aNDF, NDFd, iNDF, ADL and kd: What have we learned?

Lignin is generally accepted as the primary entity responsible for limiting the
digestion of forages (Besle et al., 1994; Van Soest, 1994). Assuming that any estimation
based on long fermentations and made at any time other than infinity is an overestimate
of the true asymptotic indigestible residue, several attempts to predict iNDF from lignin
concentration have been made (Mertens, 1973; Chandler, 1980; Conrad et al., 1984;
Weiss et al., 1992; Traxler et al., 1998). As previously mentioned Chandler et al. (1980)
estimated the indigestible fraction as lignin times 2.4, after fermentation between 90 and
120 days in methane digesters. The Cornell Net Carbohydrate and Protein System uses
the 2.4 value as ratio between ADL and NDF to estimate iNDF in forages. Despite the
small size and origin of the database, the Chandler equation performed well in
prediction of iNDF in our observations (Van Soest et al., 2005), resulting in satisfactory
regression between predicted and observed (R 2 = 0.94).
Data from Huhtanen et al. (2006) shows a general relationship between
permanganate lignin and iNDF measured by 12 day in-situ fermentation, with an overall
slope of 2.4, but that relationship does not hold among all the values from Huhtanen et
al. (2006). This was also observed by Nousianen et al. (2004) who could not develop an
acceptable prediction equation (R 2 < 0.40) for iNDF based on lignin content on different
grass silage types. The previous findings from Nousianen et al. (2004) and Huhtanen et
al. (2006) refer to cold climate grasses that might result in different relationship between
lignin and iNDF due to environmental and agronomic interactions.
Our next objective was therefore to evaluate the consistency of the 2.4 ratio among
various forage families. Previous tests for nutritional uniformity indicated an average
recovery of 86% for ADL and sintered glass filters with a 40 µm aperture might not
achieve complete recovery of fine particles (Robertson, unpublished results; Udén,
2006). Thus, we evaluated ADL and iNDF recovery to assess the ratio between ADL
and NDF after improved recovery in order to estimate iNDF. Thirty forage samples of
various species and ADL content were analyzed for ADL content in Gooch crucibles of
porosity of 40 µm, with or without glass microfiber filters (1.5 µm; Whatman, 934-AH).
The same samples were also fermented in situ for 16 d using bags of PPT
monofilament fabric with porosity of 15 µm and an open area of 8.5% (Ankom
Technology, Fairport, NY). The bags were also inserted in the rumen of two fistulated
non-lactating cows. The same samples were also fermented for 16 d in the same bags
in a Daisy Ankom System where the medium and rumen fluid were renewed every 4 d.
All bags were analyzed for NDF after 16 d. Ratios between ADL and NDF, for
estimation of iNDF, were back calculated with the iNDF obtained after the
fermentations. Recovery of ADL varied among samples, but was generally higher using
the filter paper (Table 2). Improved recoveries for iNDF followed the ADL recoveries,
with higher recoveries for lower iNDF forages, when using the filter papers from the invitros
rather than the bags from the in-situ or from the Daisy. Long fermentations were
consistent within specie for both in-situ and in-vitro procedures. The combined improved
recoveries of ADL and iNDF resulted in the back-calculated ratio not being constant and
different than the 2.4 used so far. The observed ratios were in general always larger
than 2.4. Forages lower in ADL/NDF had larger ratios and vice-versa for forages with
larger ADL/NDF values. The data suggest the lignin procedure needs to be revised to