Pollination Studies on Both Floret and Capitulum Levels in an Oil Seed Crop, Guizotia abyssinica (L.f.) Cass.

A total of thirty capitula were tagged and bagged with fine transparent muslin cloth bags to prevent insect visitation. The bags were removed carefully and the different phenological events were observed. While recording the observations bee visits to the tagged capitula were restrained considering the possibility of pollen deposition on the unopened anthers by foraging activity that may have resulted in wrong interpretation of anther dehiscence data. Each capitulum consisted of five or six rows of disc florets which followed centripetal succession for anthesis. Every day seven to ten florets per capitulum were opening, and these were used for the observing of the time of anthesis, anther dehiscence, stigmatic lobes opening and floral longevity (n=30 capitula).

To record the stigma receptivity, a set of disc florets (n=130 florets; 5 florets/capitulum) were emasculated early in the morning before 06.00 h. Hand pollination was done at hourly intervals with a camel hair brush from 6 am to 6 pm using pollen from a neighbouring G. abyssinica plant. Similarly, stigma receptivity of ray florets (female flowers) were studied. Based on the preliminary results, the procedure was repeated at hourly intervals to record the time of peak stigma receptivity. Hand-pollinated pistils were fixed in Carnoy's fixative (absolute ethanol: chloroform: acetic acid; 6:3:1) twenty four hours after pollination. Pistils were cleared and stained according to methods proposed by Huang et al. (2004) and Veereshkumar et al. (2017). The percentage of germinated pollen grain was scored for receptivity. Pollen viability was estimated using 1% acetocarmine (Stanley & Linskens, 1974).

Quantity of nectar produced in the flowers, was estimated using calibrated capillary tubes (1mm diameter with 1 microliter volume). Nectar volume was measured at two-hour intervals (n=10 florets) from morning 6 am to 6 pm with a method adopted from Shrishail et al. (2011). Every time, nectar was estimated from the same florets.

Fully opened florets opened on the same day (n=10) were brought to the laboratory and examined for detailed floral structure (Veereshkumar et al., 2021). The floral parts were measured and photographed using a stereo binocular microscope (Leica M205C with auto-mountage and Leica DFC450 camera).

Pollinator abundance was recorded during 25, 50, 75, and >90 % of floral density in the field. Each sampling day was divided into three time intervals of four hours each from 06.00 to 18.00 h. Samplings were done through visual counting of flower visitors for 15 minutes per predetermined transect of 1m² area, and this was repeated thrice in an hour with a five-minute gap between two subsequent transects. Thus, we had three samples/hour and a total of thirty-six samples/sampling day. This was repeated twice during 25, 50, 75 and >90% of floral abundance in the field (Laroca & Orth, 2002; Belavadi & Ganeshaiah, 2013; Revanasidda & Belavadi, 2019; Veereshkumar et al., 2019). The percentage of flower abundance was calculated by number of plants produced flowers/m² area (Belavadi & Ganeshaiah, 2013). The most frequent flower visitors of Niger were recognized during the sampling and further observed for their foraging behavior as described by Mattu et al. (2012), Belavadi & Ganeshaiah (2013), Sushil et al. (2013). Time spent by an individual bee per floret and capitulum (n=30 bees) and the number of capitula and florets visited by a single bee (n=30 bees) per five minute were recorded. The total number of visits received per floret were also recorded and observed, and capitulum for whole day and the average time lag between two subsequent bee visits were recorded before and after the peak stigma receptivity (n=30).

Number of pollen deposited: A set of capitulum were tagged and bagged early in the morning at 06.00 h to prevent bee visitation (n=30 capitulum). Each capitulum was unbagged and observed until an insect contacted a stigma. Insect species was recorded and then the stigma was carefully removed from the top of the style using fine-pointed forceps (Dafni, 1992). The collected stigma was transferred to 2 ml plastic vials and brought to the laboratory. Stigmatic surface was observed under stereo binocular microscope to count the number of pollens deposited by bees per visit.

