miebbareney teers

sin
Seton

sya

ISLAND BIOLOGY

© ISLAND
BIOLOGY

SHERWIN CARLOUIST

with illustrations
by Sherwin Carlquist
and Michael J]. Cole

COLUMBIA UNIVERSITY PRESS
New York and London
1974

Sherwin Carlquist is Horton Professor of Botany at

Claremont Graduate School and Research Associate at

the Rancho Santa Ana Botanic Garden in Claremont,
California.

Library of Congress Cataloging
in Publication Data
Carlquist, Sherwin John, 1930-
Island biology.
Includes bibliographies.
1. Island ecology. I. Title.
OH541.5.18C37 WG pe 5) 73-4643
ISBN 0-231-03562-4

Copyright © Columbia University Press 1974
Printed in the United States of America

To Evelyn Hooker
and Lois Langland

PREFACE

At a time when insular biotas are, in many respects, little known beyond
the floristic and faunistic level, yet when a number of these same biotas
are endangered, apology is not necessary for presentation of materials that
could be called “natural history.” The material in this book will reveal not
how much is known but how very many interesting questions remain to be
asked and answered, largely by field studies.

Past literature is covered, directly and indirectly, but the emphasis is
on recent findings. There is little value in republishing or attempting to
summarize fine works such as David Lack’s Darwin’s finches—such books
and papers speak for themselves. I have attempted to present new material
and new syntheses, or stimulating comparisons, that have not to any great
degree appeared previously in book form. At the time I wrote Island life,*
I had made observations on islands but, aside from anatomical studies, had
not published data relevant to concepts of evolution on islands. I attempted
to supply supporting data for some ideas in a series of five papers, published
under the somewhat cumbersome series title, “The Biota of Long-distance
Dispersal.”

Interest in those five papers suggested to me that, instead of con-
tinuing the series, I should present the material in book form, augmented
by new materials. The content of the papers is represented—in revised,
expanded, updated, and, I hope, improved form—in chapters 1, 2, 11, and
13. Chapter 1 serves as an introduction or précis of my concepts of the

* New York, Natural History Press, 1965.

Vil

Preface

natural history of islands. Aside from those four chapters, the content of
this book represents original material. I have deliberately attempted as far
as possible to avoid, both in content and in illustration, duplication of
material presented in Island life; Hawaii, a natural history; * or various
papers Ihave written dealing with anatomical aspects of insular plants.
Exhaustive citation of literature is not attempted; rather, key papers with
good citations are mentioned. Sources of literature on insular plants and
animals are extraordinarily diverse, and readers will, I hope, forgive me if
certain quite worthy papers have been overlooked.

This book obviously has a bias somewhat more in favor of botany
than zoology; this reflects my training but is also capable of rationalization
on the basis that zoologists have studied insular patterns in a more con-
spicuous way than botanists.

For help during my field work and for other assistance, acknowledg-
ment is due a great number of individuals. I have listed them in my earlier
books and papers, and rather than reprint a long list of names, I wish to
thank again those whom I have previously acknowledged. Those who have
made contributions to this book, and whom I have not previously men-
tioned, include José Cuatrecasas, from whose negatives I printed the
photographs of paramo plants in chapter 14. The bird, plant, and insect
drawings in chapters 4, 8, 12, and 14 are the work of Michael J. Cole.
Photographs have also been contributed by Mrs. Evelyn Humphreys and
Alex George, as credit lines indicate, in chapter 8. The remaining drawings
and photographs are my work. Identification of New Guinean alpine plants
was provided by Pieter van Royen. Specimens documenting my photo-
graphs presented here are located at the Rancho Santa Ana Botanic
Garden, Claremont, California. I am responsible if there are misidentifica-
tions or errors in nomenclature. I have deliberately omitted author citations
on binomials because these are available in the taxonomic works I cite.

* New York, Natural History Press, 1970.

Vill

CONTENTS

onE Principles of Dispersal and Evolution

two Long-Distance Dispersal: Evidence and Implications
THREE Adaptive Radiation in Islands: Definitions and Comments
FouR Adaptive Radiation: Hawauan Islands

FIvE Adaptive Radiation: Canary Islands and Other Macaronesian
Islands

stx Adaptive Radiation: Galapagos, Juan Fernandez, and Smaller
Pacific Islands

SEVEN Adaptive Radiation: New Caledonia and New Zealand
EIGHT Adaptive Radiation: The Flora of Southwestern Australia

NINE Adaptive Radiation: Continental Islands, Fringing Archipe-
lagos, Continental Islandlike Areas

TEN Insular Woodiness

ELEVEN Loss of Dispersibility in Island Plants
TWELVE Loss of Dispersibility in Island Animals
THIRTEEN Reproductive Biology on Islands
FOURTEEN Equatorial Highland Biota

FIFTEEN Selected Island Topics

Index of Biological Names

Subject Index

Ao
487
509
wan)
603
620
634

ISLAND BIOLOGY

[APT
ONE

PRINCIPLES OF DISPERSAL
AND EVOLUTION

The faunas and floras of oceanic islands possess many distinctive charac-
teristics that have long attracted attention. Interest in these lands as “evo-
lutionary laboratories” crystallized during the Darwin-Wallace era and
has not yet diminished. Floristic and faunistic studies of islands have
been frequent (see Blake and Atwood, 1942; Darlington, 1957; Thorne,
1963), although more are needed and will undoubtedly be produced. Dis-
cussions of the possibility of dispersal to oceanic islands, and of whether
particular islands are oceanic or continental, have occupied an inordinate
number of pages and created considerable controversy. Within this out-
pouring there have been a few outstanding contributions dealing with the
problems of dispersal and patterns of insular evolution. ‘Those commend-
able for their comprehensiveness and rational outlook include Gulick
(1932), Zimmerman (1948), Darlington (1957), and Thorne (1963). To a
suprising degree, the assessments which most appeal to me are those that
confirm and extend the concepts of Darwin (1859) and Wallace (1880).
Adherents of alternative proposals may still be found, but there appears to
be a growing consensus among biologists regarding many of the ideas that
I shall discuss.

Although it seems a departure from the Darwin-Wallace line of
natural history, the book by MacArthur and Wilson (1967), The Theory
of Insular Biogeography, is a natural product of these lines of inquiry.
NOTE: This chapter is a revision of a paper (Carlquist, 1966). For permission to

present materials from this paper, acknowledgment is gratefully extended to the
Quarterly Review of Biology.

Principles of Dispersal and Evolution

Wallace, for example, noted faunistic and floristic richness or impoverish-
ment on islands, and offered explanations. That such factors as distance
from source area, ecological richness, and “‘saturation” should be expressed
as a series of mathematical models (as offered by MacArthur and Wilson)
is a natural culmination of ideas advanced by Darwin and Wallace and
extended by such workers as Ernst Mayr and P. J. Darlington, Jr. Recent
papers that incorporate such considerations in analyzing insular patterns,
of various areas and of groups of organisms, are those of Soulé (1966), Wil-
son and Taylor (1967), Greenslade (1968a, 1968b), Simberloff (1969),
Simberloff and Wilson (1969) and Wilson and Simberloff (1969). Similar
papers, applying to situations other than true islands, are cited in chapter
15. Some workers have noted discrepancies between their results and par-
ticular mathematical models (Sauer, 1969; Soulé and Stewart, 1970; Lack,
1971). Model construction is derived from actual situations viewed
(usually) collectively, and cannot be expected to predict the fauna or flora
of any given island. Although the interest and intrinsic clarity of mathe-
matical models of insular situations are great, the appropriate ways for
biologists of various types to use them are open to question. While the
pathway from natural situations to mathematical models is clear, applica-
tion of a model to a particular region or taxonomic group 1s not always
justifiable, because any real situation is replete with complicating factors
that a model obviously cannot predict.

Reasons for presentation of this material at this time are several. In-
terest has shifted recently from a primarily floristic-faunistic outlook to a
concern for evolutionary processes on islands. Principles that have emerged
clearly, as well as those that can only be tentatively enunciated, are re-
viewed here because a foundation for further research seems needed and
because a critical sifting of past hypotheses and a summary of recent work
may prove helpful. An earlier generation of biologists stressed the amazing
speciation that has occurred on oceanic islands. Today’s biologists are in-
quiring into the background of this speciation: its origin, direction, and
underlying mechanisms. The basic mechanisms of evolution on oceanic
islands are the same as those on continental islands, and no new “laws”
are needed; yet the direction that evolution takes on islands, and the
products that result, are often quite distinctive, at least modally. Under-
standing of these evolutionary concepts must be based on sound principles

2

Principles of Dispersal and Evolution

of dispersal and biogeography. Therefore these topics as they apply to the
waif biotas (establishments via long-distance dispersal) must be reviewed,
however briefly. For example, the peculiar rosette trees of the Juan Fer-
nandez Islands (see chapter 10) must be interpreted as relicts by those
who see long-distance dispersal as an inadequate explanation for the Juan
Fernandez flora, and such individuals invoke short-distance dispersal across
land bridges, or cite continental drift. Those who can envisage these
islands as being recent and as populated by means of long-distance disper-
sal can, on the contrary, regard the rosette trees as innovations—the
products of recent evolution.

Literature on the evolution of island biotas is widely scattered among
zoological, botanical, evolutionary, ecological, and other journals and
books. Too often, consideration is limited to patterns of only one group,
or only animals, or only plants. There is a definite need for integrating the
data from various disciplines. Information from one field not only often
proves valuable to studies in other fields, but also must eventually be
coordinated into larger patterns.

The biota that is characteristic of oceanic islands is, ironically, not
always found on such islands and may occur in other situations as well.
Oceanic islands near continents may show an essentially continental pat-
tern. A completely glaciated island will bear a biota that is entirely oceanic
in character (Fleming, 1963). Older freshwater lakes exhibit a full range
of insular features (see chapter 9), and equatorial highlands (chapter 14)
are definitely worth consideration as “nearly insular” situations. A remote
oceanic island may be too new or ecologically too poor to have shown the
evolutionary patterns that will be described. Despite incongruities such as
these, the concept that there are oceanic and continental islands, with
differences reflecting their geological histories, still seems useful and is
generally accepted (Fosberg, 1963). No scheme for classification of islands,
can hope, however, to reflect unexceptionably both the geological history
and the nature of the biota.

Most oceanic islands are volcanic, and volcanic islands are, on a
geological scale, relatively short-lived structures. Therefore, true oceanic
islands display clearly the early and middle stages in the phylesis of a waif
biota, but such islands may vanish before evolutionary products are well
advanced (for example, before they are differentiated to the level of dis-

3

Principles of Dispersal and Evolution

tinct families). Later stages in the evolution of biotas established subse-
quent to long-distance dispersal are often shown instead on old continental
islands. For example, New Caledonia and New Zealand have been isolated
for a very long time—since the Cretaceous, if not much longer. Waif
arrivals on these islands have had a much longer time in which to evolve
than have arrivals to the Hawauan Islands, which are probably no older
than the Miocene. The discerning biologist can, in most instances, detect
which elements in the New Zealand biota are waif immigrants, and he can
follow their evolutionary patterns separately from those that probably are
relicts of an era of continental interconnection or near-interconnection.
Among the relicts are conifers and primitive flowering plants, as well as
the reptile Sphenodon and the frog Leiopelma. With an extended period
of isolation and with a sizable land mass available for occupancy, the waif
biota of New Zealand has been able to evolve extensively, to show waif
evolutionary patterns better than such biotas on most oceanic islands.

It is ironic that the evolutionary products of waif biotas are perhaps
best studied on old continental or continental-like islands; nevertheless it
is possible to cite oceanic islands on which insular patterns are well ful-
filled. Among these, the Hawaiian Islands surely deserve first place,
although part of the evolutionary history of Hawaiian organisms may have
taken place on adjacent islands that are now atolls (the Leeward Islands
of the Hawaiian chain) or now-vanished adjacent islands. The Hawaiian
Islands are favored by ecological richness, extreme remoteness, and (for
oceanic islands) relative oldness. The Galapagos Islands, although famed
for their insular biota, are relatively young, relatively poor ecologically,
and relatively close to continental areas. With the exception of the Dar-
win’s finches, the Galapagos biota best serves to show earlier stages in the
evolution of insular groups (see chapter 6). Completing this spectrum
from old to young waif biotas, the earliest stages in evolution are perhaps
best studied in animals and plants that have been introduced to islands by
man. Mice have proved good examples (Berry, 1964).

Some persons will note that waif biotas are represented in noninsular
situations as well: mountaintops, especially those of equatorial highlands
(see chapter 14), lakes (see chapter 9), caves (see chapter 15) and the like.
This is true. Islands differ in some degree from all of these situations by

4

Principles of Dispersal and Evolution

offering, at least in some cases, a wider gamut of habitats (shore to alpine,
aquatic to xeric) instead of a single extreme habitat (caves) or one that
grades off into continental patterns (the lower limits of equatorial high-
lands). Some noninsular waif biotas can be very informative, however,
and are among the examples cited in this book. I cannot emphasize too
strongly that aspects of insular patterns permeate continental situations.
Thus a waif biota can occur wherever arrival is by means of long-distance
dispersal and where isolation is high for a prolonged period of time. The
assumption made here that waif arrivals-on continental islands can be
distinguished from preinsular arrivals will no doubt be challenged. Means
of separating the two groups are inferential but should improve as our
knowledge of waif-biota characteristics becomes surer. I believe that most
groups on an old continental island can be thus discriminated and that we
should aim for such discrimination. The basis for the distinction is now,
and will continue to be, largely a comparison with oceanic islands. For
example, if conifers and flowering plants with numerous primitive charac-
teristics are subtracted from the New Zealand flora, and if differences in
climate are taken into account, the composition of that flora is remarkably
similar in families and even in many genera to the Hawaiian flora.

Hypotheses regarding the dispersal and evolution of waif biotas are
presented here as a series of principles, or, if one prefers, topic sentences.
The number of such principles could be enlarged or reduced, and ideas
placed under some headings could just as well have been entered under
others. Many of the ideas are closely related and lead me to conclude that
there is an “insular syndrome” which derives from (1) difficulties of long-
distance dispersal; (2) isolation after establishment; (3) ecological oppor-
tunities—“‘vacant niches’; and (4) to a much lesser extent, the modera-
tion of a maritime climate. The consequences of these factors are, how-
ever, manifold. Some of the principles stated in this chapter will appear
to be restatements of the obvious, but others will represent views that
have not hitherto been discussed to any appreciable extent in the litera-
ture on waif biotas. For definitions to classify islands as continental or
oceanic, the reader can consult Darlington (1957), Carlquist (1965), or
Hubbell (1968).

Principles of Dispersal and Evolution

Principles

1. Disharmony in composition of an insular biota is considered a prime
source of evidence for the occurrence of long-distance dispersal.

The distinction between harmonic and disharmonic biotas is a simple
one: a harmonic flora or fauna contains a spread of forms with poor to
excellent dispersal ability, whereas waif biotas will result from only the
more easily dispersed end of the spectrum. An excellent criterion for
dispersibility is the maximum gap of salt water that may have been crossed
by natural means by particular groups of plants and animals. A distance
measured from a continental area to an oceanic island is the most reliable.
Admittedly, the gap crossed may have been greater, because the conti-
nental source population may have been inland, or narrower if a strait
has widened since dispersal occurred. A rating of dispersibility was de-
veloped by Darlington (1957) for animals, and has been extended to plants
by Carlquist (1965). Within taxonomic groups, dispersibility is related to
size of disseminule, desiccation resistance, cold resistance, and other fea-
tures that can vary widely within a group and that depend, of course, on
the nature of the vector. The idea that particular groups have differential
thresholds, or hurdle-values (approximate maximum limits of dispersibil-
ity), is basic, however, to the concept of disharmony.

A demonstration of disharmony is rendered more difficult by the fact
that some regions, insular and otherwise, have biotic depauperation for
ecological reasons. However, poorer faunas or floras are usually deprived
of species with good dispersibility as well as of those with poor dispersibil-
ity. By coincidence, some poorly dispersing groups also happen to be re-
stricted to ecologically “good” situations (e.g., some conifers ).

Good dispersibility is often correlated with ability to occupy pioneer
habitats—a reasonable correlation because pioneer habitats are widely scat-
tered (e.g., beaches) and often open for occupancy suddenly, whereas
stable habitats, such as rain forests, are limited in extent and do not change
rapidly. The possibility of confusion between biotic depauperation owing
to distance and that due to ecological factors is great, and has caused
some workers, who are unwilling to make a distinction, to reject the con-
cept of disharmony and thus also that of long-distance dispersal. The

6

Principles of Dispersal and Evolution

alternatives they must therefore select (hypothetical land bridges or con-
tinental drift) are far more unlikely in the case of oceanic islands—and
even in situations other than oceanic islands (see chapter 2). Strong sup-
port for the concept of disharmony comes from the fact that it occurs not
once, but in as many replications as there are oceanic islands. ‘The limit of
dispersal ability of snakes is shown by their absence not only from Samoa,
but also from Hawaii and the Society Islands. Metrosideros (Myrtaceae)
is a tree that has reached all the major high islands of the Pacific, starting
from an Indo-Malesian * source or sources. Likewise, lizards (once thought
not to reach Samoa | Darlington, 1957]) may be native on Hawaii and
Easter Island.

A continental island can harbor both harmonic and disharmonic
elements. New Zealand, Fiji, and New Caledonia are floristically notable
for old harmonic elements of poor dispersibility (e.g., Araucariaceae), pre-
sumably derived from times when this island arc was connected or nearly
connected with an Antarctic route, perhaps in Cretaceous times (Thorne,
1963). Since that time, waif elements have also populated these islands.
They bulk larger, presumably both by virtue of continual immigration
(with the excellent “invasiveness” which that implies) and by the diver-
sification of these arrivals.

An oceanic island, by definition, will have only a waif biota. Some
biologists tend to overemphasize the difficulties of dispersal for groups with
which they are most familiar. For oceanic islands, however, there can be
no exceptions to overwater (including oceanic drift) dispersal. Not gen-
erally appreciated is the tendency for loss of dispersibility after arrival (see
principles 16 and 17, and the detailed accounts in chapters 11 and 12).
Consequently, the present dispersibility of any given insular species cannot
always be used as a basis for estimates of dispersibility. The pivotal con-
troversy, as to which islands are continental and which are oceanic, seems
to be dwindling. Only a few islands are still difficult to interpret, and the
growing consensus regarding the interpretation of most islands suggests

* “Malesia” is used in this book to connote the Malay Peninsula and all islands
north of Australia eastward to the easternmost Solomon Islands. “‘Indo-Malesia”
includes Malesia plus Southeast Asia as far west as India. The terms Indo-Malaysia
and Malaysia as used in older works must be abandoned because of formation of the
nation of Malaysia.

Principles of Dispersal and Evolution

that there is likewise increasing agreement on some of the principles dis-
cussed here. For a few islands that seem to have been previously connected
to or closer to the source areas, the time of connection may remain prob-
lematic, as may the degree of separation if always separated and the pos-
sibility of formerly larger or smaller area. .

2. Positive adaptations for long-distance dispersal and for establish-
ment are the key to disharmony, and disharmony is thus not a negative
concept. All elements in disharmonic biotas such as those of oceanic
islands are capable of long-distance dispersal or are derived from ancestors
that were capable of it.

Organisms that are clearly adapted to dispersal across very long or
unlimited distances include spore-bearing plants, strand plants, and some
birds and insects (see chapter 2). Species with virtually unlimited dispersi-
bility (e.g., strand plants) are a constant element on islands; they provide
no problems in interpretation, nor do they usually develop evolutionary
patterns of interest. The strand flora usually remains a strand flora, and
rarely evolves into montane species. Exceptions include Erythrina and
Acacia (Rock, 1919b) on the Hawaiian Islands. For further information
on strand plants that have evolved into inland sites, and the reasons why
this has occurred in a limited number of instances, see chapter 11 (also
certain halophytic elements in the Australian flora, chapter 9). Although
dispersal by seawater flotation has contributed little to montane floras of
islands, one must remember that some plants for which seawater dispersal
may seem doubtful are, in fact, capable of it (e.g., Gossypium |[Stephens,
1958a, 1958b, 1963, 1964], and Lagenaria {Whitaker and Carter, 1954,
1961]). In the case of Gossypium, the Tahitian and Hawauan cottons are
endemics, and there is no evidence of derivation from cultigens; in any
event, the same principles apply to other, noneconomic Malvaceae.

Positive adaptations for dispersal by virtue of habits, habitats, and
resistance to salt water occur in lizards. One widespread gecko, Hemudacty-
lus garnoti, is a uniparental species (i.e., its eggs hatch without fertiliza-
tion). This characteristic makes possible the dispersal to Hawaii and else-
where, because any single individual can start a new population. Two
species of skinks have crossed the Pacific from Asia to islands off the
South American coast, very likely without the aid of man. One of these,
Ablepharus boutoni, lives on sea beaches; it nests in epiphytes on nearby

8

Principles of Dispersal and Evolution

trees that might often be washed out to sea. One New World gecko, of
wide distribution, lays its eggs communally under bark or in tunnels of
trees, and lizard eggs are notably resistant to seawater. For further infor-
mation on the role of oceanic drift, see chapter 2.

Rafting is probably most effective over relatively short distances but is
doubtless responsible for cases of land-animal (mammal and _ reptile)
transport (for reviews and observed examples, see Darlington, 1938;
McCann, 1953; Wheeler, 1916; Zimmerman, 1948).

For the remaining species, birds or winds must be the vectors. Gres-
sitt and his co-workers are to be congratulated for extensive aerial trapping
experiments which demonstrate clearly not only that aerial transport of
insects and spiders is likely, but also that the species caught represent
groups in the same proportions as they occur in insect faunas of oceanic
islands (Clagg, 1966; Gressitt, 1956, 196la; Gressitt, Leech, and O’Brien,
1960; Gressitt, Leech et al., 1961; Gressitt and Nakata, 1958; Gressitt,
Sedlacek, Wise, and Yoshimoto, 1961; Gressitt and Yoshimoto, 1963;
Harrell and Holzapfel, 1966; Thornton and Harrell, 1965; Yoshimoto and
Gressitt, 1959, 1960, 1961, 1963, 1964; Yoshimoto, Gressitt, and Mitchell,
1962; see also the earlier literature cited in these papers and in Richter,
1970). Actual dispersal events can be witnessed with such rarity that these
experiments are as close to definitive proof as can be expected. The con-
tention of such workers as Visher (1925), Ridley (1930), Setchell (1926,
1928, 1935), Andrews (1940), Zimmerman (1948), Myers (1953), and
Fosberg (1963) that aerial dispersal is a reality over extended distances for
very small or air-floatable disseminules seems to be justified.

Dispersal of birds with migratory habits is easily explained. Land birds,
however, are notably sedentary (Mayr, 1940). As Mayr notes, however,
involuntary flight owing to storms can be invoked to explain distribution
of land birds in Polynesia. Birds that migrate moderate distances in a
seasonal pattern as different food sources become available might be
better candidates for long-distance dispersal than extremely sedentary
species. Some workers have a tendency to believe that long-distance dis-
persal of land birds is such a rare occurrence that it simply never happens.
Rare occurrences do suffice to explain dispersal, however (see principle 4).
The successful colonization of the Hawaiian Islands by only a few groups
of Jand birds is witness to their sedentary nature but also to the likelihood

9

Principles of Dispersal and Evolution

that very infrequent exceptions have sufficed for establishment. Lack
(1971) stresses that the absence of land-bird groups on islands may not
result so much from difficulty of dispersal as from lack of suitable ecologi-
cal sites; he contends that islands that appear by MacArthur and Wilson’s
(1967) predictions to be “unsaturated” may, in fact, have no open ecological
niches. For a further discussion of dispersal of land birds to islands, see
chapter 2.

Transport of fruits, seeds, and eggs externally or of fruits and seeds
internally by birds must be regarded as the dispersal mechanism of many
species (Ridley, 1930; Gulick, 1932; Zimmerman, 1948; Falla, 1960; Hold-
gate, 1960; Wace, 1960). Observational and experimental evidence for
these modes of transport is increasing. Barbed or viscid fruits or seeds have
been observed attached to feathers. The experimental work of Proctor and
others (see chapter 2) shows that internal transport of seeds by shorebirds
over long distances is entirely within the realm of possibility. In fact, dis-
persal of fruits and seeds by migratory birds is probably the method by
which most groups of angiosperms arrived on the high islands of the
Pacific (fig. 2.7). We can see this by analyzing dispersal mechanisms of
native floras of islands, as Skottsberg (1928) and Lems (1960b) have done.
Continental disjunctions, such as the famed central Chile-western North
America pattern (Constance et al., 1963) should be regarded as oppor-
tunities for discovering modes of long-distance dispersal, not as mysteries.

Ability for long-distance dispersal is shown in the Pacific by the points
at which genera terminate, in relation to source areas. Smith (1955) has
shown that 101 genera of flowering plants have their eastern terminus on
Fiji, the most easterly major island within the line delimiting andesite
rocks (which suggest the former presence of larger land areas). Of the
genera that do continue beyond Fiji—to Samoa, for example—many (or
endemic genera derived from them) extend all the way to the Society
Islands or Hawaii, a fact that suggests that the flora of Polynesia is indeed
the flora of long-distance dispersal.

Present dispersal configurations of island plants and animals may be
less favorable than those of their ancestors (see principles 16 and 17) a
possibility neglected by some authors (e.g., Remington, 1971). In some
cases, one would do better to investigate the present dispersal mechanisms
of mainland relatives of island plants and animals rather than those of the

10

Principles of Dispersal and Evolution

insular representatives themselves. As Williams (1969) and Tryon (1971)
note, only a few of the “available” species in a source area are preadapted
for successful dispersal and colonization. Brown and Alcala (1970) have
presented ingenious demonstrations of these phenomena for Philippine
reptiles.

3. Long-distance dispersal is probably not achieved invariably by only
a single individual or disseminule, but may in some instances be expected
to result from simultaneous (or nearly simultaneous) introduction of more
than one individual or disseminule. This could still be said to constitute a
single “dispersal event.”

There is a tendency to believe that introduction of a single seed can
and will suffice to establish a new plant species. ‘This possibility cannot be
denied, but several circumstances point to additional possibilities. The
requirement of more than one seed holds true for dioecious plants (see
principle 20 and chapter 13). A single individual of a plant also seems less
likely to yield progeny destined for long-term success, owing to its limited
content of genetic variability (see principle 10). Animals require at least a
breeding pair or a gravid female. A single gravid female would introduce
some heterozygosity, but a flock of adults would introduce more and would
be expected to be a better source of genetic variation.

In addition to these relatively demonstrable reasons, one can suggest
other cogent ones. The frequent observation of stragglers among migratory
birds (Munro, 1960; Lack, 1971) suggests repeated introduction, although
repeated introductions are not indicated in the West Indian anoline
lizards (Williams, 1969). Repeated stragglers succeeded recently in estab-
lishing a breeding colony of purple gallinules on Tristan da Cunha (Rand,
1955). Repeated arrival of straggling monarch butterflies on Canton Island
was observed by Zwaluwenberg (1942). Shifts in ocean currents may
throw onto beaches unprecedented deposits of seeds of a previously
absent species. Such populations of Mucuna seeds were observed on New
Zealand beaches in 1956 (Mason, 1961). Mucuna seedlings that do not
survive can be observed from time to time on the Leeward Islands in the
Hawaiian chain (fig. 2.6, bottom right). Rafting would tend to bring a
small population rather than one individual, as with the ants observed by
Wheeler (1916). The distribution pattern of some genera that occur on
virtually all the high Pacific islands, such as Metrosideros, suggests that

1

Principles of Dispersal and Evolution

one and only one introduction of a single establishable seed to each of
these islands is, viewed as a statistical matter, rather unlikely.

The rapid colonization of the Krakatau islets by animals (Dammer-
mann, 1948) and plants (Docters van Leeuwen, 1936) after the volcanic
devastation of Krakatau suggests introduction of many disseminules of a
species. ‘This is also the case on Surtsey, the newly emerged island near
Iceland (Einarsson, 1967) and on the defaunated islets in the Florida
Keys (Simberloff, 1969; Simberloff and Wilson, 1969; Wilson and Simber-
loff, 1969). The equilibrium theory of MacArthur and Wilson (1963),
which seems well justified, depends on a continuous rate of immigration
rather than a few random accidents.

Arrival of a gravid female of an animal species, or of a single seed of
a self-fertile and self-pollinating plant may seem a favorable condition for
establishment on islands—but are these absolute requirements? That all
angiosperm immigrants to islands self-pollinate seems unlikely (see chap-
ter 13). Introductions are probably not on a yearly or continual basis
except for oceanic drift or other unusually efficient dispersal means. Intro-
duction may well result from violent occurrences such as storms that might
deposit quantities of alien jetsam on an island.

4. Among organisms for which long-distance dispersal is possible,
eventual introduction to an island is more probable than nonintroduction.

This principle, together with appropriate calculations, was clearly
enunciated by Simpson (1952). Calculations demonstrating probabilities
of this sort had been offered earlier by Matthew (1915) and Darlington
(1938). The calculations of Fosberg (1948) on the origins of the Hawaiian
flora suggested the inevitability of the outcome. Attention is called to
Fosberg’s estimate that the Hawaiian species of no fewer than 23 genera
of flowering plants have resulted from two independent introductions,
whereas three or more introductions are hypothesized for Hawaiian species
of 11 other genera (and the reader must keep in mind that Fosberg delib-
erately calculated only the absolute minimum number of required intro-
ductions). Similar calculations probably could be made of hypothetical
immigrants in other groups of organisms and in other waif biotas.

Odd chance distributions do, of course, occur. Lepinia (Apocynaceae),
which occurs on Tahiti and Ponape, is probably an example (Fosberg,
1963). Peculiar distributions of this sort are discussed by Falla (1960),

7

Principles of Dispersal and Evolution

who suggests that habits of birds which are possible vectors may be respon-
sible. The distribution of Bidens in the Pacific suggests the operation of
chance (Carlquist, 1965). Other disjunct distributions on a worldwide
basis that seem to suggest chance include Pilostyles (Rafflesiaceae) and
Coriaria (Coriariaceae). Drastic disjunctions sometimes prove to be the
result of incorrect taxonomic interpretation. The supposed occurrence of
Lipochaeta (Asteraceae) on the Hawaiian Islands, Galapagos Islands, and
New Caledonia now proves to be a case of three different genera that had
not been distinguished because of insufficient study (Harling, 1962).

5. Elements are present in proportion to not only dissemination
ability but also to establishment ability. Establishment ability is a charac-
teristic of every species and group but cannot be viewed except in relation
to the ecological conditions provided by the recipient island.

Difficulties of establishment seem much greater than those of trans-
port. One can safely say that among successful introductions of dissemi-
nules in good condition, only a fraction manage to survive and establish a
continuing colony. Animals with wide food preferences and plants of
pioneer habitats and easily satisfied pollination requirements appear espe-
cially favored. In many cases, greater tolerance for aridity seems an advan-
tage—but not universally. Carnivores are less favored than herbivores, since
elements in a food chain must become established on an island before
secondary ones can be. Perhaps this explains why the insect fauna of Tris-
tan de Cunha contains an exceptionally high proportion of phytophagous
species (Brinck, 1948). According to Gressitt (1961b), insects that inhabit
plant debnis or are leaf miners or wood borers are especially successful on
oceanic islands. “Weedy” or “versatile’’ species seem to make more suc-
cessful colonists (Williams, 1969; Baker and Stebbins, 1965).

Many forest trees or other forest plants can become established only
when soils suitable for them are present. An exception is the chief forest
tree of the Hawaiian Islands, Metrosideros, which can grow and eventually
form a forest on new lava flows (Robyns and Lamb, 1939; Skottsberg, 1941;
see also fig. 2.1). Lava pioneers are especially likely to succeed on oceanic
islands and, by no coincidence, many plants of oceanic islands fall into
this category. Rangitoto, a recent volcanic island in Auckland harbor, New
Zealand, is vegetated by plants of the New Zealand flora that can grow on
bare lava. Most genera found on Rangitoto are also found on Hawaii and

13

Principles of Dispersal and Evolution

Tahiti (Carlquist, 1965), or are very much alike (e.g., the araliads Notho-
panax in New Zealand and Cheirodendron in the Hawaiian Islands).

Ecological requirements for establishment are numerous, and unsuit-
able conditions must screen out many potentially successful colonists. Ex-
amples on Canton Island, described by Zwaluwenberg (1942), show this
phenomenon clearly. Wilson (1959) has shown that, in the Pacific, ants of
open and marginal habitats have been the migrants to new island areas.
After arrival on islands, they evolve into more specialized interior sites.
Establishment may fail because an apparently suitable niche is preempted
(Lack, 1969, 1971; Williams, 1969). What seems to be an “unsaturated”
condition on an island may not be: as we understand better the ecology
of island species, we often discover that available resources are being fully
utilized. Another reason for failure, actually observed in the field, is that
animals suited ecologically for a particular island may not find there an
adequate food supply to maintain a breeding population (lack of predation
may also be instrumental here). Examples of otherwise successful animal
immigrants “overeating” their resources are given in chapter 13. Problems
of island colonization have recently been summarized by MacArthur
(UTA

Animals have the advantage of being able to seek suitable environ-
ments upon arrival, whereas plants, to survive, must be deposited upon
locations suitable for their growth and reproduction. Moreover, suitable
pollinating agents must be present, or else self-pollination must be pos-
sible. The claim that hermaphroditism or self-pollination is advantageous
at the time of establishment (Baker, 1955) seems readily understandable,
but species with long-term tenure on islands seem to show the value of
outcrossing (see chapter 13).

Perennial plants have been alleged to possess an advantage over
annuals because their longevity increases the likelihood of securing sufh-
cient pollination to produce seeds to establish and maintain a species
(Wallace, 1895), but alternative explanations for the lack of annuals on
islands may take precedence (see principle 14 and chapter 10).

6. Migration to islands is governed by probability, and ordinary con-
cepts of migratory routes and biological provinces do not apply well to
many islands.

The statement above is a criticism of the division of islands (chiefly

14

Principles of Dispersal and Evolution

in the Pacific Ocean) into provinces, subprovinces, and the like. Plants of
equatorial highlands, such as the Afroalpine flora, are also anomalous in
this respect (see chapter 14). The concepts of biological provinces and
migratory routes are primarily derived from, and therefore best applied to,
continental situations. The best systems of biological provinces for islands,
such as those of Thorne (1963) and the workers he has cited, still remain
unsatisfying to me for the following reasons:

1. Groups of plants and animals on a particular island (or equatorial
highland “‘island’’) are present roughly in proportion to the distance from
suitable source areas. Many islands have acquired their biota from several
sources. Can the Hawaiian Islands logically be put into an Oriental prov-
ince when appreciable portions of their flora and fauna are American in
origin (Fosberg, 1948)? The diverse sources of the New Zealand flora and
fauna defy categorization into anything but multiple provinces and mul-
tiple routes (Dawson, 1958, 1963; McDowall, 1964). Darlington (1957)
has shown that animals have reached the West Indies from many sources
and have used many ports of entry. These and other examples provide
difficulties that, in my opinion, override whatever merits the province and
route devices may have.

2. Although there is value in emphasizing major faunal regions, such
as those delimited by Wallace’s Line and Weber’s Lines, these concepts
apply to large land areas, not to oceanic islands. For example, the marsu-
pials, which are a chief criterion of the Australo-Papuan region, do not
reach oceanic islands or New Zealand.

3. Criteria for one province differ from those for another. Placental
mammals characterize the Malayan area, whereas biotic depauperation is
the best criterion for Micronesia. In this connection, the Tuamotus must
fall into Micronesia floristically, although geographically they are Poly-
nesian.

4. Each biologist will erect provinces and routes differently from
others because, inevitably, each person will stress some groups not stressed
by others, or will give different weight to historical, climatic, ecological, or
geographic factors. No two biologists appear to agree about the designation
of provinces or routes, nor would agreement necessarily be desirable.

>. The concepts of provinces and migration routes result from
analyses of data but in their final expression do not themselves yield infor-

15

Principles of Dispersal and Evolution

mation. They are a shorthand that cannot be translated. Because there is
no substitute for knowledge of original data, the construction of hypo-
thetical routes and provinces serves only their authors’ ideas. Other per-
sons would be well advised to acquire the original data if they want to com-
prehend distribution patterns on islands.

7. Guyots and other now-vanished high islands or lands more exten-
sive formerly than now may have aided dispersal to oceanic islands as
subsidiary source areas or “‘stepping-stones” but not as dry land bridges.
Continental drift does not explain oceanic island patterns, and even inter-
continental disjunctions are more easily explained by long-distance dis-
persal, in many cases.

One can agree tentatively with Zimmerman (1948) that disseminules
are abundant near a source and become progressively fewer with distance.
If so, the presence of an island chain would improve the dispersal possi-
bilities of a species. That this arrangement is indeed effective is shown by
the Hawaiian Islands. Although the nearest area with any appreciable land
surface is the North American continent, the Hawaiian biota is predomi-
nantly Indo-Malesian. However, the prevailing winds do not seem to favor
immigration from Indo-Malesia, at least at lower levels in the atmosphere
(jet-stream winds at higher levels do seem more favorable). The many
small islands and atolls that lie west and south of the Hawaiian chain very
likely aided, when they were larger high islands, in transmitting plants and
animals to the Hawaiian Islands. Realizing the potential importance of
these vanished lands, Zimmerman (1948) called for the development of
paleontological information about the Pacific basin. Such data are now
rapidly accumulating (Cloud, 1956; Cloud, Schmidt, and Burke, 1956;
Durham, 1963; Hamilton, 1953, 1956; Ladd, 1958; Ladd et al., 1953;
Menard, 1956; Menard and Hamilton, 1963; Stark and Schlanger, 1956).
The authors cited have contributed to a picture of large former archipelagos,
some as early as the Eocene, lying in the mid-Pacific. The maps by Menard
and Hamilton (1963) strongly suggest a dispersal potential for previously
existing islands and should serve to reduce greatly the skepticism concern-
ing long-distance dispersal as a means for populating the Hawaiian Islands
and other Pacific islands.

Stepping-stone islands are potentially less effective as a source for
dispersal than large land masses, because dispersal may be expected to be

16

Principles of Dispersal and Evolution

proportionate to the number of individuals in a source area (Johnson,
1960a, 1960b). Moreover, autochthonous loss of dispersal mechanisms
may lower the dispersal function of stepping-stone islands. That such
islands function as staging areas in dispersal cannot be denied, however.

Within archipelagos, redispersal occurs, often with interesting conse-
quences as a result of successive events of isolation and reinvasion (Hamil-
ton and Rubinoff, 1963; Greenslade, 1968a, 1968b; see also examples from
the Drosophilidae and Drepanididae in chapter 4). Many theoretical possi-
bilities of how archipelagos may affect dispersal have been summarized by
Carlquist (1965) and MacArthur and Wilson (1967). One can say that
any species with enough dispersal ability to bring it to an archipelago can
be expected to succeed in dispersing (but not, perhaps, in establishing)
itself throughout the archipelago (Tryon, 1971). A species may lose its
pioneering ecological characteristics, however (Wilson, 1959).

The implications of long-distance dispersal for continental drift the-
ories are discussed in chapter 2. Briefly, one can say that biotic distribution
patterns must be examined in the light of long-distance dispersal possibili-
ties; work that fails to do so cannot be accepted. Although relictual pat-
terns can and do occur on continents, data now at hand show how long-
distance dispersal has very likely operated in intercontinental disjunctions.

8. The size and systematic composition of insular biotas are deter-
mined by many factors that differ in relative importance from island to
island.

The factors that affect biotic composition are discussed with consid-
erable mathematical sophistication by MacArthur and Wilson (1967).
However, to say which factors are of prime importance is difficult. Even
for particular islands, factors are not easily ranked, although attempts have
been made (Hamilton and Rubinoff, 1963; for a critique, see Carlquist,
1965). Among the factors claimed to be important are island area and
altitude (Darlington, 1957; Tryon, 1971); size, nearness, and richness of
source (Zimmerman, 1948); latitude, climate, island age, and geological
events. Archipelago effects (size of island, size of neighboring island, near-
ness to neighboring island, altitude, altitude of neighboring island) are
manifold (Hamilton and Rubinoff, 1963).

Factors influential in an island’s biotic richness are measures of eco-
logical opportunity, the degree to which barriers to dispersal to an island

Uy)

Principles of Dispersal and Evolution

can be overcome, the number of barriers within an island that can serve
for isolation and therefore promote speciation, and the requirement for a
certain minimal area for maintenance and evolution of a population (see
principle 10).

9. Relicts in the strictest sense are few or absent on oceanic islands,
although every immigrant group has a history, and one can designate older
island autochthones as “recent relicts.”

Good dispersal ability is not entirely restricted to phylogenetically
“advanced” groups. If “primitive” forms have migrated to oceanic islands
but the mainland remnants have become extinguished recently, the insular
representatives appear as relicts. This is probably the case with the primi-
tive flowering plant Lactoris (Carlquist, 1964) and the fern Thyrsopteris
on the Juan Fernandez Islands. ‘These islands do not appear to be ancient,
and the remainder of their flora and fauna contains no relicts. In the West
Indies, the cycad Microcycas and the insectivore Solenodon (together with
its fossil relative Nesophontes) may be considered relicts (McDowell,
1958; Darlington, 1957). Other than the examples cited, no spectacular
relicts can be found on oceanic islands.

If one wishes to think in terms of a “Pliocene relict,” a few might
occur on oceanic islands—perhaps Hesperomannia (Asteraceae) on the
Hawaiian Islands, for example. That genus is mentioned because its closest
relatives are now withdrawn to areas not likely to be source areas, but
mutisioids similar to Hesperomannia might have been more widespread in
western North America (Mexico?) during the Pliocene. With few excep-
tions, however, we should look first at the probability that groups on vol-
canic oceanic islands have close counterparts in likely source areas and are
therefore not relictual.

Insular groups may tend to have a rapid cycle of speciation and extinc-
tion, and the latter stages may be said to contain, in a sense, relict species
(see principle 18).

Why are there so few real relicts among the waif biotas, whereas
continental islands, such as ‘Tasmania, are notable for relictism? One
answer may lie in the fact that the most successful groups in the waif
biotas are those that appear to be evolutionarily upgrade or weedy and
that have a greater degree of genetic momentum. Most antique groups
make relatively poor immigrants, both because of their generally poorer

18

Principles of Dispersal and Evolution

adaptability and because they often have poorer dispersal mechanisms.

10. Immigrant species must overcome the restriction of genetic mate-
rial related to the very small size of the initial population tf effects such
as inbreeding are to be countered; in any case, the portion of the genetic
content of a species that the insular establishment represents will influence
the nature of the resulting insular population.

This is, in effect, a restatement of Mayr’s “founder principle.” The
small size with which waif groups begin is one of the unique features of
insular existence. With a few exceptions, such as the strand flora, an immi-
grant population will receive no new genetic material from the parent
population. If such infusions occurred with any frequency, endemism on
islands would be much lower than it is. Because infusions do occur rela-
tively often among the strand species, endemism among them is in fact
quite low (Fosberg, 1963).

Genetic variability can be increased by mutation after migration. ‘The
degree to which mutations are retained depends not only on the mutability
of a species but also on the ecological opportunity. Removal of introduced
animals from the Channel Islands of California has permitted greater
expression of genetic diversity in the dwarf island fox (Urocyon littoralis)
and the island moth Estigmene (Remington, 1971). Chromosomal poly-
morphism in West Indian populations of Drosophila willistoni is greater
on islands that are ecologically diverse (Dobzhansky, 1957). Metrosideros
polymorpha, the chief forest tree of the Hawaiian Islands, is extraordinarily
diverse and has proved difficult for taxonomists. This multiform tree ranges
from bare lowlands to high bogs, where it is a small subshrub (Rock, 1917),
and it may be, according to the area, in an almost continual cycle of hy-
bridization. Polymorphism characterizes many island species; there may be
a potential value to lack of interspecific barriers on islands (see chapter
13). Notable examples of variable island species in stages of expansion
occur in the land shell Partula (Crampton, 1916, 1925, 1932; Schwabl and
Murray, 1970). There is every reason to believe that physiological and
ecological opportunities play controlling roles. A small area and ecological
poverty are cited by MacArthur and Wilson (1963, 1967) as reasons why
extinction is high on some oceanic islands. A population’s requirement
for a land area large enough to maintain genetic variability and thus to
maintain the species itself is cited by various authors. Darlington (1957)

19

Principles of Dispersal and Evolution

believes that this requirement explains the extinction of large mammals on
Ceylon; it may also explain why the pygmy mammoth, whose reduced
stature may reflect excessive inbreeding, became extinct on the California
offshore islands. Similar size reduction has occurred to a marked degree in
an American bison herd introduced to Santa Catalina Island (R. F.
Thorne, personal communication, 1972).

Weedy immigrants would be expected to have the advantage of high
mutability. Moreover, the broader ecological tolerances of weedy plants
and animals would permit occupancy by more individuals in a given island,
thus increasing potential genetic variability. Smaller body size would also
favor development of larger populations, which similarly could be geneti-
cally more viable.

Berry (1964) has demonstrated in island mice that only a fraction of
the species’ gene pool is present in island immigrants, thus influencing the
nature of the insular population.

11. Rapid evolution of island immigrants is not only possible but fre-
quent. Change after arrival is inevitable.

“Explosive” evolution is demonstrated by various groups that have
had good ecological opportunities. Among outstanding examples may be
cited the Hawaiian Drosophilidae (about 400 species according to C. Elmo
Hardy | personal communication, 1966]) or the Hawaiian species of Cyrtan-
dra (Gesneriaceae); about 130 species of Cyrtandra on Oahu alone are
claimed by St. John (1966).

Actual times have been estimated for some cases of rapid insular
evolution. Five species of endemic banana-constant moths have evolved
in the approximately 1000 years since human introduction of the banana
to Hawai (Zimmerman, 1960). In Lake Lanao in the Philippines, four
endemic genera of cyprinid fish have evolved in 10,000 years or less (Myers,
1960). Endemism has reached the level of families in the fishes of Lake
Baikal (Comephoridae [Kozhov, 1963]) and two Burmese lakes (see
chapter 3).

Biologists are increasingly aware that geological time is short on vol-
canic islands. Evolution must proceed within this span, although to this
length of time can be added that available on an archipelago as a whole,
as well as that on stepping-stone islands, if any.

Factors assisting rapid evolution on oceanic islands include lack of

20

Principles of Dispersal and Evolution

competitors and predators, presence of a wide spectrum of ecological
opportunities, and presence of isolating mechanisms (ridges, etc.) that
favor small, rapidly changing populations. Prominent geographical relief
may have been partly responsible for the remarkable speciation of Partula
(Crampton, 1916, 1925, 1932). Lava flows can subdivide populations and
provide isolation (Zimmerman, 1948), and catastrophic episodes of vol-
canism may well spur evolution. The barriers among islands of an archi-
pelago seem important in promoting speciation of insular groups (Darling-
ton, 1957). Other forms of reproductive isolation among insular species
are possible, however, and should not be overlooked (Bailey, 1956).

When distinct endemic species and genera evolve, one should expect
changes in any part of a plant or animal. For example, it is illogical to
expect that the dispersal mechanism in the extraordinary Hawauan silver-
sword, Argyroxiphium (Asteraceae), is the same as that of the ancestors
of this endemic genus. Possibilities that genes may have pleiotropic effects
should not be overlooked.

12. Situations new to immigrants will dictate their courses of evolu-
tion on islands—if they can cross ecotones into them. Adaptive radiation
is the inevitable result on an island or archipelago where a small number
of immigrants meets a broad spread of ecological opportunities.

The “genus-and-family-poor but species-rich” condition is one reflec-
tion of adaptive radiation on oceanic islands. What is not reflected in taxo-
nomic terms is the tendency for island species to evolve into ecological
niches that would be occupied by a member of an entirely different group
on a comparable mainland area. The absence of mammals on most oceanic
islands has led to the assumption of mammalian roles by birds and reptiles.
For example, the dodo of Mauritius, a large terrestrial herbivore, repre-
sented a sort of avian rabbit or sheep. Phases of adaptive radiation are
related to topics discussed under principles 14 through 17 below, in which
diversification of habits within groups is cited.

Because only a few groups such as the Darwin’s finches (Lack, 1945,
1947; Bowman, 1961) or the Hawaiian honeycreepers (Amadon, 1950)
have been discussed as examples of adaptive radiation, I have attempted
to broaden the picture of this phenomenon by discussing numerous groups
and geographical regions, in chapters 3 through 9. Understanding of adap-
tive radiation may benefit from field knowledge, but in some groups where

Zh

Principles of Dispersal and Evolution

distinctive forms suggest adaptations we have no field knowledge with
which to implement suggestions concerning potential adaptive radiation.

Time is a requisite for completion of a cycle of adaptive radiation. On
some islands, time (as well as other conditions) has been sufficient for
achievement of spectacular radiation. In some cases, not merely one but
several cycles of adaptive radiation may occur within a group, as in the
Hawaiian honeycreepers (chapter 4).

13. An immigrant group that is not faced with, or cannot enter, be-
cause of inherent limitations, a broad spectrum of ecological opportunities
on an island may evolve into one or a few niches.

In addition to “definitive” adaptive radiation, islands may bear por-
tions of a gamut of adaptive radiation. Examples of incomplete or lopsided
adaptive radiation are abundant on islands. These examples usually dem-
onstrate entry into a few new habits or mechanisms—often unusual ones
for the groups to which the particular organisms belong. ‘The Hawaiian
species of Viola and Bidens are remarkable in being herbaceous to woody
shrubs, but they do not form large shrubs or trees (see chapter 10).

Islands may play a role in fostering peculiar adaptations, such as that
of the seaweed-eating marine iguana of the Galapagos Islands, Ambly-
rhynchus (chapter 15.) New dispersal mechanisms (chapter 11) or pollina-
tion mechanisms (chapter 13) may develop in plants. The tool-using Gala-
pagos finch Cactospiza is part of a program of adaptive radiation, but it is
in addition a unique development. Developments in categories such as
these occur on mainland areas, of course, as well as on islands. One can
speculate that examples such as those given represent responses to unusual
conditions on islands and may well have been preserved in the less compet-
itive island situation.

14. New growth forms evolve among plants on oceanic islands. Most
conspicuously, there is a tendency toward increased stature.

The chief changes that seem to occur are from herb to rosette tree or
rosette shrub, herb to shrub, shrub to rosette shrub, and shrub to true tree.
Because these tendencies have not been explored and analyzed to any
appreciable extent since the essay of Hemsley (1885), I have attempted to
document them in chapter 10. The reasons that “insular woodiness” has
occurred on particular island groups and the factors that seem of para-
mount importance in particular localities are discussed at length there.

22:

Principles of Dispersal and Evolution

In addition to viewing insular woodiness from the standpoint of eco-
logical and biogeographical factors, one can view this phenomenon from
the standpoint of taxonomic groups. This, like many other insular phe-
nomena, occurs most abundantly in relatively plastic, evolutionarily up-
grade groups. For example, unusual woodiness of Asteraceae is demon-
strated on the Hawauan Islands (Rock, 1913), the Juan Fernandez Islands
(Skottsberg, 1953), St. Helena (Melliss, 1875), the Canary Islands
(Schenck, 1907; Bérgesen, 1924) and many others. The numerous taxo-
nomic groups (even to the tribes of Asteraceae, independently of each
other) that follow these trends on islands, and the numerous ecologically
suitable islands that demonstrate insular woodiness, are useful in the same
way that replications and control groups are in experimental areas of sci-
ence, and offer equivalent security of interpretation.

Insular woodiness is characteristic of certain continental areas as well
as oceanic islands. Most conspicuous are the alpine zones of equatorial
South America and equatorial Africa. The plants of these areas are dis-
cussed in chapter 14.

Although a few workers have interpreted insular rosette trees as relicts
(Skottsberg, 1956; Lems, 1960a, 1961), this interpretation seems to be a
minority view. Reasons that point to a nonrelict interpretation of insular
and equatorial highland rosette trees seem compelling and are detailed in
chapter 10. One new line of evidence that would not have been expected
comes from wood anatomy, particularly the concept of paedomorphosis
or neoteny (Carlquist, 1962). The strong tendency for change of habit in
insular plants has confused some workers, for certain insular plants may
superficially look more like each other than like their respective mainland
relatives. For example, Keck (1936) thought that the Hawaiian silver-
swords (Argyroxiphium) might be related to a Juan Fernandez genus with
similarity in habit, Robinsonia. In reality, Argyroxiphium is a member of
the tribe Heliantheae, subtribe Madiinae (Carlquist, 1959) of Asteraceae,
whereas Robinsonia belongs to the tribe Senecioneae.

15. Changes in form, size, and color of animals often occur on islands:
gigantism, dwarfism, changes in body proportions, and melanism are
among the changes represented.

The animals that show the trends well are, as one might expect, evolu-
tionarily plastic groups. Insects, birds, reptiles, and land molluscs are

23

Principles of Dispersal and Evolution

among them. Birds and insects, being mostly volant, often have alterations
in shape, size, and proportions related to transitions from flight to flight-
lessness, and are considered under principle 17. The topic of flightlessness
is reviewed in detail in chapter 12 because much evidence, particularly in
insects, has now accumulated.

Gigantism is common among insular animals as compared with their
mainland relatives (Berland, 1924; Mertens, 1934; Steven, 1953; Hill, 1959;
Cook, 1961; Berry, 1964; Darlington, 1971). Because recent studies seem
to permit more secure interpretations of this phenomenon, it is reviewed
in detail in chapter 15.

Dwarfism, however, is also notable on islands. The following reasons
are offered:

1. Smaller forms on islands may reflect the small size of immigrant
species, a fact related to their greater ease of dispersal.

2. Smaller size may be an adaptation to pressures of predators, if the
pressure is greater on a particular island (Hecht, 1952), although this
pressure has also been claimed to be effective in promoting gigantism (see
chapter 15).

3. Smaller size may be an adaptation to smaller food articles or food
supplies on an island. A small animal could maintain a population of viable
size in a situation where food is scarce.

Changes in proportion that are evident on islands include fatter
bodies, fatter tails, and thicker limbs in reptiles (Mertens, 1934). These
changes, shown well by lizards, may be forms of gigantism, and to that
extent they are explained in relation to this topic (chapter 15). Lack of
need for locomotion may underlie these trends, as may need for water
storage during dry seasons. Shorter limbs and tails are also characteristic
of many island reptiles. A good instance is reported in a well-designed
study of lizards (Kramer, 1951). Stubby legs and clumsy habits are claimed
for some island insects, as in Hawaii (Perkins, 1913). Adaptation to a ter-
restrial way of life where food supplies are secured by crawling may be
responsible. Kramer interpreted the shorter limbs and tails in the lizards
he studied as responses to the lack of predator pressure and to the short-
distance nature of locomotion on small islets.

Increase in bill size of birds on islands has been noted repeatedly and
is discussed in chapter 15. Change in plumage color of birds on islands

24

Principles of Dispersal and Evolution

(chapters 6 and 15) and melanism (chapter 15) have been subjects of
recent investigations.

16. Dispersal mechanisms and dispersal ability may be lost during
the evolution of plants on oceanic islands.

Plants owe their presence on oceanic islands to excellence of dispersal
mechanisms. In many cases, these mechanisms are still intact and demon-
strable. In other groups, they have been lost. In a comparison with repre-
sentatives that retain good dispersibility, any morphological changes that
tend to decrease dispersibility will be readily visible. After noticing nu-
merous examples of these trends on the Hawauan Islands and other Pacific
islands, I decided to investigate the phenomenon in detail. The results are
presented in chapter 1]. Although “‘loss of dispersibility” is a convenient
heading to use when discussing various changes that lead to deterioration
in dispersal mechanisms of seeds and fruits, this phenomenon is not a
single trend. What may result in lessening of dispersibility in propagules
of one group may not produce the same result in another group. Only
particular insular areas demonstrate these phenomena. However, because
numerous taxonomic groups and insular areas do show parallel changes,
one can offer some likely interpretations.

A subtle concept related to loss of dispersibility in propagules—but
extremely important nonetheless—is the probability of shift in ecological
preferences by groups on islands. Very likely, most groups become adapted
to ecological conditions on islands (or islandlike areas). To the extent that
these conditions differ from those of source areas, and adaptation occurs,
these plants (and animals) will have lost the ability to disperse success-
fully to new sites. This concept seems potentially much more important
than is the relatively small surface area of islands in explaining why suc-
cessful dispersal from islands to mainland areas does not, apparently, occur.

A reasonable corollary, however, is that an insular species can disperse
to, and establish itself on, an ecologically similar island without difficulty.
This seems to be borne out in certain instances. It may, for example, ex-
plain the surprising number of angiosperm species endemic to more than
one of the California Channel Islands. They include Crossosoma califor-
nicum, Jepsonia malvifolia, and Haplopappus canus (see Raven, 1967, for a
listing). The distances between some of the islands occupied by these
species are greater than distances to the mainland. The only explanation

25

Principles of Dispersal and Evolution

for this distribution is that species such as Galvezia speciosa are adapted
to a particular range of insular climates and can disperse long distances
from one of the Channel Islands to another yet cannot obtain a foothold
on the nearby mainland coast. A maritime species with broader tolerances
could occur both on islands and on a few coastal mainland areas; Coreopsis
gigantea, in fact, has such a distribution.

17. Flightlessness may evolve in volant groups of animals in response
to insular conditions. Ecological shift may produce equivalent restriction
to insular areas.

The number of flightless birds and insects among waif biotas on
islands is quite startling. Early noted by Darwin (1855), it has received
comment and interpretation from various authors, such as Perkins (1913),
Darlington (1943), Brinck (1948), Zimmerman (1948, 1957), Hagen
(1952), Gressitt, Leech, and Wise (1963), Holloway (1963), and Gressitt
(1964). A list of flightless birds has been assembled (Carlquist, 1965).

Darlington’s (1943) study of flightlessness in carabids of the West
Indies and the New Hampshire mountains was notable for the systematic
way in which it attempted to isolate and formulate decisive interpretations
for factors that lead to flightlessness. The Carabidae are a good group for
study, because they are widespread and are apparently sensitive to factors
that tend to promote flightlessness. More recent literature supports many
of Darlington’s interpretations. However, flightlessness in islands and
islandlike areas appears to involve additional factors, and thus a more
complicated picture appears. Darlington’s stress on a stable habitat of
limited geographical extent obviously does apply to certain instances of
flightlessness. But does it apply to the extreme conditions of the subant-
arctic islands and equatorial alpine regions? In an attempt to analyze this
picture, I have devoted chapter 12 to a discussion of flightlessness, and a
terminal section of chapter 14 to a commentary on flightlessness in the
equatorial alpine region. Flightlessness also characterizes many cave insects
(May, 1963; see also chapter 15) and some groups of parasitic insects, such
as the flies that parasitize bats (Streblidae, Nycteribiidae). The latter cir-
cumstance suggests that a host-parasite relationship may constitute, in
some respects, an insular situation.

If, as Darlington stresses, adaptation to stable habitats of limited
geographical extent is strongly related to flightlessness, does ecological shift

26

Principles of Dispersal and Evolution

by animal species at large on islands result in a loss of dispersibility? Appar-
ently it does. The Tornatellinidae (Cooke and Kondo, 1960) form an
interesting example. Although some genera of this family, such as Torna-
tellinides, range literally from one side of the Pacific to the other, they
seem unable to establish on mainland areas. Apparently they are adapted
to insular conditions, and some factors in mainland areas must be inimical
to them—either climatic factors to which they are not adapted, or the
superiority there of continental genera of land shells. As with plants (see
principle 16), ecological shift on islands may be irreversible, and increasing
specialization leads to what can be termed loss of dispersibility, although
it might also be placed under other headings.

18. Competitive ability is often decreased slightly to markedly among
endemics of oceanic islands.

The statement above has an intentional vagueness that hinges on the
inexactness of the concept of competition. The phenomenon seems real
enough, although difficult to demonstrate precisely or to quantify, and
aspects have been noticed by many biologists. One indication is the rapid-
ity with which indigenous island species yield to continental species (see
Elton, 1958; innumerable examples could, of course, be cited). Competi-
tive ability would appear to depend upon constant selective pressure, and
is thus abundantly represented on continents, where numerous aggressive
groups are evolving simultaneously. This pressure is lowered or vanishes
when a species migrates to an island where there are fewer species and
where ecological opportunities not severely contested. Each immigrant
group would, in this view, lose competitive ability after arrival, so that at
any time the total flora or fauna would be less competitive than on a con-
tinent. However, one might prefer to cite more precise factors. Certainly
some groups on mainland areas become “Jess competitive,” and the reasons
are presumably similar to the less competitive characteristics of insular
groups. Even within the native floras and faunas of islands, there are
weedier endemics. The question why insular biotas as a whole seem to
have a lower level of competitiveness is not easy to answer, although some
of the observations that follow may appeal as factors. The nature of com-
petition, with reference to geographical areas at large, has been scrutinized
and mathematically formulated by MacArthur (1972).

In the Hawaiian Islands, almost any introduced continental species

27

Principles of Dispersal and Evolution

of plant seems capable of replacing autochthonous species of comparable
ecological requirements. Of course, soils have been disturbed by man and
his domesticates, yet native island species (with some exceptions) show
poorer self-replacement after disturbance than do native species of a com-
parably disturbed continental area. The rain forests of Kauai now host a
remarkable variety of weeds, including many garden flowers, few of which
would be noxious—if they were weedy at all—in continental areas. Even
high bogs are not exempt from weeds (Rubus is now covering the bogs of
Mt. Kaala, Oahu). This situation augurs poorly for attempts at conservation
of island endemics. Not only are weeds well-entrenched in many areas of
the Hawaiian Islands, but efforts to remove them would very likely only
renew and widen the areas of disturbance and encourage more weedy
growth than before. Many plants that are now weeds in the Hawaiian
Islands can hardly be kept out of many areas because of their good dis-
persal mechanisms (e.g., Schinus and Psidium are spread by frugivorous
birds). Among the Hawaiian endemics that seem capable of occupying
disturbed sites are species of Scaevola, Pipturus, and Acacia. Even with
these, one could name more aggressive continental species of comparable
ecological requirements.

Loss of competitive ability may in part be due to genetic depaupera-
tion, a fate that seems common to most island autochthones. One would
expect that even in prehuman times some genera had lost momentum.
There do seem to be such genera in the Hawaiian biota. These include
two genera of land shells, Carelia (Cooke, 1931) and, to a lesser extent,
Achatinella (Cooke and Kondo, 1960). Among plant genera, the lobelioid
Delissea has been labeled “decadent” by Rock (1962). Other Hawaiian
plant genera that could be described in this way include Hesperomanma,
Hibiscadelphus, Kokia, Pteralyxia, Remya, Rollandia, and some species of
many other genera, such as Cyanea and Clermontia.

Irreversible adaptation to highly specialized locations or food sources
may be a key to these vanishing genera. Specializations are presumably
irreversible if loss of variability occurs. An insular phylad would be ex-
pected to be either in a process of expansion and speciation as it draws on
genetic variability (retained or acquired), or in a state of decadence, as
adaptability to new situations dwindles. The upgrade groups would be

28

Principles of Dispersal and Evolution

expected to replace the downgrade ones, so that taxa in both categories
could be found at any given time. Paucity of individuals would seem to
play a key role in loss of genetic variability. Adaptation to an ecological
zone of limited extent (and on oceanic islands, any zone would be limited
in extent) would result in a smaller number of individuals per species. Ex-
cessive specialization is exemplified among Hawaiian animals by the land
shell Achatinella, which is capable of eating only epiphytic algae and fungi,
not foliage. The Hawaiian drosophilids subsist on a very limited number
of food sources, including leaves of the endemic arborescent Araliaceae
and rotting portions of lobelioids—foods unusual for these genera (Carson
et al., 1970). Some Hawaiian insect species apparently are so restricted that
they occur only on a single tree (Zimmerman, 1948).

Most oceanic islands lack herbivores; large mammalian herbivores are
especially conspicuous in the list of absentees. The effect of a large mam-
malian herbivore is drastic and is related to a wholly different ecosystem
than that found on oceanic islands. Not only do herbivores consume
large quantities of plant material, but they also create considerable dis-
turbance by trampling and by creating and maintaining trails. Insular
autochthones that evolve in the absence of large herbivores do not have
any resistance to them, with few exceptions. For example, Hawaiian
plants are notably nonpoisonous, free from armament, and free from many
characteristics thought to be deterrents to herbivores (oils, resins, stinging
hairs, coarse texture) (Carlquist, 1970). Herbivores have been introduced
to virtually all major oceanic islands and archipelagos, with predictable
effects. A review of these topics, island by island, is hardly necessary for
field biologists who know islands. Because so many oceanic islands have
been afflicted by man-introduced herbivores, and because oceanic island
floras (and indirectly, faunas) are affected by them, insular biotas can be
said to be vulnerable compared to continental biotas.

The same situation holds for animal predators, which are absent to a
large degree on oceanic islands, and for parasites. When these organisms
are introduced to islands, the native animals—which have, in the course
of evolution, lost resistance to parasites—will be more vulnerable than con-
tinental relatives that have evolved in coexistence with predators and
parasites.

A minor aspect that could be termed a form of vulnerability is the

29

Principles of Dispersal and Evolution

fearlessness that characterizes insular animals (see principle 22 for citation
of literature).

Problems of endangered species and extinction are acute on most
islands, and a discussion of these questions forms the last section of chap-
er IS,

19. Means for outcrossing tend to be highly developed in at least the
long-term autochthonous biotas of oceanic islands. Species without poten-
tial for outcrossing are probably doomed to shorter tenures.

Baker (1955) hypothesized that self-compatibility is advantageous for
establishment of plant immigrants on islands. This concept seems opera-
tive, although it does not altogether rule out establishments of non-self-
fertile plants (see chapter 13). In any case, there is apparently a selective
advantage among island autochthones for mechanisms that tend to pro-
mote or ensure outcrossing. Where population size is limited, as it neces-
sarily is on islands, and where inflow of new genetic material is cut off by
the sea barrier, maximization of outcrossing seems a necessity.

Flowers may be wholly outcrossing, as in the case of dioecious, self-
sterile, or heterostyled species. Self-sterility and heterostyly are rare on
islands, but dioecism is exceptionally common in certain insular floras.
Other floral conditions that tend to promote outcrossing include gynodi-
oecism, monoecism, protandry, protogyny, and wind pollination. Such
conditions tend to bulk larger in the floras of oceanic islands than in
comparable mainland areas. Briefly, one can cite high outcrossing poten-
tial, by virtue of the abundance of the floral types listed above, in the Juan
Fernandez Islands (Skottsberg, 1928, 1938), the Desventuradas Islands
(Skottsberg, 1963), New Zealand (Godley, 1955; Franklin, 1962; Ratten-
bury, 1962; Dawson, 1964), and the Hawaiian Islands (Skottsberg, 1936a,
1936b, 1944a, 1944b, 1945; Fosberg, 1956; see also chapter 13).

Because floral conditions and other factors tending to promote out-
crossing are diverse and complex, I have devoted chapter 13 to this topic,
which requires detailed analysis.

20. Natural hybridization acquires a positive value in evolution of
the waif flora and, in some cases, fauna.

In a sense, this is merely an extension of principle 19, because ex-
change and circulation of genetic material, a desideratum in a finite pop-
ulation cut off from new genetic material, is achieved.

30

Principles of Dispersal and Evolution

Taxonomic literature on insular floras (and sometimes faunas) often
mentions that some species are unusually “variable,” “unstable,” or “poly-
morphic.” These conditions would suggest that hybridization, or some
aspect related to it, is operative. Yet taxonomists who have written floras
or monographs of insular genera have not often reported hybridization,
probably because until recently students in taxonomy were not trained to
detect and report hybrids. Nevertheless, there have been reports of hybrid-
ization in waif floras: by Dawson (1960) in Acaena, Fosberg (1956) in
Gouldia, Franklin (1962, 1964) in Gaultheria and Pernettya, Gillett and
Lim (1970) in Bidens, Lems (1958) in Adenocarpus, Nicharat and Gillett
(1970) in Pipturus, Rock (1919a) in Clermontia, Sherff (1935) in Dubau-
tia, and Skottsberg (1939) in Viola. “Cyclic hybridization” is reported in
the New Zealand flora by Rattenbury (1962), and numerous reports of
hybrids in the New Zealand flora can be gleaned from Cockayne and
Allan (1934) and Allan (1961). Hybridization in the land shell Partula is
clearly indicated by Schwabl and Murray (1970).

Hybrids might be said to occur between races (although some bota-
nists will argue that this violates the meaning of the term hybrid) as well as
between species on islands. A successful condition (at least in plants)
would seem to be that of a population that takes advantage of reproduc-
tive isolation in speciating into new ecological territory, but that retains
fertility with other populations. As conditions change, gene flow among a
series of semiseparate populations (and by “populations” one can include
different taxonomic species) could maintain a high level of adaptability.
Instances, modes, and extent of natural hybridization remain rewarding
avenues for investigation by students of insular botany. These investiga-
tions are particularly urgent, because they can be pursued only while
insular floras are relatively intact. I applaud the studies of Gillett (1966)
on Scaevola and Gillett and Lim (1970) on Hawaiian species of Bidens.
Further commentary on the role of hybridization in insular floras will be
found in chapter 13.

21. Pollination relationships correspond to and change with respect
to availability of insects and other pollinating agents on islands.

Wallace (1895) noted the paucity of conspicuous flowers in the
floras of the Galapagos Islands and New Zealand, a fact he correlated with
poverty of insects on those islands. This insect poverty is best described,

31

Principles of Dispersal and Evolution

however, in terms of the absence or scarcity of particular groups, such as
butterflies in the Galapagos (Wallace, 1895) or long-tongued bees in New
Zealand (Rattenbury, 1962). The Hawaiian forest flora also is abundant in
small green or whitish flowers that are poor in scent. Smaller flowers have
probably established preferentially or evolved in these habits to suit the
smaller sizes, and the habits and preferences, of available pollinators. As
will be noted, the waif insect fauna consists of immigrants belonging to
smaller size classes, and this situation may have promoted the evolution
of smaller floral sizes. That island conditions do influence pollination
mechanisms is suggested by Hagerup (1950a, 1950b, 1951). Flowers on
oceanic islands often have simple, open forms, suitable to entry by a wide
variety of potential pollinators.

Massing of flowers might serve to attract pollinators, and is theoreti-
cally advantageous if pollinators are scarce. This factor might help to ex-
plain the occurrence of insular rosette trees with large inflorescences, such
as Wilkesia and Trematolobelia (Hawaiian Is.) or Centaurodendron, Yun-
queda, and Dendroseris sect. Phoenicoseris (Juan Fernandez Is.). On
these islands other plants with massive inflorescences could also be cited.

Wallace (1895) has claimed that the perennial habit aids in securing
pollinators. That interpretation of this habit seems questionable, because
a few additional flowering seasons will probably not serve to secure pollina-
tion if suitable pollinators are not present in adequate numbers in any
given year. Probability seems to favor either the continual presence of a
suitable insect in sufficient numbers or else its complete absence from an
island. The high proportion of perennials on oceanic islands probably is
related primarily to climatic factors (see principle 14 as well as the detailed
account in chapter 10).

The attractiveness of Asteraceae to a variety of insects is alleged by
Wallace (1895) to be a reason for the abundance of this family on islands.
If operative, this reason is probably subsidiary to the good dispersibility,
weediness, and adaptability that are characteristic of this family.

Change from insect pollination to wind pollination (partly or wholly)
solves the problem of scarcity of suitable insect pollinators. Autochthonous
evolution into anemophily may have occurred on a few islands (e.g., with
Rhetinodendron on the Juan Fernandez Islands [Skottsberg, 1928]). Ane-

32

Principles of Dispersal and Evolution

mophily might help to explain the success of some genera on oceanic islands
(e.g., Plantago, Coprosma).

22. Some mutations that would be lethal or disadvantageous in con-
tinental environments have a more nearly neutral value in the less compett-
tive environment of an oceanic island.

The less disadvantageous nature of many mutations (see the closely
allied principle 18) is easy to state but difficult to demonstrate, although
many of the changes described under previous headings might qualify as
examples. One feature that clearly does qualify is the fearlessness that
typifies many insular animals. Many descriptions of fearlessness in insular
animals, particularly birds, have been offered (Lonnberg, 1920; Beebe,
1924; Rand, 1938; Greenway, 1958; Rice, 1964; Carlquist, 1965). Where
predators are absent, evasive action would seem to be an unnecessary
behavioral characteristic, perhaps wasteful, because excessive wariness
could interfere with other activities such as feeding.

Development of conspicuous color patterns may be a visual equiva-
lent of fearlessness, reflecting as it does fearless behavior and the absence
of predators. This characteristic occurs prominently among certain lizards
(Mertens, 1934; Carlquist, 1965).

Various forms of apparent reproductive inefficiency might charac-
terize some animals and plants on islands, although this subject is difh-
cult to explore. Changes in fertility may not be so much inefficiency as
alterations to the lowered rates of reproduction required to maintain a
small insular population, or adaptation to a limited food supply. Further
commentary on these topics is offered in chapter 13.

On the Brazilian island of Queimada Grande, an unusually high per-
centage of individuals of the viper Bothrops insularis are hermaphrodites
(sterile), and the percentage seems increasingly large. For an account, see
Mertens (1960), who summarizes, for example, the work of A. R. Hoge
et al. (published in the Memoirs of the Institute of Butantan, vol. 29,
1959). This situation would probably be disadvantageous in a mainland
environment. Slow maturation or delayed adulthood of island birds
(Hagen, 1952) may represent adjustment to the altered nature of repro-
duction on a small land area. Greater longevity of island animals has been
observed in several cases (see chapter 13); although longevity would itself

33

Principles of Dispersal and Evolution

be of no advantage on islands, it may represent release from predator pres-
sure. Reptiles and amphibians are usually killed before reaching full size.
If enemies and hazards are reduced on islands, more individuals would
reach full size than on mainland areas.

In the case of seeds and fruits that lose dispersibility, there appear
to be malformations that have no selective advantage, as in Yunquea,
Dendroseris, and perhaps some species of Bidens. For whatever reasons
such malformations occur, they seem quite unusual for the family to which
these genera belong, Asteraceae, and do not seem to have occurred on
mainland areas (see chapter 11).

23. Endemism, although high on oceanic islands, is not itself a
criterion for identification of an island as oceanic. There are various rea-
sons for high and low endemism. The nature of endemism may prove
informative as a way to illustrate certain features concerning insular biotas.

Oceanic islands tend to have endemics that are restricted to lower
categories, particularly species, and to a lesser extent, genera. Old conti-
nental islands may possess endemic families, even orders, as relicts from
continents. Setchell (1928) hints at this.

Endemism is a constant byproduct of evolutionary change, and the
percentage of endemics is more a measure of degree of isolation in time
and space than of mode of origin of islands. Ecological opportunity is also
involved, for shift into new ecological sites usually is accompanied by
changes that promote endemism. Atolls have a low rate of endemism not
only because they are not isolated (dispersal is easy via migratory birds
and oceanic drift) but because they are ecologically relatively uniform.

If an island biota contains phylogenetically primitive forms with poor
dispersal ability, a continental origin for the island may be suspected, or it
might be continentlike (an island composed of sedimentary rocks, but not
previously connected to a land mass; whether such islands exist is prob-
lematic, and differences between them and continental islands separated
from land masses for very long periods would be difficult to detect). For
example, the presence of Araucariaceae strongly suggests that New Cale-
donia is an old continental island.

An interesting commentary on endemism of marine shore faunas of
islands is offered by McDowall (1968). He shows that endemism of marine

34

Principles of Dispersal and Evolution

fishes is high where ocean currents do not provide an island with a supply
of species from other areas (e.g., 33 percent endemism for Easter Island,
50 percent for the Juan Fernandez Islands) or when islands are extremely
isolated (e.g., Kerguelen). Presumably, if currents came from islands of
different marine environment than recipient islands, endemism would still
be high. Islands in the paths of suitable currents, however, have very low
rates of marine fish endemism (5 percent for Bermuda). McDowall also
shows that high endemism in marine faunas in general on islands is related
to lack of pelagic stages in particular taxonomic groups.

Comments on endemism and its relation to speciation in various
insular areas will be found in chapter 3.

24. Eyolutionarily plastic groups are sensitive indicators of directions
of evolution in the biota that results from colonization by means of long-
distance dispersal.

Groups such as insects or composites (Asteraceae) may be expected
to fit themselves rapidly and closely to the templates provided by the island
environment. As an example of the usefulness of this concept, one can dis-
count Skottsberg’s (1956) claim that peculiar growth forms on the Juan
Fernandez Islands are relicts. One would expect such growth forms, if they
were relicts, to be members of relatively primitive or slowly evolving
groups. Instead, the Juan Fernandez rosette trees and rosette shrubs occur
in upgrade, predominantly herbaceous groups, characterized by weediness
and rapid evolution; for example, the Plantaginaceae, the Apiaceae, and
three tnbes of the Asteraceae.

One can hardly doubt that Chenopodium is a weedy genus of flower-
ing plants. Not surprisingly, it has reached the Hawaiian Islands and the
Juan Fernandez Islands and has formed arborescent endemic species there.
One might expect that those families that contain agricultural weeds in
abundance in Europe might be represented on oceanic islands; they are.
In some cases (Sonchus, Lepidium, Rumex, and Artemisia) the same
genera are involved.

Some groups that do not appear to be weedy are nonetheless evi-
dently excellent at colonization of pioneer areas—for example, the lobe-
loids (especially Lobelia and Clermontia) of the Hawaiian Islands. Several
species of those two genera often occur there on road cuts and new lava.

2)

Principles of Dispersal and Evolution

References

Allan, A. H. 1961. Flora of New Zealand, vol. 1. Wellington, R. E. Owen, Govern-
ment Printer.

Amadon, D. 1950. The Hawaiian honeycreepers (Aves, Drepaniidae). Bull. Amer.
Mus. Natur. Hist. 95:151-262.

Andrews, E. C. 1940. Origin of the Pacific insular floras. Proc. Sixth Pacific Sci.
Congr. 4:613-20.

Bailey, D. W. 1956. Re-examination of the diversity in Partula taeniata. Evolution
10:360-66.

Baker, H. G. 1955. Self-compatibility and establishment after “long-distance” dis-
persal. Evolution 9:347-49.

Baker, H. G., and G. L. Stebbins, Jr., eds. 1965. The genetics of colonizing species.
New York, Academic Press.

Beebe, W. 1924. Galapagos: World’s end. New York, G. P. Putnam’s Sons.

Berland, L. 1924. Araignées de Vile de Paques et des iles Juan Fernandez. Natur.
Hist. Juan Fernandez Easter I. 3:419-37.

Berry, R. J. 1964. The evolution of an island population of the house mouse. Evolu-
tion 18:468-83.

Blake, S. F., and A. C. Atwood. 1942. Geographical guide to floras of the world, pt.
1. U.S. Dept. Agr. Misc. Publ. 401. Washington, D.C.

Borgesen, F. 1924. Contributions to the knowledge of the vegetation of the Canary
Islands. Mém. Acad. Roy. Sci. Danemark, Sect. Sci., ser. 8, 6:285—395.

Bowman, R. I. 1961. Morphological differentiation and adaptation in the Galapagos
finches. Univ. Calif. Publ. Zool. 58:1—302.

Brinck, P. 1948. Coleoptera of Tristan da Cunha. Results Norwegian Sci. Exped.
Tristan da Cunha 17:1-121.

Brown, W. C., and A. C. Alcala. 1970. The zoogeography of the herpetofauna of
the Philippine Islands, a fringing archipelago. Proc. Calif. Acad. Sci., 4th ser.,
38(6):105-30.

Carlquist, S. 1959. Studies on Madiinae: Anatomy, cytology, and evolutionary rela-
tionships. Aliso 4:171—236.

Carlquist, S. 1962. A theory of paedomorphosis in dicotyledonous woods. Phyto-
morphology 12:30—45.

Carlquist, S. 1964. Morphology and relationships of Lactoridaceae. Aliso 5:421—35.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S$. 1966. The biota of long-distance dispersal. I. Principles of dispersal and
evolution. Quart. Rey. Biol. 41:247-70.

Carlquist, S. 1970. Hawaii, a natural history. New York, Natural History Press.

Carson, H. L., D. E. Hardy, H. T. Spieth, and W. S. Stone. 1970. The evolutionary
biology of the Hawaiian Drosophilidae. In Essays in evolution and genetics in
honor of Theodosius Dobzhansky, ed. M. K. Hecht and W. C. Steere, pp. 437—
543. New York, Appleton-Century-Crofts.

36

Principles of Dispersal and Evolution

Clagg, H. B. 1966. Trapping of air-borne insects in the Atlantic-Antarctic area.
Pacific Insects 8:455-66.

Cloud, P. E., Jr. 1956. Provisional correlation of selected Cenozoic sequences in the
western and central Pacific. Proc. Eighth Pacific Sci. Congr. 2:555-73..

Cloud, P. E., Jr., R. G. Schmidt, and H. W. Burke. 1956. Geology of Saipan, Mari-
anas Islands. Part I. General Geology. U.S. Geol. Survey Prof. Papers 208A:1-126.

Cockayne, L., and H. H. Allan. 1934. An annotated list of groups of wild hybrids in
the New Zealand flora. Ann Bot. 48:1—55.

Constance, L., L. Heckard, K. L. Chambers, R. Ornduff, and P. R. Raven. 1963.
Amphitropical relations in the herbaceous floras of the Pacific coast of North and
South America: A symposium. Quart. Rey. Biol. 38:109-77.

Cook, L. M. 1961. The edge effect in population genetics. Amer. Natur. 95:295—307.

Cooke, C. M., Jr., 1931. The land snail genus Carelia. Bishop Mus. Bull. 85:1-97.

Cooke, C. M., Jr., and Y. Kondo. 1960. Revision of Tornatellinidae and Achati-
nellidae (Gastropoda, Pulmonata). Bishop Mus. Bull. 221:1-303.

Crampton, H. E. 1916. Studies on the variation, distribution and evolution of the
genus Partula. Carnegie Inst. Wash. Publ. 228:1-311.

Crampton, H. E. 1925. Contemporaneous differentiation in the species of Partula
living on Moorea, Society Islands. Amer. Natur. 59:5-35.

Crampton, H. E. 1932. Studies on the variation, distribution and evolution of the
genus Partula. The species inhabiting Moorea. Carnegie Inst. Wash. Publ. 410:1-
555.

Dammermann, K. W. 1948. The fauna of Krakatau 1883-1933. Verh. kon. Ned
Akad. Wetensch. Afd. Natuurk., ser. 2, 44:1-594.

Darlington, P. J., Jr. 1938. The origin of the fauna of the Greater Antilles, with dis-
cussion of dispersal of animals over water and through the air. Quart. Rev. Biol.
13:274-300.

Darlington, P. J., Jr. 1943. Carabidae of mountains and islands: Data on the evolu-
tion of isolated faunas and on atrophy of wings. Ecol. Monogr. 13:37-61.

Darlington, P. J., Jr. 1957. Zoogeography. New York, John Wiley & Sons.

Darlington, P. J., Jr. 1971. Carabidae on tropical islands, especially the West Indies.
In Adaptive aspects of insular evolution, ed. W. L. Stern, pp. 7-15. Pullman,
Wash., Washington State University Press.

Darwin, C. 1855. Letter to J. D. Hooker, March 7, 1855. In Life and letters of
Charles Darwin, ed. F. Darwin, New York, Basic Books.

Darwin, C. 1859. On the origin of species by means of natural selection. Reprint of
first edition, 1950. London, Watts & Co.

Dawson, J. W. 1958. Interrelationships of the Australasian and South American
floras. Tuatara 7:1-6.

Dawson, J. W. 1960. Natural Acaena hybrids growing in the vicinity of Wellington.
Trans. Roy. Soc. New Zealand 88:13-27.

Dawson, J. W. 1963. Origins of the New Zealand alpine flora. Proc. New Zealand
Ecol. Soc. 10:1—-4.

Dawson, J. W. 1964. Unisexuality in the New Zealand Umbelliferae. Tuatara 12:67—
68.

37

Principles of Dispersal and Evolution

Dobzhansky, T. 1957. Genetics of natural populations. XXVI. Chromosomal varia-
bility in island and continental populations of Drosophila willistoni from Central
America and the West Indies. Evolution 11:280-93.

Docters van Leeuwen, W. M. 1936. Krakatau, 1883-1933. Ann. Jard. Bot. Buiten-
zorg, 46-47:xii + 506.

Durham, J. W. 1963. Paleogeographic conclusions in light of biological data. In
Pacific basin biogeography, ed. J. L. Gressitt, pp. 355-65. Honolulu, Bishop
Museum Press.

Einarsson, E.. 1967. The colonization of Surtsey, the new volcanic island, by vascular
plants. Aquilo, ser. Bot. 6:172-82.

Elton, C. S. 1958. The ecology of invasions by animals and plants. New York, John
Wiley & Sons.

Falla, R. A. 1960. Oceanic birds as dispersal agents. Proc. Roy. Soc. London, ser. B,
152:655-59.

Fleming, C. A. 1963. Paleontology and southern biogeography. In Pacific basin bio-
geography, ed. J. L. Gressitt, pp. 369-82. Honolulu, Bishop Museum Press.

Fosberg, IF’. R. 1948. Derivation of the flora of the Hawauan Islands. In Insects of
Hawai: Introduction, E. C. Zimmerman, pp. 107-19. Honolulu, University of
Hawaii Press.

Fosberg, F. R. 1956. Studies in Pacific Rubiaceae: I-IV. Brittonia 8:165-178.

Fosberg, F. R. 1963. Plant dispersal in the Pacific. In Pacific basin biogeography, ed.
J. L. Gressitt, pp. 273-81. Honolulu, Bishop Museum Press.

Franklin, D. A. 1962. The Ericaceae in New Zealand (Gaultheria and Pernettya).
Trans. Roy. Soc. New Zealand, Bot. 1:155-73.

Franklin, D. A. 1964. Gaultheria hybrids on Rainbow’Mountain. New Zealand
Jour. Bot. 2:34-43.

Gillett, G. W. 1966. Hybridization and its taxonomic interpretation in the Scaevola
gaudichaudiana complex of the Hawaiian Islands. Evolution 20:506-16.

Gillett, G. W., and E. K. S. Lim. 1970. An experimental study of the genus Bidens
in the Hawaiian Islands. Uniy. Calif. Publ. Bot. 56:1-62.

Godley, E. J. 1955. Breeding systems in New Zealand plants. I. Fuchsia. Ann. Bot.
19:549-59.

Greenslade, P. J. M. 1968a. Island patterns in the Solomon Islands. Bird fauna.
Evolution 22:751-61.

Greenslade, P. J. M. 1968b. The distribution of some insects in the Solomon
Islands. Proc. Linnaean Soc. London 179:189-96.

Greenway, J. C., Jr. 1958. Extinct and vanishing birds of the world. New York,
American Committee for International Wild Life Protection.

Gressitt, J. L. 1956. Some distribution patterns of Pacific island faunas. Syst. Zool.
5:11-32.

Gressitt, J. L. 1961a. Problems in the zoogeography of Pacific and Antarctic insects.
Pacific Insects Monogr. 2:1—94.

Gressitt, J. L. 1961b. Zoogeography of Pacific Coleoptera. Verh. XI Internat. Kongr.
Entomol. 1:463-65.

Gressitt, J. L. 1964. Insects of Campbell Island. Summary. Pacific Insects Monogr.
7:531-600.

38

Principles of Dispersal and Evolution

Gressitt, J. L., R. E. Leech, and C. W. O’Brien. 1960. Trapping of air-borne insects
in the Antarctic area. Pacific Insects 2:245—-50.

Gressitt, J. L., R. E. Leech, and K. A. Wise. 1963. Entomological investigations in
Antarctica. Pacific Insects 5:287—304.

Gressitt, J. L., R. E. Leech, et. al. 1961. Trapping of air-borne insects in the Antarc-
tic area. Part 2. Pacific Insects 3:559-62.

Gressitt, J. L., and S. Nakata. 1958. Trapping of air-borne insects on ships on the
Pacific. Proc. Hawaiian Entomol. Soc. 16:363—-65.

Gressitt, J. L., J. Sedlacek, K. A. Wise, and C. M. Yoshimoto. 1961. A high-speed
airplane trap for air-borne organisms. Pacific Insects 5:549—-55.

Gressitt, J. L., and C. M. Yoshimoto. 1963. Dispersal of animals in the Pacific. In
Pacific basin biogeography, ed. J. L. Gressitt, pp. 283-92. Honolulu, Bishop
Museum Press.

Gulick, A. 1932. Biological peculiarities of oceanic islands. Quart. Rev. Biol. 7:405-
a

Hagen, Y. 1952. Birds of Tristan da Cunha. Results Norwegian Exped. Tristan da
Cunha 20:1-248.

Hagerup, O. 1950a. Thrips pollination in Calluna. Kong. Danske Vidensk. Selsk.
Biol. Medd. 18(4):1-16.

Hagerup, O. 1950b. Rain pollination. Kong. Danske Vidensk. Selsk. Biol. Medd.
18(5):1-19.

Hagerup, O. 1951. Pollination in the Faroes—in spite of rain and poverty of insects.
Kong. Danske Vidensk. Selsk. Biol. Medd. 18(15):1-48.

Hamilton, E. L. 1953. Upper Cretaceous, Tertiary, and Recent planktonic Forami-
nifera from mid-Pacific flat-topped seamounts. Jour. Paleontol. 27:204-37.

Hamilton, E. L. 1956. Sunken islands of the mid-Pacific mountains. Geol. Soc.
Amer. Mem. 64:1-97.

Hamilton, T., and J. Rubinoff. 1963. Isolation, endemism and multiplication of
species in the Darwin finches. Evolution 17:388—403.

Harling, G. 1962. On some Compositae endemic to the Galapagos Islands. Acta
Horti Bergiani 20:63—120.

Harrell, J. C., and E. Holzapfel. 1966. Trapping air-borne insects on ships in the
Pacific. Part 6. Pacific Insects 8:33—42.

Hecht, M. K. 1962. Natural selection in the lizard genus Aristelliger. Evolution
6:122-24.

Hemsley, W. B. 1885. Endemic and arborescent Compositae in oceanic islands. Rep.
Sci. Results Voyage H.M.S. Challenger, Bot., 1:19-24.

Hill, J. E. 1959. Rats and mice from the islands of Tristan da Cunha and Gough,
South Atlantic Ocean. Results Norwegian Sci. Exped. Tristan da Cunha 46:1-5.

Holdgate, M. W. 1960. The fauna of the mid-Atlantic islands. Proc. Roy. Soc. Lon-
don, ser. B, 152:550-67.

Holloway, B. A. 1963. Wing development and evolution of New Zealand Lucanidae
(Insecta: Coleoptera). Trans. Roy. Soc. New Zealand, Zool. 3:99-116.

Hubbell, T. H. 1968. The biology of islands. Proc. Nat. Acad. Sci. U.S.A. 60:22-32.

Johnson, C. G. 1960a. A basis for a general system of insect migration and dispersal
by flight. Nature 186:348—50.

39

Principles of Dispersal and Evolution

Johnson, C. G. 1960b. Present position in the study of insect dispersal and migra-
tion. Rep. Seventh Commonwealth Entomol. Conf., London: 140-45.

Keck, D. D. 1936. The Hawaiian silverswords: Systematics, affinities and phyto-
geographic problems of the genus Argyroxiphium. Occas. Papers Bishop Mus.
11(19):1-38.

Kozhoy, M. 1963. Lake Baikal and its life. The Hague, Dr. W. Junk, Publishers.

Kramer, G. 1951. Body proportions in mainland and island lizards. Evolution 5:193-
206.

Lack, D. 1945. The Galapagos finches (Geospizinae). Occas. Papers Calif. Acad Sci.
Tee

Lack, D. 1947. Darwin’s finches. Cambridge, Cambridge University Press.

Lack, D. 1969. The numbers of bird species on islands. Bird Study 16:193-209.

Lack, D. 1971. Island birds. In Adaptive aspects of insular evolution, ed. W. L. Stern,
pp. 29-31. Pullman, Wash., Washington State University Press.

Ladd, H. S. 1958. Fossil land shells from western Pacific atolls. Jour Paleontol.
32:183-98.

Ladd, H. S., E. Ingerson, R. C. Townsend, M. Russell, and H. K. Stephenson. 1953.
Dnilling on Eniwetok Atoll, Marshall Islands. Bull. Amer. Assoc. Petrol. Geol.
37:2257-80.

Lems, K. 1958. Botanical notes on the Canary Islands. I. Introgression among the
species of Adenocarpus, and their role in the vegetation of the islands. Biol. Inst.
Nac. Invest. Agron. (Madrid) 39:351-60.

Lems, K. 1960a. Botanical notes on the Canary Islands. II. The evolution of plant
forms in the islands: Aeonium. Ecology 41:1-17.

Lems, K. 1960b. Floristic botany of the Canary Islands. Sarracenia 5:1-94.

Lems, K. 1961. Botanical notes on the Canary Islands. III. The life form spectrum
and its interpretation. Ecology 42:569-72.

Lonnberg, E. 1920. The birds of the Juan Fernandez Islands. Natur. Hist. Juan
Fernandez Easter I. 3:1—17.

MacArthur, R. H. 1972. Geographic ecology. New York, Harper and Row, Pub-

lishers.

MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoo-

geography. Evolution 17:373-—87.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography.

Princeton, N.J., Princeton University Press.

McCann, C. 1953. Distribution of the Gekkonidae in the Pacific area. Proc. Seventh

Pacific Sci. Congr. 4:27-32.

McDowall, R. M. 1964. The affinities and derivation of the New Zealand fresh-water
fish fauna. Tuatara 12:59-67.

McDowall, R. M. 1968. Oceanic islands and endemism. Syst. Zool. 17:346-50.

McDowell, S. B., Jr. 1958. The Greater Antillean insectivores. Bull. Amer. Mus.
Natur. Hist. 115:113-214.

Mason, R. 1961. Dispersal of tropical seeds by ocean currents. Nature 191:408-9.

Matthew, W. D. 1915. Climate and evolution. Ann. N.Y. Acad. Sci. 24:171-318.

May, B. M. 1963. New Zealand cave fauna. II. The limestone caves between Port
Waikato and Piopio Districts. Trans. Roy. Soc. New Zealand, Zool. 3:181—204.

40

Principles of Dispersal and Evolution

Mayr, E. 1940. The origin and the history of the bird fauna of Polynesia. Proc.
Sixth Pacific Sci. Congr. 4:197-216.

Melliss, J. C. 1875. St. Helena. London, L. Reeve & Co.

Menard, H. W. 1956. Recent discoveries bearing on linear tectonics and seamounts
in the Pacific basin. Proc. Eighth Pacific Sci. Congr. 2a:809.

Menard, H. W., and E. L. Hamilton. 1963. Paleogeography of the tropical Pacific.
In Pacific basin biogeography, ed. J. L. Gressitt, pp. 193-217. Honolulu, Bishop
Museum Press.

Mertens, R. 1934. Die Insel-Reptilien, ihre Ausbreitung, Variation, und Artbilding.
Zoologica 32:1—209.

Mertens, R. 1960. The world of amphibians and reptiles. New York, McGraw-
Hill Book Co.

Munro, G. C. 1960. Birds of Hawaii. 2d ed. Rutland, Vt., Ridgeway Press.

Myers, G. S. 1953. Ability of amphibians to cross sea barriers, with especial refer-
ence to Pacific zoogeography. Proc. Seventh Pacific Sci. Congr. 4:19-26.

Myers, G. S. 1960. The endemic fish fauna of Lake Lanao, and the evolution of
higher taxonomic categories. Evolution 14:232-333.

Nicharat, S., and G. W. Gillett. 1970. A review of the taxonomy of Hawaiian Pip-
turus (Urticaceae) by anatomical and cytological evidence. Brittonia 22:191—206.

Perkins, R. C. L. 1913. Fauna Hawaiiensis: Introduction. Cambridge, Cambridge
University Press.

Rand, A. L. 1938. Results of the Archbold Expedition. No. 22. On the breeding
habits of some birds of paradise in the wild. Amer. Mus. Novitates 993:1-8.

Rand, A. L. 1955. The origin of the land birds of Tristan da Cunha. Fieldiana,
Zool. 37:139-63.

Rattenbury, J. A. 1962. Cyclic hybridization as a survival mechanism in the New
Zealand forest flora. Evolution 16:348—-63.

Raven, P. H. 1967. The floristics of the California Islands. In Proceedings of the
symposium on the biology of the California Islands, ed. R. N. Philbrick, pp. 57—
71. Santa Barbara, Calif., Santa Barbara Botanic Garden.

Remington, D. C. 1971. Natural history and evolutionary genetics of the California
Channel Islands. Discovery (Yale Univ.) 7(1):3-18.

Rice, D. W. 1964. The Hawaiian monk seal. Natur. Hist. 73(2):48-55.

Richter, C. J. J. 1970. Aerial dispersal in relation to habitat in eight wolf spider
species (Pardosa, Araneae, Lycosidae). Oecologia 5:200-14.

Ridley, H. N. 1930. The dispersal of plants throughout the world. Ashford, England,
L. Reeve & Co.

Robyns, W., and S. H. Lamb. 1939. Preliminary ecological survey of the Island of
Hawai. Bull. Jard. Bot. Bruxelles 15:241-93.

Rock, J. F. 1913. The indigenous trees of the Hawaiian Islands. Honolulu, privately
published.

Rock, J. F. 1917. The ohia lehua trees of Hawaii. Terr. Hawaii Board Agr. Forest.
Bull. 4:1-76.

Rock, J. F. 1919a. A monographic study of the Hawaiian species of the tribe
Lobelioideae, family Campanulaceae. Mem. Bishop Mus. 7(2):1-394.

4]

Principles of Dispersal and Evolution

Rock, J. F. 1919b. The arborescent indigenous legumes of Hawaii. Terr. Hawaii
Board Agr. Forest. Bull. 5:1-153.

Rock, J. F. 1962. Hawaiian lobelioids. Occas. Papers Bishop Mus. 23:65-75.

St. John, H. 1966. Monograph of Cyrtandra (Gesneriaceae) on Oahu, Hawaiian
Islands. Bishop Mus. Bull. 229:1-466.

Sauer, J. D. 1969. Oceanic islands and biogeographical theory: A review. Geogr.
Rey. 59:582-93.

Schenck, H. 1907. Beitrage zur Kenntniss der Vegetation der Kananrischen Inseln.
Wiss. Ergebn. Deutsch. Tiefsee-Exped. Dampfer “Valdivia” 2(1:2):225-406.
Schwabl, G., and J. Murray. 1970. Electrophoresis of proteins in natural populations

of Partula (Gastropoda). Evolution 24:424-30.

Setchell, W. A. 1926. Les migrations des oiseaux et la dissémination des plantes.
Compt. Rend. Soc. Biogéogr. 3:54—57.

Setchell, W. A. 1928. Migration and endemism with reference to Pacific insular
floras. Proc. Third Pan-Pacific Sci. Congr. 1:869-75.

Setchell, W. A. 1935. Pacific insular floras and Pacific paleogeography. Amer. Natur.
69:289-310.

Sherff, E. E. 1935. Revision of Tetramolopium, Lipochaeta, Dubautia and Rail-
liardia.. Bishop Mus. Bull. 135:1—136.

Simberloff, D. S. 1969. Experimental zoogeography of islands. A model for insular
colonization. Ecology 50:296—314.

Simberloff, D. S., and E. O. Wilson. 1969. Experimental zoogeography of empty
islands. Ecology 50:278—-95.

Simpson, G. G. 1952. Probabilities of dispersal in geologic time. Bull. Amer. Mus.
Natur. Hist. 99:163-76.

Skottsberg, C. 1928. Pollinationsbiologie und Samenverbreitung auf den Juan Fer-
nandez Inseln. Natur. Hist. Juan Fernandez Easter I. 2:503-47.

Skottsberg, C. 1936a. The arboreous Nyctaginaceae of Hawaii. Svensk Bot. Tidskr.
30:722-43.

Skottsberg, C. 1936b. Vascular plants from the Hawaiian Islands. II]. Medd. Géte-
borgs Bot. Tradgard 10:97-193.

Skottsberg, C. 1938. On Mr. C. Bock’s collection of plants from Masatierra (Juan
Fernandez), with remarks on the flowers of Centaurodendron. Medd. Goteborgs
Bot. Tradgard 12:361-73.

Skottsberg, C. 1939. A hybrid violet from the Hawaiian Islands. Bot. Notis. (1939):
805-12.

Skottsberg, C. 1941. Plant succession on recent lava flows in the Island of Hawaii.
Géteborgs Kungl. Vetens. Vitterhets-Samhdll Handl., ser. b, 1(8):1—32.

Skottsberg, C. 1944a. On the flower dimorphism in Hawaiian Rubiaceae. Ark. Bot.
31a(4):1-28.

Skottsberg, C. 1944b. Vascular plants from the Hawaiian Islands. Medd. Goteborgs
Bot. Tradgard 15:275-531.

Skottsberg, C. 1945. The flower of Canthium. Ark. Bot. 32a(5):1-12.

Skottsberg, C. 1953. The vegetation of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:793-960.

42

Principles of Dispersal and Evolution

Skottsberg, C. 1956. Derivation of the flora and fauna of Juan Fernandez and Easter
Island. Natur. Hist. Juan Fernandez Easter I. 1:193-438.

Skottsberg, C. 1963. Zur Naturgeschichte der Insel San Ambrosio (Islas Desven-
turadas, Chile). 2. Bliitenpflanzen. Ark. Bot., ser. 2, 4:465-88.

Smith, A. C. 1955. Phanerogam genera with distributions terminating in Fij1.
Jour. Arnold Arboretum 36:273-92.

Soulé, M. 1966. Trends in the insular radiation of a lizard. Amer. Natur. 100:47-64.

Soulé, M., and B. R. Stewart. 1970. The “niche-variation” hypothesis: A test and
alternatives. Amer. Natur. 104:85-97.

Stark, J. T., and S. O. Schlanger. 1956. Stratigraphic succession on Guam. Proc.
Eighth Pacific Sci. Congr. 2:262-66.

Stephens, S. G. 1958a. Factors affecting seed dispersal in Gossypium. N.C. Agr.
Exp. Sta. Tech. Bull. 131:1-32.

Stephens, S. G. 1958b. Salt water tolerance of seeds of Gossypium species as a pos-
sible factor in seed dispersal. Amer. Natur. 92:83-92.

Stephens, S. G. 1963. Polynesian cottons. Ann. Missouri Bot. Gard. 50:1-22.

Stephens, S. G. 1964. Native Hawaiian cotton (Gossypium tomentosum Nutt.).
Pacific Sci. 18:385-98.

Steven, D. M. 1953. Recent evolution in the genus Clethrionomys. Symp. Soc. Exp.
Biol. 7:310-19.

Thorne, R. F. 1963. Biotic distribution patterns in the tropical Pacific. In Pacific
basin biogeography, ed. J. L. Gressitt, pp. 311-54. Honolulu, Bishop Museum
Press.

Thornton, I. W. B., and J. C. Harrell. 1965. Air-borne Psocoptera trapped on ships
and aircraft, 2—Pacific ship trappings, 1963-64. Pacific Insects 7:700-702.

Tryon, R. 1971. Development and evolution of fern floras of oceanic islands. In
Adaptive aspects of insular evolution, ed. W. L. Stern, pp. 54-62. Pullman,
Wash., Washington State University Press.

Visher, S. S. 1925. Tropical cyclones and the dispersal of life from island to island
in the Pacific. Amer. Natur. 59:70-78.

Wace, N. M. 1960. The botany of the southern oceanic islands. Proc. Roy. Soc.
London, ser. B, 152:475-90.

Wallace, A. R. 1880. Island life. London, Macmillan & Co.

Wallace, A. R. 1895. Natural selection and tropical nature. London, Macmillan &
Co.

Wheeler, W. N. 1916. Ants carried in a floating log from the Brazilian coast to San
Sebastian Island. Psyche 28:180.

Whitaker, T. W., and G. F. Carter. 1954. Oceanic drift of gourds—experimental
observations. Amer. Jour. Bot. 41:697-700.

Whitaker, T. W., and G. F. Carter. 1961. A note on the longevity of seed of
Lagenaria siceraria (Mol.) Standl. after floating in sea water. Bull. Torrey Bot.
Club 88:104-6.

Williams, E. 1969. The ecology of colonization as seen in the zoogeography of
anoline lizards on small islands. Quart. Rev. Biol. 44:345-89.

Wilson, E. O. 1959. Adaptive shift and dispersal in a tropical ant fauna. Evolution
13:122-44.

43

Principles of Dispersal and Evolution

Wilson, E. O., and D. S. Simberloff. 1969. Experimental zoogeography of islands.
Defaunation and monitoring techniques. Ecology 50:276-77.

Wilson, E. O., and R. W. Taylor. 1967. An estimate of the potential evolutionary
increase in species density in the Polynesian ant fauna. Evolution 21:1-10.

Yoshimoto, C. M., and J. L. Gressitt. 1959. Trapping of air-borne insects on ships
on the Pacific. II. Proc. Hawatian Entomol. Soc. 17:150-55.

Yoshimoto, C. M., and J. L. Gressitt. 1960. Trapping of air-borne insects on ships
on the Pacific. III. Pacific Insects 2:239-43.

Yoshimoto, C. M., and J. L. Gressitt. 1961. Trapping of air-borne insects on ships
on the Pacific. IV. Pacific Insects 3:556—-58.

Yoshimoto, C. M., and J. L. Gressitt. 1963. Trapping of air-borne insects in the
Pacific-Antarctic area, 2. Pacific Insects 5:873-83.

Yoshimoto, C. M., and J. L. Gressitt. 1964. Dispersal studies in Aphididae, Agromy-
zidae and Cynipoidea. Pacific Insects 6:525-31.

Yoshimoto, C. M., J. L. Gressitt, and C. J. Mitchell. 1962. Trapping of air-borne
insects in the Pacific-Antarctic area. 1. Pacific Insects 4:847—58.

Zimmerman, E. C. 1948. Insects of Hawaii, vol. 1: Introduction. Honolulu, Univer-
sity of Hawaii Press.

Zimmerman, E. C. 1957. Insects of Hawaii, vol. 6: Ephemeroptera-Neuroptera-Tri-
choptera. Honolulu, University of Hawaii Press.

Zimmerman, E.. C. 1960. Possible evidence of rapid evolution in Hawaiian moths.
Evolution 14:137-38.

Zwaluwenberg, R. H. van. 1942. Notes on the temporary establishment of insect
and plant species on Canton Island. Hawatian Planter’s Rec. 46:49-52.

APTER
TWO

LONG-DISTANCE DISPERSAL:
EVIDENCE AND IMPLICATIONS

el

The majority of biologists hold that oceanic islands have acquired their
floras and faunas by means of long-distance dispersal, and agree that these
islands include the volcanic islands of the Pacific (1.e., those north and
east of the so-called andesite line) as well as those of the Atlantic (the
Macaronesian islands—from the Azores to the Cape Verde Islands, and
including Madeira and the Canary Islands—and such islands as St.
Helena and Tristan da Cunha) and Indian Ocean (e.g., the Mascarene
islands). In addition, once-glaciated islands (the so-called subantarctic
islands, for example) are recipients of waif immigrants in the same way.
Older islands known to have been strongly isolated for very long periods
of time (New Zealand, New Caledonia) have received at least an appre-
ciable portion and probably the vast majority of their immigrants by long-
distance dispersal. The equatorial mountaintops, discussed in chapter 14,
also have been populated in this manner, as have areas of other distinctive
habitats that are separated from each other by marked disjunctions.
Reasons for belief in the operation of long-distance dispersal to oceanic
islands lie in floristics and faunistics (lines of evidence reviewed by Thorne
'1963] and others) and knowledge of the biology of the organisms. Geo-
logical research provides no rational evidence for hypothetical land links
between the volcanic islands cited above and mainland areas. Geological
studies also show how short-lived volcanic islands are. A review of the
NOTE: This chapter is a revision of a paper. (Carlquist, 1967). For permission to

present materials from this paper, acknowledgment is gratefully extended to the
Bulletin of the Torrey Botanical Club.

45

Long-Distance Dispersal: Evidence and Implications

literature is unnecessary here; the reader can consult chapter 1 and the
literature cited by Thorne (1963) and Carlquist (1970).

Other lines of evidence or inference are related to fruit and seed
morphology, or to various adaptations for dispersibility in animals. Con-
tributions on plant dispersal by Guppy (1906) and Ridley (1930) are
especially noteworthy; Zimmerman (1948) reviews some persuasive obser-
vations concerning animal distribution. In recent years, the reality of
aerial dispersal of insects and spiders has been established with remark-
able clarity by Gressitt, Yoshimoto, and their co-workers (see chapter 1
for full citation). Attention should also be called to a thoughtful and
ingenious paper by Hurst (1969). Other papers are cited under appropriate
categories below. Although I feel that the reality of long-distance dispersal
is overwhelming, obviously much more remains to be learned about the
mechanisms by which it takes place. Many advances have been made in
recent years, however, and certain aspects seem worthy of examination at
this time, both to demonstrate what is at hand and to indicate what new
types of evidence might be most meaningful.

Land-Animal Dispersal

Dispersal of insects and spiders over long distances by air has been well
established. There seems to be skepticism by a few biologists, however,
concerning dispersal of land birds. This may surprise some amateurs, who
probably regard the flying ability of land birds as an obvious mechanism of
dispersal. Yet Deignan (1963) is so pessimistic about dispersal of land
birds to islands that he claims that contemporary patterns of distribution
of land birds on islands will predict what geologists must inevitably find
concerning former land connections. To those familiar with observational
data on land birds, this is simply not true. For example, species of Dar-
win’s finches do reach some of the islands in the Galapagos where they
never are able to form breeding colonies (Lack, 1969). These examples
and others (see Lack, 1971) show that land birds, although they are indeed
“sedentary,” as Mayr (1940) claimed them to be, are not as sedentary as
the more conservative zoologists would have us believe. A compelling in-
stance—illustrating how careful observation pays large dividends—is offered
by Bannerman and Bannerman (1965). They report that Madeira has only

46

Long-Distance Dispersal: Evidence and Implications

27 resident species of land birds, compared with about 200 in Morocco.
Of the 200 species in Morocco, however, the Bannermans have observed
no fewer than 69 of those not resident on Madeira to appear as occasional
visitors on the island. The distance between Madeira and Morocco is
great enough that we can extrapolate to islands at large, and conclude that
occasional visits by land birds are easily sufficient to explain establishment
of insular breeding colonies. We must remember that there have been
relatively few islands on which ornithologists have attempted to record
occasional visitors or stragglers, so skepticism is completely unjustified.

The difficulty in establishment of land birds on islands seems much
greater than that of dispersal to islands. One establishing species may
shut out subsequent potential residents. An interesting possible example
is cited by Lack (1971). Of the two European species of kinglets, the
goldcrest (Regulus regulus) has colonized the Azores, whereas the firecrest
(R. ignicapillus) has colonized the Canary Islands and Madeira. Very likely,
the goldcrest was the first to reach the Azores, whereas the firecrest reached
the Canaries and Madeira first. Each colonization forestalled establish-
ment of the other species within an archipelago. Alternatively, one could
speculate that some subtle ecological distinction has favored the goldcrest
on the Azores and the firecrest on the Canaries and Madeira.

Important new data on problems of dispersal and establishment of
animals on islands are offered in the work of Williams (1969) on anoline
lizards in the West Indies. Williams emphasizes that vagility is linked to
habitat versatility. In other words, lizards of lowland dry forest and mar-
ginal areas will be able to travel and withstand desiccation better than
will upland lizards; the lowland lizards also, of course, are likely to arrive
upon terrain more nearly like that of their source area. Wilson (1959) has
emphasized a similar situation for ants, and the same is apparent to me in
the case of insular plants: most of the wet-forest species seem to be deriva-
tives (although one could cite exceptions, such as Cyrtandra in the Hawai-
ian Islands). Williams (1969) describes evolution into more stable or mesic
habitats as “decay of versatility.” One might also mention that not only
are dry lowland areas larger than wet-forest areas (if any) on most islands,
but they are also nearer the site of origin for a dispersal event (presumably
by rafting in the case of lizards, and often by shorebirds or even marine
birds for plants).

ai)

Long-Distance Dispersal: Evidence and Implications

Williams gives considerable attention to problems of coexistence and
competition. He, like Lack, finds that some anoline species are denied
success in colonization of a particular island because their ecological re-
quirements are too similar to those of an already established occupant.
Noting. the critical difficulty in the first phases of establishment because of
small population size, Williams feels that a potential colonist is highly
disadvantaged unless a competing species is completely absent, or a pre-
viously established resident (anoline species) occupies a wholly different
ecological niche. Even a potentially superior invader, he feels, might not
succeed against an already established colonist species with a “safe” popu-
lation level. Coexistence is possible if several colonizing species have been
preadapted in sympatry on complex source islands and then migrate simul-
taneously to a new island. “Ecological release” can occur on small islands:
where two species of anoles cannot be supported, the one successful
colonist will tend to broaden its ecological range. This evolutionary ten-
dency might seem to be beyond the scope of the present discussion, but a
species that has undergone ecological release might well forestall estab-
lishment by newer immigrants—perhaps even on the larger islands.

Williams notes that although we can easily see cases in which several
species have established, we cannot readily visualize exclusion of other
species by a successful colonist. He stresses that we have been too con-
cerned with the possibility of competition as a source of “character dis-
placement” between competing species during autochthonous develop-
ment on a single island. While character displacement may occur among
large continental populations (Brown and Wilson, 1965), congeners on an
island may have developed their mutually exclusive ecological require-
ments prior to one or both invasions of that island.

Williams is to be congratulated for his method, which was “to derive
all statements from an empirical basis’—the study of actual Anolis popu-
lations. I do object, however, to his erroneous interpretation of my (1966a)
paper, implying that I believe in lizard colonization of islands by means
of “multiple rafting.” I did not use that term, nor did I suggest any such
possibility for lizards, which I did not mention in that context at all. I
was merely attempting to suggest that, for animals suited to dispersal, a
single individual on a raft is not necessarily sufficient, nor is a single seed
brought by a flock of birds invariably the founder of a plant colony. A

48

Long-Distance Dispersal: Evidence and Implications

colonizing species may well invade by means of a single dispersal event,
but this may not necessarily involve only one individual. There is nothing
in Williams’s data to preclude the possibility that two gravid Anolis
females of a given species, rather than a single individual, successfully rode
a raft to a particular island. The results would be identical, although sev-
eral individuals brought by a single dispersal event would, potentially,
bring the presumptive advantage of greater heterozygosity. We would
probably not be able to measure whether a contemporary population
stemmed from one gravid female or three. If rafting is indeed a chance
event—as it must be, like other types of long-distance dispersal—what are
the chances that every “raft” contains only a single gravid female? It is
much more likely that some rafts contain one individual, but other rafts
several individuals that establish (see chapter 1, principle 3).

Another problem in establishment of land animals on islands is that
their ecological requirements may be met but they may gradually overeat
their food supply and then vanish. Such instances have been reported in
recent time in the case of the Australian duck, Aythya australis, a natural
immigrant to New Zealand, which died out within 30 years after first
being observed (Fleming, 1962); and for the reindeer, introduced by man
to St. Matthew Island, Bering Sea (Klein, 1968). This also apparently
happened in the “rabbit plagues” of Laysan and Lisianski (Hawauan
Leeward Is.); self-extinction and near self-extinction had occurred when
the Tanager Expedition visited these islands in 1922. Thus, for reasons
such as those cited here, Lack (1971) finds that some islands regarded as
“unsaturated” may, in fact, be “saturated” with suitable species.

With regard to land molluscs, there seems to be relatively little objec-
tion to possibilities of long-distance dispersal. Examples of attachment to
birds have been cited (see Zimmerman, 1948). The work of Malone
(1965b) is pertinent here. Gastropod eggs may be more resistant to sea-
water or desiccation than suspected; adults can, of course, resist desiccation
to varying degrees. The relatively large size of certain insular gastropods
may deter some biologists from believing in their dispersibility, but altera-
tion in size of land molluscs has undoubtedly taken place on islands, and
we have every reason to believe that gigantism has occurred during evolu-
tion of some of them (see chapter 15). A highly interesting exercise in this
regard would be a study of size distributions of continental land molluscs,

49

Long-Distance Dispersal: Evidence and Implications

insular land molluscs, and—most importantly—size of types likely to have
been ancestral to contemporary insular land molluscs. Results of such a
study might well show what the work of Gressitt and co-workers has shown
in insects: that insular populations show establishment of smaller repre-
sentatives predominantly.

Experimental work on colonization of animals has been done by
Wilson and Simberloff (1969) and Simberloff and Wilson (1969). Their
papers show the rapidity with which colonization can take place on nearby
islets, and are admirable examples of how experimental techniques can be
used in dispersal studies. These studies on islets of the Florida Keys have
been used as the basis of theoretical models by Simberloff (1969).

Plant Dispersal

In this chapter, information and speculation on long-distance dispersal of
plants have been selected and placed in a context somewhat different from
those employed by other authors. The modes of dispersal likely to have
resulted in the present native floras of Pacific islands are the primary
object of analysis here. The correlations that result are informative and
somewhat surprising, with many implications concerning dispersal to
areas other than those detailed, and dispersal of groups other than vas-
cular plants.

By calculating the proportions of a specific flora that have presumably
arrived by air flotation, internal transportation in birds, etc., and by com-
paring the percentages with those of the same categories in floras of other
islands, we can see that the ecology of the recipient island is tremendously
important. Distance from source area is to some extent secondary in
determining which modes of dispersal are effective in populating a par-
ticular island. Such problems as so-called succession on lava flows prove
to be related to differential dispersal ability (fig. 2.1, top left, nght). Dis-
persibility is the key to the surprising floristic and faunistic affinities be-
tween Polynesian and Mascarene islands. With the precedent of evidence
on dispersal to oceanic islands, many supposed difficulties in explanation
of inter- and intra-continental disjunctions seem much more easily soluble.

50

Long-Distance Dispersal: Evidence and Implications

Data on dispersal methods have been drawn from various sources,
such as Guppy (1906), Ridley (1930), Fosberg (1963) and my own obser-
vations. In using fruit and seed morphology as an indicator of probable
dispersal method, one can hypothesize a certain degree of loss of dis-
persibility during autochthonous evolution on islands, but only in par-
ticular taxonomic groups and in particular islands, such as the Hawauan
Islands (see chapter 11 for a discussion of loss of dispersibility in plants).
Direct observations of plant dispersal to islands (other than via oceanic
drift) are virtually impossible, so indirect lines of evidence must be used.
One might have expected that aerial trapping experiments, such as those
used by Gressitt, Yoshimoto, and co-workers in studying insect dispersal
might be useful to botanists. This is not true for flowering plants, because,
according to the data below, aerial flotation of seeds appears to be insignifi-
cant except rather close to source areas. There are, to be sure, a number
of papers dealing with aeroplankton (Bold, Brown, and Larson, 1964;
Brown, 1971; Chang, 1967; Fulton, 1966a, 1966b, 1966c; Fulton and
Mitchell, 1966; Gregory and Monteith, 1967; Heise and Heise, 1948;
Hirst, Stedman, and Hogg, 1966; Hirst, Stedman, and Hurst, 1967; Kohl-
meyer, 1966; Kramer and Pady, 1968; McDonald, 1962; Maynard, 1968;
Moar, 1969; Overeem, 1937; Schlichting, 1964). Although these papers
do not deal with seeds, many of their considerations may well apply to
small airborne seeds, so they cannot be disregarded in this connection.
With respect to spore-bearing plants, the papers just cited form over-
whelming evidence that long-distance dispersal is a reality.

Observations by ornithologists give useful hints about dispersal of
plants by birds. For example, the bibliography of McAtee (1947) is inter-
esting, because data in the papers he cites were not collected with the
intention of stressing or “proving” long-distance dispersal of plant dis-
seminules by birds. Such observations are indicative of how dispersal via
birds might usefully be investigated further. The importance of dispersal
via birds suggests to me that botanists interested in dispersal problems
must turn intensively to observational and experimental work with birds
if methods of long-distance dispersal are ever to be “demonstrated.” We
must be realistic in such investigations, however. Obviously, many kinds
of birds probably have no function in long-distance dispersal. The papers

51

Long-Distance Dispersal: Evidence and Implications

cited by McAtee (1947) all deal with land birds, for example. On the
other hand, waterfowl and especially shorebirds seem likely vectors. The
experimental evidence (Proctor, 1968; Vlaming and Proctor, 1968) on the
potential effectiveness of shorebirds is outstanding in showing that poten-
tial transport is far greater than many workers would have suspected. Other
interesting papers in relation to bird dispersal of seeds include those of
Guppy (1897), Low (1937), McAtee (1918), and Roessler (1936). Con-
tributions allied to this topic include those of Sauer (1957) and Wagner
(1965). |

Ironically, information on dispersal of small disseminules by birds is
relatively abundant. Among recent papers on dispersal of algae by birds
are those by Malone (1965a), Proctor (1961, 1962, 1966), Proctor and
Malone (1965), and Proctor, Malone, and Vlaming (1967). Likewise, we
are not lacking in observations on dispersal of fungi by birds, thanks to
studies by Pugh (1964a, 1964b, 1965a, 1965b, 1966a, 1966b) and the
workers he cited. The irony in these observations is that algae and fungi
are easily dispersed by wind (see aeroplankton papers cited earlier), so
that bird vectors are not as significant for study as they would be in the
case of seed transport. Of course, algae and fungi on and in birds can be
relatively easily studied, whereas occurrence of seeds in birds suited to
long-distance transport of flowering plants is difficult to study in the wild,
or in collected birds.

Processes of long-distance dispersal in plants are thoroughly credible,
but only if unsupportable hypotheses can be eliminated. We can no longer
accept Guppy’s idea that different taxonomic groups dispersed at different
times (e.g., “age of ferns,” “age of Compositae”). Nor can we agree with
the full implications of Guppy’s claim that “agencies of dispersal. . .
have ceased to be active” in the Pacific. High islands may well have existed
where only atolls or seamounts now stand, and such former high islands
may once have served as subsidiary source areas or stepping-stones for
populating more remote islands. Nevertheless, the birds that function
now in dispersal are probably much the same as those that have done so
in previous eras of the relatively recent geological past, and Guppy’s sug-
gestion that now-vanished, and presumably migratory, fruit-eating pigeons
were vectors in dispersal of plants must be largely dismissed as visionary.

52

Long-Distance Dispersal: Evidence and Implications

Although noting the improbability of some of Guppy’s ideas, one must
concede that his work provides most of the information we have on dis-
persal of Pacific plants, and the contributions since his are relatively few.
Guppy’s survey of the Hawaiian flora with respect to methods of dispersal
is understandably incomplete. If the Hawaiian flora arrived entirely via
long-distance dispersal, however, we must ultimately be able to account
for each and every arrival with respect to dispersal. We must similarly be
able to account for other Pacific floras. The geographic sources of insular
floras are reasonably well understood, thanks to the work of Fosberg
(1948), Balgooy (1960), and Thorne (1963). We need comparable infor-
mation (which must, again, be speculative) on the probable means of
arrival of the plant immigrants to oceanic islands.

The tabulations and computations presented here for the Hawaiian
flora were influenced by Fosberg’s (1948) tabular summary of Hawaiian
angiosperms. Fosberg’s list presents geographical sources, numbers of
immigrants, and numbers of contemporary species. Using his immigrant-
number estimates (with certain alterations), I computed percentages for
seven methods of arrival. Only clearly native taxa were included. Lists
similar to Fosberg’s were then prepared for other Pacific floras, and per-
centages of the methods of arrival were estimated for each of those floras.
Admittedly, some floras are better known than others, and my specula-
tions on modes of arrival may be incorrect in particular cases. Errors may
or may not tend to balance out within a flora, but I doubt that markedly
different figures for particular modes of arrival would be obtained by
other workers.

The following floras were used as sources for making the estimates:
San Clemente I., Calif.: Raven (1963a); Revillagigedo Is.: Johnston
(1931); Cocos I.: Stewart (1912); Galapagos Is.: Stewart (1911); Desven-
turadas Is.: Skottsberg (1937, 1963); Juan Fernandez Is.: Skottsberg
7p 1973 1951); Easter 1.2 Skottsbere, (19224, 1927, 1951; Oeno’ I:
St. John and Philipson (1960); Rapa I.: Brown (1931, 1935); Society Is.:
Drake del Castillo (1893), Moore (1933, 1963); Rarotonga: Cheeseman
(1902); Tonga: Yuncker (1959); Samoa: Reinecke (1898), Christophersen
(1935, 1938); Marquesas Is.: Brown (1931, 1935); Equatorial Is.: Chris-
tophersen (1927); Laysan I.: Christophersen and Caum (1931), Lamoureux

53

Long-Distance Dispersal: Evidence and Implications

(1963); Hawaiian Is.: Hillebrand (1888), Christophersen and Caum
(1931), Fosberg (1948). Certain floras were not used if too few species oc-
curred for an accurate picture of modes of arrival (e.g., Clipperton I.).

Analysis of the Hawatian Flora

Concepts of means of arrival are presented in the tabular summary that
follows. It is offered so that my decisions on probable arrival method for
each introduction to the Hawaiian flora can be examined. Only native
seed plants are considered. Vascular cryptogams may be assumed to have
arrived via air flotation of spores, with a few exceptions, such as Marsilea
(Malone and Proctor, 1965), although spores can travel via birds as well.

The desirability of using the Hawauan biota as a source for study of
biological phenomena related to oceanic islands has been stressed pre-
viously. In the present study, the fact that the Hawaiian Islands are excep-
tionally remote yet have ecological diversity broad enough to support a
wide variety of immigrants is significant. That the Hawauan Islands have
acquired their native flora by means of long-distance dispersal has not been
seriously challenged. Whereas now-vanished archipelagos may have aided
dispersal from Indo-Malesian source areas, no such former islands have
been hypothesized between the Hawaiian Islands and North America.
American sources must be claimed for no fewer than 50 successful arrivals
in the Hawaiian angiosperm flora. If long-distance dispersal can explain
these 50 (which represent, in fact, all seven modes of transport described
below), then dispersal over shorter distances can surely be explained as
well.

For each genus, interpretations are given in parentheses. In each case
the first figure indicates the number of successful introductions needed to
account for the native species now resident. This may be considered as the
minimum number of introductions necessary, and a somewhat larger
number is conceivable in some instances. Next is an abbreviation indicat-
ing presumed mode of arrival. The abbreviations are:

A = Atr flotation
BB = Birds—seeds or fruits attached to feathers mechanically by
barbs, bristles, awns, trichomes, etc.

54

Long-Distance Dispersal: Evidence and Implications

BI = Birds—seeds or fruits eaten and carried internally

BM = Birds—seeds or fruits embedded in mud on feet

BV = Birds—seeds or fruits attached to feathers by viscid sub-
stances, or viscid quality of fruit parenchyma surrounding
seeds

DF = Oceanic drift—frequent or repetitive, as suggested by
demonstrable ability of fruits or seeds to float for pro-
longed period

DR = Oceanic drift—rare, on one or only a few occasions; seeds
or fruits probably resistant to seawater but unable to float
for prolonged periods and thus likely to have arrived via
rafting; or, alternatively, plants suited to oceanic drift,
but rare arrivals because of unsuitable patterns of ocean
currents

The last item within each set of parentheses is the estimated number
of contemporary species plus subspecies and varieties, both endemic ones
(e) and wide-ranging ones (w) that are native to one or more areas in addi-
tion to the Hawaiian Islands.

In some cases, more than one type of dispersal occurs within a genus.
For example (contrary to Fosberg, 1948), I hold that at least three intro-
ductions rather than one are necessary to account for the Hawaiian species
of Scaevola (Goodeniaceae). Oceanic drift has brought the common beach
species, Scaevola S. taccada (fig. 2.6, top left, night). Scaevola glabra (in-
cluding S. kauaiensis) represents an early and rather distinctive immigrant
(regarded as a segregate genus, Camphusia, by some) in which fruit and
seed gigantism has occurred autochthonously (Carlquist, 1966c; see chapter
11), as have other morphological changes. The remainder (e.g., S. gaudt-
chaudiana, fig. 2.3, top right, and S. coriacea, fig. 2.6, top right) of the
Hawaiian species stem from a third stock.

If a clear decision as to mode of arrival could not be reached, a prefer-
ence is expressed by indicating one method in upper case letters, the
alternative in lower case. If two arrivals in a genus are claimed to have
come by different means, both choices are given in upper case.

Mode of arrival was computed not only for ancestors of the entire
angiosperm flora of the Hawaiian Islands, but also for those segments that

55

Long-Distance Dispersal: Evidence and Implications

came from each of the source areas suggested by Fosberg. By this means I
hoped, for example, to see if immigrants from America and from Indo-
Malesia came predominantly by different methods. In fact, no such differ-
ences emerged, and the methods of long-distance dispersal seem relatively
alike in different geographical areas. Among the few figures that appear to
show significant deviation are the relatively high number of arrivals in mud
on birds’ feet (6 of them) and the low number internally in birds (1) from
boreal regions. Among the pantropical species, repetitive oceanic drift
rates notably high (11 introductions out of 27), which merely shows that
many of the pantropical species have achieved their wide distribution by
virtue of the great efficiency for wide dispersal that is characteristic of
oceanic drift.

Another type of computation proved somewhat more significant, and
these results are given at the end of the compilation of genera. For each
of the seven modes of arrival recognized, the number of contemporary
Hawaiian species was divided by the number of immigrants hypothetically
necessary to account for those species. Thus, a ratio denoting degree of
speciation in relation to each mode of arrival could be obtained. This
figure is discussed below under various dispersal headings.

MONOCOTYLEDONS. Pandanaceae: Freycinetia (1, BI, le); Pandanus
(1, DF, lw). Potamogetonaceae: Potamogeton (2, BI, 2w). Ruppiaceae:
Ruppia (1, BM, lw). Naiadaceae: Naias (1, BM, lw). Hydrocharitaceae:
Halophila (1, DF, lw). Poaceae: Agrostis (2, BB, 2e, lw); Andropogon
(Fs Bs te ; Calamagrostis (2, BB, 2e); Cenchrus (1, BB, 3e); Deschampsia
(1, BB, 3e); Digitaria (1, BB, lw); Dissochondrus (1, BB, le); Eragrostis
(1, BI or . 12e); pera ‘1, BB, le); Garnotia (1, BB, le); Heteropogon
(1, BB, lw); Isachne (2, BB, le, lw); Ischaemum (1, BB, lw); Lepturus
(LSD aay.) Micerotcete (eSB Beal): Oplismenus (1, BB, lw); Panicum
(3, BI, BB, BM, 23e;) Paspalum (2, BB, 2w); Poa (1, BB, 4e); Sporanolte
(1, BY,, lw); ie (1, BB, 2e). Cyperaceae: Carex (6, BM, BI, 6e, 3w);
Cladium (and derivatives Baumea and Vincentia) (3, BI, 2e, lw); Cyperus
(8, BI, BM, 16e, I); Eleocharis (1, BM, lw); Fimbristylis (1, BM or bi,
le, lw); Gahnia (2, BI, BM, 6e); Oreobolus (1, BB, le); Rhynchospora (2,
BI or bb, le, 2w); SGirpus (5, BM or BI, le, 4w); Scleria (1, BM, le); Unez-
nia (1, BB, lw). Arecaceae: Pritchardia (1, BI, or dr, 2—30e). Flagellaria-

56

Long-Distance Dispersal: Evidence and Implications

ceae: Joinvillea (1, BI, le). Juncaceae: Luzula (1, BM, 3a). Liliaceae:
Astelia (1, BI or by, 12e); Dianella (1, BI, 3e); Dracaena (Pleomele) (1, BI,
Ze); Smilax (1, BI, 2e). Iridaceae: Sisyrinchium (1, BM, le). Orchidaceae:
Anoectochilus (1, A, 2e); Habenaria (1, A, le); Liparis (1, A, le).
DICOTYLEDONS. Piperaceae: Peperomia (3, BV, 48e, 2w). Ulmaceae:
Pseudomorus (1, BI, lw); Trema (1, BI, lw). Urticaceae: Neraudia (1, BI,
fe) tea (1, BM, lw); Pipturus (1, BV, l3e); Touchardia (1; Bl, le);
Urera (2, BI, 3e). Santalaceae: Exocarpus (1, BI, 3e); Santalum (2, BI, 6c).
Loranthaceae (Viscaceae): Korthalsella (2, BV or BI, 6e, 2w). Polygona-
ceae: Polygonum (1, BI, lw); Rumex (1, DR or bi, 2e). Chenopodiaceae:
Chenopodium (1, BM, le). Amaranthaceae: Achyranthes (1, BB, 2e);
Aerva (1, BB, le); Charpentiera (1, BI or bm, 3e); Nototrichium (1, BB,
5e). Nyctaginaceae: Boerhavia (3, BV, 3w); Pisonia (3, BV, 3e). Phyto-
laccaceae: Phytolacca (1, BI, le). Aizoaceae: Sesuvium (1, DF, lw). Por-
tulacaceae: Portulaca (3, DF, 4e, 2w). Caryophyllaceae: Sagina (1, BM,
le); Schiedea (and Alsinodendron) (1, BM, 32e); Silene (1, BI or BM, 5e).
Ranunculaceae: Ranunculus (1, BM or bb, 2e). Menispermaceae: Coccu-
lus (1, BI, 2e). Lauraceae: Cassytha (1, DF, lw); Cryptocarya (1, BI, le).
Capparaceae: Capparis (1, DR or dr, lw). Brassicaceae: Lepidium (2, BV,
4e). Droseraceae: Drosera (1, BM, lw). Saxifragaceae: Broussaisia (1, BI,
2e). Pittosporaceae: Pittosporum (1, BV, 30e). Rosaceae: Acaena (1, BB,
2e); Fragaria (1, BI, lw); Osteomeles (1, BI, lw); Rubus (1, BI, 2e). Faba-
ceae: Acacia (1, DR, 3e); Caesalpinia (2, DF, 2w); Canavalia (1, DF or
DR, 2e); Sophora (1, DR, 4e); Sesbania (1, DR, le); Strongylodon (1, DF,
lw); Vicia (1, BI, le); Vigna (2, DF, 2e, lw). Geraniaceae: Geranium (1,
BB, 6e). Zygophyllaceae: Tribulus (1, BB, lw). Rutaceae: Zanthoxylum
(Fagara) (1, BI, 14e); Pelea (1, BI, 40e); Platydesma (1, BI, 5e). Euphorbi-
aceae: Antidesma (1, BI, 4e); Claoxylon (1, BI, 2e); Drypetes (1, BI, le);
Euphorbia (1, BI, 30e); Phyllanthus (1, BI, 2e). Aquifoliaceae: Ilex (1, BI,
le). Celastraceae: Perrottetia (1, BI, le). Anacardiaceae: Rhus (1, BI, le).
Sapindaceae: Alectryon (1, BI, 2e); Dodonaea (1, DF or DR, 3e, 1 w); Sa-
pindus (1, BI, le). Rhamnaceae: Alphitonia (1, BI, le); Colubrina (2, DF,
le, lw); Gouania (1, DR, 3e). Elaeocarpaceae: Elaeocarpus (1, BI, 1e).
Malvaceae: Abutilon (2, DF, 2e, lw); Gossypium (1, DR, le); Hibiscadel-
phus (1, DR, 4e); Hibiscus (4, DR, 10e); Kokia (1, DR, 4e); Sida (2, DF,
2w). Sterculiaceae: Waltheria (1, DR, 1w). Theaceae: Eurya (1, BI, Ze).

57

Long-Distance Dispersal: Evidence and Implications

Thymeleaceae: Wikstroemia (1, BI, 20e). Flacourtiaceae: Xylosma (1, BI,
Ze). Cucurbitaceae: Sicyos (1, BB or dr, 10e). Myrtaceae: Syzygium (Euge-
nia) (2, BI, 4e); Metrosideros (1 or 2, A, 5e). Begoniaceae: Hillebrandia (1,
BM, le). Haloragaceae: Gunnera (1, BI, 7e). Araliaceae: Cheirodendron
(1, BI,.5e); Reynoldsia (1, BI, le); Tetraplasandra (1, BI, 10e). Apiaceae:
Peucedanum (1, DR, 3e); Sanicula (1, BB, 4e). Ericaceae: Vaccinium (1,
BI, 8e). Epacridaceae: Styphelia (1, BI, 2e). Primulaceae: Lysimachia (2,
DF, BM, 12e, lw). Myrsinaceae: Embelia (1, BI, le); Myrsine (1, BI, 25e).
Sapotaceae: Nesoluma (1, DR or bi, lw); Planchonella (1, BI, 6e). Ebena-
ceae: Diospyros (Maba) (1, BI, 7e). Plumbaginaceae: Plumbago (1, BV,
Iw). Loganiaceae: Labordia (2, BI, 40e). Oleaceae: Osmanthus (Olea,
Negrestis) (1, BI, le). Apocynaceae: Alyxia (1, BI, le); Ochrosia (1, DR,
le); Pteralyxia (1, DR, 2e); Rawwolfia (1, BI, 7e). Convolvulaceae: Brew-
eria’ (1, BV or dr, 2e); Cressa (1, DF, 1 w); Cuscuta (1, DE orBie)r
Ipomoea (7, DF, 4e, 5w); Jacquemontia (1, DR, le). Hydrophyllaceae:
Nama (1, BB, le). Verbenaceae: Heliotropium (2, DF, le, lw). Lamia-
ceae: Lepechinia (1, BV, lw); Phyllostegia (1, BI, 40e); Stenogyne (1, BI,
40e); Haplostachys (1, DR, 3e); Plectranthus (1, BB, 1 w). Solanaceae: Ly-
cium (1, DF, lw); Nothocestrum (1, BI, 6e); Solanum (1, BI, 7e). Scrophu-
lariaceae: Bacopa (1, BM, lw). Myoporaceae: Myoporum (1, BI, le). Ges-
nerlaceae: Cyrtandra (1 or 2; BI, 130e). Plantaginaceae: Plantago (2, BV,
19e). Rubiaceae: Bobea (1, BI, 4e); Canthium (1, BI, lw); Coprosma (3,
BI or bv, 27e); Gardenia (2, BI, 2e); Hedyotis (and Gouldia) (1, BM or BI,
50e); Morinda (1, DR, le, lw); Nertera (1, BI, lw); Psychotria (1, BI, 1w);
Straussia (1, BI, 10e). Campanulaceae: Brighamia (1, BI, le); Clermontia
(1, BI or BV, 30e); Cyanea, Delissea, and Rollandia (1, BI or BV, 48e);
Lobelia (1, BM, 15e); Trematolobelia (1, BM, 3e). Goodeniaceae: Scae-
vola (3, BI and DR, 9e, lw). Asteraceae: Adenostemma (1, BV, lw);
Argyroxiphium, Dubautia (incl. Railliardia), and Wilkesia (1, BV or BB,
37e); Artemisia (1, BV, 3e); Bidens (1, BB, 30e); Hesperomannia (1, BB,
Se); Lagenophora (1, BV, 3e); Lipochaeta (1, BB, 35e); Remya (1, BB, 2e);
Tetramolopium (1, BB, 10e).

Ratios of numbers of contemporary species in the Hawaiian flora to
the numbers of postulated immigrant antecedents for the various dispersal
types are as follows:

58

Long-Distance Dispersal: Evidence and Implications

AS 23 BVA EA

BB: 4.7 Birds (all means): 5.7
BI: 6.0 Diese
BM: 4.9 DRe Zs

Ratio for angiosperm flora as a whole: 4.8

By way of footnotes, the following comments can be offered. A num-
ber of species included in Fosberg’s list have been omitted because I feel
they are dubiously native (perhaps this was unwise: Argemone may be
native and the species endemic). The Dubautia-Argyroxiphium-W ilkesia
complex can now be regarded as American‘in affinities. Lipochaeta is also
clearly American rather than Indo-Pacific in origin. Bidens is probably
better regarded as American than Austral, despite its occurrence in the
South Pacific. Trematolobelia probably does not stem from the same event
of introduction as do the Hawaiian species of Lobelia.

Comments on Modes of Dispersal in the Hawaiian Flora

AIR FLOTATION

In the Hawaiian angiosperm flora, air flotation appears to have been
of minimal importance. The only seed plants that probably arrived by this
means are Metrosideros (fig. 2.1, top right, bottom left) and the three
genera of orchids. Even these could conceivably have arrived in mud on
birds’ feet. The complexity of Metrosideros in the Hawaiian Islands might
be more readily explained if there has been more than a single introduc-
tion. Hybridization between earlier and later immigrants could account for
the existing patterns. Of course this complexity could equally well have
resulted from differentiation of Metrosideros stocks at different sites within
the Hawaiian Islands, followed by redispersal so that distinctive products
were juxtaposed and hybridized. The likelihood that Metrosideros seeds
(fg. 2.1, bottom left) reached the Hawaiian and other Pacific islands by
means of wind dispersal is heightened by the experimental work of Corn
(1970). Her experiments showed that wind velocities of only 3 to 12 mph
are required for M. polymorpha seeds to be carried aloft, and that the
seeds can survive temperatures of —30 C for at least 6 hr. However, she
also found that Metrosideros seeds could germinate after soaking in sea-
water for 39 days plus an additional 7 days during which seeds were dry.

59

esting dispersal by wind or by mud on birds’ feet. trop LEFT, Nephrolepis
1955 Puna lava flow, Hawaii. rop ricHt, Metrosideros polymorpha

FIG. 2.1. Examples sugg
exaltata (Davalliaceae)

(Myrtaceae) seedling; 1955 Puna lava flow (photographed 1965). Borrom LEFT, Metrosideros
polymorpha seeds compared to typewnter letters. BorromM RIGHT, Lobelia cardinalis (Cam-

panulaceae) seeds compared to typewniter letters.

60

Long-Distance Dispersal: Evidence and Implications

Of the three genera of orchids, Liparis and Habenaria have reached
many islands of the world. Among the reasons these two have attained a
wide distribution may be their simple pollination requirements. ‘The Ha-
walian climate seems suited to orchids (several species have naturalized),
but perhaps the small number of native species may be explained by lack
of suitable pollinators. Another possible factor that should be investigated
is whether orchid seeds are capable of resisting desiccation and cold tem-
peratures, such as would prevail in the upper atmosphere.

Composites (Asteraceae) are often thought to be well adapted for
wind dispersal, but the Hawaiian composites do not seem to have arrived
in this manner. Those best fitted for wind dispersal (Cichorieae, Senecio-
neae, Cynareae) are the groups notably absent. Those three tribes are, how-
ever, present in the Juan Fernandez Islands, which are much nearer to
mainland source areas. Wind dispersal may be possible for composites
with plumose pappus, but the limit for such types is probably much less
than the distance between a source area and the Hawaiian Islands. Even
species with plumose pappus can be envisaged as capable of transport
while enmeshed in bird feathers. Travel of fruits of Asteraceae in bird
feathers has, in fact, been observed under ordinary circumstances (Ridley,
1930). Fruits of Layia platyglossa (fig. 2.2, bottom left) seem to have this
capability, and it is noteworthy that Layia belongs to the subtribe Madi-
inae of Heliantheae, a subtribe represented in the Hawaiian Islands by
the endemic genera Dubautia, Argyroxiphium, and Wilkesia.

Ferns in the Hawaiian flora seem suited for air flotation of spores,
except for Marsilea and possibly, in another group of pteridophytes, Sela-
ginella, which has relatively large megaspores. Marsilea has been men-
tioned as a bird-dispersed plant (Proctor and Malone, 1965). One feature
of the Hawaiian fern flora that seems exceptional is the very high rate of
endemism (70.8 percent). Fosberg (1948) gives it as 64.9 percent, but divi-
sion of his total for endemic species plus varieties by his total for all native
fern species plus varieties yields 70.8 percent. Although this percentage is
lower than for Hawaiian angiosperms (94.4 percent according to Fosberg),
it is high by comparison with rates of endemism for ferns in other areas,
insular or not. Evidently transport to the Hawaiian Islands by air flotation
is not as easy or as frequent for fern spores as has generally been supposed.
That so few seed plants appear to have arrived by this means in Hawaii

61

Long-Distance Dispersal: Evidence and Implications

indicates the difficulty involved. Possibly spores of wet-forest ferns are
more easily killed by desiccation than those of other ferns, and the long
route to the Hawaiian Islands would prove fatal for many forest species
(see, for example, Lloyd’s studies cited in chapter 11). If arrival events of
Hawaiian ferns have in fact been infrequent and nonrepetitive (only a
single introduction of a species), then isolation leading to endemism could
occur.

The significance of air flotation within the present-day Hawaiian
flora is probably different from the relatively small role it played in bring-
ing angiosperms to the islands. A clear demonstration is the pattern of
vegetation that appears on new lava flows. Despite suggestions by various
workers that “succession” in the sense of Clements is involved, the entire
matter appears to relate primarily to dispersal ability. The supposed se-
quence from lichens to trees does not exist. Lichens, many ferns (fig. 2.1,
top left), and a scattering of angiosperms appear almost simultaneously.
Among the earlier angiosperms visible on lava flows are Metrosideros (fig.
2.1, top right), Dubautia (Railliardia) ciliolata or other Dubautia species,
Rumex giganteus, and Dodonaea viscosa (or segregate species). The pappus
of Dubautia permits it to cross relatively short distances by wind, and
Rumex and Dodonaea have winged fruits. These colonists probably arrive
on new flows by air flotation and tumbling, despite the likelihood that
they arrived ancestrally in the Hawaiian Islands in other ways, except for
ferns and Metrosideros.

Hawaiian genera with fleshy, attractive fruits, such as Styphelia, Co-
prosma, and Vaccinium also appear soon on Hawaiian lava flows. These
three genera are typically dispersed by frugivorous birds. That both native
and introduced birds are quite active in dispersal of seeds is shown by the
exceptionally rapid spread of such weeds as Psidium and Schinus. Conse-
quently, species with fleshy fruits may reach new lava flows only a short
time after the wind-dispersed species. A wide variety of species can grow
on new lava flows, and succession related to development of soil profiles
is not present. Indeed, the Kona forest on Hawaii is among the most
magnificent in the Hawaiian Islands, yet it grows on minimal soil of rela-
tively undecomposed lava flows.

The Hawaiian genus Trematolobelia is evidently an autochthonous
derivative of Lobelia. Fruit and seed morphology appear to have altered

62

Long-Distance Dispersal: Evidence and Implications

in the direction of greater efficiency for wind dispersal, at least over rela-
tively short distances (Carlquist, 1962). Circumstances suggest, however,
that Lobelia arrived in the Hawaiian Islands in mud on birds’ feet. Lobelia
seeds (fig. 2.1, bottom right) are relatively small, and many species of
Lobelia favor muddy habitats.

AVIAN TRANSPORT OF SEEDS AND FRUITS

Adherence of seeds and fruits to feathers by means of hairs, barbs,
bristles, or similar appendages is one of four methods suggested for trans-
port by birds. Indeed, it is the most distinctive of the four, and instances
can be recognized with the least probable error. In the Hawaiian flora,
most grasses appear to have arrived in this fashion. Other very likely candi-
dates include Uncinia (Cyperaceae: fig. 2.2, bottom nght), Sanicula
(Apiaceae), Acaena (Rosaceae), and Bidens (Asteraceae, fig. 2.2, top left,
right). Endemism is notably high in groups dispersed in this manner
(see table, p. 58). Of the genera just listed, the species native to the
Hawaiian Islands are also apparently endemic. This fact suggests that
transport of dry fruits by appendages caught in bird feathers occurs on an
infrequent or nonrepetitive basis (see Carlquist [1966a] for a discussion of.
“single dispersal events”; also chapter 1, principle 3).

In the cases of supposed internal transport of seeds and fruits by birds,
alternative interpretations are possible. When seeds are small and fruit
parenchyma is even moderately sticky, the possibility definitely exists that
seeds could be attached to head feathers when a bird eats portions of a
fruit. Such a small-seeded fruit occurs in Clermontia (fig. 2.3, top left), in
which the baccate fruit breaks open at maturity. Concomitantly with this
rupture, latex is released and provides an additional means of adhesion.
Also in this category are the other baccate Hawaiian lobelioids (Cyanea,
Rollandia, Delissea, and possibly Brighamia). Other groups in the Hawai- -
ian flora in which seeds are very small, and thus possibly transported ex- °
ternally as well as internally, include Vaccinium (characteristically a food
of the Hawaiian goose, according to Guppy [1906]), Labordia (and. its
South Pacific relative, Geniostoma), Cyrtandra, and Pipturus. In addition,
some groups with relatively large seeds have fruit parenchyma that is nota-
bly viscid; Korthalsella and Peperomia, for example.

Large-seeded fruits in the Hawaiian flora that are likely to be eaten by

63

5

FIG. 2.2. Examples suggesting dispersal in bird feathers by means of barbs or bristles. Top LEFT,
RIGHT, Bidens pilosa (Asteraceae). TOP RIGHT shows close-up of barbed awns and hairs on fruit;
fruits are 15 mm long. BOTTOM LEFT, fruits of Layia platyglossa (Asteraceae); fruits are 10 mm
long. BoTToM RIGHT, Uncinia tenella (Cyperaceae), hooked appendages on fruits; portion
shown is 3 mm long.

64

Long-Distance Dispersal: Evidence and Implications

birds are either relatively small fruits, if single-seeded, or large because of
probable gigantism that is presumably related to loss of dispersibility (Carl-
quist, 1966b, 1966c; see chapter 11). No single-seeded fruit longer than
1 cm need be postulated as ancestral for any angiosperm in the native,
bird-dispersed flora. Most are much smaller; those of Gunnera (fig. 2.3,
bottom right), for example. Fruits that are virtually black, as in Scaevola
(fig. 2.3, top right), fruits that are greenish and seemingly inconspicuous
(fig. 2.3, bottom left), and fruits in which accessory structures form fleshy
attractive organs (fig. 2.4, top right) are among the many types that are
adapted for effective dispersal by frugivorous birds. The contrast between
a black seed and a fleshy red aril makes Alectryon (fig. 2.4, top left) an
obvious candidate for dispersal by frugivorous birds, and even shiny, non-
fleshy, black seeds in dry capsules, such as those of Zanthoxylum (fig. 2.4,
bottom left) are often sought. Acquaintance with the Hawaiian flora
convinces one that adaptation to dispersal by frugivorous birds is abun-
dantly present. Even if some small-seeded fruits have been transported
externally on birds (and that is assumed here), the number of immigrants
brought internally in birds still remains the largest single category of dis-
persal in the Hawaiian flora (38.9 percent, representing at least 101 estab-
lishments). This percentage may seem high, but it is lower than the per-
centage of internally transported immigrants needed to account for other
Polynesian floras (fig. 2.7). The fact that internal transport by birds is the
only way one can account for two Juan Fernandez natives, Coprosma and
Santalum, demonstrates the tremendous capability of this dispersal
method. Drimys of the Winteraceae is often thought to be an old, relict
genus of angiosperms, but it has fruits ideally suited for birds to eat (fig.
2.4, bottom right). It has probably reached the Juan Fernandez Islands
from the South American mainland in this way.

That events of internal transport of seeds by birds are infrequent
seems indisputable, but one need not believe that they are therefore non-
existent. Nor should one or need one depend on highly improbable means
of transport via birds. Visits by typical frugivorous land birds to the Hawai-
ian chain are undoubtedly rare; among the stragglers listed by Munro
(1960) there are no birds like pigeons. The stragglers include mostly ma-
rine birds, waterbirds, and birds of prey. Many ducks and teals are in
Munro’s list and have been sighted with great regularity, according to

65

FIG. 2.3. Fruits probably dispersed by frugivorous birds. Top LEFT, Clermontia arborescens
(Campanulaceae), fruit 3 cm in diam., seeds visible; Lanai. top r1cHT, Scaevola gaudichaudiana
(Goodeniaceae), fruits 10 cm long; Kohahuanui, Oahu. Bsorrom LEFT, Phyllostegia racemosa
(Lamiaceae), fleshy seeds (each 8 mm wide) in groups of four; Saddle Road, Hawaii. BorroMm
RIGHT, Gunnera chilensis (Haloragaceae), fruits on inflorescence; each fruit 3 mm in diam.;
cultivated, Strybing Arboretum, Calif.

66

Long-Distance Dispersal: Evidence and Implications

ornithologists who have worked in the Hawaiian Islands (Eugene Kridler,
personal communication, 1966). Ducks could presumably carry a wide
variety of fruits or seeds—particularly from boreal source areas. Ducks
would be less likely to transport plants from Indo-Pacific regions, how-
ever.

How have the 53 hypothetical Hawaiian immigrants by internal
transport in birds arrived from Indo-Malesia? On the answer to this ques-
tion literally hinges the credibility of long-distance dispersal as a whole,
for if these introductions can be satisfactorily explained, others certainly
can also. We can, to be sure, imagine atolls that were once high islands
and that therefore enhance dispersal possibilities from Indo-Malesia to
the Hawaiian Islands. However, such islands are meaningless as stepping-
stones in dispersal unless suitable bird vectors exist. As a hint of possible
birds, we may consider Peale’s observation (reported by Guppy [1906})
that the principal food of the bristle-thighed curlew (Numenius tahitensis)
is, at least at certain times and in certain localities, fruits of Canthium
odoratum (Rubiaceae). Many fruits in the Hawaiian flora are comparable
to those of Canthium in size and morphology. No single-seeded fruit
larger than that of Canthium need be postulated as ancestral to the Ha-
wauan flora.

Many botanists probably would not expect shorebirds such as the
curlew to eat fruits and seeds to any appreciable extent. Such birds can be
omnivores, however, and Peale’s report is not an isolated one. One may
cite, for example, Munro’s (1960) report of consumption of seeds by such
shorebirds as the sanderling, the ruddy turnstone, and the stilt. Shorebirds
generally are wide ranging; the bristle-thighed curlew, for example, m1-
grates between Alaska and Tahiti; other shorebirds exceed that range.

Some will recall pessimistic estimates (e.g., by Ridley [1930]) on the
brief iength of time that elapses between ingestion and excretion of fruits
by birds. On more than one occasion, however, workers have emphasized
mitigating circumstances: the contents of a bird’s digestive tract may not
be completely emptied, and a few seeds may remain for longer periods;
birds may not always fly on empty stomachs, as some workers claim they
do; storms may alter habits of digestion, excretion, and preening from
what they are normally; and a few seeds may by chance adhere to anal
regions. Those who are dubious about the potential of internal transport

67

FIG. 2.4. Fruits probably adapted to dispersal by frugivorous birds. rop Lert, Alectryon sub-
cinereum (Sapindaceae), native of Australia, black seeds 7 mm in diam. surrounded by red
arils; cultivated, Los Angeles Arboretum. TOP RIGHT, Exocarpus gaudichaudii (Santalaceae),
fruits 7 mm long, each surrounded by white fleshy cup; Hualalai, Hawaii. BOTTOM LEFT,
Zanthoxylum (Fagara) (Rutaceae), fruits showing black seeds; Z. oahuensis, seeds 8 mm long,
above (Oahu), Z. kauaiensis below. Borrom riGHT, Drimys winteri (Winteraceae), fruiting
carpel (12 mm long) and abortive carpel; cultivated, University of California Botanic Garden,
Berkeley.

68

Long-Distance Dispersal: Evidence and Implications

of seeds by birds stress the rarity of its occurrence, not realizing that they
are thereby supporting an argument for its existence and effectiveness. If
this—or any other—kind of dispersal were frequent, isolation leading to
speciation and endemism would be impossible. A means of transport does
not need to be frequent to be operative.

The potential effectiveness of shorebirds in seed transport is remark-
ably demonstrated in the experimental work of Proctor (1968) and Vlam-
ing and Proctor (1968). They show that shorebirds have exceptionally long
retention of seeds—many seeds were retained more than 100 hr in their
experiments, some more than 300 hr. Shorebirds tend to retain larger hard
seeds (2-6 mm long) better than do other kinds of birds, according to
these authors. Shorebirds do not regurgitate pellets, a characteristic that
makes them excellent candidates for seed dispersal, and seeds are excreted
in a viable form (only viable seeds were reported in the paper by Vlaming
and Proctor [1968]). If one combines experimental data like these (which
ought to be extended, however), with known flight speeds of birds (e.g.,
from Cottam, Scooter, and Williams, 1942), long-distance dispersal of
seeds to islands by shorebirds can be said to progress from possibility to
the category of reality.

Where dispersal by birds is concerned, one must not dismiss pos-
sibilities out of hand. Preest (1963) discounts any possibility for long-dis-
tance transport of Nothofagus and podocarp seeds, but most of the species
he illustrates have small seeds capable of potential dispersal by birds, and
only Podocarpus ferrugineus has seeds too large for probable long-distance
transport. Seeds of P. ferrugineus may represent loss of dispersibility by
means of autochthonous seed gigantism of the sort described in chapter
11. We have no evidence that all Nothofagus and podocarp seeds fail to
withstand digestion by birds, or that they always have failed. Indeed, we
have evidence to the contrary: Podocarpus has reached Tonga, an archi-
pelago of oceanic islands.

The most elusive category of dispersal by birds is inclusion of seeds
in mud on birds’ feet (or some similar type of external adherence). Main
lines of inference leading me to claim this method of dispersal for a par-
ticular species include small seed size and the tendency for plants to grow
in marshes, mud flats, or along streams. Seeds that could become em-
bedded in mud might, to be sure, easily be eaten by birds and distributed

69

Long-Distance Dispersal: Evidence and Implications

as described in preceding paragraphs. For this and other reasons, arrival
in mud on birds’ feet is difficult to predicate with any certainty. This
method has nonetheless served for an estimated 12.8 percent of the estab-
lishments in the Hawaiian angiosperm flora. This percentage is higher
than that claimed on most other Polynesian islands, however (fig. 2.7).
The importance of this means of transport to the Hawauan Islands may
be related to the likelihood that ducks and other waterbirds assume a
greater role, if only by default, in long-distance dispersal here than they
do elsewhere. As a corollary, one might infer that transport internally in
birds is somewhat more difficult because of remoteness of the Hawaiian
Islands.

With respect to avian transport of seeds and fruits that have viscid
surfaces, one can cite such clear examples as Boerhavia (fig. 2.5, top left,
top and center right) and Pisonia (fig. 2.5, bottom left, right). Other ex-
amples are more subtle or perhaps show a lesser degree of efficiency for
this means of travel. The fact that Boerhavia diffusa or Pisonia umbellifera
ranges over wide stretches of the Pacific without even subspecific differen-
tiation among distant islands testifies to the excellence of this dispersal
device. Repetitive reintroduction that swamps distinctive characteristics
in incipient speciation evidently occurs (apparently the principles offered
by Williams [1967] with regard to difficulty in establishment by successive
invaders owing to small population size do not apply to strand plants,
although they do seem to apply in lizards; see chapter 1). Indeed, Boer-
havia could probably not persist on some remote atolls where it grows if
reintroductions did not provide new genetic material. According to the
considerations of MacArthur and Wilson (1963), small remote islands
foster a high rate of extinction. A species must therefore be restocked fre-
quently. One may contend that a particular beach species is not “reinforced
genetically” by establishment of new arrivals on an island, but rather dis-
appears and then is restocked by a new arrival. However, that does not
explain lack of speciation or variation in pantropical strand species, and
we must therefore assume that new arrivals of a species can establish and
interbreed with individuals of an earlier establishment. One can witness
dispersal events in a plant such as Boerhayvia (see fig. 2.5, top night). Simi-
lar evidence could be cited concerning Pisonia. A relationship between
sooty terns and Pisonia on Rose Atoll, near Samoa, is mentioned by Munro

70

Fic. 2.5. Viscid fruits of Nyctaginaceae adapted for dispersal on feathers. TOP LEFT, TOP AND
CENTER RIGHT, Boerhavia diffusa; fruits 3 mm long. Top LEFT, plant growing on East I.,
French Frigate Shoals, Hawaiian Is. rop ricuT, juvenile sooty tern with fruits stuck to
feathers (several below eye), East I., French Frigate Shoals. cENTER RIGHT, fruits enlarged;
grains of sand adhering to fruits. BOTTOM LEFT, RIGHT, Pisonia umbellifera; fruits 5 cm long.
BOTTOM LEFT, fruits held between fingers to show viscid material on surfaces; BOTTOM RIGHT,

fruit embedded in feathers.

7A

Long-Distance Dispersal: Evidence and Implications

(1960). On Heron Island, Queensland, marine birds become so thickly
covered by Pisonia fruits that they are immobilized and become prey for
rats (R. F. Thorne, personal communication, 1965).

Another plant that seems to owe a wide distribution in the Pacific to
viscidity- of seeds is Lepidium. Viscidity accounts for the occurrence of
Lepidium owaihense on small and remote islands such as the islets of
Pearl and Hermes Reef. On this atoll, I observed it as one of the first
colonists on an islet, “South North Island” that was formed recently by
shifting of sand within the lagoon and had acquired, as of 1966, only four
species of angiosperms. Lepidium owaihense is also present in the Line
Islands (Christophersen, 1927). Mature seeds of this species become gelati-
nous when wetted. This characteristic also occurs in other Hawaiian gen-
era, such as Plantago (see Hyde, 1970) and Euphorbia. Plantago is not as
wide ranging within the Islands as Euphorbia, probably because Plantago
has narrower ecological requirements rather than because of poorer dis-
persibility.

Pittosporum has seeds covered with a viscid, almost oily substance.
The distribution of this genus in the Pacific includes some rather remote
islands and is probably attributable to this seed characteristic. However,
Pittosporum seeds are black or orange and may therefore be attractive to
seed-eating birds. Pittosporum rarely grows close to where marine birds
nest—although a relationship cannot be ruled out, because marine birds
can nest on wet-forest cliffs. In any case, dispersal events would be expected
to be relatively infrequent compared to those of Boerhavia, so the higher
endemism in Pittosporum species in the Pacific is understandable.

Shore- and waterbirds do migrate and do eat plant material and are
otherwise suitable as vectors, but the question of efficacy of marine birds
remains. Do marine birds ever characteristically undertake interisland
travel or do they, as is often averred, invariably return to a home island?
Recent work shows that considerable interisland travel does occur. For
example, on Laysan Island, 69,900 sooty terns were banded. Of these, 611
returns were discovered on Laysan, 7 on Johnston, 4 on Lisianski, and 2 on
Kure. Similarly, of the red-footed boobies banded on French Fngate
Shoals, 225 returns were noted, but 18 birds traveled to Johnston, 8 to
Laysan, 3 to Midway, 2 to Oahu, 2 to Kauai, 1 to Pearl and Hermes Reef,
and 1 to Wake Island (Richard Crossin, unpublished data). Interestingly,

Up:

Long-Distance Dispersal: Evidence and Implications

some of the individuals on Sand Islet of Johnston were later recorded back
on French Frigate Shoals. That in a single year so much interisland move-
ment (of which these figures surely represent a modest sampling) occurs
in marine birds means that these birds are entirely adequate as potential
vectors. The white-tailed tropic bird must travel widely, judging from the
observations offered by Munro (1960). Moreover, it nests in cliff faces, such
as those of northern Kauai, where it could come into contact with a
variety of plants. Because these are windward, north-facing cliffs, wet-forest
species (the Metrosideros forest) range much closer to sea level than one
would expect. Thus it is quite likely that a seabird, the white-tailed tropic
bird, might be a vector for upland plants. The sooty-rumped petrel, a
Hawauan endemic, can nest far inland and at altitudes from 500 m to
above 2000 m, as in Haleakala (Carlquist, 1970). Suggestions that marine
birds can nest inland and that this habit may explain some peculiar dis-
tribution patterns in New Zealand are offered by Falla (1960).

Although marine birds would not be expected to eat much vegetable
matter, they sometimes do. According to Ridley (1930), gulls eat “all
sorts of berries, and especially those of Empetrum nigrum.” Interestingly,
Empetrum is one of the plants that shows the well-known amphitropical
distribution in temperate North and South America (Constance, 1963).
The Hawaiian species of Bidens that show the least loss of dispersibility,
and that therefore suggest mode of arrival best, are those that live on
bluffs where marine birds nest (see chapter 11). Various instances in which
marine birds have been observed to carry seeds and fruits are given by
Ridley (1930). Albatrosses (Diomedea) and petrels may well have brought
disseminules to Macquarie Island and other subantarctic islands, according
to Taylor (1954). Turbott (1951) reports the red-billed gull feeding on
fruits of Meryta sinclairii (Araliaceae); Meryta is definitely a tree of forest
habitats, not a strand plant.

Despite the cited possibilities for marine birds, most of the arrivals
alleged to be bird-dispersed (fig. 2.7) on high oceanic islands are due to
shorebirds and, to a lesser extent, waterbirds. The work of Vlaming and
Proctor (1968) shows us how experimental evidence can aid in dissipating
skepticism about this matter. No doubt, further evidence will reinforce
their findings. Thus one already can view the fruition of Ridley’s opinion:
“There is at first the endeavor in the absence of precise knowledge to dis-

73

Long-Distance Dispersal: Evidence and Implications

regard the bird and to look for a land-connection. With the increase in our
acquaintance with the efficacy of bird-agency in seed distribution there is
abandonment of such a view.”

OCEANIC: DRIFT

Oceanic drift is considered here to include two phenomena—frequent
or repetitive drift, and rare or infrequent drift (rafting, for example). The
distinction was suggested by the fact that some plants, such as Ipomoea
pes-cdprae, are pantropical and show virtually no variation from one island
to another, whereas in the Hawaiian Islands there are littoral or lowland
plants that one would say should belong to the drift flora on general con-
siderations but that are higher in endemism and poor at flotation, for
example, many Fabaceae (see chapter 11). Experimental proof that some
species poor in buoyancy have seeds that nevertheless germinate after
soaking in seawater has been given in a number of instances (e.g., Stephens,
1958, 1966; Whitaker and Carter, 1961; Wilson, 1970; Corn, 1973). Fur-
ther commentary on such species is considered in explanations of the
phenomenon of loss of dispersibility (chapter 11).

In cases of repetitive drift, lack of endemism would be expected.
Plant species that are obviously adapted to flotation are, of course, familiar.
So much literature has been devoted to this topic that at this time a
bibliography seems unnecessary to demonstrate the operativeness of this
dispersal mechanism. Papers by Stephens (1966) and Gunn (1968) deserve
mention, however, for demonstrating the basis of flotation mechanisms.
Drift species illustrated here include Scaevola taccada (fig. 2.6, top left,
right) and Mucuna (fig. 2.6, bottom left, nght). The first plants on Surt-
sey, a volcanic islet that emerged from the sea near Iceland, were typical
oceanic drift species, such as Cakile edentula (Einarsson, 1967).

Less obvious in their dispersal mechanism are certain species of the
littoral flora that range widely in the Pacific but that do not show variabil-
ity or have buoyant fruits and seeds. This is the case in Portulaca lutea
(fig. 2.6, center left) and Sesuvium portulacastrum; vegetative plants prove
to be buoyant, however, and one can suspect that this faculty is responsible
for introduction of these species to remote islands. Vegetative fragments
of these plants can root in the manner of cuttings and are not damaged
by seawater, in my observations. Wide-ranging littoral species that may

74

FIG. 2.6. Capability for seawater dispersal of plants. top LEFT, Scaevola taccada (Goode-
niaceae), fruits 14 mm long on branch; Kalapana, Hawaii. rop RIGHT, fruits (exocarp removed)
of Scaevola taccada (pale) and S. plumieri (dark) float, while seeds of S. coridcea sink. CENTER
LEFT, Portulaca lutea (Portulacaceae); branchlet floating in white enamel tray of water; from

(continued)

?

Long-Distance Dispersal: Evidence and Implications

spread by flotation of vegetative portions include Lycium sandwichense
(Rapa, Easter I., Tonga, Hawaii) and Capparis sandwichiana (Hawaii, Pit-
cairn I., ‘Tuamotus, Rarotonga, Cook Is.). Vegetative portions of Lycium
can float. Guppy (1906) reports that fresh fruits and seeds of Capparis
sandwichiana do not float; my experiments showed, however, that dried
fruits can float indefinitely.

An interesting phenomenon apparently not yet appreciated is that
seeds in the drift flora are not forced to grow only on the spots where the
surf deposits them. On Necker Island, I repeatedly observed seeds of
Mucuna and Aleurites (the former viable, the latter dead) in and around
nests of boobies, far above the surf. Apparently boobies habitually pick up
these objects and deposit them at nesting sites, a behavior for which I
know no explanation. A similar phenomenon was observed among Laysan
albatrosses on Pearl and Hermes Reef and other islands of the Hawaiian
Leeward chain visited by the Ironwood scientific party (of which I was a
member) in 1966. Numerous carcasses of Laysan albatrosses were exam-
ined in which floating objects—plastic articles, pumice, charcoal—remained
within the body cavity. These objects were evidently picked up at sea by
adult albatrosses, then fed to albatross chicks that died as a result (Kenyon
and Kridler, 1969). This finding suggests that albatrosses might act con-
junctively with oceanic drift to abet arrival of floating objects, including
seeds.

Rafting or similar infrequent dispersal via oceanic drift is difficult to
assess. Guppy (1906) found in Hawaii a few instances of ecological shift of
littoral elements into inland sites. I offer the following list of Hawaian
genera in which this shift has happened: Ochrosia, Dodonaea, Mezo-
neurum, Canavalia, Erythrina, Sophora, and Colubrina. Endemic species
are thereby produced. The ecological shift suggests arrival on the Hawaiian
Islands on a very infrequent basis—one can easily hypothesize a single suc-
cessful Hawaiian establishment for the genera listed above. An allied
phenomenon is the notable absence in the native Hawaiian flora of so
many mangrove or littoral genera common in the South Pacific—Barring-
Gardner Pinnacles, Hawaiian Is. BorroM LEFT, RIGHT, Mucuna spp. (Fabaceae), seeds 3—4.5
cm in diam., Pearl and Hermes Reef, Hawaiian Is. BorroM LEFT, assortment of seeds washed

ashore. BOTTOM RIGHT, seedling on beach: germination successful but plant dying at tip, will
not survive.

76

Long-Distance Dispersal: Evidence and Implications

tonia, Rhizophora, Bruguiera, Avicennia, Suriana, Calophyllum, Termi-
nalia, and Batis, to name only a few. Suitable ecological sites are present—at
least half of the genera named have been introduced to the Hawaiian
Islands by man and have naturalized themselves. ‘The conclusion suggested
by both the littoral elements dispersed by single events and now endemic,
and the littoral genera expected but not found in the Hawaiian Islands
is that these islands are strongly isolated from the South Pacific. The drift
flora of the Pacific is primarily a South Pacific flora, which exchanges
easily among South Pacific islands and archipelagos, but apparently does
not frequently cross the transverse equatorial currents. ‘The westward-run-
ning North and South Equatorial Currents, separated by the eastward-
running Equatorial Countercurrent, would form a triple barrier to south-
to-north flotation of the drift flora.

Factors of Ecology and Distance

To place in relief certain factors that are operative in long-distance dis-
persal, I have prepared figure 2.7 and the summary in table 2.1. The num-
bers for the Hawaiian Islands represent a situation in which ecology is
richly diversified, area is relatively great, but remoteness also is very great.
Samoa differs only in degree of remoteness: it is much closer to source areas
of Indo-Malesia. The other two columns list the highest and lowest per-
centage in the Pacific for each dispersal type, and indicate on what island
or archipelago those extremes occur.

One must interpret percentages in a skeptical fashion if the flora is

TABLE 2.1.
Comparisons of Pacific islands with respect to dispersal types
Dispersal Hawaiian

type Islands Samoa Lowest Percentage Highest Percentage
A er ioe 0 (Eq. atolls) 18.0 (San Clemente I.)
BB 12.8 4.0 0 (Eq. atolls) 29.0 (J. Fernandez Is.)
BI 38.9 50.8 0 (Oeno I.) 50.8 (Samoa, Tonga)
BM 12.8 ep) 2.8 (Rarotonga) 21.0 (J. Fernandez Is.)
BV 10.3 4.3 2.1 (Cocos Is.) 5-4 (@enoy 1s)

DF 15 By 17.8 3.0 (J. Fernandez Is.) 63.6 (Eq. atolls)

DR 8.5 a5 0 (Cocos Is.) 8.5 (Hawaii)

77

Long-Distance Dispersal: Evidence and Implications

based on a small number of original immigrants. Still correlations are re-
markably clear. In the case of air flotation, a low percentage is expected
on atolls simply because atoll plants are chiefly drift plants. Oceanic drift

LAYSAN I. N=26 SAN CLEMENTE I. N=233
4.0.0
Bier, 5 emp HAWAIIAN IS. N=256
11.4 8.5
14.3 12.8
10.3
12.8

38.9 REVILLAGIGEDO IS. N=82

37 8.5

24.4
EQUATORIAL IS. N=33

6.1 96
eee, 30.5
9.1
on 6.1 Cocos Il. N=47
SAMOA N=3lil MARQUESAS IS. N=132

3.1 36

21.3
575
SOCIETY IS. N=230 43.3 .
508 GALAPAGOS IS. N=308
OENO I. N=I3 isis
ae 21.8 22.8

TONGA N=296

3.0 68.1

EASTER I. N=28
DESVENTURADAS IS. N=20

36.0 250
mr =a 20.0
5.0
25.0
50.8 25.0 ee
21.4 Be 10.0
KEY 15.0
Drift, Rare (Rafting) Air flotation JUAN FERNANDEZ IS. N=100
, : 2.0
Drift, Frequent Birds: Barbs, bristles, etc. 30geee 29.0
3.0
Birds: Viscid fruits,seeds Birds: Internal
Birds: 43.1
Mud on feet Ne 40.0

>

Fic. 2.7. Diagrammatic representation of modes of dispersal responsible for populating the
islands of the eastern Pacific and Polynesia with seed plants. Diagrams are arranged geograph-
ically, although distances are distorted. For each flora, the number of immigrants presumed
necessary to account for the contemporary native species has been calculated (N). Of these,
the percentage thought to have arrived according to each of the seven methods shown in
the key has been calculated and presented graphically. Only native species are included. The
figures for the Hawanian Islands include the Leeward chain; figures for Laysan alone are also
presented.

78

Long-Distance Dispersal: Evidence and Implications

is ecologically linked in plant species to preference for the beach habitat—
and atolls are, of course, nothing but mid-ocean beaches. Aerial flotation
would be a poor method for populating atolls and remote islands with
flowering plants because it operates so poorly across long oceanic distances.
Thus we can anticipate the very low percentages on the Hawauan Islands
and on Rapa. The high percentage on San Clemente reflects its nearness
to a continent. Relative nearness to a source area aids in air flotation of
seeds to Samoa (the high figure is based chiefly on orchids). The figure
for the Revillagigedos (8.5 percent) is moderately high, indicating a thresh-
old value for dispersal by air flotation: islands farther from the mainland
or other source areas than the Revillagigedos may be expected to have
minimal percentages for air flotation—perhaps 2 percent.

Bristly or barbed fruits seem best suited for dispersal to islands where
birds are by far the most important vectors, and the high figure for the
Juan Fernandez Islands is understandable. High figures also occur on
islands near sources—the Revillagigedos, the Desventuradas, and the
Galapagos, for example. These are dry islands, however, and the dispersal
of bristly or barbed fruits may represent successful dispersal of dry-forest
elements, as opposed to the fleshy fruit types found in wet forests. The
efficacy of transport of barbed or bristly fruits lessens with distance,
although not nearly as sharply as that of air flotation. The low figure for
atolls reflects this trend, but perhaps more importantly reflects the over-
whelming importance of oceanic drift for atolls and the necessarily lesser
significance of other means of transport.

The ecological correlation of oceanic drift with atoll habitat certainly
explains the low figure for internal transport of fruits and seeds in birds to
atolls. Fleshy-fruited species are most characteristic of wet or moderately
dry forest, so that the high percentage for Samoa is almost predictable. The
high figure for Tonga may be unexpected in view of Tonga’s less favorable
ecology (although there are forested uplands on some islands, such as Eua),
but nearness to source area undoubtedly favors this mode of transport for
Tonga, which is close to the rich flora of Fiji. That the Hawaiian Islands
show only relatively slight diminution in internal transport by birds means
that this mode of transport is entirely effective over long distances, pro-
vided the habitat for fleshy-fruited species—the wet forest, chiefly—is
present on the recipient island, as is true in Hawaii. In this connection,

79

Long-Distance Dispersal: Evidence and Implications

one cannot help noticing that the Galapagos Islands are low in arrivals by
means of internal transport in birds and also have no true rain forest. A
rather high figure might not be expected for San Clemente Island, but
this dry island is so near to mainland areas that, according to my interpre-
tations, small-seeded plants with dry capsules (e.g., Brodiaea) may have
arrived there by means of seed-eating birds. There are also, however, fleshy-
fruited species in the San Clemente flora: among them, Rhus, Hetero-
meles, and Ribes. There are even a few fleshy-fruited species on atolls—
Solanum nelsoni and Santalum ellipticum on islands of the Leeward
Hawaiian chain, for example.

Not much discrimination among islands can be claimed for the figures
on arrival in mud on birds’ feet; the figures are all relatively low and thus
deviations are not strongly significant. The low figure for Rarotonga may
reflect the limited uplands of that island, with fewer “muddy” habitats.
The Juan Fernandez Islands, on the contrary, have a proportionately high
area of wet, scrubby uplands suited for plants of muddy regions. ‘The low
figure for Samoa could merely indicate that arrival of fleshy-fruited trees
is so successful that other forms of travel are inevitably of lesser signifi-
cance. The high percentage on San Clemente may reflect closeness to a
continental area, occasional visits by passerine birds, and a nonforest
ecology (which does have a moist winter, however). The high percentage
on Easter Island (21.4) would be unexpected unless one had noted a dis-
tinctive habitat on that island: the muddy margins of a crater lake.

Viscid-fruited and viscid-seeded species are not a homogeneous group,
ecologically, and discerning significance is less easy. Some are beach plants,
like Boerhavia and Pisonia, so that the high figure on Oeno, a raised atoll,
is understandable, and low percentages on higher wet islands are to be
expected. Such forest trees as Pittosporum have this method of dispersal,
as do perhaps Korthalsella, Vaccinium, and the baccate Hawaiian lobe-
lioids. ‘This fact may explain the relatively high figure for the Hawaiian
Islands; where distance is a problem, sticky seeds—presumably capable of
almost indefinite transoceanic travel on birds—would be more successful
immigrants by default. This latter aspect, as well as wet-forest ecology,
may explain low figures for Samoa and Cocos, where distance is not a
great problem. In angiosperms generally, dispersal by means of viscid seeds
is fairly infrequent, and thus we would not expect it to bulk large at all in

80

Long-Distance Dispersal: Evidence and Implications

insular floras; it probably is greater in insular than in mainland floras be-
cause of its efficiency across great distances.

The highest percentage of arrival by means of repetitive drift is found,
predictably, on the atolls, most of which are “all beach.” There is very
little beach area on the Juan Fernandez Islands, so the low percentage
there is no surprise. A similar situation exists on Rapa. Extensiveness of
the available beach habitat on particular Pacific islands appears to be
almost perfectly correlated with their percentages of dnft plants. The high
figure for the Galapagos may be an example; also, the xeric lowlands of
the Galapagos, beach or otherwise, might be said to favor strand plants,
which are basically halophytic in ecology.

Infrequent or single-event drift (perhaps rafting in some cases) is
high in the Hawaiian Islands because these islands are so well cut off by
the transverse equatorial currents, as mentioned earlier, from entrance of
South Pacific elements. The data of Stephens (1966) are pertinent here,
because the cottons evidently do not face this problem. Dispersal of the
cottons is from North and South American coasts, via westward-running
currents: South Pacific islands are populated via the South Equatorial Cur-
rent, North Pacific islands via the North Equatorial Current. Nonetheless,
Hawaii is not truly in the range of even the North Equatorial Current, and
the endemism of its Gossypium is understandable.

The consequences of rare events of oceanic drift in permitting or
promoting ecological shift and endemism in the Hawaiian Islands and a
few other islands are discussed in chapter 11. One must also note that not
all plants of the drift flora are suited to areas where they arive. Mucuna
and Entada, although frequently found on Hawaiian beaches, do not form
mature plants if they germinate (fig. 2.6, bottom left, nght). These plants
do occur in Hawaii near the ocean, but in shady forests where vines may
overhang the water and release seeds into the sea, or into streams.

In short, the overriding role that ecology of the recipient island plays
in dispersal cannot be strongly enough stressed. This aspect of my (1967)
paper has been confirmed by the study of Whitehead and Jones (1969) on
Kapingamarangi. Odd disjunct distributions may be the result of chance
in some cases, but far more frequently they result from disjunctions in
suitable ecological conditions. We should first suspect this latter explana-
tion, since long-distance dispersal is not, apparently, at all difficult for

$1

Long-Distance Dispersal: Evidence and Implications

many groups of plants. Ecological specificity may, in turn, be based on
geology and the geological history of an area; examples are cited, for ex-
ample, by Burrows (1964, 1965) in the New Zealand flora.

Long-Distance Dispersal to Regions
Other Than Polynesia

MASCARENE ISLANDS

If, as I have stated earlier (1966a), disharmony is an indication that
certain plants and animals are positively adapted to long-distance dis-
persal, these species ought to be more conspicuous on oceanic islands. A
phenomenon that might not have been predicted, however, is the tend-
ency of the Mascarene flora and fauna to be somewhat “Polynesian” in
character. Computations demonstrate this tendency. Despite the fact that
the Seychelles, Réunion, Mauritius, and Rodriguez are close to Mada-
gascar and Africa, their floras contain many of the genera characteristic
of Polynesia. The islands named are included in Baker’s (1877) flora. Of
the 408 genera of indigenous flowering plants in this flora, 158 occur also
in Polynesia (Samoa to Easter I. on the east, Hawaii on the north, Rapa
on the south). When one compares smaller floristic units, this similarity
becomes even more striking. If one regards 210-216 genera of flowering
plants as ancestors of the Hawaiian angiosperm flora, 115, or 53-56 per-
cent, of these also occur in the Mascarene flora. To be sure, a number of
these genera are elements of the drift flora that might be expected to be
widespread on islands. But many genera common to the two areas are not
drift plants at all; for example, Dianella, Dracaena, Astelia, Pipturus, Pep-
eromid, Diospyros (including Maba), Myrsine, Psychotria, Canthium, Pit-
tosporum, Elaeocarpus, and Zanthoxylum (Fagara). One would certainly
not expect to find American and boreal elements on the Mascarene Islands,
owing to the enormous distance involved. Likewise, one would not expect
African elements in the Hawaiian flora. If one deducts these elements,
respectively, the degree of similarity between the Mascarenes and Polynesia
is really quite high, and is an expression of the portion of the Indo-Male-
sian flora that is adapted for long-distance dispersal. Interestingly, there is
a marked zoological counterpart to the floristic phenomenon. Gressitt
(1956) shows a Pacific-Mascarene distribution for several insect groups,

82

Long-Distance Dispersal: Evidence and Implications

such as Chrysomelidae-Hispinae and Cerambycidae. Mascarene-Polynesian
affinity can even be shown in land shells. Elasmias (Tornatellinidae) is an
insular Pacific genus that has an extension on the Mascarene islands
(Cooke and Kondo, 1960). The family Tornatellinidae seems strongly
adapted for long-distance dispersal, judging from the distribution maps
these authors offer.

