Abstract

Keywords

Introduction

Isolation distance seems to be a way to reduce pollen transfer due to both main pollen vectors in oilseed rape : wind and insect pollinators. Several authors have investigated this question by using various designs. In a continuous design, with recipients plants evenly placed around a small transgenic source plot , it has been demonstrated that the pollen distribution was leptokurtic and that pollen dispersal drastically decreased between 3 m and 12 m (Lavigne et al, 1998). These results are in compliance with other experiments (Levin and Kerster, 1974; Scheffler et al, 1993). In discontinuous designs, where recipient plants are situated at various isolation distances from the source, the results are less clear. Stringam and Downey (1982) found that the contamination decreased with distance (respectively, 47, 137 and 366 m) while Manasse (1992) found that this type of design favoured long-distance dispersal. Moreover, narrow barren zones 4-8m in width seemed to increase cross-pollination (Morris et al, 1994). In the latter cases, the seed contamination was attributed to insects which could fly over the non-recipient plants or over areas without any vegetation. On the other hand, studies on the foraging behaviour have often shown that honeybees individually forage on small areas (Singh, 1950; Free and Spencer -Booth, 1964; Gary et al, 1977) and that honey bees and bumble bees generally visited the nearest flowers in the same row and were strongly directional in their movements (Cresswell et al, 1995). Because, in our region (Brittany, France), honey bees are the predominant pollinators, the purpose of this study was to determine the efficiency of short-distance isolation (3-12 m width), using a barren zone, on the movement of these insects.

Material and method

The trial consisted in two identical strips (6 m x 30 m) of the spring oilseed rape ‘Fidelio’. The design had the shape of a V truncated at the base in such a way that the zone, without any vegetation, between both strips varied from 3 m width at the base of the V and 12 m width at the top. Three equal areas (60 m²=6 m x 10 m) were considered in each strip, named sector I (corresponding to the part of the design with a barren zone from 3 m to 6 m), sector II (corresponding to the medium part of the design with a barren zone from 6 m to 9 m) and sector III (corresponding to the part of the design with the largest barren zone, i.e. from 9 m to 12 m). A hive was placed at 8 m from the top of the V. Therefore, the sector III was the closest to the hive.

With the aim to compare the attractiveness of the two strips along the flowering period and the distribution of the honey bees in the sectors, the number of available flowers/m² and the number of honey bees were evaluated at 8 dates and at several times in the afternoon on both strips. Honey bees were counted by one experimenter walking regularly along the strip during 10mn.

To study the movement of the insects within and between the strips, foraging honey bees were marked with three different colours according to each sector in one of the two strips which was considered as the « Donor » strip. The other strip was named the « Recipient ». The marking was made at the beginning of the afternoon when the activity was high. Nearly all the honey bees foraging in the strip were marked. In addition, they were marked until their number was balanced between the three « Donor » sectors. The experiment was replicated on 3 days, successively using two colonies. The total number of marked honey bees could vary from 53 to 90 according to the date of experiment. The movement of honey bees within the « Donor » strip was estimated by the percentage distribution, in each sector, of the proportion of the recovered honeybees marked with one of the three colours. These observations were made 1 hour and 2 hours after the marking time. The observation of the movements between both strips was performed straight after marking : marked and unmarked honey bees were counted in the « Recipient » strip and other counts were regularly made further.

Another objective was to evaluate the renewal of the population. So, controls were also made in the « Donor » strip to determine the remaining marked honey bees after the marking phase. Counts of marked or unmarked honey bees were analysed in two ways :

• the recovery rate (number of marked honey bees recovered after the time t_x/number of honey bees initially marked at t_0,)
• the proportion of marked honey bees/unmarked honey bees.

The marking phase lasted about 15mn (with 1 experimenter in each sector) and the counting length was about 10mn (1 experimenter studying both strips). The first recovery rate in the « Recipient » strip was determined as soon as possible after marking i.e. after 15mn ( ± 15mn), whereas the recovery rate in the « Donor » strip was controlled 1 hour ( ± 15mn) after the marking time.

Statistical analysis were performed by a one-way analysis of variance followed by a Newman-Keuls test at probability 5%. In tables results followed by the same letter are not significantly different.

Results

Attractiveness of the strips

The development of the flowering was the same in both strips. The density of available flowers/m² increased from 250 flowers/m² (18 June) up to 900 (25 June) and was less than 600 at the end of the observation period (29 June). At a given time, the total number of foraging honey bees was equal on the two strips (mean «Donor »=71 versus mean « Recipient »=74). Their abundance was not different between sectors (Table 1).

Table 1 : Mean total number of foraging honey bees in each Sector ± SE (from 18 to 29 June)

	Sector I	Sector II	Sector III	Significance
Donor	25.5 ± 4.2	22.7 ± 3.4	22.4 ± 3.2	NS p=0.80
Recipient	24.5 ± 3.3	25.7 ± 3.1	23.7 ± 3.4	NS p=0.90

Experiments with marked honey bees

The experiment with marked honey bees occurred as the flowering was increasing (18, 22 and 23 June).

Honey bees movements within the Donor strip. The comparison of the percentage of honey bees marked with one colour (Table 2) showed that most of them stayed in their originating sector. Nevertheless, honey bees marked in sector I (colour I) were also found in sector II in a rather large proportion (37%). One can notice that the proportions of honey bees marked in sector II (in the middle of the strip) and recovered in the other sectors were not significantly different.

