This fact surprises many, but Hubble was actually quite critical of using red-shift as an indicator of cosmological distance. In 1953 (the year before his death) he even convinced Robert Millikan, 1923 recipient of the Nobel Prize for Physics and director of physics at the California Insitute of Technology, that it was probably wrong to interpret red-shift as an expanison of the universe.
Hubble ended his book _Observational Approach to Cosmology_ with the statement: "... if the recession factor is dropped, if red-shifts are not primarily velocity-shifts, the picture is simple and plausible. There is no evidence of expansion and no restriction of time-scale, no trace of spatial curvature, and no limitation of spatial dimensions. Moreover, there is no problem of inter-nebular material. The observable region is thoroughly homogeneous; it is too small a sample to indicate the nature of the universe at large. The univers might even be an expanding model, provided the rate of expansion, which pure theory does not specify, is inappreciable. For that matter, the universe might even be contracting."
Hubble ended his book _Observational Approach to Cosmology_ with the statement: "... if the recession factor is dropped, if red-shifts are not primarily velocity-shifts, the picture is simple and plausible. There is no evidence of expansion and no restriction of time-scale, no trace of spatial curvature, and no limitation of spatial dimensions. Moreover, there is no problem of inter-nebular material. The observable region is thoroughly homogeneous; it is too small a sample to indicate the nature of the universe at large. The univers might even be an expanding model, provided the rate of expansion, which pure theory does not specify, is inappreciable. For that matter, the universe might even be contracting."