| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a USDA-ARS, U.S. Dairy Forage Res. Center, 1925 Linden Dr. West, Madison, WI 53706-1108
b USDA-ARS, Univ. of Nebraska–East Campus, P.O. Box 830937, Lincoln, NE 68583-0937
c Dep. of Plant & Soil Sciences, Oklahoma State Univ., Stillwater, OK 74078-6028
d Dep. of Agronomy & Plant Genetics, 1991 Buford Cir., St. Paul, MN 55108-6026
e USDA-ARS, Northern Great Plains Res. Ctr., P.O. Box 459, Mandan, ND 58544
f Crop and Soil Sciences Dep., Univ. of Georgia, 111 Riverbend Dr., Athens, GA 30602
g Division of Plant Sciences, 108 Waters Hall, Univ. of Missouri, Columbia, MO 65211
h Dep. of Crop, Soil & Environmental Sciences, Univ. of Arkansas, 1366 W. Altheimer Dr., Fayetteville, AR 72704-6898
* Corresponding author (mdcasler{at}wisc.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
a USDA-ARS, U.S. Dairy Forage Res. Center, 1925 Linden Dr. West, Madison, WI 53706-1108
b USDA-ARS, Univ. of Nebraska–East Campus, P.O. Box 830937, Lincoln, NE 68583-0937
c Dep. of Plant & Soil Sciences, Oklahoma State Univ., Stillwater, OK 74078-6028
d Dep. of Agronomy & Plant Genetics, 1991 Buford Cir., St. Paul, MN 55108-6026
e USDA-ARS, Northern Great Plains Res. Ctr., P.O. Box 459, Mandan, ND 58544
f Crop and Soil Sciences Dep., Univ. of Georgia, 111 Riverbend Dr., Athens, GA 30602
g Division of Plant Sciences, 108 Waters Hall, Univ. of Missouri, Columbia, MO 65211
h Dep. of Crop, Soil & Environmental Sciences, Univ. of Arkansas, 1366 W. Altheimer Dr., Fayetteville, AR 72704-6898
* Corresponding author (mdcasler{at}wisc.edu).
Switchgrass (Panicum virgatum L.) is a warm-season native grass, used for livestock feed, bioenergy, soil and wildlife conservation, and prairie restoration in a large portion of the USA. The objective of this research was to quantify the relative importance of latitude and longitude for adaptation and agronomic performance of a diverse group of switchgrass populations. Six populations, chosen to represent remnant prairie populations on two north–south transects, were evaluated for agronomic traits at 12 locations ranging from 36 to 47°N latitude and 88 to 101°W longitude. Although the population x location interactions accounted for only 10 to 31% of the variance among population means, many significant changes in ranking and adaptive responses were observed. Ground cover was greater for northern-origin populations evaluated in hardiness zones 3 and 4 and for southern-origin populations evaluated in hardiness zones 5 and 6. There were no adaptive responses related to longitude (ecoregion). Switchgrass populations for use in biomass production, conservation, or restoration should not be moved more than one hardiness zone north or south from their origin, but some can be moved east or west of their original ecoregion, if results from field tests support broad longitudinal adaptation.
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Evolutionary processes including gene migration, random genetic drift, mutation, and natural selection combined with environmental variation due to latitude, altitude, soil type, and precipitation have resulted in significant genetic and phenotypic variation in switchgrass. Latitude of origin has a significant impact on productivity, survival, and adaptation traits of switchgrass (Sanderson et al., 1999; Casler et al., 2004). Two distinct cytotypes exist in switchgrass, upland and lowland (Hultquist et al., 1996). Upland cytotypes are more adapted to northern latitudes and lowland cytotypes are more adapted to southern latitudes. Furthermore, there is genetic variability for adaptation within each cytotype, both of which have northern and southern types within their geographic range (Casler et al., 2004). Growth rate, photoperiodism, heat tolerance, and cold or freezing tolerance regulate adaptation of switchgrass populations.
