1890.] NATURAL SCIENCES OF PHILADELPHIA. 267 



ward growth of tlie stigma, that scarcely any is left in the antlier 

 cell, and yet so delicately is it done, that the five Imes of pollen, as 

 cleaned out from the five anthers, are generally discernible. Nature 

 could not have made a better arrangement for making cross-fertili- 

 zation impossible. 



All attempts to ascertain the precise time of the opening and 

 elongation of the style failed. Though I made several successive 

 attempts with fresh flowers, none opened up to 11 P. M., unless 

 touched with the point of a pin, when the limb would part with an 

 elastic sort of spring, becoming expanded so instantaneously as to 

 defy the eye to detect the motion. In the morning, two or three 

 hours after sunrise, the room, however, being "pitch dark " from 

 shutters, a number of newly opened flowers, with their pollen- 

 covered stigmas would be found, but not with the styles much ex- 

 serted. The elongation seemed to be done gradually through the 

 day, not reaching full length till night-fall. 



Note on Unisexuality in Connection with the Order 

 OF Flowering in Willows. 



After the flower buds have been formed and in many species of 

 plants have reached a certain size, they remain at rest until the 

 axis has reached its full growth, when this has been accomplished 

 the flower buds resume activity. In the well known case of 

 Liatris, and indeed in most allied genera of Compositie, the renewed 

 development of these flower buds is from the apex downwards. 

 In other cases, some Fumariacere for instance, the lowest bud on 

 the spike or raceme starts the renewal. In most willows the re- 

 newal of growth is from the middle of the branch. If, for in- 

 stance, there may be eleven catkins to be produced from as 

 many axils along the stem. No. 6 will be the first to expand, 5 and 7 

 next, then 4 and 8, 3 and 9, 2 and 10, and finally 1 and 11. 

 Singularly enough, this order does not extend to the catkins them- 

 selves. In them the growth ceases as soon as the anthers reach their 

 full size, and the stamen remains at rest until the axis of the catkin 

 has reached its final length. Then the filaments are formed and the 

 perfect stamen assumes its full proportion but gradually from the 

 apex dow^nward. 



It seems to me still more evident that, as I have already pointed 

 out,' this arrest of growth at a certain period and subsequent resump- 



1 Proc. Acad. Nat. Sci., Phila., 1883, p. 85; 1884, p. 117; 1885, p. 117. 



