Skip to main content

The Origins of Species

ISSUE:  Winter 1940

Professor Punnett once wrote that Darwin’s great work had in point of fact been instrumental in deflecting attention away from the question of the origin of species and canalizing it into the broader problem of large-scale evolution. Today, after eighty years, the species problem has come to the forefront of biological research. This is due partly to the progress of systematics itself, the amassing and analysis of detailed collections of animals and plants from every region, and partly to the rise of new branches of biology, such as genetics, cytology, and ecology, which are illuminating the problem often from unexpected angles.

If Darwin were alive today, the title of his book would have to be not the “Origin,” but the “Origins of Species.” For perhaps the most salient single fact that has emerged from recent studies is that species may arise in a number of quite distinct ways.

From another angle, we may say that the study of species is turning into the study of evolution in action. Large-scale evolution we can only deduce, or at best follow on its vast time-scale with the aid of fossils; but small-scale evolution is proceeding here and now, and lies open to analysis with the aid of the tools of modern research. We can hope for new facts and generalizations from its study, whereas it is unlikely that any further important principles concerning large-scale evolution remain to be brought to light.

Past students of the problem have pointed out with justice that the differentiation of new species depends on three factors—variation, selection, and isolation. Variation furnishes the raw material, the building-blocks of evolution; selection is the guiding and shaping agency; and isolation provides the barriers which allow forms to separate and diverge. But until recently these terms were largely covers for our ignorance. This was notably so in regard to variation, since until well into the present century little was known as to how organisms varied, and even as to what types of variation could be inherited. Today, however, we would include under this head the nature of the hereditary mechanism and of its modes of change; and our new knowledge here has led to new results, some of them of great importance.

It is common knowledge now that the machinery of heredity is provided by the chromosomes of the cell nucleus. These exist in each species in a definite number, one half derived from the male parent, the other half from the female. Furthermore, each kind of chromosome has its own individuality, consisting of a large number of hereditary units or genes, arranged in a definite linear order. The genes, to use an old metaphor, are the cards with which the organism has to play the game of life; and normally each animal or plant has two complete packs of these genetic cards, one from its father and one from its mother.

The genes are alive in the sense that they are self-reproducing (or at least self-copying). Normally a gene persists in the same form from generation to generation. Occasionally, however, a change occurs in the gene—it mutates, as we say in technical parlance, and then it persists in its new altered form until a fresh mutation occurs. Thus each kind of card may exist in a number of sub-kinds, each sub-kind having a slightly different effect on its possessor: changing the color of its eyes, reducing its fertility, increasing its resistance to cold, modifying the shape of its limbs, and so forth.

The sexual process shuffles and redeals these cards so that every possible combination of the different types can be realized. This is one of the fundamental facts discovered by Mendel.

With this brief preamble, let us look at one or two aspects of recent work on the species problem. One of the most startling facts, which would have been regarded as impossible by earlier generations of biologists, is that new species may arise suddenly, at a single bound. This depends on another property of the hereditary machinery. Normally, when a cell divides, its chromosomes all split lengthwise and the halves separate, so that each daughter-cell receives a complete set. Occasionally, however, though the chromosomes split, the cell misses a division, so that it and its descendants have double the normal number of chromosomes.

Now consider what happens if two distinct species cross. Their offspring contain two packs of chromosomes; but these, even if of the same number in each pack, are in most cases so dissimilar that when the time comes to sort them out so that each reproductive cell contains a single pack, with one of each kind of chromosome instead of a pair, they are incapable of executing the very precise maneuvers needed to effect this properly. Accordingly, the reproductive cells receive too many of some chromosomes, too few of others, and the result is complete (or in some cases almost complete) sterility, either through the failure of the reproductive cells to form at all, or to function properly if formed, or to produce a normal individual if they should manage to unite.

But if the chromosomes have been doubled, then each can find a mate like itself; the microscopic maneuvers can take place according to the rules, and the organism is fertile. What is more, it is now largely or wholly sterile with either of its two parent species, as the offspring from such a cross will have three instead of two of each chromosome in one set, and this again upsets the maneuvers of sorting-out during the formation of the reproductive cells.

Quite a number of new species are now known which have originated in this way, some produced experimentally and some found in nature. They may even be more successful than their parents. This is the case with the rice-grass Spartina townsendii which is used by the Dutch to reclaim land from the sea: it resulted from a cross between a European species and an American one accidentally brought over by shipping.

So far, all the examples of such sudden species are from plants. It would probably be impossible for the process to occur in higher animals because of their special method of sex-determination, which would not work if the number of chromosomes were doubled.

Chromosome-doubling after crossing is a method of species-formation in which the isolation is not spatial but genetic—the barrier between the new form and the old is provided by a change in the microscopic machinery of inheritance, which prevents fertile crossing. Nor has selection played a part in modeling the new type. It arises suddenly and stands or falls on its intrinsic merits.

