Remember that for a function to be diff. at there exists a linear transformation such that as . Take then and since we have that this goes to as goes to . This proves that is diff. at and . When we can only bound by , but I can't think of a counter-example at the moment for this case.