Citation
Comparison of methodologies for determination of content and digestibility of hemicellulose and cellulose in tropical grass hays

Material Information

Title:
Comparison of methodologies for determination of content and digestibility of hemicellulose and cellulose in tropical grass hays
Creator:
Russo, Sandra L., 1948- ( Dissertant )
Mott, Gerald O. ( Thesis advisor )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1981
Language:
English
Physical Description:
xi, 109 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Carbohydrates ( jstor )
Cell walls ( jstor )
Detergents ( jstor )
Digestion ( jstor )
Forage ( jstor )
Forage quality ( jstor )
In vitro fertilization ( jstor )
Lignin ( jstor )
Rumen ( jstor )
Trucks ( jstor )
Agronomy thesis Ph. D
Dissertations, Academic -- Agronomy -- UF
Grasses -- Tropics ( lcsh )
Grasses as feed
Hay as feed ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
Experiments were conducted to: 1) fractionate cell walls of tropical grasses into: holocellulose, cellulose and hemicellulose; 2) determine in vitro and in vivo digestibility of cell wall components; and 3) study the relationship between these components and quality of forages, as measured by feeding hays to sheep. One cultivar of Cynodon dactylon (Coastcross-1) , two experimental lines of Digitaria species (X46-2 and X124-4) and two cultivars of Paspalum notatum (Argentine and Paraguay) cut at four ages (2, 4, 6 and 8 weeks of regrowth) were used in the experiments. Samples of each hay were analyzed for organic matter (OM) , crude protein (CP) and neutral detergent fiber (NDF) . NDF fractions (cellulose, hemicellulose and lignin) were determined by two methods: 1) conventional Van Soest detergent analyses (CELLV, HEMV, PLIG respectively) : and 2) classical fractionation using NaC10_ and alkali (CELL, HC, LIG respectively). There were significant differences between fractions obtained by the two methods. HC remained the same or decreased slightly with maturity but HEMV declined. CELL and CELLV showed similar patterns, increasing up to 6 weeks, then decreasing. Samples were fermented in rumen fluid for 48 hr and digested in acid-pepsin for 48 hr or fermented in rumen fluid only for 48 hr and residual NDF was fractionated as above. In vitro digestibility of NDF and HC was increased by further digestion in acidpepsin while CELL digestibility showed minimal change. CELL digestibility was generally higher than HC digestibility. Digestion coefficients for each fraction were similar ±n vivo and in vitro (r^ = .79 to .92). NDF, CP and PLIG were the best single chemical analyses for prediction of intake and digestibility, but any in vitro analysis gave better estimates of in vivo digestibility. Two and three variable models improved prediction of organic matter intake (OMI) and digestible OMI. Hemicellulose obtained by different methods was not the same. Digestion of cell walls in acid-pepsin was due to breakdown of hemicellulose and possibly, dissolution of lignin-hemicellulose bonds. Further studies on chemical bonding, physical degradability and chemical structure of the cell wall appear necessary before being able to predict animal performance from laboratory analyses.
Thesis:
Thesis (Ph. D.)--University of Florida, 1981.
Bibliography:
Includes bibliographic references (leaves 99-107).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Sandra L. Russo.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
028126479 ( AlephBibNum )
07851783 ( OCLC )
ABS1501 ( NOTIS )

Downloads

This item has the following downloads:


Full Text










COMPARISON OF METHODOLOGIES FOR DETERMINATION
OF CONTENT AND DIGESTIBILITY OF HEMICELLULOSE
AND CELLULOSE IN TROPICAL GRASS HAYS

















BY

SANDRA L. RUSSO


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL
OF THE UNIVERSITY OF FLORIDA IN
PARTIAL FULFILLMENT OF THE REQUIREMENTS
FOR THE DEGREE OF DOCTOR OF PHILOSOPHY


UNIVERSITY OF FLORIDA


1981






















To Allen,

now its your turn...
















ACKNOWLEDGE EVENTS


The author would like to express her appreciation to Dr. G. 0.

Nott for serving as chair and academic advisor for the program of

study. The author is sincerely grateful to Dr. John E. Moore,

director of the research, for guidance and support, for always having

time and usually patience and for providing research facilities and

equipment. Appreciation is extended to Drs. Donald Graetz, Joseph

Conrad and 0. C. Ruelke, members of the supervisory committee, for

their time and advice. The statistical assistance of Dr. Ramon

Littell is gratefully acknowledged.

Appreciation is extended to Dr. Murray Gaskins for putting his

time and laboratory equipment at the author's disposal. Mr. John

Funk and Dr. John Moore provided invaluable assistance at the

Nutrition Lab in designing and setting up the experiments.

Thanks are due to Drs. Donald Graetz, Ray Varnell, G. O. Mott,

Ken Quesenberry, C. Y. Ward, David Buchanan, Art Hansen and Elon

Gilbert for providing financial assistance via a series of jobs,

assistantships and contracts.

The author appreciates the intellectual and non-intellectual

support of fellow graduate students and staff at the Nutrition Lab.

In particular, Carlos Chaves, Hari Hartadi and Steve Abrams laid the

groundwork for the dissertation.








The weekly Social, Agricultural and Food Scientists meetings

kept the mission of feeding the world's hungry through cooperation

as well as research at the front of the author's thoughts and is most

gratefully appreciated.

Finally, for the continued support and sometimes, the sacrifices,

of her husband, Allen Cook, the author is especially indebted.
















TABLE OF CONTENTS


PAGE


ACKNOWL E DGMETS . . . .

LIST OF TABLES . . . . .

LIST OF FIGURES . . . .

ABSTRACT . . . . . .

CHAPTER I. INTRODUCTION . .

CHAPTER II. LITERATURE REVIEW .


. . . . . . . iii

. . . . . . . v ii

. . . . . . . ix

. . . . . . . x

. . . . . . . 1


S . 3


Forage Quality . . . . . . . . .
Digestibility . . . . . . . . . .
Intake . . . . . . . . . . . .
Relationship Between Digestion and Intake . . . .
The Structure of Plant Cell Walls . . . . . .
Hemicellulose . . . . . . . . . .
Hemicellulose: Methodology . . . . . . .
Alkali extraction . . . . . . . .
Acid hydrolysis . . . . . . . . .
Hemicellulose: Structure . . . . . . .
Hemicellulose in Plants . . . . . . . .
Hemicellulose Digestion in the Ruminant . . . .
In vitro studies . . . . . . . . .
In vivo studies . . . . . . . .

CHAPTER III. CHEMICAL COMPOSITION OF TROPICAL GRASS HAYS.


Introduction . . . .
Experimental Procedures . .
Results and Discussion. . .
Neutral detergent fiber. .
Holocellulose. . . . .
Hemicellulose. . . . .
Cellulose. . . . . .
Lignin . . . . . .
Summary . . . . . .


CHAPTER IV. IN VITRO AND IN VIVO DIGESTIBILITY OF
TROPICAL GRASS HAYS . . . . . . . .









Introduction . . . . . .
Experimental Procedures . . .
Results and Discussion . . . .
In vitro digestibility . . .
In vivo digestibility . . .
Summary . . . . . .


. . . . . .52
. . . . . 54
55
55
59
62


CHAPTER V. PREDICTION OF FORAGE QUALITY FROM CHEMICAL AND
IN VITRO ANALYSES . . . . . . . . .


Introduction . . . . . .
Experimental Procedure . . . .
Results and Discussion . . . .
Intake and digestibility . .
Prediction of forage quality from
Conclusion . . . . . .
Summary . . . . . .


. . . . . .65
. . . . . 66
.67
. . . . . 67
laboratory analyses 68
75
. . . . . 76


CHAPTER VI. GENERAL DISCUSSION . . . . . ... .77

APPENDIX. . . . . . . . . ... . . . .83

LITERATURE CITED. . . . . . . . . ... . 99


BIOGRAPHICAL SKETCH . . . .. .


. . . . . .108













LIST OF TABLES


TABLE PAGE

1 Abbreviations used in the dissertation. . . . ... 32

2 Analysis of variance of chemical composition of
forty hays. . . . . . . . ... ..... 34

3 Analysis of variance of method differences. . . ... 35

4 Correlation coefficients (r) between variables
obtained by two methods . . . . . .... 36

5 Analysis of variance of in vitro data . . . ... .56

6 Comparison of in vitro digestion in rumen fluid and
in rumen fluid plus pepsin. . . . . . . ... 57

7 Digestion coefficients of chemical fractions in vitro
and in vivo . . . . . . . .. .. .. .. 59

8 Chemical composition of feces . . . . .... 61

9 Relationship (r2) between in vivo and in vitro
digestion coefficients for six hays . . . .... 63

10 Relationship (r2) between in vivo values and laboratory
analyses, obtained with thirty-nine different hays. . 69

11 Relationship (r2) between in vivo values and in vitro
analyses, obtained with thirty-nine different hays. 70

12 Coefficient of determination (R2), probability (P)
and standard error of the estimate (s vx) for the
repression of in vivo intake values oh several
independent variables . . . . . . . ... 72

13 Coefficient of determination (R2), probability (P)
and standard error of the estimate (s ) for the
regression of in vivo digestibility values on
several independent variables . . . . . .. 73








APPENDIX TABLE PAGE

14 Chemical composition of individual hays . . .. 83

15 In vitro digestion of individual hays . . . .. 85

16 Chemical composition and in vitro digestion of
individual hays . . . . . . . . . 87

17 In vivo intake and digestibility data . . . .. 89

18 Means of neutral detergent fiber by age and hay . 91

19 Means of holocellulose by age and hay . . . .. 92

20 Means of hemicellulose by age and hay . . . .. 93

21 Means of cellulose by age and hay . . . . .. 94

22 Means of lignin by age and hay. . . . . . 95

23 Analyses of variance of feces samples and in vivo
digestibility . . . . . . . .... . 96

24 Least squares means of in vivo intake per unit body
weight (g/kg body weight) . . . . . . .. 97

25 Least squares means of in vivo digestibility data . 98


viii














LIST OF FIGURES


FIGURE PAGE

1 Change in NDF content by weeks and hay. . . ... 38

2 Change in NDF(H) content by weeks and hay .... .39

3 Change in holocellulose content by weeks and hay. 41

4 Change in alkali-soluble hemicellulose by weeks
and hay . . . . . . . . .. . 43

5 Change in Van Soest hemicellulose by weeks and hay. 44

6 Change in alkali-insoluble cellulose by weeks and
hay . . . . . . . . . . . . 47

7 Change in Van Soest cellulose by weeks and hay. .. 48

8 Change in chlorite lignin by weeks and hay. ... 49

9 Change in permanganate lignin by weeks and hay. .. 50















Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy



COMPARISON OF METHODOLOGIES FOR DETERMINATION
OF CONTENT AND DIGESTIBILITY OF HEMICELLULOSE
AND CELLULOSE IN TROPICAL GRASS HAYS

By

SANDRA L. RUSSO

June 1981

Chairman: Dr. G. O. Mott
Major Department: Agronomy

Experiments were conducted to: 1) fractionate cell walls of

tropical grasses into: holocellulose, cellulose and hemicellulose;

2) determine in vitro and in vivo digestibility of cell wall

components; and 3) study the relationship between these components

and quality of forages, as measured by feeding hays to sheep.

One cultivar of Cynodon dactylon (Coastcross-l), two experimental

lines of Digitaria species (X46-2 and X124-4) and two cultivars

of Paspalum notatum (Argentine and Paraguay) cut at four ages (2,

4, 6 and 8 weeks of regrowth) were used in the experiments.

Samples of each hay were analyzed for organic matter (OM),

crude protein (CP) and neutral detergent fiber (NDF). NDF frac-

tions (cellulose, hemicellulose and lignin) were determined by

two methods: 1) conventional Van Soest detergent analyses







(CELLV, HEMV, PLIG respectively): and 2) classical fractionation

using NaCO.2 and alkali (CELL, HC, LIG respectively). There were

significant differences between fractions obtained by the two

methods. HC remained the same or decreased slightly with maturity

but HEI.V declined. CELL and CELLV showed similar patterns,

increasing up to 6 weeks, then decreasing.

Samples were fermented in rumen fluid for 48 hr and digested

in acid-pepsin for 48 hr or fermented in rumen fluid only for 48

hr and residual NDF was fractionated as above. In vitro digesti-

bility of NDF and HC was increased by further digestion in acid-

pepsin while CELL digestibility showed minimal change. CELL diges-

tibility was generally higher than HC digestibility. Digestion

coefficients for each fraction were similar in vivo and in vitro
2
(r = .79 to .92).

NDF, CP and PLIG were the best single chemical analyses for

prediction of intake and digestibility, but any in vitro analysis

gave better estimates of in vivo digestibility. Two and three

variable models improved prediction of organic matter intake (OMI)

and digestible OMI.

Hemicellulose obtained by different methods was not the same.

Digestion of cell walls in acid-pepsin was due to breakdown of

hemicellulose and possibly, dissolution of lignin-hemicellulose

bonds. Further studies on chemical bonding, physical degradabi-

lity and chemical structure of the cell wall appear necessary

before being able to predict animal performance from laboratory

analyses.














CHAPTER I
INTRODUCTION


The world food shortage is most critical in that portion of the

world known as the tropics, which is faced with rising populations

and diminishing food supplies. The limited availability of land

suitable for arable cropping has focused attention upon the grass-

lands which occupy fifty percent of the earth's land surface. These

lands are able to provide food for people by the use of ruminant

animals to convert pasture into meat and milk.

The structural carbohydrates in pastures and forages are an

important source of energy to ruminants and, in turn, to people.

The amount and proportions of structural carbohydrates in the plant

cell wall determine the degree to which the ruminant can utilize

this energy. Changes in amounts and types of cell wall constituents

are responsible for differences in nutritive value of forages.

The cell wall polysaccharides are comprised of fibers of organ-

ized, crystalline cellulose in a matrix of hemicelluloses which are

associated physically and chemically with lignin (Jones, 1976).

Lignin influences digestibility of forages and has been considered

the primary structural inhibitor of quality in tropical grasses

(Moore and Mott, 1973). Cellulose and hemicellulose fractions are

also responsible for variations in digestibility, since they contain

the largest proportion of the indigestible substances in the cell

wall (Van Soest, 1975).








The role, structure and composition of hemicellulose(s) vis

a vis forage quality and nutritive value is not well understood.

Hemicellulose is often defined as the resultant of a certain set of

operations and it has been convenient to estimate hemicellulose

as the difference between acid detergent fiber (ADF) and neutral

detergent fiber (NDF). Acid extraction may not, however, remove

material that is hemicellulosic in nature (Bailey and Ulyatt, 1970;

Morrison, 1980). The entire hemicellulosic fraction has been

treated as a unit since no nutritionally meaningful fractionation

has been found and since hemicellulose digestibility is so closely

related to that of cellulose (Van Soest, 1975). Questions arise,

however, as to whether this grouping together of structural carbo-

hydrates is biologically rational, given the complexity of living

plants and of ruminant digestion.

The objectives of this research were: 1) to fractionate the

cell walls of tropical grasses into three components: holocellulose,

cellulose and hemicellulose; 2) to determine the in vitro and in

vivo digestibility of the three components; and 3) to study the

relationship between these components and the quality of the

forages in vivo.














CHAPTER II
LITERATURE REVIEW


Forage Quality


The best measure of the nutritive value of a forage is animal

performance. Moore and Mott (1973) have defined forage quality as

output per animal, being a function of voluntary intake and digesti-

bility when the forage is fed ad libitum as the sole source of feed.

Intake and digestibility may be limited by the rate and extent of

forage digestion in the rumen, which, in turn, are related to the

amount of cell wall constituents (CWC).

Under the best of conditions, all forages would eventually be

tested in long-term production trials on pasture (Mott and Moore,

1970) having been preceded by a few years of agronomic and laboratory

evaluations. It is often not possible for researchers to conduct

pasture studies. Instead, trials with penned or caged animals are

used to measure voluntary intake and digestibility. Even these

animal trials may be beyond the means of some researchers who have to

rely exclusively on laboratory evaluation of forages. Fortunately,

enough data has been accumulated to show high correlations between

some laboratory methods and in vivo performance, e.g., the two stage

in vitro rumen digestion procedure (Tilley and Terry, 1963), and

digestibility. Unfortunately, a laboratory procedure that predicts

voluntary intake with acceptable accuracy has not yet been developed.








Thus, there is a continuum ranging from what would be the ideal way

to measure forage quality (animal response) to a vast array of ways

to predict forage quality in the laboratory which may or may not

be acceptable under all conditions.


Digestibility


A number of laboratory methods to estimate in vivo digestibility

have been devised. However, if only one method is to be used to

predict digestibility over a wide range of forages, then the method

to choose is the two-stage in vitro digestion procedure (Oh et al.,

1966; Marten and Barnes, 1980).

Digestibility is very important in determining the quality of

a forage and is more a function of plant factors than of animal

factors. It is now accepted that as a plant matures, its digestibi-

lity decreases due to increasing lignification and changes in plant

morphology. The cell contents (CC) of forages are considered to be

completely digestible while the CWC show variable digestibilities.

Riewe and Lippke (1970) found that tropical species accumulate dry

matter more rapidly which can be directly associated with an increase

in CWC. While temperate and tropical grasses have similar amounts of

CWC at early stages of growth, the amount of CWC and its composition

changes more rapidly with maturation in tropical species. This

increase in CWC may or may not decrease digestibility in tropical

grasses more than such an increase would in temperate grasses.

Minson and McLeod (1970) showed temperate species to be 12.8 percent-

age units higher in IVDIMD than tropical species; however, Russo







(1976) did not find differences of such magnitude despite higher

CWC. Tessema (1972) noted that the overall rate of decline in

digestibility was more rapid with tropical grasses than with temperate

grasses, that hemicellulose fluctuated considerably in the tropical

grasses and that tropical grasses had higher concentrations of

lignin. Moore and Mott (1973) have suggested that the apparent

lower quality of tropical grasses may be due to lower protein values,

which will depress intake, and higher fiber values, which will depress

digestibility.

With respect to ruminant nutrition, the detergent analysis

system of Van Soest (1967) which separates forages into CC and CWC

is useful for forage quality analysis. The CC are soluble and

almost totally available to the ruminant while the CWC are completely

insoluble, with their digestibility depending on microbial action.

The CWC appear to have the most influence on the digestibility and

intake of forages. The higher CWC in tropical grasses places in-

creased emphasis on digestibility, especially with respect to its

influence on voluntary intake.


Intake


Intake is controlled by two mechanisms: distention and meta-

bolic. In forage-fed ruminants, the distention mechanism exerts

the major effect on intake. Many other factors, of course, affect

voluntary intake and have been reviewed by Capote (1975) and Golding

(1973).








Crampton et al. (1960), Conrad (1966) and others have proposed

the following concepts relating to the distention theory in ruminants:

(1) probably a specific degree of rumen load reduction is the major

determinant of hunger: (2) the rates of forage cellulose and hemicell-

ulose degradation are directly related to the rate at which the rumen

load is reduced; and (3) the time it takes for the rumen load to be

reduced (when hunger recurs) is characteristic of the specific

forage involved. Most researchers now agree thac the rumen is

where the distention control over forage is exerted. Distention

mechanism is determined by (1) rumen fill and (2) retention time

(Weston, 1966: Thornton and Minson, 1972). Van Soest's "hotel

theory" (1975) suggests that the plant cell walls, not the dry

matter of the forage, are responsible for gut fill and only when the

cell walls are degraded or move out of the rumen, is gut fill re-

duced. Retention time is affected by the rates of digestion and

passage (Jones and Bailey, 1974; Waldo et al., 1972).


Relationship Between Digestion and Intake


The detergent systems of Van Soest separate the fibrous portion

of a forage into cellulose, hemicellulose and lignin. Other clas-

sification systems exist: Minson (1976) separated the CWC into a

potentially digestible fraction and an indigestible fraction, while

work by Abrams (1980) suggests there may be at least two types of

digestible cell walls. How these cell walls are utilized by an

animal i3 determined primarily by their rate of passage through

the digestive tract.








Chemical composition and structure of the cell walls are prob-

ably the most important factors influencing rates of digestion and

passage (Akin et al., 1974; Van Soest, 1965). The chemical bonds

holding the cell walls together as well as the structure of each

cell wall component (lignin, cellulose and hemicellulose) needs to

be elucidated. Lignin, for example, may provide a physical barrier

to the breakdown of the cell walls (Dekker, 1976). The rate of

breakdown is regulated by the organizational structure of the cell

walls (Akin et al., 1974: de la Torre, 1974).

Intake and digestibility can be seen then, to be linked to-

gether with respect to forage quality. Some alternative definitions

of forage quality use the combination of intake and digestibility,

such as voluntary intake of digestible energy (Heaney, 1970). The

goal is to predict digestible nutrient intake, i.e., animal perfor-

mance, without conducting long-term pasture production trials. Few

attempts have been successful, especially with tropical grasses. As

mentioned earlier, laboratory methods currently available can predict

forage digestibility but not voluntary intake. Characterization of

the chemical structure of the plant cell wall, especially in tropical

grasses, would help our understanding of the factors affecting volun-

tary intake.


The Structure of Plant Cell Walls


The plant cell wall is of two types, a thin primary wall and a

thicker secondary wall. The primary wall is laid down by undifferen-

tiated cells that are still growing. The primary wall is transformed









into a secondary wall after the cell has stopped growing. The

primary cell walls of many higher plants appear to have very similar

structures (Talmadge et al., 1973) but this is not true of secondary

walls.

Every living cell, whether plant or animal, is surrounded by a

complex membrane made up of lipids and proteins called the plasma-

lemma. In almost all plant cells, the plasmalemma forms a cell wall

which is rigid or semi-rigid and is chemically distinct in that the

structurally important components are polysaccharides (Preston, 1974).

Cellulose, hemicellulose, pectic polysaccharide, structural protein

and lignin have been identified as the major components of the plant

cell wall. Some of these components, e.g. lignin, are not poly-

saccharides, nevertheless, the essential and sudden change at the

plasmalemma surface is from a basically lipo-protein to a basically

polysaccharide formation (Preston, 1974).

Usually only one of these polysaccharides is crystalline and

cellulose predominates. In primary cell walls, the crystalline

cellulose fibrils form a core surrounded by a non-crystalline region

consisting of other cellulose molecules, non-cellulosic polysaccharides

and glycoprotein (Talmadge et al., 1973). In secondary walls, the

cellulose fibrils are aggregated into ropelike structures called

microfibrils.

The non-cellulosic polysaccharides making up the bulk of the

primary cell wall have been defined by their presence in chemical

fractions. The two major non-cellulosic fractions are the pectic

polysaccharides obtained by extracting cell walls with boiling








water, EDTA, or dilute acid (Aspinall, 1970) and the hemicelluloses

solubilized by the subsequent extraction of the same walls with

alkali (Whistler and Richards, 1970). In the secondary wall, also,

it appears that the extractable non-cellulosic wall compounds

surrounding the microfibrils are hemicellulosic (Preston, 1974).

But even further, some of the hemicelluloses cannot be removed

without degradation of the microfibrils and are therefore integrally

mixed with them (Preston, 1974). It has been suggested (Bauer et al.,

1973) that the structural function of the hemicellulosic polysaccha-

rides is to interconnect the cellulose fibrils and the pectic poly-

saccharides of the plant cell wall, and that this function is based

on the ability of the hemicellulosic polysaccharides to bind non-

covalently to cellulose and to bind covalently, through glycosidic

bonds at their reducing ends, to the pectic polysaccharides.


Hemicellulose


Hemicelluloses are often implicitly or explicitly defined as the

cell wall and intercellular polysaccharides that can be extracted by

alkali from higher land plant tissues that are, or were, lignified

(Wilkie, 1979). Hemicellulose is often extended to include certain

carbohydrates found in cereal endosperm. Jermyn (1955) wrote "it

cannot be too strongly emphasized that 'cellulose' and 'hemicellulose'

are normally determined as the resultants of certain sets of opera-

tions, rather than as chemically defined species." (p.197) Hemicellulose

has, thus, variable definitions depending on the methods or proce-

dures used to isolate it. A more liberal definition of hemicellulose








refers to all of the types of polysaccharides (found in plants)

other than celluloses, starches and fructans (Wilkie, 1979).

It is necessary to point out that the majority of literature

on hemicellulose and its relationship to forage quality and ruminant

nutrition deals with the Van Soest hemicellulose obtained by subtract-

ing ADF from NDF. Work as early as 1970 (Bailey and Ulvatt) points

out the problems inherent in this definition of hemicellulose. Dis-

cussion of methodology differences follows. To avoid confusion, Van

Soest hemicellulose will be distinguished from alkali-soluble

non-endospermic hemicellulose (as isolated by classical methodology).


Hemicellulose: Methodology


Hemicellulose was first isolated by Schultze in 1891 (Wilkie,

1979) using dilute alkali to extract polysaccharides from plant

tissues. Work on hemicellulose, thereafter, although becoming more

sophisticated, continued to use alkali to separate the compound

from plant tissue. The continued search for a simpler, more rapid

procedure led, however, to a second extraction scheme using acid

hydrolysis, either on delignified plant tissue (Routley and Sullivan,

1958) or on whole plant tissue (Van Soest, 1965). Alkali extraction

and acid hydrolysis are two considerably different methods of

arriving at a presumed similar result. Closer examination of these

two general methodologies appears warranted.

Alkali Extraction

In grasses, some of the hemicellulose is covalently bonded to

lignin (Morrison, 1974). Earlier, the relationship of lignin and







hemicellulose was not well understood and hemicelluloses were

extracted from lignified tissues of grasses by direct treatment

with alkali (Wilkie, 1979). However, it has now become common

practice to delignify the plant tissue with sodium chlorite and

acetic acid according to the procedure first outlined by Whistler

et al. in 1948. The delignified plant tissue, termed holocellulose,

is primarily composed of hemicellulose and cellulose. The holocellu-

lose is then treated with dilute alkali under nitrogen to extract

the hemicelluloses.

It may be possible to extract some of the hemicellulose from

holocellulose with hot or cold water. The alkali solutions, either

sodium or potassium hydroxide, may range in concentration from 4 to

24% w/v. Dilute alkali extracts only part of the hemicellulose.

If exhaustive extractions are required, the concentrations of alkali

can be increased from a low concentration to a higher concentration

to extract more of the hemicellulose (Wilkie, 1979). Nevertheless,

even with this exhaustive extraction, not all of the hemicellulose

will be accounted for (Gaillard, 1958).

With filtration or centrifugation, a residue and a supernatant

are obtained; the supernatant containing the hemicellulosic materials.

With some diligence, as many as three fractions can be derived from

the supernatant. Acidification of the supernatant with acetic acid

at 0-2 C will precipitate the A fraction (Gordon and Gaillard, 1976)

which is composed primarily of long chain xylans with small amounts

of arabinose and uronic acid (Bailey, 1973). The polysaccharides

remaining in solution can be precipitated by ethanol (Gordon and







Gaillard, 1976) or recovered by dialysis and freeze-drying (Bailey,

1973) to obtain the B fraction. The B fraction contains smaller

molecular weight xylans with many side chains and more complex mole-

cules made up of galactose, glucose and rhamnose (Bailey, 1973).

These are sometimes further broken down into linear B and branched

B fractions based on their separation by iodine precipitation

(Gaillard, 1965). The ethanol in this supernatant can be evaporated

and a C fraction recovered by dialysis for 5-7 days, followed by

freeze-drying. The C fraction has not been well studied but is

composed primarily of lignin, arabinose, galactose, glucose and

xylose (Gordon and Gaillard, 1976). In addition to iodine precipi-

tation, individual mono- or polysaccharides may be isolated from

these fractions by copper precipitation, gel filtration or quater-

nary ammonium salt precipitation (Bailey, 1973). The sugars thus

isolated can then be identified by any number of methods ranging

from the colorimetric (Dubois et al., 1956) to chromatographic

(paper, thin-layer, gas or high performance liquid).

It becomes apparent, then, that examination of hemicellulose

and its components can very rapidly become a major undertaking,

particularly when it is desired to elucidate exact amount or

structures of components. Work along these lines thus falls on

a few hardy souls upon whom the rest of the researchers become

dependent for answers. It is also apparent that such procedures

have little value in a situation where a large number of samples

need to be analyzed.








Acid Hydrolysis

A fair approximation of the hemicellulosic material in a plant

can be obtained by treating depectinated plant tissue with boiling

N acid for 2 3 hours (Bailey, 1973), hot water (Routley and Sullivan,

1958) or cold 72% sulfuric acid overnight followed by 2 hours

boiling (Sullivan, 1966). The sugars in the hydrolysate may then

be identified.

With the advent of Van Soest's detergent analysis scheme, it

became an easy matter to isolate and work with the cell wall and its

constituents. The use of neutral detergent on plant tissue leaves a

residue which is essentially nitrogen-free and is composed of lignin

and the structural polysaccharides, cellulose and hemicellulose. The

use of acid detergent on a forage was originally intended as a pre-

paratory procedure for lignin (Van Soest, 1975) but has commonly

been used to obtain cellulose and lignin values by using the acid

detergent lignin (ADL) procedure (Van Soest and Wine, 1968) and to

obtain a hemicellulose value by subtracting ADF from NDF.

