FILE: <bc-46.htm>                                                                                                                                                                                       GENERAL INDEX     [Navigate to   MAIN MENU ]







----Please CLICK on desired underlined categories [to search for Subject Matter, depress Ctrl/F ]:



Differences in Population Ecology

Documentation Bias

Food Limitation

Differential Preadaptation

Life History Characteristics

Differences in Detoxification Enzymes


Differences in Intrinsic Tolerance

Genetic Systems

Differences in Genetic Variation.Fitness Cost

Van Driesche & Bellows (1996) Account

Fitness Costs




          A review of natural enemy resistance to pesticides by Tabashnik & Johnson (1999) considered that beneficial species account for less than 3% of the 447 species of insects and mites known to be resistant to one or more pesticides (Georghiou 1986). Therefore, documented cases of resistant pest species outnumber those in natural enemies by more than 30 to 1. Reasons for this disparity are due to (1) biases in documentation, (2) differential preadaptation to pesticides and (3) differences in population ecology. Tabashnik & Johnson emphasize that the various hypotheses attempting to explain why reported cases of resistance in pests greatly outnumber those in natural enemies are not mutually exclusive. Combinations of factors may operate in a given case and the relative importance of each factor may vary among species. Therefore, there is no single explanation.

          Pesticide resistance is a genetically based, statistically significant decrease in response of a population to a pesticide. The measured response is usually acute mortality, but changes in sub lethal or long-term responses are not excluded. Pesticide resistance may be demonstrated by a decrease in mortality through time for a given population, or by decreased mortality of a population relative to conspecific populations. Unlike some definitions of resistance (Georghiou 1981), that of Tabashnik & Johnson (1999) does not specify the extent of change in response to the pesticide nor does it imply the ability to survive field applications of pesticide.

          Various hypotheses proposed to explain the scarcity of documented cases of resistance in natural enemies are not mutually exclusive. The relative importance of each factor may vary among species and among pesticides; and several factors may act jointly in some cases. Thus, there is no single general explanation. A review of available evidence casts doubt on some hypotheses, supports others and indicates potentially productive areas for research. Resistance is more likely to be documented for pests than natural enemies, but the magnitude of this effect is difficult to measure. Bioassays comparing interpopulation variation for several pests and natural enemies from a given crop and region could help to assess the influence of documentation bias.

Natural enemies apparently do not lack the detoxification enzyme systems found in pests and intrinsic levels of detoxification enzymes are not consistently higher for pests than for natural enemies. For cases in which pests do have higher levels of detoxification enzymes than natural enemies, there is little evidence showing that this difference contributed to faster evolution of resistance in the pest. Versions of a preadaptation hypothesis based on intrinsic differences in detoxification ability appear unsound. The overall differences in intrinsic pesticide tolerance between pests and natural enemies are difficult to assess, but there are many cases in which pests are more tolerant than natural enemies. Higher intrinsic tolerance in pests could account in part for more documented cases of resistance in pests, particularly if the criteria for resistance include the ability to survive field applications of pesticides. Lack of genetic variation for pesticide tolerance in natural enemies could also retard their evolution of resistance.

There is indirect evidence to suggest that differences in population ecology are important in slowing resistance development in natural enemies relative to pests. However, differences in genetic systems and related factors are not apt to explain why pests evolve resistance more readily than natural enemies. Food limitation due to reduction in host or prey populations by pesticides is believed to be a major factor influencing natural enemy populations.  The decimation of arthropod populations in insectiicide treated apple orchards is known to be enormous (Please refer to Related Research ).  If this is true then there are some important implications for management. Natural enemies that are provided abundant food in artificial selection programs should be capable of evolving pesticide resistance (Hoy 1985, 1989).   Also, intensive pesticide use may disrupt biological control, even if the natural enemies are relatively tolerant (either naturally or due to selection) (Tabashnik 1986). Therefore, in order to maintain effective biological control, the use of selective pesticides should be sparing and judicious.Following is a detailed review of some of these considerations:

Documentation Bias

Pest resistance to pesticides can cause control failures and attract immediate attention. In contrast, natural enemy resistance to pesticides does not obviously create problems and may go unnoticed. Therefore, pesticide resistance is more likely to be documented in pests than in natural enemies (Georghiou 1972, Croft & Brown 1975).

It has been noted that if resistance in natural enemies appears rare due to inadequate documentation, then systematic testing of samples of natural enemies in heavily treated ecosystems should detect more cases of resistance (Croft & Brown 1975). According to extensive surveys, data available on pesticide impact on natural enemies more than doubled from 1970 to 1985 (Theiling & Croft 1988) and the number of natural enemy species reported as resistant to one or more pesticides also doubled during the same period (Georghiou 1972, 1986). The proportion of cases of resistance accounted for by beneficial species nevertheless remained at about 3% in the 1970 and 1984 surveys. These data show that the cumulative effect of factors contributing to more documented cases of resistance in pests than in natural enemies has remained consistent through time.

Cases of resistance were included in the previous mentioned surveys only if it was due to field application of pesticides and was sufficient to cause diminished mortality at field application rates (Georghiou 1981, 1986). Cases of resistance in natural enemies due to field and laboratory selection are reviewed elsewhere (Croft & Strickler 1983, Croft 1989, Hoy 1989). Before 1979 Croft & Strickler (1983) document cases of pesticide resistance among arthropod natural enemies.

Phytoseiid mites received considerable attention in the period after 1979, and had higher levels of resistance than other natural enemies. A significant difference in susceptibility between at least one pair of populations was reported in 77% of cases for phytoseiids. Maximum resistance ratios for phytoseiids exceed 10 in 40% of cases. Field survival, as indicated by a maximum LC50 greater than recommended application rates, was reported in 6 of 12 cases. For non-phytoseiid natural enemies, 74% of cases showed significant variation in susceptibility among conspecific populations, but maximum resistance ratios exceeded 10 in only 11% of cases and 35% of cases showed field survival.

Five of six cases showing ability to survive field concentrations of pesticide in non-phytoseiids was due to Chrysoperla carnea (Stephens). The ability to survive field rates in these five cases represents natural tolerance rather than resistance, because LC50's of the most susceptible populations exceeded recommended field concentrations. In fact, four of the five cases were due to tolerance to pyrethroids that were not registered for field use when the bioassays were performed (Grafton-Cardwell & Hoy 1985). Excluding the cases due to tolerance in C. carnea, field survival was reported in only one of 11 (9%) cases for non-phytoseiid natural enemies (Aphidoletes aphidimyza (Fondani) vs. azinphosmethyl).

Almost all studies of Hymenoptera have shown significant variation in susceptibility among populations (92%), but high levels of resistance were virtually absent. Only Comperiella bifasciata Howard had a resistance ratio greater than 10, and no Hymenoptera population had an LC50 greater than recommended field rates. The few cases reviewed here suggests that significant variation in susceptibility to pesticides among conspecific populations is common in natural enemies, but resistance conferring survival at field application rates is rare in natural enemies other than phytoseiid mites. Phytoseiids may evolve resistance readily because they have many generations per season, they are exposed to pesticides in all life stages, they have limited dispersal, and they subsist on plant materials when prey density is low (Georghiou 1972, Croft & Brown 1975, Hoy 1985, Croft & van de Baan 1988).

Differential Preadaptation Hypothesis

The overall pattern shows that indeed pests evolve resistance more readily than natural enemies, with several factors involved in slowing development of resistance in natural enemies. Four hypotheses are (1) that pests have better intrinsic detoxification capabilities than natural enemies (Gordon 1961, Croft & Morse 1979, Croft & Strickler 1983), (2) pests evolve resistance more readily because they have greater intrinsic tolerance to pesticides than natural enemies, (3) pests have more genetic variation in tolerance to pesticides than natural enemies (Georghiou 1972), and (4) the fitness cost associated with resistance is lower for pests than natural enemies.

These explanations are related and not mutually exclusive. All have two parts (1) there is some intrinsic (i.e., before pesticide exposure) difference between pests and natural enemies in their response to pesticides and (2) the intrinsic difference enables pests to develop resistance more readily than natural enemies.

