HAL Id: hal-00889583https://hal.archives-ouvertes.fr/hal-00889583

Submitted on 1 Jan 1996

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Biological methods to predict the nutritive value oftropical and mediterranean feedstuffs: a critical review

P Susmel, S Filacorda

To cite this version:P Susmel, S Filacorda. Biological methods to predict the nutritive value of tropical and mediterraneanfeedstuffs: a critical review. Annales de zootechnie, INRA/EDP Sciences, 1996, 45 (Suppl1), pp.39-51.<hal-00889583>

Biological methods to predict the nutritive value of tropical andmediterranean feedstuffs: a critical review

P Susmel S Filacorda

Dipartimento di Scienze della Produzione Animale, Via S. Mauro 2, I-33010 Pagnacco (UD), Italia

Summary - In Mediterranean and tropical areas, the specific characteristics of the plants andherbivores are the result of co-evolutive mechanisms and the interactions between feed andanimal are more important than under temperate conditions. The biological methods available topredict nutritive value have been proposed and developed under temperate conditions andsimulate or describe mainly feed utilisation in the rumen. In general, these methods do not produceunique and comparable results, because they are capable of measuring either the whole rumendigestion, or the fermentation process, or the degradation process or microbial synthesisindividually. In addition, the biological methods have different levels of simplification of the feedutilisation process and take into account, in different manners, the level of intake, the associationbetween feeds, the recycling of Nitrogen (N) in the rumen and the rate of rumen passage. In orderto predict the nutritive value of tropical and Mediterranean feedstuffs, biological methods havebeen compared in terms of their general and specific approach, level to robustness and sensitivity,specific requirements and level and type of predictability. In addition, the main differences betweentropical and temperate forages are described and discussed.

Introduction

In Mediterranean and tropical areas, thespecific characteristics of plants and herbivoresare the result of co-evolutive mechanisms. The

plants in tropical and subtropical areas havedeveloped a series of anatomical andbiochemical strategies to adapt to a series ofpressure factors: high temperatures, highevaporation, shortage of water and pressurefrom herbivores.

Domestic animals have developed as aresponse to limitations in the quality andquantity of the allowance of feeds, specificanatomical and physiological characteristicsand have covered a specific ecological andnutritional niche. The various species ofdomestic animals can be distinguished by theirdifferent feeding behaviour and food

preference, rumen environment, kinetics ofnitrogen recycling and transit time of feeds. Atthe same time, different management systemsrepresent an adaptation to the ecologicalconditions and simultaneously represent anevolutionary element in their own right. Undertropical and Mediterranean conditions, thenutritional habits of domestic animals dependon the feeding systems, feed allowances andfeeding behaviour and, therefore, during the

year and the day, the diet can change widely intype of feed and quality of nutrients.

This series of factors influences theestimation of the nutritional value of tropicalfeeds.

Biological methods, used to assist theevaluation of the nutritional value of tropicalfeeds, represent models which simulate a partof the feed utilisation process, with differentlevels of complexity; the information obtainedgenerally describes and predicts intake,digestibility and the protein value. Thesemethods have been developed for temperatefeeds and generally show a lower capability ofdescribing and giving useful information on thenutritional value of tropical and subtropicalfeeds. There is thus a need to compare and

study the different methods in a series ofanalytical, technical and methodologicalapproaches to identify the critical points.

Structure and chemical composition

The nutritive value of forage feeds for ruminantproduction is a function of two dependentfactors: voluntary intake and feed utilisation.Plant characteristics affecting voluntary intakeare palatability (taste, odour and surface

characteristics), leaf to stem ratio, plant habitand dimensions of stems and leaves (Stone,1994). Plant characteristics which affect feedutilisation depend upon both internal

(composition and structure) and external (plantanatomy and morphology) plant cell wallfactors (Chesson and Forsberg, 1988). Cellularconstituents, leaf and stem strength, anti-nutritional factors and toxins are additionalfactors affecting feed utilisation. Confined andgrazing animals have different feedingbehaviour and have distinct possibilities ofselection, both between and within feeds; thismeans that autoregulatory intake processescan be different.

Environmental conditions greatly affect theplant nutritive value. High growth temperaturesare an important factor in accelerating plantmaturity, but it is possibly that they also lead tochanges in plant anatomy. Differences in plantanatomy between tropical and temperategrasses are linked to the different types ofphotosynthetic pathway, C4 and C3, respec-tively (Wilson, 1994) and are measured asproportions and spatial arrangements oftissues within organs. These differences are

quite large in leaves, while in leaf sheath theyare smaller and are not expressed in stems.Temperate grass leaves have more mesophyllthan C4 grasses and this tissue appears tomore loosely arranged than that of C4 species,allowing easier penetration by microbes (Akin,1986). Tropical grass leaves have a higherproportion of thick walled tissues, such asvascular bundles and sclerenchyma thantemperate pasture grasses (Akin, 1989).Furthermore, in warm season grass leaves,vascular bundles are more closely spaced andhave a distinct, thick walled parenchymabundle sheathes (PBS) surrounding eachbundle, whereas cool season grasses havewidely spaced bundles and a less distinct PBS(Akin, 1986). In consequence, warm season

PBS, which contain more than 50 % of the leafreserve carbohydrates and protein, are slowlyor only partially digested, whereas that of coolseason grasses is rapidly and extensivelydigested. In some genera of C4 grasses, anadditional barrier (suberized lamella) oftenappears within the outer section of the bundlesheath wall, which is completely indigestible,but is not present in C3 grasses (Wilson, 1993).

Outer tangential cells of epidermis of bothtropical and temperate forages becomethickened, lignified and completely covered

with epicuticular waxes and cutin, whichrepresent an important physical barrier forpenetration of micro-organisms and gas andwater exchange. In most cultivated temperategrasses, the epidermis is attached to the bodyof the leaf through mesophyll cells and thusquickly removed by chewing and rumination. Incontrast, in tropical grass leaves and in manywild temperate grass leaves, the epidermis isattached to the leaf at vascular bundles

through thick-walled sclerenchyma cells and isnot readily lost by chewing and ruminating(Wilson, 1993). Furthermore, the walls oftropical grass epidermal cells in the paradermalview are sinuous in a form linking adjacentcells together with strong «dove-tailed» joints,whereas in most temperate grasses the wallsare straight-sided. The straight-sided walls areeasily separated along the middle lamellawhich is not the case with sinuous walls where

breakages must occur by splitting across wallsand not by separation at the middle lamella.

