Agronomy Journal Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Undersander, D. J.
Right arrow Articles by Casler, M. D.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Undersander, D. J.
Right arrow Articles by Casler, M. D.
Agricola
Right arrow Articles by Undersander, D. J.
Right arrow Articles by Casler, M. D.
Related Collections
Right arrow Forage Management
Right arrow Interseeding
Right arrow Alfalfa
Agronomy Journal 93:609-619 (2001)
© 2001 American Society of Agronomy

FORAGES

Frost Seeding into Aging Alfalfa Stands

Sward Dynamics and Pasture Productivity

Daniel J. Undersandera, David C. Westb and Michael D. Caslera

a Dep. of Agronomy, Univ. of Wisconsin, Madison, WI 53706-1597
b Consumers' Coop., Richland Center, WI 53581

Corresponding author (mdcasler{at}facstaff.wisc.edu)

Received for publication February 23, 2000.

   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Little is known about the potential to frost-seed cool-season pasture species into mature alfalfa (Medicago sativa L.). Experiments were conducted in 1995 and 1996 near Arlington, WI (four sites), and Lancaster, WI (three sites), to evaluate the establishment and response to seeding rates of five cool-season grasses and red clover (Trifolium pratense L.) frost-seeded into mature alfalfa (2- to 5-yr-old stands with 30 to 50 plants m-2). Smooth bromegrass (Bromus inermis Leyss.), orchardgrass (Dactylis glomerata L.), perennial ryegrass (Lolium perenne L.), reed canarygrass (Phalaris arundinacea L.), timothy (Phleum pratense L.), and red clover were frost-seeded into mature alfalfa stands at six seeding rates. Orchardgrass, perennial ryegrass, and red clover had higher densities and responses to seeding rate than smooth bromegrass, timothy, and reed canarygrass in the seeding year, but these differences were less pronounced in the postseeding year. Orchardgrass contributed more grass dry matter in the seeding year but was similar to smooth bromegrass and timothy and greater than perennial ryegrass and reed canarygrass in the second year. Alfalfa and weed suppression were highest with orchardgrass due to its aggressive growth habit, high occurrence, and winterhardiness. Postseeding–year mixture yields were high for smooth bromegrass, orchardgrass, timothy, and reed canarygrass, but low for perennial ryegrass and red clover. Forage yield increased with seeding rate at sites with the greatest initial establishment. The results of this study suggest frost seeding temperate pasture species into mature alfalfa can increase plant diversity and forage yield while suppressing weeds.

Abbreviations: NIRS, near-infrared reflectance spectrophotometer • SEC, standard error of calibration • SEV, standard error of validation


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
INCREASED CAPITAL COSTS without commensurate increases in produce prices have caused many farmers with livestock, especially dairy, to rely more on pasture as a low-cost source of forage. Rising land prices and increased tax rates have caused many of the same farmers to optimize both total production and seasonal distribution of production through increased management and agronomic inputs. Unmanaged pastures tend to have low yield, often contain undesirable species (Doll, 1981; Shaeffer et al., 1990), and may shift to less palatable biotypes of grass species (Falkner and Casler, 1998), while well-managed pastures maintain higher total yield, seasonal yield distribution, and forage quality (Paine et al., 1999).

Well-managed pastures are characterized by adequate fertilization, adequate pasture rest periods between grazings, and introduction of new species. Pasture improvement research has concentrated mainly on introduction of legume species that have been shown to increase dry matter production (Knight, 1970) and improve seasonal forage distribution (Evers, 1985). Grasses have been difficult to establish in existing sod due to excessive competition from resident plans (Sprague et al., 1947). However, the tremendous range in yield potential, maturity, and palatability among grass species and biotypes may justify introduction of grasses into pasture swards to improve animal performance.

No-tillage seeding allows introduction of new species while reducing erosion and minimizing risk of stand failure and yield loss in the seeding year. However, no-tillage seeding has traditionally been limited by cost and availability of specialized seeding equipment. Surface broadcasting of seed in late winter (frost seeding) provides a mechanism to renovate pastures without tillage and with minimal equipment expenditures. Aging alfalfa stands, which have become unproductive hay fields, are excellent candidates for introduction of grasses by frost-seeding techniques. Establishment of perennial grasses by frost seeding into aging alfalfa stands may allow for development of productive, persistent, and species-rich pastures without opening the sod to erosion. The objectives of this research were to determine the influence of species and seeding rates on sward component occurrence, botanical composition, and forage yield of six temperate pasture species frost-seeded into mature, or declining alfalfa stands.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Field experiments were conducted in 1995 and 1996 at the University of Wisconsin Agricultural Research Stations near Arlington, WI (43°18'N, 89°22'W), and near Lancaster, WI (42°50'N, 90°47'W). In 1995 the experiment was conducted in three alfalfa fields; one at Lancaster and two at Arlington, and all were harvested mechanically. In 1996 the experiment was conducted in four alfalfa fields—two at Arlington and two at Lancaster—with a grazed site and clipped site at each location. The soil at all Arlington sites was Plano silt loam (fine-silty, mixed, mesic Typic Argiudolls). The soil at all Lancaster sites was Fayette silt loam (fine-silty, mixed, mesic, Typic Hapludalfs).

