Guybrush

Yes, exactly. This is something that happens under certain conditions. If several queens found a colony together, the queens will fight after emergence of the first workers. Then, workers may also get involved in an early in-house war, killing or throwing out queens. This occasionally also happens in mature colonies with multiple queens.

So although ant hills may seem like superorganisms, current evolutionary understanding does predict that because they are actually made up of selfish interests working together, there would be conflicts of interest and special cases like above where eusociality might completely break down, at least for a while.

Because it's an interesting subject, I've done a little bit of reading on that subject. This article was pretty good and is co-authored among others by Bert Hölldobler who has also been a collaborator with Wilson, but who is not a proponent of group selection (group selection support seems fringe).

I don't expect people to want to wade through that like I have, but a few reasons in-colony conflicts may exist are described.

In a monogamous single-queen colony, queens are as related to her male babies as she is to her female (50%). However, workers are 75% related to sisters and only 25% related to brothers. This means that workers are incentivized to care/invest more in female offspring (sisters) than they are in males (brothers). As such, the workers may work against the queen and her male offspring, working to skew the sex ratio between males and females to 3/4 females and 1/4th males, mirroring their level of relatedness. This could possibly done by feeding male eggs to female larvae (not sure if that happens). Studies on sex-ratio seem to suggest it is close to 3 females for every male which suggests that workers are in control of this aspect of colonies. HOWEVER nature is messy and I'm sure deviates from this ratio are abundant. One thing a queen can do in order to stop workers messing with her man-babies is to mate with several males. If she does, the relatedness between sisters should drop and so should their incentive to "mistreat" males. Observation in Fornica truncorum colony with several queens showed a higher ratio of males, seeming to support this hypothesis.

Another source of conflict is origin of males as workers are often capable of laying male eggs. This has been mentioned earlier and may lead to workers policing that kind of activity in colonies where the queen is a polygamist. Also the queen is incentivized to inhibit worker reproduction. Slave-keeping ants also prevent their slaves from having male offspring (which would they would not be related to at all and so would care nothing about).

Another source of conflict seem to be in ant species where the shapes (morphology) of ants are more similar. The typical queen is ideally suited for egg-laying while a worker is not, so this would reinforce how a worker should rather spread her own genes by using the queen as her sexual proxy. However, if queens and workers are physically more similar and more alike in capabilities, that seems to increase the level of violent conflict.

A few other kinds of conflicts that may appear:

• Queens fighting other queens
• Queens begging food more often from workers with bigger ovaries (which may reproduce), thus keeping them from reproducing
• In the genus Diacamma, morphologically distinct queens do not exist. Instead, a worker tries to monopolize reproduction by mutilating larvae after emergence so that new workers cannot sexually reproduce themselves.

I still have not yet read Wilson himself, but he is referenced in this paper saying that he believes some colonies to achieve superorganism status if the level of conflict/competition within a colony dies down enough. The authors of the paper conclude that calling ant hills superorganism or suggest they may be a unit of evolution in itself is difficult as it would be hard to separate the levels at which selection occurs. They write somewhat diplomatically (to my mind) that the concept is a useful heuristic device.

To my mind, a problem with group selection arguments are that they don't explain the evolution of conflicts and related behaviours like mentioned above. It leads to misunderstandings, making people ignorant of the constant presence of conflict, exploitation and competition that exists inside populations. Animals don't do things for the good of the colony/populations, but for themselves (or rather their genes). The predictions that arise from this fits observations beautifully. I also don't (yet) know what mechanisms are suggested for group selection, so there's that.

jwb, although you may be on Wilson's team in this (?), I implore you to pick up and read The Selfish Gene Although it is quite old (its 50th anniversary will come up in a couple of years), it is still relevant and I'm relatively sure it will blow your mind in a good way. It is just a genius piece of work and my favorite bit of scientific litterature. In terms of understanding evolution, it was a big eye-opener, more so than any uni lectures I can remember.

__________________
Something Completely Different