In legumes there appear to be no groupdifferences in anatomy and cell wallcharacteristics between tropical and temperatespecies (Wilson, 1993). Legume leaf

digestibilities are mostly high and the specieswith particularly low values have been largelylinked to high tannin levels rather thananatomical problems. Tannins, which arepresent in warm season dicotyledons in

appreciable amounts, can affect both feedintake (palatability because of their

astringency) and feed utilisation (formcomplexes with feed proteins, enzymes andpolysaccharides) (Kumar and Singh, 1984;Leinmueller et al, 1991 but there are alsoreports that small quantities of tannins (3-4 %of DM) show positive effects by preventingexcessive rumen degradation of dietary proteinand reducing the risk of bloat (Mangan, 1988;Leinmueller et al, 1991 ).

Decreasing digestibilities of thick-walledplant tissues are in close relation with theirdegree and the chemical nature of theirlignin/phenolic composition (Wilson, 1990). Allforages contain lignin, but the concentrationsare higher in legumes than grasses (Jung,1989) and slightly higher in C4 than C3 grasses(Akin and Chesson, 1989). Furthermore, grassstems contain higher quantities of lignin andphenolic acids than leaves (Lavrencic et al,1994, unpublished). Warm season grasses arealso characterised by the presence of greateramounts of esterified phenolic acids such as p-

coumaric and ferulic acid. Legumes containmuch smaller amounts of phenolic acids thangrasses. The total lignin concentrationincreases in forages with the physiologicalmaturity of stem and leaf tissue. High growthtemperatures have a greater influence onincreasing lignin concentration in cool seasongrasses than warm season grasses (Akin andChesson, 1989). The chemical nature of ligninseems to be another important factor limitingcell wall utilisation (Buxton and Russell, 1988;Reeves, 1985a, 1985b).

In addition, Kerley et al (1988) andChesson and Forsberg (1988) suggested thatthe fine structure of cell wall polysaccharides,such as cellulose crystallinity andhemicellulose side-chain substitutions mayhave an influence on the rate and extent ofmicrobial digestion. Furthermore, somelinkages between cell wall constituents such asphenolic ester and/or ether bridges betweenhemicelluloses and between hemicellulosesand lignin, diisotyrosine bridges between wallproteins and linkages between lignin and wallproteins may limit the access of microbialenzymes to the substrate (Jung, 1989; Jungand Deetz, 1993; Akin and Chesson, 1989).

General aspects of biologicaldeterminations

The biological methods were created to

represent and simulate a part or a series of

parts of the digestive tract and digestionprocesses in animals. We have considered the

digestibility with in vivo trials; the two stagetechnique (Tilley and Terry, 1963), one stagetechnique (Smith et al, 1971; Lindgren, 1979;Shahjahan et al, 1993), the in situ technique(Demarquilly and Chenost, 1969, and Chenostet al, 1970, modellized by Mehrez and 0rskov,1977; 0rskov and McDonald, 1979), in vivocannulated animals (Hveplund et a1,1976;Thomas, 1978), allantoin (Topps and Elliott,1965; Rys et al, 1975), in vitro continuousculture (Czerkawski and Breckenridge, 1977;Hoover et al, 1976a, 1976b) and in vitro gasproduction (Menke and Steingass, 1988).

The «basic model» which gives the valueutilised for defining the nutritive value of a feedis the in vivo digestibility; this represents theentire process which occurs in the

gastrointestinal tract. The other methodssimulate the rumen tract and in the case of

cannulated animals and the two stagetechnique, the gastric digestion is integratedwith the rumen digestion.

The methods consider and study a seriesof biological and physical processes whichdiffer considerably between themselves.

The basic value obtained in in vivo trials isthe result of the integration of two mainprocesses: digestion and absorption; the othersystems consider degradation (in situ),fermentation (gas test) and microbial synthesis(allantoin and cannulated animals) or digestion(two stage and one stage techniques). Theconsequence of these different approaches isthat the information provided is different in

type, unit of measurement and its distributionover time. The results can have a unique valueor/and be a series of estimated parametersaccording to the different approach considered;the methods which utilise models for theestimation of parameters as a function of timecan be considered to have an explicitlydynamic approach, while the methods whichgive a single final result, without a timevariable, can be considered explicitly static;however, these methods can give uniqueinformation which is the result of an implicitlydynamic process.

Digestibility trials, cannulated animals andallantoin give a single data point for thedigestibility process which is the result of abalance from several days of measurements ofa dynamic process.

The two stage technique, which gives onlythe end of point of digestion (Tilley and Terry,1963; Theodorou et al, 1994), is an implicitlystatic approach.

An evolution of the two stage technique(Smith et al, 1971; Shahjahan et al, 1993),without the second stage and with differenttimes of observations, has been developed toobtain the extensive kinetics of degradation oforganic matter and cell wall components andcan be considered explicitly dynamic but is stillimplicitly static. The in situ technique is used toobtain an estimate of the parameters of thekinetics of degradation with a application of aexponential equation (Orskov and McDonald,1979; McDonald, 1981; Robbinson et al, 1986)or to have a series of observations at precisetimes (Judkins et al, 1990). The sametechnique can be used for the continuousculture method in which the sample can beincubated for different times to obtain a kineticor an absolute value. The last two methods can

be considered to have an implicitly dynamicapproach.

Generally in the one stage, in situ andcontinuous culture techniques, a singlehomogeneous feed component is consideredto be present and to have a first order digestionkinetic. This approach can be due to the limitednumber of observations and the easy utilisationand interpretation of the estimated parameters.This represents too simple an approach,especially for feeds which have a highpercentage of cell wall, without constantdegradation kinetics (Van Milgen et al, 1993).In vivo methods can also be used to estimatethe transit time and bacterial synthesis,obtaining a series of kinetic parametersgenerally using a non constant rate of

passage, and with the utilisation of a

multicompartmental and age dependent model(Moore et al, 1992).

The gas test represents a system adaptedto yield continuous information (Menke andSteingass, 1988; Pell and Schofield, 1993;Theodorou et al, 1994) and the high numbersof observations enable the application ofmodels with the presence of several feed

components and a sigmoidal approach(Beuvink and Kogut, 1993, Schofield et al,1994).

The methods can be also divided in

«open» or «closed» systems; the level ofopenness represents the possibility of enteringand modifying the process over time; generally,in open systems, the input can be introduced atany time, can be modified and the finalmetabolites can leave the system withoutinterfering with the biological phenomena.

Coupled to this type of division, thesystems are divided into types of family: in

vivo, in situ and in vitro methods; a partlycontinuous culture, i.e. the in vitro systems, areclosed and the in situ and in vivo methods are

open.The other important distinction is the

nature of the wall of the system which can bebiological, for the in vivo and in situ methods,or inert for the in vitro methods.

Variables to be consideredIn warm climate zones, the present

animals species differ widely in terms of theirdigestive physiology and anatomy, theirphilogeny and their relative ecological niche;these differences lead to a series of specificcharacteristics : a) different feeding behaviour

b) a specific rumen environment c) kinetics ofrecycling of nitrogen d) different transit time.These differences are not only inter-speciesbut are, in some cases, intraspecific, as afunction of breed, different area of origin orfeeding system. Research has described thedifferences in digestibility for some tropicalfeeds across species (El Hag, 1976; Doyle etal, 1984; Domingue et al, 1991; Lechner-Doll etal, 1990; Kayouli et al, 1993) and betweenbreeds (Silanikove, 1986).

