226 THE LAW OF JOHANNSEN. 



else. We have seen that it is possible to show, that the viable 

 seeds produced from unfertilized female flowers of squashes, 

 are grown out of real fertilizable germ-cells. Hoterozy- 

 gous plants produce unfertilized seeds, which by segregation 

 show to develop out of real gametes. Such seeds must necessar- 

 ily produce homozygous individuals, and the very pure lot of 

 plants produced from one such an individual is theoretically 

 the only really dependable "pure line" according to Johann- 

 sen's terminology. 



On the other hand, experiments with clones should be un- 

 dertaken with an understanding of what somatic segregation 

 may do to the material, if the members of a clone are hetero- 

 zygous. There are many plants, that could be chosen as mat- 

 erial. If one starts selection-experiments in a clone, which 

 starts from a homozygous plant, he may be sure, that he has 

 now only loss-mutation to watch out for in an interpretation 

 of his results. Wheat may be propagated asexuaUy, and so 

 may barley, and beans can be propagated by cuttings. It 

 should not be difficult to devise a means of propagating such 

 plants as the Cupid sweet-pea vegetatively. The results would 

 be more free from criticism than those of any selection exper- 

 iments in pure lines or in clones, that we have knowledge of. 



