The old Latin alphabet was taken from the Dorian alphabet of Cumae, a colony from Chaleis, and consisted of twenty-one letters, A B C D E F Z H I K L M N O P Q R S T V X, to which the original added three more, O or [Symbol] (_th_), [Symbol] (_ph_), and [Symbol] (_ch_). These were retained in Latin as numerals though not as letters, [Symbol] in the form of C=100, [Symbol] or M as 1000, and [Symbol] or L as 50. Of these letters Z fell out of use at an early period, its power being expressed by S (_Saguntum = Zákunthos_) or SS (_massa = máza_). Its rejection was followed by the introduction, of G. Plutarch ascribes this change to Sp. Carvilius about 231 B.C., but it is found on inscriptions nearly fifty years earlier. [4] In many words C was written for G down to a late period, _e.g._ CN. was the recognised abbreviation for _Gnaeus_. In Cicero's time Z was taken into use again as well as the Greek Y, and the Greek combinations TH, PH, CH, chiefly for purposes of transliteration. The Emperor Claudius introduced three fresh symbols, two of which appear more or less frequently on monuments of his time. They are [Symbol] or [Symbol], the inverted digamma, intended to represent the consonantal V: [Symbol], or anti-sigma, to represent the Greek _psi_, and [Symbol] to represent the Greek _upsilon_ with the sound of the French _u_ or German _ü_. The second is not found in inscriptions. Other innovations were the doubling of vowels to denote length, a device employed by the Oscans and introduced at Rome by the poet Accius, though Quintilian [5] implies that it was known before his time, and the doubling of consonants which was adopted from, the Greek by Ennius. In Greek, however, such doubling generally, though not always, has a philological justification. [6] The pronounciation of Latin has recently been the subject of much discussion. It seems clear that the vowels did not differ greatly, if at all, from the same as pronounced by the modern Italians. The distinction between E and I, however, was less clearly marked, at least in the popular speech. Inscriptions and manuscripts afford abundant instances of their confusion. _Menerva leber magester_ are mentioned by Quintilian, [7] and the employment of _ei_ for the _i_ of the dat. pl. of nouns of the second declension and of _nobis vobis_, and of _e_ and _i_ indifferently for the acc. pl. of nouns of the third declension, attest the similarity of sound. That the spirant J was in all cases pronounced as Y there is scarcely room for doubt. The pronunciation of V is still undetermined, though there is a great preponderance of evidence in favour of the W sound having been the original one. After the first century A.D. this semi-vowel began to develop into the labiodental consonant _v_, the intermediate stage being a labial _v_, such as one may often hear in South Germany at the present day, and which to ordinary ears would seem undistinguishable from _w_.