Use of Osmia bicornis L. for Pollination of Cyclamen persicum Mill. Cultivated in Greenhouse Environment During Winter Period

Pollination is widely recognized as a critical ecosystem service for the majority of economically relevant crops that require an effective pollination system (Klein et al., 2007). Over 75% of the major world crops and 80% of all flowering plant species rely on animal pollinators, mostly species of wild bee and other wildlife (Kluser & Peduzzi, 2007). Currently, the most abundant and most effective pollinator in agriculture is the honeybee (Apis mellifera) (Paudel et al., 2015). However, there is a notable decline in honeybee survival with many populations under severe anthropogenic stressors, including pesticide pollution (Hristov et al., 2020). Accordingly, there is an increased interest in the agricultural exploitation of alternative pollinator insects, e.g. native bees (Bosch et al., 2021; Gruber et al., 2011; Maccagnani et al., 2007; Oliveira & Schlindwein, 2009). Several native wild bee species have been identified as capable of supplementing and perhaps, in some settings, even replacing the decreasing populations of honeybees (Rader et al., 2009). For example, the red mason bee (Osmia bicornis syn. O. rufa) reproduced in controlled conditions has been increasingly postulated as the alternative pollinator for entomophilic plants (Kemp, 2000; Ladurner et al., 2004; Komissar, 2005).

Reportedly, bees of the genus Osmia (family: Megachilidae) are effective pollinators of early flowering plants in orchards (Krunic, & Stanisavljevic, 2006; Monzon et al., 2004; Ryder et al., 2020). Several managed species of mason bees have been successfully used as agricultural pollinators worldwide (Sedivy & Dorn, 2014). Osmia cornifrons is widely used in central and northern Japan as well as in the USA for the pollination of cherries, plums, peaches and pears (Matsumoto et al., 2009; Vaudo et al., 2020). South European species such as O. cornuta has recently been successfully deployed for the pollination of almond orchards in Spain (Bosch et al., 2021) as well as greenhouse strawberry cultivation due to significant tolerance to enclosed environments (Herrmann et al., 2019).

In Poland, O. bicornis rearing methods as well as effective protection against parasites were developed back in 1978 (Wójtowski & Wilkaniec, 1978). The usefulness of O. bicornis bees for local agriculture has been identified based on the analysis of pollen quantitates collected by the females in various biotopes. Gruber et al. (2011) and Schindler & Peters (2011) proved to be an excellent pollinator for fruit trees. Thus, it is successfully used for orchard pollination (Benedek, 2008; Hansted et al., 2014; Ryder et al., 2020) as well as for such seed crops as oilseed rape and onions (Wilkaniec et al., 2004 ; Jauker et al., 2012). In contrast to the honeybee, the adult O. bicornis bees thrive in greenhouses cultivation environments, providing new vistas for commercial cultivars.

In warm climates, the Persian cyclamen (Cyclamen persicum Mill.) can be used in perennial gardens, but in most of Europe, it is a major commercial ornamental plant grown in the winter months. It is reproduced from seedlings and grown in the winter, commonly planted on indoor terraces, flower boxes, hanging baskets or in large beddings. It can also be successfully used as a commercial cut flower, considering it can last for up to two weeks. The blooms come in various colors, including white, pink, red, magenta or lavender. They typically bloom from early winter to spring and are dormant during the summer (Schwartz-Tzachor et al., 2008). The C. persicum flower structure does not show a typical wind-pollinated flower, and the seeds are very rarely formed by self-pollination. To achieve satisfactory yields, pollinator animals are mandatory (Schwartz-Tzachor et al., 2006). The cyclamen flower structure and stamen architecture require a bee behavior called buzz-pollination, which induces anther vibration and pollen release (Nevard et al., 2021).

The major aim of this study was to investigate whether the solitary bee O. bicornis L. can be effectively used for C. persicum pollination in greenhouse conditions during winter months as an alternative to commonly accepted hand pollination methods. This could provide significant commercial benefits for greenhouse horticulture of C. persicum.

The investigations were conducted at a horticultural farm in Zdziechowice (to the south-east of Poznań; 52°18′15.5″N, 17°17′23.9″E) in a test greenhouse. The experimental material comprised of C. persicum ‘Halios’. Briefly, three week old seedlings were planted successively nine months before the commencement of experiments.

Plants in the flower bud stage were placed once a month between October and February in two experimental plots of 3 m² each. One plot was used for pollination with O. bicornis bees while the other one acted as a control, pollinated by hand. Each plot consisted of twenty-five plants. The total number of plants during the five-month period was 250. The plots were separated from one another by polyamide mesh stretched over a metal framework (in the size: 2 × 1,5 × 1,5 m L × W × H). Two polyamide mesh cages were used in one term, which covered twenty-five plants in each group. In this way, plants pollinated by O. bicornis were isolated before bees were introduced into the cage. Hand-pollinated plants were isolated to prevent pollination by bees that might have escaped from their cages. O. bicornis were collected from nests kept in natural conditions in the Dendrological Garden of Poznań University of Life Science (Poland).

