6. Sterility in tetraploid maize. An investigation of the possible causes of the variation in degree of sterility observed in different lines of tetraploid maize was made both from the cytological and genetical angles. In a study of microsporogenesis, both self-sterile and self-fertile lines showed a large number (8-10) of quadrivalents at diakinesis. This indicates that quadrivalent formation is not an important factor in causing sterility in tetraploid maize.

The chromosome number of the microspores varied from 14 to 24. Much of this variation was found to be due to the lagging of univalents and non-disjunction of chromosomes resulting in the formation of micronuclei, and to a lesser extent to the three to one separation of quadrivalents. From one to six chromosomes, usually in univalent groups, were seen to lag in sporocytes showing lagging. Gametes having 18 to 22 chromosomes are considered to be functional, since the chromosome numbers of the progeny of a tetraploid maize plant (4n = 40) has been shown to range from 37 to 42. The frequency of microspores having between 18 and 22 chromosomes agreed very well with the percentage of apparently good pollen in the fertile and sterile lines in which this was studied.

Four F1 populations resulting from crosses between lines with a high degree of pollen abortion (25%) and lines with a low degree of pollen abortion (10%), showed a low mean percentage of aborted pollen, suggesting a possible genic basis for this.

The coefficient of correlation between degree of pollen abortion and percentage of aborted ovules, when only fertile lines were considered, was found to be -0.651 0.025, indicating that factors causing pollen abortion are also operative in causing ovular abortion.

Evidence was obtained indicating that genetic factors for incompatibility were also involved in causing sterility in tetraploid maize. Some self-compatible lines were found to be cross-incompatible with other self-compatible lines when used as the pollen parent. This relationship was true even when the effect of different pollen was compared on two ears from the same plant, one ear being self-pollinated and the other cross-pollinated. In crosses between self-compatible and self-incompatible stocks a unimodal distribution was obtained for the F1 and a bimodal distribution for the F2 population, indicating the existence of at least one dominant or epistatic gene for self-compatibility. A study of reciprocal crosses between self-compatible and self-incompatible lines showed that self-incompatible lines were cross-compatible only when used as the pollen parent. No evidence of pollen tube competition was found in a compatible cross between a self-compatible and a self-incompatible line when mixed pollinations were made to determine this.

Some evidenee was obtained indicating that the chromosome number of the plant was not very important with respect to degree of fertility since a 38 chromosome plant was found to be 75% fertile (seed set) when self pollinated. This supports the conclusion that much of the sterility in tetraploid maize is due to genic rather than chromosome number difference.

If genes for self and cross-incompatibility are concerned in causing sterility in tetraploid maize, it is necessary to assume that these genes were present but inhibited in diploids, but became effective because of a genic unbalance resulting from chromosome doubling, i. e. upon doubling some genes increase in effectiveness and others remain static as far as their activity is concerned.

Harold E. Fischer