INTERCONTINENTAL DISPERSAL

Intercontinental dispersal is potentially very important; the degree
and nature of possible intercontinental long-distance dispersal bears a
direct relationship to the evidence on which rests the probability of con-
tinental drift, or the geological epochs at which it occurred. Also in ques-
tion are theories of dispersal between mountains on the same or different
continents. One area that has been of critical significance is the disjunction
between two pockets of mediterranean climate: western North America
and central Chile (and adjacent regions). This area has been the subject
of a review and individual papers by Chambers (1963), Constance (1963),
Heckard (1963), Ornduff (1963) and Raven (1963b). Difficulties in inter-
preting the disjunction as the product of dispersal by bird agencies have
been hinted by Constance (1963) and stated rather more strongly by
Cruden (1966). Although Cruden claims not to be motivated by disbelief
in birds as potential vectors, he evidently wishes us to consider an alter-
native hypothesis—a series of mountaintops as stepping-stones across the
tropics, bridging the two temperate regions. Although the amphitropical
disjunction needs a more thorough examination to determine whether
seeds or fruits of those particular genera and species are adapted to long-
distance dispersal (as I believe they are), the following comments can be
offered to support the concept of birds as dispersal agents. The background
of studies in Polynesia and in the Mascarenes is an essential ingredient,
for distant oceanic islands provide compelling evidence of the realities of
long-distance dispersal.

1. The distances of the North America-South America disjunction
are of the same level of magnitude as those of the America-Polynesia dis-
junction. If 20 percent (at least 52 immigrants) of the establishments ar-
rived in Hawaii from American sources, mostly via birds, then the amphi-
tropical American disjunction does not seem highly improbable on the

83

Long-Distance Dispersal: Evidence and Implications

basis of long-distance dispersal. Moreover, the North America—Hawaii or
North America—Polynesia disjunction is the result of one-way dispersal,
whereas both western North America and Chile have mutually been able
to serve as potential source areas for the amphitropical disjuncts, doubling
the chance of successful long-distance dispersal.

2. The list of plants showing amphitropical disjunction is very similar
to the list of migrants from American or boreal regions to Hawaii: Arte-
misia, Aster, Caesalpinia, Carex, Fragaria, Geranium, Lepidium, Luzula,
Poa, Ranunculus, Rubus, Sanicula, Scirpus, Stipa, Vaccinium, and Viola
(not to mention “close” groups, such as Madia in the Americas, and
Dubautia in Hawai). Moreover, fruit morphology is very similar to that of
the America-Hawaii disjuncts. The list of genera showing both disjunc-
tions indicates clearly an exceptional ability for long-distance dispersal. A
few genera in either disjunct pattern could be said to have “no obvious
dispersal mechanism,” but this would be a mistake. Anything that has
been able to cross 2000 or more miles of ocean to Hawaii must have an
effective dispersal mechanism (but not necessarily, of course, so efficient
that more than a single successful establishment would occur). We must
remember that birds may eat, and carry internally, fruits other than those
which are fleshy and brilliantly colored. For example, ducks eat and trans-
port a wide variety of fruits and seeds that are inconspicuous and often
greenish or brownish and woody in texture (Ridley, 1930). There is, of
course, the possibility that in a few cases the lack of a dispersal mechanism
might indicate a loss of dispersibility (Carlquist, 1966b, 1966c; see chapter
Is)

3. No “mountaintop stepping-stones” have been proposed for the
North America—Hawauan Islands disjunction, nor are they necessary to
explain the dispersal that has occurred. The proposal of intervening
mountaintops to serve as way stations in the amphitropical disjunction also
seems unnecessary. At best, it merely cuts long-distance dispersal into a
series of shorter segments, and thus demands successive events of long-
distance dispersal while at the same time making the process less probable,
for reasons given in comment 4.

4. Mountaintops in the tropics are not equivalent to temperate cli-
mates, montane or otherwise (see chapter 14). They differ particularly
with regard to seasonal changes in temperature and day-length, but also

84

Long-Distance Dispersal: Evidence and Implications

in rainfall patterns, soils, etc. To accept a mountaintop hypothesis, we
must expect temperate-zone plants to make numerous adaptations to
tropical mountains. This is not unreasonable, for some have done that, but
to explain the amphitropical disjunction by the mountaintop hypothesis
we must invoke something more preposterous: after the plants have
adapted to tropical mountains, we must imagine that they can readapt
to temperate-zone conditions in the opposite hemisphere—and these adap-
tations and retrograde changes are numerous. They must then disappear in
the tropics yet appear as closely related species (congeners) only in the
two mediterranean pockets. If the roster of amphitropical disjuncts did
adapt to the tropical mountaintops, why did they disappear so completely
there? The data from Polynesia show just the reverse of the mountaintop
hypothesis. They show that intervening way stations (e.g., Samoa as a
point between Fiji and Tahiti) have retained more immigrants than the
end points. Moreover, relatively little ecological change has occurred in
most angiosperm stocks on islands (the Hawaiian Islands are exceptional
in this regard). The various authors of the papers cited on the amphi-
tropical herbaceous disjuncts emphasize the genetic closeness of Chilean
to California plants, or vice versa, evidencing remarkably little evolutionary
change since the time of disjunction. However, if we accept mountaintops
as way stations, at the very least we must imagine considerable physiologi-
cal change, first progressive then retrograde, no evidence of which is now
visible. The morphology of the amphitropical disjuncts hints at no period
of residence on tropical mountaintops. The amphitropical pattern, accord-
ing to the data of the authors cited, suggests great recentness compared
with Polynesian patterns, which themselves must be fairly recent.

5. The islands of Polynesia have received a relatively large quantity
of waif disseminules; what has been severely limiting seems not lack of
transport, but ecological conditions and sufficient land area suitable for
establishment and then survival over a number of generations. The pat-
terns of figure 2.7 that were discussed earlier show that Pacific islands have
“selected out” from a flow of immigrants those that are exactly suited
ecologically to recipient islands. The disjunction between mediterranean
pockets in North and South America occurs precisely because ecological
conditions are so nearly identical in the two regions, yet are unlike in
intervening areas. If plants had to adapt to equatorial alpine peaks, some

85

Long-Distance Dispersal: Evidence and Implications

relicts of this supposedly recent migration ought to be there still. Instead,
the floras of the equatorial Andean highlands are quite different (chapter
14).

6. Cruden (1966) has assessed his evidence concerning bird groups
in a very conservative fashion. There certainly are, as his paper indicates,
some species that could serve as vectors, by virtue of their migratory pat-
terns. The point is not how many species do, in fact, have this route, but
how many individuals travel it. If one migratory bird species followed the
Cordilleran route annually in sufficient numbers, we would have the vec-
tors needed. Another question, evidently not investigated, is whether the
birds known to follow this migratory route come into contact with seeds
and fruits of any of the species of plants concerned. Unless this question
can be answered negatively, we cannot dismiss the probability that birds
utilizing the Cordilleran flyway have served as vectors.

The tropical South America—western Africa (or even eastern Africa)
disjuncts among flowering plants are often cited as corroborative evidence
—together with geology (although rarely by one who knows both fields
well)—as evidence of continental drift (e.g., Corro, 1964; Aubréville, 1969).
Do the angiosperm elements in common between South America and
Africa suggest that these continents did drift apart? Thorne (1972) has
reviewed these angiosperm patterns in a comprehensive and detailed
fashion that supersedes all prior essays on this topic. Thorne has demon-
strated that:

1. We can disregard the 110 species common to tropical West Africa
and tropical America; they are almost certainly the result of human intro-
ductions or recent long-distance dispersal by good vectors, because (if not
weeds) they grow in aquatic, maritime, riparian, or other habitats likely to
be visited by migratory birds on a regular basis. One species with a wide
range on both continents is Rhipsalis baccifera, the only cactus known to
occur in a supposedly native state in the Old World. This is not con-
ceivably a relict distribution, because Rhipsalis is a highly specialized cac-
tus, well-suited to long-distance dispersal (it may even have been intro-
duced to Africa within historical time, possibly by man), and because it
has a wide distribution in both Africa and the New World, without any
hint of subspeciation between the continents. That only one species of

86

Long-Distance Dispersal: Evidence and Implications

this family with perhaps 2000 species in 100 genera has crossed the Atlan-
tic 1s typical of many patterns we find at the generic level.

2. There are 12 angiosperm families restricted primarily to Africa
plus South America. This number is suspiciously small, yet these families
must be the ones to provide the bulk of evidence if continental drift had
acted to produce floristic disjunctions between the two continents. Eight
of the families (Bromeliaceae, Cactaceae, Caricaceae, Humiriaceae, Loa-
saceae, Mayacaceae, Rapateaceae, and Vochysiaceae) are essentially
American: they include at least 3500 species, of which no more than 10
occur in Africa. If continental drift separated portions of these families,
shouldn’t the development on both continents be more nearly balanced?
Evidently these patterns are not the result of continental drift; the barrier
to migration across the Atlantic is great and explains the patterns. One of
the eight primarily American families (Cactaceae) has been discussed
above. All 12 families seem well-suited to long-distance dispersal, so the -
finding that a few species have crossed the Atlantic is not surprising. A
typical case is Rapateaceae: the sole African representative, Maschaloce-
phalus dinklagei (the genus is monotypic), a plant of Liberian marshes, is
(for this family) a highly specialized plant very close to the genera geo-
graphically nearest it in South America—Windsorina and Potarophytum
(Maguire et al., 1958; Carlquist, 1961, 1966d). If continental drift were
involved, one would expect not only better balance on both sides of the
Atlantic, but also more primitive portions of these families in both con-
tinents, not just one specialized genus. Rapateaceae are usually marsh
plants, so dispersal mechanisms are easy to envisage. Other families also
have clear means of dispersal: Humiriaceae and Vochysiaceae have
buoyant fruits, for example; Mayacaceae are aquatic plants likely to be
distributed by waterbirds.

The four families that are about equally divided between South
America and Africa (Hydnoraceae, Canellaceae, Turneraceae, and Vello-
ziaceae) are not difficult to explain. There is one genus of Hydnoraceae
(Hydnora) in Africa, and one in South America (Prosopanche). Small-
seeded parasitic plants often have very wide distributions with large dis-
junctions—Pilostyles in the Rafflesiaceae, for example, has much greater
disjunctions than does Hydnoraceae.

87

Long-Distance Dispersal: Evidence and Implications

3. Not only is there strong imbalance in representation between
African and American portions of families and genera on both continents,
and not only are the elements that the continents have in common notably
good at long-distance dispersal (the mangrove genera Lacuncularia and
Combretum, for example), but they also are all representatives of what are
commonly termed “‘specialized” families (i.e., families with numerous
characteristics considered specialized for angiosperms). “Relict” genera
and families certainly do not link Africa and South America. Among
conifers, Araucariaceae links all continents equally, for it is present or
known in fossil form on all of them. The remaining conifers, especially
genera with primitive characteristics, are notably absent in Africa. Among
angiosperms, Australasia-South America patterns are seen in the “primi-
tive” families Winteraceae, Chloranthaceae, and Monimiaceae. Those
three families can be envisioned as achieving their Australasian—South
American distribution by long-distance dispersal: Drimys (fig. 2.4, bottom
right) is also suited for this. The “annonalean” or “ranalian” families pres-
ent in South America have, it seems, not come via Africa, and Africa is
notable for lack of these families.

4. Although there is definite affinity between South America and
one or more regions of Africa (including Madagascar) in terms of angio-
sperm genera (111 are limited to this distribution pattern), stronger affini-
ties link these continents with other regions than each other. As Thorne
(1972) states, “African floristic relationships thus are also much stronger
with Asia than with America,” and he notes that the genera in common
between Africa and South America are far outweighed by those that have
failed to cross the Atlantic.

5. On the basis of size and ecology, one would expect a much larger
affinity between South America and Africa on the basis of long-distance
dispersal alone. Both Africa and South America have broad expanses of
similar ecological zones, and both have enormous land areas that can
serve as source and target (recipient) areas for disseminules. The presence
of such a small degree of affinity between Africa and South America de-
spite their similarities may be due to the fact that both of these continents
have harmonic flora, and invasion of one by disseminules from the other
is not as easy as invasion of a relatively “open” oceanic island.

If the floras of the Hawaiian and other Pacific islands can be explained

88

Long-Distance Dispersal: Evidence and Implications

by long-distance dispersal, intercontinental disjunctions surely can be. It
is somewhat mysterious that those who study intercontinental disjunc-
tions do not appear to compare their questions and data to situations on
insular areas. The Hawauan Islands are much more remote, have a smaller
target area for disseminules, and have existed for a much shorter period
of time than have the North America-South America or the South
America—Africa land areas on which disjunctions are cited. As a footnote,
instances of long-distance dispersal within continents often escape notice
but are equally significant (Wilson, 1970). Gillett (1965) has cited a
single species (Trigonella laciniata) that has made a dramatic leap within
Africa. The African equatorial peaks and the Ethiopian highlands are
pertinent. Not only have they acquired elements from the temperate
floras of Europe and South Africa by long-distance dispersal, but also dis-
persal among these peaks (which are by no means close) has occurred
within relatively recent time (see chapter 14).

The concept of continental drift is supported by an increasing num-
ber of workers. However, most estimates of its time scale place its occur-
rence at rather ancient dates compared to the development of modern
angiosperm floristic assemblages. I wish to underline strongly that those
who compile evidence pointing toward continental drift are doing them-
selves a disservice when they cite angiosperm floras as evidence. If inter-
continental affinities can be explained by long-distance dispersal, the fabric
of evidence for continental drift is only weakened by using angiosperms as
examples. If continental drift has occurred, the reliable evidence for it
will surely come primarily from geology and geophysics, not from biology.

Concluding Remarks

Although one could become overenthusiastic about the occurrence of
long-distance dispersal, this phenomenon obviously plays a supreme role
for oceanic islands, and at least a significant role in the case of “conti-
nental” islands and continental masses. Most biologists appear to have
little difficulty in accepting concepts of long-distance dispersal, but a few
(mostly botanists) have been reluctant. Thorne (1972) notes with delight-
ful irony how some botanists find the transport of a tiny seed on a bird—

89

Long-Distance Dispersal: Evidence and Implications

a seemingly casual event—preposterous, yet plunge headlong into hy-
potheses that require incredibly drastic geological upheavals over excep-
tionally short periods of time. Oceanic islands are, to me, archives of the
fact of long-distance dispersal and the way in which it works and has
worked in the past. Matched ecology of donor and recipient areas appears
to be a requirement of prime importance, far more than efficiency of
vectors. The vectors are omnipresent. Some, to be sure, do not operate
over indefinite distances, and some plant and animal groups are poorly
adapted for long-distance dispersal. Still, when assessing floristic and
faunistic composition of islands and other areas, should we not begin by
considering the positive hypothesis that dispersal over distances can occur,
rather than beginning with a null hypothesis? Biologists seem occasionally
piqued that, in dealing with long-distance dispersal, we are dealing with
a phenomenon that involves chance and is not subject to precise and
orderly principles in the same way that embryo development of vertebrates
is. That should not detract, however, from the reality of the phenomenon.
We must also remember that all plant and animal groups are capable of
evolving with respect to dispersal mechanisms. T’oo often, I note the
tendency to imply, if only by default, that a given plant or animal group
can change in innumerable ways but its dispersal mechanism and dis-
persibility remain precisely the same. The data of chapters 1] and 12 are
offered as one corrective to this implication, and should be considered, in
fact, as a further source of evidence that long-distance dispersal is
operative.

References

Aubréville, A. 1969. Essais sur la distribution et l’histoire des angiospermes tropi-
cales dans le monde. Adansonia 9:189-247.

Baker, J. C. 1877. Flora of Mauritius and the Seychelles. London, L. Reeve & Co.

Balgooy, M. M. J. van. 1960. Preliminary plant-geographical analysis of the Pacific.
Blumea 10:385—430.

Bannerman, D. A., and W. M. Bannerman. 1965. A history of the birds of Maderra,
the Desertas, and the Porto Santo Islands. Edinburgh and London, Oliver &
Boyd.

Bold, H. C., M. R. Brown, and D. A. Larson. 1964. Airborne algae: Their abun-
dance and heterogeneity. Science 143:583-85.

90

Long-Distance Dispersal: Evidence and Implications

Brown, F. B. H. 1931. Flora of southeastern Polynesia. I. Monocotyledons. Bishop
Mus. Bull. 84:1-194.

Brown, F. B. H. 1935. Flora of southeastern Polynesia. III. Dicotyledons. Bishop
Mus. Bull. 130:1-386.

Brown, R. M., Jr. 1971. Studies of Hawaiian fresh water and soil algae. I. The
atmospheric dispersal of algae and fern spores across the island of Oahu, Hawaii.
In Contributions to phycology, ed. B. C. Parker and R. M. Brown, Jr. Houston,
Tex., Allen Press.

Brown, W. L., Jr., and E. O. Wilson. 1965. Character displacement. Syst. Zool.
5:49-64.

Burrows, C. J. 1964. Some discontinuous distributions of plants within New Zea-
land and their ecological significance. Part I. Tuatara 12:125-33.

Burrows, C. J. 1965. Some discontinuous distributions of plants within New Zea-
land and their ecological significance. Part II. Tuatara 13:9-29.

Carlquist, S. 1961. Pollen morphology of Rapateaceae. Aliso 5:39-66.

Carlquist, S. 1962. Trematolobelia: Seed dispersal; anatomy of fruits and seeds.
Pacific Sci. 16:126—34.

Carlquist, S. 1966a. The biota of long-distance dispersal. I. Principles of dispersal
and evolution. Quart. Rev. Biol. 41:247—70.

Carlquist, S. 1966b. The biota of long-distance dispersal. IF. Loss of dispersibility in
Pacific Compositae. Evolution 20:30-48.

Carlquist, S. 1966c. The biota of long-distance dispersal. III. Loss of dispersibility in
the Hawaiian flora. Brittonia 18:310-35.
Carlquist, S. 1966d. Anatomy of Rapateaceae

16:17-38.

Carlquist, S. 1967. The biota of long-distance dispersal. V. Plant dispersal to Pacific
islands. Bull. Torrey Bot. Club 94:129-62. |

Carlquist, S. 1970. Hawaii, a natural history. New York, Natural History Press.

Chambers, K. L. 1963. Amphitropical species pairs in Microseris and Agoseris (Com-
positae: Cichorieae). Quart. Rev. Biol. 38:124-40.

Chang, T. 1967. A preliminary survey of airborne algae in the Taipei atmosphere.
Taiwania 13:19.

Cheeseman, T. F. 1902. The flora of Rarotonga, the chief island of the Cook group.
Trans. Linnaean Soc., ser. 2, Bot., 6:261—313.

Christophersen, E. 1927. Vegetation of Pacific equatorial islands. Bishop Mus. Bull.
44:1-79.

Christophersen, E. 1935. Flowering plants of Samoa. Bishop Mus. Bull. 128:1—221.

Christophersen, E. 1938. Flowering plants of Samoa. II. Bishop Mus. Bull. 154:1-77.

Shstophersen, -E., and Ha L.- Gaum., 1931,. Vascular ‘plants of “the. Leeward
Islands, Hawaii. Bishop Mus. Bull. 81:1-41.

Constance, L. 1963. Amphitropical relationships in the herbaceous flora of the
Pacific coast of North and South America: A symposium. Introduction and _his-
torical review. Quart. Rey. Biol. 38:109-16.

Cooke, C. M., Jr., and Y. Kondo. 1960. Revision of Tornatellinidae and Achati-
nellidae (Gastropoda, Pulmonata). Bishop Mus. Bull. 221:1-303.

91

roots and stems. Phytomorphology

Long-Distance Dispersal: Evidence and Implications

Corn, C. 1973. Seed dispersal methods in Hawaiian Metrosideros. American Insti-
tute of Biological Sciences, 25th anniversary volume, in press.

Corro, G. del. 1964. La gondwania: el antigue contienente australe. Mus. Argentina
Cieng. Natur. “Bernardino Rivadavia’ Publ. 12:1-90.

Cottam, C., C. Scooter, and C. S. Williams. 1942. Flight and running speeds of
birds. Wilson Bull. 54:121-31.

Cruden, R. W. 1966. Birds as agents of long-distance dispersal for disjunct plant
groups of the temperate western hemisphere. Evolution 20:517-32.

Deignan, H. G. 1963. Birds in the tropical Pacific. In Pacific basin biogeography, ed.
J. L. Gressitt, pp. 263-69. Honolulu, Bishop Museum Press.

Drake del Castillo, E. 1893. Flore de la Polynésie francaise. Paris, G. Masson.

Einarsson, E.. 1967. The colonization of Surtsey, the new volcanic island, by vascular
plants. Aquilo, ser. Bot. 6:172-82.

Falla, R. A. 1960. Oceanic birds as dispersal agents. Proc. Roy. Soc. London, ser. B,
152:655-59.

Fleming, C. A. 1962. History of the New Zealand land bird fauna. Notornis
9:270-74.

Fosberg, F. R. 1948. Derivation of the flora of the Hawaiian Islands. In Insects of
Hawaii, vol. 1: Introduction, E. C. Zimmerman, pp. 107-19. Honolulu, University
of Hawaii Press.

Fosberg, F. R. 1963. Plant dispersal in the Pacific. In Pacific basin biogeography,
ed. J. L. Gressitt, pp. 273-82. Honolulu, University of Hawaii Press.

Fulton, J. D. 1966a. Microorganisms of the upper atmosphere. III. Relationship
between altitude and micropopulation. Appl. Microbiol. 14:237—40.

Fulton, J. D. 1966b. Microorganisms of the upper atmosphere. IV. Microorganisms
of a land air mass as it traverses an ocean. Appl. Microbiol. 14:241—44.

Fulton, J. D. 1966c. Microorganisms of the upper atmosphere. V. Relationship be-
tween frontal activity and the micropopulation at altitude. Appl. Microbiol.
14:245-50.

Fulton, J. D., and R. B. Mitchell. 1966. Microorganisms of the upper atmosphere.
II. Microorganisms in two types of air masses at 690 meters over a city. Appl.
Microbiol. 14:232-36.

Gillett, J. B. 1965. Trigonella laciniata L. in Zambia—a probable instance of long-
distance transport of seeds by birds. Kew Bull. 19:387-88.

Gregory, P. H., and J. L. Monteith, eds. 1967. Airborne microbes. Cambridge,
Cambridge University Press.

Gressitt, J. L. 1956. Some distribution patterns of Pacific island faunae. Syst. Zool.
5:11-47.

Gunn, C. R. 1968. Stranded seeds and fruits from the southeastern shore of
Florida. Gard. Jour. (N.Y. Bot. Gard.) Mar./Apr. 1968:43—-54.

Guppy, H. B. 1897. On the postponement of the germination of seeds of aquatic
plants. Proc. Roy. Soc. Phys. Sci. 13:344—59.

Guppy, H. B. 1906. Observations of a naturalist in the Pacific between 1896 and
1899, vol. II: Plant dispersal. London, Macmillan & Co.

Heckard, L. R. 1963. The Hydrophyllaceae. Quart. Rey. Biol. 38:117—23.

92

Long-Distance Dispersal: Evidence and Implications

Heise, H. A., and E. R. Heise. 1948. The distribution of ragweed pollen and Alter-
naria spores in the upper atmosphere. Jour. Allergy 19:403-427.

Hillebrand, W. 1888. Flora of the Hawaiian Islands. London, Williams & Norgate.

Hirst, J. M., O. J. Stedman, and W. H. Hogg. 1967. Long-distance spore transport:
Methods of measurement, vertical spore profiles and detection of immigrant
spores. Jour. Gen. Microbiol. 48:329-55.
Hirst, J. M., O. J. Stedman, and G. W. Hurst. 1967. Long-distance spore transport:
Vertical section of spore clouds over the sea. Jour. Gen. Microbiol. 48:357-77.
Hurst, G. W. 1969. Meteorological aspects of insect migrations. Endeavour 27
(104):77-81.

Hyde, B. B. 1970. Mucilage-producing cells in the seed coat of Plantago ovata:
Developmental fine structure. Amer. Jour. Bot. 57:1197-1206.

Johnston, I. M. 1931. The flora of the Revillagigedo Islands. Proc. Calif. Acad. Sci.
4th ser., 20:8-104.

Kenyon, K. W., and E. Kridler. 1969. Laysan albatrosses swallow indigestible mat-
ter. Auk 86:339-43.

Klein, D. R. 1968. The introduction, increase and crash of reindeer on St. Matthew
Island. Jour. Wildlife Manage. 32:350-67.

Kohlmeyer, J. 1966. Ecological observations on arenicolous marine fungi. Z. Allg.
Mikrobiol. 6:95-106.

Kramer, C. L., and S. M. Pady. 1968. Viability of airborne spores. Mycologia
60:448—49.

Lack, D. 1969. Subspecies and sympatry in Darwin’s finches. Evolution 23:252-63.

Lack, D. 1971. Island birds. In Adaptive aspects of insular evolution, ed. W. L.
Stern, pp. 29-31. Pullman, Wash., Washington State University Press.

Lamoureux, C. H. 1963. The flora and vegetation of Laysan Island. Atoll Res. Bull.
97:1-14.

Low, J. 1937. Germination tests of some aquatic plants important as duck foods.
B.S. Thesis, Utah State University, Logan, Utah. 27 pp.

MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoo-
geography. Evolution 17:373—87.

McAtee, W. L. 1918. Food habits of the mallard ducks of the United States. U.S.
Dept. Agr. Publ. 720, pp. 1-25.

McAtee, W. L. 1947. Distribution of seeds by birds. Amer. Midland Naturalist
38:214-23.

McDonald, J. E. 1962. Collection and washout of airborne pollens and spores by
raindrops. Science 135:435—36.

Maguire, B., and contributors. 1958. The botany of the Guayana Highland. Part
Ill. Mem. N.Y. Bot. Gard. 10:1-156.

Malone, C. R. 1965a. Dispersal of plankton: Rate of food passage in mallard ducks.
Jour. Wildlife Manage. 29:529-33.

Malone, C. R. 1965b. Killdeer (Charadrius vociferus Linnaeus) as means of dis-
persal for aquatic gastropods. Ecology 46:551—52.

Malone, C. R., and V. W. Proctor. 1965. Dispersal of Marsilea mucronata by water
birds. Amer. Fern Jour. 55:167—70.

93

Long-Distance Dispersal: Evidence and Implications

Maynard, N. G. 1968. Significance of air-borne algae. Z. Allg. Mikrobiol. 8:225-26.

Mayr, E. 1940. The origin and history of the bird fauna of Polynesia. Proc. Sixth
Pacific Sci. Congr. 4:197-216.

Moar, N. T. 1969. Possible long-distance transport of pollen to New Zealand. New
Zealand Jour. Bot. 7:424—-26.

Moore, J. W. 1933. New and critical plants from Raiatea. Bishop Mus. Bull.
102:1-53.

Moore, J. W. 1963. Notes on Raiatean flowering plants. Bishop Mus. Bull.
226:1-36.

Munro, G. C. 1960. Birds of Hawaii. 2d ed. Rutland, Vt., Ridgeway Press.

Ornduff, R. 1963. Experimental studies in two genera of Helenieae (Compositae):
Blennosperma and Lasthenia. Quart. Rey. Biol. 38:141—50.

Overeem, M. A. van. 1937. On green organisms occurring in the lower troposphere.
Rec. Trav. Bot. Neerl. 34:389-439.

Preest, D. S. 1963. A note on the dispersal characteristics of the seed of the New
Zealand podocarps and beeches and their biogeographical significance. In Pacific
basin biogeography, ed. J. L. Gressitt, pp. 415-24. Honolulu, Bishop Museum
Press.

Proctor, V. W. 1961. Dispersal of Riella spores by waterfowl. Bryologist 64:58-61.

Proctor, V. W. 1962. Viability of Chara oospores taken from migratory water birds.
Ecology 43:528-29.

Proctor, V. W. 1966. Dispersal of desmids by waterbirds. Phycologia 5:227-32.

Proctor, V. W. 1968. Long-distance dispersal of seeds by retention in digestive tract
of birds. Science 160:321-22.

Proctor, V. W., and C. R. Malone. 1965. Further evidence of the passive dispersal
of aquatic organisms via the intestinal tract of birds. Ecology 46:728—29.

Proctor, V. W., C. R. Malone, and V. L. de Vlaming. 1967. Dispersal of dis-
seminules recovered from the intestinal tract of captive killdeer. Ecology
48:672-76.

Pugh, G. J, F. 1964a. Fungal contamination of autumn migrants. Bardsey Observ.
Rep. 12:1-3.

Pugh, G. J. F. 1964b. Dispersal of Arthroderma curreyi by birds, its role in the soil.
Sabouraudia 3:275-78.

Pugh, G. J. F. 1965a. Fungi recorded on birds from Skokholm. Skokholm Bird
Obsery. Rep. 1965: 21-24.

Pugh, G. J. F. 1965b. Cellulolytic and keratinophilic fungi recorded on birds.
Sabouraudia 4:85-91.

Pugh, G. J. F. 1966a. Associations between birds’ nests, their pH, and keratinophilic
fungi. Sabouraudia 5:49-53.

Pugh, G. J. F. 1966b. Fungi on birds in India. Jour. Indian Bot. Soc. 45:296-303.

Raven, P. H. 1963a. A flora of San Clemente Island, California. Aliso 5:289-347.

Raven, P. H. 1963b. Amphitropical relationships in the floras of North and South
America. Quart. Rey. Biol. 38:151-77.

Reinecke, F. 1898. Die Flora der Samoa-Inseln. II. Bot. Jahrb. 25:578—708.

94

Long-Distance Dispersal: Evidence and Implications

Ridley, H. N. 1930. The dispersal of plants throughout the world. Ashford, England,
L. Reeve & Co.

Roessler, E. S. 1936. Viability of weed seeds after ingestion by California linnets.
Condor 38:62-65.

St. John, H., and W. R. Philipson. 1960. List of the flora of Oeno Atoll, ‘Tuamotu
Archipelago, south-central Pacific Ocean. Trans. Roy. Soc. New Zealand 88:401-3.

Sauer, J. 1957. Recent migration and evolution of the dioecious amaranths. Evolu-
tion 11:11-31.

Schlichting, H. E., Jr. 1964. Meteorological conditions affecting the dispersal of
airborne algae and protozoa. Lloydia 27:64-78.

Simberloff, D. S. 1969. Experimental zoogeography of islands. A model for insular
colonization. Ecology 50:296—-314.

Simberloff, D. S., and E. O. Wilson. 1969. Experimental zoogeography of islands.
The colonization of empty islands. Ecology 50:278-95.

Skottsberg, C. 1922a. Catalogue of Easter Island phanerogams. Natur. Hist. Juan
Fernandez Easter I. 2:63-84.

Skottsberg, C. 1922b. The phanerogams of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:95—240.

Skottsberg, C. 1927. The vegetation of Easter I. Natur. Hist. Juan Fernandez Easter
IT. 2:487—502.

Skottsberg, C. 1928. Pollinationsbiologie und Samenverbreitung auf den Juan
Fernandez-Inseln. Natur. Hist. Juan Fernandez Easter I. 2:503—547.

Skottsberg, C. 1937. Die Flora der Desventuradas-Inseln. Acta Horti Gotoburg.
5(6):1-88.

Skottsberg, C. 1951. A supplement to the pteridophytes and phanerogams of Juan
Fernandez and Easter I. Natur. Hist. Juan Fernandez Easter I. 2:763-92.

Skottsberg, C. 1963. Zur Naturgeschichte der Insel San Ambrosio (Islas Desventuradas,
Chile). Ark. Bot., ser. 2, 4:465-88.

Stephens, S. G. 1958. Factors affecting seed dispersal in Gossypium and their pos-
sible evolutionary significance. N.C. Agr. Exp. Sta. Tech. Bull. 131:1-32.

Stephens, S. G. 1966. The potentiality for long-range oceanic dispersal of cotton
seeds. Amer. Natur. 100:199-210.

Stewart, A. 1911. A botanical survey of the Galapagos Islands. Proc. Calif. Acad.
Sci., 4th ser., 1:7—288.

Stewart, A. 1912. Notes on the botany of Cocos Island. Proc. Calif. Acad. Sci.,
4th ser., 1:375-404.

Taylor, B. W. 1954. An example of long distance dispersal. Ecology 35:569—72.

Thorne, R. F. 1963. Biotic distribution patterns in the tropical Pacific. In Pacific
basin biogeography, ed. J. L. Gressitt, pp. 311-50. Honolulu, Bishop Museum
Press.

Thorne, R. F. 1972. Floristic relationships between Tropical Africa and Tropical
America. In Tropical forest ecosystems in Africa and South America: A compara-
tive review, ed. B. J. Meggars and E. S. Ayensu, pp. 27-47. Washington, D.C.,
Smithsonian Institution.

95

Long-Distance Dispersal: Evidence and Implications

Turbott, E. G. 1951. Notes on the birds of the Three Kings Islands. Rec. Auckland
Inst. Mus. 4:141-43.

Vlaming, V. L. de, and V. W. Proctor. 1968. Dispersal of aquatic organisms:
Viability of seeds recovered from the droppings of captive killdeer and mallard
ducks. Amer. Jour. Bot. 55:20-26. |

Wagner, R. H. 1965. The annual seed rain of adventive herbs in a radiation-dam-
aged forest. Ecology 46:517-20.

Whitaker, T. W., and G. F. Carter. 1961. A note on the longevity of seed of
Lagenaria siceraria (Mol.) Standl. after floating in sea water. Bull. Torrey Bot.
Club 88:104-6.

Whitehead, D. R., and C. E. Jones. 1969. Small islands and the equilibrium theory
of insular biogeography. Evolution 23:171-79.

Williams, E. 1969. The ecology of colonization as seen in the zoogeography of
anoline lizards on small islands. Quart. Rev. Biol. 44:345-89.

Wilson, E. O. 1959. Adaptive shift and dispersal in a tropical ant fauna. Evolution
13:122-44.

Wilson, E. O., and D. S. Simberloff. 1969. Experimental zoogeography of islands.
Defaunation and monitoring techniques. Ecology 50:276-77.

Wilson, R. C. 1970. Anthocarp dispersal and its ecological implications. Ph.D.
Thesis. Claremont Graduate School, Claremont, Calif.

Yuncker, T. G. 1959. Plants of Tonga. Bishop Mus. Bull. 220:1-283.

Zimmerman, E. C. 1948. Insects of Hawaii, vol. 1: Introduction. Honolulu, Univer-
sity of Hawaii Press.

96

APTER

ADAPTIVE RADIATION IN ISLANDS:
DEFINITIONS AND COMMENTS

The nature of adaptation is often discussed in books on evolution and
may, in fact, be said to constitute the major concern of some books, such
as Grant’s The Origin of Adaptations (1963). Simultaneously we see the
term “adaptive radiation” often used, sometimes briefly defined. Adaptive
radiation is one of the more intnguing concepts in evolution, because it
has manifold overtones. The concept is usually not explicitly defined—
rather, it is applied to examples, such as the Darwin’s finches of the
Galapagos Islands, and example thereby becomes definition. To see the
limits of this idea, and how it impinges on other modes of speciation and
diversification, I have decided to present what I consider a series of impli-
cations evident in usages of this term with reference to insular and island-
like situations.

1. Adaptive radiation connotes entry of a group into a variety of
habitats. This in turn implies that in a reasonably discrete geographical
region, usually an island but more particularly an archipelago, there is a
diversity of habitats available. In terms of vegetation, one often thinks of
a range such as dry coastal plain, dry inland scrub, dry upland forest, wet
montane forest, bogs, subalpine scrub, and alpine subscrub. Such a range
would be represented well, for example, on the Hawaiian Islands. A fresh-
water lake may also offer many habitats (see examples in chapter 9).

2. A plant group showing adaptive radiation would be expected to
be represented by species in many, although not necessarily all, of these
habitats. The diversity of habitats occupied is more often implied than is

97

Adaptive Radiation in Islands: Definitions and Comments

adaptation to the totality of an ecological range. Many areas, of course, do
not offer a broad range of ecological amplitude. In any case, diversifica-
tion among available habitats is considered a basic criterion of adaptive
radiation. A single habitat might be used differently by derivatives of the
same group. For example, Dubautia menziesii is an alpine shrub, Argy-
roxiphium sandwichense a rosette plant; both are members of the same
complex and are sympatric on Haleakala, Maui.

3. A particular animal group not only may show occupation of varied
climatic zones, but also may enter varied behavioral and dietary modes.
Lack (1944) showed that, among birds, closely related species tend to oc-
cupy separate habitats, or, if they do coexist, they take different foods—
which can be called “resource partitioning” in a single area. Examples of
resource partitioning occur in unexpected ways in lizards studied by
Schoener (1968). For example, perch diameter, perch height, and perch
nature (leaves as well as bark) separate Anolis species in the Bahamas. ‘The
length and volume of prey items and preference for particular taxonomic
groups of insects are criteria for resource-partitioning among the Bimini
anole species (Schoener, 1968), just as they are among the Darwin’s finches
(Bowman, 1961, pl. 6). Among the Indefatigable Island geospizids, Bow-
man illustrated visible differences in prey sizes and volumes and types for
Camarhynchus psittacula, C. parvulus, Certhidea olivacea, and Cactospiza
pallida.

In the Bimini Anolis species, Schoener (1968) describes not merely
differences, but specifically which species have more generalized require-
ments or tolerances in a given area, which have the narrowest. The effect
of sexual dimorphism in size of Anolis in relation to size of prey and
volume of prey was earlier expressed for a single Anolis species (Schoener,
1967).

Similar resource partitioning occurs on other West Indian islands, such
as Grenada and Martinique (Schoener and Gorman, 1968). In addition to
factors such as those cited above, Schoener finds that sympatric Anolis
species differ modally in head length and body temperature. Schoener and
Gorman (1968) liken differences in snout length in lizards to bill differ-
ences in birds.

We are now familiar with the phenomenon whereby two sympatric

98

Adaptive Radiation in Islands: Definitions and Comments

species of a given genus on an island will differ, for example, in the case
of a bird, in bill size and shape—differences indicating divergence in feed-
ing habits. If a single species of a given genus is present on an island,
intermediate bill size and shape (and presumably prey) occur—and the
range in bill dimensions will perhaps be as great as the combined ranges
for the two sympatric species on another island. Graphic illustrations of
this phenomenon have been offered by Lack (1947, chapter 8) in the geo-
spizids and Bock (1970) in the Hawaiian honeycreepers (see also fig. 4.5 in
this book). We can easily understand that if a single species of a genus
exists on an island, it may widen its range in diet and morphology, a
phenomenon termed “character release” (see Grant, 1966; Williams,
1969). The phenomenon whereby congeners on an island develop mutually
exclusive ecological ranges has been termed “character displacement”
(Brown and Wilson, 1956). Character displacement can involve features
other than ecological requirements. For example, two congeneric species
of frogs may have the greatest voice difference in the area of sympatry, in
which case this should be regarded as an isolating mechanism.

What is not certain, to judge from the literature, is how character
displacement occurs. The idea that two species diverge ecologically so
that selection favors individuals which do not compete with those of a
congeneric species on an island is tempting. Certainly the presence of
congeneric species would prevent character release by either species; the
competition would tend to enforce differences. There is the possibility that
the differences would be enforced most strongly at geographical interfaces
between two species. Once “character displacement” had been secured,
they could become sympatric. However, for such a situation to exist, there
must already be at least some degree of differentiation between two
species that occupy different parts of an island. Williams (1969) has rightly
stressed that an invading species, unless different in ecological require-
ments from a species already resident, will be potentially unsuccessful in
establishment. Where, then, can the new invader acquire characteristics
different from those of the previously established resident on a given
island?

Of course, mainland source areas can provide congeneric species dif-
ferent enough so that both can establish on an offshore island. This may
be the case for the Tres Marias Islands congeners cited by Grant (1966).

99

Adaptive Radiation in Islands: Definitions and Comments

Alternately, how can species with requirements different enough to be
sympatric evolve on an island? One possibility lies in the archipelago effect.
If a single bird (or plant) species establishes on an island, it will, with
ecological opportunity, undergo character release. This will mean a
greater range in its ecological capabilities. Character release can quite con-
ceivably take the form of race formation within a single island. As in the
ecotype concept, a single variable species (like Metrosideros polymorpha in
the Hawauan Islands) may range from relatively dry forest to quite wet
forest, into bogs (as a subshrub instead of a tree), and into specialized
habits on windswept ridges. Presumably a similar ecotype formation could
occur on a moderately large island in the case of a bird, where a certain
forest area would provide one sort of prey, yet a different forest area
would be rich in another. If opposite ends of this spectrum were dis-
persed to a second island in an archipelago, they would be able to coexist
in a single forest if both types of prey were available there.

4. Although the reasoning just given suggests the beginnings of
adaptive radiation, most examples feature not a pair of divergent species,
but many of them, often on a single island, such as the variety of geo-
spizid species on Indefatigable Island (fig. 6.1). How are such broader
programs of adaptive radiation realized? Bock (1970) rightly points to
“cumulative change” as a way of expressing this process. Another way of
describing it might be “successive ecological shifts.” If a group has suff-
cient time and ecological opportunity, and if it is sufficiently plastic, eco-
logical shifts (perhaps like those cited in the preceding section) can occur.
In the case of the Hawaiian honeycreepers, character release that involved
a spectrum including more nectar—few insects at one end and less nectar—
more insects at the other would represent an ecological shift. The more
nectarivorous line could radiate into various niches, each niche a secondary
ecological shift, while the insect-feeding line could radiate into other new
niches (small insects vs. large, in borings vs. on surface). Adaptive radiation
is thus a series of ecological shifts, and there is no group that exemplifies
the “perfect” stage in adaptive radiation. “Secondary cycles” of adaptive
radiation may be present, as in Amadon’s (1950) interpretation of differ-
ences within the genus Psittarostra of the Hawaiian honeycreepers, or we
may see a group with incipient adaptive radiation, like the shrubby insular
Hemizonias of California (Carlquist, 1965).

100

Adaptive Radiation in Islands: Definitions and Comments

There is, additionally, the possibility that in any group cited as an
example of adaptive radiation, the maximal differentiation has already
been reached. There is every reason to believe that such processes can
occur very rapidly (see, e.g., Myers, 1960), and that, in a finite area, niche
occupation and ecological shift have proceeded to a maximal extent. Conse-
quently, what looks like “incipient adaptive radiation” may, in fact, be
the farthest penetrations of which a group is capable in a given area.

5. Disharmony in a relatively discrete geographical region is the
implied condition predisposing to adaptive radiation, at least in most of
the examples cited. If a variety of niches available to a group are not
already efficiently occupied (e.g., saturated with species adapted to each
niche), they are invadable and speciation will take the form of diversifica-
tion into different habitats or into other biological gamuts. Oceanic islands
offer obvious examples of disharmony, but continental situations provide
many examples as well. Australia was disharmonic with respect to mam-
mals when the marsupials entered, so their radiation there is an obvious
result. One may argue that Australia (or the Australian-Papuan region,
since New Guinea, Australia, and Tasmania were connected with each
other until recently) is an insular situation. Some examples of adaptive
radiation in continental areas are cited briefly in chapter 9 as a way of plac-
ing in relief the insular examples.

However, the major continents themselves can be disharmonic with
respect to particular groups at various times in their geological history. To
cite radiation of, say, cacti or bromeliads (both are large families restricted
to North and South America, with only the most token representation
elsewhere) may seem to inflate the concept beyond logical bounds, but it
is an obvious extension of the disharmonic concept.

Keast (1961, 1968, 1971) has shown that in some areas of Australia,
open niches have been colonized—on Tasmania, for example, mostly by
species from adjacent feeding zones. Because of excessive cold in Pleisto-
cene Tasmania, that island became disharmonic and has been ecologically
“tefilled” in post-Pleistocene time. Southwestern Australia as a floristic
province is a superb example of an insular area and, for this reason, is dis-
cussed in detail in chapter 8. Likewise, the alpine and subalpine zones of
equatorial Africa, South America, and Malesia seemed to me to be in-
evitable components in an examination of insular adaptive radiation

101

Adaptive Radiation in Islands: Definitions and Comments

(chapter 14). Disharmony in these cases may not seem obvious, for one
may have the preconception of a single ecological zone. However, in both
cases there are a variety of niches, and disharmony has been strong, for
reasons stated in chapters 8 and 14.

6. Autochthonous development within a given area is virtually as-
sumed in the usage of the term adaptive radiation. If Darwin’s finches
represented a collection of mainland finches, independently dispersed to
and established upon the Galapagos, we would not cite them as examples
of adaptive radiation (unless in the broadest possible context, saying that
finches as a whole represented a case of adaptive radiation). Another way
of saying this is that the species of a genus in an area are monophyletic,
the products of a single invasion rather than several invasions (and there-
fore, for that area, polyphyletic). This may seem so obvious as to be super-
fluous, but the differences may be subtle where the possibilities exist that
products of several invasions may look like products of a single continuum.
Lemurs are now restricted to Madagascar, but they were once much more
widely distributed. How much of the diversity of Madagascar lemurs is the
result of evolution on Madagascar, how much is relictual preservation of
evolution that took place elsewhere?

7. A particular taxonomic group is basic to the concept of adaptive
radiation. It need not be endemic to the region in question; many cited
are endemic, however, like the Hawaiian honeycreepers. Adaptive radia-
tion is most often not cited at higher taxonomic levels, e.g., family and
above. However, Stebbins (1967) has used the concept of adaptive radia-
tion with respect to flowering plants at large. Although this extension is a
useful way of stressing problems of evolution in angiosperms, its usage in
this context is somewhat by way of analogy with the insular and islandlike
instances that provide the bulk of examples cited. The taxonomic levels
most often cited are subfamily (Geospizinae of Fringillidae), groups of
related genera (the Dubautia-Argyroxiphium-Wilkesia complex), genus
(Rollandia or Cyanea: chapter 4), or subgenus (Senecio subg. Dendro-
senecio: chapter 14). Diversification within a single species is not often
considered adaptive radiation—usually only an incipient form of it. The
concept of ecotype formation is more applicable within a species. It is
possible to regard all evolution as successions of adaptive radiation, as is
done by Mayr (1963), in a sense. However, in this instance, we need the

102

Adaptive Radiation in Islands: Definitions and Comments

insular examples of adaptive radiation as modules to which other cases can
be compared.

§. Disharmony leading to adaptive radiation is a product of difficulty
of dispersal to an (insular) area (if we discount depauperation of a biota
for reasons of poor ecology, of course). MacArthur and Wilson (1967,
p. 175) stress an aspect of this by saying that a group established on an island
at the limit of the group’s dispersal possibility (which they term the “radia-
tion zone’) will be likely to undergo adaptive radiation. An obvious
example would be land birds on the Hawaiian Islands. As Lack notes, only
five passerine bird establishments by natural means have occurred on the
Hawaiian Islands. Faced with the ecological richness of this archipelago,
they inevitably underwent radiation. However, we might ask why the
honeycreepers radiated so spectacularly and the other four passerine estab-
lishments did not. Were the honeycreepers older immigrants, or just more
capable of ecological shift?

The idea of a group at its dispersal limits is very appealing, but we
must keep in mind that different groups have different dispersal abilities.
If we compare the Hawaiian Islands to the Canary Islands, where many
more passerine bird establishments—and virtually no adaptive radiation in
the avifauna—have occurred, this is easy to understand. However, about
250 angiosperm establishments occurred in the Hawaiian Islands, and
adaptive radiation in Hawaiian angiosperms has been spectacular. Obvi-
ously, the opportunity for plant radiation in the Islands is much greater
than that for birds. There are several reasons why this should be true. A
breeding population of a bird species may require a much larger territory
than would a plant species. Isolation within an island may be much
greater for a plant species than an animal species. For example, Cyrtandra
grows in deep ravines, of which there are many on Oahu, and the inter-
vening ridges can serve as isolating barriers that have permitted extraor-
dinary speciation in that genus (126 species on Oahu, according to St.
John {1966]). Thus, we have to consider not only dispersal limits of a
group, but also differential dispersibility for various groups, and the eco-
logical requirements, ecological plasticity, and population characteristics
of each. We also have to consider ecological opportunity on recipient
islands—what appears to be uniform for one group of organisms may be a
series of ecological niches for another. Furthermore, recentness of coloniza-

103

Adaptive Radiation in Islands: Definitions and Comments

tion may be a factor in explaining lack of radiation. This is hypothesized
(with reservation) by MacArthur and Wilson (1967), and they cite the
possibility that recentness may explain lack of radiation of New Zealand
frogs, Antillean insectivores, Solomon Islands mammals, Fijian snakes,
and Samoan lizards. However, when we consider the rapid radiation and
speciation of groups such as the Hawauan drosophilids or the Galapagos
geospizids, the earlier examples do not seem easily explained by lack of
time. The New Zealand frog Leiopelma, related to frogs in northwestern
America, is undoubtedly not recent in New Zealand. Leiopelmid frogs are
known from Upper Jurassic (Sharell, 1966). That this frog is prevented
from radiation in New Zealand by ecological restrictions is suggested by
the following passage (Sharell, 1966):

The native frogs of New Zealand have adapted themselves most won-
derfully to life and their mode of breeding to the rigorous conditions
of their environment. Life is precarious enough on these mountain
tops, and as there is no water for tadpoles to live in and to develop
into frogs [in], their eggs are enclosed in gelatinous capsules, filled
with a watery fluid, and inside this capsule, the embryo develops, miss-
ing out the long tadpole stage, and developing directly into tailed
froglets.

Dispersal difficulty is not, of course, a matter of distance alone; we
tend to forget other factors because we cannot visualize them easily. Lack
(1947) has noted that passerine bird dispersal to the Galapagos may have
been more difficult because these islands are in a region of calm, whereas
Macaronesian islands are favored by prevailing winds.

9. Recentness of diversification is an implication partly involved in
the concept of adaptive radiation, although rarely made explicit. Recent-
ness is suggested by persistence of various types, whereas greater age would
dictate progressive extinction of at least some of them. Persistence among
recent species chiefly would be expected on oceanic islands, where popu-
lation size for each product of adaptive radiation in a group would be
expected to be finite. We would not cite a group as exemplary of adaptive
radiation unless a series of modifications were still extant.

Products of adaptive radiation may persist well on islands because
of lowered predation and competition. However, equal relictualism for all

104

Adaptive Radiation in Islands: Definitions and Comments

or even many products of an adaptive radiation seems unlikely, and only
recentness seems of overriding importance in explaining the excellent
gamut of niche occupation we see in the “good” examples of adaptive
radiation. There are a few exceptions such as the New Caledonian
conifers (chapter 7) or the marsupials of Australia, but even in those
groups, and ones like them, we see gaps, almost certainly due to extinctions
(known as fossils in the case of the Australian marsupials). Even in
the Hawaiian honeycreepers we see a difference between the subfamily
Drepanididae (probably more extinctions, fewer adaptive modes) and the
subfamily Psittarostrinae (fewer extinctions, more numerous adaptive
modes).

10. One way often utilized for defining adaptive radiation is com-
paring the products of adaptive radiation on an island or islandlike area
to mainland counterparts that would belong to a series of disparate
families. For example, a parrot would be considered the counterpart of
Psittarostra in the Hawaiian honeycreepers, a meliphagid the counter-
part of Hemignathus procerus or Drepanis, etc.

11. Despite the fact that a gamut of adaptations is basic to adaptive
radiation, discontinuities in the range of forms within a taxonomic group
are also implied in this concept. The Hawaiian species of Metrosideros
(Myrtaceae) can be said to represent a wide range of adaptations, ranging
from one of the tallest island trees in wet forests, to miniature shrubs in
bogs, to scrubby trees on windy ridges, spanning edaphic conditions from
new lava to mature soils. However, Hawaiian Metrosideros represents an
almost continuous series of forms, very probably the result of constant
hybridization. One could designate the entire complex as a single species,
M. polymorpha, or single out as many subspecies or segregate species as
one wished. In any case, the intercontinuity among infraspecific variants
evidently disqualifies Metrosideros from consideration as an example of
“adaptive radiation” in the typical sense; “good” taxonomic species
throughout most if not all of a group seem a requisite for its designation
as an example of this process.

12. Speciation (or diversification to the level of genera) progresses by
means of continual shifts into new habitats and habits, and thus a contrast
is made with vicarious species. The largest genus of angiosperms in the
Hawaiian flora is Cyrtandra. Although this genus is not lacking in diversity

105

Adaptive Radiation in Islands: Definitions and Comments

in certain characteristics, many of the species are alike in habit, and the
differences that separate them appear to be nonadaptive in many cases,
the result of drift permitted by geographic isolation. Cyrtandra is especially
characteristic of deep wet gulches; where these are numerous, isolation
and thereby drift would tend to occur. This may account for the phenom-
enal number of species on Oahu (St. John, 1966). At least the species
within each section of the genus (e.g., sect. Chaetocalyx) would be regarded
as vicarious species and would be contrasted with the products of adaptive
radiation. Strid (1970) regards Nigella (Ranunculaceae) on the Aegean
Islands as an example of “nonadaptive radiation.” These Nigella species
show morphological differences in relatively minor ways, and Strid claims
that they all have roughly the same ecological preferences. Differences
in this group appear to have resulted from the “founder principle” (namely,
the limitations of genetic content of the founding immigrants predetermine
the composition of the resultant population), with subsequent action of
genetic drift. We could call the Nigella species vicarious species rather
than products of “nonadaptive radiation.” The equatorial alpine species
of Senecio and Lobelia and the paramo species of Espeletia (chapter 14)
contain many examples of vicarious species, but adaptive radiation has
also occurred in these groups. Although the contrast is made to vicarious
species, the concept of adaptive radiation does not exclude instances of
vicarious species within a group as well: it is a matter of the preponderance
of “adaptive” instances over “nonadaptive” instances in a group.

13. There is a strong implication in the concept of adaptive radiation
that products are strikingly different, often exceptional. The diverse growth
forms of the Dubautia-Argyroxiphium-W ilkesia complex on the Hawaiian
Islands, and the fact that these genera represent growth forms and habitat
preferences different from, and beyond, those of mainland tarweeds makes
them a good example of insular adaptive radiation. A group in which
species look much alike but occupy diverse niches might not at first be
cited as an example, but could, with suitable ecological studies, be cited
thus. There is often the implication that the nature of the adaptation be
readily apparent in terms of morphology (bill size and shape in birds as
an indicator of diet; plant size, leaf size and shape, indument, and degree
of succulence as indicators of xeromorphy or mesomorphy).

Myers (1960) stresses exceptional developments, “‘supralimital special-

106

Adaptive Radiation in Islands: Definitions and Comments

izations,” as components of adaptive radiation. Among the instances he
cites in fishes are the exceptional developments of the cyprinids of Lake
Lanao and the cichlids of Lake Tanganyika (see chapter 9). Other
examples he cites include a scaleless genus of cyprinids, Sawbwa, in Lake
Inleé, Burma (Annandale, 1918), and the strikingly altered species of
Orestias in Lake Titicaca. If autochthonous products of evolution in lake
fish are supralimital and types connecting them to other fish are extin-
guished, we would have no choice but to recognize them as families.
Considering the supralimital developments in other lakes, this explanation
may apply to two monotypic genera that are sole representatives of their
families: Chaudhuria caudata (Chaudhuriidae) in Lake Inlé, Burma (An-
nandale, 1918) and Indostomus paradoxus (Indostomiidae) in Lake In-
dawygi, Burma (Prashad and Mukerji, 1929). Authors such as Myers (1960)
and Zimmerman (1948) have stressed that exceptional products of adap-
tive radiation on islands or in islandlike situations demonstrate to us bet-
ter than examples on continents how new families and orders have arisen
evolutionarily. :

14. With a few notable exceptions, adaptive radiation seems to occur
mostly in groups that could be called weedy, evolutionarily upgrade, or
plastic. That composites (Asteraceae) and insects furnish many examples
of adaptive radiation on islands is no accident. Composites, often char-
acteristic of weedy or disturbed situations on mainland areas, plainly have
a number of features that tend to promote rapid speciation and ecological
shift. This is certainly true of the family on continental areas, so its adap-
tive radiation on islands is not surprising. Chapter 4 contains examples of
adaptive radiation in both Asteraceae and insects in the Hawaiian Islands.
We should observe, however, that taxonomic groups not often considered
weeds do have excellent pioneering and character-release capabilities, and
have exhibited adaptive radiation in islands (see chapter 1). There is a
link between good dispersal and pioneer qualities.

15. As stressed by Baldwin (1953, p. 382), there is a tendency to read
as adaptive all differing or distinctive characteristics within a group cited
as an example of adaptive radiation. In an insular area, nonadaptive char-
acteristics may be established and preserved coincidentally with adaptive
ones. The Hawaiian honeycreepers are a superb example of adaptive radia-
tion, but color and pattern of plumage, as well as a few other characteris-

107

Adaptive Radiation in Islands: Definitions and Comments

tics, appear to have no adaptive significance (Amadon, 1950; Baldwin,
1953). Plumage color in the birds of the Tres Marias Islands has changed,
but apparently through decrease in selection for distinctive plumage recog-
nition patterns (see chapter 15).

16. Species involved in cases of true adaptive radiation on islands
show just as “perfect” or “complete” adaptation to environment as do
comparable continental species to their respective environments. Speaking
of the Australian marsupials, Darwin (1859) thought of them as “divided
into groups differing but little from each other, and feebly representing
carnivorous, ruminant, and rodent mammals.” I have offered a refutation
of Darwin’s judgment of the Australian marsupials (1965). Similarly, Bock
(1970) says, with respect to the Hawaiian honeycreepers, “I know of no
reason to assume that any species of drepanidid has a bill morphology
poorly adapted for its feeding methods.” The fact that products of adap-
tive radiation on islands are more readily subject to extinction, or seem
less competitive than continental types introduced by man, has been
equated by some authors with imperfect adaptation, as Darwin evidently
did. Extinction of marsupials or honeycreepers results not from imperfect
adaptation, but from other causes. These causes may include introduction
of predators, parasites, and competitors, and seems invariably to involve
destruction of natural areas by man. Groups that have evolved in the
absence of such influences are more readily vulnerable to them (see chap-
ter 1, principle 18).

17. If one group invades and radiates within an archipelago or similar
area, it may preclude not establishment of other immigrants, but adaptive
radiation in those later arrivals. This may explain why the drepanidids are
the only group of Hawaiian birds to exhibit adaptive radiation, and may
apply to the Darwin’s finches as well. Earliness of invasion may not be
advantageous if a group is unable to diversify into new ecological niches.
Thus, a genus such as Cryptocarya (Lauraceae) could well be an earlier
immigrant to Hawaii than the ancestors of the Dubautia-Argyroxiphium-
Wilkesia complex but, if so, has not been able to radiate (there is only
one species, C. mannii). Because Lauraceae (other than the parasitic Cas-
sytha) are so consistently elements of stable moist forest, one would expect
this limitation. That Cryptocarya is indeed an older immigrant in the
Hawaiian flora is suggested by its occurrence only on the oldest island
where moist forest is present, Kaual.

108

Adaptive Radiation in Islands: Definitions and Comments

Examples have been cited of instances in which successful earlier
establishments on islands may forestall later establishments (chapter 2).
Forestalling not a second establishment but the adaptive radiation of that
second establishment seems a logical extension of this concept. On the
other hand, Lack (1947) submits the possibility that if dispersal from a
source area is good, new immigrants (more efficient, presumably, than
relatively less competitive insular species) may displace products of earlier
adaptive radiation. In such a way, groups of flowering plants may have dis-
placed conifers on New Caledonia (chapter 7). Continual displacement is
basic to the MacArthur-Wilson equilibrium theory (MacArthur and Wil-
son, 1967), although one could express this in terms not of displacement
but of replacement.

18. A group with moderately poor dispersibility may conceivably
undergo adaptive radiation better than one with good dispersibility within
an archipelago. Excellence of dispersal would tend to result in lack of spe-
ciation or diversification, for continual reinvasion of a site from various
sources might tend to “swamp out” ecological shift—as it almost certainly
does in the strand flora. Excellent dispersibility may account for the fact
that the only species of lobelioid present on all major islands of the Hawai-
ian chain, Trematolobelia macrostachys, is the sole species of its genus.
Trematolobelia has an unusually efficient dispersal mechanism (Carlquist,
1962). Isolation of a population within an archipelago long enough for it
to undergo ecological shift is easy to imagine. Once ecological shift has
occurred, however, the population has “lost dispersibility” by being unable
to reinvade its original habitat.

19. Adaptive radiation is the product of invasion of an “invadable”’
situation. Given areas are invadable by particular groups at particular times
—often for geological and climatic reasons. Although hypothetical cur-
ricula could be designed here, these basic evolutionary programs are well
known. In particular, I refer the reader to the accounts of Mayr (1963,
especially pp. 481-585) and Grant (1963, especially pp. 392-423, 519-68).

“Diversity” and Adaptive Radiation

In recent years, much attention has been paid to the number of species
(genera, families) present on insular areas, and the relationship of species

109

Adaptive Radiation in Islands: Definitions and Comments

numbers to factors such as area and isolation. These topics have been so
extensively and excellently discussed by MacArthur and Wilson (1967)
that there is no need to repeat that readily available material here. ‘Two
subsequent contributions may be mentioned: Balgooy’s “A study on the
diversity of island floras” (1969), and Simberloff’s “Taxonomic diversity of
island biotas” (1970). For all of these authors, diversity is a synonym or
near-synonym for richness in number of species. Although adaptive radia-
tion clearly impinges on the “richness” concept of floras and faunas, the
two concepts differ markedly, as we shall see, and thus the material to be
offered here is not an elaboration of diversity concepts in the sense of
these authors. Rather, I am viewing diversification from the natural-history
standpoint, in order to see what ecological shifts occur, what morphological
changes occur, and in which islands or archipelagos. Simberloff’s study is
almost entirely alien to my present discussion, for he deals nearly exclu-
sively with continental islands or fringing archipelagos, whereas I have, in
effect, used “definitive” oceanic islands as a source of characteristics, and
have extended these concepts to other islands and to islandlike areas to the
extent to which they apply there in order to note changes as one shifts to
“less definitive” islands. Simberloff’s conclusions may well be correct for
his material, but because his data were derived from “near” islands, his
conclusions may or may not apply to other situations (they clearly do not
apply to the Hawaiian Islands, for example). Balgooy offers data from seed
plants which, in essence, confirm the various conclusions of MacArthur
and Wilson. Balgooy’s measure of diversity is the genera-per-family ratio,
whereas Simberloff deals with the species-per-genus ratio. What relation-
ship, if any, do these ratios bear to instances of adaptive radiation? For
the purposes of this discussion, I have presented a table based on angio-
sperm floras of various regions (table 3.1), introduced species excluded.
These floras or compilations are, of course, not equally complete in cover-
age of their respective areas or comparable in taxonomic concepts of their
authors, but these flaws alter statistical values little except where noted in
the discussion. The reader should be aware that, among these areas, our
knowledge of southwestern Australia is particularly inadequate, and we
may expect the number of species known there to increase by as much as
10 percent.

From the data on numbers of families, genera, and species, a genera-

110

Adaptive Radiation in Islands: Definitions and Comments

per-family (G/F) ratio has been computed, much as done by Balgooy
(1969). Balgooy’s preference for this ratio seems to lie chiefly in his conten-
tion that there is less disagreement among taxonomists on the limits of
genera and families than there is on species. This may or may not be true,
but, in any case, I have also calculated the species-per-genus (S/G) ratio,
as used in a number of examples by MacArthur and Wilson (1967) and
throughout the paper by Simberloff (1970). In addition, I have computed
a species-per-family (S/F) ratio, for which there is little precedent. The
S/F ratio seemed an interesting indication of speciation, and one in which
the difference among floras might be stressed in a more highly pro-
nounced form.

Infraspecific categories have been omitted with the exception of the
Hawaiian flora. Here, Fosberg’s estimate includes species plus varieties,
and is the only recent accurate estimate; Fosberg’s figure is probably not
far from the figure for species alone that would be recognized by more
“liberal” taxonomists working in that flora. However, I believe that monog-
taphers dealing with the Hawaiian flora since the time of Hillebrand (1888)
have tended to recognize species in excess of what current concepts might
indicate. Consequently, I have included, in parentheses, what I would
consider a conservative estimate for the number of Hawaiian angiosperm
species (1200), excluding subspecies or varieties. Even with a conservative
estimate, however, the position of the Hawaiian flora in the ratios remains
the same, and it still shows appreciably greater S/G or S/F ratios than do
any other insular floras.

The ratios obtained seem entirely as one would expect on the basis
of similar work by the authors cited. However, I feel that the results show
that one ratio alone does not express diversity adequately. For example,
the Hawaiian Islands have a relatively low G/F ratio (2.6), but a high
S/G or S/F ratio (at least 14.5), rating much higher than a large conti-
nental island area like the British Isles.

The conclusions, as seen from the perspective of native angiosperms,
seem obvious. Islands support a fraction of the diversity of large conti-
nental areas, with the exception of the very few islands or islandlike areas
that tend to foster speciation to a high extent: the Hawaiian Islands and
southwestern Australia. These areas would probably show similar figures
in the case of insect faunas, but not land-bird faunas, where speciation on

11]

(6191 ‘SSIIIN) BUSH “IS PUR “(PFET ‘O2%T) spuRjst vouMD Jo FIND ‘(Lggt Wye) sMALMEIN, *(OS6T BOqSHOYS +6961
‘Kooseg) sesonbieyy ‘(S961 ‘ouloY],) BUOpe[eD AXON {(9¢6] ‘d1oqs}OYS) “s] zopueuroy uen{ ‘([gp6,] weuoururz, Wor
‘eyep $,d1oqso1y]) “S] ULIIeMePT “L16[ “aemoys) “sy sosedeeg ‘(Q96T ‘stuo"q) ‘sy AreurD *(996T ‘qUIODIF) quIogoyy Aq
pougop se ‘elpeijsmy U1o}soMY INOS “(LS6L ‘310qP9H) B1oqpeH{ Aq pouyop se eIOR ouidyvory ‘(7967 ‘sinqieA\ pur “uUnny],
‘weydeD) so[sy YsHug ‘(LO6T ‘O10 pure ysneqses}g) PIUIsIA ISOA\ ‘(QQ6T ‘OYOPION PuL YWUIG) eruLoFT[LD :Pypp fo saounog

v 61 6'S II Be 99 che b6 Zuo], OBS

688 SC cal 61 Sv ve $I bus[SH 4S

Gil Cr el Ge DEC 902 v9 odour

é 89 vc 67 SOL 667 col SNe]

es One v1 6T LS GL 8S “s] Sesonbiv]

LS) ve ol Ie Al 06 th ‘s] Zopuvuso,y uen{

(SFT) (9's) (0021)

v6 8°02 08 DiC 6zLT ONC €8 SCENE NUE

6 0b 3's 07 SC 98¢ 681 Hts) ‘s] sodedyyeg

Ocl €8 81 Ly 978 89+ OOT Od Opubuiod

€°€S 9°01 ge 1's 978 L6€ 8L ‘sy AreueD

Srl Os Gal One Sia OO 8E voqeuny,
SpUup]s! IIUDIIC)

(20°06) Vsti 6 Ly 0097 799 ItI BIMOps[ED MON
PUuPjsi ]P}UAULJUOI P1O

(+06) Z'6¢ Gil oS 988¢ €hS 66 BI[LSMY UW19}sOMIINOS

é 08 Se Cae 6LZ cll SE eloy ourdjvoayy
SPOID AAI] PUPS]

0 VGll LC oP 9991 [19 eel So[sy SUE
Puprysi [PD] UAL LLOL)

0 al lis 0's Or02 099 Tél BIUIBITA JSOA\

0'8e T8¢ v's TL 6729S OeOl Sb PIUTOFTES)
SPAAD ]PJUAULJUOLT)

% Aquup snuas Aqiup} sa120qS D1aUua‘) SOU]ILUD J
wmsiuapuy . aq lag Lag
sa19adS $a19aqd§ D1QUa‘)

_AJISIOAIp,, pur se1oy wuodsorsuy
Sn ce caLenVaL

Adaptive Radiation in Islands: Definitions and Comments

insular areas is less than for angiosperms or insects. Land-snail faunas
would appear similar to angiosperm floras except that drier areas, such as
Western Australia, are less favorable for land snails than for angiosperms.
Mammals and reptiles are so poor, or even absent, on many islands that
they could not be used in a table comparable to table 3.1.

The factors for “depauperation” vary from one island group to an-
other. There is no doubt that size is the most important, but not exclu-
sively, by any means. One can note that the Canary Islands and Fernando
Po are identical (or virtually so) in number of species. The degree of isola-
tion does not seem drastically different: Fernando Po is 35 km from the
nearest point on the African mainland, the Canary Islands 100 km at the
least. The area of the Canary Islands (7500 sq. km) is 3.25 times that of
Fernando Po (2000 sq. km), and the altitudinal range of the Canaries
(3700 m) a little greater than that of Fernando Po (3000 m). However,
the nearest source area for the Canaries is very dry, and many floristic
elements of the Canaries have been acquired from areas much farther
away. Thus, isolation for the Canaries is greater than appears at first
glance. Also, the Cananan climate is much drier than that of Fernando
Po, which supports a true rain forest. The Canary Islands have appreciably
fewer families (78) than Fernando Po (100). The equity in species num-
bers has been evolved by autochthonous speciation on the Canaries, and
endemic genera (12) have developed there as well. Endemism does not
correlate reliably with diversity however. The California flora has an excep-
tionally high rate of speciation (38.1 per family) but its endemism (38.1
percent), although notably high (especially considering that the area is
defined politically rather than phytogeographically) is not nearly as high
as that of the Hawaiian Islands, New Caledonia, St. Helena, Juan Fer-
nandez, the Canary Islands, or the Marquesas.

Endemism is also very high in southwestern Australia, as defined by
McComb (1966). Southwestern Australia is much richer in speciation
than Western Australia as a whole; the remainder of that state is floristi-
cally poor; the figure of 39.2 species per family, although the highest in
the table, will probably turn out to be too low as southwestern Australia
becomes better known. If one extracts figures from Beard (1967), the S/F
ratio for southwestern Australia becomes 40.5, and that surely does not
include a number of recently discovered species. No reliable estimates of

113

Adaptive Radiation in Islands: Definitions and Comments

endemism in southwestern Australia are available, because the precise
limits of southwestern Australia as a province and ranges of species within
and beyond it are problematical. About 40 percent of the genera of south-
western Australia are endemic.

Endemism, like diversity, is controlled by various factors, chiefly area,
ecological breadth, and isolation. I very much doubt that any mathemat-
ical formula could predict endemism well, both because islands with
high—or low—rates of endemism fall into those categories for different
reasons, and because some factors, such as age, geology, and edaphic con-
ditions are very important but difficult or impossible to measure. Mac-
Arthur and Wilson (1967, p. 74) present, in fact, not a formula but a
graph comparing area to percentage of endemic species of birds for three
kinds of islands: solitary, well-isolated islands, such as New Caledonia;
single islands near mainland areas, such as ‘Tasmania; and islands in the
Gulf of Guinea. If one had graphed angiosperms instead, quite different
results would have been obtained. MacArthur and Wilson state that

the accumulation on islands of species generated within the archi-
pelago provides the best measure of what is loosely referred to in the
literature as adaptive radiation. Specifically, adaptive radiation takes
place as species, when generated within archipelagos, disperse between
islands and, most importantly, accumulate on individual islands to
form diversified associations of sympatric species. In equilibrial biotas,
then, the following prediction is possible: adaptive radiation will in-
crease with distance from the major source region and, after correc-
tions for area and climate, reach a maximum on archipelagos and
large islands located in a circular zone close to the outermost dispersal
range of the taxon.

Their definition of the conditions under which adaptive radiation occurs
seems among the better ones. However, where vascular plants are con-
cerned, sympatry seems the reverse of what has happened: in plants, radia-
tion into different ecological niches has been accompanied by an almost
inescapable allopatry. Sympatry would be difficult to find in any of the
Hawaiian species of Euphorbia, a fine example of adaptive radiation, and
it is infrequent in the Dubautia-Argyroxiphium-Wilkesia complex, pro-
vided that by “sympatry’” we mean (as logically should be the case) par-

114

Adaptive Radiation in Islands: Definitions and Comments

ticular geographical ranges on an island or islands rather than the total
island as the unit on which sympatry is based. The stress MacArthur and
Wilson place on remoteness of a large island or an archipelago is well
taken.

Endemism does appear to be a better reflection of these factors and
what I would term adaptive radiation than do most other measures. Pro-
grams of adaptive radiation (as I would apply the term) are fulfilled in
remarkably different ways in various localities. Southwestern Australia has
been included in this book because the conditions under which adaptive
radiation has occurred there are highly unusual, yet exemplary of the con-
cept. I would say that endemism of at least 50 percent on the specific level
is a threshold for adaptive radiation in flowering plants on oceanic islands,
old continental islands, and perhaps even some islandlike situations. If we
could better define phytogeographically areas like California, the endemism
level might be 50 percent on certain continental areas (adaptive radiation
is certainly present in typically California genera, but the geographic range
of the products often extends beyond California or is much restricted
within it). If we consider a threshold level of 50 percent endemism as an
indicator of likely adaptive radiation, however, we must subtract small
islands (like St. Helena) and archipelagos with relatively small total area
(the Marquesas). The Hawaiian Islands are spectacular in adaptive radia-
tion not only of angiosperms, but also of insects, land birds, and even
ferns. However, adaptive radiation may be present in New Caledonia (see
chapter 7), but certainly not in the same way or degree as in the Hawaiian
Islands—and New Caledonia cannot be said to show adaptive radiation
well in birds.

The S/F figure (table 3.1) proves very interesting with regard to
adaptive radiation. If we take as a threshold value about 8.5 species per
family for oceanic islands and old continental islands, and perhaps 30
species per family for continental areas and islandlike areas, we have above
those levels all the islands and areas where adaptive radiation can be said
to be represented well. This corresponds to the “species-rich-but-family-
poor’ concept cited earlier. A different threshold value on continental
areas would be justified because, with larger areas, more numerous in-
stances of vicarious species could be expected, and, more importantly,
several invasions of genera are probable, rather than just a few, because

115

Adaptive Radiation in Islands: Definitions and Comments

dispersal within a continent is less difficult. The S/G ratio and the G/F
ratio are very poor at predicting where we will find adaptive radiation as
defined by the instances cited in the chapters that follow. Obviously, in
dealing with a flexible concept and with many distinctive types of island
areas, precision is virtually impossible, but the descriptive materials are
intended to highlight modes of speciation and evolution on insular areas
and to call attention to pertinent factors and results, not quantify them.

References

Amadon, D. 1950. The Hawaiian honeycreepers (Aves, Drepaniidae). Bull. Amer.
Mus. Natur. Hist. 95(4):1-262.

Annandale, N. 1918. Fauna of the Inlé Lake. Rec. Indian Mus. 14:1—214.

Baker, J. G. 1887. Flora of Mauritius and the Seychelles. London, L. Reeve & Co.

Baldwin, P. H. 1953. Annual cycle, environment, and evolution in the Hawaiian
honeycreepers (Aves: Drepaniidae). Univ. Calif. Publ. Zool. 52:285-398.

Balgooy, M. M. J. van. 1969. A study on the diversity of island floras. Blumea
17:139-78.

Beard, J. S., ed. 1967. Descriptive catalogue of Western Australian plants. Perth,
W.A., King’s Park Board.

Bock, W. J. 1970. Microevolutionary sequences as a fundamental concept in macro-
evolutionary modes. Evolution 24:704—22.

Bowman, R. I. 1961. Morphological differentiation and adaptation in the Gala-
pagos finchs. Uniy. Calif. Publ. Zool. 58:1-302.

Brown, W. L., Jr., and E. O. Wilson. 1956. Character displacement. Syst. Zool.
5:49-64.

Carlquist, S., 1962. Trematolobelia: Seed dispersal; anatomy of seeds and fruits.
Pacific Sci. 16:126—34.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Clapham, A. R., T. G. Tutin, and E. F. Warburg. 1962. Flora of the British Isles.
Cambridge, Cambridge University Press.

Darwin, C. 1859. On the origin of species by means of natural selection. Reprint
of first edition, 1950. London, Watts & Co.

Exell, A. W. 1944. Catalogue of the vascular plants of S. Tomé (with Principe and
Annobon). London, British Museum (Natural History).

Grant, P. R. 1966. Ecological compatibility of bird species on islands. Amer.
Natur. 100:451-62.

Grant, V. 1963. The origin of adaptations. New York and London, Columbia Uni-
versity Press.

Hedberg, O. 1957. Afroalpine vascular plants. Symb. Bot. Upsal. 15(1):1-411.

Hillebrand, W. 1888. Flora of the Hawaiian Islands. London, Williams & Norgate.

116

Adaptive Radiation in Islands: Definitions and Comments

Keast, A. 1961. Bird speciation on the Australian Continent. Bull. Mus. Comp.
Zool. (Harvard) 123:305—495.

Keast, A. 1968. Competitive interaction and the evolution of ecological niches as
illustrated by the Australian honeyeater genus Melithreptus (Meliphagidae). Evo-
lution 22:762-84.

Keast, A. 1971. Adaptive evolution and shifts in niche occupation in island birds.
In Adaptive aspects of insular evolution, ed. W. L. Stern, pp. 39-53. Pullman,
Wash., Washington State University Press.

Lack, D. 1944. Ecological aspects of species formation in some passerine birds.
Ibis 86:260-86.

Lack, D. 1947. Darwin’s finches. Cambridge, Cambridge University Press.

Lems, C. 1960. Floristic botany of the Canary Islands. Sarracenia 5:1-94.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography.
Princeton, N.J., Princeton University Press.

McComb, J. A. 1966. The sex forms of species in the flora of the south-west of
Western Australia. Austral. Jour. Bot. 14:303-16.

Mayr, E. 1963. Animal species and evolution. Cambridge, Mass., Harvard University
Press.

Melliss, J. C. 1875. St. Helena. London, L. Reeve & Co.

Myers, G. S. 1960. The endemic fish fauna of Lake Lanao, and the evolution of higher
taxonomic categories. Evolution 14:323-33.

Prashad, B., and D. D. Mukerji. 1929. The fish of the Indawygi Lake and the
streams of the Myitkina District (Upper Burma). Rec. Indian Mus. 31:161-223.

St. John, H. 1966. Monograph of the genus Cyrtandra (Gesneriaceae) on Oahu.
Bishop Mus. Bull. 229:1—465.

Schoener, T. W. 1967. The ecological significance of sexual dimorphism in size in
the lizard Anolis conspersus. Science 155:474-77.

Schoener, T. W. 1968. The Anolis lizards of Bimini: Resource partitioning in a
complex fauna. Ecology 49:704-26.

Schoener, T. W., and G. C. Gorman. 1968. Some niche differences in three Lesser
Antillean lizards of the genus Anolis. Ecology 49:819-30.

Sharell, R. 1966. The tuatara, lizards and frogs of New Zealand. London, Collins.

Simberloff, D. S. 1970. Taxonomic diversity of island biotas. Evolution 24:23—47.

Skottsberg, C. 1956. Derivation of the flora and fauna of Juan Fernandez and Easter
Island. Natur. Hist. Juan Fernandez Easter I. 1:193-438.

Smith, G. L., and A. M. Noldeke. 1960. A statistical report on A California Flora.
Leafl. West. Bot. 9:117—27.

Stebbins, G. L., Jr. 1967. Adaptive radiation and trends of evolution in higher plants.
In Evolutionary biology, ed. T. Dobzhansky, M. K. Hecht, and W. C. Steere,
vol. I, pp. 101-42. New York, Appleton-Century-Crofts.

Stewart, A. 1911. A botanical survey of the Galapagos Islands. Proc. Calif. Acad.
Sci., 4th ser., 1:7—288.

Strasbaugh, P. P., and E. L. Core. 1964. Flora of West Virginia. Introduction. West
Virginia Bull., ser. 65, 3(1) :i-xxxi.

ny,

Adaptive Radiation in Islands: Definitions and Comments

Strid, A. 1970. Studies in the Aegean flora. XVI. Biosystematics of the Nigella
arvensis complex with special reference to the problem of non-adaptive radiation.
Opera Bot. 28:1-168.

Thorne, R. F. 1965. Floristic relationships of New Caledonia. Univ. Iowa Stud.
Natur. Hist. 20(7):1—-14.

Williams, E. E. 1969. The ecology of colonization as seen in the zoogeography of
anoline lizards on small islands. Quart. Rey. Biol. 44:345-89.

Zimmerman, E. C. 1948. Insects of Hawati, vol. 1: Introduction. Honolulu, Uni-
versity of Hawaii Press.

118

ADAPTIVE RADIATION:
HAWAIIAN ISLANDS

Numerical and Geographical Aspects of
Adaptive Radiation

The ecological diversity of the Hawaiian Islands has been described and
illustrated by a number of authors (e.g., Zimmerman, 1948a; Carson et al.,
1970; Carlquist, 1970a). Also, these and other authors have cited the
remarkably clear geological progression in age of the islands from west to
east (oldest, Kure and Midway Islands; youngest, Hawaii). There is little
need for a summary of these factors, which the reader can readily obtain
elsewhere. The discussion below will refer to them in a self-explanatory
manner. Assuming these factors, how have they acted on various groups
of organisms? What are the types of speciation we find, what are the eco-
logical adaptations, and what morphological diversification has occurred?

FLOWERING PLANTS

Table 4.1 lists, by island, the numbers of native (but not necessarily
endemic) species for several of the larger genera of flowering plants in the
Hawaiian Islands. These include the nine largest genera of flowering plants
in the Hawaiian flora, with the exception of Cyrtandra (Gesneriaceae).
In the baccate lobelioids, two smaller genera, Delissea and Rollandia, have
been included because they are very close to the larger genera Cyanea and
Clermontia. | consider Dubautia to be a single genus, but species referred
to Dubautia (sensu stricto) and to Railliardia are also listed separately for
convenience. Distribution data are for the species level, and distribution

119

-(g961 ‘ouoIS) Pajaq pur “(EEGT ‘EYOUNA) PuUoddag ‘(LSC6T ‘y10yS)

‘LS61 “6161 OU) SPIONEqo] a3e09kq

SL
0< OF
OOT
Ol

91

0
€
0c
LI
Or

(amefooyey) ¢ 9

(3)
(

8
1)
6
8

SPpuD]S1 LOY IOC 11DMD EH

IZ SL
Ib SZ
BE bZI
b 8]
tI 1Z
i Il
¢ OL
0 0
I I
Z 0Z
¢ Ol
9 I¢
¢ Ol
(Z) (Z1)
(1) (Z)
¢ bl
} 6
1DUD'T NPI

cb 8S
097 +09
6S OL
9 ZZ
0Z ZI
¢ I

S L

0 Il
0 ¢

8 Ol
9 $
tI 87
6 L
(Z) (¢)
(1) (¢)
¢ 9

6 0Z

1DYO]OJ nYyPO

audsouayg pur pisazsozjdy ‘(9S6T “OWA +7961
“(LC6[ ‘Y1oysS) suapig :pypp fo saoinog

‘(qS¢61 ‘“BIOYS) PJapyr0drT pur PiNPgnC

ON

1pnDy

d
cL
8
0

npyuN

suuzodsorsuv uvtmemeyzy Jo soroads yo wornqystc

‘Tp aIave

(ur ‘uerpout) [[eyUIeY]
(tur ‘bs) vore pues]
s[P}0,T,
DO}9d
avIoLINY
piuuo1ada
ovooviodt g
IUABOUATS
DIGAISO]AY J
IVIOLIULL'T
DIPUP]IOY]
DASSTIOAC]
paurvaAd
PIJUOWLA]D
({e}0}) avoprorpaqo"y
-svoorpnurdurd
pjapyaoqr'T
DIp{DIPIO
(0J91498 *s) pIynpgqnd
(07D) *s) PIENPgnC
suaplg
IVIOVIOSY

SnUudas PUD A]MUD |

Adaptive Radiation: Hawaiian Islands

of subspecies and varieties is included within species. The taxonomic
judgments of the authors cited have been accepted except in the lobelioids,
where names likely to be synonyms (see Carlquist, 1970a) have been
subtracted.

Viewed simply in numerical terms, the table shows several interesting
features. First, the island of Hawaii is numerically much too species-poor
(see total of species for all genera); this island would not correspond with
species-prediction equations as discussed by MacArthur and Wilson
(1967). All ecological factors on Hawaii favor speciation, and the “impov-
erished” flora seems due to the island’s newness. The geological evidence
now available shows that very little of this island is pre-Pleistocene. ‘The
appreciable speciation that has occurred on Hawaii in a relatively short
period of time reveals the effect of the rich ecology and probably also the
effect of other factors. Perhaps the isolating effect of several high moun-
tains and the innumerable successive lava flows has been effective. Eco-
logical factors such as altitude and rainfall relate closely to area in the
Hawauan Islands, so a relatively high rate of speciation is not unexpected.
One might not have predicted that Maui would be species-rich if newness
is a factor, but West Maui is much older geologically than East Maui, and
the combination of two large mountains of differing age on a single island,
together with great ecological diversity, obviously provides a variety of
niches; Maui could, with justification, be regarded and computed as two
islands rather than one. As the most ancient of the five larger islands,
Kauai might be expected to have a greater number of species than for-
mulae based on all of the islands would predict. However, remoteness is
an operative factor: the other four large islands, plus Lanai and Kahoo-
lawe, form a fairly tight group, whereas the relative isolation of Kauai
renders it a poor recipient or target for events of interisland dispersal.
There seems little doubt that Kauai has been a major “source area” be-
cause of its age, as appears clear in the case of the Hawaiian drosophilids
(Carson et al., 1970; see also fig. 4.1). With respect to angiosperms, Kauai
is slightly richer in genera compared with other islands in the chain. How-
ever, we would expect newer islands to be centers of speciation. Larger
genera in the Hawaiian Islands probably have experienced recent specia-
tion, and owe their size to lack of extinction.

If the island of Hawaii is too new for angiosperm speciation in pro-

121

Adaptive Radiation: Hawaiian Islands

portion to its size, the other islands are not. However, we find some inter-
esting patterns if we look at individual genera. The largest angiosperm
genus in the Islands is Cyrtandra. St. John (1966), in his monograph on
the Oahu species, recognizes 118 that are endemic to that island and
claims 169 for the Hawaiian chain as a whole. This would indicate that 41
occur exclusively on islands other than Oahu. Perhaps the number of non-
Oahu species might be higher if St. John had studied all Hawaiian Cyrtan-
dras. In any case, the preponderance of species obviously occurs on Oahu,
despite the fact that it is the third largest island and only slightly larger
than Kauai. Moreover, most of the Oahu species are restricted to the Koo-
lau Mts. The basis for this pattern is to be sought in the ecological prefer-
ences of Cyrtandra and the nature of the Koolau Mts. Cyrtandra is char-
acteristic of deep, wet gulches, although some species grow in other areas

0 50 100 miles
a
HAWAII
FIG. 4.1. Summary of the proposed number (22) of interisland founders in the picture-winged
Drosophilas of Hawaii. From Carson et al. (1970), in Essays in evolution and genetics in honor

of Theodosius Dobzhansky, ed. M. K. Hecht and W. C. Steere, © Meredith Corp. Repro-
duced by permission of Appleton-Century-Crofts, Educational Div., Meredith Corp.

122

Adaptive Radiation: Hawaiian Islands

of wet rain forest. The Koolau Mts. form a series of such gulches, which
not only offer numerous localities suitable for Cyrtandra but also are iso-
lated from each other by barriers of relatively dry, sunny ridges, so that an
isolating mechanism favorable to speciation in this genus is replicated
many times. The explosive speciation of the land shell Achatinella in the
Koolau Mts. is a significantly parallel example. :

If we look at the genera in table 4.1, we find that Bidens also has
speciated predominantly on Oahu. Bidens shows very nearly the inverse
of the ecological requirements of Cyrtandra, for it tends to grow on
ridges. Thus, for Bidens, the gulches are barriers that provide an isolating
mechanism.

Pelea seems clearly species-rich on the two older major islands, Kauai
and Oahu. I suspect that this woody genus speciates at a rate relatively
slower than those of the other genera shown in ‘Table 4.1, which are either
herbs (Peperomia, Phyllostegia, Stenogyne) or shrubby (occasionally arbo-
rescent) representatives of predominantly herbaceous families. One gen-
erally expects more rapid speciation in herbaceous groups than in woody
ones; the number of herbaceous individuals that can occupy a given area
is greater and the life cycle is briefer. None of the truly arboreal genera
in the Hawaiian flora has speciated to anything like the extent of the herb
or woody-herb groups listed in the table. That size of individuals is related
to speciation seems clear, and the limited size of island areas also may
well be related to it. Certainly species richness is directly related to area
and to body size of individuals. For example, Necker Island supports 37
species of insects, 5 species of flowering plants (2 small shrubs, 3 herbs),
] species of land snail, and no land birds. This example illustrates survival
rather than speciation, to be sure, but is indicative of the factors involved.
If speciation is slower in woody groups, the older islands would indeed be
expected to bear more species of Pelea. | am making the assumption that
Pelea reached the Hawaiian Islands—perhaps arriving on Kauai—early: its
greatest diversity, as well as the greatest number of species, is on Kauai.
The other genus with a high representation of species on Kauai is Dubau-
tia—especially Dubautia in the narrow sense. The paleate heads of Du-
bautia s.s. are more primitive than those of Railliardia, which have bare
receptacles. Dubautia s.s. not only has more species on Kauai, but the
species also are more diversified there. Maui appears to be the center of

123

Adaptive Radiation: Hawatian Islands

diversification for the Railliardia species, and the species on Hawaii are
either the same as those on Maui or vicarious species close to those on
Maui.

Maui is the center of species richness for several of the genera in table
4.1. There is an interesting concentration of Peperomia species on Molokai,
Maui, and Lanai. The number of species on Molokai and Maui is higher
than one would expect. Kauai 1s relatively poor in Peperomia species,
despite a high median rainfall that ought to favor Peperomia. This situation
suggests the possibility that Peperomia arrivals and establishments occurred
first on the eastern islands, in which case Kauai would have served as a
recipient island.

Kauai and Oahu are the only islands on which all four genera of bac-
cate lobelioids occur. However, Clermontia and Cyanea are much more
numerous in species on the eastern islands. Delissea, the only baccate
genus represented on the westernmost and probably oldest major island,
Nuhau, may retain the most numerous primitive characteristics in the
baccate lobelioids. Cyanea, close to Delissea, is well represented on Kauai.
Cyanea most likely is underrepresented on Oahu because Rollandia, prob-
ably a close derivative of Cyanea-like ancestors, has radiated on Oahu and
may have forestalled radiation of Cyanea and perhaps also Clermontia
there. Cyanea-like ancestors probably are the source of Clermontia, which
differs chiefly in its growth form and its few-flowered (usually 2-flowered)
inflorescences. The Clermontia species most like Cyanea are C. haleaka-
lensis (Maui), C. multiflora (Maui, Oahu), C. samuelit (Maui), and C.
wailauensis (Molokai)—all with several-flowered inflorescences. The tran-
sition from Cyanea to Clermontia may well have occurred on Maui, or
possibly Oahu. If this is true, it might explain the large number of Cler-
montia species on the island of Hawaii. Cyanea and Clermontia are wet-
forest genera, and the abundance of Clermontia species in the broad areas
of wet forest on the island of Hawaii is understandable. Kauai is poor in
Clermontia species (only one, C. clermontioides, occurs there), which sug-
gests that it has been a recipient island: C. clermontioides is close to C.
kakeana and C. oblongifolia of Oahu, from which stock it may have
originated.

Morphological aspects of radiation in the lobelioids are discussed later.

124

Adaptive Radiation: Hawaiian Islands

INSECTS
Interestingly, the geographical distribution of species of major insect
groups parallels that of angiosperms (table 4.2). Data are from Zimmerman
(1948b) and Hardy (1964, 1965). The chief differences can be seen in the
greater degree of speciation on the newer (eastern) islands. In view of the
angiosperm data, one might not have expected this finding: in angiosperm
species tabulated, Oahu and Maui exceed Hawaii, whereas in insect spe-
cies, Hawaii exceeds all other islands. Why have insect species been
formed more readily on the island of Hawaii? (Some of the species are not
endemic to that island, but of course most are endemic to the island.) The
most likely answer seems to be that insects speciate more rapidly than
angiosperms. In view of their much shorter life cycles and the fact that
insect populations can be small (in area) and localized geographically, this
is understandable. One might claim that insects disperse within an archi-
pelago more readily than angiosperms, but this is probably not a factor in
the present examples: otherwise, we would find fewer instances of one-
island endemism, whereas the proportion is, in fact, much the same for
insects and angiosperms. Moreover, although virtually all the genera of
angiosperms listed in table 4.1 have reached the island of Hawaii, they

TABLE 4.2.
Distribution of certain Hawaiian insect species, by island

Molo- Other
Family and genus Nihau Kauai Oahu kai Maui Lanai Hawaii islands
Drosophilidae (all
native species) 0 62 92 81 128 13 125 0
Dolichopodidae
(total) 0 (Z0)ae a( p>) (Bl) (B72) (6) (42 a0
Campsicnemus 0 i 3 25 30 5 35 0
Eurynogaster 0 13 19 8 7 ] 7 0
Lygaeidae
Neseis 0 2 6 7 10 3 [1 0
Nysius 0 7 8 UI 8 We alike
(both genera) 0 Oy ie (Ga) = (ls) (7) (22\ee (ley)
Totals 0 101 161 126 183 26 189 Il
Island area (sq. mi) WZ 555 1004, 4X0) WARS 141 4030

Rainfall (median,

in.) ? 99 58 42 ip) Zi 75

2 Species of Nysius on smaller islands are as follows: Kahoolawe, 2; Nihoa, 2; Necker,
2; French Frigate Shoals, 2; Lisianski, 1; Pearl and Hermes Reef, 2.

125

Adaptive Radiation: Hawatian Islands

evidently have not speciated there in proportion to the land area. Even
with insects, although the total of species by islands in table 4.2 is highest
on Hawau, the number (189) is by no means commensurate with the much
greater area of that island.

More rapid speciation of insects seems favored by conditions that do
not serve as good isolating mechanisms for angiosperms. A kipuka (forest
area surrounded by new lava flows, like an island), for example, might be
too small to harbor a genetically viable population of an angiosperm spe-
cies, but it might contain literally thousands of individuals of an insect
species.

The insect examples listed in table 4.2 are of interest because the
dipterous families (Drosophilidae, Dolichopodidae) are primarily adapted
to wet forest, whereas the heteropterous family (Lygaeidae) is character-
istic of dry localities. The ability of Nysius to colonize low, dry islands in
the Leeward chain is illustrative in this regard. Both dipterous and heter-
opterous groups in table 4.2 show similar distribution of species with re-
spect to islands, however: all three families are well represented on Hawau,
although the totals for Maui are very similar. Recentness of the island of
Hawaii must be the reason why, as with angiosperms, species number is
not proportional to area. This gives us a dramatic example of how current
mathematical formulae, which must of necessity omit factors of geology
and climate, cannot be used in particular instances. Admittedly, the island
of Hawaii is unusual in being both very large and very recent, compared
with volcanic islands at large. Its ecological diversity, however, is entirely
commensurate to its size, so we cannot cite lack of available habitats as a
reason for lack of species—although possibly its broad, relatively noneroded
surfaces do not yet offer as many isolating mechanisms as would ndges and
valleys on older islands.

In Drosophilidae, we have accumulating evidence that Kauai has
served as the staging area for invasions of the newer islands. The map pre-
sented by Carson et al. (1970) and reproduced here as figure 4.1, together
with the supporting data of those authors, provides abundant evidence in
support of this view. Systematists who have considered problems of inter-
island distribution might have hypothesized this result and brought certain
morphological lines of evidence to bear (e.g., the Dubautia-Railliardia ex-

126

Adaptive Radiation: Hawaiian Islands

ample cited above), but the evidence from cytology of the drosophilids is
conclusive. One notes from the map that there is by no means a perfect
west-to-east linear sequence, however: Kauai to Maui, followed by subse-
quent invasion of the intervening island, Oahu, is the case for some of the
species groups.

Dispersal events of the larger, expanding groups of insects in the Ha-
watian Islands are probably relatively recent. The occurrence of Nysius
very likely represents recent colonization, rather than relictualism from a
period when these islands were large, high islands. There is a parallel in
angiosperms and birds. The occurrence on Laysan of the recently (ca. 1923)
extinguished honeyeater Himatione saguinea freethi and the extant honey-
creeper (“Laysan finch”) Psittarostra cantans, together with the now-ex-
tinct sandalwood and Pritchardia palm, represent relatively recent dispersal
to Laysan from the high islands.

LAND SNAILS

The two major groups showing interesting speciation patterns on the
Hawaiian Islands are the achatinellids (Tornatellinidae, subfamily Acha-
tinellinae; also considered a family, Achatinellidae) and amastrids (Cochli-
copidae, subfamily Amastrinae; also considered a family, Amastridae).

The achatinellids are absent from Kauai, and the Molokai-Maui com-
plex of islands is considered a likely place of origin, according to Zimmer-
man (1948a, based on information from C. M. Cooke and Y. Kondo).
According to this hypothesis, a Partulina-like ancestor, perhaps an Achati-
nella, colonized that island group, then Oahu. On Oahu it radiated spec-
tacularly into numerous species (for a visual summary, see Carlquist,
1970a). As mentioned earlier, Achatinella is probably parallel to the angio-
sperm genus Cyrtandra in its speciation, and for much the same reasons.
Perkins (1913) lays stress on the tendency of Achatinella, according to his
observations, to form colonies on particular trees or groups of trees and to
stray little from characteristic host plants. With this low mobility, one
would expect that the ridges and valleys (habitats different in their host
plants) would indeed serve as effective isolating mechanisms.

Partulina has radiated on Maui and Molokai, and a Partulina complex
is also found on Oahu, indicating a second invasion of Oahu (the first is

127

Adaptive Radiation: Hawatian Islands

considered to have given rise to Achatinella, which, in its strict sense, is
endemic to Oahu). Other achatinellid genera include Perdicella (Maui
and Molokai: eight species) and Newcombia (Molokai and Maui: nine
species) .~

The amastrids, on the other hand, are centered on the older islands.
The genus Carelia (20 species), now extinct on Niihau and found in sub-
fossil form there, is restricted in the living state to Kauai, and even there
the genus is decadent, with some species extinct or known only from sub-
fossil shells. Amastra has radiated to all the major islands. The distribution
of described (but not necessarily valid) species, according to Thwing (1907)
is as follows: Kauai, 10; Oahu, 52; Molokai, 14; Maui, 17; Lanai, 12: Ha-
wail, 6. The Kauai species show a notable diversity, including the gigantic
A. knudsenti, which, by virtue of its striate columellar plait, is said to
show affinity to Carelia (Thwing, 1907). The genera with remarkably flat
shells, Pterodiscus and Planamastra, are also centered on the older islands.
Laminella (described species, not necessarily all valid) is distributed as fol-
lows: Oahu, 6; Molokai, 4; Maui, 4; Lanai, 3.

LAND BIRDS

Because the Hawaiian honeycreepers provide such superior examples
of adaptive radiation, one might expect this excellence to be reflected in
geographical distribution of species and genera. There is, to be sure, a
character-displacement effect: congeners on a single island tend to differ
in bill morphology (this will be discussed later). However, geographical
sequences in genera and species of honeycreepers, showing progressively
greater specialization, do not seem to occur. The representation of Psitta-
rostra, whose morphology suggests an end product of evolution, is heavy
on the island of Hawaii (P. bailleui, P. palmeri, P. flaviceps, P. kona); P.
psittacea is present on the six major islands, while P. cantans has reached
small western islands, Laysan and Nihoa, probably in recent time. Perhaps
the new adaptive mode of Psittarostra permitted it to invade both eastern
and western islands successfully. Loxops and Himatione, the most “primi-
tive” genera in their respective subfamilies, occur on all major islands of
the chain. Ciridops is known only from the island of Hawai. These and
other patterns (data from Amadon, 1950) suggest no geographical se-

128

Adaptive Radiation: Hawaiian Islands

quences except at the infrageneric level. Perhaps, then, the honeycreepers
have dispersed easily from island to island into all suitable niches. The
recentness of adaptive radiation in this group is difficult to estimate. Bock
(1970) concludes that radiation may have required as little as one million
years, although he concedes that ten million years have been available for
it. Although dispersal of land birds to the Hawaiian Islands has been difh-
cult, redispersal within the chain has evidently been comparatively easily
accomplished. The pattern of drepanidid distribution parallels that of
Lygaeidae among insects. To be sure, the rarity and early extinction of
some of the Hawaiian honeycreepers may have left gaps in our geographical
distribution; Ciridops might have occurred on islands other than the island
of Hawaii, for example. However, there is no reason to assume that any of
the honeycreepers ranged much more widely than our records of them in-
dicate (see Amadon, 1950). More difficult to explain are some of the other
bird distributions within the Islands. The occurrence of the Hawaiian
crow (Corvus tropicus) only in one area on the island of Hawaii might be
the result of a relatively recent establishment.

Morphological and Ecological Aspects of Hawaiian
Adaptive Radiation

Each major group of plants or animals has undergone different patterns
of adaptive radiation in the Hawaiian Islands, although there are a few
striking parallels. Some of these patterns have been studied with relative
thoroughness, as in the Hawaiian honeycreepers, for which, however, field
observations have been relatively few until recently. For such groups, I can
only offer syntheses of what has already been offered by various authors.
Other groups, such as the lobelioids, are known to taxonomists, but the
data have been expressed very little in evolutionary terms. For these, litera-
ture and original field observations can be synthesized. In still other groups,
such as certain of the land shells and insects, more thorough examinations
of ecology and other factors relating to evolution are needed or are under
way at present. For some rare, endangered, and extinct groups, definitive
data may never be available, and we will only be able to speculate about
modes and ranges of adaptation.

129

Adaptive Radiation: Hawatian Islands

FLOWERING PLANTS

Lobelioids

The Hawaiian lobelioids form an interesting series of examples of
adaptive radiation, because each genus demonstrates distinctive modes of
speciation.

Delissea. This genus, suggested by Rock (1919) and Wimmer (1956)
to be ancient and “decadent” (as Hawaiian plants go), may be similar to
the ancestral stock of the baccate lobelioids. It is distinctively a genus of
lower elevations or drier upland localities. The extreme is probably repre-
sented by the Niuhau population, named D. ntihauensis by St. John (1959).
Delissea niihauensis, now extinct, was evidently a stem succulent. One
would guess that when Niihau’s eastern ridge was populated by native
dry forest and scrub plants, such as Reynoldsia, Lipochaeta, and Erythrina,
that D. niihauensis grew in somewhat protected places and resembled to a
moderate degree the endemic stem-succulent lobelioid Brighamia. The
native locality for D. niihauensis could not have been more than several
hundred feet in altitude, for Nuhau reaches a maximum altitude of 1281
ft; it may have occupied the lowest elevation of any lobelioid except
Brighamia. Another altitudinal extreme is represented by D. undulata,
whose only known current locality is Puu Lehua, approximately 5000 ft
elevation, between Hualalai and Mauna Loa, on Hawau. This may be the
last extant population of this Delissea (Carlquist, 1970a). The Puu Lehua
locality, characterized by small cinder cones, is definitely dry forest: D. un-
dulata has unbranched stems 18-90 cm tall, and grows among Acacia koa
trees under which a few native shrubby genera, such as Lipochaeta, per-
sist. In 1966 I found plants broken by cattle, yet still alive, indicating that
this species of Delissea also has qualities of a stem succulent.

Other species of Delissea apparently occurred in localities at low eleva-
tions but probably less xeric than those of the two preceding species. Delis-
sea subcordata is distinctive in being a branched shrub, but its broad leaves
and known localities suggest a more mesic adaptation. It apparently is an
understory element (it may still be extant) in low but moist gulches on
Oahu. We do know D. rhytidosperma from elevations between 2000 and
3000 ft in the Waimea Canyon on Kauai, which would place it in rela-
tively dry Acacia koa forest or Styphelia-Dodonaea scrub. Its narrow leaves
certainly suggest a more xeromorphic adaptation than do the broad ones of

130

Adaptive Radiation: Hawaiian Islands

D. subcordata. Delissea rhytidosperma can be called a branched shrub or
small tree.

Cyanea. As mentioned earlier, Cyanea, which has lateral inflores-
cences indefinite in flower number, may represent a stock close to ancestral
stocks of the baccate Hawaiian lobelioids. If ecological shift into wet forest
represents a divergence from original adaptation, as it seems to in many
Hawaiian genera (e.g., Bidens), Cyanea could be considered a wet-forest
derivative of Delissea-like ancestors. Virtually all species of Cyanea occupy
wet forest; the range is probably between as little as 50 in. annual rainfall
in leeward koa forests for species such as C. angustifolia (Mt. Tantalus,
Manoa Trail) or C. leptostegia, in tall koa forest of the Kokee region of
Kauai. Even these koa forests are transitional to Metrosideros forests and
are usually a mixture of the two trees. In the most mesic localities for
Cyanea, rainfall may be as high as 400 in. annually, as on the northern
slopes of East Maui (Olinda Trail: C. aculeatiflora, C. hamatiflora, etc.)
or on the upper slopes of Puu Kukui, West Maui (C. atra). Some species
of Cyanea occupy deep shady gulches where rainfall may be less, but avail-
able moisture is great and evaporation low (e.g., C. solenocalyx on upper
Molokai). Although this range of moisture is broad, all the environments
could be described as mesic. The nature of adaptation in Cyanea that
is visible in terms of morphology is not so much to rainfall belts but to
different modes of vegetative structure that represent alternative ways of
occupying wet-forest areas.

The range in growth forms in Cyanea can be characterized according
to the scheme of figure 4.2. The smallest and simplest are exemplified by
C. pilosa and closely related species (C. bishopi, C. longipedunculata, C.
acuminata). These species, mostly placed in sect. Pilosae of Cyanea, typi-
cally are unbranched rosette plants with upright stems of limited length,
not much more than | m in height, with most individuals lower. They grow
as forest-floor elements in wet Metrosideros forest; leaves are broad and
thin, but hairy. In that respect, as well as in their small white flowers and
habitat preference, they mimic, in a generalized way, some of the Hawai-
ian species of Cyrtandra. The Cyaneas of sect. Pilosae may well be reduc-
tions not only in plant size but also in flower size. In any case, some un-
branched or sparsely branched plants of moderate stature probably were
basic to the taller rosette trees of Cyanea.

131

aA FIG. 4.2. Selected species of the endemic Ha-
ASAIZ waiian genus Cyanea (Campanulaceae, Lobelioi-

NS deae) to show the wide range in habit, leaf size,
; leaf morphology, leaf vesture, and leaf prickliness.
Both a juvenile leaf and an adult leaf are shown

fr
}
c:

\
f \f \

'

x
Somes
BD

vA

oe
VAG
\!

Ty y
\ XY
NAY >

Nw. Lunt \| Tals

CAN

\
dé

(

ANE ATH PPT
\/\

A

2 V\ / ps
NONGy: »

\

SATATS

TREN WAHT THEE TANNIN UIT YY
\ \ \
\

Se

—t

"C. nolimetangere’ ‘“C. rollandioides“ _—C. tritomantha C. marksii C. grimesiana

dre
C. shipmanii

132

for C. leptostegia. One typical leaf is shown for
all other plants. The alleged species ‘““C. nolime-
tangere” and “C. rollandioides’ are considered
to be juvenile stages of C. tritomantha. Separate
scales are given for habit and for leaves.

A
EE Phe 4
A Ff ed
GAL, t
vn & FB i
Et / e ZS we,
BE j iy "
, ee oA fl By
i V7 Z
wey V x ay
Sz ; we
a ie
ag.
—= f 4
4
=~ 4
ZEN £
= 5
4
} B
~ [@) tae
C. pilosa C. gayana C. leptostegia

133

Adaptive Radiation: Hawaiian Islands

A Cyanea of intermediate stature would resemble a Delissea. Spe-
cies from Kauai, such as C. gayana (fig. 4.2), C. recta, C. fissa, and C. hir-
tella, would fill this description. These are rosette herbs, somewhat woody
at their-bases, with larger leaves. While always understory elements in wet
forest, they grow to 2m or more. The leaves are thin and broad, suggesting
shaded conditions. Inflorescences are short and congested. Visiting birds
would probably perch on main stems while feeding on the flowers; or-
nithophily is suggested by size of flowers (over 2 cm long) and color (white
with purplish guidelines).

Cyanea angustifolia (fig. 4.2) and segregate or related species (C. fau-
riei, C. sylvestris) remind me both of the preceding species of Cyanea and
of Delissea. The C. angustifolia group, however, typically grows in some-
what drier situations, such as relatively low elevations of the leeward Koo-
lau Mts. behind Honolulu, where C. angustifolia can readily be seen. De-
gree of branching varies with age of plant, type of forest, and perhaps
injury to the main stem. Most individuals eventually branch in a foun-
tainlike fashion, however. An unbranched plant of C. angustifolia, with
pendant clusters of flowers of moderate size, certainly recalls species of
Delissea, especially when one considers that C. angustifolia favors some-
what drier habitats than those of other Cyanea species. The koa-ohia
forest where C. angustifolia grows is dry only by comparison to other Cya-
nea habitats. The branching habit seems related to the sunnier, almost
scrublike conditions where C. angustifolia grows. It occurs at or near the
summits of ridges where koa and ohia (Metrosideros) trees are smaller, and
where more sunlight would penetrate than in lower valleys.

Although all of the abovementioned species of Cyanea share some
aspects of a Delissea-like ancestry, those described below seem to me to
represent departures from that ancestry clearly beyond anything in Delis-
sea, and peculiar to the genus Cyanea. The most striking of these tenden-
cies is formation of giant rosette trees, represented most conspicuously
by Cyanea leptostegia (fig. 4.2) of Kauai. Similar species include C. carl-
soni (Hawaii), C. regina (Oahu), C. superba (Oahu), C. gibsoni (Hawaii),
and C. arborea (Maui). These species tend to be unbranched (with few ex-
ceptions), to grow in moderately wet forest, and to reach heights of 14 m.
The leaves are long, strap-shaped, and sessile. The height of these plants
seems related to height of surrounding forest. Cyanea leptostegia, for ex-

134

Adaptive Radiation: Hawaiian Islands

ample, grows in the relatively tall koa forest of Kokee, Kauai. The single
rosette tends, therefore, to reach better-illuminated levels of the forest.
Leaves of these species are leathery, have strong midribs, and seem better
adapted to wind action and desiccation than do those of the preceding
groups of Cyanea species. The large size of the leaves may be related to the
fact that these tall plants bear only a single cluster of leaves. A juvenile
plant of C. leptostegia has a relatively narrow stem and lobate, thin leaves,
a foliar dimorphism indicating adaptation of juvenile leaves to lower illu-
mination near the forest floor. Young plants have long internodes and nar-
row stems, eventually widened by production of a broader woody cylinder.
An adult plant of C. leptostegia has very short internodes at the level of
the leaf rosette. At the same time, the pith is very wide near the rosette,
so that the total width of the stem is nearly as great as that of the base of
the plant, which has a wider woody cylinder but a narrow pith. The upper
stem has a xylem-cylinder configuration maximizing strength, and because
of its circumference is suited to bearing more numerous leaves. If pith
(and therefore stem) diameter did not increase upwardly, too few leaves to
support a large plant might be produced, and the stem might fracture
more easily above. The secondary xylem of the upper stem (Carlquist,
1969b) does show greater mechanical strength, by virtue of thicker-walled
libriform fibers, than does the wood at the base of the plant. Inflorescences
are not long and pendant, but borne on short peduncles, among leaves.
This suggests that birds visting the flowers would perch not on the stems
but on the strong, sessile leaves, which could withstand such usage by
virtue of their stout midribs, whereas in C. angustifolia, visiting birds
would perch on the stem just below the long, pendant flower clusters and
probe upward into the inflorescences.

A series of Cyanea species that bear spiny or highly prickly leaves
and stems in the juvenile condition, and moderately prickly or merely
hirsute leaves in the adult condition has been typically grouped with the
C. leptostegia group. These prickly species—represented on Oahu and,
to a much greater extent, on the islands to the east—represent unique
innovations among Hawaiian lobelioids. They seem to be basically un-
branched rosette herbs or rosette trees of intermediate height, characteristic
of very wet forest, and may not have been derived at all from the notably
glabrous C. leptostegia group. A typical and critical example of the prickly

135

Adaptive Radiation: Hawaiian Islands

rosette-tree Cyaneas is represented by C. tritomantha (fig. 4.2), from wet
ohia forests inland from Hilo. Although a rosette tree to 5 m, it is
undoubtedly an understory element, for it grows in tall ohia forests with
Cibotium tree ferns. The leaves of adult plants of C. tritomantha are
exceptionally long and broad, and are borne on thick, succulent petioles
that may serve as perch sites for birds visiting the flowers, which are large,
and white or slightly purplish. Leaves are only moderately prickly or
slightly hirsute; internodes are very short, and if one saw only an adult
plant, one could perhaps group this species with the C. leptostegia group.
However, this would be a mistake, as I have suggested earlier (1970a).
The juvenile stages of C. tritomantha were apparently undescribed or
disregarded until my field studies in 1966. At that time, I attempted to
find a series of Cyanea species reported from the same area as C.
tritomantha: C. nolimetangere, C. rollandioides, and C. bryanii, for exam-
ple. When I did locate them, I discovered them to be, in all likelihood,
merely various stages in the development of C. tritomantha. That they
could have been named as species results only from the fact that these
juvenile stages can bear flowers. Seedlings of C. tritomantha bear broad,
highly prickly, thin leaves and have long internodes. Shoots corresponding
to these seedlings were also observed as sucker shoots, or shoots in response
to injury, on adult C. tritomantha plants. These prickly seedlings were
the basis for C. nolimetangere Rock. Very possibly they may stay in this
juvenile stage for a long period of time (they do bear flowers in this
condition), and perhaps in some cases never are able to mature into adult
plants. In clearings near the virgin Metrosideros forest where I saw C.
tritomantha, I noticed plants with longer, less prickly leaves. These plants
bore flowers and tended to branch somewhat, in this sunnier habitat. Such
individuals correspond well to what has been named C. rollandioides and
perhaps to other species, such as C. fernaldii. Minor differences in calyx
morphology may account for naming of several species that essentially
represent an intermediate stage in the development of C. tritomantha.
In the same general area, Cyanea plants to 3 m, still with relatively long
internodes, and with leaves intermediate in size and vesture between those
of “C. rollandioides” and those of C. tritomantha, were seen. These plants
have been named C. bryanii by Rock. The flowers of “C. bryanii” are
within the range of variability of C. tritomantha. If one only had herbarium

136

Adaptive Radiation: Hawaiian Islands

sheets of these stages available for study, one could, understandably, name
several species, but field studies dictate otherwise. To my knowledge, there
are no studies that follow the development of Cyanea species from seedling
to adult plant.

The situation described opens possibilities that other named prickly-
leaved Cyaneas may represent juvenile stages, but unraveling this complex
situation to its fullest extent will not prove easy for several reasons.
(1) Some of the named prickly Cyaneas are very scarce and may even be
extinct, and field studies will be difficult or impossible; although we can
suspect that some are juvenile stages of others, we cannot be positive.
(2) There is variability in floral morphology within Cyanea species, and a
juvenile stage with flowers slightly different from those of an adult plant
may cause some to support segregation. If only slight floral differences
separate them, however, one should be suspicious: calyx-lobe length and
morphology are notoriously variable in Cyanea, and might even vary with
degree of adulthood of the plant. (3) There probably are prickly species
that are permanently juvenile. One clear example seems to be Cyanea
marksii (fig. 4.2). This plant grows in the Honaunau forests of Kona, island
of Hawai, well removed from any tree-Cyaneas that could conceivably
be related to it. It takes the form of a miniature rosette tree to 2 m, very
prickly on stems and leaves. To be sure, it grows near (but not in) forest
openings, but no “versions” of this plant corresponding to an adult form
have ever been found in the Honaunau forests (although C. stictophylla,
in this region, should be reexamined to be sure there is no relationship to
C. marksii). Cyanea marksti may well be a permanently juvenile species.

The prickly species of Cyanea on Maui are poorly known, and similar
considerations may apply to them. Cyanea ferox may be a permanently
juvenile species. However, taxonomic designation of plants with differing
degrees of prickliness as varieties (Cyanea ferox var. horrida) without field
study seems inadvisable—these are probably different degrees of juvenilism.
Broad-leaved Cyaneas with sinuate leaves may well be juvenile stages (just
as sinuate leaves are juvenile in C. leptostegia). One notes in this regard
plants named Cyanea atra var. lobata—very likely juvenile stages of C. atra,
a species similar in habit, when adult, to C. tritomantha. I am suspicious
of certain taxa, collected very infrequently, that are named on the basis
of sinuate leaves, such as C. quercifolia or C. solanacea. They should be

137

Adaptive Radiation: Hawaiian Islands

compared with C. holophylla to see if they are developmental stages or
perhaps permanently juvenile species. Other prickly species from Maui,
such as C. aculeatiflora (with distinctive prickly flowers and fruits) and
C. hamatiflora are undoubtedly valid species based on adult plants, but
we do not know what the younger stages look like. The possibility remains
that juvenile stages capable of flowering might have been named as distinct
species.

The interesting feature of the prickly Cyaneas is not the taxonomic
confusion but the apparent adaptations to various conditions they seem
to represent, by virtue of marked or slight polymorphism from juvenile
to adult. A seedling of a palmiform Cyanea grows under conditions quite
different from those of an adult plant. Greater prickliness on seedlings
may represent protection from phytophagous insects—or even molluscs
(Carlquist, 1962), although this is as yet entirely speculative. Prickliness
that disappears as a plant becomes adult suggests this possibility (as in
the giant Euphorbia, E. candelabrum), but the paucity of herbivores in
the native Hawaiian fauna complicates this speculation—especially since
prickliness occurs only in some of the species of Cyanea (and some
Rollandias), not in other lobelioids of the world.

Permanent juvenilism of a sort (but other factors as well) may be
involved in a trio of interesting species: C. asplenifolia, C. grimesiana,
and C. shipmani (fig. 4.2). Of these, C. asplenifolia (Maui) and C.
grimesiana (Maui, Molokai, Oahu, Lanai, Hawaii) agree in having pinnate
leaves, not unlike those of C. ferox. Juvenile leaves of C. grimesiana, as
I have observed in the field, are pinnatifid but also markedly prickly,
showing that this group of species is probably related to the palmiform
group. The same is true of C. shipmanii, known only from Metrosideros
forests at 6000 ft on Mauna Kea. This species is more striking, however,
in that the leaves are not merely pinnate but bipinnatifid, nearly bipinnate.
The ecological adaptation of these three species is perhaps not what one
might expect. Pinnate leaves suggest a shady forest-floor habitat. However,
the pinnate-leaved Cyaneas occupy an even more apppropriate niche: deep
shady gulches, which receive less light than would an equivalent area of
flat rain forest. The stems of these species do become elongate, up to
about 2.5 m in height, which places them in better-lit portions of ravines,
but still in deep shade.

Adaptive Radiation: Hawaiian Islands

Thus, Cyanea has undergone an amazing degree of adaptive radiation
with respect to habit. Because Cyanea, like Delissea, Rollandia, and
Clermontia, is an endemic genus, we can be certain that this radiation
is autochthonous. Flower colors vary in Cyanea (white, white streaked
with purple, purple, rose, pink). In some species (C. atra, C. marksti, and
C. aculeatiflora), corollas are deep purple outside, white within. Corolla
and calyx-lobe shapes differ among the species. Corollas and calyces of
C. aculeatiflora are densely covered with prickles (Carlquist, 1962). These
floral variations alone constitute an unexpected diversity. Different pollina-
tion agents may be involved in these floral types; certainly different
positioning of pollinating birds must be involved during their feeding
activities on the various Cyaneas. Even supposing most species to be
ornithophilous, we are faced with the question, discussed later in the case
of Clermontia, why there should be diversity in floral morphology and
color.

Rollandia. Rollandia is a genus endemic to Oahu, with the single
exception of R. parvifolia Forbes, an apparently valid Rollandia collected
only once at its type locality on Kauai. Rollandia on Oahu has shown a
surprising range in habit and habitat (fig. 4.3). Although all Rollandias,
like all Cyaneas, could be called rosette herbs (or rosette trees), this brief
description covers a diversity of expressions and ecological correlations.
Some wet-forest species, such as R. crispa, have broad leaves, tomentose
to various degrees. Rollandia crispa is branched or unbranched, to 1.3 m
in height; R. calycina, which has longer racemes, is similar but taller, to
2 m. While most Rollandias can be called understory species, R. crispa
and R. calycina are forest-floor species with rose flowers. They remind
me of Cyrtandras in their habit. This is even more striking in R. hum-
boldtiana, a species 1-2 m tall, often unbranched, with long, drooping
racemes of white flowers. Rollandia humboldtiana tends to grow in very
muddy, wet, shady localities. Another arborescent species, to 3 m, is
R. lanceolata. This species occurs in moderately moist situations (Acacia
koa forests) and forms chiefly unbranched rosette “trees.” The corolla
color of R. lanceolata varies from pale green with purplish stripes to pale
red or deep purple. The relatively hairy leaves of R. lanceolata seem
adapted to its drier habitat; their broad shape, however, suggests shadiness,
as do leaves of R. crispa, R. humboldtiana, and R. calycina. Hairy leaves

139

(juvenile)

C ca FF ia, 2 R, ~
=
Re, Ce Ge Ge

SS aa

a.

RS

FIc. 4.3. Selected species of the endemic Hawaiian genus Rollandia (Campanulaceae, Lobe-
lioideae). This genus on Oahu (only one non-Oahu species is known) shows a great range in
habit, habitat, leaf size, and leaf vesture. Rollandia pinnatifida is from the Waianae Mts.; the
remaining species are from the Koolau Mts. Drawings are placed approximately in relation to
characteristic altitudes for these species, and background for each species illustrates the typical
associated vegetation. Scales same as those in fig. 4.2.

EGG OU

[Bitiatasizee

R. saintjohnii

Nee A PUNO he
eS

<= KOOLAU MOUNTAINS=——
: aa XS a

> =e

Fic. 4.3. Selected specie 2

lioidese). This c REESE of the endemic Hawaiian genus Rollandia (Campanulaceae, Lobe-

REDIE Habitat. lect oo ae (only one non-Oahu species is known) shows a great anne in

mente cine speniceoare anc ler ete Rollandia pinnatifida is from the Waianae Mts; the

characteristic altitudes feathes ae ad ae As pera aa at relation Eo
i Py = “se species, and backgro 29 ee 4 ane

| associated vegetation. Scales same as those in fig, apes for each species illustrates the typical

paisa oe

nee

ha REE Aron

ee a

R. saintjohnii

(juvenile) (adult)

AG a oite,
ee al

4 ee SS

WAIANAE MOUNTAINS — ‘fs \KOOLAU uONTANS Se ;
; ‘ 2 : ? ae Kar = ce
Cd A ir og

“io Ba x

\

Adaptive Radiation: Hawaiian Islands

occur in a population that was named R. waianeensis by St. John. This
plant may be conspecific with R. lanceolata var. tomentella, as Wimmer
(1956) suggests, but not, as Wimmer alternatively suggests, a hybrid
between R. lanceolata and R. calycina.

Two species recognized on the basis of glabrous leaves and calyx
characters, R. purpurellifolia and R. longifolia, reach 5 m and 3.3 m,
respectively. Thus R. purpurellifolia is comparable in habit to R. crispa,
R. longifolia to R. lanceolata.

An interesting tree-Rollandia that evidently has been long misunder-
stood is one called R. kaalae by Wawra, although the probable earliest
name for it is R. pinnatifida G. Don.; its synonymy includes Cyanea
selachicauda Degener. This plant grows deep in moist gulches of the other-
wise moderately dry Waianae Mts. (fig. 4.3). Juvenile foliage, markedly
sinuate, is produced for many years, and the plant can flower during this
juvenile period. Adult leaves ultimately are lanceolate, like those of R.
lanceolata or R. crispa. Lobate juvenile leaves also occur in R. longiflora
(Rock, 1919). Their significance seems the same as that of sinuate leaves
cited in the discussion of Cyanea. Occurrence of juvenile leaves of this
description in Cyanea and Rollandia makes uncertain the taxonomic status
of Delissea sinuata Hbd. and D. laciniata Hbd.; both species are known
from type specimens only. However, the occurrence of lobate leaves in
these three genera is more importantly regarded as three independent
adaptations to conditions of low light intensity.

Narrow, glabrous leaves characterize R. angustifolia (fig. 4.3). Al-
though this species is not very tall, it grows in scrubby ridge forest and
is likely to occur there as an emergent rosette plant of nearly the same
height as the shrubby species of Coprosma, Syzygium, and others with
which it grows. The shiny, thickish, but narrow leaves of R. angustifolia
are probably suited to the higher light intensity the crown of this plant
receives, on cloudy ridges of the leeward Koolau Mts. Rollandia angustt-
folia has purple or purplish-veined flowers.

An extreme in adaptation, perhaps not related to R. angustifolia but
reminiscent of it in reduction of leaf surface, is R. saintjohni. This species
occurs in the low, scrubby, krummbholz-like vegetation at the summit
ridges of the Koolau range. Constant pressure from trade winds that sweep -
with tremendous force up the steep windward cliffs (pali) of the Koolau

142

Adaptive Radiation: Hawaiian Islands

range result in a low, scrubby vegetation that consists mainly of species
which would attain greater stature in more sheltered localities. However,
R. saintjohnii (fig. 4.3) is adapted to and grows only in such windswept
localities. Its leaves are the thickest in the genus (about three times as thick
as those of R. calycina | Carlquist, 1962]), suggesting condensation of form
related to wind. A hypodermis can be said to be present in leaves of R.
saintjohni (but absent in those of R. calycina), and it has a much thicker
palisade. Prickles are conspicuous on leaves and stems of R. saintjohnii,
a fact which suggests a phase of juvenilism (Carlquist, 1962). As one might
expect, prickles of this prominence occur elsewhere in the genus Rollandia
only on the juvenile leaves of R. pinnatifida. Both the prickles and the
lobate nature of juvenile leaves underline afhnity between Rollandia and
Cyanea, separated on the basis of whether the staminal tube is united to
the corolla (Rollandia) or not.

Clermontia. As mentioned earlier, Clermontia seems to be a derivative
of Cyanea or Delissea in which number of flowers is reduced. Clermontia
is easily recognized, however, by its shrubby, much-branched habit. This
habit (present in all except Clermontia haleakalensis, which looks like a
Cyanea vegetatively) is unlike that of other Hawaiian lobelioids. There is
a certain range in Clermontia with respect to size. At margins of open
bog areas on Puu Kukui, Clermontia micrantha forms only a small shrub.
Clermontia arborescens and C. hawaitiensis can be treelike and attain
heights of more than 6 m. Most species grade imperceptibly between
these extremes.

Epiphytism characterizes a few species. Clermontia peleana is exclu-
sively epiphytic; C. clermontioides and C. parviflora are often seen as
epiphytes, although by no means invariably so. Other species of Clermon-
tia occur only occasionally as epiphytes. In range of habit, Clermontia is
not exceptional. Moreover, the size and shape of leaves are relatively un-
varied throughout the genus.

Clermontia is remarkable, however, in the diversity of floral mor-
phology, color, and size. This variation has been illustrated elsewhere
(Rock, 1919; Carlquist, 1970a) and does not need repetition here. I am
puzzled by the reasons for this diversity (as well as the floral diversity in
Rollandia and Cyanea). The nature of pollination biology would be the
possibility most readily sought by the evolutionary biologist. These appar-

143

Adaptive Radiation: Hawaiian Islands

ently ornithophilous flowers are known to be visited by the nectar-feeding
honeycreepers (Amadon, 1950; Spieth, 1966). The three species of meli-
phagids (all but Moho braccatus of Kauai are apparently now extinct
[Richardson and Bowles, 1964]) may have visited lobelioid flowers. Among
the honeycreepers, Vestiaria (one species, widespread in the islands, still
extant) and some of the species of Hemignathus (e.g., H. procerus, fig.
4.11) might be or have been effective in lobelioid pollination. The two
species of Drepanis were definitely nectar-feeders, and might have been
very effective in lobelioid pollination. However, the number of species of
bird pollinators does not exceed 10 at the most, and perhaps is as few as
seven, of which most are known only from a single island each, and per-
haps only a limited area on each island. However, there are a large number
of endemic lobelioid species. Among the baccate genera Wimmer (1956)
recognizes Delissea, 8; Cyanea, 58; Rollandia, 12; and Clermontia, 27.
Additionally, among nonbaccate Hawaiian lobelioids there are Trematolo-
belia (one species with three subspecies), Brighamia (one species, possibly
with several geographical races), and the nonendemic genus Lobelia,
which has 10 endemic Hawauan species. I regard Wimmer’s species con-
cept as too narrow and would recognize about 10 percent fewer species.
However, supposing that there are about 105 valid lobelioid species in
the Hawaiian Islands, all of them endemic, how are we to explain the
diversity in the floral patterns they show? The number of species of bird
pollinators obviously does not at all match the number of types of lobe-
hoid flowers.

There are several possibilities. Do (or did) honeycreepers and meli-
phagid individuals show flower-constancy for particular lobelioid species?
As many as three or four species of lobelioids can be found in a single area;
for example, Clermontia arborescens, C. grandiflora, and C. kakeana all
occur at Haelaau, Puu Kukui, Maui. Among observations to date, only
one, that of Perkins (1903) gives us any indication that flower-constancy
might be operative. Writing of the feeding habits of Hemignathus, he
says, “The two common species [H. lucidus and H. procerus| are both
partial to the nectar of some kinds of arborescent Lobeliaceae . . . espe-
cially to those with large corollas, whereas to some kinds they pay no
attention at all, however profusely they may be in flower.” On this slim
evidence one might make a case for flower-constancy, but only as a starting

144

Adaptive Radiation: Hawaiian Islands

point for further observations. Although flower-constancy is not expected
in birds as much as in certain groups of insects, it certainly remains a pos-
sibility.

There is relatively little separation among species of the lobelioid
genera with respect to flowering time, so temporal isolation cannot be in-
voked as a speciation mechanism. Almost all the Clermontia species, for
example, can be found in flower in July and August, and some species
flower over a period of perhaps 6 months. Some of the most abundant spe-
cies have greenish flowers (e.g., C. kakeana), which would seem least attrac-
tive to potential pollinators. One could cite a tendency for lobelioid spe-
cies of deep shade to have white, more conspicuous flowers, species of
better-lit situations to have colored flowers, but there are too many excep-
tions to credit such a suggestion seriously. Even if the correlation were
perfect, it would only explain one dimension of the diversity in floral color,
size, and morphology in these genera.

Other factors one might cite include ecological shift, genetic drift,
redispersal of variants, and multiplicity of appeal to potential pollinators.
Although some of the baccate lobelioids are sympatric, small differences
in slope exposure, moisture, and shade tolerance separate the habitats of
“sympatric” species. Unfortunately, ecological data are not readily avail-
able in monographs of the lobelioids, owing to paucity of information on
herbarium labels. However, field botanists familiar with lobelioids do know
characteristic preferences for the various species. Ecological shift is basic
to adaptive radiation in all major Hawaiian angiosperm genera, and the
lobelioids are no exception, as some of the distinctions listed suggest. Dur-
ing ecological shift, one can easily envision changes in growth form and
size, but changes in floral morphology and color may have accompanied
them. One could invoke genetic drift as an explanation for minor changes,
such as color differences. If variants are dispersed to new sites within the
islands, geographical isolation would provide a mechanism permitting
genetic drift to operate. Certainly the Hawaiian Islands are rich in such
isolating mechanisms.

Lobelioids can hybridize, and natural hybrids are known especially in
Clermontia (Carlquist, 1970a). Probably hybrids were more frequent when
pollinating birds were more abundant. That hybrids can be found at all,
in view of paucity of pollinating birds, is rather surprising. Lobelioids can

145

Adaptive Radiation: Hawatian Islands

self-pollinate and probably owe their continued presence in the Hawaiian
flora to this mechanism, although one can imagine that a high degree of
selfing might eventually be deleterious. If, however, hybrids were once fre-
quent, one can imagine that a source of variability to supply distinctive
new populations has existed. However, one must still explain development
of the extremes which can yield hybrids; the cited factors collectively do
not seem sufficient. One must not underestimate the possibility that color
and shape of lobelioid flowers, within certain limits, may not be adaptive,
just as Amadon (1950) and Baldwin (1953) have claimed that plumage
colors in the honeycreepers have no apparent adaptive significance.

One may, as a last resort, hypothesize diversity in floral size and mor-
phology as a way of presenting a “‘diversified appeal” to particular pollina-
tors, just as on a given mainland area of California there are hummingbird
flowers in a number of different families and genera. Development of
diversity of ornithophilous flowers among Hawaiian lobelioids would then
depend on two factors. First, one must assume a “disharmony” in the
Hawaiian ornithophilous floras—in other words, not all the potential ap-
peals to birds were originally present, because some groups (e.g., Bignonia-
ceae) are absent. This would explain why Hibiscadelphus and Kokia,
endemic genera of Malvaceae, evolved curved tubular flowers autoch-
thonously in the Islands. This is true also in Geranium arboreum, which I
have seen visited by Himatione sanguinea (Polipoli Rd, Haleakala). One
can imagine even that there might be character displacement, so that spe-
cies of Clermontia in a given area differ in appeal by virtue of size and
color—and that if a second species enters the range of a Clermontia or

*Cyanea, it must, as discussed in chapter 3, exploit a different range of
appeals to pollinators. It is also possible that minor differences in curva-
ture, size, and display of flowers within the lobelioids might offer precise
and different pollinating mechanisms, providing isolating factors. At any
rate, it seems evident that the lobelioids now represent several stocks, each
of which, by sheer probability, is capable of a different range of possibili-
ties, rather than all stocks trending in the same direction.

Lobelia. Although Trematolobelia shows only subspecific differentia-
tion related to geography, the Hawaiian species of Lobelia show radiation
into diversified situations. They could all be described as characteristic of
openings in wet forest. However, these “openings” are quite varied. For

146

Adaptive Radiation: Hawaiian Islands

L. gaudichaudii (sensu lato), the openings are open bogs. For L. tortuosa,
which has purple flowers, unlike the white (or pinkish) flowers of L. gaudi-
chaudu, the habitat is dripping cliffs near waterfalls on Kauai. Blue-flow-
ered species (L. hillebrandu, L. hypocleuca, L. oahuensis, L. yuccoides)
favor scrub in very moist, cloud-swept localities, often on steep slopes on
various islands. Lobelia oahuensis can be observed on road cuts on Mt.
Kaala.

Dubautia-Argyroxiphium-W ilkesia complex

These three genera (four, to authors who recognize Railliardia) prob-
ably arose from a single ancestral stock. They form undoubtedly the most
outstanding example of adaptive radiation among Hawaiian angiosperms.
The habits and habitats of species in this complex have been illustrated
rather extensively elsewhere (Carlquist, 1970a). Leaf anatomy of Argyroxt-
phium, Dubautia, and Wilkesia has been explored (Carlquist, 1957,
1959a), as has wood anatomy (Carlquist, 1958) and anatomy of the heads
(Carlquist, 1959b). Although some references are made to ecology in these
papers, a more explicit series of correlations can be presented.

There is little doubt that the primary adaptation of this complex, like
that of its mainland madiinean relatives, is rather xeric. There seems a
likelihood that the tarweeds (Asteraceae: Heliantheae, subtribe Madiinae)
were more widespread in late Tertiary times in California and Mexico.
They probably included shrubby forms, like Adenothamnus validus (north-
ern Baja California) and the fruticose species of Hemizonia on the offshore
Californian islands, or Raillardella scabrida (Lake Co., northern Califor-
nia). These appear truly like relicts, although the annual tarweeds are now
abundant and widespread in central western California.

The Hawaiian immigrants probably diverged early into the rosette
plants Argyroxiphium-Wilkesia and the shrubby Dubautia, although hy-
brids between the two groups are evidently still possible, for A. sandwich-
ense * D. menziesii hybrids have been reported and others may occur
(see chapter 13).

Argyroxiphium-W ilkesia. The silvery-leaved species of Argyroxiphium
are probably specialized in leaf anatomy. Even with regard to the indument
alone, they are exceptional in the subtribe, and only such mainland species
as Raillardella argentea, itself an alpine specialization, offer parallels. ‘The

Ay

Adaptive Radiation: Hawatian Islands

green-leaved species (A. grayanum, A. virescens, Wilkesia gymnoxiphium)
resemble ancestral types more closely. Wilkesia has less complicated leaves
(the three superposed series of bundles in Argyroxiphium leaves are un-
doubtedly a specialization [Carlquist, 1957]), but has some seemingly spe-
cialized characteristics, such as the verticillate leaves (which may be a
permanent juvenilism of sorts: production of verticillate leaves can be seen
in seedling stages of all Argyroxiphiums, and occasionally in entire shoots
of A. grayanum). Wilkesia is an upland plant but favors dry, open sites at
about 2000 ft and above, on Kauai. Argyroxiphium is clearly alpine, and
probably does not occur below 4000 ft. Argyroxiphium virescens, in high,
dry to somewhat moist, but cloudswept sites on Haleakala, East Maui,
perhaps represents an early adaptation from dry lowland conditions to dry
alpine conditions—or dry mid-elevation conditions. Argyroxiphium vire-
scens is monocarpic. Evidently the A. virescens stock has given mse to A.
grayanum, a non-monocarpic shrub species from the bogs of Puu Kukui,
West Maui, and certain boggy areas on northwestern Haleakala. It repre-
sents a rather rapid entry into highly mesic (but still montane) conditions.
The silvery-leaved species of Argyroxiphium, however, seem basically
adapted to dry conditions (although in alpine localities where rising cloud
masses may sweep through), and their leaves suggest definite adaptation
to brilliant sunshine and desiccating air. Argyroxiphium sandwichense oc-
curs on Haleakala and once occurred on all three peaks of the island of
Hawau. Argyroxiphium kauense occurs in a peculair “dry cloud forest” at
6000 ft on the southwestern rift zone of Mauna Loa. Although this open
forest locality on relatively new lava seems dry, the misty clouds that sweep
upwards daily permit such bog species as Oreobolus furcatus to occur at
this locality, although most associated species are typically alpine or sub-
alpine. The fact that A. kauense has rosettes raised on woody stems rather
than acaulescent ones, as in A. sandwichense, suggests release from the
acaulescent habit of A. sandwichense, which experiences frost to which a
ground-level rosette would be a form of resistance.

An entry into an even more extremely wet locality is represented by
A. caliginii, a silversword growing in open bog areas of West Maui. Argy-
roxiphium caliginii is almost as “silvery leaved’”’ as A. sandwichense Dut
has adopted an almost apomictic habit in the bogs. It is not monocarpic;
instead, the rosettes branch at the base, and these prostrate shoots root in

148

Adaptive Radiation: Hawaiian Islands

the bog. Very few plants flower, in my experience. Perhaps the lack of suit-
able pollination conditions in the bog has induced this characteristic. Inter-
estingly, heads are pendant in this species, and thus apparently escape wet-
ting by mist and rains. Flowers are dull yellow brown, unlike the brightly
colored flowers (yellow or purple) of the other Argyroxiphiums. Hydathodes,
to permit jettisoning of excess water, have been developed in leaves of
A. caliginii (Carlquist, 1957). That the direction of ecological shift in
Argyroxiphium has been from drier to wetter localities seems clear. Wood
anatomy alone would dictate this conclusion: all species of Argyroxiphium
have the short vessel elements that are characteristic of xeromorphic
plants; the bog species have only slightly longer vessel elements, suggest-
ing a recent shift into the bogs (Carlquist, 1958). Thick leaves, rich in
water-storing pectic accumulations in intercellular spaces, occur in the bog
species, although this pronounced xeromorphic adaptation has no signifi-
cance in the bog habitat (Carlquist, 1957); it simply is not disadvantageous,
and has not been lost. Asteraceae evidently have a capability for rapid eco-
logical shift, and Argyroxiphium demonstrates this dramatically. How can
such a seemingly abrupt shift take place so readily? The answer probably
lies in the fact that despite differences in moisture availability, the high-
elevation bogs are subject to intermittent bright sunshine and nocturnal
cold, features of dry alpine sites. Moreover, the Hawaiian bogs are prob-
ably rather strongly disharmonic (the total bog area in the Islands is not
great, and similar bogs in source areas are remote), so that invasion of bogs
by montane Hawaiian genera is not difficult.

The ancestors of Argyroxiphium and Wilkesia may well not have been
dry lowland species, and elevations no lower than 2000 ft should be hypoth-
esized for ancestors of the Dubautia-Argyroxiphium-Wilkesia complex.
These ecological preferences are likely for all temperate elements in the
Hawaiian flora, especially those probably from North America, for upland
Hawaiian climates simulate (except in day length) lowland maritime Cali-
fornia climates, for example. Interestingly, the Dubautia complex does not
seem to occur below 1000 ft anywhere in the Islands.

Dubautia. As noted in table 4.1, the species of Dubautia, sensu stricto,
are on the older western islands chiefly, while “Railliardia” tends to be on
the eastern islands. The sect. Dubautia species may be considered more
primitive, as stated earlier, because of presence of receptacular bracts.

149

Adaptive Radiation: Hawaiian Islands

However, species of sect. Dubautia are mostly in quite mesic situations.
If the Hawaiian tarweeds originally established in drier sites, the various
extant species of Dubautia s.s. must be hypothesized to have undergone
shift into moister sites, and, with few exceptions (e.g., perhaps D. planta-
ginea), the ancestral stock in drier areas has now vanished. Thus, diver-
gence from the original ancestors into Dubautia s.s. and Railliardia on the
one hand and Argyroxiphium and Wilkesia on the other probably occurred
prior to invasion of wet forest. To be sure, there are hybrids between spe-
cies of sect. Dubautia and sect. Railliardia of Dubautia (Sherff, 1935b). If
divergence of Argyroxiphium-Wilkesia from Dubautia (including Railli-
ardia) took place in drier localities, Argyroxiphium must have retained a
primitive feature in its ray flowers (lacking in Dubautia and Railliardia, as
well as in Wilkesia). The heads of Argyroxiphium and Wilkesia suggest
gigantism, perhaps related to the monocarpic habit, but an expression that
involves loss of dispersibility (see chapter 11). Loss of ray flowers can take
place easily in Asteraceae. Conspicuousness of rays in the silverswords may
relate to alpine pollinators. In any case, Argyroxiphium seems a Maui-cen-
tered complex (four species there) at present.

If Dubautia, sensu stricto, entered the rain forest early, we ought to
have some stages in this shift. Dubautia plantaginea, a representative from
moderately wet forest, has leaves with venation like that of vernal leaves
of mainland tarweeds (Carlquist, 1959a, p. 198). Leaf anatomy shows simi-
lar features in species of Dubautia s.s. that probably represent recent shifts
into the wet forest of the summit plateau of Kauai (D. paleata, D. railli-
ardioides). Lack of alteration in leaf anatomy of these species compared
with that of D. plantaginea may be explained by the fact that although
they grow in zones of higher rainfall, they characterize forest margins
rather than the wet forest itself. Dubautia laxa, D. knudsenti, D. molo-
kaiensis, and D. reticulata show reduction in number of palisade layers,
thinner leaves, and absence of bundle-sheath extensions, all of which may
be related to lower light intensity in dense forests or misty ridges where
these species typically grow. The broad, glabrous, relatively thin leaves of
D. knudsenii typify adaptation to uniformly mesic rain forest. Dubautia
knudsenizt is a true tree.

Leaves of D. latifolia are unique among tarweeds: they have reticulate

150

Adaptive Radiation: Hawaiian Islands

venation, whereas longitudinally oriented veins predominate in all others
(most notably in Wilkesia). In addition, coriaceous leaf texture in this spe-
cies is the result of fibrous bundle-sheath extensions. This odd and nearly
extinct species is unique in the group in its vining habit. It typically grows
in lower, drier areas of the upper Waimea Canyon on Kauai and forms
inflorescences when it reaches the canopy of the forest. Its leaves can only
be called those of a sclerophyllous plant. However, herbarium specimens,
as would be expected, have only leaves adjacent to the inflorescence. Such
leaves would be expected to be less mesic in construction. Leaves lower on
these vining plants may have been quite different, but unless this species
is studied in the field, we may never know.

Dubautia waialealae is notable for representing an adaptation to the
wettest locale in the Hawaiian Islands, the summit area proper of Kauai.
This species has more fibers on bundles in leaves than one might expect,
but this characteristic might be explained by the value of shorter, tougher
leaves in a bog area often buffeted by strong misty winds. Lack of a well-
defined palisade in leaves of D. waialealae is related to the almost perpetu-
ally cloudy conditions of the Waialeale bog.

The Railliardia species of Dubautia are mostly alpine and may repre-
sent an advance to upland dry sites. One may not think of Kilauea, 4000
ft elevation, as an alpine locality. However, many Hawaiian alpine ele-
ments occur there (Vaccinium, Silene). The Kilauea species Dubautia
(Railliardia) ciliolata can also be found at high elevations on the island of
Hawaii, as can another species that occurs at Kilauea, D. (R.) scabra. Du-
bautia menziesii, a Haleakala alpine, is characteristic for this site in its
leaf adaptations. Leaves are short and stiff, and especially thick—a con-
densation in form reminiscent of the condensed leaf form in the Haleakala
silversword. Leaves of D. menziesii tend to be isolateral. The large fibrous
bundle caps in leaves are correlated with stout texture. Chloroplasts are
abundant throughout the leaf, indicating adaptation to high light intensity.
Cuticle is exceptionally thick, indicating minimal water loss in the dry
alpine situation. Air spaces are smaller in this leaf than in leaves of forest
species, and secretory canals are smaller or absent. Of these features, only
the diminished secretory canals do not have an obvious correlation, as
compared with angiosperms at large, to the alpine habitat.

151

Adaptive Radiation: Hawaiian Islands

Invasion of the alpine areas by both Argyroxiphium and the Railli-
ardias is interesting, because the shrubby habit of the latter (perhaps rather
like the ancestral habit of the complex) is quite unlike the rosette habit of
the former. One cannot say which entered the alpine zone first; they may
have shifted into different alpine zones within the chain, then redispersed
so as to be sympatric. The alpine Railliardias and Argyroxiphium point up
the fact that most of the Hawaiian alpine flora is derived from species of
dry lowlands; the same genera, often the same species, tend to appear both
at low dry localities and high dry sites: Styphelia, Dodonaea, Artemisia,
Wikstroemia. There are a few primarily or exclusively alpine genera: Silene,
Geranium, Sisyrinchium, and Vaccinium, for example. However, the Ha-
walian alpine flora seems even more disharmonic than the flora as a whole,
so radiation of lower-elevation and even wet-forest elements (Sadleria, Co-
prosma) into the alpine zones seems to have occurred. This has occurred
in the Canary Islands also (chapter 5).

Since leaf size and leaf shape are sensitive indicators of mesomorphy
or xeromorphy in Dubautia, it is interesting to find that xylem shows these
trends as well (fig. 4.4). Xylem evolution, in general, tends to lag behind
external morphology as an indicator of xeromorphy or mesomorphy, but
in many Asteraceae it has adapted very sensitively to habitat. The average
dimensions of vessel elements, given in figure 4.4, show a most amazing
degree of correlation. The cell diameters probably give a better indication
of available moisture in the habitat of any given species than would even
a very good rainfall map. Areas subject to frequent cloud cover tend to
receive more rain, but they also provide increased moisture availability be-
cause of retarded evaporation. Areas with less cloud cover naturally tend
to bear less vegetation, and the bare soil dries out rapidly, a fact also re-
lated to the porosity of Hawaiian soils. Thus, there are factors which, for
a plant, make a wet Hawaiian forest “wetter” than rainfall figures would
suggest, and dry areas, in effect, “drier.” Ecological gradients are therefore
sharper than one might expect. This may help explain why vessels are so
well correlated. Short, narrow vessels indicate xeromorphy. Vessel group-
ing and density, as seen in transections of woods, are also greater in xeric
species of Dubautia, as one would expect from studies of angiosperm
woods at large (Carlquist, 1958).

152

0

|

D. scabra

D. menziesii
aS
D. ciliolata

1 EE Rs:

D. plantaginea

D. railliardioides

D. laxa ae

D. microcephala

Z2cm.

D. knudsenii

(ae ese (245)

I96 X 47a

(TS eee

223 X 32

penaN meen

229 X 38y

236 X 53y

(373)
301 X 72y

306 X 69y

(4.00)
336 X 64y

413 X 83y (496)

FIG. 4.4. Leaves, in silhouette, and vessel elements of selected species of Dubautia (Asteraceae).
Average adult leaves are shown (scale indicated). Vessel elements are drawn to represent the

average length and width (figures below vessel element in each species). Species are arranged
in order of the figure obtained by adding the average vessel width and vessel-element length
in parentheses at right for each). Data original and from Carlquist (1958).

153

Adaptive Radiation: Hawaiian Islands

Pelea
Pelea (Rutaceae) is a genus nearly endemic to the Hawaiian Islands.

Two species occur in the Marquesas. In turn, the genus is close to Meli-
cope, Evodia (Euodia), and Acronychia of the South Pacific. Independ-—
ently, Stone (1969) prepared a monograph of the genus and I devoted a
section of a book (1970a) to diversity in Pelea. In habit, Pelea ranges
from small shrubs (P. waialealae) to species 16 m tall (P. zahlbruckneri,
P. pallida, P. hawaiensis. Although most Peleas are shrubs of intermediate
height, three species are scandent and could almost be called vines (P.
ovata, P. adscendens, P. anisata). Pelea and its relatives seem basically
adapted to moderately wet forest. If so, a few species (P. pallida, P. sand-
wicensis, P. hawaiensis, P. radiata, and P. grandifolia) may have adapted
to habitats of a slightly drier nature; these species are from the rather dry
koa-sandalwood forest of low (below 2000 ft) elevations. Bog species in-
clude P. waialealae, P. recurvata, P. cruciata, P. parvifolia var. sessilis, and
P. orbiculata. Leaf size tends to be largest in less mesic species, smallest
in the bog species, intermediate in wet-forest species. Pelea provides a
good example of adaptive radiation, although the entire genus is basically
woody, and leaf texture does not vary greatly (however, I would guess that
interesting differences in leaf anatomy with relation to ecology could be
found). Some species have developed cauliflory in the Hawaiian Islands,
whereas others have large open panicles. The adaptive significance of this
variation is unclear.

That the range in habit and habitat of Pelea seems a divergence to
arborescence and to smaller shrubs from shrubs of intermediate size is in-
teresting, for shrubs tend to show less adaptive radiation than genera in the
Hawaiian flora that represent predominantly herbaceous families. More-
over, the Hawaiian Peleas are far more diverse than Melicope, Evodia, and
Acronychia together. This diversity may have developed autochthonously
more than once in the Hawaiian Islands; Stone (1969) recognizes four
sections (two are recognized in earlier treatments), none of which has a
stereotyped habit or habitat range. The Marquesan species do not belong
to sect. Apocarpa, which is presumed to be primitive in having separated
carpels. If this interpretation is accurate, the Marquesan Peleas may be
immigrants from Hawaii rather than intermediates of a group that termi-

154

Adaptive Radiation: Hawaiian Islands

nates in the Hawaiian Islands, as is considered to be the case in most

genera of plants and animals.

Coprosma

Another shrubby genus that has shown appreciable radiation is Co-
prosma (Rubiaceae). This Pacific genus is represented by about 27 species
in the Hawaiian Islands. As Oliver’s (1935) monograph suggests, the black-
fruited C. ernodeoides may have been an establishment independent of
the one that led to the remaining Hawauian species, all of which are orange-
fruited. Among the orange-fruited species, there are distinctive adapta-
tions. By far the majority of Coprosmas are broad-leaved, rain-forest spe-
cies. However, C. ochracea is a distinctive, miniature bog shrub with
coriaceous leaves (Carlquist, 1970a, p. 355). Coprosma montana is a micro-
phyllous species of dry alpine localities, such as the cinder cones of Hale-
akala, where the silversword grows, and at high elevations on the island

of Hawaii.

Cyrtandra

Gesneriaceae seems basically an herbaceous family, as does the gesne-
riad genus Cyrtandra. The family does have, however, the capacity for in-
creased woodiness. Very likely, the Hawaiian species represent a single
stock. One would suspect this from the chromosome number, n = 17,
which is otherwise unusual in the genus. Cyrtandra can be cited, as it was
earlier in this chapter, as a large genus that shows much vicarious specia-
tion in the Hawaiian Islands. However, there is a notable range in habit
among the Hawaiian species. Some can become small trees, to 6 m (for
example, C. gayana, C. kalihii, C. waiolanii). Most are little-branched, up-
right herbs, but some (C. oenobarba) are decumbent. There is some cauli-
flory in Hawaiian species (C. lysiosepala [Carlquist, 1970a, p. 127]). How-
ever, the most curious adaptation is the production of flowers on shoots at
ground level, so that inflorescences invariably trail along the ground sur-
face. This habit, which suggests a distinctive (but unknown) pollination
system, occurs in C. stupantha and C. kaulantha (St. John, 1966). This
habit has probably developed in the two Hawaiian species independently
of the same adaptation in a Borneo species, C. basiflora. Some distinctive
features in Hawaiian Cyrtandras are of unknown adaptive value, if any.
Most Cyrtandras have ellipsoid or fusiform fruits, but C. brevicornuta, C.

155

Adaptive Radiation: Hawaiian Islands

calpidicarpa, and C. cornuta have elongate fruits to 12.5 cm in length (St.
John, 1966). Other features of questionable adaptive significance include
ovaries glabrous in anthesis but hairy at maturity (C. laxiflora, etc.), densely
hairy versus glabrous leaves (species with each type can be found sympat-
rically),-and the various calyx types. Cyrtandra flowers, in the Hawaiian
Islands and elsewhere, may be actinomorphic or zygomorphic. The differ-
ence is not great, and zygomorphy is undoubtedly reversible, according
to Burtt’s data for Gesneriaceae.

Pritchardia

This palm genus probably represents an eastern-Pacific extension of
Livistona. At any rate, there are some low-elevation species in areas of less
than 25 in. annual rainfall, as P. remota on Nihoa (perhaps formerly also
on Laysan), but Pritchardias can occur in areas where rainfall exceeds 250
in. per year. ‘The population termed P. beccariana behind Hilo 1s a tall
tree, reaching forest canopy levels in the unusually tall ohia forest there.
Populations in the rather more scrubby forest of the Koolau Mts., Oahu,
tend to mimic surrounding woody plants in their short trunks; such popu-
lations have been called P. martii, P. macdanielsti, etc. Pritchardia was
probably a widespread lowland genus, as shown by fossils (Carlquist, 1970a,
p. 61), but the lowland climate may have been much wetter in former
times, as during the Pleistocene, when climatic belts were presumably
lower. The range in fruit size in Pritchardia is phenomenal, and is discussed
in chapter 11.

Euphorbia

An excellent example of adaptive radiation, Euphorbia (Euphorbia-
ceae) shows increased stature with ecological shift into successively moister
sites. As such, it is an example of “insular woodiness,” and is discussed in
that connection in chapter 10. Illustrations of the various species and their
habits have been offered earlier (Carlquist, 1970a), and need no repetition
here. However, in the time since that account, a new species, E. haelee-
leana, has been described (Herbst, 1971). This novelty appears to represent
the tallest Euphorbia (14 m) in the Hawaiian Islands, but cannot be cited
as an end-product of increased woodiness in the same line as the other spe-
cies because, according to Herbst, it is not a chamaesycoid Euphorbia, as

156

Adaptive Radiation: Hawaiian Islands

the others all are. The aspects of adaptive radiation in the Hawaiian cha-
maesycoid Euphorbias that relate to wood anatomy have been described
(Carlquist, 1970b), and are notable in that, as with Dubautia, ecologically
interpretable features of wood anatomy in these species parallel degree of
mesomorphy or xeromorphy. Leaf anatomy of the Hawaiian chamaesycoid
Euphorbias, a project currently under investigation, appears strongly re-
lated also; leaves range in texture from succulent in coastal species to
papery, of short duration, in dry forest, to leathery in rain-forest species.

Scaevola

Like Euphorbia, Scaevola (Goodeniaceae) in the Hawaiian Islands
shows a series in increased woodiness. Therefore, comments on this genus
are offered in chapter 10. Illustrations have been given in Carlquist (1970a),
and ecological correlations in wood anatomy have also been offered (Carl-
guist, 1969a). The lowland species, S. coriacea, has succulent leaves—remi-
niscent of the lowland chamaesycoid Euphorbias, although more conspic-
uous—and a sprawling habit. Coraceous leaves characterize most species;
the leaves of the dry-upland S. gaudichaudii are more restricted in size,
however, than those of rain-forest species such as S. gaudichaudiana and
S. chamissoniana. Truly coriaceous leaves occur in the species of wet forest
on older islands, S. glabra, which offers a parallel to the wet-forest Euphor-
bias such as E. rockit.

FERNS

Two fern genera have shown shift into various distinctive localities.
The endemic genus Sadleria (Blechnaceae) might be called a typical rain-
forest fern, but S. cyatheoides pioneers on new lava and can also be found
in alpine sites where it is subject to occasional mists, above 7500 ft in ele-
vation, as on Haleakala. Sadleria squarrosa is a peculiar species found only
on dripping cliffs or other wet, steep surfaces in deep shade.

Amphoradenium (an endemic Hawaiian genus of Polypodiaceae) is
always a fern of wet forest or bog. However, it exploits these habitats in a
variety of ways. The terrestrial species may be small or large, with bipin-
nate or tripinnate leaves. The epiphytes include tiny thin-leaved species,
such as A. hymenophylloides, as well as species with long, drooping leaves,
such as A. pinnatifidum (Carlquist, 1970a).

17

Adaptive Radiation: Hawatian Islands

BIRDS: THE HAWAIIAN HONEYCREEPERS

The Drepanididae are the most frequently cited example of adaptive
radiation in the Hawaiian Islands. To be sure, their fame in this regard is
by no means as great as that of the Darwin’s finches in the Galapagos
Islands. ‘The range in adaptations and its extensiveness are far greater than
those of the Darwin’s finches, and a more spectacular example of adaptive
radiation on oceanic islands would be difficult to find among avifauna. The
lesser publicity of the honeycreepers is related to two main facts. First, ex-
tinction among them has been high, and even of the extant species only a
few are at all common. Therefore, we know little about their habits. A
second problem is that the honeycreepers are, with a few exceptions, birds
of dense, wet forest, and study of them in the wild is difficult. If we knew
about their habits in depth, we would undoubtedly have a heightened
appreciation of various aspects of their adaptive radiation. Such studies
are in progress, as shown by the results of Berger (1972), and soon we may
be able to add some important dimensions to the story we already have
about these birds.

Students of the Hawaiian biota have available for inspection a number
of well-llustrated studies of the drepanidids. These have explored and
graphically expressed in various ways the information we have at present.
The monograph of Amadon (1950) summarizes all previous information,
and offers a classification and nomenclature that have been followed by
subsequent workers. Baldwin (1953), Richardson and Bowles (1964), and
Spieth (1966) have contributed field studies, and Bock (1970) has ex-
pressed the theoretical framework of evolution of this family. Those who
seek illustrations will find the early lithographs the best. Those of Wilson
and Evans (1890-1899) have the merit of representing all species of honey-
creepers. The plates of Rothschild (1893-1900) are elegant and accurately
represent both birds and their associated plants, but they do not include
all honeycreeper species. Acceptable illustrations are offered by Munro
(1960) and Carlquist (1965, 1970a).

Rather than repeat the many details related to this adaptation, I pre-
fer to present the factors on which the radiation is based—what is of prime
importance, what is secondary, and what is not relevant to the main story.

1. All authors stress bill morphology and associated feeding apparatus
(bill musculature, tongue morphology) as visible indicators. The narrow,

158

Adaptive Radiation: Hawaiian Islands

short, straight bill is considered primitive, although which of the present
genera best exemplifies this is not certain. Bock (1970) chooses Ciridops
(presumably a fruit eater) of the subfamily Drepanidinae as representing
more numerous primitive characteristics, whereas Amadon (1950) stresses
Loxops, but also Himatione.

These latter two species differ in diet: Loxops (figs. 4.6-4.9) is primarily
an insect feeder, whereas Himatione feeds on nectar plus insects. The
ancestry of the family is uncertain. Tanager-like ancestry has been sug-
gested (Amadon, 1950), but the fact that the honeycreeper tapeworm
parasite (Anonchotaenia brasiliense) is known otherwise only from a South
American icterid, Cacicus (Voge and Davis, 1953), seems compelling
evidence. Both thraupids (tanagers) and icterids are nine-primaried pas-
serines.

The two subfamilies of honeycreepers are the yellow-plumaged Psit-
tarostrinae (figs. 4.6—4.15) and the “‘red-and-black” plumaged Drepanidinae
(Ciridops, Palmeria, Himatione, Vestiaria, Drepanis). The former are
basically insectivores; the latter are basically nectarivores, but also eat
insects. Which subfamily exhibits more features that would have been
seen in the ancestral stock is open to question, but Baldwin (1953) and
Bock (1970) appear to favor Drepanidinae. Changes in plumage color in
Drepanidinae are considered the result of nonadaptive genetic drift (Bock,
1970).

2. Bock (1970) has emphasized that if a single species is present on
an island, it may be variable, as most species are, and natural selection
can maintain and widen products of this variability (character release—see
chapter 3). As long as such a population is not pressed by the environment
to subdivide, it will remain a single, highly varied unit. If two congeners
invade a particular island and coexist successfully, these two residents,
rather than being intermediate, will tend to diverge, or at least the
differences between them will be enforced (character displacement—see
chapter 3). This has been shown graphically for Loxops by Bock (1970):
see figure 4.5. On Kauai, double invasions have produced the divergent
pair Loxops parva (fig. 4.6), with a short, delicate, nearly straight bill, and
L. virens stejnegeri (fig. 4.9), which has a stouter, longer, decurved bill.
A similar phenomenon has happened on Hawaii, where L. virens virens
has a short, curved, delicate bill but L. sagittirostris (fig. 4.8) has a nearly

159

L.sagittirostris (Hawaii)

L. virens virens (Hawaii)

en (Lanai)

ae (Maui)
L. virens wilsoni

al (Molokai)

L. virens chloris (Oahu)

L. virens stejnegeri (Kauai)

L. parva (Kauai)

O |2 24mm.

FIG. 4.5. Comparative bill sizes of the Loxops virens group of species (Drepanididae). Graph
represents range of culmen length for females (horizontal bar above in each) and males (bar
below in each); vertical lines represent means. Ranges and means are not known for all.
Arrangement from top to bottom corresponds to geographical sequence of the islands. Diver-
gence in bill shapes is shown on Hawaii and on Kauai, where pairs of species of this group
are sympatric; divergence is interpreted as representing nonoverlap of ecological requirements.
Redrawn from a graph presented by Bock (1970), constructed from data of Amadon (1950).

160

Adaptive Radiation: Hawaiian Islands

Fic. 4.6. Bill and feeding habits of Loxops parva (Drepanididae). In this and succeeding
drawings (to 4.15, inclusive) of Drepanididae, subfamily Psittarostrinae, bill in profile is shown
at left, portion of bird in feeding situation at right. All are drawn to the scale shown here.
Loxops parva feeds on insects available on leaf surfaces; it also takes insects from shallow
flowers.

straight, long, stouter bill. As shown in the figures, even these small
differences result in ecological separation (resource partitioning), for food
is thereby obtained from different sites.

This example is used by Bock to indicate how more extreme products
in honeycreepers may have originated. Cumulations of divergences by
variability, isolation, double invasions, character divergence, and further
ecological shifts have produced, according to Bock, such divergent types
as the parrot-billed Psittarostra (figs. 4.14, 4.15) and the sickle-billed
Hemignathus procerus (fig. 4.11). Sympatry of these various types can be
expected to occur, because they will exploit different portions of available
resources. Sympatry implies reproductive isolation, and one can hypothe-
size a positive value for reproductive isolation as a mechanism than would
prevent merging of these types and would enforce character displacement.

3. In Psittarostrinae, a variety of diets and methods of feeding
have evolved. These are summarized in figures 4.6 to 4.15 and are self-
explanatory. However, the reason that divergences can become cumulative
is that any particular type is a preadaptation for more extreme types,
which it may reach by a sort of character release if new niches remain
available. This process might be expected to occur rapidly. For example,
the curved bills of Loxops, suited for extracting insects from relatively
shallow localities, can be considered as preadapted to, by extension, the
longer curved bills of Hemignathus, which aid in extraction of insects
from deeper locations (figs. 4.10-4.12).

161

Adaptive Radiation: Hawaiian Islands

FIG. 4.7. Bill and feeding habits of ana coccinea. This bird takes Lame and larvae
on leaf surfaces and in buds. It is able to open buds, such as the ones shown, by means of
its “crossbill” beak tip, which is also illustrated as seen from above, to the right of the bill seen
in profile.

4. In Drepanidinae, there has been a divergence in diet among the
five genera, but not nearly as strong as that in Psittarostrinae, nor are
intermediates or variable species present. However, even Drepanidinae
shows remarkable differentiation among the genera. Ciridops (monotypic)
apparently was frugivorous (fruits of the palm Pritchardia are claimed to
be its food source; if so, one of the smaller-fruited species of Pritchardia
is to be suspected). Himatione (monotypic) feeds on nectar in open,
shallow flowers (Metrosideros) and eats insects. Palmeria (monotypic)
also had a mixed diet, although the decurved beak suggests more of a
tendency toward nectar-feeding and away from a primarily insectivorous
diet. Vestiaria (monotypic) is alleged to be a nectar-feeder primarily, on
tubular, curved flowers. However, it is known to have taken small insects,

aes, (i ee
; Og tl th gos Hy | /
| Bie hi lp [7 1
\ Dor Ahi} pL 4 Al pl dA] Y
Fic. 4.8. Bill and feeding habits of overs sagittirostris, which takes insects found in mod-

erately deep crevices, which it can probe by virtue of its relatively long, stout beak, as in shaggy

bark.

162

Adaptive Radiation: Hawaiian Islands

such as Drosophila, which breed in lobelioid flowers (Spieth, 1966). One
must remember that even nectarivorous birds are likely to eat insects
(which in turn feed on pollen or nectar) present in flowers in the process of
feeding on nectar. Drepanis (two species) is alleged to have been com-
pletely nectarivorous, for example, although one can imagine it might
ingest occasional insects in the process of feeding in large, nectar-rich
flowers likely to contain insects.

Baldwin (1953) explains evolution of the honeycreepers by saying
that Drepanidinae represents an early burst of basically nectarivorous birds.

%, *
Het

Fic. 4.9. Bill and feeding habits of Loxops virens stejnegeri. This species secures caterpillars
and insects from bark, among Freycinetia stems, and in the sheathing bases of Freycinetia
leaves. It also feeds on nectar from ohia (Metrosideros) flowers, and has a tongue adapted
for this.

This could explain the relictlike condition (almost all genera monotypic).
The Psittarostrinae would be, to Baldwin, as well as Amadon, a “second
burst,” its recentness denoted by the “survival of poorly adapted inter-
mediate forms.” To Baldwin, this explains why the Psittarostrinae have
not occupied the nectarivorous habit so completely as Drepanidinae—this
niche was preempted by the older subfamily.

Psittarostrinae, in my opinion, could represent an active burst of
evolution explainable not so much by persistence of “poorly adapted
intermediate forms” but by the opportunities of insectivory. Assuming
that a breakthrough to insectivory was made, this resource can be exploited
in ad much greater variety of ways, because of the many sizes and types
of insects, and sites in which they are to be found, many of which are

163

r ey
3 aif Tage § Ke a
WA: pes

Adaptive Radiation: Hawatan Islands

ric. 4.10. Bill and feeding habits of Hemignathus lucidus. It is reported to hammer with its
bill on wood, dislodging or driving out insects (especially the weevil Oddemas). It also on rare
occasions has been seen to take nectar from ohia (Metrosideros) flowers.

illustrated in figures 4.6 to 4.15. On the other hand, nectar-feeding is a
relatively stereotyped activity, varying only in the type of flower sought
and in the inclusion of varying degrees of insect food along with nectar.
In addition, the nectarivorous opportunities in the Hawauan Islands are
more limited: not only is there competition from the meliphagids (now
extinct), but nectar availability is much less during the winter months,
whereas insects and their developmental stages are available throughout
the year. To be sure, Psittarostra (figs. 4.14, 4.15) represents evolution
“beyond” insectivory (although it does eat some large insects and cater-
pillars), but its seed-eating and foliage-eating habits may be considered a
natural consequence of the preadaptation of the stout bill of Pseudonestor
(fig. 4.13), which is adapted to crushing stout twigs in search of insects
and larvae.

5. Differentiation among drepanidids with respect to feeding apparatus
and food sources almost inevitably leads to differentiation in habitat or
type of forest or levels within forest which particular species frequent.
For example, if Pseudonestor (fig. 4.13) characteristically ate larvae and
insects obtained by crushing Acacia koa twigs, it would not occur in
wet-forest areas where nectar-rich ornithophilous flowers (Metrosideros,
lobelioids) are present, but would remain in the dry koa forest. The
larger-seeded forest trees are chiefly those of dry lowland forest (Diospyros,
Colubrina, Santalum) or subalpine forest (Myoporum, Sophora), so Psit-

164

Adaptive Radiation: Hawaiian Islands

Fic. 4.11. Bill and feeding habits of Hemignathus procerus. This species is reported to have
fed on nectar in large-flowered lobelioids, such as the Clermontia shown here. Because these
flowers are pendant, feeding is from below. Hemignathus procerus is also reported to have
sought insects in leaf bases of Freycinetia arborea, or in crevices or borings. It, like H. lucidus,
had a tapping habit in coaxing insect prey from crevices.

tarostra, if it sought the larger seeds, probably occurred primarily in such
zones rather than in deep, wet forest. Psittarostra bailleui apparently
favored subalpine forest on Hawaii, whereas the other species sought seeds
in lowland forest—except for P. psittacea, whose fondness of Freycinetia
fruits (fig. 4.14) and lobelioid fruits would place it in the moist Metrosi-
deros forests where those plants grow. Loxops parva on Kauai (fig. 4.6)

165

Adaptive Radiation: Hawaiian Islands

lower mandibles have different functions. The lower is driven forcibly and repeatedly into
wood or bark in order to dislodge insects. Bits and pieces of wood loosened in this manner
are grasped with both mandibles and torm or tossed away. The lower mandible is stout and
serves not merely as a chisel but as a lever as well. The upper could serve as a probe. This
bird has never been reported to feed on nectar.

would be expected to operate in the forest canopy, by taking insects from
leaves and stems, whereas L. virens stejnegeri, also of Kauai, would be
expected to occur at lower levels within the forest, if it exploited leaf
bases of the liana Freycinetia and extracted insects from bark of tree
trunks. Baldwin (1953) notes that

Loxops virens frequents the middle and lower layers of forest trees,
whereas Himatione and Vestiaria occupy the upper story. Loxops ven-
tures into shrubby formations to a far greater degree than do the
other two birds. . . . Loxops seeks food among leaves and bark more
than do the other birds, whereas Himatione captures flying insects
more readily. Loxops tends to remain in more restricted feeding areas
than do Himatione and Vestiaria at any given time of the year. The
development and maintenance of such dissimilarities as these must
prevent serious competition among the three species.

Evolution of nectar-feeding drepanidids in the Hawaiian avifauna is
permitted by the almost indefinite flowering season of ornithophilous
plants, although they flower more abundantly during summer. Without
sources of nectar throughout the year, nectarivory would have been
impossible. At least some individuals of Metrosideros, the chief tree of
Hawaiian wet forests, can be found in flower at any time in the year.

166

Adaptive Radiation: Hawaiian Islands

Climatic moderation is undoubtedly basic to such prolonged flowering
seasons.

Short-distance migrations by honeycreepers within a habitat, related
to flowering patterns of particular trees, have been noted by Baldwin
(1953).

6. Although drepanidid bills are undoubtedly adaptive in nature,
they are nevertheless not specific with respect to diet, and drastic changes
can occur. For example, the powerful beak of Psittarostra suggests crushing
of large seeds. However, the Laysan species, P. cantans, feeds chiefly on
seeds of a grass, Eragrostis variabilis, the most abundant plant on Laysan,
because large-seeded plants (except Tribulus, whose seeds P. cantans also
eats) are infrequent or absent on Laysan. Likewise, the treeless atoll habitat
of Laysan is rather different, at first glance, from forests on the high islands.
Although there are typical small differences in leg size and other character-
istics of P. cantans in comparison with Psittarostras of the high islands,
these are not really as sharp as in the difference in habitat (P. cantans
spends most of its time on the ground, and even perches during only part
of its feeding time). Morphology may well lag behind relatively sudden
ecological shifts. This lag was stressed in the instance of the bog silversword,
Argyroxiphium caliginii, in which only a few morphological features have

es

Fic. 4.13. Bill and feeding habits of Pseudonestor xanthophrys. This species is reported to
have crushed and broken twigs in order to obtain larvae, pupae, and immature cerambycid
beetles, but more notably Clytarlus pennatus and C. modestus, in koa trees. On trees of
Pelea, it sought larvae of Plagithmysus. In addition to crushing and breaking, it employed a
nipping and wrenching technique on bark and wood (always healthy wood, never rotten).

167

Adaptive Radiation: Hawatian Islands

shown a shift to bogs from the dry-alpine habitat of the closely related A.
sandwichense.

7. The two subfamilies of drepanidids differ in song type (Baldwin,
1953) and such other features as skull development and postjuvenile molt.
These differences are difficult to explain.

Likewise, I am unable to explain why so many species of Psittarostra
(five) should occur on the island of Hawaii, whereas the other islands
have a single species each, or none. Possibly the vast tracts of lowland,
large-seeded forest and dry, upland, large-seeded forest provided greater
scope for survival of the various species on Hawaii, and each species may
have had distinctive ecological preferences, as hinted earlier. Perhaps only
on such a relatively large island could such preferences be satisfied, for
dry-forest areas are (or were) more limited on the other islands, so that
populations of a “favored plant” for a Psittarostra species would not be
large enough on smaller islands to support a viable population. Alterna-
tively, the five species on Hawaii may represent survivals of intergrades,
survivals possible because competition is less keen where very large areas
are involved or is likely to lead to extinction less quickly. Psittarostra
bailleui appears geographically isolated from other species by its high-
altitude preference.

Lessened competition on larger land areas might explain why Drepanis

FIG. 4.14. Bill and feeding habits of Psittarostra psittacea. According to several reports, P.
psittacea fed on fruits of Freycinetia arborea, shown here. These fruits are formed in long,
conelike aggregations, and each individual fruit has a relatively small seed. Berries of lobe-
lioids were also sought, as well as leafy matter, which it chewed finely. It has been reported
to catch caterpillars and feed them to its young.

168

Adaptive Radiation: Hawatian Islands

FIG. 4.15. Bill and feeding habits of Psittarostra kona. The fruits of Myoporum, shown here,
formed a mainstay of its diet. It cracked these large hard seeds with its stout beak. Fresh
leaves and caterpillars were also eaten.

pacifica and Ciridops anna are both endemic to Hawaii; alternatively,
more abundant food supplies may have permitted their survival there. If
these are “relict” representatives of the “older subfamily” Drepanidinae,
their survival on a newer island might be explainable. However, an
alternative explanation is possible: Ciridops and Drepanis are not relictual
but are relatively recent end products, produced autochthonously on the
newer islands. There does not appear to be detailed evidence for regarding
Drepanidinae as older than Psittarostrinae, and I feel that we must still
be open-minded on phylogeny within drepanidids. For example, there is
no reason to believe that any of the living gencra is highly similar to, or
nearly identical with, the original stock of the drepanidids, yet one reads
statements that a particular genus may be primitive. What is meant is
presumably that its bill features and feeding habits are like those of the
eatly honeycreepers—but we do not know for sure what the drepanidid
ancestors were, nor is there reason to believe that any of the contemporary
genera represents survival of the ancestral stock without change.

INSECTS

There is no question that, as the most numerous group of species on
the Hawaiian Islands, insects have diversified to an exceptional extent.
According to Zimmerman (1948a), between 233 and 254 original insect
immigrants can account for the 3722 species of endemic insects (perhaps

169

Adaptive Radiation: Hawatian Islands

the figure for species would be much higher today because discoveries
have been made since 1948). ‘This compares with the very similar number
of probable immigrants (272) for angiosperms, which have yielded 1633
endemic species (perhaps 1200, according to a more conservative view).
The logical question, in view of these figures, is how much adaptive
radiation, as defined earlier, has accompanied this explosive speciation.
This question is considerably more difficult to answer than in the case of
plants, because the habits of insects, revealed only by painstaking field
studies, are more difficult to study than are angiosperm adaptations; and
because adaptive morphological changes by insects are fewer and more
subtle than those of other groups of organisms. Nevertheless, thanks to
the work of several enterprising entomologists, we can cite at least a
number of instances of adaptive radiation. These can be described under
three headings, as listed below.

Shift in Vegetation Zone and Host Specificity. ‘The basic contribu-
tion in this regard is the compilation by Swezey (1954). His work is
organized in terms of flowering-plant hosts, listed alphabetically. However,
by using the index, one can easily see the range in host plants to which
any insect group is adapted. Swezey’s compilation is of particular value
because many of the observations recorded then could no longer be made,
since a number of insect species have become extinct. One must always
be careful in using host-plant lists, for there can be a difference between
plants on which particular insects have merely been seen and those which
bear a definite nutritional relationship to the insects. With this caution,
however, we may cite some examples.

The Hawaiian Lygaeidae (Heteroptera) offer good instances. This
family has been investigated by Usinger (1942), whose study was basic
to one of Zimmerman’s (1948b) monographs in Insects of Hawai. The
three large genera of Lygaeidae are Oceanides, Neseis, and Nysius. There
is no question that these genera show host specificity by species to particular
angiosperm species, and that the range of angiosperms utilized and there-
fore habitats occupied is exceptionally large. Oceanides ranges from coast-
line to alpine localities, and different species are adapted to different
plants, in areas of markedly divergent climate and rainfall. If Lygaeidae
are a group of insects basically characteristic of moderately dry localities,
Oceanides has, during its evolution in the Hawatan Islands, made every

170

Adaptive Radiation: Hawaiian Islands

degree of transition into the wet forest. The host plants represent most
of the common plant genera within each zone. Unfortunately, we do not
know the plant species, although we can guess at some of them from the
plant genera cited, because of endemism of plant species in particular
areas. However, particular species of Oceanides have been recorded from
several unrelated plant genera in many cases, so that knowledge of these
plant species is unimportant. There appears to be considerable overlap
among Oceanides species in host-plant genera utilized, so that geographical
speciation must have been important, and isolating barriers other than
those involving plant chemistry are obviously operative.

The same patterns occur in Nysius and Neseis. Nysius appears to
favor low-elevation dry localities to a greater degree, for species occur on
most of the Hawauan Leeward Islands where host plants include the grass
Eragrostis and the native Chenopodium (C. oahuense). However, other
Nysius species have adapted to situations of much greater moisture. Nysius
communis, for example, occurs in the Puu Kukui bogs of Maui, where
the host plants include Argyroxiphium grayanum and Lobelia gaudichaudii
var. gloria-montis. Nysius communis is equally well adapted to alpine
localities, such as Haleakala, where its host plants are Argyroxiphium
sandwichense, A. virescens, and Artemisia. Neseis species show a range
from rain forest to dry alpine localities, with correlative host plants. The
three species of the subgenus Icteronysius are entirely alpine.

Other groups of insects, however, show more precise specificity, where
the plant species evidently serves as the primary isolating mechanism.
Zimmerman (1948a) says, for example, “We find such great host-specificity
among the plagithmysine longhorn beetles that if we find borings in a
species of tree unrecorded as harboring a species of the group we conclude
that the borings are those of a new species. Thus, some new species of
these beetles have been discovered, and others are known only from
borings and may be captured sooner or later.” Perkins (1913) gives an
example, “Thus, even within a few square yards, the three species of
longhorn beetles, Plagithmysus darwinianus, P. lamarckianus, and P.
varians occur. It is hardly conceivable that species can be more closely
allied than these and remain distinct. Each keeps to its own food-plant,
and though occurring on adjoining trees, do not mix or interbreed.
Plagithmysus darwinianus has been found chiefly on Sophora, P. lamarcki-

A

Adaptive Radiation: Hawaiian Islands

anus on Pipturus, P. varians on Acacia.” The feature of interest stressed
by Perkins and Zimmerman is that of an isolating mechanism. However,
I can cite this example in connection with adaptive radiation because
these three beetle species have adapted to three quite different host plants.
One may suspect that adaptation to the three host plants was aided in
previous time by geographical isolation, and, following development of
specificity, the three beetle species have become distinct. This idea would
seem to be supported by a comment of Muir (1917), “I believe that one
of the first steps in species formation among Hawaiian Delphacidae is a
change of food-plant.”

Although other features may be of greater interest in speciation of
Hawaiian drosophilids, host plants (and animals) have played a significant
role. Although most drosophilids feed in decaying bark or leaves, one
genus, Titanochaeta, has a totally different food source, the eggs of spiders
(Carson et al., 1970). The number of plant genera which Hawaiian
drosophilids feed upon and breed in is relatively small, but diversified:
fermenting leaves and bark of Cheirodendron (Araliaceae); leaves, flowers,
fruits, and bark of lobelioids (Clermontia especially); leaves of Pisonia
(Nyctaginaceae); some fungi (particularly Polyporus, a bracket fungus);
and exudates from cut stumps of tree ferns (Cibotium). The nature of
the chemistry of these diverse substrates may be less important than some
of their physical characteristics; for example, rate of decay, retention of
moisture, and presence of suitable yeasts and molds. Even so, the list of
food plants (32 genera) is diverse in one or more of these respects.

With respect to Hawaiian drosophilids, Carson et al. (1970) state that

if only one immigrant was introduced, this stock early split into the
scaptomyzids and the drosophiloids. If there were separate introduc-
tions for each of these major groups, then they were constrained by
the poverty of food sources to partially overlap in their use of sub-
strates. Both groups radiated into various ecological niches, the
scaptomyzids evolving as physically small species, able to tolerate ex-
posure to higher light intensities, and thus able to use substrates such
as the flowers of morning-glories [Ipomoea] and certain non-rotting
vegetation as well as the same fermenting materials that the dro-
sophiloids use. The drosophiloid stock evolved species that are in

172

Adaptive Radiation: Hawaiian Islands

general larger than the scaptomyzids and more restricted to the areas
of low light intensity in the dense forest.

Adaptive Radiation with Regard to Morphology. With respect to
morphology, the most obvious contrasts occur in coloring. Because there
are a number of independent instances of diversification with respect to
color in the Hawaiian insect fauna, one would tend to assume that
predation must be an important factor. The scarcity of observations on
actual predation of Hawaiian insects and the value of color that stimulates
substrate does not rule out this possibility. Field observations and _lab-
oratory experiments would be of great value.

Zimmerman (1948a, pp. 142-43) cites a number of examples of
apparent protective coloration, and they need not all be repeated here.
One of his striking examples, however, is the following:

Near the summit of Haleakala on Maui, I found specimens of the
pale, silvery-colored Geranium tridens growing with their branches
intermixed with those of the dark-leaved Coprosma montana. On the
pale Geranium were colonies of a very pale Nesosydne leafhopper.
On the dark foliage of the Coprosma was a black species of the same
genus! Although the branches of the two shrubs intermixed, the two
leafhoppers kept to their own hostplants.

