comparative forage evaluation using infrared reflectance
TRANSCRIPT
Louisiana State UniversityLSU Digital Commons
LSU Historical Dissertations and Theses Graduate School
1982
Comparative Forage Evaluation Using InfraredReflectance Spectroscopy, Microbial, Enzymaticand Chemical Analyses.Miguel Baptista CoelhoLouisiana State University and Agricultural & Mechanical College
Follow this and additional works at: https://digitalcommons.lsu.edu/gradschool_disstheses
This Dissertation is brought to you for free and open access by the Graduate School at LSU Digital Commons. It has been accepted for inclusion inLSU Historical Dissertations and Theses by an authorized administrator of LSU Digital Commons. For more information, please [email protected].
Recommended CitationCoelho, Miguel Baptista, "Comparative Forage Evaluation Using Infrared Reflectance Spectroscopy, Microbial, Enzymatic andChemical Analyses." (1982). LSU Historical Dissertations and Theses. 3752.https://digitalcommons.lsu.edu/gradschool_disstheses/3752
INFORMATION TO USERS
This reproduction was made from a copy of a document sent to us for microfilming. While the most advanced technology has been used to photograph and reproduce this document, the quality of the reproduction is heavily dependent upon the quality of the material submitted.
The following explanation of techniques is provided to help clarify markings or notations which may appear on this reproduction.
1. The sign or “target” for pages apparently lacking from the document photographed is “Missing Page(s)”. If it was possible to obtain the missing page(s) or section, they are spliced into the film along with adjacent pages. This may have necessitated cutting through an image and duplicating adjacent pages to assure complete continuity.
2. When an image on the film is obliterated with a round black mark, it is an indication of either blurred copy because of movement during exposure, duplicate copy, or copyrighted materials that should not have been filmed. For blurred pages, a good image of the page can be found in the adjacent frame. If copyrighted materials were deleted, a target note will appear listing the pages in the adjacent frame.
3. When a map, drawing or chart, etc., is part of the material being photographed, a definite method of “sectioning” the material has been followed. It is customary to begin filming at the upper left hand comer of a large sheet and to continue from left to right in equal sections with small overlaps. If necessary, sectioning is continued again—beginning below the first row and continuing on until complete.
4. For illustrations that cannot be satisfactorily reproduced by xerographic means, photographic prints can be purchased at additional cost and inserted into your xerographic copy. These prints are available upon request from the Dissertations Customer Services Department.
5. Some pages in any document may have indistinct print. In all cases the best available copy has been filmed.
University.Microfilms
International300 N. Zeeb Road Ann Arbor, Ml 48106
8229495
Coelho, Miguel Baptists
COMPARATIVE FORAGE EVALUATION USING INFRARED REFLECTANCE SPECTROSCOPY, MICROBIAL, ENZYMATIC AND CHEMICAL ANALYSES
The Louisiana State University and Agricultural and Mechanical Col PhJD. 1982
University Microfilms
International 300 N. Zeeb Rout, Ann Aitoor, MI 48106
COMPARATIVE FORAGE EVALUATION USING
INFRARED REFLECTANCE SPECTROSCOPY, MICROBIAL,
ENZYMATIC AND CHEMICAL ANALYSES
A DISSERTATION
Submitted to the Graduate Faculty of the Louisiana State University and
Agricultural and Mechanical College in partial fulfillment of the
requirements for the degree of Doctor of Philosophy
in
The Department of Animal Science
by
Miguel B. Coelho
B.S., Lisbon Technical University, 1978
August 1982
ACKNOWLEDGEMENTS
The author wishes to express gratitude and sincere appreciation to
Dr. F. Glen Hembry for the valuable friendship, advice and constructive
criticism given during the course of this program of graduate study,
and in the preparation of this manuscript.
Appreciation is also extended to Dr. P. E. Humes, Dr. D. F. Franke,
Dr. T. D. Bidner, Dr. J. W. Turner, Dr. D. L. Robinson and Dr. K. L.
Koonce for serving on the author's graduate committee and for the
critical reading of this manuscript.
Appreciation is given to Deborah Babcock for her aid in the statis
tical analysis of the experimental data.
The author also thanks Dr. "Woody" Barton, 11 for his assistance in
obtaining the Near Infrared Spectroscopy analysis.
The financial assistance of North Atlantic Treaty Organization,
Fulbright and the LSU Department of Animal Science, trtiich made this work
possible is gratefully acknowledged.
Lastly, the author wishes to acknowledge the encouragement and the
typing of this manuscript by his wife Betty N. Coelho.
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS................................................. ii
LIST OF TABLES....................................... v
LIST OF FIGURES....................................... viii
LIST OF APPENDIX TABLES..................................... ix
ABSTRACT....................................................... xi
INTRODUCTION ................................................. 1
LITERATURE REVIEW............................................. 3
I. Characteristic Differences Between Grasses andLegumes ........................................... 3
II. Characteristic Differences Between Tropical andTemperate Forages .................................. 4
III. Effect of Stage of Maturity on Nutritive Value ofTropical Forages.................................... 6
IV. Voluntary Intake and Digestibility.................... 8Control of Voluntary Intake ....... . 9Factors Affecting the Measurement of Intake ....... 10Digestion Trials.......................... 12Associative Effect in Digestion of Legumes andGrasses............................................13
V. In Vitro Fermentation Techniques....................... 14Effect of Volume of Inoculum on IVDMD............16Effect of Length of Fermentation Time on IVDMD. . . . 17Effect of Inoculum Source on IVDMD................... 18
VI. Forage Evaluation Using Enzymes ..................... 19VII, Chemical Methods of Forage Evaluation ............... 21
Proximate Analysis.................................. 21The Detergent System................................ 22
VIII. Relationship Among Chemical Components, LaboratoryValues and In Vivo Data.............................. 23
IX. Physical Methods of Forage Evaluation . . ............ 28Application of Near Infrared ReflectanceSpectroscopy to Forage Evaluation ........... . . . 28
MATERIALS AND METHODS
Page
35
I, Hays and Animals................................... 35II. Voluntary Intake and Digestion Trials ............... 35III. Chemical Analysis ................................. 36IV. In Vitro Fermentations................................ 37V. Enzyme Digestion Studies.............................. 39VI. Near Infrared Reflectance Spectroscopy Analysis . . . . 40VII. Statistical Analysis..................................41
RESULTS AND DISCUSSION ........................................ 43
I. Chemical Composition of the Hays Studied and Resultsof the Intake and Digestion Trials..................... 43
II. In Vitro Ruminal Fermentation Results................ 49III. Enzyme Digestion Results.............................. 73IV. Results From the NIR Study............................ 77V. Prediction of Forage Nutritive Value Using Simple
and Multiple Linear Regression.........................82
SUMMARY......................................................... 89
LITERATURE CITED ............................................. 93
APPENDIX TABLES.................................................104
VITA........................................................ .121
iv
24
26
30
44
45
47
48
50
51
53
55
57
62
63
63
LIST OF TABLES
Correlation Between Dry Hatter Intake (DMI) and Measures of Forage Nutritive Value ...................
Correlation Between In Vivo Dry Matter Digestibility (DMD) and Measures of Forage Nutritive Value .........
Typical NIR Wavelengths Used in Regression Analysis for Each Forage Component........... .
Chemical Composition and IVDMD of Hays Averagedfor all Periods (%)....... .........................
Digestibility of Chemical Components of Hays Averaged for all Periods (%)..................................
Voluntary Intake (DMI), Apparent Digestibility (DMD) and Digestible Energy (DE) of Hays Averaged for all Periods.............................................
Overall Correlation Coefficients (r) Among Various Parameters of Forage Quality in the Hays .............
Overall Correlation Coefficients (r) Among Digestibility of Various Parameters of Forage Quality of the Hays and Dry Matter Intake....................................
Effect of Ruminal Fluid Dilution on IVDMD of Alfalfa Hay
Effect of Ruminal Fluid Dilution on Average IVDMD for all Hays ...........................................
Effect of Further Ruminal Fluid Dilution on IVDMD of Alfalfa Hay.........................................
Effect of Length of Fermentation Time on Average IVDMD for all Hays .......................................
Forage x Time Interaction for Several Dilution Rates of Ruminal Fluid......... .........................
Predicted Effect of Hay on Rate of IVDMD (6) from 24-96 Hr Fermentation for Two Ruminal Fluid Dilutions. .
Predicted Effect of Hay on Rate of IVDMD (b) from 24-48 Hr and 48-96 Hr Fermentation for Two Ruminal Fluid Dilutions......................................
v
Table Page
16 Associative Effect of Alfalfa and Ryegrass on ForageQuality Measurements .................................. 68
17 Analysis of Variance for Effect of Inoculum Donor,Dilution Rate of Ruminal Fluid and Length ofFermentation Time on IVDMD of Alfalfa Hay................. 70
18 Comparison of In Vitro Dry Matter Disappearance ofHays Among Ruminal In Vitro Procedures................... 71
19 Effect of In Vitro Ruminal Procedure on IVDMD.............. 71
20 Validity Test for the Prediction of Dry Matter Digestibility from Different In Vitro RuminalProcedures............................................. 72
21 Comparison of In Vitro Dry Matter Disappearanceof Hays Among In Vitro Enzymatic Procedures............... 74
22 Effect of In Vitro Enzymatic Procedure on IVDMD............ 75
23 Validity Test for the Prediction of Dry MatterDigestibility from Different Enzyme Procedures . . . . . . 76
24 Validity Test for the Prediction of In Vitro Dry MatterDisappearance from Different Enzyme Procedures ......... 76
25 Summary of Multiple Linear Regression Analysis Relating the Chemical Composition of Hay andIn Vivo Data to NIR...................................... 81
26 Prediction of Dry Matter Intake from LaboratoryAnalysis with Simple Regression...........................83
27 Variables Used in Prediction of Dry Matter Intakefrom Laboratory Analysis with Stepwise Regression..........84
28 Prediction of Dry Matter Intake (g/W'^g) fromProposed Equations .................................... 84
29 Prediction of Dry Matter Digestibility from LaboratoryAnalysis with Simple Regression...........................86
30 Variables uBed in the Prediction of Dry MatterDigestibility from Laboratory Analysis withStepwise Regression...................................... 86
vi
Table Page
31 Prediction of Dry Matter D i g e s t i b i l i t y from ProposedEquations..................... . . 87
32 Prediction of In Vitro Dry M a t t e r Disappearance fromLaboratory Analysis with Simple R e g r e s s i o n ................. 88
33 Variables Used in Prediction of I n Vitro Dry MatterDisappearance from Laboratory A n a l y s i s with Stepwise Regression................ . «. . 88
vii
LIST OF FIGURES
Figure Page
1 Moat Common Locations of Wavelengths Absorbed by Forage Components Superimposed on Alfalfa HayReflectance Log (1/R) Spectra.............................31
2 Regression of In Vitro Dry Matter Disappearance of Alfalfa Hay on Percent Added Water toRuminal Fluid.............................. 52
3 Regression of In Vitro Dry Matter Disappearance of Alfalfa Hay on Increased Percent Added Waterto Ruminal Fluid.................... 56
4 Predicted Effect of 2:8:40 Ruminal Fluid Dilution onIVDMD of Several Hays.................................... 58
5 Predicted Effect of 10:0:40 Ruminal Fluid Dilution onIVDMD of Several Hays.................................... 59
6 Predicted Effect of 25:0:25 Ruminal Fluid Dilution onIVDMD of Several Hays.............. 60
7 Predicted Effect of Fermentation Length on IVDMD forRyegrass H a y .......................... 64
8 Predicted Effect of Fermentation Length on IVDMD forAlfalfa Hay..............................................65
9 Predicted Effect of Fermentation Length on IVDMD forCommon Bermudagrass Hay................................ 66
10 Predicted Effect of Fermentation Length on IVDMD forPensacola Bahiagrass Hay .............................. 67
11 Reflectance (log 1/R) Spectra of Ryegrass and AlfalfaHays ............... 78
12 Superimposition of Diet and Fecal Spectra of Ryegrassand Alfalfa Hays................................. . . . 79
13 Superimposition of Digestibility Spectra of Bahiagrassand Mature Ryegrass Hays................................ 80
viii
LIST OF APPENDIX TABLES
Table Page
1 Composition of "Kansas State Buffer" for In VitroDry Matter Disappearance Analysis ............. . . . . 105
2 Composition of Artificial Rumen Fluid (ARF) Bufferfor In Vitro Dry Matter Disappearance Analysis. . . . . . 106
3 Composition of McDougall's Buffer for In Vitro DryMatter Disappearance Analysis ......................... 107
4 Composition of Sodium Acetate : Acetic Acid Bufferfor Enzyme Assays.......................................107
5 Date of Cutting and Stage of Regrowth of the HaysStudied................................................ 108
6 Analysis of Variance for Digestible Energy................ 109
7 Analysis of Variance for Dry Matter Digestibility . . . . 109
8 Analysis of Variance for Dry Matter Intake................ 109
9 Analysis of Variance for Dilution of Ruminal Fluidwith Alfalfa Hay as the Substrate.............. '. . . . 110
10 Analysis of Variance for Further Dilution of RuminalFluid with Alfalfa Hay as the Substrate..................110
11 Effect of Dilution Rate of Ruminal Fluid and Lengthof Fermentation Time on IVDMD ....................... Ill
12 Analysis of Variance for Effect of Dilution Rate of Ruminal Fluid and Length of Fermentation Time onIVDMD of Several Hays................................... 112
13 Effect of Hay on IVDMD................................... 113
14 Effect of Inoculum Donor* Dilution Rate of RuminalFluid and Length of Fermentation Time on IVDMD ofAlfalfa H a y .............................................114
15 Effect of Length of Fermentation Time on IVDMD ofAlfalfa H a y .............................................115
16 Effect of Ruminal Fluid Dilution on IVDMD of AlfalfaH a y .................................................... 115
ix
Table Page
17 Analysis of Variance for In Vitro Ruminal Procedures . . 116
18 Analysis of Variance for In Vitro Enzymatic Procedures . 116
19 Comparison of Laboratory Values and NIR PredictedValues for Various Quality Constituents of Common Bermudagrass . . . . . . . . 117
20 Prediction Equations of Dry Matter Intake fromLaboratory Analysis with Simple Regression ........... 118
21 Prediction Equations of Dry Matter Intake fromLaboratory Analysis with Stepwise Regression ......... 118
22 Prediction Equations of Dry Matter Digestibilityfrom Laboratory Analysis with Simple Regression....... 119
23 Prediction Equations of Dry Matter Digestibility fromLaboratory Analysis with Stepwise Regression ......... 119
24 Prediction Equations of In Vitro Dry MatterDisappearance from Laboratory Analysis withSimple Regression.................................... 120
25 Prediction Equations of In Vitro Dry MatterDisappearance from Laboratory Analysis withStepwise Regression.................................. 120
x
ABSTRACT
A 6x6 Latin Square design intake and in vivo digestion trial was
conducted using six crossbred steers to evaluate the nutritive value of
alfalfa, mature ryegrass, common and Alicia bermudagrass, Pensacola
bahiagrass and a 50:50 mixture of alfalfa and ryegrass hays. Voluntary
dry matter intake (DMI) and in vivo dry matter digestibility (DMD)
ranked the hays in the same order of quality. Alfalfa hay and mature
ryegrass had the highest and lowest quality, respectively. The quality
of warm season grasses decreased as the length of regrowth period after
cutting increased.
In vitro dry matter disappearance (IVDMD) increased with increased
dilution of ruminal fluid regardless of the donor source of ruminal
fluid inoculum or hay substrate. In vitro dry matter disappearance for
all hays increased as the length of incubation time increased. Alfalfa
hay, which had the highest IVDMD, showed the least increase in IVDMD
from 48 to 96 hr. All in vitro ruminal procedures significantly pre
dicted DMD. The Moore procedure with phosphate buffer and the
Mellenberger procedure had the highest and lowest coefficient of deter
mination in predicting DMD, r2=.88 and r2“.72, respectively. Enzymatic
procedures also significantly predicted DMD, although they had a wider
range than the ruminal procedures. Pepsin-T. viride cellulase + A.
niger hemicellulase and pepsin-Onozuka cellulase preparation had the
highest and lowest coefficient of determination in predicting DMD,
r2".88 and r2".69, respectively.
Dry matter intake and digestibility were both significantly pre
dicted by crude protein (CP) and IVDMD. The best predictors for DMI
and DMD were CP and IVDMD, respectively. Acid detergent fiber (ADF)
significantly predicted DMD, but neutral detergent fiber (NDF) failed
to significantly predict DMI.
Near infrared reflectance (NIR) (log 1/R) spectra positively iden
tified differences in digestibility between alfalfa and mature ryegrass
and between bahiagrass and mature ryegrass. Near infrared reflectance
spectroscopy predicted in vivo data with highest accuracy for DMI
(r2*,84) and lowest accuracy for DMD (r2*.75). Prediction by NIR of
laboratory analysis was highest for NDF (r2".98) and lowest for crude
fiber (CF) (r2=.87). Errors in predicting animal responses to forages
were greater than those in predicting chemical composition of the
forages.
INTRODUCTION
Forages, which consist to varying degrees of potentially digesti
ble components, are an important source of nutrients for ruminants and
other animals. Fifty percent or more of the potentially useful energy
of forages can be obtained from cellulose and hemicellulose. Cellulose
is the most abundant carbohydrate in forages and can only be digested
by ruminants and a few nonruminant animals due to their symbiotic asso
ciation with gastro-intestinal microflora.
Profitability of grain feeding has been decreasing steadily during
the past decade; therefore, ruminant production is becoming more de
pendent than ever on forage utilization. Thus, it is imperative that
producers possess accurate information relative to the productive ca
pacity of forages available to them. This productive capacity depends
upon the availability of nutrients present in the forage, the quantity
of forage consumed and the efficiency of utilization of ingested nutri
ents .
The feeding of the forage is the most accurate method of forage
evaluation. While the accuracy of feeding and digestion trials can not
be denied, they are time consuming, expensive and very slow in produc
ing results. Therefore, the quest for simple, quick, reliable and in
expensive laboratory methods for prediction of forage quality contin
ues. Recently, many laboratory procedures have been developed for es
timating forage nutritive value. The objectives of the present study
were:
1. To compare nutritive values of legumes with grasses and the
1
2
value of tropical with temperate grasses.
3. To further evaluate and verify old and new laboratory methods
of forage evaluation including proximate analysis, detergent
analysis, in vitro fermentation, enzymatic techniques and
chemometric techniques such as near infrared reflectance
spectroscopy.
4. To determine correlations among laboratory estimates and in
vivo data, and to develope prediction equations for various in
vivo parameters.
LITERATURE REVIEW
I. CHARACTERISTIC DIFFERENCES BETWEEN GRASSES AND LEGUMES
Over a range of maturity, legumes have less cell wall material
than grasses. The cell wall of legumes when compared to grasses con
tains considerably less hemicellulose and more lignin (Riewe and
Lippke, 1969). Van Soest (1967) noted that hemicellulose appears to be
one of the most important fractions relating to the nutritive charac
teristics of grasses and legumes. When compared to grasses, alfalfa
contains a smaller but more highly lignified holocellulose fraction
that is considerably less digestible, thus leaving a greater proportion
of the dry matter of alfalfa free of lignin. As the cellulose content
of grasses and alfalfa are about equal, the principal species differ
ence lies in the higher proportion of hemicellulose present in the
grasses.
In forages such as alfalfa the high lignin/cellulose significantly
reduces the fermentation of the cell wall, and the slope of the diges
tion curve early in the fermentation period is greater and cumulative
digestion plateaus much more quickly than it does in grasses where the
lignin/cellulose is lower. Grass cell walls continue to be signifi
cantly degraded after legume fermentation has essentially ceased (Van
Soest, 1970).
Robles et al. (1980) studied the cell wall digestibility of
alfalfa and orchardgrass. Alfalfa contained less cell wall at the
3
4
beginning of in vivo digestion, more cell wall residue at the end of
digestion and less potentially digestible cell wall than did orchard-
grass. The cell wall digestibilities of alfalfa and orchardgrass were
44 and 61 percent, respectively.
Johnson et al. (1962) found considerable differences between
grasses and alfalfa when correlating in vitro digestibility with in
vivo digestion trial data. The correlation coefficients between in
vitro cellulose digestibility and in vivo digestibility coefficients
for dry matter, cellulose and energy were very high when only grasses
were considered. When the data from alfalfa were included in the anal
ysis, the correlation between DMD and energy and cellulose digestibil
ities were lower. The correlation between the nutritive value index
and cellulose digestibility was much higher with only grasses, than
when alfalfa was included.
II. CHARACTERISTIC DIFFERENCES BETWEEN TROPICAL AND TEMPERATE FORAGES
There is ample evidence of basic inherent differences between
tropical and temperate grasses. Taxonomically, the tropical and tem
perate grasses belong to different subfamilies, therefore the many
physiological and anatomical differences observed are mainly the result
of genetic differences. The most important difference is in the carbon
pathway for photosynthesis trtiich is generaly associated with the number
and disposition of bundle sheaths. Tropical plants have more vascular
bundles, both in the stem and leaves, and the bundles are usually dou
ble sheathed (Esau, 1977). Therefore, a sizable portion of readily di
gestible nutrients are unaccessible to microbial degradation. Leaf, as
5
a rule, has higher DMD, DM1 and palatability than stem, and its propor
tion to stem significantly influences animal performance. Leaves from
tropical grasses have less readily digested mesophyl and more of the
less digestible epidermis, vascular and sclerenchyma tissues (Wilson
and Minson, 1980). The mesophyl is less densely packed in temperate
grasses. This layer of intercellular spaces allows the rumen microor
ganisms quicker access to a larger surface area (Hanna et al., 1973).
Transpiration is higher in tropical than in temperate grasses
(Minson and McLeod, 1970). Deinum (1966) concluded that the crude
fiber content of plants was also related to their transpiration rates.
Photorespiration is lower in tropical than in temperate grasses, and
the maximum level of photosynthesis is higher in tropical than in tem
perate grasses (Moore and Mott, 1972). Moir et al. (1977) concluded
that high temperature increased the proportion of cell wall and de
creased its digestibility in both leaf and stem due partly to higher
growth rate and greater stem development. Wilson and Minson (1980)
calculated the decrease in DMD for each degree centigrade rise in tem
perature. Percent DMD decreased at the rate 0.60 (tropical grasses),
0.56 (temperate grasses), 0.28 (tropical legumes) and 0.21 (temperatet
legumes) for each centigrade degree increase.
Minson and McLeod (1970) concluded that the mean digestibility of
tropical forages was 12.8 percentage units lower than that for the tem
perate forages due to the higher fiber content in tropical grasses.
Composition data from a number of published reports show that cell wall
contents (CWC), ADF and L in tropical grasses are higher than in tem
perate grasses while 1VDMD is lower for tropical grasses as illustrated
6
below (Moore and Mott, 1972; Montgomery et al., 1979; Pendlum et al.,
1980 and Riewe and Lippke, 1969).
TEMPERATE GRASSES TROPICAL GRASSES
IVDMD 42-70 40-60
CWC 34-73 45-83
ADF 18-46 21-57
L 1-11 2-12
At similar stages of maturity, the warm season annual or tropical
grasses have a higher CWC than cool season or temperate grasses. The
warm season grasses may have a CWC lower than 50 percent in the very im
mature stages, but this rapidly increases to 70 percent or more (Moore and
Mott, 1972). Cell wall material of warm season perennial grasses in
creases more rapidly than in the cool season grasses. This increase ap
pears to be directly associated with rapid growth and dry matter accumula
tion (Riewe and Lippke, 1969).
III. EFFECT OF STAGE OF MATURITY ON NUTRITIVE VALUE OF TROPICAL FORAGES
It has repeatedly been shown that the general trend is for
fast-growing forages to decline in protein and increase in fiber as they
mature.
Burton et al. (1963) studied the effects of cutting frequency and ni
trogen application on the chemical composition of Coastal bermudagrass.
Cutting intervals of 3, 4, 5, 6, 8, 12 and 24 weeks resulted in crude pro
tein values of 18.5, 16.4, 15.4, 13.3, 10.7, 9.0 and 8.4 percent,
6 0 8 3 5 7 1 2
7
respectively. Crude fiber increased steadily from 27.0 to 33.9 percent
from the 3rd to the 24th week cutting.
As forages mature, the composition of the total cell wall material
changes. Lignification of the cell wall in legumes increases sharply,
and percent hemicellulose content in the cell wall declines with ad
vance in maturity. Montgomery et al. (1979) studied the chemical com
position of warm season grasses cut at four week intervals from May to
September. Lignin increased linearly in bahiagrass from 3.26 percent
in May to 5.28 percent in September, while the lignin content of
common, Coastal and Alicia bermudagrasses did not increase during the
same time interval. No significant change in hemicellulose content of
the grasses was observed from May to September.
When cell wall contents reach 50 to 60 percent of the forage dry
matter, the fiber mass begins to adversely affect digestibility and in
take (Moore et al., 1980). In forages with a low cell wall content,
digestibility and intake apparently are not related. Therefore, the
relationship between digestible dry matter and voluntary intake may de
pend on the proportion of digestible energy contributed by cell wall
constituents.
Recently, Akin and Burdick (1975) used light and electron micros
copy to relate microanatomy with digestibility in Coastal bermudagrass
and tall fescue. Lignified tissues of Coastal bermudagrass remained
completely undegraded after 72 hourB of incubation compared to appre
ciable removal of the same tissues in the tall fescue.
Donefer et al. (1960) reported a decrease in rate of digestion of
plant cellulose after 12 hours of fermentation. This decrease seemed
8
to be related to lignin content because the fermentation rate of puri
fied cellulose did not decrease until digestion was almost complete.
Packett et al. (1965) indicated that forage delignification always in
creased the amount of cellulose digested.
IV. VOLUNTARY INTAKE AND DIGESTIBILITY
Voluntary intake has an important place in forage evaluation.
Barnes (1966) concluded that animal performance was more related to vol
untary intake than to digestibility. The range of percent dry matter
intake from low to high quality forages was about 2 1/2 times greater
than for digestibility. The lower accuracy of using intake to measure
forage quality results from the fact that intake measurements are less
precise than these of digestibility. Heaney (1970) concluded that true
digestibility differences of 2 percentage units could be detected with
3-4 animals per measurement; whereas, about 10 animals are required to
detect true intake differences of 10 percentage units. Therefore,
neither intake nor digestibility individually are reliable evaluators of
differences in feeding value between forage species.
The proposal to use the voluntary intake of a forage as a quanti
tative measure of its nutritive value was made by Crampton (1957) and
expressed as a nutritive value index (NVI) where both relative intake
and gross energy digestibility of the feed contributed to the index
value. Crampton et al. (1962) developed this measurement into an ab
solute digestible energy intake potential of a forage in terms of Real
DE/W’ I . Heaney (1970), Jones (1972) and Milford and Minson (1965)
agreed that intake and digestibility should be combined for determining
9
quality and that digestible energy intake was the most effective method
of expressing the quality of a forage for animal performance. More re
cently, Moore et al. (1980) in a study of 41 southern forages, observed
a correlation between dry matter intake and digestibility of only .69.
Therefore, they concluded that voluntary intake and nutrient digest
ibility must be considered separately, because they are often not
closely related across forage species.
Control of Voluntary Intake. It has been well established that
when feed resources are available, mature animals consume feed at con
stant rates and maintain body weight at nearly constant levels for long
periods of time. Bines (1969) concluded that these phenomena imply a
long term regulatory mechanism of feed intake. Blaxter et al. (1966)
concluded that When digestibility was less than 67 percent, feed intake
was related to the physical control factors of body weight, undigested
residue per unit of body weight, rate of passage and dry matter digest
ibility. When digestibility was greater than 67 percent, feed intake
was related to the physical control factors, metabolic size, digest
ibility and production. Voluntary intake of feed is also limited by
the physical capacity of the gut due to its relationship with rumen
fill and rate of passage (Campling et al., 1962). Several researchers
have shown that ruminants eat forage to a constant fill (Blaxter et
al., 1961). With most forages, the physical limit is the primary
controling factor. The physical limit has been described as a
"distention" or "fill" mechanism which assumes that ruminants eat until
some part of the gut is full of undigested bulk. Thiago et al. (1979)
10
confirmed the existence of the physical control mechanism by showing
that the rumen contained the same amount of dry matter over a range of
forage qualities. Differences in intake must be due, therefore, to
differences in rumen capacity and the rate of clearence from the rumen
of forage fiber.
The "hotel theory" that Van Soest (1975) proposed, may be the most
graphic description of the relationship between cell wall degradation
and voluntary intake. This theory states that cell wall degradation is
analogous to the demolition of a hotel in that removal of the furniture
and even a few interior walls does not change the space occupied by the
outer wall of the hotel. In the rumen, forage particles occupy space
(bulk) until their structure is degraded to the point of collapse and
the resulting particle size is small enough to leave the rumen.
Factors Affecting the Measurement of Intake. Heaney et al. (1968)
reported that the coefficient of variation of intake due to animal var
iability was 16 percent for all forages, but if straws were excluded
the coefficient of variation dropped to 14 percent. The coefficient of
variation from determination on straws may be 40-80 percent higher than
for other forages. Raymond and Minson (1955) concluded that feed al
lowances which cause significant weighbacks can result in detectable
increases in forage digestibility, because the animal will naturally
select the more digestible portion of the plant material and leave the
less digestible portion. On the other hand, "the animals should be fed
enough daily feed to provide weighbacks of at least 5 percent to ensure
they are truly on an ad libitum regime. Campling et al. (1962)
11
observed that urea addition to a low quality diet increased intake and
suggested that this increase was due to a faster rate of digestion.
Elliot and Topps (1963) observed that intake was related more highly
to protein content of the forage than to its digestibility. Elliot
(1967) suggested that nitrogen content below 1.6 percent depressed vol
untary intake of forages. Weston (1971) determined that a level of 2.0
percent N prevented decreases in forage intake. Egan and Moir (1965)
concluded that nitrogen has both a metabolic and ruminal effect. At
low levels of dietary N the animal is in a state of N inbalance, lead
ing to an accumulation of non essential amino acids in the blood. Jones
(1972) proposed that when the level of these non essential amino acids
exceeds the oxidative capacity of the animal, the animal reduces intake
to alleviate this situation. Nitrogen also has a ruminal effect. A
low N diet limits the microbial growth and leads to a reduced ruminal
turnover rate due to decreased dry matter digestion. The net result is
a decrease in voluntary intake.
Familiarity and habit are also factors influencing voluntary for
age consumption. Furthermore, the rumen microbial population should be
allowed to addapt to the new feed. Thus, when measuring intake for
forage evaluation purposes, a preliminary period of 10 to 15 days
should normally be sufficient to avoid errors due to previous experi
ence (Vander Noot et al., 1965; Heaney, 1970).
Liveweight influences the level of intake. Blaxter et al. (1961)
and Crampton et al. (1960) demonstrated that if intake was expressed as
amount consumed per unit of metabolic size (W'j[g ) the effect of body
weight on the measured intake value was effectively removed.
12
Digestion Trials. The most widely used method for determining the
nutritive value of a forage is the animal digestion trial. Although it
is laborious, time consuming and expensive, it represents the base by
which chemical and artifical rumen techniques are evaluated.
Several ruminant species have been used in digestion trials, but
not all researchers agree that data obtained with sheep can be used for
cattle. Heaney (1970) measured intake of bermudagrass silage and two
alfalfa hays concurrently with steers and lambs. In every trial,
steers consumed more forage per unit of metabolic size, than did lambs.
Blaxter et al (1966) reported similar differences in intake between
cattle and sheep. Baumgardt et al. (1964) compared the digestive abil
ities of steers and goats using first growth alfalfa-bromegrass hay.
No significant differences were found between species in the digestibi
lity of dry matter, cellulose and energy. Vander Noot et al. (1965)
compared the digestibility of eight silages by sheep and cattle. When
the data for all silages were pooled, there was no significant differ
ence in extent of digestion between the two species.
Preliminary and collection periods have been fairly well standard
ized, but the tendency has been to shorten the collection period due to
increasing labor and feed costs and a need for faster results. Clanton
(1961) conducted digestion and metabolism trials using four heifers and
four rations to compare 7- and 10-day collection periods. There was no
significant difference in the digestion coefficients, metabolizable
energy or nitrogen retention determined from the two methods, regard
less of the type of ration. White et al. (1974) fed dehydrated Coastal
bermudagrass and rice straw pellets at 0, 20, 40, 60, 80 and 100
13
percent of the ration during metabolism studies. No significant reduc
tion in the standard error resulted When the collection period lasted
beyond four days.
Associative Effect in Digestion of Legumes and Grasses. The as
sociative effect or trade-off of protein/energy or nutrient/protein
have been widely documented for forage:grain mixtures, but rarely for
grass:legume mixtures. Minson and Milford (1967) conducted in vivo di
gestion trials using mixtures of grasses (Digitaria decumbens, 0.7%N)
and legumes (Medicago sativa, 3.6%N, and Trifolium repens. 4.0%N), as
well as the individual grasses and legumes. The three treatments used
were 10, 20 and 30 percent legume, respectively. The supplementary ef
fects of legume in these three experiments were 1.1, 1.8 and 2.6 di
gestibility units. McLeod and Minson (1969) using samples from the
previous hays, conducted in vitro digestibility trials using exactly
the same treatments. The r2 in vivo vs in vitro was 0.998, 0.994 and
0.987 for the 10, 20 and 30 percent treatments, respectively. From the
in vivo experiment, the authors concluded that the supplementary effect
of legumes on the in vivo digestibility of D. decumbens appeared to be
due to the legume overcoming a N deficiency in the grass. The results
from an in vitro trial did not verify this trade-off of protein/energy.
The ruminal fluid was obtained from sheep fed a grass-alfalfa mixture
that supplied sufficient N to overcome any N deficiency. In a second
in vitro trial McLeod and Minson (1969) used ruminal fluid obtained
from a sheep on a low-N diet. The authors noted that this type of
ruminal fluid had very low microbial activity and resulted in
14
incomplete forage digestion at 48 hours with great variability between
samples.
V. IN VITRO FERMENTATION TECHNIQUES
The study of forage nutritive value using in vitro rumen fermenta
tion techniques began somewhat simultaneously in several laboratories.
Clark (I960), at Purdue University, pioneered research describing the
use of ruminal fluid in an "artificial rumen11 technique. Kamstra et
al. (1958) related fairly long term in vitro cellulose digestibility to
stage of maturity and* lignification of forages. Asplund et al. (1958)
measured dry matter loss and volatile fatty acid production in vitro as
indices of forage quality.
In 1963, Tilley and Terry published the widely known two=stage
procedure, with a first stage of rumen fermentation and a second stage
of acid-pepsin solubilization of the residue from the first stage. The
acid pepsin acts to simulate the in vivo breakdown of feed and micro
bial protein by the digestive enzymes of the ruminant abomasum. Van
Soest et al. (1966) proposed the use of the neutral-detergent solution
as a substitute for acid-pepsin in the second stage of the Tilley and
Terry procedure. The neutral-detergent solution solubilizes more total
dry matter than does acid-pepsin because neutral-detergent solubilizes
bacterial cell walls and other endogenous products in addition to pro
tein. Therefore, according to Van Soest, this modification predicts
true digestibility rather than apparent digestibility. Barnes (1969,
1970) and Moore (1970) simplified the two-stage in vitro procedure by
mixing the buffer and ruminal fluid prior to inoculation. The
15
application of CO2 was reduced to 15 sec. prior to closing, the tubes
with rubber stoppers. The direct acidification method eliminated the
centrifugation step following the initial 48 hour fermentation. The
use of a phosphate buffer (Kansas State buffer, appendix table 1)
greatly facilitated the procedure through elimination of the excessive
frothing that occurs with a bicarbonate buffer. The length of incuba
tion in the acid-pepsin stage was reduced in several laboratories from
48 to 24 hours. This modification greatly facilitated scheduling for
routine analysis of large numbers of samples by allowing in vitro runs
to be completed within one normal work week. Mellenberger et al.
(1970) proposed a method which eliminated the filtering step. At the
end of incubation the samples were centrifuged and the supernatant de
canted. The pretared centrifuge tubes were then dried, thus avoiding
the error introduced both by weighing the sample on paper and weighing
the crucible.
Donefer et al. (1960) observed a close relationship between 12
hour in vitro cellulose digestibility and nutritive value index.
Johnson (1969) concluded that this relationship was accomplished by
looking at rate curves for different forages and realizing that the
shape of the sigmoid curve was not necessarily related to final digest
ibility of the forage. Thus, digestion at an earlier time period,
e.g., 12 hours, was more indicative of rates of digestion in the rumen
and in turn to intake. In vitro dry matter disappearance was especial
ly well correlated with DMD or DE (Oh et al., 1966; Mellenberger
et al., 1970). Johnson (1969) and Johnson and Dehority (1968) con
cluded that IVDMD is poorly related to intake. In vitro cellulose
16
digestibilities during shorter incubation periods proved to be quite
highly related to intake (Donefer et al., 1960; Chalupa and Lee, 1966).
Effect of Volume of Inoculum on IVDMD. Experiments by Balch
(1950) and Balch and Johnson (1950) suggested that increasing the pro
portion of water in the contents of the reticulo-rumen increased the
rate of break down of feeds. Harrison et al. (1975) fed a diet of
flaked corn and dried chopped grass to three sheep with duodenal and
rumen cannulas. The relative proportion of feed and microbial protein
entering the duodenum were measured. Then the animals were infused
intra-rurainally with artificial saliva containing polyethylene glycol
(PEG). The authors concluded that the increased dilution rate due to
the buffer infusion in the rumen, caused more ct-linked glucose polymers
and poly-unsaturated fatty acids to escape ruminal digestion. Thomson
et al. (1975) made similar observation, e.g., increased flow of
a-linked glucose polymers and microbial amino acids at the proximal
duodenum and improved efficiency of rumen microbial growth.
Bales et al. (1976) evaluated an "artificial rumen fluid" (ARF,
appendix table 2) for in vitro dry matter disappearance of forages.
The ARF was designed to supply nutrients which might occur in the rumi
nal fluid and were not present in the McDougall's artificial saliva
(appendix table 3). The ARF was evaluated using corn stalks as subs
trate and adding 1 and 10 ml of ruminal fluid inoculum. The IVDMD of
corn stalks was significantly (P<.05) higher (64.9 vs 61.1 percent) for
1 vs 10 ml of ruminal fluid. Neher (1976) used alfalfa hay as a subs
trate and diluted the ruminal fluid 1:1 with water. The percent IVDMD
17
for the diluted ruminal fluid treatment was significantly higher
(P<,001) than the undiluted, 60.7 and 52.2, respectively. Neher (1976)
speculated that the results observed may be explained by the presence
of a toxic factor in the rumen fluid which was diluted beyond its thre
shold value with the dilution rate of 1:1.
Effects of Length of Fermentation Time on IVDMD. The time related
curve for amount of dry matter digested in stage one of an in vitro
rumen fermentation system is sigmoidal with an initial lag phase of up
to 12 hours. The curve then plateaus at 18-24 hours and often becomes
assymptotic at 48 hours vftien legumes are used as the substrate. When
grasses are used, the curve increases continously up to 96 hours (Grant
et al., 1974). Due to the greater concentration of soluble cell con
tents, the initial rate of digestion in legumes is faster than that of
grasses, and the maximum degree of digestion is reached sooner for le
gumes (Marten and Barnes, 1980; Troelsen and Hanel, 1966). Based on
the same finding, Nelson et al. (1975) concluded that a 36 hour fermen
tation had the lowest standard deviation for legumes and some annual
temperate grasses, while 60 hours was the optimum fermentation time for
tropical and perennial grasses. Troelsen and Hanel (1966) studied the
effect of duration of in vitro fermentation using the Tilley and Terry
technique on the digestibility of cellulose and dry matter using
alfalfa hay and wheat straw as substrates. Cellulose in the alfalfa
hay was digested more rapidly than the cellulose in the straw, but
the total amount of digested cellulose in the straw was greater than
in the alfalfa when the fermentation period exceeded 24 hours. The
18
digestibility of the noncellulosic organic matter fraction in the al
falfa substrate at the zero-hour fermentation period revealed that most
of this fraction was digested by the acid-pepsin incubation. In the
wheat straw, only a small percentage of the corresponding fraction was
digested by acid-pepsin. The effects of in vitro fermentation length
on the amount of nutrients digested clearly illustrates the importance
of fermentation time in qualitative assays. Balch and Campling (1962)
revealed that as forages become coarser and more mature, the longer
they could be expected to remain in the rumen exposed to digestion by
the microflora. Therefore, this observation points out the need for
extension of the in vitro rumen digestion procedure to include an esti
mation of the fermentation time required to attain a digestibility
similar to that observed in vivo.
