fOrstemann] corrections, or encircled NUMERALS 411. 



yet from the part remaining that which is destroyed may be restored 

 with tolerable certainty, as follows: 3,640, 5,460, 7,280 (so far with 

 the difference 1,820), 14,560, 21,840, 29,120, 36,400 (so far with the 

 difference 7,280=4X1,820), 72,800, 109,200, 145,600 (difference, 

 36,400=20X1,820). Below this highest number (1,600X91) stands, 

 written very small in red, crowded in between the figures of 1,820, 

 a large number, the Maya numerals of which, read from the top 

 downward, are 19, 0, 4, 4. I can understand this number only by 

 substituting a 3 for the first 4; then it signifies 136,864=1,504X91. 

 1 intend to return to this number farther on. 



The numbers in a series always relate to certain days, which are 

 usually designated below, and which stand at the same distances from 

 each other as the numbers. In our case, five days belong to each num- 

 ber, which are specified as follows : 



and so on. Whenever a difference divisible by 260 is reached the 

 same days recur invariably, to wit : 



III 2 



1 



11 



3 

 XIII 20 



The three days in the middle should be regarded as having a III, 

 like the upper one; but for the present we may leave them uncon- 

 sidered, for only III 2 and XIII 20 are of immediate importance. 



Furthermore, these last-named five days are, of course, the actual 

 zero point from which the series progresses. With respect to the 

 series see also Cyrus Thomas's Aids to the Study of the Maya Codices, 

 Washington, 1888, page 327. 



THE CORRECTIONS, OR ENCIRCLED NUMERALS 



As I have shown in my treatise mentioned in the beginning, all the 

 days are computed from IV 17 onward. Therefore, it is imiDossible 

 that the above-named days. III 2 and XIII 20, should be either equal 

 to zero or equal to a number divisible b}^ 260. Actually, the day IV 

 17 is always meant here. The days under the numbers, therefore, are 

 arbitrary and merely used provisionally to measure the distances 

 between the numbers by th(; distances between the days. If one 

 wculd find the number actually corresponding to a day, a correction 



