Island animals have been an endless source of wonder and fascination for biologists for centuries, and often capture public awe as well. It is always fascinating to picture miniature elephants and gigantic rabbits adrift on dots of land in a vast ocean, flowers with unimaginable types of fruit, or communities in which "terror birds" have replaced mammals as the apex predators. These collections of bizarre and unique creatures also serve as valuable laboratories for discerning just what determines the composition and structure of natural communities.

The tenets of biogeography put forth by MacArthur and Wilson (1963) are well known and fairly ingrained in the fabric of research on island biodiversity and community composition: species diversity decreases as a function of an island’s isolation, and increases as a function of an island’s total area. The number of species represents equilibrium between immigration (determined by distance from nearby land masses) and extinction rate (determined by island size, which often determines resource abundance and habitat heterogeneity).

There is also a related phenomenon, however, which has gotten much less attention over the decades. In Carlquist’s (1966) comprehensive discussion of issues relating to dispersal and establishment in island systems, he discusses the issue of "disharmony": island species composition is often biased towards species that can travel to and propagate themselves on islands successfully. Disharmony, in a nutshell, means that insular assemblages include only a subset of the species found on the mainland. We often see this manifested in island communities with few, or even zero, non-volant mammal species, with carnivorous mammals being especially poorly represented due to the difficulty of sustaining their substantial appetites on the limited resources of a small land mass, not to mention getting their bulky bodies across stretches of ocean in the first place.

Principles such as compositional disharmony, and the "Island Rule" regarding gigantism and dwarfism on islands, often make for interesting case studies, because no two sets of islands are the same. But as with all areas of life, the main source of excitement comes when a "rule" is broken. Such are the novel cases that serve as exceptions to a principle (for example, an otherwise strictly marine mammalian taxa inhabiting a freshwater lake in Siberia), or even indicate that the principle itself may not be as solid as previously thought (see Meiri et al. 2007 for a re-evaluation of the Island Rule). These instances are often extremely informative about the processes and factors involved in a phenomenon. Not to mention just plain cool.

Which brings us to the Kuril Islands, an archipelago in the northwestern Pacific that stretches from just south of the Kamchatka Peninsula to the Japanese island of Hokkaido. These islands have been an ongoing point of contention between Japan and Russia since the end of World War II, resulting in a diplomatic tug-of-war over rightful ownership of this frigid and storm-swept chain of volcanic activity (Russia refers to the southern stretch of the archipelago as the Southern Kurils, while Japan calls them its Northern Territories).

Figure 2 – Kuril Islands

The Kurils have a complex geological history to go along with their complicated political status. The archipelago consists of 56 islands of varying ages, ranging in time of origin from the Cretaceous up through the Pleistocene. The islands fall into three general groups: 1) the northern landbridge islands, which, due to lower sea levels during the Pleistocene glaciation, were once connected to the Kamchatka peninsula; 2) the isolated volcanic islands in the center of the chain, which have never been connected to a landmass; and 3) the southern landbridge islands, which were once joined to both Hokkaido and the island of Sakhalin. The largest islands are not necessarily closest to their mainlands, a point that will become important later in this story.

While humans tussle over the right to claim this chilly necklace of land rimming the Sea of Okhotsk, the faunal assemblage has a surprising story to tell about how other, arguably less contentious mammals have distributed themselves amongst the islands. Hoekstra and Fagan (1998) examined compositional disharmony of mammal assemblages on this archipelago relative to its two mainlands to the north and south. They measured mammalian species diversity for each island, straight-line distance of each mammal-occupied island from the mainland maximum water depth between islands, and isolation—calculated by determining the "width of the water gap given a 100m decrease in sea level."

So, as expected, the roster of island mammal species was subset of those species found on the mainlands. Disharmony: check. As might be predicted from normal biogeographic trends, the smaller, central oceanic islands were completely devoid of native mammals, while all of the landbridge and two of the southernmost oceanic islands—Iturup and Urup—were home to up to four mammalian orders: Rodentia, Lagomorpha, Insectivora, and Carnivora (note there are no large-bodied herbivores such as artiodactyls). Island biogeography predictions of lowered species diversity with isolation and small area: check.

Now this is where it really gets interesting: while you would expect large, resource-hoovering carnivores to be disproportionately rare on islands more distant from the mainland (which presumably house less species diversity, i.e. less menu options for predators), the opposite is actually the case. Carnivorans were found on all but the very smallest oceanic islands, and in some cases the Carnivora was the order with the largest proportional representation among the island’s mammal assemblage.

For example, the southern island of Kunashir is home to three species of rodent, one lagomorph, and a whopping five carnivore species. On the northern end of the island chain, the island of Paramushir was home to three rodents, three insectivores, and four carnivores. The island of Iturup, with the highest degree of carnivore dominance, had a single rodent, a single lagomorph, and five carnivores, which means predators comprised over 70% of the mammal diversity.

