yield losses were apparent. As has been Why lacewings may fail to suppress . . . found in studies of several different cot- ton pests, plants that are setting bolls appear to have limited abilities to compensate for feeding damage. During Predators that eat other the late season, when bolls are opening and cotton lint is exposed, cotton aphids create problems by excreting large quan- predators disrupt cotton tities of sugary honeydew, which fall onto lint and create “sticky cotton.” Problems with sticky cotton become ap- control parent during harvest, ginning and yarn manufacturing, and threaten overseas markets and the price premiums Califor- Jay A. Rosenheim D Lawrence R. Wilhoit nia cotton has historically received. Be- cause the cotton aphid is already resis- tant to many insecticides in California and an even larger array of pesticides in The predatory green lacewing, predators may attack other predators, the southern United States, long-term with potentially negative effects on carnea, is often management of aphids will probably control. Here, we report a study de- need to rely on noninsecticidal alterna- abundant in mid- and late-season signed to determine the effectiveness of tives. cotton fields in the San Joaquin lacewing larvae, , as Cotton grown in the San Joaquin Val- Valley. However, neither these biological control agents of the cotton ley generally develops large populations aphid, , which feed on natural populations nor insecfary- of generalist predators, including big- mid- and late-season cotton in the San eyed bugs (Geocoris spp.), damsel bugs reared and mass-released lace- Joaquin Valley. We found that a number (Nabis spp.), assassin bugs ( spp. wings appear to suppress popula- of generalist predators impose heavy and others), minute pirate bugs (Orius tions of the cotton aphid. The key mortality on lacewing populations and tristicolor) and lacewings (primarily reason for the ineffectiveness of thereby render natural and augmented Chrysoperla carnea). Other predators, less biological control appears to be populations of lacewings ineffective as abundant, are also present. Lacewing biological control agents. the heavy mortality imposed on larvae, known as potential predators of The cotton aphid, a major pest of cot- aphids in many crops, are available from lacewing larvae by other general- ton in California, presents very different several commercial insectaries. Some ist insect predators. Results of a problems for cotton production at differ- growers use augmentative releases of study suggest that interactions ent times of the year. Early season popu- purchased lacewings at the recom- between different of in- lations, which develop on very small mended rate of 5,000 eggs per acre to cotton seedlings (often on plants with improve control of cotton aphids. sect predators may disrupt the fewer than six nodes), can cause crinkled biological control of pest species. leaves, partial defoliation and stunted Natural densities of lacewings plant growth. Nevertheless, research still The first question addressed was: in progress suggests that this damage is How many lacewing eggs and larvae are Rising costs of insecticides, widespread fully compensated for before harvest naturally present in late-season cotton insecticide resistance and increasing re- and that the timing of crop maturation, fields when sticky cotton is a potential strictions on insecticide use in California quantity of yield and cotton fiber quality problem? Six fields were sampled be- have spurred interest in insect manage- are unaffected. At mid-season, cotton tween August 26 and September 2,1992. ment by other means, including biologi- aphid populations are frequently low; In each field, entire plants were sampled cal control. Generalist insect predators however, during 1992, populations in and carefully inspected to count all are frequently abundant in annual crops, the southern San Joaquin Valley grew lacewings present. The number of plants including field and vegetable crops, and rapidly during July and August, and per row meter was recorded to translate have been identified as important in counts into lacewings per acre. As suppressing populations of damaging shown in figure 1, natural densities of . However, effective use of such lacewing eggs were in all cases high, natural enemies to manage pests re- ranging from 77,000 to 380,000 per acre. quires more complete understanding of Clearly, these naturally present eggs insect ecology, including the biology of were so abundant in the sampled fields insect predators. that more releases of insectary-reared In cotton, for instance, generalist lacewing eggs at the recommended rate predators may be critical in controlling (5,000 per acre) would not be useful. Ob- populations of spider mites, Lygus bugs servations of many other fields suggest that lacewings establish large densities and several worm pests. Few pest man- Y - agement experts or insect ecologists 5 of eggs in cotton fields at midseason and have studied interactions between insect 2 that these densities remain fairly stable predators, leaving the impression that 5 for nearly all of the remainder of the P insect predators feed exclusively on her- 7 growing season. Thus, before releasing bivorous . However, it is pos- Adults and nymphs of the cotton aphid, lacewings, sampling of naturally present sible in theory that some generalist Aphis gossypii. lacewings is recommended to determine

