As stated above, there are no set rules in botanical taxonomy for uniformly defining what constitutes a genus, though many workers have laid out their own views over the years. Not only does this mean that the parameters which are employed will vary according to the views of those specialists who have researched any particular family, but it also means that there is always the possibility that new enthusiasts will use new versions to tinker with long-established classifications and thereby create never ending nomenclatural turmoil.

As the question of what constitutes a genus in any plant family is thus a matter of opinion, it is perhaps desirable to periodically re-examine what one should be attempting to achieve when one assigns a species to a genus or when genera are created or modified, to ensure that the basic principles and raison d'etre of taxonomy are observed and are being applied as effectively as is possible.

The aims of placing various species together and then creating a generic name to designate that group, must surely encompass the following:

• To have a convenient name for referring to that group of plants.
• To indicate that the plants placed in that genus are thought to be more closely related to each other than to others that are excluded from that genus.
• To have a name which, when utilised, should enable a mental image to be evoked of the general appearance of the constituent members of that genus.

To achieve this, the following criteria are requisite for this process:

• All the members of a genus should be closely related, insofar as this can be deduced, and should thus be apparently descended from a relatively recent common ancestor (i.e. be monophyletic), even though there is normally no proof of this by means of a fossil record (an especial problem with succulents). (Now made easier to determine by DNA, phyto-chemical and SEM studies)
• The assigned generic name must be unique, be validly published and be widely known, and it should also be widely acceptable.
• The establishment of a genus and the placing within it of constituent species requires that such a genus must have a unique combination of characteristics that will clearly separate it from other genera in that family.
• This in turn implies that this combination of features is common to all the included species.
• It should be possible to construct a key to the family or tribe that will enable anyone to identify to which genus a particular plant belongs by noting what diagnostic features it possesses, and to utilise these to work through that key until the appropriate genus is correctly located.
• Insofar as is possible, the characters that are utilised for the taxonomic classification of any plant should be those that are easily visible or at most, should be visible with a reasonable hand-lens.
• Features that require microscopic techniques and laboratory facilities should insofar as possible only be used in a supportive capacity to confirm or rebut proposed taxonomic changes: they can however, be invaluable in disclosing relationships that are not immediately obvious to the unaided eye.

In detailing and comparing the visible features that characterise any particular plant populations, it is necessary to bear in mind the possibility that similar features may have evolved in unrelated plants by convergent evolution in response to similar environmental pressures (or even by chance) e.g. the similar appearances of the stems of Pseudolithos in the north-east corner of Africa, and those of Larryleachia (formerly the 'smooth' species of Trichocaulon) down in the opposite south-west end of the African continent: both are inhabitants of very hot arid areas and have achieved similar maximum reduction of stem surface and light reflectance to cope with the hot sun, combined with optimal camouflage as a protection from thirsty predators. However, even a quick glance at their flowers shows that there is no close relationship between these two outwardly similar genera.

Unfortunately there is no way that firm guidelines can be laid down as to how many differing or similar characters a group of species should possess in common in order to justify their separation into two separate genera, or conversely, for two existing genera to be merged: this has to be left to the discretion of individual workers, but preferably they should consult with others who have an adequate knowledge of that family and of the individual species, and they should also attempt to obtain consensus approval for the proposed changes from other appropriate persons in order to avoid a perpetual succession of new names or shuffling of species from one genus to another.

As a rule of thumb, it is generally accepted that there should be at least two, and preferably a minimum of three, significant morphological differences between two species of plants to justify the placement of each in separate genera. Such differences should furthermore not include flower colour differences unless accompanied by major pattern variations, but should comprise genetically fixed morphological characters that are shared by all the members of the proposed genus, and which thereby help point to a mutual common ancestor by virtue of all of them having inherited these same features.

The number and size of the genera that should be recognised in a particular family should not be subject to arbitrary limits that are determined by the convenience of fewer names to be remembered: genera should delineate homogeneous groups of what appear to be phylogenetically related species that exist at this moment in the evolutionary history of such families i.e. they should be based on the realities as found in nature and should not be the result of artificial shoe-horning of species into arbitrary groupings in order to satisfy preconceived ideas about the size and nature of the genera that should be recognised in any particular family.

