RESULTS OF THE FIRST JOINT

US— USSR CENTRAL PACIFIC

EXPEDITION (BERPAC)

W^r

**

Hi

W-WM^m

AUTUMN 1988

UNITED STATES DEPARTMENT OF THE INTERIOR / Fish and Wildlife Service

OtMCO

RESULTS OF THE FIRST JOINT US-USSR

CENTRAL PACIFIC EXPEDITION
H (BERPAC)

Oi
1 =

s>-
SI

a
iP-
i un

! -D
: eO
i □

I a

• i-^
i a
: m
= o

AUTUMN 1988

\&iQQQS biCLZ

Results of the First Joint
US-USSR Central Pacific Expedition
(BERPAC)

Autum 1988

John F. Turner
Director, US Fish and

Wildlife Service,
Washington, DC

Yuriy A. Izrael
Chairman. USSR State

Committee for Hydrometeorology
Moscow, USSR

Harold J. O'Connor

Project Leader, USA

US Fish and Wildlife Service

Patuxent Wildlife Research Center

Laurel. Maryland

Alia V. Tsyban

Project Leader, USSR

Institute of Global Climate and Ecology

State Committee for Hydrometeorology

Academy of Sciences

Moscow, USSR

Copies of this publication may be obtained from the Publications Unit. US Fish and Wildlife
Service, 1 849 C Street, NW, Mail Stop 130— ARLSQ. Washington. DC 20240.

Suggested Citation:

Nagel, P. A. (ed. ) ( 1 992). Results of the First Joint US-USSR Central Pacific Expedition

(BERPAC), Autumn 1988. US Fish and Wildlife Service. Washington. DC.

Disclaimer:

The opinions and recommendations expressed in this report are those of the authors and do not necessarily reflect the
views of the US Fish and Wildlife Service, nor does the mention of trade names constitute endorsement or
recommendation for use by the Federal Government.

Foreword

Included here in its entirety is the BERPAC Program paper, published in 1991 [J.F. Turner, H.J. O'Connor, Yu.A. Izrael, and
A. V. Tsyban (eds. )( 1991 )]: BERPAC — A Program for Long-term Ecological Research of Ecosystems of the Bering and Chukchi
Seas and the Pacific Ocean. National Fund for the Patuxent Wildlife Research Center, Bowie, Maryland]. BERPAC refers to
joint research of the US and USSR in the Bering Sea and central Pacific Ocean since 1 977. The Bering Sea portion of the project
has included three joint research expeditions (1977, 1984, and 1988) between the two countries. The central Pacific portion of
the project was established in 1 988 when the first joint expedition took place. The central Pacific segment, therefore, has not had
the opportunity to go through the same maturing process as that of the Bering Sea. As time goes on, and research continues, goals
and objectives to be accomplished will be further developed. The paper is included here as a Foreword to show the correlation
with this volume 's sister monograph. Results of the Third Joint US-USSR Bering & Chukchi Seas Expedition (BERPAC), Summer
1988.

BERPAC

A Program for Long-term Ecological Research of Ecosystems of the Bering and Chukchi Seas and the Pacific Ocean

Introduction

Deterioration of ecosystems on a large scale threatens
many functional equilibria in the biosphere. This problem is
particularly urgent for the World Ocean, which is the sink for
many different pollutants that can produce significant ecological
impacts.

The ocean is able to assimilate a certain amount of
anthropogenic compounds due to "self-purification" without
visible deterioration of the ecosystem. However, continuous
increase in the flux of pollutants to the ocean creates the need
for study of the resistance of marine ecosystems to anthropogenic
impacts. Investigations of ecological consequences and
elucidation of causal relationships between the impact levels
and adverse biological effects are only poorly understood for
the marine environment. The study of these interactions and
responses is interdisciplinary in character and covers different
fields of biology, ecology, chemistry, and physics of the sea.

The dynamics of marine ecosystems, including biological
and physical processes and biogeochemical cycles, are closely
related to changes in the climate of the Earth. The predicted
global warming may have a pronounced effect on certain vital
processes in the World Ocean, especially the resistance of its
ecosystems to anthropogenic contamination. This is because
the living ocean determines, to a great degree, the normal
functions of the Earth's climatic system.

Long-term observations of physical, geochemical, and
hydrobiological processes are necessary for the assessment of
ecological consequences of contamination in the ocean
environment and isolation of local anthropogenic effects
compared to the effect of climatic variability.

The Bering Sea is located between the coasts of the Soviet
Far East (USSR) and Alaska (USA), and naturally an interest
in the study of its ecosystems has been shown by Soviet and
American scientists (Izrael & Tsyban, 1983a, 1977, 1990;
Roscigno, 1990).

In spite of comprehensive studies carried out in the Bering
Sea in the last few years (Izrael etai, 1988b; Izrael & Tsyban,
1989, 1990; Coachman, 1990; Roscigno, 1990), a number of
the oceanographic, hydrochemical. and biological parameters

determining its ecosystem functions are as yet poorly known
when compared with, for instance, the Baltic, Mediterranean,
and Black Seas. For example, the joint bilateral program of
Bering/Chukchi investigations have been carried out for more
than 13 years with the production of three monographs of
cruise results. However, the as yet inadequate data on the
characteristics and processes occurring in the ecosystems of
the Bering Sea and North Pacific waters has led to the
organization and implementation of an international program:
Long-term Ecological Research of Ecosystems of the Bering
and Chukchi Seas and the Pacific Ocean (BERPAC Program).

Goals, Objectives, and Scientific Basis of the BERPAC
Program

Goals

The goal of the BERPAC Program is to examine the status
of marine ecosystems of the Pacific Ocean, Bering Sea, and
Chukchi Sea and to assess their role in determining global
climate. BERPAC will study the dynamics of these ecosystems
related to conditions of global climate change and anthropogenic
contamination.

Objectives and Scientific Basis of the BERPAC Program

Objectives of the BERPAC Program consist of the study
of the biogeochemical cycles of contaminants, related
oceanographic processes, and food-web interactions in the
North Pacific waters that flow through the Bering/Chukchi
Seas, including study of the behavior of organic pollutants at
the water/sediment interface, since sediments are sources of
the secondary pollution of ecosystems. Important topics of
study are the control and the accumulation of pollutants in
bottom deposits, and the study of their migration within the
sediments and exchange with overlying waters are important
topics of study.

1. Assessment of Ecological Consequences of Contamination

Progressively severe changes in chemical contamination

of the ocean biosphere are on the increase. Anthropogenic

in

impacts influence not only the biotic component of the marine
environment but different abiotic components as well. Such
impacts lead to even more significant changes in the World
Ocean and in the biosphere as a whole.

Specific features of the Bering Sea and other ecosystems
w ith "background" levels of contamination are such that they
are especially vulnerable because of the continual input of
small doses of pollution. This leads to a gradual accumulation
of pollutants, and may ultimately cause the degradation of the
ecosystems. Therefore, ecological investigations and
monitoring of the background regions of the ocean, especially
in such highly bioproductive zones as the Bering Sea, are of
g reat importance. In order to assess the ecological consequences
of the pollution and isolate anthropogenic effects from the
background of natural variability, it is necessary to make long-
term observations of fundamental physical, chemical and
biological processes in selected areas of the above regions.
These regions differ in their geographical location, as well as
in the subsystems of their ecosystems, and are subjected to
different anthropogenic impacts.

2. Study of the Processes Determining the Assimilative
Capacity for Contaminants in Marine Ecosystems

In the marine environment, various physical, chemical,
and biological processes occur through which contaminants
can be eliminated from the ecosystem without serious
disturbances of the biogeochemical cycles of the elements or
changes in the biota. Diverse oceanological investigations
carried out in the last few years have shown that the biotic
component is important in the fluxes of pollutants.

The ability of an ecosystem to protect itself against a
foreign interference at the expense of many biological, physical,
and chemical processes is its natural ■■immunity"' and the
measure of this immunity is its assimilative capacity.

According to the contemporary interpretation (Izrael &
Tsyban. 1983b, 1989; Izrael et a/., 1988b.c), the assimilative
capacity of a marine ecosystem is an integral function of its
ex ist ing environmental status that reflects the ability of physical,
chemical, and biological processes forelimination of pollutants
and their impacts on the biota.

When using the concept of assimilative capacity in practice,
it is necessary to bear in mind that a marine ecosystem occupies
a finite volume that may be isolated on the basis of the spatial
distribution of organisms of various trophic levels, groups of
ecologically similar species, and production/destruction
processes, as well as physical and chemical characteristics.
Hence, the assimilative capacity of each specific ecosystem
also has a value that objectively characterizes existing properties
of the marine em ironment. This value could be determined in
practice on the basis of integrated investigations and monitoring
of the marine environment, carried out in accordance with
existing methodological recommendations (Izrael & Tsyban.
1983b, 1985, 1987, 1989; Izrael etal., 1988b).

The use of this concept in the BERPAC studies will
include investigations of the following basic problems:
/. quantitative assessment ol the balance of chemical elements
in the ecosystem and possible changes in residence times due
to disturbances: 2. assessment of adverse biological effects at

the level of population and communities: and.v determination
ol the critical concentrations at whichcontaminants adversely
impact the marine organisms and communities.

Thus, a conceptual model of the assimilative capacity,
based on a better understanding of the laws of marine ecosy stem
functions, can serve as a theoretical basis for the dev elopment
of forecasts of both the immediate and long-range
consequences of anthropogenic and climatic impacts on the
ocean ecosystems.

3. Study of the Elements of the Biogeochemical Carbon Cycle
and Its Role in Global Climatic Processes

Global warming predicted in connection with the
developing greenhouse effect depends directly upon the
biogeochemical cycle of carbon — the most important process
forming the Earth's climate. The basic elements of this cycle
are carbon-dioxide and other "greenhouse gases" exchanged
within the ocean-atmosphere system, the function of the
carbonate system, and the turnover of organic forms of carbon
in the ocean.

The most intensive uptake of atmospheric CO, occurs at
high latitudes, as a result of favorable thermal and hydrological
conditions (low sea surface temperature and permanent
downwelling) in the region. These peculiarities explain the
important role of the Bering Sea, a subarctic body of water
having a large area, in the global cycle of carbon dioxide.

The relationship between the rates and directions of CO,
flow within the ocean-atmosphere system directly affects the
functioning of the carbonate system. So. in the conditions
where global warming is induced by an increase in the
concentration of atmospheric CO:. a shift of the equilibrium
between carbonate forms of carbon in seawater might occur,
which will be accompanied by a decrease of pH and.
consequently, elevation of the lysocline.

Investigations of these processes, directly affecting the
sedimentation of organic carbon and the vital functions of
marine organisms, are only possible with direct determination
of all components of the carbonate system (i.e.. HCO„ CO,,
H:CO,. andCO,).

To fully understand all of the characteristics of the oceanic
portion of the global carbon cycle, it is necessary to stud) the
processes of the circulation of its organic forms in the
composition of dissolved and particulate matter and in the cells
of living organisms (Zaitsev. 1970. 1980, 1985).

The dynamic equilibrium of dissolved and particulate
organic matter, living matter, and the content of organic carbon
within water masses depends on the relations between
production/destruction processes established in the ecosystem.
In this connection, the predicted effects of global warming on
the bioproductivity of the Bering Sea ecosystem will influence
the organic carbon cycle. In order to studv possible changes,
long-term observations of the concentrations of all organic
forms of carbon are necessary.

Thus, to establish the carbon balance in the Bering Sea
ecosystem, comprehensive long-term observations of all carbon
constituents in the aquatic interface and the study of quantitative
and qualitative composition of both the carbonate system and
organic forms of carbon are required.

IV

4. In\ estimation ot'the Physical Mechanisms Related toClimate
Variations

Existing global physical models ot'the ocean-atmosphere
system do not make it possible to predict possible climate
changes on a regional scale because of the extreme complexity
of the modeled systems. Additional investigations of the
physical development of regional models, in particular of a
model for the Bering Sea. are an important need for long-term
climate forecasting at the present time.

This problem could be solved on the basis of long-term
oceanological observations in different regions of the Bering
Sea. which are aimed at the acquisition of systematic information
on the vertical distribution of temperature, heat content of the
active layer and its variability with time, the structure and
variability of ocean circulation, heat transfer by the basic sea
currents, and heat and moisture fluxes across the sea surface.

To develop the above models it is necessary to know the
regularity of water mass formation in the deep basins of the
Bering Sea. The following issues are not yet clear: North
Pacific water must be involved in bottom water formation, but
given the topographic isolation of Bowers and the central
basins, how and where docs this lake place? Are sources the
same for the different basins' What are the flushing rates (e.g..
residence times)?

There are three hypothetical mechanisms by which bottom
water might possibly be formed: /. modification of surface
(upper layer) water within the confines of the sea by cooling
and brine enhancement through ice formation, creating water
sufficiently dense to sink to the bottom: 2. subsurface mixings
of North Pacific water with appropriate Bering Sea waters as it
crosses the sills in the Aleutian-Komandorskiy Island arc
passages; and 3. direct advection of deep North Pacific water
in through Kamchatka Strait and then sequentially through the
gaps into the other basins.

The BERPAC Program will investigate the mechanism of
deep water formation, renewal rates, and Hushing of the basins.

Area of Investigations

While selecting the study areas and location of stations in
the Bering Sea. the diversity and contrast of ecological conditions
in different regions of the sea were taken into account.

In order to reflect a variety of ecological conditions in the
Bering Sea more completely, it seems appropriate that integrated
expeditions include work on polygons located in different
areas of the sea (with the purpose of obtaining representative
data on the structure and functions of the basic marine
ecosystems) and work across transects (with the purpose of
determining the space and time variations of the key ecological
parameters).

Investigations within the framework of BERPAC will be
conducted on four polygons where investigations were carried
out in 1 98 1 (during the integrated ecological expedition aboard
the research vessel ( R/V ) Akademik Shirshov) and in 1 984 and

1988 (during the second and third Soviet-American ecological
expeditions aboard the R/V Akademik Korolev) (Izrael &
Tsyban. 1987, 1990; Izrael etcd.. 1988a; Roscigno. 1990).

Deep stations will be repeated at four centered polygons in
the four deep basins. The center station of each polygon will
also be a location for a mooring containing sediment traps and
current meters, funding permitting. Four other mooring
locations will cover the entrance from the North Pacific (in the
deep channel northwest of Komandorskiy Island), the main
gaps in the ridges north of Attu, and a location on the east side
of the Central Basin under the Bering Slope Current. The
mooring locations are also deep oceanographic stations, and
1 1 additional stations will provide continuity among the deep
waters.

In addition to polygons, observations are planned at stations
along the transects located in areas that are not yet completely
understood, such as the Gulf of Anadyr, the Chirikov Basin, the
Gulf of Alaska, the northern portion of the Pacific Ocean, and
the deep-water central and southwestern areas of the sea.
Larger scale studies in the Chukchi Sea and central Pacific
ecosystems are also planned. The program for individual
expeditions will be discussed specifically during joint symposia.

Proposed Observations

Complex observations during the ecological expeditions
include meteorological (including aerological and geophysical
studies), oceanographical. and ecological observations.
Specifically, the following observations will be made:

A. Meteorological observations will include routine
observations of meteorological parameters, such as studies of
direct solar radiation intensity and ultraviolet irradiation, cloud
and cloud type studies, and collection of samples of atmospheric
precipitation for chemical analyses. Aerological and
geophysical observations will include temperature and wind
sounding with the aid of radiosondes. Air samples will be
collected lor determination of sulfates and nitrogen oxides.
Visual observations of oil and oil product contamination on the
sea surface will be recorded.

B. Oceanographic observations at designated sampling
depths in the water column will include temperature, salinity,
nutrients, oxygen content, water color and transparency,
biogenic elements, alkalinity, and petroleum hydrocarbons.
Tracers for water mass types will include stable isotope content
of seawater (oxygen, deuterium, tritium, freons, silica, and
carbon 14). In addition, current velocity and direction will be
determined, and sediment trap collections will be made.

C. Ecological observations will include studies of the
atmosphere, sea surface microlayer, water column, and bottom
deposits in the environment.

/. Atmosphere

In rainfall. pH and the content of organic contaminants
will be determined. In dust particles, the content of organic
contaminants and metals will be determined. In the air at the

sea surface, the content of "greenhouse" gases (C02, nitrogen
oxides), oxygen, and chlorinated hydrocarbons will be
determined.

2. Sea surface microlayer. water column, and bottom
deposits

Water samples will be collected in the surface microlayer
and at standard hydrological depths and at selected experimental
depths (e.g., thermocline, pycnoline, phyto- and zooplankton
maxima and sediment-water interface) (Zaitsev, 1980).

a. In the surface microlayer, the following elements and
parameters will be determined:

— organic carbon;

- contaminants (toxic metals, and aliphatic aromatic
and chlorinated hydrocarbons), the state of neustonic
communities; determination of the structural
characteristics of bacterioplankton; total numbers,
biomass of microorganisms, most probable numbers
(MPN) of indicator groups of bacteria (e.g., paraffin-
oxidizers, PCB-transforming and neurotrophic
saprophyte groups), and indices of phyto- and
zooneuston (numbers, biomass, species, size
composition, species mass and indicator forms),
mutation (teratogenesis) of zooneuston organisms.

b. In the water column, the following parameters will be
determined:

— water optical indices;

— contaminants (toxic metals, and aromatic, aliphatic
and chlorinated hydrocarbons);

• the total concentrations of organic carbon and its
composition;

— elements of the carbonate system (CO,, HCO„ CO:);

— characteristics of bacterioplankton (total numbers,
biomass, MPN, and distribution of indicator groups);
and their biochemical and genetic capacities;
structural characteristics of phyto-, microzoo-, and
mesozooplankton ( numbers, biomass, size, and species
composition, species mass, and indicator forms);
functional characteristics of planktonic communities
(heterotrophic CO, assimilation by bacteria, bacterial
production, phytoplankton productivity); and
biosedimentation rate of particulate matter.

c. In the biota, the following parameters will be
determined:

contaminants (toxic metals, and aromatic, chlorinated
and aliphatic hydrocarbons; and
organic carbon content, stable carbon, and nitrogen
isotope content.

d. In bottom sediments, the following elements will be
determined:

determinants (toxic metals, and aromatic, chlorinated
and aliphatic hydrocarbons).
total organic carbon and nitrogen;
stable carbon and nitrogen isotopes; and

— structural characteristics of zoobenthos (numbers,
biomass, species composition, and species mass).

3. Higher trophic levels

During the expedition, zoological observations will be
carried out: numbers, distribution, and migrator)' patterns of
fish, birds, and marine mammals.

4. Model experiments

Model experiments will be performed under conditions
similar to natural situations. During these experiments, the
following parameters will be studied:

— photochemical oxidation of organic contaminants;

— biodegradation potential of bacterioplankton with
respect to organic contaminants (benzo(a)pyrene.
PCB, etc.);

— combined influence of contaminants on biological
"targets" and establishment of "critical" concentrations
of the impact on plankton communities in the
conditions of controlled ecosystems (Izrael el a!..
1988a); and

— sediment respiration and nutrient flux experiments.

Connection with Other International Programs

The BERPAC Program has much in common with other
international programs, but at the same time it has its own
particular features mentioned earlier. Wide cooperation with
other similar international projects is built within the framework
of this program — in particular, in the preparation of joint
marine expeditions. Wide data exchange is also planned.

Schedule of Activities and Applications of Results

Since 1977, successful joint investigations of Soviet and
American scientists have been carried out in the Bering Sea
within the framework of the specific theme of the bilateral
cooperation "Bering Sea" (Project "Comprehensive
Environmental Analysis;" Subproject "Comprehensive
Analysis of Marine Ecosystem State and Ecological Problems
of the World Ocean"). Important stages of this cooperation
were three joint ecological Soviet-American expeditions in the
Bering Sea on the R/V Volna (Summer 1977). R/V Akademik
Korolev ( Summer 1 984 and 1 988 ). and several symposia on the
preparation of scientific programs and analyses of the results of
these expeditions, as well as three monographs describing the
results of long-term Soviet-American investigations in the
Bering Sea (Izrael & Tsyban, 1990; Roscigno. 1990). H is
expected that these expeditions will be every four years and
followed by international symposia and joint publications.

Monographs on the results of future expeditions will be
published. It is expected that seminars and symposia within the
framework of the BERPAC Program will be conducted. Also
included in the plans are special intercalibrations, a wide
exchange of specialists, and joint experimental work.

VI

References

Coachman, L. K. (1990). Bering Sea ecosystem: basic
characteristics and prospects for long-term research. In
Research on the Bering Sea Ecosystem. Results of the
Second Soviet-American Expedition. The 37th Cruise of
the Research Vessel Akademik Korolev. June-September
1984. Gidrometeoizdat Publishers, Leningrad. 2. (in Russian)

Izrael, Yu. A. & Tsyban, A. V. (eds. ) (1983a). Research on the
Bering Sea Ecosystem. Gidrometeoizdat Publishers.
Leningrad. 157 pp. (in Russian)

Izrael. Yu. A. & Tsyban. A. V. (1983b). On the assimilative
capacity of the World Ocean. Reports of the USSR Academy
of Sciences 272(3), 702-705. (in Russian)

Izrael. Yu. A. & Tsyban. A. V. (1985). The ecology and
problems of global ocean monitoring. In Comprehensive
Global Ocean Monitoring 1. 19-48. Gidrometeoizdat
Publishers, Leningrad, (in Russian)

Izrael, Yu. A. & Tsyban, A. V. (eds.) ( 1987). Comprehensive
Analysis of the Bering Sea Ecosystem. Gidrometeoizdat
Publishers, Leningrad, 264 pp. (in Russian)

Izrael, Yu. A. & Tsyban, A. V. (1989). Anthropogenic Ecology
of the Ocean. Gidrometeoizdat. Leningrad. 528 pp. (in
Russian)

Izrael, Yu. A. & Tsyban, A. V. (eds.) (1990). Research on the
Bering Sea Ecosystem. In Results of the Soviet-American
Expedition. The 37th Cruise of the Research Vessel Akademik
Korolev. June-September. 1984. Gidrometeoizdat
Publishers. Leningrad, 344 pp. (in Russian)

Izrael, Yu. A., Tsyban. A. V., Panov, G. V., Korsak, M. N.
et al. ( 1988a). Comprehensive analysis of the Bering Sea
ecosystem. In Comprehensive Analysis of the Environment.
Proceedings from the Fifth USSR-US Symposium.
Gidrometeoizdat Publishers. Leningrad, 528 pp. (in Russian)

Izrael, Yu. A.. Tsyban, A. V., Ventzel, M. V. & Shigaev, V. V.
( 1988b). Generalized model of the assimilative capacity of
a marine ecosystem. Reports of the USSR Academy of
Sciences 380(2). (in Russian)

Izrael. Yu. A.. Tsyban, A. V., Ventzel, M. V. & Shigaev. V. V.
(1988c). Scientific basis for ecological standardization of
the anthropogenic impact on marine ecosystems (using the
example of the Baltic Sea ecosystem. Oceanology 28(2).
(in Russian)

Roscigno, P. F. (ed. ) (1990). Results of the Second Joint US-
USSR Bering Sea Expedition. Summer 1984. US Fish and
Wildlife Service Biological Report 90( 13). 347 pp.

Zaitsev. Yu. P. (1970). Marine Neustonology. Naukova
Dumka Publishers. Kiev, 264 pp. (in Russian)

Zaitsev, Yu. P. ( 1980). Zooneuston and methods for its study.
In Methods for Biological Analysis of Sea Water and Bottom
Sediments, pp. 134-139. Gidrometeoizdat Publishers,
Leningrad,, (in Russian)

Zaitsev, Yu. P. ( 1985). Biotic contours in ocean monitoring. In
Comprehensive Global Ocean Monitoring. Proceedings
from the First International Symposium 2, 76-83.
Gidrometeoizdat Publishers. Leningrad, (in Russian I

John F. Turner
Director. US Fish and

Wildlife Service
Washington, DC

Yuriy A. Izrael
Chairman. USSR State
Committee for Hydrometeorology
Moscow, USSR

Harold J. O'Connor

Project Leader. USA

US Fish and Wildlife Service

Patuxent Wildlife Research Center

Laurel, Maryland

Alia V. Tsyban

Project Leader, USSR

Institute of Global Climate and Ecology &

State Committee for Hydrometeorology
Academy of Sciences
Moscow, USSR

Protocol of the First Joint US-USSR

Central Pacific Ocean Expedition

on the R/V Akademik Korolev

In accordance with the memorandum of the 1 1th meeting
of the US-USSR Joint Committee on the Environment
Protection (Moscow, USSR, February 1988) and the
recommendation of the "Soviet-American Conference on the
Ecology of the Bering Sea" (Batumi, USSR. March 1988) and
the plan of the joint bilateral activity of 02.07-2101.
"Comprehensive Analysis of Marine Ecosystems and Ecological
Problems of the World Ocean." the Third Joint US-USSR
Bering & Chukchi Seas Expedition was held on 26 July 1988
on board the Soviet research vessel Akademik Korolev. The
second leg of this expedition was conducted from 9 September
to 3 1 October in the central Pacific Ocean and South China Sea.
The delegation was headed by Prof. Alia V. Tsyban and
Dr. Gregory J. Smith. Nearly 10,000 nautical miles of ocean
were covered during this leg of the expedition.

The Soviet delegates were represented by participants in
the cruise from the USSR State Committee for
Hydrometeorology and Control of Natural Environment; the
Academy of Sciences from the USSR; and the Academies of
Sciences from Ukraine, Belyorussia, and Estonia. A list of
participants is given as Appendix A.

Six American delegates joined the expedition during the
port of call in Hilo, Hawaii, on 9 September 1988 and were
represented by participants from the US Fish and Wildlife
Service (Department of the Interior), and the University of
Washington. A list of participants is given as Appendix A.

The principal objective of the second leg of the Third Joint
US-USSR Expedition was to conduct comprehensive studies
of the ecology of the central Pacific Ocean with an emphasis on
ecosystem processes and the effects of anthropogenic pollutants
on those systems. Included in this research was a complete
zoological, chemical, and microbiological assessment of an
isolated tropical coral atoll. Caroline Atoll, Kiribati.
Hydrological stations were included along the route from
Hawaii to Christinas Island, where a port of call was made to
bring aboard a wildlife biologist from the government of
Kiribati for studies of Caroline Atoll. During the period of
22-29 September, extensive studies of this remote atoll's
ecology were conducted and an assessment of the atoll's status
with respect to environmental contamination was made. A
complete survey of the atoll's plant and bird communities was
also done. A full ecological station was sampled offshore from
the atoll and series of four more stations were included enroute
to Ratawa where the Kiribati representative disembarked.
After Tarawa, the research vessel proceeded to I IN latitude.
Along (he 1 I transect, six ecological stations were studied
in the Marianas section. The first station was m more than
6.00(1 meters of water, the last station along this transect was
near the eastern Philippines, and the Akademik Korolev

proceeded through the Straits of Mindanao into the Mindanao.
Sulu. and South China Seas. This route avoided Typhoon
Ruby, which was located in the north Philippines and caused a
severe loss of life and property in that area. On 25 October
1 988, a three-day anchor station began in the South China Sea
at 6°N latitude, three miles west of the 107th parallel. This
long-term ecological station was the first in a series of five as
the expedition proceeded into Singapore. Arrival at Singapore
was on 31 October 1988, completing the second leg of the joint
American-Soviet expedition of 1988. The route of the
expedition is shown in the Frontispiece.

The main scientific tasks:

/. Biological, chemical, and physical fundamental data
were collected to provide a comprehensive ecological and
oceanographic profile of the central Pacific Ocean and South
China Sea. These data provided a basis for comparison with the
Bering and Chukchi Seas.

2. Cooperative studies of the methodology of collection,
quantification, and chemical analysis of plastic debris were
conducted to assess the overall hazard of plastics to marine life.

3. The ecological health of a remote coral atoll in the
central Pacific was determined. Complete studies of the bird
and plant communities were done. Studies of chlorinated
hydrocarbons, their concentrations, distribution, and
degradation by photolytic and microbiological processes were
determined, thereby providing an assessment of the effects of
anthropogenic activities.

In accordance with specialties of the expedition's
participants, working groups were organized. At these meetings,
work schedules, joint studies, and model experiments were
planned. During the expedition, meetings of the Scientific
Council Board were also held. Discussed at these meetings
were the important scientific findings of the expedition,
comparing the Pacific Ocean and the Bering and Chukchi Seas.
Also discussed were new areas of investigation that were
introduced into the research effort with the new delegation of
American scientists on the second leg. These areas
included studies of marine plastic pollution, bird ecology, and
the collection of biota for radionuclide analysis.
Cooperative studies of the degradation and photolytic products
of benzo(a)pyrene were expanding during this leg of the
expedition.

Complex ecological studies were undertaken for the first
time in the equatorial part of the Pacific Ocean. Unique joint
studies thai w ere undertaken to study the ecosystem Of Caroline
Atoll, an extremely remote area removed from the main sources
of anthropogenic influences, are especially interesting. During
the course of the cruise, results of preliminary investigations
were obtained.

VIII

The concentration of dissolved oxygen in the lagoon
waters of Caroline Atoll reached 1 15 to 126%, and the vertical
distribution of hydrochemical parameters were relatively
homogeneous. Salinity in lagoon waters was slightly higher
than in water sampled outside of the lagoon (ocean water); the
average values for lagoon and adjacent ocean water were
36.05 and 35.85 parts per thousand, respectively. Nutrient
concentrations were lower in lagoon water than in coastal
water: silicon and phosphorus were two times lower, nitrates
were one and one-half times lower. According to studies of
hydrooptical characteristics the lagoon waters do not differ
significantly from the ocean waters.

The mean rate of biosedimentation of particulate organic
matter (POM) in the water column from 0 to 100 m in the area
investigated in the tropical zone of the Pacific Ocean was found
to be 14.2 to 72.5 mg of dry matter/m'/day. In coastal waters
of the atoll this rate was 64.5 mg dry matter/mVday. The rate
of biosedimentation in the surface water was two times higher
than the rate at a depth of 100 m.

The most probable number (MPN) of heterotrophic
saprophytic bacteria in the water inside and outside of the
Caroline Atoll lagoon was 1,000 to 10,000 cells/ml. The MPN
ofparaffin-oxidizing,benzo(a)pyrene-transforming,andPCB-
transforming bacteria in the water near the atoll was 10 to
1,000 cells/ml. In the water inside of the lagoon the MPN of
microorganisms of these groups was less: 3 to 10 cells/ml. This
information gives us the ability to characterize the atoll's water
as clean, without anthropogenic pollutants.

The dominant pollutant in the atoll's ecosystem, in the
widely investigated range of chlorinated hydrocarbons, was
DDT and its metabolites (approaching 1 ng/1). Other substances
observed in the pollutant benzo(a)pyrene were more intensive
at Caroline Atoll than in the tropical zone of the ocean. In one
hour, the amount of benzo(a)pyrene destroyed approached
85%, and after three hours more than 95% of experimentally
added benzo(a)pyrene was destroyed. Ecotoxicity experiments
conducted in situ dealing with the influence of benzo(a)pyrene,
PCB's, copper, and cadmium on planktonic communities of
the atoll's waters showed higher vulnerability of planktonic
organisms to toxic metals and were relatively more resistant to
benzo(a)pyrene and PCB's in comparison with the northern
seas.

The distinctive feature of the zooneuston of the Caroline
lagoon in comparison with the adjoining oceanic waters is the
presence of a neuritic complex of organisms (larval stages of
benthic animals and the early stages of Copepoda). The mean
number of organisms was 64 ind/m\ 26 less than in the coastal
waters of the atoll. Nevertheless, the zooneuston of the lagoon
play an important role in the formation of the coastal water
fauna.

A unique Acropora-Tridacna reef, dividing the lagoon,
was discovered in the southern part of the atoll. Its length was
more than one kilometer and its width was 15 to 20 m. At
certain places along the reef, the Tridacna formed a dense
aggregation with numbers approaching 40 ind/m:.

More than 42,000 m2 of Pacific Ocean surface water were
sampled for plastic debris and macroscopic spherules. Plastics
were recovered at six of the 29 different stations sampled.

At two of these stations, tar balls were also recovered. The
density of surface plastics in areas that had positive samples
ranged from 0.00782 to 0.19481 mg plastic/nr. Because
marine growth on the surface of plastics may alter the specific
gravity of floating debris, sampling was also done at the
thermocline to determine if plastic could be contaminating this
important subsurface stratum. None of the 5 1 ,600 m3 of water
sampled at the thermocline at nine different stations contained
any plastics or anthropogenic materials. The surface water of
the Caroline Atoll lagoon was sampled at two locations using
a 102 neuston net. More than 21,500 nr of water sampled
showed no evidence of plastics or other debris. Extensive
sampling of surface and subsurface water was also done in the
South China Sea. Although a wide variety of anthropogenic
materials were recovered in surface tows, few plastic cylinders
(raw material) were found. One sample from the thermocline
contained plastic line, the most common form of plastic
recovered in this area. Another surface sample from the South
China Sea contained more than 1 64 tar balls greater than 4 mm
in diameter.

Detailed data on the plants, seabirds, landbirds, mammals,
reptiles, and human disturbance were obtained from 39 islets at
Caroline Atoll. The flora consists of 19 species (one new to the
island) organized into 4 natural and 4 anthropogenic plant
communities; 92% of the islets are pristine. Eleven species of
seabirds breed (red-tailed tropicbird is a new record): the sooty
tern, with 189,000 breeding pairs, was the most abundant
species. The long-tailed cuckoo was recorded for the first time
in the Line Islands.

At-sea observations of marine birds indicated high
densities near uninhabited islands and in known areas of high
productivity. Low densities were observed in areas of
low ocean productivity and in areas with high human
disturbance (Gilbert Islands, and the Bohol, Sulu, and South
China Seas).

Beach surveys for anthropogenic debris were done on nine
islands of Caroline Atoll. Plastic and styrofoam objects
accounted for 75 to 80% of the total number of items observed;
however, glass bottles and fishing gear were significantly more
important with respect to volume and biomass of debris.
Observations were also made at sea to determine the amount of
floating debris.

Samples were collected at Caroline Atoll and selected
ocean sampling stations for radiological analyses. These
analyses will compare natural versus anthropogenic
radioisotopes. Radioisotope concentrations at Caroline Atoll
will be compared to similar samples obtained in the Marshall
Islands that contain contamination from atmospheric nuclear
weapons testing 30 years ago.

At the end of the joint expedition on board the Akademik
Korolev, there was an exchange of preliminary data. Future
exchanges of data and results of analyses will occur in a series
of three exchanges: 1. 1 March 1989; 2. 1 June 1989; and
3. 1 October 1989.

Both sides note with satisfaction the friendly and
constructive atmosphere of the expedition's work and the
effectiveness of joint observations allowing for a variety of
oceanographic and ecological studies.

The American delegation w ishes to extend their sincerest
gratitude to the Soviet participants of the expedition.
The science staff and ship's crew provided an atmosphere
conducive not only to research but also the main friendships
that resulted from our meeting. The American delegation
would especially like to thank Professor A. V. Tsyban and

Captain O. A. Rostov tse\ for the wealth of scientific and
maritime knowledge provided the foundation for this highly
successful scientific endeavor.

This protocol was written in English and Russian and was
signed on board the research vessel Akademik Korolev,
1 November 1988. Both texts are equally authentic.

For the American side:

Leader of the American
Scientific Delegation.
Patuxent Wildlife Research Center.
US Fish and Wildlife Service.
US Department of the Interior

For the Soviet side:
Head of Expedition,

Leader of Project for

the USSR Side.
Deputy Director ol Laboratory lor

Environmental and Climate Monitoring

Laboratory .
Goskomgidromet and Academv of Sciences

Dr. (1. J. Smith Professor A. V. Tsyban

(This text is a reproduction of the protocol vv ritten on board the R/V Akademik Korolev m I C)S8. The original w as signed bv both project leader.*

Acknowledgments

We gratefully acknowledge and thank the many individuals without whose participation this monograph
may not have been published with the same quality, accuracy, and clarity.