The Van Soest procedures are widely used because they are rapid,

simple and not extremely expensive. The isolation of the cell walls

by neutral detergent is a chemically, as well as structurally, rational

procedure and is a convenient starting point for further work on cell

wall constituents. Daughtry (1976) for example, combined the use of

neutral detergent for cell wall preparation with boiling in 1 N

H2SO4 to obtain what was termed a hemicellulose hydrolysate.

It is with the other Van Soest procedures, ADF and ADL, that

problems arise. ADF minus lignin and silica, for example, is a close









estimate of cellulose (Van Soest and Wine, 1968) but has been

found to be contaminated with hemicellulosic residues (Bailey and

Ulyatt, 1970; Morrison, 1980). ADL, on the other hand, gives lignin

values quite different from lignin values obtained on the same plant

sample by other methods. Lignin, like hemicellulose, is difficult

to define and becomes defined by the procedure used to isolate it.

Many, and sometimes imaginative, adaptations have been made on the

Van Soest procedures in an attempt to get fractions that behave in a

uniform manner. However, since the cell wall fractions are not uni-

form, it is unlikely that a simple evaluation procedure will isolate

a recognizably uniform entity in all forages at all ages under all

environmental conditions.

Recalling that hemicellulose becomes defined by the procedure

used to isolate it, these two different approaches-- solubility in

alkali and hydrolysis in acid --can be seen to, in effect, act as a

check on each other. The former, by elucidating the chemistry and

structure of hemicellulose piece by piece, acts as a baseline

against which the results of the more rapid, latter methodology can

be checked for accuracy. It is with a very critical eye that infor-

mation on "hemicellulose" should be examined and compared because,

as has been seen, these hemicelluloses may or may not be representa-

tive of the hemicelluloses present in plants.


Hemicellulose: Structure


Hemicellulose was originally thought to be a precursor of cellu-

lose, hence the name hemi cellulose. While the reasoning has been








proven wrong, the name persists. Cellulose is a linear polymer of

8-1-4 linked D-glucose units; repeating unit 4-0-$-D-glucopyranosyl-D-

glucopyranose, cellobiose (Bailey, 1973). Hemicellulose is composed

of two major groups of polysaccharides: pentosans and non-cellulosic

hexosans.

Bailey (1973) reviewed the literature on chemical composition

of hemicellulose from which the following information is taken. The

pentosans are primarily based on xylose. They appear to be linear

chains of 8-1-4 linked D-xylopyranose units; repeating unit 4-0-8-D-

xylanopyranosyl-D xylopyranose, xylobiose. To these chains are

attached L-arabinose, D-glucuronic acid or its 4-0-methyl ether,

D-galactose and possibly D-glucose. These sugars are usually single

side chains but may be connecting points for further branching, par-

ticularly via L-arabinose. Here also, is where the linkages with

lignin polymers occur via ferulic acid, other phenolic acids or

their esters (Wilkie, 1979; Morrison, 1980).

The hexosans in forage grasses are primarily mannans, either

glucomannans or galactoglucomannans. These are linear chains of

8-1-4 linked D-glucose and D-mannose. The glucomannans are usually

free of other sugars, the galactoglucomannans have a-linked galactose

side chains. There has also been some evidence (Fraser and Wilkie,

1971; Wilkie and Woo, 1976) of the presence of linear chains of

s-linked D-glucose units which are not cellulosic. The glucans

have both 8-1-3 and 8-1-4 linkages. The ratio of the 1-3 to 1-4

linkages in the glucans in any one tissue falls as the plant matures

(Buchala et al., 1972). A more thorough discussion of the structural









features in hemicelluloses can be found in a review chapter by

Wilkie (1979).

Albersheim and coworkers (Bauer et al., 1973; Keegstra et al.,

1973; Talmadge et al., 1973) working with dicotyledonous plants

have developed a cell wall model. The cellulose fibers are linked

together by other polysaccharides, in particular, hemicellulosic

xyloglucans which completely cover the cellulose fibrils (Albersheim,

1978). Their preliminary work on grasses suggests that a model of

the cell wall for monocotyledonous plants will be similar to that

for dicots. The dominant hemicellulosic polysaccharide in primary

cell walls of monocots is an arabinoxylan rather than a xyloglucan

(Albersheim, 1978). A modified cell wall model (Bailey et al., 1976)

proposes that much of the hemicellulose and all of the pectin is not

linked to other CWC and only some of the hemicellulose is bonded to

glycoprotein and cellulose. They also point out that cellulose is

not the only crystalline structure and that hemicellulosic xylan is

capable of forming a crystalline structure.


Hemicellulose in Plants


Hemicelluloses comprise from 10 30% of the total dry matter

in temperate and tropical grasses (Bailey, 1973) and from 30 60%

of the cell wall in tropical grasses (Hartadi, 1980; Bailey and

Connor, 1973; Russo et al., 1981). While it is difficult to compare

hemicellulose values reported by various researchers due to differ-

ences in definitions and procedures, it appears that grasses contain

more hemicellulose than legumes (Bailey, 1973). Further, the








hemicellulose in grasses is different structurally from legumes.

Gaillard (1965) reported that the hemicellulose A fraction of

grasses was primarily an arabinoxylan while in legumes it was a

xylan with some uronic acids. Legumes seem to have more uronic

acids than grasses and some additional sugars such as rhamnose which

are not present in grasses (Gaillard, 1965; Collings and Yokoyama,

1979). Lignin has long been noted to be higher in legumes while

seemingly having less effect on the digestibility of legumes than it

does on grasses. Gordon and Gaillard (1976) suggest that this

difference could be due to the different types of hemicellulose

present in grasses and legumes and the subsequent differences in

lignin-carbohydrate bonds.

The hemicellulose content of a plant changes with growth and

maturity. The most marked change in hemicellulose composition in a

grass is in the growth period before changing from a vegetative to a

floral morphology (Bailey and Connor, 1973). It is expected that

hemicellulose content will increase with maturation. Bailey (1973)

reported on several studies with temperate grasses where the increase

in hemicellulose with age was associated with in increase in stem

tissue. For tropical grasses, however, seasonal changes did not

appear so great, the hemicellulose content remaining about the same

in spite of increasing cell wall content and increasing proportion

of stems.

The greatest change in hemicellulose content occurs during

growth. The leaves of speargrass (Heteropogon contortus) at vegeta-

tive, early seed and dormant stages had 23, 34 and 36% hemicellulose








(of the total dry matter); the stem in early-seed and dormant

stages had 37 and 42% hemicellulose (Blake and Richards, 1971).

Reid and Wilkie (1969a) found that in older, mature plants, there was

little apparent change in the hemicellulose of either leaves or stems.

Mature stems of rye, barley, orchardgrass, oats, ryegrass, timothy

and wheat had values of 23 28% for hemicellulose (Wilkie, 1979)

while mature leaves of oats, barley, wheat, kikuyugrass and pampas-

grass had values of 21 37% for hemicellulose.

Monosaccharide analysis (Buchala and Wilkie, 1973) of the hemi-

cellulose fraction found that with increasing maturity, there was an

increase in amount of D-xylose in all tissues (stems and leaves),

L-arabinose and D-glucose decreased and D-galactose remained constant

or varied slightly.

The hemicellulose changes with growth may be linked to formation

of the secondary wall. When a plant cell wall stops elongating, it

produces a thicker secondary wall and at this stage, lignin begins

to be synthesized into the wall along with continuing synthesis of

structural carbohydrates found in the primary wall (Morrison, 1979).

It is at this stage that lignin and hemicellulose become linked to

each other. With respect to cellulose, its structure does not change

during growth. Hemicellulose, on the other hand, does change its

structure during growth and with increasing lignification (Morrison,

1979). These changes have sometimes unexpected effects on digesti-

bility, which will be discussed in a later section.

Bailey (1973) reported that in temperate grasses the cellulose

content is greater than or equal to the hemicellulose content but,








in tropical grasses, the hemicellulose content is greater than or

equal to the cellulose content. A further difference between tem-

perate and tropical grasses is in monosaccharides present in the

hemicellulose fraction. Ojima and Isawa (1968) showed that two

tropical grasses (Paspalum notatum and Cynodon dactylon) had xylose:

arabinose:glucose:galactose ratios of 50:15:30:2 while two temperate

grasses (Phleum pratense and Lolium perenne) had ratios of the same

sugars of 74:12:11:2.

It is difficult to separate out environmental effects. The

possibility exists that the differences in CWC of temperate and

tropical grasses may be due to a temperature effect (Cooper and

Tainton, 1968). Anatomical and photosynthetic differences do exist.

Wilson and Ford (1971) studied three grasses (Panicum maximum,

Setaria sphacelata and Lolium perenne) under a controlled environ-

ment and found that the effect of increasing temperature on CWC is

not linear but passes through a maximum; this maximum may be due to

hemicellulose rather than cellulose differences (Bailey, 1973).

Reid and Wilkie (1969b) found little difference in monosaccharide

composition of hemicelluloses of oat plants grown indoors or out,

in light or in dark. Hemicellulose content and composition are not

constant and are apparently influenced by environmental conditions.

Almost every study on hemicellulose contains values for cultivars,

species and genera and many attempts have been made to relate hemi-

cellulose values to intraspecific or intrageneric variations. The

results can be very confusing and probably of little value. As

Wilkie pointed out (1979), "it is not possible to determine invariant,








quantitative values for total hemicelluloses . and there are

no self-evident parameters upon which to base such quantitative

comparisons, particularly at the taxa level." (p. 251) Fraser and

Wilkie (1971) re-emphasized the danger of comparative studies, che

danger arising when the procedure used in isolating materials for

study is either undesirable selective or inadequately reproducible.

With so many hemicellulose procedures in use, it appears that compa-

rative studies should be limited to within locations or laboratories.


Hemicellulose Digestion in the Ruminant


Ruminants obtain a great portion of their dietary energy from

the digestion of cell wall polysaccharides when they are fed forage

diets. The digestibility of the CWC is very crucial in determining

the quality of the forage. Intake, coo, is dependent on the break-

down of the CWC in the rumen. This digestion involves initially

enzymatic hydrolysis by various rumen microorganisms to release

sugars. These sugars can then be fermented to volatile fatty acids.

It is at the initial stages of hydrolysis of the various cell wall

polysaccharides that differences occur and when such properties as

degree of lignification (Bailey et al., 1976) or chemical structure

of the polysaccharides affect the digestion of the CWC.

In vitro Studies

Hemicellulose and its component monosaccharides can be separated

from plant tissue by chemical methods. Earlier in vitro work on hemi-

cellulose degradation in the rumen used these isolated fractions to

determine digestibility of hemicellulose. Gaillard (1965) found








that legume hemicellulose was more resistant to digestion than grass

hemicellulose; of the linear A and B fractions, A was more resistant;

and of the B fractions, the branched B fraction was more resistant

to digestion than the linear B. Sullivan (1966) also showed that

grass and legume hemicelluloses differed in their digestibilities

and suggested that it may be due to a difference in the chemical

heterogeneity of the xylans. Waite and others (1964) found that

during growth in timothy, ryegrass and orchardgrass, the xylan frac-

tion containing the uronic acids became less digestible. Dekker and

Richards (1973) found the glycans in hemicellulose to be more diges-

tible than xylan or uronic acid.

Gaillard (1965), Waite et al., (1964) and Bailey (1967) working

with temperate grasses, found the pectins, arabinans and galactans to

be extensively digested in young plants but incompletely digested in

older plants or in hay. They also concluded that xylans are incom-

pletely digested in mature plants and hay.

Other work has shown that rather than the chemical heterogeneity

of the hemicellulose being responsible for lower digestion, that the

incomplete digestion was due to the presence of lignin. Beveridge

and Richards (1975) studying speargrass, showed no significant

difference between the hemicellulose composition of fresh samples

and of samples digested for 72 hrs in the rumen. In their earlier

work (Dekker, Richards and Playne, 1972) the hypothesis had been

that branching would cause resistance to digestion but since there

were no differences between the original and "resistant" hemicellu-

loses in the later study, they attributed the poorer digestibility








to lignification. Bailey and MacRae (1970) isolated "digestion-

resistant" hemicelluloses which were capable of further digestion

after they had been delignified with NaCIO. Dehoritv et al. (1962)

delignified timothy, alfalfa and orchardgrass by ball-milling and

found that hemicellulose fermentation increased. The increases

became greater with maturity and, again, they offered the theory

that lignin provided a physical barrier preventing degradation of

the hemicellulose molecules.

In studies using isolated strains or species of rumen hemicellu-

lolytic bacteria, Dehoritv (1973) found that fermentation of hemicell-

ulose occurred in two stages. The first stage involved one species

of bacteria which degraded or depolymerized the hemicellulose but

did not utilize it. In the second stage, another type of bacteria

which had limited degrading abilities was able to utilize hemicell-

ulose that had been degraded by the first group. Dehority (1973)

also found that cellulolytic bacteria could degrade hemicellulose.

These synergistic systems may be oversimplifications but the apparent

specificity of the rumen bacteria provides another explanation for

variabilities in hemicellulose digestion.

Another avenue of in vitro studies is the use of depectinated

cell walls and mixtures of mold cellulase and hemicellulase to

study the patterns of hemicellulose digestion. With the enzyme

systems, hemicellulose hydrolysis ,as shown to differ between rye-

grass varieties and the extent of hydrolysis decreased with

maturity (Bailey and Jones, 1971; 1973). Jones and Bailey (1972)

found that prolonged drying did not affect the rate or extent of

enzymatic hydrolysis.







Studies on isolated hemicelluloses, its components or enzymes

from isolated rumen microorganisms give only a limited idea of what

is going on in the rumen. More recent research has been on total

plant hemicellulose and in vivo studies of hemicellulose digestion

in the rumen.

In vivo Studies

Beever et al. (1971) studied the sites of carbohydrate digestion

by fitting sheep with rumen cannulas and re-entrant cannulas in the

proximal duodenum and terminal ileum. Their results showed that 90%

of the cellulose and about 70% of the hemicellulose (in ryegrass)

was digested before entering the small intestine. Very little hemi-

cellulose digestion occurred in the small intestine itself but

approximately 30% of the hemicellulose was digested in the cecum

and colon. These values are averages for fresh, dried and ensiled

ryegrass. Other experiments with sheep also showed a significant

post-ruminal digestion of hemicellulose especially when the forage

was ground and pelleted (Thomson et al., 1972).

Van't Klooster and Gaillard (1976) showed that distal to the

duodenum in dairy cows hardly any digestion of hemicellulose

occurred on diets of hay, grass and/or concentrates. Digestion of

cellulose was almost completed before the small intestine.

In a review of several digestion studies, Ulyatt et al. (1975)

found that the proportion of hemicellulose digested in the large

intestine was higher than that of cellulose. An analysis of these

and other results (Bailey et al., 1976) showed that in forages

as the hemicellulose increases, the percent digested in the rumen









increases. The differences noted in all of these studies in the

contribution of the lower gastrointestinal tract to digestion of

hemicellulose may be due to species differences between cattle and

sheep, to the type of feed used or to differences in techniques.

Overall, the in vivo studies show that hemicellulose digestion

decreases with maturity and increasing lignification and that hemi-

cellulose is less digestible than cellulose. Models for cell wall

carbohydrates and their fate in the ruminant animal (Smith et al.,

1971; McLeod and Minson, 1974; Abrams, 1980) suggest the existence

of at least one digestible cell wall fraction and one indigestible

fraction, generally accounted for by the presence or absence of

lignin. Morrison (1979) discusses the physical theory vs. the chem-

ical theory. The physical theory proposes that the lignin and hemi-

cellulose form a complex which has a "cage" effect; in young tissue,

the complex is not fully developed and digestion is easily accom-

plished by the rumen microbes but in older tissue, the "bars" of

the "cage" are closer together and enzyme access is restricted.

The chemical theory suggests that in young tissue, lignification is

at a minimum and hemicelluloses are not very complex; with maturity,

complexity of the structure increases. Morrison concluded that

both theories may be involved; i.e., in youn: grass, the proportion

of side chains is the major factor controlling digestion while with

older tissue the lignin content is the dominant factor.

There are still many gaps in our knowledge of hemicellulose in

the plant and its fate in the ruminant, especially with respect to

the tropical grasses. The literature on hemicellulose digestibility









is often contradictory but there is the possibility that significant

hemicellulose digestion occurs after the rumen and that animal

performance is related to the digestibility of hemicellulose.

Because of the high hemicellulose content of tropical grasses, the

acid-soluble hemicellulose from the Van Soest method may not be an

acceptable estimate of the hemicellulose in these plants. An alter-

nate method of measuring hemicellulose is by alkali extraction from

plant cell walls. The hemicellulose obtained in this manner may be

more representative of the hemicellulose in the plant than the hemi-

cellulose obtained by the Van Soest analysis. Alkali soluble hemi-

cellulose could be a major factor influencing forage quality in

tropical grasses. More information on the hemicellulose fraction

of tropical grasses may lead to a better understanding of the

factors influencing quality. This dissertation investigates cell

wall content and digestibility in tropical grasses, with special

emphasis on hemicellulose.















CHAPTER III
CHEMICAL COMPOSITION OF TROPICAL GRASS HAYS


Introduction


TWhile the present official method of forage analysis is the

proximate analysis system first developed in the late 1800's, the

Van Soest (1965) system which partitions forages into cell contents

(CC) and cell wall constituents (CWC) is now extensively used for

evaluation of forages. Use of neutral detergent reagent on a forage

leaves a residue which is essentially nitrogen-free and is composed

of the structural components of the plant, the CWC (Van Soest, 1965;

Van Soest and Wine, 1968). The CC, which were removed by the

neutral detergent reagent, contain the sugars, starches, lipids

and proteins. The CC are essentially digestible by the ruminant

while the CWC have incomplete and differential digestibilities

(Van Soest, 1975; Abrams, 1980). The Van Soest system of forage

analysis is biologically rational and it has been recommended for

grading hays by the American Forage and Grassland Council (AFGC)

(Rohweder et al., 1976; Barnes and Marten, 1979).

Nevertheless, in spite of the widespread acceptance and use of

the Van Soest system, problems still remain, particularly with res-

pect to tropical and subtropical grasses (Moore and rMott, 1973). Due

to physiological and photosynthetic differences and environmental

influences, tropical grasses are higher in CWC than temperate








grasses (Cooper and Tainton, 1968; Wilson and Ford, 1971; Moore and

Mott, 1973). Prediction equations such as proposed by the AFGC rank

the tropical grasses very low due to their high CWC, yet animal per-

formance does not necessarily reflect the predicted low quality of

these grasses.

Due to these anomalies, questions arose as to whether the Van

Soest method, which had been developed on temperate forages, was

suitable for analysis of tropical grasses. While the separation of

CC and CWC appears rational (Van Soest, 1975), the further fraction-

ation of the CWC using acid detergent reagent to obtain cellulose,

hemicellulose and lignin, warrants further examination. The break-

down of the CWC by the acid detergent reagents appears incongruent

with earlier, classical work on hemicellulose, for example, which

used alkaline conditions to fractionate the CWC. Resolution of the

differences between the Van Soest hemicellulose and hemicellulose

obtained via classical methodology may lead to a clearer under-

standing of the factors affecting forage quality in tropical grasses.

The objective of this study was to compare the cell wall frac-

tions of tropical grass hays obtained by the Van Soest analysis with

cell wall fractions obtained by classical methodology, involving

an alkali extraction procedure.


Experimental Procedures


One cultivar of Cynodon dactylon ('Coastcross-l' bermudagrass),

two experimental lines of Digitaria species (X124-4 and X46-2) and

two cultivars of Paspalum notatum ('Paraguay' and 'Argentine'








bahiagrass) grown in an upland sandy soil in north-central Florida

during the spring-summer were used in this experiment. At the

beginning of the growing season a complete fertilizer was applied to

provide 72 kg N, 72 kg P205 and 72 kg K20/ha. Another nitrogen fer-

tilization at a rate of 66 kg N/ha was applied during July. All

plots were staged during July and August by mowing and removing the

growth; thereafter, two field replications of the 2, 4, 6 and 8 week

regrowths were harvested and artificially dried in a portable wagon

at approximately 50 C. The dry hays were stored in woven bags and

chopped prior to feeding.

Voluntary intake and digestibility were determined simultaneously

by feeding the grasses ad libitum to mature wethers. Each hay was

fed to at least cwo (and often three) wethers in separate trials,

with no wether receiving the same hay twice. Animals were treated

for internal parasites, fed a standard hay and weighed before each

trial. The seven day collection period was preceded by a fourteen

day preliminary period. Each animal was housed in a wooden pen with

slotted floors and had access to water, trace mineralized salt and

deflourinated phosphate at all times. Ad libitum feeding was

achieved by allowing a feed refusal (orts) of 200-300 g of hay

daily. Daily samples were taken of the hay offered during the

trial. The orts were removed and weighed daily. Feces were collected

in canvas bags, weighed daily and 20% kept for further analyses.

The waste (hay dropped through the floor of the cages) was weighed.

All samples were dried at 50 C. Each sample was pooled by animal,

ground to pass a 4 mm screen in a Wiley mill, mixed and a








representative portion reground through a 1 mm screen. A composite

hay sample was made by mixing proportional amounts from each animal

fed that specific hay. Animals were weighed at the end of the

trial. The composite hay samples were used for the laboratory

analyses in this experiment.

The hay samples were analyzed for dry matter (DM) and organic

matter (OM) (A.O.A.C., 1975). Neutral detergent fiber (NDF) was

determined as outlined by Goering and Van Soest (1970) using

sintered glass crucibles for filtration. NDF residues and all

subsequent residues were dried at 55 C in a forced air oven over-

night and at 55 C in a vacuum oven for 1-3 hrs.

Hartadi (1980) determined NDF (using glass wool for filtration)

termed NDF(H), acid detergent fiber (ADF), permanganate lignin (PLIG),

hemicellulose (HEMV), cellulose (CELLV) and crude protein (CP)

(A.O.A.C., 1975; Goering and Van Soest, 1970).

The NDF residues were delignified with sodium chlorite (NaCIO2)

according to a procedure adapted from Whistler et al. (1948). An

Eberbach heated shaking water bath set at 70 C was fitted with

clamps to hold six 500 ml Erlenmeyer flasks and was placed in an

explosion proof hood. Rubber tubing, attached to a manifold, was

directed above each flask to deliver CO2 continuously in order to

displace C102 formed during the reaction. NDF residues were placed

in the flasks and NaCIO2 equivalent to the weight of the sample and

100 ml of 1% acetic acid were added to each flask. The weight of

NDF residue transferred was determined by weighing the sintered

glass crucible before and after the transfer. Flasks were placed in







the water bath and shaken for 30 minutes. Flasks were removed from

the bath in order to add an additional .5 gm of NaCIO2. The flasks

were returned to the bath and shaken for an additional 15 minutes.

The flasks were then removed from the water bath to a rack above the

bath, continually flushed with CO2 and allowed to cool. Enough

ascorbic acid was added to stop the reaction (color changed from

bright yellow to pale yellow or white). Each sample was filtered

on a sintered glass crucible, and the residue obtained was termed

holocellulose (HOLO); the soluble portion was termed lignin (LIG).

The holocellulose residues were further fractionated to cellu-

lose and hemicellulose by the use of KOH as suggested by Whistler

and Gaillard (1961), Bailey et al. (1976) and Wilkie (1979). Holo-

cellulose residues were transferred to 100 ml glass bottles with

plastic screw-on caps. Weights were determined by difference, as

for NDF. Fifty milliliters of 5% KOH was added and the bottles

flushed with N2. Up to 20 samples at a time were shaken inter-

mittently in an Eberbach shaker for 24 hrs. After 24 hrs, an addi-

tional 9.5 gm KOH was added to each bottle (to bring the final

solution up to 24% KOH), the bottles were again flushed with N2 and

shaken an additional 24 hrs. The samples were filtered on sintered

glass crucibles fitted to 250 ml side-arm Erlenmever flasks. The

residue was washed with 10% acetic acid and water and was termed

cellulose (CELL). The filtrate was immediately neutralized to pH 7

by the addition of acetic acid and refrigerated. The filtrate

contained the hemicellulosic portion of the cell walls (HC). CW

components were expressed as the percentage of NDF.








A list of abbreviations is included in Table 1. Descriptions

of hays and composition data are listed in Appendix Tables 14 17.

Statistical analysis (S.A.S., 1979) included the general linear

model procedure and simple correlations.


Results and Discussion


Analyses of variance for the chemical composition of the hays

are presented in Table 2. Cultivar and age were the main effects.

Age generally did not have a consistent linear effect on chemical

parameters. For all parameters, hay (cultivar) and age were

significant. Significant interactions of hay x age occurred only

for NDF, NDF(H), HC, HEMV, CELL and PLIG. Except for HOLO and LIG

(which sum to 100 for each sample), the R2 values were .89 .94,

showing that the effects of hay and age explained most of the

variation in chemical composition.

One of the objectives of this experiment was to determine if

the HC, CELL and LIG values were different from HEMV, CELLV and

PLIG or, in other words, wasthe NaC10O-KOH method different from the

Van Soest analyses. The two procedures (NaC102-KOH vs. Van Soest)

were compared by analysis of variance of hemicellulose, cellulose

and lignin (Table 3). For all 3 variables, methods were different.

The overall means for NaClO2-KOH vs. Van Soest's analysis, respec-

tively, were hemicellulose 51.8 vs. 45.2; cellulose 40.8 vs. 43.5;

and lignin 7.4 vs. 8.6 (% of NDF).

The correlation between values obtained for the various chemical

fractions by the two methods are in Table 4. High correlations








Table 1. Abbreviations used in the dissertation.


ADF Acid detergent fiber, % of DM

CC Cell contents, 7 of DM

CELL Alkali-insoluble cellulose, % of NDF

CELLD Cellulose digestibility, in vivo, %

CELLV Van Soest cellulose (ADF PLIG), % of NDF

CP Crude protein, % of DM

CWC Cell wall constituents, % of DM

DM Dry matter, %

DNDF Digestible neutral detergent fiber, % of DM

DOM Digestible organic matter, % of DM

DOMI Digestible organic matter intake, g/kg 75/day

HC Alkali soluble hemicellulose, % of NDF

HCD Hemicellulose digestibility, in vivo, %

HEMV Van Soest hemicellulose (NDF ADF), % of NDF

HOLO Holocellulose, % of NDF

HOLOD Holocellulose digestibility, in vivo, %

IVCELLD In vitro cellulose digestion, rumen fluid only, %

IVCELLDP In vitro cellulose digestion, rumen fluid plus pepsin, %

IV`HCD In vitro hemicellulose digestion, rumen fluid only, %

IVHCDP In vitro hemicellulose digestion, rumen fluid plus pepsin, %

IVHOLOD In vitro holocellulose digestion, rumen fluid only, %

IVHOLODP In vitro holocellulose digestion, rumen fluid plus pepsin, %

IVNDFD In vitro neutral detergent fiber digestion, rumen fluid
only, %

IVNDFDP In vitro neutral detergent fiber digestion, rumen fluid plus
pepsin, %








Table 1 continued.


LIG Lignin (100 HOLO), % of NDF

NDF Neutral detergent fiber, % of DM

NDFD Neutral detergent fiber digestibility, in vivo, %

NDF(H) Neutral detergent fiber (Hartadi, 1980), % of DM

NDFI Neutral detergent fiber intake, g/kg W'75/day

OM Organic matter, % of DM

OMD Organic matter digestibility, in vivo, %

OMI Organic matter intake, g/kg W'75/day

PLIG Permanganate lignin, % of DM





































C-,

























C,
0r


0 0



o o o





















0 0 cO













0 0J 0
0 0 (
0 0 0
- oi oo
o o oo


































O O
o o o
o o Lfl






























o 0 r
0 0 0





o 0 0












-4 O







0 0 C
o o o



o 0 0









o 0 0








,T Cn <1






3= < 3Q
-4or


c
r i


ON
r r- 0







000







-T
cy 0 c-

C C-4








m-:1









-4 t
U-T



cc;









00






0yi C 0 i*-


iI.







rI


Q,





Zu "
O 1

(U
MI c






ao E






a c0
U m









L 0
0 C







w) co
P.







>
(0i
ai co
Q- T-




*-1 ^
TO J








Table 3. Analysis of variance of method differences.a


P Values

df Hemicelluloseb Cellulose Lignin


Hay (H) 4 .0001 .0001 .1364

Age (A) 3 .0001 .0001 .0001

H x A 12 .0001 .0034 .0193

Method (M) 1 .0001 .0001 .0001

H x M 4 .0001 .0029 .0001

A x M 3 .0111 .0026 .0001

A x H x M 12 .2122 .3154 .0836


C.V. 2.6 2.7 8.4

S.D. 1.3 1.2 .7

Mean 48.5 42.1 8.0


aThe two methods are NaC0l2-KOH

Values expressed as percentage


and the Van Soest

of NDF.


analysis.









Table 4. Correlation coefficients (r) between variables obtained
by two methods.a,b



NaClO1-KOH

NDFc HOLO HC CELL LIC


Van Soest

NDF(H)d .96*** .15 .10 -.06 -.15

HEMV .03 .12 .71*** -.73*** -.12

CELLV -.04 -.24 -.86*** .86*** .24

PLIG .55*** .03 -.01 .02 -.03


aThe two methods are NaClO1-KOH and the Van Soest analysis.
b
Abbreviations are found in Table 1.
CAnalyzed by Van Soest method and recovered on sintered glass.
Analyzed by Van Soest method and recovered using glass wool.

