earlier use of grassmann variables?

The earlier articles I have found invoking grassman variables are
studies of supersimmetry in the mid-seventies.
Does anybody has a clue of earlier work?

Also, any idea of which is the reference in Grassman himself speaking
about such anticommuting variables? I hope it is not in the Rig Veda
dictionary.

I am looking particularly for such discussion on approacing Grassman
variables with some kind of idefinitely small lattice, so that
supersymmetry will exist in a approximate (broken) way in such lattice.

Alejandro

.