sign in sign up
<<
Home , Bromus tectorum

Comparing Phenotype and Fitness of Native, Naturalized, and Invasive Populations of Downy Brome (Cheatgrass, tectorum)

C. Lynn Kinter Richard N. Mack

Abstract: The Eurasian grass downy brome (cheatgrass, Bromus causes loss of native , and increases fire fre- tectorum L.) was introduced into arid and semiarid bunchgrass and quency, , and siltation. In contrast, in the central shrub communities in and over 100 part of New Zealand’s South Island, downy brome is natural- years ago, but has strikingly different histories in these ranges. On ized at low levels in shrub and bunchgrass communities New Zealand’s South Island, it persists at low levels, while in (Connor 1964; Williams 1980). There, it persists primarily Western North America, it dominates vast areas. Inherent high on disturbed sites, such as roadsides or trampled areas, but fitness in founder populations may have contributed to the invasive does not become dominant over large acreages. character of North American populations, while low fitness may These performance differences might typically be attrib- have precluded a New Zealand invasion. In four common green- uted to environmental differences between the two intro- house environments, 62 populations from Western North America, duced ranges. However, in the region of New Zealand where New Zealand, and the native range in Western were downy brome has persisted for decades, the physiognomies compared for 15 phenotypic and fitness traits (including height, of the native communities, climate, and recent fire and days to flowering, vegetative ). Significant differences among grazing history are similar to those features in Western source locations were evident for most traits. North American North America where downy brome has invaded heavily populations were typically most vigorous, followed by European, (Mack 1986; Wardle 1991). Even two species that commonly and lastly New Zealand populations. North American flow- attack downy brome in its native range—head smut fungus ered earlier, and New Zealand plants later, than those from the (Ustilago bullata) and bird cherry-oak aphid (Rhopalosiphum native range. Differences in levels of fitness between the sources of padi)—have been known for at least a century in both founder populations have likely contributed to radically different introduced ranges (Lowe 1961; Pergande 1904). Conse- histories of downy brome in two of its new ranges. Our research has quently, we hypothesized that differences in performance of important implications for screening and predicting invaders and downy brome in these new ranges may be explained by controlling . differences in the founder genotypes, rather than differ- ences in the environments in the new ranges. We assessed the potential for different performance among downy brome from three ranges: the portion of its native Downy brome, or cheatgrass (Bromus tectorum), is a range in Western Europe that served as the donor region for cleistogamous, annual C grass native to Eurasia and north- 3 both North America and New Zealand (Novak and Mack ern . It was introduced from Europe (Novak and Mack 2001), and the new ranges in Western North America and 2001) into both North America and New Zealand over 100 the South Island of New Zealand. By growing all populations years ago (Kirk 1869; Mack 1981), but has radically differ- in common greenhouse environments, we investigated ge- ent histories in the new ranges. In Western North America, netically based differences in phenotype, vigor, and fitness it dominates vast areas formerly dominated by shrubs and among these three ranges. bunchgrasses (Mack 1981). On these , it changes the structure of the community, often comprising over 90 percent of the vegetative cover. It decreases forage values, Methods ______Sample Collection

In: Hild, Ann L.; Shaw, Nancy L.; Meyer, Susan E.; Booth, D. Terrance; In the native range, we based our collecting efforts on McArthur, E. Durant, comps. 2004. and dynamics in shrubland Novak and Mack’s (2001) allozyme research. They found ecosystems: proceedings; 2002 August 12–16; Laramie, WY. Proceedings RMRS-P-31. Ogden, UT: U.S. Department of Agriculture, Forest Service, that the widespread genotypes in Western North America Rocky Mountain Research Station. were most closely related to those in Europe, and less closely C. Lynn Kinter is a Botanist and Ph.D. Candidate, and Richard N. Mack related to genotypes in Eastern North America, , and is a Professor of Botany, School of Biological Sciences, Washington State University, Pullman, WA 99164-4236. FAX: (509) 335-3184; phone: (208) 373- Northern Africa. They deduced that the cheatgrass geno- 4381; e-mail: [email protected] types, which were the early founders that spread widely

