Previous Page Table of Contents Next Page

Agronomic and feeding value of white clover - grass binary mixtures for the dairy systems of Southern Lombardy (Italy)

P. Annicchiarico and N. Berardo Istituto Sperimentale per le Culture Foraggere, viale Piacenza 29, 20075 Lodi, Italy

Comparison of white clover-based vs. lucerne-based legume-grass binary mixtures
Evaluation of binary mixtures of white clover with different grass species and varieties
Evolution of feeding value in function of clover content in binary mixtures of white clover with different grass species


Italian milk production is progressively concentrating on the northern part of the country and particularly on those regions, like southern Lombardy, whose farming systems can combine large farm and herd size with rather high production per cow and which are increasingly directed to high quality production for cheese making (Zuppiroli, 1988; Gian and Mosso, 1991). In these systems, reasons related to milk quality (Annibaldi, 1988), conservation of soil fertility and reduction of nitrogen leaching (Spallacci et al., 1988) and profitability (Gian and Mosso, 1991) suggest that forage production should also rely on high yielding, mowed rotational meadows beside cropping of annual cereals for ensiling. However, only forages of high quality should be included in the rations to maintain high values of digestibility and voluntary intake of the feed and maximize, therefore, milk yields while providing the necessary amount of fibre to the animals (Succi et al., 1988; Mertens, 1992). Evaluation of fodder crops targeted to these dairy systems needs therefore to pay special attention to feeding value.

The present study aimed at: i) comparing, particularly in terms of nutritive value, a set of legume - grass binary mixtures including Ladino white clover with other mixtures including lucerne as the legume component; ii) evaluating yield and quality of binary mixtures of white clover with different species and varieties of grass.

Nitrogen fertilisation and vigour of the associated grass can influence both yield and content of clover, i.e. nutritive value, of white clover-based mixtures (Frame and Newbould, 1986; Harris, 1987). Maximization of both quantity and quality often proves difficult, and knowledge of the minimum content of clover compatible with good feeding value can provide a basis for sound technical decisions. A further aim of this study was, therefore, iii) preliminarily assessing the evolution of feeding value as a function of clover content in binary mixtures of white clover with different grass species.

Comparison of white clover-based vs. lucerne-based legume-grass binary mixtures

The white clover 'Gigante lodigiano' and the lucerne 'Cremonese' were grown in binary mixture with each of 3 varieties of cocksfoot and 3 varieties of tall fescue near Cremona (southern Lombardy) for 3 years. Though in separate trials, the white clover and the lucerne mixtures were grown contemporarily in the same field and under the same management which included border irrigation in summer and 125 kg/ha of N per year beside P and K fertilisation. Information on dry matter (DM) yield and content of legume on green weight, already published by Onofrii et al. (1988a; 1988b), is reported in Table 1 without statistical comparisons.

Forage samples were collected on each cut of the second year, dried at 60 ºC and ground to 1 mm size. They were analysed for content of crude protein (CP), as 6.25 x N by the Dumas procedure (Kirsten, 1983) and utilizing a Carlo Erba Nitrogen Analyser, and for content of crude fibre (CF), following the Weende method. Content of neutral (NDF) and of acid (ADF) detergent fibre were determined via near infrared reflectance spectroscopy using a calibration equation developed in Lodi which had proved capable to explain about 96 % and 97 % of variation for these traits in a large set of forages (N. Berardo. unpublished results).

The nutritive value of the forages was estimated: i) as digestible DM supply expressed as milk feed units (MFU), computed from CP and CF according to Andrieu et al. (1981, p. 79 Table 1) and from ADF through estimation of net energy for lactation according to Northeast Dairy Herd Improvement Association (DHIA) and its conversion to MFU (Chase, 1981); ii) as Relative Feed Value (RFV), a popular index in the United States providing a synthetic assessment of quality based both on potential DM intake of the feed, mainly depending on NDF content (Mertens, 1992), and on digestible DM estimated by ADF content. The index (Ball et al., 1991) was computed according to the equation:

RFV = ((88.9 - 0.779 ADF%) x (120/NDF%))/1.29

Since time and methods of cutting, sampling and chemical analyses were the same for the two trials, data on quality of the forage were jointly submitted to analysis of variance (ANOVA) retaining the average variation within plot replicates as the error term. The data of the randomized complete blocks included in the original experiment design were previously averaged over the 3 varieties of each grass species in order to gain information on the average response of these species.

