It is easy to see from your references of a Roman mile unit what John Michell used to construct them. The lesser value of 4838.4 Ft./Mi. would be what was intended based on 5000 units of the Greek Ft. of Samos, described at 126/125 ratio of the English Ft., reduced by ratio 24/25 to depict the Roman/Greco interconnection values to produce a Roman Ft. unit of 11.61216 ins. The greater unit mentioned would be this one extended by the 176/175 ratio to 11.678512 ins. That would produce a geodetic value of his canonical model of the Earth's Circ. of 24,883.2 English miles equivalent. Although I still regard his canonical model as being valid for some applications, whether or not it was still in use in Roman times is somewhat doubtful. Other values of it being 11.664 ins. and what I regard being more appropriate, that being in ratio to the English Ft. by 35/36 to 11.6666 ins. There existed a cyclic interconnection between the English Ft., the Egyptian and the Roman by use of comma factors that are still standard in musical intervals today. 12 ins. x 63/64 septimal comma for the Egyptian conversion I mentioned before, and further reduced to the Roman version by use of the Comma of Didymos of 80/81 to obtain the 11.6666 ins value. That may have been one route used by some other very astute metrologers who also discovered this value to obtain what I regard as the actual ancient Polar Meridian Circ. of the Earth, but there are many other ways of obtaining it also. As for what you mention in the last portion of your post of how eloquently John could express his ideas, that was his greatest art and asset of all I think.