Forage yield of smooth bromegrass collections from rural cemeteries.
Regrowth and persistence of smooth bromegrass is limited 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). Regrowth is reduced by cutting or grazing before new tillers have developed sufficiently, eventually reducing 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 on 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).
The first widespread use of smooth bromegrass in the USA occurred during the drought of the 1930s when it was an important component of hay, pasture, and conservation plantings (Casler and Carlson, 1995). Remnants of these plantings can be found in rural areas throughout the central USA. Smooth bromegrass can persist in the soil in the form of seed or rhizomes, potentially leading to long-term persistence of remnant populations from plantings made in the 1930s.
Rural cemeteries are another source of smooth bromegrass germplasm that likely occurs as remnants of plantings from the 1930s. Many rural cemeteries of the North Central USA are characterized by a Kentucky bluegrass (Poa pratensis L.) sod that is well maintained by members of a local church or cemetery association. Smooth bromegrass often survives in both the cemetery sod and the area surrounding the cemetery. In many cases, the fence or border population of smooth bromegrass is unmanaged, creating two visually distinct habitats for smooth bromegrass: a frequently mowed sod and an uncut fence or border area. When compared in a common nursery, fence and sod populations from many cemeteries are phenotypically similar to each other, suggesting that they represent a single population that has not been modified by habitat management (Casler, 2004). Migration from the fence to the sod, either by rhizomes or seed, may also contribute to maintenance of genotypic diversity, but a similar overall phenotype in fence and sod populations. In other cases, there is clear phenotypic divergence between fence and sod populations, suggesting the possibility that natural selection may be responsible for a degree of genetic differentiation between them (Casler, 2004). Natural selection pressures appeared to be greater in sod populations than in fence populations, resulting in greater among-cemetery variability for sod vs. fence populations (Casler, 2004).
Because sod plants can only reproduce by rhizomes, natural selection pressure in the sod habitat would favor genotypes with a greater tolerance for frequent defoliation and greater long-term survivorship. Preliminary evidence for this was found in the observation that, for a limited number of cemeteries, the fence population had considerably higher forage yield than the sod population under infrequent harvest, but fence and sod populations were equal in forage yield under frequent harvest (Casler, 2004). The objective of this experiment was to quantify, describe, and test the responses of paired fence and sod populations to different harvest frequencies.
MATERIALS AND METHODS
Smooth bromegrass plants were collected from 30 cemeteries in Minnesota, Wisconsin, and Iowa in 1995 and 1996. Details of the collection protocol and location of the cemeteries are provided by Casler (2004). Smooth bromegrass plants were collected only from cemeteries with the following characteristics: (i) a well-managed turf, dominated by Kentucky bluegrass, with few obvious weeds and showing no evidence of infrequent or lax mowing management, (ii) a reasonably vigorous stand of smooth bromegrass in the sod, and (iii) a good stand of uncut smooth bromegrass in the fence or border area. Plants were collected from the fence and sod of each cemetery, creating discreet populations of smooth bromegrass plants. Each population was phenotypically similar to the southern (steppe) type of smooth bromegrass.
In 1999, 25 random plants of each population (one habitat of one cemetery) were cloned into four replicates and transplanted into one of 60 isolated crossing blocks at Arlington, WI. Each crossing block contained 25 clones of a population and four replicates in a randomized complete block design with a 0.6-m plant spacing. Each crossing block was 10 m from the adjacent crossing block and winter rye (Secale cereale L.) was planted as a pollen barrier between all crossing blocks in autumn 1999. Seed was harvested from each plant in July 2000, threshed, cleaned, and bulked in equal volumes for all plants within a crossing block. Seed was tested for germination using standardized methodology in February 2001 (AOSA, 1998).
In April 2001, three separate experiments were planted at Ames, IA, and Arlington, WI. The soil types were Plano silt loam (fine-silty, mixed, mesic Typic Argiudoll) at Arlington and Nicollet loam (fine-loamy, mixed, mesic Aquic Hapludoll) at Ames. The seeding rate was 600 pure live seeds [m.sup.-2], which was equivalent to an average seeding rate of 22 kg [ha.sup.-1]. Each experiment was designed as a randomized complete block with four replicates. Plot size was 0.9 x 1.5 m with five drilled rows. Five check cultivars were also included in each experiment.
