Comparative Efficiency of Native Insect Pollinators in Reproductive Performance of Medicago sativa L. in Pakistan

Simple Summary Lucerne (Medicago sativa L.) is a cross-pollinated crop and pollination does not usually take place until the sexual column is released. The discharge of the sexual column is referred to as ‘tripping’ the flower. In this study, several bee species were investigated to facilitate the pollination of lucerne crop with a purpose to examine the pollination efficiencies of native insect pollinators that lead to a seed set of alfalfa. We designed a two-year field trial to investigate the abundance and diversity of insect pollinators along with foraging behavior in terms of stay time, visitation rate, pollen harvest and tripping efficiency. Moreover, the single-visit efficiency in terms of number of seeds per pod, germination and seed yield was evaluated. Ten major pollinators (five solitary bees, three honeybees and two flies) were tested for their pollination efficiency as effective pollinators. Findings show the solitary bees (Megachile cephalotes, Megachile hera, Amegilla sp.) can be recommended as an effective pollinator for this crop, thus increasing options available to seed growers. For lucerne seed development, more consideration should be paid to the variety of native wild bees, conservation strategies and foraging requirements in establishing a diverse pollination system rather than a single bee species. Abstract Lucerne (Medicago sativa L.) is a cross-pollinated crop and requires entomophilous pollination for tripping of flowers and subsequent pod and seed set. To discover the best pollinators for lucerne seed production, a two-year field trial was carried out at the research farm of MNS University of Agriculture, Multan, Pakistan. Abundance and diversity of insect pollinators along with the foraging behavior were recorded in terms of tripping efficiency, stay time, visitation rate and pollen harvest. The single-visit efficiency of abundant insect pollinators was also evaluated in terms of number of seeds and seed weight per raceme along with germination percentage. Ten most abundant floral visitors (five solitary bee species, three honeybee species and two syrphid fly species) were tested for their pollination efficiency. Honeybees were most abundant in both the years followed by the solitary bees and syrphid flies. Single-visit efficacy in terms of number of pods per raceme, number of seeds per raceme, 1000 seed weight and germination percentage revealed Megachile cephalotes as the most efficient insect pollinator followed by Megachile hera and Amegilla sp. Future studies should investigate the biology and ecology of these bee species with special emphasis on their nesting behavior and seasonality.


Introduction
Insect pollination is essential for the sustainability of both agricultural and natural ecosystems since 87% of flowering plants and nearly 35% of crops worldwide are dependent on insect and other pollinators [1,2]. Bees have been reported as the most efficient among

Study Site and Experiment Design
This study was conducted at the experimental farm of MNS University of Agriculture, Multan (longitude 71 • 26 35.5" E and latitude 30 • 08 50.5" N), Punjab, Pakistan. The climate of Multan is sub-tropical (arid and semi-arid) with hot summers and cold winters with low rainfall [46]. The maximum and minimum mean temperatures in summer are 42 • C and 29 • C, respectively, while these are 21 • C and 4.5 • C in winter, respectively. The average annual rainfall is around 186 mm, with the majority of occurrence during the monsoon season from July to September [47,48]. The soil of the area is classified as sandy loam that is mildly alkaline and well-drained [49]. The lucerne cultivar 'Sargodha 2000' was sown on an area of one acre during November in both the years (2019 and 2020). The row-to-row spacing was maintained at 30-45 cm. All the standard cultural practices were used in the experimental plot except the use of any pesticides [50].

Study Plant Species
Lucerne is widely grown in Pakistan as a fodder crop in different districts of Punjab and Sindh provinces. It is a perennial outcrossed open-pollinated legume pollinated by insects. Flowers are bisexual, with four petals and ten stamens arranged into racemes, and a plant can support several stems, each with multiple racemes and flowers [16,51].
Individual flowers remain open for a week if no pollination occurs and multi-seeded pods will mature six weeks after pollination. The peak bloom occurs between April and June in South Punjab, Pakistan [52].

Diurnal Abundance of Insect Pollinators
To measure the abundance of insect pollinators, three quadrats were randomly selected by throwing one square meter wooden frame during the blossom period of alfalfa. Counts were taken once a week at three-hour intervals throughout the day, starting from 8:00 am to 17:00 pm across the whole flowering season (April-June). During each time interval, a quadrat (one square meter) was observed for five-minute and all the insect visits were counted [53]. All the bee specimens were first morphotyped and then identified on wings during data recording. We also observed many other flower visitors, which did not come under systematic observation. Insect specimens were identified to the lowest taxonomic level by the experts (see Acknowledgments). All the voucher specimens were preserved and submitted to the Agricultural Museum of the Faculty of Agriculture and Environmental Sciences, MNS-University of Agriculture, Multan, Pakistan.