*** P<.001








existed between the two hemicellulose fractions (r=.71) and the two

cellulose fractions (r=.86) but there was no relationship between

the lignin values (r=-.03).

Differences due to method of analysis are more readily apparent

when presented in graphic form. The actual numerical values for the

means of each chemical analyses, grouped by age and cultivar, may

be found in Appendix Tables 18 22.

Neutral Detergent Fiber

NDF values were higher than NDF(H) values (Figures 1 and 2),

presumably due to differences in procedure. NDF(H) was filtered on

Gooch crucibles packed with glass wool, which is more rapid than

filtering on sintered glass crucibles, as the NDF samples were. NDF

and NDF(H) analyses were conducted by two different analysts, at

different times. Disregarding the numerical values of the two

determinations, it can be seen that the lines follow the same

pattern, all cultivars increasing in NDF from 2 to 4 weeks, then

leveling off or decreasing. Overall, in both cases, the two digit-

grasses are different from the bahiagrasses and the bermudagrass.

The two analyses were highly correlated (Table 4). The NDF residues

were used for subsequent fractionation in the present study.

It has often been stated that maturity is the greatest factor

affecting forage quality. In many instances, the effect of age is

linear; i.e., with increasing maturity, some factor increases and/or

decreases. Well-known effects of maturity have been the increase in

cell wall components and the decrease in crude protein and digesti-

bility. The data presented here, however, generally do not exhibit














90







/ / .. o. ..-
'%0
80 o








70
2 4 6 8
W E E S

Figure 1. Change in NDF content by weeks and hay.

O- -Argentine bahiagrass, 0- -0Paraguay bahiagrass,
*---*X46-2 digitgrass, ----' X124-4 digitgrass,

A. A Coastcross-1 bermudagrass.














85









75

LA-







65


WEEKS


Figure 2. Change in NDF(H) content by weeks and hay.
0- -0 Argentine bahiagrass, 0- -0Paraguay bahiagrass,
i----* X46-2 digitgrass, -----*X124-4 digitgrass,

A. Coastcross-1 bermudagrass.


I ,- I


..








a linear trend with maturity. One known event, extensive insect

damage (striped grass looper, Mocis latipes) at 4 weeks to the

digitgrasses and bahiagrass, may provide partial explanation for

lack of linearity.

Holocellulose

Holocellulose, the delignified plant cell wall composed prima-

rily of hemicellulose and cellulose, made up the major portion of

the NDF fraction, ranging from 90.45 94.36% of NDF (Appendix

Table 19). Holocellulose values, and all subsequent chemical compo-

sition values, were expressed as percentage of NDF rather than DM.

Holocellulose content of X46-2, X124-4 and Argentine cultivars

were not different; Coastcross-1 was different from the former three

cultivars and Paraguay fell between Coastcross-l and the rest (Figure

3). Four of the cultivars showed an eventual decline in holocellu-

lose content from 2 to 8 weeks but the interim pattern was erratic.

Earlier plant analysts felt that lignin interfered with the

analysis of structural carbohydrates (Whistler et al., 1948).

Adaption of wood chemistry methods led to the use of NaC 02-KOH

to de-lignify plant material. Packett et al. (1965) indicated that

polysaccharide structures are left intact after delignification

provided the procedure is not continued to reduce lignin values to

or less. Wilkie (1979), however, stated that delignification

may oxidize some reducing-end residues and cause partial depolymeri-

zation. Whether or not hemicellulose is bonded to lignin is still

uncertain although work by Albersheim et al.(Bauer et al., 1973;

Keegstra et al., 1973; Talmadge et al., 1973) on dicotyledonous
















94

eeoe

LL93



S92
=-Jo


91



90 -
2 4 6 8
WEEKS
Figure 3. Change in holocellulose content by weeks and hay.

) - -Argentine bahiagrass, 0- -*Paraguay bahiagrass,
*----*- X46-2 digitgrass, *---T X124-4 digitgrass,
G OCoastcross-1 bermudagrass.









plants suggest that there is bonding between hemicellulose and lignin.

Hemicellulose has been extracted from plants without prior delignifi-

cation (Bailey et al., 1976) but the completeness of the extraction

is unclear (Gaillard, 1958). It was felt on these tropical grass

hays, which are high in CWC and lignin, the best procedure would

be delignification prior to further analyses.

The holocellulose values, as an end in themselves, did not give

a clearer or better picture of the plant than did NDF values, which

are easier to obtain. The erratic pattern with maturity suggests

some caution should be used in interpretation of the LIG data

(100-HOLO).

Hemicellulose

Two methods of analyzing for hemicellulose were used NaC10 -KOH

and the Van Soest detergent system. Figures 4 and 5 present the

changes observed in hemicellulose content by cultivar and age (see

Appendix Table 20 for means). HC values, except in the case of

Coastcross-l, decreased only slightly with maturation which agrees

with other work on hemicellulose in tropical grasses (Bailey, 1973).

HEV values, on the other hand and again excepting Coastcross-l,

showed an overall greater decline with maturity.

HC values were higher as a percentage of NDF than were HEMV

values. With both methods, X124-4 was different from the other

grasses. The two bahiagrasses and X46-2 were similar in HC content

compared to the other grasses while Coastcross-1 was different from

the rest. The bahiagrasses and Coastcross-1 formed a similar group

with respect to HERV values, with the digitgrasses differing from

each other and from the rest of the grasses.












60







50








40
2 4 6 8
WEEKS

Figure 4. Change in alkali-soluble hemicellulose by weeks and hay.

- --Argentine bahiagrass, 0- -Paraguay bahiagrass,
*-- X46-2 digitgrass, ----*X124-4 digitgrass,
.ACoastcross-l bermudagrass.












60






Ll.

S50








40
s r\ ^-**' """





2 4 6 8
WEE S
Figure 5. Change in Van Soest hemicellulose by weeks and hay.

0- -Argentine bahiagrass, - Paraguay bahiagrass,
*--- X446-2 digitgrass, ----'-X124-4 digitgrass,
.A Coastcross-1 bermudagrass.









Hemicellulose has been defined as the cell wall and inter-

cellular polysaccharides that can be extracted from higher land

plant tissues that are, or were, lignified (Wilkie, 1979). More

simply, hemicellulose refers to all of the types of polysaccharides

found in plants other than cellulose, starch and fructans.

The preponderance of literature on hemicellulose deals with

the Van Soest hemicellulose obtained by subtracting ADF from NDF.

Work as early as 1970 (Bailey and Ulyatt) points out the problems

inherent in this definition, viz., that there are hemicellulosic

sugars present in the ADF residue (see also Morrison, 1980).

The hemicellulose content of a plant changes with growth and

maturity and is expected to increase with maturity. Bailey (1973)

reported on several studies with temperate grasses where the increase

in hemicellulose content with age was associated with an increase in

stem tissue. For tropical grasses, however, seasonal changes are

not as marked, the hemicellulose content remaining about the same in

spite of increasing CWC and increasing proportion of stems.

Hemicellulosic materials extracted by alkali may precipitate on

neutralization and mild acidification at 0 2 C to form hemicellu-

lose-A. The polysaccharides remaining in solution can be precipitated

with ethanol to form hemicellulose-B. Repeated attempts to obtain

a hemicellulose-A fraction failed. Scant information in the litera-

ture as reviewed by Bailey (1973) showed that hemicellulose-A

comprises 1.0 19% of DM in Heteropogon contortus leaves and stems

respectively, 1.4 3.2% of DM in Digitaria decumbens and .1 .4%

of DM in Setaria sphacelata. It would seem then, that the hemicellu-

lose-A fraction is generally low in tropical grasses, thus providing








partial explanation for our failure to obtain the fraction. HC

values presented here are representative of the hemicellulose-B

fraction which consists primarily of xylans.

Cellulose

Cellulose obtained by either method (Figures 6 and 7) showed

remarkably similar patterns (see Appendix Table 21 for means). In

both instances, Coastcross-1 was very different (having the lowest

amount of cellulose) from the other grasses; X46-2, Paraguay and

Argentine cultivars were not different, and X124-4 had the highest

amount of cellulose. CELL, collected on the sintered glass crucibles,

was a pure white fibrous substance very similar in appearance to

CELLV obtained by the ADF procedure.

Lignin

Lignin values (Figure 8) were calculated as 100 HOLO. These

lignin values are very different from the PLIG values (Figure 9;

Appendix Table 22). Lignin, like hemicellulose, is defined by the

method used to isolate it. The completeness of delignification of

NDF was tested by staining holocellulose with phloroglucinol-HCI

(a test for lignin). No stain was present in the holocellulose but

it is likely that the material oxidized by the NaCIO2 included more

than lignin. Obtaining a lignin value in this manner is not

recommended due to the possibility of additional products forming

during the oxidation of lignin from the cell walls by NaCIO2. Only

Coastcross-1 had a different amount of LIG chan the other grasses

while PLIC amounts were different for each cultivar.















50






CD


S40










30


WEEKS


Figure 6. Change in alkali-insoluble cellulose by weeks and hay.

()- -0 Argentine bahiagrass, O- -(Paraguay bahiagrass,

--- -X46-2 digitgrass, *----- X124-4 digitgrass,

A- Coastcross-l bermudagrass.


AL .


I I-


I ~r












50








40 .......A

--"





30
2 4 6 8
WE E(S

Figure 7. Change in Van Soest cellulose by weeks and hay.

0 -0 Argentine bahiagrass, O- -OParaguay bahiagrass,
*-- 46-2 digitgrass, *'----- X124-4 digitgrass,
A. .ACoastcross-1 bermudagrass.












10



9




C= 8





"0'

6



5
2 4 6 8
WEEKS

Figure 8. Change in chlorite lignin by weeks and hay.

0- -Argentine bahiagrass, - -OParaguay bahiagrass,
--- *X46-2 digitgrass, ---- X124-4 digitgrass,
. OCoastcross-1 berumudagrass.












10



9 / '. ........... O








6 /







5
2 4 6 8

WE E S

Figure 9. Change in permanganate lignin by weeks and hay.

O- -OArgentine bahiagrass, 0- - Paraguay bahiagrass,
S-- X46-2 digitgrass, *----- X124-4 digitgrass,
. Coastcross-1 bermudagrass.
Q. .Qroastcross-l bermudagrass.









Summary


One cultivar of Cynodon dactylon (Coastcross-l), two experimen-

tal lines of Digitaria species (X46-2 and X124-4) and two cultivars

of Paspalum notatum (Argentine and Paraguay) cut at four ages each

(2, 4, 6 and 8 weeks regrowths) were evaluated by chemical analyses.

Composite samples for each hay were analyzed for organic matter and

neutral detergent fiber. NDF fractions (cellulose, hemicellulose,

lignin) were determined by two methods: one, the conventional Van

Soest detergent analysis (CELLV, HEMV, PLIG); the other a classical

fractionation using hypochlorite and alkali (HOLO, CELL, HC, LIG).

Values obtained by the two methods used in the analyses were

significantly different. As percentages of NDF, HC values remained

the same or decreased slightly with age for each cultivar while

HEMV values showed a greater decline with maturity. CELL and CELLV

values had similar patterns, increasing up to 6 weeks, then decreasing

slightly. Correlation between HC and HEMV was .71 and between CELL

and CELLV was .86. Further studies of these fractions to determine

their digestibility and value in prediction equations are in subse-

quent chapters.
















CHAPTER IV
IN VITRO AND IN VIVO DIGESTIBILITY OF TROPICAL GRASS HAYS


Introduction


Ruminants derive a great portion of their dietary energy from

the digestion of cell wall polysaccharides when they are fed forage

diets. The digestibility and intake of CWC is crucial in determining

forage quality. Differences in breakdown of the CWC in the rumen are

due, in part, to differences in chemical structure of the polysaccha-

rides.

Early in vitro work on hemicellulose degradation in the rumen

used isolated fractions to determine digestibility of hemicellulose.

Two general theories arose from these studies. One theory attributes

low hemicellulose digestibility to chemical heterogeneity of the

hemicellulose structure (Caillard, 1965; Sullivan, 1966; Dekker and

Richards, 1973). Other work has shown that rather than the chemical

heterogeneity of the hemicellulose being responsible for lower

digestion, that the incomplete digestion was due to the presence of

lignin (Bailey and MacRae, 1970; Beveridge and Richards, 1973).

Morrison (1979) suggests that both theories may be involved, i.e.,

in young plants the chemical structure of the hemicellulose inter-

feres with digestion processes while in older plants, increasing

lignification causes lower digestibility.








Studies on isolated hemicelluloses and its components give

only a limited idea of what is occurring in the rumen. Beever et

al. (1971) in in vivo studies, showed that 70% of the hemicellulose

digestion occurred before entering the small intestine, very little

hemicellulose digestion occurred in the small intestine and approxi-

mately 30% of the hemicellulose digestion occurred in the cecum and

colon. A review of several digestion studies (Ulyatt et al.,1975)

found that the proportion of hemicellulose digested in the large

intestine was higher than that of cellulose. Bailey et al. (1976)

found that as hemicellulose content increased, the percent digested

in the rumen decreased. Ulyatt and Egan (1979) found 77 to 94% of

digestible hemicellulose apparently digested in the rumen, and 6 to

23% in the intestines. Some explanation of the confusing in vivo

results may be provided by Gaillard and Richards (1975) who suggest

the presence of a lignin-carbohydrate complex which passes from the

rumen in solution, is precipitated in the lower pH of the abomasum,

undergoes some redissolution at later stages in the lower gastroin-

testinal tract, and emerges in the feces.

An intermediate step, then, between studying the digestibility

of isolated hemicellulose and other CWC fractions and the complex

in vivo studies appears to be in vitro studies on whole plant tissue,

followed by analyses of the various fractions.

The objectives of this study were to determine the in vitro

and in vivo digestibilities of various fractions of tropical grass

hays.








Experimental Procedures


One cultivar of Cynodon dactylon (Coastcross-l bermudagrass),

two experimental lines of Digitaria (X124-4 and X46-2) and two

cultivars of Paspalum notatum (Paraguay and Argentine bahiagrass)

harvested after 2, 4, 6 and 8 weeks of regrowth with two field

replications were used in this experiment (Abrams, 1980). The

grasses were artificially dried and chopped.

The grasses were fed ad libitum as chopped hay to mature wethers

to determine simultaneously voluntary intake and digestibility of

organic matter (OM) and neutral detergent fiber (NDF) (Abrams, 1980).

Voluntary intakes were reported on the basis of metabolic body

weight (g/kg W 75/day). Each sample was pooled by animal, ground

to pass a 4 mm screen in a Wiley mill, mixed and a representative

portion reground through a 1 mm screen. A composite hay sample was

made by mixing proportional amounts from each animal fed that specific

hay. Feces were collected in canvas bags, weighed daily and 20%

kept for further analyses. The composite hay samples and some indi-

vidual feces samples were used for the analyses in this study.

Half gram samples of each hay were fermented in rumen fluid and

buffer for 48 hr, followed by 48 hr in acid-pepsin (Moore et al.,

1972; Moore and Hott, 1974; 1976), refluxed in neutral detergent

solution (Goering and Van Soest, 1970) and filtered on sintered glass

crucibles. The NDF residues were saved for further analyses. This

treatment was termed the rumen fluid plus pepsin (RF+P) treatment.

Another set of hay samples (.5 gm each) were fermented in rumen fluid









for 48 hr, refluxed in neutral detergent solution (Goering and

Van Soest, 1970), filtered on sintered glass crucibles and the NDF

residue saved for further analyses. This treatment was termed the

rumen fluid (RF) treatment. Each of the RF and RF+P residues were

analyzed for holocellulose (HOLO), hemicellulose (HC), cellulose

(CELL) and lignin (LIG) as outlined in Chapter III.

Fecal samples from the in vivo studies of 2 wk and 8 wk

Argentine bahiagrass, Coastcross-1 bermudagrass and X46-2 digitgrass

were analyzed for NDF, HOLO, HC, CELL and LIC (Chapter III), and acid

detergent fiber (ADF) (Goering and Van Soest, 1970). In vivo

digestion coefficients were calculated for these fractions. Abbre-

viations are listed in Table 1. Descriptions of hays and in vivo

data are listed in Appendix Tables 14 17.

Statistical analyses were performed using the statistical

analysis system (SAS, 1979) computer package for general linear

models, t-tests and correlation procedures.


Results and Discussion


In Vitro Digestibility

Analysis of variance of the in vitro data is presented in

Table 5. Values for in vitro digestion of individual hays are in

Appendix Table 15. In vitro digestibility of all fractions decreased

with increasing maturity in all hays. RF and RF+P digestibilities

were compared by t-tests (SAS, 1979) (HO:RF+P RF = 0) for each

fraction (NDF, HOLO, HC and CELL) by hay and weeks of regrowth

(Table 6).





















- r- I
0 r-1







o 0




0 0
0 01





- 0
0 OI
0 Cr
0 0





n 0D
0 0I
0 ,-.









0 01
0 0
0 -- 1


















0 0

CD I
0 0
O 01
0 0







0 0
0 0 1
0 0
1







0 0
0 0 1
0 0


17 cni (
i-t





>, Ol, V,
CO &0l
3


ir. r





Lo





4 C \O
















mo r~- <
*


\o 0

CNJ m
1tn


0 00 00








0 r- Or>

rn r










rn CM4 1
Lfm


* 0


0 0

O O
m 0






0


c-


co
,a
0





U
4-






(0
'2)





0
ru












-i-
*r-









>




1-4
0)
Cl








u-
co
9
1-1
LA


MO






















1-
mU
O



cd
S







4I-

U A
Cd-











Co
0














-





41








A
60












--1




N
i-


-4
C











-4



0

4-4
Cd


1-4
Trl































i-l







0

*rA




4-
co
Ca




















E--
r4
0


4-I




a,















n
rl








Cd

0










*-4
1-


cd 0

r- 4 a


U U


CM Cm CM4 C I
0 0 0 0
* *





00 r- 0 -1







-000
o -4 to CM
o No i







I
to ej I






r- -T --T o i
o 0 0








-N CN C4 CN







o o oC 1 1















n*
N- II I









OOCM






SN 0 3 I


0 ot
*


3 .0 CM
* to
en i


0

t 0
z Z


U ex
SU


A







4J













A







4-









4-

A
04


4-
u

N A
I CL
\0

X


0




0 r- 00














CN



-l









NDF digestibility was greater (P values ranged from .14 to

.0007, Table 6) in RF+P for all hays and regrowths. The standard

method of determining NDF digestibility (Goering and Van Soest, 1970)

uses only rumen fluid for 48 hr. While the overall increase in NDF

digestibility ranged from 2.2 to 5.7 percentage units (Table 7), for

individual hays there were differences in NDF digestibility of up to

7 percentage units. In most instances, CELL was more digestible

than HC, whether digested in RF or RF+P. HOLO and CELL digestibility

increased in RF+P only for Paraguay and Coastcross-1 and either

decreased or showed minimal change in the other grasses. HC digesti-

bility was not different for X46-2 but showed an increase of 2.1 to

3.0 percentage units for the other grasses. For individual hays,

there were increases in HC digestibility of up to 12 units by use

of acid-pepsin.

By weeks of regrowth, NDF digestibility increased with RF+P

except at 8 weeks. HOLO and CELL showed no significant difference

in digestibility. HC exhibited the greatest difference in digesti-

bility at 2 weeks; the difference decreased with increasing maturity.

NDF and HC digestibility decreased with age (Table 7 and Appendix

Table 16) which is consistent with work on temperate grasses (Gaillard,

1965; Waite et al., 1964; Bailey, 1967).

It seems, then, that further digestion in vitro of the CWC

occurs under acid-pepsin conditions. This digestion, for the most

part, is not attributable to further cellulose digestion but rather,

to hemicellulose digestion and dissolution of lignin. Whether this

increase in hemicellulose digestibility is due to solubilization by
























00 cn 01%




00 0 "
o r CM

%D Lrn V


% 0 0
~ -.7-


en %D t-


lZ r- 0Z
M. 1 y

e m


CV) c" 0
eo n i
* e *
NMrl


0r0 I 0






r- % L,
4 N H

Inll


o Ln %D r- CV %D


iA in Lrn %D in V


-.7 D '.

C-') '.D -.
'.0 -j 't-


c0 %o L*


-* C
0 0




00 cc-
cc -n


N& 4-1 u -1
u (M c 1
0) I 0) C 0)
3 %D &o co :
IT $4 0
(" x < 0 -4


rl



C 0
En r-( Cl(
-l 4J U .;
0) 1 C a0) C-
0) I ) m 0) I


oX < u o


S *o






) Lr






r- 0











Lrn -.

n <



* *

in~ -*


0 0






r 0



rl-





Lr -.T














NO
Scc





















L* *






0) (M
o 0
C O


C 0
(n) En
r0
coc
u 0
< u








the acidic conditions and whether the hemicellulose would re-precipi-

tate under neutral conditions (as suggested by Gaillard and Richards,

1975) is not known. The majority of the in vivo studies have used

HEMV as the measure of hemicellulose. Our earlier experiments suggest

that HEIXV and HC may not be the same entity. Thus, studies showing

that the greatest proportion of "hemicellulose" digestion occurs in

the rumen may be misleading, as indicated by these in vitro results.

In Vivo Digestibility

Fecal samples collected during the feeding trials of these

tropical grass hays were used in further analyses. Argentine bahia-

grass, Coastcross-1 bermudagrass and X46-2 digitgrass were chosen

from each of the three genera at 2 wk and 8 wk of age. NDF, HOLO, HC,

CELL, LIC and ADF values are found in Table 8. Analysis of variance

of the chemical composition of the fecal samples is in Appendix

Table 24.

Compared to the chemical composition of the hays (Appendix

Table 14), fecal samples were lower in NDF, HOLO and CELL, higher

in LIG and were about the same or slightly higher in HC content.

In vivo intake information was used in conjunction with the feces,

orts and waste information to calculate digestion coefficients for

NDF, HOLO, HC and CELL (Table 7; Abrams, 1980). Analysis of variance

for the digestion coefficients is in Appendix Table 23. The values

are compared with in vitro digestibility data on the same hays in RF

and RF+P (Table 7). At 2 weeks, in vivo digestion coefficients were

similar to in vitro for X46-2 digitgrass and Argentine bahiagrass.

For these two grasses, also, digestion coefficients were higher for









Table 8. Chemical composition of feces.


NDFb HOLO ADF HC CELL LIG


2 Weeks

X46-2 61.3c 89.9a 37.5b 57.2a 34.4a 10.1b

Argentine 67.8b 86.2b 38.9b 56.2a 30.la 13.8a

Coastcross-1 73.4a 86.7ab 43.4a 53.9a 32.8a 13.3ab

8 Weeks

X46-2 73.9c 86.6b 44.2a 50.5a 46.la 13.4a

Argentine 76.5b 84.3b 44.3a 49.4a 31.5b 15.7a

Coastcross-1 82.la 90.3a 45.5a 54.2a 36.0ab 9.7b


aNumbers within a column at each age followed by different letters
are significantly different from each other (Duncan's multiple range).
Abbreviations may be found in Table 1.








cellulose than for hemicellulose at 2 and 8 weeks. The Coastcross-1

bermudagrass had higher in vivo digestion coefficients. Digestibi-

lity of the various fractions was high at an early stage of growth.

Not unusually, the situation changed with maturity; at 8 weeks, the

in vivo digestion coefficients were higher than the in vitro,

ranging from at least 2 to 14 percentage points higher. The degree

of relationship (or correlation) between the in vivo and in vitro

values was high (Table 9) despite lower in vitro values at 8 weeks.

Implications are that in vitro digestibility information is more

representative of in vivo performance at early stages of plant

growth and that with maturity, the in vitro digestibility values

still accurately reflect in vivo but the absolute values will be

lower than in vivo values. The correlation of HCD with IVHCDP was

higher than with IVHCD, implying that acid-pepsin may have some

effect in vivo.


Summarv


One cultivar of Cvnodon dactvlon (Coastcross-l), two cultivars

of Digitaria species (X46-2 and X124-4) and two cultivars of Paspalum

notatum (Argentine and Paraguay) cut at four ages each 2, 4, 6 and

8 weeks of regrowth, were evaluated by in vitro digestion of neutral

detergent fiber (NDF) in rumen fluid (RF) and RF plus acid-pepsin

(RF+P) and by chemical analyses of NDF fractions. In addition,

sheep feeding trials uere run. Composite samples of each hay were

analyzed for NDF, holocellulose (HOLO), hemicellulose (HC), cellulose

(CELL) and lignin (LIG). Fecal samples of 2 and 8 week Coastcross-l,








Table 9. Relationship (r2) between in vivo
coefficients for six hays.


and in vitro digestion


2
Variables in Model r P


NDFDa vs. IVNDFD .91 .0033
NDFD vs. IVNDFDP .92 .0025

HOLOD vs. IVHOLOD .88 .0058
HOLOD vs. IVHOLODP .84 .0094

HCD vs. IVHCD .87 .0065
HCD vs. IVHCDP .91 .0034

CELD vs. IVCELD .84 .0100
CELD vs. IVCELDP .79 .0180


aAbbreviations may be found in Table 1.









Argentine and X46-2 grasses were analyzed for NDF, HOLO, HC, CELL,

LIG and acid detergent fiber (ADF).

In vitro digestibility in RF and RF+P of NDF, HOLO, HC and

CELL decreased with increasing maturity in all hays. In vitro

digestion of cellulose was generally greater than that of hemicellu-

lose. In vitro digestibility of INDF and hemicellulose was increased

by further digestion in RF+P while holocellulose and cellulose di-

gestibility showed minimal change.

Fecal samples were lower in NDF, HOLO and CELL, higher in LIG

and about the same or slightly higher in HC content. In vivo

digestion coefficients were calculated for each of the fractions

(NDF, HOLO, HC and CELL) and showed trends similar to in vitro

digestibility.













CHAPTER V
PREDICTION OF FORAGE QUALITY FROM CHEMICAL AND IN VITRO ANALYSES


Introduction


Forage quality must be related to animal performance. Moore

and Mott (1973) defined forage quality as output per animal and as

being a function of voluntary intake and digestibility of nutrients

when the forage is fed alone, ad libitum, to a specified animal.

When long-term production trials are not feasible as is often the

case, researchers turn to laboratory analyses of forage quality.

Many and varied are the laboratory procedures devised to predict

forage quality. These may be grouped into chemical and in vitro

methods. Thus far, chemical methods have not been overly successful

in predicting forage quality but the in vitro methodology, especially

with respect to predicting in vivo digestibility, has been more

successful. In a review of the many forage evaluation methods in

popular use, Barnes and Marten (1979) reported that the in vitro

system was superior to chemical analyses for predicting in vivo

digestibility of forages. Intake is much more difficult to predict

from laboratory analyses but Rohweder et al. (1976) have suggested

the use of neutral detergent fiber (NDF) to predict intake.

This study was conducted to examine the relationship between

the Van Soest chemical analyses, another chemical analysis procedure

(NaC102-KOH, see Chapter III) and/or in vitro digestion with in vivo









intake and digestibility of organic matter (OMI) and NDF in 39

tropical grass hays of three genera harvested after 2, 4, 6 and 8

weeks of regrowth.


Experimental Procedure


One cultivar of Cvnodon dactvlon (Coastcross-1 bermudagrass),

two experimental cultivars of Digitaria (X46-2 and X124-4 digitgrass)

and two cultivars of Paspalum notatum (Argentine and Paraguay bahia-

grass) harvested after 2, 4, 6 and 8 weeks of regrowth were used in

this experiment. Hays were chopped and fed ad libitum to mature

wethers to determine intake and digestibility as outlined in

Chapter III. A composite hay sample was made by mixing proportional

amounts of hay from each animal fed that particular hay, was ground

through 4 mm and 1 mm screens of a Wiley mill and was used for the

laboratory analyses.

The composite hay samples were analyzed for DM, OM, NDF, holo-

cellulose (HOLO), hemicellulose (HC, HEMV), cellulose (CELL, CELLV)

and lignin (LIG, PLIG) as outlined in Chapter III. In vitro fermen-

tation in rumen fluid (RF) for 48 hrs and in vitro fermentation in

RF for 48 hrs followed by 48 hrs in acid pepsin (RF+P) were both

followed by refluxing in neutral detergent solution to determine

in vitro NDF, HOLO, HC and CELL digestion as outlined in Chapter IV.

A list of abbreviations is presented in Table 1.

Statistical analyses were performed using the statistical ana-

lysis system (SAS, 1979) computer package for general linear models

and correlation procedures.









Results and Discussion


Intake and Digestibility

Individual values of in vivo intake and digestibility are in

Appendix Tables 17, 24 and 25. Chaves (1979), Abrams (1980) and

Hartadi (1980) presented discussions of the relationship between

intake and digestibility in vivo of these hays. They suggested that

digestibility was not the primary factor controlling intake of these

hays. Furthermore, there was a poor relationship between intake and

digestibility (OMI vs. OMD, r = .47; Chaves, 1979). Therefore, it

is unlikely that a single laboratory procedure will prove to be a

good predictor of in vivo performance.

