66 Cooperative Investigations on Plauts 



thc Hampden parcnt value from which it rose in 1900. While in Table 11. the 

 variability of the oH'spring was in eaeh case greater than that of the parents, — 

 because \ve were coniparing a capsule variability in the former, with a mean plant 

 variability in the latter, — in Table III. the parent capsules are markedly more 

 variable than the offspring capsules. This is precisely what \ve should expect, 

 if the individuality on which heredity dcpends lies in the plant as a whole and 

 not in the special capsule. We shall investigate this at greater length below. 



(4) Relative Fertility of Different Capsules. Direct experiments were not 

 raade on relative fertility, but one or two points may be noted. In exaniining the 

 Crockhain capsules, we were again* Struck by the steril ity of the low baiided, and 

 to a lesser extent of the high banded poppies. W. R. Macdonell for the second 

 Enfield crop selected 57 capsules aud sowed from the seed of these capsules 

 57 rows. Of these rows 22 were from capsules with 9 to 12 bands, 28 from 

 capsules of 13 to 15 bauds and 7 from capsules of 16 to 20 bands. 10 rows of the 

 first group, none of the second and 3 of the third were unproductive, or about 

 45 per cent. of the low banded aud 43 per cent. of the higli bamled capsules 

 failed, while none of the capsules near the modal value (about 14) were un- 

 productive. 



In the Oxford Band Series (ß), W. F. R. Weldon sowed 110 pots from 

 capsules 7 to 10, 150 pots from capsules 11 to 15, and 90 pots from 16 to 18, 

 each pot being sown with a few seeds. 43 per cent. of the pots in the first group 

 failed to show germinations, and 21 per cent. in the third group, while 8 per cent. 

 only failed in the group from 11 to 15, the modal value of the original Hampden 

 crop being 12"75 bands per capsule. The poppies were -sown on .March 25 — 26, 

 and the productive pots counted from April 12 to 20. Further, on April the 21st, 

 the first group from capsules of 7 to 10 had about 5 seodlings per pot, the second 

 group from capsules of 11 to 15 about 22 seudlings per pot, and the third group 

 from capsules of 16 to 18 had 10 seedlings per pot. 



Now of course this seedling result admits of more than one Interpretation, for 

 it may bc said that there was more seed of thc modal capsules aud that more of it 

 may have been sown. While it required comparatively few modal Hampden 

 capsules to provide a good amount of seed, almost every available low-banded 

 and many high-banded had to be dealt with to collect enough seed at all from 

 these classes of capsules, thus the seed was certainly not iu proportion to the 

 frequency of the capsules in the original crop, nor was it sown in that proportion. 

 Hence Weldon's percentage of productive pots and proportion of seedlings seem 

 on the whole to confirm Macdoneil's results, which started not with cqual quantities 

 of seed, but with the seed of single capsule.s. 



From the Seed (a) it is difficult to draw any conclusions, for we do not know 

 what proportion of the seed came from modal capsules, and it is therefore not 

 possible to test the general principles (i) that the modal capsule lias most seed 



* Grammar of Science, p. 444. 