Adaptation of switchgrass populations has important implications for both agronomic production and prairie conservation and restoration. Agronomically, it is important to utilize germplasm that has photoperiod traits, morphological plasticity, and stress tolerances that match the environmental characteristics of a particular region (Casler et al., 2004; Boe and Casler, 2005). Photoperiod, morphological, and adaptation traits are all important in prairie restoration and conservation endeavors, to ensure that populations are phenotypically similar and well adapted to local environmental conditions. For this reason, many restoration ecologists recommend that populations be drawn only from collections made locally, although "local" is often difficult to define. USDA hardiness zones, defined in 5.5°C increments of mean annual minimum temperature, provide an excellent framework for choosing germplasm for use in agronomic breeding programs and for defining "local" conditions for restoration purposes (Casler et al., 2004). However, little is known about genetic variation that conditions the response of switchgrass populations to longitude and its environmental basis. The objective of this research was to quantify the relative importance of latitude and longitude in regulating the adaptation and agronomic performance of a diverse group of switchgrass populations.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
Populations were planted at 12 locations representing four of the USDA hardiness zones (Table 2 ; Fig. 1). A Latin square design was used at all locations. Eleven locations were planted in spring 2001 and Mandan was planted in spring 2002. Plot size and soil type for each location are listed in Table 2. Plots were seeded with a drill, rows were spaced 15 to 20 cm apart, and the seeding rate was 930 PLS m–2. Weeds were controlled by use of pre- and postemergence herbicides, which varied among locations due to local conditions, needs, and restrictions. Forage growth was harvested, but no data were collected at the end of the first growing season.
|
Biomass yield data were analyzed by nearest neighbor analysis using separate row and column covariates and ignoring all population and block effects for each location-year combination (Casler, 1999). The residuals, representing all variation due to populations and blocks, were saved for each location-year combination. The appropriate location-year mean was added to each residual to rescale the residuals to represent the raw data adjusted for spatial variation. Adjusted biomass yield values were analyzed by general linear models analysis of variance, subtracting 2 df from error for each location-year combination (Smith and Casler, 2004). All other variables were analyzed by general linear models analysis of variance, treating populations as a fixed effect and all other factors as random effects.
The population x evaluation-location interaction was partitioned into contrasts to test specific differences among populations using specific combinations of locations. First, the six populations were partitioned into five single degree of freedom contrasts: ecoregion of origin (Prairie Parkland vs. Eastern Forest), Prairie Parkland-transect linear and nonlinear, and Eastern Forest-transect linear and nonlinear. The ecoregion contrast was computed from four of the six populations, eliminating some confounding effects between ecoregion and hardiness zone (Fig. 1), using Sunburst and WS98-IP from hardiness zone 4 and Blackwell and Cave-in-Rock from hardiness zone 6. Each of these five contrasts was computed for six combinations of the evaluation locations (Table 2; Fig. 1): Prairie Parkland locations (Mandan, ND; Ames, IA; DeKalb, IL; Mead, NE; Columbia, MO; Stillwater, OK), Eastern Forest locations (Spooner, WI; Rosemount, MN; Marshfield, WI; Arlington, WI; Lancaster, WI; Fayetteville, AR), and locations within USDA hardiness zones 3, 4, 5, and 6. Second, population means at each location were scaled to eliminate the main effect of location by subtracting the location mean from each value (Casler et al., 2004). Scaled means for each population were separately regressed on latitude and longitude of each evaluation location (n = 12). All regressions were formulated as linear contrasts using coefficients computed (Carmer and Seif, 1963) and tested in the general linear models ANOVAs (Steel et al., 1996).
|
RESULTS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The mean growth stage at harvest was mid-anthesis for Blackwell and Pathfinder, late anthesis for Cave-in-Rock, and postanthesis for the other three populations, corresponding to a range in heading date of 9 d. Maturity at harvest was highly and consistently correlated with heading date at all 12 locations (mean r = –0.83 ± 0.06), so the maturity rating was excluded from all further data analyses and presentations.
