Lucerne silage in high yielding dairy cows: effect of parity and analysis of reasons in mixed-parity group housing using random regression test‐day models

Abstract

A study with high-yielding dairy cows was re-analysed in order to test the suitability of lucerne silage separately for primi- and multiparous cows as an alternative to grass silage in maize-based total mixed rations (TMR). Lactation curves were fitted using random regression test‐day models for energy corrected milk (ECM) and dry matter intake (DMI) as well as for number and duration of feeder visits (NFV and DFV, respectively). Existing models for ECM and DMI were extended by animal-specific random effects, which were formulated in their dependency on days in milk. For NFV and DFV random regression models were applied for the very first time. The chosen approach of statistical analysis permitted comparisons of the lactation curves as well as of least square means for sub-periods to answer nutritional questions. Whilst primiparous cows had generally lower DMI and ECM as compared to multiparous cows, only in primiparous cows a negative effect of lucerne TMR on ECM was observed, especially in early lactation. Nutritional factors should be rejected because of very similar ECM between the various TMR in multiparous cows. Traits of feeding behaviour indicated that particle size could contribute to the decreased ECM. Even more impact on the lower ECM should be addressed to domination behaviour of multiparous cows. The resulting restlessness of primiparous cows caused a reduced intake per minute spent at the feeder. Further studies should focus on optimising the proportion and chopping length of lucerne in the diet and to improve flock management to maximise feed intake of primiparous cows. Generally, statistical analysis of lactation data became a very complex issue. It seems inevitable that nutritionists and statisticians team up to address this problem.

Keywords:

• dairy cows
• feeding behaviour
• lucerne silage
• mathematical model
• milk yield
• parity

Acknowledgements

We appreciate the referees’ comments, which led to a considerable number of improvements. The Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) is gratefully acknowledged for funding the first author.