Tải bản đầy đủ - 0 (trang)
III. The Digestibility of Forage Crops

III. The Digestibility of Forage Crops

Tải bản đầy đủ - 0trang



periods of at least 10 days are recommended (Raymond et al., 1953); this

presents the particular difficulty with fresh forages that the forage must be

cut daily, and so may change in digestibility and chemical composition

during this experimental period.

In many studies this day-to-day variation in feed characteristics has

been overcome by cutting at one time sufficient fresh forage for the complete digestibility experiment, and preserving this forage so that it can

be fed over an extended period. Storage as hay (J. R. Jones and Hogue,

1963) or after artificial drying (Kivimae, 1959) has been used but, because of the changes in digestibility possible with these methods, cold

storage of forages has been adopted by some workers (Raymond et al.,

1953; Pigden et al., 196 1 ; Minson, 1966).

The technique of storage at 5°F. has been described in detail (Commonwealth Agricultural Bureaux, 1961, pp. 88 and 150); it has been

shown to have a negligible effect on the digestibility of the dry matter or

organic matter in forage (Raymond et al., 1953) or of the rate of digestion within the rumen (Pigden et al., 1961), but slightly reduces the

digestibility of the crude protein fraction (Raymond et al., 1953; Minson, 1966).

An alternative technique, the continuous digestion trial with fresh

forage, is now being increasingly widely used (Greenhalgh er al., 1960;

Commonwealth Agricultural Bureaux, 196 1 ; Ademosum et al., 1968).

Herbage is cut and fed daily over an extended period, and the amounts

of forage eaten and feces voided are measured daily throughout the experiment. The amounts of forage eaten and feces are summed over 5day subperiods, allowing a 2-day lag for passage of the feces, and digestibility coefficients are calculated on these subperiods, each of which

serves as the preliminary (adaption) treatment for the succeeding subperiod. This technique has proved of particular use in association with

grazing experiments (see Section X,A,2), but it is less accurate than the

cold-storage technique because of the shorter balance periods used.

The measurement of the digestibility of forages conserved by natural

or artificial dehydration presents no such problem, and most of the reported data on forage digestibility relate to such feeds. Silage is generally removed daily from silos for feeding, but cold storage of silage

(Harris and Raymond, 1963) requires less labor, and eliminates any risk

of day-to-day variation in silage quality.

The many thousands of recorded determinations of forage digestibility

have been collated at intervals and provide the broad background to our

present understanding of forage nutritive value (Schneider, 1952; Leitch,

1969; tropical forages, Butterworth, 1967). However, such compila-



tions may be of limited value in indicating the digestibility of an “unknown” forage, because of the difficulty of identifying it with a particular

feed class. This problem, long recognized, led to the development of

techniques such as the Weende feed analysis for estimating nutritive

values; a major advance in the period under review has been in the development of improved laboratory techniques for predicting the nutritive value of forages, to replace wherever possible the laborious and

expensive in vivo determination.






As Van Soest (1968) has noted, animal nutrition has had a history of

inertia and complacency, each further experiment carried out with old

techniques and old terminologies making it yet more difficult to adopt

new ones. But it is still difficult to create a logical pattern from the torrent of new analytical techniques and new parameters of nutritive value

that have recently been put forward to replace these older concepts.

The requirement is to establish a relationship between a nutritional

parameter (e.g., digestibility) of forages, measured in controlled in vivo

experiments, and the chemical composition of the same forages, from

which the nutritive value of other forages can be predicted. Digestion of

forage by the ruminant is a most complex process; yet for nearly a century

the attempt was made to predict the extent of forage digestion in terms of

its proximate analysis based on Weende crude fiber, crude protein and

nitrogen-free extractives. Sullivan (1 962) and Dijkstra ( 1 966) have both

shown that when these analyses are applied to a limited range of forages

close relationships between digestibility and chemical composition can

be established, but that these relationships become less precise as the

range of forages included is increased. As a forage crop matures its fiber

content increases and it becomes less digestible; a close negative relationship between fiber content and digestibility is found. But this relationship is likely to differ from that with a different forage species (in

particular, tropical forages; Butterworth, 1963) or from that with the

same forage species at a different time of year; in each case the forage

becomes less digestible as it becomes more fibrous, but at a given fiber

content different forages can have markedly different levels of digestibility. To some extent this can be overcome by using tabulations of relationships, each based on a limited feed class (Dijkstra, 1966). But again

these pose the problem of allocation to a particular feed class; more

seriously, they add little to our basic understanding of the factors that

determine forage digestibility.



