(Department of Botany)


Relation of crossing over to mutation of Ab.


Previous studies indicate that Ab and its mutant Ad derivatives are non-linear in action. The former produces purple-plant and aleurone color; the latter are associated with pale phenotypes. Ab and Ad types produce a brown pericarp which is dominant to the red effect of A. Moreover studies involving different doses and combinations of the Ad alleles are difficult to interpret on the basis of a simple relation between gene and agent. It is possible to explain the action of these alleles if it is assumed that they are compound. in the sense that each is composed of two or more physically distinct determinants. To test this possibility in the case of Ab experiments were conducted to determine whether there is a relation between crossing over and the mutation of this allele.


The genes lg2 and et, which lie, respectively, 33 units to the left and 13 units to the right of A in the long arm of the third chromosome, were used as markers. The recombination value for the lg-et interval is 0.42. In view of this high frequency Ab a plants were employed rather than the homozygotes since this permits the use of a as a third marker gene. Since mutants of the Ad type have never been obtained from a it is certain, after testing to establish their nature, that the mutants obtained in these experiments originate from Ab.


Individuals having the constitution Ab/lg a et were grown at Princeton in 1947 and crossed with lg a et pollen. To avoid contamination the plants were started and flowered earlier than any others in the field; all pollinations were made by hand. Among the resulting progenies which contained 27,936 purple seeds (Ab contributed by the egg) seven pale seeds were obtained which produced pale plants ana whose progeny tests showed that they carried a mutant allele. Progeny tests grown at the California Institute of Technology in the summer of 1948 established that these plants carried a mutant gene of the Ad type.


Tests of the mutant individuals showed that six of the seven Ad-bearing strands contributed to the eggs by the heterozygous parents represented recombinations for a and et (Table 1). The expected frequency for recombination in this region is 0.128, taken from table 2, which summarizes the data on crossing over in the lg-a and a-et regions; the data are from ears in those families which produced the mutant individuals. Although the numbers are small, the high frequency (0.858) for recombination in the a-et region among the mutant Ad strands suggests a relation between crossing over and mutation of Ab.


Table 1. Cross: Ab/lg a et x lg a et



Constitution of the Ad-bearing strand



Crossover types

Mutant plant

Non-crossover types

Region I

Region II

Regions I and II




Lg Ad et
























lg Ad Et







lg Ad et



Table 2. Summary of the data on crossing over in the lg-a (Region I)
and a-et (Region II) segments from progenies of backcrosses
of the type: Ab/lg a et x lg a et





Region I

Region II

Regions I and II
























This interpretation is supported by evidence from experiments in 1948 conducted on a large scale. As before, marked heterozygotes carrying Ab were crossed with lg a et pollen which in addition carried Dt. The precautions against possible contamination were the same as in the previous experiment. While progeny tests of suspected mutants from these crosses are not yet available those seeds, which were pale and dotted, were selected and classified for normal versus etched (et) phenotype. Since dots have never been observed on endosperms carrying Ab, a and Dt, the presence of dots on the selected seeds many be taken as evidence of their mutant character. Of the total of 34 pale, dotted seeds obtained, 30 were crossover types for the a-et interval; as in the previous experiment there was a preponderance of the crossover class. The 34 pale seeds occurred individually on ears. From classification of the remnant seeds on these ears a frequency of 0.1399 O.0045 was obtained for recombination between a and et; this is to be compared with 0.882, the value for recombination in the same region among the 34 mutants.


In the cases of five of the total of 41 mutants from both experiments the Ad-bearing strands delivered to the eggs were non-crossover types for the a-et region. It is possible that these are due to mutations of Ab, which are not associated with crossing over at that locus; if such mutations do occur they would fall predominantly in the noncrossover class. However, in view of the relatively great length of the a-et segment (13.2 units as an average of the two experiments) it is a more plausible explanation that the apparent noncrossovers are cases of double exchange, one occurring at the locus of Ab giving rise to the mutation, the other occurring somewhere between this locus and that of et, thus reconstituting the parental combination.


The evidence suggests that Ab is composed of at least two components, separable by crossing over. If these are designated alpha and beta, the latter being more distal, the Ad mutants may be described as having alpha and lacking beta; this follows since the strands carrying Ad were predominantly of the nonparental class for et, the more distal marker.


On this basis it is possible to account for the non-linear action of Ab in terms of its compound nature. For example, the dominant brown pericarp effect of Ab may be argued to reside in the alpha component (Ad) since the Ad mutants also show this effect. It is possible that the Ad mutants, which also are antimorphic in their action, are in turn compound. This possibility, along with others which logically suggest themselves, is being investigated. It is an inviting prospect that the complex action of a number of the A alleles may be resolved on the basis of their compound nature rather than in terms of genic agents having relatively complex interactions.


John R. Laughnan