14. Let (1';'1 " , . , 1';'p)' v = 1, .. . , n, be a sample from a p-variate distribution F wit. h mean d covari . d I Z(II) - ~II Y / zero an covanance matnx ..., an et ; - "-, ,_\cl , /' ; V"-I,=IC" for some sequence of constants cl ' c2" " . Then (ZI"), . . . , Z~"» tends in law to N(O, I) provided the c's satisfy the condition (50) of Chapter 7. [By the Cramer-Wold theorem [see for example Serfling (1980)], it is enough to prove that La;Z?) -+ N(O, a'Ia) for all 0= (0\, . .. , all) with La; = I, and this follows from Lemma 3 of Chapter 7.]