- Email: [email protected]
Comparison of soil surface arthropod populations in conventional tillage, no-tillage and old field systems
Agro-Ecosystems, Elsevier Scientific
8 (1983) 247-253 Publishing Company,
247 Amsterdam
- Printed
in The Netherlands
COMPARISON OF SOIL SURFACE ARTHROPOD POPULATIONS IN CONVENTIONAL TILLAGE, NO-TILLAGE AND OLD FIELD SYSTEMS
A.Y. BLUMBERG
and D.A. CROSSLEY,
Department of Entomology GA 30602 (U.S.A.) (Accepted
Jr.
and Institute of Ecology,
University of Georgia, Athens,
13 April 1982)
ABSTRACT Blumberg, A.Y. and Crossley, D.A. Jr., 1983. Comparison of soil surface arthropod lations in conventional tillage, no-tillage and old field systems. Agro-Ecosystems, 247-253.
popu8:
Soil surface arthropod populations in conventional tillage (CT) and no-tillage (NT) sorghum and adjacent old field (OF) were compared using pitfall trap captures. Overall diversity (H) and similarity quotients (QS) between systems were calculated for each of seven 24-h sampling periods throughout the season. Diversity of the soil surface arthropods was greater in NT than either CT or OF. Although each system was distinct (any two of the systems had less than 30% of their species in common), NT was most similar to OF and least similar to CT during a period of drought and after heading of the sorghum. Percentages of individuals and species represented by spiders were similar in NT (30 and 15%) and OF (22 and 17%); percentages were substantially less in CT (11 and 8%). Yields (biomass of sorghum) in CT and NT were not significantly different despite the generally predicted higher pest populations in NT in the absence of insecticides. Leaf area grazed by insects was greater in CT (28%) than in NT (12%).
INTRODUCTION
As the crop acreage under no-tillage (NT) practices increases, research is being expanded on the composition of the arthropod complex associated with these systems in contrast to the corresponding complex in analogous conventional tillage (CT) systems (All, 1978). Basic research concerning the composition of such systems and relevant comparisons is not presently available. The objective of the present research was to compare beta or betweenhabitat diversity (Whittaker, 1975) of the soil surface arthropod populations in three systems - CT sorghum, NT sorghum and old-field (OF). Using diversity and similarity indices, the populations sampled (arthropods captured in pitfall traps) were compared. In order to determine the complete arthropod complex, insecticides were not used in any of the systems. Both herbivorous (“pests”) and predaceous (“beneficial”) arthropods are susceptible to insecticide treatments. Research of this nature has not previously been conducted in the absence of insecticides. 0304-3746/83/000~00001$03.00
0 1983 Elsevier Scientific
Publishing
Company
248 MATERIALS
AND METHODS
The systems sampled are located at the University of Georgia’s Horseshoe Bend (HSB) Research Area in Athens, Georgia. The perimeter is forested, while the center is an old field except for the cultivated plots. Prior to acquisition by the Institute of Ecology in 1964, HSB was used for forage and pasture. The area used in the present study for CT and NT was the same in which Barrett (1968) cultivated a grain crop in 1966. Although not cultivated since 1966 the area has been naturally flooded and was experimentally burned in 1970 (Odum et al., 1974). In late spring 1978, the experimental area was cleared of woody growth, rotary mowed, and divided into eight 1012 m* plots. Plots were assigned either CT or NT treatment (four replicates each) at random. In June 1978 CT plots were plowed and drag harrowed; NT plots were not cultivated. Atrazine was applied to the entire area at a rate of 1.9 liters per 4047 m* which resulted in 0.9 kg active ingredient per 4047 m*. A two-row fluted coulter planter was used to plant 7.3 kg per 4047 m* Funk’sR 516 BR hybrid grain sorghum seed in 76-cm rows. Upon completion of planting, 6-12-12 fertilizer was broadcast at the rate of 272 kg per 4047 m*. Ten pitfall traps (Morrill, 1975) were installed in a regular pattern in each of three CT and three NT plots. Five traps were installed in an adjacent 506 m* section of OF. Traps were opened, fitted with a vial of alcohol beneath the funnel and left open for twenty-four periods throughout the season; when not set, traps were covered with plastic caps. Two indices were calculated for the comparison of systems. The Shannon index of diversity (H) is given by: s R(S)
= R
= 2
JljlO&JJpi
i=l where s is the total number of species in a sample and pi is the observed proportion of individuals belonging to the ith species (i = 1,2,...,s) (Shannon, 1948). Cumulative H was calculated by increasing the sample base by one trap increments in OF and randomly chosen plots of CT and NT (Fig. 1). The resulting curves were similar to those of cumulative number of traps which provide an adequate sample (Phillips, 1959). When the percentage increase in H fell below lo%, the sample was considered sufficient. Use of this criterion suggests samples of four traps from CT, three traps from NT and four traps from OF. This compared favorably with Uetz and Unzicker (1976) who, using species curves, determined that 3.8 traps of 15-cm diameter or 9.2 traps of 6.5-cm diameter were adequate for sampling wandering spiders. Therefore, although a total of 65 traps were installed, the comparisons presented are based on analysis of five traps (9.8 cm in diameter) in 506 m* of each system. (The average standard deviation per sampling date for CT
249
dC
0123456709 Number of Traps (98cm
doom 1
Fig. 1. Shannon index of general diversity (R) for soil surface arthropods calculated cumulatively as sample is increased by one trap increments. Sampling took place concurrently in conventional tillage (CT) and no-tillage (NT) sorghum and old field (OF) over a twenty-four hour period in June, 1978.
