CCD became what appeared to be a dire threat to humans about 6 years ago, because it was so widespread and catastrophic and so much of our food requires bee pollination. Industrial-scale agriculture requires industrial-scale beekeepers, because megafields of one crop, or mega-orchards and the like have a short blossoming period when zillions of bees are needed to pollinate the flowers, or there will be no almonds, cherries, cranberries, blueberries and the like. But after the harvest, in general the fields produce nothing else for the rest of the year. Local wild bees simply would not be numerous enough to do the amount of pollination industrial ag requires, so bees are trucked in from around the country.
But if you bring in megahives, you have to then truck them elsewhere to do their next job so the bees themselves have food enough to survive. The result has been big bee-businesses that maintain hives in the tens of thousands and, on a contract basis, truck them around from crop to crop. Unfortunately, and probably contributing to CCD, each area and each crop involve different exposures to things like viruses, pesticides, herbicides and so on. So it has been very difficult (i.e., hasn't worked) to find a local virus or pesticide that is responsible for the disorder worldwide. Or even nationwide.
Controlled studies of different hives or different agricultural areas have identified different candidate risk factors. Each, under controlled experiments, seems to cause problems for at least some strains of bees. So each seems to be a legitimate factor. Yet it doesn't by itself cause the widespread CCD occurring around the world.
One response to such facts is to say that CCD is multifactorial, but that doesn't seem to wash, because if that were the case, then why would CCD have reached such proportions so quickly, with 30-40% of hives being wiped out this past winter, e.g.?
One possible explanation is that the industrial trucking of bees from location to location has brought bees who winter in very different environments, or who spend time in different series of crops, together for some particular crop--like the massive California almond crop, where bees are pollinating on the order of 1000 of trees. This is the largest single bee-related monocrop area in the world. Bees from all over, thousands of miles distant, who have spent their past seasons working in other environments with other crops involving various pesticide, virus and parasite exposures etc., are brought together. In, say, the almond groves, they can exchange pathogens they are carrying or perhaps mix in other ways, and then after the blossoming is over, be transported to the next crop, maybe half-way across the country where they encounter ore bees from other areas.
Local causation, or not?
One spokesperson on a BBC documentary program that led to this post* likened this to bringing people from all over the world to a single location, where they sniffle and sneeze and exchange their viruses, only to fly back to homes all around the world. Whatever might be causing disease would not stay local, just as CCD has not stayed local. Had it done so, local factors in a specific population of bees and hives, might have led to a clearer causal understanding.
More possible, however, is that there really is a mix of multiple contributing causes, and no local area would have had a serious, confined problem until the mixing of bees from so many areas, carrying a diversity of afflictions.
This is like a polygenic trait in humans or other species. Many different genes contribute, but no one gene contributes enough to cause the trait on its own. Only individuals who have disease-associated variants at many different genes will manifest the trait. Here we are talking about individuals, and the distribution of trait values (e.g., blood pressure related to stroke) in the population reflects the distribution of genetic variants at the many contributing genes.
Except that here it is not individuals with different genotypes in a population, but hives of different exposure types brought together in a population of hives. If this is the reason for CCD it is a kind of hyper-individual polygenic-like trait for which complexity really is the story and the manifestation is only on a population basis.
As with natural selection, over time the fraction of individuals with vulnerable genotypes can diminish if the causation is polygenic. Likewise, perhaps the fraction of hives that have the bad combination of risk factors and hence don't survive, is reduced and the disorder changes prevalence--declining--over the years. That may explain in general principle why CCD seemed to be declining, at least before the decimation that seems to have happened in much of the US this past winter.
In any case, this explanation would be one of a kind of hyper-polygenic (hyper-risk-factor) causation where it really is true that many minor factors have, by the demography of bee-keeping in modern times, comprised a kind of population in which the CCD trait has appeared.
*If you want to learn more, painlessly and even entertainingly, listen to BBC Radio4 'Discovery' broadcasts (online or as podcasts).