Home Queens Honey Production Employment Contact Us Research & Publications Tours Items For Sale LinksFAQs

By Karen Pedersen and Gil Pedersen of Pedersen Apiaries

I think we’ve all heard or said, “Queens aren’t surviving as long as they used to.”  I think we’ve also all heard, “Where are your records to show that?  Give us something to start with.”  Like every other beekeeper, we have limited time, so we only keep records that have a practical purpose and can tell us something.  In 2011, we started to notice a problem in our operation. While we have always kept good records, as the years have progressed since 2011, our records increased to track that particular problem.  This is not going to be the story of a nice experiment where we were able to hold variables constant to isolate a certain response.  Instead it is a messy story of trying to figure out the answer to a problem through observation, increased record keeping, and trying to make small changes.  Our experimentation has the advantage of drawing on a large sample, but because we’re a commercial farm and not an academic research facility, and were looking for a particular desired result (viable queens), that also resulted in certain biases.  We surpassed the level of our competence long ago, but until someone else with competence volunteers to help, it will continue to be something that we must endeavor to do.  Our purpose in writing this article is first to seek to find out if other beekeepers are seeing similar problems and secondly to find out if there is a researcher out there who might have the answer to our problem.

A scale attached to a Billet Ezyloader is used to track honey production for evaluating queens.
A scale attached to a Billet Ezyloader is used to track honey production for evaluating queens.

Pedersen Apiaries is a multi-generational commercial honey producing operation near Cut Knife, SK.  Sometime in the late 80’s, Pedersen Apiaries started grafting queen cells to raise queens for our operation.  By that point we were overwintering our bees and it made sense to breed from the best of our own stock.  By the late 1990’s we had refined our queen raising process.  We bred from queens that had survived 2 winters and had out-produced their neighbouring/sister hives the summer between those winters.  Not only were the winters and poor honey production culling out poor queens, so were our fingers as we judged them for not building up in the spring, for being defensive and for showing signs of diseases.   We grafted queen cells into swarm boxes of bees that are broodless, closed boxes with lots of pollen, liquid honey and young bees.  After 24 hours, the queen cells were transferred into the top box of cell builder hives.  Each cell builder hive was queen-right in the bottom brood box, which was separated from the top cell box by an excluder and a honey super.  The top cell box had hatching brood and lots of pollen beside the cells.  On day 14, the queen cells were separated and moved into separate tiny mating nucs to hatch and get mated.  From there they were marked and transferred to our own hives or sold.  The only real change we made to the process occurred somewhere around 2004, when we converted from using a Jenter comb to graft cells to hand grafting into homemade wax cups.  Raising our own queens had become an integral part of our commercial honey operation.

Lots of young bees are enclosed into swarm boxes with a minimum of 1 comb of pollen and 1 comb of liquid honey.
Lots of young bees are enclosed into swarm boxes with a minimum of 1 comb of pollen and 1 comb of liquid honey.


Typically, we expected at least 75-80% of the queen cells to develop into full cells.  We had no reason to track how many actually did develop into full cells and so we didn’t.  As long as most of the batches were successful, we were happy.  In 2011, the last couple of grafts had a significant drop in the success rate.  We still didn’t record it because it was the end of the season and it just seemed like a blip.  We had enough queens for the year, so it wasn’t a huge loss.  However, in 2012, our success rate of developing grafted queen cells into full queen cells had dropped to 30% and I was frantically trying to figure out what was going on. I was blaming it on the weather, but I also tried changing some other things, but it was all to no avail.


By 2013, we started systematically trying to figure out and track what we might be doing that was causing the queen cells to die.  We hypothesized that that our equipment was carrying a virus or other pathogen that was causing the queen cells to die.  We started with new cell combs, cell bars and wax to make the cell bars, making sure that it was both new wood and new wax.  I disinfected my grafting tool and the place where I graft.  I had someone else try grafting to see if I had just lost my skill or eyesight.  None of these changes had any effect. 

We contacted Geoff Wilson M.Sc., P.Ag. Saskatchewan Provincial Specialist in Apiculture for help.  He came to observe and critique our grafting on June 18, 2013.  Geoff said that he could see nothing wrong with our method.  As per his suggestions, we tried both plastic cups and wax cups to compare, we tried bee polished and not polished cups and we tried transporting the swarm boxes at a crawl.  We bought a brand new grafting tool, used fresh water in a glass jar as opposed to a plastic bottle, and used a new frying pan without a Teflon coating and more new wax to make new wax cups.  He was suggesting that perhaps we were seeing a residue problem in our process that was ending up in the water or wax.  He also suggested that we closely observe the cells when they were being transferred from the swarm boxes to the cell builder hives to determine if the cells were still alive at the 24 hour stage.  We marked those cells grafted with the changes versus those done in the same way we had always done them and waited to see if there was any difference.

