An Experimental Analysis of Grasshopper Community Responses to Fire and Livestock Grazing in a Northern Mixed-Grass Prairie

COMMUNITY AND ECOSYSTEM ECOLOGY An Experimental Analysis of Grasshopper Community Responses to Fire and Livestock Grazing in a Northern Mixed-Grass Prairie

1–3 1,3 DAVID H. BRANSON AND GREGORY A. SWORD

Environ. Entomol. 39(5): 1441Ð1446 (2010); DOI: 10.1603/EN09378 ABSTRACT The outcomes of grasshopper responses to both vertebrate grazing and Þre vary across grassland ecosystems, and are strongly inßuenced by local climactic factors. Thus, the possible application of grazing and Þre as components of an ecologically based grasshopper management strategy must be investigated in regional studies. In this study, we examined the effects of grazing and Þre on grasshopper population density and community composition in a northern Great Plains mixed-grass prairie. We employed a large-scale, replicated, and fully-factorial manipulative experimental design across 4 yr to examine the separate and interactive effects of three grazing systems in burned and unburned habitats. Grasshopper densities were low throughout the 4-yr study and 1 yr of pretreatment sampling. There was a signiÞcant Þre by grazing interaction effect on cumulative density and community composition, resulting from burned season long grazing pastures having higher densities than unburned pastures. Shannon diversity and grasshopper species richness were signiÞ- cantly higher with twice-over rotational livestock grazing. The ability to draw strong conclusions regarding the nature of species composition shifts and population changes in the presence of Þre and grazing is complicated by the large site differences and low grasshopper densities. The results reinforce the importance of long-term research to examine the effects of habitat manipulation on grasshopper population dynamics.

KEY WORDS Þre, livestock grazing, Orthoptera, population dynamics

Grasshoppers are important invertebrate herbivores trol, habitat manipulations have the potential to main- in grassland ecosystems worldwide. Herbivory by tain or promote existing ecological feedbacks that grasshoppers can place them in direct competition for keep grasshopper populations below economically resources with both native and introduced vertebrate threatening levels (Branson et al. 2006). However, grazers. The strength of these competitive interac- opportunities to conduct the long-term manipulative tions will vary, but can be particularly intense during ecological experiments necessary to examine the ef- grasshopper outbreaks or when plant resources are fects of large-scale habitat manipulations on grasshop- rare, such as occurs during droughts (Belovsky and per populations are rare. Joern 1995, Belovsky and Slade 1995, Joern 2000, Bran- Foraging by livestock and native vertebrates can son 2008). Grasshoppers are also highly responsive to directly reduce food availability for grasshoppers altered habitats (Belovsky et al. 2000, Branson et al. through competition or indirectly via changes in plant 2006, Engle et al. 2008). For example, both grasshop- community composition (Fielding and Brusven 1996, per abundance and community composition can be Belovsky et al. 2000, OÕNeill et al. 2003). In addition, affected by factors such as Þre and vertebrate grazing the structure and microclimate of grasshopper habitat (Fielding and Brusven 1996; Onsager 2000; OÕNeill et can be affected by both grazing and trampling. Prom- al. 2003; Joern 2004, 2005). Branson et al. (2006) re- cently argued that burning and livestock grazing could ising evidence in support of using livestock grazing as be used as important habitat manipulation tools for use a tool for preventative grasshopper management was in ecologically based grasshopper management strat- provided by Onsager (2000). In OnsagerÕs study, grass- egies that seek to prevent or moderate the occurrence hopper densities in pastures subjected to a twice-over of damaging outbreaks. In contrast to chemical con- rotational (TOR) grazing scheme remained stable relative to those in pastures subjected to season-long grazing where densities sharply increased during a 1 United States Department of Agriculture-Agricultural Research Service, Northern Plains Agricultural Research Laboratory, Sidney, multi-year period of population growth. Grazing-me- MT 59270. diated changes in habitat structure in the TOR pas- 2 Corresponding author: United States Department of Agriculture- tures were hypothesized to have reduced favorable Agricultural Research Service, Northern Plains Agricultural Research microclimates available for thermoregulation, thereby Laboratory, Sidney, MT 59270 (e-mail: [email protected]). 3 School of Biological Sciences, University of Sydney, Sydney, Aus- suppressing population growth by reducing their de- tralia. velopment rate (Onsager 2000).

