264 TWENTY-FIRST REPORT. 



this kind liave been called intergrading Rabbit-Ear rogues, to distinguish them 

 from the normal Rabbit-Ear rogue, which can be recognized as such at all 

 stages of development. The progeny from b<tth normal and intergrading 

 Rabbit-Ear ntgues has been in all cases normal rogue. II is of interest to 

 know that in the three cases mentioned the intergrading i-ogues occurred in 

 progenies of typical plants. No Rabbit-Ear rogues have ever been observed bj' 

 the author to give rise to any type-like plants. 



EXPERIMENTAL METHODS. 



In growing pure-lines of either type or rogue plants the tlowers were not 

 ordinarily protected fi-om cross pollination. Among the thousands of typical 

 plants grown under observation, only one case of uncontrolled crossing in the 

 field has been recognized. With the exception of the rogues liere de.scribed 

 the amount of cross-fertilization which usually takes place in peas varies 

 considerably, according to the observations of different workers. White ( 1 ) 

 examined 10,000 green-seeded peas grown at the Brooklyn Botanical Garden 

 and found no evidence of any of them having been crossed with a yellow- 

 seeded variety that was growing in close proximity. Opposed to the expe- 

 rience of White and of the present writer is that of Mr. George Starr of the 

 Rice Seed Company. Mr. Starr has informed me that in growing pure lines 

 of Gradus derived from single plant .selections he found in one season that 

 the percentage of crossed plants was very high. No exact figures were avail- 

 able, but the progenies of a number of plants indicated that out-crossing with 

 at least four other varieties had occurred. These results were typical for only 

 one season, and would indicate that under very unusual conditions dehiscence 

 of the anthers might be delayed until the flower had opened. If a delay of 

 this kind occui-rcd it would be possil>le for cross-pollination to lie effected by 

 bumble bees. 



In tlie ca>-c of (lie liabbit-Ear rogues (he percentage of volunteer crosses 

 seems normally to be larger than in the type, and in the future it seems 

 advisable to protect the plants during the flowering period. The difference in 

 the amount of normal cross-pollination which takes place in the type and in 

 the rogue plants is accounted for by the smaller flowers and the occasional 

 failure of the rogues to produce good pollen or to shed their pollen before 

 the flower opens. 



In making crosses the usual procedure in carrying on artificial pollination 

 in peas wa.«5 observed, with the additional precaution that in neai-ly all cases 

 the crossed flow(>rs were protected by a small gauze bag. The pollen on the 

 keel of an iniopened flower of the staminate parent was applied to the stigma 

 directly aflcr emasculation of the flower of the pistillate parent. The forceps 

 u.se<l were washed in U't'/, alcohol after each operation. The plants used in 

 making crosses were descend«'d fi-om parents whose progeny had l>een under 

 observation at least two generations. The variety most intensively studied has 



