CHANEY, WILLIAM EUGENE. PhD., Purdue University, August 1988. The Effect of Synthetic Pyrethroid Insecticides on Honey Bees in Indiana: Laboratory Studies and a Survey of Beekeepers and Pesticide Applicators.

This part of the study was designed to answer two basic questions important to further studies. These were: 1) what are the relative effects of four common synthetic pyrethroid insecticides on adult honey bees and 2) what impact does temperature have on each of these products. The synthetic pyrethroids chosen to represent this insecticide class were: fenvalerate (Pydrin 2.4EC), permethrin (Ambush 2E), flucythrinate (Pay-Off 2.5EC) and fluvalinate (Spur 2E). Formulated product at five dosages was fed in 50% sucrose syrup to caged adult bees held in different growth chambers at three temperatures (25 degrees C, 18 degrees C and 12 degrees C). Toxicity was measured by monitoring the number of dead bees daily for five days.

Background and Objectives

Since the introduction of synthetic pyrethroid insecticides, there has been controversy concerning their impact on honeybees. Early reports indicated that these products were safe to bees because they repelled the foraging workers (Atkins 1977, Bos 1983, Moffet 1982, Stoner 1984). Later evidence indicated that the reduced number of foragers observed in treated areas was not due to repellency, but to a disruption of the normal system of communication among foragers or to direct mortality of foraging bees (Rieth 1986, Cox 1987). Because honey bees tend to work the same nectar source until it is depleted, the gradual return of foragers to the treated area could be due to a new set of foraging bees being recruited to the area or due to a loss of repellency as the product degraded.

While this question was still unresolved, reports of serious mortality of bees and the eventual death of colonies of bees exposed to synthetic pyrethroid insecticides began to surface. Laboratory tests indicated that some of these products were toxic to bees when applied topically or ingested, yet field studies did not show the expected damaging effects (Atkins 1981, Johansen 1983, Moffett 1982, Smart 1982, Stevenson 1978). Many of the reports of damage from synthetic pyrethroid insecticides were undocumented because the beekeepers involved did not feel that there was a reasonable chance of reimbursement. Since the USDA Beekeeper Indemnification Program had been terminated, few beekeepers sought restitution for pesticide damage through other legal means (Happ 1971).

The greatest number of reported poisonings came from the midwestern and eastern states. Geographical differences in toxicity to insecticides have been documented in relation to honeybees. As an example, methomyl was found to be safe to honeybees in western states (Atkins, 1979) but highly toxic when used in Wisconsin in sunflower fields leading to massive bee kills (Krause 1983). The differences in toxicity were attributed to the higher humidity of Wisconsin and increased incidences of heavy dew on treated plants which made the pesticide more available to the foraging bees.

In addition to immediate bees kills of greater magnitude, reports by northern beekeepers indicated a greater incidence of winter mortality in colonies exposed to synthetic pyrethroid insecticides during the previous season. If indeed the foraging bees were returning to the hive with nectar or pollen contaminated with sub-lethal doses of synthetic pyrethroid insecticides, perhaps the lower colony temperature in winter was resulting in these doses becoming lethal due to the inverse relationship between the toxicity of synthetic pyrethrbid insecticides and temperature.

In winter, bees control their temperature by forming a cluster. This cluster is roughly a sphere which expands or contracts to regulate the temperature on the surface at about 8 degrees C (45 degrees F) (Owens 1971, Szabo 1985). The center of the cluster may be considerably warmer, depending on the rate of heat loss. In honeybee colonies in the Midwest, the rearing of immatures, or brood, normally begins in January or early February. There is commonly no brood present from October to that time. After the initiation of brood rearing, the temperature at the center of the cluster is maintained at 33 degrees C to 35 degrees C (92-94 degrees F) (Owens 1971, Szabo 1985). Bees move from the inner areas of the cluster to the outer surface in a slow but constant rotation.

The method of storage of food by honeybees, both honey - the carbohydrate source, and pollen - the protein source, is important in this situation. Nectar is collected from flowers and contains 15-35% sugars. The bees evaporate the excess water from the nectar to increase this sugar concentration to about 80%. During this process, the conversion of most of the 12-carbon sugars to 6-carbon sugars is accomplished by a bee supplied esterase. The resulting product, now correctly called honey, is very stable chemically and biologically. The sugar concentration makes it unsuitable for yeast, fungal or bacterial growth. If pesticides contaminated nectar, they are concentrated in the process of producing honey and end up in a quite stable environment (Barker 1980, Winterlin 1973).

