6. I am told by Mr. Barrow, who has not observed any.
7. [Grafting and certain other forms of plant propagation do not involve reproductive cells or any other means for the exchange of (what we now know to be) genetic material, so these methods do not actually involve “crossing.” – D.D.]
2
VARIATION IN NATURE
BEFORE APPLYING THE PRINCIPLES FROM THE LAST CHAPTER to living things in the wild, we need to establish whether or not they too are subject to variation. A proper treatment of this topic would involve a long catalog of dry facts, but I will reserve this for my future work. And I won’t discuss the various definitions of “species,” because no one definition satisfies all naturalists, even though everyone vaguely knows what it means. Generally, the term includes an unknown element of a distinct act of creation. “Variety” is almost as difficult to define, but in this case community of descent is often implied even though it can rarely be proven. There are also monstrosities – by which I mean considerable structural deviations that are either harmful or useless to the species and not usually propagated – but these graduate into variations. Some authors use the term “variation” in a technical sense to indicate an un-inheritable modification resulting from environmental conditions, but the dwarfed shells of brackish Baltic waters, dwarfed plants of alpine summits, and the thickened fur of animals living far north might be inherited for at least a few generations, and in such cases I would call the form a variety.
Once again, consider that there are many slight individual differences, such as those observed in offspring from the same parent. (Sometimes differences are just assumed to have arisen this way because they are frequently observed in individuals of the same species living in a confined locality.) No one would argue that all the individuals of a species are cast in the very same mold. Individual differences are important because they supply material for natural selection to accumulate, just as humans can accumulate individual differences in any direction in domestic organisms. Individual differences generally occur in what naturalists consider “unimportant parts,” but I could show a long list of examples where physiologically or classificatorily important parts also vary. I am certain that the most experienced naturalist would be surprised at how many cases of variability, even in important structural parts, can be collected from good authorities. Keep in mind that systematists do not like finding variations of important characteristics, and there are not many people prepared to laboriously compare the internal organs of many specimens from one species. I never expected that the branching pattern of nerves close to an insect’s central ganglion would vary within a species – I expected such changes to occur in small increments – yet Mr. Lubbock recently showed the variability of these main nerves in Coccus to be comparable to the irregular branching of trees. He also found that the muscles in the larvae of certain insects are far from uniform. Some authors state that important organs never vary, and in a circular argument they basically classify “important traits” as those that are invariable (a few naturalists have honestly confessed this). In this scheme there cannot be any varying important parts by definition! But under any other scheme many examples can be given.
With respect to individual differences, one point strikes me as extremely perplexing: the “polymorphic” genera, in which species vary so inordinately that few naturalists can agree which forms should be ranked as species and which as varieties. Among plants, examples include Rubus, Rosa, and Hieracium; among animals, there are several genera of insects and brachiopods. Most polymorphic genera contain some species with fixed characteristics. Also, genera that are polymorphic in one region seem to be polymorphic in others, with a few exceptions. Judging from brachiopod shells, genera remain polymorphic through time. These observations are perplexing because they suggest that this type of variability is independent of the environment. I suspect that members of polymorphic genera vary in characteristics that are of no service or disservice and have therefore not been seized and rendered definite by natural selection, as I explain later.
Those forms that could justifiably be considered species but are similar enough to other forms or linked to them by intermediates so that naturalists do not rank them as distinct species are very important for this discussion. There is every reason to believe that many of these “doubtful” forms have permanently retained their characteristics in their native regions for as long as “true” species. On a practical level, when a naturalist can link two forms through intermediates, he will treat the most common one, or the one that happened to have been described first, as the species and the other as a variety. But there are some very challenging cases – I won’t list them – concerning whether or not one form should be classified as a variety of another even if they are closely linked by intermediates. The commonly assumed “hybrid nature” of intermediates does not always solve the problem. In many cases the intermediate forms have not been found, and they are assumed by analogy to either exist undiscovered or to have become extinct; here a wide door for the entry of doubt and conjecture is opened.
It cannot be disputed that these doubtful forms are common.1 A surprising number of plants from Great Britain, France, and the United States have been ranked as species by some botanists and mere varieties by others. Mr. H. C. Watson, to whom I am grateful for all kinds of assistance, has listed for me 182 British plants that are generally considered varieties but have all been ranked by some botanists as species. He did not include many minor varieties that have nevertheless been ranked as species by some botanists; he entirely omitted several highly polymorphic genera. Under genera, Mr. Babington lists 251 species, whereas Mr. Bentham lists only 112, amounting to a difference of 139 doubtful forms! Doubtful forms of mobile animals that mate to reproduce are rare within the same region but common in separated areas. Many slightly differing North American and European birds and insects are ranked by one naturalist as undoubted species and by another as varieties. Many years ago, when comparing birds of the Galápagos Archipelago to one another and to those of the American mainland, I was struck by the vague and arbitrary distinction between species and varieties. Many of the insects on Madeira are ranked as varieties in Mr. Wollaston’s admirable book but would certainly be characterized as species by many entomologists. Even Ireland has a few animals generally considered varieties but ranked as species by some zoologists. Several experienced ornithologists consider the British red grouse a variety of a Norwegian species, while most rank it as a species peculiar to Great Britain. If there is a large distance between the homes of two doubtful forms, many naturalists rank them as species, but what distance will suffice? If that between Europe and America is enough, is the distance between the Continent and the Azores, or Madeira, or the Canaries, or Ireland also enough? Trying to discuss the demarcation between a variety and a species before establishing any definition of these terms is like hammering the air.
Several interesting arguments from geographic distribution, analogical variation, hybridism, and so forth have been developed to address this problem with respect to different cases. I will discuss only the example of the primrose and cowslip. These plants differ in appearance, flavor, odor, flowering period, and range. They grow in somewhat different habitats and ascend mountains to different heights. Finally, according to many experiments conducted over the course of several years by the careful observer Gärtner, they can be crossed only with great difficulty. There could hardly be better evidence that two forms are distinct species. But for all that, they are linked by many intermediates that are probably not hybrids. There is also overwhelming experimental evidence showing that they have descended from a common parent and are therefore varieties.
In most cases, close investigation brings naturalists to agree on how to rank a doubtful form. Yet the greatest number of doubtful forms is found in