Differences in Detoxification Enzymes.--Croft & Strickler (1983) review different detoxification enzymes. The idea is that herbivorous pest insects have evolved enzymes to detoxify plant secondary compounds and are then better preadapted to detoxify pesticides than are entomophagous natural enemies which do not eat plants. In the broadest survey to date, Brattsten & Metcalf (1970) used the synergistic ratio of carbaryl with piperonyl butoxide to compare levels of mixed function oxidase (MFO) detoxification enzymes in 53 insect species from 8 orders. Piperonyl butoxide inhibits MFO enzymes that detoxify carbaryl. Thus, the ratio of the LD50 or LC50 of carbaryl alone to that of carbaryl plus piperonyl butoxide (synergistic ratio) is an indicator of MFO activity.

Brattsten & Metcalf (1970) found that a few of the phytophagous species had extraordinarily high synergistic ratios (e.g., Pogonomyrmex barbatus (F.Smith)), the red harvester ant ratio >223), but the median synergistic ratio for 25 herbivores (3.6) was less than half the median for 15 entomophagous species (8.4). The proportion of species with low synergistic ratios (below 4.8) was significantly greater for herbivores (17/25) than for predators and parasites (3/15). The median synergistic ratios for 13 crop pests (2.3) and 27 herbivores and non-herbivorous pests (4.2) were also significantly lower than the median for entomophagous species. These findings contract the first part of the preadaptation hypothesis.

Considering the second part of the hypothesis, natural enemies that have not been reported as resistant to insecticides had greater synergistic ratios than several pests that are notorious for their ability to evolve insecticide resistance. For example, the synergistic ratios for six of seven species of parasitic Hymenoptera studied were higher than the ratios for the German cockroach, corn earworm, fail armyworm and the southern house mosquito. Thus, the data of Brattsten & Metcalf (1970) do not support the idea that phytophagous pests have higher levels of detoxification enzymes than natural enemies. Furthermore, their data suggest that low MFO activity, as indicated by the synergistic ratio of carbaryl with poperonyl butoxide, does not limit evolution of pesticide resistance. Pests with low synergistic ratios have evolved resistance whereas natural enemies with relatively high ratios have avoided resistance.

These conclusions were recently confirmed by Strickler & Croft (1985) in work that showed that piperonyl butoxide affected a predatory mite (Amblyseius fallacis) more than its herbivorous prey mite (Tetranychus urticae). Additionally, the synergistic ratio of piperonyl butoxide with propoxur for adults of the ectoparasitic braconid Oncophanes americanus (Weed) (14.0) was 8, 2 and 12-fold greater than the synergistic ratio for adults, large larvae and medium larvae, respectively, of its herbivorous tortricid host, Artyrotaenia citrana (Fernald) (Croft & Mullin 1984). The synergistic ratio for large larvae of the ectoparasite (5.5) was more than triple the synergistic ratio for adults or medium larvae of the host, but it was slightly less than the synergistic ratio of large larvae of the host (6.0).

Croft & Mullin (1984) stated that "synergist tests are useful in qualitatively estimating the availability of an MFO detoxification pathway," but they questioned the value of poperonyl butoxide synergist tests as an indicator of MFO activity across diverse arthropod species. Therefore, it is important to consider other measures of enzyme activity, such as results from in vitro assays.

In vitro inhibition studies and hydrolysis assays showed that lacewing larvae, C. carnea, have unusually active esterases that detoxify pyrethroids (Ishaaya & Casida 1981). Although direct comparative tests were not done, preparations of whole lacewing larvae were more active than larval gut or integument preparations from an herbivore, the cabbage looper, Trichoplusia ni (Hübner). Lacewing larvae were chosen for enzymatic tests because they have high natural tolerance to pyrethroids and thus were suspected to have high esterase activity (Plapp & Bull 1978). Consequently, the lacewing esterase activity data are a biased sample of natural enemy detoxification enzyme capacity.

Mullin et al. (1982) in a seminal study found in vitro detoxification enzyme differences between susceptible strains of a predatory mite, Amblyseius fallacis, and its herbivorous prey, Tetranychus urticae. MFO and trans-epoxide hydrolase were higher in the prey than in the predator, but the opposite was observed for cis-epoxide hydrolase and glutathione transferase. Esterase activity was similar in susceptible strains of the two species.  Even though the herbivorous mite had higher activity than the predatory mite for an MFO enzyme (aldrin epoxidase) and for trans-epoxide hydrolase, neither of these two enzymes were more active in resistant than in susceptible strains of the herbivore. These data suggest that resistance in the herbivore is not due to elevated levels of aldrin epoxidase or trans-epoxide hydrolase. It appears, therefore, that intrinsically higher levels of these two enzymes were not responsible for the ability of the herbivorous pest to evolve resistance more readily than its predator. On the other hand, of the five types of enzyme activity measured, only glutathione transferase was higher in resistant prey than in susceptible prey, which suggests that this enzyme could be partly responsible for resistance in the prey species. Contrary to the expectation of a preadaptation hypothesis, however, susceptible predators had more than 10-fold higher activity for this enzyme than did susceptible pests.

It appears that the detoxification enzyme differences between A. fallacis and T. urticae reported by Mullin et al. (1982) do not support the second part of the preadaptation hypothesis. It may be inappropriate to generalize from this case because predatory phytoseiid mites are the one group of natural enemies that has readily evolved pesticide resistance. Therefore, intrinsic disadvantages that slow evolution of resistance in other natural enemies may be absent or diminished in phytoseiids.

Croft & Mullin (1984) compared various life stages of a braconid ectoparasitoid, Oncophanes americanus (Weed) and its tortricid host, Argyrotaenia citrana (Fernald). Based on data from whole body preparations averaged across life stages, none of the five in vitro assays conducted showed significantly lower enzyme activity in the parasitoid compared to the host. Another in vitro study comparing adult midgut levels of three types of enzyme between a herbivorous coccinellid, Epilachna varivestis Mulsant, and a predatory coccinellid, Hippodamia convergens Guerin-Meneville, transepoxide hydrolase was virtually identical in the two species, but the predator a four-fold higher cis-epoxide hydrolase activity and 2.5-fold higher esterase activity than the herbivore (Mullin 1985). Thus this comparison between taxonomically similar beetles that differ in feeding habit does not support the first part of the pre-adaptation hypothesis.

Aldrin epoxidase and trans-epoxide hydrolase activity were lower in whole body preparations of Pediobius foveolatus (Crawford), a eulophid parasitoid, than in midgut tissues of its host, Epilachna varivestis (Mullin 1985). Because enzyme activity of midgut tissue can be more than 200-fold greater than activity from a whole body preparation of the same species (Croft & Mullin 1984), comparisons between activity in the midgut of a pest vs. the whole body of a natural enemy are not a direct test of the first part of the preadaptation hypothesis.

A review of the aforementioned in vitro studies of seven species and similar studies of four species of piercing-sucking herbivores (Mullin 1985) shows broad overlap in detoxification enzyme levels between herbivores and natural enemies. For instance, aldrin epoxidase, trans-epoxide hydrolase, and cis-epoxide hydrolase were lower in the herbivore Aphis nerii Fonscolombe than in any of the three natural enemies tested (A. fallacis, O. americanus, and P. fovelatus). Whole body levels of aldrin epoxidase for the two species listed as chewing herbivores (T. urticae and A. citrana) and the three natural enemies tested fell within the range of the three aphid species tested (A. nerii, Myzus persicae (Sulzer), and Macrosiphus euphorbiae (Thomas).