The three experiments were designed for three harvest managements, originally designated as two harvests per year (anthesis and post-killing-frost), three harvests per year (early heading, 30-cm canopy, and post-killing-frost), or four harvests per year (30-cm canopy at each harvest date). Because of mild drought in 2002 and severe drought in 2003, harvest frequencies were compressed to the following three treatments, spanning 2002 and 2003: four harvests (as originally planned), five harvests (three as planned in 2002 but only the first two harvests in 2003), and six harvests (Arlington: four as planned in 2002 but only the first three in 2003; Ames: the first three harvests in 2002 and the first two harvests in 2003). The six-harvest treatment represents an average number of harvests over the two locations (seven at Arlington and five at Ames). Because the growth stages and timing of the harvests were similar between the two locations, they differed largely in the presence or absence of the later harvests when the effects of drought were most severe.
Nitrogen fertilizer was applied early in the spring and immediately after each harvest (except the last harvest within each year) at the following rates: 112 kg N [ha.sup.-1] (four-harvest treatment), 75 kg N [ha.sup.-1] (five-harvest treatment), and 56 kg N [ha.sup.-1] (six-harvest treatment). Nitrogen rates were originally intended to be equal, on a season-total basis, across the three harvest managements, but loss of some harvests on the two most frequent harvest managements eliminated some of the nitrogen applications. This created a small amount of confounding between harvest managements and total nitrogen application rates, resulting in less nitrogen applied as the frequency of harvest increased. However, because cultivar x nitrogen-fertilization-level interactions are biologically unimportant in smooth bromegrass (Fortmann, 1953; Offutt and Hileman, 1972), this confounding should not affect the interpretation of population x harvest management interactions.
Each experiment was harvested according to the harvest management described above, with a flail harvester at Arlington and a sickle-bar harvester at Ames. Because all populations had similar maturity on any given harvest date, dry matter was determined on a bulk sample of harvested forage from 14 plots and a single dry matter value was used to adjust all plot biomass values to a dry-matter basis. We justified this procedure on the basis of previous work that demonstrated little or no genetic variation for maturity or dry matter content of smooth bromegrass (Casler et al., 2000) and the fact that each block could be harvested within 30 to 40 min, minimizing any dirunal changes in dry matter content. Samples were dried at 60[degrees]C before dry matter determination. Although ground cover (persistence) was one of the intended variables for measurement, there was no observable loss of ground cover for the duration of the experiment in any of these populations.
Forage yield data for each harvest were analyzed by nearest neighbor analysis to adjust each plot value for spatial variation, using the two-covariate, preadjustment-by-harvests method of Smith and Casler (2004). Spatial adjustment decreased effective error mean squares by 0 to 107% (average of 24%) and the spatial covariates accounted for 0 to 31% (average of 10%) of the plot-to-plot variability. Regrowth percentage was computed for each plot in each growing season using adjusted plot forage yield values. Regrowth was defined as total forage yield for all harvests following the first harvest within a growing season.
Spatially adjusted season total forage yield and regrowth percentage were analyzed by analysis of variance in which all factors (years, locations, harvest managements, and populations) were assumed to have fixed effects, except replicates. Years were considered fixed because of age of stand and dominance of drought effects. Locations were considered fixed because they each represented one site within the geographic range of the Iowa and Wisconsin cemetery collection sites. The main effect of populations was partitioned into sources of variation describing the structure of the populations: habitat, (1 df), cemetery (29 df), and habitat x cemetery (29 df). The interactions of populations with locations and years were similarly partitioned.
The population x harvest management interaction was partitioned into single-degree-of-freedom contrasts to measure the linear effect of harvest management on each population by the methods of Hill and Baylor (1983). For each of the 30 cemeteries, three orthogonal contrasts were computed: (i) the linear regression of fence population means on the number of harvests over the 2-yr period, (ii) the linear regression of sod population means on the number of harvests over the 2-yr period, and (iii) the linear response x habitat interaction, which measures the difference in slope between (i) and (ii). Linear regressions on the number of harvests over the 2-yr period were also computed for each of the five cultivars.
RESULTS AND DISCUSSION
Population x year interactions were always considerably smaller than population x location interactions and population main effects. While some population x year interaction terms were statistically significant, interpretation of results did not change between years, indicating that these interactions were of little biological significance. Therefore, all results are presented as means over 2 yr. Only two portions of the population x location interaction were significant, the location x state-of-origin and the location x site within state-of-origin interactions. Habitat did not interact with location. Results are presented as means over locations or within locations, depending on the relative importance of these portions of the population x location interaction. Higherorder interaction terms were not significant.