Another reason for skepticism concerning the ability of a single

laboratory procedure to predict forage quality lies in the effect of

maturity on the in vivo values compared to the effect of maturity on

the laboratory values. In Chapter III, it was reported that maturity

did not have a consistent effect on the chemical parameters, i.e.,

the effect of age was not linear. Expectations are that, with

maturity, some factors decrease (e.g., crude protein) and some

increase (e.g., NDF) but the hays used in this study did not follow

the expected pattern. However, the in vitro and in vivo values did

follow the expected pattern of decline with maturity in most of

the hays (see Appendix Tables 15 and 16). Argentine bahiagrass and

Coastcross-1 bermudagrass did not change in OMI and changed very

little in OMD but the other three grasses decreased from 2 8 weeks

in OMI by 4 8 units (g/kg W.75/day) and decreased in OMD by 10 12

percentage units (Abrams, 1980).









Prediction of Forage Quality from Laboratory Analyses

Simple correlations between in vivo data and laboratory

analyses were conducted (SAS, 1979). Only the best correlations

for each in vivo parameter are presented in Table 10. The highest

correlation obtained between OMI and all laboratory analyses was

with CP (r2=.20). DOMI had the highest correlation with NDF

(r2=.46) and [NDFI had the highest correlation with PLIG and CP

(r =.06). Intake has always been difficult to predict by laboratory

analyses alone. Differences in voluntary intake are often attributed

to rate of digestion and in vitro digestibility systems generally

are more accurate at predicting intake than are chemical analyses.

Table 11 presents correlations between in vivo and in vitro values.

Two in vivo intake variables, OMI and DOMI, had significant corre-

lations with all of the in vitro analyses; r values ranging from

.13 to .56 with significance probabilities of P5.10 to 5.001, respec-

tively. NDFI, which did not vary (Appendix Table 24), was not

predicted well, or at all, by any of the in vitro analyses.
?
OMD and NDFD had the highest correlations with NDF (r-=.59 and

.56, respectively) while DNDF had the highest correlation with PLIG

(r-=.23) (Table 10). Aside from NDF and PLIG, there were no chemical

analyses which accurately predicted in vivo digestibility. However,

the in vitro analyses generally gave highly significant correlations

with OMD and NDFD; r- values ranging from .61 to .81 (Table 11).

Correlations between in vitro analyses and DNDF, although significant,

were lower.









2
Table 10. Relationship (r ) between in vivo values and
analyses, obtained with thirty-nine different hays.


laboratory


Variables in Model
Variables in Model r P


OMIa
OMI
OMI

OMD
OMD
OMD
OMD
OMD

DOMI
DOMI
DOMI


vs. NDF
vs. PLIG
vs. CP


NDF
PLIG
CP
HC
CELL


vs. NDF
vs. PLIG
vs. CP


NDFI vs. PLIG
NDFI vs. CP


NDFD
NDFD
NDFD

DNDF
DNDF
DNDF
DNDF
DNDF


vs. NDF
vs. PLIG
vs. CP


NDF
HC
CELL
CELLV
PLIG


aAbbreviations may be found in Table 1.


.46
.37
.28


.0076
.0053
.0039

.0001
.0001
.0026
.2725
.2840

.0001
.0001
.0005

.1379
.1166

.0001
.0001
.0022

.0921
.0097
.0056
.0099
.0010


.56
.42
.23

.07
.17
.19
.17
.23







9
Table 11. Relationships (r ) between in vivo values and in vitro
analyses, obtained with thirty-nine different hays.a


OMID


IVNDFDa
IVNDFDP


In vivo Variables
DOM I JDFI


NDFD


D~ JOE


.18** .77*** .54*** .02
.20*** .79*** .56*** .02


IVHOLOD .15* .72*** .47*** .01
IVHOLODP .19** .71*** .52*** .01


.16** .73*** .49*** .01
.20*** .67*** .52*** .02


IVCELLD .17** .71*** .50*** .02
IVCELLDP .13* .61*** .42*** .004


.80*** .32***
.81*** .28***

.75*** .31***
.70*** .18**

.74*** .25***
.65*** .14*

.78*** .38***
.63*** .20***


P1.10
** P.01
*** P2.001


aAbbreviations may be found in Table 1.


IVHCD
IVHCDP


D:JDF








2
Interestingly, in almost every instance, r values for each

chemical fraction's in vitro digestibility in RF and RF+P (IVNDFD

and IVNDFDP, for example) were not different. In Chapter III,

it was shown that there was a significant increase in digestion

in the RF+P treatment for the NDF and hemicellulose fractions,

while holocellulose and cellulose digestibility showed minimal

change. The similar in vivo correlations with the RF and RF+P

in vitro data make laboratory decision making easier. Since there

is little difference in the two methods (RF vs. RF+P) insofar as

predicting in vivo data is concerned, the method of choice would

be RF, which is less time, equipment and reagent consuming.

Furthermore, there was no advantage to using a more complicated

procedure for residue analysis than the NDF procedure.

Tables 12 and 13 present the models obtained by general linear

models procedure (SAS, 1979) that best fitted the in vivo intake

and digestibility data. A rational, rather than an empirical or

multiple regression, approach was used to choose independent

variables for each model. Independent variables were chosen on

the basis of significant correlation with the in vivo measurement

of interest (Tables 10 and 11). In most cases, in vitro digestion

values in RF only were used unless RF+P values were much higher.

In addition, some variables were included on the basis of their

presumed biological role in the structure of the cell wall. Cell

wall structure and components play a crucial role in determining

intake and digestibility of tropical grasses.









Table 12. Coefficient of determination (R2), probability (P) and
standard error of the estimate (s ) for the regression of in
vivo intake values on several independent variables.

Dependent Independent 2 P s
Variable Variable(s) y-x

OMIa,b NDF PLIG .24 .0078 6.42
OMI NDF HC PLIG .25 .0166 6.44
OMI NDF CELL PLIG .26 .0150 6.43
OMI NDF HC CELL PLIG .25 .0347 6.51
OMI NDF HC HC2 .23 .0269 6.54
OMI HC CP PLIG .25 .0185 6.47
OMI IVNDFDP IVHOLODP .20 .0183 6.57
OMI TVNDFDP IVHOLODP TVHCDP .21 .0366 6.61
OMI IVNDFDP IVHOLODP IVHCDP IVCELD .21 .0779 6.70

DOMT NDF PLIG .53 .0001 4.89
DOMI NDF HC PLIG .53 .0001 4.94
DOMI NDF CELL PLIG .53 .0001 4.94
DOMI NDF HC HC2 .47 .0001 5.30
DOMI HC CP PLIG .42 .0002 5.47
DOMI IVNDFDP PLIC HC .58 .0001 4.71
DOMI IVNDFD IVHOLOD .55 .0001 4.81
DOMI IVNDFD IVHOLOD IVHCD .55 .0001 4.87
DOMI IVNDFD IVHOLOD TVCELD .55 .0001 4.93

NDFI HC PLIG .07 .2824 4.85
NDFI HC CP PLTG .08 .4106 4.90
NDFT IVNDFDP HC PLIG .07 .4677 4.92


aAbbreviations may be found in Table 1.
bIntake expressed as g/kg W 75/day.









Table 13. Coefficient of determination (R2), probability (P) and
standard error of the estimate (s ) for the regression of in
vivo digestibility values on several independent variables.

Dependent Independent 2
Variable Variable(s) y.x

OMDa,b NDF CP .61 .0001 3.62
OMD NDF CP PLIG .65 .0001 3.47
OMD NDF PLIG .65 .0001 3.42
01D NDF HC .59 .0001 3.69
OMD NDF CELL .60 .0001 3.68
OMD NDF HC PLIG .66 .0001 3.44
OMD NDF HC CELL .43 .0001 3.74
OMD NDF HC CELL PLIG .66 .0001 3.49

NDFD CP PLIG .43 .0001 5.14
NDFD NDF CP .59 .0001 4.38
NDFD NDF PLIG .64 .0001 4.08
NDFD NDF CP PLIG .64 .0001 4.14
NDFD CP PLIG HC .54 .0001 4.67
NDFD NDF HC PLIG .66 .0001 4.02
NDFD NDF HC CELL PLIG .67 .0001 4.05
NDFD NDF HC HC2 .58 .0001 4.46

DNDFC HC CELL .19 .0225 3.50
DNDF HC PLIG .40 .0001 3.02
DNDF HC CELL PLIG .42 .0002 2.99
DNDF PLIG CELL .42 .0001 2.96
DNDF CP PLIG .32 .0011 3.22
DNDF CP PLIG HC .40 .0004 3.05
DNDF NDF CP PLIG .32 .0035 3.26
DNDF NDF HC CELL PLIG .43 .0006 3.02


aAbbreviations may be found in Table 1.
Digestibility expressed as %.

CDigestible NDF expressed as % of DM.








Going from a model containing one independent variable to a

model containing more than one independent variable did not produce

better or more acceptable predictions of intake. OMI ranged from

46.7 to 79.7 g/kg W'75/day (Appendix Table 24). Four of the "best"

equations for predicting OMI included NDF and PLIG. R2 values and

standard error of the estimate, respectively, were .24 and 6.42

(NDF and PLIG), .25 and 6.44 (NDF, HC and PLIG), .26 and 6.43

(MDF, CELL and PLIG) and .25 and 6.51 (NDF, HC, CELL and PLIG). DOMI

(ranging from 25.6 to 54.0 g/kg W 75/day) was best predicted by two

and three variable models that included an in vitro value in the

model, as to be expected for the prediction of digestible OM intake.

The best equation for predicting DOMI included IVNDFDP, PLIG and HC
2
(R = .58 and s =4.71), but this three variable equation was not
v'x
2
much better than the equation using only IVNDFDP (R =.56).

NDFI was still predicted poorly by any equation, be it single

or multiple variable, probably because there is little variation in

NDFI (see Appendix Table 24). As R2 values increased slightly,

standard error of the estimate increased so that the best predictor

of NDFI remained crude protein alone (Table 10).

OtD and IDFD were predicted quite well by any and all of the

single variable in vitro equations (Table 11). Adding more variables

to the model decreased the R2 and increased s for OMD and NDFD.
y X
Single variable models of in vitro data used for predicting DNDF had



2
R2 values ranging from .14 to .38. The best two variable model for

predicting DNDF included PLIG and CELL (R =.42) and was not improved
by adding another variable to the model (HC, CELL and PLI, R=.42).
by adding another variable to the model (HC, CELL and PLIG, R =.42).










The best four variable model included NDF, HC, CELL and PLIG (R =.43)

but the standard error of the estimate increased with the addition

of more variables to the model statements.


Conclusions


The results of the studies indicated that NDF, CP and PLIG

were the chemical analyses most closely associated with the digesti-

bility and intake of tropical grasses. New methodology developed to

analyze cell walls did not contribute to the prediction of in vivo

performance. Most of the in vitro procedures gave fair to good

predictions of OMI, OMD, DOMI, NDFD and DNDF but not of NDFI. Two,

three and four variable models increased the R2 values for OMI, OMD

and DOMI and decreased the R2 values slightly for DNDF and more so

for NDFD. NDFI could not be predicted by any single or multiple

variable model. It is concluded that we still do not have an

accurate picture of the structure and function of the cell wall in

tropical grasses. The difficulty of predicting in vivo animal

performance remains a great challenge in evaluation of forage

quality. Twenty of these hays were analyzed at the Pennsylvania

State University by near infrared reflectance spectroscopy (NIRS)

(Russo, 1981, unpublished data). Using large multiple regression

models and two wavelengths, NIRS gave better estimates of forage

quality than any of the chemical and in vitro analyses conducted

in this study.









Summary


One cultivar of Cynodon dactylon (Coastcross-1 bermudagrass),

two cultivars of Digitaria species (X46-2 and X124-4) and two culti-

vars of Paspalum notatum (Argentine and Paraguay bahiagrass) cut at

four ages each (2, 4, 6 and 8 week regrowths) were evaluated by

chemical analyses and in vitro NDF digestibility in rumen fluid (RF)

and RF plus acid pepsin (RF+P). In addition, sheep feeding trials

were conducted. Composite samples for each hay were analyzed for

organic matter, NDF and holocellulose (HOLO). Hemicellulose (HC,

HEMV), cellulose (CELL, CELLV) and lignin (LIG, PLIG) were determined

by two different methods. Information obtained was used to generate

prediction equations for sheep in vivo intake and digestibility.

NDF, CP and PLIG were the best single chemical analyses for

prediction of intake and digestibility. OMI and NDFI were not pre-

dicted as well as DOMI. Any of the in vitro analyses were better

for prediction of in vivo digestibility (OiD, NDFD and DNDF) than

were chemical analyses. Two and three variable models which included

NDF and PLIG improved prediction of OMI and DOMI.
















CHAPTER VI
GENERAL DISCUSSION


Forage quality reflects the potential of animal production

from forage by use of the available nutrients in the forage. Forage

quality may be defined as animal performance (Moore and Mott, 1973).

Analysis of forage quality in the laboratory has almost a hundred

year history and gaps still remain in our knowledge of how the

forage plant is put together, how it functions and how it affects

animal performance.

Essentially all of the forage evaluation procedures have been

developed in temperate climates using temperate forages. Not

unlike the problems associated with transferring Western agricul-

tural expertise to developing countries is the problem of trans-

ferring forage evaluation schemes generated on temperate forages

to tropical forages. Sometimes it works, sometimes it does not.

Tropical grasses are rather different from temperate grasses.

Tropical grasses have a much greater photosynthetic capability, are

more efficient in water use, have different carbon pathways and

have different anatomies. The first product of photosynthesis is

carbohydrates and tropical grasses have specialized plastids in

the sheath surrounding the vascular bundles for starch conversion

and storage. Tropical grasses are higher in cell wall and cell wall

components than temperate grasses.








What is the cell wall? It is a highly ordered macromolecular

multilayered network made up of structural carbohydrates (cellulose

and hemicellulose), lignin and protein. Dividing plants into cell

contents and cell walls (as does the Van Soest forage evaluation

system) provides a rational basis for discussion of forage quality.

Cell contents are essentially completely digestible by the ruminant,

cell walls have incomplete digestibilities. A number of cell wall

models have been proposed. Albersheim and others (Bauer et al.,

1973; Keegstra et al., 1973; Talmadge et al., 1973) propose that the

cell wall is made up of long crystalline microfibrils (cellulose)

surrounded by hemicellulose, which is hydrogen bonded to the cellu-

lose and covalently bonded to the other cell wall components (pri-

marily lignin). Covalent bonds are more difficult to break than

are hydrogen bonds. Bailey and others (1976) propose that all of

the crystalline polymer need not be cellulose, but could be hemicell-

ulosic xylans.

The role of hemicellulose in the plant has been established

as structural but hemicellulose may exist not only in the cell wall

and have been found in the endosperm of cereal plants, for example

(Wilkie, 1979). Hemicellulose is added to the wall continuously and

is increased by increases in temperature and light. While cellulose

composition and structure does not change with maturity or under

environmental influences, hemicellulose does. Not enough information

exists, but it is likely that each genera, each cultivar, perhaps

even each plant, and certainly, different parts within a plant have

different kinds and amounts of hemicellulose. Hemicellulose content

and concentrations are variable, not constant.









The results of this research indicate that "hemicellulose" is

an elusive term. Van Soest hemicellulose is certainly not the same

hemicellulose obtained in the NaCIO2-KOH procedure despite presumed

similarities in the cellulose fraction. Lack of expertise prevented

further examination of the sugars making up the hemicellulose but

some initial observations were made. One of the experimental digit-

grasses, X46-2, which is "high" in forage quality by most standards,

appeared very low in xylans. Xylans are usually highly ordered,

even crystalline and difficult to digest (Daughtry, 1976). Tropical

grasses are generally higher in xylans than are temperate grasses

(Ojima and Isawa, 1968; Bailey, 1973) and xylans may be responsible

for lower quality of tropical grasses. This experimental line of

digitgrass should be examined for xylan content and compared with

other grasses used in this study.

In vitro digestibility results were illuminating. NDF and

hemicellulose digestibility increased in acid-pepsin conditions,

cellulose digestibility did not increase. Further breakdown of the

cell wall, then, in the lower gastrointestinal tract, could be due

to breakdown of the hemicellulose. Hemicellulose has generally

been presumed to have a low digestibility, with lignification

postulated as the limiting factor but a distinction should be made

between hemicellulose unavailability due to structural organization

of the cell wall carbohydrates or interference by cellulose, ligni-

fication, acetylation, etc. (Morrison, 1979). McLeod and Minson

(1974) propose that the cell wall carbohydrates have two fractions -

lignin-free and lignified the lignin-free being totally digestible,








the lignified fraction being undigestible. Other digestion models

(Waldo et al., 1972; Abrans, 1980) are similar. Accurate procedures

for determining hemicellulose and lignin must be developed in order

to answer some of these questions and to prove some of these theories.

It would seem that if the proposed cell wall models and digestion

models were fairly accurate, using both models would allow one to

state that hemicellulose can be "digested" from cellulose and

within the hemicellulose polymer itself but that pieces of the

hemicellulose polymer remain covalently bonded to lignin and are

more resistant to digestion. Bailey and Ulyatt (1970) and Morrison

(1980) intimate these conclusions also. To extrapolate further,

Chaves (1979) suggests that bridges are formed between the cell

wall carbohydrates and the lignin by phenolic acids which can

deleteriously affect microbial digestion.

In tropical grasses, hemicellulose, lignin and phenolic acids

are higher than in temperate grasses. Are these, then, the components

responsible for low quality in tropical grasses? Can the plant

breeder do anything about it? Breeding for low lignin brought about

the development of brown midrib corn, academically interesting but

of little use to the farmer. Is low hemicellulose or low lignin

content really desirable in a forage plant?

Forage quality analysis and prediction of intake and digestibi-

lity still has researchable areas. This research should be carried

further to determine monosaccharide composition, relate the composi-

tion to animal performance and then devise a rapid method of cell

wall monosaccharide determination. The physical and chemical forces








holding the various cell wall components together and what it takes

in terms of energy, microbes, enzymes, acids or whatever to break

those bonds could be examined. Anatomy and histology of forages as

related to intake and digestibility should be examined and the poly-

saccharides and monosaccharides in the various structures be

determined. Infrared reflectance spectroscopy may alter procedures

used by many forage testing laboratories as it becomes more accurate

and accessible for routine use. That may be for the better, allowing

more basic research to be conducted on forage plants.





































APPENDIX















So eeeOee O O OO ONWe Oeeee e e Oe
%D 0 )O' 4TL* T 00r r C C4 * C ** * **C L) ** * * **- r4 .0 C


omoocI -4L"r-4-Im o

CM r4 O 00 '~ v( ) 4 O C ] C ML O CM

mn d T m 's m' l l)I l -T(Tm s-T 'T 'T


-l T CM W % C O
CMq -T -.:T C14 W0 M O'

C4 CM i-v v1:) 0 0
N~> rI) u Wr L in n
.......00
rNrN~r ecln O


OC0 r-


oC0o om m0*m m


nO %c C r-
0oa~,c,-1


* *
r0 oN
r- 00


- r-l O l-4O
** ***r*O3
OHH0~ mOHH0


O r-I
n Cen0
*a *r
IT .Tlz


Tom -c-%oo

CM 0 T \000 i-
* * * a-in i


%oD r C% LnM r- t~


0N 0\ 0^ 0^0 CN m CNm
0 0 O~~~
, , .J ~


%0 %\0 0 % C
C 00O0D0n
* * * _
C)ahNaCN


c ooWmr-r- OOcM
r- r~- CM M M Mo r-c


00 0oo oo t~ 0' oo 0o r- P-


NOc-eeeee'ee-e
.............


HCN

00


c> C
41 -
MCI I

Cl -i 4r- r-I
rd0rcr
PI XX


a)
b>C

00 I


S0 0 0 0* 0* 0*
r~lJh 001O o


r-4 0 T 0 0 0 0w
-.: in 'T -1 CMq in r~
00Co Co
o * * *


I
E(
01
04

.)
cn i

> x


0)


M 44 -4
3 4 -C4 X
&0 C I
CB o '~


00v0
aloIo
***"a


00 00
,-1 % C0 %D
**4*l0
NC-NJ0*


I I

0 00

S41J U u
0i C4 4-J AJ CM
C Q) I 0 0 I


04 -:c x 0 0 x


-H c *H -H ',4 r: 5H -H 5A r t cO H r= 5 -r- I -H E -' H '

( (- (m m r-4 r4 CZi -1 -1 (3 (m -i (i 0 0 (C r- (-1 (Mi -1 (C 0 0 m (
41 a ma jl 4- a ca 4J cl a 4 a 4 -1 a ca rl 4 1 u a 1a Mu MJ 41 -C -0 4o 4 1 41
*rP- P- *d 4 -r-4 P -HL0 H -,A P. O H *,-I P. P. -H 0 0 T-H Ow 4 -I -H 3. Cu T-H 0 0 H




QrAQ mm QA QQ -1UG AA r-4 C4 C4 C4 C4 (y) (y) Trl rl rdLr Lr.%


o-4 CM

0 9
.T fm


C' -1mm


17-.:T OO %D %Dn O "

m C% o 0 o 'TO O % -C
o. CN .% ..C.C.(N *


30


r- o0- CM
r- r- r-



Sm 1-o

T0 00


c)
>a c
=J M4 -T
60 1.. CM
MO () T- I


c

4-)
0) CI I
(1) (a -T
-C -i cCM
,i C ,.--
<; P-i


r i-4 V) CM4


T3 Ln in) '-,
>HON~c
uogv\ M
* **lo
eMNW~~


CM CM4 r-1 l 00 CM
%0oY a o-4- CM4 V

















~~.D~~ 04 -T -. 4 ~C -.7

1 C C C C


N Du % D 00


T* ** 7
-T T ^T--


CO '0 .
00 J ",-0

0 o00












rN .-4 -n

m o O

--1 .-- o0
oO -o r


(Ni 10 .40
-7 10


r- cl r) r-
o oo
cCLAn nO


CO z CT, M C1M M c1 O

W r-4C1 0 m ND U
Sr.o o o ^ n U' n 'L.





1- c -i 0 0oM I 0Dn 0D

m T CNO N ON
04 CNI 0 (N 01 1 4
c7%C7"CNCN 7" 7"(T"C7


,- Cr00 N- O C0


CO CO C- IO CM C-
r 0oo 0 r- 0 ooo


00 0 -1 Fn -7 0 -.0


0' a' 0' a' 0' 0- 0'
a' LA C -7~


C-1 C- 00 -4 40
c rU
ooo~go


a0 -rH
=1 4-1 -3- IT
o00 C I I

ex C4 -7 C.





E E -- -1-1


Q. Ca *r- *r-I

- 40 *c- *I--
Se. QQ
~ U,
c= 1 -4.-


O-N LI rC I
0% 01 ON CN


SC Ca
CO Q I
0 sir


Co
EE E -'
r-4 ,-4 TO 0 0 0 0
mo ca 1.J *C -L -a -v
0- Q. *-4 O O O 0

CO 0 *-4 >. >. >C >C
C Cr- uu u


o 0 --in '%. r -T Ln 'D r-
, ? COOCO

















r- -IT

nC4
* i~


a,% -IT C
L<30

*n *i


- n rON ND O DO


N-NO-OONC
LfO-WN-O0N
* *0*


ON W (m3 M- o) C-4 r~-

c< _; cT4 r4 4 r0 0
* * * * *
0 ~v ~ m


r O\L
4T- '-I in


-ON~e'0omO

S N0lN~-0N


%D rl -\D %


Lor)o %oi CT\


W0OOO fcHN


c4T N 1
ins % sD


OCIo0\
O~~0hO


(% C100
r-- ) CM1
" *." '. '.


r (O m C 0o IT

in) Lni i) u


ComwV0N
0 ONCH
* * * **~6


%D C4 -00 0
SOc CM %D CO
NWCM MONC
a*1**1s1
s 4r O so


C-4 -t n Mo0 %D in


%Di c n -I t 'T %D D I
soWS inNd*As
* * * **hlr
oDoOr OMm W


- 1 0 C i IT w r


1- %D in Ln sO0 -IT I-
* *0 * *

0OMOO W0


- -m . mmmmm T T n Ln


\O n n \O D Ln


IT ) m.O
- ,- ..

.. u' <'


SC-4m

U 1Lf r-:


- SO M 003 0 CM






ro; r'- in ITa '-4 %.0 so ITa
h1\a0m 3hNN
rl\DOOMM(\




N ~oMCCONo
Hr lffOU3mN


TOOm







coos
* *


-IT %O
CO D


r~, %.0< c in "M %D m r-I r-. in
O r-| m O ON % -1 i-1 in
r4i 11; -~ r~- 1o r11> cT r4i c4 1o co
r" r" r~- %~.0 o10 r~- -1 ini m %Do <


r, Wc o r,-4 M mm o M in


*n n%0\ W) i %D W) LnD


00 000~m0


IT w> %o 0 rr o -1 r~-
HO 0NNO0NW
WoWNCHN
mbO N0 HN


inr) l C14sDCM

C4 _l; l r~
sIso in nin


" m r .Do M .0O -
in %.o a,,M -- -- oo

% nICT inO in m r- in
sO nCa-nmininin

















I-











-1





U


















C-










-I
-t
C-



















0-
a
o
c-






r_










o




r






o

r
1-1


,.D r-. 'D a, LCn Cr 'D0 r-. 00 ON


-7'- N 00 3*1 00 r-- a' 0 Co 4 C14
LI Lf4 1.0 Lt') U') LI) L4 iLr) Lf) ur) Lr






If' r)-) -7 1.0 00oI r-7 Lf)14 C7% C7%a
00 1.0 '.D IT 00 Co '-4 r- ir) -7 r-
.1 '. Co '. .1 '. .0 .4 .'
Lf4 Lf4 1.0 1.0 Lf4 Lo) Lf4 Lo) d) Lf4 -17







' %D0 -7 Ir) C4 r-- a7% Co r-- Lr


r- o o -4 r- r-Coor- o a)
-IT -.7 1.0 1.0 17 Lf4 Lf4 Lo) ir) Lf4 -Zr







C)oCo Lo) a,. 0. 1.0 C( N a' -7 N
e-- ON '0 %D0 a'% 1-4 r- -7 C. IT

Co Co o. Co Lf4 r- '.0 U) -7 Co r-
17 17 %D0 Lf4 17 -.7 LI) V) Lf4 17 -.7







'.0 C11 a C' ci 17 .. i I, 3' F ) '.0
.~C . .' Co .- a' a'
O 0) .-4 (3N -4 -7 r- 0o Co C. a
L') 124 %.0 U) ir) U') Lf4 Lf4 .fl Lf4 -7








%D0 a'% -4 -7 Co C C'n '0 C) 7 -7
Lf4 -7 '-4 Lr) Co --7 ~-4 3*1 C '7

'-4 cN eD a', o. -4 'o0 '. c7% a')C
124 t) '.0 124 Lf4 L(4 L,) 12) V24 -T -.







M~ a) C. 124 C C. 3*1 C'n .0 CN -7

a'4 3') r.F-3-4 00 m V. F-1 '4 a' Co

-.7 -7 %0 124)-.7 124 124 12) 124 -7 -.7







12n C. -4 -T '0 (N (NC '.0 Cn4
00 cn 0.C 010r ) -Co co

a' :7: Co N- C. 3' L4 Lf r Co F-
-7 -.7 124 124 Ln '1 7 Lt4 124 .24 IT7 17









Co C. -4 Cn L) 12 .0 F-- -7 424 '.0 F--
.or r-r-r- r- r- o~o a)o~o
















CM4 (3
'-4 -T-

10 CM
* *1a


0o i-A CM4 r- t~-
00 "M CM4 Ulf CM
CON NlrO





NhC14 00 0



nCM ooo CM





















N lCNQI
110 r~- P- %









m00-1 ooC 00










00 C 19 CM ril
CT^ 00 r- r--l r--l
CM4"CM M~ M






1-4 ( 3-T cM oo
CM r~ i O rs
ON C0 c') 0o r^




rLi r- r- 110 en





r-i 00 n oo r rn

moo en ene


Ca H\Dro
in n Ni
****~C


ON r NM r r- I' -i r- sTN






















~~00 00




















O~ .NOO meeee


ooD0oONNmmmNo









MOOMONOmOOCMO












%00CC0 00%0- 0
%0 r. 00 00 r~. r. r~. r- r~. r~- r~- fl- 0


ma~O
CM) a
0* C
Lo cn 110 Lo i


0I 00 r~
1-1 -1
* *
CT^ D ^D


0w % "m r- r- I D
* . .O. .


mC oCm rI. CI
rn r cI '.0 r1-
mrmm o
MNMON0\


mCM wC 0I T- ui0 I-
wM oo 1 -1 C mI o
N lWOOC

* * * **r~r


o1mm
OoC M

11 1 N


OyO
cO r






, .

*o*o


ITr -4 r-
en -4 ry-

m U\) T\o








CM Co CM








m mm
Com


***C


Omm
I m


C00 CY 00










r00 C) CIT 00









Ir I 0 I











e eNNo'.





