Biological Control 58 (2011) 222–229
Contents lists available at ScienceDirect
Biological Control journal homepage: www.elsevier.com/locate/ybcon
Does Bdellodes lapidaria (Acari: Bdellidae) have a role in biological control of the springtail pest, Sminthurus viridis (Collembola: Sminthuridae) in south-eastern Australia? John M.K. Roberts a,d,⇑, Andrew R. Weeks a,b, Ary A. Hoffmann a,c, Paul A. Umina b,c a
Department of Genetics, The University of Melbourne, Parkville, Victoria 3010, Australia CESAR, 102/55 Flemington Road, North Melbourne, Victoria 3051, Australia Department of Zoology, The University of Melbourne, Parkville, Victoria 3010, Australia d CSIRO Ecosystem Sciences, Canberra, ACT 2601, Australia b c
a r t i c l e
i n f o
Article history: Received 3 August 2010 Accepted 6 June 2011 Available online 13 June 2011 Keywords: Pesticide Lucerne ﬂea Pasture snout mite Control
a b s t r a c t Throughout southern Australia, the lucerne ﬂea, Sminthurus viridis (Collembola: Sminthuridae), is an important pest of a variety of winter grain crops and pastures. The predatory mite, Bdellodes lapidaria (Acari: Bdellidae), co-occurs with S. viridis and is reported to be a biological control agent of this pest. Using laboratory bioassays and ﬁeld experiments, we assessed the susceptibility of B. lapidaria to several pesticides and investigated its impact in controlling S. viridis. In the laboratory, B. lapidaria was found to be susceptible to the synthetic pyrethroids, a-cypermethrin and bifenthrin, but relatively tolerant to the avermectin, abamectin, and organophosphorous chemicals, omethoate and chlorpyrifos. In ﬁeld experiments, B. lapidaria was not adversely affected by applications of either bifenthrin or omethoate. Despite strong intraspeciﬁc interactions, we found no detectable impact of B. lapidaria on S. viridis numbers in the ﬁeld. These results indicate that B. lapidaria has a relatively high tolerance to several pesticides, perhaps partly through behavioural avoidance, but little impact as a biological control agent on S. viridis in southeastern Australia. Ó 2011 Elsevier Inc. All rights reserved.
1. Introduction Throughout southern Australia, the lucerne ﬂea, Sminthurus viridis (Collembola: Sminthuridae), is an important pest of winter grain crops and pastures. Emerging in autumn from summer-diapausing eggs, S. viridis is often problematic during the establishment phase of crops and pastures but can also reach damaging levels in spring (Bishop et al., 2001). Control of this pest is heavily dependent on the use of broad spectrum pesticides. However, due to the increasing pest status of S. viridis and ever present threat of pesticide resistance, there is mounting pressure for greater integration of cultural and biological control options (Hoffmann et al., 2008). Several studies have identiﬁed various predators of S. viridis, although few have been considered as control agents (Holdaway, 1927; MacLagan, 1932; Walters, 1966). Most research has been conducted on two snout mite species, the pasture snout mite (Bdellodes lapidaria, Acari: Bdellidae) and the spiny snout mite (Neomolgus capillatus, Acari: Bdellidae) (Ireson and Webb, 1995; Wallace, 1967, 1974). Despite concerted efforts at establishing these two predatory mites as biological control agents their ⇑ Corresponding author at: CSIRO Ecosystem Sciences, Black Mountain Laboratories, Clunies Ross Street, Canberra, ACT 2601, Australia. E-mail address: [email protected]
(J.M.K. Roberts). 1049-9644/$ - see front matter Ó 2011 Elsevier Inc. All rights reserved. doi:10.1016/j.biocontrol.2011.06.007
effectiveness in suppressing S. viridis on a broad scale remains questionable. There are numerous bdellid mite species in Australia (Atyeo, 1963). Many of these are native, although some including B. lapidaria and N. capillatus have been introduced. N. capillatus was introduced to Australia in the 1960s in an attempt to extend the biological control S. viridis (Wallace, 1974) but the mite has failed to become widely established in many regions and is not considered in this study. B. lapidaria was ﬁrst described feeding on S. viridis in Western Australia in 1931 and is thought to have been accidentally introduced from Europe (Womersley, 1933). The majority of studies reporting B. lapidaria as an effective biological control agent of S. viridis have originated from Western Australia (Currie, 1934; Wallace, 1967). The most convincing empirical evidence was provided by Wallace (1967), who indicated that the presence of at least 20 B. lapidaria per square metre in early winter could suppress S. viridis numbers in spring. However, in other areas of southern Australia there has been little evidence to support the importance of this predatory mite as a biological control agent (Ireson, 1984; Pescott, 1937; Swan, 1940). In south-eastern Australia, B. lapidaria has an overlapping distribution with S. viridis (Wallace and Mahon, 1971). The only study to assess the effectiveness of B. lapidaria against S. viridis in this region
J.M.K. Roberts et al. / Biological Control 58 (2011) 222–229
was conducted in Tasmania and failed to identify a signiﬁcant predator–prey relationship (Ireson, 1984). It was suggested that climatic factors and the abundance of alternative prey may limit the effectiveness of B. lapidaria. In Victoria, where the climate appears largely favourable for B. lapidaria (Wallace and Mahon, 1971), other factors might also inﬂuence this predator–prey relationship. For example, S. viridis has been found to increase after the application of some pesticides. Michael (1991) and Wallace (1967) have speculated that this could be the direct result of pesticides on numbers of B. lapidaria, although there is little information on the relative susceptibility of this predator to chemicals. Here we assess the susceptibility of B. lapidaria to several broad spectrum pesticides and compare this with S. viridis under laboratory and ﬁeld conditions. Comparison with the sympatric pest mite, Halotydeus destructor (Acari: Penthaleidae), was included as a positive control. H. destructor is considered broadly susceptible to pesticides and application rates of many registered chemicals in Australia are based on this pest (Hoffmann et al., 1997; Roberts et al., 2009). In addition, we use ﬁeld exclusion plots to examine the predator–prey relationship of B. lapidaria and S. viridis in the absence of pesticides.