Forage yield declined with increased harvest frequency, by -1.35 [+ or -] 0.15 Mg [ha.sup.-1] [harvest.sup.-1] at Arlington (P = 0.04) and by -3.17 [+ or -] 0.04 Mg [ha.sup.-1] [harvest.sup.-1] at Ames (P < 0.01). These linear responses accounted for over 99% of the variability in mean forage yield across harvest managements. These responses were largely a result of the significant reductions in first-harvest forage yield associated with the earlier harvest date and the lower N applications in early spring of the more frequent harvest managements: -2.06 [+ or -] 0.01 Mg [ha.sup.-1] [harvest.sup.-1] at Arlington (P < 0.01) and -2.59 [+ or -] 0.05 Mg [ha.sup.-1] [harvest.sup.-1] at Ames (P < 0.01). The average dates of first harvest were 26 May (jointing), 6 June (early heading), and 12 June (anthesis) for the 4-, 5-, and 6-harvest treatments, respectively (SE = 2 d). At the rates of nitrogen applied in this study, which were similar for the entire season across harvest managements, smooth brome grass was incapable of recovering lost forage yield potential associated with earlier spring harvest.
The forage yield responses to harvest management were reflected in changes in regrowth percentage across harvest managements, which increased by 21.2 [+ or -] 2.6% units [harvest.sup.-1] at Arlington (P = 0.04) and 5.6 [+ or -] 1.1% units [harvest.sup.-1] at Ames (P = 0.06). The lower response of regrowth percentage to increasing harvest frequency at Ames indicated that plots under a more frequent harvest management were much less capable of responding with increased regrowth percentage at Ames compared to Arlington. This may be an indication that the drought experienced at both sites in both years was more severe at Ames than at Arlington. In both years at both locations, drought severity increased through the growing season, becoming more pronounced after the summer solstice. Thus, regrowth forage yields were more severely affected by drought than first-harvest forage yields.
Cemetery populations derived from the three states of origin were similar in forage yield at Arlington (5.68-5.87 Mg [ha.sup.-1]) but differed at Ames (Table 1). At Ames, Iowa populations had 8.2% higher forage yield than those from Wisconsin and Minnesota (9.74 vs. 9.00 Mg [ha.sup.-1]; P < 0.01). This result suggested that populations collected in Iowa may be better adapted to Ames than to Arlington. Because most of the Iowa cemeteries were clustered around Ames, this may be evidence of an adaptive response as populations evolved at these cemeteries. While fence populations may be relatively stable because of low or nil immigration rates and lack of obvious natural selection pressures, sod populations are subject to immigration and natural selection (Casler, 2004). It is possible that natural selection pressures on smooth bromegrass populations at Iowa cemeteries may have resulted in adaptive shifts to some factor characterizing the local environment, but it is not realistic at this time to speculate on the identity or nature of this factor.
Perennial grasses are capable of adaptive genetic shifts in response to local environmental conditions. Natural selection pressures on perennial grasses can create differential adaptive changes across distances as short as 1 m (Snaydon, 1970). Population differentiation can result from spatial variation in defoliation frequency, soil type, soil nutrient levels, and the incidence of plant pathogens (Casler, 2004; Snaydon, 1987; Snaydon and Davies, 1972). Soil and/or climatic factors may have caused adaptive changes in remnant cemetery populations of smooth bromegrass, resulting in adaptive responses to local environments. Phenotypic plasticity, the ability of a plant to utilize different growth habits and/or strategies in response to its local environment (Sultan, 1987), is partially responsible for the differential phenotype of fence and sod populations in their original habitats. However, the presence of genotypic variability indicates that natural selection and/or differential origin of some populations are important phenomena contributing to differentiation of smooth bromegrass cemetery populations.
Fence populations had an average forage yield 5.5 % higher than sod populations, a difference that was fairly consistent across harvest managements, test locations, and state-of-origin (Table 1). This difference accounted for 17% of the variation in forage yield among the 60 populations. There was a general trend for forage yield of fence and sod populations to converge as harvest frequency increased at Ames, perhaps reflecting reduced genotypic variability associated with drought-suppression of regrowth forage yields at Ames. However, taken as a whole, these results provide strong evidence that smooth bromegrass populations derived from cemetery sods have reduced forage yield potential compared with populations derived from fence rows surrounding these cemeteries. This result is similar to that observed in preliminary evaluations of the parents of these populations in which it was demonstrated that sod plants had shorter heights, narrower crown diameter, and lower forage yield (Casler, 2004). The consistency of these results across most cemetery sites suggests the presence of selection pressures toward a more prostrate sod phenotype and/ or a more upright and vigorous fence phenotype. If differential origin of fence and sod populations was important, this effect would result in more-or-less random or unpredictable differences between fence and sod phenotypes in these uniform experiments.
