This file was created by scanning the printed publication. Errors identified by the software have been corrected; however, some errors may remain. QUANTITATIVE VARIATION WITHIN AND AMONG CHEATGRASS POPULATIONS: THE ROLE OF MULTIPLE INTRODUCTIONS Stephen J. Novak ABSTRACT Cheatgrass (}Jromus tectorum) is a highly selfing annual grass species that occurs worldwide in temperate grasslands. The native range of this species extends from the Mediterranean Basin, including northern Africa, to Tibet and southern Pakistan (Pierson and Mack 1990). Cheatgrass has been introduced into many locations and was first observed in eastern North America as early as 1790 (Muhlenberg 1793) and in the Intermountain West about 1889 (Mack 1981). Previous results from electrophoretic analysis of cheatgrass show little genetic variation in North American populations (Novak and others 1991, also reviewed by Pyke and Novak, these proceedings). Regional differences in the distribution of allozyme variation as well as the presence of several novel allelic variants indicate this species has experienced a complex introduction history in North America. Some populations in the West appear to have resulted from the combining of genotypes following mu\tiple introduction of different genotypes from the native range and subsequent gene flow (Novak and others, submitted). Furthermore, allozyme data suggest a higher level of population differentiation in the native range than in the introduced range. Populations from North America partition their genetic variation evenly within and among populations, most likely the result of multiple introductions (Novak and Mack, in preparation). The objectives of this study were to determine the level and distribution of variation for quantitative traits in six native and six introduced populations of cheatgrass. I was interested in how genetic variation for these traits is partitioned within and among families and populations and among regions in this widespread selfing weed. Using a comparison of heritable variation within populations from Eurasia and North America, I asked whether there is evidence for a reduction in variation for quantitative traits due to founder effects. Alternatively, would an increase in variation in quantitative traits be detected within populations as a result of multiple introductions? Variation within and among populations of cheatgra,ss (Bromus tectorum) was determined for 35 quantitative traits. Substantial population differentiation was detected for most of the 35 traits. Phenotypic plasticity accounts for the variation seen for most of these traits. Genetic variation for these traits is usually partitioned among populations and regions. Populations from North America possessed higher levels of variation among families, suggesting the possible importance of multiple introductions in increasing quantitative variation in alien species. INTRODUCTION Genetic variation for individuals and populations can be described as residing at three distinct levels: DNA sequences, single-locus traits, and multilocus (quantitative) traits. Quantitative traits such as time to emergence, growth rate, time to reproduction, and fecundity are critical components of the life history of a species because they can influence the growth and persistence of the species (Dingle and Hegmann 1982). Venable (1984) states that studies of quantitative trait variation "... present an interesting bridge between the fields of ecology and evolution because they are directly interpretable ecologically as adaptation for survival and reproduction." Additionally, variation in quantitative traits can be genetically based or the result of phenotypic plasticity (Bradshaw 1965; Clausen and others 1940). Either mechanism allows a species to survive under variable environmental conditions (Lewontin 1957). The ability to survive in a variable or novel environment is potentially made even more difficult for invading species because founder events and population bottlenecks result in a reduction in genetic variation following immigration (Barrett and Husband 1990; Barrett and Richardson 1986; Brown and Marshall1981). The potential difficulty in avoiding extinction is accentuated for selfing-introduced species since they partition most of their genetic variability among rather than within populations (Brown 1979; Hamrick and Godt 1990; Loveless and Hamrick 1984). MATERIALS AND METHODS Plant Material-Seeds from individual maternal plants (families) were collected from a total of 12 populations from the native and introduced ranges of cheatgrass (table 1). The six native populations span the Mediterranean Region: Jordan, Turkey, Yugoslavia, France, and Morocco. Populations from North America were sampled across the introduced range of cheatgrass and in several cases their locations correspond to sites at or near putative points ofintroduction (7, 9, 10, and 12). Paper presented at the Symposium on Ecology, Management, and Restoration of Intermountain Annual Rangelands, Boise, m, May 18-22, 1992. Stephen J. Novak is a Postdoctoral Research Associate, Department of Botany, Washington State University, Pullman, WA 99164-4238. 