Apis mellifera and Megachile rotundata: a comparison of behaviour and seed yield in a hybrid carrot seed crop

Abstract

The movement of managed alfalfa leafcutting bees (Megachile rotundata (F.)) in and around carrot flowers is compared with that of the honey bee (Apis mellifera L.) in a field trial to assess their relative efficacy as pollinators of hybrid carrot seed crops. Seed set is measured on carrot flowers inside cages with and without alfalfa leafcutting bees and compared with seed set in open field conditions in the presence of honey bees. Non-caged alfalfa leafcutting bees spend a similar amount of time on umbels and visit a similar number of umblets per umble as honey bees. Seed yield (number and weight of seeds) for flowers pollinated by caged alfalfa leafcutting bees is similar to that of flowers pollinated mainly by honey bees. However, it is not known how effective alfalfa leafcutting bees are compared with honey bees when not confined in cages (i.e. open pollination) under a range of commercial field conditions, particularly in the absence of plant species preferred by alfalfa leafcutting bees.

Keywords:

• Daucus carota
• seed production
• pollination
• field cages
• Apis mellifera
• Megachile rotundata

Introduction

Apis mellifera L., the honey bee, has been extensively relied on to pollinate a large number of crops (Delaplane & Mayer 2005). However, recent marked declines in honey bee populations have threatened reliable crop pollination services (Cox-Foster et al. 2007; Stokstad 2007). With the arrival of the varroa mite into New Zealand in 2000 (Zhang 2000), and reports of the mites developing resistance to miticide treatments elsewhere in the world (Strange & Sheppard 2001 and reference within), the future reliability of honey bee pollination services in New Zealand may also become uncertain. This has spurred a growing interest in the use of alternative insect species for crop pollination (Rader et al. 2009; Howlett & Donovan2010). Of particular concern is the viability of industries that grow crops that are poorly pollinated by honey bees. For example, in carrot seed crops, growers place large numbers of hives in fields (>10/ha in hybrid seed crops [Mayer & Lunden 1983]) to maximize seed set. Even then, pollination can be unreliable because foraging honey bees have preferences for certain carrot phenotypes or genotypes (Erickson et al. 1979) and bees are lured from the carrot field by other preferred flowering crops and weeds (Delaplane & Mayer 2005).

The alfalfa leafcutting bee is a major pollinator of seed alfalfa (Medicago sativa L.) in the western USA and Canada (Delaplane & Mayer 2005), but it is also considered a very good pollinator of a range of plants, particularly those in the family Fabaceae, including clover, vetch and trefoil (Delaplane & Mayer 2005). The alfalfa leafcutting bee was introduced into New Zealand in the early 1980s and soon proved to be an effective pollinator of alfalfa (lucerne) seed (Donovan & Read 1989). Alfalfa leafcutting bee visits have also been recorded on flowers from species within the families Crassulaceae, Labiatae, Liliaceae, Lythraceae, Plumbaginaceae and Umbelliferae (Small et al. 1997). A previous study demonstrated that alfalfa leafcutting bee (Megachile rotundata (F.) pollination of carrot flowers resulted in similar seed yield as honey bee pollination when each species was confined in cages on carrot plants in the field (Tepedino 1997). However, a study examining alfalfa leafcutting bee performance on a hybrid carrot field found that visits to carrot flowers by the bees were rare, and pollen in brood cells was predominantly from alfalfa flowers (Tepedino 1983). This may have been due to competition from preferred floral hosts within and surrounding the carrot field.

The purpose of this study was to compare the pollination performance of alfalfa leafcutting bees with honey bees in a carrot crop by evaluating behaviour on carrot flowers and carrot seed yield. In order to eliminate potential pollination by honey bees, alfalfa leafcutting bees were initially held in cages over carrot plants. The behaviour of alfalfa leafcutting bees was observed for both caged and non-caged alfalfa leafcutting bees.

Materials and methods

Field site and alfalfa leafcutting bees

The trial was undertaken between the summer months of January and March 2008 in a 4 ha, organically managed carrot seed crop located at Lincoln, Canterbury, New Zealand. Crops surrounding the carrot seed field included processing peas to the west (flowers present at the time of the study may have been attractive to pollinators including honey bees), potatoes (not flowering at the time of the present study) to the north, and pasture (predominantly ryegrass) to the east and south. Honey bee hives were present throughout the trial period, with four full-size hives per hectare.

Field cages

To determine the yield of carrot seeds in the absence of honey bees, a release tray of alfalfa leafcutting bees (approximately 2500 bees per tray, from a local supplier), with a nesting board, was placed inside one of two cages by 5:30pm on 7 January. Potted flowering alfalfa plants were also placed beside the nesting boards within the cages to provide nesting material. Four cages, 3 m tall with a 4×3 m floor area (1 mm² mesh), were set up in the carrot crop by 10.00am, 7 January, before the honey bees became active. Cages were separated by approximately 20 m and positioned just inside the carrot field boundary. Each cage covered up to 10 flowering carrot plants (0.75–1 m tall) with three to five male fertile and male sterile carrot plants. The treatments were alternated (alfalfa leafcutting bees or no bees) along the line of four cages. The end cages were approximately 20 m from neighbouring honey bee hives. Cages were removed on 14 January 2008.

