Dec. 6, 1915 Inheritance of Length of Pod in Certain Crosses 409 



bean, on the one hand, and those of the Lyon bean, on the other. But, 

 according to the special investigations of Strasburger and his coworkers, 

 only a sperm nucleus without cytoplasm passes from the pollen tube to 

 the egg cell in most angiosperms. If this is the case here, the progeny of 

 the Florida velvet bean X Lyon bean receives cytoplasm only from the 

 Florida parent; and the progeny of the reciprocal cross has cytoplasm 

 only from the Lyon bean. Hence, the genetic difference which deter- 

 mines the visible difference between long and short pods is a difference of 

 the nuclei, not a difference of the cytoplasms. If we call this particular 

 nuclear difference of the gametes, E — e, the nuclear difference of the 

 zygotes (the Florida velvet bean and the Lyon bean plants) will be 

 £2 — <?2- {E2 = E + E.) Since we have no definite base of measurement, 

 it is useful in many cases to take the recessive as our base and to regard 

 e as zero. This is merely a convention. 



To sum up, the Florida velvet bean and the Lyon bean have one main 

 genetic difference affecting pod length. This genetic difference segre- 

 gates in typical Mendelian fashion. 



QUANTITATIVE INVESTIGATION 



Investigators of the inheritance of differences in size have found that 

 in many cases these differences are inherited as if several genetic dif- 

 ferences (factors) were concerned and dominance was lacking. For 

 instance, in East's masterly investigation of the inheritance of flower 

 size in crosses of two species of Nicotiana (6), the first-generation mean 

 flower length was near the geometrical mean of the parent flower lengths, 

 while the second-generation mean was only slightly greater. The fre- 

 quency array of the flower lengths of the second-generation plants 

 formed a continuous series between the two grandparental means, with 

 the mode below the center. If dominance had been present, the second- 

 generation mean would have been less than the first-generation mean 

 and the first-generation mean should have approached that of the long- 

 flowered parent (supposing all factors were positive). Emerson (9) 

 obtained similar results from a cross of short and long squashes {Ciicur- 

 hiia pepo). Groth (ii) in many crosses of tomato {Lycopersicon escu- 

 lentum) found the first-generation fruit length near the geometrical mean 

 of the parent lengths. However, the strict proof of this absence of dom- 

 inance demands, I think, the isolation of a family in which only one 

 such genetic difference is segregating. 



The hypothesis that size factors act as multipliers was, I believe, first 

 applied by East (6). Groth's results are readily explicable on this 

 hypothesis. A similar assumption has been made by Punnett and 

 Bailey (14). 



To sum up, previous work favors the hypothesis that some size factors 

 show no dominance and act as multipliers. 