The methods have different levels of

complexity because of the numbers ofvariables considered; it is possible to considerfour variables and their relative effects: level of

intake, associative effects, recycling of nitrogenand transit time. These variable can be presentin the system under different forms: asconstitutional, implicit components, as

integrated components or as absent variable.The level of intake for in vivo methods is a

constitutional part of the system while in vitromethods are unable to consider these aspects.Several researchers consider that the two

stage technique is an unreliable estimate of invivo digestibility because the static approachthat does not consider the effects of intakelevel (Van Soest, 1982); this criticism can beextended to all the «closed» in vitro systems.The transit time, a variable closely associatedwith the level of intake, is an intrinsic variablein the in vivo methods and in continuousculture techniques it can be simulated and

integrated (Czerkawski and Breckenridge,1977; Hoover et al, 1976a, 1976b). In the in

situ, the passage rate is integrated in the

model used to estimate the effective

degradability (0rskov and McDonald, 1979);the value of the rate used in the in situ methodare constant and have generally beenmeasured for temperate feeds, without anyinvestigations on tropical animal species anddiets. In vitro closed systems lack the ratevariable and this represents a limit to thedynamic interpretation of the phenomena (Ellis,1978). Associative effects are considered inthe in vivo digestibility methods as an intrinsicaspect, and indeed the allantoin methodrepresents the final result of these phenomenaat the rumen level. Fistulated animals are usedto detect the associative effects in the reticulo-rumen (Archimede et al, 1995a), and in the insitu technique these effects can be studied(Flachowsky and Schneider, 1992; Archimedeet al, 1995b) but the results do not consider the

rumination process, passage time of feeds andthe phenomenon of compensation of digestion.The in vitro closed techniques are consideredinadequate to examine this effect (Mehrez etal, 1983), although there are possibilities ofusing the continuous culture method for suchinvestigations.

The recycling of nitrogen representsanother important factor for understanding thenutritional ecology and the adaptability ofspecies which eat roughages (Watson andNorton, 1982). The open systems take this intoaccount as an intrinsic factor; otherwise, theclosed in vitro methods cannot replicate thisaspect in dynamic terms and with continuousculture the NH3N represents an unstablefermentation characteristic (Mietinen andSetdld, 1989). In in vivo trials the differentbehaviour of nitrogen recycling studiedbetween species present in warm countriesshould be considered.

Factors affecting biological methods

The systems have different degrees of

sensitivity to a series of factors and havecritical points in which an apparently smallmodification in the conditions leads to a largemodification of the estimated parameters andthe overall result.

The initial conditions have different

meaning for the different techniques andshould generally represent the optimalconditions for gastrointestinal function (VanSoest, 1982) and normal feeding behaviour(Done-Curry et al, 1984). In the in vivo

methods, the initial trial conditions generallyhave a low weight because the final result isthe combination of the adaptation period andthe observation period. The other methods canbe influenced to variable degrees by the initialconditions - the more evident sensitivity withrespect to the in vivo digestibility trials is in

dependence for the different length of theperiod and times of observation, the modelapplied and the nature of the system; in in situtrials, the estimation of soluble fractions andthe degradation kinetics are very sensitive tothe initial conditions and to the firstobservation.

For the in vitro methods, the initialconditions are very important; the precisionand variability of the results depend on thecomposition and activity of the rumen liquid

(Aerts et al, 1977) and in turn these aredependent upon time of collection and type ofdiet (Judkins et al, 1990).

The basal diet represents a strong sourceof variability in the in situ and in vitro methodswhich need a source of rumen inoculum; theconcentrate to forage ratio and the level ofintake affect the rumen degradability oftemperate feedstuffs (Lindberg, 1981; Weaklyet al, 1983; Susmel et al, 1989) and tropicalfeeds (Zhao et al, 1993). In particular, anincrease in both these factors causes, for

tropical feeds, the degradability to decrease(Zhao et al, 1993). For the in vitro closed

methods, the type of diet represents a strongsource of modulation of the activity of rumeninoculum (Judkins et al, 1990; Susmel et al,unpublished).

The use of mixed diets in digestibility trialscan allow a series of phenomena to confusethe estimate of digestibility of the fibrous feed,by enhancing the activity of the amylolyticmicroflora through a change in rumen pH(0rskov, 1983), and/or by a competitionbetween the amylolytic and fibrolyticpopulations for the substrates rich in nonstructural carbohydrates (Tamminga, 1993). Inthese two cases, the fibre digestion kinetic is

depressed and modified leading to misleadingdigestibility value.

The end products represent anothersource of variability; in the in vitro closed and

open systems, they can be a disturbance factorfor the estimate, while in open systems theinfluence is limited. This is due to the

impossibility of the in vitro methods to absorbthe end products, causing a shift in themicrobial population (Mansifield et al, 1995).

Bacterial contamination can have differentlevels of influence according to the method andfeed utilised. For the in vivo methods, i.e.

digestibility trials and cannulated animals, theinfluence is lower than in situ and the one

stage and two stage in vitro techniques, ifthese avoid treatment with Neutral Detergentsolution, (Goering and Van Soest, 1970).Bacterial contamination can be represent astrong source of error, especially for poorquality feeds and for estimates of proteindegradability in situ (Varvikko, 1986). For thegas test and the allantoin technique, the effectis absent because of the nature of the estimateand the mechanism involved (fermentation andmicrobial synthesis, respectively).

The presence of antinutritional factors

(ANFs) have more significance for the closedin vitro systems because they represent asource of selection against the microbialpopulation leading to a depression of cellulosedegradation (Muinga et al, 1992). In the nylonbag technique, if a different basal diet to thefeed being tested is used, as the latter onlyrepresents a very small portion of the totalrumen contents, any adverse effects of theslow release of the ANFs on the activity of therumen microorganisms would be masked(Jones et al, 1992); the other techniques havelevel of sensitivity intermediate between the insitu and closed in vitro technique.

The methods have different sensitivity tothe amount, representativeness and physicalform of sample. Generally, the methods thatuse the mastication and rumination process inthe animal are less sensitive to the physicalform of the feed. The in vitro systems arehighly influenced by the amount and physicalform of the sample and the same phenomenainfluencing the in situ methods; the effect ofamount and physical form is the consequenceof the ratio between surface area available for

degradation and the amount of rumen liquid; inthe case of the in situ method, the ratio is

dependent upon the volume of the microhabitat in the bag and the pore size which maycause a selection of the microbial population.