*****OQ


It-i
C'J u-i


HM0NN0NmOmmmmWWW40000000


r-M 1<10 i- C)
<3 \) 110 0 '-q
r- *c4 *a
T I'D%0 I


\O





























O
C14























Cr
rl-











C.,










-in
r-i





















%\


%0 co-- T\


17 IT -:T
\cOA
***lQ
MroO


110 110 r-A OD i-A
en i-I %.o i) i-4
* * *
'a- \ f- in
r~. o ^ r~ r~















o Lr co -".

oNM 0 r
-T 'O0 Lt -7, .T


Q\N 0

***lQ
st e try


0- Ltr~ r-3 rn-


Lr i Lin L t
=nooeo


0tf0L af


' 1o 0 -r '.- .0 i 0


r-'. C4 r~-'. r -: O0
NrQ o ~ Oo
* * * **)Q


rr O


mn om o mO


in .n M- 0 1-1 '"
r 0rQr ci NrQ


00 C cn ir cM 00

Cen n Cn .-i in









rN 00 r- 'T caj rc
4 c rm -C4 cJ N-4

-.7 7T -t -3 T -T


0C0
N'0

coin
r-.r-


a'-O
-1 r,


an o in -T7 0-.

N N- NO C 0








0 cn \' i r -



IT -T -T .T -T


-n e Ci-


-. coC
hh


-4-4'0


SCO 0- 0 Nr- CO 0-


o0 0 -T C 0 Lr).
o m 0so


Co o rn n oZ rN- T iU) 'o r-
'0N--N--N-NCo~goh

















W nC - -q wC C-4C m r- mOQCD '4\













r-r Dr -%D %0P-%0Lr lu0P 0V V l Z10 DV -r Lr) .0O0








m~~oom ~ e~~cm oor-~~m-
% i iv v 4 i: c L a m o a % 4 i r r T -
44 444 m mU cN CN C'41 0
m I z z V T z z I I T : L U I I .































V)~~~ ~ ~ ~ ~ C% C)"L) 0o % v)- 0r- 1rr 0 D 0C r-C -T%
% % % 0 V % )L o ) r)tr)








1 C C r 19- C q q uJ-,
CN -Tr- rq c a ri % r m T L C) ) 4raNCv r C 0



r-%D%0r-P-%0% 0 0 r)Lr %o oL U) l Ul % % U)% % V L) D r




m C 0 r mLn m .41r-0 D 0 n - D C L) -4C1




. . . . . . . I *
,T1 M CLr r L) r4 14C1 L) r L) 0 1001. r-ic4 ;




r-1. % r % U) n D D D D n n n n n ) ) D 0 r)U' -T r)Lr I






C1 L) o N :)-icn m r)r-ooc)c- m -, L) o %o o r-lM T r






r-Ir-l 1 C C4 4 N M M M M -. -T -T ) f)U) f)Lr




Full Text

PAGE 1

COMPARISON OF METHODOLOGIES FOR DETERMINATION OF CONTENT AND DIGESTIBILITY OF HEMICELLULOSE AND CELLULOSE IN TROPICAL GRASS HAYS BY SANDRA L. RUSSO A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1981

PAGE 2

To Allen, now its your turn.

PAGE 3

ACKNOWLEDGEMENTS The author would like to express her appreciation to Dr. G. 0. Mott for serving as chair and academic advisor for the program of study. The author is sincerely grateful to Dr. John E. Moore, director of the research, for guidance and support, for always having time and usually patience and for providing research facilities and equipment. Appreciation is extended to Drs. Donald Graetz, Joseph Conrad and 0. C. Ruelke, members of the supervisory committee, for their time and advice. The statistical assistance of Dr. Ramon Littell is gratefully acknowledged. Appreciation is extended to Dr. Murray Gaskins for putting his time and laboratory equipment at the author's disposal. Mr. John Funk and Dr. John Moore provided invaluable assistance at the Nutrition Lab in designing and setting up the experiments. Thanks are due to Drs. Donald Graetz, Ray Varnell, G. 0. Mott, Ken Quesenberry, C. Y. Ward, David Buchanan, Art Hansen and Elon Gilbert for providing financial assistance via a series of jobs, assistantships and contracts. The author appreciates the intellectual and non-intellectual support of fellow graduate students and staff at the Nutrition Lab. In particular, Carlos Chaves, Hari Hartadi and Steve Abrams laid the groundwork for the dissertation. ixx

PAGE 4

The weekly Social, Agricultural and Food Scientists meetings kept the mission of feeding the world's hungry through cooperation as well as research at the front of the author's thoughts and is most gratefully appreciated. Finally, for the continued support and sometimes, the sacrifices, of her husband, Allen Cook, the author is especially indebted. IV

PAGE 5

TABLE OF CONTENTS PAGE ACKNOWLEDGMENTS iii LIST OF TABLES vli LIST OF FIGURES ix ABSTRACT x CHAPTER I. INTRODUCTION 1 CHAPTER II. LITERATURE REVIEW 3 Forage Quality 3 Digestibility 4 Intake 5 Relationship Between Digestion and Intake 6 The Structure of Plant Cell Walls 7 Hemicellulose 9 Hemicellulose: Methodology 10 Alkali extraction 10 Acid hydrolysis 13 Hemicellulose: Structure 14 Hemicellulose in Plants 16 Hemicellulose Digestion in the Ruminant 20 In vitro studies 20 In vivo studies 23 CHAPTER III. CHEMICAL COMPOSITION OF TROPICAL GRASS HAYS. 26 Introduction 26 Experimental Procedures 27 Results and Discussion 31 Neutral detergent fiber 37 Holocellulose 40 Hemicellulose 42 Cellulose 46 Lignin 46 Summary 51 CHAPTER IV. IN VITRO AND IN VIVO DIGESTIBILITY OF TROPICAL GRASS HAYS 52

PAGE 6

Introduction 52 Experimental Procedures 54 Results and Discussion 55 In vitro digestibility 55 In vivo digestibility 59 Summary "2 CHAPTER V. PREDICTION OF FORAGE QUALITY FROM CHEMICAL AND IN VITRO ANALYSES ^5 Introduction 65 Experimental Procedure 66 Results and Discussion 67 Intake and digestibility 67 Prediction of forage quality from laboratory analyses 68 Conclusion 75 Summary 76 CHAPTER VI. GENERAL DISCUSSION 77 APPENDIX 83 LITERATURE CITED 99 BIOGRAPHICAL SKETCH 103 vx

PAGE 7

LIST OF TABLES TABLE PAGE 1 Abbreviations used in the dissertation 32 2 Analysis of variance of chemical composition of forty hays 34 3 Analysis of variance of method differences 35 4 Correlation coefficients (r) between variables obtained by two methods 36 5 Analysis of variance of ±n vitro data 56 6 Comparison of in vitro digestion in rumen fluid and in rumen fluid plus pepsin 57 7 Digestion coefficients of chemical fractions in vitro and In vivo 59 8 Chemical composition of feces 61 2 9 Relationship (r ) between in vivo and 2:11 vitro digestion coefficients for six hays 63 2 10 Relationship (r ) between in vivo values and laboratory analyses, obtained with thirty-nine different hays. . . 69 2 11 Relationship (r ) between ^ vivo values and ±n vitro analyses, obtained with thirty-nine different hays. . . 70 2 12 Coefficient of determination (R ) , probability (P) and standard error of the estimate (s . ) for the y X repression of in vivo intake values on several independent variables 72 2 13 Coefficient of determination (R ) , probability (P) and standard error of the estimate (s ) for the V • X regression of in vivo digestibility values on several independent variables 73 Vll

PAGE 8

APPENDIX TABLE PAGE 14 Chemical composition of individual hays 83 15 ^ vitro digestion of individual hays 85 16 Chemical composition and ±n vitro digestion of individual hays 87 17 Ln vivo intake and digestibility data 89 18 Means of neutral detergent fiber by age and hay ... 91 19 Means of holocellulose by age and hay 92 20 Means of hemicellulose by age and hay 93 21 Means of cellulose by age and hay 94 22 Means of lignin by age and hay 95 23 Analyses of variance of feces samples and ±n vivo digestibility 96 24 Least squares means of iji vivo intake per unit body weight (g/kg body weight) 97 25 Least squares means of in vivo digestibility data . . 98 Vlll

PAGE 9

LIST OF FIGURES FIGURE PAGE 1 Change in NDF content by weeks and hay 38 2 Change in NDF(H) content by weeks and hay 39 3 Change in holocellulose content by weeks and hay. . ^1 A Change in alkali-soluble hemicellulose by weeks and hay ^3 5 Change in Van Soest hemicellulose by weeks and hay. ^^ 6 Change in alkali-insoluble cellulose by weeks and hay '^7 7 Change in Van Soest cellulose by weeks and hay. . . ^° 8 Change in chlorite lignin by weeks and hay ^^ 9 Change in permanganate lignin by weeks and hay. . . ^^ ix

PAGE 10

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COMPARISON OF METHODOLOGIES FOR DETERMINATION OF CONTENT AND DIGESTIBILITY OF HEMICELLULOSE AND CELLULOSE IN TROPICAL GRASS HAYS By SANDRA L. RUSSO June 1981 Chairman: Dr. G. 0. Mott Major Department: Agronomy Experiments were conducted to: 1) fractionate cell walls of tropical grasses into: holocellulose, cellulose and hemicellulose; 2) determine in vitro and in vivo digestibility of cell wall components; and 3) study the relationship between these components and quality of forages, as measured by feeding hays to sheep. One cultivar of Cynodon dactylon (Coastcross-1) , two experimental lines of Digitaria species (X46-2 and X124-4) and two cultivars of Paspalum notatum (Argentine and Paraguay) cut at four ages (2, 4, 6 and 8 weeks of regrowth) were used in the experiments. Samples of each hay were analyzed for organic matter (OM) , crude protein (CP) and neutral detergent fiber (NDF) . NDF fractions (cellulose, hemicellulose and lignin) were determined by two methods: 1) conventional Van Soest detergent analyses

PAGE 11

(CELLV, HEMV, PLIG respectively) : and 2) classical fractionation using NaC10_ and alkali (CELL, HC, LIG respectively). There were significant differences between fractions obtained by the two methods. HC remained the same or decreased slightly with maturity but HEMV declined. CELL and CELLV showed similar patterns, increasing up to 6 weeks, then decreasing. Samples were fermented in rumen fluid for 48 hr and digested in acid-pepsin for 48 hr or fermented in rumen fluid only for 48 hr and residual NDF was fractionated as above. In vitro digestibility of NDF and HC was increased by further digestion in acidpepsin while CELL digestibility showed minimal change. CELL digestibility was generally higher than HC digestibility. Digestion coefficients for each fraction were similar ±n vivo and in vitro (r^ = .79 to .92). NDF, CP and PLIG were the best single chemical analyses for prediction of intake and digestibility, but any in vitro analysis gave better estimates of in vivo digestibility. Two and three variable models improved prediction of organic matter intake (OMI) and digestible OMI. Hemicellulose obtained by different methods was not the same. Digestion of cell walls in acid-pepsin was due to breakdown of hemicellulose and possibly, dissolution of lignin-hemicellulose bonds. Further studies on chemical bonding, physical degradability and chemical structure of the cell wall appear necessary before being able to predict animal performance from laboratory analyses . XX

PAGE 12

CHAPTER I INTRODUCTION The world food shortage is most critical in that portion of the world known as the tropics, which is faced with rising populations and diminishing food supplies. The limited availability of land suitable for arable cropping has focused attention upon the grasslands which occupy fifty percent of the earth's land surface. These lands are able to provide food for people by the use of ruminant animals to convert pasture into meat and milk. The structural carbohydrates in pastures and forages are an important source of energy to ruminants and, in turn, to people. The amount and proportions of structural carbohydrates in the plant cell wall determine the degree to which the ruminant can utilize this energy. Changes in amounts and types of cell wall constituents are responsible for differences in nutritive value of forages. The cell wall polysaccharides are comprised of fibers of organized, crystalline cellulose in a matrix of hemicelluloses which are associated physically and chemically with lignin (Jones, 1976). Lignin influences digestibility of forages and has been considered the primary structural inhibitor of quality in tropical grasses (Moore and Mott, 1973). Cellulose and hemicellulose fractions are also responsible for variations in digestibility, since they contain the largest proportion of the indigestible substances in the cell wall (Van Soest, 1975).

PAGE 13

The role, structure and composition of hemicellulose(s) vis a vis forage quality and nutritive value is not well understood. Hemicellulose is often defined as the resultant of a certain set of operations and it has been convenient to estimate hemicellulose as the difference between acid detergent fiber (ADF) and neutral detergent fiber (NDF) . Acid extraction may not, however, remove material that is hemicellulosic in nature (Bailey and Ulyatt, 1970; Morrison, 1980). The entire hemicellulosic fraction has been treated as a unit since no nutritionally meaningful fractionation has been found and since hemicellulose digestibility is so closely related to that of cellulose (Van Soest, 1975). Questions arise, however, as to whether this grouping together of structural carbohydrates is biologically rational, given the complexity of living plants and of ruminant digestion. The objectives of this research were: 1) to fractionate the cell walls of tropical grasses into three components: holocellulose, cellulose and hemicellulose; 2) to determine the in vitro and in vivo digestibility of the three components; and 3) to study the relationship between these components and the quality of the forages in vivo.

PAGE 14

CHAPTER II LITERATURE REVIEW Forage Quality The best measure of the nutritive value of a forage is animal performance. Moore and Mott (1973) have defined forage quality as output per animal, being a function of voluntary intake and digestibility when the forage is fed ad_ libitum as the sole source of feed. Intake and digestibility may be limited by the rate and extent of forage digestion in the rumen, which, in turn, are related to the amount of cell wall constituents (CWC) . Under the best of conditions, all forages would eventually be tested in long-term production trials on pasture (Mott and Moore, 1970) having been preceded by a few years of agronomic and laboratory evaluations. It is often not possible for researchers to conduct pasture studies. Instead, trials with penned or caged animals are used to measure voluntary intake and digestibility. Even these animal trials may be beyond the means of some researchers who have to rely exclusively on laboratory evaluation of forages. Fortunately, enough data has been accumulated to show high correlations between some laboratory methods and in vivo performance, e.g., the two stage in vitro rumen digestion procedure (Tilley and Terry, 1963), and digestibility. Unfortunately, a laboratory procedure that predicts voluntary intake with acceptable accuracy has not yet been developed.

PAGE 15

Thus, there is a continuum ranging from what would be the ideal way to measure forage quality (animal response) to a vast array of ways to predict forage quality in the laboratory which may or may not be acceptable under all conditions. Digestibility A number of laboratory methods to estimate in vivo digestibility have been devised. However, if only one method is to be used to predict digestibility over a wide range of forages, then the method to choose is the two-stage in vitro digestion procedure (Oh et al., 1966; Marten and Barnes, 1980). Digestibility is very important in determining the quality of a forage and is more a function of plant factors than of animal factors. It is now accepted that as a plant matures, its digestibility decreases due to increasing lignification and changes in plant morphology. The cell contents (CC) of forages are considered to be completely digestible while the CWC show variable digestibilities. Riewe and Lippke (1970) found that tropical species accumulate dry matter more rapidly which can be directly associated with an increase in CWC. While temperate and tropical grasses have similar amounts of CWC at early stages of growth, the amount of CWC and its composition changes more rapidly with maturation in tropical species. This increase in CWC may or may not decrease digestibility in tropical grasses more than such an increase would in temperate grasses. Minson and McLeod (1970) showed temperate species to be 12.8 percentage units higher in IVDMD than tropical species; however, Russo

PAGE 16

(1976) did not find differences of such magnitude despite higher CWC. Tessema (1972) noted that the overall rate of decline in digestibility was more rapid with tropical grasses than with temperate grasses, that hemicellulose fluctuated considerably in the tropical grasses and that tropical grasses had higher concentrations of lignin. Moore and Mott (1973) have suggested that the apparent lower quality of tropical grasses may be due to lower protein values, which will depress intake, and higher fiber values, which will depress digestibility. With respect to ruminant nutrition, the detergent analysis system of Van Soest (1967) which separates forages into CC and CWC is useful for forage quality analysis. The CC are soluble and almost totally available to the ruminant while the CWC are completely insoluble, with their digestibility depending on microbial action. The CWC appear to have the most influence on the digestibility and intake of forages. The higher CWC in tropical grasses places increased emphasis on digestibility, especially with respect to its influence on voluntary intake. Intake Intake is controlled by two mechanisms: distention and metabolic. In forage-fed ruminants, the distention mechanism exerts the major effect on intake. Many other factors, of course, affect voluntary intake and have been reviewed by Capote (1975) and Golding (1973).

PAGE 17

Crampton et al. (1960), Conrad (1966) and others have proposed the following concepts relating to the distention theory in ruminants: (1) probably a specific degree of rumen load reduction is the major determinant of hunger; (2) the rates of forage cellulose and hemicellulose degradation are directly related to the rate at which the rumen load is reduced; and (3) the time it takes for the rumen load to be reduced (when hunger recurs) is characteristic of the specific forage involved. Most researchers now agree that the rumen is where the distention control over forage is exerted. Distention mechanism is determined by (1) rumen fill and (2) retention time (Weston, 1966; Thornton and Minson, 1972), Van Soest's "hotel theory" (1975) suggests that the plant cell walls, not the dry matter of the forage, are responsible for gut fill and only when the cell walls are degraded or move out of the rumen, is gut fill reduced. Retention time is affected by the rates of digestion and passage (Jones and Bailey, 1974; Waldo et al., 1972). Relationship Between Digestion and Intake The detergent systems of Van Soest separate the fibrous portion of a forage into cellulose, hemicellulose and lignin. Other classification systems exist: Minson (1976) separated the CWC into a potentially digestible fraction and an indigestible fraction, while work by Abrams (1980) suggests there may be at least two types of digestible cell walls. How these cell walls are utilized by an animal is determined primarily by their rate of passage through the digestive tract.

PAGE 18

Chemical composition and structure of the cell walls are probably the most important factors influencing rates of digestion and passage (Akin et al., 1974; Van Soest, 1965). The chemical bonds holding the cell walls together as well as the structure of each cell wall component (lignin, cellulose and hemicellulose) needs to be elucidated. Lignin, for example, may provide a physical barrier to the breakdown of the cell walls (Dekker, 1976). The rate of breakdown is regulated by the organizational structure of the cell walls (Akin et al., 1974; de la Torre, 1974). Intake and digestibility can be seen then, to be linked together with respect to forage quality. Some alternative definitions of forage quality use the combination of intake and digestibility, such as voluntary intake of digestible energy (Heaney, 1970). The goal is to predict digestible nutrient intake, i.e., animal performance, without conducting long-term pasture production trials. Few attempts have been successful, especially with tropical grasses. As mentioned earlier, laboratory methods currently available can predict forage digestibility but not voluntary intake. Characterization of the chemical structure of the plant cell wall, especially in tropical grasses, would help our understanding of the factors affecting voluntary intake. The Structure of Plant Cell Walls The plant cell wall is of two types, a thin primary wall and a thicker secondary wall. The primary wall is laid down by undifferentiated cells that are still growing. The primary wall is transformed

PAGE 19

into a secondary wall after the cell has stopped growing. The primary cell walls of many higher plants appear to have very similar structures (Talmadge et al., 1973) but this is not true of secondary walls . Every living cell, whether plant or animal, is surrounded by a complex membrane made up of lipids and proteins called the plasmalemma. In almost all plant cells, the plasmalemma forms a cell wall which is rigid or semi-rigid and is chemically distinct in that the structurally important components are polysaccharides (Preston, 1974). Cellulose, hemicellulose, pectic polysaccharide, structural protein and lignin have been identified as the major components of the plant cell wall. Some of these components, e.g. lignin, are not polysaccharides, nevertheless, the essential and sudden change at the plasmalemma surface is from a basically lipo-protein to a basically polysaccharide formation (Preston, 1974). Usually only one of these polysaccharides is crystalline and cellulose predominates. In primary cell walls, the crystalline cellulose fibrils form a core surrounded by a non-crystalline region consisting of other cellulose molecules, non-cellulosic polysaccharides and glycoprotein (Talmadge et al. , 1973). In secondary walls, the cellulose fibrils are aggregated into ropelike structures called microfibrils. The non-cellulosic polysaccharides making up the bulk of the primary cell wall have been defined by their presence in chemical fractions. The two major non-cellulosic fractions are the pectic polysaccharides obtained by extracting cell walls with boiling

PAGE 20

9 water, EDTA, or dilute acid (Aspinall, 1970) and the hemicelluloses solubilized by the subsequent extraction of the same walls with alkali (Whistler and Richards, 1970). In the secondary wall, also. It appears that the extractable non-cellulosic wall compounds surrounding the microfibrils are hemicellulosic (Preston, 1974) . But even further, some of the hemicelluloses cannot be removed without degradation of the microfibrils and are therefore integrally mixed with them (Preston, 1974). It has been suggested (Bauer et al., 1973) that the structural function of the hemicellulosic polysaccharides is to interconnect the cellulose fibrils and the pectic polysaccharides of the plant cell wall, and that this function is based on the ability of the hemicellulosic polysaccharides to bind noncovalently to cellulose and to bind covalently, through glycosidic bonds at their reducing ends, to the pectic polysaccharides. Hemicellulose Hemicelluloses are often implicitly or explicitly defined as the cell wall and intercellular polysaccharides that can be extracted by alkali from higher land plant tissues that are, or were, lignified (Wilkie, 1979). Hemicellulose is often extended to include certain carbohydrates found in cereal endosperm. Jermyn (1955) wrote "it cannot be too strongly emphasized that 'cellulose' and 'hemicellulose' are normally determined as the resultants of certain sets of operations, rather than as chemically defined species." (p . 197) Hemicellulose has, thus, variable definitions depending on the methods or procedures used to isolate it. A more liberal definition of hemicellulose

PAGE 21

10 refers to all of the types of polysaccharides (found in plants) other than celluloses, starches and fructans (Wilkie, 1979). It is necessary to point out that the majority of literature on hemicellulose and its relationship to forage quality and ruminant nutrition deals with the Van Soest hemicellulose obtained by subtracting ADF from NDF. Work as early as 1970 (Bailey and Ulyatt) points out the problems inherent in this definition of hemicellulose. Discussion of methodology differences follows. To avoid confusion. Van Soest hemicellulose will be distinguished from alkali-soluble non-endospermic hemicellulose (as isolated by classical methodology) . Hemicellulose; Methodology Hemicellulose was first isolated by Schultze in 1891 (Wilkie, 1979) using dilute alkali to extract polysaccharides from plant tissues. Work on hemicellulose, thereafter, although becoming more sophisticated, continued to use alkali to separate the compound from plant tissue. The continued search for a simpler, more rapid procedure led, however, to a second extraction scheme using acid hydrolysis, either on delignified plant tissue (Routley and Sullivan, 1958) or on whole plant tissue (Van Soest, 1965), Alkali extraction and acid hydrolysis are two considerably different methods of arriving at a presumed similar result. Closer examination of these two general methodologies appears warranted. Alkali Extraction In grasses, some of the hemicellulose is covalently bonded to lignin (Morrison, 1974). Earlier, the relationship of lignin and

PAGE 22

11 hemicellulose was not well understood and hemicelluloses were extracted from lignified tissues of grasses by direct treatment with alkali (Wilkie, 1979). However, it has now become common practice to delignify the plant tissue with sodium chlorite and acetic acid according to the procedure first outlined by Whistler et al. in 1948. The delignified plant tissue, termed holocellulose, is primarily composed of hemicellulose and cellulose. The holocellulose is then treated with dilute alkali under nitrogen to extract the hemicelluloses. It may be possible to extract some of the hemicellulose from holocellulose with hot or cold water. The alkali solutions, either sodium or potassium hydroxide, may range in concentration from 4 to 24% w/v. Dilute alkali extracts only part of the hemicellulose. If exhaustive extractions are required, the concentrations of alkali can be increased from a low concentration to a higher concentration to extract more of the hemicellulose (Wilkie, 1979). Nevertheless, even with this exhaustive extraction, not all of the hemicellulose will be accounted for (Gaillard, 1958). With filtration or centrifugation, a residue and a supernatant are obtained; the supernatant containing the hemicellulosic materials. With some diligence, as many as three fractions can be derived from the supernatant. Acidification of the supernatant with acetic acid at 0-2 C will precipitate the A fraction (Gordon and Gaillard, 1976) which is composed primarily of long chain xylans with small amounts of arabinose and uronic acid (Bailey, 1973). The polysaccharides remaining in solution can be precipitated by ethanol (Gordon and

PAGE 23

12 Gaillard, 1976) or recovered by dialysis and freeze-drying (Bailey, 1973) to obtain the B fraction. The B fraction contains smaller molecular weight xylans with many side chains and more complex molecules made up of galactose, glucose and rhamnose (Bailey, 1973). These are sometimes further broken down into linear B and branched B fractions based on their separation by iodine precipitation (Gaillard, 1965). The ethanol in this supernatant can be evaporated and a C fraction recovered by dialysis for 5-7 days, followed by freeze-drying. The C fraction has not been well studied but is composed primarily of lignin, arabinose, galactose, glucose and xylose (Gordon and Gaillard, 1976). In addition to iodine precipitation, individual monoor polysaccharides may be isolated from these fractions by copper precipitation, gel filtration or quaternary ammonium salt precipitation (Bailey, 1973) . The sugars thus isolated can then be identified by any number of methods ranging from the colorimetric (Dubois et al., 1956) to chromatographic (paper, thin-layer, gas or high performance liquid). It becomes apparent, then, that examination of heraicellulose and its components can very rapidly become a major undertaking, particularly when it is desired to elucidate exact amount or structures of components. Work along these lines thus falls on a few hardy souls upon whom the rest of the researchers become dependent for answers. It is also apparent that such procedures have little value in a situation where a large number of samples need to be analyzed.