The experimental design was a split-plot in a randomized complete block with four replicates. Whole plots were six species and subplots were six seeding rates; subplot size was 1.22 by 6.71 m. The six species used were: smooth bromegrass cv. Alpha, orchardgrass cv. Benchmark, perennial ryegrass cv. Madiera, reed canarygrass cv. Rival, red clover cv. Marathon, and timothy cv. Colt. The six seeding rates were 0, 55, 110, 220, 440, and 880 seeds m-2 on a pure live seed basis. The 880 seeds m-2 rate corresponded to an average rate of 14.5, 17.6, 6.1, 7.5, 29.4, and 3.3 kg ha-1 for red clover, perennial ryegrass, orchardgrass, reed canarygrass, smooth bromegrass, and timothy, respectively (Smith et al., 1986). The fields selected for the experiment had been in alfalfa for 2 to 5 yr. Alfalfa stand density was approximately 30 to 50 plants m-2. Field preparation consisted of clipping each site to a 5-cm stubble height in the autumn before the seeding year. Seeding took place in mid-March. A drill seeder was used with the openers elevated above the soil surface to simulate broadcast seeding while maintaining precise control of seeding rates.

Nitrogen was applied to all plots (excluding the six red clover seeding rates) at a rate of 56 kg ha-1 approximately 30 d after seeding and again the first week of August in the seeding year, and at 78 kg ha-1 in early April of the postseeding year. To reduce competition from preexisting vegetation, plots were clipped (hay treatment) or grazed (pasture treatment) to 8 cm throughout the seeding year whenever the maximum canopy height reached 35 cm. After grazing, the fields were clipped to achieve a uniform stubble height of 8 cm.

Sward-component occurrence was determined in late September of the seeding year and in late May of the postseeding year by the line-intercept method. A 90-cm transect was placed at three random positions per plot and the single species nearest to each of six evenly spaced points was recorded, giving a total of 18 observations per plot. Species were recorded in five groups: seeded species, alfalfa, annual grasses, unseeded perennial grasses (any perennial grass not intentionally seeded), and broadleaf weeds. Occurrence was computed as the percentage of the 18 recorded observations.

Forage yield was determined in late September of the seeding year and late May of the postseeding year. Forage yields from the seeding year were obtained by randomly clipping two 0.25-m2 areas per plot to a 5-cm stubble height. Forage yields for the postseeding year were obtained by using a sickle-bar plot harvester to harvest the entire plot. A 1-kg grab-sample was taken at random from each plot.

Samples were oven-dried at 60°C for approximately 3 d and used to determine dry matter content. Two stratified random groups of forage samples (n = 205 for the fall 1995 sites and n = 245 for the remaining harvests combined) were manually separated into three components: grass, legume, and other broadleaves. Separated samples were reconstituted after weighing the individual components. Samples from the fall of 1995 were ground twice to pass through a 2-mm and a 1-mm screen, respectively. The remaining samples were ground once to pass a 2-mm screen. All forage samples were scanned on a near-infrared reflectance spectrophotometer (NIRS) and separated samples were used to calibrate NIRS equations to predict the contribution of these components for all forage samples. Calibration and validation statistics (R2, SEC = standard error of calibration, and SEV = standard error of validation, respectively) for the fall 1995 harvest were: 0.52, 16.8 g kg-1, and 19.4 g kg-1 (grass); 0.77, 13.6 g kg-1, and 16.9 g kg-1 (legume); and 0.58, 15.7 g kg-1, and 16.5 g kg-1 (other broadleaves). Calibration and validation statistics (R2, SEC, and SEV, respectively) for the remaining harvests were: 0.97, 4.4 g kg-1, and 5.5 g kg-1 (grass); 0.97, 6.3 g kg-1, and 7.6 g kg-1 (legume); and 0.96, 5.9 g kg-1, and 6.9 g kg-1 (other broadleaves).