The inadequacy of crude fiber as a determinant of nutritive value was

clearly established by Norman ( 1 935). Tentative alternatives to crude

fiber were proposed: cellulose (Crampton and Maynard, 1938), holocellulose (Ely and Moore, 1955), modified acid-detergent fiber (Clancy and

Wilson, 1966). Each of these aimed to analyze a more precise chemical

grouping than crude fiber, but each perhaps reflected the same basic

thinking, that the complex process of forage digestion can be quantified

by a single chemical analysis. The relationships between these “fiber”

components and forage digestibility (reviewed by Miller, 196 1 ; Sullivan,

1962), are often more precise than those based on crude fiber; they are

still inadequate for predictive purposes.

This conclusion, which had become evident by 1960, stimulated the

two main developments discussed below: the study of chemical techniques more relevant to the digestion process, and of biological techniques that attempt to simulate the process of rumen digestion by a

laboratory technique.



Forage digestibility, Eq. (2), is the summation C% content X % digestibility of all the different chemical components in the forage. Some of

these components, such as soluble carbohydrates and organic acids, are

completely digested as the forage passes through the ruminant tract;

others, in particular the polysaccharides and lignin, are much less completely digested and comprise most of the feed residue excreted as feces.

The “classical” approach, discussed above, assumes that the extent to

which the fiber fraction is digested is directly related to the proportion

of that fraction in the forage. Detailed studies of the digestibility of different fiber fractions, based on in vivo experiments, have clearly shown

that this is not so. Thus Jarrige and Minson (1964) found that there was

no decrease in the digestibility of the cellulose in S.24 ryegrass as the

cellulose content increased from 14.1 to 19.0 percent of the dry matter in

early spring, while Gaillard (1962) and others showed that the cellulose

in alfalfa is much less digestible than that in grasses with the same content of cellulose.

This led to the development of techniques of graded extraction with

reagents of increasing concentration (Gaillard, 1958; Jarrige, 196 1 ;

Burdick and Sullivan, 1963) and of cellulose solubility in cupriethylenediamine (Dehority and Johnson, 1963) which take some account of the

chain length and resistance to digestion of the different polysaccharide

fractions. However, no single procedure is likely to give results relevant

to the wide range of polysaccharides and lignin that comprise the fiber



fraction in forages, and Gaillard (1966) has developed a more comprehensive relationship between forage digestibility and composition:

Digestibility of organic matter % = 0.37(C-19.19) - 5.51(L-5.58) - 0.51(H-18.10)


+ 4. I I(U-3.80) 65. I


which includes the percent contents of cellulose (C), lignin (L), hemicellulose (H), and anhydrouronic acid (U). More recently Gaillard and

Nijkamp (1968) have proposed a less complex analytical system, which

replaces the separate determinations of cellulose and hemicellulose with

neutral-detergent fiber (N DF, v.i.):

Digestibility of organic matter % = 66.7 - 4.64(L-5.19)- 0.14(NDF-48.05)

+ 2.95(U-3.47)


An alternative approach, developed by Van Soest (1 967) and Terry

and Tilley ( 1964a), emphasizes the contribution to total forage digestibility of the highly digestible cell-contents fraction in forages. These

workers have considered forage to contain two main fractions, the cell

contents which are almost completely digested, and the cell-wall constituents, which are only partly digested, and they have proposed analytical

systems that (a) separate these two fractions and (b) indicate the extent

to which the cell-wall fraction would be digested in the ruminant tract.

In a series of papers (summarized by Van Soest, 1967) this author has

described methods for separating a forage sample into a cell-contents

fraction soluble in neutral detergent (S), and an insoluble cell-wall fraction (neutral-detergent fiber, NDF), as well as a fiber fraction insoluble

in acid detergent (acid-detergent fiber, ADF) and lignin (L). In a key

paper (Van Soest and Moore, 1966), the digestibility of the N D F fraction was shown to be negatively correlated with log X (r = -0.98**)

where X , the concentration of lignin in the A D F fraction, effectively

measures the extent of lignification of the cellulose in the forage (in that

paper X was denoted as L, which was confused with percent lignin).