For the in situ method, experimental resultsreport a general increase of degradability formore finely milled samples (Susmel et al,1990) with different behaviour for differentfeeds (Michalet-Doreau and Ould-Bah, 1992).The methods with a micro habitat have toconfront the problem of the need for a smallquantity of feed, while the in vivo digestibilityneeds a comparatively huge of quantity offorage, and this may be difficult to obtainespecially in arid area if there is a need forfresh feed.

The importance of the quantity and thephysical features of samples is smaller in themethods which have a macro habitat and witha natural turnover rate. The representativenessof a sample is particularly important for themethods in which it represents the only sourceof feed (e.g. for the microbes), and this is

especially the case for tropical feeds in whichthere are not only differences between leavesand stems, but also between different fractionsof the stems (Navaratne et al, 1990); for all

systems, for these forages is important to havea representative sample to what the animal hasactually consumed.

Level of prediction

Each system has different predictionobjectives: in vivo digestibility trials representthe basic value for digestibility and intake of drymatter; for voluntary intake. The digestibilitytrials are often conducted with mixed dietsbecause of the impossibility of having amaintenance level; in this case, the estimate ofthe digestibility of fibre-rich feeds, given in amixed diet, should not be considered anabsolute value but a value relative to the rationin which it is included (Pigden et al, 1980).

For tropical forages, in vitro techniquesgives different values of correlation withdifferent types of feed and the coefficientsobtained are difficult to compare; the influencesof chemical variables (NDF and protein) arealso different for the different classes of feeds

(Navaratne et al, 1990). However, the level ofprediction is always lower than temperatefeeds and is often not significant. Thecontinuous culture technique gives a resultsimilar to in vivo values but these systems uselow DM inputs and liquid dilution rates whichare less than in vivo estimates (Mansifield et al,1995).

The degradability of dry matter and NDFseems to be the most important variable fordescribing the nutritive value of some tropicalfeeds (Vadiveloo and Fadel, 1992) although forfeeds rich in tannins the in situ values do not

represent a good description of the truenutritive value (Makkar et al, 1989; Vadivelooand Fadel, 1992; Jones et al, 1992).

The cannulated animals method

represents a good technique for estimatingdigestibility, especially for feeds rich in

structural carbohydrates (Archimede et al,1995b) but this technique is less accurate thanthose involving the whole tract (classicaldigestibility trial).

The allantoin method does not give anyindications about digestibility, an estimate ofbacterial synthesis. The gas production methodand its associated mathematical models canbe used to give a prediction of energy andorganic matter fermentation of tropical feeds(Krishnamoorthy et al, 1995) and detect thepresence of different components in the feed

(Beuvink and Kogut, 1993), the effectivekinetics of degradation (France et al, 1993) andto identify the different stoichiometriccoefficients for different feed components fromthe gas production (Pell and Schofield, 1993;

Filacorda and Stefanon, 1995); this methodgenerally has a higher prediction level for haysthan two stage technique, but is slightly worsethan in situ (Khazaal et al, 1993a).

For dry matter intake, the levels of

predictability are good with the parametersobtained in situ (0rskov et al, 1988; Reid et al,1988) and from the gas test (Bliimmel and0rskov, 1993; Khazaal et al, 1993a) while theone and two stage techniques have a lowcorrelation with the observed intake for hay(Khazaal et al, 1993a) shrubs andMediterranean feeds (Susmel et al, 1993a).The continuous culture, allantoin andcannulated animal techniques are generally notused for estimating intake.

The protein value of a feed is estimatedwith a specific prediction system; the valueobtained from in vivo digestibility trials can onlyrepresent a source of information for theenergy allowance for the bacterial growth butgives no information on the protein allowance.The in situ (INRA, 1988; Sniffen et al, 1992;Susmel et al, 1993b) cannulated animal (Rohret al, 1986) and allantoin methods (Topps andElliott, 1966; Rys et al, 1975) give informationwhich is utilised for the estimate of a proteinvalue. For the cannulated animals, theestimation of microbial protein yield is achievedusing microbial markers, substances which arecapable of identifying the microbial cells andwhich have a constant ratio with the DM or

nitrogen; of the markers available, the mostfrequently used are ATP, RNA, DAPA andisotopes (15N, 35S) (Demeyer and Tamminga,1987; Satter et al, 1986; Susmel et al, 1993c).If the allantoin method is used to obtain anestimate of microbial production, there is aneed to consider a series of coefficients which

appear to be species dependent (Liang et al,1994) and this aspect needs of more

investigation. The data obtained also requirescorrecting for the endogenous contribution andrepresents a static value without information ondigestion kinetics. The results obtained with thecontinuous culture (Rusitec), have been shownnot to reproduce the rumen or in vivo

degradation patterns of carbohydrates andprotein (Czerkawski and Breckenridge, 1985;Jayasuriya et al, 1987; Susmel et al, 1991;C.R. Mills, personal communication) and givedifferent estimates for the protein value withrespect to the value obtained in vivo

(Mansifield et al, 1995). The one and two stagetechniques are not adapted for roughages and

are not sufficient alone, for supplements, togive good information on protein values(Sehgal and Makkar, 1994). In any case, someauthors (Antongiovanni et al, 1993; Stefanonand Guzzon, 1995) have attempted to use thedry matter disappearance, at the end ofdigestion to determine the amino acid profile ofthe feed and the entire contents of the Tilleyand Terry digestion tube. The gas test is not

currently used to describe the protein value butcould, in the future, be a potential source ofinformation of bacterial activity and growth (Pelland Schofield, 1993).

In the case of feeds treated or/and

supplemented with nitrogen, in vivo trials givethe reference information; several authors(Kartchner and Campbell, 1979; Holechek etal, 1986) have reported inaccuracies in theclosed in vitro technique for estimatingdigestibility of N-supplemented diets, soinaccuracies can be detected in situ if differentbasal diets are used (Makkar and Singh,1993). The continuous culture (Owen et al,1991 ), allantoin and cannulated animaltechniques can be used to study these effects.

In addition to the variables described

above, the methods can detect the presence ofcertain compounds, especially volatile fattyacids (VFA), tannins and phenols andsecondary metabolites.

VFA are routinely detected with thecontinuous culture (Czerkawski and

Breckenridge, 1977; Hoover et al, 1976a,1976b) and can be studied in the one stage,gas test (BlOmmel and 0rskov, 1993) andfistulated animal techniques.

The presence and the effect of tannins and

phenols can be verified with in vitro closed

methods; these methods (gas test) can detect,with the addition of Polyvinylpyrolidone(Khazaal et al, 1994) the presence ofantinutritional compounds but this aspectrequires further investigation; in the one stagetechnique, the kinetics of phenolic compounddigestion can be studied (Shahjahan et al,1993). For the in vivo and continuous culture

techniques, it is possible to study thesecompounds, while for the in situ method theyare difficult to detect (Khazaal et al, 1993b;Khazaal et al, 1994). The secondarymetabolites with an activity on the neuralmetabolism of the host can be studied with thein vivo trials and can only be detected with thein vivo technique.