Effect of Inoculum Source on IVDMD. Marten and Barnes (1980) con
sidered the inoculum as the greatest source of uncontrolled variation
in in vitro rumen fermentation systems. Digestive capacity of rumen
inoculum may be influenced by animal species, breeds within species and
individual variation due to previous diet and time of collection.
Hungate et al. (1960) observed a faster rate of fermentation with ru
minal fluid from Zebu cattle, when inoculum taken from African Zebu and
European cattle were compared in in vitro fermentation systems. Grant
et al. (1974) compared ruminal fluid from water buffalo, Philipine
dairy cattle and Holstein (Cornell Univeristy) cattle. No differences
were observed after 72 hours of in vitro digestion among the three
types of cattle using 22 forages, but the buffalo inoculum gave the
19
highest (P<.05) digestibility after 96 hours of fermentation period.
Nelson et al. (1969) studied the effect of source of inoculum between
one Jersey and two Holstein cows using alfalfa, bahiagrass, bermuda
grass and ryegrass as substrates. A significantly (P<.05) higher sub
strate IVDMD was found when digested with inoculum from Jersey compared
to Holstein cows, 71.75 and 70.50 percent, respectively. When urea and
glucose were added to the fermentation media, no significant differ
ences were found in the IVDMD among donor animals. Bezeau (1965) com
pared the source of inoculum between an Ayrshire and a Holstein cow
using three alfalfa hays and mixed grass hay consisting of bromegrass,
red fescue and orchardgrass as substrates. There was no significant
difference in the digestibility of the cellulose, when the inoculum
came from cows fed different hays. The activity of the inoculum from
the Ayrshire cow was significantly (P<.01) higher than the inoculum
from the Holstein cow for all hays studied.
VI. FORAGE EVALUATION USING ENZYMES
Johnson and Dehority (1968) considered the in vitro rumen fermen
tation techniques as unparalleled for predicting relative digestibility
differences among a wide range of forage species and types. This tech
nique of predicting forage digestibility suffers from the disadvantage
that ruminal fluid must be obtained in a fresh state from fistulated
animals and is subject to wide variations in microbial activity which
can only be corrected by standardization.
Forage evaluation using fungal enzymes, as an alternative to rumen
fluid, was first suggested by Donefer et al. (1963) using "cellulase
20
36" (Rohm and Haas, Philadelphia, PA). The treatments used were; cel
lulase, cellulase + pepsin, acid-pepsin, distilled water and 12 hours
in vitro rumen fermentation. Eight temperate legumes and six grasses
were used as substrates. The highest correlation with in vivo was .73
for the acid-pepsin treatment. "Cellulase 36" was considered too weak
an enzyme to obtain adequate fiber degradation. Guggolz et al. (1971)
tested a more active cellulase preparation (Onozuka SS cellulase) ob
tained from fungi Trichoderma viride. The preparation contained cellu
lase, hemicellulase and several sugars as nutrients. The procedure in
volved digestion of the forage with a viride cellulase enzyme for 72
hours followed by a protease digestion of 12 hours. The enzyme solubi
lization of 29 forages was correlated (r*.90) with in vivo dry matter
digestibility. Jones and Hayward (1973) devised a one-stage procedure
based on T. viride preparation (BDH Ltd, Dorset, England) that had cel
lulase, hemicellulase and proteolytic activity. A high correlation
(r=.92, P<,0001) was found between the enzyme digestion and in vivo dry
matter digestion for a wide range of grasses. McQueen and Van Soest
(1975) tested 2 cellulases, Aspergillus niger and Onozuka preparation
of Trichoderma viride, fungal hemicellulases and proteolytic enzymes.
A high correlation (rB.80) was observed between enzyme digestion and in
vivo digestion of 18 grasses and legume hays. Onozuka preparation
showed the greatest deviation in activity from the in vitro fermenta
tion; therefore, T. viride cellulolytic enzymes may be influenced by
chemical and physical characteristics of forage species to a different
extent from those of rumen microorganisms. The authors also con
cluded that the correlation with in vivo was higher for individual
21
species or groups of species.
Recently, two other procedures have been developed with a pre
treatment before the digestion in cellulase in order to make the cell
walls readily available to the fungal enzymes and to improve prediction
equations for both grasses and legumes. Jones and Hayward (1975) in
cluded an acid-pepsin pretreatment before digestion in cellulase and
Roughan and Holland (1977) used a neutral-detergent (NDF) extraction
followed by cellulase enzyme.
VII. CHEMICAL METHODS OF FORAGE EVALUATION
Chemical analyses are the most widely used laboratory techniques
to measure nutritive value of forages. The most common procedures of
chemical analysis are the proximate analysis and the Van Soest tech
nique .
Proximate Analysis System. The Weende System of proximate analy
sis has been the most generally used chemical scheme for describing
feedstuffs. Moisture, crude protein (CP), crude fiber (CF), ether ex
tract (EE), nitrogen-free extract (NFE) and ash in feedstuffs are de
termined and from these values an evaluation of the feedstuffs can be
made, mainly through the total digestible nutrients (TDN) system. The
use of TDN in feeds and feeding is internationally accepted and more
TDN values are available than values in any other system; therefore,
this system will likely continue to be used for many years.
The carbohydrates are divided by the proximate analysis scheme
into CF and NFE. This division is intended to separate the less di
gestible portion (CF) from the more digestible carbohydrates (NFE), but
22
this division is not realistic either chemically or nutritionally for
forages (Van Soest, 1964). The CF residue does not include all hemi
cellulose (He), lignin (L) and acid-insoluble ash (AIA) (Fonnesbeck,
1976; Van Soest, 1973). Nitrogen-free extract which is supposed to
contain soluble carbohydrates, also contains variable amounts of He, L,
cellulose (C) and AIA.
Another drawback of the system is related to nutrient digestibili
ty. In many cases, CF is more digestible than NFE because the latter
contains L, He and some C. Butterworth (1967) reported that the di
gestion coefficients of CF for paragrass, bermudagrass, Guinea grass
and speargrass were 57, 66, 72 and 74 percent, respectively, vrtiereas
the digestion coefficients of NFE for the same grasses were only 51,
59, 67 and 57 percent, respectively. Therefore CF and NFE should not
be considered as nutritive entities.
The chemical estimates of proximate analysis are poorly related to
the in vivo data, and they are poor predictors of forage quality
(Butterworth, 1964; Moore and Mott, 1972).
The Detergent System. In the early sixties Van Soest proposed a
new system of feed partitioning which overcomes some of the short
comings of the TDN and proximate analysis system (Van Soest, 1964).
The system is based on the principle that dry matter of forages can be
divided into a readily available soluble fraction (cell contents, CC)
and a fibrous residue of partial availability (cell wall constituents,
CWC) by use of detergent solutions. In this system the readily availa
ble soluble fraction bears no relationship to lignification (Van Soest,
1964).
23
Feed partitioning and analysis by this system has gradually been
accepted by scientists because it classifies feeds and forages accord
ing to nutritional functions of animals (Goering and Van Soest, 1970).
Althought the detergent system has made significant progress on the se
paration scheme of nutrients, it still has several drawbacks. A large
portion of the total protein remains in the CW fraction. Pectins, tan
nins and most of the silica are extracted into the cell content por
tion. The acid-detergent solution dissolves a considerable amount of
lignin and ADF. Hemicellulose and cellulose are determined by differ
ence, therefore they contain errors from previous determinations.
Other systems of feed partitioning are available. The Fonnesbeck
system (Fonnesbeck, 1976) is a modification of the Van Soest detergent
system. This system results in a purer fiber fraction and also elimi
nates several starch interferences. The Southgate system (Southgate,
1969) was developed for human foodstuffs low in dietary fiber. In this
system the dietary fiber is fractioned into lignin, cellulosic and non-
cellulosic polysaccharides. Where sugar analysis is required, the
Southgate's system is probably the method of choice. This system and
the Van Soest 'a differ in approach but give similar results for many
feeds and foods (McConnel and Eastwood, 1974).
VIII. RELATIONSHIP AMONG CHEMICAL COMPONENTS, LABORATORY VALUES AND IN VIVO DATA
For many years, animal scientists and agronomists have searched
for the "one best component" for predicting forage quality. As there
are large variations in "quality-related" characteristics among forages
24
Table 1. CORRELATION BETWEEN DRY MATTER INTAKE (DMI) AND MEASURES OF FORAGE NUTRITIVE VALUE
Factors correlated r Forage type Reference
DMI and CP .54 gras. + leg. Van Soest (1965b)
DMI and CF -.79 gras. + leg. Wilson et al. (1966)
DMI and CW -.73 gras. + leg. Van Soest (1973)
DMI and CW -.77 gras. + leg. Mertens & Van Soest (1973)
DMI and CW -.65 gras. + leg. Van Soest (1963)
DMI and ADF -.64 gras. + leg. Van Soest (1973)
DMI and C -.75 gras. + leg. Van Soest (1965b)
DMI and L -.10 gras. + leg. Van Soest (1973)
DMI and IVDMD .51 gras. + leg. Barnes (1966)
DMI and IVCD (12hr) .83 gras. + leg. Donefer et al. (1960)
DMI and IVCD (18hr) .72 gras. + leg. Chalupa & Lee (1966)
DMI and DMD .66 trop. grasses Minson (1971)
DMI and DMD .86 gras. + leg. Chenost (1966)
25
and due to differences in species, maturity and environment, no "one
best component" will give the best prediction. Therefore, prediction
of forage quality from laboratory analysis will be successful if the
methods measure or are correlated with one or more quality-related
characteristics (Mott and Moore, 1970). Several forage constituents
have been correlated with intake (Table 1). Blaxter (1950) generalized
that ruminants increase intake with increasing concentration of nutri
ents in the diet. Intake has been correlated with CF, CW, DMD and in
vitro cellulose digestibility (IVCD).
When making comparisons across many species, Van Soest (1965b)
found that NDF (CW) was the component correlated most consistently and
highest with intake. Donefer et al. (1960) proposed that in vitro cel
lulose digestion (IVCD) at 12 hours fermentation be used as a measure
of rate of digestion.
Several forage constituents have been used to predict DMD (Table
2). Butterworth and Diaz (1970) working with a large number of forages
found that the correlation between DMD and various components of the
proximate analysis were CF, -.30; CP, -.47; EE, .37; and NFE, -.09.
Bredon et al. (1963) working with tropical grasses found high correla
tion coefficients between DMD and CP, and CF, of .94 and -.72, respec
tively. Sullivart (1962) published regression equations with CF to pre
dict DMD. Within a given class of forage the determination coeffi
cients were relatively high. The effect of CF was inconsistent between
classes of forages, as the determination coefficient dropped to only
-.49 for all groups. The lower determination coefficient is attributed
to the fact that CF is not a chemical entity. Van Soest (1964) pointed
26
Table 2. CORRELATION BETWEEN IN VIVO DRY MATTER DIGESTIBILITY (DMD) AND MEASURES OF FORAGE NUTRITIVE VALUE
Factors correlated r Forage type Reference
DMD and CP .76 legumes Oh et al. (1966)
DMD and CP .21 grasses Oh et al. (1966)
DMD and CF -.65 grasses Sullivan (1964)
DMD and CF -.54 legumes Sullivan (1964)
DMD and CF -.49 gras. + leg. Sullivan (1964)
DMD and ADF -.85 gras. + leg. Wurster et al. (1971)
DMD and ADF -.78 gras. + leg. Van Soest (1964)
DMD and C -.62 gras. + leg. Wurster et al. (1971)
DMD and C -.81 gras. + leg. Clancy and Wilson (1966)
DMD and IVDMD .83 grasses Oh et al. (1966)
DMD and IVCD .75 gras. + leg. Oh et al. (1966)
27
out that lignin, in theory, is an ideal chemical component for predic
ting digestibility, but that the chemistry of lignin and the methods of
determining lignin were not well developed. Van Soest (1965b) found
that DMI was highly correlated with CW (r=-.79) for six grass species
and alfalfa.
Van Soest and Moore (1965) proposed an index of availability based
on the ratio of the digestible portions of plant material. The index
of availability of the DM was expressed as 100-100 (percent CWC/percent
CC). Further refinement of the index of availability by Van Soest
(1967) resulted in the summative equation for predicting DMD. An aver
age digestibility for CC was estimated at .98 percent. The CWC digest
ibility was estimated with a logarithmic transformation based on lignin
as a percent of ADF. The endogenous excretion average was estimated at
12.9g per lOOg DM consumed. The complete summative equation was: per
cent apparent DMD ■ 0.98 (CC) + CWC J l . 473-0,789 log (100-percent
lignin/percent ADF)] -12.9. Duble et al. (1971) concluded that the Van
Soest summative equation overestimated the DMD of warm season perennial
grasses. Deinun et al. (1968) observed a relatively high residual
standard error of 4.06 for predicting DMD of several forages from the
summative equation.
The two-stage IVDMD has proven to be an excellent method to pre
dict DMD because it gives high predictability for DMD with very low
standard errors. Hershberger et al. (1959) reported a high correlation
(rs0.97) between in vitro and in vivo cellulose digestibility; however,
the in vitro digestion was lower than that observed in vivo. Tilley
and Terry (1963) found a high correlation (r“0.93) between digestibili
ties measured in vivo and by the two-stage in vitro technique. Oh et
28
al. (1966) found that the two-stage in vitro technique was the most re
liable predictor of forage DMD of all laboratory methods studied.
IX. PHYSICAL METHODS OF FORAGE EVALUATION
Physical methods include, infrared reflectance spectroscopy, arti
ficial mdstication of forage tissue (Troelsen et al., 1970) and the
measurement of a fibrousness index (Chenost, 1966) based on the elec
trical energy required to grind a sample. Although histological tech
niques are difficult to quantify, Barnes and Marten (1979) considered
them as promising physical procedures for the study of the morphologi
cal and structural configuration of plant tissues. Monson et al.
(1972) indicated that cutin and the waxy epidermis of certain plant
leaves can act as an effective barrier to microbial attack of leaves.
Akin and Burdick (1973) demonstrated the importance of certain morpho
logical characteristics of grass leaves in determining the digestibili
ty of individual chemical components.
The physical methods illustrate the limitations of chemical meth
ods in the prediction of quality from heterogenous and complex forage
plant tissues.
Application of Near-Infrared Reflectance Spectroscopy to Forage
Evaluation. Near-infrared reflectance spectroscopy (NIR) has already
proved itself for analysis of dry matter, protein and lipids in milk
and soybeans (Norris and Hart, 1965). This technique has only recently
been applied to forages.
Near-infrared reflectance spectroscopy is based on the fact that
29
the near infrared reflectance spectra differ for the various feed com
ponents due to chemical groups and linkages inherent to their mole
cules. Radiant energy in the near infrared portion of the spectrum
(1,100 to 2,500 nm) is applied to dry, ground samples of forage and the
detection of energy reflected from the samples depends on its
chemical/physical properties. The sample, packed into a cell and cov
ered with a quartz window, is illuminated through the' window and dif
fusely reflected radiation is collected with four lead sulfide cells.
The signal from the lead-sulfide detector is amplified with a
logarithmic-response, digitized, fed to a computer as log (1/R) and
then transformed to a second derivative of log (1/R) for correlation
with compositional data (Norris and Barnes, 1976). Multiple stepwise
regression techniques are used to determine the optimum wavelengths for
predicting each of the chemical constituents and nutritive values. The
data contained in the 2,000-point spectral curves are smoothed by a
technique which averages adjacent points and compresses the curves to
500 points. Using an iterative procedure, all wavelengths from 1,400
to 2,400 nm are tested up to nine steps for each component (Norris and
Barnes, 1976).
Plotting log (1/R) as a function of wavelength, gives a curve that
is comparable with an absorption curve having peak readings at wave
lengths that correspond to absorption bands in the sample. The first
two wavelengths used in the regression analysis of each component are
the most important. The locations of these wavelengths are summarized
in Table 3 and superimposed on typical reflectance (log 1/R) in figure
1. Most of the wavelengths used for CP and NDF are in the longer
Table 3. TYPICAL NIR WAVELENGTHS USED IN REGRESSION ANALYSIS FOR EACH FORAGE COMPONENT
Component Wavelengths Cum)
DMI 1.98, 1.81, 1.69, 2.08
DMD 1.98, 1.69, 2.20
IVDMD 2.26, 1.90, 1.60
CP 2.16, 2.11, 2.25
CHO 2.25-2.40
CF 1.61, 2.19, 1.53
NDF 2.34, 2.09, 1.70
ADF 1.66, 1.40 1.71
L 1.55, 1.64, 1.42
DMI = Dry Matter Intake DMD ■ Dry Matter Digestibility IVDMD " In Vitro Dry Matter Disappearance CP * Crude Protein CHO “ Carbohydrates -CF “ Crude Fiber NDF * Neutral Detergent Fiber ADF ■ Acid Detergent Fiber L “ Lignin
• ■ • i i i i i ” i r — i " i " 1 i
h CHO -|DMIDMDADF CP IVDMD NDF
1100 1300 1500 1700 1900 2100 2300NMFigure 1. MOST COMMON LOCATIONS OF WAVELENGTHS ABSORBED BY FORAGE COMPONENTS
SUPERIMPOSED ON ALFALFA HAY REFLECTANCE LOG (1/R) SPECTRA
32
wavelength region while those for ADF and L are at the shorter wave
lengths. The best wavelength for predicting DMD is the same as for
ADF, and the best wavelength for predicting DMI is essentially the same
as for CP and NDF. Also the best wavelength for predicting IVDMD is
the same as for NDF. The region from 2.25 to 2.40 ym usually shows a
combined spectra of protein and carbohydrates.
Monochrometer-type instruments can scan the entire NIR region and
give a linear response throughout this region. Because the optical en
ergy output of monochrometers is relatively low, their usefulness for
generating quantitative data in the NIR may be limited for certain for
age constituents (Burdick et al., 1981).
Near-infrared spectrophotometers using multiple filters instead of
a monochrome ter have also been used in forage analysis (Akin and
Burdick, 1979). Barnes and Marten (1979) concluded that unless the
instrument contains a scanning monochrometer, the key to its success
for estimating the chemical composition in forages depends upon whether
the proper wavelength filters are selected. It is apparent that quali
ty of all forages cannot be measured by one unique set of filters;
therefore, it is incorrect to assume that an instrument with a given
set of filters will be accurate for predicting all quality parameters
for all forages.
Chemometric techiques such as NIR, require calibration with a set
of samples. The calibration samples should include all the variability
in composition that might be encountered in any sample to be measured
because the technique relies on multiple reflectance measurements to
correct for interference from the different components in the sample.
33
There are several factors that affect the accuracy of the NIR tech
nique. Particle size, packing density within the sample cell, tempera
ture and homogeneity of the sample, and uniformity of the surface to be
measured are the most important factors according to Norris and Barnes
(1976). All these factors are under operator control in sample prepa
ration.
Shenk et al, (1977) considered three different error expressions
important in NIR. These are standard error of calibration (S.E.C.),
standard error of duplication (S.E.D.) and the standard error of pre
diction (S.E.P.). The standard error of calibration is the standard
error of the fit to the regression equation. The standard error of
duplication is the variability in repetitive measurements of the same
sample, and the standard error of prediction is the variability between
NIR prediction of the composition of an unknown sample and the chemical
composition as determined by chemical analysis. If the chemical analy
ses are assumed to be correct, than the prediction error states the ac
curacy of the NIR.
