Yet another major controversy has hit the troubled world of
"cereology" with the publication of the H-Glaze
Report by Dr W.C. Levengood and his co-worker John A. Burke,
in the United States. Readers will already know from lengthy
articles in The Cerealogist and The Circular about
the controversial work being conducted by Dr W.C. Levengood and
John A. Burke at Pinelandia Biophysical Laboratories (an
impressive sounding name, but in fact merely a laboratory
attached to Dr Levengood's private address). Over the past few
years a number of "Lab Reports" have been issued
proclaiming the latest discoveries by these researchers. As
someone with postgraduate training in experimental design methods
I was naturally interested in what Levengood and Burke have been
up to !
In Lab Report No 18 Levengood and Burke describe what
they call a "Technique for Examining Crop Circle
Energetics". Readers will recall that one of the major
criticisms made against the crop circle researchers by
sociologists in the Equinox documentary was this vague use
of that term "energy". So far my attempts to find out
what kinds of "energetics" are being analysed by
Levengood and Burke have met with failure. In the meantime it is
perhaps safe to say that as a professional statistician I found
their description of their methodology confusing and disquieting.
Levengood and Burke claim that they have developed two
verification methods that are capable of distinguishing
"genuine" crop circles from fakes. These two tests are
the amplitude coefficient (also referred to as the "alpha
test") and the use of seedling development rates (ie growth
rates). These tests have apparently indicated that
"something is altering the rate at which ions flow through
the affected crop". Levengood and Burke state that they have
established that trampling cannot produce the statistical results
they are discovering in "genuine" circles because they
have compared their test results with results produced by
provably man-made circles. Strangely, this finding didn't stop
them from promoting crop taken from Jim Schnabel's Dharmic Wheel
as genuine products of the rapid heat-inducing circle-making
mechanism.
In this reviewer's opinion, there are many problems with the
claims made in Lab Report No 18. To begin with, Levengood
and Burke appear to confuse the terms "sample" and
"population". Also, they appear to have exaggerated the
importance of the results they have obtained. Quoting chances of
"less than one in a million" for their test results
Levengood and Burke do not appear to appreciate that it is
inappropriate to calculate binomial probabilities when ratio data
is available.
Reading through Lab Report No 18 I must admit that I
have found it difficult to understand how these two researchers
have analysed their data. They claim that:
"Each sample run involves five alpha values per trace. The current procedure involves six replicate tests on individual bracts (each selected from a different plant if available). Controls and crop circle samples are ran [sic] in alternate tests. The 30 data points (alphas) are entered into a computer program ("Statview") which provides a convenient means of statistically analysing many aspects of the data population. The most reliable, consistent information from the thirty alpha values is based on a statistical analysis of the paired, thirty data point alpha populations."
Again Burke and Levengood use the term "population"
when they mean "sample". I have read this statement
over and over again, and I still don't understand how one can
apply a "paired" analysis of "six replicate
tests" on each plant. A "paired" analysis involves
comparing two values, not six !
The correct method of analysing the kind of data discussed in Lab
Report No 18 is to conduct a two-way analysis of variance. In
this way one can test whether or not there are statistically
significant differences between samples of crop taken inside the
formation and samples of crop taken from outside fformations,
taking into account the natural variations in the alpha values of
samples in both groups. Such an analysis would only be
representative of crop circles in general if the samples taken
were truly independent of eachother within each formation and if
these tests were repeated in numerous formations chosen at random
across the world. Unfortunately Lab Report No 18 examines
samples taken from just one formation, the 1993 ringed circle in
oats at Albertsville in Canada.
Unfortunately, by taking "six replicate tests" on
the same plants it is debatable whether or not these researchers
have collected a truly random sample. For this reason not only
have Levengood and Burke conducted the wrong statistical test but
they may well have invalidated any results they obtain because
they failed to satisfy one of the primary assumptions underlying
almost every statistical test ever conducted !
However, the greatest problem with Lab Report No 18 is
Levengood and Burkes' curious decision to alter PROX-10 from a
control reading into a circle reading in their Figure 4
(approximately reproduced in Figure 1 on page 6). This decision
cannot possibly be justified because it completely alters the
outcome of the results of the alpha test !
In the top half of Figure 1 we have reproduced Levengood and
Burkes' results by drawing the average alpha value for each
sample. Levengood and Burke have drawn a line through the highest
control average (Cont-7) to emphasise how all the average alpha
values taken inside the circle and ring are higher. However, this
decision ignores the fact that PROX-10 - a sample taken in
unaffected crop close to the formation - produces an average
alpha value which is higher than four of the six circle and ring
samples!
