Heredity 56 (1986) 397-408 The Genetical Society of Great Britain Received 20 September 1985

Genetic variation in prairie populations of Melano plus San guinipes, the migratory grasshopper

W. Chapco and M. J. Bidochka Department of Biology, University of Regina, Regina, Saskatchewan, Canada S4S 0A2.

Genetic variation at two visible loci and 10 electrophoretic loci in populations of the migratory grasshopper, Melanoplus sanguinipes, in the Canadian prairies is described. Tibia colour and femoral stripe loci exhibit considerable geographic variation in frequency. Multiple regression and correlation analyses revealed dma! variation patterns for tibia colour, with minimum annual temperature accounting for about 40 per cent of that variation. The possible direct and indirect effects of this factor are discussed. No spatial patterns were evinced for the femoral stripe locus despite its close linkage to the tibia colour gene. Whatever the nature of selection, coefficients must be large enough to maintain polymorphism and divergence in opposition to the swamping effects of gene flow. This does not appear to be the case for the allozyme loci; frequencies of the most common allele at each locus are high (average heterozygosity =45per cent) and about the same at each location. Tte mean level of polymorphism, 378 per cent, is well within the range of most insects.

The migratory grasshopper, Melanoplus to an agricultural ecosystem, which according to sanguinipes (formerly known as M. bilituratus, many (e.g., Roffey, 1970; Turnock, 1977) coincided M. mexicanus and M. atlantis—see Gurney, 1962), with—some say caused—the extinction of the most is widely distributed throughout North America, destructive grasshopper in North America, the occupying (particularly sandy aspects of) a variety Rocky Mountain grasshopper, M. spretus. During of habitats such as grasslands, roadsides, open this century, M. sanguinipes has enjoyed, at areas in forests and even northern tundra (Vickery irregular intervals, massive outbreaks in numbers, and Kevan, 1983). This, in part, is attributable to which today can be predicted largely on the basis the insect's polyphagia (Mulkern et aL, 1969; of climatic conditions (Gage et a!., 1976; Gage and Vickery and Kevan, 1983) and reported capacity Mukerji, 1977). As is typical for most members of to migrate over long distances (Parker et aL, 1955; the Orthopteran order, information on the popula- Johnson, 1969; Vickery and Kevan, 1983). Given tion genetics of M. sanguinipes is patchy (Chapco, these features one might, therefore, expect the 1983b). While there is certainly no shortage of species to be fairly rich genetically depending on reports on phenotypic variability for this species the amount of gene flow. Migratory proclivities (Shotwell, 1930; Gurney and Brooks, 1959), until need not be translated into gene flow effects recently only the polymorphism, in hind tibia (Endler, 1977), but in M. sanguinipes, at least, there colour (red versus blue), has been genetically is good evidence from past records of outbreak researched (King and Slifer, 1955; the locus was periods that migrating females do deposit eggs labelled T by Chapco, 1980a). Brooks' (1958) contributing to the next year's pullulation (Parker classification and Gurney and Brooks' (1959) et a!., 1955). Also based on historical accounts, it monograph on the genus Melanoplus give several is reasonable to assume that the current population qualitative accounts of the geographical distribu- structure was engendered by the development of tion of tibia colour variants. In Canada, the red agriculture on the prairies over a century ago, form is predominant in Quebec, Ontario, British settlement and cultivation providing both suitable Columbia and the prairie parklands (the "more pod laying sites and succulent food (Bird, 1961; humid part"), whereas the blue form is pre- Roffey, 1970; Turnock, 1977). M. sanguinipes was dominant in the prairie grasslands. No distribution probably spottily distributed prior to the transition maps are provided, although Brooks (1958) does 398 W. CHAPCO AND M. J. BIDOCHKA state that the blue form is "usually confined to a tibia allele (TR)decreaseswith about a three-fold narrow strip south of a line from Medicine Hat, drop occurring from the Manitoba-Saskatchewan Alta. to Brandon, Man., but after a series of 'good' border near Ebenezer (51° 22', 102° 27'), located in grasshopper years, or in local dry spots, this form the aspen-parkland to Regina (50° 25', 104° 35'), predominates on the grasslands". Based on rearing located in the grassland area (Chapco, 1983a). studies, Brett (1947) claimed that high humidity Since no material was obtained between these two and high temperature (also the consuming of let- points, it is not known whether the changes are tuce and corn) "produce" a preponderance of abrupt or smooth. Also examined were the distri- individuals with red caudal tibiae. The results are butions of three other traits whose genetics have taken as experimental confirmation of Hebard's previously been worked out: red back (Prolocus), (1925) assertion that in South Dakota, the red femoral stripe (Ostlocus)and lactate dehy- forms are more frequent in the "more humid sec- drogenase (Ldh locus) (Chapco, 1980a; 1980b; tions of the state; elsewhere in more arid areas, 1984). Ost exhibits considerable geographical vari- hind tibiae are mostly glaucous". Brett's experi- ation; Pro is uniformly rare everywhere; and the ments, unfortunately, are difficult to evaluate since most common Ldh allele is uniformly frequent in information on source of material, and in some all sampled areas. cases, sample sizes, and proper controls are The present study concentrates on the southern lacking. Saskatchewan area in order to (a) establish the In a preliminary survey of several North nature of the TR dine, (b) test statistically the American populations, it was shown that from hypothesis, suggested by the literature, that the Quebec to Saskatchewan, the frequency of the red distribution of T' is connected to moisture and

Table 1 Red tibia colour and femoral stripe allelic frequencies at 32 prairie locations

