3Department of Entomology
Cornell University
Ithaca, NY, 14853

Abstract

Releases of Eretmocerus eremicus Rose and Zolnerowich (Hymenoptera: Aphelinidae) at release rates of 3.0-7.5 females per plant per week successfully suppressed whitefly populations on commercial poinsettia (Euphorbia pulcherrima Willd. ex Koltz.) crops in fall of 1996 at four Massachusetts commercial producers. At two sites, the whitefly populations consisted exclusively of greenhouse whitefly, Trialeurodes vaporariorum (Westwood), and at the other two sites exclusively of silverleaf whitefly, Bemisia argentifolii Bellows and Perring. Parasitoids were received from commercial suppliers and monitored weekly to determine the sex ratio of newly emerged adults, as well as the rate of adult emergence. Commercially produced pupae were 48.1% (± 2.2 SE) female and had 58.1% (± 3.6 SE) emergence under greenhouse conditions. Plants from the four biological control greenhouses in this trial at the time of sale of the crop had an average of 0.55 (± 0.28 SE) nymphs per leaf. Chemically-protected poinsettias offered for sale at eight local retail outlets had an average of 0.16 (± 0.09 SE) nymphs per leaf. Final whitefly densities in both biological control and insecticide-treated greenhouses were acceptable to consumers. An average of 6.8 insecticide applications was applied to suppress whiteflies in chemical control greenhouses in this trial, compared to 1.7 in the biological control greenhouses. Use of biological control was 27 fold more expensive, costing $2.14 per plant compared to $0.08 for chemical control. Cost of biological control was inflated by three factors: (1) an incorrectly high estimate by one grower of number of plants per greenhouse, (2) an unusually long production period (23 weeks) for one grower, and (3) miscommunication with the insectary concerning manner of filling orders to compensate for reduced percentage of emergence of adult parasitoids from ordered parasitized nymphs. Control of these cost-inflating factors would allow some reduction in the cost of the use of this parasitoid, but not to levels competitive with current pesticides. This study is the first to demonstrate the ability of E. eremicus releases to suppress T. vaporariorum populations in commercial poinsettia crops and parasitism of T. vaporariorum by E. eremicus was 7.5-fold higher (ave. 24.8% parasitism of fourth instar nymphs in pooled seasonal samples) than that observed in B. argentifolii (ave. 3.3%).

Key Words: Eretmocerus eremicus, Bemisia argentifolii, Trialeurodes vaporariorum poinsettia, biological control, augmentative release, evaluation, cost, greenhouses

