This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright
Author's personal copy
Biological Control 48 (2009) 115–124
Contents lists available at ScienceDirect
Biological Control journal homepage: www.elsevier.com/locate/ybcon
Prey availability, pesticides and the abundance of orchard spider communities Viktor Markó a,*, Balázs Keresztes b, Michelle T. Fountain c, Jerry V. Cross c a b c
Department of Entomology, Corvinus University of Budapest, P.O. Box 53, H-1518 Budapest, Hungary Institute for Plant Protection, Georgikon Faculty for Agriculture, University of Pannonia, H-8360 Keszthel, Hungary East Malling Research, New Road, East Malling, Kent ME19 6BJ, UK
a r t i c l e
i n f o
Article history: Received 13 December 2007 Accepted 8 October 2008 Available online 17 October 2008 Keywords: Araneae Apple orchard Immigration Prey Integrated pest management (IPM)
a b s t r a c t In a 4 year study, in southern England, the abundance of apple orchard canopy spiders and their potential prey was manipulated using two pest management strategies based on broad spectrum (highly toxic both to spiders and pests) and selective (moderately toxic to spiders but highly toxic to pests) insecticides in the first part of the growing season. The spider community was left to develop freely afterwards. Apple orchard plots untreated by pesticides served as control. The effect of insecticides was detrimental to spider populations as the treatments coincided with the peak abundance of adults in May and early June. Within adults, the treatments were harmful to female spiders, whereas, male spiders were much less affected. As a result the proportion of males increased in all of the sampled spider families. The use of selective insecticides resulted in a higher spider abundance compared to the use of broad spectrum compounds while the highest spider abundance was found in the pesticide free trees, i.e. three significantly different spider abundance levels were produced in spring. Spider abundance began to increase unequally between the treatments afterwards and became identical in the two pesticide treated plots due to the immigration of juveniles from surrounding habitats. However, a similar equalisation of abundance was not observed between the pesticide treated plots and untreated control. Analysing the abundance pattern of potential prey in the plots of the studied orchard we concluded that the post-disturbance increase in spider abundance was regulated by prey availability. Ó 2008 Elsevier Inc. All rights reserved.
1. Introduction Spiders are one of the most species rich and abundant predacious arthropod groups in agroecosystems, including the ground flora and canopy of orchards (Olszak et al., 1992a; Bogya et al., 1999a; Pekár, 1999; Miliczky et al., 2000; Fountain et al., 2007). Although there are only scarce data on their influence on invertebrate orchard pests (Mansour et al., 1980; Wyss et al., 1995), laboratory feeding tests (Wisniewska and Prokopy, 1997a; Miliczky and Calkins, 2002) and their abundance in orchards suggests that they may have a significant impact on many pest species. The aim of biological and integrated pest control in orchards is to shift the balance of arthropod assemblages to beneficial arthropods, including spiders. Many components of agroecosystems and elements of agricultural practices can be managed to achieve this goal (Marc et al., 1999). One of the most important elements is the reduction of chemical disturbance to spiders. The use of broad spectrum pesticides has a strong negative effect on spider populations which may be more sensitive than the orchard pests (Mansour et al., 1980; Bostanian et al., 1984). Selective insecticides * Corresponding author. Fax: +36 1 4826072. E-mail address:
[email protected] (V. Markó). 1049-9644/$ - see front matter Ó 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.biocontrol.2008.10.002
used in organic and integrated pest management (IPM) programmes have a negligible or moderate effect on spiders (Amalin et al., 2000; Pekár, 2002; Pekár and Kocourek, 2004). Therefore, the use of selective insecticides is one of the main tools in enhancing spider communities in commercial orchards. In organic and IPM orchards the abundance and species richness of spiders is usually higher than in conventionally pesticide treated orchards, ordinarily based on applications of broad spectrum pesticides (Olszak et al., 1992a; Bogya et al., 2000; Miliczky et al., 2000; Cárdenas et al., 2006). There is evidence that spiders re-colonise orchards after the application of pesticides has ceased. However, Bogya et al. (2000) and Miliczky and Horton (2005) demonstrated that spiders abundance declined with increasing distance from the margin to the centre of the orchard, both in small and large commercial apple and pear orchards, in the spring and autumn. This implied that some abundant orchard spider species had a relatively low immigration rate and that re-colonisation of pesticide treated orchards could not be fully achieved until the end of the season. Pesticide applications affect spiders directly, but also indirectly by depleting their prey. Whilst there is some documentation on the effect of prey shortage on spiders in agricultural fields and orchards (Bogya et al., 2000; Harwood et al., 2001), many studies deal
Author's personal copy
116
V. Markó et al. / Biological Control 48 (2009) 115–124
with the effect of prey abundances on spider aggregation in microhabitats in natural ecosystems (Wise, 1993). Better food supply increases the patch residence time, whilst food deprivation results in a higher probability of dispersal behaviour (Wise, 1993; Weyman et al., 2002). Additionally, an increase of spider fecundity was shown in plots with more prey (Wise, 1993; Marc et al., 1999). In contrast, the cost of relocation when moving to a new site can be too high for some spiders compared to the probability of finding a new site with better food availability. Therefore, some spiders have lower motivation to abandon a habitat in response to a period of poor prey capture (Gillespie and Caraco, 1987). Energy cost of web construction can also determine the frequency of web-relocation especially for sheet- and funnel-weavers (Tanaka, 1989) and emigration can be costly and risky process associated with higher predation (Weyman et al., 2002). Generally, spiders can survive long periods in starvation and species may respond differently to changes in prey abundance (Wise, 1993). Species susceptibility to pesticides and immigration and emigration rates may affect both the abundance and community composition of spider assemblages in orchards with different levels of pesticide use. The aims of this study were to: (i) determine whether a zero pesticide residue (ZERO) strategy (selective pesticides only moderately harmful to natural enemies are applied before the fruit begins to develop), was less harmful to spider populations than a conventional (CONV) pest management system (based on broad spectrum insecticides) in apple orchards, (ii) test whether orchard spider assemblages in the latter part of the fruit growing season benefited from higher initial abundances in the spring, (iii) establish if prey density affected the post-pesticide disturbance recovery of spider communities. We hypothesised that the new IPM strategy (ZERO) would have a similar affect on pest populations in orchards as the conventional pest management strategy (CONV) compared to the untreated control plots (UNTR), but the ZERO treatment would have a less negative impact on spiders compared to the CONV treatment in the first part of the vegetation growing season, spring (Table 1). Three scenarios were hypothesized for the post-disturbance recovery of spider communities in the second part of the growing season, autumn, after finishing the use of pesticides highly (CONV) or moderately (ZERO) harmful to spiders. The first scenario (hypothesis A, Table 1) is characterised by a high spider immigration rate in the insecticide treated plots which is independent of prey availability. In the second scenario (hypothesis B), the ZERO treatment has a long lasting positive effect on the orchard spider communities. The use of selective insecticides (ZERO) with a lower toxicity on spiders maintains a more abundant spider assemblage in spring and as a consequence more offspring are produced in the latter part of the growing season. The role of the prey supply is limited (Table 1). Table 1 The three predicted affects of the UNTR, untreated; ZERO, zero pesticide residue treatments; CONV, conventional treatments, on abundance of spider populations and their prey at different stages in the fruit growth season. Prey abundancea
Spider abundancea Spring
CONV ZERO UNTR a
+ + +++
+ ++ +++
Autumn Hypothesis A
Hypothesis B
Hypothesis C
+++ +++ +++
+ ++ +++
++ ++ +++
+, low abundance; ++, medium abundance; +++, high abundance.
