Abstract
The objective of the paper was to see if the substitution of feed class regressions of estimated net energy (ENE) on TDN for a general regression of the combined classes would significantly affect the amount of error involved in converting TDN to ENE values. Using Morrison''s data for ENE and TDN values, 21st ed., it was found that all the feed-classes had regression coefficients that differed significantly from their combined regression coefficient. The feed classes, silage excepted, although they had similar slopes, apparently differed from the combined regression line because the feed classes operated at different levels of energy. Data in Morrison''s 22nd ed., were treated in the same manner, and results agreed satisfactorily with the previous edition except for silage and green roughage classes. Due to the many changes in TDN and ENE values between editions, the regression lines of these 2 classes differed significantly from those of the earlier edition and green roughage class now became no longer significantly different from the combined regression. Regression equations of the feed classes developed from data in each edition are given with their errors of the estimate. In converting TDN to ENE values using Morrison''s data, the appropriate feed class regression appears to be more accurate than a general regression.

This publication has 1 reference indexed in Scilit: