Thrips Densities Research Articles

Variation and Consistency in Spatial Distribution as Measured by Taylor's Power Law

To investigate the spatial isotropy of Taylor's power law, data of the aerial population densities of western flower thrips, Frankliniella occidentalis (Pergande), were compared at 6 greenhouses in northeastern Ohio. The distributions were the smne in all 6 houses despite differences in crop (host plant) and in thrips population growth and development. Comparison with data from the literature on F. occidentalis distributions obtained in houses with horticultural crops using closely similar sampling methods revealed a consistency of distribution over a range of plant architectures. The consistency of power law results means that the same sampling program may be used for F. occidentalis in a variety of greenhouse crops. In contrast, samples taken within the canopy of a cucumber crop resulted in Taylor's power laws significantly different from those obtained with samples taken above the canopy. Taylor's power law appears to remain constant provided the same statistical population is being sampled. Sampling the same reproductive population under different conditions (e.g., stage, phenology, or habitat) can result in different power law estimates, as reported by some authors. Possible reasons for power law differences and the significance of power law constancy also are discussed. Tables for optimum sample size are given for sampling F. occidentalis in greenhouse crops using industry standard yellow sticky cards. In addition, it is shown how the number of samples required for a fixed level of precision can be manipulated by varying the sampling efficiency or size of sampler.

Impact of western flower thrips (Thysanoptera: Thripidae) on growth, photosynthesis and productivity of greenhouse sweet pepper

Low, medium and high densities of western flower thrips [ Frankliniella occidentalis (Pergande)] were established in three greenhouses, in 1994 and 1995, to investigate the effects of thrips on growth, photosynthesis and productivity of greenhouse sweet pepper. Sweet pepper fruit was very susceptible to cosmetic injury which resulted in a reduction of the commercial grade of the fruit (quality). Damage by western flower thrips (WFT) on the fruit was direct and instant. Short periods of high thrips densities resulted in cosmetic damage to the fruit. The major period for fruit damage occurred in the 3rd to 5th weeks before harvest, when the fruit was growing rapidly. Approximately seven thrips per sticky trap per day from weeks #3 to #5 before fruit harvest will result in fruit being downgraded from commercial grade #1 to #2. Impact of WFT on plant growth, photosynthesis and marketable yield was indirect and cumulative. Leaf photosynthesis was reduced when plants were exposed to more than 1200 accumulated larval-days per plant and marketable yield was reduced when exposed to more than 1800 accumulated larval-days per plant. Plants recovered slowly from the thrips damage once yield loss had been detected. Therefore, thrips abundance should be maintained below 1200 larval-days per plant to prevent the negative impact of WFT on plant physiology and yield. The importance of these findings on pest management programs for WFT is discussed.

Evaluation of Thrips Densities in At-Planting Seed and Soil Treatments on a Sharkey Clay Soil, 1997

Abstract Plots were replicated four times in a RCB design and were four rows (40 inch centers) X 45 ft. ‘StoneviUe 474’ cotton seed was planted on a Sharkey clay soil on 5 May at the rate of 4 seed/row ft. Cotton seed was planted with a John Deere 7100 series planter which was equipped with 10-inch seed cones mounted to replace the seed hoppers. Granular in-furrow treatments were applied with 8-inch belt-cone applicators mounted to replace the standard granular applicators. In-furrow spray treatments were applied with a CO2-charged spray system through 25015 nozzles (l/row) mounted in front of the press wheels. The spray tips were turned to spray across the furrow and calibrated to deliver 5 gpa finished spray. Control of thrips was assessed by randomly selecting 5 plants per plot at 7, 11, 16, 19, 23, 27, 31, and 35 days after emergence (DAE). Plant samples were washed to remove insects. Major past and/or secondary pest control was initiated in Jun and continued on an “as needed” basis through Aug. Percent open bolls was determined on 3 Sep. Samples of 25 bolls/plot were collected and ginned on a laboratory gin to determine lint percentage for each plot. These data were then used to determine lint yield. Plots were harvested on 19 Sep and 2 Oct. The center two rows were mechanically harvested using a John Deere spindle-type picker. Yields were con-verted to lb lint/acre.

