPA and USDA funded National Leigh, T. F. 1963. Considerations of distribution, Consortium for Integrated Pest Management, and by a abundance, and control of acarine pests of cotton. grant to D. Gonzalez and L. T. Wilson from the UniAdv. Acarol. 1: 14-20. versity of California Systemwide Integrated Pest Man- Leigh, T. F. & V. E. Burton. 1976. Spider mite pests agement Program. We thank R. Friesen and M. P. Hoffof cotton. Division of Agricultural Sciences. Univermann for their help in gathering much of the field sity of California, Berkeley. Leaflet 2888, infestation data. Letourneau, D. K. & M. A. Altieri. 1983. Abundance patterns of a predator, Orius tristicolor (Hemiptera: Anthocoridae), and its prey, Frankliniella occidentalis (Th~sanoptera:Thripidae): Habitat attraction in References Cited po~yculturesversus monocultures. Environ. Entomol. Bartlett, B. R. 1968. Outbreaks of two-spotted spider12: 1464-1469. mites and cotton aphids following pesticide treatMcWhorter, G . M. 1982. Spider mites: a status report, ment. I. Pest stimulation vs. natural enemy destrucpp. 189-191. In 1982 Proc. Beltwide Cotton Prod. tion as the cause of outbreaks. J. Econ. Entomol. 61: Res. Conf. National Cotton Council of America. 297-303. Memphis, Tenn. Bentley, W. J., F. G. Zalom, W. W. Barnen & J. P. Penman, D. R. & R. B. Chapman. 1988. PesticideSanderson. 1987. Population densities of Tetraninduced mite outbreaks: pyrethroids and spider mites. ychus spp. (Acari: Tetranychidae) after treatment Exp. App. Acarol. 4: 265-276. with insecticides for Amyelais transitella (Lepidop- Pickett, C. H., L. T. Wilson & D. Gonzalez. 1988. tera: Pyralidae). J. Econ. Entomol. 80: 193-199. Population dynamics and within-plant distribution Botha, J. H., H. Van Ark 81A. J. Scholtz. 1986. The of the western flower thrips (Thysanoptera: Thripieffect of bollworm control with regard to populations dae), an early-season predator of spider mites infestof spider mites and some of their natural enemies on ing cotton. Environ. Entomol. 17: 551-559. cotton. Phytophylactica 18: 141-150. Stoltz, R. L. & V. M. Stern. 1978. Cotton arthropod Bower, C. C. & J. Kaldor. 1980. Selectivity of five food chain disruption by pesticides in the San Joaquin Valley, California. Environ. Entomol. 7: 703-707. insecticides for codling moth (Laspeyresia pomonella) control: effects on the twospotted spider mite (Te- Trichilo, P. J. & T. F. Leigh. 1985. The use of life tranychus urticae) and its predators (in apple ortables to assess varietal resistance of cotton to spider mites. Entomol. Exp. Appl. 39: 27-33. chards). Environ. Entomol. 9: 128-132. Carey, J. R. 1980. Ecological investigations on the 1986a. Predation on spider mite eggs by the western tetranychid mites on cotton. Ph.D. dissertation, Uniflower thrips, Frankliniella occidentalis (Thysanopversity of California, Berkeley. tera: Thripidae), an opportunist in a cotton agroecoDeBach, P., C. B. Huffaker & A. W. MacPhee. 1976. system. Environ. Entomol. 15: 821-825. 1986b. The impact of cotton plant resistance on spider Evaluation of the impact of natural enemies, pp. 255mites and their natural enemies. Hilgardia 54: 1-20. 285. In C. B. Huffaker & P. S. Messenger [eds.l The1988. Influence of resource quality on the reproducory and practice of biological control. Academic, New tive fitness of flower thrips (Thysanoptera: ThripiYork. dae). Ann. Entomol. Soc. Am. 81: 64-70. Gonzalez, D. & L. T. Wilson. 1982. A food web apTrichilo, P. J., L. T. Wilson & D. Gonzalez. 1990. proach to economic thresholds: a sequence of pests/ Relative abundance of three species of spider mites predaceous arthropods on California cotton. Ento(Acari: Tetranychidae) on cotton, as influenced by mophaga 27: 31-43. pesticides and time of establishment. J. Econ. EntoGonzalez, D., B. R. Patterson, T. F. Leigh & L. T. mol. 83: 1604-1611. Wilson. 1982. Mites, a primary food source for two predators in San Joaquin Valley cotton. Calif. van de Vrie, M., J. A. McMurtry & C. B. Huffaker. 1972. Ecology of tetran~chidmites and their natuAgric. 36: 18-20. ral enemies: a review. 111. Biology, ecology and pest Hall, F. R. 1979. Effects of synthetic pyrethroids on status and host-plant relations of tetran~chids.Hilmajor insect and mite pests of apple. J. Econ. Entogardia 41: 343-432. mol. 72: 441-446. Hoyt, S. C., P. H. Westigard & E. C. Burts. 1978. van den Bosch, R. & K. S. Hagen. 1966. Predaceous and parasitic arthropods in California cotton fields. Effects of two synthetic pyrethroids on the codling California Agricultural Experiment Station Bulletin moth, pear psylla and various mite species in north820. west apple and pear orchards. J. Econ. Entomol. 71: Wilson, L. T., D. Gonzalez, T. F. Leigh, V. Maggi, C. 431-434. Foristiere & P. Goodell. 1983. Within-plant disHuffaker, C. B., M. van de Vrie & J. A. McMurtry. tribution of spider mites (Acari: Tetranychidae) on 1970. Ecology of tetranychid mites and their natcotton: a developing implementable monitoring proural enemies, a review. 11. Tetranychid populations gram. Environ. Entomol. 12: 128-134. and their pssible control by predators: an evaluation. Wilson, L. T., D. Gonzalez & R. Plant. 1985. PreHilgardia 40: 391-458. dicting sampling frequency and economic status of Iftner, D. C. & F. R. Hall. 1983. Effects of fenvalerate
856
ENVIRONMENTAL ENTOMOLOGY
spider mites on cotton, pp. 168-170. In 1985 Proc. Beltwide Cotton Prod. Res. Conf. National Cotton Council of America. Memphis, Tenn. Wrensch, D. L. & S. S. Y. Young. 1978. Effects of density and host quality on rate of development, survivorship, and sex ratio in the carmine spider mite. Environ. Entomol. 7: 499-501.
Vol. 20, no. 3
York, G. T. 1944. Food studies of Geocoris spp., predators of the beet leafhopper. J. Econ. Entomol. 37: 25-29.
Received for publication 11 June 1990; accepted 21 November 1990.