Norris et al. (1976) recorded near-infrared reflectance spectra
for 87 samples of ground dry forages including alfalfa, tall fescue,
bermudagrass and pangola digitgrass. Crude protein, ADF, NDF, L,
IVDMD, DMD, DMI and DEI were used to calibrate the instrument. The de
termination coefficients were .98 for CP, .96 for NDF, .92 for ADF, .92
for L, .90 for IVDMD, .77 for DMD, .64 for DMI and .72 for DEI. The
standard error of prediction obtained was ±.95% for CP, ±3.1% for NDF,
±5.1% for DMD and ±7.9g for DMI. Barton and Coleman (1981) used 30
samples of forages assembled from three locations, Pennsylvania,
34
Georgia and Oklahoma. The samples included alfalfa, orchardgrass,
timothy, tall fescue and bermudagrass. Samples were analized for CP,
ADF, ADL and IVDMD. The determination coefficients were .92 for IVDMD,
.98 for CP, .88 for ADF and ,87 for ADL.
MATERIALS AND METHODS
I. HAYS AND ANIMALS
The experimental hays used in the study were U.S. number 1 grade
alfalfa hay (IFN 1-00-073)? mature ryegrass (IFN 1-04-068), common
bermudagrass (IFN 1-00-073), Alicia bermudagrass and Pensacola bahia-
grass (IFN 1-00-462). Origin of forage, date of cutting and stage of
maturity are summarized in appendix table 5. All hays, except alfalfa,
were chopped (4 to 6 cm) and rebailed in rectangular bales to facili
tate handling.
Ruminal fluid donor animals for the entire study included a
Holstein steer, a Hereford steer and two Holstein cows. Only one donor
animal was used for a specific experiment except when effect of ruminal
fluid donor was being studied as a variable in the experiment.
In vivo forage intake and digestion coefficient values were ob
tained with six cross-bred steers selected for uniformity in age and
type. The initial weight ranged from 280 to 330Kg.
II. VOLUNTARY INTAKE AND DIGESTION TRIALS
The hay treatments used in the intake and digestion trials were
mature ryegrass, alfalfa, common bermudagrass, Alicia bermudagrass,
Pensacola bahiagrass and a 50:50 mixture of alfalfa and ryegrass.
The steers were wormed two weeks before initiating the trials.
Each steer was sprayed for flies and weighed immediately before each
international feed nomenclature, International Feed Institute, Colorado State University
35
36
trial and kept in individual stalls throughout each intake and diges
tion trial. The experimental design was a 6x6 Latin Square with three
weekB separating each period so that animals could adjust to the treat
ment change. During the digestion trials the steers were watered and
fed at 6:30 a.m. and 4:00 p.m. each day. Trace mineralized salt was
available at all times. Each hay treatment was fed ad libitum on days
1-10 inclusive at a level to supply approximately 110 percent of the
previous day's consumption. Ad libitum intake was measured on days
3-10 inclusive with weighbacks taken once daily. On day 9 an amount of
feed equivalent to 90 percent of the average daily ad libitum intake
established on days 3-8 was weighed to the nearest gram. Each steer
was fed this amount daily for days 10-19 inclusive. At the time of
weighing, samples of hay fed to each steer were collected for chemical
analysis. The samples of hay were composited for each steer and period
and dried at 60 G for 48 hr in a forced-air oven. In each trial, feces
excreted by each steer on days 15 through 19 were related to forage
consumed on days 13 through 17. The feces and urine were collected
once daily and weighed to the nearest gram. A representative 5 percent
fecal sample was dried at 60 C for 48 hr in a forced-air oven. The
feces were then composited for each steer and period. After drying,
the diet and fecal samples were ground through a Wiley-Mill with a 1 mm
screen and stored in air-tight containers.
III. CHEMICAL ANALYSIS
The forage and fecal samples were analyzed for proximate compo
nents as outlined by AOAC (1980). Neutral detergent fiber and ADF were
37
determined by the method of Goering and Van Soest (1970). Hemicellu-
lose was calculated as the difference between NDF and ADF, and cellu
lose was calculated as the weight loss upon treating the ADF residue
with 72 percent ^SO^.
An automatic Farr adiabatic calorimeter was used for determining
the gross energy of feed and feces.
IV. IN VITRO FERMENTATION STUDIES
Many factors cause variation in in vitro fermentation results.
Such things as ruminal fluid fill of donor animal and length of fermen
tation time are some of these factors. Therefore, a series of experi
ments were designed to investigate these effects.
The hay samples were analyzed for IVDMD as outlined by Moore
(1970) with direct acidification. Hay samples from each hay treatment
and period were analyzed in triplicate.
An experiment was designed to study the effect of in vitro proce
dure on IVDMD. Composite samples of each hay treatment and period were
analyzed in triplicate using five different techniques. The techniques
used were the Moore technique (Moore, 1970) with bicarbonate buffer,
the Moore technique with phosphate buffer, the Van Soest modification
of Tilley and Terry (Goering and Van Soest, 1970), the Troelsen tech
nique (Troelsen, 1970) and a one-stage technique described by Baumgardt
et al. (1962) and modified by Mellenberger et al. (1970). A completely
randomized design was used to statistically analyze the results.
Previous research by Neher (1976) had demonstrated effects of ru
minal fluid dilution on IVDMD. Therefore, an experiment was designed
38
to study the effect of ruminal fluid dilution on IVDMD using alfalfa
hay as the substrate with the Moore in vitro fermentation technique.
The following amounts of ruminal fluid: distilled water: buffer were
used; 10:0:40; 8:2:40; 6:4:40; 4:6:40 and 2:8:40. The several ruminal
fluid treatments were mixed with buffer for 15 minutes and then added
to the centrifuge tubes containing the substrate. Thirty-nine repli
cates per treatment were used during each of two fermentation runs con
ducted two weeks apart. Two blanks were used for each treatment. A
randomized block design was used to statistically analyze the results.
A follow-up experiment was designed to study the effect of further
dilution of ruminal fluid on IVDMD, using the same procedure. The fol
lowing amounts of ruminal fluid: distilled water: buffer were used:
2.5:7.5:40; 2:8:40; 1.5:8.5:40; 1.0:9.0:40; 0.5:9.5:40 and 0:10:40.
Thirty-five replicates per treatment were used during each of two fer
mentation runs conducted two weeks appart. Two blanks were used for
each treatment and a randomized block design was used to statiscally
analyze the results.
A third in vitro experiment was conducted to determine if any in
teraction existed involving dilution, length of fermentation time and
forages. Ruminal fluid treatments were 2:8:40; 5:5:40; 10:0:40;
17:0:33 and 25:0:25 (ruminal fluid:distilled water:buffer). Fermen
tation lengths were 24, 48 and 96 hours. The substrates were samples
of each hay treatment composited across periods. All determinations
were made using triplicate samples. The experiment was designed as a
split plot with a 6x5 (hay, dilution) factorial as the main plot and
39
time as the subplot and two in vitro runs scheduled two weeks apart.
A fourth in vitro experiment was conducted to study the effect of
inoculum donor, dilution and concentration of ruminal fluid and fermen
tation length on IVDMD of alfalfa hay. Four cannulated animals were
used as inoculum donors. Animals were withdrawn from water six hours
prior to ruminal fluid collection. The ruminal fluid treatments were
2:8:40; 5:5:40; 10:0:40; 17:0:33 and 25:0:25 (ruminal fluid:distilled
water:buffer). Fermentation lengths were 24, 48 and 96 hours. This
experiment was designed as a split-splitplot with animals as the main
plot, dilution as subplot and time as the sub-subplot. Two in vitro
runs scheduled at two week intervals with triplicate samples for each
arrangement of treatments were conducted in this experiment.
V. ENZYME DIGESTION STUDIES
Fungal enzymes, source of organism and commercial supplier were:
cellulase (Aspergillus niger) type IV, No. C-7377, Sigma, St. Louis,
MO; cellulase (Trichoderma viride) type IV, No. C-4137, Sigma; cellu
lase Onozuka preparation (Trichoderma viride), Kanematsu-Gosho (U.S.A.)
Ltd, CA; hemicellulase (Aspergillus niger) No. H-2125, Sigma; and pep
sin (1:10,000) from Nutritional Biochemical Corp., Cleveland, OH.
Two enzyme techniques with different pre-treatments were used.
One was the Jones and Hayward (1975) technique adapted by Goto and
Minson (1977). Two hundred mg of air-dry samples were placed in 100-ml
centrifuge tubes followed by addition of 5 ml distilled water to mois
ten the samples. Substrate samples were incubated for 48 hr at 39 C
with 30 ml of 0.1 N HC1 containing 0.2 percent (w/v) pepsin to remove
40
the cell contents. The tubes were swirled three times a day. After
incubation the samples were filtered through Whatman paper No. 54 and
thoroughly rinsed with distilled water and sodium acetate:acetic acid
buffer solution (appendix table 4). The samples were transferred back
to the centrifuge tubes with the final sodium acetate:acetic acid buf
fer solution containing 2.5 percent (w/v) cellulase preparation made up
to 30 ml and incubated for 48 hours. After the incubation the contents
were filtered through a tared, coarse porosity, fritted glass crucible
and dried to a constant weight. The residue retained in the filter was
undigested dry matter.
The second technique used was the same as the one described above,
except that the cell contents were removed with NDF instead of pepsin
before the cellulase incubation (Roughan and Holand, 1977).
This experiment was designed to study the effect of enzyme diges
tion technique on IVDMD. Substrates were composite samples of each hay
treatment and period, and were run in triplicate. Enzyme treatments
for both techniques were cellulase Onozuka preparation, cellulase
T. viride type IV + hemicellulase A. niger and cellulase A. niger
type IV + hemicellulase A. niger. Six treatments combining the two
pre-treatments and the three enzyme preparations were compared. A
completely randomized design was used to statiscally analyze the
results.
VI. NEAR-INFRARED REFLECTANCE SPECTROSCOPY ANALYSIS
The instrument system for near-infrared reflectance spectroscopy
was comprised of two major instruments, the spectrometer and the com
puter. The spectrometer was a Neotec model 6100 monochromator. It
41
produced monochromatic light in the near infrared region from 1,100 to
2,500 nm with a holographically ruled grating. Wavelength detection
was accomplished by lead-sulfide detectors. Sixty-four scans of each
Bample were taken and averaged. Seven hundred data points at 2.0 nm
intervals were recorded.
The computer was a Digital Equipment Corporation mini-computer PDP
111-03, 32K. It was comprised of RL01 hard 5.2 M byte disk, dual 512K
byte floppy disks, Decwritter III and hard copy terminals. The instru
ment was located at the Field Crops Marketing and Utilization Research
Laboratory, Richard B. Russell Agr. Res. Ctr., USDA, SEA-AR, Athens,
GA.
All hay and fecal samples from each hay treatment and digestion
trial were reground through a Tecator/UDY Cyclone Mill with .25 mm
screen and uniformly packed into a Neotec sample cup. Twenty-nine sam
ples were used to calibrate the instrument, and five were used for pre
diction. A blank scan was executed just before each sample.
The operating software was developed at Pennsylvania State
University and created files of the data, added calibration data, per
formed mathematical transformations and stepwise linear multiple re
gression analysis, made statistical comparisons of actual vs predicted
data, and measured instrument noise and wavelength accuracy (Shenk et
al., 1977).
VII. STATISTICAL ANALYSIS
Differences between treatments were tested by Scheffe's test,
Duncan's multiple range test (Steel and Torrie, 1980) or contrast
comparisons (Cochran and Cox, 1950; SAS, 1982). Duncan's MRT
42
controls the type I comparisonwi.se error rate, while the Scheffe's T.
controls the type 1 experimentwise error rate. The F statistic for
pairwise comparisons using Scheffe's T. is more independent than in
Duncan's MRT. Moreover, Scheffe's T. is more conservative (less sensi
tive) than most multiple comparisons procedures for detecting signifi
cant differences among pairs of means. Duncan's MRT was used in ini
tial experiments. When the 1982 SAS package was released and the
Scheffe's T. became available, the multiple comparisons in the remain
ing experiments were tested with the Scheffe's T. Single contrast com
parisons are an essential tool when the values compared are averages of
treatment means, i.e., . 5x^ + .5x2 vs x3* Simple and multiple linear
regression, stepwise linear regression with stepwise and maximum R2 im
provement methods (Neter and Wasserman, 1974; SAS, 1979) were performed
among in vivo data, in vitro fermentation values, enzyme digestion
values and chemical components. Coefficients of determination obtained
with simple and multiple linear regression analyses, depend on number
of observations used, range of the observations and degrees of freedom
left for the error term. Therefore, these factors should be taken into
consideration when comparing coefficients of determination. Some pre
diction equations for in vivo parameters from selected laboratory
determinations were also developed. An IBM model 3033, a DEC PDP
111-03 computers and the SAS package (SAS, 1979, 1982) were used for
all statistical analyses.
RESULTS AND DISCUSSION
I. CHEMICAL COMPOSITION OF THE HAYS STUDIED AND RESULTS OF THE INTAKE AND DIGESTION TRIALS
The chemical compositions of the hays studied are shown in Table
4. The digestibilities of chemical components are summarized in Table
5. Alfalfa and ryegrass had the highest and lowest IVDMD (62.8, 43,6)
and CP (18.3, 8.8) values respectively, of the hays studied. The ma
ture ryegrass presents a characteristic chemical composition of mature
grasses with low protein and high fiber content. The chemical composi
tion of alfalfa is characteristic of a legume with high protein con
tent, low fiber, and especially low hemicellulose content. The chemi
cal composition of the warm season grasses shows the effect of longer
regrowth period than the optimum, e.g., 4 weeks. Fiber content was
higher and digestibility of chemical components was lower than the
values observed by Montgomery et al. (1979) at 4 weeks regrowth. These
data probably reflect the longer regrowth period of the forage in this
study.
Holt and Conrad (1981) working with bermudagrass hybrids, observed
a decline in in vitro dry matter digestibility from 80 to 65 percent
from April to September for 2-4 weeks regrowth, and from 65 to 55 per
cent for 4-8 weeks regrowth. Montgomery et al. (1979) concluded that
IVDMD of 4-week old common and Alicia bermudagrass was not signifi
cantly affected by harvest date. Seasonal average IVDMD of common and
Alicia bermudagrasses with 4 weeks regrowth were 56.2 and 53.2 percent,
respectively. Pensacola bahiagrass had a significant decline in IVDMD
from 62 to 55 percent from May to August for 4 weeks regrowth.
43
Table 4. CHEMICAL COMPOSITION AND IVDMD OF HAYS AVERAGED FOR ALL PERIODS (%)
IVDMD CP CF NDF ADF C He L AshHay Mean sd Mean sd Mean sd Mean sd Mean sd Mean sd Mean sd Mean sd Mean sd
Ryegrass 43.6°1.13 8.80 .29 37.6a .30 75.6b .38 50.5a .64 36.53 .73 25.0C .77 10.I3 .72 10.9a .22
Alfalfa 62.8ai.44 18.3a .66 be33.3 2.24 48.5d .71
cd38.4 1.71 30.3° 1.56 10.281.67
ab9.2 .79
be9.5 .32
Ryeg+alfb
54.4 1.85be13.7 .45
ab35.5 1.10 62.0° .48 b
44.5 1.11b33.4 1.06
d17.5 .99 ab9.7 .67 ab10.2 .06
CommonBG 55.9b1.52 14.3b .54 31.5°1.29 80.3ai.36 39.7C .34 b32.9 1.79 40.8ai.33 d6.9 .39 6.3d .43
AliciaBG b54.8 3.99 d10.3 .82 d26.5 .94 b73.8 2.22 d37.7 1.52 28.0°1.48 b36.1 1.69 be8.6 .80 9.3C .58
P bahia b52.7 2.60 c12.4 .62be ■
34.0 1.18a
80.0 1.52b
45.1 .98 36. b31.87b
34.8 .64cd
8.0 1.20d
6.0 .36
SE2 0.72 0.21 0.47 0.44 0.31 0.40 0.43 0.18 0.15
3 b Cl'Q 6 # • • •* * Values in Che same column followed by different superscripts are significantlydifferent at (P<.01)f Scheffe's T.Standard deviation by hayStandard error
Table 5. DIGESTIBILITY OF CHEMICAL COMPONENTS OF HAYS AVERAGED FOR ALLPERIODS (%)
CP NDF ADF Ci HeHay Mean sd1 Mean sd Mean sd Mean sd Mean sd
Ryegrass 45.9° 3.49 be55.3 4.96 abc 54.7 3.94 ab65.2 7.39ab
60.5 5.83
Alfalfa a73.5 2.16 be54.1 2.09ab57.2 1.37
ab69.4 3.15
c41.8 12.50
50:50ryeg+alfb
61.4 2.31c
54.7 2.05be
56.1 2.89b
67.4 3.55be
51.3 4.86
Common BGb
60.3 2.45a
66.4 1.30ab
61.2 2.32ab
73.6 4.01a
70.9 2.87
Alicia BGc
47.8 3.93ab
61.3 4.68C'
46.1 8.41ab
68.8 7.32a
70.6 5.61
Bdiiagrass 46.9° 6.47 a64.6 5.10 a63.6 4.37a
76.9 5.32ab
66.1 6.57
SE2 1.58 1.27 1.77 2.09 2.59
a* ,c Values in the same column followed by different superscripts are significantly different at (P<.01), Scheffe's T.Standard deviation by hay Standard error
46
The in vivo results are summarized in Table 6, and appendix Table
6, 7, 8, and closely follow the in vitro and chemical composition
data.
Overall correlation coefficients among various parameters of for
age quality are presented in Table 7, There were significantly high
positive correlations among DMI, DE, DMD, IVDMD and CP. High levels of
these entities are generally accepted as an indication of good quality
in forages. Therefore, these entities were negatively correlated with
all cell wall components. There was a highly significant relationship
between CF and ADF, since both contain C and L. The significant nega
tive relationship (^-.84, P<0,05) between ADF and DMD is in agreement
with other studies by Oh et al. (1966), Van Soest (1965b) and Wurster
et al. (1971) who considered ADF as the single best predictor of DMD,
due to the fact that ADF is made up of the least digestible components
in the forage. The significant negative relationship (rm-.76, P<.05)
between NDF and DMI is in agreement with other studies by Van Soest
(1965a), Rohweder et al. (1976) who considered NDF as the single best
predictor of DMI. Neutral-detergent fiber, the entity representing the
cell wall constituents, plays a major role in digestibility and rate of
passage and therefore directly affects DMI.
Among chemical components, ADF tended to be the most important
single constituent (r=-.76, P<.05) controlling energy digestibility,
presumably because C and L have a lower energy content than the cell
contents, and lignin renders several cell constituents unavailable.
This finding is in agreement with the results reported by Johnson and
Dehority (1968), where a high correlation between ADF and DE
47
Table 6. VOLUNTARY INTAKE (DMI), APPARENT DIGESTIBILITY (DMD) AND DIGESTIBLE ENERGY (DE) OF HAYS AVERAGED FOR ALL PERIODS
DMI(g/W*Vrt) DMD(%) DE(Kcal/g)Hay Mean sir1 Mean sd Mean sd
Ryegrass 42.9C 4.73 49.2d 4.16 2.03d .17
Alfalfa 88. 7S 6.58 61.5S 1.44 2.64® ino
50:50ryeg+alf b64.8 5.93 52.9C 1.80 2.25° .09
Common BG 67.2 8.68 56.9b 1.22 2.49Sb .07
Alicia BG 49.0C 7.03 be54.9 2.95 2.23° *13
Bahiagrass b59.3 8.78 ,bc53.4 4.66 be2.29 .19
SE2 2.84 1.14 0.05
8 b c d’ * ’ Values in the same column followed by different superscripts are significantly different at (P<.01), Scheffe's T.1 Standard deviation by hay2 Standard error
Table 7. OVERALL CORRELATION COEFFICIENTS (r) AMONG VARIOUS PARAMETERS OF FORAGE QUALITYIN THE HAYS
DMI DE DMD IVDMD CP CF NDF ADF C He L
DMI 1.00 ** * .93 .88 _ * .88 **.99 .00 *-.76 -.54 -.30 -.53 -.17
DE 1.00 .96****.92 **.94 -.27 -.51
*-.77 -.44 -.22 -.46
DMD 1.00**
.95*
.88 -.42 -.59*
-.84 -.63 -.26 -.38
IVDMD 1.00*
.89 -.45 -.63*
-.85 -.65 -.29 -.34
CP 1.00 -.02 -.70 -.56 -.31 -.49 -.21
CF 1.00 -.13 *.83 .81* -.48 .52
NDF 1.00 .31 .43**
.91 -.50
ADF 1.00*
.88 -.10 .50
C 1.00 .08 .11
He 1.00 -.74
L 1.00
*Significant at P<.05Significant at P<.01
49
(r™-.76, P<.01) was found.
