| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Forest & Landscape Denmark, Rolighedsvej 23, DK-1958 Frederiksberg C, Denmark
b Dep. of Agricultural Sciences, Royal Veterinary and Agricultural Univ., Thorvaldsensvej 40, DK-1871 Frederiksberg C, Denmark
* Corresponding author (sugl{at}kvl.dk).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Kentucky bluegrass germinates and emerges more slowly than red fescue and particularly perennial ryegrass (Skirde, 1967; Pommer, 1972; Bø, 1989). Hence, Pommer (1972) reported that time to emergence ranged from 16 to 22 d for Kentucky bluegrass, 13 to 18 d for red fescue, and 10 to 12 d for perennial ryegrass depending on the cultivar, and Skirde (1967) reported 11 to 35 d, 7 to 32 d, and 6 to 26 d, respectively, depending on the time of sowing. Moreover, Kentucky bluegrass has been reported to establish more slowly than perennial ryegrass with red fescue having intermediate establishment speed when evaluated by plant height at a certain time after sowing (DeFrance and Simmons, 1951; Adams and Bryan, 1974) or by the time to sward closure (Pommer, 1972).
When competition occurs between plants in a turfgrass community, the initial advantage gained by a plant subsequently increases it competitive advantage in an additive way (Beard, 1973). Among the factors affecting the competitive balance between species, the time of germination and emergence has a particularly profound effect (Black and Wilkinson, 1963; Ross and Harper, 1972). For example, in a two-species mixture of compact brome (Bromus madritensis L.) and ripgut brome (Bromus rigidus Roth), the relative time of sowing of the two species considerably affected the relative contribution to the total dry matter production of the mixture (Harper, 1961). Thus, when the sowing of ripgut brome was delayed up to 32 d compared with the sowing of compact brome, the relative contribution from compact brome increased from less than 25% to approximately 90%, whereas the total yield was not affected.
In the competitive relationship between Kentucky bluegrass, red fescue, and perennial ryegrass, it is most likely that the faster emergence of perennial ryegrass and red fescue results in a competitive advantage which may, at least partly, explain the poor establishment of Kentucky bluegrass in mixtures with perennial ryegrass and red fescue. Little is known, however, about the relationship between the timing of emergence of Kentucky bluegrass and the establishment success of this species in mixtures with faster emerging species. The aim of this study was to elucidate the role of an initial time advantage for Kentucky bluegrass establishment when establishing species mixtures. The study examines the effect of delaying the sowing of perennial ryegrass and red fescue seeds compared with the sowing of Kentucky bluegrass when establishing triple species mixtures for turfgrass. In contrast to previous work, this study takes into account the effect of soil temperature during the establishment period and quantifies on a °day scale the effect of differential sowing time on the establishment of Kentucky bluegrass.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The field experiments were performed at the Royal Veterinary and Agricultural University experimental farm Højbakkegaard at Taastrup, Denmark, on a sandy loam (246 g kg–1 coarse sand, 388 g kg–1 fine sand, 174 g kg–1 silt, 168 g kg–1 clay, 24 g kg–1 humus, pH 6.8). The cultivars used for the experiments were red fescue cv. Symphony, perennial ryegrass cv. aya, and Kentucky bluegrass cv. Andante. Seeds were provided by DLF-Trifolium, Denmark, and seeds for the 2001 experiments were produced in 2000, and seeds for the 2002 experiments were produced in 2001. The seed mixture composition by weight consisted of 30% red fescue, 40% perennial ryegrass, and 30% Kentucky bluegrass, and a seed rate of 2.5 kg seed mixture 100 m–2 was applied. The germination percentage of the applied red fescue seed lots was 84% in both 2001 and 2002, whereas perennial ryegrass germinated 96% and 94% in the two years, respectively, and Kentucky bluegrass germinated 84 and 62%, respectively. When correcting for 1000-seed weight, purity, and viability, the mixture consisted of 24% pure viable seeds of red fescue, 17% of perennial ryegrass, and 59% of Kentucky bluegrass in the 2001 experiments and 27, 23, and 50%, respectively, in the 2002 experiments.
To ensure an even distribution of seeds, the amount of seeds for each plot was divided in two equal portions, which were sown by hand with a sieve with specific mesh width for each species. Before each sowing Kentucky bluegrass, the soil was lightly cultivated to a 1-cm depth, with a hand rake. Seeds of Kentucky bluegrass were covered by raking lightly after sowing, being careful not to bury the seeds too deeply, and the plot was rolled afterwards to ensure appropriate seed-soil contact. In order not to bury or damage germinating Kentucky bluegrass seed, no soil cultivation was done before sowing red fescue and perennial ryegrass; the seeds were sown directly on the soil surface and were then covered by a layer of 8 mm topdressing sand.