103 Table 1-Populations from the native and introduced ranges of cheatgrass used in the analysis of quantitative varaition Range Native Introduced 1985) to partition total phenotypic variation within and between families, populations, and ranges. Intra-class correlation coefficients (t) were used to determine the amount of differentiation among families within each population (Zar 1974). Previous electrophoretic analyses revealed that outcrossing is exceedingly rare in cheatgrass: zero in the introduced range and 0.17 percent in the native range (Novak and Mack, in preparation). Given these results, I assume complete selfing and that the progeny from all families used in this study are genetically identical full sibs. Under the assumption of full-sib families the withinfamily variance is considered to be due to the environment (phenotypic plasticity), and the total genetic variation for each trait is partitioned into the family, population, and range variance components. The proportion of genetic variation among families (PGVAF), among populations (PGVAP), and among ranges (PGVAR) is reported as a proportion of the total genetic variance (Venable and Burquez 1989). These values allow the distribution of the total genetic variance to be compared at several hierarchical levels. Population 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. Amman, Jordan AJgada, Jordan DeGoreme, Turkey Cemerno, Yugoslavia Conquerac, France Touna Station, Morocco Philadelphia, PA Lexington, KY Provo, UT Smoot Hill, WA Rattlesnake Springs, WA Cache Creek, BC Greenhouse growouts were performed for one generation to reduce environmental effects on the expression of the traits to be measured (Roach and Wulff1987; Schall 1984). Seeds from field-collected plants of each population were sown into wooden flats (50 by 40 by 20 em) containing standard greenhouse potting media and placed in an unheated greenhouse. Plants were watered as needed, and the position of the flats on the greenhouse bench was rotated regularly until the plants reached maturity. Flats were rotated to minimize the effects of microclimate differences within the greenhouse on the traits being measured (Schall1984). In the spring the mature plants were harvested and placed separately in envelopes; 10 plants were chosen at random as the families for the subsequent experiment. In the fall of 1988, seeds from 10 families, each consisting of 10 individuals, from each of the 12 populations were individually weighed with a Cahn Instruments electrobalance to the nearest 0.001 mg. Seeds from all families and populations were randomly positioned in a grid in flats at a density of 550 seedslm2 and maintained under the same conditions described above. This sowing density is lower than plant densities usually observed in the field for cheatgrass (Mack and Pyke 1983) and was chosen to minimize intraspecific competition (Mitchell-Olds and Rutledge 1986). Plants were scored for 35 quantitative traits (Novak 1990). In this paper I will report on a subset of these traits, including initial seed weight, time to emergence, plant height at 30-day intervals and at harvest, time to flowering (defined as the days needed for the panicle to exert through the leaf sheath), and mean seed number per plant. Plants were harvested individually, when chlorophyll pigmentation was no longer observable within the panicle, and were ovendried (70 °C) for 48 h. Plants were stored with dessicant for 48-72 h before taking morphometric measurements and dry weight readings. RESULTS AND DISCUSSION Phenotypic means differ among populations for all quantitative traits except percent reprodu~tive biomass (%REPBIOM) and percent vegetative biomass (%VEGBIOM), indicating substantial population differentiation for most of these traits (data not shown). Initial seed weight was uniformly higher in native range populations than in introduced populations (fig. 1). Of the introduced populations, Provo, UT, had the largest initial seed weight (2. 76 mg), while only two populations from the native range, Cememo, Yugoslavia (2.85 mg) and Conquerac, France (2.74 mg), possessed comparable values. The initial seed weight does not appear to influence time to emergence in the cheatgrass populations used in this study (fig. 2). In fact, there may be an inverse relationship between initial seed weight and time to emergence since the population from both the native and introduced ranges with the largest initial seed weight also possessed the longest time to emergence (Algada, Jordan, and Provo, UT, respectively). Initial seed weight and plant height (at least at the early time intervals) does appear to possess a positive relationship. Mean plant heights at the 30-, 60-, and 90-day intervals for all populations of cheatgrass from the native range are larger than those observed for the populations from the introduced range (fig. 3). However, the mean plant height for the introduced populations at 120 days exceeds that observed for native populations. By the time of harvest, differences in plant heights between the native and introduced ranges