150 



MEMOIRS or THE NEW YOKK BOTANICAL GARDEN 



In the forenoon, while flowers of Taylor are in the first period 

 or female anthesis, the flowers of Panchoy are all of the second- 

 period or male opening (upper photo of plate 26), Trees of Pan- 

 choy are shedding pollen during the forenoon while the flowers 

 of trees of Taylor are most ready to be pollinated. During mid- 

 day an alternation or change in sets occurs for each and then 

 in the afternoon the flowers open on Panchoy are in the first 

 period or female opening while the flowers open on Taylor are 

 all of the second period or male opening (see lower photo of 

 plate 26). These tAvo varieties are, therefore, reciprocating in 

 their daily alternations of sex. Taylor is female while Panchoy 

 is male and then while Taylor is male Panchoy is female — all 

 during the hours of a single day. 



Extending the observation to trees of other varieties grow- 

 ing side by side one finds that the daily behavior of flowers is 

 either lilce that of Taylor or similar to that of Panchoy. This 

 is readily revealed when the daily sequence in the flower behavior 

 of a number of varieties (all clonal varieties) is shown diagra- 

 matically as in chart 1. The observations there recorded are for 

 trees growing in the same orchard or in a nearby one and are all 

 for the same day, which was one of a series of warm sunny days 

 on which the flower behavior may be considered as most nearly 

 normal. 



It may be stated that in obtaining such records as are shown 

 in the various charts accompanying this report observations 

 were made from tree to tree as quickly as possible, especially 

 during the midday shift of sets. These were checked on squared 

 paper for fifteen-minute intervals. If, for example, flowers were 

 not open on a tree at 8 :05 but were open rather abundantly at a 

 next visit at 8 :25 the start of opening was credited to 8 :15. If 

 at 12:20 first-period flowers were rapidly closing and at 12:40 

 they were all closed the time of closing was checked for 12 :30. 

 Thus in the charts there is rarely an error of 15 minute_s, and for 

 the majority of cases the start of opening and the end of closing 

 of sets is given within five minutes of the correct time. 



This record (see explanation of chart for symbols used) is 

 typical of that which may be obtained day after day during 

 favorable weather (see also charts 2, 3, and 4). Sucli records 