One may mention that the dark color of the Coprosma is striking, because
of the tendency of the branches to be covered by black lichens or fungi.

The example, also in Nesosydne, of divergence of two species with
respect to juvenile versus adult foliage of Acacia koa (Zimmerman, 1948a)
is particularly striking. On the pale-colored (glaucescent) bipinnate juvenile
leaves, one finds only Nesosydne koae. On the phyllodes (adult leaves
with flattened petiole, lacking leaf blade) of older plants one can find
N. rubescens, whose color matches the bright green or yellowish shade of
the Acacia phyllodes. Not only do adults in these two Nesosydne species
simulate colors in these respective parts of the plant, but the nymphs do
also. Nesosydne koae deposits eggs in young shoots bearing juvenile foliage.
Nesosydne rubescens, with a different ovipositor structure, lays eggs in the
edges of the leaves and in phyllodes.

Zimmerman (1948a) also stresses insect coloration by comparing the
“faunulas” found on distinctive adjacent trees. Pipturus (Urticaceae) looks

173

Adaptive Radiation: Hawaiian Islands

pale because of grayish bark and densely woolly leaves. Metrosideros
(Myrtaceae) has dark gray bark and deep green, glabrous leaves. Insect
assemblages of adjacent trees of Pipturus and Metrosideros reveal differ-
ences in color: all the Pipturus-inhabiting species of insects tend to be
light in color, whereas the Metrosideros insect species are dark.

Such color differences extend to habitats other than host plants.
Williams (1939, p. 297) states that among dolichopodid flies of the genus
Eurynogaster, species “that occur along streams, waterfalls, wet banks
bordering rapids, at springs and dripping wet banks are largely blue, green,
or coppery, while those frequenting the forest floor or trails and often the
leaves of plants, are of duller hues.” Zimmerman (1948a) expands this
example, stating that forest-floor species are brownish, foliage-inhabiting
species are greenish, and species of wet banks or pools tend to have a
metallic glitter like that of watery surfaces. Moreover, he notes that “the
species which frequent permanent streams are excellent and agile water
skaters; those found on still ponds are fair to good skaters; but the species
which hunt about areas where standing water exists only as temporary
puddles, and the flies which frequent wet ground in the absence of standing
water are poor skaters.” That these distinctions have occurred 1s interesting,
for all of the dolichopodid flies can be called wet-forest inhabitants, as
stressed by Hardy (1964); the radiation, then, is related to exploitation of
distinctive sites within the wet forest.

Adaptive Radiation with Respect to Behavior. The examples from
Dolichopodidae suggest different modes of behavior of species with respect
to water, and one genus of that family, Campsicnemus, has flightless
Hawauan species in which the wings only aid in hopping. Flightlessness
is discussed in chapter 11 (see also Zimmerman, 1948a). The most striking
examples of adaptive radiation in behavior in Hawaiian insects occur,
however, in the drosophilids. Hawaiian drosophilids are characterized by
“lek” behavior. That is, like the birds of paradise, they have well-defined
mating sites that are different from the localities at which feeding takes
place. In continental drosophilids, mating occurs at feeding sites. The
only exceptions to lek behavior in Hawaiian drosophilids occur in the
anomalipes and quasianomalipes groups. Lek behavior by itself could not
be cited as an example of adaptive radiation. Rather, the different types
of lek behavior are of interest here.

174

Adaptive Radiation: Hawaiian Islands

The remarkable patterns and diversifications in Hawauan drosophilids
are summarized by Spieth (1968). Exemplifying these behavioral patterns
is the following passage:

Each male defends a limited area, such as a single leaf, a fern pinna,
a section of a fern frond stem [= petiole], or a portion of a tree trunk.
Concurrently he advertises his presence visually or otherwise. Thus
males of some species (e.g., D. crucigera, D. grimshawi, and D. en-
gyochracea) repeatedly drag the tip of their abdomens over the sub-
strate and in doing so deposit a thin film of liquid. . . . Others
assume a ritualized posture and extrude and retract a bubble of fluid
from their anal papillae. . . . Defense of territory against intruders,
typically other males but occasionally nonreceptive females, usually
involves physical contact between the individuals, and many of the
species exhibit ritualized postures such as curling and slashing which
involves wing, leg, and body movements. . . . The most vigorous fly,
usually the largest, appears to be the victor in such contests, the
original possessor of the area often being displaced by the intruder.
The mere assumption of the aggressive posture, without physical
contact, frequently is sufficient to determine the outcome of the con-
test. If the two individuals are of disparate size, the assumption of
the aggressive posture usually causes the smaller to flee. Under field
conditions, the contestants apparently do not physically injure each
other.

Morphological adaptations related to various of these ritualized
behavior patterns occur, of course. Moreover, radiation has occurred with
respect to sites of courtship behavior. As noted by Spieth,

these sites are not randomly distributed in the rain forest where the
flies live, but each species has special preferences, apparently deter-
mined by factors such as light, humidity, temperature, and spatial
conditions. Invariably they are in close proximity to where the in-
dividuals feed and the females oviposit. . . . Each leaf on a single
branch of a given shrub may have a male on its under surface, or the
frond stems [= petioles] of a particular tree fern may have a male
patrolling and defending a two to four foot section of each [petiole].

For a more detailed account of these behavioral characteristics and
the morphological features of the flies associated with these performances,

175

Adaptive Radiation: Hawaiian Islands

‘one should consult Spieth (1968). Obviously, many more details remain
to be discovered, considering the large number of Drosophila species in
the Hawauan Islands. While one cannot hope for comparable data on
mating behavior and sites in all of these, increased knowledge will surely
result in a more complete picture of this phenomenal instance of adaptive
radiation.

MOLLUSCA

Two groups of land molluscs have speciated in a remarkable way on
the Hawaiian Islands: the amastrids and the achatinellids (see discussion
earlier in this chapter). The amastrids are largely terrestrial, whereas the
achatinellids are arboreal. ‘This divergent preference in habitats may
explain why each has undergone adaptive radiation and speciation; if both
occupied similar ecological niches, speciation of one group would likely
have been suppressed by the other.

The speciation of the achatinellids has been due, as stated earlier,
to the great degree of dissection of habitats suitable to them, and to the
tendency for colonies to migrate very little and very slowly. Although
many of the distinctive species patterns may merely be the product of
genetic drift, Perkins (1913) claimed that distinctive colors of achatinellids
might serve a protective function. In my experience, achatinellid snails
are by no means camouflaged, although they may be less conspicuous
merely by their typical position on lower surfaces of leaves. For example,
the brown-and-white striped Achatinella mustelina always occurs on green
leaves; in the genus at large, green species are infrequent, but would be
expected to be the ones best suited for protective coloration. Perkins
questions what could be a predator; he finds no really plausible bird, and
only birds would be likely to serve as predators for land snails of this size.
The concept of genetic drift was not really developed in Perkins’s time,
so his search for an adaptive significance is understandable.

There is a noticeable degree of host-plant fidelity in achatinellids.
They are absent on some tree species, present on others, in any given
area; they avoid, with few exceptions, species of plants introduced by
man. Achatinella grazes on fungi and algae on leaves, not on the leaves
themselves. Some features of leaf texture or leaf chemistry may, however,
tend to favor Achatinella species, or be preferred by them. There may be

176

Adaptive Radiation: Hawaiian Islands

some radiation with respect to host plants in this genus. If so, this matter
needs careful investigation and assessment.

In the genus Amastra, and in the closely related genera Carelia,
Pterodiscus, and Planamastra, there is remarkable differentiation in shell
size and shape (see Zimmerman, 1948a; Carlquist, 1970a). Such distinctive
forms seem to suggest an explanation in terms of adaptive radiation. So
far, however, we do not have enough field data to present distinctions in
habit and habitat that might suggest explanations. This remains an
interesting problem for investigation.

References

Amadon, D. 1950. The Hawaiian honeycreepers (Aves, Drepaniidae). Bull. Amer.
Mus. Natur. Hist. 95(4):153-262.

Baldwin, P. H. 1953. Annual cycle, environment and evolution in the Hawaiian
honeycreepers (Aves: Drepaniidae). Univ. Calif. Publ. Zool. 52:285-398.

Berger, A. J. 1972. Hawaiian bird-life. Honolulu, University of Hawaii Press.

Bock, W. J. 1970. Microevolutionary sequences as a fundamental concept in macro-
evolutionary models. Evolution 24:704-22.

Carlquist, S. 1957. Leaf anatomy and ontogeny in Argyroxiphium and Wilkesia
(Compositae). Amer. Jour. Bot. 44:696-705.

Carlquist, S. 1958. Wood anatomy of Heliantheae (Compositae). Trop. Woods
108:1—30.

Carlquist, S. 1959a. Vegetative anatomy of Dubautia, Argyroxiphium, and Wilkesia
(Compositae). Pacific Sci. 13:195-210.

Carlquist, S. 1959b. Studies on Madiinae: Anatomy, cytology, and evolutionary
relationships. Aliso 4:171—236.

Carlquist, S. 1962. Ontogeny and comparative anatomy of thorns of Hawaiian
Lobeliaceae. Amer. Jour. Bot. 49:413-19.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1969a. Wood anatomy of Goodeniaceae and the problem of insular
woodiness. Ann. Missouri Bot. Gard. 56:358—90.

Carlquist, S. 1969b. Wood anatomy of Lobelioideae (Campanulaceae). Biotropica
1(2):47-72.

Carlquist, S. 1970a. Hawaii, a natural history. New York, Natural History Press.

Carlquist, S. 1970b. Wood anatomy of Hawaiian, Macaronesian, and other species
of Euphorbia. Bot. Jour. Linnaean Soc. 63, suppl. 1:181-93. (Also as New Re-
search in plant anatomy, ed. D. Cutler. London, Academic Press.)

Carson, H. L., D. E. Hardy, H. T. Spieth, and W. S. Stone. 1970. The evolutionary
biology of the Hawaiian Drosophilidae. In Essays in evolutionary biology in honor
of Theodosius Dobzhansky, ed. M. K. Hecht and W. C. Steere, pp. 437-543.
New York, Appleton-Century-Crofts.

177

Adaptive Radiation: Hawaiian Islands

Hardy, D. E. 1964. Insects of Hawaii, vol. 11: Diptera: Brachycera II—Cyclorhapha.
Honolulu, University of Hawaii Press.

Hardy, D. E. 1965. Insects of Hawai, vol. 12: Diptera: Cyclorhapha. Honolulu,
University of Hawaii Press.

Herbst, D. 1971. A new Euphorbia (Euphorbiaceae) from Hawaii. Pacific Sci. 25:
489-90.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography.
Princeton, N.J., Princeton University Press.

Muir, F. 1917. New Hawaiian Delphacidae. Proc. Hawaiian Entomol. Soc. 3:168-
DD.

Munro, G. C. 1960. Birds of Hawaii. 2d ed. Rutland, Vt., Ridgeway Press.

Oliver, W. R. B. 1935. The genus Coprosma. Bishop Mus. Bull. 132:1—207.

Perkins, R. C. L. 1903. Fauna Hawaiiensis: Vertebrata, vol. 1, pt. 4, pp. 365-466.
Cambridge, Cambridge University Press.

Perkins, R. C. L. 1913. Fauna Hawatiensis: Introduction, vol. 1, pt. 6, xv—cexxviil.
Cambridge, Cambridge University Press.

Richardson, F’., and J. Bowles. 1964. A survey of the birds of Kauai, Hawaii. Bishop
Mus. Bull. 227:1-49.

Rock, J. F. 1919. A monographic study of the Hawaiian species of the tribe Lobe-
lioideae, family Campanulaceae. Mem. Bishop Mus. 7(2):1—394.

Rock, J. F. 1957. Some new Hawaiian lobelioids. Occas. Papers Bishop Mus. 22(5):
35-66.

Rock, J. F. 1962. Hawaiian lobelioids. Occas. Papers Bishop Mus. 23(5):65-75.

Rothschild, W. 1893-1900. The avifauna of Laysan and the Hawaiian possessions.
London, R. H. Porter.

St. John, H. 1959. Botanical novelties on the island of Niihau, Hawaiian Islands.
Hawaiian plant studies 25. Pacific Sct. 13:156—90.

St. John, H. 1966. Monograph of Cyrtandra (Gesneriaceae) on Oahu, Hawaiian
Islands. Bishop Mus. Bull. 229:1—465.

Sherff, E. E. 1935a. Revision of Haplostachys, Phyllostegia, and Stenogyne. Bishop
Mus. Bull. 136:1-101.

Sherff, E. E. 1935b. Revision of Tetramolopium, Lipochaeta, Dubautia and Rail-
liardia. Bishop Mus. Bull. 135:1—136.

Sherff, E. E. 1937. The genus Bidens. Part I. Field Mus. Natur. Hist. Publ., Bot.
ser. 16:1—346.

Spieth, H. T. 1966. Hawaiian honeycreeper, Vestiaria coccinea (Forster) feeding on
lobeliad flowers, Clermontia arborescens (Mann) Hillebrand. Amer. Natur. 100:
470-73.

Spieth, H. T. 1968. Evolutionary implications of sexual behavior in Drosophila. In
Evolutionary biology, ed. T. Dobzhansky, M. K. Hecht, and W. C. Steere,
vol. 2, pp. 157-93. New York, Appleton-Century-Crofts.

Stone, B. C. 1969. The genus Pelea A. Gray. Phanerogamarum Monogr. 3:1-180.

Swezey, O. 1954. Forest entomology in Hawai. Bishop Mus. Spec. Publ. 44:1-266.

Thwing, E. W. 1907. Reprint of original descriptions of Achatinella, with addi-
tional notes. Occas. Papers Bishop Mus. 3(1):1-196.

178

Adaptive Radiation: Hawaiian Islands

Usinger, R. L. 1942. The genus Nysius and its allies in the Hawaiian Islands. Bishop
Mus. Bull. 173:1-167.

Voge, M., and B. S. Davis. 1953. Studies on the cestode genus Anonchotaenia
(Dilepididae, Paruterininae) and related forms. Univ. Calif. Publ. Zool. 59:1-30.

Williams, F. X. 1939. Biological studies in Hawaiian water-loving insects. Part IIIB,
Diptera. Proc. Hawatian Entomol. Soc. 10:281-315.

Wilson, S. B., and A. H. Evans. 1890-1899. Aves Hawaiienses: The birds of the
Sandwich Islands. London, R. H. Porter.

Wimmer, F. E. 1956. Campanulaceae—Lobelioideae. Das Pflanzenreich IV 267b
(106):1—260.

Yuncker, G. T. 1933. Revision of the Hawaiian species of Peperomia. Bishop Mus.
Bull. 112:1—131.

Zimmerman, E. C. 1948a. Insects of Hawaii, vol. 1: Introduction. Honolulu, Uni-
versity of Hawaii Press.

Zimmerman, E. C. 1948b. Insects of Hawaii, vol. 3: Heteroptera. Honolulu, Uni-
versity of Hawaii Press.

179

CHAPTER
PIV

ADAPTIVE RADIATION:
CANARY ISLANDS AND OTHER
MACARONESIAN ISLANDS

Although the term Macaronesia is not as generally used as it should be,
and is not yet recognized officially, it is of value for referring to a bio-
geographically related group of volcanic archipelagos: the Azores, Madeira,
the Canary Islands, the Cape Verde Islands, and a few minor islets near
or between these major groups. Of these, the Canary Islands are of
predominant interest because their larger area and greater elevation make
them the prime archipelago of Macaronesia for biological diversity.
Madeira is biologically rich, considering its remoteness; it represents a
mesic outlier of the Canary Islands, in effect. It can be bracketed with
the Canaries biologically, but it is separated from them by a distance
much greater than the gaps among the Canaries, so that it does not
participate in archipelago effects except in somewhat peripheral ways.
However, some groups of plants on Madeira must be considered part of
Canarian patterns of adaptive radiation. The relatively depauperate floras
of the Azores and the Cape Verde Islands are decidedly peripheral to
Madeira-Canary patterns, although some of the same genera do occur.
The reason for this is not merely the remoteness of the Azores and the
Cape Verde Islands, but also their small area and low elevation.

Area and altitude are important factors in species abundance on an
island, and can play a part—if other suitable factors are present—in
promoting adaptive radiation. Table 5.1 has been prepared to offer a view
of these factors. Introduced as well as native vascular plants are included

180

Adaptive Radiation: The Macaronesian Islands

TABLE 5.1]
Areas, altitude, and vascular plant species of the Canary Islands
Lanza- Fuerte- Gran Tene- Go- La All
rote ventura Canaria rife mera Palma Hierro islands

Area (sq km), total 873 1725 1534 2060 378 728 I SHS
below cloud belt 873 1/25 ee OS Oh aaelleOM nz 08 363 194 5563
in cloud belt 470 445 170 260 82 1427
above 1500 m 34 445 105 1 585

Highest point, m 670 807 19502 37 1484s 24238 1520) sl

No. of species,
vascular plants 366 348 73m O79 e589 yD 30) eloe

Source: Data from Lems, 1960.

in the table. Inclusion of introduced plants seems valid for demonstrating
species richness of the various islands. Table 5.1 shows that habitat
diversity, rather than any of the individual measures, is the prime key to
biological diversity. Thus, Fuerteventura has a larger area than Gran
Canaria but less than half the number of plants, because it is essentially
a flat island compared with Gran Canaria, and has only a lowland vege-
tation. In fact it has only a fraction of the lowland flora because it lacks
the barrancos and other topographic diversity that would be present in
the lowlands of a high island. Gran Canaria and Tenerife are roughly
comparable in areas of lowland and cloud belt, but Tenerife has a much
greater alpine zone (above 1500 m), and this provides another key to
species diversity: it is not as important that each zone be present on an
island as that these zones be of sufficient size to maintain variable and
viable populations of species. This explains why La Palma, which has
appreciable areas in each of the zones, does not have exceptional species
richness: area within each zone is too small to permit much species
diversity or diversification. In this regard, one should note that Tenerife
is divided by a pass into the high, recent El Teide caldera, and the older
Anaga range, which reaches into the cloud-forest zone. Presence of all
zones on La Palma does permit it to bear a greater number of species than
Lanzarote or Fuerteventura, although the latter islands have greater areas.
The geological recentness of La Palma, compared with others of the
Canaries, may also be a reason why it has fewer species than does
Tenerife.

181

Adaptive Radiation: The Macaronesian Islands

Angiosperms

Although zoologists may find the Canary Islands and Madeira poor in
adaptive radiation, the botanist can find numerous good examples. Sonchus
(Asteraceae), Euphorbia (Euphorbiaceae), and Limonium are excellent
examples of adaptive radiation on these islands. ‘They would be considered
here except that in each of these examples there is a series in increased
woodiness, and I have chosen to include them in chapter 10, which should
be considered in part an annex of this chapter. Also, these groups show
changes in wood anatomy concomitant with radiation into more mesic
areas from drier ones (chapter 10). Increased woodiness is certainly
involved in adaptive radiation of Echium (Boraginaceae), but other aspects
of adaptive radiation are presented by this genus as well.

Echium. Continental species of Echium are small, herbaceous plants,
generally annuals. ‘This growth form is easily seen on the Canary Islands
because the weedy European E. plantagineum (fig. 5.1, top left) is a
common plant, introduced by man to the Canaries as it has been to other
areas where it has also become a weed (e.g., Australia). There are native
annual Echiums in the Canary Islands, such as the endemic E. triste.
However, a number of short-lived perennial Echiums should also be cited
among Canarian endemics. These include E. bonnettiu, E. hierrense
(fig. 5.1, top right), and E. strictum (fig. 5.2, top left). A moderated
chmate permits these species to persist for more than a single season.
For example, E. strictum, which grows in open areas in the cloud forest,
receives enough moisture so that neither protracted drought nor freezing
winter temperatures terminate its life-span. The largest species of Echium,
E.. candicans (fig. 5.2, top right) is reminiscent of a gigantic E. strictum.
Echium candicans grows in the wet laurel forests of Madeira—probably
the most mesic habitat in Macaronesia—and thereby attains greater size
and has leaves larger than those of other shrubby species. Echium
candicans reflects a pioneering quality common to Echium species by
growing in openings in the laurel forests, but these openings are small
in area and not really drier than the surrounding laurel forest. Therefore,
not surprisingly, E. candicans is almost indefinite in life-span, although
it never becomes a true tree.

182

Fic. 5.1. Species of Echium (Boraginaceae) from the Canary Islands. rop Lert, E. planta-
gineum habit; this plant is an introduced weed in the Canaries, grows during the wet season.
TOP RIGHT, E. hierrense, inflorescence with leaves, showing the dense covering of upwardly
appressed hairs on leaves. BOTTOM LEFT, E. onosmaefolium, specimen about 2 m tall; note

al trunk and shaggy bark; Barranco de Arguineguin, Gran Canaria. BOTTOM RIGHT,
E. leucophaeum var. angustissimum, shrub about 1 m tall; Barranco de San Andres, Tenerife.

183

Adaptive Radiation: The Macaronesian Islands

Perhaps if a true rain forest were present in Macaronesia, Echium
could be represented by real trees, as is Dubautia in the Hawauian Islands.
Echium candicans represents the farthest advance toward mesomorphy
that the Macaronesian climate permits (except perhaps for E. pininana).
Echium candicans has been cultivated in southern and northern California
(usually as “E. fastuosum’’), and there plants live for 10 years or more
near the coast, but at sites about 40 mi inland the plants survive for
fewer years because low humidity and higher summer temperatures cause
excessive transpiration.

Most shrubby Macaronesian Echiums are not, however, laurel-forest
species. For example, E. onosmaefolium (fig. 5.1, bottom left) and E. leuco-
phaeum (fig. 5.1, bottom right) are lowland or dry-montane species.
Echium onosmaefolium is perhaps the largest shrubby Echium in the Ca-
nary Islands. Among the mechanisms that seem related to its longevity are
its shaggy, corky bark, whose layers would reduce transpiration, and re-
duction of leaves, both in number and in size. During the dry season, a
plant can be relatively leafless. Echium leucophaeum may also produce
smaller and fewer leaves during the dry season. Significant in this species
are the very hairy, silvery leaves, which may reduce transpiration by form-
ing a “windbreak” on leaf surfaces or by reflecting incident illumination,
or both. Leaves of E. hierrense (fig. 5.1, top right) are similarly silvery or
felty, and represent potential reduction in transpiration sufficient for sur-
vival in the dry season. On Tenerife, E. giganteum represents the closest
approach in the Canarian flora to E. candicans of Madeira, and, like it,
has relatively broad leaves.

Echium pininana (figs. 5.2, bottom left, center; 5.4, bottom right) is
a striking endemic of La Palma. It is taller than other Echiums, but
achieves this height by virtue of being unbranched. Echium pininana
grows in dense laurel forests, and when young has long internodes. As it
reaches better-illuminated levels, internodes become shorter, leaves be-
come larger, and the plant flowers. Echium pininana is monocarpic,
whereas most other perennial species of Echium do not die after flower-
ing. There seems little doubt that E. pininana is an adaptation to laurel-
forest areas; unlike E. candicans, it can grow in dense areas and does not
require openings in the forest. The parallel between E. pininana and Wil-
kesia of the Hawauan Islands (see Carlquist, 1970a, for illustrations) is

184

FIG. 5.2. Species of Echium from relatively mesic localities in the Canary Islands and Madeira.
TOP LEFT, E. strictum, plant about 0.5 m tall; at head of Barranco de San Andres, Tenerife.
TOP RIGHT, E. candicans, shrub 2.5 m tall; in opening of laurel forest, Ribeiro Frio, Madeira.
BOTTOM LEFT, CENTER, E. pinindnd. BOTTOM LEFT shows specimens from the wild, with in-
florescences; Los Roques y Los Silos, La Palma; BorroM CENTER shows a specimen grown
from seeds of the wild plants; cultivated, in Santa Barbara, Calif. sorrom ricHT, E. simplex,
plants ca. 2 m tall (inflorescence included); cultivated in Jardin Canario, Gran Canaria.

185

Adaptive Radiation: The Macaronesian Islands

remarkable. If grown in cultivation (fig. 5.2, bottom center), E. pininana
does not form a tall stem, and internodes are all of intermediate length.
Apparently, under natural conditions, flowering occurs only when the leaf
crown reaches upper, better-lit levels of the forest, and presumably accu-
mulates sufficient photosynthates to flower. The same is true of Wilkesia,
which also flowers “‘prematurely” in cultivation.

Echium simplex (fig. 5.2, bottom right) is a species of relatively low
elevations on Tenerife, but in relatively mesic sites on the Anaga Penin-
sula. Like E. pininana, it is monocarpic, but acaulescent. Because sufficient
illumination is present near the ground level in its habitat, the lack of an
elongate stem in E. simplex is understandable. A thin felt of silvery hairs
is present on leaves of E.. simplex, whereas leaves of E. pininana are hispid
but not covered by “shading” or “reflecting” hairs. The large leaf size in
both E. simplex and E. pininana seems correlated with the fact that the
production of a single massive inflorescence requires a broad area of photo-
synthetic surface that is provided, in each case, by means of a rosette of
large leaves.

Similar in form to E. simplex is the alpine species of Tenerife, E.
bourgeauanum (fig. 5.3, top left), which has bright crimson flowers rather
than whitish ones, as in E. simplex. The conspicuous color may relate to
the factors that tend to produce conspicuous, brightly colored flowers in
alpine angiosperms in various parts of the world. Rosettes of E. bourgeau-
anum are narrow, rather silvery (but with some hispid hairs), and prob-
ably resist low alpine humidity and high levels of illumination; they are
sufficient, however, in the brilliant illumination of the Teide caldera, for
enough photosynthesis to produce, after several seasons, the single massive
inflorescence that terminates the life of an individual. In virtually all re-
spects E. bourgeauanum forms a striking parallel to the dry alpine species
of Argyroxiphium (chapter 4) on the Hawaiian Islands. Thus we see how
similar selective factors, acting on two unrelated genera, have operated to
produce similar results.

There is no Hawaiian parallel, however, to a second alpine species of
Tenerife, E. auberianum (fig. 5.3, top nght, bottom left). This species
occurs on cinder cones within the caldera of E] Teide. It is an acaulescent
rosette plant, but it does not form a massive rosette of leaves, then flower
and die. Each flowering shoot does die after flowering and fruiting, but

186

FIG. 5.3. Species of Echium from areas of recent volcanism in the Canary Islands. ToP LEFT,
E. bourgeauanum, plants about 2 m tall; caldera of El Teide, Tenerife. ror ricut, E. auber-
ianum, plant 1 m tall; Arenas Negras cinder cones, El] Teide, Tenerife. BorroM LEFT, same,
base of plant to show perennial rootstock, from the top of which stems arise. BOTTOM RIGHT,
E. leucophaeum growing as a reduced shrub on cinders; Volcan San Antonio, La Palma.

187

rad

FIG. 5.4. Echium species as cultivated in the Rancho Santa Ana Botanic Garden, Claremont,
photographed after a freeze (ca. —3 C) in December
1967, to show differential damage. rop LEFT, E. nervosum (Madeira), leaf tips killed, stems
intact. TOP RIGHT, E. webbiu (La Palma), little or no damage; plants still surviving well as
of 1972. Borrom LEFT, E. faramae (Lanzarote), all plants killed. Borrom ricHT, E. pininana

(La Palma), little or no damage.

Calif. (plantings by Kornelius Lems),

188

Adaptive Radiation: The Macaronesian Islands

new innovations are formed on the perennial, succulent caudex (fig. 5.3,
bottom left). This habit, which recalls other genera of Boraginaceae, such
as Cynoglossum, seems well adapted to the alpine cinder-cone habitat.
Flowering shoots are limited in duration, so that, as snow melts and the
soil dries, shoots are not exposed during the dry summer and autumn
months. Moreover, the caudex seems adapted to cinder cones; if it be-
comes buried more deeply as cinders slide downslope, it can still innovate
shoots in successive years and the shoots will reach the surface. Signifi-
cantly, both E. auberianum and E. bourgeauanum are resistant to the cold
temperatures of the high caldera. Echium bourgeauanum (like Argyroxi-
phium sandwichense) has the apical bud protected by a dense series of
leaves, and the sessile nature of the rosette prevents damage to the stem,
which is extremely short. Echium auberianum survives freezing tempera-
tures by means of its underground caudex, which innovates shoots only
during the growing season.

Echium leucophaeum (figs. 5.1, bottom right; 5.3, bottom right) can
be found on new cinder cones, as on Volcan San Antonio (fig. 5.3, bottom
right). In these plants, leaves are extremely narrow and silvery with dense
hairs, the plants being nearly prostrate. These characteristics suggest an
ecotype successful on cinder cones, which are relatively dry and brilliantly
illuminated, with much heat radiated from the cinders during daylight
hours. Plants of E. leucophaeum in dry lowlands (fig. 5.1, bottom right)
are taller, but reduction of leaf surface corresponds to lower moisture
availability.

Although this account does not include mention of all Macaronesian
Echium species, those not specified can be grouped with the adaptational
modes described. One dimension of adaptive radiation that is not visible,
however, is climatic tolerance. These factors were demonstrated by an
experimental planting of various Echium species at the Rancho Santa Ana
Botanic Garden, in Claremont, California. This planting was undertaken
in summer, 1965, by the late Kornelius Lems. In late December, 1967, an
unusually severe freeze (about 25 F') occurred at the Botanic Garden.
This freeze was highly informative, because differential cold tolerances
were illustrated graphically (fig. 5.4). Observations made in February,
1968, when damage could be clearly assessed, revealed the following:

189

Adaptive Radiation: The Macaronesian Islands

E. aculeatum (a species close to E. leucophaeum), from middle
elevations on Tenerife: half of the plants killed.

E. aculeatum, from low elevations on Hierro: all plants killed.

E. bourgeauanum, from El Teide, Tenerife: all plants killed.

E. faramae, from lowlands, Lanzarote: all plants killed (fig. 5.4,
bottom left).

E. giganteum, middle elevations, Tenerife: all plants killed.

E.. leucophaeum, from lowlands, La Palma: almost killed; some
plants survived but died soon after February, 1968.

E. nervosum, from lowlands of Madeira: leaves damaged, but
plants survived, new leaves produced (fig. 5.4, top left).

E. onosmaefolium, middle elevations, Gran Canaria: leaves and
younger stems damaged, but new shoots formed and plants
surviving well.

E. pininana, from laurel forests, La Palma uplands: tips of leaves
damaged, but plants otherwise intact (fig. 5.4, bottom right).

E.. strictum, Tenerife upper lowlands: all plants killed.

E. webbii, caldera of La Palma: all plants intact (fig. 5.4, top
right).

Lowland plants, and plants of frost-free (or relatively frost-free) areas
not far above the coast on Macaronesian islands, were subject to the
greatest damage or killing. Of these, larger, woodier plants (E. nervosum,
E.. leucophaeum, E. onosmaefolium) survived better, possibly because not
all parts of the plant were reached by freezing, or possibly because the
chances of survival of some stems on a larger plant are greater. Echium
faramae and E. strictum were clearly vulnerable, and reminded me of how
a large ornamental annual could be killed by frost in a garden. The best
survival was for higher-elevation (but not alpine) species: E. webbii, for
example. In fact, at the date of writing, these plants are still growing suc-
cessfully; they also survive hot, dry summers in Claremont, whereas the
horticulturally popular E. candicans survives Claremont summers poorly.
Obviously upland Echiums have evolved frost-resistance not present in
the lowland species. Partial frost resistance in the lowland Madeira species,
E. nervosum, may be explained by the fact that on that more northerly
island, frosts can occur nearer the coast. One might have expected E. bour-

190

Adaptive Radiation: The Macaronesian Islands

geauanum to survive, but evidently this plant was dying in the Botanic
Garden prior to the freeze. One might speculate that heavy soil, warm
temperatures on autumn and winter days, or artificial watering—combined
with pathogenic fungi—contributed to its decline.

Descriptions of growth forms in Echium have been offered (Lems,
1961). Readers should be warned that Lems changed his ideas on phylog-
eny of growth forms subsequent to that paper but did not have an
opportunity to revise his concepts. The same applies to the account of
trichomes in the genus (Lems and Holzapfel, 1968). ‘Trichomes, as well
as wood anatomy (Carlquist, 1970b), correlate closely with ecology in
Echium. There seems little doubt that there has been evolution toward
increased woodiness in Echium on Macaronesian islands, conforming to
that of many groups on islands (chapter 10). The diversification within
Echiums on Macaronesia is, however, of great interest, and one cannot
help drawing a parallel with the Dubautia-Argyroxiphium-Wilkesia com-
plex on the Hawaiian Islands. Because there are no true rain forests in
Macaronesia, equivalent climatically to those of the Hawaiian Islands, the
adaptive radiation in Echium seems “foreshortened” compared with the
Hawaiian complex.

Aeonium. A large number of species of this genus, which is nearly
endemic to Macaronesia, occur in the Canary Islands. Two species are re-
stricted to the Cape Verde Islands, two to Madeira, one is endemic to
Morocco (A. arboreum), and one (A. leucoblepharum) to Abyssinia and
Somaliland (Praeger, 1932). Even if one were to regard the mainland
African species as relicts, it seems obvious that radiation has occurred
abundantly on the Canary Islands. If one regards the mainland African
species as relicts, one is taking a difficult interpretation, for each of them
is similar to a Canarian species, whereas they are not similar to each other.
They could possibly be, in one or both cases, results of dispersal from
islands to mainland. At any rate, the Canarian Aeoniums offer a good ex-
ample of radiation in growth form on the Islands.* There is, indeed, a

* My plate of four species of Aeonium (Carlquist, 1965, fig. 8.5) misidentifies a
plant as A. glutinosum; it probably is A. holochrysum, although excessive hybridiza-
tion in cultivated Aeoniums makes any plant in cultivation suspect where indentif-
cation is concerned. Aeonium arboreum is, by its presence in that plate, implied to
be a Canarian species; it was, in fact, thought to be from the Canaries for many

101

Adaptive Radiation: The Macaronesian Islands

range from the acaulescent species such as A. tabulaeforme (Tenerife)
and A. glandulosum (Madeira) to the shrubby species such as A. manri-
queorum and A. haworthii. Some (A. glandulosum) can be annual or bien-
nial, whereas the shrubs are of indefinite duration. A contrast can be
drawn between the monocarpic species (some of which are acaulescent,
some tall) and the non-monocarpic, shrubbier species. However, even the
shrubby species are relatively determinate in growth, as shown by their
tendency to form abundant aerial roots on upper stems, which tend to fall
off and root, or root when the main stem falls over onto the ground. The
accumulation of secondary xylem is limited because of this tendency. How-
ever, the habitats occupied by Aeonium species show a remarkable range;
one might expect this genus to be characteristic only of dry situations be-
cause of its succulence, which is not really markedly different from one
species to another. To be sure, there are lowland species, but even these
tend to grow in barrancos (A. holochrysum, A. manriqueorum, A. sedifo-
lium, A. viscatum), just as A. arboreum grows in similar barrancos on the
nearby Moroccan coast (Praeger, 1932). Other species of xeric localities
below the forest belt include A. nobile, A. percarneum, A. saundersii, and
A. tabulaeforme (Lems, 1960). In the cloud forest, one can find A. gooch-
ide (laurel forest) and A. cuneatum. Aeonium also enters subalpine condi-
tions (A. caespitosum, A. spathulatum), but apparently never is a true al-
pine plant. Although its succulence would seem to qualify it for such habi-
tats, none of the Canarian Aeoniums are resistant to heavy frost. In diam-
eter, Aeonium rosettes differ widely, from 80 cm (A. canariense) to about
3 cm (A. sedifolium).

Lamiuaceae. Lamiaceae have radiated into a variety of Canarian habi-
tats. Lewcophae, a Canarian genus dubiously segregated from Sideritis, is
endemic. It ranges from near sea level to the subalpine zone. Leucophae
mdssoniana occurs in lowland barrancos, although in moister parts of them.
Cliffs are the habitat of L. gomerae, with pendulous stems. The large-
leaved L. macrostachys grows in laurel forests, while L. eriocephala is a sub-

years; it is apparently restricted to Morocco (Praeger, 1932). In my (1965) text, it is
called ‘“‘A. arborescens’’ by mistake, although it is correctly labeled on the figure.
There are Canarian species, such as A. manriqueorum, which form shrubs as large
as those of A. arboreum, and could be cited as Canarian representatives of this
growth form that have evolved there autochthonously.

192

Adaptive Radiation: The Macaronesian Islands

alpine cliff plant (Lems, 1960). The genus Micromeria can range from dry
lowland barrancos (M. densiflora, M. linkit) to pine forests (M. lanata),
to openings in laurel forests (M. lepida), to rocky subalpine localities (M.
benthamii, M. teydensis, M. tragothymus).

Dipsacaceae. The three endemic Canarian species of Pterocephalus
occupy diverse localities. Pterocephalus dumetorum (fig. 10.9, bottom left,
right) 1s a shrub of subalpine scrub, whereas P. lasiosperma grows in “sub-
alpine desert” (Lems, 1960), and P. virens is a lowland species.

Ecotype Formation in Various Families. There are instances in the
Canarian flora of highly variable species that cover a broad altitudinal
span and may form distinctive ecotypes in various zones. These include
Andryala pinnatifida (Asteraceae) and Cheiranthus virescens (Brassica-
ceae). The latter may be considered conspecific both with the alpine C.
scoparius and the Madeiran species C. mutabilis, which can be found on
banks of waterfalls as well as in subalpine grassland. These species may
represent incipient adaptive radiation. However, their adaptation to any
localities other than dry ones seems superficial and recent. Where they do
occur in more mesic localities, they seem to be chance pioneers in open-
ings, as with C. mutabilis along waterfalls in laurel forest on Madeira.
However, entry into more mesic localities as an opening phase in adaptive
radiation may occur in just such a way, so we should not neglect examples
of this sort as irrelevant to consideration of adaptive radiation.

Animals

The Canary Islands, as well as the Azores and other Macaronesian islands,
have relatively “harmonic” passerine bird faunas, This, as Lack (1947)
notes, may be the reason why these islands show no cases of adaptive radia-
tion in their avifauna. Although the avifauna of Macaronesian islands is
depauperate in land birds in comparison with mainland areas (e.g., Ban-
nerman and Bannerman, 1963), this reflects the fact that there are fewer
niches available on the islands—that is to say, fewer niches of sufficient
extent to support breeding populations. As Bannerman and Bannerman
(1963) also note, many straggling land birds from Morocco visit Madeira;
access is obviously easy, which tends to make an island fauna less strongly

193

Adaptive Radiation: The Macaronesian Islands

disharmonic, so that adaptive radiation is forestalled. Lack (1947) has
noted the possibility that successively newer immigrants, more aggressive
than the insular species already there, will tend to wipe out groups that
have undergone adaptive radiation. In Macaronesia, it would seem that
on account of easy access, adaptive radiation in birds never had a chance.
MacArthur and Wilson (1967), as noted earlier (chapter 3), hypothesize
that on remote large islands and archipelagos, adaptive radiation is likely
to occur if the distance from a source area is at a maximum range of dis-
persal for a particular group of animals. Lack finds the Galapagos Islands
more remote from the mainland than they would seem geographically,
because those islands lie in a region of calms, whereas winds favor immi-
gration to Macaronesia from source areas. Lack of adaptive radiation in
Macaronesian avifauna does not seem due to lack of time: the angiosperm
radiation has become pronounced. Apparently, what constitutes “opportu-
nity” for one group does not for another.

In the Canarian avifauna, there are cases of incipient speciation in
birds. For example, there are four geographic races in the titmice (Parus
teneriffae) according to Bannerman and Bannerman (1963). In the case
of chafinches on the Canary Islands, Lack, agreeing with Mayr (1942),
finds the two species of Fringilla to be the result of two successive intro-
ductions, not radiation. The endemic species, F. teydea, is hypothesized to
be the older. It breeds only in pine forests on Gran Canaria and on Tene-
tife. Its adaptations are discussed in chapter 15. On those two islands, F.
coelebs, hypothesized as a more recent introduction and native also in
Europe, “avoids” pine forest and lives in laurel forest and chestnut wood-
land. However, on La Palma, F. teydea is absent, and F.. coelebs occurs in
all forest types, as it does in Europe. These examples suggest that, with
respect to avifauna, resource availability is tight and unsuited to diversifi-
cation of a group: a given bird species either meets the needs of a niche
(and possibly changes somewhat to achieve a more precise adaptation to
that niche) or is excluded by the many other competitors resident on the
islands.

Radiation has undoubtedly occurred in Macaronesian insect groups,
but data on this point are very difficult to secure in a clear fashion. Wolla-
ston’s data for Madeiran Coleoptera do give a number of hints in this
direction, but his information is of greater interest in underlining the

194

Adaptive Radiation: The Macaronesian Islands

occurrence of flightlessness in Macaronesian insects, and is discussed in that
connection (chapter 12). There does not appear to be any striking radia-
tion among Macaronesian reptiles, although interesting size differences can
occur in a single species represented on various of the islands, probably
for reasons discussed in similar instances in chapter 15.

Comment on Macaronesian Patterns

As mentioned above, Macaronesian islands do not seem remote enough
that strong disharmony can foster adaptive radiation in birds—whereas the
five passerine establishments on the ecologically rich Hawaiian Islands
have had opportunity for adaptive radiation.

There have, however, been situations of opportunity for angiosperms
on Macaronesian islands, especially on the Canary Islands. Unfilled niches
in the Canarian and Madeiran floras have been occupied largely by mem-
bers of herbaceous, “weedy,” evolutionarily upgrade groups, not woody
ones. All of the examples cited (Echium, Sonchus, Euphorbia sect. Tithy-
malus, Limonium, Aeonium, Leucophae, Micromeria, and Pterocephalus)
are herbaceous or, at most, “woody herbs.” These have radiated into all
zones. Interestingly, the laurel forest contains products of radiation in sev-
eral of these groups, so it obviously is not a “closed” community. How-
ever, lowland scrub and heather scrub do seem to contain most products
of radiation. Differences in topography (e.g., barrancos) and moisture
availability have provided more ecological opportunity in the lowlands
and, of course, the total area of lowlands is greater than that of the higher
zones. Alpine zones are relatively poor in products of adaptive radiation,
but not only is this a much more severe environment, the total land area
of this habitat in Macaronesia also is limited.

Ecological opportunity for angiosperms is considerable in Macaro-
nesia. Herbaceous groups seem favored for establishment in the lowlands,
as one might expect: this zone is where virtually all man-introduced weeds
have established. Lowland groups seem favored for rapid adaptation to
alpine conditions, as on the Hawaiian Islands. For example, the lowland
Macaronesian species Plantago arborescens (fig. 10.10, top left, right) has
a Close species, P. webbii (fig. 10.10, bottom left), in the alpine zone, but

195

Adaptive Radiation: The Macaronesian Islands

Plantago is not represented in the moister forest areas. Crucifers show this
condition clearly in Macaronesia. Ecological shift of Brassicaceae into wet-
forest areas has occurred to a minimal extent. The Hawaiian Islands exceed
Macaronesia both in disharmony and in ecological richness, so that greater
radiation would be expected in Hawaii. Compared with the Hawaiian
flora, the Macaronesian flora shows a somewhat truncated (on the mesic
end of the spectrum) adaptive radiation. Nevertheless, the Macaronesian,
and especially the Canarian, flora is second only to the Hawauan flora
among plant assemblages illustrative of adaptive radiation. The Galapagos
flora is far more truncated in mesic opportunities than is the Macaronesian
flora (chapter 6). In Macaronesian floras, there does not seem to be much
displacement of “older” groups by newer groups. The laurels and perhaps
the pine are “relictual” preservations, but they have neither radiated nor
been displaced (except as a result of human activities). The laurels are
“recent relicts,” for these volcanic islands are not truly ancient at all. The
conifers of New Caledonia (chapter 7) and New Zealand are much older,
and show both more pronounced relictism and also radiation into a variety
of ecological habitats.

References

Bannerman, D. A., and W. M. Bannerman. 1963. Birds of the Atlantic Islands, vol.
I: A history of the birds of the Canary Islands and of the Salvages. Edinburgh and
London, Oliver and Boyd.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1970a. Hawaii, a natural history. New York, Natural History Press.

Carlquist, S. 1970b. Wood anatomy of Echium (Boraginaceae). Aliso 7:183-99.

Lack, D. 1947. Darwin’s finches. Cambridge, Cambridge University Press.

Lems, K. 1960. Floristic botany of the Canary Islands. Sarracenia 5:1—94.

Lems, K. 1961. Botanical notes on the Canary Islands. III. The life form spectrum
and its interpretation. Ecology 42:569-72.

Lems, K., and C. M. Holzapfel. 1968. Evolution in the Canary Islands. I. Phylo-
genetic relations in the genus Echium (Boraginaceae) as shown by trichome devel-
opment. Bot. Gaz. 129:95-107.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography.
Princeton, N.J., Princeton University Press.

Mayr, E. 1942. Systematics and the origin of species. New York, Columbia Uni-
versity Press.

Praeger, L. 1932. An account of the Sempervivum group. London, Royal Horticul-
tural Society.

196

ADAPTIVE RADIATION: GALAPAGOS,
JUAN FERNANDEZ, AND SMALLER
PACIFIC ISLANDS

The Galapagos Islands

The Galapagos Islands do indeed represent many interesting biological
phenomena and have the advantage of relatively good preservation of
many elements in the biota, with a few notable exceptions. However, we
must remember that the Galapagos biota is not a rich one at all. In fact,
its impoverishment and simplicity permit insular patterns to be seen. The
Hawaiian Islands are probably richer in all insular phenomena. Their very
complexity might have proved confusing for Darwin, although I like to
think that if the itinerary of H.M.S. Beagle had included the Hawaiian
Islands rather than the Galapagos and allowed Darwin sufficient time to
study that biota, he might have emerged with a clearer picture of adaptive
radiation and other insular phenomena. However, the very clarity and
simplicity of the Galapagos biota rendered patterns clearly visible, and
perhaps we owe to this simplicity the fact that Darwin was able to make
effective observations there. Yet, as Lack (1947) reminds us, Darwin
himself was slow in realizing the significance of the Darwin’s finches.

To be sure, the Darwin’s finches (Fringillidae, subfamily Geospizinae)
are a good example of adaptive radiation, for reasons to be discussed.
They fulfill almost all of the definitions suggested in chapter 3. The only
Galapagos plant group that partially fulfills the criteria of adaptive radia-
tion is Scalesia (Asteraceae, tribe Heliantheae). Other plant and animal

197

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

groups suggest phases of adaptive radiation, but these phases are greatly
“foreshortened”’ by the nature of the Galapagos environment and by other
factors. In analyzing why only one Galapagos example shows “classic”
adaptive radiation and others only a few partial phases of this concept,
most none at all, we can learn considerably not only about the Galapagos
environment but about the nature of adaptive radiation as well.

Darwin's Finches. We could imagine a greater diversification in adap-
tive radiation of a group of birds than the geospizids show. For example,
the nectarivory of the Hawaiian honeycreepers is missing—and for a very
good reason. Flowering seasons are short in the Galapagos, and the few
flowers available during this season are not good nectar flowers (nor would
we expect them to be, if establishment of ornithophilous plants is related
to presence of bird pollinators).

However, one of the primary points of interest in the geospizids is
their presentation of a nearly unbroken series of intermediates between ex-
tremes (fig. 6.1). This is, in a way, more instructive than the adaptive radia-
tion of the honeycreepers, where we see (especially in the subfamily Dre-
panidinae) some obvious gaps, probably due to extinctions. Geospizid
adaptive radiation features the following characteristics:

1. There has been occupation of a variety of ecological niches, not
just a few. In terms of diet, we see a range (fig. 6.1), from 100 percent in-
sectivory to mainly plant feeding.

2. The morphology of bills and the ways in which they work form
external evidences not only of diet but also of the various mechanisms
(probing, grasping, biting, crushing), by which this diet is obtained. Bow-
man (1963) constructed appealing analogies to tools like pliers of various
types (see also Carlquist, 1965), but he has analyzed form and musculature
(1961) of the bill apparatus in a sophisticated fashion.

3. Likewise, color indicates ground-feeding (blacker tones of plum-
age), which is also related to herbivory. Tawnier colors are related to feed-
ing in branches of shrubs and trees, which in turn connotes insectivory.
This is clearly shown in the scheme of Bowman (1963), reproduced here
as fig. 6.2.

4. Theoretically the Darwin’s finches might achieve diversity yet be
able to occupy only a single vegetational zone, taking advantage of dif-

198

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

Geospiza
magnirostris

e
Sn £0

“Sp Ors

Te, <9

2s

Certhidea
olivacea

ZZ Plant food
mainly
FIG. 6.1. Schematic representation of the relationships between bill structure and feeding

habits in ten species of Geospizinae from Indefatigable I., Galapagos Is. Redrawn from
Bowman (1961); reproduced by permission of the University of California Press.

ferent resources within this zone. In fact, the Darwin’s finches occupy all
conceivable zones (Lack, 1947). Geospiza tends to occupy the arid and
transitional zones. Cactospiza, Platyspiza, and Camarhynchus tend to be
more inland and upland in preferences, ranging from transitional into the
humid Scalesia forest. Certhidea ranges among all zones. Pinaroloxias,
which occurs on Cocos Island and is thus a Galapagos outlier, occurs in
the coconut forests of that wet tropical island. Even in a single species,

199

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

HERBIVOROUS INSECTIVOROUS

.
GROUND LEVEL LU
Se vere] A

| Geospiza —_| | Platyspiza | Nesuainrete | Cactospiza_| | Certhidea |

WAVY

FIG. 6.2. Correlation of male plumage condition, foraging level, and diet in five genera of
Geospizinae. Redrawn from Bowman (1963); reproduced by permission of the California
Academy of Sciences.

100

75

ABOVE GROUND FORAGING

50

Time in percent

25

i

Male
plumages

Geospiza difficilis, some subspecies are lowland, some are upland in their
adaptations (Lack, 1947, p. 27).

5. The various genera of geospizids are not merely several relictual
types, but represent a more or less continuous spread of types. There are,
in addition, “unexpected” or “‘supralimital” adaptations, such as the tool-
using finch, Cactospiza pallida.

6. The various taxa of Darwin’s finches are equivalent to different
families on mainland areas. Bowman (1963) has offered the following
“equivalents” or comparisons:

Geospiza spp.: like mainland Fringillidae
Geospiza scandens: like Icteridae
Cactospiza pallida: like Thraupidae

200

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

Certhidea olivacea: like Parulidae

Pinaroloxias inornata: like Coerebidae
Camarhynchus psittacula: like Paradoxornithidae
Camarhynchus parvulus: like Paridae

If, instead of the single invasion of the geospizids, invasions by stocks rep-
resenting all of the families listed had taken place, radiation by the geo-
spizids would undoubtedly have been forestalled. However, land birds do
not disperse easily to oceanic islands more than several hundred miles
from the mainland (and as Lack notes, the Galapagos are in a region of
calms, where winds are not likely to bring birds), so the Galapagos land-
bird fauna is undoubtedly quite disharmonic. This has permitted radia-
tion of ancestral types of the Darwin’s finches.

7. Geospizids show secondary indications of diversification. These
are perhaps much like the opening stages of adaptive radiation, in which
character release occurs (see chapter 3). For example, Geospiza fortis is the
sole Geospiza on Daphne Major Island. It has a smaller beak than does
G. fortis on Charles and Chatham Islands, where G. fortis competes with
the characteristically small-beaked G. fuliginosa (Lack, 1947, p. 832). Geo-
spiza fuliginosa, in turn, has smaller beak size on Charles and Chatham
that it has on an island where G. fortis is absent. On Crossman Island, for
example, the beak of G. fortis has a broader range in size, for it is the sole
Geospiza there, and is not subject to the character displacement effect of
congeners on a single island.

8. Although the Galapagos Islands are “habitat poor’ for many
groups, they do represent a variety of habitats for a bird group such as the
finches, which show plasticity and versatility in diet capabilities. Most
other bird—or vertebrate—groups, if we could see habitat diversity
“through their eyes,” would not have a broad range of opportunities avail-
able to them.

Other Animal Groups. The other Galapagos groups do not show good
instances of what could be called adaptive radiation. The giant tortoises
exemplify race formation, but these races or subspecies do not appear to
differ in ecology, habitat, or diet (although obviously food sources are
more abundant, and different in species content, on the various islands,
there is no evidence that any of the tortoises has developed specific prefer-

201

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

ences for the plant assemblages of any particular island). The distinctively
shaped shells of the tortoise subspecies could be thought to represent the
product of genetic drift, or the result of the “founder principle.” However,
shell differences in tortoises at large can have selective value. In the case
of the Galapagos tortoises, rarity of most of the subspecies makes selective
advantage of the various shell shapes difficult to demonstrate, but this
possibility cannot be ruled out.

This subspeciation owing to genetic drift seems true of the various
island populations of the lizard genus Tropidurus, although further studies
may reveal differences among the various island populations of more than
a “strictly taxonomic’ nature.

There is no doubt that several insect genera have speciated appre-
ciably in the Galapagos environment. There may be aspects of adaptive
radiation among these, but, if so, the available literature is weak in eco-
logical data; one cannot find the basis for ecological distinctions that
might underlie adaptive radiation. The reasons for this lack are clear: most
Galapagos entomological studies are based on museum specimens rather
than on collection of extensive data in the field. Because of the peculiari-
ties of the Galapagos environment, field studies of entomology remain one
of the more promising avenues open for investigation.

Scalesia (Asteraceae). Adaptive radiation on oceanic islands often,
although by no means invariably, seems to proceed by ecological shift from
dry or relatively dry into more mesic sites. This seems true in Scalesia. The
reasons why this should have occurred are clear. Establishment of a plant
group in the Galapagos depends, with few exceptions, upon tolerance of
the dry conditions that occupy such a large area compared with the area
of humid places suitable for trees. Ecological shift in flowering plants on
islands from wet to dry is relatively infrequent, and if it were true in Sca-
lesia, one would expect diversity among the humid-forest species, com-
bined with penetration of arid areas by only a few species. In reality, the
reverse is true: numerous dry lowland species, with decrease in number of
Scalesia species with increasing altitude. The arborescent species, S. pedun-
culata (3-20 m tall) and S. microcephala (at most, 3.5 m tall), are excep-
tional for height in their tribe, Heliantheae. Although shrubs are not infre-
quent in Heliantheae, true trees are. Let us assume that arborescence is
secondary in Scalesia, as it probably is throughout the family Asteraceae.

202

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

The altitudinal distribution of Scalesia species is as follows (data from
Arthur Cronquist, in Wiggins and Porter, 1971):

LOWLAND SPECIES

S. aspera: “near the shore but in less barren, more mesic sites
than S. affinis.”

S. atractyloides: ‘“‘pioneer on and around recent lava flows.”

S. crockeri: “on bluffs and in lava crevices near the shore”
(S. crockeri is very close to S. aspera).

S. helleri: “pioneer on cliffs and raw lava, especially near the
shore.”

S. stewartii: “in barren lava beds and cinders, from near the
coast to above 200 m” (S. stewartii is very close to S. atrac-
tyloides).

S. villosa: “a pioneer on barren cinder slopes on lava beds.”

LOW-TO-MIDDLE-ELEVATION SPECIES

S. affinis: “a pioneer on lava beds and cinder slopes and ridges,
from sea level to about 600 m.”

S. incisa: “a pioneer of cliffs and raw lava to at least 520 m.”

MIDDLE-ELEVATION SPECIES

S. cordata: “relatively mesic sites (and apparently not on raw lava)
from about 50 m upward to about 400 m or higher.”

MIDDLE-TO-UPPER-ELEVATION SPECIES

S. microcephala: “at middle and upper altitudes, from about
350 m upward, in less barren habitats than S. affinis.”

S. pedunculata: “in relatively mesic or wet sites, not on raw lava,
from about 60 m upward, often forming considerable forests,
pure or mixed with other kinds of trees and shrubs.”

In addition to its greater stature and moist upland habitat, S. pedun-
culata shows adaptation to an arboreal and more nearly mesic role in its
relatively large leaves (3-30 cm long, 12 cm wide). Most other Scalesia
species have leaves that rarely exceed 10 cm in length. For photographs
of leaves of the various Scalesia species, see Carlquist (1965).

Opuntia (Cactaceae). In their account of this genus in the Wiggins
and Porter (1971) flora, Anderson and Walkington described Opuntia as
a “highly variable genus in the Galapagos, exhibiting rapid radiation that

203

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

has resulted in marked morphological differences and establishment of
several specific and infraspecific taxa within the archipelago.” Undoubtedly
the youth of the archipelago means that these species and subspecies (as
they should probably be called, although this Flora unfortunately
utilizes “variety” as the only infraspecific category) are relatively new.
However, do these taxa exemplify adaptive radiation? In some ways, yes.
Among the populations recognized as Opuntia echios, some are trees (to
12 m in “var. gigantea’), some low and prostrate (“var. zacana’). ‘The
typical populations (“‘var. echios’’) extend from sea level to 380 m or
higher, whereas some (“‘var. zacana’’) occur only near sea level. In O. mega-
sperma, the typical population (“var. megasperma’) occurs to only 50 m
on Charles Island and nearby Champion Island, whereas the Chatham
Island population (‘“‘var. mesophytica’) occurs in the upper transition zone
and Scalesia zone at about 250 m.

These examples and others, however, suggest in part vicarious species
and in part populations that have, during the process of isolation in dif-
ferent islands, adapted to different zones (although these zones, admittedly
different, are not comparable to the much broader ecological range pre-
sented by vegetational zones in the Hawaiian Islands). ‘The fact that there
are no truly mesophytic sites in the Galapagos would predispose this mod-
erate ecological shift. One could question whether Opuntia would be able
to enter true rain forest if it were present on the Galapagos Islands. The
Galapagos Opuntias show diversity, but do not qualify as good examples
of adaptive radiation (compared, for example, to Hawaian angiosperm
genera; see chapter 4), unless we wish to define adaptive radiation in terms
of the ecological diversity of the habitat, no matter how limited that
diversity. We clearly do not have to alter our concepts of adaptive radia-
tion where the Darwin’s finches are concerned.

Acalypha (Euphorbiaceae). Acalypha is represented in the Galapagos
Islands by endemic species in ecologically different sites (Wiggins and
Porter, 1971). We do not know if the species represent products of a single
introduction, which would be useful to know if we were to assess the
nature of adaptive radiation in this genus.

Other Genera of Angiosperms. In Chamaesyce (or Euphorbia sect.
Chamaesyce) of the Euphorbiaceae, the Galapagos species range from
herbs to subshrubs—but C. (E.) viminea is apparently introduced by man

204

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

(Svenson, 1963). This problem is repeated by a genus of another family,
Alternanthera (Amaranthaceae). Alternanthera rugulosa, a small tree (3—4+
m), would be notably arborescent compared to shrubby or prostrate Gala-
pagos species, and might have been cited as an end-product of radiation in
Alternanthera. However, A. rugulosa is not an autochthone; it has recently
been collected in Peru (Eliasson, in Wiggins and Porter, 1971). The Gala-
pagos Alternantheras must represent two (or possibly more) introductions.

Many Galapagos species in various families (e.g., Piscidia carthagi-
nensis, Cordia spp.) show a full altitudinal range from coast to uplands.
Thus all altitudes, despite differences in vegetation, may be considered
basically similar in being more or less xeric. Ecological shift by Galapagos
groups may occur, but the climatic range available for shifting may well
be within the tolerances of a single species.

Commentary. The synopses given show only two examples of adap-
tive radiation, in the best sense, in the Galapagos Islands: the geospizids
and (to a lesser extent) Scalesia. My discussion of the Darwin’s finches is
only illustrative and is not to be considered a substitute for the works of
Lack (1947) and Bowman (1961, 1963), which should be read by any seri-
ous student. The key to lack of adaptive radiation is the limited or “‘fore-
shortened” ecological span of the Galapagos environment. The amount of
rainfall in the “cloud forest” and the effect of the protracted cloud cover,
which may enhance the effect of available moisture, are not known. How-
ever, even the “cloud forest” is relatively dry. If this were not true, species
corresponding to those of mainland wet forest would be expected, but they
are notable by their absence (whereas wet-forest elements are present in
the Hawaiian flora, both as immigrants from wet-forest source areas and
as adaptations from drier to wetter areas). The Galapagos environment
must be considered, in comparison with climates at large, to be variously
xeric, with mere modifications of xeric environments. The geospizids are
not affected by the dryness because they can exploit this environment in
various ways: insect and plant foods are relatively varied for them, both
within a given altitudinal level and certainly within the gamut of altitu-
dinal levels. Although the Galapagos flora is disharmonic, it is apparently
far less disharmonic than the land-bird fauna. The Galapagos flora con-
tains an assortment of drier Middle American plant groups, yet these are
presented with relatively limited habitat diversity within the Galapagos.

205

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

In literature on the Galapagos Islands biota, one notes the implica-
tion that recentness of the archipelago may be a limitation to the develop-
ment of patterns of adaptive radiation, so that one might expect additional
radiation given more time. This is, to me, unlikely. The time elapsed on
the Galapagos seems sufficient for radiation into the available environ-
ments; it has certainly been sufficient for the geospizids to occupy what
are probably all the niches available to that group. Thus the factors sup-
pressing adaptive radiation in the Galapagos as compared with other in-
sular areas seem to be, in probable order of importance: lack of strong
ecological diversity; closeness to mainland areas; and geological recentness.

The Juan Fernandez Islands

An inherent interest of the Juan Fernandez Islands is that on these rela-
tively small islands there has been any adaptive radiation at all. The small
size of the two islands (Masafuera and Masatierra) and outlying islets (e.g.,
Santa Clara) seems too limited to promote adaptive radiation. Neverthe-
less, two groups of composites (Asteraceae) show what could be called
adaptive radiation. One group is from the tribe Senecioneae: Robinsonia
and genera close to it (Rhetinodendron and Symphyochaeta); by a wider
definition, all three genera could perhaps be considereda single genus. The
other example is Dendroseris, a genus of Lactuceae (Cichorieae). As rec-
ognized here, Dendroseris includes the segregate genera (which certainly
merit subgeneric status) Phoenicoseris, Hesperoseris, and Rea. For habit
illustrations of these, the reader can consult Skottsberg (1922, 1953) and
Carlquist (1965).

Although all species of Robinsonia, Rhetinodendron, and Symphyo-
chaeta could be considered shrubs, they vary considerably in habit, ranging
from a subshrub 1-2 m tall to a true tree, 5-6 m or more. The gamut in
growth form is not overwhelming, for all could be called rosette trees.
However, the habitats do vary from obligate epiphytism on tree ferns
(R. evenia) in very moist forest to growth on montane scrubland and open
cliff faces (R. gayana). I have liquid-preserved material of all the species
of Robinsonia and of the two genera closely allied to it. This material,
collected on more than one occasion by Carl Skottsberg, was kindly given

206

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

to me by him in 1956. Upon compiling habit and habitat data for the
species and then comparing these to transections of leaves (figs. 6.3, 6.4),
I noticed a remarkable degree of correlation. Therefore I have decided
to present data on leaf anatomy here rather than wait until a more
extensive account of comparative data from other portions of the plant
could be presented. Descriptions of habits and habitats are from Skottsberg
(1953). The species are arranged here from most to least mesomorphic,
based chiefly on anatomical considerations, but the correlation with habitat
appears so close that either basis could have been used.

Robinsonia evenia (fig. 6.3, top left). Masatierra: upper
montane region, seldom below 500 m; epiphytic on Blechnum
cycadifolium and Dicksonia tree ferns; 2—3, or even 4, m tall.
Leaves very thick; 1 hypodermis layer; 2-3 layers of elongate
palisade cells; thin-walled, spongy, chlorophyllous tissue with
large air spaces; fibers few or none on bundles; | abaxial secretory
canal per vein, on average.

Symphyochaeta macrocephala (fig. 6.3, center right). Masa-
tierra: Puerto Frances to Pangal, 300 to 350 m; same habit as
Robinsonia, but more straggling; 2 m or less. Leaves very thin;
1 layer of hypodermis (2 near large veins); 2 layers of elongate
palisade cells; thin-walled, compact, chlorophyllous, spongy tissue;
no fibers adjacent to veins; 1 abaxial and 1 adaxial secretory canal
per bundle, on average; all cells of the leaf rather small.

Robinsonia masafuerae (fig. 6.3, top right). Masafuera: 650-
1370 m, preferring fern-rich gullies; hardly surpassing 2 m; closely
related to R. evenia, and sometimes epiphytic. Leaves medium
thin; 1 hypodermis layer; 2 layers of elongate palisade cells; thin-
walled, chlorophyllous, spongy tissue; some fibers adjacent to larger
bundles; few or no abaxial secretory canals adjacent to bundles.

Robinsonia gracilis (fig. 6.3, bottom left). Masatierra: ridge
between Piedra Agujemada and Laura Valleys to Cordén Sal-
sipuedes, never below 500 m; not often epiphytic, although
seedlings may start as epiphytes on Blechnum and Dicksonia;
1-2 m tall. Leaves thin, 1 layer of hypodermis; 2 layers of rather
cuboidal palisade; spongy tissue chlorophyllous, but with some

207

FIG. 6.3. Leaves of the Juan Fernandez rosette-shrub Senecioneae (Asteraceae) in transection,

adaxial face above. TOP LEFT,

R.

Robinsonia eyenid. BOTTOM LEFT, R. masdfuerde. TOP RIGHT,
Symphyochaeta macrocephala. BorTroM RIGHT, Rhetinodendron

Leaf of R. evenia is 0.8 mm thick,

scale of mas

HT,

CENTER RIG

gracilis.

eaves are photographed at the same

]

all

Carl Skottsber

and
208

berteri.

24
5-

1 by

s furnished

]

mification. Materia

5

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

resemblance to hypodermis; fibers at phloem poles of bundles
absent or nearly so; 1 abaxial secretory canal per bundle.

Rhetinodendron berteri (fig. 6.3, bottom mght). Masatierra:
open brushwood and forest along the ridge, rarely below 500 m,
of the elevated central part of the island; shrubs 2—4 m, trees
of 5 m occasional, one tree 8 m observed. Leaves medium thin;
2 layers of hypodermis; | layer of palisade; thin-walled, spongy
tissue on abaxial side resembling hypodermis cells; fibers on
bundles few or none.

Robinsonia gayana (fig. 6.4, left). Masatierra: montane scrub,
descending to about 200 m, often in crevices of cliffs; shrub to 3,
rarely 4, m tall. Leaves medium thin; 1 layer of hypodermis;
2 (—3) layers compact palisade; spongy tissue on abaxial side
thick-walled and very much like hypodermis cells; fibers present
at phloem poles (abaxial side of leaf) of veins; 1 abaxial and
1 adaxial secretory canal per bundle, on average.

Robinsonia thurifera (fig. 6.4, mght). Masatierra: slope of :
Cordon Chifladores above Puerto Frances, 350-450 m; Portezuelo

Fic. 6.4. Leaves of “less mesic” species of Juan Fernandez rosette-shrub Senecioneae (Aster-
aceae) in transection, adaxial face above. Lert, Robinsonia gayana. ricut, R. thurifera. Scale
of magnification as in fig. 6.3.

209

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

de Villagra, 550-600 m; to 5 or 6 m tall, consistently the largest
species of the genus; flowers later than any other tree or shrub
in the islands. Leaves medium thick; 1-2 layers of hypodermis;
2 layers of compact palisade; abaxial spongy tissue wide, notably
thick-walled and hypodermis-like; a few fibers adjacent to phloem
poles of veins; 1 adaxial and 1 abaxial secretory canal per vein.

To those not acquainted with concepts of mesomorphy and xero-
morphy in leaf anatomy, the following can be offered. Hypodermis, a
characteristic of brightly illuminated situations, tends to increase with
exposed situations, but there is also, in the Robinsonia group, a curious
conversion of spongy tissue to a sort of second hypodermis in the less
mesomorphic species. In effect, this makes the leaves much like isolateral
leaves, which often characterize brightly illuminated localities (often also
correlated with positioning of the leaf in manners other than horizontal,
as in some of the Eucalypts). Increased presence of secretory canals with
greater illumination and “xeromorphy,” although not uniform, seems
evident. Secretory canals and cavities are most abundant in composites
of dry localities, such as those of the southwestern United States. While
thinness would be expected to be a mesomorphic characteristic—as it is
in Symphyochaeta macrocephala, the thick leaves of Robinsonia evenia
are decidedly mesomorphic, for the thickness is not the result of compac-
tion of cells but of a very large air-space system. There is a tendency, with
decreasing mesomorphy, for increase in presence of fibers near bundles.
Fibers tend to be associated with a more leathery leaf texture and
characterized leaves of greater duration in sunnier localities.

The habitat can be called less mesomorphic if the species grows in
scrub rather than rain forest. Large size (R. thurifera) and growth on cliff
faces (R. gayana) would connote less mesomorphy, for there is greater
exposure to illumination under these circumstances. Epiphytism on tree
ferns (R. evenia, R. gracilis), the straggling habit (Symphyochaeta) and
association with ferns (R. masafuerae), all within cloud forest, are maxi-
mally mesic habitats and mesomorphic habits.

In the other group to be considered, Dendroseris (sensu lato), very
thick leaves with large hypodermis cells, 2-3 cells thick, can be said to
characterize less mesomorphic species. Such features (Carlquist, 1967)

210

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

occur in D. litoralis and D. (Rea) pruinata. Dendroseris litoralis grows
only on lowland cliffs of Morro Vinillo (an islet near Masatierra), on a
rather low islet, Santa Clara Island, and on Morro del Spartan, also a
low islet near Masatierra. Dendroseris (Rea) pruindta is a near-maritime
species, scattered in dry ravines of Masatierra and Santa Clara (Skottsberg,
1953). The rain-forest species of Dendroseris have much more meso-
morphic types of leaf construction, and the reader can consult Carlquist
(1967) for further information. Because Dendroseris occupies a wider
range of ecological situations than does Robinsonia and allies, it can better
be cited as an example of adaptive radiation—especially when one takes
into account the greater range of growth forms in Dendroseris (sensu
lato). However, the Robinsonia group clearly shows a range of diversifica-
tion that can be considered early or middle stages in adaptive radiation
(or perhaps better, adaptive radiation within limits).

Other Fernandezian genera certainly show alteration in growth form
and ecological shift, and one would cite them in connection with adaptive
radiation if only they represented a series of species instead of merely
one or two, and thereby could occupy several distinctive ranges of habitats.
It is conceivable that some of the genera may once have had more
numerous species, but diversification seems relatively recent in the Juan
Fernandez Islands, and it is, I feel, no accident that adaptive radiation is
represented in the most “active” family evolutionarily, the Asteraceae.

Other Pacific Islands

We do not yet know enough, to judge from the literature, about adaptive
radiation in the biotas of certain islands, such as in the ants of Fiji, or
(as a similar example) the frogs of the Seychelles. Where several distinctive
species occur on a small island, such as the three species of Zosterops on
Lord Howe Island, we may suspect, in agreement with Lack (1947), that
these represent independent colonizations, not autochthonous speciation.
To be sure, coexistence on a single island does tend to force minimal
overlap in habits (and, in the case of Zosterops, bill morphology) so that
a situation simulating adaptive radiation has been achieved.

We might have expected that the various Polynesian high islands—the

JANI

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

Marquesas, the Society Islands, and Samoa—would bear examples of
adaptive radiation. For the most part, they do not. Perhaps there are only
two good examples among flowering plants. One of these is Fitchia
(Asteraceae). It includes a tree in wet forests of Rarotonga, F’. speciosa
(Carlquist, 1957), but can also be a small shrub, as in F. cordata on the
scrubby ridges of Bora Bora (Carlquist, 1965; Carlquist and Grant, 1963).
The species of Tahiti are intermediate. One can cite Oparanthus (Aster-
aceae) as an example of adaptive radiation in southeastern Polynesia,
but its range is not as great as that of Fitchia, although there are interesting
divergences in leaf anatomy (Carlquist, 1957). On Samoa, the endemic
genus Sarcopygme (Rubiaceae) has several species, but they do not appear
to show distinctive ecological modes.

The land snail Partula, made famous by Crampton as a case of
speciation on insular areas, does not really illustrate adaptive radiation.
The habits of the various species seem quite similar, with geographic
isolation the main feature of speciation. The archipelagos of southern
Polynesia evidently consist of islands too small or too recent to have good
“archipelago effects” which seem so often to lead to programs of adaptive
radiation. These islands appear to have received that portion of the
Indo-Malesian (and, in the case of Fitchia and Oparanthus, American)
flora that excels at long-distance dispersal, and to nave provided isolation
sufficient for formation of endemic species but not space enough (particu-
larly with respect to sufficient areas in various ecological zones, especially
wet forest) for adaptive radiation to have occurred. To be sure, there are
interesting endemic genera of flowering plants (e.g., Apetahia, Sclerotheca)
or land snails in southern Polynesia, but if these represent products of
adaptive radiation, the related or ancestral forms are now extinct, and
they stand isolated. We do not know how much the older Leeward Islands,
now atolls, of the Hawaiian chain may have contributed to adaptive
radiation now visible on the major Hawaiian islands, but the presence of
a sequence (and we should note that the Leeward Islands form a rather
continuous sequence geologically, leading into the newest and highest
eastern islands) in the Hawaiian chain may well have been one of the
factors leading to the remarkable adaptive radiation in the Hawauan
biota. Such sequences are not present in southern Polynesia, although
older groups of islands, now atolls (the Tuamotus, for example), are

212

Adaptive Radiation: Galdpagos, Juan Fernandez, and Smaller Islands

present there. Consequently, the lack of examples of adaptive radiation
in southern Polynesian islands seems directly related to lack of extensive
land areas in appropriate ecological zones. In the case of Fi, there is
relatively little adaptive radiation mentioned; perhaps authors have not
discussed the Fijian flora in this respect, or the relatively harmonic nature
of the Fijian biota (compared with those of more remote islands) may
have forestalled adaptive radiation there.

References

Bowman, R. I. 1961. Morphological differentiation and adaptation in the Galapagos
finches. Uniy. Calif. Publ. Zool. 58:1—302.

Bowman, R. I. 1963. Evolutionary patterns in Darwin’s finches. Occas. Papers Calif.
Acad. Sci. 44:107—40.

Carlquist, S. 1957. The genus Fitchia (Compositae). Univ. Calif. Publ. Bot. 29:
1-144.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1967. Anatomy and systematics of Dendroseris (sensu lato). Brittonia
19:99-121.

Carlquist, S., and M. L. Grant. 1963. Studies in Fitchia (Compositae). Novelties
from the Society Islands, anatomical studies. Pacific Sci. 17:282-98.

Lack, D. 1947. Darwin’s finches. Cambridge, Cambridge University Press.

Skottsberg, C. 1922. The phanerogams of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:95-240.

Skottsberg, C. 1953. The vegetation of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:793—-960.

Svenson, H. K. 1963. Opportunities for botanical study on the Galapagos Islands.
Occas. Papers Calif. Acad. Sci. 44:53-58.

Wiggins, I. L., and D. M. Porter. 1971. Flora of the Galdpagos Islands. Stanford,
Calif., Stanford University Press.

DN)

CHAPTER
SIE VIN

ADAPTIVE RADIATION:
NEW CALEDONIA AND NEW ZEALAND

New Caledonia and New Zealand share certain biogeographic character-
istics: both are islands that have been in existence for longer periods of
time than the volcanic islands of the Pacific; they contain various ancient
elements, some of such poor dispersibility (Araucariaceae) that very early
contact or near-contact with other land masses is suggested. However,
both have acquired most of their biotas via long-distance dispersal. This
conclusion is suggested not only by the disharmony in their biotas, like that
of the larger volcanic islands of the Pacific, but by the “‘non-primitiveness”
of many of their families and genera (e.g., the birds on New Caledonia,
with the exception of the kagu). The insular phenomena they exhibit are
so much like those on, say, the Hawaiian Islands, that similar processes
of waif-biota evolution must have been operative, albeit for longer periods
of time. However, some elements in the biotas of New Zealand and New
Caledonia may be very recent—there is nothing to suggest that over-water
dispersal to them has ceased. In any case, both New Caledonia and New
Zealand have an admixture of old and new elements, and we can see
adaptive radiation in both categories.

New Caledonia

The flora of New Caledonia appears fairly harmonic (especially considering
its curious rock and soil types and moderately dry climate) but the fauna
is devoid of mammals and is essentially oceanic (see Thorne, 1965, and

214

Adaptive Radiation: New Caledonia and New Zealand

references cited by him). The reason for this ironic situation has mystified
some, but the explanation seems, in fact, rather simple. The fauna
probably did reach the island over oceanic distance. ‘That New Caledonia
was once a larger island and probably was, at a remote time, separated by
shorter distances from other land masses southwest of the andesite line
seems likely. If this is true, ancient floristic elements—the Araucariaceae
in particular—may have arrived at that time, whereas marsupials and
certainly placentals were not present then in the land mass contributing
immigrants to New Caledonia. Although Australia, including New Guinea,
did receive marsupials and monotremes, New Caledonia was not connected
with (or was farther from) the Malesian “mainland” at that time. Because
the animal groups present no particular problems (although there are some
interesting features in the reptiles), the nature of diversification and
radiation in the New Caledonian flora has been chosen for examination
here.

Only the Araucariaceae (Araucaria and Agathis) appear to have such
poor dispersibility that they would require short over-water distances or
land connections for arrival in New Caledonia. Araucarias such as A. cookit
obviously do disperse within New Caledonia from one suitable site to
another, so there is no reason to assume that they are incapable of short-
range dispersal. The New Caledonian genus of Taxaceae, Austrotaxus, 1s
far removed geographically from other Taxaceae, which may have receded
since this stock reached the present land mass of New Caledonia. The
Podocarpaceae, like the preceding gymnosperms, would appear primitive
and not easily dispersible. Acmopyle, with large seeds (fig. 7.6, top left),
is extant only on New Caledonia and Fiji. Dacrydium, with relatively
small seeds, would seem more readily dispersible. However, Dacrydium
seeds lack the fleshy cone, attractive to birds, that Podocarpus seeds have,
and are probably more poorly dispersed, as the narrower range of Da-
crydium compared with that of Podocarpus seems to confirm. Dacrydium
does not extend farther eastward into the Pacific than Fiji. Interestingly,
D. lycopodioides (fig. 7.4, bottom right) occurs on both New Caledonia
and Fiji. Dacrydium once had a wider extent, for fossils indicate that it
occurred in India (Jurassic), Argentina (Eocene) and Antarctica (Oligo-
cene), according to Florin (1963). Not only do Agathis, Acmopyle, and
Dacrydium among gymnosperms reach their current easternmost Pacific

25

Adaptive Radiation: New Caledonia and New Zealand

station on Fiji, but also many genera of angiosperms stop there (Smith,
I 5>));

I see no reason why New Caledonian flowering plants cannot be
hypothesized to have arrived via long-distance dispersal or, in a few cases,
shorter-distance dispersal. To be sure, there are relatively “primitive”
angiosperm groups, such as the Winteraceae and a monogeneric family,
Amborellaceae (fig. 7.6, top right). However, Amborella has small, red,
fleshy fruits that seem capable of bird dispersal, and fruits of Drimys of
the Winteraceae (fig. 2.4, bottom right) seem eminently dispersible—
although these groups now may well be relatively specialized in habitat
requirements and no longer able to succeed readily at establishing. One
endemic New Caledonian genus of Winteraceae, Zygogynum, is probably
an example of autochthonous loss of dispersibility (Carlquist, 1965). There
are other families well represented in the New Caledonian flora that could
be called primitive, such as Cunoniaceae (fig. 7.6, bottom left, right).
However, the fact that genera of Cunoniaceae have a number of primitive
features does not mean that they are poor at long-distance dispersal. The
contrary is true: Weinmannia occurs on oceanic islands such as Tahiti.
With the exception of some of the conifers, then, the New Caledonian
flora may well have resulted from long-distance dispersal. However, the
various establishments probably occurred over a relatively long period of
time. T’o say that there are “old arrivals” and “new arrivals” might be
misleading, but we can make some reasonable guesses. For example, one
can cite those taxonomic groups in the New Caledonian flora that also
occur on volcanic islands such as ‘Tahiti, Samoa, or the Hawauan Islands.
The roster of such families in the New Caledonian flora, gymnosperms
excepted, is not extraordinarily different from families present on more
remote volcanic oceanic islands.

“OLD” ADAPTIVE RADIATION

For the reasons cited in the preceding section, we can assume that the
conifers represent early immigrants to New Caledonia. The Araucariaceae
are rather interesting in this respect. Agathis occurs on New Caledonia,
Fiji, New Zealand, Queensland, the Malesian islands, and Malaya, but
does not show what could be called adaptive radiation within any of these

216

Adaptive Radiation: New Caledonia and New Zealand

localities or even in all localities considered as a whole. Araucaria on
New Caledonia is quite different: there perhaps eight species occur.
Although some of these are close, the gamut of New Caledonian species
represents almost as broad a range as the entire remainder of the genus.
New Caledonian Araucarias tend to occupy distinctive localities. Araucaria
cookii (fig. 7.1, top left) occurs near the coast, often in valleys where
seepage of fresh water occurs. In this way, it can even occur just behind
the beach and actually can come into contact with strand plants such as
Pandanus and Cerbera. The leaves of A. cookii are small (5 < 4 mm); the
habit of the tree is slender and conical. Araucaria bernieri (fig. 7.1, top
right) is very narrow, but not conical; mature trees tend to have a terminal
crown of horizontal branches. The habitat of A. bernieri is the Montagne
des Sources and low adjacent valleys on the Plaine des Lacs. A close species,
A. balansae, grows on higher ground; its leaves are extremely narrow
(4 2 mm). Araucaria muelleri (figs. 7.1, bottom right; 7.2, top left)
occupies much the same sites as A. bernieri (although it tends to grow
more on mountains around the Plaine des Lacs). The leaves of A. muelleri
(fig. 7.2, top left) are exceptional in size (30 X 20 mm), much like those
of the Chilean species A. araucana, but markedly incurved. Trees of A.
muelleri are distinctive because of their paucity of branches. Araucaria
rulei (fg. 7.1, bottom left) occupies cloud forest of higher mountains,
such as Mt. Mou. Other New Caledonian species include A. humboldtensis,
which has the habit of Dacrydium araucarioides, and grows only on Mt.
Humboldt. Araucaria biramulata, distinctive in its bifurcate branches
(Sarlin, 1954), grows only at the Mois de Mai mine on the Plaine des
Lacs. Araucdaria montana occurs on mountaintops in northern New
Caledonia. It has rounded leaves (12 X 8 mm). The habit of mature
trees of A. montana is very narrow, like that of A. bernieri; when younger,
trees look like those of A. rulei.

Supposing that all of these species are, in fact, valid and distinctive,
one is faced with the question, Could all of these species have originated
on an island the size of present-day New Caledonia (6360 sq. mi or
17,000 sq. km, one-seventh the size of Cuba)? The size of populations of
most Araucaria species seems far below the breeding size of all but a few
conifer species—and those could be called declining or relictual species.
This suggests that diversification of the neocaledonian Araucarias probably

217

FIG. 7.1. Araucaria (Araucariaceae) species from New Caledonia, showing distinctive habits
and habitats. rop Lert, A. cookii; small valley at coast, near Plum. Top ricHT, A. bernieri;
upper Riviére de Madeleine, Plaine des Lacs. sorrom LEFT, A. rulei; summit of Mt. Mou.
BOTTOM RIGHT, A. muelleri, lower Riviere de Madeleine, Plaine des Lacs.

218

Conifers from New Caledonia. trop Lert, Araucaria muelleri, foliage and young
cone. TOP RIGHT, BOTTOM LEFT, RIGHT, Neocallitropsis comptonii (Cupressaceae);
s of Riviere de Madeleine. ror ricut, habit, tree about 4 m tall. Bsorrom Lert, foliage
1 female cones. BOTTOM RIGHT, seed, showing capability of flotation.

female

IN)

Adaptive Radiation: New Caledonia and New Zealand

took place on a larger land mass—not a continent, but a “larger New
Caledonia.”

If three species of Agathis (Araucariaceae) and three of Libocedrus
(Cupressaceae) can be recognized as endemic to New Caledonia, as
claimed by Sarlin (1954), these other conifer genera reinforce the situation
just cited for Araucaria. Likewise, the New Caledonian species of Dacryd-
ium are interesting because one of them, D. lycopodioides (fig. 7.4, bottom
right), also occurs on Fiji. Dispersal of D. lycopodioides across the present
distance between New Caledonia and Fiji is not impossible, and lack of
differentiation between the New Caledonian and Fiian populations sug-
gests this has occurred, for a long period of separation would be expected to
result in at least some differentiation. The genus Dacrydium (Podocar-
paceae) may have radiated on New Caledonia, but this radiation again
seems to have taken place early. My reason for this conclusion is not
merely that gymnosperms are an ancient group. Rather, I would guess
that if Dacrydium shares, as it seems to, the slow-evolving characteristics
of most contemporary conifers, radiation would not have occurred if
numerous stocks of angiosperms had been present, competing for the
niches Dacrydium now occupies on New Caledonia.

The diversity of the New Caledonian Dacrydiums is exceptional—
a range in form exceeding that of Dacrydium elsewhere. Dacrydium
araucarioides (fig. 7.3, top left, right) appears to be a xeromorph. It grows
on purple-brown, highly mineralized soils as an element of a semi-open
scrubby woodland. Leaves are narrow, incurved, and imbricate; branches
are thick (fig. 7.3, top nght). At an opposite extreme is D. taxoides (fig. 7.3,
bottom left). Its leaves, although basically arranged in a spiral, are
positioned so as to form flattened sprays; both surfaces appear identical,
and the leaves are broad, looking quite unlike the needlelike leaves of
most Dacrydium species. Dacrydium taxoides is a small tree, shaped
irregularly by the other understory elements in the cloud-forest areas
where it characteristically grows. The sprays of broad, flat leaves seem
related to shaded conditions and to greater availability of moisture. Also
in a cloud-forest area is D. lycopodioides (fig. 7.4, bottom right). Judging
from juvenile foliage of other Dacrydiums, D. lycopodioides appears almost
like a permanently juvenile form. This would not be unexpected, for it
grows in shady levels of the Mt. Mou cloud forest. The leaves are narrow,

220

FIG. 7.3. Dacrydium (Podocarpaceae) species from New Caledonia. Top LEFT, RIGHT, D.
draucarioides. TOP LEFT, habit of tree, about 5 m tall; western Plaine des Lacs. ToP RIGHT,
foliage and female cones, inconspicuous, 2-seeded, at tips of branchlets. BorromM LEFT, D.
taxoides, foliage bearing a female cone with one seed; leaves 2—3 cm long; summit of Mt.
Mou. sotrom ricut, D. balansae, branches showing male strobili and female cones; near Thio.

7jb}\

FIG. 7.4. Dacrydium (Podocarpaceae) species from New Caledonia. TOP LEFT, RIGHT, BOTTOM
LEFT, D. guillauminii; Riviere de Madeleine, Plaine des Lacs. rop LEFT, habit, trees at edge
of river, about 2 m tall. rop ricut, foliage, including female cone, 2-seeded, at tip of branch-

let. BOTTOM LEFT, seeds, showing capability of flotation. BoTtomM ricHT, D. lycopodioides,
foliage; Montagne des Sources.

222

Adaptive Radiation: New Caledonia and New Zealand

densely imbricate, borne on slender twigs. Dacrydium balansae (fig. 7.3,
bottom night) has foliage not unlike that of D. araucarioides, but D.
balansae is densely branched and is characteristic of more mesic locations
than is D. araucarioides.

An expected adaptation is represented by D. guillauminii (fig. 7.4,
top left, nght, bottom left). This plant forms small trees—some only 2 m
tall at maturity—along the banks of rivers on the Plaine des Lacs. The
branches are few and thick, the foliage needlelike, appropriate to the
sunny climate of the Plaine des Lacs. Daerydium guillauminiu shows
obvious adaptations to the riparian habitat. The trunk is wide below,
tapering upwards, a form characteristic of some woody plants that with-
stand occasional flooding. Wood of the broad base is exceptionally hght
in texture and weight (tracheids are very thin-walled; intercellular spaces
are present), although wood of upper portions of the tree is denser. Not
unexpectedly, seeds of D. guillauminii float (fig. 7.4, bottom left), either
fresh or dried.

An obvious question—which might seem to vitiate at least part of
the preceding account—is how much radiation of Dacrydium has taken
place autochthonously on New Caledonia. Some seems clearly to have
done so, although we must remember that in addition to the present
distribution of the genus (Tasmania, Fiji, New Caledonia, New Zealand,
Malesia, southeast Asia, Chile) Dacrydium is known from fossils in Aus-
tralia, India, and possibly Antarctica (Florin, 1963). However, the great
diversity of the New Caledonian Dacrydiums and their apparent adapta-
tions to unusual and highly specialized environments suggests that virtually
all the peculiar features of the New Caledonian Dacrydiums are autoch-
thonously evolved.

The riparian syndrome shown in D. guillauminii is also represented
in the genus Podocarpus in New Caledonia by P. minor (= P. palustris).
Podocarpus minor (fig. 7.5, top left, right, bottom left) and D. guillauminit
ate the only conifers that show these marked adaptations. Podocarpus
minor becomes a little larger than D. guillauminii, is more richly branched,
and has oval leaves, which are, however, very thick, almost succulent. The
exceptionally large seeds do float (fig. 7.5, bottom left). Although nine
species of Podocarpus occur on New Caledonia, they do not represent
radiations from a single stock. As shown by Florin (1963), Podocarpus

223

FIG. 7.5. Podocarpus (Podocarpaceae) species from New Caledonia. TOP LEFT, RIGHT, BOTTOM
LEFT, P. minor (= P. palustris). ror Lert, habit, tree 3 m tall; floodplain of Riviere de
Madeleine, Plaine des Lacs. rop r1cuT, foliage, showing seeds 3 cm long. BOTTOM LEFT, seeds,
showing capability of flotation. Borrom ricHT, P. ustus, close-up of mounted herbarium
specimen, showing cones with immature seeds; foliage would be red in living plants of this
parasitic species.

224

Adaptive Radiation: New Caledonia and New Zealand

on New Caledonia is represented by the following subgenera and sections:
Stachycarpus, Microcarpus, Polypodiopsis, and Podocarpus sections B, D,
and F. This representation on such a small area suggests a remarkable
degree of relictism. However, some autochthonous radiation in Podocarpus
is probable. In addition to the mparian P. minor, the genus is represented
on New Caledonia by the world’s only parasitic gymnosperm, P. ustus
(fig. 7.5, bottom nght). These two extreme modifications appear autoch-
thonous. There are moderate streams and lakes for P. minor, certainly,
and the abundance of Podocarpaceae as host plants for P. ustus would
have favored origin of a parasitic species on this island.

On New Caledonia there are two monotypic endemic conifer genera,
Neocallitropsis (fig. 7.2, top night, bottom left, nght) and Austrotaxus.
Neocallitropsis seems related to Callitris and Actinostrobus of Australia
but, if so, retains characters primitive for the group. Neocallitropsis seems
to grow near streams (but not on the immediate margins), and floatability
of its seeds (fig. 7.2, bottom right) suggests adaptation. Austrotaxus is not
closely related to other genera of Taxaceae and is the sole southern-
hemisphere representative of the family, except for Taxus in Sulawesi
(Celebes). The large leaves of Austrotaxus, similar to those of Podocarpus
macrophylla, seem adapted to cloud forest. Neocallitropsis and Austrotaxus
underline the harmonic quality of the conifer flora of New Caledonia.
In fact, a more harmonic conifer flora—considering the small size of New
Caledonia—would be hard to imagine. The conclusion seems inescapable
that conifers were present on New Caledonia at an early time. The peculiar
soil types of New Caledonia would not even seem suited for conifers, but
the New Caledonian conifers obviously have adapted, and in the case of
Araucaria, Dacrydium (and, to some extent, the other genera) are not only
adapted to the unique climatic and edaphic regimes of New Caledonia,
but have speciated as well. The speciation and adaptation probably
occurred at a time when conifers and taxads were genetically a more plastic
and rapidly evolving group, and niche occupation by conifers and taxads
was relatively unopposed by angiosperms. The lesson of the northern
hemisphere is that angiosperms have gradually displaced gymnosperms,
and that gymnosperms retain dominance mostly at higher latitudes and
higher elevations. In the southwestern Pacific we see some replication of
this phenomenon. Consulting the distribution maps of fossil and living

225

Adaptive Radiation: New Caledonia and New Zealand

conifers by Florin (1963), we see that mainland Australia retains some
living conifers: Araucaria, Agathis, numerous Cupressaceae (Callitris and
Actinostrobus), and a few Podocarpaceae (Podocarpus and Microstrobus).
However, genera one would expect to find are absent. Phyllocladus
(Podocarpaceae) ought to be on mainland Australia today, for it is on
islands recently connected to Australia: ‘Tasmania and New Guinea. The
same applies to three genera confined to Tasmania: Diselma (Cupres-
saceae), Athrotaxis (Taxodiaceae), and Microcachrys (Podocarpaceae).
Likewise, Papuacedrus (Cupressaceae) is confined to New Guinea. Groups
absent from Australia but present on New Guinea and nearby islands
include Podocarpus sections Nageia, Polypodiopsis, and Dacrycarpus, as
well as Dacrydium. The most obvious explanation seems to be that
Australia has been a large land area that has served as a “‘target” for
angiosperm groups that have dispersed across oceanic distances. Moreover,
this area has functioned to increase genetic diversification, and aggressive
angiosperm groups (especially those adapted to dry areas, which most
conifers are not) have been so successful that many gymnosperm groups
have been unable to retain their niches.

“NEW” ADAPTIVE RADIATION

There seems to have been abundant time for transoceanic dispersal
of angiosperm groups to New Caledonia. All of the “successful” groups
in the New Caledonian flora seem suited to long-distance dispersal. Among
those with small seeds capable of wind dispersal are Hibbertia (Dil-
leniaceae), Nepenthes (Nepenthaceae), Casuarina (Casuarinaceae), all
Fpacridaceae, all Cunoniaceae, and capsular Myrtaceae. Araliaceae, Goo-
deniaceae (fig. 7.7, top left, right, bottom left), and many other New
Caledonian elements are clearly suited to dispersal by frugivorous birds.
Although some families in New Caledonia have large-seeded, fleshy fruits,
such as Sapotaceae and Clusiaceae, this is not as great an obstacle as it
might seem. Sapotaceae have reached, by natural means, such remote
islands as those of Hawaii (where Planchonella fruits exhibit various
degrees of gigantism; see chapter 11). For the reasons stated in chapter 11,
increase in fruit and seed size does not seem unlikely in New Caledonia,
where other examples of loss of dispersibility can be found (e.g., Zygogynum
of the Winteraceae).

226

Adaptive Radiation: New Caledonia and New Zealand

Radiation of over-water immigrants depends on the range of available
habitats. New Caledonia would not seem as ecologically diversified as,
say, the Hawaiian Islands. Edaphic diversity is great. New Caledonia has,
as the rather intense mining indicates, areas rich in chromium, nickel, and
other heavy metals. Uplifts of limestone are not uncommon. There are
distinctive pockets of serpentine. The broad lowlands seem parched most
of the year, but in cloud-forest areas and valleys there are quite mesic habi-
tats. The cloud forests, however, are subject to prolonged periods of sun-
shine, so that canopy shrubs and trees must be able to withstand desicca-
tion. The soils, so unlike those typical of Indo-Malesian source areas that
potentially could contribute to the New Caledonian flora, undoubtedly
have prevented establishment of numerous groups of plants (and perhaps,
directly and indirectly, of some animal groups as well). ‘These conditions
theoretically offer varied niches to each successful group; thus more abun-
dant adaptive radiation might be expected. However, even with the pecul-
iar conditions of New Caledonia that have undoubtedly deterred establish-
ments, the many years available for establishments have overcome this
bottleneck, and the New Caledonian flora is less disharmonic (and there-
fore poorer in adaptive radiation) than one might expect, compared with
Southeast Asian angiosperms.

The amazing habital similarity of New Caledonian angiosperm taxa
probably relates to the relatively foreshortened ecological diversity. There
is a modal tendency for smaller, rather sparsely branched shrubs, 2-8 m
tall, with very leathery elliptical leaves 5-20 cm long. Numerous examples
of these can be seen in the angiosperms illustrated in figures 7.6, 7.7, and
7.8. Anatomical studies would probably reveal parallel construction of
leaves among various groups, and are worth the effort. Because of the
“adaptive mode” of these angiosperms, one should keep open the possi-
bility that ancestral habits in these groups were not the same as the present-
day sclerophyll species.

Despite this mode, there is a range apparent in larger genera. In Hib-
bertia, for example, there are New Caledonian species with smaller, sclero-
phyllous leaves. However, H. badouinii (fig. 7.7, bottom right) is not a
shrub but a rosette tree to 10 m, with clusters of long, membranous leaves
at tips of branches. It is much like rosette trees in rain forests of oceanic
islands, such as Cyanea (Campanulaceae) on Hawaii or Robinsonia or Den-

227

Plants from New Caledonia. top LEFT, Acmopyle pancheri (Podocarpaceae), foli-
g one bearing a seed 2 cm in diam.; summit of Mt. Mou. rop ricut, Amborella tricho-
poda (Amborellaceae), tv fruits, each ca. 1 cm long; Plateau de Dogny. Borrom
LEFT, Pancheria insignis (Cunoniaceae), tip of stem showing compound woolly leaves, inflo-
rescences; summit of Mt. Mou. sorrom ricut, P. communis, branch with inflorescence; leaves
about 3 cm long; Plaine des Lacs.

iD s

FIG. 7.7. Plants from New Caledonia. trop LEFT, rIGHT, Scaevola beckii (Goodeniaceae);
base of Mt. Doré. top LEFT, habit; plant is 1 m tall. rop ricut, leaves and portion of inflo-
rescence. BOTTOM LEFT, S. montana, branch showing flowers and fruits; leaves are ca. 5 cm
long; lower Mt. Koghi. sorrom ricut, Hibbertia badouinii, rosette of leaves from a large
rosette shrub about 6 m tall; leaves are ca. 20 cm long, flowers 7 cm in diam.; summit of
Mt. Mou.

229

FIG. 7.8. Myodocarpus (Apiaceae) from New Caledonia. TOP LEFT, RIGHT, BOTTOM LEFT, M.
fraxinifolius. Tor LEFT, habit of tree, 6 m tall; pass between Nouméa and Plaine des Lacs.
TOP RIGHT, branch, showing pinnate habit of leaves. BOTTOM LEFT, inflorescence, showing the
flattened dry fruit. Borrom ricHT, M. involucratus; leaves are entire, ca. 20 cm long; summit

of Mt. Mou.

230

Adaptive Radiation: New Caledonia and New Zealand

droseris (Asteraceae) on the Juan Fernandez Islands. The New Caledonian
species of Scaevola (fig. 7.7, top left, nght, bottom left) mostly occur in
upland cloud forest and, like other Pacific species, have relatively thin,
broad leaves (e.g., S. montana). However, the lowland S. beckii (fig. 7.7,
top left, nght) shows modifications typical of drier lowland sites—branches
and leaves are few, leaves are thick and purple brown.

Pancheria of the Cunoniaceae (fig. 7.6, bottom left, right) is an en-
demic genus with perhaps as many as 30 species (Guillaumin, 1948). These
species are highly diversified in leaf type, and only two of many are shown
here. Of these, P. insignis (fig. 7.6, bottom left) has densely hairy, com-
pound leaves. This species occurs in the Mt. Mou cloud forest. A contrast
is formed by P. communis (fig. 7.6, bottom right) from the Plaine des Lacs.
It has small, simple leaves that are thickish, sclerophyllous, and perhaps
isolateral, in contrast to the clearly bifacial condition found elsewhere in
the genus.

Some foliar distinctions in New Caledonian angiosperms do not at
first glance seem to bear close relationships to habitat features. In Myodo-
carpus of the Araliaceae (fig. 7.8), for example, M. involucratus (fig. 7.8,
bottom right), a species of cloud forest and intermediate scrub, has simple
leaves. In M. fraxinifolius (fig. 7.8, top left, right, bottom left), a taller
rosette tree of scrubby areas adjacent to the Plaine des Lacs, pinnately
compound leaves occur. However, the leaflets prove to be thicker than
those of M. involucratus, suggesting semi-succulence. Compound leaves in
M. fraxinifolius may represent a relationship, perhaps not strong now, to
growth of a rosette tree in areas shaded by trees and shrubs. Where foliar
dimorphism occurs, more dissected leaves are characteristic of lower levels
in a forest. This may not obtain in Myodocarpus, or, if dissected leaves
are a juvenile form, they seem to have advantages in the adult plant as
well, in the case of M. fraxinifolius.

Myodocarpus exemplifies another problem typical of evolution in
certain New Caledonian plants. In its habit, Myodocarpus could not be
more typical of Araliaceae (sensu stricto). Its dry fruit (fig. 7.8, bottom left),
however, would place it in a nearby family, Apiaceae (Umbelliferae). Is
Myodocarpus a vestige of the stock from which these two families (united
by some authors) were diverging? Or is Myodocarpus an apiad that has be-
come woodier on an insular area, as has Bupleurum in the Mediterranean

ZN

Adaptive Radiation: New Caledonia and New Zealand

region? If the latter hypothesis is true, it can be true only to the extent that
Myodocarpus probably never had any truly herbaceous ancestors—as did,
by contrast, such genera as Melanoselinum on Madeira (fig. 10.2, top
right). New Caledonia, without doubt, has an unusual assemblage of plants
rich in primitive characteristics—particularly the conifers, Winteraceae,
and Amborellaceae. But we may well be in error if we are thereby influ-
enced to regard a given New Caledonian genus as more “primitive” in its
family than genera elsewhere. Even assuming this is true for a particular
genus, this does not prove that the New Caledonian flora (with the excep-
tion of conifers other than Podocarpaceae) could have arrived only across
a dry-land zone or small water gap from a source area. If New Caledonian
angiosperm genera have more numerous primitive characteristics than rela-
tives elsewhere, this probably is due to the fact that they arrived early by
means of long-distance dispersal.

ANIMALS

With the exception of the flightless kagu (Rhynochaetos jubatus),
the New Caledonian avifauna is just about what we would expect for a
not-too-old volcanic island in this geographical region, and the roster of
birds (Delacour, 1966) suggests recentness and no examples of adaptive
radiation, probably because New Caledonia is not at the limit of dispersi-
bility for land birds. The gekkonid lizards of New Caledonia have been
regarded as representing adaptive radiation (MacArthur and Wilson,
1967). Two endemic genera of Gekkonidae (Rhacodactylus and Eurydac-
tylus) and one nonendemic genus (Bavayia) contribute to this pattern.
Among skinks, New Caledonia is notable for a giant form, Riopa bocourti
(Mertens, 1934).

Assuming that among insects the phasmatid Clitarchus (Nakata, 1961)
and the longicorn Enicodini (Gressitt, 1961) may be relatively primitive,
other insect groups in New Caledonia have few relictual aspects. New
Caledonian ants are cited as an example of adaptive radiation by Mac-
Arthur and Wilson (1967).

Changes in size and form—some perhaps recent—apply to land snails
as well (Solem, 1961). We may be astonished at the large size of Placo-
stylus, and deem it incapable of long-distance dispersal—and in its present
form perhaps it does lack dispersibility. However, Placostylus may be a

232

Adaptive Radiation: New Caledonia and New Zealand

gigantism, perhaps old in origin, of a once smaller and more vagile land
shell. Gigantism is common among the land shells of the Hawaiian Islands,
and certainly among New Caledonian land shells as well (see chapter 15).
Gigantism in these groups can be regarded as part of a program of adap-
tive radiation into new habitats.

New Zealand

In view of its exceptional ecological richness and long isolation, New
Zealand would be expected to exhibit remarkable speciation and adaptive
radiation. New Zealand has probably been: separated by wide water gaps
since the time when it acquired conifers. Few or no water barriers at
an early date might be hypothesized for the arrival of some conifers.
Araucariaceae seem poorly dispersible, yet Agathis reached New Zealand,
as did Araucaria (known there as a fossil: Florin, 1963). The southern
beech, Nothofagus, may have reached New Zealand at a similar time. It
is present in Patagonia, New Caledonia, Tasmania, and Australia, a dis-
tribution somewhat like that of the Araucariaceae. An Antarctic arc of
greater land areas may once have existed and permitted these ranges. How-
ever, marsupial and placental mammals (other than bats) never reached
New Zealand, which, like New Caledonia, has a strongly disharmonic
fauna: birds, insects, land shells, a few reptiles (including the primitive
Sphenodon), a few amphibians (including the primitive Leiopelma), and
freshwater fishes (although no primary-division freshwater fishes). The suc-
cess of introduced marsupials and placentals certainly demonstrates that
potential niches for these have existed.

Favorable to adaptive radiation on New Zealand are the great range
in altitude (to 12,349 ft) and in rainfall: as low as 20 in. on the eastern
coast of South and North Islands to above 200 in. in the western coastal
and montane areas of South Island. Volcanics augment a diversified series
of metamorphic and sedimentary rocks (Hochstetter, 1959). Vegetational
formations, consequently, present a broad spectrum—too broad to summa-
rize here; Cockayne (1958) has offered a highly detailed study. Cockayne
(1927) has also presented an appealing semipopular account of the vege-
tation.

233

Adaptive Radiation: New Caledonia and New Zealand

Several features detract, however, from the potential ecological diver-
sity of New Zealand. First, New Zealand is decidedly temperate, and even
the northernmost part of North Island cannot be considered really sub-
tropical. The wet coastal forests, however, are comparable to higher eleva-
tions in the Hawaiian rain forests. Second, the Pleistocene glaciation of
New Zealand (Hochstetter, 1959) in essence reduced the available land
area severely and thus had the net effect of reducing the biota to a much
more cold-hardy flora than New Zealand would support under current
conditions. The time since glaciation has undoubtedly featured speciation
and radiation of at least some of the surviving groups, but postglaciation
time has not, in all likelihood, been sufficient for over-water immigration
of many of the less cold-hardy groups, much less been sufficient for their
speciation.

CONIFERS

Compared with New Caledonia, New Zealand is notably poor in
radiation of Araucariaceae. Arducaria became extinct in prehuman time,
and only Agathis australis represents the family there today. Perhaps New
Zealand is more nearly marginal for Araucariaceae, but certainly it must
have been more so in the Pleistocene, and the rather slowly evolving
Agathis has not speciated since that time.

Dacrydium and Podocarpus are another matter: these two genera
(which can even be regarded as a single genus) show not merely an array
of species in New Zealand, but a spectrum of growth form and habitat
preference. Probably this radiation occurred prior to glaciation, with few
if any exceptions. Most distinctive of New Zealand Dacrydiums is D. laxi-
folium, the “pygmy pine,” a sprawling or almost vinelike shrub with
branches to only 2.5 mm in diameter. This subshrubby alpine species is per-
haps the least woody conifer in the world, other than the parasitic Podo-
carpus ustus of New Caledonia. Dacrydium biforme and D. bidwillii (the
“bog pine’) also grow in alpine New Zealand but are shrubs, about 1—2 m
tall. These stand in contrast to such arboreal species of lowland or montane
forest as D. cupressinum (the common “rimu” or “red pine”), D. inter-
medium, and D. colensoi.

In Podocarpus, P. dacrydioides, to 50 m or more, can be a huge tree of
“lowland swamp forest.” One can contrast it with the prostrate alpine

234

Adaptive Radiation: New Caledonia and New Zealand

shrub, P. nivalis. However, such extremes may represent more than one
introduction of Podocarpus to New Zealand rather than autochthonous
radiation on the islands. Podocarpus nivalis appears very close to P. alpina
of Australia. Podocarpus totara, P. spicatus, P. ferrugineus, and P. hallii
are medium-sized trees most frequently found in lowlands; some indi-
viduals of these species are rounded in shape, and look more oaklike than
coniferlike at a distance. Small size of leaves in these species suggests adap-
tation to a forest that is only moderately moist. Podocarpus spicatus may
be related to the South American P. andinus, according to Florin (1951),
who apparently regards cone types in these species as primitive.

ANGIOSPERMS

In discussing angiosperms of New Zealand, one can present growth
forms and habitats and enumerate examples, or one can see the range of
niches occupied by particular genera. The former point of view is very
tempting, because growth forms and habitats are so distinctive for New
Zealand plants.

A number of these forms are illustrated by species of Hebe (Scrophu-
lariaceae), as shown in figure 7.9 (regrettably, only a small range can be
illustrated). The majority of the approximately 100 species of this genus
are native to New Zealand, the remainder to other austral areas (Allan,
1961). Hebe can be a true tree, as in H. parviflora var. arborea, to 7.5 m
tall. This variety occurs in mesic localities in the relatively warm North
Island and its offshore islands, and moderation of the maritime climate
is probably related to arborescence. Hebe salicifolia (fig. 7.9, bottom right)
is a shrub to 5 m, from South Island. Lowland and mid-elevation localities
in New Zealand host a number of somewhat smaller broad-leaved shrubs
(1-2 m). At increasingly higher elevations, one tends to find progressively
smaller shrubs with smaller leaves. Whereas mid-elevation shrubby species
tend to be branched from above the base, the alpine and subalpine species
are branched from the base or at ground level. Because New Zealand moun-
tains are rich in scree areas, most of the high-elevation species branch from
below the rocks and rubble which perpetually fall on them, and thus re-
newal buds are potentially removed from killing cold. This is true of the
three species illustrated as figure 7.9, top left and right, and bottom left.
Prostrate stems characterize some alpine species, such as H. ramosissima,

235

FIG. 7.9. Species of Hebe (Scrophulariaceae) from New Zealand. top Lert, H. hectori var.
demissa, leaves about 3 mm long; Arthur’s Pass. Trop ricuHT, H. tetrasticha, leaves about 5 mm
long; Porter’s Pass. Borrom Lert, H. epacridea, leaves 10-12 mm long; Arthur’s Pass. BorTOM
RIGHT, H. salicifolia, leaves 90-100 mm long; cultivated in Garberville, Calif.

236

Adaptive Radiation: New Caledonia and New Zealand

H. petriei, and H. haastii (Allan, 1961). Hebe macrantha is exceptional in
its very large flowers, borne on prostrate stems; the large flowers remind
me of the tendency toward large flowers in other alpine genera.

Decrease in leaf size can be shown well by alpine Hebes; this change
would be expected in these physiologically dry habitats. Water is available
as snow melts, but not as the (often porous) soils dry out in warm weather
or freeze during winter; humidity is low in alpine localities, and greater
wind action aggravates the effect of desiccation, so microphylly is an almost
inevitable adaptation. Hebe epacridea (fig. 7.9, bottom left) shows an early
step in xeromorphy, H. tetrasticha (fig. 7.9, top nght) and H. hectori var.
demissa (fig. 7.9, top left) show progressively further expressions. Species
like the latter, known as “whiplash” Hebes, are notable for their coniferlike
foliage. Because Hebe often has decussate phyllotaxy, these species re-
semble various genera of Cupressaceae. ‘The factors that favor conifers such
as Juniperus in comparable alpine localities on mainland areas are plainly
operative in evolution of Hebe as well. One may ask why, with a relatively
great number of conifers in New Zealand, coniferlike Hebe species should
evolve in New Zealand. One reason is that particularly appropriate conifer-
equivalents, such as Cupressaceae, are absent in New Zealand. Another
explanation is the ability of Hebe to occupy a greater variety of niches.
However, alpine conifers have been evolved (or relictually preserved) in
alpine New Zealand (Dacrydium laxifolium, D. biforme, D. bidwillii, and
Podocarpus nivalis). This suggests that a selective pressure for these growth
forms definitely does exist. The endemic scrophulariaceous genus of cush-
ion plants Pygmaea probably represents a Hebe that has evolved fully into
the cushion-plant habit.

In addition to distinction in habit, preferences in edaphic situations
occur in Hebe. If one reviews the data in Allan (1961), one finds that rocky
places, even disturbed places, form a predominant habitat for Hebe spe-
cies. This habitat suggests that Hebe is a pioneering group. Some species
are described as favoring “forest margins,” “grasslands,” or “scrub.” Habi-
tats like these also suggest characteristics of a pioneering group. However,
among species for which rocky habitats are specified, there is a diversity
of preferences, suggesting adaptive radiation: ‘‘sea-cliffs’” (H. speciosa, H.
obtusata); “under falls” (H. acutiflora); “maritime cliffs where there is a
maximum of spray” (H. chathamica); “on serpentine’ (H. urvilleana);

237),

Adaptive Radiation: New Caledonia and New Zealand

“sock platforms on river banks” (H. recurva); “on rocks in tussock grass-
land” (H. amplexicaulis); “limestone rocks” (H. raoulii var. maccaskillit);
and some seem to favor volcanic rocks. ‘There are many alpine species, in-
evitably in rocky places, but even these are varied in preferences; some
grow on shingle on steep slopes, some in fell fields, and one entire section
(Buxifoliatae) of the genus occurs in swampy alpine places. At least one
species (H. diosmifolia) grows in “dense woods” and thus appears to have
entered the most stable habitats. The genus Parahebe can be regarded as
a more herbaceous (subshrub) phase of Hebe, and thus represents an adap-
tation to situations suited for herbs. The great degree of hybridism in
Hebe (see Allan, 1961) is also an indication of the rapid evolution that
has occurred (perhaps following glaciation) without concomitant forma-
tion of genetic barriers. Geographical isolation and shift into various
edaphic preferences have evidently operated to promote speciation (and
the predominantly rocky habitats favored by Hebes are, by their particular
geographic nature, isolated—stream beds, for example).

Coprosma (fig. 7.10) has 45 species in New Zealand, and 45 on other
Pacific islands (Allan, 1961), so that New Zealand is obviously a primary
area of speciation of the genus. In its growth forms, it constitutes a prime
example of adaptive radiation. These include broad-leaved arboreal spe-
cies. Interestingly, as in Hebe, these tend to occupy mesic sites in northern,
warmer New Zealand (data from Allan, 1961): C. tenuifolia (to 5 m; North
I; alluvial soils); C. acutifolia (to 10 m; Kermadec Is.); C. australis (to 6
m; Three Kings Is., North I. to 42° S; marginal forest); C. robusta (to 6
m; Three Kings Is., North I., South I., Chatham Is.; forest and shrubland
on alluvial soils); C. macrocarpa (to 5 or even 10 m; Three Kings Is., North
I., and adjacent islets); C. chathamica (15 m; Chatham Is.; forest); C. pett-
olata (to 6 m; Kermadec Is.); C. repens (to 8 m; Kermadec Is., Three Kings
Is., North I., South I. to 41°30’S); C. lucida (to 6 m; North I., South L.,
Stewart I.).

Such trees probably represent secondary entrants into the truly arbo-
real category and are really large shrubs that can, under favorable condi-
tions, form trees One species not listed above, but which corresponds to
this description, is C. foetidissima (fig. 7.10, bottom right), a rain-forest
species that can attain 6 m. The shrubby species that do not attain tree-
hood tend to be much smaller leaved, and they occupy drier or colder locali-

238

Fic. 7.10. Species of Coprosma (Rubiaceae) from New Zealand. top LEFT, C. pseudocuneata,
leaves 18-20 mm long; Arthur’s Pass. rop ricut, C. brunnea, leaves 20-25 mm long; river-

bed, Franz Josef Glacier. sorrom Lert, C. crassifolia, leaves 12-15 mm in diam.; between
Westport and Greymouth. potrom ricuT, C. foetidissima, branch with male and female
flowers; leaves 90-100 mm long; between Arthur’s Pass and Hokitika.

239

Adaptive Radiation: New Caledonia and New Zealand

ties. Examples of the many species in this category are C. pseudocuneata
(fig. 7.10, top left), C. brunnea (fig. 7.10, top right), and C. crassifolia (fig.
7.10, bottom left). Of these, C. brunnea is a prostrate mat-plant of montane
river beds and cold lowlands. Coprosma pseudocuneata is an alpine shrub
of herb fields; its branches appear stunted and stiff, condensed, with almost
succulent leaves. Coprosma crassifolia is not an alpine plant but grows in
shrubland. It has orbicular leaves, spaced well apart on stout branches. ‘This
species could qualify as a “divaricating shrub” because of the branching
pattern (see below). The opposite to arborescence is illustrated by C. pe-
triei (fig. 7.11, top left). ‘This species is very small-leaved, and forms exclu-
sively prostrate stems. If the stems were crowded, short, and upright, we
could call it a true cushion plant; it does grow with cushion plants such
as Scleranthus and Raoulia. Flowers are borne virtually at ground level, so
that on female plants the large red or purple berries appear to be lying
on the ground. Like cushion plants, C. petriei grows in montane regions in
grassland, on margins of streams, and in rocky habitats.

Cushion Plants. Cushion plants are highly characteristic of the New
Zealand flora because this growth form is so well adapted to montane
regions (see also chapter 14). Cushion plants are microphyllous and thus
transpire minimally in dry summer months or during winds. The compac-
tion of stems, side by side, further reduces transpiration. Old stems and
leaves can serve as wicks, and roots from younger portions grow down
through the cushion. The cushion, as a sort of ground cover, aids in reduc-
ing water evaporation from the soil beneath it. The cushion habit is also a
form of resistance to cold. ‘The many renewal buds are below the surface
of the cushion and are not damaged by freezes. However, the numerous
stems can produce inflorescences rapidly during the summer, and produce
large numbers of flowers and fruits at cushion surfaces. Because flowers and
fruits are borne at surfaces of cushions, they emerge rapidly during the
growing season, and protection against occasional summer frost is also
afforded. Raoulia (Asteraceae) is an ideal example of autochthonous evolu-
tion of cushion plants in New Zealand. The genus is almost exclusively
native to New Zealand; a few outlier species have reached alpine New
Guinea. Some species are merely prostrate and matlike, others form leafy
upright stems, short but separate, and with moderately small leaves. The
true cushion Raoulias such as R. australis, shown in figure 7.11, bottom left

240

Win

FIG. 7.1]. Prostrate and cushion plants from New Zealand. top Lert, Coprosma petrei,
male plant in flower; Porter's Pass. rop ricut, Dracophyllum prostratum in flower, growing
in a cushion of Scleranthus (Caryophyllaceae); Arthur’s Pass. BOTTOM LEFT, RIGHT, Raoulia
australis. BOTTOM LEFT, habit of cushion; near Wanaka. BOTTOM RIGHT, view of surface of
cushion in flower; leaves 4-5 mm long.

241

Adaptive Radiation: New Caledonia and New Zealand

and right, and the famous R. eximia have developed in New Zealand, prob-
ably from ancestors with prostrate stems. Within the single genus Raoulia,
a variety of growth forms and habits is represented. However, Raoulia is,
in turn, only one phase of radiation in helichrysoid composites, a radiation
which includes shrubs as well. Raoulia leaves are not only very small, but
also woolly, which aids in frost resistance or desiccation resistance, or both.
For illustrations of Raoulia and allied genera that can form cushions, the
reader can consult Philipson and Hearn (1962).

Cushion plants, particularly those that are more matlike and prostrate,
can form “seed beds” in which other species can germinate. Thus, cushions
of Scleranthus may host seedlings of Dracophyllum, which can in turn be-
come cushions. Thus, in alpine regions suited for them, cushion plants are
“contaminated” with other cushion species. A stage in this process is illus-
trated in figure 7.11, top right. Species like Dracophyllum muscoides, D.
prostratum, D. pronum, Pernettya nana, Gaultheria depressa, Coprosma
pumila, C. petriei, Pimelea prostrata, and Myrsine nummularia are unusual
in their reduced stature, probably evolved autochthonously in New Zea-
land. Some cushion plants extend well beyond New Zealand (e.g., Oreo-
bolus, Scleranthus, Gaimardia, Donatia, Phyllachne) and although they
may not represent evolution of this growth form on New Zealand itself,
they do demonstrate existence of selective factors that have led to autoch-
thonous evolution of this form in Raoulia, Pygmaea, and others. Cotula
is unusual in New Zealand in that it is an herb group that forms mats,
whereas it is often only an annual elsewhere. These examples point up the
tendency of adaptive radiation in evolutionarily plastic groups such as
Asteraceae, Scrophulariaceae, and Rubiaceae.

Divaricating Shrubs. Although not unique to New Zealand, the so-
called divaricating shrubs, mentioned above, are remarkably abundant in
the New Zealand flora. Went (1971) claims 51 species in 23 families as
representing this habit in the New Zealand flora. Some (Dawson, 1963;
Went, 1971) have implied that divaricating shrubs are unique to New
Zealand, but they are not. A number are evident to me in the California
flora: Cercocarpus intricatus and C. ledifolius (Rosaceae), to name only
two at random. A divaricating shrub may be defined as a woody shrub in
which branches tend to be at right angles and in which apical dominance

242

Adaptive Radiation: New Caledonia and New Zealand

is lessened, although not totally absent. Divaricating shrubs tend to have
stout stems and small leaves.

Many genera also represented by “ordinary” broad-leaved shrubs or
other growth forms have evolved divaricating shrubs that are characteristic
of New Zealand’s shrublands. Coprosma crassifolia (fig. 7.10, bottom left)
was cited earlier. In the small genus Corokia (Saxifragaceae? Cornaceae?)
with three New Zealand species, C. buddleoides (fig. 7.12, bottom right)
and C. macrocarpa contrast with the divaricating shrub C. cotoneaster (fig.
7.12, bottom left). An almost identical situation is represented by Meli-
cytus (Violaceae), with three broad-leaved New Zealand species, such as
M. ramiflorus (fig. 7.13, bottom right), and one other New Zealand species,
M. micranthus (fig. 7.13, bottom left), that is a divaricating shrub, or a
close approach to that habit.

Many other genera have entered the divaricating-shrub habit. Shown
here are Pittosporum crassicaule (fig. 7.14, top left) and Aristotelia fruti-
cosa (fig. 7.14, top nght). Other divaricating species of Pittosporum in-
clude P. rigidum, P. anomalum, P. divaricatum, and P. lineare. Further
good examples include Muhlenbeckia astoni (Polygonaceae), Myrtus
pedunculata (Myrtaceae), Plagianthus divaricatus (Malvaceae), Paratrophis
microphylla (Moraceae), Pennantia corymbosa (Icacinaceae), Neopanax
anomalum (Araliaceae), Myrsine divaricata (Myrsinaceae), many species
of Coprosma (Rubiaceae), several of Olearia (Asteraceae: O. odorata, O.
virgata), and Helichrysum glomeratum (Asteraceae). Divaricating shrubs
represent tendencies, not a unitary growth form; in any case, ordinary evo-
lutionary principles can explain their development, and Went’s (1971)
“floating gene” hypothesis should be rejected.

Microphylly. Numerous New Zealand plants, including divaricating
shrubs, cushion plants, and mat plants, could be cited as examples of
microphylly. However, some small-leaved plants, such as Hymenanthera
alpina (fig. 7.13, top right) of the Violaceae and a thorn-shrub, Discaria
toumatou (fig. 7.14, bottom left), of the Rhamnaceae could come under
this heading. Virtually all of the New Zealand Epacridaceae could be
called microphyllous. The microphyllous habit suits alpine and montane
regions for reasons cited earlier for the whiplash Hebe species.

Broom Shrubs and Cladode Shrubs. Xeromorphy in dry montane
regions is found in a number of New Zealand genera that have other

TAB

7.12. Plants from New Zealand. ror Lert, r1GHT, Metrosideros perforata (Myrtaceae).
g on a tree; near Paeroa. TOP RIGHT, note wiry stems, leaves
about 15 mm in diam. Borrom Lert, Corokia cotoneaster (Cornaceae), stems have rather
small leaves 10-12 mm; from plant in cultivation, Claremont, Calif. Borrom ricHT, C. bud-
70-80 mm long; cultivated on University of Auckland

FIG.
TOP LEFT, habit, plant clim

dleioides, tip of shoot, showing leaves
campus.

244

OIG
Uy,

y
Vhs w

Zealand. top LEFT, Muhlenbeckia ephedrioides (Polygonaceae),

; cultivated at the Dominion Scientific Industrial and Research Organization,

n, Lincoln, N.Z. top ricur, Hymenanthera alpina (Violaceae), leaves about

near Queenstown. BOTTOM LEFT, Melicytus micranthus (Violaceae), leaves

long; cultivated, University of Auckland campus. Borrom ricHT, M. ramiflorus,
s ca. 7) mm long; between Greymouth and Westport.

245

Adaptive Radiation: New Caledonia and New Zealand

growth forms as well; for example, Muhlenbeckia ephedroides (fig. 7.13,
top left). Other examples include species of Carmichaelia (Fabaceae) such
as C. uniflora, C. kirkii, and C. curta. Other New Zealand legume broom-
shrub genera include Notospartium, Chordospartium, and Huttonella.
Most species of Carmichaelia and the monotypic Corallospartium (Faba-
ceae) are close to the broom habit, but they have flattened leafless branches
and should be called cladode shrubs.

Pronounced and Persistent Juvenilism. Surprisingly, the phenomenon
of juvenile leaves quite different from adult leaves is present in about 200
species of New Zealand seed plants, of which 165 remain juvenile for a
year or more (Cockayne, 1958). According to him, the 165 species belong
to 50 genera in 30 families, so here we are dealing with a phenomenon for
which selective factors must be operative. Cockayne claims that in 106
cases the juvenile appears to be more mesophytic than the mature form,
and in 97 cases more xeromorphic, while in 39 cases there is little differ-
ence: thus degree of dryness is apparently not involved here as a selective
factor in juvenilism. One possible explanation is that the New Zealand
forest is unusually dense—at least denser than the characteristic habitat
for some of the genera that compose it; for example, Hoheria (Malvaceae)
or the numerous Araliaceae, such as Psewdopanax. In other instances,
“adaptive modes” of adult plants (such as the divaricating shrubs) are so
distinctive that juvenile leaves resembling less specialized members of par-
ticular genera are almost inevitable. Cockayne (1958) notes that in the
whiplash Hebe species and the cupressoid Podocarpaceae, shoots with ju-
venile (mesomorphic) leaves may be found close to the ground on plants
growing in exceptionally moist conditions. Cockayne provides an extensive
listing of species known characteristically to flower on juvenile shoots—a
situation confusing to the inexperienced botanist. The parallels to the
juvenile stages of the Hawaiian lobelioids are striking (see chapter 4), and
the reasons are probably similar in many of the New Zealand instances.
Juvenile leaves in Hawaiian lobelioids and in most New Zealand plants with
juvenilism are broader (at least in proportion to length) and more lobed
than are adult leaves.

Climbing plants in particular seem to exhibit pronounced juvenilism
in New Zealand. One of these, Metrosideros perforata, is illustrated in fig-
ure 7.12, top left and right. The juvenile shoots form in profusion as this

246

Fic. 7.14. Divaricating shrubs (top LEFT, RIGHT) and plants with armature (BOTTOM LEFT,
RIGHT) from New Zealand. top LEFT, Pittosporum crassicaule (Pittosporaceae), leaves about
10 mm long; Arthur’s Pass. rop ricur, Aristotelia fruticosa (Elaeocarpaceae), stems in flower;
leaves 12-14 mm long; Arthurs Pass. Bsorrom tert, Discaria toumatou (Rhamnaceae),
(continued)

ei]

Adaptive Radiation: New Caledonia and New Zealand

epiphytic liana climbs a tree. The adult leaves are larger (even if they still
are smaller than those of most other species of Metrosideros), and inter-
node length is reduced as adult shoots form. Adult shoots form only when
M. perforata attains the forest canopy, and one would not easily guess the
identity of the plant from juvenile shoots alone.

Rubus shows juvenilism in narrowing of laminae on leaflets in some
species. In R. squarrosus laminae are almost absent; the bare prickly midnb
of leaflets serves as an efficient climbing-attachment organ that is sufficient,
however, for photosynthesis. Presumably Cockayne would have listed this
condition as being more xeromorphic than presence of a lamina, but xero-
morphy is not really applicable here. Shade leaves in seed plants at large
tend to have fewer photosynthetic cells than do adult leaves, so absence of
a lamina in R. squarrosus may be interpretable in this way.

Thorniness. Absence of armature in the Hawaiian flora is striking.
One might expect a similar absence in New Zealand, equally as free from
native mammalian herbivores as the Hawaiian Islands. However, a number
of New Zealand plants (like the Rubus species just mentioned) have spines,
thorns, or prickles. Two are shown here: Aciphylla sp. (fig. 7.14, bottom
right) and Discaria toumatou (fig. 7.14, bottom left). The entire genus
Aciphylla (Apiaceae) could be cited. Other examples include Hymenan-
thera angustifolia (Violaceae), Carmichaelia angustata (Fabaceae), Urtica
ferox (Urticaceae), juvenile leaves of Pseuwdopanax (Araliaceae), and Ole-
dria ilicifolia (Asteraceae). However, the tough, tangled, microphyllous
branches of many of the divaricating shrubs seem to have virtually the
same effect as thorny branches. If these various “armor” devices have a
selective value in deterring herbivores, as one would expect, we would have
to cite herbivores likely to have provided such pressure. Avian herbivores,
such as the giant extinct moas and other extant herbivorous birds, might
be one factor. Apiaceae is a notably unarmed family; the strong develop-
ment of armature in the many species of Aciphylla seems difficult to ex-
plain on other grounds, although a component of xeromorphy might be
invoked in this and other cases.

FIG. 7.14 (continued)

branches, showing thorniness, in flower; below Porter’s Pass. BoTTOM RIGHT, Aciphylla sp.
(Apiaceae), showing needlelike leaves, about 30 cm long, and inflorescence bracts; cultivated

at the Dominion Scientific Industrial and Research Organization, Botany Division, Lincoln,
N.Z.

248

Adaptive Radiation: New Caledonia and New Zealand

Scree Plants. In habit and color, New Zealand scree plants are unusu-
ally interesting. There is nothing unique about them, for scree slopes else-
where in the world bear autochthones equally striking in adaptation. The
features of interest are the large roster of species in New Zealand (see
Philipson and Hearn, 1962) and the groups to which they belong—like
Cotula and Crasspedia (Asteraceae) and Ranunculus—genera one might
not expect to have scree adaptations, but which have radiated into this
habitat under the disharmonic conditions of the New Zealand flora. Color
mimicry is striking in New Zealand scree plants, and they truly are often
difficult to discern in their natural habitat. Certain instances of scree plants
were cited earlier in Hebe.

Arborescence. For discussions of arborescence in the New Zealand
flora, see chapter 10.

INSECTS

The best examples of adaptive radiation in the New Zealand fauna
are perhaps the wetas, considered as belonging to two families of Orthop-
tera, Henicidae (ground wetas and straight-bodied wetas) and Rhaphido-
phoridae (the arched-bodied wetas—often called cave wetas but found on
trees as well—with long legs and antennae). The henicids can be very large.
Deinacrida and Hemideina consume rotting wood, and their large mouth
parts and movements (running, never springing) correspond to this habit.
Hemideina burrows in rotten wood and is often found among foliage.
Deinacrida occurs among leaves and stones in montane regions. Hemian-
drus lives in small burrows in soft logs and riverbanks. Zealandrosardus
occurs under stones on montane slopes and under logs.

The Rhaphidophoridae are smaller and much more delicate wetas,
capable of rapid jumping movements. Pharmacus montana, for example,
springs from rock cavities in mountain areas of the Southern Alps. Other
thaphidophorids live in tunnels, caves, holes in logs, and under rocks. For
illustrations and further references, see Natusch (1967).

FRESHWATER FISHES
The Galaxiidae is the largest family of freshwater fishes in the New
Zealand icthyofauna, and may represent an older element (McDowall,

249

Adaptive Radiation: New Caledonia and New Zealand

1970). Although McDowall speaks of the “limited extent of the radiation
of the New Zealand Galaxiudae” as evidence that they are a conservative
group, their adaptive radiation in New Zealand is quite evident. ‘To be
sure, none are herbivores, none are piscivorous, and none live in streams
that lack forest cover (except for alpine streams). Within their diet of
stream invertebrates, however, many adaptations are evident. Those species
that feed in rapid water tend to be more slender in form. Some (Galaxias
brevipinnis, G. postvectis) have receding lower jaws, an adaptation to feed-
ing from the stream bottom. Galaxias prognathus, however, has a notably
protruding lower jaw, related to feeding from the undersides of rocks. The
solitary and secretive nature of most species exposes them to damage from
close contact with rocks and logs, and this is evidently counteracted by the
greatly thickened fins and truncated caudal fins of those species. In those
species, pectoral fins also are in a low latero-ventral position, correlated
with resting on the bottom. In G. brevipinnis, corrugations are present on
lower surfaces of pectoral fins, permitting grasping of the stream bottom
despite rapid flow of water. The few species that live in open pools and
shoal have much more membranous fins, placed in a higher lateral posi-
tion. Shoaling species such as G. maculatus (like shoaling fishes in general)
have silvery coloration on the abdomen. Some Galaxias species in New
Zealand have superior climbing ability (G. brevipinnis, G. fasciatus, G.
postvectis), and show this in that pelvic fins are nearly flush with the ven-
tral trunk and have the laminae facing downwards. ‘T'wo lacustrine species
(G. usitatus and G. gracilis) are notable for their long gill rakers. Gill-raker
reduction occurs in the alpine species G. divergens and G. prognathus. ‘The
most striking adaptation evident in New Zealand galaxiads, and very pos-
sibly an autochthonous development following immigration of the original
galaxiad stock from Australia, according to McDowall, is the ability to live
in situations subject to drying: the genus Neochanna. Neochanna (three
species) lives in swamps, swampy creeks, and springs, and can aestivate
during droughts. The semi-burrowing habit appears in such morphological
features as lost or reduced pelvic fins, reduced eyes, long nostrils, low
median fins approximately confluent with the caudal fin, and a rounded
caudal fin. Even with the limitations that McDowall (1970) suggests, his
fine monograph clearly indicates interesting parameters of adaptive radia-
tion in galaxiads of New Zealand.

250

Adaptive Radiation: New Caledonia and New Zealand

COMMENTS ON NEW ZEALAND PATTERNS

The avifauna of New Zealand, although quite interesting on other
grounds, shows relatively little adaptive radiation. Such genera as Acanthi-
sitta, Xenicus, Cyanorhamphus, and Anthornis have formed what should
be called vicarious species, judging from the data of Oliver (1955). These
vicarious species must be recent, for differences within such genera among
species of North Island, South Island, and the Chatham Islands are small.
There may have been adaptive radiation among the moas, judging from
differences in size among the fossils, but we will perhaps never have sufh-
cient information for a definitive statement. Vertebrates other than birds
are either relictual (Leiopelma, Sphenodon), or little-speciated immigrants
of more recent epochs, such as the geckos.

Even if New Zealand had escaped glaciation, it still might not bear
as broad spectrum of adaptive radiation as do the Hawaiian Islands. New
Caledonia agrees with New Zealand in this respect. In both cases, we are
dealing with older islands. However, neither of these old islands has the
Hawauan advantage of being a true archipelago, nor is climate as favorable
as that in Hawai. There is no doubt that events of dispersal throughout
an archipelago, followed by speciation and adaptive shift on various islands,
then redispersal to island of origin can lead to diversification.

How, then, have the cases of adaptive radiation cited for New Zealand
and New Caledonia occurred? First, there is the possibility that for antique
groups, such as conifers, an archipelago—or larger island with internal bar-
riers—may have existed where only one island (e.g., New Caledonia) now
is present. Adaptive radiation in this case would have occurred long ago,
and its presumptive products have not yet been extinguished. Second,
there is the possibility that, in recent time, evolutionarily plastic groups
have speciated adaptively. If they have, it could be explained if suitable
habitats for a given genus occur in discontinuous fashion. For example,
pockets of scrubland, alpine regions, and areas of particular soil types
could, in effect, provide archipelagos, or even a series of archipelagos—and
adaptive radiation could occur (possibly following glaciation). This may
explain radiation in New Zealand species of Hebe, the wetas, or the ga-
laxiad fishes.

HE)

Adaptive Radiation: New Caledonia and New Zealand

References

Allan, H. H. 1961. Flora of New Zealand, vol. I. Wellington, R. E. Owens, Gov-
ernment Printer.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Cockayne, L. 1927. New Zealand plants and their story. Wellington, W. A. G.
Skinner, Government Printer.

Cockayne, L. 1958. The vegetation of New Zealand. 3d ed. Weinheim, Germany,
H. R. Engelmann (J. Cramer).

Dawson, J. W. 1963. A comment on divaricating shrubs. Tuatara 11:193—-94.

Delacour, J. 1966. Guide des oiseaux de la Nouvelle-Calédonie et de ses dépendances.
Neuchatel, Switzerland, Editions Delachaux & Niestleé.

Florin, R. 1951. Evolution in cordaites and conifers. Acta Horti Bergiani 15(11):
285-388.

Florin, R. 1963. The distribution of conifer and taxad genera in time and space.
Acta Horti Bergiani 20(4):121-312.

Gressitt, J. L. 1961. Problems in the zoogeography of Pacific and Antarctic insects.
Pacific Insects Monogr. 2:1—94.

Guillaumin, A. 1948. Flore analytique et synoptique de la Nouvelle-Calédonie:
Phanerogames. Paris, Office de la Recherche Scientifique Coloniale.

Hochstetter, F. von. 1959. Geology of New Zealand. Transl. and ed. C. A. Fleming.
Wellington, R. E. Owen, Government Pmnter.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography.
Princeton, N.J., Princeton University Press.

McDowall, R. M. 1970. The galaxiid fishes of New Zealand. Bull. Mus. Comp. Zool.
139(7):341-432.

Mertens, R. 1934. Die Insel-Reptilien, ihre Ausbreitung, Variation, und Artbildung.
Zoologica 32(6):1—209.

Nakata, S. 1961. Some notes on the occurrence of Phasmatodea in Oceania. Pacific
Insects Monogr. 2:107—21.

Natusch, S. 1967. Animals of New Zealand. Christchurch, Whitcombe and Tombs.

Oliver, W. R. B. 1955. New Zealand birds. 2d ed. Wellington, A. H. and A. W.
Reed.

Philipson, W. R., and D. Hearn. 1962. Rock garden plants of the southern alps.
Christchurch, Caxton Press.

Sarlin, P. 1954. Bois et foréts de la Nouvelle Calédonie. Nogent-sur-Marne, France,
Centre de Technique Forestier Tropical.

Smith, A. C. 1955. Phanerogam genera with distributions terminating in Fiji. Jour.
Arnold Arboretum 36:273-92.

Solem, A. 1961. New Caledonian land and fresh-water snails. An annotated check
list. Fieldiana, Zool. 41:415-501.

Thorne, R. F. 1965. Floristic relationships of New Caledonia. Univ. Iowa Stud.
Natur. Hist. 20(7):1-14.

Went, F. W. 1971. Parallel evolution. Taxon 20:197—226.

252

APTER
EIGHT

ADAPTIVE RADIATION: THE FLORA
OF SOUTHWESTERN AUSTRALIA

Selection of a single portion of Australia as a site for discussion of adaptive
radiation may seem a dubious exercise at first glance. Conceding that Aus-
tralia can be considered insular, can patterns of southwestern Australia be
examined independently of the remainder of the continent? Many genera
of angiosperms occur in both temperate western and eastern Australia.
The significance of the southwestern flora lies primarily in the extent to
which it is disharmonic, and in the genera that have radiated in a spectacu-
lar fashion in that region. There are a number of aspects in which south-
western Australian plants show insular characteristics in heightened form.
Western Australia was indeed an island in the Cretaceous (Noakes, 1966),
but in the Tertiary, outlines of the continent were not drastically different
from their present extents. However, Western Australia proves to be in-
sular with respect to angiosperms in many ways.

To be sure, Australian vertebrates do not show sharply insular patterns.
Some genera are confined to the West (see Keast, 1961). Some other
vertebrates, such as the marsupial Thylacinus, once occurred in Western
Australia, as fossil records indicate, so significance should not be attached
to absence of these genera. Among angiosperms, however, there is an
appreciable generic endemism and an extremely high species endemism
(see chapter 3), even if we cannot define southwestern Australia precisely
as a province. Also, this region is still poorly known. For example, during
three months of intensive field work in southwestern Australia in 1967,
I was able to augment the family Stylidiaceae by approximately 10 percent:
10 new species of Stylidium, 1 new species of Levenhookia, and 7 new

253

Adaptive Radiation: Flora of Southwestern Australia

subspecies of Stylidium (Carlquist, 1969). Very likely, new species could
be added to many of the larger genera in southwestern Australia.

Southwestern Australia, rather than Western Australia as a whole, is
a natural biogeographic region to consider. ‘To the east and northeast of
the southwestern province, rainfall tapers to extremely low levels, providing
barriers and isolation virtually as effective as seawater in demarcating the
more mesic southwestern flora. The northern portions of Western Australia
are also very dry and have a subtropical rather than temperate regime,
but also have a summer rainfall pattern, rather than a winter pattern as
in the southwest. The frost-free coastal zone extends as far south as Shark
Bay (these data and other facts on temperature, rainfall, and geology cited
below are from the Atlas of Australian resources, 1952-1966). ‘The
northern limit of the southwestern province could conveniently be placed
at Shark Bay, where the 10-in. rainfall isohyet runs, or, somewhat further
south, at the mouth of the Murchison River, which coincides with an
isohyet of 15 in. per year. Along the southern coast of Western Australia,
rainfall decreases eastwardly toward the Nullarbor Plain. The 15-in. isohyet
at Cape Arid would be a convenient boundary for the southwestern
Province. Not far to the east of Cape Arid there is also a sharp change
in geology: a boundary between the granitics of the west and the Nullarbor
limestones, which represent a giant Tertiary embayment. If one follows
the 10-in. isohyet from Shark Bay toward the southeast, one finds that it
includes such localities as Southern Cross and, very nearly, Kalgoorlie.
Localities interior to Kalgoorlie show a sharp reduction in number of plant
species. The subdesert areas from Kalgoorlie and Norseman westward
should be included in the southwestern province.

If defined roughly as above, the southwestern province is an island
of mediterranean climate, well separated from similar climates in eastern
Australia by the long stretches of desert. Rainfall forms an almost evenly
increasing gradient toward the southwestern tip of the continent at Point
D’Entrecasteaux, where rainfall exceeds 60 in. per year. Clearly, rainfall
is not distributed in such a way that any marked moisture microclimates
occur along the increasing gradient to the southwest; however, micro-
climates do result from the variation in moisture availability owing to
different soil types and drainage.

Topography forms few isolating barriers within southwestern Australia,

254

Adaptive Radiation: Flora of Southwestern Australia

for the province is remarkably flat. The Darling Scarp behind Perth is a
minor feature, and the Stirling Range, although the most notable series
of elevations in the southwest, would hardly qualify as mountains (the
highest point, Bluff Knoll, is 3640 ft). Basic rock types, as shown on
geological maps, are remarkably unvaried in the southwestern province.
Most of the area consists of a pre-Cambrian (Archaean) granitic shield.
The only exceptions to this are the Jurassic sandstone and shales, near the
coast from Geraldton south nearly to Perth, and the pre-Cambrian shales
(Proterozoic) of the Stirling Range and a few nearby areas. A strip of the
western coast consists of sand, but this is granitic sand derived from the
ubiquitous granites. The combination of topography and restricted rock
types probably accounts for the remarkable endemism of Darwinia species
on the Stirling Range (fig. 8.23), but Darwinias occur on other topography
and soil types as well in southwestern Australia.

Soils across broad stretches of southwestern Australia are not as
uniform as underlying rock types would suggest, however. Lateralization
is widespread. Laterized soils are reddish, and often contain red-brown
pebbles of laterite. Laterized soils tend to be moderately or strongly
alkaline. On the other hand, granitic sands, varying from bright white to
yellow (at Southern Cross), tend to be acidic. Sand areas, if not well
drained, form “bogs” or ‘‘swamps”’ during the winter, but ironically take on a
desertlike aspect during the summer months. Granitic sands are not, as
previously thought, derived directly from decomposition domes or other
facies, but represent wind-carried accumulations, presumably derived from
former beaches. Areas of granitic sands—known as “‘sand heaths’” to
botanists—are peculiar in their lack of micronutrients, which spared them
from farming until recent discovery of this lack.

The contrast between laterite and granite sand areas is a striking one,
because the species composition of the respective soil types is so different.
One can literally find an assemblage of perhaps a hundred species on an
area of white sand, then cross to a laterite area a few yards away and find
a totally different assemblage of species. Although there are some exten-
sive laterite areas, often these two sand types occur as a series of islands.
In some places, granite sand will be the continuous phase, in others, the
discontinuous phase. Thus southwestern Australia is edaphically a ‘“‘two-
phased archipelago,” and this, together with superimposed rainfall and

255

Adaptive Radiation: Flora of Southwestern Australia

topographic features, accounts for the rather fantastic speciation that has
occurred. Rapid evolution in a mosaic-type soil distribution is stressed by
Snaydon (1970). The relation between edaphic endemism and “catastrophic
selection” was described by Lewis (1962) and further emphasized by Raven
(1964).-Beadle (1966) notes the poor nutrient content of certain soils in
Australia. Although Beadle does not, in my opinion, prove his contention
that phosphate impoverishment leads to xeromorphy and sclerophylly, the
existence of nutrient-poor soils is itself significant.

Southwestern Australia contains areas difficult for establishment of
many angiosperm groups because of relatively low rainfall, high soil
porosity, temperature extremes, and peculiar soil types, some of which are
poor in nutrients. ‘The net consequence is that only a limited number of
establishments have succeeded in the isolated southwestern province. We
have, therefore, a disharmonic condition. For those groups preadapted to
establishment, however, there has been enormous opportunity for radiation.

Beadle (1966) assumes that we must rely on land connection with
Asia during the Cretaceous for entry of 180 angiospem families to Australia.
This is surely false. Many of these angiosperm families are capable of
long-distance dispersal, which explains why a few have strongly disjunct
distributions (e.g., Restionaceae, Rafflesiaceae) that include Australia.
There is no reason whatever to believe that Australia has been less effective
than oceanic islands as a recipient of immigrants via long-distance dispersal.
Australia presents a much larger target area than a volcanic island, and
has had greater longevity as a land surface. In addition, long-distance
dispersal has undoubtedly operated within Australia itself. It has been
operative between the two mesic and sub-mesic temperate areas, south-
western and southeastern Australia. One certainly need not hypothesize
plants growing at ecologically unsuitable mid-stations. Wind-carried seeds
can be picked up, once fallen to a land surface, and carried farther by
successive gusts, an advantage not shared by seeds carried across oceanic
distances. The difficulty of dispersal to, and establishment in, south-
western Australia is great enough, however, that we can justifiably regard
the southwestern flora as a “flora of long-distance dispersal” (that is, a flora
produced by evolution of ancestors that arrived via long-distance dispersal).

For example, Dodonaea viscosa, a common lowland plant of Pacific

256

Adaptive Radiation: Flora of Southwestern Australia

islands, may have arrived on maritime sites, such as limestone bluffs just
south of Geraldton. Localities with alkaline and xeric qualities like those
of the immediate coast occur in inland areas as “‘islands”: pockets of
alkaline laterite, margins of former salt lakes, etc. Rapid evolution in one
of these edaphic islands, followed by dispersal to others and further change
there, may explain the varied and variable species of Dodonaea in the
southwest. Likewise, it would explain the formation of numerous inland
species of Frankenia, typically a beach genus, in this province. Similar
considerations undoubtedly apply to other halophyte genera that have
speciated markedly in southwestern Australia: Bassia, Chenopodium, and
Atriplex of the Chenopodiaceae; and Kochia and Ptilotus (Trichinium)
of the Amaranthaceae. Those groups characteristic of the white sand
“winter bogs,” such as Drosera (fig. $.29, bottom left, nght), would likewise
react to the insular distribution of white sand. Numerous small populations,
interconnected by events of dispersal, form a scheme ideal for rapid
evolution, according to many evolutionists. In this case, the distribution
of white sand pockets in southwestern Australia would be perfect.

Once a species has diverged from parent stocks and been separated
by formation of sterility barriers from similar species, it can redisperse to
areas occupied by ancestral species and become sympatric. During my
field work in the southwest, I frequently found several species of Stylidium
growing together and often flowering at the same time, but interspecific
hybrids were extremely rare. I attribute this lack mainly to sterility of
hybrids. There are other means of barrier formation among species, how-
ever. In Stylidium, for example, sympatric species can have different
flowering times. Also, distinctive flower patterns with respect to flower
size, corolla-lobe shape, color, and display position, and length and position
of the column may make cross-pollination impossible in this genus
(Carlquist, 1969). Some species of Stylidium have developed selfing
mechanisms (Carlquist, 1969); flower-constancy in pollinating insects for
this genus is apparently also a reality (Erickson, 1958). There has been an
evolutionary shift to new insect pollinators in some species, such as
S. insensitivum (Carlquist, 1969).

Even viewing southwestern Australia in an oversimplified scheme of
archipelagic patterns of granitic sand and laterite, one can see other

257

Adaptive Radiation: Flora of Southwestern Australia

mechanisms by which soil patterns could promote rapid evolution. If white
granitic sand occurs in pockets within a laterite area, the fringes of the
sand pockets could offer intermediate soil types. Ecotypes tolerant to
sand-laterite mixtures could develop in a sand-inhabiting species, and from
the intermediates a laterite species could develop. In Drosera, this may
have happened; although most species are characteristic of sand heaths,
some occur on laterite soils or, like D. macrophylla, are tolerant of both
or of various soil mixtures (Erickson, 1968). In Stylidium, the peculiar
S. preissii is a sand species, although the only closely related species, S.
imbricatum, occurs on laterite gravels, but only where sand has infiltrated
the gravels. If some species of a genus are sand species, others laterite
species (as is true in Stylidium), both phases of the soil mosaic can serve
as “islands” stimulating evolution.

Even if climatic factors such as temperature and rainfall form a
gradual gradient from the southwest coast inland, these gradients can still
be entirely operative as selective factors. One can envision numerous
genera of flowering plants evolving into drier habitats, or, on the other
hand, more mesic habitats, by forming successively more xeromorphic or
mesomorphic ecotypes across distances of hundreds of miles. In some
places, these gradients can be very sharp. For example, Esperance is a
coastal locality that is relatively mesic compared with desertlike areas only
a short distance inland.

Evolutionary trends in the southwestern flora are surely not uni-
directional. There is no reason to suppose that mesomorphy is always
more primitive, xeromorphy secondary, even within a single genus. In
fact, some plants of higher rainfall areas seem to be recent invasions. One
might expect this because, in area alone, the broader zones of dry country
would offer larger bases for establishment than the relatively wet south-
western corner. Dry areas, with less or no soil profile development, would
also be expected to be “open communities,” suitable for pioneers, as
compared with the well-developed podzolized soils of the wet karri
forests of southwesternmost Australia. Shorter dispersal distances would
be required for migration of a subdesert species from east to west (or the
reverse) in temperate Australia, because more mesic areas are localized in
extreme southwest and southeast.

258

Adaptive Radiation: Flora of Southwestern Australia

Radiation in Taxonomic Groups

One can view adaptive radiation in southwestern Australia in terms of
what growth forms and other modes of structure have been produced in
particular genera, or one can view these phenomena one by one. I am
taking both viewpoints. The first illustrates most clearly that any given
genus covers a range of phenomena but not all of them. The latter,
presentation of morphological modes, suggests what significance these
may have in terms of adaptations, and how these form a remarkable array
of parallelisms. Information on plants of southwestern Australia has been
derived from my fieldwork there in 1962 and 1967, and also from the
valuable books of Blackall and Grieve (1956, 1959, 1965) and Beard (1965).
Beard offers rough data on ranges and very brief descriptions, whereas the
pictorial keys of Blackall and Grieve contain condensed visual information.

DAVIESIA (FABACEAE)

Daviesia, a genus of small shrubs, is represented by a few species in
eastern Australia. The southwestern species, however, are more numerous
and remarkably diverse (figs. 8.1-8.3). Virtually all of them represent
various types of reduction or condensation of photosynthetic surfaces—the
diversity within this general phenomenon is amazing. Few of the south-
western species have true leaves, but those that do show restriction in
form. Daviesia obtusifolia (fig. 8.1, top left) has oblanceolate leaves with
both surfaces identical (isolateral), like those of many eucalypt species. ‘The
wedge-shaped leaves of D. dielsii (fig. 8.1, top right) are also isolateral,
tather thick, each tipped by a spine. Leaves of D. teretifolia (fig. 8.1,
bottom left) are needlelike and tend to point upwards (they are thus
minimally exposed to sunlight), and each is tipped by a spine. Daviesia
pachyphylla (fig. 8.1, bottom right) is a sparsely branched shrub with
branches densely covered by leaves not unlike rose thorns, but with a more
inflated appearance. The interior of the leaves is spongy, but the leaves
are very stiff, not succulent, thus making the spine at the tip effectively
pungent. Narrower and sparser—but clearly spine-tipped and stout—are
leaves of D. polyphylla (fig. 8.2, top left). Leaves of D. polyphylla do not

259

FIG. 8.1. Species of Daviesia (Fabaceae) from Western Australia. Top Lert, D. obtusifolia,
leaves 6 cm long (from a color photograph by Fred W. Humphreys, courtesy of Mrs. Evelyn
Humphreys). rop ricut, D. dielsii, wedge-shaped leaves 15 mm long; between Eneabba and
Gingin. Bottom Lert, D. teretifolia, needlelike leaves 60 mm long; between Gibson and
Scaddan. sorrom ricut, D. pachyphylla, conical leaves 15 mm long; between Ravensthorpe
and Hopetoun.

260

FIG. 8.2. Species of Daviesia (Fabaceae) from Western Australia. top LeFT, D. polyphylla,

pinelike leaves 20 mm long; Cape Le Grand. ror ricut, D. retroflexa, spinelike leaves 25
mm long; between Ravensthorpe and Esperance. BOTTOM LEFT, D. divaricata; near Mullewa.
BOTTOM RIcHT, D. aphylla, broom shrub in flower; near Melville.

261

FIG. 8.3. Fabaceae from Western Australia. trop LEFT, Daviesia euphorbioides, stems 1 cm
in diam.; Wongan Hills. rop ricut, D. trigonophylla, stems angled because of decurrent
leaf bases; Chester Pass. Borrom Lert, D. epiphylla, flattened stems 2-5 cm wide (from a
color photograph by Fred W. Humphreys, courtesy of Mrs. Evelyn Humphreys). Borrom
RIGHT, Brachysema aphyllum, flattened branch 15 mm wide, bearing a flower 60 mm long;
near Northampton.

262

Adaptive Radiation: Flora of Southwestern Australia

appear demarcated from the stem by any external juncture. This is also
true of D. retroflexa (fig. 8.2, top right), the leaves of which point down-
wards. The angle of the leaves suggests that they might be effective as
climbing hooks, although D. retroflexa does not often appear to be a
climbing shrub. Daviesia aphylla (fig. 8.2, bottom right) and D. divaricata
(fig. 8.2, bottom left) could be called broom shrubs. Stems of D. aphylla
are leafless, as the name suggests. Stems of D. divaricata are more branched,
producing a tangle of stems, and terminate in thomlike fashion. Daviesia
divaricata should probably be termed a thorn shrub rather than a broom
shrub.

Daviesia euphorbioides (fig. 8.3, top left) is nearly leafless: the leaf
vestiges are spinelike. Stems of D. euphorbioides are expanded, chiefly by
means of a broadened pith, and the plant is thus rather like an “inflated”
broom shrub. This species cannot be called a true succulent. Daviesia
trigonophylla (fig. 8.3, top nght) reminds one of D. polyphylla, but leaf
bases are decurrent and fused into three longitudinal ridges on each stem,
so that the stems are cladode-like. More definitively a cladode plant is
D. epiphylla (fig. 8.3, bottom left), with glaucous flattened branches,
capable of photosynthesis on all surfaces. These branches can be considered
cladodes, but the spine-tipped lobes probably represent leaf portions,
whereas in a true cladode plant a flattened stem lacking leaves is ordinarily
implied.

The various forms illustrated in Daviesia do not seem related to
habitat differentially. One might expect leafy species such as D. obtusifolia
to be from a more mesic habitat, D. polyphylla to occur in xeric localities.
Although perhaps the range in climate 1s not great, D. obtusifolia extends
from the Stirling area inland to the Coolgardie subdeserts. Daviesia
polyphylla occurs at the margins of swamps, as at Cape Le Grand, but
also in shade of jarrah (Eucalyptus marginata or E. rostrata) forest else-
where. However, all the species of Daviesia are presumably subject to about
equal herbivore pressure, and all of those illustrated seem to lack tender
leaves yet have spines and thorns. The species illustrated can be interpreted
as various responses to herbivore pressure, as well as expressions of
xeromorphy. Perhaps all the habitats of these species can be considered
xeric at least during the warmest months of the year.

263

Adaptive Radiation: Flora of Southwestern Australia

BRACHYSEMA (FABACEAE)

Brachysema (figs. 8.3, bottom right; 8.4; 8.5) is characteristic of south-
western Australia, although it also occurs in northeastern Australia. Thus,
as in Daviesia, most although perhaps not all of the morphological features
of Brachysema may be autochthonous developments in the southwest.
Brachysema in southwestern Australia is notable for having the most dis-
parate types of habit. For example, B. aphyllum (fig. 8.3, bottom right) rep-
resents true cladode formation; vestigial leaves can be seen on margins of
the long cladodes, which tend to sprawl, forming a prostrate shrub. How-
ever, there are species of Brachysema with true leaves. Brachysema sericea
(fig. 8.4, top left) grows in seasonally wet sand heath, so its broad-leaved
habit does correlate with mesic habitat. This is also true of other leafy
species: B. lanceolatum (fig. 8.4, top right) tends to be scandent in other
shrubs in eucalypt woodland; B. bracteolosum (fig. 8.4, bottom left), a shrub
to 2 m in the Stirlings, also occurs with mesomorphic species. Brachysema
praemorsum (fig. 8.4, bottom left) is a subshrub, an understory element
in shady woodland. Brachysema daviesioides (fig. 8.5, top left, nght) and
B. tomentosum (fig. 8.5, bottom left, right) are both peculiar, however, in
being subshrubs with very large flowers borne on inflorescences that
represent specialized shoots formed a little below ground level. Brachysema
daviesioides is a leafless thorn shrub, with scarlet flowers forming a circle
as they emerge from the ground at the base of the plant. Brachysema
tomentosum has flattened stems, and is thereby a cladode shrub or sub-
shrub. Flowers are pale yellow green. Both of these species occur in
relatively dry, open areas, so reduction in vegetative surface is under-
standable. Brachysema daviesioides would seem more adapted to herbivore
deterrence. Why large curved flowers, seemingly ornithophilous, should
occur at ground level is rather mysterious, and is one of many unsolved
questions about this plant that observational work might be able to clarify.

BANKSIA (PROTEACEAE)

Whereas species of Banksia are shrubs of various sizes in other parts
of Australia and southernmost New Guinea, the southwestern species show
a maximal divergence in habit (fig. 8.6). Banksia grandis (fig. 8.6, top left)
is a true tree, to 50 ft. Typical of shrubby species are B. sphaerocarpa (fig.
8.6, top right), which may be under 1 ft high at maturity in some

264

Fic. 8.4. Species of Brachysema (Fabaceae) from Western Australia. Top LEFT, B. sericea,
leaves about 4 cm long; near Cranbrook. trop ricut, B. lanceolatum, leaves white on reverse,
m long; Toolbrunup. Bortom ricut, B. praemorsum, leaves about 22 mm long (from a color

about 5 cm long; cultivated, Los Angeles Arboretum. sortom LEFT, B. bracteolosum, leaves 8
photograph by Fred W. Humphreys, courtesy of Mrs. Evelyn Humphreys).

265

W ‘at

FIG. 8.5. Species of Brachysema (Fabaceae) from Western Australia. TOP LEFT, RIGHT, B.
daviesioides. TOP LEFT, habit of plant, about 0.5 m tall; near Perenjori. rop ricHT, bright red
flower, 60 mm long, at base of plant. BOTTOM LEFT, RIGHT, B. tomentosum. BOTTOM LEFT,

habit, showing flattened stems, about 0.5 m tall; near Southern Cross. BorTroM ricuT, base of
plant, showing pale yellow brown flowers 70 mm long.

266

FIG. 8.6. Species of Banksia (Proteaceae) from Western Australia. Top LEFT, B. grandis,
habit of tree, 7 m tall; near Yanchep. top ricut, B. sphaerocarpa, shrub in flower, 1.5 m
tall; near Dandaragan. spotrom LEFT, B. nutans, branches showing pendant inflorescences;

cultivated, King’s Park, Perth. sorrom ricut, B. prostrata, habit of plant (for size of leaves,
see fig. 8.13), showing head of flowers and head of fruits at ground level; near Scaddan.

267

Adaptive Radiation: Flora of Southwestern Australia

populations, to 2 ft in others; and B. nutans (fig. 8.6, bottom left), a shrub
about 1 m high in which inflorescences are nodding and tend to be borne
near the ground. At the opposite extreme from B. grandis is B. prostrata
(fig. 8.6, bottom nght), a plant that can only be called an “underground
shrub.” Its stems are stout and woody but always grow as stolons below
the soil surface, with only the large leaves exposed—a habit reminiscent
of the bracken fern, Pteridium. Inflorescences terminate shoots, and are
borne at ground level, thus providing another example of a seemingly
ornithophilous flower borne at the soil surface. The ecological significance
of the various growth forms seems clear enough. Banksia grandis occupies
relatively mesic sites along the western coast to the north and south of
Perth. Other tree species of Banksia in the southwest include B. menziesii
(fig. 8.32, top), B. prionotes, B. speciosa, and B. ilicifolia. Although these
all are plants of white granitic sands, none occurs very far from the coast,
whereas shrubby species can occur farther inland. Banksia prostrata may
represent a fire adaptation, for all renewal buds are below the ground
surface.

OTHERS

Perhaps the most outstanding example of diversity in growth form
and tendencies in other vegetative morphological details is Acacia, but
this genus is pan-Australian. The reader can obtain a picture of its scope
(which is greatest in Western Australia) by consulting Blackall and Grieve
(1959). Many other good examples of adaptive radiation in habit are
discussed below in terms of particular tendencies; such genera include
Scaevola and Dampiera (Goodeniaceae), Drosera (Droseraceae), Come-
sperma (Polygalaceae), Hakea and Grevillea (Broteaceae), Tetratheca
(Tremandraceae), and Stylidium (Stylidiaceae). This list is by no means
complete, and virtually any angiosperm genus in southwestern Australia
tends to contain a range in growth form.

Modes of Adaptation
BROOM SHRUBS
As noted by Diels (1906), broom shrubs are common in the south-
western province. Four are illustrated here (fig. 8.7). Scaevola restiacea

268

Fic. 8.7. Broom plants from Western Australia. top LEFT, Scaevola restiacea (Goodeniaceae),
flowers about 18 mm long; near Merredin. rop ricut, Dampiera juncea (Goodeniaceae),
flowers about 15 mm long; near Moora. BOTTOM LEFT, Comesperma scoparium (Polygalaceae),
flowers about 12 mm long; near Merredin. sorrom ricuT, Tetratheca efoliata (Tremandra-
ceae), flowers 15 mm in diam.; south of Southern Cross.

269

*§
%

Fae
a ie | >

oe te

Eo nrg

FIG. 8.8. Cladode plants from Western Australia. Top LEFT, RIGHT, Dampiera alata (Goode-
niaceae). TOP LEFT, habit of plant, branched from near ground, ca. 0.6 m tall; Wongan Hills.
TOP RIGHT, close view of stems, about 15 mm at widest. BorromM LEFT, Tetratheca affinis
(Tremandraceae), flowers are pendant; north of Mt. Barker. Borrom ricut, Platysace com-
pressa (Apiaceae), cladodes are thickish, about 15 mm at widest; Cape Le Grand.

270

Adaptive Radiation: Flora of Southwestern Australia

(fig. 8.7, top left) and Dampiera juncea (fig. 8.7, top right) illustrate that
this habit has been evolved independently in two genera of Goodeniaceae;
a third instance is Leschenaultia juncea. Scaevola tortuosa is a broom shrub
probably close to S. restiacea. Comesperma scoparius (fig. 8.7, bottom left)
is interesting, because the remainder of this polygalaceous genus does not
have this habit. This is also true of Tetratheca efoliata (fig. 8.7, bottom
right), the sole broom shrub of the Tremandraceae. Other broom shrubs
in the southwestern flora include Leptomeria spinosa and L. pauciflora
(Santalaceae); Exocarpus spartea and E. aphylla (Santalaceae); Choretrum
(four spp.; Santalaceae); Psammmomoya choretroides (Celastraceae); Boronta
juncea and B. defoliata (Rutaceae); Logania nuda (Loganiaceae); Sphaero-
lobium yimineum, Jacksonia capitata, and Daviesia aphyllum (fig. 8.2,
bottom right) of the Fabaceae; Stylidium junceum and S. squamosotuber-
osum (Stylidiaceae [Carlquist, 1969]). Although monocotyledons cannot
be considered shrubs, many Restiaceae and Cyperaceae could be considered
to represent this adaptive mode. Other angiosperms could easily be added
to the list. Although it is difficult to generalize, most often the broom
shrubs in southwestern Australia tend to occupy more inland, hotter
habitats than do their leafy congeners. For example, Tetratheca efoliata
occurs inland near Southern Cross, whereas other species occur closer to
the coast. Another potential value of the adaptation is that the habit
permits plants to limit transpiration greatly, so that flowering can occur
later in the season. Another possibility is that the plant does not die to
ground level as soon. Renewal buds of most of the broom shrubs mentioned
occur at or below ground level. In addition to reduction of transpirational
surface, broom shrubs tend to be very fibrous in structure and thereby may
excel at herbivore resistance; lack of leaves would definitely seem to con-
tribute to lack of visible appeal for herbivores. The proportion of plants
with broom-type construction seems higher in southwestern Australia than
elsewhere in the world, although comparable figures are not available.

CLADODE SHRUBS
Certainly southwestern Australia exceeds other areas in the propor-
tion or total number of angiosperms in which flattened stems, usually
with vestigial leaves, form the vegetative system of the plant. These occur,
for example, in Dampiera alata (fig. 8.8, top left, right). Interestingly, three

eff!

Adaptive Radiation: Flora of Southwestern Australia

other Dampiera species may have acquired this habit independently:
D. trigona var. latealata, D. lindleyi, and D. epiphylloidea. Tetratheca
affinis (fig. 8.8, bottom left) is the sole species of its family (Tremandra-
ceae) to have acquired the cladode habit. One would not expect a mem-
ber of Apiaceae to form cladodes, but Platysace compressa (fig. 8.8, bot-
tom right) clearly does. Although one might not term this perennial a shrub
because it is not woody, it has all the appearances of a shrub, forming
tangled mounds, as on the granitic domes of Cape Le Grand. The cladodes
of P. compressa are quite succulent and resist drying. Legumes (Fabaceae)
seem particularly subject to cladode formation, as shown by figures 8.3,
bottom left and nght, and 8.9. Different modes of cladode formation can
be seen. Sphaerolobium alatum (fig. 8.9, top left) has thick, woody, glau-
cous cladodes, broadly elliptical in transection. Bossiaea walkeri (fig. 8.9,
top right) has thin wings, especially in internodal regions, along the length
of the upright stems. Cladodes of Acacia glaucoptera (fig. 8.9, bottom left)
are very broad and leaflike; the marginal teeth may represent participation
of phyllode vestiges in the cladode. Cladodes of Jacksonia eremodendron
(fig. 8.9, bottom right) resemble prickly leaves very closely and could easily
be mistaken for leaves, except for their position axillary to leaf vestiges
and the fact that flowers are borne along margins of the cladodes. Other
legumes from southwestern Australia that form cladodes include additional
species of Jacksonia (J. alata, ]. compressa, ]. pteroclada, ]. carduacea,
J. decumbens, J. floribunda), two species of Templetonia (T. sulcata, T.
egena), and several legumes illustrated earlier: Brachysema aphyllum (fig.
8.3, bottom right), B. tomentosum (fig. 8.5, bottom left, right), and Davie-
sia epiphylla (fig. 8.3, bottom left). Acacia is the legume genus of south-
western Australia with the largest number of cladode-forming species, how-
ever—Acacia alata, A. diptera, A. ataxiphylla, A. trigonophylla, and A.
stenoptera, for example. A monocotyledon, Anarthria scabra (Restiaceae),
can qualify as a cladode-former because it has flattened inflorescence axes.

The significance of cladode species is not easy to assess, for they are
by no means sympatric. However, if we compare the ranges of cladode
species to those of, say, broom shrubs or thorn shrubs, we find cladode spe-
cies more nearly coastal or mesic in habitat preference. This is clear in
Acacia glaucoptera, Tetratheca affinis, Sphaerolobium alatum, Daviesia
epiphylla, and Platysace compressa. More nearly coastal preference seems

272

sth alll
em

iat

Fic. 8.9. Cladode-forming species of Fabaceae from Western Australia. Top LEFT, Sphaerolo-
bium alatum, stems about 1 cm wide; Mt. Barker. rop ricut, Bossiaea walkeri, glaucous flat-
tish branches about 1 cm wide; north of Norseman. porrom LEFT, Acacia glaucoptera, flat-
tened stems incorporating leaf bases, typically 5 cm at widest; west of Esperance. BOTTOM
RIGHT, Jacksonia eremodendron, prickly leaflike cladodes about 4 cm wide; Badgingarra.

273

Fic. 8.10. Thorn shrubs from Western Australia. TOP LEFT, RIGHT, Comesperma spinosum
(Polygalaceae). Top LeFr, habit, about 0.5 m tall; Ghooli. rop ricHT, view of twigs and
flowers, which are 7 mm long. BoTroM LEFT, Scaevola oxyclona (Goodeniaceae), showing
tanged twigs; bud of flower is 8 mm long; Spargoville. BoTTOM RIGHT, S. spinescens, twigs,
showing succulent leaves; twigs terminate in spines, not abundantly shown here; leaves about
1.5 cm long; north of Morawa.

7p

Adaptive Radiation: Flora of Southwestern Australia

evident in the cladode-forming species of Dampiera. There are exceptions:
Bossiaea walkeri is an inland species, but its thick, glaucous cladodes make
it resemble a woody succulent. In general, I am tempted to say that clad-
odes can only be expansions of stems from-ancestors with broomlike or
microphyllous habits. Cladode species might represent invasions of more
mesic areas, therefore. Comparisons with closely related species in respec-
tive genera would be very informative. Cladodes appear to have greater
duration—even for more than a single year, as in Acacia glaucoptera—
and thus represent less fugacious structures than leaves. Functionally,
cladodes are rather like phyllodes in isolateral structure, although presum-
ably with greater longevity. Cladode formation may represent moderate
increase in surface area in plants that have lost the capability of forming
normal leaves or laminae. Cladodes would seem well adapted to a medi-
terranean climate, where limited surface area can exist in conditions of
warm temperature and low humidity, and such plants would not be killed
by frosts. Virtually all of the southwestern cladode plants occur in the
region noted as having fewer than 5 days of frost per year.

THORN SHRUBS

The only obvious common denominator of thorn shrubs is that branch
tips are converted into thorns. There is an obvious alliance with plants in
which leaves are converted into spines, or even plants with prickly leaves:
the purpose of all these devices seems to be herbivore deterrence. Thorn
shrubs do have reduced leaves, and are often aphyllous. If aphyllous, they
differ little, physiologically, from broom shrubs; they differ in appearance
in that their branches form a tangle. Comesperma spinosum (fig. 8.10, top
left, right) is a good example of this growth form, as compared with C. sco-
parium (fig. 8.7, bottom left). The same is true of Scaevola oxyclona (fig.
8.10, bottom left) as compared to S. restiacea (fig. 8.7, top left) or S. tor-
tuosa. In other cases, like S. spinescens (fig. 8.10, bottom left), leaves are
present—semisucculent, rather fugacious. Thorn shrubs in the southwest-
ern Australian flora are almost too numerous to mention, but the list would
include Leptomeria exocarpoides (Santalaceae); Hybanthus epacroides
(Violaceae); Grevillea inconspicua (Proteaceae); Prostanthera tysoni (La-
miaceae); Plagianthus incanus and P. microphylla (Malvaceae); Blackallia
connata (Rhamnaceae); virtually all species of Cryptandra (Rhamnaceae);

275

Adaptive Radiation: Flora of Southwestern Australia

Psammomoya ephedroides (Celastraceae); Eriostemon deserti (Rutaceae);
Leptospermum spinescens, Melaleuca thymoides and many other Myrta-
ceae; and a large number of legumes, including Acacia exocarpoides, A.
nodiflora, A. congesta, A. cuneata, Bossiaea preissii, B. concinna, B. pedun-
cularis, B. leptacantha, B. rufa, Jacksonia foliosa (and other Jacksonia spe-
cies), Mirbelia spinosa and other Mirbelia species. ‘Thorn shrubs in south-
western Australia do connote xeromorphy, not only because of aphylly or
microphylly, but also because they suggest greater herbivore retardation,
as would be expected in drier localities where herbivore pressure on limited
available vegetation would be expected to be greatest. In fact, the thorn
shrubs listed above do, as a whole, tend to occupy sites farther inland than
do their equivalents in broad-leaved and cladode-forming habits.

TWINING PLANTS

Twining or scandent plants in southwestern Australia are surprising
because they belong to genera and even families in which the twining habit
is unknown in representatives outside of this area. To be sure, vining repre-
sentatives are to be expected in many groups; Kennedya and Hardenbergia
of the Fabaceae, for example. Likewise, Cassytha (Lauraceae) is a genus
of twining parasites; the presence of a large number of species (eight) in
the southwest is noteworthy, however. A genus such as Comesperma (Poly-
galaceae), already cited as containing a thorn shrub and a broom shrub,
might not be expected to include a vining species, but it does, C. volubile
(fig. 8.11, top left). Pittosporaceae is certainly not a family of scandent
plants, but twining representatives include Marianthus erubescens (fig.
8.11, top right) and most other species of Marianthus; Sollya fusiformuis;
Cheiranthera filifolia and C. parviflora; Pronaya elegans; and several spe-
cies of Billardiera. Drosera would not be expected to be a scandent genus
from its behavior elsewhere in the world, but the vining species (approxi-
mately 10 of them) are, in fact, the most conspicuous Droseras (fig. 8.11,
bottom left) in southwestern Australia. One would not expect vining spe-
cles in a genus commonly regarded as consisting of trees and shrubs, Aca-
cid, but two such species occur in the southwest (A. volubilis, A. inops).
Stylidium laciniatum, a narrow endemic of the Frankland River area, is
habitally unique in its family. In S. laciniatum, inflorescence axes are
markedly elongate, as in the closely related S. junceum; however, they twine

276

LEFT, Comesperma volubile (Poly-
shown, flowers about 1 cm long;
nbarker Rd., between Denmark and Mt. Barker. rop ricut, Marianthus erubescens (Pitto-

(continued)

Lie,

Adaptive Radiation: Flora of Southwestern Australia

upward on restiads or shrubs, and flower upon reaching the top of the
forest understory. Utricularia volubilis, from wet flats near the coast from
Perth to Albany, is also unique in its family (Lentibulariaceae) in its twin-
ing inflorescences, which reach 90 cm in length and climb on stems of
aquatic monocotyledons (Erickson, 1968). It grows in deep water and has
polymorphism in its insect traps, depending on how deeply they are im-
mersed beneath the water surface.

Thysanotus (Liliaceae) has various habits; most species are typical
tufted plants, with bulbs or fibrous roots and upright scapes. In T. pater-
soni (fig. 8.11, bottom right), however, the stems are twining, leafy, and
of almost indefinite extent, often forming masses on shrubs. Several resti-
ads have spiral or twining stems; Calorhaphus gracillima, Leptocarpus
tenax, and Loxocarya flexuosa, for example. ‘These may not be true climb-
ing plants, but may rather exhibit “three-dimensionalization” of the vege-
tative form for other reasons discussed below under Leaf Morphology.

Among orchids, Thelymitra flexuosa has twisted stems, and T. spiralis has
leaves with a helical form.

LEAF MORPHOLOGY

That plants of the southwest should exhibit a variety of leaf types is
not surprising. The features of interest are that (1) such widely divergent
leaf types occur within single genera, and evidently most of them represent
autochthonous developments; (2) some of these leaf types are unique, not
found elsewhere; (3) with all genera compared, there is such a high degree
of parallelism in types, or “convergent evolution”; and (4) there is appar-
ently easy reversibility in morphology, and leaf morphology, a rapidly
evolving feature, can exhibit sensitive adaptation for a particular habitat,
although it may in some cases lag behind physiological adaptation, and
various leaf types are equally successful in a given locality. Some bizarre
types may have a generalized adaptive value, but their form in detail can-
not be explained in adaptive terms of a single factor or a few factors, and
probably must involve the nature of the leaf type ancestral to that of a

FIc. 8.11 (continued)

sporaceae), twining stems and flowers, which are about 5 cm long; Miling—Pithara road.
BOTTOM LEFT, Drosera heteromera (Droseraceae), tip of twining stem with flower 2 cm in
diam.; near Toodyay. BoTromM RIGHT, Thysanotus patersonii (Liliaceae), flowers 2 cm in diam.
and twining stems; near Perenjon.

278

Fic. 8.12. Plants illustrating evolution of particular leaf types in Western Australia. Top
LEFT, Chorilaena quercifolia (Rutaceae), head of flowers and lobate leaves 5 cm long, covered
by stellate hairs. rop ricur, Dampiera hederacea (Goodeniaceae), showing lobate leaves 4
cm long, covered by stellate hairs; near Pemberton. Bortom LEFT, Scaevola crassifolia (Goode-
niaceae), broad leaves with glutinous surfaces, typically 5-6 cm long; beach at Mandurah.

BOTTOM RIGHT, Dampiera teres (Goodeniaceae), leaves terete or slightly flattened, 2 cm long;
Dandaragan.

279

B. speciosa

B. dryandroides

see,

B. boxteri

B. brownii

B. ilicifolia

B. coccinea

B. solandri

Fic. 8.13. Leaves of various species of Banksia (Proteaceae) to show diversity of species in
southwestern Australia. Further explanation, describing ecological correlations of these species,
is in the text.

280

B. baueri

B. prostrata

B. elderiana B. grandis

261

Adaptive Radiation: Flora of Southwestern Australia

contemporary species. For example, one can offer correlations between ecol-
ogy and leaf morphology in Banksia (fig. 8.13). But one cannot present
precise, detailed correlations. Banksia coccinea has leaves different from
those of B. baxteri, but the two can grow together, as at Hopetoun (how-
ever, one may note a similar area and thickness of leaf despite difference
in form).

Banksia (fig. 8.13) and Hakea (fig. 8.14) are presented here as examples
of the wide gamut to be found within particular genera—although even for
these genera, the figures by no means include all the southwestern species.
In other photographs stressing leaf morphology (figs. 8.12, 8.15-8.20), an
attempt has been made to compare similar features shown by unrelated
genera. We can note briefly the parallelisms in terete and hollylike leaves
cited by Went (1971), but each of these categories itself contains great
diversity and is not mutually exclusive but intercontinuous with other
types, and can, contrary to Went’s claim, easily be explained by ordinary
selective factors. We do not need hypotheses of “floating genes” or “‘ortho-
genesis” at all. Such hypotheses tend to thrive best when examples are
selected incompletely or in a preconceived manner.

Broad Leaves. Broad, lobate leaves are relatively infrequent in the
southwestern flora as a whole. Such leaves, which often have stellate pubes-
cence, are surprisingly characteristic of the southwesternmost forests, par-
ticularly the karn (Eucalyptus diversicolor) forests. Dampiera hederacea
(fig. 8.12, top nght) of the Goodeniaceae and Chorilaena querctfolia (fig.
8.12, top left) of the Rutaceae are typical of this mode. Other “oak-leaved”’
species of these forests include Dampiera phlebopetala and D. striata; in
Sterculiaceae, Lasiopetalum floribundum, Rulingia grandiflora, Thomasia
quercifolia, and T. solanacea. The endemic genus Tremandra (Treman-
draceae) is a good example of this syndrome. Not surprisingly, the species
of Banksia with the most oaklike leaves, B. solandri and B. coccinea (fig.
8.13), are characteristic of moist forests in and near the Stirling Range.
Banksia grandis and B. prostrata (fig. 8.13), with the greatest leaf surface
of any Banksia species, occur on moister coastal plain. The largest-leaved
species of Scaevola in southwestern Australia, S. crassifolia (fig. 8.12, bot-
tom right), S. nitida, and S. porocarya, are all coastal species. Coastal and
moister sites are the habitats occupied by Hakea species with the broadest
leaves: H. conchifolia (fig. 8.15, top left), H. cucullata (fig. 8.15, top right),

282

Adaptive Radiation: Flora of Southwestern Australia:

and H. victoriae (fig. 8.15, bottom left), as well as H. amplexicaulis, H. roet,
H. smilacifolia, H. elliptica, H. undulata, H. auriculata, H. ferruginea, H.
hookeriana, H. baxteri, and H. flabellifolia (fig. 8.14). Some narrower-leaved
species of Hakea can also be found in these habitats, but the broadest-
leaved species are never in drier, inland localities.

Hakea trifurcata (fig. 8.14) provides an interesting problem because
it produces not only trifurcate leaves with terete leaves, but also a limited
number of broad elliptical leaves, with only a few intermediate types. The
broader leaves seem to be formed during moister portions of the year, not
during the dner portion of the growing season.

Isolateral Leaves. As Diels (1906) notes, isolateral leaf construction is
extraordinarily common in southwestern Australia. Acacia and Eucalyptus
contain many examples of this form, both in the southwest and in other
parts of Australia. Of course, needle-shaped leaves, or leaves with terete
segments, are by definition isolateral. Isolateral leaves occur scattered
through genera in which bifacial leaves predominate; Dampiera teres (fig.
8.12, bottom right), for example, is an isolateral-leaved species. Examples
of isolateral leaves (acicular and otherwise) have been cited in Daviesia
and are obvious in the illustrations of Hakea (fig. 8.14). Broom shrubs
are stem equivalents of isolateral leaves. Abundance of isolateral leaves
in various genera of Australia is related to mediterranean climate: exposure
of minimal leaf surface with maximal photosynthetic layers; isolateral
leaves are, with few exceptions, displayed at various angles other than more-
or-less horizontal.

With respect to isolateralleaved species that occur in genera with
other laminar types, I have repeatedly had the impression that some of
these broader leaves have been derived phylogenetically from narrower-
leaved ancestors. If we suppose that acicular or linear leaves are ancestral
in certain southwestern Australian groups, broader leaves would represent
evolution toward mesomorphy. I have gained this impression where leaf
venation is markedly different from the venation typical for dicotyledons
(a central midvein, with pinnate arrangement of lateral veins and reticu-
late areoles). This “typical” condition is found in Hakea amplexicaulis,
A. roei, and H. auriculata (fig. 8.14). Yet how can we explain the longitu-
dinally oriented veins characteristic of leaves in the well-named H. multili-
neata (fig. $.14), or the broader-leaved species H. elliptica, H. undulata,

283

— ———

H. lehmanniana H. brookeana

H. corymbosa H. hookeriana

H. multilineata

H. baxteri H. flabellifolia H. trifurcata

FIG. 8.14. Leaves of various species of Hakea (Proteaceae) to show several evolutionary tend-
encies in this genus in southwestern Australia. Comments on the ecological correlation and
morphological significance of these various types are in the text.