Because all organisms are continually evolving and because there are always minor differences between the individuals in any given population, different populations of any species may or may not exhibit slight variations from other populations of that species. With the passage of time, and if gene flow between the different populations terminates totally because of widening geographic gaps, these differences are likely to become accentuated to a point where taxonomic recognition is justified, probably first at subspecies level, and later at species level. Here again, there are often differences between workers on any group of plants as to where to draw the line. A similar process, with similar divergence of views, also operates at the genus level.

A further problem is that in some groups of species, the various species may have very different combinations of diagnostic features with no consistently clear-cut grouping of such features. In such cases it extremely difficult to decide how they can be optimally combined into homogeneous genera that all have similar morphological characteristics and which also probably possess close phylogenetic affinities, without having to resort to making monotypic genera for each of the more aberrant and incompatible species in order to leave all the remaining species in genera with the same set of shared features. Creating a multitude of monotypic genera to cater for aberrant misfits has the inherent disadvantage that the criteria and parameters of such genera coincide with those pertaining to the single contained species.

An alternative option would appear to be the creation of larger genera but each of these would then contain a variety of species of dubious relationships. This for instance seems to be the case with species that were formerly in the genera Pectinaria, Caralluma, and Stapeliopsis: these genera were largely rearranged by Bruyns (1981) by expanding the concept of Stapeliopsis, revamping Pectinaria, and creating the new genus Ophionella in the process. However, aspects of his regrouping seem questionable in the light of further investigations, and it will be necessary to undertake another review of all the species involved.

The easiest solution in this and similar cases is to create monotypic or microgenera that contain only two or three species to accommodate the more aberrant ones that do not lend themselves to being grouped together in larger homogenous genera. Although most taxonomists prefer to avoid the creation of monotypic genera, many plant families, especially those that encompass many succulent or xerophytic species, do in fact have a high proportion of monotypic genera. Leach (1983), quoting from Clayton (1972), gave figures of 46.1% in the Cucurbitaceae and 49% in the Asclepiadaceae, with an overall average of 36.8% monotypic genera in the examined tropical and subtropical families. Airy Shaw (1973) noted that the Asclepiadaceae embraced 123 mono-specific genera (i.e. 38.4 %) and 118 genera (i.e. 37 %) with only 2-10 species. Hammer (1994) stated that a similar situation prevails in the Mesembryanthemae. In the Cactaceae, 17.2% of the genera are monotypic and 12.9% have only 2 species according to the list of genera that the I.O.S. Working Party published in Bradleya 8/1990.

At this juncture, it becomes necessary to refer to the long-standing area of disagreement between the 'lumpers' and the 'splitters'. The opposing views of these two schools of thought in the taxonomic world (they exist in zoology as well as in botany) are well summarised by Rowley (1980: 14) in 'Name That Succulent' and will not be repeated here other than to note that each approach has pro's and con's. It therefore becomes necessary to keep in mind what the taxonomic aims and objects should ideally be when classifying plants and assigning any given species to a genus. There is much to be said for having a system that conveys the maximum amount of information with the minimum number of words, i.e. the normal Linnean binomial system.

[in a later personal communication, Darrel writes:]

Further to the copy that I sent you earlier of what I had written about Vadulia longipes ssp villetii, it has subsequently occurred to me that if it was to appear in Asklepios, that I should add a bit about V. maughanii, plus some photos of it, and perhaps a shot or two of typical Pectinaria flowers, in order to emphasis just how very different the two genera are. I would also include mention that for the two species of Vadulia to have diverged so much from Pectinaria, and yet for the two species of Vadulia to have had enough time to also diverge from each other, then their split from Pectinaria and their subsequent split from each other, points to them being a very old lineage that has had ample time to become established as a separate genus that originally derived from a common ancestor for both groups.