112 Journal of Agricultural Research voi. xvii. no. 3 



retarded but did not prevent anthers from opening. This shows clearly 

 the relation of dehiscence to water. 



The fact that empty anthers close during a rain and open afterwards 

 probably has been the basis for the popular conception that rain washes 

 pollen away. 



With this statement, then, of anther action in relation to water, the 

 question arises as to what extent rain removes pollen from anthers which 

 have just dehisced. In investigating this point a branch of flowers was 

 brought into the laboratory, and after the anthers opened it was stirred 

 vigorously for 8 minutes in a pail of water. All anthers closed com- 

 pletely during the time of stirring. The larger part of the pollen lost 

 occurred with the first impact with the water. After this treatment it was 

 estimated that those anthers which were open before being put into the 

 water still contained, when they opened again, from one-quarter to two- 

 thirds of their pollen. These results agree with observ^ations made in 

 the orchard both during and after a rain. 



The effect of rain in washing pollen away, even in the quantity noted 

 above, is partly modified by the unevenness of anther opening, there 

 being in some cases as much as 3 days' difference between the first and 

 last opening of anthers. The unopened anthers have a light yellowish 

 color in contrast to the water-soaked appearance of those which have been 

 closed by rain. 



These observations show that anther action is a reversible process and 

 is controlled by water. The presence of the anther sap until the maturity 

 of the pollen creates an internal condition unfavorable to dehiscence. If 

 dehiscense takes place only after sufficient drying, there must be an inter- 

 nal control of water as well as a means for external loss. These two con- 

 ditions are met by a breaking of the epidermis at the suture and by the 

 drying or death of the cells of the filament at the point of union with the 

 anther where there is a pronounced constriction of the filament. At this 

 point the cells typically turn brown before dehiscence, a condition which 

 suggests an early cutting off of water. The browning slowly extends 

 down the filament and at the time the petals fall, 3 to 4 days after bloom- 

 ing, the filament is dead for a distance of i to 2 mm. 



Under some conditions pollen is shed more quickly than under others. 

 When anthers of Surprise were allowed to open in a dry, still room at 

 about 72° F., at the end of four days pollen had not been shed except in 

 very small amounts. This was due partly to the adhesive action of a 

 yellowish, oily substance about the pollen grains which is characteristic 

 of some varieties, and partly to the absence of w4nd. The persistence of 

 pollen is further shown by specimens of Surprise gro\Mi in the green- 

 house, which, at the time of abscission of the calyx tube, 10 days after 

 blooming, still had an abundance of pollen present. But in some 

 varieties with sticky pollen, under orchard conditions as much as one- 

 half may still be present at the time the petals drop. On the other hand, 