translaminar activity which may reduce impact on non-target organisms. The ﬁeld rate we used was 0.108 g/L, which was based on rates used for the two-spotted mite (Tetranychus urticae). To assess the response of B. lapidaria to pesticides, the laboratory bioassay described in Roberts et al. (2009) was used, based on Hoffmann et al. (1997). Brieﬂy, glass vials were coated with a solution of a particular concentration, ranging from 1 104 to 10 times the recommended ﬁeld rate, or with water for controls, and then left to dry for at least 12 hrs. A single adult arthropod was added to each dry vial, together with a vetch leaf (Vicia sativa cv. Blancheﬂeur) to provide food and maintain humidity. Vials were then sealed with laboratory Paraﬁlm MÒ (Bacto Laboratories, Mt. Pritchard, NSW) and left for 8 hrs at 18 °C. After the treatment period, individuals were scored as alive (actively moving), dead (no movement) or incapacitated (inhibited movement). Incapacitated arthropods were pooled with dead individuals for analysis as they are unlikely to survive and contribute to the next generation (Hoffmann et al., 1997). Simultaneous bioassays were conducted on S. viridis and H. destructor for comparison with B. lapidaria. Because H. destructor is generally susceptible to pesticides (Hoffmann et al., 1997; Roberts et al., 2009), comparison with this species provides greater relevance to the tolerance levels of the other species.
2. Methods 2.3. Field assessment of pesticides and predator efﬁcacy 2.1. Collection of arthropods Collections of S. viridis, H. destructor and B. lapidaria were made during the winter growing season of 2008 and 2009 from various sites in Victoria, Australia. Sites were selected that had no recorded pesticide applications for the past 5 years. For laboratory bioassays undertaken in 2008, samples were collected from established pasture at Irrewarra in western Victoria (38°170 05 S, 143°380 27 E). For the ﬁeld experiment conducted in 2009, samples of S. viridis were collected from a lucerne paddock near Inverleigh in western Victoria (37°070 23 S, 144°460 10 E) and B. lapidaria were collected from grasses under native vegetation adjacent to a pasture paddock near Lanceﬁeld in central Victoria (38°080 55 S, 144°000 56 E). All collections were made using a Stihl SH55 blower vacuum and transferred to the laboratory in containers with plant material and paper towel (to absorb excess moisture). S. viridis, H. destructor and B. lapidaria were separated and stored at 4 °C for a maximum of ﬁve days before screening in laboratory bioassays (see below). Samples used in the ﬁeld experiment were added to the ﬁeld plots within 24 hrs after collection. Suction sampling does not signiﬁcantly impact the condition of arthropods and is a commonly used sampling technique (Roberts et al., 2009; Umina and Hoffmann, 2005; Wallace and Walters, 1974).
Two ﬁeld experiments were conducted simultaneously in 2009 at Irrewarra, Victoria (38°170 05 S, 143°380 27 E). The ﬁrst experiment aimed to assess the efﬁcacy of omethoate (290 g a.i./L) and bifenthrin (250 g a.i./L) against B. lapidaria and S. viridis under ﬁeld conditions. A second experiment was performed to examine the predatory impact of B. lapidaria on S. viridis in the absence of pesticides. For each experiment, 18 ﬁeld plots (2 2 m2) were constructed in June 2009 following Weeks and Hoffmann (1998) using white corﬂute sheeting (120 cm high, 5 mm thick) inserted approximately 15 cm into the soil and attached to wooden garden stakes. Movement of B. lapidaria and S. viridis into plots was prevented by applying Tac-GelÒ (Rentokil, active ingredient polybutene) to the top edge of the sheeting and keeping vegetation in and around plots low. Vegetation consisted of subterranean clover (Trifolium subterraneum), capeweed (Arctotheca calendula) and various grasses. Variable numbers of S. viridis and low numbers of B. lapidaria (<15/m2) and H. destructor (<50/m2) were present in most plots. In July 2009, a further 500 S. viridis were collected and added to each plot. H. destructor were not added to plots because they are not prey of B. lapidaria (Michael et al., 1991) and could potentially compete with S. viridis (Michael, 1991). 2.4. Pesticide effects
2.2. Chemicals tested and laboratory bioassays In laboratory bioassays S. viridis, H. destructor and B. lapidaria were tested for their response to two synthetic pyrethroids, acypermethrin (Fastac DuoÒ, Nufarm, Laverton North, VIC, Australia) and bifenthrin (Talstar 250ECÒ, FMC, Pinkenba, QLD, Australia), as well as two organophosphorous chemicals, omethoate (Le-MatÒ, Bayer Crop Protection, Pymble, NSW, Australia) and chlorpyrifos (Lorsban 500ECÒ, Dow AgroSciences, Frenchs Forest, NSW, Australia). A selective acaricide, abamectin (VertimecÒ, Syngenta Crop Protection, North Ryde, NSW) was also examined. Bifenthrin and a-cypermethrin are broad spectrum contact pesticides with recommended ﬁeld rates of 0.1 g/L of the active ingredient. Omethoate and chlorpyrifos are also broad-spectrum contact pesticides with 0.29 g/L of omethoate and 0.35 g/L of chlorpyrifos at the recommended ﬁeld rates. Abamectin is an avermectin that acts via ingestion and contact to paralyse pests. It also has strong
Eighteen ﬁeld plots were used to assess pesticide efﬁcacy against B. lapidaria and S. viridis. In August 2009, approximately 200 B. lapidaria were collected and added to each plot. Using an INTER 16 knapsack sprayer (SPACEPAC Industries Pty. Ltd., Wollongong, NSW, Australia), a treatment of omethoate at the recommended ﬁeld rate was applied to seven plots and an application of bifenthrin at the recommended ﬁeld rate was applied to six plots on the 10th of September. An equal number of plots were intended for each treatment, but an additional plot was inadvertently treated with omethoate. The remaining ﬁve plots were left as untreated controls. A pre-treatment sample was collected on the 10th September, and four post-treatment samples were taken on the 15th September, 29th September, 9th October and 18th November. Sampling of plots involved using a Stihl SH55 blower vacuum to take three suction samples within a 30 30 cm2 frame and recording numbers of B. lapidaria, S. viridis and H. destructor in
J.M.K. Roberts et al. / Biological Control 58 (2011) 222–229
a plastic tray. All species were returned to the plots immediately after each sample recording. 2.5. Predatory effects The remaining 18 ﬁeld plots were used to assess the effect of B. lapidaria on S. viridis. In August 2009, different densities of B. lapidaria were created where ﬁve plots each received 400, 100, or 50 predatory mites, and three plots had no additional B. lapidaria added. All ﬁeld plots were then sampled on the 28th August, 10th September, 15th September, 29th September and 9th October. Sampling again involved taking three suction samples within a 30 30 cm2 and recording the number of B. lapidaria, S. viridis and H. destructor before returning all species to the plot.
