Re: Sister queens and trials
 
Early on, I used sister queens  in all trials to reduce genetic  diversity. 
 But, when we compared colony metrics for sister  queen  colonies to 
non-sister queen colonies, the variance was more or less  the same.  In fact, at 
times the sister queen sets were more variable  than the non-sister queens. 
 
We even tried sister queens all inseminated from same drone lines  - and 
the variance got worse - probably as a consequence of the handling of  the 
queens ( back then we called it artificial insemination).  This was  years ago, 
before people like Sue Colby really improved the insemination method  (and 
now its called instrumental insemination).  
 
Regardless, I no longer use sister queens - bottom line, in most trials one 
 is trying to test something representative of the real world beekeeping  
practices - and sister queens is NOT a routine part of beekeeping  operations.
 
Similarly, I try to avoid 'equalizing' colonies - regardless of what one  
does, the strong colonies again become strong, the weak ones will revert to  
weak.  Our  bee counters indicate that strong colonies gain bees from  their 
neighbors - but don't necessarily follow traditional notions of drift - a  
colony in the middle of the yard may well be the strongest, and the weak  
colonies always lose bees (based on the percent of bees that return to the 
hive  each day)).  Strong colonies may well run to ~100-105% return rates  day 
after day.  For healthy colonies, we often see a return rate  of mid-90% or 
higher.  The weaker ones often drop to 92%, even down to  high 80% returns.  
So, the strong colonies routinely gain bees, the weak  ones lose bees.  The 
obvious conclusion, the weak colonies lose bees to the  strong colonies. 
(Note, the return rates are usually higher during prime  foraging season when 
floral resources are abundant, may fall off in  early spring and fall, 
possibly due to the aging of the forager  force. Old bees dying).  One has to 
generate the mean and mode return  rates for the beeyard at any given time and 
place, then look for the outliers -  unusually high or unusually low return 
rates,
 
 
When starting a trial, rather than equalizing (adding or removing frames of 
 brood, etc. or swapping hive position), I prefer to either: 1) grade/rank 
all of  the  colonies, then assign equal numbers for each grade category to 
every  control and  treatment, randomly distributing for controls and  
treatments OR 2) grade a large number of colonies and pick a set that is  well 
matched (as they occur - with no manipulation such as adding or  removing 
frames).
 
The first option is more easily attainable and again is representative of  
real beekeeping operations, the second is better for looking at subtle 
changes -  but you need a large set of colonies from which to obtain the matched  
colonies.  To be clear, under option 1, if you establish four  colony 
grades (1, 2, 3, 5) based on metrics such as frames covered by bees,  frames of 
brood, possibly weight of colony or frames of honey, etc.), and if you  have 
controls and 4 treatment levels, you'd end up with each control set and  
each treatment composed of a 1, 2, 3, and a 4 grade colony).  All of  the 1s 
would be randomly assigned to the control and to each of four treatments,  all 
of the 2s would be randomly assigned, etc.  There would be 4 control  
colonies, and 4 colonies in each treatment.
 
 
Jerry
 
 
 

             ***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software.  For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html

Guidelines for posting to BEE-L can be found at:
http://honeybeeworld.com/bee-l/guidelines.htm