Predator Rearing

Phytoseiulus macropilis (Banks) were sub-cultured from a laboratory colony established in 1987 by R. F. Mizell, III from a wild population feeding on twospotted spider mites at Monticello, Florida. Mite identification was performed by H. A. Denmark and vouchers specimens are housed in the Florida Collection of Arthropods, Gainesville, Florida. Predators were reared in the laboratory during 1993 and 1994 at 27–34°C, 60 ± 5% relative humidity (R.H.) and a photoperiod of 16:8 light:dark (L:D). Rearing units consisted of a rectangular sponge (length = 23.5 mm, width = 16.5, thickness = 2 mm) placed in a metal tray (length = 27 mm, width = 17.5, height = 3.5 mm). The tray was filled with tap water until the sponge was saturated. A black plastic plate (length = 22 mm, width = 12.3 mm, thickness = 0.2 mm) was seated on the sponge as the rearing arena.

Lima beans, Phaseolus vulgaris L., ‘Henderson’, were grown in a greenhouse (25°C and 60% ± 5% R.H. under fluorescent lights) at the North Florida Research and Education Center in Monticello, Florida. Twospotted spider mites were reared on 1–2-week-old lima bean seedlings at 27–31°C, 60 ± 5% R.H. under fluorescent lights in the laboratory. Predators were fed on the rearing units every 24 h by adding 2–4 lima bean leaves infested with all stages of T. urticae. The old dried leaves from rearing units were discarded after removing predator eggs, nymphs, and adults. Adult female predators 3–5 d old were used in all bioassays.

Residue Bioassays

Residue bioassays were used to determine the toxic effects of selected registered pesticides on P. macropilis by a modified double-leaf-disk method. The double-leaf-disk method described by Schiffhauer & Mizell (1988) was modified so that the ring was closed with a treated leaf instead of cloth. Rings cut from standard PVC pipe (diameter = 20 mm; thickness = 2.5 mm) were attached to a 7-d-old cotyledon leaf of lima bean with contact cement (DAP Inc., Dayton, Ohio). The seedlings with the PVC rings were dipped in the test concentrations for 5 s. After drying seedlings in a hood, eggs and other life stages of T. urticae were placed inside the ring arena with a fine brush. Ten to 15 female predators from the colony were aspirated into a small straw with a vacuum device. The straw was sealed with parafilm and held in ice for about 5 min to slow down predator movement to prevent their escape. Predators were transferred into the ring arena, and the latter was closed on top by the second cotyledon leaf with contact cement. Treated and untreated control seedlings were held in beakers with their roots immersed in water. The number of concentrations tested varied by chemical, but most were tested at 5 serial dilutions of the label rates (Table 1). Six replicates of 10 predators were tested for each concentration of 12 acaricides. Untreated control groups (6 replicates of 10 predators), were included with each test. After 72 h, the top leaf from each ring was removed and mortality was assessed under a steromicroscope by a gentle touch to the predators with a fine probe. Predators that could not walk one body length when probed were counted as dead. These bioassays were carried out at 27–34°C and 50± 15% R.H. under fluorescent lighting. Toxicity was rated as in IOBC/WPRS (Hassan et al. 1987). Dose-response lines, LC₅₀ values and slopes were calculated for 5 acaricides by the Probit Procedure in SAS (SAS 1994).

Field Efficacy of Fenpropathrin and Dicofol Residues

The pesticides, fenpropathrin and dicofol, were applied to Euonymus japonica Thunb to runoff at the 1× rate (Table 1) with a backpack sprayer. Seven to 10 cm long shoots were cut from treated plants and leaf replicates and numbers of predators tested were as described in the lima bean bioassays at 0, 3, 7, 14, and 21 d in order to determine the mortality of the residues to P. macropilis. The final mortality was adjusted for untreated control mortality by Abbott's formula (Abbott 1925).

