In his letter to Carpianus, their inventor clearly yet briefly describes the purpose of his canons, ten in number. The first contains a list of seventy-one places in which all the four Evangelists have a narrative, discourse, or saying in common: the second of 111 places in which the three Matthew, Mark, Luke agree: the third of twenty-two places common to Matthew, Luke, John: the fourth of twenty-six passages common to Matthew, Mark, John: the fifth of eighty-two places in which the two Matthew, Luke coincide: the sixth of forty-seven places wherein Matthew, Mark agree: the seventh of seven places common to Matthew and John: the eighth of fourteen places common to Luke and Mark: the ninth of twenty-one places in which Luke and John agree: the tenth of sixty-two passages of Matthew, twenty-one of Mark, seventy-one of Luke, and ninety-seven of John which have no parallels, but are peculiar to a single Evangelist. Under each of the 1165 so-named Ammonian sections, in its proper place in the margin of a manuscript, is put in coloured ink the number of that Eusebian canon to which it refers. On looking for that section in the proper table or canon, there will also be found the parallel place or places in the other Gospels, each indicated by its proper numeral, and so readily searched out. A single example will serve to explain our meaning. In the facsimile of the Cotton fragment (Plate v. No. 14), in the margin of the passage (John xv. 20) we see _ΡΛΘ_ over Γ, where ΡΑΘ (139) is the proper section of St. John, Γ (3) the number of the canon. On searching the third Eusebian table we read MT. [symbol], Λ. _νη_, ΙΩ. _ρλθ_ and thus we learn that the first clause of John xv. 20 is parallel in sense to the ninetieth ([symbol]) section of St. Matthew (x. 24), and to the fifty-eighth (_νη_) of St. Luke (vi. 40). The advantage of such a system of parallels to the exact study of the Gospels is too evident to need insisting on.