Latitude Longitude Locality ('North) ('West) N %TR %Ost

1.Carnduff 4917 10183 254 14 53 2.Oxbow 4923 10218 115 13 56 3.Mankota" 4942 10707 74 11 28 4.Pangman 4965 10463 111 13 62 5.Kayville 4973 10517 42 20 65 6.Assiniboja* 4963 10598 125 12 38 7. Ormiston* 4975 10537 185 12 69 8.Bateman 5002 10675 71 12 71 9.Gull Lake 5013 10845 78 12 58 10. Rowatt 5033 10462 184 11 38 11.Whitewood 5033 10425 106 25 60 12. Jameson* 5042 10228 104 19 33 13.Grenfell 5042 10293 100 31 65 14.Indian Head* 5053 10367 148 16 33 15.Stockholm 5067 10225 51 20 72 16.Keeler 5070 10585 55 16 44 17.Bethune 5072 10513 186 12 44 18.Russell 5077 10128 48 25 62 19.Tugaske 5088 10627 173 13 59 20.Melville* 5092 10280 55 44 34 21.Craik 5105 10582 80 8 45 22.Roblin 5120 10145 46 58 100 23.Yorkton* 5122 10247 56 31 31 24.Springside 5135 10275 15 42 42 25.Ebenezer 5137 10245 345 41 31 26.Outlook 5150 10705 72 17 49 27.Good Spirit Lake Park 5157 10265 20 50 68 28.Canora* 5162 10243 51 46 30 29.Harris 5173 10745 34 18 21 30.Preeceville* 5197 10267 77 53 29 31. Biggar 52'07 10800 135 23 48 32.Hague 5250 10642 110 21 41 *Sites also sampled for variation. GENETIC VARIATION IN GRASSHOPPER POPULATIONS 399 temperature, and (c) by including several other trophorsis. A portion of the eviscerated tissue was loci, mostly electrophoretic, lay the foundation for stored at —70°C for later analysis. Two gel buffer discovering the genetic structure of what appears systems were used: (1) 0375 M tris HC1, pH 89 to be a continuously distributed species. The results (for LDH, MDH, G-6PD, ODH, SDH, ODH and should provide an interesting antithesis to the few SOD) and (2) O1 M tris borate with 2mM EDTA allozyme studies that have tended to concentrate (pH 89) for aGPDH-1, aGPDH-2, XDH, GLDH on acridids of low mobility (Tepper, 1979; Daly and ME. A 05 M tris glycine (pH 8.9) running et a!., 1981; Gill, 1981a; 1981b; 1981c). buffer was used throughout. The stains for XDH, G-6PD and SDH were essentially the same as those described by Shaw and Prasad (1970); SOD, ME, MATERIALS AND METHODS aGPDH-2 and ODH were stained (with some slight modifications) according to Ayala et al. (1972);LDH and MDH were stained according Collection to Chapco (1984), 50mg malic acid substituting Adult M. sanguinipes, scored for visible poly- for lactic acid in the latter case. Gels were run at morphisms, were collected between the months of 7 per cent acrylamide and electrophoresis per- July and August, 1978 to 1984, from 30 different formed for 2 hours at 250 volts. localities in southern Saskatchewan; two localities were in Manitoba (table 1). For electrophoretic analysis, grasshoppers from nine of the sites Analysis (marked with an asterisk in the table), situated Allelicfrequencies were estimated by gene count- along a climatic gradient (Richards and Fung, ing in the case of the enzyme loci (all alleles proved 1969) that extends from the generally hot and dry to be codominant) and in the case of the two colour (summer) short grass prairie in the south-west to traits, Hardy-Weinberg equilibrium was assumed the relatively cooler and moister aspen parkiand and the frequencies of the dominant alleles esti- in the north-east, were sampled in 1982; three mated by taking the square-root of the correspond- places were resampled in 1983. Collections were ing recessive phenotypic frequencies and subtract- made using a sweep net, usually along roadsides ing the values from 1. and/or adjacent fields. Locations were compared by applying log- likelihood ratio tests or G-tests (Everitt, 1977) to 2 (phenotypes) x 32 (locations) contingency tables Traits in the case of the colour traits and to 2 (common Animalswere scored for two colour traits: hind allele versus "other" alleles pooled) x 9 (locations) tibia colour (red allele, T' dominant to blue allele, contingency tables in the case of the elec- TB)andhind femoral stripe (orange stripe allele, trophoretic traits. Geographical diversity was Ost,dominantto stripeless allele, ost) and for measured by computing Nei's (1972) genetic iden- possible electrophoretic variations at the 11 tity indices between pairs of samples. G-tests were enzyme loci: Ldh (lactate dehydrogenase), Mdh also applied to multiway tables to evaluate interac- (malate dehydrogenase), G-6pd (glucose-6-phos- tions among several variables, e.g., genotype x phate dehydrogenase), Sod (superoxide dis- location x year. mutase), Odh (octanol dehydrogenase), Sdh (sor- Climatic information was obtained (Environ- bitol dehydrogenase), aGpdh-1 and aGpdh-2 (a ment Canada, 1983) for the weather station nearest glycerophosphate dehydrogenase), Xdh (xanthine to each sampled area; in a few cases, values dehydrogenase), Gldh (glutamate dehydro- between weather stations were obtained by inter- genase), and Me (malic enzyme). polation.Climaticvariableschosen were maximum temperature (Tmax, the average daily temperature for the hottest month—always July), Electrophoresis minimum temperature (Tmjn, the average daily Allozymepatterns were ascertained by vertical minimum temperature for the coldest month— polyacrylamide slab gel electrophoresis (Hoefer always January) and annual precipitation (Prec, Scientific Model SE 500). After recording the snow and rain combined). Nineteen- to 30-year colour traits, eviscerated flight muscle was averages (depending on the weather station) were homogenised in a 10 per cent sucrose solution. The taken, rather than values for the year in which homogenates were then centrifuged at 3000 rpm collections were made. To examine the hypothesis for 5minutesand the supernatant used for elec- that the incidence of red legged grasshoppers is 400 W. CHAPCO AND M. J. BIDOCHKA connected to temperature and moisture, the fol- sequently, for assuming genetic equilibrium lowing steps were taken: (1) the data were first (otherwise the method of estimating TR and Ost examined for clinal patterns with respect to three frequencies would be invalid) is supported by the "positional variables": latitude (Lat), longitude analysis of two data subsets. First, in the Rowatt (Long) and elevation (Elev) by multiple regression, region, neither the red tibia nor the orange stripe (2) one or more principle components were extrac- phenotypic frequencies varied significantly over ted from the correlation matrix of climatic vari- six years (1978—1983) of sampling; G-test values ables in order to try to identify "meaningful" for each 2 x 6 contingency table are 384 (P =0.57) weather components (Newham, 1968), e.g."arid- and 398 (P=0.55), respectively. Second, nine ity", and (3) multiple regression and partial corre- sites situated along the previously mentioned cli- lation analyses were applied in a manner essen- matic gradient were sampled in 1982 and 1983. tially similar to that of Oakeshott eta!.(1980) to Data for each trait arranged in a three-dimensional see if any of the positional variations could be contingency table, when analysed by a series of accounted for by the weather principle com- G-tests, failed to yield significant phenotype x ponents; since temperature and moisture were location x year interactions (G8df= 654;P =059 singled out as possible determinants, step 3 also and 952; P=030 for tibia colour and femoral examined the contributions of Tmax, Tmin and Prec. stripe, respectively) and significant phenotype x All cases were weighted by sample size. The analy- year interactions (Gldf= 123; P=027 and 284; ses presented were performed on untransformed P =OO9for tibia colour and femoral stripe, respec- frequencies; essentially the same conclusions were tively). If the two subsets are exemplary, then reached using angular transformed data. choosing the "best" years for the full analysis is not expected to produce misleading conclusions. Before examining data for systematic geo- RESULTS graphical variation and for association with one or more weather factors, each matrix (X, standar- Colour traits dised) of positional variables and climatic vari- Frequency estimates of the red tibia allele, TR, and ables was diagnosed for possible collinearity of the orange femoral stripe allele, Ost, are presen- problems or ill-conditioning created by non- ted in table 1 for each of the 32 collection sites. orthogonality. If severe, collinearity can lead to There is tremendous variation among localities for large variances and covariances for standard least both traits (2 x 32 contingency table G values are square estimators of regression coefficients (Mont- 4468 and 3647 for tibia colour and femoral stripe, gomery and Peck, 1982) and, possible, erroneous respectively; P <0001 in both cases). Frequencies inferences (e.g., see Mauriello and Roskowski, of TR range from 8 per cent at Craik, Saskatchewan 1974). In the present study most variables within to 58 per cent at Roblin, Manitoba, 300 km away; each set are significantly correlated (table 2), there- Ost frequencies range from 21 per cent at Harris fore suggesting possible collinearity problems. to 100 per cent at Roblin, 360 km away. Frequen- Several methods have been proposed for diagnos- cies quoted represent years (within 1978—84) for ing the extent of collinearity and for deciding if which sample sizes were maximal. Justification for alternative procedures such as ridge regression, are deeming year effects insignificant and con- necessary. Following Montgomery and Peck

