Australian Lentil Breeding Between 1988 and 2019 Has Delivered Greater Yield Gain Under Stress Than Under High-Yield...

The contemporary lentil (Lens culinaris ssp. culinaris) industry in Australia started in the late 1980s. Yield in farmers' fields averages 1.2 t ha^-1 nationally and has not increased over three decades. Lack of yield progress can be related to a number of non-mutually exclusive reasons: expansion of lentil to low-yielding environments, lack of genetic gain in yield, lack of progress in agronomic practices, and lack of adoption of superior technologies. The aims of this study were to (i) quantify the genetic gain in lentil yield since 1988, (ii) explore the variation in the expression of genetic gain with the environment, and (iii) identify shifts in crop phenotype associated with selection for yield and agronomic adaptation. We grew a historic collection of 19 varieties released between 1988 and 2019 in eight environments resulting from the factorial combination of two sowing dates, two water regimes, and two seasons. Across environments, yield varied 11-fold from 0.2 to 2.2 t ha^-1. The rate of genetic gain averaged 20 kg ha^-1 year^-1 or 1.23% year^-1 across environments and was higher in low-yield environments. The yield increase was associated with substantial shifts in phenology. Newer varieties had a shorter time to flowering and pod emergence, and the rate of change in these traits was more pronounced in slow-developing environments (e.g., earlier sowing). Thermal time from sowing to end of flowering and maturity were shorter in newer varieties, and thermal time from pod emergence to maturity was longer in newer varieties; the rate of change in these traits was unrelated to developmental drivers and correlated with environmental mean yield. Genetic gain in yield was associated with increased grain number and increased harvest index. Despite their shorter time to maturity, newer varieties had similar or higher biomass than their older counterparts because crop growth rate during the critical period increased with the year of release. Genotype-dependent yield increased over three decades in low-yield environments, whereas actual farm yield has been stagnant; this suggests an increasing yield gap requiring agronomic solutions. Genetic improvement in high-yield environments requires improved coupling of growth and reproduction.