(Bombus impatiens) hives under exclusion nets for
apple management pollination in orchards.
Résumé
Ayant démontré leur efficacité dans plusieurs contextes, les filets d'exclusion sont de plus en plus utilisés pour lutter contre les ravageurs des cultures. Cependant, la gestion de la pollinisation sous les filets doit être mieux étudiée afin qu’ils puissent être utilisés de façon optimale. L'efficacité de la pollinisation de ruchettes commerciales du bourdon Bombus impatiens (Hymenoptera: Apidae) a été évaluée pour des pommiers sous filets d’exclusion dans un verger expérimental du Québec, Canada, en 2016-2017. Seize parcelles de pommiers (longueur: 18,5 m, cultivar GingerGold) ont été soumises durant la floraison à l'un des quatre traitements suivants: 1) introduction d’une ruchette de bourdons placée à la fin d'un rang sous filet d’exclusion; 2) introduction d’une ruchette de bourdons placée au milieu d’un rang sous filet d’exclusion; 3) aucune ruchette sous filet d’exclusion et 4) des colonies d'abeilles domestiques à proximité, sans filet d’exclusion. Le poids et le calibre des fruits, le nombre de pépins par fruit et leur distribution dans les carpelles ont été mesurés afin d’évaluer la qualité de la récolte. Les fruits produits dans les parcelles sous filets d’exclusion pollinisées par les bourdons étaient de qualité équivalente à ceux produits conventionnellement après pollinisation par les abeilles domestiques et les abeilles sauvages provenant de l'environnement du verger. Le positionnement des ruchettes au milieu de la rangée a résulté en une meilleure homogénéité de la charge en fruits des arbres. Une discussion sur les bourdons en tant que pollinisateurs des vergers pommiers et sur les méthodes de distribution de pollen sont également mentionnées.
28
Abstract
Having demonstrated their effectiveness in many contexts, exclusion nets are increasingly used to control crop pests. However, pollination management under nets must be better studied if they are to be widely used. The pollination effectiveness of commercial bumblebee hives of Bombus impatiens (Hymenoptera: Apidae) was studied for apple production under exclusion nets in a research orchard located in Quebec, Canada during 2016-2017. Sixteen plots of apple trees (length: 18,5m, cultivar GingerGold) were subjected during bloom to one of the following four treatments: 1) introduction of a bumblebee hive placed at the end of a row under exclusion net; 2) introduction of a bumblebee hive placed in the middle of the row under exclusion net; 3) no hive under exclusion net 4) nearby bee hives without exclusion net. Resulting post-harvest fruit quality (e,g, fruit weight, size, number and distribution of seeds) was evaluated. Results suggest that under exclusion nets, bumblebees provide adequate pollination and fruit quality is equivalent to apple fruit conventionally pollinated by honeybees and wild bees in the orchard environment. We found that positioning of commercial bumblebee hives in the middle of the row provided better fruit load homogeneity in pollinated trees. Further discussion on bumblebees as apple pollinators and on pollen distribution methods are also provided.
Introduction
Apple is among the most important grown fruits in Eastern North America. Benefiting from land and climate proper to apple cultivation, the province of Quebec is the second largest apple producer in Canada. Still, apple growing can be arduous due to the diversity and abundance of pests and diseases (Vincent & Bostanian, 1998; Vincent & Roy, 1992). Without insecticide input, more than half of the north-eastern North American production would be damaged (Chouinard et al., 2006). In a typical commercial orchard of Quebec, over 14 pesticide applications are needed during the growing season to limit production losses (Morin & Chouinard, 2001). However, insecticides used are generally broad spectrum and will negatively impact a multitude of other organisms, either beneficial or damaging to the crop (Thistlewood, 1991; Bishop et al., 2000; Biddinger et al., 2013). Furthermore, intensive application and repeated use of compounds containing the same active ingredient may result in a faster natural selection of resistant individuals (Grigg-McGuffin et al., 2015).
