2020 Volume 64 Issue 2 Pages 54-60
Passion fruit plants were cultivated under greenhouse conditions in a warm-temperate zone to study the relationship between daily number of flowers (F) and cumulative hours (H) of a range of temperature regimes, and then the specific temperature regimes and the periods that affected F were determined. Multiple regression analysis with H as a variable was performed using stepwise backward elimination to develop a predictive model of F fluctuation. The F predictive model was then verified using F data collected from open orchards. The strongest correlations with H were detected at 15 days before anthesis (DBA). Significant positive and negative correlations with H<25°C and H>25°C, respectively, were detected during a period of several days around 15 DBA. Longer hours with temperatures <25°C increased F, whereas longer hours with temperatures >25°C decreased F. The strongest positive correlations were observed at H20–25°C, and the strongest negative correlations at H>30ºC. F decreased greatly when H>30ºC exceeded 6 h at 15 DBA, indicating that temperatures higher than 30°C severely affected F. F was not significantly correlated with H10–15ºC, nor H15–20ºC except for 19–21 DBA, indicating that the effect of temperatures <20°C on F was negligible. Within 1 week before anthesis, the effect of temperature on F was also negligible. Therefore, F fluctuation in passion fruit was accurately predicted about 2 weeks prior to flowering by the following equation for F as a function of H during 15 ± 3 DBA:
F = b1H10–15ºC + b2H15–20ºC + b3H20–25ºC + b4H25–30ºC + b5H>30ºC,
where b1 = 0.026, b2 = –0.0092, b3 = 0.013, b4 = 0.0066, and b5 = –0.012.
