>>>> Australia got a bum rap. <<<<
Below is an excerpt from the multi-page
article that appears in October's
Bee Culture, on the Bee Culture website,
and in a serialized form on the Hearst
Newspaper website "The Daily Green".
My hope in posting the excerpt here is to
stop any possible fistfights from breaking out
at Apimondia before they start. Ironically,
Apimondia is being held this year in Australia.
Now, in fact.
I've read the full paper. I've read multiple
versions of the paper as it slogged its way
through review, a process that took all summer.
I've also participated in the press conference,
where the authors presented a VERY different
view of their results than one might gain from
merely reading the full paper.
The short version of all this is:
a) The Mainstream Press, As Usual, Got It All Wrong.
b) The Conclusions Drawn In The Paper Are Not
Supported By The Actual Work Described in
The Paper, The Supplement, and The Press
Conference Audio Transcript
c) As I Said, Australia Got A Bum Rap.
d) This paper may be long remembered as beekeeping's
"cold fusion", even overshadowing the 2004
claims made by USDA researchers that a fungus
had been found that killed 100% of varroa in hives.
(Remember "Metarhizium anisopliae"?)
Don't get me wrong here - I STILL want to see all imported
bees sampled, tested, and inspected, as I have been saying
since 2002. Betcha the Aussies now agree that this should
be done, if for no other reason than to protect their
reputation on the world market for bees and honey.
But a "ban" on bee imports might be counter-productive.
We may NEED bees from Oz (and NZ) if CCD gets out of
hand. Where else might we get replacement packages
EXCEPT from overseas in the scenario where we start
taking casualties among the US queen and package
producers?
There may even be bans on Australian honey as a result
of the heavy-handed attempts to implicate a virus and
the Aussies in CCD. While the paper stopped short of
saying it outright, they did everything but put up
billboards accusing Australia, without even a shred
of solid evidence that the virus at issue has anything
to do with CCD. Read the analysis.
The first group to try to ban Australian honey did
not even wait for the ink to dry on the paper.
Shame on the NZ National Beekeepers Association!
Withdraw your call for a ban on honey from Oz!
Its not that NZ imports any significant amount of
honey from Oz, so its not about money or honey
prices or the usual agricultural protectionism.
And it is not about "biosecurity", since honey
from Oz would be waaay more expensive than HFCS
as bee feed, so the honey would not be fed to
bees and perhaps "transmit" the virus. Assuming
that NZ doesn't already have IAPV in their hives,
of course.
Its all about disparaging both Australian export
honey and export bees in an attempt to get OTHER
countries to ban honey from Oz, where NZ and Oz
compete for market share.
I humbly apologize on behalf of the US to every
beekeeper in Australia for the baseless accusations
and unjustified disparagement leveled in your
direction as a result of a research project where
the wheels came completely off. (This problem
seems to be related to the size of the wheels,
as big wheels are much more problematic than
little wheels in this regard.)
Bottom line, any analysis of samples that were not
first properly classified as to their "CCD" or
"Non-CCD" status is meaningless.
But don't trust me if I am telling you to not trust
a paper published by the journal "Science", read
the paper yourself. Fact-check my critique.
Here's the excerpt that I think you might want to read
if you haven't time to slog through everything else.
Beekeepers are uniquely qualified to understand the
exact nature of the fatal flaw described. The entire
paper, as a result, is much worse than merely "wrong",
it is a deliberate attempt to declare some sort of
"success" in diagnosing CCD using sampling methodology
that wouldn't get past a Junior-High School science
fair judge.
==============================================
Were Diseased Samples Consistently Classified?
==============================================
One basic step that can skew results is how one classifies the samples
as being from "diseased" or "disease-free" colonies. If samples are
labeled "diseased" when the specific colony is not diseased, the result
is a misleading sample, and the results are much less accurate. Given
that a total of only 30 "CCD Samples" were analyzed, a single
misclassified sample could make a difference of 3% in the "CCD Samples"
results.
The paper says:
"Diseased apiaries were selected based on evidence of recent collapse of
some colonies within the apiary and a lack of dead bees in collapsed
colonies. Up to three dead, collapsing, or stronger colonies were
selected for sample collection in each diseased apiary."
So, if only some of the colonies in an apiary had collapsed, samples
were collected from any colony in that apiary, as if all colonies in the
apiary were certain to be "diseased". This implies that some of the
samples called "CCD Samples" could actually be "Non-CCD Samples".
There's no way to tell, as there is no fool-proof test that confirms the
presence of CCD. The best that could have been done would have been to
only use samples collected from hives that showed obvious symptoms of
CCD. That wasn't done, so the results attributed to the "CCD Samples"
are questionable as to their accuracy.