Pollen is also collected from the flowers and stored in the cells. It has been shown to be carried to the hive contaminated with pesticides (Johansen 1972, Mayer 1983, Rhodes 1980, Winterlin 1973). The bees add nectar to this pollen to make it more workable and often seal it under a cap of honey. This situation is also very stable chemically and biologically. Unlike the honey, pollen is not concentrated and is usually consumed only by very young bees and those involved in caring for larvae. Under normal conditions in this highly social colony, most bees spend a period of their early lives caring for larvae. This period coincides with the greatest activity of the hypopharyngeal gland in the head. It is also the period when the most pollen is consumed. In winter when brood rearing is initiated, older bees are often needed to care for larvae and they again begin to consume large amounts of pollen. These older bees are also more likely to have previously been exposed to pesticides than young bees reared in the spring.

In the winter cluster, bees are able to maintain temperature by consuming honey stored from the previous season. If this honey is contaminated with pesticides, it is possible that bees moving from the center where they consumed contaminated honey to the outside of the cluster, could experience a temperature drop from 35 degrees C to 7 degrees C (95 degrees F to 45 degrees F). One of the objectives of this study was to look at the toxicity of the selected products at temperatures within this range.

There were also conflicting reports as to the relative toxicity to bees of the two most commonly available synthetic pyrethroid insecticides of the time, permethrin and fenvalerate (Atkins 1979, B.J.Erickson 1983b, Smart 1982, Stoner 1985). Manufacturers were also seeking registrations for new products for agricultural use. Two of these newer products were chosen because they represented both ends of the spectrum as to bee toxicity. These products were fluvalinate and flucythrinate.

The objectives of this first study were:
  1. To devise an easy method of testing the toxicity of synthetic pyrethroid insecticides to adult honeybees in the laboratory
  2. To determine the relative toxicity of the four selected synthetic pyrethroid insecticides
  3. To determine the effect of three temperature ranges on the toxicity of the selected synthetic pyrethroid insecticides
Materials and Methods

Adult honey bee workers were collected from a single colony headed by an Italian queen purchased from a commercial queen breeder. The queen was introduced into the colony approximately 120 days before the start of the experiments. Adult bees were collected from the honey supers during mid-morning on days when the colonies were in active flight. Collected bees were held without food in one gallon paper Fonda Cups (cardboard ice cream cartons) with screen tops for ventilation until needed.

The bees were anesthetized with CO2 and then counted out directly into one half pint Fonda cups containing a vial of 50% sucrose syrup with the assigned concentration of formulated pesticide. The bees were handled by a leg with larval forceps. Any worker bees appearing damaged or exhibited unusual behavior in any way or any drone bees were discarded.

The cups were filled in numerical order, each cup having been randomly assigned a treatment after numbering. The cups were provided with a seven-dram glass vial of sucrose solution with a perforated plastic cap to serve as a feeder when inverted. The inverted vials were held in place on the side of the cup above the floor by a piece of rubber band stapled on either end to the walls of the cup. The top was replaced by a piece of netting to provide ventilation and to facilitate observation of the bees.

The assigned pesticide treatment was prepared just prior to the introduction of the bees by mixing formulated pesticide with 50% sucrose solution at room temperature with a magnetic stirrer. Dilutions of a stock 1000 PPM solution were used to make concentrations of 100, 33, 10, 3 and 1 PPM. The formulated product was obtained from the manufacturing company without the knowledge of the company as to the nature or purpose of the experiment. Product manufactured for the current year was obtained and the concentration was assumed to be as stated on the label.

Once the cups were filled with the 25 adult bees, the bees were allowed to recover from the anesthesia and recounted. The very occasional bee which did not appear to be behaving normally was replaced. As soon as all the bees had recovered, the cups were moved to one of three environmental chambers that held at 12 degrees C, 18 degrees C or 25 degrees C (± 1,25 degrees C). Since only one temperature could be maintained in any one chamber, the design was a split plot randomized complete block design with randomization restricted by temperature. The three replications were made with each of the three chambers held at each of the temperatures.

Within cup variation was estimated by replication of a selected portion of the treatments in any given replication. These were chosen so that when combined they would represent an additional complete set of treatments. These replicates were randomized within the assigned temperatures.

The bees were held in darkness within the chambers and at 60-70% relative humidity. Mortality was determined at 12 and 24 hours the first day and each 24 hours thereafter for 5 days, for a total of 6 determinations. Bees were considered to be dead if they did not respond to a gentle puff of breath. The CO2 in human breath normally produces a fanning response in worker honeybees.