Before the experiments, the bees were relocated to a laboratory where nest tubes were dismantled, and cocoons were removed from nest cells. The cocoons were sexed according to their size and position within the nest. Each month different batches of bees were activated in the laboratory with a temperature of 26°C and juvenile hormone analogue methoprene (200 μg of methoprene dissolved in 5 μl acetone (Sigma, Poland) (Robinson et al., 1992). The specimens were then transported to the horticultural farm and introduced into an isolation cage.

For the pollination experiments, both females and males were used (100 bees / isolated plot/month; 1:1 female to male ratio). Hand pollination was performed with the standard method by local commercial cyclamen growers. Briefly, the pollen was transferred with the finger to the flower of another plant. The pollen was collected from five to six flowers by the tapping of the basal corolla ring with the thumb nail, to make the pollen drop from the anthers onto the finger. The thumb was then, moved closer to the center of the corolla to the stigma. The pollen on the finger was taken as well as from two other plants of the same control group.

The pollination effectiveness in both pollination treatments was determined based on the following parameters: (i) the number of flowers and fruits per plant, (ii) number of seeds per fruit, (iii) number of seeds per plant, and (iv) the mass of thousand seeds. The weight of 1000 seeds (g) was calculated according to the method by Ladurner et al. (2002). From the seed yield of each pollination treatment and month, four samples of 100 seeds were weighed. The calculated average value was multiplied by ten, and the average weight for 1000 seeds was obtained. The number of plants for which fruits were counted was reduced from the starting number of plants by 1, 16, 9, 8 and 10 in October; November; December; January and February (respectively), due to the withering of plants.

Statistical analyses and graphs were generated with Statistica software package (Statsoft, v. 12). To enhance the data’s normality, the Box-Cox transformation was used (Box & Cox, 1964). One-way ANOVA (with months used as an independent variable) and Student’s T-test (with pollination method as an independent variable) were then conducted to separate out means within and between experimental groups. Tukey’s test was used as a post hoc test with ANOVA for multiple comparisons of mean values among months, for bees pollination and hand pollination separately. Statistical analysis was carried out at the significance level of α=0.05, and the critical level of significance p was also determined.

There was no significant effect on the number of flowers produced by one plant in any month either for pollination by bees F (4, 120)=0.634; p=0.592) or for hand pollination F (4, 120)=0.978; p=0.413) (Tab. 1). The number of flowers per one plant between bee-pollinated and hand-pollinated plants was also not significantly different each month (Oct: t=0.28; p=0.78; Nov: t=0.51; p=0.61; Dec: t=0.18; p=0.86; Jan: t=0.68; p=0.56; Feb: t=0.69; p=0.49).

The pollination term (month) had a significant effect on the number of fruits per plant in plants pollinated by Osmia bees (F(4, 91)=2.93; p=0.024) but not hand-pollinated (F(4, 71)=0.41; p=0.791). Moreover, December was the most effective month in relation to overall fruits production, but only significantly different from January (p≤0.05, Tukey’s test) (Fig. 1). Except for October and January, significantly more fruits (per plant) were produced by plants pollinated by O. bicornis than by hand (p≤0.05) (Fig. 2 A, B). The month was also found to significantly affect the number of seeds in the capsule (fruit) (bee pollination: F (4, 521)=59.8; p<0.001; hand-pollination: F (4, 365)=15,9; p<0.001) (Fig. 3. and 4.) and total number of seeds per plants (bee pollination: F (4, 91)=115.1; p<0.001; hand-pollination: F (4, 79)=14,2; p<0.001) (Fig. 5). In the period from December to February, statistically significantly more seeds per capsule were produced by plants pollinated by O. bicornis (p<0.05; Fig. 3). Regardless of the month, pollination method did not significantly affect the weight of 1000 seeds (F = (1, 22 = 3,336; p=0.081) as it was very similar (Tab. 2).

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5

In February, the efficiency of pollination by Osmia bees expressed by the number of seeds per capsule was 28.1% higher than that of hand pollination (based on the untransformed data). In each research month, statistically significantly more seeds were produced by plants pollinated by O. bicornis (p<0.05; Fig. 5). Bee pollination was particularly efficient in December when seed yield per plant pollinated by O. bicornis was as much as 56.1% higher than from a hand-pollinated plant; in other months, these differences ranged from 39.6% to 48.4% (untransformed data).

The availability of commercial pollinators is required for high yields of entomophily crops grown in greenhouses. The effectiveness results of flower pollination as well as the seed yields confirm the importance of insects in cyclamen pollination (Schwartz-Tzachor et al., 2006). Considering the pollination requirements of C. persicum, cross-pollination is necessary for achieving high yields of seeds (Mazouz & Djeddi, 2013). Currently in temperate climates in Europe, bumblebee colonies are the only available managed pollinator during winter periods (Velthuis & Van Doorn, 2006). However, colonies require initiation and rearing in the laboratory, and this process takes about ten weeks (Treanore et al., 2021). Our experiment shows the possibility of activating O. bicornis adults early in October, with appropriate vitality and pollination activity that can be utilized for successful pollination of commercial greenhouse C. persicum plantations.