With so few samples, there isn't a lot of room for any of them to be
misclassified. The assumption that all colonies in a yard will have CCD
when only some of the colonies are showing symptoms is clearly a guess.
The resulting data is dependent upon those guesses, and cannot be said
to be any better than those guesses.
Were Healthy Samples Consistently Classified?
=============================================
While it may seem simple to tell a healthy colony from a colony
suffering from CCD, it would be premature to call the samples taken from
such colonies "disease free" unless the colonies were inspected again at
a later time to verify that they had remained disease-free.
No one knows what a colony about to collapse from CCD looks like, but it
is reasonable to conclude that there aren't any obvious symptoms, or the
many highly trained and experienced people investigating would have
noticed something unusual. The paper did not mention this aspect of the
sample collection process, so we asked Dr. Jeff Pettis of the USDA
Beltsville Bee Lab about it during the press conference where the paper
was announced to the press. He answered "No, we didn't have the luxury
of going back to those same colonies."
At risk of sounding flippant, couldn't someone at least have called the
beekeepers who owned the hives, and asked "how are the hives that were
sampled doing right now"?
They didn't even do that, so it is impossible to know if the samples
classified as "non-CCD" were taken from hives that collapsed from CCD
soon after, and thus should have been classified as samples from "CCD"
colonies. Each misclassified non-CCD hive would result in a 4.7% error
in the "Non-CCD Samples" results.
Chance, Luck, Statistics
=========================
If you think that I'm being unreasonable here, understand that being
pedantic and picky about sample collection is central to being certain
about the results. Given the lack of certainty about the samples,
NOTHING can be said about the results with any certainty. The results
could only be accurate by chance! All sorts of statistical analysis was
done, but we can ignore it all, as the samples can only be properly
classified if everyone was very lucky.
I'm surprised that something this basic slipped by the reviewers. The
problems with sample classification invalidate the entire paper in my
view. But we won't abandon the slog here, as there are other problems
to consider in this paper... (continues)
Here is the full text of the "Science" paper (watch that line wrap):
http://lists.sonic.net/pipermail/pollinator/attachments/20070906/2a8e96c
8/attachment.pdf
Here are the details - the "supporting materials" for the paper:
http://www.sciencemag.org/cgi/content/full/1146498/DC1
Here is the press conference audio transcript:
http://bee-quick.com/reprints/ccd_aaas_news_conf.mp3
(WARNING: 7 megabyte mp3 file)
The full Bee Culture coverage and analysis consists of:
The introduction, on the "Daily Green" website:
http://www.thedailygreen.com/2007/09/06/bee-disease-study-offers-nothing
-that-is-both-new-and-compelling/6251/
This page will have links to more stuff, as it happens.
(As of 09/08/07, the links are not working)
The rest of the segments are here:
http://beeculture.com/content/CCD_Analysis.cfm
Which will also have links to further breaking news
"Practical Implications For Beekeepers"
http://beeculture.com/content/PRACTICAL%20IMPLICATIONS%20FOR%20BEEKEEPER
S.pdf
"A Beekeeper Reads The Paper"
http://beeculture.com/content/Reads_The_Paper.pdf
"World Trade"
http://beeculture.com/content/World%20Trade.pdf
ADDENDUM:
Maybe you got down to here, and don't see just how faulty the sampling
process was.
Let's replace "apiaries" with "apartment buildings", "hives" with
"humans", and
"CCD" with "sudden unexplained death of humans".
Now, let's summarize the sampling protocol:
"DISEASED" SAMPLES
a) Find a building where a few or many people have died from this
strange disease.
b) Take samples from anyone in the building, dead or alive, apparently
healthy or not,
long-term resident, or someone who just walked in the door to deliver
a pizza.
c) Don't be consistent in your approach, be "random" in your selections.
"HEALTHY" SAMPLES
a) Find other buildings where no one is sick yet, and take samples.
b) Don't bother to check back after a bit to see if they stayed healthy,
or started showing symptoms later. Don't even check to see if anyone
died in those buildings before you publish your paper.
c) Recall that the disease in question has no "early symptoms".
I hope that makes things a bit more clear. :)
LEGAL MUMBO-JUMBO FOR THE ANALYSIS AND COVERAGE:
Copyright 2007 James Fischer, All Rights Reserved.
First print publication rights granted to Bee Culture
magazine for 10/01/07 issue. Unlimited rights
are hereby granted for all non-commercial uses.
Commercial rights available. Contact author.
PLAIN ENGLISH TRANSLATION:
Giving away copies for free? - Go for it!
Charging or making money? - You can't run, you can't hide.
We will find you, and you won't like it one bit.
******************************************************
* Full guidelines for BEE-L posting are at: *
* http://www.honeybeeworld.com/bee-l/guidelines.htm *
******************************************************
|
