Ecology of a Grasshopper Community in Desert Shrub-Grassland
in Southeastern Arizona

Robert M. Chew
PO Box 16306 , Portal , AZ 85632
e-mail: rchew@vtc.net
© 2006 Robert M. Chew.  Permission for use granted.  Please include author's name, article title, date, and publication.

            Abstract. Populations of grasshoppers were censused 1983-1987 in a degraded grassland dominated by creosotebush.  For adults, there were three phenological groups of species: (1) two Trimerotropis spp. were present throughout, (2) an Early Group of 2-4 species  were present before midsummer, and (3) a Late Group of 2-5 species, numerically dominated by two Melanoplus spp, were present after midsummer.   Trimerotropis and Melanoplus accounted for 53% of censused individuals out of a yearly average of 800 individuals/0.1 m².  Trimerotropis numbers peaked in late spring to early summer, as did the Early Group; the Late Group peaked in Fall. Average fresh weights of adult females ranged from 0.18 to 3.28 gm; males from 0.08-1.15 gm. Although many species were significantly different in weight, there were three groupings that were not; apparently species weight does not structure the community. There were several patterns of the time of presence of nymphs and gravid females, related to the phase(s) in which the species overwintered. Five species had a microhabitat of open ground, one possibly on a shrub, and one on a suffrutescent. Fourteen forbs and two grasses were foods of primary or secondary prevalence in crop samples. One grasshopper was monophagous on the most common suffrutescent and one almost monophagous on one grass. Three species were polyphagous for herbs, the remaining species are polyphagous for both forbs and grasses. The two pairs of congenerics: Trimerotropis spp. and Melanoplus spp. in general had different diets.  The species of the community appear to be generally separated in their ecological niches by their particular complexes of phenology,  life phases, microhabitat, food habits, and possibly body size. The numbers and biodiversity of grasshoppers was remarkably higher in 1983-1987 than in years before and after, which may partly be the result of certain different patterns of rainfall during these years.

INTRODUCTION

            Two major trajectories were observed: (a) in the 3/5ths of the site which has thin soil (ave. 25 cm) and is dominated by creosotebushes, there was a progressive decline of Larrea, an increase of subshrubs, and a slow increase of black grama and other perennial grasses, the latter becoming clear only after 25 years of removal of cattle from the site; (b) a quicker and very significant increase of grass cover on the 1/5th of the site which has deeper soil (1 m) and more open coverage dominated by Flourensia cernua and Opuntia spp. This has been related to decreasing soil compaction and increased water infiltration following removal of cattle (Valone et al., 2002).

            An important concern of periodic observations was to detect the onset of quantitative short-term or long-term biotic and abiotic changes of the site.  In 1983 it was noted that there was a definite increase in abundance and variety of grasshoppers beyond the few individuals and species that were routinely seen.  The present study of populations of grasshoppers was begun at that time and continued through 1987 when the phenomenon came to an end, not to be repeated. The major objectives were to follow the seasonal timing and fluctuations of numbers of individual species, and to compile information on their reproductive biology and food habits, within the context of changing weather and vegetation dynamics.

METHODS AND MATERIALS

            Site.  The study site is 1.5 km east of the foothills of the Chiricahua Mountains at 1370 m elevation on a gentle slope (1.8%) of alluvium largely derived from limestone and volcanic tuff. The site has two soil-vegetation types: (a) One dominated by creosotebush and subshrubs on very shallow soil (mean depth ca 25 cm) over very indurated caliche. Table 1 lists the abundances of plants therein.  [Plant names follow Kearnery and Peebles, 1961.] Study of grasshoppers was done only in this area. (b) An area dominated by tarbush and prickly pear on deeper soil (1 m or more) of sandy-gravel (Chew and Chew, 1965).                                            

            Since the site was originally semiarid grassland, its biota has components of that grassland, and of Chihuahuan Desert (just to the southeast) and Sonoran Desert (just to the west and northwest). The present landscape is much like that of shrub deserts of the Chihuahuan Desert. The landscape outside the site is a mosaic of patches of grassland remnants, especially in low lying areas, and shrub-grasslands different from the site.

            Censusing. Periodic censuses of grasshoppers were made by slowly walking a rectangular pathway, 152.4 m x 121.9 m, in the creosotebush-dominated vegetation type. As far as possible all individuals within 0.9 m of the center of the path were identified where they sat on the ground or vegetation, or from which they were flushed with an insect net. The sample area of this walking transect is approximately 0.1 ha.  Grasshoppers were netted for identification as necessary; a known number escaped positive identification. Grasshopper numbers were recorded in 30 m increments along the walking transect, and elapsed time, air and soil surface temperature and air movement were recorded at each corner of the rectangular path. Rainfall, and air temperature at a height of 5 cm above soil surface in a well shaded enclosure, were continuously recorded.

            The first census was in July 1983; periodic censuses continued until the end of active presence of no more than a few adult grasshoppers.   Censuses in 1984, 1985 and 1987, were made periodically from the beginning to the end of presence of adults.

