Click below for larger images and slideshow more...

The female flowers of Pittosporum gayanum have stamens with withered brown tips, devoid of pollen.

The male flowers of Pittosporum gayanum have long styles, but the ovgaries contain no viable ovules. The anthers are large and filled with pollen, however.

Broussaisia is a dioecious Hawaiian genus. A flower from a female plant (left) and a flower from a male plant (right) have been placed together in this picture.

Coprosma ernodeoides (Rubiaceae) is a sprawling shrub with black berries that grows on the floor of the caldera, Haleakala.

A female flower of Coprosma ernodeoides. The stigmas are very long compared to the corolla. Coprosma ernodioides is not only dioecious, it is wind pollinated, two qualities that favor survival where population size is small and pollinating insects are not abundant.

A male flower of Coprosma ernodeoides. The stamens stand well above the corolla, so that pollen will be shed directly into a wind current.

Flowers of a New Zealand Coprosma, C. aff. robusta, are bisexual. Most species of Coprosma in the world are dioecious.

Clermontia montis-loa is a Hawaiian lobelioid with rather large flowers that are purplish and have relatively wide sepals and petals.

Clermontia parviflora is a Hawaiian lobelioid with rather small white flowers.

A hybrid between Clermontia montis-loa and C. parviflora. The flowers of this plant are intermediate in size between those of the two species, and have the color of C. parviflora flowers except that the sepal and petal tips are purplish.

Another plant hybrid between Clermontia montis-loa and C. parviflora. The flowers are intermediate in size and color between the two parent species. Hybrid plants were growing in disturbed soil along the Kulani Road near Hilo. The road margins soon became invaded by weedy species that supplanted the Clermontia colony.

Plantago species occur on many islands of the world. They are favored by the fact that they are wind-pollinated. The flowers are protogynous: the stigmas are receptive before the anthers release pollen. Thus, the flowers first to open in an inflorescence must be pollinated by pollen from another plant, although flowers further down on an inflorescence might receive pollen from flowers further toward the top.

Drepanis funerea, now extinct, has a bill that matches in size and curvature flowers of some of the lobelioids, so it was thought to take nectar from some of them. Although a the honeycreeper finches and meliphagids that once visited lobelioid flowers have dwindled in numbers, the lobelioids still reproduce because their flowers can self-pollinate if they are not first cross-pollinated.

The iiwi, Vestiaria coccinea, can still be seen in Hawaiian forests, and it has been photographed taking nectar from such lobelioids as Trematolobelia and in the process pollinating them; the flowers of the ohia (Metrosideros) are its main source of nectar.

Scaevola gaudichaudiana has flowers less than 2 cm long. Like most other Hawaiian species of Scaevola, it is a shrub with fan-shaped flowers white in color, but with purple lines in the throat.

The flowers of the tree Scaevola glabra are tubular, yellow, and about 4 cm long. Their size, color, and curved shape are adaptations for bird-pollination, and are thereby unlike characteristics of any other species in the family Goodeniaceae, to which S. glabra belongs.

The flowers of Hibiscadelphus giffardianus are greenish, but rosy at the tips where they open. Thus, a bird seeking nectar in them would tend to be oriented to find the openings. The flowers, shown vertically here, are borne more nearly horizontally, so bills of pollinating birds probably angle upwards to enter them. Hibiscadelphus is an endemic Hawaiian genus, probably a derivative of Hibiscus, in which tubular curved flowers have evolved as an adaptation to bird pollination.

A species of Argemone (Papaveraceae) is native to the Hawaiian Islands. Argemone has flowers pollinated by beetles, which chew on petals, stamens, and stigmas; although they destroy some of these structures, in the process they transfer pollen to stigmas and thereby achieve pollination. Flowers that can be pollinated by beetles possess an advantage on oceanic islands, because beetles colonize islands more readily than bees, which are much more precise and reliable of insect pollinators. Compared to mainland floras, in which pollination by bees accounts for a rather large share of the pollination activity, island floras are poor in good pollinating insects.



