Am. Midl. Nat. 142:314-327

Habitat RelationshipsAmong (:)at the WesternLimit of the Great Plains in Colorado

DAVID P. CRAIG, CARL E. BOCK, BARRYC. BENNETT and JANE H. BOCK Departmentof Environmental, Population and OrganismicBiology, Campus Box 334, Universityof Colorado,Boulder 80309

ABSTRACT.-We measured grasshopperdensities on 66 shortgrass,mixed grass,tallgrass and hayfieldplots on Boulder,Colorado, open space in 1995-1996. Grasshoppersas a group, and mostspecies individually,were more abundant on relativelysparse shortand mixed grass plots than on lusherhayfields and tallgrassplots-a resultconsistent with the hypothesisthat most species require warm and sunny open ground for survivaland reproduction.Band- winged grasshoppers(Oedipodinae) were particularlyassociated with open ground in short- grassvegetation, whereas spur-throated grasshoppers () were mostabundant in mixed grass plots witha high percentage of forbvs. graminoidcover. Slant-facedgrasshop- pers () were the most uniformlydistributed among habitats,but were gen- erallyassociated with plots includingrelatively high proportionsof grassvs. forbcover. While 20 of the 25 most common grasshopperspecies on Boulder open space occurred in all four grasslandtypes, each habitatwas dominated by a distinctiveset of species. However,numer- ically dominant grasshopperson short, mixed and tallgrassplots only loosely resembled groups of dominantspecies in the same habitatselsewhere on the Great Plains. While most Great Plains grasshoppersare widelydistributed, they are apparentlynumerically responsive to combinationsof environmentalconditions expressed at local scales. Common species of grasshopperson Boulder open space in 1995-1996 were the same as those collected in the region in the 1950s. Our resultssuggest these protected grasslands,although invaded by alien vegetationand fragmentedby suburbanization,are stilleffectively conserving this par- ticularinsect group.

INTRODUCTION Insectsin general,and grasshoppersin particular,are dominantabove ground herbivores in most grassland ecosystemsbut much remains to be learned about the structureand functionof their communities (Otte, 1981; Curry,1994; Tscharntkeand Greiler,1995). Most grasshoppersof the NorthAmerican Great Plains are widelydistributed, with ranges thatencompass eastern tallgrass,western shortgrass and northernmixed grassprairies that comprise the naturalvegetation across broad expanses of this region (Daubenmire, 1978; Otte, 1981, 1984; Sims, 1988; Pfadt,1994). Local and regional studies suggestthat many Great Plains grasshoppershave distincthabitat requirements,despite their large ranges (Joern, 1982; Capinera and Thompson, 1987; Evans, 1988; Kemp et al., 1990; Quinn and Walgenbach, 1990; Welch et al., 1991; Kemp, 1992). However,we are not aware of any investigationattempting a quantitativecomparison of grasshopperdensities among short, mixed and tallgrasshabitats. One difficultyis the notoriousand apparentlyclimate-related volatilityof manygrasshopper populations (Capinera and Horton, 1989; Joernand Gaines, 1990; Kemp, 1992). These fluctuationscan make it difficultto distinguishfundamental habitatdifferences from temporal variations in widelyseparated sites sampled in different years (but seeJoern and Pruess, 1986). The arid plains of eastern Colorado are, or were, dominated by relativelyuniform short- grassprairie. However, fine-scale mosaics of short,mixed and tallgrasshabitats occur where 314 1999 CRAIG ET AL.: HABITAT 315 the plains meet the eastern edge of the Colorado RockyMountains, apparently the result of increased precipitation,topographic heterogeneity and differencesin soils and land-use history(Livingston, 1952; Branson et al., 1965; Moir, 1969; Bennett, 1997). This circum- stancepermitted a comparisonof grasshoppersoccupying the threefundamentally different sortsof Great Plains grasslandsat a local scale where theirregional populations would be experiencingsimilar patterns of temperatureand precipitation,and presumablywhere in- dividualscould move among the available habitats. Grasslandsat the base of the RockyMountains in Colorado (the so-called Front Range Corridor) are rapidlybeing fragmentedand lost to development(Mutel and Emerick,1992; Long, 1997). Some of the largestremaining stands are part of municipalopen space owned and managed by the Cityof Boulder, Colorado. In 1995-1996 we quantifieddensities of grasshopperson 66, 200-mdiameter plots on Boulder open space, placed systematicallyto include stands of short,mixed and tallgrassvegetation, as well as irrigatedhayfields which have replaced manynative grasslands within the past 150 y.We then compared grasshopper populations among the habitattypes and correlatedtheir densities with vegetative attributes of the plots. Our objectiveswere to: (1) reveal individualspecies' habitatpreferences, (2) determineif the differentgrassland plant communitiessupported distinctivegrasshopper communitiesand (3) compare our resultswith those fromother studies across the Great Plains.

Cityc m elevation, around the northern,eastern and southern citylimits, all against a Rocky Mountain foothillsbackdrop of ponderosa pine (Pinus ponderosa)savanna. Plains habitats include riparian corridorsalong streams,tallgrass vegetation and agriculturalhayfields in adjacent lowland floodplainsand mixed and shortgrassstands on upland slopes and bench- es (Bock et al., 1995). Soils are principallyclay or sandy loams (Moreland and Moreland, 1975). Averageannual precipitation(47 cm) is sufficientto supportmixed grassvegetation on all of the uplands. Shortgrassoccurs only in comparativelyheavily grazed areas. We established66,200-m diameter circular plots on Boulder open space grasslandsin the fall of 1993. Preliminaryfloristic surveys were conducted in 1994. Each plot was categorized in its predominantgrassland type based on extensiveplant frequencydata collected inJuly of 1995 and 1996 (Bennett, 1997). Botanical nomenclaturewas taken from Great Plains Flora Association (1986), whereas nativevs. exotic statusof each plant species was deter- mined fromWeber (1995). Types and numbersof plots,and theirpredominant plant spe- cies, were as follows: Shortgrassvegetation (13 plots).-These upland plots had a recent or currenthistory of grazingby cattle,horses and/or black-tailedprairie dogs (Cynomysludovicianus). The most abundant native grasses on shortgrassplots were Agropyronsmithii, Bouteloua gracilis and BuchloI dactyloides.Common native forbswere Artemisiafrigida, Plantago patagonica and Sphaeralceacoccinea. Two common exotic annual grasseswere Bromustectorum and B. Ia- ponicus,while the mostabundant exoticforbs were Alyssum minus and Convolvulusarvensis. Mixed grass vegetation(21 plots).-These upland plots were ungrazed, or lightly-grazed, and were dominated by varyingmixtures of midheightand short grasses and sedges, de- pending on soils and topography.Common nativegraminoids included Agropyronsmithii, Andropogongerardii, A. scoparius,Aristida purpurea, Bouteloua gracilis, B. curtipendula,Car- ex heliophila,Koelaria pyramidata, Panicum virgatum,Stipa comataand S. neomexicana.The mostabundant nativeforbs were Ambrosiapsilostachya, Artemisia ludoviciana, Asterfalcatus, 316 THE AMERICAN MIDLAND NATURALIST 142(2)

