This paper explores how ancient Mayan calendar specialists designed a predictive eclipse table, revising a century of interpretation. The table’s length, 405 months, was originally implemented in a general lunar calendar table of 405 successive months; within a few passes through this lunar calendar, intervals among observed eclipses could have stimulated an approximation to the series of lunar intervals that were later compiled as stations of an eclipse table; and for all and only these lunar stations to correspond to dates of upcoming eclipses, dates in successive 405-month eclipse tables would have to have overlapped. It identifies optimal procedures for the amounts of overlap that would maintain its predictions’ correctness and shows that these procedures could yield a sequence of tables that would anticipate every solar eclipse observable in the Mayan territory from a century or two after the first evidence of the Mayan lunar calendar to at least the era of the extant eclipse table, 700 years later.
5 February 2026 in Journal for the History of Astronomy
Many Mesoamerican hieroglyphic texts bear dates in indigenous calendars. Hundreds include dates in a lunation: the most ancient, in Zapotec texts, were observational; the most numerous, in Mayan texts, in a formal lunar calendar. Many Mayan lunar dates were projections into the past or future. Previous research has suggested evidence for specific computing ratios of days to months for such projections; that of 11,960 days to 405 lunations has long been accepted as the quintessential Mayan model for such calculations. This paper addresses observational evidence available to Mesoamerican calendar specialists for developing such models. It uses Lowry’s corpus of lunar first appearances at Monte Alban (650 to 0 BCE) to determine which sequences would have stood out as most reliable observationally, shows that a ratio of 1447 days to 49 months would have been by far the most reliable, and provides circumstantial evidence for its use in Mayan calendrics.