ON THE ORIGIN OF WEEKS AND SABBATHS. 333 



days,* In Sofala (East Africa), according to De Faria, a civil 

 month of thirty days was adopted, and divided into three weeks 

 of ten days each. As, however, he says that the first day of the 

 first week was the festival of the new moon, there must be some 

 mistake. It looks more as if time were reckoned by lunar months, 

 and consequently, while the first and second week might each be 

 of ten days' duration, the third would be some hours short of 

 that,f The ancient Greeks had a civil month of thirty days, 

 divided into three weeks each of ten days. The Ahantas of the 

 western districts of the Gold Coast divide the lunar month into 

 three periods or weeks, the first and second of which are of ten 

 days' duration, while the third consists of the remainder of the 

 month. The first period, called Adae, is considered lucky ; the sec- 

 ond, called Ajain-fo, unlucky ; and the third, called Adim, neither 

 lucky nor unlucky. The Yorubas of the Slave Coast of West 

 Africa reckon by nights and moons, and have subdivided the 

 lunar month into six weeks of five days each, or rather, five of 

 them actually contain five days, and the remaining one, which 

 completes the month, about four days and a half. The Javanese 

 week, before the week of seven days was adopted from the Mo- 

 hammedans, consisted of five days, J 



The Siamese seem, like the Tshi and Ga tribes of West Africa, 

 to have divided the lunar months into four periods of seven 

 days and some odd hours, but, for convenience' sake, they have 

 now made the odd months contain twenty-nine nights and the 

 even months thirty. Their week is commonly said to consist 

 of seven days, but as it is contrived that their sabbath, called 

 Vamjira, should always fall on the fourth day, and, in the first 

 week of the month, should always be coincident with the fourth 

 night of the moon, it is evident that each week must be of seven 

 days and some hours' duration, or, if three of them are exactly 

 seven days long, then the fourth must complete the lunar month 

 and be eight days and a half long. In dates, the age of the 

 moon, either waxing or waning, is reckoned by evenings, and 

 hence the day of twenty-four hours is considered to begin at 

 sunset,* This, of course, must be the case with all peoples who 

 reckon by moons and nights ; and so enduring is custom that the 

 Italians and Bohemians still reckon the day of twenty-four hours 

 from sunset to sunset. 



When we tabulate our results, we get the following subdivi- 

 sions of months, lunar or civil : 



* Johnston, The River Congo, p. 455. 

 f Astley's Collection, vol. iii, p. 397. 



X Raffles, History of Java, vol. i, p. 4*75. 



* Bock, Temples and Elephants, appendix iii. 