Perhaps the most detailed comparative study of detoxification abilities was given by Yu (1987) who measured 15 components of enzymatic detoxification capacity in the spined soldier bug, Podisus maculiventris (Say), and four of its noctuid prey, Heliothis virescens (Fab.), H. zea, Spodoptera frugiperda, and Anticarsia gemmatalis (Hübner). In vitro assays of midgut tissues from adult predators and final instar larvae of the lepidopterous prey showed that the predator lacked none of the detoxification enzyme systems that were found in the prey. The prey had greater detoxification capacity in 70% of pairwise comparisons between the predator and prey. For 11 of 15 components measured, either 3 or 4 of the prey species had greater activity than the predator. However, the reverse was true for cytochrome P-450, microsomal desulfurase, and glutathione transferase toward 1-chloro-2, 4-dinitro benzene (CDNB). Additionally, the proportion of the high-spin form of cytochrome P-450, thought to indicate allelochemical-oxidizing capacity, was higher in the predator than in any of the prey.. Topical bioassays showed that the predator was generally more susceptible to organophosphorous and carbamate insecticides, but was more tolerant of pyrethroids compared with its prey (Yu 1988). However, the organophosphate tetrachlorvinphos was much more toxic to two of the prey species, S. fruigiperda and A. gemmatalis, than to the predator. Although penetration of tetrachlorvinophos was slower for the predator than for S. frugiperda, injection tests and acetylcholinesterase inhibition tests implied that the predator's ability to survive tetrachlorvinphos was due to enhanced enzymatic detoxification. Yu (1988) concluded that the predator's high pyrethroid tolerance was probably due to reduced penetration or target site insensitivity.

The in vitro assays of Yu (1987) showing that the predator generally had lower enzymatic detoxification capacity than its prey support the first part of the preadaptation hypothesis, but several significant exceptions were also found. Later bioassays (Yu 1988) show that patterns of susceptibility to insecticides could not be reliably predicted on the basis of enzyme activities. Higher levels of certain detoxification enzymes such as desulfurase, may actually increase the toxicity of some insecticides.

Available data from in vivo and in vitro studies do not support the idea that natural enemies have consistently lower levels of all types of detoxification enzymes than pests, but this does not exclude the possibility of more subtle differences between the two groups. For instance, an alternative hypothesis is that oxidative detoxification enzymes are more active in pests than natural enemies, whereas hydrolytic detoxification enzymes are similar in pests and natural enemies (Plapp & Vinson 1977). The first part of this hypothesis, however, is contradicted by data from in vivo synergism tests and is not strongly supported by data from in vitro tests. Aldrin epoxidase data from a predatory mite and its herbivorous prey (Mullin et al. 1982) support the hypothesis, but data from whole body preparations of an ectoparasitoid and its lepidopterous host show the opposite trend (Croft & Mullin 1984). Furthermore, intraspecific variation in MFO enzymes among developmental stages sometimes even within an instar, can be greater than the typical differences found between pests and natural enemies (Gould 1984, Croft & Mullin 1984).

Hydrolytic enzymes such as esterases and epoxide hydrolases are often equally or more active in natural enemies compared with pests. Yu (1987) found, however, that hydrolytic enzyme activities were lower in a predator than in its prey in 88% of comparisons, but oxidative detoxification activities were lower in the predator than the prey in only 59% of comparisons.

An obvious trend is that the ratio of trans- to cis-epoxide hydrolase is, generally, higher for herbivores than natural enemies (Mullin & Croft 1984). The lower trans/cis ratio of natural enemies, however, does not explain their reduced ability to evolve resistance. The pests Blattella germanica and Myzus persicae, both well known for their ability to develop resistance, had lower trans/cis ratios than six of seven entomophagous species studied (Mullin 1985). Nevertheless, differences in trans/cis ratios and other biochemical differences might be useful in designing selective pesticides (Mullin & Croft 1984, 1985).

The detoxification enzymes found in pests are also present in natural enemies. Data from in vivo synergism tests show that MFO enzyme levels in natural enemies were not consistently lower than those in pests. In fact, evidence to data shows the opposite trend. Although in vitro enzyme assays have shown many cases in which pests have higher levels of detoxification enzymes than natural enemies, the reverse has been reported with nearly equal frequency. Therefore, the first part of the hypothesis is not generally supported. In those cases in which detoxification enzyme levels are higher in pests than in natural enemies, there is little or no evidence indicating that this difference contributed to more rapid evolution of resistance in the pest. Thus, there is little support for the second part of the hypothesis. Finally, the differential detoxification hypothesis does not address nonmetabolic resistance (such as reduced penetration and target site insensitivity) or resistance in nonherbivorous pests (Tabashnik 1986).

Differences in Intrinsic Tolerance.--The belief that natural enemies are intrinsically less tolerant to pesticides than pests is, effectually, a generalized version of the differential detoxification enzyme hypothesis. The concept is the same, but unlike the differential detoxification hypothesis, the mechanism causing the intrinsic difference is unspecified. Tests of this version of the preadaptation hypothesis must determine (1) whether pests have higher intrinsic tolerance to pesticides than natural enemies and if so then (2) whether such intrinsic differences in tolerance cause natural enemies to evolve resistance more slowly than pests. Two types of bias make it difficult to evaluate the first part of the hypothesis. First, widespread resistance and cross-resistance in pests can make it difficult to assess their intrinsic tolerance. Inclusion of resistant pest populations in surveys of tolerance would tend to inflate the tolerance of pests relative to natural enemies, which are less likely to be resistant. Second, researchers may concentrate efforts on pesticides that natural enemies can tolerate because such compounds are particularly useful in integrated pest management. These two biases operate in opposite directions, but their relative magnitude is not easily determined.

Brattsten & Metcalf (1970) studied the susceptibility to carbaryl, finding that some pests were very tolerant (e.g., formicid ants). However, median LD50 values did not differ significantly between 22 pests and 8 entomophagous species. Similarly, medians for carbaryl LC50 did not differ significantly between five pests. Surveys of studies, which compared the LC50 or LD50 values within natural enemy/pest complexes, also do not support the hypothesis that natural enemies are intrinsically more susceptible to pesticides than are their prey or hosts (Croft & Brown 1975, Theiling & Croft 1988). Croft & Brown (1975) found that natural enemies were more tolerant than their prey or host in 67 of 92 cases in which the same bioassay method was used to compare species. Although predators were usually more tolerant than their prey (63 of 77 cases), parasitoids were usually less tolerant than their hosts (11 of 15 cases).

Theiling & Croft (1988) calculated LC/LD50 ratios for 870 cases in which a natural enemy was compared to its prey or host. Ratios ranged widely, yet natural enemies were more tolerant than their prey or host in 57% of the cases. Furthermore, for 10 of 12 families of natural enemies, including Braconidae and Aphelinidae, the average ratio for the family showed that the natural enemy was more tolerant than its prey or host. Phytoseiidae were as tolerant as their prey and only Ichneumonidae were less tolerant than their hosts.

These results contradict the differential intrinsic tolerance hypothesis, but the authors of the reviews suggest some reasons why their surveys may make natural enemies appear more tolerant than they actually are. They note that LC/LD50 comparisons between pest/natural enemy pairs are available for only a small subset of the studies of pesticide impact on natural enemies. Natural enemies suspected to be tolerant to pesticides may be more likely to be compared to their hosts or prey in bioassays. Similarly, compounds thought to be more toxic to pests than natural enemies may be more likely to be tested in comparative studies.

Although some surveys suggest that natural enemies are not consistently less tolerant to pesticides than pests, they often are. Therefore, it is useful to consider how reduced intrinsic tolerance might affect evolution of insecticide resistance. If ability to survive field rates of pesticide is a criterion for resistance, then a natural enemy with lower intrinsic tolerance would have to increase its tolerance more substantially to be considered resistant.

Tabashnik & Croft (1985) did some simulations which suggest that under certain conditions reduced intrinsic tolerance could also slow evolution of resistance as measured by changes in the frequency of a resistance allele (R). Simulations based on a one locus model showed that 10 or 100-fold reduction in the LC50 of homozygous susceptible (SS) individuals had little impact on projected times for resistance development in natural enemies of apple pests. In contrast, reducing the LC50 of all three presumed genotypes (SS, RS, RR) by 10 or 100-fold greatly slowed resistance development in nearly all cases. When the LC50's of all three genotypes were reduced, heterozygous individuals were rendered functionally recessive, which slowed evolution of resistance (Curtis et al. 1978, Taylor & Georghiou 1979, Tabashnik & Croft 1982). The assumption of lower LC50's for all three genotypes implies that a resistance allele increases the LC50 by a fixed multiple. Thus, if the SS individuals are less tolerant, then so are RS and RR individuals. Lowering the LC50 of SS individuals without altering the same of RS or RR individuals assumes that a resistance allele provides a fixed level of tolerance, regardless of the tolerance of SS individuals.