The mechanism by which lignin reduces fiber digestibility probably

includes the effects of physical incrustation, of lignin-carbohydrate

complexes, and of molecular bonds. Van Soest (1967) also showed that

the cell-content fraction in forages is almost completely digested (98

percent) by the ruminant. However, a significant amount of material

soluble in neutral detergent occurs in ruminant feces. This is not undigested plant cell contents, but consists of endogenous materials

(mucus, salts, bile residues, and undigested bacteria) resulting from the

digestion process; digestibility as measured by Eq. (2) is not the “true”

digestibility of the forage material, but the “apparent” digestibility, (feed

- feces) measuring the amount of feed digested, less this inevitable



endogenous loss associated with the passage of the feed through the

tract. Based on in vivo results with a limited range of feeds, Van Soest

(1967) calculated this fecal loss to be 12.9 percent of the dry weight of

forage eaten.

Van Soest (1967) was then able to compute the apparent digestibility

of forage:

Apparent digestibility of dry matter % = 0.98s

+ W ( 1 . 4 7 3 - 0.789 log X ) - 12.9


comprising the almost completely digested cell-contents (S),plus the cellwall constituents (W=NDF) digested to an extent depending on the degree of lignification of the A D F fraction ( X ) , and less the endogenous


It has not yet been possible to test this relationship on a wider range

of forages than those studied by Van Soest. But by taking account of the

differing contents and digestibilities of the two main fractions in herbage,

the cell contents and the cell-wall material, Eq. (5a) clearly represents

an important advance over the more empirical methods summarized by

Miller ( 1 96 1) and Sullivan (1 962).

In the course of the development of the detergent-fiber methods,

Van Soest ( 1 965b) examined the effect of the method of drying herbage

samples before analysis on the measured levels of acid-detergent fiber

and lignin. Drying temperatures above 50”C., particularly over an extended period, significantly increased the levels of both these fractions;

this artifact “fiber” was shown to result from a nonenzymatic browning

reaction, in which protein polymerizes with products of carbohydrate

breakdown, so that the “lignin” fraction in particular contains an abnormally high percentage of nitrogen. In earlier work this had been

corrected by subtracting %N X 6.25 from the apparent lignin analysis.

However, Van Soest recognized that natural plant lignins may contain

some nitrogen, and derived a relationship that would correct only for

the nitrogenous matter which might be attributed to heat damage:

% corrected lignin (L,)


1.208 X % measured lignin (LA)- 10.75

x %N in A D F + 0.42


The acid-detergent fiber (ADF) content is then corrected:

% A D F corrected = % A D F observed - (LA- L,)


From Eqs. (6) and (7) the factor log X in Eq. (5a), based on corrected

values for A D F and lignin, can be calculated.

The need for this correction must reduce the utility (and precision) of

Eq. (5a) and emphasizes the importance of adequate drying methods for

preparing herbage samples for analysis. The method of choice must surely



be freeze-drying (lyophilization). But the great majority of freeze-driers

in current laboratory use are of small capacity (< 1 kg. water/24 hours),

and this can introduce a source of error which is seldom recognized- that

the sample of forage which is dried by this ideal method may be so small

as to be quite unrepresentative of the material sampled. This possible

contradiction between the precision of the drying method and the accuracy of sampling has been discussed (Commonwealth Agricultural

Bureaux, 1961, p. 135); until much larger freeze-driers become available, the solution in many cases may be to dry forage samples of adequate

size as rapidly as possible at 100"C.,so as to minimize the time during

which nonenzymatic browning (which occurs only in the presence of

water) can take place. The individual investigator can then test the success of his own drying method by the application of Eq. (6) to analyses

on representative samples.

Recently Van Soest and Jones (1969) suggested a further refinement

to the concept summarized in Eq. (5a), by indicating that the silica

present in plant material may exert much the same effect as lignin in reducing the digestibility of the neutral-detergent cell-wall fraction. L. H. P.