Foraging Behavior of Insect Pollinators
The tripping efficiency of abundant insect pollinators was observed in terms of the following parameters: the number of virgin flowers tripped per raceme, flowers visited but not tripped per raceme, and visiting already tripped flowers per raceme [15,24]. The number of flowers visited by a pollinator whose stigmas and anthers released from the keel (referred to as the tripping). The weekly data were recorded throughout the flowering season at 8:00, 11:00, 14:00 and 17:00 h.
Pollen harvest was carried out early in the morning. The buds were caged one day before their opening with a nylon mesh bag. These fully opened flowers were uncaged and visiting individuals were captured after a single visit. The pollen grains were counted in the laboratory under a stereo-zoom microscope by hemocytometric method [54].

Single-Visit Seed Set Efficiency
For measuring single-visit seed set efficiency, some unopened floral racemes were caged with nylon mesh bags (15 × 7 cm) before their opening. Upon opening of flowers, the mesh bags were removed and an individual of a specific pollinator species was allowed to land on the raceme and to visit the flowers. The racemes were re-caged after the insect visit and the un-tripped flowers were removed. Ten single visits were made for each pollinator species. Some flowers were also manually pollinated/tripped by using experimental hand pollination. The resultant reproductive success was recorded in terms of number of seeds per raceme, seed weight per raceme and germination percentage. Fifty open-pollinated (unrestricted insect visitation) flowers and 50 caged (no insect visitation) flowers were also maintained for the comparison. The resultant seeds were subjected to germination test by placing all seeds on moist filter paper in a Petri dish at room temperature, and germination was checked four days later [59].

Data Analysis
Data regarding pollination effectiveness of different species in terms of tripping, pollen harvest, foraging behavior (in terms of visitation rate and stay time) and single visit efficiency of abundant insect pollinator (in terms of number of pods per raceme, number of seeds set per raceme, 1000 seed weight) and germination (%) were initially tested for normality as the basic assumption of ANOVA is that error of the model should be distributed normally. By using QQ-plot and normality test, we did not find this assumption to be met for our data. So, we used nonparametric methods to test the mean equality. These methods are based upon ranking of the mean. Thus, we used Kruskal-Wallis test for analysis the data followed by Conover-Iman post-hoc test as corresponding Kruskal-Wallis null hypothesis were rejected. Data were analyzed by using computer software XLSTAT (XLSTAT. 2014 version: 5.03).

Pollinator Community/Floral Visitor Census
In this study, honeybees were most abundant followed by solitary bees and flies in both years. Apidae was the dominant family with five species, followed by two species of Megachilidae and a single species from the Halictidae. The pollinator community of lucerne was composed of 16 bee (Hymenoptera) species and six true fly (Diptera) species. A smaller proportion of these species was found during our systematic observations, i.e., five solitary bee species, three honeybees and two true fly species. Among the bees Megachile bicolor, Megachile lanata, Pseudapis sp., Xylocopa fenestrata, Xylocopa basalis, Halictus sp., Ceratina (Pithitis) smaragdula and Andrena sp. whereas, among flies Eristalinus laetus, E. megacephalus, Chrysomya rufifacies and Musca domestica were rarely seen.
The Amegilla sp., Eucera sp., Nomia (Hoplonomia) sp., Megachile hera and M. cephalotes were the most abundant visitors during both the studied years. Among solitary bees Eucera sp. and M. cephalotes were the most frequent floral visitors with totals of 215 and 206 individuals, respectively for both years. The average visitation frequency was also highest among all the observed bee species: (0.029, 0.042) and (0.022, 0.047) individuals per square meter per minute for the year 2019 and 2020, respectively (Table 1). Apis florea was most abundant (472) among honeybees while Eristalinus aeneus was the most frequent floral visitor (481) among the flies during both years studied. Diurnal dynamic patterns revealed the peak abundance of all the insect pollinator groups was attained at 11:00 am to 14:00 pm followed by a gradual decline until 17:00 pm during both years ( Figure 1).

