XXVIII. On the theory of the analytical forms called trees