Factors affecting fruit filling in the dioecious saltbush Atriplex amnicola Paul G. Wilson

Secondary salinisation is an increasing problem in Australia. Both native and exotic saltbush (Atriplex) species have been used for rehabilitation and forage production on saline soils, particularly A. amnicola, a species native to north Western Australia, which has very good forage potential. Establishment of saltbush stands is usually by direct sowing of fruits and frequent failure of establishment is due to the very low fruit fill obtained from fruit collected from existing saltbush stands (commonly less than 20%). At present there are very few dedicated seed production stands of Atriplex. In order to establish and manage such stands for maximum seed production, the causes of poor fruit fill must be determined. Earlier observations showed that fruit fill in A. amnicola declined from 70% to 80% in late winter, to between 20% and 30% prior to harvest in December. Limitation of mineral nutrients to mother plants and of assimilates to developing seeds, and environmental and genotypic influences which may account for this seed abortion, were examined in the field and glasshouse. Although initial observations did not indicate that pollination and fertilisation were major influences on fruit fill, they also were examined to determine to what extent they may limit fruit fill in A. amnicola. No incompatibility was seen in the diallele crosses made between five male and five female clones of A. amnicola, but pollen supply did limit fruit fill. A 1 to 1 sex ratio was required to optimise fruit fill and yield. Sexual lability of individual plants in a population of A. amnicola caused the population ratio to shift from a ratio of 3 males to 2 females, to a 1 to 1 sex ratio, but the change in sex ratio had no effect on fruit filling. Seed abortion which may result from nutritional limitations to mother plants and limitations of assimilate to developing seeds, were examined. Addition of nitrogen and phosphorus fertiliser did not improve fruit fill at harvest. This response is not unusual in native Australian plants that are adapted to very low nutrient soils. High levels of fruit pruning (50% and 75%) to reduce competition between sinks for assimilate increased fruit fill by 5%. In unpruned plants, earlier developing fruits in the distal region of stems had higher levels of fruit fill than proximal fruits. By contrast, in the pruned plants, the increase in total fruit fill resulted from increasing proximal fruit fill to the same level as the distal fruits. The effects of salinity, pH, soil texture and compaction, gravimetric water content and depth of water table were examined for their effects on fruit fill in A. amnicola. Duration of waterlogging was found to be the most significant factor causing seed abortion. Genotypic influences were also very important in determining fruit fill. Of the three female clones tested in the field, one clone had a mean fruit fill of 70% compared to 40% for the others. However, this clonal difference was not maintained under the most severe levels of waterlogging and salinity, indicating that environmental effects were more significant. To assess further the effects of waterlogging on fruit fill and seed abortion in A. amnicola, a controlled pot experiment was conducted, which aimed to mimic the climatic conditions in the field. Plants were waterlogged for 5 weeks in the middle of winter, for 5 weeks at the onset of spring when temperatures began to rise and for 10 weeks throughout winter and spring. Waterlogging of Atriplex amnicola, for up to ten weeks had no affect on fruit fill. This is in direct contrast with results from the field where waterlogging reduced fill by approximately 15% after 30 days and 70% after 135 days. For all treatments, fruit fill decreased from 50% to 27% in the first five weeks. This discrepancy between field and controlled experimental results may be due to the conditions of low vapour pressure deficit (VPD) which existed for the controlled experiments. The evapotranspirational demand of A. amnicola under those conditions was such that the effect of waterlogging for 10 weeks was minimal. However, the increased seed abortion seen in the field may have occurred under conditions of high VPD, where evapotranspirational demand was much greater and the waterlogged roots of the plants were unable to take up enough water. The seed abortion which occurred in the first five weeks was strongly correlated with large rainfall events which may have led to substantial dilution for short periods, of the salt solution used to maintain waterlogging. Lack of available ions to supply the rapid uptake required to maintain turgor may have accounted for the seed abortion. By contrast to the lack of effects on fruit fill, waterlogging treatment significantly depressed fruit yield. Waterlogging for ten weeks decreased yield to the same extent as the 5 week treatment in early spring. Both treatments decreased yield to a greater extent than waterlogging for 5 weeks during winter. This effect supports the results from the field where yield was reduced by nearly 80% by 135 days of waterlogging. It is concluded from the present study that the establishment of dedicated seed production plots requires selected female genotypes planted at a 1 to 1 ratio with male genotypes, on mildly saline sites that are not subject to prolonged periods of waterlogging in winter. The interactive effects of VPD, waterlogging and salinity on fruit fill in A. amnicola need to be tested to allow predictions to be made on the quality of seed crops in different seasons