Other changes sometimes take place in the genetic machinery that may assist in isolating new types, though the isolation is not so complete. For instance, a considerable section of one chromosome may become inverted end-to-end, so that the genes it contains are now in reversed order. When this happens, the genes in the inverted section cannot be recombined freely with the corresponding genes in normal chromosomes. Thus this section of the germ-plasm of the species is effectively divided into two parts, which must remain isolated from each other in subsequent evolution, even if the species itself remains single. However, if there should subsequently arise mutations which reduce the fertility of crosses between the type with the inverted and that with the normal uninverted section, the species may split into two.

Accordingly, such accidents to the chromosomes, while not immediately producing new or even incipient species, may pave the way for species-splitting later, in the same sort of way as is done by geographical isolation of a population on an island or a mountain top.

Other similar rearrangements of the chromosome machinery may occur. For instance, a bit of one chromosome may become detached and then attached to a different kind of chromosome. Such accidents, each in their own special way, may provide partial isolation and pave the way for species-splitting. This sort of thing seems to occur in many animals. Certainly the little fruit fly, Drosophila, which has yielded more information on heredity than probably all other organisms lumped together, is very prone to such happenings. The numerous different species of Drosophila are all characterized by such internal rearrangements of the chromosomes, and in many cases the rearrangements are both numerous and far-reaching.

Some species of Drosophila are so alike to look at that it was only their sterility on crossing which led to the discovery that they were separate species. In all such cases, accidents seem to have occurred to the chromosome machinery, providing some initial degree of genetic isolation to form a partial barrier between the two different stocks.

Another quite different type of barrier is that provided by ecological isolation, when groups are divided by differences in their habits or habitats. The best-analyzed cases concern what are called “biological races” of parasites adapted to different hosts, or of plant-eating insects adapted to different food plants. At the outset, such groups seem to be held apart rather incompletely by accidental experience. The moth that has lived on a plant of kind A as a grub will generally prefer to deposit its eggs on a plant of the same kind instead of on a plant of kind B. Mutations crop up later and are incorporated into the animal’s inheritance, giving it an instinctive preference for one or the other food plant; still later other mutations give it an instinctive aversion to mating with an individual of the other race. The further the process goes, the more will selection encourage such mutations, for if each race is nicely adjusted to its particular food plant, any mixture of the two races will be less closely adapted, and will therefore be at a disadvantage.

Once the isolation is fairly complete, other differences can and often will accumulate, so that the two types, after passing through a stage in which they are almost or quite indistinguishable by appearance, though they behave as good species do by exhibiting sterility when crossed, can be visibly separated as well.

Undoubtedly this sort of process on a broader basis has operated extensively in nature. For instance, in Lake Baikal the water-shrimps of the sandhopper family exist in numerous species unknown elsewhere in the world, some adapted to life in the open water and others to the depths, as well as to various more ordinary habitats; and there can be no question but that they have all diverged hi situ from some one or two ancestral forms. The same sort of thing is often found on oceanic islands—as witness the mocking birds of the Galapagos. Incipient stages in the process are also known, as with certain North American mice, where two distinct races are found in the same geographical region but in different habitats, one in woodland, the other in open country. Here the two forms are still merely subspecies, capable of fertile interbreeding if confined together, but kept apart in nature by the invisible barrier of their ecological preferences.

Finally, there remains the geographical type of isolation, in which the barrier between groups is provided by geographical features, like rivers, mountain ranges, or stretches of sea for land forms, stretches of land for water forms. The results of this sort of isolation have been the most thoroughly investigated, and are in many ways of great interest. One fact that has long struck systematists has been the much greater amount of divergence achieved on small islands as compared to large continental areas, even when the differences in environmental conditions are smaller. Thus there are almost as many different races of lizards in the Adriatic as there are islands, while on the neighboring mainland the species is uniform over large stretches. Again, the common wren remains the same throughout Great Britain and all the mainland of Western Europe. But on the islands off Scotland differentiation has set in. The Shetlands boast one quite distinct type, St. Kilda a second, and the Faroes yet another.

This excessive differentiation of isolated populations (the same thing happens in fish, as in the char of isolated lakes) has until recently remained as an empirical fact for which no adequate explanation was forthcoming. A few years ago, however, Professor Sewall Wright of Chicago showed that it was to be expected as a consequence of Mendelian inheritance. The mathematical reasoning involved is too complex to set forth here. But the results are simple enough. Briefly, if isolated populations are small enough in numbers, then mere chance will step in and largely override the effects of selection. New mutations or new recombinations of old genes will often become established even if they are not advantageous, and in some cases even if they are slightly dis-i advantageous. The result is to promote divergence which is non-adaptive and, biologically speaking, accidental and irrelevant. An analysis of the Adriatic lizards mentioned above has confirmed these deductions in a very pretty way. Other things being equal, their degree of difference from the mainland form is greater when the islands they inhabit are smaller. This is to be expected since the effects of chance will increase as the size of the group goes down.