Fig. 1. Dose–response curves of Bdellodes lapidaria to omethoate (), chlorpyrifos (j), bifenthrin (N), a-cypermethrin (), and abamectin (s). Concentration is relative to the proportion of recommended ﬁeld rates for each pesticide.
2.6. Data analysis Dose–response curves were generated from the laboratory bioassay results using logit analyses as outlined in Robertson and Preisler (1992). Chi-square tests found that logit models provided an adequate ﬁt to the data in all cases. From these dose–response curves, regression coefﬁcients and LD50 estimates with 95% conﬁdence intervals were generated. Regression coefﬁcients were compared using the GT2-method (Sokal and Rohlf, 1995). For the pesticide ﬁeld experiment, data were log10 transformed to achieve normality prior to analysis. The overall effects of treatment on arthropod numbers were assessed using a repeated measures analysis of variance (ANOVA) on post-treatment samples, using the pre-treatment sample as a covariate. Following this, differences in arthropod numbers between treatments for each sample date were assessed using one-way ANOVA’s and Tukey’s-b post-hoc tests, again with the pre-treatment sample as a covariate. For the predator efﬁcacy experiment, all data was log10 transformed to achieve normality prior to analysis. Reproductive output (RO) was calculated for each species in each plot (Stearns, 1992) by dividing the numbers in the ﬁnal sample (Nb) by the numbers in the initial sample (Na). Multiple regression was used to determine if the number of each species in the ﬁnal sample (Nb) or the RO of each species was dependent on initial numbers (Na) of either species. We controlled for initial numbers (Na) of the opposing species by including them as the second dependent variable in the multiple regression. Linear regression was also used to assess the relationship between B. lapidaria and S. viridis numbers for each sample and to determine if the RO of S. viridis was dependent on the RO of B. lapidaria. Mean S. viridis numbers in the ﬁnal sample were also compared between ﬁve plots with highest initial B. lapidaria numbers and ﬁve plots with lowest initial numbers using t-tests. All analyses were conducted in SPSS v14 for Windows (SPSS Inc., Chicago, USA). 3. Results 3.1. Dose–response curves For all pesticides, there was no mortality in the control vials for any of the three species tested. Fig. 1 shows the dose–response curves for B. lapidaria to the different chemicals expressed relative to the ﬁeld rate. The curves of both synthetic pyrethroids were shifted to the left, suggesting that B. lapidaria is particularly susceptible to these pesticides. The curves of omethoate and abamectin were shifted to the right, suggesting these pesticides are least toxic to B. lapidaria. Susceptibility to chlorpyrifos fell between these patterns. Dose–response curves of B. lapidaria for each chemical were also compared with those of S. viridis and H. destructor (Fig. 2). For the
synthetic pyrethroids, a-cypermethrin and bifenthrin, the curves of B. lapidaria and H. destructor overlapped and were to the left of the curve for S. viridis. For omethoate, chlorpyrifos and abamectin, the curve of B. lapidaria was to the right of S. viridis and H. destructor. These patterns suggest that B. lapidaria is more tolerant of abamectin and the organophosphorus chemicals than the pests S. viridis and H. destructor, but more susceptible to the two synthetic pyrethroids. 3.2. LD50 estimates and regression coefﬁcients The dose–response data for each species were analysed with logit models which provided an adequate ﬁt based on chi-square tests in all cases. From these models, regression coefﬁcients and LD50 estimates with 95% conﬁdence intervals were obtained for each pesticide (Table 1). Using the GT2-method, signiﬁcant differences in regression coefﬁcients were only observed for chlorpyrifos. The regression coefﬁcient of S. viridis (b = 8.726 ± 2.204) was signiﬁcantly higher than both B. lapidaria (b = 2.040 ± 0.376) and H. destructor (b = 3.779 ± 1.070). For the synthetic pyrethroids, LD50 estimates for S. viridis were signiﬁcantly higher than for the mite species, with B. lapidaria having only a marginally higher estimate than H. destructor. For omethoate, B. lapidaria showed the highest LD50 estimate, being 8 and 93 times greater than the estimates for S. viridis and H. destructor, respectively. Similarly, B. lapidaria had signiﬁcantly higher LD50 estimates for chlorpyrifos and abamectin, while S. viridis had lower LD50 estimates for these chemicals than H. destructor. 3.3. Pesticide efﬁcacy in the ﬁeld Numbers of B. lapidaria were not signiﬁcantly affected by either pesticide treatment (Fig. 3). Despite a small decline in numbers, the repeated measures ANOVA found no signiﬁcant overall effect of treatment for the four post-treatment samples (F2,14 = 3.34, P = 0.06). There were also no signiﬁcant differences between treatments at any post-treatment sample (P > 0.05). This result was in stark contrast to the laboratory bioassays, which indicated B. lapidaria was susceptible to bifenthrin. Although not intended in this study, the toxicity of these chemicals against H. destructor was noticeable. A signiﬁcant decrease in mite numbers in the posttreatment sample for bifenthrin (F1,10 = 7.56, P = 0.02) and omethoate (F1,11 = 16.46, P = 0.002) conﬁrmed the efﬁcacy of the two treatments. Consistent with the laboratory bioassays, bifenthrin did not signiﬁcantly reduce numbers of S. viridis. There were no signiﬁcant differences in S. viridis numbers for the ﬁrst three post-treatment samples (P > 0.05). In the last post-treatment sample, S. viridis numbers were actually higher than the controls (F2,17 = 5.30,
J.M.K. Roberts et al. / Biological Control 58 (2011) 222–229
than the control at each of the ﬁrst three post-treatment samples (P < 0.001). However, by the last post-treatment sample, S. viridis numbers had increased to similar levels as the controls.