and NT, based on 30 traps in each system, was 3.4 and 4.0 individuals, respectively.) The degree of similarity of species captured in any two systems was estimated by Sjrenson’s quotient of similarity (QS): QS = 2c/(a+b) where Q and b are the numbers of species in samples A and B and c is the number of species common to both samples A and B (Sdrenson, 1948). RESULTS
A total of 643 individuals in 146 morphospecies were recognized from pitfall traps. (Insects were identified to family, but only morphospecies (species types) were identified within family.) Spiders (with the exception of the dwarf spiders) were identified to genus and, when possible, to species. Total numbers of individuals and species captured was greatest in NT (Fig. 2). The Shannon index of general diversity for NT and OF was higher than for CT, with the exception of the August and September samples when CT and OF diversities were equal (Fig. 3). Similarity of species (S&enson’s quotient of similarity) between CT and NT was greatest shortly after planting (36%) but fell to zero during the drought and ranged between 12
250 300
100 90
250
80
A! 0 4 200
.g
$ a E z’
70
fii CL 60 v)
‘5 .0 _C 6 150
%
50
iiB a 40 E z’ 30
100
20 50 IO 0
CT
NT
0
OF
CT
NT
Fig. 2. Total numbers of individuals and species Each bar represents five traps in 506 rn’.
H
CT/NT
~+-a
CT/OF
b-.-d
NT/OF
OF
captured
during seven sampling
dates.
Months
Fig. 3. Shannon index given as a percentage) tillage (NT) sorghum planting of sorghum;
of for and R -
general diversity (8) and SQrenson’s quotient of similarity (QS, soil surface arthropods in conventional tillage (CT) and noold field (OF). Each point represents five traps in 506 m’. (P first rainfall after planting).
and 29% the remainder of the season (Fig. 3). Trends during the drought indicate that the soil surface fauna in CT was most affected during the stress of drought while NT and OF were equally affected and less dissimilar.
