Published online by Cambridge University Press: 01 January 2008
Milk production data from Holstein × Zebu cows in small farms (2.4 cows per farm on average) in Maharashtra, India, followed by Bharatiya Agro Industries Foundation (BAIF), an Indian non-governmental organisation, were analysed to evaluate the impact of simplified milk-recording systems. The aim was to investigate, in developing tropical areas, less-costly protocols compared with the one currently implemented at BAIF, used as a reference. The latter can be considered an ‘AT2’ protocol with the recording made by specialised technicians at 2-week intervals. The simplified protocols were simulated from an initial data file by sampling test days according to each protocol. Bias and accuracy on the 305-day cow milk production and on the resulting reliability of the estimated breeding value of bulls were the criteria used in the comparison with the reference protocol. One type of simplified protocol considered an increase in the interval between two tests to at least 4 and up to 8 weeks. Another alternative studied corresponded to the situation where milk yield information measured by the farmer is collected by the artificial insemination technicians themselves when visiting a farm. This could be an option in the case of very small herd sizes (two or three cows). The results suggest that simplifying the current milk-recording protocol leads to a clear decrease in accuracy of estimating 305-day cow production but it has a limited effect on the reliability of bull proofs. No economic comparison was carried out, but the results strongly suggest that properly managed simplified milk-recording schemes could permit a substantial decrease of costs of milk recording per cow without damaging the efficiency of progeny testing in tropical areas with small herd size. Moreover, with the proposed simplified milk-recording protocols, up to three to four times more bulls could be tested with the same number of records.