The experimental area was irrigated during the first weeks after sowing if there was insufficient rainfall to ensure a moist seedbed. Before establishment of each experiment, the area was fertilized with 32 kg N ha–1, 8 kg P ha–1, 24 kg K ha–1, and 48 kg Mg ha–1. Experiments established in May were subsequently fertilized with three applications of 19 kg N ha–1, 5 kg P ha–1, 14 kg K ha–1, and 2 kg Mg ha–1, whereas trials established in August did not receive any further fertilization within the year of establishment. In the year after establishment, all trials were fertilized with five applications of 18 kg N ha–1, 4.5 kg P ha–1, 13 kg K ha–1, and 2 kg Mg ha–1. Plots were mowed once or twice weekly at a height of 4 cm with the first mowing being done when grass height exceeded approximately 5 to 6 cm. Clippings were removed from the area. Hand weeding was done to avoid competition from other plant species.
Soil temperature was recorded throughout the 35 d of the establishment period of the four experiments. Two temperature sensors (model 1400-103, LI-COR, Lincoln, NE) were placed at 1-cm soil depth. Temperature was monitored every minute, and the mean temperature was logged every hour by a LI-COR data logger (model LI-1400).
The effect of difference in sowing time was evaluated in May the year after establishment, i.e., 12 and 9 mo after establishment of the May and August experiments, respectively. The botanical composition of each plot was evaluated by core sampling, which is an appropriate method for detecting frequently occurring species in a turf (Mahdi and Stoutemeyer, 1953; Lush and Franz, 1991). Eight core samples of 34-mm diameter and 10-cm depth were sampled from each plot in a predetermined pattern, which allowed representative sampling from the plot area. The eight sampling positions within a plot were fixed by a frame, which ensured the same sampling positions in all plots. The samples were taken within the central 1.0 by 1.0 m2 of the plot to avoid effects from neighbor plots. The grass plants in the core samples were allowed to grow in a greenhouse for 2 wk before counting the number of tillers of each species in each core sample to ensure correct identification of the species.
Statistical Analysis
The relationship between difference in sowing time and the percentage of tillers of each of the three species was analyzed by linear and nonlinear regression, by the GLM and NLIN procedures of the SAS package (SAS, 2000). The analysis was based on percentage tillers rather than number of tillers since these two measures resulted in similar results (data not shown) and since the relative measure is more convenient for comparisons. The response variable was percentage of tillers determined from the mean value of the eight core samples from each plot. To make the four experiments comparable, the difference in sowing time was calculated as °days with a base temperature of 0°C by multiplying the time difference in days by the soil temperature at 1-cm depth. For each species, the relationship between difference in sowing time and the percentage of tillers was analyzed by a linear model:
 | [1] |
Since the response to difference in sowing time in certain cases appeared to be nonlinear, the relationship was also analyzed individually for each species, year, month, and sowing series by a nonlinear model:
 | [2] |
 | [3] |
 | [4] |
 | [5] |
 | [6] |
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
There was a large variation in percentage tillers of red fescue, perennial ryegrass, as well as Kentucky bluegrass within each of the four experiments (Fig. 1) . The linear regression model explained 70.1, 53.0, and 78.5% of the variation in percentage tillers for red fescue, perennial ryegrass, and Kentucky bluegrass, respectively. The analysis of covariance of the linear model demonstrated that difference in sowing time contributed significantly to the variation in percentage tillers in all three species, and in perennial ryegrass and Kentucky bluegrass it explained the majority of the variation (Table 1). There were significant main effects of year in red fescue and Kentucky bluegrass and of month in perennial ryegrass and Kentucky bluegrass, whereas there was only significant effect of sowing series in perennial ryegrass. Moreover, there were a range of first, second, and third order interactions between the main effects, indicating that the effect of difference in sowing time was influenced by year, month, and/or sowing series. Visual inspection of residual plots for each of the three species indicated homogeneous variance and the plots did not imply that the prediction of the linear model deviated systematically from the observed values. In certain cases, the nonlinear model described the relationship significantly better than the linear model (Table 2, Fig. 1).