251 DISCUSSION
Ecological dogma, though not resolved on cause and effect, couples the properties of stability and diversity, and diversity has been used as an indicator of relative stability (van Emden and Williams, 1974). With regard to programs of pest management, there is considerable support for the concept that, through various practices, an increase in the complexity of agroecosystems should lead to greater stability (Rabb et al., 1974). The importance of adjacent, more floristically diverse systems has also been recognized (van Emden and Williams, 1974). These systems may act as reservoirs of faunal diversity, including both herbivores and predators. Deliberate addition of diversity may not necessarily increase stability, and the manipulation of diversity through introduction of new species must be considered on the basis of knowledge of the entire complex present. Most comparative studies of the arthropods in CT and NT have focused on one or a few pest species in insecticide-stressed systems; results from these studies indicate that specific pest species may be enhanced, deterred or not affected by NT practices (All, 1978). Classifying habitats according to increasing spatial heterogeneity (Southwood, 1976) might be expected to parallel diversity measurements. If an increase in spatial diversity does allow for greater faunal diversity, and concurrently imparts an increase in stability, then NT systems should be more stable than CT, providing not only advantages such as reduced erosion and increased soil moisture but also a system less susceptible to pest outbreaks. Pitfall traps are most often used for the qualitative comparison of species occupying different habitats (Crossley et al., 1973, Hagvar et al., 1978). The efficacy of the method has been much discussed (Uetz and Unzicker, 1976). There are several specific methodological factors pertinent to the present study. Hagvar et al. (1978) found that catches from traps placed along the outer margins of a system did not differ significantly from those of traps in the center of the pattern when traps were 5 m apart (in the present study traps were at least 7 m apart). In trapping Carabidae (Coleoptera) Greenslade (1964) found that baiting, odor (from alcohol or other preservatives), and trap color (black, white or clear) caused no significant variation in catch. The decision to use alcohol in the present trapping was made after observing predation of trap contents by birds and predation inside the traps by some of the larger predatory species. In our study, we expected OF, the system of longest duration, to have the highest overall soil surface arthropod diversity, followed by NT, then CT. However, there were nearly 50% more individuals captured in NT than in CT (samples pooled). Concurrently, yield (above-ground biomass of sorghum) was not significantly different in the two systems at the 5% level (A. Kumara, personal communication, 1980). The increase in individuals in NT concomitant with greater species richness resulted in NT having the greatest diversity of the three systems. Thus, when not treated with insecticide, the
252
increase in diversity/stability in NT due to the increase in spatial diversity provided by the litter may ameliorate any increase in pest numbers or species. This possibility has far-reaching implications concerning insecticide use. Decisions concerning the timing of insecticide application should consider the possibility that system components may be decoupled (reducing interactions). Insecticides may at times actually destabilize an agricultural system, temporarily reducing pest populations but eliminating the ability of the intact arthropod complex to effect stabilization within the system. As with the problems of resistance (leading to resurgence), insecticides in NT systems may in truth amplify pest problems, not ameliorate them. Throughout the sampling period each of the three community types (CT, NT or OF) had a distinctive ground-surface arthropod community. This community evidently has a rapid response to either CT or NT practices. Trapping efficiency doubtless became different between the treatments, but this would not explain our observed differences in species composition between treatments. Hagvar et al. (1978) suggested that only if the vegetation is more or less equally dense at the soil surface could catches in different habitats be taken as a measure of relative density. However, surface-activity may be restricted by vegetational obstructions at the soil surface. Therefore, the larger catches in the habitats with denser vegetation and a litter layer, NT and OF, indicate that the density of animals is undoubtedly higher than that in the habitat which allows unrestricted movement, CT. Immediately after plowing, the NT community was more similar to the CT, but under drought the NT community became more similar to the OF. The early similarity was probably due to a common response to clearing and mowing. The NT community clearly was better buffered against drought than was the CT community. The soil surface arthropods were more numerous in NT, probably because of the particular combination of properties derived from cropping and natural systems: the increased moisture retention, litter component and variety of plant species available in NT paralleled conditions in a natural system while the crop, still an energy-subsidized system, provided an attractive nutrient-rich food source for herbivores and thus prey populations adequate to fuel predator and parasite populations. The complex of arthropods captured in NT was composed of relatively more individuals and species of predators and parasites than in CT. Spiders, for example, comprised a larger proportion of the catch in NT (30%) than in CT (11%) or OF (22%). Leaf area removed by grazing insects was greater in CT (28%) than NT (12%). The overall increase in arthropod numbers in NT with a decrease in herbivory relative to CT suggests that stability may parallel diversity in these systems. We assume that the greater complexity of the NT food web is indicative of higher relative stability, see Pimentel (1961). We do not infer resistance to or resilience following perturbation, however; further measurements of ecosystems-level phenomena will be required to identify these stability characteristics of NT systems.