Helianthuspumilis and Psoralea tenuiflora.Common exotics on mixed grass plots included Alyssumminus, Bromus japonicus, Poa pratensis,P compressaand Tragapogondubius. Tallgrassvegetation (11 plots).-These plots occurred in lowland floodplains or on low benches with cobbly soils and high water tables. Some were winter-grazedby cattle,but all remained undisturbedthrough the summergrowing season. They were dominated by mid- height and tall grassescharacteristic of the True Prairie of the eastern Great Plains (Risser et al., 1981):' Andropogongerardii, A. scoparius,Panicum virgatum,Sorghastrum nutans and Sporobolusasper. The most abundant native forbs on tallgrassplots were Ambrosiapsilos- tachya,Aster falcatus, Dicantheliumoligosanthes, Grindelia squarosa and Prunella vulgaris. There were many exotic species on tallgrassplots, including the forbs Cichoriumintybus, Plantago lanceolata, Taraxacum officinale,Trifolium pratense and a varietyof cool-season meadow grasses planted for pasture improvement(Agrostis stolonifera, Festuca pratensis, Phleumpratense and Poa pratensis). Hayfields (21 plots).-These lowland plots were flood-irrigatedin early summer and mowed between mid-Julyand earlyAugust. Most were also winter-grazedby cattle. Some included remnants of the tallgrassvegetation they replaced, but all were dominated by exotic species such as alfalfa (Medicago sativa) and a series of cool-season pasture grasses, including Bromusinermis, Festuca pratensis, Dactylis glomerata and Phleumpratense.

METHODS Vegetationcover.-Vegetation canopy cover was sampled on each of the 66 plots in June and again in August of 1995 and 1996. During each samplingperiod 100 m transectswere establishedrunning due east and due westfrom each plot center.An optical point projector was set at 4 m intervalsalong each transectand vegetation was sampled at six random points within1 m of each setting.This resultedin 300 points per plot per samplingperiod (n = 1200 total points per plot over the 2 y). Each point was classifiedas unvegetatedor as covered by one or more of the followingfour vegetationcategories: native graminoids, exotic graminoids,native forbs and/or exotic forbs.Results from the foursampling periods were averaged to generate a single set of vegetation cover descriptorsfor each plot to compare withpooled grasshopperdata collected over the same period. Grasshopperdensities.-We collected sweep net samples of grasshoppersin each plot in the summerof 1994 to build a referencecollection. Grasshopper densitieswere then esti- mated on each plot four timesper summer,from late May to earlySeptember, in both 1995 and 1996. During each samplingperiod, 10, 0.5 m2wire hoops were distributedalong each of four 100 m radii followingcardinal compass directionsfrom the center of each plot (n = 40 hoops per plot). Hoops were undisturbed for three days. Then a single observer (alwaysthe firstauthor) approached each hoop and identifiedadult grasshoppersas they were flushedout, followingmethods in Thompson (1987). Althoughthe hoop method may be best-suitedto habitats with low and relativelysparse vegetation,we found that it was possible to confidentlyidentify grasshoppers in hoops on tallgrassand hayfieldplots, and made a special effortto flush all individualsfrom the heavier cover in these habitats.It should be noted that none of our plots on Boulder open space were in vegetationas tall as thatof the tallgrassprairie in eastern Kansas where the hoop method proved impractical (Evans et al., 1983). Because habitat associations of grasshopperswere consistentbetween years (see results), our eight 40-hoop samples were pooled to generate one estimate of average densityper hoop for each species on each plot over the 1995-1996 sampling period. These estimates were convertedto densities/100m2 to facilitatecomparisons with certain other studies (e.g., Capinera and Thompson, 1987; Fowler et al., 1991). 1999 CRAIG ET AL.: GRASSHOPPER HABITAT 317

Statistics.-Data on combined densities of three acridid subfamilies,total grasshoppers and attributesof vegetationcover met the criteriafor parametricanalyses. These results were compared among habitatsusing onewayanalysis of variance withScheffe a-posterior7i pairwise comparisons of means. Data for many individual grasshopper species failed to meet criteriafor parametricanalyses because theywere missingfrom one or more habi- tats. Therefore,we compared their ranked densities among habitats using the Kruskal- Wallis H statisticwith a-posteriori protected rank-sum tests (Welkowitz et al., 1991) between each pair of habitats. Overall faunal similaritiesamong habitat typeswere estimatedby calculatingboth Renkonen percentage similarities(Krebs, 1989) and Spearman rank cor- relations between pairs of habitatsbased on their grasshopper population densities. Fi- nally,we computed Spearman rank correlationsbetween grasshopperdensities and three vegetation cover variables important to grasshoppers in previous studies (Joern and Gaines, 1990): (1) the amount of unvegetated ground, (2) the relative canopy cover of forbsversus grasses and (3) the relative canopy cover of native vs. exotic species (each with Bonferroni's adjustment for joint significanceof the multiple comparisons; three testsper species, criticalP = 0.17).

