532 BUREAU OF AMERICAN ETHNOLOGY [bull. 28 



amountino;, therefore, in the former to iihout one-third and in the 

 hitter to abont one-fonrth of all the cases, instead of only two- 

 thirteenths. I can further add that the day IV 17 in Codex Troano- 

 Cortesianiis stands at the beginning of the tonalamatl al)out 24 times. 

 Its importance is not so plainly shown in the Dresden codex on 

 account of the smallness of the number (I knoAv only of two cases) ; 

 IV 17, however, is the day from which computation of time begins. 

 Codex Troano-Cortesianus (41 instances) gives to IV even greater 

 prominence than to I or XIII. 



Apart from these points, the week days and month days in both 

 manuscripts are purely the result of chance and caprice. 



This being so, we arrive first at two negative results : , 



(1) The tonalamatls of the Maya manuscripts do not immediately 

 follow one another like months and years; else they w^ould all have 

 to begin wdth the same day, which would ahvays recur after 260 days. 



(2) Neither can they lia^e a fixed place in the year; else their first 

 days, eA'en on the supposition that intercalary days were inserted 

 after certain periods, would easilj^ be seen to follow a definite rule. 

 The year, or at least the exact date in the year, would also occasion- 

 ally be stated, but as yet I find no traces of this. 



I have a special reason for speaking of this second point, since the 

 distinguished and untiring worker in the field of Aztec research, INIrs 

 Zelia Nuttall, at the Americanist Congress held last year at Stock- 

 holm, presented her treatise On the Ancient Mexican Calendar Sys- 

 tem, in which with great ingenuity she advances the view that with 

 the Aztecs the tonalamatl as a special festal season occui:)ied the mid- 

 dle of every year of 364 days, which was preceded and followed by 

 four weeks. I do not deny that the Mayas had such a festal season, 

 but the tonalamatls of the manuscripts surely have nothing whatever 

 to do with it. 



After these negations let us .ask Avhat these tonalamatls really are. 

 I can only arrive at the following hypothesis, which may very soon be 

 superseded by a better one: The tonalamatls of the manuscript are 

 kinds of horoscopes which were cast by the priests for the purpose 

 of foretelling the future lives of persons, classes, or tribes, as well as 

 future political events or natural phenomena. They may have been 

 so employed because they approximate periods of pregnancy. Natu- 

 rally, they had constant reference to the mythologic personages, but 

 had no connection whatever Avith the established calendar. 



This hypothesis also explains the fact that such horoscopes were 

 occasionally cast, not for only 260 days, but for multiples of this 

 period. I believe I have fouiid five cases of this in the Dresden 

 codex. I give them here in a table showing in the first column the 

 place in the manuscript, in the second the distance of the month 



