A Multivariate Approach to Modeling Shapes of Individual Lactation Curves in Cattle

Milk test-day records of 5728 lactations of Italian
Simmental cows were analyzed with multivariate factor
analysis in order to extract 2 common factors, whose
scores were used as quantitative measures of 2 main
features of lactation curve shape—i.e., the increasing
rate of yield in the first part of lactation and the rate
of decline of milk yield after the lactation peak. The 2
indices, objectively derived from the correlation matrix
of original test-day records, showed a high discriminant
power in separating lactation curves with different
shapes. The weak correlation between the 2 factors
(0.11), together with the high correlation of factors and
the total 305-d yield (about 0.70), suggests that an increase
in lactation yield could be achieved by acting
only on one of the 2 factors related to lactation-curve
shape, with the other kept constant at a medium or low
value. The suitability of the 2 factors as descriptors of
lactation patterns has been confirmed by the relationships
found between factor scores and the main environmental
effects known to affect the shape of the lactation
curve, such as parity and season of calving.