AN ATTEMPT TO IMPROVE THROUGH SELECTION THE STYLE LENGTH AND FERTILITY OF OENOTHERA BREVISTYLIS' BRADLEY MOORE DAVIS University of Michigan, Ann Arbor, Michigan Received January 2, 22 In an earlier paper (DAVIS ) there were published some observas on a selfed line of Oenothera brevistylis which had been carried for two generas through plants selected for greater varia in the direc of normal styles and stigmas. The line has now passed through five generas of selec, and not only has there been no improvement but the generas of later years have shown a retrogression so positive as to lend no encouragement to further effort. Apparently the varias in style length and fertility of Oenothera brevistylis are not due to heritable factors and consequently there can be no permanent success in selec for the improvement of these characters. The history of the first two generas has been published in detail (DAVIS, pp. -24) 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 at 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 produc, the ovaries ripening few or no seeds. There is, however, much varia 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 at 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 varia in that direc and suggest the possibility of selec for permanent improvement of this character. Such longerstyled flowers with more perfect stigmas have a distinctly greater degree Genetical studies on Oenothera XI. Contribu from the Botanical Laboratory, UNI- VERSITY OE MICHIGAN, No.. GENETICS 7: 0 N 22
STYLE LENGTH AND FERTILITY IN OE. BREVISTYLIS of fertility and it is possible by persistent selfing, in spite of many failures, to obtain from them a few seeds. The selec of an individual in each genera to carry on the selfed line was based on a test of performance of all of the plants in each culture. The test, introduced with the second genera, consisted in recording the length of style in the series of flowers opening along the main stem of each plant over a definite period of time. After the first genera the styles were listed as in three grades, long when the stigma reached above the lower tips of the anthers, medium when the stigma lay between the lower tips of the anthers and the top of the calyx tube, short when the stigma was at or below the top of the calyx tube. A full record of the performance of the plants in the first and second generas of the selfed line is given in tables and of the paper cited (DAVIS ) and will not be repeated. The best individual,.2d, IV-4, from the first genera of plants, showed a record between July and August, of long-styled and 0 short-styled flowers, a percentage of 00. In this first genera styles were recorded as long when they extended well above the top of the calyx tube and as short if the stigma was at or below the level of the calyx tube. There were selfed flowers of this best individual which set capsules with a total of 04 seeds. The poorest record of this genera was that of a plant which bore long-styled and 7 short-styled flowers over the same period of time and on which 2 selfings gave 7 capsules with a total of seeds. The second genera consisted of plants from the 04 seeds set by the selected individual,. 2d, IV-4, of the first genera, the performance of which is given in the paragraph above. There was a wide range of performance in this genera between June 0 and July 2, 7. The culture presented plants with high percentages of long-styled flowers as follows, 7., 7.0, 7.,.2,. and 4.7, but none of these was chosen as parent for the third genera because they gave either no seed when selfed or the yield was small (DAVIS, table ). The plant selected to carry forward the selfed line, 7.2, V- ripened 4 seeds from capsules after selfing 24 flowers, and produced over the period of observa long-styled, medium-styled and short-styled flowers, a record of. percent of long-styled flowers. The poorest plant of the second genera bore 0 long-styled, 2 medium-styled and short-styled flowers. The record of the culture as a whole was much below that of the first genera, as indicated in table 7, although the two generas cannot be compared with strict accuracy because of
2 BRADLEY MOORE DAVIS the different system of recording style length introduced with the second genera. The parent of the third genera was 7.2, V- described in the preceding paragraph. Its seeds, being soaked in water 24 hours, were forced to complete germina in a Petri dish following alternate exhaust and pressure up to 7 pounds applied 4 times in the course of 24 hours. As shown in table the 4 seeds from capsules gave seedlings, a germina percentage of.4, and 0 brevistylis plants reached maturity. The 0 plants of the third genera (table 2) made a record somewhat worse than that of the second genera and was far below that of the first. Its best performance, by plant.2, -, from July 7 to July 2,, was 2 long-styled, medium-styled and 4 short-styled flowers, a record of. percent long-styled flowers. There were selfed 2 flowers which set only capsules with a total of seeds, a very low degree of fertility. The second-best plant in the culture produced long-styled, medium-styled and short-styled flowers over the same period of time, a percentage of. long-styled flowers. As shown in table 2 the remaining plants of the culture had percentages of long-styled flowers ranging from. to 2., all very bad records. CULTURE TABLE PARENT PLANT PERCENTAGE OF brcvistylis SEEDS 'OWN I SEEDLINGS GERMINATION PLANTS REACHING MATURITY camules The best plant of the third genera,.2, -, was selected as parent of the fourth. Its seeds from capsules after selfing 2 flowers were forced to complete germina in Petri dishes after being soaked in water and following alternate exhaust and pressure up to 4 pounds. Complete germina gave 4 seedlings, of which 2 died early, so that only 2 plants reached maturity; the germina was 2.7 percent (table ). The fourth genera, therefore consisted of 2 plants the performance of which between July 2 and August 2, is given in table 4. The best of these plants,.2-, with 7 long-styled, medium-styled and short-styled flowers made a record of only.4 percent long-styled flowers. From 2 selfed flowers 7 capsules were set which ripened seeds. The other plant presented a record of. percent long-styled flowers. GENEITCS 7: N 22