(a) 100 80 60
3.4. Predatory assessment of B. lapidaria
40 20 0 0
0.1 0.3 1 3 10 Concentration (mg/L)
(b) 100 80 60 40 20 0 0
0.1 0.3 1 3 Concentration (mg/L)
(c) 100 80 60 40 20 0 0
0.03 0.09 0.29 0.87 2.9 8.7 29 87 Concentration (mg/L)
290 870 2900
(d) 100 80 60 40 20 0 0
4.2 7 11.2 14 42 Concentration (mg/L)
(e) 100 Mortality (%)
B. lapidaria did not exert a signiﬁcant predatory impact on S. viridis. A t-test between plots with the ﬁve highest and ﬁve lowest initial B. lapidaria numbers found no difference in the RO of S. viridis (t = 0.69, df = 4, P = 0.26) and linear regressions found numbers of B. lapidaria and S. viridis were independent at each sample. Furthermore, no signiﬁcant relationship was detected by multiple regression between ﬁnal S. viridis numbers and initial B. lapidaria numbers (R2 = 0.13, b = 0.000067 ± 0.14, t = 0.00049, P = 0.99), using initial S. viridis numbers as the second dependant variable (b = 0.34 ± 0.21, t = 1.67, P = 0.11) (Fig. 4). Similarly, the RO of S. viridis was not signiﬁcantly inﬂuenced by the RO of B. lapidaria (R2 = 0.043, b = 0.16 ± 0.16, t = 0.97, P = 0.34) or by initial B. lapidaria numbers (R2 = 0.35 b = 0.000067 ± 0.14, t = 0.00049, P = 0.99) when using initial S. viridis numbers as the second dependant variable (b = 0.66 ± 0.21, t = 3.19, P = 0.0047). In contrast, S. viridis had a signiﬁcant impact upon B. lapidaria numbers. Both the RO and ﬁnal numbers of B. lapidaria were lower in ﬁeld plots with higher initial S. viridis numbers. Multiple regression found a signiﬁcant negative relationships between the RO of B. lapidaria and initial S. viridis numbers (R2 = 0.72, b = 0.44 ± 0.18, t = 2.45, P = 0.024) (Fig. 5) and also between ﬁnal numbers of B. lapidaria and initial S. viridis numbers (b = 0.44 ± 0.18, t = 2.45, P = 0.024), using initial B. lapidaria numbers as the second dependant variable in both instances (b = 0.72 ± 0.12, t = 6.00, P < 0.0001; b = 0.28 ± 0.12, t = 2.35, P = 0.029, respectively). There was also strong intra-speciﬁc competition found for both species. The RO of S. viridis and RO of B. lapidaria were both lower when initial numbers of S. viridis and B. lapidaria, respectively, were higher. Multiple regressions found a signiﬁcant negative relationship between the RO of S. viridis and initial S. viridis numbers (R2 = 0.35, b = 0.66 ± 0.21, t = 3.19, P < 0.01), using initial B. lapidaria numbers as the second dependent variable (b = 0.000067 ± 0.14, t = 0.00049, P = 0.99) (Fig. 6), and also between the RO of B. lapidaria and initial B. lapidaria numbers (R2 = 0.72, b = 0.72 ± 0.12, t = 5.99, P < 0.001), using initial S. viridis numbers as the second dependent variable (b = 0.44 ± 0.18, t = 2.45, P = 0.024) (Fig. 7). Examination of the partial regression coefﬁcients can provide additional insight into the type of competition occurring within species (Umina and Hoffmann, 2005; Weeks and Hoffmann, 2000). Using the GT2-method (Sokal and Rohlf, 1995), the partial regression coefﬁcients of both species were not signiﬁcantly different from a slope of 1, which is indicative of contest competition (Nicholson, 1954). 4. Discussion
40 20 0 0
3.24 10.8 32.4 108 Concentration (mg/L)
Fig. 2. Dose–response curves of Sminthurus viridis (s), Halotydeus destructor (j), and Bdellodes lapidaria (N), to (a) a-cypermethrin, (b) bifenthrin, (c) omethoate, (d) chlorpyrifos, and (e) abamectin.