We thank the US Fish and Wildlife Service and the USSR State Committee for Hydrometeorology

for their continued support.

Steven Kohl and Stephanie Miller, the Coordinator and Associate Coordinator of US-USSR Programs,
US Fish and Wildlife Service (Office of International Affairs), have provided invaluable assistance
throughout every phase of this project. Their enthusiasm and energy given to this project, and the people
involved with this project, are outstanding.

Without each participant of the expedition, and each author of research results, there would be no need
for a monograph. There are far too many to name here; however, their names are listed with each subchapter
and in Appendix A in this volume. It is their interest and excitement for the research presented here, and their
spirit of cooperation so necessary for an international project, that provide the essence of the scientific
accomplishments.

We are indebted to each of the US and USSR chapter editors for their help and their patience with the
seemingly endless questions and tasks assigned to them, and last but certainly not least, for their sense of
humor which is often the only saving grace inputting together a volume of this magnitude. Their names are
listed alphabetically below:

Sergei M. Chernyak Gregory J. Smith

Cameron B. Kepler Alia V. Tsyban

Mikhael N. Korsak Terry E. Whitledge

Clifford P. Rice

The "Production Team" at Patuxent Wildlife Research Center — Kinard Boone, Patricia A. Holt,
Susan A. Liga, Robert E. Munro, Patricia A. Nagel, and John C. Sauer — deserves recognition for their
dedication to meeting the challenge of producing a quality volume in time for it to be distributed to
participants on the 1992 Expedition.

Harold J. O'Connor Alia V. Tsyban

XI

Table of Contents

Page

Foreword iii

Protocol of the First Joint US-USSR Central Pacific Ocean Expedition on the R/V Akademik Korolev viii

Acknowledgments xi

Frontispiece xiv

Chapter 1 : ECOLOGICAL INVESTIGATION OF A CORAL ATOLL IN THE CENTRAL PACIFIC .. 1

Chapter 1 Frontispiece 3

1.1 Ecological Studies of Caroline Atoll. Republic of Kiribati. South-central Pacific Ocean :

Part 1. History, Physiography, Botany, and Islet Descriptions 5

Introduction 5

Geography 6

History of Caroline Atoll 6

Pre-European History: Tuamotuan Period 6

Post-European History: 17th to 19th Centuries 6

The Late 19th and 20th Centuries 8

Methods 9

Naming Caroline's Motus 9

Structure and Topography 10

General Account 10

Reef Flats 10

Beaches 11

Lagoon 12

Substrata 13

Hydrology 13

Climate 14

Vegetation: Vascular Plants and Floristics 15

Botanical History 15

Vascular Plants of Caroline Atoll 15

Floristics and Ecology of the Motus 21

Ecological Succession 23

Basic Serai Stages 23

Ecological Succession on Motus of Different Size Classes 23

Species-Area Relationships 25

Plant Communities 26

General Account 26

Natural Herb Mat 26

Beach Scrub with Suriana 27

Pandanus Forest 27

Tournefortia Scrub and Forest 27

Cordia Forest 29

Pisonia Forest 29

Coconut Woodlands 32

Absent Plant Communities 34

Description and Ecology of the Motus 34

Nake Island 35

Long Island 36

Windward Islets 37

South Nake Islets 42

Central Leeward Islets 44

Southern Leeward Islets 47

Conclusion 44

fables 51

Figures 63

Plates 93

Appendix 1 135

Appendix 2 136

1.2 Ecological Studies on Caroline Atoll, Republic of Kiribati, South-central Pacific Ocean:

Part 2. Seabirds, Other Terresterial Anmials and Conservation 139

Introduction 139

History of Ornithological Studies 139

Methods 140

Seabird Species Accounts 14!

Other Birds on Caroline Atoll 147

Other Vertebrates 149

Coconut Crabs 150

Conservation: Attributes of International Significance 151

Tables 153

Figures 157

Plates 163

1 .3 First Records of the Long-tailed Cuckoo (Eudynamis taitensis) on Caroline Atoll, Southern

Line Islands, Republic of Kiribati 165

1.4 A Study of the Benthic Communities of Caroline Atoll (Line Islands, Pacific Ocean) 166

Chapter 1 References 171

Chapter 2: INVESTIGATIONS AND ANTHROPOGENIC ECOLOGY 177

2. 1 Distribution of Chlorinated Hydrocarbons in Ecosystems of the Equatorial Pacific 1 79

2.2 Distribution of Polycyclic Aromatic Hydrocarbons 183

2.3 The Occurrence and Microbial Transformation of Benzo(a)pyrene in the Waters of the

Tropical Pacific (Caroline Atoll, Line Islands, Phoenix Islands Transect. South China Sea) 186

2.4 Cesium-137 in the Surface Waters of the Central Equatorial Pacific 191

2.5 Quantity and Distribution of Plastics: An Analysis of Chemical Hazards to Marine Life 193

2.6 The Role of Solar Irradiation in the Oxidative Transformation of Benzo(a)pyrene 197

Chapter 2 References 203

Chapter 3: BIOLOGICAL INVESTIGATIONS IN THE CENTRAL PACIFIC 205

3.1 A Description of Bacterioplankton 207

3.2 A Study of Primary Phytoplankton Production 212

3.3 Mesozooplankton 213

3.4 Zooneuston of the Tropical Pacific 223

3.5 Observations of Seabirds along a 14.892-km Cruise Track in the Tropical Pacific Ocean and

the Bohol, Sulu, and South China Seas 225

3.6 Evidence for a Major Fall Land Bird Migration Corridor Across the South China Sea from
Indo-China to the Greater Sunda Islands 247

Chapter 3 References 253

General Conclusions 257

Appendix A 259

o

m -

o

CD

O

o

CD '

O

o

CD .

o

in .

ca
ca

■x.

-^-0^^

y.

C

O

—

'->

O

_1

t ,

C

'-)

_-

c

(D

ca

KJ

t/3

f-

_)

C

p

®

c

)

z

<

u

o

u

u
<

C

.c
U

ca

ca
ca
H.

©

®

\>

0

o

®

.

^

<?'

®

/

■8

O
c/5

3C
30

<

0.

04

UJ

aa

■a
u
n.
x

UJ

act

D

l

t/3

D

o
■a-

o

CD

O
CM

O
CM

\l\

Chapter 1: I

ECOLOGICAL INVESTIGATION

OF A CORAL ATOLL

IN THE CENTRAL PACIFIC

Editor:

CAMERON B. KEPLER

CAROLINE ATOLL

1000
I L

2000

METERS

Chapter 1 Frontispiece. Air-mosaic of Caroline
Atoll, RNZAF 6569. Reproduced by permission of
the Lands and Survey Department, New Zealand.

1.1 Ecological Studies of Caroline Atoll,

Republic of Kiribati, South-central Pacific
Ocean

Part 1. History, Physiography, Botany, and Islet Descriptions

ANGELA K. KEPLER , CAMERON B. KEPLER1, and DAVID H. ELLIS*

*t/5 Fish & Wildlife Senice, Patuxent Wildlife Researeh Center, Southeast Research Station, Athens, Georgia, USA
XUS Fish & Wildlife Senice, Patuxent Wildlife Research Center, Laurel, Maryland, USA

Introduction

On 26 July 1988, the Soviet research vessel Akademik
Korolev sailed from Vladivostok enroute to Dutch Harbor,
Alaska. There, Soviet oceanographers joined their American
colleagues to investigate the Gulf of Alaska and the Chukchi
Sea in the Third Joint US-USSR Eering & Chukchi Seas
Expedition. When the arctic research was completed in early
September, the ship headed toward the central Pacific. A
rendezvous for a second contingent of Americans took place in
Hilo, Hawaii, on 9 September. Six Americans joined the ship,
which set sail on a cruise track of 14,892 km that terminated

6 weeks later in Singapore. An important part of the expedition
was research in and around little-known Caroline Atoll, at the
southeastern edge of the Line Group. On Christmas Island, we
picked up Katino Teeb'aki, a conservation officer for the
Republic of Kiribati, who represented his government and
helped our land-based research efforts. After landing on
Caroline on 22 September, we camped in 2 locations for

7 nights, surveying the terrestrial plants and animals on all
39 islets. Caroline is a remarkably pristine atoll with its native
plant communities nearly intact on all but three islets, and
teeming seabird communities that, collectively, are second in
the Line Group only to Christmas Island ( Kiribati ) in diversity.
For several historical reasons, the natural values of this
spectacular blend of marine and terrestrial resources have been
overlooked.

Caroline Atoll (Chapter Frontispiece) is situated at 10°00'S
latitude and I50°13'W longitude in the south-central Pacific
Ocean. Caroline is the southeasternmost of the Southern Line
Islands, a group of three islands that also includes Vostok and
Flint, lying 230 km to its west and southwest, respectively.
Although anthropologically and geographically within
Polynesia, Caroline is owned by the Republic of Kiribati
(formerly Gilbert and Ellice Islands).

Caroline, 9.7 km long, 2.3 km wide at its widest point, and
26.9 km in circumference, is a crescentic coral ring with
39 islets centered on a continuous reef enclosing a relatively
shallow lagoon. Its total land area above high water is 399 ha,
with islets ranging in size from 0.02 to 107.5 ha. Islets
extend along 55% of the reef perimeter. The closed lagoon,
rich in marine life, contains a maze of patch reefs and remarkably
clear water.

The island was "discovered" by de Quiros in 1606.
Although remnants of an ancient Tuamotuan culture still exist,
the atoll apparently never supported a long-term permanent
population and has been less affected by man than most Pacific
islands. Its European history includes guano export, a
multinational expedition to observe a solar eclipse, and copra
production. It has been uninhabited since the early 1930's (a
factor contributing to its relatively undisturbed ecology), except
for the presence of one family from 1987 to 1991. The primary
factors responsible for its lack of permanent settlement are
remoteness, apparent absence of usable ground water, the
repeated failure of its coconut plantations, absence of a passage
into the lagoon, a paucity of safe boat anchorages, and an
abundance of rats.

Until the 1988 US-USSR Expedition on the Akademik
Korolev, only one fairly accurate map was available, which
named seven islets. We have drafted an accurate map based on
this field work and recent aerial photographs, also naming 32
previously unnamed islets, 4 islet groups, and an inlet. During
8 days of intensive field work, we surveyed all 39 islets,
walking 33 km in systematic cross-islet transects and around
islet perimeters.

Soils, of coral and molluscan origin, are categorized into
five types, from barren coral rubble to rubble mixed with
humus and guano. Caroline provides an excellent example of
soil development through different age and size classes of
motus.

Caroline's near-pristine, lush native vegetation supports
27 species of plants organized into 7 plant communities —
6 natural and 1 anthropogenic. The atollwide distribution of
each plant species is mapped. Its insular flora, typical of central
equatorial islands in their natural state, is 85% indigenous
(possibly up to 93% ), an extremely high figure for anywhere in
the world. Though Caroline's islets are covered with
extensive tracts of native woodland, the Pisonia grandis
forests, up to 25 m high, are particularly notable as they
constitute some of the best groves left in the Pacific. Toumefortia
argentea is abundant, and Cordia subcordata, becoming
quite rare elsewhere, is present on most islets. Cocos is
present, but only dominates one islet; 22 islets harbor wholly
indigenous vegetation.

Islets of many different age and size classes provide
excellent examples of soil and vegetation development,
accompanied by an increasing diversity of bird life.
On account of its relatively low human disturbance and rapid
recovery to a more natural state, especially over the last
70 years, Caroline is one of the few Pacific islands that is truly
an "outdoor ecological laboratory." Many of its disturbed
islets have recovered so remarkably they are almost
indistinguishable from those that have remained pristine.

An analysis of ecological succession on motus of increasing
size reveals that by the time a motu reaches 0.8 ha in size, all
the natural plant communities, most plant species, and most
species of seabirds are present. This is in striking contrast to
species-area relationships on inhabited atolls with more
introduced plant species.

Foreach motu, the main physiographic features, vegetation
patterns, seabird colonies, miscellaneous biota (coconut crabs
and rats), and the effects of human activity (if any), are
described in detail. Included are vegetation maps for each
motu, and tables and figures relating to species-area
relationships.

Permanent protection of Caroline is currently being sought
by The Nature Conservancy of Hawaii as it negotiates with the
government of Kiribati for a Southern Line Islands Wildlife
Preserve, which includes Caroline, Vostok, and Flint.

Geography

Caroline Atoll* (Chapter Frontispiece; Figs. 1,2) is a
small, low coral atoll situated at LOWS latitude, 150°13'W
longitude in the south-central Pacific Ocean. It lies 2.800 km
south of Hawaii. 830 km north of Tahiti, and 1 .030 km west of
the Marquesas Islands. Its nearest neighbors are Vostok and
Flint, 230 km to the west and southwest, respectively.

A recent geographic survey of Caroline by the ICBP 1990
Line and Phoenix Islands Expedition, using a compact satellite
navigation computer, indicates that the atoll lies one nautical
mile east of its previously charted longitude position, 1 50° 1 4" W.
Its range of coordinates are 09°54' to 10°01'S latitude and
150°12'to 1 50° 14'W longitude. The actual coordinates given.
10°00'S and 150°13'W, intersect in the lower lagoon just west
of the "blind passage."

Caroline is the southeasternmost of the Line Group
(Fig. 1 ), a widely scattered group of five atolls, five islands, and
two submerged reefs lying in the south-central Pacific Ocean
between 06°N and 12°S latitude and 162 and 1 50°W longitude.
Scattered across 250.000 km2 of ocean, the Line Group falls
naturally into two parts: the Northern Line Islands, four atolls
and one reef north of the equator, and the Southern Line
Islands, five islands, one atoll, and one reef to its south. The
name, Line Group, reflects its equator-straddling location.

■Note: Caroline Atoll is neither physically, geographically,
nor politically associated with the Caroline Islands, now part of
the Federated States of Micronesia, more than 6.000 km to the
northwest. Because of this confusion, we use the name "Caroline
Atoll" instead of "Caroline Island."

Although archaeologically and geographically within
Polynesia, the Line Group was uninhabited when discovered
by Europeans; its islands have been variously claimed by the
United States and England. With the exception of US-owned
Jarvis, Palmyra, and Kingman Reef, all are now governed by
the Republic of Kiribati (formerly Gilbert Islands).

Caroline. 9.7 km long by 2.3 km wide at its widest point,
is a crescentic coral ring 26.9 km in circumference. It is
composed of 39 islets (Fig. 2) and a few incipient islets,
centered on a continuous reef flat, submerged at high tide, that
encloses a relatively shallow lagoon. The total land area above
high water is 399 ha. Entirely of coralline origin, its geology,
soils, climate, and vegetation are typical of low-latitude atolls.
It is relatively unmodified by man.

History of Caroline Atoll

Pre-European History: Tuamotuan Period

Centuries before Europeans encountered Caroline, this
lonely atoll was inhabited by Polynesians. No oral traditions of
this occupation are known, but evidence of former habitation
was evident when de Quiros found the atoll in 1606. He noted
"an old canoe, lying on her side." and a small grove of coconuts
planted on South Island (Bennett. 1840; Markham. 1904).

No further clues to the atoll" s early history were unearthed
until Messrs. Brown, Brothers, and Arundel exposed about 50
ancient Polynesian sites in the 1870's while digging for guano
(Holden, 1884; Arundel, 1890). Although the largest two were
marked as "graves" on Arundel's 1883 map (Fig. 4). no bones,
ashes, or human remains were found. Natives living on
Caroline called them "marai" (marae). Arundel photographed
and drew plans of them (Fig. 3): depicted are a platform of coral
and conglomerate rock, surrounded by 10 smaller slabs
resembling gravestones, all arranged in a rectangular plan.
Their findings were later identified as Tuamotuan manic
( Emory. 1 947). Manic, according to ancient belief, "bound the
ancestral spirits and gods of the kindred to the land, putting it
under their eternal guardianship" (Emory. 1947). The largest
manic was on northwest Nake Island, and a smaller one was
found near the southern tip of Long Island. Both locations
conform to such prerequisites for building manic as nearby
shorelines and birds ( see Description and Ecology of the Motus
section [Nake Island]), which Tuamotuans believed housed
divine spirits (Emory, 1947. p. 123). Although members of the
ICBP 1990 Line and Phoenix Islands Expedition located them
and took photographs and measurements of the Nake site
(PI. 36). no field work by archaeologists has been conducted.

Post-European History: 17th to 19th Centuries

On 21 February 1606. the Portuguese explorer Pedro
Fernandez de Quiros. employed by Spain, "discovered" Caroline
Atoll (Markham. 1904; Stevens & Barwick. 1930). naming it
San Bernardo. Despite its remote location. Caroline was
encountered early in Pacific history, long before Tahiti.
Rarotonga, and Hawaii. This is possibly due to its location, for
early navigators tended to sail due west from South America
along lines of latitude, and 10 S was an obvious choice. De
Quiros. the last ad\ enturer in the Spanish age of discovery, was

leading his second major trans-Pacific expedition with 3 ships

and 1 50 men obsessed with finding the fabled "Terra Australis

Incognito." The descriptions of Caroline by his crew, although

atvariance with one another.still apply today(Pl. 1). Theirfirst

at-sea impression was that it was "divided into four or five

hummocks, and all the rest submerged. Its circumference

appeared to be ten leagues" (Markham, 1904). After landing,

they found that

There was a great number offish inshore, and, owing to the

water being very shallow, they were killed with swords

and poles. There were great numbers of lobster and

crawfish, and other kinds of marine animals. They found

a great quantity of cocoa-nuts in a heap at the foot of the

palm trees, many large, and of different sizes. There were

a great quantity of sea birds of several kinds, and so

importunate that they seemed to want to attack the men.

We took plenty of all these things. ..It seemed to the

Captain that on an island where there are so many trees

there could not fail to be water. (Markham, 1904)

Fresh water was crucial to de Quiros and his crew, who

were suffering from lack of food and water. Despite their

efforts, however, they failed to obtain water. Disappointed and

lacking energy, they continued their voyage the following

morning. Their demoralized state may explain one statement

that Caroline "consisted of twenty-two islets, uninhabited and

without water, trees or scrub for wood."

In 1795, Captain W. R. Broughton, on the British sloop
Providence, rediscovered the atoll while voyaging from Tahiti
to Hawaii. He named it in honor of the daughter of the First
Lord of the British Admiralty (Broughton, 1804):
The southern extremity was the highest part, covered with
trees, most probably cocoa-nut from their appearance, as
they stood in detached clumps along the shore. The
island.. .appeared to be low, and covered with trees, and if
I am right in its estimated distance, its length will be about
five miles in a north and south direction. I named it
Carolina Island in compliment to the daughter of Sir P.
Stephens of the Admiralty.
Because early navigation techniques and communication
were far less sophisticated than today, especially with regards
to longitude, Caroline was sighted or "discovered" by several
more explorers who were unsure of its identity. By 1821 the
atoll had amassed an impressive collection of coordinates and
names: San Bernardo, Island of Fish, Thornton, Hurst's,
Clark's, Independence, and Carolina (which later became
Caroline). Some navigators equated Caroline with an island
named "San Bernardo" by the Spanish explorer Mendana in
1595. Mendana's island has only recently been verified as
Puka-Puka in the northern Cook Islands (Maude, 1968).

The best early descriptive account of the atoll's flora and
fauna comes from an 1835 visit by F. D. Bennett, who was
reasonably well versed in natural history (Bennett, 1840). He
noted that the islets then, as now, were "covered with
verdure. ..surprisingly luxuriant, when compared to the arid
soil it covers." Although Bennett had visited many atolls, he
was particularly impressed with the quality of Caroline's coral
reefs. His party observed "rats of a red-brown color" and

various birds but no reptiles (Subchapter 1.2). Although he
discusses "land lobsters (Coenobita species)," no mention is
made of coconut ciabs (Birgus latro).

First Occupation: The existence of two small coconut
groves on Caroline prompted two British entrepreneurs,
representing the Tahitian firm Collie & Lucett, to establish a
stock raising venture there in 1846. This first known settlement
was located adjacent to the main coconut grove on the northwest
peninsula of South Island; a smaller grove evidently existed
"on the south-south-west side" of the same island (Lucett,
1851 ). Tahitian laborers tended pigs, hens, turkeys, and grew
many food plants, including pumpkins and melons. They dried
and salted fish, planted coconuts, and extracted coconut oil
(Maude, 1942a;Garnett, 1983 (and were evidently still therein
May 1852 (Ellsworth, 1990).

Political Annexation: Though inhabited in prehistory by
Tuamotuans, officially "discovered" by the Spanish, and visited
by British, French, and American ships, it took centuries for
Caroline to acquire a political identity. It was formally annexed
to Britain by Captain Nares, R. N., who arrived in the H.M.S.
Reindeer in 1868, finding 27 residents.

Caroline was under the control of various merchants in the
late 19th century: Lionel Brown, Captain Brothers, and later
John Arundel, a well-known businessman, trader, and guano
merchant in the Pacific. Arundel's 1883 map (Fig. 4) of
Caroline is the only reasonably correct chart published until
this paper.

The Guano Era: Though bonded under the American
Guano Act in 1860, no phosphate was dug on Caroline until
Arundel was granted a 7-year license in 1874. A few months
earlier, a set of moorings was laid off the lee side of South
Island, allowing ships of up to 1 ,000 tons to lie safely during
trade wind weather. Guano was the only successful business
venture at Caroline: approximately 10,000 tons were shipped
to California and Australia between 1873 and 1895, when
supplies became exhausted (Young, ca. 1922).

Solar Eclipse Expedition: In 1883, Caroline received
international publicity when astronomers calculated that it lay
directly under the path of a pending solar eclipse. As a result,
three parties of astronomers (American, British, and French)
set up camp on South Island, making detailed observations of
this celestial event (PI. 2a). At that time Caroline was more
famous, and housed more people, than before or since: 7
"natives," scientists, and crewmen totaled 5 1 occupants.
Legacies from former inhabitants included three houses
(PI. 2b), two sheds, three huts on smaller motus, nautical
flotsam and jetsam, and two shallow wells. To this they added
tents, observatory frames, a marble slab, flagpole, and brick
"piers" for their telescopes, most of which remained as
technological litter.

This expedition (Dixon, 1884; Holden, 1884; Holden &
Qualtrough, 1884; Trelease, 1884; Young, 1884) also marked
the first attempt to describe the topography, climate, flora, and
fauna of the atoll. Drawings included an artist's rendering of
Caroline and map of the "settlement" (PI. 3) and views along
South Island's lagoon shore (Pis. 4, 5). Another map (Fig. 5)
was drafted but is highly inaccurate.

Observations by the astronomers on vegetation, birds,
insects, reptiles, marine organisms, et cetera were sketch)
( Butler &Strecker, 1884; Dixon, 1884). Dixon, the zoologist,
listed such organisms as •'shrimp." "'hermit crabs," "gnat,"
et cetera. As with Bennett, there was no mention of coconut
crabs, even though they were evidently abundant on South
Island in 1910 (Young, ca. 1922).

The Late lVth and 20th Centuries

In 1875. C. D. Voy. a naturalist from California visited
Caroline, collecting mollusks (Pilsbry & Vanatta, 1905a.b)
and fish (Fowler, 1901).

As early as 1885, Arundel began to clear land and plant
coconuts, but his planned copra industry was unsuccessful. In
I S97 he sold his business to the Pacific Islands Company. Ltd.,
which also failed. The coconut plantations suffered from
disease and poor vitality, and populations of Polynesian rats
apparently exploded. By 1904, when the H.M.S. Icarus visited
Caroline, only six Polynesians lived there. A few months later
they were repatriated to Niue, and Caroline remained
uninhabited until 1916. when a new effort was made to develop
the coconut plantation by Messrs. S. R. Maxwell & Co., Ltd.
During the uninhabited years. South Island"s vegetation
and wildlife began to recover from the earlier forest felling
( Pis. 2-4). When Mr. J. L. Young, then managing director for
S. R. Maxwell & Co., Ltd. ( Young, ca. 1 922 ), visited the atoll
in July 1910, he described it as a wilderness, teeming with sooty
terns, fish and coconut crabs:

The ground was covered with nests of seabirds which

latter rose like a cloud when disturbed: the noise of their

shrieking was so great that one had to shout to enable

oneself to be heard by his companions. Hundreds of great

Coconut Crabs were seen: 40 large ones were caught by

the crew of the schooner in an hour. The reef and the

lagoon swarmed with fish and small sharks.

From 1 9 1 6 to 1 929. Caroline was altered more than before

ui since. All the available land on .South was deforested to

make room for thousands of palms, and laborers demolished

huge numbers of coconut crabs and seabirds. which were

thought to damage the palms (Young, ca. 1922). In addition.

coconuts were planted on all of the main windward islets.

southern Nake. and on Mannikiba. Plantation workers in great

part lived oil the land, feasting on fresh fish, seabirds. seabird

eggs, turtles, and coconut crabs.

Copra exports averaged around 1 4 tons per year from 1929
to 1 934. after which the company ran into debt. Concurrently,
the French government forbade further recruitment from Tahiti;
b> l936onl> a lew families were left (N.I.D., 1943). In 1941
(he atoll earned a price tag of 000 English pounds (Maude,
personal communication), but was never purchased.

Occupation leases for Caroline were canceled in 1943.
after which the British Western Pacific High Commission
repossessed it (Maude, 1953). However.nev. "queen's leases"
were granted to M. P. A. Bainbridge of Papeete. Tahiti
( 1 95 I - 1 964 ) ( Nicholson & Douglas. 1 969), (hen ( 'aplam ( )mer
Darr ( 1964-1989) of Moorea. French Polynesia. When the

British granted independence to the Gilbert and Ellice Islands
in 1979, a new country, the Republic of Kiribati (pronounced
"Keer-ee-bahss") assumed ownership of Caroline, along with
most of the Line and Phoenix Islands.

Apart from occasional parties of Tahitians cutting copra
and a shipw recked sailor in the early 1 880's. the atoll remained
uninhabited for over 50 vears. During this time. Caroline's
vegetation and w ildlife recovered to such an extent that, were
it not for unpublished manuscripts from Maude (Maude,
ca. 1938.ca. 1 942a. and no date (.including Young' s(ca. 1922)
"Memoranda re Tahitian Business"our detailed vegetation
analysis (including tree diameters), and the 1 990 comparisons
with Flint and Vostok. we would have been unaware of the
actual extent of previous human interference or of the rapidity
of forest recovery (the fact that 60<7r of Caroline's motus
harbored wholly indigenous vegetation seemed to point to a
relatively pristine atoll).

In 1987. the Office de la Recherche Scientifique et
Technique Outre-Mer ( ORSTOM. a French scientific research
agency) was requested by the Kiribati government to conduct
a short study at Caroline on the feasibility of pearl-shell culture
(G. Monet, personal communication). Their results concluded
that the atoll would be inappropriate for this type of development.

Also in 1987, a Scotsman. Ron Falconer, his French wife
Anne, and two small children settled on Caroline as volunteer
custodians. From October 1989 to November 1990 a new lease
was under negotiation by Felix Urima. a French businessman,
who planned to blast a channel through the reef, construct an
airstrip, build a small hotel, cut timber, and engage in various
commercial ventures including fishing, a turtle farm, and pearl-
shell culture. In April 1 990. Urima' s workers began commercial
fishing, killing turtles and coconut crabs, and clearing land
( Kepler. 1 990a). This w as a major new insult to the atoll which,
in spite of its long history of intermittent human occupation,
remains to this day "possibly one of the least spoiled o( true
atolls in the Pacific" (Stoddart. 1976). Reports from our
expeditions to Caroline (Kepler & Kepler. 1989; Kepler.
1990a.d) resulted in the short-lived cancellation of Urima' s
tentative lease in November 1990. As of this writing. Urima
has returned to unlimited fishing at Caroline's reefs. The
government of Kiribati appears amenable to negotiations for a
wildlife preserve (see Subchapter 1.2 [Conservation section],
this volume ). Conservation efforts are presently underway tor
Caroline to become part of a triple-island wildlife presen e with
Vostok and Flint (see Subchapter 1.2 [Conservation section],
this volume).

20th Century Scientific Studies: In June 1965, afield party
from the Smithsonian Institution's Pacific Ocean Biological
Survey Program ( POBSP) visited Caroline for 2 days ( Clapp &
Sibley. 1971a). Then vsurvej and specimens added much to the
pre\ ious botanical and ornithological know ledge of the island.
Other quick visits were made b\ the Line Islands Expedition on
9-10 September 1974 and by Roger Perry, then Wildlife
Warden of the Line and Phoenix Islands, on 12-13 November
1977. Eleven years later the present authors surveyed the
terrestrial env ironments of Caroline more thoroughly than had

been previously attempted. This visit was part of the Third
Joint US-USSR Bering & Chukehi Seas Expedition, a
4-month voyage that also ineluded oceanographic studies in
the Pacific Ocean and South China Sea.

In 1990, one of the authors (AKK) visited Caroline twice
with the International Council for Bird Preservation (ICBP)
1990 Line and Phoenix Islands Expedition. These visits were
primarily to discuss conservation matters with the Falconers;
introduce Caroline to Dr. and Mrs. M. Garnett, representatives
from ICBP; confirm the illegal taking of fish, turtles, and
coconut crabs by Mr. Urima's men; find the Tuamotuan
maraes; survey North Arundel Islet; collect invertebrates; and
fill in missing details from the 1988 expedition.

Methods

From 22 to 29 September 1988, Drs. A. K. Kepler, C. B.
Kepler, D. H. Ellis (USA), and Mr. Katino Teeb'aki (Republic
of Kiribati) surveyed all 39 motus at Caroline Atoll (Fig. 2),
gathering detailed information on plants, seabirds, land birds,
mammals, reptiles, coconut crabs, and human disturbance.
Some incidental data have been added from the two visits in
1990 (10-13 March, 18-28 May) by Dr. A.K. Kepler, Captain
G. Wragg, A. Garnett, M. Linsley. J. Phillips (March), and
Dr. M. Garnett (May).

Prior to the first expedition, a series of transects and known
botanical information were mapped to ensure that 5% of each
motu was sampled and to maximize our chances of encountering
all known plant species. Transects on the larger motus (Nake,
Long, South ) were spaced approximately 400 m apart and, with
one exception, were perpendicular to the long axes of each islet
(Tr. 3 on Nake extended first from east to west, then ran south
parallel to the west coast). If the motus were greater than
400 m long, we used two transects. Transects on the smaller
motus passed through their widest points (Fig. 8). Their
lengths ranged from 77 m (Azure) to 2,000 m (Tr. 3, Nake).

Considerable modifications were required when we realized
that the Clapp& Sibley ( 197 la) map (Fig. 7) was incorrect. We
redrew the transects on Arundel's 1883 map (Fig. 4), secured
just prior to the expedition. On South Island, due to impenetrable
draperies of Ipomoea vines, Tr. 3 was omitted, Tr. 5 ran only
from the lagoon south to the Ipomoea curtain (75 m), and
Trs. 4 and 6 ran north and south until we reached an impasse
(PI. 7).

Compass headings were determined by the configuration
of each island. Beginning at high water mark, all distances
(islet dimensions, widths of reef flats and substrates, and plant
communities) were measured using hip chains with
biodegradable cotton thread. These parameters were later
checked against aerial photographs in stereoscopic pairs taken
in 1 985 by the Royal New Zealand Air Force ( RNZAF), which
provided 3-dimensional overviews of every islet. Vegetation
maps for each islet were constructed by drawing the outlines
and plant communities (based on the aerial photos) on graph
paper, enlarging, then counting dots.

Data were collected in a 30-m (15 m to each side of the
transect) swath along each transect and recorded on field
forms. Within each plant community we took photographs.

assessed the relative abundance of each plant species (rare,
uncommon, common, very common, abundant, and locally
common; see Vegetation section ) and recorded plant community
width, plants collected, and substrate type. We also estimated
the maximum height of the dominant vegetation and percentage
of ground area covered by each species.

In addition to the linear transects, an additional 19,300 m
of perimeter surveys were conducted on 21 islets (Fig. 9). The
combined distance for linear and perimeter transects was
32.6 km. Seven tiny islets (Noddy Rock, Skull, Atibu, Bosun
Bird, Coral, Reef-flat, Fishball) were surveyed completely.

In 1988. we camped on the atoll for seven nights,
establishing base camps ( Fig. 8; Pis. 6,8) on the northwest point
of South (22-24 September) and southwest Long
(25-28 September). We relocated camps using a Zodiac with
outboard motor and an inflatable Sevylor canoe (PI. 9). All
transects were surveyed during daylight hours, beginning at
dawn. Walking the interislet channels was relatively easy at
low tide, but became hazardous at incoming or high tide, as the
numerous black-tipped reef sharks, Carcharhinusmelanopterus
(PI. 10), regularly charged at our legs and had to be beaten off
with sticks and coral chunks.

During the 1990 visits, we stayed on Motu Ana- Ana with
the Falconers. Work was not intensive, as in 1988. We walked
or motored an inflatable Lancer, visiting 20 motus: Ana-Ana,
Kimoa, Pisonia, Eitei, South, North Arundel, Noddy Rock,
Brothers, North Brothers, Skull, Pig, North Pig, Bo'sun Bird,
Long, Nake, Mouakena, Shark, Scarlet Crab, Bird, and Fishball.
Insects were preserved in ethyl alcohol. We used a "Magellan"
NAV 1000 to obtain accurate geographical coordinates of
Caroline.

Naming Caroline 's Motus

Previous literature has provided vague or incomplete data
on Caroline's constituent motus (Bennett, 1840; Holden &
Qualtrough, 1884;Markham, 1904; Stevens &Barwick, 1930;
Bryan, 1942; Clapp & Sibley. 1971a; Garnett, 1983). This
confusion resulted because most previous visits had been brief.
The only charts available were a quite accurate survey by
Arundel, a guano merchant who mapped the atoll in 1883
( Fig. 4; Admiralty Chart, No. 979, 1 965 ), and a map. greatly in
error, drafted by an international Solar Eclipse Party, also in
1883 (Fig. 5). There are no hydrological navigation charts.
Unfortunately, the eclipse party map has been used in all
subsequent scientific, historical, and sociological publications
(Bryan, 1942 [Fig. 6]; Maude. 1968; Clapp & Sibley, 1971a
[Fig. 7]; Garnett, 1983). It shows only 25 of the 39 islets and
many of the shapes are distorted. The 38 islets on Arundel's
map are similar to those in the RNZAF (1986) aerial
photographs. Only a few islets appear to have changed in
minor ways since 1883: major discrepancies in Arundel's map.
we believe, are due to difficulties involved in the accurate
rendition of small land areas (i.e., the South Nake Islets).
Maude (ca. 1938) counted 36 islets but never published his
information.

To aid our survey we named 32 islets and 4 islet groups
(Fig. 2). Our names, whether in English or Gilbertese, reflect
appropriate aspects of islet biology. Etymology is provided in

the Description and Ecology of the Motus section. Gilbertese
names are prefixed with motu. Any name not appearing on
Arundel's map ( Fig. 4) was given to the islets by us. They have
been sent to the British Admiralty and US Hydrographic Office
(along with corrections to the Pacific Pilot) for official
recognition.