In summary, surveys of comparative bioassay studies suggest that on the average natural enemies are not intrinsically less tolerant to pesticides than pests. Such surveys may be biased, however, and there are many cases in which pests are intrinsically more tolerant than natural enemies. Reduced intrinsic tolerance could retard evolution of resistance in some natural enemies if resistance is defined as the ability to survive field concentrations of a pesticide or if resistance alleles confer a fixed multiple of increased tolerance relative to susceptible individuals. The finding that Phytoseiidae, which are known for their ability to evolve resistance, had low selectivity ratios compared to other natural enemies (Theiling & Croft 1988) suggests that low intrinsic tolerance relative to pests is not a major impediment to evolution of resistance in natural enemies. The idea that natural enemies are not generally less tolerant to pesticides than pests differs from widely held perceptions, which may be based partly on observations that field applications of pesticides affect natural enemies more than pests. However, disruption of natural enemy populations by field applications of pesticides may be due to reduction of host or prey populations in addition to direct toxic effects. Indeed, mathematical models show that if a pest and its natural enemy are equally susceptible to a pesticide, the pesticide will have a more severe impact on the natural enemy population than the pest population (Wilson & Bossert 1971, Tabashnik 1986). Therefore, pesticide applications can be extremely detrimental to natural enemy populations even if the natural enemy's tolerance is similar to that of the pest.

Differences in Genetic Variation.--The genetic variation hypothesis maintains that natural enemies evolve resistance more slowly than pests because natural enemy populations have less genetic variation than pest populations (Huffaker 1971, Georghiou 1972). Surveys of electrophoretic data show that Hymenoptera have less variation in allozymes than most other insects. In particular the expected heterozygosity or average gene diversity for 13 species of wasps was less than half the expected heterozygosity for 158 species of Orthoptera, Homoptera, Coleoptera, Lepidoptera and Diptera (Graur 1985). These data are consistent with the idea that hymenopterous parasitoids have less genetic variation than herbivorous insects.

Although hymenopterous parasitoids have less allozymic variation than most other insects, including many pests, this measure of genetic variation may be unrelated to the ability to evolve insecticide resistance. Some pests known for their resistance development had high-expected heterozygosity (Heliothis virescens = 0.389, H. zea = 0.327, Lygus hesperus = 0.256) but others had very low values (Blattella germanica = 0.015, Myzus persicae = 0.000). The ability of some pests to readily evolve pesticide resistance, even though they display little or no electrophoretic heterozygosity, shows that this type of genetic variation is not a prerequisite for resistance development.

Genetic variation in tolerance to pesticides is required for evolution of resistance, but it has rarely been measured (Roush & McKenzie 1987, Tabashnik & Cushing 1988). How then can intrinsic differences in genetic variation influence resistance development? Single locus population genetic theory predicts that the rate of resistance evolution increases approximately linearly as the logarithm of the initial frequency of a resistance allele (May & Dobson 1986). In this sense large differences in initial resistance allele frequency have relatively small effects on rates of evolution. Most economically significant cases of pesticide resistance are thought to be under monogenic control, but there are also many examples of polygenic resistance (Roush & McKenzie 1987, Tabashnik & Cushing 1988). According to quantitative genetics theory, the rate of increase in pesticide tolerance would be directly proportional to the additive genetic variance (Via 1986). Therefore, large differences in additive genetic variance would have a major impact on resistance development.

Natural enemies (particularly parasitic Hymenoptera) may have less allozyme heterozygosity than pests, but among herbivores this index of genetic variation is not well correlated with the ability to evolve pesticide resistance. The extent of genetic variation in pesticide tolerance in pest and natural enemy populations is virtually unknown. Such variation could influence rates of evolution of resistance, but the importance of this factor cannot be assessed without more empirical information.

Fitness Cost.--It is often assumed that in the absence of pesticide a resistant individual is less fit than a susceptible individual. If this fitness cost of resistance were substantially greater for natural enemies than pests it might retard evolution of resistance in natural enemies. Review of data available for pests suggests that the fitness cost is generally not large, but it may depend on the nature of the resistance mechanism (Roush & McKenzie 1987). Little is known about the fitness cost of resistance in natural enemies. Studies of the predators Metaseiulus occidentalis and Chrysoperla carnea show little or no fitness cost associated with resistance (Roush & Hoy 1981, Roush & Plapp 1982, Grafton-Cardwell & How 1986), but data from other natural enemies are needed to evaluate this hypothesis more completely (Croft & Tabashnik 1989).

Differences in Population Ecology

Underlying the population ecology hypothesis is that concept that pesticide resistance evolves more readily in pests than natural enemies due to differences in population ecology between them. Several specific hypotheses are included in this general category: (1) resistance evolves more readily in pests because natural enemies suffer from food limitation following insecticide treatments (Huffaker 1971, Georghiou 1972); (2) differences exist between pests and natural enemies in life history traits (Croft 1982, Tabashnik & Croft 1985); (3) pests are more exposed to pesticides than natural enemies (Croft & Brown 1975); and (4) pests have different genetic systems than natural enemies (e.g., ploidy level).

As with the preadaptation hypotheses, each hypothesis has two parts (1) there is some difference in population ecology between pests and natural enemies and (2) the difference enables pests to develop resistance more readily than natural enemies.

Food Limitation.--The food limitation hypothesis is based on the population dynamics of interactions between natural enemies and pests (Huffaker 1971, Georghiou 1972, Croft & Brown 1975, Tabashnik 1986). The idea is that the few resistant pests surviving an initial pesticide treatment will have an abundant food supply. In contrast, resistant natural enemies surviving treatment will find their food supply (prey or host) severely reduced. Thus, resistance evolves more slowly in natural enemies because they starve, emigrate or have reduced reproduction following treatments that eliminate much of their food supply (Tabashnik 1989).

Pesticide treatments can reduce the food supply of natural enemies while leaving the pests' food supply intact. Thus, there is little question that pests and natural enemies differ in the way in which their food supply is affected by pesticides. However, it is difficult to determine the extent to which natural enemy populations are limited by the availability of prey or hosts after pesticide treatments. Pesticide applications reduce natural enemy populations, but in most cases, direct effects of the pesticides on a natural enemy are confounded with indirect effects of the pesticides on the natural enemy's food supply. Furthermore, the effect of food limitation on a natural enemy's ability to evolve resistance cannot be determined readily in the field.

The food limitation hypothesis could be tested directly by contrasting responses to pesticide treatments in a natural enemy population feeding on a susceptible strain of a pest vs. a population of the same natural enemy feeding on a resistant strain of the pest. The food limitation hypothesis predicts that pesticide resistance will evolve in the latter case but not the former. Although such a test has never been performed, there is other experimental, historical and theoretical evidence available to evaluate the food limitation hypothesis.

If food limitation is a major factor slowing evolution of resistance in natural enemies then it might be predicted that (1) natural enemies will evolve resistance readily when provided with abundant food in artificial selection programs; and (2) that if natural enemies can use food sources that are not greatly reduced by pesticides (e.g., plants or resistant pests), they will evolve resistance more readily than those that specialize on susceptible pests. The successful laboratory selection for pesticide resistance in natural enemies (Croft & Strickler 1983, Hoy 1985) provides some support for the food limitation hypothesis, yet laboratory selection may not produce high levels of resistance in natural enemies as readily as in pests. Such differences are difficult to assess, but they could be due to intrinsic limitations of natural enemies or to technical problems associated with sampling, rearing and selecting large numbers of natural enemies.

