Click below for larger images and slideshow more...

The takahe is a colorful flightless rail native to New Zealand.

The flightless Laysan rail, Porzanula palmeri, once lived in moderate numbers (about 2,000 recorded there in 1891) on Laysan. It became extinct when rabbits were released on the island. These two specimens were photographed through a glass case in the Bishop Museum.

The wing pattern shows this Hawaiian insect to be a lacewing. It’s Nesomicromus bellulus, an endemic of the Hawaiian Islands, probably now extinct. Even though its wings seem moderately large as wings of lacewings go, this insect could not fly.

This lacewing, Pseudopsectra lobipennis, is one of the flightless wonders of the Hawaiian insect fauna. Its wings are shell-like and too heavy for flight.

Nesothauma hakeakalae is a flightless Hawaiian lacewing, its wings relatively heavy compared with its body. The pattern of veins typically seen in lacewings has vanished here: one sees only a dense network of veins.

A flightless grasshopper from Konahuanui in the Koolau Range of Oahu.

Species of Darwinia, of the myrtle family, form an interest assemblage of plants in Western Australia. The flowers in Darwinia are not colorful, but the bracts (modified leaves) associated with the flowers are colored. This is D. neildiana.

The fruits of Darwinia oederoides are typical of most Darwinia species: they are relatively small, and have sepals that might aid somewhat in their dispersal.

The fruits of Darwinia meeboldii and the other Darwinia species endemic to peaks in the Stirling Range of Western Australia are relatively large and heavy, and do not have sepals in proportion to their size.

A member of the same family as the litchi, Alectryon subcinereum from Australia has black seeds about 4 mm in diameter, each surrounded by the bright fleshy red aril which attracts and is eaten by birds.

The Hawaiian Alectryon species have relatively huge seeds and fruits: the seeds are about 50 mm in diameter. This is A. macrococcus from the Waianae Mts. of Oahu.

The mainland Bidens pilosa is a weed famous for its achenes that get caught in socks (giving it the name of Spanish needle or beggar tick). It also gets caught in feathers of birds.

The awns of Bidens pilosa have stiff downwardly pointing barbs that run counter to the stiff upwardly pointing hairs on the bodies of the achenes—an ideal system for anchoring the achene in feathers, as well as fur or socks.

Achenes of the Hawaiian Bidens conjuncta have only the smallest vestiges of awns. The bodies of the achenes are hairy.

Achenes of the Hawaiian Bidens micrantha have a single coil. The reason for this is not understood.

Achenes of the Hawaiian Bidens torta are coiled into several gyres. This feature is known only on Hawaiian species of Bidens.

This photograph shows the rather huge achenes of Bidens macrocarpa before they dry. The relatively large volume of the achenes may reflect greater food storage in the seed, an adaptive feature where a seed germinates in shady areas such as the ones where Bidens macrocarpa grows.

The Erythrina or wiliwili native to the Hawaiian Islands grows in dry areas near the coast, as seen here on Lanai, but not on the immediate coast.

The flowers and seeds of the Hawaiian Erythrina.

The seeds of the Hawaiian Erythrina sink when put into water. Intercellular air spaces in the embryo are very small.

The seeds of the Indomalesian Erythrina variegata float when put into water. Very likely seeds with this capability are ancestral to the Hawaiian Erythrina, and ability to float was lost after establishment of Erythrina on the Hawaiian Islands. The seeds of Erythrina variegata float because they have relatively large intercellular spaces in their embryos.



