SUMMARYShort-term associations between disease and milk yield were studied in 1594 lactations occurring between 1977 and 1982 in 732 Friesian, Ayrshire and Holstein crossbred cows. Statistically significant differences in milk yield between one week before and one week after clinical diagnosis were observed for ketosis (5·l kg/d), hypomagnesaemia (4·1 kg/d), mastitis occurring after peak yield (2·1 kg/d) and lameness (1·1 kg/d). Milk yield declined for as much as 2–4 weeks before diagnosis of disease, and total losses in milk yield associated with ketosis or hypomagnesaemia were estimated to be 60–70 kg. No significant differences were found for endometritis or for mastitis when it occurred before peak yield. Milk yield was suppressed for about 4 weeks after calvings with retained placenta, but there were no corresponding effects of dystocia or hypocalcaemia on milk yield. It is suggested that if automatic daily milk recording is available it may be possible to detect deviations from normal, and hence detect subclinical disease 2 or more weeks before its clinical appearance.

One thousand, four hundred and ninety-one lactations in 770 Friesian, Holstein and Ayrshire crossbred cows have been used to study the associations between lameness and fertility. Lameness was associated with a longer interval between calving and first service and a longer interval between calving and conception. The largest increases in these intervals, of 17 and 30 days, respectively, occurred in cows with either sole or white line lesions occurring between 36 and 70 days after calving. The conception rate during the 63 days before a diagnosis of lameness was made was lower (31 per cent) than at other times (40 per cent).

A study of 468 lactations in two herds of British Friesian cows sampled weekly between 1976 and 1979 was undertaken to investigate seasonal variations in milk yield and to compare four models of the lactation curve: (i) l°g e y( n ) -l°8e a + blog e n -en; (ii) y(n) = an b e~a l ; (iii) y(n) = ae~'"' -ae~q n ; and (iv) y(n) = a-p'n -ae~q' n .Model (i) was fitted by unweighted linear regression and the remainder by maximum likelihood. Average peak yield was 29 kg/day and 305-day yield was 5720 kg.Model (ii) fitted the data slightly better on average than models (iii) and (iv), and all were better than model (i). Compared with model (i), model (ii) reduced the average residual mean square proportionally by 0-10 in cows and 004 in heifers (first calvers). Model (iv) described the initial rise in milk yield up to week 5 better than models (i) or (ii) but reached a maximum value slightly early. Models (i), (ii) and (iv) slightly underestimated and (iii) slightly overestimated maximum milk yield, but model (ii) provided the best estimate of the position of maximum yield.Peak yields for heifers occurred on average at approximately 10 weeks post calving compared with 7 weeks for cows. Mean persistency of lactation, calculated as s= -(b+\)\og e c, did not vary from the second lactation onwards, whereas the average rate of decline p' in model (iv) increased with parity.When milk yields were adjusted for stage of lactation, significant variations were found that were related to the week of recording but this seasonal variation in milk production was smaller than that found nationally in the United Kingdom.

A study of 1514 lactations that occurred between 1977 and 1982 in 790 Friesian, Ayrshire and Holstein crossbred cows was undertaken. One thousand and eighty-four of these lactations were used to consider the association between the occurrence of clinical mastitis and changes in milk yield, and to illustrate a technique that enables the previous lactation history of a cow to be used for estimating changes associated with the disease.The lactation curves were described using the model y(n) = anbe−cn from which values for peak yield, persistency, rate of decline at week 25, and recorded and extrapolated 305-day yield were calculated. A significant correlation between successive lactations of the same animal was observed for each of these variables. Pairs of successive lactations free from clinical mastitis were then compared by regression analysis with lactations in which the preceding lactation was free from the disease whilst the second lactation of the pair was infected.Significant reductions in peak yield, recorded and extrapolated 305-day yield were found when mastitis occurred before the week of peak yield. A reduction in recorded 305-day yield was also observed when the disease occurred between peak and 10 weeks after peak. Mastitis later in the lactation had no significant effect.The mean reduction in recorded 305-day yield of 540 kg when mastitis occurred before peak for a cow with a mean yield of 4830 kg, was higher than previous estimates. No evidence of recovery was observed in subsequent lactations that were free of clinical infection, indicating that once a cow has contracted mastitis it may not achieve its full milk-yield potential in the next lactation.

No abstract