SEED COAT COLOR IN GARDEN BEANS. 61 



had taken place before the foreign pollen was introduced, or at least before 

 it could take effect. Some of the crossing was done in the field and some 

 in the greenhouse during the winter. In the latter case the blossoms were 

 not covered, while in the former a one-fourth pound manilla bag was tied 

 tightly over the flower stalk for five or six days, after which it was torn 

 open, and, if the attempt seemed successful, left to indicate the seed pod 

 at the time of harvest. 



In all cases four generations from the cross have been grown. In each 

 generation except the fourth a certain number of plants chosen more or 

 less at random have been self-fertilized by enclosing them during the 

 blossoming period in cheesecloth or, in a few cases, waxed paper sacks. 

 Neither of these is satisfactory. Both weaken the plant, the waxed paper 

 sacks more than the cheesecloth ones. It has been the invariable observa- 

 tion that there is a progressive weakening of the plants through the four 

 generations. First generation crosses are invariably strong, vigorous 

 plants, often seemingly more vigorous than the parent varieties, while the 

 fourth generation plants are decidedly weak and unproductive. Whether 

 this is due to the repeated self-fertilization or to the weakening effect of 

 covering the plants does not appear. Possibly both have contributed to 

 the result observed. 



The blossoms have been more or less infested with thrips. None have 

 been observed on covered plants, but it is entirely possible that there 

 may have been cases of such infestation, and that in rare cases a grain of 

 foreign pollen was introduced in the blossom of a plant supposed to be 

 self-fertilized. A few irregularities that may have been due to such a 

 cause have been observed. Nevertheless, the probability that such cases 

 are extremely rare is indicated by a number of considerations. Bean 

 blossoms are commonly self-fertilized before they open, and, according 

 to our observation, thrips does not infest unopened buds. The pollen- 

 carrying ability of thrips cannot be large, and it appears that it does 

 not commonly enter beneath the bags, or it would have been observed 

 in the many examinations of the covered plants. And finally, cases 

 arousing suspicion of the entrance of foreign pollen are extremely rare. 



Recording Data. 

 The method of securing records of the plants has been previously 

 described (16). It consists, essentially, in assigning to the exjDression of 

 each supposed Mendelian character a special letter designation. The 

 plants have been examined for blossom color, pod color, and for seed 

 coat color. This involves going over the plants not less than three times, 

 and in most cases two examinations have been made for each character, 

 involving examination of the plants six times. In order to identify the 

 individual plants, each is assigned a number in order. The seeds are 

 planted about 6 inches apart, and a small tag bearing its number is early 

 attached to every fifth plant. Thus, in order to ascertain the number of 



