ly^^ Harris. 



this central problem to become too much in\ol\etl in a discussion of 

 hypotheses concerning the nature of selective elimination. 



Second. Extensive data bearing on \arious problems in Stapliylea 

 were not yet sufficiently ground in the mathematical mill to show 

 their full significance. I had, however, done enough with them to 

 answer to my own satisfaction the chief objections which might be made. 



Third. I had hoped to supplement the available material with 

 more, collected with special reference to the needs which had made 

 themselves apparent during the analysis of the data already tabled. 

 In this, I have been disappointed. A large series of countings to be 

 carried out dsaring the course of development of the ovaries was 

 planned for the spring of 1910. The warm weather came unusually 

 early at the Missouri Botanical Garden and the Staphylca material 

 appeared to be in the best of condition, but the work had just been 

 started when a severe cold wave brought it summarily to a close. 



Under these circumstances, it seems best to publish certain parts 

 of the data already analyzed. 



The object of this paper is to determine whether the selective 

 elimination of ovaries which occurs between the time of flowering 

 and the maturing of the fruit can be accounted tor merely by a 

 differentiation in the ovaries, due to position on the inflorescence, 

 coupled with a higher percentage of failure in more distal regions of 

 the inflorescence. 



II. Material and Methods. 



The precision to be hoped for in any kind of biological work is 

 largely dependent upon the nature of the material in hand. 



The inflorescence of Stapliylea, like every other plant organ, opposes 

 certain limitations to accuracy of results. In the mature inflorescence 

 it is impossible to determine the number of flowers formed. This can 

 be done fairly well in the young inflorescence i), but after the axis 

 becomes dried it is difficult to make out the number of fallen ovaries. 

 The position of the fruits on the inflorescence can be fairly well as- 

 certained from the persistent bases of the composite primary branches 

 of the main axis. Stapltylea would not be selected as particularly 

 desirable for special investigation of such relationships as the cor- 

 relation between the number of flowers produced, or position on the 



1) See a table showing the correlation between the number o£ flowers formed 

 and the number of fruits developing, in Biometrika. vol. VI, p. 440, 1909. 