253 ACKNOWLEDGEMENTS
We thank Drs. Robert L. Todd and J. Bruce Wallace, University of Georgia, for reviewing the manuscript. The research was supported in part by a contract, DE-AS09-76EV00641, between the U.S. Department of Energy and the University of Georgia Research Foundation, Inc. REFERENCES All, J.N., 1978. Insect relationships in no-tillage cropping. No-till Systems Conf. Experiment, University of Georgia, special publication No. 5, pp. 17-19. Barrett, G.W., 1968. The effects of an acute insecticide stress on a semi-enclosed grassland ecosystem. Ecology, 49: 1019-1035. Crossley, D.A., Jr., Coulson, R.N. and Gist, C.S., 1973. Trophic level effects on species diversity in arthropod communities of forest canopies. Environ. Entomol., 2: 1097-1100. Greenslade, P.J.M., 1964. Pitfall trapping as a method of studying populations of Carabidae (Coleoptera). J. Anim. Ecol., 33: 301-310. Hagvar, S., Ostsbye, E. and Melaen, J., 1978. Pitfall catches of surface active arthropods in some high mountain habitats at Finse South Norway. II. General results at group level with emphasis on Opiliones, Araneida, and Coleoptera. Norw. J. Entomol., 25 : 195-206. Morrill, W.L., 1975. Plastic pitfall trap. Environ. Entomol., 4: 596. Odum, E.P., Pomeroy, S.E., Dickinson, J.C. and Hutcheson, K., 1974. The effects of late winter burn on the composition, productivity and diversity of a 4 year-old fallow field in Georgia. Proc. Annu. Tall Timbers Ecol. Conf., 1973, pp. 399-414. Phillips, E.A., 1959. Methods of Vegetation Study. Henry Holt and Co., 107 pp. Pimental, D., 1961. Species diversity and insect population outbreaks. Annu. Entomol. Sot. Am., 54: 76-86. Rabb, R.L., Stinner, R.E. and Carlson, G.A., 1974. Ecological principles as a basis for pest management in the agro-ecosystem. In: F.G. Maxwell and F.A. Harris (Editors), proc., Summer Institute of Biological Control of Plant Insects and Diseases. Mississippi State University, pp. 19-45. Shannon, C.E., 1948. A mathematical theory of communication. Bell Syst. Tech. J., 27: 379-423.
Simpson, E.H., 1949. Measurement of diversity. Nature (London), 163: 688. Sdrenson, T., 1948. A method of establishing groups of equal amplitude in plant society based on similarity of species content. K. Danske vidensk. Selsk., 5: l-34. Southwood, T.R.E., 1976. Bionomic strategies and population parameters. In: R.M. May (Editor), Theoretical Ecology. W.B. Saunders, Philadelphia, 317 pp. Uetz, G.W. and Unzicker, J.D., 1976. Pitfall trapping in ecological studies of wandering spiders. J. Arach., 3: 101-111. van Emden, H.F. and Williams, G.C., 1974. Insect stability and diversity in agroecosystems. Annu. Rev. Entomol., 19: 455-475. Whittaker, R.H., 1975, Communities and Ecosystems. Macmillan, New York, 387 pp.
8 (1983) 247-253 Publishing Company,
247 Amsterdam
- Printed
in The Netherlands
COMPARISON OF SOIL SURFACE ARTHROPOD POPULATIONS IN CONVENTIONAL TILLAGE, NO-TILLAGE AND OLD FIELD SYSTEMS
A.Y. BLUMBERG
and D.A. CROSSLEY,
Department of Entomology GA 30602 (U.S.A.) (Accepted
Jr.
and Institute of Ecology,
University of Georgia, Athens,
13 April 1982)
ABSTRACT Blumberg, A.Y. and Crossley, D.A. Jr., 1983. Comparison of soil surface arthropod lations in conventional tillage, no-tillage and old field systems. Agro-Ecosystems, 247-253.
popu8:
Soil surface arthropod populations in conventional tillage (CT) and no-tillage (NT) sorghum and adjacent old field (OF) were compared using pitfall trap captures. Overall diversity (H) and similarity quotients (QS) between systems were calculated for each of seven 24-h sampling periods throughout the season. Diversity of the soil surface arthropods was greater in NT than either CT or OF. Although each system was distinct (any two of the systems had less than 30% of their species in common), NT was most similar to OF and least similar to CT during a period of drought and after heading of the sorghum. Percentages of individuals and species represented by spiders were similar in NT (30 and 15%) and OF (22 and 17%); percentages were substantially less in CT (11 and 8%). Yields (biomass of sorghum) in CT and NT were not significantly different despite the generally predicted higher pest populations in NT in the absence of insecticides. Leaf area grazed by insects was greater in CT (28%) than in NT (12%).
