Bee Culture - November, 2006

by Walt Wright

In the last few years there has been a running battle between proponents and opponents of the effectiveness of the dance language of scouts or foragers to direct recruits to a forage source. The disagreement has oscillated since Von Frisch learned to decipher the dance many years ago. At first, most agreed that his observations were a step forward in knowledge and a major contribution to the craft of beekeeping. As time wore on, and other investigators tested the concepts in different ways, doubt about the effectiveness of the dance crept in. A group of investigators emerged that set out to debunk the dance language as an effective way to induce recruited foragers to find the source of the dancers' enthusiasm. They contended that the forage source was located primarily by the equivalent of our sense of "smell." There is no doubt that the sense of smell of the honey bee is acute. We cannot even comprehend how acute.

After watching the flap unfold between the advocates and the naysayers for some time, I concluded that they both were probably correct. If we had just happened to come in during an evolutionary step forward in foraging skill, where the forager had learned to describe direction and distance, and the recruit had not learned to interpret that message, both sides would be correct. Should we come back in a couple eons, when recruits have learned to read the message, we would find them going directly to the forage source.

I'm glad that my erroneous conclusion was confined to my quarters. This year in ABJ (April) Emily Smith and Gard Otis updated us on some experiments published in more prestigious publications in recent years. It's like we dumb beekeepers didn't have a need to know. There's a lot of data generated that we don't see - even subscribing to both ABJ and this magazine. In the Smith/Otis article, not counting Von Frisches original work and a thesis, eight periodicals and two university press outlets were listed in the references, plus two entries in a German publication. Five separate references were published in Nature magazine alone. Several other periodicals had been referenced two or more times. Do you think we may be getting left in the academic dust?

The more recent experiments convinced me that the bees do use the information contained in the dance language to their advantage. But this submittal is relevant to a circumstance where the dance language is not good enough. The reproductive swarm is reported to travel in excess of five miles, sometimes, to a distant nest site selected by nest scouts. Assuming those distances to be correct, any error in the dance data is amplified the further away the nest site is from the parent colony. We'll come back to this later. This introduction is just to provide some background for a short observation yarn and a discussion of that observation.

A description of the observation follows: The truck was located to the side of a row of colonies such that the flight path of foragers could be seen. The reason for observing the flight path was just to determine the direction of most forager traffic. The colonies were backed up to some trees for afternoon shade at the top of a low ridge overlooking the river bottom less than a half mile down hill.

At that time, a hypothesis had been developed that early season colony decisions were based on forage availability. An active monitoring of field forage was an on-going interest to confirm or refute the hypothesis. Unfamiliar with forage sources across the river (eight miles by road with bridges), I wanted to know roughly what percentage of foragers were working over there. As feared, most were headed in that direction, and very few climbed up and over the trees behind them. Scratch this group of hives from evaluation of the hypothesis. This beekeeper was not going to scout the mix of forage sources across the river. But that is not what this submittal is about.

Lighting was very good for observing the "streaks" coming and going. Sun angle, glinting on the wings, made them quite conspicuous in flight. And there were many at any one time. The hives wintered well and were quite strong. The season timing was early in the swarm issue period. At any instant there might be a dozen or more streaks in my field of view.

Having the info I came for in just a few minutes, I was about to move on when an outgoing streak got my attention. Eyes are attracted to, and follow, motion naturally. That prominent streak branched into five or six individual bees about twenty yards out, and the group continued in tight formation out of sight. Whoa! Did I really see what I thought I saw? Decided to stay a little longer to see if it happened again. It didn't take long. A prominent streak came from the third hive down the row (watching closer in) and did the same thing. Staying long enough to see two more occurrences convinced me that what was seen the first time was real and not an illusion.

My conclusion at the time, and it hasn't changed since, was that what I was seeing were nest scouts leading recruits to a promising distant location. I have regretted many times not having supported that conclusion by spending a few more minutes watching the group on the landing board getting ready for take off. Confirming completion of swarm preps would have helped sell the concept. It is inconceivable to me that others have not seen this phenomenon. But then, as noted earlier, I see only a small fraction of the information published about honey bees. Have you seen anything written about honey bees flying in formation like the Blue Angels at the air show? I think not, but that's mostly speculation.