RESULTS Comparisonsbetween years.-We collected 48 species of Acrididae on Boulder open space in 1995-1996 (Appendix). All but twouncommon species (Aulocarafemoratum and Melan- oplus differentialis)were counted in both years.While most species were more abundant in 1996 than in 1995, theirhabitat associations and relativeabundances were similarbetween years.For example, the Spearman rank correlationbetween densitiesof the 48 species in 1995 vs. 1996 forall 66 plots combined was 0.898 (n = 48, P < 0.0001), indicatingthat the same species clearlydominated the grasshopperfauna in both years.Across the 66 plots, Spearman rank correlationsbetween 1995 and 1996 countsfor the threeacridid subfamilies were as follows(n = 66, P < 0.0001 in all cases): Oedipodinae, 0.871; Melanoplinae, 0.609; Gomphocerinae, 0.737. Based on these resultswe combined the two years' data for our subsequent analysesof grasshopperfaunas and habitatassociations. Grasslandfaunas.-Grasshopper species richnessaveraged higher in shortgrass(20.3 spe- cies) and mixed grassplots (18.4 species) than in eithertallgrass (12.1 species) or hayfields (8.4 species; Fig. 1). An average of 84.1 adult grasshoppers/i00m2occurred on shortgrassplots (Fig. 1). Eight species comprisedover 70% of these individuals: elliotti, kiowa, Me- lanoplus sanguinipes, simplex, obscura,Phoetaliotes nebrascensis, Ageneotettix deorumand Arphia conspersa.Of these,only the firsttwo were significantlymore abundant in shortgrassthan in any other habitat,and no species was restrictedto it (Table 1). Grasshoppersin the subfamilyMelanoplinae dominatedmixed grassplots (Fig. 1), where seven species (six melanoplines) made up over 70% of the average 122.1 individuals/lOOm2: keeleri,M. packardii,M. sanguinipes,M. femurrubrum,Phoetaliotes nebrascensis, Hypochloraalba and Eritettixsimplex. Only the firsttwo Melanopluswere significantlymore abundant in thishabitat than in any other (Table 1). Tallgrasssupported an average of 54.1 grasshoppers/100m2,70% of which were distrib- uted among fivespecies: Melanoplusfemurrubrum, M. bivittatus,Phoetaliotes nebrascensis, Eritettixsimplex and Aulocara elliotti.No grasshopperspecies was significantlymore abun- dant in tallgrassthan in any otherhabitat. However, counts of M. bivittatusand Chortophaga viridifasciatawere highestin tallgrass(Table 1). Three species made up nearly 75% of the average 41.7 individuals/lOOm2counted on hayfieldplots: Melanoplus bivittatus,M. femurrubrumand Chorthippuscurtipennis. No spe- 318 THE A-RcRicAN MIDLAND NATURALIST 142(2)

a a SPECIES RICHNESS b E SHORTGRASS a El MIXEDGRASS OEDIPODINAE b I TALLGRASS C U HAYFIELD C

a

a%////// ab MELANOPLINAE ab

a ALL GRASSHOPPERS a _ lb~~~~~~

0 25 50 75 100 125 150

GRASSHOPPERDENSITY/100m2 or SPECIES RICHNESS

FIG. 1.-Mean and standarderror of species richnessand densitiesof grasshopperson 13 shortgrass, 21 mixed grass, 11 tallgrassand 21 hayfieldplots on Boulder, Colorado, open space in 1995-1996. Values sharingany letterwere not significantlydifferent in Scheffea-posteriori pairwise comparisons of means. F-scoresand P-valuesfor oneway analyses of variance were as follows:species richness,F = 26.37, P < 0.0001; Oedipodinae, F = 22.08, P < 0.001; Melanoplinae, F = 4.93, P = 0.004; Gompho- cerinae, F = 9.96, P < 0.0001; all grasshoppers,F = 7.25, P = 0.0003

cies was statisticallymost abundant in hayfields, although we counted highest numbers of the last two of these three species on hayfield plots (Table 1). In terms of shared species densities, the faunas of shortgrass and mixed grass plots were most similar to one another (60.1%), and significantlypositively correlated (Table 2). Tall- grass and hayfields were 78.1% similar, and were also significantly positively correlated. Faunal similarities and rank correlations consistently declined across the habitat gradient from shortgrass to mixed grass to tallgrass to hayfield (Table 2). Habitat associations of species and subfamilies-Bandwinged grasshoppers (subfamily Oed- ipodinae) were most abundant as a group in shortgrass habitats and very uncommon in tallgrass or hayfield (Fig. 1). Trachyrhachyskiowa was the most abundant oedipodine in our study area, and a shortgrass specialist (Table 1). Chortophaga viridifasciata was most abun- dant in tallgrass. The other five common oedipodines were relatively abundant in short and mixed grass habitats, and rare or absent from tallgrass and hayfields. Spur-throated grasshoppers (subfamily Melanoplinae) were more abundant as a group on mixed grass plots than elsewhere in the study area, and least common in hayfields (Fig. 1999 CRAIG ET AL.: GRASSHOPPER HABITAT 319

TABLE 1.-Mean density/lOOm2(SE) of the 25 most common grasshopperson 13 shortgrass,21 mixed grass,11 tallgrassand 21 hayfieldplots on Boulder, Colorado, open space in 1995 and 1996