P = 0.02). In contrast, a repeated measures ANOVA found a signiﬁcant overall effect on S. viridis numbers from the omethoate application (F2,14 = 134.64, P < 0.001). Numbers were signiﬁcantly lower
The laboratory bioassays showed B. lapidaria has a tolerance to organophosphorous pesticides and abamectin that is signiﬁcantly higher than both S. viridis and H. destructor. Furthermore, a strong susceptibility by B. lapidaria to the synthetic pyrethroids was coupled with high tolerance by S. viridis. These ﬁndings suggest that broad spectrum synthetic pyrethroids may be harmful to this predatory mite in the ﬁeld and could exacerbate S. viridis numbers, whereas abamectin and organophosphorous pesticides may be less likely to disrupt the biological control of B. lapidaria. However, the ﬁeld experiment showed the bifenthrin treatment did not signiﬁcantly reduce B. lapidaria numbers. The disparity between the laboratory and ﬁeld data may reﬂect a reduction in pesticide contact rather than chemical tolerance. Reduced contact could be due to
J.M.K. Roberts et al. / Biological Control 58 (2011) 222–229
Table 1 LD50 estimates with 95% conﬁdence intervals and regression coefﬁcients for Sminthurus viridis, Halotydeus destructor and Bdellodes lapidaria, to ﬁve pesticides. CI = conﬁdence intervals, sem = standard error of the mean. Pesticide
S. viridis H. destructor B. lapidaria S. viridis H. destructor B. lapidaria S. viridis H. destructor B. lapidaria S. viridis H. destructor B. lapidaria S. viridis H. destructor B. lapidaria
95% CI (mg/L)
14.600 0.030 0.106 2.320 0.008 0.099 5.560 0.502 46.500 5.880 11.000 28.000 1.580 5.550 39.800
9.180 0.016 0.061 1.370 0.000 0.018 3.260 0.351 33.000 5.210 9.240 20.000 0.672 4.030 19.700
23.400 0.047 0.172 3.850 0.015 0.164 8.970 0.718 66.200 6.580 13.600 37.000 2.500 7.590 67.100
Regression coefﬁcient ± sem
138 (B. lapidaria); 486 (H. destructor)
1.139 ± 0.207 1.309 ± 0.305 1.021 ± 0.173 1.272 ± 0.272 1.407 ± 0.584 1.578 ± 0.597 1.354 ± 0.293 2.802 ± 0.710 1.743 ± 0.313 8.726 ± 2.204 3.779 ± 1.070 2.040 ± 0.376 1.474 ± 0.299 1.645 ± 0.359 1.177 ± 0.197
4 (H. destructor) 290 (B. lapidaria); 23 (H. destructor) 12 (H. destructor) 11 (H. destructor) 8 (S. viridis); 93 (H. destructor) 2 (S. viridis) 5 (S. viridis); 3 (H. destructor) 4 (S. viridis) 25 (S. viridis); 7 (H. destructor)
Mean B. lapidaria/m2
Mean S. viridis/m2
a 400 b 300
Fig. 3. Field toxicity of omethoate and bifenthrin on (a) Bdellodes lapidaria and (b) Sminthurus viridis over four post-treatment sampling dates. The pre-treatment sample (10September) was included as a covariate in the analyses. Different letters above bars denote signiﬁcantly different means between treatments (at the P < 0.05 level, Tukey’s-b post hoc tests).
vegetation height, which was approximately 10–15 cm higher than closely grazed pasture or perhaps due to the high mobility of B. lapidaria and its tendency to shelter under bark and debris (Wallace, 1971). Behavioural avoidance of pesticides has been demonstrated in several other predatory species (Amorim et al., 2005; Pekar and Haddad, 2005). Nevertheless, the ﬁeld results were still largely consistent with the laboratory data. For instance, the omethoate treatment reduced S. viridis to minimal levels but had no effect on B. lapidaria numbers. The systemic properties of omethoate may have helped to minimise the impact on B. lapidaria. The bifenthrin treatment also
had no effect on S. viridis, as predicted by the laboratory bioassays. This ﬁnding supports previous laboratory bioassays that have found a higher tolerance in S. viridis to synthetic pyrethroids (Roberts et al., 2009) and helps explain outbreaks of this pest following chemical control aimed at earth mites (Hoffmann et al., 2008). Michael (1991) also assessed the toxicity of several pesticides against B. lapidaria in the ﬁeld. Consistent with the current study, omethoate and the synthetic pyrethroid, fenvalerate, did not negatively affect B. lapidaria. However, Michael (1991) did see a reduction in numbers from chlorpyrifos, which was more toxic than omethoate in the laboratory bioassays. From the limited data
0.6 log10 R O (B. lapidaria )
log10 Nb (S. viridis)
J.M.K. Roberts et al. / Biological Control 58 (2011) 222–229
2.5 2.3 2.1 1.9 1.7 1.5 0.0
1.0 1.5 log10 Na (B. lapidaria)
Fig. 4. Linear regression for ﬁnal numbers (Nb) of Sminthurus viridis versus the initial numbers (Na) of Bdellodes lapidaria; R2 = 0.13, b = 0.000067 ± 0.14, t = 0.00049, P = 0.99. Na and Nb were log10 transformed.
0.8 0.6 log10 RO (B. lapidaria)
0.4 0.2 0.0 -0.2 1.5
-0.4 -0.6 -0.8 -1.0 -1.2
log10 Na (S. viridis)
Fig. 5. Linear regression for reproductive output (RO) of Bdellodes lapidaria versus the initial numbers (Na) of Sminthurus viridis; R2 = 0.72, b = 0.44 ± 0.18, t = 2.45, P < 0.05. RO and Na were log10 transformed.
log10 RO (S. viridis)
0.8 0.6 0.4 0.2 0.0 -0.2
-0.4 log10 Na (S. viridis) Fig. 6. Linear regression for reproductive output (RO) of Sminthurus viridis versus the initial numbers (Na) of Sminthurus viridis; R2 = 0.35, b = 0.66 ± 0.21, t = 3.19, P < 0.01. RO and Na were log10 transformed.
available, broad spectrum pesticides may not be strongly impacting on B. lapidaria in the ﬁeld. Despite this, further development of selective pesticides to minimise non-target effects should remain a priority. Abamectin and omethoate were least toxic to B. lapidaria in the laboratory and also have systemic or translaminar properties that can further reduce their non-target impact. Furthermore, the potential sub-lethal effects from broad-spectrum pesticides are not detected in standard mortality assays and require further investigation (Bernard et al., 2004).
0.4 0.2 0.0 -0.2 0.0
-0.4 -0.6 -0.8 -1.0 -1.2
log10 N a (B. lapidaria )
Fig. 7. Linear regression for reproductive output (RO) of Bdellodes lapidaria versus the initial numbers (Na) of Bdellodes lapidaria; R2 = 0.72, b = 0.72 ± 0.12, t = 5.99, P < 0.001. RO and Na were log10 transformed.