Table 2 Correlation matrix for red tibia colour (Tib) and femoral stripe (Fs) phenotypic frequencies, positional variables and weather variables obtained for each of 32 sites. "Aridity" is first weather principle component.*

Fs Lat Long Elev Tma, Tmin Prec Aridity

Tib —032 057 —O49t —049t —077 —0841 O58t —0811 Fs 005 027 023 032 —018 027 Lat 003 —050f —0611 —O66l 023 Long O681 O.59 0621 —080t O67t Elev O541 0751 —0571 067t Tmx 079t Ø.77 0941 Tmin —0.70t 091t Prec —0901 * P<005,t P

(1982), two suggested measures of collinearity Table 3 Multiple (standardised) regression coefficients for (obtained from the correlation matrix, X'X)were tibia colour and orange femoral stripe frequencies with positional variables, latitude, longitude and elevation and computed for each variable set: "condition num- with climatic variables, maximum and minimum tem- ber", or K, and a set of "variance inflation factors", perature and precipitation or VIF's (see definitions in reference). For posi- tional variables, k= 1303 and the largest VIF= Regressioncoefficients 365, and for the climatic variables, k = 1345 and Tibia colour Femoral stripe the largest VIF= 379; both sets of k's and VIF's are well below values (100 and 10, respectively) Lat 1011: —032 that would lead to estimation difficulties. Alterna- Long —1l41: 0i5 Elev 0651: <0.01 tive regression methods were, therefore, not %R2 809 163 required. Together, latitude, longitude and elevation Tmx —025 —011 Tmi,, —063t 042 account for about 81 per cent of the variation in Prec <001 001 red tibia phenotypic frequency. Each factor "cor- %R2 724 100 rected" for the other two is highly significantly contributory (table 3). Red tibia is positively 1: P<001;1: P<0001. associated with latitude and elevation and nega- tively associated with longitude. The geographic change evidently occurs (at least in the north-south distribution of red tibia phenotypic frequencies is direction) in a region bounded by Melville—Roblin shown in fig. 1; frequencies are low in the south in the north and Indian Head-Russell in the south. and west and increase in the north-east. A large Elevation, of course, is not shown on the map but

_.11 °

32 25% - — — — SASKATOON II' L 3I 30D. / I IT226 C. B. \ 22 \N ---- 20 18 •... 192l I7 R'GINA l5 A. IIC I0'—' Iz N N N Ui -S.-. - (k —S.. 4 —.5 lIp NN 2 49- _..z__1__p--. 49° - -- _L... — SASKATCHEWAN-- — — N1020 109° U.S.A. Figure 1 Sample locations (see also table 1) in southern Saskatchewan and adjoining area in Manitoba. Frequency of redtibia phenotype is given by black portion of pie diagram. Major vegetation zones (Rowe, 1983): A, short/mixed prairie; B, mid-grass prairie; C, aspen parkland; D, mixed wood. 402 W. CHAPCO AND M. J. BIDOCHKA as latitude decreases and longitude increases, ele- Table 4 Partial correlations of tibia colour frequencies with vation increases, thus accounting for the negative positional variables before and after correction for sig- sign of the simple correlation with red tibia nificant climatic factors frequency (table 2). The sign reverses when latitude Correlationcoefficient and longitude are corrected for (table 3)—would this suggest that TR frequency is expected to Position variable Before correction After correction increase with altitude for a narrow ranged moun- Lat 0861: 0661: (Aridity) tain? Given the paucity of such topography on the 0561: (Tmin) prairies, it is doubtful whether the question can Long —0821: —0611: (Aridity) ever be addressed! —0611: (Tmin) A principle component analysis of the correla- Elev 0621: 0591: (Aridity) tion matrix of climatic variables reveals that the 0611: (Tmi) first component "explains" about 84 per cent of t P<0•01, P<0001. the total climatic variation. From a consideration of the signs and magnitudes of the eigenvector coordinates: (0.37 Tmax, 0.36 Tmin, —036 Prec), the first component can be regarded as an index of account; nevertheless, the correlations remain "aridity" with large positive scores in the drier and highly significant, indicating that minimum tem- hotter south-west and large negative scores in the perature (or some factor that is related to minimum wetter and cooler north-east. The second com- temperature) is not the sole important determining ponent accounts for an additional 10 per cent of factor. Minimum temperature makes a negligible the variation and has a corresponding eigenvector: contribution to tibia colour's association with ele- (013 Tmax, l2OTmin, 135 Prec) with no evident vation. geographical interpretation. In any case, following Femoral stripe frequencies, unlike those of red Jeffers' (see Newham, 1968) suggestion that com- tibia, do not exhibit clinal patterns, despite huge ponents with eigenvalues less than one be ignored, interlocational differences. The correlation with the second (and third) component will not be latitude is significant at the 5 per cent level (table considered further. 2) but disappears with the inclusion of the other Tibia colour frequency and the main principle two positional variables (table 3). None of the component are highly negatively correlated (table correlationsinvolvingclimaticfactorsis 2). The extent to which "aridity" accounts for the significant. clinal variations is best seen from an examination of partial correlations between tibia colour and latitude, longitude and elevation, that have been Allozymicpolymorphism additionally "corrected" for by the first weather Ofthe 11 enzyme-determining loci examined in principle component (table 4). About 41 per cent 1982-83, aGpdh-1, Odh, Sod, Sdh, and Gdh were and 45 per cent of the latitudinal and longitudinal monomorphic for the same allele at all nine variation, respectively, is removed by tibia colour's sampled locations; Ldh, Mdh, Me, aGdh-2 and association with "aridity". Significant amounts of Xdh displayed some measure of genetic variability. variability, however, remain unaccounted for. The enzyme, G-6PD, was irresolvable and hence, Only 9 per cent of the elevational variability is was excluded from further consideration. Elec- accounted for by the weather component. Turning tromorphic frequency estimates are recorded in to the climatic factors themselves, each variable is table 5 for alleles (where "a" corresponds to the highly correlated with tibia colour (table 2), but "slowest running anodal allele", etc.) at each only the relationship with minimum temperature variant locus within each location. Recorded is significant as a partial correlation (table 4). values for sites, MAN, IND and PRE, are for data About 70 per cent of tibia colour frequency is obtained in 1982 or 1983, depending on sample connected with minimum temperature and this size, the larger being taken. G-tests revealed the figure increases only slightly (72 per cent) if absence of any significant inter-year variation. maximum temperature and precipitation are Males and females were pooled since there were included (table 3). The partial correlations listed no significant differences in allelic frequencies in the last column of table 4 show that sizeable between the sexes at each site. Six alleles were portions (58 per cent and 45 per cent, respectively) detected for Ldh and three alleles for each of the of the latitudinal and longitudinal variations are other loci. With respect to Ldh, alleles a and d removed by taking minimum temperature into represent newly discovered alleles; alleles b, c, e GENETIC VARIATION IN GRASSHOPPER POPULATIONS 403