Preliminary studies had shown that the test insecticides, even at concentrations of 1000 PPM in 50% sucrose solution, did not have any fumigant effect within the growth chambers. The toxicity of the insecticide-tainted sucrose solution did change the bees willingness to consume the solution. The toxicity of the insecticide held in solution for seven days was not different from that of solution made fresh the day of introduction. New solutions were prepared for each replication, however.

Results and Discussion

The relative toxicity of the four synthetic pyrethroid insecticides was found to be constant in nearly all situations after the first observation. Regardless of the temperature, concentration or day of observation, the relative order of toxicity from most toxic to least toxic was: permethrin, flucythrinate, fenvalerate and fluvalinate. The product which showed the greatest actual difference in toxicity between the temperature ranges examined was fenvalerate. It was found that the test procedure was acceptable at 25 degrees C and 18 degrees C. The bees were not observed to cluster at 12 degrees C as they would in the hive and normal behavior was not observed. At 12 degrees C the bees became so inactive that feeding was reduced significantly, resulting in lower mortalities than those observed at 18 degrees C. Even under these conditions, however, the relative order of the toxicities of the test insecticides was unchanged. An accurate test at this temperature would require the use of at least three frame broodless colonies. Normal clustering behavior would be initiated and would have to be controlled so as to not raise the temperature of the bees in the cluster.

Data Handling

The ANOVA, GLM (General Linear Model) and TTEST procedures of SAS (Statistical Analysis System) were used for data analysis as appropriate. The GLM procedure was used when unequal cell sizes were present. The TTEST procedure was used to examine the variance between cups within a treatment. The ANOVA procedure was used for all other tests. Duncan's Multiple Range Test (DMRT) at the 5% level was used to indicate significance of differences between means tested together.

Because the number of dead bees tended to converge on 100% from the higher concentrations of the more toxic products, the mean number of dead bees observed over the five days presented a clearer picture of the relative toxicities. Since the bees were exposed to the test insecticide over the entire time, this mean percentage of dead bees better represented the threat these products present to the bees in an overwintering colony. The mean number of dead bees per observation was used for all calculations unless otherwise noted and results are reported as the mean percentage of dead bees per observation.

Duplicates of one-third of each of the treatments were used to evaluate the within-treatment, between-cup variance. Analysis showed this to be such a minor contribution to total variance (0.96%) that these observations were used as a fourth replicate in further analysis, even though one-third had been run simultaneously with each of the three other replications. Analysis using only the first three replicates did not give different results than using all four.

Overall Results.

Figure 1 The mean percentage of dead bees per observation by product.​

Figure 1 shows the mean percentage of test dead bees at each observation period and the standard deviation of this measurement. This represents the mortality for each product at all five concentrations and at all three temperature ranges. The relative order of the toxicity of the products was the same in nearly all of the situations examined. Permethrin was most toxic to the bees followed by flucythrinate, fenvalerate and fluvalinate, in that order. Each of the insecticides was significantly different from the others at the 5% level using Duncans Multiple Range test. Fluvalinate was not significantly different from the control, however.

Figure 2 The mean percentage of dead bees per observation for each treatment-concentration combination​

Figure 2 shows the mean percentage of dead bees per observation for each treatment by concentration. Cups receiving pure 50% sucrose solution as a control were previously assigned to a concentration for analysis. In an analysis of variance by product, concentration was a significant source of variation at a 0.01 level for all insecticides, including fluvalinate, which was not significantly different from the control in an overall analysis.

An analysis of variance by concentration (Table 1) showed significant differences between flucythrinate and fenvalerate only at a concentration of 33 PPM, even though overall means were significantly different. The only concentration at which fluvalinate and the control were significantly different was at 100 PPM. At all concentrations permethrin was significantly more toxic than the other products. Analysis using a pooled estimate of all observations of the value for the control did not change the significance of any differences. The LC50 values calculated using a log-probit analysis were: permethrin - 2.6 PPM, flucythrinate - 8.4 PPM, fenvalerate - 14.7 PPM, and fluvalinate - 799.6 PPM. Each mean represents 6 observations over 5 days for 12 cups, 3 for each of 4 replications, for a total of 120 observations.

Temperature Effects.

Figure 3


The effect of reduced temperature was to increase the toxicity of all of the materials tested over the toxicity observed at 25 degrees C (Figure 3) . The only exception was the mean number of dead bees per observation for permethrin at 12 degrees C, which was lower than for the same product at 25 degrees C. As previously explained, the results at 12 degrees C were affected by the extremely reduced bee activity and food consumption at this temperature. Permethrin caused such a high mortality so quickly that even the relatively high mortality at 12 degrees C was lower than that observed at 25 degrees C or 18 degrees C. The insecticide which showed the greatest increase in the mean percentage of dead bees was fluvalinate, the least toxic of the four synthetic pyrethroid insecticides tested. Fenvalerate showed nearly as large an increase.

