the amount of available starch. The differ-
ences in proteolytic activity of solid-adher-
ent microorganisms between bags and the
surrounding digesta are slight (Michalet-
Doreau and Nozière, 1998). In the bags,
the proteolytic activity varies with the
incubated feeds, but differences between
feeds are observed only for longer incuba-
tion times, and would not limit the use of
the in situmethod to characterize feed N
degradability.
Microbial colonization of bag samples:
output of microorganisms
The washing of bags after rumen incuba-
tion has two main objectives: to stop
microbial activity and to free feed
residues from rumen microorganisms, with-
out increasing the loss of feed particles
through bag pores. Several mechanical
methods have been proposed for rinsing
post-incubation bags: hand rinsing of bags
until the water is clear, about 90 s per bag,
or mechanical rinsing of bags in a washing
machine for a variable duration from 2 to
15 min (Michalet-Doreau and Ould-Bah,
1992). Cherney et al.(1990) compared the
two rinsing methods. Machine rinsing
twice for 2 min or hand rinsing of bags
gave similar results, but higher DM dis-
appearance and higher standard error for
DM disappearance were observed when
bags were machine rinsed twice for 5 min
than when bags were either machine rinsed
twice for 2 min or hand rinsed.
Despite washing, the bag residue
can still contain significant quantities of
microbial matter after incubation. Different
markers of microbial population (di-
aminopimelic acid (DAPA),^15 N or^35 S) or
feed (^15 N) can be employed for measuring
the microbial contamination of incubated
feeds (see Chapter 12). The marker tech-
nique nevertheless is difficult and the
results are probably affected by the type of
marker used. Moreover, the use of a marker
for the microbial population requires a ref-
erence bacterial sample in order to calcu-
late the percentage of bacterial DM or N
from the concentration of marker in the
residue. This reference bacterial sample is
isolated either from the liquid or the solid
phase of rumen content, or directly from
nylon bag residues. The chemical composi-
tion and the marker concentration of bacte-
ria vary according to the microbial
population. Marker concentration for the
fluid-phase microbial population is higher
than that for solid-phase microbial popula-
tion (Merry and McAllan, 1983; Craig et
al., 1987). However, Olubobokun and Craig
(1990) and Beckers et al.(1995) found little
or no difference in N degradability between
values corrected for fluid or adherent
microbes.
The bacterial DM in the bag, as a
percentage of residual DM, increases
rapidly in the first hours of incubation, and
either remains steady (Beckers et al., 1995)
or declines with ruminal incubation time
(Wanderley et al., 1993). The time neces-
sary for the appearance of the peak of cont-
amination and the maximum value of this
peak, expressed as a percentage of micro-
bial DM in bag residual DM, varies consid-
erably between feeds. Several hypotheses
have been proposed to explain the micro-
bial contamination variations between
feeds. For example, the cell wall content of
bag residues: however, it was not possible
to find a relationship between the fibre
fraction of bag residues and their coloniza-
tion rate by the microorganisms. It is also
possible that the differences in particle size
between feeds could explain the differ-
ences in microbial contamination. The
microbial colonization of feed particles in
the rumen is related negatively to their size
(Gerson et al., 1988). However W.Z. Yang
and C. Poncet (France, 1991, personal com-
munication) reported no relationship
between microbial colonization of particles
and their size inside the bags. Furthermore
mastication, which increases the bacterial
colonization of particles in the rumen, did
not affect microbial colonization of parti-
cles inside the bags (Olubobokun et al.,
1990).
The bacterial contamination of bags
leads to considerable and variable under-
estimation of the N degradability of feeds,
and this underestimation is greater when
the feed N content is low (Fig. 11.2). In a
study conducted on 51 forages, theIn Sacco Methods 239