Hi all

I will try to straighten out things by describing in general terms how 
Morphometry can be used and the accuracy that can be expected.

First of all,

Morphometry is the precise study of anatomical characters by measurement.

Morphology is merely the study of form and structure, although the word 
is often wrongly used to imply morphometry.

If our objective is to sort unknown colonies of bees into racial types, 
there are various characters that can be measured that we already have 
databases for. (many of these characters can be viewed on...
http://www.dave-cushman.net/bee/morphometry.html
tools that are helpful are on...
http://www.dave-cushman.net/bee/morphometrytools.html

If your objective is to discriminate AHB from EHB, then the basic 
database is not enough, because the small body size of AHB overlaps the 
lower region of sizes that is also common to Ligustica, the genes of 
which are prevalent in US bee populations.

Some very detailed studies have been done of island populations of AHB, 
to the extent that they can tell which village the colony came from, but 
a database of this type is only of any use on the island concerned.

When two populations have portions of data that overlap we need to 
extend the database to include more characters in the measurements 
(things like inter-joint distances on legs, relationships of all wing 
veins) so that we can build an additional portion to the database. I say 
this glibly, because the amount of work involved is enormous and each 
finding has to be checked back to the original populations to make sure 
that it is valid.

I believe that such morphometric databases do exist in US research labs 
(I would love to get my hands on them), but the information is very 
extensive... To check all possibilities for a 50 bee sample runs into 
hundreds of hours of lab time, so in practice the full ID job is rarely 
done.

Then we come to FABIS

If we have reasons to suspect AHB, our suspicions are first aroused by 
the behaviour of the bees and the colony in general, because the number 
of cases where AHB is suspected is quite high, and the government 
establishments are unwilling to invest the full effort required to make 
a positive identification, a streamlined identification system (fabis) 
is used which picks a limited number of characters, but because of the 
limitation an ID can only be made on a probability basis rather than 
certainty.

I believe that cases identified as likely to be AHB were originally 
intended to be followed up with a full detailed analysis, but in 
practice, I do not think this happens.

Closed or limited populations

If we have adequate records of colony data that goes back many 
generations (of bees) for a limited population that we are already 
convinced of the racial type. Then we can look at this population to 
establish other things. Natural selection works right down to a village 
by village basis, but what can we expect to change in a bee over such 
short distances ? There are climate, geography, geology, forage that can 
all vary over short distances and we can map those changes and not on 
the map the locations of the colonies in our study. We can do a large 
number of dissections of bees measuring each part as we go. This data in 
turn can be compared to the map information of climate, geography, 
geology and forage to see if we can identify changes that match. If such 
a match is found we can then mount a study to examine that feature in 
yet more detail.

To sum up...
If you have a colony with well recorded behaviour, backed up by 
morphometry using the right database, you can have good confidence in 
your conclusions.

One or the other on their own is not enough.

I have not brought in to any of the discussion the use of DNA evidence 
(because we were talking about Morphometry). In recent years DNA has 
proved incredibly useful in confirming the high standards that 
morphometry and behaviour analysis can achieve.

The automatic identification of wing vein junctions, coupled with 
extensive computer calculations can now give good results with much less 
tedium to the operator, but we should be wary that we do not believe a 
result just because it has been generated by the computer, everything 
still needs to be capable of being tracked back to a database that was 
generated from bee colonies.


Regards & Best 73s, Dave Cushman, G8MZY
http://website.lineone.net/~dave.cushman or http://www.dave-cushman.net
Short FallBack M/c, Build 6.02/3.1 (stable)

-- Visit www.honeybeeworld.com/bee-l for rules, FAQ and  other info ---