A . Tolera, F. Sundstøl Livestock Production Science 68 2001 13 –23
19 Table 5
multiple regressions further increased the coefficient
Pearson correlation coefficients of the relationship between esti-
of determination to 95, and the addition of L in the
mated in sacco DM degradability characteristics of the diet and
multiple regression further increased the coefficient
feed intake, digestibility and body-weight change in sheep
of determination to 96.
Dry matter Dry matter
Digestible Growth
intake digestibility
DM intake rate
DMD 0.89
0.97 0.97
0.98
24 h
4. Discussion
DMD 0.85
0.95 0.94
0.95
48 h
DMD 0.64
0.74 0.73
0.78
72 h NS
NS NS
The higher washing loss in earlier harvested maize
DMD 0.67
0.47 0.55
0.41
96 h
A 0.94
0.97 0.98
0.96
stover and in the desmodium hay indicates the
B 20.75
20.88 20.85
20.88
presence of more soluble components in these feeds.
NS NS
A 1 B
20.41 20.58
20.53 20.59
The same explanation holds true for the increased
c 0.77
0.92 0.89
0.93
NS NS
estimated washing loss at earlier stages of maturity
L 0.45
0.59 0.55
0.59
of the stover and with increasing level of desmodium
ED 0.87
0.97 0.95
0.97
hay supplementation of the diets. The increased
Significance level: NS 5P.0.05; 5P,0.05; 5P,0.01;
washing loss observed with increasing level of
5P,0.001; For other abbreviations, see Table 2.
supplementation was due to the higher washing loss nation to 96. However, the use of A, B, c or the
in desmodium hay and the additive effect of the addition of L in a multiple regression could not
stover and desmodium hay in DM degradability. improve the precision of prediction any further.
Desmodium hay supplementation also improved the The DMD
and ED were the best predictors of estimated DM degradation rate of the diets due to the
24h
growth rate followed by the washing loss, DMD ,
significantly higher degradation rate of the de-
48h
degradation rate and the insoluble, but potentially smodium hay than the stovers. Moreover, forage
degradable, fraction Table 9. The DMD had the
legume supplementation enhances the degradation of
96h
lowest precision of predicting growth rate followed the basal roughage by providing readily fermentable
by the
lag phase,
potential degradability
and substrates and thereby creating a more favourable
DMD . The use of A and B in multiple regressions
environment for rumen micro-organisms. According-
72h
improved the coefficient of determination by three ly, Ndlovu and Buchanan-Smith 1985 showed that
percentage units over the use of the A value alone, to alfalfa supplementation to poor quality roughages
94, whereas the use of A and c or A, B and c in increased rates of disappearance of DM and cell wall
Table 6
0.75
Prediction of dry matter intake by sheep y of the total diet g kg W day from DM degradability of the diet
2
Factors Equation
R RMSE
P DMD
y 5 213.611.7DMD
0.78 4.1
,0.001
24 h 24h
DMD y
5 252.512.01DMD 0.72
4.9 ,0.001
48 h 48 h
DMD y
5 2177.813.8DMD 0.35
7.1 ,0.05
72 h 72h
DMD y
5 2270.314.99DMD 0.39
6.9 ,0.05
96 h 96h
A y
59.512.98A 0.87
3.3 ,0.001
B y
5130.621.3B 0.56
6.1 ,0.01
c y
533.91854.8c 0.60
5.9 ,0.01
L y
544.719.6L 0.20
8.2 50.14
A 1 B
y 5137.821.1A 1 B
0.17 8.4
50.19 ED
y 5 234.612.22ED
0.76 4.5
,0.001 A B
y 5 20.3813.17A10.12B
0.88 3.4
,0.001 A c
y 56.713.55A2225.02c
0.88 3.3
,0.001 A B c
y 5 98.0 1 3.9A 2 1.23B 2 1117.5c
0.91 3.2
,0.001 A B c L
y 5 2 218.2 1 5.8A 1 2.6B 2 104.8c 1 22.14L
0.91 3.4
,0.001 RMSE
5Root mean square error. For other abbreviations, see Table 2.
