He gives it as his opinion that it would have been well to have adopted the vau, and says that neither by the old way of writing (by uo), nor by the modern way (by uu), is at all produced the sound which we perceive:
[Quint. I. vii. 26.] Nunc u gemina scribuntur (servus et cervus) ea ratione quam reddidi: neutro sane modo vox quam sentimus efficitur. Nec inutiliter Claudius Aeolicam illam ad hos usus litteram adjecerat.
And again still more distinctly:
[Id. ib. iv. 7, 8.] At grammatici saltem omnes in hanc descendent rerum tenuitatem, desintne aliquae nobis necessariae literarum, non cum Graeca scribimus (tum enim ab iisdem duas mutuamur) sed propriae, in Latinis, ut in his seruus et uulgus Aeolicum digammon desideratur.
This need of a new symbol, recognized by authorities like Cicero and Quintilian, is not an insignificant point in the argument.
Marius Victorinus says that Cicero adds u (consonant) to the other five consonants that are understood to assimilate certain other consonants coming before them:
[Mar. Vict. I. iv. 64.] Sed propriae sunt cognatae (consonantes) quae simili figuratione oris dicuntur, ut est b, f, r, m, p, quibus Cicero adjicit u, non eam quae accipitur pro vocali, sed eam quae consonantis obtinet vicem, et interposita vocali fit ut aliae quoque consonantes.
He proceeds to illustrate with the proposition ob:
[Id. ib. 67.] Ob autem mutatur in cognatas easdem, ut offert, officit; et ommovet, ommutescit; et oppandit, opperitur; ovvertit, ovvius.
Let any one, keeping in mind the distinctness with which the Romans uttered doubled consonants, attempt to pronounce ovvius on the theory of consonant u like English (w) (!).