Resumen

Liberaciones de Eretmocerus eremicus Rose y Zolerowich (Hymenoptera: Aphelinidae) a razón de 3.0-7.5 hembras por planta por semana lograron un control efectivo de mosquita blanca en cuatro cultivos comerciales de nochebuena (Euphorbia pulcherrima Willd. ex Koltz.) de Massachusetts durante el otoño de 1996. En dos sitios, las poblaciones de mosquita blanca consistieron exclusivamente de la mosquita blanca de invernadero, Trialeurodes vaporariorum (Westwood), mientras que en los otros dos sitios las poblaciones fueron exclusivamente de mosquita blanca de la hoja plateada, Bemisia argentifolii Bellows and Perring. Los parasitoides fueron obtenidos de proveedores comerciales y monitoreados semanalmente para determinar la proporción de machos y hembras adultos recién emergidos, así como la tasa de emergencia de adultos. Las pupas producidas comercialmente fueron 48.1% (± 2.2 SE) hembras y tuvieron una tasa de emergencia de 58.1% (± 3.6 SE) bajo condiciones de invernadero. Al momento de su venta, plantas provenientes de los cuatro invernaderos de control biológico usados en este estudio tuvieron un promedio de 0.55 (± 0.28 SE) ninfas por hoja. En comparación, plantas protegidas con insecticidas tuvieron un promedio de 0.16 (± 0.09 SE) ninfas por hoja al momento de su venta en ocho locales comerciales. Las densidades finales de mosquita blanca encontradas tanto en los invernaderos de control biológico como en aquellos donde se emplearon insecticidas fueron aceptables a los consumidores. En promedio, 6.8 aplicaciones de insecticida fueron hechas para controlar a la mosquita blanca en los invernaderos de control químico usados en este estudio, comparado con 1.7 aplicaciones en los invernaderos de control biológico. El costo del control biológico fue 27 veces más caro que el del control químico ($2.14 vs. $0.08 por planta). El costo del control biológico resultó elevado debido a tres factores: (1) el cálculo erróneo (demasiado alto) por parte de un productor con respecto al número de plantas por invernadero, (2) un período demasiado largo de producción (23 semanas) en el caso de un productor, y (3) falta de comunicación con personal del insectario respecto a la manera de compensar el porcentaje reducido de emergencia de adultos parasitoides logrado por las ninfas parasitadas ordenadas. El costo del uso de parasitoides podría reducirse al corregir los errores mencionados, pero no lo suficiente para ser competitivo con el uso de insecticidas. Este estudio es el primero en demostrar la eficacia del parasitoide E. eremicus en el control de T. vaporariorum en cultivos comerciales de nochebuena. El estudio demostró que el parasitismo de T. vaporariorum por E. eremicus fué 7.5 veces más alto que el obtenido con B. argentifolii (parasitismo de 24.8% vs. 3.3% de ninfas de cuarto instar).

Silverleaf whitefly, Bemisia argentifolii Bellows and Perring, (= the "B" strain of Bemisia tabaci [Gennadius]) and greenhouse whitefly, Trialeurodes vaporariorum (Westwood), (both Homoptera: Aleyrodidae) are important pests of poinsettia (Euphorbia pulcherrima Willd. ex Koltz.) in the United States (Helgesen & Tauber 1974, Byrne et al. 1990, Bellows et al. 1994). The parasitoids most extensively used for whitefly biological control in protected floricultural crops have been Encarsia formosa Gahan and Eretmocerus eremicus Rose and Zolnerowich (formerly given as Eretmocerus sp. nr. californicus) (both Hymenoptera: Aphelinidae) (Drost et al. 1996; Hoddle & Van Driesche 1996; Rose & Zolnerowich 1997; Hoddle et al. 1996, 1997ab, 1998abc; Hoddle & Van Driesche in press).

Previous trials in small, experimental greenhouses (holding 90 plants) suggested that a Bemisia-adapted strain of E. formosa (referred to as the Beltsville strain, Heinz & Parrella 1994) and E. eremicus had the potential to provide effective silverleaf whitefly control in poinsettia crops if released at rates of 1-3 females per plant per week (Hoddle et al. 1997ab, 1998a). Trials in commercial greenhouse poinsettia crops in 1995 in Massachusetts compared the efficacy of E. eremicus and the Beltsville strain of E. formosa at a release rate of 3 females per plant per week for each species. In both summer stock plants and fall Christmas crop plants, E. eremicus suppressed silverleaf whitefly better than the Beltsville strain of E. formosa (Hoddle and Van Driesche, 1999). Poinsettias from these 1995 trials were sufficiently free of whiteflies to be acceptable to growers for use of cuttings from the summer crop (Hoddle & Van Driesche, 1999) and sale to retailers in the Christmas poinsettia market in the fall (Hoddle and Van Driesche, in press).

Here we report further results from commercial trials conducted in fall 1996 in Massachusetts in which four commercial poinsettia growers employed E. eremicus for control of whiteflies. The purpose of the trial was to assess the robustness of E. eremicus releases as a means of suppressing whiteflies in commercial poinsettia crops when applied to a wider variety of commercial conditions and when releases were made by growers. At each of four commercial greenhouse ranges, we measured the level of whitefly suppression achieved by releases of E. eremicus compared to whitefly populations treated chemically. At one study site, we made a further comparison to a caged whitefly population not subject to either biological or chemical control. The costs of biological control and chemical control were compared at all four locations.