Table 5 Allelic frequencies at five variable electrophoretic loci at nine samples sites. N is sample size. G statistics test for HWE and interlocational variability

Locus Allele Man Asn Orm Jam md Mel Yor Can Pre G10(8df)

Ldh a 0 0-011 0-103 0 0 0 0015 0015 0 b 0037 0054 0-006 0-010 0-010 0-016 0-015 0-015 0-017 c 0-813 0-806 0-851 0-846 0-890 0-887 0-818 0-794 0-890 e 0-037 0 0-013 0-019 0 0 0 0 0-008 d 0-113 0-113 0-117 0-125 0-100 0-097 0-152 0-176 0-076 f 0 0016 0 0 0 0 0 0 0-008 N 40 93 77 52 100 31 33 34 59 2-37 0-11 6811- 291 261 <0-01 0-01 0-23 2-19 10-4 Mdh a 0-025 0-005 0-006 0 0 0 0 0 0-008 b 0-963 0-995 0-987 1-000 0-988 0-984 0-985 1-000 0-983 c 0-013 0 0-006 0 0-012 0-016 0-015 0 0-008 N 40 93 77 52 80 31 33 34 59 — GHWE 1-01 3-46 2-25 — 2-28 2-38 2-44 1-99 4-70 Me a 0 0-056 0-081 0 0-020 0 0 0 0-008 b 1-000 0-930 0-878 1-000 0-975 1-000 1-000 1-000 0-975 c 0 0-014 0-041 0 0-005 0 0 0 0-017 N 40 36 37 13 100 10 13 16 59 0HWE — 0-25 3-58 0 4.52* — 1-35 17.1* csGpdh-2 a 0 0-032 0-057 0-019 0 0-016 0-015 0-029 0 b 1-000 0-968 0-930 0-953 1-000 0-984 0-985 0-971 1-000 c 0 0 0-13 0-028 0 0 0 0 0 N 40 94 79 53 100 31 33 34 59 GHwE — 0-70 0-01 3-28 — 2-38 2-44 1-47 — 23-01- Xdh a 0-015 0-016 0-026 0 0-043 0 0-045 0 0-037 b 0-985 0-984 0-974 0-981 0-957 0-968 0-955 0-956 0-963 c 0 0 0 0-019 0 0-032 0 0-044 0 N 33 93 77 52 47 31 33 34 27 GHwE 2-44 1-78 1-44 2-22 0-72 1-38 0-84 0-87 1-26 3-96 * P<0-05; 1- P<001. and f correspond to L', L2, L3 and L4 in Chapco Ldh is the only gene that is polymorphic in all (1983a). To date, the monogenic inheritance of samples. The average heterozygosity per locus LDH (Chapco, 1984), Me and cGPDH-2 (unpub- ranges from 3-3 per cent (MEL) to 6-9 per cent lished results) has been confirmed; XDH and (ORM), with a grand average of 45 per cent. Per MDH are likely also monogenically based, given location, the average heterozygosity is highest for the rather uncomplicated banding patterns Ldh (27.1 per cent) and lowest for Mdh (2.5 per obtained here. Heterozygotes for Ldh and Xdh cent). display five bands, which would suggest that the Intersite variation, assesses by performing G- enzymes are tetramers; the remaining three enzy- tests on all 9 x 2 (common allele versus "other" mes are probably dimeric, since three bands are alleles pooled) contingency tables, is not sig- obtained for heterozygotes. On the whole, nificant for loci, Ldh Mdh and Xdh, but is respec- genotypic frequencies (obtained by collapsing tively, significant and highly significant for loci, each data set to three genotypes) are in agreement Me and aGpdh-2. The biological significance of with Hardy-Weinberg expectations; two of the 35 these data, however, is seen from an examination calculated G-test values are significant, a finding of the original frequency data (table 5)—the com- probably due to chance. Based on the polymorph- monest allele is the same across all locations for ism criterion that the most common allele should each enzyme system; there is no instance of a locus have a frequency of less than 95 per cent, the in which a different allele is the common one in average frequency of polymorphic loci is 133 per different sample sites. Moreover, the lack of overall cent. This increases substantially to 37-8 per cent geographical differentiation is reflected in the very if the 99 per cent criterion is used. This discrepancy high average of 36 genetic identity estimates: stems from the fact that the commonest alleles for O•9992 (±0.0001). Grasshoppers sampled from the Mdh, Me, aGpdh-2 and Xdh "hover" about a nine sites behave as if they are part of one large frequency of 95 per cent. Based on either criterion, population. 404 W. CHAPCO AND M. J. BIDOCHKA

