ERRORS ASSOCIATED WITH HERBAGE DISSECTION ANALYSES 1. A survey of errors in subsampling plots in the field

Abstract

Errors in field subsampling of pasture plots for herbage dissection are presented from nine different sites. There were marked site differences in error level. Two different basic methods of sainpling were employed at the sites included in the survey. Subsamples were either taken from a mown herbage sample or were cut to ground level with a motorized shearing machine from the uncut plot areas following mower cuts. Standard error was compatible with an appfoximate parabolic relationship with percentage herbage dissection values and was symmetrical about 50%. Percentage errors associated with herbage components constituting up to 5% of the sample, were unacceptably large.