            Interpretation of species numbers and presence was made in the context of six “phenological seasons” as defined in Chew (2004), based on generalities of data for plant growth and flowering over 13 years of phenological data:

            (1) Winter: weeks 45-4 (calendar days 309-28) ca. November-January, a period of dormancy of perennials and germination of spring annuals; an average of 35.8% of annual rainfall came in gentle rains in Winter and on into Early Spring;

            (2) Early Spring: weeks 5-16 (days 29-112) ca. February-mid April, a period of vegetative growth, but before blooming of most species;        

            (3) Late Spring: weeks 17-26 (days 113-182) ca. Late April-June, the period of major spring flowering; 7.6% of annual rainfall;

            (4) Early Summer: weeks 27-30 (days 183-210), July, generally before the start of summer flowering and germination of  summer annuals; beginning of thunderstorm rains, which continue on into Late Summer and provide 50.8% of annual rainfall;

            (5) Late Summer: weeks 331-39 (days 211-273), ca. August-September, the time of major summer flowering;

            (6) Fall: weeks 40-44 (days 274-308), October, little or no flowering, senescence or dormancy of plants, 5.8% of annual rainfall.

            Hereafter in the text, when reference is made to these phenological periods, the periods will be capitalized, eg. “Early Spring”. When a general  reference is made to part of a year, it will not be capitalized, eg. “spring”, “winter”.


            Biological measures. Grasshoppers were collected well outside the census pathway to provide adults for measurements of fresh weights and anatomical sizes, examination of reproductive status of females, crop contents and identification of nymphs. Species were documented in reference collections.

            Observations of the locations of immobile grasshoppers were an indication of species’ preferences for sites of concealment and feeding.


            Food habits.   Reference slides for 25 of the most common species of plants (species in Table 1 and others) were prepared from samples of leaves and stems that were ground to a fineness like that of crop contents of grasshoppers.  These were used to identify materials in crops.

           A determination was made of the specific food that was most abundant in a crop’s contents, and the food that was second in abundance.

        The mandibles of each species of grasshopper were examined and categorized according to the descriptions of Isely (1944) as forbivorous, graminivorous and herbivorous (intermediate between forbivorous and granivorous).  This gives an indication of the general food preferences of a species: forbs, grasses or a mixture, within the context of what is available.  

---

RESULTS

            Species present.   Table 2 lists the species that were found at one time or another, in a census or other sampling in the Larrea-dominated vegetation.  Only species marked (*) were sufficiently abundant for any determination of their phenological patterns. 

            Beyond the frequently taken resident species, there were species taken as single or very few individuals that dispersed from locations outside the site.  For example, Taeniopoda eques, which was taken as one individual on day 294 (Fall) in 1987, was abundant, “in the hundreds” on vegetation, in gardens, and on roads 7-9 km to the east of the site.  A curve of the accumulation of species did not show any trend to asymptote.


            Grasshopper phenological patterns.  As shown in Tables 3, 4, 5, and 6, there was a basic pattern of three groupings of species within the grasshopper community:  (1) Two species of Trimerotropis  were present throughout from Early or Late Spring into Fall, and were usually numerically dominant; (2) an “Early Group” of numerically minor species appeared in Early or Late Spring and persisted only into Late Summer; (3) a “Late Group” did not appear until Late Summer and persisted into Fall or early Winter. Psoloessa texana and Cibolacris parviceps were always part of the Early Group; three other species were present in some years. Melanoplus spp. were always in the Late Group, and were then sometimes numerically equal to Trimerotropis spp.   Hippopedon capito was also present in three of four years; two other species were present once.

            A general idea of the numbers of individuals censused in each of the three groups is shown by the following listing in which  “nC*” is the number of individuals counted in all censuses in each year.                                        

 

Species numbers 1983:

             The first census in 1983  was not made until Early Summer when there were already about 50 adult grasshoppers /0.1 ha (Table 3). The nC values in the listing are the actual counts plus an estimate based on following years as to numbers before day 183, Trimerotropis californica and T. pallidipennis, dominated the grasshopper community in numbers and biomass from near the start of Early Summer through to the end of the season in mid-Fall, peaking in numbers in Late Summer.  The Early Group,  P. texana and C. parviceps persisted into the first half of Late Summer.  The Late Group began in Early Summer; Melanoplus spp. were in very small numbers, (nC = 11) together with modest numbers of H. capito.

Species numbers 1984:

            In 1984 (Table 4)  Trimerotropis spp. were present throughout, with a peak in Early Summer; T. pallidipennis was more abundant into Early Summer, after which T. californica became more numerous. Xanthippus corallipes was uniquely present this year as a consistent member of the Early Group.  Melanoplus pictus and M. desultorius  were clearly established as resident species this year in the Late Group together with H. capito, and were then slightly more abundant than Trimerotropis spp.

Species numbers 1985:

            In 1985 (Table 5) total numbers peaked in Late Spring due to T. pallidipennis (nC = 140). There were four species in the Early Group, Hesperotettix viridis for the first time, together with H. capito, which in this year only, had a developmental timing shifted into the Early Group. Melanoplus spp. were common to abundant in the Late Group; probably principally due to M. pictis.  Weather conditions during some censuses made the discrimination of the Melanoplus species difficult. Eremiacris acris was uniquely present this year.