are considerable (fig. 3). The patterns suggested by the relationship between seed weight and plant height appear to be a function of increased growth rates that are found in plants possessing large seeds. Interestingly, initial seed weight apparently has a larger influence on plant heights than on time to emergence. Time to flowering also appears to be related to large early season plant heights and generally with initial seed weight. Populations that had the fastest growth rates in Data Analysis-Analysis of variance was performed to detect trait differences among all populations and to test for family differences within each population. A nested analysis of variance was performed for all populations using the NESTED program of SAS (SAS Institute 104 5 a 25 a a g - 4 20 3 15 ~ Cl ~ ~ = 2 c 10 5 0 0 Amman Algada DoGorme Comorno Conquorac Amman Touna Sta. Algada E l: 5 25 4 20 Cl ~ a 3 Comerno Conquorac Touna Sta. Native Population Native Population Q DoGormo 15 2 10 5 0 0 Phlla. Lexington Provo Smoot Hill Phlla. Rattlesnake Cacho Cr. Lexington Provo Smoot Hill Rattlesnake Cache Cr. Introduced Population Introduced Population Figure 2-Days to emergence for native (top) and introduced (bottom) populations of cheatgrass. Different letters at the top of columns indicate significant differences (P < 0.05). Figure 1-lnitial seed weight (mg) for native (top) and introduced (bottom) populations of cheatgrass. Different letters at the top of columns indicate significant differences (P < 0.05). cheatgrass populations. The native populations generally have larger seeds, larger plant heights in the early time intervals, shorter time to flowering, and produce fewer seeds when compared to their North American counterparts. The pattern described for native range populations consistently the first populations to initiate flowering (fig. 4). The mean time to flowering for all native populations is 95.1 days; the introduced populations initiated flowering in 103.4 days. The variation in time to flowering observed for native populations circumscribes that observed for introduced populations: the Algada, Jordan, population had the shortest time to flowering (68.54 days) while the Cemerno, Yugoslavia, population had the longest (121.87 days). Time to flowering for introduced populations ranged from 93.42 days (Rattlesnake Springs, WA) to 116.14 days (Lexington, KY). A large amount of variation for mean seed number was observed in the populations used in this experiment (fig. 5). Introduced populations produced more seeds (41.9) than populations from the native range (38.8), suggesting that there is a benefit associated with delayed flowering. Indeed, the two populations from the native range with the longest time to flowering, DeQQreme, Turkey, and Cemerno, Yugoslavia, produced the largest number of seeds (41.5 and 63.4, respectively). The same trend generally held for the introduced populations, with the exception of the population from Lexington, KY. This population had the longest time to flowering in North America (fig. 4); however, it also produced the smallest number of seeds (fig. 5). Different patterns emerge when these quantitative traits are compared between the native and introduced 25 ---uE -~ 20 Cl 15 a:= 10 -; • Native Plant Height ~ Introduced Plant Height J: c 5 0 30 60 90 120 Days From Emergence Figure 3-Piant heights (em) for native and introduced populations of cheatgrass at 30 day intervals and at harvest. 105 HAR 140 70 c 120 80 t! 60 60 as % 100 l a 50 .B 40 z:I 30 E J 40 20 20 10 0 0 Amman Algada DeGorme Amman Cememo Conquemc Touna Sta. AJgada DeGorme Cememo Conquomc Touna Sta. Native Population Native Population 140 70 • 120 i 100 60 a a ~ 50 .l.i 40 E 60 :I z J 40 20 30 20 10 0 0 Phlla. Lexington Provo Smoot HID Rattlesnake Ceche Cr. Phlla. Lexington Provo Smoot HID Rattlesnake C8che Cr. Introduced Population Introduced Population Figure 4-0ays to flowering for native (top) and introduced (bottom) populations of cheatgrass. Criteria used to define the initiating of flowering are explained In text. Different letters at the top of columns indicate significant differences (P < 0.05). Figure 5-Seed numbers in native (top) and introduced (bottom) populations of cheatgrass. Different letters at the top of columns Indicate significant differences (P < 0.05). appears to be consistent with a life history adapted to extremely arid environments (Fox 1990). Plants in such environments flower quickly before the onset of summer drought, but they do so at a cost because they also produce fewer seeds. In the face of such drastic selection pressures, however, the production of even a small number of seeds ensures that these plants (and their genotypes) will be represented in the next generation. reveals a fairly even distribution of genetic variation among these hierarchical levels (among families =10 traits, among populations =15 traits, and among ranges = 10 traits). These results indicate that the majority of genetic variation for 25 of these traits is partitioned among populations and regions and suggest that genetic variation for quantitative traits in cheatgrass is