A curious remark concerning the geometric transfer map (Q791206)

In this paper the author returns to a question he discussed in several previous papers [ibid. 49, 512-528 (1974; Zbl 0289.57014); 50, 33-40 (1975; Zbl 0299.58008); see also the following review]. The question is this: Let \(D^ r\!_ n\) denote the identity component of the group of \(C^ r\) compactly supported diffeomorphisms of \({\mathbb{R}}^ n\). In the above papers, the author proves that \(D^ r\!_ n\) is perfect, i.e. is equal to its own commutator subgroup, provided \(r\neq n+1\). The case \(r=n+1\) is still open. The purpose of this paper is to present an elementary example which shows where the proof breaks down. The difficulty is the impossibility of proving the surjectivity of a certain geometric transfer map in case \(r=n+1\), and this paper presents an example which shows that this map may indeed fail to be surjective.