"simulating a hurricane by putting fans around the bathtub and studying the effect on a fleet of toy boats?"

Such experiments are done with "large scale flume" apparatuses...these are the "bathtubs", and models of boats, bridges, or whatever else one desires to test are used in the flume devices. Scaled experiments are excellent approaches to deal with large-scale problems, as long as the experimenters know enough physics to deal with the scale effects.  It's done all the time in space research, as well as in automobile and other engineering applications.

Pete, your analogy was perfect.  I'm curious as to why you apply it in fun?  The point of the bumblebee experiment, as far as I can determine from the paper, is to come up with a mathematical model that can describe mortality in eusocial bumblebee colonies.  They state clearly that bumblebees were chosen because they could track all individuals....a useful metric when putting a model together.  The results they found were interesting, confirming what many on this list have said many times.  Here I quote:

"Our results provide two explanations as to why it has been so difficult to explain colony losses. First, although we imposed a specific (pesticide) stress in our experiment, the argument about reduced colony function applies to any stressor that reduces the contribution made by individual bees to colony function. This suggests that multifactorial stress can cause colony failure (Vanbergen & The Insect Pollinators Initiative 2013), but that failure is the result of a critical stress level from the accumulation of multiple sublethal factors (e.g. disease, weather and anthropogenic influences) without the need for synergy between these effects. This can explain why finding the link between colony failures and a single specific stress factor has so far proved elusive. Second, it appears that a number of potential causes for the failure of bee colonies have been dismissed because the presence of the stressor is not a good predictor of colony failure (Cox-Foster et al. 2007). On the basis of our model, we predict that two colonies, with similar stress levels can have divergent fates through the Allee effect caused by the feedback through colony function. If the effect is strong, one colony may be growing, while the other is set on a path to failure. Even if the effect is weak, random events (such as other stresses or demographic noise) can put similar colonies
on divergent trajectories. This will weaken the correlation between the level of a stressor and colony failure. For example, the parasite Nosema ceranae can cause colonies to fail (Higes et al. 2008).
However, while N. ceranae was detected in 100% of failing colonies, it was also present in 47% of non-failing colonies (Cox-Foster et al. 2007). Our results therefore suggest that correlation statistics should be used with caution for inferring causality."

In other words, the authors state that we cannot point at any one stress factor as a sole explanation for colony losses, because even when stressors are non-synergistic, multiple stressors TOGETHER are interacting dynamically in a way that can result in divergent outcomes between neighboring colonies.  Exactly what the Bee-L participants are saying over and over (with the exception of those who think that certain pesticides do not stress bee colonies significantly).

Christina

             ***********************************************
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