528 BUREAU OF AMERICAN ETHNOLOGY [bull. 28 



time and the rate at which knowledge of the subject has in the mean- 

 time progressed emphatically demand a fresh exposition. 



The matter is the more important because a large part of the sur- 

 face of the manuscripts is covered with tonalamatls of this kind. To 

 be sure, in those sorry remains which we call Codex Peresianus I find 

 in one place only (page 17) a tonalamatl, of five parts, which seems 

 to begin with the day VII T. The Dresden codex, however, abounds 

 in such examples, since it contains in its first part (not in the second, 

 which is more astronomic) not fewer than about 70 of these tonal- 

 amatls. Their number can not be determined Avith perfect accu- 

 racy on account of the destruction of certain passages, the careless- 

 ness of the scribe, and other causes of uncertainty. Codex Troano- 

 Cortesianus, however, is richest in tonalamatls; all its parts indeed 

 abound in them. It presents not feAver than about -223 examples. 



In order that this matter nniy not be too difficult for the compre- 

 hension of the reader, I w411 here give examples, taken from Codex 

 Troano-Cortesianus, of the three kinds mentioned above : 



1. Codex Cortesianus, pages 10b to lib, tonalamatl of four parts : 



XIII IX 9 V 10 II 6 VIII 2 X 10 VII 5 XII 7 VI 7 XIII 

 19 



4 



9 

 14 



2. Cortesian codex, page 17a, tonalamatl of five parts: 



I 11 XII 12 XI 8 VI 13 VI 8 I 

 17 



9 



1 

 13 



5 



3. Troano codex, page 33b, tonalamatl of ten parts : 



IV 11 II G VIII 3 XI 6 IV 



8 18 

 14 4 

 20 10 



6 16 

 12 2 



The Roman numeral in the left-hand upper corner indicates the 

 Aveek day with which the tonalamatl begins; the Roman numerals at 

 the right of it indicate the week days with which the different parts 

 begin; the last week day (XIII, I, IV) must always be like the first, 

 as the number of days is always divisible by 13 without remainder. 

 The length of the different jjeriods is shown by the Arabic numerals, 

 and the sum of these must therefore be 05, 52, and 26. The vertical 

 row of numbers on the left gives the so-called month days, reckoned 

 from the day Kan. A^Tioever counts from Imix must set down 1, 2, 

 and 3, instead of 18, 19, and 20, respectively, and increase the other 

 numbers by 3. These days, in the three examples, are actually sepa- 