Kruskal- Species Shortgrass Mixedgrass Tallgrass Hayfield WallisH SubfamilyOedipodinae Arphia conspersa 3.6 (0.9)a' 3.0 (0.4)a 0.5 (0.2)b 0.6 (0.2)b 33.18*** Arphiapseudonietana 2.6 (0.6) a 2.4 (0.5) a 0.1 (0.1)b 0.2 (0.1)b 31.33*** Chortophagaviridifasciata 0.2 (0.1) ab 0.1 (0.1) a 0.9 (0.4)b 0.4 (0.1) ab 9.26* Dissosteiracarolina 1.6 (0.4)a 1.6 (0.9)ab 0.1 (0.1)c 0.5 (0.2)bc 12.22** Hadrotettixtrifasciata 1.0 (0.4)a 0.7 (0.2)a 0.1 (0.1)b 0.1 (0.1)b 17.41*** Spharagemonequale 1.4 (0.5)a 0.7 (0.2)a 0.2 (0.2)b 0 b 28.92*** Trachyrhachyskiowa 6.8 (1.6)a 1.4 (0.4)b 0.2 (0.1)c 0.2 (0.1)c 29.86*** SubfamilyMelanoplinae Hypochloraalba 1.8 (I.1)a 9.2 (2.8)a 0.2 (0.2)b <0.1 (<0.1)b 25.02*** Melanoplus bivittatus 0.7 (0.3)a 2.3 (0.7)a 6.4 (1.7)b 4.9 (0.7)b 24.41*** Melanoplus confusus 0.6 (0.2)a 3.5 (1.0)a 0.1 (0.1)b 0.3 (0.2)b 19.71*** Melanoplusfemurrubrum 1.6 (0.6)a 14.6 (4.3)b 20.0 (2.5)c 22.5 (2.7)c 25.84*** Melanopluskeeleri 1.5 (0.8)a 17.1 (4.8)b 0.4 (0.3)a 0 c 30.64*** Melanoplus lakinus 0.3 (0.1)a 1.1 (0.8)a 0 b 0 b 13.93** Melanopluspackardii 2.7 (I.1)a 13.0 (4.0)b 0.3 (0.3)c 0.2 (0.1)c 30.81*** Melanoplussanguinipes 18.4 (3.7)a 19.1 (5.7)a 1.9 (0.9)b 0.6 (0.2)b 46.48*** Phoetaliotesnebrascensis 5.1 (1.4)a 9.5 (3.6)a 3.6 (1.2)ab 1.3 (0.6)b 7.84* SubfamilyGomphocerinae Aeropedellusclavatus 0.3 (0.2)a 1.1 (0.3)bc 0.8 (0.3)ac 0.5 (0.2)ac 6.06 ns Ageneotettixdeorum 3.6 (I.1)a 3.3 (0.7)a 2.6 (0.9)ab 0.9 (0.4)b 12.68** Amphitornuscoloradus 2.1 (0.5)ab 1.8 (0.4)a 0.8 (0.6)b 0.1 (0.1)c 26.38*** Aulocara elliotti 7.4 (I.1)a 3.0 (0.5)b 3.0 (1.6)bc 1.7 (0.7)c 20.49*** Chorthippuscurtipennis 0 a 0.5 (0.3)a 2.7 (0.5)b 3.7 (1.3)b 25.12*** Eritettixsimplex 9.6 (I.9)a 6.9 (I.1)a 5.1 (0.8)a 2.2 (0.6)b 18.34*** Mermiriabivittata 0.3 (0.2) 1.0 (0.3) 0.7 (0.5) 0.3 (0.2) 1.53 ns 5.1 (2.6)a 0.9 (0. 3)a 1.7 (0.5)a 0 b 17.52*** Orphulellaspeciosa 1.6 (0.8) 1.0 (0.3) 0.3 (0.2) 0.3 (0.2) 2.32 ns 1 Within-speciesacross habitats,values sharingany letternot significantlydifferent in aposterioripair- wise rank-sumtests (Welkowitz et al., 1991) * P < 0.05, ** P < 0.01, *** P < 0.001, ns = not significant,Kruskal-Wallis H tests

TABLE 2.-Percent similarities(above diagonal) and Spearman rank correlations(below diagonal) betweenshortgrass, mixed grass,tallgrass and hayfieldhabitats on Boulder,Colorado, open space, based on densitiesof 48 grasshopperspecies in 1995-1996

Shortgrass Mixedgrass Tallgrass Hayfield Shortgrass --- 60.1 36.1 19.9 Mixed grass 0.586* 45.8 33.7 Tallgrass 0.253 0.255 78.1 Hayfield 0.121 0.224 0.726*

* P < 0.01, Spearman rank correlation 320 THE AMERICAN MIDLAND NATURALIST 142(2)

TABLE 3.-Mean (and standard error) of percent unvegetatedground, percent of total graminoid and forbcanopy thatwas exotic species and percentof totalgraminoid and forbcanopy thatwas native or exotic forbson 13 shortgrass,21 mixed grass,11 tallgrassand 21 hayfieldplots on Boulder,Colorado, open space, in 1995-1996

Percent Percent Percent Grassland type bare ground exotic forbs Shortgrass 54.8 (2.2) a' 44.3 (3.6) a 41.6 (3.3) a Mixed grass 56.6 (1.7) a 29.2 (4.3) a 42.8 (4.5) a Tallgrass 35.2 (3.3) b 44.7 (4.6) a 24.0 (2.5) b Hayfield 31.8 (2.8) b 84.9 (3.5) b 30.4 (3.8) ab F (P) 26.7 (<0.0001) 41.8 (<0.0001) 4.4 (0.007) 'Values sharingany letternot significantlydifferent in Scheffea-posteriori pairwise comparisons

1). Seven of the nine most abundant melanopline species followed this general pattern (Table 1, see especiallyMelanoplus keeleri,and M. packardii), or were more abundant in mixed and shortgrassthan in tallgrassor hayfield(especially Melanoplus sanguinipes). Con- spicuous exceptionswere twospecies thatwere relativelyabundant in tallgrassand hayfields: Melanoplus bivittatusand M. ftmurrubrum. Slant-facedgrasshoppers (subfamily Gomphocerinae) were the most evenlydistributed among habitats,but were more common in shortgrassthan in hayfields(Fig. 1). A pattern shared by six of the nine most abundant gomphocerineswas highestdensities in shortgrass, usuallyfollowed in decreasing order by mixed grass,tallgrass and hayfields(e.g., ,Ageneotettix deorum, Aulocara elliotti),though not all these differenceswere statis- ticallysignificant (Table 1). Chorthippuscurtipennis was an exception to thispattern, being more abundant in hayfieldand tallgrasshabitats. Correlationswith cover variables.-In addition to floristicdifferences, the four grassland habitatsdiffered in termsof three cover attributes(Table 3). First,short and mixed grass plots had significantlymore bare ground than either tallgrassor hayfieldplots. Second, hayfieldshad a much higherpercentage of exoticvegetation canopy than the otherhabitats. Hayfieldcanopy consistedprimarily of alfalfaand exotic pasture grasses.Finally, vegetative canopy on shortand mixed grassplots consistedof a higherpercentage of forbs(vs. grasses) than tallgrass,with hayfields intermediate. Across all 66 plots total adult grasshopper densitywas negativelycorrelated with per- centage of vegetativecanopy that consisted of exotic species (Spearman r = -0.451, P < 0.001). However,this relationshipwas due entirelyto low densitieson hayfieldplots; the correlationdisappeared when hayfieldplots were eliminatedfrom the analysis. We computed Spearman rank correlationsbetween grasshopper density, % bare ground, % exotic vegetationcover and % forbcover, restricting the analysisto the 45 short,mixed and tallgrassplots (Table 4). There were no significantcorrelations between any species or subfamilyand percentage of exotic vegetation.However, there were significantdifferences among species in theirSpearman rank correlationswith amounts of bare ground and with percentagesof vegetativecanopy consistingof forbsvs. grasses. Oedipodinae as a group, and three of the common species individually,were positively correlatedwith the percentage of unvegetatedground, a conspicuous exception being the tallgrassspecies Chortophagaviridifasciata. Total Melanoplinae was stronglypositively asso- ciated with % forb cover,and no individualmelanoplines were significantlynegatively as- sociated withthis variable. Densities of the Gomphocerinae as a group, and of mostspecies 1999 CRAIG ET AL.: GRASSHOPPER HABITAT 321