Forage yield, sward component occurrence, and botanical composition variables were analyzed by analysis of variance for individual sites and combined over sites. Species was considered a fixed effect, while block, site, and seeding rate were considered random. Data from each year were analyzed separately. Simple linear or log-linear regression was used to measure response of species to seeding rates. Regression models were determined by visual inspection of plots with the objective of using the best common model for all species. Linear or log-linear regression coefficients were compared by t-test (Steel et al., 1997).

A short-term economic analysis of frost-seeding success was performed using forage yield data from May of the postseeding year. For each combination of seven sites–six species–five seeding rates, the value of additional hay due to frost seeding (V = frost-seeding treatment mean-mean of the unseeded control), was computed using a hay price of $90 Mg-1. The cost of seed (C) was computed using data from Casler et al. (1999). A successful frost seeding was defined as V - C > 0. The probability of frost-seeding success was computed for each of the 30 seeding rate–species combinations as the frequency of sites with V - C > 0. These probability estimates are highly conservative, because they only account for extremely short-term profits (May of the postseeding year). Frost-seeding equipment, fuel, and labor costs were assumed to be negligible. The probability of frost-seeding success was modeled by linear, log-linear, or quadratic regression for each species.


   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Species x site and rate x site interactions were significant (P < 0.05) for all variables. Therefore, most results are presented for individual sites. Species x rate interactions were significant (P < 0.05) only for grass and legume contributions to sward dry matter and for occurrence at both September and May harvests. The species x rate x site interaction was not significant for these variables, so species x rate interaction data are presented as means over sites. Finally, despite the presence of these interactions, there were numerous significant (P < 0.01) main effects for species and seeding rate. The main effect of species resulted from reasonably consistent species rankings across sites and rates. The significant rate main effects resulted from the overwhelmingly large variation among rate means.

Variation among Species
Seeding Year
Occurrence of seeded species differed among years and sites within years (Table 1). Mean seeded species occurrence was higher in 1995 compared with 1996 (17.9 vs. 7.0%, respectively). In 1995, temperatures and rainfall were normal through the period of seedling establishment but the remainder of the growing season had unusually high temperatures and below-average rainfall. The Lancaster site may have endured greater climatic stress because of its southern exposure and substantially lower organic matter in the soil compared with the soil at Arlington (24 vs. 39 g kg-1), potentially reducing available water holding capacity. In 1996, rainfall was above normal for spring, but below normal for July and August. Temperatures were well below normal, especially in April and May, possibly reducing seedling establishment and thereby affecting seeded-species occurrence at the end of the season. Seedling density 60 d after planting (Casler et al., 1999) was generally lower for grazed sites, although this trend was not evident for seeded-species occurrence by September of the seeding year.


View this table:
[in this window]
[in a new window]
  		Table 1. Mean seeded-species occurrence following frost seeding, determined by line-intercepts in September of the seeding year at seven sites, identified by location, harvest management, and year of seeding. Means are over four replicates and five seeding rates

 
Species rankings for occurrence in September of the seeding year were highly repeatable across sites (Table 1). Perennial ryegrass or orchardgrass ranked first or second at all sites, with the exception of perennial ryegrass at Arlington-Clipped-1995 (no. 2) and orchardgrass at Arlington-Grazed-1996. Similarly, reed canarygrass ranked sixth (last) at six sites and fifth at the other site, while timothy ranked fifth at four sites and fourth at three sites. The ranking of species means corresponded closely to their ranking for seedling density 60 d after planting (Casler et al., 1999) and to the seedling aggressiveness classification made by Blazer et al. (1956).

Perennial ryegrass, due to its aggressive seedlings (Table 1; Blazer et al., 1956; Casler et al., 1999), had greater suppression of annual grasses (mean occurrence = 5.5%, data not shown) than smooth bromegrass, reed canarygrass, red clover, and timothy, which have less aggressive seedlings (mean occurrence = 7.5 to 7.9%; all P < 0.05 compared with perennial ryegrass). Suppression of annual grasses by orchardgrass was greater than smooth bromegrass (mean annual grass occurrence = 6.0 vs. 7.9%: P < 0.05) for similar reasons. Suppression of unseeded perennial grasses, primarily quackgrass [Elytrigia repens (L.) Nevski] may be more difficult because of their extensive rhizomes. Orchardgrass had greater quackgrass suppression than reed canarygrass (mean unseeded perennial grass occurrence = 10.0 vs. 12.9%; P < 0.05); there were no other species differences for unseeded perennial grass occurrence. Species with the highest component occurrence (Table 1)—such as orchardgrass, perennial ryegrass, and red clover—reduced broadleaf weed occurrence more than reed canarygrass (mean broadleaf weed occurrence = 24.2 to 26.4 vs. 33.6%; all P < 0.05).