284

H. undulata

H. auriculata

Adaptive Radiation: Flora of Southwestern Australia

and H. ferruginea (fig. 8.14), in which transverse veins are frequent but
several main “parallel” veins are predominant? Likewise, how can one ex-
plain leaves with flabellate venation, such as H. baxteri and H. flabellifolia
(fig. 8.14)? Parallel-veined leaves of some Hakea species, including H.
hookeriana (fig. 8.14), H. laurina, and H. pandanicarpa, are remarkably
reminiscent of phyllodes of Acacia, both in form and in anatomy (isolateral
structure, etc.). I feel there is a distinct possibility that certain broader
leaves in Hakea may represent expansion from acicular or linear leaves,
and may be an accompaniment of ecological shift into moister regions,
which form a relatively small area in Western Australia. If so, such
broader leaves are much like phyllodes and could perhaps be called “pseu-
dophyllodes.”” If my interpretation is correct, they would not be true
phyllodes, as in Acacia, in which the leaf surface results from phylogenetic
expansion of petioles. Some Myrtaceae (e.g., broader-leaved species of
Callistemon and Melaleuca: M. globifera, M. megacephala) may exemplify
pseudophyllodes. In the parallel-veined and flabelliform leaves cited, expan-
sion of a midrib of an acicular or linear leaf would contribute the majority
of the flattened surface, not merely the petiole (which may not be involved
at all). I suggested this interpretation earlier for broader-leaved species of
Stylidium such as S. maitlandianum, S. rigidifolium, and S. amoenum
(Carlquist, 1969), which have flabellate venation.

Another odd venation type that may be the result of expansion of
linear leaves includes leaves which form perfect or nearly perfect dichoto-
mous branching systems, each branch a linear segment of the leaf. The
genera Stirlingia (fig. 8.17, bottom right) and especially Franklandia of the
Proteaceae exhibit such dichotomous leaves in a striking fashion. This
tendency, which seems virtually endemic to southwestern Australia, is also
represented in Corynotheca micrantha and Thysanotus dichotomus of the
Liliaceae, as well as in such Proteaceae as Isopogon divergens, I. laticornis,
I. yillosus, Petrophila conifera, P. divaricata, P. semifurcata, and Adenan-
thos cunninghamii. A broader scope is introduced by consideration of a
family widespread in Australia, Epacridaceae. This family has an amaz-
ingly stereotyped parallel venation, and seems derived from, or very close
to, Ericaceae. Ericaceae have unilacunar nodes, and it is believed (I. W.
Bailey, personal communication, 1955) that Epacridaceae represent expan-
sion from unilacunar ancestry to a multilacunar condition. The broader

286

Fic. 8.15. Leaves of other species of Hakea (Proteaceae) from southwestern Australia: because
of their three-dimensional shapes, these species are shown photographically and supplement
(continued)

287

Adaptive Radiation: Flora of Southwestern Australia

leaf form of Epacridaceae would represent an external manifestation of
this expansion. At any rate, the question of pseudophyllodes and other
possibly expanded leaf types in the floras of southwestern Australia and
other parts of Australia is one that needs examination by developmental
and comparative techniques.

There is an additional category of probable surface expansion. Acacia
urophylla, an understory element of the wet karri forests in southwestern-
most Australia, appears to have unusually broad phyllodes, probably de-
rived from an ancestor with narrower phyllodes, and the nature of phyllode
venation in this species suggests exceptional expansion of the phyllode as
an accompaniment to ecological shift into unusually wet (for Western
Australia) forests. The examples given show the factors of reversibility and
independent origins of similar-appearing leaf types. More importantly,
they illustrate the great plasticity of angiosperms (a plasticity which
appears, in many cases, to exceed that of the conceptualizations of bota-
nists who study them).

Smaller Leaves. ‘That narrower leaves are correlated with xeromorphy
is almost axiomatic. The instances of leaf restriction in the Western Aus-
tralian flora are simply too numerous to cite adequately. Some of these,
at least, appear to be secondary restrictions phylogenetically of once-
broader leaf surfaces. One such phenomenon is inrolling of margins to the
abaxial surface (ericoid leaves). Tetratheca species, particularly T. conferti-
folia (fig. 8.16, top right) as compared to T. setigera (fig. 8.16, top left),
show this. In Banksia, B. occidentalis, B. sphaerocarpa (fig. 8.13) and other
species show these tendencies clearly. Many examples of ericoid leaves
from such families as Lamiaceae and Verbenaceae can be cited in the
southwest. Southwestern Rutaceae (Phebalium, Eriostemon, Philotheca,
Boronia ericifolia and other species of Boronia) are excellent examples,
as are all the species of Frankenia (Frankeniaceae) and a number of species
of Hibbertia (Dilleniaceae). Casual perusal of the southwestern flora could
add many species to the list, particularly from such families as Myrtaceae.

FIG. 8.15 (continued)

those of fig. 8.14. rop Lert, H. conchifolia, fruit can be seen within tapered, hoodlike leaf;
leaf about 6 cm long; near Eneabba. tor ricut, H. cucuilata, cuplike leaves about 8 cm in
diam., capitate infloresceuce within each; between Porongorups and Stirlings. BoTTOM LEFT,
H. victoriae, prickly leaves, about 12 cm long, which conceal axillary flowers; East Mt.
Barren. BOTTOM RIGHT, H. clayata, tip of twig showing fruit and leaves, which are 2 cm wide
and so succulent they are nearly terete; granite headlands of Cape Le Grand.

288

Adaptive Radiation: Flora of Southwestern Australia

That acicular leaves have evolved in numerous groups in Western
Australia seems very likely. This form is the ultimate in condensation, and
ericoid leaves may be an intermediate step. The genus Platytheca (Tre-
mandraceae) can hardly have any other explanation, considering the re-
mainder of this family, its characteristics and distribution. Acicular leaves
are common in Hakea (see fig. 8.14) Calothamnus (figs. 8.19, top left,
night; 8.33, bottom right), Petrophila (fig. 8.18, bottom nght), Darwinia
(figs. 8.22, top nght, bottom right; 8.23), Acacia, and many other genera.
The hypothesis proposed earlier, that pseudophyllodes have resulted from
expansion of acicular leaves, in no way contradicts evolution of acicular-
leaved plants from broader-leaved ancestors. One would indeed expect the
latter tendency to occur more often, in view of the broad dry areas of
Western Australia, and conversion of narrower leaves into broader forms
to be a relatively infrequent phenomenon characteristic of entry into the
limited wet-forest and moist coastal areas of southwestern Australia.

Another way in which leaf surface is reduced is so obvious as to be
perhaps overlooked: mere microphylly, without special change in form.
Comesperma drummondi (fig. 8.16, bottom left, right), with small glau-
cous leaves appressed to stems, is an example. However, there are more
numerous instances in such families as Myrtaceae, where Lhotzkya,
Wehlia, Micromyrtus, Thryptomene, Baeckia, Sholtzia, and Balaustion
serve as definitive examples.

Succulence. By contrast, succulence is a rather rare phenomenon in
Western Australia, rarer than one would expect. It is not unusual in halo-
phytes (Calandrinia spp., Carprobrotus spp., Arthrocnema spp., Scaevola
spinescens), but few nonhalophytes have evolved succulence. One notable
exception is Hakea clavata (fig. 8.15, bottom right), in which large, fibrous,
obviously succulent leaves that dry with extreme slowness have evolved
autochthonously in southwestern Australia. Although a few other examples
of succulence could be cited, the interesting phenomenon is that succu-
lence, where present in nonhalophytes, often occurs in underground struc-
tures, such as bulbs, tubers, or corms. These are discussed later under
mechanisms that tend to represent resistance to fire as well as to desicca-
tion. Minimal succulence in the southwestern flora may also be related to
herbivore pressure, for succulent plants would tend to be vulnerable to

289

FIG. 8.16. Plants illustrating evolution of particular leaf types in southwestern Australia. TOP
LEFT, Tetratheca setigera (Tremandraceae), leaves sparse, relatively narrow, about 2 cm long;
near Pinjarra. TOP RIGHT, Tetratheca confertifolia, leaves narrow, “‘ericoid,” ca. 12 mm long;
near Mogumber. BOTTOM LEFT, RIGHT, Comesperma drummondii (Polygalaceae). BOTTOM

LEFT, habit of plant, showing sparse branching; Ghooli. BorromM ricHT, branch showing leaves,
which are glaucous, about 15 mm long.

290

Adaptive Radiation: Flora of Southwestern Australia

herbivores—which presumably might avoid the excessively salty leaves of
halophytes.

THREE-DIMENSIONALIZATION

There seems no existing term to cover a number of angiosperm plant
forms, several of which are unique to southwestern Australia. Only field
experience can make the concept of “three-dimensionalization” real, for
herbarium specimens and even photographs do not unequivocally illustrate
the tendency of so many southwestern plants to arrange typically flat or
terete structures—especially (but by no means exclusively) spiny, thorny, or
prickly portions—in such a way as to deter intrusion into the center of a
plant. Some of these phenomena are obvious, some subtle. They seem, —
however, to be best interpreted as herbivore-deterrents. We are all familiar
with the way in which cactus spines radiate from glochids, so that an herbi-
vore would suffer puncture no matter what the angle of approach. The
same—plus other qualities—is true of three-dimensionalization of south-
western plants.

Radiating Armament. A simple type of three-dimensionalization is
created by leaves with margins the spiny teeth of which point upward and
downward with reference to the plane of the lamina. Leaves with such
spiny teeth could be termed “hollylike” except that they exceed spiny-
leaved holly forms in the degree to which teeth are divergent in their posi-
tioning. Even a single leaf may radiate spiny teeth at various angles. This
is illustrated by Synaphaea decorticans (fig. 8.17, bottom left) of the Pro-
teaceae. Most other species of Synaphaea have this habit, as do many
other Proteaceae: Hakea amplexicaulis, H. undulata, H. auriculata, and,
to a lesser degree, H. baxteri and H. flabellifolia (fig. 8.14) are examples.
In an obvious although different way, H. victoriae (fig. 8.15, bottom left)
and H. conchifolia (fig. 8.15, top left) show this phenomenon. The great
toughness—almost like thin sheet metal—of proteaceous leaves tends to
reinforce the effectiveness of the spines or prickles, giving them a stronger
positioning than a flexible leaf would allow. Thus, the tough leaves of
Banksia (fig. 8.13) are effective out of proportion to their seemingly mod-
erate teeth. Other Proteaceae in which toothed leaf margins form excellent
armor include Dryandra horrida (and, for that matter, virtually the entire
genus Dryandra), Grevillea acrobotrys, G. amplexicans, G. apiciloba, G.

291

FIG. 8.17. Leaves of Proteaceae to show tendencies in “‘three-dimensionalization” in the flora
of southwestern Australia. rop LEFT, Grevillea patentiloba, leaves about 3 cm long, terete seg-
ments spreading divaricately. rop r1GHT, Hakea varia, twig showing flattened leaves, about 2.5
cm long, which are prickly and project at various angles from twigs. BOTTOM LEFT, Synaphaea
decorticans, prickly-lobed leaves project at various angles. BOTTOM RIGHT, Stirlingia tenuifolia,
leaf (shown sideways) with dichotomous branching of the terete segments; leaf about 7 cm long.

292

Adaptive Radiation: Flora of Southwestern Australia

astericosa, G. bipinnatifida, G. cirsiifolia, G. flexuosa, G. glabrata, G. ma-
crostylis, G. muelleri, G. phanerophlebia, G. platypoda, G. pulchella, G.
rudis, G. synapheae, G. trifida, G. triloba, Hakea ceratophylla, H. cristata,
H. dolichostyla, H. glabella, H. pritzeli, Isopogon alcicornis, I. baxter, I.
roseus, I. tridens, Lambertia echinata, Petrophila biternata, P. diversifolia,
P. macrostachya, P. trifida, and Xylomelum occidentale. Examples from
other families are too numerous to cite; Hovea chorizemifolia, and various
species of Daviesia and Chorizema are among many such Fabaceae.

A heightened form of this armor is achieved when leaf segments are
linear and project at various angles. Examples shown here include Grevil-
lea patentiloba (fig. §.17, top left) and Hakea varia (fig. 8.17, top right).
Other species of these genera with remarkably divaricating, linear, spine-
tipped leaf segments include Adenanthos meissneri, A. pungens, Grevillea
annulifera, G. armigera, G. asparagoides, G. brevicuspis, G. intricata, G.
leptobotrya, G. paniculata, G. paradoxa, G. stenostachya, G. tenuiloba, G.
teretifolia, G. thelemanniana, G. tridentata, G. tripartita, G. wilsoni, Hakea
erinacea, H. purpurea, Isopogon divergens, I. formosus, I. teretifolia, Petro-
phila drummondii, and Synaphaea petiolaris, all of Proteaceae. Represen-
tatives of other families include Gompholobium baxteri and Jacksonia fur-
cellata (Fabaceae).

The simplest and most obvious form of radiating armament is display
of linear, terete, spine-tipped leaves in such a way that they, like cactus
spines, radiate in all directions from a stem. Examples are too numerous
to mention fully, but include many species of Hakea, such as H. lehman-
manda, H. brookeana, H. platysperma (fig. 8.14), and Grevillea, Acacia,
Daviesia (figs. 8.1, bottom left, right; 8.2, top left, right), Melaleuca, Caly-
thrix, Calothamnus, Bassia, and Stylidium.

Spiralization. A remarkable feature, rarely encountered in other floras
except as an anomaly, but of almost common occurrence in southwestern
Australia, is the presence of spirally (helically) twisted leaves. Hakea
stenoptera (fig. 8.18, top left) has decidedly curved lanceolate leaves.
Persoomia tortifolia (fig. 8.18, top right) has corkscrewlike lanceolate leaves.
And Petrophila linearis (fig. 8.18, bottom left) has curved lanceolate
leaves. In P. stricta (fig. 8.18, bottom right), leaves are terete, curved, and
spine-tipped. Another protead with this leaf habit is Conospermum incur-
yum. Similar leaves occur in Acacia viscifolia, A. triptycha var. subflexuosa,

293

Adaptive Radiation: Flora of Southwestern Australia

and A. merinthifolia. Helical leaves have been mentioned in the orchid
Thelymitra spiralis. In Epacridaceae, curved and often markedly helical
leaves occur in Sphenotoma gracile and many of the 22 currently recog-
nized species of Andersonia, such as A. grandiflora, A. involucrata, and A.
echinocephala.

Spiralized stems in broom shrubs have a similar net effect, and in this
category one can cite Dampiera curvula, Scaevola tortuosa, and Tetratheca
efoliata (fig. 8.7, bottom right). Among monocotyledons, one notes this
habit in several restiads: Restio megalotheca, Loxocarya flexuosa, and
Caloraphus gracillima, for example.

The tendency toward dichotomously branched leaves, mentioned
earlier, is clearly a form of three-dimensionalization. Stirlingia and Frank-
landia are so marked in this respect that individual leaves resemble aerial
axes of Psilotum (especially in F. fucosa).

The explanation for the various forms of three-dimensionalization is
not clear. However, the most likely explanation consistent with tendencies
in the remainder of the flora is that this phenomenon, like the tangled
branches of thorn shrubs, forms a sort of deterrent to herbivores. The
number of instances, in each of the categories, is too large to be explained
by chance, and the phenomenon offers an interesting opportunity for
further evolutionary speculation and investigation.

SHORT SHOOTS AND LONG SHOOTS

Presence of both long shoots and short shoots on a plant, known as
heteroblasty, is not rare in dicotyledons at large, but it has been evolved
autochthonously in several genera of the southwestern flora. It is strikingly
represented by Calothamnus pinifolius (fig. 8.19, top left, nght). The long
shoots in this species are few and branched, but almost every axil on the
long shoots bears a short shoot. Heteroblasty is strongly marked in Plagi-
anthus helmsii (fig. 8.19, bottom left, right). This cactoid-appearing malva-
ceous plant has been segregated as a separate genus, Selenothamnus. Con-
sidering the wide divergences in habit within various Australian genera,
P. helmsii is not completely incongruous when compared with other spe-
cies of Plagianthus, although it is distinctive. In P. helmsii, long shoots are
few and are densely covered with short shoots of almost determinate
length. The explanation for heteroblasty in these plants probably does not

294

FIG. 8.18. Leaves of Proteaceae, showing “‘spiralization” tendencies, from southwestern Aus-
tralia. trop Lert, Hakea stenoptera, leaves 7 cm long, curved into one or more spirals. TOP
RIGHT, Persoonia tortifolia, leaves ca. 5 cm long, twisted into numerous gyres. BOTTOM LEFT,
Petrophila linearis, showing flattened leaves curved into about one spiral each. BOTTOM RIGHT,
Petrophila stricta, spiral-tending leaves, terete and tipped by a spine.

2o5

’

FIG. 8.19. Plants of Western Australia with marked dimorphism into long and short shoots.
TOP LEFT, RIGHT, Calothamnus pinifolius (Myrtaceae). Top LEFT, habit of plant, 2 m tall;
south of Ravensthorpe. Trop ricHT, branch showing short shoots bearing needlelike leaves
(lower portion; flowers above), leaves about 4 cm long. BOTTOM LEFT, RIGHT, Plagianthus
(“Selenothamnus”’) helmsii (Malvaceae). BOTTOM LEFT, habit of plant—this plant 0.5 m tall,
mature plants reach 3 m. BoTToM RIGHT, surfaces of two branches, showing flowers (one
branch male, one female) at tips of short shoots that cover these branches.

296

Adaptive Radiation: Flora of Southwestern Australia

have to do with herbivore deterrence as much as with climatic regime, since
similar heteroblasty occurs in plants of southern California (e.g., Artemisia
californica, Eriogonum fasciculatum, Adenostoma fasciculatum), in which
growth of short shoots can take place when moderate rainfalls permit, but
long shoots are formed after winter rains, with increased temperature. Con-
densation of form in plants with heteroblasty, such as P. helmsii, may bear
some relationship to xeromorphy.

WOOLLY VESTURE

Conspicuous in many southwestern genera—including some where
one would not expect it—is an exceptional degree of covering by dense,
and usually white, nonglandular trichomes. Notably, not merely stems
and leaves, but also inflorescences and flowers can be strongly hairy. Ptilo-
tus (Trichinium) rotundifolius (fig. 8.20, top left) of the Amaranthaceae
has wool on the inner side of the perianth segments, as do other species of
this genus; leaves are closely and densely woolly. Likewise, Dampiera lutet-
flora (fig. 8.20, top right) of the Goodeniaceae has leaves and stems cov-
ered with a very dense tomentum, and even calyces are densely clad. Other
species of Dampiera are also notably hairy, some with slate gray wool on
leaves, corolla, and calyx. Other Goodeniaceae of the southwest, such as
Verreauxia reinwardtii, share these characteristics. Verbenaceae as a whole
are often densely hairy in southwestern Australia. Examples shown here
include Lachnostachys ferrugineus (fig. 8.20, bottom left), L. albicans (fig.
8.24, top nght), and Pityrodia oldfieldii (fig. 8.20, bottom right). Calyces
as well as leaves and stems commonly are densely woolly in these genera.
In Hemiphora elderi, a monotypic genus of the southwestern Verbenaceae,
herbage is white-woolly, but calyces and corollas are densely red-woolly.
Other genera in which hairiness of southwestern species seems exceptional
include Thomasia and Lasiopetalum of the Sterculiaceae (see Lasiopetalum
drummondi, fig. $8.24, top right), Hemigenia (Lamiaceae), Kochia (Ama-
ranthaceae), Bonamia (Solanaceae), Rulingia (Malvaceae), Actinotus and
Xanthosia (Apiaceae), Conospermum (Proteaceae), and Diplolaena (fig.
8.21, top left) of the Rutaceae. The hairiness of vegetative portions of all
of these seems exceptional, but one notes much more hairiness on floral
portions than one would expect. Flowers of Anigozanthos, Macropidia,
Conostylis, and Blancoa (Liliaceae sensu lato, sometimes Haemodoraceae)

297

=

FIG. 8.20. Tendencies toward exceptionally woolly vesture in plants of southwestern Australia.
TOP LEFT, Ptilotus rotundifolius (Amaranthaceae), leaves 3.5 cm long, covered with short
dense wool. ror RIGHT, Dampiera luteiflora (Goodeniaceae), leaves and calyces woolly; Ghoolli.
BOTTOM LEFT, Lachnostachys ferrugineus (Verbenaceae), inflorescence and adjacent extremely
woolly, cuplike leaves 3 cm in diam.; north of Moora. sorrom ricut, Pityrodia oldfieldti
(Verbenaceae), portion of stem showing hairy leaves, stem surface, and calyx; north of Moora.

298

Adaptive Radiation: Flora of Southwestern Australia

are quite noteworthy in this respect, and must represent autochthonous
developments in southwestern Australia, because three of these genera are
endemic to the province.

Several possible explanations can be suggested for the exceptional
woolliness of the taxa cited. There may be ecological correlations. Many
of these genera and species tend to grow in drier, sunnier localities, in
which case desiccation resistance or shading from insolation might be
operative. Such genera as Ptilotus (Trichinium), Kochia, and Lachnosta-
chys are more markedly inland in distribution than most larger angiosperm
genera in the southwestern flora. Such inland locations, moreover, are
more frequently subject to frost than are coastal locations. The inland
area of southwestern Australia in which frost exceeds 5 days per year on
the average includes the ranges of many of the taxa cited above. This area
extends roughly from Wongan Hills south to Southern Cross and east to
Coolgardie.

Excessive hairiness may prove a deterrent to herbivores. In some cases,
such as Lachnostachys, purplish colors of corollas stand out clearly against
white woolly vesture and may be related to pollination. In Hemiphora,
contrast between red wool of flowers and white wool of herbage is marked,
although this circumstance does not explain why the flowers would be
hairy. This is also true in Anigozanthos, Macropidia, Blancoa, and Cono-
stylis, where the distinctive colors of portions of these flowers are contained
in the dense hairs that cover the perianths. Obviously, there is no single
or simple explanation of woolliness in angiosperms of southwestern Aus-
tralia, and experimental evidence on possibilities such as the above would
be very desirable.

OILS, OLEORESINS, AND OTHER CHEMICAL CONTENTS

Many plant groups well represented in the southwest contain oil-rich
secretory cavities (e.g., Rutaceae and Myrtaceae) or have external cover-
ings of resinlike or varnishlike materials; Anthocercis (Solanaceae), for
example. In Anthocercis, restriction of transpiration is to be suspected
(from the degree of difficulty experienced in preparing dried specimens).
Internal oil deposits and strong scents from essential oils contained in
glandular hairs (Lamiaceae, Verbenaceae) may deter herbivores. As yet

zo

Adaptive Radiation: Flora of Southwestern Australia

we know nothing certain about these possibilities, but the abundance of
these characteristics in the flora suggests adaptive significance.

Pollination Mechanisms

Every flora has pollination mechanisms of interest, but southwestern Aus-
tralia has several exceptional modalities. Among these are the highly color-
ful nature of the flora and the large number of plants apparently adapted
to ornithophily. Although one could simply say that the southwestern
flora is exceptional in its colorfulnes, particular examples are illustrative.
For example, Amaranthaceae are often colorless in perianth parts, but
Ptilotus (Trichinium) often has flowers with large magenta perianths, as
in P. rotundifolius (fig. 8.20, top left). Blue flowers are exceptionally fre-
quent, as in Dampiera and Leschenaultia. Adaptation to particular color-
sensitive pollinators is to be suspected. This certainly is true in the orni-
thophilous groups. Obviously a large number of genera have entered the
ornithophilous mode, including Anigozanthos and the closely related Ma-
cropidia, the colorful patterns of which can only be appreciated in the field
or with good color photographs. In addition to heightened color and
lengthened or enlarged flower form, there are other obvious ornithophilous
devices. Calothamnus (figs. 8.19, top night; 8.33, bottom right) has fascicles
of bright red stamens arranged in a curved pattern. Increased prominence
of stamens related to ornithophily can be suspected in the gigantic flowers
on Eucalyptus macrocarpa. | have seen both Calothamnus and E.. macro-
carpa visited by meliphagids. The congestion of flowers into colorful heads
in proteads such as Banksia, Dryandra, and some species of Grevillea re-
lates to ornithophily, and I have also observed honeyeaters visiting these
genera.

Congestion of flowers, coupled with development of bracts around
the inflorescence (the bracts thus taking over the display function), is a
feature autochthonously developed in a number of southwestern groups.
Diplolaena (fig. 8.21, top left) of the Rutaceae exemplifies this. Most
species of Pimelea have congestion of flowers but with green, inconspicuous
bracts, as in P. floribunda (fig. 8.21, top right). However, P. physodes (fig.
8.21, bottom left, right) has long bracts that are red purple outside, cream

300

“ee

Fic. 8.21. Tendencies toward capitate inflorescences and showy bracts in southwestern Aus-
tralian plants. rop LEFT, Diplolaena dampieri (Rutaceae), one head in flower, one in fruit;
heads about 6 cm in diam.; Canal Rocks. rop ricut, Pimelea floribunda (Thymeleaceae), head
of flowers, bracts greenish, conspicuous; Moora—Jutien Bay road. BOTTOM LEFT, RIGHT,
P. physodes, head nutant, surrounded by long purplish bracts; near Ongerup. BOTTOM LEFT,
heads with bracts in natural position; bracts 4-5 cm long. Borrom ricut, head in which bracts
have been folded back in order to show old flowers.

301

Adaptive Radiation: Flora of Southwestern Australia

white within, an obvious adaptation to ornithophily. Other instances of
this phenomenon are the three species of Johnsonia (Liliaceae) in which
the papery bracts are pink or reddish and are longer than the flowers,
which are white and inconspicuous. In Andersonia echinocephala (Epa-
cridaceae), bracts surrounding the congested inflorescence are white or
even pinkish. This is true also in Grevillea bracteosa (Proteaceae).

The most outstanding example of conversion of inflorescence bracts
to an attractive device suitable for bird pollination is, however, Darwinia
(Myrtaceae). The Stirling Range species of Darwinia (fig. 8.23), for exam-
ple, all show this habit. If we examine a series of Darwinia species, we
can see the development within a single genus. Darwinia diosmoides (fig.
8.22, top left) has few-flowered, capitate inflorescences, and flowers appear
scattered over the branches. The papery white petals form the attractive
display apparatus of these flowers, which are probably not ornithophilous.
Darwinia sanguinea (fig. 8.22, top right) has larger, congested inflorescences
that are borne in such a way that they sprawl on the soil surface at the
periphery of the plant. Bnght red perianth parts form the attractive
poitions of the flowers. In D. citriodora (fig. 8.22, bottom left) both bracts
and flowers are colorful organs: flowers are orange, but the head of flowers
is enclosed in spreading, open bracts that are greenish at their tips but
reddish orange at their bases. Heads of D. citriodora are borne more or
less upright, at random over a shrub. In D. leiostyla (figs. 8.22, bottom right;
8.23), the achievement of the campanulate head is complete. Heads
terminate branches and are nodding—which means, presumably, that
visiting birds will perch below the heads and face upward in search of
nectar. The bracts are bright red, forming an obviously attractive pseudo-
perianth for the head. To be sure, flowers are red to pink, but they are
not visible unless one peers into the head. The other campanulate species
and semi-campanulate species of Darwinia offer distinct bract patterns.
For example, D. helichrysoides and D. hypericifolia (fig. 8.23) have reddish
bracts. Bracts of D. macrostegia are creamy white, veined red; in D. collina,
bracts are lemon yellow, whereas in D. fimbriata (D. squarrosa), bracts are
deep pink. Darwinia meeboldii has white bracts, greenish at the base,
tipped crimson.

Among the Stirling species of Darwinia, there is a habitat difference
between the valley species, D. leiostyla, and the remainder, which grow

302

FIG. 8.22. Tendencies in inflorescence formation in Darwinia (Myrtaceae). Top LEFT, D.
diosmoides, flowers in heads, scattered over branches (from a color photograph by Fred W.
Humphreys, courtesy of Mrs. Evelyn Hymphreys). rop ricut, D. sanguinea, heads of flowers
bozne on branches that lie on ground; Cockleshell Gully. sorrom wert, D. citriodora, heads
of orange flowers surrounded by bracts orange at bases; Darlington. sorrom ricHT, D.
lziostyla, heads of flowers nutant, surrounded by scarlet bracts to 6 cm long; Red Gum Springs.

303

Adaptive Radiation: Flora of Southwestern Australia

MT. LESUEUR D. helichrysoides

(north of Perth)

D. hypericifolia

HUME PEAK

io

PEAK

DONELLY / 7 PEAK BARNETT _
é ~ SINS YL
HESS Seek St EN eNO, \ hice
y Ss / 4 eye “a 4 TOOLBRUNUP
Se ees nae NC THE ABBEY <= Sy,
ROSS PEAK ye <a SS J,
—— A NG e SX y,
a va MONDURUP : PEAK HENTON Tt, mags J
/ = ; == “ff es,
Qesiem: / ee TALYUBERLUP mas
ee \ Gigs Se
me SS =
— ————— N
* STIRLING

FIG. 8.23. Narrow endemism by species of Darwinia (Myrtaceae) on the Stirling Range of
southwestern Australia. Each of these species, which have notably bracteate ornithophilous
inflorescences, is endemic to the few hills and peaks indicated, and has not been found in any
other areas. Darwinia leiostyla grows not on peak tops, but in valley areas between peaks. A

at the summits of the peaks. Distinctive size, color, pattern, shape, and
texture of the heads may provide isolating mechanisms among the Stirling
Darwinias. Peak-to-peak isolation may account for their narrow endemism,
although the distances involved are so small that the situation of the
Stirling Darwinias is notable and difficult to explain with certainty. Non-
Stirling Darwinias, such as D. helichrysoides, may also represent such

304

Adaptive Radiation: Flora of Southwestern Australia

D. leiostyla

(in valleys among peaks)

D. squarrosa D. collina

eae’ We WARRUNGUP \
4 HILL

“X

MT. oe
HASSELL oe
ooMiE.
TRIO 4
\ > a . ae
\ Zl ace = - pe :
Sa YUNGERMERE a a ae pS eee
as
RANGE

non-Stirling species, endemic to Mt. Lesueur, is also shown. The Stirling Range is shown as
seen from the south, hills and peaks drawn at a single scale (the highest, Bluff Knoll, reaches
3640 ft).

isolating factors. The Stirling assemblage may well be not a single
autochthonous radiation but more than one invasion of these shale peaks.

The evolution of congested inflorescences surrounded by bracts might
represent another ornithophilous adaptation: landing platforms. The stout
leaves of Hakea cucullata (fig. 8.24, bottom left), each surrounding a group
of flowers, may well accommodate claws of visiting birds. This certainly

305

Adaptive Radiation: Flora of Southwestern Australia

MT. LESUEUR D. helichrysoides

(north of Perth)

D. macrostegia D. hypericifolia

a

i
/ j »
sone J ~ PEAK BARNETT I~
LO Sie >
Zo y aw ‘ He \ Gb AS ; TOOLBRUNUP
vi SOs we VY ff NN THE ABBEY ff” /
a4 ROSS PEAK Si « J SS =
ee y MONDURUP \— PEAK HENTON ot wasos y
ae J ZirS ;
fe 3d cl Se Ne oS
9° lem Bo Y= ——~ TALYUBERLUP__>“*
__— << STIRLING

——
Fic. 8.23. Narrow endemism by species of Darwinia (Myrtaceae) on the Stirling Range of
southwestern Australia. Each of these species, which have notably bracteate ornithophilous
inflorescences, is endemic to the few hills and peaks indicated, and has not been found in any
other areas. Darwinia leiostyla grows not on peak tops, but in valley areas between peaks. A

at the summits of the peaks. Distinctive size, color, pattern, shape, and
texture of the heads may provide isolating mechanisms among the Stirling
Darwinias. Peak-to-peak isolation may account for their narrow endemism,
although the distances involved are so small that the situation of the
Stirling Darwinias is notable and difficult to explain with certainty. Non-
Stirling Darwinias, such as D. helichrysoides, may also represent such

304

Adaptive R
aptive Radiation: Flora of Southwestern Australia

D. leiostyla

(in valleys among peaks)

squarrosa

D. collina

\ Vpa~

y \ Au \
fi \VE \ A AN
)
f \ ELLEN

E is X
3 BLUFF KNOLL \ ISONGERUP
PYUNGOORUP y \ \ PEAK

\

A

vi WARRUNGUP \

} J
HILL 7 ‘
x MT. SUCCESS x s
Se = — = <
nee om z :
\ HASSELL \_ wo fe = COYANARUP / aN
\ & \ # x
\ TRIO 27. r \ ‘
as
~ = SNe
a YUNGERMERE ee

RANGE
non-Stirling species, endemic to Mt. Lesueur, is also shown. The Stirling Range is shown as

seen from tl ills ; VER An
3640 ye he south, hills and peaks drawn at a single scale (the highest, Blut Knoll, reaches

isolating factors. The Stirling assemblage may well be not a single
autochthonous radiation but more than one invasion of these shale peaks,
The evolution of congested inflorescences surrounded by bracts might
a ornithophilous adaptation: landing platforms. The stout
tn of Hakea cucullata (fig. 8.24, bottom left), each surrounding a group
Owers, may well accommodate claws of visiting birds. This certainly

305

Fic. 8.24. Unusual flower types in the flora of southwestern Australia, showing probable evolu-
tion of new pollination mechanisms. top LEFT, Lasiopetalum drummondi (Sterculiaceae),
flower parts excessively woolly with stellate hairs; Three Springs. Top ricHr, Lachnostachys
albicans (Verbenaceae), inflorescence congested and hairy; leaves densely felty; leaves 4 cm

(continued)

306

Adaptive Radiation: Flora of Southwestern Australia

would be true for H. victoriae (fig. 8.15, bottom left); every head of flowers
is axillary to a large leaf, which—perhaps serving as the attractive organ as
well as a potential landing platform—is whitish to orangeish at the base
and could serve as a “guide” to bird visitors. The bracts of Dryandra of
the Proteaceae (see D. runcinata, fig. 8.25, bottom right) are rarely colorful,
usually only brown, but may contrast with flowers (white, yellow, orange)
and possibly aid in perching of visiting birds.

Inflorescence bracts that lack attractive functions also occur. Diels
(1906) has called attention to taxa in southwestern Australia in which
bracts are papery. These include Acacia squamata, A. restiacea, Cono-
stephium pendulum, and Daviesia hakeoides. ‘Their bracts are sufficiently
inconspicuous that one cannot cite a function in pollination; although to
cite bracts as a protective device is facile, one cannot easily find other
reasons for their existence in these taxa.

Conversion to ornithophily within particular genera can be cited.
In Verticordia (Myrtaceae), flowers tend to be pastel-colored, broadly
cuplike, and small—all features suggesting bee flowers. In V. insignis and
VY. mitchelliana of the southwest, petals are large, tend to form a tube,
and are bright red. These species, moreover, tend to have branching suited
for perching, whereas other Verticordias do not.

How are we to interpret the rather numerous instances in which
clearly ornithophilous flowers are borne at ground level? Instances occur
in Brachysema aphyllum (fig. 8.3, bottom right), B. daviesioides (fig. 8.5,
top left, ight), B. tomentosum (fig. 8.5, bottom left, nght), B. chambersii,
Banksia prostrata (fig. 8.6, bottom right), and perhaps Banksia nutans
(fg. 8.6, bottom left); Darwinia sanguinea (fig. 8.22, top right) and D.
speciosa; and several species of Grevillea with what could be called “ground-
level flagelliflory.” Grevillea pectinata (fig. 8.25, top left, right, bottom left)
is a good example of these Grevillea species. Branches from a shrub tend
to sprawl; they terminate in inflorescences borne on the ground at distances
of up to 2 m from the plant base. Flowers in these plants are bright red
and congested, and, as with the others cited above, they are evidently
Fic. 8.24 (continued)
long. Bottom LEFT, Hakea cucullata (Proteaceae), close view of cuplike leaves enclosing
groups of flowers; Porongorups, Stirlings. sorrom ricut, Rhizanthella gardneri (Orchidaceae),

photograph of pickled specimen (see text for explanation of underground-flowering habit).
(From a color photograph by Alex George.)

307

FIG. 8.25. Tendencies to flowering at ground level in Proteaceae of southwestern Australia.
TOP LEFT, RIGHT, BOTTOM LEFT, Grevillea pectinata. TOP LEFT, habit; leafy stems at center,
flowering stems inconspicuous, trail up to 2 m away from plant; near Perenjori. TOP RIGHT,
leaves; flowering stems bear reduced leaves, vegetative leaves like this only in center of plant.
BOTTOM LEFT, inflorescence at tip of trailing shoot. BoTTOM RIGHT, Dryandra runcinata,

flowers surrounded by cup of brown bracts, inflorescence is about 8 cm high; near Badgin-
garra.

308

Adaptive Radiation: Flora of Southwestern Australia

ornithophilous (although actual observations seem lacking). Apparently
the ground can serve as a perch for feeding from large bright flowers borne
at that level. Although ground-level ornithophily seems fantastic compared
to ornithophily in other regions (except South Africa), there is no reason
to believe it is not operative.

Pollination relationships such as the preceding, which have probably
evolved autochthonously in southwestern Australia in most cases, are quite
unusual and are worth extensive study. Erickson’s (1958) comments on
pollination in Stylidiaceae and my own theories on evolution of pollination
types in this family (Carlquist, 1969) have endeavored to show, on the
basis of fieldwork, the almost unbelievably intricate patterns of evolution
in these respects in Stylidiaceae. Many other groups would repay study.
What is the function of the markedly fringed petals in Thysanotus (fig.
8.11, bottom rght) and Verticordia? Do flowers with marked, hairy
perianths, such as Lasiopetalum drummondii (fig. 8.23, top left), have
distinctive pollination schemes? Few efforts have been made to study
pollination in plants of the southwest. One can cite, however, the occur-
rence of pseudocopulation in the orchid Cryptostylis ovata (Morcombe,
1968). Pollination of Banksia by the marsupial Tarsipes spenseri (Mor-
combe, 1968) is unusual, but is probably a secondary phenomenon in the
primarily ornithophilous Banksia. I will cite only one more example, one
of the truly bizarre plants of southwestern Australia, Rhizanthella gardneri
(fig. 8.24, bottom right). This “underground orchid” apparently completes
its entire life cycle below the soil surface. If it emerges (and that has not
been observed in the three collections now made of this elusive plant), it
must do so briefly at best. I suspect pollination by burrowing beetles, or
self-pollination. Likewise, the eastern Australian orchid Cryptanthemis
with similar habits (but with emergence barely above ground surface) has
not been observed with respect to pollination mechanism. These are some
of the challenges to the botanist interested in evolution in southwestern
Australia.

Evolution of Dispersal Types

Likewise, many unusual seed and fruit-dispersal types occur in south-
western Australia, and little is known about them. I am presenting a few

309

Adaptive Radiation: Flora of Southwestern Australia

examples only to show interesting evolutionary tendencies and to suggest
the opportunities for further work. The flora of southwestern Australia
can be explained easily enough as originating from long-distance dispersal,
as well as shorter transport events across dry land. Prevailing winds during
summer—when seeds are likely to be dispersed—are from east to west, so
that transport of airborne seeds in that direction would be expected. ‘Those
who like to categorize seed-dispersal mechanisms would find analysis of
the southwestern flora highly interesting. Whatever the original mode of
transport to Western Australia, evolution has subsequently occurred in
dispersal mechanisms of many groups.

One family that shows a diversity of dispersal mechanisms is Goo-
deniaceae. This family, centered in Australia, has undoubtedly evolved the
various dispersal mechanisms described here within the Australian conti-
nent. Some of them have evolved autochthonously in southwestern
Australia. Anthotium (fig. 8.26, top left) is endemic to the southwest. The
calyx is adherent to the ovary. The capsule is quadrivalvate, and release
of minute seeds that could be dispersed by any of various methods is to be
expected. The sharp-tipped calyx lobes might be devices for lodging the
fruit in fur, but this is very doubtful. On the other hand, dispersal of fruits
with inferior ovaries and with numerous stout hairs is to be expected in
Verreauxia paniculata (fig. 8.26, top nght); this genus is endemic to
Western Australia and the central desert. The fruits are one-seeded.

Pentaptilon (fig. 8.26, center left) is a monotypic genus of Western
Australia. Pentaptilon fruits are the product of an inferior ovary, are
one-seeded and indehiscent. The five wings (some of which remain
underdeveloped) on Pentaptilon fruits, together with light weight of the
fruits, make them excellent for wind dispersal, although presumably only
in a tumbling fashion across relatively short distances. Pentaptilon careyt
is a highly localized species. Wind dispersal would be expected for
Calogyne, Velleia (fig. 8.26, bottom right), Symphyobasis (fig. 8.26, center
right), and Goodenia (fig. 8.26, bottom left). These have capsular fruits
(formed from superior or semisuperior ovaries) containing flat seeds of
such a size that wind dispersal is conceivable. In Velleia, Calogyne, and
some species of Goodenia, seeds are not only flat but winged, enhancing
dispersibility. An additional feature, stressed by Carolin (1966), is the
presence of mucilage cells in seed coats of these genera, with some excep-

310

Fic. 8.26. Fruits (and for some, seeds) of Goodeniaceae of southwestern Australia to show
radiation in fruit type. rop LEFT, Anthotium rubriflorum. Top ricut, Verreauxia paniculata.
CENTER LEFT, Pentaptilon careyi. CENTER RIGHT, Symphyobasis macroplectra. BOTTOM LEFT,
Goodenia incana. pottoM RIGHT, Velleia paradoxa. All at same scale (Pentaptilon fruits 10
mm in diam.).

311

Adaptive Radiation: Flora of Southwestern Australia

tions, such as Velleia montana. Presence of gelatinous cells in seed coats
would promote adherence of seeds, when wet, to fur or feathers.

Two genera close to Goodenia show an interesting divergence in fruit
and seed dispersal. Neogoodenia has a one-seeded, indehiscent fruit that
is flat, however, and mimics seeds of Goodenia (Carolin, 1966). Another
genus, once considered part of Goodenia, is Coopernookia (Carolin, 1968).
In Coopernookia, seeds are ovoid and have no flattened margins or mucilage
cells. Instead, as illustrated by Carolin (1968), they bear a strophiolate
body that might serve as a myrmecochore. If dispersal by ants occurs in
Coopernookia, this is indeed a marked departure from the Goodenia
pattern.

Catosperma, a monotypic genus of Goodeniaceae from streams, has
an indehiscent fruit derived from an inferior ovary (fig. 8.27, top left).
These fruits have no obvious dispersal mechanism. I attempted to test
them for floatability because of the habitat of this genus; those tested
sank at the end of 24 hr. The dispersal mechanism of Leschenaultia is
unclear. These fruits, derived from an inferior ovary, are linear, and
although they appear as though they might be capsular, in fact they are
indehiscent and break into loment-like segments (Carolin, 1966). Seeds of
Leschenaultia are cuboidal (fig. 8.27, top right) and their outermost layers
are derived from inner portions of the fruit wall. Although one can imagine
various types of seed dispersal, there is no obvious adaptation; mere wind
dispersal of seeds or loment-like segments, or even herbivore dispersal could
be envisioned.

Scaeyola is a very interesting genus of Goodeniaceae with respect to
fruit dispersal. Scaevola has indehiscent fruits representing one or more
ovules maturing into seeds. Some Scaevola species have relatively small
dry seeds, such as S. hookeri (fig. 8.27, center left). These do not float and
are presumably distributed by birds or (over short distances) by wind.
Some species of Scaevola have a fleshy fruit exocarp, and these are either
bird-dispersed or seawater-dispersed. The widely distributed S. taccada and
S. plumieri are distributed by oceanic drift (see chapter 2). Scaevola
calendulacea occurs on Australian (but not other Pacific) beaches, and
although fresh fruits with fleshy exocarp can float, the endocarp (fig. 8.27,
bottom right) does not float, whereas endocarps of S. taccada and S.
plumieri can float (fig. 2.6, top right). Perhaps these latter, wider-ranging

SUZ.

Fic. 8.27. Fruits (and for some, seeds) of additional Goodeniaceae species of Western Aus-
tralia to show radiation in fruit types. Scale same as in fig. 8.26. TOP LEFT, Catosperma
goodeniacearum. TOP RIGHT, Leschenaultia striatum. CENTER LEFT, Scaevola hookeri. CENTER
RIGHT, S. collaris. BorroM LEFT, S. collaris fruits float. BorromM rRicuHT, fruits of S. calen-
dulacea and S. spinescens do not float.

S16

Adaptive Radiation: Flora of Southwestern Australia

species evolved from ancestors such as S. calendulacea and by minor
changes in fruit structure evolved floatability. Most Pacific insular Scaevolas
have fruits that do not float and are instead bird-dispersed. ‘This may be
true of an Australian species (primarily Western Australian), S. spinescens,
whose fruits have a fleshy exocarp but a nonbuoyant endocarp (fig. 8.27,
bottom right).

A unique development in Western Australia is the evolution of water-
dispersed inland halophytic species of Scaevola. Scaevola collaris (fig. 8.27,
center right), which grows on the shores of Lake Austin, has relatively
large fruits with a dry, papery exocarp. This exocarp, however, contains
large air chambers, and the fruits float (fig. 8.27, bottom left). These fruits
are ribbed or somewhat winged. Because of their morphology and lightness,
a secondary dispersal mechanism of tumbling in wind gusts across stretches
of inland desert can be hypothesized. In such a way, fruits of S. collaris
could be transported from one inland salt lake to another.

Fruits of Dampiera are like those of small, dry-fruited species of
Scaevola, and are presumably dispersed in the same manner. Although
the above dispersal types (all genera of Goodeniaceae are cited) represent
a broad spectrum of fruit form and therefore dispersal devices, some of
which have evolved autochthonously in Western Australia, much more
needs to be learned about dispersal mechanisms in this family. ‘The vectors
of those Goodeniaceae that appear to be bird-dispersed or ant-dispersed
need to be discovered, and buoyancy in water is to be suspected in some
species not named above. Indeed, we know virtually nothing about dispersal
in most Australian angiosperms.

One interesting group with respect to dispersal is Darwinia of the
Myrtaceae. Because some of the Darwinias are narrowly restricted to
mountain peaks, especially those of the Stirling Range (fig. 8.23), I wanted
to discover if any show loss of dispersibility. One might expect loss of
dispersibility in narrowly restricted species. There does appear to be a
clear range in dispersibility within Darwinia (fig. 8.28). Fruits of D.
pauciflora, with prominent papery petals, are ribbed and relatively light
in comparison to volume (fig. 8.28, top left). The petals serve as the display
organ of flowers of D. pauciflora, which are borne in a noncapitate fashion,
and may serve as wings that aid in dispersal of this nonmontane species.

314

(ees 4 ~
i .

Pes meee eT =

TNT
ie ha ae

“is.

Sot
ren. a

FIG. . Tendencies in fruit morphology and potential dispersibility in species of Darwinia
Myrtaceae) from southwestern Australia. rop LEFT, D. pauciflora. Top ricHT, D. acerosa.
CENTER LEFT, D. oederoides. CENTER RIGHT, D. hypericifolia. BoTtoM LEFT, D. macrostegia.
BOTTOM RIGHT, D. meeboldii. Seeds of D. meeboldii are 10 mm long; all photographed at the

same scale.

315

Adaptive Radiation: Flora of Southwestern Australia

In contrast, fruits of D. acerosa, a capitate species, are small and have
small petals (fig. 8.25, top right). This species is not widely distributed
(Mogumber, possibly also Mt. Gibson), perhaps because of edaphic pref-
erences, but its fruits are typical of those of more widely distributed
species: of Darwinia. Small size of seeds in such species ought to aid
dispersal. Darwinia oederoides is a subshrub with capitate bracteate
inflorescences, distributed in boggy places of Denmark to Manjimup, a
short range. Seeds look poorly dispersible (fig. 8.28, center left), but their
small size would lead one to expect that they could be scattered over
longer distances. The distinct ecological preferences of this species and its
habit (very short stems; it grows as a “cryptic” understory element in bog
scrub) may help explain why it is so restricted.

Darwinia hypericifolia is a Stirling Range endemic. Its fruits (fig. 8.28,
center right, probably somewhat immature) are relatively large, suggesting
lowered dispersibility. However, two other Stirling species of Darwinia,
each restricted to a single peak, distinctly show sharply reduced dispers-
ibility. Darwinia macrostegia (fig. 8.28, bottom left) has fruits with per-
sistent perianth, but the body of the shiny fruits is large and relatively
heavy. Darwinia meeboldii has fruits in which perianth is absent, and
their large size and weight and shiny surface suggest no obvious dispersal
mechanism (fig. 8.28, bottom right). Thus there is indeed reduced dispers-
ibility in those species of Darwinia that are restricted to peaks in the
Stirling Range. Many other examples of change in dispersibility in the
southwestern flora could surely be uncovered. For example, Stylidium
characteristically has minute, wind-borne seeds. However, Stylidium galt-
oides, an endemic of East Mt. Barren, has notably large seeds. This
change may be related to restriction of this peculiar species by its
preference for quartz-block hills localized along the southwest coast. Loss
of dispersibility in this case is entirely “insular” in that its strange habitat
is just as much an island as is a true oceanic island.

Loss of dispersibility can be detected by casual perusal of certain
other southwestern Australian genera. For example, the fruits of Eucalyptus
macrocarpa are notably large for the genus. Seeds are correspondingly
large, and one may suspect loss of dispersibility. This species is, in fact,
fairly narrowly restricted to sand plains near Perth.

316

Adaptive Radiation: Flora of Southwestern Australia

Insectivorous Plants

Discussion of insectivorous plants in southwestern Australia would be
merely presentation of a botanical curiosity, except that they illustrate
well the nature of evolution in the southwest. The number of species
(especially in Drosera) is greater than in any comparable area. This may
reflect capability of insectivorous plants to cope with nutrient-deficient
soil conditions in southwestern Australia, together with opportunity in
this area to radiate. Insectivorous plants in Western Australia have been
the subject of a good, recent, semipopular book (Erickson, 1968).

Cephalotus (fig. 8.29, top left, ght), a monotypic genus, is the sole
representative of Cephalotaceae and is endemic to boggy areas of south-
westernmost Australia. The pitchers of this plant, although somewhat
reminiscent of those of Sarraceniaceae or Nepenthaceae, are a totally
independent development. It is the only pitcher plant in temperate areas
of the southern hemisphere, and is apparently a derivative of saxifragoid
ancestors. Drosera (Droseraceae) has experienced explosive evolution—
which has taken the form of adaptive radiation—in southwestern Australia.
Of the 56 species endemic to Australia, 47 are restricted to southwestern
Australia. (The total of Drosera species worldwide is 90.)

Some southwestern species of Drosera appear to be annuals, such as
D. pulchella (fig. 8.29, bottom left). Other minute rosette Droseras may
perennate; during dry periods, foliage leaves disappear, and the plant
consists only of white, fimbriate scales covering the apical bud. At another
extreme is a plant of seasonally wet sand flats, D. gigantea (fig. 8.29, bottom
right), which grows to 1 m high or more, is prominently branched, and
perennates from underground stems. Some, such as D. stolonifera, grow
by means of underground stolons. Drosera macrophylla (fig. 8.34, top left,
right) is a perennial species with semi-succulent leaves. The vining species,
such as D. heteromera (fig. 8.11, bottom left), grow by means of succulent
underground structures that have been termed bulbs (fig. 8.30, top left).

We are only beginning to understand the diversity of the many species
of Utricularia endemic to Western Australia. Bladders of these differ
markedly from one species to another, as do flowers (see Erickson, 1968).

317

FIG. 8.29. Tendencies in insectivorous plants of southwestern Australia. TOP LEFT, RIGHT,
Cephalotus follicularis (Cephalotaceae). trop LEFT, habit, showing leaves modified into pitch-
ers. TOP RIGHT, detail of pitchers, one of which is shown in longisection. BOTTOM LEFT,
Drosera pulchella (Droseraceae), an ephemeral species; leaf rosette 2 cm in diam.; Bornholm.
BOTTOM RIGHT, D. gigantea, with stems to 0.7 m tall; Cannington.

318

Adaptive Radiation: Flora of Southwestern Australia

We know virtually nothing about the significance of these varied expres-
sions, and to what prey and pollinators (flowers range from pale violet to
bright red orange) they may be related. Utricularia has speciated in the
southwest probably because of the “insular” distribution of the winter-wet
boggy areas, most of which dry completely in summer.

Adaptation to Fire

Adaptation to fire is clearly an aspect of adaptive radiation, but different
mechanisms may occur within a single plant, and so fire-resistance is best
discussed collectively. One would not exaggerate much by saying that all
native plants in southwestern Australia are adapted to fire. I can show
only a few of these to illustrate the range of adaptations, some of which
are unexpected and doubtless autochthonously evolved within the province.
Seeds of certain species are probably resistant to fire in various ways, and
may even require fire for germination, but this aspect appears to have
been examined little. Nonetheless, the variety of easily visible fire-resistance
mechanisms is astonishing, and they definitely form phases of adaptive
radiation. ‘The importance of this feature is considerable, for not only
have various groups adapted to fire in a variety of ways, but also re-
sistance to fire must have been one of the obstacles for establishment
of the orginal immigrants in the southwestern flora, an obstacle not
readily perceptible when we think about disharmony of the province.
If fire resistance has been a factor in producing disharmony, it is one of
the factors that, added to peculiarities of climate and soils, have provided
broad opportunities for radiation of those colonists able to establish.

UNDERGROUND RENEWAL ORGANS

Plants that can resist dry seasons by means of bulbs, tubers, or corms
are not unusual in temperate climates. Such organs are undoubtedly cor-
related with the dry summers of the southwest. They can also, however,
be considered as ideal preadaptations to surviving fires. Bulbs and bulblike
structures are obvious mechanisms, and those of Drosera (fig. 8.30, top
left) can be cited in this regard. Another group that typically does not
bear bulbs but has evolved them autochthonously in southwestern Australia

eA)

FIG. 8.30. Underground renewal structures, also capable of fire resistance, in plants of south-
western Australia. Top LEFT, Drosera macrantha (Droseraceae), “bulb” at base of stem; near
Perenjori. Top RIGHT, Phylloglossum drummondii (Lycopodiaceae); unearthed plant shows old
tuber, new tuber at base of leaf rosette; Cannington. BOTTOM LEFT, Chamaescilla corymbosa
(Liliaceae), base of unearthed plant, showing tubers; Three Springs. BorroM ricHT, Dampiera
wellsiana (Goodeniaceae), plant showing basal rosette which perennates from buds below
ground level; near Calingiri.

320

Adaptive Radiation: Flora of Southwestern Australia

is Stylidium. Ephemeral vernal species, like S. petiolare, can survive both
drought and fire by means of these structures (Erickson, 1958; Carlquist,
1969). Stylidium bulbiferum and other species of Stylidium with wiry
stems do not have below-ground bulbs but have renewal buds at nodes
on the wiry stems, and these can sprout after fire. Phylloglossum drum-
mondii (fig. 8.30, top nght), the curious monotypic genus of Lycopodiaceae,
has an anomalous structure, often termed a tuber, that survives both
drought and fire. This structure is evidently a sort of specialized leafless
axis, one of which is produced annually by each plant, that bends down-
ward into the soil, develops a protective epidermis, and can produce shoot,
strobilus, and roots the following year.

The ordinary type of tuber, as in Chamaescilla corymbosa (fig. 8.30,
bottom left) is obviously suited to drought and fire survival. There are
many monocotyledons in the southwestern Australian flora, such as orchids,
that have underground storage and perennating structures, tending to be
succulent. Many angiosperms, however, merely have renewal buds at or
slightly below the soil surface. Leaves can be burned (they would mostly
be dried during the summer anyway), but the plant is not killed by fire.
The rosette habit, as in many species of Stylidium, is an indication of this
adaptation to fire as well as to a mediterranean climate. Dampiera wellsiana
(fig. 8.30, bottom nght) is a typical rosette perennial. Caespitose, clump-
forming plants, such as Lomandra (fig. 8.31, top left), Conostylis, and
restiads, have nonsucculent underground stems that can provide renewal
buds. This is a very common type of fire survival that has even been
evolved autochthonously in the southwest in Banksia prostrata (fig. 8.6,
bottom right). Actinostrobus acuminatus (Cupressaceae) is unusual both
within its genus and within cypresses at large; it branches below ground,
sends up shoots, and bears cones at or near the surface of the ground—
where fire temperatures would be expected to be minimal. This habit is
perhaps also related to the fact that this (like the other species of
Actinostrobus) is a native of sand areas where growth underground is
easy, and where shifting of sand will be less harmful than to a species in
which roots could be exposed or buried too deeply.

A number of shrubs in the southwestern flora characteristically branch
from ground level. At ground level is a somewhat swollen stem that usually
gives rise to new shoots each year. Tetratheca (Tremandraceae), Dampiera

321

FIG. 8.31. Mechanisms for fire-resistance in the southwestern Australian flora. TOP LEFT,
Lomandra sp. (Liliaceae), clump with leaves burned but rhizomes sending up new leaves;
Regan’s Ford. rop ricHt, Scaevola oxyclona (Goodeniaceae), showing underground stolon;
Spargoville. Borrom Lert, Xanthorrhoea sp. (Liliaceae), plant burned but new leaves emerg-
ing; Regan’s Ford. sorrom ricut, Macrozamia riedlei (Cycadaceae), old plants with new
leaves after fire; note charred stem; Badgingarra.

322

Adaptive Radiation: Flora of Southwestern Australia

(Goodeniaceae), and Loudonia (Haloragaceae) are typical examples of
this growth form. Stems can die down to the basal “caudex” each year,
and the caudex can yield renewal buds after fire. This habit has been
evolved autochthonously in subshrubby species of Stylidium such as S.
faleatum, which I have observed sprouting from a calluslike, swollen stem
base after a fire in the Tutanning Reserve (Carlquist, 1969). Even species
of Acacia have this habit.

Stolons are an excellent source of fire-surviving renewal buds. Among
southwestern plants that have autochthonously evolved such structures are
Scaevola oxyclona (fig. 8.31, top right) and Drosera stolonifera.

WOODY PLANTS

Obviously a true shrub cannot possess renewal buds of the same kind
as a typical ““chamaephyte” or “geophyte.” Yet survival of at least some
main stems after fire would seem to have a potential selective value for
shrubs and trees. One mechanism for survival is the presence of a cork,
fire-resistant cortex, or bark. Another requisite is the ability to form buds
after fire not just by outgrowth of axillary buds but by formation of buds
de novo on whatever branches were not killed. Eucalypts are capable of
this, and tree eucalypts are surely a magnificent example of trees adapted
to fire dispersal (also with regard to fruits, as mentioned later). Other
Myrtaceae commonly have fire-resistant bark; the paperbark (Melaleuca
quinquenervia) is an excellent example. Unusual, corky, excrescent bark
has been developed in Melaleuca exarata (fig. 8.33, bottom left) and
Calothamnus gracilis (fig. 8.33, bottom right). Sprouting of branches—
protected by a thick, seemingly corky cortex and bark—after fire is shown
here for Banksia menziesii (fig. 8.32, top left). Of course any bark so
specialized with respect to fire resistance by means of cork development
must also be effective in desiccation resistance, so that separation of the
two phenomena is impossible. All of the shrubs that have fire-resistant
bark do, in fact, contain appreciable amounts of cork cells.

Some rosette trees have succulent stems together with relatively
noncombustible (or expendable) leaves. The only requirement of these
plants is that the living tissues of the trunk and the apical meristem not
be killed by fire. Macrozamia of the Cycadaceae (fig. 8.31, bottom right)
and the various species of Xanthorrhoea (fig. 8.31, bottom left) and Kingia

323

FIG. 8.32. Examples of fire-resistance seen in Proteaceae at Regan’s Ford, site of a fire the
year preceding photographing. ror LEFT, Banksia menziesii, partially burned trunk sending out
new shoots. Top RIGHT, B. menziesii tree, showing base, a swollen lignotuber from which new
shoots are emerging. BorroM LEFT, B. laricina, shrub burned and killed, but all fruits opened
by fire to release unharmed seeds. Borrom ricuT, B. laricina, conelike aggregate fruits, show-
ing one with opened woody follicles, and one with unopened follicles.

324

FIG. 8.33. Examples of fire-resistance relating to cork formation in plants of southwestern Aus-
tralia. rop LEFT, Hakea obliqua (Proteaceae), opened fruit, valves 4.5 cm long, showing promi-
nent cork excresences. Top ricHT, H. pandanicarpa, unopened fruit (6 cm long) on plant,
showing numerous small corky pyramids. Bortom LEFT, Melaleuca exarata (Myrtaceae), show-
ing stems bearing prominent corky ridges; Cape Le Grand. porrom ricut, Calothamnus

gracilis (Myrtaceae), which bears ovaries and fruits embedded within stems that develop
undant cork; Wanneru.

25

Adaptive Radiation: Flora of Southwestern Australia

of the Liliaceae (sometimes Xanthorrhoeaceae) are excellent examples.
There are very few individuals of Macrozamia or Xanthorrhoea of any
age in Western Australia that do not have blackened stems, indicating
repeated burning. This is also true of the tree genus of Loranthaceae,
Nuytsia, which survives fire much as eucalypts and tree Banksias do.

“LIGNOTUBERS””

Bases of many Western Australian trees are swollen and burllike.
These are sources for renewal buds after burning. Because these trees have
woody branches that survive the dry season intact (unless burned), ligno-
tubers cannot be considered a desiccation-resistance mechanism. They are
clearly an adaptation to fire. Regrowth after fire from a lignotuber is
illustrated for Banksia menziesii (fig. $.32, top left). The mallee eucalypts,
such as E. macrocarpa, E. tetraptera, and E. preissii, all possess lignotubers.
Lignotubers are not unlike the caudex-like base of subshrubs such as
Dampiera, mentioned earlier. Lignotubers and caudex-bearing subshrubs
are also common in the Californian flora, where again they can be regarded
as mechanisms for fire resistance (Arctostaphylos spp., Adenostoma spp.).

FIRE-ADAPTED FRUITS

A fire-adapted fruit must be able to resist igniting, or at the most to
char on the outside but be thick enough that damaging temperatures do
not reach the seeds within. Superb examples of this adaptation are
presented by Proteaceae. Several southwestern species of Hakea have fruits
covered with corky excrescences. Hakea obliqua (fig. 8.33, top left), H.
pandanicarpa, H. brookeana, and H. roei are good examples of this
tendency.

Some Hakea fruits tend to open when mature and thoroughly dry—
presumably at the end of the potential fire season. The pair of winged
seeds in Hakea fruits is protected by fire up to this point, however. Even
fruits of Hakea that do not bear cork tend to be so massive and fire-re-
sistant that fire damage to seeds at any stage prior to dehiscence is fore-
stalled. This mechanism may also be hypothesized for Myrtaceae—espe-
cially Eucalyptus—in which the thickness and nonflammability of capsules
are noteworthy.

Another category of fire-adapted fruit is offered by Banksia. Banksia

326

Adaptive Radiation: Flora of Southwestern Australia

fruits are woody follicles, mostly clustered into conelike aggregate fruits.
Banksia fruits tend to stay closed indefinitely, even when dry, until the
high temperatures created by fire open them. Examples are shown here
for B. laricina (fig. 8.32, bottom left, right). Even if entire individuals are
destroyed by fire, as they are in B. laricina, fire temperatures trigger opening
of the fruits, which react not during the fire but immediately afterward;
thus the seeds are released intact. These seeds fall onto ground burned
clear of underbrush, and their potential for successful growth is enhanced.
One might say that because of the great frequency of fires in southwestern
Australia, all plants that grow there are, in a sense, pioneer plants.

Fruits many years old yet still unopened can be seen on trunks of
Banksia trees that have not experienced burning since production of those
fruits. In this respect the fruits resemble closely—even in appearance—the
cones of California’s closed-cone pines (Pinus radiata, P. attenuata, P.
muricata). Many other species in southwestern Australia also typically
retain seeds in fruits until fire creates conditions suitable for opening of
capsules. Especially prominent among them are the bottlebrushes Calo-
thamnus and Callistemon, as well as other Myrtaceae, such as a number
of species of Melaleuca. In Melaleuca exarata (fig. 8.33, bottom left) and
Calothamnus gracilis (fig. 8.33, bottom right), C. planifolius, C. lateralis,
C. microcarpus, and C. schaueri, capsules are immersed in the rapidly
exfoliating corky bark. Thus capsules in these species are as effectively
protected from fire damage as are the cork-covered follicles of Hakea
pandanicarpa.

FIRE-INDUCED FLOWERING

A phenomenon repeatedly noticed by botanists in Western Australia
is that some plants flower rarely or not at all until they have been burned.
Xanthorrhoea (Liliaceae) is often cited in this regard. Although one can
see individuals flowering without any apparent prior burning, a burned
area containing Xanthorrhoeas almost always shows abundant flowering
of this plant the following year. In this case, we have a condition one step
removed from fruits that open only after fire. The fruits of plants induced
to flower by fire need not be fire resistant for survival of seeds. In fact, the
likelihood of fires in a given region in two successive years is small
compared to that of fires at intervals of several years, during which time

SLi],

FIG. 8.34. Examples of flowering after fire in the flora of southwestern Australia. TOP LEFT,
RIGHT, Drosera macrophylla (Droseraceae), rosettes of plants. Top LEFT, rosette of plant from
a region not recently burned; near Toodyay. TOP RIGHT, rosette of plant that has flowered the
year after its habitat had been burned; Perenjon. BOTTOM LEFT, RIGHT, Lyperanthus nigricans
(Orchidaceae). BOTTOM LEFT, one-leaved but mature plant during potential flowering: season,
but not flowering; Badgingarra. BOTTOM RIGHT, flowering plants in a region that had burned
the preceding year; Cannington.

Adaptive Radiation: Flora of Southwestern Australia

combustible matter has had a chance to accumulate. Seeds that are shed
after fire have an enhanced opportunity of growth because, as in the
Banksia laricina example cited earlier, they are shed into an area cleared
the preceding year by fire.

Examples of induction of flowering by fire include an orchid, Lyper-
anthus nigricans, and Drosera macrophylla. One can often see these
species in vegetative form, in which case L. nigricans appears as a single
leaf (fig. §.34, bottom left), D. macrophylla as a sterile rosette of leaves
4—6 in. in diameter (fig. 8.34, top left) at the season when flowering of
such apparently mature plants would be expected. In a burned area,
however, these species can be seen in flower abundantly (fig. 8.34, top
night, bottom night). The contrast with angiosperms that release seeds
after fire is interesting. Both Drosera and Lyperanthus have relatively
delicate capsules containing seeds that are not at all resistant to fire. ‘Thus,
fire as a mechanism for inducing flowering is favorable, whereas fire would
be destructive to these plants when in fruit.

Discussion

Obviously not all aspects of adaptive radiation in seed plants of south-
western Australia can be covered here. The annual habit—a growth form
abundant in mediterranean climates—is present in the southwest but
not mentioned here. My comments can be considered only an introduc-
tion to the many truly exciting phenomena yet to be discovered or ex-
plained. Because these plants are vanishing, owing to increasing pressure
on land usage, work of a “natural history” nature is much needed. The
reasons for detailed field studies of the phenomena mentioned in this chap-
ter are just as compelling as the reasons for Darwin, Lack, and Bowman to
have studied the Darwin’s finches on the Galapagos Islands.

By discussing the flora of southwestern Australia in terms of phenom-
ena primarily, I have, unfortunately, diffused the presentation of adaptive
radiation in individual taxonomic groups. I hope that the reader interested
in the Australian flora will synthesize for particular groups, such as Acacia,
Eucalyptus, and Scaeyola, a picture (all too often missing where taxonomy
is the main, or sole, focus) of the remarkable adaptational aspects of specia-
tion in this flora.

329

Adaptive Radiation: Flora of Southwestern Australia

I also regret that I could not include a balancing account of adaptive
radiation in animal groups. Plainly my own training biases my presentation
in favor of plants. Elsewhere the reader can find pertinent data on frogs
(Main, 1965), lizards (Glauert, 1962), spiders (McKeown, 1963), and birds
(Keast, 1961). Further information on these groups from an evolutionary
perspective would be desirable, and a survey of insects in Australia and
particularly Western Australia from such a viewpoint would undoubtedly
be remarkably revealing.

References

Atlas of Australian resources. 1952-1966. Canberra, Department of National De
velopment. (The various maps and booklets that comprise this series were published
and revised variously, with some contributions apparently not credited to particular
authors.)

Beadle, N. C. W. 1966. Soil phosphate and its role in molding segments of the
Australian flora and vegetation, with special reference to xeromorphy and sclero-
phylly. Ecology 47:992-1007.

Beard, J. S., ed. 1965. Descriptive catalogue of West Australian plants. Perth, W.A.,
King’s Park Board.

Blackall, W. E., and B. J. Grieve. 1956, 1959, 1965. How to know Western Austra-
lian wildflowers. 3 vols. Nedlands, W.A., University of Western Australia Press.
Carlquist, S. 1969. Studies in Stylidiaceae: New taxa, field observations, evolutionary

tendencies. Aliso 7:13-—64.

Carolin, R. C. 1966. Seeds and fruit of the Goodeniaceae. Proc. Linnaean Soc. New
South Wales 91:58-83.

Carolin, R. C. 1968. Coopernookia: A new genus of Goodeniaceae. Proc. Linnaean
Soc. New South Wales 92:209-16.

Diels, L. 1906. Die Pflanzenwelt von West-Australien sudlich des Wendkreises. Die
Vegetation der Erde, ed. A. Engler and O. Drude, vol. 7. Leipzig, Verlag Wilhelm
Engelmann.

Erickson, R. 1958. Triggerplants. Perth, W.S., Paterson Brokensha.

Erickson, R. 1968. Plants of prey in Australia. Osbomme Park, W.A., Lamb Publication.

Glauert, L. 1962. A handbook of the lizards of Western Australia. Perth, W.A.,
Western Australian Naturalists’ Club.

Keast, A. 1961. Bird speciation on the Australian continent. Bull. Mus. Comp. Zool.
(Harvard) 123:305-495.

Lewis, H. 1962. Catastrophic selection as a factor in speciation. Evolution 16:257—71.

Main, A. R. 1965. Frogs of southern Western Australia. Perth, W.A., Western Aus-
tralian Naturalists’ Club.

McKeown, K. C. 1963. Australian spiders. 2d ed. Sydney, Angus and Robertson.

330

Adaptive Radiation: Flora of Southwestern Australia

Morcombe, M. K. 1968. Australia’s western wildflowers. Perth, W.A., Landfall Press.

Noakes, L. C. 1966. Geology. In Atlas of Australian resources, 2d ser., pp. 1-27.
Canberra, Department of National Development.

Raven, P. H. 1964. Catastrophic selection and edaphic endemism. Evolution
18:336-38.

Snaydon, R. W. 1970. Rapid population differentiation in a mosaic environment. I.

The response of Anthoxanthum odoratum populations to soils. Evolution 24:257-
69.

Went, F. W. 1971. Parallel evolution. Taxon 20:197-226.

331

C

[APT BIR
NINE

ADAPTIVE RADIATION: CONTINENTAL
ISLANDS, FRINGING ARCHIPELAGOS,
CONTINENTAL ISLANDLIKE AREAS

This chapter, although intended to cover instances of adaptive radiation
not included earlier, has been designed as a way of illustrating the various
ways in which areas within continents or near continents show evolutionary
characteristics of adaptive radiation much like those discussed earlier.

Madagascar

Madagascar may deserve more consideration in relation to adaptive radia-
tion than some of the areas discussed in earlier chapters. However, I have
hesitated to present an extensive synthesis of the Madagascan biota not
only because it is so complex compared with those of other islands (and
really deserves a book by itself), but also because of my own unfamiliarity
with this island.*

Madagascar is generally considered an old continental island and, as
such, might be expected to have relict groups. Chief among these relicts
are the lemurs. Known from fossils from the Paleocene and Eocene of
North America and Europe (Hill, 1953), lemurs have escaped extinction
in Madagascar. How much of the diversity of lemurs is due to autochtho-
nous evolution in Madagascar is therefore a difficult question. One could

* Such a book is now available: G. Richard and R. Battistini, eds., Biogeography and
ecology of Madagascar, The Hague, Dr. W. Junk, Publishers, 1972.

332

Continental Islands, Islandlike Areas, and Fringing Archipelagos

argue that absence of monkeys from Madagascar has permitted radiation
of lemurs there, but one could equally well argue that this absence merely
has permitted survival of lemurs in Madagascar, and has no bearing on
whether the Madagascan lemurs represent more than a single invasion of
that island. Significant contributions on the lemurs which also cite and
summarize earlier literature include the works of Hill (1953), Petter (1962,
1969) and Jolly (1966).

Although the tenrecs (Tenrecidae) have long been known as an en-
demic Madagascar family of insectivores notable for adaptive radiation,
we have lacked a careful field study of this group. This has now been admi-
tably supplied by Eisenberg and Gould (1970). Although they were unable
to provide new observations on some of the scarcer genera, such as Limno-
gale, these authors have provided a wide variety of observations on most
tenrecs, both in the field and in captivity, and have added many new as-
pects to knowledge of the adaptive radiation of the group; their book is
deserving of study.

Unfortunately, there is no such recent, comprehensive account of the
vanga shrikes (Vangidae), which form an example of adaptive radiation
rivaling the Darwin’s finches and the Hawaiian honeycreepers in many re-
spects. Distributional and habit data may be found in the field studies of
Rand (1936); some data and good illustrations are available in the work of
Milne-Edwards and Grandidier (1878-1885).

The iguanid genera of Madagascar, Hoplurus (Oplurus) and Chala-
rodon, and the boid snakes and pelomedusid turtle are often considered
relictual (Mertens, 1934). Uroplatidae, a family of lizards endemic to the
island, has no close relatives. Geckos have diversified on mainland Mada-
gascar, and distinctive species and subspecies may be found on offshore
islands. For a nearly complete taxonomic account of the diverse skinks,
geckos, and chameleons on Madagascar, the lizard fauna of Angel (1942)
will provide a convenient summary.

With respect to plants, the striking cactoid forms of the endemic
family Didiereaceae (perhaps a relative of Portulacaceae) are an obvious
example of radiation within the drier regions of Madagascar. The older
monograph of Choux (1934) has now been supplanted by the exhaustive
studies of Rauh (1956, 1961, 1963, 1965). An endemic family of dubious

333

Continental Islands, Islandlike Areas, and Fringing Archipelagos

affinities, Sarcolaenaceae (Chalenaceae), has radiated into various habitats
and shows great diversity not only in size, leaf shape, flower details and
anatomy (Gérard, 1919; Cavaco, 1952), but even to a remarkable extent in
its pollen tetrads (Carlquist, 1964).

Fringing Archipelagos

Recent offshore islands—either those formerly connected to mainland
areas, or those that have emerged as oceanic islands—often contain merely
depauperate floristic and faunistic assemblages, differing little from compa-
rable close mainland areas. However, evolutionarily active groups may well
be able to show at least some phases of adaptive radiation on insular areas.

California Islands. Among the Pacific offshore islands of California
and Baja California, only one group of plants, the genus Hemuizonia (Aster-
aceae: Heliantheae, subtribe Madiinae), shows what could be called adap-
tive radiation. ‘The insular populations of these shrubby Hemizonias may
well be to some extent relictual. This group of species is represented on the
mainland only by a small population of H. greeneana (otherwise on Guada-
lupe I.) at Punta Banda, near Ensenada, Baja California, and by H. min-
thornii, a xeromorphic species in the vicinity of Santa Susanna Pass, south-
ern California. Even if the shrubby Hemizonias have survived more abun-
dantly on the offshore islands than on mainland areas, there is every reason
to believe that autochthonous diversification has occurred on the islands
(Carlquist, 1965). Disregarding the markedly different species of the
southern islands for the moment—H. frutescens, H. greeneana, H. palmeri,
and H. streetsti—we find that H. clementina on the northern islands seems
to be in the process of forming distinctive races. Although populations
within any given island may be variable, differentiation among the islands
does seem to be occurring. Such islands as Santa Barbara, San Clemente,
and Anacapa have populations that appear to be differentiating with re-
spect to hairiness, leaf shape, leaf size, leaf margins (toothed or not), and
other characteristics. A careful, perhaps statistically oriented, study of in-
sular populations of H. clementina would be of interest for showing modal
differences, if any, and whether any of these differences are adaptive in
nature. Species of the southern islands appear to differ in characteristics

334

Continental Islands, Islandlike Areas, and Fringing Archipelagos

that could be called adaptive, keyed to their particular habitat preferences
(Carlquist, 1965).

Gulf of California Islands. Differentiation of races of Uta stansburiana
(sensu lato; segregate species have been recognized) has occurred on islands
in the Gulf of California. These differences are chiefly in respect to body
size and scale size. Although these are clear indications of differentiation,
and can be called, as Soulé (1966) does, “radiation,” some aspects are
adaptive and some may not be. This group is reviewed in chapter 15 in re-
lation to the phenomenon of insular gigantism of animals.

Philippines. Although they are offshore islands only in the wider and
much more profound sense, mention should be made of certain apparent
instances of adaptive radiation in the Philippines. This phenomenon is not
characteristic of most plant and animal groups in the Philippines, presum-
ably because this archipelago is relatively harmonic in many respects with
the floras and faunas of the major islands of Malesia, which are themselves
relatively harmonic. However, mention should be made of the murid ro-
dents of Luzon (Dickerson, 1936). Presumably Luzon’s inaccessibility to
mammals has permitted radiation of rodents there, just as rodents have
undergone radiation in Australia (albeit not spectacular in comparison to
the marsupials there). Adaptive radiation has taken place in Brachymeles
of the Scincidae (Brown and Rabor, 1967) and perhaps in some snakes
(Brown and Alcala, 1970). However, one Philippine example of adaptive
radiation is outstanding, the cyprinid fishes of Lake Lanao (see the next
section).

Large Freshwater Lakes

Although other freshwater lakes demonstrate aspects of speciation very well,
there are five in which adaptive radiation of fishes has been spectacular:
the three great lakes of Africa—Lake Victoria, Lake Tanganyika, and Lake
Malawi (= Lake Nyasa)—for the cichlids; Lake Lanao, Mindanao, Philip-
pines, for the cyprinids; and Lake Baikal, eastern USSR, for the cottids.
Because different groups of fishes are involved, each of these situations
casts light upon the others.

Lake Lanao. From fairly conclusive evidence, we know that Mindanao,
uplifted during Miocene time, formed a central plateau by means of basalt

335

Continental Islands, Islandlike Areas, and Fringing Archipelagos

flows. Within this plateau, volcanoes formed a dam around streams,
whereby Lake Lanao was created. The damming of Lake Lanao at the
Camp Keithly outlet has been quite reliably dated (Bailey Willis, in Myers,
1960a) at the significantly recent time of approximately 10,000 years ago.
The basis for this dating is the erosion of the Agus River canyon at the
main outlet of Lake Lanao. Although this event does not completely pre-
clude possible earlier instances of damming of the lake (Myers, 1960a), it
is conceivable that virtually all of the remarkable evolution of cyprinid
fishes in Lake Lanao has taken only 10,000 years. Herre (1933) confidently
states that one cyprinid, Barbodes binotatus, still extant in the Philippines,
or possibly B. quinquemaculatus, was the ancestor of the 5 genera (Bar-
bodes and 4 endemic genera) and 18 species of cyprinids endemic to the
lake. Even more species may yet be found (Myers, 1969a).

Herre (1933) considered that the Barbodes ancestor (presumptively
B. binotatus; B. quinquemaculatus may be only a geographical race of B.
binotatus on the Zamboanga Peninsula) developed into the extant Lanao
species, B. tumba, which in turn invaded streams and lakes of the Lanao
plateau as a whole. It has presumably replaced B. binotatus, which no
longer occurs on the Lanao plateau. As Lake Lanao reached maturity, a
Barbodes flock differentiated. Barbodes tumba has become an inhabitant
of the Lanao streams and tributaries, and occurs in the turbulent outlet at
the falls above the Agus River. However, it evidently once gave nse to a
flock of species of the shallow, still margins of the lake. In these shallows,
where Potamogeton is the common aquatic plant, occur B. amara, B. lana-
oensis, B. lindog, and a derivative of B. lindog, B. disa. Another shallow-
water derivative of B. lindog has been recognized as a monotypic genus,
Spratellicyparis palata, because of its peculiar lower-jaw modifications.
Barbodes clemensi, notable for its robustness and distinctive color (green
above, yellow beneath, lustrous gold on its sides), occurs in the Potamoge-
ton zone, as does B. manalak, the largest species of Cyprinidae in the entire
Philippines.

A deep-water derivative of B. tumba is B. baoulan, which rises to sur-
face depths only during storms or in midwinter. Other endemic species of
Barbodes are restricted to shallow, muddy bays with a junglelike vegetation
of reeds and similar plants. These species are blackish or dusty brown, a
melanism evidently part of their adaptation to these darkish waters. Herre

336

Continental Islands, Islandlike Areas, and Fringing Archipelagos

(1933) cites as a parallel the strongly melanistic goby from other densely
marshy Philippine lakes, Glossogobius griseus.

A very small Lanao endemic, Barbodes sirang, is also melanistic, but
its habits are unknown. Large-headed species of Barbodes, the adapta-
tional mode of which is not yet clear, include B. flavifuscus and B. katolo,
both believed to be offshoots of B. amara. From B. katolo, in turn, two
endemic genera are believed to have arisen. The monotypic Cephalokomp-
sus (C. pachycheilus) is notable for a continual postlabial groove on the
lower jaw (this grove is interrupted in Barbodes and most other cyprinids).
Ospatulus (two species, O. truncatulus and O. palaemophagus) is distinc-
tive by virtue of its truncate lower jaw, much shorter than the upper.

Mandibularca (monotypic, M. resinus) is, according to Herre (1933),
“the most remarkable of the Lanao Cyprinidae.” It occurs only in the ex-
cessive turbulence of the Agus River outlet of Lake Lanao. The lower jaw
is exceptionally elongate, upwardly curved, shaped like a spatula, and
“fleshless.’”’ Herre believed that it originated from Barbodes clemensi, some
individuals of which have lower jaws that seem to foreshadow the condi-
tion in Mandibularca.

The various papers by Herre (1924, 1926, 1932, 1933; Herre, in Dick-
erson, 1928) have not, surprisingly, been followed by new field studies, but
his material does suggest that the Lanao cyprinids are a virtually unparal-
leled situation (see Myers, 1960a).

Lake Malawi (Lake Nyasa). In dealing with adaptive radiation and spe-
ciation in Lake Malawi and the other African lakes, we are dealing with
cichlid fishes: 78 percent of the Lake Malawi fishes, 62 percent of those in
Lake Tanganyika, and 71 percent of those in Lake Victoria are cichlids.
The number of these cichlid species in the three lakes is about 193, 134,
and 127, respectively, and the percentage of endemic cichlids in the three
lakes is 97, 98, and 98, respectively. Lake Malawi has 20 endemic genera
of cichlids, Lake Tanganyika 30, and Lake Victoria 4 (Lowe-McConnell,
1969). Thus with respect to cichlids the great African lakes are as remark-
able as Lake Lanao. The reasons for this situation are not entirely clear in
the case of Lake Malawi. Fryer (1959a) presented an account of evolution
of the Lake Malawi fishes, but Myers (1960b) felt that more data on pos-
sible ecological distinctions among the numerous cichlid species and gen-
era were required to explain this remarkable autochthonous evolution.

ao)

Continental Islands, Islandlike Areas, and Fringing Archipelagos

Fryer (1960) replied that some ecological distinctions do exist. He cited,
for example, adaptive radiation in mimicry of these fishes, as studied by
Trewavas (1947), in nest-building and other habits (Lowe, 1953, 1959;
Fryer, 1956, 1959b), and in such habits as burrowing in sand and feigning
death. Fryer also cited morphological features: Petrotilapia tridentiger has
mouth structures that facilitate scraping of algae from rocks; Lathrinops
furcifer has a protractile mouth capable of sifting chironomid larvae from
sand. However, Fryer (1960) still relied on factors he cited earlier: scat-
tered habitats around the lake margins (rocky outcrops, for example),
various types of precinctiveness, and predation. With respect to precinc-
tiveness, he noted that the Lake Malawi cichlids typically are mouth-
breeders; the young shoal together; the “Tilapia mark” on juvenile fins
serves as a signal; and the young, while losing parental orientation as they
adolesce, do not scatter. Fryer stressed that in Lake Malawi predation is
heavy, and that crossing a sandy bottom to another rocky outcrop would
render individuals visible and easier prey. Predation, as noted by Green-
wood (1964), might help coexistence of prey species by keeping them be-
low the level where they would compete excessively for food, presumably
to the point where one species would replace another.

In addition, Lowe-McConnell (1969) cites temporal and ethological
barriers for cichlids in Lake Malawi. Tilapia saka breeds in shallow water
before the rainy season; T’. lidole breeds at the same time but in much
deeper water or off open sandy beaches; T. squamipinnis (a species very
close to T. saka) breeds in deeper water during the rainy season (Decem-
ber to February). Also differences in breeding colors can be cited: T. saka
and T. lidole have black males, whereas T. squamipinnis males are sky
blue, and white at the face (Lowe, 1953).

It seems that we cannot compare evolution of the Lake Malawi
cichlids to that of the Darwin’s finches, where noncompetitive utilization
of food sources and visible geographical barriers (channels between islands)
provide obvious partitioning devices. In the cichlids, these factors do oper-
ate to some extent, but behavioral and other differences must also operate
as partitioning devices. Admittedly, fish may be more difficult to study
with respect to both behavior and ecological preferences than birds.

Lake Tanganyika. Lake Tanganyika is at least 1.5 million years old,
and may be even older (Brooks, 1950). The characteristics of its fish fauna

338

Continental Islands, Islandlike Areas, and Fringing Archipelagos

suggest a more mature stage in evolution if we compare its cichlids with
those of Lake Malawi (Poll, 1946). Lowe-McConnell (1969) claims that
most Lake Tanganyika cichlids tend to live around rocky outcrops at lake
margins. Thus, these outcrops may function as centers, if not discrete sub-
divisions, for speciation, as Fryer suggests they do in Lake Malawi. In addi-
tion, Lake Tanganyika cichlids exhibit notable morphological specializa-
tions (Myers, 1960b). Perissodus has large, double-pointed, heavy-based
teeth. Asprotilapia and Enantiopus have pelvic fins specialized for bottom-
living. Myers (1960) draws comparisons—typical in illustrations of adaptive
radiation—between Lake Tanganyika cichlid genera and non-cichlid fami-
lies they resemble: Telmatochromis is comparable to Blenniudae, Tropheus
to Girellidae, Asprotilapia to European Percidae.

Lake Victoria. In Lake Victoria, Haplochromis is a major genus, with
about 120 species. Feeding habits vary. Twelve species feed on molluscs,
but in two different ways. The least specialized (i.e., most like riverine
ancestors), about 40 species, feed on insect larvae and detritus. Four spe-
cies feed on algae and 1 on vascular plants. At least 50 species prey on
other fish: 8 of these are known to feed on eggs and fry of other Haplo-
chromis species, presumably by sucking them from the mouths of their par-
ents. With such differences, in addition to some of the evolutionary factors
cited for the other great African lakes, one can understand how adaptive
radiation and speciation may have occurred in a lacustrine “island” where
terrestrial concepts of adaptive radiation do not, at first, seem to apply—
and where methods of study must be quite different.

Lowe-McConnell (1969) cites an interesting ecological mechanism
in the Malagarazi waterlily swamps of Tanzania. Characoids feed on and
partially digest leaves, buds, and fruits of the waterlilies. This semidigested
material and its contained microorganisms form a carpet on the swamp
bottom, which four cichlids (two Tilapia species, two Haplochromis spe-
cies) feed upon. In turn, small fish of these four cichlids can be prey of
piscivorous species of cichlids.

Lake Baikal. Lake Baikal presents a different picture, if only because
of its age and the more markedly cold climatic regime. It is ancient com-
pared to other freshwater lakes; its southern basic dates from the Paleocene.
In Lake Baikal, the family Cottidae is represented (Taliyev, 1955) along
with an endemic family, Comephoridae, derived from cottids. One sub-

339

Continental Islands, Islandlike Areas, and Fringing Archipelagos

family of Cottidae is sometimes recognized as a distinct family, Cotto-
comephoridae. Comephoridae are not found along the shore, but occupy
open waters of Baikal, where they live from the surface to below 500 m.
There are two species, Comephorus dybowskii, which mates from Septem-
ber to November, and C. baicalensis, which mates from April to June
(Koryakov, 1955). Comephoridae give birth to living young.

In Cottidae, there are five endemic genera. Cottocomephorus (two
species) inhabits the coastal belt and spawns at shallow places near the
shore, on the rocky bottom. Abyssocottus (five species), as its name sug-
gests, inhabits great depths; it has a slender form and an elongate jaw. The
other endemic genera include Metacottus (monotypic), Batrachocottus
(four species), Paracottus (four species) and Procottus (monotypic). Tali-
yev’s (1955) illustrations (some reproduced by Kozhoy, 1963) and data
show these genera to be amply distinct.

As might be expected, a lake as ancient as Baikal demonstrates adap-
tive radiation in other groups as well. Fifty-six species of Mollusca are
endemic to Lake Baikal. Most of these belong to six genera endemic to
Lake Baikal and nearby Lake Kosogol (Kozhov, 1963). Two of these genera
comprise an endemic family of prosobranchs, Baicaliidae. The range of
shell forms in Baicalia is quite remarkable (Kozhov, 1951, 1963). It ranges
from ovoid to turret-shaped, smooth or with various types of ridges (B.
costata) or fine spines (B. ciliata). This radiation in form is not unlike that
of the Hawaiian amastrids, and demonstrates that isolating mechanisms
have been operative within Lake Baikal, which is, however, much older
than the Hawaiian Islands.

Another Baikal group that shows considerable diversification is gam-
marid shrimps. These range from colorless, pale-eyed, slender types of
deep waters (Abyssogammarus) to markedly spinescent genera (Acantho-
gammarus). The nature of ecological distinctions among these species does
not seem clear; all gammarids have long breeding seasons (Kozhov, 1963).

Continental Areas

Lack (1971) has hypothesized, with good reason, that there are probably
few or no open ecological niches for birds on islands at present, unless

340

Continental Islands, Islandlike Areas, and Fringing Archipelagos

habitat alteration by man occurs. If this is true for land birds, which have
moderately good dispersal to oceanic islands (most of which are recent),
one can well imagine the situation, by comparison, on continents. Conti-
nents are undoubtedly “saturated” constantly, but shifts in geology and
climate make them subject to floristic and faunistic readjustments as par-
ticular areas become temporarily disharmonic. For example, previously
glaciated areas, volcanically devastated areas, and emerged embayments
offer new territories. The process by which niches are filled is not the same,
in general, as on islands. Because of the harmonic nature of continental
biotas, invasion of new niches tends to be “buffered” since so many “en-
trants’” compete for niches. Some radiative speciation can occur, as exem-
plified by Keast’s (1971) summary of niche-occupation following glaciation
on Tasmania. Because, in this case, dispersal across Bass Straits is involved,
there is some simulation of dispersal to distant oceanic islands, but there
is less disharmony, so each of the immigrants has been able to radiate only
to a limited extent. A clear exception to buffered radiation on continents
occurs when large freshwater lakes are created, because they usually offer
a variety of habitats and must begin the process of population with a few
occupants suited to fresh water (perhaps mostly from inflowing streams,
or via long-distance dispersal from other lakes)—in other words, the situa-
tion is as disharmonic as that of an oceanic island.

Cataclysmic events leading to adaptive radiation do occur, however,
in terrestrial continental areas. The California flora offers some examples.
With the mid-Tertiary drying of the Southwest, a fertile field for more
xerophytic plants arose. Although a few boreal groups (e.g., Arctostaphylos
—perhaps not strictly a boreal example) have adapted to Californian xero-
morphy, most of the entrants are from more southerly, chiefly Mexican,
groups. For example, the invasion of California by various groups of Astera-
ceae (a family notably competent in dry situations) has led to adaptive
radiation in such genera as Eriophyllum, Haplopappus, and the genera
that comprise the tarweeds (Heliantheae, subtribe Madiinae). All of these
are not only capable of success in a mediterranean climate but also show
preferences for pioneer situations—and pioneer situations tend to be dis-
tributed in an insular fashion. The various papers of Axelrod can be con-
sulted for a fuller account of occupancy of the southwestern United
States during the Tertiary.

341

Continental Islands, Islandlike Areas, and Fringing Archipelagos

Southwestern Australia, cited in chapter 8 as a superb example of an
insular area, is perhaps the most striking continental area for adaptive
radiation, at least where angiosperms are concerned. However, Australia as
a whole is insular not merely in being the smallest of the continents, but
in being so far removed from areas of similar temperate climate and being
separated from Indo-Malesia for very protracted periods. Consequently,
one is not surprised that “heightened” examples of adaptive radiation
occur there, such as the marsupials, proteads, and eucalypts.

Many other interesting examples of adaptive radiation on continents
can be cited, such as the radiation into mimicry situations and other leaf
types by the South African “living stones” of the Aizoaceae (Conophytum,
Lithops, Pleiospilos). Radiation in plant form, floral form, and pollination
mechanisms in the stapelioids and ceropegioids (Apocynaceae) of South
Africa offer another series of examples. Euphorbia species in Africa have
experienced a wide range in diversification of growth form, as have the
cacti in the Americas—and the radiation in growth form is only one dimen-
sion of radiation in those groups. Interestingly, many of these examples
are from drier regions, which suggests that only a limited number of inva-
sive groups can thrive in such conditions, but if they do, they can radiate
widely.

Thus, phases of insular adaptive radiation can illuminate continental
situations and lend perspective to their interpretation. Various authors
have speculated that if the “supralimital” products of adaptive radiation
in insular situations (e.g., the Lanao fishes or the Hawaiian honeycreepers)
had occurred on continents, we would, in fact, be seeing how larger taxo-
nomic categories originated on continental areas. The interplay of conti-
nental and insular patterns, and their analysis separately, then synthesis,
has of recent years been highly illuminating. We can achieve more com-
prehensive constructs of evolution if different types of evidence, from dif-
ferent areas and different taxonomic groups, are fed into the synthesis.

Oceanic Areas

Shelf Areas. Some areas of ocean tend to have insular qualities, but
not in the sense of offering both definitive barriers from other ocean areas

342

Continental Islands, Islandlike Areas, and Fringing Archipelagos

and, at the same time, a wide variety of habitats. The “East Pacific Bar-
rier’ (Briggs, 1961) and other restricted shelf areas (Briggs, 1970) are “iso-
lated environments within the ocean,” although they are not within them-
selves highly diversified so as to host spectacular examples of adaptive
radiation, although endemism can be high.

Invertebrates on Reefs. Other marine organisms that are not wide-
spread but have very localized requirements could be cited as examples
of adaptive radiation. ‘The mollusc Conus and other invertebrates notable
for large numbers of “sympatric” species on reefs and evident resource-
partitioning of reef areas of given islands (Kohn, 1968, 1971) should be
cited. According to Kohn (1971), specialization to particular food sources
is of prime importance, adaptation to particular habitat niches (microhabi-
tats) of secondary importance, in permitting the numbers of “co-curring”
species on given islands. Such resource-partitioning is one of the evidences
for the occurrence of adaptive radiation.

Pediculate Fishes. The great deeps of the Pacific are topographically
“islands,” but they are also highly specialized habitats. The pediculate
fishes that have entered these deeps are particularly interesting. They can
be considered products of adaptive radiation not merely within those deeps,
but as part of a larger scheme in which slower-moving predatory fishes (ex-
actly which contemporary representatives gave rise to pediculate fishes is
not certain) developed lures or “angles” and diversified into various habi-
tats. These anglerfish may be divided into three groups, representing three
adaptational modes:

Antennariidae (frogfishes), in depths of less than 50 fathoms,
mostly shallower

Ogcocephalidae (batfishes), in depths of 50 to 150 fathoms

Ceratioidea (lantern fishes), comprising several families, in depths
of 150 to 3600 fathoms (bathypelagic)

Assuming these groups—and their ancestors—to be products of adap-
tive radiation, we may find cycles of radiation within each of them. In the
Antennariidae, radiation seems to be enhanced by the slowness in bodily
movements during feeding activities. Pectoral fins are modified for walk-
ing. As voracious predators, they must lure prey and then suck it in rapidly,
rather than pursue it. This behavior, in turn, is linked to camouflage: they

343

Continental Islands, Islandlike Areas, and Fringing Archipelagos

must be virtually invisible to prey. In Pacific, Atlantic, and Indian Ocean
species of Antennarius, speciation has been accompanied by mimicry to
resemble backgrounds of the habitats which they frequent. For example,
A. sanguifluus of the Pacific has red and brown splotches that comprise
camouflage patterns closely resembling the coral and algal backgrounds
of the reefs where it lives. Various features of Antennarius morphology
contribute to radiation in background mimicry (Allen, 1970; Jordan, 1905;
Schultz, 1957, 1964; Strasburg, 1966). Condensation of form and irregu-
larities of outline are important in this regard; fins are squat, thick, and
colored like the body; eyes are inconspicuous. Various flaps of tissue and
ciliate appendages around the mouth, as well as elsewhere on the body,
simulate algal lobes or filamentous algae or animal colonies. Scales would
decrease mimicry ability; Antennarius lacks scales and has instead dermal
papillae whose texture is capable of various resemblances.

Species of Antennarius within a given region may be assumed to have
a low dispersal rate (for fish); their endemism to particular islands and
atolls is high. Where several species occur together, we may assume the
likelihood of monophylesis or development within at least close portions
of an ocean—although this is not necessarily true. If Antennarius species
in an archipelago are monophyletic, we have a definitive example of adap-
tive radiation. This does not alter the picture of Antennarius as a whole
representing adaptive radiation for purposes of mimicry. Whichever ex-
planation obtains in the Bermuda species of Antennarius (table 9.1), we
have an illustrative example. Moreover, each of the Bermuda species of
Antennarius has a distinctive lure, or angle (the illicium, tipped by the
esca, or ‘“‘bait”’). In A. multiocellatus, the esca is bulbous, with lateral hairs.
In A. ocellatus, the esca consists wholly of hairs. The other three species
have lanceolate lobes on the esca: three in A. tenebrosus, several in A. pau-
ciradiatus, two with a central bulb in A. bermudensis. Because the illicium
is so rarely seen in action, we cannot be sure of the resemblances, created
by these forms, to animals eaten by prey of the Antennarius, but evidently
a radiation has occurred. In the Bermudan species of Antennarius, distinc-
tive color patterns and depths assure lack of overlap in ecological niches.
The predators of Antennarius are not known, but one may assume that
such precise mimicry has been the product of intense selective forces, and

Sak

Continental Islands, Islandlike Areas, and Fringing Archipelagos

TABLE 9.1.
Bermudan species of Antennarius

Probable

Depth and habitat Color and pattern resemblances

A. multiocellatus Less than 20 ft, Dark gray, with spots. Loggerhead sponges
sandy bottoms

A. ocellatus Tidal pools Brown (juveniles with Colonies of inverte-

pink spots); “ocelli’” — brates on rocks

A. tenebrosus Less than 8 ft, rocky Brown, with yellow — Rocks, with flat algal
outcrops and black spots and animal colonies

A. pauciradiatus 25-50 ft or more, Brownish yellow, Aging coral heads
coralline bottoms mottled

A. bermudensis_ Beds of coral and Rust color, with Corals and red algae

coralline red algae __ reddish spots

predators would tend to enforce the excellence of mimicry, lack of inter-
mediacy, and ecological separation.

Histrio histrio, the sargassum fish, forms a distinct subfamily in the
Antennaridae. Unlike Antennarius, it floats freely among the tangles of
Sargassum in the Sargasso Sea. Its fins are better suited to locomotion than
are those of Antennarius, but its capture of prey is not achieved by pursuit,
at which Histrio would be inadequate. Its coloration perfectly matches the
various shades of Sargassum, and flaps and tubercles resembling Sargassum
fronds occur not only on the upper surface and sides, but on the lower sur-
faces as well, for predators are, in fact, more likely to occur at levels below
it. Unlike other Antennariidae, Histrio has a mouth whose inside is a tawny
color matching Sargassum shades.

Ogcocephalidae are also widespread in Pacific and Atlantic areas, but,
as indicated earlier, favor depths intermediate between those of the anten-
narioids and the ceratioids. They are, like Antennariidae, poor at swim-
ming. Flattened laterally like the manta, these fish have pectoral and pelvic
fins modified to act as forelegs and hindlegs. Although there are distinctive
patterns and textures, batfish have radiated with respect to morphology of
the esca as well (Bradbury, 1967).

Ceratioidea is the name generally given to an assemblage of rather
tentative families and genera of lantern fishes, which inhabit great depths.

345

Continental Islands, Islandlike Areas, and Fringing Archipelagos

The uncertainty of taxonomy derives from the fact that known specimens
often represent unique finds, occasionally taken during deep-sea dredging,
rather than series of specimens of any one species. We know little or noth-
ing of their habits. Bertelson’s (1951) summary provides most of the in-
formation available on them, although additional illustrations may be
found in Jordan (1905). They represent an entry into extreme abyssal
conditions by the same general pediculate stocks from which Antennari-
idae and Ogcocephalidae have also been derived.

Modifications for life in abyssal conditions include lack of bones (ex-
cept for the large and powerful jaw), condensed form, lack of a swimming
bladder, probable capability of oxygen intake through capillaries near the
skin surface, and marked sex dimorphism in at least four, perhaps six, fami-
lies. In these families, at sexual maturity the small males attach themselves
to various localities on females, which are relatively large in proportion, and
become interconnected to the vascular system of the female and degen-
erate in several ways. The parasitic habit of males presumably assures fer-
tilization of eggs under conditions of great sparsity of population and difh-
culty of pair-occurrence when swimming ability is so limited and perpetual
dark is present.

The ceratioids have bioluminescent, stalked lures (except for Neocera-
tias and Coelophryne) and other portions of their bodies may be lumines-
cent as well. Because our concepts of their phylogeny and taxonomy are
tentative, we cannot construct anything but a fragmentary picture of adap-
tive radiation within ceratioids. Presumably the remarkable differences in
morphology of illictum and esca among the known species (Bertelson,
1951) do, however, represent adaptation to distinctive predatory behavior
and perhaps diet. The means whereby knowledge of their habits can be
improved seems technologically almost unobtainable at present. Needless
to say, ceratioids brought to the surface are dead, and their sparsity alone
would make study difficult even if technical problems of observation at
great depths could be overcome.

References

Allen, G. R. 1970. Two new species of frogfishes (Antennariidae) from Easter Island.
Pacific Sci. 24:517—22.

346

Continental Islands, Islandlike Areas, and Fringing Archipelagos

Angel, F. 1942. Les lézards de Madagascar. Mém. Acad. Malgdche 36:1-193.

Bertelson, E. 1951. The ceratioid fishes. Ontogeny, taxonomy, distribution, and
biology. Dana Rep. (Carlsburg Found.) 39:1—276.

Bradbury, H. G. 1967. The genera of batfishes (Ogcocephalidae). Copeia (1967):
399-422.

Briggs, J. C. 1961. The east Pacific barrier and the distribution of shore fishes.
Evolution 15:545—54.

Briggs, J. C. 1970. Tropical shelf zoogeography. Proc. Calif. Acad. Sci., 4th ser.,
38(7):131-38.

Brooks, J. L. 1950. Speciation in ancient lakes. Quart. Rev. Biol. 25:30-60, 131-76.

Brown, W. C., and A. C. Alcala. 1970. The zoogeography of the herpetofauna of
the Philippine Islands, a fringing archipelago. Proc. Calif. Acad. Sci., 4th ser.,
38(6):105-30.

Brown, W. C., and D. S. Rabor. 1967. Review. of the genus Brachymeles (Scinci-
dae) with description of new species and subspecies. Proc. Calif. Acad. Sct.,
4th ser., 34(15):525-48.

Carlquist, S. 1964. Pollen morphology and evolution of Sarcolaenaceae (Chlaena-
ceae). Brittonia 16:231—54.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Cavaco, A. 1952. Chlaénacées. In Flore de Madagascar et des Comores, ed. H. Hum-
bert, 126e:1—37. Paris, Musée National d’Histoire Naturelle.

Choux, P. 1934. Les didiéréacées, xérophytes de Madagascar. Mém. Acad. Mal-
gache 18:1-71.

Dickerson, R. E. 1936. Distribution of life in the Philippines. Manila, Bureau of
Printing.

Eisenberg, J. F., and E. Gould. 1970. The tenrecs: A study in mammalian behavior
and evolution. Smithsonian Contrib. Zool. 27:1—137.

Fryer, G. 1956. Biological notes on some cichlid fishes of Lake Nyasa. Rev. Zool. Bot.
Afr. 54:1-7.

Fryer, G. 1959a. Some aspects of evolution in Lake Nyasa. Evolution 13:440-51.

Fryer, G. 1959b. The trophic interrelationships and ecology of some littoral com-
munities of Lake Nyasa with especial reference to the fishes, and a discussion of
the evolution of a group of rock-frequenting Cichlidae. Proc. Zool. Soc. London
132:153-281.

Fryer, G. 1960. Evolution of fishes in Lake Nyasa. Evolution 14:396—400.

Gérard, F. 1919. Etude systématique, morphologique et anatomique des Chlaé-
nacées. Thesis, Faculté des Sciences de Paris (printed by J. Vin, Marseilles).

Greenwood, P. H. 1964. Explosive evolution in African lakes. Proc. Roy. Inst. Gr.
Brit. 40:256-69.

Herre, A. W. 1924. The Philippine Cyprinidae. Philippine Jour. Sci. 24:249-307.

Herre, A. W. 1926. Two fishes from Lake Lanao. Philippine Jour. Sci. 29:499-502.

Herre, A. W. 1932. Five new Philippine fishes. Copeia (1932):139-42.

Herre, A. W. 1933. The fishes of Lake Lanao: A problem in evolution. Amer. Natur.
67:154-62.

347

Continental Islands, Islandlike Areas, and Fringing Archipelagos

Hill, W. C. O. 1953. Primates: Comparative anatomy and taxonomy, vol. I. Strep-
sirhini: A monograph. Edinburgh, Edinburgh University Press.

Jolly, A. 1966. Lemur behavior. Chicago, University of Chicago Press.

Jordan, D. S. 1905. A guide to the study of fishes. 2 vols. New York, Henry Holt
and Co.

Keast, A. 1971. Adaptive evolution and shifts in niche occupation in island birds.
In Adaptive aspects of insular evolution, ed. W. L. Stern, pp. 39-53. Pullman,
Wash., Washington State University Press.

Kohn, A. J. 1968. Microhabitats, abundance and food of Conus on atoll reefs of the
Maldive and Chagos Islands. Ecology 49:1046-62.

Kohn, A. J. 1971. Diversity, utilization of resources, and adaptive radiation in
shallow-water marine invertebrates of tropical oceanic islands. Limnol. Oceanog.
16:332-48.

Koryakov, E. A. 1955. On the fertility and type of the spawning population of
Comephorus. Dokl. Akad. Nauk. S.S.S.R. 101:965-67 (in Russian).

Kozhov, M. M. 1951. On the morphology and history of the Baikalian endemic
molluscs of the family Baicaliidae. Tr. Baik. Limnol. Stud. Akad. Nauk. S.S.S.R.
13:93-119 (in Russian).

Kozhov, M. M. 1963. Lake Baikal and its life. The Hague, Dr. W. Junk, Publishers.

Lack, D. 1947. Darwin’s finches. Cambridge, Cambridge University Press.

Lack, D. 1971. Island birds. In Adaptive aspects of insular evolution, ed. W. L.
Stern, pp. 29-31. Pullman, Wash., Washington State University Press.

Lowe, R. H. 1953. Notes on the ecology and evolution of Nyasa fishes of the genus
Tilapia, with a description of T. saka Lowe. Bull. Brit. Mus. (Natur. Hist.) Zool.
2:349-68.

Lowe, R. H. 1959. Breeding behavior patterns and ecological differences between
Tilapia species and their significance for evolution in the genus Tilapia (Pisces:
Cichlidae). Proc. Zool. Soc. London 132:1-30.

Lowe-McConnell, R. H. 1969. Speciation in tropical freshwater fishes. Biol. Jour.
Linnaean Soc. 1:51-75. (Also as Speciation in tropical environments, ed. R. H.
Lowe-McConnell. London, Academic Press.)

Mertens, R. 1934. Die Insel-Reptilien, ihre Ausbreitung, Variation, und Artbildung.
Zoologica 32(6):1—209.

Milne-Edwards, A., and A. Grandidier. 1878-1885. Histoire naturelle des oiseaux.
In Histoire physique, politique, et naturelle de Madagascar (1 vol. text: 1885;
3 vols. plates: 1878, 1879, 1881). Paris, Imprimerie National.

Myers, G. S. 1960a. The endemic fish fauna of Lake Lanao, and the evolution of
higher taxonomic categories. Evolution 14:323-33.

Myers, G. S. 1960b. Fish evolution in Lake Nyasa. Evolution 14:394-96.

Petter, J. J. 1962. Recherches sur l’ecologie et l’ethologie des lémuriens malgaches.
Mem. Mus. Nat. Hist. Natur., n.s., ser. A, Zool. 27:1-146.

Petter, J. J. 1969. Speciation in Madagascan lemurs. Biol. Jour. Linnaean Soc.
1:77—84. (Also as Speciation in tropical environments, ed. R. H. Lowe-McCon-
nell, pp. 77-84. London, Academic Press.)

348

Continental Islands, Islandlike Areas, and Fringing Archipelagos

Poll, M. 1946. Revision de la faune icthyologique du Lac Tanganyika. Ann. Mus.
Congo Belge, Zool. (1), 4:145-364.

Rand, A. L. 1936. The distribution and habits of Madagascar birds. Bull. Amer.
Mus. Natur. Hist. 72(5):143-499.

Rauh, W. 1956. Morphologische, entwicklungsgeschichtliche, histogenetische, und
anatomische Untersuchungen an dem Sprossen der Didieraceen. Arch. Akad.
Wiss. Lit. (Mainz) 1956(6):341-444.

Rauh, W. 1961. Weitere Untersuchungen an Didiereaceen. 1 Teil. Sitzungsber.
Heidelberg. Akad. Wiss. (Math.-naturwiss. Kl.) (1960/61) no. 7:182—300.

Rauh, W. 1963. Didiéréacées. Flore Madagascar Comores 121e:1-36.

Rauh, W. 1965. Weitere Untersuchungen an Didiereaceen. 2 Teil. Sitzungsber.
Heidelberg. Akad. Wiss. (Math -naturwiss. Kl.) (1965) no. 3:217-434 (with
supplement by H. Straka: Die Pollenmorphologie der Didiereaceen, pp. 435-43).

Schultz, L. P. 1957. The frogfishes of the family Antennariidae. Proc. U.S. Nat. Mus.
107 (no. 3383):47-105.

Schultz, L. P. 1964. Three new species of frogfishes from the Indian and Pacific
Oceans with notes on other species (family Antennariidae). Proc. U.S. Nat. Mus.
116 (no. 3500):171-82.

Soulé, M. 1966. Trends in the insular radiation of a lizard. Amer. Natur. 100:47-64.

Strasburg, D. N. 1966. Golem cooperae, a new antennarid fish from Fiji. Copeia
(1966) :475-77.

Taliyev, D. N. 1955. The miller’s thumbs (Cottoidei) of Lake Baikal. Tr. Akad.
Nauk. S.S.S.R. 9:603 (in Russian).

Trewavas, E. 1947. An example of mimicry among fishes. Nature 160:120.

349

CHAPTER

INSULAR WOODINESS

Darwin (1859, p. 392) claimed that representatives of predominantly her-
baceous families (particularly dicotyledonous families) tend to be woodier
on islands than their mainland counterparts: “Again, islands often possess
trees or bushes belonging to orders which elsewhere include only herba-
ceous species.” However, Hemsley (1885), in a review of insular botany
derived from observations during the Challenger voyage, says that Darwin’s
statement is “incorrect, as may be seen from the following parallels in insu-
lar and continental floras.” Despite the apparent contradiction, neither
Darwin nor Hemsley is wrong. If we examine certain insular floras, such
as those of the Hawaiian Islands, St. Helena, or Juan Fernandez, we see
remarkably woody representatives of “typically” herbaceous families—
notably Asteraceae in the islands mentioned. We can also see equivalent
arborescence in certain continental areas, as for Hemsley’s (1885) admit-
tedly incomplete listing of arboreal continental Asteraceae. However, what
one notes if one is familiar with geography, geology, and climate of certain
mainland areas is that they exhibit many of the same features strongly
represented on islands, so that similar growth forms may be expected.
The phenomenon of paramount importance, seemingly neglected by those
who have dealt with growth forms in dicotyledonous families, is that rela-
tively few predominantly herbaceous families are incapable of evolutionary
increase in production of secondary xylem. Why are phylogenetic series
always read from woodiest to most herbaceous, with uniformitarian uni-
directionality (e.g., Meusel, 1952)? There is considerable evidence, if we
wish to investigate more deeply, that angiosperms are supremely plastic in
evolving growth forms suited to situations, and that invasion of regions

350

Insular W oodiness

suited to greater woodiness may be expected to result in evolution of that
woodiness. It would be strange indeed if all groups of dicotyledons were
doomed to progressive loss of cambial activity. To be sure, there are some
families in which ability to produce secondary xylem has apparently been
completely lost (Callitrichaceae, Podostemonaceae, Nymphaeaceae), but
many families of dicotyledons seem still to be flexible where degree of
cambial activity is involved.

The purpose of this chapter, then, is twofold: to examine insular rep-
resentatives of increased woodiness as well as noninsular localities in which
similar tendencies occur; and to attempt to find the common denominators
and evidence for this phenomenon. To present only the insular picture
would be unrealistic.

Some workers have regarded the woody insular representatives of pre-
dominantly herbaceous groups as relicts. For example, Meusel (1952) and
Lems (1961) interpreted Macaronesian “woody herbs” in this way, al-
though Lems modified his views in a subsequent paper (Lems and Holzap-
fel, 1968) and in conversations with me prior to his death. Skottsberg
(1956) also held the relict hypothesis. This concept is also accepted by
Werner (1966), who amazingly and erroneously claims all but universal
acceptance for the relict hypothesis. The reverse hypothesis, that rosette
trees, rosette shrubs, and other woody herbs have evolved autochthonously
on islands, has been held by a number of workers, including Schenck
(1907), Rikli (1912), Johnston (1953), and Carlquist (1965, 1970a). The
reasons for this latter hypothesis are as follows:

1. The moderated climate of islands favors evolution of woodier
forms there. This also holds for continental areas that are like islands in
climate and other factors. The nature of these climatic regimes will be
detailed later. There are exceptions in the cases of islands at high latitudes
(which cannot be said to have moderated climates, however), and atolls
and small islands where dryness, chiefly, deters increase in woodiness (and
dry islands also cannot be said to have moderated climates).

2. Rosette trees and other woody forms are present on oceanic islands
all over the world, suggesting independent evolution into this habit.

3. Quite different taxonomic groups represent woody herbs on differ-
ent islands, and, like replications or control populations in experimental

351

Insular Woodiness

work, serve to demonstrate the selective factors involved. Parallel evolution
from herbaceous or less woody ancestors is indicated.

4. There are, to be sure, groups of woody herbs that occur both on
islands and on narrow strips of continental coastal areas with similarly
moderated climate (Coreopsis gigantea in California, for example). How-
ever, if the relict hypothesis were correct, there ought to be more numerous
continental representatives, for volcanic islands are relatively recent and
would be expected to bear few relicts. Meusel (1952) seems to think that
relictualism is indicated because some “‘candelabrum shrubs” or “rosette
shrubs” occur on islands and similar species (usually less woody!) occur
in neighboring areas of similar climate (e.g., the Canary Islands and Medi-
terranean coastal areas). Meusel (1952) and Werner (1966) seem to imply
that “woodier” is always “more primitive,” that progressive placement of
renewal buds closer to or below the ground in successively more northern
European localities is the end product in such a series, and that the fact
that one can construct such a series constitutes a form of evidence. There
seems an obvious fallacy involved in choosing highly specialized island
forms as a starting point. Why not, for instance, choose an intermediate
mainland type as a starting point for both extremes, instead of maintain-
ing unidirectionality? Moreover, Meusel and Werner simply analyze dia-
grams of growth forms (which can, of course, be arranged in any sequence
one wishes) and geographical ranges; their work does not stress evolution-
ary depth in its approach.

5. If the relict hypothesis were correct, the rosette trees and shrubs on
islands ought to represent groups with more numerous primitive features.
However, this is not true: the insular woodier herbs all belong to families
that are highly plastic and evolutionarily upgrade.

6. Insular woodier herbs belong to groups that basically have excellent
dispersal mechanisms and would be expected to reach islands easily. There
is also a strong link between good dispersibility, ability to establish in pio-
neer conditions, and ability to evolve a variety of forms because of capabili-
ties for ecological polymorphism. Such a family as Asteraceae illustrates
this link well.

7. Although only the most striking forms, such as rosette trees, are
usually mentioned, the groups that have become woody on islands show
various degrees of woodiness; larger plants are often characteristic of rain

352

Insular Woodiness

forests. However, because various degrees of woodiness occur on islands,
we would have to say that the most nearly herbaceous were reductions if
woodiness were a relictual feature on islands. This seems awkward, be-
cause we would have to explain why herbaceous groups evolve into woody
forms on continental areas, then the woody forms disperse to islands, and
some of these undergo progressive evolution back to more herbaceous
forms again. The easier hypothesis is that herbaceous types were the ones
dispersed to islands, and that they have become progressively woodier as
they undergo ecological shift (notably into wet forests). Groups that do
show a gamut in woodiness on islands include Echium and Sonchus
(Macaronesia), lobelioids, and various groups of Asteraceae. Even one ex-
ample is illustrative: Euphorbia piscatoria is a low shrub in coastal locali-
ties on Madeira but is a tree (presumably a genetically different race) in
laurel forests on that island. These gamuts seem best explained as aspects
of adaptive radiation. The tendency for adaptive radiation to occur on
islands such as the Hawaiian Islands has never been questioned—but pro-
ponents of the relictual hypothesis seem to overlook this.

8. Woody herbs on islands tend to retain some of the pioneering
qualities we would expect of herbaceous, weedy groups. This observation
was made by Skottsberg (1953) in the case of Juan Fernandez rosette trees,
and is true elsewhere, as in Macaronesia, where wet-forest species of
Echium and Sonchus invade openings in the forest. In the Hawaiian
Islands, Lobelia hypoleuca, a shrubby species, appeared in abundance on
the bare areas created by construction of the Mt. Kaala road. It is true
that some wet-forest woodier herbs on islands lack this ability, but one
would expect a gradual loss of weediness during evolution into the wet
forest.

9. If the relict hypothesis were true, one would expect a higher pro-
portion of woody herbs on old continental islands or old oceanic islands
(e.g., New Guinea, New Caledonia). In fact, this is not true: the woody
herbs occur chiefly on recent high islands with strongly disharmonic floras.

Why Does Secondary Woodiness Occur?

The reasons for secondary woodiness depend on the geographical location
—otherwise we would expect woody representatives of predominantly her-

SB)

Insular Woodiness

baceous groups to be distributed uniformly throughout the world, as they
clearly are not. The precise proportion of factors cited below whereby
woodiness may occur cannot, unfortunately, be estimated. I doubt that
only one of these factors is operative in any particular case, and I will
attempt to elucidate, by means of a few examples that seem relatively ex-
plicable, which factors have operated to which extent.

DISHARMONY

In the case of oceanic islands remote from source areas, tree species
are less likely to arrive by natural means than are herbaceous species.
De Candolle and Darwin noted the precinctiveness to continental areas
of many trees. “Now trees . . . generally have, whatever the cause may be,
confined ranges. Hence trees would be little likely to reach distant oceanic
islands” (Darwin, 1859). The “cause” which mystified Darwin involves
poorer dispersibility of tree seeds as compared to those of herbs. Large
seed size characterizes many continental forest trees, and immediately
rules out many species (e.g., conifers are notably absent on oceanic islands).
Many genera and families of large-seeded angiosperms have no repre-
sentatives on oceanic islands. In addition, we must not underestimate the
potent but imponderable factor of lessened viability of seeds of wet-forest
trees. ‘That many such species have short viability is well known, and is
undoubtedly related to the weaker selective pressure for seed longevity in
wet forest than in dry, open habitats. Large seeds with abundant food
reserves are adapted to seedling development in shaded areas of a forest,
and if the substrate is often wet, longevity to survive long dry periods is
of no selective value. In these respects (as well as, perhaps, adaptation by
forest trees to narrow ranges of tolerance), forest trees are specialized and
would represent an evolutionary dead end except that the continental
areas to which they are adapted are extensive; they suffer no disadvantage,
but are poor for long-distance dispersal.

That seed size and viability are prime factors in absence of trees
from oceanic islands is well illustrated by the exceptions: Metrosideros
has reached virtually all the high islands of the Pacific because its
seeds are small and resistant to cold, desiccation, and even seawater (see
chapter 2). Weinmannia, the chief forest tree on Tahiti, has similar

ane

Insular W oodiness

seeds. These two genera are virtually the only tree species occurring on
reasonably remote oceanic islands for which wind dispersal is possible or
probable.

An obvious series of exceptions to limitations of seed size is formed
by littoral trees brought by oceanic drift (e.g., Barringtonia, Calophyllum),
but these are not components of inland vegetation on islands, with few
exceptions (see chapter 11). Although bird dispersal of tree species with
relatively small, fleshy fruits is quite good (see chapter 2), the nature of
long-distance dispersal definitely limits the number of immigrants when
compared with the diverse contents of a forest on a mainland source area.
Even the bird-dispersed species with fleshy fruits are often not trees but
shrubs of moderately dry forest.

Consequently, we are faced with the irony that high volcanic islands
have habitats ideal for arboreal and arborescent species, but dispersal
brings a quite different spectrum. Long-distance dispersal is best for weedy
herbaceous species. Because the pioneer habitats to which these species
are adapted “open” and “close” suddenly (e.g., slides), such habitats tend
to be discontinuous in distribution (like islands within a continent, in
essence), and are often dry, so that seed longevity is likely to be greater.

CLIMATIC MODERATION

If we accept the concept that a climate characterized by moderation
tends to foster evolution of increased woodiness, we must realize that
“moderated” insular climates are by no means alike. They range from
climates that are literally as nearly uniform throughout a year as can be
possible climatically, to situations that have occasional extremes of heat
or dryness but such effects minimized by maritime influences. I am
defining moderation here in terms of particular factors. The diagrams
from the Klimadiagramm-Weltatlas (Walter, Lieth, and Rehder, 1967)
(fg. 10.1) prove useful in such discussion, and localities cited below
correspond, in many cases, to the diagrams reproduced in that figure.
Climatic data for California are from Dale (1959); for Hawaii, from
Blumenstock (1961).

Temperature. If a relatively low temperature were uniform, one could
expect arborescence to evolve in herbaceous groups. However, areas of
relatively low temperatures are subject to fluctuations, as are any areas,

5)

23.3° 703

HONOLULU (17m)

18.6
13.3

KEANAE VALLEY (300m) 19.9° 6147
[14-20]

16.2° 955

ST. HELENA (610m)
[49]

SOKOTRA 210m Tl
[2]

WAIMEA (820m)
[11-20]

17.2° 1138

SEYMOUR ISLAND (11m) 25,.0° 102
[3]

LA LAGUNA (547m) —-16.5° 594

28.7 NANYUKI (1947m) —-16.3° 679
25.9) [6]

ashy bith
7.5

2.7

TANTALUS (408m)
[14-20]

20.0° 2520

JUAN FERNANDEZ (10m). 15.3° 979

QUEIMADAS (860m) 12.3° 2322
[13]

BOGOTA (2660m) 14.5° 1061

16-49]

=

FIG. 10.1. Climatic diagrams for localities on islands and for islandlike areas on continents.
Each diagram begins with midwinter, left, and ends with midwinter, nght, regardless of
whether in northern or southern hemisphere. The two curves represent monthly means of

(continued)

Insular Woodiness

and so freezing may be expected. Occurrence of temperatures below 0 C
need not rule out occurrence of insular woodiness. One may cite genera
of arborescent Asteraceae—for example, Olearia in New Zealand, ‘Tasmania,
Victoria, and the Chatham Islands, in areas where absolute minima below
QC occur and where mean daily minima for the coldest months are
about 4C. Another climatic regime that fulfills low temperatures but
on a uniform basis is found on equatorial mountains. Arborescence does
occur in these regions, and is discussed in chapter 14. The equatorial
climate of northeastern South America is represented by the diagram for
Bogota in figure 10.1. That of montane equatorial Africa is approximated
by the diagram for Nanyuki (fig. 10.1). However, Nanyuki is on the plateau
from which peaks such as Mt. Kenya rise, and Nanyuki represents the
climatic uniformity of this region, but mean temperatures would be lower
if there were reporting stations at higher elevations. Arborescent species
are less frequent in uniformly cool areas than they are in areas with high
mean temperatures, owing to problems of occasional frosts.

A mean temperature no lower than 10 C (Hawaii: Volcano Observa-
tory, for example) might be expected to be the lower threshold for
arborescent species in most areas. In general, a mean annual temperature
of 15-20 C would be expected to favor arborescence most. For example,
the Hawaiian rain forests fall in this category, as do the Juan Fernandez
Islands and St. Helena (fig. 10.1).

There undoubtedly is an upper temperature threshold for insular
woodiness also. The reason for this limit would appear to be that her-
baceous groups are, in general, those of temperate regions, and are limited

FIG. 10.1 (continued)

temperatures (in accordance with markings on left side of diagram; each mark = 10 C) and
monthly means of rainfall (markings at right; mark = 20 mm rainfall). Where rainfall exceeds
100 mm, curve is drawn at truncated scale (1:10) and the portion above 100 mm is in black.
Where rainfall curve is higher on diagram than temperature curve, space between the curves is
filled by vertical lining (or black, if rainfall above 100 mm) whereas if rainfall curve drops
below temperature curve, that space is filled by dotted pattern. At top left of each diagram
is name of station and altitude. In brackets below name are numbers of years on which rec-
ords are based—the first figure for temperature, the second for rainfall. At lower left are
figures for temperature minima: the upper is mean daily minimum of the coldest month, the
lower is absolute minimum recorded. Figures at upper left, if specified, are for maxima. At
upper right are figures for mean annual temperature in degrees centigrade and for mean an-
nuai sum-total of precipitation in mm. These diagrams are selected from the Klimadiagramm-
Weltatlas, by H. Walter, H. Lieth, and H. Rehder, 1967 (parts issued 1960-1967), © Verlag
Fischer, Jena. Reproduced by permission of the publisher.

pil)

Insular Woodiness

in ability to adapt to hot tropical areas. ‘The upper threshold would appear
to occur in areas with a mean annual temperature of about 23-25 C (sea
level on Fiji, Hawaii, and New Caledonia). Arborescent lobelioids (near
Hilo, Hawaii) or Scaevola of the Goodeniaceae (Fiji, New Caledonia)
occur in such localities. The arborescent lobelioids and Scaevolas are most
abundant and most arborescent in areas with mean annual temperatures
of about 20 C (Tantalus, Oahu; Keanae Valley, Maui; see fig. 10.1).

There is no doubt that proximity to the temperature-insulating effect
of the ocean moderates climate considerably, and absence of the extremes,
in effect, produces a year-long growing season that seems desirable for
arborescence in herbaceous groups. Proximity to the coast is, one must
remember, a feature of all but the largest islands. The effect of coastal
versus inland localities is shown by the following figures from California.
Avalon, on offshore Santa Catalina Island, is at sea level; Claremont, about
50 mi. inland from the coast, has an elevation of approximately 1200 ft.

Mean Annual Mean July Mean January

Locality Temperature Temperature Temperature
Avalon 60.3 F 66.0 F 53.9F
Claremont 62.1 F TAs 507 as

These means are not as important as the extremes, even though differences
are evident. There is an average of only 2 days per year with temperatures
of 32 F (0 C) at San Diego, Calif., on the coast. At Claremont, there is
an average of 28 days per year on which temperatures fall to 32 F, 8 days
per year to 28 F, and 5 days per year with minima of 24 F. Such tempera-
tures are unfavorable for many insular species (see below and chapter 5).
There are several localities in coastal southern California where frost has
never been reported.

Rainfall. With respect to total rainfall, one can say that precipitation
of about 1000 mm (ca. 40 in.) per year or more favors arborescence. This
figure is almost exactly that reported for sea-level stations on the Juan
Fernandez Islands and St. Helena (fig. 10.1) and stations somewhat above
sea level in the Hawaiian Islands (compare Honolulu, Tantalus, and
Keanae in fig. 10.1). Rainfall below 1000 mm per year probably does not
permit arborescence to any great extent, unless we take into account stem

358

Insular Woodiness

succulents. For example, there are arborescent plants, but all stem succu-
lents, on the lowland Galapagos and Sokotra (Socotra); see figure 10.1.

There is perhaps no upper limit of rainfall above which arborescence
is no longer favored (unless bog situations are formed). Many arborescent
species occur in Hawaiian areas with rainfall above 2500 mm (100 in.)
per year.

Uniformity of rainfall throughout the year is important. The occur-
rence of a strongly marked dry season vitiates the effect of high rainfall.
Exceptions occur only where humidity is great, where cloud cover and
moisture condensed on vegetation from clouds rather than by direct rainfall
is available, or where deep, shady localities (valley bottoms) remain moist
during dry seasons. Cooler temperatures during drier parts of the year
would reduce evaporation and conserve available moisture. However, this
condition is rare, and could be said to exist only on St. Helena (fig. 10.1),
where rainfall is, however, not extremely low during any period of the
year.

Unfortunately, rainfall recording stations on islands are most fre-
quently near sea level. One can only guess at the increase in moisture
with altitude, but obviously there is an appreciable effect. In the Hawaiian
Islands, there are enough stations so that this effect can be seen. For
example (fig. 10.1), Honolulu at sea level has an annual rainfall of 703 mm,
a marked dry period during the summer, and an average temperature
of 23.3 C. Tantalus, a station at 408 m only a few miles inland, has an
annual rainfall of 2520 mm, no marked dry period during the summer,
and an average temperature of 20.0 C. On Tantalus, arborescent species
of Clermontia and Cyanea (Campanulaceae, Lobelioideae), Scaevola
(Goodeniaceae) and Dubautia occur, whereas none of these (except for
the semisucculent shrubby species of beach Scaevola, S. taccada) occurs
at Honolulu or similar Hawauan localities.

Similarly, upland sites in the Canary Islands must be considerably
more mesic than lowland sites. We can suspect that the cloud forest on
Tenerife has a climate like that of upland Madeira (Queimada, fig. 10.1),
whereas lowland sites on both islands have a typical mediterranean climate,
with warm, dry summers and cool, moist winters. At sea level, there is
undoubtedly considerable moderation in temperature and humidity com-
pared with inland continental localities in the same latitude. At altitudes

359

Insular Woodiness

above the cloud forest on the Canary Islands, rainfall drops off, as it does
at higher elevations in the Hawauan Islands. The increase in rainfall in
cloud forest in the Canary Islands and Madeira is not nearly so great as
it is in the Hawaiian Islands, and the Macaronesian cloud forests are just
within the range of tolerance for arborescence of herbaceous groups.
However, shrubby representatives of herbaceous groups are favored in drier
localities, both lowland and alpine, in the Canary Islands, Madeira, and
the Hawaiian Islands, and stem succulents (Euphorbia in the Canary Is.;
Reynoldsia in the Hawaiian Is.) are also suited to such regimes. The
climate of Honolulu differs little from the oceanic mediterranean pattern:
only the slightly higher mean temperature distinguishes it from, say, the
climate of La Laguna (fig. 10.1) on Tenerife, or Santa Cruz (La Palma I.)
in the Canary Islands. The rainfall in the Canaries is not nearly as heavy
as in Hawaii, and so Macaronesia could be said to have at most a cloud
forest, whereas the Hawaiian Islands have true rain forest.

The most striking example of how cloud forest has barely permitted
evolution of arborescence is to be found in the Galapagos Islands. The
only recording station is at sea level (Seymour I., fig. 10.1). The climatic
profile of this station is very similar to that of Villa Cisneros, on the
coastal Sahara south of the Canary Islands. The mean temperature at
Seymour Island is somewhat greater than at Villa Cisneros, which is farther
from the equator, but otherwise the two are comparable. However,
Scalesia pedunculata and a few other species of Scalesia form trees in
upland situations in the Galapagos. At such stations, there is a cloud-forest
condition. The helianthoid Asteraceae as a whole seem to have a capability
for woodiness in xeromorphic conditions (i.e., shrubby species of Scalesia
occur at sea level in the Galapagos), so a moderate increase in moisture
availability in the uplands has permitted arborescence. Woody lobelioids
(which seem to require greater moisture availability) are common in
Andean South America but are absent in the Galapagos. The climate of
Sokotra (fig. 10.1) is roughly comparable to that of Galapagos sea-level
localities.

Humidity. The relatively small diameter of most oceanic islands
permits the higher humidity produced by ocean surface to affect the
entire surface of the islands. On continental areas, the reverse is true;
thermal patterns alter to a considerable extent the incoming weather

360

Insular W oodiness

patterns, and offshore winds created by land masses minimize humidity
even at coasts. The effect of humidity on islands is twofold: it acts as a
form of insulation, preventing wide fluctuations in temperature, and it
attenuates considerably the effect of moisture availability to a plant,
because transpiration is reduced and evaporation from soil is lessened. Of
course, if an inversion layer is present, as it is in the Hawaiian Islands
(Blumenstock, 1961), and suitable updrafts occur, humid air can condense
into precipitation. Higher humidity, then, is a prime factor in moderating
insular climates.

Unfortunately, data on relative humidity are gathered less frequently
than are data on temperature, or in a fashion where one station’s methods
of reporting are not comparable to those of another. One can measure
humidity at a specified hour, in which case the average humidity is not
known. Or if humidity is measured five times a day, the average may well
be quite different from when it is measured twice. That humidity is high
on insular areas can be suggested by the fact that the average humidity
(average for an entire year) at Hilo, Hawai, is 81 percent. At Orotava,
Canary Is., it is 75 percent; at Lihue, Kauai, Hawaiian Is., it is 74 percent.
While coastal areas of continents are comparable, humidity drops off
sharply inland. Los Angeles Airport, on the coast, has an average humidity
of 71 percent (with probably, however, greater fluctuations during days
and during the year than on Hawaiian coasts), but the Civic Center of
Los Angeles registers an average of 61 percent. In Los Angeles, a low of
20 percent may occur commonly during the dry season, and lows of 5
percent occur occasionally. In contrast, humidity rarely falls below 40
percent even in dry lowland localities of the Hawaiian Islands.

ECOLOGICAL SHIFT

Many if not most instances of arborescence seem related to a change
in ecological preference. As a generalization, many cases can be said to
involve change from open habitats characteristic of weedy plants to forest
habitats, particularly wet forest. In a few cases, the woody herbs even
become emergents (e.g., Brachylaena in Madagascar, Brachyglottis in New
Zealand). More commonly, however, the woodier herbs are understory
elements (e.g., Hawaiian lobelioids), occupants of more open areas within
a moist forest (Juan Fernandez Asteraceae), or shrubby elements in a scrub

361

Insular W oodiness

area (shrubby Macaronesian species of Brassicaceae, Boraginaceae). In a
few cases, stem succulence has permitted formation of a shrubby habit in
a group that otherwise would not tolerate dry lowland conditions (Sonchus
leptocephalus on the Canary Islands).

What must be stressed strongly, however, is that species with increased
woodiness are often part of a spectrum of adaptive radiation, and cannot
be considered in isolation. Thus, in the listings later in this chapter, of
genera “woodier on islands,” some genera contain nonwoody as well as
markedly woody species (e.g., Sonchus, Cyrtandra). For example, most
species of Argyroxiphium (Asteraceae) are acaulescent, but some species
of the related genus Dubautia are trees. Ecological shift is an integral
component of adaptive radiation (chapter 3).

Basic to the process of ecological shift is that an island offers some
sites comparable to those of mainland ancestors but also other habitats.
If an island did not offer these differing zones, ecological shift would not
occur, but if zones like those of continental areas did not exist as well,
an immigrant species could not establish. The fact that weeds have
established readily on oceanic islands illustrates that congruent zones do
exist.

ABSENCE OF LARGE HERBIVORES

Another explanation for arborescence of woodier herbs on islands is
the absence there of large vertebrate herbivores (herbivorous birds are not
considered to play a significant role). The reverse might obtain where
herbivore pressure is so heavy that only annual grasses and annual dicoty-
ledons can complete their life cycle. This situation does exist on Guadalupe
Island, Mexico, where herbivores were absent in prehuman times; now
goats prevent reproduction of trees and shrubs except in sites inaccessible
to them: steep cliffs and offshore islets (see Carlquist [1965] for a review
of the Guadalupe situation). However, on the Canary Islands, where goats
are plentiful and denude the islands to an appreciable extent, few species
of woody herbs have been extinguished—although this may be expected
with time. Introduced herbivores have had a serious effect on the Hawaiian
Islands, and it is impossible to imagine evolution of the present species
composition and growth forms had these herbivores been present as early,
prehuman immigrants. Recording of future extinctions in areas invaded by

362

Insular Woodiness

pigs, goats, sheep, deer, and cattle in the Hawauan Islands will give us a
fairly precise estimate of how significant is the absence of large herbivores
for evolution of woody herbs on oceanic islands.

“RELEASE” FROM SEASONALITY

Although climatic moderation is obviously instrumental to release
from seasonality, the response of any herbaceous immigrant to an island
in which moderation prevails will take the form of release from events
in the growing pattern that, on continental areas, prevent arborescence.
Release from winter cold would permit positioning of innovations (renewal
buds) above the surface of the ground. It would also involve continual
growth, rather than cessation during the winter (which, in herbaceous
plants of cold regions, involves overwintering either in the form of seeds
or as renewal buds at or below ground level). Moreover, if an insular climate
is more mesic, not only can growth continue—instead of being interrupted
by seasonal dryness—it can thereby form a tree. Percentages of trees in
relationship to other growth forms (shrubs, annuals) are low in floras of
areas with mediterranean climate. This can be simply correlated with the
fact that summer drought prevents a root system from being able to support
an appreciable leaf surface and growth throughout the summer; the plant
must either die or cease growth until sufficient ground water becomes
available again. Release from any of these events that limit or terminate
growth of an herb will tend to favor arborescence. Presumably, adjustment
to the new situation will involve genetically fixed adaptation to a less
seasonal environment, with all that this connotes (increased production
of wood, for example).

Does disappearance of thermoperiodic responses accompany increased
woodiness or adaptation to insular climates where day-length differences
and seasonal fluctuations in temperature are less marked? In the Hawaiian
Islands, well-defined flowering seasons occur in some plants but not in
others. The most marked retention of flowering seasonality may be in dry
lowlands and dry alpine localities, but even in rain forests, seasonality in
flowering can be observed (although to what degree and in what groups
is a subject appropriate for further studies). Photoperiodicity has, likewise,
not been studied in Hawaiian native plants to any appreciable degree.
One relevant study, however, is that of Ray and Alexander (1966) on the

363

Insular Woodiness

widespread weed Xanthium strumarium (Asteraceae), a plant noted for
photoperiodicity. They found clinal variation that was proportional to
latitude within North American populations. Strains from Hawaii, how-
ever, showed notably tardy short-day response, with a critical night length
of about 11 hr. This suggests that photoperiodicity might be in the process
of disappearing there. Separating periodicity in response to rainfall, day
length, temperature fluctuations, and light availability from each other in
insular plants may not be easy, but these problems very much deserve
study.

Some woody insular plants do survive frost. This is well known in
the Afroalpine Senecios and Lobelias, and probably occurs in Espeletia
of the paramos of Colombia as well (see chapter 14). However, the shrubby
Silene struthioloides (Caryophyllaceae: alpine Maui) and Plantago webbu
(Plantaginaceae: alpine Tenerife) seem to protect vegetative buds by means
of the upwardly appressed leaf position at stem tips.

OTHER POSSIBLE FACTORS
Explanations for insular woodiness by Darwin and Wallace seem
dubious. Darwin (1859), on the basis of the Galapagos genus Scalesia,
suggested that competition among individuals within a species has favored
taller individuals. This clearly could not be true for lowland Scalesias, in
which individuals are shrubby and widely spaced. Scalesia pedunculata
and S. microcephala are arborescent and can form dense stands in their
native upland regions. However, these Scalesias provide one of the few
examples of an insular woody representative of an herbaceous family that
forms stands. None of the Hawaiian woody herbs do this—they are
scattered throughout a forest community composed of woody species.
Even in this latter case, one may doubt that competition for light is a
selective factor. The more logical explanation of their degree of arbores-
cence is that they grow in areas moist enough to support trees—admittedly,
Scalesia can form a tree in moderately dry cloud forest whereas other
genera (e.g., Cyrtandra) would not even survive there. However, the best
explanation seems to be that degree of arborescence is directly proportional
to available moisture. The only exceptions on islands would seem to be
those species that can form arborescent stem succulents, Opuntia and

364

Insular W oodiness

Euphorbia, for example. However, continental species of these genera are
at least as arborescent as insular species and occur in open country in any
case, not in closed stands.

Wallace (1878) claimed that woodiness of island plants connotes
greater longevity, which permits repeated flowering and thereby a greater
chance for cross-pollination in situations where paucity of pollinating
insects would make cross-pollination of more ephemeral plant species less
likely. He assumes that paucity of insular pollinating insects is a bottleneck.
This assumption is dubious: plants may self-pollinate (see chapter 13), at
least in a number of cases. Moreover, if canditions are suitable for an
insect species with pollinating capabilities, population density of the insect
should, if anything, be greater than in a mainland species (assuming the
nature of insect faunas on islands is disharmonic), or else a suitable insect
would be expected to be entirely absent—but rarely the intermediate
condition Wallace predicates. If Wallace’s assumption were true, more
ephemeral species would be expected to evolve on islands as suitable insects
became available, but ephemeral species are the exception on oceanic
islands, as seen later in this chapter. Climatic factors, again, seem to be
overriding in evolution of growth forms on islands. Less pronounced
seasonality in flowering does occur in some insular plants, but this extension
of flowering time appears to be the result of adjustment to less seasonal
insular climates, not a factor for establishment of groups.

Measures of Selection and Adaptation
TRANSPLANT STATIONS

Thus far, no ngorously chosen and monitored transplant stations,
such as those employed by the Carnegie Institute of Washington in the
various studies by Clausen, Keck, and Hiesey, have been used for the
study of insular plants and their tolerances. Cultivation of plants of
Echium (Boraginaceae) from Macaronesia was described in chapter 5.
Some insular species have been cultivated in the following California
localities, however, and even though the results are crude, they may suggest
the value of carefully performed studies:

1. San Francisco (Golden Gate Park, ca. 2 mi. inland from coast).
Both Wilkesia (Asteraceae: from Kauai, 2000-3000 ft) and Lobelia

365

Insular Woodiness

gibberoa (Campanulaceae: Mt. Kenya, ca. 9000 ft) have been cultivated
to maturity. Wilkesia gymnoxiphium seems to have flowered prematurely
(after only 3-4 years, as opposed to about 10 years or more in the wild),
perhaps because of greater sunlight in open conditions or more continual
availability of water. Echium pininana (Boraginaceae: Canary Is.) per-
formed similarly in this respect; Echium candicans (Madeira) also grows
successfully here. The climate of San Francisco is probably nearly equiva-
lent to that of the Macaronesian laurel forest.

2. Santa Barbara (garden of my parents, Mr. and Mrs. R. M. Bauer,
ca. 2 mi. inland). A wide variety of Hawaiian forest species has been
grown successfully to maturity here, including Tetraplasandra waimeae
(Araliaceae), Wilkesia gymnoxiphium (Asteraceae), Bidens waianaeensis
(Asteraceae), Scaevola gaudichaudiana (Goodeniaceae), and Pisonia um-
bellifera. Species grown successfully but not yet to flowering include
Osmanthus sandwichensis (Oleaceae) and Mezoneurum kauatense (Faba-
ceae). Scaevola floribunda (Fiji) grows well but has not flowered after 10
years of cultivation, and the same is true of Meryta lanceolata (Araliaceae:
Bora Bora). Plants from Guadalupe I., Mexico, grown to flowering include
Talinum guadalupense (Portulacaceae), shown in figure 10.15, top left
and right, and Lavatera lindsayi (Malvaceae). Stephanomeria blairii (Aster-
aceae) from San Clemente I., Calif., also succeeds well here. Macaronesian
species that have been grown with total success include Echium pininana
(Boraginaceae, fig. 5.2, bottom right), E. simplex, E. candicans, Plantago
arborescens, Laurus indica (Lauraceae), and Centaurea arbutoides (Astera-
ceae). Dendriopoterium menendezii (Rosaceae: Gran Canaria) has not
been able to reach maturity. One Juan Fernandez woody herb, Dendroseris
macrophylla, has been cultivated successfully at the nearby University of
California, Santa Barbara campus.

3. Claremont (my residence; also at Rancho Santa Ana Botanic
Garden). Californian insular species do reasonably well here, some excel-
lently. Stephanomeria blairii (fig. 10.3, top left, mght) succeeds well,
although it loses its leaves during the hottest summer days. Among the
Guadalupe Island woody herbs, Talinum guadalupense and Lavatera lind-
sayi do not reach maturity. Hawaiian species are notably unsuccessful:
Wilkesia gymnoxiphium readily succumbs to heat and low humidity.
Macaronesian plants can be grown, but results are different from those at

366

Insular W oodiness

Santa Barbara. Dendriopoterium menendezii has grown to be larger than
plants I have seen on Gran Canaria; apparently its success in Claremont 1s
related to the fact that it is from highlands of Gran Canaria, where the
climate more closely resembles that of Claremont than that of Santa
Barbara. Echium pininana has grown to maturity well and flowered.
Echium candicans is successful, but for a much shorter period of time
than in coastal localities, where its longevity is appreciably greater. Plantago
arborescens has grown readily in my garden during the winter and even
naturalized, but plants have died in the hot, dry summer; it is therefore
a facultative annual in Claremont, but grows as a perennial in Santa
Barbara, just as it does in Macaronesia. I have managed to keep alive a
plant of Dendroseris macrophylla (fig. 10.2, bottom right) in Claremont,
although each summer it loses all or nearly all its leaves for the hottest
months; it is distinctly unhealthy compared with plants at Santa Barbara.

The preceding accounts do not include truly woody shrubs and trees.
Many such insular species cultivate well in California, although even in
these (with a few exceptions, such as Pinus canariensis, from relatively
high elevations in the Canary Islands), insular plants from Macaronesia
and New Zealand succeed better in coastal localities.

GROWTH-FORM SPECTRA AS CRITERIA OF NATURAL SELECTION

The plant species introduced by man to islands may or may not
establish themselves and become weeds and escape into wild situations.
If they do, one may assume that they simulate prehuman establishments
in all respects except that they lack the mechanisms suited to natural
dispersal to islands. One might claim that introduced species tend to be
therophytes (annuals and biennials) mostly, since we tend to equate
annuals with agricultural and horticultural weeds. However, large numbers
of other growth forms have also been introduced by man to islands. In fact,
the huge number of ornamentals other than annuals introduced to the
Hawaiian Islands would suggest that shrubs and trees may, in some cases,
have even been introduced preferentially.

However, we can compare human introductions to the “normal”
spectrum of life forms (namely, angiosperms as a whole), and use this
figure as a “control” for both the naturalized introductions and the native
floras. Such comparisons were used by Skottsberg (1953), based on the

367

FIG. 10.2. Rosette herbs on islands, showing trunk formation owing to noncessation of growth.
TOP LEFT, cabbage (Brassica oleracea); near Orotava, Tenenfe, Canary Is. rop r1GHT, Tauschia
(Melanoselinum) decipiens (Apiaceae), with inflorescence; plant about 1.7 m tall; between
Ribeiro Frio and Funchal, Madeira. Bsorrom LEFT, Geranium maderense (Geraniaceae), plant
about 0.3 m tall; cultivated, Strybing Arboretum, Calif. sorrom ricur, Dendroseris macro-
phylla (Asteraceae), plant about 0.4 m tall, 4 yr. old; cultivated, Claremont, Calif.

368

Insular W oodiness

as
pa tie
Fic. 10.3. Plants from Califorma and its insular areas. TOP LEFT, RIGHT, Stephanomeria
(Munzothamnus) blairti (Asteraceae), an endemic of San Clemente I. Trop LEFT, plant flower-
ing at end of first season of growth. ToP RIGHT, same plant 1 yr. 9 mo. later, just prior to its
third flowering. Bortom LEFT, Lotus scoparius var. scoparius (Fabaceae), ca. 0.4 m_ tall,
removed from ground to show branching from base, sparse and small leaves; Claremont.
BOTTOM RIGHT, Lotus scoparius var. dendroideus, plant ca. 1.2 m tall; note trunk and density
of crown; near Avalon, Santa Catalina I.

concepts of Raunkiaer (1934). I have modified the categories to follow
Lems (1960). I have, however, recalculated Lems’s data on the Canarian
flora in order that native species and introduced, naturalized (adventive)
species could be separated. I have similarly calculated data for the Hawaiian

369

‘Jugjnoons wiojs = ¢§ ‘eur = J ‘aydydida = | ‘([eruUa
iq Jo [enuue) ay4ydo10y} = J, ‘(posiourqns spnq [eMouo1) o}4ydorpAy = FH *([PA9] [LOS Mopaq spnq [eMoud1) 9}A4ydoo3
= 5 {(JaAo] [10s Je spnq [eMouor) oyAydoydArornmoy = FY ‘(UW ¢°Q PUL [9Ad| [IOS UddM\Joq spnq [eMouI) oyAydoruLYyD = D

(wt 7—-¢'9 ‘qniys) oJ4ydoraurydouru = Qy ‘(WI Z 10A0 993}) a}4ydordULYydoIokUL = JY :SMOT[OF Sv d1e PosN so110B9}vd IL], v
(191) (eee ©) == | (2h) () (Gin) a) G0) (1) (6)
Gall Lise ie a4 ST €°6 197 90 89 9°S DATJUDApL
(ZL+1) (1) (be) é (91) =) (sz) (zoos) (91) (¢19) (192)
T0 SE é all €°0 Lal Tre Del OIF LEN dAT}LU
‘ST ULITEMPE ET
(L8) = = (65) se) = (9G) eG) (¢) (Z)
== ie Te 6:09 [+ = 6°62 5G ve £7 DATUIAPL
(9FT) = re (DQ) (G) (Z) rn 2.) en (7) (Se)
— — L0 $¢ VI ao 0°77 €°SZ O'rZ (ENSUE, OAT CU
‘ST ZIPULUTOI urn[
(09) (8) (<) = v (ozy) ) (Se (ZODe (OZ) AD) (8)
Se 50 =e +69 L‘0 67 tI Cae 87 am | OAT} UDApe
(978) (@) = NG = rei) (G8) iS) ei) ee) Wea
Tl SI as mal Ol 09 6+rl £97 797 S°9 oAT}LU
“sy Arvued
Z I ¢ €] Z + 97 6 ST 4 (p[z0a)
_uinsiqoads [PULION,,
saiaads S AL, a Alb HH ) H 2) N W
1770.1,

p Saga} UL1of-afr'T
(sasayjuored ut siaquinu sor1oods [enjoe ‘saseyuso10d ur possaidxa)
svloy} ul sumtodsoisue jo vij00ds WI0J-OV'T
‘LOL S1avL

Insular W oodiness

flora. For native species, I have used a modification of Fosberg’s data in
Zimmerman (1948). Figures for the adventive Hawaiian flora are derived
from Hillebrand (1888) with a few additions; they are definitely incom-
plete, but no current listing of adventive Hawaiian species is available—
and the number is continually increasing. The other data in table 10.1 are
from Skottsberg (1953), and the “normal spectrum” is from Raunkiaer
Cho):

Skottsberg (1953) believes that the life-form spectrum of the Juan
Fernandez flora ought to resemble that of the Canary Islands (or Madeira,
which has a spectrum very similar to that of the Canary Islands). When
Skottsberg adds the adventive Juan Fernandez flora to the native, the
combined figures, expressed in percentages, resemble the Macaronesian
spectrum more closely than does the native flora alone, and he believes the
spectrum based on native + adventive species is a more accurate reflection
of the Juan Fernandez climate than the native flora presents. The higher
percentages of trees, rosette trees, and shrubs in the native as opposed to
the adventive flora Skottsberg attributes to the “anomaly” of persistence
of relict elements.

Contrary to Skottsberg, I interpret the native floristic spectrum of
the Juan Fernandez Islands as an accurate indication of the climate—or,
if one prefers, the ecological niches provided by the areas of climatic zones
in these islands. Surely time has been sufficient on these and the other
islands of table 10.1 for evolution into growth forms precisely suited to
localities of the native species. At sea level, the rainfall of the Juan
Fernandez Islands is at least twice that of sea-level localities on the Canary
Islands, and clearly the Fernandezian rain forest must be much wetter than
the Canarian cloud forest.

If we calculate the percentage of phanerophytes (renewal buds above
0.3 m: trees and shrubs), the native Hawaiian flora is obviously highest,
61.7 percent; the native Fernandezian flora second (48.2 percent); and
the Canarian native flora third (35.6 percent). This is precisely what one
would expect on the basis of rainfall. I have estimated trees in the Hawai-
ian flora somewhat conservatively as compared with Skottsberg’s estimate
of what constitutes a tree (macrophanerophyte) in the Juan Fernandez
Islands; a more liberal estimate in the Hawaiian flora might well make
the figure for trees the same as or greater than the Juan Fernandez figure.

BI

Insular Woodiness

In any case, the 261 species of trees in the Hawaiian flora are many more
than the 34 in the Juan Fernandez flora. The broad lowlands, the bogs,
and the alpine zone of the Hawaiian Islands favor hemicryptophytes or
cryptophytes; if Hawaiian topography matched that of the Juan Fernandez
Islands in narrowness of lowlands and lack of alpine and bog localities,
then these categories would be much lower in the Hawaiian flora, and per-
centages for phanerophytes relatively higher.

In any event, both the Hawaiian Islands and the Juan Fernandez
Islands have a higher proportion of phanerophytes than the world flora
taken as a whole (46 percent).

The most important feature to me, however, is that the spectra of
growth forms in the Hawaiian, Fernandezian and Canarian native floras
represent not relictual features but adaptive modes (and therefore, in-
directly, optima of natural selection) precisely or nearly suited to the cli-
matic regions of these respective archipelagos. To believe otherwise would
be to believe that angiosperms can occur in localities to which they are
not adapted.

The spectra of life forms for adventive floras in these archipelagos are
much alike—but markedly different from those of the native floras. The
adventive floras are rich primarily in therophytes, secondarily in hemicryp-
tophytes. These, presumably, grow mostly in disturbed and lowland sites.
However, even if the life-form spectra of the adventive floras are not pre-
cisely like those of the original prehuman plant immigrants to these archi-
pelagos, they undoubtedly show some resemblance to the makeup of ances-
tral stocks. There is no reason to believe that the present native spectra
are precisely the same as the immigrant spectra. For example, the adven-
tive trees in the Canary Islands include Castanea, Quercus spp., and
Prunus spp., and these are types very unlikely to have arrived via long-dis-
tance dispersal. We would not even expect the “normal” or world spec-
trum to simulate immigrants to these islands, because of the poor dispersi-
bility of trees—yet trees have “exceeded expectations” based on the world
flora as a whole.

At least some portion of the discrepancies between adventive floras
and native floras on these islands can be considered expressions of selective
pressures. Thus, in all three floras, there is selective pressure favoring trees
and shrubs and (to a lesser extent) chamaephytes—at the expense chiefly

372

Insular Woodiness

of therophytes. In all three floras, the percentage of native shrubs is at
least six times the percentage of adventive shrubs, the percentage of native
trees at least three times the percentage of adventive trees.

Although the concepts suggested by table 10.1 must of necessity be
somewhat vague, there does seem to be a selective pressure toward trees
and shrubs, and an indication that these are not what establish most
easily at first. Thus, we can expect evolutionary shift into greater woodi-
ness—a shift that also corresponds to shift into wetter forest. Conversion
from herbs into shrubs is relatively easy, and may help to explain why
shrubs bulk large in the proportions of growth forms on these islands. In
at least some cases, the native trees of these islands have evolved from an-
cestors that were woody shrubs, in all likelihood, at their time of original
establishment. If genetic systems were a consideration, therophytes ought
to bulk very large in insular floras, for they would represent maximum pop-
ulation size, which ought to be valuable where space is limited and in-
breeding potentially a liability. However, outbreeding trees and shrubs are
what we find, chiefly, and the ecology of the island seems of prime impor-
tance in governing what growth forms will establish—but more impor-
tantly, toward what modes evolution will take place. Increased woodiness
seems the inevitable interpretation of these data, particularly if one takes
into account factors of anatomy and systematics.

Ways in Which Woody Growth Forms Evolve

If one considers the various herbaceous groups that are woodier on islands,
one soon concludes that there are a wide variety of ways in which this
process can occur, as suggested below.

1. Rosette herbs, experiencing prolonged vegetative growth over a
period of years, develop elongate stems and eventually flower. ‘The mono-
carpic rosette tree is relatively infrequent on islands, but examples can be
cited. Outstanding among these are Wilkesia (Asteraceae: Hawaii) Tre-
matolobelia (Campanulaceae: Hawaii); Lobelia spp. (Campanulaceae: Ha-
wall); Echium pininana (Boraginaceae: Canary Is. [fig. 5.2, bottom left,
right|); Tauschia (Melanoselinum) decipiens (Apiaceae) on Madeira (fig.
10.2, top right); Dendroseris (Phoenicoseris) pinnata and D. regia (Astera-

BE,

Insular Woodiness

ceae, tribe Cichorieae: Juan Fernandez Is.); Centaurodendron dracaenoi-
des, C. palmiforme, and the closely related, monotypic Yunquea tenzi
(Asteraceae, tribe Cynareae: Juan Fernandez; for photographs of these
Juan Fernandez taxa, see Skottsberg [1953] and Carlquist [1965]). Mono-
carpic species of Lobelia occur in Afroalpine areas (see chapter 14). The
monocarpic species of Aeonium on the Canary Islands could be called
miniature, monocarpic rosette trees.

In relation to rosette plants, one may note that the commonly culti-
vated cabbage (fig. 10.2, top left) takes this form when grown on subtropi-
cal islands. One might also cite Geranium maderense (fig. 10.2, bottom
left), which is not monocarpic, but appears to be merely a rosette plant
that forms a somewhat elongate stem by virtue of growth through a suc-
cession of seasons. Lowe (1868) states that Tauschia (Melanoselinum)
decipiens does not always die upon flowering: rarely lateral branches form
near the base of the main stem. This pattern suggests the origin of the
transitional type discussed next.

2. Development of the ability to form lateral branches may convert
a monocarpic rosette plant into a shrub. This seems plainly what has hap-
pened in the case of Stephanomeria (Munzothamnus) blairit (fig. 10.3, top
left, right). I was able to observe this phenomenon by maintaining a plant
in cultivation. Many species of Stephanomeria (e.g., S. virgata) are autumn-
flowering annuals. During the first year, S. blairit forms a terminal inflo-
rescence (fig. 10.3, top left), much as would an annual Stephanomeria,
differing only in its paniculate inflorescence. At this point, the plant, in-
stead of dying, produces innovations from the base during the winter,
which is mild enough in its native habitat, San Clemente I., Calif., that
such shoots are not killed by frosts. During the second year these innova-
tions grow out, each ultimately tipped by an inflorescence. Some of these
second-year shoots may die, but some of them produce lateral shoots dis-
tally during the third year; during the third year, additional basal shoots
can form (fig. 10.3, top right). Thus a shrubby or suffrutescent form is
achieved, although there is no trunk per se. Some of the lateral branches
eventually become woodier than others, and the caudexlike base continues
to increase in size. This shrublike habit can be seen as failure of the plant
to die from winter frosts. Moisture is perhaps also a factor; the broad leaves
(compared with other species) suggest lowered transpiration in the deep

374

Insular Woodiness

canyons of San Clemente and in the greater humidity of insular condi-
tions. Species of Lobelia in the Afroalpine flora may follow this pattern—
L. gibberoa, for example, appears like a monocarpic, tall Lobelia, but I
have observed innovations forming near the bases of plants in flower on
Mt. Kenya. This is occasionally true of other Lobelias, such as L. keniensis
(Hedberg, 1957).

3. Evolution in habitats free from frost can mean that innovations
need not take place from below or at ground level, but can occur above
ground level on one of the main stems of a plant. A good example of this
tendency is found in Lotus scoparius. Lotus scoparius var. scoparius (fig.
10.3, bottom left) is a microphyllous broom shrub of mainland southern
California, branched from ground level. On Santa Catalina I., southern
California, Lotus scoparius var. dendroideus (fig. 10.3, bottom nght) be-
comes a miniature tree, almost invariably from a single trunk. The crown
is richly branched and bears a dense tangle of herbage with relatively large
leaves.

4. Consistent growth of lateral branches from nodes just beneath an
inflorescence can take place. The resultant growth form varies, but a dis-
tinction from the Stephanomeria example above is evident, in that Steph-
anomeria forms basal innovations at least as frequently as it does branches
just below inflorescences.

The ability to form axillary inflorescences would appear a more direct
pathway to prolonged growth of stems but is evidently a more difficult
process to achieve evolutionarily in a number of taxonomic groups. For
example, none of the Macaronesian species of Sonchus (figs. 10.4—10.8)
has developed axillary inflorescences. During the first year of growth, these
Sonchus species may flower, as shown for a young plant of S. radicatus (fig.
10.5, top nght). At this stage, a Macaronesian Sonchus resembles an an-
nual species, such as the widespread weed S. oleraceus. However, in most
of the species a single innovation grows out after flowering, from just be-
low the inflorescence. This, as it matures, pushes the dead inflorescence
axis to one side so that a relatively straight continuation of the original
branch is formed—a rather striking substitute for an axillary inflorescence
position. In rosette-tree species such as S. abbreviatus (fig. 10.6, top left,
tight), which rarely branches, this condition has been perfected most
clearly. Branching does occur in most of the Macaronesian Sonchus spe-

aD

Insular W oodiness

= lg

i SS
FIG. 10.4. Species of Sonchus (Asteraceae) endemic to the Canary Is. and Madeira. TOP LEFT,
S. ustulatus, plant about 0.2 m tall, acaulescent, leaves 6—8 cm long; Machico, Madeira. Top

RIGHT, S. brachylobus, plant branching but small, ca. 0.3 m tall, leaves 6-8 mm long; northern
coast of Gran Canaria. BOTTOM LEFT, S. congestus, plant acaulescent, with large leaves, to 1
m; middle elevations, Barranco de San Andres, Tenerife. Borrom ricHT, S. platylepis, acaules-
cent plant with leaves ca. 15 cm long; summit region, Gran Canaria.

cies, least often perhaps in the acaulescent S. congestus (fig. 10.4, bottom
left), occasionally in the short-stemmed species S. ustulatus (fig. 10.4, top
left), S. brachylobus (fig. 10.4, top right), and S. platylepis (fig. 10.4, bot-
tom nght). Branching occurs most readily in S. leptocephalus (fig. 10.8),
which is often branched freely from near the base, but elongate branches

376

%

oz
ae ae

es

F1G.10.5. Sonchus radicatus (Asteraceae), an endemic of the Canary Is. top Lert, habit of
older plant, ca. 0.5 m tall; cliffs of upper Barranco de San Andres, Tenerife. Top RIGHT,
younger (probably 1 yr. old) plant in flower, showing nature of leaves, ca. 12 cm long; lower
Barranco de San Andres. BoTToM LEFT, transection of secondary xylem. BOTTOM RIGHT, tan-
gential section of secondary xylem. Scale of magnification of wood sections is shown at bottom
of fig. 10.16.

377

\
har
on
aayeu
Oe
eR0S 8.
4 HS
+4 es
’

Pot
SG Soe SD

PRESTR TIE TET D
= A) hel
+. Tiley : .
‘S
ea Se
eT aaa UIE es tea
i

SSS
~

eae athe " “ of Sheet } f

UP lg A
FIG. 10.6. Rosette-tree species of Sonchus endemic to the Canary Is. TOP LEFT, RIGHT, BOTTOM
LEFT, S. abbreviatus; Anaga Peninsula near Bailladero, Tenerife. rop Lert, habit of plant, ca.
3 m tall. rop ricHT, flowers from plant; open heads measure about 4 cm in diam. BOTTOM
LEFT, transection of secondary xylem. BOTTOM RIGHT, S. jacquinii, tangential section of sec-
ondary xylem; between Icod de los Vifios and Puerto de la Cruz, Tenerife. Scale of magnifica-
tion of wood sections shown with fig. 10.16.

378

Fic. 10.7. Rosette-tree species of Sonchus from the Canary Is. and Madeira. TOP LEFT, RIGHT,
S. pinnatus; near Machico, Madeira. tor LEFT, habit of plant, ca. 1.5 m tall. rop RIGHT, por-
tion of plant in flower; leaves are 15-30 cm long. BOTTOM LEFT, CENTER RIGHT, BOTTOM
RIGHT, S. fruticosus; near waterfall, Ribeiro Frio, Madeira. Borrom Lert, habit, plant is 4.5
m tall. CENTER RIGHT, stem of same plant, showing formation of aerial roots. BOTTOM RIGHT,
base of same plant, showing maximum diameter of this stem; hand indicates scale.

379

rt : f
bi
*

7
7
ec «=

j

ite

ik
| i vole | | N
| OH NCAA COR
AAS sta a
WAAAY Hie tie a
AMPA) AWE a
BAAN i Ya
FIG. 10.8. Sonchus leptocephalus, an endemic of the Canary Is. top Lert, habit of plant,
ca. 1.2 m tall; between Santa Cruz de Tenerife and San Andres. TOP CENTER, view of stem
tips in leafless condition during dry season, same plant. TOP RIGHT, plant with leafy stems;
leaves are ca. 10 cm long, leaf segments terete; near Puerto de la Cruz, Tenerife. Borrom
LEFT, transection of secondary xylem. BOTTOM RIGHT, tangential section of secondary xylem.
Scale of magnification of wood sections shown with fig. 10.16.
380

Insular Woodiness

can also be formed by the mechanism described above. In the species with
little branching, the stems bear a succession of old and almost impercep-
tible weathered inflorescence bases, noticeable as irregularities along what
seems at first glance to be an axis that has been built from continual
growth of an unbranched axis.

The Macaronesian Sonchus species show a variety of growth forms,
so that this mechanism of bearing exclusively lateral inflorescences ob-
viously lends itself to a wide variety of growth forms from acaulescent
rosette shrub, to broomlike shrub, small rosette shrub, and _ sparsely
branched rosette tree.

Stephanomeria and Sonchus both belong to the tribe Cichorieae of
Asteraceae. The Juan Fernandez Cichorieae conform to the same pattern
(except for monocarpic species), as the diagrams of Skottsberg (1953, p.
851) show for Dendroseris litoralis (see also the habit of D. litoralis as
illustrated by Skottsberg [1922, pl. 18] or Carlquist [1965]).

Insular Asteraceae other than Cichorieae also show this pattern, as
diagrammed by Skottsberg (1953) for Robinsonia, Rhetinodendron, and
Symphyochaeta (Senecioneae).

In Macaronesia, woody species of Centaurea (fig. 10.9, top left, right)
and Carlina, of the tribe Cynareae of Asteraceae, agree with the pattern
described. Meusel (1952) has diagrammed ‘‘candelabrum shrubs” in which
this form is achieved by innovations from below inflorescence bases; he
illustrates Carlina, Echium, and Euphorbia. Werner (1966) has shown
this growth habit diagrammatically for Macaronesian species of Isoplexis
(Scrophulariaceae). Pterocephalus of the Dipsacaceae forms shrubby spe-
cies in Macaronesia (fig. 10.9, bottom left, right) conforming with this
type.

Macaronesian Brassicaceae such as Parolinia (fig. 10.10, bottom right),
Cheiranthus, Crambe, and Sinapidendron have this scheme. The Hawai-
ian Lepidium serra (Brassicaceae) is a small rosette shrub (see Carlquist,
1970a) resembling Sonchus radicatus in habit.

In the Hawaiian Islands, genera of Asteraceae such as Dubautia and
Argyroxiphium (some species) follow this pattern. A series of growth forms
including rosette herb (A. caliginii), sprawling matlike shrub (D. scabra),
rosette shrub (D. plantaginea), scandent shrub (D. latifolia), and even
trees (D. paleata, D. reticulata) has been formed on this basis. The sole

381

FIG. 10.9. “Rosette shrubs” of the Canary Is. TOP LEFT, RIGHT, Centaurea arbutoides
(Asteraceae). Top LEFT, habit of plant, 1.7 m tall; cliffs above Agaete, Gran Canaria. TOP
RIGHT, flower heads, 4 cm in diam. BOTTOM LEFT, RIGHT, Pterocephalus dumetorum (Dipsaca-
ceae). BOTTOM LEFT, habit of plant (Giinther Kunkel shown standing by plant). BoTrom
RIGHT, heads of flowers, 3.5 cm in diam.

Insular Woodiness

Fic. 10.10. Woody herbs from the Canary Is. Top LEFT, RIGHT, Plantago arborescens;
middle elevations, Barranco de San Andres, Tenerife. rop Lert, habit of plant, 0.4 m tall.
TOP RIGHT, leaves and inflorescences; leaves are 3—4 cm long. BorromM LEFT, P. webbii, habit
of plant, 0.3 m tall; Los Roques, El Teide, Tenerife. sorrom ricut, Parolinia ornata (Bras-
sicaceae), showing habit of this sparsely branched species, 2 m tall; middle elevations, Barranco
de Arguineguin, Gran Canaria.

exception appears to be D. knudsenii, a true tree with a single trunk, but
in fact inflorescences are terminal. The various species of Argyroxiphium
and Dubautia have been illustrated by Carlquist (1970a).

All of the Asteraceae native to St. Helena (see Carlquist, 1965, pp.
202-3 for illustrations) show the Sonchus type of innovations. These spe-

383

Insular Woodiness

cies vary from shrubs to sizable trees, and belong to the tribes Astereae,
Senecioneae, and Heliantheae. The South Pacific helianthoid genera
Fitchia and Oparanthus (Asteraceae) also show this type of innovation
(Carlquist, 1957, 1965). The Galapagos helianthoid Scalesia also has exclu-
sively terminal inflorescences, and varies from shrub to tree (Howell, 1941;
Harling, 1962).

Dicotyledons with anomalous secondary thickening also show ter-
minal inflorescences. Chenopodium arborescens (Juan Fernandez Is.) and
C. oahuense (Hawaiian Is.) can be cited. Nototrichium (Amaranthaceae:
Hawaiian Is.) can be added. Another amaranthaceous genus with consid-
erable stature and terminal inflorescences is Bosea (fig. 10.12, top left,
right), a Canarian endemic. Bosea, however, tends to be branched often
near the base, thus forming a fountain-shaped shrub.

Although not arborescent, the relatively large (for its genus) stem suc-
culent Talinum guadalupense (Portulacaceae) can be added to this mode
of growth (fig. 10.15, top left). The Afroalpine species of Senecio have ter-
minal inflorescences (see chapter 14).

5. Axillary inflorescences occur in some insular woodier herbs, and
thus axes can continue growth directly. The Hawaiian lobelioids (other
than Lobelia and Trematolobelia) exemplify this type. Illustrations of
these can be found in Rock (1919) and Carlquist (1970a). Thus, evolution
of the tall palmiform species of Cyanea is merely a matter of continued
growth, as it is in the related genera Delissea and Rollandia. Clermontia
species are mostly shrubby and much branched, usually with a single
trunk; branching is not related to production of inflorescences. Flowers
are axillary in the South Pacific lobelioid genus Apetahia.

Axillary inflorescences are characteristic of the genus Plantago (Plan-
taginaceae), on islands and elsewhere. Illustrative of this are the Macaro-
nesian Plantagos (fig. 10.10, top left, right, bottom left). These are shrubs
of dry localities. On the other hand, single-stemmed or little-branched,
macrophyllous, rain-forest species on Hawaii (P. princeps) and Juan Fer-
nandez (P. fernandeziana) still have the same plan with regard to inflo-
Tescences.

Other axillary-inflorescence shrubs and trees that can be cited include
Pacific species of Scaevola (Goodeniaceae); Nothocestrum (Solanaceae:
Hawaii); Lysimachia (Primulaceae: Hawaii); Convolvulus (Convolvu-

384

‘
|
'
.
|

Fic. 10.11. Woody representatives of herbaceous families from the Canary Is. TOP LEFT,
RIGHT, Convolyulus scoparius (Convolvulaceae). Top LEFT, habit, plant 3.2 m tall; Barranco
de Arguineguin, Gran Canaria. Top RIGHT, flowers of this broom shrub; flowers 2 cm in diam.
BOTTOM LEFT, Limonium rumicifolium (Plumbaginaceae), old plant 1 m tall; Era del Cardon,

Gran Canaria. BOTTOM RIGHT, L. arborescens, habit of plant 1.4 m tall; cliffs of northwestern
Tenerife.

385

Fic. 10.12. Insular plants with anomalous secondary thickening. TOP LEFT, RIGHT, Bosea
yervamora (Amaranthaceae). TOP LEFT, habit of plant, 5 m tall; La Calzada, Gran Canana,
Canary Is. rop r1cuT, fruits from same plant; leaves are 3.5—4 cm long. BOTTOM LEFT, RIGHT,
Charpentiera obovata (Amaranthaceae), sections of stem; Kokee, Kauai, Hawaiian Is. Borrom

LEFT, transection. BOTTOM RIGHT, tangential section. Wood sections at magnification shown
with fig. 10.16.

386

Insular W oodiness

laceae: Macaronesia, fig. 10.11, top left, ght); Charpentiera (Amaran-
thaceae: Hawaii and Austral Is.); Limonium (Plumbaginaceae: Macaro-
nesia, fig. 10.11, bottom left, right); and Espeletia (Asteraceae: Colombian
paramos; see chapter 14).

6. Viola has axillary flowers. However, the Hawaiian species of Viola
appear to have achieved woody stature by an unusual change in growth
form. North temperate Viola species are typically prostrate and rhizoma-
tous, as are the Hawaiian bog species. Viola robusta, a wet-forest Hawauian
species, begins with an unbranched prostrate stem, but then the stem turns
upward, so that a mature plant looks entirely upright unless one studies
the base. The most nearly arborescent species, V. tracheltifolia of the Ha-
waiian Islands, also shows vestiges of this change in orientation. For illus-
trations, see Carlquist (1970a).

7. The Hawaiian species of Euphorbia (with the exception of a
single, recently described species, E. haeleeleana D. Herbst) belong to the
section Chamaesyce, sometimes termed a genus (also called Euphorbia
sect. Anisophyllum). Continental species of this section assume a matlike
form because the terminal shoot aborts and the lateral shoots become
prostrate stems. Lowland Hawaiian species (e.g., E. degeneri) also show
this form. Other Hawaiian chamaesycoid Euphorbias are trees and shrubs.
Study of early stages in these species shows that they all undergo abortion
of the terminal shoot, but a lateral innovation in the seedling turns upward
and forms a new main axis which can eventually become a woody trunk,
as in the tree E. rockii. This process of upright shoots based on lateral in-
novations was described by Degener (1937). For illustrations of these Eu-
phorbias, see Degener (1937) and Carlquist (1970a). These species all have
axillary inflorescences.

8. The endemic genus of Cucurbitaceae on Socotra, Dendrosicyos,
has a large erect trunk. Hypocotylary tissue may be involved, but a young
plant in cultivation at the University of California, Berkeley, illustrates
that thickening occurs in regions involving internodes above the cotyle-
dons. This water-tank construction is unique among examples of insular
woodiness and is remarkable for its occurrence in a family of vining plants
(for illustration, see Carlquist, 1965). To be sure, there are other water-
tank plants on the island of Socotra, but these are no different from water-
tank plants on the African mainland.

387

Insular Woodiness

9. Although ancestrally herblike groups of angiosperms are the pri-
mary concern of this chapter, in certain groups shrubby species have in-
creased in stature during their evolution on insular areas. One clear ex-
ample is Sarcopygme, a rubiaceous Samoan endemic rosette tree probably
related to the genus Morinda. Other instances are more subtle: are Ha-
waiian Araliaceae, Rubiaceae, or Urticaceae more arborescent than their
nearest noninsular relatives? In some cases this is almost certainly true.
For example, Platydesma cornuta is a cauliflorous rosette tree with gigantic
leaves. It almost surely represents an autochthonous alteration in size and
shape, although its ancestors in Rutaceae (Pelea or Melicope, perhaps)
were most likely woody shrubs, perhaps with smaller stature but surely
with smaller leaves and narrower branches.

The Canarian species of Bencomia (Rosaceae) are probably woodier
than their alchemilloid ancestors; Hagenia, in high African cloud forest
(see chapter 14), has almost certainly increased in woodiness compared
with other alchemilloids or with Bencomia. The subtlety of increase in
woodiness on islands of a group of basically woody plants is a problem, for
one must survey the world representatives of a genus to see if insular spe-
cies are larger. Heights for each species are difficult to estimate (or un-
known). There are, nevertheless, a number of examples of increase in size
of shrubby species on islands. This is the simplest of all paths to increased
woodiness, of course. Analysis is complicated by the fact that we would
not expect any notable changes in wood anatomy between less woody non-
insular species and more woody insular species in such groups.

10. A factor suggested earlier but not clearly enunciated is that eco-
logical shift into wet forest may accompany increased woodiness in some
groups, but others become woody under dry conditions. These latter in-
stances appear to be plants adapted to growth under very dry conditions.
In the Hawaiian flora, one can cite Nototrichium (Amaranthaceae), Che-
nopodium (Chenopodiaceae), Kokia and Hibiscadelphus (Malvaceae), and
Nothocestrum (Solanaceae). Some shrubby species of Lipochaeta (Astera-
ceae, Heliantheae) occur in Hawaiian lowlands. This is true of Scalesia
in the Galapagos (same family and tribe); very likely, few other genera
could have formed trees under the relatively dry cloud-forest conditions
of the Galapagos, but Scalesia (S. pedunculata) has. Sida eggersti (Malva-
ceae) is a true tree in dry regions of Jost van Duke, Culebra, and other

388

Insular Woodiness

West Indian islands. Didiereaceae on Madagascar are a spectacular ex-
ample of this pattern.

11. Expressions of insular woodiness will vary according to the climatic
regime of any given island or archipelago. Although I cannot take space
to discuss these modalities extensively, I can mention a few examples. In
Macaronesia, arborescence or shrubbiness in herbaceous groups often takes
the form of stem succulents (Sonchus, Euphorbia), broom shrubs (Convol-
yulus) or microphyllous shrubs (Plantago) in dry regions. In dry lowlands
of the Hawaiian Islands, groups with increased woodiness are broad leaved
and generally not very succulent. Only a few groups with increased woodi-
ness have entered the Macaronesian laurel forest, whereas many good ex-
amples of woodier herbs may be found in wet Hawaiian forests. ‘These
distinctive modalities in different zones on islands in different climatic
regimes illustrate not only the diversity of expression of insular woodiness,
but also that there are probably various pathways in this evolution; in
other words, palmiform species of Cyanea on the Hawaiian Islands have
developed from herblike ancestors in ways and for reasons entirely different
from those that have been basic to evolution of the Afroalpine Dendro-
SeNnecios.

Anatomical Correlations of Insular W oodiness
CAMBIAL ACTIVITY

Axiomatically, increased woodiness of insular representatives of her-
baceous groups involves an increase in. production of secondary xylem.
This can occur in several ways.

1. A sluggish cambium gradually produces a wider ring of secondary
xylem. This usually involves retention of a large pith and a large cortex.
The Afroalpine Senecios (e.g., S. keniodendron) or monocarpic rosette
trees (Lobelia gaudichaudii of Hawaii) exemplify this mode of construction.

2. During the life-span of an herb capable of secondary xylem activity,
the amount of secondary xylem accumulated increases with the size of the
plant. Except for lessened mechanical strength and the anatomical fea-
tures discussed below, these plants can mimic truly woody species in
growth form. Examples include Sonchus fruticosus and Euphorbia rockii.

3. An herb capable of secondary xylem activity becomes a stem suc-

389

Insular Woodiness

culent, in which case the mechanical strength of xylem is often of less
significance (cortical and pith parenchyma serve for support, and the pres-
ence of few or no leaves lessens wind resistance). Examples include Eu-
phorbia balsamifera, Sonchus leptocephalus, Coreopsis gigantea, and Tali-
num guadalupense.

4. Anomalous secondary growth, characteristic of a particular family,
produces successive cambia to an extent greater than would be character-
istic of ancestors. Species of Centrospermae (Caryophyllales) would be in-
cluded in this category, such as insular Amaranthaceae (Nototrichium,
Charpentiera) and Chenopodiaceae (Chenopodium spp.).

FEATURES OF WOOD ANATOMY

The anatomical mode of secondary xylem construction will obviously
not be the same for all the groups just listed. Some features are common
to all types, some are not. In a series of studies on wood anatomy of various
predominantly herbaceous groups of dicotyledons, I have attempted to dis-
cover principles of construction underlying woody members of these groups
(Carlquist, 1962a; Carlquist, 1966, and the various preceding papers on
wood anatomy of Asteraceae summarized therein; Carlquist, 1969a, 1969b,
1970b, 1970c, 1970d, 1971).

Vessel-element Length. Length of vessel elements is a simple and
relatively accurate measure of length of fusiform cambial initials. Early
in my studies of woody herbs, I found that they often have vessel elements
longer than one would expect. However, the ontogenetic dimension of
this phenomenon plainly became an important key, leading to the concept
of paedomorphosis (Carlquist, 1962a). The ontogeny of only some of the
species discussed here was investigated, although I feel a sufficiently large
number of species have been investigated that I can state the following
principles with regard to woodier herbs:

1. Greater length of fusiform cambial initials, characteristic of the
opening stages of cambial activity, is continued with little change as pro-
duction of secondary xylem continues.*

*wNoTE: The significance of longer elements proves to be primarily in the greater
mechanical strength of imperforate elements (libriform fibers); in the plants under
consideration, longer elements therefore relate to greater plant size, which in turn
tends to relate to mesomorphy (Carlquist, unpublished data).

390

Insular Woodiness

2. The greater length (as compared with decreased length character-
istic of early secondary-xylem production in “‘truly woody” plants) is either
ontogenetically lessened, or more commonly changed very little (gradual
ontogenetic increase characterizes most truly woody plants, on the con-
trary). Although we have no clear way of knowing, there is the possibility
that phylogenetic increase in length of fusiform cambial initials might
occur within some herbaceous groups.

3. From earlier data on a broad sampling of Asteraceae (Carlquist,
1966), there is every reason to believe that mesomorphy is related to longer
vessel elements (and therefore to fusiform cambial initials), and the reverse
is true of xeromorphy. If a genus capable of paedomorphosis evolves into
an insular habitat more mesic than that of ancestral types, longer vessel
elements are to be expected. Two genera have been selected for presenta-
tion here—Euphorbia and Sonchus (table 10.2). ‘The Hawaiian species of
Euphorbia sect. Chamaesyce show a wide range in habitat, whereas the
Macaronesian species of Sonchus and of Euphorbia sect. Tithymalus do
not have available to them the “more mesic” end of the spectrum that
occurs in Hawaii.

