Spectral Theories, IV, p. 331
Let n∈N. Let μn denote Lebesgue measure on Rn.
a) Let H1 be the image of the mapping ∂1:D(Rn)→D(Rn); it is the space of φ∈D(Rn) such that
∫Rφ(x,y)dμ1(x)=0
for every y∈Rn−1.
b) Let φ0∈D(R) whose integral over R is equal to 1. For every φ∈D(Rn), there exists φ1∈H1 such that
φ(x,y)=φ1(x,y)+φ0(x)∫Rφ(t,y)dμ1(t)
for every (x,y)∈R×Rn−1.
c) For every distribution f∈D′(Rn), there exists g∈D′(Rn) such that ∂1g=f. If g1 and g2 in D′(Rn) satisfy ∂1g1=∂1g2=f, then there exists a distribution h∈D′(Rn−1) such that
⟨g1−g2,φ⟩=⟨h,ψ⟩
for every φ∈D(Rn), where ψ∈D(Rn−1) is defined by
ψ(y)=∫Rφ(t,y)dμ1(t)
d) Let U be a connected open set of Rn. Let f∈D′(U) be a distribution such that ∂if=0 for 1⩽i⩽n. The distribution f is the distribution associated with a constant function.
Nothing.
Nothing.