DISCUSSION inheritance have not yet been elucidated (Willey, personal communication), some features can be Ona large scale, the present investigation confirms mentioned for comparison. Most of the traits dis- Brooks' (1958) assertion that predominantly red- play geographic variation and in some cases, clinal legged M. sanguinipes occur in the eastern park- variation. Differentiation appears to be greater in land of Saskatchewan and predominantly blue- mountain populations, where gene flow is greatly legged forms occur in the grasslands, but until now restricted, than in plains populations. Traits are nothing had been known about areas of transition (partially) correlated with at least one of the fac- or the nature of the transition. The results demon- tors, latitude, longitude and altitude and in this strate that well within the aspen-parkland, an area connection, Schennum (1975) presents a variety of mixed grassland and forest (Bird, 1961), that of selection arguments, based mostly on ecogeo- slowly intergrades with the midgrass prairie graphical rules or their inverses. A large portion (Rowe, 1983), a major change in red tibia of geographical variation, however, remains unac- frequency takes place, at least in a north-south counted for, about 80 per cent on the average. Of direction and in a north-east-south-west direction particular interest is the spacial distribution of (statements about rates of change in a westerly wing colour morphs which are monomorphic at direction cannot be made owing to a lack of several population "centres". The centres are sep- samples). For example, phenotypic frequencies in arated by fairly narrow intergrade zones within Roblin and Russell, 40 km south, are 82 per cent which are found intermediate wing colours. This and 44 per cent, respectively; also pronounced is is guardedly offered along with some historical the difference between Melville (69 per cent) and data as evidence for secondary integradation fol- Indian Head (29 per cent), 75 km to the south-west. lowing allopatric differentiation, at least for this Given the species' capacity for mass movement trait. (see below), the existence of a sharp dine is rather Perhaps with good reason, but noticeably surprising. absent from the above authors' deliberations, are In marked contrast, Dearn (1981) reported a arguments based on genetic drift and gene flow. smooth latitudinal dine extending over a distance One has to be impressed with the results of Endler's of approximately 1000 km in the frequency of an (1977) gene flow—drift simulations resulting in inherited dorsal stripe morph in Phaulacridium what appear to be long lasting (500 generations in vitattum. Commonly referred to as "the wingless some cases) "stable" (boundaries fluctuate each grasshopper", P. vitattum does have a macrop- generation) geographical patterns, some clinal. terous phase which, according to Dearn (1978), Conceivably these patterns might correlate tem- ensures a high degree of gene flow. Disruptive porarily with environmental variables by chance selection coupled with latitudinally varying gene (Johnson et a!., 1969). Nevertheless, with respect flow from adjacent habitats, is offered as a possible to M. san guinipes, genetic drift cannot be regarded explanation for the trait's distribution. Gill (1979; as an important factor, for the species simply 1981d), providing several examples of color pat- occurs too abundantly (Beirne, 1972). The distribu- tern polymorphisms in a grasshopper of limited tion of the TR frequencies has therefore, to be mobility, Chorthippus brunneus, favors crypsis as discussed in the context of selection and gene flow. a selection mechanism to account for both large Reports of mass movement in this species are and small morph frequency differences that are legend (see Vickery and Kevan, 1983), but one can found over relatively short distances in Britain. only assume that the few mark—recapture studies Work on a number of related species sympatric in which only limited movement was observed with C. brunneus also led Gill (1981e) to a similar (Riegert eta!., 1954; Baldwin et al., 1958; Edwards, conclusion regarding selective predation. 1961) were performed either during non-outbreak Behaviour and body size differences were con- years or with non-migrating animals. No data on sidered important in explaining observed sexual the migratory status (there are reportedly mor- dimorphisms (there was no evidence of such phometric differences between migratory and non- dimorphism in tibia colour in M sanguinipes). migratory M. san guinipes—see Paul and Putnam, Schennum (1975) and Schennum and Wiley (1979) 1960; Beirne, 1972) were given. Reports on causes investigated geographic variation of a suite of mor- of movement (these usually centre on the availabil- phometric and colour traits in another grasshopper ity of food), role of wind, and direction of move- of low vagility, Arphia conspersa, in several popula- ment are contradictory (Corkins, 1922; Shotwell, tions in southern Colorado, northern New Mexico 1930; Parker et a!., 1955; Bird et a!., 1966; Riegert, and eastern Utah. Although the traits' modes of 1968; Beirne, 1972; Turnock, 1977); reports on GENETIC VARIATION IN GRASSHOPPER POPULATIONS 405 direction are assumed to be accurate. Of particular variation is unexplained by minimum temperature interest is Turnock's statement that there have been and therefore, other selective forces need to be easterly movements towards the parkland area and explored. Two possibilities are predation and para- the subsequent establishment of populations sitism. It is difficult, however, to see how leg colour (northerly, northwesterly and southerly move- could influence the risk of capture although it ments have also been recorded with respect to the might be worthwhile to examine stomach contents study area). How the conclusion of "establish- of known predators such as western meadowlarks ment" was reached without genetic data is enig- (Shotwell, 1930). Bidochka (1984) in a small survey matic. What is outstanding, is that despite the found no evidence of differential parasitic infec- movements, which possibly have occurred in tion with respect to tibia colour even though several directions and likely at irregular intervals, the incidence of parasites varied geographically natural selection has, at least for 30 years (since (see also Paul and Putman, 1960). In light of the Brooks' paper in 1958), sustained the general recent discovery that a nondiapause strain of eastern parkland—prairie separation. Intensive M. San guinipes possesses heteromorphic chromo- sampling over several years would be required to somes (Zhan et a!., 1984), it is possible that the determine the stability and width of the transition colour polymorphism and, perhaps, distribution, zone. Although the nature of the actual selective might be caused by linkage disequilibrium between mechanisms is unknown, the multiple regression an inversion, say, and the tibia colour locus, a and correlation analyses belie the hypothesis that hypothesis to be researched further. The finding tibia color distribution is connected to moisture of three-way interactions involving body colour, a and (summer) temperature. This appears to contra- fourth chromosomal inversion and heat stress in dict the finding that the first principle component, populations of the Australian grasshopper, Aus- interpreted as "aridity", accounts for a sizeable troicetes interioris (Nankivell, 1974) is of particular proportion of the latitudinal and longitudinal vari- interest although no attempt was made to explain ation, a quantitative result which probably under- the geographical distributions. lies the qualitative observations of earlier workers. Selection pressures not at all related to weather The principle components approach, while useful or location must be responsible for the distribution (Johnson et a!., 1969), does not identify which of of the femoral stripe allele, Ost, despite its close the contained variables is/are directly or indirectly linkage with the T locus (Chapco, 1980b). The two influencing the trait and which variables can be colour loci are neither associated intra- nor interlo- discounted. Of the three climatic variables, cationally. Significant geographical concordances minimum temperature is clearly the most contribu- among colour pattern genes have been described tory (in a statistical sense). One interpretation is in Chorthippus brunneus (Gill 1979; 1981d), in that the association is directly causal. In the areas which traits combined to produce camouflage where there is little or no snow cover, soil tem- effects appropriate for particular habitats. Some peratures can be lethal to eggs (Riegert, 1967). associations were reinforced by linkage and epis- Cold resistant polygenes have been reported in tasis, a result reminiscent of Nabours' (1950 and other organisms (e.g. Drosophila pseudoobscura; references therein) extensive work on colour- Parsons, 1973); if such genes are in linkage dis- pattern genes in several grouse locust species. equilibrium with the tibia color locus in M. Whatever the selection forces are for the Ost locus, sanguinipes populations, the polymorphism might they must be powerful enough to withstand the be explained. Pleiotrophy is also a possibility attenuating effects of gene flow. (Wright, 1978) although it is difficult to envisage The level of allozyme polymorphism in M. how there could be differential egg mortality with sanguinipes is well within the range of values for respect to an adult colour trait. A more likely most insects (excluding Drosophila; Nevo et a!., interpretation is that the association is indirect and 1984; Table 6). Heterozygosity, however, is low, that agents such as vegetation or soil factors, that as it is in Chorthippus brunneus (Gill, 1981a, covary with minimum temperature, are causal. The 1981b), resulting from high frequencies of the most species is capable of exhibiting choices with common allele (about 98 per cent of all common respect to food (Mulkern eta!., 1969) and soil type allele frequencies are greater than 0.80). By con- (Edwards and Epp, 1965); if preferences are trast, two other grasshopper species, Trimerotropis genotype-dependent, then blue-legged grasshop- graci!is (Tepper, 1979) and Ca!edia captiva (Daly pers that find themselves in parkland for instance, et a!., 1981) have average heterozygosities more in simply may not be finding suitable resources. In agreement with the norm. Stochastic forces might any case, a substantial portion of the positional account for the low variability, but this is unlikely 406 W. CHAPCO AND M. J. BIDOCHKA

