Scientific American 189(6): 92f (1953)
Experiments in Acquired Characteristics
Ostrich calluses and many other inherited features are difficult to explain by chance mutation. The
author believes that the work of modern embryology now makes a Darwinian explanation possible
by C. H. Waddington
If a man does more than the average amount of muscular work, his muscles do not get worn out by the extra wear and tear, as an automobile would, but on the contrary become larger and stronger. A burrowing animal such as the mole possesses, even from birth, exceptionally well-developed shoulders and forefeet, conveniently built for digging. Every kind of creature is endowed with or develops qualities—we call them "adaptations"—which are neatly tailored to the requirements of its special mode of life.
How these adaptations come into being is one of the oldest and still one of the thorniest problems of biology. Aristotle and other Greek philosophers discussed the question in Athens 2,000 years ago. A hundred and fifty years ago the French naturalist the Chevalier de Lamarck (who invented the word "biology") tried to explain it by his theory of the inheritance of acquired characters. He supposed that if a type of animal responds, generation after generation, to nature's demands by becoming more efficient, the improvement will eventually become hereditary. Fifty years later Charles Darwin expounded the theory which still provides the most useful general explanation of adaptation. In his great work On the Origin of Species by Means of Natural Selection, he argued that the whole of evolution depends on random changes in the hereditary constitution and the selection of helpful changes by the environment. If a change, which we nowadays call a gene mutation, happens to make an animal better adapted and thus more efficient, that animal will leave more offspring than its fellows and the new type of gene will increase in frequency until it finally supplants the old.
There are still some difficulties, however, for which the strict Darwinian solution is not very satisfactory, and the debate goes on. Lamarck's theory has been revived in a modified form by various scientists, most recently by the Trofim D. Lysenko school in the U.S.S.R., who were in a hurry to adapt crop plants to the rigors of the Russian climate. Some types of adaptation are indeed difficult to account for in Darwinian terms. In this article I shall consider a type of phenomenon which has often been thought to demand some sort of Lamarckism. I hope to show that, if data from modern experimental embryology are brought into the picture, Darwin's theory not only can provide a plausible explanation of this kind of adaptation but can be made even more powerful than it has hitherto been thought to be.
Let me first make clear what the area of controversy is. The kind of adaptation illustrated by the first example I cited—the development of muscles by use—of course is not an evolutionary phenomenon and has nothing to do with Darwin's or Lamarck's theories. This phenomenon is known as "direct adaptation." It occurs in response to circumstances, usually external, during an individual's lifetime, and examples of it can be multiplied indefinitely. If an animal has one kidney removed, the other kidney enlarges until it can deal with the animal's excretory needs. If a young puppy loses its forelegs, the bones of the hind legs develop in a way which enables the dog to hop about on its hind legs more easily and efficiently. Just why or how the developing bones react in this way to the stresses placed upon them is not known. Indeed, the processes underlying direct adaptation in general are still mysterious; very probably they involve fundamental biological activities, such as those of enzymes. But there is nothing hereditary about them.
|WING OF THE FRUIT FLY Drosophila normally has two cross veins (left). When Drosophila pupae 21 to 25 hours old are subjected to a temperature of 104 degrees Fahrenheit for two hours, cross veins of the adult flies are sometimes broken (right).|
|*I must now explain what I mean by this statement: the environment affects the shape and organisation of animals, that is to say that when the environment becomes very different, it produces in course of time corresponding modifications in the shape and organisation of animals.
It is true if this statement were to be taken literally, I should be convicted of an error; for, whatever the environment may do, it does not work any direct modification whatever in the shape and organisation of animals.— Lamarck (1809)
On the other hand there are a great many adaptations which the Darwinian theory has always found hard to explain. Certain adaptations in animals and plants, which are demonstrably hereditary and therefore products of evolution, are of exactly the same kind as changes which can be produced in a developing individual by a direct effect of the environment. A classic example is the calluses of the ostrich. The bird sits on rather peculiar parts of its anatomy: the two load-bearing points are at the front of the breast and near the tail. At these two places the ostrich has large, thick callosities. Anyone who works with his hands or walks in bare feet knows that continual pressure and rubbing cause the skin to grow thicker and tougher, i.e., to form calluses. But the callosities of the ostrich, at the present stage of its evolution, are certainly not produced in this way. They appear on the chick while it is still in the egg, before it has sat on anything. They must in fact be hereditary. Now the orthodox Darwinian theory suggests that there is no essential connection between this hereditary adaptation and any direct environmental effect, that the adaptation could only have arisen by a chance mutation. This is asking us to believe a lot. Can we really be satisfied with a theory which suggests that, purely by chance, a hereditary change has turned up which produces callosities in just the right places, and that the sitting habit of the ostrich had nothing to do with it?