Correlation coefficients among digestibility of various parameters
of forage quality and voluntary intake are summarized in Table 8.
Crude protein digestibility was significantly and positively correlated
with DMI, DMD and IVDMD. The important role of CP in microbial activi
ty and amino acid balance and metabolism makes CP an important factor
both in feed intake and digestibility. Neutral detergent fiber diges
tibility was significantly and positively correlated with C and He di
gestibilities, because C and He are NDF components. Acid-detergent
fiber digestibility was significantly (F<,05) correlated with C diges
tibility, due to the fact that a large proportion of ADF is made up by
C. Prediction of forage nutritive value using single and/or multiple
factors is presented and discussed in a later section.
II. IN VITRO RUMINAL FERMENTATION RESULTS
The effect of ruminal fluid dilution on IVDMD of alfalfa hay is
presented in Table 9, and appendix Table 9. The IVDMD for the 4:6:80
and 2:8:40 (ruminal fluid:water:buffer) dilution was significantly
higher (P<.05) than the mean values for the 10:0:40 and 8:2:40 dilu
tion. The regression of percent IVDMD on percent added water to rumi
nal fluid is presented in Figure 2. The regression equation of IVDMD
on percent added water (X) to the ruminal fluid is y*75.047 + .029X,
with r2*.85. There was a significant increase in IVDMD with increasing
dilution of ruminal fluid. Similar results were obtained vften using
six different hays (Table 10) and four inoculum donors (appendix
Table 8. OVERALL CORRELATION COEFFICIENTS (r) AMONG DIGESTIBILITY OF VARIOUS_________PARAMETERS OF FORAGE QUALITY OF THE HAYS AND DRY MATTER INTAKE
DMI DMD IVDMD CPdig NDFdig ADFdig Cdig Hcdig
DMI 1.00 • 00 °°*
00 °°* *.94 -.18 .34 .17 -.66
DMD 1.00**
.95*
.81 .09 .18 .31 -.39
IVDMD 1.00*
.81 .00 .07 .25 -.41
CPdig1 1.00 -.38 .11 -.13 -.73
NDFdig2 1.00 .49*
.86*
.82
ADFdig3 1.00*
.84 .05
Cdig4 1.00 .46
Hcdig-* 1.00
Significant at P<.05 Significant at P<.01 “CPdig = crude protein digestibility TlDFdig = neutral detergent fiber digestibility 3ADFdig = acid detergent fiber digestibility Cdig = cellulose digestibility ^Hcdig = hemicellulose digestibility
Table 9. EFFECT OF RUMINAL FLUID DILUTION ON IVDMD OFALFALFA HAY
Ruminal fluid: waterrbuffer
Number of samples
Mean IVDMD (%)
10:0:40 39 74.72a
8:2:40 39 75.99ab
6:4:40 39 be76.43
4:6:40 39 76.87°
2:8:40 39 77.23da b e d* * * Means with different superscripts are different (P<.05), Duncan's MRT
% IVDMD
52
75.047 + ,029x
0 6040
% Added H20
Figure 2. REGRESSION OF IN VITRO DRY HATTER DISAPPEARANCE OF ALFALFA HAY ON PERCENT ADDED WATER TO RUMINAL FLUID
Table 10. EFFECT OF RUMINAL FLUID DILUTION ONAVERAGE IVDMD FOR ALL HAYS
Ruminal fluid: water:buffer
Number of samples
Mean IVDMD (%)
a2:8:40 36 51.34b5:5:40 36 52.91a10:0:40 36 51.06b17:0:33 36 54.09a25:0:25 36 50.16
a,bMeans with different superscripts are different (P<.05), Duncan's MRT
54
Table 16). These results are contraditory to those obtained by McLeod
and Minson (1969) for several tropical grasses, using 25, 15, 10, 5 and
2.5 ml of ruminal fluid made up to 50 ml with buffer. The two studies
were somewhat different in that the dilution that McLeod and Minson
used, was obtained by changing only the ruminal fluid/buffer without
any addition of water. The increase in 1VDMD with increased dilution
may be explained both by an increase in microbial activity and reduc
tion in the concentration of end-products of fermentation, toxic fac
tors and levels of volatile fatty acids produced at any given time
(Johnson et al., 1958; Warner, 1956).
The effect of further dilution of ruminal fluid on IVDMD of alfal
fa hay is presented in Table 11, and appendix Table 10. The regression
is presented in Figure 3. The data from the two experiments could not
be combined because they were obtained in different in vitro runs.
Ruminal fluid dilutions of 1.5:8.5:40, 0.5:9.5:40 and 0:10:40 resulted
in significantly (P<.05) lower IVDMD values than the mean values for
the 2.5:7.5:40 and 2.0:8.0:40 dilution. The regression equation of
percent IVDMD on water(X) added to ruminal fluid in this dilution range
is Y=-82.058 + 3.757X - 0.023X2 , with R2“.40. The regression shows a
decrease in IVDMD with further dilution of ruminal fluid.
The effects of dilution rate of ruminal fluid and fermentation
length on IVDMD for several hays are summarized in Table 12 and appen
dix Tables 11 and 12. In vitro dry matter disappearance for all hays
increased as the length of incubation time increased (Table 12 and fig
ures 4, 5 and 6). Interaction of forage with time was not significant
Table 11. EFFECT OF FURTHER RUMINAL FLUID DILUTION ONIVDMD OF ALFALFA HAY
Ruminal fluid: waterrbuffer
Number of samples
Mean IVDMD (%)
2.5:7.5:40 35 73.29**
2.0:8.0:40 35 73.513
1.5:8.5:40 35 72.55b
1.0:9.0:40 35 73.90a
0.5:9.5:40 35 72.52b
0:10:40 35 66.15Ca,b,cMeans with different superscripts are different (P<.05), Duncan’s MRT
X IVDMD
56
72 -
70 -
68 -
66 -
Figure 3
-82.058 + 3.757x - 0.023x
y I------------ 1______________ I_____________ l—70 80 90 ioT
% Added H20
REGRESSION OF IN VITRO DRY MATTER DISAPPEARANCE OF ALFALFA HAY ON INCREASED PERCENT ADDED WATER TO RUMINAL FLUID
Table 12. EFFECT OF LENGTH OF FERMENTATION TIMEON AVERAGE IVDMD FOR ALL HAYS
Treatment Number Meanfermentation of IVDMDlength (hr) samples (%)
a24 60 43.33b48 60 53.56
96 60 58.86a,b,c
Means with different superscripts are different(P<.05), Duncan's MRT
72
68
64
60
56
52
48
44
40
36
32
28
a
alfalfa (r.84)Alicia BG (r
.91)ryegrass (r72)common BG (r'
bahiagrass (r2 ■ .82)
49 - i24 48
Fermentation length (hr)
PREDICTED EFFECT OF 2:8:40 RUMINAL FLUID DILUTION ONIVDMD OF SEVERAL HAYS
64
60
56
52
48
44
40 ’
36 ■
32 ■
28 ■J*
59
alfalfa (r 46)
Alicia BG (r .96)
. c « ryegrass (r2 » .82)
d « common BG (r2 = .80)
e « bahiagrass (r2 - .96)
_______________________ l_24 48 96
Fermentation length (hr)
PREDICTED EFFECT OF 10:0:40 RUMINAL FLUID DILUTION ON IVDMDOF SEVERAL HAYS
60
72
68
64
60
56
52
48
44
40
36
32’
28 r
alfalfa (r2 = .34)
Alicia BG (r .92)
ryegrass (r2 = .72)c
1 1 i h
Figure 6.
24 48
d = common BG (r = .83)2e = bahiagrass (r = .51)
96Fermentation length (hr)
PREDICTED EFFECT OF 25:0:25 RUMINAL FLUID DILUTION ON IVDMDOF SEVERAL HAYS
o* n.
61
for high ruminal fluid concentrations (25:0:25) (Table 13). Alfalfa
hay, which had the highest IVDMD, demonstrated the least increase in
IVDMD from 48 to 96 hr (Tables 14 and 15 and Figures 4 and 5) indica
ting that the majority of the digestion occurred early in the fermenta
tion time. These results indicate a presence of more cell wall materi
al and thicker bundle sheaths in the grasses, making the readily di
gestible tissues less available than in legumes.
There was no interaction between forage and dilution of ruminal
fluid (appendix Table 12 and Figures 7, 8, 9, 10). Plotting in vivo
dry matter digestibility on the corresponding in vitro dry matter di
gestibility scale in Figures 7, 8, 9 and 10 gives an indication of rate
of forage passage or the length of effective digestion time of the for
age in vivo. Alfalfa had the fastest rate of passage and ryegrass the
slowest rate. This observation explains the discrepancy observed
between in vivo and in vitro digestibility data for ryegrass (49.2 vs
43.6) and the similarity in the two values for alfalfa (61.5 vs 62.8).
Therefore IVDMD evaluation of forages should take into consideration,
by in vivo or in vitro methods, the rate of passage of each forage in
order to more accurately predict in vivo digestibility.
The associative effects of alfalfa and ryegrass are summarized in
Table 16, Where the arithmetic mean of in vivo dry matter digestibility
and in vitro dry matter disappearance of alfalfa and ryegrass are com
pared to the observed in vivo and in vitro dry matter digestibility of
50:50 ryegrass + alfalfa hay treatment. Mixing alfalfa and ryegrass
demonstrated practically no associative effects, although a trend to
ward decreased digestibility did exist in the in vivo data. There may
Table 13. FORAGE x TIME INTERACTION FOR SEVERALDILUTION RATES OF RUMINAL FLUID
Fermentationperiod
Ruminal fluid: water:buffer
IVDMD Forage x time Prob of F
24-48 2:8:40 .020*1
24-48 10:0:40 .060
24-48 25:0:25 .194*
48-96 2:8:40 .050*
48-96 10:0:40 .039
48-96 25:0:25 .510it
24-96 2:8:40 .047**
24-96 10:0:40 .001
24-96 25:0:25 .225•ffSignificant at P<.05 Significant at P<.01^Regression coefficients by hay are given in
Tables 14, 15
63
Table 14. PREDICTED EFFECT OF HAY ON RATE OF IVDMD (b) FROM 24-96 HR FERMENTATION FOR TWO RUMINAL FLUID DILUTIONS
_______ Ruminal fluidiwater:buffer_______ 2:8:40__________ 10:0:40
Hay_________________ b1 b1
Ryegrass .28** ««**.22
Alfalfa .06 .02
Common BG .29** .23**
Alicia BG .24** .15*** **
Bahiagrass .33 .25
Significant at P<.05 **Significant at P<.01 Predicted regression coefficient
Table 15. PREDICTED EFFECT OF HAY ON RATE OF IVDMD (b) FROM 24-48 AND 48-96 HR FERMENTATION FOR TWO RUMINAL FLUID DILUTIONS
24-48 hr fermentation period 48--96 hr2:8:40 10:0:40 dilutioni 2:8:40 10:8:40
Hay bi b1 b1 b 1
Ryegrass .50** .46** _.*#r.24 .10
Alfalfa *.26 .11 .01 .03
Common BG **.71 **.47 .10 .17**
Alicia BG .68** .20 *.19 .18**
Bahiagrass .79** .51** *.16 .07
Significant at P<.05 Significant at P<.01 Predicted regression coefficient
64
In Vivo
(ruminal fluid:water:buffer)
2:8:40 '(i2 - .91)
25:0:25 (r2 - .72)
10:0:40 (r2 - .82)
24
Fermentation length'(hr)
Figure 7. PREDICTED EFFECT OF FERMENTATION LENGTH ON IVDMD FORRYEGRASS HAY
65
36
32
— •a— *c
-* b
In Vivo
Cruminal fluld:water:b„ffer)
a = 2;8:^ Cr2 * .64) b * 25-'0:25 (r2 * ,34) c » 10:0:40 (r2 . ,46)
fermentation length (hr)— ■*©«.u ^nr;Figure 8, PreDIcted Effect
alfalfa hay racT 0F amenyation M cth ob IyBffi F0R
% IVDMD
66
In Vivo
(ruminal fluid:water:buffer)
2:8:40 (r2 72)
25:0:25 (r2 83)
10:0:40 (r2 80)
24
Fermentation length (hr)
Figure 9. PREDICTED EFFECT OF FERMENTATION LENGTH FOR COMMONBERMUDAG3ASS HAY
67
60
In Vivo52
(ruminal fluid:water:buffer)
2:8:40 (r .83)
25:0:25 (r .51)
10:0:40 (r .95)
28
Fermentation length (hr)
Figure 10. PREDICTED EFFECT OF FERMENTATION LENGTH ON IVDMD FORBAHIAGRASS HAY
68
Table 16. ASSOCIATIVE EFFECT OF ALFALFA AND RYEGRASS ON FORAGE __________ QUALITY MEASUREMENTS______________________________
ContrastObserved values Arithmetic mean Mean
Constituent of 50;50ryeg+alf vs of 50:50ryeg+alf difference
DMl(g/W*J|) 64.8 65.8 -1.0
DE(Kcal/g) 2.33 2.25 .08
DMD(%) 52.9 55.4 -2.5*
IVDMD (5!) 54.4 53.2 1.2
*Significant at P<.05
69
truely be no associative effect of ryegrass and alfalfa on one or the
other. Or, there may be a trade-off of nutrient/protein between rye
grass and alfalfa, but the increase in rate of passage due to the rapid
digestibility of alfalfa may have offset the theoretical increase in
digestibility due to the improved nutrient/protein. The in vitro data
may have masked the associative effect as the protein content of the
inoculum artificially increased the effective nitrogen content of the
ryegrass substrate.
The effects of inoculum donor, dilution rate of ruminal fluid and
length of fermentation time on IVDMD of alfalfa hay are summarized in
Table 17 and appendix Tables 14, 15 and 16. There was a significant
effect of length of fermentation time, but there was not a significant
effect of dilution rate and inoculum donor among the four animals.
There was no interaction of dilution with donor animal which demon
strates that there is no variation in ruminal fluid dilution although
all animals were withdrawn from water six hours prior to ruminal fluid
collection.
Tables 18 and 19 summarize the effect of in vitro ruminal proce
dure on IVDMD (appendix Table 17). The Van Soest and Troelsen proce
dures had the highest and lowest IVDMD values across hays, respective
ly. The criteria used to establish the validity of the procedures was
based on coefficient of determination (r2), sampling error, coefficient
of variation end standard error of estimate (Table 20). The only
one-stage in vitro procedure used had the lowest correlation with in
vivo data. The Troelsen procedure and the Moore procedure with phos
phate buffer gave the highest coefficient of determination, lowest
70
Table 17. ANALYSIS OF VARIANCE FOR EFFECT OF INOCULUM DONOR, DILUTION RATE OF RUMINAL FLUID AND LENGTH OF FERMENTATION TIME ON IVDMD OF ALFALFA HAY
Source DF SS F
Rep 1 342.2927 4.67
Animals 3 193.5218 0.88
Error a 3 220.0578
Dilution 4 178.7831 1.84
Animal x dilution 12 401.1185 1.37
Error b 16 388.9872
Time 2 795.5382 146.16
Dilution x time 8 46.1391 2.12
Animal x time 6 29.0300 1.78
Animal x dilution x time 24 136.3668 *2.09
Error c 40 108.8584
*Significant at P<.05**Significant at P<.01
71
Table 18. COMPARISON OF IN VITRO DRY MATTER DISAPPEARANCE OFHAYS AMONG RUMINAL IN VITRO PROCEDURES
Hay
Procedures
Van Soest Troelsen Mellenberger MooreMoore w/ phosp.b.
e e d d dRyegrass 46.1 38.3 43.6 43.5 40.9a a a a aAlfalfa 64.8 56.7 60.5 62.2 62.0b b b b b
50:50rye+alf 57.0 48.5 53.4 54.9 53.4be c b b bCommon BG 56.1 43.9 54.0 55.4 54.2c d b b b
Alicia BG 54.9 45.1 52.8 55.0 53.1d e c c cBahiagrass 53.1 44.1 50.0 53.2 51.2
£1}b )C*d}6 Means in the same column followed by different super" scripts are different (P<.01), Scheffe's T.
Table 19. EFFECT OF IN VITRO RUMINAL PROCEDURE ON IVDMD
In Vitro procedure Number of samples Mean IVDMD(%)
Van Soest 18 a55.4« ■ •
Troelsen 18 c45.4
Mellenberger 18 b52.4
Moore 18 ab54.1
Moore w/ phosp.b. 18 b52.5
a>b»c Means with different superscripts are different (P<.01), Duncan's MRT
Table 20. VALIDITY TEST FOR THE PREDICTION OF DRY MATTER DIGESTIBILITY FROM DIFFERENT IN VITRO RUMINAL PROCEDURES
Procedure r*a Repeatability*3 C.V? S.E.E?
Van Soest .86 .18 3.17 .13
Troelsen .88 .34 2.95 .11
Mellenberger .72 .12 3.19 .13
Moore .86 .13 3.19 .13
Moore w/ phosp.b. .87 .24 2.78 .12£^All values significant at (P<.01)Sampling error mean square estimates standard error of duplicate^Coefficient of variation of procedure Standard error of estimate
73
coefficient of variation and lowest standard error of estimate. The
lower variation observed in the Troelsen procedure is probably due to
the addition of urea and glucose to the fermentation media. Nelson et
al. (1969) also observed a significant reduction in the standard devia
tion of IVDMD with addition of urea and glucose. The Moore procedure
with phosphate buffer had a lower variation, because the phosphate buf
fer facilitates the direct acidification by avoiding the excessive
frothing that occurs with a bicarbonate buffer.
III. ENZYME DIGESTION RESULTS
The effect of enzyme procedure is summarized in Tables 21 and 22,
and appendix Table 18. Enzyme procedures using T. viride type IV cel-
lulase had a significantly lower IVDMD than the procedures using T.
viride Onozuka and A. niger type IV cellulase because the Onozuka pre
paration is made up of more active enzyme strains. Procedures with
pepsin pre-treatment had a significantly (P<.0001) lower percent final
IVDMD than procedures with NDF pre-treatment. The criteria used to
establish the validity of the procedure was again based on coefficient
of determination, sampling error, coefficient of variation of the pro
cedure and standard error of estimate (Tables 23 and 24). The
pepsin-T.viride type IV cellulase + A. niger hemicellulase resulted in
the highest coefficient of determination. Across enzymes the T.
viride type IV cellulase + A. niger hemicellulase had the highest cor
relation average. There was no significant difference across
pre-treatments (pepsin vs NDF) in predicting DMD. These results do not
agree with the claim made by Roughan and Holland (1977) that NDF was a
better pre-treatment than pepsin.
74
Table 21. COMPARISON OF IN VITRO DRY MATTER DISAPPEARANCE OF HAYS AMONG ENZYME PROCEDURES
Hay
ProceduresPepsinOnozuka
PepsinT.virideC.A.nigerHc.
PepsinA.nigerC.A.nigerHc.
NDF NDF Onozuka T.virideC.
A.nigerHc.
NDF A.nigerC. A.nigerHc.
e d e c d fRyegrass 46.5 30.4 42.5 50.4 31.3 43.8
a a a a a aAlfalfa 67.8 58.0 65.8 68.3 58.9 66.4
b b b b b b50:50rye+alf 56.8 43.3 52.4 58.1 44.8 54.5
c b be be c cCommon BG 52.0 42.7 50.3 54.0 40.8 52.5
c b cd be c dAlicia BG 52.0 42.3 48.1 52.9 40.1 50.4
d c d c c eBahiagrass 49,0 40.9 46.2 50.1 38.8 47.9
Means in the same column followed by different superscripts are diffferent (P<.05), Scheffe’s T.