Materials and Methods

Study Sites and Experimental Design

The study was conducted at four commercial greenhouse growers. Two growers were from the Connecticut River Valley in the western part of Massachusetts (Fairview Farms, Whately; Westover Greenhouses, Chicopee) and two were from eastern Massachusetts (Loosigian Farms, Methuen; Konjoian Greenhouses, Andover). The trial was conducted on the Christmas poinsettia crop between 3 July and 13 December 1996, with cropping periods varying from 17 to 23 weeks among sites. At each of the four locations, weekly observations were made in two greenhouses, one managed with biological control and one with insecticides. In the biological control greenhouses, our intent was to make weekly releases of 3 female E. eremicus per plant. In the chemical control greenhouses, the growers managed pests with pesticides. At 3 sites (Loosigian, Konjoian, and Westover), growers ordered parasitoids directly from commercial insectaries and made releases themselves. At one site (Fairview), we ordered and received parasitoids instead of the grower so we could assess the quality of weekly shipments in terms of number of parasitoid pupae shipped (compared to number ordered) and sex ratio of emerging adult parasitoids. At this site, we made releases and retrieved parasitoid exuviae weekly from release cups in greenhouses to determine the percentage emergence under greenhouse conditions. Greenhouse dimensions, names of poinsettia cultivars grown, numbers of plants and potting arrangements per greenhouse are given in Table 1. ("Plants" refers to independently rooted poinsettias; pots may contain one or several plants.)

To formally demonstrate, at least at one site, that whitefly populations on poinsettia increase sharply if left uncontrolled, a control cage was installed in the biological control greenhouse at Fairview Farms that received neither E. eremicus nor conventional insecticides for whitefly management. The control cage (153 cm long by 92 cm wide and 117 cm tall) was constructed of PVC pipe and covered with fine polyester screening (95 micron dia openings) capable of excluding entrance of whiteflies and parasitoids. The control cage contained 5 pots (20.3 cm), each with 3 poinsettia plants (total, 15 plants per cage). To initialize the caged whitefly population, we inspected all leaves on 100 plants from the greenhouse and chose plants that bore the number of whitefly nymphs and pupae needed to match the density of the whitefly population in the whole greenhouse as determined by a count of whiteflies on the potted cuttings at the start of the trial (see Initial whitefly density). Because initial whitefly densities at this site were very low, we augmented the silverleaf whitefly population in the biological control greenhouse using silverleaf whitefly-infested plants produced by using adult whiteflies from our laboratory colony of this species (see Whitefly augmentation). The numbers of whiteflies in the control cage were also augmented at the same rate so that they had the same starting density as the biological control greenhouse.

While control cages were not installed at the other three trial sites, we have shown in previous trials that silverleaf whitefly on poinsettia typically increases to high densities if not controlled (Hoddle & Van Driesche 1996; Hoddle et al. 1997ab, 1998a). No cage controls were included at sites that proved to be infested with greenhouse, rather than silverleaf, whitefly. Consequently control data showing unrestricted growth for that species in the absence of chemical or biological control were not collected in this trial. However, such growth has been observed in other trials (Helgesen & Tauber 1974, Rumei 1982).

During the trial we collected data on (1) the weekly numbers released, percentage emergence and sex ratio of E. eremicus, (2) the weekly whitefly densities in each greenhouse, (3) the species of whitefly present at each grower, (4) insecticide usage during crop production by each grower, and (5) the quality of plants at harvest (in terms of whitefly infestation).

Crop Management

Source of cuttings, potting dates, spacing, plant removals. Three of four greenhouses potted cuttings between 7 and 20 August. One location (Westover Greenhouses) potted on 3 July in order to produce large ("tree") poinsettias. Table 1 provides details on greenhouse type, size, numbers of plants, pot sizes, and cultivars.