The application of higher temperatures at the end of diapause and post-diapause may cause bees to emerge. This process proceeds slowly, which makes flights and blooming difficult to synchronize. Therefore, apart from the temperature application of a juvenile hormone analogue, methoprene was successfully used to accelerate the end of diapause (Wasielewski et al., 2011; Giejdasz et al., 2021). This enabled us to reduce the Osmia overwintering period, which provided evidence that it can become a very effective pollinator in greenhouse environments even during the winter. Such application may provide significant commercial benefits for the greenhouse horticulture of C. persicum or other covered crops.

The administration of juvenile hormone or its analogue methoprene into adult insects can exert both positive and negative effects on development and reproduction. For instance, a recent report has demonstrated that the glycogen content of sugar-fed female Aedes aegypti after methoprene treatment significantly decreased (Denlinger & Armbruster, 2014). This suggested that treatment with juvenile hormone or its analogue (methoprene) induced vitellogenesis by mobilizing nutritional reserves or causing mortality when the reserves were depleted (Denlinger & Armbruster, 2014). Our previous studies had confirmed the positive impact of methoprene on reproductive traits in O. bicornis (Giejdasz et al. 2021). Although the evaluation of male and female vitality was not the aim of our work, it can be concluded indirectly on the basis of pollination effectiveness that the administration of methoprene did not have any negative effects on the activity of both sexes.

Open pollination of cyclamen flowers is only possible in the region where C. persicum occurs, while artificial hand pollination is used in greenhouse crops, and in breeding where flowers of female parents could be emasculated (Masouleh & Moghaddam, 2020). This pollination method is also used in research, so the cited reports of pollination efficiency refer to the manual method.

Ewald & Schwenkel (1997) and Reinhardt et al. (2008) confirmed that C. persicum has a low fruit yield and the number of fruits is highly variable, which is due to the high inflorescence abortion rate. In the ovary only an average of 20% of the ovules developed into mature seeds and yielded 30.5–56 seeds per capsule in the diploidal C. persicum cultivars (Reinhardt et al., 2008). In our study, we obtained a similar number of seeds per capsule for the red mason bee pollination (range: 27.6–48.8; data before transformation), which was on average 50% higher compared to hand pollination.

However, it is difficult to estimate the importance of pollination for seed set in C. persicum, because the final yield is significantly affected by the quality of the ovules and the germination capacity of the pollen (Kermanshahani et al., 2014; Reinhardt et al., 2008).

Although cross-pollination of C. persicum does not guarantee high yields, it is required to provide a large number of well-germinating and genotypically appropriate pollen (Ewald & Schwenkel, 1997; Kermanshahani et al., 2014). Our study shows that using O. bicornis as a pollen vector is more efficient than using hand pollination.

Compared to single pollination, repeated hand pollination and increasing deposition of pollen grains on the stigma caused a slight increase in the number of seeds per capsule and a very significant increase in the number of fruits, and resulted in twice as many seeds per plant (Ewald & Schwenkel, 1997). It can be speculated that in our study the bees visited the flowers more than once, but our results indicate that C. persicum flowers pollinated by O. bicornis not only set more fruits but the capsules produced more seeds compared to hand-pollinated. Finally, seed yield from a plant pollinated by Osmia bees was twice as high (range: 168–275; mean: 200; data before transformation) as hand-pollinated (range: 85–111; mean: 99).

The highest seed yield obtained in December for both pollination methods suggests a month-plant effect that is possibly a consequence of the ecotype of C. persicum cultivars. Despite being planted in a different habitat, plants retain their flowering regime because their phenological type is determined genetically (Schwartz-Tzachor et al., 2008). Controlling cyclamen phenology in crops is important because yield potential interacts with the life cycle of the plants (Masouleh & Moghaddam, 2020).

In our experiment, the pollination method did not influence the weight of 1000 seeds, which ranged from 6.11 to 7.78 g (untransformed data). For onion and oilseed rape crops, the weight of 1000 seeds obtained as the result of pollination by the mason bee and pollinated with the exclusion of bees was very similar in a cage pollination experiment (Jauker et al., 2012; Wilkaniec et al., 2004). Other reports stipulated that there were significant differences in the weight of 1000 seeds of C. persicum (5.80–2.41 g) when seeds were yielded from plants with various genotypes (Ewald & Schwenkel, 1999). Notably, the seed weight of C. persicum can be also determined by the age of the plant, and the weight of 1000 seeds (3.8–6.0 g) from plants aged 1 to 3 years was lower than in our study (Ayala-Garay et al., 2008). Providing O. bicornis cocoons for different consumers is currently performed only in the spring during the normal physiological activity of bees and summer when wintering is prolonged artificially. Appropriate development control of this bee species will allow pollinators to be obtained during wintertime. We conclude that this new bee management protocol can reduce pollination deficits during winter and improve the profitability of companies cultivating different plants even during the winter months.