--Virginia Walbot

A cluster of purple kernels on an ear containing a mutable gene of the anthocyanin pathway is likely to represent a single reversion event that occurred during ear ontogeny. The larger the sector, the earlier the event. If the frequency of excision is constant during development then a simple "powers of two" model would predict the number of clusters of each size, i.e. twice as many single kernel events as two kernel events, twice as many two kernel events as four kernel events, etc. Yet it is obvious that there is an excess of single kernel events. Does this mean that reversion frequency is much higher late in development?

Single kernel events are a complex class: an unknown number of cell divisions intervene between separation of lineages for individual kernels. Of events that occur just prior to meiosis, half will be lost during meiosis as a result of cell death. Among events that occur somewhat earlier, half will be lost as a result of lower floret abortion. Thus a single kernel may be the only visible event in a small sector; when a reversion event occurred early enough to affect 8 or more kernels, then it is highly likely that 2 purple kernels will be found and the sector will be correctly scored as occurring long before megaspore formation.

To determine the bias for single kernel events, the distribution of single kernels vs. clusters was scored for a1-m4 Ac crossed by a1 tester in two populations of 33 and 16 ears, with an estimated 12,250 kernels scored. This test ignores the possibility of coincidental reversion of the two alleles contributed by the female parent and of contiguous lineages each undergoing a reversion event, because overall recovery of revertant kernels was low (191/12,250).
 

Sector Size	Number of Events
One kernel	156
Two kernels	19
3-4 kernels	9
5-8 kernels	3
9-16 kernels	1
17-32 kernels	3 (17, 19, 29 kernels)

There is clearly a bias for single purple kernels. With the larger sectors there is a trend that fits a "powers of two" notion of ear development: the events with two purple kernels are about twice as frequent as the 3-4 kernel class, etc. There is, however, considerable uncertainty in classifying the smaller sector sizes. Does a single kernel really represent a two or even four kernel event masked by segregation at meiosis? Do 3 kernels represent an event involving 4 or 8 kernels, etc.? Chance events should affect all categories equivalently, but a change in the frequency of reversion or in the number of mitotic events separating when it is possible for two kernels vs. one to be revertant from a single excision event will preferentially affect specific categories. More information on the ontogeny of the ear is required to resolve this dilemma. In my opinion it seems likely that there are more cell divisions in the lineages restricted to producing a single kernel than in the developmental events separating a two kernel from a four kernel lineage. At least in aleurone tissue, however, the frequency of excision changes over developmental time (A. A. Levy and V. Walbot, 1990 Science 248: 1534) so perhaps there is an increase in Ac activity in the terminal stages of kernel development and/or gametogenesis as well.

Return to the MNL 65 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage