A remark on the Hadamard variational formula, II