INDIVIDUAL GRAIN WEIGHT RESPONSES TO GENETIC REDUCTION IN CULM LENGTHIN WHEAT AS AFFECTED BY SOURCE-SINK MANIPULATIONS

This paper describes the effects of Rht alleles in isogenic lines of M aringa spring wheat (Triticum aestivum L.) on number of grains per spi ke, average grain weight and weight of individual grains from differen t positions within the spike. Plants were grown under five environment al conditions in the field (in the 1991, 1992 and 1993 growing seasons , with two sowing dates in the first two years), aiming to determine c auses of lower average grain weight commonly observed in semi-dwarf (S D) compared with standard-height (SH) cultivars. The number of grains per spike was significantly greater in SD and dwarf (DD) lines than in the SH line due to differences among lines in number of grains per sp ikelet. Therefore, the relative contribution of proximal grains was af fected negatively by Rht alleles while the contribution of distal grai ns was affected positively. Average grain weight was reduced with incr eases in the dose of Rht alleles, and this trait was negatively correl ated to number of grains per spike. To understand the possible causes of this negative relationship, individual grains from specific positio ns within the spike were analysed. Basal grains in central spikelets w ere heavier than those in near apical and near basal spikelets. Within central spikelets, the ranking of individual grain weights were grain s 2 > 1 > 3 > 4, numbered from the most proximal to the most distal po sitions. Proximal grains (1 and 2) were heaviest in SH, lightest in DD , and intermediate in the SD line. In general, the relative difference s between the lines were smaller than that found for the average of al l grains of the spike, in particular when grains of different position s within the central spikelets were considered. For all these grains, differences in weight due to Rht alleles were due to differences in ra te of grain filling, with the effective duration of grain filling bein g almost unchanged. An increased source-sink relationship did not sign ificantly modify the weight of any of the analysed grains in SD and SH lines (and just slightly increased rate of grain filling in the DD li ne). Therefore, the smaller grains in SD and DD lines could hardly hav e been due to an increased competition for assimilates. Alternatively, SD and DD lines had a greater proportion of grains from distal positi ons than the SH line, and these grains were always smaller than proxim al grains. The reduction in average grain weight produced by Rht allel es therefore was due to a combination of effects on the potential size of each grain and on the contribution of grains from distal positions within the spike, with no effects attributed to an increased competit ion for assimilates.