80 



SCIENCE 



[N. S. Vol. XL. No. 1020 



enough to sow a parcel of a moderate ex- 

 tent. Or he limits his choice to one ear 

 only, which will take him one year more to 

 obtain the necessary quantity of grains. 

 The first method is the one which is still 

 commonly followed, the second was intro- 

 duced some twenty years ago by Nilsson. 



The real nature of the first method may 

 be explained by means of the careful 

 studies of Eimpau, who applied it for the 

 improvement of his rye. The group of ears 

 of the first choice will evidently be itself a 

 mixture, although of a lesser number of 

 types. In choosing year after year a hand- 

 ful of the best ears Rimpau must gradually 

 have purified this mixture, until after 

 twenty years he succeeded completely in 

 isolating the very best one of them. 

 From this time his race must have been 

 pure and constant, no further selection 

 being possible. Using the method of Nils- 

 son the same result may be reached by a 

 single choice, and therefore in one year. 

 The new race is produced by a jump and 

 not by the slow and gradual improvement 

 by small and almost invisible steps, which 

 was assumed by Rimpau and Darwin. 



From these discoveries the question 

 arises, whether natural selection also pro- 

 ceeds by jumps and leaps, and not, as was 

 commonly assumed, by imperceptible steps. 

 The answer may be deduced from the ob- 

 servations of Jordan and others on the 

 existence of elementary species in nature. 

 Almost every wild species consists of some 

 of them, and in special cases their number 

 increases so as to embrace dozens or even 

 hundreds of sharply distinguished types. 

 Sometimes these are found in widely dis- 

 tant stations ; at other times, however, they 

 are growing in mixtures. Natural selec- 

 tion will, of course, under changed condi- 

 tions, simply multiply one or two of the 

 types to the exclusion of the others. As a 

 whole, the species will make progress in 



the desired direction, but in reality there 

 will be no change of forms. 



From all these and many other considera- 

 tions it follows that the basis, which the 

 practise of artificial selection seemed to 

 afford to the theory of natural selection, is 

 a fallacious one, and that the idea of evo- 

 lution by means of slow and almost imper- 

 ceptible steps must therefore be abandoned. 

 But if this is conceded, how are species 

 really produced in nature? 



The theory of mutations answers that 

 species are produced by means of jumps 

 and leaps, exactly in the same way as vari- 

 eties in horticulture. Varieties are only 

 beginning species, says Darwin, and the 

 same laws must govern the origin of both 

 of them. Now, in horticulture, it is well 

 known that varieties usually arise at once. 

 In a field of a species with blue or red 

 flowers some day an individual with 

 white flowers is seen. Ordinarily it is only 

 one, and it is not surrounded by transi- 

 tions or by flowers of intermediate colors. 

 Sometimes there may be two or three, but 

 then their flowers are of the same degree of 

 whiteness. One seed of the species has been 

 transformed into a variety, and this is its 

 whole origin. A single season suffices to 

 produce the effect, no slow and gradual 

 improvement being required. Moreover, 

 the seeds of the first individual, if fertilized 

 and saved separately, will reproduce the 

 variety wholly pure. The same rule pre- 

 vails for large groups of other cases ; every- 

 where varieties arise by jumps, requiring 

 only one year for their arrival. 



The same rule also holds good in nature. 

 But in order to show this, direct experi- 

 ments are required. For this object I have 

 cultivated a large number of wild species 

 in my experiment garden, trying to see them 

 produce varieties and to be enabled there- 

 by to study the laws of this process. Let 

 me adduce two instances, the origin of the 