|
|
|
Effect of Difference in Sowing Time
In red fescue, the effect of difference in sowing time on percentage tillers was larger in the experiments in 2001 than in the experiments in 2002 (P < 0.001) (Fig. 1), indicated by the mean slope coefficients of –0.0176 and –0.0045% tillers °days–1 for the 2 yr, respectively. The effect did not differ between months (P = 0.854) or series (P = 0.491) (Table 1). The slope coefficient varied from –0.0231 to +0.0010% tillers °days–1, and the coefficient was significantly negative for five of the eight relationships (Table 2). Hence, if Kentucky bluegrass was sown 500 °days before red fescue and perennial ryegrass, the predicted percentage of red fescue tillers would decrease by up to 11.6%. When the slope was significantly negative, tmin indicated that a minimum difference in sowing time ranging from 133 to 339 °days was required to significantly decrease the percentage of red fescue tillers (Table 2). For sowing series A in May 2001, the effect was better described by the nonlinear model and the effect was relatively larger when the difference in sowing time was larger (Fig. 1).
In perennial ryegrass, difference in sowing time also caused a larger effect on percentage tillers in 2001 than in 2002 (P = 0.002) (Fig. 1) with mean slope coefficients of –0.0353 and –0.0192% tillers °days–1, respectively. The effect was smaller for series A than series B (P = 0.002) with slope coefficients of –0.0191 and –0.0354% tillers °days–1, respectively, but there was no difference between months (P = 0.657) (Table 1). The slope coefficient varied from –0.0638 to –0.0049% tillers °days–1 and was significantly different from zero for six of the eight relationships (Table 2). The predicted percentage tillers of perennial ryegrass would hence decrease by up to 31.9% if Kentucky bluegrass was sown 500 °days before red fescue and perennial ryegrass. When the slope was significantly negative, a minimum difference in sowing time ranging from 61 to 344 °days was required to significantly decrease the percentage of perennial ryegrass tillers (Table 2).
As for red fescue and perennial ryegrass, the effect of difference in sowing time was also larger in 2001 than in 2002 for Kentucky bluegrass (P < 0.001) (Fig. 1) with mean slope coefficients of 0.0529 and 0.0237% tillers °days–1, respectively. The effect was smaller for series A than for series B (P = 0.006) with slope coefficients of 0.0315 and 0.0451% tillers °days–1, respectively, whereas there was no effect of month (P = 0.752) (Table 1). For all eight relationships, the slope coefficient was significantly larger than zero and varied from 0.0164 to 0.0768% tillers °days–1 (Table 2). Thus, the percentage of tillers of Kentucky bluegrass was predicted to increase by up to 38.4% if this species was sown 500 °days before red fescue and perennial ryegrass. A minimum difference in sowing time ranging from 60 to 454 °days with a mean of 210 °days was required to significantly increase the percentage of Kentucky bluegrass tillers (Table 2).
The positive relationships for Kentucky bluegrass generally corresponded well to the negative relationships for red fescue and especially perennial ryegrass (Fig. 1), reflecting that the percentage of tillers of all three species adds up to 100%. For both perennial ryegrass and Kentucky bluegrass, the nonlinear model was superior to the linear model for sowing series A in May 2001 and sowing series A and B in August 2001 (Table 2), suggesting that the increase in percentage tillers was relatively larger at larger differences in sowing time (Fig. 1).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The difficulties of establishing Kentucky bluegrass are clearly emphasized when comparing the 59 and 50% of pure viable Kentucky bluegrass seeds sown in 2001 and 2002, respectively, with the mean tiller frequencies of 18.0 and 4.9%, respectively, in the established turfgrass mixture. The results confirm those of Niehaus (1976) who reported that tiller frequencies of Kentucky bluegrass in mixtures with perennial ryegrass never exceeded 70.5% and often were much lower despite Kentucky bluegrass constituting from 79.7 to 98.1% of the number of seeds in the seed mixture. It was noted that seeds of Kentucky bluegrass germinated after perennial ryegrass had developed a relatively heavy cover, resulting in small Kentucky bluegrass seedlings and high seedling mortality. Mehnert (1981) found that a seed mixture with 50% by number of Kentucky bluegrass and 50% of perennial ryegrass resulted in less than 10% Kentucky bluegrass in the established turfgrass three months after sowing, but fifteen months after sowing the proportion had increased to approximately 70%. This stresses that the timing of the botanical analysis can be of great importance. The composition of a turfgrass community is continuously in a state of change, but the rate of change varies depending on factors such as the age of the community as well as the environmental and cultural factors it is exposed to. Since the balance between species in a turfgrass mixture may never reach equilibrium, the result of the botanical analysis is likely to depend on–among many other factors–the timing of the analysis. Perennial ryegrass often dominates the turfgrass composition in the early stages after establishment but decreases in later stages (Skirde, 1975; Engel and Trout, 1980), although it may also increase (Schmidt and Taylor, 1981). Consequently, the difference in mean tiller frequencies for perennial ryegrass and Kentucky bluegrass between May and August may be explained by the age of the turf, since turfgrass sown in May had a 3-mo longer period for growth before botanical analysis. Alternatively or additionally, the difference may reflect an effect of sowing date, which is often noticed (Watschke and Schmidt, 1992).