In the third scenario (hypothesis C), the basis of the post-pesticide application recovery would be immigration and/or within orchard population growth, with both being regulated by prey availability. This hypothesis suggests a low spider population increase in the ZERO treatment, where the spider/prey ratio is higher. A higher population increase of spiders would occur in the CONV treatments, where the spider/prey ratio would be lower. As the prey supply would be higher in the UNTR plots spider abundance should be also higher (Table 1). Because of species specific pesticide sensitivity and migration, we also predicted differences in spider community composition between the three treatments.
2. Methods The experimental orchard was at East Malling Research, Kent, England and the studies were made in 2001, 2002, 2004 and 2006. The orchard was 1.14 ha and planted in spring 1995. The orchard had 12 plots which consisted of 12 rows of 12 dwarf (M9 rootstock) apple trees (rows were spaced 4 m apart and trees 1.75 m apart). The plots were separated by headlands and alder (Alnus cordata (Loisel.)) windbreaks. The orchard contained plots of disease susceptible and disease resistant apple varieties, but every plot contained the variety Discovery on which the sampling was conducted. A normal commercial weed-free strip was maintained by herbicide spraying in the tree rows and the grass alleys were mown regularly. The orchard was surrounded by arable fields, other fruit orchards, including pear and cherry, hedges and small areas of ruderal and woodland. In the 12 orchard plots three different pest management systems were applied in a randomised block design (four replicate plots) from 2001 to 2006. The treatments (Appendix A) were: 1. zero pesticide residues (ZERO); integrated pest management programme, where pesticides only moderately harmful to spiders were used and only in the early stage of the growing season. During the fruit development period the pest and disease management relied on biocontrol agents, 2. conventional (CONV); full pesticide treatment, where broad spectrum insecticides were applied, 3. untreated control (UNTR); no pesticides applied. Of the insecticides used pirimicarb, fenoxicarb, Bacillus thuringiensis and codling moth granulosis virus are harmless to spiders (Powell et al., 1985; Mansour, 1987; Pekár, 2002; Nicholas et al., 1999; Amalin et al., 2000), whilst diflubenzuron, in laboratory tests, had a low to moderate toxicity (Hassan et al., 1994; Amalin et al., 2000). Methoxyfenozide and thiacloprid have a moderate toxicity to many beneficial arthropods (e.g. Schmuck, 2001; Villanueva and Walgenbach, 2005). In contrast, chlorpyrifos is highly toxic to spiders (Mansour, 1987; Amalin et al., 2000; Fountain et al., 2007). In the ZERO plots insecticides with moderate effect on spiders were used up to the end of spring (2nd June). In the CONV treatment, chlorpyrifos, highly toxic to spiders, was also used and applications were applied further, if that was necessary, up to the first part of August (5th August). After these times, only substances harmless to spiders were applied to the crop and spider populations were free to develop (Appendix A). Apart from the second part of the growing season (August and September) when pesticides harmless to spiders were applied, CONV treated plots received more applications of more spider toxic insecticides for a longer time period (Appendix A). Beat samples from the canopy were taken every two weeks with a beating funnel (70 43 cm size, 50 cm depth) from early May to early October in 2001, 2002 and 2004. Two beat samples
Author's personal copy
117
V. Markó et al. / Biological Control 48 (2009) 115–124
in May and two in September were collected (and analysed) in 2006. In the summer of 2006 only a short visual inspection was made on some trees (not analysed). The whole canopies of twelve trees were sampled on every occasion from each plot. The samples were summarized within plots and months or years (total spider abundance). To avoid the saturation of species richness at higher sampling effort it was calculated from the subsamples (4 trees per plot) cumulated yearly. For comparison of the decreasing effect of the ZERO and CONV treatments on the abundance of adult females and males and juvenile spiders within year total abundances in the CONV and ZERO plots were divided by the total abundances in the UNTR (control) plots and after adjusted rank Welch test, pairwise stochastic equalities were tested with Bonferroni correction. Because the abundance of females in some of the UNTR plots, which had been treated with broad spectrum insecticides in the year before the experiment began, was low the first year’s data of females and males were excluded. The potential prey groups (Auchenorrhyncha, Psylla spp., Neuroptera, Heteroptera, Coleoptera, Lepidoptera and parasitic Hymenoptera adults) were collected by yellow sticky traps (two, twosided traps per plot in the middle of the canopy of apple cultivar Discovery) in 2001 and 2002. The same potential prey groups and aphids from the beat samples described above were counted in 2004 and 2006. For the comparison of the abundances of potential prey groups in the latter part of the fruit growing season (August, September and first week of October) the samples were summarized within plots. Additional results on the leafhoppers from the same study were presented by Bleicher et al. (2003) and on the aphids and other pest species by Cross and Berrie (2005). All of the spider and potential prey data were analysed after ln(x + 1) transformation. The yearly data of total abundance, species richness (based on identification of adult individuals) and prey abundance were subjected to robust Welch ANOVA test and Games–Howell pairwise comparison of means was used to detect treatment differences. For spider species richness, abundance of the most common genera and linyphiids (in spring: May and June; and autumn: August, September and first week of October) and prey abundance in the latter part of the growing season, robust two-way ANOVA was also conducted to test the difference between treatments and years (Welch-test) and the interaction of these two factors (Johansen-test). Pairwise comparisons were tested using a Tukey–Kramer post hoc test. As the effect of years was significant in all comparisons, and usually there was no interaction between the years and treatments, only the results on effects of treatments are presented. Metric ordination, principal coordinates analysis (PCoA), based on the Horn similarity index (less sensitive to sample size, Krebs, 1999), was applied to compare the composition of spider assemblages of differently treated plots. For this comparison, samples from (i) the different years, (ii) from the first (May and June) and second (August, September and first week of October) part of the season and (iii) samples separated into adults and juveniles were used. As many juveniles can only be identified to genus level (except some Linyphiidae juveniles, identified to family level), the comparisons were based on the abundances of genera. Eigenvalues as percentage of variance explained by the axes are shown in the figures along the axes. Stepwise multiple regressions analysis was used to find the most important prey groups that predict the abundance of the most abundant spider genera and the whole spider community in autumn. Additionally, the relationship between the abundance of potential prey groups, total prey abundance and spider abundance was also evaluated separately by using Kendall’s tau correlation. As pesticide treatments in the first part of the growing season
