| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
USDA-ARS, U.S. Dairy Forage Research Center, Madison, WI 53706-1108
* Corresponding author (mdcasler{at}wisc.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: NPD, normalized phenotypic distance • PD, Euclidean phenotypic distance
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Regrowth and persistence of smooth bromegrass is limited largely by the timing of new tiller development. Development of new tillers in smooth bromegrass is largely determinant, with synchronized elevation and elongation of new apical meristems above the soil surface (Krause and Moser, 1977). Cutting or grazing before new tillers have developed sufficiently reduces regrowth and persistence (Eastin et al., 1964; Reynolds and Smith, 1962). During reproductive development, this critical time occurs from culm elongation to late heading. Cutting before culm elongation (to avoid removal of apical meristems) or well after heading (when new tillers have begun to emerge) leads to increased forage yields and persistence (McElgunn et al., 1972; Paulsen and Smith, 1968). Apical dominance in smooth bromegrass is strong until anthesis, when auxin activity declines and tillering is normally resumed (Eastin et al., 1964). Because smooth bromegrass produces true culms with elevated apical meristems upon regrowth, timing of subsequent harvests may also be critical for smooth bromegrass regrowth and persistence. Regrowth of smooth bromegrass is not closely related to carbohydrate reserves in roots and crowns (Eastin et al., 1964; Paulsen and Smith, 1969; Raese and Decker, 1966; Reynolds and Smith, 1962).
Use of smooth bromegrass in grass–legume mixtures is limited by its synchronized tiller development. First harvest in smooth bromegrass–alfalfa (Medicago sativa L.) mixtures typically occurs during the critical late-jointing phase. This suppression of smooth bromegrass regrowth potential, combined with shading from the rapidly recovering alfalfa canopy leads to rapid smooth bromegrass stand losses (Casler, 1988; Smith et al., 1973). Breeding and selection for persistence of smooth bromegrass in mixture with alfalfa under a three- or four-cut management system has been somewhat successful. Populations selected for persistence had 40% greater ground cover and 42% faster recovery after cutting than unselected cultivars (Casler, 1988). The cultivar Alpha, a product of this program, had 10% greater survival after 2 yr in mixture with alfalfa across five locations than the second-ranked cultivar (Casler, 1988; Casler and Walgenbach, 1990). Despite these successes, smooth bromegrass cultivars, including Alpha, have relatively low persistence under management-intensive rotational grazing systems (Casler et al., 1998).
In June 1995, while visiting my maternal grandparents' gravesite, I discovered a thick and vigorous stand of smooth bromegrass growing in a sod dominated by Kentucky bluegrass (Poa pratensis L.). The turf was well-managed and frequently mowed to maintain its visual appearance, suggesting that this smooth bromegrass population may have adaptive traits allowing it to survive and vegetatively reproduce under frequent defoliation. Subsequent investigations identified numerous such cemeteries in Minnesota, Wisconsin, and Iowa. Many of these cemeteries also had a wire fence or border area dominated by smooth bromegrass. Fence and border areas were unmanaged, suggesting that fence and border populations of smooth bromegrass may have been subjected to different natural selection pressures than sod populations.
The objective of this study was to characterize smooth bromegrass plants collected from sod and fence habitats of 30 well-managed rural cemeteries in Iowa, Minnesota, and Wisconsin. It is impossible to know the origin of founder plants of these fence and sod populations or to be certain that paired populations from a cemetery are of similar origin. Smooth bromegrass was used extensively in rural areas of these three states in the 1930s (Casler and Carlson, 1995) and may have been the major component of rural cemetery sods in this region. The advent of turfgrass breeding in the 1950s and the development of seed markets and seeding methods led to widespread mechanical renovation of turf areas. Many smooth bromegrass populations likely survived this renovation, resulting in remnant survivors in these rural cemeteries. Fence populations would have had a distinct advantage over sod populations because of less intensive interspecific competition and lack of mowing management. Thus, it is possible that fence and sod populations have evolved into morphologically and/or adaptively different phenotypes.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Despite all efforts to create a favorable environment for collected tillers, some plants did not survive transplanting in the greenhouse or the field. Twenty-five random plants of each population (30 sod populations and 30 fence populations) were randomly selected from among the survivors by taking the first 25 plants in each row of the clonal nursery. Two clonal ramets of each clone were transplanted to a split-split-plot randomized complete block experiment with two replicates in May 1998. Whole plots were represented by the 30 cemeteries, subplots were a single row of 25 plants from one of the two habitats (fence or sod), and sub-subplots were individual clones within each of the 60 populations. All plants were spaced on 0.9-m centers. The soil was a Plano silt loam (fine-silty, mixed, superactive, mesic, Typic Argiudolls). Clonal propagules were cylinders 10 cm in diameter, approximately 10 cm deep, and containing 10 to 15 live tillers. All transplants were watered several times immediately after transplanting and those with poor establishment were replaced with another ramet of the same clone within the first 4 wk after transplanting.
