Automatic milking system in Mediterranean buffaloes: experimental results and considerations

The scarce implementation of automatic milking systems (AMS) in buffalo farming despite its primarily benefit, i.e. relief from the milking routine, is mainly due, beside to the initial investment, to concerns about adaptability of buffalo cows to robotic milking and a potential detrimental effect on milk quality. This study aimed to shed light on the operational and functional aspects of AMS in buffaloes and their relationships with milk yield and quality. The research was carried out over 3 years on a dairy buffalo farm equipped with both a traditional tandem parlour and four AMS, each serving about 50 cows. Efficiency of the AMS was evaluated by taking into consideration three issues: (1) yield and quality of milked milk; (2) system capacity, i.e. number of buffaloes that can be milked by AMS per year; (3) number of milking-failure visits, i.e. robot failures to attach milking cluster. The assessment of milk traits was performed by comparing (ANOVA) over the whole lactation milk yield and quality of buffaloes milked with the traditional tandem (TT, 45 cows) or the AMS (45 cows). The cows milked by the AMS had higher number of milkings/days and higher milk yield (p<.05) than those milked with TT system, whereas no differences were found for milk composition. The analysis of system capacity was carried out over two years by comparing (ANOVA) the performances of buffaloes milked in AMS serving 50 or 70 cows. Results highlighted that the number of buffaloes per robot may be increased up to 70 without impairing milking performance (p<.05), so making it possible to recoup the initial investment earlier. Factors causing milking-failure visits were studied over one year by installing, for each AMS, one camera on the robot arm applying teat cups and another one on the pre-selection gate. The video-recordings were downloaded and analysed remotely every week. The number/ year of milking failure visits were 4053 ± 550 (mean ± SD), corresponding to about 10% of AMS working time. About 27% of the failure visits was due to the cow’ behaviour, while 72% was related to malfunction of the milking robot (e.g. unexpected size, and/or orientation of one or more teats; obstructed view of the teats). It follows that the operational and functional aspects of AMS in buffaloes can be better adapted to the buffalo’s characteristics. Overall results indicate that AMS may be a suitable alternative to conventional milking, offering new management prospects for buffalo farming.