Daniel Murphy: > No, they were still O(n) worst case, for a single bucket with a > degenerate binary tree. I see. I was unable to hit this degenerate case in my testing code, but I guess that was possible. Thank you. Bye, bearophile