|REAR SEGMENT of Drosophila treated with ether at the age of three hours tends to develop into middle segment (left). The hereditary effect is slight. At right is a normal fly.|
|CALLUSES OF THE OSTRICH are at two points upon which it rests its weight while sitting: at the front of the breast and near the tail. These growths are present in the embryo.|
To be sure, there seems to be no limit to the possible forms of genetic mutation, and theoretically a mutation producing callosities in the right places to be useful to the ostrich could occur by chance if me waited long enough. If the ostrich case stood alone, one might be willing to accept it as a fortunate accident. But nature is full of similar strange coincidences, even in this matter of callosities. The skin on the soles of our own feet becomes thickened in the foetus before birth, and so do the pads on the feet of logs, cats and other animals. The African wart-hog, which has the habit of kneeling on its wrists while feeding, is born with thickened skin in those places. Other organs of animals exhibit the same sort of anticipatory adaptations. A striking example is the second molar tooth of he dugong, a tropical cousin of the whale and the manatee. The herbivorous dugong's molar is not conical, as its carnivorous relatives' molars are, but flattened—the better to crush its vegetable food. This flattening develops in the dugong embryo; its molar starts with a pointed tip, but the tip is gradually resorbed during the embryo stage.
Phenomena of this kind certainly seem at first thought to argue for the Lamarckian theory of inheritance of acquired characters; they suggest that the way in which the organ is used has in the course of time, after molding the developing organ generation after generation, brought about a hereditary change. However, the Lamarckian explanation would require us to believe that two implausible things are true. We have to suppose that the stresses which mold a developing organ can (1) alter the genes, and (2) alter them in precisely the way that is required to cause the appearance of the same characteristics as the stresses originally produced in the developing body. There is no convincing experimental evidence for the first of these hypotheses, let alone for the second. It is true that certain very special types of external agents (e.g., X-rays and some chemicals) have been shown to produce mutations, but even these unusual changes fail to fulfill the second part of the theory: the changed traits resulting from the mutations are in no way related to the inducing agent.
Thus neither the conventional Darwinian theory nor the Lamarckian theory provides a satisfactory account of adaptations of this kind. But in my opinion the Darwinian theory seems unable to explain these cases only because something has-been left out of the usual statements of it. The point which has been forgotten is this: Evolution does not necessarily mark time until a chance mutation produces a required modification. If an environmental stress modifies the development of an animal and causes it, during its lifetime, to become adapted to deal with the situation, this response itself becomes subject to evolutionary processes. In a large population of animals, natural selection will favor those whose hereditary constitution makes them best able to respond to a particular environmental stress with an appropriate adaptation. This selection will cause the population of animals to evolve in such a direction that the change necessary to make the adaptation hereditary is much more likely to occur.
|MOLAR of a dugong embryo is first pointed (left). Then, without use, it becomes flat.|
Before going any further with the theoretical argument, let me describe some experimental facts bearing on the situation. My experiments were made with the favorite animal of the geneticist, the fruit fly Drosophila. They dealt with an environmental effect, and in order to have something very definite to study I selected an effect which is unusually clear-cut. When Drosophila pupae 21 to 25 hours old are subjected to a temperature of 104 degrees Fahrenheit for two hours, some of them respond to this unusual environmental stress by developing a gap in one of the wing veins—the posterior cross vein. This change probably has no adaptive value in nature, but in the experiment it was treated as if it were useful and also as if it were harmful. In the first case, only those flies that responded to heat by developing a broken vein were selected for reproduction, and this selection was continued in succeeding generations. Meanwhile another line was bred by selection of those flies that failed to develop a broken vein in response to heat. As generation succeeded generation, the two selection lines pulled apart, that in which the response was favored showing an ever increasing proportion of flies which developed a gap in the vein, and the other line an ever decreasing proportion. In the first line, which is the more important one, the selection was in effect gradually improving the animal's hereditary ability to respond suitably to the environmental stress.
|DEVELOPMENTAL PATHS of an organism before environmental selection are depicted by the model at the left. There is a main path separated from a secondary path by a threshold. After selection secondary path has been deepened and threshold lowered (right)|
Now in this line some of the offspring in each generation were not given the high-temperature treatment. At first none of the untreated flies showed a broken cross vein. But after 12 generations a few were found with gaps in the vein. In the next generations there were more, and when these were selected and bred from, the selected strains eventually came to have a high percentage of flies with broken veins, even though they were never subjected to the heat stress. In them the break in the vein had become fully hereditary and no longer depended on the heat treatment.