Jones and Handreck (1967) discussed the forms and reactions of silica

in the food chain from soil to plant to animal. They pointed out that

silica absorbed by plant roots is carried in solution to the actively metabolizing tissues. As the transporting water is transpired, solid silica is

deposited on to the cell walls so that as these develop the polysaccharides

are intimately associated with encrusting silica as well as lignin. From

examination of the digestibility in vitro of forage samples of silica content

ranging from 0.5 percent to 5.4 percent, Van Soest and Jones (1969) proposed a modified form of Eq. (5a):

Apparent digestibility of dry matter % = 0.98s + W(1.473 - 0.789 log X)

- 3.O(SiO2)- 12.9


As yet the evidence is restricted to relatively few forages, but further

study may clearly indicate the need for refinement of the biological concepts implicit in Eqs. (5a) and (5b).





The inclusion of silica as a further component which may influence

forage digestibility illustrates the trend toward multicomponent chemical techniques for predicting digestibility. In effect, this accepts that no

single component can quantify the complex process of ruminant digestion,

and that this must be treated as a series of stages, each described by a

logical chemical evaluation, as in the decreasing digestibility of the N D F

fraction as the fiber becomes more lignified.



The inclusion of silica also illustrates a basic problem with chemical

methods of evaluation, that a relationship such as Eq. (5a), which is

found to be adequate with one population of forages, may give inaccurate prediction of the digestibility of other forages-in this case, of

forages of unusually high silica content. This could arise from two causes:

(a) the original relationship did not include all the components that exert

a significant effect on forage digestibility and (b) chemical methods measure the content of different components in forage samples; they do not

measure the physical distribution and organization of these different components within the plant, which must to some extent determine how far

the plant fibers are digested by the microorganisms within the rumen. The

chemical approach tends to treat a forage as a homogeneous material, an

increase in lignin content, for instance, being visualized as an increase

in lignification throughout the whole plant. In practice the forage plant is

more realistically considered as made up of morphologically “distinct”

fractions, each of which can be changing in chemical composition and

digestibility in a way not necessarily related to the other fractions, so

that chemical analysis (an average of the whole plant material) may well

not describe the summation of the individual plant fractions that make up

the digestibility of the whole plant.

Thus, parallel to the development of chemical methods of forage evaluation, described in the previous section, has been the development of

biological methods of evaluation, the artificial rumen or in vitro digestion

methods. Essentially these have attempted to simulate the process of

ruminant digestion by methods that can take account both of the overall

chemical composition of the forage plant and of the distribution and

physical interrelations of the chemical components within the different

morphological parts of the plant.

With the recognition that the digestibility of the “fiber” fraction of

forages would be most affected by these physical characteristics the initial

investigations of biological methods were concerned with fiber digestibility, and in particular with the digestibility of the cellulose fraction in

forages. Although details of technique differed, all these methods were

based on the incubation, under controlled conditions, of a sample of the

test forage with a mixed culture of the microflora taken from the rumen

of a forage-fed animal; the aim was to standardize the conditions of incubation so that the fiber in the forage sample was digested to the same extent as in the same forage when fed in an in vivo experiment (Quicke et al.,

1959; Lefevre and Kamstra, 1960; Karn et al., 1967). These techniques

were also used to measure the extent to which the dry matter (Clark and

Mott, 1960), organic matter (R. L. Reid el al., 19601, or energy content

(R. L. Reid et al., 1960; Baumgardt el al., 1962; Naga and El-Shazly,



1963) in forage was digested in vitro. In most cases the extent of digestion

in vitro was found to be less than that in vivo, and regression equations

were developed to allow prediction of in vivo values.

In an alternative system, a sample of dried forage is enclosed in a nylon

or dacron mesh bag suspended within the rumen in vivo, and digestibility

and rate of digestion are measured by the loss of dry matter or of cellulose from the sample (Lusk et al., 1962; Hopson et al., 1963). This technique could have the advantage that a normal microfloral population will

be maintained, although this will tend to be that characteristic of the

feed eaten by the host animal, rather than of the sample under test. However the technique does permit the comparison of large numbers of feed

samples under standard conditions, and it could be of use in ranking

forage samples in a breeding selection program.

