Published online by Cambridge University Press: 27 February 2018
In nutrition, as in any other branch of quantitative applied science, two essential criteria must be met before research findings can be transferred successfully into practice. (1) The fundamental concepts must be essentially sound and (2) the predictions emerging from that conceptually sound system must, in a practical situation, constitute a useful increase in accuracy. The new method proposed by the Agricultural Research Council (ARC, 1980) for expressing protein requirements and the value of feeds for ruminants is undoubtedly an enormous improvement in a conceptual sense on ARC (1965). Indeed, the division of protein inputs into rumen degradable nitrogen (RDN) and undegradable dietary nitrogen (UDN) cannot be faulted since it is the best possible simple description of what happens. Even the decision to retain the concept of crude protein (N x 6.25) and refer therefore to RDP and UDP is harmless. The new ARC protein system therefore fulfils the first criterion for successful practical application. The second criterion can only be fulfilled if ARC (1980) accurately describes both the requirements of animals and the nutritive value of feeds in terms of RDP and UDP. In the case of dairy cows, this is open to considerable doubt.