There is only one study that directly compared evolution of pesticide resistance in a pest and its natural enemy in the laboratory. Croft & Morse (1981) contrasted responses to selection for resistance to azinphosmethyl in a predatory mite, Amblyseius fallacis, and its prey, Tetranychus urticae. A susceptible strain of the predator initiated from only a few individuals did not evolve resistance after seven selections. However, a composite susceptible strain initiated with 600 adult females from three predator strains, developed an 80-fold resistance in 22 selections. Another concurrent experiment with a susceptible non-composite strain of the pest showed only a 20-fold increase in LC50 in 22 selections. In the only experiment to compare resistance development in the predator and prey with susceptible strains of both species in contact, the composite strain of the predator evolved 80-fold resistance whereas the pest strain evolved only a five-fold resistance. Such results suggest that the pest did not develop resistance more readily than its predator, but interpretations are complicated by several factors. First a non-composite pest strain developed resistance more readily than a noncomposite predator strain. Second, even though the proportional increase in LC50 was higher for the predator than the pest in some experiments, the final LC50 after selection was always higher for the pest because of the pest's initially higher LC50. Third, the food limitation hypothesis was not tested directly because predators fed on leaf nectaries and survived even in the absence of prey. Therefore it was not possible to determine if lack of food would retard evolution of resistance in the predator.

The ability of phytoseiid mites to derive nourishment from plant materials and their ability to evolve resistance in the field (Georghiou 1972, Croft & Brown 1975) is consistent with the second prediction from the food limitation hypothesis. Also consistent with this prediction is the general pattern that natural enemies usually become resistant only after their prey or host develops resistance (Georghiou 1972, Croft & Brown 1975, Tabashnik & Croft 1985).

It was concluded that food limitation was not a key factor affecting resistance development in Aphytis melinus (Rosenheim & Hoy 1986). It was noted that a non-resistant host population can survive pesticide applications if it is not contacted by treatments or if it is naturally tolerant. Thus, treating the periphery of citrus trees with dimethoate to control citrus thrips or with chlorpyrifos to control orangeworms should not severely reduce populations of California red scale which are distributed throughout the tree. Additionally, California red scale populations are naturally tolerant of relatively low concentrations of dimethoate used to control citrus thrips. Therefore, food limitation should not restrict development of resistance to dimethoate or chlorpyrifos in Aphytis melinus. On the other hand, food limitation should slow development of resistance in A. melinus to carbaryl, malathion, and methidathion because these insecticides are used to control scales .           An analysis of the range in LC50's showed that resistance in A. melinus to dimethoate and chlorpyrifos was not consistently greater than resistance to carbaryl, malathion and methidathion. These results support the conclusion that food limitation was not a major determinant of rates of resistance development in the parasitoid. Rosenheim & Hoy (1986) noted that this is a limited test of the hypothesis because many factors other than food limitation (e.g., dross-resistance and variation among insecticides in the duration and extent of use) could have affected the outcome.

The potential impact of food limitation on the population dynamics and ability to evolve resistance of natural enemies has been tested with mathematical models. According to the Lotka-Volterra equations of predator-prey population growth, equivalent mortality will suppress a predator population more than its prey (Wilson & Bossert 1971). This occurs because the predator's birth rate and the prey's death rate are proportional to the product of the population sizes of both species. On the other hand, the predator's death rate and the prey's birth rate are not affected by the population size of the other species. Thus, a pesticide treatment that kills 90% of predator and prey populations reduces the predator's birth rate and the prey's death rate by a factor of 100, but reduces the predator's death rate and the prey's birth rate only by a factor of 10. More refined models also show that natural enemy populations are more severely suppressed by pesticides than are pest populations, even though the immediate mortality is similar for both populations (Waage et al. 1985).

Considering whether the suppression of natural enemy populations affects their ability to evolve resistance, May & Dobson (1986) emphasized the general distinction between overcompensating and undercompensating density-dependence. Pests generally rebound above their long-term average or equilibrium levels following pesticide treatments and thus show overcompensating density-dependence. On the contrary, natural enemies recover slowly, showing undercompensating density-dependence. Undercompensating density-dependence reduces the average population size, thereby increasing the impact of immigration of susceptible individuals (Comins 1977, Taylor & Georghiou 1979, Tabashnik & Croft 1982). Therefore, in the presence of immigration, pests with overcompensating density-dependence will develop resistance faster than natural enemies with undercompensating density-dependence (May & Dobson 1986).

Simulation studies of 12 natural enemies or orchard pests showed that incorporation of a simplified version of the food limitation hypothesis substantially improved the correspondence between predicted and reported times for resistance development (Tabashnik & Croft 1985). Natural enemies were assumed to begin evolving resistance only after their prey or host had become resistant. Although the results supported the food limitation hypothesis, this approach oversimplified dynamic ecological and evolutionary processes. Other simulation studies included evolutionary potential for resistance in both predator and prey, as well as coupled predator-prey population dynamics (Tabashnik 1986). The key assumption of these simulations was that low prey density reduced the predator's rates of consumption, survival and fecundity. Predator functional response and the effects of food shortage on predator survival and fecundity were partly based on experimental data from mites (Dover et al. 1979). Even though the predator and prey were assumed to have equal intrinsic tolerance and equal genetic potential for evolving resistance, intensive pesticide use caused rapid resistance development in the pest (prey), but either suppressed resistance development or caused local extinction of the natural enemy (predator). These theoretical results imply that food limitation is sufficient to account for pests' ability to evolve pesticide resistance more readily than natural enemies.

Pesticide treatments reduce the food supply of natural enemies more than of pests. Severe reductions in food supply can slow resistance development in natural enemies. Indirect support for the food limitation hypothesis is provided by the success of laboratory selection programs in which natural enemies are provided abundant food, by the general trend that natural enemies evolve resistance only after their prey or host becomes resistant, and by theoretical models. However, food limitation does not seem to explain patterns of development of resistance to various insecticides in Aphytis melinus, a parasitoid of the California red scale.

Life History Characteristics.--Included here are the number of generations per year,a the rate and timing of reproduction, survivorship, development rate and sex ratio. Theoretical work suggests that the rate of resistance development increases as reproductive capacity increases, particularly the number of generations per year (Tabashnik & Croft 1982, May & Dobson 1986). Historical patterns show a positive correlation between the number of generations per year and rate of resistance development (Tabashnik & Croft 1985, Georghiou & Taylor 1986, May & Dobson 1986). Variation in life history traits is sufficient to explain variation in rates of resistance development among apple orchard pests and their  enemies (Croft & Stickler 1983, Tabashnik & Croft 1985). If reproductive capacity, especially the number of generations per year, is consistently higher for pests than natural enemies, then this might explain why pests evolve resistance more readily than natural enemies.

This may be examined for 24 species of apple pests and natural enemies (Tabashnik & Croft 1985). Although the average number of generations per year was slightly higher for pests than natural enemies, there was also a broad overlap. Seven of 12 pest species had less than three generations yearly, but only five of 12 natural enemies had fewer than three generations per year. Generations per year ranged widely for each group. Thus no consistent difference between pests and natural enemies in generations per year is evident. The apple arthropods may be a biased sample because such a high proportion of the natural enemies have developed resistance. Broad surveys of fecundity in parasitic Ichneumonidae and Tachinidae show a range from 20 to 5000 eggs per female within these groups (Price 1984).

Exposure.--Rate of resistance development is a function of selection intensity, which is determined in part by the extent of exposure to pesticides. Croft & Brown (1975) hypothesized that natural enemies are often less intensively selected than pests because pesticides are directed at pests; natural enemies contact pesticides only because they occupy the same habitat as their prey or host. However, mobile predators and parasitoids might contact more toxicant and thus suffer greater mortality from residual deposits than would sedentary pests in the same habitat (Croft & Brown 1975).

Genetic Systems.--Evolutionary considerations about pesticide resistance are generally based on the assumption that organisms are diploid and sexually reproducing, but insects and mites have a variety of genetic systems. Differences in such systems between pests and natural enemies could influence their relative ability to evolve resistance. For instance, many parasitic Hymenoptera are haplo-diploid. In a modeling study, however, resistance evolved faster under haplo-diploidy than diplo-diploidy (Horn & Wadleigh 1987). Thus, haplo-diploidy does not seem to be a factor slowing evolution of resistance in parasitic Hymenoptera  (Related Research).

