On continuous cellularity (Q1090982)

Let \(M^ n\) be an n-manifold without boundary. A subset in M is cellular in M if it is the intersection of a sequence of topological n-cells \(D_ i\), where \(D_ i\subseteq\) int \(D_{i-1}\subseteq\) \(M^ n\). A proper map \(f: M\to Y\) is cellular if \(f^{-1}(y)\) is cellular in M for each \(y\in Y\). Let C(M) denote the space of all maps of M onto itself with the compact-open topology. Define \(G=\{(f(x),x):\) \(x\in M\}\), the graph of f. The central theorem in the paper is the following. If \(f: M\to Y\) is a cellular map, then the following conditions are equivalent: (1) dim Y\(<\infty\); (2) there exists a map \(\Phi\) : \(Y\to C(M)\) such that \(f^{- 1}(y)\) for each \(y\in Y\) is a unique nondegenerate fiber of \(\Phi\) ; (3) the projection p: (Y\(\times M)-G\to Y\) is a locally trivial bundle map. This result generalizes the work of R. D. Edwards and F. D. Ancel.