A Study of Lactation as Affected by Heredity and Environment

1. A review of the existing knowledge regarding the inheritance of milk yield indicated the need for a more extensive investigation of the subject, particularly with reference to (i) the inheritance of the various components of the lactation yield, and (ii) the differentiation of hereditary from environmental variations. 2. In the present thesis an account is given of such an investigation, the requisite data being obtained from six prominent herds of tuberculin tested pedigree Ayrshire cattle. The Official milk records of all the recorded cows in these herds were used as a basis of study, (i.e. the investigation was not limited to published records and therefore to selected data). The available data comprised a total of over 2,800 lactation period milk records. The portion of the lactation curve studied was between the period 15 days to eight and a half months after calving and the milk production of each record in successive 30-day intervals over this period was computed. 3. The first step was to find a satisfactory mathematical expression to describe the shape of the lactation curve. After reviewing the previous work, the use of the exponential curve was decided upon. As a result of preliminary work, however, it was discovered that, although this curve gave a good fit. with some lactations, the fit was extremely poor for others. The parabolic exponential curve was next tried and it was found that with most cows the addition of the extra constant considerably improved the fit of the exponential curve. As a result of this work it became clear that, to describe the shape of individual curves satisfactorily, two constants were required: (i) the linear (exponential), which measures the average slope of the curve, and (ii) the parabolic, which describes how on an average the rate of this slope varies from period to period. With most lactations the former term was, however, found to have a preponderating influence in graduating the time-change of milk yield, and in view of this fact and of the very high variability exhibited by the parabolic term, the latter was not included in the subsequent work (except in the study of the effect of time of calving). 4. Because of the poor fit of the exponential curve in some lactations, the maximum yield as given by the fitted theoretical curve was very different from the observed maximum. It was, therefore, decided to use the latter instead of the theoretical value to measure the maximum yield of any lactation. 5. A special study was made of two environmental factors which are known to influence milk production and on which quantitative data were available, namely month of calving and age. 6. Two methods were adopted in studying the influence of month of calving, one of which was based on the lumped data for all lactations and the other confined to intra-cow comparisons. The results obtained by the two methods are discussed in the text and an explanation offered of the observed variations. Correction factors were worked out for persistency, maximum yield and total yield. 7. The influence of age was next investigated from data corrected for the effect of month of calving, two alternative methods again being employed. The influence of selection was studied and allowed for in working out the correction factors for persistency, maximum yield, and total yield. A study was also made of the correlation between the production of a cow at different ages. The bearing of these latter results on the question of predictability of the future performance of a cow, the value of each production record as a measure of a cow's genetic constitution, and the method of standardisation of records for the influence of age was indicated. 8. The predictability of the performance of a cow in a future lactation on the basis of her production in all the earlier lactations was studied by the method of multiple regression. It was found that the performance during the lactation immediately preceding that to be predicted gives nearly as good an estimate of the probable production as the yield for all the previous lactations