Ground cover measurements, taken twice per year for 3 yr, had an autoregressive correlation structure in which measurements made following shorter time intervals were more highly correlated than measurements made following longer time intervals. Averaged over locations, the mean phenotypic correlations of ground cover at 40 mo postplanting with the other ground cover measurements were: r = 0.61 ± 0.11 for 12 mo, r = 0.72 ± 0.06 for 16 mo, r = 0.63 ± 0.12 for 24 mo, r = 0.81 ± 0.06 for 28 mo, and r = 0.95 ± 0.02 for 36 mo. Based on these results and their consistency across locations, ground cover at approximately 40 mo was used in all further analyses and presentations.
The three populations originating in the Prairie Parkland ecoregion were later in heading than the three populations originating in the Eastern Forest ecoregion (Table 3 ). This difference was similar for trials at Prairie Parkland and Eastern Forest locations, and there were no consistent trends for variation in this effect measured on a north-south transect across hardiness zones 3 to 6. The ecoregion effect accounted for an average of 24% of the variation among the six populations.
|
The Prairie Parkland populations had greater dry matter concentration than the Eastern Forest populations for all but one of the location groups (Table 4 ). This effect accounted for an average of 19% of the variation among the six populations. There were distinct trends toward a greater effect measured at Prairie Parkland locations and at the more southern locations.
|
The Prairie Parkland populations produced 3.4 to 7.2% greater biomass yield for all six of the evaluation-location groups (Table 5 ). Each of these differences was significant, except for HZ3 (P = 0.09), but this was only due to the low precision associated with one location within this hardiness zone. These effects were highly uniform across the range of location groups and accounted for an average of 13% of the variation among the six populations.
|
Populations originating in the Prairie Parkland were significantly greater (P < 0.01) in ground cover than populations originating in the Eastern Forest for all six location groups (Table 6 ). This effect accounted for 17 to 56% of the variation among populations and was generally uniform across location groups.
|
|
|
|
|
DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Latitude and Hardiness Zones
The populations chosen for this study represented two latitudinal transects, one encompassing hardiness zones 4, 5, and 6 through the Prairie Parkland region (Bailey's ecoregions 332, Great Plains Steppe, and 252, Prairie Parkland), and the other encompassing hardiness zones 3, 4, and 6 through the Eastern Forest region (Bailey's ecoregions 221, Laurentian Mixed Forest, and 222, Eastern Broadleaf Forest Continental; Bailey, 1997, 1998). Both transects were characterized by significant levels of phenotypic variability associated with linear responses to latitude of origin. These responses were simple and generally uniform for heading date, dry matter concentration, and biomass yield, but were complex for ground cover, suggesting differential adaptation of populations.
Populations originating from southern latitudes were later in heading than populations originating from northern latitudes, a result that was consistent with previous studies (McMillan, 1959, 1965; Casler, 2005). Similarly, populations originating from southern latitudes tended to have lower dry matter concentration at harvest than populations originating from northern latitudes, also consistent with previous observations (McMillan, 1965). The relationship between heading date and dry matter concentration was consistent for the Eastern Forest locations, but did not exist for the Prairie Parkland locations. Switchgrass populations from the Prairie Parkland ecoregion are more heterogeneous than other populations (McMillan and Weiler, 1959) and this region appears to be an important center of diversity for this species (McMillan, 1959). Reduced genetic variability or historical genetic bottlenecks (changes in the population due to reduced population size) in Eastern Forest populations may have resulted in a fixed relationship between heading date and dry matter concentration. Larger populations and greater genetic diversity in the Prairie Parkland may support a greater diversity in physiological responses to environmental factors such as photoperiod and temperature, resulting in populations that may be later in heading, but not necessarily lower in dry matter concentration. The overall mean for dry matter concentration was highest at the Prairie Parkland locations and relatively little variation was observed among populations (Table 4). Hot, dry, windy weather during late summer or early autumn may have conditioned this response.