0046-225X/10/1441Ð1446$04.00/0 ᭧ 2010 Entomological Society of America 1442 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 5

Burning has also been shown to affect grasshopper are described in Onsager (2000). Block 1 (T145N, community composition and population density. Di- R99/98W), referred to here as Grassy Butte, consisted rect effects of Þre on grasshoppers include killing of three TOR pastures (140, 142, and 210 ha), a 162-ha individuals present in the habitat at the time of the SL pasture, and a 10-ha CO exclosure. At block 2 burn (e.g., Bock and Bock 1991) as well as egg mor- (T148N, R103W), referred to in this study as Sidney, tality as a result of elevated soil temperatures (Branson there were four TOR pastures (631, 519, 398, and 354 and Vermeire 2007). Possible indirect effects include ha), a 127-ha SL pasture, and a 10-ha CO exclosure. At changes in host plant quality or plant community com- block 3 (T150N, R103/102W), referred to in this study position (Porter and Redak 1996, Vermeire et al. as Alexander, there were four TOR pastures (228, 190, 2004a), changes in natural enemy abundance or be- 256, and 223 ha), a 323-ha SL pasture, and an 8-ha CO havior (Branson et al. 2006), and postÞre changes in exclosure. Both grazing treatments were designed to soil moisture or temperature regimes affecting egg result in 40Ð60% biomass utilization, which is consid- developmental phenology or mortality (Meyer et al. ered full grazing utilization. Controlled burns were 2002). Grasshopper responses to Þre can also be spe- conducted on 2.3Ð3.0 hectare plots in each CO exclo- cies speciÞc and heavily dependent upon the timing sure, TOR, and SL pasture in late October 2002, mim- and intensity of the burn (Joern 2004, Branson and icking a fall dormant season wildÞre. As part of a larger Vermeire 2007). study funded by the United States Forest Service, peak Importantly, the outcomes of grasshopper re- standing biomass, biomass utilization by livestock, sponses to both vertebrate grazing and Þre have re- basal cover, and species composition were examined peatedly been shown to vary across grassland ecosys- before burning in 2002 and after burning in 2003 tems. Furthermore, the effects are strongly inßuenced through 2005 using a combination of transects and by local climactic factors (Fielding et al. 2001, re- small exclosure cages (Plummer 2005, Clark 2006). viewed in Branson et al. 2006). Thus, the possible Grasshopper Sampling. Sites were sampled for application of grazing and Þre as components of an grasshopper population densities and species compo- ecologically based grasshopper management strategy sition every 3Ð4 wk between June and early Septem- must necessarily be investigated in regional studies. In ber, with Þve sampling dates in 2002 and 2003 and four this study, we examine the effects of grazing and Þre sampling dates in 2004 and 2005. Sampling took place on grasshopper population density and community on sunny days when air temperature was Ͼ23ЊC. Total composition in a northern Great Plains mixed-grass grasshopper population densities were estimated by prairie. We employed a large-scale, replicated, and counting the number of grasshoppers within 40, 0.1-m2 fully-factorial manipulative experimental design aluminum wire rings in burned and unburned plots in across 4 yr to examine the separate and interactive the three grazing treatments, following the methods of effects of three different grazing systems in burned Onsager and Henry (1977). Rings were permanently and unburned habitats. placed in four transects in each replicate, located Ϸ5 m from each other, and held in place with landscape staples. A sweep net sample, consisting of at least 150 Materials and Methods fast sweeps in the upper part of the plant canopy and Study Site. This study was conducted on the Little 150 slow sweeps at ground level, was taken to establish Missouri National Grasslands in western North Da- grasshopper community composition (Berry et al. kota, managed as part of the United States Forest 2000). Sweep net samples were frozen and grasshop- Service Dakota Prairie Grasslands. The landscape is pers were identiÞed to species and developmental eroded, characterized by wide summits and networks instar. Common grasshopper species at each of the of gullies (Butler et al. 1986). The plant community is three sites are presented in Table 1. As population a mixed grass prairie dominated by western wheat- densities were low throughout the study, cumulative grass (Pascopyrum smithii), blue grama (Bouteloua densities were calculated by summing densities over a gracilis), needle and thread (Hesperostipa comata), season to generate a measure of seasonal abundance and green needlegrass (Nassella viridula). The region that was less affected by random sampling error. is semiarid and receives Ϸ355Ð400 mm of annual pre- Grasshopper densities were sampled before treatment cipitation, most of which occurs during the growing application in 2001, and pretreatment grasshopper season from May to September, but is irregular in densities and species diversity metrics did not differ distribution. Mean daily temperatures range from signiÞcantly between treatments (P Ͼ 0.2 in all cases). Ϫ17.2ЊC in winter to 29.4ЊC in summer, with temper- Statistics. Patterns of grasshopper species diversity atures from 2002 through 2004 similar to long-term were examined using numerical species richness, Sh- averages. Temperatures during the summer of 2005 annon index of species diversity, and Simpson even- were above average. ness index to assess the relative abundance of species A three- or four-pasture TOR grazing system, sea- (Magurran 2004). Because of the low population den- son-long (SL) grazing treatment, and ungrazed con- sities during the study, treatment differences in grass- trol (CO) exclosure were initiated on each of three hopper densities were assessed by comparing the cu- separate blocks of similar soil and ecological types mulative densities, the sum of all ring density counts within a 2,352-km2 region in McKenzie County, ND, within 1 yr. Cumulative density data were square root using a randomized 2 ϫ 3 block design in the spring of transformed to achieve normality. Generalized linear 2002 (Plummer 2005, Clark 2006). Grazing treatments models treating year as a Þxed variable were initially October 2010 BRANSON AND SWORD:GRASSHOPPER RESPONSES TO FIRE AND GRAZING 1443