Evaluation of Thrips Densities in At-Planting Seed and Soil Treatments on a Gigger Silt Loam Soil, 1997

Abstract Plots were replicated four times in a RCB design and were 4 rows (40 inch centers) X 45 ft. ‘Stoneville 474’ cotton seed was planted on a Gigger silt loam soil 7 May. Cotton seed was planted at a rate of 4 seed/row ft with John Deere 7100 series planter which was equipped with 10-inch seed cones mounted to replace the seed hoppers. Granular in-furrow treatments were applied with standard granular applicators. In-furrow spray treatments were applied with a CO2-charged spray system through 80015 nozzles (l /row) mounted in front of the press wheels. The spray tips were turned to spray across the furrow and calibrated to deliver 5 gpa finished spray. Control of thrips was assessed by randomly selecting 5 plants per plot at 7, 11, 16, 19, 23, 27, 31, and 35 days after emergence (DAE). Plant samples were washed to remove insects. Major pest and/or secondary pest control was initiated in Jun and continued on an “as needed” basis through Aug. Percent open bolls was determined on 4 Sep. Samples of 25 bolls/plot were collected and ginned on a laboratory gin to determine lint percentage for each plot. These data were then used to determine lint yield. The center two rows were mechanically harvested on 30 Sep and 20 Oct using a John Deere spindle-type picker. Yields were converted to lb lint/acre.

Evaluation of Thrips Densities in At-Planting Seed and Soil Treatments on a Commerce Silt Loam Soil, 1997

Abstract Plots were replicated four times in a RCB design and were 4 rows (40 inch centers) X 45 ft. ‘Stoneville 474’ cotton seed was planted on a Commerce silt loam soil on 6 May. Cotton seed was planted at the rate of 4 seed/row ft with a John Deere 7100 series planter which was equipped with 10-inch seed cones mounted to replace the seed hoppers. Granular in-furrow treatments were applied with 8-inch belt-cone applicators mounted to replace the standard granular applicators. In-furrow spray treatments were applied with a CO2-charged spray system through 25015 nozzles (1/row) mounted in front of the press wheels. The spray tips were turned to spray across the furrow and calibrated to deliver 5 gpa finished spray. Control of thrips was measured by randomly selecting 5 plants per plot at 7, 11, 16, 19, 23, 27, 31, and 35 days after emergence (DAE). Plant samples were washed to remove insects. Major pest and/or secondary pest control was initiated in Jun and continued on an “as needed” basis through Aug. Percent open bolls was determined on 4 Sep. Samples of 25 bolls/plot were collected and ginned to determine lint percentage for each plot. These data were than used to determine lint yield. The center two rows were mechanically harvested on 19 Sep and 2 Oct using a John Deere spindle type picker. Yields were converted to lb lint/acre.

A new sampling technique for estimating the population densities of greenhouse thrips

A new sampling technique for estimating the population densities of greenhouse thrips

Environmental Influences on Feeding Damage Caused by Western Flower Thrips (Thysanoptera: Thripidae) to Chrysanthemum

The effects of test-location, daylength, thrips density, and choice or no-choice test on feeding damage by Frankliniella occidentalis (Pergande) were studied in chrysanthemum, Dendranthema grandiflora Tzvelev. Under all circumstances, tested cultivars differed significantly in the level of both silver and growth damage caused by thrips feeding on old and young chrysanthemum leaves, respectively. Thrips resistance to both types of feeding damage, measured under the different test circumstances was highly correlated. This means that thrips resistance in chrysanthemum measured in our standardized laboratory experiment gives a good prediction for chrysanthemum breeders who want to select cultivars for thrips resistance under commercial greenhouse conditions. Only test location had an effect on the level of growth damage. In the greenhouse, all chrysanthemum cultivars were more susceptible to growth damage than in the laboratory. The level of silver damage, however, remained constant. Silver damage was only influenced by daylength. All cultivars showed higher levels of silver damage under short daylength compared with long daylength. A density of 10 thlips per plant was found to be high because only 1.6 times more silver damage and 1.1 times more growth damage was found at a 4 times higher thrips density. At the highest density, many thrips died, which resulted in about the same thrips numbers for both treatments at the end of the experiment. Thrips tended to choose among cultivars. However, this did not result in a higher variance in resistance in the free-choice experiment compared with the no-choice experiment.

Effect of elevated CO2 on interactions betwe en the western flower thrips, Frankliniella occidentalis (Thysanoptera: Thripidae) and the common milkweed, Asclepias syriaca.