Pollination Effectiveness
There was a difference among the three pollinator groups in terms of their tripping efficiency (chi-square = 32.69, DF = 39, p = 0.0000) and also visiting already tripped flowers (chi-square = 32.09, DF = 39, p = 0.0000). Tripping efficiency of solitary bees was higher

Pollination Effectiveness
There was a difference among the three pollinator groups in terms of their tripping efficiency (chi-square = 32.69, DF = 39, p = 0.0000) and also visiting already tripped flowers (chi-square = 32.09, DF = 39, p = 0.0000). Tripping efficiency of solitary bees was higher than the honeybees and flies. Flies and honeybees preferred to visit the already tripped or open flowers during their foraging.
Pollinator species varied with respect to single visit pollen harvest from the lucerne flowers (chi-square = 74.18, DF = 9, p = 0.0000). The pollen harvest was found to be highest for two Megachile species (M. cephalotes and M. hera) followed by three honeybee species (A. mellifera, A. dorsata and A. florea) while it was lowest for syrphid fly (E. aeneus) ( Table 2). The probabilities of tripping flowers change over time: more flowers were tripped by pollinators on an inflorescences at 11:00 am (33.63 ± 4.45) and by 8:00 am (33.87 ± 3.73) while no difference was observed between 14:00 pm and 17:00 pm. Moreover, pollinator visits to already tripped flowers were higher at all times of the day except at 17:00 am as shown in Table 3.

Single Visit Effectiveness of Pollinators
There was a difference among the pollinator species in terms of number of pods produced per raceme (chi-square = 74.22, DF = 23, p = 0.0001), number of seeds per raceme (chi-square = 78.62, DF = 23, p = 0.0001), 1000 seed weight (chi-square = 49.60, DF = 11, p = 0.0001) and germination (chi-square = 47.08, DF = 11, p = 0.0001) among the tested pollinator species in a single visit. Single visit efficacy proved M. cephalotes as the most efficient insect pollinator followed by M. hera and Amegilla sp. while the flies (E. arvorum and E. aeneus) were the least efficient insect pollinators. Open pollination treatment resulted in highest number of pods per raceme, number of seeds per raceme, 1000 seed weight and germination. Hand tripping also proved less effective than the single visits of three most efficient bee species. No pod development was observed in pollination exclusion treatments (Table 5).