Olfactometer and Response to Prey-host Emitted Volatiles

A glass Y-tube olfactometer of diameter 1.1 cm with arms 8.5 cm long (Mizell & Schiffhauer 1991) was used to observe the behavioral response of gravid female P. macropilis to plant-and prey-emitted volatiles. The central arm of the Y-tube was connected to an air outlet and a Nal-gene water vacuum pump (11.5 L/minute capacity, NO. 6140–0010) on a water spigot via a plastic tube. Each end of the Y-tube was connected to adjustable plastic containers that contained the test treatments. Air flowed through a canister of activated charcoal into the adjustable plastic containers and through the Y-tube. Air flow was adjusted with flowmeters at 100 mL/min in both arms of the Y-tube. Treatments were placed upwind in the plastic containers. The Y-tube was placed horizontally on a wooden block over white paper to improve observation of predator movement. Adult female predators were placed in the main arm of the olfactometer with a fine probe.

Predatory Mites

Adult female P. macropilis reared on T. urticae Koch on ‘Henderson’ lima bean were starved for 24 h before the tests. Thirty-six predators (used only once) were tested individually for each treatment. Each predator was observed for a maximum of 10 min. After 10 predators were tested, the treatments in the olfactometer arms were rotated to remove possible positional effect(s) and the olfactometer was washed with acetone. Predators that did not choose one of the arms of the olfactometer after 10 min were removed and scored as no response. Predators were scored ‘+’ or ‘-’ when they reached the upper end of one of the Y-arms of the olfactometer. The tests were carried out at 27–34°C and 60–80% R.H., under fluorescent light from 1130–1800 EDT The following treatments were tested to observe the response of P. macropilis to kairomonal cues from the prey and host plant: (1) blank versus blank (to test the validity and potential positional bias of the olfactometer); (2) blank vs uninfested lima bean leaves; (3) T. urticae infested lima bean leaves vs blank; (4) T. urticae-infested bean leaves vs uninfested lima bean leaves; (5) lima bean leaves previously infested by T. urticae but with all traces of mites removed with a light brushing vs blank; (6) blank vs uninfested juniper leaves; (7) O. ununguis infested juniper leaves vs uninfested juniper leaves; (8) T. urticae on lima bean leaves vs O. ununguis on juniper leaves; (9) T. evansi infested tomato leaves vs blank; (10) T. evansi on filter paper vs blank; and (11) T. urticae on lima bean leaves vs T. evansi on tomato leaves. The results were statistically analyzed by the Sign-test at P = 0.05 (Conover 1971).

Alternative Food Bioassays

Selected non-prey food: water, water and sugar solution (16:1), daylily pollen Hemerocallis sp., eggs of false oleander scale Pseudaulacaspis cockerelli (Cooley), and Phylloxera sp. nymphs were presented to adult female predators to study their effect on survival and fecundity. The experimental units consisted of an inverted plastic Petri dish (5.3 mm dia.) closed with silicone (Alex Plus, DAP Inc. Dayton, Ohio 45401). A moistened filter paper placed at the bottom retained humidity and also served as the arena for the subjects. The inverted dish had 2 holes at the base; one was used for predator and food placement (fitted with a cork) and the other was fitted with a stub of filter paper to supply moisture. Food was provided every 24 h. The number of eggs and live predators were recorded per 24 h. Five replicates of 6 satiated, gravid female P. macropilis were used for each treatment. The treatments were continued until the tested predators died. A control group of 5 replicates of 6 predators with all life stages of T. urticae as food was included with each treatment. Bioassays were conducted at 27–34°C and 60–80% R.H. Results were analyzed by the randomization test at P = 0.05 (Conover 1971).

TABLE 1.

PESTICIDES TESTED FOR THEIR EFFECT ON ADULT PHYTOSEIULUS MACROPILIS.

Residue Bioassays

Synthetic pyrethroids were highly toxic as permethrin, fluvalinate, and fenpropathrin at1X label rates caused 100% mortality (Table 2). However, in the field efficacy test, fenpropathrin toxicity decreased to 30% at 21 d after application (Table 3). Other studies of synthetic pyrethroid toxicity to phytoseiids reported similar results (Croft et al. 1983; Riedl & Hoying 1983; Mizell & Schiffhauer 1991). Fenpropathrin under field conditions showed high toxicity during the 2 weeks after application, but its toxicity declined to 30% after 21 d (Table 3). Thus, P. macropilis might be released into nurseries 3–4 weeks following treatment with fenpropathrin.