Recently, exclusion devices have been used to prevent pest damage and reduce economic losses in many organic crops (Granatstein et al., 2016). The characteristics and effectiveness of exclusion systems have recently been adapted for fruit trees protection and appear promising in protecting orchards without the need for insecticides, as nets provide an impenetrable barrier that prevent the entry of harmful species. In the Quebec orchards, exclusion nets are generally installed in a row-by-row fashion and are closed above the base of the trunks (Chouinard et al., 2016; Chouinard et al., 2017). Therefore, the soil is excluded from the system, which prevents the development of most pests (e.g. codling moth (Cydia pomonella), apple maggot (Rhagoletis pomonella) and plum curculio (Conotrachelus nenuphar)) that complete their life cycle in the soil or at the base of the trunk (Agnello et al., 2006). Furthermore, nets are installed before bud break to protect trees against early pests (e.g. tarnished plant bug (Lygus lineolaris)). However, as exclusion nets provide a barrier against pests, they also prevent pollinators from reaching the flowers.
Apple trees have self-sterile flowers and require pollen from another cultivar to ensure fruit set (Ramírez & Davenport, 2013). Pollen transfer is carried out exclusively by pollinating insects since apple pollen is too heavy to be propagated by wind (Boyle-Makowski & Philogène, 1985). The ability of pollinators to allow cross-pollination is paramount for apple fruit set. Without cross-pollination, fertilization and seed initiation do not take place; the flower dries and falls without forming a fruit. In North America, honeybees (Apis mellifera) are the most popular commercial pollinator used by fruit tree producers, including apples. They are effective pollinators of apple flowers and have been used for many years as the main pollinator in commercial orchards (Gardner & Ascher, 2006).
30
Other wild bees, such as bumblebees (Bombus spp., Hymenoptera: Apidae) are however proving to be excellent pollinators of fruit tree orchards (Thomson & Goodell, 2001). Many studies have already stated that bumblebees are an effective complementary pollinator to honeybees when added in various crops, resulting in higher fruit set, fruit size, yield and seed number per fruit (Calzoni & Speranza, 1996; Dag et al., 2006; Desjardins & De Oliveira, 2006;Zisovich et al., 2012; Brittain et al., 2013; Garibaldi et al., 2013-2014; Mallinger & Gratton, 2015). In fact, bumblebees exhibit a daily activity in open field that exceeds that of honeybees (Garibaldi et al., 2013-2014). On favorable days, bumblebees tend to fly later during the evening and often resume their activity early in the morning (Prys-Jones & Corbet, 1991), while honeybees work during the central part of the day. It is also believed that they are even more effective than honeybees at pollinating some fruit trees such as apple and pear (Thomson & Goodell, 2001). Strong winds, cloudy weather, light intensity and fluctuations in temperature and humidity are all parameters that may affect pollination activity. Being larger than honeybees and covered with dense fur, bumblebees are less affected by adverse conditions and are able to carry heavier pollen or nectar loads (Goulson, 2010). Unlike honeybees, they will fly even if the weather is overcast, rainy or the wind is over 16 km/h (Boyle-Makowski & Philogène, 1985). Furthermore, bumblebees initiate activity at lower temperatures than honeybees, which require temperatures above 10°C (Burrill & Dietz, 1981; Boyle-Makowski & Philogène, 1985). In temperate regions, such as Quebec, temperatures can vary between 5 and 25°C during bloom and weather can be unpredictable at this time of year. Being able to pollinate under these conditions is an important advantage because apple blossoms appear early in the spring and last for a short period of time.
Exclusion nets appear to be a good alternative to insecticides for growers and may become the norm for organic orchard management in the future (Chouinard et al., 2017). Currently, nets have to be manually opened and closed for a few hours every day during bloom to ensure successful pollination and fruit set. Honeybees do not tolerate being restricted to an enclosed space such as the interior of nets and are thus not suitable pollinators of apple under nets (Vaissière et al., 2000). On the other hand, bumblebees are commonly used for pollination in sheltered crops such as greenhouses and high tunnels (see table IV in Velthuis & Doorn, 2006) and appear promising for pollination under exclusion nets. It is essential to adapt exclusion nets in order to allow the action of pollinators while retaining the maximum protective effect of nets. The exact capability of these insects under net conditions, particularly with fruit trees remains unknown and urgently needs to be studied to ensure the full efficiency of the exclusion method.
Here we report on the first study to focus exclusively on the pollination of apple trees by bumblebees under complete exclusion nets. The aim of the study was to test the effectiveness of commercial bumblebees
Bombus impatiens in net secluded hives to pollinate apple blossoms without requiring the opening of exclusion
under nets to that provided by pollinators (primarily honeybees and wild bees) naturally present in the environment of the orchard (open field). Based on previous studies (e.g., Boyle-Makowski & Philogène, 1985 ; Thomson & Goodell, 2001) that compared bumblebees and honeybees pollination we hypothesized that apple trees pollinated by commercial bumblebees from hives in plots under exclusion nets : 1) will offer a satisfactory fruit set (equivalent or greater than 5%); and 2) will provide equivalent fruit quality (i.e. weight, size and number of seeds). The second objective was to compare the efficacy of commercial bumblebee from hives under exclusion nets in two placements, either at the end or in the middle of the tree row. Tested hypotheses were as follow: 1) the location of hives under exclusion nets will not influence the quality of fruit (i.e. weight, size and number of seeds) produced by apple trees; and 2) fruit load per tree will decrease as the distance between commercial bumblebee hives and trees to be pollinated increase.
Materials and Methods
1. Experimental setupThis two-year study (2016-2017) was conducted in the experimental orchard of the Research and Development Institute for the Agri-Environment (IRDA), located in Saint-Bruno-de-Montarville, Québec, Canada (45°31’59.84°N, 73°20’59.55°W). Eight adjacent rows of dwarf apple trees, cv. GingerGold on B-9 rootstock, were selected from a three-year-old high-density plot. Each row was divided in two experimental units (length: 18,5 m, height: 3 m; Fig. 2) containing 14 trees each. Four treatments were compared: 1) an exclusion plot (nets) with a commercial bumblebee hive placed at the end of the row; 2) an exclusion plot (nets) with a commercial bumblebee hive places in the middle of the row; 3) a negative control plot (exclusion nets with no pollinators); and 4) an agronomical control plot (no nets) with honey bee hives nearby. Treatments were apply randomly and repeated four times. Pre-selected trees were sampled in all 16 experimental units.
2. Cultivar
The experiment was conducted using GingerGold cultivar. This is a very productive cultivar (40-50 t/ha) that produces relatively large-sized fruits (diameter 70-80 mm), uniform and whose visual qualities and taste are very appreciated from customers (Charest, 2006). It is an early cultivar, which is not prone to biennial bearing and is ready to pick around the first week of September. However, this cultivar requires strict phytosanitary control because of its highly susceptibility to scab (Venturia inaequalis (Cooke) Winter) and fire blight (Erwinia amylovora (Burrill) Winslow et al.).
3. Exclusion net system
In early April of 2016, vertical stakes from the planting's training system were changed and wooden structures were kept to support the netting system (see Appendix 1). Six black polyamide wires (Dura-Line®, Dubois Agrinovation, Saint-Rémi, Québec, Canada) were also installed on the wooden structure in order to
32
create a hexagonal tunnel to keep the net away from the branches (Chouinard et al. 2017) (see Appendix 1). The exclusion nets, made of clear high-density polyethylene nets (18,5 x 9 m; mesh size: 1,90 mm x 0,95 mm; 60 g/m2; 87% light transmission) (ProtekNet®, Dubois Agrinovation, Saint-Rémi, Québec, Canada) were installed before bloom (May 3rd in 2016 and April 24th in 2017). To close the bottom and extremities of each exclusion plot, sides were rolled together and sealed with clips (EasyKlip®, Dubois Agrinovation, Saint-Rémi, Québec, Canada). Particular attention was given to seal the nets as closely as possible around each tree trunk to prevent insect from climbing and reaching the interior of the net. The nets remained in the plot during the entire growing season and were removed after harvest (mid-September).
4. Pollination technique
Ten commercial bumblebee hives of Bombus impatiens were purchased from Koppert Canada Limited (Natupol©, Scarborough, Ontario, Canada, https://www.koppert.ca/). Each Natupol hive contained a full colony with a queen, plus 50 workers, and brood (pupae, eggs and larvae). Hives were received five to seven days prior to full bloom and were equipped with feeders that provided nectar solution. During that time, each colony was weighed to confirm their vigour and a black plastic pollen dispenser (PM 3D printing, https://pm3dprinting.com/) was added to the lid (see Appendix 2). Bumblebee hives were placed under nets three days before full bloom (May 18th in 2016 and May 15th in 2017) in eight experimental units; four at the end and four in the middle of the row. In addition, a white plastic board was installed on all hives as insulation and added protection against harsh weather. Hives were opened and bumblebees were allowed to acclimate to the net environment for half a day. Apple pollen from a compatible pollinizer (cv. Rome Beauty, Firman Pollen, Yakima, Washington, USA, http://www.firmanpollen.com/) was used for cross-pollination because only one cultivar was protected under the nets. During bloom, fresh pollen was added to the pollen dispenser each sampling day (three days in 2016 and two days in 2017). Quality testing (i.e. pollen germination tests) of stored pollen (-18°C) was performed in 2017 to ensure its viability (R. Reisbick, Firman Pollen, Pers. comm.).