INTRODUCTION
As the crop acreage under no-tillage (NT) practices increases, research is being expanded on the composition of the arthropod complex associated with these systems in contrast to the corresponding complex in analogous conventional tillage (CT) systems (All, 1978). Basic research concerning the composition of such systems and relevant comparisons is not presently available. The objective of the present research was to compare beta or betweenhabitat diversity (Whittaker, 1975) of the soil surface arthropod populations in three systems - CT sorghum, NT sorghum and old-field (OF). Using diversity and similarity indices, the populations sampled (arthropods captured in pitfall traps) were compared. In order to determine the complete arthropod complex, insecticides were not used in any of the systems. Both herbivorous (“pests”) and predaceous (“beneficial”) arthropods are susceptible to insecticide treatments. Research of this nature has not previously been conducted in the absence of insecticides. 0304-3746/83/000~00001$03.00
0 1983 Elsevier Scientific
Publishing
Company
248 MATERIALS
AND METHODS
The systems sampled are located at the University of Georgia’s Horseshoe Bend (HSB) Research Area in Athens, Georgia. The perimeter is forested, while the center is an old field except for the cultivated plots. Prior to acquisition by the Institute of Ecology in 1964, HSB was used for forage and pasture. The area used in the present study for CT and NT was the same in which Barrett (1968) cultivated a grain crop in 1966. Although not cultivated since 1966 the area has been naturally flooded and was experimentally burned in 1970 (Odum et al., 1974). In late spring 1978, the experimental area was cleared of woody growth, rotary mowed, and divided into eight 1012 m* plots. Plots were assigned either CT or NT treatment (four replicates each) at random. In June 1978 CT plots were plowed and drag harrowed; NT plots were not cultivated. Atrazine was applied to the entire area at a rate of 1.9 liters per 4047 m* which resulted in 0.9 kg active ingredient per 4047 m*. A two-row fluted coulter planter was used to plant 7.3 kg per 4047 m* Funk’sR 516 BR hybrid grain sorghum seed in 76-cm rows. Upon completion of planting, 6-12-12 fertilizer was broadcast at the rate of 272 kg per 4047 m*. Ten pitfall traps (Morrill, 1975) were installed in a regular pattern in each of three CT and three NT plots. Five traps were installed in an adjacent 506 m* section of OF. Traps were opened, fitted with a vial of alcohol beneath the funnel and left open for twenty-four periods throughout the season; when not set, traps were covered with plastic caps. Two indices were calculated for the comparison of systems. The Shannon index of diversity (H) is given by: s R(S)
= R
= 2
JljlO&JJpi
i=l where s is the total number of species in a sample and pi is the observed proportion of individuals belonging to the ith species (i = 1,2,...,s) (Shannon, 1948). Cumulative H was calculated by increasing the sample base by one trap increments in OF and randomly chosen plots of CT and NT (Fig. 1). The resulting curves were similar to those of cumulative number of traps which provide an adequate sample (Phillips, 1959). When the percentage increase in H fell below lo%, the sample was considered sufficient. Use of this criterion suggests samples of four traps from CT, three traps from NT and four traps from OF. This compared favorably with Uetz and Unzicker (1976) who, using species curves, determined that 3.8 traps of 15-cm diameter or 9.2 traps of 6.5-cm diameter were adequate for sampling wandering spiders. Therefore, although a total of 65 traps were installed, the comparisons presented are based on analysis of five traps (9.8 cm in diameter) in 506 m* of each system. (The average standard deviation per sampling date for CT
249
dC
0123456709 Number of Traps (98cm
doom 1
Fig. 1. Shannon index of general diversity (R) for soil surface arthropods calculated cumulatively as sample is increased by one trap increments. Sampling took place concurrently in conventional tillage (CT) and no-tillage (NT) sorghum and old field (OF) over a twenty-four hour period in June, 1978.
and NT, based on 30 traps in each system, was 3.4 and 4.0 individuals, respectively.) The degree of similarity of species captured in any two systems was estimated by Sjrenson’s quotient of similarity (QS): QS = 2c/(a+b) where Q and b are the numbers of species in samples A and B and c is the number of species common to both samples A and B (Sdrenson, 1948). RESULTS
A total of 643 individuals in 146 morphospecies were recognized from pitfall traps. (Insects were identified to family, but only morphospecies (species types) were identified within family.) Spiders (with the exception of the dwarf spiders) were identified to genus and, when possible, to species. Total numbers of individuals and species captured was greatest in NT (Fig. 2). The Shannon index of general diversity for NT and OF was higher than for CT, with the exception of the August and September samples when CT and OF diversities were equal (Fig. 3). Similarity of species (S&enson’s quotient of similarity) between CT and NT was greatest shortly after planting (36%) but fell to zero during the drought and ranged between 12
250 300
100 90
250
80
A! 0 4 200
.g
$ a E z’
70
fii CL 60 v)
‘5 .0 _C 6 150
%
50
iiB a 40 E z’ 30
100
20 50 IO 0
CT
NT
0
OF
CT
NT
Fig. 2. Total numbers of individuals and species Each bar represents five traps in 506 rn’.
