A. Elgersma, H. Schlepers and J.H. van WenumWageningen Agricultural University, Department of Agronomy, Haarweg 333, 6709 RZ Wageningen, The Netherlands
Combinations of varieties of perennial ryegrass and white clover with contrasting growth habits are evaluated under cutting and grazing to study white clover persistence in the Netherlands. This paper presents the first results of the cutting trial.
The cutting trial was established in April 1991 on heavy river clay at Wageningen, The Netherlands. Previous use had been arable land. The soil was fertilised with P and K, but no N was applied.
Certified seed was used of the perennial ryegrass varieties Condesa (tetraploid) and Barlet (diploid, erect), and of the white clover varieties Alice (large-leaved), Retor (medium-leaved) and Gwenda (small-leaved). There were six mixtures and two grass monocultures. The experimental design was a split-plot with cutting frequency as main factor, mixture as subfactor and three replications. In total there were 48 plots. Each plot was 1.5 x 10 m. In the mixtures, grass and clover were sown mixed at a row width of 8 cm. Five kg/ha clover seed and 20 kg ha-1 Barlet or 30 kg ha-1 Condesa were used, to compensate for the different seed weight of the grass varieties. In the monocultures the same quantities of grass seed were sown.
Starting in 1992, all entries were cut at either approximately 1200 or 2000 kg ha"' DM (treatments will be abbreviated as F1200 and F2000, respectively). The expected yield was estimated by measuring sward height with a rising plate. When the average expected yield of the mixtures was 1200 or 2000 kg ha-1, all eight entries at the relevant cutting frequency were harvested on one day, although the grass monocultures yielded less. The plots were cut with a Haldrup forage harvester at 5 cm sward height. A core sample of the cut material was taken for assessment of the percentage dry matter. A pluck sample of about 0.125 kg was separated into grass, clover and weeds. The samples were dried at 70 °C during 24 hours. Dry matter (DM) yield of grass, clover leaves and clover flowers (if present) were calculated per plot.
Analyses of variance were carried out for DM total yield, grass yield and clover yield. The dried materials were bulked per grass/clover combination, grass and clover separately, and ground to determine the N-concentration. The N-fixation was calculated by subtracting the N-yield of the monocultures from the total N-yield of the mixtures. The N-transmission was calculated by subtracting the N-yield of the monocultures from the N-yield of the grass harvested in the mixtures.
Establishment and management
Because of the dry April and cold May month, germination was slow and somewhat irregular. The whole trial was cut in the middle of June, mainly to remove weeds. Other cuts were carried out in July, mid September and the end of October. The winter 91/92 was mild and no winter damage was detected. Apart from some slug damage and Sitona weevil dagame in the leaves, no diseases were observed. During winter field mice were poisoned.
During 1992, the summer was again very warm. The trial was not irrigated. The F1200 was cut 9 times and the F2000 7 times. The monocultures performed very poorly due to shortage of N and to drought. The mixtures performed very well and were much less affected by drought. Retor showed more symptoms of drought stress than Alice and Gwenda, and Alice performed best during the drought.
Total yield
At all harvests, the yield of the monocultures was far less than that of the mixtures. The mixtures with Alice yielded most during the whole season, for both cutting frequencies. The mixtures with Retor and Gwenda had similar yields in spring and summer, but in autumn Retor yielded less than Gwenda. Table 1 shows the effect of cutting frequency on yield.
Table 1. The effect of cutting frequency on dry matter yield (t ha-1) of grass/clover mixtures.
|
|
F1200 |
Condesa |
F2000 |
Condesa |
|
Barlet |
Barlet |
|||
|
Clover |
|
|
|
|
|
Alice |
12.7 |
12.9 |
14.1 |
14.6 |
|
Gwenda |
11.0 |
10.8 |
12.8 |
12.5 |
|
Retor |
10.6 |
10.6 |
12.6 |
11.7 |
|
none |
2.0 |
1.2 |
2.0 |
1.7 |
There was a highly significant (P <. 0.001) effect of mixture, which was caused by the clover variety. There was no effect of the grass variety. No interaction was found between cutting frequency and mixture. The mixtures with Alice yielded significantly more than the other mixtures, and all mixtures yielded much more than the monocultures. The effect of cutting frequency was just not significant (P = 0.052); the F2000 yielded more than the F1200 except in Barlet.
Grass and clover yield
The effect of cutting frequency on the yield of grass and the yield of clover in the six mixtures was not significant, although the F2000 had higher yields. The effect of mixture was highly significant (P: S 0.001) for both grass yield and clover yield. There was no mixture by cutting frequency interaction. The mixtures with Alice yielded significantly less grass and more clover. Condesa with Alice and Gwenda yielded more grass than Barlet with Alice and Gwenda, respectively. The clover yield was not affected by the grass variety. The clover content ranged from 60 - 90 % (Fig. 1).
Nitrogen
The N-concentration in the F1200 was always slightly higher than that in the F2000. The grass harvested from the mixtures contained 2.5 - 4 % N, whereas the grass from the monocultures contained 1.6 - 3 % N. The total N-yield, N-fixation and N-transmission are presented in Table 2.
Fig. 1. The percentage white clover DM in the harvested material in six grass/clover mixtures.
Table 2. The effect of cutting frequency on N-yield, N-fixation and N-transmission (kg N ha-1) of grass/clover mixtures.
|
|
F1200 |
F2000 |
||||||||||
|
Barlet |
Condesa |
Barlet |
Condesa |
|||||||||
|
Clover |
|
|
|
|
|
|
|
|
|
|
|
|
|
Alice |
559 |
513 |
40 |
559 |
534 |
78 |
565 |
524 |
47 |
599 |
565 |
61 |
|
Gwenda |
480 |
434 |
62 |
460 |
435 |
88 |
516 |
475 |
69 |
480 |
455 |
79 |
|
Retor |
446 |
400 |
65 |
453 |
428 |
79 |
497 |
456 |
74 |
457 |
423 |
81 |
|
none |
46 |
|
|
25 |
|
|
41 |
|
|
34 |
|
|
A lower yield of the monocultures was expected, but their yield was extremely low due to summer drought. Also, the organic matter content of the soil, being former arable land, was low. The N from the monocultures has come from mineralisation and deposition, the latter being estimated at about 45 kg N ha-1 y-1 in Wageningen. The relatively high yield of Barlet monoculture at frequent cutting can be explained by its erect growth habit. The plots with Barlet all were rather open, with erect, separate, grass plants. Companion grass variety did not affect the clover content in the mixtures, but there was an effect on the grass content: Condesa had a higher grass yield when combined with Alice and Gwenda, than Barlet combined with Alice and Gwenda, respectively.
In general the effect of grass variety was much less pronounced than the effect of clover variety. The mixtures with Alice produced the highest total yield, the highest clover yield and the lowest grass yield. This is not surprising, Alice being the clover variety with the largest leaves. Large-leaved varieties are productive and well adapted to cutting. The clover content was higher than desirable, but it is expected that the grass yield will increase in 1993 because more N will probably be available. N-fixation and N-cycling will be studied further. Retor (a Dutch variety) declined in the autumn of 1992, and its spring growth in 1993 was also delayed compared to that of Alice and Gwenda (both Welsh-bred varieties). The experiments will be continued to study persistence.
