HOLMES ANNIVERSARY VOLUME 



months were interpolated at hotun endings, it is equally clear that if 

 this same proportion of the 30-day months to the 29-day months had 

 obtained throughout the sequence, the lunar calendar would very 

 soon have run ahead of the actual lunar revolutions by nearly two days 

 a year. In other words, such an arrangement would have over-cor- 

 rected the calendar. Moreover, we have the evidence of the Dresden 

 Codex to the contrary, that there was no such preponderance of the 

 30-day months over the 29-day months, as indicated by the current 

 months in the hotun endings. Indeed we may safely conclude that 

 hotun endings were particularly favored times for the interpolation 

 of the extra 30-day months necessary to keep the lunar calendar in 

 accord with the true lunar periods, and furthermore that this two-to- 

 one predominance of 30-day months at these points does not extend 

 to other parts of the sequence. 



Acting on the writer's suggestion, Mr Carl Guthe, of the Division 

 of Anthropology at Harvard University, made a number of arbi- 

 trarily arranged sequences of the numbers 29 and 30 in order to deter- 

 mine if any arrangement could be devised that would agree with the 

 sequence of the lunar months as they are recorded : at Quirigua in the 

 fourteen successive hotun endings given in Table XI. Unfortunately 

 the results of this investigation were not entirely satisfactory, and 

 indeed it now seems probable that another important factor must be 

 recognized as having heavily influenced the sequence of the lunar 

 months among the ancient Maya. 1 



Professor Willson has pointed out that pages 51-58 of the Dresden 

 Codex are probably tables of possible solar eclipses, the ends of the 

 successive groups being times at which solar eclipses were particularly 

 likely to occur. A full presentation of his elaborate thesis, strongly 

 corroborated by the numbers actually recorded, would require too 

 much space here, 2 but it is evident that the total numbers of days 

 from the beginning of the series to the ends of the successive groups so 

 closely parallel the numbers of eclipse days on page 358 of Schram's 

 table of lunar phases, that these can hardly deal with any other phenom- 

 enon. The arrangement of the successive lunations there recorded into 

 groups of five or six lunations each can hardly be accidental, and we 

 must recognize the eclipse phenomenon as one of the determining 

 factors in the sequence of the lunar months among the ancient Maya. 



'In a recent letter to the writer, Mr Guthe states that he has worked out two sequences 

 which agree with the recorded sequences at Copan and Quirigua, the successful arrangements 

 being adaptations of the lunar sequences of the Dresden Codex. 



2 The writer earnestly hopes that Professor Willson may soon publish his important discover- 

 ies in the Dresden Codex, the astronomical significance of which points the way to the decipher- 

 ment of a large group of heretofore unknown glyphs. 



[394] 