The experiment was fertilized with 112 kg N ha–1 in early spring and following the first harvest of 1999 and 2000. Plant height was measured on all plants in early June, just after all plants were fully headed. Forage from each row of 25 plants was harvested in bulk by a flail harvester and converted to a dry matter basis by a bulk dry matter sample taken from the entire field. Cutting height was 7 cm. Plant diameter was measured at the widest part of the crown immediately after first harvest. Regrowth vigor was measured as canopy height approximately 2 wk after first harvest. Forage was harvested again in September as described for first harvest.
The experiment was fertilized with 45 kg N ha–1 in early spring 2001 and 2002. Each plot was harvested with a flair harvester when the grass canopy was approximately 20 cm tall. Five harvests were made in each year, all when the canopy height was 15 to 20 cm. Cutting height and dry matter determination were as described above. The experiment was fertilized with 45 kg N ha–1 after each of the first four harvests. Total frequent-harvest forage yield was the sum of dry matter yield across five harvests (early May to late October).
Plant height, plant diameter, and regrowth vigor were subjected to ANOVA within each of the 60 populations. Clones, blocks, and years were all assumed to have random effects. Variance components for clones were estimated by equating mean squares to their expectations (Gaylor et al., 1970). Pooled clone variance components for states and habitats were created from pooled analyses of variance in which states and habitats were considered to have fixed effects. Confidence intervals for clone variance components were computed according to Milliken and Johnson (1984).
Subplot means for plant height, plant diameter, and regrowth vigor and raw data for reproductive forage yield, vegetative forage yield, and frequent-harvest forage yield were subjected to ANOVA in which states and habitats were fixed, while sites within states, blocks, and years were random. Variance components for sites within states were computed as described above. Fence vs. sod comparisons were made by contrasts. A combined ANOVA was computed for total forage yield across 4 yr, based on the total of all harvests within each year, considering total forage yield for frequent and infrequent harvest managements to be a single variable. The population x year interaction (3 df) was partitioned into population x management (1 df) and population x year/management (2 df).
The 60 population means for all six variables were subjected to principal components analysis. Euclidean phenotypic distance (PD) values were computed among all 60 populations by the formula
 |
= 1 before computation of PD. Phenotypic distances were converted to normalized phenotypic distances (NPD) by dividing each value by the mean phenotypic distance. The NPD was adapted from Smouse and Peakall (1999). |
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Within-population genotypic variance was significant (P < 0.10) for at least one trait (plant height, plant diameter, and regrowth vigor) in 59 of 60 populations (data not shown). However, for three of these 59 populations, all traits failed to show significant within-population variance at the P < 0.05 level. Thus, four populations (WI-3-sod, WI-9-fence, WI-10-fence, and IA-28-fence) appear to have very low amounts of genotypic variation. This suggests that these populations were founded by a relatively small number of genotypes or there has been considerable natural selection and mortality leading to a very small number of genotypes remaining in these populations. Many of the 25 clones of these four populations may be vegetative propagules of each other. Because smooth bromegrass can spread by rhizomes, individual genotypes have the potential to spread across a large area, resulting in multiple samples of a single clone even with a careful and systematic sampling strategy.
