Performance of spring wheat lines near-isogenic for the reduced-tillering ‘tin’ trait across a wide range of water-stress environment-types

• Effects of the tiller-inhibition gene tin were quantified in 19 pairs of wheat lines near-isogenic for tin, and in up to 28 water-limited environments.
• G (‘Tin’) × E interactions were significant for grain yield, total biomass, spike number, grains/spike, and number of grains/m2.
• Reduced-tillering plasticity was associated with increased plasticity in numbers of grains/spike and, to a lesser extent, grain weight.
• The marginal yield response to increases in grains/m² was greater in tin-containing lines, i.e. greater plasticity in grains/spike and grain weight can compensate for reduced spike numbers
• The hypothesis of a general yield advantage of tin-containing lines in more severe water-stress environment-types was not confirmed.

Agronomically-important traits were studied in 19 genetically-diverse sets of spring wheat (Triticum aestivum L.) lines near-isogenic (NILs) for the tiller inhibition gene (tin) in contrasting water-stress environment-types (ETs) in the Australian grainbelt between 2010 and 2014. The combination of NILs and ETs generated a wide range of spike numbers (103 to 885 spikes/m2), grain numbers per m2 (GNO; 1400 to 22800 grains/m2), and grain yields (0.3 to 8.2 t/ha). On average, the presence of tin significantly reduced spike numbers (319 vs. 381 spikes/m2), increased the numbers of grains/spike (30 vs. 27 grains/spike), and thousand grain weight (TGW; 38.4 vs. 36.9 g) at maturity. Variation in final biomass and GNO each explained over 70% of the genotypic variation in grain yield. Differences in anthesis biomass explained 38% of the variation in grain yield. The marginal yield response to increases in GNO was slightly greater in the tin-containing lines while the response to increases in final biomass was greater in the non-tin lines (p ≤ 0.016). The harvest index varied with changes in ET (p < 0.001) only. The interaction between presence/absence of tin and environment-type was significant for grain yield, total biomass, spike number, grains/spike, and GNO. Grain yields were similar between NIL groups in the lower-yielding ETs (mean yield <2 t/ha) whereas non-tin lines were higher-yielding in more favourable ETs (mean yield >2 t/ha). The rate of the response in grain yield, total biomass, spike number, and GNO to changes in environmental conditions from ‘unfavourable’ to ‘favourable’ was significantly greater in the non-tin lines, i.e. tin-containing genotypes were less responsive or plastic. Reduced-tillering plasticity was associated with increased plasticity in numbers of grains/spike and, to a lesser extent, TGW. The tin lines produced more grains/spike in ETs averaging 28 to 33 grains/spike while NIL groups had similar grains/spike in ETs with lower means. A slightly greater marginal yield response to increases in GNO in tin lines indicated that greater plasticity in the number of grains/spike and TGW can compensate for reduced spike numbers. It is concluded that large variability in adaptation exists within the group of tin-containing genotypes providing opportunities for further selection to improve wheat performance in environments yielding less than 2 t/ha. Further work on the analysis of GNO and yield determination in tin wheat should also consider changes in resource utilisation and hence competition between plants as mediated by planting arrangement.