Populations originating from southern latitudes tended to have greater biomass yields than populations originating from northern latitudes. Although this effect was not observed uniformly across all evaluation locations, it was observed for both the Eastern Forest and the Prairie Parkland population transects and it is similar to results from two previous studies (Sanderson et al., 1999; Casler et al., 2004). In the Southern Great Plains of the USA, biomass yield is largely a function of plant height (Redfearn et al., 1997; Casler et al., 2004). The longer growing season favors plants with later heading dates and an ability to retain photosynthetically active leaf area longer through the growing season (as indicated by lower dry matter concentration), resulting in more phytomers (more leaves) and taller plants compared to northern genotypes (McMillan, 1964; 1965; Casler et al., 2004; Boe and Casler, 2005). Northern populations grown at southern latitudes tend to flower early and mature more rapidly, reducing their ability to take advantage of the longer growing season. Conversely, southern populations grown at northern latitudes remain vegetative for a longer period of time, flowering later and allowing them to take advantage of longer days and produce higher biomass yields (Newell, 1968a). The linear decline in dry matter concentration of Blackwell (origin in hardiness zone 6) with increasing latitude, relative to the other populations, was consistent with these observations.
McMillan (1959) hypothesized that three gene pools, or primary gene distribution centers, were responsible for repopulation of the tall-grass prairies after the retreat of the Pleiostocene glaciers. A western montane population, originating at southern latitudes, but higher altitudes would have rapidly migrated north, filling the northern ecological zones with early-heading plants capable of sexual reproduction under long-day conditions in a short growing season. A southern population, originating in Texas and Oklahoma, would have possessed considerable genetic variability for photoperiod response, rapidly migrating north throughout the Great Plains and evolving a range of photoperiod and temperature responses as it migrated to higher latitudes and colder climates. A southeastern population would have been responsible for filling ecological niches within the various Eastern Forest ecosystems. Hybridization and mixing along contact lines would have occurred frequently and may be partly responsible for much of the genetic variability observed in both chromosome number and morphological traits within many prairie-remnant switchgrass populations. McMillan's theory sufficiently explains the existence of latitudinal variability for relatively simple traits such as heading date, dry matter concentration, and biomass yield, that are largely photoperiodic and consistent along latitudinal transects through two distinct ecoregions, the Prairie Parkland and the Eastern Forest.
This was not the case for ground cover. Both latitudinal transects showed distinct adaptive responses for ground cover, suggesting that natural selection is an important factor regulating survival of switchgrass plants and populations. Populations originating from northern latitudes were higher in ground cover, measured in hardiness zones 3 and 4, compared to populations originating from southern latitudes. Similarly, WS98-IP, originating in hardiness zone 4, increased linearly in ground cover relative to the other populations as latitude of evaluation location increased. Blackwell and Cave-in-Rock, originating in hardiness zone 6, both decreased linearly in ground cover relative to the other populations as latitude of evaluation location increased, consistent with results of Berdahl et al. (2005). Conversely, populations originating from northern latitudes were lower in ground cover, measured in hardiness zones 5 and 6, compared to populations originating from southern latitudes.
A previous study, based on 5 of the 12 locations utilized in the current study, demonstrated southern adaptation of lowland populations vs. northern adaptation of upland populations, and that northern- and southern-adapted populations can be distinguished within both lowland and upland cytotypes (Casler et al., 2004). The results of this previous study were confirmed by the current study for upland populations of switchgrass, confirming that prairie-remnant populations from hardiness zones 5 and 6 are better adapted to more southern latitudes, while populations from hardiness zones 3 and 4 are better adapted to more northern latitudes, as measured by ground cover 40 mo after planting. Natural selection for survivorship within switchgrass populations is most likely controlled largely by photoperiod and perhaps disease resistance at southern locations, favoring plants that can respond to shorter days with later heading, delayed moisture loss, and an extended photosynthetically active period. It is likely that southern populations also have greater heat tolerance than northern populations. At northern locations, natural selection for survivorship is likely controlled by cold or freezing tolerance. Most switchgrass mortality occurs during winter months (Casler et al., 2002; unreported data from the current study).
