Oilseed rape pollen dispersal by insect pollinators in agricultural landscape R. Chifflet, B. Vaissière, A. Ricroch, E. Klein, C. Lavigne, J. Lecomte Good afternoon, my name is Rémy Chifflet and I am a PhD Student at the INRA Avignon research center and I will talk to you today about pollen dispersal by insect pollinators at the lanscape scale in the case of oilseed rape. This study was conducted in a program called GMBIOimpact sponsored by the French NationalAgency for Research. 0
Oilseed rape (OSR) Brassica napus Third oilseed crop in the world GM OSR area = 19% of OSR world crops Partially self-pollinated (< 50%) Pollen dispersed by both wind and insects still debated 1 Oilseed rape ranks third in the world for edible oil after soybean and palm. Genetically modified oilseed rape makes up 19% of all oilseed rape area in the world. Oilseed rape is a partially self-pollinating species and its pollen is dispersed both by wind and insects in proportion which are still debated, especially over long distances. 1
Genetically Modified (GM) crops Increase of GM crops area (x 75 between 1996 and 2008) Impacts on the environment Decline of biodiversity Changes plant & animal communities Alter the evolutionary potential of plants Co-existence with non-gm crops Isolation distance Clustering of fields Gene flow at landscape scale? 2 So we can ask about pollen-borne gene flow across landscapes. 2
Pollen vectors Wind Insects Bees Syrphid flies Butterflies Beetles & others 3 Rape pollen can be transported by two main vectors, mainly wind or insects. Among insects, the main pollinators are bees, honey bees and also wildbees, but syrphid flies, butterflies, beetles and some others insects could also play a significant role in pollen movement. Clearly different vectors may lead to differentspatial scales of pollen dispersal. 3
Spatial scale of pollen dispersal Most pollination events occur nearby a source field Sharp decrease at a short distance and slower decrease thereafter over 1000 m Wind component assumed from local to landscape scale but insect pollen dispersal considered mainly local Current models do not differentiate between wind and insect-borne pollen dispersal Long distance pollen flow by insects? 4 This is why we studied pollen flows over long distance by insects. 4
Foraging range of bees Correlated with size ; Varies among species and with resource spatial distribution ; Assessed with pollen load analysis, harmonic radar, sucrose feeders Pollen-borne gene dispersal? 5 And this is why we looked at the pollen-borne gene dispersal by insects 5
Study site Paris Selommes Montpellier Large production area of winter oilseed rape 6 Pollen flow was observed in an intensive agricultural landscape within a large production area of winter oilseed rape. The experimentation took place in north-central France during the flowering period in 2008. 6
Experimental layout 400 m 20 m 10 m 55 m 150 m 6 distances from closest OSR field ; 1100 m 0 1000 2000 3000 meters N 5 sampling dates in April and May 2008 7 The study area was a 10 by 10 km square where all feral populations and fields of oilseed rape were mapped. Six points were selected at different distances to the nearest field of oilseed rape and for each point, 4 male sterile plants were located to attract insects visiting oilseed rape flowers. Wecaught insects on five dates in April and May 2008. 7
Assessing pollination effectiveness Insects caught by net Anesthetised with CO 2 and transported at 6 C Hand pollination by scrubbing insects on MS plants in pollenproof greenhouse 8 Insects visiting male sterile flowers on sample points were caught by net, anesthetised with CO2 and transported at 6 C to maintain pollen viability during transport as foragers were transported from the study site to agreenhouse where hand pollination were made by scrubbing each insect on the stigmas of other male sterile plants. Seeds were harvested at maturity and insects were identified. 8
Results 1. Insect visitors 2. Fruiting & seed production 3. Modelling of fruiting 9 Then we assessed the pollination efficiency of each insect by the fruiting success and seed production that resulted from these hand pollinations. 9
1. Flower-visiting insects 71 insects caught : - 3 orders Hymenoptera : Bees, Sawflies, Wasps Diptera : Syrphid flies Coleoptera : Melligethes aneus - 26 species Only one honey bee (no colony in study area) 10 We caught 71 insects in 3 orders, first X hymenopterans with bees, sawflies and wasps. Second Y Dipterans with syrphid flies and third Z Coleopterans with rape beetles. 26 species were indentified and only one honey bee was caught and no colony was found in study area. 10
1. Flower-visiting insects Inter-tegular span to characterize insect size Taxon IT span (mm) Andrena sp. 1.58 Lasioglossum malachurum 1.69 Dolerus haematodes 2.31 Episyrphus balteatus 2.43 Andrena cineraria 2.78 Apis mellifera 3.67 Bombus lapidarius 6.14 Bombus terrestris 6.90 Xylocopa violacea 7.12 11 As we caught many species but few specimens per species, wemeasured the intertegular span to characterize the size of each insect. Some specimens had an inter-tegular span about 1 mm such as some small Andrena or Lasioglossium malachurum,medium-sized insects such as Dolerus haematodes, Episyrphus balteatus, Andrena cineraria or also honey bees,and the larger insects included bumble bees or carpenter bees such as Xylocopa violacea. 11
