“…One method that has been widely used in the context of the L p -theory of elliptic operators relies on the exploitation of certain bounds, especially of Gaussian type, that hold for the kernels of associated integral operators, namely semigroup or resolvent operators ( [9], [1], [20], [7], [10], [13], [14]). If these operators are not integral operators, which is in general the case if p > 1 and q < ∞, then r-independence of the spectrum still holds under the assumption of certain weighted norm estimates ( [15]- [18]). It is the latter kind of assumptions, i.e., weighted norm estimates, that we shall use in this paper.…”