PAGE 24

13 Acid Hydrolysis A fair approximation of the hemicellulosic material in a plant can be obtained by treating depectinated plant tissue with boiling N acid for 2-3 hours (Bailey, 1973), hot water (Routley and Sullivan, 1958) or cold 72% sulfuric acid overnight followed by 2 hours boiling (Sullivan, 1966). The sugars in the hydrolysate may then be identified. With the advent of Van Soest's detergent analysis scheme, it became an easy matter to isolate and work with the cell wall and its constituents. The use of neutral detergent on plant tissue leaves a residue which is essentially nitrogen-free and is composed of lignin and the structural polysaccharides, cellulose and hemicellulose. The use of acid detergent on a forage was originally intended as a preparatory procedure for lignin (Van Soest, 1975) but has commonly been used to obtain cellulose and lignin values by using the acid detergent lignin (ADL) procedure (Van Soest and Wine, 1968) and to obtain a hemicellulose value by subtracting ADF from NDF. The Van Soest procedures are widely used because they are rapid, simple and not extremely expensive. The isolation of the cell walls by neutral detergent is a chemically, as well as structurally, rational procedure and is a convenient starting point for further work on cell wall constituents. Daughtry (1976) for example, combined the use of neutral detergent for cell wall preparation with boiling in 1 N H-SO, to obtain what was termed a hemicellulose hydrolysate. It is with the other Van Soest procedures, ADF and ADL, that problems arise. ADF minus lignin and silica, for example, is a close

PAGE 25

14 estimate of cellulose (Van Soest and Wine, 1968) but has been found to be contaminated with hemicellulosic residues (Bailey and Ulyatt, 1970; Morrison, 1980). ADL, on the other hand, gives lignin values quite different from lignin values obtained on the same plant sample by other methods. Lignin, like hemicellulose, is difficult to define and becomes defined by the procedure used to isolate it. Many, and sometimes imaginative, adaptations have been made on the Van Soest procedures in an attempt to get fractions that behave in a uniform manner. However, since the cell wall fractions are not uniform, it is unlikely that a simple evaluation procedure will isolate a recognizably uniform entity in all forages at all ages under all environmental conditions. Recalling that hemicellulose becomes defined by the procedure used to isolate it, these two different approaches — solubility in alkali and hydrolysis in acid — can be seen to, in effect, act as a check on each other. The former, by elucidating the chemistry and structure of hemicellulose piece by piece, acts as a baseline against which the results of the more rapid, latter methodology can be checked for accuracy. It is with a very critical eye that information on "hemicellulose" should be examined and compared because, as has been seen, these hemicelluloses may or may not be representative of the hemicelluloses present in plants. Hemicellulose; Structure Hemicellulose was originally thought to be a precursor of cellulose, hence the name hemi cellulose. While the reasoning has been

PAGE 26

15 proven wrong, the name persists. Cellulose is a linear polymer of 3-1-4 linked D-glucose units; repeating unit 4-0-3-D-glucopyranosyl-Dglucopyranose, cellobiose (Bailey, 1973). Hemicellulose is composed of two major groups of polysaccharides: pentosans and non-cellulosic hexosans. Bailey (1973) reviewed the literature on chemical composition of hemicellulose from which the following information is taken. The pentosans are primarily based on xylose. They appear to be linear chains of 6-1-4 linked D-xylopyranose units; repeating unit 4-0-B-Dxylanopyranosyl-D xylopyranose, xylobiose. To these chains are attached L-arabinose, D-glucuronic acid or its 4-0-methyl ether, D-galactose and possibly D-glucose. These sugars are usually single side chains but may be connecting points for further branching, particularly via L-arabinose. Here also, is where the linkages with lignin polymers occur via ferulic acid, other phenolic acids or their esters (Wilkie, 1979; Morrison, 1980). The hexosans in forage grasses are primarily mannans, either glucomannans or galactoglucomannans. These are linear chains of 3-1-4 linked D-glucose and D-mannose. The glucomannans are usually free of other sugars, the galactoglucomannans have a-linked galactose side chains. There has also been some evidence (Fraser and Wilkie, 1971; Wilkie and Woo, 1976) of the presence of linear chains of 3-linked D-glucose units which are not cellulosic. The glucans have both 3-1-3 and 3-1-4 linkages. The ratio of the 1-3 to 1-4 linkages in the glucans in any one tissue falls as the plant matures (Buchala et al. , 1972) . A more thorough discussion of the structural

PAGE 27

16 features in hemicelluloses can be found in a review chapter by Wilkie (1979). Albersheim and coworkers (Bauer et al., 1973; Keegstra et al., 1973; Talmadge et al., 1973) working with dicotyledonous plants have developed a cell wall model. The cellulose fibers are linked together by other polysaccharides, in particular, hemicellulosic xyloglucans which completely cover the cellulose fibrils (Albersheim, 1978) . Their preliminary work on grasses suggests that a model of the cell wall for monocotyledonous plants will be similar to that for dicots. The dominant hemicellulosic polysaccharide in primary cell walls of monocots is an arabinoxylan rather than a xyloglucan (Albersheim, 1978). A modified cell wall model (Bailey et al., 1976) proposes that much of the heraicellulose and all of the pectin is not linked to other CWC and only some of the hemicellulose is bonded to glycoprotein and cellulose. They also point out that cellulose is not the only crystalline structure and that hemicellulosic xylan is capable of forming a crystalline structure. Hemicellulose in Plants Hemicelluloses comprise from 10 30% of the total dry matter in temperate and tropical grasses (Bailey, 1973) and from 30 60% of the cell wall in tropical grasses (Hartadi, 1980; Bailey and Connor, 1973; Russo et al., 1981). While it is difficult to compare heraicellulose values reported by various researchers due to differences in definitions and procedures, it appears that grasses contain more hemicellulose than legumes (Bailey, 1973). Further, the

PAGE 28

17 hemicellulose in grasses is different structurally from legumes. Gaillard (1965) reported that the hemicellulose A fraction of grasses was primarily an arabinoxylan while in legumes it was a xylan with some uronic acids. Legumes seem to have more uronic acids than grasses and some additional sugars such as rhamnose which are not present in grasses (Gaillard, 1965; Collings and Yokoyama, 1979). Lignin has long been noted to be higher in legumes while seemingly having less effect on the digestibility of legumes than it does on grasses. Gordon and Gaillard (1976) suggest that this difference could be due to the different types of hemicellulose present in grasses and legumes and the subsequent differences in lignin-carbohydrate bonds. The hemicellulose content of a plant changes with growth and maturity. The most marked change in hemicellulose composition in a grass is in the growth period before changing from a vegetative to a floral morphology (Bailey and Connor, 1973). It is expected that hemicellulose content will increase with maturation. Bailey (1973) reported on several studies with temperate grasses where the increase in hemicellulose with age was associated with in increase in stem tissue. For tropical grasses, however, seasonal changes did not appear so great, the hemicellulose content remaining about the same in spite of increasing cell wall content and increasing proportion of stems. The greatest change in hemicellulose content occurs during growth. The leaves of speargrass (Heteropogon contortus ) at vegetative, early seed and dormant stages had 23, 34 and 36% hemicellulose

PAGE 29

18 (of the total dry matter); the stem in early-seed and dormant stages had 37 and 42% hemicellulose (Blake and Richards, 1971). Reid and Wilkie (1969a) found that in older, mature plants, there was little apparent change in the hemicellulose of either leaves or stems. Mature stems of rye, barley, orchardgrass, oats, ryegrass, timothy and wheat had values of 23 28% for hemicellulose (Wilkie, 1979) while mature leaves of oats, barley, wheat, kikuyugrass and pampasgrass had values of 21 37% for hemicellulose. Monosaccharide analysis (Buchala and Wilkie, 1973) of the hemicellulose fraction found that with increasing maturity, there was an increase in amount of D-xylose in all tissues (stems and leaves), L-arabinose and D-glucose decreased and D-galactose remained constant or varied slightly. The hemicellulose changes with growth may be linked to formation of the secondary wall. When a plant cell wall stops elongating, it produces a thicker secondary wall and at this stage, lignin begins to be synthesized into the wall along with continuing synthesis of structural carbohydrates found in the primary wall (Morrison, 1979). It is at this stage that lignin and hemicellulose become linked to each other. With respect to cellulose, its structure does not change during growth. Hemicellulose, on the other hand, does change its structure during growth and with increasing lignification (Morrison, 1979). These changes have sometimes unexpected effects on digestibility, which will be discussed in a later section. Bailey (1973) reported that in temperate grasses the cellulose content is greater than or equal to the hemicellulose content but.

PAGE 30

19 in tropical grasses, the hemicellulose content is greater than or equal to the cellulose content. A further difference between temperate and tropical grasses is in monosaccharides present in the hemicellulose fraction. Ojima and Isawa (1968) showed that two tropical grasses ( Paspalum notatum and Cynodon dactylon) had xylose: arabinose: glucose: galactose ratios of 50:15:30:2 while two temperate grasses ( Phleum pratense and Lolium perenne ) had ratios of the same sugars of 74:12:11:2. It is difficult to separate out environmental effects. The possibility exists that the differences in CWC of temperate and tropical grasses may be due to a temperature effect (Cooper and Tainton, 1968), Anatomical and photosynthetic differences do exist. Wilson and Ford (1971) studied three grasses ( Panicum maximum , Setaria sphacelata and Lolium perenne ) under a controlled environment and found that the effect of increasing temperature on CWC is not linear but passes through a maximum; this maximum may be due to hemicellulose rather than cellulose differences (Bailey, 1973). Reid and Wilkie (1969b) found little difference in monosaccharide composition of hemicelluloses of oat plants grown indoors or out, in light or in dark. Hemicellulose content and composition are not constant and are apparently influenced by environmental conditions. Almost every study on hemicellulose contains values for cultivars, species and genera and many attempts have been made to relate hemicellulose values to intraspecif ic or intrageneric variations. The results can be very confusing and probably of little value. As Wilkie pointed out (1979), "it is not possible to determine invariant.

PAGE 31

20 quantitative values for total hemicelluloses . . • and there are no self-evident parameters upon which to base such quantitative comparisons, particularly at the taxa level." (p. 251) Fraser and Wilkie (1971) re-emphasized the danger of comparative studies, the danger arising when the procedure used in isolating materials for study is either undesirably selective or inadequately reproducible. With so many hemicellulose procedures in use, it appears that comparative studies should be limited to within locations or laboratories. Hemicellulose Digestion in the Ruminant Ruminants obtain a great portion of their dietary energy from the digestion of cell wall polysaccharides when they are fed forage diets. The digestibility of the CWC is very crucial in determining the quality of the forage. Intake, too, is dependent on the breakdown of the CWC in the rumen. This digestion involves initially enzymatic hydrolysis by various rumen microorganisms to release sugars. These sugars can then be fermented to volatile fatty acids. It is at the initial stages of hydrolysis of the various cell wall polysaccharides that differences occur and when such properties as degree of lignif ication (Bailey et al., 1976) or chemical structure of the polysaccharides affect the digestion of the CWC. In vitro Studies Hemicellulose and its component monosaccharides can be separated from plant tissue by chemical methods. Earlier ^ vitro work on hemicellulose degradation in the rumen used these isolated fractions to determine digestibility of hemicellulose. Gaillard (1965) found

PAGE 32

21 that legume hemicellulose was more resistant to digestion than grass hemicellulose; of the linear A and B fractions, A was more resistant; and of the B fractions, the branched B fraction was more resistant to digestion than the linear B. Sullivan (1966) also showed that grass and legume hemicelluloses differed in their digestibilities and suggested that it may be due to a difference in the chemical heterogeneity of the xylans. Waite and others (1964) found that during growth in timothy, ryegrass and orchardgrass, the xylan fraction containing the uronic acids became less digestible. Dekker and Richards (1973) found the glycans in hemicellulose to be more digestible than xylan or uronic acid. Gaillard (1965), Waite et al., (1964) and Bailey (1967) working with temperate grasses, found the pectins, arabinans and galactans to be extensively digested in young plants but incompletely digested in older plants or in hay. They also concluded that xylans are incompletely digested in mature plants and hay. Other work has shown that rather than the chemical heterogeneity of the hemicellulose being responsible for lower digestion, that the incomplete digestion was due to the presence of lignin. Beveridge and Richards (1975) studying speargrass, showed no significant difference between the hemicellulose composition of fresh samples and of samples digested for 72 hrs in the rumen. In their earlier work (Dekker, Richards and Playne, 1972) the hypothesis had been that branching would cause resistance to digestion but since there were no differences between the original and "resistant" hemicelluloses in the later study, they attributed the poorer digestibility

PAGE 33

22 to lignification. Bailey and MacRae (1970) isolated "digestionresistant" hemicelluloses which were capable of further digestion after they had been delignified with NaC10„. Dehority et al. (1962) delignified timothy, alfalfa and orchardgrass by ball-milling and found that hemicellulose fermentation increased. The increases became greater with maturity and, again, they offered the theory that lignin provided a physical barrier preventing degradation of the hemicellulose molecules. In studies using isolated strains or species of rumen hemicellulolytic bacteria, Dehority (1973) found that fermentation of hemicellulose occurred in two stages. The first stage involved one species of bacteria which degraded or depolymerized the hemicellulose but did not utilize it. In the second stage, another type of bacteria which had limited degrading abilities was able to utilize hemicellulose that had been degraded by the first group. Dehority (1973) also found that cellulo lytic bacteria could degrade hemicellulose. These synergistic systems may be oversimplifications but the apparent specificity of the rumen bacteria provides another explanation for variabilities in hemicellulose digestion. Another avenue of iii vitro studies is the use of depectinated cell walls and mixtures of mold cellulase and hemicellulase to study the patterns of hemicellulose digestion. With the enzyme systems, hemicellulose hydrolysis was shown to differ between ryegrass varieties and the extent of hydrolysis decreased with maturity (Bailey and Jones, 1971; 1973). Jones and Bailey (1972) found that prolonged drying did not affect the rate or extent of enzymatic hydrolysis.

PAGE 34

23 Studies on isolated hemicelluloses, its components or enzymes from isolated rumen microorganisms give only a limited idea of what is going on in the rumen. More recent research has been on total plant hemicellulose and ^^ vivo studies of hemicellulose digestion in the rumen. In vivo Studies Beever et al. (1971) studied the sites of carbohydrate digestion by fitting sheep with rumen cannulas and re-entrant cannulas in the proximal duodenum and terminal ileum. Their results showed that 90% of the cellulose and about 70% of the hemicellulose (in ryegrass) was digested before entering the small intestine. Very little hemicellulose digestion occurred in the small intestine itself but approximately 30% of the hemicellulose was digested in the cecum and colon. These values are averages for fresh, dried and ensiled ryegrass. Other experiments with sheep also showed a significant post-ruminal digestion of hemicellulose especially when the forage was ground and pelleted (Thomson et al., 1972). Van't Klooster and Gaillard (1976) showed that distal to the duodenum in dairy cows hardly any digestion of hemicellulose occurred on diets of hay, grass and/or concentrates. Digestion of cellulose was almost completed before the small intestine. In a review of several digestion studies, Ulyatt et al. (1975) found that the proportion of hemicellulose digested in the large intestine was higher than that of cellulose. An analysis of these and other results (Bailey et al., 1976) showed that in forages as the hemicellulose increases, the percent digested in the rumen

PAGE 35

24 increases. The differences noted in all of these studies in the contribution of the lower gastrointestinal tract to digestion of hemicellulose may be due to species differences between cattle and sheep, to the type of feed used or to differences in techniques. Overall, the in vivo studies show that hemicellulose digestion decreases with maturity and increasing lignif ication and that hemicellulose is less digestible than cellulose. Models for cell wall carbohydrates and their fate in the ruminant animal (Smith et al., 1971; McLeod and Minson, 1974; Abrams, 1980) suggest the existence of at least one digestible cell wall fraction and one indigestible fraction, generally accounted for by the presence or absence of lignin. Morrison (1979) discusses the physical theory vs. the chemical theory. The physical theory proposes that the lignin and hemicellulose form a complex which has a "cage" effect; in young tissue, the complex is not fully developed and digestion is easily accomplished by the rumen microbes but in older tissue, the "bars" of the "cage" are closer together and enzyme access is restricted. The chemical theory suggests that in young tissue, lignification is at a minimum and hemicelluloses are not very complex; with maturity, complexity of the structure increases. Morrison concluded that both theories may be involved; i.e., in young grass, the proportion of side chains is the major factor controlling digestion while with older tissue the lignin content is the dominant factor. There are still many gaps in our knowledge of hemicellulose in the plant and its fate in the ruminant, especially with respect to the tropical grasses. The literature on hemicellulose digestibility

PAGE 36

25 is often contradictory but there is the possibility that significant hemicellulose digestion occurs after the rumen and that animal performance is related to the digestibility of hemicellulose. Because of the high hemicellulose content of tropical grasses, the acid-soluble hemicellulose from the Van Soest method may not be an acceptable estimate of the hemicellulose in these plants. An alternate method of measuring hemicellulose is by alkali extraction from plant cell walls. The hemicellulose obtained in this manner may be more representative of the hemicellulose in the plant than the hemicellulose obtained by the Van Soest analysis. Alkali soluble hemicellulose could be a major factor influencing forage quality in tropical grasses. More information on the hemicellulose fraction of tropical grasses may lead to a better understanding of the factors influencing quality. This dissertation investigates cell wall content and digestibility in tropical grasses, with special emphasis on hemicellulose.

PAGE 37

CHAPTER III CHEMICAL COMPOSITION OF TROPICAL GRASS HAYS Introduction While the present official method of forage analysis is the proximate analysis system first developed in the late 1800' s, the Van Soest (1965) system which partitions forages into cell contents (CC) and cell wall constituents (CWC) is now extensively used for evaluation of forages. Use of neutral detergent reagent on a forage leaves a residue which is essentially nitrogen-free and is composed of the structural components of the plant, the CWC (Van Soest, 1965; Van Soest and Wine, 1968). The CC, which were removed by the neutral detergent reagent, contain the sugars, starches, lipids and proteins. The CC are essentially digestible by the ruminant while the CWC have incomplete and differential digestibilities (Van Soest, 1975; Abrams, 1980). The Van Soest system of forage analysis is biologically rational and it has been recommended for grading hays by the American Forage and Grassland Council (AFGC) (Rohweder et al,, 1976; Barnes and Marten, 1979). Nevertheless, in spite of the widespread acceptance and use of the Van Soest system, problems still remain, particularly with respect to tropical and subtropical grasses (Moore and Mott, 1973). Due to physiological and photosynthetic differences and environmental influences, tropical grasses are higher in CWC than temperate 26

PAGE 38

27 grasses (Cooper and Tainton, 1968; Wilson and Ford, 1971; Moore and Mott, 1973). Prediction equations such as proposed by the AFGC rank the tropical grasses very low due to their high CWC, yet animal performance does not necessarily reflect the predicted low quality of these grasses. Due to these anomalies, questions arose as to whether the Van Soest method, which had been developed on temperate forages, was suitable for analysis of tropical grasses, l/hile the separation of CC and CWC appears rational (Van Soest, 1975), the further fractionation of the CWC using acid detergent reagent to obtain cellulose, hemicellulose and lignin, warrants further examination. The breakdown of the CWC by the acid detergent reagents appears incongruent with earlier, classical work on hemicellulose, for example, which used alkaline conditions to fractionate the CWC. Resolution of the differences between the Van Soest hemicellulose and hemicellulose obtained via classical methodology may lead to a clearer understanding of the factors affecting forage quality in tropical grasses. The objective of this study was to compare the cell wall fractions of tropical grass hays obtained by the Van Soest analysis with cell wall fractions obtained by classical methodology, involving an alkali extraction procedure. Experimental Procedures One cultivar of Cynodon dactylon ('Coastcross-1' bermudagrass) , two experimental lines of Digltaria species (X124-4 and X46-2) and two cultivars of Paspalum notatum ('Paraguay' and 'Argentine'

PAGE 39

28 bahiagrass) grown in an upland sandy soil in north-central Florida during the spring-summer were used in this experiment. At the beginning of the growing season a complete fertilizer was applied to provide 72 kg N, 72 kg P«0 and 72 kg K^O/ha. Another nitrogen fertilization at a rate of 66 kg N/ha was applied during July. All plots were staged during July and August by mowing and removing the growth; thereafter, two field replications of the 2, 4, 6 and 8 week regrowths were harvested and artificially dried in a portable wagon at approximately 50 C. The dry hays were stored in woven bags and chopped prior to feeding. Voluntary intake and digestibility were determined simultaneously by feeding the grasses ^ libitum to mature wethers. Each hay was fed to at least two (and often three) wethers in separate trials, with no wether receiving the same hay twice. Animals were treated for internal parasites, fed a standard hay and weighed before each trial. The seven day collection period was preceded by a fourteen day preliminary period. Each animal was housed in a wooden pen with slotted floors and had access to water, trace mineralized salt and def lourinated phosphate at all times. Ad libitum feeding was achieved by allowing a feed refusal (orts) of 200-300 g of hay daily. Daily samples were taken of the hay offered during the trial. The orts were removed and weighed daily. Feces were collected in canvas bags, weighed daily and 20% kept for further analyses. The waste (hay dropped through the floor of the cages) was weighed. All samples were dried at 50 C. Each sample was pooled by animal, ground to pass a A mm screen in a Wiley mill, mixed and a

PAGE 40

29 representative portion reground through a 1 mm screen. A composite hay sample was made by mixing proportional amounts from each animal fed that specific hay. Animals were weighed at the end of the trial. The composite hay samples were used for the laboratory analyses in this experiment. The hay samples were analyzed for dry matter (DM) and organic matter (OM) (A.O.A.C, 1975). Neutral detergent fiber (NDF) was determined as outlined by Goering and Van Soest (1970) using sintered glass crucibles for filtration. NDF residues and all subsequent residues were dried at 55 C in a forced air oven overnight and at 55 C in a vacuum oven for 1-3 hrs. Hartadi (1980) determined NDF (using glass wool for filtration) termed NDF (H), acid detergent fiber (ADF) , permanganate lignin (PLIG) , hemicellulose (HEMV) , cellulose (CELLV) and crude protein (CP) (A.O.A.C, 1975; Goering and Van Soest, 1970). The NDF residues were delignified with sodium chlorite (NaC10~) according to a procedure adapted from Whistler et al. (1948). An Eberbach heated shaking water bath set at 70 C was fitted with clamps to hold six 500 ml Erlenmeyer flasks and was placed in an explosion proof hood. Rubber tubing, attached to a manifold, was directed above each flask to deliver C0» continuously in order to displace CIOformed during the reaction. NDF residues were placed in the flasks and NaClO. equivalent to the weight of the sample and 100 ml of 1% acetic acid were added to each flask. The weight of NDF residue transferred was determined by weighing the sintered glass crucible before and after the transfer. Flasks were placed in

PAGE 41

30 the water bath and shaken for 30 minutes. Flasks were removed from the bath in order to add an additional .5 gm of NaClO . The flasks were returned to the bath and shaken for an additional 15 minutes. The flasks were then removed from the water bath to a rack above the bath, continually flushed with CO^ and allowed to cool. Enough ascorbic acid was added to stop the reaction (color changed from bright yellow to pale yellow or white) . Each sample was filtered on a sintered glass crucible, and the residue obtained was termed holocellulose (HOLD) ; the soluble portion was termed lignin (LIG) . The holocellulose residues were further fractionated to cellulose and hemicellulose by the use of KOH as suggested by Whistler and Gaillard (1961), Bailey et al. (1976) and Wilkie (1979). Holocellulose residues were transferred to 100 ml glass bottles with plastic screw-on caps. Weights were determined by difference, as for NDF. Fifty milliliters of 5% KOH was added and the bottles flushed with N„. Up to 20 samples at a time were shaken intermittently in an Eberbach shaker for 24 hrs. After 24 hrs, an additional 9.5 gm KOH was added to each bottle (to bring the final solution up to 24% KOH) , the bottles were again flushed with N_ and shaken an additional 24 hrs. The samples were filtered on sintered glass crucibles fitted to 250 ml side-arm Erlenmeyer flasks. The residue was washed with 10% acetic acid and water and was termed cellulose (CELL) . The filtrate was immediately neutralized to pH 7 by the addition of acetic acid and refrigerated. The filtrate contained the hemicellulosic portion of the cell walls (HC) . CW components were expressed as the percentage of NDF.

PAGE 42

31 A list of abbreviations is included in Table 1. Descriptions of hays and composition data are listed in Appendix Tables 14 17. Statistical analysis (S.A.S., 1979) included the general linear model procedure and simple correlations. Results and Discussion Analyses of variance for the chemical composition of the hays are presented in Table 2. Cultivar and age were the main effects. Age generally did not have a consistent linear effect on chemical parameters. For all parameters, hay (cultivar) and age were significant. Significant interactions of hay x age occurred only for NDF, NDF(H), HC, HEMV, CELL and PLIG. Except for HOLO and LIG 2 (which sum to 100 for each sample), the R values were .89 .94, showing that the effects of hay and age explained most of the variation in chemical composition. One of the objectives of this experiment was to determine if the HC, CELL and LIG values were different from HEMV, CELLV and PLIG or, in other words, was the NaClO -KOH method different from the Van Soest analyses. The two procedures (NaClO^-KOH vs. Van Soest) were compared by analysis of variance of hemicellulose, cellulose and lignin (Table 3). For all 3 variables, methods were different. The overall means for NaClO^-KOH vs. Van Soest 's analysis, respectively, were hemicellulose 51.8 vs. 45.2; cellulose 40.8 vs. 43.5; and lignin 7.4 vs. 8.6 (% of NDF). The correlation between values obtained for the various chemical fractions by the two methods are in Table 4. High correlations

PAGE 43

Table 1. Abbreviations used in the dissertation. 32 ADF Acid detergent fiber, % of DM CC Cell contents, % of DM CELL Alkali-insoluble cellulose, % of NDF CELLD Cellulose digestibility, in vivo , % CELLV Van Soest cellulose (ADF PLIG) , % of NDF CP Crude protein, % of DM CWC Cell wall constituents, % of DM DM Dry matter, % DNDF Digestible neutral detergent fiber, % of DM DOM Digestible organic matter, % of DM DOMI Digestible organic matter intake, g/kg W /day HC Alkali soluble hemicellulose, % of NDF HCD Hemicellulose digestibility, in vivo , % HEMV Van Soest hemicellulose (NDF ADF) , % of NDF HOLO Holocellulose, % of NDF HOLOD Holocellulose digestibility, in vivo, % IVCELLD _In vitro cellulose digestion, rumen fluid only, % IVCELLDP In vitro cellulose digestion, rumen fluid plus pepsin, % IVHCD In vitro hemicellulose digestion, rumen fluid only, % IVHCDP In vitro hemicellulose digestion, rumen fluid plus pepsin, % IVHOLOD In vitro holocellulose digestion, rumen fluid only, % IVHOLODP 2ll vitro holocellulose digestion, rumen fluid plus pepsin, % IVNDFD ^ vitro neutral detergent fiber digestion, rumen fluid only, % IVNDFDP In vitro neutral detergent fiber digestion, rumen fluid plus pepsin, %

PAGE 44

Table 1 continued. 33 LIG Lignin (100 HOLD), % of NDF NDF Neutral detergent fiber, % of DM NDFD Neutral detergent fiber digestibility, in vivo, % NDF(H) Neutral detergent fiber (Hartadi, 1980), % of DM NDFI Neutral detergent fiber intake, g/kg W /day (M Organic matter, % of DM OMD Organic matter digestibility, ±n vivo , % OMI Organic matter intake, g/kg W /day PLIG Permanganate lignin, % of DM

PAGE 45

34 CO CO u o o •H 4-1 •H CO O Cu e o o CO o •H B ai x: o 4-1 O 0) O c CO •H Vj CO > o CO •H CO CO C < CM QJ .-I XI CO H

PAGE 46

35 Table 3. Analysis of variance of method differences.

PAGE 47

36 Table 4. Correlation coefficients (r) between variables obtained by two methods.^'"

PAGE 48

37 existed between the two hemicellulose fractions (r=.71) and the two cellulose fractions (r=.86) but there was no relationship between the lignin values (r=-.03). Differences due to method of analysis are more readily apparent when presented in graphic form. The actual numerical values for the means of each chemical analyses, grouped by age and cultivar, may be found in Appendix Tables 18 22, Neutral Detergent Fiber NDF values were higher than NDF(H) values (Figures 1 and 2), presumably due to differences in procedure. NDF(H) was filtered on Gooch crucibles packed with glass wool, which is more rapid than filtering on sintered glass crucibles, as the NDF samples were. NDF and NDF(H) analyses were conducted by two different analysts, at different times. Disregarding the numerical values of the two determinations, it can be seen that the lines follow the same pattern, all cultivars increasing in NDF from 2 to 4 weeks, then • leveling off or decreasing. Overall, in both cases, the two digit/ grasses are different from the bahiagrasses and the bermudagrass. The two analyses were highly correlated (Table 4) . The NDF residues were used for subsequent fractionation in the present study. It has often been stated that maturity is the greatest factor affecting forage quality. In many instances, the effect of age is linear; i.e., with increasing maturity, some factor increases and/or decreases. Well-known effects of maturity have been the increase in cell wall components and the decrease in crude protein and digestibility. The data presented here, however, generally do not exhibit

PAGE 49

38 90 I 80 k ^ 70 :<^':'ll^^& 4 6 WEEKS 8 Figure 1. Change in NDF content by weeks and hay. Q_ O^i^gentine bahiagrass, 9 ©Paraguay bahiagrass, * *X46-2 digitgrass, it^X124-4 digitgrass, A. . ^ Coastcross-1 bermudagrass.

PAGE 50

85 r 39 ^ 175 ^ 65 2 4 6 WEEKS 8 Figure 2. Change in NDF(H) content by weeks and hay. O -Q Argentine bahiagrass, O -©Paraguay bahiagrass, 'A' * X46-2 digitgrass, ^ TiirXl24-4 digitgrass, A • .A Coastcross-l bermudagrass.

PAGE 51

40 a linear trend with maturity. One known event, extensive insect damage (striped grass looper, Mocis latipes ) at 4 weeks to the digitgrasses and bahiagrass, may provide partial explanation for lack of linearity. Holo cellulose Holocellulose, the delignified plant cell wall composed primarily of hemicellulose and cellulose, made up the major portion of the NDF fraction, ranging from 90.45 94.36% of NDF (Appendix Table 19). Holocellulose values, and all subsequent chemical composition values, were expressed as percentage of NDF rather than DM. Holocellulose content of X46-2, X124-4 and Argentine cultivars were not different; Coastcross-1 was different from the former three cultivars and Paraguay fell between Coastcross-1 and the rest (Figure 3) . Four of the cultivars showed an eventual decline in holocellulose content from 2 to 8 weeks but the interim pattern was erratic. Earlier plant analysts felt that lignin interfered with the analysis of structural carbohydrates (Whistler et al., 1948). Adaption of wood chemistry methods led to the use of NaClO -KOH to de-lignify plant material. Packett et al. (1965) indicated that polysaccharide structures are left intact after delignification provided the procedure is not continued to reduce lignin values to 2% or less. Wilkie (1979), however, stated that delignification may oxidize some reducing-end residues and cause partial depolymerization. Whether or not hemicellulose is bonded to lignin is still uncertain although work by Albersheim et al. (Bauer et al., 1973; Keegstra et al., 1973; Talmadge et al., 1973) on dicotyledonous

PAGE 52

94 41 u. 93 ^ 92 91 90 4 6 WEEKS 8 Figure 3. Change in holocellulose content by weeks and hay. OO Argentine bahiagrass, 9 Paraguay bahiagrass, * :ArX46-2 digitgrass, -^ :iirX124-4 digitgrass, O* • -QCoastcross-l bermudagrass.

PAGE 53

42 plants suggest that there is bonding between hemicellulose and lignin. Hemicellulose has been extracted from plants without prior delignification (Bailey et al., 1976) but the completeness of the extraction is unclear (Gaillard, 1958). It was felt on these tropical grass hays, which are high in CWC and lignin, the best procedure would be delignification prior to further analyses. The holocellulose values, as an end in themselves, did not give a clearer or better picture of the plant than did NDF values, which are easier to obtain. The erratic pattern with maturity suggests some caution should be used in interpretation of the LIG data (100-HOLO) , Hemicellulose Two methods of analyzing for hemicellulose were used NaClO -KOH and the Van Soest detergent system. Figures 4 and 5 present the changes observed in hemicellulose content by cultivar and age (see Appendix Table 20 for means). HC values, except in the case of Coastcross-1, decreased only slightly with maturation which agrees with other work on hemicellulose in tropical grasses (Bailey, 1973). HEMV values, on the other hand and again excepting Coastcross-1, showed an overall greater decline with maturity. HC values were higher as a percentage of NDF than were HEMV values. With both methods, X124-4 was different from the other grasses. The two bahiagrasses and X46-2 were similar in HC content compared to the other grasses while Coastcross-1 was different from the rest. The bahiagrasses and Coastcross-1 formed a similar group with respect to HEMV values, with the digitgrasses differing from each other and from the rest of the grasses.