2. Fruiting and seed production 71 insects 455 hand pollinations 227 pods 4047 seeds (from 28 insects) Melligethes aneus caught at each point but gave no seed (n = 13) 12 We did a total of of 455 hand pollinations, which resulted in 227 pods that contained 4047 seeds, and these originated from 28 insects among 71 caught. It is interesting to note that all hand pollination made with rape beetles did not result in any fruit set regardless of the distance to the nearest oilseed rape field considered. 12
2. Fruiting and seed production 4047 seeds from 28 insects Andrena nitida at 55 m gave 1898 seeds (IT span = 3.6 mm) Xylocopa violacea at 1100 m gave 1 seed (IT span = 7.1 mm) Andrena sp. at 1100 m gave 3 seeds (IT span = 1.6 mm) 13 Among these 28 insects, it is interesting to note that only one Andrena nitida with a medium body size led to the production of 1898 seeds, and two insects trapped at 1100 m from the nearest oilseed rape field gave seeds: A capneter bee gave one seed and one Andrena gave 3 seeds. 13
2. Fruiting and seed production n is the number of insects caught at each distance point Varied significantly among the 6 distances (P < 0.001) Decreased from 59% at short distance ( 150 m) to 6% at greater distances Context Protocole Results - Conclusion 14 Here we can see that the distance to the nearest field had a significant effect on the fruit set of hand-pollinated flowers : fruiting success decreased from 59% over the first 150 m from the pollen source to 6% at greater distances. 14
3. Modelling fruiting success To refine hypothesis testing : Distance to the nearest fields; Order/Species : different foraging behaviour; Insects size (IT span) : Amount of pollen carry-over; Level of flowering (sampling date) : resource spatial distribution 15 Initially our working hypothesis for modelling pollen dispersal was that fruiting success varied with : the distance to the nearest field because empirical data showed such a decrease with the distance from the pollen source; the taxonomic position of the insect because different order or species don t have the same foraging behaviour and flight range; the size of the insect because some studies show a link betweensize and flight range and also larger insects could carry more pollen than smaller ones and thus pollinate more flowers more efficiently; finally the level of flowering because insects could have higher flight distance when there are few flowers available in the landscape. 15
3. Modelling fruiting success Fruiting success (0 or 1) Distance to the nearest fields; Order/Species : different foraging behaviour; Insects size (IT span) : Amount of pollen carry-over; Level of flowering (sampling date) : resource spatial distribution 16 With regard to modelling pollen dispersal, neither the order northe genus of the specimens affected significantly the fruiting success. 16
3. Modelling fruiting success Effect of insects size Greater fruiting success for larger insects Fruiting success > 0 for distance over 1100 m 17 We see a decrease in the fruiting success with the distance to the nearest oilseed rape field. And also the pollination with larger insects had a greater chance of producing a fruit than with smaller insects. Intrestingly, the predicted fruiting success wasis non null beyond 1100 m and seeds were harvested at this point. 17
3. Modelling fruiting success Effect of sampling dates Large variations among sampling dates Effect not related to level of flowering (temperature?) 18 Here we see that the fruiting success had large variations amongsampling dates, especially at short distances. Our first hypothesis was that these variations were due to different flowering levels, but we note that fruiting success was lower at the third date when flowering was greater thanat the first sampling date, which suggests that weather conditions may have an effect as well. 18
Conclusions Insects can play a significant role in long-distance effective pollen dispersal, and therefore in long-distance pollen-born gene dispersal; Greater dispersal for larger insects; Flowering affected effective pollen dispersal; 19 To conclude, Insects can play a significant role in long-distance effective pollen dispersal, and therefore in long-distance pollen-borne gene dispersal. Size is key factor affecting the probability of fruiting more than the taxonomic status of the insects, as large insects have a greater probability of fruiting even far from pollen sources Furthermore, environmental parameters such as flowering levels and weather conditions affect pollen dispersal. 19
Special thanks Sébastien Bauger, Eric Dufrênes, Robert Fonfria, David Genoud, Laurent Neu, Thierry Noblecourt, Alain Pauly, Jean-Pierre Sarthou, Elodie Poirrel To finish, I would like to thank people who helped me in the course of my experiments and for the identification of the insect spcimens. 20