92 Hiitton Webster 



Our English " fortnight" (O. E. feozuertyne iiiht) preserves the 

 memory of a similar practice. 



Peoples sufficiently developed in culture to require weekly 

 cycles being already familiar with the length of the moon's syno- 

 dic revolution,^" it is reasonable to assume that in most, perhaps 

 all, cases civil weeks of 5, 6, 7, 8, 9, and 10 days would arise as 

 divisions, either of the true lunation or of the conventional month. 

 The origin and early history of the week, remain, however, in- 

 volved in much obscurity. W. H. Roscher, whose extensive re- 

 searches have greatly contributed to our knowledge of the sub- 

 ject, argues that the sidereal month (averaging 2'/d. yh. 43m. 

 11.55s.) and the "light" month or period of the visible moon 

 (assuming this at 28d.), have furnished the basis for the 9-day 

 and 7-day divisions ; the synodic month the basis for the division 

 into 3 decades. ^^ This ingenious explanation is open to several 

 objections. The calculation of the sidereal month implies an 

 extent of astronomical knowledge which can scarcely be pred- 

 icated of the peoples who have employed 9-day divisions. 

 Again, the length of the " light month " is a highly variable 

 quantity which cannot be said even to average 28 days. As a 

 matter of fact months of 27 and 28 days which we should ex- 

 pect to find associated with the 9- and 7-day divisions are very 

 rare. By far the commoner arrangement is the calendarizing 

 of the lunation by alternate periods of 29 and 30 days for the 

 reasons mentioned above. Still another objection to Roscher's 

 theories requires more extended discussion. 



Since the lunar month begins with the new moon it follows 



each part is placed in large letters the word ATENOUX, indicating the 

 night of the full moon, "great night" (Hastings, op. cit., iii. 82). 



^ The synodic revolution of the moon is the time between two successive 

 conjunctions with the sun, and may be measured from new moon to new 

 moon or from full to full. It varies about 13 hours by reason of eccen- 

 tricities of the moon's orbit and of that of the earth about the sun, but 

 its mean value is 2gd. I2h. 44m. 3s. 



" " Die enneadischen und hebdomadischen Fristen und Wochen der 

 altesten Griechen," Abhandlungen der philologisch-hisforischen Klasse der 

 koniglich-sachsischcn Gcsellschaft der Wisscnschaften, 1903, xxi. no. 4, 

 pp. 5 sq., 14, 68 sq. 



92 



