Odors, Wind and Colony Foraging (3 of 3): Insights from Beehunting

[1998 Wenner, A.M. Odors, wind and colony foraging — Part III of three parts: Insights from beehunting. Am. Bee J. 138:897-899 (Dec. issue).]

967 Garcia Road
Santa Barbara, California 93103
(805) 963-8508

“…in the drift of the years I by and by found out that a Consensus examines a new [idea] with its feelings rather oftener than with its mind. You know, yourself, that is so. Do those people examine with feelings that are friendly to evidence? You know they don’t.”
Mark Twain

Part I of this series (October issue) treated some of the points that Steve Taber made about honey bee recruitment in the April, May, and June issues of this journal. That first part documented the fact that bees which leave their colonies after contact with “dancing” bees need an odor marker or they cannot locate the sought-after food. Von Frisch had made that same point most emphatically in the late 1930s (Wenner, 1993). Natural odors suffice in normal circumstances; experimenters necessarily include an odor (deliberately or unintentionally) in or with the sucrose solution they use for their experiments.

Part II of this series (November issue), based mostly on research by Friesen (1973), revealed how wind direction greatly influences recruited bees in their search for a target food source. As indicated therein, the direction and distance of the food with respect to prevailing wind conditions matter a great deal. In essence, odors of food sources located downwind from the hive travel only further down-wind (away) from the hive. Even so, searching bees have little trouble if the crop is within a couple hundred yards. That is because recruited bees apparently begin their search downwind (e.g., Wenner, et al., 1991) and not necessarily in the direction expressed in the dance maneuver.

Friesen’s studies further revealed that searching bees rarely manage to find a target food source located any great distance (more than 300 yds) downwind from their colony. That situation holds true unless a great many foragers provide an “aerial pathway” between hive and food (e.g., Wells and Wenner, 1974).

Searching bees can more readily find rather distant upwind food sources (Friesen, 1973), a success that contrasts markedly with results obtained when target sources are several hundred yards downwind. The results provided in Parts I and II of this series thus do not mesh with what one should expect from the original von Frisch dance language hypothesis (e.g., Wenner and Wells, 1990; pages 63 and 64).

Another point pertains here – a truism in science: Good scientific hypotheses eventually have practical benefit. In this case one would expect value to emerge for beekeepers, but the dance language hypothesis has not met that expectation. At least, to the best of my knowledge, no beekeepers have come forth with testimonials about how their operations have improved this past half century due to the existence of the language hypothesis.

On the other hand, the odor-search model of honey bee recruitment (e.g., Wenner, et al., 1991; see Southwick in the October 1992 issue of The American Bee Journal) has immense potential – though ignored by bee researchers since first introduced a quarter century ago (e.g., Wells and Wenner, 1974). In this Part III of the series, I will illustrate how our knowledge of the importance of odor and wind direction aided us in a long-term research project on Santa Cruz Island, 25 miles offshore from Santa Barbara.

Beehunting Exercises and Recruit Search Behavior

The excitement of searching for “wild” (feral) bee colonies has persisted for centuries (e.g., Wenner et al., 1992; Mangum, 1998). Beehunting is not only fun and challenging; one can learn much about honey bee foraging patterns and recruit search behavior by engaging in that process.

In 1988, Robbin Thorp of the University of California campus at Davis and I began a long-term research project on Santa Cruz Island, a 96 square mile (25,000 hectare) uninhabited island offshore from Santa Barbara (Wenner and Thorp, 1994; Wenner, Thorp, and Barthell, 1999). Among other accomplishments – by the use of improved beehunt techniques (Wenner, et al., 1992) – we located 136 feral colonies between 1988 and 1993. In doing so, we effectively repeated the essence of the Bogdany and Taber 1979 study (as outlined by Taber in the May issue of this journal) a great many times.

