Seeds of the Hawaiian ohia tree, Metrosideros polymorpha, compared to typewriter letters (12 point Courier in terms of computer printing). This tree probably reached the Hawaiian Islands via the jet stream. It also reached many other high islands in the Pacific.
“Sea beans” is the name given to these large legume seeds that commonly wash ashore on Pacific beaches and atolls. The seeds are those of Mucuna (larger) and Strongylodon (smaller), vining legumes that can be found in Indomalesia. The vines sometimes overhang streams, so that the seeds wash out to sea. These seeds were collected on a sandy island in Pearl and Hermes Reef.
These plants of Boerhaavia diffusa are growing in the coarse coral sand of French Frigate Shoals, an atoll in the Leeward Islands of the Hawaiian chain. The flowers and fruits are formed at the ends of stalks about 5 to 8 cm tall—heights that would place them at the level of the feathers of most birds walking on this habitat.
Heimerliodendron brunonianum has fruits with pinkish ridges that are amazingly sticky. The viscid substance on the fruits looks like and stretches like rubber cement. Unlike rubber cement, this substance takes years to dry. If a fruit of Heimerliodendron becomes attached to feathers, it does not fall off. In fact, birds that have caught several fruits on their feathers may become immobilized and die.
The seeds of Lepidium, of the mustard family, would not seem adapted to long-distance dispersal until one soaks them in water: then a coating of gel appears. This gel can attach them to the surface of a bird. The genus Lepidium may have reached the Hawaiian Islands, where there are two endemic species, by means of this mechanism.
Sharply pointed or barbed appendages on fruits can penetrate skin and tend to stay lodged there. Skin on the feet or bodies of birds may pick up such fruits. Bristly fruits and seeds can attach to feathers of birds.
Those who were skeptical that shore birds could bring seeds to the Hawaiian Islands should have paid more attention to the Pacific golden plover. Its annual migratory route goes from the northwest of North America to Hawaii and Tahiti. Shore birds travel well inland: this individual was photographed in Haleakala crater, Maui—well inland and at a high elevation.
The fruits of Astelia (Asteliaceae, formerly Liliaceae) are orange and fleshy, and would prove attractive to any bird that eats fruits. That may account for the amazing range of Astelia, which ranges from volcanic islands of the Indian Ocean to Pacific islands including Tahiti and Hawaii; it even occurs in Chile.
Pipturus (Urticaceae) has tiny fruits (achenes, each enclosed by a calyx). When mature, a condensed stem (receptacle) beneath the achenes enlarges into a fleshy structure that is attractive to birds. That undoubtedly accounts for the range of Pipturus, which is similar to that of Astelia, extending from the Mascarene Islands to a number of Pacific high islands.
This diagram from the 1967 paper (1967-The biota of long…V) shows “glyphs” for various Pacific Islands, arranged according to the geographical relationships of those islands. The key to the arms of the “glyphs” is at lower left. For each of the islands or island groups concerned, the percentages for categories of probable arrival modes of original colonizers are given. Not all guesses are necessarily correct! The text of the paper given is the original 1967 paper; a revised version is a chapter in the book “Ísland Biology.”
Tiny size characterizes most of the genera of land snails native to the Hawaiian Islands. This is Tornatellides, a genus native from offshore islands of Hong Kong to the Juan Fernanzez Islands. Evidently this genus is very good at long-distance dispersal but cannot compete on continental areas.
A butterfly, Vanessa kamehamehae, is native to the Hawaiian Islands, probably because butterflies are migratory and might have been carried to the Islands during an unusual storm system. Most of the Lepidoptera native to the Hawaiian Islands are very small, and may have been carried in the jet stream.
Seeds of Lobelia. Seeds this small (12-point typewriter letters shown) and rough are suited for various types of dispersal. Attempts to refine how particular genera with seeds such as these might have traveled may not be entirely successful, although circumstantial evidence may be able to narrow the probabilities.
