Van Soest, digestibility, cell wall, nitrogen content, hemicelluloses, crude fibre, interferences, estimate, interference, energy values, proximate analysis, American Association of Cereal Chemists, feeds, human nutrition, dry matter, AOAC, plant cell wall, lignin, tannin content, European Economic Community, permanganate, hemicellulose, developing countries, Development Research Centre, ADF, energy value, the Government of Canada, analytical techniques, nutrition, plant cell, dietary fibre, microbial products, Lignin Pentosans Cellulose, NFE, John Gorham, Decrease Van Soest, fibre sources, bacterial fermentation, Standardization, feed evaluation, Ottawa CA International Union of Nutritional Sciences, The International Development Research Centre, Ottawa, Canada, organic residues, IDRC, International Development Research Centre, International Agency for Research on Cancer
:' :' CANADA
0" V4(IO'V O'Y
I cc'. .
The international development
Research Centre is a public corporation created by the Parliament of Canada in 1970 to support research
designed to adapt science and technology
to the needs of Developing Countries
. The Centre's activity is concentrated in five sectors: agriculture, food and nutrition science
s; health sciences; information sciences; social sciences
; and communications. IDRC is financed solely by the Government of Canada
; its policies, however, are set by an international Board of Governors. The Centre's headquarters are in Ottawa, Canada. Regional offices are located in Africa, Asia, Latin America
, and the Middle East.
©l980 International Development Research Centre Postal Address: Box 8500, Ottawa, Canada KIG 3H9 Head Office: 60 Queen Street, Ottawa
Pigden, W.J. Baich, C.C. Graham, M. IDRC, Ottawa CA International Union of nutritional science
s IDRC-134e Standardization of analytical methodology for feeds : proceedings of a workshop held in Ottawa, Canada, 12-l4 March 1979. Ottawa, Ont. IDRC, 1980. 128 p. : ill.
/IDRC publication/. Compilation on /animal nutrition/ /nutrition research/ applied to the /evaluation/ of energy values of Jfeed/s and the / standardization / of analytical / methodology / discusses / biochemistry/ aspects, practical rationing systems, /nitrogen/ evaluation, /sugar cane/ feeds /classification/, /trade/ and /legal aspect/s of /technique/s. /List of participants /.
Microfiche edition available
3ou IDRC- 134e Standardization of Analytical Methodology for Feeds Proceedings of a workshop held in Ottawa, Canada, 12-14 March 1979 Editors: W.J. Pigden, C.C. Balch, and Michael Graham Cosponsored by the International Development Research Centre and the International Union of Nutritional Sciences
Foreword 3 Participants 5 Summary and Recommendations 7 Evaluation of the energy value of feeds: overall appreciation A.J.H. van Es 15
Problems of standardization P.W. Moe 25
Application of practical rationing systems Alderman 29
Feed evaluation systems for the tropics of Latin America 0. Paladines 36
A new technique for estimating the ME content of feeds for poultry I.R. Sibbald 38
Sheep as pilot animals D.P. Heaney 44
Systems of analysis for evaluating fibrous feeds P.J. Van Soest and J.B. Robertson 49
Prediction of energy digestibility of forages with in vitro rumen fermentation and fungal enzyme systems Gordon C. Marten and Robert F. Barnes 61 Relationships of conventional and preferred fractions to determined energy values D.J. Minson 72
Description of sugarcane feeds: nomenclature and nutritional information E. Donefer and L. Latrille 79
Appreciation of the nitrogen value of feeds for ruminants R. Vйritй 87
Trade and legal aspects
of analytical techniques for feeds C. Brenninkmeijer 97
Standardization of procedures Elwyn D. Schall 106
Relationship to INFIC: feed data documentation and standardized methods Haendler 114
Systems of Analysis for Evaluating Fibrous Feeds P.J. Van Soest and J.B. Robertson'
Crude fibre has been and remains a common means of evaluating fibrous feeds. It is, however, grossly misleading and involves large errors on the basis of nutritional and biochemical criteria. In searching for a replacement there is a conflict among criteria that evaluate the analytical parameters as indicators of nutritive value. The favoured criterion is that which evaluates on the basis of recovery of refractory and indigestible residues. Another important consideration is economics of the methodology, which needs to be competitive with proximate analysis in terms of laboratory cost and tech- nician time. At the present time two systems are in use: the detergent system of fibre analysis and that devel- oped by Southgate in England. The Southgate methodology is used mainly in human nutrition. There are a number of important and somewhat divergent modifications of the detergent system in use. The establishment of any one methodology as a standard system of analysis will require some refinement and further collaborative study. The lead in this direction has been taken by the human nutritionists: International Agency for Research on Cancer (IARC) of the European Economic Community
. This group has sponsored a committee on fibre methodology, which has conducted one collaborative study. Another committee sponsored by the American Association
of Cereal Chemists has adopted a modification of neutral-detergent fibre.