Various genetic systems are known in phytoseiid predatory mites, including thelytoky, parahaploidy (embryos of both sexes are diploid, but males lose one chromosome set during embryonic development), and something akin to arrhenotoky (unfertilized eggs produce haploid males, fertilized eggs produce diploid females) (Hoy 1985). Parahaploid phytoseiids such as Metaseiulus occidentalis and Phytoseiulus persimilis Athias-Henriot may have some advantages of both haploidy (exposing haploid individuals to selection) and diploidy (recombination) (Hoy 1985). This might explain why phytoseiids evolve resistance to pesticides more readily than other natural enemies, but it does not support the idea that the genetic systems of natural enemies retard their resistance development.

Other factors that could affect evolution of resistance in pests and natural enemies are the extent of sexual vs. asexual reproduction, inbreeding and coloniality. Colonial insects would not be expected to evolve resistance readily because they have small effective population size (few reproductives), slow generation turnover, and their reproducing individuals usually have limited exposure to pesticides. These factors may explain the paucity of documented cases of resistance in social Hymenoptera and Isoptera (Georghiou 1981), many of which are pests. However, they do not explain the lack of resistance in nonsocial parasitic Hymenoptera. Genetic systems and related factors may influence resistance development in pests and natural enemies, but it does not seem that there are consistent differences in these traits that would favor resistance development in pests compared to natural enemies.

Arthropod Resistance to Pesticides

(Van Driesche & Bellows (1996) Account)



Use of Agricultural Chemicals

Pesticide-induced Mortality

Effects of Plant Properties on Natural Enemies



Van Driesche & Bellows (1996) observed that pesticide resistance develops in a population when certain individuals possess genes which allow them to better avoid or survive contact with pesticides. Treating such a population with a pesticide confers differentially greater survival or fitness on these tolerant individuals, and the frequency of the resistant genotype increases when the tolerant individuals reproduce. For species in which these surviving individuals remain together as a new breeding group, undiluted by addition of susceptible individuals from outside the pesticide-treated area, pesticide resistance may develop. An increasing number of pests of several types have become resistant to pesticides. When pests develop resistance, agriculturists may respond by increasing dosage, changing or alternating pesticides, or combining several pesticides. If resistance is sufficiently severe to prevent control of the pest, chemical control may be abandoned and management systems based on biological control, including the conservation of native natural enemies, may finally be implemented instead. On the other hand, when natural enemies develop resistance to pesticides commonly used on a crop, this resistance may make it possible to conserve such natural enemies as important mortality agents contributing to the control of pests in crops even with continued pesticide use.

Use of Agricultural Chemicals

Pesticides can reduce natural enemy effectiveness either by directly causing mortality or by influencing the behavior, foraging, or movement of natural enemies, their relative rate of reproduction compared to that of the pest, or by causing imbalances between host and natural enemy populations such as catastrophic host synchronization (Table 7.1) (Jepson 1989; Waage 1989; Croft 1990).

Pesticide-induced Mortality

Many classes of pesticides are directly toxic, to one degree or another, to some categories of natural enemies. Insecticides and acaricides, for example, are likely to be damaging to most parasitoids and predacious anhropods (Bartlett 1951, 1953, 1963, 1964b, 1966; Bellows and Morse 1988, 1993; Bellows et al. 1985, 1992a, 1993; Morse and Bellows 1986; Morse et al. 1987), while fungicides would generally not affect these organisms but may inhibit fungi pathogenic to pest arthropods (see Yasem de Romero 1986; Saito 1988; Majchrowicz and Poprawski 1993) or fungi antagonistic to plant pathogens (Vyas 1988). Agricultural chemicals such as soil sterilants drastically alter soil microbial, fungal, and invertebrate communities, affecting the influence of such soils on plant pathogens. Other materials may be toxic outside of their intended scope of use, A bird repellent, for example, may also be insecticidal. A fungicide may also kill arthropods (sulfur is damaging to phytoseiid mites) or affect their reproduction or movement. Herbicides may kill beneficial nematodes applied for insect control (Forschler et al. 1990). It is thus important to assume that any pesticide, of whatever type, might affect a natural enemy until data are available to demonstrate that it does not (Hassan 1989a). Even materials often thought of as nontoxic, such as soaps or oils, which may be safe to humans, may be harmful to natural enemies. Oils may reduce the emergence of parasitoids of scale insects as well as cause scale mortality (Meyer & Nalepa 1991).

The degree of effect on a natural enemy population caused by any given pesticide will depend on both physiological and ecological factors. Physiological selectivity consists of the intrinsic relative toxicities of the compound to the pest and the natural enemy. Chemicals vary greatly in their inherent toxicity to a species (Jones et al. 1983; Smith & Papacek 1991). Some insecticides or acaricides have been found that are effective against pests and also mostly harmless, physiologically, to some arthropod natural enemies. Toxicity varies with species of natural enemy, but some examples include primicarb, toxins of Bacillus thuringiensis, fenbutatin oxide, and diflubenzuron (Hassan 1989a); certain plant alkaloids, mevinphos, and cryolite (Bellows et al. 1985; Bellows & Morse 1993); avermectin and narrow range oils (Morse et al. 1987); and the systemic materials demeton and aldoxycarb (Bellows et al. 1988). Ecological selectivity results from those aspects of the use of the material that determine the degree of contact that actually occurs between the pesticide and the natural enemy (Van Driesche & Bellows 1996). Contact is affected by the formulation and concentration applied, the persistence of the material in the environment (as affected by such abiotic factors as temperature and rainfall), the mode of action of the chemical (contact vs. Ingestion), the spatial pattern of application, and the timing of application (Van Driesche & Bellows 1996).

Effects of Plant Properties on Natural Enemies

The characteristics of plants influence natural enemies in a wide range of ways, some of which are just beginning to be recognized. It is common for a particular host species to be subject to varying levels of attack by natural enemies when it feeds on different plant species. Plant features influence foraging success and reproductive fitness of natural enemies. Knowledge of such interactions is important for conducting foreign exploration (to obtain species or strains of natural enemies adapted to attack the target pest on the target crops); for planning conservation and augmentation programs; and for guiding changes made to crop plants through plant breeding (so that natural enemy action is enhanced rather than reduced during creation of new crop varieties, (see Boethel and Eikenbary (1986). Strong et al. (1984) scussed relationships between plants and insects and concluded that competition for space from natural enemy attack was important in shaping herbivore communities. Price (1986) presented a classification of how plants can affect natural enemies through either semiochemically mediated effects, chemically-mediated effects, or physically-mediated effects.

In some instances, plant compounds are used directly by natural as cues for habitat location. In other instances, plant compounds may either be hidden by herbivores or may be released by plants under herbivore attack, resulting in the attraction of natural enemies directly to the host. At the level of plant communities, associated plants may produce compounds that either enhance attractiveness to natural enemies or may make the detection of host plants by natural enemies more difficult (Van Driesche & Bellows 1996).

Chemically mediated effects of plants on natural enemies include the use by natural enemies of plant products such as nectar and pollen as food sources. Nutritional qualities of plants also affect natural enemies indirectly by influencing the rate of growth and survival of herbivores which feed on them. Herbivores, which develop on plants of reduced nutritional quality, are likely to require a longer period to develop and may remain in stages susceptible to natural enemy attack longer than herbivores developing on more nutritious hosts. Plant qualities that cause mortality to an associated herbivore affect rates of the herbivore's survival in its various life stages. This in turn will affect the survival of immature parasitoids associated with the stage of the herbivore, and also will affect the number of hosts in subsequent stages that are available for other parasitoids or predators to attack (Van Driesche & Bellows 1996). Finally, plant compounds may be sequestered by herbivores and used as defenses against natural enemy attack. Protection of monarch butterflies (Danaus plexippus (L.)) from predation by birds through sequestering by monarch larvae of plant-derived cardiac glycosides is a well-known example (Brower 1969).