distributed over an enormous geographical area (three continents). The ranges of environments and habitats from which these plants were collected are extensive and suggest that differentiation among populations and ranges may be the result of local selection. The analysis of adaptive significance for these traits is confounded because the populations were grown in the greenhouse (Venable 1984). It should be noted, however, that this was not the goal of the current study; I was interested in detecting genetic variation among and within populations and the nearly uniform greenhouse environment makes this possible (Schall1984). Hierarchical ADalysis-The nested ANOVA was used to partition the total phenotypic variance for all 35 quantitative traits into a hierarchical series (individuals, families, populations, and ranges) (data not shown). The range variance component accounts for differences among the native and introduced ranges for a trait, while the population component accounts for differences among populations. The within-population variation is divided into the family and individual variance components. For 30 of 35 of these traits the individual variance component accounts for most of the phenotypic variation for these traits. The individual variance component is assumed to be due to the environment and indicates that for these 30 traits most of the variation is due to phenotypic plasticity. The sum of the family, population, and range variance components comprises the total genetic variation for each trait. The total genetic variation was partitioned among families, among populations, and among regions and Among-Family Variance-Intraclass correlation (t) analysis was used to identify among-family differences for all 35 quantitative traits within each native and introduced population. Genetic variation within populations of selfing species is usually characterized by the amount of variation found among families of that population 106 populations was similar, and generally high. In contrast, higher than expected (for a selfing species) levels of within-population variability were detected in North American populations for both electrophoretic and quantitative traits. Both results suggest that multiple introductions can play an important role in influencing the level of genetic variation within populations of introduced plant species. Table 2-Signlficant intraclass correlation coefficients (t) for each native and introduced population of cheatgrass. A total of 35 traits were examined in each population Range Native Introduced Population Amman tvalues 25 AIgada 15 DeGoreme Cememo Conquerac Touna Station 20 8 14 13 Philadelphia Lexington Provo Smoot Hill Rattlesnake Springs Cache Creek 23 25 23 REFERENCES Allard, R. W.; Jain, S. K.; Workman, P. L. 1968. The genetics of inbreeding populations. Advances in Genetics. 14:55-131. Barrett, S. C. H.; Husband, B. C. 1990. The genetics of plant migration and colonization. In: Brown, A H. D.; Clegg, M. T.; Kahler, A L.; Weir, B. S., eds. Plant population genetics, breeding, and germplasm resources. Sunderland, MA: Sinauer: 254-277. Barrett, S. C. H.; Richardson, B. J. 1986. Genetic attributes of invading species. In: Groves, R. H.; Burdon, J. J., eds. Ecology of biological invasions: an Australian perspective. Canberra: Australian Academy of Sciences: 21-33. Bradshaw, A D. 1965. Evolutionary significance of· phenotypic plasticity in plants. Advances in Genetics. 13: 115-155. Brown, A H. D. 1979. Enzyme polymorphism& in plant populations. Theoretical and Population Biology. 15: 1-42. Brown, A H. D.; Marshall, D. R. 1981. Evolutionary changes accompanying colonization in plants. In: Scudder, G. C. E.; Reveal, J. L., eds. Evolution today. Pittsburgh, PA: Hunt Institute for Botanical Documentation, Carnegie-Mellon University: 351-363. Clausen, J.; Keck, D. D.; Heisey, W. M. 1940. Experimental studies on the nature of species. I. The effect of varied environments on western American plants. Washington, DC: Carnegie Institute. 520 p. Clay, K.; Levin, D. A 1989. Quantitative variation in Phlox: comparison of selfing and outcrossing species. American Journal ofBotany. 76:577-588. Dingle, H.; Hegmann, J. P. 1982. Evolution and genetics of life histories. New York: Springer-Verlag. 250 p. Fox, G. A 1990. Drought and the evolution of flowering time in desert annuals. American Journal of Botany. 77: 1508-1518. Hamrick, J. L.; Godt, M. J. W. 1990. Allozyme diversity in plant species. In: Brown, A H. D.; Clegg, M. T.; Kahler, A L.; Weir, B. S., eds. Plant population genetics, breeding, and germplasm resources. Sunderland, MA: Sinauer: 43-63. Imam, A G.; Allard, R. W. 1965. Population studies in predominantly self-pollinated species. VI. Genetic variability between and within natural populations of wild oats from different habitats in California. Genetics. 51:49-62. Kannenberg, L. W.; Allard, R. W. 1967. Population studies in predominantly self-pollinated species. VIII. Genetic variability in the Festuca microstachys complex. Evolution. 21: 227-240. 