The predictability of transit time is

associated with the use of exogenous markersutilised with fistulated animals and the in situ

technique. The flow of solid and liquid phasesfrom the rumen is estimated with markers ofthe particulate phase, such as chromium or therare earth elements (Yb, Ce, La,) mordanted tothe fibre, while the liquid phase is often markedwith polyethyleneglycol, cobalt or chromium inEDTA complexes (Demeyer and Tamminga,1987; Rohr et al, 1986; Satter et al, 1986).These markers are also used in the in vivo

digestibility to define the flow of solid and liquidalong the gastro-intestinal tract and in the

rumen, to detect the different kinetics of feedand between the animals species present intropical areas (Lechner-Doll et al, 1990;Kayouli et al, 1993).

Specific requirements of the methodsand recommendations

The methods have different experimentalhousing requirements and have differentdegrees of utilisation of animal. In vivo

digestibility trials need not less than threeentire animals to have a good repeatability ofthe results (Heaney, 1979): the time required isabout three weeks to one month of adaptationand one week of measurements and theamount of feed needed is large. With anexperimental design defined to cover the effectof time, the other in vivo methods needdifferent amounts of feed. In in vivo and in situ

trials, the level of intake should represent themaintenance level; this condition for poor orunpalatable feeds is difficult to obtain,although, at the same time, this probably doesnot represent the true feeding conditions in thetropics. Under these experimental conditions, itis important to use a type and percentage ofsupplement which does not give an abruptalteration of the «normal» rumen environment,but which enhances the optimal digestion ofdry matter and structural carbohydrates.

The use of cannulated animals shouldconsider the higher RSD and the need of asufficient number of animals (Archimede et al,1995a). For this method, from a practical pointof view, there are considerable difficulties and,for some aspects, criticisms, especiallyconcerning the accuracy of the evaluationobtained with microbial markers and flow of therumen phases. Considerable variations ofmicrobial flow are often reported to be due to

the type of marker, the sampling site andanalytical procedures (Demeyer and

Tamminga, 1987; Satter et al, 1986).The closed in vitro methods require the

presence of 2 animal donors and the trials canbe conducted from 48 to 96 hours with 3

replications and two sequential weeks ofanalysis; only small quantities of feed areneeded, together with forage standards, withknown digestibility, to allow the control ofvariation in the conditions. The rumen liquidcan be substituted with a pure culture obtainedin vitro or with the Rusitec liquid effluent (Owenet al, 1991); the values obtained, even forsubtropical roughages, are comparable tovalues measured with rumen liquid with aslightly negative systematic difference.

Most discussion is related to the time ofincubation (Holechek et al, 1986) and level ofsupplementation of N in the inoculum;however, timing and amount of supplementhave not been well established (Galyean et al,1987) and depend on forage, diet quality andbotanical composition, as well as accuracy ofworking. There is the need to preserveanaerobiosis and to have a sufficientconcentration of nitrogen in the medium tomaintain the activity of the bacterial population;this is particularly the case for feeds with a lowpercentage of nitrogen.

Gas production does not have a constantproportional coefficient with the degradedsubstrate, especially in the first few hours offermentation and after the 36th hour offermentation. This could be due to the growthand adaptation of the bacteria in the first hourof fermentation and the accumulation of toxicresidues in the later period (Filacorda andStefanon, 1995); for this reason, the kineticparameters should be utilised with caution.

The in vitro systems, used routinely, shouldnot require a large input of electricity but thereis a need for an uninterrupted supply (Owen etal, 1991). ).

The in situ method needs not less thanthree fistulated animals and 10 days ofadaptation and 140 hours of incubation; forfeeds with slow degradability, it should beconsidered that a change of one point of transitrate can considerably affect the results, asobserved by Kamatali et al (1992) andparticular attention should be given to the useand estimation of the lag phase (Kaitho et al,1993; Kamatali et al, 1992) and the use ofsoluble fractions (France et al, 1993; Dewhurstet al, 1995).

For the continuous culture technique, freshsheep faeces can be used as a source ofmicro-organisms for commencing the Rusitec(El Shaer et al, 1987) and the time needed isfrom 7 to 15 days.

Generally, in the systems withoutmastication by the animals, the definition of thedimension of milling, for different animalspecies, should be accounted for by differentsystems of mastication and specific particulatesizes present in the rumen.

More attention should be given to the useof the estimated parameters and care shouldbe exercised in applying predicting equationsfor the nutritive value of feeds: some authors

(Navaratne et al, 1990; Shem et al, 1995;Khazaal et al, 1993a) utilise the parametersprincipally on the basis of their statisticalrelevance (maximising the coefficient ofdetermination) rather than any biologicalmeaning. The consequence is that single ormultiple regressions obtained can be difficult tounderstand and compare from a biological andnutritional point of view.

Conclusions

The different biological methods try to

represent and simulate only one or a few sitesand mechanisms of feed utilisation, withdifferent levels of complexity. The informationobtained from them is often impossible tocompare and is often used not for the

biological meaning but simply for a statisticalcontribution to regressions with data obtainedfrom in vivo trials.

These methods do not give general andreliable predictions of the nutritive value offeeds from tropical and sub-tropical areasbecause of a series of factors: lack of

systematic investigation, incapability ofreproducing or considering the interactionbetween animal and plant, a static approachand/or unstable behaviour in response to

specific characteristics of feeds. The biologicalmethods can and should be used in

combination with investigations on possiblestrategies for feed utilisation. The methodsshould be integrated to obtain more completeinformation. The feed should be studied andassembled for its structural characteristicsbefore being tested with any given method.

Literature cited

Aerts JV, de Branbader DL, Cottyn BG, Buyesse FX(1977) Comparison of laboratory methods forpredicting the organic matter digestibility offorages. Anim Feed Sci Technol2, 237-249

Akin DE (1986) Interaction of ruminal bacteria andfungi with southern forages. J Anim Sci 63, 962-977

Akin DE (1989) Histological and physical factorsaffecting digestibility of forages. Agron J 81, 17-25

Akin DE, Chesson A (1989) Lignification as the majorfactor limiting forage feeding value especially inwarm conditions. In: XVI Inter Grassl Cong ,Nice, France, 1753-1760

Antongiovanni M, Stefanon B, Susmel P (1993)Estimation of nutritive value of protein forruminants with a simple laboratory technique. AttiSeminario ASPA-RAISA Roma, Italia, 127-133

Archimede H, Sauvant D, Hervieu J, Poncet C,Dorl6ans D (1995a) Digestive interactions in theruminants: relationships between whole tract andstomach evaluation. Anim Feed Sci Technol 54,327-340

Archimede H, Sauvant D, Dorl6ans D, Chapoutot P,Poncet C (1995b) Influence of the nature offorage and concentrate on the digestiveinteractions measured in sacco and in vivo. Anim