PAGE 54

43 60 50 40 4 6 WEEKS 8 Figure 4. Change in alkali-soluble heraicellulose by weeks and hay. O" -O Argentine bahiagrass, 0-0 Paraguay bahiagrass, * ir:i^6-2 digitgrass, ^ =5^X124-4 digitgrass, ^. • .^ Coastcross-1 bermudagrass.

PAGE 55

44 60 ^ 50 40 4 6 WEEKS 8 Figure 5. Change in Van Soest hemicellulose by weeks and hay. O O Argentine bahiagrass, ^ -^Paraguay bahiagrass. * *X46-2 digitgrass, -ji^ ^X124-4 digitgrass, ^. . .^ Coastcross-1 bermudagrass.

PAGE 56

45 Hemicellulose has been defined as the cell wall and intercellular polysaccharides that can be extracted from higher land plant tissues that are, or were, lignified (Wilkie, 1979). More simply, hemicellulose refers to all of the types of polysaccharides found in plants other than cellulose, starch and fructans. The preponderance of literature on hemicellulose deals with the Van Soest hemicellulose obtained by subtracting ADF from NDF. Work as early as 1970 (Bailey and Ulyatt) points out the problems inherent in this definition, viz., that there are hemicellulosic sugars present in the ADF residue (see also Morrison, 1980). The hemicellulose content of a plant changes with growth and maturity and is expected to increase with maturity. Bailey (1973) reported on several studies with temperate grasses where the increase in hemicellulose content with age was associated with an increase in stem tissue. For tropical grasses, however, seasonal changes are not as marked, the hemicellulose content remaining about the same in spite of increasing CWC and increasing proportion of stems. Hemicellulosic materials extracted by alkali may precipitate on neutralization and mild acidification at 2 C to form hemicellulose-A. The polysaccharides remaining in solution can be precipitated with ethanol to form hemicellulose-B. Repeated attempts to obtain a hemicellulose-A fraction failed. Scant information in the literature as reviewed by Bailey (1973) showed that hemicellulose-A comprises 1.0 19% of DM in Heteropogon contortus leaves and stems respectively, l.A 3.2% of DM in Digitaria decumbens and .1 .4% of DM in Setaria sphacelata . It would seem then, that the hemicellulose-A fraction is generally low in tropical grasses, thus providing

PAGE 57

46 partial explanation for our failure to obtain the fraction. HC values presented here are representative of the hemicellulose-B fraction which consists primarily of xylans. Cellulose Cellulose obtained by either method (Figures 6 and 7) showed remarkably similar patterns (see Appendix Table 21 for means) . In both instances, Coastcross-1 was very different (having the lowest amount of cellulose) from the other grasses; X46-2, Paraguay and Argentine cultivars were not different, and X124-4 had the highest amount of cellulose. CELL, collected on the sintered glass crucibles, was a pure white fibrous substance very similar in appearance to CELLV obtained by the ADF procedure. Lignin Lignin values (Figure 8) were calculated as 100 HOLO. These lignin values are very different from the PLIG values (Figure 9; Appendix Table 22). Lignin, like hemicellulose, is defined by the method used to isolate it. The completeness of delignification of NDF was tested by staining holocellulose with phloroglucinol-HCl (a test for lignin). No stain was present in the holocellulose but it is likely that the material oxidized by the NaClO included more than lignin. Obtaining a lignin value in this manner is not recommended due to the possibility of additional products forming during the oxidation of lignin from the cell walls by NaC10„ . Only Coastcross-1 had a different amount of LIG than the other grasses while PLIG amounts were different for each cultivar.

PAGE 58

47 50 ^ 40 30 •A4 6 WEEKS 8 Figure 6. Change in alkali-insoluble cellulose by weeks and hay. O" ~ "O Argentine bahiagrass, ®-^Paraguay bahiagrass, * -"A'XAe-Z digitgrass, ix ^X124-4 digitgrass, J^' • .^ Coastcross-1 bermudagrass .

PAGE 59

48 50 40 30 A 2 4 6 WEEKS 8 Figure 7. Change in Van Soest cellulose by weeks and hay. 0~ ~ "O Argentine bahiagrass, Q-^Paraguay bahiagrass, "At ^X46-2 digitgrass, ^ =5irX124-4 digitgrass, A• -ACoastcross-l bermudagrass.

PAGE 60

49 10 I 9 •

PAGE 61

50 o 10 8 ^ =^ 7 5 4 6 WEEKS Figure 9. Change in permanganate lignin by weeks and hay. 0~ ~ "O^'^S^'^tine bahiagrass, O-©Paraguay bahiagrass, tAt ^A" X46-2 digitgrass, "jlf ^X124-4 digitgrass, O" • •QCoastcross-1 berinudagrass .

PAGE 62

51 Summary One cultivar of Cynodon dactylon (Coastcross-1) , two experimental lines of Digitaria species (X46-2 and X124-4) and two cultivars of Paspalum notatum (Argentine and Paraguay) cut at four ages each (2, 4, 6 and 8 weeks regrowths) were evaluated by chemical analyses. Composite samples for each hay were analyzed for organic matter and neutral detergent fiber. NDF fractions (cellulose, hemicellulose, lignin) were determined by two methods: one, the conventional Van Soest detergent analysis (CELLV, HEMV, PLIG) ; the other a classical fractionation using hypochlorite and alkali (HOLO, CELL, HC, LIG) . Values obtained by the two methods used in the analyses were significantly different. As percentages of NDF, HC values remained the same or decreased slightly with age for each cultivar while HEMV values showed a greater decline with maturity. CELL and CELLV values had similar patterns, increasing up to 6 weeks, then decreasing slightly. Correlation between HC and HEMV was .71 and between CELL and CELLV was .86. Further studies of these fractions to determine their digestibility and value in prediction equations are in subsequent chapters.

PAGE 63

CHAPTER IV IN VITRO AND IN VIVO DIGESTIBILITY OF TROPICAL GRASS HAYS Introduction Ruminants derive a great portion of their dietary energy from the digestion of cell wall polysaccharides when they are fed forage diets. The digestibility and intake of CWC is crucial in determining forage quality. Differences in breakdown of the CWC in the rumen are due, in part, to differences in chemical structure of the polysaccharides. Early iji vitro work on hemicellulose degradation in the rumen used isolated fractions to determine digestibility of hemicellulose. Two general theories arose from these studies. One theory attributes low hemicellulose digestibility to chemical heterogeneity of the hemicellulose structure (Gaillard, 1965; Sullivan, 1966; Dekker and Richards, 1973). Other work has shown that rather than the chemical heterogeneity of the hemicellulose being responsible for lower digestion, that the incomplete digestion was due to the presence of lignin (Bailey and MacRae, 1970; Beveridge and Richards, 1973). Morrison (1979) suggests that both theories may be involved, i.e., in young plants the chemical structure of the hemicellulose interferes with digestion processes while in older plants, increasing lignif ication causes lower digestibility. 52

PAGE 64

53 Studies on isolated hemicelluloses and its components give only a limited idea of what is occurring in the rumen. Beever et al. (1971) in in vivo studies, showed that 70% of the hemicellulose digestion occurred before entering the small intestine, very little hemicellulose digestion occurred in the small intestine and approximately 30% of the hemicellulose digestion occurred in the cecum and colon. A review of several digestion studies (Ulyatt et al. ,1975) found that the proportion of hemicellulose digested in the large intestine was higher than that of cellulose. Bailey et al. (1976) found that as hemicellulose content increased, the percent digested in the rumen decreased. Ulyatt and Egan (1979) found 77 to 94% of digestible hemicellulose apparently digested in the rumen, and 6 to 23% in the intestines. Some explanation of the confusing in vivo results may be provided by Gaillard and Richards (1975) who suggest the presence of a lignin-carbohydrate complex which passes from the rumen in solution, is precipitated in the lower pH of the abomasum, undergoes some redissolution at later stages in the lower gastrointestinal tract, and emerges in the feces. An intermediate step, then, between studying the digestibility of isolated hemicellulose and other CWC fractions and the complex in vivo studies appears to be in_ vitro studies on whole plant tissue, followed by analyses of the various fractions. The objectives of this study were to determine the ±n vitro and in vivo digestibilities of various fractions of tropical grass hays.

PAGE 65

54 Experimental Procedures One cultivar of Cynodon dactylon (Coastcross-1 bermudagrass) , two experimental lines of Digitaria (X124-4 and X46-2) and two cultivars of Paspalum no ta turn (Paraguay and Argentine bahiagrass) harvested after 2, 4, 6 and 8 weeks of regrowth with two field replications were used in this experiment (Abrams, 1980). The grasses were artificially dried and chopped. The grasses were fed ad libitum as chopped hay to mature wethers to determine simultaneously voluntary intake and digestibility of organic matter (CM) and neutral detergent fiber (NDF) (Abrams, 1980). Voluntary intakes were reported on the basis of metabolic body weight (g/kg W' /day). Each sample was pooled by animal, ground to pass a 4 mm screen in a Wiley mill, mixed and a representative portion reground through a 1 mm screen. A composite hay sample was made by mixing proportional amounts from each animal fed that specific hay. Feces were collected in canvas bags, weighed daily and 20% kept for further analyses. The composite hay samples and some individual feces samples were used for the analyses in this study. Half gram samples of each hay were fermented in rumen fluid and buffer for 48 hr, followed by 48 hr in acid-pepsin (Moore et al., 1972; Moore and Mott, 1974; 1976), refluxed in neutral detergent solution (Goering and Van Soest, 1970) and filtered on sintered glass crucibles. The NDF residues were saved for further analyses. This treatment was termed the rumen fluid plus pepsin (RF+P) treatment. Another set of hay samples (.5 gm each) were fermented in rumen fluid

PAGE 66

55 for 48 hr, refluxed in neutral detergent solution (Goering and Van Soest, 1970), filtered on sintered glass crucibles and the NDF residue saved for further analyses. This treatment was termed the rumen fluid (RF) treatment. Each of the RF and RF+P residues were analyzed for holocellulose (HOLO) , hemicellulose (HC) , cellulose (CELL) and lignin (LIG) as outlined in Chapter III. Fecal samples from the in vivo studies of 2 wk and 8 wk Argentine bahiagrass, Coastcross-1 bermudagrass and X46-2 digitgrass were analyzed for NDF, HOLO, HC, CELL ^nd LIG (Chapter III) , and acid detergent fiber (ADF) (Goering and Van Soest, 1970). In vivo digestion coefficients were calculated for these fractions. Abbreviations are listed in Table 1. Descriptions of hays and jJl vivo data are listed in Appendix Tables 14 17. Statistical analyses were performed using the statistical analysis system (SAS, 1979) computer package for general linear models, t-tests and correlation procedures. Results and Discussion In Vitro Digestibility Analysis of variance of the 1^ vitro data is presented in Table 5. Values for in vitro digestion of individual hays are in Appendix Table 15. In vitro digestibility of all fractions decreased with increasing maturity in all hays. RF and RF+P digestibilities were compared by t-tests (SAS, 1979) (H :RF+P RF = 0) for each fraction (NDF, HOLO, HC and CELL) by hay and weeks of regrowth (Table 6).

PAGE 67

56 CO 0) 3 r-l CO > to O U u •H > o CO in 01 CO H P-i Q hJ hJ W u > M O w o > o Q O I-) o o o o Q O CO O P M O o •

PAGE 68

57 CO acu CO 3 a •H 3 C OJ B 3 ^1 T3 C to -a •H 3 o c o tfl H S-i CO a e o o CO El

PAGE 69

58 NDF digestibility was greater (P values ranged from .14 to .0007, Table 6) in RF+P for all hays and regrowths. The standard method of determining NDF digestibility (Goering and Van Soest, 1970) uses only rumen fluid for 48 hr. While the overall increase in NDF digestibility ranged from 2.2 to 5.7 percentage units (Table 7), for individual hays there were differences in NDF digestibility of up to 7 percentage units. In most instances, CELL was more digestible than HC, whether digested in RF or RF+P. HOLD and CELL digestibility increased in RF+P only for Paraguay and Coastcross-1 and either decreased or showed minimal change in the other grasses. HC digestibility was not different for X46-2 but showed an increase of 2.1 to 3.0 percentage units for the other grasses. For individual hays, there were increases in HC digestibility of up to 12 units by use of acid-pepsin. By weeks of regrowth, NDF digestibility increased with RF+P except at 8 weeks. HOLD and CELL showed no significant difference in digestibility. HC exhibited the greatest difference in digestibility at 2 weeks; the difference decreased with increasing maturity. NDF and HC digestibility decreased with age (Table 7 and Appendix Table 16) which is consistent with work on temperate grasses (Gaillard, 1965; Waite et al., 1964; Bailey, 1967). It seems, then, that further digestion in vitro of the CWC occurs under acid-pepsin conditions. This digestion, for the most part, is not attributable to further cellulose digestion but rather, to hemicellulose digestion and dissolution of lignin. Whether this increase in hemicellulose digestibility is due to solubilization by

PAGE 70

59 o u w u in 1-^ 00 o> O in 00 00 u-i J w u a. Q w u > o w u > o ON o 00 ON H r^

PAGE 71

60 the acidic conditions and whether the hemicellulose would re-precipitate under neutral conditions (as suggested by Gaillard and Richards, 1975) is not known. The majority of the in vivo studies have used HEMV as the measure of hemicellulose. Our earlier experiments suggest that HEMV and HC may not be the same entity. Thus, studies showing that the greatest proportion of "hemicellulose" digestion occurs in the rumen may be misleading, as indicated by these in vitro results. In Vivo Digestibility Fecal samples collected during the feeding trials of these tropical grass hays were used in further analyses. Argentine bahiagrass, Coastcross-1 bermudagrass and X46-2 digitgrass were chosen from each of the three genera at 2 wk and 8 wk of age. NDF, HOLO, HC, CELL, LIG and ADF values are found in Table 8. Analysis of variance of the chemical composition of the fecal samples is in Appendix Table 24. Compared to the chemical composition of the hays (Appendix Table 14), fecal samples were lower in NDF, HOLO and CELL, higher in LIG and were about the same or slightly higher in HC content. In vivo intake information was used in conjunction with the feces, orts and waste information to calculate digestion coefficients for NDF, HOLO, HC and CELL (Table 7; Abrams, 1980). Analysis of variance for the digestion coefficients is in Appendix Table 23. The values are compared with in vitro digestibility data on the same hays in RF and RF+P (Table 7). At 2 weeks, in vivo digestion coefficients were similar to in vitro for X46-2 digitgrass and Argentine bahiagrass. For these two grasses, also, digestion coefficients were higher for

PAGE 72

61 Table 8. Chemical composition of feces. NDF^ HOLO ADF HC CELL LIG 2 Weeks X46-2

PAGE 73

62 cellulose than for hemi cellulose at 2 and 8 weeks. The Coastcross-l bermudagrass had higher in vivo digestion coefficients. Digestibility of the various fractions was high at an early stage of growth. Not unusually, the situation changed with maturity; at 8 weeks, the in vivo digestion coefficients were higher than the in vitro , ranging from at least 2 to 14 percentage points higher. The degree of relationship (or correlation) between the in vivo and in vitro values was high (Table 9) despite lower in vitro values at 8 weeks. Implications are that in vitro digestibility information is more representative of i^ vivo performance at early stages of plant growth and that with maturity, the in vitro digestibility values still accurately reflect in vivo but the absolute values will be lower than in vivo values. The correlation of HCD with IVHCDP was higher than with IVHCD, implying that acid-pepsin may have some effect in vivo . Summary One cultivar of Cynodon dactylon (Coastcross-l) , two cultivars of Digitaria species (X46-2 and X124-4) and two cultivars of Paspalum notatum (Argentine and Paraguay) cut at four ages each 2, 4, 6 and 8 weeks of regrowth, were evaluated by in vitro digestion of neutral detergent fiber (NDF) in rumen fluid (RF) and RF plus acid-pepsin (RF+P) and by chemical analyses of NDF fractions. In addition, sheep feeding trials were run. Composite samples of each hay were analyzed for NDF, holocellulose (HOLO) , hemicellulose (HC) , cellulose (CELL) and lignin (LIG) . Fecal samples of 2 and 8 week Coastcross-l,

PAGE 74

63 Table 9. Relationship (r ) between 2:11 vivo and ±n vitro digestion coefficients for six hays. Variables in Model NDFD vs . NDFD vs . IVNDFD IVNDFDP .91 .92 .0033 .0025 HOLOD vs. IVHOLOD HOLOD vs. IVHOLOD? ,88 ,84 .0058 .0094 HCD

PAGE 75

64 Argentine and X46-2 grasses were analyzed for NDF, HOLO, HC, CELL, LIG and acid detergent fiber (ADF) . In vitro digestibility in RF and RF+P of NDF, HOLO, HC and CELL decreased with increasing maturity in all hays. In vitro digestion of cellulose was generally greater than that of hemicellulose. 2lL v^tro digestibility of NDF and heraicellulose was increased by further digestion in RF+P while holocellulose and cellulose digestibility showed minimal change. Fecal samples were lower in NDF, HOLO and CELL, higher in LIG and about the same or slightly higher in HC content. In vivo digestion coefficients were calculated for each of the fractions (NDF, HOLO, HC and CELL) and showed trends similar to in vitro digestibility.

PAGE 76

CHAPTER V PREDICTION OF FORAGE QUALITY FROM CHEMICAL AND IN VITRO ANALYSES Introduction Forage quality must be related to animal performance. Moore and Mott (1973) defined forage quality as output per animal and as being a function of voluntary intake and digestibility of nutrients when the forage is fed alone, ad libitum , to a specified animal. When long-term production trials are not feasible as is often the case, researchers turn to laboratory analyses of forage quality. Many and varied are the laboratory procedures devised to predict forage quality. These may be grouped into chemical and in vitro methods. Thus far, chemical methods have not been overly successful in predicting forage quality but the in vitro methodology, especially with respect to predicting In vivo digestibility, has been more successful. In a review of the many forage evaluation methods in popular use, Barnes and Marten (1979) reported that the in vitro system was superior to chemical analyses for predicting in vivo digestibility of forages. Intake is much more difficult to predict from laboratory analyses but Rohweder et al. (1976) have suggested the use of neutral detergent fiber (NDF) to predict intake. This study was conducted to examine the relationship between the Van Soest chemical analyses, another chemical analysis procedure (NaC10„-K0H, see Chapter III) and/or in vitro digestion with in vivo 65

PAGE 77

66 intake and digestibility of organic matter (OM) and NDF in 39 tropical grass hays of three genera harvested after 2, 4, 6 and 8 weeks of regrowth. Experimental Procedure One cultivar of Cynodon dactylon (Coastcross-1 bermudagrass) , two experimental cultivars of Digitaria (X46-2 and X124-4 digitgrass) and two cultivars of Paspalum not a turn (Argentine and Paraguay bahiagrass) harvested after 2, 4, 6 and 8 weeks of regrowth were used in this experiment. Hays were chopped and fed ad libitum to mature wethers to determine intake and digestibility as outlined in Chapter III. A composite hay sample was made by mixing proportional amounts of hay from each animal fed that particular hay, was ground through 4 mm and 1 mm screens of a Wiley mill and was used for the laboratory analyses. The composite hay samples were analyzed for DM, OM, NDF, holocellulose (HOLO) , hemicellulose (HC, HEMV) , cellulose (CELL, CELLV) and lignin (LIG, PLIG) as outlined in Chapter III. In vitro fermentation in rumen fluid (RF) for 48 hrs and in vitro fermentation in RF for 48 hrs followed by 48 hrs in acid pepsin (RF+P) were both followed by refluxing in neutral detergent solution to determine in vitro NDF, HOLO, HC and CELL digestion as outlined in Chapter IV. A list of abbreviations is presented in Table 1. Statistical analyses were performed using the statistical analysis system (SAS, 1979) computer package for general linear models and correlation procedures.

PAGE 78

67 Results and Discussion Intake and Digestibility Individual values of ±n vivo Intake and digestibility are in Appendix Tables 17, 24 and 25. Chaves (1979), Abrams (1980) and Hartadl (1980) presented discussions of the relationship between intake and digestibility in vivo of these hays. They suggested that digestibility was not the primary factor controlling Intake of these hays. Furthermore, there was a poor relationship between intake and digestibility (OMI vs. OMD, r = .47; Chaves, 1979). Therefore, it is unlikely that a single laboratory procedure will prove to be a good predictor of In vivo performane. Another reason for skepticism concerning the ability of a single laboratory procedure to predict forage quality lies in the effect of maturity on the ijx vivo values compared to the effect of maturity on the laboratory values. In Chapter III, it was reported that maturity did not have a consistent effect on the chemical parameters, i.e., the effect of age was not linear. Expectations are that, with maturity, some factors decrease (e.g., crude protein) and some increase (e.g., NDF) but the hays used in this study did not follow the expected pattern. However, the i^ vitro and ±n vivo values did follow the expected pattern of decline with maturity in most of the hays (see Appendix Tables 15 and 16) . Argentine bahiagrass and Coastcross-1 berraudagrass did not change in OMI and changed very little in OMD but the other three grasses decreased from 2-8 weeks in OMI by 4 8 units (g/kg W^^/day) and decreased in OMD by 10 12 percentage units (Abrams, 1980)

PAGE 79

68 Prediction of Forage Quality from Laboratory Analyses Simple correlations between i^ vivo data and laboratory analyses were conducted (SAS, 1979). Only the best correlations for each ±n vivo parameter are presented in Table 10. The highest correlation obtained between OMI and all laboratory analyses was 2 with CP (r =.20). DOMI had the highest correlation with NDF 2 (r =.46) and NDFI had the highest correlation with PLIG and CP 2 (r =.06). Intake has always been difficult to predict by laboratory analyses alone. Differences in voluntary intake are often attributed to rate of digestion and ±n vitro digestibility systems generally are more accurate at predicting intake than are chemical analyses. Table 11 presents correlations between in vivo and in vitro values. Two in vivo intake variables, OMI and DOMI, had significant corre2 lations with all of the ijn vitro analyses; r values ranging from .13 to .56 with significance probabilities of Pi.lO to -.001, respectively. NDFI, which did not vary (Appendix Table 24), was not predicted well, or at all, by any of the ±n vitro analyses. 2 OMD and NDFD had the highest correlations with NDF (r =.59 and .56, respectively) while DNDF had the highest correlation with PLIG 2 (r =.23) (Table 10). Aside from NDF and PLIG, there were no chemical analyses which accurately predicted iji vivo digestibility. However, the i^ vitro analyses generally gave highly significant correlations 2 with OMD and NDFD; r values ranging from .61 to .81 (Table 11). Correlations between in vitro analyses and DNDF, although significant, were lower.

PAGE 80

69 Table 10, Relationship (r ) between in vivo values and laboratory analyses, obtained with thirty-nine different hays. 2 Variables in Model r P OMI vs .

PAGE 81

70 2 Table 11. Relationships (r ) between in vivo values and in vitro analyses, obtained with thirty-nine different hays.^ OMI In vivo Variables OMD DOMI NDFI NDFD DNDF IVITOFD^ IVNDFDP

PAGE 82

71 2 Interestingly, in almost every instance, r values for each chemical fraction's in vitro digestibility in RF and RF+P (IVNDFD and IVNDFDP, for example) were not different. In Chapter III, it was shown that there was a significant increase in digestion in the RF+P treatment for the NDF and hemicellulose fractions, while holocellulose and cellulose digestibility showed minimal change. The similar in vivo correlations with the RF and RF+P In vitro data make laboratory decision making easier. Since there is little difference in the two methods (RF vs. RF+P) insofar as predicting In vivo data is concerned, the method of choice would be RF, which is less time, equipment and reagent consuming. Furthermore, there was no advantage to using a more complicated procedure for residue analysis than the NDF procedure. Tables 12 and 13 present the models obtained by general linear models procedure (SAS, 1979) that best fitted the in vivo intake and digestibility data. A rational, rather than an empirical or multiple regression, approach was used to choose independent variables for each model. Independent variables were chosen on the basis of significant correlation with the j^ vivo measurement of interest (Tables 10 and 11). In most cases, in vitro digestion values in RF only were used unless RF+P values were much higher. In addition, some variables were included on the basis of their presumed biological role in the structure of the cell wall. Cell wall structure and components play a crucial role in determining intake and digestibility of tropical grasses.

PAGE 83

72 Table 12. Coefficient of determination (R^) , probability (P) and standard error of the estimate (s ) for the regression of in vivo intake values on several independent variables.

PAGE 84

73 Table 13. Coefficient of determination (R^) , probability (P) and standard error of the estimate (s _ ) for the regression of in vivo digestibility values on several independent variables.

PAGE 85

74 Going from a model containing one independent variable to a model containing more than one independent variable did not produce better or more acceptable predictions of intake. OMI ranged from 46.7 to 79.7 g/kg W'^^/day (Appendix Table 24). Four of the "best" 2 equations for predicting OMI included NDF and PLIG. R values and standard error of the estimate, respectively, were .24 and 6.42 (NDF and PLIG), .25 and 6.44 (NDF, HC and PLIG), .26 and 6.43 (NDF, CELL and PLIG) and .25 and 6.51 (NDF, HC, CELL and PLIG). DOMI .75 (ranging from 25.6 to 54.0 g/kg W* /day) was best predicted by two and three variable models that included an i^ vitro value in the model, as to be expected for the prediction of digestible OM intake. The best equation for predicting DOMI included IVNDFDP, PLIG and HC 2 (R = .58 and s =4,71), but this three variable equation was not yx ^ 2 much better than the equation using only IVNDFDP (R =.56). NDFI was still predicted poorly by any equation, be it single or multiple variable, probably because there is little variation in 2 NDFI (see Appendix Table 24). As R values increased slightly, standard error of the estimate increased so that the best predictor of NDFI remained crude protein alone (Table 10) . OMD and NDFD were predicted quite well by any and all of the single variable rn vitro equations (Table 11) . Adding more variables 2 to the model decreased the R and increased s for OMD and NDFD. y-x Single variable models of in_ vitro data used for predicting DNDF had 2 R values ranging from .14 to .38. The best two variable model for 2 predicting DNDF included PLIG and CELL (R =.42) and was not improved 2 by adding another variable to the model (HC, CELL and PLIG, R =.42).

PAGE 86

75 The best four variable model included NDF, HC, CELL and PLIG (R^=.43) but the standard error of the estimate increased with the addition of more variables to the model statements. Conclusions The results of the studies indicated that NDF, CP and PLIG were the chemical analyses most closely associated with the digestibility and intake of tropical grasses. New methodology developed to analyze cell walls did not contribute to the prediction of in vivo performance. Most of the in vitro procedures gave fair to good predictions of OMI, OMD, DOMI , NDFD and DNDF but not of NDFI. Two, three and four variable models increased the R values for OMI, OMD 2 and DOMI and decreased the R values slightly for DNDF and more so for NDFD. NDFI could not be predicted by any single or multiple variable model. It is concluded that we still do not have an accurate picture of the structure and function of the cell wall in tropical grasses. The difficulty of predicting in vivo animal performance remains a great challenge in evaluation of forage quality. Twenty of these hays were analyzed at the Pennsylvania State University by near infrared reflectance spectroscopy (NIRS) (Russo, 1981, unpublished data). Using large multiple regression models and two wavelengths, NIRS gave better estimates of forage quality than any of the chemical and in vitro analyses conducted in this study.

PAGE 87

76 Summary One cultivar of Cynodon dactylon (Coastcross-1 bermudagrass) , two cultivars of Digitarla species (X46-2 and X124-4) and two cultivars of Paspalum no ta turn (Argentine and Paraguay bahiagrass) cut at four ages each (2, 4, 6 and 8 week regrowths) were evaluated by chemical analyses and in vitro NDF digestibility in rumen fluid (RF) and RF plus acid pepsin (RF+P) . in addition, sheep feeding trials were conducted. Composite samples for each hay were analyzed for organic matter, NDF and holocellulose (HOLO) . Hemicellulose (HC, HEMV), cellulose (CELL, CELLV) and lignin (LIG, PLIG) were determined by two different methods. Information obtained was used to generate prediction equations for sheep in vivo intake and digestibility. NDF, CP and PLIG were the best single chemical analyses for prediction of intake and digestibility. OMI and NDFI were not predicted as well as DOMI. Any of the in vitro analyses were better for prediction of in vivo digestibility (OMD, NDFD and DNDF) than were chemical analyses. Two and three variable models which included NDF and PLIG improved prediction of OMI and DOMI.