Longitude and Ecoregions
This is the first study of switchgrass populations specifically designed to test populations for differential adaptation to longitude or ecoregions as defined by Bailey (1998). Other studies have evaluated switchgrass populations across a more limited longitudinal gradient of evaluation locations and found strong (Hopkins et al., 1995a) or weak (Hopkins et al., 1995b) population x location interactions. In the current study, all four variables were characterized by relatively simple plastic responses, indicating that longitude or ecoregion is not a major factor in regulating adaptive responses of switchgrass.
Prairie Parkland populations were consistently later heading than Eastern Forest populations, a result that is inconsistent with observations made by McMillan (1959), perhaps owing to different germplasm samples. However, four of the six populations had significant linear responses to longitude of the 12 evaluation locations. Two Eastern Forest populations became earlier in heading, relative to the other populations, as they were moved west, while two Prairie Parkland populations became later in heading as they were moved west. These responses resulted in a wide range of heading dates at Prairie Parkland locations, compared to eastern locations. Responses for dry matter concentration were similar in nature to those for heading date, but not as uniform, frequent, or extreme. For both of these traits, it is not possible to differentiate whether these responses resulted from Prairie Parkland populations adapting to the eastern environments, Eastern Forest populations adapting to the western environments, or a combination of the two. Nevertheless, the plasticity of these two traits indicated a clear lack of stability across ecoregions.
Prairie Parkland populations had greater biomass yield and ground cover regardless of the evaluation locations. In contrast, a previous study, which included five Prairie Parkland cultivars and one Eastern Forest cultivar (Cave-in-Rock), demonstrated a population x location interaction that could be considered adaptive in nature (Casler and Boe, 2003). Cave-in-Rock ranked highest in biomass yield and ground cover in southern Wisconsin, but fifth in biomass yield and sixth in ground cover in eastern South Dakota. Furthermore, the correlation between the two locations was r = 0.67 for biomass yield of all six populations, a value that increased to r = 0.97 when Cave-in-Rock was removed from the data set. Cave-in-Rock consistently ranks highest in biomass yield relative to cultivars from the Prairie Parkland ecoregion when evaluated in eastern Canada (Madakadze et al., 1998, 1999).
However, the majority of literature supports a lack of adaptive responses associated with longitude or ecoregion. Numerous soil and edaphic factors that differ between the Prairie Parkland and Eastern Forest ecoregions have been investigated as environmental factors that might explain population x location interactions of switchgrass. Soil texture, soil pH, soil cation exchange capacity, soil N availability, populations of arbuscular mycorrhizal fungi, and precipitation or moisture availability during the growing season were all important factors discriminating among locations in one or more switchgrass studies (Nixon and McMillan, 1964; Hopkins and Taliaferro, 1997; Brejda et al., 1998; Cassida et al., 2005; Lee and Boe, 2005). In each case there was no evidence for genotype x environment interaction or interactions could not be described by any of the soil or edaphic factors discriminating the different environmental conditions. Because there is genetic variation for transpiration efficiency in switchgrass (Byrd and May, 2000), there may also be genetic variation for drought tolerance, which would likely have an impact on population x location interactions of field studies over a larger geographic area, such as westward into the rain shadow of the Rocky Mountains.
