Fig 1.
Floral reversion phenotype in Arabidopsis suecica.
A: Normal flower. B: Reverting flower with a bulging carpel inside an otherwise normal flower. C: Isolated carpel from reverting flower. Note the unevenness of the size of the two valves at the basal end, and the stretched replum (R) between the valves. D: Reverting flower after petal fall. One petal remains marking the floral scar (FS). Note the burst-open carpel with approximately six additional flowers inside. E: Elongated siliques of the flowers inside a reverting carpel (asterisks). F: Reverting flower with the new inflorescence fully expanded. Four additional siliques are formed, three additional flowers are at petal fall stage. Note the elongated gynophore (EG) between the FS and the base of the reverting carpel. Scale bar = 1 mm (A-E) or 2 cm (F).
Fig 2.
Floral reversion in A. suecica originates from the replum.
A: The top valve of a reverting carpel was removed and the tissue imaged using environmental SEM. Multiple new floral stems (light green) originate out of the replum (dark green), which is connected to the gynophore (bottom right). The original stigma of the reverting flower is colored yellow for orientation. B: The same reverting carpel as in A. In this view the new inflorescence was unfolded and placed above its remaining carpel valve. Gynophore and replum are fused (dark green), and several new inflorescence stems branch off of the replum (light green). Stigma is colored in yellow as above. Colors are false colors. Size bar in B is the same as for A = 25 mm.
Fig 3.
The total seed number is increased in reverting compared to non-reverting flowers.
Seeds were harvested and counted from individual typical, non-reverting siliques (N = 27), and compared to the average of the sum of seeds from all flowers on new inflorescences (N = 21) erupting from reverting flowers. A two-sample Welch t-test, assuming unequal variances, indicated a statistically significant difference between the samples (t = 6.319; p < 0.001). Error bars reflect SE.
Fig 4.
Floral reversion frequency increases in short days.
Individuals from each population were grown in incubators (20°C; and day lengths set at 8h light/16h dark; 12h light/12h dark; 16h light/8h dark; 24h light) until senescence. Reverting flowers were counted along the inflorescence axis from the first (oldest) to the 80th flower position and the proportion of total reverting flowers was calculated for each condition. Here, data from all eleven populations were analyzed together. The plot on the left includes data from all plants from the study, the plot on the right excludes plants that exhibited no case of floral reversion. Statistically significant differences in reversion rates between treatments were determined using ANOVA and Tukey posthoc testing. Treatments not connected by a letter are statistically significantly different from each other (p< 0.05). Boxplots show data range, including median (bold line), the 25th, 75th, and 5th or 95th percentile of data distribution (lower and upper end of box, and lower and upper whiskers, respectively), plus outliers (circles). N = 148 (reverting), 190 (never reverting).
Fig 5.
Floral reversion is more likely to occur earlier in development when days are short.
To determine if reversion was equally likely for each position along the inflorescence from the oldest to the 50th flower, “weighted reversion frequencies” (see Methods for details) were calculated for each day length condition and plotted. Statistical significance between the positions of the peaks was calculated using ANOVA and Tukey post hoc testing. Statistically significant differences in peak positions for reversions were found only for the two extreme values of 8h and 24h of light.
Fig 6.
Flowering time is significantly increased in shorter days.
Plants of eleven populations were grown in growth chambers at the indicated day lengths. The first day of flowering was defined as the day when the first flower opened or the inflorescence was 1 cm tall, whichever occurred first. ANOVA, followed by Tukey post hoc testing was performed for the combined data set of all populations of A. suecica (N = 270; data for all individual populations can be found in S3 Fig). Treatments not connected by the same letter are statistically significantly different. (For general boxplot description see Fig 4.)
Fig 7.
Floral reversion is positively correlated with flowering time.
Flowering time and reversion data were plotted for each individual plant that showed any reversion, and the Pearson correlation coefficient was calculated (p< 0.0001, R2 = 0.169, N = 148).