It's Feast Or Famine: Predators May Drive Lemming Cycles, Science Researchers Say

While it's a myth that lemmings hurl themselves en masse over cliffs, the reality has been nearly as bewildering for Scandinavians and others at northern high latitudes who periodically find the land awash in these small rodents. Depending on the species, populations of lemmings and their vole cousins can explode by 100 or even 1,000 times their original size and then crash on a regular basis.

The collared lemming, which lives in the high-arctic tundra, is the single prey in one of the world's simplest vertebrate predator-prey relationships. The stoat, arctic fox, snowy owl, and a seabird called the long-tailed skua all dine on this hamster-like animal.

Scientists now suggest that these four predators may be solely responsible for the four-year population cycle in eastern Greenland and possibly in many other collared lemming populations. In contrast to previous hypotheses, food or space shortages didn't appear to be involved, the researchers report in their study, which appears in the 31 October issue of the journal Science, published by AAAS, the non-profit science society

"This question of lemming cycles has been open for almost a century. Different schools have argued about this. It has been a very, very hot issue," said study author Olivier Gilg of the University of Helsinki in Finland and Center for Biology and Management of Populations (CBGP), in Montferrier, France.

The field of small mammal ecology was born in 1924, when the eminent British ecologist Charles Elton published one of his seminal papers on rodent population cycles. Researchers have been studying these cycles ever since, in part because these simple animal communities make good models for more complex ecosystems, and because an explanation for the phenomenon has been so elusive.

"Population ecologists are in the business of explaining the changes in population size of plants and animals," said co-author Ilkka Hanski of the University of Helsinki in Finland.

"If ecologists can't explain what appear to be the most regular patterns, then we really aren't very far yet in our challenge of understanding the population dynamics of all species. So, lemming cycles may appear an esoteric problem, but to me it's a kind of test case of our ability to understand the forces that shape natural populations," Hanski said.

Gilg and Benoît Sittler at University of Freiburg in Germany have been monitoring the numbers of lemmings and their predators in a Greenland valley over the last 15 years. The researchers have spent the summer months scanning a 75 square-kilometer region of tundra, aided in no small measure by the continuous daylight and open landscapes.

For the Science study, Gilg, Hanski and Sittler compiled the 15 year's-worth of data for each of the five species.

The results showed the stoats' population cycles lagging behind those of the lemmings. This was expected, according to the authors, because the stoat is a specialist predator that depends exclusively on lemmings for prey. The only predator that remains in the study area year-round, the stoat also reproduces more slowly than lemmings do.

The numbers of foxes, owls, and skuas -- all generalist predators that eat lemmings only when they are abundant -- fluctuated right in sync with the lemming numbers, the authors found.

Gilg and his colleagues hypothesized that the generalist and specialist predators work in concert to limit the lemming populations, which would otherwise grow until the rodents ran out of food or habitat.

The stoats, with their reproductive lag, are the key to the lemmings' characteristic four-year cycle, according to the researchers.

During a peak lemming year, the generalist predators help keep the rapidly multiplying lemmings in check until the stoat numbers can catch up. At that point, predation is intense enough to drive the lemming numbers down to the low point in the cycle.

Once the three generalists have turned to alternative, more plentiful prey or left the area, the lemmings can multiply fast enough to reach their peak once again.

To test their hypothesis, the researchers constructed a numerical model that incorporated many aspects of each species' population dynamics, such as how fast the animals reproduce, how long they live and how many lemmings they eat. The model didn't take into account the amount of food or space available to the lemmings, but nonetheless predicted the very same trends for the lemmings that the researchers had observed in the field.

"Our model shows that all the predators are important," Gilg said.

"Now of course we can't easily infer anything about other more complex systems. But that certainly doesn't decrease probability that in some other complex system predators would also be very important," Hanski said.

In a related "Perspective" article that accompanies the study, Peter Hudson and Ottar Bjørnstad of Pennsylvania State University write that the new findings represent "one of those rare instances when nature appears to reflect basic theory, and authors of biology textbooks will no doubt incorporate these findings as a case study in future editions."

###

The study was supported by the French Polar Institute, the Academy of Finland, Deutsche Forschungsgemeinschaft and Groupe de Recherches en Ecologie Arctique.

Founded in 1848, AAAS has worked to advance science for human well-being through its projects, programs, and publications, in the areas of science policy, science education and international scientific cooperation. AAAS and its journal, Science, report nearly 140,000 individual and institutional subscribers, plus 272 affiliated organizations in more than 130 countries, serving a total of 10 million individuals. Thus, AAAS is the world's largest general federation of scientists. Science is an editorially independent, multidisciplinary, peer-reviewed weekly that ranks among the world's most prestigious scientific journals. AAAS administers EurekAlert! (http://www.eurekalert.org), the online news service, featuring the latest discoveries in science and technology.