This phenomenon, which we call "genetic assimilation of an acquired character," seems to provide an explanation of the whole category of adaptations which were so difficult for conventional Darwinism to deal with. How does the genetic assimilation work? Roughly speaking, the answer must be that the genetic constitution becomes so ready to make this particular response—is set on such a delicate hair trigger to do so—that finally the response occurs on its own without requiring the environment to touch it off. Recent work in experimental embryology has shown that developing tissues very commonly get into unstable states in which comparatively minor influences will shunt them into one or another of various possible paths of development. For instance, in the early embryo of a vertebrate the outer layer of tissue (the ectoderm) is capable of becoming either skin or nerve or tissue of the middle layer (mesoderm). Some slight stimulus from the tissues with which it is in contact decides which path a given piece of ectoderm will follow in its development.
We are probably dealing with a similar situation in genetic assimilation. At an early stage in the evolution of an adaptive character, we can picture the relevant part of the animal's developmental system as consisting of a main developmental track which leads to the normal nonadapted adult condition with a rather ill-defined alternative leading to a roughly adapted condition. Development will go on along the main track unless at the appropriate time an environmental stress pushes it over into the alternative path. Now suppose that an external stress is acting on a population where natural selection is favoring the members that respond best to this stress. After a time we shall find that the path leading to the adapted condition is better defined than the main path, and also that it has become easier for development to choose that path. The threshold between the adapted alternative and the original main track will have been lowered. If this lowering goes far enough, the alternative will become the main track, and genetic assimilation will be complete.
So far very few cases of genetic assimilation have been studied experimentally, and there is still a great deal to be discovered about it. One point still to be settled arises in connection with the examples of the ostrich and the dugong. Their anticipation, in the embryonic stage, of the need for their peculiar adaptations has not yet been imitated experimentally. But I think there is no difficulty in explaining it. It is clearly an advantage to an animal if it has its adaptive features ready before they are needed. In exactly the same way as natural selection will cause the genetic assimilation of an adaptive character, it should eventually cause the character to appear early enough to be of maximum use. There is, in fact, a general tendency for developmental modifications in an animal to occur earlier and earlier in its life history as its evolution proceeds. Of course one should expect that it will take much longer for an anticipatory adaptation to evolve than for the adaptive feature itself, and this may explain why it has been impossible so far to produce one in the laboratory.
A more important problem is the question whether genetic assimilation takes place through the rise of new genes by mutation. Lamarckians would, I imagine, insist that it does. The experiments so far do not disprove that idea, but some of the data argue against it. For instance, the assimilated broken-vein character in the fruit flies depends on the action of a rather large number of genes. It is one thing to suggest that the heat treatment might produce one or two mutations acting in the correct way, but surely it is very unconvincing to suppose that it produces large numbers of them.
Actually the assumption that mutations take place is unnecessary; we may more plausibly assume that the genes on which the assimilated character depends were present in the original population, though scattered in it. If this is the case, the process of genetic assimilation can go only as far as the genes contained in the initial population will permit. The real test of whether mutations occur would be to try to produce genetic assimilation in an inbred strain, which should contain very little genetic variability. If mutations are not involved, genetic assimilation should not occur in such a strain. This experiment is under way, but so far no definite conclusion can be reached.
Another experiment, also uncompleted, does suggest, however, that the genetic materials available in the initial population set a limit to the progress which can be made. In this experiment three-hour-old embryos of Drosophila were treated with ether. Under this environmental stress the third segment of the thorax tends to develop into the organs normally formed by the second segment, that is, into the main body of the fly and the wings. Selection of flies which show the most readiness to make this response has produced some degree of genetic assimilation. But the hereditary character produced in this way is only a feeble version of the direct effect of ether on the embryo; usually it consists of no more than a slight enlargement of the "balancer" borne in the third segment, and a tendency for it to show some of the characteristics of a wing. It seems probable that the initial stock did not contain genes which could produce any higher grade of the abnormality than this; also that the treatment has not made such genes appear by mutation. But selection is still continuing and perhaps they will eventually turn up.