STYLE LENGTH AND FERTILITY IN OE. BREVZSTYLZS The seeds from the best plant of the fourth genera,.2-, after 24 hours in water, were subjected to alternate exhaust and pressure up to 0 pounds 0 times in 24 hours. Complete germina yielded seedlings, a percentage of.4 (table ) ; plants reached maturity. TABLE 2 Performance in the third genera of brevistylis in a selfed line. Observas from July 7 to 2,. PLANTS FLOWERS OB- SERVED LONG- MEDIUM- SHORT- E R C E NT A C OF LONG- FLOWERS YIELD OF SEED PBOY SELFED FLOWERS.2, -.2,II-.2, -.2, -4.2,II-4.2, -2.2,II-.2,II-.2, -.2,II-2 4 2 0 2 24 0 2 0 7 0 0 4 7.. 2. 2.0 2... 2... 2 flowers selfed gave capsules with seeds TABLE CULTURE.2 I.2,I- PARENT PLANT brewistylis I SEEDS 'OWN capsules SEEDLINGS 4 PERCENTAGE OF PLANTS REACHING GERMINATION MATURITY 2.7 2 TABLE 4 Performance in the fourth genera of bre&tylis in a selfed line. Observas from July 2 to August 2,. PLANTS FLOWERS OB- SERVED LONG- MEDIUM- SHORT- 'ERCENTAGI OF LONG FLOWERS YIELD OF SEED PROM SELPED FLOWERS.2-.2-2 7.4. 2 flowers selfed gave 7 capsules with seeds
4 BRADLEY MOORE DAVIS The performance of the plants in the fifth genera (table ), taken between July and August, 2, was somewhat better than that of the fourth, but not so good as that of the third, and it was far below the records of the first and second generas. Its best plant, with longstyled, medium-styled and 0 short-styled flowers, made a record of 7. percent long-styled flowers as contrasted with records of 7. percent and. percent for the best plants of the second and third generas, respectively. The other five plants of the culture had records for longstyled flowers ranging from 22.2 to.7 percent (table ), which were very poor performances. TABLE CULTURE 2.2.2- (7 cadsules) SEEDLINGS PERCENTAGE OF PLANTS REACHING GERUfNATION MATURITY.4 TABLE Performance in the jifth genera of brevistylis in a selfed line. Observas from July to August, 2. PLANTS 2.2-2.2-2.2-2 2.2-2.2-2.24 FLOWERS OBSERVED - 7 LONG- PERCENTAGE OF MEDIUM- SHORT- LONG- FLOWERS 2 0 2 7. 22.2.2 7..4.7 Thus instead of improvement in the length of style as the result of selec for these five generas there took place a marked retrogression. The records of the third, fourth and fifth generas were far below those of the first and second. This retrogressive movement is clearly expressed in table 7 which gives the distribu of the plants for all generas grouped in percentages of long-styled flowers. There is no sugges in this table of probable or permanent advance in style length. As has been stated, there is correlated with the poor development of the style and stigma of Oenothera brevistylis a very low degree of seed produc. On this subject it is difficult to obtain much data because so few capsules set any seed that a very large number of pollinas must he GENETICS 7: N 22
STYLE LENGTH AND FERTILITY IN OE. BREVISTYLIS made to obtain harvests worth considera. In choosing parent plants for the five generas of this selfed line it was of course necessary to pick individuals that produced enough seed to ensure a progeny and in some cases plants with better records of performance in style length could not be used because they ripened either no seed at all or so few seeds that there was no prospect of a satisfactory succeeding genera. This was particularly true of the large second genera (DAVIS, table ) in which for this reason plants were discarded although their percentage records of long-styled flowers were somewhat higher than that of the plant selected to carry forward the line. TABLE 7 Distribu of the plants of bremstylis through generas, grouped as to the percentages of long-styled flowers. PERCENTAGES IN CLASSES 0-0 - 2-0 -40 4-0 -0-70 7-0 -0-00 First......... Second... 7 2 4...... GENERATIONS Third Fourth 2... 2... Fifth Table gives the comparative fertility of the plants selected as parents of the five generas of this selfed line of brevistylis. It presents a record consistent with that of table 7, showing that fertility fell off in the later generas as did the percentages of long-styled flowers. Thus the average number of seeds per capsule in those set by the five parent plants were.,.o, 4.,.O and 2.. The percentages of selfed capsules that set seed likewise declined with the later generas, e.g.,.7, 7., 4., 2. and 2.0. The percentages of germina fluctuated without apparent significance. These conclusions, because of the very low fertility, are necessarily based on a small amount of data, but it is a matter of interest that they show a close correla between higher degrees of fertility and higher percentages of long-styled flowers.
BRADLEY MOORE DAVIS TABLE Comparative fertility of the plants selected as parents of thefive generas of this selfed line of brewistylis. SELECTED PLANT FWWEBS SELFED CAPSULES THAT SET SEED I TOTAL PERCENTAGE AVERAGE XUMBEROF OFSELFED NVkCBEROF CAPSULES TEAT SEEDS PER SEEDS SET SEW CAPSULE PEPCENTAGE OF GERMWA- TION Parent plant of thc selfed line.d, - From first genera.2dj IV4 ~- 04.7. 7. 4. 4. From second genera 7.2, V- 24 4 4. 4..4 From third genera.2,i- From fourth genera.2-2 2 7 2 2. 2. 2.7.4 Thus after a five-years study of a selfed line of brevistylis I am brought to the conclusion that the factors which are responsible for the reduced style length and stigma structure in this plant, and with the correlated very low fertility, are fixed, in the sense that they express themselves through fluctuas within distinctly limited ranges. The lower ranges are close to the usual condis in the plant, which are those of complete or almost complete sterility. The higher ranges may be found by close observas on performance and tests for fertility. They express a small but definite advance in style length and stigma structure and in somewhat increased fertility, but there is no evidence that such improvement will be inherited. Selecs for increased style length and increased fertility failed to hold the gains and the results were so discouraging to further attempts to improve style length and fertility that this selfed line wjll not be continued. REFERENCE CITED DAVIS, B. M., The segrega of Oenothera brevistylis from crosses uith Oenothera Lamarckiana. Genetics : 0-. GENETICS 7: N 22