Pollen production in olive cultivars and its interannual variability

Abstract

Background and aims: Olive (Olea europaea subsp. europaea var. europaea) is the most extensively cultivated fruit crop worldwide. It is considered a wind-pollinated and strictly outcrossing crop. Thus, elevated pollen production is crucial to guarantee optimum fruit set and yield. Despite these facts, the variability of pollen production within the cultivated olive has been scarcely studied. This study aimed to characterize this feature by analysing a representative set of worldwide olive cultivars.

Methods: We evaluated the average number of pollen grains per anther in 57 principal cultivars over three consecutive years. We applied a standard generalized linear model (GLM) approach to study the influence of cultivar, year and the previous year's fruit load on the amount of pollen per anther. Additionally, the K-means method was used for cluster analysis to group cultivars based on their pollen production capacity.

Key results: Pollen production per anther was highly variable among olive cultivars. The cultivar significantly accounted for 51.3 % of the variance in pollen production and the year for 0.3 %. The interaction between the two factors explained 8.4 % of the variance, indicating that not all cultivars were equally stable in producing pollen across the years. The previous year's fruit load and its interaction with the year were significant, but barely accounted for 1.5 % of the variance. Olive cultivars were classified into four clusters according to their capacity to produce pollen. Interestingly, the fourth cluster was composed of male-sterile cultivars, which presumably share this character by inheritance.

Conclusions: Pollen production per anther varied extensively within the cultivated olive. This variation was mainly driven by the cultivar and its interaction with the year. The differential capacity of olive cultivars to produce pollen should be considered not only for designing new orchards but also gardens where this species is used as an ornamental.

Keywords: Olea europaea; androsterility; biennial bearing; clustering; fruit load; pollen per anther; variability; wind pollination.

Fig. 1.

Distribution of olive pollen production during three consecutive years represented as density charts. Cultivars ‘Hojiblanca’ and ‘Mari’ presented a normal distribution, while ‘Kato Drys’ and ‘Royal de Sabiñán’ showed a bimodal distribution.

Fig. 2.

Frequency distribution of pollen production shown by olive cultivars during the three evaluated years. The relative frequency was calculated for the average of all the repetitions for the same cultivar. Each graph shows the mean number of pollen grains per anther (red dashed line; 2019 = 53 228 ± 16 339; 2020, 46 118 ± 15 757; 2021, 56 469 ± 17 052) and density (light grey area).

Fig. 3.

Frequency distribution of average pollen production of the cultivars during the three evaluated years (2019–21). Relative frequency was calculated for the average of all repetitions for the same cultivar. Different colours show the three homogeneous clusters according to their pollen production capacity: Cluster I in blue, Cluster II in green and Cluster III in red. Letters indicate significant differences after Bonferroni correction (P < 0.05).

Fig. 4.

Inflorescences and undeveloped pollen grains stained with fuchsin (scale bar = 50 μm) in the androsterile cultivar ‘Vera’ (A and C) and in the excellent pollen donor ‘Arbequina’ (B and D).