There were 51, 51, and 27 populations that showed significant within-population genotypic variation for plant height, plant diameter, and regrowth vigor, respectively (data not shown). There were few differences in within-population variation between fence and sod populations, partly because of low degrees of freedom. Furthermore, for plant height and diameter, there was no relationship between the variance component for fence and sod populations; highly variable sod populations were not associated with highly variable fence populations for these two traits. However, the opposite was observed for regrowth vigor; variance components for sod and fence populations were positively correlated (r = 0.37, P < 0.05). This is circumstantial evidence for disruptive selection and migration between fence and sod habitats in cemeteries with large amounts of variability. Disruptive selection, favoring different genotypes in fence and sod habitats, and migration of alleles between habitats would combine to maintain large amounts of genotypic variance within populations sampled from each habitat.
For regrowth vigor, pooled variance components were nearly identical for fence and sod populations (Table 2). However, for plant height and diameter, sod populations had generally greater genotypic variability than fence populations, suggesting the possibility that migration may be maintaining large amounts of genotypic variability in some sod populations. Migration from fence to sod populations can occur by rhizome growth or seed dispersal followed by seedling recruitment, whereas rhizome growth is the only mechanism for migration from sod to fence populations. Migration from sod to fence by rhizomes is unlikely, because all fencelines contained solid and vigorous populations of smooth bromegrass tillers. Thus, migration from fence to sod may be the mechanism maintaining larger amounts of genotypic variability for some traits in sod populations compared with fence populations.
|
External pollen may contribute to migration of alleles into the fence population, but this is likely to be a small source of genotypic variability because smooth bromegrass pollen rarely travels over large distances (Hittle, 1954; Knowles, 1969) and there were no additional sources of smooth bromegrass pollen within visual sight of these cemeteries. Furthermore, migration of alleles via pollen would require the additional step of seedling recruitment, which is unlikely in the fence habitat because of competition from existing vegetation and low light conditions at the soil surface.
Among-Population Genotypic Variability
Population x year interaction was significant for reproductive forage yield, plant height, and plant diameter (P < 0.05). However, for these three traits, the variance component for populations was 6.8 to 9.9 times higher than the population x year interaction variance component. Therefore, population x year interactions were relatively unimportant. All variance analyses were based on the expected mean squares including population x year interaction as a source of variation. All analyses of means and effects were based on means across years.
For total yield analyzed across all 4 yr, 89% of the overall population x year interaction was due to differences in management between 1999–2000 and 2001–2002. The phenotypic correlations between years (n = 60) were r = 0.89 (P < 0.01) for 1999 vs. 2000 (two harvests), r = 0.63 (P < 0.01) for 2001 vs. 2002 (five harvests), and r = 0.21 to 0.28 (0.03 < P < 0.10) for 1999 or 2000 vs. 2001 or 2002. Because there were no differences in weather patterns or fertilization levels between 1999–2000 and 2001–2002, this correlation structure is likely due to the change in harvest management. These results suggest that drastic changes in harvest frequency can significantly alter the ranking and relative differences among smooth bromegrass populations.
Populations differed for all six traits at P < 0.01. Principal components analysis resulted in the first two components that described 68% of the variability among populations (data not shown). Component 1 (50%) was largely associated with high forage yield for both harvests, tall plants, and highly spreading plants. These four variables were all positively correlated with each other (r = 0.47 to 0.74, P < 0.01). Component 2 (18%) was largely associated with high regrowth vigor and high frequent-harvest forage yield. Regrowth vigor was not correlated with vegetative forage yield, plant height, and plant diameter, and was negatively correlated with reproductive forage yield (r = –0.28, P < 0.05). Frequent-harvest forage yield was correlated only with reproductive forage yield under infrequent harvest (r = 0.29, P < 0.05). This covariance structure suggests that populations with high reproductive forage yield have partitioned insufficient carbohydrate reserves into crowns and roots where they would be needed for rapid regrowth. Following Cut 1, rapid regrowth (measured by regrowth vigor) is not correlated with extent of regrowth (measured by vegetative forage yield). Thus, high vegetative forage yield in these populations is probably achieved by different mechanisms, one of which may be rapid tillering and initial regrowth resulting in rapid leaf area development. The high correlation of vegetative forage yield with reproductive forage yield and plant height suggests a second mechanism may involve more uniform growth rates that are sustained throughout the growing season.