Dienochlor at the 1X concentration caused high mortality, but at lower concentrations mortality declined sharply (Table 3). These results agree withreports of Mizell & Schiffhauer (1991) for N. collegae De Leon, and Malezieux et al. (1992) for N. fallacis (Garman). The LC₅₀ value was 335.54 mg a.i./L, which is more than half of the registration field rate (480 mg a.i./L). Phytoseiulus macropilis is relatively unharmed by exposure to dienochlor, which is an organochlorine compound no longer available for use (Table 4).

TABLE 2.

RESPONSE OF PHYTOSEIULUS MACROPILIS FEMALES AFTER 72 H TO COMMERCIAL PESTICIDES AT LABEL RATES IN A LABORATORY BIOASSAY.

The insecticide Diazinon surprisingly caused only 18% mortality whereas acephate caused only 5% mortality at 1× rate and was the safest of all pesticides tested (Table 2). Nevertheless, Hassan et al. (1987) reported acephate as highly toxic to Phytoseiulus spp. Similarly, Diazinon has been observed as highly toxic to A. hibisci (Bartlett 1964) and A. fallacis (Croft & Nelson 1972). However, results similar to ours were reported by Babcock & Tanigoshi (1988) on T. occidentalis and Zacharda & Hluchy (1991) on T. pyri Scheuten. Apparently both insecticides could be used in nurseries targeted to pest species other than mites and cause little harm to P. macropilis.

The acaricide dicofol at the 1× rate caused 100% mortality and the LC₅₀ was very low 0.083 mg a.i./L in comparison to the label rate. Under field conditions, however, residual toxicity declined 7 to 14 d after application to 12% and 1%, respectively (Table 3). Heretofore, this chemical has been reported as nontoxic (Rock & Yeargan 1971; Theiling & Croft 1988), slightly toxic (MacPhee & Sanford 1961), moderately toxic (Bartlett 1964; Zacharda & Hluchi 1991) and highly toxic to phytoseiids (Van de Vrie 1962; Hassan et al. 1987; Mizell & Schiffhauer 1991). Based on our observation of dicofol in the field, it appears that P. macropilis could be safely integrated with this chemical for tetranychid suppression in nurseries 7–14 d after its application.

The acaricides abamectin and hexakis at the 1X rate caused high P. macropilis mortality (Table 4). The LC₅₀ value for hexakis was 74.54 mg a.i./L and for abamectin 0.7513 mg a.i./L. Propargite at the 1X rate caused low mortality (Table 4). The acaricide propargite was harmless, both at the label rate and at concentrations 2, 3, and 6-fold of the label rate. In contrast, propargite was reported slightly toxic to A. hibisci (Bartlett 1964) and A. fallacis (Croft & Nelson 1972). However, these results are similar to those reported by Rock & Yeargan (1971) on T. occidentalis, Babcock & Tanigoshi (1988) on N. fallacies, and Mizell & Schiffhauer (1991) on N. collegae. The LC₅₀ was recorded as 8917 mg a.i/L, more than 10 times the label rate.

TABLE 3.

RESIDUE TOXICITY OF 5 ACARICIDES TO PHYTOSEIULUS MACROPILIS.

TABLE 4.

RESPONSE OF P. MACROPILIS TO RESIDUES OF FENPROPATHRIN AND DICOFOL (AT LABEL RATES) UNDER COMMERCIAL NURSERY CONDITIONS.

TABLE 5.

RESPONSE OF ADULT FEMALE PHYTOSEIULUS MACROPILIS TO ODORS IN AN OLFACTOMETER PRODUCED BY TETRANYCHID MITE SPECIES. STATISTICAL SIGNIFICANCE IS SCORED WITH RESPECT TO PREFERENCE FOR THE FIRST TREATMENT IN THE COUPLETS. ZERO AND N.S. APPLIES TO NO SIGNIFICANT DIFFERENCE BETWEEN THE TREATMENTS, ++ EXPLAINED AS COMMENTS.

The fungicides thiophanate and thiophanate-methyl at the 1X rates caused 58% and 35% mortality, respectively (Table 2). The predator's tolerance to tested fungicides is important relative to its use in nurseries, as fungicides are frequently applied for disease control.