Table 6 Comparison of average heterozygosity (H), polymorphism (P) and genetic identity (I) values in Melanoplus sanguinipes with those of other Orthopterans

Species Site No. Loci No. %P %H %J Author(s) Acridids M. sanguinipes 9 10 37'8 4'5 999 present Trimerotropis gracilis 1 19 632 156 — Tepper (1979) Chorthippus brunneus 42 15 33'O 3'2 99'9 Gill (1981a,b) Caledia captiva 2-5 20 320—45O 90—13'S 83'5—99'0 Daly et a!. (1981) Cave crickets Ceuthophilus gracilipes 8 26 80 20—33 84'6—98'8 Cockley et a!. (1977) Dolichopoda sp. 23 - — 11'3—28'0 — Sbordoni et aI. (1981) Troglophilus cavicola 2 17 469 8'2 985 Sbordoni et a!. (1981) T andreinii 3 17 67'O 113 928 Sbordoni eta! (1981) All insects (excluding Drosophila) — — 350 8'9 — Nevo et a!. (1984)

in C. brunneus and certainly is not the case in M. imply uniform selection coefficients. In Caledia sanguinipes with its large population numbers. captiva, four chromosomal taxa are distributed Population figures are not provided by Gill but the parapatrically in eastern Australia. Identity values rather uniform allozyme frequency distribution in for populations within taxa are somewhat lower C. brunneus (see below) argues against an explana- than those of M. sanguinipes and C. brunneus, but tion based on genetic drift. Low genetic variability are considerably higher than values between popu- as predicted by Levins (1968) and demonstrated lations of different taxa. Allozyme divergence is in several organisms by Bryant (1974), might be attributed to low gene flow and founder effects imposed by unvarying environments; a likely rather than selection (Daly et aL, 1981). Orthopteran example (table 6) is Ceuthophilus At one time, Brooks (1958) considered as major gracilipes, a cave cricket occupying relatively con- and minor "forms" the prairie and parkland popu- stant habitats (Cockley et a!., 1977) (Sbordoni et lations, and populations in forested areas north of a!. (1981) attribute differences in heterozygosity in the parkland, separate subspecies; Vickery (1979) two other cave crickets, Troglophilus cavicola (low) regarded the latter designation as premature and T andreinii (high), to differences in age of without further data. This paper demonstrates that population since establishment by founders). Nevo with respect to allozyme loci, there is no difference et a!. (1984), in their compendium of H values and between the major and minor forms; it might, life history traits, invoke Levins' hypothesis in an however, be fruitful to examine the more northerly attempt to explain why species with a low "arid populations, electrophoretically. index" have low H values (at the other extreme, species with high indices have large H values). Acknowledgements Financial support from the Natural Notwithstanding theattractability of this Sciences and Engineering Council of Canada (Grant A-0485 hypothesis, M. Sanguinipescannot be seen as an to W. Chapco) and from the Saskatchewan Research Council exemplar. The insect is not exclusively an arid (Scholarship to M. J. Bidochka) is gratefully acknowledged. species, and environmental variability, at least We thank Dr E. J. Chapco for comments on the writing. climatic, is anything but uniform (Environment Canada, 1983). Selection is clearly sustaining high allelic frequencies at each locus, but for reasons as yet unclear. REFERENCES Selection forces need not be uniform interloca- tionally to account for the near unity genetic iden- AYALA,F. J., POWELL, J. R., TRACY, M. L., MOURAO, C. A. AND tity values. Gene flow effects may be "strong" PEREz-SALAS, S. 1972. Enzyme variability in the Drosophila enough to overcome tendencies for genetic diver- wil!istoni group, IV: Genie variations in natural popula- tions of Drosophila willistoni. Genetics, 70, 113—135. gence at the enzyme loci, but as already mentioned, BALDWIN, W. F., RIORDAN, D. F. AND SMITH, R. w. 1958. Note not "strong" enough to attenuate differential selec- on dispersal of radioactive grasshoppers. Can. Ent., 90, tion effects at the visible loci. If gene flow is exten- 374—376. sive, one cannot determine whether the selection BEIRNE, B. P. 1972. Pest Insects ofAnnual Crop Plants in Canada. Part V. Orthoptera. pp. 38—56. coefficients are, in fact, homogeneous or spacially BIDOCHKA, M. .i. 1984. Genetic variation in natural populations variable. In C. brunneus, gene flow is low and of the migratory grasshopper, Melanoplus sanguinipes. M.Sc. therefore, the near unity identity values (table 6) thesis. University of Regina. GENETIC VARIATION IN GRASSHOPPER POPULATIONS 407