TABLE 4.-Spearman rank correlationsbetween densitiesof the 25 most common grasshoppersand three vegetationmeasures on 45 short,mixed, and tallgrassplots on Boulder, Colorado, open space in 1995 and 1996

Percent Percent Percent Taxon bare ground' exotic2 forbs3 SubfamilyOedipodinae Arphia conspersa 0.17 -0.21 0.54*** Arphiapseudonietana 0.43* -0.10 0.21 Chortophagaviridifasciata -0.31 0.17 -0.24 Dissosteiracarolina 0.24 0.14 0.27 Hadrotettixtrifasciata 0.40* 0.05 0.08 Spharagemonequale 0.32 0.02 0.31 Trachyrhachyskiowa 0.46** 0.23 0.19 Total subfamily 0.49** 0.02 0.39* SubfamilyMelanoplinae Hypochloraalba 0.22 -0.27 0.43* Melanoplus bivittatus -0.38* 0.05 -0.05 Melanoplus confusus 0.32 -0.22 0.20 Melanoplusftmurrubrum -0.28 0.16 -0.19 Melanopluskeeleri 0.29 -0.24 0.15 Melanoplus lakinus 0.13 0.20 0.59*** Melanopluspackardii 0.37* -0.21 0.42* Melanoplussanguinipes 0.42* 0.01 0.58*** Phoetaliotesnebrascensis 0.09 0.24 0.26 Total subfamily 0.17 0.01 0.55*** SubfamilyGomphocerinae Aeropedellusclavatus -0.09 -0.16 -0.39* Ageneotettixdeorum -0.03 -0.09 0.03 Amphitornuscoloradus 0.21 0.02 0.11 Aulocara elliotti -0.19 0.19 0.02 Chorthippuscurtipennis -0.55*** 0.13 -0.43* Eritettixsimplex 0.12 -0.16 0.07 Mermiriabivittata 0.20 0.02 -0.41* Opeia obscura 0.11 0.13 -0.09 Orphulellaspeciosa 0.25 -0.29 0.22 Total subfamily 0.22 -0.04 -0.10 1 Percentageof the ground withoutvegetation canopy 2 Percentageof total graminoidand forbcanopy thatwas comprisedof exotic species 3 Percentageof total graminoidand forbcanopy thatwas comprisedof native or exotic forbs * P < 0.05 (P < 0.017 before Bonferroniadjustment for multiplecomparisons within species) ** P < 0.01 (P < 0.003 before Bonferroniadjustment) * P < 0.001 (P < 0.0003 before Bonferroniadjustment)

individually,were not significantlyassociated withvegetative attributes of the plots (Table 4). Exceptionswere Mermiriabivittata and Aeropedellusclavatus, whose densitieswere great- est on plots withrelatively high proportionsof grassvs. forb cover,and Chorthippuscurti- pennis,a tallgrassand hayfieldspecies most abundant on plots withlittle bare ground and much grass cover. 322 THE AMERICAN MIDLAND NATURALIST 142(2)

DISCUSSION Grasshoppercommunities in eastern Montana, eastern Colorado and the Nebraska san- dhillswere each dominatedby differentsets of species despite havingmost taxa in common (Kemp, 1992). Kemp attributedthese patternsto differencesin individualspecies' resource requirementsrather than to the existence of distinctgrasshopper communities predictably associated with differentsorts of grasslands.Results of the present studystrongly support thisconclusion at a local scale, for reasons described below. Among the most abundant grasshoppers,20 of 25 occurred in all fourhabitats, while an additional three occurred in all grasslandsexcept hayfields(Table 1). This resultis consis- tent with the wide ranges of most Great Plains grasshoppers(Otte, 1981, 1984; Capinera and Sechrist,1982a; Pfadt,1994) and does not support the recognitionof habitat-specific grasshopperfaunas. Nevertheless,short, mixed and tallgrassplots were numericallydomi- nated by differentsets of species on Boulder open space. One might argue that these representedcharacteristic grasshopper communitiesif those same species dominated in short,mixed and tallgrassprairie regions elsewhereacross the Great Plains. However,that appears not to be the case, as illustratedby the followingexamples. Capinera and Sechrist(1982b) sampled grasshoppersin shortgrassprairie at the Central Plains ExperimentalRange (CPER) in northeasternColorado in 1980-1981. The fivedom- inant species were Melanoplusgladstoni, Opeia obscura,Trachyrhachys aspersa, Arphia pseu- donietanaand Phlibostromaquadrimaculatum. Welch et al. (1991) sampled shortgrasshabi- tats elsewhere at CPER in 1989, and found four species to be most abundant: Eritettix simplex,Phlibostroma quadrimaculatum, Psoloessa deliculataand Trachyrhachysaspersa. On Boulder open space in 1995-1996 the fivemost abundant species on shortgrassplots were Melanoplus sanguinipes,Eritettix simplex, Aulocara elliotti, and Opeia obscura(Table 1), a list that includes only one dominant species each fromthe two CPER studies (which themselvesshared only two dominants). Evans (1984, 1988, 1989) conducted extensivestudies of grasshoppersin tallgrasshabitats on the Konza Prairie in eastern Kansas. The most common species was the grass-feeding melanopline,Phoetaliotes nebrascensis. While thisspecies was the fourthmost abundant spe- cies on our tallgrassplots, it was more common on Boulder open space in shortand mixed grass habitats(Table 1), and not correlatedwith grass vs. forbcover across individualplots (Table 4). This was an unexpected result,although Mulkern et al. (1964) found P nebras- censisassociated withmixed forb-grassand earlysuccessional grasslandsin North Dakota. Other common grasshoppersin Konza Prairietallgrass were Orphulellaspeciosa, Hypochlora alba, Melanoplus scudderi,M. keeleri,M. femurrubrumand M. bivitattus.Only the last two of these were among the dominantspecies on Boulder tallgrassplots (Table 1). As in the preceding two examples, grasshoppercommunities on Boulder mixed grass plots were onlygenerally similar to those of mixed grassprairie regions on the GreatPlains (Joern, 1986; Joern and Pruess, 1986; Kemp et al., 1990; Quinn and Walgenbach, 1990; Fowler et al., 1991; Kemp, 1992). One factormay be the lack of sandy soils on Boulder open space (Moreland and Moreland, 1975). This probablyexplains the absence of species such as Melanoplus angustipennisthat are particularlycommon in sand prairies (Joern, 1986). The Boulder Valleyis outside, or at the extrememargins, of the ranges of species thatare typicalelsewhere in mixed grassprairies across the Great Plains (Otte, 1981, 1984; Pfadt,1994), such as Melanoplusinfantilis in NorthDakota (Fowleret al., 1991) or Camnula pellucida in Montana (Kemp, 1992). There are numerous other cases where we cannot explain the rarityor apparent absence of a grasshopperspecies fromour mixed grassstudy plots on Boulder open space whichwere dominated by various species of Melanoplus (e.g., 1999 CRAIG ET AL.: GRASSHOPPER HABITAT 323