This approach was analogous to that with the earlier chemical methods,

in attempting to predict the complex process of forage digestion by a

single procedure. As with the chemical methods, the accuracy of prediction was found to decrease as the range of forages examined was widened;

in particular marked divergences were found between results for grasses

and legumes (Shelton and Reid, 1960). Tilley and Terry (1963) suggested

that these discrepancies might be the result of correlating data from a

single digestion with rumen organisms with those from digestion within

the animal, which involves a mainly bacterial digestion within the rumen

followed by a mainly enzymatic digestion in the remainder of the digestive

tract. Within the rumen, the “digestible” polysaccharides, carbohydrates,

and protein in the feed are broken down by the action of the microorganisms there; some of the products of digestion are absorbed directly

through the lumen wall, but a considerable part serves as the substrate

for microbial growth, and is resynthesized into protein, polysaccharides,

and lipids within the proliferating bacterial and protozoal population.

These microorganisms, entrained in the residues of undigested fiber and

other feed components, then pass from the rumen to the abomasum and

duodenum. In these organs the digesta are acidified and further digested

by secreted enzymes that hydrolyze much of the bacterial and residual

plant proteins to amino acids. These are then absorbed as the main

source of amino acids for the metabolism of the host animal.

The undigested residue from the in vitro digestion of forage material

with rumen microorganisms is thus seen to contain, in addition to undigested feed, the rumen organisms which, in vivo, would be enzymatically

digested in the ruminant hind tract. Tilley and Terry ( 1 963) proposed that

this second stage should be simulated by subjecting the residue from the

in vitro bacterial digestion to a second enzymatic digestion. They ex-



amined several enzymes and concluded that a two-stage procedure comprising digestion by rumen microorganisms followed by digestion by

acid-pepsin gave the closest agreement with in vivo digestibility values

for the dry matter and organic contents in forages. This method showed a

correlation of 0.97 between in vitro and in vivo values when tested on a

wide range of forages, including grasses fertilized with different levels of

nitrogen, and legumes:

Digestibility in vivo = 0.99 X digestibility in vitro - 1.O I

(S.E. = f 2.3 I )


Similar high degrees of correlation have been found by O’Shea and Wilson ( I 965; r = 0.94), Wedin el al. ( 1966; r = 0.996) and Ademosum et af.

( 1 968; r = 0.96); Dent ( 1 963) found close agreement between two-stage

in vitro and in vivo digestibility results with brassicas and forage maize.

In a number of studies the precision of prediction of digestibijity in

vivo by this in vitro technique has been compared with the chemical

methods already discussed. Armstrong et al. (1964a) found that the

metabolizable energy and net energy contents of a series of dried grass

feeds were more accurately correlated with organic matter digestibility

in vitro than with cellulose or lignin contents; Bosman (1967) and

Ademosum ef al. (1968) have reported the in vitro method to correlate

more closely with in vivo digestibility than the chemical methods tested;

Engels and Van der Merwe (1967) have found the same result with lowdigestibility hays in South Africa. However, to date no direct comparison

has been reported with the improved chemical technique proposed by

Van Soest ( 1 967, Eq. 5 ) , and it is possible that these two techniques

differ little in the precision with which they allow prediction of forage

digestibility in the laboratory.

In fact the stage may be approaching at which little further improvement in precision can be expected. In interpreting the error terms of these

relationships, it is important to recognize that this error does not arise

solely from deficiencies in the laboratory technique used (chemical or

in vitro), but that errors are also associated with the actual measurement

of the in vivo digestibility of the forages, and with the fact that digestibility

in vivo is not a constant parameter of a particular forage. Thus digestibility determined in an animal experiment may depend on the amount of

forage fed, digestibility decreasing as the level of feeding increases (Moe

et al., 1963, and can be significantly reduced if the animal is parasitized

with stomach worms-probably the rule rather than the exception with

sheep (Spedding, 1954; Shumard et al., 1957). The standard deviation of

an estimate of digestibility is between 1.0 and 1.3% (Raymond et al.,



1953); as most digestibility determinations are made with only 2 or 3

sheep it is evident, even where the factors noted above are standardized,

that much of the error in the relationships noted must result from errors

in the in vivo determination, rather than in the concept or precision of

the laboratory determination.

