literature is exactly the length of the nine-night week, the term used being nómad, genitive feminine nómaide, '(the) ninth (night),' as in co cend nomaide, 'till the end of (a) ninth,' that is to say, to the end of the nine-night week. This is continued in Welsh with the incorrect substitution of day for night, for the favourite Welsh period is naw diwrnod, or nine days; as in fact it is in certain cases in English likewise, as when one speaks of 'the nine days' wonder.' From this point of view, the Germans are more correct with the space of acht tage, or eight days, to which they colloquially give a decided preference.
What, it may be asked in passing, should have led anybody to fix on a week of nine nights and eight days as a unit of time? It would be useless to demand an answer from the moon, and one should rather look at the fingers on one's hands: the half of a nine-night week would be the Irish noinden of five nights and four days; that is to say, a hand of nights, if you reckon the nights alone, as the ancient Celts must have done; and just as a third in music added to another third yields not a sixth but a fifth, so two hands of nights reduced to one sum make not ten nights but nine. But why the two hands should have been preferred as a unit to the single hand, I cannot say, though it may be guessed that the latter was too short a reckoning to be as useful as the longer one. The nine-night reckoning of eight days to the week could not, of course, be made in any way to coincide with the months as measured by the moon; but that cannot be urged as an objection. In fact, the more hopeless the discrepancy appeared, the more room it gave for the interference of the professional man, one of the
