Can a full moon really affect our behavior?

No celestial object has inspired more myth and legend than the moon - pale, mysterious, waxing and waning, inspiration for lovers, creator of tides and trigger for werewolves; personified as a deity or a demon or just the man in the moon (though some cultures see a rabbit or a frog or a toad). Wild behavior is blamed on the full moon - and the word "lunacy" comes from the Latin name for the moon goddess.

In his new book, "Moonstruck: How Lunar Cycles Affect Life," Ernest Naylor writes that this history of fantasy and fable has made us too skeptical of actual moon-related effects on earthly beings. While modern science accepts that the sun's 24-hour circadian pattern is wired into the genes of many species, he says, "scientists fearing ridicule" are wary of studying similar moon-related phenomena.

Naylor, a professor emeritus of marine biology and ocean science at Britain's Bangor University, sets out to counter that reluctance by citing patterns, mostly in the marine animals that are his specialty, of moon-related behavior. Crayfish, for example, adjust their activity levels in response to the phases of the moon, even when the sky is overcast and they can't perceive moonlight. Turtles and horseshoe crabs reproduce on moon-related cycles.

But do such links have anything to do with humans? Naylor says "rigorously controlled" studies in sleep laboratories give evidence that human sleep patterns vary with the phase of the moon - even when people have no way of knowing what phase the moon is in. This, he says, raises the possibility that we have genes with a "circalunar" clock as well as ones that follow a circadian cycle.

One area in which he does not find a strong link between the moon and human biology is the menstrual cycle. That link has been assumed rather than studied, he says, and there is actually no real evidence of a connection. He notes that this cycle "varies considerably" among women and points out that our closest primate relatives have menstrual cycles ranging from 10 to 50 days. What evidence there is, he says, points toward only a "superficial similarity" between menses and moon phases, and "when a correlation occurs, it is by chance."