266 



SCIENCE 



[N. S. Vol. XXIX. No. 737 



lamarchiana and being indistinguishable from 

 it in manner of branching and large number 

 of flowers produced. A fourth member of this 

 group bloomed late in the season and did not 

 attain its full stature. The buds of these 

 peculiar hybrids might have been classed as 

 stout lamarchiana or slender gigas. The 

 sepals remained conspicuously yellow, attain- 

 ing near the end of the season a faint tint of 

 red, like extracted latas of this cross. The 

 foliage resembled gigas. 



The second half of this group consisted of 

 four plants somewhat resembling the first in 

 size and branching habits but, in three of the 

 four individuals, distinguished from it by the 

 deep red color of the sepals — strikingly re- 

 sembling rubrinervis in this regard. The 

 fourth plant had less deeply colored sepals and 

 possibly should be placed in a subdivision of 

 its own. 



The appearance among the offspring of 0. 

 lata X 0. gigas of plants with pronounced 

 lamarchiana characters is puzzling. The 

 combination of these with equally pronounced 

 gigas characters, further associated with the 

 intermediate number of chromosomes, pre- 

 cludes the possibility of the pistil parent hav- 

 ing been accidentally fertilized by lamarchiana 

 pollen. MacDougal" succeeded in fertilizing 

 lata with its own pollen for the first time 

 on record, and obtained seed which " gave rise 

 to a progeny which showed only the conn 

 stitutents usually found in a progeny of this 

 plant when fertilized by lamarchiana." 



Study of the pollen of these hybrids is as 

 yet in the initial stages, but some interesting 

 observations have already been made. 



Pollen grains of 0. lamarchiana, 0. lata, 

 0. nanella, 0. rubrinervis and 0. cruciata are, 

 so far as I have observed from the cultures 

 growing at the station this summer, character- 

 istically 3-lobed as figured by Gray' for 

 CEnothera. Limiting observations to the first 



° " Mutations, Variations and Relationship of 

 the CEnothera," by D. T. MacDougal, A. M. Vail 

 and G. H. Shull. The Carnegie Institution of 

 Washington, Papers of Station for Experimental 

 Evolution No. 9. 



° Gray's " Lessons and Manual of Botany," re- 

 vised edition, 103, fig. 316. 



two forms mentioned, ordinarily about 1 in 

 1,000 grains has been found to have 4 or 

 more lobes, although as high as 15 per cent, 

 has been observed in normal, typical indi- 

 viduals. Seven representatives of pure gigas 

 were examined and the 3-lobed grain was 

 found to occur as rarely in this species as 

 the 4 and 4+ grains in lamarchiana. Four- 

 lobed grains prevailed for most individuals of 

 gigas, although these were commonly found 

 mixed with 5-, 6-, 7- and even 8-lobed grains, 

 usually decreasing in frequency with the in- 

 crease in the number of lobes. A further 

 interesting point brought out was the fact 

 that certain individuals showed these extra 

 lobes in all stages. 



The study of the pollen of the hybrid 

 progeny of 0. lata^C^O. gigas was therefore 

 of absorbing interest. 



Class I. (two extracted latas) showed 3- 

 lobed pollen with an exceptional extra-lobed 

 grain, as is characteristic for a lata mutant. 



Class H., composed of two adults with 

 gigas number of chromosomes gave each 4 

 and 4 -\- grains with an exceptional 3-lobed in- 

 dividual as is characteristic for pure gigas. 



Class HI. (hybrid number of chromosomes), 

 pollen of seven individuals studied each gave 

 a mixture of 3- and extra-lobed grains. The 

 proportions have not yet been ascertained, 

 but it is clear that the former are usually 

 considerably in excess of the latter. 



The sterility of lata is due not only to the 

 scarcity of pollen produced, but to the large 

 percentage of bad grains. The amount yielded 

 by any form was found to vary with the indi- 

 vidual, with the fiower of the individual, and 

 with the anther of the flower. Even the most 

 fertile forms produce a surprisingly large per- 

 centage of bad pollen, and some of the above 

 hybrids, notably certain members of Class 

 m., subdivision: 3, have been found to have 

 as high as 90 per cent., although pollen was 

 produced in moderate abundance. Inasmuch 

 as lata yields but very little pollen, mostly bad, 

 and gigas, while producing a considerably 

 larger quantity, has in most instances a low 

 percentage of good grains, it is not at all sur- 

 prising that it was utterly impossible to arti- 

 ficially self-pollinate the majority of these 



