Main Article Content
The question of whether tillers are a burden or a resource in durum wheat is of concern in the variable Mediterranean climates. The contribution of tillers to grain yield was investigated in commercial cultivars differing in time to anthesis, tillering and spike size, in response to three sowing dates:mid-autumn (recommended), winter, and early spring. The thermal time of phenological phases was calculated, and yield-components and floret production were analysed separately in main culm and tillers. Tiller spikes showed higher spikelet abortion coupled to lower spikelet fertility and mean kernel weight, so that grain yield was 40-60% lower than in main culm spikes. Despite this, tillers contributed 35 to 50% to plant yield. The sowing date affected tiller number rather than one tiller yield. In winter sowings (December), lower main culm yield was fully compensated by increased tiller yield, whereas shifts of sowing date to early spring (February) reduced tillering, which caused a yield loss ranging from 12 to 20%. Cultivars differed in one tiller yield rather than in tiller number, and higher grain yield of tillers was primarily due to increased grain recovery. A more equal partitioning of resources within main culm and tillers corresponded to better yield stability across sowing dates. Starting from this, we suggest that early anthesis, a long stem elongation phase, a high primordium initiation-rate and small spikes, could be positive traits for durum wheat yield stability in changing environments, since they allow plants directing more time and resources to floret production and grain filling both in main culm and tiller spikes. From a methodological point of view, our results show that the number of fertile florets per spike is highly correlated with the average floret number of five given spikelets.