could affect both the abundance of prey and spiders in autumn, the spider abundance in May was incorporated into the analysis to determine if this added to the prediction power of the model (multiple regression) or if removing its monotonic effect would affect the correlation between the prey and spider abundances (Kendall tau partial correlation). By dividing the abundance in the different plots by the total catch in the 12 plots the data from the first (2001, 2002) and last (2004, 2006) years of the study were transformed to a common scale and were analysed together. As the sample size was low, the data of the most common genera were not analysed separately in the years 2001 and 2002. RobStat statistical package (Vargha, 2007) was used for all statistical analyses except metric ordination, which was performed by Syntax 2000 (Podani, 2001). 3. Results 3.1. Total abundance and different effects on gender During the four year study a total of 5958 individuals comprising 51 species were collected in the canopy. Both the CONV and ZERO treatments significantly reduced the total arboreal spider abundance compared to the UNTR plots in all of the investigated years. The ZERO programme did not result in an increase in total spider abundance compared to CONV strategy even after six years of applications. Similar results were found for species richness measurements (Table 2). There were substantial differences of the effects of the pesticide regimes on the gender of spider and the juvenile numbers. Both in the CONV and ZERO treatments, the abundance of females and juveniles decreased significantly compared to the males (Table 3). This was especially obvious in the CONV plots where females were reduced by 65% while males by only 19% compared to the untreated control (Tests of pairwise stochastic equalities: Afemale/male = 1.000, BM(4, 0) = 999.000, p < 0.0001) (Table 3). As a consequence, the proportion of males in the arboreal spider assemblages increased from 32% in the UNTR plots to 47% and 46% in the CONV and ZERO plots, respectively (Table 3). The difference was
Table 2 Abundance (total abundance/12 trees) and species richness (number of species/4 trees) of spiders (±SD) in the orchard canopy of the orchard plots treated with broad spectrum insecticides (CONV), zero pesticide residue treatments (ZERO), untreated control (UNTR). CONV
ZERO
UNTR
Probability
Abundance 2001
76 (17.88) a
68 (14.97) a
108 (17.25) b
2002
41 (7.61) A
50 (8.17) A
75 (4.72) B
2004
201 (54.77) A
211 (48.37) A
399 (87.03) B
2006
146 (27.01) A
127 (12.40) A
253 (13.48) B
W(2; 5.9) = 6.182, p = 0.0355 W(2; 5.2) = 12.227, p = 0.0109 W(2; 5.9) = 9.957, p = 0.0127 W(2; 5.1) = 76.285, p = 0.0002
Species richness 2001 1.6 (0.66) a
2.4 (0.84) a
1.6 (1.15) a
2002
3.2 (0.88) a
3.0 (0.73) a
3.9 (3.90) a
2004
5.1 (0.70) a
4.8 (0.76) a
6.4 (0.66) b
2006
2.1 (0.97) a
3.1 (0.59) a
3.6 (0.99) a
2002–2006a
3.4 A
3.6 A
4.6 B
W(2; 5.9) = 1.213, p = 0.3621 W(2; 5.3) = 2.161, p = 0.2061 W(2; 5.8) = 5.507, p = 0.0457 W(2; 5.4) = 1.721, p = 0.2638 F(2; 12) = 6.323a, p = 0.0133
Means followed by different capitals or different lowercase letters within the row represent significant p < 0.05 or marginally significant p < 0.10 difference, respectively. a Two-way ANOVA (treatments versus years).
Author's personal copy
118
V. Markó et al. / Biological Control 48 (2009) 115–124
The abundance of juvenile and adult spiders was highest in the UNTR plots throughout the whole growing season (Fig. 1). However, the difference between the UNTR and pesticide treated (CONV and ZERO) plots was significant only in May, August and September (including first week of October) (Table 4). In addition, in May, the spider abundance was higher in the ZERO plots compared to CONV plots in 2001–2002, 2004 and 2006 by 30%, 105% and 33%, respectively (Fig. 1, Table 4). In the latter part of the growing season the abundance of juveniles increased rapidly, however, the increase was two to three times greater in the CONV plots compared to the ZERO plots (Table 4). As a consequence, there was no significant difference between CONV and ZERO treatments in August or September (Table 4). The difference in population increase is apparent if the four year total abundance of spiders in the CONV and ZERO plots is expressed as a proportion of the catch in the UNTR control plots. Thus, the abundance of spiders in the CONV plots was 44% of the UNTR plots in May and this value increased to 62% up to September. In contrast, in the ZERO plots, the relative abundance decreased from 67% in May to 52% in September (Table 4). The spider assemblages in all four years were dominated by Araniella (mainly A. opistographa (Kulczyn´ski) and fewer A. cucurbitina (Clerck), Theridion (T. varians Hahn with some T. pallens (Blackwall)), Neottiura bimaculata (Linnaeus), Philodromus (P. cespitum (Walckenaer) and some P. praedatus O.P. Cambridge) and Tetragnatha (mostly T. extensa (Linnaeus). These five genera comprised 82% of the total catch. A further 5% and 1.5% of the total catch were Linyphiidae (mainly Entelecara acuminata (Wider) and Lepthyphantes tenuis (Blackwall)) and Xysticus (mainly X. cristatus (Clerck) and some X. kochi Thorell), respectively. Spiders from the genera Araniella, Theridion and Philodromus showed a similar pattern of abundance to the whole spider assemblage in the spring and autumn (Fig. 2). The use of selective insecticides resulted in a greater number of specimens in the
Table 3 Proportion of males (%) of the main spider families [with the total catch of adults], and relative abundance (%, ±SD) of male, female and juvenile spiders compared to the untreated control (100%). CONV, treated with broad spectrum insecticides; ZERO, zero pesticide residue treatments and UNTR: untreated control. Families
CONV
ZERO
UNTR
Proportion of males (%) and [total number of individuals] Araneidae 54 [58] 49 [80] Linyphiidae 40 [43] 33 [36] Theridiidae 51 [53] 53 [70] Other families 36 [22] 41 [29] All spiders 47 [176] 46 [215] Only in May 56 [88] 54 [124]
41 20 37 25 32 35
Relative abundance (%) compared to UNTRa Male 81.4 (26.6) a Female 34.7 (7.3) b Juvenile 59.8 (8.6) a
100 100 100
a
83.7 (6.7) a 46.9 (12.1) b 61.8 (17.1) ab
[83] [100] [134] [20] [337] [218]
Different letters within a column represent significant differences (p < 0.0001).
apparent during the pesticide applications in May when the sex ratio shifted to a male bias in the pesticide treated plots (Table 3). Though the sample size was too low for detailed comparison, it seems that the increase of proportion of males in the pesticide treated plots is characteristic for all spider families investigated (Table 3). 3.2. The annual activity Spider abundance showed a similar pattern in all years and treatments. Numbers of spiders increased in May and the first week of June, and to a greater extent in September. 23% of the total catch was collected in May, whilst 53% were from September. The greatest number of adults was collected in May, whilst juveniles dominated the catch in September (Fig. 1).
500
Juveniles
Abundance
400
300
2006 2004 2001
200
2002
100
0 100
CONV ZERO UNTR
Adults
Abundance
80 60
2004
40
2001
20
2006
2002
04.09.
18.05.
10.09.
13.08.
15.07.
17.06.
20.05.
18.09.
22.08.
24.07.
28.06.
30.05.
01.05.
07.09.
08.08.
20.07.
19.06.
24.05.
0
Date Fig. 1. Total numbers of juvenile and adult spiders in the canopy of the apple trees (individuals/48 trees) over time. CONV, conventional treatments (broad spectrum insecticides); ZERO, zero pesticide residue treatments (selective insecticides); UNTR, untreated control. Dashed lines, estimation.
Author's personal copy
119
V. Markó et al. / Biological Control 48 (2009) 115–124
Table 4 Mean number of canopy spiders (individuals/12 trees ±SD) and their net increase (mean abundance in September minus mean abundance in May) during the fruit growing season. CONV, conventional treatments (broad spectrum insecticides); ZERO, zero pesticide residue treatments (selective insecticides); UNTR, untreated control.
2001+02 CONV ZERO UNTR
2004 CONV ZERO UNTR
2006 CONV ZERO UNTR
May
June
July
August
September
Increase
33 (7.1) A 42 (7.8) B 57 (7.0) C
6 (1.5) a 11 (3.7) a 11 (6.1) a
4 (1.7) a 4 (1.0) a 6 (0.8) b
12 (3.1) a 10 (5.1) a 19 (6.1) b
65 (19.2) A 52 (9.8) A 90 (8.4) B
32 10 33
W(2; 5.6) = 10.147, p = 0.0136
W(2; 5.5) = 2.436, p = 0.1744
W(2; 5.0) = 6.236, p = 0.0441
W(2; 5.7) = 4.011, p = 0.0816
W(2; 5.1) = 13.373, p = 0.0096
23 (9.5) A 49 (13.6) B 64 (20.1) B
9 (3.4) a 8 (4.1) a 15 (9.1) a
11 (1.0) a 6 (1.3) a 15 (7.6) a
43 (23.5) a 45 (12.9) ab 100 (32.1) b
115 (29.8) A 104 (30.5) A 205 (31.8) B
W(2; 5.9) = 7.685, p = 0.0229
W(2; 5.6) = 1.273, p = 0.3499
W(2; 4.7) = 10.910, p = 0.0168
W(2; 5.3) = 4.770, p = 0.0658
W(2; 6.0) = 11.243, p = 0.0094
30 (6.8) A 41 (7.6) B 74 (19.9) C
No regular sampling — — —
— — —
— — —
116 (28.0) A 86 (18.1) A 179 (8.1) B
W(2; 5.9) = 10.799, p = 0.0106 Meana CONV ZERO UNTR
92 55 141
96 45 105
W(2; 4.9) = 43.413, p = 0.0008
29 Ab 44 B 65 C
8a 9a 13 a
7a 5a 10 a
27 A 27 A 60 B
98 Ac 80 A 158 B
F(2; 17) = 28.726, P < 0.0001
F(2; 11) = 1.483, p = 0.2691
F(2; 10) = 3.334, p = 0.0777
F(2; 10) = 10.203, p = 0.0038
F(2; 15) = 16.422, p = 0.0002
69 36 93
Means followed by different capitals or lowercase letters within a column in a year represent a significant p < 0.05 and marginally significant p < 0.10 difference, respectively. a Two-way ANOVA (treatments & years). b Abundance in May: TCONV/ZERO(3; 30) = 5.14, p < 0.01; TCONV/UNTR(3; 27) = 10.80, p < 0.01; TZERO/UNTR(3; 29) = 5.51 p < 0.01. c Abundance in September: TCONV/ZERO(3; 29) = 0.44, p > 0.1; TCONV/UNTR(3; 25) = 5.97, p < 0.01; TZERO/UNTR(3; 27) = 7.24, p < 0.01.
35
35
25
A B
20
UNTR
A
B
C
15 10
A
A A
B
B
B
A
5
A
AB
CONV
30
CONV ZERO
A A
Mean abundance
30
Mean abundance
B
B
25 20 15
UNTR
A A
B
B
B
10 5
ZERO
A A
A A A
B AB
0
0
spring
autumn
Araniella
spring
autumn
Theridion
spring
autumn
Philodromus
spring
autumn
Linyphiidae
spring
autumn
Neottiura
spring
autumn
Tetragnatha
Fig. 2. Mean abundance of three spider genera in the canopy of the apple orchard plots in the spring (May and June) and the autumn (August, September and first week of October). Different letters represent significant differences p < 0.05. CONV, conventional treatments (broad spectrum insecticides); ZERO, zero pesticide residue treatments (selective insecticides); UNTR, untreated control.
Fig. 3. Mean abundance of the Linyphiidae, Neottiura bimaculata and Tetragnatha in the canopy of the apple orchard plots in spring (May and June) and autumn (August, September and first week of October). Different letters represent significant differences p < 0.05. CONV, conventional treatments (broad spectrum insecticides); ZERO, zero pesticide residue treatments (selective insecticides); UNTR, untreated control.
ZERO compared to the CONV plots in the spring (Araniella: TCONV/ZERO(3; 29) = 4.19, p < 0.05; Theridion: TCONV/ZERO(3; 30) = 3.82 p < 0.05; Philodromus: TCONV/ZERO(3; 29) = 1.55, p > 0.1). However, the abundance of spiders in these two treatments was fairly similar by the autumn (Araniella: TCONV/ZERO(3; 29) = 1.34, p > 0.1; p > 0.1; Philodromus: Theridion: TCONV/ZERO(3; 29) = 0.02, TCONV/ZERO(3; 22) = 0.46, p > 0.1) (Fig. 2). 94% of the Xysticus individuals were collected in the autumn and there was no difference in their abundance between treatments (total abundance/plot (±SD) CONV: 7.6 (1.7), ZERO: 8.0 (0.8), UNTR: 7.3 (2.6), W(2; 5.0)=0.29, p = 0.76). An unexpected pattern of abundance was detected in three of the spider groupings: Linyphiidae, Neottiura bimaculata and Tetragnatha. Linyphiids compensated for the higher distur-
bance to spiders in the CONV plots in the spring (Fig. 3). A similar pattern was found in the spring for the adults of the most abundant Linyphiid species, E. acuminata (96% of the E. acuminata adults were collected in this period): total abundance/plot (±SD) CONV: 3.8 (3.9), ZERO: 1.8 (2.1), UNTR: 17.0 (14.5), W(2; 5.9) = 5.4, p = 0.05. The abundance of Linyphiidae remained low in all treatments in the autumn, particularly in the CONV (24%) and ZERO (19%) plots compared to the UNTR (100%) plots. N. bimaculata juveniles and adults, and Tetragnatha juveniles occurred in the canopy almost exclusively in the autumn. Their abundance in the CONV plots was significantly higher than in the ZERO plots (N. bimaculata: TCONV/ZERO(3; 28) = 4.86, p < 0.01; Tetragnatha: TCONV/ZERO(3; 25) = 4,40, p < 0.05) and was not different from the
Author's personal copy
V. Markó et al. / Biological Control 48 (2009) 115–124
The community composition of canopy spiders was different between years whilst the effect of the different treatments on the species composition was low (Fig. 4). Metric ordination of the cumulated data of the four years revealed that the genus composition of adult and juvenile assemblages, independent from the treatments, differed significantly (Fig. 4). The proportions of the main genera and Linyphiidae in the total catch (within adults and juveniles) were: Araniella (26%, 33%), Theridion (27%, 21%), Neottiura (bimaculata) (0.3%, 14%), Philodromus (5%, 14%), Linyphiidae (26%, 3%), Tetragnatha (0.1%, 4%), Xysticus (0%, 1.5%) and other genera (15.6%, 9.5%). Smaller differences were found between the treatments within adult and juvenile assemblages (Fig. 4). The spring and autumn spider assemblages were separated along Axis 1 while the composition of spiders from different treatments showed minor differences along Axis 2 (Fig. 4). 3.4. Pattern of potential prey and spider abundance During the four year study the most abundant prey organisms in May, independent from the collecting method used, were beetles, parasitic wasps, apple suckers (Psylla mali Schmidberger) and leafhoppers. Leafhoppers and parasitic wasps (2001), apple suckers (2002) and beetles, predacious bugs and apple suckers (2004) dominated the samples in June and July. Leafhoppers and parasitic wasps (2001, 2002) and predacious bugs and beetles (2004) were common in the canopy in August. The main potential prey group was leafhoppers during the second population peak of spiders in September and first week of October in all years. The aphid abundance was significantly higher in the UNTR plots compared to CONV and ZERO plots especially in May and June in 2001 and 2006, the years with higher aphid infestation (Cross and Berrie, 2005). No significant difference was found between CONV and ZERO plots during the whole season with the exception of 2001, when aphids (Dysaphis plantaginea (Passerini)) were more abundant in the ZERO plots in May and June (Cross and Berrie, 2005). The total abundance of potential prey was higher in the UNTR plots compared to CONV and ZERO plots throughout the whole season. The abundance in CONV and ZERO plots were more similar (not shown). The abundance of potential prey was reflected by the main prey groups just before and during the period of peak spider abundance in August, September and first week of October (Appendix B). Analysis of the abundances by two-way ANOVA (prey and years) revealed that more leafhoppers, beetles, bugs, apple suckers and aphids were collected in the canopy in the UNTR plots than in the CONV and ZERO plots. The potential prey abundance was identical in the CONV and ZERO treatments (Appendix B) with the exception of parasitic wasps where the abundance differed from year to year. However, in general, more parasitic wasps were collected in the UNTR and CONV compared to the ZERO plots (Appendix B). Multiple regressions on the potential prey data retained the variables Auchenorrhyncha (‘total spider abundance in autumn’ and Theridion), Heteroptera (Philodromus), Psylla and Hymenoptera
CONV 2002
0.15
UNTR
0.1
ZERO
CONV
UNTR
0.05
Axis 2 (16%)
3.3. Composition of canopy spider assemblages
0.2
2004 0
ZERO
UNTR ZERO CONV
-0.05
UNTR ZERO
2006
-0.1
2001
-0.15 -0.2
CONV
-0.25 -0.3
-0.2
-0.1
0
0.1
0.2
0.3
Axis 1 (58%) 0.25 0.2
AD UNTR
0.15
Axis 2 (10.5%)
abundance in the UNTR plots (N. bimaculata: TCONV/UNTR(3; 23) = 1.95, p > 0.1, TZERO/UNTR(3; 19) = 5.00, p < 0.01; Tetragnatha: TCON/ UNTR(3; 24) = 3.01, p > 0.1, TZERO/UNTR(3; 30) = 1.01, p > 0.1) (Fig. 3). All six genera and Linyphiidae analysed compensated, or overcompensated the higher initial pesticide disturbance in the CONV plots compared to ZERO plots by the autumn, but only Xysticus and Tetragnatha compensated the relatively higher disturbance in ZERO plots compared to UNTR plots.
0.1
JUV ZERO
AD ZERO
0.05 0 -0.05
JUV CONV JUV UNTR
-0.1
AD CONV
-0.15 -0.2 -0.4
-0.3
-0.2
-0.1
0
0.1
0.2
0.3
0.4
Axis 1 (49.3%) 0.2 0.15
SPR UNTR AUT UNTR
0.1
Axis 2 (14.5%)
120
0.05
SPR ZERO
AUT ZERO
0 -0.05 -0.1
SPR CONV
AUT CONV
-0.15 -0.3
-0.2
-0.1
0
0.1
0.2
0.3
Axis 1 (50.7%) Fig. 4. (A) Similarity of the genus composition of canopy spider assemblages in the four year’s of the study. (B) Similarity of the genus composition of adult (AD) and juvenile (JUV) spider assemblages and (C) the canopy spider assemblages in spring (SPR) and autumn (AUT). CONV, conventionally treated; ZERO, zero pesticide residue; UNTR, untreated control plots.
(Neottiura) in the models, accounting for 31–69% of the total variation (Table 5). No additional variance was explained by adding the ‘spider abundance in May’ variables to the models with exception of genus Araniella where variable ‘Aphididae’ was the most power-
Author's personal copy
121
V. Markó et al. / Biological Control 48 (2009) 115–124
Table 5 Relationship between the abundance of the total spider community, the most common genera and the studied potential prey groups in autumn by stepwise multiple linear regression. (df = 23). Standardized Beta Coefficients are shown if the relationship is significant. Total abundance
r2 F p Auchenorrhyncha Coleoptera Heteroptera Hymenoptera Aphididae Psylla Spider abundance in May
Araniella
2001 + 2002
2004 + 2006
2004 + 2006
0.313 11.462 0.003 0.585** nsa
0.690 52.301 0.000 0.839*** nsa ns nsa ns ns ns
0.446 10.255 0.001 ns nsa ns nsa 0.601** nsa 0.361*,a
ns
ns
Theridion
Philodromus
Neottiura
0.465 20.994 0.000 0.699*** nsa ns nsa nsa ns ns
0.390 15.718 0.001 ns nsa 0.646** nsa ns nsa nsa
0.404 8.796 0.002 nsa nsa nsa 0.436* nsa 0.459*,a nsa
Significant relationships are indicated by asterisks. ns, not significant. * p < 0.05. ** p < 0.01. *** p < 0.001. a Investigating separately the independent and dependent variable no correlation was found (p > 0.05).
ful predictor followed by ‘Araniella abundance in May’ (Table 5). It is important to emphasize that many of the predictor variables were intercorrelated: for example between the variable Auchenorrhyncha, Heteroptera, Aphididae and Psylla but also between the Coleoptera and Hymenoptera positive correlation was observed, and spider densities in May and ‘total prey abundance’ were often not independent too (not shown). By contracting the prey groups into one variable (‘total prey abundance’) and removing the monotonic effect of the variable ‘spider abundance in May’ we found that there is a significant positive correlation between the ‘total prey abundance’ and ‘spider abundance in autumn’ both in 2001–2002 (Kendall’s partial tau = 0.382, p = 0.009) and in 2004–2006 (tau = 0.577, p < 0.001). Conversely, there was no monotonic trend between the variables ‘spider abundance in May’ and ‘spider abundance in autumn’ if the effect of the variable, ‘total prey abundance’, was removed (2001 + 2002: tau = 0.212, p = 0.147; 2004 + 2006: tau = 0.203, p = 0.164). Overall, the results indicate that the spider abundance followed the pattern of the prey abundance in the investigated orchard plots in autumn while spider abundance in May played limited role in forming the abundance pattern of autumn spider community.