Structure and Topography

General Account

Caroline, one of the oceanic islands contributing to
Darwin's theory of atoll formation (Darwin. 1842), is a low
island derived entirely from coral reefs and mollusks, living
and dead. Although undoubtedly resting upon an ancient
basaltic base, today's atoll reveals no visible fragments of its
volcanic heritage. "The atoll consists of a chain of twenty-five
[sic] little islets, well covered with trees and shrubbery, the
whole forming a quiet scene of grove and lake, charmingly set
off by the contrasting ocean" (Holden & Qualtrough, 1884).
This description could apply equally well today. Caroline is
actually composed of a ring of 39 "permanent" and 3 incipient
islets whose total area is 398.94 ha. Most are well wooded, but
four tiny ones, less than 0. 1 ha in size, are scarcely more than
coral rubble piled on the reef, supporting sparse patches of
Tournefortia and Heliotropium. One islet. Noddy Rock
(PI. 19), is a vestige of a former reef segment.

Our terminology is based on Tracey et al. (1955) as cited
by Wiens (1962), to which we have one addition. The name
motu, Polynesian for "islet" or "small island," is now a technical
term for detrital reef islands (Danielsson, 1954; Stoddart &
Gibbs, 1975). In this paper the terms motu and islet are used
interchangeably; however, as motu is now a bona fide English
word and not italicized, it may be pluralized by adding an "s"
(normally Polynesian words are not pluralized with "s").

Caroline's overall shape is like a flattened crescent. 9.7 km
long on its north-south axis. Its perimeter is 26.9 km. measured
along the outer reef. Its greatest breadth ( including both reefs),
2.3 km, lies centrally along an east-west transect that includes
Motu Mannikiba and lower Long Island. The longest islet.
Long, extends 4.23 km from north to south, while South Island,
extending 1.2 km from east to west, claims the widest stretch
of land.

The motus, lying upon a wide, continuous reef flat which
encloses an elongated, relatively shallow lagoon, fall naturally
into groupings of 3 large islands (South, Nake, Long) and
4 groups of smaller islets! 13 Windwards, 5 Southern Leewards,
1 I Central Leewards, 7 South Nakes). There are four basic
motu shapes:

/. long, linear, and parallel to the reef axis (e.g.. Long
Island);

2. small, linear or oval, and perpendicular to the reel axis
(e g., Southern Leeward Islets i;

3. triangularorcrescentic.withtheapex lacing the seaward
reel (e.g.. most of the Windward Islets); and

4. large and quadrangular, occupying the ends of the atoll
(e.g.. South. Nake).

All individual motus are discussed m detail in the
Description and Ecology of the Motus section.

Reef Flats

Caroline's peripheral reefs, which completely surround
the lagoon and upon which the motus rest, are consistently wide
(average 562 m, range 396-759 m, N = 100). The windward
and leeward reefs differ in structure and dimensions. Neither
are entirely dry, even at the lowest tides. They consist primarily
of barren calcareous rock, which on other atolls generally
represents the erosional surface of an older reef. Jagged
"mushrooms" of newer (but dead) reef dot the leeward reefs
(PI. 11 ). There are no passes from ocean to lagoon, a typical
feature of central Pacific atolls (Wiens, 1962). The combined
area of intertidal and subtidal marine environments that they
enclose is several times the areaoccupied by terrestrial habitats.

In the Southern Hemisphere, reef flats tend to be widest in
the southwest sector and narrowest in the northeast (Wiens,
1962). Caroline's reefs are quite wide throughout the west
(PI. 12), and definitely narrowest in the northeast off Nake
(PI. 13).

The reef rim. irregularly dentate and 26.9 km in
circumference, is surmounted by islets for 55% of its length.
On 72% of all Pacific atolls, less than half the reef circumference
is occupied by land (Wiens. 1962). Caroline lies within a 28%
minority in which one-half to two-thirds of the reef rim contains
land. Corresponding values for 2 Tuamotuan atolls. Rangiroa
and Raroia. are 33% and 35% (Stoddart & Sachet. 1969).
Where motus exist, the reef flat is divided into the seaward reef
flat (PI. 12), islet, and lagoon reef flat (PI. 14).

At low tide all reef fiats are wadable. Black-tipped reef
sharks were a threat to our safety in most areas in 1988, but by
1 990 dozens had been killed. The reef segments that separated
the Southern and Central Leeward groups and the Leeward and
South Nake groups were particularly hazardous. Within these
islet clusters, names such as Blackfin, Shark, and Danger
reflected this ubiquitous feature of Caroline.

The only tidal measurements taken were by the Solar
Eclipse Party (Holden & Qualtrough. 1884). who noted that in
May 1 883 the greatest daily fluctuation ranged from a maximum
of 475 mm (1' 7") to a minimum of 125 mm (0' 5"), similar to
that (around 2') in the Tuamotus (Stoddart & Sachet. 1969);
Arundel's map (Fig. 4) gives 0.5 m ( 1.5') for Caroline, which
we have tentatively used in the schematic profiles
(Figs. 34-36) as the difference between low and high spring
tidal levels.

Windward Reef Flats: Constantly pounded by surf
(PI. 16). the windward reefs are typically narrower than those
to leeward, averaging 5 19 m ( range 396-759 in), though this is
less evident from a map than in the field.

The windward reefs are 13.5 km long, surmounted by 16
motus that total 63% of its length. This is not surprising, as a
recurrent pattern on central Pacific and Tuamotuan atolls is that
motus are more frequent along windward reef rims (Thomas.
1961; Wiens. 1962). The longest islets are Nake ( 1.980 inland
Long (4.226 m), both formed from the coalescence of two or
more smaller islets (Figs. 37-39). The rest vary from 1 S m to
858 m in length.

The character of these reef Hats differs, depending on the
presence or absence of land, interisland distances, lagoon
depth, and recent weather conditions. In February 1990. part

10

of a hurricane hit Caroline, rearranging tons of coral rubble and
sand on the windward beaches and motus, tearing out
Tournefortia scrub, obliterating much of the native herb mat,
exposing beachrock, depositing large chunks of broken reef on
the seaward reef flats, and changing the sizes and shapes of the
beach crests.

Reef Rim w ith Motu: The w idth of the seaward flats is
quite uniform, averaging 307 m (range 1 93-396 m), occupying
517c of the rim width. It consists of a slightly raised algal ridge
bearing the brunt of incessant wave action and a rubbly reef flat,
partly drying at low water, which sweeps up to the motu's
beach (PI. 16). Adjacent to Nake, in the northeast, the seaward
margin forms a shallow moat separating the land and algal
ridge (PI. 13).

The motus differ considerably in width, ranging from the
narrow tip of Long, merely 30 m wide, to Windward, 290 m
wide. Nake and South islands, forming "caps" to the atoll at its
upper and lower ends, respectively, exhibit characteristics
more typical of windward than leeward motus. Because coral
debris accumulates wherever atoll reefs turn sharply (Thomas.
1961), these two motus are the widest on the atoll (Pis. 16.17).
A comparison of maps a century apart ( Figs. 2.4) indicates that
several layered ridges of coral debris have accumulated on
northern Nake since 1883.

Reel Rim without Motu: Zonation within the reef flat is
less marked where there is no land. Within these interislet reef
flats, however, certain areas of high water transport have
carved surge channels and grooves. These are particularly
e\ ident at the north and south ends of islets ( PI. 18), between
islets, and within the longer flats. In all these areas, tidal tans
extend into the lagoon especially at its northern end where
sedimentation is most active. Caroline has no deep pass or
navigable channels into the lagoon nor a ship anchorage
beyond the reef, though small boats may anchor within the
close lee of South Island during normal trade winds and low
seas. Landing in an inflatable is best made across the reef
slightly north of the "boat entrance" (marked by an upright
anchor I.

The reef flat between Tridacna and South Island, serrated
with 6 erosional grooves, one labelled "blind passage." is of
particular importance to navigators. Its most southerly channel
is a narrow diverticulum 380 in long within a reef 430 m wide.
On all previously published maps this passage is drawn as
though it completely connects ocean and lagoon (Figs. 4-7).
However, it is a true blind channel (Fig. 50), serving as a
sheltered anchorage for motored yachts near its lagoon end. hut
cannot be entered or exited during high winds or moderate-to-
high surf.

Leeward Reef Flats: These are wider, flatter, gentler, more
consolidated, and less filled with rubble than the windward
reefs(Pl. 1 1 ). Everywhere except within the surge channels, an
orange-colored alga blankets the coralheads. chunks of upraised
coral (PI. 11). carbonate rock, and giant clams. This alga is
found on many atolls — for example Enewetak and Rangiroa
(Stoddart & Sachet. 1969: US Department of Energy. 1987).
Living coral is sparse.

Surge Channels: These occur in a variety of shapes and
sizes, depending on the distances between motus, the extent
and buildup of reef flats adjacent to land, and lagoon depth.
Surge channels and reef grooves (hoa) are deeper on the
windward side. Aerial photos indicate that the vigorous
currents washing daily into Caroline's lagoon have created
largerdebris fans between windward motus than between those
to leeward (see Chapter Frontispiece).

Beaches

Caroline's beaches — the zones lying between low water
mark and the inland limit of wave-deposited debris — are entirely
of reef origin. There is, however, considerable variation in the
sizes of coral rubble, and the proportions of sand and silt with
which they are mingled. In general, the windward beaches and
surge channels, subject to winds and swells and in a constant
state of erosion or deposition, support the greatest variety of
sediments: well-sorted sands; gravels of coralline, algal, and
molluscan origin; and a wide variety of sizes of coral rubble.

Almost all exposed rubble on Caroline is colored from
pale to dark gray , aconsequence of penetration by cyanobacteria
(Fosberg, 1953). Typically the oldest rubbles, highest up the
beach and extending into the interior, are darkest. A marked
beach crest rises — gently or abruptly — from the windward
beaches, at the crest of which is deposited an assortment of
flotsam and jetsam: bottles, plastic, wood, coconuts, et cetera
(Pis. 16,20). No large chunks ofdisengaged reef were found on
or near land in 1988. but in 1990 many of these littered the
windward reefs and shores, the result of cyclonic weather in
February of that year. Similarly, in 1990, considerably thicker
deposits of coarse sand had overlain the rubbly windward
beaches and interislet channels of 1988.

Alterations to Caroline's beaches provide the major areas
of change in islet shape. The main areas of aggradation are on
the northwest and southwest leeward points of the
windward islets. This is particularly true of the largerones such
as Brothers (which has joined with a separate islet
mapped by Arundel ). and Windw ard and Tridacna ( which have
added more sediment to their southwest points during the last
century).

Beachrock: These narrow, elongated strata of eroded
upraised reef, brown consolidated sands, and reef detritus are
not abundant on Caroline. Occurring as strips at the low water
mark, they flank the windward beaches of Nake, Long, and
South (Figs. 37.38.50; PI. 58) and a few of the leeward islets.
They become more exposed alter violent storms. Extensive
areas of a coarser conglomerate (PI. 21 ) hug the inner reef flats
of western South Island.

Upraised Reef: In a lew areas, jagged, eroded upraised
reef {champignon or /<•<<) is evident — for example, the lower
quarter of Long. A thin soil cover supports a forest of lower
stature than would otherwise be expected. In this minikarst
area the rocky substrate is pitted with holes of varying sizes and
undermined with subterranean tunnels in which at least two
species of land crabs (Birgus latro, Cardisoma spp. ) shelter
( PI. 22). Noddv Rock (PI. 1 9). the smallest motu (0.02 ha), and

11

many jagged coralline "mushrooms" found on the reel Hats
(PI. 11 ). are probably remnants of former reef flats formed
when sea levels were several feet higher than present.

Lagoon

A notable attribute of Caroline's lagoon fromaconservation
perspective is its outstanding clarity and beauty. Throughout,
but especially near the Central Leeward Islets, the twisted reef
configurations studded with white sandy channels and deeper
circular openings generate a spectrum of stunning colors —
turquoise, apple green, tawny, azure, and royal blue (PI. 25).

Caroline's lagoon. 8.9 km long, is closed. Though it
appears to dominate the atoll, its total area is less than that of
the combined reef flats. The lagoon is relatively shallow,
tapering in shape and depth at each end, and is crisscrossed with
living coral. Its bathymetry is unknown.

In the north the lagoon is more sheltered, as the presence
of continuous vegetated land buffers the easterly trades. At its
northern extremity, merging reef flats squeeze the lagoon until
it disappears east of Pandanus Islet. A filled-in portion of the
former lagoon penetrates Nake for 300 m as a fishhook-shaped
mudflat, Sandy Inlet (Fig. 37. PI. 23). before succumbing to
encircling vegetation. At the lagoon's southern end, where
winds whip through the "blind channel." it is choppy, having
more sediment and slightly less visibility. However, w ithin the
lee of South Island' s north-central curve, the lagoon is frequently
quiet and reflective (PI. 24).

Lagoon Hydrology: Although Caroline's hydrology has
not been studied, it has been closely observed for over two
years by Ron Falconer, especially the south end of the lagoon
and "blind passage.'" He has noticed that the lagoon water is
tv pically "perched" at a level above that of all but the daily high
tides. High tide water flows rapidly over the reef flats into the
lagoon but is held back by the reefs as the tide lowers. Ron has
noted that lagoon water at low tide is about 0.3 m higher than
water in the "blind passage." Water moves out of the lagoon
through a few channels that, although deep in places, form
broad, shallow troughs over the reef flats. A major channel
with a current flowing west at several knots passes along the
northwest point of South Island, although water passage is
impeded hv the reel flats west of South Island. If a channel
u ere to he blasted through the reef flats, .is has been proposed,
this delicate hydrolog) would be disrupted. For example, the
high tide w ater is never more than 20 cm above the coral heads
and reels m the lagoon. A man-made reef channel for vessels
could lower water levels 30-40 cm. thereby exposing and
killing the extensive Acropora Tridacna reefs within the lagoon.

The "blind passage" northeast ol South Island (Fig. 50) is
sustained by a powerful northward How of water along the east
coast of South Island and a strong southward flow of water
along the seaward reels of Tridacna Islet. The South Island
flow is apparently augmented bv water draining from a large
shallow basin on the reef Hats south of the island. Water spills
into the "blind passage" and drams east at about 4 knots against
(he pre\ ailing Hade w inds and surf. There is much less current
at the west (inner) end of the passage, where less water is
collected, and throughout the passage at low tide i there being
essential!) no water How out from the lagoon).

Patch Reefs: A complex series of patch reefs and coral
knolls ( primarily Acropora spp. i. circular and elongated, flank
the smaller motus and crisscross most of the lagoon's area (see
Chapter Frontispiece). They are particularly evident in the
southern two-thirds of the lagoon, where they approach and/or
break the surface. Coral limestone bedrock, surmounted by a
variety of living coral, mollusks. and other invertebrates,
provides their basic structure (Fig. 10). The atoll's perimeter
reefs shelter the knolls from storms, surf, and excessive erosion.

Caroline's lagoon is gradually filling in with ever-
expanding patch reefs and debris washed in from the fringing
reefs. Since Arundel's time, the effects of detrital deposition
can be discerned as changes in the shapes of islets such as Nake.
Danger, and Arundel (compare Figs. 2 and 4 1.

Such change is typical of atoll evolution. Geologically.
Caroline is a few steps behind one of its "neighbor" Line
Islands, Christmas, where sediments and coral growth have
converted the original lagoon into a maze of supersaline.
minilagoons and tiny islets, mostly cut off from the sea. Further
steps in evolution are exemplified by completely filled-in atolls
such as Jarvis and Vostok, where not even salty pools remain.

Tridacna- Acropora Reef: Though the giant clam
(Tridacna maxima) is an abundant component of Caroline's
lagoonside reefs, exceptional aggregations flank the most
southerly windward motus (Brothers through Tridacna.
Figs. 44 to 48 ). Two especially outstanding reefs extend across
the lagoon from Tridacna to Ana-Ana (Fig. 10: PI. 26) and
Tridacna to Kimoa (Fig. 48). where Tridacna are attached to
Acropora spp. corals, a favored substrate (Braley. 1987).
Abundant inshore Tridacna on all these islets suggest that their
density is similar to that on the main reef: up to 20/. 25 nr
(i.e.. 80/m2), averaging 35/nv for the entire area surveyed
(Sirenko& Koltun. Subchapter 1. 4. this volume). This exceeds
the highest densities known: up to60/nv at Reao Atoll. Tuamotu
Archipelago ( Richard. 1985). Densities of 6-20/m2, at Takapoto
Atoll (Tuamotus) are considered high. Throughout Caroline,
the clams averaged 18 x lOcminsize. Several species of giant
clams have suffered greatly in the Indo-Paeific from poaching
and ov erharv esling: few undisturbed populations exist ( Braley .
1987). Caroline is thus a special refuge for T. maxima.

Lagoon Reel Flats: These varv considerably but are
narrower and more gently sloping than the seaward reef flats.
They are typically covered with fine coral gravel and coarse
sand In sheltered areas dower Long. Windward. Crescent.
South, upper end of lagoon), lush shrubbery — Cordia,
Toiirnefortia, Pisonia, Cocos — overhangs the lagoon. Here
fine silt. sand, and/or an algal slime are common (PI. 28). In
ll)88. narrow, sandy beaches occurred only on the north shore
of South (PI. 24) and east side of Shark (PI. 29). but in 1990,
sand was more common throughout Caroline.

Where the lagoon shorelines arc less sheltered and
vegetation does not overhang the lagoon, unvegetated rubble
and sparse herb mats are typical. Here, lagoonside rubble
av erages less than 2 m wide ( PI. 3 1 ). This contrasts with their
seaw aid reef Hats, w Inch av erage 2 1 m w ide (PI. 12).

Lagoon Reef Fauna: A Brief Summary : Caroline's
marine environment is rich vet essentially undocumented;
know ledge of its ecosystems is limited to small, outdated lists

12

offish and invertebrates (Dixon, 1884; Fowler, 1901; Pilsbry
&Vanatta. 1905a.b). However, preliminary studies on mollusks,
benthic invertebrates, and lagoon plankton were begun by
Soviet scientists in 1988 (Tsyban & Smith, 1988; Sirenko &
Koltun. Subchapter 1.4. this volume). The lagoon and reefs are
remarkably pristine, having changed little since their first
discovery; all early travelers remarked on their beauty,
abundance, and variety (Bennett, 1840; Markham, 1904).

The usual assemblage of tropical invertebrates —
echinoderms. mollusks, crustaceans, porifera, corals, tunicates,
et cetera — are present. Large numbers of black sea cucumbers
(PI. 10). about 20 cm long, are particularly abundant on the
lagoon reef flats of the southern windward islands. They have
been tentatively identified as Ludwigothuria sp. (B. Sirenko,
personal communication). Conspicuous fish families include
paiTot fish (Scaridae), butterfly fish (Chaetodontidae),
surgeonfish (Acanthuridae), damselfish (Pomacentridae),
pufferfish (Tetraodontidae). and wrasses (Labridae).

Substrata

Throughout the atoll, the substrata reflect a coralline
origin. There is little "soil" in the accepted sense. Various
grades of jagged, eroded coral and mollusean rubble (from fist-
sized to tiny pebbles), together with sand, coralline algae, and
small proportions of organic litter, humus, and guano, are
present. Such accumulations of reef and terrestrial debris are
similar to those of other low, coral atolls ( Fosberg, 1953;
Stone. 1953: Wiens, 1962; Niering. 1963; Stoddart & Sachet.
1969; Reese, 1987; Garnett, 1983).

Generally speaking, atoll soils are calcareous and extremely
immature, a consequence of their limited age and frequent
disturbance by storms. Barely modified beyond the reef that
spawned their presence, they are composed primarily of calcium
and magnesium carbonates. Water retention, if any, is due to
accumulated organic matter and its associated chemical changes.
This is particularly important with respect to guano, which
reacts with coral sand and humus to form phosphatic hardpan
and nitrogen-rich "soils" (Fosberg. 1953).

Reese (1987) categorizes atoll "soils" into five types, all of
which occur, in different proportions, at Caroline. The degree
of organic matter, decomposition, amount of humus, and the
depth of the "soil" strata are directly correlated with age and
size of the motus.

/. Accumulations of coral rubble, mainly of stone size.
These youngest of "soils" are most evident around the edges of
the motus. acting as a substrate for pioneer herb mats. Often
extending well inland, they can support surprisingly lush
Tournefortia scrub.

2. Unaltered coral sand and gravel. Although exposed
sand was uncommon at Caroline in 1 988, this substrate occurred
intertidally where the lagoon was filling in and on actively
growing sandbars, primarily in the upper lagoon (Pis. 23,28),
northeast and northwest South (Pis. 24,32), and the lagoon ward
edge of Shark (PI. 29). In 1990, sand was more ubiquitous on
Caroline as a result of the deposition of tons of sand during the
severe February storm.

3. Soils with a weakly developed A-horizon, with color
i mly slightly darker than the unaltered sand below, but with no

evidence of structural development. Especially evident in
1988 within the ancient interislet channels that compose Long
Island (PI. 33), much of this substrate is now storm-eroded and
overlain by fresh sand.

4. Soils with a more developed A-horizon. deeper and
darker than above, with some structural development. This
stage defines areas where the rubbly/sandy substrate approaches
a true, but poor, "soil." As such, it represents older, more stable
parts of each island. It is common within the islet interiors
where Pisonia is (or was) present. Humus and guano fill the
gaps within the irregular shapes of eroded coral. Its composition
may be likened to a coarse mixture of gravel, sand, bones, and
shells, all mixed with sparse amounts of partly-decomposed
litter. Land crabs are particularly numerous, helping break
down organic matter into finer particles.

5. Soils with an accumulation of raw humus on the surface
and with a relatively deep A-horizon. During this stage
phosphatic hardpan may develop. These true soils, though
somewhat depleted by guano diggers, cover significant areas
on South and Nake. Cocos and/or Pisonia debris adds greatly
to their dark color and moisture content. This earthy substrate
is composed primarily of rotting Cocos fronds and fibers that
have been shredded by coconut crabs. Patches of blackish
muck on South Island support local patches of Polynesian
arrowroot (Tacca leontopetaloides).

On Caroline, hardpan (PI. 76) was present in several areas
(primarily South, Nake, Long, Emerald, and Mannikiba),
supporting herb mats and Tournefortia scrub. This substrate
may be likened to an old asphalt road.

Caroline provides an excellent example of the progression
of soil development through islets of different age and size
classes (see Ecological Succession section). From a wave-
washed mound of coral rubble, barely above sea level (Fig. 5),
the substrate gradually improves in texture and fertility as the
emerging islet ages and organic matter accumulates. Pioneer
plants are hardy, salt-tolerant, low-lying mats consisting
primarily of Heliotropium, and later, Tournefortia. Increasing
numbers of shrubs provide shade and branches for nesting
seabirds. Larger trees (Pisonia. Cordia, Morinda) add more
shade and thereby increase humidity, as well as provide
opportunities for additional organic "fallout": leaves and bird
remains (nests, eggs, chicks, droppings, regurgitated food,
dead adults).

Each stage of substrate development accelerates the
accumulation of organic material and helps to define an
emerging, deeper A-horizon. Soil maturity is indicated by
more organic matter, improved soil texture, and a lowered
volume of coralline and mollusean debris. Caroline's soils
barely exceed several centimeters in depth and are always
intermingled with coral fragments. As a result, they are
unsuitable for burrowing seabirds, such as petrels and
shearwaters.

Hydrology

Hydrological information is essentially lacking. No
standing fresh water exists. The quality, extent, and salinity of
the freshwater lenses, as well as their variability according to
tide, season, and rainfall, are unknown. At the time of Caroline's

13

discovery" ( 1606), de Quiros and his party were desperate for
fresh water. After noting how lush and green Caroline was,
they expected to find good water supplies, but there was
"nothing but salt water in the holes they dug" (Markham,
1904). Maude ( 1968) suggested, in retrospect, that had they
waited longer the salt water in their shallow wells might have
run fresh, as has been his experience on some other atolls.
During the 19th century, three wells were used — one on Nake
and two on South (Holden & Qualtrough, 1884). One South
Island well contained fresh water at 1.5 m depth in 1974
(Garnett. 1983). We saw no wells but located concrete cisterns,
one built near the northwest point of South in 1937 and rebuilt
by the Falconers in 1989. and another uncovered one (dating
from 1938) within a Cocos—Pisonia grove along Tr. 2, about
200 in east of the southwest corner of Nake.

Caroline's paucity of fresh water may be partly responsible
for the lack of a permanent population. The annual rainfall in
1989 (Appendix 2) was 1,242 mm (48.9"). However, like the
similarly lush Nikumaroro and Orona (Phoenix Islands),
Caroline's rainfall may vary greatly from yearto year, resulting
in undependable water supplies. In the past, residents relied on
rainfall catchment for fresh water (Maude, ca. 1938; R. Falconer,
personal communication).

Shallow sources of fresh or brackish water may be present
on most islets, as Pisonia forests occur on 14% of them, but
very little is known of freshwater lenses supporting Pisonia
forests (Wiens, 1962). If we assume that Pisonia is not salt-
tolerant, small water lenses may be present on motus as small
as 0.2 ha. This is further discussed in the Ecological Succession
section.

Climate

Meteorological records for Caroline were sparse until
1989, when Ron Falconer began daily records of rainfall and
wind direction (Appendix 2). Some data is available from the
plantation years 1916-1920 (Young, ca. 1922) and during the
1883 Solar Eclipse Expedition (20 April-8 May) (Upton.
1884), when 203.2 mm (8") fell. The best generalizations on
weather conditions in this area are found in the Geographical
Handbook Series: Pacific Islands, Volume 2 (N.I.D.. 1943).
Wiens (1962), Seelye (1950), Taylor (1973), and various
papers on the Tuamotus (Stoddart & Sachet, 1969; Sachet,
1983). Islands in the Line Group experience a wide range of
climates. In general, those near the equator are dry, with
rainfall increasing with increasing latitude north or south.

Caroline experiences a tropical oceanic climate that varies
little during the year. Temperatures are uniformly warm to hot,
normally tempered by trade winds from the southeast to
northeast. Falconer (personal communication) has recorded an
annual average of 30°C (86°F) (range 26-3PC [78-88°F|).
Mean annual temperatures for the Central Equatorial Islands
lie between 24 and 29 C (75-85°F). Surface temperatures
increase rapidly in early morning and remain hot throughout
the day. forest interiors are humid. The daily range of
temperatures exceeds the annual fluctuation in the daily mean.

Atmospheric pressure, sunshine, and cloud cover are
probably similar lo the northern Tuamotus — uniform except

during storms. During our 1 988 surveys the sky and air were
extremely clear, but we experienced much heavy cloud and
rain in March 1 990, the aftermath of cyclones "Peni" (centered
near Vostok) and "Ofu" (centered further west).

Wind and Rainfall: Caroline is dominated by trade winds.
As on all low atolls, land topography has no appreciable effect
on weather. Although it lies within an area primarily influenced
by southeast trades, there is a small annual oscillatory movement
northward and southward, so that in reality winds blow from
the east, northeast, and southeast. This accounts for the east
and northeast winds that puzzled the eclipse party, who were
expecting winds from the southeast (Upton. 1884). Data from
Falconer (Appendix 2) indicates that, at least for 1989 and
1990, winds blow primarily from the north and northeast, and
rarely from the southeast (April-August).

The atoll lies within a belt of variable rainfall, along with
Vostok, Flint, and the northern Tuamotus. Young (ca. 1922,
p. 13) notes that Caroline's rainfall is "certainly less than that
of Flint," giving exact figures for 1919 (2,172 mm) and 1920
( 1.854 mm) and estimates that there was "probably not more
than 50" (1270 mm)" during 1916. 1917. and 1919. These
estimates were based on exact figures from Flint ( 1 .600. 1 .346,
1,295 mm, respectively). Falconer measured 1.242.1 mm
(48.9") in 1989 and 2,209.8 mm (87") in 1990. An unusually
stormy February in 1990 brought 640 mm (25.2") of rain.
Rainfall distribution isohyets (Taylor. 1973) assign Caroline
an annual precipitation of approximately 1,500 mm (60"), a
perfect average ofthe above 6 years (x= 1,513mm). Ingeneral,
"winter" (May-October) corresponds roughly to a dry season
and "summer" (November-April) to a wet season.

Hurricanes and Tsunamis: Atoll islets are active structures,
undergoing repeated death and rebirth. Violent storms contribute
to ongoing erosional and rebuilding processes. Storms deposit
debris not only along the shores of the windward islets
(Pis. 17,20) and across reef flats into the lagoon but sweeps it
far inland.

Although the south-central Pacific is relatively free of
cyclonic storms (cyclones, typhoons, hurricanes), they do
occur with enough frequency and devastating force that any
discussion on climate should include them. Although detailed
records of hurricanes and tropical storms exist for the inhabited
Tuamotus since European discovery, many of these may not
have affected Caroline. However, the following evidence
suggests that Caroline experienced two major hurricanes last
century and that periodic violent storms can modify the atoll
substantially:

/. Between IH22 and 1825. When de Quiros visited
Caroline in 1 606. the northwesterly Cocos plantation on South
Island was healthy. When Bennett arrived in 1834. he noted
that all the palms were "of dwarf stature." and that "amidst the
original groves, the number of vigorous seedlings fully
confirmed Captain Stavers' statement [who had visited the
atoll in 1828] that these palms had increased greatlv since his
last visit to the spot" (Bennett. 1840).

A few years before 1 828. therefore, something had affected
the palms. By 1834 they were all of an even height and quite
short, vet bore nuts. French records indicate that two devastating

14

storms whipped through the Tuamotus during this time — in
1 822 and 1 825 (Sachet, 1 983 ). At least one of these could have
affected Caroline.

2. The 1878 cyclone. The first unambiguous record of
major devastation at Caroline comes from the letter of a certain
J. M. Salmon, dated 1883 and reproduced in Holden (1884).
Speaking of the time when Messrs. Brown and Brothers took
possession of Caroline (somewhere between 1865 and 1872).
he stated that "it seemed as if there had been a storm or
hurricane at some short period previous, which had desolated
the place." Arundel ( 1 890) attributed this to a tidal wave that
swept across the Pacific from South America to New Zealand
and Australia in 1868 (Arundel, 1890). but atolls do not
generally suffer greatly from tsunamis because they lack
focusing relief. Hydrographer of the Navy (1931. Vol. Ill,
p. 154), however, referring to Caroline, clearly states that in
" 1 878 a cyclone passed over the islands, destroying most of the
coconut trees."

The Great Britain Naval Intelligence Division (N.I.D..
1 943. p. 490). in reference to Caroline, also states that "in 1 878
a hurricane wrought g-eat destruction." This was possibly the
storm of 6-7 February 1878. an extremely violent one which
killed I 17 persons on Kaukura Atoll. 750 km southeast of
Caroline in the Tuamotus (Sachet. 1983).

3. The 1990 storms. We know from Arundel's chart
( Fig. 4) that no major islet-altering storm has hit Caroline Atoll
since 1883. However. oursecondvisittoCarolinewas2 weeks
after cyclone "Peni," centered near Vostok (February 1990).
affected the atoll. Violent winds, torrential rain, and high seas
had uprooted vegetation in some windward areas and greatly
altered Caroline's shorelines, interislet channels, tidal fans.
and incipient islets from our 1988 visit. Sand and rubble had
been rearranged on both windward and leeward islets, Motu
Atibu virtually disappeared, and the main interislet channel
that divides Long Island had lost its herb mats and many
Tournefortia shrubs, becoming smothered with fresh sand.

Because islets on coral atolls rarely exceed 5 m in elevation,
the tidal surges associated with Class IV orClass V hurricanes.
often exceeding 5 m in depth, can overwhelm them, not only
altering or destroying the vegetation, but in extreme cases
completely removing them from the coral rim (Frisbie, 1944).
It is essential to consider the ephemeral nature of Caroline's
islets in the discussions that follow.

Sea Conditions: Because the most extensive coral rubble
deposits occur around northern Nake and southern South
Island, and because the Cocos plantation of northwest South
was so badly hit by storms last century, the following Tuamotuan
generalities (Newell. 1956) probably also apply to Caroline:
/. prevailing trade winds from the east give heavy seas on
the northeast or windward side;

2. southern ocean swells generated in the sub-Antarctic
break heavily on the south or seaward side; and

3. occasional hurricanes or tropical storms strike in the
northwest or stormward quarter.

Vegetation: Vascular Plants and Floristics

Botanical History

All early visitors to Caroline described a well wooded atoll
with numerous islets whose vegetation extended to the shoreline.
It has changed little in the 384 years since its Western discovery.
The first botanical collection and notes were those of Bennett
in 1835 (Bennett. 1840), who recorded 10 flowering plants and
a fern and planted Tahitian chestnut, sweet potato, and
Polynesian arrowroot. The location of his plant collection, if
it still exists, is unknown (Clapp & Sibley. 1971a). Evidently
only a single nonnative species (Cocos), surviving as two small
groves, persisted until the late 19th century. Beginning in
1885. coconuts were planted extensively on South Island and
south Nake. but the copra industry failed twice, and from 1929
to 1987 the atoll was essentially uninhabited.

Dixon made the first true botanical collection in 1883
during the Solar Eclipse Expedition (in Trelease, 1884). All
specimens were from South Island except Laponea ruderalis.
His collection included several ornamentals and vegetables
that have not been reported since, an important point as these
temporary introductions have since been cited in the literature
as part of Caroline's 35 plant species. Many were not found by
the POBSP party, yet because no scientific investigations had
been conducted for 80 years, they were counted as part of the
atoll flora (Clapp & Sibley. 1971a). Three more visits to
Caroline, plus periodic searching by the Falconers, have also
failed to uncover most of these ornamentals. Since Caroline's
occasional occupants tended gardens (Lucett, 1851), it is
evident that many introduced plants have died out, lacking
constant care.

Vascular Plants of Caroline Atoll

Plant Collections: To avoid duplicating Long's plant
collection (Clapp & Sibley. 1971a). we collected only
5 specimens in 1988 and 33 in 1990. Dr. D. Herbst assisted
with identification, prepared and deposited the specimens with
Long's in the Bernice P. Bishop Museum. Honolulu. Hawaii,
with duplicates in the U.S. National Museum. Washington.
D.C. Collection numbers preceded by 'K' were collected by
A. Kay Kepler; those preceded by 'L' are those of the late
C. R. Long. Earlier collections of Bennett in 1835 (Bennett.
1 840) and Dixon in 1 883 ( in Trelease. 1 884) are noted by date
only.

Working with Long's location records for some species
has proven difficult. He was working with an incorrect map
(Fig. 7). which showed only 25 islets instead of 39. Much of
his work was done at night, which in some places would have
made it hard for him to determine his exact location. His
references to South, Long, and Nake are undoubtedly correct,
and presumably the following: "second islet south of Long" =
Crescent; "islet northeast of South Island" = Tridacna: and
"fourth islet north of Bird Islet" = Emerald. Long records
Pandanus on the "second islet south of Nake Island." which

15

lacked Pandanus when we surveyed the island. Moreover, the
first islet south of Nake supports an extensive grove of large
Pandanus trees on its eastern (lagoon) shore, and we feel
confident in ascribing Long's specimen to this island, which
we had named "Pandanus" because of this grove. To be
consistent, we have ascribed all his other "second islet"
specimens to Pandanus Islet as well and assume he made no
collections on the actual second islet (Danger).

Species Lists. Annotated Checklist, and Maps of Terrestrial
Vascular Plants: Following recent authors Sachet & Fosberg,
19S3;Lamberson, 1987), we do not considerCaroline's transient
or extinct vascular flora (Table 1) or the vegetables and
ornamentals in the Falconer's garden as part of Caroline's
viable flora. Table 2 summarizes the current flora, detailing the
relative abundance of each species within each plant community .
These tables are based on sight records supplemented by all
collections, past and present. No beach drift seeds are known
from Caroline apart from those species already represented.
English and Gilbertese names in Tables 1 and 2 are from
Thaman ( 1987), St. John (1973). and Perry & Garnett (n.d.).
If no common name is available, the Hawaiian name, familiar
to many students of Pacific botany, is used.

One new species record for the atoll, as yet unidentified, is
called Species A (K-90-23. 24): a single sterile shrub 2.5 in
high, collected from southwest Motu Mannikiba in coarse
strand rubble, is similarto Clerodendrum inerme, with leathery
leaves and arching stems 4-5 m long. It was found by John
Phillips.

Table 3 lists the distribution and abundance of plant
species ( with subdivisions into tree, shrub, and herb components)
onallmotus. Figures 1 1-25 map the entire atoll distribution of
each species according to data from transects and aerial maps.
Families are arranged phylogenetically. according to
losherg & Sachet ( 1 987 ). with species arranged alphabetically
within each family. The taxonomy of vascular plants follows
W. Wagner etal. ( 1990). and ferns follow H. Wagner (personal
communication). "'< cover" means the percentage of the
ground area covered by a particular plant species. In all text and
tables, the following symbols apply:
New record for Caroline

Indigenous — plants native to Caroline but also occurring
elsewhere (I)

Aboriginal introduction — useful plants brought by

Polynesians in pre-historical times (AI)

# Recent introduction — plants of accidental or deliberate

introduction alter Western discovery of the atoll (Rl)

A Abundant — generally the major or dominant species in

a given area
VC Very common — often seen but not quite as abundantly

as above
C Common — generally distributed throughout a given

area in large numbers
UC Uncommon — observed uncommonly but more than 10

times in a given area
0 Occasional — here and there, often widely scattered but

not forming a major component of the vegetation
R Rare — observed 2 10 times in a given area
S Single — only one specimen observed

L Local — found only or principally in one or more restricted

areas
D Drift seedling — plant derived from a water-borne seed
+ Not seen 1988-1990 but probably still present

PSILOTACEAE

* Psilotum nudum (L.) Beauv (Fig. 1 1 )

Formerly Known Distribution: L-3233 from Nake
Present Distribution: Cosmopolitan, common on remote
islands, rare on Caroline. K-90-15 from South. In 1965.
common on wet base of Cocos only on Nake Island. In 1988
and 1990. a few clumps found similarly on South Island in
shady, damp locations, close to lagoon, northwest sector.
Cocos canopy was 18 m.

POLYPODIACEAE

* Phymatosorus scolopendria (Burm. /) Pichi-Sermolli
(Fig. 12: PI. 34)

Phymatodes scolopendria (Burm./) Ching

Polypodium phymatodes L.

Polypodium scolopendrium Burm./.

Microsorium scolopendria (Burm.) Pichi-Sermolli
Formerly Known Distribution: Recorded 1840. collected
1884; L-3244, L-3250, L-3287 from Nake, Long, and South
Islands.

Present Distribution: Range extension from 3 to 1 1 motus.
Rarely a continuous ground cover, usually locally rare to
abundant. Commonest on Nake. with cover 10-80%. Well
represented on South, especially in open areas of the interior,
where soils are moister. On other motus local distribution
varied from less than 1 to 80%; accurate mapping is difficult.
Absent from motus less than 0.6 ha in size, where habitats
cannot provide cover, moisture, and substrate for both
sporophyte and gametophyte generations.

Ecology: Hardy. Leaves burn in sun but can withstand
very dry conditions. Primarily in Toumefortia scrub, mixed
forests with Pisonia and Pandanus. or Cocos plantations.
Associated with Cordia, Morinda, Suriana. In open clearings
within dying Cocos forests, occurs in dense mats intermingled
with Boerhavia, Ipomoea, and Portulaca. Sometimes gathers
in thick bands at the interface of Toumefortia and Pisonia
forests. Prefers shelter, high humidity, "soil." and relative lack
of wind, but absent from deeply shaded forests. Rhizomes
never exposed on ground surface or epiphytic on trunks, as in
wetter islands such as Hawaii or Samoa (personal observation)
or m the moister Line Islands (Wester. 1985). indicating that
Caroline's habitats are suboptimal. Although most ferns are
not halophytic, tins species grew (rather stunted) in ll>NN
amongst sparse herb mats ( 1% cover) on older beach sands of
an ancient reef channel on Long Island (Tr. C). where rainfall
pi ov ides the sole fresh water, but was ( temporarily?) obliterated
in February 1990. Rare to uncommon in outer beach strand,
and beach scrub w ith Suriana on South, Arundel, and Shark.
Substrata: Dry coral rubble, sand and gravel, rubble with
sparse humus, lagoon mud, relatively fertile humus, older
beach sands.

PANDANACEAE

* **? Pandanus tectorius Park. (Fig. 11; Pis. 35-38,50)

Formerly Known Distribution: Recorded 1840,
unidentified Pandanus; L-3227, Pandanus Islet, seen on Nake
by Long.

Present Distribution: A minor plant community ( Plant and
Communities section), Pandanus is primarily associated with
Tournefortia or Pisonia on the leeward motus. Most common
on Nake. with Cocos and Ipomoea. Range extension from two
to seven motus.

Phenology: Flowers and fruit in October, March, and
May.

Substrata: Variable. Prefers lagoon mud, pure sand, and
rubble-humus, but survives in almost pure rubble.

GRAMINAE

* +? Digitaria species

Collected 1883 and recorded as IPanicum (Digitaria)
marginata. Examined by Long, who believes it a Digitaria
identical to his L-3235. Not found by the authors.

* Lepturus repens (Forst. f.) R. Br. (Fig. 13; PI. 2a)

Formerly Known Distribution: Collected 1883; L-3211,
322 1 . 3236, 3238, 3247, 3259, 3286 from Windward, Tridacna.
Nake, Long, Emerald, Crab, and South Islands, respectively.

Present Distribution: On most dry Pacific atolls. K-88-4,
5 ; K-90- 1 , 2, 1 9 to 2 1, 25 from South, Tridacna, and Ana- Ana.
On Caroline, range extension from 6 to 26 motus.

Ecology: Patchy, rare to locally common. Usually in
exposed herb mats with Heliotropium, Laportea, Portulaca,
and low Tournefortia scrub. Abundance 1-5% cover where
not in thick patches. Occasionally inland under Tournefortia,
Cordia, or Cocos, fitting the generalization that Lepturus,
though a pioneer, will often persist as undergrowth in forests.
Tufts tiny (few centimeters), dry and scrappy in exposed areas,
but to 3 dm where shaded. Never in tall, upright clumps, or with
the same abundance as on the drier, filled-in equatorial atolls or
islands with sandier habitats (Christopherson. 1927; Fosberg,
1953, personal observation). Never forms a turf.

Substrata: Able to survive in coral rubble of varying
coarseness, down to high water mark, but preferred habitat is
part sand. L-3286 was from "numerous clumps under Suriana
scrub on South Island," perhaps the low, sandy portion of the
northwest point ( PI. 45 ). our best Lepturus site. Comparison of
Arundel's chart (1883), recent aerial surveys, and earlier
photographs indicate that several motus have altered shape
since 1883. The amount of open area on South Island has also
decreased markedly since 1883. The distribution of Lepturus
parallels these changes; there is clearly much less on South
Island, and more in newly-created islet fringes.

Since 1965 the lagoon shore of South Island has become
overgrown by Cocos, so much that both Suriana and Lepturus
are much less common than previously (Pis. 39,40). However,
sand and debris will always be shifting, so that Lepturus will
move from place to place, establishing wherever conditions
permit. In the second situation, a comparison of Pis. 2a and 24

from 1 883 and 1988, respectively, shows that a century ago the
lagoon-facing shores of South Island were far more open than
the dense Cocos plantations of today. The clumped grass in the
foreground of PI. 2a is undoubtedly Lepturus, probably mixed
with introduced grasses not seen since that time (Eleusine
indica, Eragrostis plumosa) and the dubious Digitaria sp.,
above.

PALMAE

**# Cocos nuciferaL. (Figs. 14,36,51; Pis. 2,6,18,24,34,37,44)

Formerly Known Distribution: Recorded 1840, 1884;
L-3285 from South Island, extensive groves on South and
Nake, scattered on north portion of Long.

Present Distribution; Range extension from 3 to 1 5 motus.
Planted groves on South, Nake, and Long; the rest derived from
drift.

Phenology: Flowers and fruit year round.

Ecology: Forms a major vegetation type (Plant
Communities section). Primarily South and Nake, where
closed canopy forests average 2 1 m high.

TACCACEAE

# **? Tacca leontopetaloides (L.) O. Kuntze (Fig. 1 1 )

Tacca pinnatifida Forster

Formerly Known Distribution: Normally an aboriginal
introduction on Pacific islands, but on Caroline is first mentioned
as planted in 1834 (Bennett. 1840); L-3213 and 3219, and
K-90-7 and 90-16 from moist muck, South Island. L-3234,
common under Cocos and numerous patches found in muck,
south end, Nake.

Present Distribution: Common in northwest South. None
in flower; each plant had two to three leaves, possibly dying
back. None found on Nake, despite searching the south end.
Has large underground tubers, dies back, and though cultivated,
still occurs spontaneously in Cocos groves on many atolls.
Currently harvested by the Falconers.

Ecology: Needs fine, moist soil and shade. Though its
seeds float for months (Guppy, 1906), it will probably not
become established on any other motu, due to the prevalence of
rubbly substrates.

Phenology: Flowers and fruit in March and May, dies
back in October.

URTICACEAE

* Laportea ruderalis (Forst. 0 Chew (Fig. 15)

Fleurya ruderalis (Forst. f.) Gaud, ex Wedd
Formerly Known Distribution: Reported 1840, collected

1884. L-3215 common in shady areas South Island;

L-3229 scattered on exposed coral and sand, west side Crescent

Islet. L-3253 under shade of Cocos and Pisonia on north side

of Long Island.

Distribution and Abundance: K-88-3 South Island, Tr. 1,

elevation 0.3 m, under old Cocos plantation, in humus and

rubble. Range extension from 3 to 32 islets (Table 3).

Commonest and most widespread ground cover,

17

patchily distributed. Rare to locally abundant, percentage
cover from less than 1% in herb mats of tiny motus to 60% in
tall Pisonia forest. Best represented on Nake, Long, Brothers,
South, Pisonia, Eitei, and Mannikiba, where coverage exceeded
50% in appropriate habitats. To 1.1 m tall on Kimoa.

Ecology: Largest specimens found under Tournefortia.
Pisonia, Cocos, or Pandanus. Tiny (1-2 cm) and tougher in
sunny, exposed sites. Halophytic, pioneering in herb mats on
islets less than 0.75 ha in size (e.g., Fishball). Optimum habitat
is Tournefortia scrub, in sunny clearings, or belts behind beach
scrub. Uncommon in Pisonia forest. Occurs in both windward
and leeward sites, but in greater density leeward. Will persist
through several stages of plant succession if given adequate
shade.

Phenology: Flowers and fruit in October. March, and
May.

Substrata: Primarily beach gravel or coarse rubble. Also
rubble-sand mixtures; not lagoon silt.

OLACACEAE

* (#?) Ximenia americana L. (Fig. 1 1 )

Never previously collected. K-90-170 South Island.
50-100 m north of cistern, elevation 0.3 m. 10-20 m from
coastal Tournefortia fringe, within Cocos plantation. Collected
by crew of the yacht Amanita and posted by Anne Falconer to
AKK.

Distribution and Abundance: Locally abundant in one
location, about 50 bushes ( 3—4 m high, 2-3 m wide ) spread over
about 100 m. Adjacent to indigenous scrub, on edge of Cocos
plantation near old settlement.

Phenology: Flowering in July 1990.

AMARANTHACEAE

* Achyranthes canescens R. Br. (Fig. 16, PI. 41 )

Never previously collected. K-88-1 South Island,
Tr. 5, to 0.7 m, elevation 0.3 m, in Tournefortia fringe, coral
rubble.

Distribution and Abundance: Quite widespread, primarily
in interior scrub and forest of 19 motus (Table 3), from tiny,
barely vegetated Fishball (0.73 ha) to the largest. South
(1 06 ha). Density variable: from less than 1 % in Tournefortia
scrub to 50% local ground cover in mixed Pandanus forest.
Primarily associated with Tournefortia. May be locally abundant
in clearings in Pisonia forests, pure or mixed. Often in a zone
dividing Tournefortia and Pisonia trees, especially on Pig,
Brothers, and Nake.

Ecology: Never in natural herb mats. Needs shade but
requires some direct sun: rare in pure stands of Cocos and
Pisonia. Prefers small, sunny openings in forest or scrub.
Drought-resistent and probably partly halophytic. Dies back
annually in the dry season and reappears with winter rains
(Anne Falconer, personal communication). To 1.5 m tall.
Little or no capacity for dispersal by sea. On other islands,
seeds carried by birds, especially fruit pigeons (Guppy, 1906),

but pigeons are absent from the Line Islands. Perhaps dispersed
by the long-tailed cuckoo (Ellis et al., 1990).

Phenology: Flowers and fruit present in October, March,
and May.

Substrata: Lushest growth in humus soils of forest interiors.
Often grows in pure rubble.

NYCTAGINACEAE

* Boerhavia repens L. (Fig. 17, PI. 34)

Boerhavia diffusa L.

Boerhavia hirsuta: Sensu Bennett, 1840

Boerhavia species: Dixon. 1884

Formerly Known Distribution: Reported 1 840, collected
1 884; L-32 1 0. 3324, 3239, 3225, 3252, 3262, 3289, 329 1 from
Windward, Tridacna, Nake. Long, Emerald, and South,
respectively.

Present Distribution: Cosmopolitan, widespread in the
Pacific. K-90-164and 165 from Ana- Ana. Range extension on
Caroline from 6 to 33 motus (Table 3).

Abundance: Present in every habitat, leeward and
windward, ranging from less than 1 to 80% cover. Often in
unpredictable patches. Best locations (>50% cover) on Nake,
Long, Windward. Pig, Brothers, Arundel. Tridacna. South.
Ana-Ana, Pisonia. and Pandanus Islets.

Ecology: Mostly found beneath Tournefortia. either in
pure scrub or mixed with Pisonia. Cordia. Morinda. Suriana,
or Cocos. Not in deep Pisonia shade; rarely in herb mats. Thick
ground cover in indigenous scrub ( Shark ) or within clearings in
old Cocos-Ipomoea forest (South), where it mingles with
Phymatosorus, reaching a high density (PI. 34) and large size
(rooting at nodes, vines exceeded 1 mlong). BIRDS: Bristle-
thighed curlews fed within the Boerhavia mat in old Cocos
forests. South. Sticky seeds (.32 cm [one-eighth inch] long)
found entangled in preened down and adhering to contour
feathers of a juvenile great frigatebird (PI. 42). Species is
customarily dispersed around large oceanic areas and within
atolls by seabirds such as red-footed boobies (Guppy. 1906;
Ridley. 1930).

Phenology: Small mauve flowers and seeds present in
October, March, and May.

Substrata: Coral rubble with sand or humus, rarely pure
beach rubble. Lushest growth in humus-and-guano-laden
rubble clearings where Pisonia forest once grew.

* Pisonia grandis R. Br. (Fig. 18; PI. 43)

Formerly Known Distribution: Collected 1884; L-3280
4 m tree, north shore. South. Small grove, north end. Long.

Present Distribution: Cosmopolitan, pan-Pacific. Caroline
range extension from 2 to 29 motus (Table 3).

Abundance: A major plant community (see Plant
Communities section). Caroline^ Pisonia forests, some of the
last remaining groves in the Pacific, are of special conservation
value.

Substrata: Occupies, and contributes to. best soils on atoll:
mixture of rubble, humus, and guano.

18

PORTULACACEAE

* Portulaca lutea Solander ex Forster F. (Fig. 19; Pis. 34,38)

Formerly Known Distribution: Reported 1840 and 1884;
L-3233 and 3292, 3231 . 3237, 3255, 3257, from South,
Pandanus. Nake. Long, and Emerald, respectively, in open
coral, rubble, gravel, and exposed areas, to 1.5 dm high.

Present Distribution: Range extension from 5 to 33 islets
(Table 3).

Abundance: Along with Heliotropium anomalum is a
component of the plant community. Natural Herb Mat (see
Plant Communities section). Widespread, predictable on coast
and former reef channels but local inland. Covered from one
to 60% of land area on almost every transect, windward and
leeward, especially facing lagoon. Best areas are Long, Tr. 4
(36-m wide meadow); South, north end of Tr. 6 (50 m wide):
Brothers, lee, almost pure mat covering 20% ground (6 m
wide); Kimoa. north side (8 m wide), 10 cm high; Eitei, north
side, 5 cm high.

Ecology: Primarily occurs along edges of motus in rubble
mat and open Tournefortia scrub, averages 12 cm high.
Prominent in sparsely vegetated areas, extending seaward to
high tide level. Halophytic; highly tolerant of sun. A Hat mat
in exposed areas but lusher inland, rising to 2 dm tall. Generally
found with Heliotropium, Lepturus, Boerhavia, or Laportea
but may form pure mats. Uncommon in Tournefortia scrub
patchy in clearings within Pisonia forests up to 13 m high
Exceptionally common in old Cocos groves with Boerhavia
et cetera (PI. 34); otherwise rare or absent from closed canopy
Cocos plantations. Pinker stems found in sunny sites. BIRDS
Provides nesting cushion for masked booby, sooty tern, brown
noddy. On noddy rock, brown noddies nest on a thick mat of
pure Portulaca. Feeding location for shorebirds.

Phenology: Flowers and fruit October, March, and May.

Substrata: Coral rubble and gravel, fine to very coarse.
Healthier on older sands and coral-humus.

ZYGOPHYLLACEAE

* Tribulus cistoides L. (Fig. 1 1 )

Formerly Known Distribution: Collected 1884. L-3245in
open sandy area among Tournefortia shrubs. Long Island. Not
seen elsewhere on atoll.

Present Distribution: Not seen on our surveys, but present
in 2 sites on west-central Long Island. K-90-161 (collected by
Anne Falconer), probably from one of same sites as 1965
collection. Flowers in March.

SURIANACEAE

* Suriana maritima L. (Fig. 20; Pis. 6,21,39,40,44)

Formerly Known Distribution: Collected 1884. L-3220,
shrub to 1.8 m, east edge of Tridacna Islet.

Present Distribution: K-90-5, 6 from South Island. Range
expansion from one to 9 motus (Table 3).

Abundance: Occasional on Caroline. Forms a vegetation
unit. Beach Scrub with Suriana ( see Plant Communities section).

Phenology: Flowers in March and May.

Substrata: Best sites in sand but also on coral rubble.

EUPHORBIACEAE

# Phyllanthus amarus Schum. and Thonn. (Fig. 1 1 )

Formerly Known Distribution: Collected 1884. L-3283,
herb Phyllanthus niruri L. (Trelease, 1884) to 4 dm, common
on north side of South Island.

Present Distribution: K-90-10-13, herb, 2 small patches,
South Island. Limited to a few square meters in the atoll's only
weedy area, less than 10 m: in two small clearings by the
recently-renovated cistern. South. A fairly common weed in
the Society and Tuamotu Islands, therefore probably arrived
with Tahitian copra-cutters and perhaps again within the last
2 years. Caroline's only established "weed" (excluding
Polynesian introductions such as Cocos).

MALVACEAE

o * (**?) (#?) Hibiscus tiliaceus L. (Fig. 1 1 )

Never previously collected. K-90-8, 90-9 from South
Island, northwest peninsula, in Cocos plantation near old
settlement and "landing," in coral rubble and humus, 0.6 m in
elevation.

Present Distribution: Two or three large spreading trees in
heavy Cocos shade, 1 0 m tall, with recumbent branches forming
an impenetrable thicket. This species, culturally important to
Polynesians, is either indigenous, an early Polynesian
introduction, or an ornamental brought by 1 9th century settlers.

° * (**'?) (#?) Thespesia populnea (L.) Soland. ex Correa
(Fig. 1 1 )

Never previously collected. K-90-22, 154. 155 from
South Island, in Cocos plantation and in lagoon strand, northwest
peninsula, near "landing."

Present Distribution: Two trees ( 10 m tall), one near the
cistern, the other in a fringe of native vegetation bordering the
lagoon. The history of this species is probably the same as
Hibiscus tiliaceus (above).

* SidafallaxWalp. (Fig. 1 1)

Formerly Known Distribution: Collected by Dixon, 1884,
who found one specimen.

Present Distribution: Not seen for 106 years. K-90-156,
157, 158 from South Island, at edge of cistern, north side. One
clump located in a sunny clearing, recently enlarged by the
Falconers.

CONVULVULACEAE

* Ipomoea macrantha R & S (Fig. 21; Pis. 34,37)

Ipomoea tuba (Schlecht.) G. Don
Formerly Known Distribution: L-3228 and 3293, 3242,
325 1 on South, Nake, and Long, respectively. Trailing vines,
white flowers, stems to 25 m long climbing over Tournefortia.
Morinda. and Cocos.

19

History: Not collected last century, though plantation
records indicate that it was a major reason for the abandonments
of the coconut plantations: "The Pohue Vine |misidentifiedas
Tuumfetta(=Triumfetta)procumbens], which is the worst pest
on the island, was reported in 1921 to be under control"
(Young, ca. 1922). Today it still causes severe damage to
Cocos on South Island, strangling about two-thirds of the
plantation (54 ha).

Present Distribution: Range extension from three to seven
motus, five Windward and two Southern Leeward Islets
(Table 3).

Abundance: Forms part of a vegetation subunit. Dying
Cocos-lpomoea Forest (PI. 34, Plant Communities section).
An indigenous, nonparasitic vine, becoming abundant and
strangling in disturbed areas. Rampant growth over most of the
interior of South Island, where it forms dense tangles up to
25 m high. Less dense thickets on southern Nake drape
Pandanus, Tournefortia, Morinda, and Cocos to 10 m. Our
transects on Nake were not rerouted or abandoned, as on South.
Coverage scant elsewhere, generally 2-5%, except in two
Pisonia sites, where its coverage was 20% (Long Island. Tr. B:
Windward Islet, Tr. 1 ).

Ecology: Lush in dying Cocos forests and mixed forest
with Pandanus, because of relatively fertile soils, moisture,
humidity, and partly sunny clearings. Strangles all but the
tallest Pisonia and Cordia. Typically sea-dispersed to atolls
(seeds germinate after floating up to 1 year in seawater), crawls
inland, progressively dropping seeds, to attain full size in
interior forests (Guppy, 1906; Ridley, 1930). Seeds of /. pes-
caprae are known to be ingested by white terns in the Marshall
Islands, perhaps as gizzard stones (Fosberg, 1953). Possibly
these same terns, abundant at Caroline, once aided the seed
dispersal of /. macrantha. Also characteristic of Cocos
plantations elsewhere in the Pacific (Fosberg. 1965: Stoddart
& Sachet. 1969: Lamberson. 1987).

Substrata: Prefers humus-laden rubble but can grow in
coarse rubble and sand, especially in leeward areas.

BORAGINACEAE

* Cordia subcordata Lam. (Fig. 22: PI. 27)

Formerly Known Distribution: Collected in 1X84. L-3213
and 326 1 a, 3228, 3246. and 326 1 b on South. Pandanus. Long,
anil Emerald, respectively; flowering trees to 4.5 m high in
leeward coral rubble or along lagoon.

Present Distribution: Africa to Polynesia. K-90-3 from
South Island, lagoon edge. Range expansion on Caroline from
5 to 23 motus (Table 3).

Phenology: Peak flowering November through April,
fruits collected in March and Max

Abundance: A separate, though minor, plant community
(Plant Communities section). Caroline's Cordia forests,
typically small and mixed \\ ith other emergents. are some of
the last remaining gun es in the Pacific and are thus of particular
importance to conservation.

* Heliotropiumanomalum H. & A. (Fig. 23: Pis. 17.33.45-47)

Formerly Known Distribution: Recorded (as
H. curassavicum) in 1840, collected in 1884. L-3222 and 3288.
3240, 3248, 3256, 3288 on South, Danger. Long, and Emerald,
respectively.

Present Distribution: Pantropical. K-90-17 from Ana-
Ana. In coral gravel, leeward, and windward shores. Range
extension on Caroline from 4 to 34 motus (Table 3).

Abundance: Forms part of a major vegetation unit. Natural
Herb Mat (Plant Communities section), often associated with
Laportea, Lepturus, or Boerhavia. Area coverage ranges from
less than 1% to 50%. Widespread, predictable on wind- and
salt-blown, low flats where vegetation does not overhang edge
of motu. Also in ancient reef channels and newly evolving land
connecting islets. Covers major areas of islets — that is. those
less than 1.0 ha (e.g., Fishball. Skull, and Bo'sun Bird. Best
developed on Skull. Tridacna. South. Emerald, and Mannikiba
(50% coverage, western seaward rim).

Ecology: Halophytic pioneer. Heights to 22 cm, averaging
7 em. Thrives in heat and exposure.

Phenology: Flowers and fruits year round.

Substrata: Primarily coral rubble and rubbly sand. Marginal
habitats extend down to high tide line in areas of coarse coral
chunks, where it is tiny and leathery.

* Tournefortia argentea L. (Fig. 24; Pis. 8.37.47.48)

Messerschmidia argentea (L.f.) Johnston

Formerly Known Distribution: Collected 1884. L-3216.
3226, 324 1 , 3249, 3258 from South, Tridacna, Nake, Long, and
Emerald Isle; shrub to 3 m high, edge of lagoon and above high
tide, with white flowers,

Present Distribution: Range extension from 5 to 38 motus
(Table 3). Widespread in the Pacific, especially on small islets.
Caroline's large tracts are excellent examples of relatively
undisturbed, pure Tournefortia scrub and forest.

Abundance: Dominates the atoll woodlands, forming the
major \ egetation type ( Plant Communities section). On almost
every motu ranging from a spattering of exposed shrubs within
herb mats, through scrublands and taller forests to 14 m high.

Ecology: Supports seven species of breeding seabirds;
provides feeding habitats for reef herons (Egretta sacra).
shorebirds, land crabs, and rats.

Phenology: Flowers and fruits year round.

Substrata: Pure coral clinker; mixtures of rubble, gravel,
sand, and humus.

BRASSICACEAE

* Lepidium Indentation Montin (Fig. 1 1 )

Formerly Known Distribution: Reported in 1S25: "a boat
load of pepper-grass and pursley" (Paulding. 1X31) and in
1835. "a Lepidium of luxuriant growth" (Bennett. 1840).
Collected by Dixon as /.. piscidium Forst in 1883.

Present Distribution: Widely distributed throughout the
North and South Pacific. K-90-169 and 171 (collected by

20

Alexandre Falconer), on Tridacna and Pisonia, most probably
in coastal Tournefortia scrub.

RUBIACEAE

* Morinda citrifolia L. (Fig. 25; PI. 48)

Formerly Known Distribution: Reported 1840, collected
1884. L-32 14. 32 17 and 3282: 3232; 3254 on South, Nake, and
Long, respectively.

Present Distribution: K-90-4, 18 from South Island,
lagoon edge, and Ana-Ana. respectively. Range extension on
Caroline from 3 to 30 motus (Table 3).

Abundance: Coverage from 2% to 50%. Basically an
inland species, widespread and predictable in scrub and forest
understory. Rarely a component of the canopy, except on
Raurau. where Morinda grows 12 m tall in a 13 m Pisonia
forest. Essentially associated with established Tournefortia
woodlands on motus greater than one hectare in size. Quite
common on South Island despite major disturbance, occurring
within beach strand, Cocos plantations, and Cocos-Ipomoea
interior. Best locations (40-60% coverage): Nake. Tr. 3;
Tridacna, both transects; Long, Tr. 8; Raurau and Ana-Ana.

Ecology: Appears early in plant succession: inToumefortia
scrub as an early pioneer (Stage I), then from Stages II to IV,
progressively becoming more common and robust. Not in pure
Pisonia forest (Stage V). Much less common in Pandanus
stands. Although it thrives best in light to heavy shade,
preferably growing in moist "soil," one leathery seedling (7 cm
high) had gained a foothold in exposed rubbly Heliotropium
flats on Fishball Islet.

Biogeographical Note: Generally considered a naturalized
aboriginal introduction on most Pacific islands. Morinda could
be native to Caroline, as theorized for the northern Line Islands
(Wester, 1985). Although possibly introduced by early
Tuamotuan settlers, its present distribution strongly suggests
that it is indigenous. Throughout the atoll Morinda occurs in
the greatest densities on motus with no anthropogenic forests
or in areas distant from historical settlements (Fig. 26). On
Nake, Morinda occurs frequentl — in places abundantly — within
the interior Pisonia forests, yet its coverage is only 5-10% in
mixed Pandanus— Cocos forests in the southern sector. It also
appears to be part of natural biological succession (Table 6).
Further support for this theory comes from nearby Flint.
Though there is no direct archaeological evidence that Flint
was settled in prehistoric times (Garnett, 1983), our 1990
surveys found Morinda in all habitats (mixed woodland, native
coastal scrub. Cocos plantations, and abandoned settlement).

Originating in Polynesia, Morinda has been widely
dispersed by man but has apparently also spread, unaided by
man. "widely by sea in the Malayan and Polynesian Islands"
(Ridley, 1930). Its air-filled, buoyant pyrenes can float for at
least 53 days and "its seeds are almost certainly disseminated
by birds and bats" (Guppy, 1906). It could also be disseminated
by Coenobita crabs and rats within and between motus, as has
been found elsewhere by Ridley.

Phenology: Flowers and fruits year round.

Substrata: Coral rubble, gravel, sand, and humus. Rarely
found in coarse clinker. On larger motus. prefers moist soils
under tall forests.

GOODENIACEAE

c * Scaevola sericea var. sericea Vahl (Fig. 1 1 )

Scaevola taccada var. sericea (Vahl) St. John

Never previously collected. K-88-2, Windward Islet,
central-windward side, elevation 0.3 m.

Distribution and Abundance: One wind- and salt-sheared
"hedge," found by K. Teeb'aki on Windward Islet, was growing
on a coarse rubble beach. "The saltbush..., being recorded for
the first time too from the island. ..covered approximately 3%
of the islet's land area [this probably can be translated as "3%
of the area covered at that location on the transect." as we
understood from Mr. Teeb'aki's description that it was quite
small], occupying the mid-windward side. The patch grew
very low — only up to 2' high with its foliage forming an
extended raised mat canopy all along the area it occupied"
(Teeb'aki, 1988). We have been unable to return to this spot to
observe and photograph it directly.

Because Scaevola is hardy, halophytic, and widespread in
the Pacific, it is surprising that it is so rare on Caroline.
However, none occur on Vostok, and only one clump is known
from Flint (Clapp & Sibley, 1971b; Garnett, 1983). Fosberg
(1953) noted that Scaevola seeds are transported by bristle-
thighed curlews (Numenius taitensis) in the Marshall Islands:
curlews are common on Caroline ( Subchapter 1 .2, this volume)
and could have brought seeds from elsewhere.

Substrata: Coarse rubble, windward beach.

° * Scaevola sericea var. tuamotensis (Si. John)Fosb. (Fig. 1 1 )
Scaevola taccada tuamotensis St. John

Never previously collected. K-90-168 (collected by
Alexandre Falconer), northeast peninsula. South Island, in
coral rubble.

Present Distribution: One individual, of unknown size.
with Suriana and Heliotropium, northeast peninsula, South
Island, facing the inner side of the "blind passage."

Floristics and Ecology of the Motus

Size of the Flora: Atoll floras characteristically lack
diversity. Numbers of species range from 3 to around 150 in the
Pacific and 284 in the Indian Ocean. The flora of the Southern
Line Islands is particularly impoverished because of /. their
easterly location ( far from the major source areas of Australasia);

2. low profiles (most only rise a few meters above sea level);

3. lack of topographic diversity (most have a very limited range
of habitats); 4. low to medium rainfall (approximately
1,500 mm p. a.); and 5. edaphic factors such as salinity, highly
calcareous soils, et cetera. Long-distance dispersal and hardiness
are important factors in establishing a flora, especially since the
closest high island, Tahiti, is 830 km away, and the ultimate
source of its flora, the Malayan-Melanesian region, is over
8.000 km away. South America, the closest continent, is

21

approximately 9,000 km distant. The motus of Aitutaki, for
example, at a similar latitude but further west and wetter, are
considered depleted with 45 species. Fanning, at a similar
longitude but wetter, has 123 species. Tarawa, 3,900 km to the
northwest, receives a similar rainfall but supports 109 species.

Where an atoll's potential flora is larger, the increased
shade and greater protection from wind, salt spray, and storms
result in a greater number of natural plant species on its larger
motus. However, such atolls are generally inhabited and
alterations by both aboriginal and modern man have modified
their original flora. Caroline's isolation, variety of islet areas,
and minimal human disturbance all contribute to its excellence
for the study of atoll evolution.

The number of species presently established on Caroline's
39 motus is 27 (Tables 2, 3). The previous expedition in
1 965 (Clapp & Sibley, 1971a) collected 20 species, of which 4
were new to the atoll. Their total of 35 species, however,
incorporating reports and collections from the 1800's, is
misleading. Our total, 6 of which were new records, would
have brought the atoll total to 44 (plus about 15 more
unestablished, mostly garden, plants). However, following
recent custom (see Vegetation section ), we have listed transient
orextinct members of the flora separately (Table 1). To include
them would obscure the near-pristine nature of the atoll and
bias our analyses of species-area relationships.

The 1 883 drawings of the South Island settlement, inhabited
when most of Caroline's species were catalogued, shows that
the island was vastly different (compare Pis. 2 and 24). A
century ago homes were set amidst large grassy clearings; now
the site is completely obliterated beneath shady 21-m-tall
coconut palms. Nine exotic plant species have not been seen
for over a century (Table 1 ). Evidently most ornamentals and
domestic vegetables perished during uninhabited periods. The
present residents struggle to keep garden plants alive because
of poor soils, irregular rainfall, and foraging land crabs. A few
native species might also have been eliminated during the
guano and copra-harvesting years.

Numbers of Indigenous Plants: A comparison of the
percentage of indigenous species between different island
groups (Table 4) shows that Caroline, with 85%* (N = 23)
indigenous, is unusually high. Only 11 of 44 Pacific atolls
reviewed have more than 75% of their species indigenous. Of
these, nine ( including Caroline) are remote and lack permanent
human occupation.

The Tuamotu Islands ( 149° to 134°W) lie east and south
of Caroline, yet they harbor considerably larger floras. Rainfall
is similar. Three of them average 121 species (Table 4).
averaging 42 indigenous species. When the variables rainfall
and distance from a colonization source to the west are
considered, the proximity of the Tuamotus to the diverse high

* Note: Perhaps as high as 93% ; the Digitaria sp., if still extant,
is of unknown identity and origin, and Species A has yet to be
determined.

islands of the Societies seems to play a major part in
determiningtheir indigenous flora. A similar situation exists in
the southern Cook Islands. Caroline and other remote Line and
Phoenix Islands are sufficiently isolated from high volcanic
islands that they exhibit a much simpler flora. Tahiti, the
closest high island (830 km south), is in the wrong direction for
direct currents, winds, or vagrant birds to bring seeds to
Caroline.

Composition of the Flora (Tables 2. 3 ): Caroline's botanical
affinities lie with other southern Line Islands and the Tuamotus.
Although the strand and inland floras consist of pan-Pacific or
pantropical species, there are several widespread species and
communities that are notably absent. Those that survive have
withstood the atoll tests of time — poor soils, scarcity of fresh
water, periodic inundation by salt water, intermittent cyclonic
storms and hurricanes, harsh climate, and high seedling
mortality. Caroline provides an excellent ecological laboratory
in which floristic correlations with variations in habitat, motu
size, and leeward/windward aspect may be studied. Fosberg
(1985) and Sachet (1967) have noted the importance of such
details in understanding the biogeography and taxonomy of
Pacific plants.