M. sanguinipes,M. keeleri)that are knownto preferdisturbed ground rich in forbs(Table 1; Pfadt,1994). Despite the absence of distinctgrasshopper communities associated with shortgrass, mixed grass and tallgrassvegetation, density differences among Boulder habitatsand cor- relationswith grassland characteristics on our individualplots are similarto what is known about habitatpreferences of acridid species and subfamilieselsewhere. Our estimatesof densityand species richnessof grasshoppersin the Boulder Valleywere greaterin relativelyopen and low-statureshort and mixed grassplots thanin denser tallgrass or hayfields(Fig. 1). Assuming these results are not an artifactof conspicuousness (see methods), theysupport the hypothesisthat most grasshopperspecies require dryand warm open ground for both survivaland reproduction(Capinera and Horton, 1989; Joern and Gaines, 1990; Coxwell and Bock, 1995). Only fourof 48 species (Chortophagaviridifasciata, Chorthippuscurtipennis, Melanoplus bivittatusand M. femurrubrum)were associated with tallgrassand hayfields.Each of these species has been characterizedas preferringrelatively lush and moist grasslandsand meadows (Otte 1981, 1984; Capinera and Sechrist,1982a; Pfadt,1994). In The Netherlandsvan Wingerden et al. (1992) found grasshopperspecies richnessand abundance declined in response to fertilizationof hayfields,which theyattri- buted to the resultingincreases in dense grass cover. Habitat associationsof the threemajor acridid subfamiliesin the Boulder Valleyconform to resultsof other studies.Band-winged grasshoppers (Oedipodinae) are known to prefer areas of relativelylow and sparse vegetation(Capinera and Sechrist,1982a; Otte, 1984), as theyclearly did in our studyarea (Fig. 1, Table 4). Most spur-throatedgrasshoppers (Me- lanoplinae) are forb-feeders,frequently associated withsites having relatively high propor- tions of forbvs. graminoidvegetation (Joern and Pruess, 1986; Evans, 1988; see Table 4). On Boulder open space melanoplines were most abundant and the dominant group on mixed grass plots, as theywere in ungrazed mixed grass prairiesin South Dakota (Quinn and Walgenbach,1990). Slant-facedgrasshoppers (Gomphocerinae) feed mainlyon grasses and sedges (Otte, 1981) and were less stronglyassociated withforb-dominated plots than were melanoplines in our studyarea. Ranked relativeabundances of grasshopperspecies were statisticallysimilar over multiple years on two Nebraska prairies,despite wide fluctuationsof absolute densities (Joern and Pruess, 1986), as was the case in the present study.If this relationshipis generallytrue, differencesin grasshoppercommunities on broader habitat areas across the Great Plains cannot be attributedsolely to the years when theywere sampled. Various studies suggest thatlocal grasshoppercommunities have predictablestructure based on diet, microhabitat selection, interspecificcompetition and predation (Joern and Lawlor, 1981; Joern, 1982, 1992; Joern and Klucas, 1993; Belovskyand Slade, 1993, 1995). The question remains:over whatgeographic range do such patternshold? Resultsof the presentstudy and othersacross the Great Plains suggestthat, while most species are widelydistributed in the region,grass- hopper communitiesassociated with particularsorts of grasslandsmay be dominated by predictablesets of species only at a local scale. Alexander and Hilliard (1969) collected grasshoppersin upland grasslandsnear Boulder between 1949 and 1960, when the whole of the Boulder Valleywas much less suburbanized than it is today (Mutel and Emerick, 1992; Long, 1997). Grasshopperdensities were not quantified,but species were categorizedas "resident"vs. "accidental," apparentlybased on frequencyof collection.Although Alexander and Hilliard collected ten species not detected during our study,all were listed as accidental. Furthermore,each of the 25 most common species found in the presentstudy was listedas residentin Boulder grasslandsby Alexander and Hilliard. These resultssuggest that the grasshopperfauna of the Boulder Valley has 324 THE AMERICAN MIDLAND NATURALIST 142(2) been relativelystable over the past half-century.Today Boulder open space grasslandsare significantlyinvaded by alien vegetationand fragmentedby suburban development.These have had negativeconsequences forcertain native plants, songbirds and birdsof prey(Berry et al., 1998; Bock and Bock, 1998; Bock et al., in press), but apparentlynot forgrasshoppers.

