Horn-faced bees vs. indigenous pollinators in blueberries

Final Report for FNE05-554

Project Type: Farmer
Funds awarded in 2005: $4,304.00
Projected End Date: 12/31/2007
Region: Northeast
State: West Virginia
Project Leader:
Robert McConnell
McConnell Berry Farm
Expand All

Project Information


Introduction and Background

Most small blueberry growers rely on wild or indigenous bees to pollinate their crop. In the past, wild honey bee colonies were plentiful, but parasitic mites have severely reduced the incidence of wild honey bee colonies so the only wild pollinators are bumble bees, carpenter bees, and lesser known solitary bees of several species. Generally speaking, the density of the wild species is fairly low due to lack of suitable nesting sites. Smaller growers usually get a reasonable crop set from wild pollination, but it is not known how close this is to optimum.

The purpose of this study was to introduce large numbers of solitary pollinators, specifically the hornfaced bee, into small blueberry plantings to see if an increased number of pollinators would increase the number of blueberries set and increase the size of the berries by improving the pollination.

Tests have shown that HF bees are excellent pollinators for apples. It take about 300 female horn-faced bees to pollinate and acre of apples while it takes a hive of 20,000-30,000 honey bees to achieve the same level of pollination. Tests in blueberries are much more limited, but it is clear the HF bees do work blueberries. Hornfaced bees have an additional advantage in that they prefer fruit blossoms and are less distracted by weed flowers than honey bees, which are more general pollinators.

Management of HF bees is less demanding than honey bees and consists of putting the bees out in the spring and putting them back in storage at the end of bloom period. Winter storage requires maintaining the temperature above about 20 degrees and keeping them below about 40 degrees in the spring until placing them in the field. The bees nest in inexpensive paper tubes which must be protected from wet weather. Field shelters can be quite simple and inexpensive. The only additional task is to count out the required number of tubes. Thus, even small growers can maintain their own pollinators.

The hornfaced bee is imported from Japan and is a close relative of the native blue orchard bee. Management practices for the hornfaced bee are well established but have not been very successful when applied to the blue orchard bee in the eastern US. Management of other species of solitary bees has been successful on a small scale, but usually relies on specific local conditions which are often not available or easy to maintain by the small grower.

The Experimental Process

We wanted to run a series of field experiments to compare the rate of pollination and the size of berries between plants pollinated by wild pollinators only and plants pollinated by both wild pollinators and horn-faced bees. These tests were carried out on three farms in the Appalachian region all growing for the u-pick market. Because hornfaced bees only fly about 100-200 yards from the nesting site, isolation was achieved by observing one set of plants near the nest site and observing another set of plants at least 300 yards away. Both sets of plants were the same variety (Bluecrop) and under the same cultural practices at each farm. The objective of this experiment was to determine the usefulness of hornfaced bees in blueberries by observing the percentage of blossoms set with and without HF bees, and the resulting average berry size under each case.

A field box of bees was prepared for each farm with approximately 200 filled tubes of bees from the previous season and an additional 600 empty nest tubes. We expected approximately 400-500 female bees which will be sufficient for 1-2 acres of fruit, based on apple density requirements. The number of bees in this experiment was based on the need to saturate the area near the bee box with active female pollinators while keeping the number low enough so the bees would not be forced to fly long distances for forage. The large number of empty nest tubes ensured that the females had adequate nesting space available and could be actively pollinating throughout the blueberry bloom period. In the field, these boxes were mounted on posts approximately 4 feet above the ground to prevent wild animal predation of the bee tubes.

The nest boxes were placed at the end of a field where we could get maximum isolation while simultaneously ensuring samples of the same variety and cultural practice with and without the HF bees. Because it takes one warm day for emergence to begin and about a week for all the females to become fully active, we placed the bees in the field about 3 days before significant bloom. The bees were set out in the field when the first few blossoms opened in late April: April 25 at Crow's Nest and White Oak farms, and April 23 at MBF. Because of the late spring weather, these dates are at least one week later than normal. At that time, we selected 10 plants closest to the nest box and selected a branch with five fruit/bloom clusters on each plant. We tied a ribbon around each branch. We repeated this process on ten plants of the same variety as far away from the nest box as possible.

In mid May; May 17 at MBF and May 18 at Crow's Nest and White Oak farms, an assay of pollinators was taken by Mr. McCutcheon with an insect sweep net. The collected insects were placed in plastic bags and put in ice. The insects were identified and numbers recorded later.

