What makes honeybee colonies Collapse?

Frank explores a bee hive. Credit: KJHvM. See the whole album here.

Honeybee Colony Collapse Disorder has always interested me, because I’m interested in insect pathology – and this is probably the most important insect-pathology related event we’ll see in our lifetimes. I’ve written about CCD here at Biofortified, first in my post Colony Collapse Disorder: an Introduction. I followed this up with Are Neonicotinoids the Cause of Colony Collapse Disorder, where I talked about why the pesticide topic was a lot more complicated than neonicotinoid topic alone.
I’ve not been happy with media narratives which focus exclusively on neonicotinoids, because I think the picture is a lot more complicated than one group of pesticides. There are a lot of things which make bees sick, and a lot of these things change the social structure of bees in ways which are negative for the health of the colony. Honeybees also have problems finding food in many areas, which makes these problems worse. So, to restate something I’ve said in previous posts – I don’t think pesticides are entirely blameless, but I think many popular science articles on the topic lay too much blame on pesticides. CCD is multifactorial, with a lot of factors which interact to cause problems.
One question which I’ve had for awhile is: What happens when honeybee colonies Collapse*? In other words, why do the bees leave? A paper in PNAS, Rapid behavioral maturation accelerates failure of stressed honey bee colonies, seems to have answered the question, at least partially.

To understand what’s going on in this paper, we need to know a little bit about honeybee colony structure and how it changes both over time and in response to environmental variables. Honeybees have a distinct behavioral progression as they age. Directly after emerging as adults, they care for the young and perform housekeeping tasks for a few weeks. Awhile later, they begin to forage and bring food back to the colony. This behavioral progression is regulated by social cues provided by the queen and the foraging bees.
When bees are stressed, for essentially any reason (starvation, illness, wax deprivation, CO2 narcosis, poisoning, etc.), this behavioral progression gets bumped forward. Foragers abandon their roles, and the nurse and housekeeping bees begin foraging in their place. In the case of illness, this is likely adaptive because this moves the most infectious bees outside of the colony.
It’s important to remember that CCD is a specific condition which results in specific symptoms. CCD is a lot more complicated than simply a lack of bees. There is a checklist which must be met to diagnose the condition. I’ve brought up this checklist before, in relation to Chensheng Lu’s work.
Pery-Lu comparisonCCD goes beyond a sudden loss of the hive’s population. At the very least, there needs to be evidence of healthy brood, evidence that the remaining adults are biased towards younger bees, a healthy queen present, and evidence that the hive wasn’t killed by Nosema or Varroa mites. Absence of certain parasites, or avoidance of the hive by scavenging bees is another characteristic. Lu’s paper, along with others, claimed to replicate CCD but did not meet most of these characteristics.
The fact that the remaining bees are young is important, because young bees don’t leave the hive. If you’re investigating a hive collapse, a great question to ask is this: Are the bees young because the hive is Collapsing, or does the hive Collapse because the hive is young?
Clint Perry tackled this question by manipulating the ages of the bees in the colony. They compared hives started with younger bees to hives started using a mix of bees. They wanted to see if colony progression would be different between the two colonies. There are a lot of differences between older and younger bees which aren’t readily apparent, so it’s possible colony progression would be different.
There are a lot of complicated models and figures in this paper, but they did find some very interesting results. To be brief, it turns out that young bees aren’t very good foragers. Because the young bees aren’t very good foragers, the hive becomes nutritionally stressed and food stores shrink. As the bees get hungrier, more and more bees begin to forage. As the colony shrinks, it attains the symptoms of CCD described above.
CCD appears to happen when enough older bees are removed from the colony that younger bees need to pick up the slack. If food stores are sufficient enough for the bees to survive 14 days, Collapse may be delayed or averted.
Perry 2015 figure
Image from Perry (2015)