BIRD, R. D. 1961. Ecology of the aspen parkiand of western GAGE, S. H., MUKERJI, M. K. AND RANDELL, B.. L. 1976. A Canada in relation to land use. Canada Department of predictive model for seasonal occurrence of three grasshop- Agriculture. Res. Br. Pubi., 1066, 1—155. per species in Saskatchewan. (Orthoptera: Acrididae) Can. BIRD, R. D., ALLEN, W. AND SMITH, D. s. 1966. The responses Ent., 108, 245—253. of grasshoppers to ecological changes produced by agricul- GILL, P. 1979. Colour-pattern variation in relation to habitat tural development in Southwestern Manitoba. Can. Ent., in the grasshopper Chorthippus brunneus (Thunberg). Ecol. 98, 1191—1205. EntomoL, 4, 249—257. BRETr, C. H. 1947. Interrelated effects of food, temperature, GILL, '.1981.Heterozygosity estimates in the grasshopper and humidity on the development of the lesser migratory Chorthippus brunneus (Thunberg). Heredity, 46, 269- grasshopper, Melanoplus mexicanus mexicanus (Saussure) 272. (Orthoptera). Tech. Bull. Okla. Agric. Exp. Stn. No. T-26, GILL. P. 1981b. Enzyme variation in the grasshopper Chor- pp. 1-50. thippus brunneus (Thunberg). BioL J. Linn. Soc., 15, BROOKS, A. B.. 1958. Acridoidea of Southern Alberta, Saskat- 247—258. chewan and Manitoba (Orthoptera). Can. Ent. Supp., 9, GILL, i'. 1981c. Allozyme variation in sympatric populations of 1-92. British grasshoppers—evidence of natural selection. Biol. BRYANT, E. H. 1974. On the adaptive significance of enzyme J. Linn. Soc., 16, 83-91. polymorphisms in relation to environmental variability. GILL, P. 1981d. The genetics of colour-patterns in the grass- Amer. Natur.,108,1-19. hopper Chorthippus brunneus. Biol. J. Linn. Soc., 16, CHAPCO, w. 1980a. Inheritance of an unusual color variant in 243—259. the grasshopper Melanoplus sanguinipes. Can. J. Genet. GILL, P. 1981e. Colour-patterns and ecology of British grass- CytoL 22, '315—318. hoppers. Acrida, 10, 145-158. CHAPCO, w. 1980b. Genetics of the migratory grasshopper, GURNEY, A. B. 1962. On the name of the migratory grasshopper Melanoplus sanguinipes: Orange stripe and its association of the United States and Canada, Melanoplus sanguinipes with tibia color and red-back genes. Ann. Entomof Soc. (F.) (Orthoptera: Acrididae). Proc. BioL Soc. Wash. 75, Am. 73, 319—322. 189- 192. CHAPCO, w. 1983a. The genes, TR, Ost, p.0RandL in natural GURNEY, A. B. AND BROOKS, A. R. 1959. Grasshoppers of the populations of Melanoplus sanguinipes. Can. J. Genet. Mexicanus group, Genus Melanoplus (Orthoptera: Cytol., 25,88-92. Acrididae). Proc. U.S. NatI. Mus. 110, 1-93. CHAPCO, w. 1983b. Genetics and population genetics of grass- HEBARD, M. 1925. The Orthoptera of South Dakota. Proc. Acad. hoppers and locusts—a bibliography. Bib!.Entomo!.Soc. Nat. Sci. Philadelphia, 77, 33-155. Am., 2, 11-16. JOHNSON. C. G. 1969 Migration and Dispersal of Insects by CHAPCO, w. 1984. Inheritance of lactic acid dehydrogenase in Flight. Methuen, London. Melanoplus sanguinipes (F.) (Orthoptera: Acrididae). Ann. JOHNSON. F. M., SCHAFFER, H. E., GILLASPY, J. E. AND ROCK. Entomol. Soc. Am. 77, 11-13. WOOD, F. S. 1969. Isozyme genotype-environmental COCKLEY. D. E., GOOCH, J. L. AND WESTON, D. P. 1977. Genic relationships in natural populations of the harvester ant, diversity in cave-dwelling crickets(Ceuthophilus Pogonomyrmex barbatus from Texas. Biochem. Genet., 3, gracilipes). Evolution, 31, 313—318. 429-450. CORKINS, C. L. 1922. Notes on the migration of Melanoplus KING, B.. L. AND SLIFER, E. H. 1955. The inheritance of red atlantis Riley, in Northern North Dakota in 1920. Can. and blue hind tibiae in the lesser migratory grasshopper. Ent., 54, 1-4. J. Heredity, 56, 7-10. DALY. J. C., WILKINSON, P. AND SHAW, D. D. 1981. Reproduc- LEVINS, R. 1968. Evolution in Changing Environments. Princeton tive isolation in relation to allozymic and chromosomal University Press, Princeton, N.J. differentiation in the grasshopper Caledia captiva. Evol- MAURIELLO. D. AND ROSKOWSKI, J. i'. 1974. A reanalysis of ution, 35, 1164-1179. Vuilleumier's data. Amer. Natur., 108, 711-714. DEARN, j.M.1978. Polymorphism for wing length and colour MONTGOMERY. D. C. AND PECK, A. 1982. Introduction to pattern in the grasshopper Phaulacridium vittatum (Sjost.). Linear Regression Analysis. Wiley, New York. J. Aust. Entomol. Soc. 17, 135-137. MULKERN, G. B., PRUESS, K. P., KNUTSON, H., HAGEN, A. F., DEARN, j.M.1981. Latitudinal dine in a colour pattern poly- CAMPBELL. J. B. AND LAMBKY, .j.o.1969. Food habits morphism in the Australian grasshopper Phaulacridium and preferences of grassland grasshoppers. BulL N. Dakota vittatum. Heredity, 47, 111—119. Agr. Exp. Sta., 481, 1-48. EDWARDS. R. L. 1961. Limited movement of individuals in a NABOURS. R. K. 1950. Cytogenetics of the grouse locust Apotet- population of the migratory grasshopper, Melanoplus ox eurycaphalus Hancock. Kansas. Agric. Exp. Stn. Tech. bilituratus (Walker) (Acrididae) at Kamloops, British Bull., 67, 1-116. Columbia. Can. Ent., 93, 628-631. NANKIVELL, R. N. 1974. Interactions between inversion poly- EDWARDS, B.. L. AND EPP, H. T. 1965. The influence of soil morphisms and the colour pattern polymorphism in the moisture and soil type on the oviposition behaviour of the grasshopper Austroicetes interioris (V/bite & Key). Acrida, migratory grasshopper, Melanoplus sanguinipes 3,93—111. (Fabricius). Can. Ent., 97, 401-409. NEt. M. 1975. Molecular Population Genetics and Evolution. ENDLER, J. A. 1977. Geographic Variation, Speafication, and North-Holland Publishing Co. Amsterdam. Clines. Princeton University Press. Princeton, New Jersey. NEVO, E., BElLES, A. AND BEN-SHLOMO, R. 1984. The ENVIRONMENT CANADA 1983. Canadian climate normals evolutionary significance of genetic diversity: ecological, 1951-1980; Temperature and precipitation—Prairie prov- demographic and life history correlates. In Levin S. inces. (ed.) Lecture Notes in Biomathematics 53: Evolutionary EVERFVr, B. S. 1977. The Analysis of Contingency Tables. John Dynamics of Genetic Diversity. Springer-Verlag, New Wiley, New York. York. GAGE, S. H. AND MUKERJI. M. K. 1977. A perspective of grass- NEWHAM, B.. Nt. 1968. A classification of climate by principle hopper population distribution in Saskatchewan and inter- component analysis and its relationship to tree species relationship with weather. Env. Entomol., 6, 469—479. distribution. Forest Science, 14, 254-264. 408 W. CHAPCO AND M. J. BIDOCHKA