We measured the effect of elevated CO2 on populations of the western flower thrips, Frankliniella occidentalis and on the amount of leaf damage inflicted by the thrips to one of its host plants, the common milkweed, Asclepias syriaca. Plants grown at elevated CO2 had significantly greater aboveground biomass and C:N ratios, and significantly reduced percentage nitrogen. The number of thrips per plant was not affected by CO2 treatment, but the density of thrips (numbers per gram aboveground biomass), was significantly reduced at high CO2. Consumption by thrips, expressed as the amount of damaged leaf area per capita, was significantly greater at high CO2, and the amount of leaf area damaged by thrips was increased by 33%. However overall leaf area at elevated CO2 increased by 62%, more than compensating for the increase in thrips consumption. The net outcome was that plants at elevated CO2 had 3.6 times more undamaged leaf area available for photosynthesis than plants at ambient CO2, even though they had only 1.6 times the overall amount of leaf area. This study highlights the need for measuring the effects of herbivory at the whole-plant level and also the importance of taking herbivory into account when predicting plant responses to elevated CO2.

The proportion of viruliferous individuals in field populations of Frankliniella occidentalis: Implications for tomato spotted wilt virus epidemics in tomato

Tomato spotted wilt tospovirus (TSWV) infected plants and western flower thrips (Frankliniella occidentalis Perg., WFT) adult population densities were monitored during 1993 and 1994 in field tomatoes in Northeastern Spain. The proportion of viruliferous WFT adults in field populations was quantified. A significant association has been found between early population densities of WFT adult thrips and final TSWV incidence for early transplanted tomato crops. In contrast, for late transplanted tomato crops, whereas similar high final disease incidences of TSWV could be attained, very low WFT adult population densities were always detected. The significantly higher infectious potential of WFT populations collected during the early growth stages of late transplanted tomatoes could be relevant for the TSWV incidences attained in spite of the low thrips numbers detected.

Evaluation of Thrips Densities in At-Planting Seed and Soil Treatments on Gigger Silt Loam, 1996

Abstract Plots were replicated four times in a RCB design and were four rows wide (40 inch spacing) X45 ft. Stoneville 474 cotton seed of the same seed lot was planted on a Gigger silt loam soil on 7 May. The seeding rate was 4 seed/row ft. Cotton seed was planted with a John Deere 7300 series planter which was equipped with 10 inch seed cones mounted to replace the seed hoppers. Granular in-furrow treatments were applied with standard granular applicators. In-furrow spray treatments were applied with a CO2 charged spray system through 25015 nozzles (1/row) mounted in front of the press wheels. The spray tips were turned to spray across the furrow and calibrated to deliver 5 gpa finished spray. Control of thrips was assessed by randomly selecting 5 plants per plot on 20, 24, 28, and 31 May and 5 and 9 Jun. Plant samples were processed by using whole plant washing procedures to remove insects. Major pest and/or secondary pest control was initiated in Jun and continued on an “as needed” basis through Aug. Percentage of open bolls was determined by counting all open and unopened bolls within a 3.0 m section of a center row on 18 Sep. Plots were harvested on 26 Sep and 10 Oct. These data were subjected to appropriate calculations to determine percent first harvest. The center two rows were mechanically harvested using a John Deere spindle type picker. Yields were converted to lbs seed cotton/acre.

Evaluation of Thrips Densities in At-Planting Seed and Soil Treatments on a Commerce Silt Loam Soil, 1996

Abstract Plots were replicated four times in a RCB design and were four rows wide (40 inch spacing) X45 ft. Stoneville 474 cotton seed of the same seed lot was planted on a Commerce silt loam soil on 7 May. The seeding rate was 4 seed/row ft. Cotton seed was planted with a John Deere 7100 series planter which was equipped with 10 inch seed cones mounted to replace the seed hoppers. Granular in-furrow treatments were applied with 8 inch belt cone applicators mounted to replace the standard granular applicators. In-furrow spray treatments were applied with a CO2 charged spray system through 25015 nozzles (1/row) mounted in front of the press wheels. The spray tips were turned to spray across the furrow and calibrated to deliver gpa finished spray. Control of thrips was assessed by randomly selecting 5 plants per plot on 20, 24, 28, and 31 May and 5 and 9 Jun. Plant samples were processed by using whole plant washing procedures to remove insects. Major past and/or secondary pest control was initiated in Jun and continued on an “as needed” basis through Aug. Percentage of open bolls was determined by counting all open and unopened bolls within a 3.0m section of a center row on 16 Sep. Plots were harvested on 23 Sep and 8 Oct. These data were subjected to appropriate calculations to determine percent first harvest. All four rows were mechanically harvested using a John Deere spindle type picker. Yields were converted to lbs seed cotton/acre.