            Censuses were not made in 1986 in the creosotebush vegetation due to a paucity of grasshopper numbers and species. Censuses were made at another site 9 km to the east, where there was a greater diversity of vegetation,  in an attempt to relate grasshopper presence to vegetation type (in manuscript). When grasshopper numbers partly recovered in 1987 a final series of censuses were made.

Species numbers in 1987:

             Trimerotropis spp. were as before (Table 6); the Early Group had three species. The Late Group had many fewer individuals than 1985, particularly for Melanoplus spp.  Ageneotettix deorum was uniquely present with two other species.

            The phenological groupings provide a way for species to  avoid each other in temporal niches.. Even the two Trimerotropis  are slightly different. T. californica  adults appear somewhat later than T. pallidipennis (by 26-84 days in different years) and disappear later (by 7-28 days).  The adults of Early and Late groups are broadly separated from each other.


            Nymphs, adults and females with eggs. Figure 1 summarizes the times of presence of nymphs and adults, and the presence or absence of eggs in dissected females, for seven common species, combining observations of 1983 through1987. There is a variety of patterns. For T. pallidipennis and H. capito, nymphs were not seen until just prior to, or until, adults were first present; in the other species, nymphs were observed well before adults.   For Melanoplus spp. nymphs were seen seventeen weeks before the first adults,  although a few adults were collected outside the census area two weeks earlier. For C. parviceps, gravid females were present in Late Spring; for other species they were not present until Early Summer, and restricted to this season, or present as late as Fall.


            Fresh weights of adult grasshoppers. Table 7a summarizes data on fresh weights of adults collected outside the census area and weighed the same day.  The five species with the heaviest females (ca. 1.1 to 3.3 gm) were taken in small numbers, and are probably not important in the site ecosystem. Females of the other 13 species ranged from average weights of 0.175 to 0.646 gm.

            Males are consistently of significantly less weight than females of the species, by a median male/female ratio of 0.44, range 0.27-0.73.  For 12 of 13 species the mean coefficient of variation of males is  less than that of females.

            The variation of weight within a category can be due to: (1) the adult size to which an individual grew, as affected by its nutrition and temperature and moisture experience through nymphal stages to adult; (2) the reproductive status of the female,  i.e. with ovariole threads only, or with eggs in different stages of development; (3) the  fill of the digestive tract when the individual was captured.

            Several series of individuals through time were sufficiently large to test (1) and (2). In Trimerotropis pallidipennis and T. californica, the next to the largest of the common species, fresh weight (gm) (FW) of females and males was significantly related to pronotal length (cm) (PL), which is taken as an index of body growth:                                                                                   

            FW femaleT. p. = -0.624 + ( 2.111 x PL)   R2 = 0.53

            FW female T. c. = -0.056 + (1.211 x PL )   R2 = 0.70

            FW male T. p. = -0.143 +( 0.936 x  PL)    R2 = 0.46

            Female fresh weights also showed a significant relationship to the length (cm) of eggs (EL) present:

            FW female T. p. = -0.438 + (1.922 x EL)   R2 = 0.44

            In Psoloessa texana, the smallest species, the fresh weight of females was not significantly related to pronotal length, but was significantly related to egg length:

            FW female P. t.  = 0.151 + 0.13EL R2 = 0.36

            The heaviest of the common species, Cibolacris parviceps, was unusual in that there was a significant difference of fresh weight of samples in 1983 and 1987. Females showed no relationship of fresh weight and egg size, but the number of individuals was very low. Male C p., but not females, had a significant relationship of fresh weight to pronotal length.

           Table 7b gives an estimation of the biomass per 0.1 ha for eight species, based on the weights in Table 7a and on nC* summed from Tables 4-7.  This provides a comparison of the relative importance of these species in the site ecosystem, supplementing comparisons based on numbers.


            Food Habits.  Mandible analysis. The mandibles of individuals were examined and categorized according to the three general categories described by Isley (1944): forbivorous, graminivorous, and herbivorous, which is between the previous two. Pinder and Jackson (1988) add another category: “florivores-forbivores” for species eating flowers and leaves.                                                                                    

             My analysis of the mandibles of some of the species is summarized in Table 8.  Seven were clearly forbivorous, 10 graminivorous, 3 herbivorous, and the two Trimerotropis  were various. In five cases where a comparison is possible, my categorization is the same as Isely (1944); only one case is different.


            Grasshopper diets. The small number of observed diets are summarized in Table 9, with species ordered in descending fresh weight. For angiosperms, 69% were specifically identified and 31% were unknown; for grasses it was 58% identified versus 42% unknown. Only one grasshopper,  Hesperotettix viridis, is possibly monophagous, its diet samples showing only one suffrutescent, Gutierrezia sarothrae.   Hippopedon capito ate only one grass, Tridens pulchellus, and two herbs. Three species:  Conozoa carinata,  Melanoplus pictus  and  M. desultorius, are polyphagous for herbs, but their samples had no grass. The remaining species, particularly the two Trimerotropis, are truly polyphagous, samples showing principally herbs, but also some grasses.

