It is a natural conclusion that the reckoning in weeks had its origin in the use of the rune-staff. Since the week-day letters on these are repeated the whole year through, the weeks offered an easy means of reckoning. This conclusion is certainly correct, but still we may venture to ask why the week-day letters were admitted into the national calendar by the North especially, and why the reckoning in weeks should be adopted in popular use only there. The reason can only be that the counting in weeks was already in use before the rune-staff was introduced. This mode of counting, which in Iceland had been developed into a curious form of year, was in Scandinavia adapted to the Julian calendar and remained bound up with this. The leap-week was therefore unnecessary. The old basis is however still preserved in the points of departure, the summer and winter nights. It is the same system as the Icelandic, built up on the week and the year, but differently modified: the idea of any borrowing cannot be entertained. The basis of this calendar, therefore, was once common to all Scandinavia, and the calendar must go back to heathen times. Under the influence of the popular lay astrology the week was early spread among the Germanic peoples: on it and on an approximate knowledge of the length of the year, such as could easily be acquired in the lively intercourse with Christian lands during the Viking period, the system of the Icelandic calendar is built up. An indigenous element however appears, the half-year reckoning, and indeed the great probability is that the limitation of the half-year to a fixed number of days was first achieved as a result of this systematising of the calendar. Winter and summer, like all natural seasons, had at first no fixed limits. The quarters arose in the course of the reckoning, the people counting forwards in the first half of the half-year and backwards in the other half. The middle points of the half-year, mid-winter and midsummer, fell where both reckonings met. This agrees with the popular objection to high numbers. The Germanic tribes of the south, in accordance with their milder climate, commonly reckoned five months for winter. In the north the dead season is longer, about six months, and this fact has contributed to the half-year reckoning which, as has already been remarked, is widely characteristic of northern peoples. That the limits between both seasons were unstable and could be moved forward according to circumstances is in my opinion shewn by the names of the initial days of the half-year--_sumarmál_ (plural) and _vetrnaetr_, ‘the winter nights’. Where a definitely determined day is in question the plural is out of place: it is used to describe a period, for instance _jol_ (plur.) denotes Christmas-time[338].