Fortunately, we worked during those years under several advantages not available to Bogdany and Taber (1979) in their experiments, including the following: 1) We did not have the confounding problem of step-training our foragers. That is, extraneous searching bees lost during the training process can sometimes learn visual and odor cues that could help them in their search later. Instead, to obtain natural flight data, we converted foragers to our feeding dishes from blossoms or water. 2) We worked with large, natural colonies, not a small observation hive. 3) The few bees we began work with at any one time (as against the 30 foragers used in the Bogdany and Taber experiments) did not provide a significant aerial pathway for searching bees (see Wells and Wenner, 1974). 4) We located colonies at distances shorter than, equal to, and often exceeding those in the Bogdany and Taber experiments. 5) Extreme drought years prevailed in our region between 1986 and 1991, with no rain between March and November and very little during the winter months (Wenner and Thorp, 1994). That situation made our artificial food sources especially attractive.

In all those years, we never experienced appreciable downwind or crosswind recruitment at any great distances. Under those conditions, truly naive searching bees had apparently paid no attention to any of the distance or direction information they presumably had obtained from dance maneuvers before leaving their colony (see below).

Fortunately, the results from our crucial experiments in the late 1960s and early 1970s (Chaps. 9 and 10 in Wenner and Wells, 1990) proved a boon. We understood well what von Frisch meant when he wrote, in the late 1930s (e.g., Weaner, 1993):

“…I put some other dishes farther and farther away in the meadow, observing whether they would be found or not. The farther they were, the longer time it took till they were found by the bees sent out by the dancer. In the last experiment they were found after 4 hours in a meadow a full kilometer from the hive….lt is clear from a long series of experiments that, after the commencement of the dances, the bees first seek in the neighborhood, and then go farther away, and finally search the whole flying district.”

Consider now a few concrete examples from our experiences at locating feral colonies on Santa Cruz Island:

1) Black Point Colony. A student converted a foraging bee to his scented feeding dish, marked that bee, and gathered data on flight times and homeward bearing. By noting the bearing and calculating round trip times (Wenner, et al., 1992), he learned that he was about 1300 yds (1200m) downwind from the parent colony. During the entire first day and a half, he obtained no recruits, despite drought conditions while working in a grassland area.

In the middle of the second day the student set up an auxiliary station within 200 yds of the estimated colony location. Immediately, he was deluged with new recruits. (Compare that experience with the Friesen results – Figs. 3 and 4 in Part II of this series – and with the above von Frisch quotation). He then picked up the dish with scented sugar solution and walked slowly downwind toward the initial feeding dish location. Upon arriving there, he found a great many bees investigating the stock bottle of scented food in his vehicle – something that had not happened earlier.

The volunteer had, in fact, thus moved the body of already searching bees downwind, bees that had been stimulated to leave the hive (perhaps repeatedly) the previous day and a half after their contact with the lone forager. Obviously, those recruited bees had earlier failed to use any of the direction and distance information contained in the dance maneuver.

2) TNC Camp View Colony. Several individually marked bees routinely foraged all week at our feeder in mid-August, 1991, a drought year, with no recruitment from the parent colony – estimated to be about 1200 yds away across wind, over a canyon, and over a 700′ ridge.

When we eventually went upwind (down canyon) from that colony, we could hear bees searching everywhere; a small dish of food set down in that area had about a hundred recruits within a half hour. Again, compare that experience with part of the above quotation by von Frisch: “the bees first seek in the neighborhood…”

3) Navy Facility East Colony. In mid-August of the same 1991 drought year, a half dozen individually marked bees routinely foraged at our feeder from their parent colony, about 1550 yds distant (across wind, over a canyon as deep as the one the Bodgany and Taber group used, but with sharper topography) for two full days. No recruits appeared on either day.

4) N. Cyn FS/A Colony. This was one of the first real eye-opening experiences for our volunteers. In mid-June of 1988, we had a half dozen individually marked bees travelling between our feeder at ~250′ elevation to their colony across the main stream bed (crossways to the wind flow and later determined to be 850 yds away at an elevation of 550′).

Day after day our volunteers watched those few faithful foragers make their routine trips and fail to gain any recruits. We found the target colony only by placing auxiliary feeding stations rather near and upwind (down canyon) from that colony.

5) Prisoners Stream East Colony. By converting bees from water (150′ elevation) to our feeding dish in mid-September of 1989, we determined (again, by obtaining flight direction and round trip times) that the parent colony was only about 275 yds up a side canyon (just under 500′ elevation). We scrambled up the steep slope and set up an accessory station on a ridge crest only about 65 yds crosswind from the colony (as determined later – and at a somewhat higher elevation).

