“20% of the people do 80% of the work.” – Pareto Principle

credit: Yellowstone Wolf Project Annual Report 2007

Among social predators, hunting success initially increases with group size but then levels off and even decreases with larger groups. This is a commonly observed trend that has been noted in everything from insects to felines, wolves and chimps.

There are two hypotheses for the observed decline in performance among large groups: 1) the interference and 2)  the free-rider hypothesis. Renamed for fun:

• Too Many Cooks in the Kitchen – proposes that overcrowding is responsible for reduced hunting success (Hn ). It also predicts that the greatest drop in (Hn ) is expected with inexperienced hunters as they are more likely to interfere with each other. With interference, we would expect to see wolves getting in each other’s way in all phases of the hunt but the decline should be greatest with inept hunters.
• The Freeloaders – suggests that hunting success is non-linear because individual hunters withhold meaningful participation and take part only to be close enough to enjoy the spoils. It predicts that the level of competence is immaterial, and that withholding effort will be higher in hunting phases with greater individual cost (e.g. injury). The last prediction is that non-breeding wolves have greater incentive to freeload.

$The\ Study$

With individually known wolves (N=94), the authors collected hunting behavioral data for 8 years (1995-2003). Using the same basic methods briefly described in Wolves Start Getting Old at 2. This time the analysis focused on how group numbers affected individual and pack (Druid Peak, Leopold, Mollie’s and Rose Creek) hunting success.

MacNulty et al 2007

As in the previous study, the individual wolves were scored on whether they completed each of the 3 predatory tasks. If a wolf successfully participated in sequential foraging states then it was scored as a successful completion of the predatory task.

• approach => attack-group = Attacking
• attack-group => attack-individual = Selecting
• attack-individual =>capture = Killing

$The\ Findings$

Group Hunting Success

Graphs (a)-(c) show the non-linear effects of group size on the three hunting tasks. Graph (d) shows that the overall probability of success reaches a maximum with a group of 4 wolves.

Main effects of hunting group size on the probability that wolf packs attack (a), select (b), and kill (c) elk. Graph (d) shows the overall probability of success for a given encounter and the net effect of group size on group hunting success. (MacNulty 2011)

Individual Hunting Success

Crunching the data for individuals indicates that individual performance decreases with group size. For any individual wolf, the probability that it will kill an elk for a given encounter mirrors what we see in the group analysis.

Main effects of hunting group size on the probability that individual wolves attack (a), select (b), and kill (c) elk . Graph (d) shows the overall probability that an individual kills an elk given an encounter and thus the net effect of group size on individual-level predatory performance (MacNulty 2011)

$The\ Conclusion$

The folkloric view held by some is that of wolves as supreme social predators with some magical qualities beyond those of other animals. The reality is that wolves are just like any other group predator.

Like other social carnivores, birds, insects and primates – including (specially?) humans, bigger hunting groups does not increase hunting success. Increasingly larger groups show ever decreasing levels of cooperation. Token participation allows the individual to partake in the benefits of the group, while withholding effort minimizes the individual cost.

A similar pattern of reduced cooperation with increasing numbers has been reported in group conflicts among free-ranging dogs (Bonanni et al 2010).

The authors give three lines of evidence that support the withholding hypothesis.

1. Individual performance decreased as group size increased regardless of age or hunting ability [recall that these are individually known wolves]
2. Individual decline in performance is linked to the danger of the task. [more danger meant more withholding]
3. Decline in individual performance with increasing group sized related was slower in wolves with pups.

The authors also checked the data to see if fatigue from a previous hunt or the previous task could account for a wolf’s under-performance, for example, a wolf that took part in selecting might be too tired to participate in killing. The data showed the opposite was true and that a wolf was more likely to participate in a task if it had performed it earlier that day.

Breeder wolves were also more likely to select and kill than non-breeder wolves. Wolves with dependent pups were less likely to withhold effort regardless of group size.

One finding also went against the alternative explanation of inference. The rate of decline was greater among competent groups than less competent groups; this is the opposite of what is predicted by the interference hypothesis.

Many social also maintain groups several times larger than their optimal hunting numbers indicating that hunting does not play a role in the formation of large groups.

For wolves the optimum hunting group size is 4, greater numbers don’t make them better hunters.

The authors conclude:

“Regardless of the mechanism(s) involved, the widespread tendency for hunting success to level off with increasing group size suggests that the influence of group size on hunting success per se is unlikely to promote the formation and maintenance of large predator groups.”