20 A
. Tolera, F. Sundstøl Livestock Production Science 68 2001 13 –23 Table 7
0.75
Prediction of digestible dry matter intake by sheep y of the total diet g kg W day from DM degradability of the diet
2
Factors Equation
R RMSE
P DMD
y 5 284.712.7DMD
0.93 3.4
,0.001
24 h 24h
DMD y
5 2149.913.3DMD 0.88
4.8 ,0.001
48 h 48 h
DMD y
5 2367.316.4DMD 0.49
9.2 ,0.01
72 h 72h
DMD y
5 2366.916.04DMD 0.23
11.4 50.066
96 h 96h
A y
5 2 44.8 1 4.6A 0.95
2.9 ,0.001
B y
5 151.1 2 2.1B 0.72
7.2 ,0.001
c y
5 2 11.2 1 1447.97c 0.79
6.3 ,0.001
L y
5 5.7 1 17.1L 0.30
11.4 50.066
A 1 B
y 5 182.0 2 2.03A 1 B
0.28 11.5
50.077 ED
y 5 298.313.1ED
0.90 3.6
,0.001 A B
y 5 2 17.02 1 4.1A 2 0.33B
0.96 2.9
,0.001 A c
y 5 2 42.05 1 4.03A 1 221.12c
0.96 2.9
,0.001 A B c
y 5 2 2.1 1 4.2A 2 0.54B 2 169.3c
0.96 3.0
,0.001 A B c L
y 5 2 168.2 1 5.2A 1 1.5B 1 362.7c 1 11.6L
0.91 3.2
,0.001 RMSE
5Root mean square error. For other abbreviations, see Table 2. Table 8
Prediction of dry matter digestibility by sheep y of the total diet from DM degradability of the diet
2
Factors Equation
R RMSE
P DMD
y 5 291.213.4DMD
0.94 3.7
,0.001
24 h 24h
DMD y
5 2173.514.0DMD 0.91
5.0 ,0.001
48 h 48 h
DMD y
5 2437.017.8DMD 0.50
11.1 ,0.01
72 h 72h
DMD y
5 2365.516.3DMD 0.14
14.5 50.123
96 h 96h
A y
5 2 40.7 1 5.5A 0.93
4.3 ,0.001
B y
5 201.2 2 2.6B 0.77
7.8 ,0.001
c y
5 2 2.85 1 1821.2c 0.84
6.5 ,0.001
L y
5 16.7 1 22.5L 0.35
13.3 50.043
A 1 B
y 5 252.7 2 2.7A 1 B
0.34 13.4
50.046 ED
y 5 2138.614.5ED
0.94 3.9
,0.001 A B
y 5 23.7 1 4.29A 2 0.8B
0.95 3.7
,0.001 A c
y 5 2 33.1 1 3.95A 1 619.3c
0.96 3.6
,0.001 A B c
y 5 2 16.4 1 4.0A 2 0.2B 1 456.3c
0.96 3.9
,0.001 A B c L
y 5 2 182.2 1 5.0A 1 1.8B 1 987.1c 1 11.6L
0.96 4.1
,0.001 RMSE
5Root mean square error, For other abbreviations, see Table 2.
from nylon bags. Similarly, Kabatange and Shayo these parameters. The poor correlation of potential
1991 reported that supplementation of Leucaena degradability with DMI is consistent with the find-
leucocephala hay to maize stover resulted in in- ings of Carro et al. 1991. The A value, ED,
creased DM disappearance from nylon bags at DMD
, DMD and the c value were the best
24h 48h
different incubation times and in improved degra- predictors of DMI, DMD, DDMI and growth rate in
dation characteristics. a simple linear regression equation. In contrast, the
The DMD , DMD
, A, ED and c values lag phase, potential degradability, DMD
and
24h 48h
96h
showed strong positive correlation whereas the B DMD
were relatively poor predictors of feed
72h
value showed a negative correlation with DMI, intake, digestibility and growth rate.
