We describe the structure of distributions supported on conical surfaces and calculate the Fourier transform for some cones. The results are represented as convolutions with particular kernels. We use transmutation operators owing to which it is possible to clarify connections between the change of variables for a distribution and its Fourier transform. Bibliography: 17 titles.A lot of examples of distributions supported on various type surfaces in m-dimensional spaces can be found in [1,2]. However, in the literature, there are no results concerning distributions in the general form (counterparts of the Schwartz theorem on the general form of a distribution supported at a point in the one-dimensional case [3]). This paper is motivated by the recent results of [4]-[8], where pseudodifferential equations are studied in domains with conical points on the boundary in the multidimensional (m 3) case.
Distributions and Change of Variables1.1. Choice of test functions. Let C be an acute convex cone in R m containing no entire line. Assume that the conical surface is given by the equation x m = ϕ(x ), x = (x 1 , . . . , x m−1 ), where ϕ : R m−1 → R is a smooth function on R m−1 \ {0} such that ϕ(0) = 0. We introduce the change of variables tand denote by T ϕ : R m → R m the change operator. It is obvious that this transformation is smooth except for the origin. We introduce the change of variables for the following class of distributions. For the space of test functions we take the Lizorkin space Φ(R m ) [9] which is a subspace of the Schwartz space S(R m ) of infinitely differentiable functions that are rapidly decreasing at infinity and vanish at the origin, together with all its derivatives. If Φ (R m ) and S (R m ) denote the corresponding spaces of distributions, then Φ (R m ) ⊃ S (R m ) and all operations with distributions in Φ (R m ) are legitimate for distributions in S (R m ).