Ninety-four silages were made over 5 years from predominantly perennial ryegrass swards using a range of cutting dates (19 May to 18 September), wilting periods (0 to 48 h) and additives (none, acids, inoculants, sugar, sugar + acids, sugar + inoculants). A wide range of silage composition was achieved (CV for dry matter (DM), crude protein (CP), digestible organic matter (DOMD), lactic acid, total volatile fatty acids (VFA) and sugar were 0.22, 0.19, 0.07, 0.43, 0.84 and 0.69 respectively). Silage dry-matter intake (SDMI) was measured for 88 silages using lambs (mean live weight (M) 28 kg) given silage as their sole diet in four incomplete block design experiments using four lambs per silage and a standard hay given every third period for covariance correction. Thirty-four of the silages were also evaluated using early lactation cows (M, 561 kg and milk yield 27 kg/day) with 7 kg/day of concentrate in eight incomplete block change-over experiments each using 12 cows.Intakes (SDMI mean, range, s.d. g/kg M(0.75)) were 56, 25 to 84, 13.7 for lambs and 90, 64 to 119, 13.4 for cows. Scaling lamb SDMI by M(1.47) accounted best for the effect of lamb weight on intake (mean, range 5.07, 2.43 to 7.68). Silage predictors were grouped as follows: traditional values (BASAL) -DM, CP, organic matter (OM), DOMD, neutral-detergent fibre (NDF), acid-detergent fibre (ADF), ammonia nitrogen (NH(3)N), pH, acid hydrolysed ether extract (AHEE); silage fermentation values obtained by high-performance liquid chromatography (HPLC); or by electrometric titration (ET); and near infra-red reflectance spectra (NIRS) obtained on either 100 degrees C dried (NIRSdry) or fresh samples (NIRSwet1 using a vertical transport mechanism and NIRSwet2 using a rotating cup). The most useful predictors within each group were firstly identified by step-wise multiple linear regression and models were then derived by partial least squares.Standard errors of cross validation (SECV) obtained by the 'leave one out' method were for lamb SDMI (g/M(1.47)) 0.81, 0.81, 0.75, 0.52, 0.82 and 0.56 for BASAL, BASAL + HPLC, BASAL + ET, NIRSdry, NIRSwet1 and NIRSwet2 respectively. Corresponding values for cows (g/M(0.75)) were 7.3, 7.3, 5.9, 5.1, 6.2, and 2.5. Inclusion of fermentation measurements made by ET, but not by HPLC, improved SDMI prediction over that obtained from the BASAL set. However, NIRSdry and NIRSwet2 were the most accurate methods giving values for s.d. (reference population)/SECV of 2.27 and 2.13 for lambs and 2.65 and 5.28 for dairy cotes. Use of these methods in advisory silage evaluation should substantially reduce the errors of predicting the intake potential of grass silages.

To investigate the effect of dietary fat and metabolizable energy (ME) on milk protein concentration, an experiment was carried out using 12 multiparous early-lactation Holstein-Friesian dairy cows. Three diets were offered in a complete Latin-square change-over design, based on ad libitum access to grass silage. One of three concentrates was offered at a rate of 12 kg/day, each formulated to supply one of two levels of ME (12·1 and 13·6 MJ/kg dry matter (DM)) and one of two levels of fat (31 and a mean of 88 g acid hydrolysis ether extract per kg DM): low energy, high fat (LEHF); low energy, low fat (LELF); and high energy, high fat (HEHF). The concentration of milk protein was significantly higher from animals offered the LELF concentrate (32·5 v. a mean of 31·2 (s.e.d. 0·45) g/kg, P < 0·05), because of lower milk yields (31·0 v. a mean of 33·4 (s.e.d. 0·63) kg/day, P < 0·05). Animals offered the HEHF concentrate produced the highest yields of milk protein but their milk had the lowest concentrations of fat (32·5,34·4 and 31·9 g/kg for LEHF, LELF and HEHF respectively; s.e.d 1·07; P < 0·05 for difference between LELF and HEHF). Silage DM intake was significantly increased by animals offered the LEHF concentrate (9·1, 8·6 and 8·7 (s.e.d. 0·19) kg/day, P < 0·05 for differences between LEHF and the other two concentrates). Urinary purine derivative excretion, used as an index ofmicrobial protein supply, was highest from animals offered the LELF and HEHF concentrates, which both supplied similar amounts of fermentable ME. It is hypothesized that increased de novo synthesis offatty acids on the low fat diet reduced the availability of glucose for lactose synthesis, leading to reduced milk yields and hence increased milk protein concentrations.