The object of laboratory feed analysis is to derive compositional information from which estimates of animal responses to dietary inputs can be made. The purpose of this paper is to review systems of analysis for fibre, which has served as a negative index of energy availability in feeds and forages. Recently, a new dimension has been added with the suggestion that human diets in the developed countries are fibre-deficient. This has put emphasis on the positive aspects of dietary fibre quality that parallels the effort to utilize fibrous waste as ruminant feed. The need for better methods of fibre analysis stems from the deficiencies of crude fibre analysis and associated proximate analysis. The weaknesses of the crude fibre method have been known for a long time but have only recently become recognized as a real concern in monogastric and human nutrition. Most of the effort to improve fibre analysis has been in the ruminant field. The replacement of crude fibre with a more accurate and scientific method
is a problem involving perhaps more politics and economics than chemistry. Crude fibre is the legal method in 'Department of Animal Science
, F.B. Morrison Hall, Cornell University
, Ithaca, New York, USA 14853.
the UNITED STATES
and many other countries, and legislative action may be required. At the same time an adequate system of feed analysis must not only satisfy scientific criteria but also be sufficiently convenient and economical so as to be competitive with the proximate system of analysis including crude fibre. This latter objective has been a difficult one in view of the problem that true dietary fibre cannot be adequately described by any one single procedure. There are at present two other interNational Committee
s charged with developing and recommending a suitable technique for evaluating and characterizing human and animal diets. The Medical ReSearch Committee
of the European Economic Community (EEC) in conjunction with the International Agency for Research on Cancer (IRAC) of the World Health Organisation sponsors a fibre work group. This EECIRAC fibre group has had two meetings, one in Lyon, France in December 1977 and the other at Cambridge, England
in December 1978. It has conducted one collaborative study on dietary fibre methods. The other group is the International Organization for Standardisation (ISO) with headquarters in Amsterdam, The Netherlands
. In addition, there is a fibre committee of
the American Association of Cereal Chemists (AACC), which has adopted an amylase modification of the neutral-detergent method (Schaller 1976), and a nonnutritive residues referee of the Association of Official Analytical Chemists (AOAC). The EEC-IRAC fibre committee has so far taken the lead in fibre studies and has made the recommendation that crude fibre be dropped even if no other alternatives are immediately available. It feels that crude fibre values are so misleading that they are of little use in human nu- tritional studies. The group has not recom- mended any one procedure although many of the members are using some modification of the de- tergent system (Goering and Van Soest 1970) or the Southgate (l969a,b) methodologies.
Laboratory analysis of forages and feedstuffs is necessary to have a conveniently quick answer to feed quality. However, chemical analysis
also provides some understanding of the nature and effects of a feed. Laboratory evaluation of forage is essentially aimed at obtaining analytical data that predict the extent of biological degradation under specified conditions. The biochemical problem becomes that of assaying the limiting factors in the substrate; hence, there is emphasis on lignins and other components of the fibrous cell wall
of plants that provide resistance to digestion. Table 1. Correlations of various forage components with in vivo voluntary intake and digestibility for 187 forages of diverse species (from Van Soest and Mertens 1977).
Digestibility (in vivo) Digestibility (in vitro)a
Intake +0.61 +0.47
Lignin Acid-detergent fibre Crude protein
-0.08 -0.61 +0.56 -0.75
-0.61 -0.75 +0.44 -0.56
Cell wall Hemicelluloses
aTWO stage procedure of Tilley-Terry as modified by Goering and van Soest (1970).
The applied objective is to be able to formulate diets from compositional information that will elicit predictable animal responses. This desire leads to the net energy concept of feed evaluation. However, net energy is a complex of digestibility, intake, and efficiency phenomena that is in turn dependent on a variety of factors. Nutritional parameters of feed quality are variably correlated with digestibility or intake in ruminants (Table I). Cell wall is better related to intake than digestibility even though it recovers the undigestible frac- tions in the feces. In the case of monogastrics, cell wall is probably a better indicator of digestibility. Fibre quality is an important variable and evaluation will require analysis for components includ- ing lignin, cellulose, and hemicelluloses. Although these components are important in themselves, they cannot be individually a substitute for a total fibre analysis. This is a relevant criticism of those who wish to replace crude fibre with cellulose. The basic requirement for a feed fraction to have an effect on digestibility is for its content to be correlated with its digestible quantity in the diet. This is the so-called Lucas test and is an essential component of modeling studies. Further, for any component to have a consistent effect upon digestibility it must have a true causative effect. Lignin is such a component, but its effect is limited to the Plant Cell
wall, which is in turn another variable. These points are the foundation of the summative equation for estimating digestibility (Van Soest and Jones 1968). Crude fibre and cellulose are not such prime factors but are associated secondarily with digestibility through association with plant age or maturity. Cellulose is correlated to digestibility only to the extent that it is correlated to lignin. This association fails in many cases, particularly in tropical forages and aftermath cuttings of temperate forage (Van Soest et al. 1978). This lack of association is perhaps the single most unsatisfactory aspect of the use of the proximate system in evaluating forage from developing countries. Definition of Fibre The recent interest in the role of fibre in human nutrition has led to the advancement of the concept that total dietary fibre is the polymeric substances from plants that are resistant to mammalian digestive enzymes. This definition contains more than lignin, cellulose, and hemicel- luloses and includes pectins, gums, galactans, etc., which are relatively soluble materials. These are for the most part completely degraded by rumen bacteria and the bacteria of the lower tract of