            In only one species is the observed diet different from what was expected; samples for Xanthippus corallipes show both herbs and grasses, although I diagnosed its mandibles a graminivorous.

---

            Methods.  Onsager (1977) compared five methods of making density estimates of grasshoppers in rangeland: visual counting over an estimated area (as herein) and four methods of mechanically delimiting an area enclosing a sample.  The latter four methods were all “superior” to the visual, but as he states, “accuracy and precision of...sampling methods were inversely related to convenience.” Because I worked alone and wished to make repeated censuses over a period of years, and the shrubby vegetation precluded mechanical devices, methodology was limited to visual counts while walking over a 0.1 ha pathway, which was reasonably delimited by the range of the sweep net on each side of pathway.

            Practice can make this method reliable for the desired objectives of this study, within the limits imposed by changing weather conditions and the different behaviors of grasshoppers to hide from the observer. Brusven (1972) comments that Hesperotettix viridis are “nearly imperceptible” when on shrubs, but Scroggan and Brusven (1972) comment several times that their nymphs (which may be more elusive than adults) “jump vigorously” when disturbed, making easy to count.


            Phenology.  It is to be expected that there will be “patterns” of phenology within a grasshopper community, considering how species differ in the stage of overwintering, when eggs are laid, how fast they mature, when they hatch, possible diapause,  and when the nymphs emerge (Chapman and Joern, 1990).  Data are sketchy for most species, but some speculation can be drawn from Figure 1. T. pallidipennis, which has adults present throughout the seasons, has nymphs also present except in Fall, and gravid females are present from Late Spring through Fall.  This is consistent with Ball, et al. (1942)  that this species can overwinter in any stage, but most commonly as eggs; Scroggan and Brusven (1972) and Otte (1984) concur with its overwintering as eggs. Ball, et al. (1942) state that Psoloessa texana and Cibolacris  parviceps, members of the Early phenological group, overwinter as nymphs, consistent with Fig. 1, where these two species have nymphs present before adults, gravid females in Late Spring and Early Summer, and disappearance of adults in Late Summer.  Presumably their eggs hatch in Fall or Winter.  Ball, et al. (1942) state that Hippopedon capito (=Rhenita capito in Ball, et al.)) overwinters as eggs, Fig. 1 suggests that their eggs do not hatch until Late Spring and eggs are laid in Fall and Winter.  If Hesperotettix viridis is like H. curtipennis in Ball, et al. (1942), it also overwinters as eggs.  Ball, et al. (1942) suggest that in general Melanoplus spp. overwinter as eggs.

            The data of Fig.1 modestly improve understanding of these species.

            The only comparable study of phenology in a related ecological situation is that of Joern (1979b), for arid Bouteloua eriopoda grassland near Big  Bend , Texas .  Because there are more grasshopper species at his site, there is a more extensive phenological “pattern”: 4 species present through all seasons, including T. pallidipennis and P. texana; 3 species in an Early Group, including C. parviceps and Xanthippus corallipes; 4 species in a Late Group, with M. desutorius and H. viridis; 6 species in a very late group, and 7 species present only in the middle of my six phenological seasons. There are thus some coincidences of species presences in the two sites, and one exception, P. texana being present all seasons.


            Densities. The densities of grasshoppers that triggered this study in 1983 were exceptional with respect to the previous states of the grasshopper community.  They are comparable to values found by Joern (1979b).   He measured densities of the more abundant grasshoppers by a mark-and-recapture study from July to August 1975. Partially these are, in comparison to the present study: 

            However, such values are very much lower than those of other studies in Arizona .  Barnes (1960) measured grasshopper densities (method not given) in uncultivated land in the Salt River Valley 1937-1944. He found:  T. pallidipennis,  29.3 to 86.1/0.1 ha, in a total community of 105 to 545 grasshoppers per 0.1 ha.

            Nerney (1961) measured densities on shortgrass prairie (Bouteloua gracilis,  Eriogonum spp., Gutierrezia sarothrae) 1953--1960, using a calibrated sweep net method. In April-July, he estimated a minimum of 239 and maximum 55,000 grasshoppers per 0.1 ha.  Pfadt (1982), working in the same area during an “outbreak” year,  using visual counts of 0.09 square meter plots in mid May, estimated 3200 to 83,600 grasshoppers per 0.09 square meters, for 24 different sites.


            Fresh weights of adults.  As in Table 7a, there is a step-wise sequence of body weights in the grasshopper community, of females and males. In all cases, males are significantly smaller than the females of a species. If body size is one of the factors determining ecological niches, there is some basis therein for niche separation.

            Although there is a sequence of significant differences for both   females and males, there are several groupings of species in which sizes are not significantly different, e.g.: (1) female C. parviceps, T. californica and T. pallidipennis; (2)  female M. pictus, M. desultorius and H. capito; and (3) male T. californica, T. pallidipennis, M. pictus, M. desultorius and H. viridis. Within these groups there is here no evidence of adaptation of body size to avoid competition.

