“…Then we wjould like to say that y M a (X -vP)I/p and z = b (yq _ wq)I/q for some a, b, v, w, p and q. But the implied overall striking-residual connection is then z = ab [(xp _ vP)q/p -(w/a)q]l/q which does not conform to (2). This lack of closure under composition, while not stri ly an impairment to the present concern of supplying effective representations of velocity relationships, does preclude full matheinaiical consistency in the consideration of multiple plate targets.…”