256 University of California Publications in Botany [Vol. 5 



normally matured seed. From 200 to 300 .seeds were placed in a 

 watch glass and covered with strong Eau de Javelle. In twenty-four 

 hours the characteristic dark brown color of the seed-coats was bleached 

 out and when placed on a dark background the question as to whether 

 or not the seed-coats were empty could be answered at a glance. In 

 every case a control was made by similarly treating the selfed seed of 

 "Nic. tahaccum Cuba" obtained this past season. This control served 

 to show that the length of time for which the bleaching solution was 

 allowed to act was not sufficient to dissolve the endosperm and embryo 

 and served also as an aid in distinguishing between the normal and 

 abortive seed produced by the treated flowers. Although the chalky 

 white seed containing endosperm could be distinguished at once from 

 the transparent functionless seed, quantities of the bleached seed were 

 examined under magnification in every case. The slightest trace 

 of endosperm could thus be detected and by gentle pressure on the 

 cover-glass over the preparation the embryos of the normal seeds 

 could be forced out of the endosperm through the softened seed-coats. 

 In a number of cases seed, which according to the bleaching test con- 

 sisted of empty shells only, was germinated as a check upon the 

 method but no germination was noted at the end of three weeks. All 

 seed packets which were indicated by the bleaching test to contain 

 seed with endosperin or endosperm and embryos were also germinated. 

 In every case all the seed remaining in the packets was sown accord- 

 ing to the method elsewhere described (Goodspeed, 1913). 



The following table indicates so far as possible all significant 

 facts concerning the 112 treatments of flowers which yielded abortive 

 or normal seed, together with the proportions of seed with endosperm 

 and embryos and seed with endosperm only. 



