24 THE PLANT WORLD. 



must be on the alert for every possible source of contamina- 

 tion of pollen and of seed, or of mistaken labels and rec- 

 ords. Briefly the method I am following at the Station for 

 Experimental Evolution is as follows: 



Seeds of individual plants, whose pollination has been con- 

 trolled when that has been practicable, are collected and to 

 the seeds of each such individual a pedigree-number is as- 

 signed, giving reference to a record, or description of parent 

 and parentage. These are sown in earthen pans of sterilized 

 soil at such time that the offspring may be expected to reach 

 maturity at the close of the first season. For most biennials 

 and perennials this planting must be done before the ist of 

 March, while the annuals are not sown until about the ist of 

 May or later. When sufficiently developed, the young seed- 

 lings are set in 3-inch pots, each of which receives a wooden 

 label bearing the pedigree-number of the mother plant. As 

 a rule no selections are made in removing from the seedpan. 

 If all are not to be reset the requisite number is secured by 

 taking all consecutively, beginning at one side of the pan, 

 until the desired number has been removed, after which the 

 contents of the pan are dumped. 



At the proper season such as are destined for out-of-door 

 culture are reset to the garden, and those which are to be 

 used for indoor culture may or may not be reset to larger 

 pots, as the needs of each particular case requires. A label 

 once attached to a plant accompanies it throughout its life, 

 and is copied on the seed packet which contains its seeds. 

 The chance of error is slight after the plants have been 

 potted singly and each has received its appropriate label. 

 The most critical points in the technique are the control of 

 fertilization, the sowing of the seeds, and care of the seed- 

 pans. These matters, together with the question of labels 

 and records, may be discussed with profit to others engaged 

 in pedigree work, or contemplating taking it up. 



In regard to the control of fertilization the general 

 method consists (a) in preventing self-fertilization or cross- 