nonruminants, and therefore do not contribute to the true indigestible fecal fraction. They will, however, act as a substrate for the intestinal microflora and affect the quantity of microbial products voided. The EEC-IARC committee has re- commended determination of the insoluble (neutral-detergent residue) and soluble components separately as these may have varying effects upon digestion. The insoluble undigested fibre is the principal fraction promoting passage of food residues in man (Van Soest et al. 1978). The work on fibre in humans has provided a model for application to other nonruminant species where work with fibrous diets is less advanced. Here it is important to state the principle that the recommendation of a world committee must consider an overall view of the role of fibre in animal nutrition in regard to its methological choices. Categories of organic residues from foods rela- tive to the dietary fibre definition include the fol- lowing: Matter that is available but which escapes through fast passage and slow rate of digestion to the lower digestive tract. Further competition between bacterial fermentation rate and passage determines the amount that may escape and be ex- creted in the feces. Matter unavailable to mammalian digestive enzymes but which is potentially fermentable may be lost to the feces through competition between passage and fermentation rates. These frac- tions are recovered in neutral-detergent fibre. Unavailable and unfermentable matter that is affected only by passage rate and excreted in the feces. Ruminant studies show that obliately unfermentable cellulose and hemicellulose is about 2.5 times the amount of dietary lignin (Smith et al. 1972; Mertens 1973). These carbohydrates will become available to fermentation if the lignin-carbohydrate bond is broken by chemical pretreatment of the dietary fibres. Microbial matter not originally present in the diet, but generated through microbial action on dietary residues and endogenous secretions. Gross microbial composition includes major amounts of protein and lipid, which may represent the main sources of increased excretion of fecal dry matter when fibre sources are fed. About 30% of the microbial dry matter is composed of cell wall or capsular matter that is resistant to digestion by mammalian enzymes. It is composed of muramic acid complexes including glucosamine and diaminopimelic acid in the polymer (Mason 1969). Enzymatic methods of analysis will include this fourth fraction as a part of the fecal fibre excre-
tion and may fail to distinguish it from genuine fibre fractions that survive from the diet. The microbial cell wall is soluble in neutral and acid detergent. The detergent methods are specific for plant cell wall and thus are useful for making the separation of microbial and plant-derived resi- dues. Criteria For Evaluating Analytical Systems Two contrasting and conflicting sets of criteria have been used to evaluate analytic procedures for their relevance to nutritive evaluation. One is the recovery of unavailable residues in the fibre residue and the other is the degree of correlation of digestibility with the measured parameters. Ironically, the residue that recovers the indigestible fractions is the neutral-detergent fibre, which is poorly correlated with digestibility in ruminants (Table 1). Cell wall content of the diet correlates highest with forage intake of ruminants. Plant cell wall is a better estimate of indigestibility of nonruminant diets (Henry 1976). The degree of correlation is generally an unsatisfactory criterion, although it reflects the practical desire of obtaining the most accurate estimates of nutritive value from composition. The unsatisfactory aspect arises because most standardized feeds and forages (through digestion trials) are unrepresentative of the environmental and physiological Factors Affecting
plant composition (Van Soest et al. 1978). Any standard set developed on the basis of cutting dates at a university farm are apt to reflect only that environment and practice, which are often ideal relative to that which may be sampled in the adjacent countryside. In a statistical sense such a standard would reflect no variation in quality due to soil and climate as well as that due to variable practices in management. The lack of this variation would in turn bias prediction equations relative to field samples. It would be better to develop a standard set for calibration by random selection and evaluation of farmers' products. If the objective is to obtain the most accurate estimate of digestibility, more than one analysis becomes essential. The effects of lignification are restricted to the plant cell wall allowing a model to be constructed that assumes non-cell-wall (lOO-NDF) is completely digestible; whereas, the digestibility of the cell wall itself is estimated by its lignin content. This system, the summative equation (Goering and Van Soest 1970), will satisfactorily estimate digestibility of mixed forage populations from diverse environments (standard error
of about 3.6). It will not improve the evaluation of first-cut single species relative to in-
dividual analytical parameters. The summative system requires determination of cell wall (NDF), ADF, and lignin. A silica or insoluble ash correction may also be needed. For a single measurement of digestibility in mixed populations the Tiley-Terry in vitro rumen procedure or a modification of it remains the most accurate. The limit of practical prediction is a standard error of about 3-5, which is the practical level of animal variability under producing conditions. The animal error is variable and depends in part on level of intake and the specific diet (Van Soest l973b). The reliability of regression systems depends on whether the standard forages, upon which the equation is founded, reflect the balance of species and environmental interactions characteristic of the forages to be tested. Strict standards must be kept regarding the forage populations used to test the system. There should not be less than 20 forages of determined animal digestibility. Legumes and grasses should be equally represented and several species of each included. Reference forages should come from localities similar to those
of the forages to be tested. Aftermath cuttings should be included as well as first cuttings with age of plant. The reliability of the regression system should be tested on a population other than that from which the equation was derived, but from similar localities. This allows the possibility of ascertaining two types of error: the standard error of an estimate; and the bias of systems to over or underestimate the correct value. When prediction systems are tested against a properly selected group of forages, more realistic estimates of the predictive errors are obtained (Table 2). Systems based on crude fibre and protein involve large errors because of the failure to assess the effect of environment. Protein is asso- ciated positively with digestibility through its decline with age of the plant, but nitrogen fertiliza- tion increases crude protein content without greatly altering digestibility. Equations based on proximate analyses, while using a large data base (Schneider et al. 1952), suffer from the limitations of the proximate system and the historical nature of much of the data.
