Yield response to frost in a set of historic wheat varieties.

Low temperature during the critical period of wheat can cause floret sterility, grain abortion, yield reduction, and economic losses. We tested the hypothesis that selection for yield improved frost tolerance. We measured phenology, yield and its components, including the distribution of grains within the spike, in a factorial experiment combining 12 wheat varieties released from 1973 to 2015, two sowing dates (mid-April and early or mid-May), two thermal regimes, ambient control and frost-protected, and two seasons. To protect crops from frost we used moveable, lightweight passive heating chambers before each frost event (≤ 0 °C). The rate of genetic gain in yield aligned with frost severity: it ranged from negligible under severe frost to 0.79 ± 0.05% y−1 in unfrosted crops. The chambers diminished but did not fully prevent frost; across sources of variation grain yield was 179 ± 20.1 g m−2 in ambient control and 264 ± 13.3 g m−2 in frost-protected treatments. Frost intensity was calculated as the cumulative, canopy-level minimum temperature below 2 °C during the period from −300 to + 100 °C d centred at anthesis. For the pooled data, yield declined with frost intensity at 1.24 ± 0.12% °C−1. Across varieties and environments, frost-protected crops set 29 ± 3.0 grains spike−1 compared with 19 ± 2.7 in ambient control. We did not find random grain abortion within spikelets, but random sterile spikelets. Average grain weight (weighed by the relative contributions of these grains to the number of grains per spike) in main shoot spike did not vary with thermal regime. Selection for yield did not improve frost tolerance. Preventing spring radiation frost improved spike fertility by increasing the proportion of grains in distal positions with no reduction in grain weight potential within spikelets in relation to controls. We advanced damage functions relating yield and frost, which are useful for modelling, and maps of grain set in spikes that help understanding the physiology of yield in response to frost. • We advance damage functions relating yield and frost, which are useful for modelling. • We quantified frost intensity as the cumulative minimum temperature below 2 ℃ during the critical period. • Wheat grain yield declined, on average, at 1.24% °C−1 with frost intensity. • The rate of genetic change in yield, from 1973 to 2015, declined with frost severity. • Frost-protection improved the proportion of grains in distal positions with no reduction in grain weight. [ABSTRACT FROM AUTHOR]