Impact of abiotic factors on the foraging behavior of two honeybee species on canola in Bahawalpur, Punjab-Pakistan

This study was conducted to understand the effect of abiotic factors on the foraging behavior of two native honeybee species ( Apis dorsata and A. florea ) andits ultimate impact on reproductive success of Brassica napus L. Foraging behavior of honeybees was studied in terms of visitation rate, visitation frequency, stay time on flower, stigma contact events, nectar robbing and pollen grain deposition on stigma. Simple linear regression model was used to find relationship between foraging behavior parameters and abiotic factors including ambient temperature, light intensity, wind speed and relative humidity. The results showed that A. dorsata had statistically higher visitation rate (54.23±1.307 flowers/300 seconds), number of stigma contact events (0.98±0.008) and pollen deposition (287.72±6.58 pollens) than A. florea i.e. 22.65±0.808 flower/300 seconds, 0.51±0.032 events and 154.83±7.83 pollens, respectively. The linear regression analysis showed that the stay time and nectar robbing of A. dorsata was not affected by any of the abiotic factors. Visitation frequency had negative relationship with all the abiotic factors. Stigma contact time decreased with the increase in temperature and increased with the increase in light intensity. In case of A. florea , stigma contact time had no relationship with any of the abiotic factor. Visitation rate and stay time were affected by light intensity. However, the value of R 2 did not exceed 0.25 in all these significant relationships. Nectar robbing events increased with the increase in relative humidity and decreased with the increase in wind speed. Apis dorsata was more efficient pollinator of B. napus than A. florea in terms of seed weight per pod and germination percentage.


Introduction
Insect species belonging to the order Hymenoptera are affected by a number of abiotic factors such as rainfall, humidity, temperature, wind velocity and light radiation (Southwick and Moritz, 1987).For example, in case of bees and wasps, the most important abiotic factor is ambient air temperature (Heinrich, 1984).Insect pollinators respond very rapidly to the external temperature as they are endothermic in nature.Variation in temperature and rainfall strongly predicts their relationship with host plants (Hilário et al., 2000).Besides this, temperature also affects the survival, fecundity, development, population size and migration of several insect species ( Régnière et al., 2012).The flying activity of most of the insect species is directly proportional to temperature up to a certain optimum limit and beyond that, it is inversely proportional (Taylor, 1963).Both the density dependent and independent factors determine different life history traits of insects like geographical distribution, relative abundance, growth rate, overwintering, interspecific competition, number of generations (Savopoulou-Soultani et al., 2012) and population dynamics (Shivanna et al., 2011).The foraging and defensive behavior of honeybees is strongly affected by certain density independent factors i.e., temperature, wind speed, light, rainfall and humidity.Recent studies have provided the optimum ranges of abiotic factors for honeybee activities along with corresponding impacts of high or low levels of these factors (Southwick and Moritz, 1987).Pollination is an important ecosystem service delivered by insects responsible for production of twothirds of the crop plants (Ricketts, 2004).From total consumption of the world food, 90% is being pollinated from the activities of honey bees and almost from 100 valuable crops of the world, 70% are being pollinated by honeybees (Moritz et al., 2005;Klein et al., 2007).In Pakistan, Brassica napus (canola) ranks third after sunflower and cotton contributing 17% of the total domestic edible oil; grown on an area of nearly 0.402 million acres (GOP, 2009).Its flowers are entomophilous in nature capable of both self and cross-pollination.The out-crossing rate ranges from 12 to 47% depending on cultivar (Williams et al., 1986;Becker et al., 1992).The cross-pollination by insects can increases the seed weight of B. napus by 18% and its market value by 20% (Bommarco et al., 2012).Honeybees have been recognized as the most common visitor of B. napus flowers.The shallow settlement of its visible nectar and yellow color mostly attracts bees, flies and butterflies (Kunin, 1993).Insect pollinators, especially the bees, ensure the successful reproduction of plants through crosspollination (Lane and David, 2006).However, foraging activities (i.e.foraging behavior) of bees are strongly influenced by unpredictable environmental variables (Tripath, 2011;Tirado et al., 2013) and timing and location of food (Biesmeijer and Ermers, 1999).All the daily foraging activities and foraging patterns of honeybees sharply change with weather conditions.Bees therefore can accordingly alter their floral resource preferences depending on temperature, rainfall, wind, relative humidity and sunshine.However, each social bee species has its own set of microclimate in which its foraging activity can be sustained (Contreras et al., 2013).Conserving and utilization of alternate native pollinators, especially the bees therefore appear to be a right choice for successful B. napus production (Sajjad et al., 2008).To understand whether a particular honeybee species could be an effective pollinator in a given climate, it is imperative to understand its foraging behavior in response to the local climatic factors.The current study was carried out to evaluate the impact of density independent factors on foraging behavior of two native honey bee species, A. dorsata and A. florea and its ultimate impact on reproductive success of B. napus.