I was unable to find papers on why free-loaders are tolerated –  reasonable given the newness of this paper. It might be that Canis lupus can’t keep track of cooperators and free-loaders, or that they lack capacity to formulate a punishment for cheaters. It might also be that they are gaining some fitness benefit in other non-hunting related tasks like defending a kill or maintaining territory.

That’s the way it goes with research, you get one answer which leads to a whole bunch of more questions.

REFERENCES

Bonanni, R., Valsecchi, P., and Natoli, E. (2010). Pattern of Individual Participation and Cheating in Conflicts Between Groups of Free-Ranging Dogs. Animal Behaviour 79: 957-968.

MacNulty DR, Smith DW, Mech LD, Vucetich JA, Packer C (2011). Nonlinear effects of group size on the success of wolves hunting elk. Behavioral Ecology 23:75–82.  doi:10.1093/beheco/arr159

MacNulty DR, Mech LD, Smith DW (2007).  A Proposed Ethogram of Large-Carnivore Predatory Behavior, Exemplified by the Wolf. USGS Northern Prairie Wildlife Research Center. Paper 105. http://digitalcommons.unl.edu/usgsnpwrc/105

• Actually that is a finished blog that is coming right after this one and hopefully I’ll be done with wolves.

• CG says:

Great, looking forward to it!

1. Makes sense that both scenarios would occur. Sad that free loading can happen in other parts of the animal world too. Did they also study where in the pack the free loaders stood? Also if it becomes apparent to the other wolves and they take action to limit their pack?

• None of the papers that came out of this long term study mention that. And from there appears to be no limiting of pack size; at one time the Leopolds were once in the mid 20s before they split into two. At one point the Druids hit 37 (in 2002) members but they also ended up splintering into several groups. So maybe there might be some mechanism keeping the packs from growing.

As to who is slacking, the authors didn’t look into status or affiliative relationships, but they do note that free-riding was age-independent. The other study that I mention about free-ranging dogs did find that the more affiliative relationships the less likely the dog is to hold back in a conflict.

• Thanks for addressing these questions. Pack hierarchy seems very complex and it would seem they would have a way to identify and address the slackers.

2. Wolf packs are usually mom, dad and their kids. It favors their genes if there are more in the pack which can spread and form new packs, thus it is a selective advantage with bigger packs, even if it does not increase hunt efficiency. The question when is if freeloaders can turn into efficient hunters when leaving the pack to form their own.
When a pack kills an elk it is food for all, thus a low cost for maintaining freeloaders. At least for most of the season.

• There seem to be some points of confusion. I don’t see how having more (adult) mouths to feed can be considered an adaptive advantage. We are not talking about incompetent hunters. The wolves are simply holding back from participating in any meaningful way because they are risk averse. Recall from the previous blog post that the authors had age-dependent performance data; what they found in this study was that performance decreases faster among a competent group than with an inept group.

3. Martin says:

That is really interesting. Are the free loaders always free loaders? Do they ever participate in a meaningful way?

• I gave the supplemental data a second look when Garden Walk Garden Talk asked about which wolves are witholding and I was unable to get any sense that the same wolves are always the freeloaders. Remember that the tendency to freeride increases with group size, but it’s not an either/or situation. It is also possible that there are some perpetual slackers in the pack but that MacNulty didn’t have enough data to make that determination.

4. As attacks by huge packs resulted in NO kills it seems to me more likely that the explanation is the too many cooks hypothesis than freeriders. The bars in diagrams probably indicate “errors” thus not very exact data. This is data from the growth phase at Yellowstone, when probably plenty of prey, thus hunting efficiency not critical. Maybe big pack hunting was for social reasons, and high efficiency was thus not important at this occations.
Are there data if the wolves were well fed at hunting? Hungry wolves may hunt better than well fed, for which pleasure and social activities are more important.
Wolf pack may have several meanings, it may be the hunting group or the social group. The hunting group is referred to and usually smaller than the social group. I mixed this up in my earlier comment (till now I have only thought in terms of social packs) and thus the comment was not very relevant for this situation.

5. DagL, recall the above as to why the authors rejected the interference hypothesis
“The authors give three lines of evidence that support the withholding hypothesis.

1. Individual performance decreased as group size increased regardless of age or hunting ability [recall that these are individually known wolves]
2. Individual decline in performance is linked to the danger of the task. [more danger meant more withholding]
3. Decline in individual performance with increasing group sized related was slower in wolves with pups.