http://www.collembola.org/doc/generagl.htm - Last updated on 2001.01.18 by Frans Janssens

Peter F. Bellinger (), Department of Biology, California State University, Northridge, CA 91330
Kenneth A. Christiansen, Department of Biology, Grinnell College, PO Box 805, Grinnell, IA 50112-0806

Preamble

Frans Janssens
1998.10.16

Genera

The scientific function of genera is the organization of information about species in such a way as to make it accessible to biologists and others.
This is their primary function, and where it conflicts with other functions the latter must give way. The primary function requires stability but not stasis.
When a classification reflecting phylogeny promotes stability, by giving a rational rather than arbitrary basis for classification, it is to be preferred to other arrangements. However, it is probably impossible and certainly undesirable to attempt a classification which reflects phylogeny perfectly. Any such attempt at the generic level would make almost all genera monotypic and thus of little use in classification.

With this in mind we believe that the functions of genera are to:

1. organize information about species in such a way as to make it accessible to biologists and others;
2. serve as a useful tool for the identification of species;
3. facilitate ecological and biogeographic analysis;
4. give an indication of most close evolutionary relationships among species.

If these functions are accepted, certain guidelines for the erection of genera follow.

1. Insofar as possible, genera should be separable on characteristics which are readily observed and recognized by non-taxonomists. This does not mean that the adaptive characters (Gisin's terminology) which are the most easily observed and the least useful phylogenetically, should be used exclusively. But erection of genera and recognition of genera do not necessarily involve the same features. It should be possible to base generic separation on non adaptive characters (ex.: the trunk chaetotaxy of Pseudosinella species) but to find associated characters which are more readily visible to use for identification, or, if this is impractical, to clarify the characters used in such a way as to make them accessible to biologists in general.
2. Classifications based on single characters should be avoided so far as possible. Such characters may be useful for phylogenetic arrangements, but the latter could be indicated by the use of subgenera without interfering with the primary function of generic names.
3. Every effort should be made to avoid erection of monotypic genera based on exaggerated characters, useful though these may be for recognition. An example of a genus erected on such characters, but later expanded to include related forms with less extreme modifications, is Dicranocentroides. Other such genera have been erected recently, e.g. Ongulonychiurus, Bessoniella. It may be impossible, or impractical, to extend the scope of these genera to encompass related but less aberrant species. But so long as these genera remain monobasic, their phylogenetic position is in no way indicated by their names.
4. No new genus should be erected which is clearly polyphyletic. This means in particular that genera should not be based on characters subject to frequent convergent or parallel evolution, such as reduction in eye number, elongation of the unguis or reduction in tenent hairs. Furthermore, existing genera based on such characters, such as Pseudosinella, should be recognized as composite. However, existing polyphyletic genera should be fragmented only with great care and following analysis at least of the majority of the species included. It is suggested that if a phylogenetic classification of such a genus is attempted, the segregates should be regarded as subgenera at least until it is possible to extend the classification not only to the greater part of the genus but also to related species in the genus from which the polyphyletic genus is derived (e.g. Lepidocyrtus-Pseudosinella).
5. Newly discovered species which do not fit perfectly in existing genera should not be placed in new genera unless it is impossible to accommodate them in the existing genera with minor modifications of the diagnoses of the latter. A claim that a segregate identified by a single character, even a synapomorphic character, deserves its own generic name is excessive, and such a segregate is very conveniently recognized by subgeneric ranking (ex.: Sensillanura-Americanura).

An author proposing a new genus should, before publication, ask him (her) self if the principles stated above have been taken into consideration. Ignoring this usually destroys generic utility for function 2) and 3) and often leads to erroneous conclusions by non collembolists.

Peter F. Bellinger,
Kenneth A. Christiansen