THE PURE LINE HYPOTHESIS. 547 



Thus in eight famihes ranging from 2-6 (average four) in 

 number the mean number of good pollen grains in the parent was 

 31%; in 10 famihes from 9-15 (average 12) the mean 

 number of good grains was 71% ; while in three families with a 

 mean number of 20 offspring the average number of good grains 

 was 97%. It is therefore clear that when the fertility of the 

 plant was low the pollen was defective in quality. 



Accordingly it may be stated that on self-fertilising these 

 hybrids it was found that the number of offspring that could be 

 raised from each of the 21 plants varied considerably, and on the 

 average in those parent plants where the pollen was largely bad, 

 the number of offspring raised was low. This poor quality of 

 the pollen is to be considered as a concomitant effect of the 

 hybridism, and not in itself a cause of the fewness of the off- 

 spring that could be obtained, since with 30% of good grains 

 there would be an enormous superfluity of such with the slightest 

 smear on the stigma. It implies that where the pollen was of 

 poor quality the ovules were also weak or in some way defective. 

 Thus in selfing the hybrids there was partial sterility arising from 

 (i) the effect of hybridism, and (2) the bad effect of self- 

 'fertilisation ; and with the data to hand it is not possible to 

 separate sharply the effects of the two unfavourable influences. 



7. The Nature of the Inheritance of the Various Characters. 

 (i) General Hubit of Plant. 



The original seed-plant (Tropccolum majus) was tall or 

 straggling, and the original pollen-plant (Tropa:olum minus), 

 variety known as " Queen,'' was dwarf. 



The parent plants appeared to be homozygous with reference 

 to this character; since on self-fertilising the tall plant. 20 off- 

 spring were raised, and they were all tall ; and on selfing the 

 dwarf plant, seven offspring were obtained, and they were all 

 dwarf. 



By crossing the two plants 21 offspring, or hybrids, were 

 raised, and they all tended to be tall. 



The offspring resembled the seed-plant, and consequentlv in 

 Mendelian phraseology the straggling habit is to be regarded as 

 dominant, and the dwarf habit recessive. It was not easy to deter- 

 mine whether the straggling habit of the offspring was in any 

 way less pronounced than in the seed-plant. On the whole it 

 appeared to be rather less : and even in the absence of a special 

 investigation of the point it is probable that dominance was not 

 perfect. The habit could not, however, be regarded as a mean 

 between the habits of the two i)arents ; the offspring were cc"- 

 tainly much closer to the seed-plant than to the pollen-plant in 

 this character. 



By self-fertilising these hybrids 21 families were obtained 

 and 200 seedlings were raised. Of these, on account of death and 

 other causes, it was possible to observe the habit in only 180 

 plants. Three grades in the habit were distinguished — dwarf, 

 semi-dwarf, and the straggling or tall. These grades were not 