22 25 19 (Allard and others 1968). The mean number of significant t values in native populations was 16; however, there is considerable variability in the number of significant values for these populations (table 2). The population from Amman, Jordan, had the highest number of significant t values (25), while the population from Cememo, Yugoslavia, had the lowest (8). The range of significant t values for introduced populations is much less; the highest number of significant value was detected for the Lexington, KY, population (25) and the lowest value was detected for the Cache Creek, BC, population (19). The introduced range populations (23) had a greater mean number of significant t values when compared to native populations (16). Large amounts of within-population genetic variation (among-family differences) were detected for most of the cheatgrass populations analyzed. Considerable quantitative genetic variation has previously been reported within populations of other highly selfing plant species (Clay and Levin 1989; Iman and Allard 1965; Kannenberg and Allard 1967; Law and others 1977; Venable and Burquez 1989). There clearly are large differences in the distribution of qualitative and quantitative genetic variation within populations of selfing plants such as cheatgrass. Equally surprising was the observation that the number of traits with significant intraclass correlation coefficients (among-family variation) was higher for populations from North America. Founder effects and genetic bottlenecks associated with the introduction and range expansion of this species would be expected to reduce genetic variation in North America (Nei and others 1975); yet the opposite was observed. The higher within-population genetic variation in North American populations compared to home range populations of cheatgrass may be the result of multiple introductions. This hypothesis attributes the observed among-family variability within populations in the introduced range to the presence of different quantitative genotypes from the native range. These results for the distribution of quantitative genetic variation provide independent support for the conclusions obtained using starch gel electrophoresis. The distribution of quantitative and electrophoretic variation among 107 Law, R.; Bradshaw, A. D.; Putwain, P. D. 1977. Life history variation in Poa annua. Evolution. 31: 233-246. Lewontin, R. C. 1957. The adaptation of populations to varying environments. Cold Spring Harbor Symposium on Quantitative Biology. 22: 395-408. Loveless, M. D.; Hamrick, J. L. 1984. Ecological determinants of genetic structure in plant populations. Annual Review of Ecology and Systematics. 15: 65-95. Mack, R.N. 1981. The invasion ofBromus tectorum L. into western North America: an ecological chronical. Agro-Ecosystems. 7: 145-165. Mack, R. N.; Pyke, D. A. 1983. The demography of Bromus tectorum L.: variation in time and space. Journal of Ecology. 71: 69-93. Mitchell-Olds, T.; Rutledge, J. J. 1986. Quantitative genetics in natural plant populations: a review of the theory. American Naturalist. 127: 379-402. Muhlenberg, H. 1793. Index florae Lancastriensis. Transactions of the American Philosophical Society of Philadelphia. 3: 157-184. Nei, M.; Maruyama, T.; Chakraborty, R. 1975. The bottleneck effect and genetic variability in populations. Evolution. 29: 1-10. Novak, S. J.1990. Multiple introduction and founder effects in Bromus tectorum L.: an analysis of Eurasian and North American populations. Pullman, WA: Washington State University. 219 p. Thesis. Novak, S. J.; Mack, R.N. [In press]. Genetic variation in Bromus tectorum (Poaceae): a comparison between native and introduced populations. Heredity. Novak, S. J.; Mack, R. N.; Soltis, D. E. 1991. Genetic vari-ation in Bromus tectorum (Poaceae): population differentiation in its North American range. American Journal of Botany. 78: 1150-1161. Novak, S. J.; Mack, R.N.; Soltis, P. S. [Submitted]. Genetic variation in Bromus tectorum (Poaceae): multiple introductions and estimation of gene flow. Genetics. Pierson, E. A.; Mack, R. N. 1990. The population biology of Bromus tectorum in forests: distinguishing the opportunity for dispersal from environmental restriction. Oecologia. 84: 519-525. Roach, D. A.; Wulff, R. D.1987. Maternal effects in plants. Annual Review of Ecology and Systematics. 18: 209-235. SAS Institute. 1985. SAS User's guide: statistics. Cary, NC: SAS Institute. Schall, B. A 1984. Life-history variation, natural selection, and maternal effects in plant populations. In: Dirzo, R.; Sarukhan, J., eds. Perspectives on plant population biology. Sunderland, MA: Sinauer: 188-206. Venable, D. L. 1984. Using intraspecific variation to study the ecological significance and evolution of plant lifehistories. In: Dirzo, R.; Sarukhan, J., eds. Perspectives on plant population biology. Sunderland, MA: Sinauer: 166-187. Venable, D. L.; Burquez, A. 1989. Quantitative genetics of size, shape, life-history, and fruit characteristics of the seed-heteromorphic composite Heterosperma pinnatum. I. Variation within and among populatio~. Evolution. 43: 113-124. Zar, J. H. 1974. Biostatistical analysis. Englewood, NJ: Prentice-Hall. 718 p. 108