HOWARD, HOWARD AND ABDUR RAHMAN. 311 



tunities for natural crossing were therefore possible at Piisa for 

 two seasons only, and these opportunities were limited by the fact 

 that but few plants of each kind were allowed to flower and the 

 period of flowering of the various kinds did not in all cases overlap. 

 Tn 1908, observations were made on the uniformity of the cultures 

 raised from one parent plant and any individuals which varied 

 from the rest in the slightest degree were noted. In every case 

 these aberrant plants were selfed and grown the next year when 

 they gave rise to a large number of forms, often very diverse in 

 habit and in leaves and which were obviously the products of natural 

 cross-fertilization. In over a hundred cultures of this species 

 in 1908, about 20 per cent, of the rows contained aberrant plants, 

 thus showing to what a great extent natural crossing is possible 

 when the various kinds are grown in close proximity and allowed 

 to flower freely. Some of the aberrant plants only differed very 

 slightly from the rest of the culture and much care was necessary 

 to distinguish them. Sometimes they appeared only a little 

 more robust than the rest, but in the case of every plant which 

 showed the slightest variation in 1908, its progeny after bagging 

 gave rise to uneven cultures made up of many different types. 



Too much emphasis cannot be laid on the difficulty of observ- 

 ing the first generation of a cross between two varieties of tobacco 

 in the field. The differences between many tobacco types are ex- 

 ceedingly small and can only be detected with precision after con- 

 siderable study in places where such types are grown side by side 

 and allowed to flower. Crosses between types unless they are very 

 different would not be easy to observe in the Fj generation even in 

 cultures from single plants. In the field, where the great majority 

 of the plants are topped, these differences would, in all probability, 

 be overlooked. In the second and succeeding generations, when 

 s])litting takes place, the various types resulting from crossing 

 could for the first time be observed. 



In great contrast to tlie variable character of the progeny of a 

 selfed aberrant plant is the great uniformity of the produce raised 

 from the seed of an ordinary plant raised under bag. 



