The Honey Bee Dance Language Controversy: The Search for “Truth” vs. The Search for Useful Information

[1987 Wenner, A.M. and P.H. Wells. The honey bee dance language controversy: The search for "truth' vs. the search for useful information. Am. Bee J. 127:130,131.]

Department of Biological Sciences
University of California, Santa Barbara
Santa Barbara, California 93106 – (805) 961-2838

Department of Biology
Occidental College
Los Angeles, California 90041 – (213) 259-2676

It was while attempting to build a “model” dancing bee, by which we hoped to be able to direct bees out to crops, that we first obtained experimental results which cast doubt on the bee language hypothesis.

We read with great interest the articles, “The dance language of the honey bee: The controversy and its resolution” (Gene Robinson – March, 1986), as well as an earlier article, “A Computer Program that Teaches Dance Language” (Dewey M. Caron – January, 1986).

Whereas readers of the AMERICAN BEE JOURNAL may be quite familiar with the usual bee language story, as treated in those articles, they probably do not recall an April 1971 article in the AMERICAN BEE JOURNAL which outlined our case on why we feel the use of odor is sufficient to explain how new bees are recruited to food sources. The experiments described there have never been refuted, despite claims to the contrary. We provide here additional information about this most interesting problem of communication and recruitment among honey bees and the alleged “resolution” of the controversy.

Some of us in science do not agree that science is all cut and dried, as others would like to believe. Frederick Schram, a paleontologist at the San Diego Natural History Museum, recognized that fact in a little book, The Myth of Science or the Fantasy of Truth (Schram, 1979). In that book, he wrote:

. . . in the scientific method the outcome of an “experiment” must be evaluated. The common misconception is that one sets out to prove a hypothesis . . . (but) one can only disprove a scientific statement . . . If some results of an experiment tend to support a deduction, a hypothesis is safe only temporarily. Schram then listed three options available to a scientist when confronted by negative results.

He continued:

Besides rejecting or modifying a hypothesis in fight of experimental results, there is a third alternative. That is to simply ignore the results and stick with the hypothesis. This may sound silly, but this third alternative is used more often than one might imagine.

We found that proponents of the bee language hypothesis have consistently ignored important aspects of the problem. The following 3 points are examples of that selective viewpoint:

1) Our experiments yielded results which cannot be reconciled with the language hypothesis. For instance, when bees from two hives in different locations visited a single array of dishes, recruit arrival patterns from those hives were identical to one another, despite the supposed difference of “language” information provided in their hives by foragers (Johnson, 1967; Wenner, 1967). Later, when experiments were ran, in which new recruits could use either odor or “language” information, the results indicated that they had ignored any language information they might have obtained (Wenner, et al., 1969). We also found that disoriented dances were as equally effective as oriented dances in recruiting bees to a scented food source (Wells and Wenner, 1973).

2) Proponents of the language hypothesis have also ignored their own negative (“unfavorable”) results. In an often cited experiment, 87% of marked bees observed to follow dances and then leave their hive failed to arrive at any feeding dish. Of those that did find a dish, one-third were 240 m distant from their predicted goal; those that did arrive at the predicted site required more than 30 times the 25 seconds necessary for a direct flight (Gould, et al., 1970).

3) R. Rosin (1978, 1980a, 1980b, 1984) published several criticisms of the approach used by proponents of the language hypothesis because of the lack of adequate controls in all experiments run by those proponents. Rosin (1978) also argued that, if we have a choice between an “insect level” explanation (odor) and a “human level” explanation (“language”) for what we observe, we in science should accept the more simple interpretation. Furthermore, Rosin (1980a, 1980b) noticed that Gould’s claim (Gould, 1975) to have finally resolved the issue in favor of “language” is actually refuted by the results of his own earlier experiments (Gould, et al., 1970).

It was while attempting to build a “model” dancing bee, by which we hoped to be able to direct bees out to crops (an attempt to gain “useful” information) that we first obtained experimental results which cast doubt on the bee language hypothesis. We had to abandon our previous support and felt compelled to challenge that hypotheseis, starting in 1967.

We were able to present our views in scientific journals for a brief period of time; the traditional scientific process of free inquiry and expression then broke down (as it fairly often does in science). Suddenly we found that others were free to challenge us in print, but we were permitted no opportunity to reply to such challenges.

Our inability to publish our views denied beekeepers (among others) the opportunity to consider those alternative results and (we think) to benefit from the practical applications we felt could arise from our experimentation. Schram (1979) addressed (p. 39) the fundamental problem which arises during such restriction of academic freedom:

Science never answers questions by constraining investigation, but the lesson always has to be repeatedly learned.

When we realized that we could not express our opinions or get additional experimental results into print without an undue amount of effort and/or time expenditure, we took a “leave of absence” from research on bee communication and recruitment and studied other aspects of biology.

