September io, 1903] 



NATURE 



455 



At present there appears to be some doubt whether the 

 workers in these subjects are really agreed as to the general 

 propositions of the problems, the definitions of the terms 

 employed, and the standard of proof that is requisite in 

 each step of progress. It is true that in most, if not in all, 

 biological problems we are at the disadvantage of being 

 unable to define or measure anything with the same mathe- 

 matical accuracy that our friends, the chemists and 

 physicists, are accustomed to. We cannot say for example 

 that the chela of a particular species of crab is so many 

 millimetres in length, in the manner the chemist determines 

 the atomic weight of a new metal, as the length of the chela 

 is found to vary within a certain range in all species that 

 have been investigated ; nor can we define such common 

 expressions as a species, a variation, or even a cell with 

 the same conciseness as a physicist defines the ohm, the 

 volt, specific gravity, or the mechanical equivalent of heat. 

 As a consequence it is not surprising that when our problems 

 have been studied and a solution reached the resultant 

 " laws " e.xhibit so many exceptions that they are really 

 not worthy to be called " laws " at all. We may see the 

 truth, but we see it as through a glass, darkly. 



There is perhaps no word in the whole of our vocabulary 

 which is used in so many different senses as the word 

 " variation," and yet when it is used an attempt is only 

 rarely made to define the sense in which it is employed. 



When we study the adult progeny of a single pair of 

 parents we notice that they differ from one another as 

 regards any one particular character within a certain range. 

 Thus the eight children of a single pair of human parents 

 may vary in weight from, say, 130 lbs. to 200 lbs., and we 

 may find that the average weight of the eight children is 

 approximately the same as the average weight of the two 

 parents. If parents and children were all of exactly the 

 same weight — an impossible supposition — it would be said 

 that they exhibited no variation in this respect, but, as they 

 always do vary in weight, it is said that they exhibit 

 " variations " in weight. Now, such variations may be 

 due partly to differences in the muscular training, the 

 nourishment, the general health, and other post-natal 

 causes ; but it is assumed, and there are doubtless good 

 reasons for the assumption, that if all these post-natal 

 influences had been equal throughout life there would still 

 remain variations in weight of lesser amplitude than is 

 usual, but nevertheless substantial. 



The variation of the adult in weight, therefore, is a com- 

 pound quantity, partly due to the influence of external con- 

 ditions upon the growing body, and partly due to a quality 

 or character present at birth and usually supposed to be 

 inherited with the germ-plasm from one or both parents. 

 The former may be called the artificial part of the variation, 

 or for brevity the artificial variation, and the latter the 

 natural or inherited variation. In the character of weight 

 in human beings there can be no doubt that artificial 

 variation is predominant, the character being a very 

 fluctuating one and liable to profound modification in the 

 varying vicissitudes of civilised human life. 



In the character of stature the artificial variation is prob- 

 ably much less predominant. The children of tall parents 

 grow into tall men and women, however handicapped in 

 early life by ill-health or insuflicient nourishment, and the 

 children of short parents remain short in adult life, how- 

 ever healthy and well fed in their youth. Nevertheless, he 

 would be a bold man who would assert that the character 

 of stature is uninfluenced by the environment, and that the 

 short people would not have been taller had the conditions 

 of their life in childhood been more favourable, or the tall 

 people shorter if the conditions in their early life had been 

 less favourable. 



Finally, we have, in the colour of the iris, the shape of 

 the ear, and the size of the teeth, characters which are 

 usually considered to be unmodified by post-natal conditions, 

 or at least so slightly modified by them that the differences 

 observed in them may be regarded as almost pure natural 

 variations. Now, if we turn our attention to characters 

 such as weight, which we feel certain are influenced very 

 profoundly by the environment, we might be tempted to 

 exaggerate the importance of the environment in moulding 

 or forming the characteristic features of the adult organism, 

 as. in (he opinion of many authorities, Lamarck did, and 

 many of his followers are still doing. If, on the other hand, 



NO. 1767, VOL. 681 



we confine our attention to such characters as the colour 

 of the iris or the shape of the ear, we might be tempted to 

 under-estimate the influence of the environment. 



This brings us to the important question whether the 

 characters of the adult that are due to the influence of the 

 environment, and that part or degree of any character 

 which is more or less modified by the conditions of the 

 earlier stages of life are or are not transmitted by parents 

 to their ofTspring. Time will not permit me to discuss this 

 difficult problem here. Rightly or wrongly, I agree with 

 those who maintain that acquired characters are not in- 

 herited, and I intend to assume for the purpose of the 

 argument that follows that they are not inherited. I will 

 also assume, and I must say that the facts seem to be con- 

 clusive in favour of this assumption, that the characters 

 which are usually supposed not to be influenced, or to be 

 only slightly influenced, by the environment are capable of 

 transmission by heredity. 



We have, then, in most variations a part that can be 

 transmitted and a part that cannot be transmitted by heredity 

 from parents to ofTspring, and we find in every plant and 

 animal an enormous difference in the proportions of these 

 two parts in different organs. It is not difficult to see the 

 general reasons for these differences. It is clearly important 

 that some organs should be plastic — i.e. capable of changing 

 in form and size to meet the varying changes in the 

 environment, and that others should remain relatively 

 constant in spite of changes in the environment. 

 Thus the shape and size of the branches of an oak 

 in a plantation will vary enormously, according to the 

 light and space they have for their development, whereas 

 the anthers, the pistils and fruit will be relatively constant 

 in form and colour. It is clearly important for a chama-leon 

 that the colour of its skin should vary according to the 

 colour of its environment ; but it is none the less important 

 that the shape and muscular organisation of its tongue 

 should remain relatively constant throughout life. 



An essential point, however, for us to consider is whether 

 there are any characters in animals or plants upon which 

 the environment exercises no influence at all or exercises 

 such a slight influence that it may be safely neglected. 

 The method to adopt in order to settle this point would be 

 to compare at a definite period of their lives the statistics 

 of variation in a family or population which has been brought 

 up under identical circumstances with those of a similar 

 family or population at the same period of life which has 

 been brought up under differing circumstances. If this 

 were done we could determine with considerable accuracy 

 the proportion of the variation of any character of the 

 individuals that is due to the environment and that which 

 is natural and inherited. 



Unfortunately it is impossible to bring up a population 

 under identical circumstances. If we take, for example, 

 th'! individuals of a single hive of bees, which have the 

 same parents, pass through the early stages of their develop- 

 ment in cells which are almost identical in size and are 

 regularly fed by the workers during the whole of their 

 larval life, there is still a considerable probability that the 

 individuals do not have a treatment which can. with any 

 pretence to accuracy, be called identical. The food that is 

 collected by the worker-bees frequently comes from varied 

 sources or from flowers in different stages of their growth, 

 and it is impossible to believe therefore that it has always 

 identical nutritive properties ; the larvre are not of the same 

 »age, and seasonal changes may affect the larvae differently, 

 some being checked in the early stages of their development 

 more than others. 



But even if we could, with justice, assume that the con- 

 ditions of life for the individual bees in a hive are identical 

 from the time of hatching up to the time when the adult 

 characters are assumed, there still remain two sets of 

 variable conditions which must affect the development in- 

 dependently of the influences brought by the two parents 

 in the germ-plasms. 



In the c^^^ of the bee there is a considerable quantity of 

 yolk, and this yolk is the food material upon which the 

 embryo is nourished throughout the earlier stages of its 

 development. There is no evidence that the yolk in the 

 eggs of this or of any other animal is constant either in 

 quality or quantity. On the other hand, the extraordinary 

 variations or. abnormalities, as they are usually termed, 