The lack of differential adaptation of eastern- vs. western-origin switchgrass populations suggests that McMillan's theory of three gene pools or centers of deployment following retreat of the Pleistocene glaciers must have been followed by extensive homogenization among the three gene pools (McMillan, 1959, 1964). The principal forces promoting homogenization would be hybridization at population boundaries and gene migration, the latter occurring by wind-borne pollen transport or seed transport by birds, mammals, and humans. The forces promoting homogenization appear to be considerably stronger than any forces promoting adaptive responses to longitude or ecoregions.
Synthesis: Regional Gene Pools
Switchgrass breeding programs and the seed industry have combined to create a seed marketing and distribution system that encourages and facilitates movement of switchgrass seeds across large regional areas. Due to low profit margins, the seed industry favors high-volume cultivars, which are more likely to result from germplasm that is broadly adapted across multiple ecological zones. Furthermore, there are very few cultivars developed from germplasm that originates east of the Mississippi River, leading to a heavy reliance on a small number of eastern cultivars and a broad distribution of cultivars from the Great Plains (historical Prairie Parkland ecoregion).
Our research has validated the Plant Adaptation Region proposal of Vogel et al. (2005) as fully compatible with empirical agronomic, adaptation, and stability of switchgrass populations. Plant Adaptation Regions, combining hardiness zones (Cathey, 1990) with ecoregions (Bailey, 1997, 1998), are the functional units that define adaptation of switchgrass populations, incorporating photoperiod, average minimum temperature, historic vegetation, and regional soil type into an effective germplasm classification system for both cultivated and natural germplasm.
Some populations are broadly adapted beyond their Plant Adaptation Region of origin, such as Cave-in-Rock, which is adapted to hardiness zones 5, 6, and 7 throughout ecoregion Province 251 and east through most ecoregions to the Atlantic Ocean (Madakadze et al., 1998, 1999; Vogel, 2004). This broad adaptation indicates that cultivars such as Cave-in-Rock have a genetic composition that provides for a robust responsiveness to environmental variables, a valuable characteristic for use in both livestock and feedstock production systems. Germplasm with broad adaptation potential also will be very valuable for long-term conservation plantings if predicted climatic changes occur.
As fossil fuel reserves become more depleted and their effects on Earth's atmosphere become more prominent, the need for renewable and cleaner energy sources increases. Switchgrass will be an essential component of a new paradigm in sustainable energy production systems in North America. In the USA alone, 16 million ha of productive farmland is set aside every year for conservation purposes (Perlack et al., 2005). This land, combined with many more hectares of marginal cropland, could support bioenergy production from switchgrass, providing many additional socioeconomic benefits derived from permanent grassland (Paine et al., 1996; Vogel, 1996, 2004). Our research indicates that a comprehensive program of developing switchgrass cultivars for use in bioenergy production will require some level of regional breeding in North America (Sanderson et al., 2006). Efforts should continue to focus on large-scale regional testing of new candidate cultivars as the only means of identifying broadly adapted cultivars such as Cave-in-Rock and the limits to their adaptation range. Breeding efforts should continue to focus on regional gene pools, defined by hardiness zones, gathering germplasm from throughout the region, conducting selection in representative environments, and developing large networks of collaborators to support field trials throughout the target hardiness zones.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Received for publication December 12, 2006.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. M. Tobias A Genome May Reduce Your Carbon Footprint The Plant Genome, March 1, 2009; 2(1): 5 - 8. [Full Text] [PDF]
|
|
|
|
 |
|
 K. P. Vogel and R. B. Mitchell Heterosis in Switchgrass: Biomass Yield in Swards Crop Sci., November 24, 2008; 48(6): 2159 - 2164. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. M. Martinez-Reyna and K. P. Vogel Heterosis in Switchgrass: Spaced Plants Crop Sci., July 1, 2008; 48(4): 1312 - 1320. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Casler and E. C. Brummer Theoretical Expected Genetic Gains for Among-and-Within-Family Selection Methods in Perennial Forage Crops Crop Sci., May 1, 2008; 48(3): 890 - 902. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