Cemetery sites accounted for 62% of the variation in forage yield among the 60 populations. Mean forage yield ranged from 6.64 to 8.15 Mg [ha.sup.-1] among the 30 cemetery sites. Most of these values were within the range of variability for the three cultivars with the lowest mean forage yield (Peak, Radisson, and Rebound with means of 8.01, 8.06, and 7.65 Mg [ha.sup.-1], respectively) but lower than the means for Alpha and Lincoln at 8.45 and 8.31 Mg [ha.sup.-1], respectively. Alpha and Lincoln are consistently high in forage yield among smooth bromegrass cultivars (Casler et al., 2000). These results indicate that the majority of germplasm collected from these cemeteries is relatively low in forage yield potential and it may require many years of selection and breeding to increase its forage yield potential to be competitive with the best cultivars available.
Results for regrowth percentage were less consistent than for total forage yield (Table 2). Differences between fence and sod populations were highly inconsistent, accounting for less than 1% of the variability in regrowth percentage among the 60 populations, suggesting that it is not possible to generalize differences in regrowth percentage between habitats. Cemetery sites accounted for 56% of the variation in regrowth percentage among the 60 populations. Mean regrowth percentage ranged from 33.0 to 38.7% among the 30 cemetery sites, with most sites falling within the range of cultivar means (33.9-37.0%). Populations from Minnesota had the highest regrowth percentage compared to populations from Iowa and Wisconsin (36.1 vs. 34.9%; P < 0.01).
Variation in linear responses to harvest management made up 65% of the harvest management x population interaction for forage yield (P < 0.01). Averaged over locations, fence populations ranged from -3.03 to -1.71 Mg [ha.sup.-1] harvest a and sod populations ranged from -2.52 to -1.52 Mg [ha.sup.-1] [harvest.sup.-1] in their response to increasing harvest frequency (Table 3). Cemetery sites accounted for 39% of the variation among the 60 populations, which was reflected in a positive correlation between fence and sod responses (r = 0.27, P < 0.05). Habitat accounted for only 4% of this source of variation, reflecting a small, but statistically significant effect (-2.34 vs. -2.18 Mg [ha.sup.-1] [harvest.sup.-1] for fence vs. sod, respectively, P < 0.01).
Nevertheless, the large amount of variation among cemetery sites indicated that such generalizations cannot be extended to each population or site. For nine of the 30 cemeteries, the linear response of forage yield to harvest frequency differed between habitats at P < 0.05 (Table 3). For seven of these nine sites, the sod population had a greater slope (value closer to zero) than the fence population (Fig. 1). The net result for these seven sites was a superiority of the fence population under the four-harvest management and a gradual convergence of responses to the six-harvest management, with a reversal of ranking occurring at Sites 6 and 18. These results indicate that, for seven cemetery sites, the sod population was more stable across the three harvest managements. Several of these sod populations had linear responses that were higher (closer to zero) than for all or most of the five cultivars, most notably for Sites 15 and 18. These results further suggest that natural selection is likely a more important phenomenon than differential origin of fence and sod populations, which would have resulted in more random or less predictable differences between fence and sod populations.
[FIGURE 1 OMITTED]
Variation in linear responses to harvest management made up 77% of the harvest management x population interaction for regrowth percentage (P < 0.01). Fence populations ranged from 7.6 to 16.3% units [harvest.sup.-1] and sod populations ranged from 10.6 to 18.3% units [harvest.sup.-1] in their response to increasing harvest frequency (Table 3). Cemetery sites accounted for 45% of the variation among the 60 populations, which was reflected in a positive correlation between fence and sod responses (r = 0.32, P < 0.05). Habitat only accounted for 7% of this source of variation, reflecting a small, but statistically significant effect (12.8 vs. 14.0% units [harvest.sup.-1] for fence vs. sod, respectively, P < 0.01).
As observed for forage yield per se, the variability among cemetery sites was large and significant. For 11 of the 30 cemeteries, the linear response of regrowth percentage to harvest frequency differed between habitats at P < 0.05 (Table 3). For nine of these 11 sites, the sod population had a greater slope than the fence population (Fig. 2). Thus, for these nine sites, the sod population responded more favorably to harvest frequency, with a greater increase in regrowth percentage than the respective fence population. Numerous sod populations and a small number of fence populations had linear responses of regrowth percentage to harvest frequency that numerically exceeded that of Radisson, the cultivar with the highest linear response for regrowth percentage. For four sites (4, 6, 10, and 18), the sod population had a more favorable response than the fence population to harvest frequency for both forage yield and regrowth percentage. A greater increase in regrowth percentage with increasing harvest frequency would be expected for sod populations, resulting from their more prostrate growth habit, if natural selection is responsible for differentiation between fence and sod populations.
