The table consists of 12 columns (headed with the numbers 2 to 12 in sequence) and 19 rows (headed with the numbers 1 to 19 out of sequence). Each cell of the table contains one or two letters (lowercase Roman alphabet) and two numbers (left number from 5 to 10, right number from 0 to 6).
Above the table are two lines of abbreviated Latin, and there is one below. The first line suggests that the table was started in 1505 and finished in 1523 (I think - my Latin does not stretch to abbreviated or calligraphic forms too well).
The column headers (numbers 2 to 12) are simple to interpret, particularly since the author has provided us with pictures of two dice to the left of the table. One dice has the numbers 3, 4 and 6 showing, the other numbers 1, 2 and 5. Clearly, the sum of two dice is to be indexed on the columns, and that random letters/numbers are the result. It is worth noting that the sum of two dice produces a symmetrical distribution favouring the central numbers of the sequence-
Sum of Frequency
two dice (nearest 1%)
2 3%
3 6%
4 8%
5 11%
6 14%
7 17%
8 14%
9 11%
10 8%
11 6%
12 3%
Thus, if the table was used to generate random letters or numbers, the central values (occurring in the 6, 7 and 8 columns) would dominate the results. This could lead to the impression of structure within a random sequence.
The row headers (numbers 1 to 19) are much harder to interpret. Generating numbers within this range with dice is in no way simple, so the numbers almost certainly refer to something else. The sequence of the numbers is-
16, 9, 1, 14, 6, 2, 19, 8, 11, 5, 17, 3, 10, 7, 12, 18, 4, 13, 15
No two sequential numbers are next to each other (or, in fact, within two 'places' of each other ? a fact which eliminates a chance arrangement. Try writing the numbers 1 to 19 in a list such that no number is within two stops of a number one higher or one lower and you'll probably agree). It is possible that the author created the contents of the table prior to assigning the row headers, and sought to 'space out' the row designations as far as possible.
The letters within the table's cells are limited to just 11 patterns; a, ag, b, ba, d, dr, e, f, fe, g and r. Some rows contain many recurring letters (e.g. row 9 reads a-b-f-r-ba-r-e-e-e-e-r).
The pairs of numbers within the table seem to follow a vague set of rules or patterns. Columns 9 and 10 have identical numbers. It can also be seen that, wherever the pattern 10-1 occurs, the same pattern occurs immediately underneath it, 7-6 always has 5-5 below it, 8-0 always has 9-0 below it, etc.
The letters seem to match the numbers. For example, the letter 'a' is invariably seen with 9-0, the letter 'e' with 5-4 or 9-4, the letter 'd' and the pattern 'dr' with 8-3, etc.
The pairings of letters and numbers are thus:
a(9-0)
ag(8-0)
b(6-1)
b(10-1)
ba(10-1)
d(7-3)
d(8-3)
dr(8-3)
e(5-4)
e(9-4)
f(6-5)
fe(5-5)
g(7-6)
r(7-2)
It may be significant that the numbers in the letter pairs range from 0 to 10 (there being 11 letter forms found in the grid).
Anyway, so much for this afternoon,
Rob