A seedling of Mucuna, dug up and shown on the rough coral beach of Pearl and Hermes Reef where it was growing. There evidently was enough fresh water to permit it to germinate. However, the bright sun, heat, or possibly some intrusion of salt water has burned the tip of the plant, which would never have grown to maturity (although many seeds of Mucuna reach the Hawaiian Leeward Islands, no mature plant has ever been recorded on those islands). This plant is mute testimony to the fact that events of dispersal may not lead to successful colonization. For every successful arrival of a seed on an island, there must be numerous events of plants that have not been able to establish on an island.
DISPERSAL TO ISLANDS
I certainly didn’t originate the concept of long-distance dispersal. But I did use the term and the concept often. Long-distance dispersal is basic to all of my interpretations of island plants (and animals), because the rather amazing changes that have evolved on oceanic islands would not have occurred if dispersal were frequent. Even in the case of old continental islands, new immigrants do arrive and establish from time to time. There can be no question now that all of the immigrants on such islands as the Hawaiian Islands or the Society Islands arrived by means of long distance dispersal. The geological evidence and especially the DNA evidence have become overwhelming and compelling.
In the 1960s, when I was doing most of my work on island biology, the situation was by no means so decisive. In a sense, long-distance dispersal is easy to demonstrate: the plants and animals are sitting there on oceanic islands, and what has established on those islands differs to a certain degree, because no two islands are alike in position, age, climate, etc, but the similarities are more conspicuous. The volcanic Bonin Islands, south of Tokyo, for example, lack many genera typical of mainland Japan (the conifers), but the genera that have arrived are clearly a subset of the Japanese flora and other floras that is suited for long-distance dispersal. And a surprising number of genera on the Bonin Islands are similar to or the same as genera that reached the Hawaiian Islands, therefore. One can see events of dispersal of beach plants: one sees seeds or fruits adapted to seawater flotation washed up on beaches. But with respect to upland plants that arrive with great rarity on islands, and which therefore often turn into endemic species over time, witnessing long-distance dispersal is virtually impossible. So one must rely on circumstantial evidence of how plants and animals arrived on islands by means of long-distance dispersal.
Thus, my task in looking at floras and faunas of oceanic islands was explaining not a few of the arrivals, but literally all of them. That’s what I tried to do for Pacific Islands in my paper, “The biota of long-distance dispersal. V. Plant dispersal to Pacific Islands.” [ PDF ]. If one can explain arrivals on Pacific oceanic islands, one can explain arrivals on any oceanic islands. My method was remarkably simple. First, I reduced each of the Pacific floras I studied to the minimal number of immigration events necessary to account for the species presently there. If there are five species of a genus on the Hawaiian Islands, that means one arrival event unless there is good evidence for more than one arrival event. Then, for each of those hypothetical immigrants, I guessed the probable method of travel based mostly on morphology of seeds and fruits. And in the case of spores, size. And in the case of beach species, floatability of seeds and fruits.
The results were not what the average person would expect. The average person, if asked to guess, thinks of wind dispersal and dispersal by seawater flotation. Dispersal by wind certainly works for spores, and collection of pollen and spores by devices sent aloft in balloons and carried in airplanes long ago showed the presence of pollen and spores in the atmosphere. (Some island fern spores are larger than their mainland relatives, but that is because of the phenomenon of loss of dispersibility in island plants, covered in another section of this website). Dispersal by air flotation has worked for only a few native Hawaiian flowering plants. The ohia, Metrosideros, is one of them. As a tree species, Metrosideros is the exception that proves the rule. Most mainland forest tree species have seeds too large or otherwise unsuited to long-displace dispersal by wind currents. Testing my hypothesis, Carolyn Corn (1972) showed that Metrosideros seeds can withstand temperatures as cold as those found in the jet stream and can even withstand exposure to salt water. Orchid seeds are tiny enough to travel in the jet stream, but are probably damaged by such cold temperatures. (The right pollinators for orchids may not arrive on an island when needed, also). The number of orchid species on Samoa, much greater than that on Hawaii, shows the probable effect of prolonged exposure to dry, cold, temperatures and other factors that diminish dispersal with increased distance.