There was little difference in overall mean grass contribution between years (212 g kg-1 in 1995 and 243 g kg-1 in 1996), but considerable variation among sites within years (Table 2). As with seeded-species occurrence, variation among species was as great as variation among sites. Orchardgrass contributed more grass in 1995 than smooth bromegrass, reed canarygrass, and timothy due to its aggressive growth and tolerance to heat and drought. Despite its high occurrence, perennial ryegrass provided dry matter contributions similar to those for smooth bromegrass, reed canarygrass, and timothy at Arlington in 1995, where drought conditions were less evident than at Lancaster. This is attributed to the short stature of perennial ryegrass compared with the upright growth of smooth bromegrass, reed canarygrass, and timothy. Species differences were less evident in 1996 as a result of reduced establishment compared with 1995 (Casler et al., 1999) and cooler temperatures, resulting in more uniform growth of all seeded grasses. Across all seven sites, only three species showed evidence of significant (P < 0.05) increases in grass dry matter compared with the unseeded check plots: smooth bromegrass at two sites, orchardgrass at three sites and averaged across sites, and perennial ryegrass at three sites and averaged across sites.


View this table:
[in this window]
[in a new window]
  		Table 2. Mean grass and legume contributions to sward dry matter following frost seeding, determined by NIRS analysis of samples clipped in September of the seeding year at seven sites, identified by location, harvest management, and year of seeding. Means are over four replicates and five seeding rates

 
Differences among species for legume dry matter contributions (Table 2) were generally inverse of those for seeded-species occurrence (Table 1). All five grass species led to reduced legume dry matter compared with the red clover seedings (P < 0.05) at a minimum of three sites and averaged across sites. However, there were few significant differences among grass species; the only reasonably consistent difference was: orchardgrass < timothy, generally at sites with higher temperatures, lower rainfall, and/or higher legume contributions, conditions under which orchardgrass was better able to compete than timothy. Orchardgrass appeared to be the most effective competitor against the established alfalfa plants.

For all of the seven sites, there were no differences among seeded species or for seeded species vs. unseeded check for total-sward forage yield in September of the seeding year (data not shown).

Postseeding Year
Species ranks for seeded-species occurrence were consistent among the seven sites, except for Arlington-Grazed-1996, where smooth bromegrass values were inflated due to the presence of smooth bromegrass before frost seeding (Table 3). However, there were some notable changes in species abundance from that observed in the seeding year (Table 1). Orchardgrass became the most prevalent species, having the greatest occurrence at all sites, except Arlington-Grazed-1996. This was due to its aggressive growth in the seeding year (Table 1; Casler et al., 1999), coupled with earlier spring growth and increased response to supplemental N compared with the other species. The overall mean occurrence of orchardgrass was 2.3 times greater in the postseeding year compared with the seeding year.


View this table:
[in this window]
[in a new window]
  		Table 3. Mean seeded-species occurrence following frost seeding, determined by line-intercepts in May of the postseeding year at seven sites, identified by location, harvest management, and year of seeding. Means are over four replicates and five seeding rates

 
The occurrence of perennial ryegrass was less than orchardgrass, similar to smooth bromegrass and red clover, and usually greater than timothy and reed canarygrass. Perennial ryegrass was not as prevalent in the postseeding year, especially at the Arlington 1996 sites where reduced duration of snow cover, compared with Lancaster, increased the potential for winter injury. Overall mean occurrence of perennial ryegrass in the postseeding year was 1.4 times greater than in the seeding year, the lowest increase of all species. Substantial over-winter increases of perennial ryegrass are not likely in this temperate climate due to its lack of adaptation to Wisconsin winters with low potential for snow cover (Casler, 1988; Casler and Walgenbach, 1990). This suggests the potential benefits of frost seeding perennial ryegrass may be seen more in the seeding year and persistence is likely to decrease in subsequent years, prompting the need for occasional reseeding.

Red clover had a 2.2-fold increase in occurrence from the seeding year to the postseeding year, similar to that for most of the grasses (Table 3). Smooth bromegrass, timothy, and reed canarygrass were similar in that they had the lowest establishment (Casler et al., 1999), but they also had the greatest increase in occurrence from the seeding year to the postseeding year (2.9, 3.8, and 3.4 times greater, respectively). These three species have the greatest long-term persistence of the six species used in this study. These results indicate the general negative association between short-term establishment rate and long-term persistence. The mechanisms that promote long-term persistence in these three species—rhizomes or haplocorms—most likely require additional energy inputs and development time, reducing the establishment rate for these species. These results support the suggestion that these species can be successfully frost-seeded at lower rates than orchardgrass, red clover, or perennial ryegrass, due to their increased capacity for vegetative reproduction (Casler et al., 1999).