Vessel-element Width. That vessel-element width is an indicator of
degree of mesomorphy or xeromorphy in groups that show insular woodi-
ness has emerged from a series of studies (Carlquist, 1966, 1969a, 1969b,
1970b, 1970c, 1970d, 1971). Vessel width is not under the control of length
of fusiform cambial initials, and therefore serves as confirmation of meso-
morphic or xeromorphic adaptation. Wider vessel elements seem clearly
correlated with greater mesomorphy.

In table 10.2, figures for average vessel-element length and width are
presented. If length and width are independent indicators of moisture
availability, conceivably either might be significant. Consequently, a com-
bined width + length figure is presented for each species. Sequences
within genera (and sections, in Euphorbia) in table 10.2 have been ar-
ranged according to this combined figure.

In insular Euphorbias, the correlation with rainfall is startlingly per-
fect and needs no comment. I am reminded of a similar correlation with
leaf area in Dubautia (fig. 4.4). The correlation in Sonchus (table 10.2) is
nearly as perfect. The rainfall figures for the Hawaiian Islands are rela-
tively accurate, because of the numerous recording stations. The Canarian

391

C9E
Sve
672é
vie
8LZ

£97

8SZ
£S7
OFZ
DEG

bis
I8¢

S89
C84
OLY
604
e8e
vee

C672
8Ec

yyeua] +
UtPIAN

S8 LLZ
Sik OLZ
8L LSC
68 $72
cOl 9LI
68 L02
VS +07
oF L0Z
OL OLT
eZ €£$
29 ZS+
09 Zé
19 $79
6€ €7b
b9 90+
€S 9c¢
96 LZ
ch Z8Z
bh SbZ
ce 902
a Wl “148Ua]
Y}PIM 1assaa quailaja
ABDLOAY -JO880A
IBDLOAY

00SZ
0002
00SZ
089
Osh

0O0b

Oss
OSPF
OSE
Ore

00SZ
OSS

0008
OOSZ
007
0002
OOOT
Ocr

OSZ
v00E

LU
]pfuip4s
aqDUull

-xolgdgy

‘(T96[) YOosuouIN[g wos spuejsy ULITEML_T UO LLC v

ojuoua [, “eseuy

PILIOPRI ‘SPITOIY] Op [RING

PILOPRI] ‘OL OILOQIyyT

dflI9Ud T, “e[GuIeY op ueN{ ues
dflI9Ud J,

‘soIpuy Urs op OoULLIEg
VLUIBUPD)

UPID) ‘UOIOP[LD 9p O1qouaDg

BILIOPPJY “PARI: OLOGY

PULURD ULI ‘OJORSY DAOqE

BIIBUBD) UBIFy “EPPZ[PD PT

vILOpeyy ‘Oorporyy

PITOPLYY ‘OL, OLLOGNy
LILOpe|] ‘PALIG OILOGNy

nyegd ‘nnpeung joddn
nyrg ‘Mmuerenyeuoy IoMoy
IEnBY ‘OOYOH

Ieney ‘IyOY

mneyy ‘enyeypedny, a

Teue'y UIO}SPIY}IOU

Ieney ‘UOAURD POLITE AA
nyeg ‘nndeyeyy

Aq1DI0'T

SNyD1AalqgP *S
snypuuidg *
snsoo1jnif
nuinpapl

$njsapuoa

snjpjnjsn *
sidajAqpjg *
snypydas0}qa] °
snjpuuid *S
(uvisouoreoRyy) snyau0sg
P1LOPISIG *4]
D1LOJDISIG *"T
snppuldy}1 J, *JOos “Piqioydniy
M0 Sar
Dyofusnjo “iT
MUPLULAIPY *]
ppodo.a}¢ “Iva 14a “oT
SISUAINDLUL “ILA SAP1O1]SP]AI “5
SUuD]Ia]qG UD
“IBA SAP1OL}SP]AI "iT
DIJOf1L0] “ILA SAP101}8P]29 “oT
1auapap "I
JIASDUDYT) “09S “DIqLoygniy

S
S
S
Sy
snqo]cyopig *S
S
S
S

SULIOASOISUL IL[NSUI UL SUOISUIWIP [ASSIA JO SUOT}RIOIION [PIIBO[ON

‘TOL XIGvL

Insular Woodiness

data are much less reliable, and I have made estimates, based on altitude,
leeward versus windward exposure, aspect of vegetation at the various
localities, and the few recording stations. One must remember in either
case that at stations of low rainfall, high evaporation lowers the available
soil moisture still further. At stations of high rainfall, cloud cover and
shade from trees increase available soil moisture appreciably, so that the
gamut in rainfall is a foreshortened expression of conditions experienced
by the plants themselves.

Several other factors must be taken into account in relating tracheary
element length with moisture availability. If, as seems frequent, there is a
decrease in element length with age in woody herbs, a large wood sample
might be expected to have elements shorter than the moisture correlation
would dictate. A number of examples can be cited: in Lobelioideae (Carl-
quist, 1969a), Cyanea pilosa, which forms little wood, has notably long
tracheary elements, but Clermontia clermontioides, which has the largest
stems I have collected in Lobelioideae, has rather short tracheary elements.
To test this correlation, measurements were taken from sections near the
pith of a basal sample, from the periphery of a basal sample, and from an
upper branch. Data (Carlquist, 1969a) on Clermontia arborescens, C. ha-
watiensis, C. kakeana, C. montis-loa, Cyanea leptostegia, and others con-
firm this tendency: in lobelioids as in other woody herbs (Carlquist, 1962a;
Carlquist, 1970b) there is a decrease in element length with age. Interest-
ingly, upper branches often have longer elements than does the inside
portion of the base. The explanation for this difference is possibly stress
on strength in the branches.

At any rate, woodier species in a group are likely to occur in rain
forests, less woody ones in lowlands. Therefore, sections from the former
may exhibit shorter elements than expected on the basis of their meso-
morphy. This probably accounts for the fact that vessel elements in Son-
chus fruticosus and S. abbreviatus are not spectacularly longer than those
of the lowland Sonchus species.

Even with these restrictions, the vessel-element dimensions, derived
as they are from independent taxonomic groups and independent island
groups, show clearly that evolution into situations of increased water
availability is accompanied by increase in vessel-element length and width.
Surely no one is going to propose the reverse hypothesis: for example,

393

Insular Woodiness

Euphorbia degeneri resembles the wide-ranging E.. atoto stock that is prob-
ably basic to this complex, whereas E. rockii is unique; beginning a series
with EF. rockii seems preposterous. In insular species that occupy rather
xeric localities, rather short tracheary elements would be expected. This
proved true in most Macaronesian taxa (Carlquist, 1970b, 1970c, 1971).
Brassicaceae have evolved into the cloud forest to a negligible degree in
Macaronesia, and Echium shows only a few forest species (which do, how-
ever, have slightly longer tracheary elements).

Length of libriform fibers in all cases parallels vessel-element length
for a given species. Collection of data on this feature (see the various
papers cited earlier) served little purpose except to reinforce the picture
of vessel-element length and to demonstrate that length of fusiform cam-
bial initials is basic in length considerations.

A photographic indication of vessel-element dimensions in Sonchus
species is given in the bottom pictures of figures 10.5, 10.6, and 10.8.

Ray Histology. If vessel-element length reflects the condition of the
vascular cambium, we might look to ray initials and the ray cells they pro-
duce to see if similar conditions apply. They do indeed, and transverse
divisions of both ray initials and fusiform cambial initials lag in woods
that show paedomorphosis. Where vessel elements are long for a particular
group, a higher percentage of erect ray cells as opposed to procumbent
cells was almost invariably observed. Predominance of erect ray cells, or
presence of erect and square ray cells to the exclusion of procumbent cells,
is unusual in truly woody dicotyledons; in fact, Kribs (1935) did not even
include such ray types, common in woody herbs, in his survey of rays of
dicotyledons. These ray types seem a protraction of a juvenile condition,
in which case, as with vessel-element length, derivation from more herba-
ceous ancestors 1s indicated.

Raylessness. Raylessness characterizes woody herbs (Barghoorn, 1941;
Carlquist, 1961, 1970c), many of which are insular. I advanced the hy-
pothesis (Carlquist, 1970c) that in a species which is herbaceous and has
specialized wood (very short fusiform cambial initials), increase in xylem
production under appropriate conditions (involving paedomorphosis) may
often lead to production of libriform fibers like ray cells in height, and
thus so short as to be histologically equivalent to ray cells. As Barghoorn
(1941) noted, continual production of xylem sometimes does lead to sub-

394

Insular Woodiness

division in ray initials, so that rays eventually appear in stems that begin
with a rayless condition.

These considerations apply to two rayless woods, hitherto unreported.
Viola tracheliifolia (fig. 10.14, top left, nght) is probably the most woody
species of its genus, and shows complete absence of rays. Although up to
10 ft tall, this species is not truly arborescent (taller individuals grow in a
leaning posture among other shrubs); production of secondary xylem may
be relatively finite, and eventual production of rays impossible. This may
also be true of Lysimachia kalalauensis (fig. 10.14, bottom left, nght), also
a Hawaiian species, which possibly has greater accumulation of secondary
xylem than other Primulaceae. Lysimachia kalalauensis is a rosette shrub,
chiefly branched from the base, which grows in underbrush of semiexposed
areas In upper Kauai.

Among insular species of Plantago, rays have developed only in the
Macaronesian species, in which secondary xylem production is maximal
for the genus. Even here, however, rays are relatively few and small (Carl-
quist, 1970c).

Raylessness in young stems, followed by eventual production of rays,
characterizes Hawaiian species of Geranium (Barghoorn, 1941). Another
group in which this occurs is the Hawaiian species of the gesneriaceous
genus Cyrtandra (fig. 10.13, top left, nght). The transection illustrated (fig.
10.13, top left) shows initiation of rays. Other species of Cyrtandra are
similar. Ray production begins soon, so that in all samples of Cyrtandra
studied (10 species), rays oceurred in outer portions of the secondary
xylem. In the species illustrated, rays are few and small, however (fig. 10.13,
top right).

Ixanthus, a monotypic Canarian genus of Gentianaceae, shows an
almost rayless condition (fig. 10.13, bottom left, right). The inner portions
of secondary xylem are rayless, and the oldest stems available show only
a few inconspicuous rays in their outermost portions.

Although the physiological significance of raylessness is difficult to
assess, we might say that the value of rays as transport or storage tissues in
woody species with larger secondary xylem cylinders is minimized in her-
baceous species incapable of indefinite production of xylem. Thus the
herbaceous species of relatively small and finite xylem production are
most likely to be rayless, whereas others (Geranium, Cyrtandra) which

395

a,e 4° rks had

8 tans oiste laste

ees eta’ 4 ed

nG hen’ ; H gis i
nite bea'e gic R -2h,7 ae oe seh Be.

ML TLR #

i233

LAG ¢ ae
4x58 a8
au: ¢ ox;
yt =

: rs

od a

yo
~

4 4"

Al " 5 ele
get: 6atneet
ii ags' ile e ageiecs

typ
Siyhhon'e
yeeiate 8)
CAAA

FIG. 10.13. Sections of woods of insular plants that are rayless at first, eventually develop rays.
TOP LEFT, Cyrtandra sp. (sect. Chaetocalyx), transection of secondary xylem; Kahana Valley,
Oahu, Hawaiian Is. rop ricut, Cyrtandra cf. propinqua (Gesneriaceae), tangential section of
secondary xylem; Waikane, Oahu, Hawaiian Is. BorroM LEFT, RIGHT, Ixanthus viscosus
(Gentianaceae); laurel forest of Anaga Peninsula, Tenerife, Canary Is. BorroM LEFT, transec-

tion of secondary xylem. BOTTOM RIGHT, tangential section of secondary xylem. Magnification
shown by scale with fig. 10.16.

396

eat
et aata
Th

Sere
vs)
De aay TA

Fic. 10.14. Rayless woods from the Hawaiian flora. Trop LEFT, rIGHT, Viola tracheliifolia
(Violaceae), sections; Waimea, Kauai. Top LEFT, transection of secondary xylem. TOP RIGHT,
tangential section of secondary xylem. BOTTOM LEFT, RIGHT, Lysimachia kalalauensis (Primu-
laceae), sections; Kalalau Lookout, Kauai. sorrom Lert, transection of secondary xylem.
BOTTOM RIGHT, tangential section of secondary xylem. Magnification shown by scale with fig.
10.16.

397

by

&

tant |

, Ha
Se PeeSss

oy

FIG. 10.15. Talinum guadalupense (Portulacaceae), an exceptionally large stem succulent for
its genus; Outer Islet, Guadalupe I., Mexico. ror Lert, habit of plant, ca. 0.8 m in diam.
(shown sideways). TOP RIGHT, flower, 4 cm in diam., and leaves. BOTTOM LEFT, transection of

secondary xylem. BOTTOM RIGHT, tangential section of secondary xylem. Magnification for
wood sections shown by scale with fig. 10.16.

398

Insular Woodiness

eventually develop larger secondary xylem cylinders are likely to acquire
rays ontogenetically. Raylessness is by no means an exclusively insular phe-
nomenon. However, considerations implicit above (for a lengthier discus-
sion, see Carlquist, 1970c) suggest that such indications of paedomorphosis
as raylessness, rays composed wholly or predominantly of erect cells, and
relatively long vessel elements are interpreted best as criteria for herbaceous
or more nearly herbaceous ancestry. Woods with these phenomena would
be difficult to interpret as products of a more woody ancestry.

Parenchymatization. Increased parenchymatization is a common phe-
nomenon in insular woody herbs. This might, at first glance, appear to run
counter to the trend of raylessness. However, stems rich in axial paren-
chyma are larger than those which exhibit raylessness; increased parenchy-
matization is particularly characteristic of stem succulents. Talinum guada-
lupense (fig. 10.15, bottom left, right) is a good example. In this species,
rays are exceptionally wide and probably little altered from the conforma-
tion of pith rays in the primary stem. Ray cells are nonlignified. In the
axial xylem, nonlignified parenchymal cells are present to the exclusion of
libriform fibers; the axial xylem consists wholly of vessel elements and
parenchyma. Brighamia insignis (Carlquist, 1969a) also contains these fea-
tures except at the base of the plant and in roots, where the narrow bands
of axial xylem do contain thin-walled fibers.

In another stem succulent, Coreopsis gigantea (fig. 10.16, bottom
left), libriform fibers are present but axial parenchyma is abundant in
patches and bands. Rays are wide. Macropiper excelsum (fig. 10.16, top
left, right) could be described as a stem succulent. In this species, wide
rays are present almost exclusively. As new rays are initiated in an enlarg-
ing stem, they are converted rapidly to wide rays resembling those already
present. A number of stem succulents (e.g., the cactoid Euphorbias [Carl-
quist, 1970d]) do not have wide rays, and have parenchyma chiefly in the
cortex.

Increase in axial parenchyma is characteristic of some other succulents
or semisucculents—Scaevola coriacea, S. crassifolia, and S. spinescens of
Goodeniaceae, for example (Carlquist, 1969b). There are rain-forest spe-
cies of Scaevola (S. glabra) in which a marked increase in axial parenchyma
does not seem related to succulence.

Bands of parenchyma (“marginal parenchyma’) are characteristic of

399

seR SE

vera

SBS!

ee.
8

nae ens

Es
ae
EY

x

\y

b

POY

Tig ie

karan
=

»
: ane oe
ma

qn
—

.*e
oe
“c+

or ee

;

FIG. 10.16. Anatomical details of woody herbs. Top LEFT, RIGHT, Macropiper excelsum
(Piperaceae), sections; between Westport and Greymouth, New Zealand. Top LEFT, transection.
TOP RIGHT, tangential section. BOTTOM LEFT, CENTER, Coreopsis gigantea (Asteraceae), a species
from California islands and adjacent coast; Malibu, Calif. sorrom LEFT, transection of secondary

(continued)

400

Insular W oodiness

certain insular species. They have been observed chiefly in Asteraceae.
They are shown here in Sonchus radicatus (fig. 10.5, bottom left), whereas
other species of Sonchus lack them (figs. 10.6, bottom left; 10.8, bottom
left). Other examples include the insular California species of Hemizonia
and species, chiefly alpine, of Dubautia (Carlquist, 1958). In these, bands
of parenchyma are believed to have resulted from “fiber dimorphism”
(Carlquist, 1961). Because, in Asteraceae, libriform fibers may not differ
from axial parenchyma except in length and a little in wall thickness, a
minor shortening and a persistence of a nucleus result in a cell, type that
we must term axial parenchyma. The significance of parenchyma in these
Asteraceae is not obvious, although it appears in insular species with
growth forms that stress mechanical strength less and those that live in a
climate with marked dry seasons.

Woods with anomalous secondary xylem, such as Charpentiera (fig.
10.12, bottom left, right), Bosea (fig. 10.12, top left, right), Nototrichium,
Chenopodium, and Convolyulus (fig. 10.11, top left, right), obviously have
a mode of parenchymatization related to successive cambium formation.
Genera characterized by anomalous secondary thickening would seem good
candidates for insular woodiness because the process of anomalous second-
ary thickening seen in annual representatives of Chenopodiaceae, Amaran-
thaceae, and others can be continued indefinitely in situations where cli-
matic moderation permits.

Libriform-fiber Characteristics. Although one can select a few examples
of woodier insular herbs in which libriform fibers are thick walled, thin-
walled fibers characterize the great majority of them. The species of Son-
chus illustrated here are notable in this regard (bottom left and right of
figs. 10.5, 10.6, 10.8). Fibers are so thin walled in this genus that sectioning
is extremely difficult. This is also true in insular species of Echium (Carl-
quist, 1970b), Lobelioideae (Carlquist, 1969a), Plantaginaceae (Carlquist,
1970c), most Brassicaceae (Carlquist, 1971), Euphorbia (Carlquist, 1970d),
Cyrtandra (fig. 10.13, top left, right) and many Asteraceae (see papers cited
ric. 10.16 (continued)
xylem. BOTTOM CENTER, vessel from tangential section of wood. sorrom ricut, Dendroseris
berteriana (Asteraceae), section of secondary phloem and adjacent cortex to show lack of
fibers in bark; specimen collected by Carl Skottsberg on Masatierra, Juan Fernandez Is. Scale

of magnifications for all except bottom center is shown at bottom. Vessel in bottom cen-
ter, X ca. 600.

401

Insular Woodiness

in Carlquist, 1966). The exceptions to the rule are interesting. In Lobe-
lioideae, thick-walled fibers could be said to characterize only upper por-
tions of Cyanea leptostegia. Thick-walled fibers definitely seem correlated
with mechanical strength in this tall, palmiform species (Carlquist, 1969a).

Among the exceptional insular species of herbaceous groups with hard
wood (owing to thick-walled libriform fibers), one can cite Parolinia ornata
(fig. 10.10, bottom nght). However, predominance of thin-walled fibers
suggests that selective pressure for mechanical strength is lessened. How-
ever, if insular woodier herbs have, in general, longer fusiform cambial
initials and therefore longer libriform fibers, this length might offer greater
mechanical strength.

Gelatinous fibers or fiber-tracheids characterize some insular groups.
Most notable are species of Euphorbia (Carlquist, 1970d) and Goodenia-
ceae (Carlquist, 1969b).

Pitting. Occurrences of scalariform or near-scalariform pits on vessels
in woods of herbaceous plants that are woodier on islands are too numer-
ous to mention here. These instances have been discussed in virtually all
of my recent papers dealing with woods (1962a, 1962b, 1969a, 1969b,
1970d, for example). Such pitting occurs in Coreopsis gigantea (fig. 10.16,
bottom center). An almost perfect example is offered by Macropiper excel-
sum (Carlquist, 1962a). I have advanced two explanations, which can be
regarded as complementary rather than mutually exclusive, for this phe-
nomenon. First, this pitting pattern is a protraction into secondary xylem
of metaxylem pitting patterns (Carlquist, 1962a). Second, it tends to occur
in groups in which stem succulence or other factors place low stress
on mechanical strength. “Release” from narrowness of pits is possible
where vessel elements of greater mechanical strength are not of selective
value (Carlquist, 1969a, 1970d). The role that stem succulence plays is
obvious from the presence of wide, gaping, elongate pits, often with mini-
mal wall area between pits, in such families as Crassulaceae and Cactaceae
(notably cereoids) at large.

Growth Rings. Degree of growth-ring development in insular woody
herbs is related, with few exceptions, to seasonality of the environment in
temperature or rainfall. Insular woods also have fewer growth rings than
do their continental counterparts. In general, Macaronesian woody herbs
show relatively marked growth-ring development except for laurel-forest

402

Insular Woodiness

species, such as Sonchus abbrevidtus (fig. 10.6, bottom left). Growth rings
in Macaronesian woody herbs are a response to dry summers; stem succu-
lents, however, do not show growth rings very much, as one would expect,
for a stem succulent is not affected directly by soil moisture.

Growth rings are relatively mild in Hawaiian woods; they are usually
expressed in the form of somewhat wider vessels in early woods. The Ha-
walian lobelioids (Carlquist, 1969a) and Goodeniaceae (1969b) are rela-
tively poor in growth rings. Interestingly, the least demarcation of growth
rings was observed in Afroalpine Lobelia species (Carlquist, 1969a), Afro-
alpine Senecio species (Carlquist, 1962b), and the Espeletias of the equa-
torial Andes (Carlquist, 1958). The uniform temperature regime through-
out the year in these regions evidently is responsible. In addition, there
is an aspect of stem succulence in these groups. Equatorial alpine bog spe-
cies have, of course, a constant source of water.

PHLOEM AND CORTEX

Herbs are often plants with minimal mechanical strength. To the ex-
tent that this generalization is true, it applies not only to xylem but to
extraxylary fibers as well. Among families and genera that are woody on
islands and in which I found little or no extraxylary fiber formation are
Boraginaceae, Brassicaceae, Campanulaceae, Goodeniaceae, Solanaceae,
and many groups of Asteraceae (e.g., Dendroseris | Phoenicoseris| bertert-
ana, fig. 10.16, bottom right). Families with anomalous secondary thicken-
ing and succulence of stems can be added to this list. Extraxylary fibers are
definitely present in some groups of woody herbs on islands, but even in
these, they rarely reach the abundance or strength of fibers in families
of truly woody dicotyledons, such as Fagaceae, Moraceae, and Sterculia-
ceae. Lack of extraxylary fibers may be one indication of an herbaceous an-
cestry for a particular group.

A Catalogue of Woody Herbs on Islands

The lists below have been compiled to demonstrate groups in which at
least some taxa are generally greater in size than are their continental
counterparts. This accounts for omission of some arborescent insular

403

Insular Woodiness

groups in which continental relatives often exceed insular species in size
(cactoid Euphorbias and cacti, for example). Because oceanic islands are
not the sole sites for evolution of woody herbs, a discussion of continental
areas with insular characteristics follows. I believe these continental areas
are important, because their woody herbs show some portions of the insular
syndrome. Climatic regions on islands are not exclusively different from
those on mainland areas, even if islands do offer more definitive examples.
Our understanding of insular woodiness would be incomplete if we failed
to take into account continental situations where woody herbs have
evolved. Arrangement of the insular areas is geographical: Pacific, Atlantic,
then Indian Ocean islands, followed by equatorial montane situations and
subtropical continental climates with certain degrees of climatic modera-
tion, disharmony, and other features that tend to promote insular woodi-
ness. Estimates of size of plants and of numbers of species are approximate.
Fndemic genera are indicated by “e” within the parentheses listing species
and their sizes.

HAWAIIAN ISLANDS

The list is based on F. R. Fosberg’s data in Zimmerman (1948), on

Hillebrand (1888), Rock (1913, 1919), Sherff (1935, 1937a), and

Carlquist (1970a).

Amaranthaceae: Charpentiera (4 spp.; 2-6 m); Nototrichium (e, 3
spp. + vars.; 2-6 m). Asteraceae: Argyroxiphium (e, 4 spp. acaulescent, but
A. grayanum a shrub to 2 m); Bidens (ca. 40 spp.; some to 3 m, others
subshrubs or suffruticose); Dubautia s.]. (e, ca. 30 spp. + vars.; D. knud-
senit and D. arborea to 8 m, others shrubby to decumbent); Hesperomannia
(e, 3 spp. + subsp.; to 5 m); Lipochaeta (e, ca. 25 spp. + vars.; to 1.5 m,
some species decumbent); Remya (e, 2 spp.; to 2 m); Wilkesia (e, 1 sp.; to
5 m). Brassicaceae: Lepidiuwm (3 spp.; L. arbuscula and L. serra to 1.5 m).
Campanulaceae, tribe Lobelioideae: Brighamia (e, 1 sp.; stem succulent
to 1.5 m); Clermontia (e, ca. 40 spp.; all shrubs, some—C. arborescens—to
8 m); Cyanea (e, ca. 50 spp.; 2-14 m); Delissea (e, ca. 6 spp.; to 5 m in
D. undulata); Lobelia (10 spp.; some to 5 m—L. gaudichaudii; some sub-
herbaceous); Rollandia (e. ca. 8 spp.; to 6 m); Trematolobelia (e, 1 sp. +
vars.; to 2m). Caryophyllaceae: Alsinodendron (e, 2 spp.; to 2 m); Schiedea
(e, ca. 30 spp. + vars.; to 1.5 m but mostly herbaceous); Silene (5 spp.;

404

Insular Woodiness

S. struthioloides and S. alexandri shrubby, to 1 m, others herbaceous).
Chenopodiaceae: Chenopodium (1 sp.; C. oahuense to 6 m). Euphorbia-
ceae: Euphorbia (ca. 16 spp. + vars.; E. haeleeleana to 14 m, E. rocki
and several others to 7 m, others shrubby or decumbent). Geraniaceae:
Geranium (ca. 6 spp.; G. arboreum to 4 m, others shrubs, one decumbent).
Gesneriaceae: Cyrtandra (ca. 140 spp.; some to 7 m, some subherbaceous).
Goodeniaceae: Scaevola (ca. 7 spp.; several to 6 m, S. coriacea decumbent).
Malvaceae: Hibiscadelphus (e, 3 spp.; 2-8 m; one a shrub); Kokia (e, 3
spp.; trees to § m). Nyctaginaceae: see discussion below. Plantaginaceae:
Plantago (ca. 10 spp. + vars.; P. princeps to 2 m, others rosette herbs).
Primulaceae: Lysimachia (ca. 13 spp.; to 2 m in L. Ralalauensis, others
shrubs or subshrubs). Polygonaceae: Rumex (2 spp.; R. giganteus to 2 m
as a shrub, to 12 m scandent). Solanaceae: Nothocestrum (e, ca. 6 spp.;
to 10 m in N. breviflorum, others at least 3 m); Solanum (ca. 7 spp.; S.
kauaiense to 5 m, others shrubby). Violaceae: Viola (7 spp.; V. trachelt-
ifolia to 5 m, others smaller).

Charpentiera is known only from the Hawaiian and Austral Islands;
its closest affinities appear to be with Chamissoa from Mexico. Among
Asteraceae, the Argyroxiphium-Dubautia-W ilkesia complex is discussed in
chapter 4+. Hawaiian Bidens species appear close to such North American
species as B. pilosa but may have been derived via the southeastern
Polynesian species; Bidens is more arborescent on the Hawaiian Islands
than elsewhere. The affinities of Hesperomannia are imprecisely known,
but must lie with actinomorphic-flowered American mutisioids such as
Gochnatia, Anastraphia, Stenopadus, and Stifftia. Lipochaeta belongs to
Heliantheae, subtribe Verbesininae, and may be allied to Zexmenia,
Wedelia, and Verbesina from western North America. Reports of Lipo-
chaeta elsewhere have been discredited (Harling, 1962). Remya, today
nearly extinct, has obscure affinities with Haplopappus or other astereoid
genera of North America. In Brassicaceae, the Hawaiian species of
Lepidium may have had various origins; the genus is pan-boreal. The
same can be said for Hawaiian species of Lobelia (Campanulaceae), and
therefore also Trematolobelia, a probable derivative of Lobelia. Clermontia,
Cyanea, Delissea and Rollandia (see also chapter 4) form a close grouping
and have their closest affinities with similarly baccate Andean genera such
as Burmeistera and Centropogon, which can become woody at their bases.

405

Insular Woodiness

Brighamia, although not truly woody, is probably greater in stature than
its ancestors; such genera as Hippobroma and Isotoma have been suggested,
but there are possible resemblances to Apetahia of southeastern Polynesia.
In Caryophyllaceae, Alsinodendron and Schiedea are almost intercontinu-
ous, but their relationships are obscure and could lie with any of several
boreal herbaceous genera, such as Minuartia. Chenopodium oahuense
(Chenopodiaceae) is the most arborescent species of its genus; its origins
may be boreal. The Hawaiian Geraniums are distinctive and have even
been segregated as an endemic genus (Neurophyllodes), but are best
regarded as sect. Neurophyllodes of Geranium. Geranium jahnii (sect.
Paramenses) may be close. Cyrtandra (Gesneriaceae) is an Indo-Pacific
genus, and only a few Hawaiian species are exceptionally woody. Scaevola
(Goodeniaceae) is clearly Australian in origin, but the Hawaiian species,
which include the tallest shrubs and trees, probably reached the Hawauan
Islands via South Pacific way stations. In Malvaceae, Hibiscadelphus is
close to Hibiscus, from Indo-Malesia. Kokia is alleged to be related to
Gossypium, in which case its affinities would lie with New World species
such as G. arboreum. Abortopetalum, claimed to be an endemic Hawauan
genus, is perhaps only an exceptional species of Abutilon. Lysimachia
(Primulaceae) is a boreal genus, but Hawaiian species are exceptional in
woodiness. Shrubby species of Rumex (Polygonaceae) can be found in
various parts of the world, but probably derive from boreal ancestors.
Nothocestrum (Solanaceae) seems close to Andean species of Cestrum,
which are shrubs; Hawaiian Solanums could have had various origins. ‘The
same can be said for Hawaiian species of Viola, which are related to boreal
Violas.

One could add Nyctaginaceae to the list. Pisonia umbellifera and
P. sandwicensis form sizable trees; however, they are related to P. inermis,
an Indo-Malesian strand species that may have become arborescent on
continental beaches rather than on islands. The problem of how much
increase in size can be attributed to Hawaiian Rubiaceae, Rutaceae, and
Araliaceae has been mentioned; Gunnera, although gigantic in the Hawai-
ian Islands, is not truly woody—it has little secondary xylem activity in its
polystelic stem and is just a “giant herb.”

The chief site for evolution of arborescence in the Hawaiian flora
appears to be the wet (Metrosideros) forest, although an appreciable

406

Insular Woodiness

number of the species listed above occur in the somewhat drier Acacia
forests. Arborescence in lowland Amaranthaceae, Chenopodiaceae, and
Malvaceae has been discussed earlier in this chapter. Woodiness has
increased in bogs only in Argyroxiphium grayanum and Lobelia gaudi-
chaudii, and the opposite effect is seen in Pelea waialealae and the bog
ecotypes of Metrosideros. The distribution of growth forms of Dubautia
(chapter +) in various ecological zones shows that maximal arborescence
is related to wet forest. Even Nothocestrum, characteristically a lowland
genus, has entered the rain forest (N. peltatum). Because much of the
Hawaiian forest is scrubby (e.g., the Koolau Range Metrosideros forest),
evolution toward shrubby species rather than tall trees is to be expected.
In tall Hawauan forest, or in gulches, woodier herbs tend to occur as
understory elements rather than emergents.

SOUTHERN POLYNESIA

The lists are based on data from Brown (1935), Carlquist (1957,

1969b), Carlquist and Grant (1963), Christophersen (1928), Moore

(1933), Sherff (1937a, 1937b), and Wimmer (1953, 1956).

Marquesas Islands. Asteraceae: Bidens (9 spp. [e]; mostly to 1 m);
Oparanthus (O. albus, 1-3 m). Campanulaceae: Apetahia (A. longistigmata
to 1 m). Gesneriaceae: Cyrtandroidea (e, C. jonesii to 8 m). Goodeniaceae:
Scaevola (2 close spp.; to 2 m).

Society Islands. Asteraceae: Bidens (8 spp. [7 e]; B. mooreensis to
3.6 m, others somewhat smaller); Fitchia (5 spp. + 1 subsp.; F. nutans
to 8 m, others at least to 2 m). Campanulaceae: Apetahia (A. raiateensis
to 2 m); Sclerotheca (3 spp.; S. arborea to 8 m, others at least 2 m).
Goodeniaceae: Scaevola (S. tahitensis to 2 m). Onagraceae: Fuchsia (F.
cyrtandroides to 2 m).

Rapa. Asteraceae: Fitchia (F. rapensis to 8 m); Oparanthus (3 spp.;
O. rapensis to 6 m, O. intermedius to 4 m). Campanulaceae: Sclerotheca
(S. margaretae to 8 m).

Austral Islands. Amaranthaceae: Charpentiera (C. australis to 11 m).

Rarotonga, Cook Islands. Asteraceae: Fitchia (F. speciosa to 10 m).
Campanulaceae: Sclerotheca (S. viridiflora to 2 m).

Henderson Island. Asteraceae: Bidens (B. hendersonensis + 1 var.,
to 7 m).

407

Insular Woodiness

Oeno Island. Asteraceae: Bidens (B. hendersonensis var. oenoensis
to 3m).

Pitcairn Island. Asteraceae: Bidens (B. mathewsii to 1 m).

Mangareva, Gambier Islands. Asteraceae: Fitchia (F. mangarevensis
(c@) 5) san),

Samoa. Goodeniaceae: Scaevola (S. nubigena to 2 m). Rubiaceae:
Sarcopygme (e, 5 spp.; at least two reach 7 m).

The total number of species showing insular woodiness in southern
Polynesia is not very great. Cyrtandra has been omitted from the listings
because none of the southern Polynesian species are exceptionally woody.
Cyrtandroidea (originally placed in Campanulaceae, Lobelioideae by
F’. B. H. Brown) is included because it is exceptional in height. Fitchia and
Oparanthus are interesting endemics of southern Polynesia (Carlquist,
1957; Carlquist and Grant, 1963); they are without any doubt related to
Bidens and may be considered an early introduction of a Bidens-like
stock into this region. Apetahia and Sclerotheca, also endemics of south-
eastern Polynesia, are capsular lobelioids presumably derived from Lobelia-
like ancestors. Some Bidens species in southern Polynesia are notably tall;
B. saintjohniana, a species from Marotiri, near Rapa, was not included
since it is a subshrub. Fuchsia cyrtandroides is included, although it is less
woody than F’. excorticata of New Zealand, to which it is related; it is,
however, still woodier than many species of Fuchsia. Sarcopygme is derived
from woody ancestors—perhaps Morinda—but because of its peculiar
rosette-tree habit (see Carlquist, 1965), it is judged to be increased in
height and altered in habit. The smaller number of examples of insular
woodiness in southern Polynesia may be due to limited land area (especially
limited area of wet forest), remoteness of the islands from one another,
and remoteness from source areas. The ecological gamut of the southern
Polynesian islands is not as great as that of the Hawaiian Islands.

MELANESIA

Data are from Guillaumin (1948), Krause (1912), and Leenhouts

(1964).

If we consider islands from Fiji westward to the Asiatic mainland,
except for Australia, New Zealand, and Taiwan, we find only one genus
we can clearly cite as an example of insular woodiness: Scaevola (sect.

408

Insular W oodiness

Scaevola, the fleshy-fruited species) of the Goodeniaceae. Scaevola flort-
bunda is a shrub or tree to about 8 m on Viti Levu, Fiji. On New
Caledonia there are also several species of comparable size—S. montana
and S. indigofera; the lowland S. beckii never exceeds about 1 m in height.
Additional Scaevola species, some of questionable taxonomic status, have
been reported from New Caledonia: S. racemigera, S. coccinea, S. cylin-
drica, and S. balansae. On Mt. Doorman, New Guinea, there is the recently
descnbed S. pauciflora Leenhouts. There are also in this series S. micrantha
(Palawan, Talaud Is., Mt. Kinabalu in Sabah) and the recently described
(Leenhouts, 1964) Borneo species, S. verticillata. There is, to be sure, a
variable species of Scaevola, S. oppositifolia (segregate species variously
recognized), which ranges from Queensland into New Guinea, Indonesia,
and the Philippines, but it belongs to sect. Enantiophyllum and is not
very woody, in any case. It tends to be scandent, and woody only at its
base. Brachionostylum (Asteraceae) is a monotypic genus reported from
New Guinea; it is a shrub close to (perhaps even congeneric with) Senecio,

The lack of insular woodiness in Melanesia probably represents the
harmonic nature of Melanesian floras compared to those of Polynesia.
Suitably large and wet areas are certainly present, and have been for long
periods of time. The reasons why Polynesia is so much more strongly

disharmonic than Melanesia have been discussed elsewhere (Carlquist,
1965).

NEW ZEALAND

Data are from Allan (1961).

Asteraceae: Brachyglottis (e, 1 sp. + vars.; tree to 6 m or more);
Cassinia (5 spp.; C. retorta to 5 m, others smaller); Olearia (32 spp.; shrubs
or trees 2-6 m); Senecio (40 spp.; 23 are shrubs or trees; S. stewartiae, S.
huntu, S. reinoldti—all from southernmost New Zealand—to 6 m or more);
Traversia (e, 1 sp.; shrub to 1 m). Caprifoliaceae (?): Alsewosmia (e, 1 sp.;
shrub to 2 m). Coriariaceae: Coriaria (7 spp.; C. arborea a shrub to 6 m,
branched from base). Malvaceae: Hoheria (e, 5 spp.; all trees to ca. 6 m);
Plagianthus (2 spp.; P. betulinus a tree to 15 m). Onagraceae: Fuchsia (4
spp.; F'. excorticata a tree to 12 m). Piperaceae: Macropiper (1 sp.; M.
excelsum to 6 m). Scrophulariaceae: Hebe (ca. 80 spp.; many are shrubs,
some trees, H. parviflora var. arborea to 7.5 m).

409

Insular Woodiness

A number of larger genera are only marginally includable in this
listing. Some that might have been included are Homalanthus polyandrus
(Euphorbiaceae), the only New Zealand species of this genus, a tree to
8 m; and Carmichaelia, a genus of Fabaceae endemic to New Zealand
and Lord Howe Island—C. arborea is a shrub or tree 3-5 m tall. Although
Hebe was included above, many species are diminutive shrubs, of course.
The taxonomic position of Alsewosmia may not be in Caprifoliaceae; it
has rayless wood, which suggests herbaceous ancestry.

The presence of an appreciable number of woodier herbs in New
Zealand reflects greater disharmony and greater ecological opportunity
than in Melanesia. Glaciation may have contributed to this disharmony.
One should also note that herbaceous groups are typically temperate, so
that New Zealand is a more likely site for their establishment than are
the tropical islands of Melanesia. The woodier herbs listed occur in pioneer
or more strongly disharmonic portions of New Zealand. Interesting in this
regard is that the most arboreal species of Senecio are characteristic of the
offshore islands. Coriaria appears on glaciated rubble, new lava, and road
cuts. With a few exceptions (e.g., Fuchsia), Australia seems to be the source
of most of the herbaceous groups that have become notably woody on New
Zealand. Some basically woody groups, such as Araliaceae, may have
increased in stature on New Zealand.

JUAN FERNANDEZ ISLANDS

Data are from Skottsberg (1922, 1953, 1956).

Apiaceae: Eryngium (4 spp.; E. bupleuroides with main stem to | m
long, E. fernandezianum to 0.75 m, others to 0.5 m). Asteraceae: Centauro-
dendron (e, 2 spp.; to 4 m); Dendroseris (including Phoenicoseris, Rea,
and Hesperoseris) (e, 9 spp.; D. macrantha to 5 m, others most typically
2m tall); Erigeron (5 spp.; E. fruticosus, to 1 m, is tallest); Rhetinodendron
(e; R. berterti to 5 m); Robinsonia (e, 7 spp, incl. Symphyochaeta; R.
thurifera to 6 m, others mostly to 4 m); Yunquea (e; Y. tenzti to 3 m).
Boraginaceae: Selkirkia (e; S. berteroi a “miniature tree” to 1 m). Cam-
panulaceae: Wahlenbergia (5 spp.; shrubs or subshrubs 0.5-1 m). Cheno-
podiaceae: Chenopodium (3 spp.; shrubs, C. nesodendron to 2.8 m).
Plantaginaceae: Plantago (P. fernandezianum to ca. 1 m).

Some questionable groups could have been added. Cuminia is an

410

Insular W oodiness

endemic genus of Lamiaceae, but its species are shrubs, and shrubby
Lamiaceae are common on mainland areas. Likewise, shrubby species of
Erigeron are certainly not lacking on continents. Fernandezian species of
Haloragis are not notably woody; Gunnera is not included (see discussion
earlier under Hawaiian flora). That Asteraceae bulk so large as examples
of woodier herbs is interesting. Three distinct tribes (Lactuceae, Cynareae,
and Senecioneae) are involved; four (Astereae) if Erigeron is included.
The good representation of woody herbs derives from the fact that the
Juan Fernandez Islands are temperate, suitable for establishment of tem-
perate weedy groups; the climate in the cloud forest is uniformly moist,
providing excellent growing conditions. Some lowland species (Dendroseris
litoralis, Chenopodium spp.) have achieved woodiness, perhaps by virtue
of succulence or drought resistance. The distance of the Juan Fernandez
Islands from South America is great enough to insure disharmony of this
flora, but close enough to permit arrival of Asteraceae and others. Despite
the presence of the apparently relictual genera Lactoris and Thyrsopteris,
there is every reason to believe that the groups listed above have achieved
woodiness autochthonously.

DESVENTURADAS ISLANDS

Data are from Skottsberg (1937).

Thamnoseris, an endemic genus with probably only one species, T.
lacerata, could be called an instance of insular woodiness, or a sort of
“giant lettuce” (Asteraceae, tribe Cichorieae, probably related to the
Fernandezian genus Dendroseris). However, Thamnoseris tends to be
acaulescent or have aspects of a stem succulent rather than a true shrub.
The only other Desventuradas plant worthy of mention is Lycapsus
tenuifolius (Asteraceae), a monotypic endemic genus; plants reach 70 cm.
The affinities of Lycapsus are uncertain; if a member of Heliantheae, it
certainly is not a plant of exceptional size for that tribe.

GALAPAGOS ISLANDS
Data are from Stewart (1911), Howell (1931, 1941), Harling (1962),
Dawson (1962), Anderson (1967), and Eliasson (1968).
Asteraceae: Darwiniothamnus (e, 2 spp.; to 3.5 m); Macraea (e; M.
laricifolia to 3.5 m); Scalesia (e; the tallest species, S. pedunculata, to

411

Insular Woodiness

20 m, S. cordata to 8 m; larger shrubs—S. atractyloides, S. darwinii, and
S. stewartii, to about 3 m). Cactaceae: Jasminocereus (3 spp.; to ca. 6 m);
Opuntia (O. echios, O. galapageia, and O. megasperma qualify as trees).
Portulacaceae: Calandrinia (C. galapagosa is an unusual succulent rosette
shrub to at least 60 cm).

The Galapagos flora includes only a few striking instances of arbores-
cence, of which the truly outstanding example is Scalesia pedunculata.
The remaining species of Scalesia range downward to relatively small
shrubs. Possible relatives of Scalesia include such genera of Heliantheae,
subtribe Verbesininae, as Wulffia, Wedelia, Perymenium, and Mirasolia.
Some of these are shrubby. Macraea is close to Wedelia, which can be a
shrub, so autochthonous increase in woodiness is not marked. Similarly,
Darwiniothamnus, although perhaps not as close to Erigeron as once
thought (Harling, 1962), is not exceptionally woody for Astereae when one
considers the arborescent species of Olearia in Australia and New Zealand.
Likewise, the endemic genus Lecocarpus (Asteraceae, Heliantheae, sub-
tribe Melampodiinae) is only a small shrub.

Visitors to the Galapagos Islands find the tree-Opuntias of lowland
areas striking. Likewise, Jasminocereus (a Cereus segregate of questionable
validity) can form candelabrum-shaped trees. However, the fact remains
that arborescent species of Opuntia and Cereus (or Cereus segregates) are
common in dry areas of Mexico, and Central and South America, and
Galapageian populations (e.g., O. echios var. zacana) can be low and
shrubby. The possibility remains that there has been some autochthonous
increase in arborescence among Galapagos Opuntias. In any case, the
Galapagos Islands are notable for the few instances in which woody herbs
have increased in size. The obvious explanation is the dry climate; only a
limited number of groups could achieve tree stature under these circum-
stances.

CALIFORNIA OFFSHORE ISLANDS

Data are from Thorne (1969).

Under the category of offshore islands are recognized islands that lie
off of both California and Baja California (Pacific coast). Some, such as
San Clemente and Guadalupe, are volcanic oceanic islands; others can be
called continental islands but have been populated at least in recent

412

Insular Woodiness

geological time by overwater dispersal (for a discussion, see Thorne, 1969).

Asteraceae: Coreopsis (C. gigantea to 2 m, on mainland also); Erio-
phyllum (1 insular sp., E. nevinii, to 1 m); Haplopappus (2 insular spp.,
H. canus and H. detonsus, to 1 m); Hemizonia (5 insular spp. [4 e], to
about 1 m; H. greeneana also on mainland); Perityle (P. incana [e, Guada-
lupe], to 1 m); Senecio (S. palmeri [e, Guadalupe], to 1 m; S. lyoni to 1 m,
on both mainland and islands); Stephanomeria (S. blairii |e, San Clemente],
to 2 m; S. guadalupense |e, Guadalupe], to 1 m). Brassicaceae: Erysimum
(2 insular spp., E. insulare and E. moranii, to 1 m). Fabaceae: Lotus (L.
scoparius var. dendroideus with single trunk, to 1.5 m). Papaveraceae:
Eschscholzia (E. palmeri and E. frutescens e on Guadalupe, to 1 m).
Portulacaceae: Talinum (T. guadalupense, to 1 m, e on Guadalupe).
Polygonaceae: Eriogonum (E. arborescens to 2 m; E. gigantewm, endemic
to islands, to 4 m). Rubiaceae: Galium (e; G. angulosum, to 1 m,
Guadalupe).

Climatic moderation obviously plays a strong role in permitting
greater size of these insular species as compared with their mainland
relatives. Larger leaves and shrub size can be seen in other genera, such
as Dendromecon, Cercocarpus, and Ceanothus, on these islands. Lavatera
of the Malvaceae should probably have been included in the list, although
present ranges of Lavatera species may be more restricted now than
formerly. This is certainly true of the rosaceous tree Lyonothamnus, now
an insular endemic but formerly known (as fossils reveal) from mainland
areas, and presumably surviving on islands only because of the maritime
climate. In addition, lack of herbivore pressure may have permitted survival
of some insular species. The degree of increase in size in the insular species
listed as compared with mainland relatives is not great, probably reflecting
that the moderated climate of the California offshore islands is only
slightly more favorable than that of the mainland—certainly not at all
spectacular.

CANARY ISLANDS
Data are from Lems (1960) and from original observations.
Amaranthaceae: Bosea (e; B. yervamora to 4 m). Asteraceae: Allago-

pappus (e, 1 or 2 spp.; to 2 m); Andryala (3 spp.; A. pinnatifida to 1 m);
Carlina (C. canariensis to 2 m, C. salicifolia to 3 m); Centaurea (11 spp.;

glk

Insular Woodiness

C. arbutifolia and some others to 4 m); Chrysanthemum (13 spp.; C.
broussenetii and others forming shrubs, to 2 m); Gonospermum (e, 4 spp.;
G. canariense to 4 m); Kleinia (K. nertifolia a succulent shrub to 2 m);
Schizogyne (e; S. sericea to 2 m); Sonchus (20 spp.; S. abbreviatus to 5 m,
others ranging downward in height); Tolpis (8 spp.; herbs to subshrubs 1
m); Vierea (e; V. laevigata a sprawling shrub to 1 m). Boraginaceae:
Echium (18 spp.; mostly shrubs 1-2 m tall; E. pininana a monocarpic
rosette tree to + m). Brassicaceae: Cheiranthus (3 spp.; shrubs to 1.5 m);
Crambe (7 spp.; some to 1 m); Descurainia (6 spp.; subshrubs to 1 m);
Parolinia (e, 1 sp.; to 4 m); Sinapidendron (2 spp.; shrubs to 1 m). Caryo-
phyllaceae: Dicheranthus (e, monotypic; to 1 m). Convolvulaceae: Con-
volyulus (8 spp.; some shrubs or fountain-shaped trees to 4 m, as in
C. floridus, C. scoparius). Crassulaceae: Aeonium (31 spp.; about 10
shrubby, to 1 m or more). Dipsacaceae: Pterocephalus (3 spp.; P. dume-
torum a shrub to 4 m). Euphorbiaceae: Euphorbia (8 spp. of sect.
Tithymalus; the common E. balsamifera to 2 m, E. regis-jubae to 3 m,
FE. mellifera to 5 m; cactoid spp. not listed because mainland species are
equally woody). Gentianaceae: Ixanthus (e, monotypic; I. viscosus to 1.5
m). Hypericaceae: Hypericum (5 spp.; some are shrubs to 2 m). Lamiaceae:
Bystropogon (3 spp.; shrubs to 3 m). Plantaginaceae: Plantago (15 spp.;
mostly rosette herbs, but P. arborescens and P. webbii “miniature trees”
to 1 m). Plumbaginaceae: Limonium (18 spp.; some are shrubs to 2 m—
L. arborescens, L. fruticans, L. preauxii—others smaller). Polygonaceae:
Rumex (2 spp.; R. lunaria a shrub to 2 m). Rosaceae: Bencomia (2 spp.;
trees to 4 m); Dendriopoterium (e; close to Bencomia, a tree to 4 m);
Marcetella (1 sp.; to 3 m). Scrophulariaceae: Isoplexis (2 spp.; to 1.5 m);
Lytanthus (L. salicinus, a shrub to 2 m). Solanaceae: Withania (3 spp.;
shrubs to 3 m). Urticaceae: Gesnouinia (e, 2 spp.; trees to 4 m).

In the Canarian flora, few typically herbaceous groups have achieved
anything like true arborescence. As the sizes above indicate, evolution
from herb to shrub is common, however. Interestingly, all the genera
endemic to the Canary Islands or to the Canary Islands plus one or more
other Macaronesian islands (Marcetella, Bencomia, Lytanthus, Isoplexis,
and Sinapidendron) are in the list, with the exception of two herbaceous
genera, Greenovia (Crassulaceae) and Legendraea (Convolvulaceae). Thus,
those groups in which woodiness has increased and which are endemic have
very likely evolved this characteristic autochthonously. If the endemic

414

Insular W oodiness

genera were relictual, one would expect at least some of them to be from
typically woody families. Only Lauraceae could be called relictual elements
in the Macaronesian flora, and even these must have arrived via long-
distance dispersal. Bencomia, Marcetella, and Dendriopoterium have been
included because they seem appreciably woodier than their closest relatives
(Poterium, Sanguisorba, Acaena). The affinities of Parolinia (Brassicaceae)
are not readily evident (Diceratella of Socotra, Somali Republic, and
Southern Iran?), but it appears woodier than any possible relatives. Literally
all zones of the Canarian flora contain expressions of increased woodiness—
least perhaps in alpine zones and laurel forest. For example, Brassicaceae
have not penetrated (as shrubs) into the laurel forest, and Aeonium is
absent (presumably because of frost susceptibility) in the alpine zone. ‘The
degree of increased woodiness in the laurel forest (Echium pininana,
Sonchus abbreyiatus, Ixanthus, Bencomia spp., Gesnouinia, Isoplexis) is
perhaps greater than one would expect, considering the limited area of
this forest (see chapter 5).

MADEIRA

Data are from Lowe (1868) and original observations.

Apiaceae: Tauschia (Melanoselinum) (M. decipiens, monocarpic, to
2 m); Tauschia (Monizia) (M. edulis, large-leaved, short-stemmed succu-
lent). Asteraceae: Andryala (2 spp.; A. crithmifolia a subshrub); Centaurea
(C. massoniana a shrub to 2 m); Sonchus (3 spp., probably with introgres-
sants: S. fruticosus a rosette tree to 5 m, S. pinnatus to 2 m, S. ustulatus to
0.6 m); Tolpis (4 spp.; T. succulenta a shrub to 1 m). Boraginaceae: Echium
(E. nervosum a shrub to 1.5 m, E. candicans a broad shrub to 3 m). Brassica-
ceae: Cheiranthus (3 spp.; shrubs to 1 m); Crambe (C. fruticosus a shrub
to 2 m); Sinapidendron (3 spp.; shrubs 1-2 m). Campanulaceae: Musschia
(e, 2 spp.; M. wollastonii a rosette tree to 4 m, M. aurea an acaulescent cliff
rosette herb). Convolvulaceae: Convolvulus (3 spp.; only C. massonii a
woody shrub). Euphorbiaceae: Euphorbia (E. piscatoria a shrub at sea level,
to 4 m in laurel forest). Rosaceae: Bencomia (B. caudata, also on Canaries,
to 4m).

The limited number of woody herbs on Madeira as compared with
the Canaries simply reflects that Madeira is a single island and has offered
relatively little “intramural” isolation for speciation to occur within the
genera; at least some of the species listed may be natural immigrants from

al

Insular Woodiness

the Canarian flora (Sonchus pinnatus, Bencomia caudata). Introgression
has occurred among species in several genera (Sonchus, Cheiranthus,
Sinapidendron), and some of Lowe’s species are probably only ecotypes.
The laurel forest contains the most spectacularly woody species—Sonchus
fruticosus, Echium candicans, Melanoselinum decipiens, and Musschia
wollastoni. The climate of Madeira is much like that of the Juan Fernan-
dez Islands (see fig. 10.1), and only the lack of land area and archipelagic
effects has favored Madeira less.

CAPE VERDE ISLANDS

Data are from Chevalier (1935).

Apiaceae: Tauschia (Melanoselinum) (5 spp.; range from an annual
to a short-stemmed herb, M. insulare). Asteraceae: Sonchus (S. gorgadensis
to 1 m; S. daltoni to 2 m). Boraginaceae: Echium (3 spp.; E. hypertropi-
cum a “small tree to 2 m” and E. vulcanorum a shrub to 2 m). Brassica-
ceae: Sinapidendron (5 spp.; less woody than the Madeiran species).
Euphorbiaceae: Euphorbia (E. tuckeyana to 1.5 m). Scrophulariaceae: Lyt-
anthus (L. amygdalifolius a decumbent shrub to 0.5 m).

AZORES

Data are from Palhinha (1966).

Apiaceae: Tauschia (Melanoselinum) (M. decipiens, also on Madeira).
Asteraceae: Tolpis (T. succulenta, also on Madeira).

The depauperate woody-herb floras of both the Cape Verde Islands
and the Azores are readily apparent, despite the fact that archipelagic
status might be thought to favor the Azores, which have laurel-forest and
heather belts. Remoteness, lack of high mountains, and lack of ecological
diversity may militate against a greater representation of the Macaronesian
woody herbs on the Azores. The Cape Verde Islands are relatively dry and
far from Mediterranean source areas, but their proximity to the Canaries,
and similarity in climate, may explain why there are more Canary-Madeira
elements on the Cape Verde Islands than on the Azores.

ST. HELENA

Data are from Melliss (1875).

Asteraceae: Aster (3 spp.; all shrubs 1.5-3 m); Commidendron (e,
monotypic; C. robustum a flat-topped tree to 7 m); Melanodendron (e,

416

Insular Woodiness

monotypic; M. integrifolium a spreading tree to 5 m); Petrobium *
(e, monotypic; P. arboreum a tree to 4 m); Psiadia (1 sp.; P. rotundifolia a
tree to 7 m); Senecio (1 sp., also called Pladaroxylon; a tree to 5 m). Cam-
panulaceae: Wahlenbergia (3 spp.; 2 shrubby to 1.25 m). Frankeniaceae:
Frankenia (1 sp.; subshrub to 0.75 m). Geraniaceae: Pelargonium (1 sp.;
P. cotyledonis an acaulescent cliff plant with large, gnarled base). Planta-
ginaceae: Plantago (P. robusta to 1 m). Rubiaceae: Hedyotis (1 sp.; H.
arborea to 7 m). Solanaceae: Mellissia (e, monotypic; M. begoniaefolia a
shrub to 3 m).

The preponderance of arboreal composites is evident. Commidendron
and Melanodendron are probably derived from Conyza-like ancestors. Pe-
trobium, despite its resemblances to Oparanthus (Carlquist, 1957), is prob-
ably a derivative of Bidens (to which Oparanthus may be independently re-
lated). Lobelia scaevolifolia, a St. Helena endemic, may be woodier than
mainland relatives, but subshrubby Lobelias do occur in both Africa and
South America. Hemsley (1885) discounts the “insular woodiness” of
Frankenia portulacaefolia on St. Helena, since there are shrubby Frank-
enias in Australia. However, the Australian Frankenias are probably paral-
lel instances of insular woodiness. The Hedyotis (like those in Hawaii)
seems likely to have increased in woodiness over its ancestors. The same
can be said for Mellissia if it is, in fact, a derivative of Physalis. The Pe-
largonium is included only because it forms a striking parallel to Monizia
(Apiaceae) on Madeira and Brighamia (Campanulaceae) on the Hawaiian
Islands. Possibly the St. Helena species of Phylica, including the segregate
genus Nesiota (Rhamnaceae), are larger than their ancestors. This also is
possible in Melhania (Sterculiaceae) on St. Helena. The abundance of
woody herbs there is amazing, considering the small area of the island.
Perhaps St. Helena is of relatively great age, as volcanic islands go.

WEST INDIES

Cuba. Data are from Alain (1953-1962).

Asteraceae: Ekmania (e, monotypic; to 3 m). Campanulaceae: Lobe-
lia (7 spp.; only L. cubana is shrubby, to 3 m). Euphorbiaceae: Euphorbia,
sect. Cubanthus (sect. endemic: 3 spp.; shrubs or small trees to 4 m); Eu-

* The correct name for this tree is Laxmannia, and conservation of Petrobium
has failed; however, Laxmannia continues in use for an Australian genus of Liliaceae.

alkyl

Insular Woodiness

phorbia sect. Euphorbiodendron (7 spp.; shrubs or small trees to 5 m).
Gesneriaceae: Rhytidophyllum (12 spp.; shrubs or small trees to 4 m; an
Antillean genus). Polygalaceae: Polygala (24 spp.; 1, also known as Phle-
botaenia cuneata, a tree to 15 m). Solanaceae: Espedaea (e, 1 sp.; E. amo-
ena a small tree); Henoonia (e, 2 spp.; shrubby).

Hispaniola.

Asteraceae: Narvalina (e, monotypic; shrub to 2 m); Piptocoma (e,
monotypic; shrub). Euphorbiaceae: Euphorbia (of several spp., only E. pett-
olaris a tree, to 7 m). Solanaceae: Coeloneurum (e, 2 spp.; to 3.5 m).

Puerto Rico. Data are from Little and Wadsworth (1964).

Asteraceae: Eupatorium (of native spp., E. portoricense a tree, to 7
m). Euphorbiaceae: Euphorbia (of native spp., E. petiolaris a tree to 7 m).
Malvaceae: Montezuma (e, monotypic; M. speciosissima a tree to 15 m).
Polygalaceae: Polygala (1 of native spp., also known as Phlebotaenia cowel-
lii, a tree to 10 m). Solanaceae: Goetzia (e; C. elegans a tree 8-10 m).

Virgin Islands. Data are from Little and Wadsworth (1964).

Euphorbiaceae: Euphorbia (of native spp., E. petiolaris a tree to 7 m;
extends from Hispaniola to the Lesser Antilles and the Venezuelan island,
Santa Margarita). Malvaceae: Sida (of native spp., S. eggersii, on Culebra,
Tortola, and Jost van Dyke; a tree to § m).

The arborescent species of the West Indies are notable, when com-
pared to those of other islands, for two reasons: the West Indian list is
relatively small and has few mesic species. If the Antilles are oceanic
islands with a relatively broad spread of ecological conditions, why should
this be true? Apparently the wet forest of the Antilles is relatively har-
monic. ‘The Antillean wet forest contains many of the same species seen
in mainland wet forests of comparable conditions. Thus, there is much
less opportunity for herbs to become arborescent during shift into wet
forest. Lowland forest, however, is relatively scrubby, and apparently gen-
era of dry habitats have become arborescent in this zone, as has happened
with Kokia, Chenopodium, Nothocestrum, and Nototrichium in dry low-
land Hawaiian forest. The West Indian woody genera listed seem, in some
cases, derivatives of littoral elements. Montezuma may be a relative of
Thespesia, and Sida is generally a coastal element. The four genera of Sola-
naceae, Goetzia, Espadaea, Coeloneurum, and Henoonia, belong to the
tribe Cestreae, as does the Hawaiian Nothocestrum.

418

Insular W oodiness

The low number of Asteraceae on West Indian islands—especially
arborescent species—is notable; herbaceous species are frequent on the
Antilles. Apparently ecological shift into wet forest is required for woodi-
ness, but the harmonic wet forest has not permitted this to occur.

MADAGASCAR AND COMORO ISLANDS

Data are from Choux (1934), Perrier de ]a Bathie (1950a, 1950b, 1960,

1962, 1963), and Rauh (1963).

Asteraceae: Apodocephala (8 spp.; 3 are shrubs and the remainder
trees 6-30 m tall); Brachylaena (5 spp. trees, to 40 m in B. merana); Di-
coma (4 spp. shrubs to trees); Oliganthes (4 spp. small trees to 6 m); Psi-
adia (28 spp.; mostly shrubby; P. altissima a small tree to 5 m); Senecio
(many native spp.; 4 spp. shrubs; S. humblotii and S. myricaefolius trees
to 15 m); Vernonia (of native spp., at least 9 are shrubs or small trees, most
to 5 m); Vernoniopsis (1 sp.; V. caudata to 8 m). Didiereaceae (e): Allu-
audia (6 spp.; lowest, A. dumosa, to 5 m; tallest, A. adscendens, to 12 m);
Alluaudiopsis (2 spp.; to 4 m); Decaryia (monotypic; shrub or small tree
to 6 m); Didierea (2 spp.; the taller, D. madagascariensis, to 6 m). Hyperi-
caceae (s.s.): Eliaea (e, monotypic; E. articulata a small tree); Haronga (e;
H. madagascariensis a “small tree’); Psorospermum (25 native spp.; 10
spp. trees to 10 m or more). Malvaceae: Hibiscus (H. mangindranensis a
tree to 30 m; 8 other spp. trees but shorter); Humbertianthus (e, mono-
typic; H. cardiostegius a tree); Macrostelia (e, 3 spp.; trees to 5 m or more);
Neohumbertiella (e, 3 spp.; shrubs or small trees). Turneraceae: Piriqueta
(3 spp. endemic to Madagascar; 2 are small trees).

This list cannot be considered exhaustive, because widespread genera
like Vernonia and Hibiscus are represented by arboreal species in regions
other than Madagascar, and we do not know the growth forms of the
stock ancestral to the Madagascan species. This statement certainly applies
also to the family Sarcolaenaceae (Chlaenaceae). However, there definitely
is increased woodiness in several genera. Brachylaena (Asteraceae) is rep-
resented in South Africa and Madagascar, but the Madagascan species
include the tallest for the genus; B. merana may even be the largest tree-
composite in the world. Other genera of Asteraceae could have been added
to the list; the preponderance of increased woodiness among Asteraceae

419

Insular Woodiness

is not surprising in view of similar tendencies by this family on other
islands.

Didiereaceae are exceptionally interesting, because the family is en-
demic to Madagascar, and it is difficult to imagine that it did not evolve
woodiness autochthonously there. Didiereaceae belongs to Caryophyllales
(Centrospermae), and its closest affinities appear to lie with such herba-
ceous families as Portulacaceae. Didiereaceae also shows the less frequent
phenomenon of increased woodiness in a dry area rather than in a rain-
forest area.

The fact that insular woodiness occurs in Madagascar is curious, be-
cause the flora seems so large that one would expect it to be relatively
harmonic. Evidently it is not, or was not at times when rather invasive
groups such as Asteraceae established there. The Madagascan wet forest,
one must remember, is rather distant from moist forest areas on conti-
nents, so disharmony is not unexpected. Even if once connected to Africa,
Madagascar has been isolated for such a long time that extinction of some
elements and establishment of other new ones have undoubtedly taken
place, and radiation among the new establishments has been inevitable.

SOCOTRA

Data are from Balfour (1888).

Apiaceae: Nirarathamnos (e, monotypic; N. asarifolius a small tree).
Cucurbitaceae: Dendrosicyos (e, monotypic; D. socotrana to 4 m); Mora-
ceae: Dorstenia (1 sp.; D. gigas, to 2 m).

The listing may be misleading, because some remarkable arborescent
plants other than these occur on Socotra. For example, the bizarre stem
succulent Adenium sokotranum (Apocynaceae) and the giant liliaceous
tree Dracaena cinnabari deserve mention, but these genera have equally
arborescent counterparts elsewhere. Nirarathamnos is probably only an
unusual segregate of Bupleurum. Dendrosicyos does seem unique (except
for Acanthosicyos) as a water-tank plant in a family of vines. Stem succu-
lence, rather than simple arborescence, is the hallmark of Socotra’s larger
and more conspicuous plants. As we have seen, Socotra’s climate (fig. 10.1)
is not one that would permit arborescence except for stem succulents. The
Didiereaceae of dry Madagascar offer a good parallel.

420

Insular Woodiness

MASCARENE ISLANDS (MAURITIUS, REUNION, RODRIGUEZ)

Data are from Baker (1877).

Asteraceae: Faujasia (3 spp.; shrubs to 2 m); Monarrhenus (2 spp.;
shrubs to 1.5 m). Campanulaceae: Heterochaenia (e, monotypic; H. ensz-
folia a shrub to 80 cm). Turneraceae: Mathurina (e, monotypic |Rodri-
guez|; M. penduliflora a small tree).

Additional woody herbs that could be cited include Cylindrocline
(endemic, monotypic; Asteraceae, tribe Inuleae) and shrubby Helichry-
sums. Faujasia (endemic to Madagascar and the Mascarenes) is obviously
close to Senecio, and Monarrhenus (endemic to Madagascar and the Mas-
carenes) perhaps is close to Helichrysum. Heterochaenia seems a derivative
of Campanula, and Mathurina is apparently allied to the widespread genus
Piriqueta. Thus, the list of plants woodier on the Mascarenes (excluding
Madagascar) is not great, nor are these plants exceptionally woody or dis-
tinct from their closest relatives. The impoverishment of this list recalls
that of southern Polynesia, and the reasons may be similar—the land areas
may be too small to have fostered evolution and persistence of more
numerous arborescent taxa, or the Mascarene flora may be relatively har-
monic. I cannot claim any clear instances of insular woodiness for the
Seychelles. Here, other elements such as palms seem to have preempted
niches into which waif flora derivatives might have evolved.

By way of concluding the listings for insular areas, one may note that
tree ferns and palms are, in fact, abundant on many of the islands, at least
the wetter ones (see Holttum [1964] on Cyathea in the Pacific). In New
Zealand, the pioneering qualities of tree ferns are clearly evident from
their appearance in overgrown pastures and other disturbed sites. In New
Zealand, and elsewhere, the waif flora that has evolved arborescence tends
to be represented more often in semi-open areas of wet forest. This may
not be typical for the Hawaiian flora, but it is true of some arborescent
genera there, Clermontia and Scaevola, for example.

AUSTRALIA

Australia—or parts of it—can qualify as an island or islands in several
respects. Many taxonomic groups have probably increased in woodiness in
Australia. The most obvious is Eucalyptus. Because this genus contains
the tallest tree species in angiosperms, one can readily conclude that euca-

421

Insular Woodiness

lypts represent some autochthonous gigantism in Myrtaceae, despite the
fact that this family is woody in any case. Eucalypt seeds at their smallest
seem readily wind dispersed; ancestors of the eucalypts may have reached
Australia in this fashion at a time when the flora was highly disharmonic
and, by virtue of the amazing drought-tolerance of this genus, radiated
widely throughout the continent.

Aside from major floristic elements such as the eucalypts, however,
there are some genuine instances of insular woodiness in special areas of
Australia. The Tasmania-Victoria region summarized below is one. Com-
paratively recent disharmony in this area may be related to glaciation,
opening up large areas to temperate plants. The relatively moist climate
of this region provided opportunity for ecological shift into the wet forest,
but wet forest moderated by maritime climates. Thus the Tasmania-Vic-
toria region provided opportunities for insular woodiness not unlike
those in New Zealand or even the Juan Fernandez Islands, which are
rather cool. Data on the Tasmania-Victoria plants listed here are from
Curtis (1956-1967) and Ewart (1930).

Asteraceae: Bedfordia (2 spp.; B. salicina a tree to 10 m, B. linearis
somewhat smaller); Cassinia (numerous spp.; most notably arborescent
are C. longifolia and C. trinervia, to 7 m); Helichrysum (spp. formerly
known under the generic name Ozothamnus, now a section, are notably
shrubby; at least 7 spp. reach 3 m); Olearia (many spp. prominently woody;
most notable is O. argophylla, 5-15 m tall, but others would qualify as
trees); Senecio (S. centropappus a tree 2-4 m tall, on Tasmania). Goode-
niaceae: Goodenia (G. ovata, an understory shrub to 2 m). Lamiaceae:
Prostanthera (P. lasianthos a small tree to 6 m, P. melissifolia to 4 m, P.
rotundifolia to 5 m, others large- to medium-sized shrubs); Westringia (W.
angustifolia a shrub to 3 m, other species smaller). Solanaceae: Solanum
(many shrubby species; S. laciniata, to 3 m, may be tallest).

One could add many other species from these and other families, and
the listing may be considered a token one. Lamiaceae have been neglected
in listings of insular areas because of the frequency of shrubs on mainland
areas in this family; Prostanthera is clearly exceptional within the family
in size, however. The entire family Gyrostemonaceae might be cited as an
example, for some species of this family are trees; ancestors of Gyrostemo-
naceae appear likely to have been herbaceous or nearly so. Goodeniaceae,

422

Insular W oodiness

a basically herbaceous family, and strongly represented in Western Aus-
tralia, might be cited in greater numbers: Scaevola spinescens and other
species of Scaevola, Leschenaultia, and Verreauxia might be listed (see
Krause, 1912). Stylidiaceae form an interesting example; the only woody
species of this family (and these are only slightly woody by virtue of anoma-
lous secondary thickening) are Stylidium laricifolium (New South Wales,
to 2 m), S. glandulosum (known with certainty only from the Recherche
Archipelago off southwestern Australia) and S. imbricatum (Stirling Range,
W.A.)—all relatively mesic or maritime localities (Erickson, 1958). One
could add some exceptionally woody species of basically woody groups:
for example, Richea pandanifolia from ‘Tasmania forms trees 2-12 m tall,
and other Epacridaceae from this region could be cited. In Loranthaceae,
the tree-mistletoe Nuytsia floribunda of Western Australia, which has
anomalous secondary thickening, clearly deserves mention.

SOUTH AMERICA AND AFRICA

Chapter 14 is devoted to the vegetation of equatorial uplands in
South America and Africa, which form a special case. In chapter 14, atten-
tion is paid not only to the alpine belts, which contain the famous rosette
trees Espeletia and Senecio sect. Dendrosenecio, but also the subadjacent
vegetational belts which exhibit part of the same program of insular woodi-
ness.

However, medium elevations south of the equator—particularly in
Peru—constitute an area much like the true insular areas cited above. The
uplift of the Andes is relatively recent, so that it represents a pioneer area.
Certain herbaceous groups have been able to establish by virtue of good
dispersibility and become woodier under the equable, cool conditions.
These represent temperate groups; apparently, tropical groups, as in the
equatorial alpine and subalpine situations, are not able to adapt so readily.
The roster of herbaceous groups with woody representatives is strikingly
like that for oceanic islands, and is presented here to demonstrate that
secondary woodiness is a very real phenomenon on particular continental
areas as well. The listing below has been compiled largely from MacBride’s
(1937 et seq.) “Flora of Peru.”

Apiaceae: Eremocharis (E. piscoensis and E. triradiata to 2 m). Aster-
aceae: many genera and species, including Flotovia, Gynoxys, Senecio.

423

Insular Woodiness

Begoniaceae: Begonia (B. peruviana to 2.5 m). Campanulaceae: Burmei-
stera (several spp. shrubs to 2 m); Centropogon (several spp. 2-3 m; C.
densiflorus “‘arborescent, 4-8 m’”’); Siphocampylus (several spp. shrubs to
4 m). Caricaceae: Carica (several spp., some are sparsely branched, small
trees); Jacaratia (J. digitata, “‘tall tree’). Convolvulaceae: Ipomoea (I. calo-
dendron a tree 4-10 m); Maripa (several spp., “lianas when unsupported be-
coming trees”). Euphorbiaceae: Acalypha (A. macrostachya a tree to 12
m; A. samydaefolia to 10 m; several others small trees); Euphorbia (E.
tessmannti a tree to 3 m). Flacourtiaceae (s.l.): Cochlospermum (C. ort-
nocense “‘a large tree”). Gentianaceae: Lisianthus (L. revolutus a small
tree to 5 m, L. corymbosus to 4 m, L. ovalis 2-3 m); Macrocarpaea (M.
pachystyla “‘arborescent,” M. guttifera a shrub 2-3 m). Hydrophyllaceae:
Wigandia (W. crispa to 6 m; W. urens to 4 m). Lythraceae: Cuphea
(many shrubby spp., such as C. cordata, “most frequent shrub of the E.
Andes between 1800-2500 m’’). Malvaceae: Abutilon (A. arboreum to
6 m). Plantaginaceae: Plantago (P. sericea a subshrub). Polemoniaceae:
Cantua (at least 6 spp. to 3 m or more); Huthia (2 spp.; to 1 m). Solana-
ceae: Cestrum (many shrubby spp., notably C. longiflorum, “known to
attain 8 m,” and C. laevifolium, a “tree about 4.5 m”); Cyphomandra
(C. endopogon a tree to 7 m; C. pendula to 5 m; others small trees);
Nicotiana (N. tomentosa to 7 m, N. setchellii to 5 m, others small trees);
Solanum (many arboreal and arborescent spp., such as S. verbascifolium,
10 m or taller). Valerianaceae: Valeriana (several shrubby spp). Violaceae:
Hybanthus (H. sprucei a small shrub); Viola (several shrubby spp.).

A number of additional examples could be cited: perhaps some species
of Heliotropium (Boraginaceae), Croton (Euphorbiaceae), Salvia (Lamia-
ceae), or the giant Puya, P. raimondii (Bromeliaceae).

To lesser degrees, other continental areas show insular woodiness and
serve to remind us that secondary increase in xylem production occurs in
particular situations. Ipomoea arborea of the Mexican highlands would
certainly qualify, as would the woody Asteraceae of that region.

?

References

Alain, H. 1953-1962. Flora de Cuba, vols. 3—5. Vol. 3, 1953; vol. 4, 1957; Havana,
Imp. P. Fernandez y Cia. Vol. 5, 1962, Rio Piedras, Universidad de Puerto Rico.

424

Insular W oodiness

Allan, H. G. 1961. Flora of New Zealand, vol. 1. Wellington, R. E. Owen, Govern-
ment Printer.

Anderson, E. F. 1967. An account of my cactus-collecting trip to the Galapagos
Islands. Cactus Succulent Jour. 39:186-99.

Baker, J. G. 1877. Flora of Mauritius and the Seychelles. London, L. Reeve & Co.

Balfour, I. B. 1888. Botany of Socotra. Trans. Roy. Soc. Edinburgh 31:1-446.

Barghoorn, E. S. 1941. The ontogenetic and phylogenetic specialization of rays in
the xylem of dicotyledons. III. The elimination of rays. Bull. Torrey Bot. Club
68:317-25.

Blumenstock, D. I. 1961. Climates of the states. Hawaii. U.S. Dept. of Commerce
Weather Bureau. Washington, D.C., U.S. Government Printing Office. 20 p.

Brown, F. B. H. 1935. Flora of southeastern Polynesia. III. Dicotyledons. Bishop
Mus. Bull. 130:1—386.

Carlquist, S. 1957. The genus Fitchia (Compositae). Univ. Calif. Publ. Bot. 29:1-144.

Carlquist, S. 1958. Wood anatomy of Heliantheae (Compositae). Trop. Woods
108:1-30.

Carlquist, S. 1961. Comparative plant anatomy. New York, Holt, Rinehart &
Winston.

Carlquist, S. 1962a. A theory of paedomorphosis in dicotyledonous woods. Phytomor-
phology 12:30—-45.

Carlquist, S. 1962b. Wood anatomy of Senecioneae (Compositae). Aliso 5:123-46.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1966. Wood anatomy of Compositae: a summary, with comments on
factors controlling wood evolution. Aliso 6(2):25-44.

Carlquist, S. 1969a. Wood anatomy of Lobelioideae (Campanulaceae). Biotropica
1(2):47-72.

Carlquist, S. 1969b. Wood anatomy of Goodeniaceae and the problem of insular
woodiness. Ann. Missouri Bot. Gard. 56:358—90.

Carlquist, S. 1970a. Hawaii, a natural history. New York, Natural History Press.

Carlquist, S. 1970b. Wood anatomy of Echium (Boraginaceae). Aliso 7(2):183-99.

Carlquist, S. 1970c. Wood anatomy of insular species of Plantago and the problem
of raylessness. Bull. Torrey Bot. Club 97:353-61.

Carlquist, S. 1970d. Wood anatomy of Hawaiian, Macaronesian, and other species
of Euphorbia. Biol. Jour. Linnaean Soc. 63 (suppl. 1):181-93.

Carlquist, S. 1971. Wood anatomy of Macaronesian and other Brassicaceae. Aliso
7(3):365-84.

Carlquist, S., and M. L. Grant. 1963. Studies in Fitchia (Compositae): novelties from
the Society Islands; anatomical studies. Pacific Sci. 17:282-98.

Chevalier, A. 1935. Les iles du Cap Vert. Flore de l’archipel. Rev. Bot. Appl. 15:733-
1090.

Choux, P. 1934. Les Didiéréacées, xérophytes de Madagascar. Mém. Acad. Mal-
gache 18:1-71.

Christophersen, E. 1938. Flowering plants of Samoa—II. Bishop Mus. Bull. 154:1-
Vide

425

Insular W oodiness

Curtis, W. M. 1956, 1963, 1967. The student’s flora of Tasmania, vols. 1-3. Hobart,
Tasmania, Government Printer.

Dale, R. F. 1959. Climates of the states: California. U.S. Dept. of Commerce
Weather Bureau. Washington, D.C., U.S. Government Printing Office. 37 p.
Darwin, C. 1859. On the origin of species by means of natural selection. Reprint

of first edition, 1950. London, Watts & Co.

Dawson, E. Y. 1962. Cacti of the Galapagos Islands and of coastal Ecuador. Cactus
Succulent Jour. 34:67—74, 99-105.

Degener, O. 1937. Euphorbiaceae. In O. Degener, Flora Hawaitiensis (parts pub-
lished variously by the author).

Eliasson, U. 1968. Studies in Galapagos plants. VI. On the identity of Calandrinia
galapagosa St. John. Svensk Bot. Tidskr. 62:365-68.

Erickson, R. 1958. Triggerplants. Perth, Paterson Brokensha.

Ewart, A. J. 1930. Flora of Victoria. Melbourne, Government Printer.

Guillaumin, A. 1948. Flore analytique et synoptique de la Nouvelle-Calédonie:
Phanérogames. Paris, Office de la Recherche Scientifique Coloniale.

Harling, G. 1962. On some Compositae endemic to the Galapagos Islands. Acta
Horti Bergiani 20(3):63-120.

Hedberg, O. 1957. Afroalpine vascular plants. A taxonomic revision. Symb. Bot.
Upsalienses 15(1):1-411.

Hemsley, W. B. 1885. Report on present state of knowledge of various insular
floras. Rep. Sci. Results Voyage H.M.S. Challenger, Bot. 1:1-—75.

Hillebrand, W. 1888. Flora of the Hawaiian Islands. London, Williams & Norgate.

Holttum, R. E. 1964. The tree ferns of the genus Cyathea in Australasia and the
Pacific. Blumea 12:241-74.

Howell, J. T. 1931. Cacti in the Galapagos Islands. Cactus Succulent Jour. 5:513.

Howell, J. T. 1941. The genus Scalesia. Proc. Calif. Acad. Sci., 4th ser., 22(11):221-
a.

Johnston, I. M. 1953. Studies in the Boraginaceae. XXV. A revaluation of some
genera of the Lithospermeae. Jour. Arnold Arboretum 34:258-99.

Krause, K. 1912. Goodeniaceae. In Das Pflanzenreich, ed. A. Engler, 54:1—207.
Berlin, Akademie Verlag.

Kribs, D. A. 1935. Salient lines of structural specialization in the wood rays of
dicotyledons. Bot. Gaz. 96:547—57.

Leenhouts, P. W. 1964. A new montane Scaevola from Borneo: S. verticillata
(Goodeniaceae). Blumea 12:317-18.

Lems, K. 1960. Floristic botany of the Canary Islands. Sarracenia 5:1—-94.

Lems, K. 1961. Botanical notes on the Canary Islands. III. The life form spectrum
and its interpretation. Ecology 42:569-72.

Lems, K., and C. M. Holzapfel. 1968. Evolution in the Canary Islands. I. Phylo-
genetic relations in the genus Echium (Boraginaceae) as shown by trichome
development. Bot. Gaz. 129:95-107.

Little, E. L., Jr., and F. H. Wadsworth. 1964. Common trees of Puerto Rico and
the Virgin Islands. U.S.D.A. Forest Service Agricultural Handbook no. 249.
Washington, D.C.

426

Insular W oodiness

Lowe, R. T. 1868. A manual flora of Madeira. London, John Van Voorst.

MacBride, J. F. (and collaborators). 1937 et seq. Flora of Peru. Field Mus. Natur.
Hist. Publ., bot. ser., vol. 13 (numerous parts, currently not complete).

Melliss, J. C. 1875. St. Helena. London, L. Reeve & Co.

Meusel, H. 1952. Uber Wuchsformen, Verbreitung, und Phylogenie der einiger
mediterran-mitteleuropaischen Angiospermen Gattungen. Flora 139:333-93.
Moore, J. W. 1933. New and critical plants from Raiatea. Bishop Mus. Bull.

102:1-53.

Palhinha, R. T. 1966. Catdlogo dos plantas vasculares dos Acores. Lisbon, Sociedade
de Estudos Acorianos, Alfonso Chaves.

Perrier de la Bathie, H. 1950a. Turneraceae. In Flore de Madagascar et des Comores,
ed. H. Humbert. Paris, Musée National d’Histoire Naturelle.

Perrier de la Bathie, H. 1950b. Hypericacées, Guttiféres. In Flore de Madagascar et
des Comores, ed. H. Humbert. Paris, Musée National d’Histoire Naturelle.

Perrier de la Bathie, H. 1960. Composées, vol. I. In Flore de Madagascar et des
Comores, ed. H. Humbert. Paris, Musée National d’Histoire Naturelle.

Perrier de la Bathie, H. 1962. Composées, vol. II. In Flore de Madagascar et des
Comores, ed. H. Humbert. Paris, Musée National d’Histoire Naturelle.

Perrier de la Bathie, H. 1963. Composées, vol. III. In Flore de Madagascar et des
Comores, ed. H. Humbert. Paris, Musée National d’Histoire Naturelle.

Rauh, W. 1963. Didiéréacées. In Flore de Madagascar et des Comores, ed. H. Hum-
bert. Paris, Musée National d’Histoire Naturelle.

Raunkiaer, C. 1934. The life forms of plants and statistical plant geography. Oxford,
Clarendon Press.

Ray, P. M., and W. E. Alexander. 1966. Photoperiodic adaptation to latitude in
Xanthium strumarium. Amer. Jour. Bot. 53:806-16.

Rikli, M. 1912. Lebensbedingungen und Vegetationsverhdltnisse der Mittelmeer-
lande und der Atlantischen Inseln. Jena, Verlag Kramer.

Rock, J. F. 1913. The indigenous trees of the Hawaiian Islands. Honolulu, privately
published.

Rock, J. F. 1919. A monographic study of the Hawaiian species of the tribe
Lobelioideae, family Campanulaceae. Mem. Bishop Mus. 7(2):1-394.

Schenck, H. 1907. Beitrage zur Kenntniss der Vegetation der Kanarischen Inseln.
Wiss. Ergebn. Deutsch. Tiefsee-Exped. Dampfer “Valdivia” 1898-1899 2(1:2):
225-406.

Sherff, E. E. 1935. Revision of Tetramolopium, Lipochaeta, Dubautia and Rail-
liardia. Bishop Mus. Bull. 135:1-136.

Sherff, E. E. 1937a. The genus Bidens. Part I. Field Mus. Natur. Hist. Publ., bot.
ser., 16:1—346.

Sherff, E. E. 1937b. Some Compositae of southeastern Polynesia (Bidens, Coreopsis,
Cosmos, and Oparanthus). Occas. Papers Bishop Mus. 12(19):1-19.

Skottsberg, C. 1922. The phanerogams of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:95-240.

Skottsberg, C. 1937. Die Flora der Desventuradas-Inseln. Acta Horti Gotoburg.
5(6): 1-88.

427

Insular W oodiness

Skottsberg, C. 1953. The vegetation of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:793-960.

Skottsberg, C. 1956. Derivation of the flora and fauna of Juan Fernandez and Easter
I. Natur. Hist. Juan Fernandez Easter I. 1:193-438.

Stewart, A. 1911. A botanical survey of the Galapagos Islands. Proc. Calif. Acad. Sct.,
4th ser., 1:7—288.

Thorne, R. F. 1969. The California Islands. Ann. Missouri Bot. Gard. 56:391-408.

Wallace, A. R. 1878. Tropical nature and other essays. London, MacMillan & Co.

Walter, H., H. Lieth, and H. Rehder. 1967. Klimadiagramm-W eltatlas. Jena, Ver-
lag Fischer.

Werner, K. 1966. Die Wuchsformen der Gattungen Isoplexis (Lindl.) Benth. und
Digitalis L. Bot. Jahrb. 85(1):88-149.

Wimmer, F. E. 1953. Campanulaceae-Lobelioideae. Teil II. Das Pflanzenreich IV,
276b (107 Heft). Berlin, Akademie Verlag.

Wimmer, F. E. 1956. Campanulaceae-Lobelioideae. Teil I. Das Pflanzenreich IV,
276b (106 Heft). Berlin, Akademie Verlag.

Zimmerman, E. C. 1948. Insects of Hawaii, vol. 1: Introduction. Honolulu, Univer-
sity of Hawaii Press.

428

LOSS OF DISPERSIBILITY
IN ISLAND PLANTS

While working with evolutionary phenomena in the Hawaiian Islands
and other Polynesian islands, I observed types of fruits and seeds that
seemed incongruously poor at dispersal. The floras of these islands must
have arrived by long-distance dispersal, yet during evolution on the island
areas various groups of plants have lost dispersibility. My recognition of
this fact stemmed not merely from observation of unexpected fruit and
seed types and comparison of these with source-area relatives, but from
cognizance of flightless insects and birds on islands. Although I did not
deliberately attempt to find a counterpart in plants, my observations soon
indicated that a parallel phenomenon did exist, and for much the same
reasons.

At the outset, I wish to stress that loss of dispersibility is by no means
limited to islands. More striking examples may, in fact, occur in certain
mainland situations. If one takes the broadest possible view, large-seeded
forest trees such as mangos (Mangifera indica), avocados (Persea ameri-
cana), and a host of others are adaptations in which the large size of seeds
represents a definite loss of dispersibility. Although botanists are familiar
enough with the fact that in any given family a variety of dispersal types
may exist, they sometimes do not express clearly that not only dispersal
mechanisms but dispersal capability shows evolutionary change. Some

NOTE: This chapter is a revised version of two papers (Carlquist, 1966a, 1966b). For
permission to present portions of these papers, acknowledgment is extended to
Evolution and Brittonia. Preliminary description of these concepts may also be
found in my book Island life (1965).

429

Loss of Dispersibility in Island Plants

changes in dispersibility yield marked improvement in capability for cross-
ing long distances of land, water, or both. Some show retrograde tendencies.

There is no reason to believe that only a single tendency occurs within
a given group. In a particular genus, therefore, we may have difficulty in
saying whether ease of dispersibility is primitive or specialized, and whether
both trends may prevail. For example, Sauer (1964) includes data on
permeability and buoyancy of seeds in his monograph of Canavalia
(Fabaceae). The vast majority of species have seeds that are nonbuoyant
and permeable by water. Are they all examples of loss of dispersibility?
Probably not. One can envision that Canavalia might have evolved littoral
species that, as part of their adaptation to growth near shores, developed
increasingly buoyant and impermeable seeds. One or more such maritime
species may have reached other beaches in the world, and lost capability
for seawater dispersal after establishment.

The endemic Hawaiian species of Canavalia (sufficiently distinctive
to constitute a subgenus, Maunaloa) offer no problem, however, in inter-
pretation. All five of these species have large, nonbuoyant, permeable
seeds (Sauer, 1964). Because 12 species of Canavalia have seeds that are
buoyant and/or impermeable, arrival of Canavalia on the Hawaiian Islands
by oceanic drift seems likely. One suspects, however, that the stock of
Canavalia that gave rise to the endemic species did not arrive repeatedly
by drift, but only on a single occasion, or at most a limited number of
times. If this were true, the Canavalia stock could undergo evolutionary
shift into inland sites and concomitantly lose buoyancy and permeability.
If continually reintroduced, a maritime Canavalia would probably swamp
out ecological shift and prevent the accompanying loss of dispersibility via
oceanic drift. Thus, isolation is required for loss of dispersibility—not
merely in Canavalia, but in all other instances cited as examples below.

Ecological shift does seem inextricably related to loss of dispersibility.
For this reason, and because my field experience has been more extensive
in the Hawaiian Islands than elsewhere, chiefly Hawaiian examples are
cited below. The Hawaiian flora appears to have been derived primarily
from moderately dry regions, and the wet Hawaiian forest seems composed
chiefly (but certainly with exceptions) of plants that have evolutionarily
adapted to more mesic habitats. This ecological shift, so pervasive in
instances of loss of dispersibility on islands, is possible in a region with

430

Loss of Dispersibility in Island Plants

such abundant areas of rain forest (and relatively great duration of time—
5 million years, perhaps more) as the Hawaiian Islands offer. If ecological
shift is a prime accompaniment to loss of dispersibility, one can then
understand why one would see this phenomenon on the Hawauan and
other Pacific islands, but not to any appreciable degree on the Galapagos
or the Canary Islands (or other Macaronesian islands). ‘These latter
archipelagos have no true rain forest—only, at best, a cloud forest. Also,
equatorial alpine plants would not be expected to show loss of dispersibility.
These are plants of temperate (usually temperate alpine) climates, and
their occurrence in an equatorial region represents a minimal change:
adaptation to a climate that is still temperate alpine, but with a prolonged
growing season (the entire year, provided a species has resistance to the
nightly frost at higher altitudes on equatorial peaks),

I must emphasize that the observations offered here are limited in
nature, and the student of island biology can uncover many interesting
facts I have not investigated. For example, shortened viability is almost
certainly involved in loss of dispersibility, and tests of viability of species
of wet Hawauan forest in comparison to their non-Hawaiian relatives
would probably enlarge appreciably the number of examples cited here.
In the case of species hypothesized as having been introduced to islands
by oceanic drift, I tested only buoyancy. Whether these species also show
loss of resistance to seawater (permeability) has not yet been determined
except in a few obvious cases, such as the Seychellean “double coconut,”
Lodoicea, known from beaches all over the world yet never found in a
viable condition. Another possible means of loss of dispersibility concerns
the nature of seeds or the endocarp of fruits in bird-dispersed species.
Typically, for successful transport by birds one would expect that the
seed coat or fruit endocarp would be altered little, or only to the point of
permitting more reliable germination by action of their digestive tract.
In loss of dispersibility, possibly the seeds may be more vulnerable to
destruction by action of the digestive tract of birds. Other factors not
investigated here that would be criteria of loss of dispersibility would
include loss of resistance to extremes of cold, heat, and desiccation.
Fertility of a plant may be expected to alter under insular conditions. I
suspect that in some insular species the fertility is lowered to whatever
point maintains the species in the area of ecological conditions suitable to

431

Loss of Dispersibility in Island Plants

its adaptation. Because of the numerous variables, fertility is dificult to
measure, but suitable measures probably would show a lessening in some
insular groups. The results of Sohmer (1972) in Charpentiera on the
Hawauan Islands point clearly to lowered fertility. Also, if a species has
fewer, larger fruits than its ancestors, and if seed number is proportionately
lowered per plant, this could be said to constitute loss of dispersibility. The
trend to fewer seeds per plant can be seen in Hawaiian species of
Zanthoxylum (= Fagara), Plantago (P. princeps), Breweria, and Stenogyne.

One need not hypothesize complete loss of dispersibility, however,
and perhaps I should have termed this phenomenon “‘lessened dispersi-
bility” rather than “loss.” Obviously, some species that have lost dis-
persibility on islands still can disperse within an island and within
an archipelago. An example is Bidens sandyicensis (Asteraceae), which
certainly shows diminished dispersibility (fig. 11.3), but which occurs
nonetheless on most major islands in the chain—Kauai, Oahu, Maui, and
Hawaii.

Hypotheses for Loss of Dispersibility

The operative factors that result in loss of dispersibility are several and
may differ in various cases. They cannot, therefore, be rated as to which
is more, which is less important in the instances as a whole. One factor
of probable importance may be described as precinctiveness, a tendency
noted by Beccari and Rock (1921) in the palm genus Pritchardia. Briefly
defined, precinctiveness denotes reproduction in such a fashion that most
propagules are shed within a limited zone of stable conditions, such as
wet forest, to which a species has a definitive and narrow degree of
adaptation. Potentially, this permits survival of more numerous propagules
than would random, widespread dissemination into a variety of habitats.
One can assume that species adjust their reproductive abundance to a
rate suitable for occupation of favorable sites. In strand plants and weedy
species, reproductive rates may be high because longevity of individuals
may be shorter, and because favorable sites are widely scattered and change
rapidly. In species of stable forest, a lower reproductive rate would be
expected to suffice.

432

Loss of Dispersibility in Island Plants

An extreme function of precinctiveness would be the tendency for
lowered dispersibility to prevent the majority of propagules from being
blown away from an island. Excessive loss of propagules seems unlikely in
Hawaiian flowering plants, where most species with lowered dispersal
ability occupy not windy sites, but relatively protected places in forests.
Likewise, wind pressure seems insufficient to explain development of flight-
lessness in Hawaiian insects (Perkins, 1913). Wind pressure is, however, a
very real factor for plants with minute propagules, such as ferns.

A second probable factor in loss of dispersibility is the tendency for
immigrants to oceanic islands to be originally weedy plants, capable of
establishing in pioneering conditions, and for these immigrants then to
evolve characteristics of stable forest species. A continental type of forest
can be characterized as having relatively poor dispersal ability in its species
because forest trees tend to have large seeds (the abundant storage tissue
suited to growth of seedlings in shady forest floors), shorter seed viability,
and a high degree of ecological specialization. Plants of pioneer habitats,
on the contrary, are suited to long-distance dispersal, corresponding with
the scattered nature of disturbed or new habitats, and they typically
live in sunny areas, have less seed storage and smaller fruits and seeds
(smallness contributes to their dispersibility), and are mostly herbaceous,
relatively short lived, and establish easily because of plasticity of ecological
requirements. Shift from the pioneering-plant syndrome to the forest
syndrome is a prime source of ecological change on islands. The evolution
of Asteraceae from herbaceous, weedy types to shrubs and trees of forest
regions is a prime example (chapter 10). This ecological shift is possible
under insular conditions because wet forest is a relatively uncontested zone,
a fact related to the few wet-forest-species propagules successfully estab-
lished on islands after long-distance dispersal. Increase in fruit and seed
size is an adjustment to this change in ecological preferences, and loss of
dispersibility is a by-product. Ecological shift is invoked in explaining loss
of dispersibility of insular insects by Darlington (1943) and Wilson (1959).

A third possible factor can be described as failure of a plant to
maintain the link with the dispersal mechanism that brought an ancestral
stock to an oceanic island. An outstanding example of this tendency is
evident in Bidens (Asteraceae). The Pacific island species of Bidens (figs.
11.2, 11.3) that have the best dispersal mechanisms, and appear most

433

Loss of Dispersibility in Island Plants

similar to ancestors, live on cliffs where seabirds typically nest or once
nested. This suggests that attachment of fruits to bird feathers was the
vector for immigration of Bidens, and that ancestral species occupied areas
near nests of seabirds. As a population of Bidens evolved into upland forest
sites, where marine birds are absent, the contact with the dispersal vector
would be lost, and either the old mechanism would be lost, a new one
developed, or both would occur. Evolution into inland sites by plant groups
brought to islands by oceanic drift would also involve loss of contact with
a dispersal vector, as mentioned above for Canavalia. Thus, floatability
has either a neutral or negative value in an inland site. The value of
floatability could be said to be negative if this tended to reinforce the
littoral preference and thus restrict a species permanently to a narrow
coastal zone and deprive it of occupying broad zones of inland forest.
Loss of contact with a dispersal vector is a by-product of evolutionary
opportunity and ecological shift, which in turn are produced by the
disharmonic nature of oceanic island floras.

The number of species occupying a given acre of forest on an oceanic
island is likely to be limited for reasons of disharmony. In an island
forest, such as those of the Hawaiian Islands, a relatively poor rate of
reproduction and dispersal will suffice to maintain a species. In a com-
parable continental forest area, the number of species is much higher, the
probable attrition of propagules of each species high, so despite adaptation
to the forest habitat, a continental species might be expected to maintain
a higher rate of reproduction and dispersal. This would explain why a
number of genera represented both in the relatively unspecialized forests
of continents (or continental islands) and also in forests on oceanic islands
would tend to reach maximum fruit size in the island locality (e.g.,
Alectryon, Fagara, and Tetraplasandra in the Hawaiian Islands). It would
also tend to explain why little gigantism or loss of dispersibility can be
observed west of the andesite line in the Pacific, in floras such as that of
Fiji, where many of the same genera present in the Hawaiian flora occur.

Other factors may involve pleiotropic genes. Assuming that there is
a selective value for gigantism in flowers, as in the gigantic heads in
Argyroxiphium (Asteraceae), enlargement of flower size may result in
changes in proportions of the flower, and the dispersal mechanism (pappus
in Argyroxiphium) might deteriorate accordingly. Or one can imagine that

434

Loss of Dispersibility in Island Plants

if a dispersal mechanism is discontinued (by loss of link with a vector, for
example), and the associated structure has a neutral value, its loss may be
an inevitable result of economy in the energy budget of a plant. One
may hypothesize that continued selective pressure is required to maintain
a particular dispersal mechanism, and once this pressure is reduced by
altered dispersal modes, change or deterioration of structure or construction
is inevitable. The grotesque fruits of Dendroseris litoralis (fig. 11.8G),
D. pinnata (fig. 11.8F) and Yunquea tenzii (fig. 11.7C) probably represent
forms “tolerated” by lack of selective pressure for typical fruit forms in
Asteraceae. Odd-shaped seeds occur in Hawaiian species of Dianella,
Haplostachys (fig. 11.11), Stenogyne, Pittosporum (fig. 11.12) and Alsino-
dendron (fig. 11.9). Improvement in dispersibility of island plants is also
conceivable, and may have occurred in Lecocarpus (fig. 11.5E) and Tre-
matolobelia (Carlquist, 1962). In any case, I would expect alteration to a
new dispersal mechanism to occur rapidly once any of the various factors
suggested has become operative.

Methods and Problems of Comparisons

In order to assess loss of dispersibility, one must be able to compare
insular species with their closest continental relatives. In the case of the
Hawaiian flora, the comments of Hillebrand (1888) on what the relation-
ships of endemic Hawaiian genera might be, proved helpful. For other
floras, affinities were determined on the basis of the closest morphological
and phytogeographic affinities I could determine. Dispersal mechanisms
seem clear for some species, but several possibilities must be conceded for
others. Despite the wealth of information offered by Guppy (1906), Ridley
(1930), and others, mechanisms of dispersal in most plants are suppositions
based upen the morphology of fruits and seeds compared to those of
species for which good observational data are available.

Phenomena of lessened dispersibility may not be subject to experi-
mental proof in the ordinary sense, but the array of data given here does
provide “controls.” If a number of different taxonomic groups on a single
archipelago, such as the Hawaiian Islands, show parallels, one can say that
controls or replications in the experimental sense exist. Likewise, different

435

Loss of Dispersibility in Island Plants

island groups on which a particular group (e.g., Bidens) is represented and
shows, on each, similar tendencies, serve as controls or independent
“proofs.”

Adaptations for dispersal in particular plant groups have been outlined
earlier, as for the Hawauan flora in chapter 2. One can hypothesize that
the maximum size of fleshy fruits dispersed by frugivorous birds to the
Hawaiian and other oceanic islands is about 1 cm, and that if the fruit is
drupoid the contained seed is 8 mm long or less. Some groups have shiny
black seeds not accompanied by fleshy pulp (seeds of Rutaceae, for
example); these seeds are nevertheless attractive to seed-eating birds.
Gigantism in fruits and seeds of bird-dispersed plants would be considered
a loss of dispersibility.

Oceanic drift only accounts for a small proportion of the floras of
most high islands, but it accounts for a higher proportion of the Hawaiian
flora (fig. 2.7). Guppy emphasizes a “Fijian difficulty,” whereby coastal
strand plants fail to evolve into forest sites and forest plants do not evolve
into coastal sites. For this reason, Guppy seems reluctant to acknowledge
the existence of ecological shift from strand to forest in the Hawauan
flora. One should note that Fiji, although insular, retains an essentially
continental (harmonic) flora, whereas the Hawaiian flora is strongly
disharmonic. In Hawaii, immigrant groups were offered more numerous
ecological opportunities that would have been preempted on Fiji by the
many and varied genera native there. Hawaiian groups (other than those
which are now exclusively strand plants) that may be suspected of having
immigrated via seawater include all the native Fabaceae and Malvaceae,
at least some Apocynaceae, and the native species of Colubrina, Dodonaea,
Gouania, Haplostachys, Myoporum, and Peucedanum (see chapter 2).
For these, absence of both buoyancy and impermeability by seawater
constitutes an easily measurable loss of dispersibility.

For spores or seeds suspected of being dispersed by wind or in mud on
feet of birds, evolutionary increase in size would curtail dispersibility.
Adherent fruits are less likely to be dispersed if barbs or hooks are lost,
sticky seed coatings disappear, or gigantism in fruits or seeds renders
adherence mechanisms less effective than they would be for small «seeds
or fruits. A feature worthy of note for types of transport other than oceanic
drift is that only a slight decrease in transportability would markedly

436

Loss of Dispersibility in Island Plants

increase the precinctiveness of a species. This is particularly true of wind-
dispersed seeds or spores.

The instances selected for presentation represent, in my opinion,
easily interpreted cases. More detailed statistical treatment of other groups
might reveal additional valid examples. Data and observations were derived
from material collected in the field and from specimens studied in the
herbarium. Living materials were utilized wherever possible. Experiments
to determine floatability of seeds and fruits were undertaken, following the
methods of Guppy (1906). Such tests were considered valid only if seeds
contained normal embryos, and if seeds or fruits either sank readily or
floated for prolonged periods. In the illustrations, scales of magnification
are consistent within families, although different from one family to
another. Sources of non-Hawalian species are indicated in parentheses.
All species for which no source is cited in the drawings are endemic to the
Hawauan Islands.

Ferns

Warren H. Wagner, Jr., kindly contributed data on spore size, ecology,
and nomenclature of Hawaiian ferns. A selection from these data is given
in table 11.1. Data on spore size in non-Hawaiian Schizaeas is from Brown
and Brown (1931) and from Selling (1944, 1946). Data on the fossil
Schizaeas (S. skottsbergii and var. mauiensis) are from Selling (1946).
Selling’s measurements apparently are uniformly too large by about 10
percent and have been corrected in accordance with Wagner’s guidance
(personal communication, 1965). Spore sizes represent an average of the
greatest dimension of spores from one or more collections. Exotic ferns
have been added to the table where, in a given family, there are no Hawai-
ian representatives in low forest, and a general picture of typical spore sizes
in a family is necessary for comparison with species of wet Hawaiian forest.

The examples given in table 11.1 indicate a clear correlation between
ecology and spore size. Because average size in most fern spores centers
around 30 », the dimensions of the spores in the wet-forest species are
truly exceptional. This may be regarded as a specialization with reference
to wet forest except in Athyrium, where the small-spored A. molokaiense

437

Loss of Dispersibility in Island Plants

TABLE 1].].
Comparison of spore size with ecology for Hawaiian native ferns,
with spore sizes given for selected non-Hawaiian species
for purposes of comparison

Family Species of low forest, or widespread species

Schizaeaceae Schizaea dichotoma (Old World Tropics), 60
Schizaea digitata (Old World Tropics), 42 »

Gleicheniaceae Dicranopteris emarginata, 32.4 p

Dicranopteris linearis, 28.7
Hicriopteris pinnata, 26.3 p

Hymenophyllaceae | Vandenboschia cyrtotheca, 38.4
Vandenboschia davallioides, 41.4 u

Adiantaceae Microlepia speluncae, 26.1 »
Pellaea ternifolia, 41.1 p
Pityrogramma calomelanos (New World Tropics), 40.3 p
Pteris excelsa, 42.9 w
Pteris vittata, 34.8 p

Blechnaceae Blechnum occidentale, 42.2
Doodia kunthiana, 39.2 p
Doodia lyonii, 40.9 p

Aspidiaceae Athyrium molokaiense, 36.4

Rumohra adiantiforme (New Zealand), 32.2 pw
Rumohra aristata (Michigan), 33.4 »
Rumohra hispida (New Zealand), 37.0 »

may represent the reverse adaptation to dry lowland conditions. Hawatan
species of Cibotium (Cyatheaceae) are not included in the table. These
species do have notably large spores (C. chamissoi, 54.6 »; C. glaucum,
49.3 »; C. st.-johnii, 46.5 »; C. splendens, 52.2 »). This large spore size
characterizes not merely Hawaiian Cibotiums, however, but all Cibotiums.
The genus is adapted to wet forest almost uniformly, and thus both
Hawaiian and non-Hawaiian Cibotiums may be said to show correlation
between wet-forest habitat and large spore size.

The explanation for the good correlation shown in table 11.1 seems

438

Loss of Dispersibility in Island Plants

Species of intermediate forest

Vandenboschia draytoniana, 48.6 p
Goniocormus minutus, 53.4 p

Sadleria cyatheoides, 49.6 »
Sadleria pallida, 48.9
Sadleria souleyetiana, 46.9

Athyrium kaalanum, 39.8
Athyrium marginale, 41.0 p,
Athyrium meyenianum, 42.6 p
Athyrium microphyllum, 36.4 p
Athyrium sandwichianum, 40.3 p

Source: Carlquist, 1966b.

chiefly one of precinctiveness: the large spores of wet-forest species will
fall nearer the parent plant and will be more likely to survive, for this
specialized habitat is quite limited in geographical extent. Although the
ferns listed represent a selection from the Hawaiian fern flora, no cases
which definitely run counter to the trend of increasing spore size in wetter
areas can be said to exist. There are large groups of native Hawaiian ferns
which show in spore size neither a positive nor a negative correlation with
forest ecology. These—mostly Aspleniaceae, Polypodiaceae, and Gram-
mitidaceae—have spores that range mostly between 30 and 45 ». Signifi-

Species of high, wet forest

Schizaea skottsbergit, 100 p
S. s. var. mauiensis, 90 pw

Sticherus owyhensis, 36.9

Callistopteris baldwinit, 54.8
Mecodium recurvum, 57.5 p
Sphaerocionium lanceolatum, 54.8 p

Pteris lidgatii, 74.1 p

Sadleria squarrosa, 62.0 p
Sadleria unisora, 56.3 p

Rumohra curvifolia, 58.8 p,

439

Loss of Dispersibility in Island Plants

cantly, these ferns are almost all epiphytes, whereas the ferns listed in
table 11.1 are terrestrial. The epiphytic habit represents occupancy of an
“open habitat,” one which is widely scattered. Small- or medium-sized
spores seem better adapted by virtue of their vagility to reach the many,
widely. dispersed sites suitable for growth of epiphytes. Trees bearing
epiphytes are limited in life-span, suggesting another reason for vagility of
spores of epiphytic ferns.

Polyploidy and apogamy do influence spore size, of course. However,
none of the ferns listed in table 11.1 is known to be apogamous, nor do
chromosome numbers form a good correlation with the spore-size trends
(Wagner, personal communication, 1965). Indeed, according to Wagner,
the endemic Hawaiian ferns, among which the high-forest species may
be cited, tend to be diploid, whereas wider-ranging native species in the
Hawaiian fern flora contain the bulk of the species in which Wagner has
now established the existence of polyploidy. The explanation for larger
spore sizes in the Hawaiian fern flora seems clearly related to ecology.

Another noteworthy feature that Wagner has established for Hawaiian
ferns is the exceptional degree to which they possess paraphyses in sori.
Although the functions of paraphyses and the reasons for abundance
of these structures in Hawaiian ferns cannot be elucidated at present,
paraphyses might have the effect of hindering free discharge of spores, and
might thereby increase precinctiveness of ferns which possess them.

The large size of Sadleria spores is related to the fact that only 16
are borne per sporangium (Robert Lloyd, personal communication, 1967);
the genus does reproduce sexually, however. Precinctiveness in some ferns
on islands as well as on mainland areas may be aided by the fact that
chlorophyllous spores (such spores average only 48 days’ viability, according
to Lloyd) may be a form of lessened dispersibility. Nonchlorophyllous
spores average 2.8 yr. in viability, according to Lloyd. Among the fern
families in the Hawaiian flora with chlorophyllous spores, one can cite
Hymenophyllaceae. For some hymenophylls, spores even germinate within
a sporangium (Lloyd and Klekowski, 1970). Lloyd and Klekowski believe
that for some groups of ferns at large, the green-spored condition may be
primitive (Osmundaceae), but in others, relaxation of selective pressure
for longevity (chiefly in wet-forest areas) has resulted in the attainment of
chlorophyllous spores.

440

Loss of Dispersibility in Island Plants

Gymnosperms

Because few gymnosperms are present on oceanic islands, this group could
be expected to provide few examples of lessened dispersibility. However,
one notes that Podocarpus ferrugineus on New Zealand and P. minor
(= P. palustris) on New Caledonia have exceptionally large seeds. These
examples may constitute instances of autochthonous evolution of seed
gigantism on these islands.

Dicotyledons

Apiaceae. Dispersal mechanisms of most umbellifers appear obscure.
Whereas the hooked fruits of Sanicula or the barbed fruits of Drusa or
Bowlesia are obvious enough, the dry, corky mericarps of most Apiaceae,
often without wings, do not appear readily dispersible at first glance.
However, birds may consume these fruits. That some Apiaceae without
“obvious” dispersal mechanisms have wide ranges, including many of the
world’s oceanic islands, speaks for itself. However, some insular Apiaceae
may have an autochthonously lessened degree of dispersibility. Perhaps
Myodocarpus on New Caledonia and Peucedanum on the Hawaiian Islands
fall into this category.

Apocynaceae. Among Hawaiian Apocynaceae, only the endemic genus
Pteralyxia suggests a fruit size possibly greater than that of ancestral types.
Unfortunately, the relationships of Pteralyxia have not yet been clearly
established. Hillebrand (1888) claimed affinity with the American genus
Vallesia, whereas Schumann (1895) emphasized resemblances to the wide-
ranging Pacific genus Alyxia. The single-seeded drupes of Alyxia do show
a close resemblance to those of Pteralyxia in features such as the large
locule, filled with ruminate endosperm. Fruits of Alyxia do not float and,
because of fleshy exocarp and small size, are probably dispersed by frugivo-
rous birds. Fruits of Pteralyxia, on the other hand, are fusiform and
much larger—ranging from about 5 X 2 * 2 cm in P. kauaiensis to about
6 x 4X 3 cm in P. macrocarpa. This large size would seem clearly to
preclude bird dispersal, despite the fleshy exocarp. If Pteralyxia is, in fact,

441

Loss of Dispersibility in Island Plants

related to Alyxia, this would constitute a remarkable case of loss of dispersal
mechanism. However, dried fruits of Pteralyxia are capable of floating
indefinitely, either with or without the fleshy exocarp. This feature is found
in such Apocynaceae as Ochrosia oppositifolia (= O. parviflora), a wide-
spread. species of the Pacific which is, very likely, seawater dispersed.
Ochrosia sandwicensis, a Hawaiian endemic, probably stems from seawater-
dispersed ancestors like O. parviflora, a possibility conceded by Guppy
(1906). If Pteralyxia is, indeed, related to seawater-dispersed ancestors like
Ochrosia, there may be relatively little loss of dispersibility, although seeds
of Pteralyxia might be rendered inviable by soaking in seawater. Indeed,
the woody endocarp is relatively thin and fragile, especially at the peduncle
end. The large size of fruits of P. macrocarpa suggests difficulty of transport
on land, in any case. Pteralyxia remains an enigmatic genus with regard
to ancestral dispersal type, and information about its relationships is very
much needed.

Araliaceae. Among Hawaiian Araliaceae an almost classical case of
loss of dispersibility is furnished by the genus Tetraplasandra. The re-
mainder of Hawaiian araliads have dupes of relatively small size, suggesting
ease in dispersal by frugivorous birds. Within the Hawaiian species of
Tetraplasandra, however, a trend toward marked gigantism of fruits and
seeds is evident (fig. 11.1). Of the species recognized by Rock (1913) or
Sherff (1955), most have fruits which range from 7 to 10 mm in length,
averaging 8 mm, and generally fusiform. Typical of these is T’. oahuensis

T. waialealae T. waimeae

i
Tetraplasandra philippinensis
(Philippines)

FIG. 11.1. Comparison of fruits and seeds of Tetraplasandra (Araliaceae). All species except T.
philippinensis are endemic to the Hawaiian Is.

442

Loss of Dispersibility in Island Plants

(fig. 11.1). This fruit size seems clearly ancestral among Hawaiian species,
for it is shared by the non-Hawaiian species of Tetraplasandra: T. philip-
pinensis (Palawan, Philippine Is. [fig. 11.1]), IT. paucidens (New Guinea),
and T. koerdersii (Celebes). Hawaiian species with slightly larger fruit in-
clude T. hawatiensis (10 mm in diameter), T. waianensis (10-12 mm long,
3-4 mm thick), T. meiandra (13-14 mm long, 5-6 mm thick), and T. kau-
aiensis (ca. 14 mm in diameter [fig. 11.1]). More marked gigantism is
shown by two Kauai species: T. waialealae and T. waimeae (fig. 11.1).
Drupes of T. waialealae are fusiform, 25-32 mm long, whereas those of
T. waimeae are spheroidal, ca. 30 mm in diameter. Pyrenes (seeds) in these
species are in proportion to size of fruits, as shown in figure 11.1. There
seems to have been a trend toward gigantism in Tetraplasandras with re-
lation to upper wet rain forest. The two species with the largest fruits are
native to wet forest on the oldest major island, Kauai.

Asteraceae. Of all families, the composites show tendencies in loss
of dispersibility most clearly, and so they bulk largest in this account. My
field work in Polynesia was useful in demonstrating relationships. Addi-
tional ecological data of a simple sort were gleaned from labels of speci-
mens and from the works of Brown (1935), Harling (1962), Papy (1954-
1955), Sherff (1937b), Skottsberg (1953, 1954), Stewart (1911), and Zim-
merman (1948). Nomenclature for Pacific Asteraceae is according to the
following authors: for Bidens, Sherff (1937b); for Oparanthus, Sherff
(1937a); for Fitchia, Carlquist (1957) and Carlquist and Grant (1963); for
Scalesia, Howell (1941); for Macraea, Harling (1962); for Lipochaeta,
Sherff (1935b); for Argyroxiphium and Wilkesia, Carlquist (1959); for
Juan Fernandez Asteraceae, Skottsberg (1922, 1951, 1958), although the
segregate genera of Dendroseris are not used here.

The genus Bidens (tribe Heliantheae, subtribe Coreopsidinae) is well
represented by endemic species in the Pacific (figs. 11.2, 11.3). Bidens
pilosa, a widespread tropical weed, has been introduced by man to many
Pacific islands and is included here only as representative of a mainland
species. The Pacific species of Bidens may be regarded as close to the B.
pilosa complex, and this species (fig. 11.2, top left) probably represents a
reasonable approximation of the stock or stocks of Bidens that established
on Pacific islands aboriginally. For each of the insular species shown in
figures 11.2 and 11.3, a typical achene is shown (or, for a few, variables

443

B. pilosa
(AMERICAN MAINLAND)

af

MEXICO

aa |

B. henryi B. collina

¥ 5 eg ee
x j ? iS ,
y K fo R\ a
X y = €
2 & r
aD aS ©
i) S «
. aS > &
ws Y
B orofenensis B.aoraiensis 8 glabrata
RAPA ;
< AUSTRAL Is. / h
+ MAROTIRI A \
| fr
\

B. beckiana , 8: Polycephala g jardinii B bipontina 8. cordifolia

EIAO o °
NUKU HIVA

UA POU — @ HIVA OA
MARQUESAS IS.

Sociery IS. YB. uapensis

B. glandulifera
g a

/

B. deltoidea /

a

vee /

B. raiateensis WW)

B. mooreensis B. lantanoides

B. mathewsii

B. australis
TAHITI >)
MOOREA
TAHAA

TONGA

B. hendersonensis

B. hendersonensis var. oenoensis
B. saintjohniana

FIG. 11.2. Comparison of fruits of endemic species of Bidens (Asteraceae) from southeastern
Polynesia, shown geographically. These species, together with those from the Hawauan Is.,
represent the extent of this genus in the Pacific. Inset, top nght, an unnamed species native
to Socorro I., Revillagigedo Is., Mexico. Inset, top left, a typical achene of B. pilosa; this
fruit probably approximates the achene type ancestral to the Pacific species of Bidens. Species
omitted from this map include B. ahnei, which may be conspecific with B. polycephala, and
B. hivoana, which is a synonym of Oparanthus albus.

444

Loss of Dispersibility in Island Plants

in morphology are illustrated). Achene morphology does vary considerably
within some species of Bidens.

The dispersal mechanism of Bidens is evidently quite efficient, for
the genus has reached, by natural means, many of the smaller islands of
Polynesia. However, the genus appears to have “bypassed’’ many seem-
ingly suitable islands (Rapa, Mangareva, and the Austral Islands other
than Marotiri); this suggests that chance has operated in dispersal of
Bidens. On Tonga a single collection of B. australis has been made; al-
though not unlikely, this record needs confirmation. If B. australis is pres-
ent on Tonga, one would expect this or other Bidens species on Samoa
and other islands.

None of the endemic Pacific species of Bidens seemingly has dispersal
mechanisms quite as efficient as B. pilosa. Five features may be said to
comprise the dispersal mechanism of B. pilosa: (1) awns long compared
to the body of the achene; (2) awns spreading; (3) awns armed with stiff
retrorse barbs (unicellular, thick-walled trichomes); (4) upward-pointing
hairs on lateral margins of the achene; (5) upward-pointing hairs on the
dorsiventral margins (flattened faces of the achene). The length of awns
seems important because it controls the number of barbs and thus the
likelihood of attachment to animal fur or bird feathers. Even without
barbs, some degree of adhesion can be effected by long awns, especially
if they are spreading. The presence of retrorse barbs on the awns in con-
junction with upward-pointing hairs on the achene body seems important.
Onented counter to each other, these structures probably secure a firmer
adhesion than either alone would achieve. Any depreciation in the five fea-
tures listed may be said to constitute a lessening of dispersibility.

Among the South Pacific species of Bidens, only a few retain the speci-
fied features to a marked degree. Those best equipped seem to be the trio
of species from the southeasternmost islands: B. hendersonensis, B. ma-
thewsi, and B. saintjohniana. All three species have numerous hairs on the
achene body (dorsiventral as well as lateral margins, except for some popu-
lations of B. hendersonensis). Similarly well equipped is the unnamed spe-
cies from Socorro Island, Mexico (fig. 11.2, top right). Awns are short and
not spreading in these species, however.

Second in potential dispersibility among South Pacific species are
those with short, but still barbed, awns and with hairs on the achenes

445

Loss of Dispersibility in Island Plants

(lateral margins only): B. australis, B. lantanoides, B. mooreensis, and B.
cordifolia. Efficiency of dispersal in these species is confirmed by the fact
that two (B. australis, B. lantanoides) occur on more than one island,
whereas all other South Pacific species are restricted to single islands.

A third category is represented by species with awns barbed but
achene body glabrous: B. glandulifera, B. aoraiensis, and B. glabrata (p.p.).
Bidens orofenensis and B. glabrata (p.p.) lack barbs on awns but have,
like B. aoraiensis, markedly spreading awns.

Lack of awns (except as ineffective vestiges) is shown by the Marque-
san species other than B. cordifolia. All the Marquesan Bidens, however,
have hairy achenes (lateral margins only, except for B. collina). Similarities
in achene morphology and other features suggest that the Marquesan spe-
cies originated from a single introduction. The presence of achene hairs
alone would probably suffice to secure transport via bird feathers through-
out that archipelago.

The two species from Raiatea show the greatest loss of dispersibility
among South Pacific Bidens species, for they lack any hairs or appendages
(except for vestigial awns and a few hairs near the achene summit in B.
deltoidea).

Within the Hawaiian species of Bidens (fig. 11.3), a wide gamut in
achene morphology can be seen. For purposes of illustration, these have
been grouped into categories based upon features of morphology related
to potential dispersibility. The four species of A are equipped with all
features. One (B. hillebrandiana) has spreading awns, another (B. nema-
tocera) notably long awns. Diminution of awn length is shown by the spe-
cies of B, whereas those of C have awns of varying length but lack achene-
body hairs. Near-absence of awns and restriction of hairs to the base of
the achene is shown in the species of D. A few hairs occur near the sum-
mit of the achene body in B. cuneata and B. forbesii. Lack of awn vestiges,
combined with a hairy achene body, characterizes the species of E, which
thus resemble the Marquesan species. In F are species that possess a gla-
brous achene body and only the most vestigial awns. Awn vestiges in B.
menziesii and B. mauiensis are flat and toothlike, and are seemingly func-
tionless. A few hairs crown the achenes of B. cosmoides, which are other-
wise glabrous.

The species in the upper half of figure 11.3 have no peculiarities not

446

i B. skottsbergii Cc

YW
B. micrantha igh

B. hawaiiensis

J
B. distans i.)
ik

Le 7 B hillebrandiana A

B. molokaiensis \
B. waimeana ¥

n
A
\

(

B. nematocera
B. stokesii

>

4 A B populifolia B
B. micranthoides U
B. amplectens =
B. fecunda me
B. valida WwW
B.

conjuncta tf =|

B. coartata E

AN

ea

4.) 8. napaliensis ~ B. salicoides F

B. perverso

Lit
B. menziesii in Tas
VN

B. pulchella

B. obtusiloba ih

B. asplenioides B. mauiensis

B. cervicata se

B. graciloides A

B. cosmoides

B. campylotheca G B. torta H

‘ B. sandvicensis
B. waianensis >
(
B. fulvescens iD)

B. ctenophylla I

1} | 8B. magnidisca L
is l 5 mm |

ric. 11.3. Comparison of fruits of endemic Hawaiian species of Bidens. The species are
arranged into groups according to achene morphology. The groupings shown are not intended
to represent natural groups, although in some cases they apparently represent closely related
species. In each grouping shown here, one achene is drawn in its entirety; achenes of the re-
mainder within a grouping agree in details of achene body, but differ in details of awns, etc.,
at the achene apex. Achenes may differ in size within a grouping, however. Some species of
Bidens are variable in achene morphology; a “typical” achene is shown for each. Scale (bottom
right) applies to all figures.

acti

Loss of Dispersibility in Island Plants

also seen in the South Pacific species. In the lower half of figure 11.3, how-
ever, are features of achene morphology unique to the Hawaiian Islands.
Achenes of B. campylotheca (G) are glabrous, have vestigial awns, and
are bent into an arc. Achenes of B. degeneri have no awn vestiges (only
callous margins at the achene summit) and may be straight or curved like
those of B. campylotheca. More prominently curved are the species of H,
which often exhibit achenes twisted into three or more helices. Although
achenes in these species are narrow in diameter, the embryo is relatively
large and occupies nearly the entire length of the achene. Because several
species have curved or helical achenes, there may be a selective value for
this form, but I am not willing to speculate as to the possible function of
this shape. Bidens campylotheca and B. sandyicensis each occur on four
of the major islands. Therefore this achene form can be effective over rela-
tively short distances. Achenes of the species in H differ from each other
only in minor respects.

The remaining Hawaiian species (I-L) have large, flat achenes that
can sometimes be described as winged and that have awns vestigial, lack-
ing, or highly modified. Bidens ctenophylla (1) has vestigial awns, if any,
on broad black achenes that occasionally bear a tooth along the lateral
margins. Achenes of B. wiebkei (J) are flat, have awns vestigial or lacking,
and bear a few stiff hairs on lateral margins of achenes. ‘These hairs have
broad, multiseriate bases. Some achenes of B. wiebkei are distorted in
shape and have wavy lateral margins. Achenes of B. macrocarpa (K) are
the largest among Hawaiian species. They have broad yellowish wings and
are glabrous. Flat awnlike appendages are modified in conjunction with
the formation of wings, and are thus different from true awns. A few hairs
are present on these appendages, but most hairs are not retrorse. Achenes
of B. magnidisca (L) are like those of B. ctenophylla but have awns like
those of B. macrocarpa, with yellowish winglike webbing between the
awns and the body of the achene; or awns may be lacking, with narrow
yellowish wings at the achene summit. The species of I through L may be
flat enough to suggest transport by wind, but such transport would be
effective over short distances at best.

Ecology correlates clearly with morphology of achenes for the insular
Bidens species. Of the South Pacific species (fig. 11.2), those with the best
dispersal mechanisms are all coastal species. Bidens saintjohniana occupies

448

Loss of Dispersibility in Island Plants

Marotiri, which can be regarded as nothing more than a few rocky islets
inhabited by seabirds. Oeno and Henderson, occupied by B. hendersonen-
sis, are smal] low islands on which seabirds are abundant. Bidens mathewsit
grows on exposed sea cliffs of Pitcairn Island. In other South Pacific Bidens
species, greater diminution of dispersibility has occurred; these are spe-
cies which grow in localities inland from the immediate coast. Those
species with awns still barbed occupy relatively low elevations in open
scrubby forest or on grassy ridges. The unnamed Bidens from Socorro
Island grows in open dry forest at about 1000 ft. Bidens australis has been
collected at low elevations on Tahaa recently; it has not been collected
recently on the other islands of its range, perhaps because of destruction
of low-elevation forests by burning and farming. This may be true of B.
lantanoides also. Bidens aoraiensis and B. orofenensis are not low-elevation
species, but they do grow on exposed rocky outcrops of ridges high on
Tahiti and could not in any way be considered forest species. Bidens glan-
dulifera occurs on exposed grassy ridges of montane Bora Bora.

The Marquesan species occupy grassy ridges and open forest of middle
elevations, as do B. deltoides and B. mooreensis of the Society Islands.
Bidens raiateensis, with the poorest dispersal means of any South Pacific
Bidens, grows in wet upland forest. Thus, in the South Pacific species, the
farthest penetrations into forest sites are accompanied by greatest diminu-
tion of dispersal mechanism.

Among the Hawaiian species, similar ecological correlations occur.
The species of A in figure 11.3 are occupants of seacoast bluffs (B. hille-
brandiana, B. molokaiensis, B. nematocera) or open, low-elevation slopes
(B. waimeana). The species of B through F occupy low-elevation sites ex-
cept for B. micranthoides (open places in mid-elevation dry koa forest,
Kauai), B. hawaiiensis (open lava fields of Mauna Kea, Hawaii), B. con-
juncta (wet but open grassy sites of higher Oahu and Mau) and B. cos-
mozides (rain forest of upper Kauai). The species in the bottom half of figure
11.3 mostly occupy relatively wet upland forests. ‘These species, in general,
have large embryos suited to germination in shady forest sites. A notable
exception is B. ctenophylla (open lava slopes near Puuwaawaa, Hawaii).
This might be a case of a species which has been derived from a forest-
inhabiting ancestor rather than a coastal one. Another exception is B.
degeneri (arid regions of lower Oahu, Molokai, and Maui).

449

Loss of Dispersibility in Island Plants

Despite the exceptions noted, the progress of Hawaiian species of
Bidens from seacoast to interior and the accompaniment of this adaptive
shift by loss of dispersal mechanism are clear. One can hypothesize that
Bidens originally established in coastal sites where seabirds nested. ‘The
fact that many coastal areas of the Hawaiian Islands are not now occu-
pied by seabirds is probably a result of human interference. One should
not expect perfect ecological correlation, and there seems to be a degree
of flexibility within many species in expression of the dispersal mechanism
—a flexibility that suggests weak selectivity for “good” dispersal mecha-
nisms. Interestingly, George W. Gillett (personal communication, 1966)
has demonstrated that features indicating good dispersibility (e.g., like
B. pilosa) are genetically dominant, and that the various loss-of-dispersi-
bility features are genetically recessive.

Pacific species of Bidens are well suited for entry into Polynesian for-
ests because they are perennial and can become large shrubs up to about
3 m in height. They are not, however, strongly woody, and seem not to
have advanced as far toward arborescence as, for example, Fitchia. There
seems little question that the woodier species of Bidens do represent spe-
cialization—adaptation to forest situations.

Oparanthus (fig. 11.4) consists of four species native to the Marquesas
and Rapa. This recently recognized genus (Sherff, 1937a) may be consid-
ered a derivative of an early immigrant stock belonging to Heliantheae,
subtribe Coreopsidinae. The distribution of Oparanthus suggests relictual-
ism of a recent order or poor dispersal capacity. Anatomical studies (Carl-
quist, 1957) show relationships to such Coreopsidinae as Bidens and Core-
opsis, as well as to Fitchia. Oparanthus has fertile ray achenes (at right in
each species, fig. 11.4) and sterile (male) disc achenes. The disc achenes—
which have no function in reproduction—retain a single awn in O. rapensis
and O. coriaceus. The ray achenes are flattened (curved in O. rapensis),
although not broadly winged. Awns are modified into winglike structures
in O. albus, reminiscent of Bidens macrocarpa or B. magnidisca. Achene
body and awns in Oparanthus are completely glabrous. Thus, achenes
have a low dispersibility. This correlates well with ecology and habit of
Oparanthus. All species are shrubs or trees, markedly woody, and thus
have a growth form frequent in Polynesian forests. Oparanthus seems to

450

O. albus ;

MARQUESAS IS.

NUKU HIVAS ,
© HIVA OA
at

FE cuneata F cuneata F tahitensis F nutans
subsp. tahagensis

E cordata

BORA BORA ° TuAMOTUS

<< ;
RS O yaHiTt pee th
Coox Is. e if

SOciETy Is.

RAROTONGA °

* o MANGAREVA

AUSTRAL IS.

O. rapensis

} E speciosa fn ey
ity O. intermedius (

Bidens pilosa

Lomm |

F rapensis F mangarevensis

Fic. 11.4. Comparison of fruits of all known species of Fitchia and Oparanthus (Asteraceae),
shown geographically. Note that the scale of magnification is half that of figs. 11.2 and 11.3.
An achene of Bidens pilosa is shown, bottom center, for comparison of size and morphology.

451

Loss of Dispersibility in Island Plants

represent a more complete adaptation to the rain forest than does Bidens,
and probably signifies a longer history in Polynesia.

Fitchia (fig. 11.4) seems clearly related to Coreopsidinae, but by virtue
of homogamous capitula and other features it seems worthy of recognition
within a separate subtribe, Fitchiinae (Carlquist, 1957). Fitchia is a truly
woody genus, ranging from small shrubs to true trees. I suspect that it
represents a South American ancestor whose mainland relatives have now
vanished. The achenes of Fitchia are relatively large; the smallest occur in
the Society Islands species (particularly small ones, smaller than those fig-
ured, have been observed recently on Tahiti by Van Balgooy [personal
communication, 1971]). Fitchia mangarevensis and F. rapensis have fruits
of intermediate size, whereas F. speciosa has gigantic achenes—probably
the largest in the family Asteraceae. All species are provided with long
awns, but, except in F’. speciosa and F. cordata, the awns seem thin and
break easily from the achene body. Achenes in all species are densely
covered with upwardly appressed hairs, as are the awns. Lack of retrorse
barbs on the awns does not seem strongly disadvantageous in dispersal.
Fruits with hairs oriented in a single direction disperse quite well (e.g.,
grasses). ‘The ancestors of Fitchia probably were brought to Polynesia in
feathers of birds, perhaps seabirds.

One might speculate that the ancestors of Fitchia had smaller fruits
than do the current species. The relatively large size of achenes in the
Society Islands species (except as noted above) seems no severe disadvan-
tage, for this group has become distributed throughout the high islands
of the archipelago. ‘The southern pair of species, F’. rapensis and F’. manga-
revensis, are very close and their intermediate fruit size may have origi-
nated in southeastern Polynesia. However, ancestors have ‘evidently suc-
ceeded in dispersing across the distance between Mangareva and Rapa,
for it is difficult to imagine that achenes of this size originated independ-
ently on the two islands when other features of the two species are so
much alike. The large achene size of F. speciosa seems likely to have
originated on Rarotonga autochthonously, perhaps from ancestors with
achenes like those of F. cordata. Not only does F. speciosa represent the
farthest penetration of the genus into the Pacific, it is also the most arbo-
Tescent species in the genus.

Fruit size correlates well with size of plant in Fitchia. The Society

452

Loss of Dispersibility in Island Plants

Islands species are chiefly shrubs which occur in scrubby forest on midges.
Fitchia rapensis and F. mangarevensis are small trees, whereas I’. speciosa
is a relatively tall tree of deep forest. The large fruit size of F’. speciosa is
reminiscent of the large fruits of other forest trees. It seems an ultimate
point in adaptation to rain forest.

Fitchia seems a more primitive genus than Bidens or Oparanthus and
may have, during a longer occupancy in Polynesia, perfected this forest
adaptation (just as it has evidently evolved ornithophily, and the other
genera have not). Even if ancestors of modern Fitchias had fruits as large
as those of the Society Island species, subsequent increase in fruit size
has undoubtedly occurred within Polynesia, and this increase is correlated
with degree of entry into forest habitats—a farther penetration than that
of Bidens. Although no direct evidence can be adduced, I suspect that an-
cestors of Fitchia did arrive in Polynesia by means of fruits smaller than
those of modern species, and perhaps rather like those of Bidens. The ab-
sence of Fitchia from the Marquesas, Hawaiian, and Austral Islands (other
than Rapa), all of which offer suitable ecological sites (F’. speciosa has es-
caped from cultivation on Oahu), suggests its present dispersibility is effec-
tive only over rather short ranges. A feature not related to fruit size but
probably deleterious to dispersibility is the nodding habit of the heads.
Heads shatter at maturity, but instead of displaying fruits above foliage,
tend to drop fruits beneath the parent plant.

Scalesia, endemic to the Galapagos Islands, belongs to Heliantheae,
subtribe Verbesininae. It is probably closely related to several verbesininoid
genera, of which Viguiera (fig. 11.5A) can be considered an example. Of
the 18 species of Scalesia, only a few, like S. microcephala (fig. 11.5B), have
one or two awns, usually rather short. Scalesia cordata has the longest awns
(1-2 mm) but does not seem markedly more dispersible than the other
species. The remainder of the species have only small, “callous” pappus
rudiments, as in S. aspera (fig. 11.5C), or none at all. The achenes are not
hairy. Hairs on achenes of Asteraceae serve not only for adhesion to dis-
persal agents, but for release from the head as well. At maturity, achene-
body hairs flex outward when dry, forcing the fruits out of the head. Thus,
lacking hairs, achenes of Scalesia have no device for release. There seems
little question that Scalesia has a somewhat diminished dispersibility. Its
dispersal throughout the Galapagos might be due to seed-eating birds.

453

Loss of Dispersibility in Island Plants

YE

Fic. 11.5. Comparison of achenes of Asteraceae, tribe Heliantheae. Subtribe Melampodiinae is
represented by E, all others from subtribe Verbesininae. All achenes are from disc flowers un-
less otherwise stated. Mainland species are included for comparison with the insular species.
A, Viguiera linearis; Mexican mainland. B, Scalesia microcephala; Albemarle I., Galapagos Is.
C, S. aspera; Indefatigable I., Galapagos Is. D, Macraea laricifolia; Charles I., Galapagos Is.
E, Lecocarpus pinnatifidus; Charles I., Galapagos Is.; the ray achene and the bract with
peltate wing that permanently enfolds the achene are shown in longitudinal section. F,
Zexmenia brevifolia; Mexican mainland. G, Lipochaeta rockii; Molokai, Hawaiian Is. H,
L. remyi; Oahu, Hawaiian Is. I, L. lavarum; Maui, Hawaiian Is.

Some of the species appear to have reached their present stations after loss
of awns, because awnless species occur on many of the islands; perhaps
awns are not as important in dispersal as they are in Bidens.

Adaptation to new ecological situations does not appear to have
played a dominant role in loss of dispersibility in Scalesia. To the extent
that dispersibility has been lost in this genus (and that is difficult to esti-
mate, since the vector and mode of transport are not at all clear), it might
be a by-product of other evolutionary changes.

Macraea, a monotypic genus from the Galapagos (fig. 11.5D), was
formerly united with Lipochaeta but is worthy of recognition as a separate
genus (Harling, 1962). Harling claims that Macraea is close to American
species of Wedelia and Aspilia, whereas Lipochaeta has generally been
conceded to have close affinities to Zexmenia (Sherff, 1935b). All of these
genera belong to subtribe Verbesininae of Heliantheae, and are perhaps
not distant from one another. Macraea laricifolia occurs at lower eleva-
tions of the Galapagos. Achenes are winged and crowned by a circle of
short scales. Outer achenes in each head are broader and warty. The dis-
persal mechanism is unknown but apparently is effective over interisland
distances in the archipelago. A few achenes in specimens of Macraea have
short awns like those of Viguiera (fig. 11.5), so Macraea may be in the
process of losing these awns.

Lecocarpus (fig. 11.5E), a genus endemic to the Galapagos, seems to be
in the process of altering to a new mode of dispersal. As in other members

454

Loss of Dispersibility in Island Plants

of Heliantheae, subtribe Melampodiinae, ray achenes are fertile; each is
enclosed within a highly modified bract which bears a peltate wing. The
wing is thin and leaflike. The two genera closest to Lecocarpus are Melam-
podium, in which fruit-enclosing bracts are like an additional seed coat
and are not winged, and Acanthospermum, in which enclosing bracts bear
hooks (Hoffmann, 1890-1893). The morphology and weight of bract-en-
closed fruits of Lecocarpus suggest wind dispersal across relatively short
distances. Lecocarpus occurs on two of the Galapagos Islands. Either the
winged bract has been evolved since the arrival of the ancestral stock, or
Lecocarpus has now vanished from the American mainland. Further studies
might elucidate this point.

Lipochaeta (fig. 11.5G—I) contains 24 species endemic to the Hawaiian
Islands. A species alleged to be native to New Caledonia proved to be
Wedelia uniflora (Harling, 1962). Lipochaeta characteristically occupies
lowland sites. Compared with relatives such as Zexmenia (fig. 11.5F) spe-
cies of Lipochaeta have what appear to be relatively poor dispersal mecha-
nisms (although one must concede that the precise mode of dispersal is not
known). Some Lipochaetas, such as L. rockii (fig. 11.5G), retain both aristae
and wings on achenes, but most have short wings and no aristae, as in L.
remyi (fig. 11.5H), or large fruits with no wings and only vestigial aristae
(L. lavarum, fig. 11.51). Fruits of Lipochaeta, unlike those of Zexmenia,
lack hairs.

The genera Argyroxiphium, Wilkesia, and Dubautia (including Rail-
liardia) must be considered endemic Hawaiian representatives of Helian-
theae, subtribe Madiinae. These “tarweeds” are otherwise restricted to
western North America, with one species in Chile (Carlquist, 1959). Some
mainland tarweeds, such as Madia and Layia (fig. 11.6A, B), have fertile
ray achenes that have little pappus but that are enfolded within sticky
bracts, and thus have the dispersal mechanism transferred from the achene
to the bract. This seems true, to some extent, also of Argyroxiphium (fig.
11.6D, E), but few ray flowers are present in that genus, and fertile disc
achenes have no such adherent bracts. All achenes of Wilkesia (fig. 11.6C)
are disc achenes, without adherent bracts. The scales which crown the
gigantic achenes of Argyroxiphium and Wilkesia seem relatively inefficient,
but might become caught in birds’ feathers. Hairs are few or absent on
these achenes. When mature, heads in these genera shatter, and most

455

Loss of Dispersibility in Island Plants

Fic. 11.6. Comparisons of achenes of Asteraceae, tribe Heliantheae, subtribe Madiinae. All
achenes shown are disc achenes. A, Madia nutans; California mainland. B, Layia platyglossa;
California mainland. C, Wilkesia gymnoxiphium; Kauai, Hawaiian Is. D, Argyroxiphium
virescens; Maui, Hawaiian Is. E, A. sandwichense; Maui, Hawaiian Is. F, Dubautia latifolia;
Kauai, Hawanan Is. G, D. knudsenti; Kauai, Hawaiian Is.

fruits fall near the parent plant. Wéilkesia is restricted to Kauai, but Argy-
roxiphium occurs on both Maui and Hawaii. Both Wilkesia and Argyroxt-
phium seem gigantic versions of mainland tarweeds. With such relatively
large fruits, an exceptionally good dispersal mechanism would be required
if long-distance dispersibility were to be retained. Thus, dispersibility of
Argyroxiphium and Wilkesia appears to have fallen well behind that of
ancestral forms.

Dubautia (fig. 11.6F, G) apparently has retained better dispersal mech-
anisms, owing to possession of the plumose-bristle type of pappus. This
genus may be related closely to the preceding two, and a hybrid between
Argyroxiphium and Dubautia has been reported. Some species, such as D.
knudsen (fig. 11.6G), however, do have pappus rather short in proportion
to the achene; these bristles (or scales) do not spread well at maturity.
Moreover, D. knudsenii has relatively few achene-body hairs, and bracts
of the head do not reflex well to release the achenes. The heads are nod-
ding and tend to drop achenes to the ground. Appropriately, D. knudsen
is a rain-forest tree, whereas species with longer bristles on achenes and
heads that tend to open better, such as D. latifolia (fig. 11.6F), are charac-
teristic of exposed situations. There does appear to be a dispersal diminu-
tion in relation to ecology in Dubautia, and the genus may be in a state of
transition in this respect. Better dispersibility in Dubautia may account
for the fact that the genus occurs the length of the chain of major Hawai-

456

Loss of Dispersibility in Island Plants

ian Islands, and has more species (27) and has occupied more diverse eco-
logical sites than Argyroxiphium or Wilkesia.

The sole entry of the tribe Cynareae into the Pacific is a pair of
closely related genera, Centaurodendron (two species) and Yunquea
(monotypic) endemic to the Juan Fernandez Islands. These genera are
probably close to Centaurea sect. Plectocephalus, which have fruits as
shown in figure 11.7A. In such Centaureas, pappus bristles are numerous
and strong, although somewhat caducous, and generally are retained in
large numbers at the achene summit. Pappus bristles spread outward well
when dry, forcing mature achenes out of the head. Bracts of the head re-
flex well at maturity, liberating the achenes. In Centaurodendron (fig.
11.7B), on the contrary, bracts of the involucre do not reflex appreciably,
so that fruits are well retained within the head. Furthermore, achenes
bear relatively short pappus bristles which “come off at the slightest touch,
and there can be little doubt that they are useless for the transport of the
achene” (Skottsberg, 1938). Moreover, the bristles break easily into tiny

Fic. 11.7. Comparison of achenes of Asteraceae, tribes Cynareae (A-C) and Senecioneae
(D-G). Continental species are shown for comparison with insular ones; all achenes are disc
achenes. Where pappus bristles or setae are shown to the right and above the achene (A-E),
a portion of the setae is caducous. In B, C, and E, only a small number of setae remain
attached to the achene at maturity. In D, some setae may break from the fruit, but the
majority are persistent. Setae are shown as broken lines in B and C because they tend to
crumble easily and are rarely intact on mature achenes. A, Centaurea tweediei; Argentina. B,
Centaurodendron dracaenoides; Masatierra, Juan Fernandez Is. C, Yunquea tenzii; Masatierra,
Juan Fernandez Is. D, Senecio macrantha; Australia. E, Robinsonia gayana; Masatierra, Juan
Fernandez Is. F. Rhetinodendron berterii; Masatierra, Juan Fernandez Is. G, Symphyochaeta
macrocephala; Masatierra, Juan Fernandez Is.

457

Loss of Dispersibility in Island Plants

fragments. Mature fruits are even larger than those of Centaurea, and
would require an even more efficient pappus to equal its dispersal. On all
accounts, Centaurodendron exhibits greatly lessened dispersibility. Fruits
are released only with gradual weathering of the heads. Achenes of Yun-
quea (fig. 11.7C) are similar to those of Centaurodendron in all respects
except that they have an oddly convolute surface and are irregular in shape
—features which would certainly hinder dispersal. Such distortions of the
achene may represent mutations which, although they would be unfavor-
able in a species where efficiency of dispersal is a requisite, are not dele-
terious in Yunquea, where dispersibility is already minimal. Both Centau-
rodendron and Yunquea are rosette trees of rain forest. This adaptation
is an unusual one for Cynareae, and the adaptation of these genera appears
to be “complete.”

Three genera of Senecioneae, closely related to each other, are also
endemic to the Juan Fernandez Islands: Robinsonia (fig. 11.7E), with five
species, and Rhetinodendron and Symphyochaeta (fig. 11.7G), each mono-
typic. T'ypical mainland Senecioneae, such as the Senecio macrantha illus-
trated in figure 11.7D, have many fine pappus setae that are longer than
the achene body, are persistent to somewhat caducous, and flex outward
at maturity, liberating achenes as well as providing a parachutelike struc-
ture for flotation in wind. Release of achenes is aided by reflexing of in-
volucral bracts. In Robinsonia, achenes are relatively small, but the pappus
bristles are notably short and clearly caducous. Mature female heads of
Robinsonias are cupules that do not open well and contain nearly bare
fruits that disperse mainly by falling from the head:

Rhetinodendron (fig. 11.7F) has female heads with narrow involucres
usually containing only three fruits each. The achenes are hairy, but these
hairs do not spread outward appreciably upon drying, and achenes tend
to remain within the involucres. The short pappus bristles, although per-
sistent, do not reflex but are permanently upright. Involucres open only a
little at maturity.

In Symphyochaeta (fig. 11.7G), female achenes are notably large. The
achenes are not hairy, and ridges on the achene body are knobby. The
pappus bristles are short and united into a tube, thus definitively canceling
any possibility of wind dispersal. Involucral bracts do not reflex at maturity.

All three genera of Juan Fernandez Senecioneae show advanced stages

458

Loss of Dispersibility in Island Plants

Fic. 11.8. Comparison of achenes of Asteraceae, tribe Cichorieae. A mainland species (A) is
shown for comparison with insular species. A, Lygodesmia juncea; southwestern United States.
B, Stephanomeria (Munzothamnus) blairii; San Clemente I., Calif. C, Thamnoseris lacerata;
San Ambrosio I., Desventuradas Is. D, Dendroseris (Hesperoseris) gigantea; Masafuera, Juan
Fernandez Is. E, Dendroseris (Rea) micrantha; Masatierra, Juan Fernandez Is. F, Dendroseris
(Phoenicoseris) pinnata; Masatierra, Juan Fernandez Is. G, Dendroseris (s.s.) litoralis; Santa
Clara, Juan Fernandez Is.

in dispersal loss. As in Centaurodendron and Yunquea, this alteration is
accompanied by entry into the rain-forest habitat, an ecological preference
of relatively few Senecioneae.

The tribe Cichorieae of Asteraceae is represented on a few Pacific
islands. The achene of Lygodesmia juncea (fig. 11.8A) is typical of main-
land forms. The achene is crowned by numerous pappus setae; these reflex
at maturity, as do the involucral bracts. The setae in this species are per-
sistent. In the same subtribe is an insular endemic, Stephanomeria (Mun-
zothamnus) blairti, of San Clemente Island, off the coast of California.
Stephanomeria blairii (fig. 11.8B) does not seem to show appreciable dimi-
nution of dispersibility. However, pappus setae are more caducous than in
many of the tribe.

The remaining genera of Lactuceae (Cichorieae) shown in figure 11.8
belong to subtribe Dendroserinae, a subtribe claimed to be close to Steph-
anomeriinae (Stebbins, 1953). Thamnoseris (fig. 11.8C) is endemic to
the Desventuradas Islands, Chile, where it grows on barren lava slopes
(Skottsberg, 1963). Involucres of Thamnoseris are composed of bracts
with large swollen bases; involucres apparently do open when mature and
dry to release fruits (Skottsberg, 1947). Pappus setae are short and “fall
very easily” from the achene body (Skottsberg, 1937). These pappus fea-
tures suggest some diminution of dispersal capability.

459

Loss of Dispersibility in Island Plants

The Juan Fernandez Cichorieae (fig. 11.8C—G) form an unusually rich
assemblage which Skottsberg (1951) concluded were best regarded as four
genera, although Stebbins (1953) has demurred. Dendroseris (Hespero-
seris) gigantea (fig. 11.8D) has achenes with fairly good potential dispersal
ability, although pappus setae are relatively caducous. In Dendroseris (Rea)
micrantha (fig. 11.8E), pappus setae are very short and caducous. Involucral
bracts are not caducous in D. (R.) micrantha, although they are in D. (R.)
pruinata.

Dendroseris (Phoenicoseris) pinnata has achenes which are oddly
angled, sometimes triquetrous, but which have strikingly convolute sur-
faces (fig. 11.8F). This is, in part, related to embryo volume, but the con-
siderations cited above for Yunquea may also apply here. Pappus setae in
D. pinnata are too short to offer appreciable lift to achenes released into
the air. Setae do not reflex well in mature dry achenes and are quite cadu-
cous. Moreover, mature involucres open poorly when dry, and seeds are
tardily and irregularly released.

All four species of Dendroseris sect. Dendroseris have achenes with
fantastically sculptured, flattish achenes (fig. 11.8G). The shape and size of
the achenes is related to the broad cotyledons of the embryos, as noted by
Skottsberg (1922). In their large size, fruits of these Dendroseris species
represent an evolutionary culmination in Lactuceae comparable to that of
Fitchia speciosa in Heliantheae. Also contributing to minimized dispersi-
bility in the sect. Dendroseris species is the tuft of short, caducous pappus
setae, far too short and too few either to release achenes from the involucre
or to aid in their aerial flotation. Achenes are, in fact, retained at maturity
within the well-closed involucres composed of massive involucral bracts.
The reasons for highly contorted achene shapes may, in part, be the same
as those given for Yunquea, although the value of a larger embryo might
relate to forest conditions.

Genera of Astereae in the Pacific show a few instances of apparently
diminished dispersibility. Achenes of Remya (Hawaiian Is.) bear a few
vestigial bristles. Remya is difficult to discuss in connection with dispersi-
bility, however, for its affinities are unclear. Many species of the genus
Tetramolopium on the Hawaiian Islands appear to have a pappus typical
of the tribe Astereae, but at least one species (T. remyi) has few setae.

Campanulaceae. Attempts to germinate seeds of the baccate Hawai-

460

Loss of Dispersibility in Island Plants

ian lobelioids tend to meet with poor success, probably because of very
short viability (J. F. Rock, personal communication, 1960); this could be
considered a phase of loss of dispersibility. Seeds of one of the baccate
genera, Delissea, are relatively large. On the other hand, as mentioned
earlier, Trematolobelia, a capsular Hawaiian genus of lobelioids, has prob-
ably evolved improved dispersibility autochthonously by virtue of its
markedly winged seeds (Carlquist, 1962).

Caryophyllaceae. The Hawaiian endemic genus Schiedea would seem
to retain in at least some species the same dispersibility of ancestral types
that immigrated to the Hawauan Islands. Seeds in this genus are most
frequently very small and sculptured in various ways that would enhance
their adhesion within mud or debris on birds’ feet. Capsules open so that
the valves spread widely, providing unimpeded release of seeds. Schiedea
globosa (fig. 11.9) serves as an example of good retention of dispersibility,
as do other lowland species. Exceptions in regard to seed sculpturing are
found in species of somewhat higher altitudes, S. amplexicaulis, S. kealiae,
and S. diffusa, which have smooth seeds. This trend is also shown in S. vis-
cosa (fig. 11.9) and S. lychnoides, both from middle elevations on Kauai;
these species not only have somewhat larger seeds than do the other
Schiedeas, but they also have prominent sepals which enfold the capsule
at maturity, hindering the release of seeds. This suggests a shaker type of
dispersal, in which seeds are disseminated more gradually. Hillebrand
(1888) claimed that S. viscosa and S. lychnoides are transitional to another
endemic Hawaiian genus, Alsinodendron; possibly they may be included in
the latter (Sherff, 1944). Alsinodendron trinerve (fig. 11.9) is notable not

i 5mm. , fruits

Schiedea (Alsinodendron) viscosa Alsinodendron trinerve

Euphorbia celastroides Euphorbia clusiaefolia

Fic. 11.9. Comparison of fruits and seeds of the endemic Hawaiian genera Schiedea and
Alsinodendron (Caryophyllaceae) and of species of Euphorbia (Euphorbiaceae) endemic to the
Hawaiian Is. Schiedea viscosa has also been treated under Alsinodendron and may be inter-
mediate between the two genera. Dotted line around seed of E. celastroides indicates presence .
of a gelatinous sheath when seeds are wet. Scales are shown for fruits and seeds of Caryo-
phyllaceae and for fruits and seeds of Euphorbiaceae.

461

Loss of Dispersibility in Island Plants

only for seeds which are markedly larger than those of Schiedea, but also
for sepals which enclose the capsule and are fleshy, becoming dark blue
at maturity. This suggests a drupelike structure that might well represent
a shift to dispersal by frugivorous birds. Certainly A. trinerve represents
the furthest penetration in this series into a forest habitat, for it grows in
wet scrubby forest near the summit of Oahu’s highest mountain, Mt.
Kaala.

Convolvulaceae. The Hawaiian Breweria menziesii is notable for its
large ovoid seeds, 8-9 mm long, which are often borne singly in capsules.
The precise affinities of B. menziesii need to be established. Other species
of Breweria, however, have smaller seeds. Breweria media of Australia has
seeds 2 mm long, two or more per capsule. Breweria cordata (= Bonamia
semidigyna) of Indonesia has four seeds per capsule, each 5-6 mm long.
The fleshy outer testa of seeds of B. menziesii suggests attractiveness to
birds, a feature that may or may not have been present in ancestral stocks.
This testa is said to be mucilaginous, but is probably not sufficiently so
to permit adhesion to bird feathers. Seeds of Breweria do not float.

Jacquemontia sandwicensis has nonbuoyant seeds. This characteristic
was interpreted by Guppy (1906) as indicating that seeds arrived in crevices
of logs that rafted to the Hawaiian Islands. Ridley (1930) reached what
appeals to me as the more logical conclusion, that J. sandwicensis has been
derived from a littoral species from America that had buoyant seeds.

Cucurbitaceae. Guppy (1906) noted that some Hawaiian species of
Sicyos have hooked fruits, whereas others do not. Although the vector for
Sicyos is not clear, hooked fruits would seem to aid in dispersibility by
birds.

Ebenaceae. The two Hawaiian species of Diospyros differ appreciably
in fruit and seed size. The smaller-fruited species, D. ferrea, has fruits ca.
18 mm long. These fruits are elongate, and thus much smaller in volume
than those of the other Hawaiian species, D. hillebrandii. Fruits of D. hille-
brandi are 20-25 mm long, and 15-18 mm wide. Size of seeds parallels
that of fruits: a typical seed of D. ferrea is 13 X 7 X 4 mm, whereas in
D. hillebrandii a comparable seed is 17 X 10 X 6 mm, approximately
twice as great in volume. Appropriately, D. hillebrandti is much more re-
stricted in distribution and ecology, occurring at about 2000 ft in certain
valleys on Oahu, whereas D. ferrea grows at lower elevations, even down

462

Loss of Dispersibility in Island Plants

to near sea level, on all of the major islands. The fruit and seed sizes of
other Pacific species of Diospyros, such as ““Maba elliptica” and “M. savat-
iensis’” of Samoa, are slightly smaller than those of D. ferrea (Christopher-
sen, 1935). Evolutionary increase in size of fruits and seeds may be sus-
pected in D. samoensis (non Maba samoensis Hieron.), the fruits of which
are 45-50 mm long, with seeds about 30 X 9 X 5 mm.

Elaeocarpaceae. The Hawaiian species of Elaeocarpus, E. bifidus, has
fruits which are ovoid, 22-30 mm long, 18-20 mm wide. This size seems
best regarded, in view of the trends demonstrated in this chapter, as an
increase over fruit size in ancestral species. Large-fruited species of Elaeo-
carpus also occur on Samoa: 4 cm long in E. magnifolius, 5 cm long in E.
ulianus (Christophersen, 1935). Fruit gigantism may characterize a num-
ber of genera in the Samoan flora, however. Species with much smaller
drupes are scattered throughout the Pacific. Elaeocarpus rarotongensis of
Rarotonga has drupes 12 mm long (Brown, 1935). Elaeocarpus hookerianus
of New Zealand has drupes 8 mm long, and several Australian species have
fruits in this range. Smaller-fruited species such as these may be suspected
of being ancestral to the Hawaiian E. bifidus and to the Samoan species.
Frugivorous birds are probably the agent of dispersal. Fruits of Elaeocarpus
do not float.

Euphorbiaceae. The native Hawaiian species of Euphorbia (except E.
haeleeleana) belong to the subgenus Chamaesyce (Anisophyllum of some
authors [see Sherff, 1937c]) and range from seacoast to wet rain forest.
Seacoast species, e.g., E. degeneri and E. skottsbergii, suggest in morphol-
ogy a relationship to species of southeastern Polynesia: E. atoto, E. taiten-
sis, and E. pitcairnensis. Such coastal species as these may disperse by
oceanic drift. Guppy (1906) lists seeds of E. atoto as floatable by virtue of
air space within the seed coat. Mid-altitude Hawaiian species of Euphorbia
may represent a different introduction; they are clearly different in their
basic dispersal mechanism—which may be the same as that by which their
ancestors arrived on the islands. Seeds of the low-altitude (but inland) spe-
cies E.. celastroides (fig. 11.9) are small (1.5 X 1 X 1 mm) and are ridged
and covered, when wet, by a prominent gelatinous coating. This suggests
adherence to birds’ feathers as a means of dispersal. Species of high, wet
forests, however, show marked fruit and seed gigantism. In E.. clusiaefolia
(fig. 11.9), capsules range from 6 to 9 mm in length, and seeds are 3 mm in

463

Loss of Dispersibility in Island Plants

diameter. Not only are these seeds smooth, rather than markedly pitted
or ridged, but no appreciable gelatinous coating can be observed on them.
This would seem clearly to indicate loss of dispersibility. Seeds of neither
E. celastroides nor E. clusiaefolia float, and there is no reason to believe
that their ancestors had floatable seeds. The Hawaiian species which pos-
sesses the largest capsules is E. rockii. Unfortunately, all capsules available
for study contained nothing but markedly shrunken seeds that were clearly
infertile. Because the capsules of E. rockti are so much larger (13-22 mm
in length) than those of E. clusiaefolia, one would expect larger seeds as
well. In any case, the large-seeded species of Euphorbia, which are re-
stricted to upland wet-forest areas, have clearly diminished dispersibility.

Fabaceae. Because native Hawaiian legumes are chiefly found in
lower dry forest—down to the seacoast in many cases—and because legume
seeds are notably capable of seawater transport without loss of viability
(Guppy, 1906; Ridley, 1930), one might suspect that the Hawaiian legumes
as a whole stem from such a history of immigration and have to various
degrees evolved into forest habitats. Moreover, almost all genera in the
indigenous Hawaiian legume flora contain non-Hawauian species that are
widespread beach species, known to be capable of seawater flotation. Loss
of floatability within the Hawaiian legumes would represent a discontinua-
tion of this means of dispersal. Some genera within the Hawauan legumes
do have seeds which are black and shiny or colorful, suggesting attractive-
ness to seed-eating birds. Pods in most Hawaiian legumes open poorly,
however. Moreover, the large size of seeds in these species—like that in
related non-Hawaiian species—suggests that such seed-eating birds would
have to be large. Such birds have not played any appreciable role,
apparently, in long-distance dispersal, and so oceanic drift seems the best
hypothesis.

Seeds of the endemic Erythrina sandwichensis (fig. 11.10) will not
float, although seeds of other Erythrina species do float. Guppy (1906)
reported buoyancy of seeds in E. indica and E. ovalifolia, but failure of
flotation in E. sandwichensis (then regarded as conspecific with the
Tahitian E. monosperma). I have found that seeds of E. variegata (fig.
11.10) also float. Moreover, the legumes of E. variegata remain closed and
are themselves a flotation mechanism. Guppy attributed floatability of
Erythrina seeds to a “light kernel,” since the testa does not float and

464

Loss of Dispersibility in Island Plants

Colubrina asiatica an
(Paleotropical incl. Hawaii) C. oppositifolia

Erythrina variegata (Paleotropical) E. sandwichensis

Fic. 11.10. Seeds of endemic Hawaiian Fabaceae (Erythrina sandwichensis) and Rhamnaceae
(Colubrina oppositifolia) compared with wide-ranging species (E. variegata, C. asiatica) that
are dispersed by seawater. Insets beside seeds of the Erythrinas are drawings of parenchymas of
cotyledons. A prominent intercellular space system (black) is present in E. variegata but not
in E. sandwichensis. Transections of seeds are shown beside the drawings of entire seeds in
the Colubrina species. Prominent spaces between cotyledons and endosperm in C. asiatica are
indicated in black. Scales are shown for seeds of Fabaceae and of Rhamnaceae.

mature seeds contain no air space. The difference in specific gravity
between floatable Erythrina seeds and those of E. sandwichensis was
investigated by means of hand sections of living seeds. My sections (fig.
11.10) show that a prominent air-space system in cotyledons of E. variegata
is probably responsible for flotation; air spaces are much smaller in
cotyledons of E. sandwichensis. Legumes of E. sandwichensis could serve
for seawater dispersal if they remained closed but, in fact, they do not.
The only probable agent for dispersal might be the seed-eating honey-
creeper Psittarostra or some similar bird; if so, a shift in dispersal mecha-
nism might be hypothesized. There seems little doubt that ancestors of
E. sandwichensis arrived on the Hawaiian Islands by oceanic drift. The
near-identity of E. sandwichensis with the Tahitian E. monosperma and
the tendency of both to occupy coastal stations, stations so remote
geographically from those of the nearest other Erythrina species, would
tend to support this idea.*

A similar situation obtains with the Hawaiian koas (Acacia koa and
A. kauaiensis). Seeds of these species will not float, and unopened mature
pods float only for a day or two. Dried branches bearing unopened but
mature pods might be washed down to sea during torrential storms, and
such “rafting” might be envisioned as a means of dispersal for A. koa.
Acacia koa is very similar to A. heterophylla of Mauritius (Indian Ocean);
the distance between the stations for these species alone would suggest

* Study of the type specimen of E. monosperma shows that E. sandwichensis should
be regarded as a synonym of E. monosperma.

465

Loss of Dispersibility in Island Plants

the likelihood of seawater dispersal. A phyllodial Acacia similar to, but
perhaps not immediately related to, A. koa occurs on Samoa and Tonga:
A. simplicifolia. The legumes of this species, unlike those of A. koa, do
not open readily at all, and because each seed is sealed into a loment-like
segment which contains an air space adequate to float an enclosed seed,
flotation would appear likely. Seeds of A. simplicifolia will not float when
removed from the legume. The legumes, or segments thereof, will float for
several days, but eventually become wetted and sink. Acacia simplicifolia
is a beach species and, like the Hawaiian Acacias, doubtless owes its arrival
in Polynesia to oceanic drift. Probably Polynesian Acacias came from a
source in or near Australia, where phyllodial Acacias are abundant. Pods
of A. koa seem to show vestiges of adhesion between the two halves of
the legume in zones which separate adjacent seeds, suggesting the A.
simplicifolia condition. ‘The fact that most pods of A. koa open readily
means that legumes of the species are now ill-suited for seawater dispersal.
Koa seeds may have been distributed over short distances among the
Hawaiian Islands by birds.

The endemic Hawaiian Sophora, S. chrysophylla, forms another
parallel. Seeds of S. chrysophylla will not float, a fact noted by Guppy
(1906) and confirmed by me. Legumes of this species, however, do not
open readily. They are, moreover, formed with narrow constrictions be-
tween some of the seeds, so that either the entire legume or parts of it
can float. Guppy claims that the legume in this species is too delicate
to provide good flotation. Nevertheless, Sophora undoubtedly owes its
original advent on the Hawaiian Islands to oceanic drift. Both legumes
and individual seeds of S. tomentosa and S. tetraptera float well. Legumes
in these species are more durable than those of S. chrysophylla. At least
some species of Sophora are still littoral, such as S. tomentosa. Loss of
dispersibility in S. chrysophylla is shown not only in fragility of legumes
and failure of seeds to float, but also in its adaptation to inland dry forest
primarily; to the extent that it has left the coastal areas, it could not
disperse by seawater regardless of floatability (except for rafting of branches
washed down in streams). Sykes and Godley (1968) find that S. prostrata
has nonbuoyant seeds, which would constitute a loss of dispersibility in
this New Zealand species. Sophora microphylla (New Zealand, Gough I.,
southern Chile) has various degrees of floatability in seeds (Sykes and

466

Loss of Dispersibility in Island Plants

Godley, 1968)—perhaps this species is in the process of losing dispersibility.
Sophora chrysophylla in the Hawaiian Islands may represent a change to
a new mode of dispersal if, as Guppy (1906) claimed, seeds are eaten by
birds and successfully transported short distances within the archipelago.

The example of Canavalia was discussed in the introductory portion
of this chapter. Sauer (1964) reports nonbuoyant seeds in the endemic
Hawaiian species C. galeata, C. molokaiensis, C. hawatiensis, C. kaudiensis,
and C. pubescens. I confirmed nonbuoyancy for the latter three species.
Sauer seems to believe that species such as these were derived from a
wide-ranging ancestor with buoyant seeds, an interpretation my studies
would endorse. Other Pacific species that are restricted to one island or
several nearby islands and that have nonbuoyant seeds (thereby suggesting
loss of dispersibility) include C. vitiensis (Fiji), C. raiateensis (Raiatea)
and C. megalantha (Marianas Is.). Wide-ranging species of the Pacific
with buoyant seeds include C. sericea and C. maritima (Sauer, 1964).

Seeds of the Hawaiian endemic Mezoneurum kauaiensis showed no
floatability in my experiments. Mixed results were obtained with seeds
of Strongylodon lucidus: of five seeds tested, only one floated. Seeds of
Mucuna gigantea (a coastal element in the Hawaiian flora) did float.

These results suggest that loss of buoyancy—and thereby dispersi-
bility—has occurred quite conspicuously in endemic legumes of Pacific
islands. Loss of dispersibility in these seems to have occurred almost in
direct proportion to the extent to which the species have adapted to
upland forest conditions.

Goodeniaceae. The Hawaiian species of Scaevola represent three main
groups: S. taccada (= S. sericea), which is widespread on beaches through-
out the Pacific; species such as S. gaudichaudiana and S. mollis with small,
ovoid, purple drupes, which are representatives of a montane group of
species that extends from Australia to the Hawaiian Islands; and S. glabra
(including S. kauaiensis), a curious and isolated species of high wet forest
on Oahu and Kauai, which has sometimes been treated under a segregate
genus, Camphusia. Among the second group of species, some variation in
size of drupe and endocarp does occur, but drupes mostly range from
4 to 10 mm in length, with the woody endocarp 3-7 mm long. These
species seem easily dispersible by frugivorous birds. Well outside that
range of size is the fruit of S. glabra, which is spheroidal or pyriform,

467

Loss of Dispersibility in Island Plants

15-20 mm long. The endocarp is typically ovoid, ca. 13 X 8 mm. The
fruit dimensions of S. glabra clearly exceed those of any other Scaevolas
of montane Pacific islands. Scaevola glabra probably represents an early
immigrant to the islands, a hypothesis suggested both by its curious
morphology and its restriction to relatively older mountains. If so, the
greater time available for its evolutionary change on the Hawaiian Islands
may have permitted it to approximate more nearly the large-fruited habit
characteristic of trees of wet forest.

Scaevola coriacea, although a species from coastal and near-coastal
areas in the Hawaiian Islands, has fruits incapable of flotation (chapter 2).
It is probably a derivative of the montane species and never had buoyancy,
despite a superficial resemblance to the widespread S. plumieri, a purple-
fruited species with buoyant seeds. Scaevola coriacea is probably dispersed
by birds. Likewise, S. spinescens, an Australian species with nonbuoyant
seeds (chapter 8), may never have had seed buoyancy, despite its occurrence
near salt lakes. For further comments on dispersibility of Goodeniaceae,
see chapter 8.

Occasionally I have been asked whether the purple-fruited montane
Scaevolas of the Pacific are derivatives of S. taccada, the beach species
with buoyant, white-fruited seeds. There seems little doubt on this point:
S. taccada is a relatively uniform species, and although it is like the
montane, purple-fruited species in some respects, the montane species are
closely related to each other, and each is endemic to an island or an
archipelago. ‘There seems no likelihood that S. taccada could have given
rise to each of the montane species independently, yet left no trace of any
intermediates.

Haloragidaceae. The Hawaiian species of Gunneva cannot be said to
be large-fruited in comparison with forest trees, but they do represent a
marked increase in fruit size over species of other Pacific islands. Dried
fruits of G. mauiensis (Hawaiian Is.) are 3 mm in diameter, whereas those
of G. macrophylla (New Guinea) are ca. 1.5 mm in diameter and those of
G. prorepens (New Zealand) ca. 1.2 mm in diameter.

Lamiaceae. Phyllostegia is endemic to the main Hawaiian Islands
except for P. variabilis on Laysan Island and P. tahitensis on ‘Tahiti.
Phyllostegia and an endemic Hawaiian genus, Stenogyne, have been placed
together in the tribe Prasieae. This tribe is distinctive in possessing genera

468

Loss of Dispersibility in Island Plants

with fleshy drupelets rather than dry nutlets. Phyllostegia and Stenogyne
may have evolved from such ancestors as the Indo-Malesian Gompho-
stemma, which also possesses drupelets. Because of fruit morphology,
dispersal by frugivorous birds may be suspected. Among the 24 species
of Phyllostegia recognized by Sherff (1935a), there is variation in size of
the drupelets, but data are difficult to obtain because mature drupelets
are unknown in some species, and although the size of dry endocarp 1s
known for the remainder, dimensions of fresh fruits are rarely recorded.
The average length of dried drupelets in Phyllostegia appears to be about
2 mm, with fresh drupelets 4 mm long. The related genus Stenogyne
(fg. 11.11) shows a more dramatic series in increase of drupelet size.
Species belonging to the section Microphyllae of Sherff (1935a), such as
S. diffusa, have small drupelets. Dried drupelets of these species are ca. 2-4
mm long. These species inhabit dry, grassy areas on recent volcanic craters,
such as Haleakala, Mauna Kea, and Mauna Loa. Most species, such as
S. calaminthoides (fig. 11.11), have drupelets which are 3-5 mm long when
dried. A few species (usually those of very wet forest areas) have drupelets
6-8 mm long. These species grow in patches of forest adjacent to the
high wet bogs of the Hawaiian Islands: S. kaalae (Mt. Kaala, Oahu), S.
kamehamehae (Puu Kolekole, Molokai), and S. purpurea (Alakai Swamp,
Kauai). The large-fruited Stenogyne shown in figure 11.11 is presumably
referable to S. purpurea, although it differs in leaf shape and in the size
of drupelets, which are appreciably larger than indicated for that species
by Sherff (1935a). The series in fruit sizes within Stenogyne appears closely
correlated with ecology and perhaps indicates tendencies toward precinc-
tiveness. Increase in fruit size is accompanied in some species by a tendency

Haplostachys linearifolia H. munroi

Fic. 11.11. Fruits of the Hawaiian genera Stenogyne and Haplostachys (Lamiaceae). Seeds
with fleshy exocarp removed are shown to the right in each Stenogyne species. Fruits are
shown with the facing portion of the enclosing calyx removed, except for Stenogyne purpurea.

469

Loss of Dispersibility in Island Plants

to bear fewer drupelets per flower. Another indication of lowered dispersi-
bility might be the habit of some species of Stenogyne (e.g., S. rotundifolia)
of enveloping drupelets within a persistent tubular calyx which may hinder
shedding of fruits.

The Hawaiian genus Haplostachys differs from Stenogyne and Phyl-
lostegia in its dry nutlets. Fosberg (1948) suggests a common origin of
this genus with the other two. Most species of Haplostachys, such as
H. linearifolia (fig. 11.11), occupy dry grassy areas only a few feet above
sea level. Haplostachys munroi (fig. 11.11), with larger fruits, grows in dry
forest areas on Lanai. Fruiting calyces and also individual nutlets of all
species of Haplostachys tested are capable of flotation in water, a result
inferable from the coastal habitat of the species, but also suggesting
possible derivation of the genus independently of Phyllostegia and Steno-
gyne. The corky nature of the nutlet exocarp is responsible for floatability
of the fruits. The larger seed size in H. munroi seems related to the entry
of that species into the forest habitat. Occasional fruits of H. munroi are
composed not of four triquetrous nutlets but of two larger hemispherical
ones. Nutlets of H. munroi are often somewhat misshapen, with irregular
margins,

Lauraceae. Cryptocarya in the Hawaiian Islands is represented by two
close species, C. oahuensis and C. mannii. Both are unusual in the genus
in their large fruit size. Drupes in these species are spheroidal, 30-35 mm
in diameter when dry; the figure of 16 mm given by Rock (1913) is
probably in error, perhaps based on immature material. These fruits are
covered by a thin, fleshy exocarp which overlies a thick (2-4 mm) woody
endocarp. These features are in contrast with those of other Cryptocarya
species, in which fruits are much smaller and have a thin, almost papery
endocarp which can be broken easily. In the Australian species C. angulata
and C. hyposporlia, fruits are ovoid, 11-12 * 8 mm when dry. Dried fruits
of the Fijian C. hornei are somewhat larger, 16 X 14 mm. The presence
of a fleshy, though thin, blackish exocarp suggests that frugivorous birds
would be responsible for dispersal of these fruits. If so, the large size of
fruits in the Hawaiian Cryptocaryas would seem to have foreclosed this
avenue of dissemination. Large-fruited Cryptocaryas are also present in
Samoa: C. samoensis has drupes 40 X 17 mm, and C. glaucescens has
similarly large fruits (Christophersen, 1935). Dried fruits of Cryptocarya

470

Loss of Dispersibility in Island Plants

containing well-developed embryos are capable of flotation, but the fragile
nature of endocarp in some species—as well as the probability that they
will not be viable when thoroughly dried and washed out to sea—suggests
that oceanic drift is a relatively unlikely means of dispersal. Cryptocaryas
are not lowland forest trees, which one would expect if they were, in fact,
seawater dispersed. In the Hawaiian Islands, Cryptocarya occurs only in
the geologically oldest areas (Kauai; Waianae Mts., Oahu), a fact which
suggests that time has been available for evolution of the gigantic fruit size.

Malvaceae. The problem of whether the Hawaiian and Tahitian cot-
tons are native has been discussed by Stephens (1963, 1964), who con-
cludes that they are probably native and that a hypothesis of prehuman
natural arrival by means of oceanic drift seems most acceptable. Guppy
(1906), Stephens (1958), and I (1966b) have indicated relatively poor
floatability for seeds of the Hawaiian cotton, Gossypium sandvicense. How-
ever, the matter of seed floatablity and resistance to salt water in cottons
has been more recently and exhaustively investigated by Stephens (1966).
Seeds of G. sandvicense (Hawaiian Is.) and G. hirsutum (various Pacific
islands; Puerto Rico) show sufficiently long buoyancy to explain their
arrival on these various islands. Saltwater tolerance is retained in G. sand-
vicense and in G. hirsutum from Wake Island. Stephens (1966) seems
mystified by loss of saltwater tolerance in G. hirsutum from Samoa and
the Marquesas. The explanation is clearly a loss of dispersibility for rea-
sons such as suggested in the opening pages of this chapter. Seeds of most
cottons retain great viability after soaking in seawater (Stephens, 1958),
and this fact, together with their typical preference for habitats along
coasts, strongly suggests oceanic drift as the means of dispersal basic to the
genus. In addition to lack of subspecific variation, the presence of G. sand-
vicense on all major islands of the Hawaiian chain shows that it can utilize
oceanic drift for interisland dispersal.

The peculiar endemic Hawaiian genus of tree-cottons, Kokia, may
have been derived from a Gossypium-like ancestor. With Kokia, the prob-
lem is much like that of Gossypium. Kokia species occupy low arid sites,
characteristic of species introduced by oceanic drift. Seeds of K. cookii and
K. rockti float for several days, but sink as soon as the fibers become thor-
oughly wetted. This degree of floatability might suffice for dispersal of
the genus throughout the major Hawaiian islands. Capsules in Kokia float

47]

Loss of Dispersibility in Island Plants

only for a few hours. The capsules ordinarily flare widely, releasing seeds.
Seed flotation, as with Gossypium, seems possible for Kokia, but one must
hypothesize a loss of dispersibility if ancestors did arrive by means of
oceanic drift.

Precisely the same results as with Kokia were obtained with another
Hawauan genus, Hibiscadelphus. ‘Vests were made on seeds and capsules
of H. giffardianus, H. hualalaiensis, and H. wilderianus.

Of the endemic species of Hibiscus tested (H. arnottianus, H. brack-
enridgei, H. immaculatus, and H. waimeae), none had flotation ability in
either seeds or capsules. If these species did stem from seawater-dispersed
ancestors, this ability has been entirely lost. Seeds of the widespread
littoral Hibiscus of the Pacific, H. tiliaceus, do float and have undoubt-
edly dispersed in this fashion. In addition, tests on seeds of the non-
insular H. mutabilis showed that seeds, even when deprived of their cov-
ering of hairs, float indefinitely. Dispersal of Hibiscus species throughout
the Pacific, with loss of dispersibility following establishment, might still
be postulated. Dispersal of hispid seeds in bird feathers is another possi-
bility that cannot be ignored. Seeds of a few native Hawaiian Hibiscus
species still possess some stiff hairs. In either case, autochthonous loss of
dispersal capability seems to have occurred in Hawaiian species, a loss
correlated with evolution into moist upland sites.

Formation of endemic species in Gossypium and other Malvaceae, as
in insular legumes, especially those of Hawaii, may be explained by infre-
quency in events of oceanic drift. This concept and reasons why oceanic
drift may operate on rare occasions to particular islands are discussed in
chapter 2.

Myrtaceae. An excellent series in loss of dispersibility related to
mountaintop endemism is demonstrated by species of Darwinia from
southwestern Australia (chapter 8). Attention is also called to other in-
stances of loss of dispersibility cited there, as in Eucalyptus macrocarpa
and Stylidium galioides.

Nyctaginaceae. The native Hawaiian Nyctaginaceae probably show
little loss of dispersibility, if any, with respect to fruit morphology. The
fact that Pisonia (Ceodes) umbellifera is absent from some suitable islands
in the Pacific might be due to the large size of the fruiting perigones, al-
though the perigones are certainly markedly viscid. The same may apply

472

Loss of Dispersibility in Island Plants

to Pisonia (Rockia) sandwicensis, which is endemic to the Hawaiian Islands.
I have cited this species as an example of loss of dispersibility (Carlquist,
1965). The sticky perigones suggest retention of ability for dissemination
by birds, but the ecological preference of the species—forests between
2000 and 4000 ft altitude—would place it well out of the range of seabirds,
its best means for long-distance dispersal. The loss of dispersibility in P.
sandwicensis may consist not of morphological change but of ecological
shift and attendant cessation of contact with long-range vectors. Its en-
demism is suggestive of loss of the link with vectors.

Pittosporaceae. Fruiting structures that are exceptionally viscid can
be quite large and still presumably retain good dispersibility. Increase in
size of seeds in the Hawaiian species of Pittosporum may not, therefore,
actually cancel their ability to be trznsported, although it would seem to
diminish the probabilities. Species of Pittosporum with very large seeds
have evolved in the Hawaiian Islands, a gigantism that may be primarily
telated to ecological adaptation to shady forests. Also, adaptation to high
forests puts Pittosporum out of the range of seabirds. According to Rock
(1913), the Hawaiian crow seeks Pittosporum seeds; the very localized dis-
tibution of the Hawaiian crow rules it out as a vector of Pittosporum
seeds, but other birds might carry them. Seeds of most of the Hawaiian
species of Pittosporum recognized by Sherff (1942) are somewhat discoid
and 5—7 mm in diameter. Species with larger seeds (7-9.5 mm) include P.
sulcatum, P. dolosum, P. cauliflorum, P. hawaiiense and P. hosmeri (fig.
11.12). Notably long seeds occur in a few, such as P. confertiflorum (fig.
11.12). The gigantism is evident when one compares other Pacific Pitto-
sporums: P. niueana (4-5 mm); P. taitense (ca. 4 mm [fig. 11.12]); P. raro-

] 1 cm. J

800 Cad BE@

Pittosporum tenuifolium
(New Zealand) P taitense (Tahiti)

Acaena exigua

P hosmeri longifolium

P confertiflorum mannii A. californica

(California)

Fic. 11.12. Comparison of seeds of Hawaiian and non-Hawaiian species of Pittosporum

(Pittosporaceae) and of Acaena (Rosaceae). Viscid coating is not shown on Pittosporum seeds.

Fruiting calyces of A. exigua have prickles and hairs, but lack retrorse barbs on prickles that

are characteristic of A. californica and other species of Acaena. Scales are shown for seeds of
Pittosporum and for fruits of Acaena.

473

Loss of Dispersibility in Island Plants

tongense (4.5-5 mm), and P. rapensz (5.5-6 mm). Seeds of Australian and
New Zealand Pittosporums are much smaller by comparison: many of
them have seeds no more than 3 mm long, as in P. tenuifolium (fig. 11.12).
Species in the western Pacific also have relatively small seeds: P. mega-
carpum of the Philippines (4-6 mm), P. resiniferum of the Philippines
(3-4 mm), and P. moluccanum of the Philippines and Indonesia (3 mm).

Plantaginaceae. Of the two Hawaiian species of Plantago with the
widest distribution in the archipelago, P. grayana has lost the mucilaginous
quality of the seed coat whereas P. augusta has retained it. Of the four
Plantago species sympatric on Mt. Waialeale, Kauai, two have a non-
mucilaginous seed coat (Melvern Tessene, unpublished data).

Primulaceae. Although some Hawaiian species of Lysimachia have
large woody capsules (such as L. forbesii) that seem to open poorly, seeds
are no longer than 2 mm, a size which suggests ease in dispersal. Gigantism
in fruits without corresponding increase in seed size also characterizes
other Hawaiian genera: Labordia, of Loganiaceae, compared with the
closely related South Pacific genus Geniostoma; Cyanea and Clermontia
of the Campanulaceae, compared among themselves.

