494 WHITE— STUDIES OF INHERITANCE IN PISUM. 



is ordinarily unnecessary. The few recorded changes of chance 

 crossing are probably due to the pea weevil {Bruchiis pisi) (60, 78) 

 or to thrips (3, 78). In case the pollen of a flower is ineffective, the 

 stigma may extend itself beyond the keel and chance crossing come 

 about in this way. No cases of the latter type are recorded and the 

 possibility of error from this source is rare (60). 



The source of error from chance crossing in a locality may be 

 tested out by growing several hundred plants of a variety with 

 green cotyledons side by side with a row of a pure yellow cotyledon 

 strain. When the seed of the green cotyledon strain are mature, the 

 per cent, of crossing can be calculated from the number of yellow 

 seeds found on the green-seeded plants. In an examination of 

 over 10,000 seeds of several green-seeded varieties at the Brooklyn 

 Botanic Garden, not a single case of cross fertilization came to light. 

 Bateson and his students (5), Messrs. Sutton (5), Tschermak (81) 

 and Mendel (60) each record a few cases, the per cent, in each case 

 being much less than ^ per cent. The few non-conformables in 

 Lock's experiments (54) on cotyledon color are attributed by him 

 to errors in labeling, planting and to improper maturing. 



Because peas are naturally self-fertilizing, pure lines may be 

 selected from almost any of the commercial varieties with the as- 

 surance that they will be constant as regards visible characteristics 

 and relatively free from heterozygosis almost at once. Most of the 

 varieties at the Brooklyn Botanic Garden have given constant strains 

 after at most two years of selection, while the great majority were 

 constant from the start. In judging constancy, only characters such 

 as flower color, seed shape and color, foliage color and shape of 

 pod, which are but slightly influenced by small environmental 

 changes, were used. Tschermak (78), Macoun (57.5), Hurst (42), 

 Sherwood {72), Knight (50), Darwin (22) and many other experi- 

 menters have often remarked upon the exceptional constancy of 

 pea varieties. It should be noted that pea varieties commonly grown 

 for forage purposes have generally become very much mixed 

 mechanically with each other as well as with various vetches through 

 carelessness in handling and harvesting. Often it is possible to 

 select ten or more constant varieties from a handful of such seed. 



Labeling and Recording. — The system of labeling used in the 