4. Discussion The use of target specific selective pesticides is a key element of environmentally safe plant protection in orchards (Cross, 2002). Currently, there are several compounds that according to laboratory tests have only moderate toxicity to spiders. In field studies, the abundance and species richness of spider communities has been shown to be higher in organic and IPM orchards where selective compounds were used (Olszak et al., 1992a; Bogya et al., 2000; Cárdenas et al., 2006). This reduction of impact on spider populations is most certainly due to the lower direct toxicity of these pesticides and the retention of prey availability in these systems. In our study the species richness, and total annual spider abundance did not differ between the two pesticide treatment practices. This result differs from other orchard studies where the abundance of spiders was found to be higher in orchards treated with selective insecticides compared to conventional applications throughout the whole growing season (Olszak et al., 1992a; Bogya et al., 2000; Cárdenas et al., 2006). Other studies have demonstrated that spider abundance, at least in some years, can be higher
in orchards treated with less selective pesticides (Olszak et al., 1992a; Pekár, 1999) and it was suggested that the rapid increase of spider abundances in the second part of the growing season was due to immigration (Olszak et al., 1992a) or increased fecundity of adults (Pekár, 1999). Our results showed that the effect of insecticides was particularly harmful to spiders as the treatments coincide with the peak adult abundance in May and early June. Komorek and Vogt (2000) and Wisniewska and Prokopy (1997b) documented a similar increase in adults in the early part of the season, in German and Massachusetts apple orchards, respectively. We also demonstrated that the pesticide treatments had a detrimental affect on female spiders. Males were less affected, probably because they are more mobile (searching for females) and less dependent on prey. As a result, the sex ratio shifted from a female bias towards a male bias in the pesticide treated plots. Similar tendencies were found in all spider families sampled in this study. The analysis of annual activity of spiders and abundances of potential prey may explain why, in spite of its lower chemical disturbance to spiders, the ZERO treatment had not resulted in a higher total annual abundance compared to CONV plots, even after six– years application of pest management programmes (Table 2). In May, with the use of insecticides with different toxicity to spiders, but detrimental to pests, we created higher spider abundance (by 30–105%) in the ZERO compared to the CONV plots. However, the positive effects of the lower chemical disturbance on spider abundance, in contrast to our hypothesis ‘B’, diminished rapidly. This indicated that if applications of pesticides harmful to spiders were restricted to the first part of the growing season then juveniles may be able to colonise the orchard. Additionally, we revealed that the post-disturbance increase in abundance was 1.7–3.2 times higher in the CONV plots and 2.3–3.3 times higher in the UNTR plots compared to ZERO plots. For example the difference between the spider abundance in CONV and ZERO, and ZERO and UNTR plots was similar in May 2004 (16 and 15 individuals). However, the difference diminished only between the CONV and ZERO plots by autumn and the distance between the spider abundances in ZERO and UNTR plots even increased (Table 4). To explain this pattern we studied the potential prey abundance and its potential effect on spider densities. In this study, the potential availability of prey for spiders was similar in the two pest management technologies, especially in the latter part of the fruit growing season (during increase of spider abundances in late summer/autumn), compared to the UNTR plots
Author's personal copy
122
V. Markó et al. / Biological Control 48 (2009) 115–124
where the prey abundance was higher. Multiple regression and Kendall’s partial tau correlation analyses revealed that the pattern of spider abundance followed the pattern of prey in this part of the season according to hypothesis ‘C’ (Table 1). A high immigration rate of spiders (hypothesis A) may explain the similar spider abundances in the CONV and ZERO plots in autumn; or a low immigration rate (hypothesis B) may explain the lower abundance in the pesticide treated plots compared to UNTR; but only a high immigration rate and strong regulation by prey abundance (hypothesis C) can explain both. In a similar study, Wisniewska and Prokopy (1997b) demonstrated a 2- to 3-fold increase in autumn spider abundances in orchards where pesticide treatments were restricted up to midJune compared to those applied throughout the season. Comparing the apple pest abundance in the same orchards Prokopy et al. (1996) found no differences in the abundance of leaf-rollers, apple maggot fly, aphids, spider mites and their predators, whilst the leafhoppers (Edwardsiana rosae (Linnaeus), Empoasca fabae (Harris)) migrated into the orchards where pesticide treatments had finished in mid-June. It is likely that for spiders the principal advantage of halting pesticide applications after mid June was a reduced effect on potential spider prey populations, particularly leafhoppers in the latter part of the growing season (Prokopy et al., 1996; Wisniewska and Prokopy, 1997a,b). Our study also indicated that leafhoppers could be the most important prey for the autumn spider communities. They were not only the most abundant potential prey but also the best predictors of total spider abundance in the orchard plots in the autumn. In contrast, the influence of coleopterans on spider abundance was moderate. All seven spider groups in this study compensated for the higher initial pesticide disturbance in the CONV plots compared to ZERO plots by the autumn, thus no genera showed the pattern predicted by hypothesis ‘B’ (Table 1). Three genera (Araniella, Theridion and Philodromus), showed a pattern similar to the total spider communities in agreement with hypothesis ‘C’. One genus, Xysticus, was not affected by the pesticide treatments in agreement with hypothesis ‘A’ (Table 1). In addition, the abundance of two canopy species, Neottiura bimaculata and Tetragnatha (extensa), was higher in the CONV than the ZERO plots, implying a rapid colonisation into the CONV plots. The pattern of some prey (e.g. parasitic hymenopterans) or effect of intraguild predators could explain their