Acknowledgments.-Thisstudy was supportedby the Cityof Boulder Open Space Departmentand by the Biological Resources Division of the U. S. Geological Survey.We thank the followingindividuals for advice and assistance: H. Alden, M. Berry,J. Crain, T. Crist,D. Kuntz, C. Miller,K. Mingo, T. Nauman, R. Schroeder,T. Seasted and K. Vierling.

LITERATURE CITED

ALEXANDER, G. AND J. R. HILLIARD, JR. 1969. Altitudinaland seasonal distributionof Orthopterain the RockyMountains of northernColorado. Ecol. Monogr,39:385-431. BELOVSKY, G. E. AND J. B. SLADE.1993. The role of vertebrateand invertebratepredators in a grass- hopper community.Oikos, 68:193-201. AND . 1995. Dynamics of two Montana grasshopper populations: relationshipsamong weather,food abundance and intraspecificcompetition. Oecologia, 101:383-396. BENNETT,B. C. 1997. Vegetationon the Cityof Boulder open space grasslands.Ph.D. Dissertation, Universityof Colorado, Boulder, Colorado. 108 p. BERRY,M. E., C. E. BOCKAND S. L. HAIRE.1998. Abundance of diurnalraptors on open space grasslands in an urbanized landscape. Condor,100:601-608. BOCK, C. E., J. H. BOCK AND B. C. BENNETT. 1995. The avifauna of remnant tallgrass prairie near Boulder, Colorado. PrairieNat., 27:147-157. Songbird abundance in grasslandsat a suburban interfaceon the Colorado high plains. In: P. D. Vickeryand J. R. Herkert (eds.). Grasslandbirds of the westernhemisphere-ecology and conservationof a diminishingresource. Stud. Avian Biol., in press. BOCK,J. H. ANDC. E. BOCK. 1998. Tallgrassprairie: remnants and relicts.Great Plains Res.,8:213-230. BRANSON, F. A., R. F. MILLER AND I. S. MCQUEEN. 1965. Plant communities and soil moisture relation- ships near Denver,Colorado. Ecology,46:311-319. CAPINERA,J. L. ANDD. R. HORTON. 1989. Geographic variationin effectsof weatheron grasshopper infestation.Environ. Entomol., 18:8-14. AND T. S. SECHRIST.1982a. Grasshoppers (Acrididae) of Colorado. Colorado State University ExperimentStation, Bull. No. 584S. Fort Collins, Colorado. 161 p. AND . 1982b. Grasshopper (Acrididae)-host plant associations:response of grasshopper populations to cattle grazingintensity. Can. Entomol.,114:1055-1062. ANDD. C. THOMPSON.1987. Dynamicsand structureof grasshopperassemblages in shortgrass prairie. Can. Entomol.,119:567-575. COXWELL, C. C. ANDC. E. BOCK. 1995. Spatial variationin diurnal surface temperaturesand the dis- tributionand abundance of an alpine grasshopper.Oecologia, 104:433-439. CURRY,J.P. 1994. Grasslandinvertebrates. Chapman and Hall, London. 437 p. DAUBENMIRE,R. 1978. Plant geography,with special referenceto NorthAmerica. Academic Press,New York.338 p. EvANS,E. W. 1984. Fire as a natural disturbanceto grasshopperassemblages of tallgrassprairie. Oikos, 43:9-16. 1988. Grasshopper(Insecta: Orthoptera:Acrididae) assemblagesof tallgrassprairie: influences of firefrequency, topography, and vegetation.Can. J. Zool., 66:1495-1501. 1989. Interspecificinteractions among phytophagousinsects of tallgrassprairie: an experimen- tal test.Ecology, 70:435-444. R. A. ROGERSAND D. J. OPFERMANN.1983. Sampling grasshoppers(Orthoptera: Acrididae) on burned and unburned tallgrassprairie: night trappingvs. sweeping.Environ. Entomol.,12: 1449-1454. FOWLER, A. C., R. L. KNIGHT, T. L. GEORGE AND L. C. McEW N. 1991. Effects of avian predation on grasshopperpopulations in North Dakota grasslands.Ecology, 72:1775-1781. 1999 CRAIG ET AL.: GRASSHOPPER HABITAT 325