Having failed to support the observation at the time with more data, and considering others might be interested, we have resorted to arithmetic to support the need. The dance language is considered to have too much error to direct nest scout recruits to a distant location. Finding that distant potential nest site location by scent (smell) is out of the question. A tree hollow in acres of forest is not anything like a forest with dispersed blooming trees. If the recruit finds any one of the scattered trees in bloom the dance has accomplished its mission. She can home in on the fragrance once in the area. A potential cavity in a residence located in miles of subdivision houses is also low in scent value. Forget smell.

With smell ruled out, that leaves the dance language. We have not the foggiest notion of how accurate that is. It's only recently that its use has been confirmed(?). The skeptics are not convinced yet. The controversy persists. Perhaps some of the smell advocates would like to take a crack at the question we are addressing here.

Back to accuracy of the dance language: I try not to guess much and report only what I see. If what I see has an obvious conclusion, I might report that as a conclusion, rather than a fact. You have the option to draw your own conclusion from the observation and ignore mine. In this case, data from another source will be used as a basis for the conclusion that will be supported with some arithmetic. We have for purposes of this discussion arbitrarily selected ±1%. We need a number to work with for the arithmetic. Two bees are involved - the dancer and the recruit. Each can add error in the message. The dancer can be a little off, and the recruit a little off in both her reading of the message and the flight application of the message. The ±1% error selected looks like a large slice of 360 degrees (360°), but is well within the error of the experimental data. One of the experiments brought forward by Smith/Otis was the work of Riley, and others in 2005. Recruits were collected on exit and fitted with a tiny transponder. They could then release the recruit and track the flight path with radar. The worst case angular error was greater than 10° on both sides. The selected 1% error of 3.6° is therefore conservative. Assuming the sharp change of direction on most of the plots reflects the change from dance directions to the odor search mode, the distance error is well outside our discussion error of 1%.

Figure 1. Area to be Searched

The shaded area of the sketch reflects the area to be searched by the recruit, if the nest site is three miles from the parent colony. Remembering the pi factor of circle area calculations has a 3.14 multiplier, further distances build area numbers at a rapid rate.

Without walking you through the calculations, the ±1 % error yields almost 14 acres of trees. Using a tree spacing of twenty feet between medium to large trees per acre (also arbitrary) yields about 1500 trees that need examination by the scout armed only with dance directions. The scout looking for a nest cavity must check the whole tree trunk, all the way from the ground up, and all the way around. She is looking for a dark spot that may be an opening. She will check dark spots more closely that do not turn out to be openings. The search is a time-consuming effort.

That is not a practical way to go about it. Honey bees are efficiency experts. I remember writing somewhere that conservation of resources is a cornerstone of their survival format. That concept has applicability here. To send a nest scout to a forested area to look for the tree hollow that has already been found by a nest mate makes no sense at all. What better way to get a review of the site by others than to have the original finder lead them there? That group then, in turn, can lead others to the site.

Another factor to be considered is the urgency of nest site scouting. The swarm that left the parent colony, without having selected a place to go, is in a hurry. They left the parent colony with limited resources, and do not have much time for the "voting" process. Multiple potential nest sites must be examined by many bees in a short time to arrive at a consensus. There is not much time to cycle bees to multiple potential sites for comparison. Taking them there in groups by a guide that knows the way should definitely speed up the voting process.

I call this phenomenon "shotgunning." Watching the streak get larger until it was comprised of several individual bees reminded me of the dispersal of shotgun pellets. The pellets exit the gun in a tight group and disperse with distance to make a large pattern to insure a hit on the target game. If a description of the phenomenon exists under some other name, I'll be happy to make the adjustment.

Shotgunning was only seen on this one occurrence about ten years ago. That same year investigation of checkerboarding started with a strong outyard in a different location. In the intervening ten years, only two colonies swarmed. Both of those swarms were the result of failing to follow my own recommendations. Other priorities caused me to get a little sloppy in maintaining empty space at the top. This paragraph is not intended to brag but to point out that I may never see shotgunning again.


If you still have swarming, you might have an opportunity to see it yourself. Look for it on the settled swarm cluster that didn't move to a pre-selected nest site the same day. That cluster has increased motivation to get the vote finalized. And leave the swarm box at that location - at least until dark to collect nest scouts returning for nourishment. Those senior foragers will be an asset during the establishment phase.