18 USDA Forest Service Proceedings RMRS-P-31. 2004 Comparing Phenotype and Fitness of Native, Naturalized, and ... Kinter and Mack across Western North America, were from Central and dried at room temperature for 16 weeks. Throughout the Western Europe, probably around Slovakia, Czech Repub- study, any extraneous plant material or collection bags lic, or the eastern part of Germany. For this study, we were autoclaved before disposal to prevent escape of ge- sampled heavily in that region, as well as north into Sweden netic material. and west across The Netherlands, where Novak and Mack While the plants were actively growing, we recorded days did not sample but where many immigrants and shipments to emergence, height of the tallest leaf every 30 days follow- left for the New World. We also included samples from Spain ing emergence, width of the tallest leaf at 60 and 120 days and Italy. In the native range, downy brome is found at low after emergence, tiller number at 60 days after emergence, levels in native plant communities, such as stabilized dunes, and days from emergence to flowering. At senescence, we and on disturbed sites, such as railroad beds, farmsteads, recorded days from flowering to senescence, number of and cereal grain fields. panicles, and first internode length of the tallest panicle. In Western North America, we sampled three populations After harvest and drying, first internode diameter of the each in the vicinity of six founder locations identified by tallest panicle, aboveground vegetative biomass, panicle Novak and Mack (2001)—from and north to biomass, and presence of glume hairs were recorded. Three British Columbia. In New Zealand, founder information is traits were also recorded before were sown: weight of not known, so we sampled three populations in the region 50 seeds, seed length, and awn length. around Lake Pukaki and Bendigo (Canterbury and Otago Counties) where herbarium specimens have been collected Statistical Analyses since the 1930s. In total, we sampled 62 populations: 41 in Europe, 18 in Western North America, and three in New To assess trait differences among ranges, we compared Zealand. From each population, we randomly collected 30 responses using Multivariate Analysis of Variance individuals, and ultimately chose six of those at random for (MANOVA) and Analysis of Variance (ANOVA) in Proc GLM use in experiments. (SAS 2000) with a two-stage nested design: population nested within range, and six replicates nested within popu- Common Greenhouse Experiments lation. Both population and range were fixed effects. Data from each environment were analyzed separately due to In December 2000, we planted seeds from six individuals small but significant range by environment interactions. We from each of 62 populations in four common greenhouse conducted a multivariate analysis of seven of the response environments: control, low water, low nutrients, and low variables (height, leaf width, and tiller number at 60 days; temperature. These seeds were one generation grown out culm diameter, vegetative biomass, panicle biomass, days from the wild to minimize maternal effects (Roach and from emergence to flowering) using Canonical Discriminant Wulff 1987; Schaal 1984). No mortality occurred during the Function Analysis (SAS 2000). All European and New grow-out phase, and all parent plants produced numerous Zealand populations were included, but only the geographi- seeds for experiments. In the experimental phase, seeds cally central population from each of the six North American were sown in standard potting medium (60 percent peat, 20 founder sites was included to facilitate interpretation of the percent pumice, 20 percent sand; pH 6.9–7.0) in individual scatterplot by reducing the amount of overlap in symbols. In 15-cm fiber pots on greenhouse benches under ambient all comparisons, differences were determined to be signifi- light. Temperatures were 24/16 ∞C (12h/12h) for 6 weeks cant when P < 0.05. after sowing, followed by a 9-week cold period at 4/2 ∞C (10h/14h) to ensure vernalization. Plants in the control environment were watered daily to container capacity, and Results and Discussion ______fertilized weekly with Peters 20:20:20 delivered in line We found significant differences among ranges for nearly during watering. Each of the remaining experimental en- every trait assessed in each of the four environments. For vironments differed from the control in only one factor. example, in all but the early cold treatments, plants from Plants in the low-water environment were watered to North America were significantly taller, those from New container capacity when the soil moisture content of 10 Zealand were shorter, and those from Europe were inter- randomly chosen pots had fallen 50 percent. Plants in the mediate (fig. 1). Similarly, North American plants had wider low-nutrients environment received only the nutrients leaves, New Zealand plants had narrower leaves, and Euro- initially in the potting medium and trace minerals in tap pean plants were intermediate (fig. 2). For aboveground water. Three weeks after the seeds were sown, plants in the vegetative biomass at senescence, North American plants low-temperature environment were placed in an adjacent had higher values and New Zealand plants had lower values greenhouse room for a cold period of 4/2 C (10h/14h) that ∞ than did the European plants in each treatment; however, was 3 weeks longer than in the other environments. With differences were not significant in every comparison (fig. the exception of plants in the low-temperature environ- 3). The small number of populations from New Zealand ment during these 3 weeks, plants were arranged in a (n = 3) reduced test power and significant differences in completely randomized design, and rotated and re-ran- some cases. domized weekly to minimize block and bench effects. Water For most of the remaining phenotypic traits we found and fertilizer treatments continued until all plants had significant differences among ranges in each experimental senesced. Each plant’s aboveground biomass was har- environment. In each of the four environments, North Ameri- vested when chlorophyll pigmentation in the panicles was can plants flowered earliest and New Zealand plants latest no longer visible—167 to 259 days after emergence—then (fig. 4)—147 days for North America, 151 for Europe, and

