If there is a value that is outside the main diagonal is very high (in absolute value), some variants are associated if most of them will be close to zero (R) PCA Cellular. That is not really useful.
If the value is outside the main diagonal are much higher (in absolute value) some variables most correlated If these values are close to zero (R) PCA Cellular useless really.
If the address is outside in the main diagonal are very high (in absolute value) of some variables correlated if most of these values will be close to zero. R) PCA Cellular useless thing!