don't think it is possible. since range is between -1 & +1. If it is phenotypic corrln, it should be intermediate to genetic and envtl corrln, as it is partitioned that way. If it is genetic corrln, (same genes affecting two traits in the same way) strength of relationship can maxm be one. I don't think any corrln can go above one - practically.but when u divide covariance by standard deviations, theoretically can it go above one ??- I think u should look at the standard error before coming to a conclusion.