This prose was fortunately a fairly musical thing by nature. In comparing the earliest Latin word-forms with those of the Gracchan days we find that they had improved very much in musical quality, due in part, no doubt, to the fact that the Etruscans and Sabines, who had temporarily dominated Rome, had slurred over harsh collocations of consonants till they fell away, and partly to the fact that the plebeians, who were of course less conservative in speech than the patricians, had won great positions in the fourth century. _Jouxmenta_ of the Stele inscription had now softened to _jumenta_; _stlis_ had become _lis_; _stlocus_, locus; _forctis_, fortis; _scandsla_, scala; and so on, in hundreds of words. In many positions the harsh sibilants had been eliminated: _cosmis_ had become comis; _dusmos_, dumus; and intervocalic _s_ had become _r_: _eram_ was better than _esam_. This elimination of harsh sounds had wrought so effectively between 500 B.C. and 100 B.C. that a language that was once as rough as Gothic had acquired the mellifluous quality of Italian. Though it still contained too many sibilants for ideal speech and the final _m_ occurred so frequently as to invite monotony, it had few sounds that could jar upon the most delicate ear. The vowels were relatively pure, and because of the abundance in inflections of the sonorous vowels _a_, _o_, and _u_ (=_oo_), they gave the language an orotund quality. The Indo-European _i_ is on the whole apt to be shrill, and the great vowel shift of sixteenth-century England which altered it to the much more musical _i_ (=_aye_) undid its benefit by raising English _e_ to the thinner sound of the old _i_. Latin retained the old sound, but in _i_-stems it frequently went over by analogy to _e_, and the _a̅i̅_, _a̅ĭ_ diphthongs fortunately softened to the mellow _æ_. In all this, mechanistic forces of the speech organs were at work, but one cannot help thinking that a delicate auditory guidance helped select the desirable sounds. Another great advantage inherent in the Latin language from the beginning was that quantities were carefully observed by it and were in fact the determining factor in its rhythm; and since time rather than stress is the guiding principle of music in human song, as in the flute and organ, Roman speech was to an unusual degree suited to modulated utterance. To be sure, in the century before Plautus, stress had threatened for a while to gain dominance in vulgar speech—enough in fact to question the rights of measured verse—, nevertheless the timely spread of the conservative, aristocratic pronunciation through political and forensic oratory, as it was heard almost daily in the open forum during the second century, gradually checked the process and standardized a precise observance of longs and shorts.