My work on loss of dispersibility in island plants had a logical beginning.  I had read about the flightless cormorant on the Galapagos in Beebe’s “Galapagos: World’s End” when I was about 10.  There was no explanation there for why a cormorant should be flightless on the Galapagos Islands in that book, however.  Perhaps I was always fascinated not so much what books said, but what they didn’t say.  What books tell you as facts is good to know, but what books don’t tell you provides opportunities for original research. 
   When I was in high school, I bought, via mail order, publications of the California Academy of Sciences.  The California Academy of Sciences published numerous papers on the flora and fauna of the Galapagos Islands.  These were far beyond my comprehension at that time, and I’m not sure why I bought them.  Certainly my mother didn’t like my spending money (“squandering money”) on them, so I always hoped that I could take in the mail that included a sending of them before she got home from work.  I guess I bought them to maintain an interest in island creatures at a time when there was no way of my learning much about them. 
   I do remember thinking about one of them, a long paper on the Coleoptera of the Galapagos.  A catalog of Galapagos beetles, really, with a few descriptive words.  I remember noticing that quite a number of the beetles were recorded as wingless or with their wing covers “soldered” together so that flight would have been impossible.  But the author didn’t seem intrigued by that, there was no explanation for why that should be.  There was no discussion of the ecology or the habits of the beetles.  In those days, plants and animals were collected, made into specimens, and archived in the back rooms of museums. Those who collected them weren’t trained to look for evolutionary stories, and obviously weren’t paid to do so.   
   Somehow, I learned or figured out over time that flightless island birds and insects could compete well on islands because flight to evade predators was unnecessary if mainland predators of these birds or insects weren’t present on islands.  Some did claim (Darwin) that flightlessness in insects might prevent them from being blown out to sea.  Nice idea, but not applicable in most cases, as it turns out.  Flightless insects are mostly in protected places like forest floors where they are unlikely to be buffeted by winds.  (Flightless moths on subantarctic islands may benefit from not being in strong winds, however).  Flightlessness in birds and insects is mostly related to economy.  Producing wings requires energy, so if they are not needed, the limited energy budget in development of a bird can be devoted to other items.  But more importantly, flight itself requires enormous expenditure of energy.  Presence of predators does maintain natural selection for flight.  But something else does, too: location of food resources.  Take rails (coots, mudhens) for example.  These birds migrate, but not large distances—they are thought of as “poor” fliers, so that they might be blown off course and land on an island during a major storm.  But once in a feeding ground, rails don’t feed by flying: everything that they need can be eaten without flight.  So if a rail species reaches an island by chance and establishes there, lack of predators and availability of food on the ground make flight too costly, and over time, flight is lost, and then wings are reduced to vestiges.  The takahe in New Zealand is a colorful island rail.  And there were once two flightless rails in the Hawaiian Islands—both now extinct.  Flightless rails were present on other islands, also.  A number of them have vanished because of animals introduced onto islands by humans.  I put observations like these into my book “Island Life.”  “Island Life” was the first comprehensive attempt to explore in one book the phenomena of dispersal to islands and then how plants and animals evolved with respect to island environments.  I know, one is not supposed to put original scientific observations and syntheses into a semipopular book, but I did.  The only earlier book about life on islands of the world was Wallace’s “Island Life,” which merely tells which organisms are on which islands and something about them, but doesn’t look at evolutionary phenomena.  Why the evolutionary approach was delayed so long, I don’t know.  The Introduction to Fauna Hawaiiensis by Perkins did have some excellent evolutionary observations, as did the Introduction to Insects of Hawaii by Zimmerman, and I am much indebted to them for ideas.  They had very little to say about plants—understandably, since both men were zoologists.
   The lack of evolutionary interpretation of island plants by botanists was not only amazing, it was, in a sense, an inspiration.  The great stories of adaptive radiation in the Hawaiian silversword complex, the Echiums of the Canary Islands, and many other plant groups had not been told.  So I had enormous opportunity for hypothesis-making and interpretation, but in a number of cases, I had to invent the frameworks myself.  That was the case with loss of dispersal in island plants.  In my early years of field work on the Hawaiian Islands, I began to notice that a number of fruits and seeds were larger than those of their continental relatives.  Were the same evolutionary scenarios operating in island plants as in island birds and insects?
   Large fruits and seeds are not at all unusual in continental forests.  Oak acorns are rather large, and in the tropics, such seeds as those of the avocado, the mango, and the Brazil nut show a large investment in food storage in seeds.  And for a good reason: seeds of forest trees usually germinate in shady places, so stored food permits the stem of a seedling to reach into better-illuminated levels—every centimeter upwards counts!—to where the seedling can get enough light to photosynthesize sufficiently so that it can continue growing.upward.  One can find instances where seed and fruit gigantism occurs on species of mountains—“sky islands”—but not in their downslope relatives on continental areas.  Darwinia in Western Australia is just such an example—more about that in the captions of the pictures.
   Large seeds and fruits don’t get to islands, except for some beach plants (coconuts, for example).  Beach plants, with rare exceptions, always stay beach plants and don’t evolve into inland plants, so they couldn’t account for large seed size in plants of inland areas. 
   One striking example I saw involves Alectryon of the Sapindaceae.  This genus is present on continental areas adjacent to the Western Pacific and on various islands in the Pacific.  The seed size of the continental Alectryon species, such as A. subcinereum from Australia, is 5 mm or less.  Seed size in the two Hawaiian species of Alectryon is about 50 mm—10 times the diameter of seeds of a continental species!  The irony is obvious: the largest seeds in the genus occur on islands remote from the source area.  Such minimally dispersible seeds on the most remote islands!  A situation like this begs for an explanation.  How can anyone look at such a situation and not want to answer the question it poses?  But Alectryon is just one genus, and one genus might be enough to generate a hypothesis, but not enough to generate circumstantial evidence sufficient for a convincing explanation.