A plot of the first two principal components revealed a general pattern to fence and sod populations of most sites (Fig. 2) . For Minnesota and Iowa sites, fence and sod populations formed nearly distinct clusters, with the only exceptions being MN-17-fence and IA-29-sod. Sod populations generally had lower values of PRIN1 and/or PRIN2 than their respective fence population. Fence populations averaged 7.6% higher in reproductive forage yield, 9.5% higher in vegetative forage yield, 6.0% taller, 8.4% wider plant diameter, 4.7% higher regrowth vigor, and 6.9% higher frequent-harvest forage yield (Table 3). For plant height and regrowth vigor, differences between fence and sod populations were consistent across the three states. For all forage yield variables and plant diameter, the difference between fence and sod populations at Iowa sites was one-third to one-half the difference observed for Minnesota and Wisconsin sites. This resulted from a greater among-state variability for sod populations compared with fence populations, which tended to be more uniform among states. Wisconsin sites were the most variable in phenotype, particularly for sod populations, which did not tend to cluster together. Furthermore, the general pattern that PRIN1 and PRIN2 scores tended to be lower for sod populations was not observed for several of the Wisconsin cemeteries.
|
|
Clones collected from fence and sod habitats did not differ in overall phenotype; on the whole, for 1500 clones, neither habitat resulted in unique phenotypes which were not present in the other habitat (Fig. 3) . The most extreme individual-clone phenotypes for plant height, plant diameter, and regrowth vigor were found within both habitats. Differences between population means for fence and sod habitats arose from frequency shifts within the distribution of clonal phenotypes.
|
|
Third, stabilizing selection would tend to eliminate extreme (and unique) individuals from each habitat because of reduced fitness of extreme individuals at each end of the distribution. However, disruptive selection, acting to accentuate phenotypic differences between the two habitats, precludes stabilizing selection. Furthermore, extremely vigorous individuals (as measured by forage yield, plant height, or plant diameter) are unlikely to have reduced fitness in both habitats.
Fourth, habitat may have little or no effect on natural selection pressures and evolution of phenotype within these populations. The large phenotypic differences between fence and sod populations and their generally consistent level across cemeteries (Table 3; Fig. 2), combined with the strucure of NPD values (Table 4), suggest that this hypothesis is unlikely.
Normalized phenotypic distances among the 60 populations ranged from 0.02 to 5.81. The 30 fence populations tended to be phenotypically similar to each other, with a maximum NPD of 3.84 and mean NPD values for state groups ranging from 0.56 to 0.81 (upper left corner of Table 4). Conversely, the 30 sod populations tended to be more phenotypically distinct from each other compared with the fence populations, with a maximum NPD of 5.81 and mean NPD values for state groups ranging from 0.66 to 1.54 (lower right corner of Table 4). Mean NPD values for specific state groups were always higher for sod populations than for fence populations. The differential in NPD values between fence and sod pairs was greatest for Wisconsin, intermediate for Minnesota, and least for Iowa cemeteries. This reflects the pattern of principal components for which sod populations appeared to be more variable than fence populations in Minnesota and Wisconsin (Fig. 2).
Habitats with greater internal genetic variability would have greater potential to express among-site phenotypic and genetic variability because of a greater potential for the local environment to favor different phenotypes and, ultimately, different genotypes. Both selection and migration could act to maintain greater phenotypic diversity among cemeteries for sod populations compared with fence populations. As discussed above, migration likely is unidirectional from fence to sod habitats, maintaining higher potential levels of genetic variability within sod populations compared with fence populations. In addition, selection pressure is likely to be greater within sod habitats because of the extreme and potentially stressful nature of the management regime. Smooth bromegrass evolved in natural grasslands without intensive grazing pressure or frequent defoliation. Smooth bromegrass is relatively poorly adapted to frequent mowing, particularly in a competitive environment (Casler, 1988; Eastin et al., 1964; Reynolds and Smith, 1962). Most natural selection is driven by environmental stress per se or by fluctuating environmental stresses (Wright, 1932, 1949), suggesting that the most stressful and/or unstable environment has the greatest potential for selection. Selection may counteract the effects of migration within a particular cemetery sod, resulting in dominance of a relatively few highly fit genotypes. However, these dominant genotypes would likely vary among cemeteries, resulting in genetic variability among cemetery sods.