Alfalfa occurrence was greatly reduced from the seeding year (503 to 399 g kg-1 from September to May, data not shown) as a result of natural stand mortality, N applications favoring grass growth, and the advancing effects of grass colonization. Generally, the grass species with the greatest occurrence had the most suppressive effect on alfalfa occurrence (data not shown). Research has shown orchardgrass to be more aggressive to alfalfa in binary mixtures than other cool-season grasses (Casler, 1988; Jung et al., 1982). This frost-seeding research can be interpreted similarly, but occurrence of the colonizing species may be as important as their growth and establishment characteristics. Adjusting seeding rates to obtain a desired density, regardless of species, may help minimize the differences in suppression.

Orchardgrass and perennial ryegrass each suppressed annual grasses more than the other species (annual grass occurrence = 4.4 vs. 6.5 to 9.6%; all P < 0.05; data not shown). The early spring growth of orchardgrass and dense ground cover of both orchardgrass and perennial ryegrass provided an unfavorable environment for annual grasses. Reed canarygrass had the lowest suppression (annual grass occurrence = 9.6%) because of its low occurrence, providing adequate space for annual grasses to grow. Quackgrass occurrence increased from the levels found in the seeding year because of vigorous rhizomes and supplemental N applications intended to promote growth of the seeded species. Orchardgrass, with its aggressive growth habit and high plant occurrence suppressed quackgrass better than reed canarygrass or smooth bromegrass (perennial grass occurrence = 13.9 vs. 24.5 and 23.0, respectively; both P < 0.05). Control of these weeds has proven difficult, especially in forage systems where legumes or other short-lived species decrease in occurrence and provide a niche for weed invasion. By manipulating the plant environment with the addition of aggressive and/or high occurrence species, the extent of weed invasion may be reduced.

Grass contribution to dry matter yield increased from the seeding year to the postseeding year at all sites, except Arlington-Grazed-1996 and Lancaster-Clipped-1996 (Table 4 vs. Table 2). Orchardgrass, smooth bromegrass, and timothy increased in grass dry matter from September to May, by an average of 53 to 92%. Reed canarygrass showed no change, while perennial ryegrass dry matter declined by 22%, most likely due to winter injury.


View this table:
[in this window]
[in a new window]
  		Table 4. Mean grass and legume contributions to sward dry matter following frost seeding, determined by NIRS analysis of samples clipped in May of the postseeding year at seven sites, identified by location, harvest management, and year of seeding. Means are over four replicates and five seeding rates

 
The contribution of frost-seeded red clover to legume dry matter was greater than for all frost-seeded grasses at all 1995 sites, but only at Lancaster-Grazed-1996 (Table 4). This was due to the generally greater stands of seeded species at the 1995 sites, both immediately after establishment (Casler et al., 1999) and in September of the seeding year (Table 2). Greater frost-seeding establishment success of both grasses and red clover resulted in a greater differential between grass and legume contributions for the grass vs. red clover seeding treatments in 1995.

Seeded-species rankings for broadleaf dry matter contribution were highly inconsistent among sites (data not shown). There were numerous significant differences among species, but they often involved large changes in rank among sites. Over all sites, there were no differences among species or between seeded species and unseeded checks in broadleaf weed contributions to the sward.

Forage yield differences among species were variable among sites, but perennial ryegrass and red clover seedings usually had lower yield compared with smooth bromegrass, orchardgrass, timothy, and reed canarygrass seedings (Table 5). This was observed for perennial ryegrass at four sites where establishment and/or seeded-species occurrence was relatively high, and probably resulted from its short stature. In addition, the occurrence of dandelions (Taraxacum officinale Weber), which reduce yield potential due to their rosette growth pattern, was usually the highest in perennial ryegrass plots (data not shown). The dense canopy of horizontal leaves associated with red clover has been shown to shade more prostrate species, thus reducing sward yield potential (Harris, 1974). Smooth bromegrass ranked highest in forage yield at three of seven sites and, overall, was significantly higher in sward forage yield than all seeded species, except timothy (P < 0.05). Smooth bromegrass was the only species with mean forage yield higher than the unseeded check averaged over sites, due to its consistent high ranking across sites.


View this table:
[in this window]
[in a new window]
  		Table 5. Mean forage yield in May of the postseeding year for six temperate pasture species following frost seeding at seven sites, identified by location, harvest management, and year of seeding. Means are over four replicates and five seeding rates

 
Seeding Rate Responses
Simple linear regressions of mean botanical component densities on seeding rate did not fit well. On the whole, the best overall model was the linear regression of mean component densities on the natural logarithm of seeding rate—all such regressions were fitted with this model. Forage yield and botanical composition data fitted best to simple linear regressions.