Not only are they a large proportion of the fauna, these carnivores are relatively large animals: only two of the eight carnivores occupying the Kurils, a pair of weasel species, are relatively diminutive. The others range from the substantial to the gigantic: sables (a medium-sized mustelid similar to a marten), otters, wolverines, foxes, wolves, and the kingly grizzly bear—which is found on five islands, always in sympatry with at least one and sometimes both species of canid. Also, it is notable that the researchers did not count introduced species such as the Arctic fox, Alopex lagopus, but one may want to keep in mind that at times there is yet another carnivore sharing these habitats as well.

As a result of this unusual distribution, the Kuril Islands actually show an increase in average mammalian body mass with increasing island isolation, the opposite of what one would expect based on commonly cited trends. This is largely due to the presence of large-bodied predators on distant islands. As Hoekstra and Fagan report, "the proportional contribution of carnivorans to the total mammal fauna is an increasing function of isolation from the mainland" [emphasis added]. For carnivores, it seems here that island area mattered more than isolation in limiting diversity. For example, foxes were absent from three small landbridge islands, but were found on more distant yet larger oceanic islands.

These surprising trends raise interesting questions: 1) why is the compositional disharmony of these islands skewed towards large carnivores, and 2) how are so many species of large predators subsisting on these isolated islands, where few to zero other mammal species occur?

First of all, the distance issue. Remember that Island Biogeography holds that species diversity is negatively correlated to distance from mainland and positively correlated to island area. The isolation of an island matters largely due to the fact that the farther it is from another landmass, the harder it is for non-volant animals to reach it. Granted, some terrestrial mammals are shockingly good swimmers (elephants are thought to have settled Sri Lanka by swimming 20+ miles/32 km from southern India, and even camels, which evolved in some of the driest habitats on earth, can swim when necessary). Yet still, crossing stretches of ocean is largely out of the realm of possibility for most mammals, largely limiting non-volant mammal dispersal to species small enough to find other methods, such as drifting over on rafts of vegetation and debris.

The caveat here is that island biogeography was largely conceived based on research in the middle latitudes of the tropics and subtropics, where water and air temperatures tend to be amenable to even small, minimally insulated species. The Kuril Islands, however, cover a chilly stretch of sea between Japan and Russia. In January, the average temperature on the southern end of the archipelago is -6°C (21°F), and this average decreases to a frigid -24°C (-11°F) on the northernmost islands.

Beyond thermoregulation issues, the physics of dispersal is also especially challenging in the Kurils: the ocean currents surrounding the archipelago run perpendicular to the island chain, essentially acting like a conveyor belt to sweep even a determined swimmer away from its destination and out into the great blue sea.

Thus, we can see how the challenging environs of the northwestern Pacific are distinctly intractable for the dispersal of small mammals, either by rafting or swimming. In order to cross between islands—and to survive once you reach a patch of land—you must have a good amount of muscle and insulation and relatively low surface to volume ratios. It’s hard to picture small mammals cruising over to these islands as they did to reach Madagascar from mainland Africa, but one can imagine a tenacious wolverine or bear making the trip, especially in past ages when sea levels were lower and the distance between islands was shorter. Lomolino (1993) referred to this phenomenon as "winter filtering."

So, carnivores are well-equipped to colonize these chilly islands, and smaller species have a more difficult time. Thus, we have a theory to partially explain why we see the mammal diversity dominated by predators. But as we will see, the same physical advantages that allow them to disperse should, in theory, make it difficult for them to subsist.

Hoekstra and Fagan found that while island isolation did not seem to be a significant barrier to carnivore dispersal, island area was a critical factor, as predicted by the classic tenets of island biogeography. Carnivores are missing from the smaller islands close to the mainlands, yet are present on more distant ones as long as they are of substantial size (the smallest island occupied by at least one carnivore is 231 km2, and the largest in the chain is 2,479 km2). For example, no predators are found on 80 km2 Habomais, just 8 km from Hokkaido, yet they are highly diverse—three mustelids, a canid and the grizzly—on the 6,725 km2 Iturup, an intimidating 124 km from the mainland. In comparison, only two non-carnivore species—the very largest rodent and the very largest lagomorph—made it as far as Iturup. (For comparison’s sake, Hokkaido is 83,453 km2)

Figure 3 – (Hoekstra and Fagan 1998)

In a nutshell, carnivores can go the distance as long as there is area enough to support them at their destination. This results in the unusual trend of increasing average body mass with distance from mainland. Hoekstra and Fagan make no mention of gigantism amongst the small mammals that do occupy the far-flung islands, and attribute this pattern of increasing average body mass to the proportional dominance of carnivore species.

All of which carries a significant implication: the carnivores can reach islands that smaller animals, i.e. much of their common prey, cannot. This seems to answer the question of why carnivores dominate diversity, but not how: they’re here, now what’s for dinner?