USDA Forest Service Proceedings RMRS-P-31. 2004 19 Kinter and Mack Comparing Phenotype and Fitness of Native, Naturalized, and ...

Control Control 50 North America 9 North America Europe Europe New Zealand 8 New Zealand 40

7 30 6

20 5

10 4

Low temperature Low temperature 50 9

8 40

7 30 6

20 5

10 4

Low nutrients Low nutrients 50 9 Height (cm)

Leaf width (mm) 8 40

7 30 6

20 5

10 4

Low water Low water 50 9

8 40

7 30 6

20 5

10 4 30 60 90 120 150 60 120 Days after emergence Days after emergence Figure 1—Mean height (cm) of downy brome Figure 2—Mean leaf width (mm) of downy populations from North America, Europe, and brome populations from North America, New Zealand at 30-day intervals after Europe, and New Zealand at 60-day intervals emergence in each of four test environments. after emergence in each of four test environ- Error bars indicate standard error. ments. Error bars indicate standard error.

20 USDA Forest Service Proceedings RMRS-P-31. 2004 Comparing Phenotype and Fitness of Native, Naturalized, and ... Kinter and Mack

found less phenotypic variability among North American 14 and New Zealand populations than European populations, as would be expected with the restricted genetic variability 12 following founder events based on a small number of intro-

10 duced genotypes (Barrett and Husband 1990). In analyzing all phenotypic traits simultaneously using 8 Canonical Discriminant Function Analysis in SAS, we found a a ab b b that the centroids for each range were significantly different 6 b from each other at P < 0.0001 (fig. 5). In each of the four a 4 b ab environments, the introduced phenotypes are essentially a a subset of the native phenotypes. North American pheno- b 2

