Dự báo khả năng tiêu hóa cỏ của động vật nhai lại bằng phương pháp enzym

Prediction of digestibility of North Vietnamese grasses for
ruminant by the enzymatic method with "persine-cellulase"
of De Boever: first results

Paul Pozy, Vu Chi Cuong, Doan Thi Khang

and Daniel Dehareng


1. Introduction
The digestibility of feed for ruminants is classically determined by
measuring the quantity ingested and the faecal matter excreted by animals.
However, even with sheep this method is long and laborious. It is for that reason
that laboratory methods (in vitro) were developed to predict digestibility in vivo,
and on that basis determine the energetic value of feedstuffs. Among these
methods, one technique based on the use of "pepsine" and "cellulase", developed
in Europe by De Boever (1986) is often used. Recently, this method was
introduced in NIAH's chemical analysis laboratory. The present study follows two
objectives. First to see that De Boever's method has been correctly transferred to
NIAH. If that proves to be so, then determine the precision of the predictability of
in vivo digestibility based on in vitro measures according to De Boever's method.

2. Meterials and methods
Following practical and theoretical training of Vietnamese staff in
Louvain's laboratory of animal physiology in Belgiun, the method was introduced
in NIAH in such away that techniques, equipment and enzymatic products were
identical in both laboratories. Reference samples of tropical and temperate origins,
whose in vitro digestibility had been previously been determined by De Boever's
labotatoty, were sent to Hnoi to calibrate the method. Indeed, the method to
determine in labotatoty the digestibility of organic matter (in vitro) is an indirect
method; it requires a correction (a calibration) of the parameters measured with

reference to standard samples of known digestibility.

In this study, all the fodder studies were harvested under the practical
conditions of dairy farms around Hanoi. Part of them was subjected to in vitro
measures of digesbitility in UCL and in Hanoi (1
st
objective of the study; many
sample from 3 harvests of narural grasses made in April, June and August 1999).
Another part was subjected in Hanoi to in vitro digestibility measures (on sheep)
and in vitro, to determine intensity of the relation linking the two types of
measures, that is to evaluate the predictability of digestibility in vivo from in vitro
measures of digestibility (2
nd
objective). In this case, the correlation analysis has
focused, for practical purposes, on two groups of feed: narural glasses
harvested for 5 experiences in vivo throughout one year (17 pairs of measures)
and cultivated grasses (Pennisetum Purpurum, Tripsacum laxum, maize) including
rice straw (21 pairs and 7 experiences in vivo).

3. Results
The comparison of digestibility values in vitro (IVD
MO
) measures in UCL
and Hanoi (Table 1) on samples of narural grasses harvested in April (HNO4),
June (HNO6) and August (HNO8) shows that the difference is less than 2.5%, less
than the error generally allowed for the manipulations. The method is operational
in NIAH's labotatory.

Grass Digestibility in vitro (IVD
MO
)



MS

MO

UCL Lab.

NIAH Lab.

HNO4 49.7

88.1

57.64

57.60

56.97

56.13

55.73

56.27

58.20

HNO6

27.7

86.8

63.75

64.72

63.24



68.18

67.14

63.22

HNO8

32.3

85.1

57.46

57.35

55.51



55.81

58.89

55.35

T-test = 0.4888 with 2 degrees of freedom


Morever, the analysis of correlations has shown that a weak relation
(P<0.01) (R
2
=0.442) between the coefficient of digestive utilisation of the natural
grasses compiled from the in vivo method (CUD
MO
) and its comparative value
from the in vitro method (IVD
MO
):

Y (CUD
MO
) = 0.7807*X (IVD
MO

) + 14.3953 avec r = 0.6645**

In contrast, the relation (P<0.01) between the two digestive utilisation
coefficient is fairly strong (R
2
= 0.807) in the case of cultivated grasses:

Y (CUD
MO
) = 0.7778 * X(IVD
MO
) + 21.7623 avec r = 0.8985**

This difference in the strength of correlations established for the two groups
of feed, and therefore the precision of the predictability of the in vivo digestibility
from in vitro measures, probably comes from the greater variability in the
chemical composition (MO, NDF, MAT) of natural grasses than in cultivated
grasses.

Nevertheless, at this stage of use of the in vitro method, we can reasonably
state that in practice, the precision of the prediction is sufficient only for use in the
formulation of more accurate food rations.

4. Conclusions
The transfer of the De Bowver method has been successful and NIAH has
now a quick and performing method to predict the energetic value of feed for dairy
cows from the knowledge of the digestibility of their organic matter.

However, the continuation of the study of other feeds by the in vivo nad in vitro
methods, but the accumulation of digestibility data, will permit the consolidation

of the established correlations.



Acknowlegements

Thanks to the UCL lab. The Vietnamese and Belgian goverments financed
this study.



Summary



A study was undertaken to establish the possible correlations between in
vivo and in vitro digestibility of ruminat feeds . Parallelly, a series of in vivo
digestibility trials with sheep and in vitro digestibility experiments using pepcine-
cellulase technique developed by De Bover (1986) in Belgium was conducted. It
revealed that:

The difference in digestibility values in vitro (IVD
MO
) measures in UCL
and Hanoi on samples of narural grasses was less than 2.5% (accepted error), less
than the error generally allowed for the manipulations. The method is operational
in NIAH's labotatory.

The is a weak relation (P<0.01) (R
2

=0.442) between the coefficient of
digestibility of the natural grasses from the in vivo method (CUD
MO
) and its
comparative value from the in vitro method (IVD
MO
):

Y (CUD
MO
) = 0.7807*X (IVD
MO
) + 14.3953 avec r = 0.6645**

In contrast, the relation (P<0.01) between the two coefficients of
digestibility coefficient is fairly strong (R
2
= 0.807) in the case of cultivated
grasses:

Y (CUD
MO
) = 0.7778 * X(IVD
MO
) + 21.7623 avec r = 0.8985**