Table 2. Predictive errors associated with systems to estimate digestibility from composition. Evaluation is with a balanced group of legumes and grasses of vary- ing geographic origin (Van Soest and Jones 1968; Van Soest 1973).
Method of estimation
Value Biasa S.D.b predicted units of digestibility
Acid-detergent fibre DDM
Equations based on
crude fibre and
Legume and grass TDN
With silica cor-
In vitro rumen digest-
True digestibility DDM
aMean difference between predicted and observed values. bstandard deviation
from regression. CDigestibility of dry matter. dEquations of Adams et at. 1964 based on the system of Axellson. eEquation of Goering and Van Soest 1970. t'Tilley-Terry procedure. 5Modification of Tilley-Terry according to Goering and Van Soest (1970).
Critique of Laboratory Methods Analytical parameters for forage and feedstuffs are of unequal value in providing useful dietary information and there is no such thing as a best method because most nutritive aspects of quality are complex. For example, a method that estimates digestibility may be unsatisfactory for evaluating intake or efficiency. Inmost forages, digestibility is a function of both plant cell wall and lignification. However, in some species, other factors have a role (tannins, silica, etc.). An individual analysis will be unsatisfactory if substantial feed variation is due to another unassayed factor. Therefore, an adequate system of analyses must attempt to assay the relevant limiting factors in feeds and forages. The following critique is developed from that point of view as well as some practical sense of laboratory economy and utility. Laboratory analyses can be divided into several categories: those that determine chemical entities; in vitro estimations of quality; and empirical tests. In the first category are analyses for specific feed entities such as lignin, cell wall, cellulose, etc., and in the second, enzymatic techniques. The empirical tests include crude fibre and the various dry matter solubility measurements. The first two categories satisfy some biochemical standards and provide the most valuable information in the form of actual composition or biodegradability. The third category is the least useful because the results can only be correlated with nutritive quality and the statistical associations are dependent
Table 3. The percentages of original feed lignin, pentosans, and cellulose dissolved in the crude fibre determination (from summary by Van Soest 1977).
Lignin Pentosans Cellulose
Legumes Range Average
aGymnospes and angiosperms exclusive of legumes and grasses.
on time and environmental interactions that may not be reproducible. Further, methods must be considered for their utility. There is a conflict of interest in efforts to modify methods according to ideal criteria because modifications almost invariably increase the length and complexity of a procedure. Crude Fibre and the Proximate Analysis The crude fibre method is of uncertain origin (Tyler 1975) and has been in use for at least 150 years. The earliest published analysis that is extant was done on Indian corn by John Gorham of Harvard in 1820 (Gorham 1820). Many authors attribute crude fibre to the German chemist H. Einhof. However, recent historical research (Tyler 1975) does not support this, and Einhof's published values, obtained by a maceration procedure (Einhof 1806) correspond to modern cell wall values (Van Soest 1977). Cell wall values, which represent the sum of lignin, cellulose, and hemicelluloses are higher than crude fibre in varying degrees depending on the food source. The error in crude fibre arises from the sequential extraction with hot dilute acid followed by hot dilute alkali. In this extraction sequence, 50-90% of the lignin, 0-50% of the cellulose, and upwards to 85% of the hemicelluloses are dis- solved (Table 3). The error through these losses is variable depending upon the proportions of lignin, cellulose, and hemiceliuloses in the fibre and can be as high as 700%. In the case of wheat bran, the most common source of fibre in human food, the true fibre value is about four times that indicated by the crude fibre value. For over a century we have known about the losses resulting from the crude fibre method and
its failure to recover lignin and other genuine components of fibre (Henneberg and Stohmann 1864; Van Soest 1975). Nevertheless, the Wiley Committee of the AOAC was instrumental in obtaining the approval of crude fibre as a legal official method in 1887 (AOAC 1887). Since that time the main effort of the AOAC has been to ensure analytical reproducibility within and among laboratories. Over the past 50 years there have been a number of attempts to develop improved fibre methods, none of which thus far has managed to dislodge crude fibre. This effort has been dominated, but not exclusively, by the fields of ruminant nutrition and grassland husbandry, where fibre utilization has been a main objective of research on forage quality (Raymond 1969). The methodology itself must be directed towards two separate goals that are not entirely compatible: for research purposes
