Woodet al[4] have criticized Kevan's [1]analyses of blueberry production in New Brunswick as related to insecticide applications. Their criticisms have elements of validity, but do not refute in general Kevan's conclusions. That is not to say Kevan's original analyses were not overly simple. However, we present here further arguments which lend strong credence to the essential validity of Kevan's position [1, 2] that blueberry production in New Brunswick has suffered since 1970 because of applications of fenitrothion directed at spruce‐budworm infestations in the forest, but which also have adversely affected native pollinating bees of the blueberry lands. The crucial feature of the criticism is Kevan's choice of the co‐variate, Nova Scotia, with which to compare New Brunswick. Woodet alobject to the choice of Nova Scotia because of the presence of positive trend in blueberry production. In order to clarify this objection it is necessary to point out that Kevan used the time period 1960‐75, wheras Woodet aluse 1951‐77. The trend coefficients and their significance levels for 1960‐75 for New Brunswick and Nova Scotia were —0.017 (p = .849) and 0.251 (p= .047), respectively. These estimates support Woodet al. However, Woodet algo on to argue that Kevan should have used Maine as the comparison; their correct logic up to this point is now us