If we redraw Figure 1 by correctly treating PROX-10 as a
control sample (rather than a sample taken from inside the
formation), then the true test result becomes clear. There is
little evidence that the average alpha values are significantly
higher inside the circles and rings than in surrounding,
untouched crop. In other words, the alpha test provides no
evidence of unusual effects. This decision to alter PROX-10 from
a control sample to a "circle" sample is scientifically
dishonest, for it alters the whole outcome of the experiment. It
is true that two of the alpha values are higher than the PROX-10
average (CIR-1 and RING-6), but in this reviewer's opinion it
must surely be expected that a two-way analysis of variance will
demonstrate that there is no statistically significant difference
between the average alpha values found inside the formation and
those found in the surrounding crop. I say this because it is
clear that there are wide variations between the average alpha
values in both groups (e.g. the average alpha readings in the
control samples vary from approx. 0.022 and 0.075, whilst the
average alpha values of samples taken within the circle vary from
0.038 to 0.090).
Unfortunately Levengood and Burke have failed to publish the
data they used in Lab Report No 18 so I cannot test this
conclusion properly.
One of the problems with this research is that it is apparent
from their own published work that Levengood and Burke seem
blithely unaware of the true extent of hoaxing in Britain.
Levengood's recent promotion of Jim Schnabel's Dharmic Wheel
formation seems an excellent example of the way in which the crop
circle myth continues to flourish because of the mass suppression
of pro-hoax evidence by leading cerealogists. Of course Levengood
and Burke claim to be searching for an infallible method of
distinguishing real from fake - something we would all love to
see - but this is no excuse for not having done their homework on
recent events.
In July 1994 an even bigger controversy broke with the
publication of what has been called the H-Glaze Report.
The author, John A. Burke, begins by claiming that he and
Levengood have made an "extraordinary discovery"
following their analysis of some reddish-brown glazed chalk found
by Peter Sorensen in two formations that lay close to the 1993
Cherhill pictogram.
Sorensen would have preferred to examine these circles
immediately but - unfortunately - Busty Taylor had to return home
that evening for an appointment. Sorensen returned to the site
two days later, accompanied by a neighbour. According to an
amicable farmer the circles had arrived a week or so earlier and
that originally parts of the circles had been covered by "a
dark grey mist" which had been largely washed away by heavy
rain. When Sorensen arrived both formations had been harvested.
The first formation was shaped like a tear-drop (in fact like a
"Nautilus") and exhibited multiple swirls and complex
layering effects. Sorensen noted that the dust was concentrated
inside the swirls and resembled soot. As he videoed the formation
Sorensen largely dismissed the possibility of a prank because the
dust appeared "almost accidental". However, as he
looked more closely Sorensen discovered a "reddish-brown,
dull glaze" on lumps of chalk and pebbles. A smaller
concentration of dust and coated chalk was discovered in the
second formation, a circle with an arc, which lay close by.
According to the H-Glaze Report, Levengood subjected
the glaze to a spectroscopic analysis. He discovered that the
particles were composed of iron and oxygen (FE). According to
Levengood's reasoning this didn't make any sense, because had
these originated from the soil there should have been traces of
calcium and silicon as well, but strangely there was none.
Microscopic study revealed that the glaze was composed of
"thousands of partially-fused tiny spheres" which
contained both magnetite (Fe O) and hermatite (Fe O). As the
particles were magnetized, the "glaze" acquired an
"H" - the chemical symbol for magnetism. Finding no
evidence of a "terrestrial system" that could account
for such unusual particles Levengood and Burke mounted an
"extensive" literature search to discover if such
material had been discovered before. Astonishingly they concluded
that the only way particles containing both iron and oxygen could
have appeared in a crop formation was if it had been deposited
during a meteor shower ! In their preliminary report Levengood
and Burke go into great detail about how the surface of a
meteorite would become molten as it enters the earth's
atmosphere. During this state the outer surface of the meteorite
is blown off and solidifies into tiny spheres that oxidise (rust)
and fall to earth. Somewhat conveniently this process is said to
take days or even weeks.
Levengood and Burke hypothesize that this dust was released
during an unusually intense Perseid meteor shower, which
apparently peaked nearly two weeks earlier. During their
microscopic examination of the particles they noticed
"mud-crack" patterns and bubbles where the molten
meteoritic droplets had partially refused. Attempting to explain
why the molten droplets had failed to burn the wheat Levengood
and Burke propose that the moisture inside the stems evaporated
and produced water droplets on the stems, thus insulating them
from the effect of the heat.
