Dear John,
In cosmology, the speed of light is used as the denominator when you use comformal coordinates. For the sake of simplicity let us consider the case of euclidean sections (k=0) and the x coordinate only. Then, the Friedmann - Lemaître - Robertson - Walker metric reads ds^2=[R(t)]^2[c^2dt^2-dx^2]. Calling 1 the first galaxy and 2 the second galaxy, the condition of null geodesic for the light gives cdt=dx and, in turn, t1-t2=(x1-x2)/c. But this is NOT the proper time which one uses to compute the redshift. It is ONLY a coordinate time. The infinitesimal proper time dT is given by the root square of [R(t)]^2dt^2=[R(t)]^2dx^2/c^2, i.e. dT= R(t)dx/c. You must integrate this last quantity in order to compute the variation of proper time and, in turn, the redshift. In other words, you cannot merely divide for c in order to get the proper time. Instead, you must know the function R(t) and, in order to do this, you must solve Einstein Field Equation by inserting the Friedmann - Lemaître - Robertson - Walker metric components in such an equation. For galaxies which are not too much distant each other, one approximates R(T)= constant=R obtaining T1-T2=R(x1-x2)/c.
Cheers,
Ch.