Dec. 6. 191 5 Inheritance of Length of Pod in Certain Crosses 



407 



The most probable single ratios have been calculated on the hypothesis 

 that there are three chances for the long pod to one chance for the short 

 f>od. However, by the theory of probability, a deviation from the 

 whole numbers nearest to these calculated ratios is far more likely to 

 occur than not. The most probable deviation has been calculated by 

 the conventional formula,^ and is given in the last column of Table II. 

 Since the actual are not greater than the calculated deviations, it is 

 probable that there is no interference with the random segregation of 

 the long and the short pod, with three chances for the long to one chance 

 for the short pod. 



The third-generation families of the Florida velvet bean X Lyon bean 

 were grown in an elimination field among crowding sorghum, where there 

 was some selective elimination of short-podded plants (3). Hence the 

 ratios are useless here. Two long-podded parents, however, of those 

 whose families were grown on poles gave a total of 49 long- podded to 13 

 short-podded (Calculated, 46.5 ± 2.3 : 15. 5^2. 3). In the third genera- 

 tion of the Lyon bean X Florida velvet bean, 17 families of more than 8 

 members each from long-podded parents were grown on poles. The 

 totals of the 11 segregating families among these amounted to 231 long- 

 podded and 76 short-podded plants, the calculated nearest whole num- 

 bers being 230 and yy. The long-podded homozygotes could not be 

 distinguished by inspection from the heterozygotes. These results are 

 given in Table III. The abbreviations used in this and the subsequent 

 tables in this paper are "V" for Florida velvet bean and "L" for the 

 Lyon bean. 



Table III. — Length of pods in third-generation bean crosses from long-podded parents 



Calculated ratio. 



LV-92.. 

 LV-548. 

 LV-S69. 

 LV-558. 

 Lv-27. . 

 I.V-311. 



LV-80.. 

 LV-II3. 



LV-279. 

 I.V-486. 

 LV-91 . . 

 LV-II4. 

 LV-310. 

 LV-468. 

 LV-527. 

 LV-461 . 

 LV-392. 



Total. 



Long. 

 23 

 30 

 38 

 20 

 28 

 9 



231 



Short. 

 O 



27-75 



21 



22.5 



28.5 



18.75 



12.75 



25-5 



18.75 



17-25 



27 



9-25 



7 



7-5 



9-5 



6.25 



4.25 



8-5 

 6.25 



5-75 



9 



3-5 



76 



230. 25 



76-75 



-2-75 

 + 1.0 



+ 1-5 



+2.5 



+2.25 



+0. 25 



+0.5 



-3-75 



-2. 25 



-hi.o 



+O.S 



+0.75 



:l-5 

 :i. 6 



±1-5 

 ±1. 2 

 ±1.7 

 ±1-5 

 rfci.4 

 ±1.8 

 ±1. I 



±5-1 



1 1 have used the ordinary formula for probable deviation, which, however, does not seem to be appropri- 

 ate (except with large numbers) to any but a i to i segregation. East and Hayes's practical test of this 

 formula with large numbers (7) shows that it will in that case fit a 3 to i segregation with sufficient accuracy. 

 Hence, the calculated probable deviations in Table III, where the numbers are small, are not reUable. 