Caroline's present established flora includes only one
weed species (Phyllanthus amarus), represented by a tiny
patch less than 2 m: in size. There are two Polynesian
introductions (Cocos, Tacca). Pandanus tectorius, Morinda
citrifolia, Thespesia populnea, Ximenia americana, and
Hibiscus tiliaceus, though indigenous, may have been
introduced by Polynesians or 1 9th-century settlers. (See section
on Vegetation.)

Trees: Seven species present. Only three — Pacificwide
natives — are widespread: Pisoniagrandis. Morinda citrifolia.
and Cordia subcordata. Two are locally abundant: Cocos
nucifera and Pandanus tectorius, while the rest, Thespesia
populnea and Hibiscus tiliaceus. are rare and limited to the old
settlement site. The absence of typical Pacific species such as
Calophyllum inophyllum and Guettarda speciosa is notable, as
they occur naturally on atolls such as Rangiroa. further east
(Stoddart & Sachet. 1969).

Shrubs: Five species present, at least four indigenous.
Only Tournefortia argentea is abundant: its most abundant size
class is under4 m. Scaevola and Suriana. tough and widespread
elsewhere, are poorly represented on Caroline. It is noteworthy
that two varieties of Scaevola sericea are present. Species A is
represented by a single individual, Ximenia americana. by a
single, large patch. Pemphis acidula. though common on atolls
of similar latitude and climate, is absent from most of the Line
and Phoenix groups (Stoddart & Gibbs, 1975; Fosberg &
Sachet, no date). This may be due to the paucity of its preferred
habitats: low rocky substrates (reef rock, conglomerate rock)
and sand-gravel ridges.

Herbs: Fifteen species present, at least 12 indigenous. Of
these only seven are common: Heliotropium anomalum.
Boerhavia re pens. Portulaca lutea, Laportea ruder alls,
Achyranthes canescens, Lepturus repens, and Phymatosorus

22

scolopendria. Ipomoea macrantha and Tacca leontopetaloides
are locally abundant, while Phyllanthus amarus, Tribulus
cistoides, Lepidium bidentatum, and Psilotum nudum are rare
and localized. Digitaria sp. may be extinct. The fact that Sida
fallax has only been recorded twice in 106 years is curious.

Ecological Succession

We have attempted to trace the development of Caroline' s
flora from the smallest to largest motus, using field data and
aerial photos, which reveal past geological processes. Three
tables provide this analysis of ecological succession: Table 5
presents Caroline's motus in order of ascending size, together
with the numbers of plant species and major plant communities.
Since the atoll's total land area is small, our data provide
relatively complete floristic lists for each islet and detailed
maps of their plant communities (Figs. 37-57). The number of
species varied from 3 growing on 4 tiny islets (0.02 ha each) to
23 on South ( 104.41 ha). Because the total number of species
for the entire atoll (27) is also small, the addition of one or two
rare species contributes significantly to the total flora. Such
additions must be kept in perspective when evaluating plant
succession.

Table 3 provides a summary of plant species distribution
by islet in decreasing order of abundance, and Table 6 is a
summary of plant species distribution and relative abundance
with respect to islet area and the primary mode of seed dispersal.

Basic Serai Stages

Islets appear, grow, mature ecologically, or vanish in
violent storms. Many interacting factors, including geographical
(islet area, atoll shape, distance from high islands and continents),
geological (changes in sea level), chemical (nitrates from bird
droppings, leaf fall, et cetera), climatological (wind, droughts,
storms, microclimates), and biological (seabirds, rats, land
crabs, and man, both aboriginal and modern ), constantly interact
to change conditions. The relative influence of some of these
factors is evident when comparing the floras on motus of
different sizes.

Seed-dispersal mechanisms (Table 6) and the presence of
underground fresh water are also vital. Unfortunately, the
relationships between groundwater salinity, species distribution,
and vegetation patterns on atolls are poorly understood ( Fosberg.
1985). The presence and relative salinity of permanent water
depends on Ghyben-Herzberg lenses of varying thickness on
different islets, and this in turn depends upon island dimensions
(especially width), soil porosity, rainfall, tidal fluctuation, and
other hydrological factors. Though groundwater supplies have
been studied on many atolls (Wiens, 1962; Maude, 1953), each
island group is so unique that it is unwise to extrapolate
information from one to the other.

Caroline's 39 motus fall naturally into 4 size classes:
motus with areas of a) <0.2 ha, b) 0.2 to 0.7 ha, c) 0.8 to
25.0 ha, and d) >25.0 ha. These size groupings harbor all 5 of
the serai stages identified on Enewetak Atoll (Lamberson,
1987), tailored to reflect Caroline's particular geography,
geology, and impoverished flora. Each stage may be the sole
example of ecological succession on an islet or may occur as

one of several stages. Typically the early stages cover the
peripheral rubble and scrubby outer zones, while the later ones
appear as a series of roughly concentric bands progressing
inland.

Stage 1: Early pioneers on sandbars, spits, or small rubbly
islets subject to storm damage and washover. Harsh conditions,
intense sun, drying winds, salt spray. High salt concentration
in the substrate. Lack of fresh water and nutrients. Plant genera
present include Heliotropium, Portulaca. Lepturus, Boerhavia,
and seedling oropenTo!/r/ie/o/-f/a scrub. No Cocos. This stage
is found on many small motus (Noddy Rock, Fishball) and
former interislet channels (e.g., Long Island).

Stage II: Thick scrub of mixed genera, often impenetrable.
Its protective barrier allows for the development of vegetation
on the larger islets. Seabirds begin to contribute to the soil
(guano, eggs, regurgitated fish, decaying nesting material).
Plant genera include Toumefortia, Suriana, Cordia, and
Laportea. If Cocos present, accompanied by coconut crabs.
Very common around the periphery of most motus just inland
of the native herb mats or flanking sheltered shores adjacent to
the lagoon (South. Kota).

Stage III: Trees larger, seabirds add further to soil fertility.
Open grassland may develop in sunny clearings (Tridacna).
Added plant communities and Cordia-Tournefortia,
Tournefortia-Morinda, and Pisonia— Toumefortia forests.
Occurs in the next inner concentric zone of vegetation to
Stage II on larger motus (Nake, Long) or, more commonly, the
entire interior of smaller ones (Pandanus, Southern Leeward
Islets).

Stage IV: Pisonia dominates the older mixed forest.
Morinda and Toumefortia reach for the sun. Forests are more
open. Undergrowth mostly a ground cover of Laportea,
Boerhavia, Lepturus, and Portulaca. Covers the main portion
of larger islets. If Cocos and Pandanus present, forms a mixed
forest with vines (southern Nake, Shark). Coconut crabs
common. A widespread stage in the center of most motus
(Central Leeward, Windward Islets).

Stage V: Pisonia takes over. Other trees are confined to
the forest edges. Always in the deep interior of the larger islets.
Little or no ground cover. Abundant nesting black noddies
(Anous minutus). A more restricted stage (Brothers, Raurau,
central Nake, Pig).

Ecological Succession on Motus of Different Size Classes

To assist discussions of succession on Caroline's motus,
refer to the individual vegetation maps and graphs
(Figs. 27-57) and photographs (Pis. 13-80). particularly
Figs. 27-30. The latter figures summarize the amounts of each
islet's surface covered by each major plant community, as well
as providing the numbers and percentages of indigenous species
for each islet.

a) Motus with Areas <0.2 ha (Figs. 27,31; PI. 49;
Tables 5.6): Caroline has four motus in this category, three
windward and one leeward, whose combined area totals
0. 15 ha. There are also three incipient islets which, because of
their temporary character, have not been counted in Caroline's
overall total (Fig. 2; PI. 15). With the exception of Noddy
Rock — a jagged, upraised limestone plateau — all consist

23

predominantly of coarse coral rubble (75-98% coverage).
These liny motus are the simplest ecosystems on the atoll,
representing early Stage I in plant succession. The number of
plant species per motu averages three, all hardy, sea-dispersed,
and salt-tolerant pioneers (Heliotropium, Portulaca, Lepturus,
Tournefurtia). The sole plant community is a natural herb mat
of varying thickness and extent. Tournefortia, though stunted
and scattered, is not sufficiently common to form a separate
scrub habitat. Indigenous vegetation covers 2 to 22% of the
islet areas. Seabirds. especially brown noddies and red-tailed
tropicbirds, may attempt to nest.

b) Motus with Areas 0.2 to 0.7 ha ( Figs. 28,3 1 : Tables 5.6):
There are five leeward motus in this category whose combined
areas total 2.21 ha. Their vegetative cover is more extensive
and diverse than in size class a, with herb mats and Tournefortia
scrub and forest, but open rubble is still abundant (30-55%
cover). Plant succession corresponds to late Stage I and
Stage II. The average number of species is 8.2 (range 6-11),
one-third of Caroline's total. All vegetation on these motus is
indigenous except for a few Cocos palms. Seeds are dispersed
by sea, wind, and birds.

With the appearance of shrubs, the number of species
increases markedly, and woodlands, primarily of Tournefortia,
form and expand to create dense thickets averaging 5 m tall and
covering 25% of the land area. Canopies of 10 m occur on
motus Nautonga and Kota. Seabird colonies of up to six
species (brown and red-footed boobies, great frigatebirds,
black and brown noddies, white terns) are present.

A low herb mat, dominated by Heliotropium, Portulaca,
Boerhavia, and, more rarely, Lepturus, develops first, after
which Tournefortia quickly becomes established. Shade,
producing locally humid conditions, and better "soils" derived
from guano, decomposing leaves, and the activities of land
crabs and rats, provide appropriate habitat for Laportea and
occasional Phymatosorus and Achyranthes. The major tree
species — Pisonia, Morinda. Cordia, and Cocos — subsequently
appear but are relatively rare. Pisonia, typically an inland
species assumed to need companion trees and underground
water (Wiens, 1962; Spicer & Newbery. 1979), could well be
salt tolerant as it occurs on motus as small as 0.2 ha (Tables 5,6 ).
In this size class Pisonia occupies only 2-6% of the total islet
areas.

c) Motus with Areas from 0.8 to 25.0 ha (Figs. 29,31;
Tables 5,6): All 27 motus in this category share a similar
complement of species and plant communities (Tables 5, 6).
Their combined area totals 1 24.35 ha. They are well wooded
(Fig. 29), although the leeward motus have a higher proportion
of rubble and herb mats, and forests are higher to windward.
Unvegetated rubble covers less land area (219? ) than in size
classes a and b (87% and 39%, respectively). Within the
woodlands of these motus, substrates mature from basic rubble
to primitive "soils" with small, but significant, structural
development. Then flora shows increasing diversity with size,
and almost the lull complement of seabirds may nest.

All natural ecosystems are firmly established; canopy
heights range from 4 to 2 1 m. On Booby Islet (0.84 ha). Pisonia
suddenly becomes very common, and the Pisonia forests on
North Brothers (1.71 ha) and Pig (7.25 ha), at 21 m. are the

tallest on Caroline. As rich guano and dead foliage accumulate,
a layer of phosphate-rich humus enables those species already
present but poorly represented on the small motus (Pisonia.
Morinda, Boerhavia, Laportea, Achyranthes) to increase in
abundance and stature (Table 6). Additional species are Suriana,
Pandanus, Scaevola, Ipomoea, Lepidium, and Species A.

Plant succession, ranging from Stage III to Stage V in the
interior, primarily involves forest maturity rather than the
addition of large numbers of species. On the larger islets, the
numberof plant species increases by relatively small increments,
filling out the subcanopy layers and, in the cases of Cordia and
Pandanus, adding variety to the canopy.

The average number of plant species is 1 1 .0, ranging from
4 to 15. If we divide the motus into smaller size classes, we find
that their species numbers increase slightly with increasing
size: 8.0 species for areas 0.8-1.0 ha, 9.8 species for areas
1.1-2.0 ha, 1 1.5 species for areas 2.1-4.0 ha. 1 1.3 species for
areas 4.1-10.0 ha, and 12.0 for areas 10. 1-22 ha. An increase
in herbs (range 3-9) is primarily responsible for these higher
averages (Table 5).

Despite the large range of motu sizes in this category, plant
communities are essentially natural (Table 5). Their overall
species composition is 96% indigenous. Seventeen of the
motus lack Cocos, the only introduced species in this area
category, which is represented by small, isolated clumps or
individual palms.

On the larger motus, and within the taller forests, more
species of birds, especially red-footed boobies, great frigatebirds.
white terns, and black noddies, nest in increasingly large
colonies, furnishing more minerals to the developing soils,
especially where Pisonia covers large areas.

In summary, by the time a motu on Caroline has reached
0.8 ha in size, all the natural plant communities, most species
of trees, shrubs, and herbs, and most species of seabirds are
present. In Caroline's depauperate flora there are few species
left to increase floral diversity on the larger islets, regardless of
their size. This is very different from the inhabited atolls such
as Kapingamarangi.

Although we do not know when true freshwater lenses
develop, they may occur in motus of this size class. If we
assume that Pisonia is not salt-tolerant, limited fresh water
must be available on motus as small as 0.2 ha. and actual
freshwater lenses may begin forming at around 0.7 ha, as
indicated by the sudden proliferation of Pisonia forest
(Tables 5,6). However, the Falconers have not been successful
in finding any underground fresh water on Motu Ana-Ana
(2. 16 ha), which suggests that Pisonia may be salt-tolerant.

d) Motus with Areas >25.0 ha (Figs. 30,31; Tables 5.6):
On Caroline, no motus fall between 22 and 75 ha in size. Thus
the three motus in this category (Nake. South, Long) cover a
limited range of sizes: 75.98 to 107.50 ha. They average 18.0
plant species. The floral components and forest heights of
these larger motus (Figs. 32,33; Table 5) are essentially the
same as for class c. There are no additional ecosystems
( mangrove swamps, salt Hats, grasslands, et cetera ) or understory
layers. Ten species, all rare or uncommon, are present only on
the larger motus (Table 3): Scaevola. Tribulus, Hibiscus,
Thespesia, Ximenia, Psilotum, Tacca (introduced in 1834),

24

Phyllanthus, Sida, and the dubious Digitaria. Four, possibly as
many as eight, are indigenous. In 1965, one vine of the
indigenous Ipomoea pes-caprae was also found, but three
subsequent surveys failed to locate it.

Species-Area Relationships

The relationship between the numbers of plant species and
island size has long fascinated biologists (Fosberg, 1949;
Wiens, 1 962; MacArthur& Wilson, 1967; Whitehead & Jones,
1969), yet data from uninhabited islands is scant. The studies
from Kapingamarangi (Niering, 1956; Wiens, 1956) and
Aitutaki ( Stoddart & Gibbs, 1975 ) treat atolls with long histories
of human occupancy. SomeofthevillagesonKapingamarangi's
23 motus date to 1200 A.D. Aitutaki's 16 uninhabited motus
lie adjacent to a westernized volcanic island in an "almost-
atoll." People on both these atolls have profoundly influenced
their flora.

Caroline provides an opportunity to compare the numbers
of species on motus of different sizes in an uninhabited atoll,
then to compare the results with Kapingamarangi, Aitutaki,
and uninhabited islands in the Line and Phoenix groups that
have no introduced species and have experienced minimal
human contact.

Comparisons of Species-Area Relationships with Other
Atolls: Studies of Kapingamarangi (Niering, 1956 (contributed
greatly to theories of island biogeography (Mac Arthur &
Wilson, 1967). Because its motus coverthe same range of sizes
as Caroline, the two atolls might be expected to exhibit similar
patterns. However, their species-area relationships are
completely different. On Kapingamarangi, islets less than
1 .4 ha showed a constant, small number of species, after which
islets up to 100 ha showed a direct correlation of area with
numbers of species. On Caroline, a motu of 1 .4 ha supports
almost two-thirds of the total number of species, and plant
diversity on islets up to 107 ha shows only a slight, but not
necessarily steady, increase (Table 5).

Species-area relationships on the motus of Aitutaki
(Stoddart & Gibbs, 1975, Figs. 33 and 34 of that paper)
conformed to the Caroline model; the number of species
increased only slightly on motus from 4 to 71 ha. Unfortunately,
Aitutaki had only one motu less than 1 .4 ha, so comparisons for
smaller islets cannot be made. The floras of all three atolls have
been impacted by man, but Caroline far less so than the others.
Much of the floral diversity on larger islets at Kapingamarangi
is derived from plants introduced by man and cannot be
considered normal. Caroline and Aitutaki provide much better
samples of natural plant species-area relations on atolls.

Six islands in the Line and Phoenix groups (Maiden,
Starbuck, McKean, Phoenix, Vostok, Birnie) are uninhabited.
Their flora is entirely native. All are Caroline's "neighbors" in
an oceanic sense, and all except Vostok are dry , receiving about
750 mm (30") of rain yearly. They are old, essentially filled-
in atolls, containing hypersaline central lagoons or no lagoon
at all. Although the largest island (Maiden) has the greatest
diversity, there is only a very small linear increase in plant
species with increasing area (Table 8). Plant diversity is more
a function of climate (hot and dry) and distance from source
areas, than size, similar to the situation on Caroline.

The Question of Fresh Water: The Kapingamarangi data
were analyzed with availability of fresh water in mind (Wiens,
1962;Whitehead&Jones, 1969). These authors suggested that
1 .4 ha is the threshold at which a freshwater lens can develop.
Below this size only halophytes can survive. They argue that,
as there are only a limited number of salt-tolerant species, the
floral composition on islets below 1 .4 ha is relatively constant.
On larger islets, species numbers increase in direct proportion
to land area, because permanent groundwater promotes the
survival of an increasing variety of nonhalophytic plants.

The groundwater versus plant model does not apply to
depauperate Caroline for a number of reasons: first, the number
of plant species is not constant on islets below 1.4 ha: in fact,
species are added faster on motus from 0.02 to 1 .4 ha than any
other size range.

Second, on Kapingamarangi, the number of species
increased in direct relation to islet size from 1.4 ha to 100 ha.
On Caroline, species numbers increased only slightly from
1.4 to 22 ha and exhibited another minor increase from 70 to
108 ha (see Fig. 31; Tables 5,6; and Ecological Succession
section). Thus, Caroline's data do not support the area-
diversity theory.

Third, Whitehead & Jones (1969) argued that the flora on
"small" motus lacking a freshwater lens (i.e., < 1 .4 ha) consists
only of salt-tolerant strand species. This is not true on Caroline
(Table 6). In addition to harboring the usual strand species
(Tournefortia, Portulaca, Laportea, Heliotropium, Boerhavia,
Lepturus), Caroline' s "small" motus also support inland species
that are generally considered nonhalophytic (Pison ia, Morinda,
Achyranthes, Cordia, Phymatosorus). Either these latter five
species are moderately salt tolerant, or on Caroline the minimum
islet size with a freshwater lens is much less than 1 .4 ha, or both.

Fourth, Whitehead & Jones (1969) postulated that the
nonhalophytic species are those that control overall species-
area associations. This may be a good generalization for less
remote islands but does not hold up for atolls with depauperate
floras (Table 6). For example, on Caroline the halophytic
Ipomoea macrantha, I. pes-caprae, Scaevola sericea, Sida
fallax, Lepidium bidentatum, Hibiscus tiliaceus, Thespesia
populnea. and Tribidus cistoides, which theoretically should
only occur as strand species on the smaller islets, occur only on
larger islets. In addition, several nonhalophytes (e.g., Morinda)
were found at Caroline on small motus where one might only
expect to find strand species.

Fifth, the above authors do not mention bird-dispersal of
seeds, which is probably a factor that needs to be taken into
account on remote islands: at Caroline, Pisonia and Boerhavia
contribute to the floral diversity of islets from 0.2 ha to 108 ha.

Sixth, Caroline does not have an assemblage of nonstrand
plants that only occur on larger motus; the only naturally
occurring, nonstrand plant is Psilotum.

Seventh, the greatest factor complicating our understanding
of Kapingamarangi's natural evolutionary processes is the
presence of numerous exotics: of its 98 vascular plants, only
38 (39%) are indigenous. Its exotics include numerous weedy
herbs and food plants, which occupy gardens, abandoned
house sites, taro patches, and plantations (Cocos, Pandanus,
Artocarpus). These man-made habitats are particularly

25

prevalent on larger islands. Such an abundance of exotics, both
in species and area covered, renders a discussion of natural
processes on Kapingamarangi almost impossible. Undisturbed
habitats such as those on most of Caroline's motus, and on
other uninhabited Pacific islands whose quota of indigenous
plants exceeds 75%, provide far better data on species-area
relationships.

Motu Size in Relation to the Distribution of Trees, Shrubs
and Herbs: As one progresses from small to large islets
(Table 5), the number of tree species rises from 0 to 7, the
number of shrubs from 1 to 4, and the number of herbs from
2 to 12. Caroline's trends are similar to those at Aitutaki
(Stoddart & Gibbs, 1975), where the numbers of trees and
shrubs are relatively constant over a wide range of motu sizes
(3.8-7 1 ha), while the number of herbs shows a slight increase.
There are too many recent exotics on Kapingamarangi for
comparisons to be valid. We believe that if Niering's data were
reanalyzed, using only indigenous species, similar
generalizations would be found, viz.: most species on atolls
establish rapidly on small motus, after which a few additions
occur on motus of increasing size until the maximum number
of potentially available species is reached. Cursory examination
of Niering's Fig. 31, detailing the breakdown of total species
numbers into indigenous and nonindigenous components, bears
out this hypothesis.

Plant Communities

General Account

The total area covered by vegetation on Caroline is
357.55 ha, fully 90% of the combined areas of all the motus. Of
this, two-thirds (289.82 ha) is woodland. Substantial areas of
Caroline's native woodlands and herb mats are pristine, and
85% (possibly as high as 93%) of its plant species are indigenous.
Twenty-three (60%) of its 39 motus harbor wholly indigenous
vegetation (Figs. 27-30). Atolls that support substantial areas
of native forest are typically remote and uninhabited. Where
people are present, native vegetation is usually confined to the
smallest motus or the extremities of larger ones — areas with
marginal human usefulness.

Typical of atolls, but unusual for the tropics, are monotypic
stands of shrubs and trees. Caroline is rich in such woodlands
(Figs. 18,22,24.27-30; Table 9). The present vascular flora of
Caroline, 27 species, is organized into 7 plant communities
( 1 1 subcommunities) defined principally by dominant species
(Fosberg, 1953. 1977a). Eight subcommunities are natural,
three are anthropogenic (Table 5 ). The subcommunities include
a mix of dominant species, which are discussed in the major
Community sections below.

NATURAL COMMUNITIES:
Natural Herb Mat
Toumefortla Scrub and Forest
Beach Scrub with Suriana
( 'ordia Forest
Paiulamis forest
Pisonia Forest

ANTHROPOGENIC COMMUNITY:
Coconut Woodlands

Natural Herb Mat (67.73 ha) (Figs. 19.23: Pis. 20.33,
34.45,46,47)

Widespread and predictable on wind- and salt-blown
coastal coral rubble and incipient motus, these mats are
composed primarily of Heliotropium and Portulaca. They are
pioneers on newly emergent motus, cover most of the ground
area of small motus, extend inland along ancient reef channels,
and typify newly evolving land that connects or augments
established islets. Natural herb mats may persist through all
five stages of plant succession as long as sunny openings occur.
Caroline's motus illustrate two general principles: /. the
smaller the area of an islet, the more extreme is the strand
character of its vegetation, and its corollary; 2. as islet areas
enlarge, strand flora becomes less important (Fosberg, 1949).

The following species are present (see Table 2 for
abundance indices):

Trees: Morinda citrifolia (one drift seedling on one motu);
Shrubs: Tournefortia argentea, Suriana maritima, and
Scaevola sericea; and

Herbs: Heliotropium anomalum, Portulaca lutea, Boerhavia
repens, Lepturus repens, Laportea ruder alls, Lepidium
bidentatum, and Ipomoea macrantha.

Near the high water mark, the herb mats are recumbent,
leathery, and somewhat desiccated. As environmental
conditions improve further inland, they spread more laterally
and average up to 7 cm in height. Their rubbly habitat, often
sprinkled with Tournefortia, resembles a low savannah.
Although these prostrate herbs can tolerate dazzling sunshine,
they grow most vigorously when slight shade, and hence a
higher relative humidity, is present. Under these conditions
they may attain a height of 22 cm and form a fairly thick mat.
With too much shade the mats disappear or their species
proportions and abundance changes according to the presence
or absence of sunny clearings. Thus, natural herb mats may be
found in patchy clearings within forests up to 1 3 m tall. They
are common in the abandoned Cocos plantations of South
Island, where Boerhavia tends to proliferate into thick mats
that completely cover the substrate, vying with Phymatosorus
and Ipomoea for"lebensraum" (PI. 34). A thick, exposed mat
of succulent herbs is found on Noddy Rock, where Portulaca
is the primary component.

Herb mats occurred on almost every transect, windward
and leeward, ranging from 1% to 60% coverage (Figs. 19,23).
The most extensive areas (coverage 35-50%) were on Skull.
Tridacna, South (Trs. 1.4,6). Emerald, and Mannikiba. Mats
predominated in sparsely vegetated areas. Their widths varied
according to the age, shape, exposure, and geographic position
of the motu but were widest on seaward-facing shores
(Table 7).

Wide bands of herb mats may encircle an entire motu: to
windward they average 36 m (Table 7), while, bordering the
relatively placid and intermittently shaded lagoon, they shrink
to a mere 0.9 m. On leeward motus, the corresponding figures
are 18.5 m and 4.2 m.

26

Although reef flats are typically wider wherever an islet
turns sharply, it is not unexpected that these perimeter bands
are the most extensive on the extremely exposed shores of
northern Nake( PI. 17)andsouthernSouthIsland. On the latter,
they are up to 59 m wide. Similarly, on small exposed motus
(e.g.. Skull. Noddy Rock), they carpet most of the area
(Fig. 27). Under such conditions, Portulaca and Boerhavia
develop much redder stems, possibly due to the presence of a
chemical "sunscreen."

Associations with Birds: Whether bordering the edges of
established islands or composing the entire ground cover of
tiny motus and ancient reef channels, herb mats are nesting
sites for red-tailed tropicbirds, masked and brown boobies,
sooty terns, and brown noddies. Herb mats are often used as
foraging grounds for shorebirds.

Beach Scrub with Sunana ( 1 .49 ha) (Fig. 20; Pis. 6,24,39,44)
Uncommon on Caroline, beach scrub with Suriana is
typically foundon sandorsandy rubble bordering Tournefortia
or Cocos. On Caroline, it is evidently limited by the paucity of
low-lying sand and gravel sheets, with which it is normally
associated elsewhere (Fosberg, 1953; Wiens, 1962;Stoddart&
Gibbs, 1975).

The following species are present (see Table 2 for
abundance indices):

Shrubs: Tournefortia argentea, Suriana maritima; and
Herbs: Heliotropium anomalum, Boerhavia repens,
Portulaca lutea, Laportca ruderalis, Phymatosorus
scolopendria, and Lepturus repens.

This plant community was found on 10 motus (Fig. 20),
either in thick bands or as scattered shrubs. Suriana is most
robust on sandy substrates, especially fringing the lower lagoon
on South Island (Fig. 36; Pis. 6,24,39) and on windward
Tridacna. The fringe, repeatedly interrupted by other species,
grows to 1 2 m wide and 1.8 m high. Here the shrubs are closely
appressed and slightly entangled, forming dense shade, which
supports a sparse understory. On South, where its roots are
submerged at high tide, it is being shaded out by overhanging
Cocos (compare Pis. 39 and 40), having retreated since 1965.
Suriana also occurs as scattered individuals or in open bands in
coarse rubble. Beach strand up to 60 m wide, containing herb
mats, Tournefortia, and scattered Suriana, were found on
South (Tr. 1, PI. 21 ). Long (Tr. C), Brothers. Matawa, Long,
and the Southern Leeward Islets.

Pandanus Forest (3.38 ha [this figure is pure Pandanus forest.
Mixed forests containing Pandanus account for a further
14.96 ha]) (Fig. 11; Pis. 18.35-38)

Although several species of Pandanus are native to the
Line Islands, and their seeds are common components of
Pacific sea-drift (Ridley, 1930; Stone, 1968), it is probable that
the groves of P. tectohus on Caroline represent both naturally
established forests and cultivars transported by early
Polynesians. Its largest acreages are on two islands that
contained historical settlements (Nake, South). However, its
presence within the interior forests of a few motus lead us to
conclude that it may have experienced a dual introduction. On
Emerald Isle, 3.20 ha (38% of the islet) supports a mixed forest

of Tournefortia, Pisonia, and Pandanus. Similarly, Shark
Islet's interior woodlands of Tournefortia, Pisonia, and Cordia
(5.52 ha, 70% of the islet's area) also contain a substantial
amount of Pandanus. The only record of habitation for these
motus was a possible hut on Shark. The occurrence of Pandanus
groves or lone trees on other islets (Fig. 1 1 ) is easily attributable
to drift seedlings. Dried Pandanus seedpods are the most
conspicuous litter along Caroline's lagoon beaches (PL 38); its
seeds last for months in seawater (Guppy, 1906) and are
probably distributed locally by rats and land crabs, as noted
elsewhere (Ridley, 1930). Carpels from Nake's southern
mixed woodlands undoubtedly established the grove on
Pandanus Islet.

The mixed forest with Pandanus on south Nake (with
Cocos, Cordia, Pisonia, and Tournefortia) contains up to 50%
Pandanus attaining heights of 1 2 m (Fig. 37). It does, however,
look disturbed.

Many Pandanus trees were felled on South Island during
the coconut planting era (ca. 1 873- 1 925 ), as we know that they
were "somewhat numerous" in 1834(Bennett, 1840), but only
"one or more of the screw pines were found growing in various
parts of the island" in 1883 (Trelease. 1884). A drawing in this
latter paper (PI. 50) depicts a grove from South Island denser
than any remaining today, where Pandanus is uncommon in
the beach scrub bordering the Cocos plantation.

Trees were fruiting abundantly in September 1988,
especially on Nake. The green phalanges, 17.5-20 cm in
diameter, ripen to yellow and orange when they fall to the
ground. They are eaten by hermit crabs (Coenobita perlatus
[PI. 381).

Tournefortia Scrub and Forest (125.25 ha) (Figs. 24.34;
Pis. 5.8,20,30,47.51)

General Distribution: Characteristic of many Pacific
islands, Tournefortia, a broadleafed evergreen, dominates the
wooded motus of Caroline, forming 40% of its total vegetative
cover (Fig. 24). Its pale foliage and hemispherical canopies (to
14 m tall) typically surround the taller, darker canopies of
Pisonia and Cordia.

A hardy halophyte. Tournefortia occurs on every motu
and in every habitat except pure Pisonia forest. It is tallest,
widest, and lushest on the windward motus, particularly on
those where Pisonia is also best developed. Without direct sun
though, as under dense Pisonia or Cocos, it withers (PI. 24).

On other atolls Tournefortia forms a narrow or interrupted
belt inland of the beach or is a component of mixed scrub
(Fosberg, 1953). However, given the floristic poverty on
Caroline, especially of shrubs and trees, Tournefortia not only
has expanded into niches that might elsewhere be occupied by
combinations of Scaevola. Pemphis, Suriana. Terminalia,
Hernandia, Thespesia, Hibiscus, et cetera, but frequently occurs
in pure stands ( 1 13.03 ha) that extend well inland. It thus
occupies a much higher percentage of the islet areas on Caroline
than on atolls with greater biodiversity. For example, Nake, the
largest islet, has the greatest amount of Tournefortia (79.68 ha)
of any islet: 28.9 ha of pure scrub and forest, 18.28 ha of
"savannah," 17.48 ha with Cordia, 8.99 ha with Pisonia, and
6.03 ha mixed with Cocos, Pandanus, and Pisonia.

27

Overall, we classify Tournefortia as a shrub (Stoddart &
Gibbs, 1975). However, following Mueller-Dombois et al.
( 1 98 1 , p. 58), we also distinguish between its shrub (scrub) and
tree communities. Because they intergrade. sometimes we
lump them together ( vegetation maps and schematic profiles of
the motus) and at other times treat them separately
(Tables 2.5, 1 0 and ecological discussions |:
/. Tournefortia Scrub: <5 m high (x = 2 m). <60% canopy
coverage (Pis. 20,30,33,47 (.This open scrub growth is typically
confined to islet perimeters or emergent reef channels and
covers much of the vegetated rubble on smaller islets. Its
species composition is similar to that of the taller forest, except
that herbs are more prominent.

2. Tournefortia Forest: >5 m high ("x = 8 m), >6C7f canopy
coverage (PI. 48). This taller, closed forest, with maximum
height 1 5 m. develops as a second belt of woody vegetation
approaching the interior of the larger islets. Figure 34 depicts
a schematic profile through pure Tournefortia scrub and forest,
while Fig. 35 diagrams a profile of a larger islet where
Tournefortia is represented only on its periphery.

Species Diversity in Tournefortia Woodlands: The
following species occur in both scrub and forest. Those marked
"*" occur primarily in the scrubland (Table 2).
Trees: Pisonia grandis, Morinda citrifolia, Pandanus
tectorius. Cocas nucifera, Cordia subcordata;
Shrubs: Suriana maritima, Tournefortia argentea, Scaevola
sericea. Species A;

Herbs: *Heliotropium anomalum, *Boerhavia repens,
*Portulaca lutea, *Lepturus repens, *Laportea ruderalis,
*Achyranthes canescens, Phymatosorus scolopendria,
Ipomoea macrantha.

Caroline's tallest Tournefortia stands (12-15 m) occur on
Nake. On all other windward motus, the Tournefortia canopies
vary between 6 and 9 m tall, shorter than expected if their
forests were virgin. This has historical significance: we do not
know the extent of forest felling (if any ) on the Windward Islets
(Crescent through Tridacna) during the guano era, but we do
know that 4.587 coconut palms were planted during 1 9 1 9-20,
and that "misses" (dead seedlings) were fastidiously replaced
over the following 2 years (Young, ca. 1922). Thus, their
forests, though weed-free today, comprise secondary growth
around 60 years old. It is not surprising that Achyranthei
canescens and Lepturus repens, both weedy (though
indigenous), are particularly common inland on some windward
motus (Figs. 13.16). Tournefortia' s rapid recovery illustrates
that ecosystems in the pioneer stage generally recover their
original condition rapidly when left alone (Fosberg. 19S3).

Stature and Area Coverage: Forming an umbrellalike
canopy, a typical Tournefortia forest is very simple. Its twisted
branches ami gnarled trunks stretch about untidily over an open
understory. The lower branches die off as the trees increase in
stature. Sometimes a scant coverofherbs develops in restricted
pockets of better soil, such as a clearing where a dead tree fell,
a semishaded spot beneath a colony of seabirds, or a site where
a storm deposited a leu dead fish.

Tournefortia is abundant throughout the atoll. Areas with
90-10095 canopy cover were found on Nake (Tr. 4). Long
(Trs. B.C.4.6. 1 0. 1 2 ). North Pig, Pig. North Brothers, Brothers.

Crescent, Arundel (Fig. 34). Tridacna (Trs. 1.2). South
(Trs. 1,4), all 5 Southern Leeward Islets, all Central Leewards
over 0.5 ha, and Pandanus Islet. Tournefortia is present across
the entire width of some small motus — for example. Fishball
(144 in wide). Even on larger motus such as Mannikiba
(280 m wide). Tournefortia blankets nearly all the land ( PI. 70).
Long (75.98 ha) is a composite motu: long, narrow, and
derived from the coalescence of at least five former islets.
Because 7barn€/brria encircled the perimeters of these ancient
islets, it is now present in five sets of concentric circles.
connected by herb mats, down the length of the island ( Fig. 39).

