European reference method for the in vivo determination of metabolisable energy with adult cockerels: reproducibility, effect of food intake and comparison with individual laboratory methods.

1. A common reference procedure adopted by several European laboratories to determine apparent metabolisable energy corrected to zero-nitrogen balance (AMEn) is described. 2. Reproducibility has been estimated by comparing results from 4 diets measured in 7 laboratories. The standard deviations between laboratories of dry matter, gross energy, Kjeldahl nitrogen and AMEn were respectively, on average, 11.29 g/kg, 0.242 MJ/kg, 1.56 g/kg and 0.380 MJ/kg dry matter; the corresponding coefficients of variation between laboratories were 1.27, 1.29, 4.39 and 2.92%. 3. Effect of food intake was tested by comparing AMEn from birds fed either 0.90 or 0.45 of ad libitum: the amount of food intake had a negligible effect on AMEn values. Endogenous energy losses corrected (EELn) or not (EEL) to zero nitrogen balance were estimated either by regression between excreted energy and ingested energy or in fasted and glucose-fed cockerels. EELn estimated by regression was, on average, 8.7 kJ/bird/d.; it was independent of diet composition. EEL and EELn determined in fasted or glucose-fed birds were higher than those determined by regression. 4. AMEn values measured by individual laboratory procedures were very close to those obtained by the reference method, except from laboratories using a tube feeding procedure, where deviations were probably a consequence of overestimated EEL.