10 The Plant World. 



that interbreed. Johannsen denies this idea of Plate's. At 

 this point we must discriminate carefully to avoid confusion. 

 According to our author's hypothesis, the pure line is a fixed 

 type and thus far Plate's criticism is correct, though such fixed 

 types may have been in existence for a long period. But therj 

 is nothing to show that a variety of beans, such as the Princess 

 beans, from which Johannsen secured his pure lines, is less 

 variable than a variety of corn which constantly interbreeds. 

 The standard deviations and the coefficients of variability may 

 be as great in the variety of com as in the bean variety. 



In a later article * Johannsen seeks to show that when 

 different varieties of beans are crossed, the variability of the 

 hybrids is no greater than that of either parent. His results 

 are certainly interesting but are scarcely conclusive. He deals 

 here with the seeds of plants, using the seeds as a measure of the 

 variability. Again he brings into consideration the partial 

 variations induced by the accidental environmental conditions 

 of the 3'ear. These would have arisen under any condition and 

 theii presence in the data may be insufficient to mask the germi- 

 nal variations, which arc evidently those in which Johannsen 

 is interested. The data of the next generation should have been 

 taken and the means foi the characters established for different 

 plants individually. 



Johannscn's work has been of inestimable value in bringing 

 out clearly the existence of pure lines of biotypes. He has 

 shown that they are of enough importance to be reckoned with 

 in heredity studies. He has shown in a statistical manner that 

 the variety can no longer be considered as a unit and that sta- 

 tistics dealing with populations, while they may present facts, 

 have little application for the students of heredity. The facts 

 do not speak in terms of evolution. He has shown that a variety, 

 a rac5, a population, does not have a biological mean and that 

 selection can not be made upon the basis of the law of ancestral 

 heredity. His paper is almost epoch making in its relation to the 

 theory of unit characters in self-pollinated forms. The idea, 

 to be sure, has been dominant at Svaloef since 1892, and was 

 championed in a way by DeVries, but Johannsen gave the idea 



*Rpt. 3rd Int. Conf. on Genet, London. 



