I assume that relative ratios between different-length...

I assume that relative ratios between different-length measures are to some significant degree appreciable in and of themselves.