INHERITANCE— IMPROVEMENT BY BREEDING 229 



The fundamental principle of the method described above CHAP, 

 is the rigid elimination of undesirable types, which may appear 

 owing to reversion, and the mating together in the breeding 

 and propagation plots of the most desirable types. 



183. The Breeding Plot should be so selected that the soil 

 will be typical of that on which the main crop will be grown. 

 The same preparation should be given as for the main crop, 

 no extra care or fertilizer being used. The object is to find 

 out which plants will give the best results under normal con- 

 ditions ; if they do well, then they may be expected to do 

 better on well -fertilized soils. 



It is absolutely necessary that the plot should be isolated 

 at least 400 yards from any other sort, or from strains of the 

 same sort, flowering at the same time. Any stray plants from 

 previous plantings must be carefully rogued out before they 

 have a chance to tassel. 



The rows should be 200 or more yards long and of exactly 

 the same length. The seed should be planted on the principle 

 of one row to each ear. After removing the tips and butts 

 from the selected ears, the rest of the grain should be planted 

 in a single row. It is better to drill the seed than to check- 

 row it, as it is difficult to isolate suckers from main stalks 

 when more than one plant occupies a place. Each row should 

 be numbered consecutively and labelled with a stake at the 

 end. With the seed left over from the rows three or four 

 border rows may be planted all round the breeding plot to 

 protect the plot to some extent from depredation. 



184. Devices to Prevent or Detect Cross-pollination. — As a 

 means of minimizing the amount of cross-pollination between 

 breeds of maize grown near to each other on similar soils, the 

 following devices, among others, have been resorted to at the 

 Botanical Experiment Station, Pretoria ; none of them, how- 

 ever, has been entirely successful. 



Planting one week and two weeks apart was tried. Several 

 of the breeds tested were new to us, and their relative time of 

 flowering in that climate was not known, so that some of the 

 later-sown* flowered earlier than those planted before them. 

 The danger of cross-pollination was minimized by bagging 

 and hand-pollination, but the number of plants of each that 

 could be treated this way was small, and the amount of seed 