My paper on “genetic systems” in island floras [ PDF ] was a response to my interest in how small populations on islands probably survive over time, despite the obvious potential danger of inbreeding.  During field work on islands, I repeatedly saw ways in which plants were—or-weren’t—confronting these problems, and decided to summarize what I knew and what could be easily observed.  One weak point of the paper was that I did not experimental work.  Ideally, such a paper would be the result of 20 years of concentrated research.  So, as the introduction to the paper says, it’s more of an invitation for further study, showing what might be found, rather than a definitive statement.  Another weakness of the paper is that I do not present comparable data from mainland floras or florulas that would show to what degree and in what ways plant reproductive systems on islands differ from those on mainland areas.  And still another weakness of the paper is that I had never been trained or mentored in plant reproductive biology and had no credentials in the field, so what I said was likely to be disregarded.  So why did I write the paper and of what value is it? 
   I wrote the paper because I wanted to construct a coherent picture of all of the evolutionary changes that occur on oceanic islands.  I had done that in “Island Life” in an undocumented way, with occasional examples.  But there is a big difference between citing an occasional example and looking at a flora as a whole.  The Hawaiian flora is the ideal flora, of course, because it is relatively old as floras of oceanic islands go (the island of Kauai is more than five million years old, and the Leeward Chain extends the history by about another five).  The islands of the Hawaiian chain were all, before eroding down, ecologically rich, so more colonizers have arrived and have been able to fulfill evolutionary destinies quite diverse in comparison to those of colonists on islands with monotonous ecology.  The Hawaiian flora is an ideal source for finding out how plants deal not only with inbreeding, but with altered pollination conditions.  Pollinating insects that are effective on mainland areas often do not reach islands: long-tongued bees, for example.  Small beetles such as staphylinids play a much larger role in pollination than they would on mainland areas.  I could see that, but to document that would take careful observation of at least a few species in the field.  That still has not been done, yet a sufficient number of island species and island insects are still available, so that such studies could be done.  The reason why studies on pollination biology are not being done is that financial support for studies in pollination biology on islands supposedly is not available.  Perhaps such money would in fact be available if there were workers with a declared interest in this field, but such workers are lacking.  Sometimes particular areas of study go out of fashion in science, and the reason is usually not because those are well studied and need no further attention, nut rather because newer fields attract money and interest.
    The prevailing wisdom (Baker, 1955: “Baker’s Law”) was that self-compatibility was an aid to colonization on islands.  This seems intuitively true, and can be demonstrated in the beach flora of islands quite readily.  In the beach flora, self-compatibility that leads to inbreeding isn’t a problem, because there are other populations of those species, and seeds from those other populations wash ashore and can supplant any population that might be failing because of inbreeding.  Beach plants form, in a sense, one huge population that collectively has enough genetic diversity and enough occasional outbreeding to remain evolutionarily viable.  But what about inland populations?  Are all of these self-compatible?  Do all of these self-pollinate?  I said I didn’t think so.  My thought was that for island populations, mechanisms that tended to promote outbreeding were valuable where population size was small.  Thus, adaptations for at least a small amount of outbreeding would be valuable and if not present on arrival of a plant, would be valuable during its evolution.  Thus, long-term residents of islands such as the Hawaiian Islands would be expected to have mechanisms for outbreeding.   Some plants must have arrived in a dioecious condition (Coprosma, clearly C. ernodeoides), a clear violation of Baker’s Law.  How many other violations were there?  Today, we know that the most spectacular case of adaptive radiation on the Hawaiian chain, that of the Argyroxiphium—Dubautia—Wilkesia complex, must have originated from a self-incompatible colonizer (I reported Wilkesia as self-sterile in the 1966 paper, based on growing it in my garden).  This makes sense: a built-in mechanism for outcrossing should give a group of organisms a boost where evolutionary capabilities are concerned.   The short-term advantage of self-pollination for establishment is outweighed by the long-term advantage of outbreeding. 
   When one does field work in the Hawaiian Islands, one sees the result of this.  In Hawaiian species of Pittosporum, one seeds fruits on some individuals, not on others.  If one examines the flowers, one sees why: flowers are effectively male on some plants, female on others.  The development of dioecy is apparently not complete in Hawaiian species of Pittosporum, but it’s so well advanced that one can see that this condition, which enforces outcrossing more effectively than any other sexual condition in flowering plants, has taken hold quite strongly.  Elsewhere in the world, Pittosporum flowers are bisexual.  So, one concludes that considerable progress from hermaphroditism to dioecy must have taken place during evolution of Pittosporum on the Hawaiian chain.  Similar considerations apply to Coprosma.  There is one species with hermaphroditic flowers in Hawaii, but the other Coprosma species there are dioecious.  Coprosma ernodeoides, result of a colonization probably different from that which led to the other Hawaiian Coprosma species.  A relative of C. ernodeoides, C. moorei of Australia, has bisexual flowers, so that colonization might have stemmed from plants with bisexual flowers.   However, most Coprosma species in the world are dioecious, so perhaps at least some of those migrations involved species that were dioecious before migrating. 