, one needs a detailed system of structural analysis that is definitive in characterizing individual plant fibre sources; and for surveys or quality control work, the methods must be rapid and convenient even though some detail may be sacrificed. Whatever system is adopted, if it is to be competitive with the crude fibre method it must permit the handling of large numbers of samples, yet at the same time, yield more than a single measurement. Fibres are variable in their composition and properties, and it is not possible to describe the characteristics and amount of fibre with a single value. Nitrogen-free extract (NFE) The greatest and most fundamental error in the proximate system of analysis is the division of the carbohydrates between NFE and crude fibre. All attempts to unseat and replace crude fibre have attacked in one way or another the problem of carbohydrate fractionation and analysis. The AOAC recommended as far back as 1940 that reporting of NFE be discontinued. The NFE contains the cumulative errors of all the other determinations, the largest of them being due to the solubility and loss of much lignin and hemicelluloses in the preparation and determination of crude fibre. Even cellulose is not wholly recovered and the behaviour of different plant materials is quite variable (Table 3). Generally, the net solubility of lignin in grasses is greater than that in legumes. The error caused by the inclusion of cell wall fractions in the NFE is lowest in the case of concentrate foods where about three-quarters of the NFE is starch and soluble carbohydrates. In alfalfa, available carbohydrate and organic acids are about 50% of the NFE; whereas, in mature grasses and straws, very
Table 4. Comparison of relative digestibilities of crude fibre and NFE with the proportion of cell wall components (Van Soest 1975).
Cases where dig. CF>NFE
No. of samples
% No. of samples Cell wall Hemicellulose Lignin
hulls Forage (total)
temperate legume tropical legume
annual grasses temperate grass
tropical grass straws
little of the NFE is available carbohydrate. The effect of this error is to cause the apparent digesti- bilities of NFE to be less than those of crude fibre in a significant number of cases (Table 4). The presence of a prominent metabolic fraction in fecal NFE contributes greatly to this effect. The proportion of cases where digestibility of crude fibre equals or exceeds digestibility of NFE is about 30% for all feedstuffs, but is notably greater in the forages that contain more hemicelluloses and lignin. The error is largest in tropical grasses and straws. A further problem with NFE lies in the use of a factor (6.25) to convert nitrogen into an estimate of protein content. True protein forms only about 70-80% of feed nitrogen and only a very little of feces nitrogen so that the application of the 6.25 factor to all feed and feces nitrogen constitutes an error that is reflected mainly in the NFE. The magnitude of this error depends on the nitrogen content of the nonprotein nitrogen compounds and their deviation from the 6.25 ratio. This error is most serious in fecal analysis where little or no true protein at all is ordinarily found and the main nitrogenous constituents are microbial cell
walls, which contain 7% nitrogen. Most feces yield considerable NFE upon analysis and calculation but do not ordinarily contain any watersoluble carbohydrates. Insoluble starch is the only nonstructural carbohydrate likely to appear in feces, and then only at high intakes does it appear in substantial amounts. The Detergent System The system of analysis employing detergents was originally developed to solve analytical problems relative to ruminant diets, more specifically forages. Since that time (early 1960's) applications have been made to many animal species. The objective of the analysis is the fractionation of foods of plant origin relative to their nutritive availability and fibre content. The extention of the system to a general treatment of herbivorous diets for both ruminants and nonruminants leads to a new set of problems that are presently being faced. The truly indigestible components of feed are recovered in the neutral-detergent residue (NDF), while acid-detergent divides these into fractions soluble and insoluble in 1 N acid. The acid-sol-
NDF Lignin Cellulose Hemicellulose
ADF Kjeldahl---* Indig. protein
(bound in crude
lignin as Maillard
V ADF Lignin less tannin Cellulose (pectin free)
V Cellulose Cutin Pectin Silica
Crude Lignin True lignin Maillard products Cutin Leather
V Cellulose Ash
Lignin Cutin Ash
Fig. 1. Sequences of analytical treatments offeed samples subjected to the detergent system:pretreatmen: with neutral detergent dissolves tann ins, pect ins, and opaline silica that would otherwise contri- bute to acid-detergent fibre; permanganate removes zann ins but not cutin.
Table 5. Interferences in the estimation of hemicellulose as the difference between neutral-detergent fibre (NDF) and acid-detergent fibre (ADF).