Table 1. Three-year DM yield and legume percent on green weight, and feeding value estimated as milk feed units per kg DM (MFU) and Relative Feed Value (RFV) in the second year of the crop cycle, in binary mixtures including white clover (WC), lucerne (Lu), cocksfoot (Co) and tall fescue (TF)

First cutc

Mean of 2nd to 4th cutc


DMb yield t/ha

Legumeb %







WC - Co



0.77 a

0.75 a

98 a

0.79 b

0.77 a

103 a




0.76 a

0.74 a

97 a

0.82 a

0.79 a

107 a

Lu - Co



0.78 a

0.69 b

96 a

0.80 ab

0.70 c

105 a

Lu - TF



0.79 a

0.73 a

101 a

0.81 ab

0.74 b

108 a

a averaged over 3 varieties for each grass species

b from Onofrii et al. (1988a; 1988b)

c means with same letters do not differ at P £ 0.05 according to Newman-Keuls test

d according to Andrieu et al. (1981)

e according to Northeast Dairy Herd Improvement Association (Chase, 1981)

f according to Ball et al. (1991)

White clover was generally at rather strong competitive disadvantage (Table 1), as commonly found when grown in association with a grass (Frame and Newbould, 1986). Its mixtures, compared to the lucerne-based ones, showed on the whole a similar nutritive value though retaining just about a half of the content of legume. This was likely possible because of the outstandingly low crude fibre content of white clover. The relatively higher MFU of the white clover-based mixtures according to DHIA's estimation may be possibly due to a relatively lower portion of ADF within the fibre fraction of the clover. The lucerne-grass mixtures had a crude protein content about 1 % lower at the first cut and about 1.6 % higher on the average of the other cuts. Since forages of good and of just fair quality should have RFV values respectively above 125 and 103 (Ball et al., 1991), all mixtures at the first cut could be ranked as low quality. Their nutritive value could be improved by anticipating that cut (Succi et al., 1988). Combining DM yield and quality information, the white clover-cocksfoot mixtures seemed somewhat worse performing than the others.

Evaluation of binary mixtures of white clover with different grass species and varieties

Within the framework of a competition study, six white clover genotypes mostly belonging to the Ladino type were grown together both as a monoculture and in mixture with each of eight grasses (Table 2) for 2 years in Lodi (southern Lombardy). Beside P and K, all theses were given 40 kg/ha of N at establishment of the trial. In addition, the mixtures received 40 and 70 kg/ha of N respectively during the first and the second year of the crop cycle while the grass monocultures, also included in the experiment, were given a double amount of these rates. The trial was regularly irrigated during summer. DM yield was recorded on 2 cuts in the first year and 3 cuts in the second year. The last cut, representative of an ordinary summer cut, was chosen for evaluation of feeding values of the associations whose agronomic performance had proved valuable. The chemical analyses were executed as already described for the previous set of mixtures except for determination of NDF and ADF content, carried out according to the methods of Goering and Van Soest (1967).

Table 2. DM yield, clover percent on DM and Relative Yield Total (RYT) over 5 cuts, and clover percent on DM, feeding value estimated as milk feed units per kg DM (MFU) and Relative Feed Value (RFV) at the last cut, in binary mixtures of with clover with different grass species and varieties

Companion grass


Last cutb



DM yield t/ha

Clover %


Clover %











0.83 a




25.2 b









42.0 a









36.3 a






112 b







0.76 b

0.77 b

106 b

M. Kasba


29.1 b













0.84 a

0.83 a

131 a







0.84 a

0.84 a

132 a

a Co = cocksfoot; TF = tall fescue; IR = Italian ryegrass; PR = perennial ryegrass

b letters refer to separation of the top (a) and bottom (b) ranking means; same letters do not differ at P £ 0.05 according to Newman-Keuls test,

c according to Andrieu et al. (1981) according to Northeast Dairy Herd Improvement Association (Chase, 1981)

d according to Ball et al. (1991)