In early June, near the end of bloom, but before any un-pollinated berries dropped, each cooperator counted the number of blossoms and blossom buds on each branch. These numbers were recorded in a spreadsheet. Shortly after the end of bloom, mid-June, a mesh bag was tied around each selected branch to prevent predation and loss of berries from other causes such as wind drops and customer picking. The bees had finished their life-cycle by this time and were removed from the field and returned to storage.

In early July the berries began to ripen and were picked on approximately a weekly basis for four weeks. Berry counts and weights were kept for each branch as the berries ripened. At the end of the harvest period, any unripe berries were counted to be included in percentage pollination data, but not weighed or used in berry weight data. Apparently the mesh bags interfered with berry development and ripening as will be discussed in the conclusions.


Pollinator Assay and comments:

5/17 - Sampled the McConnell Farm, MBF. Upper field (Near Field) had many pollinators. Many hornfaced bees were observed and some collected. The lower field (Far Field) had fewer pollinators. It was sampled for one half hour. No hornfaced bees were collected although one was observed sunning itself, apparently a male.

5/18 - The O'Dell Farm, Crow's Nest Farm, was sampled 15 minutes in the near field and 15 minutes in the far field. Hornfaced bees were observed and not collected in the near field and no hornfaced bees were observed or collected in the far field. The number of pollinators in the far field was significantly lower than the near field. Honey bees were the predominate pollinator at the O'Dell farm.

The Robinson Farm, White Oak Farm, was sampled 15 minutes in the near field and 15 minutes in the far field. Hornfaced bees were observed and not collected in the near field and no hornfaced bees were observed or collected in the far field. There were more pollinators in the far field than the near field, but that was probably due to the wild cherry trees being in bloom located near the near filed.

Other notes - Hornfaced bees are very fast fliers and shy. They are very hard to collect with a sweep net especially when one tries not to damage the blueberry plants, blooms, and berries. Only when they were in high numbers (McConnell upper field) was I able to collect them. They are distinctive enough to recognize them through observations.


In some cases, branches were damaged after the blossoms were counted and berries picked. These entire branches were removed from the data on pollination, but any berries picked were used in the berry weight data. The above table shows very inconsistent results. In fact, in two cases we counted more berries picked than there were blossoms counted. Clearly, this data is invalid, at least for the cases shown. While the specific branch data is not shown here, in the underlying data, there are branches in each group and on each farm that show greater than 100% pollination. It was difficult to count blossoms on each branch because of the density of the blossoms. In all likelihood, our inability to count blossoms accurately was the root of the problem. It is also possible that late emerging blossom buds contributed to the problem.


These data show inconsistent results also. Two cases show improved size while the third shows reduction is size of berries using hornfaced bees. Again, the process used to collect data was probably the culprit. We used mesh vegetable bags to prevent predation and loss of berries. We noticed that in many cases, the berries inside the bags were considerable smaller than berries on the same branch but not inside the bag. In some cases, this size difference was substantial. One observation indicates that the size was related to the number of leaves inside the bag; the more leaves, the larger the berries. The most likely explanation is that the light reduction and/or constriction of the leaves caused by the mesh bag was enough to reduce photosynthetic action resulting in less nutrition going into the berries. Those bags with a large amount of growth and leaves inside still had sufficient energy to produce full sized fruit, while those with fewer leaves did not. Another factor may have been the process used to select the branches for data collection. In this case we selected branches with exactly five fruit buds. We may have inadvertently been selecting the poorest branches.

Observation of weather during pollination revealed information that may have played a role in the results. Within a few days of placing the bees in the field, the weather turned very cold, cloudy, windy, and wet for about 7-10 days. While not below freezing except at night, the temperatures were not warm enough for the bees to forage significantly. Following this period, we noticed a substantial reduction in bee activity, indicating some mortality from the weather. There were still bees flying, but not as many. Resulting filled tubes from egg-laying activity indicates the loss of bees was 50% -75%.


We can only conclude that we did not obtain valid results. The results were probably affected by weather limiting the population of bees for pollination, but the deciding factors were our inability to obtain accurate blossom counts for the percent pollination and major reduction in berry size caused by the mesh bags used to prevent berry loss.

If this study is to be tried again, we recommend that blossom counts be done more than once, and that a better method of preventing berry loss be used.

The motivation for this project came from the cooperators who tested these pollinators the previous year, but without actual bloom and fruit measurements. We note these growers' observations that in both these years, fruit yields and berry size were improved in the fields nearest the pollinator boxes compared to former years.


Participation Summary
Any opinions, findings, conclusions, or recommendations expressed in this publication are those of the author(s) and do not necessarily reflect the view of the U.S. Department of Agriculture or SARE.