            The role of body size in structuring communities has been studied in widely different groups of animals  (Ernest, 2005). Ernest found that body size did not play a role in structuring communities of small mammals.


            Microhabitat.  My limited observations of immobile adults show that some species have a preference for a particular substrate.  Lightfoot (1985), working in a Chihuahuan Desert site, concluded that species’ preferences often have a 90-100% fidelity.  The following are my substrate assignments for some of the species of my site.

            Bare ground: Trimerotropis pallidipennis                                                

                                    T. californica

                                    Psoloessa texana

                                    Cibolacris parviceps

                                    Derotmema haydeni

            Shrubs: possibly Melanoplus pictus on Flourensia cernua

            Subshrubs: Hesperotettix viridis, principally on Gutierrezia sarothrae.

            T. californica may at times feed and rest in grass clumps (Otte 1981) and probably all species at times jump or fly into plants when escaping disturbance.

             Four comparisons of the numbers of T. pallidipennis and T. californica in successive 30 m units along the census line showed no significant differences in their distribution.


            Food habits.  As in Table 9, the number of observations of crop contents is limited, but the results are consistent with the literature.  X. corallipes was most dependent on grasses, and secondarily used an annual herb and one suffrutescent.  In northern Colorado its diet was 77-93% grasses and 7-17% forbs (Ueckert 1968). C. parviceps used six forbs from different plant families, although Otte and Joern (1977) cite it as oligophagous.

            Although both Trimerotropis are polyphagous they were well separated in their diets. There was only one herb species in common between the six species eaten by T. californica and the five by T. pallidipennis. One grass was eaten by T. pallidipenis, none by T. californica.  However, both species had a high score of unknowns. T. pallidipennis showed a high usage of flowers in six samples, particularly of one common annual, Gilia longiflora; T. californica had flower fragments in three samples.  "Florivory" is rarely mentioned in the literature. Otte and Joern (1977) found T. californica (=T. strenua) to have a larger mean diet breadth (3.94 species per Arizona locality) than T. pallidipennis (3.06). Although they never found the predominant weedy grass, Tridens pulchellus (14% cover in present site) in any crops of desert grasshoppers, presumably to be due to its hairiness and leaf toughness, here it was the primary food in one crop of T. pallidipennis. Because of its broad geographical and ecological distribution, T. pallidipennis can be expected to be a broadly based polyphagous species; in one study  Joern (1979a) scored it as eating 68% forbs (14 species) and 21% grasses (7 species), and in another study of 101 crops, 62% forbs and 33% grasses.

            Melanoplus spp. had limited separation of diets. All three species of forbs eaten by M. pictis were also used by M. desultorius. Neither had samples containing grass. Joern (1979a) scored 25 crops of M. desultorius as 93% with forbs (14 species) and 3% grasses (1 species). 

            In Table 9 there are two species limited to one plant of primary and secondary importance: six crops of Hippopedon capito with only the grass Tridens pulchellus, and four crops of Hesperotettix viridis with only the suffrutescent Gutierrezia sarothrae. The oligophagy of H. viridis is noted by Ball, et al. (1972), Brusven (1972) and Otte and Joern (1977).

            The smallest species, Psoloessa texana, used one species of grass in 11 of 17 identified cases, versus three species of forbs in 6 of 17. Joern (1979a) found this species to have grass, principally one species, in 87% of 87 crops.

            Overall there is no evidence for Hansen and Ueckert (1976) that smaller species of grasshoppers have a more restricted diet than larger ones.

            Creosotebush (Larrea tridentata) is an anomaly for the grasshopper community since no species used it as food or substrate. This is  unexpected, since Cibolacris parviceps is closely associated with this shrub and has been found to feed readily upon, and sometimes depend upon Larrea, (Otte and Joern, 1977).  Ligurotettix coquilletti, an oligophagous feeder on Larrea, was not present, although it is present 16 km or less to the east. This species is sensitive to the quality of creosotebushes (Greenfield et al, 1987); this may explain its absence from the site, which has a  declining population of Larrea.  Bootettix argnatus, which “lives exclusively on Larrea”  and is monophagous for it, is present throughout southern and southeastern Arizona , and can have high population densities (Otte 1981), but I have not encountered the species locally.


            Weather and Grasshoppers.  Changes in grasshopper abundances are commonly attributed to changes in weather, generally examined using correlation analyses.  This is not possible with the present data, and such correlations are problematical, considering the many life stages involved and the possible interactions of abiotic and biotic factors.  However, for heuristic reasons one can speculate as to what may have been involved in the increase of grasshoppers in 1983, the slump in numbers in1986 and the termination of any increase after 1987.

