334 Belling. 



Nothing can be learned from LV-206 and LV-475, except that 

 they did not contradict their full pods by producing a plant with semi- 

 sterile pollen. The other parent plants have a sufficient number of 

 progeny. 



Thus fourteen plants with fertile ovules (and doubtless fertile 

 pollen, though the pollen was not examined) gave 2tj4 progeny, all with 

 fertile pollen. Seven plants with semi-sterile ovules gave a progeny of 106 

 plants with sound pollen, and 102 i)lants with half aborted pollen. This 

 ratio is undoubtedly 1:1. (The flowers of five seuii-sterile third- 

 generation plants, LV-279-34, 279-36, 569-3, .569-7, and 569-27, showed 

 however a large excess of empty gi-ains.) 



To sum up: — In the third generation, fertile second- 

 generation parents give progeny all with sound pollen-grains; 

 and semi-sterile parents give a progeny half of whose mem- 

 bers have sound pollen-grains, and half have semi-aborted 

 pollen-grains. 



Ovules of the Third Generation. 



Velvet by Lyon. — Notliiug can be said as to the iiilieritance of 

 semi-sterility in this generation, since the Fa crops were mostly small 

 on account of the crowding with sorghum in the elimination test. Three 

 families, from fertue second-generation plants, were gi'own on poles in 

 1912, and all tlieir members had fertile ovules. 



Lyon by Velvet. — Six fresh pods were examined in 1913 from 

 all available plants of the 21 families, between September 18 and 

 Octobei' 11. The families of LV-436 and LV-206 consisted of black 

 plants; and these, together with the many black plants of LV-486, and 

 the few black plants of the othei' families that bore pods, are kept 

 separate, because of their often imperfect pollination. 



Progenies of fertile second -generation plants. — The following 

 table gives the numbers of undeveloped ovules for the progenies of the 

 fertile parents. Only the distal halves of the pods are considered. 

 Two progenies, those of LV-92 and LV-468, have the highest percen- 

 tages of aborted ovules. Among the parent plants also, LV-92 and 

 LV-468 had the highest percentages of aliorted ovules (except the 

 black plant LV-486). This points to tlie inheritance of some zygotic 

 factor in LV-92 and LV-468 adverse to the nutrition of the ovaries. 

 The segiegated plants with white pubescence of the progeny of LV-486 

 show but little abortion. 



