Farm Animal Metabolism and Nutrition

error in estimating this fraction introduces
a bias into the estimates of fractional rate
and lag time. Choice of an incubation time
which is too short may produce an over-
estimation of the indigestible residue,
an underestimation of the potentially
digestible fraction and an overestimation of
its degradation rate. Sometimes an incuba-
tion which is too long results in a concave
semi-logarithmic curve indicating than
more than one digestible pool exists.
Observations during lag time must be
removed from the data set before semi-
logarithmic regression, to prevent an
underestimation of the degradation rate. If
the discrete lag time cannot be determined
visually, an iterative approach may have to
be used to determine which data to use for
determining the rate after an initial lag
time (Mertens, 1993).
In most cases, NLIN regression
methods are preferred because they result
in the smallest residual sum of squared
deviations from the model. Model para-
meters are adjusted to the exponential
model with (Dhanoa, 1988) or without
(Ørskov and McDonald, 1979) lag time,
through an iterative procedure until the
change in the residual sums of square
meets a convergence criterion, assuming
equal error at each observation (Mertens,

1993). Advantages of this technique are

that all parameters are estimated simul-

taneously. The Marquardt method in the

NLIN procedure of SAS (1985) is used

most often.

The different methods of calculation of

degradation rates of DM, N and cell wall

components have been compared by some

authors. The same data and model may

give different parameter estimates when fit

by lnLIN compared with NLIN regression,

but both will provide acceptable estimates

of kinetic parameters when data are

collected during well-designed experiments,

i.e. a large number of incubation times

adequately spaced out. Optimal and

minimal incubation schedules have been

proposed by Mertens (1993) (Table 11.1).

Determination of effective degradability

The nylon bag technique offers the possi-

bility of measuring the extent and rate

of degradation of feed during specific

intervals. However, digestion rate relative

to rate of passage is a critical dynamic

property affecting digestibility: most non-

cell wall components in feeds are digested

rapidly with rates that are 3–10 times faster

than rates of passage; conversely, the rate

of cell wall digestion typically is of the

same magnitude as the rate of passage

In Sacco Methods 243

Table 11.1.Recommended incubation intervals (h) to obtain a precise estimate of sample degradation
kinetics for degradation models containing three parameters (Mertens, 1993).

Rapidly degradable feeds Slowly degradable feeds

Optimum Minimum Optimum Minimum

0 a 0 a

0000

2233

4466

8899

12 12 12 12

16 18

20 20 24 24

24 30

32 32 36 36

40 48

48 48 72 72

64 64 96 96

aDetermination of the water-soluble fraction.