140 JJnwersity of Calif omia Publications in Botany [Vol. 5 



Here again we have a very small degree of variation in the 

 corolla diameters of the parents. The measurements were very 

 carefully taken and at times a few measurements would be made 

 upon the parents, then upon the hybrid plants, and again upon 

 the parents. It was possible in 1911 to watch the plants as they 

 first came into flower and every effort was made to follow the 

 development, from the bud to the fully opened flower, of every 

 individual flower measured either on hybrid or parent plants. By 

 this is meant that, for example, on September 28 it was observed, 

 in the case of plant 1 of the cross III JX II d^. that two flowers 

 were in a proper condition (see page 134) and they were picked 

 off and measured. On the same date and on this plant three 

 flowers were almost in the proper condition to be measured — i.e., 

 the anther lowest in the tube shedding pollen and the corolla 

 fully extended but the remaining anthers unopened. There were 

 also four flowers all with the anthers green and the corolla only 

 partially opened as well as a considerable number of rather 

 small buds. Now in this case, it was plain that the anthers of 

 the first three flowers mentioned above would be shedding pollen 

 during the afternoon of September 29. and that on the follow- 

 ing morning these two flowers would be in the proper condition 

 to be measured. It was also evident that some of the remaining 

 four fully formed buds might be properly developed by that 

 time. At eight a.m., September 30, only three flowers could be 

 measured. One of the three flowers almost ready the morning 

 before had evidently begun to shed pollen from all the anthers 

 before noon, for in this flower two of the anther cells were 

 almost free from pollen, the brownish coloration at the back of 

 the pollen receptacles was showing, and the stigma was covered 

 with pollen. In the other two the anthers were balls of light, 

 fluffy pollen and no pollen could be noticed on the stigma. All 

 three of these flowers probably could rightly have been included 

 in the measurements so far as development in corolla diameter 

 is concerned, but only the last two were considered to be in 

 exactly the proper condition and the third was picked off and 

 thrown away. Of the four flowers that had been quite unde- 

 veloped on September 29, two showed the anthers lowest in the 

 tube shedding pollen, the third flower showed all the anthers 



