“…He emphasised that, although the line was possibly curved for very low or very high populations of borers, it was straight for the range of populations that was of practical concern. Many other authors have calculated linear regressions of yield on pest numbers, injuries or injured plants (e.g., Bardner, Maskell & Ross, 1970;Buckley & Burkhardt, 1962;Deay et al, 1949;Gould & Graham, 1969;Kinzer & Henderson, 1968;Rautapaa, 1966Rautapaa, , 1968aRautapaa, ,6, 1969Rautapaa, , 1970Wilson et al, 1969). It is common practice to estimate the yield of unattacked crops by extrapolating the regression on yield to the point of zero infestation, but this can be seriously misleading if there is an unsuspected threshold level, or the upper part of the line is markedly curved.…”