Discussion
Pollination efficiencies of leguminous crops (alfalfa and clover) differed greatly among pollinator species. In our study, honeybees were the highest in proportional abundance (45%) in the first year, while honeybees and solitary bees (35%) had the similar proportional abundance in the second year. However, in both years, syrphid flies were the least abundant. Some previous studies have also reported higher abundance (85-98%) of solitary bees visiting lucerne flowers [24,26,30,45,60]. Moreover, from the same region, the higher abundance of honeybees and solitary bees has also been reported from other crops, i.e., canola [58], pumpkin [61]; luffa gourd [62] and radish [63]. In our study, during both years, bees from the family Megachilidae were more abundant (44-47%) as compared to the other bee families. Worldwide, higher abundance of Megachile bees has been linked with the enhanced alfalfa seed production due to their higher pollination efficiency [5,16,22,30,64].
Pollinator species varied in terms of tripping lucerne flowers and it was higher for solitary bees than the honeybees and flies. Among solitary bees, two Megachile sp. (M. cephalotes and M. hera) and Nomia (Hoplonomia) sp. were superior in tripping as compared to other most-abundant bee species. These tripping rates varied among pollinator species [24]. Some previous studies have also reported the better tripping efficiency of Nomia melanderi and other Megachile species (M. rotundata and M. abluta) on alfalfa flowers [15,45]. Social characters of Megachile sp. to collect pollen from the lucerne, its maximum foraging period coinciding with the lucerne blooming, also enable it to trip the lucerne flowers more efficiently [65,66]. Since most of the female solitary bees are pollen foragers, they exploit flower by ensuring tripping of Medicago flowers [31]. Moreover, tripping efficiency of Nomia sp. is due to its buzz pollination character, production of several generations in a year, high abundance and foraging time synchrony with lucerne blooming [25]. Tripping varied with time of day; our results are in line with the previous studies illustrating that visits of bees peaked from 10:00 am to 14:00 pm, and this period coincided with the complete release of the pollen [32]. Moreover, all the pollinators were at their peak abundance at 11:00 am to 14:00 pm during both years. This is attributed to the pollinators' activity, which is predicted by the time of anthesis, the time when a large amount of nectar and pollen resources are available [33].
In our study, both managed and wild honeybees were least effective (4-10%) in terms of tripping the flowers. Some previous studies have also reported the inefficiency of managed honeybees in tripping the alfalfa flowers due to different reasons. First, honeybee workers quickly learn to access the nectar without tripping the flower due to their side feeding habit that minimizes the chances of tripping lucerne flowers [31,67]. Second, lucerne pollen is lacking the essential amino acid (isoleucine), which means honeybees suffer protein stress when only pollinating lucerne [30]. Third, some adaptive behavioral characters also lead honeybees towards poor tripping, resulting in 1% tripping in open field conditions [29,68,69] and only 22% under caged experiments [15]. Moreover, tripping inefficiency (5% of the visited flowers) of wild honeybees (A. dorsata, A. florea) has also been reported [24,70].
Similarly, syrphid flies (E. aeneus and E. arvorum) also proved ineffective in terms of tripping lucerne flowers. Previously, tripping inefficiency of syrphid fly (Eristalinus tenax) and other flies (Calliphora vicina and Lucilia sericata) has also been reported from Medicago pollination [71]. Syrphid flies feed mainly on nectar and often return to feed on previously visited flowers, despite being caged and restricted in space or resources [72][73][74].
In the current study, solitary bees harvested more pollen on a single visit than those of honeybees and flies. The two Megachile species (M. cephalotes and M. hera) harvested the highest number of pollen grains followed by A. mellifera and A. dorsata. This pollen harvest indicates a passive pollen load by the bees on a single visit. The pollen-foraging individuals can be more effective pollinators than the mere nectar foragers [75,76] and pollen foraging leads to greater tripping of alfalfa flowers [77,78]. In our study, Nomia sp. and Eucera sp. were less efficient in terms of harvesting pollen grains although their tripping efficiency was equivalent to the other high pollen harvesting species. Contrarily, a previous study reported high pollen harvest (2100) for N. melanderi after their 15 visits in M. sativa [15].
Solitary bees proved best in terms of single visit pod set per raceme and seed setting as compared to honeybees and flies. Among the solitary bees, M. cephalotes resulted in maximum pods and seed set followed by M. hera and Amegilla sp. Probabilities of pod and seed setting are higher for the insect pollinators with higher tripping rates [15]. Pod setting done by Nomia (Hoplonomia) sp. and M. hera was 66 and 62%, respectively. Previously, a single study reported effectiveness of M. rotundata and N. melanderi) as most efficient in terms of single-visit pod set efficiency in the range of 40-57% [15]. Megachile rotundata and N. melanderi are the intensively managed bees on commercial level for lucerne crop [20]. Moreover, from South Punjab region, the effectiveness of other solitary bees in terms of single-visit seed/fruit set have been reported for other crops, i.e., canola [58], bitter gourd [56], pumpkin [61], luffa gourd [62] and falsa [79].
The effect of solitary pollinating insects was remarkable and confirms the preponderant role of solitary bee pollinators on seed set of Medicago flowers whereas honeybees were confirmed to be poor pollinators of this crop, which collect pollen from Medicago under certain circumstances. Pollen-gathering honeybees are usually better pollinators than nectar-gathering honeybees, which often forage by 'side-working' the flowers. During those visits, the bee does not make contact with the stigmas. Solitary bees, on the other hand, usually work the flowers from the tip, irrespective of foraging activity, almost always contacting the anthers and stigmas [80][81][82][83]. Moreover, flies (E. aeneus and E. arvorum) were also ineffective in their single-visit seed set in both the years. Seed setting of flies E. tenax (Syrphidae), C. vicina and L. sericata (Calliphoridae) were lower because of insufficient pollen transfer and pollen adhering to their bodies [71].
In open pollination, seed set was 85% higher as compared to the self-pollination (0%). Self-pollination is not fruitful in the case of lucerne seed formation due to the abortion of seeds [15]. Previously, open pollination along with solitary bees produced significantly more seeds as compared to honeybees and pollination-exclusion flowers [18,26,84]. Moreover, open pollination also produced 87% higher seed set than the hand-tripped flowers.

Conclusions
In conclusion, the solitary bees (M. cephalotes, M. hera, Amegilla sp.) proved to be the most efficient pollinators for lucerne seed production. Conservation of these most-efficient bees (through provisioning of foraging and nesting resources) can lead to sustainable alfalfa seed production. Future studies should focus on exploring the population threshold of these bee species for better pollination and seed production in lucerne.