Manchester Memoirs, Vol. liv. (1909), No. 3- 5 



determined by chance. It is rarely in the same direction 

 on each of the five stamens in the same flower. I have 

 also noticed that in some cases no such rotation occurs, 

 especially in dull or cold weather. On those days in 

 June and July upon which we had bright sunshine and a 

 temperature of about /S'^F. in the greenhouse, the move- 

 ments above described from the opening of the petals 

 took place in about a couple of hours, and during this time 

 the filaments continued to move in a radial plane, so as 

 to make an angle of 90^ with the gynophore. 



There appears to be a fairly obvious reason for the 

 second movement of the anther. Pollen is transferred to 

 the head of the insect visitor as it moves from ray to ray 

 of the outer circlet of rays around the circular opening of 

 the nectar chamber. Since the anthers are in a tangential 

 position, it follows that the insect's head will come into 

 contact with some portion of an anther at every stage 

 throughout its progress. The function of the locking 

 mechanism is also apparent, for, otherwise, the push of 

 the insect's head would tend to replace the anther in the 

 radial plane, but by the system of knobs, this must be 

 prevented to a large extent, or, should it occur, the 

 stamen would automatically return to the tangential 

 position. 



At the conclusion of the first stage of anthesis the 

 stamens rapidly fall, thus completing a movement of 180° 

 for the filament and anther together, and the styles move 

 downwards, so that the receptive stigmas can dust the 

 pollen from the heads of the insect visitors (Sprengel, 

 Miiller, and Knuth). It is this movement of the stamen 

 as a whole through 180'' in the vertical plane of its 

 filament, that I take to be the first of the two movements 

 described by Warnstorf. If he meant that movement by 

 which the anthers become extrorse, I am at a loss to 