Carlquist (1966) asserted, "Difficulties of establishment seem much greater than those of transport." Especially if you’re a meat-eater. Secondary consumers (predators, in other words), sitting at the apex of the food chain, require an alarming amount of resources. The commonly cited rule is that only 10% of available energy at a given trophic level will be transferred to the next, meaning that secondary consumers have access to only about 0.01-0.001 of the primary production of an island (depending on whether they are eating herbivores or smaller carnivores; the problems with a strictly hierarchical food web model are worthy of their own blog post). And the primary production of a small, mountainous volcanic island—especially in a frigid and storm-swept climate—does not strike one as being all that auspicious.

The Kurils, however, despite their challenges, do offer some advantages. The large islands in this chain feature rivers, which is habitat for salmonid fishes and freshwater copepods. They are also home to nesting colonies of sea birds. Even more fortuitously, the nesting season occurs at a different time of year than the "salmon run" glut of fish in the rivers, creating a fairly consistent supply of protein throughout most of the year. As Hoekstra and Fagan also point out, the subzero temperatures on these islands make it possible for predators to cache their prey with minimal decomposition during the winter months. An early study on island carnivores also found that predators will broaden their niche breadth—relative to mainland relatives—to make maximal use of resource and minimize competition on islands (Crooks and Vuren 1995), and the mechanism for maintenance of carnivore diversity on the Kurils may be similar.

Inadvertent (or sometimes all-too advertent) experiments resulting from species introductions often invite insights into community dynamics (see the stories of the cane toad in Australia and feral pigs on California’s Channel Islands as ominous examples). Nearly a century ago, a species of vole (Microtus oeconomous), which occurs only on landbridge islands, was introduced to several of the Kurils’ tiny (<3 km2) oceanic islands, and has persisted there ever since. These pinpricks of land apparently did have the resources to support the voles, yet the little guys couldn’t get there on their own, supporting the "winter filtering" model. Two carnivores, the arctic fox and the American martin, have also been introduced to the Kurils, to bolster the fur industry. These have been deposited on some islands that are totally devoid of other mammals, and their persistence shows that the avian and freshwater protein sources are indeed abundant enough to support predator populations in the absence of mammalian prey, further support for Hoekstra and Fagan’s explanations regarding the distribution of native species within the Kurils.

Are climate and ocean currents the only factors at play here? Probably not. Many of the Kuril Islands have high volcanic peaks, and the southern island of Kunashir, which reaches 1,819 m/5,968 ft (the second highest point in the archipelago) also has the highest carnivore diversity. This is just one of many additional conditions that may or may not interact with the complex set of factors that influence species distributions. Ecology is rarely simple, and it is likely that there are many ingredients in the recipe that led to the rare carnivore-biased disharmony found in this system.

The take-home message of this story seems to be that we cannot become complacent in assuming that established ecological rules, principles, and assumptions will play out as expected. Nature is far too diverse and interesting for that. The issues surrounding animal distributions are complex, and the importance of a given factor (such as area or isolation) may differ according to environmental conditions and the traits of a given taxa of animals. Patterns that make you stop and do a double-take—such as a mammal assemblage overwhelmingly dominated by large predators—often have the most to teach us about the intricacies involved in developing the fascinating patterns of biodiversity that we see around us.

*Hoestra and Fagan note that Hokkaido was deemed large enough to qualify as a mainland because it is >25 times larger than the largest body of land in the archipelago. References:

Carlquist, S. 1966. The biota of long-distance dispersal. I. principles of dispersal and evolution. The Quarterly Review of Biology,41:247–270.

Crooks, K. and D. Vuren. 1995. Resource utilization by two insular endemic mammals, the island fox and the island spotted skunk. Oecologia 104: 301–307.

Hoekstra, H. E. and W. F. Fagan. 1998. Body size, dispersal ability and compositional disharmony: the carnivore-dominated fauna of the Kuril Islands. Diversity and Distributions 4: 135–149.

Lomolino, M. V. 2005. Body size evolution in insular vertebrates: generality of the island rule. Journal of Biogeography 32: 1683–1699.

MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoogeography. Evolution 17:373–387.

Meiri, S., N. Cooper, and A. Purvis. 2007. The island rule: made to be broken? Proceedings of the Royal Society B 275: 141-148.

Island Biogeography graph source:

About The Author: Anne-Marie Hodge is currently working on her master’s degree at the University of North Carolina–Wilmington. She is a 2009 graduate of Auburn University, receiving a bachelor’s degree in Zoology, with a concentration in Conservation/Biodiversity and a minor in Anthropology. While at Auburn, she founded and served as president of Alabama’s first chapter of the Society for Conservation Biology. Her current research is on carnivore ecology in the eastern Andean foothills of Ecuador, and her field experiences in Mexico, Belize, Arizona, and Ecuador are a perfect example of "winter filtering" for a cold-intolerant mammal. When she is not tracking cats up and down volcanoes, she blogs at Endless Forms on the Nature Network.


The views expressed are those of the author and are not necessarily those of Scientific American.