Probably the main area for improvement lies then in the better standardization of the in vivo digestibility experiments, of the preparation of

the forage samples for analysis, and of the conduct of the laboratory


R. L. Reid et al. (1964) and Noller et al. ( 1966) have both reported

significantly higher levels of dry-matter digestibility in vitro in forages

prepared by freeze-drying than by oven-drying, presumably because of

the production of indigestible artifacts during oven-drying, as suggested

by Van Soest ( I965b). Tilley and Terry (personal communication), however, found no advantage of freeze drying compared with rapid ovendrying at 1Oo”C., although the same authors (1963) had reported considerable depression of digestibility in vitro in samples dried at temperatures above 1 10°C.

The concept that fiber digestibility is limited by physical incrustation

with lignin indicated that digestibility should be increased by fine subdivision, and both Dehority et al. (1962) and Tilley and Terry (1 963)

reported increases of up to 50 percent in digestibility as a result of grinding forage samples in a ball mill before in vitro digestion. However, within

the range of fineness of grinding found in forage samples ground by hammermill, Tilley and Terry (1 963) found no significant effect of particle


The most serious problem arises, however, in the lack of standardization of in vitro procedures between different laboratories. Barnes (1967)

reported the results of a collaborative study in which the in vitro digestibilities of the dry matter and cellulose in three forages were measured

at 17 laboratories. The mean values for cellulose digestibility after 24

hours ranged from 40.0 to 63.9 percent, reflecting the use of different

techniques in terms of sample size, preparation of the rumen inoculum,

pH control, etc. In contrast, Raymond and Terry (1966) have reported

close agreement between in vitro results from two laboratories using

identical procedures, and have stressed the importance of different

laboratories using the same “standard” forage samples as an additional

check on the reproducibility of the method.

Tilley and Terry ( 1 963) found that rumen liquors taken from donor

animals fed on several contrasting forages were of similar digestive

efficiency in the two-stage in vitro system, and Troelsen and Hanel(l966)



have reported that the potency of liquors taken from different sheep

differs less than the in vivo digestive efficiency between sheep. I n general the most important consideration seems to be that the diet of the

donor animals should contain at least 10 percent of crude protein (see

below), that it should give a sample of rumen contents from which a

strained liquor can readily be separated (i.e., the animals should be

fed on a coarsely chopped hay rather than on a pelleted ground feed),

and that the sample should be kept in an anaerobic condition and be prepared for inoculation of the digestion tubes as rapidly as possible.

It must be accepted, however, that this in vitro procedure, developed

with temperate grass and legume species, may not be directly applicable

with other temperate species, or with tropical forage species, and Drew

(1 966) has stressed that the system should wherever possible be checked

with relevant samples of known in vivo digestibility. Thus Raymond and

Terry (1 966) reported low in vitro digestibility levels when both the test

forage (0.7 percent N) and the feed of the donor animal were of low

nitrogen content, as can often occur with tropical forage species. The

higher level of digestibility in vivo resulted from the animal’s ability to

recycle urea via salivary and ruminal secretions, whereas digestibility

in vitro was limited by a deficiency of nitrogen in the combined sample

and inoculum. Addition of 6 mg. of N , as urea, to the in vitro system increased sample digestibility to the in vivo level.

Engels and Van der Merwe (1967) found that the difference between

in vitro and in vivo digestibility values of veldt grasses became greater

as the nitrogen content of the test forages decreased. Addition of 20 mg.

of urea N to each digestion tube gave in vitro values in close agreement

with those in vivo. In a modification of the method of Tilley and Terry

( 1963), Alexander and McGowan ( 1966) have included ammonium sulfate in the buffer added to each tube. However in the author’s opinion

this is advisable only where depressed levels of in vitro digestibility result from low nitrogen contents. Engels and Van der Merwe (1967)

showed a marked depression in digestibility when 60 mg. of urea N was

included in the digestion system, and a similar depression might occur

when urea is added to an in vitro digestion of a forage sample already of

high nitrogen content.

Although some modification may be necessary in particular situations,

the study reported by Barnes (1967) does emphasize the importance of

the general adoption of standardized in vitro digestibility procedures,

without which results reported by different laboratories cannot be directly

comparable. The two-stage procedure described by Tilley and Terry

(1963) is now used by many laboratories, and there seems a strong case

Tài liệu bạn tìm kiếm đã sẵn sàng tải về

III. The Digestibility of Forage Crops

Tải bản đầy đủ ngay(0 tr)