We determined that the grafted queen cells were dead within 24 hours of being grafted.  They died within the swarm boxes. We learned that plastic cups gave us a better success rate than wax cups, but still did not make a huge difference to the success rate.  We ascertained that we still had not identified the cause of the queen cell deaths.  One batch had an 80% success rate, but we could not correlate it to any changes we had made or not made.  The other batches ranged from a 6% to 50% success rate and also seemed to have no correlation to the changes made or not made.  The difference in the plastic cups versus the wax cups was the only clue that led us to looking for other possibilities of chemical residues that might be making a difference.  By that time our short season was over.  We thought we had enough queens for ourselves, but barely. 

Mating nuc with cells just introduced.
Mating nuc with cells just introduced.

Unfortunately, the pain continued the next year as the queens we had raised proceeded to die over winter or supersede.  They just didn’t last like they should.  By this point, we were tracking the success rate of the queen cells from grafting to mating nuc, but we still weren’t tracking right from the grafted batch to the queen in the hive. We knew their parentage, but could not correlate it to the exact batch of cells and the success or failure of that batch or at what point in the season they had been raised.


In 2014, we hypothesized that there was a chemical residue in the pollen being fed to the queen cell larva that was causing them to die.  We further hypothesized that the residue was from a fungicide since the problem of queen cells dying seemed to coincide with when widespread fungicide spraying of cereal crops and canola began in our area.  We also wondered if nosema spores might be causing the problem since emerging research on nosema cerana seemed to indicate that the beekeeping industry understood little about it. 

We collected combs of pollen in the spring of 2014 from our deadouts, separating it into 2 lots.  Lot 1 came from two beeyards where there was much less availability of cultivated crop land (and thus canola) for the bees.  Lot 2 came from the rest of our beeyards that are predominantly surrounded by agricultural cropland.  (We will refer to these 2 groupings as Lot 1 and Lot 2 in the rest of the article to refer to the separation of the pollen by location. In future years, the pollen was mostly sourced from live hives during the season.)  We collected two random pollen samples from each lot to send to a lab to be tested for residues.  We were unable to find a lab that could test for all of the possible chemicals that were sprayed in our area, but we found a lab that would test for a lot of them.  The results came back saying that there was no measurable residue above the acceptable limits in either sample.  However, since to my knowledge there is no research confirming a safe dosage of chemicals fed to apis mellifera larvae (particularly queen larvae), the test told us nothing useful. 

Queen cells in plastic cups coming out of cell builders.
Queen cells in plastic cups coming out of cell builders.

The pollen that we had separated out from the hives in Lot 1 was then used in the swarm boxes with the grafted queen cells.  Every week, with the exception of one, we prepared the swarm boxes with pollen from Lot 1 and grafted into plastic cups in our traditional manner.  I will admit that because we are still commercial honey producers, rather than professional researchers, we specifically used pollen from Lot 1 rather than Lot 2.   Our first goal was to get queens.  Our second goal was to figure out what the problem was and we make no apologies for that. With the exception of that one week, our success rate ranged from 75 – 85%. In the one week that was an exception, we used pollen accidentally from both Lot 1 and Lot 2.  That week, the success rate dropped noticeably, ranging from 30-80% in the different boxes.  This certainly wasn’t a conclusive answer, but after 2 years of dismal failure, it was enough to give us hope and encourage us to believe that we might have at least partially identified our problem.   In 2014, we also fed all of our hives one dose of Fumagillin B in syrup the spring.  We then fed 50% of the first batch of swarm boxes, Fumagillin B while not feeding the other 50%.  We could see no difference in the Fumagillin B feeding.


In 2015, we continued to separate the pollen from the yards into Lot 1 and Lot 2.  Then we used pollen from both lots in the swarm boxes with the grafted queen cells while tracking where the pollen was from and what percentage of the queen cells developed into full cells.  The lowest batch coming out of the swarm boxes was 36% which still indicated a problem, but the next lowest was 69%.  These 2 low batches were from boxes that used pollen from Lot 2.  In contrast, the lowest success rates of batches from Lot 1 were 75% and 78%.  However, while indicating that we might be on to something the results were inconclusive because the average success rate coming out of all of the swarm boxes was 94% and that included pollen from both Lot 1 and Lot 2.

Cell builder hives.
Cell builder hives.

So, we weren’t able to recreate a consistently high death rate of the cells in the swarm boxes.  Since realizing that the cells were dying in the swarm boxes, the cell builder hives had not been of much concern.  They mostly raised the live cells that they received.   However, in 2015, cell development in the cell builder hives became much more inconsistent. The lowest development rate in the cell builder hives was 8%, and ranged up to 100%, with an average of 76%.  We did not have enough collected identified pollen to use in the cell builder hives, so mostly random unknown pollen was used in those hives. Therefore, we started to systematically collect pollen from hives during the summer and fall of 2015 noting where it was from and during which season it was collected.  We don’t have the capacity or knowledge to identify the pollen microscopically, but we are in the hives often enough that we can identify pollen by season and colour with some accuracy.  For example, bright yellow pollen in the spring is most likely willow and would not be canola, but bright yellow pollen in July would more likely be from sweet clover or canola, but would not be willow. By this point, we were identifying pollen as to where it was from, what season it was collected and whether or not it was canola pollen.  We labelled each comb of pollen by which bee yard it came from and either what type of pollen it most likely came from or the time of year that it came from.  Our hypothesis had narrowed to the point that we thought it was canola pollen that was the problem.  Whether that was because of fungicides, another type of chemical, or a varietal issue, we had no idea. 