Pesticide use. At three sites, the biological control greenhouse was treated only with fungicides and plant growth regulators. At one site, Konjoian Greenhouses, insecticides were sprayed on 23-30% of the plants in the biological control greenhouse. The infestation on these plants occurred because whitefly-infested plants from another greenhouse on the property were placed directly beneath the intake vent of the biological control greenhouse early in the cropping cycle, leading to a heavy, localized infestation on benches near the air intake fans. Plants sprayed with insecticides were excluded from sampling for the remainder of the trial.

Chemical and biological control greenhouses at all sites were treated with plant growth regulators and fungicides. Names and application dates of insecticides used to control foliar insects in chemical control greenhouses (and a portion of one biological control greenhouse) are presented in Table 2.

Parasitoid releases. Biological control greenhouses at all four sites received weekly releases of E. eremicus for whitefly control. The intended weekly release rate was 3 female parasitoids per plant. When plants were removed from biological control greenhouses for sale, numbers of parasitoids released per greenhouse were reduced accordingly. To avoid conflicts with parasitoids, yellow sticky cards (which are highly attractive to E. eremicus, Sanderson, unpub. data), used by growers to monitor whiteflies and fungus gnats, were not placed in any of the biological control greenhouses.

Whitefly Species Composition, Initial Density Estimate, and Augmentation

Whitefly species. Both B. argentifolii and T. vaporariorum infest poinsettia in Massachusetts. To determine the whitefly species present in each test greenhouse, ten heavily infested leaves were collected at each location in middle of the trial (mid-October). In the laboratory, all fourth instar nymphs, pupae and pupal cases were examined under a dissecting microscope and identified to species. Voucher specimens of whiteflies were not retained as no opportunity exists for taxonomic confunsion in our case. Trialeurodes vaporariorum is distinct in the context of greenhouse crops from all other whiteflies, and all Bemisia whiteflies in poinsettia greenhouse crops in North America are strain B of B. tabaci (= B. argentifolii), as the A strain was known only from outdoor crops and even there has disappeared over the last decade in North America, being replaced completely by the B strain.

Initial whitefly density. In order to determine if initial whitefly population densities in greenhouses designated as biological control greenhouses were within an acceptable range for management using parasitoids (considered by us to be 1.0 or fewer live nymphs, pupae and adults combined per cutting, for B. argentifolii, based on levels seen in our earlier trials, Hoddle et al. 1996, Hoddle and Van Driesche in press), population densities were estimated on cuttings at the time of potting. At each location, all nymphs, pupae, and adults on all leaves of 50 potted cuttings in the biological control greenhouse were counted within 1-2 days of the potting date (see Table 1), and numbers of leaves per cutting were recorded.

Whitefly augmentation. Because no whiteflies were seen on cuttings (n = 100) examined from the biological control greenhouse at one site (Fairview Farms), the whitefly population there had to be augmented by introducing whitefly-infested plants from our laboratory. Our intention was to add a number of immature whiteflies sufficient to bring the per plant density at this site up to the average value of the three other sites. To infest plants, we chose six uninfested poinsettia plants and used ventilated, clip-on leaf cages (2.5 cm dia) to enclose 4-5 pairs of whitefly adults over leaves for 2 days to produce eggs. We then counted the eggs produced and removed excess numbers. Infested plants each had three infested leaves; each infested leaf (after egg removal) bore an average of 105 B. argentifolii eggs (± 8.6 SE, n = 10 leaves counted). Infested plants were placed in the biological control greenhouse at Fairview Farms on 19 August. Initially, all infested leaf patches remained protected from attack by parasitoids within clip cages. One clip cage on each plant was removed on each of 19, 23, and 29 August, allowing for a gradual introduction of the whiteflies into the crop. A total of 1890 eggs (6 plants × 3 patches × 105 eggs per patch) were added to this greenhouse, which contained 1500 plants. We assumed 79% survival to the settled crawler stage (based on cohort survival data in Hoddle et al. 1998a), giving a projected augmented nymphal density of 1.0 nymph per plant, meeting our objective of a density comparable to the average density of the other three biological control greenhouses in the trial (1.05 nymphs per plant).

Parasitoid Sources, Application Methods, and Release Rates

The E. eremicus we used were purchased from commercial suppliers and shipped as parasitized T. vaporariorum fourth instar nymphs packed in sawdust, except for the material used at Fairview Farms. Sawdust was omitted from shipments send to our laboratory for use at this site in order to allow us to retrieve parasitized whiteflies for estimation of parasitoid number per unit weight, sex ratio, and percentage emergence. Over the course of the trial, parasitoids were obtained from two suppliers. From the start of the trial until 4 October, parasitoids were supplied by Beneficial Insectary, Inc. (14751 Oak Run Rd., Oak Run, CA 96069). This colony was discontinued mid-way through the trial, but the same population of E. eremicus was available from Koppert Biological Systems, Inc. in the Netherlands, and parasitoids from this source were used to complete the trial. Koppert's production of this species was initiated with the same material that had been used by Beneficial Insectaries (O. Minkenberg, pers. comm.), so the genetic composition of the parasitoids used in the trial was consistent throughout. Specimens from material sold by Koppert, Inc. as E. californicus (previous name for E. eremicus) were submitted for taxonomic confirmation to Michael Rose (specialist on the genera Eretmocerus and Encarsia, formerly of Texas A & M University) and were confirmed to be E. eremicus. Voucher specimens have been deposited in the insect collection of the University of California, Riverside campus.

Parasitoid pupae were shipped directly to three of the four participating growers because it was intended that processes used in the trial be as close to commercial as possible. Therefore, at three locations growers received parasitoid shipments and placed shipped material in release containers in greenhouses. These growers received parasitized fourth instar T. vaporariorum nymphs mixed with sawdust. This mixture was placed in styrofoam release cups (6 cm tall, 5.5 cm wide at bottom, 8.5 cm wide at top) that had the bottoms cut out and replaced with organdy (mesh 0.95 microns) to allow for drainage. Cups were attached 10 cm above the canopy to wooden stakes (50 cm long) placed in the potting media. In each biological control greenhouse, there were 15 release cups distributed evenly throughout the crop. Each week, growers emptied sawdust and any unemerged parasitoids from the previous week's release into pots of plants on benches where cups were located and then added the new material to the same cups. Watering was done so as to avoid wetting parasitoid pupae in release cups (either drip irrigation was used or workers were advised not to wet sawdust in release cups when hand watering).

To estimate numbers of parasitoids released, parasitoids for use at Fairview Farms were sent to our laboratory for subsampling before release. To estimate the number of parasitoids released, we measured the number of pupae per unit weight of material sent by the supplier, the weight of the shipment, the sex ratio of emerging adults, and the percentage of pupae from which parasitoids successfully emerged under greenhouse conditions.

Estimating number of E. eremicus pupae received from suppliers. Each week before taking parasitoid pupae to Fairview Farms to be released, we counted the number of live parasitoid pupae in each of ten 20 mg subsamples under a stereomicroscope at 25×. The average number of pupae per 20 mg was multiplied by 50 (to get the number per gram) and then by the weight (in g) of all pupae received to determine the total number of pupae actually shipped by the supplier in particular orders. The percentage deviation between this value and the number ordered was noted.

Estimating E. eremicus sex ratio. Each week, 200-300 parasitoid pupae from the shipment sent to our laboratory were placed in a petri dish in a growth chamber at 22°C and long day light regime (16:8 L:D) and held for emergence. One week after receipt, samples were frozen, and 15 groups of 10 adult parasitoids were examined at 50× with a stereomicroscope and their sex determined. Sexes were recognized based on the clubbed antennae of the female (Rose & Zolnerowich 1997).