Aboveground vegetative biomass (g) c types most closely match those from Austria, Slovakia, and Southeastern Germany, while New Zealand phenotypes 0 Control Low Low nutrients Low water mostly closely match those from Northwestern Germany and The Netherlands—areas that lie geographically to the Figure 3—Aboveground vegetative biomass for northwest of the North American matches. downy brome populations from North America While plants from the three ranges clearly differ in pheno- (black), Europe (gray), and New Zealand (white) in type and vigor, fitness differences may occur as well. We each of four environments. Error bars indicate found differences in mortality during a single time period standard error. Within environments, differing letters when plants in the low water treatment became extra dry in indicate significant differences (P < 0.05) among a sudden spell of very hot days. North American plants had ranges. a significantly lower death rate, 8.33 ± 2.66 percent, com- pared to European and New Zealand plants at 22.95 ± 2.69 percent and 27.78 ± 10.60 percent, respectively. A set of experimental stress tests under different environments may show that North American plants typically have higher 160 survival. They may also be more water use efficient, and we plan to compare water use efficiency among the ranges using 140 carbon isotope analysis. To further assess fitness, we are conducting seed germina- 120 tion trials and cleaning seeds from panicles to get a measure c of seed biomass produced in the four treatments. We will use b 100 a b b c c these two measures to calculate relative fitness values for a a b a b each population. Although these data are still being col- 80 lected, some insight may be found in a previous analysis we conducted in a single environment with one North American 60 Emergence to flowering time (days) population from Smoot Hill, WA, and 10 populations from Central and Western Europe. In this pilot study, the North 40 Control Low Low nutrients Low water American plants ranked second for relative fitness between populations from first-ranked Bratislava, Slovakia, and Figure 4—Time from emergence to flowering for third-ranked Vienna, Austria. Plants from these two Euro- downy brome populations from North America pean populations were among those that displayed pheno- (black), Europe (gray), and New Zealand (white) in types most similar to those from North America in the four each of four environments. Error bars indicate greenhouse environments, and which were most genetically standard error. Within environments, differing letters similar based on allozyme data (Novak and Mack 2001). If indicate significant differences (P < 0.05) among the pattern found in the pilot study corresponds to the larger ranges. test in four environments, we expect plants from Western North America to rank among those with highest fitness from the native range. 156 days for New Zealand—when averaged across the four environments. Not only are these differences statistically significant, they are likely to be ecologically significant, Conclusions______because previous work (Mack and Pyke 1983, 1984; Rice and others 1992) has shown that a few days difference in - To summarize our findings, plants from the invasive ing time can allow seed filling in arid environments. range in North America were typically largest and most Though population-level variability is not the key focus of vigorous, followed by those from the native European range, this study, we note that for nearly every trait assessed, we and lastly those from the naturalized New Zealand range. found significant differences among populations within a These and other phenotypic differences we have documented single range. This parallels findings by other researchers among ranges indicate that genetic differences, rather than working on Western North American populations and mea- environment alone, explain many performance differences suring traits such as biomass and seed (Meyer between native and introduced populations. Our findings and Allen 1999; Rice and Mack 1991a,b). Additionally, we highlight the importance of studying representative genotypes

USDA Forest Service Proceedings RMRS-P-31. 2004 21 Kinter and Mack Comparing Phenotype and Fitness of Native, Naturalized, and ...

Can2 Control e

e

Z Z e 2 Z Z e e e A A e e e eZ Z e e eAA e AA e e e e e e ZZ e e eee e eAA e 1 eZZ e Z e eA e eZ e e e e e e AA A e e eZ e e e e A ee e Z e A e e Z eAA e A e e A A ee e e e eee e Z ee e e e eeeeeeAA e eA 0 e ee eZ e ee e ee e e e ee ee e e A e e eee e e eee e eA ee e e e eA e e e e e ee e e e e ee e ee e A e e eee e ee ee A -1 ee e ee e eee e e e e e e ee e e e e e e e e ee e e e e e e e e e e e e e A -2 e A e e e e

e

-3 A

-3 -2 -1 0 1 2 3 Can 1

Figure 5—Canonical Discriminant Function Analyses of seven phenotypic measurements of downy brome in the control environment. Centroids of each range are indicated by large letters: New Zealand (Z), Europe (e), and North America (A). from across a species’ range in screening for invasive poten- Acknowledgments ______tial, testing for effectiveness of and biocontrol agents, or similar studies. These results have ramifications We thank J. R. Alldredge, M. A. Evans, and M. S. Minton for for quarantine regulations because, at present, an intro- statistical assistance; R. S. Bussa, C. A. Cody, M. A. Darbous, duced species is often not considered to be a threat if it has M. Erneberg, S. K. Foster, G. K. Radamaker, B. D. Veley, and not been a problem in the past. Our study illustrates that the K. D. Walker for technical assistance; and J. Baar, L. Barkan, outcome of an introduction can be very different based on C. E. Hellquist, A. Klaudisova, N. L. Mack, H. C. Prentice, random genetic sampling in the native range. Serious eco- R. R. Pattison, K. D. Rode, T. Sebastia-Alvarez, A. Tudela, logical consequences may have occurred in New Zealand if and L. J. van der Ent for seed collection assistance. We ac- favorable on the South Island had gained plants knowledge helpful discussions with R. A. Black, L. D. Hufford, from the populations that we unfortunately gained in West- M. Morgan, S. J. Novak, P. S. Soltis, and M. S. Webster. Fund- ern North America. ing sources include Betty W. Higinbotham and Noe Higinbotham Research Awards, Hardman Foundation Award,

