So if the JIT is already able to optimize the method so well, why wouldn't you want to create the anonymous classes? It seems like that would enable the lambdas to automatically fit into the existing optimizations. llewellyn