Issue 89: Blue-Banded Bees Pass the First Hurdle

July/August – 2006
Author: Steven Carruthers

Blue Banded Bee on basil flower. Photo courtesy David Radel.

STEVEN CARRUTHERS looks at the latest published research to develop the native blue-banded bee as an alternative to bumblebees for pollinating greenhouse tomatoes. He writes that while some progress has been made, researchers are still many years away from reaching a commercial outcome.

Commercially reared bumblebees are used safely in over 30 countries to pollinate greenhouse tomato crops, but this technology is not available in Australia. Pressure from NZ imports, with recent approval for importation of Dutch tomatoes, and with Chinese imports on the horizon, means that if the industry hopes to match production standards with its international competitors, all of which use bumblebees, then access to this technology can no longer be ignored.

Following a three-year Environmental Impact Study on Tasmania’s flora and fauna, where bumblebees were inadvertently introduced in 1992, the Australian Hydroponic & Greenhouse Association (AHGA) can find no reason why bumblebees should not be allowed to be imported onto the Australian mainland to pollinate greenhouse tomato crops. Despite the gloom and doom scenario painted by a few individuals, bumblebees have had no adverse effects in the island State. Additionally, an independent CLIMEX modelling study only found limited opportunities for bumblebees to establish on the mainland should they escape to the wild. Subsequently, the AHGA applied to the Department of Environment and Heritage (DEH) to allow their import onto the Australian mainland.

In the meantime, blue-banded bee researchers have been working around the clock over the past four years to develop an economical and viable alternative to bumblebee technology. A newly published study assessing the ability of the native bluebanded bee Amegilla holmesi to buzz pollinate tomato plants does little to reassure growers that blue-banded bees are an economical and viable alternative to proven bumblebee technology. The single experiment, using only four bees, was conducted in a small greenhouse with two chambers to compare blue-banded bee pollination with mechanical pollination and with control plants with no supplementary pollination. The study, recently published in the Journal of Economic Entomology, concludes that the percentage of fruit set of bee-pollinated plants was not significantly different from the percentage fruit set of mechanically pollinated plants. So far so good. This research was conducted in 2002-03.

The experiment was conducted in two adjacent chambers in a glasshouse at the University of Western Sydney, Hawkesbury Campus, NSW, during summer from December 2002 to April 2003. The chambers measured 5.25 x 3 x 4.3m (22.58sqm) and were illuminated by ambient light. The temperature was maintained for optimum tomato production at 23°C during the day and 17°C at night.

Six nesting bricks were stacked in two columns on top of hollow, concrete Besser blocks at the end of each chamber for bees to nest. Mud collected from a site where Amegillanaturally nested was used to construct the nests in the Besser block.

Bees used in this investigation were collected from the wild as prepupae and allowed to develop in an incubator to the winged stage. When the bees were ready to hatch, two females and two males were randomly selected and placed on the nesting blocks in each chamber for emergence. The bees were observed daily and immediately replaced if mortality occurred. The study does not indicate the mortality rate or reason(s) for mortality. Because tomato flowers produce little or no nectar, the bees were provided with sucrose-water solution supplied on blue sponges.

Thirty tomato plants grown to first truss stage were placed in each chamber and arranged in four rows of seven to eight plants with a metre-wide aisle between the inner rows. Plants were randomly allocated to the three treatments – bee pollination, mechanical pollination (with a vibrating wand), and control (no supplementary pollination). As trusses developed they were pruned to four flower buds. Those receiving mechanical pollination or no supplementary pollination were bagged before the flowers opened. Pollination bags were removed as soon as the last flower was set. Trusses receiving mechanical pollination were vibrated with a commercial electric pollinator every second day between 10:00 am and 2:00pm.

Pest and diseases were controlled using methods safe for bees. Encarsia formosawere introduced every two weeks to control greenhouse whitefly (Trialeurodes vaporariorum), and plants were sprayed with 1% petroleum oil every two to three weeks to control aphids and powdery mildew.

Tomatoes were harvested when the fruit were orange-red and considered mature, then weighed using an electronic scale, and their maximum and minimum diameters measured with digital vernier callipers. Seeds from individual fruit were separated from fruit pulp, air-dried then counted. Only fruit grown on trusses 2-6 were used to determine the pollination efficacy.

The study reports both blue-banded bee and mechanical pollination treatments significantly influenced all the parameters assessed – fruit set, weight, roundness and number of seeds – but they did not differ significantly from each other (Table 1). The pollination treatments resulted in 94% fruit set, which was significantly greater than the 82% fruit set for the control treatment, but reported erroneously as notsignificantly different. The fruit was also heavier and had larger min/max diameters than those produced from flowers in the control treatment. Flowers pollinated by bees and mechanical vibrator also produced fruit that was significantly rounder and seedier than those fruits produced with no supplementary pollination.