We found no evidence of a predator–prey relationship between B. lapidaria and S. viridis. In untreated ﬁeld plots, there was no association between ﬁnal numbers or the RO of S. viridis and numbers of B. lapidaria. Levels of S. viridis were also independent of B. lapidaria in the pesticide treated plots. Although not directly established to inform on predatory effects, S. viridis numbers recovered quickly in the omethoate treated plots despite the presence of B. lapidaria. This result supports those of Swan (1940) and Ireson (1984) who were also unable to detect any signiﬁcant predatory impact on S. viridis in south-eastern Australia. In contrast, Wallace (1967), who described S. viridis as the principal host of B. lapidaria, found that in Western Australia the presence of these predatory mites in winter was sufﬁcient to suppress S. viridis numbers in spring. Interestingly, in the last post-treatment sample, S. viridis numbers were higher for the bifenthrin treatment than the controls. However unlike previous studies, where S. viridis has often increased after treatment with synthetic pyrethroids (Bishop et al., 1998; Michael, 1991), this difference was due to decreased pest numbers in the control. As S. viridis was naturally dying off with the onset of summer, variability in numbers is more likely at this last posttreatment sample. Previous observations in the laboratory and ﬁeld also indicate B. lapidaria does actively feed on S. viridis (Michael et al., 1991; Wallace, 1981). It is possible that B. lapidaria has a preference for alternative prey, thereby reducing its effectiveness in pest control (Symondson et al., 2006; Symondson et al., 2002). B. lapidaria feeds on a range of collembolan species. Ireson (1984) identiﬁed several species as alternative prey for B. lapidaria including Entomobrya multifasciata, which was present in our ﬁeld plots. Wallace and Walters (1974) suggested that S. viridis is the preferred prey species, however this was based primarily on the high abundance of S. viridis relative to other species. Ireson (1984) found no evidence for preference of S. viridis. Considering the relatively large size of S. viridis, prey body size and its effect on predation success are more likely inﬂuencing prey preference in B. lapidaria. Wallace (1967) noted that predation of S. viridis was greatest when a higher proportion of small nymphs were present. Preference for smaller prey has also been found for other arthropod predators such as the assassin bug, Zelus longipes (Cogni et al., 2002), and the carabid beetle, Poecilius cupreus (Lang and Gsödl, 2001). Despite the lack of an observable predator–prey interaction between these species, S. viridis negatively impacted the reproductive output of B. lapidaria. This ﬁnding seems counter intuitive. Perhaps indirect effects from competitive interactions with other species are adding to the complexity of this system. Barker (2006) found competition occurring between S. viridis and the collembolan Bourletiella hortensis reduced feeding damage and that the presence of
J.M.K. Roberts et al. / Biological Control 58 (2011) 222–229
multiple non-herbivorous collembolan species ameliorated the negative effects of S. viridis. Perhaps similar competitive interactions occurred in this study, with S. viridis displacing other collembolan species which were preferred prey of B. lapidaria. The present experiment was also conducted in spring, over a period when more diapause eggs of S. viridis are laid (Wallace, 1968). This may have resulted in fewer small nymphs being available as the experiment progressed and may have led to an overall reduction of prey resources for B. lapidaria. Signiﬁcant intra-speciﬁc competition was detected in S. viridis and B. lapidaria. Based on the partial regression coefﬁcients for reproductive output regressed against initial numbers, both species appear to display contest competition (Nicholson, 1954). Under this competition model, a proportion of individuals, typically through behavioural interference, obtain a sufﬁcient share of resources needed for reproduction and survival to the detriment of other individuals. In predatory mites, aggressiveness, cannibalism and conspeciﬁc egg consumption are common competitive behaviours (Schausberger, 2003). For S. viridis, competition may occur for preferred plant hosts and for oviposition sites. Walters (1968) found reduced egg-laying by S. viridis with increasing densities and suggested this was due to greater disturbance during oviposition. Wallace (1967) also found a density-dependent effect in S. viridis related to the behaviour of nymphs feeding on dead adults, except this resembled scramble-like competition. Newly hatched nymphs feeding on dead adults are thought to ingest a stored toxic substance, such as uric acid, and at high densities this phenomenon is suggested to cause dramatic population reductions. Contest competition could also occur through ﬁtness differences among individuals and has been suggested in the co-occurring pest mite, Penthaleus major (Weeks and Hoffmann, 2000). What are the implications of these ﬁndings for biological control of S. viridis in south-eastern Australia? While B. lapidaria can predate on S. viridis, it appears unable to exert a signiﬁcant predatory impact. Although pesticides may in some instances affect B. lapidaria, the presence of alternative prey is most likely reducing the impact of predation on S. viridis. In situations where there is less alternative prey or a greater proportion of S. viridis nymphs available, B. lapidaria may become more effective. Further research into the effects of alternative prey and the effectiveness of B. lapidaria earlier in the season are still required before determining the usefulness of this predatory mite in the biological control of S. viridis. However, it is likely that a more effective approach is to employ a suite of predators rather than rely on one predatory mite. There are numerous generalist predators in agro-ecosystems