Seeding Year
A log-linear response of seeded-species occurrence to seeding rate (P < 0.01) was observed at five of seven sites (Table 6). Among sites, log-linear responses showed a sevenfold range of variation. Sites with the highest mean seeded-species occurrence had the greatest response to seeding rate. Increased seeding rate led to reduced occurrence of alfalfa, annual grasses, and broadleaf weeds at nearly all sites, but the magnitude and significance of these responses were highly variable. Significance (P < 0.05) of regressions for these components occurred only at one to four of the five sites with significant regressions for seeded-species occurrence, and were not closely related to mean occurrence of these components.


View this table:
[in this window]
[in a new window]
  		Table 6. Summary statistics for linear regressions of sward component occurrences (%) measured in September of the seeding year on the natural logarithm of seeding rate (seeds m-2) following frost seeding at seven sites, identified by location, harvest management, and year of seeding. Regressions were computed from means over four replicates and six species

 
Averaged over sites, red clover, perennial ryegrass, and orchardgrass showed significantly contrasting seeding-rate responses for the contributions of grass vs. legume dry matter to swards (Fig. 1). For these three species, responses of grass and legume dry matter to seeding rate were all significant (P < 0.05). These three species also showed the greatest establishment 60 d following frost seeding (Casler et al., 1999). Although none of the grasses became the dominant component of the sward, grass and legume concentration of the sward were equal at the highest orchardgrass seeding rate. The short stature of perennial ryegrass, resulting in shading by the existing canopy, may have reduced its ability to respond to increasing seeding rate, reflecting its lower response compared with orchardgrass. Competition among plants at high densities further decreases plant height due to reduced availability of resources (Donald, 1963), further reducing the ability of perennial ryegrass to show grass component dry matter responses. Increased red clover seeding rates led to increased legume and reduced grass component dry matter through greater numbers of red clover plants at high seeding rates, which provided increased competition and shading of preexisting grasses.



View larger version (33K):
[in this window]
[in a new window]
  		Fig. 1. Grass (closed circles) and legume (open circles) contributions to sward dry matter in September of the seeding year, 6 mo after frost seeding with six temperate pasture species. Means are over four replicates at each of seven sites. Presence of * or ** indicates significance of linear regressions for grass (G) or legume (L) dry matter on seeding rate, or significance of the difference between linear regression coefficients for grass (ßG) vs. legume (ßL) at P = 0.05 or 0.01, respectively

 
Conversely, the three species that were slowest to establish from frost seeding—reed canarygrass, smooth bromegrass, and timothy—generally did not have significant responses of grass or legume dry matter contributions to increased seeding rate in September of the seeding year (Fig. 1). Species with low establishment capacity and seedling growth rate apparently have little ability to express short-term responses to increased seeding rate in terms of contributing to sward dry matter. Short-term growth of seedlings for these species is most likely focused on mechanisms that promote long-term survival and vegetative reproduction (rhizomes and haplocorms), rather than on producing excessive aboveground dry matter during the establishment year.

Postseeding Year
The log-linear response of seeded-species occurrence to seeding rate was significant (P < 0.01) at all seven sites (Table 7). Among sites, log-linear responses showed a sevenfold range of variation and, as observed for September of the seeding year, were positively associated with mean seeded-species occurrence. At all sites, the mean seeded-species occurrence increased from September of the seeding year to May of the postseeding year (Table 7 vs. Table 6); each of these increases was significant (P < 0.01). Increases in mean seeded-species occurrence ranged from 79 to 200% of the September mean. At all but two sites (the two grazed sites), the log-linear response to seeding rate increased (P < 0.05) from September of the seeding year to May of the postseeding year (Table 7 vs. Table 6). Increases in seeded-species occurrence response to seeding rate in May ranged from 65 to 177% of the response observed in the previous September. Reductions in alfalfa occurrence associated with increased seeding rates were similar to those observed in September of the seeding year.


View this table:
[in this window]
[in a new window]
  		Table 7. Summary statistics for linear regressions of sward component occurrences (%) measured in May of the postseeding year on the natural logarithm of seeding rate (seeds m-2) following frost seeding at seven sites, identified by location, harvest management, and year of seeding. Regressions were computed from means over four replicates and six species

 
Annual grasses and broadleaf weeds were significantly (P < 0.05) reduced by higher seeding rate at all sites in which these weeds were present in May of the postseeding year, with the single exception of Arlington-Clipped-1996 (Table 7). This result is probably explained by a combination of increased seeded-species ground cover at high seeding rates between September and May, and the earlier spring growth of perennial forages vs. annual weeds. Similar principles apply to suppression of unseeded perennial grass, but due to their aggressive nature, similar growth habit, and similar reproductive cycle to the seeded species, unseeded perennial grasses were reduced at only four of the seven sites, those sites with the highest seeding rate response for seeded-species occurrence. Thus, frost seeding of perennial forages can suppress existing perennial grasses when establishment conditions are sufficient for frost seeding to be successful (Casler et al., 1999). Suppressed occurrence of perennial grasses does not appear to be dependent on their initial occurrence, as noted by the lack of relationship between mean perennial grass occurrence and their response to increased frost-seeding rates of the introduced species (Table 7).