5. Management
For the duration of the study, no pesticides or any other chemical management, with the exception of ground fertilizers, were applied. However, due to heavy spring rains in 2017, fungicide treatments (captan) against apple scab (V. inaequalis) needed to be carried. In 2016, experimental units were manually thinned according to standard commercial practices. However, in 2017, because of high scab incidence, no thinning was performed to limit fruit losses.
6. Sampling
Sampling was performed during favourable days, i.e. sunny, temperatures above 15 °C and winds below 11 km/h (May 21th to 23rd in 2016 and 17th-18th in 2017) (Martins et al., 2015). Because multiple
experimental units were observed at the same time, data collection was rigorously planned to avoid any bias that could arise due to lack of time or adverse climatic events.
a) Floral visitation rate
In 2017, ten individual bumblebee workers were visually tracked for 3 min in each plot that contained a commercial hive (for a total of 30 min of observation per plot). Observations were made from outside the plot since foraging workers could be clearly seen while causing the least interference (Pyke, 1978; Goulson, 2010). The number of flowers visited was recorded for each bee (Javorek et al., 2002). A visit occurred when the bee touched the stamens and/or the pistil of the flower. The presence (I) or absence (0) of a pollen load in the pollen basket to confirm a pollen transfer. Floral visitation rate (number of flower visited per minute) was calculated by dividing the number of floral visits by the duration of the observation period (Javorek et al., 2002).
b) Pollination efficacy
For this study, the pollination efficacy of bumblebees under nets was defined as the effect of pollination on 1) fruit load, 2) fruit quality and 3) seed yield.
Fruit load: In the spring, the total number of flower buds (at the pink stage) was evaluated on select trees (6-8 apple trees in 2016 and 5 in 2017) within each treatment (Amarante et al., 2008-2011; Sheffield, 2014). In mid- June, number of fruits was also evaluated on the same trees to assess total fruit load (Amarante et al., 2008- 2011; Sheffield, 2014).
Fruit quality: Harvest took place during the first week of September (September 6th in 2016 and September 1st in 2017) and quality tests were performed on 50 marketable fruits from each experimental unit except for the negative control plot where no fruit were produced due to absence of pollination. Fruits were randomly selected and were evaluated for any kind of damage. Those presenting developmental aberrations (smaller than 70 mm) as well as external or internal wounds caused by insects or more than 25% of the skin covered by scab were removed from sampling according to commercial standard for fresh market. Fruits were then measured to establish weight, size, firmness, sugar level and total number of seeds per fruit (Iglesias & Alegre, 2006; Amarante 2008-2011; Garratt et al., 2014a-b; Sheffield, 2014). Fresh weight (g) of the fruits was assessed using an electronic scale (Mettler Toledo, 0.1g). The maximum diameter (mm) of fruits at their equator (widest point) was evaluated using an electronic vernier (Mastercraft, 0.01mm). The firmness of the fruit flesh (lb) was evaluated on two sides using a penetrometer (Fruit pressure tester FT327, 1lb). The estimated amount of sugar (°Brix) contained in the fruits was evaluated using a refractometer (Hanna HI 96801, 0.2%).
34
Seed yield: Following seed count, seed distribution pattern for each fruit was assigned to one of eight possible categories of fertilized carpel (Fig. 3) (Sheffield, 2014) and only fruits with five carpels were kept for analyses. The number of seeds per fruit was also noted.
Within three days of harvest, all quality measurements had been taken to prevent bias due to post-harvest maturation.
c) Bumblebee hive location
Distance (in meters) between commercial bumblebee hives and selected trees were calculated for each plot.
