Previous Page Table of Contents Next Page


Wageningen Agricultural University, Department of Agronomy, Haarweg 333, 6709 RZ Wageningen, The Netherlands


The persistence of white clover in grass-clover mixtures was studied in plots, continuously stocked with cattle. The effect of grass cultivar on clover content and animal production were studied at three sites. Throughout the years, mixtures with the prostrate diploid perennial ryegrass cv. Wendy formed a dense sward with less weeds than mixtures with the tetraploid cv. Condesa or the erect diploid cv. Barlet. Tetraploid cv. Condesa had the lowest tiller density. In 1992-1994 the clover content was highest in mixtures with Wendy. These differences between mixtures did not result in differences in animal production. Despite incidental 'clover crashes' the small-leaved white clover cv. Gwenda persisted in all paddocks more than seven years after sowing under the continuous stocking management imposed.

Keywords: grazing, stolons, white clover, grass cultivars, animal production


It is very difficult to maintain desired levels of white clover (Trifolium repens L.) within the sward over an extended period. Sudden severe reductions in clover content ('clover crashes') which seem to occur in a three to four year cycle have been reported (Fothergill et al., 1996). Clover content and persistence are affected by many factors; Schwinning and Parsons (1996) distinguished between intrinsic oscillations with a phase of three to four years, and external sources of legume variation. In our experiment the effect of companion grass on botanical composition was studied on two soil types. Grass cultivars with contrasting growth habits were included. Animal output was measured and related to clover content in the pastures.


Grazing trials were established on two river clay sites in Wageningen (HN, former grassland, and HW, former arable land) and on sandy soil at Achterberg (ACH), The Netherlands (Elgersma and Schlepers, 1994). ACH and HN were sown in April and HW in August 1991. The experimental design was a randomized block with three replications on HN and HW and two on ACH. Plot size was approximately 3 300 m². Perennial ryegrass cultivars with a contrasting growth habit: Condesa (tetraploid), Barlet (diploid, erect) and Wendy (diploid, prostrate), were sown in mixture with the small-leaved white clover cultivar Gwenda. The soil was fertilized with P and K, but no N was applied. The pastures were continuously stocked during the growing season with steers in 1991-1993, and with pregnant heifers during 1994-1997 (Elgersma and Schlepers, 1996).

Average sward height was maintained at 7-8 cm by adjusting the number of animals weekly. Rejected areas were topped at 8 cm height every month during the grazing season. Dry matter (DM) yield of clover leaves and stolons, stolon length and tiller numbers of various grass species were assessed in 20 circular 0.25 dm² cores per plot taken in April prior to the turnout of the animals and again in November. Stolon abundance (m stolon.m-2) was used to estimate clover performance. Animal liveweight gain was calculated from monthly weighings of two test animals which remained in the plot for nearly the whole growing season. Net production was calculated in terms of net energy requirement expressed as Dutch Feed Units for beef cattle (VEVI ha-1) (Elgersma and Schlepers, 1996).


From 1992-95 there were clear effects of grass cultivar on clover abundance, which were consistently lowest in mixtures with Wendy (Figure 1). In later years, however, there was no consistent cultivar effect and the ranking order of mixtures varied between sampling dates and sites. Perennial ryegrass cv. Wendy had the highest tiller density and Condesa the lowest. Despite the lower ryegrass tiller density of Condesa compared with Barlet, mixtures with Condesa did not contain more clover. Thus, also factors other than ryegrass tiller density affect clover growth in mixed swards. Mixtures with Barlet and Condesa had about the same number of tillers of the weed grass Poa annua, but mixtures with Wendy had significantly less weeds. The plots on sandy soil (ACH) contained most weeds. The clover stolon abundance varied strongly between sites and years (Figure 1), albeit not always in three to four year cycles as reported by Fothergill et al. (1996) and Schwinning and Parsons (1996). The pattern of fluctuations was different in the three trials. In ACH the peak stolon abundance occurred in October 1992, one and a half years after sowing, followed by a severe crash the next winter. Only from April 1994 onwards, was there a recovery of the stolon abundance. A second crash was in 1996 after a severe winter, followed by a quick recovery in 1997, a very favourable year for grassland production with sufficient moisture. In HN the peak did not reach the level of that in ACH. There was a peak plateau from October 1992 until October 1993, a slow decline until April 1995, and a recovery thereafter. A second, fast decline was in 1996, followed by a quick recovery in 1997. In HW the peak occurred from October 1992 (one year after sowing) until April 1993; a low stolon abundance was measured in April 1994 with a recovery thereafter; a second, lower peak was observed from October 1994 until April 1995, followed by a sharp, deep decline in 1995/96. However, as in the other sites, this was followed by a quick recovery in 1997. As the locations were within 7 km distance, the weather conditions were similar and differences will mainly have been caused by soil type and soil ecosystem associated with it. Besides an effect on the timing of fluctuations, soil type also had a large effect on the levels of species abundance and animal output.

The amount of clover DM (stolons plus leaves) in autumn (Figure 2) also depicts the low clover content in mixtures with Wendy compared to that in other mixtures within 1992-1994. However, there were no significant differences in animal performance or net energy production among the mixtures within each location. No relation was found between clover content and net energy production (Figure 2).

Under the continuous stocking management imposed for seven years, despite 'clover crashes' the small-leaved white clover cv. Gwenda had persisted in all paddocks. At present (during 1998) the plots are being rotationally stocked with cattle, integrated with silage cuts.

Figure 1. Stolon abundance of white clover (m.m-2) and tiller density of perennial ryegrass (number.m-2) in continuously stocked pastures. In the upper figures the symbol 'square' (B) indicates white clover stolon in mixtures with grass cv. Barlet; in the lower figures grass tillers in mixtures with Barlet.

Figure 2. The relation between the net energy (kVEVI ha-1) during the growing season and clover in the sward (g DM m-2), measured in autumn, for three mixtures during six years (1992-1997) at three locations.


Elgersma, A. & Schlepers, H. 1994. Contrasting perennial ryegrass/white clover mixtures under cutting and grazing. In: Editors: Mannetje L. 't & J. Frame: Grassland and Society. Proceedings of the 15th General Meeting of the European Grassland Federation. Wageningen Pers: 69-72.

Elgersma, A. & Schlepers, H. 1996. Cattle production on white clover - perennial ryegrass mixtures. In: Groen A.F. & J. van Bruchem (Eds), 1996. Utilization of local feed resources in dairy cattle. EAAP Publication no. 84, Wageningen Pers, Wageningen. pp. 142-144.

Fothergill, M., Davies, D.A., Morgan, C.T. & Jones, J.R. 1996. White clover crashes. In: Younie D. (Ed.) Legumes in sustainable farming systems, Occasional symposium of the British Grassland Society, no. 30, pp. 172-176.

Schwinning, S. & Parsons, A.J. 1996. In: Younie D. (Ed.) Legumes in sustainable farming systems, Occasional symposium of the British Grassland Society, no. 30. pp. 153-163.

Previous Page Top of Page Next Page