Figure 1.
Effect of water stress and competition on phenological traits and fitness.
INT: interval between bolting and anthesis, ANT: anthesis, FLO: flowering, RP: reproductive period duration, FRR: flowering-to-reproductive period ratio, FITNESS: total silique length as a proxy of seed production. INT, FT, FP and RP are expressed in days. FITNESS is expressed in millimeters. Ctl: ‘control’ treatment. W−: ‘moderate water stress’ treatment. W+: ‘severe water stress’ treatment. C−: ‘moderate competition’ treatment. C+: ‘intense competition’ treatment. For each phenotypic trait, different letters indicate different phenotypic means among treatments after a Tukey's test of multiple comparisons of means (P = 0.05). Data are not available for RP and FRR in the two ‘water stress’ treatments (see Material and Methods).
Table 1.
Phenotypic selection analysis with selection differentials (S), selection gradients (β), quadratic selection coefficients (γ) for phenological traits in each treatment.
Table 2.
Genotypic selection analysis with selection differentials (S), selection gradients (β), quadratic selection coefficients (γ) for phenological traits in each treatment.
Figure 2.
Relationship between bolting time and fitness for each treatment at both phenotypic and genotypic levels.
For illustration purposes, a polynomial regression including both linear and quadratic terms described either in Table 1 (phenotypic level) or in Table 2 (genotypic level) was first performed including all traits but bolting time. Then, a second polynomial regression including the linear and quadratic terms associated with bolting time was run on the residual fitness of the first polynomial regression. The black lines were drawn using the parameters from this second polynomial regression. BT: bolting time. BT is expressed in standardized values.
Figure 3.
Relationship between flowering and fitness for each treatment at phenotypic and genotypic levels.
For illustration purposes, a polynomial regression including both linear and quadratic terms described either in Table 1 (phenotypic level) or in Table 2 (genotypic level) was first performed including all traits but flowering duration. Then, a second polynomial regression including the linear and quadratic terms associated with flowering duration was run on the residual fitness of the first polynomial regression. The black lines were drawn using the parameters from this second polynomial regression. FLO: flowering. FLO is expressed in standardized values. The absence of a black line indicates that there is no significant linear or quadratic relationship between fitness and flowering (see Tables 1 and 2).
Table 3.
Estimates of broad-sense heritability (H2) for seven phenological traits in Arabidopsis thaliana in five treatments.
Table 4.
Correlations among phenological traits in five treatments.
Table 5.
Evolution of phenological traits in experimental evolution populations of Arabidopsis thaliana in the ‘water stress’ and ‘competition’ treatments.
Figure 4.
Evolution of bolting time in experimental populations of Arabidopsis thaliana in the ‘water stress’ treatments.
Control: ‘control’ treatment. W1–W4 correspond to the four water stress intensities simulating mild to severe drought (i.e. bottom-watering stopped either 46, 39, 32 or 25 days after sowing for W1, W2, W3 and W4, respectively). BT (bolting time) is expressed in days. G0: initial experimental generation, G1–G4: four successive experimental generations. For each treatment, different letters indicate different phenotypic means among generations after a Tukey's test of multiple comparisons of means (P = 0.05). For each treatment, a box-plot representing raw data is given for each experimental generation.