Research and study area
Canola, Brassica napus L. was sown during first week of November 2018 at the research farm of Regional Agriculture Research Institute (RARI) Bahawalpur, Punjab, Pakistan.Blooming and our systematic observations were from 1 st February to 15 th April.The climate of this region is sub-tropical with mean daily minimum and maximum temperature range of 15°C to 25°C and 30°C to 35°C, respectively.The average annual rainfall is about 300 to 500 mm (Khan et al., 2010).The area is about 181 meter above sea level.We selected an area of 2 acres for experimentation during the peak flowering period of B. napus.

Foraging behavior of honeybees
Different parameters of foraging behavior of A. dorsata and A. florea were recorded i.e., visitation frequency, visitation rate, stay time per flower, stigma contact time and side feeding (nectar robbing).Visitation rate is defined as number of flowers visited in a unit time (i.e., five minutes in this study).Visitation frequency is defined as number of bees visited on a specific plant in a unit time (i.e., five minutes in this study).Stay time is the duration of stay on a flower in a single visit.Sometime pollinators tend to feed on nectar without coming in contact with stigma is called nectar robbing e.g.side feeding by honeybees.

Visitation rate
To record visitation rate of a bee species at five different times of a day (i.e., 10:00 hrs, 12:00 hrs, 14:00 hrs, 16:00 hrs and 17:00 hrs), we randomly selected ten plants and counted the number of flowers visited by A. dorsata and A. florea in five minutes.A stop watch was used for recording five minutes of observations.In this way there was a total of 50 minutes of observation in a day and we repeated taking such observations four times in a season i.e. usually with the interval of seven days.Before recording visitation rates, the temperature and relative humidity was recorded with the help of a field thermohygrometer, light intensity with the help of a lux meter and wind speed with the help of an anemometer.

Visitation frequency
To record visitation frequency of honeybees at five different time intervals of a day (i.e., 10:00 hrs, 12:00 hrs, 14:00 hrs, 16:00 hrs and 17:00 hrs) 10 plants were randomly selected and each plant was observed for five minutes to record the number of individuals of each bee species, visiting on the flowers.Stop watch was used for this purpose.A total of 250 minutes of observation in a day was recorded four times in the season usually with the interval of seven days.

Stay time
The stay time of both the honeybee species was recorded on 10:00 hrs, 12:00 hrs, 14:00 hrs, 1600 hrs and 1700 hrs in a day.At each interval, twenty individuals of each bee species were observed for the time spent on a single flower.A stopwatch was used for this purpose.This procedure was repeated four times in a season usually with the interval of seven days.Weather conditions were always recorded before systematic observations.

Nectar robbing and stigma contact duration
We determined nectar robbing events (side feeding without coming in contact with stigma) and stigma contact duration in both the honeybee species at different time intervals of a day i.e., 10:00 hrs, 12:00, 14:00 hrs, 1600 hrs and 1700 hrs.A stopwatch was used to measure the length of time an individual bee remained in contact with the stigma during a single visit.

Role of honeybees in pollination
The single visit efficiency of A. dorsata and A. florea was recorded in two different ways i.e. in terms of pollen deposition on stigma and in terms of plant reproductive success.In case of single visit efficiency, we caged 40 floral buds of B. napus (20 for each bee species) with nylon mesh bags 24 hours before their opening.The flowers were unveiled during the peak activity timing (10:00 hrs to 14:00 hrs) of pollinators.Once a flower had been visited by the honeybee species, the stigma was excised and dipped into a drop of ethyl alcohol placed on glass slide.Subsequently, a cover slip was placed on it and number of pollen grains were counted under 40X stereo-microscope (Akram et al., 2019).Similarly, in case of plant reproductive success, we caged 40 floral buds of B. napus (20 for each bee species) with nylon mesh bags 24 hours before their opening.The flowers were unveiled during the peak activity timing (10:00 hrs to 14:00 hrs) of pollinators.Once the flowers had been visited by the honeybee species, the flowers were re-caged and tagged.The nylon mesh bags were removed once pods had been formed.Upon the maturity, the pods were removed and measured for pod length, pod weight, number of seeds per pod, seed weight per pod and germination percentage.We also maintained 20 floral buds which were caged throughout flower opening period and regarded it as self-pollinated while maintained 20 other floral buds which were kept open (without cage) and regarded as cross-pollinated.