Table 22. EFFECT OF IN VITRO ENZYME PROCEDURE ON IVDMD
Enzyme procedure No. of samples Mean IVDMD(%)
Pepsin-Onozuka1 18ab
54.0
Peps in-T.vi r.C.+A.nig.He. 18d
42.9
Pepsin-A.nig.C.+A.nig.Hc. 18c
50.9
NDF-Onozuka 18a55.6
NDF-T.vir.C.+A.nig.Hc. 18d
42.4
NDF-A.nig.C.+A.nig.He. 18be
52.6
Means with different superscripts are different (P<.05), Scheffe's T.1 Contrast of combined Pepsin vs combined NDF procedures (P<.000l)
76
Table 23. VALIDITY TEST FOR THE PREDICTION OF DRY MATTER DIGESTIBILITY FROM DIFFERENT ENZYME PROCEDURES
Procedure r2 3 2br' cRepeatability dC.V. eS.E.E.
Peps in-Onozuka .69 .32 4.68 .15
Pepsin-T.vir.C.+A.nig.Hc. .88 .972 .30 2.81 .08
Pepsin-A.nig.C.+A.nig.Hc. .79 .49 3.78 .12
NDF-Onozuka .72 .35 4.41 .15
NDF-T.vir.C.+A.nig.Hc. .79 .968 .84 3.85 .10
NDF-A.nig.C.+A.nig.Hc. .81 .33 3.66 .12
aAll values significant at (P<.01)“Within pretreatment^Sampling error mean square estimates standard error “Coefficient of variation of the procedure eStandard error of estimate
of duplicate
Table 24. VALIDITY TEST FOR THE PREDICTION OF IN VITRO DRY DISAPPEARANCE FROM DIFFERENT ENZYME PROCEDURES
MATTER
Procedure r 2a cRepeatability C.V*! S.E.E.
Pepsin-Onozuka .72 .32 6.80 .21
Pepsin-T.vir.C.+A.nig.Hc. .94 .947 .30 3.08 .08
Pepsin-A.nig.C.+A.nig.Hc. .79 .49 5.84 .18
NDF-Onozuka .76 .35 6.22 .20
NDF-T.vir.C.+A.nig.Hc. .85 .882 .84 5.07 .13
NDF-A.nig.C.+A.nig.Hc. .83 .33 5.31 .17
aAll values significant at (P<.01)^Within pretreatraentcSampling error mean square estimates standard error of duplicate ^Coefficient of variation of the procedure eStandard error of estimate
77
IV. RESULTS FROM THE NIR STUDY
Reflectance (log 1/R) spectra of ryegrasB and alfalfa hay are pre-
sented in Figure 11. The difference in levels between curves A and R
are a result of particle size and moisture content, rather than of a
difference in composition. The major difference in the spectra is in
the region from 2.25 to 2.40 ym, known to show a combined spectra of
protein and carbohydrates with a NDF peak at 2.34 ym (Norris and
Barnes, 1976; Shenk et al., 1977). A smaller difference in the spectra
is also observed in the region from 1.65 to 1.85 ym due to the CP dif
ference between the hays.
Alfalfa and ryegrass fecal spectra are superimposed on the spectra
of both hays in Figure 12. The alfalfa fecal spectra clearly indicates
a higher dry matter digestibility in the region between 2.25 and 2.40
ym, except for the NDF peak at 2.34 ym, compared to ryegrass which had
very few spectral changes in this region,
Bahiagrass digestibility spectra is superimposed on the spectra of
mature ryegrass digestibility in Figure 13. Again, the major spectral
differences are in the 2.25 to 2.40 ym region, illustrating a higher
protein and carbohydrate digestibility in bahiagrass.
The comparision of laboratory values and NIR predicted values for
various quality constituents of common bermudagrass are presented in
appendix Table 19 and the respective multiple-linear regression analy
sis of second-derivative reflectance data is summarized in Table 25.
Errors in predicting animal responses from NIR data were greater than
those in predicting chemical composition. The magnitudes of the deter
mination coefficients and the standard errors compare well with
1--- 1--- 1--- 1--- 1--- 1--- 1--- 1--- 1--- 1--- 1--- I--- 1--- 1--- 1
RYEG DIET (R> ALFAL DIET (a) 03-APR-82
RYEG
ALFi
1 I1 i lii I l1100 1300 1500 1700 1$00 2100 2300 2500
m
Figure 11. REFLECTANCE (LOG 1/R) SPECTRA OF RYEGRASS AND ALFALFA HAYS-si00
I— *r-"‘ | |--- ir » M .r ,----T r r ^ ■ T |
RYEG DIET (r) ALFAL DIET (A) 03-APR-82
RYEG FECAL <*>ALFAL FF.CAL (a) 0S-APR-82
RYEGDIET
ALFi
FECAL
1106 1306 1566 1760 1900 2160 2308 . 2500H M
Ficure 12. SUPERIMPOSITION OF DIET AND FECAL SPECTRA OF RYEGRASS AND ALFALFA HAYS
I 1--- 1----1--- 1--- 1--- 1--- 1--- 1--- 1--- 1--- 1----1----1----I
RYEG DIET-RYEG FECAL (R) DEFAULT SCALING§AfilA £1aSS OIET-BAHIA GRASS FECAL (b)DEFAULT SCALING8-APR-82
1- 1 . 1 I L l r i i » _ _ _ _ ] _ _ _ _ 1 1 , 1118S 13S0 1500 1700 1900 2100 23O0 2500
NMFigure 13. SUPERIMPOSITION OF DIGESTIBILITY SPECTRA OF BAHIAGRASS AND MATURE RYEGRASS HAYS
81
Table 25. SUMMARY OF MULTIPLE-LINEAR REGRESSION ANALYSIS RELATING THE CHEMICAL COMPOSITION OF HAY AND IN VIVO DATA TO NIR
Regression data __________ Wavelengths (um)______Variable N R2 SE__________ A i * A2________As_______ U
DMl(g/W *k|) 29 .84 6.68
DE(Kcal/g) 29 .75 0.11
DMD(Z) 29 .69 2.66
ivdmd(%) 29 .90 2.02
c p(%) 29 .97 0.59
CF(%) 29 .87 1.35
NDF(Z) 29 .98 1.56
ADF(Z) 29 .96 1.22
L(%) 29 .73 .69
2.298 1.750 1.258 1.154
1.522 2.406 1.362 1.210
1.326 2.266 1.522
1.494 2.306 1.702 1.358
1.726 1.622 2.154 1.166
2.266 1.978 1.358 1.230
2.290 2.062 2.206 2.414
2.270 1.702 2.198 1.534
2.374 1.514 2.278 1.806
aActual wavelengths (ym) selected by NIR to predict each hay constituent
82
published values (Norris et al., 1976; Barnes, 1973).
The regression data presented, suggest that infrared reflectance
measurements can successfully predict forage laboratory analysis and in
vivo data.
V. PREDICTION OF FORAGE NUTRITIVE VALUE USING SIMPLE AND MULTIPLE LINEAR REGRESSION
In the present study there were different relationships onong in
vivo parameters and laboratory values for the various hays studied.
Therefore, some laboratory methods or their combination may be better
predictors of nutritive value than others.
The prediction of dry matter intake from laboratory analysis using
simple regressions is summarized in Table 26, and appendix Table 20.
The intake of the hays could be significantly predicted by IVDMD and
CP, r2“.78 and r2*.996, respectively. The high coefficient of determi
nation obtained with CP is partially due to the wide range in CP ob
served in the hays. Neutral-detergent fiber did not significantly pre
dict DMI, but demonstrated a relatively high coefficient of determina
tion, r 2m. 5 3 . Van Soest (1965b) considered NDF as a good predictor of
DMI. Using stepwise regression, there were several combinations of
chemical components that had significant correlation with DMI (Table 27
and appendix Table 21). Hemicellulose was present in all combinations,
followed by C and L (all cell wall components) and CP. Table 28 sum
marizes some proposed equations to predict DMI. The predicted values
were analyzed against the observed values. Both equations by Wilson
and McCarrick (1966) and Minson (1972) were good predictors of DMI.
83
Table 26. PREDICTION OF DRY MATTER INTAKE FROM LABORATORY ANALYSIS WITH SIMPLE REGRESSION
Constituent na r2 bb S.E.E?
IVDMD 6 .78* 2.28 0.6
CP 6 .99* 4.81 0.1
CF 6 .00 0.02 2.1
NDF 6 .53 -0.93 0.4
ADF 6 .29 -1.76 1.4
C 6 .09 -1.44 2.3
He 6 .28 -0.71 0.6
L 6 .03 -2.42 6.7
®Mean of a 6x6 LSD Significant (P<.05) Regression coefficient
cStandard error of estimate
84
Table 27. VARIABLES USED IN THE PREDICTION OF DRY MATTER INTAKE FROM LABORATORY ANALYSIS WITH STEPWISE REGRESSION8
nb- Constituent
6 CP, He
6 CP,Hc,L
6 C,He|L
6 CP,C,Hc,L
6 CP,NDF,C,Hc,La.With stepwise and max R improvements Mean of a 6x6 LSD
Table 28. PREDICTION OF DRY MATTER INTAKE (g/W’Kg) FROM __________ PROPOSED EQUATIONS__________________________
Source Equation bn r2 bc SEEd
Wilson and McCarrick (1966) 1.3(DMD)-30.1 6*
.77 2.62 .70
Minson (1972) .75(DMD)+7.4 6 *.77 4.56 1.22
Rohweder et ml. (1976) 54.8+1.22NDF-.018NDF2 6 .41 0.78 .46
The predicted values obtained by the proposed equations were analyzed xgainst the observed values obtained with the six forages ■?Mean of a 6x6 LSD cSignificant at (P<.05)^Regression coefficient Standard error of estimate
85
Dry matter digestibility (DMD) of the hays could be significantly
predicted with CP, IVDMD and ADF (Table 29 and appendix Table 22). In
vitro dry matter disappearance is the most widely used predictor of DMD
(Oh et al., 1966; Van Soest et al., 1966 and Van Soest, 1965b). Dry
matter digestibility has been significantly predicted by ADF (Van
Soest, 1965a; Yu and Thomas, 1973 and Rohweder et al., 1976). Dry mat
ter digestibility of the hays could be predicted by several combina
tions of chemical components (Table 30 and appendix Table 23). Crude
protein was included in all combinations followed by He, L, ADF and C.
Table 31 summarizes proposed equations to predict DMD. The predicted
values obtained with the equations were regressed against the observed
values. Dry matter digestibility was significantly predicted by equa
tions proposed by Bredon et al. (1963) based on CP, Reid et al. (1973)
based on IVDMD, and the known Van Soest summative equation, tfiich pre
dicted DMD with r2-.9999 (PC.01) and SEE-.0015.
In vitro dry matter digestibility, like DMD, was significantly
predicted by CP and ADF (Table 32 and appendix Table 24). In vitro dry
matter disappearance of hays could also be predicted by several combi
nations of chemical components (Table 33 and appendix Table 25). Crude
protein and CF were included in all combinations, followed by L and
NDF.
Table 29. PREDICTION OF DRY HATTER DIGESTIBILITY FROMLABORATORY ANALYSIS WITH SIMPLE REGRESSION
Constituent an r* bb S.E.E®
IVDMD 6 **.90 0.64 0.10
CP 6 *.77 1.10 0.29
CF 6 .18 -0.46 0.49
NDF 6 .35 -0.20 0.14
ADF 6*
.71 -0.70 0.23
C 6 .40 -0.77 0.17
He 6 .08 -0.09 0.17
L 6 .14 -1.33 1.63aAMean of a 6x6 LSD ^Significant (P<.05) kSignificant (P<.01) cRegression coefficient Standard error of estimate
Table 30. VARIABLES USED IN THE PREDICTION OF DRYMATTER DIGESTIBILITY FROM LABORATORY ANALYSIS WITH STEPWISE REGRESSIOlf
n Constituent
6 CP,ADF
6 CP,ADF,He
6 ADF,He,L
6 CP,ADF,He,L
6 CP,C,Hc,L
6 CP,NDF,C,He,L£kWith stepwise and max R improvements Mean of a 6x6 LSD
Table 31. PREDICTION OF DRY MATTER DIGESTIBILITY FROM PROPOSED EQPATIONS a
Source Equation bn r2 bc SEEd
Bredon et al. (1963) 41.81+1.63CP 6 .74* 0.22 0.07
Bredon et al. (1963) 120.03-1.778CF 6 .18 0.26 0.28
Reid et al. (1973) 24.51+.631IVDMD 6 .90 1.01 0.16
Rohweder et al. (1976) 34.8+2.56ADF-.0491ADF2 6 .69 0.40 0.43
Deinum and Tan Soest (1969) 1.81-.9661og(100x(L/ADF))-20 6 .42 5.42 3.17
Tan Soest (1965) .98CC+CWC(1.473-.7891ogL/ADFxl00)-12.9 6 **.99 0.77 0.00
aThe predicted values obtained with the proposed equations were analyzed against the observed values obtained with the six forages ^Mean of a 6x6 LSD *Significant at (P<.05)^Significant at (P<.01) cRegression coefficient ^Standard error of estimate
88
Table 32. PREDICTION OF IN VITRO DRY HATTER DISAPPEARANCE FROM LABORATORY ANALYSIS WITH SIMPLE REGRESSION
Constituent na r2 bb s.E.ti':
CP 6*
.79 1.65 0.43
CF 6 .20 -0.73 0.72
NDF 6 .40 -0.31 0.19
ADF 6*
.72 -1.07 0.33
C 6 .42 -1.19 0.69
He 6 .08 -0.15 0.24
L 6 .12 -1.77 2.40
aMean of a 6x6 LSD Significant (P<.05) Regression coefficient
cStandard error of estimate
Table 33. VARIABLES USED IN THE PREDICTION OF IN VITRO MATTER DISAPPEARANCE FROM LABORATORY ANALYSIS WITH STEPWISE REGRESSION'1
DRY
nb Constituent
6 CP.CF
6 CP,CF,L
6 CP,CF,NDF,L
6 CP.CF.NDF.C.Ln.With stepwise and max R improvements Mean of a 6x6 LSD
SUMMARY
Six in vivo conventional intake and digestion trials were con
ducted with alfalfa hay, mature ryegrass, common and Alicia bermuda-
grass, bahiagrass and a 50:50 mixture of alfalfa and ryegrass, using
six crossbred steers. Intake and digestion trials ranked the hays in
the same order of quality. Alfalfa hay and mature ryegrass had the
highest and lowest quality, respectively. The quality of warm season
grasses ranked according to length of regrowth period. Bahiagrass with
the longest (6 weeks) period of regrowth, had a lower quality than com
mon and Alicia bermudagrass; although steers consumed a larger amount
of bahiagrass dry matter than Alicia bermudagrass dry matter probably
due to a higher protein content (12.4 vs 10.3%).
There were highly significant positive correlations among DMI, DE,
DMD, IVDMD and CF, as they are all considered to be positively associa
ted with high quality in forages. All these forage entities were nega
tively correlated with the more poorly utilized cell wall components.
There were also significant relative relationships between cell wall
components and in vivo data. Acid-detergent fiber was significantly
correlated with in vivo digestibility, while NDF failed to be signifi
cantly correlated with DMI, although there was a relatively high corre
lation (r".73).
The results from the in vitro ruminal fermentations indicate that
there was a significant increase in IVDMD with increased dilution of
ruminal fluid, both across hays and inoculum donors. The rate of
increase was reduced considerably tfien the volume of added water
exceeded the volume of ruminal fluid used and IVDMD was drastically
89
90
decreased with further dilution of ruminal fluid. In vitro dry matter
disappearance for all hays increased as the length of incubation time
increased. There was a significant interaction of forage with fermen
tation time. Alfalfa hay irtiich had the highest IVDMD, showed the least
increase in IVDMD from 48 to 96 hours. The forage x time interaction
was not significant for high ruminal fluid concentrations. No interac
tion between forage and dilution of ruminal fluid was observed. There
waB a discrepancy between in vivo and in vitro digestibility data for
ryegrass (49.2 vs 43.6%) and close agreement for alfalfa (61.5 vs
62.5%). These differences are probably due to the very low and very
high intakes observed for ryegrass and alfalfa, respectively, and show
the necessity for any laboratory or in vitro techniques for predicting
in vivo digestibility to consider intake or rate of passage.
No associative effect was observed between alfalfa and mature
ryegrass when mixed in a 50:50 ratio, although there was a trend toward
a depressed digestibility observed in the in vivo data. The in vitro
data may have masked the negative associative effect since the protein
content of the inoculum probably artificially increased the low nitro
gen content of ryegrass.
All in vitro ruminal procedures presented highly significant coef
ficients of determination in predicting in vivo digestibility. The
Troelsen procedure and the Moore procedure with phosphate buffer showed
the highest correlation and the lowest standard error of estimate.
Enzymatic procedures also had significant coefficients of determi
nation in predicting in vivo digestibility; however, they resulted in a
wider range of values than the ruminal procedures. Pepsin-T. viride
91
cellulase + A. niger hemicellulase and pepsin-Onozuka cellulaae prepa
rations showed the highest and lowest coefficient of determination,
respectively, in predicting in vivo digestibility. There were no dif
ferences in prediction due to pre-treatments (pepsin vs NDF).
In vivo data were significantly predicted by several chemical com
ponents. Dry matter intake and digestibility were both significantly
predicted by CP and IVDMD. The best predictors for DMI and DMD were CP
and IVDMD, respectively. Acid-detergent fiber significantly predicted
DMD, tfiile NDF failed to significantly predict DMI. Stepwise linear
regression showed that CP, He and L were present in most all combina
tions of chemical components to predict DMI and DMD. Van Soest summa-
tive equation was a highly accurate predictor of DMD (r2 “1.00,
SEE-.00)
Reflectance (log 1/R) spectra was successful in graphically iden
tifying differences in digestibility between alfalfa and mature rye
grass, and between bahiagrass and mature ryegrass. Multiple linear re
gression analysis of second-derivative reflectance data predicted in
vivo data with highest accuracy for DMI, r2-.84 and lowest for DMD,
r2-.69. Prediction of laboratory analysis was highest for NDF, r2-.98
and lowest for CF, r2-.87. Errors in predicting animal responses were
greater than those in predicting chemical composition. The magnitude
of the coefficient of determination and the standard error of estimate
indicate that infrared reflectance measurements can successfully pre
dict forage laboratory analysis and in vivo data. These results indi
cate that chemometric techniques, such as near infrared reflectance
spectroscopy, are able to successfully compete with conventional animal
92
trials and wet chemistry analysis in determining forage value. The
near infrared technique is fast, adapted to field analysis and with
perfection and better calibration may become the preferred forage eval
uation method in the near future.
This study serves as a developmental basis for chosing evaluation
techniques depending on the criteria (highest correlation and lowest
standard error of duplicate) and constraints (time, expense, labor and
number of samples) that one sets forth.
LITERATURE CITED
Akin, D. E. and D. Burdick. 1973. Microanatomical Differences of Wartn- Season Grasses Revealed by Light and Electron Microscopy. Agron. J. 65:533.
Akin, D. E. and D. Burdick. 1975. Percentage of Tissue Types in Tropical and Temperate Grass Leaf Blades and Degradation of Tissues by Rumen Microorganisms. Crop. Sci. 15:661.
Akin, D. E. and D. Burdick. 1979. Preliminary Study on the Analysis of Forages with a Filter Type Near Infrared Spectrometer. J. Agr. Food Chem, 27:1248.
A.O.A.C. 1980. Official Methods of Analysis. 13th Edition. Association of Official Agricultural Chemists, Washington D. C.
Asplund, J. M., R. T. McElroy and W. J. Pidgen. 1958. Dry Matter Loss and VFA Production in the Artificial Rumen as Indices of Forage Quality. Can. J. Anim. Sci. 38:171.
Balch, C. C. 1950. Factors Affecting the Utilization of Food by Dairy Cows. I. The Rate of Passage of Food Through the Digestive Tract. Brit. J. Nutr. 4:361.
Balch, C. C. and R. C. Campling. 1962. Regulation of Voluntary Food Intake in Ruminants. Nutr. Abstr. Revs. 32:669.
Balch, C. C. and V. W. Johnson. 1950. Factors Affecting the Utilization of Food by Dairy Cows. II. Factors Influencing the Rate of Breakdown of cellulose in the Rumen of the Cow. Brit. J. Nutr. 4:389.