H
CT/NT
~+-a
CT/OF
b-.-d
NT/OF
OF
captured
during seven sampling
dates.
Months
Fig. 3. Shannon index given as a percentage) tillage (NT) sorghum planting of sorghum;
of for and R -
general diversity (8) and SQrenson’s quotient of similarity (QS, soil surface arthropods in conventional tillage (CT) and noold field (OF). Each point represents five traps in 506 m’. (P first rainfall after planting).
and 29% the remainder of the season (Fig. 3). Trends during the drought indicate that the soil surface fauna in CT was most affected during the stress of drought while NT and OF were equally affected and less dissimilar.
251 DISCUSSION
Ecological dogma, though not resolved on cause and effect, couples the properties of stability and diversity, and diversity has been used as an indicator of relative stability (van Emden and Williams, 1974). With regard to programs of pest management, there is considerable support for the concept that, through various practices, an increase in the complexity of agroecosystems should lead to greater stability (Rabb et al., 1974). The importance of adjacent, more floristically diverse systems has also been recognized (van Emden and Williams, 1974). These systems may act as reservoirs of faunal diversity, including both herbivores and predators. Deliberate addition of diversity may not necessarily increase stability, and the manipulation of diversity through introduction of new species must be considered on the basis of knowledge of the entire complex present. Most comparative studies of the arthropods in CT and NT have focused on one or a few pest species in insecticide-stressed systems; results from these studies indicate that specific pest species may be enhanced, deterred or not affected by NT practices (All, 1978). Classifying habitats according to increasing spatial heterogeneity (Southwood, 1976) might be expected to parallel diversity measurements. If an increase in spatial diversity does allow for greater faunal diversity, and concurrently imparts an increase in stability, then NT systems should be more stable than CT, providing not only advantages such as reduced erosion and increased soil moisture but also a system less susceptible to pest outbreaks. Pitfall traps are most often used for the qualitative comparison of species occupying different habitats (Crossley et al., 1973, Hagvar et al., 1978). The efficacy of the method has been much discussed (Uetz and Unzicker, 1976). There are several specific methodological factors pertinent to the present study. Hagvar et al. (1978) found that catches from traps placed along the outer margins of a system did not differ significantly from those of traps in the center of the pattern when traps were 5 m apart (in the present study traps were at least 7 m apart). In trapping Carabidae (Coleoptera) Greenslade (1964) found that baiting, odor (from alcohol or other preservatives), and trap color (black, white or clear) caused no significant variation in catch. The decision to use alcohol in the present trapping was made after observing predation of trap contents by birds and predation inside the traps by some of the larger predatory species. In our study, we expected OF, the system of longest duration, to have the highest overall soil surface arthropod diversity, followed by NT, then CT. However, there were nearly 50% more individuals captured in NT than in CT (samples pooled). Concurrently, yield (above-ground biomass of sorghum) was not significantly different in the two systems at the 5% level (A. Kumara, personal communication, 1980). The increase in individuals in NT concomitant with greater species richness resulted in NT having the greatest diversity of the three systems. Thus, when not treated with insecticide, the
252
increase in diversity/stability in NT due to the increase in spatial diversity provided by the litter may ameliorate any increase in pest numbers or species. This possibility has far-reaching implications concerning insecticide use. Decisions concerning the timing of insecticide application should consider the possibility that system components may be decoupled (reducing interactions). Insecticides may at times actually destabilize an agricultural system, temporarily reducing pest populations but eliminating the ability of the intact arthropod complex to effect stabilization within the system. As with the problems of resistance (leading to resurgence), insecticides in NT systems may in truth amplify pest problems, not ameliorate them. Throughout the sampling period each of the three community types (CT, NT or OF) had a distinctive ground-surface arthropod community. This community evidently has a rapid response to either CT or NT practices. Trapping efficiency doubtless became different between the treatments, but this would not explain our observed differences in species composition between treatments. Hagvar et al. (1978) suggested that only if the vegetation is more or less equally dense at the soil surface could catches in different habitats be taken as a measure of relative density. However, surface-activity may be restricted by vegetational obstructions at the soil surface. Therefore, the larger catches in the habitats with denser vegetation and a litter layer, NT and OF, indicate that the density of animals is undoubtedly higher than that in the habitat which allows unrestricted movement, CT. Immediately after plowing, the NT community was more similar to the CT, but under drought the NT community became more similar to the OF. The early similarity was probably due to a common response to clearing and mowing. The NT community clearly was better buffered against drought than was the CT community. The soil surface arthropods were more numerous in NT, probably because of the particular combination of properties derived from cropping and natural systems: the increased moisture retention, litter component and variety of plant species available in NT paralleled conditions in a natural system while the crop, still an energy-subsidized system, provided an attractive nutrient-rich food source for herbivores and thus prey populations adequate to fuel predator and parasite populations. The complex of arthropods captured in NT was composed of relatively more individuals and species of predators and parasites than in CT. Spiders, for example, comprised a larger proportion of the catch in NT (30%) than in CT (11%) or OF (22%). Leaf area removed by grazing insects was greater in CT (28%) than NT (12%). The overall increase in arthropod numbers in NT with a decrease in herbivory relative to CT suggests that stability may parallel diversity in these systems. We assume that the greater complexity of the NT food web is indicative of higher relative stability, see Pimentel (1961). We do not infer resistance to or resilience following perturbation, however; further measurements of ecosystems-level phenomena will be required to identify these stability characteristics of NT systems.