Rancunculaceae. ‘The two endemic species of Ranunculus, R. maut-
ensis and R. hawaiiensis, differ in the degree to which achenes are beaked.
Neither species has a stigmatic beak that is hooked. A Ranunculus with a
hooked achene may not, however, have been ancestral to these species, and
travel by means other than mechanical attachment is conceivable. Spe-
cies with the same achene size (2-4 mm) and similar morphology include
R. fibrosus, R. plebius, and R. sieboldii (Java).

Rhamnaceae. As was noted by Hillebrand (1888), the Hawaiian Alphi-
tonia ponderosa has larger fruits (ca. 13 mm in diameter) than A. zizyph-
oides of ‘Tahiti and Tonga (ca. 10 mm in diameter) or A. franguloides of
Fiji. These distinctions in fruit size do not appear great. However, one
should note that in A. zizyphoides and other species the fleshy exocarp and
the woody cocci which the exocarp encloses fall from the mature fruit,
leaving a pair of brownish seeds, surrounded by contrasting red arils, ex-
posed upon the pedicel. This presentation would seemingly offer an excel-
lent display for seed-eating birds. However, Hillebrand (1888) notes that
in the Hawaiian A. ponderosa the exocarp does not fall away, so that seeds

474

Loss of Dispersibility in Island Plants

are not, in fact, exposed. To the extent to which this fruit dehiscence fails
to occur, A. ponderosa can be said to have lost dispersibility.

Two species of Colubrina are native to the Hawaiian Islands. Of
these, C. asiatica is wide-ranging in the Pacific and occurs near the sea-
coast on the many islands where it is found. Colubrina oppositifolia, how-
ever, is endemic to the Hawaiian Islands and occurs inland in the lower
scrub forest. The two species differ in floatability of seeds, as noted by
Guppy (1906) and confirmed by me. However, Guppy’s observation that
seeds of C. oppositifolia would float “for a week or two” may be based on
faulty material. Specimens I tested had germinable seeds; these seeds
sank immediately upon being dropped into water. Seeds of C. asiatica
float indefinitely, as Guppy claims. Anatomically, tissues of seeds in the
species appear similar. Ability of seeds of C. asiatica to float is probably
due to a prominent space that forms between cotyledons and endosperm
(fig. 11.10). I observed no such space in my seeds of C. oppositifolia (fig.
11.10). When the testa is removed from seeds of C. asiatica so that water
can penetrate the air space, they sink. In addition, the seeds of C. oppositi-
folia are larger than those of C. asiatica. As noted by Hillebrand (1888),
fruits of C. oppositifolia do not dehisce well: “The epicarp woody and not
separating from the endocarp, therefore the cocci separating tardily and
imperfectly.” Seeds of C. asiatica, on the contrary, are readily released.
Colubrina oppositifolia may be assumed to have originated from ancestors
like C. asiatica that reached the Hawaiian Islands through oceanic drift.
Colubrina oppositifolia is a good example both of loss of dispersibility and
increase in seed size. The latter characteristic may represent an adjust-
ment to the forest habit, for C. oppositifolia is a tree whereas C. asiatica
can only be regarded as a shrub.

Rosaceae. The Hawaiian Acaena, A. exigua, appears closer in mot-
phology to the California A. californica than to any of the Acaenas of the
South Pacific. On New Guinea, New Zealand, and the Juan Fernandez
Islands are species of Acaena in which the fruit-enfolding calyx bears only
a few long prongs, each tipped by a prominent crown of retrorse barbs. On
both A. californica and A. exigua, however, the calyx is covered by a scat-
tering of short prickles (fig. 11.12). In A. californica, these prickles bear
retrorse barbs. In A. exigua, the barbs are lacking. The small size of fruit-
ing calyces in A. exigua suggests that it presents no problem for dispersal.

475

Loss of Dispersibility in Island Plants

However, loss of barbs may well have occurred after establishment on the
Hawaiian Islands. If so, it would denote a loss of dispersal mechanism simi-
lar to the loss of retrorse barbs in Bidens (Asteraceae).

Of the two endemic species of Rubus on the Hawaiian Islands, R.
macraei is conspicuous because of its large fruits, 2-3 cm. in diameter. The
endocarp of each achene in this species measures ca. 4 X 3 X 3 mm, a
size appreciably larger than that of R. hawatiensis (4X 2 X 1 mm) or
South Pacific species of Rubus, such as New Zealand’s R. australis (2.5 X
Des) SX A,saneaa))

Rubiaceae. Fruits of most drupaceous Hawaiian Rubiaceae are 1 cm
or less in diameter and thus may be considered still readily dispersible by
frugivorous birds. Selected comparisons within these genera would prob-
ably show larger fruit and seed size in the species of wet forest, as well as
greater overall size of fruits and seeds in Hawauan rubiads as compared
with those from continental areas of the western Pacific. For example,
Timonius polygamus from Rimitara (Austral Is.) has drupes ca. 8 mm in
diameter. ‘The seeds are about 2.4 * 2.5 x 1 mm. In Bobea elatior, repre-
sentative of an endemic Hawaiian genus closely related to Timomuus,
drupes are slightly larger (ca. 10 mm in diameter) with fewer and larger
(6 x 3 X 1.5 mm) seeds. Among non-Hawaiian Polynesian Rubiaceae, the
Samoan genus Sarcopygme contains an interesting series of increasing
fruit and seed size.

Rutaceae. Hawaiian Rutaceae belong to three genera: Pelea, Fagara
(perhaps best now called Zanthoxylum), and Platydesma. Pelea is endemic
to the Hawaiian Islands and the Marquesas and is probably most closely
related to Evodia (Euodia) of the South Pacific, although Melicope and
Acronychia of that region also show many resemblances. Platydesma, an
endemic Hawaiian genus, probably evolved from a Pelea-like stock. Fagara
is well represented in the Old World tropics. All three genera represented
in the Hawaiian Islands have follicular fruits that open at maturity, re-
vealing shiny black seeds which, although not fleshy, are doubtless attrac-
tive to birds. Within each of these genera, a trend toward gigantism of
fruits and seeds is evident.

Pelea shows a notably wide range of fruit and seed sizes. The three
species shown in figure 11.13 suggest some extremes and an intermediate

476

Loss of Dispersibility in Island Plants

Evodia samoensis

(Samoa)

Pelea waialealae GO)

Pelea clusiaefolia

Pelea pseudoanisata

I 2 cm. |

Fagara oahuensis °

Fagara ailanthoides (Japan)

Fagara dipetala

Platydesma cornuta @

Platydesma remyi

Fic. 11.13. Comparison of fruits and seeds of Hawaiian and non-Hawaiian Rutaceae. The
genus Pelea, endemic to the Hawaiian Is. and the Marquesas, is closely related to Evodia
(Euodia), one species of which is shown for comparison. For each species of Evodia and
Pelea, a view toward the summit of the fruit and a view of a longitudinal section of a fruit
are shown. There are most often two seeds per follicle in Pelea; because follicles are shown
sectioned, only a single seed is visible. Habits of fruits and seeds are shown for two Hawaiian
and one non-Hawaiian species of Fagara (perhaps better termed Zanthoxylum now). One or
two follicles may develop in the Hawaiian species. Two species of the endemic Hawaiian genus
Platydesma, which may be closely related to Pelea, are shown. Scale applies to all drawings.

size for the genus. Pelea waialealae, with very small seeds, might at first
sight, if the principles suggested here were to apply, seem as though it
should be a lowland species instead of a native of high bogs, as it is. Actu-
ally, the correlation of small seed size with the open bog habitat is under-
standable, because smaller seeds seemingly characterize the bog plants of
the Hawaiian Islands, whereas seeds of forest plants are much larger. This
suggests that one factor operative in increase in fruit and seed size in Pelea
has been adaptation to the large-seeded forest-tree habit; precinctiveness
may also play a role. A feature that appears to accompany increase in fruit
size in Pelea is the tendency for seeds not to be exserted, as they are in
Evodia samoensis (fig. 11.13). The funiculus that in Evodia aids in display
of seeds is present in Pelea but does not have that function. Mature fruits
in Pelea exhibit loculicidal dehiscence of follicles, followed by shattering
of the valves and fall of seeds. Increase in seed size can be seen progres-

Ail

Loss of Dispersibility in Island Plants

sively by the species of Pelea illustrated in figure 11.13, terminating with
P. pseudoanisata.

In the genus Fagara, a wide range of fruit and seed sizes is visible, as
suggested in figure 11.13. The seed size shown for F’. oahuensis represents a
minimal size for Hawaiian Fagaras, as well as one that characterizes all
Hawaiian species except F. dipetala. The same as for F. oahuensis can be
seen in seeds of F. pinnata (Tonga), F. megistophylla (Fiji), and F. gille-
spieana (Fiji). The seed size in the Hawaiian F. dipetala (20 X 14 X 12
mm) is remarkable, and represents a definitive lessening of dispersibility.
Contrasingly minute fruits and seeds are shown for the Japanese species
F’. ailanthoides (fig. 11.13).

Within Platydesma, a marked increase in fruit and seed size may be
seen by comparing P. cornuta and P. remyi (fig. 11.13). A feature in Platy-
desma which is of interest in connection with loss of dispersibility was
observed by Rock (1913) for P. campanulatum (= P. spathulatum): “The
capsule often rots away but the seeds remain attached to the placenta for
some time.”

Santalaceae. ‘The Hawaiian species of Santalum form a good series
with respect to size of drupes. At maturity, drupes in Santalum are reddish
or purplish. The majority of the Hawaiian sandalwoods have drupes 8-12
mm in diameter, a size which, together with the nature of the fruit, makes
dispersal by frugivorous birds likely. Species of Santalum elsewhere, such
as S. album, also have drupes in this size range. Rather small drupes occur
in S. ellipticum var. littorale (6-7 mm long) and S. cuneatum (8 mm in
diameter); both of these are lowland Hawaiian plants, growing on the
coastal plain. Notably large-fruited is S. pyrularium—characteristic of dry
but upland forests of Kauai. In S. pyrularium, drupes attain 23 mm in
length, and ca. 15 mm in width. The endocarp measures about 15 X 10 X
10 mm. This large size suggests that dispersal by birds might be more
difficult.

Sapindaceae. Of all examples of gigantism in fruits in the Hawaiian
flora, perhaps none is so dramatic as Alectryon. This genus of about 30
species ranges from Australia, New Zealand, New Guinea and the Philip-
pines through Melanesia; in Polynesia, it is represented in Samoa and
the Hawaiian Islands. Most species have relatively small fruits, as repre-

478

Loss of Dispersibility in Island Plants

@ Bese

Alectryon carinctus
(New Caledonia)

A. samoensis (Samoa)

Fic. 11.14. Comparison of fruits and seeds of non-Hawaiian species of Alectryon (Sapindaceae)
with those of the Hawaiian A. macrococcus. One or both cocci of the fruits may develop in
all species. At matunty cocci are broken open by the expanding red aril, at the tip of which the
seed is borne. For each species, an opened monococcate and an unopened bicoccate fruit are
shown; also two views of each seed: one lateral view (chalazal end of seed above) and one
polar view (surface opposite chalazal end). Dotted lines around lateral views of seeds of A.
samoensis and A. macrococcus indicate outline of aril. Scale applies to all drawings.

sented in figure 11.14 by A. carinatus and A. excelsus. Cocci of A. carinatus
measure 10 mm in diameter if borne singly; if borne in pairs, each coccus
is ca. 8 mm long, 6 mm wide. In A. excelsus single cocci are 12 mm in
diameter, paired cocci 10 mm each. As the drawings of various species of
Alectryon show, mature cocci open so as to expose the fleshy red aril sur-
rounding the black seed (see photograph of the Australian species A. sub-
cinereum, fig. 2.4, top left). The relatively small seed size of these smaller-
seeded Alectryons, together with the contrast with the red aril, suggests
ease in distribution by frugivorous birds. ‘The seeds in the two species men-
tioned are only 5-6 in diameter, and would be easily swallowed by a bird
seeking the arils. If so, the remainder of the species shown in figure 11-14
may be said to form a series in loss of dispersibility. Alectryon grandifolius
of Fiji has single cocci 22 mm in diameter, paired cocci 18 mm each. The
seed measures about 12 x 10 X 8 mm, and thus by itself would be as
large as the smallest drupe (including fleshy exocarp) that one would ex-
pect birds ordinarily to have carried over long distances. Alectryon samo-

479

Loss of Dispersibility in Island Plants

ensis 1s well beyond this size range. Single cocci measure ca. 35 mm in
diameter, and paired cocci about 30 X 28 mm each. The seeds are about
22 X 20 X 20 mm. The Hawaiian species of Alectryon represent the logi-
cal conclusion of this trend toward gigantism. Alectryon macrococcus and
a closely related segregate species, A. mahoe, have fruits somewhat variable
in size but of enormous proportions. In these species, cocci borne singly
may reach 60 mm in diameter. When borne in pairs, each coccus is ca.
34 X 40 mm. Seeds in these species can easily be 32 X 30 X 20 mm. Dis-
persal of such seeds by frugivorous birds—especially those of the Hawaiian
Islands—is unthinkable. Seeds of this size would tend to drop within a
short range of the parent plant, so that a high degree of precinctiveness
is likely. Tendency to reproduce within the moderately dry mid-altitude
forests where this tree is native may explain the gigantism. The value of
a large amount of storage tissue within seeds does not seem preeminent in
the relatively open forests where the Hawaiian Alectryons live, although
it may be a factor. One can postulate that the immigrant ancestors of the
Hawaiian Alectryons had fruit and seed sizes comparable to those of A.
excelsus. Gigantism in the Fijian and Samoan Alectryons has probably
been achieved parallel to that of the Hawaiian species.

Sapindus oahuensis is also notable for relatively large fruits and seeds.
In this species, only one of the two potential drupelets per flower ordi-
narily matures. Each drupelet when fresh measures ca. 33 X 24 X 20 mm.
The seed measures ca. 21 X 12 X 10 mm. The fleshy exocarp, which has
a datelike odor and texture when drying, suggests that the ancestors of S.
odahuensis were sought by frugivorous birds. There are also species with
large fruits elsewhere in the Pacific, such as S. mukorossi, which occurs in
the Bonin Islands. The wide-ranging S. saponaria, however, has fruits 15—
18 mm in diameter, with most seeds less than 10 mm in length. Such a
species—or one with even smaller fruits—likely was ancestral to S. oahu-
ensis. Neither fruits nor seeds of Sapindus are known to be capable of
flotation.

According to Guppy (1906), Hawaiian species of Dodonaea vary in
the ability of their seeds to float. This suggests ecological shift toward in-
land sites by a typically lowland or littoral genus.

Sapotaceae. The genus Planchonella (Sideroxylon of earlier authors)

480

Loss of Dispersibility in Island Plants

is represented in the Hawaiian Islands by a single highly polymorphic
species, P. sandwicensis. Segregate species were recognized by Rock
(1913). There are distinctive populations within P. sandwicensis, some
notable for distinctive fruit shapes and sizes. Some fruits of P. sandwicen-
sis are only 2 cm in diameter at maturity, with seeds about 17 X 6 X 3
mm. Others, as in the population recognized by Rock as Sideroxylon
rhynchospermum, are 5.5 cm in length, 3.5 cm or more in width, and con-
tain seeds 35 X 14 X 3 mm. Thus, within the Hawaiian Planchonellas
there may be tendencies toward fruit and seed gigantism. The smaller
fruit and seed sizes mentioned are about the same as those of species of
Planchonella on other Pacific islands, such as P. costata (= P. vitiensis) of
Fiji, P. garberi of Samoa, and P. novozelandica of New Zealand. The
sizes of fruits in these species seem excessive for overwater dispersal by
frugivorous birds. Although fruits can be viscid, the seeds are exceptionally
smooth and would not adhere easily to birds. However, tests on flotation
prove that dred fruits of Planchonella float almost indefinitely, perhaps
because the papery endocarp with which each seed is surrounded contains
a generous air space. Individual seeds also will float, at least for several
days. Because Planchonella is a component of lowland forests, one might
suspect that at least some species are distributed by oceanic drift. Also,
the occurrence of the same subspecies (according to Lam’s determination)
on Fiji and on Rurutu in the Austral Islands is suggestive of oceanic drift.

There are a number of Sapotaceae in the New Caledonian flora with
notably large fruits and seeds. One can suspect autochthonous gigantism
in at least some of these.

Strasburgeriaceae. The affinities of the monotypic genus Strasburgeria,
sole genus of the family and a New Caledonia endemic, are not clear. How-
ever, these massive fleshy fruits, which contain fibrous endocarps (for illus-
trations, see Carlquist, 1965), have no evident dispersal mechanism and in
all likelihood constitute an instance of loss of dispersibility.

Winteraceae. | earlier cited Zygogynum, a New Caledonia endemic,
as an example of loss of dispersibility and illustrated the fruit (1965).
Zygogynum seems an outstanding example of this phenomenon. I suspect
that further analysis of the New Caledonian flora would reveal a number
of other instances of loss of dispersibility.

481

Loss of Dispersibility in Island Plants

Monocotyledons

Arecaceae (Palmae). Pritchardia is a palm that ranges from Fiji to the
Hawaiian Islands and that may be regarded as an extension of Livistona of
Asia and Indo-Malesia. The westernmost species, P. pacifica and P.
thurstonii of Fiji, have the smallest fruits, 12 and 7 mm in diameter,
respectively. The Hawaiian Pritchardias were split into a large number
of species by Beccari and Rock (1921); only a fraction of these are likely
to be recognized by most workers. The polymorphic nature of Pritchardia
in the Hawaiian Islands is evident, however. Most of the Hawaiian species
recognized by Beccari and Rock have fruits that vary from 20 to 30 cm in
diameter. Ten of the species they recognize have fruits that exceed 40 cm
in diameter. Of these, some have quite gigantic seeds (6 X 4 cm in P.
lowreyana; 5 X 3 cm in P. rockii; 5.5 * 4 cm in P. montis-kea). Beccari
and Rock (1921) note the irony in the fact that large-fruited species such
as these tend to occur in pockets high in wet rain forest, situations to
which access by such cumbersome fruits seems unlikely. Dried fruits,
however, have a relatively low specific gravity and might be transported
short distances by violent winds. To explain the development of such
large fruits in Hawaiian Pritchardias, Beccari and Rock invoke the concept
of precinctiveness, which seems quite acceptable. Regarding the possible
mode of introduction of Pritchardia to the Hawaiian Islands, one could
say that the relatively small size of fruits in the Fijian P. thurstonit would
suggest edibility and therefore transport by frugivorous birds. However,
there is also a distinct possibility of oceanic drift. My tests of fruits of
P. martii (summit of Castle Trail, Oahu) showed that fresh fruits sink
immediately when placed in water. However, these same fruits, if allowed
to dry for several days and then placed in water, are capable of floating
indefinitely (in my tests, 2 weeks, at which time the experiment was
terminated). The occurrence of Pritchardia on such a low and dry island
as Nihoa, in the Hawaiian Leeward chain, as well as on low small islands
in southeastern Polynesia, suggests the possibility of seawater dispersal, at
least on rare occasions. The Nihoa species, P. remota (also once apparently
on Laysan Island), is relatively small fruited (18-20 x 17-19 mm). The
large-fruited species from the major Hawaiian islands are all characteristic

482

Loss of Dispersibility in Island Plants

of rain forest. Thus, a correlation between fruit size and ecology can be
said to exist in Pritchardia.

Liliaceae. Exact systematic affinities are difficult to establish for the
Hawaiian Dracaenas, of which two species, Pleomele aurea and P. fernaldii,
have been recognized. Pleomele, an endemic genus, is a segregate of
Dracaena, and some authors prefer to retain the Hawaiian species under
Dracaena. With regard to species of Dracaena elsewhere in the world,
in most species the fruits are 5-9 mm in diameter and contain several
seeds, each 3-5 mm long. Such fruits might have been ancestral to the
Hawaiian Dracaenas. Such fruits would be edible by birds; most Dracaena
fruits have a reddish fleshy exocarp. Fruits of Dracaenas are not known to
float. Pleomele fernaldii has fruits 10-12 mm long, 5-10 mm in diameter
if one carpel develops, 10-17 mm in diameter if two or three carpels
develop. Seeds are about 4-8 mm in diameter. In P. aurea, fruit and seed
size vary somewhat, but fruits are usually 14-25 mm in diameter, with
seeds 9-18 mm in diameter when fresh. Pleomele aurea is a tree that
ranges from dry lowland lava areas to well into higher portions of dry
(Acacia-Metrosideros) forest. Relatively large fruits can be seen on trees
in the latter region. Pleomele exhibits probable gigantism of fruits and
seeds.

Geography of Loss of Dispersibility

The Hawaiian Islands have predominated in the examples cited, for
reasons outlined in the introductory section of this chapter: all of the
probable reasons for loss of dispersibility are fulfilled in Hawaiian condi-
tions. A number of examples from the Juan Fernandez Islands (inde-
pendent instances in Asteraceae) and Samoa (Diospyros, Elaeocarpus,
Alectryon) suggest that these islands, with wet forest and moderate area,
may show loss of dispersibility chiefly because of ecological shift into the
wet forest. There is a scattering of examples throughout southeastern
Polynesia, all of which could be said to involve ecological shift into wetter
areas from dry coastal forest. A few examples can be cited from Fiji, several
prominent instances occur on New Caledonia, and a very few on New
Zealand. These islands are more nearly harmonic than the Hawaiian

483

Loss of Dispersibility in Island Plants

Islands, so there has been less opportunity for ecological shift into forest
areas, which are more nearly “saturated.”

Again, attention should be called to the role of loss of dispersibility
on continental areas. Although large-seeded trees of wet forests, especially
in the tropics, represent this phenomenon, one can find it elsewhere as
well. In particular, I would like to mention the work of Wilson (1970),
who showed some examples of degrees and types of dispersibility, as well
as drastic loss of dispersibility (Abronia alpina) in Nyctaginaceae of western
North America.

References

Beccari, O., and J. F. Rock. 1921. A monographic study of the genus Pritchardia.
Mem. Bishop Mus. 7(1):1-77.

Brown, E. D. W., and F. B. H. Brown. 1931. Flora of southeastern Polynesia. II.
Pteridophytes. Bishop Mus. Bull. 89:1—123.

Brown, F. B. H. 1935. Flora of southeastern Polynesia. III. Dicotyledons. Bishop
Mus. Bull. 130:1—386.

Carlquist, S. 1957. The genus Fitchia (Compositae). Univ. Calif. Publ. Bot.
29:1—-144.

Carlquist, S. 1959. Studies on Madiinae: Anatomy, cytology, and evolutionary rela-
tionships. Aliso 4:171-—236.

Carlquist, S. 1962. Trematolobelia: Seed dispersal; anatomy of fruits and seeds.
Pacific Sci. 16:126—34.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1966a. The biota of long-distance dispersal. II. Loss of dispersibility in
Pacific Compositae. Evolution 20:433-55.

Carlquist, S. 1966b. The biota of long-distance dispersal. III. Loss of dispersibility
in the Hawauian flora. Brittonia 18:310-35.

Carlquist, S., and M. L. Grant. 1963. Studies in Fitchia (Compositae): Novelties
from the Society Islands; anatomical studies. Pacific Sci. 17:282-98.

Chnistophersen, E. 1935. Flowering plants of Samoa. Bishop Mus. Bull. 128:1—221.

Darlington, P. J., Jr. 1943. Carabidae of mountains and islands: Data on the evolu-
tion of isolated faunas, and on atrophy of wings. Ecol. Monogr. 13:37-61.

Fosberg, F. R. 1948. Derivation of the flora of the Hawaiian Islands. In Insects of
Hawan, vol. 1: Introduction, E. C. Zimmerman, pp. 107-19. Honolulu, Uni-
versity of Hawaii Press.

Guppy, H. B. 1906. Observations of a naturalist in the Pacific 1896-1899, vol. 2:
Plant dispersal. London, Macmillan & Co.

Harling, G. 1962. On some Compositae endemic to the Galapagos Islands. Acta
Horti Bergiani 20:63-120.

Hillebrand, W. 1888. Flora of the Hawaiian Islands. London, Williams & Norgate.

484

Loss of Dispersibility in Island Plants

Hoffmann, O. 1890-1893. Compositae. In Die natiirlichen Pflanzenfamilien, ed.
A. Engler and K. Prantl, 4(5):87-402. Leipzig, Verlag Wilhelm Engelmann.

Howell, J. T. 1941. The genus Scalesia. Proc. Calif. Acad. Sci., 4th ser. 22:221-71.

Lloyd, R. M., and E. J. Klekowski, Jr. 1970. Spore germination and viability in
Pteridophyta: Evolutionary significance of chlorophyllous spores. Biotropica
2(2):129-37.

Papy, H. R. 1954-1955. Tahiti et les iles voisines. La végétation des iles de la
Société et de Makatea. Trav. Lab. Forest. Toulouse 5(2:1):1-386.

Perkins, R. C. L. 1913. Fauna Hawaiiensis: Introduction. Cambridge, Cambridge
University Press.

Ridley, H. N. 1930. The dispersal of plants throughout the world. London, L.
Reeve & Co.

Rock, J. F. 1913. The indigenous trees of the Hawaiian Islands. Honolulu, privately
published.

Sauer, J. 1964. Revision of Canavalia. Brittonia 16:106-81.

Schumann, K. 1895. Apocynaceae. In Die natiirlichen Pflanzenfamilien, ed. A.
Engler and K. Prantl, +(2):109-306. Leipzig, Verlag Wilhelm Engelmann.

Selling, O. 1944. A new species of Schizaea from Melanesia and some connected
problems. Svensk Bot. Tidskr. 38:207-21.

Selling, O. 1946. Studies in Hawaiian pollen statistics. Part I. The spores of the
Hawaiian pteridophytes. Bishop Mus. Spec. Publ. 37:1—87.

Sherff, E. E. 1935a. Revision of Haplostachys, Phyllostegia and Stenogyne. Bishop
Mus. Bull. 136:1-101.

Sherff, E. E. 1935b. Revision of Tetramolopium, Lipochaeta, Dubautia and Rail-
liardia. Bishop Mus. Bull. 135:1-136.

Sherff, E. E. 1937a. Some Compositae of southeastern Polynesia (Bidens, Coreopsis,
Cosmos, and Oparanthus). Occas. Papers Bishop Mus. 12(19):1-19.

Sherff, E. E. 1937b. The genus Bidens. Part I. Field Mus. Natur. Hist. Publ., Bot.
ser. 16:1—346.

Sherff, E. E. 1937c. Revision of the Hawaiian species of Euphorbia L. Ann. Missouri
Bot. Gard. 25:1-75.

Sherff, E. E. 1942. Revision of the Hawaiian members of the genus Pittosporum
Banks. Field Mus. Nat. Hist. Publ., Bot. ser. 22:467—566.

Sherff, E. E. 1944. Some additions to our knowledge of the flora of the Hawaiian
Islands. Amer. Jour. Bot. 31:151-61.

Sherff, E. E. 1955. Revision of the Hawaiian members of the genus Tetraplasandra
A. Gray. Fieldiana, Bot. 29:47—142.

Skottsberg, C. 1922. The phanerogams of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:95—240.

Skottsberg, C. 1937. Die Flora der Desventuradas-Inseln. Acta Horti Gotoburg.
5(6):1—-88.

Skottsberg, C. 1938. On Mr. C. Bock’s collection of plants from Masatierra (Juan
Fernandez), with remarks on the flowers of Centaurodendron. Acta Horti Goto-
burg. 12:361-73.

485

Loss of Dispersibility in Island Plants

Skottsberg, C. 1947. Eine kleine Pflanzensammlung von San Ambrosio (Islas Des-
venturadas, Chile). Acta Horti Gotoburg. 17:49-57.

Skottsberg, C. 1951. A supplement to the pteridophytes and phanerogams of Juan
Fernandez and Easter I. Natur. Hist. Juan Fernandez Easter I. 2:763—92.

Skottsberg, C. 1953. The vegetation of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:793-960.

Skottsberg, C. 1954. A geographical sketch of the Juan Fernandez Islands. Natur.
Hist. Juan Fernandez Easter I. 1:89-192.

Skottsberg, C. 1958. Uber Yunquea tenzii Skottsb. Ber. Deutsch. Bot. Ges. 71:45—52.

Skottsberg, C. 1963. Zur Naturgeschichte der Insel San Ambrosio (Islas Desven-
turadas, Chile). 2. Bliitenpflanzen. Ark. Bot., ser. 2, 4:465-88.

Sohmer, S. H. 1972. Evolutionary trends in the genus Charpentiera (Amarantha-
ceae). Brittonia 24:283-312.

Stebbins, G. L., Jr. 1953. A new classification of the tribe Cichorieae, family Com-
positae. Madrofio 12:33-64.

Stephens, S. G. 1958. Salt water tolerance of seeds of Gossypium species as a pos-
sible factor in seed dispersal. Amer. Natur. 92:83—92.

Stephens, S. G. 1963. Polynesian cottons. Ann. Missouri Bot. Gard. 50:1—22.

Stephens, S. G. 1964. Native Hawaiian cotton (Gossypium tomentosum Nutt).
Pacific Sci. 18:385-98.

Stephens, S. G. 1966. The potentiality for long range oceanic dispersal of cotton
seeds. Amer. Natur. 100:199-210.

Stewart, A. 1911. A botanical survey of the Galapagos Islands. Proc. Calif. Acad.
Sci., 4th ser., 1:7—288.

Sykes, W. R., and E. J. Godley. 1968. Transoceanic dispersal in Sophora and other
genera. Nature 218:495-96.

Wilson, E. O. 1959. Adaptive shift and dispersal in a tropical ant fauna. Evolution
13:122-44.

Wilson, R. C. 1970. Anthocarp dispersal in Abronia and its ecological implications.
Ph.D. thesis, Claremont Graduate School, Claremont, Calif.

Zimmerman, E. C. 1948. Insects of Hawait. vol. 1: Introduction. Honolulu, Univer-
sity of Hawaii Press.

486

LOSS OF DISPERSIBILITY
IN ISLAND ANIMALS

The most obvious loss of dispersibility in insular animals is the loss of
flying ability in insects and birds whose ancestors were capable of flight
when they immigrated to islands. Before proceeding with that conspicuous
phenomenon, I would like to discuss some more subtle ways in which
dispersibility can be lost in insular animals.

As an example of a “nonflying” animal that is nonetheless definitely
capable of aerial dispersal, one may cite a remarkable range of dispersal
abilities among spiders. In the genus Pardosa, Richter (1970) has rated
populations of the species according to the percentages of individuals
which, on the basis of their behavioral characteristics, are ‘‘potential
aeronauts.” These percentages range from a low of 0.05 percent in P.
lugubris to a high of 65.00 percent in P. palustris. Richter finds a close
correlation between low dispersal ability and preference for stable habitats.
However, when he rated habitats as to their abundance or rarity, he found
high dispersibility correlated with “rare” habitats. These findings are
eminently acceptable on the basis that species of far-flung habitats of
limited extent—like beach plants among angiosperms—have high dispersal
capacity, and that these same habitats also tend to be unstable. Thereby,
a selective advantage for high dispersibility is characteristic for species
inhabiting unstable, dispersed habitats of limited extent. Such species
would make good immigrants to islands—indeed they can be regarded as
living on a series of ecological islands within continents. If such a species
should migrate to an oceanic island and adapt to stable conditions (as are
characteristic of most islands in prehuman times), one would expect a

487

Loss of Dispersibility in Island Animals

loss of dispersibility. Available habitat is then “abundant” in a sense—the
new island population is a solitary population, not one of many populations
among which interchange can take place. This principle applies to high
islands—atoll plants and animals remain vagile and form populations
among which interchange occurs.

In carabid beetles, both Lindroth (1945-1949) and Boer (1970) believe
that “species living in unstable habitats invest extensively in dispersal”
(quoted from Richter, 1970), a concept in effect advanced somewhat earlier
by Southwood (1962). The inverse seems obvious: the more stable the
habitat, the less the pressure for dispersibility. Darlington (1943) stated
this, in essence, when he said, “The principal function of flight is found
to be to maintain sparse, unstable populations in large, unstable areas. In
small, stable areas, where populations are dense and stable, flight presum-
ably loses this function, permitting existence of flightless . . . Carabidae,
especially geophiles.”

These considerations also apply to various other animal groups. On
the Hawaiian Islands, Perkins (1913) noted the remarkable tendency of
Achatinella (Gastropoda) colonies to reside in a single tree over a period
of months. Clarke and Murray (1969), noting that the Partula species of
Moorea tend to remain isolated (although hybridization does occur), cited
geographical barriers within the island but believed that other factors must
be responsible for maintaining integrity of the species of these land shells.
If Partula resembles Achatinella in its precinctiveness, lack of mobility may
well provide the mechanism that could explain maintenance of species.
Indeed, in Achatinella the numerous species, each restricted to a very few
valleys, must have been produced and maintained by both geographical
isolation and precinctiveness.

Dispersibility of plants and animals is subject to evolutionary modi-
fication. Once a population (other than that of a repetitively introduced
beach species or the like) has been established on an oceanic island, the
factors that favor precinctiveness are many, the factors that favor increased
dispersibility are few and exceptional. Different groups and different islands
will not provide identical pressures for precinctiveness. In extremely windy
and cold subantarctic islands, precinctiveness will be favored for reasons
different from those operative in deep forest of tropical islands. As
Darlington stated (1943), “Whether or not flight is useful, and whether

488

Loss of Dispersibility in Island Animals

or not wings atrophy, depends on a balance of factors very much more
complex than Darwin guessed.” I will attempt to assess which factors may
be favored in which situations. Nevertheless, it appears to me that
precinctiveness does operate in insular plants and animals and that we
must consider flightlessness only one manifestation of it. It is entirely
possible for an insect species capable of flight to be highly precinctive also,
but there are only a few observations at present that note this (see,
however, some examples below from subantarctic islands).

In the discussion that follows, continental areas are considered
peripherally, and flightlessness is also briefly considered in chapter 14 in
a discussion of the equatorial alpine biota. As Darlington (1943) com-
pellingly showed for carabid beetles, flightlessness can be definitely
characteristic of certain continental areas. Likewise, loss of dispersibility
in plants is far from an exclusively insular phenomenon, as suggested in
the preceding chapter. The majority of instances, in fact, occur on
continents.

Flightless Birds

In my account of flightless birds on islands (1965, pp. 224-41), a listing
of flightless birds and references to them (pp. 416-18) are given, together
with illustrations and brief descriptions. Because most of these (other
than the large continental ratites and the penguins) are endangered or
extinct, the account of Greenway (1958), the I.U.C.N. Red Data Book
(vol. 2, Aves), and the literature they cite are the most complete sources
of material. Attention is called to De Beer’s (1956) account on the
evolution of the ratites. To my 1965 listing I can add that the Ascension
Island rail (listed as “Crecopsis sp.” in the I.U.C.N. Red Data Book) is
now known from complete skeletal material (Olson, 1971), and can be
added to the list of now-vanished insular Rallidae of the world. My
mention of the flightless duck or teal might be amended to indicate that
two subspecies are now recognized, Anas aucklandica aucklandica on the
Auckland Islands and A. a. nesiotis on Campbell Island. Personal observa-
tions during my 1966 visit to Laysan Island show that the Laysan teal can,
in fact, be frightened into flight rather easily, but the birds merely fly to

489

Loss of Dispersibility in Island Animals

another portion of the lagoon. Geographically, they can still be said to be
quite sedentary.

In reviewing my (1965) account of flightless birds, I find that several
features become apparent. First, no truly arboreal birds or bird groups are
in the list, with the exception of the flightless parrot (Strigops habroptilus)
and the flightless wren (Xenicus lyallii), both of New Zealand—and these
flightless representatives are, of course, terrestrial in their habits. The
relatively abundant time available in New Zealand, as well as the absence
of mammals, is undoubtedly a factor in the evolution of so many flightless
birds there. The Mascarene dodos and solitaires are of obscure ancestry,
but may well stem from nonarboreal birds.

To find that flightless birds are derived from groups that are ground-
feeders is no surprise. We must make an exception, at first glance, for
flightless waterbirds: the penguins, the giant auk (Pinguinus impennis),
the Galapagos flightless cormorant (Phalacrorax harrisii), and the flightless
grebe of Lake Titicaca, Rollandia micropterum (= Centropelma microp-
terum). These are all analogous to ground-feeding birds, however, in that
swimming (together with underwater diving) is all that is required, not
flight. With these exceptions, flightless birds are descended from terrestrial-
feeding bird groups, and these habits are merely maintained and heightened
on islands. Another feature common to the flightless insular birds (other
than the waterbirds mentioned) is their virtually omnivorous feeding habit.
This is particularly true of rails. In view of the limited size of the islands
on which many of the rails exist or once existed, broad food tolerances are
necessary for survival of viable populations. Another feature Rallidae
exemplifies is the tendency to fly not so much in long migratory patterns,
but merely from one suitable feeding area to another. On islands, this
would predispose them not to use flight as a tactic to any appreciable
degree. During feeding per se in such terrestrial birds as rails, flight does
not play a role, as it would for a frugivorous bird or nectarivorous bird.
Flight does occur in Rallidae as an evasive mechanism where predation
occurs.

The preadaptation of Rallidae to dispersal to islands, survival on
islands, and evolution into flightlessness there becomes obvious from the
preceding. Because they are somewhat migratory in their feeding habits
and not territory-bound as passerine birds typically are, Rallidae are likely

490

Loss of Dispersibility in Island Animals

to occur as stragglers to islands, more so, one would expect, than passerine
birds. On islands, they will survive because of their broad food tolerances.
Their feeding at ground level or in shallow water continues. Flight to new
feeding grounds is impossible because of oceanic distance. If an island can
support a rail population at all (and evidently small islands can), the food
supply can be exploited without resort to flight. The only further requisite
for evolution into flightlessness is then the absence of predators. Even
flight for evasive purposes may not be necessary. ‘The flightless ‘Tasmanian
native hen (Tribonyx mortieri, of Rallidae) can run at speeds up to
30 mph, faster than can the Rallidae of Tasmania that are still capable of
flight (Sharland, 1958). One cannot stress too strongly that of all natural
methods of locomotion, flight is the most highly expensive of energy;
therefore, in any condition under which flight is not a necessity for a
volant organism, there will be a positive selective pressure for flightlessness.

The continental ratites (emus, ostriches) substitute rapidity of running
for flight as an evasive mechanism. They are also able to use feet, legs,
and beak defensively; the cassowary, in addition, can use its horned crest.
Penguins, subject to predation by porpoises, use “underwater flight” as a
means of escape. Survival of flightless birds has either depended on
lack of predation or, where predation occurs, has involved compensatory
evolution of evasive mechanisms other than flight. Although one tends to
think of flightlessness as a retrogressive tendency, it can also be considered
a new and even efficient adaptive mode that is possible under certain
circumstances. One can cite exploitation of habitats in the fashion of
large herbivores (dodos, moas, ratites) as features of efficient ground
locomotion made possible by “release” from wing formation and main-
tenance.

Flightless Insects

Darlington (1943) views flightlessness in carabid beetles as an inevitable
trend wherever flight is not of high selective value in order to maintain
large, sparse, unstable populations over large, unstable areas. ‘he smallness
of occupiable area and the stability of the environment are, in his
considerations, of prime importance. Going farther, Darlington cites an

491

Loss of Dispersibility in Island Animals

instance in weevils where flightless individuals may, in fact, show better
viability than winged individuals within a single species (Jackson, 1928),
and tends to regard flightless insects as having “inherent superiority” due
to “simplicity and vitality.” One might wish for some physiological
measures to demonstrate actual types of advantage in reduced-wing insects,
other than the obvious one that an economy is involved if wings and
even hapteres are lost. Spencer (1932) did show that vestigial-winged
Drosophila mutants survived better under conditions of food and water
deprivation than normal-winged populations. However, Lindroth (1945-
1949), with the aid of some experimental work, contradicts the idea that
flightlessness improves viability. Brinck (1948) would seem to agree with
Darlington, for he states that “on oceanic islands, where the conditions
occasionally become very unfavorable, a vestigial | flightless] type might
even oust the normal.” However, Brinck’s discussion obviously favors the
selective effect of wind pressure as operative in producing a high proportion
of flightless insects on Tristan da Cunha. Darlington does not believe wind
pressure is a factor in evolution of flightless insects, but the subantarctic
insects may be an exception (see below). Tristan da Cunha has some of
the aspects of a subantarctic island.

Perkins (1913) subscribed to the “disuse” theory to explain flightless-
ness in Hawauan insects. Interestingly, a high proportion of arboreal
Coleoptera are flightless in the Hawaiian Islands. Among instances Darling-
ton cites where flight is useful in carabids are hydrophiles or arboreal
species; species in areas likely to be flooded; species on low tropical islands;
and species on dry lowlands of tropical islands. Small, low, tropical islands
(e.g., coral platforms like the Cayman Islands) are, as Darlington notes,
often recent, with no endemism. However, applying the considerations
of MacArthur and Wilson (1963), one can say that extinction rate is high
on a small, low island, so that influx of new stocks by a constant stream
of immigrants is required to maintain a given number of species. In the
case of carabids, fully winged, widespread, tropical species would presum-
ably be the immigrants to such islands. If repetitive and continual
immigration to a small low island occurs, even without extinction, new
immigrants with wings would doubtless swamp out any winglessness,
assuming (as seems likely) that brachyptery and aptery are the results of
recessive genes.

492

Loss of Dispersibility in Island Animals

Darlington (1943) would explain lack of flightless carabids in tropical
lowlands (insular and otherwise) by the fact that these dry areas, poorer
in food resources, fall under the category of the “large and unstable areas”
he designates as correlated with selective advantage for flying forms.

Darlington discounts altogether the possibility that wind pressure
may exercise a selective force leading to flightlessness. His data on montane
carabids demonstrate, in his opinion, no appreciable difference in flying
versus flightless in exposed or protected situations, respectively. Wind
pressure was claimed by Darwin to make flightlessness advantageous in
insular insects, in that excessive numbers would be blown out to sea (on
mountains, wind pressure would presumably blow insects to ecologically
unsuitable locations according to the Darwinian reasoning, although
Darwin did not comment on flightlessness in montane situations). An
alternative explanation would be that where wind pressure is excessive,
wings are disadvantageous because wind would tend to hinder flight during
feeding. Darlington (1943) analyzed data for carabids of the Presidential
Range, New Hampshire. His results show that when hydrophiles, non-
resident, and arboreal species are excluded, 70 percent of the forest
geophiles and 64 percent of the exposed-area geophiles above 3000-ft
elevation are flightless. This he attributes to the geophilous habit and
the concentrated area of suitable habitat—the latter related to the concept
of precinctiveness that I have cited earlier. Flightless Carabidae for New
Hampshire as a whole amount to 13 percent, so there is, without doubt,
a much higher percentage in montane carabids in Darlington’s data.
Greenslade (1968) finds similar data in Argyll, Scotland, although the
percentages are much less striking than in Darlington’s data—and the
number of species of carabids involved is relatively small also.

Whether wind pressure is a factor in winglessness has not been
resolved, to judge from the literature. Downes (1962) noted reduction of
mating flight in insects of far northern Canada. The experiments of
L’Heritier, Neefs, and Teissier (1937) with brachypterous and normal-
winged Drosophila races in culture dishes open to wind showed that
normal-winged flies tended to be blown away, while the flightless flies had
a low rate of attrition. This experiment, however, may well not be definitive
in demonstrating the effect of wind.

498

Loss of Dispersibility in Island Animals

SUBANTARCTIC ISLAND INSECTS

Either insects of subantarctic islands add data to illuminate the
question of wind pressure in relation to flightlessness, or else they form a
special case. Gressitt (1970) says that Darlington’s hypotheses “do not
apply” to the subantarctic islands, and he also finds the models of
MacArthur and Wilson (1963) inapplicable. Cold makes flying much more
difficult for insects, and the subantarctic islands are exceptionally cold—
cool even during the summer. This, together with the added “burden”
of wind pressure, may prove a definitive selective pressure for flightlessness
and eventually aptery (see the last portion of chapter 14 for a discussion
of these points).

In any case, subantarctic and Antarctic insects have been the subject
of intense study during the past decade. For references to biological work
of all kinds on the “true” subantarctic islands, Gressitt’s (1970) summary
is invaluable. Somewhat more northern islands of the far south are
summarized by Gressitt (1962, 1964), and Gressitt, Leech, and Wise (1963)
have provided a summary for Antarctic entomology. Attention should be
called to monographic papers, not cited here, which accompany Gressitt’s
papers (1962, 1964, 1970).

The comments of Gressitt (1962, 1964) and the rich base of data on
which he operates are of especial significance to Antarctic entomology.
In particular, the effect of climate—presumably strong winds and cold—is
examined. Gressitt’s data (1964, 1970) show the following degrees of
flightlessness for insect faunas of subantarctic islands:

Total Total
number of Percentage number of Percentage
species flightless species flightless
Campbell I. 183 40 Kerguélen 26 85
Crozet Is. 36 89 South Georgia 16 WD
Marion I. 18 89 Heard I. 9 100
Macquarie I. 20 aD)

On Antarctica itself, only a single insect capable of flight is known
(out of 20 land arthropods, mostly insects): a chironomid midge, Parochlus
steineni (Gressitt, 1970). Gressitt’s data are subdivided further than my
table shows; he indicates various degrees of wing reduction from moderate
to complete. The strong degree of flightlessness is evident. In addition,

494

Loss of Dispersibility in Island Animals

the degree of nonflying behavior in insects capable of flight is manifest in
his observations. He cites, on Campbell I. (Gressitt, 1964), many species
known to fly but never taken in trapping experiments. These include
pyralid moths, geometrid moths, and an ichneumon wasp. The smaller
wasp Campoplex disjunctus was, however, taken in nets. Gressitt states,
“These species which were not trapped are weaker fliers than some of the
types commonly caught in the nets, like calliphorid flies, muscids, and
others, yet are stronger fliers than some which were taken abundantly in
the nets, like the aphids, psychodids, sciarids, and others.” According to
Gressitt and Weber (1959), wingless or partially winged subantarctic
insects as a whole may be termed “active,” whereas winged species are
“sluggish.” The highly sedentary aspect of the far-southern insect fauna
is strongly evident.

Lepidoptera are particularly interesting in respect to flightlessness.
Lepidoptera, with few exceptions, have evolved flightlessness only on the
subantarctic islands (Enderlein, 1909; Viette, 1948, 1952a, 1952b, 1954,
1959). This, together with the tendency for strong-flying insects not to fly
during periods of high wind, suggests that extreme wind pressure may
indeed be a factor in the evolution of the subantarctic fauna. The most
important feature may not even be that an excessive number would be
carried out to sea (although lack of flightless forms in nets would seem to
offer support for this), but that there might be interference with locomotion
by wind pressure on wings or that there would be loss of heat through
wings during conditions of low temperatures.

Three Campbell Island flightless moths are shown here: Tinearupa
sorenseni of the Hyponomeutidae (fig. 12.1); Campbellana attenuata,
also of Hyponomeutidae (fig. 12.2); and Exsilaracha graminea of the
Pyralidae (fig. 12.3). Exsilaracha is nearly twice as large as the others, and
has a pronounced jumping habit that substitutes for flying (Munroe, 1964).
Exsilaracha lives exclusively in grass tussocks, and its hopping habits
resemble those of grasshoppers, according to Salmon and Bradley (1956).
Tinearupa (fig. 12.1) is found on coastal rocks, where its coloring and
markings form a perfect camouflage, according to these authors. Sorensen
(quoted in Salmon and Bradley [1956]) finds that Campbellana is a tussock
insect, also with grasshopperlike habits. Strong development of legs is

495

Loss of Dispersibility in Island Animals

fy "
WA)

QS eS Ns ee MP RB OE
Fic. 12.1. Tinearupa sorenseni, a flightless Fic. 12.2. Campbellana attenuata, a flightless
moth from Campbell I. moth from Campbell I.

SSSR, “ay, Rey
: Se . =

SS

= ; [o) 2mm.

lay
y 4

ee %e

FIG. 12.3. Exsilaracha graminea, a flightless Fic. 12.4. Pringleophaga kerguelensis, a
moth from Campbell I. flightless moth from Kerguélen I. It is shown
on a leaf of Pringlea antiscorbutica, the “Ker-

guélen cabbage.”

evident in both Exsilaracha (fig. 12.3) and Campbellana (fig. 12.2). The
abundance of flightless moths on Campbell Island may be related to the
closeness to New Zealand, which could have served as an efficient source
area for entry of moths into these southerly islands. Relative absence of
moths from remote Macquarie Island is cited by Gressitt (1970) as a

496

Loss of Dispersibility in Island Animals

reason why the proportion of flightless insects is lower there than one
might expect on the basis of the extreme climate.

A curious feature of the Campbell Island moths is their tendency to
sham death or seek refuge in tussocks or crevices despite the seeming lack
of predatory birds—the pipit is the only insectivorous bird on Campbell
Island (Gressitt, 1964).

One notable flightless moth on a remote subantarctic island is
Pringleophaga kerguelensis on Kerguélen (fig. 12.4). As its generic name
suggests, this insect’s host plant is the curious Kerguelen “cabbage,”
Pringlea antiscorbutica. Comparison of many flightless moths of sub-
antarctic islands shows remarkable similarity in that the wings are reduced
by narrowing, although shortening is obviously involved as well. Extreme
wing reduction is shown in the moth Dimorphinoctua cunhaensis of
Tristan da Cunha (Viette, 1952a) and Brachyapteragrotis patricet of
Amsterdam Island (for illustrations, see Carlquist, 1965).

A tendency noted by various authors is that in subantarctic moths
the males have somewhat less wing reduction than females. For compari-
sons, see Viette (1948), Salmon and Bradley (1956) and Carlquist (1965,
p. 222). This is also true in the one Hawaiian moth with flightless
tendencies, Hodegia apatela. One is tempted to speculate that the males
are more “expendable” than the females. Wind pressure alone might be
a selective agent, but predation on males that are more conspicuous by
virtue of longer wings might also be evoked. Field investigation of this
dimorphism would, of course, be difficult yet very valuable.

In subantarctic Diptera, there is great variation in wing development,
even within a single species. On Macquarie Island, the entire endemic
species Schoenophilus pedestris (fig. 12.5) is flightless, yet some individuals
have longer wings than others; there is no evident correlation with habits
or habitats (Gressitt, 1964). Schoenophilus pedestris moves by hopping
and tends to occur in tussock plants, such as Colobanthus (Kohn, 1962).
Kohn compares Schoenophilus to a Hawaiian member of the same family,
Dolichopodidae, in its wing reduction. A dolichopodid of Campbell Island,
Acropsilus borboroides, has the ultimate in wing reduction: lack of wings
and even halteres (Oldroyd, 1955).

Apataenus watsoni (Coelopidae) is another dipteran from Macquarie
Island (fig. 12.6). It tends to be found under stones, particularly those

497

Fic. 12.5. Schoenophilus pedestris (Diptera: Dolichopodidae), from Macquarie I.

i
N}

<
FIG. 12.6. Apataenus watsoni (Diptera:
Coelopidae), from Macquarie I.

498

COT ob
Reap macr ye oo

te

[SRE Lee

Fc. 12.7. Belgica antarctica (Diptera: Chiro-
nomidae), a flightless midge from the Antarc-
tic mainland.

Loss of Dispersibility in Island Animals

of bird rookeries (Hardy, 1962). Its wings may be up to half as long as
the thorax. Extreme wing reduction is found in a midge from the Antarctic
continent, Belgica antarctica of the Chironomidae (fig. 12.7). It occurs in
penguin rookeries and can even be found on the penguins themselves
(Gressitt, Leech, and Wise, 1963).

With respect to flightless Diptera of subantarctic islands, Seguy
(1940) claimed several structural trends: robust pharynx, long malpighian
tubules, reduced gastric coeca, short hind intestine, abundant adipose
tissue; tough pupaia; adults with heavily sclerotized mouthparts, atrophied
notopleural callus, small thorax, and reduced wing and leg muscles. How-
ever, Gressitt (1964) found that on Campbell Island an increase in leg
size and musculature occurred with reduction in wing size. This is certainly
obvious in his figures of the dipteran Baeopterus robustus and Icaridion
nasutum (both of Coelopidae). The latter has no wings or halteres.

Kleidotoma subantarctica of the Encoilidae (fig. 12.8) is the first
record of a cynipid wasp from any subantarctic island (Yoshimoto, 1964).
It shows an interesting series in wing reduction, and one individual with
fully developed wings has even been collected. Kleidotoma subantarctica
is an endemic of Campbell Island. It lives on algae-covered rocks, to judge
from collection data, but its feeding habits are as yet unknown (Yoshimoto,
1964).

Hymenoptera are not abundant in flightless species in areas other
than the subantarctic islands, but they clearly occur with frequency on the
far-southern islands. Various hymenopteran families are represented. Anti-
podencyrtus procellosus of Campbell Island (fig. 12.9) is a representative
of Encyrtidae. It is evidently a parasite on a soft-bodied coccid of tussock
grass (Kerrich, 1964). Occurrence of a parasite under the difficult condi-
tions for establishment on subantarctic islands is interesting. Another
curious feature of Antipodencyrtus is the presence of very long antenna
branches.

Aucklandella flavomaculata (fig. 12.10) is an ichneumonid wasp dis-
covered during an early survey of the subantarctic islands of New Zealand
(Chilton, 1909). It occurs in Carnley Harbor of Auckland Island and may
have a habitat similar to that of Kleidotoma subantarctica. There is an-
other flightless ichneumonid on Campbell Island, Gelis campbellensis.
Other flightless subantarctic Hymenoptera include two diapriids, Antarc-

499

FIG.

FIG. 12.9. Antipodencyrtus procellosus (En- FIG.

12.8. Kleidotoma subantarctica (Eucolidae), a flightless hymenopteran from Campbell I.

2k

Aucklandella flayomaculata

cyrtidae), a flightless hymenopteran of Camp- (Ichneumonidae), a flightless hymenop-
bell I. teran from the Auckland Is.

500

Loss of Dispersibility in Island Animals

topria latigaster of Macquarie Island, and a subspecies, A. 1. campbellana,
on Campbell Island (Gressitt, 1964).

Judging from insect groups that have flightless representatives in in-
sular situations elsewhere in the world, one would not expect, a prion,
that subantarctic islands would bear numerous flightless Lepidoptera,
Hymenoptera, and Diptera, yet only a scattering of Coleoptera. Coleoptera
are probably underrepresented because larger-bodied insects travel poorly
over long distances and in cold wind currents—which also would account
for the absence of Orthoptera and Neuroptera, and the paucity of Homop-
tera other than widespread aphids on the subantarctic islands. Interest-
ingly, according to Beardsley (1964) the family Pseudococcidae of Homop-
tera is represented on subantarctic islands by several genera and species
in which males are flightless and completely apterous (female wingless-
ness would be expected in this family in any case).

An interesting feature of Coleoptera discussed by Brinck (1948) is the
tendency for the ratio of carnivorous to phytophagous species to vary with
habitat. Although the ratio is high in northern Europe, including some
insular areas (7.53 in the Faeroe Is., 6.08 in the Shetland Is., 4.5 in Green-
land, 3.05 in Norway), the proportion is lower farther south (1.17 on Cor-
sica, 1.33 on Madeira, 2.14 on the Cape Verde Is.). However, subantarctic
islands have an extraordinarily high number of phytophagous insects, so
that the ratios are 0.17 on Tristan da Cunha and 0.12 on Kerguélen. The
explanations for this are not at all clear, although Brinck (1948) offers a
number of possibilities.

MADEIRAN AND OTHER ATLANTIC ISLAND INSECTS

Darwin, upon reading Wollaston’s Insecta Maderensia (1854), was
struck by the high proportion of flightless forms. As we have seen earlier,
Darwin’s explanation has seemingly been supplanted by Darlington’s
(1943) detailed and compelling considerations. However, assuming that
precinctive populations and geophily are, as Darlington suggests, the most
important factors in evolution of flightlessness of Coleoptera, it is interest-
ing to review Wollaston’s data to see how these factors operate on Ma-
deira. Wollaston’s work is of particular interest because he studied these
species extensively in the field and dealt with their habits in great detail.
Although Wollaston’s monograph is outdated by subsequent faunistic

501

Loss of Dispersibility in Island Animals

Fic. 12.1]. Eurygnathus latreillei, male, a Fic. 12.12. Loricera wollastoni, a flightless
flightless carabid (Coleoptera) from Madeira carabid (Coleoptera) from Madeira (from
(from Wollaston, 1854). Wollaston, 1854).

work, it can still be admired for its care to ecological detail. Indeed, such
information would be perhaps impossible to secure today. For purposes of
precise reference to Wollaston’s volume, I am retaining his nomenclature.

Eurygnathus latreillei (fig. 12.11) is interesting to us for several rea-
sons. Not only is this carabid beetle apterous, but its elytra are connate—
or “soldered’”’ as some have said—definitely canceling the power of flight.
Its ancestors were, with little doubt, carabids capable of flight. In addition
to this striking expression of flightlessness, E. latreillei exhibits different
size tendencies in different localities. Although found beneath stones’ on
the small islands of Porto Santo and on the Desertas (both offshore islet
groups of Madeira), specimens on Deserta Grande are much larger than
those on Porto Santo, according to Wollaston. This is also true of another
carabid, Olisthopus maderensis, much larger on Deserta Grande than on
Madeira itself. Wollaston erects a new genus, Thalassophilus, for a flight-

502

Loss of Dispersibility in Island Animals

FIG. 12.13. Cossyphodes wollastoni (Colydia- Fic. 12.14. Cholovocera madaerae (Coleop-
dae), a flightless and eyeless coleopteran tera: Lathridiadae), a flightless beetle from
from Madeira (from Wollaston, 1854). Madeira (from Wollaston, 1854).

less reduced-wing carabid of gigantic size compared with its closest relative,
Aépus. Bembidium obtusum, a widespread species, attains a much greater
size on Madeira than in Europe. Comments on gigantism on islands, which
may be relevant to the preceding instances, will be found in chapter 15.

The carabid Loricera wollastoni (fig. 12.12) has vestigial wings but
normal elytra. It lives chiefly under logs and leaves. Wollaston suggests
that its peculiar antennae and long legs are virtually “supralimital’”’ char-
acteristics within the genus Loricera.

Another Madeiran carabid, Harpalus vividus, is wingless, with elytra
connate to various degrees. Lowland and high-elevation populations have
soldered elytra and are dull in color, whereas elytra are separate and body
colors are brighter in mid-elevation populations.

Other families of Coleoptera on Madeira show flightlessness inde-
pendently. Tarphius (Colydiadae) is an example, with reduced wings.
Wollaston notes that Tarphius species are strictly nocturnal. Interestingly,
Greenslade (1968) notes that nocturnalism and dark colors are much more
common in Coleoptera with reduced wings, whereas diurnalism and
brighter colors are commoner in normal-winged species. Another genus
of Colydiadae, Cossyphodes (fig. 12.13), a Madeiran endemic, is interest-
ing not merely in its winglessness and exceptional shape, but in its loss of

503

Loss of Dispersibility in Island Animals

eyes. It feeds in nests of an ant, Oecopthora pusilla. Although the legs of
Cossyphodes wollastoni are apparently small in relation to body size and
it is blind, it runs with exceptional velocity. Another endemic colydiad
genus, Pleiosoma, is also wingless and occurs “in the very center of moist
decaying wood.”

In the family Lathridiadae, Cholovocera maderae (fig. 12.14) is nota-
ble in its vestigial wings. Although not blind, its eyes are much reduced,
with many fewer facets than are characteristic for beetles. It is not surpris-
ing, therefore, that it, like Cossyphodes, is predatory in nests of ants. Re-
duced eyes of this type occur in an endemic Madeiran species of another
ant-predaceous genus, Metophthalmus, which has European species as
well.

In Ciidae, the Madeiran endemic species Cis wollastoni is exceptional
for its genus by virtue of its large and elongated body. Among flightless
curculionids, numerous Madeiran species could be cited, such as Atlantis
vespertinus (fig. 12.15). Wollaston created this endemic genus for flightless
species in which wings are absent and elytra firmly united. These features
are also shared by another endemic curculionid genus that Wollaston
named Echinosoma. Echinosoma porcellus (fig. 12.16) is exceptionally
setose and has an unusual “robust” rounded form, yet it has delicate an-
tennae. Atlantis lives beneath moss on trees and is strictly nocturnal.
Echinosoma occurs beneath stones and logs.

LAKE BAIKAL INSECTS

Lake Baikal exhibits many insular features, as do other freshwater
lakes (see chapter 9), so it is not surprising to find certain situations of
flightlessness there among a family of aquatic caddis flies, Limnophilidae.
In Lake Baikal, there are two endemic genera, Baicalina and Thamastes,
while the only other genus of the tribe Baicalinini, Radema, lives in the
Lena and other creeks that flow into Lake Baikal. Larvae of Baicalina and
Thamastes live on the rocky bottom of the littoral of Lake Baikal, Radema
larvae occur underwater in streams. As the water rises in spring and sum-
mer in Lake Baikal, adults formed from larvae occur at a distance from
the newly expanded shoreline, toward which they swim in droves. Wings
and legs of the three endemic caddis-fly genera are modified for swimming
(Martynov, 1929, 1935). The wings are not capable of flight but are used

504

Loss of Dispersibility in Island Animals

Fic. 12.15. Atlantis vespertinus (Coleoptera: 1c. 12.16. Echinosoma porcellus (Coleop-
Curculionidae), a flightless weevil from Ma-_ tera: Curculionidae), a flightless weevil from
deira (from Wollaston, 1854). Madeira (from Wollaston, 1854).

as “paddles.” Kozhovy (1963) attributes the wingless habit to the limited
habitat available. He states that fully winged insects with a typical caddis-
fly life cycle would be bound to be driven far from the shore and could
not have offspring. This is reminiscent of the Darwinian interpretation of
the effect of wind pressure on insular insects. Certainly, at least precinctive-
ness and a limited stable habitat are involved. Lake Baikal is unusual in
its low temperature, and larvae of the endemic limnophilid genera are
adapted to this and might not survive under other temperature regimes.

NEW ZEALAND INSECTS

Among the many interesting flightless insects of the mainland islands
of New Zealand, the wetas (Orthoptera) have received abundant comment
because of their flightlessness, hopping habits, and adaptive radiation in
wet recesses in caverns and forest situations (see chapter 7). Among hop-
ping Orthoptera, the wetas are also unusual in their capability to regen-

505

Loss of Dispersibility in Island Animals

erate appendages of near-normal size (Ramsay, 1964). Is this ability re-
lated to their total reliance on legs for locomotion, or does the relative
longevity of the wetas (two years or longer) play a role? This is one of the
many intriguing questions possibly related to loss of flight on insular areas.

References

Beardley, J. W. 1964. Insects of Campbell Island. Homoptera: Coccoidea. Pacific
Insects Monogr. 7:238-52.

Boer, P. J. den. 1970. On the significance of dispersal power for populations of cara-
bid-beetles (Coleoptera, Carabidae). Oecologia 4:1-28.

Brinck, P. 1948. Coleoptera of Tristan da Cunha. Results Norwegian Sci. Exped.
Tristan da Cunha 1937-1938 17:1-120.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Chilton, C., ed. 1909. The subantarctic islands of New Zealand. 2 vols. Wellington,
John Mackay, Government Printer.

Clarke, B., and J. Murray. 1969. Ecological genetics and speciation in land snails of
the genus Partula. Biol. Jour. Linnaean Soc. 1:31-42. (Also as Speciation in Tropi-
cal Environments, ed. R. H. Lowe-McConnell. London, Academic Press.)

Darlington, P. J., Jr. 1943. Carabidae of mountains and islands: Data on the evolu-
tion of isolated faunas, and on atrophy of wings. Ecol. Monogr. 13:37-61.

De Beer, G. R. 1956. The evolution of the ratites. Bull. Brit. Mus. (Natur. Hist.)
Zool. 4(2):1-63.

Downes, J. A. 1962. What is an Arctic insect? Can. Entomol. 94(2):143-62.

Enderlein, G. 1909. Die Insektfauna der Insel St. Paul. Die Insektfauna der Insel Neu
Amsterdam. Deutsch. Siidpolar-Exped., Zool. 2(4):481-92.

Greenslade, P. J. M. 1968. Habitat and altitude distribution of Carabidae (Coleop-
tera) in Argyll, Scotland. Trans. Roy. Entomol. Soc. London 120(2):39-54.

Greenway, J. C., Jr. 1958. Extinet and vanishing birds of the world. New York, Amer-
ican Committee for International Wild Life Protection.

Gressitt, J. L. 1962. Insects of Macquarie Island. Introduction. Pacific Insects
4:905-15.

Gressitt, J. L. 1964. Insects of Campbell Island. Summary. Pacific Insects Monogr.
7:531-600.

Gressitt, J. L. 1970. Subantarctic entomology and biogeography. Pacific Insects
Monogr. 23:295-374.

Gressitt, J. L., R. E. Leech, and K. A. J. Wise. 1963. Entomological investigations
in Antarctica. Pacific Insects 5:287—304.

Gressitt, J. L., and N. A. Weber. 1959. Bibliographic introduction to Antarctic-sub-
antarctic entomology. Pacific Insects 1:441—80.

Hardy, D. E. 1962. Insects of Macquarie Island. Diptera: Coelopidae. Pacific Insects
4:963-71.

506

Loss of Dispersibility in Island Animals

L’Heritier, P., Y. Neefs, and G. Teissier. 1937. Aptérisme des insectes et sélection
naturelle. Compt. Rend. Acad. Sci. Paris 204:907-9.

Jackson, D. J. 1928. The inheritance of long and short wings in the weevil, Sitonia
hispidula, with a discussion of wing reduction among beetles. Trans. Roy. Soc.
Edinburgh 55:665-735.

Kertich, G. J. 1964. Insects of Macquarie Island. Hymenoptera: Encyrtidae. Pacific
Insects Monogr. 7:504-6.

Kohn, M. A. 1962. Insects of Macquarie Island. Diptera: Dolichopodidae. Pacific
Insects 4:959-62.

Kozhoy, M. 1963. Lake Baikal and its life. The Hague, Dr. W. Junk, Publishers.

Lindroth, C. H. 1945, 1946, 1949. Die Fennoskandischen Carabidae, eine tiergeo-
graphische Studie. Medd. Goéteborgs Mus. Zool. Avd. 109:1-707, 110:1-277,
122:1-911.

MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular bio-
geography. Evolution 17:373-87.

Martynov, A. 1929. Ecological prerequisites for the zoogeography of freshwater ben-
thonic animals. Russk. Zool. Zh. 9(3):3-38.

Martynov, A. 1935. The caddis-flies of the Amur region. Tr. Zool. Inst. Akad. Nauk.
S.S.S.R. 2(2—3):205—395 (in Russian).

Munroe, E. 1964. Insects of Campbell Island. Lepidoptera: Pyralidae. Pacific Insects
Monogr. 7:260-71.

Oldroyd, H. 1955. A wingless dolichopodid (Diptera) from Campbell Island. Rec.
Dominion Mus. Wellington 2(4):243—46.

Olson, S. L. 1971. A flightless bird, lost and found. Smithsonian 2(3):63-65.

Perkins, R. C. L. 1913. Fauna Hawaiiensis: Introduction. Cambridge, Cambridge
University Press.

Ramsay, G. W. 1964. Regeneration of appendages in some New Zealand wetas
(Insecta: Orthoptera). Trans. Roy. Soc. New Zealand, Zool. 4(6):139-44.

Richter, C. J. J. 1970. Aerial dispersal in relation to habitat in eight wolf spider species
(Pardosa, Araneae, Lycosidae). Oecologia 5:200-214.

Salmon, J. T., and J. D. Bradley. 1956. Lepidoptera from the Cape Expedition and
Antipodes Islands. Ree. Dominion Mus. Wellington 3(1):61-81.

Seguy, E. 1940. Dipteres. Mém. Mus. Nat. Hist. Natur. Paris, ser. A., Zool. 14:203-
67.

Sharland, M. 1958. Tasmanian birds, 3d ed. Sydney, Angus and Robertson.

Southwood, T. R. E. 1962. Migration of terrestrial arthropods in relation to habitat.
Biol. Rey. 37:171-214.

Spencer, W. P. 1932. The vermillion mutant of Drosophila hydei breeding in nature.
Amer. Natur. 66:474-79.

Viette, P. 1948. Croisiére du Bougainville aux iles australes Francaises. 20. Lepidop-
tera. Mém. Mus. Nat. Hist. Natur. Paris, ser. A., Zool. 27:1—28.

Viette, P. 1952a. Lepidoptera. Results Norwegian Sci. Exped. Tristan da Cunha
1937-1938 23:1-19.

Viette, P. 1952b. Lepidoptera. Sci. Results Norwegian Antarctic Exped. 33:1—4.

507

Loss of Dispersibility in Island Animals

Viette, P. 1954. Une nouvelle espéce de Lépidoptere de Vile Campbell. Entomol.
Medd. 27:19-22.

Viette, P. 1959. Lépidopteres de Vile d’Amsterdam (récoltés de Patrice Paulian, 1955-
1956). Bull. Entomol. Soc. France 64:22-29.

Wollaston, T. V. 1854. Insecta Maderensia. London, John Van Voorst.

Yoshimoto, C. M. 1964. Insects of Campbell Island. Hymenoptera: Cynipoidea:
Eucolinae. Pacific Insects Monogr. 7:509-12.

508

APTEIR

REPRODUCTIVE BIOLOGY
ON ISLANDS

As yet we know little about modes of reproductive biology on islands. One
can cite the fact that field observations are required to establish data in
this general topic, and that prolonged stays on many insular areas are, un-
fortunately, difficult. In other instances, no such reasons exist, and one can
only note the absence of qualified investigators who are willing to study
these very important “natural history” aspects on insular areas where re-
search institutions are present. Characteristic modes of speciation, hybridi-
zation, floral biology, pollination relationships, shift in seasonal phenom-
ena of reproduction, and changes in fertility and longevity of individuals
are among the neglected phenomena. | take pleasure in citing results ob-
tained thus far, but concede that these studies are as yet inadequate in
number and scope, and invite other workers to enter this interesting field.

The data and commentary that follow are presented to suggest the
significance of such phenomena as the high degree of dioecism in insular
plants, polymorphism in insular species, and seemingly low rates of repro-
duction in island plants and animals. These suggestions have been organ-
ized so as to place them into the context of what may be called the “insular
syndrome” of interrelated evolutionary phenomena. Analysis of single
phenomena on islands has not always led to sound or thoroughgoing inter-
pretations. Floral morphology, breeding systems, ecological preferences,
growth forms, dispersal mechanisms, fertility, and other features of insular

NOTE: Portions of this chapter are revised segments of a paper (Carlquist, 1966). I
wish to express appreciation to Evolution for permission to use this material.

509

Reproductive Biology on Islands

groups are closely linked together. When considered conjunctively, these
features appear more comprehensible.

Reproductive Biology in Plants

The emphasis in this chapter is the presentation of data from the Hawaiian
Islands (modified from Carlquist, 1966) and citation of information on
other insular areas where data are available—particularly New Zealand and
the Galapagos Islands. he New Zealand flora is a waif flora in post-Meso-
zoic time, and probably earlier as well (Darlington, 1965). Thus, the New
Zealand flora may be considered almost entirely a waif flora, differing from
that of the Hawaiian Islands in its greater age and the climatic conditions
to which plants have adapted.

Long-distance dispersal to an island is a drastic event in the history of
a species. In virtually all groups that establish by means other than oceanic
drift, genetic contact with the parent population is irretrievably broken.
The subsequent evolution of such a stock may be regarded as involving
compensation for this broken contact. To the extent to which the species
has evolutionary plasticity, is endowed with favorable characteristics for
maintenance on a relatively small land area, and manages to maximize out-
crossing, it may be able to perpetuate a lineage of descendents into a rela-
tively long career. With the variability of mainland relatives no longer
accessible, the variability that remains or can be developed within an
island population must, for best success, be circulated throughout the en-
tire population. Dioecism, monoecism, and various other diclinous floral
conditions tend to ensure the occurrence of this process. Self-sterility is of
prime importance, and wind pollination can be regarded as a mechanism
broadening the degree of, and geographical extent of, outcrossing. To a
lesser extent, dichogamy (protandry and protogyny) and even showiness of
flowers may aid in increasing the level of outcrossing. These various phe-
nomena may be expected to reach a high pitch on oceanic islands. Such
conditions do not exist independently of other factors, however. As sug-
gested by Grant (1958), for example, poor seed dispersal would limit the
amount of recombination which would occur. The topographical diversity
within an island, the effectiveness of interisland isolation, and the size of

510

Reproductive Biology on Islands

land areas suitable for a particular species are other dimensions that would
influence the genetic diversity likely to be contained within a species.
The floral condition and the genetic nature of each immigrant species
will dictate which species are best suited for long-term success in an island
situation, and what changes will maximize this occupancy. Dioecious spe-
cies may seem disadvantaged by the requirement that two propagules
rather than one establish when a species colonizes. In the Hawaiian flora,
some arrivals seem clearly to have been dioecious prior to arrival, whereas
others very likely evolved the dioecious condition on the Islands. In addi-
tion to diclinous conditions, other pathways for maintenance of genetic
interchange within a phylad are open. Genetic barriers among island spe-
cies, especially those which have resulted from adaptive radiation of a
single immigrant, will be expected to be virtually absent. Reasons for this
absence include the tendency of oceanic island floras to emphasize peren-
nials, especially woody species. Such growth forms are favored in the mod-
erate climate of many islands, and woody species characteristically develop
few interspecific sterility barriers (Baker, 1959). Also, geographic isolation
seems to predominate over sterility barriers as the main mechanism of
isolation among species in genera on islands (observation and experimental
evidence are cited below). Moreover, genetic barriers would be deleterious,
for they would subdivide a stock into portions having less heterozygosity
(and thus less genetic momentum for long-term survival in isolation) than
the whole. Other factors that tend to maintain integrity of insular species
include distinctive ecological adaptations, lessened dispersibility (precinc-
tiveness), flower constancy of pollinators, and a tendency toward a high
level of autogamy. Where one factor predominates, another may be lack-
ing. Loss of dispersibility occurs prominently on oceanic islands (chapter
11). A dioecious species will harbor a high level of outcrossing. An alter-
native scheme, however, would be a group of interfertile species, capable
of outcrossing, which reproduce mostly by autogamy. Hybrid swarms will
form occasionally in such species. Lack of genetic barriers and other factors
will tend to produce varied and fertile recombinants. Hybrid types can
stabilize and persist. With good dispersibility in such a species group, hy-
brids may be disseminated to form new colonies—as may the parental
types. The unit within which genetic interchange occurs in such a case is
not the species, but the entirety of a genus as represented on an archi-

511

Reproductive Biology on Islands

pelago. The vehicle for genetic interchange is not mere outcrossing, but
successive events of hybridization. In addition to the two examples given—
dioecious species and hybrid complexes—innumerable compromises in
genetic and floral systems may be imagined, but all systems that succeed
over long periods of time will contain means for effective outcrossing.

Rattenbury (1962) has stressed the value of hybridization as a means
of retaining genetic vigor in the restrictive environments of the Pleistocene
in New Zealand. The small land areas available during that period may
have been significant in raising the selective value of hybridization and
outcrossing. However, either a very large portion or all of the New Zealand
flora was established via long-distance dispersal (Davis, 1950; Darlington,
1965). Therefore, New Zealand plants may, in fact, be compensating for
loss of genetic contact with mainland relatives by a high degree of out-
crossing (of which hybridization is a heightened form) within the island
populations. If continental populations are larger than insular ones, only
a small proportion of individuals need to participate in outcrossing to
maintain a relatively high level of heterozygosity. If, on the other hand,
island populations are smaller, a greater proportion of individuals must
be involved in outcrossing if a similar level of heterozygosity is to be main-
tained. The mechanisms described in this chapter may be considered ways
of ensuring interbreeding among as many individuals as possible within a
population.

Hawaiian ferns form hybrids, but with relative infrequency (W. H.
Wagner, Jr., personal communication, 1965). Although morphological
change has been great in Hawaiian ferns, genetic systems may have
changed little. Interspecific barriers characteristic of continental ferns may
have been retained. This situation would be expected because relatively
little autochthonous speciation has taken place among the native Hawai-
ian ferns. According to Wagner (in Fosberg, 1948), the number of prob-
able immigrants (135) is only slightly smaller than the number of con-
temporary species plus varieties (168). This contrasts with angiosperms,
where about 250 immigrants have given rise to between 1400 and 2000
species plus varieties native to the Hawaiian Islands (data from Fosberg,
1948). Ferns established easily in the Hawaiian flora, and the compara-
tively large number of immigrant groups may have inhibited the adaptive

BZ

Reproductive Biology on Islands

radiation of each stock. In the Hawaiian angiosperm flora, the relative
paucity of successful immigrants coupled with great ecological opportunity
has resulted in abundant radiation. The products of such radiation are
interfertile species that are “ripe” for events of hybridization. As Ratten-
bury (1962) claims for the New Zealand flora, these “‘species’’ could be
regarded as subspecies. Taxonomy should take into account the fact that
insular species may differ modally in genetic nature from continental
species.

Hybridization may be expected to contribute to the high degree of
polymorphism among the waif floras. This tendency has been noted not
only for the Hawaiian flora, but also for those of the Galapagos Islands
and New Zealand. Polymorphism may be a product of diversification
within a species, but many cases seem derived from events of hybridization.
Polymorphic species may be hybrid complexes in which segregation into
distinctive populations has not taken place.

Most oceanic islands capable of supporting rich floras (Hawauan
Islands, Canary Islands) enjoy a relatively uniform maritime climate. That
tends to prolong flowering times over a large portion of the year (see
Skottsberg, 1928). Thus, outcrossing within species and hybridization
among species are encouraged.

Autogamous species are definitely not barred from establishment on
islands. Baker (1955) has justifiably claimed that such species may be
favored for establishment in insular situations. The evolutionary careers
of autogamous species may be expected to be short on islands, however,
unless occasional outcrossing can occur, as mentioned earlier, or occa-
sional reintroduction of the species from source areas can occur, bringing
in new genetic material. The second possibility would be satisfied by spe-
cies which arrive via oceanic drift. Littoral species in the Hawaiian flora
and other insular floras are, in fact, mostly suited for autogamy. Autogamy
is a condition characteristic of species of pioneer habitats, which open up
and alter rapidly. Occasional reintroduction of a species would compensate
for the genetic depauperation that would inevitably occur in the island
situation. Reintroduction would also tend to swamp any ecological shift
of a beach species and would inhibit radiation in the original population.
These factors would tend to keep a beach species confined to the shore
zone. Species brought by oceanic drift could change if reintroduction did

513

Reproductive Biology on Islands

not occur, or occurred only after ecological change to the forest had been
accomplished; this curriculum can be expected in only a few groups. In
the Hawaiian flora, endemic Fabaceae and Malvaceae seem to be good
examples (see chapters 2 and 11] for explanation and description of inland
shift of the drift flora).

Many species may retain flexibility in ability for exogamy and autog-
amy. Species that are autogamous or nearly so may establish; they may
then maximize exogamy under appropriate circumstances. The partly or
wholly exogamous flora will be semipermanent, provided that land area
is sufficient to foster genetic variability over long periods. Special cases
may be presented on particular island groups. The Faeroes suggest one
such case; there autogamy permits reproduction in species that otherwise
might be deprived of pollinators on account of extreme climatic condi-
tions (Hagerup, 1951). In this case, repopulation of islands by subsequent
immigrants of a native species could retain a species in the flora.

Because floral morphology is an essential part of the framework
within which genetic recombination occurs, it becomes a significant mat-
ter in the preceding considerations. Anemophilous species have no polli-
nation problem that would deter their establishment on islands. For this
reason, they might be expected to bulk large in the floras of oceanic
islands. Entomophilous species can establish most easily if their floral
structure lends itself to access by a variety of pollinating agents. Absence
of long-tongued bees on islands disfavors many angiosperms with larger
and more colorful flowers (other than omithophilous species). Simulta-
neous establishment of an ornithophilous flowering plant and its accus-
tomed bird pollinator is unlikely, so most ornithophilous plants on islands
would be expected to have very versatile floral forms (e.g., Metrosideros)
or else to have evolved autochthonously on islands. Interestingly, Williams
(1927) has shown that Metrosideros is often visited by the bee Nesopro-
sopis, showing that an ornithophilous flower can indeed be versatile. Lin-
sey (1966), citing the work of Perkins (1906, 1913) and Williams (1927),
also suggests another important pollination factor: any given pollinating
insect may visit only a few species of flowering plants (oligolecty) on a
mainland area, but broadens its range of angiosperm species visited on
insular areas.

514

Reproductive Biology on Islands

SURVEY OF THE HAWAIIAN FLORA

Basic to an understanding of genetic systems on islands is a sound
knowledge of floral conditions and floral biology in an insular flora. The
definitive flora for this purpose is that of the Hawaiian Islands. ‘The Gala-
pagos flora, where self-sterility was not encountered during tests (Rick,
1966), may be more recent and not show insular tendencies as well, or may
be inhibited from speciation and radiation because of the dryness of the
archipelago.

The listing of families and their floral characteristics given below is
offered as a basis for discussion. For this purpose, the listing of genera and
species given by Fosberg (1948) has basically been followed. The numbers
of species and varieties given here are somewhat more conservative than
Fosberg’s figures. Information on sexual conditions is not entirely reliable,
and field and herbarium studies were required to reassess their interpreta-
tions. Both Hillebrand (1888) and Skottsberg (1936a, 1936b, 1944a, 1944b,
1945) were keen observers of floral conditions in Hawaiian angiosperms,
and their descriptions proved good starting points for this survey.

Recent information has been incorporated, making changes from my
(1966) paper necessary. Sohmer’s (1972) work on Charpentiera (Amaran-
thaceae) has shown that that genus is functionally dioecious and is there-
fore not truly gynodioecious. George W. Gillett (personal communication,
1971) has called my attention to the fact that ray flowers in Hawaiian spe-
cies of Bidens (Asteraceae) are sterile, not female, so my earlier listing of
this genus as gynomonoecious has been changed. However, Gillett has
shown in his experimental work that at least three Hawaiian species of
Bidens (e.g., B. skottsbergii) are, in fact, dioecious.

In the list that follows, data are given within parentheses after each
genus name. The first figure is the estimated number of species plus varie-
ties. This figure is, of course, an entirely tentative estimate. However, the
percentages in table 13.1 would probably not be altered very much by
different estimates of species numbers. Introduced species were not in-
cluded in the list. Sexual conditions and other information are given within
the parentheses according to the scheme of abbreviations below. Capital
letters indicate that the condition designated characterizes the entire genus,
lower case is used where only a portion of the genus has the condition.
Designations are tentative and oversimplified in some cases (e.g., not all

5

Reproductive Biology on Islands

Poaceae may be fully anemophilous). Additional instances of protandry
and protogyny doubtless occur. More significant, many genera and species
have not been tested with respect to self-sterility.

The last figure within the parentheses is a rating for degree of poten-
tial outcrossing ability according to the scale of values listed. Because
knowledge is incomplete, the ratings must be considered minimal in some
cases. Some combinations of characters tend to reinforce each other to
promote outcrossing. However, the characters have not been treated addi-
tively in designation of the ratings; the feature with the highest numerical
value is the one expressed as a rating for each genus.

Rating Characteristics and abbreviations

5 Dioecism (D); functional dioecism: sterile pollen on some
plants, sterile ovules on others (FD); self-sterility (SS)

4 Anemophily (W); water-pollination (WA)

3 Monoecism (M); polygamodioecism (PD); gynodioecism
(GD); andromonoecism: male + bisexual flowers (AM);
gynomonoecism: female + bisexual flowers (GM); poly-
gamomonoecism: male + female + bisexual flowers (PM)

2 Dichogamy: protogyny (PG) or protandry (PA); spatial sep-
aration of anthers and style so that flower tends to be non-
selfing (NS)

1 Flowers showy (S)

0 ‘Tendency to occupy pioneer habitats (P)

MONOCOTYLEDONS: Arecaceae: Pritchardia (2, S, 1). Cyperaceae: Bau-
mea (1, W, 4); Carex (9, W, PA, 4); Cladium (1, W, 4); Cyperus (23, W,
4); Eleocharis (1, W, 4); Fimbristylis (2, W, 4); Gahnia (6, W, 4); Oreo-
bolus (1, W, 4); Rhynchospora (3, W, 4); Scirpus (5, W, 4); Scleria (1, W,
M, PA, 4); Uncinia (1, W, M, PA, 4); Vincentia (1, W, 4). Flagellariaceae:
Joinyillea (1, W, 4). Hydrocharitaceae: Halophila (1, WA, P, 4). Iridaceae:
Sisyrinchium (1, S, 1). Juncaceae: Luzula (3, W, 4). Liliaceae: Astelia (12,
D, 5); Dianella (3, S, 1); Pleomele (2, S, 1); Smilax (2, D, 5). Naiadaceae:
Naias (1, WA, M, 4). Orchidaceae: Anoectochilus (2, NS, S, 2); Habenaria
(IS NS, 2); Liparis (1, NS} S, 2). Pandanaceae: Freycinetia (Vl D-ase-
Pandanus (1, D, S, P, 5). Poaceae: Agrostis (3, W, 4); Andropogon (1, W,

516

Reproductive Biology on Islands

AM, 4); Calamagrostis (2, W, 4); Cenchrus (1, W, 4); Deschampsia (3, W,
4); Digitaria (1, W, 4); Dissochondrus (1, W, 4); Eragrostis (12, W, 4);
Festuca (1, W, 4); Garnotia (1, W, 4); Heteropogon (1, W, AM, 4);
Isachne (2, W, 4); Ischaemum (1, W, AM, 4); Lepturus (1, W, 4); Micro-
laena (1, W, 4); Oplismenus (1, W, 4); Panicum (23, W, AM, 4); Paspa-
lum (2, W, 4); Poa (4, W, 4); Sporobolus (1, W, 2); Trisetum (2, W, 4).
Potamogetonaceae: Potamogeton (2, PG, 2).

DICOTYLEDONS: Aizoaceae: Sesuvium (1, P, 0). Amaranthaceae:
Achyranthes (2, W, P, 4); Aerva (1, W, P, 4); Charpentiera (7, W, FD, 5);
Nototrichium (5, W, P, 4). Anacardiaceae: Rhus (1, PM, 3). Apiaceae:
Daucus (1, AM, P, 3); Hydrocotyle (1, P, 0); “Pewcedanum” (2, AM, P, 3);
Sanicula (4, AM, 3); Spermolepis (1, P, 0). Apocynaceae: Alyxia (1, NS, 2);
Ochrosia (1, NS, P, 2); Pteralyxia (2, NS, 2); Rauwolfia (7, NS, S, 2). Aqui-
foliaceae: Ilex (1, D, S, 5). Araliaceae: Cheirodendron (5, S, 1); Reynoldsia
(1, S, 1); Tetraplasandra (10, S, 1). Asteraceae: Adenostemma (1, PA, 3);
Argyroxiphium (6, PA, GM, S, 3); Artemisia (5, W, GM, p, 3); Aster (1,
PA, GM, S, 3); Bidens (30, PA, d, S, p, 2); Dubautia (40, PA, S, 3); Gna-
phalium (1, PA, GM, P, 3); Hesperomannia (6, PA, SS, S, 5); Lagenophora
(3, PA, GM, 3); Lipochaeta (30, PA, GM, S, p, 3); Remya (2, PA, GM, 3);
Tetramolopium (10, PA, GM, S, P, 3); Wilkesia (1, PS, S, 2). Begoniaceae:
Hillebrandia (1, M, S, 3). Boraginaceae: Heliotropium (2, P, 0). Brassica-
ceae: Lepidium (4, P, 0). Campanulaceae: Brighamia (1, PA, S, P, 2);
Clermontia (45, PA, S, 2); Cyanea (60, PA, S, 2); Delissea (10, PA, S, 2);
Lobelia (8, PA, S, 2); Rollandia (10, PA, S, 2); Trematolobelia (5, PA, S,
2). Capparaceae: Capparis (1, S, P, 1). Caryophyllaceae: Alsinodendron
(3, PA, 2); Schiedea (30, PA, gd, p, 2-4); Silene (4, P, 0). Celastraceae:
Perrottetia (1, W? PD, 3). Chenopodiaceae: Chenopodium (1, W, P, 4).
Convolvulaceae: Breweria (2, S, P, 1); Cressa (1, P, 0); Cuscuta (2, P, 0);
Ipomoea (9, S, P, 1); Jacquemontia (1, P, 0). Cucurbitaceae: Sicyos (8, M,
P, 3). Droseraceae: Drosera (1, S, 1). Ebenaceae: Maba (2, D, 5). Elaeo-
carpaceae: Elaeocarpus (1, SS? 5?). Epacridaceae: Styphelia (2, GD, 3).
Ericaceae: Vaccinium (8, S, 1). Euphorbiaceae: Antidesma (6, W, D, 5);
Claoxylon (2, W, m, d, 4-5); Drypetes (1, W, D, 5); Euphorbia (31, M,
p, 3); Phyllanthus (2, W? M, 4?). Fabaceae: Acacia (3, AM, S, 3); Caesal-
pia (1,8, P, 1); Cassia (1; P, 0); Entada (1,S, P, 1); Erythring (1,8, P, 1);
Mezoneurum (1, S, 1); Mucuna (2, S, P, 1); Sesbania (1, S, 1); Sophora (1,

SU

Reproductive Biology on Islands

S, P, 1); Strongylodon (2, S, 1); Tephrosia (1, P, 0); Vicia (1, S, 1); Vigna
(3, S, P, 1). Flacourtiaceae: Xylosma (2, W? D, 5). Gentianaceae: Centau-
rium (1, S, P, 1). Geraniaceae: Geranium (6, S, 1). Gesneriaceae: Cyrtandra
(130, PA, gd, s, 2-3). Goodeniaceae: Scaevola (5, NS, S, P, 2). Haloraga-
ceae: Gunnera (9, W, FD? PA, 5?). Hydrophyllaceae: Nama (1, P, 1).
Lamiaceae: Haplostachys (5, PA, S, P, 2); Lepechinia (1, PA, S, P, 2);
Phyllostegia (55, PA, S, 2); Plectranthus (1, PA, S, P, 2); Stenogyne (30,
PA, S, 2). Lauraceae: Cassytha (1, PG, 2); Cryptocarya (2, PG, 2). Logania-
ceae: Labordia (40, D, S, 5). Loranthaceae (see Viscaceae). Malvaceae:
Abutilon (3, PA, S, P, 2); Gossypium (1, PA, S, P, 2); Hibiscadelphus (4,
RASS, B, 2); Hibiscus (10-PA, S, p,.2); Kokia (4; PA, SP) 2)2Sidaa(Zalaae
S, P, 2). Menispermaceae: Cocculus (2, D, 5). Moraceae: Pseudomorus (1,
M, P, 3). Myrsinaceae: Embelia (2, PD, 3); Myrsine (25, D, 5). Myrtaceae:
Metrosideros (12, S, 1); Syzygium (4, S, 1). Nyctaginaceae: Boerhaavia (1,
P, 0); Pisonia (3, d, 5). Oleaceae: Osmanthus (1, NS, 2). Papaveraceae:
Argemone (1, S, P, 1). Pittosporaceae: Pittosporum (50, D, 5). Plantagina-
ceae: Plantago (10, W, ss, PG, 4-5). Plumbaginaceae: Plumbago (1, S, P,
1). Polygonaceae: Polygonum (1, PA, 2); Rumex (2, PA, 2). Portulacaceae:
Portulaca (6, P, 0). Primulaceae: Lysimachia (13, ns, S, 1-2). Ranuncula-
ceae: Ranunculus (2, S, 1). Rhamnaceae: Alphitonia (1, AM, 3); Colu-
brina (2, P, 0); Gouania (4, PM? P, 3?). Rosaceae: Acaena (1, W, 4); Fra-
garia (1, PG, s, P, 2); Osteomeles (1, S, P, 1); Rubus (2, S, 1). Rubiaceae:
Bobea (4, D, 5); Canthium (1, FD, 5); Coprosma (30, W, D, 5); Gardenia
(2, S, 1); Gouldia (25, GD, 3); Hedyotis (Kadua) (50, D, 5); Psychotria (3,
PA, 2); Straussia (8, D, 5). Rutaceae: Pelea (80, D, 5); Platydesma (6, ss,
S, 5); Zanthoxylum (Fagara) (14, d, m, 3-5). Santalaceae: Exocarpus (3,
W, PM, P, 4); Santalum (6, PG, FD, p, 5). Sapindaceae: Alectryon (2, FD,
5); Dodonaea (8, d, pd, P, 3-5); Sapindus (1, P, 0). Sapotaceae: Nesoluma
(1, GD, P, 3); Planchonella (Pouteria, Sideroxylon) (1, PG, 2). Saxifraga-
ceae: Broussaisia (2, GD or D, S, 5). Scrophulariaceae: Bacopa (1, P, 0).
Solanaceae: Lycium (1, PG, P, 2); Nothocestrum (6, PG, 2); Solanum (7,
PG, S, 2). Sterculiaceae: Waltheria (1, P, 0). Theaceae: Eurya (2, D, 5).
Thymeleaceae: Wikstroemia (18, D or pd, 4-5). Ulmaceae: Trema (1, W,
M, 4). Urticaceae: Boehmeria (1, W, M, 4); Hesperocnide (1, W, M, 4);
Neraudia (9, W, D, 5); Pilea (1, W, M, 4); Pipturus (13, W, m or d, 4-5);
Touchardia (5, W, D, 5); Urera (3, W, D, 5). Verbenaceae: Vitex (1, S,

518

Reproductive Biology on Islands

P, 1). Violaceae: Isodendrion (4, s, p, 0-1); Viola (7, s, 0-1). Viscaceae:
Korthalsella (8, W, M, 4). Zygophyllaceae: Tribulus (1, S, P, 1).

TABE REL oli
Floral conditions and pollination types in the native Hawaiian flora

Genera Genera Species Species
Sexual condition no. % no. %
Dioecious 34 153 412 Lie
Monoecious 16 ee. 1 5.0
Andromonoecious 9 4.1 36 2.5
Gynomonoecious 8 3.6 61 4.4
Gynodioecious 6 Ze 45 2.6
Polygamodioecious 2 0.9 6 0.4
Polygamomonoecious 4 1.8 7 0.5
Total, diclinous 79 34.8 639 44.1
Flowers monoclinous but
dichogamous 3] 14.7 445 29.8
Diclinous + dichogamous 110 4553) 1084 139
Wind-pollinated fiz 32.4 295 19.8

A feature which emerges clearly from the tabulation is that species
which show few features designed for exogamy are often pioneer species.
Many of these, if studied, would doubtless prove to be self-fertile. This is
true for the Hawaiian cotton, Gossypium sandvicense (= G. tomentosum),
according to Stephens (1964). The beach Scaevola, S. taccada (= S. sericea)
is cleistogamous according to George W. Gillett (personal communica-
tion, 1965). Baker (1967) speculates that the Hawaiian Pandanus may be
apomictic. Flowers of Convolvulaceae are notably suited for self-pollina-
tion. Of the species termed pioneer-habitat species, the vast majority are
rated 0 or 1 in potential outcrossing ability. Only a scattering are rated 4
(mostly owing to wind pollination). Because Asteraceae are typically plants
of pioneer habitats, they may be expected in such sites regardless of ability
to outcross. Of the beach species native to the Hawaiian Islands, most
fall in rating 0 or 1. In the Galapagos flora, autogamy has been demon-
strated in Lycopersicon (Rick, 1963) and many other genera (Rick, 1966).
The Galapagos flora may be considered a pioneer flora almost entirely;
true wet forest is absent there.

Unexpectedly low ratings characterize a few genera from the forest
flora (Cheirodendron, Cryptocarya, Elaeocarpus, Syzygium, Gardenia,

519

Reproductive Biology on Islands

Lysimachia, Metrosideros, Osmanthus, Tetraplasandra). Study of these
genera may demonstrate features that would tend to prevent inbreeding.
A combination of showy inflorescences, fragrance, and abundance of nec-
tar (e.g., Cheirodendron) may be effective in securing some degree of out-
crossing ‘even if plants are self-fertile. Low ratings for some plants of the
high bogs (Drosera, Viola, Vaccinium) suggest that bogs may be, in a
sense, pioneer habitats and thus show a lack of exogamy, as in beach spe-
cies. Hawaiian bog plants appear relatively close to bog plants elsewhere,
and with few exceptions seem not to have evolved or radiated greatly in
the Hawaiian Islands; instances of loss of dispersibility (chapter 11) are
few in the bog flora.

DIOECISM

Grant (1958) states that evolution of dioecism requires a high selec-
tive pressure. The high percentage of dioecism in the New Zealand flora
has been noted by various workers. Millener (1961) states that 25 percent
of New Zealand plants show unisexuality, but Thomson (1880) claims 46
percent; “unisexuality”’ in Thomson’s sense is probably equivalent to
“Diclinous” in table 13.1, where the 44.1 percent would correspond closely
to his estimate for New Zealand. Cockayne (1921) claims 49 percent
diclinous in the lowland forest of New Zealand. The percentage of dioe-
cious species in New Zealand is claimed to be 14.5 percent (Godley,
quoted by Rattenbury, 1962). This figure is well below the Hawauan flora’s
27.7 percent. These insular floras have high dioecism compared with other
floras, however: 4.4 percent of species are dioecious in southwestern Aus-
tralia (McComb, 1966); the figure is 3.1 percent in the British Isles (Mc-
Comb, 1966) and 2.6 percent in California (Baker, 1967). At any rate, the
Hawaiian flora has a dioecism rate well above those previously reported.

Other oceanic islands also are notable for dioecism. In the Juan Fer-
nandez Islands, the frequency of dioecism is perhaps less striking than
the groups in which it occurs. Two endemic genera of Asteraceae, tribe
Senecioneae, are dioecious: Robinsonia (Symphyochaeta is sometimes
recognized as a segregate of this) and Rhetinodendron; the latter even
suggests an approach toward wind pollination (Skottsberg, 1928). The en-
demic genus Cuminia (Lamiaceae) may have evolved dioecism autochtho-

520

Reproductive Biology on Islands

nously on the Juan Fernandez Islands. The presence of a dioecious palm
there, Juania, parallels the dioecious Lodoicea of the Seychelles.

On St. Helena Island, an endemic genus of Asteraceae, tribe Helian-
theae, Petrobium (Laxmannia) is dioecious or nearly so (Carlquist, 1957).
Senecio rapae of Rapa appears to be functionally dioecious (Brown, 1935).
The only other dioecious Senecios known are S. antennaria and S. dicli-
nous of the high Andes—an area that might be regarded as insular in
nature (chapter 14). Examples of groups that are monoclinous on conti-
nental areas but dioecious in the New Zealand flora could be mentioned,
as in the Apiaceae (Dawson, 1964).

Pittosporum is dioecious in the Hawaiian Islands but represented by
monoclinous species, P. taitensis and P. orofenensis, in southeastern
Polynesia. Most New Zealand Pittosporums are monoclinous, but a few
are dioecious. The Hawaiian species of Pittosporum (fig. 13.1) are all, as
far as is known, dioecious.

A similar situation obtains in the Polynesian species of Ilex (Aqui-
foliaceae). The Hawaiian Ilex sandwichensis is entirely dioecious but I.
marquesensis (Marquesas Is.) is “polygamous” (polygamodioecious) accord-
ing to Brown (1935), as is, apparently, the Tahitian I. tahitensis.

The genus Myrsine (Rapanea) in the Hawaiian Islands forms another
interesting case. Hawaiian species of Myrsine are described as having
“flowers mostly polygamous” by Hillebrand (1888), although unisexuality
of flowers was not observed by Skottsberg, who was ordinarily acute in
these matters, nor is it discussed by the monographer Hosaka (1940).
Sexual conditions are not easy to establish in Hawaiian Myrsines, because
flowering occurs in winter, when few collections are made, and most
species are collected in fruit. I was able to demonstrate dioecism for
M. knudsen (fig. 13.1) and M. lanaiensis. Anthers contained no pollen
grains at all in the individuals which could be designated as female. In
the individuals with functional anthers, ovaries appeared appreciably
smaller.

In Santalaceae, there is the possibility of cryptic dioecism: microtome
sections I prepared from flowers of one individual of Santalum haleakalae
uniformly showed sterile anthers but well-formed embryo sacs, suggesting
functional dioecism.

The genus Charpentiera (Amaranthaceae) provides an interesting case.

ZI

Reproductive Biology on Islands

< ; SEX ** Schiedea globosa
“\.Charpentiera ovata

FIG. 13.1. Flowers of dioecious species endemic to the Hawaiian Is. Male flowers shown at
left, female flowers at right in each pair. Some sepals, petals, and stamens are removed (cut
surfaces indicated) on the facing side of flowers to show nature of ovaries. Pittosporum acu-
minatum (Pittosporaceae), 3.8; Myrsine knudsenit (Myrsinaceae), X 8.4; Pelea clusiaefolia
(Rutaceae), 2.7; Eurya sandwicensis (Theaceae), X 4.2.

Flowers in this genus were reported by Hillebrand (1888) as bisexual;
Skottsberg termed them gynodioecious. Morphologically, they are gyno-
dioecious (fig. 13.2), and I reported them as such (1966). This characteristic
would be expected, because the genus to which Hillebrand (1888) claims
Charpentiera is allied, Chamissoa, has gynodioecious flowers (Miiller,
1870). However, recent studies by Sohmer (1972) show that plants with
morphologically bisexual flowers in fact act as staminate plants and set few
or no seeds. Consequently, Sohmer (1972) concludes that “functionally,
the members of the genus are often dioecious.” Here, then, we have an
instance of transition between gynodioecism and dioecism.

According to my observations, the genus Eurya (fig. 13.1) is entirely
dioecious in the Hawaiian Islands.

The excellent observations of Skottsberg (1936a, 1936b, 1944, 1945)

522

Reproductive Biology on Islands

on sexual conditions of the native Hawaiian flora do not need repetition
here. They show to what an amazing degree dioecism is present. A
notable example is Wikstroemia (Thymeleaceae). Exploring the hint given
by Hillebrand that the genus is effectively dioecious in the Hawaiian
Islands although monoclinous elsewhere, Skottsberg (1936a, 1944a, 1964)
located a few individuals with bisexual flowers in some Hawaiian species,
but even in those species dioecism was also present, and the remaining
species were found to be exclusively dioecious. A parallel to this behavior,
also in Thymeleaceae, occurs in New Zealand. There the genus Pimelea
is dioecious in at least some species, although the total picture is not
entirely clear.

The Hawaiian Rubiaceae (Skottsberg, 1944b, 1945) present a nearly
uniform picture of dioecism, almost as though dioecious species had
been selected out from Rubiaceae at large. The few native species
of Gardenia, Nertera, and Psychotria lack dioecism. Nertera is wind
pollinated, Psychotria is protandrous. Fosberg (1956) finds that Gouldia
is not truly dioecious, for flowers with connivent stigmas are still capable
of setting fruit. In this case, Gouldia is to be regarded as gynodioecious.
Paralleling the dioecism of Hawaiian species of Hedyotis (formerly recog-
nized as a segregate genus, Kadua), the single species on St. Helena
Island, H. arborea, is dioecious. The remarkable success indicated by
speciation of Coprosma on both New Zealand and the Hawaiian Islands
may, in part, be due to the efficient outcrossing mechanism provided by
dioecism coupled with wind pollination.

The endemic Hawaiian genus of Loganiaceae, Labordia, is a uniformly
dioecious extension of the South Pacific genus Geniostoma. Likewise,
Pelea (fig. 13.1) of the Rutaceae is entirely dioecious in the Hawaiian
Islands, although its South Pacific relatives (Melicope, Acronychia) have
not attained true dioecism.

Astelia (Liliaceae), a widespread genus which is chiefly insular, was
reported to be dioecious by Skottsberg (1934a, 1934b). However, exceptions
do exist. Fruit set can occur (although never abundantly) on male plants
of a few New Zealand species of Astelia (Moore, 1966; Moore and Edgar,
1970). Baker (1967) seems to feel for this reason that Astelia might
have arrived in the Hawaiian Islands in a condition other than dioecious.
However, no exceptions to dioecism have been found in Hawaiian species

Spee

Reproductive Biology on Islands

of Astelia, or in those of other islands, such as Tahiti, Mauritius, and New
Guinea, where this genus occurs. If Astelia has not reached these islands
in a dioecious condition, we have to hypothesize that dioecism has been
independently evolved in all of these (presumably recent) insular species.
The probability seems to me much less than the probability that Astelia
reached these various islands in a dioecious condition, and I feel we must
count the Hawaiian establishment of Astelia as having taken place in the
dioecious condition.

Similar considerations apply to Coprosma (Rubiaceae). If Oliver
(1935) is correct in implying that the Hawaiian C. ernodeoides (dioecious)
is related to the Australian C. moorei (hermaphroditic), a hermaphroditic
ancestry for that one Hawaiian species could be assumed. However, the
remaining Hawaiian species of Coprosma form a natural group which
seems not at all closely related to C. ernodeoides, and thus probably was
derived from a second ancestor. These Hawaiian species all are dioecious,
and are apparently related to dioecious non-Hawaiian species. The fact
that there are monoecious plants of several species of Coprosma in New
Zealand (see Allan, 1961) still faces us with the same situation as in
Astelia. It is much easier to envision that Coprosma has arrived on the
many islands where there are only endemic dioecious species (see Oliver,
1935) in a dioecious condition than to imagine that dioecism has been
independently evolved from monoecism in all of these. A notable exception
is C. pumila, which is dioecious in its broad range in the South Pacific
except for Macquarie Island, where it is monoecious (Taylor, 1954).

All of the Hawaiian Peleas are, as noted, dioecious (fig. 13.1). The
Marquesan Peleas are alleged to be monoecious by Brown (1935), although
my studies on his material in the Bishop Museum could not confirm this.
The species of Pelea claimed for New Caledonia are probably referable
to Melicope. Maba (Ebenaceae) is apparently dioecious in the Hawaiian
Islands, but data on species elsewhere in the Pacific are not available.

Genera that are native in the Hawaiian Islands and are entirely
dioecious on a worldwide basis as far as is known include Pandanus and
Freycinetia (Pandanaceae), Smilax (Liliaceae), Antidesma and Drypetes
(Euphorbiaceae), and Cocculus (Menispermaceae). In addition, the Ha-
walian species of Eurya (Theaceae) belong to the section Proteurya, which
is entirely dioecious. The Hawaiian species of Xylosma (Flacourtiaceae),

524

Reproductive Biology on Islands

like the related Tahitian, Marquesan, Samoan, and Fijian species, are
dioecious, providing good circumstantial evidence for immigration of dioe-
cious stocks rather than independent dioecism in all these cases. Pisonia
(Ceodes) umbellifera is dioecious (Skottsberg, 1936b) not only in the Ha-
waiian Islands but on all other Pacific islands where it is native.

In other genera, however, where non-Hawaiian relatives are non-
dioecious, we can propose the possibility or probability that dioecism
evolved autochthonously in the Hawaiian Islands. This can be suspected
or is conceivable in Hawaiian species of the following genera: Charpentiera
(Amaranthaceae), Ilex (Aquifoliaceae), Claoxylon (Euphorbiaceae), Gun-
nera (Haloragaceae), Labordia (Loganiaceae), Myrsine (Myrsinaceae),
Pittosporum (Pittosporaceae), Bobea (Rubiaceae), Canthium (Rubiaceae),
Hedyotis (Rubiaceae), Straussia (Rubiaceae), Zanthoxylum (Rutaceae),
Santalum (Santalaceae), Dodonaea (Sapindaceae), Broussaisia (Saxifraga-
ceae), Wikstroemia (Thymeleaceae), Touchardia (Urticaceae), and Urera
(Urticaceae). Pisonia (Rockia) sandwichensis (Nyctaginaceae) is a dioecious
species that may have evolved this condition autochthonously. We do
not know for sure whether the Hawaiian Fragaria chiloensis is dioecious;
this species is dioecious on coastal America (Baker, 1967).

The fact that Hawaiian species of at least 10 genera must be inter-
preted as dioecious prior to immigration (other interpretations are much
less likely) is very interesting. It suggests that the value of dioecism in
insular circumstances is so great that these genera have established and
survived despite the requirement for more than a single propagule. As
suggested in chapter 1, this requirement is probably no strong disadvantage,
for occasional colonization by a group of propagules seems more likely
than the hypothesis (implicit although never expressed) that all plant
colonizations on islands begin with one and only one seed. Animals face
a similar problem, unless we hypothesize that 100 percent of animal
colonizations occur in the form of one and only one gravid female. Baker
(1967) finds the 10 genera of plants that with little doubt reached the
Hawaiian Islands in a dioecious condition to be a minor exception.
Although even more than 10 may have, I felt justified in stating these as
exceptions to “Baker's law” (1955) in my 1966 paper. Baker’s “law,”
which he has graciously downgraded to a “rule” (1967), is an acceptable
statement of the potential advantage of autogamous stocks in establish-

525

Reproductive Biology on Islands

ment on islands. By stressing the advantage of outcrossing in long-term
survival (1966), I did not intend to exclude the potential advantage of
autogamy for establishment.

The great value of dioecism for long-term survival on islands, where
land area and population size are limited and outcrossing is therefore
of presumptive value, is suggested by a fact which emerges from the
figures in table 13.1. The number of species per genus in dioecious
Hawauan genera (12.8) is nearly twice the number of species per genus
in the flora at large (6.7). The number of species per genus is also very
high (13.6), however, in the Hawaiian genera that are monoclinous but
dichogamous. These advantages are on an insular area; in contrast, on
large continental areas such as Western Australia or the British Isles the
number of species per genus is about the same, or higher, in hermaphrodite
plants (McComb, 1966). In the Hawaiian Islands, the number of species
per genus in diclinous genera is 8.0, versus 5.9 for hermaphrodite genera.
In southwestern Australia, exclusively diclinous groups have 4.6 species
per genus, exclusively hermaphrodite groups 9.0 species per genus (Mc-
Comb, 1966). In the British Isles, exclusively diclinous groups have 3.2
species per genus, exclusively hermaphrodite groups 2.8 species per genus
(McComb, 1966), which suggests little if any advantage there for non-

hermaphrodite plants.

GYNODIOECISM

The occurrence of female flowers on some plants of a species, and
bisexual flowers on other individuals, is a characteristic which seems
prominent on islands. In the Hawaiian flora, seven genera contain species
which are partly or wholly gynodioecious. Styphelia tameiameiae (fig.
13.2) and its varieties have some flowers with sterile anthers and others
with fertile anthers. Ovaries appear identical in the two types of flowers.
I have observed fruits on both types of plants. Herbarium specimens
(e.g., Sleumer 4559, RSA) have flowers with well-formed pollen grains,
with fruits formed on the same branches; such specimens have been
annotated as bisexual by the monographer M. O. Sleumer. The species
cannot, therefore, be regarded as dioecious, as claimed by Godley (1957).
Godley’s figures for other Epacridaceae—Cyathodes acerosa and C. colen-

526

Reproductive Biology on Islands

Myrsine knudsenii

Pittosporum acuminatum

Pelea clusiaefolia Eurya sandwicensis

Fic. 13.2. Flowers of gynodioecious species (Broussaisia arguta and Charpentiera spp. are
functionally dioecious, however), from the Hawaiian Is. Bisexual (or functionally male)
flowers at left, female flowers at right in each pair. On facing side of flower, perianth, sepals,
corolla portion, and anthers have been removed to show nature of ovaries. Ovaries in male
flowers of Broussaisia arguta (in section, near center) contain normal ovules, although few of
them; only ovaries of female flowers (in section, bottom right) appear to mature into fruits.
Flowers of Schiedea globosa contain structures considered staminodes by some authors, al-
though they may also serve as nectaries. Styphelia tameiameiae (Epacridaceae), X 7.6; Brous-
saisia arguta (Saxifragaceae), X 3.0; Charpentiera ovata (Amaranthaceae), X 9.5; Schiedea glo-
bosa (Caryophyllaceae), « 4.0.

soi—do suggest dioecism. Gynodioecism has been reported in Leucopogon
melaleucoides of the Epacridaceae (McCusker, 1962). Godley reports func-
tional dioecism for the New Zealand species Leucopogon fasciculatus. In
the related family Ericaceae, either dioecism or gynodioecism may occur
in New Zealand species of Gaultheria and Pernettya (Godley, 1957; Frank-
lin, 1962).

The Hawaiian monotypic genus Broussaisia of the Saxifragaceae (a
second species, B. pellucida, is sometimes recognized) has traditionally
been regarded as dioecious, and functionally does appear to be so (fig.
13.2). No fruits like those of female plants are, in fact, formed on male
plants, the flowers of which apparently wither without fruiting. Female

527

Reproductive Biology on Islands

flowers lack anthers and petals. Only the smallest vestiges of petals are
present; minute staminodes may or may not be present. Female flowers
have inferior ovaries, whereas ovaries are superior on male flowers. Male
flowers appear hermaphroditic at first glance; petals and stamens are well
formed, and the ovary appears normal, although smaller than that on a
female plant. Sections of ovaries from male flowers reveal that normal
ovules are present, although they are fewer in number than those of fe-
male flowers (fig. 13.2). Thus, there is a possibility a few seeds might be
formed in “male” flowers, even though the fruits do not seem to mature
normally. Broussaisia appears to have arrived at dioecism or near-dioecism
via gynodioecism. ‘The fact that anthers are virtually absent in female
flowers, whereas well-formed ovules are present in male flowers, suggests
this conclusion.

Charpentiera (Amaranthaceae) of the Hawaiian Islands and the Aus-
tral Islands is a genus that looks gynodioecious (fig. 13.2), and has been
cited as an example of this. However, as a result of Sohmer’s (1972) work
we must, as stated earlier, regard this genus as well along in its progress
from gynodioecism to dioecism, and it is now probably better cited as an
example of dioecism.

Nesoluma polynesicum (Sapotaceae), confined to the Hawaiian
Islands and Austral Islands, is gynodioecious according to the data of Lam
(1938). Gynodioecism characterizes another species of Nesoluma, N. saint-
johnianum (Henderson I.), although apparently N. nadeaudi (Society Is.)
is hermaphroditic. Unisexual flowers have been reported in Planchonella
and Pouteria (Lam, 1938) but are not definitely known in Hawaiian repre-
sentatives of those genera.

Skottsberg (1944a) reported and illustrated in Cyrtandra (Gesneria-
ceae), a genus with bisexual flowers, two Hawaiian species in which in-
dividuals bearing female flowers were observed. I have reexamined Skotts-
berg’s specimens in the Bishop Museum. The specimen of C. platyphylla
var. brevipes (Skottsberg 3178B) alleged to have only staminodes proved
to have reduced anthers, but one fertile anther bearing apparently good
pollen was seen. In Cyrtandra gayana, the specimen Skottsberg 3015 which
was alleged to be female did, in fact, have only sterile staminodes in
flowers. Thus, at least C. gayana may be said to be gynodioecious. Al-

528

Reproductive Biology on Islands

though more material must be studied to confirm that these are not
merely chance occurrences of malformed flowers, the Hawaiian Cyrtandras
may actually show the beginnings of gynodioecism; unisexual flowers do
occur in non-Hawaiian species of Cyrtandra.

Because Fosberg (1956) has found that flowers of Gouldia (Rubia-
ceae) which Skottsberg (1944b) termed male can actually set some fruit,
Gouldia must be regarded as an instance of gynodioecism, transitional to
dioecism.

‘The genus Schiedea (Caryophyllaceae) is endemic to the Hawaiian
Islands. Although no mention of sexual conditions other than hermaphro-
ditic occurs in Hillebrand (1888), Sherff (1945), or Degener’s Flora Hawai-
iensis, my observations show that at least several species deviate from the
bisexual condition. When I examined a flowering population of S. globosa
(fig. 13.2) on the slopes between Hanauma Bay and Koko Head, Oahu,
two kinds of flowers, bisexual and female, were apparent. As far as could
be determined, each plant had only one type of flower, and thus the con-
dition was gynodioecious. As can be seen from figure 13.2, the female
flowers lack stamens; the small petaloid structures which are present in
both female and bisexual flowers have been designated as staminodes by
Sherff (1945). Liquid-preserved specimens of S. globosa revealed that fer-
tile seeds are formed in both kinds of flowers, even though the ovaries are
somewhat smaller in bisexual flowers. Study of herbarium specimens also
revealed the presence of both female and bisexual flowers in S. menziesii
and in S. ligustrina. Specimens of most species of Schiedea are not numer-
ous, and field studies are much needed to determine if gynodioecism pre-
vails in other species as well. At least some species of Schiedea are probably
monoclinous.

A peculiar condition that is functionally near gynodioecism occurs
in another genus of Caryophyllaceae, Sanctambrosia, which is endemic
to San Ambrosio in the Desventuradas Islands (Skottsberg, 1963). In this
genus, some plants bear female flowers, other plants bear both male and
bisexual flowers. Nesocaryum, a genus of Boraginaceae also endemic to
San Ambrosio, has a truly gynodioecious condition (Skottsberg, 1963).
The occurrence, in a very small oceanic-island flora, of two genera that
are in effect gynodioecious is most remarkable. On the Juan Fernandez

529

Reproductive Biology on Islands

Islands is the notable annonalean relict Lactoris, a monotypic genus which
is gynodioecious or nearly so (Skottsberg, 1928; Carlquist, 1964).

In the New Zealand flora, gynodioecism has been carefully studied
in three endemic species of the grass Cortaderia (Connor, 1963, 1965), and
four endemic species of Pimelea of the ‘Thymeleaceae (Burrows, 1960).
Fuchsia excorticata and F. perscandens prove to be gynodioecious (Godley,
1955); Godley describes F. procumbens of New Zealand as “‘trioecious”
and the Tahitian species, F’. cyrtandroides, as either gynodioecious or trioe-
cious. Gingidium (Apiaceae) in New Zealand has been reported to be
gynodioecious (Dawson, 1964).

The number of gynodioecious species in these insular floras is not
dramatically high, but the presence of these, combined with the high
proportion of dioecism, does call for some explanations. Baker (1963)
finds no obvious selective force that would induce gynodioecism. Possible
explanations which can be given are that (1), gynodioecism represents a
transitional stage from bisexuality to dioecism, and may be the pathway
through which dioecism must be achieved in some groups; and (2), gyno-
dioecism promotes a high degree of outcrossing, while seed fertility is re-
tained on all plants, thus providing a potentially high rate of reproduction.
The two explanations may be interconnected, for if a plant were evolving
from monoclinous to diclinous, retention of seed fertility on both female
and bisexual plants would seemingly be more valuable than retention of
pollen fertility on both male and bisexual plants (androdioecism is the
alternative transitional condition). A study of the gynodioecious Leucopo-
gon melaleucoides (McCusker, 1962) showed that female plants had a
fruit set of 62 percent, whereas hermaphroditic plants had 43 percent.
Androdioecism is extremely rare in flowering plants at large, so probably
gynodioecism is the most successful pathway to dioecism. Gynodioecism
is probably an unstable condition over long periods of time. Of the seven
genera in the Hawaiian flora cited above, none appears well stabilized in
this condition. Broussaisia, Charpentiera, Styphelia, and Gouldia tend to-
ward dioecism, while gynodioecism is only incipient in Schiedea, Nesoluma,
and Cyrtandra. Among these seven genera, the gynodioecious condition
was almost certainly present in the stocks that arrived in the Hawauan
Islands in the cases of Charpentiera and Nesoluma, and possibly also in
Styphelia.

530

Reproductive Biology on Islands

MONOECISM AND ALLIED CONDITIONS

The occurrence in the Hawaiian flora of monoecism in Euphorbiaceae
and Urticaceae, gynomonoecism (heterogamy) in Asteraceae, and andro-
monoecism in Apiaceae is not surprising since these conditions occur in
those groups on mainland areas as well. Those groups, however, have suc-
ceeded well in the Hawaiian flora. With regard to Asteraceae, note should
be taken that although many species in the family at large have a pollen-
presentation mechanism which permits self-pollination to occur if cross-
pollination has not already occurred, other features of the family are
suited to outcrossing—such as gynomonoecism and protandry.

Among interesting floral conditions that deserve comment are those
in Acacia, Alectryon, and Alphitonia. Skottsberg (1944a) has demonstrated
andromonoecism in the Hawaiian Acacias, A. koa and A. kauadiensis. Ac-
cording to St. John and Frederick (1949), the two Hawaiian species of
Alectryon (Sapindaceae), A. macrococcus and A. mahoe, must be regarded
as polygamodioecious, because apparently bisexual flowers have been found
in addition to ones that are functionally male or female. However, in his
description of the genus Alectryon, Radlkofer (1931-1933) discounts the
validity of hermaphrodite flowers, so the genus would have to be monoe-
cious or dioecious. In Alphitonia ponderosa, Hillebrand (1888) reported
short-styled flowers mingled with normal long-styled flowers, a report I
have confirmed. Anthers are functional in both kinds of flowers. If the
short-styled flowers are functionally male, A. ponderosa is andromonoecious.

In Hillebrandia (Begoniaceae), plants are monoecious, but as inflo-
rescences open, female flowers only are receptive at first; male flowers (and
additional female flowers) open later; thus, in order to be fertilized, the
first female flowers must outcross.

HETEROSTYLY

Heterostyly is not, apparently, present in the Hawaiian flora (some
Hawaiian genera were erroneously reported as heterostyled, on the basis
of Hillebrand’s vague terminology, in the first printing of my 1965 book,
but that error was corrected in the second printing). The genetic basis for
heterostyly appears more complicated than that for dioecism, and conse-
quently the former has rarely evolved autochthonously on islands. Dioe-
cism, on the contrary, is apparently evolved easily and rapidly in particular

53]

Reproductive Biology on Islands

groups (Lewis, 1942). An example of a genus with both monomorphic
and heteromorphic flowers, represented on continental and insular areas,
is Limonium of the Plumbaginaceae (Baker, 1953). Baker’s data reveal
that, of the species confined to relatively small oceanic islands, 94.5 per-
cent have dimorphic flowers; on islands of all kinds, the percentage is 87.0;
in the genus as a whole, 79.6 percent have dimorphic flowers (my calcula-
tions). Baker (1967) seems to object to my (1966) citing of these percent-
ages, but his objections are not specified, and I think that these percentages
can still be cited as exemplifying the greater value that outcrossing has on
limited land areas. During my 1967 visit to the Canary Islands, I studied
populations of Limonium and was struck by the fact that the endemic
Limonium species all seem confined to small population sizes, where out-
crossing would be an advantage. Limonium tuberculatum, for example,
occurs only in a small marshy part of the sand dunes near Maspalomas,
Gran Canaria (it is probably extinct now because of construction). Limo-
nium rumicifolium occurs in a few very small barrancos near Era del Car-
don, Gran Canaria.

DICHOGAMY

The tabulation for the Hawaiian flora (table 13.1) shows an appreci-
able percentage of species that are either protogynous or protandrous; this
percentage could doubtless be increased as a result of further studies.
Whether the total figure would be greater or the same on a mainland
area remains to be determined. Skottsberg (1928) noted dichogamy in
the Juan Fernandez flora for Eryngium and Colletia, and implied it for
Gunnera. Lauraceae tend to be protogynous, and the flowers of Cryptocarya
in the Hawaiian Islands, as well as those of the various laurels in Macaro-
nesia, may prove to have this condition. There can be little doubt that
dichogamy appreciably increases the possibility of outcrossing. In an
inflorescence of a protogynous species, such as Plantago, the flowers that
open first can only form seed by outcrossing (or, of course, possibly by
pollen from an older inflorescence on the same plant).

WIND POLLINATION
Anemophily will tend not only to promote outcrossing, but also to
promote it over a relatively wide area. Anemophily characterizes 19.8

532

Reproductive Biology on Islands

percent of the Hawaiian angiosperm species (32.4 percent of the genera),
34 percent of Juan Fernandez angiosperms (Skottsberg, 1928), and 29
percent of the New Zealand flora (Thomson, 1880). Although these
percentages are not strongly different, one could say that they support
Skottsberg’s (1928) claim that the high percentage on the Juan Fernandez
Islands is related to windy subantarctic weather. Interestingly, calculation
of pollination mechanisms in the Florida Keys, using Small’s (1913) flora,
reveals that 19.7 percent of species are anemophilous—the same as in the
Hawauan Islands. The value of wind pollination in promoting outcrossing
is obvious, but furthermore, because no animal vector is required in
pollination, establishment of anemophilous groups should be facilitated
on islands. The Hawaiian figures are interesting, because about a third
of the hypothetical immigrants (roughly equating genera in the Hawaiian
flora with the number of original establishments) are anemophilous,
although these have not speciated as well as have the entomophilous and
ornithophilous establishments, so that the percentage of anemophilous
species is much lower than the percentage for genera. The number of
species per genus in the anemophilous portion of the Hawaiian flora is
4.1; if two large genera (which might not be entirely wind pollinated),
Coprosma and Peperomia, are deducted, the figure becomes only 2.8
species per genus. These figures are well below the average number of
species per genus in the Hawaiian angiosperm flora as a whole (6.7 percent,
a percentage identical to the ratio in the California flora as a whole).
Perhaps wind pollination lowers isolation and thereby speciation. Dioecious,
entomophilous flowers would be ideal for securing outcrossing yet would
allow isolation that would tend to promote speciation: in the Hawaiian
flora, dioecious plants have a species-per-genus ratio of 12.8.

Interestingly, Plantago grayana, a native of the very wet Alakai Swamp
of Kauai successfully dehisces its anthers under experimental conditions
of 100 percent humidity, whereas P. lanceolata (a boreal weed) does not
under these conditions (Melvern Tessene, unpublished data).

PATHWAYS TO OUTCROSSING

There seems little doubt that, as Baker (1967) suggests, dioecism is
easy to achieve genetically in comparison to self-incompatibility and
heterostyly. Gynodioecism, monoecism, andromonoecism, and gynomo-

533

Reproductive Biology on Islands

noecism are presumably also easy to achieve. In Hawaiian species
of Plantago, a normally protogynous genus, Tessene (unpublished data)
has discovered occasional gynomonoecious spikes. It will be interesting to
see if any appreciable number of hermaphroditic angiosperms do occur
on oceanic islands. T’essene (unpublished data) has shown that Plantago
grayana of the Hawauan Islands is, in fact, self-incompatible. I would guess
that this condition may have evolved autochthonously on the Hawaiian
Islands. As Baker (1967) suggests, however, the high proportion of dioecism
in the Hawaian flora (and presumably other insular floras) may well
compensate for the lack of genetic self-sterility in hermaphroditic plants,
a compensation representing the difficulty of evolving self-sterility but also
illustrating the high value that outcrossing has on insular areas.

Although gynodioecism probably arises from hermaphroditism and
does not, in most situations, continue on to dioecism (Lewis, 1942), Ross
(1970) credits my contention (1966) that gynodioecism has continued on
to dioecism in the Hawaiian Islands. Evidently the value of ensured
outcrossing there (and apparently in New Zealand, where male plants
outnumber females in the majority of dioecious species [Godley, 1964}])
overrides the retention of a high level of fertility.

Hybridization

Reports of hybridization in insular floras depend on the tendency of
systematists to recognize the presence of hybrids or to neglect them. There
seems no question that with the recent interest by systematists in natural
hybridization, and with an increasing number of botanists trained to
discern hybrids, we may expect the roster of known and probable hybrids
on islands to lengthen greatly. Emphasis here on insular hybrids should
not be interpreted as a denial of the importance of hybrids on continental
areas.

New Zealand botanists appear to have been very alert in reporting
instances of hybridization and species polymorphism. The list of Cockayne
and Allan (1934) has been supplemented by the accounts, under various
dicotyledonous genera, of Allan (1961). Recent or well-documented studies
include those of Oliver (1935), Dawson (1960), and Franklin (1962).

534

Reproductive Biology on Islands

Rattenbury (1962) has offered a hypothesis concerning the survival value
of groups with hybridization and polymorphism during New Zealand’s
Pleistocene.

On the Juan Fernandez Islands, Skottsberg (1922) reports an inter-
generic hybrid, Margyracaena (Margyricarpus setosus < Acaena argentea)
and interspecific hybrids in Gunnera (G. peltata < G. bracteata) and
Wahlenbergia (W. grahamae <X W. larraini), On the Canary Islands,
hybridization probably occurs in Aeonium and Sonchus, and has been
reported in Adenocarpus by Lems (1958). During my fieldwork in the
Canary Islands in 1967, I discovered a hybrid swarm involving three species
of Micromeria (Lamiaceae) at Tamadaba, Gran Canaria.

In the Hawauan Islands, hybrids have been reported in Rubiaceae
for Gouldia (Fosberg, 1937, 1956), and Hedyotis and Coprosma (Fosberg,
1943). In Asteraceae, hybrids have been noted in Bidens (Sherff, 1937;
see, however, a more definitive list in Gillett and Lim, 1970); Dubautia,
including Railliardia (Sherff, 1935); Lipochaeta (Sherff, 1935); Argyro-
xiphium (Keck, 1936); and there is even a report of a hybrid between
two very distinctive genera, Dubautia and Argyroxiphium (Sherff, 1944).
Hybnds between Plantago pachyphylla and P. hillebrandii occur on Puu
Kukui, Maui (M. Tessene, unpublished data). In Campanulaceae, Rock
(1919) reported hybrids between Clermontia montis-loa and C. parviflora.
I have examined this colony (Kulani Camp Road, Hawaii) and found it
to be a clear instance of hybridization, in which all degrees of recombina-
tion between the two parental species were represented and abundant.
In the same region, C. montis-loa hybridizes with C. hawaiiensis (Carlquist,
1970, p. 113). The genus Metrosideros (Myrtaceae), notorious for poly-
morphism in the Hawaiian Islands, may have radiated to a certain extent
by forming distinctive ecotypes, but hybridization of some sort must be
occurring (Fosberg, 1943), because the variability and types found within
a given population suggest active segregation and recombination. Addi-
tional instances will probably be revealed in such “difficult” and “poly-
morphic” groups as Pittosporum and Euphorbia. Potentially very important
in this regard is that many Hawaiian genera contain not merely recent
hybrids, but the products of ancient hybridizations as well. Older events
of hybridization are not easily amenable to analysis, however.

The first example that has been carefully analyzed in the Hawaiian

535

Reproductive Biology on Islands

flora is Scaevola mollis < S. gaudichaudiana (Gillett, 1966). Aside from
the pantropical beach Scaevola, S. taccada, and a distinctive (and probably
old) high-mountain species, S. glabra, most populations of the commoner
native species of Scaevola seem to involve some degrees of hybridization,
recent or old. Gillett’s analysis of the S. gauwdichaudiana hybrids is interest-
ing because apparently no sterility barriers exist in this group; chromosome
studies indicate perfect pairing in obviously hybrid individuals. Hybrid
populations do not seem to segregate markedly toward parental types, as
would be expected in “typical” introgressive behavior where a_ partial
sterility barrier often occurs. These findings suggest lack of generic barriers,
and free exchange of genes is potentially possible among the vast majority
of montane Hawaiian Scaevolas.

A population of Scaeyola on Kauai, termed S. procera by certain
authors, is particularly interesting. ‘This population is nearly uniform, at
least in the area studied by Gillett. It must be regarded as being hybrid,
however. Scaevola “procera” seems probably the result of natural intro-
duction to Kauai of a hybrid from one of the other islands. However, the
intermediate quality of this hybrid has been retained within rather narrow
limits, without marked segregation toward “typical” S. gaudichaudiana
or S. mollis. A population such as S. “procera” could, now that it is isolated
from its parents, develop distinctive characteristics not found in parental
populations. This example shows some ways in which species may have
arisen in various Hawaiian genera.

A recent study of the Hawaiian species of Pipturus (Urticaceae) by
Nicharat and Gillett (1970) shows extensive hybridization, no apparent
genetic barriers, and a combination of a few “source” species to form
many mote or less distinctive and “‘specieslike” hybrids. Skottsberg (1934a)
does concede two hybrids in Pipturus, but some of the species he recognizes
are probably “stable hybrids.” Nicharat and Gillett find that the 13 species
recognized by earlier authors can be considered races derived from hybrids
between two valid species.

Gillett and Lim (1970) have experimentally analyzed Hawaiian species
of Bidens. ‘Their summary of experimentally produced F: and F2 hybrids
shows a notable absence of generic barriers among the species, except
for the highly distinctive Kauai species B. cosmoides. Fully fertile hybrids

536

Reproductive Biology on Islands

between species endemic to different islands were easily obtained in many
cases. This shows clearly that geographical isolation is the only major
factor that prevents the numerous species (or “‘species”) of Bidens from
occurring as a single variable species (except for B. cosmoides). In the
Hawaiian species of Charpentiera, Sohmer (1972) finds no sterility barriers;
he believes that differing ecological requirements may be responsible for
maintaining the integrity of the several species.

Although hybnds have not yet been produced in a controlled experi-
mental manner for other insular genera, one can cite the garden hybrids
of Aeonium (Praeger, 1932). These are so numerous and interbreed so
freely that accurate naming of many Aeonium plants in culture is
impossible.

Self-pollination would tend to maintain hybrid characteristics. If
Hawaiian species of Cyanea and Clermontia (Campanulaceae), Cyrtandra
(Gesneriaceae), or Bidens (Asteraceae) are frequently autogamous but
occasionally outcross, a hybrid, once produced, might tend to maintain
itself. Such a hybrid, if disseminated to a new locality, could found a
distinctive population. Rick’s (1963) analysis of Galapagos tomatoes and
his subsequent studies on the flora of the Galapagos show a lack of genetic
barriers, a high degree of autogamy, and the possibility that these features,
combined with occasional hybridization events, have produced distinctive
populations.

The genetic significance of a moderate to high degree of hybridism
in insular groups is like that of outcrossing within a single species: both
tend to maintain a high level of heterozygosity, which permits evolutionary
flexibility despite loss of genetic contact with mainland relatives, and
small land area. Hybridization probably does, as Rattenbury (1962) sug-
gests, help a group survive a “bottleneck” of dwindling land area and
climatic stress. Volcanic eruptions on oceanic islands could both tempo-
tarily diminish available land area and create new barriers. New land open
to hybrid individuals is also thereby created.

Land area is limited on islands, especially when one considers that
only a portion of an island is ecologically suitable for a particular species.
Where land area is small, the total genetic variability of a species or a
species-complex may be insufficient to maintain the species over a long

537

Reproductive Biology on Islands

period of time. MacArthur and Wilson (1963) have postulated a high rate
of extinction for small and remote islands. This would seem to have a
sound genetic basis. If so, the Hawaiian Islands probably lie above the
minimal area necessary to foster radiation and long-term maintenance of
the products of that radiation. The relatively poor (compared to Hawai)
floras of the Society Islands, the Marquesas, and other islands nearer
important Indo-Malesian source areas than the Hawaiian Islands probably
derive not so much from youth or ecology of southeastern Polynesia but
from inability of species on these smaller land areas to maintain requisite
genetic variability. Whereas the Hawauan flora and fauna contain truly
distinctive products of adaptive radiation (chapter 4), the oceanic islands
of southeastern Polynesia with few exceptions merely contain species that
are endemic but not exceptional within their genera or families. Lowered
adaptive radiation can be shown in the smaller number of species per
genus in the floras of southeastern Polynesia than is the case in the
Hawaiian Islands.

Lack (1971) appears to wonder why plant species seem to withstand
the “small island” effect better than, say, birds. Birds, of course, depend
directly or indirectly on plant species. Land birds, with territorial require-
ments, require much larger land areas than do plants, which have the
additional advantage of being autotrophic and thus are dependent directly
on suitable area, not food supplies. A population of 50 plants of a species
can occur on a small island, like Portulaca lutea on Gardner Pinnacles.
If excessively inbred, this population might eventually die out and be
replaced by a new stock brought by oceanic dnft. However, vegetative
reproduction of a genetically “healthy” Portulaca lutea (and this species
can reproduce by rooting of fragmented stems) could keep such a popula-
tion continuing almost indefinitely by apomixis.

With a plant-eating animal, like the reindeer introduced to St.
Matthew Island that died by overeating their food supply (Klein, 1968),
we see that not only must an animal have an appropmiate ecological niche
and a breeding population above a minimal size, it must also have a
sufhcient food supply to keep that population in equilibrium. Thus, two
species of a given genus of passerine birds cannot be tolerated on some
small islands (Lack, 1971; Lack and Southern, 1949).

538

Reproductive Biology on Islands

Floral Biology

The species native to waif forest floras of islands have flowers that are
notable for their inconspicuousness. These species are adapted mostly to
cross-pollination, however. The inconspicuousness of insular flowers (with
the exception of ornithophilous genera and species) is correlated with the
paucity of large Lepidoptera, long-tongued bees, and various other polli-
nators on islands. Studies are much needed on pollinator relationships
of insular species; since the pioneering work of Thomson (1880) and the
observations offered and summarized by Cockayne (1921) on the New
Zealand flora, there have been only a few major contributions. These
include Heine’s (1937) work on the New Zealand flora, Skottsberg’s (1928)
study of pollination in the Juan Fernandez Islands, and surveys by Linsey
(1966), Rick (1966), and Linsey, Rick, and Stephens (1966) on the Gala-
pagos Islands. In the Hawaiian Islands, we can admire the field-biologist
approach of Perkins (1913), whose observations on floral biology and other
aspects of the Hawaiian biota have, unfortunately, few sequels. A paper
by Williams (1927) is worthy of note, however.

With respect to conspicuousness of flowers, a factor related primarily
to size, Thomson finds that nearly 49 percent of the New Zealand angio-
sperm flora falls in the third, or least conspicuous, of three flower classes.
On Masatierra in the Juan Fernandez Islands, 49 percent of angiosperms
fall in the latter two, or less conspicuous, of three categories. In the
Hawaiian flora, flowers are notably inconspicuous, with a few prominent
exceptions. Smallness of flowers in these floras is due not only to anemo-
phily, but also to smallness of the pollinators of entomophilous flowers.
This, in turn, is related to the fact that small size classes of insects have
been dispersed to islands, in the main. With small insular insects, angio-
sperms with suitably small flowers might be supposed to be most successful
at establishment. An alternative possibility, that smallness of flowers has
evolved autochthonously on islands, seems less likely. In either case, insect
size appears a better explanation for inconspicuousness of entomophilous
insular flowers than does the idea favored by Thomson and by Wallace
(1691), that paucity of insects on islands influences floral characteristics.

With respect to color classes, the following data have been gathered:

539

Reproductive Biology on Islands

New Zealand (data from Thomson, 1880): white, 33 percent;
yellow, 11; red (of all shades), 5; blue or purple, 2.5; green, 48.5.
(Ornithophilous and anemophilous flowers included.)

New Zealand (data from Cockayne, 1921): white, 36 percent;
yellow, 18; reds, 11; blue or purple, 22; green, 11; brown, 2.
(“Lowland forest, attractive” flowers only, anemophilous species
apparently excluded.)

Juan Fernandez Islands (data from Skottsberg, 1928): white, 46
percent; yellow, 24; red, 5; blue, 10; green, 10; brown, 5.
(Ornithophilous and anemophilous flowers excluded.)

Hawaiian Islands (data original): white, 33 percent; yellow, 24;
red (including purple, rose, scarlet), 11; blue, 2; green, 29;
brown, 1. (Anemophilous and ornithophilous flowers excluded.)

Saint Helena Island (data from Hemsley, 1885): white, 50 per-
cent; whitish, tinged pink or bluish, 13; white and yellow
(composites with white rays, yellow disc flowers), 17; yellow,

5

10; red, 7; green, 3. (Entirely native angiosperm flora.)

The similarity among these island groups is evident. ‘The predominance
of white, followed by green, then yellow and red, is clear. White and
yellow are heavily predominant in the Galapagos flora (my observations).

Fragrance is claimed by Thomson to characterize only 22 percent of
New Zealand flowers. A similar lack of fragrance may be found in the
Hawaiian flora as well. Families in which one would expect to find strong
scents—Rubiaceae, Loganiaceae, Lamiaceae—are virtually devoid of them.
Fragrance and nectar are prominent in some Hawaiian plants, however,
perhaps compensating for lack of conspicuousness (Cheirodendron, Tetra-
plasandra, Schiedea).

Many of the entomophilous flowers in the Hawaiian flora have a
simple cuplike, bell-like or tubelike shape that would be suited to a wide
variety of insects. This suggests that easy access by insect pollinators may
have been a major factor in successful establishment of entomophilous
species. The Hawaiian Islands are notably poor in orchids. Although this
lack may derive from short seed viability, it probably is due in part to
the specialized pollination requirements of most orchid flowers. Lack of
specialized pollination requirements in the Hawauan flora might tend to

540

Reproductive Biology on Islands

increase outcrossing and hybridization. Lack of flower-constancy among
pollinating birds and insects would also have this effect.

Other factors that might tend to secure pollination include massing of
flowers into large inflorescences, an insular tendency noted by Wallace
(1891). Wallace’s explanation that this secures pollination where the
insular insect fauna is very poor seems possible. However, there is an
alternative explanation, that congestion of flowers makes a plant in flower
more conspicuous, and therefore induces pollination where individuals of
a particular plant species may be scattered throughout a forest. Congestion
of flowers would compensate for the small flower size that is, as noted
earlier, typical of many insular species. Moreover, conspicuousness by
virtue of clustering would aid in securing pollination where outcrossing is
mandatory—as in dioecious species.

A long flowering season, a tendency noted by Skottsberg (1928) in the
Juan Fernandez Islands and also quite apparent in the Hawaiian Islands,
seems related to the mild maritime climate of many islands. A long
flowering season may tend to promote outcrossing and _ hybridization;
phenological barriers seem unlikely to develop under such conditions.

Difficulties of pollination might account for preponderance of male
over female plants in dioecious species. Precisely this situation was found
in a survey of the New Zealand flora (Godley, 1964), where results were
analyzed for statistical significance. Studies of this sort should be done in
the Hawaiian Islands.

Fertility in Insular Plants

The observations by Godley (1964), as well as the fact that in New
Zealand and Hawaiian genera, gynodioecism has given rise to dioecism,
and that dioecism (whatever its origin) is so extraordinarily common, all
suggest that a high degree of seed fertility is not of prime selective
advantage on those islands. Another subsidiary indication is that in so
many Hawaiian genera, such as Alectryon (Sapindaceae), Pritchardia
(Arecaceae), and many others, gigantic seeds are produced, but relatively
few to a plant (see chapter 11).

In addition, attempts to grow wild-collected seeds of certain Hawaiian

541

Reproductive Biology on Islands

genera have shown remarkably low germination percentages in a number
of cases. Among those in which I and others have found consistently low
germinability, regardless of culture method, are Argyroxiphium sand-
wichense, Wilkesia gymnoxiphium, and Hesperomannia arbuscula oahu-
ensis (Asteraceae). ““Weedier” natives, such as many of the Hawaiian
species of Bidens, appear to have higher germinability. One might well
guess that in many genera, fertility has decreased to an appreciable extent.
Reliable measures of fertility are, of course, dificult to define and obtain.
One might speculate that in the disharmonic Hawaiian situation, where
sO Many genera compete only to a moderate extent compared with their
mainland equivalents, a lower rate of reproduction suffices. Another
possible explanation could lie in the high percentage of woody species. A
relatively low set of seed each year would still, over a period of years,
succeed in reproducing any plant quite adequately.

Reproductive Biology in Animals

FERTILITY IN ISLAND VERTEBRATES

Small clutch size in insular birds as compared to their mainland
relatives has been observed repeatedly (Lack, 1946, 1948, 1954; Hagen,
1952). The data on this phenomenon are graphic. The average clutch size
of robins in Europe and North Africa provides an example (Lack, 1954,
p. 38). The extreme low among areas studied is reached in the Canary
Islands (3.5 eggs per clutch). The Azores have a lower figure (4.7) than
anywhere in mainland Europe (for example, 5.1] in England, 5.5 in
Scotland, 5.9 in Germany, 6.3 in Sweden, 6.3 in Finland). Although the
exact manner of physiological control of clutch size is not known, there
seems little doubt from Lack’s massive evidence that lower clutch size on
islands is governed primarily by availability of food for the young. Although
food availability on islands might, at first guess, appear favorable, food
supplies are lower than on comparable continental areas. Lack (1971) has
extended his interpretation of lower clutch size of birds on islands to
include the uniform climatic conditions typical of many tropical and
subtropical islands. Whereas in areas with marked seasonality, birds will
raise large families at the flush of the growing season, on islands with
uniform climates there is no seasonal abundance of food, and so because

542

Reproductive Biology on Islands

birds must live throughout the year at the limits of a moderate food
supply, clutch size would be indirectly related to climate.

Interestingly, the wall lizards (Lacerta sicula) on offshore islands of
Italy show similar tendencies. The clutch contains 2-4 eggs in insular
populations, 4+-7 on nearby mainland areas (Kramer, 1946). Kramer even
demonstrated clutches of intermediate size by hybridizing mainland and
insular individuals; thus, the feature is a heritable one. The island popula-
tions have larger as well as fewer eggs. Lack (1954) interprets this as a
compensation for poorer food supplies on the small islands, and perhaps
also as a mechanism that enables the young to escape cannibalism—a
tendency of island lizards but not mainland ones. Smaller egg size in
mainland lizards, according to Lack, may be related to the relatively
abundant food supplies available to newly hatched lizards on mainland
areas, whereas the young of insular lizards are faced with sparser food
sources, and therefore provision of greater food supply in the egg enables
them to survive better the difficult juvenile period. Also, there is more
predation upon lizards of mainland areas, so that each individual that
escapes predation can be said to have available the food supply that
individuals removed by predation would have consumed. This makes the
correlations of clutch size and egg size quite comprehensible. Island
populations of at least some bird species are denser than on comparable
continental areas (Grant, 1971); this would seem to support Lack’s idea
that island bird species live at the limit of their food supplies. ‘The
Bahaman hutia, endemic to small East Plana Cay, is notable for its small
litter size: only a single young per litter is typical (Clough, 1969). The
hutia has a high population density.

LONGEVITY OF ANIMALS

In plants, greater longevity on islands occurs in the case of arborescent
representatives of typically herbaceous groups. Lack of seasonal extremes
that would terminate the life of an herbaceous plant may have “released”
some plants into longer life-spans, as suggested in chapter 10. In animals,
we deal with somewhat different principles where longevity is concerned.
Lack (1954) cites an inverse correlation between size of clutch and
longevity. In other words, the more young produced each year, the higher
the mortality rate has to be. He cites figures for the blue tit: in England,

543

Reproductive Biology on Islands

the average clutch is 11, and the annual adult mortality is 73 percent; in
Spain and Portugal, the average clutch is 6, adult mortality 41 percent; in
the Canary Islands, the average clutch is 4.25, adult mortality is 36
percent.

The same correlation occurs in the Lacerta lizards mentioned earlier,
as we might expect (Kramer, 1946). Multiplying clutch size by brood
number, we find that mainland lizards lay 24 eggs each year on the
average, and individuals have an average life-span of 1.9 years. On the
offshore islands, an average of 11 eggs are laid per year, and life-span
averages 4.4 years. The factors discussed above (food supply, predation)
may be called into play in explaining these aspects of mainland versus
insular differences.

References

Allan, H. H. 1961. Flora of New Zealand, vol. I. Wellington, R. E. Owen, Govern-
ment Printer.

Baker, H. G. 1953. Dimorphism and monomorphism in the Plumbaginaceae. II.
Pollen and stigmata in the genus Limonium. Ann. Bot., n.s., 17:433-45.

Baker, H. G. 1955. Self-compatibility and establishment after “long-distance” dis-
persal. Evolution 9:347-49.

Baker, H. G. 1959. Reproductive methods as factors in speciation in flowering plants.
Quant. Biol. 24:177-91.

Baker, H. G. 1963. Evolutionary mechanisms in pollination biology. Science
139:877-83.

Baker, H. G. 1967. Support for Baker’s Law—as a rule. Evolution 21:853—-56.

Brown, F. B. H. 1935. Flora of southeastern Polynesia. III. Dicotyledons. Bishop
Mus. Bull. 130:1—386.

Burrows, C. J. 1960. Studies in Pimelea. I. The breeding system. Trans. Roy. Soc.
New Zealand 88:29-45.

Carlquist, S. 1957. The genus Fitchia (Compositae). Univ. Calif. Publ. Bot. 29:1-
ints

Carlquist, S. 1964. Morphology and relationships of Lactoridaceae. Aliso 5:421—35.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1966. The biota of long-distance dispersal. IV. Genetic systems in the
floras of oceanic islands. Evolution 20:433-55.

Carlquist, S. 1970. Hawaii, a natural history. New York, Natural History Press.

Clough, G. C. 1969. The Bahaman hutia: A rodent refound. Oryx 10:106-8.

Cockayne, L. 1921. The vegetation of New Zealand. Die Vegetation der Erde, ed.
A. Engler and O. Drude, vol. 14. Leipzig, Verlag Wilhelm Engelmann.

Cockayne, L., and H. H. Allan. 1934. An annotated list of groups of wild hybrids in
the New Zealand flora. Ann. Bot. 48:1—55.

544

Reproductive Biology on Islands

Connor, H. E. 1963. Breeding systems in New Zealand grasses. IV. Gynodioecism
in Cortaderia. New Zealand Jour. Bot. 1:258-64.

Connor, H. E. 1965. Breeding systems in New Zealand grasses. VI. Control of gyno-
dioecism in Cortaderia richardii (Endl.) Zotov. New Zealand Jour. Bot. 3:233-42.

Darlington, P. J., Jr. 1965. Biogeography of the southern end of the world. Cam-
bridge, Mass., Harvard University Press.

Davis, J. H. 1950. Evidences of trans-oceanic dispersal of plants to New Zealand.
Tuatara 3:87-97.

Dawson, J. W. 1960. Natural Acaena hybrids from the vicinity of Wellington. Trans.
Roy. Soc. New Zealand 88:13-27.

Dawson, J. W. 1964. Unisexuality in the New Zealand Umbelliferae. Tuatara 12:67-

Fosberg, F. R. 1937. The genus Gouldia (Rubiaceae). Bishop Mus. Bull. 147:1-82.

Fosberg, F. R. 1943. The Polynesian species of Hedyotis (Rubiaceae). Bishop Mus.
Bull. 174:1-102.

Fosberg, F. R. 1948. Derivation of the flora of the Hawauian Islands. In Insects of
Hawaii, vol. 1: Introduction, E. C. Zimmerman, pp. 107-19. Honolulu, University
of Hawaii Press.

Fosberg, F. R. 1956. Studies in Pacific Rubiaceae. I-IV. Brittonia 8:165-78.

Franklin, D. A. 1962. The Ericaceae in New Zealand (Gaultheria and Pernettya).
Trans. Roy. Soc. New Zealand, Bot., 1:155-73.

Gillett, G. W. 1966. Hybridization and its taxonomic implications in the Scaevola
gaudichaudiana complex of the Hawaiian Islands. Evolution 20:506-16.

Gillett, G. W., and E. K. S. Lim. 1970. An experimental study of the genus Bidens
(Asteraceae) in the Hawauan Islands. Unity. Calif. Publ. Bot. 56:1-63.

Godley, E. J. 1955. Breeding systems in New Zealand plants. I. Fuchsia. Ann. Bot.,
n.s., 19:549-59.

Godley, E. J. 1957. Unisexual flowers in the Ericales. Nature 180:284-85.

Godley, E. J. 1964. Breeding systems in New Zealand plants. 3. Sex ratios in some
natural populations. New Zealand Jour. Bot. 2:205-12.

Grant, P. R. 1971. Comment on Simberloff’s letter. Amer. Natur. 105:194-97.

Grant, V. 1958. The regulation of recombination in plants. Cold Spring Harbor
Symp. Quant. Biol. 23:337-63.

Hagen, Y. 1952. Birds of Tristan da Cunha. Results Norwegian Sci. Exped. Tristan
da Cunha 1937-1938 20:1-248.

Hagerup, O. 1951. Pollination in the Faeroes—in spite of rain and poverty in insects.
Kongl. Danske Vidensk. Biol. Selsk. Medd. 18(15):1-47.

Heine, E. M. 1937. Observation on the pollination of New Zealand flowering plants.
Trans. Roy. Soc. New Zealand Bot. 67:133-48.

Hemsley, W. B. 1885. Report on the present state of knowledge of various insular
floras. Rep. Sci. Results Voyage H. M. S. Challenger, Vol. 1, Bot., pp. 1-75. Lon-
don and Edinburgh, Her Majesty’s Stationery Office.

Hillebrand, W. 1888. Flora of the Hawaiian Islands. London, Williams & Norgate.

Hosaka, E. Y. 1940. A revision of the Hawaiian species of Myrsine (Suttonia, Rapa-
nea), (Myrsinaceae). Occas. Papers Bishop Mus. 16:25-76.

545

Reproductive Biology on Islands

Keck, D. D. 1936. The Hawaiian silverswords: Systematics, affinities and phytogeo-
graphical problems in the genus Argyroxiphium. Occas. Papers Bishop Mus.
11(10):1-38.

Klein, D. R. 1968. The introduction, increase and crash of reindeer on St. Matthew
Island. Jour. Wildlife Manage. 32:350-67.

Kramer, ‘G. 1946. Verainderungen von Nachkommenziffer und Nachkommengrosse
sowie der Altersverteilung von Inseleidechsen. Z. Naturforsch. 1:700-10.

Lack, D. 1946. Clutch and brood size in the robin. Brit. Birds 39:98-109, 130-35.

Lack, D. 1948. Further notes on clutch and brood size in the robin. Brit. Birds
41:98-104, 130-37.

Lack, D. 1954. The natural regulation of animal numbers. London, Oxford Univer-
sity Press, Clarendon Press.

Lack, D. 1971. Island birds. In Adaptive aspects of insular evolution, ed. W. L. Stern,
pp. 29-31. Pullman, Wash., Washington State University Press.

Lack, D., and H. N. Southern. 1949. Birds on Tenerife. Ibis 91:607—26.

Lam, H. J. 1938. Monograph of the genus Nesoluma (Sapotaceae). Occas. Papers
Bishop Mus. 14:127-65.

Lems, K. 1958. Botanical notes on the Canary Islands. I. Introgression among the
species of Adenocarpus, and their role in the vegetation of the islands. Bot. Inst.
Nac. Invest. Agron. (Madrid) 39:351-70.

Lewis, D. 1942. The evolution of sex in flowering plants. Biol. Rev. 17:46—67.

Linsey, E. G. 1966. Pollinating insects of the Galapagos Islands. In The Galdpagos,
ed. R. I. Bowman, pp. 225-32. Berkeley and Los Angeles, University of California
Press.

Linsey, E. G., C. M. Rick, and S. G. Stephens. 1966. Observations on the floral rela-
tions of the Galapagos carpenter bee (Hymenoptera: Apidae). Pan-Pacif. Entomol.
42:1-18.

MacArthur, R. H., and E. O. Wilson. 1963. An equilibrium theory of insular zoo-
geography. Evolution 17:373—87.

McComb, J. A. 1966. The sex forms of species in the flora of the south-west of West-
ern Australia. Austral. Jour. Bot. 14:303-16.

McCusker, A. 1962. Gynodioecism in Leucopogon melaleucoides A. Cunn. Proc.
Linnaean Soc. New South Wales 87:286-89.

Millener, L. H. 1961. Our plant world. In New Zealand junior encyclopedia, pp.
310-36. Melbourne.

Moore, L. B. 1966. Australasian asteliads (Liliaceae) with special reference to New
Zealand species of Astelia subgenus Iricella. New Zealand Jour. Bot. 4:201—40.

Moore, L. B., and E. Edgar. 1970. Flora of New Zealand, vol. II. Wellington, A. R.
Shearer, Government Printer.

Miller, F. 1870. Umwandlung von Staubgefassen in Stempel bei Begonia. Uebergang
von Zwitterbliithigkeit in Getrenntbliithigkeit bei Chamissoa. Triadrisch Varietat
eines monadrischen Epidendrum. Bot. Z. 28:149-53.

Nicharat, S., and G. W. Gillett. 1970. A review of the taxonomy of Hawaiian Pip-
turus (Urticaceae) by anatomical and cytological evidence. Brittonia 22:191-206.

Oliver, W. R. B. 1935. The genus Coprosma. Bishop Mus. Bull. 132:1—207.

546

Reproductive Biology on Islands

Perkins, R. C. L. 1906. Aculeate Hymenoptera from the Hawaiian Isles and Austra-
lia. Proc. Hawaiian Entomol. Soc. 1:111-13.

Perkins, R. C. L. 1913. Fauna Hawaiiensis: Introduction. Cambridge, Cambridge
University Press.

Praeger, R. L. 1932. An account of the Sempervivum group. London, Royal Horti-
cultural Society.

Radlkofer, L. 1930-1933. Sapindaceae, vol. I. In A. Engler, ed., Das Pflanzenreich
4(165):1-1018.

Rattenbury, J. A. 1962. Cyclic hybridization as a survival mechanism in the New
Zealand forest flora. Evolution 16:348-63.

Rick, C. M. 1963. Biosystematic studies on Galapagos tomatoes. Occas. Papers Calif.
Acad. Sci. 44:59-77.

Rick, C. M. 1966. Some plant-animal relations on the Galapagos Islands. In The
Galdpagos, ed. R. I. Bowman, pp. 215-24. Berkeley and Los Angeles, University
of California Press.

Rock, J. F. 1919. A monographic study of the Hawaiian species of the tribe Lobelioi-
deae, family Campanulaceae. Mem. Bishop Mus. 7(2):1-394.

Ross, M. D. 1970. Evolution of dioecy from gynodioecy. Evolution 24:827—-28.

St. John, H., and L. Frederick. 1949. A second Hawaiian species of Alectryon (Sapin-
daceae): Hawaiian plant studies, 17. Pacific Sci. 3:296-301.

Sherff, E. E. 1935. Revision of Tetramolopium, Lipochaeta, Dubautia and Railliardia.
Bishop Mus. Bull. 135:1-36.

Sherff, E. E. 1937. The genus Bidens. Part J. Field Mus. Natur. Hist. Publ., Bot.
ser. 16:1-346.

Sherff, E. E. 1944. Some additions to our knowledge of the flora of the Hawaiian
Islands. Amer. Jour. Bot. 31:151-61.

Sherff, E. E. 1945. Revision of the genus Schiedea Cham. & Schlect. Brittonia
5:308-35.

Skottsberg, C. 1922. The phanerogams of the Juan Fernandez Islands. Natur. Hist.
Juan Fernandez Easter I. 2:95—240.

Skottsberg, C. 1928. Pollinationsbiologie und Samenverbreitung auf den Juan Fernan-
dez Inseln. Natur. Hist. Juan Fernandez Easter I. 2:503—47.

Skottsberg, C. 1934a. Astelia and Pipturus of Hawaii. Bishop Mus. Bull. 117:1-77.

Skottsberg, C. 1934b. Studies in the genus Astelia Banks et Solander. Kungl. Svenska
Vetenskapsakad. Handl., ser. 3, 14(2):1—-106.

Skottsberg, C. 1936a. Vascular plants from the Hawaiian Islands. II]. Acta Horti
Gotoburg. 10:97-193.

Skottsberg, C. 1936b. The arboreous Nyctaginaceae of Hawaii. Svensk Bot. Tidskr.
30:722-43.

Skottsberg, C. 1944a. Vascular plants from the Hawaiian Islands. IV. Acta Horti
Gotoburg. 15:275-531.

Skottsberg, C. 1944b. On the flower dimorphism in Hawaiian Rubiaceae. Ark. Bot.
31A(4): 1-28.

Skottsberg, C. 1945. The flower of Canthium. Ark. Bot. 32A(5):1-12.

547

Reproductive Biology on Islands

Skottsberg, C. 1963. Zur Naturgeschichte der Insel San Ambrosio (Islas Desventur-
adas, Chile). 2. Bliitenpflanzen. Ark. Bot., ser. 2, 4:465-88.

Skottsberg, C. 1964. Wikstroemiae novae Hawaiienses. Svensk. Bot. Tidskr. 58:177—
83.

Small, J. K. 1913. The flora of the Florida Keys. New York, privately published.

Sohmer, S. H. 1972. Evolutionary trends in the genus Charpentiera (Amaranthaceae).
Brittonia 24:283-312.

Stephens, S. G. 1964. Native Hawaiian cotton (Gossypium tomentosum Nutt.). Pa-
cific Sci. 18:385—98.

Taylor, B. W. 1954. An example of long-distance dispersal. Ecology 35:569-72.

Thomson, G. M. 1880. On the fertilization, etc., of New Zealand flowering plants.
Trans. New Zealand Inst. 13:241-88.

Wallace, A. R. 1891. Natural selection and tropical nature. London, Macmillan and
Go;

Williams, F. X. 1927. Notes on the habits of the bees and wasps of the Hawaiian
Islands. Proc. Hawaiian Entomol. Soc. 6:425-64.

548

EQUATORIAL HIGHLAND BIOTA

At first glance, plants and animals of equatorial peaks and highlands
would appear not to merit consideration in an insular context. ‘Temperate
mountains, or even continental islands would seem to merit more con-
sideration. Equatorial highlands do tend to offer rather predictable climatic
conditions. These conditions can be considered too uniform for comparison
with the varied ecology of oceanic islands only if we select a single zone
on an equatorial peak. However, if we consider the equatorial highlands
as isolated areas of temperate flora and fauna from, say, 1000 m to the
uppermost limits of vegetation, they form a gamut in ecological conditions
very much like those of oceanic islands. Although this gamut is not as
wide as, for example, that of the Hawaiian Islands, an appreciable
amplitude is presented, and some genera, like Espeletia, are able to show
within that amplitude a surprising range of growth forms and adaptations.
The lower limits of the equatorial upland flora tend to be uneven, merging
into the uppermost tropical elements, whereas on oceanic islands the
shoreline provides a very discrete limit. If we take this into account,
however, we are dealing with a decidedly insular situation.

To be sure, species within the larger or more important genera of
equatorial highlands often represent vicarious species that show only minor
differences from one peak to another, differences that do not appear to
be adaptive in nature. However, adaptive ranges do occur within the area
described above, and the nature of these adaptational modes forms the
primary concern of this chapter.

Obviously, I am defining “equatorial highland flora’ more broadly
than most botanists, who focus on the higher vegetational belts. For
example, Hedberg’s Afroalpine flora (1957) and Coe’s (1967) monograph

549

Equatorial Highland Biota

cover the levels from about 3500 to 4500 m. Such elevations correspond
to the “typical” paramos in Colombia and Venezuela. However, my own
observations in the African highlands suggest that temperate elements
dominate vegetation much lower than that, and if we omit these subalpine
zones, we hinder our understanding of the temperate equatorial flora and
fauna in all its insular aspects. The alpine zone represents a series of
extreme adaptations which, although very interesting in themselves, are
only one phase of the temperate equatorial biota.

That tropical highlands tend to have temperate elements, rather
than cool-tolerant representatives from groups in nearby tropical lowlands,
is well known. An analysis of the Afroalpine flora (Hedberg, 1965) shows
this clearly. Thus, the equatorial highland flora is the product of long-
distance dispersal and, if we define broadly the zones occupied by these
elements, shows all the aspects of oceanic islands. In this chapter, I am
considering three areas:

(1) ‘The high equatorial mountains of Africa (Mt. Kenya, Mt. Elgon,
Mt. Kilimanjaro, Mt. Meru, Arberdare, Ruwenzori, and to a lesser extent
other eastern African highlands).

(2) Highlands of Malesia (the New Guinean highlands; Mt. Kinabalu;
mountains of Java—Gedeh, for example).

(3) Highlands of Colombia and Venezuela (paramos and subadjacent
regions.

Despite the fact that these areas are widely separated from each other
as well as from the temperate zone, they have an astonishing similarity
in temperate genera: Gentiana, Ranunculus, Viola, and Geranium, for
example, are typically temperate genera found in two or all three of these
regions.

In regarding these areas as insular, two additional features need to
be stressed: that they have cool but relatively uniform climates through-
out the year (see diagrams for Nanyuki and Bagota in fig. 10.1), and that
the flora and fauna of these highlands have been derived by means of
long-distance dispersal. In both of these statements, however, there is need
for qualification, because the situation presented by equatorial highlands
is distinct from that of the typical oceanic island.

The equatorial alpine climate is indeed “cool,” but in a unique man-
ner. The best way to summarize the equatorial alpine (or even subalpine)

550

Equatorial Highland Biota

regime is to say, “Summer every day, winter every night.” The conse-
quences of this climate are highly important. First, plants of tropical low-
lands are adapted to a regime in which night temperatures are only
slightly lower than day temperatures. Conversely, temperate regions, espe-
cially montane ones (during the growing season) not only can experience
marked difference between maximum and nightly minimum on a given
day and fluctuation from week to week in this regard, but also their plants
and animals are all equipped with physiological preadaptations to the ex-
tremes of cold. Thus, temperate species might be predicted to be much
better adapted to equatorial highlands than even the most versatile tropi-
cal species. For climatic information on the alpine zone of Mt. Kenya,
see Coe (1967, pp. 51-65) or Hedberg (1964). Climatography of lower
elevations in African highlands is fairly well known, and the diagram of
Nanyuki in figure 10.1 suggests such a regime. In the alpine zones of equa-
torial mountains, temperatures drop below 0 C every night or almost every
night, providing an extreme to which temperate plants and animals have
special adaptive mechanisms.

Even if temperature is relatively uniform month by month in the
equatorial highlands, precipitation is not (see Coe, 1967, and Hedberg,
1964). This variation causes special problems, for not only is there physio-
logical dryness due to freezing, there can also be dry seasons, which make
the climate additionally xerophytic. The opposite problem characterizes
some equatorial highland areas: excessive moisture, resulting in boggy
conditions, often with continuous mistiness, causes difficulties in pollina-
tion of angiosperms, for example.

Also with regard to dispersal, equatorial highlands are exceptional
when compared to islands. The spectra of dispersal types on particular
islands are discussed in chapter 2. Hedberg (1971) has extended this type
of analysis to the Afroalpine flora. His results for dispersal spectra in the
highest alpine zone contrast markedly with the island situation; presump-
tive dispersal by air flotation bulks very large (27-31 percent) compared
to that on oceanic islands, as does dispersal in mud or on feet of mammals
or birds (55-56 percent). Although transport in feathers or fur is of medium
importance, comparable to oceanic islands (12-14 percent), internal trans-
port of seeds and fruits in birds is almost negligible (1-2 percent). Lower
zones, according to Hedberg’s data, show little change in proportions of

551

Equatorial Highland Biota

these dispersal types, but in the lowest alpine zone (2400-3000 m), trans-
port internally by birds begins to increase. Presumably it increases with
decrease in altitude, for tropical lowlands and rain forest would be expected
to show a high proportion of seeds carried internally in birds. It is also
interesting that the two modes of transport most effective in the African
highlands are relatively ineffective in the colonization of oceanic islands.
To be sure, other authors cited by Hedberg (1971) have postulated in-
crease in anemochory with increased altitude. In this connection, one can
note that for air flotation to be effective across oceanic distances, the dis-
tance must be relatively short, the seed very small and capable of with-
standing the low temperatures of the rapid, high-altitude air streams that
carry seeds for prolonged distances. Where equatorial highlands are con-
cerned, however, there is much less problem of seeds carried by air flota-
tion terminating fatally by falling into a zone of ocean. Should such a seed
fall to the ground, it could ultimately reach a destination in a series of
short hops, a circumstance that vastly increases chances of success over a
dispersal event that must transport a seed totally without contact with
ocean for thousands of miles. Moreover, since alpine plants and especially
their seeds are adapted to cold, transport by cold, high-altitude jet streams
and other extreme air systems is much more feasible than for airborne
seeds of tropical plants, such as orchids.

The possibility of transport in “hops” rather than as “nonstop” dis-
persal events also favors transport of seeds in mud, for a bird or mammal
involved in dispersal to an equatorial highland zone is not required, for
successful transport, to make a nonstop journey in which contact of the
adherent seeds with seawater would prove fatal.

Again, one must stress that events of dispersal do not need to be fre-
quent to account for the equatorial alpine flora. If they were frequent,
genera and species would be identical with those of temperate sources, and
within each of the three areas discussed below, species would not be en-
demic to particular mountains or plateaus, as they often are. That generic
endemism is very low in the three highland areas may suggest that dis-
persal from temperate areas has been rather good, but it may also reflect
the fact that the East African mountains, the mountains of Indo-Malesia,
and the Andes are all relatively recent geologically. The Guayana High-

552

Equatorial Highland Biota

lands (discussed at the end of this chapter) have a much higher endemism
rate, which probably reflects greater geological age.

East African Highlands

The various well-known mountain peaks of East Africa, volcanic in origin
and relatively recent (see Coe, 1967), rise from a plateau. ‘The plateau has
a climate that can be closely compared with that of an oceanic island. At
Nairobi (1675 m), for example, the mean temperature is 19.5 C and the
coldest temperature ever recorded is 6.7 C, figures almost identical with
those for Lord Howe Island, which, however, receives more rainfall, dis-
tributed more evenly through the year. Las Palmas (Canary Is.) has similar
temperature figures, but with more marked dry seasons than Nairobi. ‘The
even temperature throughout the year is notable at Nairobi and other
localities on the Kenya plateau, but one must remember that diurnal-
nocturnal fluctuation does occur and is a temperate aspect.

Submontane Plateau. The temperate—but uniform—conditions of the
East African plateau should make the vegetation of this region much like
that of an oceanic island. In fact, if we examine a flora that includes this
region, such as that of Dale and Greenway (1961), and subtract species of
the lowland Kenya region, we find virtually the same phenomena—even
some of the same groups—as on oceanic islands. To be sure, the Acacia
savannah does not look insular, but it is evergreen and resembles the low-
land savannah of introduced Prosopis in the Hawaiian Islands. More im-
portantly, the genera are temperate with few exceptions. One sees large,
lichen-festooned trees of Podocarpus gracilior. On distribution maps,
Podocarpus may look like a tropical genus, but it grows typically in cool
tropical uplands. Dodonaea viscosa, an abundant element in certain dry
oceanic islands, occurs frequently here. More significantly, one sees ex-
amples of “insular woodiness.” For example, the giant cactoid Euphorbia
trees (fig. 14.1, top left) betoken a frost-free environment in which an
essentially herbaceous genus has responded to uniform, frost-free, tem-
perate conditions. Other typically herbaceous families with woody repre-
sentatives abound on the Kenya Plateau at about 1800 m, as at Naro Moru
and somewhat higher. For example, Heteromorpha (Apiaceae) is excep-

553

Equatorial Highland Biota

tional in its family in becoming arboreal, a truly woody small tree (fig.
14.1, top right). This is also true of the apiad Steganotaenia (a tree 10-15
ft). A number of arborescent and arboreal Asteraceae occur in this region:
shrubby and tree species (some to 20 ft) of Vernonia; two woody species
of Psiadia; Brachylaena merana (a tree to 80 ft); and Tarchonanthus (a
shrub or tree to 20 ft). The absence of herbaceous Asteraceae in the pla-
teau zone is quite notable. Other interesting arborescent and arboreal rep-
resentatives of herbaceous groups here include Cylicomorpha parviflora, a
tree to 80 ft, of Caricaceae; in Solanaceae, three species of Solanum and
one of Withania; and several genera of Lamiaceae.

An interesting and central feature is that if one omits the plateau flora
and considers only the Afroalpine flora, one omits a number of species in
genera that occur also in the alpine zone, and thus one obtains a false pic-
ture of adaptive radiation in these genera. These temperate genera may
well have immigrated not to the high alpine zone but to the plateau lands,
and evolved up into the alpine zone by means of adaptive radiation—or at
least established in intermediate altitudes and radiated both upward and
downward in altitude (the high alpine species appear to be specialized
end products of a radiation). For example, there are woody understory
species of Lobelia in this zone that look like some of the nearly herbaceous,
much-branched Lobelias depicted by Hauman (1933): L. petiolata, L.
squarrosa, and L. intermedia. The genus Crassocephalum (fig. 14.1, bot-
tom left, right) is only a segregate of Senecio, and can be considered a
Senecio with fewer-flowered heads. Is Crassocephalum not merely a pla-
teau tree (to 25 ft) representing the same lines—or something like them—
that have produced the alpine tree-Senecios? If Hauman (1935) had not
restricted his account of Senecio in Africa merely to sect. Dendrosenecio
(the alpine species), we would have seen some of the lower-elevation spe-
cies of Senecio that are very likely related to the Dendrosenecios.

Lower Montane Forest. When we consider lower elevational belts on
Mt. Kenya and the other mountains, the phenomena just stressed become
clearer yet. Hedberg (1957) does not include Lobelia gibberoa (fig. 14.2,
bottom left, right) in his flora, but this species is clearly part of the radia-
tion that has produced the alpine species of Lobelia. Lobelia gibberoa
tends to occur in a wet-forest zone, at approximately 2000-3000 m (Hed-
berg, 1951), in the bamboo-Podocarpus milanjianus forest (a forest that

554

Fic. 14.1. Plants with probably increased woodiness from the plateau region of Kenya. Top
LEFT, Euphorbia candelabrum (Euphorbiaceae), tree 8 m tall; near Lake Nakuru, 1700 m.
Top RIGHT, Heteromorpha trifoliata (Apiaceae), branch from tree; Naro Moru, 1800 m. Bort-
TOM LEFT, RIGHT, Crassocephalum mannii (Asteraceae); Rift Valley near Nairobi, 1400 m.
BOTTOM LEFT, habit, small tree 4 m tall. Borrom ricut, inflorescence and associated leaves;
heads are about 8 mm long.

555

Fic. 14.2.

RIGHT, Hypericum keniense (Hypericaceae). TOP LEFT, habit of tree, ca. 8 m tall; at about
3000 m. Trop ricHT, flower, + cm in diam., and foliage. BOTTOM LEFT, RIGHT, Lobelia gib-
beroa (Campanulaceae). BorroM LEFT, habit, ca. 5 m tall, including inflorescence; at about
2800 m. BOTTOM RIGHT, flowers on inflorescence; corolla pale green, staminal column purple.

556

Plants showing increased woodiness from lower elevations of Mt. Kenya. TOP LEFT,

Equatorial Highland Biota

also contains relatively “primitive” woody. plants,in contrast, such as
Ocotea kenyensis of the Lauraceae). In viewing evolution of Lobelia in
Africa, we cannot neglect L. gibberoa merely because it is not in the alpine
zone. It is a giant herb (taller than the alpine Lobelias) that has stems
dying when they flower, but producing basal offshoots (as do the alpine
Lobelias). Its leaves show no frost-resistance mechanisms. Lobelia gibberoa
is, however, just as much a product of insular woodiness as 1s Lobelia gaudi-
chaudii on the Hawauan Islands.

At the upper borders of the podocarp-bamboo zone, one finds two
other tree species that can hardly be interpreted as anything but examples
of insular woodiness. They are in a uniformly cool, but presumably frost-
free, cloud forest (Hedberg, 1951): Hagenia abyssinica (see shrubby plants,
not yet trees, in foreground of figure 14.2, bottom left) and Hypericum
keniense (fig. 14.2, top left, nght).

Upper Forest. Hagenia is a genus, related to Bencomia and Dendrio-
poterium of the Canary Islands, which shows marked arborescence within
the potentilloid line (broadly construed) of Rosaceae. Trees of Hagenia
attain 50 ft. Trees of Hypericum keniense attain 40 ft. This size is truly
exceptional within the shrub or subshrub genus Hypericum (Hyperica-
ceae). Thus, the Hypericum-Hagenia zone is quite comparable to the wet
cloud forest on the Canary Islands, although in their insular woodiness
these two genera exceed what occurs on the Canaries. The more constant
availability of moisture, along with the cool, uniform climate, has pre-
sumably permitted the amazing arborescence in Hagenia and Hypericum
in the upper forest zone of Mt. Kenya.

Moorlands. Above the Hypericum-Hagenia forest is a bog area; here
we see the effects of high-elevation bog, where wet, cold conditions (evi-
dently with occasional frosts) permit less arborescence. ‘There are, how-
ever, a number of woody species, some from temperate woody groups
(Erica and Philippia of Ericaceae, for example), and some arborescent or
shrubby forms from typically herbaceous groups. On rocky outcrops in the
bog region we can find these ericads as well as shrubby species of Heli-
chrysum (Asteraceae, tribe Inuleae) and Nidorella (a tree to 20 ft: Astera-
ceae, Inuleae). The open bog areas are notable for their species of Lobelia.
On Mt. Kenya, L. keniensis (fig. 14.3, top left, right, bottom left) is a
good example. Although these plants have elongate stems, the stems tend

557

Equatorial Highland Biota

to be prostrate and to root in the bog (fig. 14.3, bottom left) and bend up-
ward into leafy shoots (fig. 14.3, top left) which flower, then die (as in-
dividual shoots); basal innovations may form other shoots from a given
plant. Lobelia keniensis has prominent bracts concealing the purple
flowers. ‘These bracts are landing platforms for the pollinating sunbird,
Nectarinia johnston. Photographs of a visit by the sunbird have been
presented by Hedberg (1964, p. 40). The claw marks left in the bracts of
inflorescences are universally present (fig. 14.3, top right) in plants with
opened flowers, or toward the end of flowering, because repeated visits are
made to each flower in an inflorescence. Nectarinids also visit, according to
my observations in 1967, Lobelias of the lower regions, and I have seen
them on inflorescences of L. gibberoa, in which the numerous dense flowers
provide footholds during visits. Lobelia keniensis extends from the bogs
into the alpine zone, and its ability to close leaves at night (see below)
shows ability to withstand frost (Hedberg, 1964, pp. 57, 61).

The other main component of the bog region on Mt. Kenya is Senecio
brassica (fig. 14.5, bottom left, right). Its growth habits are remarkably
similar to those of L. keniensis. Like L. keniensis, it is frost resistant, the
leaf rosettes can close at night (Hedberg, 1964), and the woolly indumen-
tum on the lower leaf surfaces provides an excellent insulation, retaining
a temperature above freezing in the apical bud (see Hedberg, 1964) even
if the temperature on the outside of the rosette goes below freezing.

The bog species of Lobelia and Senecio are interesting in that they
form such an extensive series of vicarious species on the African volcanoes.
Lobelia keniensis on Mt. Kenya has close counterparts in L. deckentt on
Mt. Kilimanjaro (fig. 14.3, bottom right), L. elgonensis on Mt. Elgon
(Hedberg, 1964, p. 60), L. sattimae on the Aberdares (Dale and Green-
way, 1961), and L. bequaertii on Ruwenzori (Hauman, 1933, p. 21). The
differences among these species are relatively minor, and form a definitive
example of vicarious species. It is interesting that L. gibberoa is found on
various mountains in the form of subspecific (varietal) variants (Hauman,
1933). These may be considered vicarious subspecies, and perhaps are not
distinct species because L. gibberoa occupies lower, broader zones less iso-
lated from one another on different volcanoes than are the bog habitats
of Lobelia keniensis and its vicariants.

558

Fic. 14.3. High-elevation species of Lobelia (Campanulaceae) from African mountains. TOP
LEFT, RIGHT, BOTTOM LEFT, L. keniensis. rop LEFT, habit of flowering plant in bog; ca. 3300
m, Mt. Kenya. top ricut, bracts of inflorescence, showing claw marks made by the bird
Nectarinid. BOTTOM LEFT, portion of prostrate stem from bog, showing formation of large
lateral roots. BoTTOM RIGHT, L. deckenii; from about 3400 m on Mt. Kilimanjaro.

Day

FIG. 14.4. Species of Lobelia from high elevations on Mt. Kenya. top LEFT, L. keniensis, to
show open habit of rosettes during day; rosettes are about 0.7 m in diam. TOP RIGHT, BOTTOM
LEFT, RIGHT, L. teleku; from Teleki Valley, ca. 4100 m elevation. Top ricHT, leaf rosette in
early morning, to show “closed position’ of rosette during nighttime cold. BOTTOM LEFT,
habit of plant in flower, 1.5 m tall. Borrom ricuHT, portion of inflorescence (pith hollow),

showing flowers, very short, ca. 2—2.5 cm long, and hairy bracts, some of which have been

cropped by an herbivore.

560

FIG. 14.5. Subalpine species of Senecio (Asteraceae) from African mountains. TOP LEFT,
RIGHT, S. meruensis; Mt. Meru. top tert, habit of plant, to show relatively slender stem,
broad thin leaves, and lack of accumulation of old, dead leaves on stem. TOP RIGHT, rosette
and massive inflorescence. BOTTOM LEFT, RIGHT, S. brassica; from about 3500 m in bog on
Naro Moru track, Mt. Kenya. porrom LEFT, rosettes among grass in bog, showing woolly
undersurface of leaf. porrom ricuT, plant in flower, about 2 m tall.

561

Equatorial Highland Biota

Senecio brassica is also notable for relationship to vicarious species.
Among these are S. brassicaeformis on the Aberdares. There appears to
have been much more specific differentiation among the Dendrosenecio
species of Senecio, so that vicariads are harder to identify in the tree-Sene-
cios of ‘the higher alpine regions. For example, young individuals of S.
friesiorum on Mt. Stanley look like S. brassica but grow into tall trees of
the S. keniodendron type (Hauman, 1935). Hedberg (1969b) has shown
that species of Senecio from a given climatic belt on one mountain are
not clear-cut vicariants of species from the same belt on other East African
mountains. ‘here is the possibility that vicarious species occur upon a
single mountain, he feels (e.g., a bog species may be most closely related
to a high-altitude species on the same mountain), and thus there would
be vicarious species by zone. ‘There is no question that adaptation (prob-
ably toward higher elevations) has occurred. The question Hedberg
(1969b) raises, in essence, is How many times? In other words, high-eleva-
tion species like S. keniodendron may have originated from lower-eleva-
tion ancestors more than once. Dispersal of Afroalpine Senecio species to
comparable belts on other peaks has undoubtedly occurred, but does this
account for only part of the speciation? As Hedberg indicates, the situa-
tion is not entirely clear, and deserves close further study.

Other bog elements include abundant Festuca tussocks (fig. 14.3,
top left), and, as understory plants, Alchemilla spp., Anagallis serpens,
Viola spp., Swertia spp., and Eriocaulon spp., panboreal genera that form
prostrate stems (in the case of the first three) or dense rosettes (Eriocaulon)
and are presumably frost resistant because of their growth near the ground
in conjunction with other vegetation that would insulate them from frost.
In the wet ground of the bogs, tussock formation and prostrate stems rep-
resent the chief adaptational modes.

Alpine Zone. The alpine zone is the region (mostly 3500-4200 m
|Hedberg, 1968]) noted for giant Senecio species, and indeed their adap-
tations suggest that they and all other elements of this zone must exist
under the highly taxing conditions of nightly frost. Hedberg (1964) has
adequately summarized these, and I shall draw on his account only by way
of showing how temperate genera, mostly those that also occur at lower
elevations, have narrowly restricted adaptations that permit survival in
this extreme environment.