Influence on hemicellulose estimate Reference
Cell wall protein Recovered
Increase Keys et al. (1969)
Biogenic silica Considerable solution Quantitative recovery Decrease Van Soest and Jones (1968)
Decrease Bailey and Ulyatt (1970)
Precipitation as protein complex
Decrease Robbins et al. (1975)
uble fraction includes primarily the hemicellu- Generally, acidic polysaccharides are more likely loses and cell wall proteins, while the residue to be insoluble in acid detergent through precipi-
(ADF) recovers cellulose and the least digestible tation as the quaternary ammonium detergent
noncarbohydrate fractions. Acid-detergent has salts. Pectic acids from legumes, citrus, etc. tend
the advantage of removing substances that inter- to precipitate giving high values for acid-deter-
fere with the estimation of the refractory compo- gent fibre. The pectins of other plants, e.g. Bras-
nents so that the ADF residue is useful for the se- sica, remain soluble, however (Bailey et al. 1978),
quential estimations of lignin, cutin, cellulose, in- and they have recommended that for many pur-
digestible nitrogen, and silica (Fig. 1). Silica, in poses where purity of the acid-detergent fibre is
contrast with neutral-detergent, is quantitatively sought, neutral-detergent extraction should pre-
recovered in the acid-detergent residue (Van cede acid-detergent. Preextraction will remove
Soest and Wine 1968).
the interferences of pectin, tannins, and silica, al-
Plant cell wall as measured by neutral-detergent fibre has proven to be the most fundamental feed characteristic determining feed value. However, it gives quite a poor relationship with digestibility because of the highly variable digestibility of plant cell walls. It follows then that the problem of digestibility prediction is that of estimating cell wall digestibility.
though in the case of silica its quantitative mea- surement will be lost. However, the amount of silica solubilized by neutral-detergent is an estimate of opaline silica. Similarly, estimates of the tannin content can be obtained by the comparison of acid-detergent fibre prepared with and without preextraction with neutral-detergent. A reverse situation exists with regard to cell wall proteins that are soluble in acid-detergent but not in neu-
Acid-detergent fibre is widely used as a quick tral-detergent.
method for determining fibre in feeds, often substituting for crude fibre, but used much on the same basis as proximate analysis. Nitrogen-free extract calculations based on ADF have appeared in the literature although such use has no scientific validity, the hemicelluloses, metabolic fecal matter, and available carbohydrates having been confounded. The use of ADF as a predictor of digestibility is not founded on any solid theoretical basis other than statistical association.
Sequential extraction allows the possibility of alternate routes of analysis which, if performed in parallel, offer the possibility of further differential analysis (Fig. 1). For example, the removal of tannin-protein complexes with neutral-detergent allows parallel lignin analysis to become a means whereby the tannin content can be estimated from the difference obtained between a direct route
and that following a preextraction step. Preextraction will allow a more accurate estimate
The intended purpose of ADF is as a prepara- of hemicelluloses, and by difference with the di-
tive residue for the determination of cellulose, rect, an estimate of interferences that may include
lignin, Maillard products, and biogenic silica. pectins, alginates, tannins, etc. A suggested route
The Maillard products are formed by complexing of analytical sequence is shown in Fig. 2.
of protein and carbohydrate upon heating or drying feeds. The heat damaged protein is totally unavailable through digestion and is recoverable in the fibre, specifically in the lignin fraction. The estimation of heat damage and unavailable nitrogen is rapidly assayed by preparation of aciddetergent fibre and, sequentially, its nitrogen content (Goering et al. 1972).
The analyses for lignin with permanganate or 24 N sulfuric acid (Klason lignin) offers the possibility of separation of the plant cuticle (resistant to KMnO4 oxidation) from the phenolic matter that is oxidizable in permanganate. Preparation of cellulose by oxidative means allows a cuticular fraction to contaminate the cellulose residue.
These problems illustrate the difficulty of de-
signing a single system of analysis for all condi-
Interferences in estimating hemicellulose by difference Neutral-detergent dissolves pectin, tannins, and a variable amount of silica; whereas, acid-
tions, and a necessary result is that a single analytical protocol cannot satisfy all conditions. An outline of possible sequences of analysis is shown in Fig. 1. Tannin-protein complexes contaminate
detergent recovers silica, the tannin-protein complexes, and pectin partially. Acid-detergent residues are usually lower in protein (nitrogen) than neutral-detergent residues. The influence of these effects on the estimate of hemicelluloses by dif-
the crude lignin fraction unless preextraction by neutral-detergent is done. The alternative lignin procedures by permanganate and sulfuric acid do not measure entirely the same fraction. Cutin is resistant to permanganate oxidation and is also
ference is shown in Table 5. Some of these errors insoluble in 72% H2SO4. Its estimation is, there-
are partly self-canceling when used statistically. fore, accomplished by the appropriate sequence.
1-g air-dry sample ND extraction (Robertson and Van Soest 1977) CMoCdeillfiseoldubles ND residue detergent CAHceimdicelluloses AD residue Lignin-cutin complex
(Van Soest 1973 a,b). Decalin has also been omit- ted from the preparation of neutral-detergent fibre. Lipid interference Fats and oils do not interfer at low levels as long as the detergent can form a stable emulsion. However, at higher levels of lipid (>10%) a separate phase can form. As the detergents (both cetyltrimethylammonium bromide and sodium lauryl sulfate) are soluble in the lipid phase, increased values of fibre can be obtained due to inadequate amounts of detergent inthewaterphase. An additional problem of high lipid materials is their greasy character and the resultant difficulty in grinding a dry sample. To solve both of these problems, we have employed the treatment of fresh sample with 4 volumes of acetone or ethanol to prepare a material that can be ground and is sufficiently low in lipid content to avoid interference in the detergent analysis. It is important that this step not involve the use of heat because this will affect the nitrogen content of the fibre.