and many have been recorded feeding on S. viridis (Eden et al., 2005; Holdaway, 1927; MacLagan, 1932). By promoting predatory species generally, effective biological control may still be achieved. For instance, predator diversity and abundance can be greatly increased through the establishment of more diverse habitat within and around crops (Nash et al., 2008; Tsitsilas et al., 2006; Tsitsilas et al., 2011). Reductions in the use of broad spectrum pesticides can also help predator populations become and remain established (Holloway et al., 2008). Conversely, greater predator diversity may reduce pest suppression through increased intraguild predation (Rosenheim et al., 1995), although there is much evidence that predator diversity and biological control can coexist (Straub et al., 2008). In addition to improving agro-ecosystems for the promotion of predators, there is still one biological control agent yet to be fully utilised in south-eastern Australia. N. capillatus has proven highly effective at suppressing S. viridis populations in both Western Australia and Tasmania (Ireson and Webb, 1995; Wallace, 1974). Field experiments and laboratory feeding trials have both demonstrated a high predatory efﬁcacy against S. viridis and preference for this pest (Ireson et al., 2002; Michael et al., 1995). Aside from a limited
introduction to south-eastern Victoria in the early 1990’s, there has been no concerted effort to establish this predatory mite in the region. Introduction of N. capillatus to mainland south-eastern Australia could dramatically improve biological control efforts, although there are concerns with its potential to disperse beyond areas of introduction (Michael et al., 1995). In any case, it is necessary to consider the climate suitability and possible disruption to existing trophic links in potential introduction areas. In summary, this study has shown that B. lapidaria has low susceptibility to pesticides relative to pests in the laboratory and ﬁeld, apparently reﬂecting both inherent tolerance and avoidance. There was also no evidence of a detectable predator–prey relationship between B. lapidaria and S. viridis, questioning the usefulness of this predatory mite for biological control in south-eastern Australia. Acknowledgments We thank Dr. Michael Nash and Angelos Tsitsilas for assistance with erecting ﬁeld plots. Two anonymous reviewers provided valuable critique of the manuscript. We also thank John Martin for the use of the ﬁeld site. This study was supported by the Grains Research and Development Corporation, the National Invertebrate Pest Initiative and the Australian Research Council through their Special Research Centre program. References Amorim, M.J.B., Rombke, J., Soares, A., 2005. Avoidance behaviour of Enchytraeus albidus: effects of Benomyl, Carbendazim, phenmedipham and different soil types. Chemosphere 59, 501–510. Atyeo, W.T., 1963. The Bdellidae (Acarina) of the Australian realm. Part II. Australia and Tasmania. Bulletin University Nebraska State Museum 4, 167–210. Barker, G.M., 2006. Diversity in plants and other Collembola ameliorate impacts of Sminthurus viridis on plant community structure. Acta Oecologica-International Journal Of Ecology 29, 256–265. Bernard, M.B., Horne, P.A., Hoffmann, A.A., 2004. Developing an ecotoxicological testing standard for predatory mites in Australia: acute and sublethal effects of fungicides on Euseius victoriensis and Galendromus occidentalis (Acarina: Phytoseiidae). Journal of Economic Entomology 97, 891–899. Bishop, A.L., Harris, A.M., McKenzie, H.J., 2001. Distribution and ecology of the lucerne ﬂea, Sminthurus viridis (L.) (Collembola: Sminthuridae), in irrigated lucerne in the Hunter dairying region of New South Wales. Australian Journal of Entomology 40, 49–55. Bishop, A.L., McKenzie, H.J., Barchia, I.M., Spohr, L.J., 1998. Efﬁcacy of insecticides against the lucerne ﬂea, Sminthurus viridis (L.) (Collembola: Sminthuridae), and other arthropods in lucerne. Australian Journal of Entomology 37, 40–48. Cogni, R., Freitas, A.V.L., Amaral, B.F., 2002. Inﬂuence of prey size on predation success by Zelus longipes L. (Het, Reduviidae). Journal of Applied Entomology 126, 74–78. Currie, G.A., 1934. The Bdellid mite Biscirus lapidarius Kramer, predatory on the lucerne ﬂea (Sminthurus viridis L.) in Western Australia. Journal of the Council for Scientiﬁc and Industrial Research Australia 7, 9–20. Eden, T.M., Wilson, D.J., Hackell, D.L., 2005. Assays to determine the predatory ability of Pergamasus against clover ﬂea. New Zealand Plant Protection 58, 131– 134. Hoffmann, A.A., Porter, S., Kovacs, I., 1997. The response of the major crop and pasture pest, the red-legged earth mite (Halotydeus destructor) to pesticides: Dose–response curves and evidence for tolerance. Experimental and Applied Acarology 21, 151–162. Hoffmann, A.A., Weeks, A.R., Nash, M.A., Mangano, P.G., Umina, P.A., 2008. The changing status of invertebrate pests and the future of pest management in the Australian grains industry. Australian Journal of Experimental Agriculture 48, 1481–1493. Holdaway, F.G., 1927. The bionomics of Smynthurus viridis Linn., or the South Australian lucerne ﬂea. Coun. Scient. Indust. Res. Aust. (Pamphlet no. 4). Holloway, J.C., Furlong, M.J., Bowden, P.I., 2008. Management of beneﬁcial invertebrates and their potential role in integrated pest management for Australian grain systems. Australian Journal of Experimental Agriculture 48, 1531–1542. Ireson, J.E., 1984. The effectiveness of Bdellodes lapidaria (Kramer) (Acari, Bdellidae) as a predator of Sminthurus viridis (L) (Collembola, Sminthuridae) in northwest Tasmania. Journal of the Australian Entomological Society 23, 185–191. Ireson, J.E., Holloway, R.J., Chatterton, W.S., McCorkell, B.E., 2002. Further investigations into the efﬁcacy of Neomolgus capillatus (Kramer) (Acarina: Bdellidae) as a predator of Sminthurus viridis (L.) (Collembola: Sminthuridae) in Tasmania. Australian Journal of Entomology 41, 88–93.