562

Equatorial Highland Biota

Arborescence is possible only if terminal buds can be protected from
frost, if stems can be protected from freezing, and if leaves can be resistant
to both desiccation and frost. Senecio keniodendron illustrates all of these
features clearly (fig. 14.6, top left, right, bottom left). Stems are heavily
clothed with leaves (fig. 14.6, top left). Absence of old leaves on particular
plants is apparently the result of burning. These old leaves form an insula-
tion that keeps the tree trunks slightly above freezing at night (Hedberg,
1964). Even without leaves, the thick corky bark (fig. 14.6, top right) pro-
vides an appreciable degree of insulation. The massive leaves of S. Renio-
dendron are thick, and when they form a closed bud at night (figure 14.6,
bottom left), they protect the apical bud. Hedberg (1964) has shown that
when temperatures on the outside of the leaf rosette are below freezing
(—4 C), they are above freezing (+1.5 C) within the rosette. This is im-
portant for flowering, because the massive inflorescence, which is terminal,
completes much of its growth within the rosette before emerging. What
is not clear is how the inflorescence, once emerged, is able to flower and
set fruit under conditions of nightly frost. It may be that, as with the
leaves of the rosette, higher osmotic pressures within mature tissues are
less subject to freezing than are primordia. Also, more freezing would be
expected at ground level than at the height of an inflorescence on S. kenio-
dendron. The massive rosette of durable leaves and extremely short inter-
nodes suggests a slow growth rate, and in fact Hedberg (1969a) has shown
an annual increase in height of only 2.5 cm.

Resistance to Freezing in the Alpine Zone. Senecio keniodendron
illustrates how the entire morphology of a plant can be adapted to frost
resistance. The other large rosette plants show some of the same features.
All of the alpine Senecios and Lobelias (see fig. 14.4, top right, Lobelia
telekii) are able to close leaf rosettes at night and open them (fig. 14.4,
top left) during the day. Hedberg (1964) believes that the very hairy bracts
of the inflorescence in L. telekii (fig. 14.4, bottom left, right) may serve
as an insulation. Some of the Senecios have leaves that are notably hairy,
as in S. cottonti (fig. 14.6, bottom right)—undoubtedly an insulation, as
such leaves may be for species of Espeletia in the Colombian paramos,
discussed later in this chapter. Hauman (1935) presents drawings of Sene-
cio leaves that show the hairy vesture. Where leaves are relatively thin
and hairless in Senecios of African volcanoes, and where their old leaves

563

FIG. 14.6. Alpine species of Senecio (Asteraceae) from African mountains. TOP LEFT, RIGHT,
BOTTOM LEFT, S. keniodendron; at 4100 m in Teleki Valley, Mt. Kenya. Top Lert, habit of
old, branched tree about 6 m tall, showing accumulation of old leaves. Top RIGHT, surface
of bark, to show thick, furrowed accumulation of corky tissue. BOTTOM LEFT, rosette of young
tree, showing leaves in “closed” position typical of nighttime; photograph taken in early
moming. BOTTOM RIGHT, S. cottonii; trees about 5 m tall; at 4200 m on Mt. Kilimanjaro.

564

Equatorial Highland Biota

do not accumulate and stems are relatively slender, as in Senecio meruensis
(fig. 14.5, top left, right) and S. erici-rosenii (Hedberg, 1964), I suspect
frost-free growing conditions. These species grow not in the high alpine
zone but in the ericaceous belt (Hedberg, 1965). The mechanisms for frost
resistance by Lobelia wollastoni and L. lanuriensis are not readily appatr-
ent. These two Lobelias grow in company of Senecios much like S. kenio-
dendron in morphology (Hauman, 1933), yet they look like Lobelia gib-
beroa, which probably grows in frost-free localities.

Aside from the giant rosette plants, the flora of the high alpine zone
consists of plants that could be called cryptophytes—with renewal buds
located at or near ground level (fig. 14.7, top left, right, bottom left), and
with very short internodes. The flowers or inflorescences of these plants
are virtually sessile. Many conflicting explanations have been advanced for
the “nanism”’ of these plants (see Hedberg, 1964); this habit seems at first
glance illogical, because night temperatures would be expected to be
lowest near ground level. Hedberg (1964) advances a new explanation,
that viscosity of water and consequent difficulty of transport of water
within the plant are characteristic where temperatures fall to 0 C. There-
fore, difficulty in translocation would dictate shorter internodes, pedicels,
and peduncles. However, other explanations may be operative. If the buds
of the giant Senecios are protected from freezing by a rosette of leaves, a
rosette plant at ground level may share this ability. More significant, how-
ever, is the probability—since these low rosette herbs or cushion plants
often grow in wet places—that ground water, during freezing, gives off
considerable heat which prevents freezing of plants that grow on moist
surfaces or near water. This physical phenomenon is well known and is
used by horticulturists and agriculturists, who use watering as a means
of preventing freezing of plants. If operative, the heat given off by freezing
water extends to a localized area, so that plants with elongate internodes
would not benefit from it. One would also expect that osmotic pressures
in these plants are high, for they are covered by hoarfrost during the
night (fig. 14.7, top right) yet are not, apparently, damaged. Young flowers
and inflorescences develop within the protection of vegetative rosettes
(fig. 14.7, bottom left) and emerge only as they mature, often becoming
elongate only in fruit. Fruits and seeds are, of course, much more resistant
to subfreezing temperatures than are flowers, so their emergence above

565

FIG. 14.7. Herbaceous plants in the flora of Mt. Kenya, to illustrate adaptations and char-
actenstics of alpine flora. rop LEFT, clumps of grass through which stream runs; flat-topped
nature may be due to herbivores; narrowing of clumps below is probably due to freezing of

(continued)

566

Equatorial Highland Biota

the rosette plant at this stage is understandable. Further morphological
features of small rosette plants, cushion plants, or trailing plants suggest
frost resistance. Renewal buds are often located below ground level or
within rosettes. In the case of cushion plants, new roots often grow not
into open soil likely to be frozen, but down through the cushion itself,
which does not freeze, and ultimately reach soil below the cushion, which
is likewise protected by the cushion from freezing. Further comments on
cushion plants are offered later in this chapter.

Xerophytism. Adaptations of Afroalpine plants to xeromorphy include
the narrow, inrolled leaves of various grasses—Festuca, Agrostis—to expose
minimal surface (fig. 14.7, top left). The microphylly of the Afroalpine
species of Ericaceae could be cited in this regard. Microphylly characterizes
the shrub Alchemilla argophylla (fig. 14.7, bottom right) as compared with
the broad, cordate leaves of the prostrate Afroalpine species of Alchemilla.
This tendency is true in Helichrysum as well. Hairy or woolly leaf cover-
ings or thick cuticle—both distinctive in many Afroalpine plants—may also
deter transpiration where water availability is low. If permafrost occurs,
relatively shallow root systems would be expected. High osmotic pressure
of leaf tissues, hypothesized above, would counter not only freezing but
low humidity as well.

Protection Against Radiation. Intense sunlight in the alpine zone
provides special problems. Although mist and clouds are frequent, sun-
light is extremely brilliant during the hours when it does occur. The same
hairiness that provides insulation in some alpine species may also function
in filtering sunlight. This seems true in the Afroalpine species of Helichry-
sum (Hedberg, 1964), and would certainly explain the silvery leaves of
Alchemiulla argophylla (fig. 14.7, bottom right). Many plants have charac-
teristic leaf positions other than horizontal, and can be shiny and there-
fore highly reflective, or glaucous. ‘The grasses demonstrate some of these
characteristics. Even the giant rosette plants do not display leaves in a

FIG. 14.7 (continued)

water between clumps at night. Trop ricHT, Ranunculus sp., hoarfrost on plant early in morn-
ing; leaves are not damaged by it, however. Bottom LEFT, Haplocarpha rueppellii (Asteraceae),
formation of heads sessile within rosettes; stem is below ground level, rosette is actually
sunken beneath ground level. sorrom ricut, Alchemilla argophylla (Rosaceae), showing
leaves 2-2.5 cm long, covered with shiny felt of hairs that probably resists solar radiation and
possibly serves as insulation and a desiccation-resistance mechanism. All photographs taken in
Teleki Valley, 4100 m, Mt. Kenya.

567

Equatorial Highland Biota

horizontal position; most leaves are more nearly vertical than horizontal.
Restricted leaf surfaces (Ericaceae) or growth in shady places (Alchemilla
spp. shaded by clumps of grasses) are other mechanisms that would lower
the effects of intense insolation.

Fire and Herbivore Pressure in the Afroalpine Zone. Because of
drought in these mountains, the Afroalpine belt is subject to burning, often
by man. Succulent plants are not affected: the giant Senecios lose their
sheaths of dead leaves but survive. Cryptophytic plants are unaffected, be-
cause fire is not generally a problem for ground-level rosette plants. The
Ericaceae do appear to suffer from fire, but grasses seem, on the contrary,
to be benefited (Hedberg, 1964). Often in alpine regions in various parts
of the world, extensive grasslands may be the result of burning by man
rather than a wholly natural phenomenon (this is almost certainly true in
New Guinea).

Herbivore pressure in the Afroalpine flora is noticeable chiefly in
grasses, which can have “trimmed-off” rosettes. Rodents and the Hyrax
are the chief herbivores of this zone. Although some larger mammals (ele-
phants) occasionally penetrate the alpine zone, the plateau herbivores do
not. ‘The typical plateau herbivores, such as impalas and various species
of gazelles, are very specific in their preferences for particular grass species,
and this fact may restrict them to lower zones.

Pollination in the Afroalpine Flora. In respect to pollination, we have
an interesting parallel with oceanic islands. As perusal of Hedberg’s (1957)
Afroalpine flora shows, colorful flowers are rare. Exceptions occur in the
lobelioids, with purplish flowers almost hidden (except for L. gibberoa, a
subalpine species with greenish flowers); these species are ornithophilous,
however. A fairly high degree of anemophily is present in this zone be-
cause of predominance of grasses and sedges. ‘The remainder of plants have
white or yellow flowers, some greenish and inconspicuous (Alchemilla,
Haplosciadium). Reds, oranges, and blues are notably absent. Even Arabis
(Brassicaceae) has pale flowers. The reasons for lack of colorful flowers
may be the same as on oceanic islands (see chapter 13): absence of long-
tongued bees, relative scarcity of butterflies. Insects are also remarkably
sedentary in this zone (see later section, Insects of Equatorial High-
lands). This situation contrasts with that of temperate mountains, where
alpine flowers are notably colorful. The difference would seem to lie in the

568

Equatorial Highland Biota

fact that temperate mountains have a relatively warm, frost-free growing
season (albeit rather short) during which pollinators adapted to colorful
flowers can become active. In the Afroalpine climate the nightly frosts
would provide a different regime altogether. We would benefit greatly
from a study of pollination in the Afroalpine region (or other equatorial
highlands). Such studies have perhaps not been undertaken to date be-
cause of the discomfort and logistic difficulty involved in living in these
regions, for extended stays are required to secure data on pollination.

Sources of the Afroalpine Flora. Hedberg (1965) has summarized the
affinities of various plants in the African alpine and subalpine flora. The
vast majority, of course, are either north-temperate or “pantemperate.”’
The fact that the temperate lands of the northern hemisphere represent
an enormous source area compared to land areas of the south temperate
zone leads almost inevitably to high representation of boreal elements in
the Afroalpine flora. Hedberg lists only 10 species with affinities to the
Cape region of South Africa, 17 to South Africa other than the cape (most
of these are subalpine and not alpine plants in any case). There are even
five species Hedberg regards as having Himalayan affinities. Hedberg
(1965) recognizes an “‘Afroalpine endemic element” and an “Afromontane
element.” I find these interesting and perhaps necessary categories, but
regrettable, because even if species in these categories are endemic, their
sources inevitably lie elsewhere than on the high volcanoes themselves.
The Dendrosenecios, for example, are endemic to the high African moun-
tains, but their ancestors must ultimately have been Senecios of some
temperate region. Admittedly, we cannot easily unravel the affinities of
the Afroalpine and Afromontane endemics, so it is easy to see why Hed-
berg is not more explicit about affinities of these elements (many are en-
demic species, not genera, as in the case of Helichrysum). The chief lesson
is that size of temperate source area and distance from the African vol-
canoes are the factors that have governed immigration to the volcanoes.
This reminds one a great deal of such Pacific islands as Rapa, where flo-
ristic affinities are governed by distance from source areas and ecological
similarity to source areas (see chapter 2).

In my earlier book (Carlquist, 1965), I selected Thlaspi alliaceum as
an example of a plant species that shows affinity to European montane
areas. This appears to be an unfortunate example if Hedberg (1971) is

569

Equatorial Highland Biota

correct in claiming that Thlaspi may be man-introduced. I might better
have selected an example from Hedberg’s “Boreal element,” such as Ceras-
tium, Arabis, Astragalus, Cardamine, Sedum, Heracleum, Galium, or
Carduus.

Malesian Highlands

Alpine regions occur on several of the Malesian islands. Most notable,
perhaps, are the central highlands of New Guinea, which reach 5029 m.
The photographs of New Guinean highland plants presented here (figs.
14.8; 14.9, top left, nght, bottom left) are from Mt. Piora, 3719 m, about
7° south of the equator. Other notable Malesian island peaks include
Kinabalu (Sabah), 4101 m; Kerintji (Sumatra), 3805 m; and Smeroe (Java),
3670 m. There are numerous peaks above 3000 m on Java and several on
Sulawesi (Celebes), but the greatest area above 3000 m occurs on New
Guinea, and the Malesian highland and alpine flora is most richly repre-
sented on that island. Unfortunately, there is at present no single mono-
graph on the Malesian alpine flora. Stapf (1894) and Meijer (1965) have
offered accounts of the flora of Kinabalu. There are a number of mono-
graphs (notably Royen, 1964a, 1964b) and reports of collections in the
journal Nova Guinea. The Flora Malesiana will ultimately contain ac-
counts of all families represented in the Malesian alpine and subalpine
flora. ‘There are several studies that deal chiefly with Mt. Gedeh on Java
(Rock, 1920; Faber, 1927). These, as well as my personal observations on
Mt. Piora and Mt. Salawaket, New Guinea, are the sources of the com-
mentary below.*

Malesian highlands below the alpine zone consist of forests rich in
Podocarpaceae: Podocarpus brassii, P. brevifolia, P. cupressina, P. imbri-
catus, Dacrydium falciforme, D. gibbsiae, and Phyllocladus hypophyllus,
for example. The presence of podocarp forest below an alpine zone is
reminiscent of the vegetation belts on the African volcanoes. Intermixed
with the podocarps are temperate genera such as Drimys (Winteraceae;
chiefly D. piperita), Olearia spp. (Asteraceae), Haloragis spp. (Haloraga-

* A well-illustrated book, The mountain flora of Java, by A. Hamzah, M. Toha, and
C. G. G. J. van Steenis (Leiden, E. J. Brill, Publisher, 1972) is now available.

570

Equatorial Highland Biota

ceae), Eurya (Theaceae), Elaeocarpus (Elaeocarpaceae), Metrosideros
(Myrtaceae), and numerous species of Rhododendron (Ericaceae). In New
Guinea, scraggly trees of Podocarpus imbricatus, often bearing Dendro-
bium psychrophyllum epiphytically, surround the open, grassy alpine
swales. This grassland is at least partly the result of burning by man.
Within the grassland, there are various shrubby species of such genera as
Vaccinium, Rhododendron, Styphelia, and Coprosma. ‘Tree ferns, such
as Cyathea hymenophylloides, may be abundant (Carlquist, 1965). Within
the grass tussocks, various temperate herbaceous plants may be found.
Among those are cushions of Centrolepis philippinensis (fig. 14.8, top left,
right), Eriocaulon montanum (fig. 14.8, bottom left), Plantago aundensis
(fg. 14.8, bottom nght), Potentilla papuana (fig. 14.9, top left), Astelia
papuana (fig. 14.9, top nght), and Gentiana piundensis (fig. 14.9, bottom
left). I must emphasize that although these genera occur on various moun-
tains of Malesia, the species vary from one mountain to another, and some
may be endemic to a single massif. For example, Royen (1964a) lists 21
species of Gentiana from alpine New Guinea. On Kinabalu, there is yet
another, G. lycopodioides. Among these herbaceous genera, vicarious spe-
cies are as numerous as they are in the Afroalpine flora. Among other
genera common in Malesian alpine areas are Carex (Cyperaceae), Drapetes
(Thymeleaceae), Epilobium (Onagraceae), Euphrasia (Scrophulariaceae),
Gahnia (Cyperaceae), Gleichenia (Gleicheniaceae), Hypericum (Hyperi-
caceae), Lagenophora (Asteraceae), Lycopodium (Lycopodiaceae), Myoso-
tis (Boraginaceae), Ranunculus (Ranunculaceae), Rubus (Rosaceae), and
Tetramolopium (Asteraceae). Many characteristic alpine plants of Malesia
could be added to this list.

However, the features of interest with respect to the Malesian alpine
flora are the various characteristics these plants show that are related to
alpine existence. In the Malesian alpine areas, as on Afroalpine peaks,
temperatures can drop below 0 C at night. Thus, one should expect some
of the same adaptations. That is precisely what occurs—although mostly
in genera different from those of Africa. Tree ferns of Cyathea remind
one somewhat of the Dendrosenecio habit—there are no rosette-tree com-
posites on the Malesian peaks. Perhaps the drooping old leaves and the
numerous leaves covering the apical bud in Cyathea offer frost protection
in much the same way as they do in the Dendrosenecios, although prob-

571

FIG. 14.8. Alpine plants from Mt. Piora, New Guinea. TOP LEFT, RIGHT, Centrolepis philip-
pinensis (Centrolepidaceae). TOP LEFT, habit of cushions, seed from surface. TOP RIGHT, side
view of cushion broken open, showing inflorescences at surface level of cushion, and roots, be-
low, that grow down through cushion. Bottom LEFT, Eriocaulon montanum (Eniocaulaceae),
rosette showing short inflorescence; leaves about 3 cm long. BoTToM RIGHT, Plantago aunden-
sis (Plantaginaceae), rosettes flattened at ground level; leaves 3 cm long.

WI
~I
tN

Fic. 14.9. Alpine plants from New Guinea (top left and right, bottom left) and Costa
ica (bottom right). rop Lert, Potentilla papuana (Rosaceae); leaves silvery, flowers on short
pedicels; Mt. Piora. trop ricut, Astelia papuana (Liliaceae), rosettes showing fruits among
leaves; leaves 8-10 cm long; Mt. Piora. Borrom LeFr, Gentiana piundensis (Gentianaceae),
stems ca. 4 cm tall; Mt. Piora. Borrom ricut, Blechnum werckleanum (Blechnaceae), tree
fern dominating Paramo Ojo de Agua; Macizo de Buena Vista, Cordillera de Talamanca,
Costa Rica. (Photograph by José Cuatrecasas.)

573

Equatorial Highland Biota

ably Cyatheas are not subject to such extreme frost as are the Dendro-
senecios (the same may be said of tree ferns in the genus Blechnum, as in
the highlands of Costa Rica [fig. 14.9, bottom nght] and Colombia). The
cushion-plant habit is represented by Centrolepis philippinensis (fig. 14.8,
top left, nght), and undoubtedly has the same ecological adaptation as
mentioned for cushion plants in Afroalpine and paramo regions.

As in the Afroalpine flora, one notes many rosette plants that else-
where would have elongate inflorescences, but in Malesian alpine areas
have inflorescences that project little, if at all, above the leaves until fruit-
ing. Among the examples shown here are Eriocaulon montanum (fig. 14.8,
bottom left) and Plantago aundensis (fig. 14.8, bottom right), Potentilla
papuana (fig. 14.9, top left) and Astelia papuana (fig. 14.9, top right). In
Plantago aundensis, flowering occurs within the woolly rosettes. The in-
florescences at anthesis are nearly sessile. During fruiting, the succulent,
nonfibrous peduncle elongates rapidly, exposing fruits and seeds for dis-
persal. This habit, unusual for Plantago, suggests the high selective value
of flowering where protection is afforded by wool and leaves of the rosettes.
The wool may insulate, the leaves may give off heat during the coldest
portion of the night. Because the ground is usually quite moist in Malesian
alpine localities, heat generated by freezing might well prevent freezing in
low herbs. Also, the shelter of grass tussocks, as in Afroalpine areas, may
deter freezing. Potentilla papuana (fig. 14.9, top left) is a prostrate plant
among grass tussocks.

Microphylly is characteristic of numerous genera. In addition to the
tree that occurs around grasslands, Podocarpus imbricatus, one can cite
such examples as Gentiana piundensis (fig. 14.9, bottom left), Coprosma,
Rhododendron, Vaccinium, Styphelia, Drapetes, and Tetramolopium. The
hairiness of Potentilla papuana (fig. 14.9, top left) and Astelia papuana
(fig. 14.9, top right) may serve to reflect bright sunshine. Hairiness might
reduce transpiration somewhat in the dry alpine air, but low humidity
certainly seems related to the microphyllous habit. Microphylly might also
be related to lack of seasonality. Conditions are relatively uniform through-
out the year, so that rapid expansion of large leaves, followed by their
death at the end of a season (as in temperate-zone montane plants) is not
to be expected. Instead, leaves suited to relatively harsh but uniform con-
ditions must be formed, leaves suited to indefinite duration. Flowering in

DHE

Equatorial Highland Biota

the Malesian alpine region takes place over an extended season, perhaps
even throughout the entire year (see, for example, Royen, 1964a).

With respect to the geographical afhnities of the alpine plants of
Malesia, one can again cite distance from source and similarity in ecolog-
ical preference as predetermining factors for floristic composition, just as
they are on islands such as the Hawauan Islands. There is, as one might
expect, distinct affinity with temperate southern-hemisphere floras: Astelia,
Coprosma, Styphelia, Centrolepis, Brachycome, and Drapetes, for ex-
ample. Those without clear southerly relatives appear to belong mostly to
what would be called pantemperate genera like Ranunculus, Gentiana,
Swertia, Eriocaulon, Veronica, Myosotis, Cardamine, Vaccinium, Hyperi-
cum, Adenostemma, Erigeron, and Plantago. There are few, if any, genera
characteristic only of boreal temperate regions. For example, although
Epilobium is most highly developed in the northern hemisphere, it is a
large genus in New Zealand, and one might guess that the New Guinean
species have been derived from the south rather than from the north.
There are resemblances to the New Zealand alpine flora, or even floras of
subantarctic islands; but there are numerous differences, too. The subant-
arctic flora is indeed a flora of cool regions, but not uniformly cool as in
alpine Malesia. One genus of remarkable distribution is Tetramolopium
(Asteraceae) with 33 species: 22 occur in upland New Guinea, 11] in the
Hawauan Islands (Koster, 1966). Endemism in Malesian uplands is mostly
at the species level, but one can cite the genus Piora of the Asteraceae
(Koster, 1966). Scyphostegia, a monotypic genus and sole species of its
family, is a Kinabalu endemic but occurs at lower elevations, about 700 m
(Stapf, 1894).

There is considerable irony in the fact that the cool upland forest of
Malesia contains many genera from woody families often regarded as rich
in primitive features: Podocarpus, Dacrydium, and Phyllocladus of the
Podocarpaceae; Papuacedrus of the Cupressaceae; Drimys and Bubbia of
the Winteraceae; and several genera of Monimiaceae. Even though the
Malesian uplands are geologically rather recent, these “primitive” genera
are found there. However, the alpine flora, at a somewhat higher altitude
than the cool-forest elements mentioned above, is composed of “‘special-
ized” genera—most of which are widespread herbs of temperate regions.
The Asteraceae bulk large, as one might expect in a region of climatic ex-

575

Equatorial Highland Biota

tremes, somewhat xeric regime, and in general a pioneer nature. The same
contrast between composition of upland woody forest and alpines occurs
on the African volcanoes (Podocarpus forest containing Ocotea), and
similar considerations apply in the subparamo versus the paramo regions
of Colombia described next.

The Paramos of South America, and Subadjacent
Vegetation Types

The paramos are high cool regions, with bogs and humid ground, in the
central parts of the Andes. The lower limits of the paramos are around
3200 m in Colombia, and they extend upward to regions of permanent
snow (4700-5000 m is the usual upper limit of vegetation in Colombia).
Maximum registered temperatures approximate 12 C, and minima are
about —2 C (Cuatrecasas, 1957). The “subparamo” regions (2500-3500 m)
contain, like the subalpine forests of Africa and Malesia, many families and
genera of woody plants considered to be rich in primitive features: Podo-
carpus (Podocarpaceae), Drimys (Winteraceae), Ternstroemia (‘Theaceae),
Weinmannia (Cunoniaceae), Nectandra and Ocotea (Lauraceae), Brunel-
lia (Brunelliaceae), Clusia (Clusiaceae), Saurauia (Saurauiaceae), Escal-
lonia (Saxifragaceae, Escallonioideae), and Ilex (Aquifoliaceae). To be
sure, these forests do not exclude representatives of specialized families
such as Asteraceae, Rubiaceae, and Myrtaceae (Cuatrecasas, 1934, 1957,
1968).

The paramo proper is a relatively treeless region almost defined by
the genus Espeletia, the “‘frailejon” (figs. 14.10-14.13). The paramos are
comparable to the two equatorial highland regions just considered, but
(although comparable weather data are not really available) are somewhat
moister and more perpetually fogbound; because of the cloud cover, day-
time temperatures are more depressed, and frosts much less frequent, evi-
dently. Among the few trees (unless one considers the taller, rosette-tree
species of Espeletia to be trees) are Senecio vaccinioides, S. pulchellus, and
S. carolii-tercti (Asteraceae); Hesperomeles ferruginea (Rosaceae); Oreo-
panax tolimanus (Araliaceae); and Vaccinium floribundum of the Enica-
ceae (Cuatrecasas, 1934). The presence of a large Vaccinium reminds me

576

Fic. 14.10. Species of Espeletia (Asteraceae) with varied habits, in Colombia. trop Lert, E.
neriifolia var. columbica; below Paramo de Tama in Andean forest, 2800 m, easternmost
Colombia. top ricuT, E. caldasii; Paramo de Santurban, 3400 m, Eastern Cordillera. Borrom
LEFT, E. phaneractis boyacensis; Paramo de Canutos, 3300 m, Boyaca, Eastern Cordillera.
BOTTOM RIGHT, E. santanderensis; Paramo de Santurban, 3800 m, Santander, Eastern Cordillera.
All photographs by José Cuatrecasas.)

De,

FIG. 14.11. Acaulescent species of Espeletia from paramos of Venezuela and Colombia. Top
LEFT, E. conglomerata; Paramo de Santurban, 3800 m, Santander, Eastern Cordillera, Colom-
bia. TOP RIGHT, E. incana; Paramo de la Rusia, 3550 m, Boyaca, Eastern Cordillera, Colom-
bia. BOTTOM LEFT, E. schultzii, Paramo de Mucuchies, 4100 m, Mérida, Venezuela. BOTTOM
RIGHT, E. rositae; Paramo de Hiiina, Boyaca, Eastern Cordillera, Colombia. (All photographs
by José Cuatrecasas.)

578

FIG. 14.12. Rosette-tree species of Espeletia from paramos of Venezuela and Colombia. Top
LEFT, E. timotensis; between Paramo de Mucuchies and Llano Grande de Mifaf, 4300 m,
Mérida, Venezuela. top ricut, E. uribei; Macizo de Bogota, 3350 m, Eastern Cordillera,
Colombia. BOTTOM LEFT, RIGHT, E. pleiochasia; Paramo de la Rusia, ca. 3000 m, Boyaca,
Eastern Cordillera, Colombia. norrom ert, habit, showing broad leaves of this subparamo
endemic. BOTTOM RIGHT, inflorescences; ray flowers are yellow. (All photographs by José
Cuatrecasas.)

579

FIG. 14.13. Rosette-tree species of Espeletia from paramos of Colombia. TOP LEFT, RIGHT,
E. discoidea; Paramo de Canutos, Boyaca, Eastern Cordillera. rop Lert, E. discoidea plants
in landscape that features Weinmannia fagaroides (Cunoniaceae) trees, and shrubs such as
Senecio vaccinoides and Hypericum laricifolium. Tor RIGHT, habit of a plant in flower. BOTTOM
LEFT, RIGHT, E. hartwegiana. BOTTOM LEFT, tall trees in landscape; Paramo de Barragan,
3580 m, Valle, Central Cordillera. Borrom ricut, E. hartwegiana in flower; western slope
of Nevado del Huilo, Cauca, Central Cordillera. (All photographs by José Cuatrecasas.)

580

Equatorial Highland Biota

of Vaccinium varingiifolium of alpine Java (Faber, 1927), and woody spe-
cies of Gaultheria (Ericaceae) likewise suggest a similarity between uplands
in equatorial South America and Malesia.

Espeletia. Although a common denominator of the paramos of Vene-
zuela and Colombia, Espeletia is by no means always the rosette tree that
one often associates with that generic name. To be sure, there are several
such species in the genus (figs. 14.12, 14.13), but Espeletia ranges in form
from a branched tree at least 4-5 m high, as in E. neriifolia (fig. 14.10,
top left), to a virtual cushion plant, forming sessile rosettes, such as E.
caldasii (fig. 14.10, top night). Interestingly, E. neriifolia occupies lower-
forest altitudes (1500-3300 m) in the Andes of Venezuela, from Miranda
westward to Ménda and easternmost Colombia. Other branched species,
shrubs or small trees, include E. glossophylla (Colombia: Sierra Nevada
de Santa Marta, 3200-4000 m) and E. occulta (Venezuela: paramos of
Menida, 3000-3500 m). Espeletia neriifolia is believed to have more numer-
ous primitive characters for the genus (Smith and Koch, 1935). If so, one
might suspect a woodier shrub or tree of forest subadjacent to paramos to
be ancestral for the genus as whole. A few specializations, however, would
be expected to occur in E. neriifolia. The shrubby species tend to be at
the northeastern portion of the range of Espeletia, E.. neriifolia especially
so. If E. nertifolia represents origin of the genus in regions subadjacent to
paramos, it may be like Crassocephalum (Asteraceae) and Lobelia parvi-
bracteata and L. gibberoa, mentioned earlier as lowland relatives of
African upland Senecios and Lobelias. In any case, Espeletia does show
adaptive radiation. There seems no reason to look for origins of any of
these genera in the high-alpine representatives. Cuatrecasas (1954c) states
that in Espeletia there are “15 related species, each of which occupies a
restricted area” within the “original pre-historic area of E. nertifolia.”

Unlike the situation in the Afroalpine Senecios, neither the large
rosette-tree species of E’speletia nor those with large rosettes and short
trunks tend to branch. The difference between Espeletia and the Dendro-
senecios in this respect seems to be largely a matter of inflorescence posi-
tion. In Espeletia, inflorescences are lateral (see figs. 14.10 to 14.13),
whereas in the Dendrosenecios they are terminal, and branching, when
it occurs, results from production of more than one innovation below the
terminal inflorescence.

581

Equatorial Highland Biota

Cuatrecasas (1954c) makes a distinction between the Venezuelan
paramo species of Espeletia, which have a corymbose or racemose type of
inflorescence, and those of the Colombian paramos. Species of the eastern
Colombian Andes have a corymbose dichasial inflorescence; species of the
central~Colombian cordillera are strictly dichasial. Venezuelan species
shown here are E. schultzii (fig. 14.11, bottom left), a widespread paramo
and subparamo species of the states of Mérida and Trujillo; and E.
timotensis (fig. 14.12, top left), a high-paramo species from Mérida.
Species from the Eastern Cordillera of Colombia shown here include,
in addition to E. nertifolia, FE. caldasii (fig. 14.10, top right), E. phaneractis
boyacensis (fig. 14.10, bottom left), E. santanderensis (fig. 14.10, bottom
right), E. conglomerata (fig. 14.11, top left), E. rositae (fig. 14.11, bottom
right), EF. uribei (fig. 14.12, top right), E. pleiochasia (fig. 14.12, bottom
left, nght), and E. discoidea (fig. 14.13, top left, right). Espeletia hartwe-
giana (fig. 14.13, bottom left, right) is from the Central Cordillera of Co-
lombia. Cuatrecasas (1954c) envisions origins of E’speletia in subparamos
of Venezuela, followed by adaptation to paramo. When climatic belts
were lowered in the Pleistocene, migration of Espeletia across depressions
between the high ranges was facilitated, he feels. One such migration,
according to his speculations, included a crossing of ancestors of E.
hartwegiana to the south, continuing onwards “with prodigious vitality”
as far south as the Paramo del Angel (Ecuador). Espeletia certainly seems
capable of reasonably good dispersal, yet there is no doubt that strong
barriers in the form of deep valleys between paramos have resulted in a
high degree of differentiation and endemism among its species. That this
speciation is largely induced by geographic isolation is suggested by the
fact that hybrids form in disturbed areas, with consequent polymorphism
(Cuatrecasas, 1954c). This situation is reminiscent of the situation in
Hawaiian species of Dubautia and Bidens, also members of Asteraceae
(chapter 4). Hybrids are also known to occur among the Afroalpine
Dendrosenecios (Hedberg, 1957). Thus, an insular type of speciation is
very much at work in the Asteraceae of equatorial highlands. Narrow
endemics in Espeletia shown here include E. discoidea (fig. 14.13, top
left, nght), E. pleiochasia (fig. 14.12, bottom left, nght), E. uribei (fig.
14.12, top left), E. rositae (fig. 14.11, bottom nght), E. incana (fig. 14.11,

top right) and E. caldasii (fig. 14.10, top right). These species are all
582

Equatorial Highland Biota

endemic to the respective paramos given in the figure legends. In some
cases, similar-appearing pairs (e.g., E. conglomerata and E. incana; E. uribet
and E. pleiochasia) are produced by the islandlike geographic distribution
of paramos within a single cordilleran range, and can be cited as examples
of vicarious species.

There seem to be diversified aspects of adaptive radiation within
Espeletia. In addition to the branched trees or large shrubs of the scrubby
subparamo forest, such as E. neriifolia, there are small rosette species such
as E. caldasii (fg. 14.10, top right) with one-flowered inflorescences, and
sessile rosette species with very narrow leaves, such as E. phaneractis
subsp. boyacensis (fig. 14.10, bottom left) and E. santanderensis (fig. 14.10,
bottom right). Broader, woolly leaves occur in E. conglomerata (fig. 14.11,
top left), E. incana (fig. 14.11, top right), E. schultzi (fig. 14.11, bottom
left) and E. rositae (fig. 14.11, bottom right). Espeletia rositae has short
stems, as does E. timotensis (fig. 14.12, top left), a species with notably
broad, succulent stems. Taller rosette-tree species shown here include
E. uribei (fig. 14.12, top right), E. pleiochasia (fig. 14.12, bottom left,
right), E. discoidea (fig. 14.13, top left, nght), and E. hartwegiana (fig.
14.13, bottom left, right). The taller rosette-tree species seem characterized
by broader leaves—an obvious correlation, for a large plant with few leaves
will be likely to have large leaves, as in the tall unbranched Hawaiian
lobelioids (Cyanea leptostegia, chapter 4). As in the Dendrosenecios,
woolliness might serve as insulation against freezing, but (because the
paramos tend to freeze less at night than the high Afroalpine zones) there
are probably better interpretations—or a combination of them. Protection
against excessive radiation is one possibility. It is also possible that woolli-
ness might resist wetting (and therefore rotting) in the excessively misty
conditions. As a parallel, one can cite the exceptional woolliness of
Plantago krajinai, endemic to the excessively wet and misty summit bog
of Waialeale, Kauai, Hawaiian Is. (see Carlquist, 1970). It is interesting
to note that the species of Espeletia with the least woolliness, for whatever
reason, tend to occur at lower elevations: E. neriifolia (fig. 14.10, top left)
and E. pleiochasia (fig. 14.12, bottom left, right), for example. Many
species of Espeletia seem characterized by nutant heads: for example,
E. caldasti (fig. 14.10, top right), E. conglomerata (fig. 14.11, top left),
E. incana (fig. 14.11, top right), E. rositae (fig. 14.11, bottom right), E.

583

Equatorial Highland Biota

uribei (fig. 14.12, top right), E. discoidea (fig. 14.13, top left, right), and
E. hartwegiana (fig. 14.13, bottom left, right). There is a striking parallel
in another paramo plant, Senecio canescens subsp. boyacensis (fig. 14.14,
top left, right). The phenomenon of nutant heads may well be a response
of flowering under conditions of excessive mistiness. Interestingly, nutant
heads characterize, and have probably been evolved autochthonously in,
the two bog species of Argyroxiphium in the Hawaiian Islands, A. caliginit
and A. grayanum (Carlquist, 1970). Nutant heads are readily accessible to
pollinators yet are not likely to fill with water, which would hinder insect
visits and other aspects of pollination—water in upright heads would tend
to wash pollen grains away.

There obviously can be different adaptations within the genus Espe-
letia to what do not seem, at least at first glance, markedly different
conditions. The tall rosette-tree Espeletia species occupy much the same
altitudinal area as the species with sessile rosettes.

Herbaceous Components of the Paramos. Just as the open grasslands
of alpine New Guinea contain numerous herbaceous elements such as
Asteraceae and Gentiana, so do the paramos. Interestingly, the herbaceous
equatorial flora of South America even contains some of the same genera.
On the paramos, for example, one sees Gentiana sedifolia; a rosette
Eryngium that reminds me of apiads in alpine Africa and New Guinea;
Erigeron ecuadoriensis, a parallel to Tetramolopium in alpine New Guinea;
and Calamagrostis, a tussock grass very characteristic of the paramos
(fig. 14.15, top left). In the same locality shown in figure 14.15, top left,
one can also find Lycopodium and Espeletia hartwegiana. Calamagrostis
effusa and Geranium multiceps are shown in figure 14.15, top right.
The Afroalpine flora is rich in species of Alchemilla; Alchemilla orbiculata
of the Colombia paramos is shown in figure 14.15, bottom left. Hypochaerts
sessiliflora with its short peduncles (fig. 14.15, bottom left) is a parallel
to the similarly sessile heads of many other alpine Asteraceae, such as
Haplocarpha (fig. 14.7, bottom left) in alpine Africa. Gunnera magellamica
(fig. 14.15, bottom right) reminds us that the low, prostrate herbaceous
elements characteristic of low elevations in subantarctic latitudes occur
in moist localities at high elevations in the equatorial Andes.

Strikingly woolly herbaceous plants in the paramo include the justly

584

Errata: pages 585, 586, and 589

The photographs above the legends to figures
14.14, 14.15, and 14.17 have inadyertently been
transposed. The photographs shown above 14.14
should be above 14.15; the photographs shown
above 14.15 should be above 14.17; the photographs
shown above 14.17 should be above 14.14.

a a, APs ~
1% y i att

a
spe? "Ge

boyacensis; Paramo de Chisica, 3430 m, Cundinamarca, Eastern Cordillera, Colombia.
FT, acaulescent habit of plant. rop ricuT, inflorescence, showing nutant head. BorroM
ecio niveo-aureus, leaves and inflorescence; paramo at 3700 m, Sierra Nevada del
stern Cordillera, Colombia. porrom ricut, Senecio rigidifolius, branches of shrub,

of Paramo del Jabon, 3300 m, Trujillo, Venezuela. (All photographs by José

585

YH,

ee
Y),
ly,

ye ae
G. -oo*

mgium

Ery
s in Calama-

ombia. TOP LEFT,

l

aramos of Co

c

of the p

entiana sedifolia (

Paramo d

S

nt

compone

S

erbaceou

€ae), €

14.15. H
\pia
assland;

IG.
humil
grosti

F

Sl

ron ecuddorien

2
ge

and Eri

anacede€),

Genti

c

(J

e
gr

, Geranium

lera. TOP RIGHT

l

al Cordi

entr

C

50 m, Valle,

24

gan, 2

Barra

, €

s

continued)

(

86

Equatorial Highland Biota

famous Lupinus alopecuroides (fig. 14.16, top left), a species of the super-
paramo (above 4400 m, usually). Other superparamo elements shown here
include Senecio canescens subsp. boyacensis (fig. 14.14, top left, nght)
and Senecio niveo-aureus (fig. 14.14, bottom left). ‘The woolliness in these
species can be interpreted as insulation, because superparamo regions are
subject to frost. Woolliness also characterizes some cushion plants.

Cushion Plants. For reasons cited earlier in connection with the
Afroalpine flora, cushion plants can be expected in the equatorial Andes
(see also the discussion of cushion plants in New Zealand, chapter 7). A
magnificent example is Raouliopsis (fig. 14.17, top left, nght). This is such
a superb mimic of Raoulia eximia of New Zealand (see Carlquist, 1965)
that it is an outstanding example of parallel evolution; it belongs to the
same tribe of Asteraceae as does Raoulia (Inuleae). Other cushion-forming
plants in the paramo flora include some species of Arabis (Brassicaceae),
as shown in figure 14.17, top left; Azorella (Apiaceae), notable for its
“woody” cushions; Paepalanthus (Eriocaulaceae); Plantago rigida (Plan-
taginaceae), which forms dense cushions (Cuatrecasas, 1968); Distichia
tolimensis (Juncaceae); Oreobolus (Cyperaceae); and various species of
Mniodes (Asteraceae, tribe Inuleae), a cushion-plant genus that strikingly
parallels Haastia of New Zealand (see monograph of Mniodes by Cuatre-
casas, 1954a). Cushion plants in the equatorial Andes tend to be limited
to higher elevations, the superparamo. An interesting footnote is provided
by the cushion plants that have been developed in high elevations of
Bolivia by the family Isoétes: I. andicola and I. gemmifera, formerly segre-
gated as a separate genus, Stylites (Kubitzki and Borchert, 1964; Rauh and
Falk, 1959). The usually unbranched genus Isoétes has, in the “Stylites”
species, developed special rhizomorph modifications related to branching
and the cushion-plant habit.

Shrubby Components of the Paramo. In the paramo proper, some
shrubs tend to have narrow leaves and thus exemplify microphylly; this is
shown well by Hypericum laricifolium of the Hypericaceae (fig. 14.18, top

FIG. 14.15 (continued)

multiceps (Geraniaceae) with Calamagrostis effusa; Paramo de Cruz Verde, 3450 m, Eastern
Cordillera. sortom Lert, Hypochoeris sessiliflora (Asteraceae) heads in a mat of Alchemilla
orbiculata (Rosaceae); Macizo de Bogota, 3000 m. sorrom ricutT, Gunnera magellanica
(Haloragaceae) in flower; Paramo de Las Vegas, 3700 m, Central Cordillera. (All photographs
by José Cuatrecasas.)

587

FIG. 14.16. Various high-elevation plants of Colombia. top Lert, Lupinus alopecuroides
(Fabaceae); paramo at 4300 m, Nevado de Santa Isz
devallia coriacea (Orchidaceae

Macizo de Bogota, 3400 m, Eastern Cordillera. BoTTOM LEFT,
Diplostephium phylicoides (Asteraceae

del, Central Cordillera. rop ricuT, Mas-

shrub; paramos near Bogota, Eastern Cordillera. Bor-
TOM RIGHT, Chuquiraga jussieut

: Asteraceae), heads of flowers and leaves; Paramo de Las
Vegas, 3700 m, Valle, Central Cordillera. (All photographs by José Cuatrecasas.)

588

ric. 14.17. Plants of the superparamo of Colombia. Top LEFT, RIGHT, Raouliopsis seifrizit
Asteraceae), a high-elevation cushion plant. rop LEFT, habit of cushions, with flowering plants
of Draba sancta-martae (Brassicaceae); slopes around Laguna Naboba, 4400 m, Sierra Nevada

(continued)

589

Equatorial Highland Biota

left), a common element in the Espeletia “woodland,” and by Aragoa
abietina of the Scrophulariaceae (fig. 14.18, bottom left). Relatively small,
condensed leaves of Chuquiraga jussieui (fig. 14.16, bottom nght) of
Asteraceae, tribe Mutisieae. Foliar narrowness due to recurved leaf margins
is shown by Diplostephium phylicoides (fig. 14.16, bottom left) of Astera-
ceae and Arcytophyllum caracasana (fig. 14.18, bottom mnght) of the
Rubiaceae. Leaves not notably narrow, but restricted in size and excep-
tionally thick, are shown by Senecio rigidifolius (fig. 14.14, bottom right),
which illustrates leaf condensation that could be termed sclerophylly as
well as microphylly.

Extreme microphylly is shown by superparamo shrubs. Loricaria
thuyoides of the Asteraceae, tribe Inuleae, is an outstanding example
(fig. 14.17, bottom left). This genus (Cuatrecasas, 1954b) can only be
described as cupressoid or juniperlike. The small size of the evergreen
leaves is a perfect mimic of those of Cupressaceae, a family lacking in
these high Andean regions. The “cupressoid” habit is suited to frost
resistance and minimization of transpiration and the effects of alpine light
conditions. While the gymnosperm family Podocarpaceae in alpine New
Zealand illustrates cupressoid stature, so do species of Hebe of the Scroph-
ulariaceae (see chapter 7). Not only does Loricaria suggest Cupressaceae
in habit, but its ultraspecialized wood (by abundance of vascular tracheids)
mimics, in a way, primitive conifer woods (Carlquist, 1961). It would be
interesting to know if this mimicry occurs in another microphyllous shrub
of the superparamo, Valeriana karstenti (fig. 14.17, bottom might). For
other examples of paramo microphyllous shrubs, chiefly Asteraceae, see
Cuatrecasas (1968).

Monocotyledons of the Paramo. Tussock species of grass (Calama-
grostis) and the cushion plants Distichia (Juncaceae) and Paepalanthus
(Eriocaulaceae) have been mentioned. One may cite the presence in
paramos of Cyperaceae (Oreobolus spp., Rhynchospora ruiziana, Carex

pinetorum [Cuatrecasas, 1934]). The presence in paramos of an orchid

FIG. 14.17 (continued)

de Santa Marta, Magdalena. top ricHT, surface of cushion of Raouliopsis, same locality.
BOTTOM LEFT, Loricaria thuyoides (Asteraceae) shrubs, in Festuca tolucensis grassland; slopes
of Volcan Puracé, +300 m, Central Cordillera. Borrom ricut, Valeriana karstenii (Valeria-
naceae); Laguna Naboba, 4400 m, Sierra Nevada de Santa Marta, Central Cordillera. (All
photographs by José Cuatrecasas.)

590

Y,

4 A) f
m 4
| ANG

Z Lp

Fic. 14.18. Paramo shrubs from Colombia, showing tendencies in reduction of leaf surface.

top LEFT, Hypericum laricifolium (Hypericaceae); 3900 m, Nevado de Tolima, Central Cor-

dillera. top ricHtT, Hypericum magniflorum; Paramo de Tama, 3200 m, Eastern Cordillera.

BOTTOM LEFT, Aragoa abietina (Scrophulariaceae); Macizo de Bogota, 3400 m, Eastern Cor-

lillera. Bottom ricut, Arcytophyllum caracasana (Rubiaceae); 3800 m, western slopes of
vado del Nuilo, Central Cordillera. (All photographs by José Cuatrecasas.)

591

Equatorial Highland Biota

genus usually epiphytic in moderately cool, neotropical uplands, Masde-
vallia, is interesting. Masdevallia coriacea (fig. 14.16, top right) shows
adaptation to the paramos in its thick, leathery leaves. The rosette habit
of monocotyledons preadapts them to the paramos, where the rosette habit
plainly has a selective advantage. In a moist paramo swale, one almost
has the equivalent of an epiphytic community at ground level. This may
be operative in the manner in which Puya of the Bromeliaceae (fig. 14.19)
has been able to enter the paramo habitat. Species of Puya range from
2450 m for P. hamata (fig. 14.19, bottom right); 3100 m for P. nitida
(fig. 14.19, top left), inhabitant of areas where the rosette tree Espeletia
grandiflora grows; 3700 m for P. trianae (fig. 14.19, bottom left); to 3800 m
for P. cuatrecasasu (fig. 14.19, top nght). Bromeliaceae can be called a
typically tropical family, and its entrance into paramo situations is an
exception to the predominance of temperate families and genera. The
elevations cited show that it does not enter superparamo, however.
Evidently ornithophily is possible at the paramo altitudes, for Puya is
characteristically a bird-pollinated plant. Of course, ornithophily occurs
in the Afroalpine flora in the case of Lobelia, as mentioned earlier. The
presence of conspicuous bracts in some paramo species of Puya (fig. 14.19,
bottom left, right) suggests a resemblance to the bracts that serve as
landing platforms for Nectarinia in its visits to Lobelia keniensis and the
vicarious species related to it (fig. 14.3). However, the function of bracts
in the paramo Puyas may not be precisely parallel; until pollination modes
in the paramos are studied, little but speculation can be offered. There is
a possibility that overarching bracts of P. trianae (fig. 14.19, bottom left)
and the very woolly bracts of P. hamata (fig. 14.19, bottom right) could
be mechanisms whereby the flowers—seemingly borne nearly horizontally
in at least some paramo species of Puya—could serve to prevent entry of
rainwater into flowers. The relatively low elevations at which P. hamata
grows do not seem to imply the value of woolliness in insulation, although
this possibility does exist.

Further Studies in the Paramos. Although much work of a taxonomic
nature has been done in the paramos, the need for work of a “natural
history” or ecological nature has been stressed in this section. Anatomical
studies on the various groups of plants are desirable. Analyses of other

592

Fic. 14.19. Species of Puya (Bromeliaceae) from paramos of Colombia. Top Lert, P. nitida,
habit in swale with Calamagrostis; paramos above Bogota, 3100 m, Eastern Cordillera. Top
RIGHT, P. cuatrecasasii; paramo at 3800 m, western slope of Nevado del Huila, Cauca, Central
Cordillera. porrom LEFT, P. trianae, inflorescence; Paramo de Chisaca, 3430 m, Eastern
Cordillera. sortom ricut, P. hamata, upper part of inflorescence, showing notable woolliness;
Paramo de Barragan, 2450 m, Valle, Central Cordillera. (All photographs by José Cuatrecasas.)

593

Equatorial Highland Biota

types are possible, too. For a view of birds of North Andean highlands as
an insular model, see Vuilleumier (1970).

The Guayana Highlands

The Guayana Highlands, older than the paramos, are also much lower:
1000 to 3000 m (Maguire, 1971). These flat-topped mountains, or “tepuis,”
can be said to have insular qualities because of their distinctive geology
and their almost insular distribution. Not surprisingly, they have a high
endemism. Maguire (1971) seems to regard them as a relict area, but his
case 1s poor: the tablelands of Guayana are, in fact, much like islands and
seem clearly to have acquired their flora by dispersal over distance, perhaps
by truly long-distance dispersal. To be sure, some relatively primitive
features are to be seen in certain Asteraceae there (Carlquist, 1957).
However, many of the Guayana Asteraceae are highly specialized. The
occurrence of high endemism and distinctive adaptive features not in
“primitive” families but in very specialized, “weedy,” upgrade families
such as Asteraceae (Maguire, 1956) is indicative. ‘That the tepuis are older
than volcanic islands seems clear, but their floras are only moderately
relictual in nature. The primitive families Maguire (1971) cites as charac-
teristic of the “Highland Area” of Guayana are, in fact, not really in the
highlands to an appreciable extent: they are chiefly on the low and middle
elevations below the steep cliffs, not on the summit areas of the tepuis.
Maguire cites Ocotea and Persea as examples of how such a primitive
family as Lauraceae has survived in the Guayana Highlands, but these two
genera have reached Madeira and the Canary Islands by long-distance
dispersal. To be sure, these Lauraceae on Madeira and the Canaries rep-
resent an element that has now vanished in the Mediterranean mainland
with drying climates, but, as on all volcanic islands, their relictualism there
is only “recent relictualism.”

I note with interest the lack of conifers on the Guayana Highlands.
If these tepuis were truly the relict area of South America, one would
expect Araucariaceae to persist there. Instead, South American Araucarias
are restricted to the southern Brazilian highlands and to the southern

594

Equatorial Highland Biota

Andes. Likewise, the “primitive” taxa of Podocarpaceae, such as Saxe-
gothaea, persist in Chile despite recentness of the Andes. Among the
highly specialized families of angiosperms other than Asteraceae that show
high speciation and endemism on the Guayana Highlands are Cyperaceae,
Rapateaceae, Bromeliaceae, Myrtaceae, Melastomaceae, and Rubiaceae.
Interestingly, the families well represented on the Guayana Highlands are
well adapted for long-distance dispersal, as the listing above may suggest.
Even where a moderate degree of loss of dispersal can be hypothesized,
the tepuis have a flora reminiscent of oceanic island floras in dispersibility
of seeds and fruits. Of course, this dispersibility does not imply high levels
of vagility—if it did, one would not expect the high endemism which
occurs on the tepuis. The very unusual climatic and edaphic features of
the tepuis contribute to a high degree of endemism (probably preventing
easy establishment even if long-distance dispersal occurs frequently); these
tablelands can be considered a natural biotic province.

In effect, I am saying that the Guayana Highlands are much like an
archipelago of oceanic islands, and with many of the same features. That
their flora is somewhat older than that of most volcanic islands may be
hypothesized, but to a moderate degree only. In fact, the Guayana High-
lands would very much repay study as an example of an archipelago, and
I hope that biogeographers familiar with both continental and insular
patterns will devote analysis to this region, which is indubitably highly
interesting. For references to the Guayana flora, see Maguire (1971) and
Steyermark (1966, 1967).

Equatorial Highland Insects

Because a recent book (Mani, 1968) on the subject of high-altitude insects
has two chapters devoted to insects of equatorial highlands, there is no
need for an extensive review of this topic here. The flightlessness of high-
altitude insects is reviewed in his book (pp. 54-61). The question I wish
to discuss is whether flightlessness in equatorial highland insects is compa-
rable in evolutionary significance with that on tropical islands or temperate
mountains, as indicated by Darlington (1943), or whether equatorial high-
land insects are a special case. In chapter 12, the concepts of Darlington

595

Equatorial Highland Biota

were discussed, as was the possibility (see Gressitt’s conclusions cited there)
that flightlessness in subantarctic insects may form a special case.

One distinctive feature of the subantarctic environment is that not
merely the winter is extremely cold there, but midsummer temperatures
are quite cool also. On equatorial highlands, generally low temperatures
prevail, often below 0 C in some areas. Thus, exceptional cold, with no
prolonged periods of moderate warmth, occurs in both areas, albeit in
different modes. Rehn (1945) notes that in tropical and subtropical
climates, cockroaches of both sexes fly easily, but females cannot fly in
north temperate climates despite normal wing formation. Kramer (1958)
found that most cockroaches cannot fly at temperatures below 21 C.
Some of the work of Dewitz (1902, 1920) shows a correlation between
cold and loss of ability to fly in insects. Flightlessness and aptery may be
expected to occur in exceptionally cold localities, on the basis of such
results.

The effect of low atmospheric pressure and more particularly low
oxygen availability would seem to play an important role. The results of
Dewitz (1902, 1920) seem to show loss of flying ability under such circum-
stances. Although Darlington (1943) discounts the effect of wind pressure
on reduction of flight, the results of Digby (1958) would seem to indicate
that wind pressure cannot be wholly discounted. However, one can say that
if low temperature, low atmospheric pressure, and low oxygen availability
all tend to decrease flying ability markedly in equatorial highland situations,
the effect of increased winds might form a selective pressure in insects
for either unusually strong flight or loss of flight with consequent wing
reduction. The obstacles to any marked increase of flight efficiency in
such conditions seem overwhelming when one considers the combined
effect of these factors. The factors stressed by Darlington (geophily,
advantage of flightlessness in stable communities of small extent) apply
also to equatorial highlands, to be sure. The treelessness characteristic of
these highlands makes geophily a necessity, and of course mountains
provide areas of limited extent, more limited in area at each higher belt.

Winglessness is evident in Afroalpine insects, as Salt (1954) noted.
Parasphenus pulchripes (fig. 14.20) is a completely wingless orthopteran
from above 4500 m on Mt. Kilimanjaro. Orders other than Coleoptera
(which are usually sensitive indicators of flightlessness, and demonstrate

596

Equatorial Highland Biota

/

FIG. 14.20. Parasphenus pulchripes, a flightless orthopteran from the higher alpine zone of
Mt. Kilimanjaro, Tanzania.

the tendency more easily than other insect groups) seem unusually subject
to loss of flight on African peaks. In the northwest Himalayas, for example,
70 percent of Orthopera are apterous, and an additional 25 percent are
flightless (Mani, 1968). Also from above 4000 m on Kilimanjaro, one finds
a flightless tipulid, Tipula subaptera (fig. 14.21). As noted, if any insects
become flightless, we would expect them to be Coleoptera. Above 4000 m
on Kilimanjaro, one finds the apterous beetle Plocamotrechus kilimanus
(fig. 14.22). Other flightless Coleoptera from these elevations on Kiliman-
jaro include Peryphus sjoestedti (micropterous) and Parasystatiella agrestis
(elytra fused). Interestingly, flightless Lepidoptera also occur high on
Kilimanjaro: Saltia acrophylax is brachypterous. Occurrence of flightless

597

Equatorial Highland Biota

1
KV

&
SS
Ss
™iypy))

Nik

| iv) —a =
= MN))))) yas

Z 7M KY) ZZ
Mj,“ NULZZ fll

Ct
ae ‘ ty Mila tw Se EF — =
Zz LY Diyiy ly, ce
gz LZ Lp MY,

Fl

jaro, Tanzania.

lepidopterans on equatorial highlands recalls the presence of numerous
flightless moths on subantarctic islands (chapter 12). We can say, in
conclusion, that the aggravations of severe climatic conditions both on
equatorial highlands and on subantarctic islands seem overriding considera-
tions; at the very least, they are additive to the reasons cited by Darlington
(1943) for flightlessness in carabids.

With the prominence of vicarious plant species on mountains within
East Africa, Malesia, and the Andean paramos, one might expect similar
occurrence of vicarious species among insects. This is certainly true in the
Afroalpine region, as Mani (1968, pp. 170-71) clearly indicates.

598

Equatorial Highland Biota

Other phenomena in equatorial highland insects seem related to cold.
Mani (1968) relates melanism to increased receptivity to solar radiation,
thus minimizing the effect of cold. This phenomenon, however, is typical
of high-altitude insects in general, not just those of equatorial highlands.
Hedberg (1964) reports that insects tend to seek the shelter of the thick
accumulation of leaves on trunks of the Dendrosenecios, and attributes
this to the same insulation effect that, as his experiments show, keeps
temperatures above the freezing level within this leafy mass.

Although many facts about evolution on equatorial highlands can be
obtained from short visits and study, special efforts to study them will
undoubtedly reveal much of an ecological, physiological, and behavioral
nature, and perhaps some of the speculative comments and vague infer-
ences that one is forced to use in speaking of the curious characteristics of
these plants and animals can be superseded by informative studies.

ti =/TSANE
il! (pull ¢
es OO TT

/_) ES

pep z EEE iy,
EE Le LG res ZZ Wi Tite EME. Ta Cle

Fic. 14.22. Plocamotrechus kilimanus, a flightless micropterous beetle from the high alpine
zone of Mt. Kilimanjaro, Tanzania.

599

Equatorial Highland Biota

References

Carlquist, S. 1957. Anatomy of Guayana Mutisieae. Mem. N.Y. Bot. Gard. 9:441-
76.

Carlquist, S. 1961. Wood anatomy of Inuleae (Compositae). Aliso 5:21-37.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1970. Hawaii, a natural history. New York, Natural History Press.

Coe, M. J. 1967. The ecology of the alpine zone of Mount Kenya. The Hague, Dr.
W. Junk, Publishers.

Cuatrecasas, J. 1934. Observaciones geobotanicas en Colombia. Trab. Mus. Nac.
Cienc. Natur. (Madrid), ser. Bot., 27:1-144.

Cuatrecasas, J. 1954a. El género Mniodes. Folia Biol. Andina 1:1-7.

Cuatrecasas, J. 1954b. Synopsis der Gattung Loricaria Wedd. Feddes Repert. 56:
150-72.

Cuatrecasas, J. 1954c. Distribution of the genus Espeletia. Proc. Eighth Int. Bot.
Congr. (Paris), sect. 4:121-32.

Cuatrecasas, J. 1957. A sketch of the vegetation of the north-Andean province. Proc.
Eighth Pacific Sci. Congr. 4:167-—73.

Cuatrecasas, J. 1968. Geo-ecology of the mountainous regions of the tropical Ameri-
cas. Colloq. Geogr. 9:163-86.

Dale, I. R., and P. J. Greenway. 1961. Kenya trees and shrubs. Nairobi, Buchanan’s
Kenya Estates.

Darlington, P. J., Jr. 1943. Carabidae of mountains and islands: Data on the evolution
of isolated faunas and on atrophy of wings. Ecol. Monogr. 13:37-61.

Dewitz, J. 1902. Der Apterismus bei Insekten, seine kiinstlichen Erzeugung und
seine physiologische Erklarung. Arch. Anat. Physiol. (Abt. Physiol.) (1902): 61-67.

Dewitz, J. 1920. Uber die Enstehung rudimentiarer Organe bei den Tieren. 4. Die
Beeinflussung der Fliigelbildung bei Insekten durch Kalte und Blausaduregase. Zool.
Jahrb., Allg. Zool. 37:305—12.

Digby, P. S. B. 1958. Flight activity in the blowfly Calliphora erythrocephala in rela-
tion to wind speed, with special reference to adaptation. Jour. Exp. Biol. 35:776-95.

Faber, C. F. von. 1927. Die Kraterpflanzen Javas. ’s-Lands Plantentuin (Buitenzorg,
Java: Arb. Treub Lab.) 1:1-119.

Hauman, L. 1933. Les “Lobelia” géants des montagnes du Congo belge. Mém. Inst.
Colon. Belge 2:1-52.

Hauman, L. 1935. Les “Senecio” arborescents du Congo. Rev. Zool. Bot. Afr. 38(1):
1-76.

Hedberg, O. 1951. Vegetation belts of the East African mountains. Svensk Bot.
Tidskr. 45:140—202.

Hedberg, O. 1957. Afroalpine vascular plants. A taxonomic revision. Symb. Bot.
Upsal. 15(1):1-411.

Hedberg, O. 1964. Features of Afroalpine plant ecology. Acta Phytogeogr. Suec.
49 :1-144.

600

Equatorial Highland Biota

Hedberg, O. 1965. Afroalpine flora elements. Webbia 19:519-29.

Hedberg, O. 1968. Taxonomic and ecological studies on the Afroalpine flora of Mt.
Kenya. Hochgebirgforsch. 1:171-94.

Hedberg, O. 1969a. Growth rate of the East African giant Senecios. Nature 222:
163-64.

Hedberg, O. 1969b. Evolution and speciation in a tropical high mountain flora. Biol.
Jour. Linnaean Soc. 1:135-48. (Also as Speciation in tropical environments, ed.
R. H. Lowe-McConnell. London, Academic Press.)

Hedberg, O. 1971. Evolution of the Afroalpine flora. In Adaptive aspects of insular
evolution, ed. W. L. Stern, pp. 16-23. Pullman, Wash., Washington State Uni-
versity Press.

Koster, J. T. 1966. The Compositae of New Guinea. I. Nova Guinea, Bot. 24:497-
614.

Kramer, S. 1958. Pigmentation in the thoracic musculature of cockroaches and related
Orthoptera and the analysis of flight and stridulation. Proc. Tenth Int. Congr.
Entomol. 1:569-79.

Kubitzki, K., and R. Borchert. 1964. Morphologische Studien an Isoetes triquetra
A. Braun und Bemerkungen iiber das Verhaltnis der Gattung Stylites E. Amstutz
zur Gattung Isoetes L. Ber. Deutsch. Bot. Ges. 77:227—233.

Maguire, B. 1956. Distribution, endemicity, and evolution patterns among Com-
positae of the Guayana Highland of Venezuela. Proc. Amer. Phil. Soc. 100:467-75.

Maguire, B. 1971. On the flora of the Guayana Highland. In Adaptive aspects of
insular evolution, ed. W. L. Stern, pp. 63-78. Pullman, Wash., Washington State
University Press.

Mani, M. S. 1968. Ecology and biogeography of high altitude insects. The Hague,
Dr. W. Junk, Publishers.

Meijer, W. 1965. A botanical guide to the flora of Mt. Kinabalu. In Symposium on
ecological research in humid tropics vegetation, pp. 325-64. Kuching, Sarawak,
Government of Sarawak and UNESCO Science Cooperation Office for Southeast
Asia.

Rauh, W., and H. Falk. 1959. Stylites E. Amstutz, eine neue Isoetacee aus den
Hochanden Perus. Sitzungsber. Heidelberg. Akad. Wiss. (1959):1-160.

Rehn, J. A. 1945. Man’s uninvited fellow-traveller—the cockroach. Sci. Monthly
61:265-76.

Rock, J. F. 1920. The forest of Mt. Gedeh, west Java. Hawaiian Planter’s Rec. 22:67—
104.

Royen, P. van. 1964a. Sertulum Papuanum 10. Gentianaceae. Nova Guinea, Bot.
17:369-416.

Royen, P. van. 1964b. Sertulum Papuanum 11. Plantaginaceae. Nova Guinea, Bot.
18:417-26.

Salt, G. 1954. A contribution to the ecology of upper Kilimanjaro. Jour. Ecol. 42:
265-423.

Smith, A. C., and M. F. Koch. 1935. The genus Espeletia: A study in phylogenetic
taxonomy. Brittonia 1:479—530.

601

Equatorial Highland Biota

Stapf, O. 1894. On the flora of Mount Kinabalu in North Borneo. Trans. Linnaean
Soc. London, ser. 2, Bot. 4:69—263.

Steyermark, J. A. 1966. Contribuciones a la flora de Venezuela. Parte 5. Acta Bot.
Venez. 1:9-256.

Steyermark, J. A. 1967. Flora del Auyan-Tepui. Acta Bot. Venez. 2:5-368.

Vuilleumier, F. 1970. Insular biogeography in continental regions. I. The northern
Andes of South America. Amer. Natur. 104:373-88.

602

SELECTED ISLAND FOPRICS

Gigantism in Animals

The phenomenon of gigantism has been noted repeatedly for birds, rep-
tiles, and insects of islands. In birds, increase in size is most conspicuous
in those that are flightless. Adoption of an irreversible ground-feeding habit
by flightless birds seems an entry into a relatively unlimited ecological situ-
ation. The limits may well lie in problems of egg size as well as other limits
of a physical and physiological nature that tend to preclude progress be-
yond a certain point in body size (see De Beer, 1956). Among normally
winged birds, smaller body size is typical on low-latitude islands, whereas
large body size characterizes birds of higher latitudes (Grant, 1968). ‘This
variation is interpreted by Grant as “physiological adjustments to energy
sources and losses.”

Large body size is characteristic of rodents living on islands and
mountaintops in Europe (Zimmermann, 1950). Grant (1965a) considers
this a pathway to “facilitate a greater exploitation of environmental re-
sources.” Grant also thinks that such gigantism is possible on islands (and
islandlike areas) because of absence of predators.

In reptiles, gigantism has long been noted, especially in island lizards,
but in other insular reptiles as well. The summary of Mertens (1934), upon
which I drew (1965) for my account of this phenomenon, contains the
widest variety of examples. One could also cite the work of Schuster (1950)
and of Bole, Brelih, and Zei (1961). Since the review of Mertens, many
contributions have compared mainland lizards with those on offshore
islands or islets. Notable among these are the studies of Kramer and Mer-
tens (1938), Kramer (1946, 1951), and Mertens (1961). Kramer (1961) de-

603

Selected Island Topics

scribes situations on offshore islands of Italy. Radovanovi¢ (1956) studied
lizards on offshore islands of Yugoslavia. His results show larger body size
for insular populations of Lacerta melisellensis, but for Lacerta sicula (in
which Kramer found gigantism in the Italian islands), Radovanovi¢ gener-
ally finds smaller body sizes in the island populations compared with the
mainland populations. However, Radovanovic notes exceptions to these
trends in both species. This result is not surprising, because he studied
lizards of such a large number of islets. More recently, we have the study
of Soulé (1966) on lizards of islands off Baja California (almost exclusively
those in the Gulf of California).

Soule’s study (1966) is of particular interest because he has compared,
for insular and mainland populations of lizards, such dimensions as body
size, island size, population density, and size of scales. Because his data
were broadly based and because correlations were sought, he could present
some meaningful conclusions. Soulé found an inverse relationship between
island size and lizard size: larger lizards on smaller islands. The reason for
this correlation is not fully evident. There is, to be sure, also a relationship
between island size and number of resident species, according to Souleé’s
data. Soulé (1966) cites greater longevity for the island lizards; this agrees
with the data of Kramer (1946) (see also chapter 13). Is absence of preda-
tion the reason for greater size and longevity of the insular populations?
We do not have predation data on the Gulf of California islands, but low-
ered predation does seem to have been demonstrated sufficiently in the
case of the offshore Italian islands. Lack of competitors on smaller islands
may not be a factor. If a given food supply is available on an island and
only a single species can take advantage of it, presumably population size
will increase to the limits of that food supply. Would there be a concomi-
tant increase in body size? Soulé (1966) tends to believe that an “ecological
release’ by lack of competitors may well lead to an “evolutionary release”
into larger body size. There is a possibility that increased size might be of
advantage in mating battles, but this should be true on any area. Although
correlations seem fairly clear, the actual pathways in evolution of larger
body size in insular lizards have not been demonstrated. Such further in-
vestigations should take into account those exceptional instances in which
insular lizards are smaller than mainland counterparts; instances have
been reported by Mertens (1934) and Radovanovié (1956).

604

Selected Island Topics

There seems to be a relationship between body size and scale size in
the Gulf of California insular lizards (Soule, 1966). The smaller the island,
the smaller and more numerous the scales per unit area. Soulé finds that
larger scales tend to be strongly keeled and markedly overlapping; these
features, he claims, radiate excess heat and decrease tendencies toward
overheating. The smaller scales of lizards from small islands are smooth and
less overlapping, as opposed to those of mainland lizards. This contrast
assumes that climates of small islands are cooler and overheating would be
less likely to occur. The widespread phenomenon of melanism in insular
lizards (Mertens, 1934, 1961; Radovanovic, 1956) seems to be correlated
with this. Darker colors would seem to promote heat absorption where
island conditions are cooler and foggier than on the mainland. Larger body
size would also make more efficient heat absorption advantageous. How-
ever, sunny tropical islands can heat to relatively high levels during the
afternoon. The lizards’ response seems to be a mid-day retreat (Soule, 1968,
and the literature cited therein).

An interesting although not easily explained change in body charac-
teristics on the Gulf of California islands is the tendency for lizards to
develop greater degrees of asymmetry on small islands than on large islands
or mainland areas (Soulé, 1967).

Gigantism in insects often seems to occur on islands. A number of
examples could be cited, including the carabid Scarites iti Puerto Rico
(Darlington, 1971), drosophilids in Hawaii (Spieth, 1968), the fossil giant
earwig of St. Helena Island, Labidura loveridgei, the wetas of New Zealand
(see Tillyard, 1926), and the various examples from Madeiran Coleoptera
given by Wollaston (mentioned in chapter 12), Many more could be cited,
especially the gigantic representatives of various groups in Australia: Peta-
lura gigantea (Odonata: Petaluridae); Macropanesthia rhinoceros and
other Blattidae; Palophus titan, Acrophylla tessellata, Podacanthus wilkin-
som and other phasmids; various giant moths (Hepialidae), and many
others (Tillyard, 1926; Department of Entomology, 1970).

The reasons for insect gigantism may be several. Schoener and Janzen
(1968) cite larger body size as better adapted to coping with predators;
however, one would expect predator pressure to be less on at least some
insular areas. Another factor cited by Schoener and Janzen seems to apply
more generally: slower development. The equable climatic conditions of

605

Selected Island Topics

islands would appear to permit longer periods of development compared
with sharply seasonal continental conditions. New Zealand wetas are
notable for their longevity (see chapter 13). In the case of the giant flight-
less Scarites on Puerto Rico, which also have more poorly developed eyes,
Darlington (1971) cites ecological shift into wet forest as a leading factor.
Hawaiian drosophilids have notably large size; Spieth (1968) has cited
mating battles as a possible factor whereby larger size would be advan-
tageous.

Obviously, causes of insect gigantism on islands may be many. Any
particular case may be explained on the basis of factors, or a syndrome of
factors, that may not apply in other instances. This would seem to under-
line the need for fieldwork.

Many interesting changes in body shape occur in insular insects.
Darlington (1971) has cited the Colpodes bromeliarum group of carabids
on Jamaica, adapted to living in the sheathing leaf bases of bromeliad
rosettes by virtue of their flattened form. A precisely analogous develop-
ment has occurred in the delphacid leafhoppers of the Hawaiian Islands.
Of 139 native species, only one, Nesodryas freycinetiae, has a flattened
nymph. This is an adaptation to living in the closely fitting leaf bases of
leaf rosettes of Freycinetia (Zimmerman, 1948).

Changes in form, color, and habits make the Australian Cerambycidae
(Coleoptera) remarkable mimics of such diverse insects as creid bugs,
thynnid wasps, and even ants (Tillyard, 1926). Other examples of so-called
Batesian mimicry occur in the Australian insect fauna (Department of
Entomology, 1970). In addition, insects that mimic habitat are common
in Australia: many Hemiptera, larvae of the butterflies Papilio aegeus and
Eucyclodes spp., the grasshopper Acripeza reticulata, and the “leaf-insects”
(Phyludae of Phasmatodea) (Department of Entomology, 1970).

One possible reason for gigantism of insects on oceanic islands is that
it “reestablishes” a more normal distribution of sizes, comparable to those
on mainland areas. The work of Gressitt and his co-workers on insect dis-
persal in the Pacific (chapter 1) shows that immigrants to islands tend to
be relatively small in body size, because small size is a requisite for dis-
persal. During evolution of the immigrants that do establish, one might
expect that larger-sized descendants might evolve, forming a size range

606

Selected Island Topics

more nearly like that on continents as they occupy the niches that are
available.

This reasoning might also apply to land molluscs. “Gigantism’’—but
also “dwarfism’”—has been noted for insular gastropods by Bole, Brelih,
and Zei (1961). One can point to numerous examples. The amastrid and
achatinellid snails of the Hawaiian Islands provide clear instances. Ama-
strinae is an endemic subfamily of Cochlicopidae; Achatinellinae is an
endemic subfamily of Tornatellinidae. The widespread members of both
of these families have very small shell size compared with these two en-
demic Hawaiian subfamilies (Zimmerman, 1948). Within amastrids,
trends toward gigantism can be seen in the case of the genus Carelia, as
well as within the genus Amastra itself (e.g., Amastra knudseni; for illus-
trations, see Carlquist [1970]). Among New Caledonian land molluscs,
similar cases can be mentioned. Solem’s (1961) data suggest gigantism in
Paryphantidae (Ougapia raynali vs. O. apoana), and within Microphyura
(Paryphantidae) and Andrefrancia (Endodontidae). The famous Placosty-
lus (Bulimulidae) almost certainly represents gigantism both as a genus
and within the New Caledonian species.

Changes in Appendages of Animals

Although the birds of the Tres Marias Islands (Mexico) are relatively small
in body size, they have larger bills and tarsi than do their equivalents on
the mainland (Grant, 1965a, 1971). According to Grant, the Tres Marias
Islands lack mainland species with requirements similar to those of species
resident on the Islands. Therefore, the species on the Tres Marias may be
said to have extended their adaptations into new niches or even habits.
The larger bill sizes enable these birds to take food of a greater variety of
sizes, and the larger tarsus would enable a greater variety of perches to be
utilized. This phenomenon is found on other islands as well. Larger bills
(compared with mainland relatives) have been reported for insular birds
of North America by Murphy (1938), for the Canary Islands by Vols¢e
(1955), for the Gulf of Guinea islands by Amadon (1953) and for the
Philippines by Rand and Rabor (1960).

In instances of adaptive radiation on islands, bill size tends to show

607

Selected Island Topics

marked differences, as in the Darwin’s finches (chapter 6) or the Hawaiian
honeycreepers (chapter 4). Figure 4.5 shows divergence between con-
generic species of Loxops on the islands of Hawaii and Kauai. These pairs
“avoid” competition by having strongly marked divergences in food prefer-
ences. In the case of Loxops, one can consider this as a secondary (or
tertiary) cycle of adaptive radiation—a sort of module of how adaptive
radiation operates.

However, many islands have congeneric species that do not represent
parts of programs of adaptive radiation. Grant (1968) gives examples of
such pairs on various islands. ‘hese represent double invasions, as hypothe-
sized for Loxops. Grant’s (1968) figures show that the two sympatric spe-
cies of particular genera on various islands differ markedly in bill length
and wing length. Such differences are undoubtedly enforced strongly
where a finite food supply is partitioned by two species. This is precisely
what we would expect, and various factors contribute to this “enforce-
ment.” For example, population densities of animals tend to be greater
on islands than on continental areas (Lack and Southern, 1949; Crowell,
1962; Grant, 1965a). That in most cases both wing and bill sizes are
affected is interesting. The percentile difference in wing size between the
congeneric pairs is, however, much smaller than the difference in bill size
in virtually all instances. This suggests that bill-morphology differences
are more critical in preventing ecological overlap than are wing sizes—as
one might expect.

Differences in bill length between species of particular bird genera on
islands are much greater, in general, than differences in bill length be-
tween species of the same bird genera on mainland areas (Schoener,
1965; Grant, 1968). The nature and abundance of specific food items ap-
pears to govern the nature of divergence of bills between congeneric
species.

An example of how an unusually large bill may develop as an adapta-
tion to a new food source is shown by the finch Fringilla teydea of Gran
Canaria and Tenerife, Canary Islands. This bird has apparently evolved a
larger beak in relation to feeding upon the rather large seeds of the Canary
pine (Pinus canariensis), endemic to the Canary Islands (Volsge, 1951).

Although differences in bills seem related primarily to differences in
articles of diet, they may also serve a subsidiary function. Lack (1947)

608

Selected Island Topics

showed in the Darwin’s finches that an individual can recognize a member
of his own species (in the form of a stuffed specimen) by bill characteris-
tics; plumage differences in this group are probably too minor to serve
for recognition purposes.

In lizards, the genus Brachymeles (Scincidae) of the Philippines is
notable for presenting a series in limb reduction (illustrated in Carlquist,
1965). This has been documented in detail by Brown and Rabor (1967).
The more highly evolved species, in which reduction of limbs is great and
some digits are lost, have a more limited distribution within the Philip-
pines. They occupy only one or two, or a close group, of islands within the
larger group. Differences in habits among the species have not been docu-
mented well, so the significance of degrees of loss of limbs is difficult to
establish with any certainty. Clearly Brachymeles is limited to forest habi-
tats and is secretive within them. A sort of precinctiveness, reinforced by
loss of limbs, may be contributing to further speciation within the genus.

Changes in body proportions in the lizard Lacerta on Italian offshore
islands have a different explanation (Kramer, 1951). The island races have
shorter, fatter appendages and shorter tails. Evidently these represent
reductions in agility that have developed in the absence of predators. Also,
there is the possibility that with small islet size, efficiency in locomotion
is no longer advantageous. If gigantism occurs in insular reptiles as com-
pared with mainland ones, changes in proportions of limbs are often in-
volved (Mertens, 1934).

Sexual dimorphism in structural characteristics is not unusual among
animals at large, but there are some instances known in which sexual
dimorphism occurs only in insular representatives of the group. The huia
(Heterolocha acutirostris) of New Zealand, now extinct, had dimorphism
in bills related to cooperative feeding of a pair; several published reports
(see Oliver, 1955) suggest the validity of this interpretation. Phillipps
(1963) claims that this cooperative feeding tended to insure greater intake
by females, a fact he correlates with greater need for nutrition by females
because of egg-production. Sexual dimorphism has also been noted in the
proterhinid beetles of the Hawaiian Islands. In this group, males have
heads that are weakly beaklike, but a pronounced beaklike extension is
present in females. The ecological significance of this difference is obscure.
Sexual dimorphism might, in theory, have the advantage of broadening

609

Selected Island Topics

the range of food articles taken (for a discussion, see Soulé and Stewart
[1970]), but sexual dimorphism in feeding apparatus is so widespread geo-
graphically that it cannot be said to be an insular phenomenon at all.

Changes in Coloration Patterns

The most obvious example of island-related color patterns is the Paradi-
saeidae, which are too well known and thoroughly explored to be resur-
veyed here. There are several much more subtle examples that deserve
mention. Notable among avian examples described recently is the account
of Grant (1965b), which deals with the phenomenon of notably dull plum-
age in birds of the Tres Marias Islands. Grant describes this as “a by-
product of genetic reorganization.” He envisions “simplified” plumage
patterns, a sort of retrograde tendency permitted by absence of particular
species from the Tres Marias Islands. Reinforcement of bright and distinc-
tive color patterns is characteristic of continental areas, where sympatry of
congeneric species makes recognition patterns of higher selective value. If
only one species of a genus is present on the Tres Marias, such distinctive-
ness no longer is of high selective value, and may be expected to be lost
incidentally.

The Darwin’s finches do show distinctions in color patterns (see fig.
6.2). These are, as described, proportionate to time spent feeding on black
lava surfaces versus time spent feeding in brownish trees. This suggests
coloration of a protective sort, and indeed predators, such as the Galapagos
hawk, are present.

The variety of color in plumage in the Hawaiian honeycreepers is diffi-
cult to explain. Why should the subfamily Drepanidinae exhibit blacks,
reds, and various mottled patterns of such distinction? Why, conversely,
does the subfamily Psittarostrinae have yellow green colors virtually iden-
tical throughout the highly distinctive genera that compose it? The great
variety in color patterns in the Hawaiian land shell Achatinella also is not
easily explained.

In contrast, there is a wealth of color patterns, evolved autochtho-
nously, in many different Hawaiian insect groups that can only be ex-
plained as protective colorings that match host plants or substrates in a

610

Selected Island Topics

remarkable way. Some of these are reviewed in chapter 4, for wide diver-
gences in substrate or host-plant mimicry are aspects of adaptive radiation.
I refer the reader to the resumé of Zimmerman (1948, pp. 142-43) where
the various insect groups that show protective coloration are described
engagingly.

Mertens (1934) and other European herpetologists have noted in-
creased melanism in insular reptiles. This melanism may provide screening
from a component of solar radiation or may be related to sea salts in the
diet, but is probably not directly related to thermoregulation. In any case,
melanism is characteristic not just of insular reptiles, but of those in supra-
tidal situations as well. For a review of this phenomenon, see Neill (1958).

Changes in Habits and Habitats

Hawaiian insects tend to show “‘supralimital” evolution in some respects.
Usinger (1942) indicated this for the Nysius species, and Zimmerman
(1948) cited a number of interesting examples. One of these that particu-
larly arouses Zimmerman’s enthusiasm is an odonatan, Megalagrion
oahuense. Unlike all other members of this order, it has nymphs adapted
not to aquatic existence but to foraging for prey on land. Likewise, bugs
of the genus Saldula are characteristically riverine and lacustrine through-
out the world. The Hawaiian Saldulas, however, are strictly arboreal. I
can imagine an explanation for such shifts from aquatic to terrestrial or
arboreal. Even though rainfall on the Hawaiian Islands is heavy, high
porosity of soils and rocks results in a notable absence of permanent
streams or lakes. Although enough aquatic habitats exist for establishment
of a typically aquatic group, radiation from this niche would almost have
to be terrestrial or arboreal, and there would be considerable selective
pressure in that direction. We must not make the mistake of considering
the floor of a Hawaiian rain forest as “dry land,” however. It consists of
almost perpetually wet litter or bryophytes. The same is true of tree trunks
in certain localities, on which the thick coverings of epiphytes are related
to frequent baths of mist, cloud, or rain, and which are constantly moist
(the occurrence of Hymenophyllaceae, a family of ferns intolerant to
desiccation, is a good indicator of moisture where these ferns are epiphytic

611

Selected Island Topics

on tree trunks). Such habitats would provide suitable zones for ecological
shift of an aquatic organism.

Likewise, the shift from terrestrial to arboreal habitats occurs in
Hawaiian insects. An exceptional number of Hawaiian Coleoptera are
arboreal rather than terrestrial (Darlington, 1943). Larvae of the family
of crane flies, Tipulidae, typically live in soil, decaying vegetation, or even
water. However, one Hawaiian species, Limonia foliocuniculator, is a leaf
miner and inhabits Cyrtandra of the Gesneriaceae (Swezey, 1915). At
least three Hawaiian species of the weevil genus Proterhinus have shifted
from boring in dead wood (the typical habit of the genus) to leaf-mining.
Of these, one can be found in leaves of. Astelia (Liliaceae), one in leaves
of Broussaisia (Saxifragaceae), and one in living stems of Broussaisia (Zim-
merman, 1948). How ecological shift of this sort might occur is evident on
the basis of my field experience in the Hawaiian Islands. Astelia is often
an epiphyte on rotting bark; Broussaisia grows in very wet forest, and, as
in many shrubs and trees of the Hawaiian forest, both rotten and living
portions may be found on a single shrub. Thus the ecotone that must be
crossed is not as drastic as one might imagine, although the end product is
rather remarkable.

The Odynerus wasps typically build mud nests in crevices or beetle
burrows, or tunnel into wood or soil. One Hawaiian species, O. oahuensis,
constructs entire independent mud cells (Williams, 1927).

Shifts in feeding habits of Hawaiian insects are seen in the drosophi-
lids cited in chapter 4; many of these utilize food sources entirely alien to
the typical slime-flux feeding habits of mainland drosophilids. One genus,
Titanochaeta, is parasitic on spider eggs. Shifts in feeding habits are seen
in Hawaiian hylaeid bees of the genus Nesoprosopis, some species of
which have developed commensal and semiparasitic modes of existence
(Zimmerman, 1948).

The Galapagos iguanas have become famous, although they do not
represent remarkable features except in a few ways. There is nothing par-
ticularly noteworthy in the herbivorous habits of the land iguana, Conolo-
phus, except that it is capable of consuming spiny fruits and stems of the
cactus Opuntia. The marine iguana, Amblyrhynchus, is much more re-
markable because of its various adaptations to life as a marine herbivore

612

Selected Island Topics

(Carpenter, 1966). In feeding upon algal beds, it is able to dive to approxi-
mately 11 m. The laterally compressed tail enables it to swim efficiently,
and it is capable of adjusting its buoyancy to different depths. The short
snout and flat, triseriate teeth enable it to graze algae efficiently from
rocks. Intake of saltwater during its feeding is inevitable, but it has com-
pensated for this by increased efficiency of its salt-extracting gland.

Behavioral features of Amblyrhynchus also show alteration (Carpen-
ter, 1966). Since it is a littoral animal of large size, congregating at feeding
sites is an efficient use of resources, but this runs counter to the territori-
ality typical of lizards. Consequently, territoriality in male marine iguanas
occurs only at mating time, and vanishes after the mating season has
waned. For females, territorial antagonism occurs during competition for
nesting sites in beaches, and disappears as soon as egg-laying has been
completed. An insular rodent, the Bahaman hutia (known only from East
Plana Cay), shows a notably low level of aggressiveness (Clough, 1969).
Clough regards this placidity as an adaptation related to high population
density.

Soulé (1966) mentions that the lizard Uta on San Pedro Martir Island
feeds on marine invertebrates to a large degree, so that flies are not the sole
article of subsistence, as in other Uta populations. Population density is
greater in the Utas of San Pedro Martir Island than on any other island
Soule examined. Thus we can see how a shift to a marine food source
might take place, with concomitant abandonment of territoriality except
during reproduction, as in Amblyrhynchus.

New Zealand skinks and geckos are notable for their habit of vivipary.
Only one New Zealand skink, Leiolopisma suteri, has been found to be
oviparous (Whitaker, 1968). New Zealand is not exceptional in this high
proportion of vivipary, however. It occurs at northern and southern limits
of lizard distribution elsewhere in the world. This adaptation, related to
intolerance of cool temperatures during egg incubation, permits extension
of lizards farther south in the southern hemisphere and farther north in
the northern hemisphere (Darlington, 1948). The single oviparous New
Zealand skink, Leiolopisma suteri, is not an exception to the rule, because
it is restricted to six islets off the Auckland-area coast of New Zealand; all
the islets are north of the 37th parallel. The viviparous habit has permitted
skinks to penetrate to 47°S. The two viviparous skinks there, Leiolopisma

613

Selected Island Topics

zelandica and L. lineo-ocellatum, may represent the southernmost lizard
populations in the world.

Analyses of Islandlike Situations

That areas or conditions other than true islands present insular aspects
does not need emphasis at this point. A number of biologists have en-
deavored to apply the considerations of MacArthur and Wilson (1967) to
these situations. We have, for example, the comparison by Janzen (1968)
of plant hosts and their insect parasites as an insular situation. Potentially
informative as this analogy is, analysis of host-parasite relationships is fore-
stalled, as Janzen notes, by lack of suitable data. Thus, although the
analogy is a compelling one, collection of data that would build toward
comparison with mathematical modes is a complicated and difficult task
yet to be performed. Brown (1971) has found that mammals of continental
mountaintops in the southwestern United States do not conform to an
insular situation. The species-area curve is much steeper than the Mac-
Arthur-Wilson curve for insular biotas. Brown finds that the rate of
immigration for mammals to these areas is “effectively zero.” He explains
the situation by hypothesizing that these mountains were colonized during
the Pleistocene, when intermountain areas did not present the climatic
barriers they do today. After post-Pleistocene isolation, diversity of faunas
was reduced. Because extinction-probability is inversely related to popula-
tion size, body size, diet, and habitat, extinction has been low. The present
residents of these mountaintops (which Brown literally terms “islands”)
are considered to be true relicts, beyond the MacArthur-Wilson equilibria
between colonization and extinction.

On the other hand, Vuilleumier’s (1970) analysis of northern Andean
regions (considered as island regions) does bear out the MacArthur-
Wilson equilibria. This is also true of cave arthropods, according to Culver
(1970), who uses Simberloff’s (1969) noninteractive model of island bio-
geography as the basis for his analysis. Those interested in caves as insular
situations will want to consult the extensive reviews of Vandel (1965) and
Barr (1968). Although cave animals are well known for such tendencies as
depigmentation, loss of eyes, and endemism, there are other less well

614

Selected Island Topics

known tendencies with insular overtones in some cave animals. The
tendencies toward flightlessness and aptery are strong. Gigantism charac-
terizes Many cave animals, especially Coleoptera. However, a tendency
toward dwarfism occurs in a few others. Biotic tendencies of equatorial
highlands have been presented in extended treatment in this book because,
if construed broadly (i.e., as more than as a single alpine zone), equatorial
highlands simulate many features of insular biotas. Caves, in comparison,
present rather restricted habitats, although obviously there are certain
ecological niches within them and caves do differ ecologically one from
another.

Extinction

If we look at the geography of extinct or endangered species (e.g., Green-
way, 1958; or the IL.U.C.N. Red Data Books), we quickly see that insular
species figure with exceptional prominence in the listings. In seeking the
reasons for this, small land area will appeal as a prime cause of extinction.
However, before we apply the MacArthur-Wilson curve to confirm that,
as expected, smaller land areas suffer higher rates of extinction, we must
remember that the roster of endangered and extinct species is not the
product of the factors MacArthur and Wilson cite, because these are man-
induced extinctions and near-extinctions, not “naturally occurring” ones
over indefinite periods of time. Small land area is clearly involved, for a
number of reasons. To be sure, population size is, in general, smaller, and
there are no scattered pockets, as there might be on a continent, where
survival might be possible. Adaptation to insular conditions seems irrevers-
ible, for various reasons, and the organisms must survive within the limited
habitat or habitats available to them or suffer extinction. Encroaching
influences—weeds, introduced pests, parasites, and predators—have an
effect geometrically greater on a small land area than a large one. Smaller
populations, in other words, are “easier prey.” One can cite the presumptive
genetic depauperation of a small endemic insular population, and it cannot -
be denied. One may express this as “loss of competitiveness” or say that
insular organisms have “limited evolutionary potential” (Simpson, 1953;
Darlington, 1948; W. L. Brown, 1957). Unfortunately we have no exact

615

Selected Island Topics

way of predicting vulnerability of any given species, or its ability to recover
if land usage is discontinued and an area is allowed to return to its
“natural” state. However, McDowall (1969) has shown that the higher
the level of endemism of an organism’s group (e.g., endemic family versus
endemic subspecies), the more prone an insular organism is to extinction.
The corollary seems evident, that the older the autochthone is, the more
prone it is to extinction. This, of course, applies most clearly on islands,
whereas more numerous exceptions can be cited on continental areas. The
subject of extinction on insular areas and the factors that contribute to it
has been formulated in detail by MacArthur (1972).

In “saving” insular species from extinction, we must discriminate
between short-term sanctuary and long-term survival. Breeding in zoos and
maintenance in botanic gardens so far give no indications of long-term
success on their own terms. Long-term success is possible only if sufficient
areas of the original habitat can be preserved, or if an area can be allowed
to “renaturalize” and an endangered organism can then be re-released into
it. These principles seem obvious, but practical demonstrations appear
necessary. For example, the Hawaiian nene (Branta sandvicensis) was
rescued from extinction and successfully bred in pens, but release into a
known former habitat (Haleakala, Maui) failed because that habitat was
no longer sufficiently intact. Those interested in prevention of extinction
on insular areas will have to make decisions as to whether their actions are
likely to lead to long-term survival, or only to short-term maintenance.

In the latter case, or in the case of organisms for which no protective
measures at all are being taken, biologists have only a single choice: obtain
as much data, preferably by field studies, as soon as possible. Studies such
as that of ‘T’omich (1971) on the Hawaiian crow—apparently now extinct—
are to be congratulated. Because increasingly rapid extinctions are probable,
the failure of biologists to study endangered insular organisms may justi-
flably be more criticized in the future than their failure to take part in
well-meaning but ill-fated efforts to save species.

References

Amadon, D. 1953. Avian systematics and evolution in the Gulf of Guinea. Bull.
Amer. Mus. Natur. Hist. 100:397—431.

616

Selected Island Topics

Barr, T. C., Jr. 1968. Cave ecology and the evolution of troglobites. In Evolutionary
biology, ed. T. Dobzhansky, M. K. Hecht, and W. C. Steere, vol. 2, pp. 35-102.
New York, Appleton-Century-Crofts.

Bole, J., S. Brelih, and M. Zei. 1961. Les pulmonés et les lézards insulaires et le
probléme de leur speciation dans l’archipel de Rovinj (Rovigno). In Le peuplement
des iles méditeranéenes et le probléme de Vinsularité. Paris, Colloques Interna-
tionaux du Centre National de la Recherche Scientifique.

Brown, J. H. 1971. Mammals on mountaintops: Nonequilibrium insular biogeogra-
phy. Amer. Natur. 105:467—78.

Brown, W. C., and D. S. Rabor. 1967. Review of the genus Brachymeles (Scincidae)
with descriptions of new species and subspecies. Proc. Calif. Acad. Sci., 4th ser.,
34(15):525-47.

Brown, W. L. 1957. Centrifugal speciation. Quart. Rev. Biol. 32:247-77.

Carlquist, S. 1965. Island life. New York, Natural History Press.

Carlquist, S. 1970. Hawaii, a natural history. New York, Natural History Press.

Carpenter, C. C. 1966. The marine iguana of the Galapagos Islands, its behavior and
ecology. Proc. Calif. Acad. Sci., 4th ser., 24(6):329-76.

Clough, G. C. 1969. The Bahaman hutia: A rodent refound. Oryx 10:106-8.

Crowell, K. L. 1962. Reduced interspecific competition among the birds of Bermuda.
Ecology 43:75-88.

Culver, D. C. 1970. Analysis of simple cave communities. I. Caves as islands. Evolu-
tion 24:463-74.

Darlington, P. J., Jr. 1943. Carabidae of mountains and islands: Data on the evolu-
tion of isolated faunas and on atrophy of wings. Ecol. Monogr. 13:37-61.

Darlington, P. J., Jr. 1948. The geographical distribution of the cold-blooded verte-
brates. Quart. Rey. Biol. 23:1—26, 106-23.

Darlington, P. J., Jr. 1971. Carabidae on tropical islands, especially the West Indies.
In Adaptive aspects of insular evolution, ed. W. L. Stern, pp. 7-15. Pullman,
Wash., Washington State University Press.

De Beer, F. R. 1956. The evolution of the ratites. Bull. Brit. Mus. (Natur. Hist.)
Zool., 4(2):1-63.

Department of Entomology, Commonwealth Scientific and Industrial Research Or-
ganization. 1970. The insects of Australia. Melbourne, Melbourne University Press.

Grant, P. R. 1965a. The adaptive significance of some size trends in island birds.
Evolution 19:255-67.

Grant, P. R. 1965b. Plumage and the evolution of birds on islands. Syst. Zool. 14:
47-52.

Grant, P. R. 1968. Bill size, body size, and the ecological adaptations of bird species
to competitive situations on islands. Syst. Zool. 17:319-33.

Grant, P. R. 1971. Variation in the tarsus length of birds in island and mainland
regions. Evolution 25:599-614.

Greenway, J. C., Jr. 1958. Extinct and vanishing birds of the world. New York,
American Committee for International Wild Life Protection.

Janzen, D. H. 196%. Host plants as islands in evolutionary and contemporary time.
Amer. Natur. 102:592-95.

617

Selected Island Topics

Kramer, G. 1946. Veraénderung von Nachkommenziffer und Nachkommengrésse
sowie der Altersverteilung von Inseleidechsen. Z. Naturforsch. 1:700—10.

Kramer, G. 1951. Body proportions of mainland and island lizards. Evolution 5:193-
206.

Kramer, G., and R. Mertens. 1938. Rassenbildung bei westistrianischen Inseleidechsen
in Abhangigkeit von Isolierungsalter und Arealgrésse. Arch. Naturgesch. (N.F.)
7:189-234.

Lack, D. 1947. Darwin’s finches. Cambridge, Cambridge University Press.

Lack, D., and H. N. Southern. 1949. Birds of Tenerife. Ibis 91:607—26.

MacArthur, R. H. 1972. Geographical ecology. New York, Harper and Row, Pub-
lishers.

MacArthur, R. H., and E. O. Wilson. 1967. The theory of island biogeography.
Princeton, N.J., Princeton University Press.

McDowall, R. M. 1969. Extinction and endemism in New Zealand land birds.
Tuatara 17:1-12.

Mertens, R. 1934. Die Insel-Reptilien, ihre Ausbreitung, Variation, und Artbildung.
Zoologica 32(6):1—209.

Mertens, R. 1961. Die Inseleidechsen des Golfes von Salerno. Senckenbergiana Biol.
42:31-40.

Murphy, R. C. 1938. The need of insular exploration as illustrated by birds. Science
88:533-39.

Neill, W. T. 1958. The occurrence of amphibians and reptiles in saltwater areas, and
a bibliography. Bull. Mar. Sci. Gulf Caribbean 8:1-97.

Oliver, W. R. B. 1955. New Zealand birds. 2d ed. Wellington, A. H. & A. W. Reed.

Phillipps, W. J. 1963. The book of the huia. Christchurch, Whitcombe and Tombs.

Radovanovicé, M. 1956. Rassenbildung bei den Eidechsen auf adriatischen Inseln.
Osterr. Akad. Wiss. (Math.-Naturwiss. Kl.) Denkschr. 110(2):1-82.

Rand, A. L., and D. S. Rabor. 1960. Birds of the Philippine Islands: Siquijor, Mount
Malindang, Bohol and Samar. Fieldiana, Zool. 35:223-441.

Schoener, T. W. 1965. The evolution of bill size differences among sympatric con-
generic species of birds. Evolution 19:189-213.

Schoener, T. W., and D. H. Janzen. 1968. Notes on environmental determinants of
tropical versus temperate insect size patterns. Amer. Natur. 102:207—24.

Schuster, O. 1950. Die klimaparallele Ausbildung der K6rperproportionen bei Poi-
kilothermen. Abhandl. Senckenberg. Naturforsch. Ges. 482:1-89.

Simberloff, D, S. 1969. Experimental zoogeography of islands. A mode of insular
colonization. Ecology 50:296-314.

Simpson, G. G. 1953. The major features of evolution. New York, Columbia Uni-
versity Press.

Solem, A. 1961. New Caledonian land and fresh-water snails. An annotated check
list. Fieldiana, Zool. 41:413—501.

Soule, M. 1966. Trends in the insular radiation of a lizard. Amer. Natur. 100:47-64.

Soulé, M. 1967. Phenetics of natural populations. I]. Asymmetry and evolution in a
lizard. Amer. Natur. 101:141-60.

618

Selected Island Topics

Soulé, M. 1968. Body temperatures of quiescent Sator grandaevus in nature. Copeia
(1969) :622—23.

Soulé, M., and B. R. Stewart. 1970. The “niche-variation” hypothesis: A test and
alternatives. Amer. Natur. 104:85-97.

Spieth, H. T. 1968. Evolutionary implications of sexual behavior in Drosophila. In
Evolutionary biology, ed. T. Dobzhansky, M. K. Hecht, and W. C. Steere, vol. 2,
pp. 157-99. New York, Appleton-Century-Crofts.

Swezey, O. H. 1915. A leaf-mining cranefly in Hawaii. Proc. Hawaiian Entomol. Soc.
3(2):87-89.

Tillyard, R. J. 1926. The insects of Australia and New Zealand. Sydney, Angus and
Robertson.

Tomich, P. O. 1971. Notes on nest and behavior of the Hawaiian crow. Pacific Sci.
25:465-74.

Usinger, R. L. 1942. The genus Nysius and its allies in the Hawaiian Islands. Bishop
Mus. Bull. 173:1-165.

Vandel, A. 1965. Biospeleology. The biology of cavernicolous animals. Transl. B. E.
Freeman. London, Pergamon Press.

Volsge, H. 1951. The breeding birds of the Canary Islands. I. Introduction and
synopsis of the species. Vidensk. Medd. Naturhist. Foren. Kjobenhavn 113:1-153.

Volsge, H. 1955. The breeding birds of the Canary Islands. II. Origin and history of
the Canarian avifauna. Vidensk. Medd. Naturhist. Foren. Kjobenhayn 117:117-78.

Vuilleumier, F. 1970. Insular biogeography in continental regions. 1. The northern
Andes of South America. Amer. Natur. 104:373-88.

Whitaker, A. H. 1968. Leiolopisma suteri (Boulenger), an oviparous skink in New
Zealand. New Zealand Jour. Sci. 11:425-32.

Williams, F. X. 1927. Notes on the habits of bees and wasps of the Hawaiian Islands.
Proc. Hawaiian Entomol. Soc. 6(3):425—64.

Zimmerman, E. C. 1948. Insects of Hawaii, vol. 1: Introduction. Honolulu, Univer-
sity of Hawaii Press.

Zimmermann, K. 1950. Die Randformen der Mitteleuropaischen Wiihlmaduse. In
Syllegomena Biologica Festschrift, ed. A. Jordans and F. Peus, pp. 454-71. Leipzig,
Kleinschmidt Verlag.

619

INDEX OF BIOLOGICAL NAMES

Numbers in italics refer to photographs.

Ablepharus boutoni, 8

Abortopetalum, 406

Abronia alpina, 484

Abutilon, 518; A. arboreum,
406, 424

Abyssocottus, 340

Abyssogammarus, 340

Acacia, 7, 28, 57, 172, 268,
283, 286, 289, 293, 323, 329,
466, 483, 517, 530; A. alata,
272; A. ataxiphylla, 272;
A. congesta, 276; A. cuneata,
276; A. diptera, 272; A. exo-
carpoides, 276; A. glaucop-
tera, 272, 273, 275; A. het-
erophylla, 465; A. inops,
276; A. kauaiensis, 465, 531;
A. koa, 130, 139, 164, 173,
465, 531; A. merinthifolia,
294; A. nodiflora, 276; A
restiacea, 307; A. simplici-
folia, 466; A. squamata, 307;
A. stenoptera, 272; A. tri-
gonophylla, 272; A. triptycha
var. subflexuosa, 293; A.
urophylla, 288; A. viscifolia,
293; A. volubilis, 276

Acaena, 31, 57, 63, 415, 518;
A. californica, 473, 475; A.
exigua, 473, 475

Acalypha, 204; A. macro-
stachya, 424; A. samydae-
folia, 424

Acanthisitta, 251
Acanthogammarus, 340
Acanthosicyos, 420
Acanthospermum, 455
Achatinella, 28-29, 123, 127-—

28, 176-77, 488, 610; A.

mustelina, 176
Achatinellidae, 127
Achatinellinae, 607
Achyranthes, 57, 517
Aciphylla, 247, 248

Acmopyle, 215; A. pancheri,
228

Acripeza reticulata, 606

Acronychia, 154, 476, 523

Acrospsilus borboroides, 497

Actinostrobus, 225, 226, 321;
A. acuminatus, 321

Actinotus, 297

Adenanthos cunninghamii,
286; A. meissneri, 293; A.
pungens, 293

Adenium sokotranum, 420

Adenocarpus, 31, 535

Adenostemma, 58, 517, 575

Adenostoma, 326; A. fasci-
culatum, 297

Adenothamnus validus, 147

Adiantaceae, 438

Aeonium, 191, 192, 195, 374,
Ala ATS: 53950903). Ae ar
boreum, 191, 192; A. “ar-
borescens,” 192; A. caes-
pitosum, 192; A. canariense,
192; A. cuneatum, 192; A.
glandulosum, 192; A. glu-
tinosum, 191; A. goochiae,
192; A. haworthii, 192;
A. holochrysum, 191, 192;
A. leucoblepharum, 191; A.
manriqueorum, 192; A.
nobile, 192; A. percarneum,
192; A. saundersii, 192; A.
sedifolium, 192; A. spathula-
tum, 192; A. tabulaeforme,
192; A. viscatum, 192

Aepus, 503

Aerva, 57, 517

Agathis, 215, 216, 220, 226,
233; A. australis, 234

Agrostis, 56, 516, 567

Aizoaceae, 57, 342

Alchemilla, 562, 568; A. argo-
phylla, 566, 567; A. orbicu-
lata, 584, 586

620

Alectryon, 57, 65, 68, 434, 478,
479, 480, 483, 518, 531; A.
carinatus, 479; A. excelsus,
479, 480; A. grandifolius,
479; A. macrococcus, 479,
480, 531; A. mahoe, 480,
531; A. samoensis, 479, 480;
A. subcinereum, 68, 479

Aleurites, 76

Allagopappus, 413

Alluaudia, 419; A. adscendens,
419; A. dumosa, 419

Alluaudiopsis, 419

Alphitonia, 57, 518, 531; A.
franguloides, 474; A. pon-
derosa, 474, 475, 531; A.
zizyphioides, 474

Alseuosmia, 409, 410

Alsinodendron, 57, 406, 435,
461, 517; A. trinerve, 461,
462

Alternanthera, 205; A.
ulosa, 205

Alyxia, 58, 441, 442, 517

Amaranthaceae, 57, 205, 257,
297, 298, 300, 404, 407, 413

Amastra, 128, 177, 607; A.
knudsenii, 128, 607

Amastridae (Cochlicopinae,
subfam. Amastrinae), 127-
28, 340, 607

Amblyrhynchus, 22, 612, 613
Amber 216; A. trichopoda,
22

Amborellaceae, 216, 228, 232

Amphoradenium, 157; A. hy-
menophylloides, 157; A. pin-
natifidum, 157

Anacardiaceae, 57

Anagallis, 562; A. serpens,
562

rug-

Anarthria scabra, 272
Anas aucklandica aucklandica,
489; A. a. nesiotis, 489

Anastraphia, 405

Andersonia, 294; A. echino-
cephala, 294, 302; A. grandi-
flora, 294; A. involucrata,
294

Andrefrancia, 607

Andropogon, 56, 516

Andryala, 413, 415; A. crithmi-
folia, 415; A. pinnatifida,
193, 413

Anigozanthos, 297, 299, 300

Anisophyllum, 463

Anoectochilus, 57, 516

Anolis, 47-49

Anonchotaenia brasiliense, 159

Antarctopria latigaster, 501;
A. 1. campbellana, 501

Antennariidae, 343, 345, 346

Antennarius, 344, 345; A. ber-
mudensis, 344, 345; A. mul-
tiocellatus, 344, 345; A. ocel-
latus, 344, 345; A. paucira-
diatus, 344, 345; A. sangui-
fluus, 344; A. tenebrosus,

, 345

Anthocercis, 299

Anthornis, 251

Anthotium, 310; A.
florum, 311

Antidesma, 57, 517, 524

Antipodencyrtus, 499; A. pro-
cellosus, 499, 500

Apataenus watsoni, 497, 498

Apetahia, 212, 384, 406, 408;
A. longistigmata, 407; A.
raiateensis, 407

Apiaceae, 35, 58, 63, 84, 230,
231, 247, 248, 270, 272, 297,
410, 415, 416, 420, 423, 587

Apocarpa, 154

Apocynaceae, 1
441

rubri-

2, 58, 342, 420,

Apodocephala, 419

Aquifoliaceae, 57

Arabis, 568, 570, 587

Aragoa abietina, 590, 591

Araliaceae, 29, 58, 73, 172,
226, 231, 243, 246, 248, 410,
442, 517

Araucaria, 215-17, 218, 220,
225, 226, 233, 234; A. arau-
cana, 217; A. balansae, 217;
A. bernieri, 217, 218; A.
biramulata, 217; A. cookii,
215-17, 218; A. humbold-
tensis, 217; A. montana, 217;
A. muelleri, 217, 218, 219;
A. rulei, 217, 218

Araucariaceae, 7, 34, 87, 214
16, 218, 220, 233, 234

Arctostaphylos, 326, 341

Arcytophyllum caracasana,
590, 591

Arecaceae, 56, 482, 516

Argemone, 59, 518

1
Argyroxiphium, 21, 23, 58, 61,
147-50, 152, 186, 191, 362,
381, 383, 404, 407, 434, 443,
455, 456, 457, 517, 535, 584;
A. caliginii, 148, 149, 167,
381, 584; A. grayanum, 148,

Index of Biological Names

171, 404, 584; A. kauense,
148; A. sandwichense, 98,
147, 148, 168, 171, 189, 456,
542; A. virescens, 148, 171,

Aristotelia fruticosa, 243, 247
Artemisia, 35, 58, 84, 152, 171,
17; A. californica, 297

Arthrocnema spp., 289

Aspidiaceae, 438

Aspilia, 454

Asprotilapia, 339

Astelia, 57, 82, 516, 523, 524,
575, 612; A. papuana, 571,
573, 574

Aster, 84, 416, 517

Asteraceae, 21, 23, 32, 35, 58,
61, 120, 147, 149, 152, 153,
158, 182, 193, 197, 202, 206,
208, 209, 211, 212, 231, 240,
242, 243, 248, 249, 341, 362,
401, 403, 404, 405, 407, 408-
11, 413, 415, 419, 421-23,
443, 456-60, 517, 587

Asteraceae: Heliantheae,
tribe Madiinae, 334

Astragalus, 570

Athrotaxis, 226

Athyrium, 437; A. kaalanum,
439; A. marginale, 439;
meyenianum, 439; A. micro-

sub-

phyllum, 439; A. molo-
kaiense, 438; A. sandwich-
ianum, 439

Atlantis, 504; A. vespertinus,
504, 505

Atriplex, 257

Aucklandella —_ flavomaculata,
499, 500

Austrotaxus, 215, 225
Avicennia, 76-77
Aythya australis, 49
Azorella, 587

Bacopa, 58, 518

Baeckia, 289

Baeopterus robustus, 499

Baicalia, 543; B. ciliata, 340;
B. costata, 340

Baicaliidae, 340

Baicalina, 504

Baicalinini, 504

Balaustion, 289

Banksia, 264, 267, 268, 280,
282, 288, 291, 300, 309, 326,
327; B. coccinea, 282; B.
grandis, 264, 267, 268, 282;

B. ilicifolia, 268: B. laricina,

324, 327, 329; B. menziesii,
268, 323, 324, 326; B. nu-
tans, 267, 268, 307; B. oc-
cidentalis, 288; B. prionotes,
268; B. prostrata, 267, 268,
282, 307, 321; B. solandri,
282; B. speciosa, 268; B.
sphaerocarpa, 264, 267, 288

Barhodes, 336, 337; B. amara,
336, 337; B. baoulan, 336;
B. binotatus, 336; B. clem-
ensi, 336, 337; B. disa, 336;
B. flavifuscus, 337; B. katolo,

621

337; B. lanaoensis, 336; B.
lindog, 336; B. manalak, 336;
B. quinquemaculatus, 336;
B. sirang, 337; B. tumba, 336

Barringtonia, 76-77, 355

Bassia, 257, 293

Batis, 76-77

Batrachocottus, 340

Baumea, 56, 516

Bavayia, 232

Bedfordia linearis, 422; B. sali-
cina, 422

Begonia, 424; B. peruviana,
424

Begoniaceae, 58, 424

Belgica antarctica, 498, 499

Bembidium obtusum, 503

Bencomia, 388, 414, 415, 557

Bidens, 13, 22, 31, 34, 58, 59,
63, 64, 120, 123, 131, 404,
407, 408, 417, 443, 444, 447,
450, 453, 454, 476, 517, 535,
536, 537, 542, 582; B. ahnei,
444, 444; B. aoraiensis, 446,
449; B. australis, 445, 446,
449; B. campylotheca, 447,
448; B. collina, 446; B. con-
juncta, 449; B. cordifolia,
446; B. cosmoides, 446, 447,
449, 536, 537; B. cteno-
phylla, 447, 448, 449; B.
cuneata, 446, 447; B. de-
generi, 447, 448-49; B. del-
toidea, 446, 449; B. forbesii,
446, 447; B. _ glabrata,
446; B. glandulifera, 446,
449: B. hawaiiensis, 449; B
hendersonensis, 407, 445,
449; B. hendersonensis var.
oenoensis, 408; B. hille-
brandiana, 446, 447, 449; B.
hivoana, 444, 444; B. lanta-
noides, 446, 449; B. macro-
carpa, 447, 448, 450; B.
magnidisca, 447, 448, 450;
B. mathewsii, 408, 445, 449;
B. mauiensis, 446, 447; B.
menziesii, 446, 447; B. mi-
cranthoides, 449; B. molo-
kaiensis, 449; B. mooreensis,
407, 446, 449; B. nemato-
cera, 446, 447, 449; B.

orofenensis, 446, 449: B.
pilosa, op 433, 434, 444,
444, 445, 451; B. polyce-

phala, 444, 444; B. raia-
teensis, 449; B. saintjohniana,
408, 445, 448; B. sandvi-
censis, 432, 447, 448; B.
skottsbergii, 515; B. waianae-
ensis, 366; B. waimeana,
449; B. wiebkei, 447, 448
Bignoniaceae, 146
Billardiera, 276
Blackallia connata, 275
Blancoa, 297, 299
Blechnaceae, 157, 438
Blechnum, 207, 574; B. cycadi-
folium, 207; B. occidentale,
438; B. werckleanum, 573
Blenniidae, 339

Bobea, 58, 518, 525; B. elatior,
476

Boehmeria, 518
Boerhavia, 57, 70-72, 80, 518;
B. diffusa, 70, 71

Bonamia, 297
182, 183, 189,

Boraginaceae,
410, 414, 415, 416, 424

Boronia, 288; B. defoliata,
271; B. ericifolia, 288; B.
juncea, 271

Bosea yervamora, 384, 386, 386,
401, 413

Bossiaea: B. concinna, 276;
B. leptacantha, 276; B.
peduncularis, 276; B. preissii,
276; B. rufa, 276; B. walkeri,
MY BYES, US

Bothrops insularis, 33

Bowlesia, 441

Brachionostylum, 409

Brachyapteragrotis patricei,
497

Brachycome, 575

Brachyglottis, 361, 409

Brachylaena, 361, 419; B.
merana, 419, 554

Brachymeles, 335, 609

Brachysema, 264, 265, 266; B.
aphyllum, 262, 264, 272, 307;
B. bracteolosum, 264, 265;
B. chambersii, 307; B. da-
viesioides, 264, 266, 307; B.
lanceolatum, 264, 265; B.
praemorsum, 264, 265; B.
sericea, 264, 265; B. tomen-
tosum, 264, 266, 272, 307

Branta sandvicensis, 616

Brassicaceae, 57, 193, 196, 404,
405, 413-416, 568, 570, 587

Brassica oleracea, 368

Breweria, 58, 432, 517;
data (= Bonamia_ semidi-
gyma), 462; B. media, 462;
B. menziesii, 462

Brighamia, 58, 63, 130, 144,
404, 406, 417, 517; B. in-
signis, 399

Brodiaea, 80

Bromeliaceae, 87, 424, 592,
593

Broussaisia, 57, 518, 525, 527,
530, 612; B. arguta, 2/5 Be
pellucida, 2

Bruguiera, 76—77

Brunellia, 576

Brunelliaceae, 576

Bubbia, 575

Burpleurum, 231, 420

Burmeistera, 405, 424

Buxifoliatae, 238

Bystropogon, 414

B. cor-

Cacicus, 159

Cactaceae, ath 203, 402, 412

Cactospiza, 22, 159; C. pallida,
98, 200

Caesalpinia, Sh te SLY

Calamagrostis, 56, 517, 584,

Index of Biological Names

586, 590, 593; C.
584, 586

Calandrinia, 412; spp., 289; C.
galapagosa, 412

Callistemon, 286, 327

Callistopteris baldwinii, 439

Callitris, 225, 226

Calogyne, 310

Calophyllum, 76-77, 355

Calorhaphus gracillima, 278,
294

Calothamnus, 289, 293, 300,
327; C. gracilis, 323, 325,
B29 Cs) lateralisy S25) 1Gs
microcarpus, 327; C. pini-
folius, 294, 296; C. plani-
folius, 327; C. schaueri, 327

Calythrix, 293

Camarhynchus, 199; C. par-
vulus, 98, 201; C. psittacula,

8, 201

Campanula, 421

Campanulaceae, 58, 120, 140,
227, 404, 405, 407, 408, 410,
415, 417, 424, 460, 461

Campbellana, 495, 496; C. at-
tenuata, 495, 496

Camphusia, 467

Campoplex disjunctus, 495

Campsicnemus, 125, 174

Canavalia, 57, 76, 430, 434,
467; C. galeata, 467; C.
hawaiiensis, 467; C. kauaien-
sis, 467; C. maritima, 467;
C. megalantha, 467; C. mo-
lokaiensis, 467; C. pubes-
cens, 467; CC. raiateensis,
467; C. sericea, 467; C. viti-
ensis, 467

Canellaceae, 87

Canthium, 58, 67, 82, 518, 525

Canuta, 424

Capparaceae, 57, 517

Capparis, 57, 76, 517; C. sand-
wichiana, 76

Caprifoliaceae, 409, 410

Carabidae, 26, 488, 491, 492,
493

Cardamine, 570, 575

Carduus, 570

Carelia, 28, 128, 177, 607

Carex, 56; 8455165847 aiG-
pinetorum, 590

Carica, 424

Caricaceae, 87, 424

Carlina, 381, 413; C. canarien-
sis, 413; C. salicifolia, 413

Carmichaelia, 216, 410; C. an-
gustata, 248; C. arborea,
410; C. curta, 246; C. kirkii,
246; C. uniflora, 246

Carpobrotus spp., 289

Caryophyllaceae, 57, 241, 404,
406, 414, 461, 462, 517, 529

Cassia, 517

Cassinia, 409, 422; C. longi-
folia, 422; (es retorta, 409;
C. trinervia, 422

Cassytha, 57, 108, 276, 518

Castanea, 372

Casuarina, 226

622

effusa,

Casuarinaceae, 226

Catosperma, 312; C. gooden-
iacearum, 313

Ceanothus, 413

Celastraceae, 57, 271, 276

Cenchrus, 56, 517

Centaurea, 381, 382, 413-15,
457, 458; C. arbutifolia, 414;
C. arbutoides, 366; C. mas-
soniana, 415

Centaurium, 518

Centaurodendron, 32, 458, 459;
C. dracaenoides, 374, 457;
C. palmiforme, 374

Contr olen idaccee Sil, S74,
57

Centrolepis, 575; C. philip-
pinensis, 571, 572, 574
Centropelma micropterum, 490
Centropogon, 405,
densiflorus, 424
Centrospermae, 420
Cephalokompsus, 337; C. pa-
chycheilus, 337
Cephalotaceae, 317, 3/8
Cephalotis follicularis,

.
>

317,

Cerambycid beetles, 167

Cerastium, 570

Ceratioidea, 343-46

Cerbera, 217

Cercocarpus, 413; C._ intri-
catus, 242; C. ledifolius, 242

Cereus, 412
199; C. olivacea,

Certhidea,
98, 201

Cestrum, 406, 424; C. laevi-
folium, 424; C. longiflorum,
424

Chaetocalyx, 106, 396

Chalarodon, 333

Chamaescilla corymbosa, 320,
21

Chamaesyce, 204, 387, 463; C.
viminea,
Chamissoa, 405, 522
Charpentiera, 57, 387, 390,
401, 404, 405, 515, 517, 521,
522, 525, 527215285 S30 NO SIE
C. australis, 407; C. obovata,
386; C. ovata, 527
Cheiranthera filifolia, 276; C.
parviflora, 276
Cheiranthus, 381, 414-16; C.
mutabilis, 193; C. scoparius,
193; C. virescens, 193
Cheirodendron, 14, 58, 172,
517, 519, 520, 540
Chenopadiacede, 57, 257, 405—
Chenopodium, 35, 57, 171,
257, 388, 401, 405s5iie
arborescens, 384; C. neso-
dendron, 410; C. oahuense,
171, 384, 405, 406
Chironomidae, 498, 499
Chlaenaceae (see Sarcolaena-
ceae)
Chlorathaceae, 88
Cholovocera madaerae, 503,
504

Chordospartium, 246
Choretrum, 271

Chorilaena quercifolia, 279,
9Q9

Chorizema, 293

Chrysanthemum, 414; G:
broussenetii, 414

Chuquiraga jussieui, 588, 590

Cibotium, 136, 172, 438; C.

chamissoi, 438; C. glaucum,
438: C. st. johnii, 438; C.
splendens, 438

Cichlid fishes, 337, 338, 339

Cichorieae (see Lactuceae)

Ciidae, 504

Ciridops, 1285 129) 159; 162;
169; C. anna, 169

Cis wollastoni, 504

Cladium, 56, 6

Claoxylon, 57, 517, 525

Clermontia, 28, 31, 35, 58, 63,
66, 119, 120, 124, 139, 143-
46, 165, 172, 384, 404, 405,
421, 517, 537; C. arborescens,
66, 143, 144, 393, 404; C.
clermontioides, 124, 143,
393; C. grandiflora, 144; C.
haleakalensis, 124, 143; C.
hawaiiensis, 143, 393, 535;
C. kakeana, 124, 144, 145,
393; C. micrantha, 143; C.
montis-loa, 393, 535; C.
multiflora, 124; C. oblongi-
folia, 124; C. parviflora, 143,
535; C. peleana, 143; C.
samuelii, 124

Clitarchus, 232

Clusia, 576

Clusiaceae, 226, 576

Clytarlus: C. modestus, 167;
C. pennatus, 167

Cocculus, 57, 518, 524

Cochlicopidae, 127, 607

Cochlospermum, 424; C. or-
inocense, 424

Coeloneurum, 418

Coelophryne, 346

Coelopidae, 497, 498, 499

Coerebidae, 201

Coleoptera, 194, 492, 501, 502,
503, 505, 597, 612, 615

Colletia, 532

Colobanthus, 497

Colpodes bromeliarum, 606

Colubrina, 57, 66, 164, 436,
475, 518; C. asiatica, 465,
Be C. oppositifolia, 465,

Colydiadae, 503

Combretum, 88

Comephoridae, 20, 339, 340

Comephorus: C._ baicalensis,
340; C. dybowskii, 340

Comesperma, 268, 276; C.
drummondii, 289, 290; C.
scoparium, 269, 271, 275;
C. spinosum, 274, 2715; C
volubile, 276, 277

Commidendron, 416, 417; C.
robustrum, 416

Compositae (see Asteraceae)

Index of Biological Names

Conolophus, 612
Conophytum, 342
Conospermum, 297; C.

vum, 293
Conostephium pendulum, 307
Conostylis, 297, 299, 321
Conus, 343

Incur-

Convolvulaceae, 58, 384-87,
385, 414, 415
Convolvulus, 58, 384, 385,

386, 389, 401, 414, 415, 462;
C. floridus, 414; C. massonii,
415; C. scoparius, 385, 414

Conyza, 41

Coopernookia, 312

Coprosma, 33, 58, 62, 142,
152% 1555 17S" 238392359) 241%
243. 51855235924) 583 hS0%
571, 574, 575; C. acutifolia,
238; C. australis, 238; C.
brunnea, 239, 240; C. chat-
hamica, 238: C. crassifolia,
239: 2405 32433) 1G) “erno=
deoides, 155, 524; C. foetidis-
sima, 238, 239; C. lucida,
238; C. macrocarpa, 238; C.
montanan A555.) 1733 1G:
moorei, 524; C. ochracea,
15530 Ga ppetiolata, 9238: G.
petriei, 240, 241, 242; C.
pseudocuneata, 239, 240; G:
pumila, 242, 524; C. repens,
23831G: robusta, 238; C. ten-
uifolia, 238

Corallospartium, 246

Cordia, 205

Coreopsidinae, 450, 452

Coreopsis, 413, 450; C. gigan-
tea, 26, 352, 390, 399, 400,
402, 413

Coriaria, 13, 410; C. arborea,
40

Coriariaceae, 409

Cornaceae, 243, 244

Corokia, 243; C. buddleoides,
243, 244; C. cotoneaster,
243, 244; C. macrocarpa,
243

Cortaderia, 530

Corvus tropicus, 129

Corynotheca micrantha, 286

Cossyphodes, 503, 504; C. wol-
lastoni, 503, 504

Cottidae, 339, 340

Cottocomephoridae, 340

Cottocomephorus, 340

Cotula, 242, 249

Crambe, 381, 414, 415; C. fru-
ticosus, 415

Crassocephalum, 554, 581; C.
mannil,

Crasspedia, 249

Crassulaceae, 402, 414, 415

Crecopsis sp. 489

Cressa, 9, 58, 517

Crossosoma californicum, 25

Croton, 424

Cruciferae (see Brassicaceae)

Cryptandra, 275

Cryptanthemis, 309

Cryptocarya, 57, 108, 470, 471,

623

518, 519, 532; C. angulata,
470; C. glaucescens, 470; C.
hornei, 470; C. hyposporlia,
470; C. mannii, 108, 470; C.
oahuensis, 470; C. samoen-
sis, 470

Cryptostylis, 309

Cucurbitaceae, 58, 387, 420,
462

Cuminia, 410-11

Cunoniaceae, 216, 226, 228,

231, 354
Cuphea, 424; C. cordata, 424
Cy presstceee) 219, 220, 226,

Curculionidae, 504, 505

Cuscuta, 58, 517

Cyanea, 28, 58, 63, 102, 119,
120, 124, 131, 132, 133-39,
227, 384, 389, 404, 405, 474,
517, 537; C. aculeatiflora,
131, 138, 139; C. acuminata,
S18 G angustifolia, lhl
134, 135; C. arborea, 134:
C. asplenifolia, 138; C. atra,
I3t eis7, 13935G. atra var:
lobata, 137; C. bishopi, 131;
C. bryanii, 136; C. carlsoni,
134; C. fauriei, 134; C. fer-
naldii, 136; C. ferox, 137,
138; C. f. var. horrida, 137;
C. fissa, 134; C. gayana, 134;
C. gibsoni, 134; C. grimes-
iana, 138; C. hamatiflora,
131; 138; €. hirtella;, “134:
C. holophylla, 138; C. lep-
tostegia, 131, 133, 134-37,
393, 402, 583; C. longipe-
dunculata, 131; C. marksii,
137, 139; C. nolimetangere,
133, 136; C. pilosa, 131, 393;
C. quercifolia, 137; C. recta,
134; C. regina, 134; C. rol-
landioides, 133, 136; C. sela-
chicauda, 142; C. shipmanii,
138; C. solanacea, 137;
solenocalyx, 131; C. sticto-
phylla, 137; C. superba, 134;
C. sylvestris, 134; C. trito-
mantha, 133, 136, 137

Cyanorhamphus, 251

Cyathea, 421, 571; C. hymeno-
phylloides, 571

Cyatheaceae, 438

Cyathodes, 526; C. acerosa,
526; C. colensoi, 526-27

Cycadaceae, 322, 323

Cylicomorpha parviflora, 554

Cylindrocline, 421

Cynareae, 61, 411, 457

Cynoglossum, 189

Cyperaceae, 56, 64, 271, 516

Cyperus, 56, 516

Cyphomandra, 424; C. endo-
pogon, 424; C. pendula, 424

Cyrtandra, 20, 47, 63, 85, 105,
LOG SATO I2O 1D 3 2s
155, 156, 364, 395, 396, 401,
405, 406, 408, 518, 528, 529,
530, 537, 612; C. basiflora,
155; C. brevicornuta, 155;

Cyrtandra (Cont.)
C. calpidicarpa, 156; C.
cornuta, 156; . gayana,
155, 528; C. kalihii, 155; C.
kaulantha, 155; C. laxiflora,
156; C. lysiosepala, 155; C.
oenobarba, 155; C._platy-
phyla var. brevipes, 528; C.
propinqua, 396; C. stupan-
tha, 155; C. waiolanii, 155

Cyrtandroidea jonesii, 407, 408

Dacrycarpus, 226

Dacrydium, 215, 220, 221, 222,
DD 322206234, Sos;
araucarioides, 217, 220, 221,
223; D. balansae, 221, 223;
D. bidwillii, 234; D. biforme,
234; D. colensoi, 234; D.
cupressinum, 234; D. fal-
ciforme, 570; D. gibbsiae,
570; D. guillauminii, 222,
223; D. intermedium, 234;
D. laxifolium, 234; D.
copodioides, 215, 220,
D. taxoides, 220, 221

Dampiera, 268, 272, 275, 297,
300, 314, 321, 326; D. alata,
270, 271; D. curvula, 294;
D. epiphylloidea, 272: D.
hederacea, 279, 282; D. jun-
cea, 269, 271; D. lindleyi,
272: D. luteiflora, 297,
298; D. phlebopetala, 282;
D. striata, 282; D. teres, 279,

y -
222;

283; D. trigona var. latea-
lata, 272; D. wellsiana, 320,
321

Darwinia, 255, 289, 302, 303,
304, 304, 314, 315, 316, 472;
D. acerosa, 315, 316; D. ci-
triodora, 302, 303; D. col-
lina, 302; D. diosmoides,
302, 303; D. fimbriata (D.
squarrosa), 302; D. heli-
chrysoides, 302, 304; D. hy-
pericifolia, 302, 315, 316; D.
leiostyla, 302, 303, 304; D.
macrostegia, 302, 315, 316;
D. meeboldii, 302, 315, 316;

D. oederoides, 315, 316; D.
pauciflora, 314, 315; D. san-
guinea, 302, 303, 307; D.

speciosa, 307
Darwiniothamnus, 411, 412
Daucus, 517

Daviesia, 259, 260, 261, 262,
263, 283, 293; D. aphylla,
261, 263, 271; D. dielsii,

259, 260; D. divaricata, 261,
263; D. epiphylla, 262,
263, 272; D. euphorbioides,
262, 263; D. hakeoides, 307;
D. obtusifolia, 259, 260, 263:
D. pachyphylla, 259, 260; D.
polyphylla, 259, 261, 263; D.
retroflexa, 261, 263: OD.
teretifolia, 259, 260; D. tri-
gonophylla, 262, 263

Decaryia, 419

Deinacrida, 249

Index of Biological Names

Delissea, 28, 58, 63, 119, 120,
124, 130, 131, 134, 139, 143,
144, 384, 404, 405, 461, 517;
D. lacinata Hbd., 142; D.
niihauensis, 130; D. rhyti-
dosperma, 130, 131; D. sin-
uata Hbd., 142; D. sub-
cordata, 130, 131; D. un-
dulata, 130, 404

Delphacidae, 172

Dendriopoterium, 414, 415,
557; D. menendezi, 366, 367

Dendrobium — psychrophyllum,
Sil

Dendromecon, 413

Dendrosenecio, 102, 423, 554,
571

Dendroseris, 32, 34, 206, 211,
227 V4 Ok Ali 4435p)
berteriana, 400-401, 403;
D. gigantea, 459, 460; D.
(Hesperoseris) gigantea, 459,
460; D. litoralis, 211, 381,
411, 435; D. macrantha, 410;
D. macrophylla, 366, 367,
3600s Ds (Phoenicoseris)
berteriana, 400-401, 403; D.
(Phoenicoseris) pinnata, 373)
435, 460; D. (Rea) micran-
tha, 459, 460; D. (Rea)
pruinata, 211, 460; D. regia,
373

Dendrosicyos, 387, 420; D.
socotrana, 420

Deschampsia, 56, 517

Descurainia, 414

Dianella, 57, 82, 435, 516

Diapriidae, 501

Diceratella, 415

Dicksonia, 207

Dicoma, 419

Dicranopteris emarginata, 438;
D. linearis, 438

Didierea, 419; D. madagas-
cariensis, 419
Didiereaceae, 333, 419, 501

Digitaria, 56, 517
Dilleniaceae, 226, 288

Dimorphinoctua cunhaensis,
197

Diospyros, 58, 82, 164, 462,
463, 483; D. ferrea, 462,

463; D. hillebrandii, 462; D
samoensis, 463
Diplolaena, 297, 300; D. dam-
pieri, 301
Diplostephium
588, 590
Dipsacaceae, 193, 414
Diptera, 497, 498, 499, 501
Discaria toumatou, 243, 247,
248
Diselma, 226
Dissochondrus, 517
Distichia, 590; D.
587

phylicoides,

tolimensis,

Dodonaea, 57, 76, 130, 152,
256-57, 436, 480, 518, 525;
D. viscosa, 62, 553

Dolichopodidae, 125, 126, 174,
497, 4

624

Donatia, 242

Doodia_ kunthiana, 438; D.
lyonii, 438

Dorstenia, 420; D. gigas, 420

Draba _ sancta-martae, 589

Dracaena, 57, 82, 483; D.
cinnabari, 420 (see also Pleo-
mele)

Dracophyllum, 242; D. mus-
coides, 242; D. pronum,
242; D. prostratum, 241, 242

Drapetes, 571, 574, 575

Drepanididae (Drepanidae,
Drepaniidae), 105, 158-69,
160, 161, 198

Drepanis, 105, 144, 159, 163,
169; D. pacifica, 169

Drimys, 65, 68, 88, 216, 570,
575, 576; D. piperita, 570;
D. winteri, 68

Drosera, 57, 257, 258, 268,
276, 317, 319, 329, 517, 520;
D. gigantea, 317, 318; D.
heteromera, 277, 317; OD.
macrantha, 320; D. macro-
phylla, 258, 317, 328, 329;
D. pulchella, 317, 318; D.
stolonifera, 317, 323

Droseraceae, 268, 277, 317,
318, 320, 328

Drosophila, 59, 122, 163, 176,
492, 493; D. crucigera, 175;
iD: engyochracea, 175; D.
STITT aU Vides willistoni,

Drosophilidae, 20, 29, 125,
126% 127

Drusa, 441

Dryandra, 291, 300, 307; D.

horrida, 291; D.
307, 308

Drypetes, 57, 517, 524

Dubautia, 31, 58, 61, 84, 119,
120, 123, 126, 147, 149-53,
153, 157, 184, 199, 359% 362%
381, 383, 397, 401, 404, 455,
517, 535, 582; D. arborea,
404; D. knudsenii, 150, 383,
404, 456; D. latifolia, 150,
381, 456; D. laxa, 150; D.
menziesii, 98, 147, 151; D.
molokaiensis, 150; D. pa-
leata, 150, 381; D. planta-
ginea, 150, 381; D. (Rail-
liardia) ciliolata, 62, 161;
D. (Railliardia) scabra, 151,
381; D. railliardioides, 150;
D. reticulata, 150, 381; D.
waialealae, 151

runcinata,

Ebenaceae, 58, 462

Echinosoma, 504; E. porcellus,
504, 505

Echium, 182-91, 183, 185, 187,
188, 195, 353, 365, 381, 394,
401, 414-16; E. aculeatum,
100; E. auberianum, 186,
187, 189; E. bonnettii, 182;
E. bourgeauanum, 186, 187,
189, 190; E. candicans, 182,
184, 185, 190, 366, 415, 416;

E. faramae, 188, 190; E.
fastuosum, 184; E. gigan-
teum, 184, 190; E. hierrense,
182, 183, 184; E. hyper-
tropicum, 416; E. leuco-
phaeum, 184, 187, 189, 190;
E. leucophaeum var. angu-
stissimum, 183; E. nervosum,
188, 190, 415; E. onosmae-
folium, 183, 184, 190; E.
pininana, 184, 185, 186, 190,
366, 367, 373, 414, 415; E.
plantagineum, 182, 183; E.
simplex, 185, 186, 366; E.
strictum, 182, 185, 190; E.
triste, 182; E. vulcanorum,
416; E. webbii, 188, 190

Ekmania, 417

Elaeocarpaceae, 247, 463, 483,
517

Elaeocarpus, 57, 82, 483, 517,
571; E. bifidus, 463; E. hook-
erianus, 463; E. magnifolius,
463; E. rarotongensis, 463;
E. ulianus, 463, 519

Elasmias, 83

Eleocharis, 56, 516

Eliaea, 419; E. articulata, 419

Embelia, 58, 518

Empetrum nigrum, 73

Enantiophyllum, 409

Enantiopus, 339

Encoilidae, 499, 500

Encyclodes, 606

Encyrtidae, 499, 500

Endodontidae, 607

Enicodini, 232

Entada, 75, 81, 517

Epacridaceae, 58, 226, 243,

286, 288, 294, 302, 423, 526
Epilobium, 571, 575
Eragrostis, 56, 171, 517;-E.

variabilis, 167
Eremocharis, 423; E. piscoen-

sis, 423; E. triradiata, 423
Erica, 557
Ericaceae, 58, 286, 557, 567,
568, 584
Erigeron, 410, 411, 412, 575;

E. ecuadoriensis, 584, 586;
E. fruticosus, 410
Eriocaulaceae, 571 ff.
Eriocaulon, 562, 575; E. mon-
tanum, 571, 572, 574
Eriogonum, 413; E. arbores-
cens, 413; E. fasciculatum,
297; E. giganteum, 413
Eriophyllum, 341, 413; E.
nevinii, 413
Eriostemon, 288; E. deserti,
276
Eryngium, 410, 523; E. bu-
pleuroides, 410; E. fernandez-
ianum, 410; E. humile, 586
Erysimum, 413; E. insulare,
13; E. moranii, 413
Erythrina, 7, 76, 130, 464, 465,
517; E. indica, 464; E. mono-
sperma, 464, 465; E. ovali-
folia, 464; E. sandwichensis,

Index of Biological Names

464, 465; E. variegata, 464,
465

Escallonia, 576

Escallonioideae, 576

Eschscholzia, 413; E. frutes-
cens, 413; E. palmeri, 413

Espedaea, 418; E. amoena, 418

Espeletia, 106, 364, 387, 423,
549, 563, 576-84, 577, 590;
E. caldasii, 577, 582; E.
conglomerata, 578, 582, 583;
E. discoidea, 580, 582-84;
E. glossophylla, 581; E.
grandiflora, 592; E. hart-
wegiana, 580, 582-84; E.
incana, 578, 582, 583; E.
neriifolia var. columbica,
577, 581-83; E. phaneractis
boyacensis, 577, 582, 583;
E. pleiochasia, 579, 582, 583;
E. rositae, 578, 582, 583; E.
santanderensis, 577, 582,
583; E. schultzii, 578, 582,
583; E. timotensis, 579, 582,
583; E. uribei, 579, 582-84;
E. vaccinoides, 580

Estigmene, 19

Eucalyptus, 283, 326, 329, 421;
E. diversicolor, 282; E. mac-
rocarpa, 300, 316, 326, 472;
E. marginata, 263; E. preis-
sii, 326; E. rostrata, 263; E.
tetraptera, 326

Eupatorium, 418; E. portori-
cense, 418

Euphorbia, 57, 114, 138, 156,
157, 182, 195, 204, 342, 360,
364, 381, 387, 391, 392, 401,
402, 405, 414-18, 46], 463,
AVL SSN. “BH, Biss o 7 de
sect. Anisophyllum, 387;
E. atoto, 394, 463; E. bal-
samifera, 390, 414; E. sect.
Cabanthus, 417; E. can-
delabruri 138. 55952 E-
celastroides, 461, 463, 464;
E. celastroides vars. amplec-
tans, lorifolia, and mauien-
sis, 392; E. sect. Chamae-
syce, 204, 392; E. clusiae-
folia, 392, 463, 464; E. de-
generi, 387, 392, 394, 463;
E. sect. Euphorbiodendron,
417-18; E. haeleeleana, 387,
405; E. halemanui, 392; E.
mellifera, 414; E. petiolaris,
418; E. piscatoria, 353, 392,
415; E. pitcairnensis, 463; E.
regis-jubae, 414; E.
var. pteropoda, 392 -aaar.

rockii, 157, 387, 389, 392,
394, 405; E. taitensis, 463;
E. tessmannii, 424; E. sect.
Tithymalus, 392; EE. tuc-
keyana, 416

Euphorbiaceae, 57, 156, 182,
204, 414-18, 424, 461, 463,

517, 524
Euphrasia, 571 ;
Eurya, 57, 518, 524, 571; E.
sandwicensis, 522

625

Eurydactylus, 232
Eurygnathus latreillei, 502
Eurynogaster, 125,
Evodia, 154, 476, 477; E.
samoensis, 477
Exocarpus, 518; E.
271; E. spartea, 271
Exsilaracha, 495, 496; E.
graminea, 495, 496

Fabaceae, 57, 74, 75, 246, 248,
259, 260, 261, 262, 264, 265,
266, 271, 272, 273, 276, 293,
He, 413, 436, 464, 514, 517,

5

Fagara, 434, 476, 477, 478; F.
ailanthoides, 477, 478; F. di-
petala, 478; F. gillespieana,
478; F. megistophylla, 478;
F. oahensis, 478; F. pinnata,
478 (see also Zanthoxylum)

Faujasia, 421

Festuca, 56, 517, 529, 562, 567;
F. tolucensis, 589

Fimbristylis, 56, 516

Fitchia, 212, 384, 418, 443, 450,
451, 452, 453; F. cordata,
212, 452; F. mangarevensis,
408, 452, 453; nutans,

; F. rapensis, 407, 452,

; F. speciosa, 212, 407,

452, 453, 460

Fitchiinae, 452

Flacourtiaceae, 58, 424

Flagellariaceae, 56-57, 516

Flotovia, 423

aphylla,

. Fragaria, 57, 84, 518; F. chi-

loensis, 525

Frankenia, 257, 288, 417; F.
portulacaefolia, 417

Frankeniaceae, 288, 417

Franklandia, 286, 294; F. fu-
cosa, 294

Freycinetia, 56, 163, 165, 166,
516, 524, 606; F. arborea,
165, 168

Fringilla, 194; F. coelebs, 194;
F. teydea, 194, 608

Fringillidae, 102, 197, 200

Fuchsia, 410; F. cyrtandroides,
407, 408, 530; F. excorticata,
408, 409, 530; F. perscan-
dens 530; F. procumbens,

Gahnia, 56, 561, 571

Gaimardia, 242

Galaxias, 250; G. brevipinnis,
250; G. divergens, 250; G.
fasciatus, 250; G. gracilis,
250; G. maculatus, 250; G.
postvectis, 250; G. prog-
nathus, 250; G. usitatus, 250

Galaxiidae, 249-51

Galium, 413, 570; G. angulo-
sum, 413

Galvezia speciosa, 26

Gammaridae, 340

Gardenia, 58, 518, 519, 523

Garnotia, 56, 517

Gastropoda, "488

Gaultheria, 31, 527, 581; G.
depressa, 242

Gekkonidae, 232

Gelis campbellensis, 499

Geniostoma, 63, 474, 523

Gentiana, 550, 571, 575, 584;
G. lycopodioides, 571; G.
piundensis, 571, 573, 574;
G. sedifolia, 584, 586.

Gentianaceae, 414, 424, 518,
SS Ohm dLs 584

Geospiza 198-201; G. difficilis,

; G. fortis, 201; G. fuli-

ae 20IEKG scandens, 201

Geospizinae (Geospizidae), 46,
79, 102, 158, 197-201, 199,
200, 205, 206

Geraniaceae, 57, 405, 417, 518

Geranium, 57, 84, 152, 173,
395, 405, 406, 518, 550; G.
arboreum, 146, 405; G.
jahnii (sect. Paramenses),
406; G. maderense, 308, 374;
G. multiceps, 584, 586; G.
tridens, 173

Gesneriaceae, 58, 119, 155, 156,
405-7, 418, 612

Gesnouinia, 414, 415

Gingidium, 530

Girellidae, 339

Gleichenia, 571

Gleicheniaceae, 438

Glossogobius griseus, 337

Gnaphalium, 517

Gochnatia, 405

Goetzia, 418; G. elegans, 418

Gompholobium, 293

Gomphostemma, 469

Goniocormus minutus, 439

Gonospermum, 414; G. canar-
iense, 414

Goodenia, 310, 312, 422; G.
incana, 311; G. ovata, 422

Goodeniaceae, 58, 75, 157, 226,
2295 26832695 2705) 271,
279, 297, 298, 306, 310, 311,
313, 320, 322, 323, 406-9,
467

Gossypium, 7, 57, 81, 406, 471,
472, 518; G. arboreum, 406;
G. hirsutum, 471; G. sand-
vicense, 471, 519; G. tomen-
tosum (see G. sandvicense)

Gouania, 57, 436, 518

Gouldia, 31, 58, 518,
530, 535

Gramineae (see Poaceae)

Greenovia, 414

Grevillea, 268, 293, 300, 307;
G. acrobotrys, 291; G. am-
plexicans, 291; G. annuli-
fera, 293; G. apiciloba, 291;
G. armigera, 293; G. aspara-
goides, 293; G. astericosa,
293; G. bipinnatifida, 293;
G. bracteosa, 302; G. brevi-
cuspis, 293; G. cirsiifolia,
293; G. flexuosa, 293;
glabrata, 293; G. incon-
spicua, 275; G. intricata, 293;
G. leptobotrya, 293; G.

Index of Biological Names

macrostylis, 293; G. muel-
leri, 293; G. paniculata, 293;
G. paradoxa, 293; G. paten-
tiloba, 292, 293; G. pecti-
nata, 307, 308; G. phanero-
phlebia, 293; G. platypoda,
293; G. pulchella, 293; G.
rudis, 293; G. stenostachya,
293; G. synapheae, 293; G.
tenuiloba, 293; G. teretifolia,
293; G. thelemanniana, 293;
G. tridentata, 293; G. trifida,
293; G. triloba, 293; G. tri-
partita, 293; G. wilsoni, 293
Gunnera, 58, 65, 66, 406, 411,
468, 518, 525, 532, 535; G.
bracteata, 535; G. chilensis,
66; G. macrophylla, 468; G.
magellanica, 584; (586; G.
mauiensis, 468; G peltata,
535; G. prorepens, 468
Guttiferae (see Clusiaceae)
Gynoxys, 423
Gyrostemonaceae, 422

Haastia, 587
Habenaria, 57, 61, 516
Haemodoraceae, 297
Hagenia, 388, 557; H.
sinica, 557
Hakea, 268, 282, 283, 284, 286,
287, 289, 293, 326; H. am-
plexicaulis, 283, 291; H.
auriculata, 283, 291; H. bax-
teri 2834 2865 291):
brookeana, 293, 326; H. cer-
atophylla, 293; H. clavata,
287, 289; H. conchifolia,
282, 287, 291: Hi. cristata,
293; H. cucullata, 282, 287,
288, 305, 306; H. dolicho-
styla, 293; H. elliptica, 283;
H. erinacea, 293; H. ferru-
ginea, 283, 286; H. flabelli-
folia, 283, 286, 291; H. gla-
bella, 293; H. hookeriana,
283, 286; H. laurina, 286;
H. lehmanniana, 293; H.
multilineata, 283; H. obli-
qua, 325, 326; H. pandani-
carpa, 286, 325, 326, 327;
H. platysperma, 293; H.
pritzelii, 293; H. purpurea,
293: H. roei, 283, 326; H.
smilacifolia, 283; H. stenop-
tera, 293, 295; H. trifurcata,
283; H. undulata, 283, 291;
H. varia, 292, 293; H. vic-
toriae, 283, 287, 291, 307
Halophila, 56, 516
Haloragaceiae (Haloragida-
ceae), 58, 323, 468, 518
Haloragis, 411, 570
Haplocarpha rueppellii, 567
Haplochromis, 339
Haplopappus, 341, 405, 413;
H. canus, 25, 413; H. deton-
sus, 413
Haplosciadium, 568
Haplostachys, 58, 435,
469, 470, 518; H.

626

abys-

436,
lineari-

folia, 469, 470; H. munroi,
469, 470

Hardenbergia, 276

Haronga, 419; H. madagas-
cariensis, 419

Harpalus vividus, 503

Hebe, 235, 236, 237, 238, 243,
246, 249, 251, 410; H. diosmi-
folia, 238: Ele epacridea, 236,
237; H. hectori var. demissa,
236, 237; H. parviflora var.
arborea, 235, 409; H. ramo-
sissima, 235; H. raoulii var.
maccaskillii, 238; H. re-
curva, 238; H._ salicifolia,
23): 236; Fe tetrasticha, 236,

ee 58, 417, 518, 523,
238 5 SySiee 'H. arborea, 417,

Heliantheae, 61, 147, 197, 202,
341, 412, 450, 454, 455, 460

Helichrysum, 421, 422, 557,
307, 569; H. glomeratum,

Heliotropium, 58, 424, 517

Hemiandrus, 249

Hemidactylus garnoti, 8

Hemideina, 249

Hemigenia, 297

Hemignathus, 144, 161; dH.
lucidus, 144, 164, 165; H.
procerus, 105, 144, 161, 165;
H. wilsoni, 166

Hemiphora elderi, 297, 299

Hemizonia, 100, 147, 334, 401,
413; H. clementina, 334; H.
frutescens, 334; H. green-
eana, 334, 413; H. minthor-
nii, 334; H. palmeri, 334; H.
streetsii, 334

Henicidae, 249

Henoonia, 418

Heracleum, 570

Hesperocnide, 518

Hesperomannia, 18, 28, 58,
404, 405, 517; H. arbuscula,
542; H. oahuensis, 542

Hesperomeles, 576

Hesperoseris, 206, 410

Heterochaenia ensifolia, 421

Heterolocha acutirostris, 609

Heteromorpha, 553; H. tri-
foliata, 555

Heteropogon, 56, 517

Heteroptera, 170

Hibbertia, 226, 227, 228; H.
badouinii, 227, 229

Hibiscadelphus, 28, 57, 146,
388, 405, 406, 472, 518; A.
giffardianus, 472; H. huala-
laiensis, 472; H. wilderianus,
472

Hibiscus, 57, 406, 419, 518;
H. arnottianus, 472; 4H.
brackenridgei, 472; H. im-
maculatus, 472; H. man-
gindranensis, 419; H. muta-
bilis, 472; H. tiliaceus, 472;
H. waimeae, 472

Hicriopteris pinnata, 438

Hillebrandia, 58, 517, 531

Himatione, 128, 159, 162, 166;
H. sanguinea, 146; HW. s.
freethii, 127

Hippobroma, 406

Histrio, 345; H. histrio, 345

Hodegia apatela, 497

Hoheria, 246, 409

Homalanthus polyandrus, 410

Homoptera, 501

Hoplurus (Oplurus), 333

Hovea chorizemifolia, 293

Humbertianthus, 419; H. car-
diostegius, 419

Humiriaceae, 87

Huthia, 424

Huttonella, 246

Hybanthus, 424; H. epacroides,
275; H. sprucei, 424

Hydnora, 87

Hydnoraceae, 87

Hydrocharitaceae, 56, 516

Hydrocotyle, 517

Hydrophyllaceae, 58, 424

Hymenanthera alpina, 243,
245; H. angustifolia, 248

Hymenophyllaceae, 438, 440,

Hymenoptera, 499, 500, 501
paipentcaceae, 414, 419, 587,
591

Hypericum, 414, 557, 571, 575;
H. keniense, 556, 557; H.
laricifolium, 580, 587, 591;
H. magniflorum, 591

Eipochaeris sessiliflora, 584,
5

Hyponomeutidae, 495
Hyrax,

Icacinaceae, 243

Icaridion nasutum, 499

Ichneumonidae, 499, 500

Icteridae, 159, 200

Icteronysius, 171

Mex Sis) 517s) 521, (525: | 576:
I. marquesensis, 521: I: sand-
penis, SPAN Tk tahitensis,

Inuleae, 240, 587

Ipomoea, 58, 74, 172, 424, 517;
if arborea, 424: I. caloden-
dron, 424

Iridaceae, 37, 16

Isachne, 56, 517

Ischaemum, 56, 517

Isodendrion, 519

Tsoétes, 587; I andicola, 587;
E gemmifera, 587

Isoplexis, 381, 414, 415

Isopogon alcicornis, D932. Of;
baxteri, 293; divergens,
286, 293; I. formosus, 293;
I. laticornis, 286; I. roseus,
293; JI. teretifolia, 293; I.
tridens, 293; I. villosus, 286

Isotoma, 406

Ixanthus, 395, 414, 415; I. vis-
cosus, 414

Jacaratia, 424; J. digitata, 424
Jacksonia, 272, DAG nd QLALA,

Index of Biological Names

272; J. capitata, 271; J. car-
duacea, 272; J. compressa,
223 if decumbens, 272;
J. eremodendron, 272, 273;
J. floribunda, 272: Lh foli-
osa, 276; J. furcellata, 273;
J. pteroclada, 272

Jacquemontia, 58, 517; J. sand-
wicensis, 462

Jasminocereus, 412

Jepsonia malvifolia, 25

Johnsonia, 302

Joinvillea, 57, 516

Juncaceae, 57, 516, 587, 590

Juania, 521

Kadua, 518,
Hedyotis)
Kennedya, 276

Kingia, 323
Kleidotoma subantarctica, 499,
500

523 (see also

Kleinia neriifolia, 414

Kochia, 257, 297, 299

Kokia, 28, 57, 146, 388, 405,
406, 418, 471, 472, 518; K.
cookii, 471; K. rockii, 471

Korthalsella, 57, 63, 80, 519

Labiatae (see Lamiaceae)

Labidura loveridgei, 605

Labordig. 58, 63, 474, 518,
523,

tees 2209: L. melisellensis,
604; E. sicula, 543, 604

Lachnostachys, 299; L. albi-
cans, 297, 306; L. ferrugi-
neus, 297, 298

Lactoris, 18, 411, 530

Lactuceae (= Cichorieae), 61,
206, 411, 459, 460

Lacuncularia, 88

Lageneria, 8

Lagenophora, 58, 517, 571

Lambertia echinata, 293

Lamiaceae, 58, 120, LODE Dil:
288, 297, 299, 411, 414, 422,
424, 468

Laminella, 128

Lasiopetalum, 297; L. drum-
mondii, 297, 306, 309; L.
floribundum, 282

Lathridiadae, 503, 504

Lathrinops furcifur, 338

Lauraceae, 57, 108, 267, 276,
415, 470, 518

Laurus indica, 366

Lavatera, 413; L. lindsayi, 366

Laxmannia, 417, 521

Fayia platyglossa, 61, 64, 455,

Lecocarpus, 412, 435, 454, 455;
L. pinnatifidus, 454, 455

Legendraea, 414

Leguminosae (see Fabaceae)

Leiolopisma lineo-cellatum,

14; L. suteri, 613; L. zelan-

dica, 613-14

Leiopelma, 4, 233, 251

Lentibulariaceae, 278

Lepechinia, 58, 518

627

Lepidium, 35, 57, 72, 404, 405,
517; L. arbuscula, 404; L.
owaihense, 74, 82; L. serra,
381, 404

Lepidoptera, 495-501

Lepinia, 12

Leptocarpus tenax, 278

Leptomeria exocarpoides, 275;
L. pauciflora, 271; L. spi-
nescens, 276; L. spinosa, 271

Lepturus, 56, 517

Leschenaultia (Lechenaultia),
300, 312, 432; L. juncea, 271;
L. striatum, 313

Leucophae, 192, 195; L. erio-
cephala, 192; L. gomerae,
192; L. macrostachys, 192;
L. massoniana, 192

Leucopogon fasciculatus, 527;
L. metaleucoides, 527, 530

Levenhookia, 253

Lhotzkya, 289

Libocedrus, 220

Liliaceae, 57, 277, 278, 286,
297, 302, 320, 322, 326, 327,
482, 516

Limnogale, 333

Limnophilidae, 504, 505

Limonia foliocuniculator, 612

Limonium, 182, 195, 414, 532;
L. aborescens, 385, 414; L.
fruticans, 414; L. preauxii,
414; L. rumicifolium, 385,
532; L. tuberculatum, 532

Liparis, 57, 61, 516

Lipochaeta, 13, 58, 59, 120,
130, 388, 404, 405, 455, 517,
535; L. lavarum, 454, 455;
L. remyi, 454; L. rockii, 454,
455

Lisianthus, 424; L. corymbo-
sus, 424; L. ovalis, 424; L.
revolutus, 424

Lithops, 342

Livistona, 156, 482

Loasaceae, 87

Lobelia, 35, 58, 59, 61-62, 106,
144, 146, 147, 373-75, 384,
404, 405, 408, 417, 517, 545,
554, 557-60, 559, 560, 592;
L. bequaertii, 558; L. cardi-
nalis, 60; L. cubana, 417;
L. deckenii, 558, 559; L. el-
gonensis, 558; L. gaudichau-
dii, 147, 389, 404, 407, 557;
L. gaudichaudii var. gloria-
montis, 171; L. gibberoa,
365-66, 554, 556, 557, 558,
565, 568, 581; L. hillebrandii,
147; L. hypoleuca, 147,
353; L. intermedia, 554; L.
keniensis, 375, 557-59, 559,
592; L. lanuriensis, 565; L.
oahuensis, 147; L. parvibrac-
teata, 581; L. petiolata, 554;
L. sattimae, 558; L. scaevo-
lifolia, 417; WL. squarrosa,
554; L. telekii, 560, 563; L.
tortuosa, 147; L. wollastonii,
565; L. yuccoides, 147

Lobeliaceae (see Lobelioideae)

(subfam. of
120, 129,

Lobelioideae
Campanulaceae),
130, 140, 144

Lodoicea, 431, 521

Loganiaceae, 58, 271, 518, 523,
525)

Logania nuda, 271

Lomandra, 321, 322

Loranthaceae, 57, 326

Loricaria thuyoides, 589

Loricera wollastoni, 502, 503

Lotus, 413; L. scoparius var.
dendriodeus, 369, 375, 413;
L. scoparius var. scoparius,
369, 37

Loudonia, 323

Loxocarya flexuosa, 278, 294

Loxops, 128, 159, 161, 166,
608; L. coccinea, 162; L.
parva, 159, J61, 165; L.
sagittirostris, 159, 162; L.
virens, 160, 166; L. v. stej-
negeri, 159, 163, 166; L. v.
virens, 159

Lupinus alopecuroides, 587,
588
Luzula, 57, 84, 516
tenuifolius,

Lycapsus, 411; L.
411

Lycium, 58, 75, 518; L. sand-
wichense, 76
Lycopersicon, 519
Lycopodiaceae, 320, 321
Lycopodium, 571, 584
Lygaeidae, 125, 126, 129, 170
Lygodesmia juncea, 459
Lyonothamnus, 413
Lyperanthus nigricans, 328,
329

Lysimachia, 58, 384, 406, 474,
518, 520; L. forbesii, 474;
L. kalalauensis, 395, 397,
405

Lytanthus, 414, 416; L. amyg-
dalifolius, 416; L. salicinus,
414

Lythraceae, 424

Maba, 82, 517, 524; M. ellip-
tica, 463; M. savaiiensis, 463
(see also Diospyros)

Macraea, 411, 412, 443, 454;
M. laricifolia, 411, 454

Macrocarpaea, 424; M. gutti-
fera, 424; MM. pachystyla,
424

Macropanesthia rhinoceros,
605

Macropidia, 297, 299, 300

Macropiper excelsum, 399,
400, 402, 409

Macrostelia, 419

Madia, 84, 455; M. nutans, 456

Madiinae, 147, 341

Malvaceae, 8, 57, 146, 243,
246, 275, 296, 297, 405-9,
413, 418, 424, 470, 514, 518

Mandibularca resinus, 337

Mangareva, 408

Mangifera indica, 429

Marcettella, 414, 415

Index of Biological Names

Margyracaena, 535

Margyricarpus setosus, 535

Marianthus_ erubescens, 276,
207.

Maripa, 424

Marsilea, 54, 61

Maschalocephalus
87

Masdevallia coriacea, 588, 592

Mathurina penduliflora, 421

Maunaloa, 430

Mayacaceae, 87

Mecodium recurvum, 439

Megalagrion, 611

Melaleuca, 286, 293, 327; M.
exarata, 323, 325, 327; M.
globifera, 286; M. mega-
cephala, 286; M. quinque-
nervia, 323; M. thymoides,
276

Melampodiinae, 412, 455

Melampodium, 455

Melanodendron, 416, 417; M.
integrifolium, 417

Melanoselinum, 232, 415, 416
(see also Tauschia); M. de-
cipiens, 416

Melhania, 417

Melicope, 154, 388, 476, 523,
524

Melicytus, 243; M. micranthus,
243, 245; M. ramiflorus, 243,
245

Mellissea begoniaefolia, 417

Menispermaceae, 57, 518, 524

Meryta, 73; M. lanceolata, 366;
M. sinclairii, 73

Metacottus, 340

Metophthalmus, 504

Metrosideros, 7, 58, 59, 62, 73,
105, 131, 134, 136, 138, 162,
163, 164 165, 166, 174, 248,
354, 406, 407, 514, 518, 520,
535, 571; M. perforata, 244,
246, 248; M. polymorpha,
19, 60, 100, 105

Mezoneurum, 76, 517; M.
kauaiensis, 366, 467

Microcachrys, 226

Microcarpus, 225

Microcyas, 18

Microlaena, 56, 517

Microlepia speluncae, 438

Micromeria, 193, 195, 535; M.
benthamii, 193; M. densi-
flora, 193; M. lanata, 193;
M. lepida, 193; M. linkii,
193; M. teydensis, 193; M.
tragothymus, 193

Micromyrtus, 289

Microphyllae, 469

Microphyura, 607

Microstrobus, 226

Microzamia, 323, 326; M. ried-
lei, 322

Minuarti, 406

Mirabelia spinosa, 276

Mirasolia, 412

Mniodes, 587

Moho braccatus, 144

628

dinklagei,

Mollusca (see names of families
and genera)

Monarrhenus, 421

Monimiaceae, 88

Monizia, 415, 417

Montezuma speciosissima, 418

Moraceae, 243, 420

Morinda, 58, 388, 408

Mucuna, 11, 75, 76, 81, 517;
M. gigantea, 467

Muhlenbeckia astonii, 243; M.
ephedroides, 245, 246

Munzothamnus (see Stepha-
nomeria blairii)

Musschia, 415; M. aurea, 415;
M. wollastonii, 415, 416

Myodocarpus, 230, 231, 232,
41; M. fraxinifolius, 230,
231; M. involucratus, 230,
231

Myoporaceae, 58

Myoporum, 58, 164, 169, 436

Myosotis, 571, 575

Myrsinaceae, 58, 243, 518

Mpyrsine, 58, 82, 518, aon
525; M. divaricata, 243;
knudsenii, 521, 522; Me

lanaiensis, 521; M: num-
mularia, 242 /
Myrtaceae, 7, 58, 105, 174,

226, 243, 244, 276, 286, 288,
289, 296, 299, 302, 303, 304,
307, 314, 315, 323, 325, 326,
3275422" 472

Myrtus pedunculata, 243

Nageia, 226

Naiadaceae, 56, 516

Naias, 56, 6

Nama, 58, 518

Narvalina, 418

Nectandra, 576

Nectarinia, 592; N. johnstonii,
558, 559

Negrestis, 58

Neocallitropsis, 225; N. comp-
tonii, 219

Neoceratias, 346

Neochanna, 250

Neogoodenia, 312

Neopanax, 243

Nepenthaceae, 226, 317

Nepenthes, 226

Neraudia, 57, 518

Nertera, 58, 523

Neseis, 125, 170, 171

Nesiota, 417

Nesocaryum, 529

Nesodyne koae, 173; N. rubes-
cens, 173

Nesodryas freycinetiae, 606

Nesoluma, 58, 518, 528, 530;
N. nadeaudi, 528; N. poly-
nesicum, 528; N. saint john-
ianum, 528

Nesophontes, 18

Nesoprosopis, 514, 612

Neurophyllodes (= Hawaiian
spp. of Geranium), 406

Neuroptera, 501

Newcombia, 128

Nicotiana, 424; N.. setchellii,
424: N. tomentosa, 424

Nidorella, 557

Nigella, 106

Nirarathamnos asarifolius, 420

Neohumbertiella, 419

Nothocestrum, 384, 388, 405-
7, 418, 518; N. breviflorum,
405; N. peltatum, 407

Nothofagus, 69, 233

Nothopanax, 14

Notospartium, 246

Nototrichium, 57, 384, 388,
390, 401, 404, 418

Numenius tahitensis, 67

Nuytsia, 326; N. floribunda,
423

Nyctaginaceae, 57, 172, 405,

406, 472, 518
Nycteribiidae, 26
Nysius, 125, 126, 127, 170, 171;
N. communis, 171

Oceanides, 170, 171

Ochrosia, 58, 76, 517; O. op-
positifolia, 442; O. parvi-
flora, 442; O. sandwicensis,

9

Ocotea, 576, 594; O. ken-
yensis, 557

Odynerus oahuensis, 612

Oecopthora pusilla, 504

Ogcocephalidae, 343, 345, 346

Olea, 58

Oleaceae, 58, 518

Olearia, 243, 409, 412, 422,
570; O. argophylla, 422; O.
ilicifolia, 248; O. odorata,
243; O. virgata, 243

Oliganthes, 419

Olisthopis maderensis, 502

Onagraceae, 407, 409, 571,
575

Oodemas, 164
Oparanthus, 212, 384, 408,
417, 443, 450, 451, 453; O.

albus, 407, 444, 450; O.
coriaceus, 450; O._ inter-
medius, 407; O. rapensis,

407, 450

Oplismenus, 56, 517

Oplurus (Hoplurus), 333

Opuntia, 203, 204, 364, 412,
Gi: 10: echios, 204, 412;
O. echios var. zacana, 204,
412; O. echios var. gigantea,
204; O. galapageia, 412; O.

megasperma, 204, 412; O.
megasperma, vat. mesophy-
tica, 204

Orchidaceae, 57, 306, 328, 516
Oreobolus, 56, 242, 516, 587,
590; O. furcatus, 148
Oreopanax tolimanus, 57
Orthoptera, 249, 501, 505
Osmanthus, 58, 518, 520; O.
sandwichensis, 366
Osmundaceae, 440
Ospatulus, 337; O. palaemo-
Ba gUs, 337; 0. truncatulus,

Index of Biological Names

Osteomeles, 57, 518

Ougapia apoana, 607;
nali, 607

Ozothamnus, 422

Paepalanthus, 587, 590

Palmae (see Arecaceae)

Palmeria, 159, 162

Pancheria, 231; P. communis,
2285 231) “PS insignis), 228;
231

Pandanaceae, 56, 516

Pandanus, 56, 217, 516, 524

Panicum, 56, 517

Papaveraceae, 413, 518

Papilio aegeus, 606

Papohus titan, 605

Papuacedrus, 336, 575

Paracottus, 340

Paradisaeidae, 610

Paradoxornithidae, 201

Parahebe, 238

Parasphenus pulchripes, 596,
597

O. ray-

Parasystatiella agrestis, 597

Paratrophis microphylla, 243

Pardosa lugubris, 487; P. pa-
lustris, 487

Paridae, 201

Parochlus steineni, 494

Parolinia, 381, 38 3, 414, 415;
P. ornata, 383, 402

Partula, 21, 31, 212, 488

Partulina, 127

Parulidae, 201

Parus teneriffae, 194

Paspalum, 56, 517

Pelargonium, 417; P. cotyle-
donis, 417

Pelea, 57, 120, 123, 154-55,

167, 388, 476-78, 518, 522—
524; P. adscendens, 154; P.
anisata, 154; P. clusiaefolia,
522°) UPA acruciatanw oA:

grandifolia, 154; P. hawai-
ensis, 154; P. orbiculata,
154; P. ovata, 154; P. pallida,
154; P. parvifolia var. ses-
silis, 154; P. pseudoanisata,

478; P. radiata, 154; P. re-
curvata, 154; P. sandwic-
ensis, 154; P. waialealae,

154, 407, 477; P. zahlbruck-
neri, 154
Pellaea ternifolia, 438
Pennantia corymbossa, 243
Pentaptilon careyi, 310, 311
Peperomia, 57, 63, 82, 120,
123 124" 533
Percidae, 339
Perdicella, 128
Perissodus, 339
Perityle incana, 413
Pernettya, 31, 527; P. nana,
242

Perrottetia, 57, 517

Persea americana, 429
Persoonia tortifolia, 293, 295
Perymenium, 412

Peryphus sjoestedti, 597
Petalura, 605

629

Petaluridae, 605

Petrobium (Laxmannia), 417,
$21

Petrophila, 289; P. biternata,
293; P. conifera, 286; P.
divaricata, 286; P. diversi-
folia, 293; P. drummondii,
293; P. linearis, 293, 295;
P. macrostachya, 293; P.
semifurcata, 286; P. stricta,
293, 295; P. trifida, 293

Petrotilapia tridentiger, 338

Peucedanum, 58, 436, 441, 517

Phalacrorax harrisii, 490

Pharmacus montana, 249

Phebalium, 288

Philippia, 557

Philotheca, 288

Phlebotaenia (= Polygala in
part), 418; P. cowellii, 418; P.
cuneata, 418

Phoenicoseris, 32, 206, 400-
401, 410

Phylica, 417

Phyllachne, 242

Phyllanthus, 57, 517

Phyllocladus, 226, 575; P.
hypophyllus, 570

Phylloglossum drummondii,
320, 321

Phyllostegia, 58, 120, 123, 468-
70, 518; P. racemosa, 66; P.
tahitensis, 468; P. variabilis,
468

Physalis, 417

Phytolacca, 57

Phytolaccaceae, 57

Pilea, 57, 518

Pilostyles, 13, 87

Pimelea, 300, 523, 530; P.
floribunda, 300, 301; P. phy-
Sones, 300, 301; P. prostrata,

42

Pinaroloxias, 199; P. inornata,
201

Pinguinus impennis, 490

Pinus attenuata, 327; P. canari-
ensis, 367, 608; P. muricata,
327; P. radiata, 327

Piora, 575

Piperaceae, 57, 120

Piptocoma, 418

Pipturus, 28, 31, 57, 63, 172,
173, 174, 518, 536

Piriqueta, 419, 421

Piscidia carthaginensis, 205

Pisonia, 57, 70, 71, 80, 172,
518; P. inermis, 406; P. sand-
wicensis (Rockia), 406, 473,
525; P. umbellifera (Ceodes),
406, 366, 472, 525

Pittosporaceae, 57, 247, 276,
277, 473
Pittosporum, 72, 80,

82,
243, 435, 518, 521, 1520) 525,
SBR Je acuminatum, 522; P.
anomalum, 243; P. cauli-
florum, 473; P. conferti-
florum, 473; P. crassicaule,
243, 247; P. divaricatum,
243; P. dolosum, 473; P.

Pittosporum (Cont.)
hawaiiense, 473; P. hosmeri,
473; P. lineare, 243; P. mega-
carpum, 474; P. moluc-
canum, 474; P. niueana,
473; P. orofenensis, 521; P.
rapense, 474; PP. raro-
tongense, 473-74; P. resinif-
erum, 474; P. rigidum, 243;
P. salcatum, 473; P. taitense,
473, 521; P. tenuifolium,
473, 474

Pityrodia oldfieldii, 297, 298

Pityrogramma calomelanos, 438

Placostylus, 332, 607

Pladaroxylon (also called Sen-
ecio), 417

Plagianthus, 294; P. betulinus,
409; P. divaricatus, 243; P.
helmsii, 294, 296, 297; P.
incanus, 275; P. microphylla,
275

Plagithmysus, 167; P. darwin-
ianus, 171; P. lamarckianus,
171; P. varians, 171, 172

Planamastra, 128, 177

Planchonella, 58, 226, 480,
518, 528; P. costada (also
called P. vitiensis), 481; P.
garberi, 481; P. novozeland-
ica, 481; P. sandwicensis, 481

Plantaginaceae, 35, 58, 384,
405, 410, 414, 417, 518

Plantago, 35, 58, 196, 384, 389,
395, 410, 414, 417, 424, 432,
Silks SBVy Se ey Sybe Je
arborescens, 195, 366, 367,
383, 414; P. aundensis, 571,
572, 574; P. fernandezi-
anum, 410; P. grayana, 533,
534; P. hillebrandii, 535;
krajinai, 583; P. lanceolata,
533; P. pachyphylla, 535;
P. princeps, 384, 405; P.
rigida, 587; P. robusta, 417;
P. sericea, 424; P..webbii,
195, 364, 383, 414

Platydesma, 57, 476, 477, 478,
518; P. campanulatum (also
called P. spathulatum), 478;
P. cornuta, 388, 478; P.
remyi, 477, 478

Platysace compressa, 270, 272

Platyspiza, 199

Platytheca, 289

Plectocephalus, 457

Plectranthus, 58, 518

Pleiosoma, 504

Pleiospilos, 342

Pleomele, 483, 516; P. aurea,
483; P. fernaldii, 483

Plocamotrechus kilimanus, 597,
599

Plumbaginaceae, 58, 518

Plumbago, 58, 518

Poa, 56, 84, 517

Poaceae, 56, 171, 516, 517,
567
Podacanthus_ wilkinsoni, 605

Podocarpaceae, 215, 220, 221,

Index of Biological Names

222, 224, 225, 226, 228, 232,
246

Podocarpus, 215, 223, 224,
225, 226, 234, 235, 553, 575,
576; P. alpina, 235; P.
andinus, 235; P. brassii, 570;
P. brevifolia, 570; P. cupres-
sina, 570; P. dacrydioides,
234; P. ferrugineus, 69, 235,

441; P. gracilior, 553; P.
hallii, 235; P. imbricatus,
570, 571, 574; P. macro-

phylla, 225; P. milanjianus,
554; P. minor (also called
P. palustris), 223, 224, 225,
441 P: nivalis, 235; P. spi-
catus, 235; P. totara, 235;
P. ustus, 224, 225, 234
Polemoniaceae, 424
Polygala (see Phlebotaenia)
Polygalaceae, 268, 269, 271,
274, 276, 277, 290, 418
Polygonaceae, 57, 243, 245,
405, 406, 413, 414
Polygonum, 57, 518
Polypodiaceae, 157
Polypodiopsis, 225, 226
Polyporus, 172

Portulaca, 57, 518; P. lutea,
qe 7 SeXs

Portulacaceae, 57, 333, 398,
413

Potamogeton, 56, 336, 517

Potamogetonaceae, 56

Potarophytum, 87

Potentilla papuana, 571, 573,
574

Poterium, 415

Pouteria (see also Plancho-
nella), 518, 528

Prasieae, 468, 469

Primulaceae, 58, 397, 405, 406,
474

Pringlea antiscorbutica, 496,
497

Pringleophaga kerguelensis,
496, 497

Pritchardia, 56, 127, 156, 162,
A432 A89" 483" S16. 541: 9P
beccariana, 156; P. low-
reyana, 482; P. macdanielsii,
156; P. martii, 156, 482; P.

montis-kea, 482; P. pacifica,
482; P. remota, 156, 482;
P. rockii, 482; P. thurstonii,
482

Procottus, 340

Pronaya elegans, 276

Prosopanche, 87

Prosopis, 553

Prostanthera, 422; P. lasian-
thos, 422; P. melissifolia,
EDIE J 22. rotundifolia, A22= IP:
tysoni, 275

Proteaceae, 264, 267, 268, 275,
280, 284, 286, 287, 291, 292,
293, 295, 297, 302, 306, 307,
308, 324, 325, 326

Proterhinus, 612

Proteurya, 524

Prunus spp., 372

630

Psammomoya choretroides,
71; P. ephedroides, 276
Pecudo cece ieee (Homoptera),

Pseudomorus, 57, 518
Pseudonestor, 164
Pseudopanax, 246, 248
Psiadia, 417, 418; 554; P. al-
tissima, 419; P. rotundifolia,

417

Psidium, 28, 62

Psilotum, 294

Psittarostra, 100, 105, 128, 161,
164, 165, 167, 168, 465; P.
bailleui, 128, 165, 168; P.
cantans, 127, 128, 167; P.
flaviceps, 128; P. kona, 128,
169; P. palmeri, 128; P. psit-
tacea, 128, 165, 168

Psittarostrinae, 105, 159, 161,
163, 164, 165, 167, 168, 169

Psorospermum, 419

Psychotria, 58, 82, 518

Pteralyxia, 58, 441, 442, 517;
P. kauaiensis, 441; P. ma-
crocarpa, 441

Pteridium, 268

Pteris excelsa, 438; P. lidgatii,
439; P. vittata, 438

Pterocephalus, 193, 195, 381,
5} : dumetorum,
193, 382, "414: P. lasiosper-
ma, 193; P. virens, 193

Pterodiscus, 128, 177

Ptilotus (Trichinium), 257,
299, 300; P. rotundifolius,
297, 298, 300

Puya, 424, 592, 593; P. cuatre-
casasii, 592, 593; P. hamata,
592, 593; P. nitida, 592, 593;
P. raimondii, 424; P. trianae,
592, 593

Pygmaea, 242

Pyralidae, 495

Quasianomalipes group, 174
Quercus spp., 372

Radema, 504

Rafflesiaceae, 13, 87, 256

Raillardella argentea, 147; R.
scabrida, 147

Railliardia, 58, 119, 120, 123,
124, 126, 147, 149, 150-52,
455, 535 (see ‘also ‘Dubautia

spp.)

Rallidae, 489, 490, 491

Ranunculaceae, 57, 106, 474

Ranunculus, 57, 84, 249, 518,
550, 567, 571, 575; R. fibro-
sus, 474; R. hawaiiensis, 474;
R. mauiensis, 474; R. plebius,
474; R. sieboldii, 474

Raoulia, 240, 242, 587; R. aus-
tralis, 240, 241; R. eximia,
242, 587

Raouliopsis, 587, 589; R. sei-
frizii, 589

Rapa, 407

Rapateaceae, 87

Ratitae, 489-91

Rauwolfia, 58, 517

Rea, 206, 410

Regulus ignicapillus, 47; R.
regulus, 47

Remya, 28, 58, 404, 405, 460,
517

Restiaceae, 256, 271, 272

Restio megalotheca, 294

Reynoldsia, 58, 130, 360, 517

Rhacodactylus, 232

Rhamnaceae, 57, 67, 164, 243,
247, 275, 417, 436, 474, 518

Rhaphidophoridae, 249

Rhetinodendron oberteri, 32,
206, 208, 209, 381, 410, 457,
458, 520

Rhipsalis, 86

Rhizanthella gardneri, 306, 309

Rhizophora, 16-117

Rhus, 57, 80, 517

Rhynochaetos jubatus, 232

Rhynchospora, 565 15163) UR:
ruiziana, 590

Rhytidophyllum, 418

Richea pandanifolia, 423

Riopa bocourti, 232

Robinsonia, 23, 206, 207, 210,
211, 227, 381, 410, 458, 520;
R. evenia, 206, 207,
210; R. gayana, 206, 209,
210, 457; R. gracilis,
208, 210; R. masafuerae, 207,
208, 210; R. thurifera, 209,
210, 410

Rockia (see Pisonia sandwic-
ensis)

Rollandia micropterum (grebe;
also called Centropelma mi-
cropterum), 490

Rollandia (lobelioid), 28, 58,
64, 102, 119, 120, 124, 139-
43, 140, 384, 404, 405, 517;
R. angustifolia, 140, 142; R.
calycina, 139, 142, 143; R.
crispa, 139, 142; R. hum-
boldtiana, 139; R. kaalae,
142; R. lanceolata, 139, 142;
R. lanceolata var. tomen-
tella, 142: R. longiflora, 142;
R. pinnatifida, 140, 142, 143;
R. purpurellifolia, 142; R.
saintjohnii, 140, 142, 143; R.
waianeensis, 142

Rosaceae, 57, 242, 415, 562,
567, 568, 584, 586
Rubiaceae, 58, 155, 212, 239,

242, 243, 406, 408, 413, 417,
476, 518, 519, 570, 590, 591

Rubus, 28, 57, 84, 248, 476,
518, 571; R. australis, 476;
R. hawaiiensis, 476; R. ma-
craei, 476; R. squarrosus,
248

Rulingia, 297; R. grandiflora,
282

Rumex, 35, 57, 406, 414, 518;
R. giganteus, 62, 405; R
lunaria, 414

Rumohra_ adiantiforme, 438;
R. aristata, 438; R. curvi-
folia, 439; R. hispida, 438

Index of Biological Names

Ruppia, 56

Ruppiaceae, 56

Rutaceae, 57, 120, 154, 271,
276, 279, 282, 288, 297, 299,
300, 301, 406, 476

Sadleria, 152, 157, 440; S. cy-
atheoides, 157, 439; S. pal-
lida, 439: S§. souleyetiana,
439; S. squarrosa, 157, 439;
S. unisora, 439

Sagina, 57

Saldula, 611

Saltia acrophylax, 597

Salvia, 424

Sanctambrosia, 529

Sanguisorba, 415

Sanicula, 58, 63, 84, 441, 517

Santalaceae, 57, 271, 275, 478,
518

Santalum, 57, 65, 164, 478,
518, 525; S. album, 478; S.
cuneatum, 478; S. ellipticum,
80, 478; S. haleakalae, 521;
S. littorale, 478; S. pyrular-
ium, 478

Sapindaceae, 57, 65, 68, 130,
152, 256-57, 436, 478, 479

Sapindus, 57, 518; S. mukoros-
si, 480; S. oahuensis, 480; S.
saponaria, 480

Sapotaceae, 58, 226, 480

Sarcolaenaceae, 334, 419

Seco Uy aT 212, 388, 408,
47

Sargassum, 345

Sarraceniaceae, 317

Saurauia, 576

Saxegothaea, 595

Saxifragaceae, 57, 243, 576

Scaevola, 31, 55, 58, 65, 157,
231, 268, 282, 312, 314, 329,
358, 384, 399, 405-9, 421,
423, 518, 519, 536; S. bal-
ansae, 409; S. beckii, 229,

231, 409; S. calendulacea,
312, 313, 314; S. chamis-
soniana, 157; S. coccinea,

409; S. collaris, 313, 314; S.
Ccoriaceas 28) 73, iol, 399;
405, 468; S. crassifolia, 279,
282, 399; S. cylindrica, 409;
S. floribunda, 366, 409; S.
gaudichaudiana, 66, 157,
366, 467, 536; S. glabra, 157,
399, 467-68, 536; S. hookeri,
312, 313; SS. indigofera,
409; S. kauaiensis, 467; S.
micrantha, 409; S. mollis,
467, 536; S. montana, 229,
231, 409; S. nitida, 282; S.
nubigena, 408; S. oppositi-
folia, 409; S. oxyclona, 274,
275, 322, 323; S. pauciflora,
409; S. plumieri, 312; S.
porocarya, 282; S. procera,
536; S. racemigera, 409; S.
restiacea, 268, 269, 271, 275;
S. sericea, 519: S. spinescens,
274, 275, 289, 313, 314, 399,
423, 468; S. taccada, 312,

631

359, 467, 468, 519, 536; S.
tahitensis, 427; S. tortuosa,
271, 275, 294; S. verticillata,
4

09

Scalesia, 197, 202-5, 364, 368,
388, 411, 412, 453, 454; S.
affinis, 203, 443; S. aspera,
203, 453, 454; S. atractyloi-
des, 203, 412; S. cordata,
203, 412, 453; S. crockeri,
203; S. darwinii, 412; S. hel-
leri, 203; S. incisa, 203; S.
microcephala, 202, 203, 364,
453, 454; S. pedunculata,
202, 203, 360, 364, 388, 411,
412; S. stewartii, 203, 412;
S. villosa, 203

Scarites, 605, 606

Schiedea, 57, 404, 406, 461,
462, 517, 529, 530, 540; S.
amplexicaulis, 461; S. dif-
fusa, 461; S. globosa, 461,
527, 529; S. kealiae, 461; S.
ligustrina, 529; lychnoi-
des, 461; S. menziesii, 529;
S. viscosa, 461

Schinus, 28, 62

Schizaea dichotoma, 438; S.
digitata, 439; S. skottsbergii,
437, 439; S. skottsbergii var.
mauiensis, 437, 439

Schizaeaceae, 438

Schizogyne sericea, 414

Schoenophilus pedestris, 497,
498

Scincidae, 8, 335, 609

Scirpus, 56, 84, 516

Scleranthus, 240, 241, 242

Scleria, 56, 516

Sclerotheca, 212, 408; S. ar-
borea, 407; S. margaretae,
407; S. viridiflora, 407

Scrophulariaceae, 58, 235, 236,
242, 409, 410, 414, 416, 518,
Sil 590, 591

Scyphostegia, SHS)

Sedum, 570

Selaginella, 61

Selenothamnus, 294, 296

Selkirkia berteroi, 410

Senecio, 102, 106, 384, 403,
409, 410, 413, 417, 419, 421-
23, 554, 558, 561-564, 584;
S. antennaria, 521; S. bras-
sica, 558, 561, 562; S. bras-
sicaeformis, 562; S. canes-
cens subsp. boyacensis, 584,
585, 587; S. carolii-tercii,
576; S. centropappus, 422; S.
cottonii, 563, 564; S. sect.
Dendrosenecio, 423, 562;
S. diclinous, 521; S. erici-
rosenii, 565; S. friesiorum,
562; S. humblotii, 419;
huntii, 409, 417; S. kenio-
dendron, 389, 562-65, 564;
S. lyoni, 413; S. macrantha,
457, 458; S. meruensis, 561,
565; S. myricaefolius, 419;
S. niveo-aureus, 585, 587; S.
palmeri, 413; S. pulchellus,

Senecio (Cont.)

