EVOLUTION, VARIATION AND HEREDITY 469 



of heredity ; suggesting the lines of work and also the technical 

 methods to be employed. Numerous investigations have been made 

 along these and similar lines, and many useful results achieved ; to 

 some of them we shall return later. It cannot yet be said, however, 

 that Mendelism has supplied the key to all problems in heredity, and 

 it would appear that certain characters so far have not been analysed 

 in Mendelian terms. Before considering the direct extension of 

 Mendel's work we may turn aside to consider two other discoveries. 

 An important contribution to our knowledge of heredity 

 was made by de Vries, one of the rediscoverers of Mendelism, whose 

 important book, " The Mutation Theory," was published in 1903. 

 This author found wild in a potato field hundreds of specimens of 

 the evening primrose (Enothcra lamarckiana, which seemed to 

 exhibit more than ordinary variability. These he removed to the 

 gardens at Amsterdam and bred carefully. In a few years he found 

 he had produced seven distinct types, " elementary species," as he 

 termed them, all so different from the original that each would have 

 been described as a new species had it been found in Nature. More- 

 over, he found that each of these would breed true to type generation 

 after generation. The most remarkable point about them was that 

 they appeared suddenly and without warning and then persisted. 

 They were so different from the original that they would find no 

 place in the normal curve of the species. For these suddenly arising 

 but persistent forms de Vries proposed the name mutant, and the 

 process was described as mutation. The cause of these mutations is 

 practically unknown, and although there has been much theoretical 

 speculation concerning them we have not time to enter into it here. 



It is quite clear from this, however, that we shall have to re- 

 consider what we mean by variation. We have seen that ordinary 

 variations may be expressed by a normal curve, but these are varia- 

 tions about a mean, and no amount of breeding from among them 

 would produce forms outside the limits of the curve. For such 

 variations, which may be described as " normal variations," the 

 more strict term " fluctuations " has been proposed. Against these 

 we may set the mutations of de Vries. The term " variation," then, 

 as it was used, and as we first used it, included both fluctuations 

 and mutations, and, according to de Vries, it is only the selection 

 of the latter that can have led to the appearance of new species. 



Further interesting observations were made by Professor 

 Johannsen, who investigated the phenomenon known as the " pure 

 line." He found in beans, where self-fertilisation is possible, that 

 all the descendants of a single plant showed normal variation about 

 a certain type. This type need not be the same as the type for the 

 general population of all beans of the same species. Moreover, if 