By 2016, 5 years after we first noticed the problem, we were now tracking which pollen we used in both the swarm boxes and cell builder hives and the percentage of queen cells that were surviving in each.  We were now mostly stealing combs of pollen from live hives and storing it until the next season so that we actually had a better idea of when it was collected and what it was.  We were differentiating and tracking whether pollen was from Lot 1 and Lot 2, as well as whether it was from spring or summer.  We were still prioritizing getting queens over doing research, but by then we had realized it was also important to track the pollen right through the life of the queen cell that survived.  Was the longevity of the queen determined by the pollen that it was raised on?

In 2016, the average survival rate coming out of the swarm boxes was 93%, however, the median was 97%.  The average cell development coming out of the cell builder hives was 81% and the median was 90%.  There were only a few batches that were low in 2016, but 4 batches, in particular were exciting to find.  In 2 of those batches, cells from two different queen mothers, put into 2 different cell builder hives, dropped from a 94% success rate coming out of the swarm boxes, to a 50% success rate coming out of the cell builders.  In the end, from those 2 batches, only 7% of the originally grafted queen cells survived to become a 2016 queen going into winter. Then other two batches of cells of note came from the same queen mother.  They were put into a swarm box and came out of the swarm box at a 50% success rate that further dropped until only 4% of the grafted cells were still alive to go into mating nucs and none made it to queens.  The common denominator between those 4 batches of cells in both the cell builder hives and the swarm boxes was canola pollen that came from one beeyard.  We had no other pollen that came from that particular beeyard.  Therefore, we put 2 combs of pollen into cold storage at a research facility, one from the swarm box with problems and one from a swarm box without problems.  We think we have identified a problem and maybe even a smoking gun.  The problem now is to find who has the technology to test for what the problem is.  Sending them to a lab that will not test for everything and will not test below the set acceptable limits is a total waste of time and money.  We are definitely past our level of competence, but what else can we do?

We are still learning how to track the pollen straight through from queen development to the queen longevity.  It has added quite a bit of detail to our record keeping, but it is looking promising.  Obviously, we have to continue tracking to see whether or not we can find another comb of pollen that shows such striking differences in queen development.  From a research and investigation viewpoint, the challenge is that the more we learn, the less we want to risk using canola pollen in queen development. 


Hive ready to be wrapped for winter.
Hive ready to be wrapped for winter.

The winter of 2016-2017 was a long winter with a cold spring which was hard on our bees.  Our losses meant that we couldn’t afford to risk losing more hives on the possibility of non-viable queens.  As a result, we specifically chose to avoid feeding the combs of pollen that we had collected from Lot 2 or more specifically we chose to avoid using combs of pollen that we thought might be predominately from canola. While we couldn’t completely avoid canola in our area, we tried to avoid using pollen from beeyards with significant exposure to canola.  The average survival rate coming out of the swarm boxes was 89%, however the median was 94%.  The average cell development coming out of the cell builder hives was 82% and the median was 88%. This year, the lowest percentage of cells coming out of a swarm box was 56%.  That swarm box was fed by summer pollen so we could not rule out that it had canola pollen in it.    The first batches coming out of the cell builders averaged 58% which we judged was because of the cold spring and the cell builder hives not being strong enough.  Later in the season, we had only one other batch of cells that stood out.  They dropped to 35% and 56% in the cell builders that they were put into.  Again, they were fed summer pollen that could have had canola pollen.  We neither conclusively proved nor disproved our hypothesis.


While we know we have a problem, we are a long way from being able to identify its cause or solution.  It would be easy enough to argue that our hypothesis is doubtful because we still had problems in 2017.  We would like nothing better than for someone to definitively prove that our hypothesis is wrong.  That would save us countless hours going down the wrong path trying to track this problem. It would certainly simplify our lives if we didn’t have to worry about avoiding canola.  However, at this time canola pollen is the only factor that we have been able to identify as possibly holding the key to answering our problem.  At least we have a large sample size and some of the records that might help narrow the search.  If you have noticed similar problems, please let us know.  It should be widespread if our hypothesis is correct.  If you have experienced the problem and found a solution, we would really like to hear that.  This isn’t clean academic research.  We are trying to find out if anyone else is groping around in the dark with us.


Pedersen Apiaries Logo
Revised: March 1, 2018.
Copyright 2002 Pedersen Apiaries. All rights reserved.
Original Design & Graphics by Karen Pedersen
Photographic images are under copyright and used with permission of John Pedersen or Karen Pedersen