The objective of this study was to investigate the effects of silage characteristics on water intake of lactating dairy cows and to examine the prediction of water intake. Sixteen grass silages, differing in fermentation and intake characteristics, were offered ad libitum to dairy cows in early lactation supplemented with 7 kg/day of concentrate (13·3 MJ metabolizable energy per kg dry matter (DM) and 216 g crude protein per kg DM). Four silages were offered in each of four incomplete change-over design experiments, consisting of three 3-week periods. Water intakes were recorded through individual Kent water meters and press water bowls over the final week of each period. Tree (drinking) water intake ranged from 20·1 to 89·9 (mean = 45·2; s.d. = 12·96) I/day whilst total water intake (also including food water) ranged from 48·4 to 123·8 (mean = 87·3; s.d. = 14·12) I/day. Water intake increased with increasing silage DM concentration, however free water replaced silage water at a rate less than 1. Milk yield and silage D value (digestible organic matter, g/kg DM) were strongly positively correlated with free water intake (r = 0·751 and 0·595 respectively), though fermentation indices were not good single predictors of water intake. Further analysis revealed problems owing to collinearity within the predictors of water intake: DM intake, silage D value and milk yields being significantly correlated, as were pH and volatile fatty acids as a proportion of total fermentation acids. The ridge regression technique was used to reduce collinearity problems and produce stable equations. The best prediction equations for water intake involved a combination of both animal and analytical information: diet DM concentration, milk yield and silage pH. The use of fermentation information, whether from titration or high-performance liquid chromatography did not describe real variation in water intake beyond that described by silage pH. Free water intake was higher with higher diet DM concentrations, higher milk yields and higher silage pH.

Sixteen varied grass silages metabolizable energy (ME): 9 . 76 to 11 . 99 MJ/kg ethanol-corrected toluene dry matter (TDM); crude protein (CP: 149 to 211 g/kg TDM; lactic acid: 3 . 5 to 134 . 7 g/kg TDM; butyric acid 0 . 4 to 46 . 7 g/kg TDM) were offered ad libitum to early-lactation daily cows (12 per experiment) along with a fixed allocation of 7 kg/day of a standard concentrate. Four silages were offered in each of four incomplete change-over design experiments with three 21-day periods. This design meant that each cow tons allocated to receive three of the four silages evaluated in that experiment. ME intake ranged from 108 to 262 MJ/day (mean 177 (s.d. 30 . 2)). Similar variation was obtained with milk yields (mean 26 . 5 (s.d. 4 . 36) kg/day), fat content (mean 37 . 7 (s.d. 5 . 60)g/kg) an_d protein content (mean 29 . 0 (s.d. 2 . 36)g/kg). Urinary purine derivative/creatinine ratio (PD/C), an index of microbial protein measured in spot samples (two per day) averaged 2 . 92 (s.d. 0 . 757) mol/mol. Allantoin made up an almost constant molar proportion of PD excretion (mean 0 . 876 (s.d. 0 . 0377)), with a small but significant (P < 0 . 001) decline of 0 . 0132 (s.d. 0 . 003) per unit increase in PD/C. Maximal utilization of silage nitrogen occurred with silages having higher ME and lower CP concentrations. Urinary PD/C suggested that microbial protein yield varied in a way which would not be predicted in current schemes and that it was a major source of variation in milk protein yield under the conditions of the present experiment. Principal components regression confirmed independent effects of ME supply and MP supply (indexed by urinary PD/C) on milk protein yield. Further work should pursue the possibility of using the urinary PD/C technique to refine protein feeding at the farm level.

To investigate the effects of energy source and protein level of diets on milk protein content, 12 multiparous Holstein-Friesian cows were used in a 4 × 4 Latin square change-over experiment with 4-week periods. Four diets were offered, with ad libitum silage as proportionately 0·40 of the diet, and the remaining 0·60 as one of four concentrates, two based on barley and two on molassed sugar-beet pulp. Two protein levels were achieved by altering the amounts of digestible undegraded protein in the concentrates, with all diets formulated to supply equal quantities of rumen degradable protein. There was no effect of diet on dry-matter intakes. Both starch and high dietary protein levels significantly increased milk protein concentration (P < 0·05), but had no effects on milk fat and lactose concentrations. Mean milk yields were significantly higher (P < 0·05) with increased dietary protein. Dietary protein significantly affected the yields of milk protein (P < 0·01) and lactose (P < 0·05) but not that of fat. Urinary allantoin excretion was significantly greater with both high protein (P < 0·05) and starch-based diets (P < 0·05). No significant interaction effects were found. It is concluded that dietary effects were due largely to differences in supply of rumen degradable protein; increases in milk protein concentration were therefore brought about by increasing the protein supply to the animal.

The greatest error in formulating rations is due to the inaccuracy of prediction of silage dry matter intake (SDMI). Until recently, predictions have been based on die method of Lewis (1981) which predicts intake from traditional silage analysis :- dry matter (DM), crude protein (CP), digestible organic matter in the dry matter (DOMD) and ammonia N. Recently, the incorporation of new feed characterisation data, obtained from electrometric titration (ET), has unproved predictions (Offer et al., 199S). A 4 year study has yielded data to evaluate alternative methods for the prediction of SDMI using traditional, ET and HPLC data and spectral information obtained by near infra-red reflectance spectroscopy (NIRS) of fresh and dried samples.