Table 2 provides information on yield and quality of the mixtures. Relative Yield Totals (De Wit and Van den Bergh, 1965) were also computed, showing that the associations always outyielded the average of the corresponding monocultures. Total DM yield was distinctly associated to aggressiveness of the grass regardless of its species, as indicated both by the values of yield and clover content reported in Fig. 1 and by the correlation (r = 0.92, P £ 0.01) between total yield and grass content of the mixture. Grass aggressiveness was strictly related also to grass vigour measured as DM yield in monoculture (r = 0.93, P £ 0.001). The large differences found between the 3 cocksfoot varieties for level of competitive stress exerted on the clover support the hypothesis (Harris, 1987) that the suppressing effect on white clover often reported for this species could mostly be just due, at least under a mowing regime, to adoption of vigorous varieties.

Figure 1. Total DM yield and clover percent on DM over 5 cuts of binary mixtures of white clover with different grass species and varieties (see Table 1 for indication of grass species)

Evaluation of quality was performed on all mixtures but those of 'Dama' and 'Maris Kasba', considered too low yielding, and of 'Dora', too clover suppressing (Table 2). Following the RFV information, only the associations including either of the two ryegrass varieties combined good DM yield with high quality of the forage. The lower quality values found on average, for a given grass species and a similar clover content, in the summer cuts of the previous set of mixtures could be partly attributed to the inclusion in those mixtures of a certain amount of low quality weeds (Onofrii et al., 1988a; 1988b) beside the fact that clover content in that case referred to green weight.

Evolution of feeding value in function of clover content in binary mixtures of white clover with different grass species

This study was limited to those mixtures and that forage cut already considered for evaluation of quality in the previous section of the paper. The same analytical determinations were executed in this case on samples obtained by combining the components of the mixtures, previously separated and oven dried, according to clover to grass DM ratios of 100: 0, 60: 40, 40: 60, 20: 80 and 0: 100.

In each mixture, the feeding value distinctly increased as a function of clover content according to all the quality parameters. The milk feed units (Pig. 2) approached a level of 0.99 per kg DM corresponding to 100 % clover whereas rather large differences among mixtures became apparent only under 60 % clover. Minimum clover contents of about 10 %, 20%, 30 % and over 35 % were needed in the mixtures including respectively the Italian ryegrass, the perennial ryegrass, the cocksfoot and the tall fescue varieties to reach good feeding value defined (Ball et al., 1991) as RFV not lower than 125. Contents about 5 % higher were necessary when good quality was defined (Succi et al., 1988; Ball et al., 1991) as NDF content not higher than 47. The tall fescue 'Magno' needed almost 5 % more clover than 'Festorina' to reach a similar feeding value suggesting that differences for minimum clover content may exist not only between but also within species of companion grass. Giovanni (1990), studying a rotational pasture exploited by sheep at 6 week intervals, had already found that 30 % white clover in the DM was compatible with animal needs in the associations including a perennial ryegrass or a cocksfoot but it was too low in the mixture with a tall fescue. About 30 % clover in the annual DM yield is usually regarded as a desideratum in white clover - perennial ryegrass associations (Frame and Newbould, 1986). Clover content should not drop anyway below 20 % on an annual DM basis if its contribution to N fixation is to be relevant (Thomas, 1992).

Figure 2. Evolution of milk feed units (MFU; Andrieu et al., 1981) in function of clover percent on DM in binary mixtures of white clover with different grass species (see Table 1 for indication of grass species)

The present findings needs confirmation in particular from data referring to first seasonal cuts. However, the occurrence in that cut: of the heading of cocksfoot and tall fescue is not expected to reduce the relative differences for the minimum needed clover content between these species and the ryegrasses. Given the large year-to-year variation for clover content typical of the white clover-based mixtures (Frame and Newbould, 1986; Annicchiarico and Onofrii, 1993), the capability of the ryegrasses to minimize the probability of obtaining a low quality forage emphasizes their value as companion grass of white clover for the dairy systems of the region. The association with Italian ryegrass, though tending to disfavour clover growth rather severely, can provide the further advantage of facilitating a possible ensiling due to its high content of soluble carbohydrates (Demarquilly et al., 1982).


