566 THE PUKE LINE HYPOTHESIS. 



In the offspring of the hybrids great diversity in the mean leaf- 

 widths of the individuals occurred, the means ranging from 20- 

 91 mm., while in the hybrid parents the range of the means was 

 only 50-84 mm. There was some sign of an imperfect segrega- 

 tion into plants with small leaves and plants with large leaves, the 

 break in the series tending to occur in the neighbourhood of a 

 leaf-width of 40-50 mm. Among these offspring of the hybrids 

 there was a definite tendency for the three characters, dwarf habit, 

 variegated leaves and small leaves to occur together in the same 

 plant (i.e., there was a strong correlation of these characters), 

 but they were not of necessity combined, in that variegated, strag- 

 gling plants with large leaves and also non-variegated, dwarf 

 plants, with small leaves, did sometimes occur. 



17. Chromoplasts. — There was clear discontinuity in the 

 iiiheritance of this character. The original seed-plant had no 

 chromoplasts on the general surface of the petals, but on self- 

 ifertilisation two offspring had deep yellow flowers, owing to the 

 presence of abundant chromoplasts throughout the whole surface 

 of the petals, while 18 oft'spring resembled the parent. The 

 original pollen-plant had no chromoplasts on the general surface, 

 but it produced four oft'spring with chromoplasts and two like 

 the parent. The 21 hybrids included nine plants with chromo- 

 plasts and 12 without. The hybrids which had no chromoplasts 

 produced only individuals in which chromoplasts were absent or 

 scarce, but the families of the hybrids which had chromoplasts 

 included 50 individuals with chromoplasts and 16 without, and 

 these figures are close to tlie 3 : i ratio. Further breeding is 

 necessary in order to determine fuHv the nature of the inheritance 

 of the character, but it is obvious that segregation is well-marked. 



18. Red Colouration. — -The redness of the flowers appears to 

 be due to the action of an oxidising agent on the pale yellow 

 chromogen ; but sometimes the oxidising agent, although present, 

 is prevented from acting by the presence of one or more inhibi- 

 tors. Accordingly we can understand how it is that some yellow 

 plants on being sel'f-fertilised are capable of producing both 

 yellow and red oft'spring, while others can only produce yellow 

 offspring; and likewise some red jilants can produce both red and 

 yellow offspring, and others are able to produce only red plants. 



iq. The seed-plant was pale yellow, and on selfing produced 

 three red offspring and 17 non-red. The pollen-plant was red, 

 and produced on selfing seven red and no non-red offspring. The 

 hybrids of the above two plants included 20 red off'spring and one 

 non-red. Among these red plants the intensity of the red coloura- 

 tion varied very widely. A scale of colour intensity was formed, 

 and to each plant a grade was assigned, expressed as a percentage. 

 The range of colour intensity in the 20 red hybrids was 6-73 

 per cent. Families were raiscfl from all 21 hybrids. The single 

 non-red hybrid producerl onlv non-red oft'spring. Of the other 

 hybrids 14 produced families containing both red and non-red 

 offspring, and six formed families including onh' red offspring. 