Cutin + minerals з5OOoC Ash Fig. 2. Flow diagram for sequential analysis. Procedural Modifications to the Detergent System A number of procedural alterations have developed out of attempts to overcome certain analytical differences; in particular, the contamination of fibre residues with protein and starch, interferences in the hemicellulose estimate, and difficulties in filtration and handling-related problems in the analysis of certain foods. Some modifications have been made by other groups, including the modification of neutral-detergent fibre by Schaller (1978) and the system devised by Fonnesbeck (1976). Elimination of decalin Originally added to overcome foaming problems, this reagent was omitted as a result of the collaborative study on acid-detergent fibre and lignin, where it appeared that decalin increased the fibre yield and contributed to difficult filtering
Protein interference Protein may cause variation in the analysis of detergent fibres when the protein content is very high such that the sample exceeds the capacity of the detergent to form soluble complexes. This can occur on analysis of samples in excess of 30% protein. It becomes desirable to use a digestion with a protease in this instance. Specific addition of a protease is not ordinarily necessary, because the bacterial amylase employed in one modification of NDF (Robertson and Van Soest 1977) has significant proteolytic activity. Not all protein or other nitrogen can be re- moved from vegetable fibres by proteases. Indeed the resistant fraction is more or less recoverable in feces and appears to be indigestible. This nitrogen, which is recoverable in acid-detergent fibre and lignin, is comprised of several fractions, one of which is indigenous to the mature plant, another due to Maillard reactions and heat damage to protein in cooking and baking as well as the tannin-protein complexes already mentioned. It is because the Maillard products are so easily formed as an artifact in sample preparation that drying procedures should be kept below 65 °C. Sodium Sulfite This reagent was used to reduce the protein content of neutral-detergent fibre. Sulfite reduces fibre nitrogen content through its ability to cleave disulfide linkages in proteins. This capacity allows it to be a very effective means of eliminating keratinaceous tissues from animal-
derived foods and of such excretions in fecal analysis (Van Soest 1968). However, sulfite unfortunately attacks lignin and causes a significant loss. Another means of reducing nitrogen content of fibre is through the use of detergent-stable pro-
gestibility of these forages. Subsequent study in Britain using other standards more carefully prepared have not borne out the original observation (Alderman, Personal communication
). The MADF procedure includes oven-drying at 95 °C as a preliminary step. Unfortunately, this treat-
teases. However, the enzymes will not degrade resistant keratinized animal tissue. There appears at the present moment no satisfactory solution to the analytical problem of separating animal keratin from plant lignin, which is a particular prob-
ment sacrifices the use of acid-detergent as a means of assaying for heat damage and unavailable protein, which is one of the more valuable applications of ADF (Goering et al. 1972).
lem of analysis of diets of mixed animal and plant The Fonnesbeck System
origin. The present consensus is to omit the use of This system of feed analysis (Fonnesbeck 1976) sulfite except as required in specific instances. is very similar to the detergent system and is es-
Starch interference One of the main problems of filtering neutral- detergent fibre is starch that tends to form viscous solutions in hot neutral-detergent. The difficulty in filtration is aggravated by cooling during slow filtrations which increases viscosity. Direct NDF preparations in cereals and cereal products often show positive tests to starch (iodine test) indicating that the fibre values are elevated by this contamination. Two procedures utilizing amylases have
sentially derived from it. Determination of plant cell wall is conducted at pH 3.5 with sodium lauryl sulfate, having been preceded by a pepsin digestion. This is done to reduce the nitrogen content of the fibre and to eliminate starch interference. The analysis proceeds sequentially, hemicellulose being extracted with 4% H2SO4 then the lignocellulose residue being treated with 72% H2SO4 to remove cellulose and isolate lignin. Lignin is determined as loss in weight on ashing, the residue being acid insoluble ash.
evolved to overcome the starch problem. One developed by Schaller (1976) was a hog-pancreas enzyme, while that of Robertson and Van Soest (1977) was an amylase derived from Bacillus subli/is. The Schaller procedure requires separate treatment with the enzyme at pH 4.5 and filtration after detergent extraction. The shorter procedure of Robertson and Van Soest takes advantage of the compatibility and stability of the enzyme in hot neutral-detergent, allowing a more rapid and convenient procedure. The detergent reagent (actually the EDTA in it) inactivates the a 1-6 activity but does not restrict the solution of starch. Problems with the method are the need for repeated treatments of the residue
Critique The procedure sacrifices some of the speed of the detergent system in order to obtain purer fibre fractions. It is not certain that removal of nitrogen from cell walls is a benefit as it appears that this entity is real and associated with insoluble protein and the fraction promoting maximal protein output from the rumen (Pichard and Van Soest 1977). The sequence does not provide the alternatives of the detergent system where tannins, cutin, and Maillard products can be fractionated out of the crude lignin. Results from this procedure have not been compared with other modifications of the detergent system.
to remove starch in some instances and the resistances of modified starches to the enzyme. Comparison of the two methods in the collaborative work of Schaller shows essentially identical results for most food samples but slightly higher results on foods containing modified starch with the Robertson and Van Soest modification (Schaller 1978).
