of the feed. Attempts to test or verify the
validity of equations obviously founder if
consistent values cannot be produced.
Extension of this philosophy to dietary
ingredients also requires the establishment
of appropriate prediction equations which
relate ME to chemical composition or,
perhaps, to some other quality control para-
meter (e.g. in the case of wheat, density).
Here, progress is greatly facilitated if data
from different laboratories can be combined,
and this results in variations in techniques
being highlighted, especially if they are seen
to lead to different biases. It might have
been hoped that the introduction of an
effective rapid bioassay based on robust
scientific principles would have resulted in
its standardization and universal adoption.
There appear to have been two main reasons
why this has not happened; firstly, rapid
assays almost invariably require birds to be
starved, and this has proved controversial;
secondly, it is clear that some laboratories
experience problems in adopting published
techniques, and this has resulted in the
introduction of a number of major and
minor modifications. Consequently, there
are probably more different methods being
used to derive ME now than at any time in
the past, and the prospect of establishing a
single standardized assay is probably as
remote as it has ever been.
Definition of ME
Although ME is frequently believed to be a
property of a diet, it is actually a charac-
teristic of the animal to which the diet is
fed. ME measurements relate to complete
diets, and values for dietary components or
ingredients must, in most cases, be obtained
by comparing data from two or more appro-
priate diets (so-called substitution or
replacement methods). In such derivations,
the assumption that ME values amongst
feedstuffs are additive is essential, and very
little progress can be made if this is not
upheld. Energy is, of course, a useful cur-
rency for describing mass conversion of the
components of the food in the bird. There is
another set of problems, similar to those
discussed here for ME, in determining the
‘metabolizability’ of any nutrient; lipid,
protein or carbohydrate. For many purposes,
and especially for prediction, it would be
invaluable if both ME values and digesti-
bility coefficients for the main chemical
components were measured concurrently,
but this is only done on rare occasions.
The terminology used in the topic of
ME is relatively free from dispute and
ambiguity. The widely used convention,
mainly in agreement with Sibbald (1982,
1986) will be followed here. The term ME
is used in a general sense rather than to
mean bioavailable energy (Sibbald, 1982),
and the expression endogenous energy loss
(EEL) is usually defined, not as a biological
entity, but as an empirical quantity, i.e. the
energy loss from a starved bird. This is
convenient and need not be confusing. The
almost invariable convention of ignoring
gaseous losses resulting from fermentation
is also followed.
Some years ago, Pesti and Edwards
(1983) proposed that ME nomenclature
should be changed quite radically to reflect
the methods which had been used in the
evaluation experiments. The approach must
have been considered to be unhelpful
and unnecessary because it has not been
adopted; however, their proposition that
more care should be taken in relating
experimental observations to well-defined
biological elements should be mandatory. In
this context, there is probably not enough
discipline exercised by editors of scientific
journals. ME values are not measured or
observed but are the results of derivations
from a number of measurements. Because
too little basic information is reported, it is
often impossible to make critical com-
parisons between different experiments. By
tabulating results in more detail, greater use
could be made of existing data and allow
outcomes of experiments in different
laboratories to be combined.Classical Methods for Deriving ME
Two approaches traditionally have been
followed to generate ME values and these308 J.M. McNab