Estimating E. eremicus emergence rate. Each week before adding new parasitoid pupae to release cups in the biological control greenhouse at Fairview Farm, whitefly nymphs with parasitoid exit holes and remaining dead nymphs in cups from the previous week were retrieved and returned to the laboratory, frozen, and used to estimate the parasitoid emergence rate. From the material returned to the laboratory from each week of the trial, 15 samples of 10 parasitoid "pupae" (comprised of whitefly nymphs containing dead parasitoid pupae and whitefly nymphal integuments with parasitoid emergence holes) were examined at 25× under a stereomicroscope, and classified as dead or emerged based on the presence of parasitoid emergence holes. The percent emergence was calculated as the number of whitefly nymphs with parasitoid emergence holes divided by the total number examined (nymphs containing dead parasitoid pupae plus whitefly nymphal integuments bearing parasitoid emergence holes).

Calculating release rates of E. eremicus. For one site (Fairview Farms) we used the above information on number of parasitoid pupae per unit weight, together with sex ratio and percent emergence, to adjust the number of parasitoid pupae actually released to achieve the intended release rate. At the other three sites, growers received shipments directly and made their own releases, and quality control checks were not made. At these sites, we estimated the number of parasitoid pupae that would be needed to achieve our intended release rate (3 females per plant per wk) by assuming a 50% female sex ratio and a 60% emergence rate. The sex ratio value was based on advice from the supplier and the emergence rate was based on quality control checks we made in greenhouse trials in 1995. Based on these assumptions, 10 parasitoid pupae per plant per week were ordered for each participating grower, with exact numbers being calculated from numbers of plants in each biological control greenhouse. Subsequent to the trial, we calculated the actual release rate achieved by reference to quality control data collected from samples taken for the Fairview Farms site.

Whitefly Population Sampling

Densities of whitefly life stages (adult whiteflies, live and dead nymphs and pupae) were estimated weekly throughout the cropping season by examining leaves on arbitrarily selected plants. At Westover Greenhouses, Konjoian Greenhouses, and Loosigian Farms, two arbitrarily selected mature leaves from each of the upper and lower halves of the plant from each of 30 arbitrarily selected plants (120 leaves total) in each greenhouse were inspected for whiteflies on each sample date.

Numbers of leaves examined in the biological control greenhouse at Fairview Farms differed from that of the other three sites because this greenhouse was also part of a separate, concurrent experiment with a more intense level of sampling. At Fairview Farms in the biological control greenhouse, three leaves (1 from the bottom third of the plant, 1 middle, and 1 top) on 90 plants (270 leaves total) were inspected. In the control cage in the biological control greenhouse at Fairview Farms, three leaves on each of eight plants were inspected in a similar manner, weekly. At Fairview Farms, the chemical control greenhouse was sampled for a shorter period than the biological control greenhouse. Three arbitrarily chosen leaves from each of 20 plants (60 leaves total) were inspected weekly, from 29 August to 13 November only. For figures in which whitefly densities are plotted on log scales, 0.001 was added to all counts to avoid zero values.

Measurement of Parasitoid-Caused Mortality

Whitefly nymphs killed by parasitoids through host feeding were included in counts of dead nymphs or pupae made to estimate densities (see above). Deaths from host feeding could not be distinguished from physiological death. Successful parasitism was scored by noting numbers of visibly parasitized fourth instar whitefly nymphs seen weekly on leaves on which whitefly stages were counted. Because parasitism was rare, weekly samples were not analyzed separately by date because of low sample sizes. Instead, season-long rates of parasitism were computed for each of the four biological control greenhouses by summing the number of visibly parasitized fourth instar whitefly nymphs across all sample dates. Parasitism was computed as the total number (A) of parasitized fourth instar whitefly nymphs summed across all dates within one location, divided by this same value (A) plus the summed value in the same samples of all whitefly pupae (B), (% parasitism = 100[A/ A +B]). Younger whitefly stages (various nymphs) were not included in the estimation of parasitism, as these stages were too young for any parasitism they might have had to have become visible in samples. Parasitism rates were compared statistically between the combined samples of the two biological control greenhouses with T. vaporariorum and those of the two with B. argentifolii.