Perry’s hypothesis, outlined in the figure shown here, explains a lot of what’s seen with CCD. For example, hives can avoid CCD by robbing other colonies. If there aren’t any food stores in the hive, it’s likely that rival hives would avoid afflicted hives because they wouldn’t be worth the effort of robbing. These results do show that nutritional stress is very important to setting off CCD. However, they don’t demonstrate any one factor is responsible for the removal of adult bees.
Perry managed to replicate CCD without pesticides, and if we compare the symptoms of their hives to the symptoms of Lu’s hives, there are some very telling differences between the two studies. Pesticides are still probably a problem for bees, but I think this study shows that CCD symptoms aren’t uniquely a pesticide-related problem. The removal of adult bees is clearly the issue, and pathogens explain these symptoms very well. Furthermore, the Lu studies failed to replicate the symptoms of CCD using chronic sublethal pesticide poisoning – which I think says something about the CCD-pesticide hypothesis.
So, in light of these results, journalists should be asking “What factors are likely to remove adult bees from their colonies?” rather than the immensely popular “How do pesticides cause CCD?” I’m not exactly optimistic that the right questions will begin to be asked in the popular sphere – but the scientific community is asking these questions, and that’s what counts in the long run.
* Note: When writing about CCD, I always denote the condition with a capital C to set it apart from a generalized collapse of social structure.
Perry CJ, MR Myerscough & AB Barron (2015). Rapid behavioral maturation accelerates failure of stressed honey bee colonies, Proceedings of the National Academy of Sciences, 201422089.


  1. Joe, thanks, that is good information. I don’t know very much about the issue, but I can see why pesticides remain up high on the list of potential causes. There are so many different pesticides and they are so widespread in the environment and that they affect so many different physiologic pathways in the body that it is pretty logical to focus first on their potential effects. It certainly could be that the associations are many and could include many factors other than pesticides. The more we find out about disease and toxic causation, we find that multiple toxicant exposures are ubiquitous, making any causation assessment ultimately likely to involve multiple toxicants. So, it makes sense to keep in mind that multiple causative factors would be additive, and often even synergistic. Focusing only on neonics is simpleminded from the get go, but it is daunting to try to focus more on the multiple causation potentials… Even science has the distinct shortfall of having to oversimplify to conduct toxicity research, which contributes to the false impression of emphasis on one toxicant sourcing.

  2. Two things I learned during my honey bee workshop that might contribute to winter kill, were
    1) No winter thaw days above 50 degrees to do a cleansing flight, so the colony dies from the bee equivalent of dysentery.
    2) The winter cluster cannot find, or is inadvertently prevented from moving to a stored honey frame, and the colony starves. Mice can make this happen by nesting in the hive, and contaminating stored honey.
    Major symptom though, would be dead bees, not the CCD “few dead bees” characteristic.
    >>Winter kill/starvation
    Overwintering colonies sometimes run out of honey or the cluster is positioned so that the bees cannot reach the honey, and the bees starve to death. This is characterized by large numbers of dead bees piled on the bottom board and/or a cluster of dead bees found with their heads in the cells. Mold may be evident on the comb from decaying brood and/or adult bodies.<<
    Long PDF here, but you can search for "Winter"

  3. Aside from CCD, people are concerned about the effect of neonicotinoids and other pesticides on bees, and on bird species as well. You’re a scientist and a specialist, and it’s important for you to be precise about drawing conclusions. It would be a mistake to base claims about the cause(s) of CCD on unfounded beliefs, so I’m glad you take the time to explain how complicated the issue of CCD really is.
    But as to neonics, there’s evidence that they do harm bees. Some neonics are worse than others. So people can be thinking about CCD and neonics separately without being unreasonable.
    Seed coating with a neonicotinoid insecticide negatively affects wild bees
    Neonicotinoid Seed Treatments and Honey Bee Health
    Declines in insectivorous birds are associated with high neonicotinoid concentrations
    (sorry for the crazy address!)

Comments are closed.