Physical aspects of plants may affect natural enemies in several ways. Spatial dispersion of plants can affect the ability of herbivores and natural enemies to locate and, in some cases, successfully colonize the plants. Plant structures can provide herbivores with physical protection. Insects in the centers of large fruits or galls, for example, are less accessible to parasitoids with short ovipositors. Some plant features may directly shelter natural enemies. Domatia (pits and pockets) on leaves are used by phytoseiids and plants with such features harbor higher numbers of these predatory mites (Walter and O'Dowd 1992; Grostal and O'Dowd 1994).

Plant festures such as leaf toughness or hairiness, which in some cases defend plants against herbivores, may also affect natural enemies. Increased trichome density on leaves is associated with reducing walking speed and lowered rates of foraging, making some natural enemies less effective in finding hosts (Hua et al. 1987). Some predators, such as chrysopid larvae, may also experience reduced walking speed on hairy leaves (Elsey 1974), but other predators, such as larvae of the coccinellid Adalia bipunctata, forage for prey more effectively on leaves with single scattered hairs, versus glabarous, waxy leaves. This because hairs force more frequent thrning and cause the larvae to move across the leaf surface rather than to only follow the veins and leaf edge (Shah 1982). Some phytoseiids have been found to be more abundant on grape varieties with hairy undersides, perhaps because of more favorable microclimate and protection from rain (Duso 1992). The influence of leaf pubescence on entomophagous species was reviewed by Obrycki (1986). In a broad sense, the shape of plant leaves and the arrangement of branches and other plant parts affect how natural enemies structure their foraging on the plant and between sets of plants (Ayal 1987; Grevstad & Klepetka 1992). Coccinellid larvae of several species, e.g., drop off leafless pea plants (Pisum spp.) Less frequently than from normal plants because tendrils of leafless plants are easier for beetles to grasp (Kareiva & Sahakian 1990). Plant canopies in which leaves overlap are associated with increased rates of dispesal of coccinellids, in the absence of prey, than crops in which plant canopies are discrete (Kareiva & Perry 1989). Parts of plants with distinct characteristics may be searched differently by species of natural enemies. Bananas (Musa spp.), e.g., are searched differently for banana aphid (Pentalonia nigronervosa Coquillet) by lysiphlebus testaceipes (Cresson), which searches open surfaces but avoids concealed areas, incontrast to Aphidius colemani Viereck, which searches in both zones (Stadler & Völkl 1991).


REFERENCES:       [Additional references may be found at:  MELVYL Library]

Bellows, T. S., Jr. & T. W. Fisher, (eds) 1999. Handbook of Biological Control: Principles and Applications. Academic Press, San Diego, CA.  1046 p.

Brattsten, L. B. & R. L. Metcalf. 1970. The synergistic ratio of carbaryl with poperonyl butoxide as an indicator of the distribution of multifunction oxiddases in the insecta. J. Econ. Ent. 63: 1347-48.

Comins, H. N. 1977. The development of insecticide resistance in the presence of immigration. J. Theor. Biol. 64: 177-97.

Croft, B. A. 1982. Developed resistance to insecticides in apple arthropods: a key pest to pest control failures and successes in North America. Ent. Exp. Appl. 31: 88-110.

Croft, B. A. & M. T. Aliniazee. 1883. Differential resistance to insecticides in Typhlodromus arboreus Chant and associated phytoseiid mites of apple in the Willamette Valley, Oregon. Environ. Ent. 12: 1420-23.

Croft, B. A. & A. W. A Brown. 1975. Responses of arthropod natural enemies to insecticides. Ann. Rev. Ent. 10: 285-335.

Croft, B. A. & J. G. Morse. 1979. Recent advances on pesticide resistance in natural enemies. Entomophaga 24: 3-11.

Croft, B. A. & C. A Mullin. 1984. Comparison of detoxification enzyme systems in Argyrotaenia citrana (Lepidoptera: Tortricidae) and the ecotoparasite, Oncophanes americanus (Hymenoptera: Braconidae). Environ. Ent. 13: 3330-35.

Croft, B. A. & K. Strickler. 1983. Natural enemy resistance to pesticides: documentation, characterization, theory and application. In: G. P. Georghiou & T. Saito (eds.), Pest Resistance to Pesticides. p. 669-702. Plen Press, New York.

Croft, B. A. & H. E. van de Baan. 1988. Ecological and genetic factors influencing evolution of pesticide resistance in tetranychid and phytoseiid mites. Exp. Appl. Acarol. 4:

Croft, B. A. & S. W. Wagner. 1981. Selectivity of acaricidal pyrethroids to permethrin-resistant strains of Amblyseius fallacis. J. Econ. Ent. 74: 703-06.

Croft, B. A., S. W. Wagner & J. G. Scott. 1984. Multiple- and cross-resistances to insecticides in pyrethroid-resistant strains of the predatory mite, Amblyseius fallacis. Environ. Ent. 11: 161-64.

Curtis, C. F., L. M. Cook & R. J. Wood. 1978. Selection for and against insecticide resistance and possible methods of inhibiting the evolution of resistance in mosquitoes. Ecol. Ent. 13: 273-87.

Dover, M. J., B. A. Croft, S. M. Welch & R. L. Tummala. 1979. Biological control of Panonychus ulmi (Acarina: Tetranychidae) by Amblyseius fallacis (Acarina: Phytoseiidae) on apple: a prey-predator model. Environ. Ent. 8: 282-292.

Georghiou, G. P. 1972. The evolution of resistance to pesticides. Ann. Rev. Ecol. Syst. 3: 133-68.

Georghiou, G. P. 1981. The occurrence of resistance to pesticides in arthropods: an index of cases reported through 1980. F.A.O., Rome.

Georghiou, G. P. 1986. The magnitude of the resistance problem. In: Pesticide Resistance: Strategies & Tactics For Management. pp. 14-43. National Acad. Sci., Wash., D. C.

Georghiou, G. P. & C. E. Taylor. 1986. Factors influencing the evolution of resistance. In: Pesticide Resistance: Strategies and Tactics For Management. pp. 143-46. National Acad. Sci., Wash. D. C.

Gordon, H. T. 1961. Nutritional factors in insect resistance to chemicals. Ann. Rev. Ent. 6: 27-54.

Gould, F. 1984. Fixed function oxidases and herbivore polyphagy: the devil's advocate position. Ecol. Ent. 9: 29-34.

Grafton-Cardwell, E. E. & M. A. Hoy. 1985. Intraspecific variability in response to pesticides in the common green lacewing. Chrysoperla carnea (Stephens) (Neuroptera: Chrysopidae). Hilgardia 53: 1-32.

Grafton-Cardwell, E. E. & M. A. Hoy. 1986. Genetic improvement of the common green lacewing, Chrysoperla carnea (Neuroptera: Chrysopidae): selection for carbaryl resistance. Environ. Ent. 15: 1130-36.

Graur, D. 1985. Gene diversity in Hymenoptera. Evolution 39: 190-99.

Havron, A. 1983. Studies toward selection of Aphytis wasps for pesticide resistance. Ph.D. Thesis, Hebrew Univ. of Jerusalem, Rehovot, Israel.

Horn, D. J. & R. W. Wadleigh. 1987. Resistance of aphid natural enemies to insecticides. In: Aphids, Their Biology, Natural Enemies and Control. P. Harrewijn & A. K. Minks, (eds.). Elsevier Publ., Amsterdam.

Hoy, M. A. 1985. Recent advances in genetics and genetic improvement of the Phytoseiidae. Ann. Eve. Ent. 30: 345-70.

Hoy, M. A. 1989. Pesticide resistance in arthropod natural enemies: variability and selection. In: Pesticide Resistance in Arthropods. CRC Press, Boca Raton, Florida.

Hoy, M. A., D. Flaherty, W. Peacock & D. Culver. 1979. Vineyard and laboratory evaluations of methanol, dimethoate and permethrin for a grape pest management program in the San Joaquin Valley of California. J. Econ. Ent. 72: 250-55.

