Should only live diatoms be used in the bioassessment of small mountain streams?