In this study a double-leaf-disk method was used to avoid ‘runoff’ mortality. The method used was similar to Schiffhauer & Mizell (1988), with the difference that the ring was closed with a treated leaf instead of a cloth. This modification provided more exposure of the treated surface to the predators. Phytoseiulus macropilis, like many other phytoseiids, has a wandering tendency, and is difficult to confine in a treated arena during experiment set up, often trying to escape the treated arena and is killed in the contact glue. This problem was countered by drawing predators into straws, and putting the latter in ice for about 5 min. This cooling technique (Mizell & Schiffhauer 1991) facilitated predator transfer into the treated arena. However, the double-leaf-disk method, like any other residue testing technique, has some tradeoffs. For example, behavioral response of subjects to pesticide residue cannot be observed. The possible fumigation effect inside the arena may cause some mortality among confined subjects that may not occur on an open treated surface. However, possible fumigation effects in this study were deemed negligible in comparison to the problems with predator escapes from other arenas.

Olfactometer Bioassays

Of the 38 P. macropilis tested, 22 responded significantly (P < 0.05) to lima bean leaves infested with T. urticae, whereas none of the P. macropilis responded to uninfested lima bean leaves (P < 0.05). The response was also significantly different (P < 0.05) when 1 treatment was lima bean leaves infested with T. urticae and the other blank. The P. macropilis response to T. urticae-in-fested lima bean versus a blank (25:2) and to lima bean leaves that were previously infested by T. urticae (20:0) to a blank were both significantly different (P < 0.05) (Table 5). The P. macropilis did not respond significantly to O. ununguis on juniper leaves both versus uninfested juniper leaves and versus the blank (Table 5). Phytoseiulus macropilis were not attracted significantly to T. evansi either on tomato leaves or on filter paper (Table 5).

The significant response of adult P. macropilis to lima bean leaves infested with T. urticae Koch, both versus blank and uninfested lima bean leaves, was similar to the results reported by Hislop & Prokopy (1981). Predators did not respond significantly to uninfested lima bean leaves. Sabelis et al. (1984) also reported that P. persimilis did not respond to uninfested lima bean leaves. However, P. macropilis responded significantly to infested lima bean leaves from which T. urticae were removed.

Takabayashi & Dicke (1992) observed that phytoseiids reared on twospotted spider mite on lima bean leaves respond to volatiles from lima bean leaves but the predators reared on the twospotted spider mites from cucumber leaves did not respond to lima bean volatiles. Phytoseiulus macropilis used in this study were reared on ‘Henderson’ lima beans for about 7 years. Thus, the significant response of P. macropilis to the kairomones of T. urticae on lima bean leaves versus the negative response to O. ununguis and T. evansi on juniper and tomato leaves, respectively, may possibly be explained by the rearing history. The response of P. macropilis to T. evansi was not significant, even when the latter were offered on filter paper, which removed the effect of the tomato leaves. Thus, it appears odors produced by T. evansi on tomato do not attract P. macropilis, or perhaps volatile(s) carried over by T. evansi from feeding on tomato leaves made them repellent. Results regarding the response of P. macropilis to these 2 tetranychid species warrant further investigation (van Wijk et al. 2008). The significant response of P. macropilis to one of the predominant phytophagous pest of nursery ornamentals T. urticae indicates that P. macropilis has potential as a biological control agent for twospotted spider mite and warrants further study.

Alternate Food Bioassays

There was no significant difference between the survivorship and fecundity (P < 0.05) of predators fed on alternate food and the controls after 24 h (Table 6). However, after 48 hr, mortality increased and fecundity decreased sharply for P. macropilis in the alternate food treatments, whereas the P. macropilis controls fed spider mites remained alive and continued ovipositing (Table 6). These results suggest that P. macropilis cannot survive or oviposit when fed only on non-prey food. The rate of mortality, however, varied among predators fed on different alternate foods (Table 6). Interestingly, the highest observed survivorship for predators after 48–72 h was for those fed on the solution of sugar and water. The predators provided with sugar-water solution survived up to 120 h, which was the maximum survival period in any of the alternate food treatments. Similarly, egg production decreased rapidly for treatments after 24 h except for the predators that were fed on the eggs of false oleander scale P. cockerelli and on sugar-water solution where they oviposited for up to 72 h. Apparently, the decreased egg production among treated predators indicates either that the predators' food reserves were exhausted after 24 h or foods other than T. urticae are nutritionally deficient for egg production. These results are similar to Kennett & Homai (1980), who reported that P. persimilis and T. occidentalis provided with food other than tetranychid prey did not oviposit.