There are lots of reasons for wanting to know how many bees the are in a hive, gauging the possibility of a swarm for instance, estimating the capacity to pollinate a crop, or measuring a colony before some experimental work. Various ways exist to make an estimate, short of killing or anaesthetising the whole colony and counting every single bee.
How do you count bees?
Early science papers would describe a ‘gravimetric analysis’, a fancy way to say they weighed the bees (weight is the result of gravity). Packages and swarms are still described as a weight. If you know how much one bee weighs (about a gram) you can calculate the number from the weight of all of them. Being a fraction of a gram out, say, a quarter of a gram for 30,000 bees, meant you were wrong by more than two frames worth. If you weigh a small sample of bees and then count them you should get a more accurate average weight with which to calculate the whole, and if you really want to be accurate you do it at night when all the bees are home. Some people would weigh the entire hive and then shake all the bees out into a spare box, re-weigh the hive, dump the bees back in, and work out the weight of bees by subtraction. You can imagine how popular that was.
A more practical method was to look at the surface area bees occupied on the frames, and use some prior measure of their density to calculate the quantity. Instead of knowing the area a single bee (or a known number of bees) covered and dividing, by the 1950s researchers could use a prepared series of calibrated photographs in which the number had already been carefully checked1. For example, it had been established that one side of a frame that was half covered in bees (50% covered) contained 750 bees. By comparing real numbers of euthanised and counted bees with the estimates provided it could be shown it didn’t take much training to become proficient at estimating reasonably accurately.
I still have a set of frame photos, although I have to make a mental adjustment from British National 14”x 8.5” frames (75% of a NZ Langstroth) partly because I’m far too lazy to make a Langstroth set (and other country’s Langs are slightly bigger). Keen bee counters use frame-sized transparent grids to assist the area size estimate; serious bee counters nowadays get a computer to analyse the photographs, but the principle is the same. However it is done, the actual counting isn’t always the main source of error, other variables could affect the number of bees present; time of day, the season, temperature, solar radiation, wind, and so on.
Too complicated?
Either of these methods might be suitable for researchers, but they are not much use for day-to-day apiary use. In the late ‘80s Medhat Nasr and his colleagues at UC Davis in the US published and validated a ‘cluster-count’ method, with estimating colony strength for pollination hives in mind2. They wanted something that would provide a rapid, consistent, comparable figure and which could be used easily by someone without much practice, and they tested it against the frame area counting method. In essence, early in the morning (when the bees aren’t flying but ‘clustered’) they just split the colony boxes and counted the number of frames that wholly or partly covered in bees – they assessed the cluster by counting the frames in the top box covered at the bottom, and the frames in the bottom box that were covered at the top. No frames were lifted.
Except where the hives were very small, or very large, the cluster count proved to be as accurate (or as inaccurate!) as frame examinations, but could be done reliably in a fraction of the time by a pair of beekeepers (one minute rather than eight to ten checking frame-by-frame). There has always been some uncertainty about how the hive state or the weather would affect the estimates, but the cluster-count, or some interpretation of it, has been shown to be useful in practice, I was taught to do it more than 20 years ago but with a more modern adaption. Where the cluster-count looked at the covered tops and bottoms of frames we should now count the ‘inter-frame bees (IFB)’, what I call the bee ‘seams’. In general use the IFB proved to be reliably correlated with the hive’s bee population. A few years back this method was tested, validating the results for reproducibility under a wide range of circumstances, different skill levels, and against the night-weighting method3.
Science, applied...
The researchers used beekeepers with 181 ten frame Dadant and Langstroth hives in Avignon (France), with and without supers, counting in March, May July, and October. The colonies ranged in size from around 3000 – 53,500 bees. The beekeepers counted the seams of bees occupying the space between the frames, as long as they covered at least half the length of the frame, so a maximum of 10 IFB ‘units’ in a brood box, and 9 IFB units in the supers. The smallest unit could be a half, so 0.5. They had to adjust the counts to allow for the different sizes of frames with a simple size ratio, and they recorded the meteorological conditions when counting. The counts were taken when viewing from above the box, from below, and both counts averaged. In the case of the supers they always used a mean of the top and bottom IFB count (and halved the value because the super frames were half the size). There were different groups counting, some experienced, and some not.
For the Langstroth hives (counted in October and which were not supered) the relationship between the frame-space IFB count and the actual bee population was linear, and most accurate using the averaged top and bottom counts. However, these hives were small, between 4 and 8 IFBs, it probably isn’t quite linear. This was different for the Dadants, when large bee populations were underestimated and small ones overestimated unless a logarithmic relationship was used. It was interesting that the error you would expect when some bees were foraging was negated by the rest of bees dispersing less densely into the inter-frame spaces. For the Dadants the count from the bottom apparently provided a more robust estimate under different light or temperature conditions. Experienced counters did fare a little better than inexperienced ones, but from what I can see plus or minus 10%. In a report full of formulae and statistics I did expect to see a straightforward comparison table between the bee population as converted from the count and the bee population weighed out, but there wasn’t one. It looks like a frame or two, say +/-5,000 bees would be right 95% of the time, so practically equivalent to weighing them all anyway.
