1905.] On the Inheritance of Heterostylism in Primula. 585 
when fertilised with pollen of long- and equal-styled plants, No. 6 gave 
14 short, 5 long. 
The evidence so far is, therefore, that the pollen of No. 6 gave a mixture of 
longs and shorts, and consequently was carrying both characters, while all the 
plants raised from it as female were shorts. The numbers alone are of course 
too few to justify any conclusion, had it not been that a closely similar result 
appeared in the next generation. 
By self-fertilisation No. 6 gave a short-styled plant, No. 37. This, on 
self-fertilisation, gave 22 shorts and no longs. Fertilised by pollen of long- 
styled plants, it gave 14 short, 24 long. But when used as a male parent, its 
pollen applied to long- and equal-styled plants gave 148 shorts and only 
4 longs, of which one was recorded as “ doubtful.’”* 
Taking their offspring together, Nos. 6 and 37, when fertilised by long- 
and equal-styled, gave 29 long, 28 short; while when the same two plants 
were used as males, the total offspring were 184 shorts and 4 (? 3) longs. We 
have, therefore, the remarkable phenomenon of plants which, judged by the 
female gametes, were ordinary heterozygotes, while their male gametes were 
almost exclusively bearing the dominant character. Pending further investi- 
gation, we can offer no further comment on this singular case. It will be 
noted that, since the mixture was given by the female side, no hypothesis of 
parthenogenesis will meet the case. 
Results of Double Pollination. 
In addition to the experiments described above, an attempt was made to 
investigate another possibility respecting the consequences of legitimate and 
illegitimate unions. Darwin, and after him many others, proved that in 
Primula more seeds are produced when plants with styles of dissimilar types 
are united (legitimately) than when similars are united (illegitimately). 
Nevertheless, illegitimate unions are not necessarily sterile, but, especially in 
the case of P. sinensis, may produce a good deal of seed. 
For examples we may refer to the average numbers given by Darwin.t 
Taking the average for legitimate unions at 100, the 13 illegitimate unions in 
the genus Primula give an average of 53 seeds per capsule, and we have found 
a similar proportion maintained with some constancy in our own fertilisations. 
Some egg-cells (about half) are therefore fertilised by illegitimate pollen, 
* The nature of the doubt is not recorded. Until the results were added and classified 
no special interest had been attached to this family. Each plant as it began to flower 
was recorded and thrown away to make room. Probably this individual was recorded 
before the flower completely matured. 
+ Loc. cit., p. 246. 
