SPECIFIC POINTS II: What about the "robot bee" experiments?
Lars Hansen asked: 1) [According to Kirchner and Towne] "Novel
experiments, such as training bees to respond to sounds and recruiting them
using a robot, have ended several debates surrounding the dance language."
and 2) Could you, Adrian Wenner, please put your piece on the table and
sum up very shortly: What precisely was wrong with the experiments with
"robot bees" and/or the conclusions made?
********
Lars raised a VERY important point. Prior to the robot bee experiments,
all evidence in support of bee language had been circumstantial (INDIRECT
EVIDENCE), with a great deal of negative evidence continually accumulating
that counters that interpretation.
An artificial bee that could direct recruits out to a point source
(without the confounding element of odor) would provide DIRECT evidence for
the hypothesis.
*********
My co-workers and I already included an analysis of all the results from
the [1989] robot bee experiments in the following publication:
Wenner, A.M., D. Meade, and L. J. Friesen. 1991. Recruitment, search
behavior, and flight ranges of honey bees. AMERICAN ZOOLOGIST
31(6):768-782 (an invited review paper).
For illustration, I will cover here only the results of the distance
experiments (as shown in Fig. 3 of the original 1989 NATURWISSENSCHAFTEN
publication).
To obtain the real results (instead of as the percentages shown in their
figure) I measured the lengths of the bars in their graphs with a ruler and
converted the information into real numbers of bees. A summary of results
from that set of four experiments follows (you might want to draw out a
line and mark the distances along that line to keep the results straight):
1. ROBOT BEE DANCING FOR A 250m DISTANCE. In two 3-hour runs, a total of
27 and 38 recruits, respectively, arrived at the four stations, located at
100m, 250m, 500m, and 1000m from the hive. However, less than a third
(31%) of them had arrived at the "designated" 250m station. (Simply at
random, one would expect 25% at each of the four stations.)
2. ROBOT BEE DANCING FOR A 500m DISTANCE. In two 3-hour runs, a total of
24 and 101 recruits, respectively, arrived at those four stations.
However, only 13 (about 10%) of them arrived at the 500m target station.
3. ROBOT BEE DANCING FOR A 1000m DISTANCE. In two 3-hour runs, a total of
27 and 41 recruits arrived at the four stations, with only 2 bees (3%)
arriving at the 1000m target station.
4. TWO REAL BEES FORAGING AT 500M. With two live foragers flying between
hive and feeding place for 3 hours (all odor artifacts included), a total
of 369 recruits arrived at the four stations However, only 30% (109) of
them had arrived at the target station --- not the overwhelming majority as
predicted by theory.
After contacting the robot bee, searching recruits obviously did not
perform as if they had "flown directly out" to the indicated target.
More important, perhaps, is the fact that searching bees who contacted
the two live dancing bees also did not perform as predicted.
The researchers thus actually obtained results at odds with the language
hypothesis.
***********
Also one should note that the authors wrote: "During each experimental
session...the wax on the model was given an additional, faint scent ....
The same scent ... was added in minute quantities to the sugar water and to
a piece of filter paper in baits placed at various locations in the field."
Odor thus remained a confounding factor in the experimental design.
********
I have great faith in the integrity of Kirchner, Towne, Lindauer,
Anderson, and others. I am sure their results are valid --- particularly
since the results in the sum total of their efforts falls into a neat
mathematical distribution (random lognormal, as displayed in Fig. 4 of our
1991 AMERICAN ZOOLOGIST paper).
How, then, could they conclude that they had succeeded?
Maeterlinck (1901) encountered the same problem. He had conducted an
experiment that yielded results counter to the notion of "language" use
(see p. 43 of our 1990 book). However, a beekeeper friend repeated his
experiment and obtained four "favorable" results. Maeterlinck concluded,
"But I am convinced that my friend was misled by his desire, a very natural
one, to see the experiment succeed." (translation from a French edition of
THE LIFE OF THE BEE).
Adrian
Adrian M. Wenner (805) 963-8508 (home phone)
967 Garcia Road (805) 893-8062 (UCSB FAX)
Santa Barbara, CA 93106
********************************************************************
*
* "History teaches that having the whole world against you
* doesn't necessarily mean you will lose."
*
* Ashleigh Brilliant's Pot-Shot # 7521, used by permission
*
********************************************************************
|
