Changing angle gauges in variable plot sampling: Is there a bias under ordinary conditions?

Changing the basal area factor in variable plot or angle count sampling to give tree counts in an acceptable range is biased. The very reasonable sounding assumption that changing basal area factors will cause an inverse and proportional change in tree count will "prove" that there is no bias, but this assumption is not likely to be true in actual forest conditions. Changing prisms should not be done unless the user has shown, by field tests, that the bias is within acceptable limits. Reducing the fieldwork by splitting the plot and measuring a random half is unbiased and practical. Zero-count plots should not be changed. Lower average tree counts are preferred to the bias caused by edge effect or by missing scattered large trees. Methods are suggested for verifying these effects locally with minimal effort to the cruising staff.