[FIGURE 2 OMITTED]
Differential responses to harvest frequency of parents (Casler, 2004) and progeny populations (Table 3; Fig. 1 and 2) were manifested as more stable and uniform forage yield across harvest frequencies, i.e., a greater tolerance to frequent harvest in some sod populations and less tolerance to frequent harvest in some fence populations. These responses indicate that there are natural selection pressures in rural cemeteries which have resulted in adaptive changes to smooth bromegrass remnant populations. Defoliation, particularly frequent defoliation that occurs in these cemetery sods, creates a stressful environment for perennial grasses (Harris, 1978; Snaydon, 1987). Perennial grasses, including smooth bromegrass, can undergo relatively rapid genetic changes in response to frequent defoliation (Falkner and Casler, 2000; Vaylay and van Santen, 1999).
Previous work suggests that sod populations are subject to selection pressure for survival under the frequent mowing management of rural cemeteries (Casler, 2004). This selection pressure appears to have resulted in some genetic shifts toward a more stable forage yield across harvest managements, most likely as a result of mortality of genotypes intolerant of frequent mowing. Conversely, fence populations are likely subject to selection pressure in the opposite direction, manifested as mortality of genotypes better adapted to a frequent mowing regime. This disruptive selection to differential management appears to be responsible for divergence of fence and sod populations at some rural cemetery sites. This divergence is balanced by migration from the fence population to the sod population, most likely by direct rhizome ingression along the boundary between the fence and sod and/or by seed dispersal and seedling recruitment under sparse cover or following the establishment of new graves. At some cemetery sites, characterized by homogeneous fence and sod populations, migration and selection appear to be balanced in a quasi-equilibrium state. Some of these sod populations, particularly those with higher forage yield and regrowth percentage under the highest harvest frequency, represent a valuable germplasm resource for improving tolerance to frequent defoliation in commercial smooth bromegrass.
Association of Official Seed Analysts. 1998. Rules for seed testing. AOSA, Beltsville, MD.
Bittman, S., and D.H. McCartney. 1994. Evaluating alfalfa cultivars and germplasms for pastures using the mob-grazing technique. Can. J. Plant Sci. 74:109-114.
Casler, M.D. 2004. Variation among and within smooth bromegrass collections from rural cemeteries. Crop Sci. 44:978-987.
Casler, M.D., and I.T. Carlson. 1995. Smooth bromegrass, p. 313-324. In R.F Barnes et al. (ed.) Forages. Vol. I. An introduction to grassland agriculture. Iowa State Univ. Press, Ames, IA.
Casler, M.D., D.J. Undersander, C. Fredericks, D.K. Combs, and J.D. Reed. 1998. An on-farm test of perennial forage grass varieties under management intensive grazing. J. Prod. Agric. 11:92-99.
Casler, M.D., K.P. Vogel, J.A. Balasko, J.D. Berdahl, D.A. Miller, J.L. Hansen, and J.O. Fritz. 2000. Genetic progress from 50 years of smooth bromegrass breeding. Crop Sci. 40:13-22.
Eastin, J.D., M.R. Teel, and R. Langston. 1964. Growth and development of six varieties of smooth bromegrass (Bromus inermis Leyss.) with observations on seasonal variation in fructosan and growth regulators. Crop Sci. 4:555-559.
Falkner, L.K., and M.D. Casler. 2000. Genetic shifts in smooth bromegrass under grazing: Changes in nutritional value and preference for surviving vs. original genotypes. Grass Forage Sci. 55:351-360.
Fortmann, H.R. 1953. Responses of varieties of bromegrass (Bromus inermis Leyss.) to nitrogen fertilization and cutting treatments. New York (Cornell) Exp. Stn. Memoir 322.
Harris, W. 1978. Defoliation as a determinant of the growth, persistence and composition of pasture, p. 67-85. In J.R. Wilson (ed.) Plant relations in pastures. CSIRO, East Melbourne.
Harrison, T., and J.T. Romo. 1994. Regrowth of smooth bromegrass (Bromus inermis Leyss.) following defoliation. Can. J. Plant Sci. 74: 531-537.
Hill, R.R., Jr., and J.E. Baylor. 1983. Genotype x environment interaction analysis for yield in alfalfa. Crop Sci. 23:811-815.
