Alternative snippet sampling methods for rapidly estimating fibre diameter
profiles (FDPs) are investigated in this paper using sheep grazing in two e
nvironments. FDPs were measured using staples from 40 Merino ewes from Chis
wick in a temperate environment (n=20) and Bakers Hill in a Mediterranean e
nvironment (n=20). These ewes were the progeny of two sires within each of
two bloodlines grazed at the two environments. Technique 1 was a simple red
uced profiling technique, which generated profiles based on a subset of the
original snippets. The FDP characteristics of maximum fibre diameter (Max)
, minimum fibre diameter (Min), difference between the maximum and minimum
fibre diameter (Diff), fibre diameter variation along the profile (AstCV),
average fibre diameter of the FDP (Profmean) and position of the Max estima
ted by technique I were highly correlated with those calculated from the fu
ll profile (r > 0.96). The correlations for the positions of Min were high(
r> 0.83) while the correlations for the rates of fibre diameter change betw
een the Max and the Min were lower, ranging from 0.65 to 0.90. This techniq
ue did not allow for the shape of the FDP and as a result often calculated
a rate of change that was not significantly related to the original FDP.
A profile prediction technique was evaluated which utilised cubic spline fu
nctions to generate predicted profiles based on reduced profiles. This tech
nique utilised more FDP characteristics and allowed for the calculation of
two rates of fibre diameter change for each FDP. The actual fibre diameter
values of the predicted profiles were all not significantly different (P>0.
05) and highly correlated (r> 0.91) with the original full profile at all l
evels of snippet inclusion. The measures of absolute fibre diameter and alo
ng-staple variation in fibre diameter within the FDP were accurately estima
ted (r> 0.80) using only 1 in 10 snippets ton average approximately 13% of
original snippets). Using approximately 27% of the original snippets all FD
P characteristics can be estimated with acceptable accuracy (r > 0.80). Env
ironment, bloodline and sire had significant effects on most FDP characteri
stics. These differences did not adversely affect the accuracy of the reduc
ed profiling procedure.