Table 1. The 15 most common grasshopper species by percentage composition at each of the three study sites from 2002 through 2005

Grassy Butte Sidney Alexander Species % Species % Species % Melanoplus dawsoni 27.05 Psoloessa delicatula 18.32 Eritettix simplex 22.08 Eritettix simplex 12.88 Ageneotettix deorum 12.25 Psoloessa delicatula 15.89 Opeia obscura 11.33 Melanoplus sanguinipes 9.99 Melanoplus sanguinipes 11.72 Orphulella speciosa 10.09 Opeia obscura 9.70 Opeia obscura 7.95 Melanoplus femurrubrum 5.85 Eritettix simplex 7.50 Ageneotettix deorum 7.28 Melanoplus sanguinipes 5.81 Philbostroma quadrimaculatum 7.23 Melanoplus infantilis 5.09 Aeropedellus clavatus 4.04 Melanoplus infantilis 6.47 Melanoplus femurrubrum 4.82 Ageneotettix deorum 3.23 Melanoplus gladstoni 4.23 Melanoplus gladstoni 3.98 Hypoclora alba 2.90 Encoptolophus costalis 4.06 Melanoplus keeleri 2.90 Trachyrachys kiowa 2.74 Aeropedellus clavatus 3.87 Phoetaliotes nebrascensis 2.82 Psoloessa delicatula 2.63 Melanoplus femurrubrum 3.87 Aeropedellus clavatus 2.26 Encoptolophus costalis 2.48 Trachyrachys kiowa 3.80 Melanoplus dawsoni 2.19 Phoetaliotes nebrascensis 1.62 Phoetaliotes nebrascensis 1.42 Orphulella speciosa 2.19 Melanoplus gladstoni 1.46 Melanoplus packardii 1.32 Trachyrachys kiowa 1.42 Melanoplus infantilis 1.44 Amphitornus coloradus 1.19 Philbostroma quadrimaculatum 1.40 used to assess differences in cumulative grasshopper m2 in 2002 to 9.0 per m2 in 2003 and remained at densities, species richness, Shannon diversity, and similarly low levels until the end of the study. How- SimpsonÕs evenness index. As year ϫ grazing, year ϫ ever, despite the overall low densities of grasshoppers, Þre, or year ϫ Þre ϫ grazing interactions were not signiÞcant Þre by grazing interactions and main graz- signiÞcant for any model, year appropriately treated as ing effects were detected on cumulative density, spe- a random variable and mixed model analyses were cies diversity, and community composition. Based on conducted using Systat 12. Because of the predicted a mixed model analysis, there was a signiÞcant Þre by amount of intersite variability resulting from the large grazing interaction effect on cumulative density (Ta- spatial extent of the study area (a 2,352-km2 region) ble 2). This resulted from burned SL pastures having and low replication levels, a signiÞcance level of 0.1 for higher densities than unburned SL pastures (P ϭ all analyses was deemed appropriate a priori. 0.03). Grasshopper densities dropped over the course Two methods were used to assess treatment differ- of the study, with an average 60% reduction in across ences in grasshopper community composition, using cu- all sites from 2002 to 2003. In this study, differences in mulative grasshopper sweep samples within a given year. average densities between treatments were typically First, nonmetric multidimensional scaling ordination less than one grasshopper per square meter (Fig. 1), (NMS) was used to assess differences in grasshopper which most likely obscured the ability to detect some community composition resulting from grazing and Þre grazing and Þre treatment effects, given the large treatments (McCune and Grace 2002). PC-ORD 5 (Mc- conÞdence intervals that result from few individuals Cune and Mefford 2006) was used for all NMS analyses. being detected in subsamples at a site (Onsager 1981). NMS was conducted using the Bray-Curtis distance mea- Branson (2005) found grasshopper populations de- sure with a random starting point and 70 runs of data. The clined strongly after a 1999 dormant season Þre in the dimensionality of the data set was assessed using the vicinity of the Sidney site, due largely to reductions in autopilot function in PC-ORD and indicated a three- densities of the Gomphocerine subfamily. Air tem- dimensional optimal solution. Second, a permutational peratures were low during the period of the pre- distance-based multivariate analysis of variance scribed burns in 2002, most likely leading to reduced (MANOVA) (PERMANOVA v1.6), using a Bray-Cur- tis (Sorensen) distance measure, was used to assess differences in community structure resulting from the Table 2. Results from mixed-model analyses of grazing and ﬁre treatment effects on cumulative density (fourth root transformed) treatments (Anderson 2001, 2005). Data were square and species diversity metrics root transformed, and species that occurred in fewer than 5% of samples were deleted before analysis (Mc- Effect df FP Cune and Grace 2002). Analyses were conducted Cumulative density Graze 2, 63 0.691 0.50 using permutation of raw data, because of small sam- Fire 1, 63 0.101 0.75 ple sizes (Anderson 2005). Because relatively few Graze ϫ Þre 2, 63 3.037 0.05 unique permutations were possible in analyzing pair- Species richness Graze 2, 65 2.949 0.06 wise comparisons, Monte Carlo asymptotic P values Fire 1, 65 1.374 0.24 Fire ϫ graze 2, 65 1.383 0.26 were used as a more conservative approach in all SimpsonÕs evenness Graze 2, 63 1.315 0.28 permutational analyses (Anderson 2005). Fire 1, 63 0.250 0.62 Fire ϫ graze 2, 63 0.492 0.61 Shannon diversity Graze 2, 65 3.644 0.03 Results and Discussion Fire 1, 65 0.209 0.65 Fire ϫ graze 2, 65 0.287 0.75 Grasshopper Density. Cumulative densities aver- aged across all sites dropped signiÞcantly from 22.6 per Year was treated as a random factor in the mixed-model analyses. 1444 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 5