   One senses that those who talk about long-distance dispersal may want to think in terms of island colonization events that are begun by just one individual.  Although that very likely occurs, one suspects that many colonizations began with more than one individual.  Any event that might bring one seed to an island, or one individual, might bring flocks of them. So arrival of two seeds, one male and one female, of a plant such as Coprosma, does not seem so preposterous.  In the case of animals, to think that colonizations invariably begin with one female carrying fertile eggs seems improbable, in fact.  Arrivals of such a small number of individuals of a species, consisting of a few males and a few females, seems more likely. 
   There are many other examples of sexual conditions that promote outcrossing in the Hawaiian flora.  Monoecism (male and female flowers on the same plant) does that. Gynodioecism occurs in some plants (Styphelia)—female flowers on some plants, bisexual flowers on others.   A number of flowers are dichogamous (pollen presented before stigmas are receptive, or vice versa).  Although I did not have good data on all species of Hawaiian flowering plants, I estimated that 79.2% of Hawaiian species had some floral mechanism that promotes outcrossing.
   Wind pollination, which often results in outcrossing, was estimated for 19.7% of Hawaiian species.  Grasses and sedges are wind-pollinated flowering plants to be expected almost anywhere in the world.  However, Plantago occurs on Hawaii and on many other islands or island groups of the world.   Thus, almost 90% species of Hawaiian plants are adapted for at least some outcrossing--in some cases, insured outcrossing (dioecism).   One might note in passing that ferns require water for fertilization, so they are not disadvantaged on islands that are reasonably moist.
   The high proportion of flora conditions promoting outcrossing on the Hawaiian Islands is rather amazing, because vectors for pollination (other than wind-pollination) are relatively poor, because of the lack of insects that characteristically account for much pollination on islands.  Thus, if one is assessing the proportion of outcrossing-promoting floral conditions in comparison with those on mainland areas, one cannot do a simple Hawaiian percentage vs. Mexican percentage or the like.  The disharmonic nature of the Hawaiian insect fauna must be taken into account.
   One noteworthy characteristic of the Hawaiian flora is the lack of interspecific barriers.  All of the major genera for which crossability has been tested, such as Bidens, show fertility among the species.  Has evolution in small population sizes favored interspecific fertility, or is there merely a neutral selection for interspecific barriers?  The retention of interspecific fertility certainly is not an evolutionary disadvantage.  In areas that change in relation to geology and age, and in areas with clines from dry to wet, interspecific fertility would be a condition that would permit survival of more individuals, and thus enlarge population size and degree of gene exchange.  One can observe natural hybrids in the Hawaiian flora, such as the Clermontia montis-loa X C. parviflora hybrids on the Kulani Road near Hilo.  Natural hybrids have been reported in some of the Argyroxiphium—Dubautia complex—despite marked differences in the appearances of the species.  We do not know to what degree lack of interspecific barriers has helped in rapid evolution of flowering plant groups on islands or in their long-term survival on islands.  Natural hybridization is very common in the New Zealand flora (a fact which may relate to the glaciated landscape of New Zealand).  .
   Adaptation to new pollinators is one of the interesting features in the Hawaiian flora.  The fit between the curved flowers of the lobelioids and the bills of honeycreeper finches such as Vestiaria or Drepanis has long been noted.  How much of that adaptation has taken place on the islands is uncertain, of course, but some of the shift in shape of both flowers and drepanid bills may be postulated. One should also mention the melphagids, which, although rare today, were birds with sickle-shaped beaks once rather abundant in Hawaiian forests.  There is clear evidence in the case of some Hawaiian plants.  Notable among these is Scaevola glabra, which has relatively large, tubular yellow flowers.  These flowers contrast with those of the other Hawaiian species of Scaevola (indeed, they are unlike flowers of anything else in the family Goodeniaceae).  The other Pacific Scaevola species are adapted to fly or bee pollination, and have relatively small fan-shaped flowers, often white with purplish lines leading down to the nectary areas.  Also notable for shift toward honeycreeper or meliphagid pollination  are two genera of Malvaceae, Hibiscadelphus and Kokia.  Beginning with Hibiscus-like flowers (perhaps species of the genus Hibiscus itself), these endemic genera have developed flowers with a curved shape.  In the case of Hibiscadelphus, not only is the shape curved, the flower is much more tubular and less than those of Hibiscus species.
   I might have devoted more time to studying reproductive biology and other aspects of the Hawaiian biota after I finished the book “Island Biology.”  Island biology is enormously interesting to people, why didn’t I continue working in Hawaii?  For one thing, I didn’t have the tools.  I’m not an ecologist.  I could have learned some appropriate methodologies.  However, I was hired to be a comparative plant anatomist in Claremont, not an ecologist or an island biologist.  My field work on island areas of the world was funded on the basis that I had published papers on wood anatomy, so I asked for grants to study more wood anatomy—often on island areas or islandlike areas, sometimes not.  Also, wood anatomy was a much neglected field in the 1960s.  Study of wood of shrubs and of herbs was virtually unknown, I had that field virtually to myself.  And by collecting my own wood materials in the 1960s and 1970s, I knew the ecology of the species whose wood anatomy I studied, and thereby I could add an ecological dimension to my interpretations of each group of woods I studied.  At that time, analysis in wood studies was limited to considerations based on those of Frost and Kribs, who analyzed phylogenetic advance in woods without relationship to ecology.  A lucky circumstance I did not foresee was that at the time I abandoned island studies, techniques for DNA sequencing and cladistic analysis of them were changing rapidly, and by the 1990s, large numbers of students were involved in those pursuits.  Looking for problems to study, they often chose island plants.  So my island books were used as sources of ideas and plant groups by those who had new tools and methodologies.  The timing was just right.  I am fascinated by their new findings in island biology.