A. B. Stout 75 



A glance over this table shows that a rather large proportion of 

 plants are self- fertile. Especially is this true of series {E3 x A)-Ji.-^-, 

 which in this respect is perhaps the most highly self-fertile of the various 

 series thus far grown. 



The more complete summary of the record for the family {E3 x A) 

 is presented in Table VII. As there indicated, the family is not a large 

 one. From the cross between the two self-sterile plants A'-> and A, 

 21 plants were grown in the F^, of which four were self-fertile with fer- 

 tilities of 2, 4, 13, and 48. Progeny were grown only fi'om one plant 

 having the highest fertility. Of the 18 grown, 10 were self-fertile, 

 and the fertilities determined for 9 of these ranged from 3 to 50°/„, 

 with an average at 26. Selection for parents for the next generation was 

 confined to the two plants most highly self-fertile.' 'Thus the immediate 

 parents of the two series were 43 and 50 °l ^ self-fertile respectively, and 

 the common Pj parent was 48 °/^ self-fertile. The selection has here 

 been continually of plants with highest fertility. The regression to 

 a condition of complete self-sterility and to feeble self-sterility is most 

 noticeable. The number of plants is perhaps not sufficient to determine 

 the mathematical expression for such regression with accuracy, but the 

 records indicate an irregular and incomplete inheritance of self-com- 

 patibility. 



However, the proportion of self-fertile plants is higher in the series 

 grown from self- fertile parents than that of the self- fertile plants 

 appearing sporadically among the progeny of the original self-sterile 

 parents. The average fertility for each of the two series from two 

 generations of self-fertile ancestry is lower than that of the preceding 

 generation {Eo x A)-^-. (A series was grown in each of two different 

 years from seed of the plant {E3 x A) tw. 4-', data for both are here 

 compiled as for a single series.) The range of self-fertility has, how- 

 ever, been extended in the case of one plant to 60°/„. This plant, 

 however, is of the series which, as a whole, is of lowest average fertility. 

 The summary for the progeny of the plant {E3 x A)-4--, as a whole, 

 shows that 65 "/, of all the f)lants were self-fertile in some degree, with 

 a distribution on the basis of self- fertility that is decidedly irregular 

 and skew, and with an average fertility of 0'197. It has already been 

 suggested by the writer that there is evidence that complete self- 

 sterility may involve different intensities of incompatibilities. If there 

 were some means of determining comparative values for these, the 

 distribution for a family and for the different series might be found 

 to be more in agreement with a normally fluctuating variability. As 



