Dhanoa et al. (Reference Dhanoa, Lopez, Sanderson and France2009) described an elegant and novel approach to estimate effective rumen degradability (E) assuming zero-order kinetics, which would increase the capacity to process samples using the in situ nylon bag method. The degradation and passage kinetic model for estimating E was basically the same as the original Ørskov & McDonald (Reference Ørskov and McDonald1979) model. Although the described method is mathematically sound, the authors ignored two erroneous assumptions of the basic kinetic model. First, the model assumes that the ‘wash fraction’ is instantly degraded. Part of the immediately ‘degraded’ fraction can be physical loss of small particles from the bags. Although the effect of particle loss may be corrected mathematically, the assumption of infinite degradation rate for the truly soluble fraction is not correct. There is ample experimental evidence based on different in vitro and in vivo techniques (Chen et al. Reference Chen, Sniffen and Russel1987; Hristov & Broderick Reference Hristov and Broderick1996; Peltekova & Broderick Reference Peltekova and Broderick1996; Choi et al. Reference Choi, Ahvenjärvi, Vanhatalo, Toivonen and Huhtanen2002; Volden et al. Reference Volden, Mydland and Olaisen2002; Ahvenjärvi et al. Reference Ahvenjärvi, Vanhatalo, Huhtanen and Hristov2007; Reynal et al. Reference Reynal, Ipharraguerre, Lineiro, Brito, Broderick and Clark2007) demonstrating that a quantitatively significant proportion of soluble non-ammonia N escapes from the rumen in the liquid phase. Another false assumption is the first-order one-compartment passage model. Marker kinetic studies based on duodenal sampling have clearly demonstrated that passage kinetics of ruminal digesta are not well described by a single exponential model (for references see Huhtanen et al. Reference Huhtanen, Ahvenjärvi, Weisbjerg, Norgaard, Sejrsen, Hvelplund and Nielsen2006), not even for concentrate feeds. These two false model assumptions would exaggerate differences in ruminal protein degradability between feeds. The insignificant effects of increasing rumen undegraded protein (RUP) supply on milk production (Ipharraguerre & Clark Reference Ipharraguerre and Clark2005) support this statement. In a recent meta-analysis (Huhtanen & Hristov Reference Huhtanen and Hristov2009) based on 1800 treatment means, ruminal protein degradability only marginally improved predictions of milk protein yield or milk N efficiency when used as a third independent variable in regression models with energy and crude protein intake. The experimental evidence strongly suggests that RUP is overvalued in current protein feeding systems, at least partly because of the above mentioned two false assumptions of the Ørskov & McDonald (Reference Ørskov and McDonald1979) model. Unless more correct models are used to derive E from kinetic parameters, simplified in situ procedures are of little value in feed evaluation despite increased labour effectiveness.