Flowering and shoot elongation of lychee in eastern Australia

We investigated the effects of the timing of shoot elongation on the flowering of lychee (Litchi chinensis Sonn.) in eastern Australia. Trees of cv. Kwai May Pink growing in Alstonville (lat. 28.9° S) were pruned during spring and summer, and subsequent shoot elongation was measured until the following spring. New shoots grew by discrete flushes, with the trees initiating 3, 2, or 1 vegetative shoots prior to winter, according to the pruning sequence. Shoots were vegetative when the mean temperature during early flush development was above 17-19°C, and floral at lower temperatures. Trees with successive flushes commencing in February (late summer) and June (early winter) were more likely to flower than trees with flushes commencing in April and August, because the weather conditions in June were cooler than those in August and more likely to favour induction. The importance of cool weather conditions during early flush development for floral determination was not significantly affected by the number of vegetative flushes to develop between pruning and winter. Having shown that the phase of recurrent flushing affects flowering, we sought to model the process in order to recognise reproductive and non-reproductive cycles along Australia's north-eastern seaboard, and to develop a management strategy for the promotion of flowering. From the results of the Alstonville pruning trial, the interval between successive flushes was regressed against the mean product of daily irradiation and mean daily temperature (°C.MJ/[m2.day]) during the interval. The regression was used in conjunction with long-term weather records to estimate the flush commencement dates required for the completion of 1 or 2 vegetative flushes by the winter solstice at different latitudes. The earliest date for the completion of 1 flush ranged from 16 February in northern New South Wales (lat. 30° S) to 13 March in northern Queensland (lat. 17° S). To test the model, a pruning trial was conducted near Mareeba (lat. 17° S). Trees pruned on 10 February, estimated to produce ≈ 1.5 flushes prior to winter (i.e. flushes in late autumn and early spring, but not in winter), flowered poorly and had low yields. In contrast, trees pruned on 11 March, estimated to produce 1 vegetative flush by winter, had good flowering and yields. Thus, strategic pruning after harvest can be used to manipulate flushing times, so that new, potentially flowering shoots emerge in winter. Cool temperatures are still required for successful flowering, and we provide estimates of the likelihood of such weather in the major growing areas by calculating the annual number of days with a mean temperature <20°C. For Cairns (lat. 16.9° S) the number of such days varied from 0 to 39 from 1888 to 1993, which is consistent with the irregular flowering of lychee in coastal northern Queensland. Our work is the first demonstration for any species that the phase of recurrent flushing affects flowering, and emphasises the interplay between a plant's endogenous developmental cycle and seasonal variations in weather.