Krause, J.W., and L.E. Moser. 1977. Tillering in irrigated smooth bromegrass (Bromus inermis Leyss.) as affected by elongated tiller removal, p. 189-191. In E.W. Jahn and H. Thons (ed.) Proc. 13th Intl. Grassl. Congr., Leipzig, Germany. 18-27 May. Akademie-Verlag, Berlin.
Lawrence, T., and R. Ashford. 1969. Effect of stage and height of cutting on the dry matter yield and persistence of intermediate wheatgrass, bromegrass, and reed canarygrass. Can. J. Plant Sci. 49:321-332.
McElgunn, J.D., D.H. Heinrichs, and R. Ashford. 1972. Effects of initial harvest date on productivity and persistence of alfalfa and bromegrass. Can. J. Plant Sci. 52:801-804.
Offutt, M.S., and L.H. Hileman. 1972. Effect of nitrogen, phosphorus, and potassium on yield and composition of smooth bromegrass. Arkansas Agric. Exp. Stn. Bull. 776.
Paulsen, G.M., and D. Smith. 1968. Influences of several management practices on growth characteristics and available carbohydrate content of smooth bromegrass. Crop Sci. 8:375-379.
Paulsen, G.M., and D. Smith. 1969. Organic reserves, axillary bud activity, and herbage yields of smooth bromegrass as influences by time of cutting, nitrogen fertilization, and shading. Crop Sci. 9: 529-534.
Raese, J.T., and A.M. Decker. 1966. Yields, stand persistence, and carbohydrate reserves of perennial grasses as influences by spring harvest stage, stubble height, and nitrogen fertilization. Agron. J. 58:322-326.
Reynolds, J.H., and D. Smith. 1962. Trend of carbohydrate reserves in alfalfa, smooth bromegrass, and timothy grown under various cutting schedules. Crop Sci. 2:333-336.
Smith, D., A.V.A. Jacques, and J.A. Balasko. 1973. Persistence of several temperate grasses grown with alfalfa and harvested two, three, and four times annually at two stubble heights. Crop Sci. 13:553-556.
Smith, K.F., and M.D. Casler. 2004. The use of spatially adjusted herbage yields during the analysis of perennial forage grass trials across locations. Crop Sci. 44:56-62.
Snaydon, R.W. 1970. Rapid population differentiation in a mosaic environment. I. The response of Anthoxanthum odoratum populations to soils. Evolution 24:257-269.
Snaydon, R.W. 1987. Population response to environmental disturbance, p. 15-31. In J. van Andel et al. (ed.) Disturbance in grasslands. Causes, effects and processes. Dr. W. Junk Publ., Dordrecht.
Snaydon, R.W., and M.W. Davies. 1972. Rapid population differentiation in a mosaic environment. II. Morphological variation in Anthoxanthum odoratum. Evolution 26:390-405.
Sultan, S.E. 1987. Evolutionary implications of phenotypic plasticity in plants. Evol. Biol. 21:127-178.
Vaylay, R., and E. van Santen. 1999. Grazing induces a patterned selection response in tall fescue. Crop Sci. 39:44-51.
M. D. Casler * and E. C. Brummer
Michael D. Casler, USDA-ARS, U.S. Dairy Forage Research Center, Madison, WI 53706-1108: E.C. Brummer, Dep. of Agronomy, Iowa State Univ., Ames, IA 50010. This research was partially supported by the College of Agricultural and Life Sciences, Univ. of Wisconsin and the Raymond F. Baker Center for Plant Breeding at Iowa State Univ. Received 4 Feb. 2005. * Corresponding author (mdcasler@ wisc.edu).
Table 1. Mean season total forage yield of smooth bromegrass populations collected from fence or sod habitats of rural cemeteries in Iowa, Minnesota, or Wisconsin and evaluated under three management regimes (four, rive, or six harvests over a 2-yr period) at two locations. Arlington, WI Ames, IA Source of germplasm Four Five Six Four Five ([dagger]) harvest harvest harvest harvest harvest Mg [ha.sup.-1] Iowa Fence 7.48 ** 5.95 ** 4.71 ** 13.36 ** 9.94 ** Sod 7.12 5.55 4.43 12.78 9.33 Minnesota Fence 7.62 ** 5.76 * 4.73 ** 12.57 ** 9.16 * Sod 6.96 5.57 4.39 11.73 8.83 Wisconsin Fence 7.46 ** 5.63 ** 4.60 ** 12.32 ** 9.11 * Sod 6.75 5.30 4.34 11.90 8.79 Ames, IA Source of germplasm Six ([dagger]) harvest Mean Mg [ha.sup.-1] Iowa Fence 6.65 * 8.01 * Sod 6.37 7.60 * Minnesota Fence 6.10 * 7.66 * Sod 5.81 7.21 * Wisconsin Fence 5.89 7.50 * Sod 5.76 7.14 * * Fence and sod means within a pair are significantly different at P < 0.05. ** Fence and sod means within a pair are significantly different at P < 0.01. ([dagger]) Number of cemeteries: Iowa (9), Minnesota (9), Wisconsin (12). Table 2. Mean regrowth percentage of smooth bromegrass populations collected from fence or sod habitats of rural cemeteries in Iowa, Minnesota, or Wisconsin and evaluated under three management regimes (four, five, or six harvests over a 2-yr period) at two locations. Arlington, WI Arlington, WI Source of germplasm Four Four Four Four ([dagger]) harvests harvests Mean harvests harvests Mg [ha.sup.-1] Iowa Fence 16.6 28.8 54.9 30.7 38.2 ** Sod 16.2 28.5 55.5 29.2 35.7 Minnesota Fence 16.6 ** 29.8 58.5 * 32.0 ** 38.8 * Sod 13.4 29.8 60.1 29.0 40.6 Wisconsin Fence 15.3 ** 28.5 56.5 ** 30.4 38.2 Sod 13.6 28.5 60.1 29.1 39.6 Arlington, WI Source of germplasm Four ([dagger]) harvests Mean Mg [ha.sup.-1] Iowa Fence 40.8 * 35.0 Sod 37.9 * 33.8 Minnesota Fence 42.5 36.4 Sod 41.7 35.8 Wisconsin Fence 41.3 35.0 Sod 42.3 35.5 * * Fence and sod means within a pair are significantly different at P < 0.05. ** Fence and sod means within a pair are significantly different at P < 0.01. ([dagger]) Number of cemeteries: Iowa (9), Minnesota (9), Wisconsin (12). Table 3. Linear regression coefficients for the regressions of season total forage yield or regrowth percentage on the average number of harvests over 2 yr for 60 smooth bromegrass populations collected from fence or sod habitats in 30 rural cemeteries of Wisconsin (WI), Minnesota (MN), or Iowa (IA) and five smooth bromegrass cultivars. Values reported are based on means over two locations and 2 yr. Season total forage yield P value ([double State Site Fence ([dagger]) Sod dagger]) WI 1 -2.39 [+ or -] 0.26 -2.16 [+ or -] 0.22 0.2432 WI 2 -2.49 [+ or -] 0.12 -2.17 [+ or -] 0.25 0.0974 WI 3 -1.94 [+ or -] 0.08 -2.52 [+ or -] 0.16 0.0032 WI 4 -3.03 [+ or -] 0.60 -2.38 [+ or -] 0.21 0.0008 Wl 5 -2.36 [+ or -] 0.04 -2.08 [+ or -] 0.15 0.1450 WI 6 -2.38 [+ or -] 0.04 -1.81 [+ or -] 0.05 0.0031 WI 7 -1.92 [+ or -] 0.17 -2.43 [+ or -] 0.27 0.0082 WI 8 -2.29 [+ or -] 0.17 -2.05 [+ or -] 0.32 0.2289 WI 9 -1.71 [+ or -] 0.12 -1.52 [+ or -] 0.17 