In another place (xix. 36) the question is answered by saying that the notes of a scale stand in a certain relation to the Mesê, which determines them with reference to it ([Greek: hê taxis hê hekastês êdê di' ekeinên]): so that the loss of the Mesê means the loss of the ground and unifying element of the scale ([Greek: arthentos tou aitiou tou hêrmosthai kai tou synechontos])[1]. These passages imply that in the scale known to Aristotle, viz. the octave _e - e_, the Mesê _a_ had the character of a Tonic or key-note. This must have been true _a fortiori_ of the older seven-stringed scale, in which the Mesê united the two conjunct tetrachords. It was quite in accordance with this state of things that the later enlargement completed the octaves from Mesê downwards and upwards, so that the scale consisted of two octaves of the form _a-a_. As to the question how the Tonic character of the Mesê was shown, in what parts of the melody it was necessarily heard, and the like, we can but guess. The statement of the _Problems_ is not repeated by any technical writer, and accordingly it does not appear that any rules on the subject had been arrived at. It is significant, perhaps, that the frequent use of the Mesê is spoken of as characteristic of _good_ melody ([Greek: panta ta chrêsta melê pollakis tê mesê chrêtai]), as though tonality were a merit rather than a necessity. Another passage of the _Problems_ has been thought to show that in Greek music the melody ended on the Hypatê. The words are these (_Probl._ xix. 33): [Greek: Dia ti euarmostoteron apo tou oxeos epi to bary ê apo tou] [Footnote 1: So in the Euclidean _Sectio Canonis_ the propositions which deal with the 'movable' notes, viz. Paranêtê and Lichanos (Theor. xvii) and Parhypatê and Tritê (Theor. xviii), begin by postulating the Mesê ([Greek: estô gar mesê ho B k.t.l.]).] [Greek: bareos epi to oxy; poteron hoti to apo tês archês ginetai archesthai? hê gar mesê kai hêgemôn oxytatê tou tetrachordou; to de ouk ap' archês all' apo teleutês.] 'Why is a descending scale more musical than an ascending one? Is it that in this order we begin with the beginning,--since the Mesê or leading note[1] is the highest of the tetrachord,--but with the reverse order we begin with the end?'