Various source and sink related traits participate in the response to elevated CO2
Observed trait means and analysis of variance (ANOVA) are presented in Table S2 (supplemental materials). Figure 3 presents correlation heatmaps for the observed traits under ambient (A) and elevated (B) CO2 levels, whereby the CO2 effect onA max and grain yield (trait variation) was included in both matrices. The A max and grain yield variation between the two environments were positively correlated (P<0.01) with each other, and both were negatively correlated with LSSR as also shown in Fig. 2. They were similarly correlated with two other calculated, potential, proxy traits for source-sink ratio: the LSSR(sink) that uses the product of panicle spikelet number and the genotypic filled-grain weight (instead of just the spikelet number) as an estimate of the sink; and [LSSR(sink) * SPAD] which factors the area leaf chlorophyll content into the equation to strengthen the source term (Table S1). In fact, A max and the electron transport rate ETR were positively correlated with SPAD (P<0.01) under ambient but not under elevated CO2 levels (Fig. 3). The ETR was negatively correlated with SLA (P<0.05) in both environments. Leaf sucrose concentration was weakly, positively correlated (P<0.05) with the tentative proxy traits LSSR, LSSR(sink) and LSSR(sink)*SPAD under elevated CO2, but not ambient CO2 level. The spikelet number per panicle was positively correlated with flag leaf ETR (P<0.05). The SPAD chlorophyll content was generally negatively correlated with SLA but positively with most of the other morphological traits such as leaf dimensions, panicle size traits and total shoot dry matter.