A field experiment was performed with seven pollination treatments (T1–T7) to determine the mating system of Niger and the role of flower visitors in seed set. T1) Florets compatibility: hollow polythene tube (7 mm diameter, 2 cm length) (Supplementary Fig.1) was placed on individual florets in a capitulum before the anther dehiscence. Pollen from the respective floret was dusted on stigmatic surface with the help of a camel hair brush (n=30). T2) Capitulum compatibility: florets within a capitulum were randomly pollinated using pollen from other florets of the same capitulum and were covered with butter paper bags (n=30 capitula). T3) Compatibility of capitula from same branch: capitula from the same branch were tagged and hand pollinated with the pollen from other capitula of the same branch (n=30). T4) Compatibility between capitula from different branches: capitula were hand pollinated with the pollen from a capitulum of different branch of the same plant (n=30). T5) Compatibility between capitula from different plants: capitula were hand pollinated with pollen of another plant (n=30). T6) Caged plants: Plants were caged with a nylon mesh net to prevent pollinator visits (n=30) T7) Open pollination: capitula were allowed for open pollination (n=30). Flowers treated with different sources of pollen were in the same stage of development. We recorded the percentage of seed set in each pollination treatment.

In the experiment of controlled bee visits, Apis cerana and A. florea (1, 2, 3, 4 and 5 bee visits per capitulum) (n=30 each) were allowed to perform a fixed number of visits on tagged capitula and the percentage of seed set was recorded. A set of capitula were bagged randomly and then were later exposed individually to a single natural visit of Apis cerena. After the visitation, they were bagged again with nylon mesh to prevent further bee visitations and allowed for seed set. Similarly, for another set of capitulum, A. cerana was allowed to visit two times the same capitulum. Likewise, three, four and five visits were allowed to a set of capitulum and recorded percentage of seed set (n=30 capitula each). During the process, other flower visitors were chased away with the help of a sweeping net. This experiment was performed similarly with A. florea.

The data are presented as Mean±SE in the results. The abundance of flower visitors was further subjected to a simple correlation coefficient analysis with prevailing weather conditions and the capitula abundance (Thriveni, 2019). The percentage of seed set was calculated as described by Revanasidda & Belavadi (2019). Before ANOVA was performed, arcsine transformation (percent data) was used to normalize the data. One-way ANOVA was used for the analysis of data on the percentage of seed set influenced by pollination treatments. An analysis of the transformed data was conducted, and the untransformed means±standard error are presented (supplementary). Two-way ANOVA was used to analyse the number of bee visits, bee species and their interactive effects on percent seed set using PROCGLM (SAS version 9.3, 2011; SAS Institute, Cary, NC, USA). When ANOVA indicated a significant F-value (α<0.05), the pollination treatment effects were further separated using Tukey's post hoc test.

Niger is an annual herb and produced 22±0.14 (Mean±SE) primary branches (n=30 plants) with an average of 25±0.18 capitula per branch (n=100 branches) (inflorescence of Asteraceae containing set of florets). Each plant produced 336 to 729 capitula, each capitulum is 1.5 to 2.2 cm in diameter (Fig. 1A), yellowish in color and with 5–6 rows of disc florets which followed centripetal succession for anthesis. There were 41–58 disc florets in a capitulum surrounded by a row of ray florets (7–12) and five green bracts. The disc florets were hermaphrodites with five yellowish orange united petals, five anthers and a centrally located style originating from the ovary, and the stigma was densely hairy. Each floret measured 11.90±0.08 mm (n=30) in length, each disc floret was surrounded by a green ligule and ray florets were pistillate, measuring about 16.61±0.43 mm (n=30) in length. The longevity of each capitulum, ray floret, and disc floret was 8.66±0.11, 6.63±0.10, and 2.47±0.09 days (n=30), respectively. Capitulum, disk floret and floral phases are represented in Fig. 1A–E. Anthesis occurred between 06.00 and 08.00 h (peak between 06.30 and 07.30 h), afterwards, the anther lobe projected outside and extended up to 2.89±0.28 mm (n=30) within three to four hours and then dehiscence (Fig. 1C) started eight to eleven hours after anthesis (between 14.00 to 17.00 h with peak dehiscence between 15.00 to 15.30 h). The style extended beyond the anther lobe up to 2.41±0.20 mm (n=30) within two hours of anther dehiscence and then the stigmatic lobes split open to form a bilobed surface (Fig. 1D). It was also observed that if a floret was not pollinated, the lobes of stigma curved further downwards to expose a greater surface area for pollinators (Fig. 1E).

Fig. 1

Stigma became receptive nine hours after anthesis and lasted for 24–25 hours. The peak receptivity with 68.2% pollen germination was observed twenty-eight hours after anthesis between 10.00 and 14.00 h on the following day. Pollen grains remained viable for up to twenty-eight hours after anther dehiscence from 14.00 h (day one) to 18.00 h (day 2). Nectar cavity was located at the base of the style (Fig. 2A), and the peak nectar secretion (0.28±0.08 μl per floret (n=10)) coincided with peak stigma receptivity (Fig.7). Ray florets did not produce nectar. Illustrations of floral biology are given in Fig. 3.

Fig. 2

Fig. 3

A total of eighteen insect species representing six families visited Niger flowers (Supplementary Tab. 1). Hymenopterans were the most abundant (93.05%) followed by Diptera (6.16%) and Lepidoptera (0.79%). Among flower visitors, Apis cerana and A. florea constituted 90.34%, and so only these two species were observed for foraging behavior. These bees commenced their activity around 07.30 h; A. cerana had two peaks of activity, one between 10.00–11.00 h and the other between 16.00–17.00 h, while A. florea had a single peak of activity between 11.00–14.00 h (Fig. 4). The abundance of bees was positively correlated with capitulum abundance (r=0.99) and maximum air temperature (r=0.89), while it was negatively correlated with relative humidity (r=−0.47) and rainfall (r=−0.92) (Supplementary Fig. 2).

Fig. 4

The mean time spent per floret and capitulum by A. cerana was 1.35±0.04 and 9.63±0.69 seconds (n=30) and that by A. florea was 1.83±0.12 and 14.9±0.94 seconds (n=30), respectively (Fig. 5). On average, 23.33±0.37 and 47.50±0.83 visits were received for a single floret and capitulum, respectively, in a day. A single A. cerana forager visited 151±1.89 florets and 29.00±0.77 capitula per five minutes, whereas A. florea visited 135.66±8.62 florets and 26.75±1.03 capitula per five minutes. An A. cerana forager deposited a slightly greater number of pollen grains per visit (8.15±1.05 (n=10 florets)) on the stigmatic surface compared to an A. florea forager (6.90±0.72 (n=10 florets)).

Fig. 5

The time lag differed between two subsequent bee visits recorded before and after peak stigma receptivity. Bee visitation was more frequent during the peak stigma receptivity. A floret was visited by bee once in every 3.13±1.04 minutes at peak stigma receptivity and once every 8.58±1.26 minutes before and after the peak stigma receptivity. Similarly, the rate of nectar secretion was more (0.28±0.08 μl nectar/disc floret) at the time of peak stigma receptivity compared to before and after the peak stigmatic receptivity period (09±0.02 μl nectar/disc floret) (Fig. 6).

Fig. 6

The percentage of seed set significantly varied among the pollination treatments (F=4285.12, df=6, 203; P<0.001) (Tab. 1). The maximum (78.33±0.14) was recorded in T7 (P<0.05) when flowers were allowed for open pollination, and the lowest was recorded in T6 (0.33±0.67) when the whole plant was caged to exclude pollinator visits. However, no seed set was observed during either floret selfing or capitulum selfing (Fig. 2B).

The role of A. cerana and A. florea visits on pollination was studied with controlled bee visits (1–5 visits per capitulum). The seed set significantly varied with a corresponding increase in the number of bee visits (Fig. 7) (F=78.17, df=4, 90; P<0.001) and also with the bee species ((F=12.69, df=1, 90; P=0.0006). There was no significant difference between the number of bee visits and bee species (F=0.64, df=4,90; p=0.63).

Fig. 7