Fig. 2. Scatter plot of cumulative grasshopper density over the course of 1 yr versus numerical species richness found at a given site. Fig. 1. Average grasshopper density (number/m2)by treatment from 2003 through 2005, after the October 2002 The results from nonmetric multidimensional scal- Þre. ing ordination indicate that differences in grasshopper community composition were greatest between sites Þre intensities and lowering both direct mortality and (Fig. 3), with no apparent visual patterns associated soil temperature elevations that cause egg mortality in with Þre or grazing treatments. Sites varied greatly in some grasshopper species (Branson and Vermeire terms of the relative species composition (Table 1), 2007). With respect to detecting grazing effects, On- but both Þre and grazing signiÞcantly affected grass- sager (2000) found grasshopper densities were much hopper community composition based on a permuta- lower under TOR grazing when compared with SL tional MANOVA using Bray-Curtis distances (Table grazing in the vicinity of the Sidney and Cartwright 3A). Although a signiÞcant Þre by grazing interaction sites during a study with much higher and increasing occurred, an examination of the grazing main effect grasshopper densities. It appears unlikely that either indicated differences in community composition ap- SL or TOR livestock grazing would have created treat- peared to exist between controls and both SL and TOR ment differences in food limitation and exploitative grazing (Table 3B). The Þnding of a signiÞcant Þre by competition, given the low grasshopper densities grazing interaction on community composition is sim- present during the study (Belovsky and Joern 1995, ilar to that found with cumulative densities, with dif- Joern 2000, Branson 2008). ferences in community composition between SL Species Diversity and Community Composition. burned and unburned plots (Table 3B). In addition, Although total sample sizes were often less than optimal for determining relative species compositions because of the low population densities (Berry et al. 2000), signiÞcant Þre by grazing interactions and main grazing effects were detected on Shannon diversity, species richness, and community composition. Based on mixed model analysis, livestock grazing appeared to affect numerical species richness (Table 2). Although more species were present in random catch samples in TOR pastures compared with ungrazed controls or SL pastures (P Ͻ 0.1), the percentage difference between treatments was Ͻ10%. Joern (2005) observed a strong positive relationship between grasshopper density and species richness in a tallgrass prairie site, but there was not a statistically signiÞcant relationship between cumulative density and species richness in this study (Fig. 2; r ϭ 0.13, P ϭ 0.26). Livestock grazing also signiÞcantly affected Shannon diversity (Table 2), with Shannon diversity higher in TOR pastures than in either SL pastures (P ϭ 0.02) or CO exclosures (P ϭ Fig. 3. NMS of grasshopper composition on each of the 0.06). Treatment differences in Shannon diversity val- ϭ ϭ Ͻ three blocks (sites) in the study (1 Grassy Butte; 2 ues were small, averaging 0.2, and perhaps an artifact Sidney; 3 ϭ Alexander). Site differences were apparent. The of the low overall grasshopper densities. There was no ordination is most stable with three dimensions and a Þnal signiÞcant effect of either Þre or grazing treatments on stress of 12.04 and Þnal instability of 0.00047 with 56 itera- SimpsonÕs evenness (Table 2). tions. October 2010 BRANSON AND SWORD:GRASSHOPPER RESPONSES TO FIRE AND GRAZING 1445

Table 3. (A) Results from a permutational multivariate analysis of variance examining treatment effects on grasshopper community composition; data were square root transformed, with analysis based on Bray-Curtis dissimilarities; (B) pairwise comparisons using Monte Carlo asymptotic P values

A Factor df SS MS FP Year 3 6,650.14 2,216.71 2.74 0.002 Graze 2 4,971.09 2,485.55 11.50 Ͻ0.001 Fire 1 2,792.08 2,792.08 5.87 0.004 Year ϫ graze 6 1,296.90 216.15 0.27 1.0 Year ϫ Þre 3 1,425.97 475.32 0.59 0.899 Graze ϫ Þre 2 2,236.60 1,118.30 5.32 Ͻ0.001 Year ϫ graze ϫ Þre 6 1,260.70 210.12 0.26 1.0 Residual 48 808.36 38,801.26 Total 71 59,434.75

B Effect Factor Comparison tP Graze EX vs SL 1.68 0.062 EX vs TOR 2.02 0.019 SL vs TOR 1.31 0.178 Fire ϫ graze Burn EX vs SL 2.18 0.016 EX vs TOR 1.89 0.026 SL vs TOR 1.45 0.104 Unburned EX vs SL 1.38 0.135 EX vs TOR 1.94 0.021 SL vs TOR 1.60 0.082 EX Burn vs unburned 1.40 0.129 SL Burn vs unburned 2.21 0.015 TOR Burn vs unburned 0.92 0.487

