Category Archives: Uncategorized

My old comb building experiment

A few years ago I was intrigued by much of the chatter online regarding various ways and means of co-ercing the bees to build more comb, faster. One premise that I read over and over, bees will build foundationless faster than building on foundation, and they will only build comb on plastic if absolutely forced. I did a very crude little experiment to test this hypothesis for myself. Granted, my experiment was not statistically significant, N=1 is not a valid statistical set, but, this was rather enlightening for me. My method was simple, very simple, present the bees with the foundationless, and plastic foundation, on the same frame. The test setup


Some time later, we removed the frame from the hive and took a photo.


A quick measurement of how much comb we found in the empty space, and how much was on the plastic foundation gave me the answer I was looking for, and that answer was, no appreciable difference between the two halves of this frame, placed in this hive.

This last weekend we attended the BCHPA annual meeting and the associated education days. One of the sessions I sat in on was a presentation by Randy Oliver describing the correct way to go about a real ‘experiment’ and produce acceptable valid conclusions. My simple comb building experiment was essentially ‘all wrong’, about the only thing right with this one, indeed there was a test case, and a control, subject to identical conditions. That part was easy, the test and control were both on the same frame, placed in the same hive. the result isn’t truely valid because we had only one frame in one hive, so, no other hives replicating this result, and N=1 is not a valid statistical answer. But I also learned, this would be a relatively strait forward experiment to repeat, but, in a more formal manner to produce a result that would be accepted as a real ‘practical research result’.

For those that know me, they know I am big on doing applied research to get practical results that have implications for our business. Next season, during the main flow, we will have more than enough colonies building in 4 over 4 nucleus configurations to repeat this experiment in a manner that can generate a statistically significant result.

The question we asked before, will bees build more comb in a foundationless frame than on a frame with plastic foundation? Randy would say, the correct answer is ‘I dont know’, because we dont have data to show a definitive answer. I have _some_ data, but not enough to be definitive, so I’ll qualify my answer as ‘I dont think so’, but now I’m a bit inspired to produce a real answer, and this is a real answer that’s easily within reach. I’ve got 6 months to write up a proper experiment protocol which can be executed next season during our spring flow, and produce a real answer to this question.

This is going to be fun….

Another Hive scale

Last fall we were intrigued by the simplicity of the hive scale system from the folks at BroodMinder. I ordered one of the scales and a couple of the hive temperature and humidity monitors. Initially there were some problems with battery life, and I set the project on the shelf for the winter, didn’t want to be opening a colony weekly all winter to change batteries in the temp monitor. Last month I dug all that stuff out again when Jeff Lee from the BCHPA stopped by for an afternoon, and it left me inspired to finish setting up and starting a serious evaluation of the gadgets.

My conclusion now, the battery issues I had were just one defective unit, the other two units ran fine all winter stuffed in the closet. The scale went under a hive beside the already existing hive scale, and the temp / humidity gadget is now in the hive that has had a scale for years.

The Broodminder scale is a relatively simple gadget, lots of information at the Broodminder website. It is essentially a bar scale with two load sensors packaged up in a form factor similar in size to a 2×4. To use it, cut a piece of 2×4 to the width of the hive, and set that under the front or back of the hive, then set the scale under the other end. The easiest way to retrieve data from the unit is to simply walk up to it and use a smartphone with bluetooth.

For our installation I wanted live data coming to the website the same way the other scale works, so ofc I did not go the easy route for fetching data from these units. I put together a raspberry pi computer with bluetooth and set it beside the hives in a box, then wrote a small program that retrieves the data from the bluetooth LE advertisements coming from both the scale and the temperature sensor, then fed that data back into the same database we use for the first scale we set up.

I wanted to get a good evaluation of the unit in action, so it’s under a hive right beside the existing scale hive, and I’ve put live graphs from this one together with the original on this website. Looking at the data now that it’s been collecting for a while, I think this unit is very useable. It is by no means ‘perfect’ and it will take some effort in understanding to get ‘science grade’ data out of this one, but that’s really a minor detail. There is some temperature sensativity in the load cells apparent in the graphs, but it should not be difficult to account for, the scale includes a temperature sensor, so it’ll just be a task of getting the temperature co-efficients for those particular load cells. The preliminary graphs I have posted include both the left and right load cell readings, as well as the average of those two in 3 separate lines. The big spike a few days ago is another dump of snow that came down.

Preliminary graphs from the BroodMinder scale.

The theory behind this unit is fairly simple. If we have a scale under the back side of the hive, and a 2×4 pry under the front to keep it level, the scale should be registering roughly half of the hive weight. The theory is sound, and if you are just after accurate approximations of hive progress during a flow, the Broodminder scale ‘out of the box’ is more than sufficient. If you want science grade data from the unit, that too is fairly easy to accomplish, just buy two and use a second scale instead of a 2×4 pry under the other end of the hive, then add the two weights together. I haven’t started to investigate temperature corrections yet, but it shouldn’t be difficult. With temp corrections applied for the load cells, graphs from this unit should indeed be as smooth as they are from the original scale which does have temperature corrections applied to it’s data.

When we first set up the hive scale, the original intent was to get an accurate understanding of the flows in our new area after we bought this property and started turning it into Rozehaven Farm. For that purpose, getting to understand the specifics of flows in any given location, this scale is more than adaquate. We track the dates when we see various blooms, and correlate that with data from the scale. This data set allows us to better understand which blooms produce the nectar that turns into honey in our colonies. When we first moved here, the mantra on this part of the island was ‘blackberries are the main flow’. Today, I dont agree with that older mantra, and we have hard data to back it up. Over 3 years of data collection, our hives on the lot here have never gained any significant weight over the blackberry flow. The early honey comes from the series of blooms that preceed the blackberry bloom, and in our case, the start of the blackberry bloom is now our cue to remove honey supers and extract the early honey. We take the time after pulling supers to prepare our colonies for the move up to Fireweed.

For somebody interested in honey production, I think the Broodminder scale can be an invaluable asset in evaluating yard locations. It’s fully self contained and requires no outside infrastructure to use. Simply place a colony at a new location, set the Broodminder scale under it, then collect your data later using your phone. You will get a nice pretty graph showing exactly when the colony was bringing in enough surplus nectar to store it as honey. We are considering a couple of new outyards as we slowly expand our bee operation, and I think those locations will get a colony this summer that is sitting on a Broodminder scale to give us a detailed record of how well each location can do in terms of honey production over the season, and help us with learning to better time hive placement for catching flows.

Scale problems solved – I think

The erratic readings from the scale under a hive have been a problem since some time in December, and it seemed to get a lot worse after we moved the system from a spot on the lawn behind the house, out to the hive stands which now have a power plug installed.  At first I thought it may be wind or rain affecting things differently in the new location, but, in the end I became convinced the scale itself was the root cause of the problem, so I ordered a replacement.  Same scale from the same source, just a brand new one.

The new scale is in place, and like magic, the readings seem to have stablized dramatically.  We are using an inexpensive platform scale for this project, it’s not built to withstand the weather, and it is a single point load cell.  I’m not sure this type of load cell is meant to have a weight on it constantly over the long term, and I suspect over time the cell has possibly just bent a little internally, ie we essentially wore it out.  In the end it doesn’t really matter why the system became unstable, it had to be fixed.  Purchase cost of a replacement scale is in the same ballpark as buying one package or nuc in the spring.  Live data from a hive thru the spring flow has become a very important input for our strategies managing and in some cases micro-managing the hives for honey production.  With 15 hives set to produce honey this spring, scale data has more value to me than another hive, so this was money well spent.