Seeded species showed an increased occurrence response to seeding rate between September of the seeding year and May of the postseeding year, with the exception of perennial ryegrass (Fig. 2). This response was greatest for the three species with the lowest initial occurrence: reed canarygrass, smooth bromegrass, and timothy, which ranged from 263 to 344% increase in rate response. Increases in seeding rate responses between September and May were likely due to increased root development and tillering, which were likely offset by winter injury to perennial ryegrass.



View larger version (38K):
[in this window]
[in a new window]
  		Fig. 2. Seeded-species occurrence in September of the seeding year (open circles) or May of the postseeding year (closed circles) following frost seeding with six temperate pasture species. Means are over four replicates at each of seven sites. Presence of * or ** indicates significance of linear regressions of seeded-species occurrence on the logarithm of seeding rate, or significance of the difference between log-linear regression coefficients for September (ßSep) vs. May (ßMay) at P = 0.05 or 0.01, respectively

 
The seeding-rate response of grass and legume sward component contributions for perennial ryegrass frost seedings had disappeared by May of the postseeding year (Fig. 3). Similarly for red clover, the legume-component response was half that observed in September of the seeding year and the response of the grass component for red clover seedings had disappeared (Fig. 3 vs. Fig. 1). The reduced response for these seeded species most likely reflects mortality between September and May, an effect that was greatest at the highest seeding rates of perennial ryegrass and red clover. Swards with greater numbers of established plants of these two relatively nonpersistent species appear to be more susceptible to winter stand losses than swards with fewer numbers of established plants.



View larger version (34K):
[in this window]
[in a new window]
  		Fig. 3. Grass (closed circles) and legume (open circles) contributions to sward dry matter in May of the postseeding year, 14 mo after frost seeding with six temperate pasture species. Means are over four replicates at each of seven sites. Presence of * or ** indicates significance of linear regressions for grass (G) or legume (L) dry matter on seeding rate, or significance of the difference between linear regression coefficients for grass (ßG) vs. legume (ßL) at P = 0.05 or 0.01, respectively

 
Seeding-rate responses of grass and legume sward component contributions were all significant for orchardgrass, smooth bromegrass, and timothy (Fig. 3). For orchardgrass, the change in response from September of the seeding year to May of the postseeding year reflected a 45% increase in the difference between the grass and legume component regression coefficients. By May of the postseeding year, orchardgrass had become the dominant component of swards at the three highest seeding rates. Thus, mature alfalfa fields can become orchardgrass-dominant pastures approximately 1 yr after frost seeding at rates of >=220 seeds m-2.

For smooth bromegrass and timothy, the change in response from September of the seeding year to May of the postseeding year reflected a 238 and 228% increase, respectively, in the difference between the grass and legume component regression coefficients. These swards had become grass-dominant, or nearly so, at the two highest seeding rates, 440 seeds m-2 or higher. These seeding rates were considered to be economically advantageous for establishment of timothy, but not for smooth bromegrass, due to its large seed size/seed cost ratio (Casler et al., 1999). For smooth bromegrass, economical seeding rates of 100 to 200 seeds m-2 would require additional time before grass dominance is achieved.

Finally, for reed canarygrass, a significant seeding-rate response of the grass component dry matter contribution was observed in May of the postseeding year (Fig. 3). However, the rate of response was considerably lower than for any of the other grasses, except perennial ryegrass. Aging alfalfa fields frost-seeded to reed canarygrass may eventually become grass-dominant, but it will require significantly more time than for orchardgrass, smooth bromegrass, and timothy.

Forage yield in May of the postseeding year increased with seeding rate at three of the seven sites (Table 8). These three sites generally had highest values of most measures of frost-seeding success: seedling density and percentage establishment (Casler et al., 1999), mean seeded-species occurrence and component contribution to sward dry matter, and log-linear response of seeded-species occurrence or sward contributions to increased seeding rate. At these sites, there was sufficient establishment to colonize openings in the alfalfa canopy and to provide competition from the seeded species to suppress and/or replace the existing vegetation with new species. Sites with moderate or poor establishment of the seeded species did not provide sufficient establishment levels to effectively increase forage yield, despite a log-linear response of seeded-species occurrence to seeding rate.


View this table:
[in this window]
[in a new window]
  		Table 8. Mean forage yield in May of the post seeding year for six seeding rates of temperate pasture species following frost seeding at seven sites, identified by location, harvest management, and year of seeding. Means are over four replicates and six species

 
Red clover had extremely low probability of frost-seeding success (mean probability = 0.03; Fig. 4), because it generally resulted in decreased sward yields (Table 5), making it a poor choice to renovate aging alfalfa fields. Reed canarygrass also had a very low mean probability of short-term success, due to its high seed cost and low seedling aggressiveness. Timothy had a uniformly high probability of frost-seeding success (mean probability = 0.80), largely because it resulted in forage yield increases (Table 5) and its cost per seed is extremely low (Casler et al., 1999). Perennial ryegrass had moderate probabilities at low seeding rates, largely because stand losses due to winter injury resulted in a lower cost investment. Finally, orchardgrass and smooth bromegrass had high probabilities of frost-seeding success at low seeding rates, due to their combination of high seedling aggressiveness, high tillering capacity, and low susceptibility to winter injury.



View larger version (20K):
[in this window]
[in a new window]
  		Fig. 4. Probability of frost-seeding success (P), as measured by the frequency of sites for which the value of increased May forage yields exceeded the cost of seed, as a function of seeding rate (SR). Regressions were: Perennial ryegrass, P = 0.951 - 0.144ln(SR), R2 = 0.72, P = 0.07; Orchardgrass, P = 0.722 + 3.4 x 10-5SR - 7.9 x 10-7SR2, R2 = 0.99, P < 0.01; Reed canarygrass, P = 0.391 - 0.062ln(SR), R2 = 0.75, P = 0.06; and Smooth bromegrass, P = 1.787 - 0.268ln(SR), R2 = 0.98, P < 0.01

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
Frost seeding temperate pasture species into aging or mature alfalfa stands can increase plant diversity and forage yield and reduce weed pressure. Species differed in their ability to contribute to yield based on their establishment, growth habit, and winterhardiness. Over the 14-mo duration of the seven experiments reported here, frost seeding resulted in successful stand establishment of five of the six species. Red clover, perennial ryegrass, and orchardgrass became successfully established and an important component of swards within the seeding year. For orchardgrass, aggressive tillering and competitive ability were probably the most important factors in its rapid establishment. For red clover and perennial ryegrass, superior seedling vigor was most likely the greatest single factor contributing to this response. By May of the postseeding year, some loss of stand was observed at the highest seeding rates for red clover and perennial ryegrass. While red clover was very amenable to frost seeding, it was not suitable for renovating aging alfalfa fields, due to reduced forage yields.

For smooth bromegrass and timothy, the energy required to develop rhizomes and haplocorms, respectively, appears to limit the ability of these species to produce new tillers and a high level of aboveground dry matter in the seeding year. However, both species had become nearly the dominant sward component in May of the postseeding year, for the highest seeding rates. Thus, their investment in an underground carbohydrate storage and vegetative reproductive system may delay their rapid establishment from frost seeding, but it should prove beneficial for long-term stand development. Both species showed increased sward forage yields as early as 14 mo after frost seeding. Reed canarygrass showed similar trends to smooth bromegrass and timothy, but on a considerably delayed schedule. The length of time required for reed canarygrass to become dominant after frost seeding into aging alfalfa fields is unclear from this research. While its capacity for vegetative reproduction was observed within 14 mo after frost seeding, it is possible that other seeding methods, such as no-till drilling, might result in more rapid establishment of reed canarygrass.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Crop Sci.Home page
M. D. Casler and D. J. Undersander
Selection for Establishment Capacity in Reed Canarygrass
Crop Sci., April 25, 2006; 46(3): 1277 - 1285.
[Abstract] [Full Text] [PDF]

Home page
 Agron. J.Home page
D. C. Cummings, R. C. Berberet, J. F. Stritzke, and J. L. Caddel
Sod-Seeding and Grazing Effects on Alfalfa Weevils, Weeds, and Forage Yields in Established Alfalfa
Agron. J., September 1, 2004; 96(5): 1216 - 1221.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Web of Science (3)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Undersander, D. J.
Right arrow Articles by Casler, M. D.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Undersander, D. J.
Right arrow Articles by Casler, M. D.
Agricola
Right arrow Articles by Undersander, D. J.
Right arrow Articles by Casler, M. D.
Related Collections
Right arrow Forage Management
Right arrow Interseeding
Right arrow Alfalfa

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Crop Science 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