90 



set, when the last flowers closed, were spent in a continuous round 

 from tree to tree. Less continuous observations to determine cer- 

 tain points of interest were made on subsequent days. The eight 

 trees fell into three well-marked groups. 



I. Two trees growing side by side agreed essentially in their 

 floral behavior. Both belonged to class A. The fresh flower buds 

 began to open in the morning at 9 :20, and by 9 :30 some were com- 

 pletely expanded. The stigmas were fresh and apparently receptive, 

 the filaments declined and the anthers closed. Thus they remained 

 during the remainder of the morning. By 12 noon some flowers 

 began to close ; by 12 :30 many were closed and others still closing. 

 The flowers then remained closed for slightly more than 24 hours, 

 or until about 1 p.m. the following day, when they began to expand 

 for the second time. The stigmas of most had withered but in a 

 few cases they still appeared fresh. The anthers began to shed their 

 pollen about 15 minutes after the flowers opened. At 4:40 p.m. 

 the flowers began to close, at 5 they were about one-half closed, 

 and at 5 :20 completely closed. 



These trees, especially the smaller, set a fair quantity of fruit. 



II. Two large trees belonged to class B. In both the flowers 

 opened for the first time between 2 :30 and 3 :20 p.m. and remained 

 in anthesis with receptive stigmas during the latter part of the 

 afternoon, different individuals closing for the night between 4 :45 

 and 6 p.m. They opened the following morning, some with stigmas 

 still fresh in appearance, between 7 :30 and 8, and having shed their 

 pollen, made their final closing between 1 :10 and 2:30 p.m. 



One of these trees, when I left Almirante in June gave promise 

 of a fair crop, while the second bore only a few green fruits. 



III. Four of the trees, while giving indications of floral be- 

 havior which would place them in class B, were extremely erratic 

 in their anthesis. A small percentage of the flowers opened sporadi- 

 cally during the late afternoon, between 3 :30 and 4 on different 

 trees, and remained open for less than two hours, for by 5 :30 or 

 5 :45 all were again tightly closed. The degree of spreading of the 

 perianth of these flowers was variable; many opened only very 

 slightly, few of them spread fully in the normal manner. The fol- 

 lowing morning, between 7 :30 and 8, all of those flowers which 

 showed any degree of opening on the previous afternoon, together 

 with many which now for the first time expanded, opened with 

 stigmas still fresh, and soon the valves of the anthers lifted and 

 allowed the pollen to escape. These flowers now remained open for 



