On the connexion of the physical sciences