df, degrees of freedom; SS, sum of squares; MS, mean square; Ex, exclosure; TOR, twice over rotational grazing; SL, season long grazing. pairwise tests among levels of the factor graze within Conclusion. During this study period with low sites that were burned indicated a signiÞcant effect of grasshopper population densities, Shannon diver- grazing on community composition in sites with SL sity and grasshopper species richness were signiÞ- grazing, but not in pastures with TOR grazing or un- cantly higher with TOR livestock grazing, and a grazed controls (Table 3B). signiÞcant Þre by grazing interaction existed in cu- At the sites under consideration in this study, Clark mulative density and community composition. (2006) found no direct effect of burning on the com- However, the ability to draw strong conclusions position of grasses, forbs, or shrubs after Þre, which regarding the nature of species composition shifts was attributed to the low Þre intensity resulting from and population changes in the presence of Þre and a dormant season-prescribed burn. Grazing treat- grazing is complicated by the large site differences ments also did not affect vegetation composition or and low densities. Studies conducted at a smaller species diversity during the study (Clark 2006). As a spatial scale would beneÞt by reducing variation. result, these rangeland systems appear to be quite Low population densities during the study likely resilient to dormant season Þre and moderate grazing. obscured the ability to detect some grazing and Þre Without substantial changes in food availability or treatment effects, given the large conÞdence inter- habitat characteristics during a period with low grass- vals that result from few individuals being detected hopper densities, treatment effects on grasshoppers in density subsamples at a site (Onsager 1981) and could be weak. However, there was a sustained effect of SL grazing in plots that were burned. Both biomass low numbers of individuals in community compo- removal by cows and the percentage of bare ground sition sweep samples (Berry et al. 2000). Impor- were higher in SL grazed burned plots, when com- tantly, the grasshopper assemblages in this system pared with SL unburned plots. This led to a signiÞcant appear quite resilient to dormant season Þre and interaction between Þre and grazing on biomass and moderate levels of grazing intensity, during a period ground cover throughout the study (Clark 2006). Ver- with low grasshopper densities. Branson et al. meire et al. (2004b), in the southern plains, found that (2006) argued that habitat management techniques cattle grazing increased substantially in patch burn are predicted to have the largest differential impacts areas within larger pastures. As a result, the signiÞcant on grasshopper population dynamics when grass- interactions found between grazing and Þre on grasshopper populations are increasing in the region of hopper community composition and cumulative den- the study, given the cyclical nature of grasshopper sity, resulting from differences between burned and population dynamics. The results of this study re- unburned SL pastures, could be indicative of a mech- inforce the importance of long-term research and anistic link between habitat characteristics and grass- monitoring to examine the effects of habitat manip- hopper population densities. ulation on grasshopper population dynamics. 1446 ENVIRONMENTAL ENTOMOLOGY Vol. 39, no. 5