            Table 10 shows mean monthly rainfalls 1980-1994.  There are some differences in the period of 1980-1983 and on into 1987 that may have affected grasshopper numbers. The winter rains (December through January) are, in sequence:  1982 60 mm, 1983 114, 1984 104, 1985 101, 1986 12, 1987 125, 1988 75.  Winter is the time of germination of winter annuals and if there is very low rainfall there is no germination (Chew 2004). If one accepts a direct coupling of rainfall, annual productivity and grasshopper reproduction there is a basis in the rainfall values for good growth of annuals 1983-1985 supporting increased grasshopper numbers then, a serious decline in numbers in1986, revived by good annual growth in 1987, but not sustained by poorer annual growth 1988 (75 mm) and 1989 (73 mm). The values of the rains to annual growth is not reflected in adult numbers until Late Spring and Early Summer, when Trimerotropis numbers peak. The Early Group also peaked at this time in 1984 and 1985.      

            The generally expected dry months of February-June, between winter and summer rains, occurred 1980-1986, with 2-5 months with less than 8.7 mm /month; 1984 is the most extreme, having only 2.8 mm of rain in May and no rain in the other four months.  However, there is one exception: March of 1982 and 1983 had high rainfalls; these March rains supplementing good winter rains, may have resulted in better growth and quality of plants in June to August. The relevance of the dry months to the success of grasshoppers is obscure. Adults appeared in censuses as early as April and could reach peak numbers in June.  There is no hint  in the records of summer rainfalls of a reason for the build up of grasshopper numbers and species in 1983, which prompted this study. The summer rains, July-September, providing on average 49.7% of the annual total, were the lowest recorded in 1982 (Table 10) and July-August rains of 1983 were second lowest.  However, September 1983 had the highest rainfall (106 mm) by far of any year. The latter may have stimulated an additional increase in growth and quality of suffrutescents and semishrubs, a basis for sustaining the 1983 populations through 1985. The summer rains of 1983 continued in the following months, merging summer rains into the start of the winter rains of November 1984. This is a possible basis for  the second peak of Trimerotropis in August to October, and the peaks by the Late Group in October 1984 and 1985.           

             A record dry period then began Dec. 1983 through June 1984. A delayed effect of this might be seen in the decline in numbers in 1986, in spite of the 1985  summer rains again merging into winter rains.  The exceptional summer rains of 1988 were not followed by good winter rains. Hence, possibly there was no recovery of the grasshopper community.

            Study of daily temperature records was limited to two aspects:  (1) the frequency of weeks in which the mean weekly minimum temperature was below 33 F (0.6 C), and (2) the frequency of weeks in which the mean weekly maximum was over 100 F (37.8 C).  These aspects of air temperature are important to the large grasshopper, Taeniopoda eques (Whitman 1986). This species requires 850 degree days to complete its development from hatch to oviposition, but southern Arizona averages only 692 degree days. T. eques must be able to make up the difference of 158 degree days by thermoregulation, as by basking in the sun and body postures. This might be impossible if cold weeks are too frequent. At the other extreme, when air temperature near the ground is over 38 C ; a grasshopper species is susceptible to heating beyond the body temperature at which it loses motor control.  Considering the behavioral options grasshoppers have to avoid overheating (Chappell and Whitman, 1990), this must usually be unlikely. Chappell (1983) found that for T. pallidipennis  heat torpor begins at body temperature of 49-51 C, and its maximum voluntary tolerated temperature is approximately 45-48 C, i.e. just below heat torpor. He notes that T. pallidipennis in the Colorado Desert “must restrict its activity for much of the day to avoid overheating, and can easily overheat in sunlight”. Very little is known about temperature relationships of other species of Table 7.

            Table 11 shows the limited  analysis of air temperatures for 1982-1985. These are for temperatures measured in the shade at 5 cm above ground surface.  The temperature at ground surface in the sunlight would be higher. Temperatures taken during each census show that maximum soil surface temperature by end of each census was: in 1984 a median of 42.5 C, range 29.2-55.6 C (highest in July), in 1985 a median of 43.0 C, range 26.4-55.6 (highest in July), and in 1987 a median 37.5 C, range of  30.5-58.9 (highest in June). There were then occasions when soil surface temperatures were in the range of heat torpor for T. pallidipennis.

             As in Table 11, in 1982 there were 13 consecutive weeks when mean weekly air temperatures exceeded 38 C.  In 1983 and 1984 there were only 7 such weeks, and these were not all consecutive, hence the potential danger of overheating was less for grasshoppers in 1982, before the increase in numbers, and in 1983, the first year of increase. In 1982-1983 there were  6 weeks with means below 33 C, and in 1983-84 there were 11 such weeks, i.e. less potential hazard prior to first year of increased grasshopper numbers. It is obvious that there is little information about the “ecological”  survival limits of body temperature in natural populations of grasshoppers.


            Resource use.  Overall the previous speculations on differences in grasshopper phenology, body weight, microhabitats and diets support Mulkern’s (1980) conclusions from an eleven year study of population fluctuations of 28 species of grasshoppers on permanent grassland plots.  He found that for 17 numerous species “there was little overlap of resource use of the species, as they were separated by time and/or space and/or food requirements, and were not competing with each other.” No single or group of factors determined population fluctuations of any species. “Each species appeared to respond independently to conditions with at least one species able to increase in numbers under any set of conditions.”

---

Ball, E. D., E. R. Tinkham, R. Flock, and C. T. Vorhies. 1942. The grasshoppers and other Orthoptera of Arizona . Technical Bulletin, University of   Arizona College of Agriculture . 93:256-373.

Barnes, O. L. 1960. Observations on the desert grasshopper, Trimerotropis pallidipennis pallidipenis, in Arizona. J. Economic Entomology 53:721-724.

Brusven, M. A. (1972). Differentiation and ecology of some common Catantopinae and Cyrtacanthacridinae nymphs (Orthoptera: Acrididae) of Idaho and adjacent areas. Melanderia 9:1-31.

Chapman, R. F., and A. Joern, Eds. 1990. Biology of grasshoppers. John Wiley & Sons, New York .                                                                                       

Chappell, M. A. 1983. Metabolism and thermoregulation in desert and montane grasshoppers.  1:126-131.Oecolohgica (Berl.)

Chappell, M. A., and D. W. Whitman. 1990. Pages 143-172 in Chapman, R.F., and A. Joern. Biology of grasshoppers. John Wiley & Sons, New York .

Chew, R. M. 1961. Ecology of the spiders of a desert community. J. New York Entomological Society 69:5-41.

Chew,  R. M. 1979. Some ecological characteristics of the ants of a desert-shrub community in southeastern Arizona . American Midland Naturalist 98:33-49.

Chew, R. M., 1982. Changes in herbaceous and suffrutescent perennials in grazed and ungrazed desertified grassland in southeastern Arizona .  1958-1978. American Midland Naturalist 108:159-169.

Chew, R. M. 2002a. The effect of removal of creosotebushes from a Larrea-dominated shrub grassland. larrea-plot.com, Papers by Robert M. Chew.

Chew, R. M. 2002b. The food habits of the sympatric horned lizards Phrynosoma modestum and P. cornutum.  larrea-plot.com, Papers by  Robert M. Chew.

Chew, R. M. 2004. Flowering phenology and diversity of dicots in desert-shrub grassland, south eastern Arizona .  larrea-plot.com, Papers by Robert M. Chew.

Chew, R. M., and A. E. Chew. 1965. The primary productivity of a desert shrub (Larrea tridentata) community. Ecological Monographs 40:1-21.

Chew, R. M., and A. E. Chew. 1970. Energy relationships of the mammals of a desert shrub (Larrea tridentata) community. Ecological Monographs 40:1-21.

Chew, R. M., and J. DeVita. 1980. Foraging characteristics of a desert ant assemblage: functional morphology and species separation. J. of Arid Environments 3:75-83.

Ernest, S. K. M. 2005. Body size, energy use, and community structure of small mammals. Ecology 86:1407-1413.

Greenfield , M. D., T. E. Shelly, and K. R. Downum. 1987. Variation in host plant quality: implications for territoriality in a desert grasshopper. Ecology 68:828-838.

Isely, F. B., 1944. Correlation between mandibular morphology and food specificity in grasshoppers. Ann. Entomological Society of America 37:47-67.

Joern, A. 1979a. Feeding patterns in grasshoppers (Orthoptera: Acrididae): factors influencing diet specialization. Oecologia 38:325-347.

Joern, A. 1979b. Resource utilization and community structure in assemblages of arid grassland grasshoppers (Orthoptera: Acrididae). Trans. American Entomological Society 105:253-300.

Kearney , T. H., and R H. Peebles,  1960. Arizona Flora. Second edition with supplement. University of California Press, Berkeley.

Lightfoot, D. C. 1985. Substrate utilization and guild structure in desert grasshopper assemblages. M. S. Thesis, New Mexico State University .

Mulkern, G. B. 1980. Population fluctuations and competitive relationships of grasshopper species (Orthoptera:Acrididae). Trans. American Entomological Society 106:1-42.

Nerney, N. J, 1961. Effects of seasonal rainfall on range conditions and grasshopper population, San Carlos Indian Reservation, Arizona . Journal of Economic Entomology 53:382-386.

Onsager, J. A. 1977. Comparison of five methods for estimating density of rangeland grasshoppers. Journal of Economic Entomology 70:187-190.

Otte, D. 1981. The North American grasshoppers. Volume 1. Acrididae. Gomphocerinae and Acridinae. Harvard University Press, Cambridge.

Otte, D. 1984. The North American grasshoppers Volume II. Acrididae. Oedipodinae. Harvard University Press, Cambridge .

Otte, D., and A. Joern. 1977. On feeding patterns in desert grasshoppers and the evolution of specialized diets. Proc. The Academy of Natural Sciences of Philadelphia 128:89-126.

Pfadt, R. E. 1982. Density and diversity of grasshoppers (Orthoptera: Acrididae) in an outbreak on Arizona rangeland. Environmental Entomology 11:690-694.

Pinder, J. E., and P. R. Jackson. 1988. Plant photosynthetic pathways and grazing by phytophagous orthopterans. American Midland Naturalist 120:201-211.

Scrogan, A. C., and M. A. Brusven. 1972. Differentiation and ecology of common immature Gomphocerinae and Oedipodinae (Orthoptera: Acrididae) of Idaho and adjacent areas. Melanderia 8:1-76.           

