An analysis of the fac-simile given in Mr. Humphreys’ work yields the following results. About 1430 separate types in the one page gives 2860 as the number required for two. The following figures (twice the number occurring in the specimen page) shew the numbers required of each of the letters most commonly used, a 44, e 122, i 182, o 146, u 74, d 44, h 28, m 60, n 100, s 84, t 82; there are, besides, the following duplicate and triplicate characters, of which no other printed work shews so large a number,--an, ca, cā, cc, ce, ch, ci, cī, co, cō, ct, cti, cu, cū, cp, cy, da, dā, de, dē, do, du, dū, eē, et, ect, fa, fā, fe, ff, fi, fl, fo, fr, fu, ga, ge, gē, gi, go, gu, gū, gp, gr, gy, ii, ib, in, la, le, lē, li, ll, lle, llz, œ, ori, no, nō, nu, pe, pp, ra, rā, re, rē, ri, ro, rō, ru, rū, ria, sa, se, si, so, ss, st, ssi, ssz, ste, ta, tā, te, tē, ti, to, tu, tū, tri,--varying in the frequency of their occurrence from twice to twenty-two times, leaving but 1082 other letters for the rest of the alphabet, including the capitals: and of these last from 6 to say 40 would be the utmost of each required. It is thus shewn, that out of the whole number of matrices, upwards of 300, which would be required for a complete fount, not more than _eight_ would be used up to or beyond the point where the fine strokes (supposing the matrices to have been of soft lead) would begun to thicken; and of these it would be a most easy matter to provide duplicates or triplicates, in order to preserve the uniformity of character aimed at by the first printers, in imitation of the manuscripts they intended their works to supersede. _All the letters thus cast, would moreover, be fac-similes of each other_, and would not, nay could not, present those dissimilarities of appearance observable in specimens of the same letter occurring in every line of the _Speculum_. Koning’s idea is thus proved to be erroneous.