Dispersal by seawater flotation is plausible precisely because one can see it—the floating coconut, the branch floating in a pond. But the frequency of flotation observations means that they work for areas touched by water—which on islands, is essentially beaches and estuaries. And, as noted elsewhere (Loss of Dispersibility), once a beach species, always a beach species with rare exceptions (the Hawaiian Erythrina, for example). When dealing with atolls—which are nothing more or less than mid-ocean beaches—one is essentially dealing with beach species. Seawater flotation is therefore the predominant means of plant arrival on atolls. A few plants on islands are coastal but have seeds that don’t float (the Hawaiian Gossypium, or cotton, for example). These can be hypothesized to represent flotation by means of a branch or a whole plant or an unopened fruit, and cotton seeds apparently are more tolerant to salt than most other seeds. So I invented a category of “infrequent seawater flotation” or “rafting” to account for species such as the Hawaiian cotton.
I concluded that most of the species that have arrived on high islands of the Pacific have arrived on or in birds. This is counterintuitive, because forest tree birds are mostly nonmigratory or migrate only short distances. Marine birds do migrate to various extents, and most shore birds undertake migrations of considerable distances. Thus, the paradox: forest birds migrate little but eat fruits and seeds, but shore birds, which migrate most, eat animals along beaches and frequent those habitats, not forests. Or so the thinking went. In fact, the situation is not so simple at all. Shore birds do go inland, and often nest in cliffs or other inland locations. Stomach contents of shore birds and even marine birds may include mostly animal material, but about a 10% content was steadily reported for stomachs of shore birds—which typically gorge before migrating. Several shore bird species, such as the Pacific golden plover, range from Alaska and the Pacific northwest to Hawaii and Tahiti and back each year. Some forest birds, such as the Cape York pigeon, migrate considerable distances in search of fruits and seeds. Pigeons do not migrate to more remote Pacific islands today, and saying that they might have in the past may not be very convincing, but shore birds by themselves offer enough possibilities so that over a time of five to ten million years, arrival of much of the present-day native Hawaiian flora is not only possible, but probable.
Viscid and sticky seeds are obvious candidates for travel on shore birds and marine birds. Boerhaavia diffusa is ideally designed for travel on feathers of birds, and by lucky chance, I photographed a juvenile sooty tern on French Frigate Shoals in 1966 that had just picked up some of the very sticky fruits of this species on its head feathers. The fact that it was a juvenile permitted me to grasp in, hold it at arm’s length in one hand, and photograph it with the other. How many such events of attachment occur one can only wonder, yet some well-known workers actually believed that the Hawaiian and other floras of wholly volcanic islands could have been derived only by some kind of land connection. Imagining that all birds visiting such islands from mainland areas would be entirely sterile and free from seeds and fruits is difficult! The wide range of Boerhaavia diffusa in the Pacific is mute evidence of what happens. Heimerliodendron brunonianum grows in areas at least somewhat inland from the immediate beach, often well inland. The five extremely sticky pink ridges on its fruits adhere remarkably to almost any surface. Heimerliodendron brunonianum ranges from Australia and New Zealand to Hawaii.
One thinks of birds eating seeds, but many seeds or small fruits travel externally on birds. Some, like those of Bidens, have hooklike barbs or stiff hairs ideal for lodging in feathers. We tend to forget about the many seeds and fruits that puncture skin and stay attached to birds that way—the Cardionema shown here is just one example. One should remember that the webbed feet of shore birds, marine birds, and waterfowl are idea surfaces for puncture by sharply pointed and barbed seeds and fruits. This accounts for the travel externally on birds characteristic of seeds and fruits of a certain percentage of atoll plants.