DMD, DDMI and growth rate. On the other hand, The A value represents the readily soluble materi-
the DMD , potential degradability and the lag
als that could affect cell-wall degradation rates
96h
phase generally showed very low correlation with through their effects on the rumen’s microbial popu-
A . Tolera, F. Sundstøl Livestock Production Science 68 2001 13 –23
21 Table 9
Prediction of growth rate y of sheep g day from DM degradability of the diet
2
Factors Equation
R RMSE
P DMD
y 5 2271.216.7DMD
0.95 7.0
,0.001
24 h 24h
DMD y
5 2434.318.0DMD 0.91
9.9 ,0.001
48 h 48 h
DMD y
5 21008.4116.3DMD 0.57
20.6 ,0.01
72 h 72h
DMD y
5 2717.3111.1DMD 0.09
29.8 50.18
96 h 96h
A y
5 2 168.5 1 10.8A 0.91
9.6 ,0.001
B y
5 310.8 2 5.2B 0.77
15.5 ,0.001
c y
5 2 96.0 1 3651.9c 0.86
12.3 ,0.001
L y
5 2 55.6 1 44.5L 0.35
26.5 50.045
A 1 B
y 5 419.15 2 5.5A 1 B
0.35 26.3
50.042 ED
y 5 2360.218.8ED
0.95 7.3
,0.001 A B
y 5 2 26.97 1 8.2A 2 1.7B
0.94 8.4
,0.001 A c
y 5 2 149.9 1 7.0A 1 1509.8c
0.95 7.8
,0.001 A B c
y 5 2 221.6 1 6.8A 1 0.96B 1 2211.0c
0.95 8.2
,0.001 A B c L
y 5 2405.8 2 9.0A 2 31.1B 2 6202.9c 2 184.0L
0.96 7.9
,0.001 RMSE
5Root mean square error. For other abbreviations, see Table 2.
lation and its enzyme systems, because they are a rate. According to Yalc¸in et al. 1998, the potential
degradability of cereal straws was not related to source of nutrients for the rumen’s micro-organisms
growth rate and was poorly related to DM intake. Carro et al., 1991. Thus, it is logical that the A and
Fonseca et al. 1998 also reported that the potential c values of the feed have a relatively high precision
degradability had the lowest precision of predicting of predicting feed intake, digestibility and growth
DMI, DDMI and live weight change. The potential rate. Ferri et al. 1998 also showed that both the A
degradability represents the fraction that would be and c values had high positive correlation with DMI,
degraded if samples were incubated in the rumen for DDMI and DMD. Carro et al. 1991 found signifi-
a prolonged period and is not the same as the in vivo cant relationships between voluntary feed intake and
situation. Roughages with similar potential de- the soluble fraction of the DM, the rate of degra-
gradabilities could have different rumen retention dation of DM and the rate of degradation of NDF.
times and, consequently, different in vivo digestibili- However, the best fit was obtained when both the
ty and voluntary intake Carro et al., 1991. soluble fraction and the rate of degradation of the
The highly significant relationship of ED with feed DM were included in a multiple regression. On the
intake, digestibility and growth rate, found in the other hand, contrary to what would normally be
present study, is in conformity with the findings of expected, Chermiti et al. 1996 could not find any
Stensig et al. 1994a who showed a significant correlation between DMI and the degradation rate of
2
hays and straws. linear relationship R
50.77 between ED and in Reports in the literature are inconsistent regarding
vivo ruminal digestibility of NDF. The calculation of the precision of potential degradability in predicting
ED takes into account the assumed fractional feed intake, digestibility and growth rate. For in-
outflow rate in addition to the different degradation stance, contrary to the results of the current study,
parameters. In this respect, the use of ED in the Hovell et al. 1986 reported a strong linear correla-
prediction of feed intake and digestibility appears to tion between the DMI and potential degradability of
be more appropriate compared to the other degra- some hays. Ørskov et al. 1988b also showed that
dation parameters. According to Stensig et al. the potential degradability of cereal straws was
1994b and Madsen et al. 1997, the use of positively correlated with DMI, DDMI and growth
individual degradation parameters for prediction of rate. On the other hand, Kibon and Ørskov 1993
feed intake has the limitation that some feeds do not found that the potential degradability was not a
fit the degradation model and that the prediction only reliable predictor of intake, digestibility and growth
holds true for the feeds used to predict the relation-
22 A
. Tolera, F. Sundstøl Livestock Production Science 68 2001 13 –23
ship. Thus, Madsen et al. 1994 and Stensig et al. References
1994b developed a method where the rate of
Carro, M.D., Lopez, S., Gonzales, J.S., Ovejero, F.J., 1991. The
degradation in the rumen is combined with the rate
use of the rumen degradation characteristics of hay as predic-
of outflow to give an estimate of the physical fill of
tors of its voluntary intake by sheep. Anim. Prod. 52, 133–139.