7. Analysis
The R software version 3.4.1 GUI 1.70 El Capitan (R core team, 2017) was used for all statistical analyses. Contrast analysis was used to compare pollination efficacy between the two treatments of the following two groups 1) honeybees (no nets) and bumblebees (nets) and 2) bumblebee hives for two positions in the row (at the end or the middle). For both, we used a linear mixed-effects model (lmer() function) with Satterthwaite correction for fruit quality (i.e. fruit weigh and size and number of seeds) (Kuznetsova et al., 2016). We addressed the relation between fruit weight and the total number of seed for each treatment of bee pollination (i.e. nets vs. no net) using a linear model and we also compared each category of carpel using an analysis of variance (ANOVA) with a post-hoc Tukey’s HSD test (glht() function) (Hothorn et al., 2008). Finally, we compared total yield per tree in relation to its distance from of each bumblebee hive under exclusion nets (i.e. for hives located at the end of rows and at mid-row) using a generalized linear mixed-effects model (glmer() function) with a Poisson distribution (Bates et al., 2017). For each analysis, significance level was set at 0.05.
Results
1. Floral visitation rate
Results indicate that bumblebee workers visited on average 12 .03±6 flowers per minute (160 observations). Nets did not appear to interfere negatively with bumblebee foraging activity. Upon opening the hives, bumblebees tended to fly upward to probe the nets and try to find an exit. Approximately one hour after opening, most bumblebees were observed foraging on apple blossoms. Other field observations have shown that bumblebee workers under nets spent about 15±5 minutes of foraging before returning to their hive (40 observations).
2. Pollination efficacy
Fruit quality:
Bumblebees under nets provided equivalent (2016) or greater (2017) fruit quality than bees in the orchard environment (Fig. 4a-b). In 2016, sampled fruits weighted 248±11 grams and measured 82±1 mm on average when pollinated by bumblebees compared to 232±12 grams and 80±1 mm under conventional pollination. However, fruit weight (t(2.965) = -1.356, P = 0.269) and size (t(2.984) = -1.011, P = 0.387) did not
statistically differ in 2016 (Fig. 4a). In 2017, despite the incidence of apple scab, sampled fruits weighted 235±17 grams and measured 78±2 mm on average via bumblebee pollination compared to 188±8 grams and 73±1 mm (Fig. 4a) for conventional pollination. However, fruits produced via conventional pollination were statistically smaller (weight: t(2.971) = -3.879, P = 0.031 and size: t(2.967) = -3.33, P = 0.045). For both years,
bumblebee pollination produced a significantly lower number of seeds (t(3.09) = 4.08, P = 0.025 | t(30.768) = 5.637,
P = 3.55e-06) than experimental units pollinated by honeybees and wild bees in the orchard environment (Fig.
4b). Seed yield:
For both years, fruit weight increased slightly with the number of seeds per fruit (Fig. 5). In 2016, correlation between weight and number of seeds was stronger for fruits produced under nets (R2 = 11,6%, P <
0.001) than fruits produced following conventional pollination (R2 = 3.9%, P = 0.006). A similar pattern was
observed in 2017 for fruits produced under nets (R2 = 3.3%, P = 0.003) but not for conventional pollination (R2
= 0.008%, P = 0.909).
Overall, we did not observe any effect of the seed distribution patterns (Fig. 3) on fruit weight in 2016 (Bees: F(5,188) = 0.4492, P = 0.8135 ; Bumblebees: F(6,329) = 2.5449, P = 0.02013) and 2017 (Bees: F(6,166) =
2.0622, P = 0.06029 ; Bumblebees: F(6,267) = 1.1129, P = 0.355) (Fig. 6). Category A and B, which represent
well pollinated fruit, were more abundant in all treatments (F(6) = 2.1703, P = 0.04365). However, seeds
seemed to be more evenly distributed with conventional pollination while bumblebee pollination result in more irregularities (C, D, E, G categories) that could result in asymmetrical fruit.
3. Bumblebee hive location
In general, hives placed in the middle of the row resulted in a better fruit load per tree (Fig. 7a-b). In 2016, fruit load was more evenly distributed along the row when the hive was placed in the middle (R2 = 0.2%,
P = 0.131) while it decreased significantly when the hive was placed at either extremity (R2 = 11,3%, P <
0.001) (Fig. 7a). In 2017, fruit load was also marginally more uniform when the hive was in the middle of the row (R2 = 5.7%, P = 0.697) compared to when the hive was at the end (R2 = 10.7%, P = 0.205) (Fig. 7a).