CALIFORNIA AGRICULTURE, SEPTEMBER-OCTOBER 1993 7 food limitation, was causing heavy mor- bugs, prey on lacewing larvae. Other tality of young lacewing larvae. predators that were abundant, for example damsel bugs, were never ob- Lacewing egg releases served feeding in the field on any prey; Methods. Experiments to assess the thus, we were unable to determine if effect of lacewing egg releases were con- they preyed on lacewing larvae. To de- ducted at the Kearney Agricultural Cen- termine if generalist predators caused ter and the UC Cotton Research Station substantial mortality of lacewing larvae, at Shafter in 1991. Two treatments, the and to quantify their effect on the bio- release of 5,000 insectary-reared lace- logical control of the cotton aphid, we wing eggs and no eggs released, were performed a manipulative experiment. replicated ten times in a randomized Methods. This experiment, con- block design. Our release rate, about 260 ducted August 21-30, 1992, at the times the recommended rate of 5,000 Kearney Agricultural Center, was de- eggs per acre, was chosen to determine signed to isolate the influences of each of whether any effect from lacewings could three predators, big-eyed bugs, damsel be discerned; smaller releases would bugs and assassin bugs, with and with- have been dwarfed by naturally present out the presence of lacewings, on lace- populations. Releases at these rates are wing survival and biological control of not, however, commercially feasible. the cotton aphid. Individual plants har- Each plot was 10 feet by 6 rows of cot- boring aphid populations were chosen ton, cultivar GC-510. Two days before for study. Each plant was carefully in- the lacewings were released and weekly spected to count all aphids and to re- thereafter, aphid densities in each plot move all nymphal and adult predators. were quantified by sampling ten leaves Polyester mesh sleeves were then per plot and counting all aphids. To de- placed over the plants to create a small termine the impact of the lacewings, we cage (either the entire plant or the top used the ratio of postrelease aphid portion of the plant was enclosed, de- counts to prerelease aphid counts. pending on aphid density). Before seal- Results. Because the effects of the ing the sleeve cage, eight treatments, lacewing releases were similar in the each replicated 6 to 10 times, were then two experiments, the results are com- applied to the caged plants: (1)no pred- bined in figure 2. For 3 weeks following ators, an ”aphids only” treatment; (2) the lacewing releases, aphid densities aphids plus two big-eyed bug adults; remained approximately constant in the (3) aphids plus two damsel bug adults; release plots but expanded in the non- (4) aphids plus two assassin bug adults; release, control plots. By the second (5) aphids plus five C. carnea lacewing week, the aphid population in the re- lease plots was less than before release (ratio = 0.85 k 0.16), while the aphid population had more than doubled in the control plots (ratio = 2.37 & 0.67; t = 2.50, P = 0.03). This trend continued for a third week, but for the remaining period there was little or no difference between the two treatments (fig. 2). Thus, releases of extremely high densities of lacewing eggs produced only a modest and tran- sient suppression of aphid populations. We suspect that releases conducted at the recommended rate would not have produced detectable results. Fig. 1. Naturally occurring densities of A few days after the lacewing eggs lacewing eggs and larvae in six fields were released, few lacewing larvae sampled during the late season (August could be found in any of the plots. This 26 September 2,1992) in San Joaquin Adult assassin bug, . Assas- - observation was consistent with many Valley cotton. Five plants were sampled in sin bugs were found to be major predators each field except for field S1, in which 48 of lacewing larvae. other observations where many adult lacewings and their eggs were seen, but plants were sampled. Field K had been few or no larvae were detected (fig. 1). treated with methamidophos June 24 for whether releases will substantially in- However, a number of other generalist Lygus control, and field S3 had been crease densities in the field. predators were present. treated with chlorpyrifos August 4 for Although egg densities were uni- aphid control; all other fields were never Predators attack lacewings treated. Shown are means plus one stan- formly high, in four of the six fields we dard error of the mean. No lacewing larvae recovered no lacewing larvae (fig. I), de- Direct observations in the field in were recovered from fields WS1, K, S2 spite the fact that in several of the fields 1991 and 1992 revealed that generalist or S3. WS = West Side Field Station; K cotton aphid prey were abundant. This predators, including nymphal big-eyed = Kearney Agricultural Center; S = UC suggested that some factor, other than bugs and nymphal and adult assassin Cotton Research Station, Shafter.

