leaving the rumen increases with time.
Gamma 2 two-compartment models are
recommended for small and medium
particles, and the gamma 3 two-
compartment models for large particles
(Quiroz et al., 1988). Algebraic or multi-
compartment models have also been used
in an attempt to improve the description of
marker excretion curves (Pond et al., 1984;
Dhanoa et al., 1985; Quiroz et al., 1988;
Lallés et al., 1991). The multicompartment
models assume instantaneous mixing
within and steady-state flow through
successive compartments. Comparison of
models showed that the two-compartment
models require good initial parameter
estimates but gave similar rumen turnover
rates to that of the multicompartment
models (Beauchemin and Buchanan-Smith,
1989). However, apparent turnover rates in
the caecum and proximal colon were twice
as long and transit times were 35% longer
using the multicompartment model com-
pared with the two-compartment model.
Reese et al. (1995) developed a multi-
compartment digesta flow model with
gamma-distributed retention times which
allows the estimation of total digesta com-
partment retention time, rather than
marker retention time. This model is used
mainly to assess changes in digesta flow
relative to changes in the environment, diet
and physiological condition of the animal.
Susmel et al. (1996) fitted results
obtained from a pulse dose of Cr-mordanted
forage to both a gamma age-dependent
model and a multicompartmental model to
predict faecal output. Similar output predic-
tions were obtained, but the multicompart-
mental model produced more accurate
results.
Markers such as chromium oxide,
which tend to behave independently of
both particulate and fluid phases, will give
erroneous estimates of flow rate when used
on their own. These limitations are over-
come by the commonly used double-marker
technique of Faichney (1992, 1993). It
involves the administration of a particulate
phase marker, such as Ru-phe, and a liquid
phase marker, such as Cr-EDTA, in the feed
or through a fistula either as a pulse dose or
at a constant rate until a steady-state has
been reached. Samples are taken at a
point(s) where flow is to be measured. This
procedure requires uniform distribution of
marker within phases, but does not require
the exclusive association of each marker
with each phase.
The rare earth elements are used exten-
sively for measuring the retention of solids
in the gastrointestinal tract, while Cr-EDTA
and Co-EDTA are used widely for measuring
the retention time of the liquid phase.
Rumen turnover rates based on ytterbium-
labelled NDF are faster than those based on
Cr-mordanted fibre as analysed by multi-
and two-compartment models, and probably
reflect a difference in particle size of the
marked feed (Beauchemin and Buchanan-
Smith, 1989). Rare earths tend to bind to
particles in proportion to their surface area,
giving mean retention times biased towards
that of smaller feed particles in the rumen.
Rare earth markers partly dissociated from
the feed particles also tend to leave the
rumen more rapidly than the feed particles.
Direct comparison of rare earth- and
chromium-mordanted-based passage rates is
therefore not advisable. However, dissocia-
tion of rare earth markers in the abomasum
may not invalidate kinetic estimates based
on faecal marker concentrations. Rare earth
elements may reduce the digestibility of
feed particles. Labelling silage with
ytterbium at a concentration of 35 g kg^1
DM reduced its rate of disappearance from
the rumen by 26% (Beauchemin and
Buchanan-Smith, 1989), but concentrations
<6 g kg^1 DM had no effect on digestion
(Siddons et al., 1985).
The excretion curve of radionuclide-
labelled microspheres in sheep fitted to a
two-compartment model gave a mean
retention time of 55 h compared with 51.2 h
for the liquid phase marker, Co-EDTA,
61.8 h for Cr 2 O 3 , and 61.9 and 55.1 h for
the particulate markers, ruthenium and
ytterbium, respectively (Young et al., 1991).
The mean retention time in the rumen,
estimated using bacterial spores as markers,
which may be more indicative of the
transport of rumen digesta than inorganic
markers, was found to be 5.5 h longer thanUse of Markers 265