Evaluation of Thrips Densities in At-Planting Seed and Soil Treatments on Sharkey Clay, 1996

Abstract Plots were replicated four times in a RCB design and were four rows wide (40 inch spacing ) X45 ft. Stoneville 474 cotton seed of the same seed lot was planted on a Sharkey clay soil on 6 May. The seeding rate was 4 seed/row ft. Cotton seed was planted with a John Deere 7100 series planter which was equipped with 10 inch seed cones mounted to replace the seed hoppers. Granular in-furrow treatments were applied with 8 inch belt cone applicators mounted to replace the standard granular applicators. In-furrow spray treatments were applied with a CO2 charged spray system through 25015 nozzles (1/row) mounted in front of the press wheels. The spray tips were turned to spray across the furrow and calibrated to deliver 5 gpa finished spray. Control of thrips was assessed by randomly selecting 5 plants per plot on 20, 24, 28, and 31 May and 5 and 9 Jun. Plant samples were processed by using whole plant washing procedures to remove insects. Major pest and/or secondary pest control was initiated in Jun and continued on an “as needed” basis through Aug. Percentage of open bolls was determined by counting all open and unopened bolls within a 3.0 m section of a center row on 17 Sep. Plots were harvested on 24 Sep and 8 Oct. These data were subjected to appropriate calculations to determine percent first harvest. All four rows were mechanically harvested using a John Deere spindle type picker. Yields were converted by lb seed cotton/acre.

Management of Western Flower Thrips on Cucumber withDicyphus tamaninii(Heteroptera: Miridae)

Late instar nymphs ofDicyphus tamaniniiWagner (Heteroptera: Miridae) were released in a 1:10 and a 3:10 ratio of the predator to prey in a high initial infestation of 5 adult thrips per leaf and in a low initial infestation of 1 adult thrips per leaf in greenhouse cucumbers. The 3:10D. tamaninii:Frankliniella occidentalis(Pergande) ratio kept thrips populations under the economic injury level. This ratio is recommended for controlling the high thrips populations that are common in Mediterranean greenhouses. The 1:10D. tamaninii:F. occidentalisratio worked well when thrips populations were low but not when thrips density was high. The efficiency ofD. tamaniniiin low thrips populations is very interesting for developing IPM programs in which only limited releases ofD. tamaniniican be done. Sampling cucumber flowers to evaluate thrips densities has not been a very effective method because a variable proportion of the thrips population is located in flowers, whileD. tamaniniiis unlikely to be found in flowers.

Abundance of Frankliniella spp. (Thysanoptera: Thripidae) on Winter and Spring Host Plants

Abundance of western flower thrips, Frankliniella occidentalis (Pergande), and tobacco thrips, F. fusca (Hinds), was determined during winter and early spring in northern and central Florida using weekly samples collected from hairy vetch, Vicia villosa Roth; crimson clover, Trifolium incarnatum L.; and wheat Triticum aestivum L. Abundance of F. occidentalis was greater in northern than in central Florida, and population densities greater in the terminal buds compared with the leaves. A total of 3.9 and 5.1 generations developed in northern Florida vetch during 2 winter/early spring seasons, which agreed well with predictions of 3.5 and 4.8, respectively, based on degree-day accumulations. Crimson clover was a poor host for F. occidentalis and F. fusca at both locations. Abundance of F. fusca in wheat was similar in northern and central Florida, and population densities of immature thrips were greater on branches compared with the leaves. Three complete and 2 partial generations of F. fusca developed on wheat, which was one more than predicted using degree-day accumulations.

Emergence of Pear Thrips (Thysanoptera: Thripidae) and its Relation to Foliar Damage

Adult emergence of pear thrips, Taeniothrips inconsequens (Uzel), from soil in forests predominating in sugar maple, Acer saccharum Marshall, was monitored, using 7.6- cm-diameter traps, at sites in north central, northeastern, central and southwestern Vermont from 1989 to 1991. The onset of emergence ranged from 2 April to 15 May, but generally occurred in mid-April in all regions. We measured the number of thrips in soil in September and how many emerged in the spring, and monitored the timing and rate of emergence to determine if damage could be predicted from these factors. Thrips density in the soil was not consistently correlated with damage, making it an unreliable predictive measure. A correlation between thrips damage and the timing or duration of emergence was also not detected. The number of thrips per emergence trap was positively correlated with damage in all study years, and from this correlation the predicted number of thrips per trap for moderate damage to occur was 18 in 1989, 13 in 1990 and 8 in 1991. Annual fluctuations in the threshold number of thrips required for damage could limit the use of thrips emergence data for predictive purposes.

Influence of temperature and vapor pressure deficit on the rate of predation by the predatory mite, Amblyseius cucumeris, on Frankliniella occidentalis

AbstractThe influence of temperature and vapor pressure deficit (VPD) on the rate of predation by the predatory mite, Amblyseius cucumeris (Oudemans) on Frankliniella occidentalis (Pergande) was determined under controlled laboratory conditions. The survival of first instars of F. occidentalis was initially determined by themselves. Then the number of first instars that were killed by a single female adult of A. cucumeris was determined for the same density of thrips. The differences in the mortality between the two experiments were used to calculate the rate of predation by A. cucumeris under a range of temperature and VPD regimes. Rate of predation was expressed as the number of thrips killed per h to account for the different time periods that the trials lasted. A quadratic model was fitted to the data. At a constant temperature, the rate of predation decreased with increasing VPD for VPDs ≤ 1.24 to 1.44 kPa. Above these VPDs, the rate of predation started to increase again. At a constant VPD, the rate of predation decreased slightly at the lower temperatures and increased at the higher temperatures. The rate of change was dependent upon the VPD. By using the quadratic model, the rate of predation can be predicted for the range of hygrothermal conditions that would be encountered in the greenhouse. The optimal condition for maximum rate of predation by A. cucumeris on first instars of F. occidentalis in the greenhouse is a VPD ≤ 0.75 kPa at the recommended production temperatures and VPDs (17 to 25°C and ≤ 1.5 kPa).

Effect of Thrips tabaci (Thysanoptera: Thripidae) on Yellow Onion Yields and Economic Thresholds for Its Management

The proper management of onion thrips, Thrips tabaci Lindeman, on yellow onions requires the determination of their effect on yields and the establishment of an economic threshold. In field plot experiments, we observed no difference in onion thrips population densities among 6 yellow onion cultivars during 1988 and 1989. Onion thrips infestations resulted in onion yield losses of 43 and 34.5% with maximum seasonal averages of 149.2 and 172.0 thrips per plant in untreated plots in 1988 and 1989, respectively. No effect on onion yields was observed in 1990 when seasonal thrips densities did not exceed 35.2 thrips per plant. The effect of thrips on onion yields was best described by nonlinear models and by the use of the cumulative number of thrips-days per plant as a measure of thrips pressure. Plants were relatively insensitive to thrips feeding early and late in phenological development. Important differences in the onion plant response to thrips damage were observed between 1988 and 1989. These differences were likely linked with the duration and intensity of the water deficit recorded during both seasons. Economic thresholds of 2.2 and 0.9 thrips per leaf were estimated during the summers of 1988 and 1989, respectively. These thresholds can be used for onions sold at harvest or stored because costs associated with storage are compensated for by a higher return price. The use of the 0.9 thrips per leaf economic threshold in 1990 would have resulted in unnecessary treatments, making this threshold suitable for severe drought situations only.

