366 TEXTBOOK OF BOTANY 



heads on the average are large, though there will be differ- 

 ences in size due to external conditions ; but the grains from 

 the plants of other races will not, on the average, produce 

 large-headed plants. So, although the average size of head 

 may have been increased by selection, the variety is still 

 mixed because it contains both large-headed and small- 

 headed races. If a new selection of the same sort is made 

 each year, the average of the variety may slowly be_raised 

 for a number of seasons, and afterward kept up by continued 

 selection ; but we are not likely to produce in this way a pure 

 large-headed variety. 



A pure variety may be obtained, however, by using the 

 more modern method of selection. This consists in select- 

 ing, as before, the grains of several or many large-headed 

 plants, but in sowing those from each plant in a separate plot. 

 The offspring of each original plant are kept separate, and by 

 studying the offspring for one, two, or more generations it 

 is possible to determine which of the plants that were first 

 selected really belonged to large-headed races and which did 

 not. Then, by breeding from the descendants of that origi- 

 nal plant which passed on in the highest degree the tendency 

 to bear large heads, a variety is established which consists 

 of a single race; it is a thoroughbred variety and has the same 

 advantages that a thoroughbred variety of horses or cattle 

 has. When such a variety is obtained, it is less likely to 

 deteriorate than a variety obtained by the older method, 

 which is almost certain to be mixed. However, as we shall 

 see, selection cannot be entirely neglected even after a pure 

 variety has been established. ^ 



1 Strictly speaking, a "pure variety," in the sense in which the expression is 

 used here, is possible only in the case of a plant which, hke many of tjie wheats, is 

 regularly self-pollinated. In species or varieties that are ordinarily cross-pollinated, 

 any plant is sure to combine within itself qualities derived from various races, and 

 with each generation new quaUties are likely to be introduced. However, even 

 in cross-poUinated species, it is possible to obtain by close inbreeding between nearly 

 related or carefully selected plants races that are approximately pure, at least with 



