Notes
Outline
New Forage Analyses
New Forage Analysis
Currently use fiber to estimate energy
relationship not adequate for high producing herds
relationship varies with level of intake
No estimate of bypass protein
Digestion of Two Forages Over time
Digestion of Two Forages Over time
Digestion of Two Forages Over time
Three approaches for predicting digestibility
Empirical equations
Summative equations
Rumen digestion kinetics
Empirical equations
simple linear regression equations
regressor: chemical constituent (ADF, NDF)
easy to measure
negatively correlated with digestibility
population specific
poor sensitivity
Comparison of  In vitro Digestibility and DDM predictions from 104 mixed forages
Summative equations
partition of feeds into fractions of similar nutritive characteristics (cell wall, cell contents)
cell contents (CP, NFC, Fat) ... 98% digestibility
cell wall (NDF) ... variable digestibility
endogenous excretions (MFE)
TDN: (Sum of digestible NDF, CP, NFC, Fat) - MFE
Summative equations
based on cause-effect relationships
sensitive to changes in nutrient concentration
not population specific
excessive number of assays
need to measure lignin
static model
no rates of digestion or passage
Rumen digestion kinetics
reproducing the reactions or environment occurring in the rumen
in situ -- in vitro
rumen disappearance: digestion + passage
rate of passage        intake        production level
Rumen digestion kinetics
more accurate and precise predictor of in vivo digestibility
direct measurement of the rate and extent of forage digestion in animals at different intakes
not practical for commercial forage testing laboratories (cost, time...)
Slide 13
Pearson correlation between chemical, empirical
and in situ digestion kinetic parameters for 50 samples
Slide 15
Conclusions
Predictions of digestibility by the empirical and in situ digestion kinetic approaches were not consistent (low correlations)
NIRS showed potential to accurately predict rumen digestion kinetics
as good as chemical values
A, C, ISDMD very accurate
B and kd lower accuracy and affected by particle size
In vivo study
Animals: 12 lactating dairy cows (6 cannulated)
Experimental design: 3x3 Latin square
Treatments:   L-ADF1  L-ADF2   H-ADF  (58:42)
Measurements:
digestibility, intake, rate of passage
milk production and composition
rumen physical and chemical characteristics
Slide 18
DM digestibility measured in vivo and predicted by four different approaches for three alfalfa hays
(predictions at 3 times maintenance intake).
Relationship between digestibility measured in vitro, predicted by empirical equations and in vitro-NIR equations
Conclusions
Validation statistics of in vitro-NIR equations:
equations for A, C, IVDMD are accurate
equations for B, kd lack accuracy (by NIRS)
all parameters were biased
slopes showed that trends were consistent with in vitro
Both empirical and in vitro-NIR equations are correlated with in vitro digestibility
Better results with in vitro-NIR equations
Digestibility at
24, 48, and 72 hours
Estimation of Rumen Undegraded Protein (RUP)
Better determine RUP of feedstuffs
Avoid overfeeding protein
Balance rations more cost effectively
Determine if soybean roasting adequate
Estimation of Rumen Undegraded Protein (RUP)
Pattern of rumen degradation determined over time
Each sample digested in 8 milking cows
Digestion for 24 hours
Residue corrected for microbial protein
Results predicted by NIR
Slide 25
Slide 26
Slide 27