In the meantime, we feel that it should have become obvious to beekeepers by now that they have not benefited from the bee language hypothesis during its 4 decades of existence. This point deserves expansion. In his book Schram (1979) also wrote (p. 8):

“In all these cases we are differentiating in a basic way between exploitable knowledge for some practical use, and stepping-stone knowledge of a somewhat impractical nature . . .”

In physics, genetics, molecular biology, or medicine, research which does not result in “useful” knowledge soon becomes suspect. We provide here an example to illustrate that point.

In 1750 Needham published the results of a seemingly irrefutable experiment which indicated that life had arisen from a nutrient broth without prior innoculation of any organisms (Lawson, 1974). Spallanzani challenged Needham by conducting more rigorous experiments which yielded results negating Needham’s claims. Needham then claimed that Spallanzani’s techniques were faulty and that Spallanzani had “weakened” and damaged “a vegetative force” in the broth while conducting his experiments.

Decades passed before the matter was studied by Louis Pasteur. Pasteur conceived of an experimental design which met the objections raised by Needham and which produced results similar to those obtained by Spallanzani. However, even then there was great resistance by some scientists. Many of them still endorsed the notion of spontaneous generation, despite the compelling evidence obtained by Pasteur. Although one can still find fault with (or alternative explanations to) the experiments run by Pasteur, we now have “pasteurized” milk, aseptic hospital techniques, and sterile surgery, tributes to the “usefulness” of the knowledge obtained by this type of experimentation.

After living with the bee language hypothesis for 4 decades, however, we still do not have a “dance language management system” for beekeepers (or any equivalent practical application). Surely beekeepers should have benefited from the “language” research by now if it had merit, considering the rapid progress of science in these times and the large amount of grant support which has been allotted to this problem. Our inability to build a “model bee,” by which we had hoped to direct bees to crops, led us to really test the language hypothesis (the first and only such real test). The hypothesis failed those tests.

Now that a sufficient period of time has elapsed, we are writing a book about our experiences during the bee language controversy, under the premise that the controversy has not been “resolved” at all. However, our book is not on whether bees have a “language.” Rather, our book concerns the history, philosophy, sociology, and polities of biology.

A central theme of that book will be a presentation of the difference in scientific approach by: 1) those who believe that “truth” is absolute and can be proved or discovered, as against 2) those who feel that science at best yields “useful” information. Biologists will then have the opportunity to study history in the making, rather than relying on sketchy accounts of earlier controversies such as the one involving Needham, Spallanzani, and Pasteur. And beekeepers should be pleased that this insect they often marvel at has once again been so useful in aiding the progress of science.


Caron, Dewey M. 1986. A computer program that teaches dance language. American Bee Journal. 126: 39-41.
Gould, James L. 1975. Honey bee recruitment: The dance language controversy. Science. 189: 685-693.
Gould, J. L.; Henerey, M.; and MacLeod, M. C. 1970. “Communication of Direction by the Honey Bee.” Science 169: 544-64.
Johnson, Dennis L. 1967. Honeybees: Do they use the direction information contained in their dance maneuver? Science. 155: 844-847.
Lawson, Anton E. 1974. “Human Traits vs. Crucial Experiments.” The American Biology Teacher. September. pp. 334-336, 348. (A concise summary of the Needham, Spallanzani, Pasteur controversy.)
Robinson, Gene E. 1985. “The Dance Language of the Honey Bee: The Controversy and Its Resolution.” American Bee Journal. 126: 184-189.
Rosin, R. 1978. “The Honey Bee ‘Language’ Controversy.” J. Theor. Biol. 72: 589-602.
Rosin, R. 1980a. “Paradoxes of the Honeybee ‘Dance Language’ Hypothesis.” J. Theor. Biol. 84: 775-800.
Rosin, R. 1980b. “The Honey-bee ‘Dance Language’ Hypothesis and the Foundations of Biology and Behavior.” J. Theor. Biol. 87: 457-481.
Rosin, R. 1984. “Further Analysis of the Honey Bee ‘Dance Language’ Controversy. I. Presumed Proofs for the ‘Dance Language’ Hypothesis, by Soviet Scientists.” J. Theor. Biol. 107: 417-442.
Schram, Frederick R. 1979. The Myth of Science or The Fantasy of Truth. Vantage Press, N.Y.
Wells, Patrick H. and Adrian M. Wenner, 1973. “Do Bees Have a Language?” Nature. 241: 171-174.
Wenner, Adrian M. 1967. Honeybees: Do they use the direction information contained in their dance maneuver? Science. 155: 847-849.
Wenner, A. M.; Wells, P. H.; and Johnson, D. L. 1969. “Honeybee Recruitment to Food Sources: Olfaction or Language?” Science. 164: 84-86.
Wenner, A. M. 1971. “Animal Communications.” American Bee Journal. 111: 140-141.

Reprinted from February, 1987, American Bee Journal
Vol. 127 (2): 130 – 131