An Attempt to Improve through Selection the Style Length and Fertility of OENOTHERA BREVISTYLIS.

In an earlier paper (DAVIS 1918) there were published some observations on a selfed line of Oenothera brevistylis which had been carried for two generations through plants selected for greater variation in the direction of normal styles and stigmas. The line has now passed through five generations of selection, and not only has there been no improvement but the generations of later years have shown a retrogression so positive as to lend no encouragement to further effort. Apparently the variations in style length and fertility of Oenothera brevistylis are not due to heritable factors and consequently there can be no permanent success in selection for the improvement of these characters. The history of the first two generations has been published in detail (DAVIS 1918, pp. 520-524) and requires but brief review. Oenothera brevistylis differs from Oe. Lamarckiana chiefly in its short style and reduced stigma which generally lies near the top of the calyx tube instead of a t a point above the tips of the anthers. With this poor development of the style and stigma is correlated a very low degree of seed production, the ovaries ripening few or no seeds. There is, however, much variation among different plants of brevistylis and between flowers on the same plant with respect to the length of style, and occasionally the stigma will be present a t the level of the lower tips of the anthers or somewhat above them. In these flowers the stigma lobes are generally found to be more normal in size and appearance. Although such styles are far from the full length which in Lamarckiana is above the anthers, they nevertheless do present a marked variation in that direction and suggest the possibility of selection for permanent improvement of this character. Such longerstyled flowers with more perfect stigmas have a distinctly greater degree