Feed Sci Techno154, 341-356

Beuvink JMW, Kogut J (1993) Modelling gasproduction kinetics of grass silages incubatedwith buffered ruminal fluid. J Anim Sci 71, 1041-1046

BlOmmel M, 0rskov ER (1993) Comparison of in vitrogas production and nylon bag degradability ofroughages in predicting feed intake in cattle.Anim Feed Sci Techno140, 109-119 g

Buxton DR, Russell JR (1988) Lignin constituents andcell wall degradability of grass and legumestems. Crop Sci 28, 553-558

Chenost M, Grenet E, Demarquilly C, Jarrige R(1970) The use of nylon bag technique for thestudy of forage digestion in rumen and for

predicting food value. Proc XI Intern GrasslCong, 7,697-701

Chesson A, Forsberg GW (1988) Polysaccharidedegradation by rumen microorganisms. In: Therumen microbial ecosystem (Hobson PN ed) 251-284. Elsevier, London, New York

Czerkawski JW, Breckenridge G (1977) Design anddevelopment of a long-term rumen simulationtechnique (Rusitec). BrJNutr38, 371-384

Czerkawski JW, Breckenridge G (1985) Metabolismof protein supplements studied by the rumen

simulation technique (Rusitec). Arch TierernahrBerlin 20, 261-278

Demarquilly C, Chenost M (1969) Etude de la

digestion des fourrages dans le rumen par lam6thode des sachets de nylon. Liaison avec lavaleur alimentaire. Ann Zootech 18, 4, 419-436

Demeyer DI, Tamminga S (1987) Microbial Proteinyield and its prediction. In: Feed evaluation and

protein requirement systems for ruminants.(R Jarrige, G Alderman, eds) Seminar EEC,Brussels, CEC, 129-141

Dewhurst RJ, Hepper D, Webster AJF (1995)Comparison of in sacco and in vitro techniquesfor estimating the rate and extent of rumenfermentation of a range of dietary ingredients.Anim Feed Sci Technol 5 211-229

Domingue BMF, Dellow DW, Barry TN (1991)Voluntary intake and rumen digestion of a lowquality roughages by goats and sheep. J AgricSci 117, 111-120

Done-Curry RJ, Hecker JF, Wodziska-TomaszewskaM (1984) Behaviour of sheep transferred frompasture to an animal house. Appl Anim BehavSci 12, 121-130

Doyle PT, Egan JK, Thalen AJ (1984) Intake digestionand sulfur retention in angora goats and merinosheep fed herbage diets. Austr J Agric AnimHusb 24, 165-169

El Shaer HM, Omed HM, Chamberlain AG, AxfordDFE (1987) Use of faecal organism from sheepfor the in vitro determination of digestibility. JAgric Sci 109, 257-259

El-Hag GA (1976) A comparative study betweendesert goat and sheep efficiency of feedutilisation. World Review of Anim Prod 13, 43-48

Ellis WD (1978) Determinants of grazed forage intakeand digestibility. J Dairy Sci 61, 1828-1840

Filacorda S, Stefanon B (1995) Description of foragefermentation characteristics with the gas testtechnique. Atti XI Congresso Nazionale ASPA,223-224

Flachowsky G, Schneider M (1992) Influence ofvarious straw-to-concentrate ratios on in sacco

dry matter degradability, feed intake andapparent digestibility in ruminants. Anim Feed SciTechno138, 199-217 7

France J, Dhanoa MS, Theodorou MK, Lister SJ,Davies DR, lsac D (1993) A model to interpretgas accumulation profiles associated with in vitrodegradation of ruminant feeds. J Theor Biol 163,99-111 I

Galyean ML, Krysl LJ, Estell RE (1987) Marker-basedapproaches for estimation of faecal output anddigestibility. In: Feed intake by Beef Cattle: Symp(Owens FN, ed) MP-121, Oklahoma Agric ExpSta, 121-131

Goering HK, Van Soest PJ (1970) Forage fiber

analyses (Apparatus, reagents, procedure andsome applications). Agric Handbook, ARS,USDA, Washington, DC, 379

Heaney DP (1979) Sheep as pilot animal in

Standardisation of analytical methodology forfeeds (Pigden WJ, Balch CC and Graham M,eds) IDRC-134e,44-48

Holechek JL, Wofford H, Arthun D, Galyean ML,Wallace JD (1986) Evaluation of total faecalcollection for measuring forage intake. J RangeManage 39, 2-4

Hoover WH, Crocker BA, Sniffen CJ (1976a) Effectsof differential solid-liquid removal rates onprotozoa numbers in continuous cultures ofrumen contents. J Anim Sci 43, 528-534

Hoover WH, Knowlton PH, Stern MD, Sniffen CJ

(1976b) Effects of differential solid-liquid removalrates on fermentation parameters in continuous

cultures of rumen contents. J Anim Sci 43, 535-542

Hveplund T, Moller PD, Madsen J, Hesselholt M(1976) Flow of digesta through the

gastrointestinal tract in the bovine with specialreference to nitrogen. In: Year-book , Royal VetAgr Univ, Copenhagen, 173-192

INRA (1988) Alimentation des bovins, ovins etcaprins. INRA Publ, Paris

Jayasuriya MCN, Hamilton R, Rogovic G (1987) Theuse of an artificial rumen to assess low qualityfibrous feeds. Biol Wastes 20, 241-250

Jones RJ, Lefeuvre RP, Palyne MJ (1992) Losses ofdry matter, nitrogen, minerals and fibre fractionfrom nylon bags containing Leucaena

leucocephala and two Calliandra species in the

rumen. Anim Feed Sci Technol37, 297-307

Judkins MB, Krysl LJ, Barton RK (1990) Estimatingdiet digestibility: a comparison of 11 techniquesacross six different diets to rams. J Anim Sci 68,1405-1415 5

Jung HG (1989) Forage lignins and their effects onfiber digestibility. Agron J 81, 33-38

Jung HG, Deetz (1993) Cell wall lignification anddegradability. In: Forage cell wall structure anddigestibility (Jung HG, Buxton DR, Hatfield RD,Ralph J, eds) American Society of Agronomy,Madison, 315-346

Kaitho RJ, Tamminga S, Van Bruchem J (1993)Rumen degradation and in vivo digestibility ofdried Calliandra calothyrsus leaves. Anim FeedSci Technol 43, 19-30

Kamatali P, Teller E, Vanbelle M, Collignon G, FoulonM (1992) In situ degradability of organic matter,crude protein and cell wall of various treeforages. Anim Prod 55, 29-34

Kartchner RJ, Campbell CM (1979) Intake anddigestibility of range forage consumed bylivestock. Montana Agric Exp Sta Bull 718 8