Bales, G. L., D. W. Kellogg and D. D. Miller. 1976. Small Volume ofInoculum with an Artificial Rumen Fluid for in vitro digestion of Forage. J. Dairy Sci. 59:1950.
Barnes, R. F. 1966. The Development and Application of In Vitro Rumen Fermentation Techniques. Proc. 10th Int. Grass1. Congr. Helsinki, Finland, p.434.
Barnes, R. F. 1969. Collaborative Research with the Two-Stage In Vitro Rumen Fermentation Technique. Proc. Natl. Conf. on Forage Eval. and Util. Lincoln, Nebraska.
Barnes, R. F. 1970. In Vitro Dry Matter and Organic Matter Digestion. In: Nutrition Research Techniques for Domestic and Wild Animals. Vol I. L. E. Harris (Ed). Logan, Utah.
93
94
Barnes, R. F. 1973. Laboratory Methods of Evaluating Feeding Value.In: Chemistry and Biochemistry of Herbage. Vol III. G. W. Butler and R. W. Bailey (Ed). Academic Press, New York.
Barnes, R. F. and G. C. Marten. 1979. Recent Developments in Predicting Forage Quality, J. Anim. Sci. 48:1554.
Barton, F. E., II, and S. W. Coleman. 1981. Potential of Near Infrared Reflectance Spectroscopy for Measuring Forage Quality. Oklahoma Agr. Exp. Sta. Misc. Publ. 108:73.
Baumgardt, B. R., J. L. Cason and M. W. Taylor. 1962. Evaluation of Forages in the Laboratory. V. Comparative Accuracy of Several Methods. J. Dairy Sci. 45:59.
Baumgardt, B. R. and H. K. Oh. 1964. Evaluation of Forages in theLaboratory. IV. Within and Among Trial Variability of Wisconsin Artificial Rumen Procedure. J. Dairy Sci. 47:263.
Bezeau, L. M. 1965. Effect of Source of Inoculum on Digestibility of Substrate in In Vitro Digestion Trials. J. Anim. Sci. 24:823.
Bines, J. A., S. Suzuki and C. C. Balch. 1969. The QuantitativeSignificance of Long-Term Regulation of Food Intake in the Cow. Brit. J. Nutr. 23:695.
Blaxter, K. L. 1950. Energy Feeding Standards for Dairy Cattle. Nutr. Abstr. Rev. 20:1.
Blaxter, K. L., F. W. Wainman and J. L. Davidson. 1966. The Voluntary Intake of Food by Sheep and Cattle in Relation to their Energy Requirements for Maintenance. Anim. Prod. 8:75.
Blaxter, K. L., F. W. Wainman and R. S. Wilson. 1961. The Regulation of Food Intake by Sheep. Anim. Prod. 3:51.
Bredon, R. M., K, W. Harter and B. Marshall. 1963. The Nutritive Value of Grasses Grown in Uganda when Fed to Zebu Cattle. J. Agr. Sci. 61:101.
Burdick, D., F. E. Barton, II and B. D. Nelson. 1981. Prediction of Bermudagrass Composition and Digestibility with a Near Infrared Multiple Filter Spectrophotometer. Agron. J. 73:399.
Burton, B. W., J. E. Jackson and R. H. Hart. 1963. Effects of Cutting Frequency and Nitrogen on Yield, In Vitro Digestibility and Protein, Fiber and Carotene Content of Coastal Bermudagrass. Agron. J. 55:500.
95
Butterworth, H. H. 1964. The Digestible Content of Some Tropical Forages. J. Agr. Sci. 63:319.
Butterworth, M. H. 1967. The Digestibility of Tropical Grasses. Nutr. Abstr. Revs. 37:349.
Butterworth, H. H. and J. A. Diaz. 1970. UBe of Equations to Predict the Nutritive Value of Tropical Grasses. J. Range Manage. 23:55.
Campling, R. C., M. Freer and C. C. Balch. 1962. Factors Affecting the Voluntary Intake of Food by Cows. Brit. J. Nutr. 16:115.
Chalupa, W. and D. D. Lee, Jr. 1966. Estimation of Forage Nutritive Value from In Vitro Cellulose Digestion. J. Dairy Sci. 49:188.
Chenost, M. 1966. Fibrousness of Forage: Its Determination and Its Relation to Feeding Value. Proc. 10th Int. Grassl. Congr. Helsinki, Finland, p.406.
Clancy, M. J. and R. K. Wilson. 1966. Development and Application of a New Chemical Method for Predicting the Digestibility and Intake of Herbage Samples. Proc. 10th Int. Grassl. Congr. Helsinki, Finland, p.445.
Clanton, D. C. 1961. Comparison of 7- and 10- Day Collection Periods in Digestion and Metabolism Trials with Beef Heifers. J. Anim. Sci. 20:640.
Clark, K. W. and G. 0. Mott. 1960. The Dry Matter Digestion In Vitro of Forage Crops. Can. J. Plant. Sci. 40:123.
Cochran, W. G. and G. M. Cox. 1950. Experimental Designs. 2nd Edition. John Willey & Sons, Inc., New York.
Crampton, E. W. 1957. Interrelations Between Digestible Nutrient and Energy Content, Voluntary Dry Matter Intake and the Overall Feeding Value of Forages. J. Anim. Sci. 16:546.
Crampton, E. W., E. Donefer and L. E. Lloyd. 1960. A Nutritive Value Index for Forages. J. Anim. Sci. 19:538.
Crampton, E. W., E. Donefer and L. E. Lloyd. 1962. Caloric Equivalent of the Nutritive Value Index. J. Anim. Sci. 21:628.
Deinum, B. 1966. Influence of Some Climatological Factors on the Chemical Composition and Feeding Value of Herbage. Proc. 10th Int. Grassl. Congr. Helsinki, Finland, p.415.
96
Deinum, B., A. J. H. Van Es and P. J. Van Soest. 1968. Climate, Nitrogen and Grass. II. The Influence of Light Intensity, Temperature and Nitrogen on In Vivo Digestibility of Grass and the Prediction of these Effects from Some Chemical Procedures. Neth. J. Agr.Sci. 16:217.
Deinum, B. and P. J. Van Soest. 1969. Prediction of Forage Digestibility from Some Laboratory Procedures. Neth. J. Agr. Sci. 17:119.
Donefer, E., E. W. Crampton and L. E. Lloyd. 1960. Prediction of the Nutritive Value Index of a Forage From In Vitro Rumen Fermentation Data. J. Anim. Sci. 19:545.
Donefer, E., P. J. Niemann, E. W. Crampton and L. E. Lloyd. 1963. Dry Matter Disappearance by Enzyme and Aqueous Solution to Predict the Nutritive Value of Forages. J. Dairy Sci. 46:965.
Duble, R. L., J. A. Lancaster and E. C. Holt. 1971. Forage Characteristic Limiting Animal Performance on Warm-Season Perennial Grasses. Agron. J. 63:795.
Egan, A. R. and R. J. Moir. 1965. Nutritional Status and Intake Regulation in Sheep. Australian J. Agr. Res. 16:437.
Elliot, R. C. 1967. Voluntary Intake of Low-Protein Diets by Ruminants.I. Intake of Food by Cattle. J. Agr. Sci. 63:375.
Elliot, R. C. and J. H. Topps. 1963. Voluntary Intake of Low ProteinDiets by Sheep. Anim. Prod. 5:269.
Esau, K. 1977. Anatomy of Seed Plants. 2nd Edition. John Wiley & Sons. New York.
Fonnesbeck, P. V. 1976. Estimating Nutritive Value from Chemical Analysis. In: P. V. Fonnesbeck, L. E. Harris (Ed). 1st Int. Symp. On Feed Composition, Animal Nutrient Requirements and Computerization of Diets. Utah Agr. Exp. Sta. Utah State Univ. Logan, Utah.
Goering, H. K. and P. J. Van Soest. 1970. Forage Fiber Analysis (Apparatus, Reagents Procedures and Some Applications). Agr. Res. Serv. USDA. Agriculture Handbook No. 379.
Goto, I. and D. J. Minson. 1977. Prediction of the Dry Matter Digestibility of Tropical Grasses Using a Pepsin-Cellulase Assay. Anim. Feed. Sci. Technol. 2:247.
Grant, R. J., P. J. Van Soest and R. E. McDowell. 1974. Influence of Rumen Fluid Source and Fermentation Time on In Vitro True Dry Matter Digestibility. J. Dairy Sci. 57:120.
97
Guggolz, J., R. M. Saunders, G. 0. Kohler and T. J. Klopfenstein.1971. Enzymatic Evaluation o£ Processes for Improving Agricultural Wastes for Ruminant Feeds. J. Anim. Sci. 33:167.
Hanna, W. W., W. G. Monaon and G. W. Burton. 1973. Histological Examinations of Fresh Forage Leaves After In Vitro Digestion. Crop. Sci. 13:99.
Harrison, D. G., D. E. Beever, D. J. Thomson and D. F. Osbourn. 1975. Manipulation of Rumen Fermentation in Sheep by Increasing the Rate of Flow of Water from the Rumen. J. Agr. Sci. 85:93.
Heaney, D. P. 1970. Voluntary Intake as a Component of an Index to Forage Quality. Proc. Natl. Conf. on Forage Eval. and Util. Lincoln, Nebraska.
Heaney, D. P., G. I. Pritchard and W. J. Pidgen. 1968. Variability in Ad Libitum Forage Intake by Sheep. J. Anim. Sci. 27:159.
Hershberger, T. V., T. A. Long, E. W. Hartsook and R. W Swift. 1959. Use of the Artifical Rumen Technique to Estimate the Nutritive Value of Forages. J. Anim. Sci. 18:770.
Holt, E. C. and B. E. Conrad. 1981. Factors Affecting Bermudagrass Production and Forage Quality. Texas Agr. Exp. Sta. PR-3868.
Hungate, R. E., G. D. Phillips, D. P. Hungate and A. MacGregor. 1960.A Comparison of the Rumen Fermentation in European and Zebu Cattle. J. Agr. Sci. 54:196
Johnson, R. R. 1969. Reasons for Forage Evaluation. Proc. Natl. Conf. on Forage Eval. and Util. Lincoln, Nebraska.
Johnson, R. R. and B. A. Dehority. 1968. A Comparison of Several Laboratory Techniques to Predict Digestibility and Intake of Forages. J. Anim. Sci. 27:1738.
Johnson, R. R, B. A. Dehority and 0. G. Bentley. 1958. Studies on the In Vitro Rumen Procedure: Improved Inoculum Preparation and the Effects of Volatile Fatty Acids on Cellulose Digestion. J. Anim. Sci. 17:841.
Johnson, R. R., B. A. Dehority, H. R. Conrad and R. R. David. 1962. Relationship of In Vitro Cellulose Digestibility of Undried and Dried Mixed Forages to Their In Vivo Dry Matter Digestibility. J. Dairy Sci. 45:250.
98
Jones, G. M. 1972. Chemical Factors and Their Relation to Feed Intake Regulation in Ruminants: A Review. Can. J. Anim. Sci. 52:207.
Jones, D. I. H. and M. V. Hayward. 1973. A Cellulase Digestion Technique for Predicting the Dry Matter Digestibility of Grasses. J. Sci. Food Agr. 24:1419.
Jones, D. I. H. and M. V. Hayward. 1975. The Effect of Pepsin Pretreatment of Herbage on the Prediction of Dry Matter Digestibility from Solubility in Fungal CellulaBe Solutions. J. Sci. Food Agr. 26:711.
Kamstra, L. D., A. L. Moxon and 0. G. Bentley. 1958. The Effects ofStage of Maturity and Lignification on the Digestion of Cellulose in Forage Plants by Rumen Microorganisms In Vitro. J. Anim. Sci. 17:199.
Marten, G. C. and R. F. Barnes. 1980. Prediction of Energy Digestibility of Forages with In Vitro Rumen Fermentation and Fungal Enzyme Systems. Symp. on Standardization of Analytipal Methodology for Feeds. Int. Devel. Res. Ctr. Otawa, Canada.
McLeod, M. N. and D. J. Minson. 1969. Sources of Variation in the In Vitro Digestibility of Tropical Grasses. J. Brit. Grassl. Soc. 24:244.
McConnel, A. A. and M. A. Eastwood. 1974. Comparison of Methods ofMeasuring Fiber in Vegetable Material. J. Sci. Food Agr. 25:1451.
McQueen, R. E. and P. J. Van Soest. 1975. Evaluation of Forage with Cellulolytic Enzymes From Trichoderma Viride. J. Dairy Sci.56:680.
Mellenberger, R. W., L. D. Satter, M. A. Millett and A. J. Baker. 1970. An In Vitro Technique for Estimating Digestibility of Treated and Untreated Wood. J.Anim. Sci. 30:1005.
Mertens, D. R. an P. J. Van Soest. 1973. Prediction of Voluntary Forage Intake. J. Anim. Sci. 37:296.
Milford, R. and D. J. Minson. 1965. Intake of Tropical Pasture Species. Proc. 10th Int. Grassl. Congr. Helsinki, Finland, p.815.
Minson, D. J. 1971. The Nutritive Value of Tropical Pastures. J. Australian Ind. Agr. Sci. 37:225.
99
Minson, D. J. 1972. The Digestibility and Voluntary Intake by Sheep of Six Tropical Grasses. Australian J. Exp. Agr. 12.21.
Minson, D. J. and M. N. McLeod. 1970. The Digestibility of Temperate and Tropical Grasses. Proc. 11th Int. Grassl. Congr. Queensland, Australia, p.719.
Minson, D. J. and R. Milford. 1967. The Voluntary Intake and Digestibility of Diets Containing Different Proportions of Legume and Mature Pangolagrass. Australian J. Exp. Agr. and Anim. Husb. 7:546.
Moir, K. W., J. R. Wilson and G. W. Blight. 1977, The In Vitro Digested Cell Wall and Fermentation Characteristics of Grasses as Affected by Temperature and Humidity During Their Growth. J. Agr.Sci. 88:217.
Monson, W. G., J. B. Powel and G. W. Burton. 1972. Digestion of Fresh Forage in Rumen Fluid. Agron. J. 64:231.
Montgomery, C. R., M. Allen, J. G. Kowalczuk, P. E. Schilling andL. Mason. 1979. Nutritional an Agronomic Comparisons of Four Warm- Season Perennial Grasses. Louisiana Agr. Exp. Sta. Bui.715.
Moore, J. E. 1970. In Vitro Dry Matter or Organic Matter Digestion. In: L. H. Harris (Ed) Nutrition Research Techniques for Domestic and Wild Animals. Logan, Utah.
Moore, J. E., J. C. Burns, A. C. Linnerud and R. J. Monroe. 1980. Relationship Between the Properties of Southern Forages and Animal Response. Proc. Southern Pasture Forage Crop. Improv. Conf. Nashville, Tennessee.
Moore, J. E. and G. 0. Mott. 1972. Structural Inhibitors of Quality in Tropical Grasses. Symp. on Antiquality Components in Forage, ASA. Miami Beach, Florida.
Mott, G. 0. and J. E. Moore. 1970. Forage Evaluation Techniques inPerspective. Proc. Natl. Conf. on Forage Eval. and Util. Lincoln, Nebraska.
Neher, N. C. 1976. The Influence of Rumen Fluid Collection Procedures and Fermentation Techniques on In Vitro Rumen Dry Matter Digestibility. M. S. Thesis. Dept. Anim. Sci. Louisiana State Univ.
Nelson, B. D., H. D. Ellzey, C. Montgomery and E. B. Morgan. 1969.Factors Affecting the Variability of an In Vitro Rumen Fermentation Technique for Estimating Forage Quality. J. Dairy Sci.55:358.
100
Nelson, B. D., C. R. Montgomery, P. E. Schilling and L. Mason. 1975. Effect of Fermentation Time on In Vivo/In Vitro Relationship. J. Dairy Sci. 59:270.
Neter, J. and W. Wasserman. 1974. Applied Linear Statistical Models.1st Edition. Richard D. Irwin Inc. Homewood, Illinois.
Norris, K. H. and R. F. Barnes. 1976. Infrared Reflectance Analysis of Nutritive Value of Feedstuffs. In: P. V. Fonnesbeck, L. E. Harris (Ed). 1st Int. Symp. on Feed Composition, Animal Nutrient Requirements, and Computerization of Diets. Utah State Univ. Logan, Utah.
Norris, K. H., R. F. Barnes, J. E. Moore and J. S. Shenk. 1976. Predicting Forage Quality by Infrared Reflectance Spectroscopy. J. Anim. Sci. 43:889.
Norris, R. H. and J. R. Hart. 1965. Direct Spectrophotometric Determination of Moisture Content of Grain and Seeds. Proc. 1963 Int. Symp. on Humidity and Moisture. Humidity and Moisture 4:19.
Oh, H. K., B, R. Baumgardt and J. M. School. 1966. Evaluation of Forages in the Laboratory. V. Comparison of Chemical Analysis, Solubility Tests, and In Vitro Fermentation. J, Dairy Sci. 49:850.
Packett, L. V., M. C. Plumlee, R. Barnes and G. 0. Mott. 1965. Influence of Hemicellulose A and B on Cellulose Digestion, V.F.A. Production and Forage Nutritive Evaluation. J. Nutr. 85:89.
Pendlum, C., J. A. Boling, L. P. Bush and R. C. Buckner. 1980. Digestibility and Metabolism of Kenhy Tall Fescue Harvested at Three Stages of Physiological Maturity. J. Anim. Sci. 51:704,
Raymond, W. F. and D. J. Minson. 1955. The Use of Chromic Oxide for Estimating the Fecal Production of Grazing Animals. J. Brit. Grassl. Soc. 10:282.
Reid, R. L., A. J. Post, F. J. Olsen and J. S. Mugerwa. 1973. Studies on the Nutritional Quality of Grasses and Legumes in Uganda. I. Application of In Vitro Digestibility Techniques to Species and Stage of Growth Effects. Trop. Agr. 50:1.
Riewe, M. E. and H. Lippke. 1969. Considerations in Determining the Digestibility of Harvested Forages. Proc. Natl. Conf. on Forage Eval. and Util. Lincoln, Nebraska.
101
Robles, A. Y., R. L. Belyea, F. A. Marts and M. F. Weiss. 1980. Effect of Particle Size Upon Digestible Cell Wall and Rate of In Vitro Digestion of Alfalfa and Orchardgrass Forages. J. Anim. Sci. 51:783.
Rohweder, D. A., R. F. Barnes and N. Jorgensen. 1976. The Use of Chemical Analysis to Establish Hay Standards. In: P. V. Fonnesbeck,L. E. Harris (Ed). 1st Int. Symp. on Feed Composition, Animal Nutrient Requirements and Computerization of Diets. Utah Agr. Exp. Sta. Utah State Univ. Logan, Utah.
Roughan, P. G. and R. Holland. 1977. Predicting In Vivo Digestibilities of Herbages by Exhaustive Enzymic Hydrolysis of Cell Walls. J.Sci. Food Agr. 28:1057.
Shenk, J. S., K. H. Norris, R. F. Barnes and G. W. Fissel. 1977. Forage and Feedstuff Analysis With Infrared Reflectance Spectro Computer System. Proc. 13th Int. Grassl. Congr. Leipzig. G.D.R.
Southgate, D. A. T. 1969, Determination of Carbohydrates in Foods. J. Sci. Food Agr. 20:326.
Sullivan, J. T. 1962. Evaluation of Forage Crops by Chemical Analysis.A Critique. Agron. J. 54:511.
Sullivan, J. T. 1964. Chemical Composition of Forages in Relation to Digestibility by Ruminants. CropB Research, USDA. ARS. p.34.
Steel, R. G. and J. H. Torrie. 1980. Principles and Procedures of Statistics. A Biometrical Approach. 2nd Edition. McGraw-Hill Book Co. New York.
S.A.S. 1979. SAS User's Guide. 1979 Edition. SAS Inst. Inc. Raleigh, North Carolina.
S.A.S. 1982. SAS User's Guide. 1982 Edition. SAS Inst. Inc. Raleigh, North Carolina (In Press).
