P. 184.—Alphabets (_abgitoria_ or _elementa_): Tír. 308₁₃, 320₂₈, 322₁₅, 326₃₀, 327₂₀, 328₂₈. The suggestion that these were figurative alphabets, “the A B C of the Christian doctrine” (tentatively put forward by Stokes, _Vit. Trip._ p. cliii.), can hardly be entertained, unless it is shown that _abgitir_ was used in this sense without the addition of _crabaith_ (= _fidei_). Tírechán saw a psalter written by Patrick for Sachall—doubtless at Baslick (301₈). Justus, a deacon of Patrick, is said to have possessed a baptismal liturgy which Patrick gave him, _libros baptismatis_ (_ib._ 318). He gave Mucneus _libros legis septem_, which seems to mean the Heptateuch? _ib._ 326₁₇. Tírechán has a story that in a district of Connaught Patrick and his companions had tablets, written _more Mosaico_, in their hands, and that pagans took them for swords, saying that they seemed to be wooden in the day-time, but were really of iron for shedding blood. It is inferred that these tablets were wooden staves, roughly resembling in shape the short swords of the ancient Irish: compare Graves, _Hermathena_, iii. 237 and 228 _sqq._, and Todd, _St. Patrick_, 509. [It may be noted that in this passage of Tírechán _cum uiii aut uiiii uiris_ points to the use of a written source; the numeral was not clear in the MS.] P. 185.—The ogam alphabet. A word must be said on its structure, as it bears upon my contention that Latin letters must have been used in Ireland before the fifth century. The twenty-one ogam scores form four groups of five, with one letter (p) as a supernumerary. The vowels form one group; the other groups are (2) h d t c q; (3) b l v s n; (4) m g ng f r. It has been suggested with great probability that the elements of the second group were selected as being the initials of the first five Irish numerals; but the principle of arrangement in groups 3 and 4 has not been discovered. The choice of _p_ as the letter to be excluded from the groups and treated separately must have depended on that principle. The supposition that the inventor of the cipher contemplated only twenty symbols, and that the symbol for _p_ was subsequently added when found indispensable, is in itself unlikely (for why should he have fixed the number twenty?), and breaks down on the fact that, while _ng_ might have been treated by a shift, _p_ was required for the archaic word _poi_, which occurs on several of the extant sepulchral inscriptions.[350] Including _p_, and excluding _ng_, which was clearly a native invention, we have simply a cipher of the Latin alphabet up to _u_, with a single modification. The symbol _u_, which did double duty in Latin as both consonant and vowel, is represented by two symbols, one for each function. The fact that only the last three letters (x y z) of the Latin alphabet are discarded, and that all the others are represented, renders this explanation of the ogam cipher very much simpler than a derivation