In the herb mats, Tournefortia is small (x = 1.4 m) and
widely scattered (Table 10). It may be of typical hemispherical
shape or irregularly windshorn (Pis. 13.45). On windward
coasts they typically form a tight wind barrier one or two trees
thick. Moving inshore from the seaward fringe, the trees
become progressively taller ( x = 6 m) with a more open
understory. Cordia often mixes with Tournefortia, either as
scattered individuals in the understory or canopy, or as small
groves. On the Southern Leeward Islets, such belts border the
seaward scrublands.

Though still widespread in the Pacific. Tournefortia is far
less abundant than formerly. On inhabited islands it exists
primarily in relict patches or as edging around anthropogenic
forests. It rarely covers most of the land area of islets: two
exceptions are TaongK Marshall Islands) and GaferuK Caroline
Islands), both in Micronesia (Fosberg. 1956: Wiens. 1962).
The finest quality Tournefortia forests on Caroline Atoll (15m,
80% cover) occupy central and northern Nake (Fig. 37); given
Caroline's history of occupation, these could well be virgin.
These 15-m Tournefortia compare favorably with 18-m
specimens found at Jemo Island by Fosberg ( 1956). Perhaps
Jemo's trees are at the upper size limit for the species, as
Tournefortia is generally recorded as 3 to 6 m tall (Wiens.
1962).

Ecology: Tournefortia is an integral part of the atoll's
evolution and ecology. Bearing seeds capable of floating for at
least 4 months in the sea (Guppy, 1906), it is the first woody
plant to establish on tiny motus (<0. 1 ha), appearing immediately
after the native herbs have begun to germinate in the coarse
coral rubble. Requiring little or no soil and adequate rainfall,
it can grow up to 2 m a year (Fosberg. 1959). Tournefortia' s
leaves contribute to soil development, paving the way for plant
succession from Stages I through IV. for it only persists in soils
that are conducive to the growth of its mesophytic competitors
(Fosberg. 1953). The most mature trees (x = 9.5 m) occur at the
Tournefortia— Pisonia interface, but die off as Pisonia expands.
When Tournefortia has reached its maximum height, most of
its lower branches have fallen, leafage is reduced, and flowers
and fruits are few. Tournefortia usually drops out after one
generation. Seedlings are rarely seen in heavy shade, and fallen
trees are fairly common on the edge of the interior forests where
Pisonia replaces it.

An example of complete replacement of Tournefortia by
Pisonia is illustrated by nearby Vostok. It has heretofore been
assumed that Vostok's sole tree species was Pisonia grandis
(Fosberg. 1936: Bryan. 1942: Clapp& Sibley, 1971b; Garnett,
1983). However, Young (ca. 1922) stated that when Captain. I.

28

Larsen. of the schooner Papeete, planted 1 00 coconuts there on
3 1 May 1922. he found "Pukatea and Tauhinu trees, et cetera
60 to 80 feet high;" that is, Pisonia grandis and Toumefortia
argentea, but no "Tou" trees (Cordia subcordata). By 1935
only Pisonia remained (Fosberg, 1936); thus, the last natural
Toumefortia must have been eliminated by Pisonia.

Along some coasts (Long, Nake, South). Toumefortia
overhangs the water, its roots immersed at high tide. We found
floating debris up to 20 m inland within dense Toumefortia
forest, indicating that this hardy shrub can withstand periodic
storms and high tides. If a rosette of Toumefortia leaves is
placed in fresh water, it droops within an hour, indicating that
its tissues require a high salt concentration in order to maintain
turgidity (personal observation). Perhaps decreased salinity in
the ground water, coupled with reduced light intensity in
advanced serai stages, contribute to the eventual disappearance
of Toumefortia in the center of coral islands.

Associations with Birds: Toumefortia is a favored roosting
and breeding site for most of Caroline's seabirds. The taller the
trees, the greater the bird diversity they harbor: scrub contained
four species (36%) and forest, nine (82%). Sooty terns nest in
tight colonies in its shade, its canopies support large populations
of red-footed boobies (PI. 51) and great frigatebirds
(Subchapter 1.2, this volume), and its branches are favored by
white terns (Figs. 34-36). Toumefortia leaves provide nesting
material for noddies.

Cordia Forest (1.39 ha) (Fig. 22; PI. 27)

General Distribution: Cordia does not form "the main
native woodland" on Caroline Atoll, as implied by Clapp &
Sibley ( 197 la) and stated by Stoddart & Gibbs (1975, p. 104).
It occupies far less area than Toumefortia or Pisonia (Table 9).
Cordia is generally mixed with other emergents: monotypic
Cordia forest covers only 1.39 ha, while Toumefortia or
Pisonia containing substantial amounts of Cordia total
25 .89 ha. In toto. this is less than 1 0% of Caroline' s woodlands,
and Cordia is usually subdominant. We treat Cordia forest as
a separate plant community because of its increasing rarity on
Pacific atolls, which makes Caroline's groves an increasingly
important resource in need of conservation. Cordia forest
occurs primarily on Nake, Windward. Crescent, North Pig,
Pig, Danger, Shark, and the Southern Leeward Islets.

History: Bennett (1840) recorded "two species of
Toumefortia" on Caroline, possibly referring to Toumefortia
and Cordia. There are no other 19th-century records. From
Cordia 's present distribution we can infer that it was formerly
more extensive on South and Nake. Scattered trees within and
bordering the Coeos plantations suggest that its history is
similar to the species on Flint: both Flint and Caroline were
worked simultaneously by the same companies for guano
( 1872-1890) and copra (into the 193()'s). Pisonia and Cordia
forests were felled to make room for coconuts. Several
hundred Cordia logs were exported from Flint to San Francisco
to be used for furniture. The last logs were exported in 1896,
6 years after the guano supplies were depleted, but coconuts
were still being planted (Young, ca. 1922). Today, the belt of
indigenous vegetation bordering Flint's coconut plantation
still has many large Cordia trees (Kepler, 1990b), unlike

Caroline, where today Cordia is rare on South Island. However,
some of Flint's Cordia trees today may well be those "few tiny,
struggling. ..trees. ..recently planted" (St. John & Fosberg, 1937).

Abundance and Distribution: Cordia seeds are dispersed
by ocean currents and can germinate after 40 days in seawater
(Guppy. 1906). Requiring the presence of other species
(Fosberg, 1953), on Caroline it develops both as an understory
shrub and forest emergent (to 15 m high). It typically occupies
the woodland periphery, occurring in small circular or linear
groves, or mixing with Toumefortia and/or Pisonia (Table 5).
On many other Pacific atolls Cordia forms a mixed scrub with
Scaevola (Fosberg, 1949). Cordia may form tall, straight-
trunked trees (PI. 27) or sprawl like Hibiscus tiliaceus. In dry
rubble sites it may become chlorotic ( PI. 79 ) or semideciduous.
The tallest groves are on Pig (PI. 27), where six trees averaged
12.6 m tall, 1 16 cm circumference at 1.5 m high, and 99.8 cm
around the trunk base. Lush Cordia groves sheltered parts of
the upper lagoon on Long Island (Tr. 10).

Flowering times are unpredictable: In November 1989,
flowers were abundant, extending through March, yet in
November 1990 not one flower was observed (Anne Falconer
and AK.K, personal observation). Two flowers were seen in
September 1988 (personal observation).

Associations with Birds: Black and brown noddies,
frigatebirds, and white terns nest in Cordia wherever it is a
forest component. Great frigatebirds and red-footed boobies
favor roosting in the lush, lagoonside forest of Cordia and
Pisonia near the south end of Long Island.

Pisonia Forest (62.17 ha) (Figs. 18,32,33,35,39,41;
Pis. 43,52,53; Tables 11-14)

General Distribution: Although Pisonia grandis is recorded
as "present" in the two previous scientific accounts of Caroline
(Trelease, 1884; Clapp & Sibley, 197 la), the quality and extent
of its forests has not been recognized. Some stands on this atoll
are outstanding representatives of a major ecosystem that was
formerly far more widespread in the Pacific.

Common throughout the atoll, Pisonia occurs on 29 motus.
Well developed groves, 10-21 m tall and up to 359 cm
circumference at 1.5 m, are present on 23 of these (Table 1 1 ).
Although present on motus less than one hectare in size
(Table 5). it typically occupies interior forests (schematic
profile. Fig. 35), with individual trees or groves contributing
from 5% to 100% of the canopy. In general, Caroline's
windward motus support the lushest forests: the maximum
height of windward Pisonia forests is 2 1 m; of leeward forests,
15 m.

Mature Pisonia forests are monocultures of grandeur. The
trees bear one to several stout boles of irregular shape, whose
rotting cavities often harbor large coconut crabs or miniponds
alive with mosquito larvae. Their scraggly branches
occasionally bend over and reroot. It is dark and humid after
the glare of the beach. Walking is easy because the forest floor
is open except for exposed roots and a scattering of broken
branches. Few seedlings occur. Overhead, aconstant cacophony
of bird calls overwhelms the sound of the trade winds, and
guano spills everywhere. Polynesian rats scurry underfoot. It
is a curious habitat for a tropical island.

29

In September 1988 we saw no flowers or fruit. Anne
Falconer reported flowers on Motu Ana- Ana in August 1990.
Pisonia was beginning to bloom on Vostok in March 1990
(A. Kepler, in prep.).

A Historical Perspective: Some of Caroline's most mature
Pisonia groves (to 21 m tall, 660 cm circumference at 1.5 m.
multiple trunks) appear to be virgin. Overall dimensions of the
trees, low species diversity, and general character of the plant
community are similar to the known virgin groves on Vostok
(personal observation, Table 12). We do not have dimensional
data (other than height) of these particular areas on Nake, but
their level of maturity can be seen in PI. 43.

Despite the advanced stage of ecological succession of
many groves, especially to windward, planting records from
1916-1922 indicate that Cocos was planted throughout, not
only South, but also on Nake. Long, and all the major Windward
Islets (Young, ca. 1922). Given the standard planting density
of one tree every 8.5 nr (28 x 28 ft) ( Young, ca. 1922), we have
calculated the approximate area on each islet given over to
Cocos plantations, based on the number of coconuts planted
(Table 13) times the area required for each tree (73 nr). We
have then compared this to the usable areas based on today's
forest cover. On all nine Windward Islets, Cocos covered
79 to 100% of usable ground; in several cases the amount
calculated for Cocos exceeds the amount of potentially usable
ground. Thus, Cocos was so intensively cultivated on the
Windward Islets that essentially all the Pisonia and most
Toumefortia forests must have been felled.

Two remarkable points emerge from Table 13: 1. Scarcely
any Cocos remains today on the nine Windward Islets; seven
of them bear no trace of the former plantations
(Figs. 43,44,47,48); and 2. recovery of Caroline's natural
ecosystems. Stages I through V ( Ecological Succession section ).
on the windward side has been rapid and, at least on Brothers
Islet (Fig. 46), reasonably complete with regard to ecological
succession and species diversity. Today the Windward Islets
have the lushest and tallest plant communities, with higher
species diversity than the leeward islets (Table 3). which have
evidently experienced far less human disturbance.

This differential disturbance on the w indward and leeward
sides of the atoll explains enigmas such as 20-m-tall Pisonia
forest on the leeward Booby Islet (0.84 ha), taller than most of
the windward forests; the absence of Pisonia on windward
Tridacna Islet (9.08 ha), which, being close to South Island,
probably supported Cocos, which was managed longer than the
more distant windward islets; and the patchy distribution of
Pisonia in the interiorof several islets (e.g.. Windward, Arundel).
This last point also applies to Mannikiba ( 2 1 .49 ha), the largest
leeward islet. According to Young (ca. 1922), 6,000 seed sets
were brought from Flint to Caroline in 1920 and kept on
Mannikiba. This "nursery stock" was used to replant "misses"
on other islets, due mostly to destruction by coconut crabs and
poor planting. Today. Mannikiba's total acreage of Pisonia
(Fig. 53) is very small and fragmented relative to the islet's
si/e: 1.13 ha. 5'< of the total land area. Compare this with Bird
Islet (Fig. 55). which, as far as we know, has never been
disturbed: 1.70 ha Pisonia, 42' i of the islet's land area.

On both Caroline and Flint there is much variation in the
quality of the regenerated Pisonia forests (Table 12). Some
trees bear enormous, partly rotting boles, black algae smothering
the bark, multiple trunks, and few or no understory herbs.
Other trees are tall and straight-trunked, with characteristic
whitish bark, and bear no rotting holes in their bases. Such
observations suggest that when their indigenous forests were
felled, only minimal cutting was done, and many Pisonias were
able to regenerate quickly by sprouting from rooted stumps and
fallen branches. This speculation is supported by the fact that
some of Vostok's Pisonia trees regenerated similarly. Maude
(1953, p. 96) stated that "there is room for 8,000 palms on
Vostok, but only 100 have been planted and most of these have
been choked in the luxuriant 'buka' (Pisonia grandis) forest:
no attempt having been made to exploit the island since the
initial planting."

Pisonia, a soft, pulpy wood, has a well-known ability to
sprout or send up suckers from dismembered branches or fallen
trunks ( Fosberg, 1 953), and it has been noted that older trees are
virtually indestructible, fire being the only effective means of
clearing forests (Wiens, 1962, p. 397). The senior author has
photographed leaf sprouts from partly burned twigs as small as
1 m long and 5 to 6 cm in diameter.

Since the existing Cocos plantations on South Island and
southwest Nake contain few Pisonias, it seems that forest
clearing was more thorough on the atoll's larger islets than on
the smaller ones, which today manifest scant traces of their
former history. Fortunately for Caroline, its coconut plantations
were plagued by a number of problems, which resulted in their
being abandoned twice: coconut crabs, seabirds, rats,
Ipomoea vines, and an unknown disease (see under Coconut
Woodlands, this section).

A footnote in Young (ca. 1922, p. 15) stated that "the larger
portion of the 30,000 trees planted were either badly planted or
smitten with some disease as in 1927 it was reported by
Mr. Bunckley that most of them had perished." In 1929 only
1 3,2 1 5 trees were left and more were being planted. Considering
the distribution of both palms and natural forests today, it
appears that plantations continued on South and Nake and were
abandoned on the smaller islets, allowing for a better recovery
than might be expected had the Cocos grown to maturity.

Once a Cocos plantation has been well established and
subsequently abandoned, Pisonia regrowth is more difficult.
This is characteristic of mans islands in the tropics. For
example, on Cousin Island (an ICBP wildlife preserve since
1968, Seychelles Islands. Indian Ocean), where Pisonia is
currently reestablishing within adeterioratingCf)cc.v plantation,
Phillips & Phillips ( 1990. p. 37) envisioned "centuries rather
than decades before something like a natural ecosystem
develops." We predict a similar time frame for areas on
Caroline and Flint where Cocos canopy is over 7095 cover.

Annual Growth Rates: Data on Pisonia grandis growth
rales are evidently lacking. On Cousin Island, vegetation
changes, including Pisonia and Cocos, have been monitored
since 1 474 (Phillips. 1984; Phillips & Phillips. 1990). However.
as in) measurements of any tree dimensions are included,
growth rates cannot be deduced.

30

Because of this paucity of data on Pisonia, and because its
forests have diminished significantly on coral islands this
century, we are presenting our data from Caroline, with
comparisons with Vostok and Flint, in the hopes that it might
inspire more research.

One point is clear: on all three of the Southern Line
Islands, Pisonia grandis has recovered fast from disturbance
(except for total forest elimination), reaching close to its
maximum height and ecological maturity in 70 years or less.
Mature Pisonia, under optimal conditions of soil, temperature,
and rainfall, may attain 35 m, as on Fanning and Washington
(Garnett, 1983 and personal communication). However, in the
Southern Line Islands, canopies of similarly virgin Pisonia on
Vostok rarely exceed 25-30 m tall (Kepler, 1990c).

Caroline's prime grove — 21 m tall, with circumferences
(at 1.5 m) to 660 cm, and bearing multiple trunks and root
suckers — we now know date back only to the 1920's. They
have thus averaged a growth rate of 0.32 m per year since, say,
1925 (65 years). Growth was undoubtedly fastest during the
first few years.

Further evidence of fast growth rates is provided from
Flint Island. In 1934 only one small Pisonia was recorded
(St. John&Fosberg, 1937). Fosberg( personal communication)
recalled that only scant traces of native vegetation existed at the
time, virtually the entire island (324 ha) being planted with
Cocos. In 1990, transect surveys (Kepler, 1990b,d), coupled
with an analysis of aerial photographs, revealed that single to
multi-trunked Pisonia trees, now quite common on the windward
side of Flint, attained maximum heights of 30 m, having
circumferences at the base and at 1 .5 m of 1 ,000 cm and 200 cm,
respectively (Table 12). These compare favorably with one
large Pisonia, presumably virgin, measured on Atafu Island
(Tokelau) by the US Exploring Expedition in 1 840, which was
more than 600 cm in circumference at its base and about 12 m
tall (Wilkes, 1845, Vol. V, p. 9). Furthermore, indigenous
forests (Pisonia, Cordia, Guettarda), with canopies of
4 to >20 m, covered 57 ha, 28% of Flint' s vegetated area. Thus,
numerous Pisonias have not only established themselves since
the plantation was abandoned in the late 1930's but have
averaged approximately 0.5 m growth per year. This faster
growth rate than on Caroline may be due to Flint's higher
rainfall and greater relative humidity due to the presence of a
more successful coconut plantation inland: Caroline's annual
output of copra was 15 tons, compared to 230 tons for Flint
(Young, ca. 1922; Maude, 1953).

Species Diversity in Pisonia Forests: Caroline's motus
harbor every stage in the development of a Pisonia forest, from
stately monoty pic groves to a single tree. The plant communities
between these extremes harbor the greatest species diversity
and most luxuriant growth on the atoll. The following species
are present (Table 2):

Trees: Morinda citrifolia, Cordia subcordata, Cocos
nucifera, Pandamis tectorius, Pisonia grandis;
Shrubs: Tournefortia argentea; and

Herbs: Boerhavia repens, Portulaca lutea, Laportea
ruderalis, Lepturus repens, Achyranthes canescens,
Phymatosorus scolopendria, and Ipomoea macrantha.

The number of species within Pisonia forests ranges from
1 to 14 (Table 14). As Pisonia becomes more dominant, their
trees are taller (21 m), and species diversity is less (Table 14).
Here, the average number of species is 3.4. Species diversity
is also very low at the other extreme of Pisonia development:
in one young motu (Azure), only a single 6-m-tall Pisonia tree
is present (x = 4.0 m). The smallest islet on which we found
Pisonia, Azure is only 0.20 ha in area and 77 m wide (Fig. 55,
PI. 53); more than half of it is rubble. The width of its scrub is
only 38 m. Along a transect within the majestic Pisonia grove
( 100% canopy cover) on Brothers (Fig. 46), we found no other
plant species, an extreme case of the barrenness of Pisonia
understory. This grove, 13mtallandextending42mfromeast
to west, was sharply delineated from the 6-m-high Tournefortia
forests on both sides and provides a striking example of
complete ecological succession since its Cocos plantation days
of the 1920's.

The highest species diversity occurred with mixed co-
dominants (Tournefortia, Cordia), and Pisonia coverage
25-50% (Table 14). Here, the average number of species was
6.2 (range 3-10). Regardless of the area or width of the motu
on which they occurred, these mixed stands (x = 7 m tall) were
always shorter than pure Pisonia forest.

Ecology: On Caroline, most plant species are established
early in the evolution of individual motus, increasing
in abundance and stature while the motus are quite small.
Pisonia typifies this pattern: single trees occur on 2 motus
whose areas are only 0.2 ha (Table 6). This suggests that
Pisonia is partly salt-tolerant, at least in its early growth stages.
In general, however, motus less than 0.7 ha on Caroline have
little Pisonia (Table 6). It is difficult to imagine a freshwater
lens on Motu Nautonga (1 ha), where an 11-m-tall Pisonia
forest is found (Table 1 1). Further evidence for the salt-tolerant
nature of Pisonia comes from Vostok, where a Pisonia
forest, the sole woodland, extends to the edge of the shoreline
rubble and herb mat. The trees, tightly pruned by wind and salt,
have no buffer of coastal scrub. During storms, seawater
reaches Vostok' s interior forest, yet this 24-ha island supports
one of the largest and tallest (25 m high) groves in the Pacific
(Clapp & Sibley, 1971b; Fosberg, 1977b and personal
observation).

Many Pisonia trees were heavily infested with scale
insects (Coccidae) and Neuropteran larvae (Chrysopa sp.),
identified by Dr. Scott Miller (Bishop Museum, Honolulu,
Hawaii). This appears to be a natural phenomenon, as they
were also abundant on the virgin Pisonia forests on Vostok and
also on secondary Pisonias at Flint.

Relationships Between Pisonia Forest Height and Motu
Dimensions: Contrary to expectations, the tallest, most mature
forests did not all occur on the largest motus (Table 1 1 ). The
three prime forests (90-100% canopy cover) are on Nake
(107.46 ha), Pig (7.21 ha), and Booby (0.84 ha). Trees on
Booby are smaller in girth than those on Nake and Pig, but their
height (20 m) is impressive; as far as we know, Booby was
never cleared. Fine forests occur on other small, undisturbed
motus; for example, Pisonia grew to 14 m on Raurau (3.48 ha)
and to 1 1 m on Kimoa ( 1 .80 ha).

31

A positive correlation exists between Pisonia height and
island width (Fig. 32). On Caroline, motus were 90 m wide
before closed canopies of 13 m developed (Fig. 32), and tree
height increased to 21 m with islet width up to 200 m (Pig,
topmost star in Fig. 32 ). Further increases in islet width did not
result in taller trees. However, even on motus with sufficient
w idth, Pisonia did not develop unless other environmental
conditions were suitable. For example, on Long, Pisonia only
occurred in the centers of its former islets, not in the scrubby
areas where coalescence is more recent. Tridacna. seemingly
excellent for Pisonia, has not yet recovered from its Cocos
plantations.

Pisonia-Seabud Relationships: Seabirds are an integral
part of Pisonia ecology. Its sticky seed capsules adhere to the
feathers of, and are thus dispersed by, seabirds such as terns,
boobies, and frigatebirds; thus, its early appearance on small
motus is not surprising.

On Caroline, six species of seabirds nest in its branches,
dropping considerable guano to the ground below. Black
noddies, amassing in dense colonies, nest almost exclusively in
Pisonia, along with brown noddies, white terns, great and
lesser frigatebirds, and red-footed boobies. Pig Islet, with
7.25 ha of excellent Pisonia forest, supported a dense colony of
nearly 2,000 pairs of black noddies (Subchapter 1.2. this
volume). Bristle-thighed curlews feed on the ground beneath
its open understory. and the long-tailed cuckoo forages within
ils canopy.

Seabirds may be so much a part of Pisonia ecology that a
debate exists as to whether Pisonia actually requires guano for
successful germination and establishment of seedlings (Shaw,
1952; Fosberg. 1953;Wiens. 1962). Very high phosphate and
nitrogen levels arc associated with mature Pisonia, and
concurrently the development of Pisonia forest results in
great!) modified soils that perpetuate its existence (Wiens,
1962; Spicer & Newbery. 1979). The formation of a highly
acid raw humus on the surface of the ground, sometimes in
association with phosphatic hardpan. has also been documented
on several atolls by Fosberg ( 1953. 1956). including Vostok
(AKK and John Phillips, personal observation). Perhaps
Pisonia's present distribution, primarily restricted to uninhabited
islands ( Shaw, 1 952; Wiens. 1 962 l.isin part due to the fact that
its primary seed carriers, seabirds. rarely coexist for long with
num.

Remnant Pisonia Forests in the Pacific: Though naturally
and widely distributed from the western Indian Ocean to the
eastern Pacific (excluding Hawaii), Pisonia grandis has become
increasingly rare this century (Fosberg. 1953 and personal
i ommunication). Occupying the interior of most atolls, it may
have formerly covered the greatest area of any tree species in
the Pacific (Wiens, 1962). Shaw (1952), summarizing its
distribution, stated that it only occurs on remote, generally
uninhabited islands ranging from the western Indian Ocean to
the eastern Pacific, including Malaysia. However, more recent
studies, particularly bv Fosberg, indicate that because its habitat
occupies, and is in part responsible for. the most fertile areas of
inhabited islands, its formerly extensive forests have been
largely replaced by coconuts. Though Pisonia's soli wood is

of little use to either atoll inhabitants or to the timber industry,
its soils were rich sources of phosphate fertilizer and w ere thus
greatly disturbed during the guano mining era.

One of the most extensive Pisonia stands in the Pacific
( 1 3.5 ha on Vostok) was partly burned in 1977 by members of
a "scientific expedition" (Fosberg. 1977b). The Royal New
Zealand Air Force found it smoldering 3 months later (Fosberg,
1977b, personal communication). In a March 1990 visit to
Vostok, we found that approximately 1.5 ha were completely
cleared ( Kepler. 1 990c ), and a further unknown amount of land
was affected. Other excellent groves exist on Palmyra and
Washington (northern Line Group land on Nikumarorol Phoenix
Group). Flint (Southern Line Group); Christmas (northern
Line Group); Bikar, Jemo, and Ujae (Marshall Islands): and
Aitutaki, Penrhyn. Suwarrow, and Manihiki (Cook Islands)
have relatively small stands. Caroline, with 62.73 ha in Pisonia
forest (36.94 ha in monotypic groves) holds some of the finest
representatives of this ecosystem in the Pacific, even though
much of it is not virgin.

Coconut Woodlands (96.14 ha) (Figs. 14.36: Pis. 18.23.24.
28-30.33.34.37.39.40.44)

General Distribution: Cocos. although present on 15
motus and known historically from another 4. covers significant
areas only on Caroline's 2 largest islets. South and Nake
(Table 13). Individual trees and small groves elsewhere are
drift-derived or remnants of plantings made from 1916-1 920.

The following species occur in habitats containing Cocos
(Table 2):

Trees: Pisonia grandis, Morinda citrifolia, Pandanus
tectorius, Cordia subcordata, Cocos nucifera, Thespesia
populnea, Hibiscus tiliaceus;

Shrubs: Tournefortia argentea, Ximenia americana; and
Herbs: Boerhavia repens, Portulaca lutea, Laportea
ruderalis.Achyranthescanescens, Phymatosorusscolopendria,
Ipomoea macrantha, Lepturus repens, Taccaleontopetaloides,
Psilotum nudum, Phyllanthus amarus, and Sida fallax.

The distribution of Cocos (Fig. 14). in order of decreasing
abundance is as follows: South: Forests old and neglected.
Healthiest palms line the lagoon, currently shading out strip of
native scrub. Nake: Southern forests (50-80% Cocos) healthier,
younger, with more native trees and Pandanus than on South:
grove of about 50 palms on northeast. Long: Range from < I ' i
cover (Tr. C) to dense fringe adjacent to lagoon. Emerald;
Northeast and center-west patches. Mannikiba: Main grove.
northeast: 40 palms. 20 m high, another patch in south centei
Ana- Ana: House site, northeast point. Bird. Blackfin, Brothers.
Nautonga. North Brothers. Pig. Pisonia. Raurau. Shark: lew
trees each, primarily in Tournefortia. Lone Palm: One tree,
central forests.

History: A relatively small coconut grove was planted on
South Island prior to the 16th century by Tuamotuan settlers
(Emory, 1947; Maude, 1968). In 1 606. deQuiros noted -plenty
of palms" and "many cocoa-nuts" (Markham, 1904). Since
then, every visitor has recorded them as they grew, and still
grow . adjacent to the boat "landing." A smaller grove evidently
also existed in the south-southwest portion of South Island

32

(Lucett, 1851). According to Maude (ca. 1938). palms were
also periodically planted — and destroyed — "by whalers and
other chance visitors to the island."

Until Arundel's arrival in 1885, Cocos was basically
confined to this single grove in the northwest sector of South
Island (Maude, ca. 1942a). In 1885, land clearing began, and
from then till 1929, nearly 38,000 palms were planted, 29,480
between 1916 and 1920 and another 7.000 young trees after
1927 to replace thousands that had perished (Young, ca. 1922).
Arundel's initial license gave him the exclusive rights to
occupy Caroline and Flint, planting coconuts and other trees
for 2 1 years, in return for an annual rental of 50 pounds ( Maude,
ca. 1942a). In 1929. 13,215 trees remained, after which no one
has counted them. Our field work and scrutiny of aerial
photographs indicate that far fewer exist today.

Caroline's plantations produced copra periodically from
1 873 to 1934, but never profitably. They suffered greatly from
the atoll's abandonment from 1901-1916. Dying and poorly
planted palms presented continual setbacks ( Young, ca. 1 922),
and in 1878 a hurricane wrought great destruction (N.I.D.,
1943). In addition, plantation managers lamented their poor
productivity due to choking "by undergrowth and Pohue Vine
[said to be Tuumfetta (- Triumfetta) procumbens, most likely
a misidentification of lpomoea mac rant ha], destruction of
inflorescences by great numbers of seabirds which roosted in
the tops and broke off the flowers as they appeared," disease,
and ruination of nuts by Polynesian rats and coconut crabs. As
a result of this, the resident laborers slaughtered many crabs,
and "greatly reduced the numbers of sea birds, who migrated to
unoccupied islets." The rat problem was never resolved and
appears to be the major reason for repeated failure of the
plantations on both Caroline and Flint. Their enormous numbers
and voracious eating habits greatly reduced both the crops of
potentially healthy nuts as well as the volume of dried copra. In
1920, 4,600 were trapped on South Island, and hundreds more
were killed by small terriers introduced specifically to control
them (Young, ca. 1922). Maude (personal communication)
recalls that one terrier still survived in the 1940's. Rats still
abound, especially within coconut groves and Pisonia forests.
Another serious problem was due to coconut crabs digging up
recently planted nuts and also their habit of pinching off young
developing shoots. Evidently after the palms had attained one
year's growth this was no longer a problem (Young, ca. 1922).

Before abandonment (1902 to 1916, and after 1934),
Caroline's plantations were owned by several companies whose
average annual copra output was approximately 14 tons. From
1934 to the 1970's, copra was harvested sporadically by small
parties from Tahiti (Garnett. 1983). but within the last 2
decades it stopped altogether.

Despite the relatively fertile soils of South Island, the
problems in the plantations hampered the establishment of
permanent settlements on Caroline. In the 1930's, Maude
estimated that the atoll could support 400 Gilbertese. increasing
to over 1 .000 "when the island has been fully planted" ( Maude,
ca. 1938). However, colonists were never established, leaving
Caroline "one of the least spoiled islands in the Pacific"

(Stoddart, 1976). As Young's (ca. 1922) unpublished
"Memoranda" indicate, Caroline is not as pristine as it appears;
however, the rapid comeback of many of its natural forests on
the windward side is remarkable (see discussion under Pisonia
Forests, this section).

Distribution and Abundance: We recognize four
subdivisions of the coconut woodlands: Cocos Plantations,
Dying Cocos-lpomoea Plantation, Scattered Groves on Small
Motus, and Mixed Forest with Cocos.

1. Cocos Plantations (34.07 ha)

Palm forests now dominate South Island and southwestern
Nake. Although the planting of Cocos on South altered most
of its original habitats, Nake escaped with less damage: whereas
Cocos covers 77% of the area on South, it takes up only 6% of
Nake (11% including mixed forests). The 60 to 100-year-old
trees form tall, closed canopy woodlands (PI. 24) 21-25 m
high, the customary maximum height recorded for old
plantations (Fosberg, 1953). Figure 51 shows the distribution
and abundance of plant species along a transect running centrally
through the island, while Fig. 36 depicts a schematic profile of
the same swath.

Pure coconut plantations (like all habitats on Caroline)
harbor relatively few species: up to 7 trees, zero to 2 shrubs, and
5-11 herbs. The ground vegetation and shrub layers are
composed almost exclusively of indigenous species, an unusual
feature. However, skirting the edge of South Island's lagoon,
tall palms overhang the water and crowd out native plants;
there were considerably fewer Suriana and Tournefortia in
1988 (PI. 29) than in 1965 (PL 40).

2. Dving Cocos-lpomoea Plantation (53.92 ha)

Mature plantations characteristically become overgrown
with shrubs and vines (Fosberg, 1953. 1956). lpomoea
macrantha. the sole vine on Caroline, forms tangled,
impenetrable thickets. Indigenous, nonparasitic, and widely
dispersed by ocean currents, it occurs naturally in small numbers
in natural habitats on Caroline, but grows rampantly in disturbed
areas. These vine-covered coconut woodlands cover two-
thirds of South Island's interior (Fig. 50). The Dying Cocos-
lpomoea forest is moribund. It is bordered by a belt of living
palms, which in turn are sheltered by a narrow rim of indigenous
vegetation (Figs. 36.51).

While surveying the South Island transects, the authors
stomped over intertwining thickets up to 3 m high (PI. 7) and
crawled through tightly-knit masses of vines descending from
the crowns of old palms, Pisonia, and Morinda bushes, until
this too, proved impenetrable. In sunny clearings dotted with
dead or dying palms, lpomoea, Boerhavia, and Phymatosorus
proliferated luxuriantly, lpomoea, one of the prime reasons for
the double abandonment of copra production, will continue to
destroy the coconuts, encouraging natural ecological succession
to begin anew.

3. Scattered Groves on Small Motus (0.82 ha)
Drift-derived palms were observed as long ago as 1834

(Bennett, 1840). In 1916, when planting operations were
commenced after a break of 14 years, about 40 trees grew
beyond the plantations (Maude, ca. 1942a). Today, small

33

Cocos groves, up to 50 palms, drift-derived and plantation
remnants, generally close to the shoreline (Pis. 29, 30), occur
on 1 1 motus.
4. Mixed Forest with Cocos (6.24 ha)

This forest type is a simplified version of more complex
and varied mixed forests that occur on most inhabited atolls.
Composed of both anthropogenic and native elements, it
contains a high proportion of Cocos (50-80%) mingled with
variable proportions of Tournefortia. Pisonia and Pandanus.
This forest type occurs primarily in southern Nake (Fig. 14),
but also on Emerald, Shark, and southwest Long, where it
mixes with Cordia and Tournefortia.

House Site: A single clearing on Motu Ana-Ana,
approximately 40 m x 70 m, contains a few Cocos adjacent to
a vegetable garden and thatched living quarters (PI. 53).

Associations with Seabirds: Coros-dominated habitats
were the most depauperate on Caroline: only brown noddies
and white terns breed there (PI. 54). The noddies nest high
within the frond and inflorescence bases, whereas the white
terns preferred lower sites. The absence of other species
suggests that the anthropogenic Cocos forests seriously inhibit
seabird use and may continue to do so for decades until they are
replaced by native vegetation.

Absent Plant Communities

Caroline's impoverished flora and simple geology has
resulted in a limited variety of ecosystems. The atoll is thus
notable not only for its grand Pisonia forests, extensive
monotypic stands of Tournefortia, and presence of Cordia
groves, but also for the absence of several ecosystems that are
generally considered typical of Pacific atolls:
/. Sesuvium flats;

2. Pemphis, Scaevola, and Sida scrub (two Scaevola plants
are present, and the only two Sida records are from 1884 and
1990);

3. Barringtonia, Calophyllum, Guettarda, Hernandia, and
Ochrosia forests;

4. Plant associations (except Cocos) typical of native cultures
on atolls: breadfruit groves (Artocarpus alt His), taro pits
( Cyrtosperma chamissonis, Colocasia esculenta, Xanthosoma
sagittifolia), cultivated ornamentals (Hibiscus rosa-sinensis.
Plumeria spp., et cetera), or weedy grasslands/wastelands
(Paspalum, Sporobolus. Wedelia, Vigna, et cetera). Even
widespread introduced strand species such as Terminalia
catappa and Casuarina equisetifolia are absent.

In addition, there are no mangroves, peat bogs, marshes,
ponds, salt flats, or other habitats associated with fresh or
brackish water. Poorly represented are:
/. Lepturus grassland. Although Lepturus is present in
coastal herb mats, and occasionally in patches w ithin the forest
understory, it does not form a separate plant community.
However, it may once have covered the extensive clearings on
South Island (PI. 2).