The study concludes that these results are similar to those reported for bumblebee pollination (Banda and Paxton 1991, Ravestijn and van der Sande 1991, Pressman et al. 1999), and for stingless bee pollination (Cauich et al. 2004).

When interpreting the results of the study, it should be remembered that this is a single experiment conducted in a small greenhouse with two chambers of 22.58sqm using only four bees.

There were 30 plants at first truss stage placed in each chamber, with 10 plants per treatment in each. They were grown through to 6 trusses. The treatments were (i) two female and two male blue-banded bees per chamber, (ii) manual pollination and (iii) self-pollination. Trusses were bagged for the two last treatments so the bees only had access to 10 plants with flower trusses in each chamber.

Trusses were pruned to four flower buds, so the total number of flowers per chamber available to blue-banded bees is 240 flowers (6 x 4 x 10) over a period of 3-5 months (actual dates are not given, only December 2002 – April 2003). If we take a minimum of 90 days, this is 2.7 flowers/day available for two female blue-banded bees, or 1.35 flowers per day per bee (only female bees collect pollen; two males were included with the two females in each chamber to ensure that they were fertilised and therefore collecting pollen). As the only source of food other than artificial nectar, one might guess that this would not only be inadequate for brood production, but is a very high stocking rate per flower: perhaps a starvation diet.

The researchers report only 2-6 trusses were used in the analysis. There are vague comments in the discussion section about bees initially only collecting nectar for brood cell construction, which suggests that the first truss was not adequately pollinated. Why was the first truss omitted from the analysis?

Data for each plant for trusses 2-6 was combined before the treatment analysis, thus obscuring any difference relating to truss position. These differences could be quite informative.

Some bees died and were replaced, but the researchers do not elaborate on their mortality; only that the majority of female bees survived for the duration. There is no mention of brood production and new bees, so presumably we are only dealing with four bees in total?

It’s also worth noting that the bees were confined to an area of 22.58sqm per chamber, so they had very limited distance to travel to find flowers.

There are several reporting errors in Table 1. Percentage fruit set is given as 13.7% for both mechanical and blue-banded bee pollination. Presumably, this should be 93.7%. There are also conflicting claims that there is or is not a significant difference from the control treatment.

The researchers calculate from Morandin et al.’s Canadian data that one bumblebee can pollinate 11-24sqm of greenhouse tomatoes, and they compare this with one blue-banded bee able to pollinate 7.9sqm. An enigma is how they arrived at this calculation from a 22.58sqm chamber. The study ignores the fact that there were only 1.35 flowers/7.9sqm/day = 0.17 flowers/sqm/day per blue-banded bee available. In a commercial situation, bumblebees pollinate 5-7 flowers/sqm/day (D. Griffiths, pers. comm.). Also, we should not forget the substantial differences in travelling distance.

Clearly, this study needs to be replicated on a much larger scale to be credible. The only claim that can be made is that in a small-scale experiment, blue-banded bees were able to pollinate greenhouse tomatoes and achieve comparable fruit set to manual pollination every two days. While some progress has been made, researchers are still many years away from reaching a commercial outcome.

Some facts about bumblebee stocking rates
The number of hives needed at any one time will vary with crop type (cherry tomatoes have more flowers than beefsteak), the season (more needed in summer), crop density, greenhouse covering material (bees work best under high UV light), greenhouse size, Bombus species and sub-species etc. For Bombus terrestris, it is generally recommended that about 5-15 colonies, each with 50-60 worker bees and one queen, are employed initially per hectare, with a colony life of 8-10 weeks. On average, this is one bee per 20sqm, but some bees are tending the nest so only a percentage of workers are actually foraging in the crop.

In Ontario, for Bombus impatiens, it has been calculated that 2000 bee trips/ha/day give sufficient pollination of tomatoes (Morandin et al. 2001). Under high UV light, which was optimal, there were 4.8 trips per bee per day.

The stocking rate of one bee per 20m2 contrasts with claims that a worker bumblebee can pollinate at least 500 tomato plants or 250sqm per day (van Ravestijn and van der Sande, 1991), but might be so if only some of the bees are collecting pollen.

Bell, M.C., Spooner-Hart, R.N. & Haigh, A.M.
Pollination of greenhouse tomatoes by the Australian Bluebanded bee Amegilla (Zonamegilla) holmesi (Hymenoptera: Apidae).
Journal of Economic Entomology99: 437-442.

Morandin, L.A., Laverty, T.M. and Kevan, P.G.
2001 Bumblebee (Hymenoptera: Apidae) activity and pollination levels in commercial greenhouses.
Journal of Economic Entomology94: 462-467.

About the author
Steven Carruthers is the Managing Editor of Practical Hydroponics & Greenhouses magazine and Vice-President of the Australian Hydroponic & Greenhouse Association. Email:

Tags: , , , , , , ,