To the extent it is reasonable to use such estimates of within farm error later when within-farm replication does not exist is debatable, as variances of measurements on small trees may be an underestimate of the variances of measurements on
older larger trees, and spatial variances structures may not be consistent through time.