0.3116 WI 10 -2.63 [+ or -] 0.11 -2.16 [+ or -] 0.32 0.0145 WI 11 -2.16 [+ or -] 0.01 -1.93 [+ or -] 0.02 0.2353 WI 12 -2.57 [+ or -] 0.05 -2.47 [+ or -] 0.13 0.6270 MN 13 -2.48 [+ or -] 0.13 -2.38 [+ or -] 0.19 0.6043 MN 14 -1.94 [+ or -] 0.13 -1.87 [+ or -] 0.16 0.7239 MN 15 -2.64 [+ or -] 0.13 -1.76 [+ or -] 0.22 0.0000 MN 16 -2.26 [+ or -] 0.02 -2.43 [+ or -] 0.05 0.4074 MN 17 -2.41 [+ or -] 0.35 -2.45 [+ or -] 0.11 0.8258 MN 18 -2.38 [+ or -] 0.75 -1.54 [+ or -] 0.41 0.0000 MN 19 -1.93 [+ or -] 0.08 -1.96 [+ or -] 0.06 0.8929 MN 20 -2.25 [+ or -] 0.36 -2.32 [+ or -] 0.22 0.7020 MN 21 -2.76 [+ or -] 0.30 -2.39 [+ or -] 0.05 0.0595 IA 22 -3.03 [+ or -] 0.01 -2.30 [+ or -] 0.16 0.0002 IA 23 -2.04 [+ or -] 0.10 -1.97 [+ or -] 0.15 0.7151 IA 24 -2.61 [+ or -] 0.38 -2.17 [+ or -] 0.29 0.0242 IA 25 -2.54 [+ or -] 0.04 -2.37 [+ or -] 0.11 0.3871 IA 26 -2.20 [+ or -] 0.38 -2.45 [+ or -] 0.15 0.1893 IA 27 -1.87 [+ or -] 0.20 -2.22 [+ or -] 0.51 0.0698 IA 28 -2.14 [+ or -] 0.35 -2.50 [+ or -] 0.12 0.0672 IA 29 -2.33 [+ or -] 0.42 -2.06 [+ or -] 0.11 0.1614 IA 30 -2.58 [+ or -] 0.32 -2.45 [+ or -] 0.19 0.4794 Regrowth percentage State Fence Sod P value WI 11.9 [+ or -] 2.8 13.3 [+ or -] 0.3 0.2343 WI 13.3 [+ or -] 2.2 16.1 [+ or -] 0.3 0.0170 WI 16.0 [+ or -] 0.7 11.4 [+ or -] 0.4 0.0001 WI 10.6 [+ or -] 0.9 13.4 [+ or -] 2.5 0.0170 Wl 12.0 [+ or -] 1.6 13.6 [+ or -] 0.8 0.1817 WI 13.1 [+ or -] 0.6 16.6 [+ or -] 1.1 0.0025 WI 16.3 [+ or -] 1.0 17.6 [+ or -] 0.9 0.2698 WI 16.0 [+ or -] 2.5 16.0 [+ or -] 0.9 0.9913 WI 14.4 [+ or -] 1.2 15.4 [+ or -] 3.2 0.3858 WI 7.6 [+ or -] 2.1 14.5 [+ or -] 1.5 0.0000 WI 12.6 [+ or -] 0.8 18.3 [+ or -] 1.7 0.0000 WI 12.6 [+ or -] 2.4 12.8 [+ or -] 2.2 0.8955 MN 10.3 [+ or -] 1.9 10.6 [+ or -] 0.3 0.7695 MN 13.8 [+ or -] 2.7 16.8 [+ or -] 2.5 0.0101 MN 13.4 [+ or -] 2.2 15.4 [+ or -] 1.6 0.0934 MN 14.2 [+ or -] 2.2 15.9 [+ or -] 0.4 0.1574 MN 15.3 [+ or -] 1.1 13.6 [+ or -] 1.5 0.1441 MN 12.2 [+ or -] 2.1 16.6 [+ or -] 1.1 0.0002 MN 13.3 [+ or -] 1.6 15.2 [+ or -] 0.3 0.1013 MN 13.7 [+ or -] 0.8 16.0 [+ or -] 1.5 0.0461 MN 11.8 [+ or -] 1.7 13.8 [+ or -] 2.7 0.0998 IA 11.7 [+ or -] 1.4 13.1 [+ or -] 0.3 0.2143 IA 13.1 [+ or -] 0.2 11.6 [+ or -] 0.8 0.1979 IA 10.8 [+ or -] 1.4 11.6 [+ or -] 2.0 0.4985 IA 10.5 [+ or -] 2.0 13.5 [+ or -] 3.4 0.0106 IA 13.5 [+ or -] 2.5 11.8 [+ or -] 2.7 0.1498 IA 14.6 [+ or -] 0.8 11.2 [+ or -] 1.4 0.0034 IA 13.5 [+ or -] 0.2 12.2 [+ or -] 2.7 0.2784 IA 10.0 [+ or -] 1.3 10.9 [+ or -] 1.6 0.4339 IA 11.1 [+ or -] 2.6 11.9 [+ or -] 0.7 0.4742 ([dagger]) Standard errors were computed from the linear regressions on number of harvests over 2 yr (1 df). ([double dagger]) P value for fence vs. sod linear regression coefficients obtained by contrast F tests in ANOVA.
|Printer friendly Cite/link Email Feedback|
|Author:||Casler, M.D.; Brummer, E.C.|
|Date:||Nov 1, 2005|
|Previous Article:||Characterization of resistance to Soybean mosaic virus in diverse soybean germplasm.|
|Next Article:||Carbon isotope discrimination accurately reflects variability in WUE measured at a whole plant level in rice.|