PAGE 88

CHAPTER VI GENERAL DISCUSSION Forage quality reflects the potential of animal production from forage by use of the available nutrients in the forage. Forage quality may be defined as animal performance (Moore and Mott, 1973). Analysis of forage quality in the laboratory has almost a hundred year history and gaps still remain in our knowledge of how the forage plant is put together, how it functions and how it affects animal performance. Essentially all of the forage evaluation procedures have been developed in temperate climates using temperate forages. Not unlike the problems associated with transferring Western agricultural expertise to developing countries is the problem of transferring forage evaluation schemes generated on temperate forages to tropical forages. Sometimes it works, sometimes it does not. Tropical grasses are rather different from temperate grasses. Tropical grasses have a much greater photosynthetic capability, are more efficient in water use, have different carbon pathways and have different anatomies. The first product of photosynthesis is carbohydrates and tropical grasses have specialized plastids in the sheath surrounding the vascular bundles for starch conversion and storage. Tropical grasses are higher in cell wall and cell wall components than temperate grasses. 77

PAGE 89

78 What is the cell wall? It is a highly ordered macromolecular multilayered network made up of structural carbohydrates (cellulose and hemicellulose) , lignin and protein. Dividing plants into cell contents and cell walls (as does the Van Soest forage evaluation system) provides a rational basis for discussion of forage quality. Cell contents are essentially completely digestible by the ruminant, cell walls have incomplete digestibilities. A number of cell wall models have been proposed. Albersheim and others (Bauer et al., 1973; Keegstra et al,, 1973; Talmadge et al., 1973) propose that the cell wall is made up of long crystalline microfibrils (cellulose) surrounded by hemicellulose, which is hydrogen bonded to the cellulose and covalently bonded to the other cell wall components (primarily lignin) . Covalent bonds are more difficult to break than are hydrogen bonds. Bailey and others (1976) propose that all of the crystalline polymer need not be cellulose, but could be hemicellulosic xylans. The role of hemicellulose in the plant has been established as structural but hemicellulose may exist not only in the cell wall and have been found in the endosperm of cereal plants, for example (Wilkie, 1979). Hemicellulose is added to the wall continuously and is increased by increases in temperature and light. While cellulose composition and structure does not change with maturity or under environmental influences, hemicellulose does. Not enough information exists, but it is likely that each genera, each cultivar, perhaps even each plant, and certainly, different parts within a plant have different kinds and amounts of hemicellulose. Hemicellulose content and concentrations are variable, not constant.

PAGE 90

79 The results of this research indicate that "hemicellulose" is an elusive term. Van Soest hemicellulose is certainly not the same hemicellulose obtained in the NaClO -KOH procedure despite presumed similarities in the cellulose fraction. Lack of expertise prevented further examination of the sugars making up the hemicellulose but some initial observations were made. One of the experimental digitgrasses, X46-2, which is "high" in forage quality by most standards, appeared very low in xylans. Xylans are usually highly ordered, even crystalline and difficult to digest (Daughtry, 1976). Tropical grasses are generally higher in xylans than are temperate grasses (Ojima and Isawa, 1968; Bailey, 1973) and xylans may be responsible for lower quality of tropical grasses. This experimental line of digitgrass should be examined for xylan content and compared with other grasses used in this study. In vitro digestibility results were illuminating. NDF and hemicellulose digestibility increased in acid-pepsin conditions, cellulose digestibility did not increase. Further breakdown of the cell wall, then, in the lower gastrointestinal tract, could be due to breakdown of the hemicellulose. Hemicellulose has generally been presumed to have a low digestibility, with lignification postulated as the limiting factor but a distinction should be made between hemicellulose unavailability due to structural organization of the cell wall carbohydrates or interference by cellulose, lignification, acetylation, etc. (Morrison, 1979). McLeod and Minson (1974) propose that the cell wall carbohydrates have two fractions lignin-free and lignified the lignin-free being totally digestible.

PAGE 91

80 the lignified fraction being undigestible. Other digestion models (Waldo et al., 1972; Abrams, 1980) are similar. Accurate procedures for determining hemicellulose and lignin must be developed in order to answer some of these questions and to prove some of these theories. It would seem that if the proposed cell wall models and digestion models were fairly accurate, using both models would allow one to state that hemicellulose can be "digested" from cellulose and within the hemicellulose polymer itself but that pieces of the hemicellulose polymer remain covalently bonded to lignin and are more resistant to digestion. Bailey and Ulyatt (1970) and Morrison (1980) intimate these conclusions also. To extrapolate further, Chaves (1979) suggests that bridges are formed between the cell wall carbohydrates and the lignin by phenolic acids which can deleteriously affect microbial digestion. In tropical grasses, hemicellulose, lignin and phenolic acids are higher than in temperate grasses. Are these, then, the components responsible for low quality in tropical grasses? Can the plant breeder do anything about it? Breeding for low lignin brought about the development of brown midrib corn, academically interesting but of little use to the farmer. Is low hemicellulose or low lignin content really desirable in a forage plant? Forage quality analysis and prediction of intake and digestibility still has researchable areas. This research should be carried further to determine monosaccharide composition, relate the composition to animal performance and then devise a rapid method of cell wall monosaccharide determination. The physical and chemical forces

PAGE 92

81 holding the various cell wall components together and what it takes in terms of energy, microbes, enzymes, acids or whatever to break those bonds could be examined. Anatomy and histology of forages as related to intake and digestibility should be examined and the polysaccharides and monosaccharides in the various structures be determined. Infrared reflectance spectroscopy may alter procedures used by many forage testing laboratories as it becomes more accurate and accessible for routine use. That may be for the better, allowing more basic research to be conducted on forage plants.

PAGE 93

APPENDIX

PAGE 94

J3 CO H -a c 0) K-l w u o J o PS 3

PAGE 95

84 -a 3 C d o o o M W O i-J O X p o B u > •H 3 U2

PAGE 96

85 CO

PAGE 97

86 o > Q W U > Q U Q Q O hJ O •

PAGE 98

87 CO CO CO 3 •H > •H -a c o CO (U 60 c CO c o •H 4J •H CO o a S o o CO o •H e o
PAGE 99

88 o PL, > w fn ^ 13 T— io>o i^vovOvor^vor^vor^oooo iHOr~~vDLnnv£)rooo>3-o LommvOLOioroOiHc-JiH oorofOmcN!ooo\ONcsi<)-r-4ir)rniHvD\c vOCNvOO<}•vOvOa^O^^O r^ooiriKooooo C O CD O a. B o o (0 c o > cu u CO c o CO c CO a. X (U o CO PS '~Dr--r^r~~r-»r-^r^oocx)oooo to

PAGE 100

89 (0 4J CO •a C 0) (0 Q O O o o ^

PAGE 101

90. T3
PAGE 102

91 Appendix Table 18. Means of neutral detergent fiber by age and hay. NDF

PAGE 103

92 Appendix Table 19. Means of holocellulose by age and hay.

PAGE 104

93 Appendix Table 20. Means of hemicellulose by age and hay. HC

PAGE 105

94 Appendix Table 21. Means of cellulose by age and hay. CELL

PAGE 106

95 Appendix Table 22. Means of lignin by age and hay. LIG

PAGE 107

96 4J ^ •H U CO 0) 00 •H •a o > T3 c to to 0) .H D. B n) en to 0) o (U l4-^ <4-l O CO Q) CO >. CO G < CO H 13 0) a. o

PAGE 108

97 Appendix Table 2 A. Least squares means of in vivo intake per unit body weight. (g/kg body weight) Maturity (Weeks Regrowth) OMI 2 4 6 8 22.2 21.6 22.1 21.4 21.5 19.7 23.1 23.7 22.2 23.3 21.4 21.9 17.3 20.0 21.5 18.2 17.2 17.4 17.1 17.3 15.7 19.2 19.6 18.4 19.3 17.8 18.2 Coastcross-1 16.0 14.2 16.6 17.8 ^Abrams (1980) X46-2

PAGE 109

98 Appendix Table 25. Least squares means of in vivo digestibility data. Maturity (Weeks Regrowth) (MD 2 4 6 8 58.3 61.8 59.2 65.2 68.5 60.6 55.0 55.4 60.1 54.5 57.0 54.1 58.1 59.4 55.2 63.2 65.6 61.2 70.2 73.1 64.3 57.7 58.0 62.9 57.3 59.9 57.5 Coastcross-1 60.9 61.1 61.1 55.4 ^Abrams (1980) X46-2

PAGE 110

LITERATURE CITED Abrams, S. M. 1980. Evaluation of kinetic models of ruminant intake and digestibility utilizing tropical grasses. Ph.D. dissertation. University of Florida. Akin, D. E., D. Burdick and H. E. Amos. 1974. Comparative degradation of Coastal bermudagrass. Coastcross-1 bermudagrass and Pensacola bermudagrass by rumen microorganisms revealed by scanning electron microscopy. Crop Sci. 14:537-541. Albersheim, P. 1978. Concerning the structure and biosynthesis of the primary cell walls of plants, p. 127. IN Biochemistry of carbohydrates II. Intl. Rev. of Bchm. Vol 16. D. J. Manners (ed.) University Park Press, Baltimore. A. 0. A. C. 1975. Official methods of analysis, 12th Ed. Assoc. Official Agric. Chemists. Washington, D. C. Aspinall, G. 0. 1970. Pectins, plant gums, and other plant polysaccharides. IN W. Pigman and D. Horton (eds.) The carbohydrates, Vol. IIB. Academic Press, New York. pp. 515-536. Bailey, R. W. 1967. Loss of ingested plant carbohydrates from the reticulo-rumen. N. Z. J. Agric. Res. 10:15-21. Bailey, R. W. 1973. Structural carbohydrates. IN G. W. Butler and R. W. Bailey (eds.) Chemistry and biochemistry of herbage. Vol. 1, Academic Press, London, pp. 157-212. Bailey, R. W. , and H. E. Connor. 1973. Structural carbohydrates in the arundinoid genus Cortaderia . Phytochemistry 12:1197-1198. Bailey, R. W. , and D. I. H. Jones. 1971. Hydrolysis of ryegrass structural carbohydrates with carbohydrases in relation to rumen digestion. N. Z. J. Agric. Res. 14:847-857. Bailey, R. W. , and D. I. H. Jones. 1973. Enzymatic hydrolysis of cell wall preparations from various ryegrasses. N. Z. J. Agric. Res. 16:197-203. Bailey, R. W. , and J. C. MacRae. 1970. The hydrolysis by rumen and caecal microbial enzymes of hemicelluloses in plant and digesta particles. J. Agric. Sci., Cambridge 75:321-326. 99

PAGE 111

100 Bailey, R. W., J, A. Monro, S. E, Pickmere and A. Chesson. 1976. Herbage hemicellulose and its digestion by the ruminant. 1^ Carbohydrate research in plants and animals. Misc. Papers 12. pp. 1-15. Bailey, R. W. , and M. J. Ulyatt. 1970. Pasture quality and ruminant nutrition II. Carbohydrate and lignin composition of detergent-extracted residues from pasture grasses and legumes. N. Z. J. Agric. Res. 13:591-604. Barnes, R. F., and G. C. Marten. 1979. Recent developments in predicting forage quality. J. An. Sci. 48(6) : 1554-1561. Bauer, W. D,, K. Talmadge, K, Keegstra and P. Albersheim. 1973. The structure of plant cell walls II. The hemicellulose of the walls of suspension cultured sycamore cells. Plant Physiol. 51: 174-187. Beever, D. E., D. J. Thomson, E. Pfeffer and D. G. Armstrong. 1971. The effect of drying and ensiling grass on its digestion in sheep. Sites of energy and carbohydrate digestion. Br. J. Nutri. 26:123-134. Beveridge, R. J., and G. N. Richards. 1975. Investigation of the digestion of cell-wall polysaccharides of spear grass and of cotton cellulose by viscometry and X-ray diffraction. Carbohydrate Res. 43:163-172. Blake, J. D., and G, N. Richards. 1971. An examination of some methods of fractionation of plant hemicelluloses. Carbohydrate Res. 17:253-268. Buchala, A. J., C. G. Eraser and K. C. B. Wilkie. 1972. Extraction of hemicellulose from oat tissues during the process of delignification. Phytochemistry 11:1249-1254. Buchala, A. J., and K. C. B. Wilkie. 1973. Uronic acid residues in the total hemicellulose of plants. Phytochemistry 12:655-659. Butler, G. W. , and R. W. Bailey. 1973. Chemistry and biochemistry of herbage. Academic Press, London. Vol. 1. Capote, F. R. 1975. Voluntary intake as affected by age and size of sheep and quality of forage. Ph.D. dissertation. University of Florida. Carbohydrate research in plants and animals. 1976. Misc. Papers 12. Landbouwhogeschool Wageningen, the Netherlands. H. Vreeman and Zonen B. V. Chaves, C. M. 1979. Lignin saponification products and their relation to forage quality. Ph.D. dissertation. University of Florida.

PAGE 112

101 Collings, G. F., and M. T. Yokoyama. 1979. Analysis of fiber components in feeds and forages using gas-liquid chromatography. J. Agric. Food Chem. 27 (2) : 373-377. Conrad, H. R. 1966. Symposium on factors influencing the voluntary intake of herbage by ruminants. Physiological and physical factors limiting feed intake. J. An. Sci. 25:227-235. Cooper, J. P., and N. M. Tainton. 1968. Light and temperature requirements for the growth of tropical and temperate grasses. Herb. Abstr. 38:167-176. Crampton, E. W. , E. Donefer and L. E. Lloyd. 1960. A nutritive value index for forages. J. An. Sci. 19:538-544. Daughtry, C. S. T. 1976. Factors affecting the concentration, composition and in vitro disappearance of tall fescue and orchardgrass hemicellulose. Ph.D. dissertation, Purdue University. Dehority, B. A. 1973. Hemicellulose degradation by rumen bacteria. Fed. Proc. 32(7) :1819-1825. Dehority, B. A., R. R. Johnson and H. R. Conrad. 1962. Digestibility of forage hemicellulose and pectin by rumen bacteria in vitro and the effect of lignif ication thereon. J. Dairy Sci. 45(4) :508512. Dekker, R. F. H. 1976. Hemicellulose degradation in the ruminant. IN Carbohydrate research in plants and animals. Misc. Papers 12. pp. 43-45. Dekker, R. F. H. , and G. N. Richards. 1973. Digestion of polysaccharide constituents of tropical pasture herbage in the bovine rumen. II. Speargrass ( Heteropogon contortus ) . Carb . Res. 27:1-6. Dekker, R. F. H., G. N. Richards and M. J. Playne. 1972. Digestion of polysaccharide constituents of tropical pasture herbage in the bovine rumen. Pt. 1. Townsville Stylo ( Stylosanthes humilis ). Carb. Res. 22:173-185. de la Torre, R. 1974. Micro-histological characteristics of three warm season grasses in relation to forage quality. Ph.D. dissertation. University of Florida. Dubois, M. , K. A. Gilles , J. K. Hamilton, P. A. Rebers and F. Smith. 1956. Colorimetric method for determination of sugars and related substances. Anal. Chem. 20:350-353. Fraser, C. G. , and K. C. B. Wilkie. 1971. 6-glucans from oat leaf tissues at different stages of maturity. Phytochem. 10:1539-1542.

PAGE 113

lo: Gaillard, B. D. E. 1958. A detailed summative analysis of the crude fiber and nitrogen-free extractive fractions of roughages. II. The analyses of straw, hay grass and mangold. J. Sci. Food Agric. 9:346-353. Gaillard, B. D. E. 1965. Comparison of the hemicelluloses from plants belonging to two different plant families. Phytochemistry 4:631-634. Gaillard, B. D. E., and G. N. Richards. 1975. Presence of soluble lignin-carbohydrate complexes in the bovine rumen. Garb. Res. 42:135-145. Goering, H. K. , and P. J. Van Soest. 1970. Forage fiber analysis. Apparatus, reagents, procedures and some applications. USDA Agric. Handbook No. 379. Washington, D. C. Golding, E. J., III. 1973. Formulation of hay: grain diets based on predicted quality of four bermudagrass hays for ruminants. M. S. thesis. University of Florida. Gordon, A. J., and B. D. E. Gaillard. 1976. The relationship between lignin and carbohydrate in the hemicellulose A, B and C fractions extracted from lucerne and wheatstraw with alkali, IN Carbohydrate research in plants and animals. Misc. Papers 12. pp. 55-56. Hartadi, H. 1980. Prediction of the quality of tropical grasses for ruminants by laboratory analyses and summative equations. M. S. thesis. University of Florida. Heaney, D. P. 1970. Voluntary intake as a component of an index to forage quality. Paper Q. IN Proc. Natl. Conf. Forage Qual. Eval. and Util. Neb. Ctr. Cont. Ed., Lincoln, Nebraska. Jermyn, M. R. 1955. Modem methods of plant analysis. Vol. 2 K. Paech and M. V. Tracey (eds.) Springer-Verlag, Berlin. Jones, D. I. H. 1976. Fungal cellulases and hemicelluloses and their application to the analysis of forage. IN Carbohydrate research in plants and animals. Misc. Papers 12. pp. 67-77. Jones, D. I. H., and R. W. Bailey. 1972. Hydrolysis of the cellwall polysaccharides from forage-dried and oven-dried herbage by rumen and mould carbohydrases . J. Sci. Fd. Agric. 23:609-614. Jones, D. I. H. , and R. W. Bailey. 1974. Hydrolysis of the cellwall carbohydrates of grasses by carbohydrases in relation to voluntary intake by sheep. J. Agr. Sci. 83:105-111.

PAGE 114

103 Keegstra, K. , K. W, Talmadge, W. D. Bauer and P. Albersheim. 1973. The structure of plant cell wall. III. A model of the walls of suspension cultured sycamore cells based on the interconnections of the raacramolecular components. Plant Physiol. 51:188-196. Marten, G. C. , and R. F. Barnes. 1980. Prediction of energy digestibility of forages with ^ vitro rumen fermentation and fungal enzyme systems. W_ W. J. Pigden, C. C. Balch and M. Braham (eds.) Standardization of analytical methodology for feeds. IDRC, Ottawa, Canada. McLeod, M. N. , and D. J. Minson. 197A. Differences in carbohydrate fractions between Lolium perenne and two tropical grasses of similar dry-matter digestibility. J. Agric. Sci. 82:449-454. Minson, D. J. 1976. Relation between digestibility and composition of feed. D^. Carbohydrate research in plants and animals. Misc. Papers 12. pp. 101-114. Minson, D. J., and M. N. McLeod. 1970. The digestibility of temperate and tropical grasses. 719-722. Intl. Grassl. Congr. Proc, 11th Surfers Paradise, Australia. Moore, J. E. , and G. 0. Mott. 1973. Structural inhibitors of quality in tropical grasses. 53-98. IN A. G. Matches (ed.) Antiquality components of forages. CSSA Special Publ. 4. Crop Sci. Soc. Am., Madison, Wisconsin. Moore, J. E. , and G. 0. Mott. 1974. Recovery of residual organic matter from in vitro digestion of forages. J. Dairy Sci. 57: 1258-1259. Moore, J. E. , and G. 0. Mott. 1976. Fermentation tubes for in vitro digestion of forages. J. Dairy Sci. 59:167-169. Moore, J. E., G. 0. Mott, D. G. Duncan and R. W. Omer. 1972. Large capacity iri vitro organic matter digestion procedure. J. An. Sci. 34:232. (Abstr.) Morrison, I. M. 1974. Lignin-carbohydrate complexes from Lolium perenne . Phytochemistry 13:1161-1165. Morrison, I. M. 1979. Symposium on 'Carbohydrate Metabolism in the Ruminant. ' Carbohydrate chemistry and rumen digestion. Proc. Nutr. Soc. 38:269-274. Morrison, I. M. 1980. Hemicellulose contamination of acid detergent residues and their replacement by cellulose residues in cell wall analysis. J. Sci. Food Agric. 31:639-645.

PAGE 115

104 Mott, G. 0., and J. E. Moore. 1970. Forage evaluation techniques in perspective, Paper L. IN Proc. Natl. Conf. Forage Qual. Eval. and Util Neb. Ctr. Cont . Ed., Lincoln, Nebraska. Oh,H.K., B. R. Baumgardt and J. M. Scholl. 19661 Evaluation of forages in the laboratory. V. Comparison of chemical analyses, solubility tests, and in vitro fermentation. J. Dairy Sci. 49: 850-855. Ojima, K., and T. Isawa, 1968. The variation of carbohydrates in various species of grasses and legumes. Can. J. Botany 46: 1507-1511. Packett, L. V., M. L. Plumlee, R. Barnes and G. 0. Mott. 1965. Influence of hemicellulose A and B on cellulose digestion, volatile fatty acid production and forage nutritive evaluation. J. Nutr. 85(1):89-101. Preston, R. D. 1974. The physical biology of plant cell walls. Chapman and Hall, London. Reid, J. S. G. , and K. C. B. Wilkie. 1969a. Polysaccharides of the oat plant in relationship to plant growth. Phytochemistry 8: 2045-2051. Reid, J. S. G., and K. C. B. Wilkie. 1969b. Total hemicellulose from oat plants at different stages of growth. Phytochemistry 8: 2059-2065. Riewe, M. E. , and H. Lippke. 1970. Considerations in determining the digestibility of harvested forages. Paper F. IN Natl. Conf. Forage Qual, Eval. and Util. Neb. Ctr. Conf. Ed., Lincoln, Nebraska. Rohweder, D. A., R. F. Barnes and N. Jorgensen. 1976. The use of chemical analyses to establish hay market standards. 1st Intl. Syrap . Feed Comp . , An. Nutr. Req. and Computerization of Diets. P. V. Fonnesbeck, L. E. Harris and L. C. Kearl (eds.) Utah AES, USU, Logan, Utah. pp. 249-257. Routley, D. G., and J. T. Sullivan. 1958. The isolation and analysis of hemicelluloses of bromegrass. J. Agric. Fd. Chem. 6(9):687-692. Russo, S. L. 1976. Assessment of laboratory techniques as estimators of quality of tropical and temperate grass species. M. S. thesis. The Pennsylvania State University. Russo, S. L., J. E. Moore, H. Hartadi and G. 0. Mott. 1981. Hemicellulose content and in vitro digestibility in tropical grasses. Southern Assoc. Agri. Scientists. Agry Div. Abstracts of Tech. Papers. No. 8:7.

PAGE 116

105 SAS. 1979. SAS User's Guide, 1979 Edition. J. T. Helwig and K. A. Council (eds.) SAS Institute, Inc., Raleigh, North Carolina, Smith, L. W., H. K. Goering, D. R. Waldo and C. H. Gordon, 1971. In vitro digestion rate of forage cell wall components. J. Dairy Sci. 54:71-76. Sullivan, J. T. 1966. Studies of the hemicelluloses of forage plants. J. An. Sci. 25(l):83-86. Talmadge, K. W. , K. Keegstra, W. D. Bauer and P. Albersheim. 1973. The structure of plant cell walls. I. The macromolecular components of the walls of suspension cultured sycamore cells with detailed analysis of the pectic polysaccharides. Plant Physiol. 51:158-173. Tessema, S. 1972. Nutritional value of some tropical grass species compared to some temperate grass species. Ph.D. dissertation, Cornell University. Thomson, D. J., D. E. Beever, J. F. Coelho da Silva and D. G. Armstrong. 1972. The effect in sheep of physical form on the sites of digestion of a dried lucerne diet. I. Sites of organic matter, energy and carbohydrate digestion. Br. J. Nutri. 28:31-41. Thornton, R. F., and D. J. Minson. 1972. The relationship between voluntary intake and mean apparent retention time in the rumen. Aust. J. Agr. Res. 23:871-877. Tilley, J. M. A., and R. A. Terry. 1963. A two-stage technique for the in vitro digestion of forage crops. J. Brit. Grassl. Soc. 18:104-111. Ulyatt, M. J., D. W. Dellouz, C. S. W. Reid and T. Bauchop. 1975. Structure and function of the large intestine of ruminants. W I. W. McDonald and A. C. I. Warner (eds.) Digestion and metabolism in the ruminant. Proc. IV. Intl. Sym. Rum. Phys . U. New England Publ. Unit, Armidale. Ulyatt, M. J., and A. R. Egan. 1979. Quantitative digestion of fresh herbage by sheep. V. The digestion of four herbages and prediction of sites of digestion. J. Agric. Sci., Cambridge 92: 605-616. Ulyatt, M. J., and J. C. MacRae. 1974. Quantitative digestion of fresh herbage by sheep. I. The sites of digestion of organic matter, energy, readily fermentable carbohydrate, structural carbohydrate and lipid. J. Agric. Sci., Cambridge 82:295-307.

PAGE 117

106 Van Soest, P. J. 1965. Symposium on factors influencing the voluntary intake of herbage by ruminants: voluntary intake in relation to chemical composition and digestibility. J. An. Sci. 24:834-843. Van Soest, P. J. 1967. Development of a comprehensive system of feed analyses and its application to forages. J. An. Sci. 26: 119-128. Van Soest, P. J. 1975. Physico-chemical aspects of fiber digestion. IN Digestion and metabolism in the ruminant. Univ. New England Publ. Unit, Armidale. pp. 351-365. Van Soest, P. J., and R. H. Wine. 1968. Determination of lignin and cellulose in acid-detergent fiber with permanganate. J. Assoc. Off. Agric. Chem. 51:780. Van't Klooster, A. Th. , and B. D. E. Gaillard. 1976. Digestion of carbohydrates in the intestines of dairy cows. IN Carbohydrate research in plants and animals. Misc. Papers 12. pp. 79-94. Waite, R. , M. J. Johnston and D. G. Armstrong. 1964. The evaluation of artificially dried grass as a source of energy for sheep. I. The effect of stage of maturity on the apparent digestibility of ryegrass, cockfoot and timothy. J. Agric. Sci. 62:391-398. Waldo, D. R., L. W. Smith and E. L. Cox. 1972. Model of cellulose disappearance from the rumen. J. Dairy Sci. 55:125-129. Weston, R. H. 1966. Factors limiting the intake of feed by sheep. I. The significance of palatibility , the capacity of the alimentary tract to handle digesta, and the supply of glucogenic substrate. Aust. J. Agr. Res. 17:939-954. Whistler, R. L., J. Bachrach and D. R. Bowman. 1948. Preparation and properties of corn cob holocellulose. Archives of Biochemistry. 19:25-33. Whistler, R. L. , and D. B. E. Gaillard. 1961. Comparison of xylans from several annual plants. Arch. Bchm. and Bphys. 93: 332-334. Whistler, R. L. , and E. L. Richards. 1970. Hemicelluloses. IN W. Pigman and D. Horton (eds.) The carbohydrates. Vol. IIA. Academic Press, New York. pp. 447-469. Wilkie, K. C. B. 1979. The hemicelluloses of grasses and cereals. IN Advances in carbohydrate chemistry and biochemistry. R. Stuart Tyson and D. Horton (eds.) Academic Press, New York. 36:215-264.

PAGE 118

107 Wilkie, K. C. B., and S. L. Woo. 1976. Non-cellulosic 6-D-glucans from bamboo, and interpretive problems in the study of all hemicelluloses. Carb. Res. 49:399-409, Wilson, J. R. , and C. C. W. Ford. 1971. Temperature influences on the growth, digestibility and carbohydrate composition of two tropical grasses, Panicum maximum var. trichoglume and Setaria sphacelata and two cultivars of the temperate grass Lollum perenne . Aust. J. Agric. Res. 22:563-571.

PAGE 119

BIOGRAPHICAL SKETCH Sandra L. Russo was born November 24, 1948, at the Great Lakes Naval Training Base, Illinois. She graduated from La Porte High School, La Porte, Indiana, in June 1966. From September 1966 until June 1968 she attended Purdue University — North Central Campus, Westville, Indiana, on a part-time basis. From June 1968 until August 1971, she attended Purdue University, West Lafayette, Indiana. She worked in the Agronomy Department for Drs. Robert Barnes and Ray Bula from September 1968 to August 1971. She received the Bachelor of Science degree in animal science in August 1971. From 1971 to 1973 she worked as a teaching technologistlaboratory technician at Purdue University — North Central Campus, Westville, Indiana. In 1974 she entered the Pennsylvania State University in agronomy, where she obtained the Master of Science degree in agronomy in August 1976. While in Pennsylvania, the author was co-founder and president of the Centre County Rape Crisis Center. In 1977 she entered the University of Florida to continue graduate studies in agronomy and animal nutrition. Presently she is a candidate for a certificate in Tropical Agriculture and the degree of Doctor of Philosophy in the Department of Agronomy, University of Florida. 108

PAGE 120

109 The author has been a member of the Social, Agricultural and Food Scientists group since its inception in 1978. She is a member of the American Society of Agronomy. The author is married to Allen F. Cook and is the stepmother of three children, a mother-in-law and a grandmother.

PAGE 121

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. erald 0. Mott, Chairmaii Professor of Agronomy a^C^ I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. ^."h itr/^A^ John E. Moore ^Professor of Animal Science I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. 0. C. Ruelke Professor of Agronomy I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy.

PAGE 122

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. •/ ' :,M: ^ y. Donald A. Graetz ' Associate Professor of Soil Science This dissertation was submitted to the Graduate Faculty of the College of Agriculture and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June 1981 VOcA ^' c// Dean, College of Agricul (/ Dean, Graduate School