76; S. rapae, 521; S. rein-
oldii, 409; S. rigidifolius, 585,
590; S. stewartiae, 409; S.
vaccinioides, 576

Senecioneae, 61, 206, 208, 209,
411, 458, 459

Sesbania, 57, 517

Sesuvium, 57, 517; S. portula-
castrum, 74

Sholtzia, 289

Sicyos, 58, 462, 517

Sida, 57, 418, 518; S. eggersii,
388, 418

Sideritis, 192

Sideroxylon (see Planchonella),
480, 518; S. rhynchosperm-

um, 481
Silene, 57, 151, 152, 404, 517;
S. alexandri, 405; S. struth-
ioloides, 364, 405
Sinapidendron, 381, 414-16
Siphocampylus, 424
Sisyrinchium, 57, 152, 516
Smilax, 57, 516, 524
Solanaceae, 58, 297, 299, 405,
406, 414, 417, 418, 422, 424,

554

Solanum, 58, 422, 424, 518,
554; S. kauaiense, 405; S
laciniata, 422; S. nelsoni, 80;
S. verbascifolium, 424

Solenodon, 18

Sollya fusiformis, 276

Sonchus, 35, 182, 195, 353, 362,
375, 376, 377, 378, 379, 380,
381, 383, 389, 391, 392, 393,
401, 414-16, 535; S. abbre-
viatus, 375, 378, 392, 393,
403, 414, 415; S. brachylobus,
376, 392; S. congestus, 376,
392; S. daltoni, 416; S. fru-
ticosus, 379, 389, 392, 393,
415, 416; S. gorgadensis,
416; S. jacquinii, 378,

; S. leptocephalus, 376,
380, 390, 392; S. oleraceus,
375; S. pinnatus, 379, 392,
415, 416; S. platylepis, 376,
392; S. radicatus, 377, 381,
401; S. ustulatus, 376, 392,
415

Sophora, 57, 76, 164, 171, 466,
517; S. chrysophylla, 466,
467; S. microphylla, 466;
S. prostrata, 466; S. tetrap-
tera, 466; S. tomentosa, 517

aan clo air lanceolatum,

Sphaerolobium alatum, 272,
273; S. vimineum, 271
Sphenodon, 4, 233, 251
Sphenotoma gracile, 294
Sporobolus, 56, 517
Spratellicyparis palata, 336
Stachycarpus, 225
Steganotaenia, 554
Stenogyne, 58, 120, 123, 432,
435, 468, 469, 469, 470, 518;
S. calaminthoides, 469; S. dif-
fusa, 469; S. kaalae, 469; S.

Index of Biological Names

kamehamehae, 469; S. pur-
purea, 469; S. rotundifolia,
470

Stenopadus, 405

Stephanomeria, 375, 381, 413;
S. blairii, 366, 369, 374, 413,
459; S. guadalupense, 413;
S. virgata, 374

Sterculiaceae, 57, 282, 297,
306, 417, 518

Sticherus owyhensis, 439

Stifftia, 405

Strigops habroptilus, 490

Stipa, 84

Stirlingia, 286, 294; S. tenui-
folia, 292

Strasburgeria, 481

Straussia, 58, 518, 525

Streblidae, 26

Strongylodon, 57, 518; S. lu-

cidus, 467
253, 268, 271,

Stylidiaceae,
309, 423

Stylidium, 253, 254, 257, 258,
268, 286, 293, 316, 321, 323;
S. amoenum, 286; S. bulbif-
erum, 321; 8S. falcatum,
323; S. galioides, 316, 472;
S. glandulosum, 423; S. im-
bricatum, 258, 423; S. insen-
sitivum, 257; S. junceum,
271, 276; S. laciniatum, 276;
S. laricifolium, 423; S.
maitlandianum, 286; S. pe-
tiolare, 321; S. preissii, 258;
S. rigidifolium, 286; S. squa-
mosotuberosum, 271

Stylites, 587

Styphelia, 58, 62, 130, 152,
Si Set Sil SYES SYER AY
tameiameiae, 526, 527

Suriana, 76-77

Swertia, 562, 575

Symphiobasis, 310; S. macro-
plectra, 311

Symphyochaeta macrocephala,
206, 207, 208, 210, 381, 410,
457, 458, 520

Synaphaea, 291; S. decorticans,
291, 292; S. petiolaris, 293

Syzygium, 58, 142, 518, 519

Talinum, 413; T. guadalupense,
366, 384, 390, 398, 399, 413
Tarchonanthus, 554
Tarphius, 503
Tarsipes spenseri, 309
Tasmannia (see Drimys)
Tauschia, 415, 416; T. decipiens
(Melanoselinum decipiens),
368, 373, 374, 415, 416; T
edulis (Monizia edulis), 415;
T. insulare (Melanoselinum
insulare), 416
Taxaceae, 215, 225
Taxodiaceae, 226
Taxus, 225
Telmatochromis, 339
Templetonia, 272; T.
272; T. sulcata, 272
Tenrecidae, 333

632

egena,

Tephrosia, 518

Terminalia, 76-77

Ternstroemia, 576

Tetramolopium, 58, 460, 517,
pe 574, 575, 584; T. remyi,

Tetraplasandra, 58, 434, 442,
443, 517, 520, 540; T. hawai-
iensis, 443; T. kauaiensis,
442, 443; T. koerdersii, 443;
T. meiandra, 443; T. oahu-
ensis, 442; T. paucidens,
443; T. philippinensis, 442,
443; T. waialealae, 443; T.
waianensis, 443; T. waimeae,
366, 442, 443

Tetratheca, 268, 288, 321; T.
affinis, 270, 272; T. confer-
tifolia, 288, 290; T. efoliata,
269, 271, 294; T. setigera,
288, 290

Thalassophilus, 502

Thamastes, 504

Thamnoseris, 411; T. lacerata,
411, 459

Theaceae, 57, 518, 576

Thelymitra flexuosa, 278; T.
spiralis, 278, 294

Thespesia, 418

Thlaspi, 570; T. alliaceum, 569

Thomasia, 297; T. quercifolia,
282; T. solanacea, 282

Thraupidae, 159, 200

Thryptomene, 289

Thylacinus, 253

Thyrsopteris, 18, 411

Thymeleaceae, 300, 30/, 523,
530, 536, 571-75

Thysanotus, 278, 309; T. di-
chotomus, 286; T. patersonii,
277, 278

Tilapia, 338, 339; T. lidole,
338; T. saka, 338; T. squa-
mipinnis, 338

Timonius polygamus, 476

Tinearupa, 495; T. sorenseni,

Tipula subaptera, 597, 598

Tipulidae, 612

Titanochaeta, 172, 612

Tithymalus, 195, 391, 414

Tolpis, 414, 416; T. succulenta,
415, 416

Tornatellinidae, 27, 127

Touchardia, 57, 518, 525

Traversia, 409

Trema, 57, 518

Tremandra, 282

Tremandraceae, 268, 269, 270,
271, 272, 282, 289, 290, 321

Trematolobelia, 32, 58, 59, 62,
63, 109, 144, 146, 373, 384,
404, 435, 461, 517; T. ma-
crostachys, 109

Tribonyx mortieri, 491

Tribulus, 57, 167, 519

Trichinium (see Ptilotus)

Trigonella laciniata, 89

Trisetum, 56, 517

Tropheus, 339

Tropidurus, 202
Turmeraceae, 27, 419, 421

Ulmaceae, 517, 518

Umbelliferae (see Apiaceae)

Uncinia, 56, 63, 64, 516; U.
tenella, 64

Urera, 57, 518, 525

Urocyon littoralis, 19

Uroplatidae, 333

Urticaceae, 57, 172-74, 248,
414, 518

Urtica ferox, 248

Uta, 613; U. stansburiana, 335

Utricularia, 317, 319; U. volu-
bilis, 278

Vaccinium, 58, 63, 80, 84, 151,
152, 517, 520, 571, 574-76;
V. floribundum, 576; V. var-
ingtifolium, 581

Valeriana (Valerianaceae),
424; V. karstenii, 589, 590

Vallesia, 441

Vandenboschia cyrtotheca,

38; V. davallioides, 438;
V. draytoniana, 439

Vangidae, 333

Velleia, 310; V. montana, 312;
V. paradoxa, 311

Velloziaceae, 87

Verbenaceae, 58, 288, 297, 298,
299, 306

Verbesina, 405

Verbesininae, 405, 412, 454

Index of Biological Names

Vernonia, 419, 554
Vernoniopsis, 419; V. caudata,
419

Veronica, 575

Verreauxia, 423; V. paniculata,
310, 331; V. reinwardtii, 297

Verticordia, 307, 309; V. in-
signis, 307; V. mitchelliana,
307

Vestiaria, 144, 159, 162, 166

Vicia, 57, 518

Vierea laevigata, 414

Vigna, 57, 518

Viguiera, 453, 454; V. linearis,
454

Vincentia, 56, 516

Viola, 22, 31, 84, 405, 406,
424, 519, 520, 550, 562; V.
robusta, 387; V._ trachelii-
folia, 387, 395, 397, 405

Violaceae, 243, 245, 248, 275,
405, 424

Viscaceae, 57, 519

Vitex, 508

Vochysiaceae, 87

Wahlenbergia, 410, 417, 535;
W. grahamae, 535; W. lar-
raini, 535

Waltheria, 57, 518

Wedelia, 405, 412, 454; W. uni-
flora, 455

Wehlia, 289

Weinmannia, 216, 354; W.
fagaroides, 580

633

Westringia angustifolia, 422

Wigandia, 424; W. crispa, 424;
W. urens, 424

Malsiroenia, 152, 518, 523,

Wilkesia gymnoxiphium, 32, 58,
61, 147-51, 184, 186, 191,
365, 366, 373, 404, 405, 445,
455, 456, 457, 517, 542°

Winteraceae, 57, 216, 226,
232, 481, 570

Withania, 414, 554

Wulffia, 412

Xanthium strumarium, 364
Manthomhoea, 322, 393) 326,

Xanthorrhoeaceae, 326
Xanthosia, 297

Xenicus, 251; X. lyallii, 490
Xylomelum occidentale, 293
Xylosma, 58, 518, 524

Yunquea tenzii, 32, 34, 374,
410, 435, 457-460

Zanthoxylum, 57, 68, 82, 476,
477, 518, 525; Z. kauaiensis,
68; Z. oahuensis, 68

Zealandrosardus, 249

Zexmenia, 405; Z. brevifolia,
405, 454, 455

Zosterops, 211

Zygogynum, 216, 226, 482

Zygophyllaceae, 57, 167, 519

SUBJECT INDEX

Numbers in italics refer to photographs. Certain geographical localities which are

referred to repeatedly have been abbreviated: C.Is. Canary Islands;

H.Is. Hawaiian Islands;
N.G. New Guinea;

island; S.W.A. Southwestern Australia;

pagos Islands;
New Caledonia;

Abyssal depths, 343-46; fish
life adaptations, 346; lumi-
nescent specialization in,
346; type of habitat, 343-44

Acacia: diversity of growth
forms, Australia, 269; mor-
phological aspects, 465-66

Acaulescent habit, 148, 381,
578, 579 (see also Cushion-
plant habit)

Achatinellids (Achatinella)
(see also Land snails); adap-
tive radiation, H.Is., 127;
genetic drift, 176; gigantism,
607; host-plant fidelity, 176,
488

Achenes, of some Asteraceae,
compared, 454, 456, 457,
459

Acicular leaves, 283, 286, 289
(see also Leaf types)

Adaptation, conceptual aspects,
20-25, 97-98, 170, 603-16

Adaptive radiation, process of,
97-116; allopatry of plants,
114; of angiosperms and land
snails, H.Is., compared, 127;
in archipelagos, 108-9, 114—
15; autochthonous develop-
ment, 102; behavioral as-
pects, in fishes, 343, 346; of
birds on islands, 103, 108,
128-29, 607-8; cataclysmic
events, 341; competition, 220;
in continental biotas, 108,
332-42; ecological factors,
21, 26-27, 100-101, 202, 204;

endemism and, 102, 114-15;
of fishes (see Cichlid fishes;
Cottid fishes; Cyprinid fishes;
Pediculate fishes); of flower-
ing plants (see Angiosperms;
also. specific genera and
species in “Index of Biologi-
cal Names’); of gekkonid
lizards, N.C., 231-32; of
geospizids, G.Is., 198-220;
harmonic biota factors, 193-
94, 227; in honeycreepers,
H.Is. (see Drepanids); in
insect behavior, 174-76
in insect groups: in HLIs.,
169-76; in Macaronesia,
194-95; in N.Z., 249
insular biota factors, 21-22,
353 (see also Insular biota
patterns); isolation a factor,
103, 171-72, 340, 511; land
area requirements, 103, 538;
in late arrivals on islands,
108; of lizards on islands
(see Lizards); long-distance
dispersal as stimulus, 510;
low in some _ Polynesian
islands, 4, 7, 211-12 (see also
Continental island biotas);
maximal differentiation, 100—
101; secondary cycles 100;
“small island” effect, 206,
538; ‘“supralimital’” specia-
tion, 106-7; time require-
ments, 103-4
Aeonium species, adaptive radi-
ation, Madeira Is., 191-92

634

J.F.Is. Juan Fernandez Islands;
N.Z. New Zealand; S.H. Saint Helena
W.A. Western Australia.

GilsarGala=
NE@:

Aerial dispersal, 9, 18, 32, 50-
2. , WS 352 asim
Afroalpine flora, 568; of
algae and fungi, 52; as ar-
rival mode, comparison, 53-
55; as dispersal mechanism,
assessed, 77—78; distance from
source areas, 15, 51, 78; in
equatorial highlands, 552; of
fern spores to H.Is., 60-61;
of flora to H.Is., 18, 59-63,
78; insect distribution by, 9,
46, 51; to Rapa I., 79; to
Revillagigedo Is., 77, 79; to
San Clemente I., 77, 79; of
spiders, 46, 51, 487; of
spores, 46, 51

Aeroplankton, studies of, 51

Aestivation (air-breathing), of
galaxiad fishes, N.Z., 250

Africa (see also Afroalpine
flora; East African high-
lands): amphitropical distri-
butions with, 86-89; conifers
not shared with, 88; equato-
rial highland zones, 550;
Macaronesian ties with, 191-
93

Afroalpine flora (see also East
African highlands; Equa-
torial highlands): adapta-
tions and characteristics, Mt.
Kenya, 566; affinities and
sources, 569-70; bird polli-
nation, 569; color factors,
568; dispersal, 551-52; en-
demic elements, 568-69; fire

survival, 568; frost survival,
563-67; herbivore pressures,
568; Himalayan affinities,
569; inflorescences, 563, 565,
574

increased woodiness: lower
elevations, Mt. Kenya,
556; plateau Tegion. of
Kenya, 555
Lobelia spp. from high eleva-
tions, 559, 560; microphylly,
567; pollination patterns,
568-69; Senecio (Aster-
aceae) spp., 561, 564; xero-
phytic adaptations, 567

Afroalpine insects, flight-ability
loss, 596-98

Air-breathing fishes (Neo-
channa), N.Z., 250

Air dispersal (see Aerial dis-
persal; Wind dispersal)

Albatross, 73, 76

Alectryon (Sapindaceae): affin-
fties, 479; fruit and seeds
compared, 479

Algae, dispersal by birds, 52

Allopatry, in adaptive radiation,
114 (see also Sympatry)

Alpine environment (see Equa-
torial alpine habitat)

Altitude: adaptation to, 558,
562; and angiosperm varia-
tion, G.Is., 203-5; and cara-
bid beetle color variation,
503; rainfall and, 359

Amadon, D. on_ secondary
adaptive radiation, 100

Amastrid snails (Amastridae):
adaptive radiation, H_Is.,
127-28, 176-77, 340; distri-
bution, H.Is., 128; isolating
mechanisms, 340

American flora, dispersed to
Polynesia, 212

Amphitropical disjunctions (see
Intercontinental disjunctions)

Amsterdam Island, moth-wing
Teduction on, 497

Ancestral ype, onecral role
of, 429,

Andean ent Paae.
belioids in, 360

Andesite line, 10, 45, 215, 434

Andromonoecious species, 516—
19, 531, 533

Anemochory, 552
Aerial dispersal;
persal)

Anemophily (see Wind pollina-
tion)

Angiosperms (see also Dicoty-
ledons; Flora; Monocotyle-
dons; particular families and
genera): adaptive diversity
and growth forms, 109-16,
370; African-American dis-
tribution, 88; Australasian—
South American distribution,
88; Australian overland en-
try, 256
competition

woody lo-

(see also
Wind dis-

with gymno-

Subject Index

sperm flora: in Australia,
226; in N.C., 220, 225
conspicuous aspect, 539; dis-
tribution of species to H.Is.,
120; distribution of taxo-
nomic types to islands, 112;
diversity, tabulated, 112;
early forms, in lava flows,
60, 62; endemic speciation,
61, 112, 512-13; evolution-
ary plasticity, 278 ff.; families
and genera, computed by
areas, 112; Bynulespenie dis-
placement 'by, 2 25-26; hydro-
phyte habit, ‘compared by
island biotas, 370 (see also
specific plant. growth habits);
large seed types not on
islands, 353, 354; life-form
spectra, table, 370; natural
selection in growth forms
and habits, 367-73; Pacific
boundary of dispersal, 10,
215-16; plasticity for wood-
iness, 351; seed dispersal, 50—
90 (see also Plant dispersal;
Seed dispersal by birds);
species per genera, computa-
tion by regions, 112; stem-
succulents, in plant-form
spectra, by islands, 371 (see
also Plant growth types; Stem
succulents); therophytes, in
plant-form spectra, by
islands, 371 (see also Plant
growth types; Therophytes);
vessel dimensions in insular
species, and ecological cor-
relations, 392
Anglerfishes (Pediculati), 343-
Annual plant, growth forms,
3
Anoline lizards (Anolis), 11,
47, 98, 543 (see also Lizards)
Antarctic: entomology, 494
(see also Subantarctic);
land-bridge connection hy-

pothesis, 7, 233; midge, flight-
less, 498
Antennarius (Antennariidae)

(frogfishes), adaptive radia-
tion, 343-45
Antillean insectivores, 104
Ants: adaptive radiation, in
J.F.Is., 211; long-distance
rafting, 11, 14; in N.C., 231-
32; predation, by sightless
insects, 503-4
Aphylly, in thorn
S.W.A., 275-76
Apiaceae, dispersibility, 441
BD OCD ACERS, dispersibility,
4

Apomixis, 148, 519 (see also
Self-pollination)

Appendages, modifications of,
607-10 (see also Body size
changes)

Aptery, 492, 494, 502 (see also
Insect flight loss)

635

shrubs,

Araliads (Araliaceae): arbo-
rescence, 29; dispersibility,
441-42; lava pioneers, 13-14

Araucaria _(Araucariaceae) :
diversification in N.C., 34,
215, 217; species with distinc-
tive habits and habitats, 2/8,
219

Arborescence (see also Insular
woodiness) : archipelago
effect favors, 372; bog habitat
unfavorable, 559; climatic
factors influencing, 352-53,
355-61; in cloud-forest habi-
tat, 360; competition for
light, 364; on continental
beaches, 406; ecological shift,
361-62, 373; in East African
highlands, 553, 554, 557-58
in herbaceous plant groups,

358; on CIs., 189, 192,
372, 373, 383, 385, 388,
413-15; on G.Is., 202-5,
360, 364, 411-12; on H_Is.,
123, 366, 381, 384, 387,
388, 404-7; on J.F.Is., 206—
10, 371-72, 384, 410-11;
on Macaronesian Is., 182,
184, 371, 381, 415-16; on
N.C., 409; on N.Z., 357,
409-10; on Socotra, 359,
420; in S.W.A., 421-23
mediterranean climate count-
er to, 363; rainfall and, 358—
60; seasonal variation, 363-
64; in stem succulents, 358—
9

Arboreal habitat, insect adap-
tations to, 611-12
Archipelago effect: adaptation
and radiation, 108-9, 114-15,
212, 250; arborescence fa-
vored, 372; dispersal factors,
16-17; floristic, 258; growth-
form spectra in, 372; in Gua-
yana highlands, 595; habitat
diversity, 97, 180; land-bird
colonization, 47; in plant dis-
persal, 12, 17; in soil condi-
tions, S.W.A., 255-58
Archipelagos, fringing, 334-35
Arecaceae (Palmae), dispers-
ibility, 482-83
Argyroxiphium (Asteraceae)
(see also Dubautia- Argyro-
xiphium-Wilkesia complex):
adaptive radiation, H.Is.,
148-49; alpine habitat adap-
tation, 148, 152; ecological
shift, 149; leaf morphology,
147-48; loss of dispersibility,
455, 456, 457
Armament, radiating, 291-93
(see also Thorn shrub habit)
Armature; plants with, N.Z.,
24

Arrival on insular area,
after long-distance dis-
persal: adaptive radia-

tion limited where recent,
108; concepts and criteria,

Arrival on insular area (Cont.)
4-5, 11-12, 54-56; conditions
for successful, 11-13, 109; on
continental islands, 5; ecolog-
ical factors, 50, 81; evolu-
tionary diversification follow-
ing, 20-21; of ferns, H_Is.,
54, 61; in H.Is., from North
American sources, 83; insect
body size and, 106; insular
biota, characteristics, 550;
modes, H.Is., 53-59, tabu-
lated, 54-55; references to
sources, 53-54; repeated in-
troductions, 11; role of re-
cipient habitat, 81; second-
ary, more efficient, 109;
sequence factor, 5, 102, 109;
speciation and modes of, 58—
59; of waif elements, 4—5
(see also Waif elements)

Ascension I. rail (‘“Crecopsis
sp.”), 489

Asia—Australia land bridge, 256

Asteraceae (see also specific
genera, and “Index of Bio-
logical Names”): adaptive
radiation: in H.Is., 147-52;
in J.F.Is., 206-10, 208, 209,
arborescence, 202, 206-10;

dispersibility loss, J.F.Is.,
483; distribution of gen-
era, H.Is., 120; in habitat
extremes, 575-76; mor-
phology of dispersibility,
443-50

Atmospheric pressure, and in-
sect flight ability, 595

Atolls equatorial, 77-81; dis-
persal role, 16-18, 80, 167;
origins, 4, 67, 212

Auckland Island, 489, 499, 500

Australia (see also South-
western Australia; Western
Australia): conifer displace-
ment from, 226; disharmonic
for mammals, 101; endemism
of flora, 253; increased
woodiness, 421-23; insect
gigantism, 605; overseas dis-
persal of flora to, 256; verte-
brate distribution, 253

Australian—South American
angiosperm distribution, 88

Australo-Papuan region: cri-
teria and limits, 15; insular
biota, character of, 101

Autochthony (see Endemism)

Autogamy, 511, 513, 537 (see
also Self-pollination)

Avalon (Calif.) temperature
range, 358

Axillary inflorescence, in plants
woodier on islands, 384—-87

Azores (see also Macaronesian
islands fauna; Macaronesian
islands flora): herbaceous
species with increased wood-
tees re: land-bird dispersal
to,

Subject Index

Baja California: lizards on
islands near, 504—5; tarweeds,
Tertiary distribution, 147

Baker, H. G.: on autogamy in
insular situations, 513; on
dioecism in H.Is. flora, 533-
34; on self-compatibility in
immigrant species, 30; on
self-pollination in dispersal,
525-26

Baker’s “law,” 525

Baldwin, P. H.: on evolution of
H.Is. honeycreepers, 163; on
feeding and movement
among honeycreepers, 167;
on group differentiation as
adaptive, 107; on sympatry
among H.Is. honeycreepers,
1

Balgooy, M. M. J., on diversity
of island floras, 110

Banana-constant moths, diver-
sification, 20

Banksia (Proteaceae), 264-68;
fire adaptation, 326-27; fruit
maturation at high tempera-
ture, 327; leaf morphology
and ecology, 282; leaves
showing species diversity,
280; loss of dispersibility,
446-48; species from W.A.,
267

Bannerman, D. A., and W. M.
Bannerman: on bird migra-
tion, Morocco to Madeira,
193-94; on long-distance
dispersal of land birds, 46—47

Barbodes (see also Cyprinid
fishes), evolution, 336—37

Barringtonia dispersal by ocean
drift, 355

Batfishes
343-46

Beach habitat: adaptations to,
6, 81, 513; arborescence on
continental beaches, 406

Beadle, N. C., on angiosperm
overland entry to Australia,
256

Beccari, O., and J. F. Rock:
on Hawaiian palms, 482; on
precinctiveness, 432, 482

Beetles (see Carabid beetles;
Coleopteran insects)

Behavioral adaptions: in fishes,
338, 342-43, 346 (see also
Cichlid fishes; Freshwater
fishes; Pediculate fishes); in
insects, 174-76 (see also In-

(Ogcocephalidae),

sects); in reptiles, 612-13
(see also Iguanas)
Bermudan frogfish (Anten-

narius) (see also Pediculate
fishes), table of character-
istics, 345

Berry, R. J., on genetic limits,
isolated mouse population,

20
Bidens (Asteraceae): achenes

of various species compared,
446, 448; dispersal, in birds’

636

feathers, 64; dispersal capa-
bilities, 443-50; ecological
requisites for speciation, 123;
fruits of endemic H_Is.
species, 447; fruits of endem-
ic species from southeastern
Polynesia, 444; woodiness
increased in, 450
Biogeographical boundaries,
15; Wallace’s Line, 15;
Weber’s Lines, 15
Biogeographical investigation,

Biogeographical provinces (see
specific localities), criteria
of, 15

Biotas (see also Habitat types) :
archipelagos as (see Archi-
pelago effect); caves as, 4-
5, 614-15 (see also Cave
habitat); characteristics, 3-5,
6-8, 112; continental type,
3-6, 27, 115-16, 340-42
(see also Continental main-
land biotas); depauperation,
6, 10, 15, 19; disharmonic
aspects (see Disharmony,
in biota composition); dis-
persal patterns (see Plant
dispersal); equatorial high-
lands as insular type, 3,
549-42, 615; harmonic as-
pects (see Harmony, in biota
composition); insular type,
characteristics, 3-4, 5, 17-18,
27-29, 341, 342, 614-15 (see
also Insular biota patterns;
Oceanic island biotas); lakes
as, 4-5, 335; long-distance
dispersal between, 6-11, 35,
81-89, 101-2, 107-9; source-
area relations between, 15;
species saturation, 49; “‘te-
puis” of Guayana as, 595;
types, 6-8

Birds (see Bird-bill forms; Bird
pollination; Land birds; Seed
dispersal by birds)

Bird banding, in plant-disper-
sal studies, 72

Bird-bill forms: in adaptation
to new feeding habits, 608;
adaptive modes, 167-68, 607—
9; adaptive variability, in
H.Is. honeycreepers, 158-69;
in Darwin’s finches, 198-99
(see also Geospizids); en-
largements, on Tres Marias
Is., 607
feeding habit relationships

of: among geospizids, 199;
among H.Is. honeycreep-
ers, 158-69 (see details
under Drepanids)
food supply characteristics
and, 608-9; sexual dimor-
phism in, of huia, N.Z., 609;
size changes in islands, 24;
sizes, compared, in Loxops
virens group, 160

Bird clutch size, on islands,
542-44

Bird dispersal to islands (see
Land birds)

Bird dispersal of plants and
seeds (see Plant dispersal,
by bird vectors)

Bird feeding habits (see Bird-
bill forms)

Bird flight loss (see Flightless
birds; Land birds)

Bird gigantism, 603, 607-9 (see
also Gigantism; Land birds;
Waterbirds)

Bird plumage: adaptations:
in geospizids, G.Is., 198; of
honeycreepers, H.Is., 159
changes in, on islands, 24—

25; conditions of, foraging
level, and diet, in geospi-
zids, 200; selective value
of bright, 610

Bird pollination, 31-32 (see
also Ornithophily; Pollina-
tion; and specific genera);
in Afroalpine zone, 568-69;
bract inflorescences attract,
301; disharmony in appeal,
146; endemism, on islands,
514; floral conspicuity, 32,
307-9, 539; floral constancy,
144-45; at ground level,
307-9

Bird wings, larger on islands,
6

07

Bock, W. J.: on time needed
for bird adaptation, H_Is.,
129; on variability in honey-
creepers (drepanids), 159

Body size changes: adaptive
aspects, 20, 23-24, 607-10
(see also Dwarfism; Gigan-
tism); in birds, relative to
latitude, 603 (see also Land
birds); in insects, 499, 595—
96, 606 (see also Insect flight
loss); in lizards, 24, 605, 606,
612-13 (see also Iguana;
Lizards)

Bog habitat: arborescence not
favored, 599; Argyroxyphium
adaptation, 148-49; in East
African highlands, 557-58,
562; phanerophyte occur-
Fence, 372

Bosote, climatic diagram, 356,
5

Bora Bora, 212

Boreal region, plant migration
to H.Is., 84

Borneo, herbaceous woodiness
in, 409

Brachyptery, 492, 494, 497 (see
also Flightless birds; Insect
flight loss; Land birds; Wing
reduction)

Brachysema (Fabaceae): adap-
tation and _ diversification,
S.W.A., 264; ornithophilous
flowering at ground level,

Subject Index

307; species from W.A., 265,
266

Bracts, showy, and capitate
inflorescences, S.W.A., 301
Branching innovations in wood-
ier growth forms, 375-84
Briggs, J. C., on ocean shelf

areas as insular biotas,
quoted, 343
Brinck, P., on _ viability of

flightless insects, quoted, 492

Bromeliads, 101

Broom-shrub habit: adaptation,
features of, 268, 271; mon-
ocot likeness, 271; in N.Z.
flora, 243, 246; in S.W.A.,
268-71; spiralization, 294;
thorn-shrub habit — distin-
guished from, 264; in W.A.,
269

Brown, J. H., on mountaintops
as noninsular situations, 614
Buds: protection against frost,
563 (see also Frost survival);
renewal after fire, 320, 321,
326 (see also Fire survival)
Buffering, in adaptive radiation
on continents, 341
Buoyancy of seeds, in plant dis-
persal, 74-75, 312, 431, 434
Butterflies: arrivals repeated,
Canton I., 11; scarcity affects
flower presence, G.Is., 32

California, 113, 147, 326

California, islands offshore:
adaptive radiation of Hemi-
zonias, 334-35; angiosperm
species, dispersal, 25; plants
from, 369; tarweeds, 147;
woody herbs dispersed to,
412-13

California flora: adaptative
radiation, with Tertiary desic-
cation, 341; dioecism, 520;
divaricating shrubs, 242; fire
survival devices, 326

Calophyllum, dispersed by
oceanic drift, 335

Camouflage (see also Color;
Melanism): flightless moth,
495; insects, 173-74, 495,
497, 606, 611; pediculate
fishes, 344-46; scree plants,
N.Z., 249; tree snails, HIs.,

176

Campbell Island: flightless
duck, 489; flightless moth,
495, 496, 497; hymenop-

terans, flightless, 500, 501;
insect species, flightless, by
percentages, 494; wing re-
duction in insects, 499
Canada, insect flight loss in, 483
Canary Islands (see also Maca-
ronesian islands fauna; Maca-
ronesian islands flora) : adap-
tive radiation, 103, 180-96;
arborescence in herbaceous
plants, 414-15; areas, alti-

637

tude, and vascular plant
species, table, 181; bird-bill
adaptations, 608
climatic conditions, 359;
compared to J.F.Is., 371
cloud-forest habitat, 360;
Echium spp., 183, 185, 187;
endemic flora, 413-14; flora,
compared with J.F.Is., 371;
growth forms in flora, tabu-
lated, 370; herbivore mam-
mals, 362; hybridization re-
ported, 535; land-bird dis-
persal to, 47; rosette habit,
herbs and shrubs, 368, 372;
Sonchus species, endemic,
376, 377, 378, 379; trunk
formation in herbs, 368;
vascular plants, relative to
altitude, 181; woody herbs,
383, 385, 413-15
Canavalia, nonbuoyant and
buoyant seeds, 430, 467
Canton Island, 11, 14
Cape Verde Islands, 180, 191,
416 (see also Macaronesian
islands fauna; Macaronesian
islands flora)
Carabid beetles (see also Cole-
opteran insects): flightless,
26 (see also Insect flight
loss); in Madeira, 501-4,
502, 503; in New Hamp-
shire, 493; in Scotland,
493; by selective adapta-
tion, 491-93; and _ sight-
less, in Madeira, 502, 503
gigantism: in Madeira, 503,
605 (see also Insect gigan-
tism); in Puerto’ Rico,
605-6
habitat and dispersibility, 488
Carson, H.C., on drosophilid
environment and evolution,
172-73
Caryophyllaceae, dispersibility
of, 461-62
Cave habitat: endemism in,
614; flight loss by insects,
26; gigantism, noted, 615;
insular biota, 5, 614-15; as
waif biota, 4
Celebes, 570
Ceratioidea (lantern fishes),
343-46; adaptation to abyssal
depths, 346; luminescent
lures, 346
Ceylon, large-mammal extinc-
tion on, 20
Chamaephyte habit: fire sur-
vival adaptation, 323; in se-
lected islands, computed,
370; 372
Chamaesycoid habit, 156
Chameleons, on Madagascar,
noted, 333
Characold fishes (Characidae),
33

Character release, 99-100; in
bird pollination of lobelioids,

Character release (Cont.)

6; in drepanids
(honeycreepers), 128; in
pioneer and weedy plants,

Chatham Island, woodiness on,
at low temperatures, 357

Chemicals, adaptive effect as
plant components, 299-300

Chile-western North America,

intercontinental disjunction,
10, 83-84
fpromosomal polymorphism,
Cichlid fishes, 335, 337-39;
adaptive radiation: in Lake
Malawi (Lake Nyasa),
337-38; in Lake Tangan-
yika, 338-39; in Lake
Victoria, 339; in Mala-

garazil swamps, 339
behavioral adaptation, 338;
breeding colors, of Tilapia
males, 338; endemic species,
count of, 337; feeding habits,
in Lake Victoria, 339; melan-
ism, 336-37; mouth breed-
ing, 338; nest building, 338

Cichoriaceae genera, dispersi-
bility, 459-60

Ciridops (see under Drepanids)

Cladode habit: characteristics,
263, 264, 271; habitat associ-
ations, 264, D2 21> leat,

morphology, 271-72, 275;
legumes (Fabaceae) and,
DDR 2TS3
mode of adaptation § in
shrubs: in N.Z., 246; in
S.W.A., 271-75; in W.A.,
270, 273
Claremont (Calif.) tempera-
ture range, 358
Cleistogamy, 519 (see Self-
pollination)
Clermontia, adaptive radiation,
H.Is., 143-46

Climatic diagrams for islands
and insular contexts, 356
Climatic factors (see also
Habitat characteristics): ar-
borescence affected, 355,
358-59, 363-64; bird clutch
size and, 542—43; coastal and
inland compared, 358; dia-

gram, on islands, 356, 357;
dispersibility and, 431; en-
demism affected, 19, 34-35,

76, 552-53, 595; in equatorial

highland habitats, 553-54,

557, 562-63

in flight-ability loss, 26, 595-
99; by birds, 487 (see also
Land birds); by insects,
493-95 (see also Insect
flight loss)

in floral conspicuity, 186;

heteroblasty and, 294, 296,

297; humidity, 360-61; hy-

bridization affected, 513; in-

sect gigantism affected, 605—

Subject Index

6; in insular biotas, 351, 353,

355-61; measurement of,

355-57, 356, 361; in paramo

habitat, 576; and woodiness

of island plants, 351, 355-
0

Cloud-forest habitat: arbores-
cence in, 360; characteristics,
210, 359; of C.Is., 359, 360;
of G.Is., 205, 360; of J.FIs.,
210; leaf adaptations to, 225;
of Macaronesian is., 182, 360

Clutch size: of birds on islands,
542-44; of lizards on islands,
504-5,

Cockroaches, flight loss due to
cold,

Cocos Islands: plant dispersal,
77, 80; wet-forest habitat, 80

Cold temperature adaptation
(see also Frost survival):
cupressoid habit as, 590;
insect flight loss as, 596;
insect speciation vicarious
in, 596-97; plant speciation
vicarious in, 558, 571-75,
581-92

Coleopteran insects (see also
Carabid beetles): altitude
and differences among, 503;
flightless, Mt. Kilimanjaro,
597; flightless and sightless,

Madeira, 501-4, 502-3;
gigantism, Madeira, 605;
nocturnalism, 503; wing and

color loss, 503

Colombia: equatorial highland
zone, 550; Espeletia spp.,
577, 579, 580; herbaceous
components of paramos, 586;
high-elevation plants, 588;
paramo shrubs, 59/, 593;
plants from superparamos,
589; tree-fern frost adapta-
tion, noted, 574

Colonization by plants:

affecting, 3-5, 11-14,
modes of arrival,

s., 54-55; recentness, 103-
4; reproductive advantages
of dioecious flowers, 525-—
26; source-area species avail-
able, 11

Color, adaptation and change
(see also Camouflage; Mel-
anism): altitude and, in
beetles, 503; in bird plumage,
610-11 (see also Drepanids;
Geospizids); in conspicuity
of alpine flora, 186, 568; in
dolichopodid flies, 174; in
feeding habits of Darwin’s
finches (geospizids), 198,
200, 610; in fishes, Philippine
Is., 336-37; in flightless in-
sects, 503; of flowers (see
Flower color); on_ islands,
610-11; insect displays of,
173-74, 495, 606, 610-11;
nocturnalism in, 503, 504;
nonprotective in achatinellids

638

condi-

(land snails) 176; pollination
functions, 300, 540-41; pro-
tective functions, 173-74,
176, 610; seed dispersal by
birds and,

Competition: as adaptive
mechanism, 27-29, 99, 608;
in angiosperm displacement
of gymnosperms, 220, 225-
26; avoidance of, by con-
generic birds, 166, 608; char-
acter displacement and, 99;
decreased between endemics
on islands, 27, 33; in estab-
lishment ability, 48; genetic
depauperation and loss of,
28; honeycreeper avoidance
of, 166; in insular biotas,
27-28; in insular lizard pop-
ulations, lost, 604

Compositae, composites
Asteraceae)

Congeneric species (see Sym-
patry )

Conifers (see also Gymno-
sperms) : absence from
oceanic island biotas, 354;
dispersal capabilities, 6, 87—
88, 225-26; Guayana high-
lands lacking, 594
occurrences: in Panes 88;

in Australia, 226; Gs
216-26; in NiG= "326: in
N.Z., 4, 233-35; in Tas-
mania, 226
relictualism in, 47) 88)e229—
26; riparian adaptation, in
N.C. species, 223-25
Continental drift: Africa-South

(see

America, indications, (86;
angiosperm distribution,
evidence of, 86-87; inter-

continental dispersals related
to, 83-89; oceanic islands
not explained by, 16-17; as
plant dispersal mechanism,
3, 10, 16-17, 83-84, 86-89

Continental elements, ‘displac-
ing insular elements, 27, 109

Continental herbaceous ele-
ments, woodier on_ islands,
404-7 (see also Insular
woodiness )

Continental island biotas, 4,
332-34 (see also Continental
mainland biotas; Oceanic
island biotas); adaptive radi-
ation process, 101-2, 115,
251; angiosperm flora, 112,
115; Araucariaceae, presence
indicative of, 34, 215; dis-
harmonic situations, 101;
dispersibility of plants, 34,
115; Fiji Is) as; 75 213) 436:
geological aspects, 34; har-
monic aspects, 7; isolation,
4: Madagascar as, 332-34;
New Caledonia as, 7, 34,
45, 115, 215-21; relictualism
characteristic of, 18, 104-5,
217, 332-33; Tasmania as,

18; waif element arrival, 5;
waif element presence, 6—7;
woody herbs lacking, 353
Continental mainland biotas,
3-5, 340-42 (see also Conti-
nental island biotas; Oceanic
island biotas); adaptive radi-
ation characteristics, 115-16,
341-42; angiosperm floras
and “diversity,” 112; animal
populations, 608; cataclysmic
events, 341; coastal zones,
with insular climate, 352;
disharmonic factors, 101;
dispersibility lessened, 484;
endemic speciation, 112, 115;
harmonic characteristics,
341, 484; humidity character-
istics, 361; insect flight loss,
489; insular factors, 3-5,
108, 342, 352, 404; island
biota pattern compared with,
111-13; lizard population
and body size, 604-5; pioneer
biota traits, in contrast with,
433; reproduction rate, 434;
South Africa as instance,
342; speciation, 111-13, 434;
S.W.A. as instance, 342
Convolvulaceae, loss of dis-
persibility in, 462
Coprosma: adaptive radiation,
H.Is., 155; arborescence,
N.Z., 238, 240; leaf anatomy,
238, 240; species, N.Z., 239
Cork formation: in fire resist-
ance, 323; in frost survival,
ae8s in plants of S.W.A.,

Cormorant, flightless (Phala-
crorax harrisii), G.Is., 490
Corn, C., experiments in wind

dispersal by, 59
Conia, establishment in H_Is.,
12

Cotton (Gossypium): endemic
in H.Is. and Tahiti, 8, 81,
471-72; routes of Pacific
dispersal, 81

Costa Rica: alpine plants, 573;
tree ferns (Blechnum), 573,
574

Cottid fishes (Cottidae), of
Lake Baikal (U.S.S.R.), 335;
339-40; adaptive radiation,
340; endemic species, 339-
40

Crane fly (Tipula subaptera),
flightless, Mt. Kilimanjaro,
598

Cretaceous period, 4 d
Cronquist, A., on altitudinal

distribution, Scalesia spp.,
G.Is., 203
Crossin, R., on travel by ma-

tine birds,
Crozet Island, insect flight loss

R.W.:
plant dispersal,

on birds in
86; on dis-

Subject Index

junctions in plant distribu-
tion, 83

Cryptocarya spp., dispersibility,
470-71

Cuatrecasas, J., on Espeletia
species, 582

Cuba, increased woodiness in,
417-18

Cucurbitaceae, _ dispersibility,
462

“Cumulative change,” 100

Cupressoid habit, in gymno-
sperms, 590

Cushion-plant habit: character-
istics, 240, 242, 574, 581,
587; endemism, N.Z., 240,
241, 242; in equatorial high-
land zones, 581, 587; frost
survival and, 565; in Ma-
lesian alpine zone, 572, 574;
in N.G. highlands, 572, 574

Cyanea: adaptation and radia-
tion, H.Is., 120, 124, 131-39,
132, 133; bird pollination,
134, 137; wet-forest adapta-
tion, 131, 134-36

Cyathea species, alpine adap-
tations, 571

“Cyclic hybridization,” in N.Z.
flora, 31

Cynareae genera, dispersibility,
457-58

Cynipid wasp (Encoilidae),
wing reduction, 499

Cyprinid fishes (Barbodes),
Lake Lanao (Philippines):
adaptive radiation, 336-37;
evolution of endemic genera,
20, 335-37

Cyrtandra: adaptation and
distribution, H.Is., 119-23,
155-56; ecological adapta-
tions, 122; increased wood-
iness, 155

Dacrydium (Podocarpaceae),
species in N.C., 220-25, 221,
222

Dampiera (Goodeniaceae), of
W.A., cladode habit in, 270,
271-72

Darlington; PP: J:, Jr; 2; on
animal dispersibility, 6; on
animals in West Indies, 15s
on classification of islands,
5; on flight loss in waif
biotas, 26; on insect flight
loss, 26, 291-93, 501, 595-
96; on utility of flight, and
wing atrophy, quoted, 488-
89

Darwin, Charles, 1, 329, 501;
on adaptation by Australian
marsupials, quoted, 108; on
competitive selectivity in
Scalesia, 364; on imperfect
adaptations, 108; on trees
mainly on continents, 354;
on wind pressure in insect
flight loss, 493; on wood-

639

iness increased on _ islands,
350

Darwin’s finches (see Geospi-
zids)

Darwinia (Myrtaceae), S.W.A.,
302-16; adaptation and di-
versification, 302-9; bird
pollination, 302; dispersal,
314-16; endemism, by
species, 304-5; fruit mor-
phology and _ dispersibility,
SUS: dnorescence formation,
303,

Dies PAPA) W.A.,
259-63, 260, 261, 262; leaf
morphology, 259, 263, 283,
293; photosynthetic surfaces
reduced, 259; as response to
herbivore pressures, 263

Deepsea fishes, 343-46

Deignan, H. G., on dispersal
of land birds to islands, 46

Delissea (Lobelioideae), H.Is.,
130-31; distribution of
species by islands, 120, 124;
leaf morphology, 130

Dendroseris (Asteraceae),
J.F.Is.: dispersibility factors,
460; leaf morphology and
adaptation, 210-11

Depauperation: genetic, 28,
513; of species, on islands,
112-13

Desiccation survival, 323 (see
also Fire survival)

Desventuradas Islands: en-
demic species, 411; floral
outcrossing, 30; herbaceous
species with increased wood-
iness, 411; plant dispersal by
birds to, 79

Dichogamy, 516-19, 532

Dichotomous branching sys-
tems, in leaves, 296

Diclinous species, 510, 519,
520

Dicotyledons, 441-81; dispers-
ibility lowered, 441-81; mode
of arrival, HIs., . 57-58;
venation in leaf evolution,
S.W.A., 283-88

Didiereaceae, Madagascar, 333

Diet (see Food supply )

Dimorphism, in fauna (see
Sexual dimorphism)

Dimorphism, flora, 532; in
W.A., 296

Dioecious species: advent on
islands, 511, 524-26; in
British Isles, percentage, 520;
in California, percentage,
52; in Central Andes, 521;
characteristics, 11, 509-11,
516, 520-26; gynodioecious
and, 522, 523, 534, 541
in H.Is., 511, 516-19, 521-

26, 522; percentages of,
533-34
insular occurrence high, 30,
509-10, 534; in J.F.Is., 520-

Dioecious 4 spevies (Cont.)
21; in N.Z. flora, high, 520;
outcrossing and, 30, 511,
533-34; in S.H. flora, 522:
strategic value on_ islands,
510-11, 525

Diptera (flies): flight loss, on
subantarctic islands, 498,
499, 501; wet-forest adapta-
tion, H.Is., 126; wing-size
variation, 497

Disharmony, in biota compo-
sition, 6-8; adaptation in,
101- 3, DSO S95. 4345) in
continental (mainland) bio-
tas, 7, 101, 340-42; in con-
tinental island, 7, 101; in
equatorial highlands, 101-2

of fauna: in N.C., 233; in
N.Z., 233, 249

of flora: in G.Is., 205; in
H.Is., 146, 436, 541-42;

in S.W.A., 253-59, 319; in
Tasmania, 101
in Malesian area, 101;
oceanic islands, 7, 18, 101,
341, 434; reproduction and,

434, 542; woodiness' on
islands and, 354-55
Disjunctions, intercontinental

(see Intercontinental disjunc-
tions; Continental drift)
Dispersibility, 6-8, 45-95, 109
(see also specific types,
families, and genera); adap-
tive radiation, at the limits
of, 103; by air and wind
(see Aerial dispersal; Wind
dispersal); of birds to islands
(see Land birds); by bird
vectors (see Seed dispersal
by birds); conditions affect-
ing, 6-8, 14-15, 609-10; in
displacement by competitors,
28; ecology and distance in,
77-82; evidence of, 14-16,
50-54: of ferns, 439-40, 512!
of forest trees, 353-54; of
fruit (see Fruit dispersal);
group limitations, 103; of
insects (see Insect ‘dispersal );
insular factors affecting, 5,
7, 18, 509-10; insular wood-
iness, B52 loss (see Plant
dispersibility loss); opera-
tional factors, 429-32; pio-
neer factors, 6, 17, 107, 352,
433; plant adaptation for, 5,
6-18, 50-54; probability
factor in, 14
Dissemination, of
fruit (see Plant
Distance factor:
persal, 77-82;
radiation, 104
“Disuse” theory of insect flight
loss, 492
Divaricating shrubs (see also
Radiating armament; Thorn-
shrub habit): California in-
stances, 242; characteristics,

seed and
dispersal )
in plant dis-
in adaptive

Subject Index

242-43; New Zealand in-
stances, 242-43, 247
Bede: Mascarene, affinities,
Dolichopodid flies, color vari-
ation, 174
Downes, J. A., on insect flight
loss in Canada, 483
Dracaena (Pleomele) (Lilia-
ceae), Hawaiian species, 483
Drepanids (Drepanididae),
(Hawaiian honeycreepers) :
adaptation and_ variation,
22-23, 128-29, 158-69, 608—
9; antecedent affinities, 159
bill form, diet, and feeding
habits, 158-69; of Hemi-
gnathus lucidus, 164; of
H. procerus, 165; of H.
Wilsoni, 166; of Loxops
coccinea, 162; of L. parva,
161; of L.. sagittirostris,
162; of L. virens, 160; of
L. virens stejnegeri, 163;
of Pseudonestor xantho-
phrys, 167; of Psittarostra
kona, 169; of P. psittacea,
168
character displacement and
release, 128, 159; Ciridops,
earlier distribution, 128;
coloration in plumage, 159,
610; competition, avoidance
of, 166, 608; diet (see bill
form above); evolutionary
development, 163, 169; ex-
tinction, 129, 158; flower-
pollinator constancy, 144~—
45; Himatione, primitive
genus, 128; illustrated studies,
references, 158; lobelioid
pollination, 144; Loxops as
primitive genus, 128; non-
adaptive factors in changes,
146; nonadaptive genetic
drift, 159; Psittarostra, adap-
tive radiation, 128; sympatry
among, 166; wet-forest adap-
tation, 158
Drift flora (see Oceanic drift)

Drosera (Droseraceae): adap-
tive radiation, ne ende-
mism, S.W.A

Drosophilas in H. Is. summary

of inter-island founders in
the picture-winged, /22

Drosophilid flies, H.Is.: adap-
tive radiation in behavior,
174-75; breeding environ-
ment, 172; gigantism, 605-6;
host-plant specificity, 172; lek
behavior, 174; specialization
excessive, 29; wind role in
survival, 403

Drought survival mechanisms,
in flora, S.W.A., 320

Dry climate (see Xeric habitat;
Xeromorphy )

Dry-forest elements,
dispersal, 78

640

in plant

Dubautia (Asteraceae), ‘“tar-
weeds,” 119-53, passim
Dubautia (sensu lato), 119,

120 (see also Railliardia)

Dubautia (senso stricto):
adaptive radiation, H_LIs.,
149-52; alpine adaptation,

151-52; differentiation from
related genera, 119, 120, 123,
149, 150, 151 (see also Argy-
roxiphium, Dubautia-Argy-
roxiphium-Wilkesia complex,
Railliardia, and Wilkesia);
distribution, by islands, tabu-
lated, 120; leaves in silhou-
ette, and vessel elements,
153; loss of dispersibility,
456-57; primitive inflores-
cences, 123

Dubautia- Argyroxiphium-Wil-
kesia complex (‘‘tarweeds”’),
102, 108, 114, 405; adaptive
radiation, H.Is., 147-53; in-
sular adaptive radiation ex-
emplified, 106; provenance,
59; relations among genera,
119, 123, 149-51

Dwarfism, on islands, 23-24,
607

East African highlands, 550,
553-70 (see also Equatorial
highlands and specific habitat
zones indicated); alpine
habitat and flora, 23, 553-
54, 562-69, 564, 566 (see
also Afroalpine flora); bog
habitat and flora, 557-58,
562; dispersal patterns, 551—
53; endemic and vicarious
flora, 552-53, 558, 562, 569;
frost survival of flora, 563,
565, 566, 567; lower montane
habitat and flora, 554, 556,
557; and Malesian highlands
compared, 571; moorland
habitat and flora, 557-62,
559; protection against radi-
ation, 567-68; submontane
habitat and flora, 553-54,
555; upper forest habitat
and flora, 557; woodiness in
herbaceous species, 423-24,
553, 562-63; xerophytism,

Easter Island: endemism in
marine fauna, 34-35; flora
dispersed to, 76, 80

Ebenaceae, morphology
dispersibility, 462-63

Echium (Boraginaceae): adap-
tive radiation, Macaronesian

and

is., 182-91; effects of frost,
tabulated, 190; species and
adaptations, 183, 185, 187,

188; woodiness increased on
islands, 353

Ecological correlations, _of
vessel dimensions in insular
angiosperms, tabulated, 392

Ecological factors (see also
Climatic factors; Habitat
types): in adaptive radiation,
21-27, 97-109; angiosperm
growth-forms affected , 372-
73; bird-bill adaptations,
607-9: broom-shrub habit
and, 268, 271; character dis-
placement and release af-
fected, 99-100, 107, 128,
159; cladode-shrub habit
and, 271-72, 275; competi-
tion between species affected,
13-14, 46-47; conifer adap-
tations, N.C., 216-26, 2/8,
219, 221, 222, 224; in dis-
persibility and establishment
of species, 13-14, 48, 77-82,
85, 251, 569; fern spore size
and, 438-40; floristic spec-
trum represents, 371; genetic
variability and, 19, 511;
geology a determinant of,
82; geospizid adaptations to
range of, 198-201, 199, 200;
in insular biota, 433, 509,
549; in insular woodiness,
373 (see also Insular wood-
iness); in integrity of insular
species, 511; in interfertility,
$11, 537; in J.F.Is. and HIs.,
compared, 372; leaf form
adaptations, S.W.A., 282-94;
in lizard dispersal and sur-
vival, 47, 98, 232, 502-4; in
long-distance plant dispersal,
14, 77-82; polymorphism in,
linked to dispersibility, 352;
range of, affecting adaptation
on islands, 81-82, 97-98; in
secondary woodiness, 353-—
61, 363-65; speciation af-
fected, 105-7, 121, 181 (see
also Speciation and diversi-
fication)

Ecological release: defined, 48;
in lizard growth on islands,
104

Ecological shift:
from terrestrial habits, by
insects, 611-12; character
release in, 99-100, 107; dis-
persibility and, 25-27, 109,
430-31, 433-34; drier to
moister habitat, 147-51, 202,
S759550; 00 ferns; -H-Is:,
157; harmonic biotas reduce
opportunities for, 6, 7, 88,
193-94, 483-84; in insect
gigantism, 606; insular con-
ditions conducive, 211; of
littoral elements to inland

to arboreal

sites, 76; and morphological
change, 167-68; into wet
forests, 483, 606

Ecotones, 21

Ecotype formation, 100-103

Eisenberg, J. F., and E. Gould,
on tenrec speciation, in Mad-
agascar, 333

Subject Index

Elaeocarpaceae, dispersibility,
3
Emu, 490, 491 ~ F
Endangered species: flightless
birds as, 489; population a

factor, 615; preservation,
616
Endemism, 34-35 (see also

other specific families, gen-
era, and localities); adapta-
tion and, 102, 113-15; in
Afroalpine flora, 558-69
in angiosperm flora, of areas
by percentages, 112; of
Australia, 253, 256, 259
68, 276-78; of G.Is., 202-
5; of H.Is., 56-58; statis-
tics, 512; of N.Z., 441
in Antennarius (frogfish),
Bermuda, 344; in Apocyna-
ceae, H.Is., 441-42; in Aster-
aceae, Guayana _ highlands,
594-95; in Bidens, on Pacific
islands, 443-50, 444, 447;
in caddis flies (Limnophili-
dae), Lake Baikal, 504; on
California off-shore islands,
412-13; in cave habitat, 614;
in C.Is. herbs, 413-15; in
Clermontia, H.Is., 139, 143-
46; climatic factors affecting,
19, 34-35, 76, 552-53, 595;
in continental areas, 112, 115,
211; on continental islands,
34, 101-2, 115, 211; in cottid
fishes, Lake Baikal, 339-40;
in cotton (Gossypium), H.Is.
and Tahiti, 8, 471-72; in
Cuba, 417-18; cultigens in
contrast with, 8; in Cyanea,
H.Is., 131-39, 132-33; in
cyprinid fishes, Lake Lanao
(Philippines), 20, 335-37;
in Dacrydium, N.C., 223-25;
in Darwinia, S.W.A., 255,
302-16, 304-5; in Delissea,
H.JIs., 139, 142; in Den-
drosenecios, African high-
lands, 569; of Desventuradas
Is., 411; in Didiereaceae,
Madagascar, 333, 420; in
dioecious species, 520-26;
in dispersal patterns, S.W.A.,
14; in equatorial highland
flora, 552-53; in Fabaceae
and Rhamnaceae, H.Is., 465;
in ferns, H.Is., 61, 157, 512;

in flightless insect (Cos-
syphodes), Madeira, 503;
in _gekkonid lizards, N.C.,

232; genetic factors, 19-
20, 510-12; in Guayana
highlands, 594: in herbaceous
types woodier on islands,
406, 408, 410 (see also In-
sular woodiness); hetero-
blasty in, 294-95; in high-
land biotas, 552; in His-
paniola, 418; in insects, Lake
Baikal, 504; in insects, H.Is.,
169-76 (see also Drosophilid

641

flies, H.Is.); isolation factor
in, 62; in J.F.Is., 210-11,
410; in lemurs, Madagascar,
102, 332-33; in leaf speciali-
zation, S.W.A., 283-94; in
lizards (Uroplatidae), Mada-
gascar, noted, 333; in Mada-
gascar and Comoro Is., 419-
20; in Madeira Is., 415-16;
in Malesian highlands, 575;
in Marine fauna, J.F.Is., 34—
35; in Marquesas Is., 115,
407; in Mascarene Is., 421;
in molluscs, Lake Baikal,
340; in mountaintop habi-
tats, 472; in N.Z., 234-49,
409; oceanic drift a factor,
81, 513— 14; in oceanic island
biotas, 34_ 35; in Opuntia
(Cactaceae), G.Is., 203-4;
percentage of, in angiosperm
flora of various biotas, 112;
prediction of, 114; in Puerto
Rico, 418; of Rollandia,
HIs., 139-43; on Samoa, 408;
on S.H., 115, 416-17; on So-
cotra, 420; of Sonchus (As-
teraceae), C.Is. and Madeira,
375-81; speciation versus,
113-14, 212; strand flora
low in, 19; in succulents,
S.W.A., 289; sympatry and,
114-15: taxonomic group-
ing and, 102; in woody herbs
on islands (see Insular wood-

iness )

Entomophily, 514 (see also
Insect pollination)

Environments (see specific
types, also listed under Habi-
tat types)

Eocene fossils, of gymno-
sperms, 215

Eocene geography, of the

Pacific, references, 16
Epiphyte habit: characteristics,
440; in J.F.Is., 206-7; in
lobelioids, H.Is., 143; in
paramos, 592; in various
islands, computed, 370
Equatorial alpine habitat, 549-
52, 562-67, 575-76
Equatorial countercurrent, 77
Equatorial highlands, 549-602;
aerial dispersal of seeds,
552; in Africa, 550, 553-70
(see also East African high-
lands); arborescence, 357;
climate, 355, 357, 550-51,
596; Colombia, 550 (see also
Colombia; Paramos); cush-
ion-plant habit, 574, 581,
587; as disharmonic biota,

101-2; endemism in flora,
552-53; factors _charac-
teristic of, 549-53; flora,

549-50, 584 (see also Afro-
alpine flora); in Guayana
highlands, 594-95; increased
woodiness, 423-24, 559, 560,

Equatorial highlands (Cont.)
561, 564, 566; insect flight
loss, 595-99; in Malesian
islands, 550, 570-76; moor-
lands zone, 557-62; in N.G.,
570, 575 (see also Malesian
highlands; New Guinea);
in paramos, 576-94; subal-
pine habitat, 570-71; sub-
montane habitat, 553-54;
temperate zone elements,
550-51; Venezuela, 550, 576—
94, passim (see also Par-
amos)

Erythrina (Fabaceae), seed
dispersibility, 464-65
Espeletia (Asteraceae): acau-

lescent species from paramos
of Venezuela and Colombia,
578; adaptation and radia-
tion, 576-84; alpine habitat
adaptation, 581-84; defini-
tive of paramos, 576; rosette-
tree species, 579, 580; varied
habits, 577

Establishment ability, after dis-
persal: competitive factors,
13, 48-49; dispersal ability
versus, 17; heterozygosity
and, 49; in land animals on
islands, 13, 49; in immigrant
species, 14-15, 48, 70, 227;
pioneer traits and, 6, 17, 107;
pollination capabilities, 14,
32, 524-26; rivalry among
invading species, 99, 108-9;
role of recipient habitat, 13-
14, 81-82

Euphorbia CBuphorbiaceac)
adaptive radiation, :
156-57; dispersibility, 463-64
insular woodiness: in HLIs.,

136; in Macaronesian is.,
360 :

Evolutionary plasticity: angio-
sperm leaf adaptations dis-
playing 278-90, 279, 284-85,
290; insular settings, 23-24,
510-11, 615; in long-distance
dispersal, 35

Evolutionary process: adaptive
radiation as, 20-27, 102; in
birds and fishes, compared,
338; on islands, 2-3, 4-5, 20—
22, 47-48, 510-11; loss of
dispersibility and, 429-30

Extinction of species, 615-16;
due to ecological poverty
on islands, 19; due to human
agencies, 615; land area
criteria, 70, 538, 615

Fabaceae: cladode-forming
species, S.W.A., 272, 273;
dispersibility of seeds, 464—
67; seeds capable of seawater
dispersal, 75; species from
SW A. and W.A., 260, 261,

Falla, R. A., on the effects of

Subject Index

inland-nesting marine birds,

Fearless behavior, in insular
animals, references, 33
Feeding habits and dietary
modes of birds: in adapta-
tion to food supplies, 98-
99, 608 (see also Resource
partitioning); beak mor-
phology and (see under Bird-
bill forms); in Darwin’s
finches (see Geospizids); in
H.Is. honeycreepers (see
Drepanids); plumage, forag-
ing level, and diet of geo-
spizid genera correlated, 200
Ferns, 437-41 (see also Fern
spores); adaptive radiation,
Huls!; 1575 (60s arrival’ in
H.Is., 54, 61; dispersibility,
437-41; epiphytic habit, 439,
440; hybrids, 512; in J.F.Is.,
206; paraphyses, in H_LIs.
species, 440; specialization

of habitat, 438-39

Fern spores (see also Ferns):
aerial flotation to H.Is., 54,
61; comparison of sizes, ‘with
environments, 437-40; epi-
phytic habit and size, 439-
40; precinctiveness and size,
4 0

Fertility: and dispersibility of
plants, 431-32; of insular
plants, 541-42; in insular
vertebrates, 542-43

“Fijian difficulty,” 436

Fiji Islands: andesite forma-
tion, easterly limit, 10; ants,
adaptive radiation indicated,
211; continental island char-
acter of, 7, 436; eastern dis-
persal boundary for many
plants, 10, 215, 216; floral
adaptive radiation, not docu-
mented, 213; harmonic ele-
ments in flora, 7, 436

Fire survival: adaptation to, of
flora, S.W.A., 319-29; 321,
322, 324; in Afroalpine zone,
568; bud renewal, 320, 321,
326; chamaephyte habit, 323;
cork formation, 323, 325;
flowering induced, 327, 328,
329; fruits adapted, 326-27:
geophyte habit, 323; “‘lig-
notubers,” 324, 326; rosette
trees, 323-26; underground
structures, 319-23, 320

Fishes (see ‘Antennarius; Cera-
tioidea; Cichlid fishes; Cottid
fishes; Cyprinid fishes; Frog-
fishes: Galaxiad fishes: Neo-
channa; Ogcocephalidae;
Pediculate fishes)

Fitchia (Asteraceae): adaptive
radiation, to Pacific islands,
212; arborescence, 452; bird
pollinator adaptation, 453;
fruits of various species, 451;
morphology and dispersi-

642

bility, 452-53; rain-forest
adaptation, 453

Flight, function in animal dis-
persal, 488

Flight ability loss: adaptation
to insular conditions, 26-27,
487-91, 595; by birds, 489-
91 (see also Land birds);
conditions affecting, 488- 93;
geophily, 501, 596; by in-
sects, 491-506 (see also In-
sect flight loss; Moths, flight-
less); precinctiveness, 488,

; 5; survival mech-

anism, 491-92

“Floating gene”
243, 282

Flora on islands: adaptive
radiation, 97-116, 353, 538;
angiosperms with large ‘seeds,
few, 353-54; annuals few,
14; conspicuity low, ESD),
539; dioecism, 30, 509-11,
520-26; diversification fac-
tors, 510-15; diversity, 110;
gymnosperm dispersibility
lessened, 441; herbs to
shrubs and trees, ratio of,
373; integrity of species, 511;
massing for pollination, 541;
outcrossing, 30, 511-13, 533-
34; perennial growth fa-
vored, 511; pollination pat-
terns (see Pollination); re-
productive factors, 510-34;
self-sterility, 515, 516; speci-
ation, 2-3, 21, 100, 104-5,
111-13, 510-11, 537 (see
also Speciation and diver-
sification); stature increased,
388; waif characteristics (see
Waif elements); wind polli-
nation, 532-33 (see also
Wind pollination) wood-
iness, 404-23, 511 (see also
Insular woodiness)

hypothesis,

Florida Keys: recolonization
rapid, 12; wind pollination,
533

Florin, R.: on gymnosperm
displacement in southwest-
em | Pacific); 9225 67aon
Podocarpus_ diversity in
NIG 223 me25

Floristic distributions: char-

acteristic of equatorial high-
land zones, 549-50, 584; by
life-form traits, 370; ratios
of species, to genera, to
families, by areas, 109-16,
tabulated, 112

Flower(s) (see Angiosperms;
Flower color; Flower con-
spicuity; flower types)

Flower color: in Afroalpine
environment, 568; of lobe-
lioid species, H.Is., 145; pol-

lination role, 300, 307, 540—-
41; range of hues, on specific
islands, by percentages, 539-

40; variations in Cyanea
spp., 139

Flower conditions and pollina-
tion types, in native H_Is.
species, table, 519

Flower conspicuity: endemism,
300; fragrance, 540; ground-
level tendencies, of Pro-
teaceae, S.W.A., 308; insect
pollinators and, 31-32, 514,
539; in J.F.Is., 539; lacking
on islands, as ‘tendency, a=
32, 539; massing to attract
pollinators, 32, 541: in N.Z.,
31-32, 539; as pollination
factor, 31-32, 144, 257, 300,
539-41, 568-69

Food supply: adaptation to,
in niches on reefs, 343; bird-
bill adaptations, 608-9 (see
also Bird-bill form); and
bird clutch size, 542-43;
congeneric species partition-
ing, 608; dwarfism as adap-
tation, 24; establishment suc-
cess and, 13-14, 19; in in-
sular contexts, 542-43; lizard
population and, 543, 604; re-
productive efficiency and, 33

Forest habitat: continental
type, 434; dispersibility of
trees, 354; dry upland, 79,
164, 165, 393; plant dis-

persibility loss, 433, 453,
467; rainy (see Rain-forest
habitat; Wet-forest habitat)

Fosberg, F. R.: on endemism
in H.Is. flora, 12, 61; tabular
summaries of H.Is. flora, 53,
11 2

Fossils: earwig, S.H., 605;
gymnosperms, from Jurassic
in India, 215; insectivore
(Nesophontes), 18; marsu-
pial (Thylacinus), 352
“Founder principle,” of Ernst
-Mayr, 19, 106, 202

French Frigate Shoals (H.Is.),

Peer fishes (see Cichlid

fishes; Cottid fishes; Cyp-
tinid fishes)

Freshwater lakes, 3, 4, 335-40

Frogfishes (Antennariidae),
343-45

Frogs, adaptive radiation: in
Seychelles, 211; lack of in
N.Z., 104

Frost-free habitat, 375, 553

Frost survival: adaptive as-
pects, 375, 562-67, 587;
arborescence, 563; cork, 563:
cryptophyte habit, 565; cu-
pressoid habit, 590; cushion-

plant habit, 565; in East
African highlands, 562-67,
566; in Echium, tabulated,

190; in insular woody species,
364: leaf structure adapta-
tions, 563-67; in Melanesian
highlands, 570-75, 572, 573;

Subject Index

in paramo flora, 583, 587;
of temperate flora in ‘alpine

zones, 562-67; tree-fern
adaptation, 574

Fruits and seeds, types of:
Alectryon spp., Hawaiian

and non-Hawaiian com-
pared, 479; Alsinodendron
SpPp., endemic in H.Is., 461;
Banksia spp., W.A., 267;
Bidens spp. endemic in H.Is.,
447; Bidens spp. endemic in
southwestern Polynesia, 444;
Darwinia spp., S.W.A., struc-
ture and dispersibility, 315;
Euphorbia spp. endemic in
H.Is., 461; Evodia_ spp.,
Hawaiian and non-Hawatian
compared, 477; Fagara spp.,
Hawaiian and non-Hawaiian
compared 477; Fitchia spp.,
451; Goodeniaceae, S.W.A.,
311, 313; Haplostachys spp.,
H.Is., 469; © Myodocarpus
spp., N.C., 230; Oparanthus
spp., 451; Pelea spp., Ha-
walian and non-Hawaiian
compared, 477; Pittospor-
um spp., Hawaiian and non-
Hawaiian compared, 473;
Platydesma spp., Hawaiian
and non-Hawaiian compared,
477; Scaevola spp., N.C.,
229; Schiedea spp., endemic
in H.Is., 46]; Stenogyne
Spp., Hs., 469; Tetrapla-
sandra spp., H.Is. and Philip-
pines, 442
Fruit dispersal: adaptations
for, in species of S.W.A.,
309-16, 3/5; analytical as-
pects, 51, 62, 80-81; bristles
and barbs, 54, 63, 64, 78,
buoyancy (see Oceanic
drift); by frugivorous birds,
55-59, 62-63, 65, 66, 67, 68,
226, 442, 446 (see also Seed
dispersal by birds); loss of
(see Plant dispersibility
loss); role of size (see Fruit
size enlargement: Gigantism
of flora); structure of Dar-
winia for, 315; viscidity, 10,
55, 70, 71, 72, 80-81, 310-
12, 436, 473 (see also Seed
dispersal; Plant dispersal )
Fruit size enlargement: Alec-
tryon spp., H.Is., 478, 479,
480; Cryptocarya, in ex-
tended evolution, 471; eco-
logical adaptation to forest
habitat, 473
ecological correlations: in
Pritchardia spp., 482-83;
in Stenogyne spp., 469
Elaeocarpus, Samoa, 463;
Gunnera spp., 468; and less-
ened dispersibility, 431, 433,
442-43 (see also Plant dis-
persibility loss)
in rain-forest

643

adaptations:

Fitchia, 453; Pritchardia,

482-83; Rubiaceae, 476;
Rutaceae, 476 :
Fryer, G., on evolution of

fishes in Lake Malawi, 337-
38

Fungi, dispersal by birds, ref-
erences, 52

Galapagos Islands: adaptive
radiation on, 197-206, 364;
angiosperm flora, altitudinal
distribution, 203-5; arbo-
rescence, 204, 358, 360;
archipelago effect, 204; biota
characteristics, 4, 20, 32, 81,
197, 199, 205, 512; bird dis-
persal of plants to, 79-80;
climatic conditions, 205, 360;
Darwin’s finches (see Geo-
spizids); floral elements and
characteristics, 31-32, 205,
411-12, 515, 519; genetic
factors, 537; giant tortoises,
201-2; iguanas, 22, 612-13;
insect speciation, 202; wood-
iness of herbaceous species,
360, 411-12

Galaxiad fishes (Galaxiidae),
N.Z., 249-50; aestivation
during drought, 250

Gammarid shrimps,
cation of, Lake
(U.S.S.R.), 340

Gastropods (see Land snails)

Geckos (see also Iguanas; Liz-
ards): adaptation for dis-
persal, 8, 9; diversification,
Madagascar, 333; endemism
and adaptive radiation, N.C.,
232

Genera per family of angio-
sperm flora, calculated by
regions, 112

Genetic drift, factors affecting:
in adaptive radiation, 19-20,
28-30, 510-13; chromosomal
polymorphism, 19; in color-
ation of achatinellids, 176;
competitive ability and, 28—
29, 30; concept (“founder
principle’), 19, 106, 202;
in drepanid birds (honey-
creepers), 159; in establish-
ment of insular species, 19—
20, 30-31, 34, 511-13, 537;
in hybridization in limited
population, 19-20, 29, 511-
13; land area requirements,
538

Genetic potential: in body size,
20 (see also Dwarfism; Gi-
gantism); in establishment
on islands, 19, 30, 512; of
giant tortoises, G.Is., 202;
of land snails, H.Is., 28

Genus formation, in response
to habitat diversity, 105

Geological factors: in age of
H.Is., 119; in biota character-
istics, 3-4: in continental

diversifi-
Baikal

Geological factors (Cont.)
drift hypothesis, 86-89; in
continental island biotas, 3,
34; in ecological conditions,
82; in endemism of highland
biotas, 552-53; in oceanic
island biotas, 3; in time span
on volcanic ‘islands, 3, 20

Geophily: _ concept, 488— 89;
conditions affecting, 596; in
insect flight loss, 493, 501,
5)

Geophyte habit: in fire-survival
adaptations, 323; occurrence
on selected islands, tabu-
lated, 370

Geospizids (Geospizinae, ‘“‘Dar-
win’s finches”): adaptive
radiation, G.Is., 100, 102,
197-98, 199, 200, 201, 205;
bill form and feeding habits,
198, 199; coloration and
feeding habits, 198, 610;

compared with drepanid
birds (Hawaiian honey-
creepers), 158; compared

with mainland families, 200—
201; Darwin’s observations
of, 197; ecological niches
occupied, 198, 199; feeding
habits, 198, 199, 200: inter-
island dispersibility, 46; male
plumage, diet, and foraging
level correlated, 200; re-
source partitioning, 98;
secondary diversification,
201

Germination: after seawater
exposure, 74; rates, 541-42

Gigantism of fauna: adaptive,
23-24, 603-7; birds, 603,
607-9 (see also Flightless
birds; Land pirds):; carabid

beetles, 503, 605-6 (see also
Carabid beetles; Insect gi-
gantism); in caves, 615;

coleopterans, Madeira, 605:
drosophilid flies, H.Is., 605,
606; in insect flight loss, 499.
503, 606—7; in island settings,
23-24, 603; land snails, 49,
128, 232-33, 607; lizards,
on islands, 24, 47, 603-5;
moths, 605; tortoises, G_Is.,
201-2; wetas, N.Z., 5
Gigantism of plants: angio-
sperm species endemic in
N.Z., 441; Argyroxiphium
and Wilkesia spp., 150; flow-
ers, 31-32, 434, 539; fruits,
431, 433, 443-45, 453, 463,
474, 476, 478-80, 482-83 (see
also Fruit size enlargement);
fruit without increased seed
size, 474; loss of dispersi-
bility in, 429-30, 431, 433,
442-43 (see also Plant dis-
persibility loss); seeds, 65,
69, 429-30, 436, 442-43,
480-83, 541 (see "also Seed
dispersal mechanisms)

Subject Index

Gillett, G. W.: on cleisto-
gamous self-pollination, 519;
on sterile ray flower of
Bidens, 515

Gillett, J. B.: on long-distance
dispersal in Africa, 89

Glaciated island, 3

Godley, E. J.: on seed fertility
as selective advantage, 541

Goodeniaceae: climatic factors
favoring, 358; dispersibility
of genera, 467-68; fruit
types, radiation in S.W.A.,

Gossypium (cotton): dispersi-
bility, 471-73; endemic in
Hels. and Tahiti, 8, 81, 471-
7

Grant, P. R.: on bird plumage
color loss in genetic reorgan-
ization, quoted, 610; on bird
plumage color, size, and
physiology, quoted, 603; on
congeneric bird species, Tres
Marias Is., 99

Grant, V.: on evolution of
dioecism, 520; on seed dis-
persal efficiency, 510

Great auk (Pinguinis impen-
nis), 490

Grebe, flightless (Rollandia
micropterum), 490

Greenslade, P. J. M.:
bid adaptations,
493; on color,
and nocturnalism
optera, 503

Greenwood, P. H., on coexis-
tence of prey fishes, 338

Gressitt, J. L.: on aerial trans-
port of insects, 9, 606; on
capture of subantarctic in-
sects, quoted, 495; on colon-
ization by phytophagous in-
sects, 13; on climate in insect
flight loss, 494-95; on insect
distribution, 45; on wing re-
duction, in subantarctic

on cara-
Scotland,
wing size,
in Cole-

Diptera, 499; on Pacific-
Mascarene insect distribu-
tions, 82-83
Ground-dwelling insects (see
Geophily )

Ground- level branching, 320-
21

Sromb -form spectra of plants,
Guadalupe Island (Mexico),
arborescence limited, 362
Guayana highlands, 594-95;
archipelago and insular char-
acter, 595; endemism high,
5 “tepui” formations
in, 594-95
Gulf of California islands (see
California, islands offshore)
Gulf of Guinea islands, bird-
bill enlargement, 607
Guppy, H. B.: on dispersal of
fruits and seeds, 45. 52-53,
76; on “Fijian difficulty,”

644

436; on seawater dispersal,
462; on seed and fruit float-
ability studies, 437; on shift
oe littoral elements inland,

rye in dispersal to islands,

Gymnosperm flora (see also
Conifers): dispersal to N.C.,
215-16, 220; dispersibility
lost in, 441; displacement by
angiosperms, in Pacific, 225—
26; fossils, in Jurassic in
India, 215; in N.C., 215-
26; in N.G., 226; Pacific dis-
tribution limits, 10, 215-16;
parasitic species, N.C., 224,
225; relictualism, 225-26

Gynodioecious species: char-
acteristics, 30, 515-16, 519,
526-30, 527; evolutionary
development, 522, 534, 541;
fertility and, 541; in H_Is.
flora, tabulated, 515-19; in
insular contexts, 526-30;
N.Z. flora, 530; sites
and, 30, 530

Gynomonoecious species: char-
acteristics, 515, 516, 533-34;

in H.JIs. flora, tabulated,
516-19
Habitat characteristics (see

also Biotas; Climatic factors;
Ecological factors; Habitat
types): adaptive aspects, 6—
8, 13-14, 16-18, 21-25, 97-—
98, 105-7, 180-81, 278, 286;
dispersibility factors and,
6-7, 48, 77-82, 85, 251 (see
also Plant dispersal mech-
anisms); diversity of, in
genera formation, 105; estab-
lishment ability and, 13-14,
81-82; genetic variability
and, 19, 511; in insular situ-
ations (see Insular biota
patterns; Oceanic island
biotas); precinctiveness, 488—
89; stability, 487-88, 505
Habitat types: abyssal depths,

343-46; alpine (see equa-
torial alpine below); archi-
pelago (see Archipelago
effect); atoll, 77-81, 167;

beach, 6, 81, 406, 513; bog,
272, 520, 557-58, 562, 599;
cave, 4, 26, 614-15; cloud
forest, 182, 205, 2105) 225;
359-60; continental (see
Continental island biotas;
Continental mainland bio-
tas); epiphytic, 206-7, 370,
440, 592; equatorial alpine,
550-52, 562-63, 570-71, 576;
equatorial highlands, 549-
602 (detailed under Equa-
torial highlands); frost-free,
375, 553; halophytic, 7, 81,
289: insular (see Insular
biota patterns; Oceanic is-

land biotas); karri forest,
282; kipuka, 126; lake, 3-5,
86, 335-40; lava, 13- 14, 62,
126; lower montane forest,
554-57; mediterranean, 5,
83, 85, 363; mesic, 152, 204,
210, 264, 272; moist forest,
206, 282-83; montane scrub-
land, 206, 209; moorland,
557-62; pioneer, 6, 13,
237; rain forest, 6, 184, 238,
357, 360, 393, 407 (see also
Rain-forest habitat): Tipar-
ian, 36, 86, 222, 233, 224,
225; sandy, 264, 320; scrub
forest, 210; subalpine, 164—
65, 550-54, 570-75 (see
also Paramos); subant-
arctic, 494-501, 596; sub-
montane, 553-55; upper
woodland, 575-76; volcanic,
3-4, 18, 20, 45; wet forest,
208, 288-89, 361-62; wet
scrubby upland, 80; xeric,
205, 210, 289, 359, 360, 567
Hagerup, O., on insular factors
affecting pollination, 32
Hakea (Proteaceae), of
S.W.A.: fire adaptation, 326;
leaf morphology and diver-
sity, 283, 284-85, 286, 287;
leaves showing evolutionary
tendencies, 284-85
Hetophyric adaptation, The

Haplochromis (cichlid fishes),
Lake Victoria, 339

Harmony, in biota composi-
tion: adaptive radiation and,
7, 193-94, 227, 341; in con-
tinental mainland habitats,
6, 341, 481; in Malesia, 335;
aTIGIN: CIs In INEZ
233, 249; in Philippines, 335;
plant dispersibility lessened,
7, 88, 227, 483-84; in south-
eastern Polynesian islands,
483

Hawaii (H.Is.), 120-28
Hawaiian Islands fauna: adap-

tation and_ diversification,

20, 125-29, 158-77; banana

moths, 20; birds, 9, 128-29,

158-69 (see also Drepanids);

Diptera, wet-forest adapta-

tion, 126

drosophilid flies, 172-76, 605
(see also Drosophilid
flies); inter-island found-
ers, summary of proposed
number,

herbivore mammals lacking

in indigenous, 138

insects, 125-27, 169-76, 605,
610-12; adaptive radiation
of, compared to angio-
sperm flora, 125-27; dis-
tribution by islands, num-
bers of species, 125

land snails (shells), 28, 127—

28, 176-77, 212, 488, 602

Subject Index

(see also Land snails); liz-
ards, 7

Hawaiian Islands flora: adap-
tive radiation, 125-26, 130-
57, 512-13; adventitious
species, 369-72; aerial flota-
tion, 18, 54, 59-63, .78; an-
cestral elements recutrent in
Mascarene Is. flora, 82;
andromonoecious types, 516—
19; angiosperm diversity,
109-11; tabulated, 112, 120;
arborescence, 387,: 404-7;
arrival and distribution, 59-
97, 119-24; arrival modes,
54-59; deterrents to large
herbivores lacking, 29; di-
chogamous types, 516-19;
dioecious types, 511, 516-19,
521-26, 522, 533-34; dis-
harmonic elements, 536, 541;
dispersal consequences, ratio
of contemporary to immi-
grant species, 58-59; dis-
persal modes and '° mecha-
nisms, 54-77; drift arrival of
species, 76; endemics, 81,
139, 140-41, 142-43, 512,
5225 endemics persistent in
disturbed sites, 28; ferns (see
Ferns; Fern spores); flower
colors, ratios, 540; flowering
seasonality, 363; forest
species, limited by available
pollinators, 32; genetic de-
pauperation, 28; grasses,
arrival, 63; gynodioecious
types, 516-19, 527, 528-29;
hemicryptophytes, 372: hu-
man _ introductions, 17: hy-
bridization, 535; immigrants
to, numbers of, 512; lava
pioneers in, 13-14; macropha-
nerophytes, 370-72; mam-
mal herbivore effects, 262—
63; migrants from boreal re-
gions to, 81; monocotyledon
arrivals, 56-57; New Zea-
land similarities to, 5; North
American elements, 54; out-
crossing, 30, 516-19; phan-
erophytes, 370-72 (see also
Trees); pollination modes
and mechanisms, 515-19,
tabulated, 519; rayless woods
from H.Is. species, 397; re-
productive elements, 515-42,
tabulated, 516-19; seeds and
fruit of endemic Schiedea,
461; seeds of endemic Faba-
ceae and Rhamnaceae, 465;
self-pollination factors, 519;
speciation, 56, 58, 59, 121,

123, 129-47, 535; spectra
of angiosperm life-form
types, compared, 367-72,

tabulated, 370; strand-plant
evolution, 7; vicarious spe-
ciation, 103-6, 118, 122, 155;
wind-pollinated species, per-

645

centages, 533; woodiness, 50,
353, 387, 404-7

Hawaiian Islands habitat (see
also Hawaiian Islands fauna;
Hawaiian Islands flora; Ha-
waiian Islands Leeward
chain): adaptive radiation,
115, 129-77, 512-13; birds,
128-29, 158-69 (see also
Drepanids; Land birds; Plant
dispersal, by bird vectors);
dispersibility lost, 25, 483-
84; ecological diversity, 54,
121, 372; genetically impov-
erished genera, 28; geological
positions, 4, 119-20; humid-
ity, 361; Indo-Malesian char-
acteristics, 16; insects, 125-—
27 (see also Insects; Hawai-
ian Islands fauna); isolation
from South Pacific elements,
76-77, 81; land _ snails
(shells), 28, 127-28, 176—
717 (see also Land snails);
rainfall factor, 159-60; rain

forest, 357, 360, 611
Hawaiian Islands Leeward
chain: adaptive radiation,

212; atolls now, once islands,
4, 67, 212; bird dispersal of
seeds, 67, 76, 80; geological
history, 4, 119; insect coloni-
zation, 126

Heard Island, flightless insect
species, percentage, 498

Hebe (Scrophulariaceae): adap-
tive radiation, N.Z., 235-38;
hybridism, 238; species, 236

Hedberg, O.: on adaptations
for frost survival, 563, 565;
on adaptive radiation in
equatorial highland flora,
552-54, 558, 562; on affin-
ities in Afroalpine flora, 550,
569-70; on insulating effect
of leaf masses, 599

Hemicryptophyte habit, 370,

72; compared to other an-
giosperm forms on selected
islands, 370

Hemsley, W. B., on insular
woodiness, 22, 350

Herbaceous plants woodier on
islands (see Insular wood-
iness )

Herbivore mammals: adaptive
response to, in plants, 263,
275-76, 289, 294, 299-300;
advantages for, in becoming
established, 13; arborescence
limited by, 362; chemical
components of plants, deter-
rents to, 299-300; oceanic
islands _ lack, 29; plant
growth-forms (woodiness)
deterrent to, 289, 291, 362—
63, 568

Herbs, on islands (see Insular
woodiness )

Herbs and therophytes,
to trees and shrubs, 373

ratio

Herre, A. W., on evolution of
cyprinid fishes, 336-37

Heteroblasty: in W.A. flora,
294, 296; xeromorphy as
factor in, 297

Heteropteran insects, shift in
vegetation zone and _ host
specificity, H.Is., 170-72

Heterostyly, in insular flora,
531_

Heterozygosity, 49, 511, 537

Himalayan area: affinities with,
in Afroalpine flora, 569;
insect flight loss and wing-
lessness in, 597

Himatione (see Drepanids)

Hispaniola, increased wood-
iness on,

Homoptera flight loss, 501

Honeycreepers (see Drepanids)

Honolulu, rainfall at,

Host-parasite relations, as in-
sular situation, 614

Host specificity, in insects,
169-71

Human agency: in changing
environment, 571; in evolu-
tion of insular biotas, 4, 367;
in Melanesian highland de-
forestation, 571; in mouse
dispersal, 4; in plant disper-
sal, 77, 86-88, 195, 367-73,
443; in species endangerment,
615

Humidity factor: in arbores-
cence, 359-61; in continental
mainland areas, 361; in H.Is.,
361; in insular climate, 360—
61; as survival factor in
Echium, 184

Hutia, high population density,
613

Hybridization, 30-31, 512-13,
534-41 (see also Interfer-
tility; Outcrossing) ; of
Bidens, H.Is., 535-36; and

bird pollination, 145; in
@xls5359) cy.clicalss in’ INEZ.
flora, 31; of Dubautia, H.Is.,
150; factors which affect,
144-46; of ferns, H.Is., 512;
floral constancy of pollina-
tors, 540-41; genetic drift
(see Genetic drift); genetic
vigor for, in limited popula-
tions, 511-12; of Hebe, N.Z..,
Ree) in H.Is., reported, 535—
in J.F.Is., 535; of lobe-
iieiae by honeycreepers, 144—
46; of Metrosideros, H.Is.,
59; in N.Z., studies of, 534
S5On Pipturus, Hs. 535:
polymorphism in waif flora
and, 513, 519; _ selective
value, in N.Z. flora, 512; as
survival factor, 537; volcan-
ism as stimulus for, 537; in
waif flora, 30-31
Hybrids: autogamy in preserva-
tion of, 537; conditions for
stabilization of, 511

Subject Index

Hydrophyte habit, character
and occurrence, 370

Hymenopterans (flightless) ,
499-501; Aucklandella fla-
vomaculata, Auckland IL.,
500; Kleidotoma subantarc-
tica, Campbell I., 500

Iceland, 74

Iguanas (see also Anoline liz-
ards; Geckos; Lizards):
body and _ feeding-habit
changes, 612-13; genera, in
Madagascar, 333; marine-
adapted, G.Is., 22, 612-13

Immigrant flora, 4, 20-22, 104,
512

India, fossil gymnosperms from
Jurassic, 215

Indo-Malesian area,
7n

Indo-Malesian flora, as dis-
persal source, 7, 212, 342
disjunctive dispersals of, 82—

83:5 to) Ise all6. 5651677:
to Polynesia, 212

and N.C. flora compared,
227]

Indo-Malesian mountains, 552

Inflorescence: as attraction to
bird pollinators, 201; axil-
lary, in insular woodier
plants, 384-87; formation
tendencies in Darwinia, 303

Insect dispersal: by aerial flo-
tation, 45-46, 51, 74, 487;
experiments, 9, 606; gigan-
tism, 605-7; over long dis-

concept,

tances, 46, 82-83, 125-27;
Pacific-Mascarene _ disjunc-
tion, 82-83; spiders, 46, 51,
487

Insect flight loss: in Afro-
alpine species, 596, 597; in
Antarctic, 494; of aquatic

caddis flies (Limnophilidae),
504; atmospheric pressure
and, 596; body changes,
499, 505-6, 595-96, 606;
brachyptery 492, 494, 497;
on Campbell I., percentage
of species, 494; camouflage
495; in Canada, 493; in
carabid beetles, 26, 491-93,
501-4, 605-6; in caves, 26;
climatic factors, 494-501,
595-99: in cockroach fe-
males, 596
in Coleoptera: Madeira, 501-—
4, 502; Mt. Kilimanjaro,
597
color change (see Camou-
flage; Color, adaptation and
change; Melanism); in con-
tinental areas, 489; in crane
flies, moorland zone, Mt.
Kilimanjaro, 598; on Crozet
I., percentage of species,
494; of dipterous species, on
subantarctic islands, 498,

646

499; “disuse”
Perkins, 492; in equatorial
highland zones, 595-99;
geophily, 493, 501, 596;
gigantism (see Gigantism;
Insect gigantism); habitat
stability, 487-88, 505; Heard
I., percentage of species, 494;
of Homoptera males, on
subantarctic islands, 501; of
Hymenoptera, on subantarc-
tic islands, 499, 500; ichneu-
monid wasp, Auckland I.,
499; Lake Baikal region,
504-5; of Lepidoptera, 495—
97; Madeira and other At-
lantic islands, 501-4, 502; of
micropterous beetles, Mt.
Kilimanjaro, 599; of midge,
Antarctic mainland, 498; of
moths, 495-97; in N.Z., 505—
6; nocturnalism, 503-4; of
Orthoptera, Mt. Kiliman-
jaro, 596-97, 597; precinc-
tiveness, 488- 89, ‘501, 505;
selectivity, 491-93: ‘sexual
dimorphism and, 497, 501,
596, 609; sight loss and, 503-
4, 606; in subantarctic is-
lands, percentage by species,
494; survival factor, 491-93;
on Tristan da Cunha, 492:
in wetas, N.Z., 249, 505-6:
wind pressure. in, 492-94,
497, 505, 596; wing reduc-
tion, 497, 499

Insect gigantism: Australian
species, 605; carabid beetles,
499, 503, 605, 606 (detailed
under Carabid beetles); cole-
opterans, Mandeira Is., 502—
3, 605; dipterous species,
Campbell I., 499; drosophilid
flies, H.Is., 605-6; fossil ear-
wig, S.H., 605; insular fac-
tors in, 605-7; moths, 605;
predator pressure, 605;
wetas, N.Z., 605-6

Insectivorous fauna (see also
Insectivorous flora): Dar-
win’s finches as (see Geo-
spizids); honeycreepers, 157—
69 (detailed under Drepan-
ids); tenrecs, Madagascar,
333

Insectivorous flora (see also In-
sectivorous fauna): aquatic
twining plants as, 278; in
S.W.A., 317, 318, 319

Insect pollination: flower con-
spicuity and, 31-32, 514,
539, 541; paucity of, and
hybridization, 365; selectiv-
ity in, 514; speciation ad-
vantages of, 533

Insects (see also Insect dis-
persal; Insect flight loss;
Insect gigantism; Insect pol-
lination) :

theory, of

adaptation: for population
maintenance, 491-92; from
terrestrial to arboreal
habitats, 612
adaptive radiation: in be-
havior, 174-76; compared
with angiosperm flora,
125-27, 169-70; in mor-
phology, 173-74
in Afroalpine habitats, 596—
98; body size changes, 499,
505-6, 595-96, 606; camou-
flage ‘and mimicry, 173-74,
495, 497, 606, 610-11; crane
fly ‘(Tipula subaptera), Mt.
Kilimanjaro, 598; cypnid
wasp. (Encoilidae), 499;
Diptera (see Diptera); Dro-
sophila (see Drosophilid
flies); ecological shift among,
H.Is., 611-12
endemism: in Lake Baikal,
504; in H.Is., 169-76
in G.Is., 202; ground-dwell-
ing (geophilous), 488-89,
501, 596; in H.Is., 125-27,
169-76, 605, 610-11; host-
plant mimicry, 173-74, 610-
ifr host-plant specificity,
170-73; isolation factor,
171-72; in Macaronesian is.,
194-95; in Madeira Is., 501,
502, 503, 504, 605; melan-
ism at high altitudes, 599;
moths, 20, 495-97, 605; in
N.C., 232; in N.Z., 249; phy-
tophagous species, 501; pred-
atory pressures, 503-4, 605-—
6; sight loss, 503-4; in sub-

antarctic islands, 494-501;
vicarious speciation, 598;
wingless (see Insect flight
loss )

Insular biota patterns (see also

Oceanic island biotas) abys-
sal depths display, 34446;
adaptation, 17-18, 22-25,
34-35, 214, 615; adaptive
tadiation, 21-22, 97-116,
342, 353, 538; in Andean
highlands, 614; autogamous
reproduction favored, 513-
14; barriers as evolutionary
factors, 21; biogeographical
models, £2 27; breeding
systems, 509, 542-44: caves
resemble, 5, 614-15; com-
petition reduced, 20-22, 27-
28, 33, 615; composition of
habitats, 17-18; disharmony,
6-8, 18, 101, 341; crest,
bility factor, 5, Ae 489 509-—
10; diversification of species,

20-21, 111-13; dwarfism,
23-24, 607; endemism, 22,
34-35, 114-15; equatorial

highlands resemble, 4, 549-
50, 595; evolutionary de-
velopments, 2, 20-24, 510-
11; extinction recurrent, 70,
538, 615-16; faunal popula-

Insular woodiness:

Subject Index

tions, 29, 608; flight ability
loss, 24-27, 487-91, 595;
food supply factor, 542-43;
freshwater lakes resemble, 4,
335-40, 504-5; genetic de-
pauperation, 28, 113, 513;
gigantism, 23 yl 232, 503,
603-7; glaciated islands and,
sip Guayana highlands re-
semble, 594-95; host-parasite
relations resemble, 26, 614;
hybridization, 534_41; isola-
tion factor, 21, 104-5, 111-
13; long-distance dispersal,
4-5, 16=17, 45; 214)~510;
Malesian alpine flora re-
sembles, 575; mathematical
model, 1-2, 27; melanism,
23-25, 605, 611; mountains
resemble, 521, 614; in non-
island situations, 3-5, 614—
15; paramos resemble, 582—
83; population size limited,
104; precinctiveness favored,
488-89; reproductive pat-
terns, 33, 509, 533-34, 603-5;
in shelf areas in oceans, 342—
43; in S.W.A., 101, 253-56;
in “tepuis’’ of Guayana, 594—
95; in W.A., 352

adaptation
and selection, measures for,
365-73; in Afroalpine flora,
555, 567-70; anatomical
correlations, 389-403; angio-
sperm plasticity for, 351; in
Australia, 421-23; axillary
inflorescences, listed, 384—
87; in Borneo, 409; in Cali-
fornia islands, 412- 13 -sein
Cape Verde Is., 416; cata-
logue, 403-24 (see also spe-
cific localities); on Chatham
I., at low temperatures, 357;
ia (CH Ree Slo Iv stetshy wieloy,
413-15; climatic factors fa-
voring, 351, 355-61, 401;
continental habitats which
display, 23, 404, 423-24; in
Cuba, 417-18; in Desven-
turadas Is., 411; disharmony
factor, 354-55; dispersibility
linked with, 352; in dry habi-
tats, listed, 388-89; in East
African highlands, 423-24,
557, 562-63; ecological fac-
tors favoring, 373; ecological
shift, 361-62; endemic in-
stances, identified, 404-24;
in equatorial alpine habitats,
423-24, 553-94; evolutionary
patterns, 351-52, 557; fac-
tors which produce, 353-65;
in G.Is., 360, 411-12; growth
elements, 402-4; in HLIs.,
404-7; in Hispaniola, 418;
humidity factor, 359-61;
hypothesis of, 352-53; in
J.F.Is., 410-11; in Kenya
plateau region, 555; in Mac-
aronesian is., 353; in Mada-

647

gascar and Comoro Is., 419-
20; in Madeira, 359-60; in
Melanesia, 408-9; measures
of, by selection and adap-
tation, 365-73; in N.Z., 357;
in paramos, 587; pioneering
qualities, 352-53; relict hy-
pothesis of, 351-52; rosette
forms, 351-52 (see also
Rosette habit); speciation
decreases with, 123; temper-
ature effects, 357-58

Insulation of leaves: against
frost, 563-67 (see also Frost
survival); against radiation,
567-68

Intercontinental disjunctions:
Africa—South America, 86—
88; African—Asian flora, 88;
America—Polynesia, 83-84;
birds, role of, 83 (see also
Plant dispersal by birds);
dispersal vectors and mech-
anisms, 3, 10, 16-17, 83-89
(see also Long-distance dis-
persal); ecological factors,
81-82, 85; Hawaiian Is., role
in, 83, 88-89; in Indo-Ma-
lesian flora, 82; in insects, 82—
83; intervening-mountaintop
hypothesis, 84-85; land
snails, 83; Mascarene—Poly-
nesian instances, 82-83; med-
iterranean climate pockets,
83-85; plants displaying,
listed, 73, 84; western North
America-—central Chile, 10,
83-84

Interfertility (see also Hybridi-
zation; Outcrossing): ecolog-
ical factors, 511, 537; taxo-
nomic issues raised by, 513

Invertebrates on reefs, adap-
tative radiation, 343

Islands (see also Continental
island biotas; Insular biota
patterns; Oceanic island
biota): climatic diagram,
356; references to definitions,

Isoétes, cushion-plant form in
Bolivia, 587

Isolation: He adaptive radiation,

103-4; of insects, 170- 74

in archipelago situation, 204;
genetic drift reinforced, 145;
in insular patterning, 5; in
insular speciation by plants,
511; in land-snail speciation,
340; in lizard speciation and
dispersal, 127, 488; in specia-
tion of fauna and flora com-
pared, 103

J.U.C.N. Red Data Book (Vol.
II, Aves), 489

Janzen, D. H., on parasite-host
relations, 614

Java, 570

Johnston Island, 72

Juan Fernandez Islands: adap-
tive radiation: of Astera-
ceae, 206-10, 208, 209;
of flora 206-11, 371
climate, 80, 206, 356, 358;
compared with C.Is., 371
cloud-forest habitat, 210;
dichogamy in flora, 532;
dioecism in flora, 520-21;
ecological factors and leaf
morphology, 207-11, 371;
endemism high in marine
fauna, 35; epiphytic habitat
206-7
flora of: color ratio, 540;
compared with C.Is., 371;
conspicuity, 539; spectrum
of growth forms, 370-72
hybridization, 535; leaves of
Senecioneae rosette shrubs,
in transection, 208, 209;
plant dispersal by birds to,
77, 80; rainfall, 356-58; rain
forest, 210-11, 357; woody
herbs, 410-11
Jurassic period, gymnosperm
fossil of, 215
Juvenilism (see also Paedomor-
phosis): in Argyroxiphium
leaves, 148; in cichlid fishes,
338; in Cyanea, H.Is., 135—
38; in Myodocarpus leaves,
N.C., 231; in Rollandia
foliage, 142-43; in seed
plants, N.Z., 246, 248

Kagu, flightless (Rhynochaetos
jubatus), 232

Kahoolawe (H.Is.), 120, 125

Karri forest habitat, 282

Kauai (H.Is.), 120-27

Keast, A., on colonization of
open niches in Australia, 101

Kerguélen Island: endemism
in marine fauna, 34-35;
flightless moth, 496, 497:

insect flight loss, percentage
of species, 494
Kinglets (Regulus),
tive dispersal, 47
Kipuka habitat, 126
Klimadiagramm — Weltatlas
(Walter, Leith, and Rehder),
355

Koa forest habitat, 131, 135

Kohn, A. S., on adaptation to
habitat niches on reefs, 343

Kokia (tree cottons), affinities
of, 471

Kona (H_.Is.), 137

Koolau Mts. (H.Is.), 122, 123,
142, 156

competi-

, on habitat pres-
sure. in flight loss, 505

Krakatau islets, repopulation
of, 11

Kramer, G., on lizard body-
proportion changes, 24

Krummholz-like vegetation,

142
Kure (H.Is.), 119

Subject Index

Lack, D.: on absence of bird
groups from islands, 10, 340-
41; on adaptive radiation
obliterated by new immi-
grants, 109, 193-94; on
Charles Darwin’s thinking,
197; on lizard clutch size,
542-43; on passerine birds,
H.Is., 103-4; on saturated
vs. unsaturated island con-
ditions, 49

“Lacustrine island,’ 339

Lake Baikal (U.S.S.R.), 335;
cottid fishes, 339-40; gam-
marid shrimps, 340; insect
flight loss, 504-5; molluscs
endemic, 340

Lake Lanao (Mindanao, Phil-
ippines), 20; cyprinid fishes,
335-37; formation of, 335-36

Lake Malawi (Lake Nyasa),
335; cichlid fishes, 335, 337—
39

Lake Tanganyika, 335; cichlid
fishes, morphological spe-
cialization, 338-39

Lake Titicaca, flightless grebe
(Rollandia micropterum)
on, 490

Lake Victoria (East Africa),
335; cichlid fishes, diversi-
fied feeding habits, 339

Lamiaceae, adaptive radiation,
in Macaronesian is., 192, 193

Wanar i(EEIs»), 1205 1224

Land animals: establishment
on islands, 13, 49; evidence
of dispersal, 46-50

Land area requirements (see
also. Territorial factors):
for adaptive radiation, 103,
538; in flight ability loss,
491-92, 502; in genetic vari-
ability, 538

Land birds: arboreal anteced-
ents obscure, 490; in Azores,
C.Is., and Madeira, 490;
colonization of archipelagos,
47; dispersal, 9, 46-47, 104,
193-94, 201, 489-91 (see also
specific localities); dispersi-
bility questioned, 46-47;
dodo (Raphus), Mascarene
Is., 490; endangered and ex-
tinct species, 489; establish-

ment, on islands, 47; feed-
ing habits, 490-91, 608;
nonmigratory adaptations

by, 490-91; parrot (Strigops
habroptilus), INEZ 490;
Rallidae, 488-91; running
ability, 491; sedentary be-
havior, 46; size increase in,
603; solitaire, Mascarene Is.,
490; survival, 491; territorial
Tequirements, 538; in waif
biotas, 4, 26; wren (Xenicus
lyalli), N.Z., 490 (see also
Drepanids; Emu; Geospizids;
Meliphagids, etc.)

648

Land bridges (see also Intercon-
tinental disjunctions): Ant-
arctic, 7; Australo-Asian, for
angiosperm dispersal, con-
jectured, 256; geologically
not indicated for oceanic
islands, 45; role in long-dis-
lance plant dispersal, 3, 6-7,

Land shells (see Land snails)
Land snails: Achatinella
(Achatinellidae), 127, 176,
488, 607; adaptation and
speciation, H.Is., 127-28,
176; adaptive radiation,
compared to insects and
flora, 127
Amastra (Amastridae): adap-
tive radiation, H.Is., 128,
177, 340; number of spe-
cies by islands, 128
birds as predators of, 176;
Bulimulidae, 607; Carelia
spp. noted, 128; Cochlicop-
idae, 607; coloration, non-
protective, 176; dwarfism,
607; ecological differentia-
tion, H.Is., 176; genetic im-
poverishment, H.Is., 28;
gigantism, 49, 128, 2ap=aah
607; in HLIs., 127-28, 176—
77, 488, 607; ‘Laminella Spp.,
by islands, 128; long-distance
dispersal, 49; loss of dispersi-
bility, H.Is., 488; Mascarene-
Polynesian affinities, 83; in
N.C., 232-33, 607; New-
combia spp.,_ distribution,
128; Partulina, adaptive ra-
diation, 127-28; Perdicella
spp., distribution, 128; pro-
tective coloration probable,
176; survival in dispersal,
49; sympatry, H.Is., 176

Lantern fishes (Ceratioidea),
343-46

Lauraceae, dispersal between
Guayana and Macaronesian
is.,

Lava flows: establishment of
plants in, 13-14; relation to
plant dispersal, 50, 62

Laysan (H.Is.): boobies and
terns, 72; dispersal of species
to, 127; flightless teal, 489-
90; rabbit plague, 49; tree-
less atoll habitat, 167

Leaf anatomy and adaptation
(see also Leaf types): an-
giosperms, N.C., 227-31; Ar-
gyroxiphium-Wilkesia, 147-—
48: Banksia, 282; Bidens,
443-50; Brachysema, 264;
cloud-forest habitat in, 225;
Coprosma, N.Z., 238-40;
Cyanea, H.Is., 132, 1335 Cyr-
tandra, H.Is., 155-56; Da-
viesia, S.W.A., 259, 260,
262, 263, 283; Dendroseris,
J.F.Is., 210-11; diversifi-
cation of types, S.W.A., 278;

Dubautia, H.Is., 150-52, 153;
ecological correlations with,
J.F.Is., 207-11; Euphorbia,
H.Is., 156-57; evolutionary
convergence of, W.A., 278,
279; Fitchia spp., 452-53;
Hakea spp., S.W.A., 283,
284; Lecocarpus spp., 454—
55; Lipochaeta spp., 455;
Macraea spp., 454; Oparan-
thus spp., 450-52; Pelea,
H.Is., 154; Railliardia, 151;
reversibility in, 288; Rhe-
tinodendron, J.F.Is., 209;
Robinsonia spp., J.F.Is., 207—
10, 208, 209; Rollandia spp.,
H.Is., 139-43, 140, 141;
Scalesia spp., 453-54; Sym-
phyochaeta, J.F.Is., 207
Leaf types (see also Leaf anat-
omy and adaptation; Plant
growth forms): acicular,
283, 286, 289; broad, 282-—
83; “convergent evolution”

in, 278; ericoid, 288-89, 290
evolution: in S.W.A., 290;
in W.A., 279

evolutionary plasticity in,
278-91, 290; flabelliform.
286; frost resistant, 563-67;
glaucous, 289, 290; isolateral,

283-88; lobate, 282-83:
microphyllous 288-89 (see
also Microphylly); parallel

development of, 288; phyl-
lodes, 286; seendephylledse!
286, 289; succulent, 289-91
(see also Stem succulents);
in S.W.A., 278-91, 287, 290;
terete, 282, 283; trifurcate,
283; venation, 283, 286;
xeromorphy in, 288
Leeward Hawaiian Islands (see

Hawaiian Islands Leeward
chain)
Legumes: in cladode forms,

272, 273; dispersibility, 464—
67 (see also Fabaceae)

Lek behavior, in drosophilids,
174

Lems, K., on dispersal mech-
anisms of island flora, 10
Lemurs (Lemuridae), of Mad-
agascar: adaptive radiation
questioned, 102, 332-33, ref-
ences to, 333; relictualism,
332-33; strategy of survival,
332-33

Epon rela, flight loss by,
495, 5

UHetticn Pee et fale on
tole of wind in Drosophila
survival, 493

Liana, occurrence by islands,
compared with other plant
forms, 370

Life-form spectrum, of angio-
sperm growth, 367-73 (see
also Plant growth forms)

Light exposure, adaptations
for reduction of, 259

Subject Index

Eienctubers, for fire survival,
324, 326

Lindroth, C. H., on viability
and flight loss by insects, 492

Linsey, E. G., on insect selec-
tivity in ypaboeeern pollina-
tion, 514

Lisianski Island (H.Is.), 125

Lizards (see also Iguanas; Rep-
tiles): adaptation to habitat
by, 47, 98, 222;
radiation, 202, 232, 13395
603-7; agility loss on islands,
609; anoline (Anolis), 11,
47, 98, 543; on Baja Cali-
fornia ‘islands, 504-5; body
changes on islands, 24, 605,
609; chameleons, Madagas-
Cas335 clutch size and
longevity, 504-5, 543; com-
petition on islands lacking,
604; continental patterns in
number and size, 604-5;
Easter I. natives,. 7; ecolog-
ical release on islands, 604;
endemic to Madagascar
(Uroplatidae), 333; geckos,

85 9823252335 333. gek-
konids, N.C., 232; as H_LIs.
natives, 7; limb reduction,
Philippines, 609; longevity,

604; melanism, 605; re-
source partitioning, 98; scale
changes on islands, 605; Uta,
marine invertebrate feeding
by, 613

Lobelia (Campanulaceae, sub-
fam. Lobeliaceae) (see
also Lobelioids): adaptive
radiation, H.Is., 146-47; al-
pine habitat adaptations,
563, 565; high-elevation
species, in African moun-
tains, 559,

Lobelioids (Campanulaceae)
(see also specific genera):
adaptation and _ diversifica-
tion, H.Is., 129-47; adaptive
radiation, East Africa, 558;

affinities, 124, 131, 134, 139,
145-47, 149, 151; Andean
occurrence, 360; arbores-

cence, 357-59; bird pollina-
tion, 144, 146; character
displacement, 144, 146; dis-
tribution of genera and spe-
cies by islands, 119-24;
endemic species enumer-
ated, 144; floral conspicuity
and color, 145; genetic drift,
145; hybridization by birds,
139-44; juvenilism, 135-38,
142, 148; moisture and ar-
borescence, 359; ornitho-
phily, 134, 137, 144-46
Long-distance dispersal, 45-95;
adaptations for, 7-11, 16-17,
83-89, 510; Afroalpine flora,
551-52, 569; by air (see
Aerial dispersal; Wind dis-
persal); angiosperm flora,

649

50-55 (see also specific types,

genera, and localities); to
Australia, 256-57; birds,
flightless, 46-47 (see also

Land birds); concepts and
principles, 3-26, 45-95, 435-
37; continental drift and, 3,
10, 16-17, 83-89 (see also

Amphitropical disjunctions;
Continental drift); credi-
bility, 14, 67; disharmony,

6-8, 103-4; to Easter I., 76,
80; ecological factors, 13-14,
48, 72-82, 85, 251; equatorial
highlands, 45, 551-53; Eu-
rope in, 89; evolutionary
aspects, 3, 6, 35; fauna, 6-7,
9-15, 46-50 (see also specific
types); ferns, 54, 61; fruit
(see Fruit dispersal; Seed
dispersal by birds); geckos,
8-9; to H.Is., flora in, 52-
59, 77, 84, 120, 405-7;
insects (see Insect dispersal);
intercontinental, 83-89 (see
also Intercontinental disjunc-
tions; Land _ bridges); to
J.F.Is., 77, 80, 411; land
birds, 46-47, 193-94, 489-91
(see also Land birds); land
snails, 49; to Mascarene Is.,
82-83; to N.C., 215-16, 220,
226, 232-33; to N.Z., 45,
214, 233; pioneer characteris-
tics, 6, 13, 172, 237, 433;
plants by birds (see Seed dis-
persal by birds); Polynesia,
77-82; to regions other than
Polynesia, 82-89; to Samoa,
77-80; sea routes (see Oce-
anic drift; Rafting); seed
morphology as evidence, 46;
South Africa in, 89; uniform
patterns, in various regions,
56

Longevity: birds, 543-44; liz-
ards, 604; perennial plant as
adaptation for, 14; response
to predatory pressures, 33-
34; in wetas, N.Z., 505-6

Lord Hews, Island, speciation
on, 211

Los Angeles (Calif.), humidity
in,

Lowe- McConnell, R. O., on
speciation of cichlid fishes,

338-39
Loxops (Drepanididae) (see
also Drepanids): L. coc-

cinea, bill and feeding ha-
bits, 162; L. parva, bill and
feeding habits, 1/6/; L. sagit-
tirostris, bill and feeding
habits, 162; L. virens, com-
parative bill sizes, 160; L. v.

stejnegeri, bill and feeding
habits, 163; as _ primitive
genus, 128

Luminescent lures, of abyssal
fishes, 346

Luzon (Philippines), 335
Lygaeidae (Heteroptera), H_.Is.,
host specificity, 170

Macaronesian islands flora (see
also Azores; Canary Islands;
Macaronesian islands habi-
tat; Madeira Islands):
adaptive radiation, 182-93,
369-71; African species, 191,
193; angiosperm growth
forms, compared, 370;
Echium spp., 183, 185, 187;
H.Is. resemblances, 196, 371;
J.F.Is. resemblances, 371;
pioneering qualities, 353;
relictual elements, 415;
woodiness, 353, 415

Macaronesian islands habitat
(see also Macaronesian
islands flora): adaptive radi-
ation of fauna, 193-95;
archipelago effect, 47, 180;
bird speciation, 194; cloud
forest, 360; environmental
zones, 180-81; insect groups,
194-95; land-bird dispersal,
46-47, 490; mediterranean
climate, 359: passerine birds
harmonic, 193- 94; region
defined, 180; xeric condi-
tions, 360

MacArthur, R. H., on mathe-
matical formulation of com-
petition by areas, 27

MacArthur, R. H., and E. O.
Wilson: on adaptive radia-
tion: of ants, N.C., 232;

in archipelagos, quoted,

114; of lizards, N.Z., 231
on biota composition, 17;
on colonization displacing
radiation, 104; on continual
displacement of island popu-
lation, 14, 17, 109, 492; on
extinction on islands, 19, 70,

538: on insular situation
concept, ae
McDowall, on endem-

ism and eyunction 616; on
endemism in marine fauna,
34-35; on radiation of ga-
laxiads, N.Z., quoted, 250

Macquarie Island, insect flight
loss on: instances, 496-501,
498; percentages, 494

Macrophanerophyte trees, oc-
currence in island flora, com-
pared, 370-72

Madagascar: adaptation and
speciation, 332-34; conti-
nental island features of,

332-34 (see also Continental
island biotas); herbaceous
species displaying increased
woodiness, 419-20; lemurs,
102, 323-33; lizards, 332-33;
relictualism, 332; skinks, 333;

tenrecs (Tenrecidae), 333;
vanga shrikes (Vangidae),
333

Subject Index

Madeira Islands (see also Ma-
caronesian islands flora); Ma-
caronesian islands habitat:
adaptive radiation, 180-96;
carabid beetles (coleopter-
ans) flightless, 501-4, 502,
503; climate, 359; Echium
spp. from mesic localities,
185; herbaceous plants with
increased woodiness, 415-16;
insect flight loss, 501-4;
insect gigantism, 503, 605;
land-bird dispersal to, 46-47;
rainfall factor in woodiness,
359-60

Maguire, B., on primitive flora
of Guayana highlands, 594

Malagarazi swamps (Tan-
zania), symbiotic competi-
tion among cichlid fishes,
339

Malayan area, 15

Malesian highlands (see also
New Guinea): alpine flora,
570-75; environment, 570—
alte frost survival, 571, 574;
as harmonic biota, 335; mor-
phological adaptations, 571-
74; subalpine flora, 570-71;
upper woodland habitat,
575-76

Malesian “mainland,” 215

Mammals: Australia dishar-
monic for, 101; extinction
of large, Ceylon, 20; her-
bivore limits to arbores-
cence, 362-63; New Cale-
donia lacking, 214-15

Mani, M. S.: on insects at high
altitudes, 595; on melanism
in highland insects, 599

Marion Island, flightless insects,
percentage, 494

Marquesas Islands, 407; adap-
tive radiation absent, 211-12;

endemism low, 115; Pelea
spp., 154

Marsupials: absent from for-
mative N.C., 215; fossil

Thylacinus, W. A., 352
Masafuera Island, 206-11 (see
also Juan Fernandez Islands)
Masatierra Island, 206-11 (see
also Juan Fernandez Islands)
Mascarene Islands: dodos and
solitaries, 490; Polynesian
species elements, 82—83; role
in long-distance dispersal,
45, 82-83, 421; woodiness,
421
Mascarene-Polynesian
in land snails, 83
Mathematical models,
geographical
a

affinity

of bio-
relationships,

1-2, 27
Matlike shrub, 381 (see also
Plant growth forms)
Matthews, W. D., on effective-
ness in long-distance dis-
persal, 12

650

Maui (H.Is.), 120-27

Mauna Kea (H.Is.), 138

Maximal differentiation, as
incipient radiation, 101

Mayr, Ernst, noted, 2; “founder
principle” of, 19; on land
birds in Polynesia, 9; on
limited genetic content of
immigrants, 19; on sedentary
character of land birds, 46

Mediterranean habitat: arbo-
rescence limited by, 363;
characteristics, 359, 363;
Chilean-Californian disjunc-
tion as, 83-85; isolateral
leaves in S.W.A. associated
with, 283; in Macaronesian
is.,

Melanesia: defined, 408; in-
sular woodiness, 408-9

Melanism: adaptive efficiency,
605; of fishes, 336-37; of
insects at high altitudes, 599:
in island settings, 23-25:
of lizards on islands, 605;
of reptiles on islands, 611

Meliphagids, as pollinators of
lobelioids, 144

Mertens, R., on hermaphrodit-
ism of vipers, 333

Mesic habitat, 152, 204, 210,
264, 272

Mesomorphy: adaptation of
leaves toward, 152, 283; con-
cept, 204, 210

Methodological considerations,
in biogeographical inquiry,
1-3

Metrosideros, dispersal, 7, 354

Meusel, H., on relictualism of
woody herbs,
351

Mexico (see also Baja Cali-
fornia; Guadalupe Island;
Tres Marias Islands): tar-
weed distribution in Tertiary,
147

Mice, human introduction of,
4

Macaronesia,

Micronesia,
ation,
Microphylly: adaptation to
xerophytic conditions, INGZ=
237, 240; in ‘Afroalpine flora,
567: cushion plants, 240;
leaf reduction mechanism,
289; leaf types for, 288-89;
in Malesian highland flora,
574; in N.Z. flora, 237-40;

biotic depauper-

paramo shrubs, 587, 590;
thorn shrubs, S.W.A., 276
Midge, flightless, Antarctic

mainland, 498
Midway Island, 72, 119, 120
Migration, and habitat, 487
Migratory birds: habitat fa-
vored by, 86; Longs distance
Po dispersal by, 9-10, 51-
WASHOE movements (see
long-distance dispersal)

Migratory routes, concept of,
relevance for inter-island
dispersal queried, 15 (see
also Intercontinental disjunc-
tions; Continental drift)

Mimicry (see Camouflage)

Miocene times, Mindanao I.
uplift in, 335

Moas, 246, 251

Models of insular patterns, 1-2

Moist forest habitat: in J.F.Is.,
206; leaf form development,

282: in S.W.A., 282-83

Moisture availability, and in-
sular woodiness, 359-64,
406-7

Molluscs (Mollusca) (see also
Land snails): adaptive radi-
ation, Lake Baikal region,
4

Molokai (H.Is.), 120-27

Monkeys, absence of, Mada-
gascar, 333

Monocotyledons: arrival modes
to H.Is., calculated, 56—57;
“broom-shrub” adaptations,
271; cladode form, 272; in
paramos, 590-92; spiraliza-
tion, 294

Monoecious species (see also
Self-pollination): character-
istics, 516, 533-34; in H_Is.
flora, 519; outcrossing, 30;
in plant reproduction, 510-11

pe opbylesis, concept of, 102,

Monotremes,
of, 215
Montane habitat and flora (see
also Equatorial highlands):

criteria for adaptive radia-
tion by species in, 8; endem-
ism in, 472, 552- 53, 557-58,
NO2s STD 594; lower forest,
554-57

Montane scrubland habitat, in
J.F.Is., 206, 209

Moorland zone habitat, East
African, 557-62; arbores-
cence in, 557-58, 562; bird
pollination in, 558: crane fly
(Tipula subaptera) in, 598

Morocco, 46-47, 191

Morphological ’ change,
ecological shift, 167-68

orks, banana-constant, H_Is.,

Moths, flightless: Campbell L.,
495, 496, 497; Kerguélen I.,
496, 497; sexual dimorphism,
497; wing reduction, 497

Moths, gigantism, 605

Mountains, as ansulat contexts,
4, 83- 85, 521,

Mouth breeders, xchlid fishes
as,

Mucuna: deposited by ocean
currents, 11; failure on H_Is.
beaches, 75, 81; seed dis-
poet by boobies, 76; seeds,

early dispersal

and

Subject Index

Muir, F., on insect species for-
mation, quoted, 172

Munro, G. C., on seed dis-
persal by birds, 70-73

Murid rodents, Luzon, 335

Myers, G. S.: on comparative
radiation of freshwater fishes,
339; on endemism of cichlid
fishes, 337-38; on “supralim-
ital specializations” in fishes,
106-7

Myodocarpus, woodiness of,

N.C., 231

Nanophanerophyte shrubs, oc-
currence in island flora,
compared, 370, 371

Natural selection, in plant
growth forms, 367-73 (see
also Plant growth forms)

Necker Island (H.Is.), 76, 123,

125; numbers of = species
represented, 123 :
Nectarivorous birds: drepanids

as, 159-69 (detailed under
Drepanids); Galapagos Is.
lacking, 198; in lobelioid
pollination, 144

Neochanna, air-breathing fishes,
N.Z., 250

Neoteny, 23 (see also Juvenil-
ism; Paedomorphosis)

AS eulldne, by cichlid fishes,
33

New Caledonia: angiosperm
dispersal to, 226, 232; angio-
sperm diversification, 214-
25; animal adaptive radia-
tion, 232-33; ant radiation,
232; Araucaria, adaptation
and diversification, 34, 215-
17, 218; arborescence, 409;
archipelaee effect lacking,

conifer flora, 215-26, 2/8,

29; 221,222, 224° har-
monic character of, 215,
225

continental ‘‘old’ island

habitat, 7, 34, 45, 115, 215,
220; Dacrydium spp., 220-
25, 221, 222; disharmony of
fauna, 233; ecological shifts
into wet-forest habitat, 483;
gekkonid lizards endemic,
232; gymnosperm dispersal
to, 215-16, 220; harmonic
biota elements, 7, 227; in-
creased woodiness, 231, 409;
insect types, 232; land-snail
gigantism, 607; leaf mor-
phology, 227-31; mammals
absent, 214-15; marsupials
absent, 215; Paleo-formation
of, 215; plants from, illus-
trated, 278, 219, 221, 222,
224, 228, 229, 230; primitive
character of flora, 232; ripar-
jan adaptations, 222, 223,
224, 225; skink gigantism

651

(Riopa bocourti), 232; soil
conditions, 227
New Guinea (see also Malesian

highlands): alpine environ-
ment, 570; alpine plants
from, 572, 573; conifers,
226; cushion-plant habit,

572, 574; Epilobium spp.,
affinities, 571, 575; increased
woodiness, 409

New Hampshire, carabid adap-
tations, 493

New Zealand, 233-51; adaptive
radiation, 214, 233-51; air-
breathing fishes (Neochan-
na), 250; alpine flora, 575;

angiosperm flora, 235-49;
arborescence, 357, 409-10;
archipelago effect lacking,

251; armature, plants with,
247; cladode shrubs, 246;
conifers, 4, 233-35; conti-
nental “old” island biota, 7,
45; Coprosma (Rubiaceae)

spp., 239; cushion-plant
habit, 240, 241, 242; “‘cyclic
hybridization” in flora, 31;

dioecism high, 520; dishar-
mony of fauna, 233) 49;
divaricating shrubs, 247:
ecological shift to wet forest,
483; endemism, 240, 242,
441; Epilobium spp., affini-
ties, 570-71; fauna, 249-51;
fishes, galaxiad (Galaxiidae),
249-50; flightless birds, 251,
490; floral colors, percent-
ages, 540; floral conspicuity,
31-32, 539; floral types, 30,
235-49; frogs, 104; geology,
82; gigantism of seeds, 441;
gynodioecism, 530; habitat,
233; Hebe spp., 236; and
H.Is. flora compared, 5, 510;
hybridization, 31, 534-35;
insect flight loss, 249, 505-6;
land birds, 490; long-distance
dispersal ‘to, 45, 214, 233;
microphylly, 237, 243; moas,
246, 251; parrot, flightless
(Strigops habroptilus), 490;
prostrate and cushion plants,
241; relictualism, 4; scree
plant mimicry, 249; sex ratio
in flora, 541; sources of
species, 15; waif character-
istics, 4, 510, 512; wetas
(Orthoptera), 249, 505-6,
606; wet forests, 483
woodiness increased, 409-10;
low temperatures in, 357

xeromorphic shrubs, 243,
246

Nihoa (H.Is.), 156

Niihau (H.Is.), 120-30

Nocturnalism, of insects, 503-4

Nonadaptive elements: con-
ditions of change, 106-7; of
Cyrtandra, H.Is., 155-56;

of Daviesia, Australia, 263;
of drepanids, 146, 159, 168;

Nonadaptive elements (Cont.)
of flora, N.C., 231; floral
characteristics as, 146;
genetic factors as (see Ge-
netic drift); of lobelioids,
H.Is., 146

Noninsular situations, 614 (see
also Continental mainland
biotas)

North America, 54, 56, 83, 101

Now Equatorial Current, 77,

Nutant heads, adapted to ex-
cess moisture, 583-84

Oahu (H.Is.), 120-27, 156

Ocean areas, as insular biotas,
342-43

Ocean currents, and marine
fauna endemism, 34-35

Oceanic drift, and plant and
seed dispersal (see also Plant
dispersal): in adaptive radi-
ation, 513-14; autogamous
species reinforced, 513; Bar-
ringtonia, 355; beach habitat,
77-78, 81; buoyancy factor,
74-75, 312, 430, 434, 471,
475, 482 (see also Seed
buoyancy); Calophyllum
spp., 355; Colubrina spp.,
475; continuity factor, 12,
55; cottons (Gossypium),
81, 471-73; diagrammatic
representation of, 78; as dis-
persal mechanism, 7, 74-77,
312, 430, 442, 464-65; en-
demism affected, 81, 513-14;
to high islands, 436; littoral
tree species, 355; Mucuna
spp., 74; Pacific barriers, 77;
of pantropical species, 56;

plant capability for, 75;
Pritchardia spp., 482-83;
rafting, 9, 11, 48-49, 76;

repetitive types, 55-56, 74—
76, 80-81; Scaevola_ spp.,
312-14, 467-68; to Surtsey,
Iceland, 74

Oceanic island biotas (see also
Continental island biotas;
Insular biota patterns; and
specific localities): adapta-
tion and radiation, 21-24,
97-116, 353, 510-15, 538;
analytical considerations,
1-3; animal predators low,
29; annual plants low, 14;
arborescence (see Insular
woodiness); in Australia,
253-54; Australo-Papuan re-
gion, 15, 101; bird clutch
size, 542-43; bird morphol-
ogy and change, 24-25 (see
also under Bird-bill form;
Drepanids; Geospizids) ;
characteristics definitive of,
5, 7, 14-18; climatic char-
acter 202. 351. 353: 355—
61, 433; conifers low, 354;
continental island biota com-

Subject Index

pared, 1, 3, 4, 108; dishar-
monic character, 6-8, 18,
101, 341; endemism, 34-35,
114-15; evolution, 3-4; evo-
lutionary development of
species, 20-24, 510-11;
faunal populations, 5, 23-25,
29, 604-14 (see also specific
types and genera); flowers
(see Flora on islands; Flower

color; Flower conspicuity;
flower types); food sup-
plies, 542-43; geological

features, 3, 16-17, 34, 45;
herbivore mammals, 29, 362-
63; human agency in evolu-
tion of, 4, 367; humidity
factor, 360-61; hybridiza-
tion, 534-41; insects (see
Insect flight loss; Insect
gigantism; Insects); insular
characteristics (see Insular
biota patterns); literature on,
1-5; lizard populations, 603-
5; long-distance dispersal, 1,
16-17, 45, 214-15, 510;
(see also Dispersibility;
Long-distance dispersal; and
relevant cross-references);
mutations, 33-34; perennials
dominant, 511; predatory
pressures lowered, 24, 33-34,
104-5, 604; relictism, 18,
104-5; reproductive patterns,
33, 509, 533-34; speciation
patterns, 2, 21, 104-5, 111-
13, 537 (see also Speciation
and diversification); volcan-
ism, 3-4, 18, 45; waif biota
elements, 3—5 (see also Waif
elements)

Offshore islands, 334 (see also
Baja California; California,
islands offshore)

Ogcocephalidae
343-46

“Old” continental islands (see
Continental island biotas)

Oliver, W. R. B., on establish-
ment of Comprosma, H.Is.,
155

Oparanthus: adaptation to
rain forest, 450-52; adap-
tive radiation in Polynesia,
212; dispersal capabilities,
450-52; fruits compared, 451

Opuntia (Cactaceae), diversi-
fication, G.Is., 203-4

Orchids: dispersal factor, 61;
in paramos, 590, 592; pau-
city, H.Is., 540

Ornithophily (see also Bird

(batfishes),

pollination): Afroalpine
flora, 568-69; Brachysema,
S.W.A., 264; Clermontia

spp., H.Is., 144; Cyanea spp.,
H.Is., 134, 137; Fitchia spp.,
453; flowers adapted for,
32, 144-45, 307-9, 514; lobe-
lioid spp., H.Is., 144-46; in

652

592; in S.W.A.,

Orthogenesis theory, of F. W.
Went, 282

Orthopterans,
Mt. Kilimanjaro,
N.Z. (see Wetas)

Ostriches, 490, 491

Outcrossing (see also Hybridi-
zation): bog flora poten-
tials, H.Is., 520; conditions
for, 510; dichogamy, 532;
floral types and, 30; genetic
factors, 511-12; gynodioe-
cism, 30, 533-34; in H_Is.
species, rating of, 516-19;
in wet-forest flora, H_Is.,
519-20; with wind pollina-
tion, 510, 532-33

paramos,
07

flightless: on
597; in

Pacific barriers, to south-north
sea flotation, 77

Pacific frontiers, of east-west
dispersal, 10, 45, 215-16,
434

Pacific great depths, insular
habitat, 343

Pacific-Mascarene distribution
of insects, 82-83

Paedomorphosis, 23, 390, 391,
394, 399

Paleocene era, 339

Paleo-formation, N.C., 215

Palms (Pritchardia): adaptive
radiation, H.Is., 156; classi-
fication and_ distribution,
482-83; pioneering qualities,
421; precinctiveness, 432, 482

Paramos, 576-94; bird pollina-
tion in, 592; climate, 576;
elevation, 550; epiphytism,
592; Espeletia spp. as charac-
teristic trait, 576, 581-84;
frost survival, 583, 587; herb-
aceous component, 584-94;
insular characteristics, 582-
83; microphylly, 587, 590;
monocotyledons, 590-92;
shrubby components, 587,
590; vegetational character-
istics, 576, 581; woodiness,
576, 587

Parasitic adaptation: of dro-
sophilid Titanochaeta, H.Is.,
612; of flightless insects, 26,
499: of gymnosperm, Podo-
carpus ustus, 224, 225; as
insular relationship, 26, 614;
in oceanic islands, rare, 29;
of pediculate fish males, 346;
of tapeworm in drepanid
birds, 159; of twining plants,
276

Parenchymatization, in
succulents, 399-401

Partula (land snail): hybridi-
zation, 31; speciation, 21,
212; species isolation, 488

Partulina (land snail): adap-
tive radiation, H.Is., 127,

stem

138; affinities and descent,

a birds, 80, 104, 490,
538; harmonic in Macarone-
sian is., 193-94

Pearl and Hermes Reef (H.Is.),
125

Pediculate fishes: adaptive ra-
diation, 343-46; behavioral
adaptations, 343, 346; Ber-
mudan species of Anten-
narius, table, 345; camou-
flage, 342, 343, 345; lures
displayed by, 343, 344; mim-
icry, 344; parasitic male
dimorphism, 346

Pelea (Rutaceae): adaptive
radiation, H.Is., 154~55;
distribution by islands, 120,
123; morphology of disper-
sion, 477-78; in Marquesas
TIs., 154

Peperomia (Piperaceae), dis-
tribution by islands, 120, 124

Perennials: adaptation for
longevity, 14; factors in high
occurrence on islands, 32;
success in establishment, 14

Perkins, R. C. L.: on coloni-
zation of host plants by
Achatinella, 127; on disuse
theory in insect flight loss,
492: on flower constancy of
Hemignathus, 144; on host-
plant specificity of insects,
171-72; on land-snail dis-
tributions, H.Is., 488; on
land-snail predation, 176

Phanerophyte (tree) habit, oc-
currences, 370-72

Philippine Islands: adaptive
radiation, 335; bird-bill en-
largement reported, 607;
cyprinid fishes endemic, 20,
335-37; as harmonic biota,
335; lizard body changes,
609

Phillipps, W. J., on sexual
dimorphism in bill of huia,
N.Z., 609

Photoperiodicity, adaptation to
latitude, 363-64

Photosynthesis, adaptations to
reduce, 259, 263

Phyllodes, development in
ecological shift, 286, 288

Phylogenetic relations in leaf
types, 283

Phytogeographical area, 115

Picture-winged drosophilas,
Epis. 122

Pioneer habitat: arrival and
establishment, 6, 13, 172;
autogamy, 513; continental
habitat compared, 433; dis-
persal ability and, 6, 172,
237, 352, 433; fire survival
and, 327; hybridization, 519;
insular woody herbs, 352,
353; lava flow, 13-14, 50;

Subject Index

deproducve capabilities,

Pitcairn Island, 408

Pittosporum (Pittosporaceae):
dispersibility, 473-74; seeds,
compared, 473

Placental mammals, 15, 215

Plantago (Plantaginaceae) : al-
pine adaptation, 574; evolu-
tionary plasticity, 35; wind
pollination of, 33

Plant dispersal: by air and
wind (see Aerial dispersal;
Wind dispersal); analytical
considerations for compar-
ing, 435-37; ancestral adap-
tations, 433-34; by bird
vectors, 9-10, 51-52, 63-74
(see also Seed dispersal by
birds); bristles and _ barbs,
51, 54, 62-63, 64, 78, 312-
16, 434; in chancy and
peculiar distributions, 11-12;
concepts and principles, 6—
26, 45-46, 50-90; in equa-
torial highland habitats, 552
(see also Long-distance
dispersal) ; evolutionary
changes, 429; fruits adapted
for (see Fruit dispersal); to
H.Is., range of types, 54—
56; in H.Is. flora, modes
and frequency, 54-77; host
plants as isolating mecha-
nisms, 127; human agency in,
77, 86-88, 195, 367-373,
443; jand-oridee hypothesis,
6-7, 45, 256
modes: in eastern Pacific and

Polynesia, diagrammed,
78; to Pacific islands,

tabulated, 77
oceanic currents, 7, 11, 74—
79, 312, 430, 442, 464-65,
510 (see also Oceanic drift;
Rafting); plumose pappus,
61-62; range of efficiency by
types, 429-32; repeated ar-
rivals, 11; to Samoa, types
compared, 77; seawater flo-
tation, 7, 74, 75, 430, 464
(see also Oceanic drift; Raft-
ing); similarities, in different

areas, 56; “stepping-stone”
islands, 16-17; typology of
modes, 54-56; viscid fruit

and seeds for, 10, 55, 70, 7/,
72, 80, 310-12, 436
Plant dispersibility loss, 429-
86; Acacia spp., 465-66;
Acaena spp., 475-76; as
adaptation, 26, 487-88;
Alectryon spp., 478-80, 479;
andesite line boundary, 434;
Apiaceae spp., 441; Apocy-
naceae, 441; Araliaceae,
442-43; Ve Nall spp.,
434, 454-56; Asteraceae,
432, 443-50, 459- 60; Bidens
spp., 432- 34, 443— 50; Brew-
eria spp., 432; buoyancy, 74—

653

75, 314, 341, 429-32, 467;
Campanulaceae, 460-61;
Canavalia spp., 430, 467:
chemical and physical fac-
tors, 431; in CIs. 431;
Colubrina spp., 475; concept,
429-32; in continental con-
texts, 429, 484; Convolvul-
aceae, 461; Cynareae, 457,
457-58; Dendroseris_ spp.,
460; desiccation, 431; dicoty-
ledons, 441-81; disharmonic
conditions and, 434; Dubau-
tia spp., 455-56; ecological
shift, 26-27, 109, 430-36,
483; Elaeocarpaceae, 483;
endocarp enlargement, 431,
433, 436, 442-43, 468, 477;
Erythrina spp., 464-65;
Euphorbiaceae, 463-64; evo-
lutionary aspects, 51, 429-
35; explanations, 74, 84,
429-37; Fabaceae, 464-66;
Fagara spp., 478; ferns, 437—
40; fertility factor, 431-32;
Fitchia spp., 452-53; in
forest habitat, 433, 453, 467;
geography of, 483-84; ger-
mination and, 74; gigantism
and, 434, 436, 442-43, 477;
Goodeniaceae, 467-68; Gun-
nera spp., 468; gymnosperms,
441; habitat and, 487; Hap-
lostachys spp., 470; har-
monic conditions and, 7,
88, 227, 483-84; Hibiscus
spp., 472; hypothesis for,
432-35 (see also explana-
tions above); inland shift,
434; insular character, 489;
isolation, 430; Kokia spp.,
471-72; large-seeded plants,
429-30; Lauraceae, 470-71;
Lecocarpus spp., 454-55;
Lipochaeta, 455; Macraea
spp., 454; Malvaceae, 471;
methods of comparison, 435—
37; monocotyledons, 482-—
84; on oceanic islands, 7,
489, 511; Oparanthus spp.,
450-52; Pelea spp., 476-78,
477; Phyllostegia spp., 469;
Planchonella spp., 480-81;
Plantago spp., 432, 474;
Platydesma spp., 478; pleio-
tropic gene factor, 434;
Podocarpus spp., 441; pre-
cinctiveness, 432-33; Pteraly-
xia spp., 441-42; in rain-
forest habitat, 431; Sandipus
spp., 480; Santalum spp.,
478; Scaevola spp., 467-68;
Scalesia spp., 454; Schiedea
spp., 461, 461-62; seed en-
largement, 429-31; Senecio-
neae, 458-59; Sophora spp.,
466-67; Stenogyne spp., 432,
469, 470; viability, 431; in
wot ones habitat, 429-33,

Plant growth forms (detailed

under specific headings):
armament, radiating, 291-
93; broom shrub, 243,

246, 263-71, 269;
phyte, 323, 370; chamesy-
coid, 156-57; cladode, 263,
264, 270-75, 270, 273; cork,
323, 325, 563; cushion, 240,
241, 242, 565, 572, 574, 587;
dicotyledons, 57-58, 383,
386, 441-81; divaricating
shrub, 242-43, 247; epiphyte,
370, 592; geophyte, 323, 370;
hemicryptophyte, 370, 372;
hydrophyte, 370; isolateral
leaves, 210; liana, 370; ligno-
tuber, 324, 326; macrophan-
erophyte, 370-72; matlike
shrub, 381; mesomorphic,
152, 204, 210, 283; micro-
phylly, 288-89; percentages,
by islands, 370; phanero-
phyte, 370-72; rosette,
139, 206, 351-52 (see
also Rosette habit); spirali-
zation, 293-94, 295; succu-
lent, 204, 289-91, 358-59,
370-71, 389-90, 398, 399-
403, 523, 568 (see also Stem
succulents); therophyte, 367,
371-73 (see also Thero-
phytes); thorny, 248, 263,
274, 275 (see also thorn-
shrub habit); twining, 276—
78, 277
Pleiotropic genes,
ability loss, 434

in dispersal

Pleistocene glaciation, N.Z.,
234

Pleistocene period, effect on
Pritchardia, 156

Pleomele, affinities and _ dis-
persal, 483 (see also Dracae-
na)

Pliocene relicts, on oceanic
islands, 18

Plumage: changes on islands,

24-25; coloration, 610; pro-
tective value of color, 610

Plumose pappus, in aerial dis-
persal, 61-62, 456, 457

Podocarpus (Podocarpaceae),
221, 222; dispersal ability
loss, 441: diversity, N.Z.,
223-25, 224; parasitic species,
224, 225, 234; in subalpine
zone, Melanesian highlands,
570-71, 575

Pollination: in Afroalpine
flora, 568-69; by birds (see
Bird pollination); color and,
300, 540-41; disharmony
and, 146; evolution of,
S.W.A., 306, 309; flower
constancy and, 14445:
flower massing and, 31-32,
SU S14, «541; 568—69:
flower types for, 514-19;
forest-habitat limitations on,
32; hybridization and, 540—-

Subject Index

41; insects in, 514 (see also

Insect pollination); insular

biota influences, 31-32, 568—

69; in S.W.A. flora, 300-

309

types of structures, 516-19;
comparative success of,
533

wind in, 568, (see also Wind
pollination); woolliness in,
299

Polymorphism: of Cyanea,
H.Is., 138; ecological factors,
193 hybridization, S135 yun
insular species, 509; in in-
sular woodiness, 352; proc-
ess, 535; in waif flora, 513

Polynesia (see also Polynesian
flora; and other Pacific
localities): adaptive radia-
tion low, 76-77, 84, 211-12;
archipelago effect low, 212;
boundary islands, 82; dis-
persal to, long-distance, 10,
S08 710.) 78s) (8255212) i(see
also Long-distance dispersal;
Plant dispersal); ecological
shift, 483 (see also Wet-
forest habitat); land-snail
affinities, 83 (see also Land

snails); Mascarene affinities,
82-83

Polynesian flora: adaptive
radiation, Oparanthus spp.,

212; affinities with Mascar-
ene flora, 50, 82-83; bird
dispersal, 65, 70, 78; disper-
sal, long-distance, 10, 50, 70,
78, 82, 212 (see also Plant
dispersal; Seed dispersal by
birds); fruit of endemic
Bidens, 444; increased wood-
iness, 407-8; Indo-Malesian
component dispersed to, 212
Polyphyletic species, 102
Eolyploigy, in fern spore size,

Population size, in island con-
8

texts, 104,

Preadaptation: concept, 161;
in drepanid birds (honey-
creepers), 161-64 (see also

Secondary
tation)
Precinctiveness (see also Insect

cycles of adap-

flight loss; Plant dispersi-
bility loss): Antennarius,
Bermuda, 344-45; cichlid
fishes, 338; concept, 432,
482, 488-89; fauna, 487-—
508; fern spore size, 438-
40; habitat role, 477-78;

hypothesis for loss of dis-

persibility, 432-37; insect
flight loss and, 501, 505;
insular _ factors, 488-89;

plant dispersibility loss, 432—
33; seed dispersibility loss,
436-37; seed size and, 477—
78; in wind-dispersed species,
437

654

Predatory behavior (see also
Predatory pressure): of
animals on islands low, 24,
29, 491; of deep-sea (pedicu-
late) fishes, 344-46; of in-
sects, 503-4, 605

Predatory pressure (see also
Predatory behavior): adap-
tive response, 24; on ants,
503-4; in cichlid fish adap-
tation, 338; dwarfism as
adaptation, 24; gigantism as
adaptation, 605; on insects
on islands, 503-4, 605-6;
islands display reduced, 24,
33-34, 104-5, 491, 604

Preest, D. S., on seed disper-
sal by birds, 69

Prickly leaves and fruit (see
Thorn shrub habit)

Primitive elements (see Re-
lictual elements)

Primitive flora: in Guayana
highlands, 594; in Malesian
highlands, 575; in N.C., 232

Primitivism: concept, 169; in
drepanids, 169

Primulaceae genera, dispersi-
bility, 474

Pritchardia, in H.Is.: adaptive
radiation, 156; morphology

and ecology, 483
Probability, in plant dispersal,
14

Proctor, V. W., on long-dis-
tance seed dispersal by
shorebirds, 10, 69

Prostrate habit, 240 (see also
Cushion-plant habit): bog
adaptation, 562; in equato-
rial highland flora, 557-58,
562; in N.Z. flora, 24]

Protandry, 30 (see also Out-
crossing); in H.JIs. flora,
516-19

Protective coloration (see Cam-
ouflage )

Protogyny, 30 (see also Out-

crossing); in H.Is. flora,
516-19
Pseudophyllodes, 286, 289
Psittarostra (Drepanididae),
128, 168, 169 (see also

Drepanids)
Pteralyxia (Apocynaceae), loss
of dispersibility, 441-42
Puerto Rico: endemism in,
418; giant flightless carabid,

Scarites, 605-6; insular
woodiness, 418
“Pygmy pine” (Dacrydium

laxifolium), N.Z., 234

Radiating armament, 291-93

Radiation, protection against:
in alpine habitat, 567-68; in
paramos, 583

“Radiation zone” and adap-
tation potentials, 103-4

Rafting, in long-distance plant
dispersal, 9, 11, 48-49, 76

Railliardia spp. of Dubautia
(Asteraceae), 151

Rails (Rallidae), 489-90

Rainfall: altitude factor, 359-
60; arborescence favored,

358-60, 371-72; in C.Is. and
J.F.Is. compared, 371; on
islands, diagram, 356; on

islands, records, 359; in in-
sular woodiness, 359-60; in
S.W.A. flora, 254

Rain-forest habitat (see also
Wet-forest habitat): arbores-
cence, 184, 3, 407; dis-
persibility loss by seeds, 429-—
31; Fitchia fruit enlarge-
ment, 453; in J.F.Is., 210-
11; in N.Z., 238; Oparanthus
adaptation, 450-52; Prit-
chardia fruit enlargement,
428-29: seed _ dispersibility
loss, 429-31; temperatures
in, 357

Rangitoto Island (N.Z.),
pioneers on, 13

Raoulia, cushion-plant evolu-
tion, N.Z., 240

Raouliopsis, cushion-plant evo-
lution, paramos, 587, 589

Rapa, plant dispersal to, 76,
78, 81, 445, 452

Rarotonga: Fitchia speciation,

212, 452; plant dispersibility,

76, 80

Ratites (Ratitae), 489-91

Rattenbury, J. A.: on cyclical
hybridization, N.Z. flora, 31;
on hybridization as survival
factor, 537; on_ selective
value of hybridization, N.Z.
flora, 512; on taxonomy, in-
fertile ‘‘species,” 513

Ray, P. M., and W. E. Alex-
ander, on photoperiodicity,
363-64

Rays, in woody herbs, 394-95

Reefs, invertebrates on, 343

Regulus (kingbirds), dispersal
to Azores and C_Is., 47

Rehn, J. A., on cockroach
flight loss, 596

Relictual elements: alternate
explanations, 3-4, 23, 104~
5, 251-53; conifers, N.C.
and N.Z., 4, 215-16, 225-—
26, 234-35; continental “‘old”
island occurrence, 4, 18,
332-33 (see also Continental
island biotas); distribution
of various, 3-4; in drepanids
(honeycreepers), 169
in flora: in J.F.Is., 18, 371;

in Macaronesia, 191, 413-
16; in N.Z., 4, 233-34

in geospizids, 200; in Gua-
yana highlands, 594-95; in
gymnosperm genera, 225-26;
insular woodiness, 251—53;
of Lauraceae, Madeira, 594;

lava

Subject Index

of lizards, Madagascar, 333;
on oceanic islands, 18, 594;
in Tasmania, 18; in waif
biotas, 3-4, 18; woody herbs
as, 351-53
Repetitive drift,
Oceanic drift)
Reproduction, on islands: of
animals, 542-44; autogam-
ous, 513-14 (see also Self-
pollination); in colonization,
525; disharmony in rate, 542;
ecological factors, 432; floral
morphology, 514-16, 519;
floral pollination types, 514,
519; food supply factor, 33;
Hawatian floral types, 515-—
42; inefficiencies, 33; vege-

74 (see also

tative, 538; woodiness, 541
Reptiles (see also Geckos;
Iguanas; Lizards; Snakes):

behavioral adaptations, 612—
13; body proportion changes,
24; in Madagascar, 532-33;
melanism, 611
Research opportunities: on
endangered species, 616; in
Guayana highlands, 595;
insect flight loss, subantarc-
tic, 494; insect gigantism,
606; insect speciation, G.Is.,
201; insular biota character-
istics, 509
land ‘snails, H.Is., 177; island
and continental sizes of,
49-50
leat anatomy, of Euphorbias,
57

pollination: Afroalpine, 569;
insular species, 539; in
paramos, 592

Resource partitioning: adaptive
radiation as, 98, 343; by
drepanids, 261; by geospi-
zids, 98; by lizards, 98; by
pediculate fishes, 343; in

sympatry, 98-99

Revillagigedo Islands,
persal to, 77-78
Richter, C. J. J., on aerial dis-

persal of nonflying insects,

487; quoted, 488

Ridley, H. N., on bird role in

air dis-

seed dispersal, quoted, 73-
74

Riparian habitat, 86, 222-25,
222, 224

Robinsonia, adaptive radiation,
J.F.Is., 206-11, 208, 209

Rodents: body size on islands,
603; in dense populations,
613

Rollandia: adaptive radiation,
H.Is., 124, 129-43, 140, 141;
arborescence, 139; juvenil-
ism in foliage, 142-43

Rosette habit: alpine adapta-
tion, 574; autochthonous on
islands, 351; characteriza-
tion, 139, 206; on C.Is., 368,

655

378, 379, 382; characteriza-
tion, 139, 206; of Dubautia,
381; evolutionary aspects,
3, 35, 373-74; fire survival
adaptation, 321, 323-26; frost
survival adaptation, 562-67;
in J.F.Is., 3, 35, 208, 209;
mediterranean adaptation,
320; monocarpic shrub, 374;
monocarpic tree, 373; mono-
cotyledon adaptations i in par-
amos, 592; on oceanic is-
lands, 351; paedomorphosis,
23): Senecioneae, J.F.Is., 208,
209; Sonchus spp., endemic
in C.Is., 378, 379; transection
of, Senecioneae, 208, 209;
trunk formation, 368; wood-
iness, 351-52

Rubiaceae: dispersibility, 476;
pout and seed enlargement,

Rutaceae, fruit and seed en-
largement, 476, 477

Saint Helena (island): climatic
conditions, 359; dioecious
flowers, 520; endemism, 115;
flower colors, percentages,
540; fossil earwig, 605; in-
sular woodiness, 416-17;
long-distance dispersal, 45;
rainfall, 356, 358; rain-forest
conditions, 357

Saint Matthew Island (Bering
Sea), 49

Saltwater tolerance, Gossypium
seeds, 471

Samoa: adaptive radiation ab-
sent, 212; dispersal station,
84; fruit dispersal to, 79-80;
fruit enlargement, Elaeo-
carpaceae, 463; plant dispers-
ibility loss, 483; wet-forest
habitat, 80, 483; woodiness,
408

San Clemente Island (Calif.):
aerial dispersal, 77; fruit and
seed dispersal, 80

San Francisco (Calif.), island
species transplanted to, 365—

Santa Barbara (Calif.), island
species transplanted to, 366

Santa Catalina Island (Calif.):
American bison on, 20; tem-
perture range, 358

Santalum spp., dispersibility
loss, 478

Sapindaceae spp., dispersibility
loss, 478-80, 479

Sapotaceae spp., dispersibility
loss, 480-81

Sarcolaenaceae spp., diversi-
fication, Madagascar, 333-34

Sarcopygme spp., diversifica-
tion, Samoa, 212

“Saturation” factor, 2

Scaevola (Goodeniaceae) :
adaptive radiation, H.Is.,

Scaevola (Cont.)
157; affinities, 467; dispersi-
bility loss, 467-68; float-
ability, 312-14; woodiness,
157, 358

Scalesia (Asteraceae): adap-
tive radiation, G.Is., 202-3;
altitudinal distribution, 202—
3, listed, 203; arborescence,
368) dispersibility loss, 453-

Scarites, carabid gigantism,
Puerto Rico, 605-6

Schiedea (Caryophyllaceae):
dispersibility, 461-62, 46];
endemics, H.Is., 461, 519;
floral characteristics, H_lIs.,
517, 527, 529-30; fragrance
prominent, 540; woodiness,
H.Is., 404

Schoener, T. W., on body dif-
gerences in sympatric lizards,
8

Schoener, T. W., and D. H.
Janzen, on insect gigantism,
605

Seasonality, release from, 363-
64

Seawater dispersal (see Oceanic
drift; Rafting; Seed buoy-
ancy)

“Secondary cycles’ of adap-
tive radiation, 100, 161; in
drepanid birds (honeycreep-
ers), 163; in geospizids
(Darwin’s finches), 201

Secondary woodiness, 353-65

Sedentary birds (see Land
birds)

Seed buoyancy (see also
Oceanic drift; Plant dis-
persal): Acacia spp., 465-—
66; Canavalia spp., 430;
Catosperma, tested, 312;

Colubrina spp., 475; Crypto-
carya spp., 470-71; in dis-
persal, long-distance, 431;
ecological shift, 434; Eryth-
rina spp., 465; germination
and, 74, 475; Gossypium
spp., 471; Hibiscus spp., 472;
inland shift and, 434; Kokia
spp., 471; littoral plants,
355; loss, for Canavalia, 430;
palms (Pritchardia), 482;
Pteralyxia spp., 442; Scae-
vola spp., 312, 314; seawater
affects, 75, 431-32, 482; and

viability, 74, 431-32, 471,
475
Seed dispersal by birds: of

algae, noted, 52; in amphi-
tropical disjunctions, 83-84;
of angiosperm flora, 52-58,
63-74, 82-89, 429-83 (see
also specific genera, and “In-
dex of Biological Names’’);
barbs and bristles, 54, 63,
64, 441-76; color factor, 72,

Subject Index

436-37, 474; concepts and
principles, 9-10, 51-53, 431-
32; ducks and waterbirds,
67, 70, 73, 84; to Easter I.,
80; eaten and carried inter-
nally, 55, 63-69, 77-80, 431,
464, 551-52; external modes,
10-12, 54, 56-59, 60, 63-65;
in feathers, 10, 55, 63-65,
64, 70-72, 71, 80, 310-12,
434, 436, 452, 463, 473-74,
473; frugivorous birds, 28,
55-59, 62-69, 66, 68, 79-80,
436, 442, 461-78; of fungi,
noted, 52; germination af-
fected, 74; to G.Is., 79-80;
to H.Is., 54-58, 63-74, 77-
82; on H.Is. Leeward chain,
76; from Indo-Malesia to
H.Is., 67; to J.F.Is., 65, 77—
78, 80; long-distance migra-
tions in, 9-10, 51-52, 226,
490; to Macquarie I., 73;
marine birds, 72-73, 436; in
mud on feet, 55-56, 60, 69—
70, 78, 80, 436, 461; percent-
ages, by types, 65, 77-80; to
Rarotonga, 77, 80; refer-
ences to, 52; to San Cle-
mente I., 80; shorebirds, 10,
52, 67; 69. 73: to Tonga)
69, 76-79, 445; viscid fruit
in, 55, 70, 71, 80, 310-312,
473, 474 (see also Fruit dis-
persal)

Seed dispersal mechanisms (for
specific families, genera, and
species see Plant dispersal;
Plant _dispersibility —_loss;
the “Index of Biological
Names’): achene, 443-74,
444, 447, 451, 454, 457,
459; air and wind (see
Aerial dispersal; Wind dis-
persal); analytical consider-
ation, 429-32; in archipel-
agos, 432; awns, 445-53;
barbs and bristles, 54, 63,
64, 441-76; birds as agents
(see Seed dispersal by
birds); buoyancy (see Oce-
anic drift; Plant dispersal;
Seed buoyancy); capsules,
462, 474; changing modes,
455, 467; in continental
type forests, 433-34; eco-
logical shift, 430-31, 433
(see also Ecological shift;
Wet-forest habitat); evolu-
tion, 430-33, 455, 467, 482-
83; ferns, 437-40 (see also
Ferns; Fern spores); fire
affects, 326-27 (see also Fire
survival); floatability (see
Oceanic drift; Plant disper-
sal; Seed buoyancy); in for-
est habitat, 432-34 (see also
Rain-forest habitat; Wet-for-
est habitat); fruit size af-
fects, 432-83; loss, 429-85
(see details under Plant dis-

656

persibility loss); methods of
investigation, 435-37; oce-
anic drift (see Oceanic
drift); pappus, 61-62, 456-
60; of pioneer plants, 433;
pods, 466; reproductive
rates and, 432; shakers, 461;
of S.W.A. flora, 309-19;
unknown for certain genera,
454-55, 470-71; viability
factor, 430-32 (see also Fruit
dispersal )

Selective factors, in adapta-
tion: for arborescence, 364;
climate, in S.W.A. flora, 258;
for floral gigantism, 434; in
juvenilism, seed plants, N.Z.,
246; in pollination, 514; in
seed dispersal mechanisms,
435
in seed fertility: on islands,

541; in wet forests, 353
in speciation, S.W.A. flora,
256; in thorniness, 248; in
tree and shrub development,
372-73; of wind in insect
flight loss, 596

Self-extinction, due to over-
eating of range, 49

Self-incompatibility, 534

Self-pollination (see also Au-
togamy; Cleistogamy): ad-

vantages of: in dispersal,

525-26; in establishment

of species, 14; in insular

situations, 513-14
in H.Is. flora, 146, 519; in
hybrid preservation, 537; in
pioneer species, 513; tend-
ency to high degree of, 511

Self sterity, in flora, 510, 515,
51

Senecio (Asteraceae): Africa:
alpine habitat adaptation,
562-63, .564; subalpine
species, 561
paramo species, 585
Senecioneae genera: dispersibil-
ity loss, 458-59; rosette
shrubs, J.F.Is., leaf transec-
tions, 208, 209
Sexual dimorphism: in bill
form, huia, N.Z., 609; in
cockroach flight loss, 596;
of fishes, in abyssal depths,
346; in insect flight loss, 497,
501, 596, 609; in insular
representatives of taxonomic
groups, 609-10; in lizard
Anolis, 98; in proterhinid
beetles, H.Is., 609; in sub-
antarctic moths, 497
Seychelles Islands: ‘double
coconut” (Lodoicea), 431;
frog adaptation noted, 211
Shelf areas of oceans, as insu-
lar biotas, 342-43
Shrikes, vanga (Vangidae), in
Madagascar, noted, 333

Shrubs (see also Broom-shrub
habit; Cladode habit; Thorn-
shrub habit): aborescent
tendencies on islands, 388—
89; evolution from herb
types, 373; favored by drier
habitat, 360; favored by se-
lective pressures, 372-73;
formation, 374-75; in ratio
to herbs and _ therophytes,
se woodiness and stature,

Sight loss by insects, 503-4,
606

Simberloff, D. S., on diversity
of island biotas, 110-11

Simpson, G. G., on probability
in long-distance dispersal, 12

Skinks (Scincidae): adapta-
tion for dispersal, 8-9; giant
form (Riopa _ bocourti),
N.C., 232; in Madagascar,
333

Skottsberg, C.: on anomaly of
telicts in J.F.Is. flora, 371;
on dispersal mechanisms in
island flora, 10; on hybrids
in J.F.Is. flora, 535; on
J.F.Is. materials, 206-11; on
J.F.Is. rosettes as relicts, 35;
on resemblances of J.F Is.
and Macaronesian floras, 371

“Small island” effect: on
adaptive radiation, 206, 538;
on bird populations, 538;
extinction rate high in, 70,
538; on migration and pop-
ulation, 492

Snakes: adaptive radiation
lacking in Fijian, 104; dis-
persibility, 7; relictual Boi-
dae, Madagascar, 332

Society Islands: adaptive radi-
ation lacking, 211-12; floral
dispersal to, 10; insular
woody species, 407

Socorro Island (Mexico), seed
dispersal on, 444, 445, 449

Socotra, arborescence on, 359,

420

Sohmer, S. H.: on dioecious
flowers, 515, 522; on eco-
logical factors in interfer-
tility, 537

Solitaires, of Mascarene Is.,
490

Sonchus (Asteraceae): eco-

logical correlations of ves-
sel dimensions, 392, 393;
endemic species, C.Is. and
Madeira Is., 376, 377, 378,
379, 380; growth forms, 381;
increased woodiness, Maca-
ronesian is., 414-16

Sophora (Fabaceae), dispersi-
bility, 466-67

Soulé, M.: on ecological re-
lease, 604; on lizard size,
604-5

Subject Index

Source areas for plant disper-
sal: of Afroalpine flora,
569-70; of congeneric island
species, 99; determinant of
biota composition, 15; dis-
tance expressed mathemati-
cally, 2, 15; distance in
aerial flotation, 51, 78; of
elements in H.Is. flora, 53-
59; Indo-Malesia as, 7, 77,
212

South Africa: continental case
of adaptive radiation, 342;
long-distance dispersal role,
89

South America, 101; angio-
sperm families shared with
Africa, 86-87; climatic dia-
gram, 356, 357; equatorial
highlands (see Paramos)

South Equatorial Current (Pa-
cific), 77, 81

South Georgia Island, insect
flight loss, percentage, 494

South Pacific drift flora, 76-77;
elements not in H.Is., 81

Southwestern Australia, 253-
330; adaptive radiation of
flora, 259-68; aphylly in
thorn shrubs, 275-76; arbo-
rescence, 421-23; Banksia
leaves showing species diver-
sity, 280; biogeographical
province, 253-54; Brachy-
sema, leaf morphology, 264;
broom shrubs, 268, 271;
cladode shrubs, 271-75; cli-
matic factors in floral di-
versification, 258; dishar-
monic factors, 256, 319; en-
demism in flora, degree of,
113, 253-54; floral diversity,
by taxonomic groups, 259-
68; floral diversity, ratios,
110-14; flowering after fire,
328; Goodeniaceae, fruit and
seed, 3/1, 313; habitat and
environment, 254-56; Hakea
spp., leaves of, 284-85, 287;
inflorescences and showy
bracts, 301; insular biota
characteristics, 101, 253-54;
leaf morphology, 278-91,
280-81, 284-85, 290, 292,
295; legumes, in cladode
forms, 272; microphylly in
thorn shrubs, 276; moist
forest habitat, 282-83; speci-
ation factors, 256; spiraliza-
tion of leaves, 293-94, 295;
thorn shrubs, 275-76; twining
plants, 277; underground re-
newal structures, 320; woodi-
ness in flora, 421-23; wool-
liness in angiosperms, 297—
99, 298

Speciation and diversification:
disharmony factor, 101; dis-
persibility of plants, 510-15;
ecological factors, H.Is., 121;

657

endemism in H.Is. flora, cal-
culated, 55-58; endemism in
relation to, 113, 212; habi-
tat and, Macaronesian is.,
181; insular patterns, 21, 28—
29, 104-5, 111-13, 537; in
land snails, 212; mechanisms
and processes, 18, 100-101,
108-9; wind pollination a
limit on, 532-33; woodiness
counter to, 123

Species integrity, autogamy in,

(see also Self-pollina-

tion)

Species maintenance, land
snails, H.Is., 488

Species number, in proportion
to area, 126

Species on islands, H.Is., 120-
22

Species-per-family ratios, 110-
16

Species-per-genus ratios, 110-
16

“Species rich, family poor,”
concept of, 111-12, 115

Spectrum of plant life forms,
367-73

Speith, H. T., on drosophilids:
behavior patterns, quoted,
175; body size, H.Is., 606;
flight loss, 492

Spiders, long-distance disper-
sal, 46, 51, 487

Spiralization, leaf adaptation
by, S.W.A., 293-94, 295

Spore-bearing plants, dispersal,

Spores: chlorophyllous, 440;
dispersal by birds, 54; fern,
437-40 (detailed under Fern
spores)

Stebbins, G. L., Jr., on adap-
tive radiation concept, 102

Stem succulents (see also Suc-
culence): in angiosperm
floras, selected islands, 370;
arborescent plants as, 358—
59; development and mor-
phology, 399-403, 398

Stenogyne (Lamiaceae): fruit
size, H.Is., 469; loss of dis-
persibility, 468-70

Stephens, S. G.: on cotton seed
dispersal, 81; on demonstra-
tion of flotation, 74; on
provenance and_ dispersal,
Polynesian cottons, 471

“Stepping-stone” islands, 16-17

Stolon adaptation, for fire sur-
vival, S.W.A.,

Strand flora: characteristics,
7; dispersibility adaptation,
7; endemism low, 19; estab-
lishment difficulties, 70; G.Is.
habitat favors, 81; reproduc-
tion rates, 432; speciation
low, 70

Strid, A., on nonadaptive ra-
diation in Nigella, 106

Subalpine habitat (see also
Afroalpine flora; Equatorial
highlands): Argyroxiphium
adaptations, 148; climatic
preadaptations in temperate
species, 551; forest, H.Is.,
164-65; in Melanesian high-
lands, 570-71

Subantarctic islands: birds
flightless, 489-91; climate,
596; cypnid wasp (Encoili-
dae) wing reduction, 499;
dipteran body change and
flight loss, 498, 499; homop-
teran males, flight loss, 501;
hymenopteran species, flight-

less, 499-501, 500; insect
flight loss, 494-501, 496, 498,
500, 596; lepidopterans,

flightless, 495; moths, flight-
less, 496, 487; moths, sexual
dimorphism, 497; parasitic
insects, 499; precinctiveness,
488; wind pressure _ influ-
ences, 494-501; wing reduc-
tion, 497, 499

Submontane plateau,
Africa, 553-54

Succulence: as _ fire-survival
adaptation, 322, 323; in leaf
types, 289; 291

Sumatra, 570

Superparamos (see also Equa-
torial highlands; Paramos):
cushion plants, 587; frost-
survival mechanisms, 587;
microphylly, 590

“Supralimital specializations,”
in adaptive radiation, 106-7

Surtsey Island (Iceland): drift
species first, 74; rapid colo-
nization, 12

Survival factors, in adaptation
on islands, 27-29, 33; flight
loss by insects and birds as,
491-93; of lemurs, Mada-
gascar, 332-33; in mutations
disadvantageous elsewhere,
33; in variability, 510

Sympatry and adaptive radia-
tion, 114-15; archipelago
effect in development of, 99—
100, 114-15; of bird species
on islands, 98-101, 114~-15,
160, 164-66, 607-9 (see also
Drepanids; Resource parti-
tioning); of flora, S.W.A.,
257; of land snails, H_Is.,
176; of lobelioids, H.Is., 145;
of Plantago spp., H.Is., 474;
reproductive isolation in,
161; resource partitioning,
98-99

Symphyochaeta, adaptive ra-
diation, J.F.Is., 206-10

East

Tahiti: adaptive radiation, 212;
cottons endemic, 8, 471-72;
lava pioneers, 13-14; long-
distance dispersal station, 84

Subject Index

Talinum guadalupense_ (Por-
tulacaceae), stem succulent,
398

Tarweeds (Asteraceae) (see
also Argyroxiphium, Dubau-
tia, Hemizonia, Railliardia,
Wilkesia): adaptation and
distribution, H.Is., 146-53;
Tertiary distribution, Baja
California, 147

Tasmania: continental island,
18; disharmony, 101; flight-
less hen (Tribonyx mortieri),
491

Taxonomic considerations in
adaptive radiation, 102, 344;
ambiguity of Cyanea, H.Is.,
137-38; of Ceratioidea, 345-—
46; in interfertility in insular
flora, 513; in juvenile spe-
cies, 137-38; in monophy-
lesis, 102, 344

Teal, Laysan, 489-90

Temperate flora: in alpine
habitats, 562-67; in equa-
torial subalpine habitat, 550-
51

Temperature: and _ arbores-
cence, 355-58; and insular
woodiness, 357-58; on
islands, diagram, 356

Tenerife (C.Is.), cloud-forest
habitat, 359

Tenrecs (Tenrecidae), adap-
tive radiation, Madagascar,
333

“Tepuis,” Guayana highlands,
594-95

Terete leaves, 282-83

Terrestrial birds (see
birds)

Territorial aggression, 613

Territorial requirements, of
flightless birds, 538

Tertiary period: tarweed dis-

Land

tribution, Mexico, 147; in
W.A., 253

Tetramolopium spp. distribu-
tion, 575

Tetraplasandra spp., fruits and
seeds compared, 442

Thermoperiodic responses, 363

Therophytes: compared _ to
other angiosperm forms, 367,
370; displacement by trees,
tendency, 372-73; popula-
tion in island flora, 373;
ratio of, with herbs, to trees
and shrubs, 373

Thorne, R. F.: on biological
provinces in islands, 15-16;
on flora shared by Africa
and tropical America, 86-88

Thorn-shrub habit: absence
from H_Is. flora, 248; adap-
tation in flora, S.W.A., 275-—
76; characteristics, 275; as
herbivore deterrence, 263,
275-76; in N.Z. flora, 247,
248; probable juvenilism of

658

Cyanea, H.Is., 137-38; selec-
tive values, 248, 263; in
W.A., 263, 274, 275

“Three dimensionalization,” in
angiosperm flora,
291-94

Thwing, E. W., on land-snail
distribution, H.Is., 128

Tilapia (cichlid fishes), 338

eer P. Q., on H.Is. crow,

5

Tonga, plants dispersed to, 69,
77-719, 445

Tortoises, giant, race forma-
tion in, G.Is., 201-2

Transplant stations, for island
plant studies, 365-67

Tree ferns: adaptation to frost,
574; in Melanesian high-
lands, 571; pioneering quali-
ties, 421

Trees: absence from oceanic
islands, 353; bog habitat and
occurrence, 372; continental
range, 354; dispersibility,
354-55; epiphyte effect on,
439; expected occurrences,
353, 370-72; ratio, with
shrubs, to herbs and thero-
phytes on islands, 373 .

Tres Marias Islands (Mexico),
99, 108, 607

Trifurcate leaves, 283

Tristan da Cunha: insect flight
loss, 492; long-distance dis-
persal role, 45; moth wing
reduction, 497; repeated ar-
rivals, purple gallinules, 11

Trunk formation, in rosette
herbs, 368

Tryon, R., on species pre-
adapted for colonization, 11

Tuamotu Islands, now atolls,
2D)

Turbott, E. G., on marine-bird
feeding on forest-tree fruit,
73

5)

Tussock grass: adaptation to
bog habitat, 562; in paramos,
590

Twining plants, W.A., 276-78,
277

Underground plant structures,
fire resistant, S.W.A., 319-
235520

“Underground shrub,” 268

“Unsaturated” condition, in
species establishment, 14

Upland forest habitat, Mela-
nesia, 575-76

Upper forest
African, 557

montane habitat, J.F.

habitat, East

Uta, lizard, adapted to marine
invertebrate feeding, 613
“Vacant niches,’ in insular

pattern, 5

Vanga_ shrike (Vangidae),
Madagascar, 333

Vascular plants (see also An-
giosperms; Ferns; Gymno-
sperms), in C.Is., 181; cryp-
togams, assumed arrivals, in
H.Is., 54 .

Venezuela: equatorial high-
lands, 550; paramo eleva-
tion, 550 ;

Vertebrates: Australian distri-
bution, 253; fertility of, in
islands, 542-43 (see also
Birds; Lizards; Mammals;
Reptiles; Snakes)

Vessel-element width: in insu-
lar woodiness, 391-94; eco-
logical correlations with,
tabulated, 392

Viability: of | seawater-dis-
persed seeds, 430-32, 464;
of trees bearing epiphytes,
439

Vicarious species: concept,
106; cushion plants in equa-
torial highlands, 571, 587;
Cyrtandra, H.Is., 103, 106,
18591225 155

Victoria (Australia), woodi-
mess at low temperatures,

Virgin Islands, increased
woodiness in, 418

Vlaming, V. L., and V. W.
Proctor, on shorebird effec-
pVenEss in seed dispersal, 69,

Volcanic factor: in oceanic
island habitat, 3-4, 18, 20,
45; stimulus for evolutionary
speciation, 21; stimulus, po-
tential, for hybridization,
537

Vuilleumier, F., on northern
Andean regions as insular,
614

Wagner, W. H., Jr., on ferns
and spores, H.Is., 437-40,

Waif elements: arrival and
evolution on islands, 3—5, 35,
214; in caves, 4; character-
istics, 3, 5-7, 18; dispersal,
6, 7, 12; ecological factors
in establishment, 85; flight
loss considerable, 26; floral
reproduction, 510-11, 539;
in freshwater lakes, 4; ge-
netic components, 19; in
G.Is., 4; in H.Is. develop-
ment, 4, 214, 510; hybridi-
zation, 30-31; in noninsular
habitats, 4-5; in N.Z., 4, 214,
510, 512; in oceanic island
ee, 3-7; relictualism rare,

Wake Island, 72
“Walking fishes”
tiidae), 343-45

(Antenna-

Subject Index

Wallace, A. R., 1; on correla-
tions between floral conspic-
uity and insect presence, 31—
32, 541; on woodiness of
island plants, 365

Wallace’s Line, 15

Wasps: on subantarctic islands,
499; nest-building changes,
H.Is., 612

Waterbirds (see also Land
birds): cormorant (Phala-
crorax harrisii), G.Is., 490
duck (Anas aucklandica): of

Auckland I., 489; of

Campbell I. (Anas auck-

landica nesotis), 489
flightless types, 490; great
auk (Pinguinis impennis),
490; grebe (Rollandia mi-
cropterum), Lake Titicaca,
490; long-distance seed dis-
persal by, 70-73; teal, on

Laysan, 489
Water dispersal, Scaevola,
W.A

Water pollination, 516 E

Watertank plant formation,
387

Weber’s Lines, 15

Weedy plants: adaptive radia-
tion, 107; character release,
99, 100, 107; displacement of
endemics, H.Is., 27-28; es-
tablishment ability, 12-14,
20; evolutionary plasticity,
35, 107; genetic momentum,
18, 20; immigrant capabili-
ties, compared with antique
types, 18; in Macaronesian
is., 195; reproductive rates
for dispersibility, 431; in
wet-forest habitat, 353

Weevil flightless, Madeira Is.,
505

Went, F. W.: “floating gene”
hypothesis of, 243, 282; on
parallelism in leaf morphol-
ogy, 282

Werner, K., on relict hypothe-
sis of insular woodiness, 351

Western Australia: Banksia
spp., 267; Brachysema spp.,
265, 266; broom plants, 268,
269, 271; cladode shrubs,

271-75, 273; climate,

in Cretaceous times,
253; Daviesia spp., 259-63,
260, 261, 262; dimorphism
of plants, 296; fossil marsu-
pial Thylacinus, 352; insu-
larity of angiosperm species,
253; leaf morphology and
evolution, 279

West Indies: anoline lizard,
47; increased woodiness,
417-19; insectivores, 18, 104

Wetas (Orthoptera), N.Z.,
249, 505-6, 605-6; flight
loss of, 505-6

Wet coastal habitat, S.W.A.,

659

Wet-forest habitat (see also
Rain-forest habitat): charac-
teristics, 80, 205, 288-89,
361-62, 406; Cibotium adap-
tation, 438; cloud-forest
comparison, 360; in Cocos
Is., 89; Cyanea adaptation,
131-36; dipterous insects,
H.Is., adaptation, 126; dis-
persibility of seeds lessened,
429-33, 483; drepanid (hon-
eycreeper) adaptation, H.Is.,
158; Dubautia adaptations,
H.Is., 150; in East African
highlands, 554-57; ecological
shifts into, 482-83, 606; fern
spore size enlargement, 437—
40; fruit size enlargement,
476, 482; gigantism of Tet-
raplasandra, 443; in HLIs.,
also. Hawaiian

habitat); insect
shifts into, 606; in N.C., 483;
ie N.Z., 483; in S.W.A., 288—

yee scrubby upland, J.F Is.,

Whitehead, D. R., and C. E.
Jones, on habitat of recipi-
ent island, 81

Williams, E. E.: on ecological
requirements of invading
species, 99; empirical ap-
proach of, 48; on establish-
ment of successive invaders,
11, 47-48, 70

Williams, F. X., on color adap-
mben in dolichopodid flies,

Willis, B., on dating the evo-
lution of cyprinid fishes, 336

Wilson, E. O.: on ant migrants,
14, 47; on loss of pioneer-
ing capabilities, 17

Wind dispersal: adaptive radi-
ation affected, 104; of As-
teraceae, 61; capability for,
lost, 436; examples, 60; in
H.Is. flora, 433; of Leco-
carpus, 455; of Metrosideros,
59; Pentaptilon adaptation,
310; precinctiveness of seeds,
437; of Pritchardia, 482;
tree seeds dependent on, 355

Wind pollination: angiosperms
dependent on, percentage,
532-33; floral size and, 539;
in Florida Keys, 533; in H.
Is., 516-18, 533; insular es-
tablishment and, 514; island
instances, percentage, 532-—
33; outcrossing and, 30; of
Plantago, 33; plant vari-
ability and, 51; speciation
and, 533; survival success
and, 31-33

Wind pressure: in fern and
plant dispersal, 433; in in-
sect flight loss, 492-96, 505,
596; in subantarctic faunal
evolution, 494-501

Wing modification, for swim-
ming, 504

Wing reduction (see also In-
sect flight loss; Land birds):
of cypnid wasp (Encoili-
dae), 499; of insects, 492-
97, 498, 500, 501-6; of
moths, 496, 497; nocturnal-
ism and, 503; sexual dimor-
phism in, 497

Wollaston, T. V., on insect
flight loss, Madeira, 501-4

Woodiness (see also Arbores-

cence; Insular woodiness;
Woody herbs): anatomical
correlations, 389-403, 400;

anomalous secondary thick-
ening, insular plants, 386;
axial parenchyma, 401; of
Bidens, 450; catalogue of, on
Oceanic islands, 404—23; of
Cyanea, H.Is., 135; extraxy-
lary fibers, on islands, 403;
factors in development, 253;
herbivore mammal pressure
and, 362-63; hypothesis of
insular development, 350-
53; libriform fibers and,
394, 399, 401-2; measures
of selection and adaptation
for, 365-73; mediterranean
climate and, 363; moisture
availability and, 364, 393,
406-7; in paramo cushion
plants, 587; parenchymati-
zation, 399-401; polymor-
phism, 352-53; ray develop-
ment, 394-95, 396; rosette
formations, 351-52; second-
ary factors in development
of, 353-65; secondary xylem,
350-51; selective pressures

Subject Index

favoring, 372-73; tempera-
ture factor, 355-58; tenden-
cies, in noninsular contexts,
351; xeric habitat and, 351,
388-89

Woody herbs (see also Arbo-
rescence; Insular woodiness;
Woodiness): in Africa, 423-
24; anatomical details, 400;
autochthonous evolution on
islands, hypothesis, 350-53;
in Australia, 421-23; axillary
inflorescences, 384-87; in
Azores, 416; in California
offshore islands, 412-13; in
Cape Verde Is., 416; in C_Is.,
413-15; in Cuba, 417-18; in
Desventuradas Is., 411; in
equatorial montane plateau,
553-54; evolution of growth

forms, 373-89; fire resist-
ence, S.W.A., 323-26 (see
also. Fire survival); frost

survival, 364 (see also Frost

survival); in G.Is., 412; in
H.Is., 405-7; in Hispaniola,
418; in LF Is., 410-11; in

Macaronesian is., 415; in
Madagascar and Comoro
Is., 419-20; in Madeira,
359-60, 415-16; in Mas-
carene Is., 421; in Melanesia,
408-9; in N.C., 409; in N.G.,
409; in N.Z., 409-10; in
Puerto Rico, 418; in Saint
Helena, 358, 416-17; in
Socotra, 420; in South Amer-
ica, 423-24, 576; in southern
Polynesia, 408; in Virgin Is.,
418-19

Woody plant speciation, less
than herbaceous, 123

660

Woolliness: in _ equatorial
highland flora, 584, 587; as
herbivore deterrent, 299:
insulation in paramo plants,
aap in superparamo flora,

World flora, natural spectra of
growth forms compared, 370

Wren (Xenicus lyallii), flight-
less, N.Z., 490

Xeric habitat: arborescence
in, 359; climate, 360; as
deterrent to woodiness, 351;
in G.Is., 205, 360; leaf ’adap-
tations ‘to, S.W.A., 278-91,
292, 295; shrubbiness, Ma-
caronesian is., favored, 360

AETOMOEP DY, 210, ei, sil,

Xylem: adaptation to habitat,
152, 192, 390-403; second-
ary, development “of, 350-
51, 389-95

Yoshimoto, C. M., on insect
ane spider dispersal, 45, 51,

Zimmermann, E. C.: on color
adaptation in dolichopodid
flies, quoted, 174; on ende-
mic insects, H.Is., 169-70;
on protective coloration of
H.Is. insects, 173; on speci-
ficity between insect species
and plant hosts, quoted, 171

Zwaluwenberg, R. H. van: on
repeat arrivals of monarch
butterflies, Canton I., 11; on
role of ecology in dispersals
to Canton I., 14