FUTUEE OF THE SCIENCE OF BEEEDING 165 



more ttan one-quarter of the red rounds. In other words-, 

 the factors for purple and for long are linked together so that 

 the two classes of gamete, purple long and red round, are 

 about seven times as numerous as the two classes purple round 

 and red long. Without this complication the four classes of 

 gamete would of course have been produced in equal numbers. 

 It will be noticed that the two types of gamete which pre- 

 ponderate among those formed on the first-cross plant are the 

 two original parental types — the new combinations purple 

 round and red long being in the minority. Now make the 



Purple, 1 ^ f "Rad 

 "Round] I [Long 



PuTpie LoTvq 



FiQ. 5. — For explanation see text. The numbers in parentheses indicate 

 the proportion of the four forms which would be expected if the case were 

 one of simple MendeKan nature. The line of figures below gives the actual 

 proportions of the four forms as found by experiment. 



cross the other way about, crossing a red long with a purple 

 round (Fig. 5). The hybrids, or F^ plants, are purple longs, 

 indistinguishable in appearance from the purple longs which 

 arise from the union of a purple long and a red round gamete ; 

 nevertheless these purple longs when bred from give an entirely 

 different kind of family. The four types are all present, but 

 the purple longs are about twice as numerous as the red longs 

 or the purple rounds, while the red rounds are extremely 

 rare, less than J per cent of the total. This result can be 

 explained on the assumption that the purple round and the 

 red long gametes arising from the first-cross plants are about 

 seven times as numerous as the purple long and the red round 