The values for the LC50 of the three most toxic products at 25 and 18 degrees C are shown in Figure 4. This figure points out the increasing numerical differences in toxicity at 18 degrees C compared to 25 degrees C for the less toxic materials. The values for fluvalinate are not shown because of the great difference in magnitude. The values for fluvalinate dropped from an LC50 at 25 degrees C of 800 PPM to 615 PPM at 18 degrees C, however, consistent with the previous observation. This decrease in LC50 for fluvalinate is a 23% drop in LC50 compared to drops in LC5O's of 83.4%, 86.1%and 96.2% for fenvalerate, flucythrinate and permethrin, respectively.

Permethrin was significantly more toxic than any of the other treatments at all temperatures (Table 2). Flucythrinate was significantly more toxic than fenvalerate and fluvalinate was significantly more toxic than the control at 18 degrees C only. The relative order of toxicities was the same as had been seen at each concentration and overall for all of the temperatures examined. Each mean represents 6 observations over 5 days of 20 cups, 5 in each of 4 replications, for a total of 120 observations.


Permethrin was found to be the most toxic of the four synthetic pyrethroid insecticides tested at all temperatures and at all concentrations. It was also found to have the largest percentage increase in toxicity with a temperature change from 25 degrees C to 18 degrees C. The extremely low LC50 at 18 degrees C of 0.202 PPM for permethrin certainly suggests that even very small concentrations of permethrin in nectar, collected at temperatures often 25 degrees C or higher, then concentrated by the bees in the process of ripening of nectar to honey, could produce significant mortality in overwintering colonies of bees. If one considers the relative importance of an individual member of a colony of bees in December or January compared to the value of an individual worker in May, June or July, the implications for a significant impact of even the smallest contamination of nectar by permethrin become even more staggering.

Even using optimistic estimates of the rate of degradation of the insecticide while in storage in a solution that is nearly 80% sugars, naturally antibiotic, and sealed within a dark wax cell, the concentration of permethrin in nectar which would result in a 0.202 PPM concentration in the resulting honey, is very small. If we assume a best case scenario of 30% sugar in nectar and a degradation of 80% of the permethrin over the period from collection to consumption, a concentration of only 0.38 PPM permethrin in nectar will result in honey containing the LC50 of permethrin at 18 degrees C.

Flucythrinate was found to be significantly more toxic than fenvalerate overall and at a concentration of 33 PPM, while at other tested concentrations the differences were not significant. Flucythrinate was found to result in a numerically greater mean percentage of dead bees per observation than fenvalerate at all concentrations. The LC50 of the fenvalerate was approximately double that of flucythrinate at both 25 degrees C and 18 degrees C, however. The percentage of change between the two temperature ranges were approximately equal. Using the same best case scenario as outlined for permethrin, a concentration of 8.8 PPM of fenvalerate or 3.45 PPM of flucythrinate in nectar would result in the LC50 of each being reached at 18 degrees C. These concentrations are in the range of the concentrations of fenvalerate found by Erickson (personal commun.) in nectar being brought back to hives in Wisconsin by bees foraging on sunflowers treated with Pydrin (fenvalerate).

Fluvalinate was significantly less toxic than any of the other synthetic pyrethroid insecticides tested at any concentration or at any temperature It was not found to be significantly different from the control of 50% sucrose except at a concentration of 100 PPM. The LC50 of 615 PPM at 18 degrees C should indicate that even in a worst case scenario considering no degradation of the product after being collected by the bees in nectar and concentrated into honey, the likelihood of a bee receiving a lethal dose is very small. The fact that the product did exhibit some increased toxicity to bees at 18 degrees C should be noted however, especially in light of the recent registration of fluvalinate as a miticide for use in live colonies of bees for control of the varroa mite, Varroa jacobosoni (Oudemans) (Herbert 1987).

The threat to colonies of honey bees in Indiana from synthetic pyrethroid insecticides as suggested by reports by beekeepers was supported by the results of these tests. The LC50's calculated from the data are certainly not difficult to visualize as a possibility in honey made by colonies foraging on plants treated intentionally or unintentionally with synthetic pyrethroid insecticides. Despite the failure of this methodology to adequately measure the toxicity at temperatures below 18 degrees C, it is reasonable to assume that the same products would show even lower LC50's at the lower extreme of the temperature range a bee might encounter in a hive over the winter in Indiana.