particles (Van Bruchem et al., 1990). Udén
et al. (1980) mordanted plant fibre in a
solution with chromium at a concentration
of 120–140 g kg^1 of fibre. In many subse-
quent studies, the chromium concentration
in mordanting solutions was reduced to
decrease the chromium content of
mordants and minimize changes in feed
density. Chromium concentrations of 10 g
kg^1 on hay and 23 g kg^1 on feed pellets
were considered sufficiently low not to
affect adversely the density and kinetics of
the feed in the digestive tract (Moore et al.,
1992). However, the disappearance from
nylon bags in the rumen of silage and hay
containing chromium at a concentration of
31 and 23 g kg^1 dry matter, respectively,
was reduced by 36.7 and 56.4%, respec-
tively (Beauchemin and Buchanan-Smith,
1989).
The digestibility of mordanted plant
material has a marked effect on the
recovery of marker in the faeces. The
lowest recovery is associated with samples
with the highest digestibility. Furthermore,
the particle size of Cr-NDF has a marked
effect on determined passage rates. A
particle size of 0.6–1.0 mm gave a fractional
passage rate of 2.0% h^1 , compared with a
passage rate of 4.1% h^1 for particles
<0.3 mm (Bruining and Bosch, 1992).
CHROMIUM-ETHYLENEDIAMINE TETRAACETIC
ACID (CR-EDTA).Cr-EDTA, which is readily
soluble in water, is used as a liquid phase
marker. Udén et al. (1980) prepared Cr-
EDTA as its lithium salt (the only salt that
could be crystallized) by dissolving
chromium(III)acetate monohydrate, EDTA
and lithium hydroxide monohydrate in
water on a steam bath. Cr-EDTA was
crystallized by cooling and the addition of
ethanol. The compound gave a yield of
about 90% and was stable to drying at
100°C. Cr-EDTA is usually infused con-
tinuously into the rumen (Remillard et al.,
1990). If radiolabelled^51 Cr-EDTA is used, it
is first diluted with carrier Cr-EDTA.
Like polyethylene glycol (PEG), Cr-
EDTA is slightly absorbable through the
rumen wall, but occupies a larger fluid
space in the rumen than PEG. It is affected
by fluctuations in osmotic pressure which
could lead to an overestimation of the
inflow or outflow of ruminal water. Low
concentrations of Cr-EDTA may also bind
to particulate matter in the rumen.
Chromium can be analysed by means
of a titrimetric procedure (Christian and
Coup, 1954), by spectrophotometry (Fenton
and Fenton, 1979) or by atomic absorp-
tion spectrophotometry (Arthur, 1970).
Unlabelled chromium can also be assayed
as^51 Cr after neutron activation (Udén et
al., 1980) or plasma emission spectroscopy
(Combs and Satter, 1992). Since^31 Si, the
main contaminant, has a half-life of only
2.6 h, it does not interfere with the
analysis. Advantages of this method are
that radioactive contamination of animals
is avoided and the dry faecal samples can
be counted without prior preparation.Cobalt
Cobalt-ethylenediamine tetraacetic acid
(Co-EDTA) is prepared either as the sodium
or lithium salt of the monovalent Co-EDTA
anion (Udén et al., 1980). The yield of
LiCo-EDTA was about 90%, which was
higher than the yield of the sodium salt.
Both salts are stable to drying at 100°C.
Cobalt is quantified by atomic absorption
spectrophotometry, by neutron activation
analysis or by plasma emission spectro-
scopy (Combs and Satter, 1992).Ruthenium
Ruthenium is used as a particulate
marker, usually in the form of ruthenium
phenanthroline (Ru-phe), which has no
adverse effects on the metabolic activity
of the rumen microbiota at the concentra-
tions used (Tan et al., 1971). It is prepared
by converting ruthenium chloride to
potassium pentachlorohydroxyruthenate
and refluxing the hydroxyruthenate with
1,10-phenanthroline and hypophosphite
to form Ru-phenanthroline chloride (Tan
et al., 1971).
Ru-phe is infused continuously into
the rumen, often as the isotope,^103 Ru-phe,
or administered as a single dose. The dose
can be prepared by dissolving the marker
in water, mixing it with milled grass in a260 J.P. Marais