water intake and the random nature of
caecal evacuation. Sibbald (1976) origi-
nally proposed an initial starvation period
of 24 h followed by a collection window
of 24 h (24 h + 24 h assay), but latterly rec-
ommended 24 h + 48 h for regular use. In
our laboratory, 48 h + 48 h is used rou-
tinely, with 48 h + 72 h for some raw mate-
rials (high-protein or particularly bulky
ingredients). The longer periods are clearly
more stressful, and factors such as bird
size and the administration of glucose
come into consideration (McNab and Blair,
1988).
It has been demonstrated unequivocally
that 12 h starvation before feeding is
insufficient to clear the digestive tract of
the residues of the previous diet. Although
in some cases extension beyond 24 h has
been argued to exert only a small effect on
TME values, a direct investigation in our
laboratory has shown measurable differ-
ences between the energy content of the
residues remaining after 24 (20.25 kJ) and
48 h (2.06 kJ) starvation (McNab and Blair,
1988). This finding and the logic of
equalizing the pre- and post-feeding starva-
tion periods encourage the application of a
48 h + 48 h assay and the adjustment of
other factors to cope with any increased
stress. In general, we find clearance rates
are quite variable between feedstuffs and
the amounts fed and it seems sensible
that a constant diet comprising highly
digestible components should be used for
the maintenance of the birds. To some
extent, however, a correction is made for
any carry-over of energy from the previous
diet in calculating the TME value of a feed-
stuff, because it seems reasonable to expect
that a comparable error will occur in both
the fed and negative control birds (those
used to derive EEL).
The time required to ensure complete
residue clearance from the digestive tract,
especially when single ingredients are fed
in type 3 assays, is a complex and largely
unresolved issue. Our experience with 50 g
intakes leads us to believe that 48 h is
insufficient for some ingredients. Lower
intakes obviously alleviate the problem,
but probably at the cost of both reduced
accuracy and increased influence of
endogenous factors. It is our view that
complete residue clearance is harder to
achieve with high-protein, and especially
finely divided, animal products; materials
of low density, which are packed very
tightly in the crop, can also cause
problems, where wetting of the feedstuff in
the crop may be a complicating factor.
Palatability may also be involved
because, for example, when blood meal-fed
birds were given water, distaste appeared
to be experienced and regurgitation
occurred. The sudden introduction of some
ingredients may induce gut stasis; attempts
to evaluate coffee residues by tube feeding
had to be abandoned because the material
did not pass from the crop. Less severe but
similar responses have been experienced
with quinoa. At the present time, it is only
possible to advise caution, particularly
with unfamiliar ingredients, to look out for
the excretion after the end of the balance
period and to be wary about accepting
unexpected results without further careful
scrutiny. While routinely extending the
balance period to 72 h could provide an
empirical solution to at least some of these
problems, it would be at the cost of
increased stress on the birds. Longer
periods where no food is given also result
in higher endogenous:exogenous energy
ratios, and the relative importance of the
error introduced by correcting for EEL
assumes greater importance.
The significance of these factors in
type 2 assays is difficult to judge. With
higher intakes, such as 70–80 g, collection
periods shorter than 48 h may be in-
sufficient for certain feedstuffs. On the
other hand, the use of complete diets and
allowing the birds access to them for 1 h,
may reduce any problems in comparison
with tube feeding single ingredients. It
seems reasonable to presume that, because
the crop will be unlikely to become so
tightly packed with dry food under the
conditions prevailing, water intake and the
passage rate of food residues might be more
normal.
The impact of water consumption in
these assays is another neglected research312 J.M. McNab