Differences in botanical composition between the two experimental years are very distinct for red fescue and Kentucky bluegrass (Table 2). Besides of general variation between years, this difference is presumably related to the seed quality of Kentucky bluegrass. First of all, the percentage of pure viable seeds in the sown mixture was 59% in 2001 and 50% in 2002, and a lower percentage of Kentucky bluegrass tillers must be expected in 2002. Second, the laboratory germination percentage of Kentucky bluegrass seeds was 84% in 2001 and 62% in 2002, i.e., the seed lot used in 2002 had an unusually low germination percentage. Low germination percentage is often accompanied by low seed vigor and poor germination under field conditions (Ellis and Roberts, 1980; Roberts, 1986). Consequently, a slower emergence and lower emergence percentage could be expected for the viable seeds sown in 2002 than for the viable seeds sown in 2001, resulting in a smaller proportion of Kentucky bluegrass tillers in the 2002 experiment. Interestingly, the smaller proportion of Kentucky bluegrass was accompanied by a corresponding increase in the proportion of red fescue but no marked change in the proportion of perennial ryegrass (Table 2). This may indicate that perennial ryegrass was not markedly affected by the competition from Kentucky bluegrass whereas red fescue establishment was favored by a lower frequency of Kentucky bluegrass.
Effect of Difference in Sowing Time
Poor establishment of Kentucky bluegrass in mixture with other grass species can potentially be due to one or more of the following factors: (i) low germination percentage due to poor seed vigor; (ii) low emergence percentage of germinated seeds due to poor seedling vigor; (iii) reduced development and increased mortality of emerged seedlings due to slow emergence; and (iv) reduced development and increased mortality of emerged seedlings due to slow seedling growth of emerged seedlings. In the present study, the timing of emergence and the emergence percentages were not recorded, and it is not possible to distinguish between the different factors. Previous work, however, has shown that the grass species differ markedly in emergence percentage. Hence, Davies (1927) observed mean emergence percentages of 33% for red fescue, 53% for perennial ryegrass, and 11% for Kentucky bluegrass, whereas Skirde (1967) reported 57%, 75%, and 37%, respectively. Low emergence percentage is, therefore, very likely to contribute to the relatively low frequency of Kentucky bluegrass in mixtures with red fescue and perennial ryegrass. On the other hand, the obvious effect of delaying the sowing of red fescue and perennial ryegrass (Table 2, Fig. 1) suggests that slow emergence and/or slow seedling growth of Kentucky bluegrass is also involved in the poor establishment of Kentucky bluegrass in mixtures. In the present study, the effect of slow emergence and slow seedling growth of Kentucky bluegrass cannot be distinguished. It is well known that Kentucky bluegrass emerges more slowly than particularly perennial ryegrass (e.g., Pommer, 1972; Bø, 1989). It has also been shown that Kentucky bluegrass has a slower seedling growth after emergence compared to perennial ryegrass (Arnott and Jones, 1970; Henderlong, 1971). It is, therefore, most likely that poor establishment of Kentucky bluegrass is due to a combined effect of slow emergence and slow seedling growth.
Although the covering with topdressing sand may have had a negative effect on Kentucky bluegrass establishment, delayed sowing of red fescue and perennial ryegrass clearly had an increasing effect on the percentage of Kentucky bluegrass and in most cases a decreasing effect on the percentage of red fescue and particularly perennial ryegrass when measured after nine or twelve months (Table 2, Fig. 1). This is consistent with results of Ross and Harper (1972) who studied the effect of emergence time on growth of individual plants of cocksfoot (Dactylis glomerata L.). First of all, early emergence of a plant resulted in a longer growing period compared with that of later emerging plants. Moreover, growth of later emerging plants was further reduced by competition from the early emerging plants, and a difference of approximately 10 d in emergence time resulted in a 1000-fold difference in plant weight. This is due to asymmetric competition in which the early emerging plants pre-empt more space and attain a higher position in the dominance hierarchy (Begon et al., 1996), primarily because of competition for light (Wilson, 1988). Such asymmetric competition is most likely also affecting the competition between Kentucky bluegrass and faster emerging species. The delayed seeding of red fescue and perennial ryegrass probably improves the establishment and the position in the dominance hierarchy of Kentucky bluegrass partly by allowing a longer growth period and partly by reducing the competition during the establishment phase. Kentucky bluegrass cultivars may, however, differ significantly in their competitive ability as exhibited by variable aggressiveness when growing cultivar blends in various conditions (Lickfeldt et al., 2002). Choosing a cultivar other than Andante may, therefore, also have affected the competitive balance between the three species.
