> mite fall counts can be used to trigger a particular management
strategy, 
> which is appropriate for them and their bees, in their locations at
that 
> point in their seasonal cycle. 

My objection is simply that they are advocating checking far too late, 
reducing their "management" to the level of triaging the doomed from 
the survivors rather than actually treating colonies that could be
"saved".
(Monty Python and The Holy Grail: "Bring Out Your Dead!")

If you dismiss the above as "too radical a view", I can offer the
less confrontational statement "We want hives that THRIVE, rather
than barely SURVIVE".  (Catchy, eh?)

I don't question that they think their approaches are acceptable.
But let's ask them - do they use the SAME threshold numbers?
Betcha they don't.  Let's hear how long they leave their sticky
boards in, and what they count, and how they count.  Betcha they 
don't follow the same protocol in getting their drop counts, either

a) 3 days and divide by 3, or one day and count the total?"
b) Count all mites, or ignore white (immature) mites?
c) Count during rain, when bees are confined, or count on foraging days?
d) Use the Great Lakes "checkerboard" card, or a plain card?

I could go on, but all the factors listed above clearly impact the 
numbers, so if one is going to toss a number around, one needs to
at least define some context for the number in terms of how it
is MEASURED.

So what's some OTHER guy to do?  Run his own little experiment to 
determine what "threshold" seems to be reasonable for him in
his location, with his bees?  Naw, he's gonna accept whatever
number he finds first, as bet his bees, his crop, and perhaps
his farm on a contextless scalar from someone with very different
conditions from his own.

I challenge you - do a Google search for "varroa threshold"
and just see how many different numbers are floating around
out there, each and every one presented as "authoritative".
Which one do you use?

Now, let's take a typical population graph for both bees and
varroa, like this one:

http://www.oie.int/eng/normes/mmanual/2.9.05-2.JPG
(Bee population is the solid line, mite population is the dotted line)

So, given that even queen producers are unable to control varroa
sufficiently to prevent the shipment of queens infected with
multiple viruses, what can we say about the "best practices"
of beekeeping as a whole?  We can at least say that a good
number of people have no idea WHERE THEIR HIVES ARE ON THAT
GRAPH.  Clearly some sort of graph like this would be of value
even if one only counted mites, and ignored bee population
levels.  Look at the ramp-up of mite population.  When do YOUR
varroa populations start to "go exponential"?    Well, that's
when you start to get multiple varroa per brood cell, which
is when the virus situation goes completely nuts.  You want
to knock back the varroa population BEFORE then, if you want
a hive that will thrive, rather than merely survive.

And what happens in a year with conditions more conducive to varroa
reproduction?  So much of beekeeping is tied to weather, and varies
from season to season.  Hives can "crash" from varroa infestation
before either one of them will even get around to looking at mite
drop counts.  

So, let's review:

1) A single measurement is a point.  It raises more questions than
   it resolves, as it lacks any context.

2) Two points can be connected by a line.  What's the SLOPE 
   of the line? Now that we have a slope, we can talk about 
   a rate of change!

3) Three points can reveal a "curve" of sorts.  Now we can
   talk about the rate of change in the slope, look at the
   area under the curve, do integrals, and know lots lots more
   about the trend we are tracking.

4) Measurement is inherently "sloppy", so any one point is
   not to be trusted.  But even with "sloppy" numbers, one
   can see a trend.

So, we can conclude from the above that 3 counts would be a much
better idea than one, and the marginal cost of doing 3 counts
is pure semi-skilled labor - something that can be done by someone 
who needs only approach hives in the evening, need not open hives, 
and can be completely unskilled in "beekeeping".  Sounds to me like
a great part-time job for a high-school student who might
otherwise be looking for lawns to cut.

> As I understand it, the aim is to reduce total mite burdens to non 
> critical levels that will ensure colony survival and productivity
> for the following parts of their seasonal cycle. 

This simply is not possible when their approach is to wait until
"September" (fall up here above the equator).  Mite levels are sure
to reach and exceed "critical levels" in some cases before 2nd or 
3rd pull of supers.

Waiting for Fall to do one's first mite count means that one has
let varroa have its way with your colonies all season, and one
is merely triaging "doomed", "treatable", and "very lucky" hives.

Any program that does not include taking hives (or entire yards)
out of production to treat them for excessive infestation is merely
a method of counting one's losses before winter begins.

> The management strategies... are again based on their relevant 
> experiences. So what's wrong with that? 

Well, let's take the statements made and apply them to a context 
that will make the level of fuzzy thinking more obvious:

Let's try the cliche of the "Cowboy and Indian" movie, with
the wagon train moving across territory that can only be 
described as "hostile".  They have bought into a policy of 
only worrying about Indians if they see 55 or more approaching.
But they don't even LOOK for Indians until they are nearly
done their journey, and ignore the different sizes of wagon
trains they lead, some small enough to be wiped out by a mere
dozen attackers. 

Does this sound like a fool-proof way to get the settlers
to their destinations with their scalps intact?

> Isn't that what we all do, all the time, in making management
decisions 
> for our livestocks' welfare and productivity? I hope I haven't missed 
> anything here.

What has been missed is the simple concept of making management
decisions based upon an understanding of:

- The life-cycle of the pest.
- The reality of less-than-perfect results from treatment.
- The ability of varroa infestation levels to get out of hand early.
- The inherent variability of one season from another

> it appears to me that you are railing against the concept of a "one
size 
> fits all" mite drop threshold, because anybody relying on such a
figure 
> which materialises out of the "research" ether with no regard to their

> bees location and their own management practices and objectives, is
doomed 
> to a future of catastrophic hive losses. 