Hoy, M. A. & N. F. Knop. 1981. Studies on pesticide resistance in the phytoseiid Metaseiulus occidentalis in California. In: Recent Advances in Acarology, Vol. I., J. G. Rodriguez, (ed.). pp. 89-94. Academic Press, New York.

Hseih, C.-Y. 1984. Effects of insecticides on Diaeretiella rapae (McIntosh) with emphasis on bioassay techniques for aphid parasitoids. Ph.D. Thesis, Univ. of Calif., Berkeley.

Huffaker, C. B. 1971. The ecology of pesticide interference with insect populations. In: Agricultural Chemicals--Harmony or Discord For Food People Environment. J. E. Swift (ed.) pp. 92-107. Univ. of Calif, Div. Agric. Sci., Publ, Berkeley.

Ishaaya, I. & J. E. Casida. 1981. Pyrethroid esterase(s) may contribute to natural pyrethroid tolerance of larvae of the common green lacewing. Environ. Ent. 10: 681-84.

Kapentanakis, E. G. & J. E. Cranham. 1983. Laboratory evaluation of resistance to pesticides in the phytoseiid predator Typhlodromus pyri from English apple orchards. Ann. Appl. Biol. 103: 389-400.

Mansour, F. 1984. A malathion-tolerant strain of the spider Chiracanthium mildei and its response to chloropyrifos. Phytoparasitica 12: 163-66.

May, R. M. & A. P. Dobson. 1986. Population dynamics and the rate of evolution of pesticide resistance. In: Pesticide Resistance: Strategies and Tactics For Management, p. 170-93. Nat. Acad. Sci., Wash. D. C.

Morse, J. G. & B. A. Croft. 1981. Developed resistance to azinphosmethyl in a predator-prey mite system in greenhouse experiments. Entomophaga 26: 191-202.

Mullin, C. A. 1985. Detoxification enzyme relationships in arthropods of differing feeding strategies. In: P. A Hedin (ed.), Bioregulators For Pest Control. p. 267-78.

Mullin, C. A. & B. A. Croft. 1984. Trans-epoxide hydrolase: a key indicator enzyme for herbivory in arthropods. Experientia 40: 176-78.

Mullin, C. A. & B. A. Croft. 1985. An update on development of selective pesticides favoring arthropod natural enemies. In: M. A. Hoy & D. C. Herzog (eds.), Biological Control of Agricultural Integrated Pest Management Systems. Acad. Press, New York.

Mullin, C. A., B. A. Croft, K. Strickler, F. Matsumura & J. R. Miller. 1982. Detoxification enzyme differences between a herbivorous and predatory mite. Science 217: 1270-71.

Plapp, F. W., Jr. & D. L. Bull. 1978. Toxicity and selectivity of some insecticides to Chrysopa carnea, a predator of the tobacco budworm. Environ. Ent. 7: 431-34.

Plapp, F. W., Jr. & S. B. Vinson. 1977. Comparative toxicities of some insecticides to the tobacco budworm and its ichneumonid parasite, Campoletis sonorensis. Environ. Ent. 6: 381-84.

Price, P. W. 1984. Insect Ecology, 2nd ed. John Wiley & Sons, New York.

Rosenheim, J. A. & M. A. Hoy. 1986. Intraspecific variation in levels of pesticide resistance in field populations of a parasitoid, Aphytis melinus (Hymenoptera: Aphelinidae): The role of past selection pressures. J. Econ. Ent. 79: 1161-73.

Roush, R. T. & M. A. Hoy. 1981. Laboratory, glasshouse, and field studies of artificially selected carbaryl resistance in Metaseiulus occidentalis. J. Econ. Ent. 74: 142-47.

Roush, R. T. & J. A. McKenzie. 1987. Ecological genetics of insecticide and acaricide resistance. Ánn. Rev. Ent. 32: 361-80.

Roush, R. T. & F. W. Plapp, Jr. 1982. Biochemical genetics of resistance to aryl carbamate insecticides in the predaceous mite, Metaseiulus occidentalis. J. Econ. Ent. 75: 708-13.

Schoonees, J. & J. H. Giliomee. 1982. The toxicity of methidathion to parasitoids of red scale, Aonidiella auranti (Hemiptera: Diaspididae). J. Ent. Soc. South Africa 45: 261-73.

Scott, J. G., B. A. Croft & S. W. Wagner. 1983. Studies on the mechanism of permethrin resistance in Amblyseius fallacis (Acarina: Phytoseiidae) relative to previous insecticide use on apple. J. Econ. Ent. 76: 6-10.

Strickler, K. & B. A. Croft. 1981. Variation in permethrin and azinphosmethyl resistance in populations of Amblyseius fallacis (Acarina: Phytoseiidae). Environ. Ent. 10: 233-36.

Strickler, K. & B. A. Croft. 1985. Comparative rotenone toxicity in the predator, Amblyseius fallacis (Acari: Phytoseidae), and the herbivore, Tetranychus urticae (Acari: Tetranychidae), grown on lima beans and cucumbers. Environ. Ent. 14: 243-46.

Tabashnik, B. E. 1986. Evolution of pesticide resistance in predator-prey systems. Bull. Ent. Soc. Amer. 32: 156-61.

Tabashnik, B. E. & B. A. Croft. 1982. managing pesticide resistance in crop-arthropod complexes: interactions between biological and operational factors. Environ. Ent. 11: 1137-44.

Tabashnik, B. E. & N. L. Cushing. 1988. Quantitative genetic analysis of insecticide resistance: variation in fenvalerate tolerance in a diamondback moth (Lepidoptera: Plutellidae) population. J. Econ. Ent. (in press).

Tabashnik, B. E. & M. W. Johnson. 1999. Evolution of pesticide resistance in natural enemies. In: Bellows, T. S., Jr. & T. W. Fisher, (eds) 1999. Handbook of Biological Control: Principles and Applications. Academic Press, San Diego, CA.

Tanigoshi, L. K. & B. D. Congdon. 1983. Laboratory toxicity of commonly-used pesticides in California citriculture to Euseius hibisci (Chant) (Acarina: Phytoseiidae). J. Econ. Ent. 76: 247-50.

Taylor, C. E. & G. P. Georghiou. 1979. Suppression of insecticide resistance by alteration of gene dominance and migration. J. Econ. Ent. 72: 105-09.

Theiling, K. M. & B. A. Croft. 1990. Pesticide effects on arthropod natural enemies: a database summary. Agric. Ecos. & Environ. (in press).

Trimble, R. M. & D. J. Pree. 1987. Relative toxicity of six insecticides to male and female Pholetesor ornigis (Weed) (Hymenoptera: Braconidae), a parasite of the spotted tentiform leafminer, Phyllonorycter blancardella (Fabr.) (Lepidoptera: Gracillariidae). Canad. Ent. 119: 153-57.

Van Driesche, R. G. & T. S. Bellows, Jr. 1996. Biological Control.. Chapman & Hall, NY. 539 p.

Via, S. 1986. Quantitative genetic models and the evolution of pesticide resistance. In: Pesticide Resistance: Strategies & Tactics For Management, p. 222-35. Nat. Acad. Sci., Wash., D. C.

Waage, J. K., M. P. Hassell & H. C. J. Godfray. 1985. The dynamics of pest-parasitoid-insecticide interactions. J. Appl. Ecol. 22: 825-38.

Warner, L. A. & B. A. Croft. 1982. Toxicities of azinphosmethyl and selected orchard pesticides to an aphid predator, Aphidoletes aphidimyza. J. Econ. Ent. 75: 410-415.

Wilson, E. O. & W. H. Bossert. 1971. A primer of population biology. Sinauer Associates, Sunderland, Massachusetts.

Yu, S. J. 1987. Biochemical defense capacity in the spined soldier bug and its lepidopterous prey. Pestic. Biochem. Physiol. 28: 216-33.

Yu, S. J. 1988. Selectivity of insecticides to spined soldier bug (Heteroptera: Pentatomidae) and its lepidopterous prey. J. Econ. Ent. 81: 119-22.