253 ACKNOWLEDGEMENTS
We thank Drs. Robert L. Todd and J. Bruce Wallace, University of Georgia, for reviewing the manuscript. The research was supported in part by a contract, DE-AS09-76EV00641, between the U.S. Department of Energy and the University of Georgia Research Foundation, Inc. REFERENCES All, J.N., 1978. Insect relationships in no-tillage cropping. No-till Systems Conf. Experiment, University of Georgia, special publication No. 5, pp. 17-19. Barrett, G.W., 1968. The effects of an acute insecticide stress on a semi-enclosed grassland ecosystem. Ecology, 49: 1019-1035. Crossley, D.A., Jr., Coulson, R.N. and Gist, C.S., 1973. Trophic level effects on species diversity in arthropod communities of forest canopies. Environ. Entomol., 2: 1097-1100. Greenslade, P.J.M., 1964. Pitfall trapping as a method of studying populations of Carabidae (Coleoptera). J. Anim. Ecol., 33: 301-310. Hagvar, S., Ostsbye, E. and Melaen, J., 1978. Pitfall catches of surface active arthropods in some high mountain habitats at Finse South Norway. II. General results at group level with emphasis on Opiliones, Araneida, and Coleoptera. Norw. J. Entomol., 25 : 195-206. Morrill, W.L., 1975. Plastic pitfall trap. Environ. Entomol., 4: 596. Odum, E.P., Pomeroy, S.E., Dickinson, J.C. and Hutcheson, K., 1974. The effects of late winter burn on the composition, productivity and diversity of a 4 year-old fallow field in Georgia. Proc. Annu. Tall Timbers Ecol. Conf., 1973, pp. 399-414. Phillips, E.A., 1959. Methods of Vegetation Study. Henry Holt and Co., 107 pp. Pimental, D., 1961. Species diversity and insect population outbreaks. Annu. Entomol. Sot. Am., 54: 76-86. Rabb, R.L., Stinner, R.E. and Carlson, G.A., 1974. Ecological principles as a basis for pest management in the agro-ecosystem. In: F.G. Maxwell and F.A. Harris (Editors), proc., Summer Institute of Biological Control of Plant Insects and Diseases. Mississippi State University, pp. 19-45. Shannon, C.E., 1948. A mathematical theory of communication. Bell Syst. Tech. J., 27: 379-423.
Simpson, E.H., 1949. Measurement of diversity. Nature (London), 163: 688. Sdrenson, T., 1948. A method of establishing groups of equal amplitude in plant society based on similarity of species content. K. Danske vidensk. Selsk., 5: l-34. Southwood, T.R.E., 1976. Bionomic strategies and population parameters. In: R.M. May (Editor), Theoretical Ecology. W.B. Saunders, Philadelphia, 317 pp. Uetz, G.W. and Unzicker, J.D., 1976. Pitfall trapping in ecological studies of wandering spiders. J. Arach., 3: 101-111. van Emden, H.F. and Williams, G.C., 1974. Insect stability and diversity in agroecosystems. Annu. Rev. Entomol., 19: 455-475. Whittaker, R.H., 1975, Communities and Ecosystems. Macmillan, New York, 387 pp.