The strength of this random access identification key is the fast and strict elimination of possible species for any given set of characters and states, with the possibility of the computer suggesting the best separating characters. A drawback can be that creating a good matrix with taxa and characters/states demands that descriptions are consistent. Constructing such a matrix can be time consuming.

However, when the matrix is complete (checked and corrected), it is extremely powerful to support fast and secure identifications without the problems for the user of using structured keys. The following suggestion might be worthwhile.

The same data can be used to make separate files to build a hierarchical structure of identification matrices, e.g. a key to genera and separate files from genus to species. The files can be linked together. When this is done, the user is flexible in his choice to choose what file to use: the file containing all species or a file for one genus. Some practical remarks:

• A special character (space or *) in front of the name will place one file automatically on top of the list.
• States can be illustrated with pictures, sounds or video.
• It is important to realize that distribution characters (values of standard deviation) can not be exported into a NEXUS file.
• Extra spaces in the text must be avoided.
• Do not give the Identify file and your Linnaeus II application identical names. This can cause errors when opening the Identify file.