However, we gained no recruits at that close a distance, despite the fact that we had a few individually marked bees routinely foraging there. The reason for failure later became obvious – winds rise from ridge crests and thus carry odors upward, not along the ground. By contrast, a feeding dish placed 100 yds down canyon (upwind) from the colony immediately became deluged with searching bees that had obviously been stimulated to leave the hive sometime earlier.

In all five of the above examples, practical results demonstrate the importance of odors and wind patterns for recruitment of naive bees to a food source and fall in line with expectations of the odor search model (e.g., Wenner, et al. 1991). Yet, not one of those experiences would have been expected according to the 1946 von Frisch dance language hypothesis (Wenner and Wells, 1990; pages 63 and 64).

Simple Experimental Designs for the Truly Curious

As a start, one can readily repeat the Friesen experiment that formed the basis for Figures 1 and 2 in Part I of this series. After step-training the bees (e.g., Weaner, 1961; Excursus GT in Wenner and Wells, 1990; Taber in the April issue of this journal), gradually move the feeding station crosswind to an appreciable distance from the hive (e.g., more than 300 yds). At one location at a time, establish routine foraging, mark ten or fewer bees, kill all excess foragers, and gain information about recruitment patterns.

Once on location, continue to kill and tally all new recruits (unmarked bees). After switching from scented to unscented food halfway through an experimental period, recruitment should cease. If unmarked bees continue to arrive, then determine and eliminate whatever contaminant odor is responsible for the continuing success of searching bees.

Alternatively, one can step-train foragers to a feeding station 550 yds (500m) downwind from a hive with one odor marker in the food, mark ten of the regular foragers with one color of paint, kill the others, and continue to kill and tally all unmarked bees for a couple of days. At that time, step-train other foragers to a station 500m upwind from the hive with a second distinctive odor marker in the food. Again, mark ten of them with a different color of paint and continue to kill and tally all unmarked arrivals for a couple of days.

Gather data for a few days on relative recruitment rates at the upwind and downwind stations. After ten bees have foraged for a day or two at the two 500m stations (both upwind and downwind from the hive), reduce the number of foragers at each station to five. Then switch the odor at the upwind station to the same odor as that used at the downwind station. Continue to capture, kill, and tally all unmarked arrivals at both upwind and downwind stations.

According to the dance language hypothesis (Wenner and Wells, 1990; pages 63 and 64), recruitment should remain high at both stations. According to the odor-search hypothesis (e.g., Wenner, et al., 1991), recruitment at the downwind site (from our experience, likely already at a very low level) should plummet to nothing or next to nothing. That would be true, I feel, even if one would double the sugar concentration at the downwind station.

Finding a suitable location for those experiments might be difficult, though. In the Santa Barbara area and in the ridge and canyon systems of the West, daily wind patterns are very predictable in the summer season. However, areas where storm systems move through all summer (with concomitant shifts in wind direction) could pose problems.


Von Frisch generated his dance language hypothesis more than a half century ago to replace an odor-search hypothesis he had embraced earlier. Unfortunately, the vast majority of bee researchers and beekeepers apparently still remain locked into one of two notions: 1) Von Frisch “discovered the language” of bees, or 2) Von Frisch “proved” that bees have a language. However, hypotheses in science remain just that, conclusions; they remain valid only as long as they provide the basis for reliable predictions about what happens in Nature.

A third of a century ago we executed far more scientifically rigorous experiments (double controlled and strong inference – see Chapters 9 and 10 in Wenner and Wells, 1990) than those conducted by von Frisch. In doing so, we obtained results not consistent with his hypothesis. Language proponents (e.g., Gould, 1975) promptly rationalized away the results of our experiments (the first real tests of the von Frisch hypothesis) without repeating our experiments.

Instead, apologists again resorted to the use of less well-controlled experiments than those we had used and gained additional supportive evidence for the hypothesis they embraced. Gaining supportive evidence for the von Frisch hypothesis is easy, though – one can either consciously or unwittingly provide an odor cue at the location supposedly indicated in the dance maneuver, one that searching bees can exploit in their search. Or, one can embrace results that support existing dogma and reject results not in agreement with theory (e.g., Wenner, 1997).

