Pennsylvania queen bee improvement program and the Heartland Honey Bee Breeders Coop: Testing for mite-biting behavior

2015 Annual Report for FNE15-819

Project Type: Farmer
Funds awarded in 2015: $14,999.00
Projected End Date: 12/31/2016
Grant Recipient: Always Summer Herbs
Region: Northeast
State: Pennsylvania
Project Leader:
Jeffrey Berta
Always Summer Herbs

Pennsylvania queen bee improvement program and the Heartland Honey Bee Breeders Coop: Testing for mite-biting behavior


This project tested if there is a relationship between mite-biting behavior MBB levels and the total number of Varroa mites in a honeybee colony. MBB is a recently described grooming behavior trait that has been documented at high levels in some honeybee stocks, but it has yet to be proven as effective and practical tool for controlling/reducing Varroa mite populations.
The study group encompassed Pennsylvania, Ohio, West Virginia, and Ohio, consisting of 35 colonies of bees, from these we collected and evaluated 127 samples.  A formal statistical analysis was performed which indicated weak correlation between MBB vs mite populations. There are several possibilities that can explain this outcome, which are currently under advisement.  Anecdtolly, we observed that MBB colonies produced more honey, and overall increased vigor.  This leads us to conclude, that more work is warranted but with a different experimental design.

Objectives/Performance Targets

Our objective was to measure the relationship between mite-biting behavior (MBB) and the total mite count in honeybee colonies. We used an assay has been developed by Dr. Greg Hunt at Purdue which accurately measures the proportion of chewed mites in a colony.

The method has two major components, collection of mites, and evaluation of chewed legs. The collection of mites in the field by beekeepers is simple: (1) an oiled board is inserted over the bottom board, allowed to remain for 24-72 hours, then the mites are scraped off of the and onto petri dish;  (2) view each mite on the slide under a low power microscope/15X-handlense and count how many mites have chewed/missing legs, then calculate the MBB, as a percentage, example: 12 chewed mite/ 36 total = 33% .

We compared three groups: 1) PA-bred queens Artificially Inseminated with semen from MBB stocks raised in Purdue  2) queens from best local survivor stock in the HHBBC (3) queens obtained from commercial package breeders in the Southeast US

The beeyards locations used are in the HHBBC members respective states: J Berta PA, M Gingrich PA, Dan O’Hanlon WV, D Wells OH, and D Schenefeld IN.  Each yard will include 4 colonies from each group.

We measured the project colonies on 3 timepoints. Starting August, and ending in October, approximately 2 weeks apart. Note: some beekeepers started early, other later in the season.



We measured MBB percentages in 35 colonies, located in Pennsylvania, Ohio, and West Virginia; Note, Indiana is part of the study but no data could be aquired.

We collected 127 samples, and measured MBB percentages. The raw and calculated data entered into an excel spreadsheet, and was statistically analyzed.

The project had a substantial outreach consisting mostly of spreading the advanced MBB genetics, and informing hobbyist beekeepers how to select and produce their own stock.

Outreach- June 27, 2015 Queen cell exchange at Slippery Rock, PA attended by queen producers from PA and OH.  Master Beekeeper Joe Kovaleski presented advanced queen rearing techniques. MBB genetics were exchanged by 95 queen cell larvae.

Outreach- July 11, 2015 Queen mating nuc producing feild day in Linesville, PA. Almost 100 queens were mated using MBB genetics using queen cell cups inserted into participants own nuclues colonies which they brought. Over 65 attended.

Outreach- August 1, 2015 Mated and virgin queen cell exchange at the PA State Beekeepers Picnic. Over 30 queens were exchanged.

Report and outreach- November 13-14,2015 PSBA annual mtg in Lewisburg, PA.  Queen producers from approx 35 counties were updated on the 2015 progress.

Impacts and Contributions/Outcomes

It was interesting to note that the Purdue MBB bees seemed to produce more honey, than the control bees, although we did not weigh the hives. We had a Penn State graduate student, Mehmet Ali Doke perform a formal statistical analysis. The following is an excerpt of the results:
“Here’s a summary what we have:

1) MBB % versus total mite count (i.e. did biting colonies drop fewer mites?).
There are two important aspects of statistical analyses; significance and strength. The analysis done here (bivariate analysis/linear regression) was significant (p values lower than 0.05 for both full data and censored data) but the correlation is very weak (Rsq values lower than 0.07 for both data sets – a good correlation that you can freely report in scientific setting would be higher than 0.50, the higher the better, 1.0 being absolute correlation).




2) Does genetic background affect biting rates?


I used ANOVA to test this. The results are in the same direction with the first question we tackled. There is significant difference but the R values are very low making them, statistically speaking, “unimportant” since genetic difference cannot explain much of the variance in biting rates.




3) Then I wanted to see, independent from biting rates, “does genetic background affect mite drop?”


Once again I used ANOVA. The results are not different from the 2nd question. There is significant difference but the R values are very low.




***In any of the analyses, to my surprise, excluding certain data points did not make a great difference in the outcome.




For most of the analyses, group 1 and group 3 were significantly different and the difference between them was the driving force behind the overall low p values (hence, statistically significant but not strong results). That didn’t hold when I excluded low mite drop colonies and compare daily drop rates of the groups. In that case, Group 2 had the highest drop rate and groups 1 and 3 were not significantly different. One big concern I have is, even when I excluded group 2 and only compared groups 1 and 3, their difference was significant but still very weak (Rsq less than 0.10).”

We still have some other types of field data that we are going to look at, like alcohol wash mite counts, and Winter mortality.  This may be presented in the final report in March 2016.




Dan O'Hanlon
Bee Breeder
Carniolan Ct
Parkersburg, WV 26102
Office Phone: 3046910548
Dwight Wells
1305 Kerr Rd
Troy, OH 45373
Office Phone: 9374779251
Mark Gingrich
7070 Bull Rd
dover, PA 17315
Office Phone: 7178171398
Dr. Christina Grozinger
Penn State
501 ASI Bldg,
Univ Park, , PA 16802-0000
Office Phone: 8148651895