Counting cells with stereology: Random versus serial sectioning