However this may be, the most singular feature of this fragment remains to be mentioned, viz. the agreement between the musical notes and the _accentuation_ of the words. We know from the grammarians that an acute accent signified that the vowel was sounded with a rise in the pitch of the voice, and that a circumflex denoted a rise followed on the same syllable by a lower note--every such rise and fall being quite independent both of syllabic quantity and of stress or _ictus_. Thus in ordinary speech the accents formed a species of melody,--[Greek: logôdes ti melos], as it is called by Aristoxenus[1]. When words were _sung_ this 'spoken melody' was no longer heard, being superseded by the melody proper. Dionysius of Halicarnassus is at pains to explain (_De Comp. Verb._, c. 11), that the melody to which words are set does not usually follow or resemble the quasi-melody of the accents, _e.g._ in the following words of a chorus in the _Orestes_ of Euripides (ll. 140-142):-- [Greek: siga siga leukon ichnos arbylês tithete, mê ktypeite; apoprobat' ekeis' apopro moi koitas,] [Footnote 1: _Harm._ p. 18 Meib. [Greek: legetai gar dê kai logôdes ti melos, to synkeimenon ek tôn prosôdiôn, to en tois onomasi; physikon gar to epiteinein kai anienai en tô dialegesthai].] he notices that the melody differs in several points from the spoken accents: (1) the three first words are all on the same note, in spite of the accents; (2) the last syllable of [Greek: arbylês] is as high as the second, though that is the only accented syllable: (3) the first syllable of [Greek: tithete] is lower than the two others, instead of being higher: (4) the circumflex of [Greek: ktypeite] is lost ([Greek: êphanistai]), because the word is all on the same pitch; (5) the fourth syllable of [Greek: apoprobate] is higher in pitch, instead of the third. In Mr. Ramsay's inscription, however, the music follows the accents as closely as possible. Every acute accent coincides with a rise of pitch, except in [Greek: hoson], which begins the melody, and in [Greek: esti], for which we should perhaps read the orthotone [Greek: esti]. Of the four instances of the circumflex accent three exhibit the two notes and the falling pitch which we expect. The interval is either a major or a minor Third. In the other case ([Greek: zês) the next note is a Third lower: but it does not seem to belong to the circumflexed syllable. All this cannot be accidental. It leads us to the conclusion that the musical notes represent a kind of recitative, or imitation of spoken words, rather than a melody in the proper sense of the term.