180 



MEMOIRS OF THE XEW YORK BOTANICAL GARDEX 



EXPLANATION OF PLATE 27 



Tpper photo. Flowers of the Trapp variety for which the normal 

 cycle of dianthesis of each flower and the daily sequence of sets are as 

 follows : 



A set of flowers begins to open for the first time late in the aftenioon 

 — usually not earlier than 3 : 00 P. >\L Then they appear as shown at 2' 

 on the branch at the left. An hour or so later the flowers are as seen 

 at 2 on the branch in the middle. These flowers close during the night 

 and then the same flowers open for their second or male opening the fol- 

 lowing forenoon, when they are as at 4 on branch to the right. Only one 

 set of flowers is in action at a time and there is no overlap of sets with 

 opportunity for close-pollination. The opening for the first time late in 

 the afternoon almost prohibits cross-pollination from any of the A varie- 

 ties (see CHARTS 1 to 4) throughout most of the blooming season. But 

 the cycle of sets of flowers is much shorter than that of A varieties 

 and of some B varieties (see B 1, chart 5) so it seems highly probable 

 that fruit may set on Trapp to self-pollination wliile flowers are shedding 

 pollen. 



Lower photo. The flowers of the Collinson variety have two periods of 

 opening quite as in other varieties. Stamens are present but they shed 

 no pollen. 



At the left is a flowering branch of this variety with four flow^ers fully 

 open for the second opening. At the right is showTi somewhat enlarged 

 a flower of Collinson (left) by the side of a second-period flower of Taylor 

 (right) as the two appear in the afternoon. The uplifted valves carrying 

 out 7")onen may be seen as fingei'-like projections from anthei's of the 

 Taylor flower. No such valves open from anthers of flowers of Collinson. 

 The anthers remain intact and studies of their structure show them to be 

 merely masses of sterile ti.ssue. The Collinson variety is male sterile. 