Counting your own
This is how I was taught to estimate colony size. At the time, the aim was to monitor and equalise colonies so that they would meet a specification for pollinating hives. There had to be an agreed number available by a particular date so having a reliable measure of how they were developing was important. If some were not going to make the grade we had to know so that we could do something about it. You may just want compare another beekeepers, or judge when to add more boxes. Six easy steps...
Drift a little smoke into the entrance, one gentle puff from a properly lit smoker is plenty, the secret is to wait for it, and the message it conveys, to spread. Standing at the back of the hive gently remove the lid and put it on the ground, upside down. With your hive tool crack the hive mat (hive cover, crown-board, etc) and tilt the back end up towards vertical so you can see the tops of the frames.
A ‘seam’ of bees describes the gap between two frames that you can see occupied by the bees from both frames (the ‘IFB’). A ‘full’ seam is fully occupied by bees (80-100%) from one end of the frames to the other. Both gaps between a frame and the side-wall are only 1/2 seams. Estimate the full seams of bees that you can see (to the nearest 1/2) and remember the number (for example, say, ‘3’) or write it on the lid.
Put the hive mat back down. Now crack the join between the top box and the next. Once the box has been loosened, wriggle the box back towards you about 50mm so that the front of the box sits on a ledge made by the lower box. Now lift only the back of the box, tilting it up using the ledge you made at the front as the hinge point, if you can, without letting the hive mat slide off (if it does put it back on when the box is flat). If you try to tilt a box without making a ledge it will slide off the front!
While the top box is up you can count and remember the number of seams you can see on the top of the bottom box (eg, 7). Now look at the bottom of the frames in the top box and count the seams of bees you can see and remember the number (call it 2). If it’s that time of year, once you have counted them you can have a look for swarm cells on the bottom edge of the frames and you might need to drift a little smoke up onto the frame bottoms so you clear the bees out of the way and get a good look.
Lift the top box you have just looked and put it on the upturned lid, at an angle to it so that there are four points-of-contact with the lid, one for each face of the box. Put the hive mat on top to cover the box if it isn’t there already. Always work to avoid squashed bees and reduce the number that take flight.
Use your hive tool to break the seal between the bottom box and the floor, and, just as you did with the top box, shuffle it back 50mm and tilt the back end up. You can count the occupied seams along the bottom of the frames, check for swarm cells, and then put it down again. Lets say there were 5 seams. In the bottom box were 7 + 5 seams (12).
So far, in the time it has taken to read the instructions, you have managed a basic check of the colony’s size. If you have to inspect the frames themselves (maybe you spotted swarm cells, or want to check something else) this is where you should start, in the box on the floor – where the entrance is that any bees you disturb will be returning to, and before the box has been filled with returning bees from other boxes. Always start at the bottom and work up. If you start at the top and work down you’re making life difficult for you and your bees.
How many bees have I got?
For a moment, let go back and think about the counting. The bottom box had twelve seams, the top box had five, so seventeen in all. A full depth frame, completely covered in bees in both sides will have around 3000 bees on it. By approximating the top and bottom ‘halves’ as the seams of bees we counted our 17 halves is worth 8.5 full frames, or 25,500 bees, 18,000 in the bottom, 7,500 in the top. A 10 frame full depth hive box could house 30,000 bees (10 x 3,000), so this 2 box hive isn’t quite half full of bees. If you use 3/4 or 1/2 depth frames I’m sure you can make the adjustment, it’s a size ratio. And yes, in principle a ‘top only’ count should work for TBHs but you’ll have a bit of work to do, it’s the price of being unique! For our purposes the conversion to a number of bees doesn’t actually matter; to compare inspections, or hives, all we need is the number of full seams. As long as we count in exactly the same way every time we have an excellent proxy for absolute colony size.
Jeffree, E.P., 1951. A Photographic Presentation of Estimated Numbers of Honeybees ( Apis Melufera L.) On Combs in 14 × 8½ Inch Frames. Bee World 32, 89–91. https://doi.org/10.1080/0005772X.1951.11094703
Nasr, M.E., Thorp, R.W., Tyler, T.L., Briggs, D.L., 1990. Estimating Honey Bee (Hymenoptera: Apidae) Colony Strength by a Simple Method: Measuring Cluster Size. Journal of Economic Entomology 83, 748–754. https://doi.org/10.1093/jee/83.3.748
Chabert, S., Requier, F., Chadoeuf, J., Guilbaud, L., Morison, N., Vaissière, B.E., 2021. Rapid measurement of the adult worker population size in honey bees. Ecological Indicators 122, 107313. https://doi.org/10.1016/j.ecolind.2020.107313