Craig Packer, a lion expert and University of Minnesota professor, led a just-published study about lion attacks that found such attacks on humans are much more likely to occur on the evenings following a full moon. Read more about that study here.

Packer answered a few questions via email from Tanzania:

Why study this?

We were concerned about the high level of danger to rural villagers in southeastern Tanzania, so we wanted to identify the associated risk factors and develop mitigation strategies. It was already known that lions were less successful in catching ungulate prey around the full moon; we wanted to know the effects of moonphase on human predation.

Did you expect to find what you did - that "the full moon is a reliable indicator of impending danger"? If so, why?

We were astonished to see how accurately the lunar cycle predicted risks of man-eating. When lions catch wildebeest and zebra, their success rate is symmetrical around the full moon: they are least successful a few days before and after the full moon. But for humans, the risks of predation switch from low to high immediately after the full moon. This is because people only move around outside in the evening whereas ungulates are outside all night long -- thus the lions can catch wildebeest, etc, in the hours before dawn in the week before the full moon and in the evening the following week. But humans are safely in bed in the morning and only accessible in the evening.

You've spent a great deal of time in Africa. Were you picking up on an association before you drilled into the data?

No, I had no idea the association existed until analyzing the data. The risks are so much higher, I'm quite surprised that no one had ever noticed. But on the other hand, these sorts of events are so frightening and horrible, that people mostly dwell on the details of each case rather than think about them in aggregate.

In addition to overlaying the data, your team spent time talking to survivors and victims' families. Why? What did those interviews tell you that the data did not?

The interviews confirmed the precise time and date of each attack, so they were essential for this particular analysis. But more generally we wanted to learn as much as possible about the precise context of each attack to better understand all the associated risk factors so as to better protect people in the future.

How might this research translate into practices that might protect people?

Street lights would obviously be a good idea, except that these areas are so impoverished that they rarely even have electricity. We are continuing to work in some of these communities and are warning them of the risks of the darkness following the full moon.