Everybody knows that one of the ancestors of the sudoku is the Latin Square.
The Latin Square consists of inserting in a different sizes square scheme the numbers without repeating them in line and column.
Less known, but certainly more interesting, it is the Greek-Latin Square.
In the Greek-Latin Square we need to insert in every cell two elements, a letter and a number without repeating every letter and every number in line and column.
Besides every couple of letter and number (example A1) must not be repeated in the scheme.
Anciently instead of the number a letter of the Greek alphabet was used, from which the name.
The game has attracted the interest of the mathematicians for some properties and particularity.
The great mathematician Eulero, had proposed for example the following problem, known note as "problem of Eulero": to build a Greek-Latin square of size 6.
Many tried to solve without succeeding, only to the beginning of the '900 it have been shown that cannot exist a Greek-Latin square of size 6, and later it has been shown that the Greek-Latin squares can be only of odd size or of multiple size of 4.
Returning to us, the game is very similar to the sudoku, that is to fill the scheme departing from some elements already inserted.
Already from size 4 we can have grids not so easy at all.
Those of size 5 are decidedly interesting, those of size 6 cannot exist, those of size 7 uselessly seem to me mastodontic and few amusing, besides requiring too much time to the computer to be produced.
We have focused therefore on the schemes of sizes 4 and 5.
For those of size 4, many of the grids are very simple, we propose one that is not so easy and that can be solved through "checked hypothesis."

B3 can be only in R2C1 or in R1C3.
If B3 in R2C1 then D3 di R3C4 and 3 not more possible in row 1.
Then B3 in R1C3.