Kayouli C, Jouany JP, Demeyer Di, Ali-Ali Taoueb H,Dardillat C (1993) Comparative studies on thedegradation and mean retention time of solid andliquid phases in forestomachs of dromedariesand sheep fed on low-quality roughages fromTunisia. Anirn Feed Sci Techno140, 343-355

Kerley MS, Fahey GC, Gould JM, lannotti EL (1988)Effects of lignification, cellulose crystallinity andenzyme accessible space on the digestibility ofplant cell wall carbohydrates by the ruminant.Food Microstructure 7, 59-65 5

Khaazal K, Boza J, 0rskov ER (1994) Assessment ofphenolics-related antinutritive effects in

Mediterranean browse: a comparison betweenthe use of the in vitro gas production techniquewith or without insoluble polyvinylpyrrolidone ornylon bag. Anim Feed Sci Techno149, 133-149

Khaazal K, Dentinho MT, Ribeiro R, 0rskov ER(1993a) A comparison of gas production duringincubation with rumen contents in vitro and nylonbag degradability as predictors of the apparentdigestibility in vivo and voluntary intake of hays.Anim Prod 57, 105-112

Khaazal K, Markontonatos X, Nastis A, 0rskov ER(1993b) Change with maturity in fibre

composition and levels of extractable

polyphenols in Greek browse: effects on in vitro

gas production and in sacco dry matter

degradation. J Sci Food Agric 63 (2), 237-244

Krishnamoorthy U, Soller H, Steingass H, Menke KH(1995) Energy and protein evaluation of tropicalfeedstuffs for whole tract and ruminal digestionby chemical analyses and rumen inoculumstudies in vitro. Anim Feed Sci Technol 52, 177-188

Kumar R, Singh M (1984) Tannins: their adverse rolein ruminant nutrition. J Agric Food Chem 32, 447-453

Lechner-Doll M, Rutagwenda T, Schwartz HJ, ShultkaW, Engelhardt WV (1990) Seasonal changes ofingesta mean retention time and forestomachfluid volume in indigenous camel, cattle, sheepand goats grazing a thornbush savannah pasturein Kenya. J Agric Sci 115, 409-420

Leinmueller E, Steingass H, Menke KH (1991)Tannins in ruminant feedstuffs. Anim Res and

Dev 33, 9-62

Liang JB, Matsumoto M, Young BA (1994) Purinederivative excretion and ruminal microbial yield inMalaysian cattle and swamp buffalo. Anim FeedSci Technol47, 189-199

Lindberg JE (1981) The effect of basal diet on theruminal degradation of dry matter, nitrogenouscompounds and cell wall in nylon bags. Swed JAgric Res 11, 159-167

Lindgren E (1979) The nutritional value of roughagesdetermined in vitro and by laboratory method. In:

Rep 45, Department of Animal Nutrition, SwedishUniversity of Agricultural Sciences, Uppsala, 1-61

Makkar HPS, Singh B, Negi SS (1989) Relationship ofrumen degradability with microbial colonizationcell wall constituents and tannin levels in some

tree leaves. Anim Prod 49, 299-303

Makkar HPS, Singh B (1993) Effect of storage andurea addition on detannification and in sacco drymatter digestibility of mature oak (Ouercusincana) leaves. Anim Feed Sci Technol 41, 247-259

Mangan JL (1988) Nutritional effects of tannins in

animal feeds. Nutr Res Rev 1, 209-231

Mansifield HR, Edres MI, Stern MD (1995)Comparison of microbial fermentation in therumen of dairy cows and dual flow continuousculture. Anim Feed Sci Techno155, 57-66

McDonald ID (1981) A revisted model for estimatingof protein degradability in the rumen. J Agric Sci96, 691-693

Mehrez AZ, 0rskov ER (1977) A study of the artificialfibre bag technique for determining the

digestibility of feed in the rumen. J Agric Sci 88,645-650

Mehrez AZ, EI-Schinnawy MM, El-Ashry MA, EadHME (1983) Assessment of the associative effectof roughages and concentrates. J Anim Sci 57(Suppl 1 ), 452-453

Menke KH, Steingass H (1988) Estimation of theenergetic feed value obtained from chemicalanalysis and in vitro gas production using rumenfluid. Anim Res Dev 28, 209-221

Michalet-Doreau B, Ould-Bah MY (1992) In vitro andin sacco methods for the estimation of dietarynitrogen degradability in the rumen: a review.Anim Feed Sci Techno140, 57-86

Mietinen H, Setala J (1989) Design and developmentof a continuous culture system for studyingrumen fermentation. J Agric Sci Finland 61, 475-488

Moore JA, Pond KR, Poore MP, Goodwin TG (1992)Influence of Model and Marker on digesta kineticestimates for sheep. J Anim Sci70, 3528-3540

Muinga RW, Thorpe W, Topps JH (1992) Voluntaryfood intake, live weight change and lactationperformance of crossbred dairy cows given adlibitum Pennisetum purpureum (Napier grass var.Bana) supplemented with Leucaena forage in thelowland semi-humid tropics. Anim Prod 55, 331-337

Navaratne HVRG, Ibrahim MNM, Shiere JB (1990)Comparison of four techniques for predictingdigestibility of tropical feeds. Anim Feed SciTechnol29, 209-221

0rskov ER, McDonald IW (1979) The estimation ofprotein degradability in the rumen frommeasurements weighted according to rate ofpassage. J Agric Sci 88, 645-650

fdrskov ER (1983) Supplementation of low quality

roughages diet for optimal microbial and hostanimal nutrition. In: Utilisation of low qualityroughages with special reference to developingcountries Proc workshop on applied researchAlexandria (Egypt), 84-87

0rskov ER, Reid GW, Kay M (1988) Prediction ofintake by cattle from degradation characteristicsof roughages. Anim Prod 46, 29-34

Owen E, Jayasurya MCN, Hamilton R, Lalenta M(1991) Use of a long-term rumen simulationtechnique (Rusitec) to provide micro-organism forin vitro digestibility assays. J Agric Sci 116, 297-301

Pell AN, Schofield P (1993) Computerized monitoringof gas production to measure forage digestion invitro. J Dairy Sci 76, 1063-1073

Pigden WJ, Balch CC, Graham M (1980)Standardisation of analytical methodology forfeeds; Summary and recommendations. IDRC,134e, 7-14 4

Reeves JB (1985a) Lignin composition of chemicallytreated feeds as determined by nitrobenzeneoxidation and its relationship to digestibility. JDairy Sci 68, 1976-1983

Reeves JB (1985b) Lignin composition and in vitro

digestibility of feeds. J Anim Sci 60, 316-322

Reid GW, 0rskov ER, Kay M (1988) A note on theeffect of variety, type of straw and ammoniatreatment on growth rate of steers. Anim Prod 46,29-34