Table 2 : Distribution (in percent ± SE) of the marked honey bees in each sector

	Sector I	Sector II	Sector III	Significance
Colour I	54.1 ± 10.3 (a)	37.0 ± 11.2 (a)	8.3 ± 5.3 (b)	S. p = 0.0103
Colour II	28.2 ± 8.4 (b)	56.9 ± 10.4 (a)	14.8 ± 2.8 (b)	S. p = 0.0078
Colour III	5.6 ± 5.5 (b)	19.4 ± 16.4 (b)	75.0 ± 15.9 (a)	S. p = 0.0061

Honey bees movements between the Donor strip and the Recipient strip. Because it was done about 1h after marking, the Donor recovery rate was underestimated compared with the Recipient because the probability of flying back to the hive became higher as the time was passing. Nevertheless, it was shown (Table 3) that the rate of honey bees recovered in the donor strip (18.2%) was 5 times higher than in the recipient strip (p = 0.0013). The most significant result is that 3.7% of the honey bees marked in one strip were quickly recovered in the other strip.

Table 3 : Recovery rate (number of recovered marked honey bees/marked honey bees ) in each strip.

Donor	60mn ± 15 after marking	18.2%
Recipient	15mn ± 15 after marking	3.7%

Moreover, a comparison of other recovery rates, based on observations made later on the « Donor » and « Recipient » strip, but within short time limits between both observations (10mn+10mn = 20mn), gave nearly similar results (Donor = 18.2% ; Recipient 4.9%).

The calculation of the proportion of marked honey bees/total number of marked and unmarked honey bees in both strips (after the same time length) which includes the effect of the renewal of the foraging population showed that the marked honey bees were diluted by unmarked new honey bees after a while. It also showed (Table 4) that honey bees recently coming from the “Donor” strip represented a low part of the total foraging individuals (1.7%). In few cases, one hour later this proportion could reach 8.8% in a sector. On average the proportion was 2.3% (30 observations = 3 sectors x 10 replicates).

Table 4 : Proportion of marked honey bees/total number of marked + unmarked foraging honey bees in each strip.

Donor	60mn ± 15 after marking	31.4%
Recipient	15mn ± 15 after marking	1.7%

Conclusion

Results showed that both strips were of equal attractiveness and that no gradient in honey bees density existed along the strip according to the distance to the hive. In our experimental conditions (2 x 180m²), the individual foraging area was generally limited to two contiguous sectors (120m²) and these results are in agreement with other results obtained in the past in other plant species and in various landscapes. Although honey bees tended to stay in a confined area, some of them flew from one strip to the other. The proportion of marked honey bees recovered in the other strip after a while could not be inconsiderable (about 4%). Nevertheless the results obtained immediately after marking (1.7%) gives a more suitable reply to the question of pollen transfer by honey bees from a plant genotype to another because the most important risk occurs when both genotypes are visited consecutively (Cresswell et al, 1995). The analysis of the recovery rates also showed that it is difficult to mark and observe honey bees in appropriate time scales because the foraging population is always moving and renewing. This preliminary study had led us to conceive a more adequate protocol to solve this problem and to write a probabilistic model to analyse the flow of honey bees between the 6 sectors. Despite these restrictions, it is obvious that a barren zone of 3m to 12m width between two similar small strips considerably reduced the movement of honey bees but it is not sufficient to totally eliminate pollen transfer by these insects.

Acknowlegments

The authors thank C. Guion and H. Picault for their technical assistance.

References

Cresswell J.E., Bassom A.P., Bell S.A., Collins S.J., Kelly T.B., 1995. Predicted pollen dispersal by honey-bees and three species of bumble-bees foraging on oil-seed rape: a comparison of three models. Functional Ecology, 9, 829-841.

Free J.B., Spencer-Booth Y., 1964. The effect of distance from pollinizer varieties in an orchard of dwarf apple trees. Journal of Horticultural Science, 39, 78-83.

Gary N.E., Witherell P.C., Lorenzen K., 1997. The interfield distribution of honey bees foraging on carrots, onions, and safflower. Environmental Ecology, 6, 637-640.

Lavigne C., Klein E.K., Vallée P., Pierre J ., Godelle B., Renard R., 1998. A pollen dispersal experiment with transgenic oilseed rape. Estimation of the average pollen dispersal of an individual plant within a field. Theoretical Applied Genetics, 96, 886-896.

Levin D.A., Kerster H;W., 1974. Gene flow in seed plants. Evolutionary Biology, 7, 139-220.

Manasse R.S., 1992. Ecological risks of transgenic plants: effects of spatial dispersion on gene flow. Ecological Applications, 2, 431-438.

Morris W.F., Kareiva P.M., Raymer P.L., 1994. Do barren zone and pollen traps reduce gene escape from trangenic crops? Ecological Applications, 4, 157-165.

Scheffler J.A., Parkinson R., Dale P.J., 1993. Frequency and distance of pollen dispersal from transgenic oilseed rape (Brassica napus). Transgenic Research, 2, 356-364.

Singh S., 1950. Behaviour study of honeybees in the gathering nectar and pollen. Mem. Cornell Agricultural Experimental Station, 288.

Stringam G.R., Downey R.K., 1982. Effectiveness of isolation distances in seed production of rapeseed (Brassica napus). Agronomic Abstracts, 136-137.