Some seeds are tiny and embedded in sticky fruits. Pipturus is an example of this kind of dispersal unit—and it occurs on Pacific islands, although most other genera of the nettle family do not have fleshy fruit structure and are not found on Pacific islands. The huckleberry/blueberry Vaccinium, conspicuous in cooler parts of Hawaii and Tahiti, has such tiny seeds in a sticky pulp. Such tiny seeds can travel indefinitely on feathers of birds. One must remember that while we as mammals can be skillful (thanks to hands and to mouth muscles) in seeing that a small-seeded berry gets entirely into our mouths and any excess is wiped off from our faces, birds have no such advantage in excluding small seeds from facial areas.
The fact that birds feed with bills (unaided by the use of a handlike structure) also explains why they would gulp down seeds that we might spit out—such seeds are often difficult to separate from fleshy parts of fruits. And yet the internal carriage of seeds in birds for long distances has been questioned. Birds were thought to evacuate seeds too rapidly for transport of seeds to be likely across more than short distances. Vernon Proctor (1968) decided to test ability of a shorebird found commonly in Texas (killdeer) to retain fruits and seeds. He fed them seeds and fruits and observed how long they were retained. Interestingly, some relatively large seeds, perhaps in the size range of seeds of Astelia or larger, stayed longer in the birds and were ultimately regurgitated. The times that they were retained in the birds were compatible with the distances to islands and rate of flight of migratory birds. One can always add in unusual storm conditions.
In fact, occasional hurricanes must be hypothesized as the way in which some forest birds, such as the ancestor of the honeycreeper finches, reached Hawaii. Birds presumably stay aloft while being transported these long distances. Unusual wind conditions also account for transport of other animals—and perhaps some plants—to remote islands such as the Hawaiian chain. (The Hawaiian Islands are the acid test for long-distance dispersal, because they are the most remote archipelago in the world with a wide range of ecology from very dry to extremely wet, and they have never been attached to any continent. Thanks to the work of Clague, we now know that the islands preceding Midway in the chain were worn down to atolls before the newer islands emerged from the sea, so that seamounts could not have played any “steppingstone” short-distance dispersal role to the present-day Hawaiian chain for anything but beach plants). Most insects of Hawaii probably owe their arrival to traveling not in the occasional hurricane, but in constant high-altitude winds, the jet stream. The effectiveness of the jet streams in carrying insects was shown by J. Linsey Gressitt, who chartered a super-Constellation airplane and equipped it with a trap that caught, on a mesh, anything in the air larger than the openings in the mesh. A surprisingly large number of insects were caught, as well as small bits of rock and plant fragments. Not surprisingly, the insects caught were mostly of families that have reached the Hawaiian Islands from Indomalesia—the jet streams travel from Indomalesia eastward toward Hawaii. The insects native to the Hawaiian Islands are mostly small, suggesting travel in upper air.
There are a few exceptions to the small size of insects native to Hawaii. A butterfly (Vanessa kamehamehae), some grasshopper relatives, and a dragonfly are among these exceptions. These are all migratory insects. Although regular migration to Hawaii is unlikely for these, they are good candidates for being caught in major storms and carried far out to sea.
The land shells native to the Hawaiian Islands belong to small size classes. They may have traveled on birds, either as entire shells or as eggs in the dried mucuslike secretions of the shells.
Dispersal of animals is mentioned here by way of showing that the means of long-distance dispersal to islands for plants are rather different from those for animals, so that one cannot rely on plants dispersal as a template for how animals travel. And even though I have created categories for how plants travel to islands, each plant dispersal event is probably a distinctive one in some ways. This is underlined by a certain number of plants for which I suggested that dispersal of seeds in mud on the feet of birds might be a possible mechanism. When I suggested that, I thought that small seed size, texture of seed, and occurrence of sticky substances in the environment of a bird might be just as significant as mud. For small seeds and for spores, being caught in the barbules of feathers is a possibility. Although wind dispersal is suggested as a probable means of transport for spores, they could certainly travel in feathers, and presence of spores in feathers has been observed.