Whitefly Densities on Plants at Harvest

To compare the quality of plants in the trial to that of plants offered for sale in Massachusetts, we determined the densities of live nymphs, pupae, and adults on plants from the biological control and chemical control greenhouses and on poinsettia plants at 8 retail outlets in Massachusetts in December 1996. Numbers of whiteflies on plants at retail outlets were measured using a standardized market survey sampling protocol used previously in Massachusetts, in which six leaves (2 bottom, 2 middle and 2 top) on 15 arbitrarily selected plants were examined for live whitefly nymphs, pupae, or adults (Hoddle et al. 1997ab, 1998a).

Cost Analysis

To compare the costs of biological and chemical control, we computed the costs of parasitoids versus pesticides used for whitefly control in the biological control and chemical control greenhouses at each trial site. To compute the cost of chemical pest control, grower spray records were examined and all applications of materials to suppress whiteflies were noted. Using 1995 catalog prices for insecticides and label application rates and methods, we computed amounts and cost of insecticide applied in each application. Seasonal expenditures for pesticides were then summed and divided by the number of plants in each greenhouse to obtain a seasonal insecticide cost per plant. To compute the cost of parasitoids we used the 1996 commercial price of $11 per thousand pupae and an application rate of 10 pupae per plant (equal to 3 females per plant, based on an assumed 50/50 sex ratio and 60% emergence rate). Costs of labor for application were not considered for either chemicals or parasitoids (after Hoddle & Van Driesche 1996).

Statistical Analyses

Average seasonal values of parasitoid emergence rate, sex ratio, and release rate at Fairview Farms were compared to assumed or intended values with Student's t test. Densities of whitefly nymphs were compared between chemical and biological control greenhouses (and in one location, to whitefly nymphal densities in a control cage) using nested ANOVAs. A Chi Square test was used to compare rates of parasitism of greenhouse whitefly and silverleaf whitefly nymphs. This comparison was performed on data after pooling across all sample dates for the pair of locations with each whitefly species. A nested ANOVA was used to compare whitefly nymphal densities on leaves from the biological control and chemical control greenhouses to whitefly densities on leaves of plants offered for sale at retail outlets.

Results

Crop Management and Pesticide Use

In the chemical control greenhouses, from 1 to 10 insecticide applications were made per greenhouse for whitefly control (avg. 6.8 ± 1.8 SE, Table 2), with an ave. of 8 applications against T. vaporariorum at two sites and 5.5 applications against B. argentifolii at the remaining two locations. In biological control greenhouses, three growers used no insecticides and one made 7 applications to a portion (about 30%) of the greenhouse (Table 2) to suppress whiteflies drawn in through the air intake vents.

Whitefly Species Composition, Initial Density, and Augmentation

Whitefly species. Of 216 nymphs and 404 pupal exuviae collected 17 October at Loosigian Farms and of 798 nymphs and 242 pupal exuviae collected on the same date at Konjoian Greenhouse, all were T. vaporariorum. In contrast, at Fairview Farms and Westover Greenhouse, all fourth instar nymphs and pupae seen in samples during the trial were B. argentifolii.

Initial whitefly density on potted cuttings. Mean numbers of live nymphs plus pupae per leaf (± SE) found in the initial count on poinsettia cuttings in the biological control greenhouses varied from 0.0 to 1.6 (Fairview Farms [0.0 initially, 1.0 after augmentation], Konjoian Greenho