distribution in the orchard plots (Table 5, Appendix B). The linyphiids (mainly E. acuminata) differed from the pattern observed in other spiders. The higher pesticide pressure (CONV) did not cause a decrease in their abundance compared to ZERO plots in the spring and their abundance remained low in the autumn in all of the pesticide treated and control plots. There were two peaks of spider abundance in the orchards, one in the spring dominated by adults, and another in the autumn dominated by juveniles. Between mid-June and the beginning of August spider abundance was very low in all of the orchard plots. We found substantial differences between the composition of adult and juvenile spider assemblages and in connection with this also between the spring and autumn spider assemblages both in the pesticide treated and untreated plots. This suggests significant restructuring in spider assemblages during the season, i.e. movement between the adult and juvenile habitats, different number of offspring and different mortality rates in different spider species. The genera Neottiura, Tetragnatha and Xystichus, occurred in the orchard canopy in the latter part of the growing season and
consisted almost exclusively of juveniles. The adults of these genera are common in many habitats, e.g. arable and potato fields (Luczak, 1979) and X. cristatus and X. kochi were also common on the ground in the orchard between April and July (data not shown). Juveniles of these species probably migrated into the canopy of apple trees, i.e. they changed their habitat towards the end of the growing season. The proportion of adults within linyphiids (mainly E. acuminata) was high, even in the autumn, when their abundance remained low. Linyphiid juveniles may leave the apple orchard when their preferred prey (e.g. aphids migrating to summer hosts) becomes scarce in the canopy. Between-habitat movement of the genera Araniella, Theridion and Philodromus is also possible as all three are most likely common in the surrounding habitats (Olszak et al., 1992b). In our study, the spider communities were very similar in the differently treated orchard plots. Bogya et al. (1999b) and Brown et al. (2003) reported that pesticide treatments affected the composition of orchard spider communities only marginally, while the orchard environment played a more significant role. 5. Conclusion In our study, in southern England, the use of insecticides less harmful to spiders in zero pesticide residue pest management practices resulted in higher spider abundance compared to the conventional pest management strategy based on broad spectrum insecticides in the first part of the fruit growing season. However, as the applications of less selective pesticides is restricted to the first part of July the abundance of spider communities become identical by the autumn due to the immigration of spiders from surrounding areas. The pesticide treatments were particulary harmful to female spiders while males were less affected. Consequently, the proportion of males increased in the pesticide treated orchards. The post-pesticide disturbance population increase of spiders (immigration, emigration and probably within orchard reproduction) was regulated by prey availability. The spider communities restructured after July probably because of habitat change of the juveniles and between-species differences in offspring and mortality. This restructuring occured both in non-disturbed and pesticide treated plots of the studied orchard. In conclusion, spider populations were only affected short-term by direct toxicity from pesticide; prey reduction regulating the re-colonisation had a far greater influence on spider populations in the studied apple orchard. Further research is needed to determine the role of spiders in controlling orchard pests and to define the mechanisms regulating the organisation of spider assemblages of orchards in different geographical regions and with different size and surrounding vegetation. Acknowledgments We thank Krisztina Bleicher, Csaba Nagy and Peter Sipos for their help in the sampling and Ferenc Samu and the anonymous reviewers for their comments on an earlier version of this paper. This study was funded by DEFRA (No. HH3122STF) and partly by OTKA (No. 46380).
Author's personal copy
123
V. Markó et al. / Biological Control 48 (2009) 115–124
Appendix A Insecticides used in the conventional (CONV) and zero pesticide residue (ZERO) pest management apple orchards.
a
Black highlighting: insecticide highly toxic to spiders. Grey highlighting: insecticides moderately toxic to spiders. No highlighting: insecticides harmless to spiders.
b c
Appendix B Mean density (±SD) of potential prey arthropod groups in the canopy assessed by yellow sticky traps (2001 and 2002) and beating (2004 and 2006) in the latter part of the fruit growing season (August, September and October). CONV, conventional treatments (broad spectrum insecticides); ZERO, zero pesticide residue treatments (selective insecticides); UNTR, untreated control. Auchenorrhyncha
Hymenoptera
Coleoptera
Heteroptera
Psylla
Aphididae
2001 CON ZERO UNTR
268 (131.9) a 237 (72.1) a 511 (127.1) b
72 (20.4) a 98 (21.0) b 144 (37.3) c
12 (3.8) A 12 (3.5) A 31 (10.1) B
1 (0.5) A 4 (3.3) B 27 (8.1) C
2 (1.0) 2 (1.3) 5 (4.1)
No sampling No sampling No sampling
2002 CON ZERO UNTR
174 (38.1) A 147 (42.1) A 343 (72.7) B
194 (60.3) a 158 (26.4) a 172 (27.9) a
36 (12.9) A 35 (10.4) A 48 (9.9) A
1 (1.7) 2 (1.9) 4 (3.0)
4 (1.3) A 10 (6.0) B 19 (7.4) C (continued
No sampling No sampling No sampling on next page)
Author's personal copy
124
V. Markó et al. / Biological Control 48 (2009) 115–124
Appendix B (continued) Auchenorrhyncha
Hymenoptera
Coleoptera
Heteroptera
Psylla
Aphididae
2004 CON ZERO UNTR
45 (15.8) A 27 (12.4) B 255 (44.1) C
39 (8.4) a 24 (9.8) b 39 (3.3) a
57 (7.5) A 41 (9.1) B 86 (22.0) C
44 (10.6) A 52 (11.0) A 120 (41.6) B
3 (3.0) a 2 (2.1) a 44 (39.8) b
14 (2.8) ab 11 (4.4) a 18 (5.0) b
2006 CON ZERO UNTR
23 (5.0) A 18 (8.6) A 172 (31.4) B
16 (5.9) A 8 (1.6) B 20 (5.8) A
34 (19.4) A 21 (3.0) A 42 (20.8) A
11 (4.7) a 8 (3.2) a 18 (3.4) b
0 (0.0) A 0 (0.5) A 15 (7.9) B
25 (5.0) a 25 (6.5) a 48 (17.2) b
Mean* CON ZERO UNTR
127 A 107 A 320 B
80 A (a) 72 B (b) 93 A (c)
35 A 27 A 52 B
19 A 21 A 55 B
2.25 A 3.833 A 26.08 B
20 A 18 A 33 B
*
Unweighted mean of the years’ catches; two-way ANOVA (treatments & years), means followed by different capitals or lowercase letters within a column and year represent significant p < 0.05 or marginally significant p < 0.10 differences, respectively.