What you describe as "catastrophic losses" are being called "normal"
over here on this side of the planet.  That's the problem.  People are
accepting losses like 20% as "normal".  Using "thresholds" is a big part

of what results in these "catastrophic losses".  The losses from varroa 
alone are bad enough that there is an actual argument bouncing around
that 
"CCD does not exist", and "the losses are due to poor varroa
management".

> I am sure both Eric and Randy would agree with you on that. 

No, they want to call their approaches not just "sufficient", but
"successful".  The "one-size-fits-all" approach is exactly what
they openly promote to others.  

> From experience Eric has arrived at "a rule"; for him a mite drop
count 
> of >55 per 24 hours in early September requires action, and this is
also 
> based on an eyeballing of relative population strength.

> He [Eric] also states;
>> "Are there special factors (variables) that could interfere with the
rule?  
>> Sure.  Will I have hives that are big enough to withstand >55 mites?
I'm 
>> sure I will.  Could the 55 number be unnecessarily conservative?
Perhaps, 
>> but more than enough hives (on average a majority) come under that 
>> threshold for it to still be very useful.  Is every winter going to
be the 
>> same?  Of course not.  Is every strain of bee and every strain of
mite 
>> going to follow the same rules?  Of course not.  Are there going to
be 
>> regional differences?  Sure.  In short, there are lots of variables,
but 
>> they don't render the information we do have meaningless.".......

Yeah, when you ask some questions, suddenly the excuses appear, and the 
backpedaling is so hard, the chain comes off their bicycle.
So with all those exceptions, what does the number "55" mean?  
I submit that it has no meaning at all.
It is a scalar trying to do the job of a vector or a curve.

But there are many, many more variables that a "threshold" approach
ignores:

1) No treatment is going to give you a 100% kill, so different hives
   are going to start in early spring with different varroa levels.
   Without some sort of check in Spring, these higher varroa populations
   are going to get completely out of hand well before they get around
   to doing their first (and ONLY) mite drop count in Fall.

2) Some hives tend to have more drone comb than others, and their
   varroa populations ramp up more quickly as a result.  These hives
   become the source of higher infestation for all colonies in the
   same yard, due to the "drift" factor, which is a much bigger
   component in varroa management than many beekeepers will admit.

3) Some queens just don't stop laying as soon as others, so one does
   not get the kill rate one expects, and that hive has a much higher
   varroa level after treatment.  Someone who only has one data point
   on their graph will never see this.

I could go on, you get the idea.

>> "The concept of thresholds only fails if they are used as a magic
number 
>> that you wait for until you treat your colonies."

But that's EXACTLY WHAT IS BEING ADVOCATED!!!

>> They are very useful as predictors if you're on a dangerous
trajectory, 

No, a "trajectory" would imply that one has done more than one mite
count,
and has some idea of a TREND.  To predict a TREND, one needs at least a 
LINE! A single point, in isolation, only allows one to compare this
colony 
with that one at THIS point in time, which is a useless bit of trivia.

>> It's not at all hard to adjust them for colony strength by
eyeballing.
>> The most important thing is to use a sampling technique that is
accurate and 
>> appropriate for the season."

Ah, yet more backpedaling.  So one needs many different thresholds, do
they?
One for "weak", one for "average", and one for "robust" hives?  Funny,
Eric
said "greater than 55", without qualification.  Looks like Randy and
Eric 
have very different evaluation methods for their "thresholds".


> So please Jim, for the varroa-illiterate, including me, what would we
actually do instead?

1)  You admit that varroa do not infest hives, but entire yards.

2)  You use sentinel colonies, and you test at regular intervals,
    especially in early spring, when it is easy to get a good
    kill due to small brood areas and confined bee populations.
    (Nip the problem in the bud!  What's so radical about that?)
    
3)  You use queen cages to assure that you are treating during a
    "broodless period".  You pick your dearth, and you manage
    your BEE POPULATION as a way to manage the varroa population.

4)  You take advantage of winter and early spring as a time when
    you can check EVERY colony, perhaps not with an accurate count,
    but at least with a quick glance at a sticky board to see if
    you have any colonies that are "varroa factories", and will
    be sources of reinfestation for the other colonies in the yard.

5)  You test weak colonies when you see weak colonies.  Sure, it may
    be nothing but a lousy queen, but wouldn't you rather know?
    The impact on the yard is what you care about.

There's a whole lot of good techniques out there, and I did not come
up with ANY of these approaches myself.

The extension community tried to convince beekeepers to do real IPM for
YEARS.
(Search for "IPM" in the archives of Bee-L, and you will see how prudent
IPM
was met with blank stares and whining by the beekeeping community.  The
extension community learned the hard way that they were wasting their
time
trying to teach IPM to beekeepers, so what we ended up with is the
current
"Special Olympics" approach to IPM.

> OK, which variables? How many measurements?

Well, no one has mentioned even tracking weather data.
Clearly weather has an impact on every other IPM program
ever developed, so I don't need to "justify" this do I?

But I don't want to push too hard here - it is enough to insist
that one can't see a "trend" with a single point on a graph, and
that waiting until "Fall" to test is way, way too late.

This is very reasonable, and is basic to the entire concept of "IPM".

Recall that the word "data" is a collective noun.
To have "data", you have to have more than one value.
So, treatment decisions are being made WITHOUT DATA!!!!

Gack!  What's so hard to grock about that?

******************************************************
* Full guidelines for BEE-L posting are at:          *
* http://www.honeybeeworld.com/bee-l/guidelines.htm  *
******************************************************