do, and placed the weekly market on the day with which it
began (-noundinae-). Along with this it once for all fixed the
first quarter in the months of 31 days on the seventh, in those
of 29 on the fifth day, and the full moon in the former on the
fifteenth, in the latter on the thirteenth day. As the course of
the months was thus permanently arranged, it was henceforth necessary
to proclaim only the number of days lying between the new moon and
the first quarter; thence the day of the newmoon received the name
of "proclamation-day" (-kalendae-). The first day of the second
section of the month, uniformly of 8 days, was--in conformity with
the Roman custom of reckoning, which included the -terminus ad
quem- --designated as "nine-day" (-nonae-). The day of the full
moon retained the old name of -idus- (perhaps "dividing-day").
The motive lying at the bottom of this strange remodelling of the
calendar seems chiefly to have been a belief in the salutary virtue
of odd numbers;(9) and while in general it is based on the oldest
form of the Greek year, its variations from that form distinctly
exhibit the influence of the doctrines of Pythagoras, which were
then paramount in Lower Italy, and which especially turned upon a
mystic view of numbers. But the consequence was that this Roman
calendar, clearly as it bears traces of the desire that it should
harmonize with the course both of sun and moon, in reality by
no means so corresponded with the lunar course as did at least on
the whole its Greek model, while, like the oldest Greek cycle, it
could only follow the solar seasons by means of frequent arbitrary
excisions, and did in all probability follow them but very imperfectly,
for it is scarcely likely that the calendar would be handled with
greater skill than was manifested in its original arrangement.
The retention moreover of the reckoning by months or--which is the
same thing--by years of ten months implies a tacit, but not to be
misunderstood, confession of the irregularity and untrustworthiness
of the oldest Roman solar year. This Roman calendar may be regarded,
at least in its essential features, as that generally current
among the Latins. When we consider how generally the beginning of
the year and the names of the months are liable to change, minor
variations in the numbering and designations are quite compatible
with the hypothesis of a common basis; and with such a calendar-system,
which practically was irrespe