This knowledge base as been set up to exploit the benefits of using a computer based parallel key. As such, the grouping of characters follows a convention of considering those that are easy to observe, rather than any sort of taxonomic grouping. The major groupings of first, second .... and so on, are best applied in that order. As features within characters are entered, each taxon is ranked in order of match guiding the process of identification to taxon level. Other references will always be required for confirmation.
When using the key, note that identification most improved by selecting features in a range of characters, rather than multiple features within the same character. This is because the individual features are treated as logical "OR", but the characters are treated as logical "AND". Clever fuzziness in the method of ranking gives tolerance to errors in the selection of features.