Acknowledgments Fielding, D. J., and M. A. Brusven. 1996. Livestock grazing and grasshoppers: an interregional perspective. Univer- We thank Donovan Craig for setting up the sites used in sity of Idaho College of Agriculture Bulletin 786: 1Ð12. this study and conducting much of the sampling. We also Fielding, D. J., M. A. Brusven, B. Shaﬁi, and W. J. Price. 2001. thank Mo OÕMara and Nicole Davidson for their help in Spatial heterogeneity of low-density populations of Mela- organizing and condensing the data. noplus sanguinipes (Orthoptera: Acrididae) associated with grazing and vegetation treatments. Can. Entomol. References Cited 133: 843Ð855. Joern, A. 2000. What are the consequences of non-linear Anderson, M. J. 2001. A new method for non-parametric ecological interactions for grasshopper control strate- multivariate analysis of variance. Austral Ecol. 26: 32Ð46. gies?, pp. 131Ð143. In J. A. Lockwood, A. V. Latchininsky, Anderson, M. J. 2005. PERMANOVA: a FORTRAN com- and M. G. Sergeev (eds.), Grasshoppers and Grassland puter program for permutational multivariate analysis of Health: Managing Grasshopper Outbreaks without Risk- variance. Department of Statistics, University of Auck- ing Environmental Disaster. Kluwer Academic, Boston, land, Auckland, New Zealand. MA. Belovsky, G. E., and A. Joern. 1995. The dominance of dif- Joern, A. 2004. Variation in grasshopper (Acrididae) den- ferent regulating factors for rangeland grasshoppers, pp. sities in response to Þre frequency and bison grazing in 359Ð386. In N. Cappuccino and P. Price (eds.), Popula- tallgrass prairie. Environ. Entomol. 33: 1617Ð1625. tion Dynamics: New Approaches and Synthesis. Aca- Joern, A. 2005. Disturbance by Þre frequency and bison demic, New York, NY. grazing modulate grasshopper assemblages in tallgrass Belovsky, G. E., and J. B. Slade. 1995. Dynamics of two prairie. Ecology 86: 861Ð873. Montana grasshopper populations: relationships among Magurran, A. E. 2004. Measuring biological diversity. Black- weather, food abundance and intraspeciÞc competition. well Publishing, Malden, MA. Oecologia 101: 383Ð396. McCune, B., and J. B. Grace. 2002. Analysis of ecological Belovsky, G. E., M. A. Brusven, D. J. Fielding, and L. Manske. communities. MJM Software Design, Gleneden Beach, 2000. Grasshopper habitat manipulation, pp. VII.15.1Ð OR. VII.15.5. In G. L. Cunningham and M. W. Sampson (eds.), McCune, B., and M. J. Mefford. 2006. PC-ORD: multivariate Grasshopper Integrated Pest Management User Hand- analysis of ecological data, version 5.21. MJM Software, book: APHIS Technical Bulletin 1809. United States Gleneden Beach, OR. Department of Agriculture, Animal and Plant Health In- Meyer, C. K., M. R. Whiles, and R. E. Charlton. 