Thiago, L. R. L., R. C. Kellaway and J. Leibholz. 1979. Kinetics of Forage Digestion in the Rumen. Ann. Res. Vet. 10:329.
Thomson, D. J., D. E. Beever, D. C. Murcell, M. L. Elderfield and D. G. Harrison. 1975. The Effect of Altering Dilution Rate on the Pattern of Fermentation in the Rumen. Proc. Nutr.' Soc.34:111A.
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.
102
Troelsen, J. E. 1970. In Vitro Digestion of Forage Samples. In: Nutrition Research Techniques for Domestic and Wild Animals. Vol I.L. E. Harris (Ed). Logan, Utah.
Troelsen, J. E., J. B. Campbell and D. H. Heinrichs. 1970. The Effect of Physical Breakdown on the Voluntary Intake of Coarse Roughage by Sheep. Proc. 11th Int. Grassl. Congr. Australia, p.747.
Troelsen, J. E. and D. J. Hanel. 1966. Ruminant Digestion In Vitro as Affected by Inoculum Donor, Collection Day and Fermentation Time. Can. J. Anim. Sci. 46:149.
Van Soest, P. J. 1963. The Use of Detergents in the Analysis of Fibrous Feeds. II. A Rapid Method for the Determination of Fiber and Lig- nin. J. A.O.A.C. 46:829.
Van Soest, P. J. 1964. New Chemical Procedures for Evaluating Forages. Symp. on Nutrition of Forage and Pastures. J. Anim. Sci. 23:838.
Van Soest, P. J. 1965a. Symp. on Factors Influencing the Voluntary Intake in Relation to Chemical Composition and Digestibility. J. Anim. Sci. 24:834.
Van Soest, P. J. 1965b. Comparison of Two Different Equations For Prediction of Digestibility from Cell Contents, Cell Wall Constituents and Lignin Content of Acid Detergent Fiber. J. Dairy Sci. 48:815.
Van Soest, P. J. 1966. Nonnutritive Residues. A System of Analysis for the Replacement of Crude Fiber. J. A.O.A.C. 49:546.
Van Soest, P. J. 1967. Development of a Comprehensive System of Feed Analysis and its Application to Forages. J. Anim. Sci. 26:119.
Van Soest, P. J. 1970. The Chemical Basis for the Nutritive Evaluation of Forages. Proc. Natl. Conf. on Forage Eval. and Util. Lincoln, Nebraska.
Van Soest, P. J. 1973. The Value of Laboratory Tests in the Estimation of the Productive Value of Feedstuffs. Proc. Wisconsin Conf. on Use of Lab. Anal, in Feeding. Univ. of Wisconsin. Madison, Wisconsin.
Van Soest, P. J. 1975. Physico-chemical Aspects of Fibre Digestion. In: Digestion and Metabolism in the Ruminant. J. W. McDonald (Ed). Univ. New England Pub. Armidale, Australia, p. 351.
103
Van Soest, P. J. and L. A. Moore. 1965. New Chemical Methods of Analysis of Forages for the Purpose of Predicting Nutritive Value.Proc. 9th Int. Grassl. Congr. Brazil, p.783.
Van Soest, P. J., R. H. Wine and L.A. Moore. 1966. Estimation of the True Digestibility of Forages by the In Vitro Digestion of Cell Wall. Proc. 10th Int. Grassl. Congr. Helsinki, Finland, p.438.
Vander Noot, G. W, R. H. Cordts and R. Hunt. 1965. Comparative Nutrient Digestibility of Silages by Cattle and Sheep. J. Anim. Sci. 24:47.
Warner, A. C. I. 1956. Criteria for Establishing the Validity of In Vitro Studies with Rumen Microorganisms in So-called Artificial Rumen Systems. J. Gen. Microbiol. 14:733.
Weston, R. H. 1971. Factors Limiting the Intake of Feed by Sheep. V. Feed Intake and the Productive Performance of the Ruminant Lamb in Relation to the Quantity of Crude Protein Digested in theIntestines. Australian J. Agr. Res. 22:307.
White, T. W., P. E. Schilling, F. G. Hembry and W. L. Reynolds. 1974. Comparison of Various Length Collection Periods in Digestion Trials With Steers Fed Various Levels of Roughage. J. Anim. Sci. 38:839.
Wilson, J. R. And D. J. Minson. 1980. Prospects for Improving the Digestibility and Intake of Tropical Grasses. Trop. Grassl. 14:253.
Wilson, R. K., T. A. Spillone and M. J. Clancy. 1966. The Influence of Fiber Content on Herbage Intake by Ruminants. Irish J. Agr. Res. 5:143.
Wilson, R. K. and R. B. McCarrick. 1966. Apparent Dry Matter Digestibility, Voluntary Food Intake and Yields of Dry Matter of Mixed Swards, Conserved as Artificially Dried Grass and Tetrapod Hay, at Progressive Stages of Maturity. Proc. 10th Int. Grassl. Congr. Helsinki, Finland, p.371.
Wurster, M. J., L. D. Kamastra and J. G. Ross. 1971. Evaluation of Cool-season Grass Species and Varieties Using In Vivo and In Vitro Techniques. Agron. J. 63:241.
Yu, Y. and J. W.Thomas. 1973. Estimation of Forage Nutritive Value. J. Anim. Sci. 37:361.
APPENDIX TABLES
table 1. COMPOSITION OF "KANSAS STATE BUFFER" FOR IN VITRO DRY MATTER DISAPPEARANCE ANALYSIS8
Solution A (g/i)
KH2PO4 10.0
Mg2^04.7H2O 0.5
NaCl 0.5
CaCl2.2H20 0.1
Solution B (g/lOOml)
Na2C03 15.0
Na 9S.9H 9O 1.0aJust prior to use add 20 ml of solution B to each liter of solution A.
106
Table 2. COMPOSITION OF ARTIFICIAL RUMEN FLUID (ARF) BUFFER FOR IN VITRO DRY MATTER DISAPPEARANCE ANALYSIS
Constituent_____________ (g/1)
(NH4)2HP04 0.820
CaCl2.2H20 0.227
(nh4)2so4 0.250
MgS04.7H20 0.200NaHC03 13.600
KC1 0.660
CoC1.6H20 0.002Casein hydrolysate 1.000
Cysteine 0.500
Biotin 0.00005
PABA 0.0001Isobutyric acid 0.058
Valeric acid 0.113
Acetic acid 2.860
Urea 0.750
HC1 1.570
Table 3. COMPOSITION OP MCDOUGALL'S BUFFER FOR IN _______ VITRO DRY MATTER DISAPPEARANCE ANALYSIS
Constituent (g/1)NaHC03 9.80
Na2HP0.7H20 7.00
KC1 0.57
NaCl 0.47
MgS04.7H20 0.12CaCll 0.04
aAdded just before use
Table 4. COMPOSITION OF ACID BUFFER FOR
SODIUM ACETATE: ACETIC ENZYME ASSAYS8
At OC At 25CpH Ab A
3.6 702 650
3.8 460 428
4.0 309 288
4.2 213 2004.4 153 145
4.6 115 110®Ionic atrength“0.05A * ml 1 M acetic acid in 50 ml of 1 M NaOH
108
Table 5. DATE OF CUTTING AND STAGE OF REGROWTH OF THE HAYS STUDIED
Hay SourceDate of cutting
Stage of growth or weeks regrowth
Ryegrass Ben Hur Exp. Sta. LA 1977 Mature
Alfalfa Western Texas 7/1978 Early-Bloom
Common BG West LA Exp. Sta. LA 6/1979 5 Weeks
Alicia BG Alexandria Exp. Sta. LA 1978 4-5 Weeks
Fens, bahiagrass West LA Exp. Sta. LA 7/1979 6 Weeks
109
Table 6. ANALYSIS OF VARIANCE FOR DIGESTIBLE ENERGY
Source d.f. Sum of Squares F
Hay 5 1.3627 19.97
Period 5 0.1048 1.54
Steer 5 0.1149 1.68Error 20 0.2730■jlgjHfSignificant at (P<.01)
Table 7. ANALYSIS OF VARIANCE FOR DRY MATTER DIGESTIBILITY
Source d.f. Sum of Squares F
Hay 5 512.1792 14.21**
Period 5 77.6425 2.15
Steer 5 50.9658 1.41
Error 20 144.1400
**Significant at (P<.01)
Table 8. ANALYSIS OF VARIANCE FOR DRY MATTER INTAKE
Source d.f. Sum of Squares F
Hay 5 7758.3992 **33.37
Period 5 263.9925 1.14
Steer 5 277.8091 1.19
Error 20 929.9067
Significant at (P<.01)
110
Table 9. ANALYSIS OF VARIANCE FOR DILUTION OF RUMINALFLUID WITH ALFALFA HAY AS THE SUBSTRATE
Source d.f. Sum of squares F
Rep 1 879.9377 **437.26
Treat 4 149.6526 **18.59
Rep x treat 4 14.1468 1.76
Error 185 372.2954
Significant at (P<.01)
Table 10. ANALYSIS OF VARIANCE FOR FURTHER DILUTION OF RUMINALFLUID WITH ALFALFA HAY AS THE SUBSTRATE
Source d.f. Sum of squares F
Rep 1 232.3893 232.87**
Treat 4 52.0949**
13.05
Rep x treat 4 11.4554 2.87
Error 170 169.6501** . Significant at (P<.01)
Ill
Table 11. EFFECT OF DILUTION RATE OF RUMINAL FLUID AND LENGTH OF FERMENTATION TIME ON IVDMD
Incubation Dilution (ruminal fluid:water’.buffer)time (hr)_______2:8:40 5:5740 10:0:40 17:0:33 25:0:25
Alfalfa hay24 61.4 63.0 62.5 66.4 58.648 67.6 67.7 65.2 66.3 62.696 68.0 68.0 66.5 65.1 64.2
Ryegrass hay24 29.7 33.5 31.5 37.7 32.848 41.7 44.7 42.5 43.4 41.396 53.2 54.8 47.2 53.6 48.1
50:50 Rye + Alf hay24 46.7 46.3 45.8 48.3 46.248 59.8 56.2 53.3 58.1 52.396 56.9 57.2 56.6 60.0 57.1
Common Bermudagrass hay24 36.8 39.7 39.4 40.9 36.848 53.7 54.0 50.6 55.7 52.196 58.3 62.4 58.9 59.5 59.4
Alicia Bermudagrass hay24 36.2 41.3 45.9 50.3 43.048 52.6 53.0 50.2 57.8 48.696 61.8 60.0 59.2 62.2 55.7
Bahiagrass hay24 31.0 35.5 35.2 38.9 38.748 50.0 53.0 47.4 51.4 52.796 57.9 61.8 61.1 58.1 52.9
112
Table 12. ANALYSIS OF VARIANCE FOR EFFECT OF DILUTION RATE OF RUMINAL FLUID AND LENGTH OF FERMENTATION TIME ON IVDMD OF SEVERAL HAYS
Source DF SS F
Rep 1 105.2793**
7.65
Dilution 4 354.7956 **6.44
Forage 5 8293.7258A*
120.49
Dilution x forage 20 189.7554 0.69
Error a 29 399.2276
Time 2 7477.1934 809.20**
Dilution x time 8 232.0906 **6.28
Forage x time 10 1278.0913 **27.66
Dilution x forage x time 40 299.6176 *1.62
Error b 60 277.2067
^Significant at (P<.05) Significant at (P<.01)
Table 13. EFFECT OF HAY ON IVDMD
HayNumber
of samplesMean
IVDMD (%)
Alfalfa 30 64.90a
Ryegrass 30 42.36®
50:50 rye+alf 30 57.38b
Common B6 30 50.58C
Alicia BG 30 51.92°
Bahlagrass 30 48.33 da,b,c,d,e 4' Means with different superscripts aredifferent (P<.05), Duncan's MRT
114
Table 14. EFFECT OF INOCULUM DONOR, DILUTION RATE OF RUMINAL FLUID AND LENGTH OF FERMENTATION TIME ON IVDMD OF ALFALFA HAY
Incubation Dilution (ruminal fluid:water;buffer)time (hr) 2:8:40 5:5:40 10:0:40 17:0:33 25:0:25
Holstein Steer24 61.6 64.7 63.2 66.9 67.848 67.9 69.7 66.9 72.1 71.096 71.1 70.9 67.8 76.5 71.5
Hereford Steer24 63.8 67.7 62.0 61.4 53.848 67.3 70.8 64.1 67.3 64.796 72.1 71.1 67.2 70.9 65.3
Holstein Cow I24 66.5 65.5 66.3 63.5 61.748 70.5 69.1 68.3 64.5 64.096 71.2 69.7 69.4 70.3 69.9
Holstein Cow II24 66.2 67.0 62.0 70.4 66.048 72.0 67.4 68.6 73.0 69.296 73.5 68.2 71.2 74.0 70.6
115
Table 15. EFFECT OF LENGTH OF FERMENTATION TIME ONIVDMD OF ALFALFA HAY
Fermentation Number Meanlength (hr) of samples IVDMD (X)
24 40 64.40
48 40 68.42b
96 40 70.62°a,b,c Means with (P<,05), Duncan's
different superscripts MRT
are different
Table 16. EFFECT OF RUMINAL FLUID OF ALFALFA HAY
DILUTION ON IVDMD
Ruminal fluid: water:buffer
Number of samples
Mean IVDMD (X)
2:8:40 24 68.64a
5:5:40 24 68.48a
10:0:40 24 66.42b
17:0:33 24 69.29a
25:0:25 24 66.2^* Means with different superscripts are different (P<.05), Duncan's MRT
116
Table 17. ANALYSIS OF VARIANCE FOR IN VITRO RUMINAL PROCEDURES
Source d.f. Stun of squares F
Procedure 4 1062.1133A*
35.47
Hay 5 2726.7872**
72.85
Proc. x hay 20 149.7133
Error 60 372.2954**
Significant at (P<.01)
Table 18. ANALYSIS OF VARIANCE FOR IN VITROENZYME PROCEDURES
Source d.f. Sum of squares F
Procedure 5 2920.2871**
79.60
Hay 5 5697.6705**
155.31
Proc. x hay 25 183.4279
Error 71 89.5950
Significant at (P<.01)
117
Table 19. COMPARISON OF LABORATORY VALUES AND NIR PREDICTED VALUES FOR VARIOUS QUALITY CONSTITUENTS OF COMMON BERMUDAGRASS
Sample1 2 3 4
Constituent LAB NIR LAB NIR LAB NIR LAB NIR
DMI (g/W-g) 56.1 63.7 66.3 60.8 81.9 61.3 59.3 60.3
DMD (Z) 56.0 58.3 54.4 55.9 68.7 57.2 57.5 55.0
IVDMD (Z) 55.6 59.6 54.6 58.7 56.3 58.5 58.2 55.2
CP (Z) 14.9 13.9 13.7 12.6 14.9 13.4 14.3 12.6CF (Z) 31.4 28.0 32.1 28.9 30.0 27.1 29.8 30.8
NDF (Z) 80.4 79.7 82.2 81.1 79.4 78.9 82.1 77.9
ADF (Z) 39.7 35.6 39.7 38.1 39.3 35.8 39.8 41.3
L (Z) 6.9 7.1 7.6 7.8 6.9 7.0 6.6 8.1
118
Table 20. PREDICTION EQUATIONS OF DRY MATTER INTAKE FROMLABORATORY ANALYSIS WITH SIMPLE REGRESSION
Equation Forage Type na r2 bb SEEC
DMI - -61.46+2.28 IVDMD gras. + leg. 6 .77* 2.28 0.61
DMI - -0.38+4.81 CP gras. + leg. 6 **.99 4.81 0.13
Mean o£ a 6x6 LSD Significant at (P<.05)
^Significant at (P<.01) “Regression Coefficient cStandard error of estimate
Table 21. PREDICTION EQUATIONS OF DRY MATTER INTAKE FROM LABORATORY ANALYSIS WITH STEPWISE REGRESSION
Forage type na Equation a b SEEbgras. + leg. 6 DMI - 2.52
5.71 CP .13-0.06 He .04
gras. + leg. 6 DMI - 272.69-0.28 C .03-1.95 He .01-16.93 L .12
gras. + leg. 6 DMI - 336.89-1.12 CP .20-0.35 C .01-2.39 He .07-20.94 L .71
gras. + leg. 6 DMI - 312.55-0.83 CP .00-0.31 NDF .00-2.24 C .00-19.56 He .00
1.43 L .00a Mean of a 6x6 LSD b Standard error of estimate
119
Table 22. PREDICTION EQUATIONS OF DRY HATTER DIGESTIBILITYFROM LABORATORY ANALYSIS WITH SIMPLE REGRESSION
Equation Forage Type na r2 b SEE?2DMD - 20.38 + 0.64 IVDMD gras. + leg. 6 .90** 0.64 0.10DMD - 84.81 - 0.70 ADF gras. + leg. 6 .71* -0.70 0.23
DMD - 40.56 + 1.10 CP gras. + leg. 6 .77* 1.10 0.29
aMean of a 6x6 LSD Significant at (P<.05) Significant at (P<.01) Standard error of estimate
Table 23. PREDICTION EQUATIONS OF DRY MATTER DIGESTIBILITYFROM LABORATORY ANALYSIS WITH STEPWISE REGRESSION
Forage type na Equation a b SEEbgras. + leg. 6 DMD - 63.08
0.75-0.42
CPADF
.19
.13gras. + leg. 6 DMD “ 63.91
0.73-0.43-0.01
CPADFHe
.32
.19
.08gras. + leg. 6 DMD - 105.23
-0.46-0.30-2.58
ADFHeL
.16
.091.01
gras. + leg. 6 DMD - 436.77-5.91-0.77-2.62
CPADFHe
3.60.221.42
-22.91 L 12.39gras. + leg. 6 DMD - 537.03
-7.34-0.97-3.30
CPCHe
1.370.100.55
-30.22 L 4.92gras. + leg. 6 DMD - 574.94
-7.770.53-1.60-4.80
CPNDFCHe
0.000.000.000.00
-33.34 L 0.00j* Mean of a 6x6 LSD Standard error of estimate
120
Table 24. PREDICTION EQUATIONS OF IVDMD FROM LABORATORY ANALYSIS WITH SIMPLE REGRESSION
Equation Forage type ns r* bb SEEC
IVDMD - 32.63 + 1.65 CP gras. + leg. 6 .79* 1.65 0.43
IVDMD - 99.68 - 1.07 ADF gras. + leg. 6 .72* -1.07 0.33
®Mean of a 6x6 LSD Significant at (P<.05)
bRegression coefficient cStandard error of estimate
Table 25. PREDICTION EQUATIONS OF IVDMD FROM LABORATORY __________ ANALYSIS WITH STEPWISE REGRESSION___________________
Forage type na Equation a______ b_______ SEE^gras 7 + leg. 5 IVDMD ■ 56.23
gras. + leg. 6 IVDMD » 53.57
gras. + leg. 6 IVDMD “ -168.77
gras. + leg. 6 IVDMD ■ -146.60
1.64 CP .14-0.74 CF .13
1.68 CP .13-0.81 CF .130.60 L .43
6.28 CP .49-2.16 CF .151.40 NDF .1513.10 L 1.34
5.82 CP 0.00-2.13 CF 0.001.25 NDF 0.000.12 C 0.0011.91 L 0.00
® Mean of a 6x6 LSD b Standard error of estimate
VITA
Miguel B. Coelho was born in Lisbon, Portugal on September 1,
1955. The author was reared in Lisbon, Portugal, where he graduated
from D. Joao de Castro High School in May 1972. He then attended
Lisbon Technical University, graduating in May 1978, with a Licence
(6 year program) in Agronomy and a minor in Animal Science. At the
present he is a candidate for the degree of Doctor of Philosophy in
the Department of Animal Science, Louisiana State University.
The author was married to Betty J. Neal of Albany, Louisiana on
February 20, 1982.
121
EXAMINATION AND THESIS REPORT
Candidate: Miguel B. Coelho
Major Field: Animal Science
Title of Thesis: Comparative Forage Evaluation Using Infrared Reflectance Spectroscopy, Microbial, Enzymatic and Chemical Analyses
Approved:
Major Professor and Ian
Dean of the Graduate School;
EXAMINING COMMITTEE:
<Juj /
eaJL
Date of Examination:
July 20, 1982