NOVA Online (click here for NOVA home)
Garden of Eden


Frog with ant The giant tortoise and the ant-sized frog, both native to the Seychelles, exemplify what can happen to species that wash ashore on islands free of their own kind. (Note the ant clinging to the frog's hindquarters.)
Why Do Islands Breed Giants (And Sometimes Dwarfs)?
by Peter Tyson

The giant tortoise and the frog the size of an ant, both native to the Seychelles, exemplify something that happens to animal species that succeed in colonizing new islands: Over time, in their new home, many of them tend to become either gigantic or dwarfed.

Island giants are aplenty: Komodo has its dragons. Madagascar has its giant hissing cockroach. Until about 1,000 years ago, New Zealand had its colossal bird, the moa. Of dwarves, the world has witnessed everything from foxes, rabbits, and snakes that are smaller than their mainland counterparts, to that ultimate oxymoron, the pygmy mammoth, which once existed in various forms from California's Channel Islands to Wrangel Island in the Siberian Arctic.

Why does this happen? What factors encourage a species to alter its dimensions on islands? What determines whether a creature will get Brobdingnagian or Lilliputian?

The island rule

Number of Species or Subspecies
Smaller Same Larger

Marsupials 0 1 3
Insectivores 4 4 1
Lagomorphs 6 1 1
Rodents 6 3 60
Carnivores 13 1 1
Artiodactyls 9 2 0

Table showing relative sizes of island-dwelling mammal species and subspecies as compared with their presumed mainland relatives.

Source: Foster, J. Bristol, "Evolution of Mammals on Islands," Nature, Vol. 202, April 18, 1964.
The first attempt to account for this apparent evolutionary roulette was made in 1964 by a young biologist named J. Bristol Foster. Fresh out of a doctoral program at the University of British Columbia, Foster published a brief but influential paper in the journal Nature entitled "Evolution of Mammals on Islands." Claiming "much confusion and contradiction" in the scientific literature on the size of mammals on islands, Foster undertook a survey of 116 insular (island-dwelling) species or subspecies living mostly off the coasts of western North America and Europe. He summarized his findings in the table at right, which indicates whether an island critter is daintier, similarly sized, or brawnier than its presumed mainland ancestor.

The table reveals some interesting trends. Rodents tend toward gigantism, while carnivores, lagomorphs (rabbits and hares), and artiodactyls (deer, hippos, and other even-toed ungulates) are more likely to become dwarfed. Overall, amongst mammal species that colonize islands, big ones have a tendency to shrink while small ones are apt to bulk up. Biologists have come to call Foster's generalization the "island rule."

Granite island J. Bristol Foster only began to unravel what factors determine whether a species of animal newly arrived on an island will become gigantic, dwarfed, or neither.

With the caution of a child testing a hot bath with an exploratory toe, Foster went on to offer tentative explanations as to how at least some of these extraordinary transformations occurred. Islands, he argued, contain fewer species than mainlands and thus fewer numbers of both predators and competitors that might face a newcomer. "In such situations," he wrote, "it appears that the larger rodent has an advantage." But how to explain those species that diminish in size? Foster offered one possible answer, and that for just one group, the artiodactyls. While rodents are able to control their populations in the absence of predators, hippos and deer and their kind cannot. As a result, artiodactyls are "especially susceptible," he wrote, to exhausting food resources and occasioning malnutrition and stunting in their young. If, in succeeding generations, smaller individuals met with greater reproductive success, then eventually evolution might begin to favor them, leading to dwarfism.

Foster's modest paper represented, as David Quammen puts it in his book The Song of the Dodo, "a sort of prerevolutionary innocence, standing on the distant side of a major upheaval." That upheaval was the 1967 publication of The Theory of Island Biogeography, by Robert MacArthur and Edward O. Wilson. The book not only launched an entirely new field of scientific endeavor—the study of how insular plants and animals got to be where they are today—but spurred a host of young biologists to tackle the gigantism/dwarfism question.

Chuckwalla Three times larger than its closest mainland relative, the roughly two-pound giant chuckwalla has gone in for gigantism in, well, a big way.
One of those young scientists was Ted J. Case. Like Foster, Case published a seminal paper early in his career (1978) in a leading scientific journal (Ecology). But his paper was much longer—19 pages compared to Foster's two—and, against Foster's lone table, it offered numerous graphs, tables, and complex mathematical equations. In his paper, Case acknowledged Foster's pioneering work, then pointed out where his predecessor's analysis fell short. For example, new data indicates that Foster's generalization about island-dwelling meat-eaters may not hold up, Case stated, and Foster never even mentioned the tendency for elephants to dwarf on islands. Case, who studies an iguana-like reptile known as the giant chuckwalla, also noted that no one had yet published a tally of size change in reptiles like that Foster had done for mammals.

Most important, Case stressed glaring exceptions to the island rule: how the same lizard or rodent could be relatively large on some islands but not on others, and how one island may have gigantic forms of one type of lizard or rodent and dwarf races of another. As an example of the latter scenario, he cited the curious case of two rattlesnake species that cohabit Angel de la Guarda, one of the sun-baked desert islands in the Gulf of California where he studies the chuckwalla. On the nearby Mexican mainland, Crotalus ruber is roughly twice the size of C. mitchelli, but on Angel de la Guarda, the situation is exactly reversed, with C. mitchelli about two times as big as C. ruber. How did this happen? Judging from a close look at the two species, C. mitchelli appears to have diverged more from its mainland progenitor than has C. ruber, Case says, which implies that C. mitchelli arrived first on Angel de la Guarda. In order to make use of all available prey, C. mitchelli went in for a larger size. When C. ruber finally reached the island, it found the big-rattlesnake niche already taken. So it had to settle on the little-rattlesnake niche, and it evolved a smaller frame to do so.

Tortoise From the looks of it, nothing has stopped the species of giant tortoise found in the Seychelles from growing just as big as it wants.

Bearing such anomalies in mind, Case concluded in his paper that "[a]ny theory proposing to account for these insular size trends must also be consistent with their numerous exceptions." In those 18 pages of small print, Case elaborately spun out such a theory. The chief factor that underlies all modifications in body size, he argued, is the net amount of energy that an animal can gain in a given amount of time. All changes proceed from that, yet all kinds of other factors also come into play to decide whether that animal becomes brawnier or daintier. And, practicing what he preached, Case allowed for exceptions to his modification of Foster's rule. For example, if an ever-increasing size in species undergoing gigantism eventually were to interfere with, say, that creature's ability to fly (in birds) or climb (in geckos) or burrow (in rodents), then that species would enlarge only up to the point where these other factors become of overriding importance.

Continue: Blazing the Trail

Seychelles Through Time | Saving the Magpie Robin
Why Do Islands Breed Giants? | Build an Island
Resources | Transcript | Site Map | Garden of Eden Home

Editor's Picks | Previous Sites | Join Us/E-mail | TV/Web Schedule
About NOVA | Teachers | Site Map | Shop | Jobs | Search | To print
PBS Online | NOVA Online | WGBH

© | Updated November 2000
/wgbh/nova/eden/textindex.html /wgbh/nova/eden/ /wgbh/nova/eden/