8 CALIFORNIA AGRICULTURE, VOLUME 47, NUMBER 5 larvae (second instar); (6) aphids plus when other predators were added. AI- logical control researchers. Densities of five lacewing larvae plus two big-eyed though big-eyed bugs had a minimal ef- lacewing eggs are naturally high in San bug adults; (7) aphids plus five lacewing fect (P = 0.15), the addition of damsel Joaquin Valley cotton fields; thus, aug- larvae plus two damsel bug adults; (8) bugs allowed aphid populations to start mentative releases of lacewings should aphids plus five lacewing larvae plus growing again (P = 0.01),and the addi- only be contemplated after field scout- two assassin bug adults. Lacewing lar- tion of assassin bugs appeared to re- ing has shown that natural populations vae were obtained from a commercial in- move most of the suppressive effect of are not present in high densities. Our sectary. After 7 to 8 days, the plant the lacewings (P = 0.001; fig. 3b). small plot releases suggest that releases stems were cut and the sleeve cages Clearly, these predators did not act in an of lacewing eggs at extremely high rates brought into the laboratory, where all additive way to suppress aphid num- (much higher than is commercially vi- predators and aphids present were bers. Rather, the heavy by able given the cost of $3 to $4 per 1,000 counted. damsel bugs and assassin bugs on eggs) produced only modest suppres- Results. Survival of lacewing larvae lacewing larvae allowed aphids to es- sion of aphids. Recommended release varied strongly across the four treat- cape the effective biological control pro- rates would probably not produce de- ments to which lacewings were added duced by the lacewings alone. This re- tectable results. (fig. 3a). Survival decreased from 47% in sult was surprising, especially for The key reason why naturally present the absence of predators to 20% when damsel bugs, which are known as major populations and augmentatively re- big-eyed bugs were present (P = 0.06), predators of aphids. leased lacewing eggs do not control and to 0% when either damsel bugs aphids appears to be the heavy mortality (P = 0.001) or assassin bugs were Conclusions imposed on lacewing larvae by other present (P = 0.003). We can infer that Biological control of the cotton aphid generalist predators. Because our con- the decreased lacewing survival was a by generalist predators involves com- clusions are based on only 2 years of result of predation rather than competi- plex interactions among predators that study, additional work is required to de- tion for aphid prey, because the avail- have not previously been studied by bio- termine how general our results are. We ability of aphid prey was greater where need to understand fully the interactions predatory bugs were present. This nega- between lacewings and other generalist tive effect of the generalist predators on predators to predict late-season out- lacewing survival influenced the level of breaks of the cotton aphid. biological control of the cotton aphid Can we manipulate this system to in- (fig. 3b). The first five treatments shown crease the effectiveness of lacewings? in figure 3b provide a test of whether We do not wish to suppress populations each of the four predators considered of big-eyed bugs or damsel bugs because alone has an impact on aphid population these predators probably help control growth. When aphids were caged alone, other cotton pests, including Lygus bugs, their populations expanded rapidly, as spider mites and worms. Thus, we may expected. Assassin bugs and big-eyed need to search for other biological con- bugs alone exerted minimal influences trol agents for the cotton aphid that are on the rate of aphid population growth not susceptible to these predators, or de- (P > 0.10). Damsel bugs slowed aphid velop other control tactics, such as host population growth substantially (P = 0.03), plant resistance or cultural techniques. but did not control the populations. Lace- Research continues on all of these fronts. wing larvae were, however, able to cause aphid populations to decrease rather than expand P = 0.001); the me- 1. A. Rosenheim is Assistant Professor and dian response was a 91.2% decrease of L. R. Wilhoit was Postgraduate Researcher aphid numbers. (currently an Environmental Research This highly successful level of bio- Scientist for the Department of Pesticide logical control produced by lacewing Regulation, California EPA), Department larvae feeding alone was disrupted of Entomology, UC Davis. The authors thank Christine Armer and Paul Wynholdsfor technical assistance; Thomas Leigh for the use offield cages; Kent Daane, Les Ehler and Thomas Leigh for helpful discussions; and Sinthya Penn and William Lalor for comments on this manu- script. They thank the staffof the Kearney Agricultural Center and the UC Cotton Re- search Station, Shafler,for growing the experi- mental cotton plantings, and the Rincon- Fig. 3. Influences of generalist predators Vitoua lnsectay for providing lacewing eggs on (A) lacewing larval survival and (B) per- for testing. capita aphid population growth rates. Per- capita aphid population growth was calcu- This study was funded by cotton-grower lated as (final aphid numbers - initial funds made available through Cotton lncor- Fig. 2. Field experiment assessing the aphid numbers)/(initial aphid numbers); a porated, in cooperation with the California efficacy of augmentative releases of green value of 0.0 indicates no change in aphid State Support Committee, Rick Wegis, lacewing eggs for biological control of the populations. Aphids were present in all chairman, and the University of California cotton aphid. treatments. Statewide lPM Project.

CALIFORNIA AGRICULTURE, SEPTEMBER-OCTOBER 1993 9