Behaviour of bees

The behaviour of alfalfa leafcutting bees (inside and outside cages) and honey bees was assessed by recording the movement of bees over umblets and between umbels for up to 2 mins for 12–14 individuals of each species. Between 11:00am and 2:00pm on 9 January, six alfalfa leafcutting bees per cage were assessed and on 22 January 2008, 1 week after the cages had been removed, 12 honey bees and 14 alfalfa leafcutting bees were assessed. The number of umblets on each umbel visited, the time on the umbel, and estimated time on each umblet were not analysed formally, but were summarized with simple summary statistics as the data did not warrant formal analysis due to the small size of the trial. The estimated time per umblet was calculated as time on an umbel divided by number of umblets visited per umbel.

Carrot seed yield

To evaluate carrot seed set following the introduction of alfalfa leafcutting bees to the cages, three to five primary umbels (one per male sterile plant) were marked in each cage. Additionally, five primary umbels (one per male sterile plant) approximately 5 m from the neighbouring honey bees were also marked to evaluate seed set by honey bees and other potential pollinators. Before the cages were removed on 14 January, the marked umbels (caged and non-caged) were covered with mesh sleeves (50 cm long, 25 cm wide, mesh 0.25–2 mm²) to inhibit any further large insect pollination.

Umbels for seed set evaluation were removed from the plant while still within the mesh sleeves on 28 February. The diameter of each umbel, and weight and number of seeds per umbel were recorded 2 weeks after removal from the field. Umbel diameter, seed weight and numbers were analysed using analysis of variance. Numbers of seeds were square-root transformed to stabilize the variance. Analyses were carried out using GenStat (GenStat Committee 2008).

Results and discussion

Behaviour of bees

Alfalfa leafcutting bees (non-caged) and honey bees spent a similar amount of time on umblets and umbels and visited a similar number of umblets per umbel (Fig. 1). Alfalfa leafcutting bees observed inside the cages on 9 January spent more time on umblets (median = 15 s) and umbels (median = 58 s) and visited fewer umblets over the observation period for each individual (median = 2 umblets) than those observed outside cages on 22 January (medians; 7 s per umblet, 34 s per umble, 5 umblets visited over the observation period) (Fig. 1). Inside the cages on 9 January, the average maximum wind speed (measured at around 1.5 m above the ground) was 6 km/h. On 22 January, wind was gusting to an average of 17 km/h (measured at around 1.5 m above the ground), which may have contributed to the shorter time non-caged alfalfa leafcutting bees spent on a given umbel. High winds have been implicated in foraging alfalfa leafcutting bees being unable to relocate shelter, leading to high bee losses (Peterson et al. 1992).

Fig. 1  Individual observations (one dot is one observation) of honey bees (n = 12) and alfalfa leafcutting bees outside (n = 14) and inside cages (n = 12) of total time spent on an umbel (A) or umblet (B), and number of umblets visited per umbel (C). The numbers of observations are greater than the sample size because some bees visited more than one umbel during the observation period.

Carrot seed yield

Seed weights and numbers were not significantly different (P >0.05) for honey bees and alfalfa leafcutting bees (Table 1), but both were higher (seed wt., F=12.46; 2, 3d.f., P =0.039; seed no., F=12.58; 2, 3d.f., P =0.035) than those from plants in cages without either bee species (Table 1). There was no significant difference for umbel diameter (F=0.74; 2, 3d.f., P=0.64) between the three treatments (honey bee, alfalfa leafcutting bee, no bees) (Table 1). Consistent with this study, Tepedino (1997) recorded similar carrot seed weights when comparing pollination by caged honey bees and caged alfalfa leafcutting bees.

Table 1  Mean umbel diameter, seed weight and number of seeds per umbel for umbels exposed to honey bees and other pollinators, alfalfa leafcutting bees, or no large insect pollinators (no bees). Numbers in brackets are means on the square root transformed scale

Treatment	Umbel diameter (mm)	Weight of seeds (g)	Number of seeds
Honey bee	57.2	7.12	1003 (31.7)
Alfalfa leafcutting bee	50.4	6.91	850 (29.2)
No bees	50.0	5.20	185 (13.6)
LSD (d.f.=3)	21.1	1.09	(12.4)

Around 200 g (equates to approximately 2000 bees) of alfalfa leafcutting bee cells and adults were collected in late March. This was around 40% of the original number of bees placed in the field in early January.

Conclusion

Alfalfa leafcutting bees will visit carrot flowers and, under open field conditions, their flower visitation behaviour is similar to honey bees. However, it is not known how alfalfa leafcutting bees would compare with honey bees in open pollination under a range of commercial field conditions, particularly in the absence of plant species preferred by alfalfa leafcutting bees.

Acknowledgements

The authors wish to thank Richard Chynoweth (Foundation for Arable Research) for support; Ross Wilson (Midlands Seeds) for advice regarding experimental design and information on carrot seed production in New Zealand; Tim Chamberlain for providing access to the carrot seed crop; Georgie Meredith for technical assistance; and Corina Till for technical advice. This work was funded by the Foundation for Arable Research, New Zealand.