Julian O'Dea asked for comments about his draft paper

> The "dance" of the honeybee (Apis mellifera L.) is one
> of the most famous of all animal behaviours...
> An alternative hypothesis is that locality odour alone, not
> dance movements, is the basis of the communication of
> the whereabouts of resources.

While odor and taste certainly are possible components
of overall bee forager-to-recruit communication, how can a
bee on the dance floor discriminate the locality odor from
a specific returning forager, given that there are:

a) many other returning foragers present, each which, under
    the theory put forth, has their own unique locality odor?

b) many house bees that offload returning foragers present,
    each presumed to be smelling of whatever they last handled?

c) some of each type of bee is often closer to the recruited
    bee than the forager/dancer?

d) What about the often powerful scent of evaporating nectar
    flooding the hive as a whole?  How does a recruit bee "ignore"
    this odor, but detect the relatively faint locality odor from a
    single returning forager?  What if the two are very similar odors?

As an analogy, (a) through (c) above reminds me of a crowded
cocktail party, with dozens of women, each wearing a different
perfume.  To make matters more complex, in (d) the host is
burning incense, creating a strong "perfume" that permeates the room.
I can attest to the fact that my nose gets quickly overloaded in such
surroundings.  Wouldn't a far more sensitive sense of smell, like that
of a bee, simply be overwhelmed more quickly in such a setting?

What explanation could be offered as to how bees could
follow "odor plumes" when flying >>with<< the wind from hive
to foraging locations?  Even when flying into the wind, how
can a bee detect the odor of a single feeding station at the
distances that "dance" has been tested with feeding stations?

Doesn't "locality odor" seems to require:

a) That a unique odor or blend of odors exists for specific
    sites, and that such odors are unique enough to allow
    one to distinguish between one site and another from afar?

b) A very large set of unique odors to identify?  Given that
    bees have been found to forage as far as 10 to 15 miles
    from their hives (this may be an extreme, and the "world
    record" distance), how many unique odors must exist to
    describe an area as large as 138 square miles at a
    resolution that is useful to bees?

c) That such mixes of odors are somehow able to travel
    on the breezes, and arrive intact at points distant from
    the source?

d) Given (c) how would odors, made up of volatile chemicals,
    not evaporate away to "nothing" long before traveling very far?

e)  Given (c) and (d), how might such odors travel upwind?

Here's a real-world application of (a) through (e) above:

1)  I have many wild rose bushes growing along fencelines that
    surround a 6-acre, a 15-acre, and a 500 acre field, all heavily
    seeded with a mix of two types of clover.  The bees clearly
    work the clovers at one time of day, and the wild roses at
    another. What would be the difference in locality odors
    between the three adjacent fields?

2)  How would the odors be different, given that the 3 fields
    contain exactly the same plants?  (All three are cut for hay,
    and are managed identically.)

3)  The local terrain is such that the sun hits the 500-acre field
    first, the 6 acres 2nd, and the 15 acres 3rd, and in the
    afternoon, the 15 acre field is shadowed long before the others.
    Given that nectar availability varies between fields with time,
    doesn't this imply that bees must communicate which area is
    good forage >>>now<<< ?  How might they do this with odors
    from identical mixes of plants?

4)  More to the point, given that the roses do not provide nectar at
    the same temperatures/time of day as the clover, what would
    be the locality odor that would differentiate between the
    different >>>perimeter<<< areas, where the roses are, as
    opposed to the interior areas, where the clover is?

If "locality odour alone, not dance movements, is the
basis of the communication", then why is dance:

a) So consistently complex?

b) So consistently done by bees upon return from good forage?

c) So easily translated into distance and direction vectors by
    observers?

d) So highly correlated to where marked recruited bees go, when
    they can be proven to have never been to that forage area before?

> Esch et al. (4) concluded that bees actually measure distance
> from the amount of "optic flow" on their trip, that is "the total
> amount of image motion en route to the food source".

Since bees fly both with and against the wind, they must use
"dead reckoning" to estimate distance and navigate.  Anyone
who sails knows the massive limitations of dead reckoning, but
a true sailor needs no compass.  He needs only the sun and the
stars.  Should bees need more?

> But, at the same time, it implies that honeybees must be poor
> at measuring distance in an absolute sense.

Isn't everyone is "poor" at measuring distance in ANY sense?
Isn't "optic flow" a very tidy explanation for how "distance" is
measured so that it may be communicated to bees that may
fly at different speeds, and/or during "gusty" wind conditions?

> If Esch et al. (4) are correct, their findings raise problems for
> the "dance language" hypothesis.  As they note themselves,
> the apparent lack of absolute accuracy in the information about
> distance - supposedly conveyed in the dance - is a problem for
> the proposed communicative mechanism.

How do inexact measurement of distances raise problems for
"dance"?  How does less than perfect accuracy undermine the
high degree of correlation between dance, the human interpretation
of the dance, and the verifiable actions of recruited bees after the dance?

> The inaccuracy in the supposed means of communication makes
> it less likely that the dance movements of honeybees have anything
> to do with communication of the whereabouts of resources. The
> alternative, locality odour hypothesis (1, 6) therefore gains in credibility.

