The third paragraph of wikipedia's article on AVL trees says: "Because AVL trees are more rigidly balanced, they are faster than red-black trees for lookup-intensive applications."
So, shouldn't TreeMap be implemented using AVL trees instead of red-black trees(as there will be more look up intensive applictions for a hashing based data structure ) ?
Red-Black trees are more general purpose. They do relatively well on add, remove, and look-up but AVL trees have faster look-ups at the cost of slower add/remove. Java's general policy is to provide the best general purpose data structures. It's also the same reason Java's default Array.sort(Object[] a) implementation is stable,adaptive ,iterative merge sort instead of quicksort.