2002. Life spection Service, United States Department of Agricul- history, secondary production, and ecosystem signiÞ- ture, Washington, DC. cance of Acridid grasshoppers in annually burned and Bock, C. E., and J. H. Bock. 1991. Response of grasshoppers unburned tallgrass prairie. Amer. Entomol. 48: 52Ð61. (Orthoptera: Acrididae) to wildÞre in a southeastern O’Neill, K. M., B. E. Olson, M. G. Rolston, R. Wallander, D. P. Arizona (USA) grassland. Amer. Midl. Natural. 125: 162Ð Larson, and C. E. Seibert. 2003. Effects of livestock graz- 167. ing on rangeland grasshopper (Orthoptera: Acrididae) Branson, D. H. 2005. Effects of Þre on grasshopper assem- abundance. Agric. Ecosys. Environ. 97: 51Ð64. blages in a northern mixed-grass prairie. Environ. Ento- Onsager, J. A. 1981. A note on the Poisson distribution in mol. 34: 1109Ð1113. integrated pest management. Bull. Entomol. Soc. Am. 27: Branson, D. H. 2008. Inßuence of a large late summer pre- 119Ð120. cipitation event on food limitation and grasshopper pop- Onsager, J. A. 2000. Suppression of grasshoppers in the ulation dynamics in a northern Great Plains grassland. Great Plains through grazing management. J. Range Man- Environ. Entomol. 37: 686Ð695. age. 53: 592Ð602. Branson, D. H., and L. T. Vermeire. 2007. Grasshopper egg Plummer, J. 2005. Dormant season burning and grazing in- mortality mediated by oviposition tactics and Þre inten- teractions on mixed grass prairie. M.S. thesis, North Da- sity. Ecol. Entomol. 32: 128Ð134. kota State University, Fargo. Branson, D. H., A. Joern, and G. A. Sword. 2006. Sustainable Porter, E. E., and R. A. Redak. 1996. Short-term recovery of management of insect herbivores in grassland ecosys- the grasshopper communities (Orthoptera: Acrididae) of tems: new perspectives in grasshopper control. Bio- a California native grassland after prescribed burning. science 56: 743Ð755. Environ. Entomol. 25: 987Ð992. Butler, J. L., H. Goetz, and J. L. Richardson. 1986. Vegeta- Vermeire, L. T., R. B. Mitchell, S. D. Fuhlendorf, and D. B. tion and soil: landscape relationships in the North Dakota Wester. 2004a. Selective control of rangeland grasshop- badlands. Am. Midl. Nat. 116: 378Ð386. pers with prescribed Þre. J. Range Manage. 57: 29Ð33. Clark, L. M. 2006. Late-season burning and grazing inter- Vermeire, L. T., R. B. Mitchell, S. D. Fuhlendorf, and R. L. actions on mixed-grass prairie and woody draws. M.S. Gillen. 2004b. Patch burning effects on grazing distri- thesis, North Dakota State University, Fargo. bution. J. Range Manage. 57: 248Ð252. Engle, D. M., S. D. Fuhlendorf, A. Roper, and D. M. Leslie. 2008. Invertebrate community response to a shifting mo- saic of habitat. Range. Ecol. Manage. 61: 55Ð62. Received 29 December 2009; accepted 1 June 2010.