Friesen published the results of his experiments (some of them summarized in the first two parts of this series) a quarter century ago. Although one can sometimes find his contribution mentioned, I have yet to find any bee researcher address the significance of his results. Nor have language advocates included our own experimental results in their reviews of the subject. That means that we have now had a half century of bee research that has essentially ignored wind direction and its role in honey bee recruitment (as covered in Friesen’s studies). Some might consider that circumstance an appalling situation, especially in light of the potential importance of such research in pollination studies.

Will bee language proponents do either of the simple experiments described above? I doubt it (as indicated, Mark Twain recognized that type of behavioral block in scientists a century ago). Any mentor who sponsored such experiments, however, could provide an excellent and refreshing opportunity for young bright students to test scientific ideas and to make meaningful contributions. As a first step, though, such students should be advised to avoid use of terms such as “the language of bees” or “their language” (phrases of interpretation or commitment), but instead concentrate on a study of the relevance of the “dance maneuver” or the “waggle dance” of bees (description or fact).

In the meantime, as long as bee language apologists continue to focus on positive results and dismiss negative results, we will remain at an impasse with respect to progress in honey bee foraging research. Scientists in other fields, though, view the continuing controversy as an exciting example of science in action – though somewhat locked in slow motion (e.g., Wenner, 1997, 1998).


Bogdany, F.J. and S. Taber. 1979. The significance of odor for bees orienting across a canyon. Apidologie. 10(1):55-62.

Friesen, L.J. 1973. The search dynamics of recruited honey bees. Apis mellifera ligusrica Spinola. Biological Bulletin. 144:107-131.

Gould, J.L. 1975. Honey Bee Communication: The Dance-Language Controversy. Ph.D. dissertation, Rockefeller University. NY.

Mangum, W.A. 1998. Honey Bee Biology: The hunted. American Bee Journal. 138(6):441-444.

Southwick, E.E. 1992. Bee Research Digest: Foraging, recruitment and search behavior of honey bees. American Bee Journal. 132(10):641,642,644.

Twain, M. (Essay of unknown date) 1963. Dr. Loeb’s incredible discovery. Page 590 in Neider, C., (ed.) The Compete Essays of Mark Twain. Doubleday, New York.

Wells, P.H. and A.M. Wenner. 1974. How recruited bees find food. Gleanings in Bee Culture. 102:110,111,127.

Wenner, A.M. 1961. A method of training bees to visit a feeding station. Bee World. 42:8-11.

Wenner, A.M. 1993 [with K. von Frisch]. The language of bees. Bee World. 74:90-98.

Weimer, A.M. 1997. The role of controversy in animal behavior. Pp. 3-37 in Greenberg, C. and E. Tobach (eds.). Comparative Psychology of Invertebrates:The Field and Laboratory Study of Insect Behavior. Garland Publishing. New York.).

Wenner, A.M. 1998. Honey bee “dance language” controversy. Pp. 859-872 in Greenberg, C. and M. Hara, (eds.), Handbook of Comparative Psychology. Garland Publishing, New York.

Wenner, A.M., J.E. Alcock, and D.E. Meade. 1992. Efficient hunting of feral colonies. Bee Science 2:64-70.

Wenner, A.M., D. Meade, and L.J. Friesen. 1991. Recruitment, search behavior, and flight ranges of honey bees. American Bee Journal. 31(6):768-782.

Wenner, A.M. and R.W. Thorp. 1994. Removal of feral honey bee (Apis mellifera) colonies from Santa Cruz Island. Pp. 513-522 in: Halverson, W.L. and G.J. Meander (eds.), Fourth California Islands Symposium: Update on the status of resources. Santa Barbara Museum of Natural History. Santa Barbara, CA.

Wenner, A.M., R.W. Thorp, and J.F. Barthell. 1999 (submitted). Removal of European honey bees from the Santa Cruz Island ecosystem. Fifth California Islands Symposium: Update on the status of resources. MBC Applied Environmental Sciences, Costa Mesa, CA.

Wenner, A.M. and P.H. Wells. 1990. Anatomy of a Controversy: The Question of a “Language” Among Bees. Columbia University Press.

* Reprinted from Volume 138, No. 10, October, 1998 American Bee Journal