2. Mixed forest. Though 6.24 ha of mixed forest (with
Cocos) occurs (primarily on Nake), it is of such minor
importance to Caroline's overall vegetation that it is treated as
a subsection of coconut woodlands.

Description and Ecology of the Motus

These islet accounts synthesize the history, physiography,
vegetation patterns, ecology, seabird colonies, miscellaneous
biota, and the effects of human activity (if any) on Caroline's
39 motus (Fig. 2). Mapping is based on the coast-to-coast
transects, perimeter surveys, complete surveys (smaller motus),
color transparencies, and aerial photographs.

All motus are detrital reef islets representing many
evolutionary stages from barely emerged coral rubble to large
islets with relatively fertile "soils" supporting lush vegetation.
There is one tiny old reef platform in its final stages of erosion.

We discuss and map them in geographic order beginning
in the north with Nake and progressing down the windward reef
through Long and the 13 Windward Islets to South Island.
Beginning anew in the north, we move south through 7 South
Nake Islets, 1 1 Central Leeward Islets, and finally the
5 Southern Leeward Islets.

Because of the variety of islet shapes, "long" or "length"
refers to the longest dimension lying parallel to the outer reef
edge (normally north-south) and "wide" or "width" to the
longest dimension perpendicular to the outer reef edge (normally
east-west). South Island, the only exception, is considered to
lie adjacent to the southern reef edge, so its "length" is measured
east-west. Seabird numbers are from Subchapter 1 .2. Table 1,
this volume. For convenience in locating particular islets, the
order is as follows:

/. Nake Island (Fig. 37)
2. Long Island (Fig. 38)

indw;

ird Islets

3.

Bo'sun Bird (Fig. 42)

4.

Windward (Fig. 43)

5.

Crescent (Fig. 43)

6.

Atibu (Fig. 43)

7.

North Pig (Fig. 44)

8.

Pig (Fig. 44)

9.

Skull (Fig. 441

10.

North Brothers (Fig. 44)

II.

Brothers (Fig. 44)

12.

Noddy Rock (Fig. 47)

13.

North Arundel (Fig. 47)

14.

Arundel (Fig. 47)

15.

Tridacna (Fig. 48)

16.

South Island (Fig. 50)

South Nake Islets

17.
18.
19.

20.
21
22
23

Pandanus (Fig. 52)
Danger (Fig. 52)
Booby (Fig. 52)
Coral (Fig. 52)
Lone Palm (Fig. 52)
Kota(Fig. 52)
Mouakena (Fig. 52)

34

Central Leeward Islets

24. Mannikiba (Fig. 53)

25. Blackfin (Fig. 54)

26. Matawa (Fig. 54)

27. Emerald (Fig. 54)

28. Shark (Fig. 55)

29. Scarlet Crab (Fig. 55)

30. Nautonga (Fig. 55)

31. Azure (Fig. 55)

32. Reef-flat (Fig. 55)

33. Bird (Fig. 55)

34. Fishball (Fig. 55)

Southern Leeward Islets

35. Raurau(Fig. 57)

36. Eitei (Fig. 57)

37. Pisonia(Fig. 57)

38. Kimoa(Fig. 57)

39. Ana- Ana (Fig. 57)

/. NAKE ISLAND (91.72 ha) (Figs. 30,37; Pis.
35-37,43)

18,23.

History: Nake's large size and underground water lens,
coupled with topography and soils more varied than elsewhere
on Caroline, attracted early Polynesian settlers. Because early
European visitors stayed primarily on South Island, there is
only a single reference to Cocos prior to the late 19th century
(one tree seen in 1825 by Paulding [1831]).

The far northwest of Nake (also called North Island in
Young, ca. 1922) houses the most important archaeological
site on Caroline — a large marae (Figs. 3,37; PI. 36). Discovered
during the guano era, the site is marked as "graves" on Arundel' s
map. Arundel, who was living on the atoll when the marae was
discovered, describes it thus: "On the north-west end of
Caroline are some curious old native remains, whether places
of burial or of sacrifice I cannot determine. I opened one of
these, but could find no indication whatever to guide me in a
decision" (Arundel, 1890). The senior author, R. Falconer, and
G. Wragg located, measured, and photographed this marae in
1990. The entire courtyard was approximately 18 m long by
14 m wide. All 10 peripheral stones and the central one were
easily identifiable from the 1883 plan (Fig. 3), although a few
had fallen over or broken due to encroaching vegetation. The
lower wall, partly destroyed by Arundel, had not been
reconstructed. It is probable that this marae had not been seen
since the 1 880' s; though discussed by Emory (1947), he never
visited Caroline personally.

Northwest Nake is particularly suitable for a place of
worship and sacrifice: it fits most of the environmental criteria
indispensable to ancient Tuamotuan religious ritual (Emory,
1947). First, flat ground was necessary, preferably lying at
right angles to, or parallel to, the lagoon. Second, it was
important to have the wind blowing across the marae to waft
away the smells of sacrificed animals. Third, ceremonial items
included branches of the Pisonia tree, leaves of Cocos (for leaf

charms/" rosaries"), and the aerial roots oiPandanus. Fourth,
feathers from "black terns" (black noddy), frigatebirds, and
red-tailed tropicbirds were also necessary for rituals. Rather
than a smooth substrate, the early Polynesians would have had
to be content with leveled coral rubble and distance from the
lagoon. The only organism not living near the marae today is
the tropicbird; however, their elongated tail feathers could
have been plucked from adults nesting on nearby motus.

Since marae are sacred places, there is possibly a meaning
to the location of the main "courtyard" close to the atoll's
northern tip. Generally the northern extremities of islands were
auspicious places for Polynesians; it is here, they believed, that
disembodied spirits were whisked to the netherworld.

Physiography: Largest in area, Nake is the northernmost
motu, separated from Long by a 40-m channel (PI. 18). With
maximum dimensions 2,000 m long and 685 m wide, it is
basically rectangular with rounded corners and a peninsula-
like extension in the southeast.

Nake lies north of the lagoon, having a southern "bay"
(which we named Sandy Inlet), in which silt, sand, and fine
coral debris are being actively deposited (PL 23). Sandy Inlet,
, a hard, flat expanse of fine lagoon mud and sand, is 145 m wide
at its mouth and extends 200 m north into the main islet. Its
3.50 ha provide a favorite feeding location for shorebirds,
especially bristle-thighed curlews. If Arundel's chart (Fig. 4)
is correct, Sandy Inlet has increased its land area during the last
century.

On the reef flats off the west side are extensive remnants
of jagged upraised reef (PI. 1 1) and occasional beachrock. The
exposed beaches and reef flats at Nake's north point are
especially broad, characteristic of reef flats at the exposed
corners of islands. Comparisons of the northern sweep of
rubble on recent aerial photos with Arundel' s map indicate that
much coral debris, in the form of raised ridges (PI. 17), has been
added since 1883. In the deep, fine coral rubble mixed with
sand east of the marae, the 1990 expedition found three old
turtle nests. Overall, unvegetated coral rubble, mud, and sand
account for only 6% of the land area. In addition, some sparsely
vegetated expanses of hardpan were noted in the south-central
sector, just inland of the coast within a belt of Tournefortia
forest.

Nake's windward coast, complete with a peaked beach
crest and discontinuous beachrock, is 30 m wide in the north,
narrowing to 3 m in the south. Offshore, submerged reef flats
form a sandy moat bordered by a barrier reef upon which waves
pound incessantly.

In the distant past, what we now call Nake consisted of two
separate motus. Aerial photos (Chapter Frontispiece) reveal an
oblique, ancient channel about two-thirds of the way down the
islet. It is now well vegetated in the center but scrubby near the
coastlines.

Vegetation: There are 16 plant species (5 trees, 1 shrub,
10 herbs), 59% of Caroline's flora. Nake is the lushest motu.
Its woodlands (82.39 ha) are about 80% native and 20% with
Cocos (PI. 37). Although in 1916 there were about 260 palms,
and the entire island was evidently planted with 10,544 palms
in 1918-1919(Young,ca. 1922, Table 13), substantial tracts of
each major vegetation type occur today. Its interior is rich in

35

Pisonia, with the largest acreage (20.79 ha) and some of the
tallest trees (20 m high) on the atoll (PI. 43. Table 1 1 ). In
addition, Cordia is well represented: two major groves of
Cordia-Toumefortia forest occupy 11.8 ha, 2% of Nake"s
area. Extensive pioneer herb mats. Hanked on their inner sides
by Tournefortia scrub, occur in the north and east. The
remaining Cocos, essentially in the southern quarter, comprise
Caroline's second largest coconut grove.

Birds: Nake, with 80% of Caroline's breeding seabird
species, shows a direct correlation between islet size and bird
species diversity. Nine species of seabirds breed, all with
larger populations (pairs) than previously reported (Clapp &
Sibley, 1971a): masked booby (105), brown booby ( 1). red-
footed booby (496), great frigatebird (522), lesser frigatebird
1 56), brown noddy ( 390), black noddy (8 14), sooty tern (nesting
in 1989; Anne Falconer, personal communication), and white
tern (1,094).

2. LONG ISLAND (75.98 ha) (Figs. 30,35,38-41;
Pis. 8,13,18,20,28,33,47,58)

Third largest in area, this longest of motus covers nearly
one-third of the atoll's windward side. In the north it is
separated from Nake by a narrow channel; from its southern tip
a chain of smaller motus extends south along the windward
reef.

Physiography and History: Long — 4.226 m long and
330 m wide — is somewhat snake-shaped, with an enlarged
northern "head" and attenuated "tail." From a distance its
vegetation appears as a series of humps. Long has experienced
a fairly complex geological history, noted by the Solar Eclipse
Party: "On some of the islands there are spaces void of
\ egetation, extending from lagoon to beach, which indicate the
existence at a former time of a water separation" (Holden &
Qualtrough. 1884).

At present. Long is composed of five distinct former islets
separated by sparsely vegetated channels of coarse sand and
coral gravel. Aerial photographs also reveal further, older
subdivisions (discussed below). Coalescence and fracturing of
the original motus have probably occurred repeatedly. Since
erosion proceeds faster on an atoll's windward reefs, providing
coral fragments, coralline algae, and pulverized mollusks, il is
no surprise that the first series of Caroline's motus to fuse were
those facing this rich source of parent material.

Long's coarse rubble beaches (Pis. 13.20) are a mirror
image of those on Nake: southward, the) widen progressively.
The swath of unvegetated rubble above high tide line in the
upper two-thirds of Long averages S m wide, while in the lower
third il is 40 m wide. Unvegetated coral debris accounts for
10% of the island's area (Fig. 30). Beachrock, Hanking the
windward shoreline for most of its length, is more abundant
than elsewhere on the atoll (PI. 58).

I ong's lagoon Hank is edged w ith submerged sand and silt
and is line of the most sheltered parts of Caroline. Sand and
rubble deposition off the south point has formed an islet in the
lagoon ( Bo'sun Bird |, w Inch could, in the future, coalesce with
Long's south point to form a hook.

An uncommon substrate on Caroline, upraised reef
( makatea > forms a low rampart ( generally < 1 m high ) paralleling
the ridge crest inside the vegetation for much of the lower
quarter of Long.

In 1 990 Graham Wragg found some scattered large stones,
similar to those of the marae on Nake. located centrally 100 m
north of the southern tip of Long, confirming the report of the
remains of a smaller marae on Long Island (Holden &
Qualtrough, 1884). Wragg noted that the marae was indeed
smaller than that on Nake. w ith dimensions approximately 3 m
wide by 8- 1 0 m long. Its orientation appeared to be northeast-
southwest. The wall on one end was evidently smashed by
storm waves. Only two of the peripheral upright stones were
still standing; they were of similar size to those on Nake. The
platform was in reasonable condition, with a huge Pisonia tree
growing through it. Some rock slabs were large (2 x 2 m). The
entire marae was situated within a Pisonia grove, with some
Cocos but no Pandanus nearby. We do not know if the nearby
coconut grove ( 1 .6 ha) was present before 1 .343 palms ( 20% of
the islet's area) were planted in 1918-19 (Young, ca. 1922).
The sheltered location and a Pisonia— Cocos forest, which
suggests an old clearing, further indicate prior occupation.

In 1990. G. Wragg also uncovered an RNZAF survey
marker just inland of Long's southernmost tip.

Vegetation: There are 15 plant species (4 trees. 2 shrubs.
9 herbs) on Long. 56% of the total flora. Long's variety of
habitats, vegetation heights, substrata, and birds make it the
most diverse islet on Caroline. Only 3% of its area remains in
Cocos. All the atoll's seabirds have bred here. Its ecology is
best understood with reference to Figs. 35 and 39—1 1 .

Within the basic pattern of five coalesced motus. it may be
seen that:

/. From north to south (measured from the midpoint of
each former channel) the motus, of divergent size and shape,
are approximately 320. 620. 700. 1.840. and 100 m long.

2. Each former islet, crowned by a Pisonia forest, contains
concentric rings of decreasing fertility around its core and is
morphologically similar to islets surrounded by water, except
that the coarse coral gravel along the former perimeter is less
marked. More specifically, beach sands and gravel extend for
200-300 m north and south o( the old channels, after which
they increasingly accumulate coral rubble, humus, and guano.

3. The dominant vegetation is Tournefortia, interspersed
with 4 patches of taller Pisonia forest and scattered clumps of
Cocos md Cordia. Interrupted herb mats parallel the windward
coasl and often extend across the island along former channels
(PI. 33). Vegetation height varies from 2 cm to 15 m.

4. Plant species diversity is highest in Tournefortia-
Pisonia and lowest in Pisonia forests.

5. Long's tallest, most mature Pisonia groves (up to L009S
Pisonia) occur on the largest of the former islets. The Pisonia
forest near the south end (Tr. 10). although healthy, is only
12 m tall. This may be due to its impoverished makatea
substrate of pitted reef rock barely covered with "soil." Since
it lies adjacent to Long's most luxuriant Cocos grove, its land
could well have been cleared in 1918-19. with the Pisonia
forest taking longer than elsewhere to recuperate. Because tern

36

guano increases soil fertility and is important for Pisonia
growth (Fosberg. 1953). it is of interest that neither black nor
brown noddies nested here.

ft. Deep dips in Fig. 39 (lower graph) correspond to east-
west corridors formed from old channels. Vegetation in these
relatively infertile, sandy flats is low. similar to that on small
developing motus (i.e.. native herbs with scattered Tournefortia
<2 in high). One sandy channel (Tr. C; PI. 33) supported a
sparse population of Suriana. During the February 1990
cyclone, all vegetation was either uprooted, washed away, or
smothered with fresh sand and coral gravel along Trs. A and C
(personal observation, March 1990). Storm erosion was
particularly marked within the channel that almost bisects the
island (Tr. A).

7. Secondary dips mark even older interislet channels
("ancient channels"), visible on aerial photographs (Chapter
Frontispiece) but barely recognizable in the field. They are
overgrown with Tournefortia and/or Pisonia.

8. Sharp dips within established forests or herb mats
denote relatively recent channels gouged out by storms ("recent
storm cuts"). These were also altered during the winter 1990
storm.

Figures 40 and 41 illustrate some differences between the
windward and leeward coasts. Transect C (Fig. 40) crosses the
north end of Long through an old interislet channel now filled
with sand and rubble. Its low profile reflects the simple habitat
harboring halophytic herbs and Tournefortia shrubs less than
2 m high. Although the shrubs are scattered, the lagoon half of
the transect passes through slightly higher ground, which
encourages denser Tournefortia. This transverse section is
similar to that of a formative motu such as Fishball (Fig. 56).
This exposed, scrubby swath, 300 m wide, harbors red-footed
boobies, great frigatebirds, and a discrete population of masked
boobies. Approximately 127,000 pairs of sooty terns nested in
a similar sandy channel 740 m to the south (Tr. A, PI. 59) in
1988.

Transect 8 (Fig. 4 1 ) crossed the islet nearer the southern tip
(Fig. 8). This profile departs significantly from the usual
parabolic cross-section seen on most of the small motus and
which exists further north on Long Island. From east ( windward )
to west, there is first a wide expanse of coarse, unvegetated
rubble, followed by rubble dotted with herbs, then Tournefortia
scrub increasing to 9 m high. Further inland, a forest of
10-m-high Tournefortia. Pisonia. and Cordia continues
westward to the lagoon. This leeward margin of Long, extending
southward nearly to its tip, is the only location on Caroline
where tall, indigenous vegetation overhangs and shelters the
lagoon. No herb mat is present.

In summary. Long contains examples of all major plant
communities, as well as two minor ecosystems, Pisonia-
Cordia (3.2 ha) and Cocos-Cordia (0.82 ha). Its woodlands
total 49.60 ha. Coconut crabs inhabit all areas containing
Cocos and Pisonia; our rough estimate of their population is
200 crabs.

Birds: In 1988, Long supported 9 ( 10 in 1965) species of
breeding seabirds. as follows (pairs): red-tailed tropicbird (5).
masked booby (69), brown booby (12), red-footed booby
(659), great frigatebird (808), sooty tern (179,800). brown

noddy (207), black noddy (986), and white tern (751 ). From
1988 through 1990, sooty terns occupied 19 large colony sites
(Fig. 11, Subchapter 1.2).

Comments: Polynesian rats were abundant, especially in
Cocos and Pisonia habitats. It was often possible to see 3 or 4
simultaneously while conducting daily surveys and 20 or more
around camp. At night, their numbers increased substantially.
Azure-tailed skinks (Emoia cyanura) were noted.

Windward Islets

This chain of 13 islets occupies the southern half of
Caroline's east coast. All rest on the same reef flat, separated
by surge channels varying in width and depth. They can be
waded with care at low tides, but most harbor black-tipped reef
sharks: up to four were visible in the shallows within 50 m of
an observer.

The motus range in size from Noddy Rock (0.02 ha) to
Windward (1 1.42 ha). They support every major vegetation
type, from simple herb mats to magnificent Pisonia forests,
21 m tall. Because of their constant exposure to trade winds,
the seaward vegetation is wind- and salt-shorn. Though
appearing completely natural, all of the Windward Islets were
intensively planted with Cocos (Table 13) from 1916-1920
(Young, ca. 1922). However, these incipient plantations
experienced difficulty and appear to have been abandoned
within a few years (see Plant Communities section).

Flanking the lagoon of the southern motus (Brothers
through Tridacna) and extending westward are coral reefs
densely studded with giant clams, whose iridescent,
multicolored mantles add to Caroline's outstanding natural
assets (PI. 26; Subchapter 1.2, Conservation section).

3. BO'SUN BIRD ISLET (0.86 ha) (Figs. 29,42; PI. 9)

We named this motu for its red-tailed tropicbirds,
commonly called bo' sun birds. The sizeable population is the
largest on Caroline. In addition, our 1988 records constituted
the first known breeding of this species on the atoll.

Physiography: Bo'sun Bird Islet, 165 m west of Long's
southern tip. is the only motu lying "within" Caroline' s lagoon.
It shares the same reef as Long, however, and is not a true
"lagoon motu."

Amoeboid in shape, Bo'sun Bird is greatly affected by the
tidal waters that spread across the shallow reef flats and gush
through the surge channels that separate Long and Windward.
Because it sits near the inner edge of a wide windward reef flat,
the layering of sediments around it is complex and transitory;
our observations indicate that more rubble was deposited on
the islet's western edge since the aerial photos were taken in
1985. Its western shoreline rises gradually to a high water
mark, and slight changes in water level greatly change its
overall size and shape. At high tide its perimeter resembles the
shape of Pinocchio' s head — ovoid with a long, expanded nose.
The "head" is approximately 70 m wide and 115 m long, while
the "nose" is 45 m long and 15 m wide.

Vegetation and Birds: Bo'sun Bird Islet, composed of
coral rubble and sand, supports only natural herb mats
(Heliotr opium, Portulaca. Lepturus) and Tournefortia scrub

37

(to 4 m tall). These two simple plant communities cover 35%
and 55% of the land area, respectively. For its size, the motu
is sparsely vegetated, with only four plant species (one shrub,
three herbs), 15% of Caroline's total flora. There are no
introductions.

Bo'sun Bird's most notable attributes are its 4 species of
breeding seabirds: red-tailed tropicbird (47 pairs in 1988, but
130 pairs seen in 1990), sooty tern (8,400 pairs), brown noddy
( 10 pairs), and white tern (6 pairs).

4. WINDWARD ISLET ( 1 1.42 ha) (Figs. 29,43)

We named this "Windward" because it is the first major,
and largest. Windward Islet.

Physiography: Windward is broadly crescentic in shape.
508 m long by 287 m wide. It parallels the reef's longitudinal
axis and is set close to the lagoon. Its seaward beach is quite
narrow (3 m wide); there is no lagoon beach.

Vegetation: Windward has 1 1 species of plants (3 trees,
1 shrub, 7 herbs), 41% of the total flora. A windward crescent
of halophytic herbs borders a zone of Tournefortia scrub,
which mixes quite densely with Pisonia and Cordia over most
of the interior in a bilobed pattern. These latter forests,
reaching 14 m high in the south and 9 m in the north, total
8.67 ha. This unusual distribution of central forests undoubtedly
reflects Pisonia' s recovery from 1 00% land clearing for Cocos —
1,299 palms— in 1920 (Young, ca. 1922, Table 13). It is
remarkable that not one Cocos remains as a legacy of this
disturbance.

The east-west profile of Windward, similar to that of
Tr. 8, Long Island (Fig. 41 ), is typical of most motus, except
that lagoon-facing herb flats are almost nonexistent. Scaevola
sericea sericea, a new plant record for the atoll, is unique to this
motu, although S. s. tuamotensis was found on South Island in
1990.

Birds: Five species of breeding birds were present, all in
appreciable numbers (pairs): red-footed booby (163), great
frigatebird (207). brown noddy (20), black noddy (28), and
white tern (134).

Comments: In May 1990, AKK noted a possible motu
midway between Windward and Crescent Islets during midtide.
It appeared to be upraised reef like Noddy Rock, but because
of extensive shallow reefs in this area, it has not yet been
confirmed.

5. CRESCENT ISLET (3.10 ha) (Figs. 29.43)

We named this islet for its cupped shape.

Physiography: Crescent Islet is 1 90 m long by 225 m wide.
It is almost entirely composed of coral rubble, with a little
humus in the interior. The seaward beach is variable (up to
50 m wide), the lagoon beach, insignificant.

Vegetation: There are 10 species (3 trees. 1 shrub.
6 herbs), 37% of Caroline's flora. No introduced plants occur.
Plant diversity is poorer than on Windward, a reflection of
small size, poor soils, and scant herb mats. However, woodlands
cover two-thirds of its area, and the centra] stand of Pisonia and

Cordia is 87 m wide and up to 1 3 m high. Crescent was heavily
planted (80% of total area, 228 palms) in Cocos in 1920. but
today none remain.

Birds: Crescent Islet was used by the following numbers
of breeding pairs: red-footed booby (28), great frigatebird (5).
brown noddy (36), black noddy (60), and white tern (8).

6. MOTU ATIBU "Coral Rubble Islet" (0.02 ha)
(Figs. 27,43)

Motu Atibu was Caroline's smallest and least vegetated
islet. Third in the windward chain, it measured 13 m x 18 m.
We named it for its basic rubble character. Vegetation covered
only 2% of the land surface and consisted of a few Tournefortia
shrubs (<1 m high) encircled by narrow swaths of low herbs
and rubble. Its three plant species (one shrub, two herbs) —
1 1% of Caroline's flora — were among the most meager on the
atoll. Atibu' s profile was similar to that of Fishball (Fig. 56).
There were no breeding birds.

Comments: Since a February 1990 storm. Atibu has
apparently disappeared, having been reduced to a thin strip of
coral gravel below high tide level.

7. NORTH PIG ISLET (5.44 ha) (Figs. 29,44; Pis. 60.61 )

We named the fourth windward islet "North Pig" for its
location immediately north of Pig Islet.

Physiography: Classically crescentic. North Pig is 350 m
long and 230 m wide. Though approximately half Pig's area
and less wooded overall. North Pig has a similar distribution of
substrates (including sand on the lee side), vegetation, and
breeding birds. Profiles of the two motus are nearly identical
(Fig. 45).

Vegetation: There are 1 1 plant species (3 trees, 1 shrub.
7 herbs), 41% of Caroline's flora. No introduced plants are
present. Proceeding south along the windward islets, lagoon-
side herb mats develop and islet cross-sections assume a more
perfect symmetry — low at the edges and forming a hump in the
middle.

North Pig's three vegetation zones are predictably
symmetrical: a peripheral band of herbs (more extensive on the
"horns"), curved belts of Tournefortia, and a spacious central
forest of mixed Pisonia, Cordia, and Tournefortia. The latter
(to 20 m tall) covers more than one-half the islet's width and
one-third its area and includes fine Cordia groves (Fig. 44).
This excellent forest is surprising because 402 Cocos palms
were planted on 93% of North Pig's usable land in 1420
(Young, ca. 1922, Table 13). Measurements from 25 Pisonia
trees (main trunks) averaged 19 m in height, 221 cm in
circumference (at 1.5 m), and 261 cm in base circumference
(Table 12).

Birds: Five species of seabirds bred: red-footed booby
( 3 1 pairs), great frigatebird ( 1 7 pairs), brown noddy (76 pairs).
black noddy (3,199 pairs), and white tern (1 10 pairs). The
largest colony of black noddies on Caroline nested in the tall
Pisonias.

Comments: Rats and coconut crabs were common.

38

8. PIG ISLET (7.21 ha) (Figs. 29,44; Pis. 27,41,52,60,61)

Number 5 down the chain. Pig was named prior to 1883.
Domestic pigs were introduced to Caroline in 1 828 by Captain
Stavers but evidently died out before 1834. Reintroduced in
1848 with the first recorded settlers, it is not known how long
they lasted. One would expect that they were only on South
Island, but the statement that "about one-third the distance up
the lagoon a canvas hut exists on one of the smaller islets on the
eastern side of the lagoon" (Holden & Qualtrough, 1884)
suggests that perhaps domestic animals also inhabited Pig.
Though this is weak evidence, there must have been some
reason for this curious name. Today, fortunately, no pig
devastation is evident here or elsewhere on the atoll.

Physiography: Pig, shaped like a fat kidney bean, is
330 m long and 255 m across. It is separated from North Pig
by a channel 60 m wide.

Vegetation: The islet has 1 1 plant species (4 trees, 1 shrub,
6 herbs), 4 1 % of Caroline' s flora. Cocos, the only introduction,
is rare (0.03 ha). In 1920, 538 palms were planted (Young,
ca. 1922), which covered approximately 79% of Pig's usable
area (Table 13).

Pig's vegetation profile (Fig. 45) is classic: a wide,
windward herb mat, bordered by Tournefortia and Cordia,
which, in turn, grades rapidly into an outstanding Pisonia
forest (to 21 m tall, 3.36 ha), one of Caroline's best groves.
Measurements from five trees, mostly multiple-trunked,
averaged 16 m in height, 338 cm in circumference (at 1.5 m),
and 282 cm in base circumference (Table 12). This Pisonia
also occupies the largest proportion (46%) of any islet area. It
is striking that such quality forests could regenerate in about
65 years (see section on Plant Communities, Pisonia Forests).
In the Cordia forest (PI. 27), also the finest on Caroline, six
trees averaged 12.6 m in height, 1 16 cm in circumference (at
1.5 m), and 99.8 cm around their bases. On the lee side of Pig,
Tournefortia extends directly to the lagoon shore.

Birds: Five species of seabirds bred: red-footed booby
(14 pairs), great frigatebird (118 pairs), brown noddy
(82 pairs), black noddy (1,928 pairs), and white tern (164
pairs).

Comments: Rats and coconut crabs were common. In
1 990 agrayish gecko (possibly mourning gecko, Lepidodactylus
lugubris) was seen by A. Garnett.

9. SKULL ISLET (0.02 ha) (Figs. 27,44; Pis. 46,49)

Sixth in the windward chain, we named Skull Islet after
finding the skull, tail feather, and eggshell of a red-tailed
tropicbird, the first evidence that this species bred on the atoll.
A low shelf of coral rubble and sand, barely above high tide
mark, this motu is barren except for a small herb mat under five
Tournefortia bushes ( 1 m high) on the lagoon side. Only 2%
of the surface area is vegetated. There are three plant species
(one shrub, two herbs), 1 1% of the atoll's flora. Although
appearing young, the islet must be more than 100 years old, as
it is marked on Arundel' s chart (Fig. 4). After February 1990,
several large reef fragments had washed into the channel close
to Skull Islet.

In 1988 there were no birds. However, in March 1990, a
colony of 150 brown noddies was in a prelaying phase,
accompanied by 6 sooty terns, a brown booby and a wandering
tattler.

70. NORTH BROTHERS ISLET (1.71 ha) (Figs. 29.44: PI. 60)

The seventh windward motu, we named this islet North
Brothers because of its location directly north of the named
motu, Brothers.

Physiography: North Brothers is shaped like an oval that
curves lagoonward toward Brothers, 40 m away. The concave
shorelines and lack of herb mats on the opposite shorelines of
these 2 islets suggest that they might have been
formerly connected. Composed primarily of rubbly substrates,
with slightly better soils centrally, it is 95 m long and 250 m
wide.

Vegetation: Plant species number 10 (3 trees, 1 shrub,
6 herbs), 37% of Caroline's flora. A few Cocos trees are
present, remains of the 180 planted in 1920 (Young, ca. 1922),
which covered 100% of all available land on the islet
(Table 13). Plant communities on North Brothers are simple:
Tournefortia (more open in the west) rises to an excellent
Pisonia forest, 80 m wide and 18 m tall, on the east end.
Average measurements from three Pisonia trees were height
18 m, base circumference 314 cm, and number of trunks, 2.3
(Table 12).

Birds: Five species of seabirds bred on the islet in 1988
(pairs): red-footed booby (25), great frigatebird (9), brown
noddy (23), black noddy (40, plus hundreds of old nests), and
white tern (69). In September 1989, sooty terns nested on the
windward beach (Anne Falconer, personal communication),
and in May 1990, a prebreeding swirl of thousands of sooty
terns swarmed above Brothers and North Brothers.

Comments: Gecko eggs were seen on Pisonia trunks in
1990.

77. BROTHERS ISLET (4.31 ha) (Figs. 29.44.46: Pis. 30.60)

The eighth windward motu. Brothers Islet was named last
century after Captain Brothers, who managed a stock-raising
venture on Caroline. In 1 873, his rights to the atoll passed into
the hands of John Arundel.

Physiography: Shaped like a molar tooth, with roots
extending toward the lagoon, Brothers Islet lies about two-
thirds of the way down Caroline's windward reef. It is 198 m
long x 178 m wide through the center. A Tridacna reef extends
westward from it almost completely across the lagoon.

An interesting aspect of Brothers' structure is that Arundel's
chart (Fig. 4) indicates a tiny, separate motu off the southwest
point. Our survey and the 1985 aerial photos show that this
motu is now joined to Brothers Islet. Its former identity is
marked by a small patch of Tournefortia, around which the
recently deposited sand and rubble is sparsely dotted with
native herbs.

Vegetation: There are 1 1 plant species (4 trees, 2 shrubs,
5 herbs), 41% of Caroline's flora. Cocos, along the leeward
shore, is the only introduced plant. Three distinct plant

39

communities are present: peripheral herb mats (including
leeward Portulaca with Suriana), Toumefortia scrub and
forest ( to 6 m high ) bordered with Cordia, and a central Pisonia
forest. Larger trees had up to 15 trunks and multiple root
suckers. Measurements of 10 trees (main trunks) averaged
15 m in height, 140 cm in circumference (at 1 .5 m), and 243 cm
base circumference. Distances to nearest neighbor for 10 trees
averaged 4.2 meters. As on its neighbor islets, the Pisonia
forest on Brothers is striking, especially since it has matured to
a closed-canopy monotypic stand devoid of any subcanopy
species (Fig. 40), evidently in about 65 years. In 1920, Brothers
Islet was planted with 315 Cocos palms, which covered
approximately 97% of the usable land area (Table 13).

Birds: Four species of seabirds bred: red-footed booby
(25 pairs), brown noddy (8 pairs), black noddy (15 pairs), and
white tern (50 pairs). In May 1990, large numbers of sooty
terns swirled over Brothers and North Brothers Islets.

Comments: Many of the mature Pisonia trees contained
capacious cavities in their boles that housed large coconut
crabs. In March 1990, several of these holes had feathered
skeleta of sooty terns (and possibly also brown noddies)
outside their entrances, along with freshly-snipped Pisonia
branches (see Subchapter 1.2, Coconut Crabs section).

12. NODDY ROCK (0.02 ha) (Figs. 27.47; PI. 19)

We named this ninth motu in the Windward Islets for its
only known breeding seabird, the brown noddy. In September
1988. at least 80 pairs were incubating their eggs on the
Portulaca mat that covers its central lee section.

Noddy Rock, an eroded limestone plateau (feo), is 26 in
wide by 9 m long. It is windswept and salty, with waves
splashing over its eastern edge on most days (it rises only
0.5 m above high water). During storms it is completely awash
( Anne Falconer, personal communication ). Only three species
of plants (11% of Caroline"s flora) grow here, thinly covering
the western (leeward) third of the island in the following
proportion: 75% Portulaca, 20% Lepturus, and 5%
Toumefortia).

13. NORTH ARUNDEL (0.91 ha) discussed below
(Figs. 29,34.47; Pis. 14.62)

14. ARUNDEL (7.34 ha)

Arundel Islet was named last century in honor of John T.
Arundel. A British trader and guano digger. Arundel was one
of the leading figures in the Pacific phosphate industry . directing
guano and coconut planting operations on Caroline and other
islands from 1873 to 1897. His most valuable contributions,
however, were his excellent surveys and maps of several
central Pacific islands, including Caroline (Fig. 4). The islet
immediately to its north. Arundel's "cap." we named North
Arundel.

Physiography: Arundel's shape is a I at crescent, with
wedge-shaped North Arundel lying across a short channel
immediately to Us north. North Arundel is 80 m long x 1 30 m

wide, while Arundel is 375 m long x 210 m wide. They are
composed almost exclusively of coral rubble and Hanked on
their inneredges by Acropora reefs heavily laden with Tridacna
clams. Arundel's inner "horns" have evidently added more
sand and rubble since 1883 (Fig. 4).

Vegetation: There are 1 1 plant species (3 trees. 1 shrub.
7 herbs) on Arundel, 41% of Caroline's total. There are no
introductions. North Arundel has 1 1 (4 trees, 1 shrub. 6 herbs),
41% of Caroline's flora, including one introduction, Cocos.

The vegetation on this pair of motus, along with Tridacna
to the south, consists of extensive herb mats, low scrub and
small interior forests (Fig. 34), slightly less lush than the more
northerly windward motus. Their woodlands are primarily
Toumefortia, with a thin belt of Cordia and central Pisonia
groves (a bilobed pattern on Arundel). Morinda is unusually
common in Arundel's central forests, and Achyranthes is
especially abundant on North Arundel. Pisonia occupies only
13% of the land area on Arundel, compared to 46% on Pig.
Soils are extremely rubbly . with scant organic matter, a possible
legacy of the guano era.

Both North Arundel and Arundel were heavily planted
with Cocos in 1919-20 (69 and 646 palms, respectively). All
usable land was cleared (Table 13). Despite the extreme
paucity of Cocos today, the relatively scant Pisonia present
today on these islets, compared to those further north, suggests
that the Cocos plantations were more successful here. Their
proximity to South probably also guaranteed better maintenance.

