2.That associative effects (e.g. the effects
of soluble sugars on fibre digestion) can be
ignored.
3.That extraction with either aqueous
ethanol or NDF does not remove sub-
stances inhibitory to in vitrodigestion.
Extraction with aqueous ethanol at
room temperature is, both chemically and
physically, a mild process and would not
seem likely to affect the digestibility of the
EIR fraction provided that the whole forage
does not contain ethanol-soluble inhibitory
compounds (see below). The NDF fraction
is prepared by extraction of the whole
forage with neutral detergent in an auto-
clave at 105°C. Small increases of the order
of 5% in fibre digestibility may result from
this treatment with temperate forages
(Schofield and Pell, 1995a; Doane et al.,
1997b). Comparative studies on NDF diges-
tion in the whole forage versus in the
isolated NDF suggest that both rates and
extents of digestion are similar (Doane et
al., 1997b). Associative effects in this in
vitro system appear to be small. Good
agreement was seen between the digestion
curves for whole lucerne and bromegrass
and for the sum of the separately digested
water-soluble and water-insoluble fractions
(Stefanon et al., 1996; Cone et al., 1997).
Errors in curve subtraction resulting from
associative effects will cause an under-
estimationof the A fraction and an over-
estimationof the B2 fraction. Assumption
3 above is certainly invalid for many
tropical legumes that may contain tannins
and saponins (see Plant Secondary
Compounds, p. 227).
The pools studied in the curve subtrac-
tion approach described above differ from
those derived by multi-pool analysis of a
whole forage gas profile. Within the limita-
tions of curve subtraction, these pools
represent actual chemical entities and each
pool may be a complex mixture and require
a multi-pool kinetic analysis. The object of
curve subtraction is to obtain a digestion
curve for a given, chemically defined,
forage fraction that cannot itself be isolated
and fermented readily. This digestion curve
may then be analysed using any of the
mathematical models already discussed. In
contrast, the mathematical pools obtained224 P. Schofield
Fig. 10.8.A diagrammatic view of curve subtraction. Three digestion curves are measured experimentally
for a given forage (whole forage, EIR and NDF). Hypothetical curves for the A and B1 fractions are then
obtained by subtraction.