This "Leidenfrost effect" insulated the stems from
burning. Levengood and Burke were so excited by their discovery
that they quickly circulated the H-Glaze Report to
numerous sources, urging cereologists to "make magnets a
standard part of their field equipment" to locate more
meteoric dust. Furthermore, the authors claim that "This
incident provides rare, direct evidence for a theoretical model
of crop formation - the plasma vortex - that had previously been
indicated only in an indirect way." They go on to cite
confirmation of their results by stating that the affected wheat
stems exhibited "dramatic differences" to control
samples in terms of the alpha test and measured growth rates. In
their conclusions Levengood and Burke grandly claim to represent
"the scientific community of the world" and they
challenge hoaxers to explain how they managed to "scavenge
the atmosphere for meteoric dust, re-heat it and lay it down just
right with no contamination". They predict that crop
formations will appear more frequently following meteor showers
than at any other time.
Well, if all the claims made by Levengood and Burke were
really supportable we would have a major breakthrough which would
make one giant conceptual leap in our understanding of the crop
circle phenomenon. However, as we have come to expect in this
business, the circlemakers were not about to let Levengood and
Burke get away with such an astonishing claim without some kind
of fightback - oh no !
On July 25th 1994 Robert Irving wrote to John Burke. Irving's
letter stated :
"It is not our primary interest to contradict your findings ... It is instead our intention to use your report as textual source material for an upcoming exhibition to be held on behalf of The Agency Gallery, in London. The piece in question (entitled 'Fe3') will comprise a museum style glass cabinet with text displayed on the glass. Inside the cabinet, beyond the text, will be a standard Oxford University chemistry laboratory bottle containing fine-grade iron filings. This bottle was originally addressed, labelled, and postmarked to correspond with the crop formation which constitutes the subject of your report ... and will be displayed in it's original state. Remaining samples of the 'grey dust' will also be shown. All text will be fully credited to you, citing the tests and conclusions of Dr W.C. Levengood. The context of the piece can be loosely summarised by the following theoretical equation: If science is incongruous to mysticism, and the mystical is represented through art, should 'bogus' science be elevated to an art form ? Certainly the gallery concerned seems to think so, and our fingers feeling the pulse of a growing trend towards millenialist awareness would seem to confirm this."
We have reproduced Irving's own photograph of the laboratory
bottle on page 8. This bottle was exhibited at a London Art
Gallery on the South Bank during September and the accompanying
text is reproduced on page. BBC2's The Late Show took an
interest in the iron fillings exhibit and they filmed an
interview with Irving during September [for proof, ring Matthew
Collings at the Beeb]. Meanwhile, a furious argument has
developed between Levengood and Burke, on the one hand, and
Irving and Montague Keen, on the other.
Irving has sent samples of the original batch of iron filings
to Montague Keen and offered them to Levengood and Burke, who so
far have failed to accept this offer. Irving's intention is to
allow all three to compare these samples with the glaze
discovered in the Cherhill formations. Keen has very sensibly
suggested that these samples, and those found in the Cherhill
formations, be subjected to an independent test by a reputable
laboratory to establish whether or not they are one and the same
thing.
Tellingly, at the time of writing, Burke and Levengood have
yet to respond to this offer. Furthermore, both Burke and
Levengood have failed to supply full answers to a series of
detailed statistical questions I sent to them during late
September (letters available as usual).
It is perhaps not surprising that these researchers have
refused to be drawn into this affair any further considering
their promotion of "dramatic differences" between
Irving's iron filing-coated seeds and controls. Were they to do
so, and if Irving's claims are true, then the fallacy of the much
vaunted alpha tests would be exposed for all to see.
The H-Glaze Report is yet another amusing story in the
long-running crop circle hoax, another testament to the failure
of researchers to attain true objectivity in their work, and
another telling lesson to the power of the anomaly myth. No one
can doubt the sincerity of Levengood and Burke, and their
dedication to their work deserves praise. But this work is
fatally flawed for two primary reasons - the desperate desire to
find an anomalous explanation on the part of Levengood and Burke,
and their seeming naivety when it comes to understanding the true
extent of the hoax evidence and the mass cover-up of that
evidence by the believer groups these past few years. Oh well,
all's fair in love and war !
Dr Levengood has had an article published in Physiologia
Plantarum 92 - a properly refereed scientific journal of the
kind that even the Wessex Skeptics presumably take seriously.
This article again promotes the alpha test and enhanced growth
rates as measures of how to verify "genuine" crop
circles. A full article will appear in our next issue discussing
this astonishing development.