The Southgate System This system (Southgate 1969a,b) was developed for human foodstuffs low in dietary fibre. In this method (Fig. 3) the dietary fibre is fractionated into lignin, cellulosic polysaccharides, and noncellulosic polysaccharides. The latter fraction may be subdivided into water soluble and water insoluble noncellulosic poly-
Modified acid-detergent fibre The MADF was developed in Ireland using forage standards that had been dried postfeeding at a high temperature in the process of sample preparation. The investigators found that increasing the acid strength and prolonging the boiling time improved the relationship with di-
saccharides. The polysaccharides are estimated as their constituent simple sugars by chemical rather than gravimetric means with the choice of spectroscopy, gas-liquid chromatography, or high performance liquid chromatography depending on the degree of sophistication desired. About 5 g of the sample is extracted with 85% methanol to remove the free sugars, then about
'-.'5-g air-dry sample
Water soluble sugars
I,зCExDtirsaccartd4X 85% methanol insoluble residue
Hydrolyze with amyloglucosidase, add I 4 vol. ethanol
Residue Extract iN H2SO4
Residue Extract 72% H2SO4
tion. However, the system does not lend itself to rapid analysis and the precision of the chemical methods may not justify the time and labour required. Although the analytical methods
, especially GLC and HPLC are very precise, the extractions are not definitive in their fractionation of the carbohydrate. However, where sugar analysis is required, Southgate's system is probably the method of choice unless more exacting extractions can be justified. McConnell and Eastwood (1974) have reported that the ADF procedure is as precise as the Southgate methods for cellulose and lignin, and perhaps preferable if preceded by a ND extraction because no artifact lignin is produced. An integration of the Van Soest and Southgate methodology may be the route to a rapid, precise system of analysis. Recommendations and Future Needs While the desirability for a uniform system of feed analysis is great, the complexity of different purposes and applications may preclude recommendation of a single system at the present time. There are at the present time a number of groups working toward standardization of food and feed analysis and a political problem exists in coordinating the efforts of diverse groups in different places. Coordination is essential if a unified system is to emerge.
Residue - Weigh Sat. KMnO4 Lignin Residue - Weigh - Loss in Weight = Lignin Fig. 3. Flow diagram for sequential analysis in Southgate's system. 0.3 g of the methanol extracted residue is incubated with amyloglucosidase to remove the starch. After removal of the starch the residue is sequentially extracted with 1 N H2SO4 to remove the noncellulosic polysaccharides and 72% H2SO4 to hydrolyze the cellulose. The residue is then weighed, oxidized with saturated KMnO4, and reweighed to give an estimate of the lignin. The sugars in the various extracts are then measured and the dietary fibre estimated by summation of the noncellulosic polysaccharide, cellulosic polysaccharide, and lignin fractions. Critique This system has the benefit of recovering the pectins in the noncellulosic polysaccharide frac-
Dietary fibre has been defined as the plant polymeric substances resistant to animal digestive enzymes (Van Soest 1978) a definition endorsed by the EEC-IARC working committee in dietary fibre. This group has recommended the discontinuance of crude fibre. The AOAC recommended the discontinuance of nitrogen free extract (NFE) in 1940. Another suggestion for the replacement of crude fibre is the determination of cellulose (ISO). This suggestion is inadequate because cellulose represents only a portion of the total fibre and is a variable portion of it. The need is for an account of unavailable residues. The two most commonly used systems in present use are the detergent system of Van Soest and the system devised by Southgate in England. These systems while different in approach give similar results for many feeds and foods (McConnel and Eastwood 1974). The principal difference between the two systems is in regard to the soluble components that are resistant to mammalian digestive enzymes. The EEC-IARC Committee has recognized the problem of soluble substances including pectin and gums and points to the need for method development.
Further definition and refinement of procedures should make careful consideration of the problem of laboratory economy. The needs of detailed research differ from those of the quality control laboratory for broad surveys. Here the need arises for two compatible systems: a rapid one for surveys; and a second for composi- tional detail. A detailed system of analysis will preferably entail component analysis of cell wall constituents via suitable chromatographic and identification procedures.
The systems in question are available in modifi- cations that allow applications to almost all kinds of feedstuffs. Because data derived by these procedures are far superior to that obtained by proximate analysis, it is recommended that where choice is possible the newer methods should replace crude fibre. The attitude of the EEC-IARC Committee is that crude fibre be abandoned even if an alternative analysis is unavailable. Crude fibre should be deleted from existing tables of composition because its use has been mischievous and misleading in human nutrition.