OAKESHOYF, J. G., GIBSON, J. B., ANDERSON, P. R., KNIBB, W. SBORDONI, V., ALLEGRUCCI, ci., CACCONE, A., CESARONI, D., R., ANDERSON, D. G. AND CHAMBERS, ci. K. Alcohol SBORDONI. M. C. AND IDE MATTHAEIS, E. 1981. Genetic dehydrogenase and glycerol-3-phosphate dehydrogenase variability and divergence in cave populations of Trog- dines in Drosophila melanogaster on different continents. lophilus cavicola and T andreinii(Orthoptera, Evolution, 36,86-96. Rhaphidophoridae). Evolution, 35,226-233. PARKER, J. R., NEWTON, R. C. AND SHOTWELL, R. L. 1955. SCHENNUM, w. E. 1975. A Geographical Analysis of Quantita- Observations of mass flights and other activities of the tive Morphological Variation in the Grashopper Arphia con- migratory grasshopper. U.S. Dept. Agr. Tech. Bull. 1109, spera. Ph.D. Dissertation. University of Illinois, Chicago. 46 pp. SCHENNUM, W. E. AND WILLEY, R. B. 1979. A geographical PARSONS, P. A. 1973. Behavioural and Ecological Genetics. A analysis of quantitative morphological variation in the Study in Drosophila. Clarendon Press, Oxford. grasshopper Arphia conspera. Evolution, 33, 64-84. PAUL, L. C. AND PUTNAM, L. ci. 1960. Morphometrics, parasites SHAW, C. R. AND PRASAD, R. 1970. Starch gel electrophoresis and predators of migrant Melanoplus bilituratus (Wik.) of enzymes—A compilation of recipes. Biochem. Genet., 4, (Orthoptera: Acrididae) in Saskatchewan in 1940.Can. 297—320. Ent., 92, 488-493. SHOTWELL, R. L. 1930. A study of the lesser migratory grasshop- RICHARDS, J. H. AND FUNG, K. 5. 1969. Atlas of Saskatchewan. per. U.S. Dept. Agric. Tech. Bull., 190, 1-34. University of Saskatchewan, Saskatoon. TEppER, C. S. 1979. Chromosomal and Electromorphic Variation RIEGERT, i'. w. 1967. Associaton of subzero temperatures, snow in the Grasshopper Trimerotropis gracilis gracilis. MS. cover, and winter mortality of grasshopper eggs in Saskat. Thesis. San Diego State University. chewan. Can. Ent. 99,1000-1003. TURNOCK, W. .j.1977.Adaptability and stability of insect pest RIEGERT, P. W. 1968. A history of grasshopper abundance poulations in prairie agricultural ecosystems. In Kulman, surveys and forecasts of outbreaks in Saskatchewan. H. M. and Chiang, H. C. (eds.) Insect Ecology Papers Memoirs of the Ent. Soc. Canada, No. 52. pp. 99. Presented in the A. C. Hodson Lectures. U Minn. Agr. Exp. RIEGERT, P. W., FULLER, R. A. AND PUTNAM, L. ci. 1954. Studies Stn. Tech. Bull., 310, 89-101. in dispersal of grasshoppers (Acrididae) tagged with Phos- VICKERY, V. R. AND McE, KEVAN, D. K. 1983. A Monograph phorus-32. Can. Ent., 86,223-232. of the Orthopteroid Insects of Canada and Adjacent Regions. ROFFEY, .s. 1970. The effects of changing land use on locusts Lyman Entomol. Mus. and Res. Lab Memoir No. 13. Vol. II. and grasshoppers. Proc. 1st. Study Conf Current and Future VICKERy, V. R. 1979. Notes on some Canadian Acrididae Problems of Acridology, London. (Orthoptera). Can. Ent., 111, 699-702. ROWE, J. S. 1983. Landscapes: A Guide to the Land Forms and WRIGHT, S.. 1978. Variability Within and Among Natural Popu- Ecology of Southern Saskatchewan. Reference Manual. Sask. lations. Vol. 4. Univ. of Illinois, Chicago Press, Chicago. Env. Public Information and Education Branch. Regina, ZHAN, T. S., PATHAK, S. AND LIANG, .t. C. 1984. Induction of Saskatchewan. G-bands in the chromosomes of Melanoplus sanguinipes (Orthoptera, Acrididae). Can. J. Genet. Cytol., 26,354—359.