J.M.K. Roberts et al. / Biological Control 58 (2011) 222–229 Ireson, J.E., Webb, W.R., 1995. Effectiveness of Neomolgus capillatus (Kramer) (Acarina, Bdellidae) as a predator of Sminthurus viridis (L) (Collembola, Sminthuridae) in northwestern Tasmania. Journal of the Australian Entomological Society 34, 237–240. Lang, A., Gsödl, S., 2001. Prey vulnerability and active predator choice as determinants of prey selection: a carabid beetle and its aphid prey. Journal of Applied Entomology 125, 53–61. MacLagan, D., 1932. An ecological study of the ‘lucerne ﬂea’ (Smynthurus viridis Linn.). Bulletin of Entomological Research 23, 101–145. Michael, P.J., 1991. Season-long effects of four chemicals on redlegged earth mite and lucerne ﬂea. Proceedings of a National workshop on redlegged earth mite, lucerne ﬂea and blue oat mite, pp. 63–65. Michael, P.J., Dutch, M.E., Pekin, C.J., 1991. A review of the predators of redlegged earth mite, blue oat and lucerne ﬂea. Plant Protection Quarterly 6, 178–180. Michael, P.J., McDonald, G., Hoffmann, A.A., 1995. Biological control of redlegged earth mite and lucerne ﬂea by the predators Anystis wallacei and Neomologus capillatus. Plant Protection Quarterly 10, 55–57. Nash, M.A., Thomson, L.J., Hoffmann, A.A., 2008. Effect of remnant vegetation, pesticides, and farm management on abundance of the beneﬁcial predator Notonomus gravis (Chaudoir) (Coleoptera: Carabidae). Biological Control 46, 83– 93. Nicholson, A.J., 1954. An outline of the dynamics of animal populations. Australian Journal of Zoology 2, 9–65. Pekar, S., Haddad, C.R., 2005. Can agrobiont spiders (Araneae) avoid a surface with pesticide residues? Pest Management Science 61, 1179–1185. Pescott, R.T.M., 1937. Insect pests of subterranean clover. Journal of the Department of Agriculture Victoria 35, 371. Roberts, J.M.K., Umina, P.A., Hoffmann, A.A., Weeks, A.R., 2009. The tolerance of the lucerne ﬂea, Sminthurus viridis (Collembola: Sminthuridae), to currently registered pesticides in Australia. Australian Journal of Entomology 48, 241– 246. Robertson, J., Preisler, H., 1992. Pesticide bioassays with arthropod. CRC Press, Boca Raton, Florida. Rosenheim, J.A., Kaya, H.K., Ehler, L.E., Marois, J.J., Jaffee, B.A., 1995. Intraguild predation among biological control agents – Theory and evidence. Biological Control 5, 303–335. Schausberger, P., 2003. Cannibalism among phytoseiid mites: a review. Experimental and Applied Acarology 29, 173–191. Sokal, R.R., Rohlf, F.J., 1995. Biometry: The Principles and Practice of Statistics in Biological Research, third ed. W.H. Freeman, New York. Stearns, S.C., 1992. The evolution of life histories. Oxford University Press, New York. Straub, C.S., Finke, D.L., Snyder, W.E., 2008. Are the conservation of natural enemy biodiversity and biological control compatible goals? Biological Control 45, 225–237. Swan, D.C., 1940. The lucerne ﬂea: its life history and control in South Australia. Journal of the Department of Agriculture South Australia 43, 462–471. Symondson, W.O.C., Cesarini, S., Dodd, P.W., Harper, G.L., Bruford, M.W., Glen, D.M., Wiltshire, C.W., Harwood, J.D., 2006. Biodiversity vs. biocontrol: positive and
negative effects of alternative prey on control of slugs by carabid beetles. Bulletin of Entomological Research 96, 637–645. Symondson, W.O.C., Sunderland, K.D., Greenstone, M.H., 2002. Can generalist predators be effective biocontrol agents? Annual Review of Entomology 47, 561–594. Tsitsilas, A., Hoffmann, A.A., Weeks, A.R., Umina, P.A., 2011. Impact of groundcover manipulations within windbreaks on mite pests and their natural enemies. Australian Journal of Entomology 50, 37–47. Tsitsilas, A., Stuckey, S., Hoffmann, A.A., Weeks, A.R., Thomson, L.J., 2006. Shelterbelts in agricultural landscapes suppress invertebrate pests. Australian Journal of Experimental Agriculture 46, 1379–1388. Umina, P.A., Hoffmann, A.A., 2005. Competitive interactions among four pest species of earth mites (Acari: Penthaleidae). Journal of Economic Entomology 98, 307–316. Wallace, M.M.H., 1967. The ecology of Sminthurus viridis (L.) (Collembola) I. Processes inﬂuencing numbers in pastures in Western Australia. Australian Journal of Zoology 15, 1173–1206. Wallace, M.M.H., 1968. The ecology of Sminthurus viridis (Collembola) II. Diapause in the aestivating egg. Australian Journal of Zoology 16, 871–883. Wallace, M.M.H., 1971. The inﬂuence of temperature and moisture on diapause development in the eggs of Bdellodes lapidaria (Acari: Bdellidae). Journal of the Australian Entomological Society 10, 276–280. Wallace, M.M.H., 1974. An attempt to extend the biological control of Sminthurus viridis (Collembola) to new areas in Australia by introducing a predatory mite, Neomolgus capillatus (Bdellidae). Australian Journal of Zoology 22, 519–529. Wallace, M.M.H., 1981. Tackling the lucerne ﬂea and red-legged earth mite. Journal of Agriculture Western Australia 22, 72–74. Wallace, M.M.H., Mahon, J.A., 1971. The ecology of Sminthurus viridis (Collembola) III. The inﬂuence of climate and land use on its distribution and that of an important predator, Bdellodes lapidaria (Acari: Bdellidae). Australian Journal of Zoology 19, 177–188. Wallace, M.M.H., Walters, M.C., 1974. The introduction of Bdellodes lapidaria (Acari: Bdellidae) from Australia into South Africa for the biological control of Sminthurus viridis (Collembola). Australian Journal of Zoology 22, 505–517. Walters, M.C., 1966. The yellow dung-ﬂy Scaiophaga stercoraria (L.) (Diptera: Cordyhnidae), as a predator of the lucerne springtail, Sminthurus viridis (L.). South African Journal of Agricultural Science 9, 739–740. Walters, M.C., 1968. A study of Sminthurus viridis (L.) (Collembola) in the Western Cape Province. Entomology Memoirs. Department of Agriculture and Forestry South Africa, 4–99. Weeks, A.R., Hoffmann, A.A., 1998. Intense selection of mite clones in a heterogeneous environment. Evolution 52, 1325–1333. Weeks, A.R., Hoffmann, A.A., 2000. Competitive interactions between two pest species of earth mites, Halotydeus destructor and Penthaleus major (Acarina: Penthaleidae). Journal of Economic Entomology 93, 1183–1191. Womersley, H., 1933. A possible biological control of the clover springtail or lucerne ﬂea (Sminthurus viridis L.) in Western Australia. Journal of the Council for Scientiﬁc and Industrial Research Australia 6, 83.