Eight of the 30 sites (WI-1, WI-3, WI-8, WI-9, WI-10, MN-13, MN-18, and MN-19) had fence and sod populations that were significantly different (P < 0.05) for four or five of the six variables measured (Table 5). Five sites (WI-2, WI-5, WI-12, IA-28, and IA-29) had fence and sod populations that were not significantly different for any of the five variables measured. Several Wisconsin sites were exceptions to the generalized phenotypic difference between fence and sod populations described in Table 3. There were two obvious exceptions from Minnesota, sites MN-17 and MN-18.
|
Population WI-9-sod had highest reproductive forage yield and the lowest regrowth vigor of all 60 populations (P < 0.01 for comparison to all other populations), and was among the highest sod populations in plant height and diameter (Table 5). The low regrowth vigor of this and several other sod populations indicates that rural cemetery sods are not a universal source of smooth bromegrass germplasm with superior regrowth potential. For regrowth vigor, 15 of the 18 lowest-ranked populations vigor were sod populations; for vegetative forage yield, 14 of the 18 lowest-ranked populations vigor were sod populations (Table 5). Within their sod habitat, these smooth bromegrass plants generally appeared to be highly vigorous, with a canopy typically 3 to 8 cm above that of the Kentucky bluegrass sod. However, under the infrequent harvest management, only one sod population exceeded its respective fence population in regrowth vigor (WI-3-sod) and no sod population exceeded its respective fence population in vegetative forage yield. Furthermore, no sod population exceeded its paired fence population in forage yield under the frequent harvest management. Thus, adaptation to frequent mowing does not necessarily confer improved regrowth potential under either a frequent or infrequent harvest management.
Sites MN-17 and MN-18 were unusual for different reasons. Population MN-17-fence ranked 29th or 30th of the 30 fence populations for reproductive and vegetative forage yield, plant height, and plant diameter (Table 5). Conversely, population MN-18-sod ranked 29th or 30th of the 30 sod populations for reproductive and vegetative forage yield, plant height, and plant diameter.
Finally, the low correlations between forage yield for infrequent (1999–2000) vs. frequent (2001–2002) harvests revealed potential adaptive differences between some sod and fence populations. Nine fence populations (WI-1-fence, WI-8-fence, WI-10-fence, MN-18-fence, MN-19-fence, MN-20-fence, IA-23-fence, IA-24-fence, and IA-25-fence) averaged 27.4% higher in reproductive forage yield and 21.9% higher in vegetative forage yield under the infrequent harvest management compared with their respective sod populations. These fence populations were similar to their respective sod populations in total forage yield under the frequent harvest management, suggesting that they were unable to express their genetic potential under the frequent harvest management.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Tolerance to frequent defoliation probably results from rapid initial regrowth following mowing. Frequent mowing may select plants that can rapidly produce new tillers with a small number of leaves, creating just enough leaf area to produce carbohydrate reserves sufficient to support tiller development following the next mowing event. When mowing is withheld from these sod populations for longer periods of time, either on a two- or five-harvests-per-year schedule, a slower long-term growth rate and/or lower photosynthesis rate is likely responsible for their relatively low regrowth vigor and regrowth forage yield. As the regrowth cycle lengthens, sod populations lag further behind fence populations, resulting in similar forage yields or reduced forage yields for the sod populations relative to the fence populations. Nevertheless, some sod populations have similar forage yield potential to fence populations under frequent-harvest management, suggesting a similar short-term growth rate. This germplasm may have potential value for the development of a multipurpose smooth bromegrass cultivar adapted to both infrequent and frequent harvest managements.
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication February 19, 2003.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in Crop Science:
This article has been cited by other articles:
|
|
 |
|
  M. D. Casler Among-and-within-Family Selection in Eight Forage Grass Populations Crop Sci., March 19, 2008; 48(2): 434 - 442. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Casler and E. C. Brummer Forage Yield of Smooth Bromegrass Collections from Rural Cemeteries Crop Sci., October 27, 2005; 45(6): 2510 - 2516. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. D. Casler Agronomic Performance of Timothy Germplasm from Forage and Turf Sods under Two Harvest Managements Crop Sci., August 26, 2005; 45(5): 1990 - 1996. [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 | |||
