Following a 14-day uniformity period when they were stall-fed to appetite on a grass silage-grass ration, 18 cows in mid lactation were allocated to three treatments on which they received during the following 28 days either 100%, 75% or 50% of their previous voluntary feed intakes.
Treatment effects were measured on milk yield and composition, live weight, liveweight change and fasted live-weight change. Twice during uniformity period, and on the 7th, 14th, 22nd and 26th days of the treatment period, samples of jugular blood were obtained from each cow immediately before and 3 h after the first meal of the day. These were analysed for percentage packed cell volume, blood acetone and β-hydroxybutyrate levels, plasma glucose, non-esterified fatty acid, and total fatty acid concentrations.
Alternative methods of collecting blood samples, involving cannulation of the vein and venipuncture, were compared. Obtaining samples through cannulae took nearly three times as long as by venipuncture, but the method of sampling did not significantly affect the concentrations of individual blood metabolites. Blood samples collected before as contrasted with after feeding had significantly higher concentrations of nonesterified fatty acids and glucose, and lower concentrations of acetone and β-hydroxybutyrate.
The feed requirements of the cows were assessed using data on live weights, fasted live-weight change and milk production by calculating the regression of estimated net attributable energy/kg LW0·75 on digestible energy intake/kg LW0·75, from data for the full 28-day period. The maintenance requirement was 124 kcal net energy/kg LW0·75, and the efficiency of conversion of digestible energy to milk and body-tissue loss or gain was 60%. These data were used with each animal's live weight, milk yield and liveweight change appropriate to each blood-sampling day to calculate individual feed requirements on every occasion. Ratios of actual feed consumed to estimated requirements were calculated subsequently to provide quantitative indices involving all cows and feeding levels. Regressions of these indices on individual blood metabolites were examined to assess the value of the latter as predictors of the relative nutritional status of the cows as defined in this way. Of the blood metabolites studied, only plasma NEFA were consistently significantly correlated with these indices, and relationships were specific to each feeding level. Thus although, at a common feeding level, plasma NEFA can be used to categorize lactating cows according to degree of nutritional stress, they will have limited values as predictors if applied indiscriminately to the wide range of feed availabilities and intakes which exist between herds at similar stages of lactation.
The results have been discussed relative to previously published work, mainly with sheep and beef cattle.