Apologies for a mild digression, although whilst this is not entirely on orchids, it is certainly not about US politics! I am adding this after writing it, pondering whether it is too long, too abstract and just plain dull. However, it may amuse some, so here it is.
Reiseberg et al (Science 440 p524 23 March 2006) begin with words guaranteed to raise some hackles: 'Many botanists doubt the existence of plant species, viewing them as arbitrary constructs of the human mind, as opposed to discrete, objective entities that represent reproductively independent lineages or "units of evolution".' That is, we see species much the way we see individual mountains in a range - naming them as independent things - "Mount Everest" - when they are in fact mere ripples on an underlying range. The paper tests this view. Anyone who likes plants knows that the range of forms within a given species - or even within the population of a species that is present in a field - will be quite wide. There are tall ones and short ones, brightly coloured examples and washy little things. This raises two issues. First, the "species" is supposed to be the centre of weight of this natural variation, but assessed in what way? Second, where is it legitimate to draw a distinction between this cloud of variation and that one, saying that the two are somehow different entities? Most taxonomy has relied upon authoritative judgement as a way to settle this. The result has been debate, some of it acrimonious, as reputations do hang on who a peer group see as the most "authoritative" when there is a difference of views. Science has tried to get away from this, substituting objectivity for judgement. There are, however, two established but rival objective methods of assessing distinctiveness. One of these methods is essentially statistical. The properties of a large number of instances of a possible species are measured, using features such as petal length. These data are compared to similar measurements that are taken from those organisms that are similar to the one under study, but from which it might be distinct. Standard statistical techniques are used to extract an abstract set of composite factors which characterize these measurements, usually called the "principal components". These components make up a space, which one can think of as a plane, or cube; and each instance of an organism is then a point in this space. Lots of instances create a cloud of points, with the centre of weight representing the average plant, and the spread showing the range of variation that has been found. Another, adjacent cloud shows the contrasting variation and average of the possibly-separate species that is under assessment. There are, then, standard tests to ask whether the overlap between these clouds is the result of random variation or real difference. If the distance between the two centres of weight is great enough, then the two phenotypes are thought to be distinct species. Biologists distinguish the phenotype from the genotype of an organism. The genotype is the ideal form, expressed in an utterly equable and competition-free environment. It is a Platonic ideal of the organism, never realised but latent in the genes. The phenotype is what reality does to us, and is the result of interactions between competitors, less than ideal and variable growing conditions and diseases may deliver. Some variation with in a breeding group is genotypic - because all do not (usually) share identical genes - but the greater part is usually phenotypic. The statistical approach to sorting between variable population is in fact intended to eliminate phenotypic variation from consideration, and it is therefore misleading that it is called a "phenotypic" analysis. The other, rival approach to this type assessment is termed 'phylogenic'. This attempts an even more rigorous approach, attempting to grasp the nature of the genetic relationships that underpin similarity and difference. It does this in two ways. First. it relies upon a more or less objective assessment of the way in which clades - descendant and diverging genetic 'mountain ranges' - segregate. This branching and segregation is often shown as a tree, in which closer twigs are more related than distant ones. A typical orchid cladogram is given in Dressler, for example, but see also Bert Pressman on http://www.cosspecies.org/February%202002.htm Second, although with many problems of assessment, it draws on detailed measurements of genetic similarity and divergence. (Unhappily, similarity that is established through the study of one gene may completely contradict the similarities established from the study of another one.) Indeed, the phylogenic approach really denies the existence of species as anything but a convenient label for the current carriers of bundles of genes that they share overwhelmingly with other so-called species. Much work has been done along both of these paths. A major problem is that both of these two objective schemes do not agree at all well with each other, and neither of them agrees with the more traditional and subjective system that has given their names to almost all plant species. About half of all plant and animal species are not 'concordant'. There are perfectly rational reasons to say that A and B are different; or that they are not. Reiseberg et al try to bring some light to bear on this. Looking at reproductive isolation - the probably most fundamental definition of speciation - they find that about 70% of taxonomic plant species (75% of phenotypic species) are "real" (as opposed to only 40% of animal species!) Ferns are the most valid, in this sense, and birds the least - perhaps only 20-30% of bird species are valid, in the sense of being incapable of interbreeding. Over-splitting by traditional taxonomists is a major source of unreal distinctions. Some areas - such as ferns, above - have received much less attention than do others (for example birds, but dare I say orchids?) and so the tendency to spilt or to lump varies greatly. "Popular" categories are almost certainly wildly over-split, as with birds - and perhaps so, too, with popular orchid genera, to name no names. When we thing about whether some entity should be seen by one polysyllabic binomial or another, therefore, we should exercise no small degree of caution. It seems likely that we are both over-split and wrongly split, and that the fairly immediate future will see some considerable consolidation. Essentially, the lesson has to be that what works for you is likely to be about right, for now. ______________________________ Oliver Sparrow +44 (0)20 7736 9716 www.chforum.org _______________________________________________ the OrchidGuide Digest (OGD) [email protected] http://orchidguide.com/mailman/listinfo/orchids_orchidguide.com