Isn't the phase "lack of absolute accuracy" a tacit admission
that the distance information is at least "useful", if not perfect?

How does any distance inaccuracy in "dance" add any credibility
to the "locality odor" theory?  Mustn't each theory stand on its
own merits and be evaluated independently? Even if "dance" were
somehow proven completely wrong, how would "odor" gain any
credibility as a result?

Wouldn't the inaccuracy have to be enough to reduce the correlation
between "dances" and recruited bee actions to "no better than chance"
in order to cause a reasonable person to question the validity of "dance"
as a coherent, consistent, and testable explanation of recruitment?

What is the testable connection between "odor" and "location" that
would reveal a location from odor alone?

Wouldn't the first step towards proving that "locality odor" exists require
multiple air samples from both the hive area and forage areas, a high
performance liquid chromatograph, and the patience to sort through
"parts-per-billion", if not "parts-per-trillion" analysis of both the air
samples and samples collected from returning foragers?

> Dancing movements and sounds emitted by bees returning from
> foraging trips may only serve to attract the attention of hivemates
> so that they can be made aware of the odours associated with
> desirable food sites

Why then is dancing alone enough to consistently direct humans
to the forage site, and allow them to verify that marked bees, seen
to "witness" the dance, are foraging in that area?

I'd submit that most of the "inaccuracy" is introduced by the lack of
state-of-the art hardware available to researchers who work with bees.
Below are my suggestions for anyone who might want to gather dance
data that would be more accurate, and less prone to inaccuracy:

1) Lasers and Electric Guitar Pickups

  If someone wants to eliminate the subjective component of "measuring"
  dance, they should get a laser, attach a reflector to the comb on which
  the dances are done, and record the vibrational component of the dance
  as propagated on the comb.  This would result in cleaner, more accurate
  data, and data that is both easier to gather and easier to analyze, and
  eliminates the possible accusations of "subjective interpretation".

  The vibrational component of bee dances is clearly important for reasons
  that should be clear to any beekeeper.  The combs that act as "dance
  floors" are nearly always chewed away from the bottom bar of the frame,
  starting at the edges of the frame, and working towards the center.
  The result of this isolation of the comb from the frame reduces the
  "damping" effect of the frame on vibrations, and results in better propagation
  of vibrations across the comb.

  This is easy to test at home - if you can't afford a laser, get a magnet and
  an electric guitar pickup, and measure for yourself. (Imbed the magnet in the
  comb, and use the pickup to measure magnetic field strength, and hence,
  vibrational motion.  Use an LM386 amplifier chip running on a 9-volt battery
  from Radio Shack to amplify the signal, and plug it into a PC's sound card
  microphone jack. Use the free "demo" version of CoolEdit 2000 to gather
  the data and look at the signals in terms of frequency and amplitude.)

For those who do not read research papers in bed at night as if they
were mystery novels, a fellow named Sandeman tore the legs off some
bees in 1996, thereby grossing out nearly everyone, and somehow
attached them to tiny electrical wires to measure exactly what ranges
of vibrations they could detect.  Bee legs are not good at detecting
higher-frequency vibrations, but are very good at picking up low
frequency vibrations that just happen to be the same range of
frequencies are produced by dances.  While he did pull off bee legs,
he was not pulling our legs.  The data he generated is very complete,
and very consistent.

2)  Infra-Red Cameras

  The "direction" vector is best recorded on video, but this is old hat.
  High-speed video does a better job, but this is also old hat.
  Infra-red cameras are now cheap, so there is no need to disturb
  the bees with lights.

3)  Waveform Analysis

  The audible sounds created during dance have been often analyzed with
  klunky old "sonographs".  The problem with sonographs is that they
  introduce artifacts, in the form of apparently skewed amplitudes for
  harmonics of the fundamental frequency.  Junk the sonograph gear, and
  start using Fast Fourier Transforms and spectrum analyzers.  One can
  buy low-cost off-the-shelf software that runs on any PC to do both.
  While you are at it, forget the silly harmonics, and pay attention to
  the fundamental frequencies.  When the harmonics are between 8 dB
  and 30 dB lower than the fundamental, one should recall that for
  every 3 dB down, the loudness (power) is halved.

  One must also recall that bees have never even been seen to react
  consistently to airborne sound frequencies above about 700 Hz
  (see Kirchner, et al  in J Comp Physiol A, 1991 168:85-89), so the
  harmonics are apparently not even noticed by the bees.

Again, for those who don't read papers, it should be stressed that,
rumors to the contrary, bees CAN hear airborne sounds, and that
their "ears" have been located.  In 1993, two fellows named Kirchner
and Dreller published a paper in "J Comp Physiol A" that likely even
grossed out Mr. Sandeman.

They cut off antenna from live bees, and they even shaved the hairs
off bees (presumably with tiny little razors) to determine how bees
detect airborne sound.  They found that an organ at the base of the
antenna called the "Johnston's Organ" is what detects the very tiny
vibrations caused by airborne sound waves as picked up by the antenna.

        jim