Data analysis
T-test was used to compare the means of two independent samples of A. dorsata and A. florea in terms of their visitation rate, visitation frequency, stay time, stigma contact time, stigma contact events, stigma contact duration, nectar robbing events and pollen deposition.Simple linear regression analysis was applied to see relationship between foraging behaviors (visitation rate, stay time, visitation frequency, stigma contact duration and nectar robbing) and abiotic factors (ambient temperature, relative humidity, light intensity and wind speed).In order to compare reproductive performance of A. dorsata and A. florea along with self-pollinated and crosspollinated flowers, one way ANOVA was applied on reproductive success parameters.The means were compared by Tukey's post hoc test at alpha 0.05.Pearson's correlation coefficient was also used to see the relationship between different reproductive success parameters.

Results
Results of T-test showed that A. dorsata and A. florea were statistically significantly different in terms of visitation rate, visitation frequency, stay time, stigma contact duration, stigma contact events, nectar robbing and pollen deposition.Visitation rate, stigma contact events and pollen deposition were highest in A. dorsata whereas visitation frequency, stay time, stigma contact duration and nectar robbing were highest in A. florea (Table 1).The linear regression analysis showed that the visitation rate of both A. dorsata and A. florea were significantly affected by wind speed, temperature and light intensity.Wind speed was negatively related with visitation rate in case of A. dorsata while it was positively related in case of A. florea.Temperature and light intensity were positively related with visitation rate of both A. dorsata and A. florea.Relative humidity, however, had no relationship with visitation rate of both the species (Table 2 & 3).Abiotic factors depicted the visitation frequency of A. dorsata better than A. florea.In case of A. florea wind speed, relative humidity and temperature had no relationship with visitation frequency.On the other hand, they were negatively related with visitation frequency of A. dorsata.Contrary to A. dorsata, light intensity had a positive relationship with visitation frequency of A. florea.All the four abiotic factors failed to predict stay time of A. dorsata while temperature and light intensity had a strong negative relationship with stay time of A. florea.Stigma contact time of A. florea largely remained unaffected by abiotic factors as there was a significant relationship between abiotic factors with stay time.In case of A. dorsata, temperature had a strong negative while light intensity had a strong positive relationship with stay time.Apis dorsata rarely robbed the nectar in this study.However, weather factors did not affect the nectar robbing in A. dorsata.On the other hand, wind speed significantly reduced the nectar robbing while relative humidity significantly increased it in A. florea (Table 2 & 3).However, the value of R 2 did not exceed 0.25 in all these significant relationships.There were statistically significant differences among pod weight, pod length, number of seeds per pod, weight of seeds per pod and germination percentage.The maximum pod weight was recorded in A. dorsata (0.51g) followed by A. florea (0.50g), open (0.12g) and caged (0.09g).The maximum pod length was recorded in A. dorsata (7.16 cm) followed by A. florea (7.03 cm), open (6.64 cm) and caged (4.13 cm).The highest number of seeds per pod was recorded in A. dorsata (28.55) followed by A. florea (27.60), open (16.89) and caged (8.50).The maximum seed weight per pod was recorded in A. dorsata (0.12g) followed by A. florea (0.10g), open (0.06g) and caged (0.05g).Apis dorsata pollinated seeds were highest in germination percentage whereas minimum germination percentage was recorded in A. florea pollinated seeds (Table 4).The Pearson's correlation matrix among different reproductive success parameters of canola crop as a result of different pollination treatment i.e. A. dorsata,A. florea,open and caged is shown in Table (5).Except germination, all the reproduction parameters had a strong significant positive relationship among each other.