Although not significant in all cases, the relationship between difference in sowing time and percentage tillers often appeared to be better described by the nonlinear function than the linear function, especially for Kentucky bluegrass and perennial ryegrass (Fig. 1). The nonlinear relationships generally showed a relatively larger effect when difference in sowing time was large. These results may indicate that a minimum difference in sowing time of more than approximately 400 °days is required for substantially increasing the percentage of Kentucky bluegrass (Fig. 1, e.g., August 2001 Series A and B, August 2002 Series A). The estimates of tmin indicated that a minimum sowing time advantage ranging from 60 to 454 °days was required to significantly increase the proportion of tillers in the established turf, corresponding to 4 to 29 d within the present temperature regime. A logical explanation of these minimum requirements may be that Kentucky bluegrass often requires up to three weeks to emerge, and if seedlings had not emerged when sowing the seeds of the other species and covering with the layer of topdressing sand, the effect may have been low. The relationship between difference in sowing time and percentage contribution to dry matter production for compact brome and ripgut brome exhibited a sigmoidal pattern, reaching an asymptote at approximately 90% (Harper, 1961). In the present study, it is conceivable that the relationship would also be sigmoidal, if the difference in sowing time was prolonged infinitely with a theoretical asymptote of nearly 100% Kentucky bluegrass. This would resemble a situation where Kentucky bluegrass was established in pure stand and subsequently overseeded with other species, and this would potentially result in a very large proportion of Kentucky bluegrass, provided there was no severe invasion of other species during the establishment phase.
The linear relationships indicated that the proportion of Kentucky bluegrass tillers could be increased by up to 0.08% for every °day the species was sown before red fescue and perennial ryegrass (Table 2). Thus, within the temperature regime of the experiments, it was possible to increase the percentage Kentucky bluegrass tillers by up to 1.2% per day of sowing difference. These results demonstrate a large potential for increasing the proportion of Kentucky bluegrass in mixture by use of differential sowing time for the species.
An important factor for the effect of differential sowing is the soil temperature between the two sowings. Since the effect is related to the accumulation of °days, a larger effect would be obtained from a certain time difference when soil temperature is high than when soil temperature is low. Consequently, if the sowings in the present experiment were performed at a colder period of the year than May and August, e.g., in April or September, a larger difference in sowing time would be required to obtain the same effect. The largest effect of differential sowing time must, therefore, be expected when soil temperature is optimal for germination of Kentucky bluegrass.
A number of studies have investigated the possibilities of enhancing germination and establishment of Kentucky bluegrass by seed priming treatments (e.g., Lush and Birkenhead, 1987; Yamamoto et al., 1997a; Pill and Necker, 2001). Priming can reduce the time to germination of Kentucky bluegrass by some days, depending on the germination temperature, but the time to germination is not reduced to the level of nonprimed perennial ryegrass seeds (Lush and Birkenhead, 1987; Yamamoto et al., 1997a). Besides, seedling growth after emergence is not affected by priming (Yamamoto et al., 1997b). Lush and Birkenhead (1987) showed that priming of Kentucky bluegrass seeds could increase the proportion of tillers from 58 to 73% in mixtures with perennial ryegrass. The present results demonstrate, however, that the effect obtained by priming may only have a limited effect on the frequency of Kentucky bluegrass, since a larger time advantage is generally required.
From a practical perspective, it may be inconvenient to sow individual species of a mixture on separate sowing dates. In particular, there is a practical challenge in sowing seeds of perennial ryegrass and red fescue without burying to deeply or damaging previously sown seeds of Kentucky bluegrass. A potential way of accomplishing this may be to carry out the delayed sowing by hydroseeding, which may not affect seedling growth of Kentucky bluegrass (Macphail et al., 1980; Pill and Nesnow, 1999).
|
ACKNOWLEDGMENTS |
|---|
Received for publication August 7, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Related articles in Crop Science:
This article has been cited by other articles:
|
|
 |
|
  S. U. Larsen and B. M. Bibby Differences in Thermal Time Requirement for Germination of Three Turfgrass Species Crop Sci., August 26, 2005; 45(5): 2030 - 2037. [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 | |||