Robbinson PH, Fadel JG, Tamminga S (1986)Evaluation of mathematical models to describeneutral detergent residues in terms of its

susceptibility to degradation in the rumen. AnimFeed Sci Technol 15, 249-271

Rohr K, Lebzien P, Schafft H, Schultz E (1986)Prediction of duodenal flow of non-ammonia

nitrogen and amino acid nitrogen in dairy cows.Livest Prod Sci 14, 29-40

Rys R, Antoniewicz A, Maciejewicz J (1975) Allantoinin urine as an index of microbial protein in therumen. In: Tracer Studies on Non-proteinNitrogen for Ruminants, II, Intern Atomic EnergyAgency, Vienna, 95-98

Satter LD, Combs DK, Lopez-Guisa JM, Nelson WF(1986) Use of markers for measurement of feeddigestibility in ruminants. In: Nuclear and RelatedTechniques in Animal Production and HealthProc Symp, Vienna, Austria

Schofield P, Pitt RE, Pell AN (1994) Kinetics of fiberdigestion from in vitro gas production. J Anim Sci72, 2980-2991

Seghal JP, Makkar GS (1994) Protein evaluation in

ruminants in vitro, in sacco, in vivo proteindegradability and microbial efficiency of differentprotein supplements in growing buffalo calves.Anim Feed Sci Technol45, 149-165

Shahjahan M, Mosihuzzamam M, Mian AJ (1993) Invitro digestibility studies of some local and highyielding varieties of rice straw (Oryza sativa).Anim Feed Sci Technol42, 121-130

Shem MN, 0rskov ER, Kimambo AE (1995)Prediction of voluntary dry-matter intake,digestible dry-matter intake and growth rate ofcattle from the degradation characteristics oftropical foods. Anim Sci 60, 65-74

Silanikove N (1986) Interrelationships between feedquality, digestibility, feed consumption, andenergy requirements in desert (Beduin) andtemperate (Saanen) goats. J Dairy Sci 69, 2157-2162

Smith LW, Goering HK, Waldo DR, Gordon CH(1971) In vitro digestion rate of forage cell wallcomponents. J Dairy Sci 54, 71-76

Sniffen CJ, O’Connor JD, Van Soest PJ, Fox DG,Russel JB (1992) A net carbohydrate and proteinsystem for evaluating cattle diets: 11

Carbohydrate and protein availability. J Anim Sci70, 3562-3577

Stefanon B, Guzzon P (1995) Effect of substrate onsynthesis and quality of microbial protein in vitro.Atti XI Congresso Nazionale ASPA, 223-224

Stone BA (1994) Prospects for improving the nutritivevalue of temperate perennial pasture grasses. NZ J Agric Res 37, 349-363

Susmel P, Stefanon B, Piasentier E (1989) Effect offorage and concentrate intake level on rumendegradability of protein sources having differentin vitro rates of N solubilisation. Anim Feed Sci

Technol26, 231-249

Susmel P, Stefanon B, Mills CR, Piasentier E (1990)Impiego di modelli matematici diversi ed effettodella macinazione e della setacciatura sulla

degradabilit- in situ della sostanza secca e

dell’N. Zoot Nutr Anim 16, 157-166

Susmel P, Stefanon B, Mills CR (1991) Generalproblems in assessing the nutritive value ofMediterranean forages. In: Fourrages et sous-produits mediterraneens (JL Tisserand, G Alibes,eds) Opt M6diter S6rie A, 16, 17-23, CIHEAM,Saragoza

Susmel P, Mills CR, Spanghero M, Stefanon B(1993a) The prediction of nutritive value anddegradability of mediterranean forages by in vitrogas production. Seminario ASPA-RAISA, Roma,Italia, 135-142

Susmel P, Spanghero M, Mills CR, Stefanon B(1993b) A comparison of the French system andthe Italian revised proposal for ruminant feedprotein evaluation. Atti Seminario ASPA-RAISARoma, ltalia, 105-112

Susmel P, Plazzotta E, Mills CR, Stefanon B (1993c)Determination of RNA and ATP in the rumen

liquid of cows fed with diets differing in forage toconcentrate ratio. J Sci Food Agric 63, 39-45

Tamminga S (1993) Influence of feeding managementon ruminant fiber digestibility. In: Forage cell wallstructure and digestibility 571-602, ASA-CSSA-SSSA, Madison, USA, 571-602

Theodorou MK, Williams AB, Dhanoa MS, McCallanAB, France J (1994) A simple gas productionmethod using a pressure transducer to determinethe fermentation kinetics of ruminant feeds. Anim

Feed Sci Technol48, 185-197

Thomas PC (1978) Measurement of the flow ofdigesta in the post-ruminal gut. In: Ruminant

digestion and Feed evaluation (Osbourn DF,Beever DE, Thomson DJ, eds) Agr Res Council,London, England, 3.1-3.10 0

Tilley JMA, Terry RA (1963) A two stage technique forin vitro digestion of forage crops. J Br Grassl Soc18, 104-111

Topps JH, Elliott RC (1965) Relationship betweenconcentration of ruminal nucleic acids andexcretion of purine derivatives by sheep. Nature205, 498-499

Vadiveloo J, Fadel JG (1992) Compositional analysesand rumen degradability of selected tropicalfeeds. Anim Feed Sci Techno137, 265-279

Van Milgen J, Berger LL, Murphy NR (1993) Digestionkinetics of alfalfa and wheat straw assumingheterogeneity of the potentially digestiblefraction. J Anim Sci 71, 1917-1923

Van Soest PJ (1982) Nutritional ecology of the

ruminant. O&B Books, Corvallis, OR, 374 pVarvikko T (1986) Microbially corrected amino acid

composition of rumen-undegraded protein andamino acid degradability in the rumen of feedsenclosed in nylon bags. BrJNutr56, 131-140

Watson C, Norton BW (1982) The utilisation ofpangola grass hay by sheep and Angora Goats.Proc Austr Soc Animal Prod 14, 467-470

Weakly DC, Stern MD, Satter LD (1983) Factorsaffecting disappearance of feedstuffs from bagssuspended in the rumen. J Anim Sci 56, 493-498

Wilson JR (1990) Influence of plant anatomy ondigestion and fibre breakdown. In: Microbial and

plant opportunities to improve lignocelluloseutilisation by ruminants (Akin D, Ljungdahl LG,Wilson JR, Harris PJ, eds) Elsevier, New York,99-117 7

Wilson JR (1993) Organisation of forage plant tissues.in Forage cell wall structure and digestibility.(Jung HD, Buxton DR, Hatfield RD, Ralph J, eds)American Society of Agronomy, Madison, 1-32

Wilson JR (1994) Cell wall characteristics in relationto forage digestion by ruminants. J Agric Sci 122,173-182

Zhao JY, Shimojo M, Goto 1 (1993) The effects offeeding level and roughages/concentrate ratio onthe measurement of protein degradability of twotropical forages in the rumen of goats using nylonbag technique. Anim Feed Sci Technol4l,261-269