Ueckert, D. N. 1968. Seasonal dry weight composition in grasshopper diets  on Colorado herbland. Annals of the Entomological Society of  America 61:1539-1544.

Valone, T. J., M. Meyer, J. H. Brown, and R. M. Chew. 2002. Timescale of perennial grass recovery in desertified arid grasslands following livestock removal. Conservation Biology 16:995-1002.

Whitman, D. W. 1988. Function and evolution of thermoregulation in the desert grasshopperTaeniopoda eques.  Journal of Animal Ecology 57:369-383.41.

Figure 1

Table 3 1983 Adult and nymph grassshopper numbers / 0.1 ha. Total includes species not abundant or frequent enough to be listed in columns. M = Melanoplus spp.; UA = unkown adults; UN = unknown nymphs; other species as in Table 2. nC = number of individuals censused in each phenological group. - TP TC PT CP HC M  UA UN TOTAL 7-11  d 182 4 18.6 14.6 0.8 0 0 7.4 2  46.6 7-17  d 198 3.8 11 5.8 1.0 0 0 1.9 1 25.3 7-21  d 202  1.9  10.8 5.8 3.2 0.6  0 1.2 1 24.5 7-25  d 205 1 18 1.3 2.3 2.3 0.7 4.9 0 32.4 7-29  d 210 2.7 18.1 3.4 0.7 3.4 0 4.7 1.3 34.3 8-5    d 217 8 40 4 2 7 0 0 1 63 8-15  d 227 5.3 34.7 2 2 2 0 2 0 52 8-26  d 237 16.8 53 3.8 0 7.6 0 2.6 1.3 83.8 9-8    d 250 17.7 58.3 0 1 5 1 3 2 87 9-21  d 263  20.5 23.5 0 1 6 1 12 0 65 9-27  d 269 3.9 23.1 0 0 2 0 6 0 35 10-11 d284 19 19 0 0 2 8 22 0 70 nC 595* 107* 49 - - -          * This is the counts listed above plus an estimate of numbers that could have been counted before day 182, based on an average for 1984, 1985 and 1987. Estimate was made only for these four species.

Table 7a Fresh weights of adult grasshoppers.  nC* = number of individuals counted in all censuses 1983-1985 and 1987; n = number of individuals in calculation of statistics. l = species with different letters are significantly different, females A-E, males a-f. Fresh weights, gm
.Species	.Year.	nC*	n	Mean	SD	95% limits	l
Xanthippus corallipes	.	25	f 2	3.288	0.064	3.039-3.483	A
.	.	..	m 8	1.050	0.124	0.949 - 1.151	a
Cibolacris parviceps	1983	..	f 10	0.646	0.095	0.579 - 0.713	B
..	..	..	m 12	0.172	0.027	0.155 - 0.189	c
.	1987	.	f 17	0.614	0.078	0.574 - 0.654	B
.	.	.	m 14	0.271	0.093	0.218 - 0.324	d
Trimeritropis californica	.	747	f  46	0.631	0.137	0.590 - 0.672	B
.	.	.	m 94	0.270	0.036	0.263 - 0.277	c
Trimeritropis pallidipennis	.	911	f 33	0.592	0.144	0.541 - 0.643	B
.	.	.	m 40	0.264	0.059	0.245 - 0.283	c
Conozoa carinata	.	40	f 18	0.448	0.081	0.408 - 0.488	C
.	.	.	m 25	0.150	0.019	0.142 - 0.158	d
Melanoplus pictus	.	178	f 23	0.391	0.095	0.350 - 0.432	C
.	.	.	m 16	0.219	0.026	0.205 - 0.233	c
Melanoplus desultorius	.	178	f 4	0.376	0.080	0.330 - 0.422	C
.	.	.	m 6	0.188	0.021	0.167 - 0.209	c,d
Hippopedon capito	.	154	f 10	0.360	0.087	0.299 - 0.421	C,D
.	.	.	m 20	0.120	0.019	0.111 - 0.129	e
Hesperotettix viridis	.	59	f 12	0.300	0.047	0.270 - 0.330	C,D
.	.	.	m 4	0.220	0.032	0.176 - 0.264	c,d
Psoloessa texana	.	269	f  56	0.175	0.030	0.167 - 0.183	E
.	.	.	m 33	0.077	0.012	0.073 - 0.081	f
-

Table 7b Species biomass estimates (gm/0,1 ha), Calculated from data in Table 7a as: Biomass = (proportion of n) x nC* x wt.
Average
Species	Sex	Fresh wt	Biomass	Species total
T. pallidipennis	f m	0.592 0.264	244 139	383
T. califomica	f m	0.631 0.270	155 135	290
Melanoplus spp.	f m	0.384 0.204	85.7 27.1	113
X. corallipes	f m	3.29 1.05	16.4 21.0	37.4
P. texana	f m	0.175 0.077	29.6 7.7	37.3
C. parviceps	f m	0.63 0.22	23.0 7.8	30.8
H. cap/to	f m	0.360 0.120	18.5 12.3	30.8
C. carinata	f m	0.448 0.150	7.5 3.5	11.0