LI>'>'EAX SOCIETY OF LO>"DOX. 75 



plants than among animals. This is due to the fact that many 

 plants are hermaplirodites and self-fertiHzers, while animals, as a 

 rule, have to pair to raise a family. 



To study heredity, one therefore does best to begin with the 

 descendance of a self-fertilized plant. This Jobannsen has done 

 more carel'ully than anybody else. He has chosen some pure 

 bean individuals and carefully compared the progeny of each 

 jiarent with the latter. The result has been that the progeny 

 dilfered fi-om the parent, but only in such characters as are not 

 transmittable. Apart from this modification, there was no differ- 

 ence between the descendants, so that selection among them had 

 ]io effect whatever. Whether the small beans produced, or the 

 largest ones were sown, the descendants liad the same mean. 



One might yet object that this failure of selection was due to 

 the short time at its disposal. Fortunately we now know of a 

 case to test this objection. Louis de A'ilmorin started, in ]8-J0, 

 a living museum of different so-called varieties of wheat. He at 

 once applied selection to their progeny, and this selection, of the 

 jnost rigorous kind, has, uninterruptedly, been continued up to 

 the present moment. Each year the best ear grown was selected, 

 and the seeds of this ear alone were sown. Thus, among forms 

 with branched ears, the best branched one, among such with long 

 ears the longest one, etc., were selected. In 1911, the collection 

 of original ears of !1840 was found at Verrieres, and at once com- 

 ])ared with the forms still cultivated there. It was shown that 

 36 of the original varieties were yet present. The best ear of each 

 variety grown in 1911 was compared with the original ear of that 

 variety, and it was shown that not one of these had changed their 

 characters one little bit ; so that more than half a century of selec- 

 tion among the descendants of a pure individual has no eff'ect 

 whatever*. 



From all this but one conclusion seems to me possible: — 



A pure individual, self-fertilized, has a descendance absolutely 

 identical ivith it, as far as hereditary properties are concerned. 



This fundamental fact I will call Johannsen's principle. We 

 nurst now ask: Does every self-fertihzed individual produce so 

 uniform a descendance ? We must answer: Xo ; there are self- 

 fertilized individuals which give a multiform descendance, uamely, 

 the hybrid individuals. 



We must therefore conclude : — 



Pure individuals have a uniform, liyhrid individucds a multiform, 

 descendance. 



This seems strange. One might conclude from it that pure 

 individuals transmit their properties as an inseparable whole, 

 while hybrid individuals do not. Is there really so fundamental 



* This statement was based on the communication by Mrs. C. and ])r. A. 

 L. Hagedoorn, " Selection in pure Lines," Am. Breeders' Magazine, vol. iy. 

 )). 165 se§. This communication, however, is erroneous, as Mr. Ph. de Vilinorin 

 pointed out ; one did not select the seeds from one plant, the best, but from 

 several plants together, which were the most typical ones. 



