Of this curious work,[337] which was published in 1676, it will suffice to say here, it is a work intended not so much for the letter-cutter as for the sign-board and inscription painter. Taking the Van Dijk letters as his models, the writer attempts to demonstrate that each letter is a combination of geometrical figures, bearing regular proportions one to another; and by sub-division of the square of each letter into forty-two equal parts, he professes to be able to erect in any other square, similarly sub-divided, the same letter in precise proportion and harmony. This theory he illustrates by copper-plate figures of the various letters {183} of the Roman, Italic and Black Alphabets, and their sub-divisions. The result is not pleasing. The letters are stiff, and in some cases distorted; although this we believe to be the fault not so much of the theory itself as of the rules of proportion for the different parts of each letter predicated in the first instance. The book, as we have observed, is clearly not intended as a guide to punch-cutting. We regard it rather as an interesting attempt to reduce to precise mathematical rules a set of characters which never have and never will yield themselves entirely to such treatment.[338] At the conclusion of the section devoted to “the ordering of Inscriptions”, Moxon says (p. 11), “But of this and several other Observations of this Nature, I have written more at large in a book I intend to publish on the whole Art of Printing.” From this it is evident that, as early as 1676, his treatises on Typography, which formed the second volume of the _Mechanick Exercises_ and were published in 1683, were already written.