22 USDA Forest Service Proceedings RMRS-P-31. 2004 Comparing Phenotype and Fitness of Native, Naturalized, and ... Kinter and Mack

Sigma Xi—Scientific Research Society Grant-in-Aid-of-Re- Meyer, S. E.; Allen, P. S. 1999. Ecological genetics of seed germina- search, and Washington State University College of Sciences tion regulation in Bromus tectorum L. I. Phenotypic variance among and within populations. Oecologia. 120: 27–34. and School of Biological Sciences. Novak, S. J.; Mack, R. N. 2001. Tracing plant introduction and spread: genetic evidence from Bromus tectorum (cheatgrass). Bioscience. 51: 114–122. References ______Pergande, T. 1904. On some of the aphides affecting grains and grasses of the . USDA Division of Entomology Barrett, S. C. H.; Husband, B. C. 1990. The genetics of plant Bulletin. 44: 5–23. migration and colonization. In: Brown, A. H. D.; Clegg, M. T.; Rice, K. J.; Black, R. A.; Radamaker, G.; Evans, R. D. 1992. Kahler, A. L.; Weir, B. S., eds. Plant population genetics, breed- Photosynthesis, growth, and biomass allocation in ecotypes ing, and genetic resources. Sunderland: Sinauer Associates. of cheatgrass (Bromus tectorum). Functional . 6: 32–40. Connor, H. E. 1964. Tussock communities in the Mackenzie Rice, K. J.; Mack, R. N. 1991a. Ecological genetics of Bromus Country, South Canterbury, New Zealand. New Zealand Journal tectorum. 1. A hierarchical analysis of phenotypic variation. of Botany. 2: 325–351. Oecologia. 88: 77–83. Kirk, T. 1869. On the naturalized plants of New Zealand, especially Rice, K. J.; Mack, R. N. 1991b. Ecological genetics of Bromus with regard to those occurring in the Province of Auckland. tectorum. 3. The demography of reciprocally sown populations. Transactions Proceedings of New Zealand Institute. 2: 131–146. Oecologia. 88: 91–101. Lowe, A. D. 1961. The cereal aphid in wheat. New Zealand Wheat Roach, D. A.; Wulff, R. D. 1987. Maternal effects in plants. Annual Review. 8: 21–26. Review of Ecology and Systematics. 18: 209–235. Mack, R. N. 1981. Invasion of Bromus tectorum L. into Western SAS. 2000. Version 8.1. Cary, NC: SAS Institute, Inc. North America: an ecological chronicle. Agro-Ecosystems. 7. Schaal, B. A. 1984. Life-history variation, natural selection, and Mack, R. N. 1986. Alien plant invasion in the : maternal effects in plant populations. In: Dirzo, R.; Sarukhan, J., a case history. In: Mooney, H. A.; Drake, J. A., eds. Ecology of eds. Perspectives on plant population ecology. Sunderland: Sinauer biological invasions of North America and . : Associates: 188–207. Springer-Verlag: 192–213. Wardle, P. 1991. Vegetation of New Zealand. Cambridge University Mack, R. N.; Pyke, D. A. 1983. The demography of Bromus tectorum Press. L.: variation in time and space. Journal of Ecology. 71: 69–93. Williams, P. A. 1980. Vittadinia triloba and Rumex acetosella Mack, R. N.; Pyke, D. A. 1984. The demography of Bromus tectorum: communities in the semi-arid regions of the South Island. New the role of microclimate, predation and disease. Journal of Ecol- Zealand Journal of Ecology. 3: 13–22. ogy. 72: 731–748.

USDA Forest Service Proceedings RMRS-P-31. 2004 23