References Amalin, D.M., Pena, J.E., Yu, S.J., McSorley, R., 2000. Selective toxicity of some pesticides to Hibana velox (Araeae: Anyphaenidae), a predator of citrus leafminer. The Florida Entomologist 83, 254–262. Bleicher, K., Markó, V., Cross, J.V., Orosz, A., 2003. Characterizing the leafhopper (Auchenorrhyncha) biodiversity in an apple orchard with reduced pesticide management at East Malling, UK. IOBC/wprs Bulletin 26, 15–20. Bogya, S., Szinetár, Cs., Markó, V., 1999a. Species composition of spider (Araneae) assemblages in apple and pear orchards in the Carpathian Basin. Acta Phytopathologica et Entomologica Hungarica 34, 99–122. Bogya, S., Markó, V., Szinetár, Cs., 1999b. Comparison of pome fruit inhabiting spider assemblages at different geographical scales. Agricultural and Forest Entomology 1, 261–269. Bogya, S., Markó, V., Szinetár, Cs., 2000. Effect of pest management systems on foliage- and grass-dwelling spider communities in an apple orchard in Hungary. International Journal of Pest Management 46, 241–250. Bostanian, N.J., Dondale, C.D., Binns, M.R., Pitre, D., 1984. Effects of pesticide use on spiders (Araneae) in Quebec apple orchards. Canadian Entomologist 116, 663– 675. Brown, M.W., Schmitt, J.J., Abraham, B.J., 2003. Seasonal and diurnal dynamics of spiders (Araneae) in West Virginia orchards and the effect of orchard management on spider communities. Environmental Entomology 32, 830– 839. Cárdenas, M., Ruano, F., García, P., Pascual, F., Campos, M., 2006. Impact of agricultural management on spider populations in the canopy of olive trees. Biological Control 38, 188–195. Cross, J.V. (Ed.), 2002. Guidelines for integrated production of pome fruits in Europe. IOBC Technical Guideline III. IOBC wprs Bulletin Bulletin OILB srop 25, 1–8. Cross, J.V., Berrie, A.M., 2005. Producing apples free of residues. Proceedings of the British Crop Production Council International Congress, 775–782. Fountain, M.T., Brown, V.K., Gange, A.C., Symondson, W.O.C., Murray, P.J., 2007. The effects of the insecticide chlorpyrifos on spider and Collembola communities. Pedobiologia 51, 147–158. Gillespie, R.G., Caraco, T., 1987. Risk-sensitive foraging strategies of two spider populations. Ecology 68, 887–899. Harwood, J.D., Sunderland, K.D., Symondson, W.O.C., 2001. Living where the food is: web location by linyphiid spiders in relation to prey availability in winter wheat. Journal of Applied Ecology 38, 88–99. Hassan, S.A., Bigler, F., Bogenschütz, H., Boller, E., Brun, J., Calis, J.N.M., CoremansPelseneer, J., Duso, C., Grove, A., Heimbach, U., Helyer, N., Hokkanen, H., Lewis, G.B., Mansour, F., Moreth, L., Polgar, L., Samsøe-Petersen, L., Sauphanor, B., Stäubli, A., Sterk, G., Vainio, A., van de Veire, M., Viggiani, G., Vogt, H., 1994. Results of the sixth joint pesticide testing programme of the IOBC/WPRSworking group; pesticides and beneficial organisms. Biocontrol 39, 107–119. Komorek, M., Vogt, H., 2000. Investigations of side-effects of two insect growth regulators and an organophosphate on dominant spiders in an apple orchard. Pesticides and Beneficial Organisms, IOBC/wprs Bulletin 23, 111–126. Krebs, C.J., 1999. Ecological Methodology. Benjamin Cummings, Menlo Park, California. pp. 1–620. Luczak, J., 1979. Spiders in agroecosystems. Polish Ecological Studies 5, 151–200. Marc, P., Canard, A., Ysnel, F., 1999. Spiders (Araneae) useful for pest limitation and bioindication. Agriculture, Ecosystems and Environment 74, 229–273. Mansour, F., 1987. Effect of pesticides on spiders occurring on apple and citrus in Israel. Phytoparasitica 15, 43–50. Mansour, F., Rosen, D., Shulov, A., 1980. A survey of spider populations (Araneae) in sprayed and unsprayed apple orchards in Israel and their ability to feed on larvae of Spodoptera littoralis (Boisd.). Acta Oecologica/Oecologia Applicata 1, 189–197.
Miliczky, E.R., Calkins, C.O., Horton, D.R., 2000. Spider abundance and diversity in apple orchards under three insect pest management programmes in Washington State, USA. Agricultural and Forest Entomology 2, 203–215. Miliczky, E.R., Calkins, C.O., 2002. Spiders (Araneae) as potential predators of leaf roller larvae and egg masses (Lepidoptera: Tortricidae) in Central Washington apple and pear orchards. Pan-Pacific Entomologist 78, 140–150. Miliczky, E.R., Horton, D.R., 2005. Densities of beneficial arthropods within pear and apple orchards affected by distance from adjacent native habitat and association of natural enemies with extra-orchard host plants. Biological Control 33, 249–259. Nicholas, A., Thwaite, W., Spooner-Hart, R., 1999. Arthropod abundance in an Australian apple orchard under mating disruption and supplementary insecticide treatments for codling moth, Cydia pomonella (L.) (Lepidoptera: Tortricidae). Australian Journal of Entomology 38, 23–29. Olszak, R.W., Luczak, J., Neimczyk, E., Zajac, R., 1992a. The spider community associated with apple trees under different pressure of pesticides. Ekologia Polska 40, 265–286. Olszak, R.W., Luczak, J., Zajac, R., 1992b. Species composition and numbers of spider communities occurring in different species of shrubs. Ekologia Polska 40, 287– 313. Pekár, S., 1999. Effect of IPM practices and conventional spraying on spider population dynamics in an apple orchard. Agriculture, Ecosystems and Environment 73, 155–166. Pekár, S., 2002. Susceptibility of the spider Theridion impressum to 17 pesticides. Journal of Pest Science 75, 51–55. Pekár, S., Kocourek, F., 2004. Spiders (Araneae) in the biological and integrated pest management of apple in the Czech Republic. Journal of Applied Entomology 128, 561–566. Podani, J., 2001. SYN-TAX 2000. Computer program for data analysis in ecology and systematics. User’s manual. Scientia, Budapest, Hungary, pp. 1–53. Powell, W., Dean, G.J., Bardner, R., 1985. Effects of pirimicarb, dimethoate and benomyl on natural enemies of cereal aphids in winter wheat. Annals of Applied Biology 16, 235–242. Prokopy, R.J., Mason, J.L., Christie, M., Wright, S.E., 1996. Arthropod pest and natural enemy abundance under second-level versus first-level integrated pest management practices in apple orchards: a 4-year study. Agriculture, Ecosystems and Environment 57, 35–47. Schmuck, R., 2001. Ecotoxicological profile of the insecticide thiacloprid. Pflanzenschutz-Nachrichten Bayer 54, 161–184. Tanaka, K., 1989. Energetic cost of web construction and its effect on web relocation in the web-building spider Agelena limbata. Oecologia 81, 1432–1939. Vargha, A., 2007. The statistical menu system of RopStat. Available from:
. Villanueva, R.T., Walgenbach, J.F., 2005. Development, oviposition, and mortality of Neoseiulus fallacis (Acari: Phytoseiidae) in response to reduced-risk insecticides. Journal of Economic Entomology 98, 2114–2120. Weyman, G.S., Sunderland, K.D., Jepson, P.C., 2002. A review of the evolution and mechanisms of ballooning by spiders inhabiting arable farmland. Ethology Ecology & Evolution 14, 307–326. Wise, D.H., 1993. Spiders in Ecological Webs. Cambridge University Press, Cambridge, UK. pp. 1–342. Wisniewska, J., Prokopy, R.J., 1997a. Do spiders (Araneae) feed on rose leafhopper (Edwardsiana rosae; Auchenorrhyncha: Cicadellidae) pests of apple trees? European Journal of Entomology 94, 243–251. Wisniewska, J., Prokopy, R.J., 1997b. Pesticide effect on faunal composition, abundance, and body length of spiders (Araneae) in apple orchards. Environmental Entomology 26, 763–776. Wyss, E., Niggli, U., Nentwig, W., 1995. The impact of spiders on aphid populations in a strip-managed apple orchard. Journal of Applied Entomology 119, 473–478.