GREATPLAINS FLORA ASSOCIATION. 1986. Flora of the Great Plains. The UniversityPress of Kansas, Lawrence,Kansas. 600 p. JOERN, A. 1982. Vegetationstructure and microhabitatselection in grasshoppers(Orthoptera: Acridi- dae). Southw.Nat., 27:197-209. . 1986. Experimentalstudy of avian predation on coexistinggrasshopper populations (Orthop- tera: Acrididae) in a sandhillsgrassland. Oikos,46:243-249. . 1992. Variable impact of avian predation on grasshopperassemblies in sandhills grassland. Oikos,64:458-463. -AND S. B. GAINES.1990. Population dynamicsand regulationin grasshoppers,p. 415-482. In: R. F. Chapman and A. Joern (eds.). Biologyof grasshoppers.John Wileyand Sons, New York. 563 p. AND G. KLUCAS. 1993. Intra-and interspecificcompetition in adults of two abundant grasshop- pers (Orthoptera:Acrididae) froma sandhillsgrassland. Environ. Entomol., 22:352-361. ANDL. R. LAWLOR.1981. Guild structurein grasshopperassemblages based on food and mi- crohabitatdifferences. Oikos, 37:93-104. AND K. P. PRUESS. 1986. Temporal constancyin grasshopperassemblies (Orthoptera: Acrididae). Ecol. Entomol.,11:379-385. KEMP,W. P. 1992. Temporal variationin rangeland grasshopper(Orthoptera: Acrididae) communities in the steppe region of Montana, USA. Can. Entomol.,124:437-450. S. J. HARVEYAND K. M. O'NEILL. 1990. Patternsof vegetationand grasshoppercommunity composition.Oecologia, 83:299-308. KREBS, C. J. 1989. Ecological methodology.Harper and Row,New York.654 p. LIVINGSTON, R. B. 1952. Relict true prairiecommunities in centralColorado. Ecology,33:72-86. LONG, M. E. 1997. Colorado's FrontRange. Nat. Geogr,190(5):80-103. MOIR, W. H. 1969. Steppe communitiesin the foothillsof the Colorado FrontRange and theirrelative productivities.Am. Midl. Nat., 81:331-340. MORELAND, D. C. AND R. E. MORELAND. 1975. Soil surveyof the Boulder Countyarea, Colorado. U.S. Departmentof Agriculture,Soil ConservationService, Washington, D. C. 86 p. MULKERN, G. B., D. R. TOCZEK AND M. A. BRUSVEN. 1964. Biologyand ecology of North Dakota grass- hoppers. II. Food habitsand preferencesof grasshoppersassociated with the sand hillsprairie. North Dakota AgriculturalExperiment Station, Research Report No. 11. 59 p. MUTEL, C. F. AND J. C. EMERICK. 1992. From grasslandto glacier: the natural historyof Colorado and the surroundingregion. Johnson Books, Boulder, Colorado. 178 p. OTTE, D. 1981. The North American grasshoppers.Volume I. Acrididae: Gomphocerinae and Acridi- nae. Harvard UniversityPress, Cambridge, Massachusetts. 275 p. 1984. The North American grasshoppers.Volume II. Acrididae: Oedipodinae. Harvard Uni- versityPress, Cambridge, Massachusetts. 366 p. PFADT, R. E. 1994. Field guide to common westerngrasshoppers (second edition). WyomingAgricul- turalExperiment Station, Bulletin 912, Universityof Wyoming,Laramie, Wyoming. 41 p., plus factsheets. QUINN, M. A. AND D. D. WALGENBACH. 1990. Influence of grazinghistory on the communitystructure of grasshoppersof a mixed grassprairie. Environ. Entomol., 19:1756-1766. RISSER, P. G., E. C. BIRNEY, H. D. BLOCKER, S. W. MAY, W. J. PARTON AND J. A. WIENS. 1981. The True Prairie ecosystem.Hutchinson Ross PublishingCompany, Stroudsburg, Pennsylvania. 557 p. SIMS,P. L. 1988. Grasslands,p. 265-286. In: M. G. Barbour and W. D. Billings (eds.). NorthAmerican terrestrialvegetation. Cambridge UniversityPress, Cambridge, England. 434 p. THOMPSON,D. C. 1987. Sampling rangeland grasshoppers,p. 219-233. In: J. L. Capinera (ed.). Inte- grated pest managementon rangeland: a shortgrassprairie perspective. Westview Press, Boul- der, Colorado. TSCHARNTKE, T. AND H.-J.GREILER. 1995. communities,grasses, and grasslands.Annu. Rev. En- tomol.,40:535-558. VAN WINGERDEN, W. K. R. E., A. R. VAN KREVELD AND W. BONGERS. 1992. Analysisof species composition 326 THE AMERICAN MIDLAND NATURALIST 142(2)

and abundance of grasshoppers(Orth., Acrididae) in naturaland fertilizedgrasslands. J. Appl. Entomol.,113:138-152. WEBER, W. A. 1995. Checklistof vascularplants of Boulder County,Colorado. NaturalHistory Inventory of Colorado, Publ. No 16. Universityof Colorado Museum, Boulder, Colorado. 66 p. WELCH, J. L., R. REDAK AND B. C. KONDRATIEFF. 1991. Effectof cattlegrazing on the densityand species of grasshoppers (Orthoptera:Acrididae) of the Central Plains ExperimentalRange: a reas- sessmentafter two decades. J. Kansas Entomol.Soc., 64:337-343. WELKOWITZ, J., R. B. EWENAND J. COHEN. 1991. Introductorystatistics for the behavioral sciences (fourthedition). Harcourt Brace Jovanovich,Fort Worth,Texas. 391 p.

SUBMITTED 18 SEPTEMBER 1998 ACCEPTED 8 MARCH 1999 1999 CRAIG ET AL.: GRASSHOPPER HABITAT 327

APPENDIX.-Forty-eight species of grasshoppers'counted on 66 grasslandplots on Boulder,Colorado, open space in 1995-1996, ranked by decreasingdensities

Mean No. plots Species density/lOOm2 occupied Melanoplusfemurrubrum 15.44 56 Melanoplussanguinipes 10.22 45 Melanopluskeeleri 5.79 23 Eritettixsimplex 5.64 61 Pheotaliotesnebrascensis 5.06 35 Melanopluspackardii 4.80 31 Melanoplus bivittatus 3.50 50 Aulocara elliotti 3.44 44 Hypochoraalba 3.33 23 Ageneotettixdeorum 2.46 43 Arphia conspersa 1.93 46 Trachyrhachyskiowa 1.90 34 Chorthippuscurtipennis 1.78 25 Opeia obscura 1.56 22 Arphiapseudonietana 1.34 39 Melanoplus confusus 1.34 29 Amphitornuscoloradus 1.15 35 Dissosteiracarolina 1.01 41 Orphulellaspeciosa 0.79 38 Aeropedellusclavatus 0.68 24 Mermiriabivittata 0.59 19 Spharagemonequale 0.53 24 Hadrotettixtrifasciatus 0.46 22 Melanoplus lakinus 0.41 11 Chortophagaviridifasciata 0.31 8 Acrolophitushirtipes 0.30 16 Mermiriapicta 0.22 7 Hesperotettixviridis 0.20 7 Spharagemoncollare 0.16 9 Xanthippuscorallipes 0.16 8 Derotmemahaydeni 0.14 6 Hesperotettixspeciosus 0.11 2 Aulocarafemoratum 0.11 1 Melanoplusgladstoni 0.09 8 Trimerotropispallidipennis 0.09 8 Melanoplus differentialis 0.09 4 Orphulellapelidna 0.07 8 Cordillacrisoccipitalis 0.06 6 Brachystolamagna 0.04 6 Paropomala wyomingensis 0.04 6 Mestobregmaplattei 0.03 4 Metatorpardalinus 0.03 4 Hippiscus ocelote 0.02 3 Melanoplusfoedus 0.01 4 Aeoloplidesturnbulli 0.01 2 Dactylotumbicolor 0.01 2 Phlibostromaquadrimaculatum 0.01 2 Pardalophorahaldemanii <0.01 1 Nomenclaturefollows Pfadt (1994), then Capinera and Sechrist(1982a) and Otte (1981, 1984)