Identification of diatom frustules, a vital step during ecological analyses, is greatly influenced by the magnification at which samples are examined. The striking differences between Light Microscopy (LM) and Scanning Electron Microscopy (SEM) reveal inconsistencies in the identification of diatoms at low magnifications. These inconsistencies hinder comparative analyses between data sets collected by several research centers and may be hiding information of potential value in the assessment of ecosystem health. Additionally, SEM research has had a marked impact on diatom taxonomy rendering traditional identification methods insufficient for recognition of newly created taxa. Many diagnostic features of small representatives of these newly erected taxa can not be recognized even at higher LM magnifications. The consequent lack of taxonomic resolution at the LM level may be leading to overestimation of geographical distributions, ranges of tolerance to environmental parameters, and optimal conditions for growth. Qualitative aspects of the differences between LM and SEM approaches are discussed in the present paper.