the feed. Furthermore, Madsen et al. 1997 rec-
Chen, X.B., 1995. Neway Excel: An Excel Application Program
ommended a system based on the physical fill of
for Processing Feed Degradability Data. International Feed
feed in the rumen to be an appropriate system for
Resources Unit, Rowett Research Institute. Chermiti, A., Nefzaoui, A., Teller, E., Vanbelle, M., Ferchichi, H.,
evaluation of tropical feeds for ruminants as the
Rokbani, N., 1996. Prediction of the voluntary intake of low
dominant limitation for intake is physical fill of
quality roughages by sheep from chemical composition and
partly digested plant fibres in the rumen.
ruminal degradation characteristics. Anim. Sci. 62, 57–62.
The higher precision of DMD and DMD
Dhanoa, M.S., 1988. On the analysis of dacron bag data for low
24h 48h
than DMD and DMD
in predicting DMI,
degradability feeds. Grass Forage Sci. 43, 441–444.
96h 72h
Ferri, C.M., Jouve, V.V., Stritzler, N.P., Petruzzi, H.J., 1998.
observed in the present study, is consistent with the
Estimation of intake and apparent digestibility of kleingrass
findings of Yalc¸in et al. 1998. In contrast, Fonseca
from in situ parameters measured in sheep. Anim. Sci. 67,
et al. 1998 reported that DMD and DMD
96h 72h
535–540.
were the best predictors of in vivo organic-matter
Fonseca, A.J.M., Dias-da-Silva, A.A., Orskov, E.R., 1998. In
digestibility. The use of the A value together with the
sacco degradation characteristics as predictors of digestibility and voluntary intake of roughages by mature ewes. Anim. Feed
B and or c value in a multiple linear regression
Sci. Technol. 72, 205–219.
further improved the precision of prediction of feed
Hovell, F.D.Deb., Ngambi, J.W.W., Barber, W.P., Kyle, D.J., 1986.
intake, digestibility and growth rate over the use of
The voluntary intake of hay by sheep in relation to its
the A value alone. However, further inclusion of the
degradability in the rumen as measured in nylon bags. Anim.
lag phase in the multiple regression equation did not
Prod. 42, 111–118. Kabatange, M.A., Shayo, C.M., 1991. Rumen degradation of
have an advantage. Previous studies Ørskov and
maize stover as influenced by Leucaena hay supplementation.
Ryle, 1990; Khazaal et al., 1993; Kibon and Ørskov,
Livest. Res. Rural Develop. 32, 3 pp.
1993; Shem et al., 1995 also showed that the
Khazaal, K., Dentinho, M.T., Ribeiro, J.M., Ørskov, E.R., 1993. A
separate use of A, B and c values in multiple
comparison of gas production during incubation with rumen
regression resulted in high precision of predicting
contents in vitro and nylon bag degradability as predictors of the apparent digestibility in vivo and the voluntary intake of
feed intake, digestibility and growth rate.
hays. Anim. Prod. 57, 105–112. Kibon, A., Ørskov, E.R., 1993. The use of degradation charac-
teristics of browse plants to predict intake and digestibility by
5. Conclusion