ANDRIEU J., DEMARQUILLY C. and WEGAT-LITRE E. (1981) Prevision de la Valeur Nutritive des Aliments des Ruminants. INRA, Versailles.

ANNIBALDI S. (1988) Alcuni aspetti della qualità casearia del latte. In: II Futuro della Foraggicoltura Prativa nella Pianura Padana, pp. 75-88, 1st. Sper. Colture Foraggere, Lodi.

ANNICCHIARICO P. and ONOFRII M. (1993) Influence of climatic factors on clover content of white clover - tall fescue rotational meadows in southern Lombardy (Italy). Proceedings White Clover Meeting of the FAO Sub-network on Lowland Pastures and Fodder crops, Arhus, Denmark (in press).

BALL D.M., HOVELAND C.S. and LACEFIELD G.D. (1991) Forage quality. In: Southern Forages, pp. 124-132. Potash and Phosphate Institute and Foundation for Agronomic Research, Atlanta.

CHASE L.E. (1981) Energy Prediction Equations in USA at the NY DHIA Forage Laboratory. Production Agricultural Training School, Ithaca.

DEMARQUILLY C., RAYMOND P. and DEDENON N. (1982) Valeur alimentaire et utilisation par les ruminants de la luzerne et du trèfle violet. Fourrages, 90, 181-202.

DE WIT C.T. and VAN DEN BERGH J.P. (1965) Competition between herbage plants. Netherlands Journal of Agricultural Science, 13, 212-221.

FRAME J. and NEWBOULD P. (1986) Agronomy of white clover. Advances in Agronomy, 40, 1-88.

GIAN B. and MOSSO A. (1991) Economia delle utilizzazioni dei prati e del pascoli. Rivista di Agronomia, 25, 170-183.

GIOVANNI R. (1990) La prairie graminée - trèfle blanc I. Valeur alimentaire du trèfle blanc et de l'association. Fourrages. 121, 47-63.

GOERING H.K. and VAN SOEST P.J. (1967) Forage Fiber Analysis. Agricultural Handbook 379, USDA, Washington.

HARRIS W. (1987) Population dynamics and competition. In: M.J. Baker and W.M. Williams (eds) White clover, pp. 203-297. CAB International, Wallingford.

KIRSTEN W.J. (1983) Rapid, automatic, high capacity Dumas determination of nitrogen. Microchemical Journal. 28, 529-547.

MERTENS D.R. (1992) Nonstructural and structural carbohydrates. In: H.H. Van Horn and C.J. Wilcox (eds) Large Dairy Herd Management, pp. 219-235. American Dairy Herd Association, Champaign.

ONOFRII M., TOMASONI C. and BERARDO N. (1988a) Consociazioni binarie di graminacee e trifoglio ladino. Confronto tra miscugli ed epoche di sfalcio. Agricoltura Ricerca, 87, 49-56.

ONOFRII M., TOMASONI C. and BERARDO N. (1988b) Consociazioni binarie di graminacee ed erba medica. Confronto tra miscugli ed epoche di sfalcio. Agricoltura Ricerca, 87, 57-64.

SPALLACCI P., PAPINI R., MICLAUS N., GREGORI E, COSTANTINI E.A.C., BRANDI G., BRANDI D., SASSO A. and TOMASONI C. (1988) Effetti di sistemi foraggero-zootecnici su alcuni aspetti di fertilità del terreno. Agricoltura Ricerca, 87, 71-90.

SUCCI G., CROVETTO G.M. and SANDRUCCI A. (1988) Il ruolo del foraggio nell'evoluzione delle tecniche alimentari della bovina da latte nella Pianura Padana lombarda. In: Il Futuro della Foraggicoltura Prativa nella Pianura Padana. pp. 97-126. 1st. Sper. Colture Foraggere, Lodi.

THOMAS R.J. (1992) The role of the legume in the nitrogen cycle of the productive and sustainable pastures. Grass and Forage Science, 47, 133-142.

ZUPPIROLI M. (1988) Gli scenari zootecnici e foraggicoli della Pianura Padana. In: II Futuro della Foraggicoltura Prativa nella Pianura Padana. pp. 31-46. 1st. Sper. Colture Foraggere, Lodi.

Previous Page Top of Page Next Page