Discussion
Apis dorsata in present study showed significantly higher visitation rate, stigma contact events and pollen deposition in a single visit than A. florea.On the other hand, A. dorsata was inferior to A. florea in terms of visitation frequency, stay time and stigma contact duration.The information on quantitative components of the pollinator visitation (i.e.visitation rate and frequency) together with the knowledge of pollination effectiveness could allow for identification of the potential pollinators of particular crop species (Albano et al., 2009).Some previous studies confirm the significantly higher visitation rate of A. dorsata as compared to A. florea in B. napus (Ali et al., 2011), onion (Sajjad et al., 2008) and pumpkin (Ali et al., 2016)  for among-plant differences in visitation rates.On the other hand, Albano et al. (2009) concluded that visitation frequency of a particular pollinator species varies considerably across regions, fields and years and therefore the results obtained in such studies should not be directly validated to other cases.Bee species which are intensive foragers (having high foraging rates) usually work on rapid pace but stay for a shorter period of time on flowers than species with less foraging rates (Sajjad et al., 2008;Ali et al., 2011;Ali et al., 2014;Ali et al., 2016;Zameer et al., 2017).This was also obvious in our study as A. dorsata had much higher visitation rate than A. florea and showed much lower stay time on flowers than A. florea.Apis florea showed significantly higher nectar robbing events than A. dorsata.Contrary to this, Villalobos and Shelly (1996) found that nectar robbing A. florea had lower visitation rate but higher stay time than A. dorsata.In the flowers of canola, the availability of pollen and nectar decreases mostly at 12:00 pm (Meyerhoff, 1954) this may explain why the A. florea robbed nectar.
In the present study, pollen deposition on stigma during a single visit was significantly higher in A. dorsata than A. florea.The visitation frequency and visitation rate are the most important factors when comparing the pollination efficiency.In general, the greater the visitation frequency and visitation rate, the higher chances of pollination (Singh et al., 2006).Pollen load and pollen harvesting also matters in pollination effectiveness (Canto-Aguilar and Parra-Tabla, 2000).Several studies have documented bees as vital pollinators based on their ability to deposit more pollen on stigma than flies and butterflies (Ali et al., 2011;Willmer et al., 2017;Zameer et al., 2017).However, some recent accounts have argued that pollen deposition is not always a reliable measure of pollinators' effectiveness.Wang et al. ( 2017) noticed that pollen deposition was generally inconsistent with reproductive success parameters of plants i.e., seed production.They did not find any correlations between pollen deposition and seed production as a result of single visits made by bees.
In our study, the visitation frequency of A. dorsata was negatively associated with wind speed, relative humidity, temperature and light intensity whereas in case of A. florea it was positively related with light intensity and only positively related with temperature and relative humidity.Similar findings were reported from India by Abrol (2010) who noticed that A. florea had significantly positive correlation with air temperature, light intensity, solar radiation and negatively with relative humidity.Apis dorsata proved to be a better pollinator than A. florea in terms of seed weight per pod and germination percentage.Number of seeds per pod, pod length and pod weight were statistically similar for both the bees.Similar findings were obtained by Ali et al. (2011) and they also found similar correlation among attributing components as we obtained in this study.In the present study, the reproductive performance of canola -number of seeds per pod, seed weight per pod and pod weight-was better in single visits than multiple visits (open-pollinated flowers).This is might be due to pollination limitations associated with quantity (number) and quality of pollen grains (Aizen and Harder, 2007).Usually the seed set increases with the number of pollen grains deposited on the stigma up to certain limit and attains an asymptote after certain number of pollen grains (Sorensen and Webber, 1997).
Therefore number of pollen grains deposited on stigma is not merely the indicator of good seed set.Moreover, viability of pollen grains and number of heterospecific pollen grains also affect seed set in angiosperms (Bellusci et al., 2010).The behavior of pollen and/or nectar feeding is also important in this regard.In Brassica compestris, single visits of pollen and pollen/nectar feeding pollinator were enough in achieving the maximum seed set.On the other hand, all the single visits of nectar feeding bees resulted in no seed set (Sihag, 2018).

Conclusion
Apis dorsata had higher visitation rate, number of stigma contact events, pollen deposition, number of seeds per pod and germination percentage as compared to A. florea.Abiotic factors i.e. ambient temperature, relative humidity, wind velocity and light intensity strongly affect the foraging behavior of both A. dorsata and A. florea.

*
means sharing similar lettering are statistically nonsignificant at alpha 0.05 A number of biotic and abiotic factors could affect the foraging behavior of a particular pollinator species.Sánchez-Lafuente et al. (2005) and Grindeland et al. (2005) reported floral display as mainly responsible

Table - 4. Comparison of pollination effectiveness of honeybees in terms of reproductive performance of Brassica napus Treatments No. of seed per pod Seed weight/pod (grams) Pod length (cm) Pod weight (grams) Germina tion (%age)
Saeed et al. (2012)Zameer et al. (2017)found significantly lower visitation rate of A. dorsata than A. florea in radish whereasSaeed et al. (2012)did not find any significant difference between visitation rates of two honeybees in bitter gourd.On the other hand, we found higher visitation frequency of A. florea as compared to A.
dorsata in the present study.Contrary to our finding, Ali et al. (2011) observed higher visitation frequency of A. dorsata than A. florea in canola whereas Saeed et al. (2012) observed higher visitation frequency of A. florea than A. dorsata in bitter gourd crop.

Table - 5. Relationship between reproductive success parameters of canola using Pearson's correlation at alpha 0.05 No of seeds Seed weight Pod length Pod weight
In case of A. dorsata, stay time in this study was not affected by any of the weather factor while it was negatively related with temperature and light intensity.
However, Stabentheiner and Kovac (2014)reported that duration of stay of honeybees, decreases with ambient air temperature and increases with body temperature.Moreover, foraging in the sun-light enables the foragers to reduce the energetic costs per stay by about 19% to 76%.Nectar robbing in A. florea increased with relative humidity and decreased with wind speed.Previously no such relationship has been established between potential nectar robbers and environmental factors.