Aggregation Indices and Sample Size Curves for Binomial Sampling of Flower-Inhabiting Frankliniella Species (Thysanoptera: Thripidae) on Tomato

Western flower thrips, Frankliniella occidentalis (Pergande), females inhabit tomato flowers and insert eggs into developing fruit, causing cosmetic damage. The species is a vector for tomato spotted wilt virus. Only larval thrips acquire the disease, and larvae developing to adult on infected plants can disperse and infect additional plants in a field. Tobacco thrips, F. fusca (Hinds), another vector species, and flower thrips, F. tritici (Fitch), a nonvector species, also inhabit tomato flowers but apparently do not reproduce in the flowers. Our objectives were to determine dispersion characteristics of these thrips in tomato flowers and to determine relationships between density and incidence for developing a presence/absence sampling program. Taylor’s power law indices indicated that populations of F. occidentalis and F. tritici were aggregated. Larvae of F. occidentalis were more aggregated than adults. Populations of F. fusca were aggregated in 1988, but randomly dispersed in 1989. Aggregation indices were used to determine relationships between incidence and density and to develop sample size curves for use in a binomial sampling program. Adults of F. occidentalis and F. tritici were estimated using presence/absence sampling at densities <1.4 per flower; binomial sampling was reliable for F. fusca at densities <2.0 per flower. Sixteen to 18 flower samples were needed to estimate thrips densities at the 25% precision level.

THE SIGNIFICANCE OF OUTCROSSING IN AN INTIMATE PLANT-HERBIVORE RELATIONSHIP. I. DOES OUTCROSSING PROVIDE AN ESCAPE FROM HERBIVORES ADAPTED TO THE PARENT PLANT?

Sexual reproduction may be advantageous for hosts that are preyed on or parasitized by enemies that are highly adapted to them. Sexual reproduction can create rare genotypes that may escape predation by virtue of rarity and can create variable progeny that may escape predation if enemies are specialized to only one genotype of host. Populations of the herbivorous thrips, Apterothrips apteris, have been shown to be adapted to individual Erigeron glaucus clones. Here, we show that thrips adapted to the parental clone could better use plant progeny of the "home" clone produced through selfing than progeny derived from selfing of other clones. Thus, despite recombination, progeny produced by selfing presented a resource that was similar to the parental phenotype with respect to use by adapted thrips. We also show that E. glaucus susceptibility to thrips has a genetic basis and then ask whether outcrossing provides a means for E. glaucus clones to escape attack by adapted thrips. When we compared the success of thrips on progeny produced by selfing or outcrossing of the home clone, we found that the merits or disadvantages associated with outcrossing were dependent on the susceptibility to infestation of the parental clones. Selfing by clones characterized by low infestations of thrips appeared to preserve resistant genotypes; all outcrossed progeny had, on average, higher infestation levels than selfed progeny. In contrast, outcrossed progeny of clones characterized by high infestations of thrips had either the same thrips density as progeny from selfing, when the pollen donor was a highly infested clone, or lower density, when the pollen donor was a low infestation clone. The advantages of outcrossing were caused by the alleles contributed to progeny rather than to progeny variability or rarity.

Influence of Planting Date and Insecticide Treatment on Insect Pest Abundance and Damage in Dryland Cotton

Objectives of this study were to measure the influence of planting date and insecticidal control on pest abundance and damage in dryland cotton. ‘Paymaster 145’ cotton was planted in late April, late May (which is the recommended delayed uniform planting date for management of the boll weevil, Anthonomus grandis grandis Boheman), and late June for 4 yr from 1986 to 1989 in the northern Texas Rolling Plains. Subplots within each planting date were no insecticide treatment, and treatment with aldicarb at planting and application of an appropriate insecticide when a pest attained its damage threshold. Insects sampled were thrips (including the western flower thrips, Frankliniella occidentalis [Pergande]); cotton fleahopper, Pseudatomoscelis seriatus (Reuter); bollworm, Helicoverpa zea (Boddie); cotton aphid, Aphis gossypii Glover; and boll weevil. Thrips densities were highest in cotton planted in late April; densities were reduced by aldicarb treatments in late April and late May cotton but not in late June cotton. Highest densities of cotton fleahoppers occurred in cotton planted in late May; densities were reduced by aldicarb treatments in late May and late June cotton but not in late April cotton. Densities of overwintered boll weevils were highest in April-planted cotton and lowest in June-planted cotton. Boll weevil damage during the growing season was lowest in cotton planted in late May, and insecticides applied for boll weevil control during the growing season suppressed damage for all planting dates. Bollworm densities were not influenced by planting date or by insecticide treatment. Cotton aphid densities were highest in cotton planted in late June. Delayed uniform planting for boll weevil management does not adversely affect damage by the other important cotton pests in dryland cotton in the northern Texas Rolling Plains. Planting in late May reduces damage caused by thrips, cotton aphids, and boll weevils.