4o6 



Journal of Agricultural Research 



Vol. V. No. lo 



grown were selected with the aim of obtaining useful agricultural plants. 

 A fairly complete third generation was raised , but the fourth generation 

 was the result of selection and was the opposite of a random sample. 



QUALITATIVE INVESTIGATION 



The Florida velvet bean (Stizolobium deeringianum) was crossed both 

 ways with the Philippine Lyon bean (S. niveum). A pertinent descrip- 

 tion of these plants has been given in my account of the inheritance of 

 semisteriHty (4). The Florida velvet bean has a short pod (PI. XL, fig. 

 5), while the pod of the Lyon bean (Pi. XL, fig. C) is about half as long 

 again and is broader. The pods of the first-generation hybrid plants 

 were as long as, or slightly longer than, those of the Lyon. The progeny 

 of the hybrids in the second generation could be divided by inspection 

 into short-podded plants and long-podded plants. The short pods could 

 be identified, even when young, by their greater proportional width. 

 Although both short pods and long pods varied greatly in size on different 

 second-generation plants, yet no case was met with where the classifi- 

 cation could not be carried out when all the pods on a plant were taken 

 into account. Plate XL, figures A and D, shows typical pods of second- 

 generation plants with pods shorter than the Florida velvet bean and 

 longer than the Lyon bean pods. The difference between short and 

 long pods was sharply marked in all the segregating third-generation 

 families. 



Tables I, II, and III give the results of field inspection, checked by 

 examination of the pods after harvesting. 



Table I. — Length of pods in first-generation bean crosses 



<" The pollen parent is given last throughout this article. 

 Table II. — Length of pods in second-generation bean crosses 



