Historically, cow selection criteria were developed for conventional milking systems that have regular milking intervals (MI). However, in automatic milking systems (AMS), there is variability in MI within and between cows. These sources of variability provide an opportunity to identify cows with high daily milk yield (DY) and long MI. An extended MI (longer than 16 h in pasture-based systems) has a negative effect on DY. Cows that tolerate extended MI and maintain high DY can be considered more efficient than cows with low DY and long MI, or with high DY but short MI, thereby improving robotic system use. Knowledge of the behavior and parameters of lactation curves of cows in AMS could help farmers to identify cows with a specific lactational phenotype. The objective of this study was to identify individual cows with high DY and long MI within herds, which could reflect increased tolerance to milk accumulation under AMS. A database containing records for 773,483 milking events for one year (July 2016–June 2017) from 4 pasture-based AMS farms was used. Lactation curves within each herd were fitted using several mixed models including fixed effects for the parameters of the lactation curve and random cow effects. Predicted curves of average DY according to parity (multiparous and primiparous) were obtained. The best linear unbiased prediction of the random cow effect allowed us to categorize lactations as having either high or low milk production. The median MI of each lactation was then used to categorize cows as having either short or long MI. Daily yield at the peak of lactation, days to peak and 305-d cumulative milk production were used to compare the effect of DY and MI categories, as well as the DY × MI interaction. Milk production by multiparous and primiparous cows with high DY and long MI was between 35 and 45% higher than that of the low DY and short MI. From all lactations analyzed, the incidence of animals with high DY and long MI across farms was 7.5%. We have identified and quantified a new, AMS-specific, phenotype (the combination of a relatively higher DY with relatively longer MI) with potential to increase use of AMS units. Identifying more efficient animals should help generate new approaches for differential management and for selecting cows in AMS.

历史上，为具有规则挤奶间隔（MI）的常规挤奶系统制定了母牛选择标准。但是，在自动挤奶系统（AMS）中，奶牛内部和之间的MI均存在差异。这些变异性来源提供了机会来识别每日产奶量高（DY）高且MI较长的母牛。MI延长（在基于牧场的系统中长于16小时）对DY有负面影响。与低DY和长MI或高DY但MI短的母牛相比，可以耐受MI延长并保持高DY的母牛效率更高，从而改善了机器人系统的使用。了解AMS中奶牛泌乳曲线的行为和参数可以帮助农民识别具有特定泌乳表型的奶牛。这项研究的目的是确定牛群中具有高DY和长MI的个体母牛，这可能反映出AMS对牛奶积累的耐受性增加。使用了一个数据库，该数据库包含来自4个基于牧场的AMS农场一年（2016年7月至2017年6月）的773,483次挤奶事件的记录。使用几种混合模型拟合每个猪群内的泌乳曲线，包括泌乳曲线参数的固定效应和随机奶牛效应。获得了根据奇偶校验（多胎和初生）的平均DY预测曲线。随机奶牛效应的最佳线性无偏预测使我们可以将泌乳量分为高或低产奶量。然后，将每次泌乳的中位MI分为具有短MI或长MI的母牛。泌乳高峰期的日产量 高峰期的天数和305天的累积产奶量用于比较DY和MI类别以及DY×MI相互作用的影响。高DY和MI长的多头和初产奶牛的牛奶产量比低DY和MI短的母牛高35％至45％。从所有泌乳期分析，整个农场中具有高DY和长MI的动物发生率为7.5％。我们已经确定并量化了一种新的，特定于AMS的表型（相对较高的DY与相对较长的MI的组合），并有可能增加AMS单位的使用。确定效率更高的动物应有助于产生新的差异管理方法和AMS中的母牛选择方法。高DY和MI长的多头和初产奶牛的牛奶产量比低DY和MI短的母牛高35％至45％。从所有泌乳期分析，整个农场中具有高DY和长MI的动物发生率为7.5％。我们已经确定并量化了一种新的，特定于AMS的表型（相对较高的DY与相对较长的MI的组合），并有可能增加AMS单位的使用。确定效率更高的动物应有助于产生新的差异管理方法和AMS中选择母牛的方法。高DY和MI长的多头和初产奶牛的牛奶产量比低DY和MI短的母牛高35％至45％。从所有泌乳期分析，整个农场中具有高DY和长MI的动物发生率为7.5％。我们已经确定并量化了一种新的，特定于AMS的表型（相对较高的DY与相对较长的MI的组合），并有可能增加AMS单位的使用。确定效率更高的动物应有助于产生新的差异管理方法和AMS中的母牛选择方法。表型（相对较高的DY和相对较长的MI的组合）具有增加AMS单元使用率的潜力。确定效率更高的动物应有助于产生新的差异管理方法和AMS中的母牛选择方法。表型（相对较高的DY和相对较长的MI的组合）具有增加AMS单元使用率的潜力。确定效率更高的动物应有助于产生新的差异管理方法和AMS中的母牛选择方法。