However, even on large continental masses some differentiation may take place, with mere distance and difference in climatic conditions as the isolating factors. The majority of widespread small birds and mammals, for instance, can be classified into distinctive subspecies, each with its own area of distribution. In many cases the process has gone further and split an originally single group into two or more “good” species. An excellent example is that of the Eastern and Western European tree creepers. While separated by the Ice Age, they diverged to such an extent that even though they now overlap in Central Europe, they never interbreed.

Subspecies often interbreed freely where their areas touch, but the zones of mixture are almost always confined to narrow belts. This is at first sight puzzling. Why, if they meet and interbreed, is there not a continuous gradation from one extreme to the other, instead of two more or less stable subspecies separated by a narrow zone? Why is there not a smooth slope instead of a staircase of change? Here again genetics provides the probable answer. The two subspecific types are adaptive, not only in their relations to the outer world but in their internal constitutions. They differ in a considerable number of genes, and each set of genes forms a harmonious stabilized whole, adapted to give the maximum vigor and viability in the circumstances. When they meet, they can still interbreed. Rut as a result of their interbreeding, the harmonious constellations of genes are taken apart and recombined in all sorts of ways, which will almost invariably be less favorable than the two parent combinations. Thus, by adverse selection, the new combinations will be prevented from spreading and the mixed zone will be kept narrow.

Subspecies have often been stated to be species in the making. Undoubtedly many of them are. But, equally undoubtedly, many of them are not. Many widespread species are permanently divided into a number of these partially isolated subspecies, still exchanging a few genes with each other by interbreeding, but each relatively stable on the whole. And this condition, as Professor Wright has shown, is the most favorable one for rapid evolution.

It appears that there are two positions of relatively stable equilibrium in the process of evolutionary divergence. There is the stage of species or of complete biological discontinuity, and there is the stage of interbreeding subspecies or of partial biological discontinuity.

Finally, we must remember that in most cases, both subspecies and species are adaptive, in the sense that they are adjusted, often very closely, to their way of life or to the climatic conditions of their environment. Even when their visible characters do not seem adaptive, experiment shows that selection has been at work upon their invisible but much more important physiological characters, such as temperature-resistance, ductless glands, or metabolism.

We are now in a position to view the species problem in rather a new light. In the animal kingdom alone, about a million distinct species are already known, and the number is being increased every year by ten thousand or so new ones as the result of new exploration and discovery. Here is indeed an astonishing diversification of life. How is it related to the broad processes of long-range evolution? The answer seems to be that it is largely independent of them, or irrelevant.

Long-range evolution, guided by selection, produces divergent specialization of types over tens of millions of years: the placental mammals, for instance, gradually radiated out into carnivores and insectivores, bats and ungulates, rodents, cetaceans and primates. It leads to the widespread extinction of older types and their replacement by new types which radiate and specialize in their turn. It leads, in a few and ever lessening number of lines, to true evolutionary progress.

Superposed on this, selection also sees to it that each type becomes adapted to different climates and to minor differences in habitat and environment. The garment of life in which the globe is clothed is thus adjusted in detail, as a suit of clothes is fitted by a tailor to the peculiarities of a client.

But on these processes of adjustment and progressive adaptation, major and minor, a series of discontinuities is superimposed. The cloth of life is divided up into a mass of snippets. Partly this discontinuity is imposed by accidents of the outer world. A mountain range or an arm of the sea, produced by subsidence, an ice age or other geographical event, separates population. Other groups are isolated by ecological accidents, in the shape of differences between habitats—woodland and open country, pond and swamp, high ground and low ground. But partly the discontinuity is imposed by accidents of the organism’s internal constitution—by doubling of the whole chromosome-complement, by inversion or translocation of chromosome-sections, by the development of harmoniously stabilized gene-combinations which automatically restrict the spread of other combinations. And finally the two agencies may cooperate, as when geographical barriers isolate small populations, and then useless accidental characters automatically accumulate.

The result is that life finds its expression in the form of almost innumerable separate groups, some fully separate, like good species, some on the way to full separation, like geographically isolated subspecies, some at the half-way equilibrium point of partial separation, like continental subspecies still interbreeding at their margins.

It is quite irrelevant to the slow processes of long-range evolution whether the European tree creepers should exist in the form of one or of many species. Owing to the accident of the Ice Age, they happen to exist as two species. It is equally irrelevant that the lizards of the Adriatic should have become divided into a large number of subspecies: they owe this to the geographical accident of the submergence of a mountainous coast with the resultant formation of many small islands.

Evolution in the broad sense consists of a few kinds of long-range trends. But these are cut up by isolation into species and subspecies, whose enormous numbers bear no relation to the major underlying trends. And even the adaptive nature of these small units is largely obscured by the frills and furbelows of non-adaptive accident which can lodge in these discontinuous group-units—mere diversification abundantly but meaninglessly superposed on the adaptive meaning and slow advance of life.


This question is for testing whether or not you are a human visitor and to prevent automated spam submissions.

Recommended Reading