91 



a long period, about six hours, and did not close until between 

 1 and 2 p.m., when many of the stigmas were still unwithered. 



I made observations to determine what proportion of the flowers 

 opened partially or completely on the afternoon before expanding 

 for what would normally have been their second period of anthesis 

 on the following morning. At 4:20 p.m., February 18, the middle 

 of the desultory first period of anthesis, I marked a twig on tree 2 

 which bore ample clusters of buds but only a single open blossom. 

 On the following morning over 100 flowers on this twig were open 

 and shedding pollen. Other twigs on this tree and its neighbor 

 tree 3 showed a similar low proportion of the flowers with two 

 periods of anthesis. The other two trees in this group had a larger 

 proportion of flowers which opened in the afternoon, but still well 

 under 50 per cent, of the total number. Thus on February 18, at 

 4:10 p.m., 25 flowers on a marked twig of tree 4 were more or less 

 open. The following morning at 10 :40 this twig supported 69 

 blossoms in full anthesis. 



It occurred to me later that all, or at least a much greater pro- 

 portion of the flowers than the observations just recorded would 

 indicate, might have a very transitory afternoon opening. If such 

 were the case, and the flowers did not all open at the same time, a 

 single examination even in the middle of the first period of an- 

 thesis would not reveal their true numbers. Accordingly, on Feb- 

 ruary 25, certain chosen branches on these trees were visited at 

 twenty-minute intervals, and every flower which showed any de- 

 gree of opening was tagged. A single example will suffice. At 

 1 :10 p.m. the chosen branch on tree 1 bore 107 open flowers in 

 their second period. Since the sky was overcast most of the day, 

 the closing of these flowers was delayed ; it began at about two and 

 was not completed until three o'clock. By 4:15 p.m. a few first 

 period flowers had opened on other accessible branches of the same 

 tree, but none on the selected branch. Between 4:55 and 5 :15 one 

 flower on this branch opened halfway. It never expanded beyond 

 this point and by 6 :25 was practically closed again. Yet next morn- 

 ing this same branch bore 49 open flowers, including that which had 

 partly opened on the preceding afternoon. Similar observations on 

 trees 3 and 4 gave comparable results. On the succeeding after- 

 noon, which was bright and warmer, a larger number of flowers 

 opened on these trees, but still a small percentage of those which 

 expanded the ensuing morning. 



Although in a large proportion of the flowers of these four 