Inaccurate prediction of silage intake is a major source of error in ration calculation. New routine methods such as electrometric titration (ET) and near infra-red reflectance spectroscopy (NIRS) have the potential to improve silage intake prediction (Offer et al., 1994). Development of new advisory models is slow because of the need to do large numbers of intake measurements. It is also essential that models are blind-tested (validated) on data not used for calibration. At SAC, sufficient animal intake data is now available for both calibration and validation although it is intended to continue to accumulate data for these purposes.

Inaccurate prediction of silage intake is a major source of error in ration calculation. Improvement in accuracy depends on the introduction of new routine methods to characterise better the silage fermentation. Electrometric titration has the potential to fill this role as it measures the main fermentation end-products in silage, residual sugar and gives absolute data on buffering characteristics (Offer et al., 1993).Results are presented for years 1 and 2 of a 3 year study of the use of electrometric titration of silage juice to improve the prediction of silage true DM (TDM) intake (SDMI). Fifty seven silages have been made using a variety of grass and grass-clover mixtures and a wide range of cutting dates (between 25 May and 29 September) and ensilage methods. The following additives were used : none, formic acid (2.5 l/tonne), formic acid (>5 l/tonne), Lactobacillus inoculant, molasses (16 l/tonne), molasses + inoculant. A range of wilting times (0 to 48 hr) was used and in some cases grass was left uncovered in a heap (up to 72 hr) before ensiling.

Inaccurate prediction of silage intake is often the greatest source of error in ration formulation but improvements have been hindered by an inadequate description of the silage fermentation in advisory practice. Until recently, only pH and ammonia N have been routinely measured as indicators of silage fermentation characteristics. A new method, based on research in Finland (Moisio and Heikonen, 1989) has been developed and is now in use by SAC. This involves automated titration of juice squeezed from the silage to pH 2 (with HC1) followed by stepwise titration (with NaOH) to pH 12. Concentrations of juice constituents are predicted from the buffering capacities measured over segments of the titration curve (pH 2 to pH 12). Calibrations have been obtained empirically by the addition of known increments of standards to a range of silage juices. Proportions of variance (R2) of measurements made by reference methods accounted for by predictions from titrations for a validation set of 93 silages were 0.90, 0.78, 0.92, 0.82 and 0.82 for lactic acid, acetic+butyric acids (VFA), soluble N, sugar and ammonia N respectively.

Seventy-two, 4-month-old, British Friesian steers were used to investigate the effects of feeding a supplement of fish meal on the voluntary intake and live-weight gain by young growing cattle given a well preserved ryegrass silage. The silage was offered either alone or mixed with 50, 100 or 150 g fish meal per kg silage dry matter (DM) and the diets were offered either ad libitum or intakes were restricted to 16, 19 or 22 g dietary DM per kg live weight (LW). Intakes were recorded daily, LW weekly and in vivo apparent digestibility over one 7-day period during the 132-day trial. For animals fed ad libitum, the absolute intake of dietary DM increased linearly with an increase in the level of fish-meal supplementation such that intake when the highest level of fish meal was given was significantly higher (P < 0.01) than when silage was given alone. However, DM intake per unit LW (approx. 24 g DM per kg LW) was not affected significantly (P > 0.05). Inclusion of fish meal in the diet did not affect the apparent digestibility of dietary DM, organic matter, acid-detergent or neutral-detergent fibre (NDF) although there was a trend for slightly higher (P > 0.05) gross energy apparent digestibility when fish meal was given. Increasing the level of feeding reduced NDF digestibility. The coefficients measured at the 22 g and ad libitum levels of intake were lower (P < 0.01 and P < 0.05 respectively) than that measured at the 16 g DM per kg LW level. Animals given silage alone to appetite achieved LW gains of 0.6 kg/day. LW gains increased linearly with increasing level of feeding (P < 0.001) and increasing level offish-meal supplementation (P < 0.001).

This paper reviews the fate of nitrogen in the feed of lactating dairy cows and considers the consequences of the various transactions for urinary N excretion. A simple computer model was established and used to investigate the major factors influencing urinary N. In setting the model up it was clear that a number of areas of uncertainty about nitrogen transactions in microbes or host tissues remain. The basal animal and area efficiencies for a herd producing 5000 litres of milk per cow per annum are 0.68 g milk N per g urinary N and 97.7 kg urinary N per hectare per annum. Feeding concentrates at 0.3 kg per litre of milk to obtain yields of 6700 litres per cow per annum increased these to 0.74 g and 191.0 kg respectively. Two major areas of uncertainty remain, the efficiency of capture of rumen degraded N by rumen microbes and the efficiency of utilisation of amino acids by host tissues. Simulations of the possible ranges in these variables illustrate the potential effects on nitrogen waste and the dietary crude protein required to support a given level of production. A major practical problem is that whilst it is very difficult to assess these variables under experimental conditions it is impossible to identify them at the farm level. This can lead to incorrect feeding decisions which merely exacerbate the problem. Given the simple assumptions of this model, the strategy that results in the lowest return of urinary N per forage hectare is an extensive system of production, particularly if this can be achieved using animals which have a high efficiency of transfer of amino acids into milk protein. Nevertheless, there is an urgent need to provide more consistent predictive relationships, and/or diagnostic tests for farm-use, in order to advance models and reduce urinary N excretion.