Abstract Wheat ( Triticum aestivum L .) breeding efforts have increased grain yield predominantly by raising grain numbers per spikelet, rather than grain weight or spikelet number. However, across a single spike large differences exist in the number of grains per spikelet. The central spikelets produce the highest number of grains in any given genotype while apical and basal spikelets are less productive. Basal spikelets are delayed in development just after initiation and are smaller and less advanced than central spikelets already by the glume primordium stage. However, basal spikelets continue to develop and produce florets until much later in the wheat growth cycle. The precise timings or the cause of their growth cessation, and subsequent abortion, is largely unknown. In this study we investigated the underlying causes of rudimentary basal spikelet abortion. We investigated basal spikelet development in four UK winter wheat varieties as well as a set of near-isogenic lines for VRT-A2 ( VEGETATIVE TO REPRODUCTIVE TRANSITION 2 ) using shading applications in the field. We propose that basal spikelet abortion is likely the consequence of complete floret abortion as both occur at the same time and have the same response to shading treatments. Furthermore, we found that the developmental age of florets pre-abortion is an important factor for their likelihood to survive and develop viable seed. Previously, it had been proposed that reduced assimilate availability in the base of the spike leads to increased abortion. Re-analysis of published data alongside data presented here, however, does not support this model. We found that rather than assimilate availability, it is the reduced developmental age of basal florets before abortion that correlates with increased abortion. Using the floret Waddington developmental stage pre-abortion, we were able to predict final grain set per spikelet across the spike, alongside the characteristic gradient in number of grains from basal to central spikelets. We found that advancing past Waddington stage 5.5 seems to be important for floret survival and that most florets in basal spikelets had not reached this stage at the onset of floret abortion. The abortion of all florets could therefore be the reason for their rudimentary appearance in the mature spike, suggesting that basal spikelet abortion is simply the consequence of all florets inside the spikelet being aborted and thus all other spikelet structures (e.g., lemma, rachilla, glume) also ceasing to develop. Future efforts to improve spikelet homogeneity across the spike could thus focus on improving basal spikelet establishment and increasing floret development rates pre-abortion.

Delayed development of basal spikelets in wheat explains their increased floret abortion and rudimentary nature