    Stages in how it happens.  [ PDF ] The genus Bidens is represented by numerous species on the Hawaiian Islands, and most amazingly, these represent various degrees of loss of dispersibility and these degrees correlate with ecological factors.  A diverse and nuanced picture of loss of dispersibility thus emerges.  Bidens pilosa can be taken as the probable ancestral type of Bidens that enteredthe Pacific from the New World.  Bidens pilosa has achenes bearing two awns, and these are retrorsely barbed.  The body of the achene has upwardly pointing stiff pointed hairs.  Thus, no matter which way a Bidens achene enters feathers, fur, or socks, it lodges there securely.  In the Hawaiian lowlands and in openings in dry forest as well as in localities near the coast in the, there are a few native Hawaiian Bidens species with achenes similar to those of Bidens pilosa, although with short awns (B. menziesii). Most Hawaiian Bidens species, however, have no awns.  They may have various degrees of hairiness on achenes; a number of species have few or no hairs.  Some species have achenes coiled into one cycle (B. micrantha) or several gyres (B. torta).  In fact, Asa Gray set up the genus Campylotheca to denote those species with curved fruits (which is what Campylotheca infers in Greek).  And then there are some species with large hairless achenes that have no true awns, such as Bidens macrocarpaBidens macrocarpa occurs in shady places on ridges of the Koolau Mountains, Oahu.  In these places, the large achenes contain large seeds with more food storage—suited to growing in shady localities.  Fitchia reminds one of Bidens macrocarpa in having large seed size.  Fitchia is like a Bidens that has evolved over a longer period of time, more perfectly suiting the shady forest habitat.  All of the native Hawaiian species of Bidens have lower dispersal ability than Bidens pilosa, and therefore exemplify loss of dispersibility, although they still have some dispersal ability. 
   Some explanations [ PDF ]. One should keep in mind that on oceanic islands, there is a shortage of forest immigrants. Most forest trees migrate poorly to islands.  Weedy plants of open spaces, like Bidens pilosa, bulk relatively large as island immigrants.  There are therefore many opportunities for evolution into wet forest habitats, and loss of dispersibility in Hawaiian plants involves plants of open lands or dry forest taking on the fruit and seed characteristics (larger size, chiefly) one sees more commonly in wet forests of mainland areas.  In species with widespread habitats limited in area, like willows, which grow in streams, seeds tend to be abundant, small, and very efficiently dispersed.  Natural selection maintains those characteristics in plants with few, widespread habitats.  Weedy or invasive plants fall into that category, because disturbed habitats open for relatively short periods of time, due to landslides, etc (in pre-human times).  Greater seed abundance means smaller seed size, and that lowers the chances of survival in shady habitats.  Thus, if selection for seed abundance and excellence of dispersal is lowered, as it is once a plant arrives on an island, the advantages of larger seed size come into play. 
   One can cite something that has been called “precinctiveness.”  A vague concept that essentially means that most seedlings from a plant tend to germinate relatively close to the parent plant, rather than be scattered far from the parent.  The advantage of this is that most of the favorable habitats for the species occur close to the parent rather than farther away from it.  Thus, if a species is adapted to a habitat of limited size, even if it’s not a wet forest habitat, anything that tends to keep seeds closer to the parent plant may be evolutionarily advantageous.  This may be what is occurring in the case of the series of Darwinia species endemic to peaks in the Stirling Range.  These are adapted to very limited mountain tops with particular climatic regimens—no other such places exist outside of the Stirling Range in Western Australia.  Therefore, a Darwinia like D. meeboldii, adapted to a particular mountaintop, is advantaged if the fruits and seeds are larger and lack dried sepals at maturity—such relatively nondispersible seeds would be more densely deposited on the mountaintop and thus would compete better than if, say, half of the seeds rolled downslope to unsuitable localities.  
   Erythrina, the coral tree, offers an interesting example.  The native Hawaiian Erythrina, known as the wiliwili, grows in dry lowland areas.  The areas where it grows are not beach areas, however.  The wiliwili is probably descended from a species like Erythrina variegata, which occurs near shores in Indomalesia.  Seeds of such an Erythrina probably drifted to Hawaiian beaches—certainly Erythrina variegata seeds float.  Some seeds were probably dispersed to dry areas inland of the beaches.  Inland areas are much bigger in extent than beaches, and therefore an Erythrina would be advantaged by reproducing in inland dry places.  By losing the ability for its seeds to float, the Hawaiian wiliwili stays in these inland areas.  Species adapted to beaches rarely do this.  Once a beach species, always a beach species, with rare exceptions like the wiliwili.   Beach species are not merely dispersed by water (which carries them to another beach), they are adapted to the other features of a beach such as saltiness and alkalinity.  If the syndrome of beach-adapted features is kept intact, a beach species can succeed on numerous tropical beaches.  If water flotation of seeds is lost, as in the wiliwili, a species is released to adapt to inland locations and probably eventually loses the entire syndrome of beach-adapted features. 
   Some plants may lose dispersibility in subtle and invisible fashion.  Hawaiian plants are relatively poor in toxic and distasteful compounds that their mainland relatives produce to deter mammalian herbivores, absent on most oceanic islands (in pre-human times).  Production of such compounds is selectively disadvantageous in the island habitat, because deterrent compounds are, on islands, a wasteful expenditure of a plant’s energy budget.  So an island plant lacking a toxic compound, if introduced to a continental area where herbivores deterred by that compound exist, would not compete well.   Hawaiian plants that have shifted into wet forest from coastal localities, such as most Bidens species, would not grow in the dry, open habitats of their continental ancestor.