1915] Goodspeed: PartJienogenesis, Parthenocarpy, Ph&nospermy 265 



found that seed of this same size and appearance, produced occasion- 

 ally in our cultures of other iV. Tabacum varieties, will give up to or 

 over fifty per cent germination. The small seed, normal in appearance, 

 in almost all cases was found to consist of entirely emptj^ seed-coats. 

 In three cases, however, considerable amounts of endosperm could 

 be distinguished under magnification but, though the seeds were 

 carefully dissected, no trace of embrvos could be found. 



I know of no special term to describe the production of seed 

 which appears to be normal but consists of empty seed-coats only — 

 i.e., the type of seed usually referred to when the word "abortive" 

 is used. At Professor Setchell's suggestion the term phenospermy 

 is proposed to cover this production of "abortive seed." The seed 

 produced by Wellington and Hartley as a result of stimulation fol- 

 lowing castration may thus be spoken of as plienospermic. 



The total number of seeds, all of which were of size similar to 

 the self-fertilized seed of "Nic. tahaccum Cuba" and a portion of 

 which, after bleaching with Eau de Javelle, were seen to contain 

 endosperm and embryos normally formed, was approximately 50. 

 Unfortunately, a rather large percentage of this number was des- 

 troyed in the bleaching test. As the seed was shaken from the seed 

 packets into the watch glasses (cf. p. 256) a large part of the heavier 

 normal seed fell out in each case. Only 18 seeds, among those re- 

 maining in the seed packets which were determined by the bleach- 

 ing test to contain normal seed, were available for germination. 

 After three weeks eight of them had germinated and six seedlings 

 are normally developing. Over twenty germination tests of seed from 

 packets which gave no indication of normal seed with the bleaching 

 test have not shown any signs of germination. The self -fertilized seed 

 of ''Nic. tahaccum Cuba" gave germination tests averaging 83 per 

 cent after three weeks, and practically all the seeds that did not 

 germinate were phenospermic (cf. Thomas, 1909, p. 4). 



It is worthy of note that, of the castrations and mutilations which 

 produced normal or phenospermic seed, approximately 30 per cent 

 occurred on the terminal inflorescence; 50 per cent on the terminal 

 inflorescence and four or five bald suckers normal to this variety, 

 and 20 per cent only on the lower leafy laterals. Castrations and 

 mutilations made within two weel« after the opening of the first 

 flower gave normal or phenospermic seed in 65 instances, as com- 

 pared with 56 instances for operations performed more than two weeks 

 after the opening of the first flower. There was also no significant 