Birds: Five species of seabirds bred on Arundel: red-
footed booby (37 pairs), great frigatebird (on territory,
September 1988; breeding confirmed, early 1989 by Anne
Falconer), brown noddy (1 1 pairs), black noddy (249 pairs).
and white tern (227 pairs). In May 1990, thousands of sooty
terns swarmed above Arundel and North Arundel.

15. TRIDACNA ISLET (9.08 ha) (Figs. 29.48,49;
Pis. 1,26.48.62.63)

The 13th and southernmost motu in the windward chain
was named by the present authors and Boris Sirenko for its
outstanding coral reef studded with giant clams {Tridacna
maxima).

Physiography: Somewhat crescentic. measuring 446 m
long and 250 m wide, Tridacna is one of the largest motus on
Caroline. Its ground surface is heavily littered with coral
rubble, having a sandy strip above the beach crest on the
windward edge.

Vegetation: There are 1 3 plant species (2 trees, 2 shrubs,
9 herbs), 48% of the atoll's flora. For its size, Tridacna's
\ egetation is surprisingly lacking in tall forests, a legacy of the
910 Cocos palms planted on 82% of its available land area
(Table 13). Vegetation patterns follow the usual concentric
zonation: peripheral herb mats border a discontinuous belt of
Suriana (windward side), while the large central mass is
dominated by scrubby Toumefortia— Morinda woodlands,
which cover 88% of the islet's area, yet only attain 7 m in
height. In cross-section (Fig. 49). the short woodlands are
particularly noticeable. Compare the present lack of Cordia.

40

paucity of Pisonia, and richness of herbs, both in species
numbers and abundance, with Pig (Fig. 45) and Brothers
(Fig. 46). Although there are no introduced plants, thick
patches of Lepturus also reflect past forest clearing.

Birds: Fourspeciesof seabirds were nesting in 1988: red-
footed booby (111 pairs), brown noddy ( 1 1 pairs), black noddy
(249 pairs), and white tern (227 pairs). Tridacna is periodically
a major breeding area for sooty terns. Clapp& Sibley (1971a)
found 4 main colonies totaling 250,000 birds, and large numbers
nested along the windward beach in August 1 989 (A. Falconer,
personal communication). Nests were located under shrubs, or
in open areas bordering them, and were evidently preyed upon
by coconut crabs.

16. SOUTH ISLAND (104.41 ha) (Figs. 30,36,50;
Pis. 1-7,12,16,24,34,39,44,45,50,62)

History: The history of South Island (called Rimapoto in
Young, ca. 1 922) is essentially the history of Caroline, for most
information about the atoll prior to 1965 is from here. It is the
second largest islet, and the staging area for trips up-lagoon as
it lies adjacent to both the "boat landing" and "blind passage."

South Island was inhabited in prehistory by Tuamotuans,
who planted the first small coconut grove on its northwest
point. The first Europeans to land, in 1606, found coconuts,
fish, lobsters, and seabirds in abundance. They dug for fresh
water in vain. Two hundred years later, in the decade after a
cyclone in 1825, pigs, sweet potatoes, arrowroot, and South
Sea chestnut were introduced. However, "the unfriendly
character of the soil, and the number of land crabs that infest it,
gave us but little hope of the experiment succeeding" ( Bennett,
1840). The pigs expired within a few years. The arrowroot,
tenacious and adapted to island environments, still exists today
(unless later immigrants brought it). Of the others — plus many
other later food plants and ornamentals — no trace exists
(Table 1). (In 1990, we found a few Hibiscus tiliaceus,
Thespesia populnea, and Ximenia americana. All could be
indigenous. The first 2 species were often planted by Polynesians
in copra-cutting settlements [e.g., Flint Island]). Tropical heat,
droughts, storms, excessive shade from Cocos, poor
germination, poor soils, terrestrial crabs, and lack of care all
undoubtedly contributed to their demise.

The first recorded settlement on Caroline, and first for the
Line Islands, was in 1846, on the northwest point. These
settlers, as well as subsequent ones, eked out a spartan living by
raising stock, drying fish and copra, and digging for guano.
Their managers built "proper" dwellings, so when US. British,
and French astronomers arrived to observe the solar eclipse in
May 1 883. South Island was quite "civilized," far more than it
is today. Three houses and two sheds "were in good repair,"
and a variety of "anchors, chains, spars, and pieces of the
woodwork of vessels" littered its reefs ( Holden & Qualtrough,
1884). Large grassy clearings adjacent to the lagoon
accommodated several European-style houses (Pis. 2—4). The
astronomers' account of South Island, illustrated with pen-
and-ink drawings (Pis. 2-5,50), is the only record of buildings
on Caroline, apart from mention of perhaps the same dwelling,
the manager's house, reported in 1936 by the "H.M.S."

Wellington to be "in excellent condition and spotlessly clean"
(Maude, ca. 1938), and a copra shed seen by Clapp & Sibley
(1971a). Arundel also took photographs, including some of the
marae on Nake, which we have not examined (Arundel's
memorabilia [photos, letters, diaries, a microfilm, et cetera] are
deposited in the Rare Book Collection, National Library and
Pacific Manuscripts Bureau, Research School of Pacific Studies,
Australian National University, both in Canberra. A.C.T.,
Australia).

Today, the houses, sheds, brick piers (constructed in 1883
for telescopes and observatory frames), signboard, flagpole,
marble slab with inscription "U.S. Eclipse Party, 1883,
May 6," and all but one of the introduced plants have
disappeared. In three trips we found no traces of the copra shed,
nor have the Falconers, after repeated visits over 2 years. All
that remains of the formerly large clearings are two small palm-
shaded "flats." in 1988 used by the US and Soviet scientists for
a base camp and work area. In 1987, the Falconers cleared one
of these for living quarters, and in 1 990 fishermen expanded the
other by burning an area 35 x 22 m, then erecting a tin shack,
cookhouse, and fishtrap, which were destroyed in a summer
1990 storm.

Our "civilization list" probably covered all that could be
seen on South Island without digging: a 26-foot wrecked sloop
( AK 669 1 J. ), complete with trail to a "Robinson Crusoe-type"
campsite strewn with remnants of radio and navigational
equipment, sail, cans, clothing, et cetera (southeast coast);
assorted flotsam and jetsam ( whisky bottles, Japanese fishballs,
plastic debris, et cetera); a large rubber ship fender; a bench
mark from the 1 985 RNZAF survey team; a recently renovated
concrete cistern (by the landing); and an old wooden canoe
lying on its side just like de Quiros found in 1606!

We assume that all the Polynesians, ancient and recent
(Tuamotuans, Tahitians, Niueans as far as is known), lived in
native thatched huts similar to the ones on Ana-Ana today.
Fashioned from coconut palms and pandanus trees, they
disappear quickly when abandoned. The largest number of
inhabitants recorded for Caroline ( probably all on South Island )
was "two managers and 52 laborers" in 1873 (The Commercial
Advertiser, 1873).

The history of South Island's coconut plantations from
1885 to 1901 and from 1916 to 1929 is discussed under
Coconut Woodlands (Plant Communities section).

Physiography: South Island forms the base of the thinly
crescentic isosceles triangle whose limits define the atoll. Its
own shape is that of an irregular parallelogram
858 m wide x 1 ,254 m long at its longest points (Fig. 50). The
north coast, a curved bay, forms the lower boundary of the
lagoon. This shore, along with the adjacent northwest
peninsula, has been the most trodden by man, but the scars
have healed, leaving few traces beyond the presence of
coconut palms.

The reef flats surrounding the outer three sides of South are
the widest on the atoll, averaging 23 1 m, 578 m, and 363 m on
the east, south, and west, respectively. The windward and
leeward reefs immediately to its north are 530 m wide. To
leeward is the small boat "landing," and to windward, the
"blind passage."

41

Aesthetically, the lagoon fringe is one of the most
picturesque spots on the atoll. Lush palms overhang a narrow
beach of blinding white sand and coral gravel, affording idyllic
views of the azure lagoon and its encircling wooded motus
(Pis. 7,24,39).

Vegetation: There are 23 plant species (7 trees, 3 shrubs,
13 herbs), 85% of the atoll's flora. Cocos dominates South
Island, occupying 77% of its area. The healthy, but old, closed-
canopy plantations (21 m tall ) that border its coastlines give no
indication of the vast extent of the overgrown, dying groves
that occupy two-thirds of its interior ( Fig. 50; Pis. 7, 34). Here,
three species of herbs (Boerhavia repens, Portulaca lutea,
Phymatosorus scolopendha) have proliferated unnaturally to
form dense carpets, and the vine lpomoea macrantha climbs,
in tangled, strangling masses, to the tops of the highest palms.

The natural communities that prevail on other motus are
only minor components on South (Fig. 50): herb mats (13% of
the total area), coastal scrub with Suriana ( 1 % ), and Tournefortia
scrub (4%). Conspicuously absent are prime scrublands and
forests of Tournefortia, Pisonia, and Cordia, which undoubtedly
once swept in a lush expanse from shore to shore, stratified and
zoned as on other motus. Canopy heights of the plantations are
uniform (21 m). and the outer fringe of indigenous scrub
(Tournefortia, Cordia, Suriana) and herbs (fleliotropium,
Portulaca) occupy a small proportion of the island's width
(Figs. 36,51). Note the abrupt transition of canopy heights as
they drop to the level of coastal scrub on both sides of the
plantation (Fig. 51). Pandanus, too, is less extensive than
formerly: Bennett (1840) called Pandanus "somewhat
numerous" and PI. 50 reproduces an 1883 painting of a grove,
denser than any existing today on South. During our survey , we
observed only one small Pandanus grove and a few scattered
trees. Bennett also noted that the island was "covered with
verdure," and there were "trees attaining the height of twenty
feet." However, it is well to recall that 9 or 10 years previous
to Bennett's visit a violent storm had whipped over the atoll.
Drawings from 1883 (Pis. 4,5) depict remnant Tournefortia
and Pisonia trees larger than this.

Apart from the coastal buffer zone, little native forest
remains. Other sizable trees (Pisonia. Cordia), up to 17 m tall,
are rare, but Morinda, tolerant to both sun and shade, is still
quite common. Though we have not been able to trace any
records to Caroline, it is possible that shiploads of Cordia logs
were exported to San Francisco on guano ships, as was the case
on Flint, worked simultaneously by Arundel's company (Young,
ca. 1922).

A final noteworthy aspect of South Island is that, despite
its history of sporadic occupation and extensive forest felling
for coconut plantations, only one "weed," a tiny patch of
Phyllanthus amarus, and no vegetable or garden ornamentals
(excluding Polynesian introductions) have survived. (We are
unsure of the status of Hibiscus tiliaceus, Thespesia populnea,
or Ximenia americana). The 19th century gardens, once
drenched in sunshine, have long been buried beneath the deep
shade of palm groves (compare Pis. 2 and 24). In addition,
periodic storms, droughts, irregular rainfall, nutrient-poor soils,
rats, land crabs, and the harsh salty environment must have
contributed to the eradication of all exotics except traditional

native food and medicine plants, which are specifically adapted
for atoll environments. Studies on other atolls, even those near
high islands (Stoddart & Fosberg, 1972; Stoddart & Gibbs,
1 975 ), have demonstrated also that exotics survive, despite the
proximity to source areas containing garden ornamentals and
weed plants. We learned on our last two visits ( March and May
1990), however, that a small sunny clearing around the cistern
has attracted one clump of a weed not previously reported
(Kyllinga brevifolia), also the location of Phyllanthus. This
area is now used as an extension of the Falconers' vegetable
garden on Motu Ana-Ana. Kyllinga is listed as a temporary
species (Table 1 ).

Birds: Only 2 species of birds bred on South in September
1988, a reflection of its paucity of natural habitats: brown
noddy (163 pairs) and white tern (381 pairs). Bristle-thighed
curlews are very common, gathering in small flocks on the
rubbly shores (Subchapter 1 .2, this volume). They also forage
in the open Ipomoea-Cocos forest, perching on dead coconut
stumps 6-10 m high, then flying down to feed in the thick herb
mats.

Terrestrial Crabs: Caroline's highest population of coconut
crabs, having many huge individuals decades old. occupies the
open Cocos forests (Pis. 22,56,57 ). A crude minimum estimate
for South Island is 500 mature individuals (since March 1990.
these have become much reduced due to killing and preserving
in formalin for curios). We also found a fist-sized blue hermit
crab within a Turbo shell, possibly Coenobita brevimanus
(Yaldwyn & Wodzicki, 1979; E. Reese, personal
communication). As elsewhere on the atoll, land crabs such as
the reddish-purple Cardisoma sp. and scarlet hermit crabs,
Coenobita perlatus (in Turbo argyrostomus shells), were
abundant ( PI. 38). A Geograpsus sp., closer to the shore, was
less common.

Rats: Polynesian rats were abundant on South, active both
day and night. The rats were constantly afoot in broad daylight,
and at night a small flashlight beam often revealed a half dozen
at a time.

South Nake Islets (Fig. 52)

This chain of seven islets extends 1,500 m south from
Nake on the west side. They range in size from 0.64 ha ( Kota)
to 7.36 ha (Pandanus). All are well wooded and support every
natural plant community. Proceeding south, the overall plant
cover thins somewhat, but not to the dryness and openness of
the Central Leeward Islets. The herb mats are more extensive
than on the windward islets, especially to seaward. Aboriginal
introductions (Cocos, Pandanus) are sparse. We have found
no historical records indicating human disturbance to these
islets, thus their vegetation, with the possible exception of
Pandanus Islet, is evidently natural. The two scrawny Cocos
are probably drift-derived.

On the Solar Eclipse Party's map of Caroline (Fig. 5), only
the top two islets of this group are drawn. The South Nake Islets
constitute the only cluster of motus that show appreciable
differences between Arundel's chart (Fig. 4) and the 1985
aerial photos: most were shown as smaller, and with slightly
different shapes, by Arundel. The interior vegetation on these
motus includes mature forests of Tournefortia. Pisonia, and

42

Pandanus, so it is unlikely that these differences reflect changes
to the center of the motus. However, since the islets now appear
larger, accretions of coral rubble and sand that may have
occurred in the past 105 years, and are now barren or covered
only with herb mats, could account for most of the differences
(see Coral Islet discussion).

Although we have no actual records of sooty tern colonies
on this chain of islets, in May 1990 AKK observed pre-
breeding swirls of this species over Lone Palm, Kota, and
Mouakena (Subchapter 1.2, Fig. 1 1 ).

17. PANDANUS ISLET (7.36 ha) (Figs. 29,52; PI. 64)

This motu was named by the present authors for its coastal
Pandanus grove, probably a drift-derived offshoot from a
parent colony on Nake.

Physiography: Pandanus Islet, first in the chain, is
irregularly oval, 400 m long and 258 m across. It is nearly twice
the size shown on Arundel's map (ca. 3.4 ha). It occupies a
sheltered spot at the apex of the lagoon. Sand, actively filling
in the adjacent lagoon, is an important component of the
substrate on Pandanus, extending one-third of the way across
the islet. Although tidal reef flats are absent on the lagoon edge,
they average 75 m wide on the seaward side, producing a fairly
high proportion of rubble compared to the total land surface
(32%).

Vegetation: Plant species total 10 (3 trees, 1 shrub,
6 herbs), 37% of Caroline's flora. Cocus, surprisingly, is
absent, despite the close proximity to Nake. Pandanus Islet has
four basic vegetation zones: natural herb mats, Toumefortia
scrub (with Pandanus), Tournefortia-Pisonia forest, and pure
Pisonia. Woodlands cover 62% of its area. The widest pioneer
mats ( 13 m) of any leeward motu occupy its east edge and
though sparsely vegetated (20% Heliotropium, 5% Lepturus,
5% Portulaca) reflect active growth toward the lagoon.
Proceeding west across the island, Toumefortia scrub (2 m
high), with pockets of pure Pandanus (10m high), merges into
Tournefortia-Pisonia forest (to 14 m high), whose bimodal
distribution suggests that the islet was once divided. The
seaward coast supports open Toumefortia (5 m high), beneath
which herbs eventually thin out onto the extensive reef flats.

Birds: Five species of seabirds breed: masked booby
(2 pairs), red-footed booby (32 pairs), great frigatebird
(26 pairs), brown noddy (26 pairs), and white tern (52 pairs).

Comments: Skinks and rats were observed, along with the
ubiquitous Coenobita and Cardisoma crabs.

18. DANGER ISLET (2.71 ha) (Figs. 29,52; PI. 65,68)

We named Danger Islet to commemorate the deep, shark-
infested channel to its north, a barrier that aborted our first
(dusk) attempt to survey the South Nake Islets.

Physiography: Danger, shaped like a thickened comma, is
approximately 150 m long and 215 m wide. It is composed
almost entirely of coral rubble; interior humus is scant. Its reef-
channel flats are 21 m (north) and 14 m (south) wide. The east

and west beaches, narrow and wide respectively, are typical of
all the leeward motus.

Vegetation: Danger has 10 plant species (3 trees, 1 shrub,
6 herbs), 37% of the total flora. There are no introductions. The
usual concentric vegetation is clearly zoned: herb mats,
Toumefortia scrub and forest, central Pisonia, and Cordia in
the southwest. The herb mats are wide, extending 22 m and
15 m on the north and south shores, respectively.

Birds: Four species of nesting seabirds were present in
1988: red-footed booby (139 pairs), great
frigatebird (26 pairs), brown noddy (26 pairs), and white tern
(52 pairs).

19. BOOBY ISLET (0.84 ha) (Figs. 29,52; PI. 66)

We named this motu, third in the chain, for its two species
of boobies, the common red-footed and rarer masked booby.

Physiography: Booby, shaped like a teardrop, is 70 m long
and 1 25 m wide. Its coral rubble flats extend 1 0 m and 30 m on
the north and south sides, respectively.

Vegetation: Despite its small size, the most notable
feature of Booby is its Pisonia forest. 20 m tall and undoubtedly
virgin. It occupies the exact center of the islet in a circle about
40 m in diameter. Surrounding this is Toumefortia scrub (to
8 m tall), thinning out to peripheral bands of coral rubble.
Although less than one hectare in size. Booby Islet's woodlands
occupy two-thirds of this area. Booby Islet has nine species of
plants (two trees, one shrub, six herbs), 33% of Caroline's
flora, and no introductions.

Birds: Five species of seabirds breed: masked booby
(7 pairs), red-footed booby (52 pairs), brown noddy (2 pairs),
black noddy ( 1 pair), and white tern (6 pairs).

20. CORAL ISLET (1.70 ha) (Figs. 29,52; PI. 66)

Fourth from the north. Coral Islet was named for its reef-
derived coralline substrate.

Physiography: Shaped like an arrowhead, Coral is
approximately 130 m long by 200 m wide, more than three
times the size mapped by Arundel (Fig. 4). Most of its area is
barely higher than the surrounding interislet channels. The
shallow reef flats between Coral and its two southern motus are
only several centimeters deep at low tide; all three may be
destined to unite. Unless closely inspected, they appear to have
already merged, a fact which, together with Bryan's incorrect
map (Fig. 6), helps account for the widely differing number of
motus attributed to Caroline.

Vegetation: There are nine species of plants (two trees,
one shrub, six herbs), 33% of Caroline's flora, and no
introductions are present. Plant communities comprise a small
Pisonia forest (0. 1 3 ha), which is surrounded by the predominant
Toumefortia, which in turn is fringed with a narrow band of
native herbs. "Soils" are extremely coarse.

Birds: Five species of seabirds bred in 1988: masked
booby (1 pair), red-footed booby (28 pairs), great frigatebird
(2 pairs), brown noddy (6 pairs), and white tern (15 pairs).

43

21. LONE PALM ISLET (1.99 ha) (Figs. 29,52; Pis. 66-68)

We named Lone Palm, fifth in the chain, for its single
coconut palm, which towers, Hag-like, above a dense mound of
Towmefortia.

Physiography: Similar to Kota (to its south), Lone Palm
is sausage-shaped, 97 m long and 240 m wide, and four times
the size mapped by Arundel. Although composed almost
entirely of coral rubble, some sand borders the lagoon. Following
a pattern prevalent on all the leeward motus, its lagoon beach
is 2 m wide, while the seaward beach is 17 m.

Vegetation: Eleven species of plants are present (three
trees, one shrub, seven herbs). 46% of Caroline's flora. Lone
Palm's plant communities are simple: a wide band of herb mats
and open Toumefortia flanks an oval of Towmefortia forest (to
1 0 m tall ). A line of Pisonia trees, with a lone Cocos surmounting
the scrub, easily identifies this islet from lagoon or ocean.

Birds: Three species of seabirds bred in 1988: masked
booby (2 pairs), red-footed booby (48 pairs), and white tern
(9 pairs). In May 1990, we saw a large prebreeding swarm of
sooty terns.

22. MOTU KOTA -Red-footed Booby Islet" (0.64 ha)
(Figs. 28,52: Pis. 66.68)

We named this motu for its high density of red-footed
boobies (kota in Gilbertese).

Physiography: Sixth in line south of Nake, sausage-
shaped Motu Kota is 50 m long and 175 m wide. At low tide
it is almost connected to Motu Mouakena. Both surveys
indicate that coral rubble, the islet' s predominant substrate, had
further accumulated on its south side since the 1985 aerial
photos and also since 1988.

Vegetation: Though barely wooded. Kota has 1 1 species
of plants (3 trees, 1 shrub, 7 herbs). 41% of Caroline's flora.
One introduction is present, a single, tattered Cocos, partly
hidden by vegetation. Two plant communities are present:
peripheral herb mats and central Toumefortia scrub (to 10 m
tall |. with a few Pisonia.

Birds: Three species of seabirds bred in 1988: brown
booby (1 pair), red-footed booby (12 pairs), and white tern
(3 pairs). In May 1990, a single masked booby was on territory,
and sooty terns swirled overhead.

23. MOTU MOUAKENA "Masked Booby Islet" ( 1 .00 ha)
(Figs. 29,52; Pis. 15.69)

This islet was named for its nesting masked boobies, a
relatively uncommon seabird on Caroline.

Physiography: Somewhat U-shaped, Motu Mouakena is
seventh, and southernmost, in the South Nake chain of islets.
Both sides of the "U" were, in the recent past, separate islets.
By joining on the west, a narrow, V-shaped inlet was created on
the lagoon side. Motu Mouakena. 100 m long and 1 60 m wide.
is extremely nubbly ami infertile; much rubble was reorganized
during the February 1990 storm. Seventeen meters to its south
lies a newly emerging shoal of sand and gravel (PI. 15), perhaps
destined to be Caroline's fortieth motu or perhaps part of

Mouakena' s southern shore. Since the above storm, rubble has
further accumulated on this shoal, its adjacent reef Oats, and the
channel separating it from Mouakena. It already supports one
Toumefortia shrub, two dozen Heliotropium plants, and very
scattered Lepturus and Portulaca.

Vegetation: The number of species is eight ( one tree, one
shrub, six herbs), 30% of Caroline's flora, with no introductions.
Mouakena is thinly vegetated with open Toumefortia scrub (to
9 m tall, 26% cover), a few small Pisonia. and very sparse herb
mats.

Birds: Mouakena has less vegetation and fewer birds than
might be expected from a consideration of its area because
much of it is unshaded, coarse coral rubble. Though
unproductive botanically, this provides ideal nesting grounds
for masked boobies, one of the two species of breeding seabirds
on the islet in 1988: masked booby (3 pairs) and red-footed
booby (8 pairs). In May 1990, we saw one great frigatebird nest
with eggs and a swirl of sooty terns.

Central Leeward Islets

This chain of 1 1 motus occupies the central west side of
Caroline. All are separated by channels, wadable only at low
tide but prowled by belligerent sharks. Approximately
1.600 m south of Motu Mouakena lies a sandy shoal (0.5 m
high. 7 m wide. 4 m long), close to the lagoon edge of the reef
flats and connected only by a thin thread of rubble to Motu
Mannikiba to its south.

The islets range in size from Mannikiba (28.50 ha), the
most northerly, to Fishball (0.46 ha), the most southerly. All
support good seabird populations and. though quite well wooded.
are nonetheless the least lush motus on Caroline. Historical
records of the Central Leewards are very scant: much of
Mannikiba's forest was felled to make room for a Cocos
seedling "nursery" (Young, ca. 1922). The bulk of "40 trees on
other islets." in Young's totals, were most likely from Shark
and Emerald. The rest of this group is evidently pristine; the
natural communities on Bird Islet, in particular, are in excellent
condition.

Shark Islet boasts the best sandy beach on the atoll.
In common with all the motus on Caroline's west rim.
their lagoonside beaches are narrow and leeward reef
Hats wide. The leeward flats are composed of a greater
variety of substrata than the former, including coral rubble
of several grades (always gray), upraised reef, and
beachrock. Periodically, thousands of nesting sooty
terns occupy their open spaces (Clapp & Sibley. 1971a:
AKK, personal observation; Anne Falconer, personal
communication).

24. MOTU MANNIKIBA "Seabird Islet" (21.49 ha)
(Fig. 29; Pis. 70-72)

We named this motu for its teeming seabirds, mannikiba
in Gilbertese.

Physiography: Largest and most northerly of the Central
Leeward Islets. Mannikiba is somewhat rectangular with
rounded corners. Its reel Hats, containing an incipient islet,
stretch 2.0 km north to the South Nake Islets.

44

Mannikiha's maximum dimensions are 700 m long and
375 m wide. On the lagoon side, the serub skirts high water, but
when the tide drops, a strip of blinding white sandy coral lines
the lagoon. To seaward, upraised reef, beachrock, and successive
layers of gray coral rubble stretch in a wide swath (40 m)
toward the outer reef, 130 m distant. Throughout the islet the
substratum is gray coral rubble, with some exposed reef flat
hardpan in the northeast. Having numerous seabirds, this motu
might have also contained productive guano deposits.

Vegetation: Mannikiba, the fourth largest motu, harbors
1 3 plant species: (4 trees, 2 shrubs, 7 herbs), 48% of Caroline's
flora. One of its shrubs (Species A), a new record for Caroline,
has yet to be identified. The only introduction is Cocos,
occupying 0.1% of the land area.

Mannikiba' s vegetation, denser toward the north end, is
clearly zoned: herb mats, Tournefortia scrub and forest, and
scattered Pisonia groves. The few clumps of peripheral Cocos
are probably not drift-derived but the remnants of 6,000 "seed
sets" brought from Flint Island in June 1920. These were stored
on Mannikiba and "used to replant misses on other islets"
(Young, ca. 1922).

Pisonia, though present, occupies only 5% of the land
area, a small percentage for such a large islet. This suggests that
a large portion of the interior forests were felled to accommodate
the coconut "sets." This is also confirmed by the presence of
several old cut stumps in the interior. A century ago, Holden
& Qualtrough ( 1 884) noted that "About one-third the distance
up the lagoon a canvas hut exists on one of the smaller islets on
the eastern side of the lagoon, and two wooden huts stand on
one of the western islets, some distance further up the lagoon."
Mannikiba, the largest western islet, situated about halfway up
the lagoon, was most likely the site of the wooden huts, erected
around 1 920 and used for the following few years when the new
company, S. R. Maxwell & Co., Ltd., was anxious for the
success of Caroline's plantations. Although nothing more is
known of Mannikiba's history, collection of guano from its
numerous seabirds, including large populations of frigatebirds
and sooty terns, may account for further past disturbance.

Transect 1 ( north-central sector, PI . 7 1 ) passed through the
heart of a fine interior forest, while Tr. 2 (south-central sector)
passed through scrub and herb mats, which may represent part
of the former Cocos "nursery." Profiles through these two
cross-island transects resemble those from Brothers (Fig. 46)
and an old interisland channel on Long (Fig. 40), respectively.

The low, peripheral herb mats (absent from the lagoon
side) are composed of 30% Heliotropium, 20% Boerhavia,
1 5% Tournefortia, and less than 1 % of Portulaca and Laportea.
They are best represented in the southern sector. The
Tournefortia forest, 6 m high on both sides, is thick, having
95% canopy coverage. The Pisonia forests, though fragmented
(12m high, 1 00% canopy cover), contain Morinda, Boerhavia,
Achyranthes, Laportea, and Phymatosorus, but none cover
more than 1 0% of the ground area.

Seabirds: Six species of seabirds are known to breed: red-
footed booby ( 184 pairs), great frigatebird (287 pairs), brown
noddy (161 pairs), black noddy (176 pairs), and white tern

(195 pairs). No sooty terns nested on this islet in 1988, but
Clapp & Sibley ( 1 97 1 a) estimated 2,500 pairs in 1 965, and the
Falconers reported large colonies on Mannikiba, Blackfin and
Matawa in July-August 1990.

Comments: Coconut crabs live in the Cocos grove.
Azure-tailed and snake-eyed skinks (Cryptoblepharus
poecilopleurus), as well as a gecko, were noted in 1990 (DHE,
G. Wragg, personal observation).

25. BLACKFIN ISLET (2.62 ha) (Figs. 29,54; Pis. 3 1 ,73 )

We named this motu, second in the Central Leeward chain,
for two exhilarating shark attacks (near misses) within its
northern surge channel.

Physiography: Blackfin, shaped like conjoined ovals, is
140 m long and 190 m across. Coral rubble covers 30% of its
surface; all beaches and upper reef flats are of variable widths,
due in part to the fact that it has. in the recent geological past,
incorporated a smaller, circular motu into its northern confines.

Vegetation: Blackfin Islet has nine species of plants (three
trees, one shrub, five herbs), 33% of Caroline's flora. The only
introduction, Cocos, is rare. Four plant communities were
identified. Herb mats are well represented, especially around
the newly incorporated islet. The Tournefortia scrub, 21 m
v/ide in the east, is short (to 2 m), but approaches the stature of
a forest (to 6 m) in the west. The central forests of Cordia and
Pisonia (0.41 ha) are 9 m high.

Birds: Three species of seabirds bred in 1988: great
frigatebird (4 pairs), brown noddy (37 pairs), and white tern
( 1 1 pairs). In May 1990, one red-footed booby sat tight on a
nest, while two months later large numbers of sooty terns began
laying.

26. MOTU MATAWA "White Tern Islet" (1.71 ha)
(Figs. 29.54; PI. 55: Subchapter 1.2, PI. 3)

On arriving at this islet, the authors were swarmed by
1 5 white terns, all hovering within arm's reach and exhibiting
the ethereal grace that inspired their former common name,
fairy tern. I-Kiribati (Gilbertese) call them matawa.

Physiography: Of oval shape, Motu Matawa is third from
the north in the Central Leeward chain. It is 105 m long and
190 m wide. The entire motu, like all of Caroline's small to
medium islets, is built of coral rubble of varying grades, whose
unvegetated portion comprises one-fourth or more of the land
area. Its lagoon beach is 2.5 m wide, while the seaward beach
(sparsely vegetated) is 6 m wide.

Vegetation: Matawa has 10 species of plants (4 trees,
2 shrubs, 4 herbs), 37% of Caroline's flora. There are no
introductions. The usual plant communities were present,
dominated by Tournefortia (to 7 m), which covers half the islet.
Vegetation is less lush and more open as one progresses south
on the leeward side. Coral rubble, flanking the beaches and
extending further inland, also becomes more evident. The east-
central Pisonia-Cordia forest (to 8 m) rises barely higher than
its surrounding Tournefortia.

45

Birds: Four species of seabirds bred: red-footed booby
(5 pairs), great frigatebird ( I pair), brown noddy (3 pairs), and
white tern ( 1 3 pairs). Most conspicuous were white terns, with
9 pairs breeding on the 30-m-wide transect swath. One dark
morph reef heron fished in the shallows. In summer 1990,
sooty terns bred.

27. EMERALD ISLE (8.34 ha) (Figs. 29,54; Pis. 25,74-76)

Fifth down the chain, we named Emerald for the richly
colored, translucent lagoon waters that fringe its shorelines.

Physiography: Of thickened crescentic shape, Emerald is
330 m long and 240 m wide. Its lagoonside reefs, patch reefs,
and coral knolls are irregularly patterned with sandy channels.
It is here that the verdure of the lagoon is most intense.

Vegetation: Emerald Isle has 12 species of plants (5 trees,
1 shrub, 6 herbs), 44% ofCaroline's flora. The only introduction
is Cocos. Four plant communities, with a fairly high species
diversity, are present: the herb mats, covering one-fourth of its
land area, are composed almost exclusively of Heliotropium
( 35% cover) with scattered low Tournefortia (30% cover). The
Tournefortia scrub and forest attains a maximum height of 8 m
and. for a little variety, is mixed about equally with Pandanus
over most of its width ( 144 m) on the seaward side.

The interior forest (to 1 1 m tall) is also mixed, with
Pandanus, Tournefortia, Pisonia, and a little Cordia (PI. 75).
This condition is similar to the mixed forest on Nake, but
because Cocos is absent, it appears more natural. The existence
of this 3.20 ha mixed forest, as well as a similar one on Shark,
prompted us to suggest that Pandanus may be both native and
Polynesian-introduced. Cocos is present as two small groves,
complete with coconut crab sign (mounds of shredded fibers,
PL 57), beside the east and midwest shores.

We have been unable to trace the history of Emerald's
forests: the presence of Cocos on the west side and fragmented
Pisonia suggest past disturbance.

Birds: Six species of breeding seabirds were present: red-
tailed Tropicbird (1 pair), red-footed booby (3 pairs), great
frigatebird (230 pairs), brown noddy (7 pairs), black noddy
( 150 pairs), and white tern (83 pairs).

Although we did not locate any red-tailed tropicbird nests,
two adults circled steadily overhead. Two reef herons (one
dark morph, one light ) also foraged in the inshore reef shallows.

28. SHARK ISLET (7.98 ha) (Figs. 29,55; Pis. 29,77)

We named this islet to commemorate a particularly
pugnacious shark who was so anxious to procure a human foot
that it charged shoreward and leaped up onto the beach.

Physiography: Stoutly crescentic. Shark Islet is 280 m
long and 310 m wide in the center. The sandy lagoon beach and
rubbly seaward beach are each 3 m wide. Beyond high water
the seaward reel Hats extend for 280 m. Like Emerald. Shark's
reefs and surrounding lagoon waters reflect particularly stunning
colors, perfect complements to the sparkling pink coral sand of
Caroline's prime beach.

Vegetation: There are 12 species of plants (5 trees,
1 shrub, 6 herbs), 44% of the atoll's flora. Shark has one
introduction, Cocos, forming 3 clumps along the lagoon beach
( 1 % of the islet's area). Shark's rings of vegetation approximate
the islet's outline. Herb mats dot the fine sand lagoonward.
while to seaward they emerge from coarse rubble. The
Tournefortia (to 7 m tall) eventually gives way to a 12-m-high
Pisonia forest studded with Cordia and Pandanus. Centrally
this mixed forest is unnaturally open, suggesting possible past
disturbance.

Birds: Four species of seabirds bred in 1988: great
frigatebird (118 pairs), brown noddy (37 pairs), black noddy
(125 pairs), and white tern (44 pairs), red-footed boobies were
nesting in 1990. The notable colonies of great frigatebirds and
black noddies are due in part to the extensive Pisonia forest,
covering half of the islet.

29. SCARLET CRAB ISLET (0.46 ha) (Figs. 28,55)

This motu was named by the authors in honor of Coenobita
perlatus, the scarlet, fist-sized hermit crab that is abundant both
here and on the entire atoll.

Physiography: Scarlet Crab, sixth in the chain and only
40 m long by 1 25 m wide, is a young motu shaped like a closed
pair of lips. It skirts the southern shore of Shark, separated from
it by a channel 1 6 m wide. Because its eastern end points into
the lagoon, there is no true lagoon beach. Together with the
next three islets. Scarlet Crab's seaward reef flats (480 m) are
the most extensive on Caroline's lee side.

Vegetation: Vegetative cover is slight: less than 1%
area coverage of Heliotropium and Laportea, interspersed
with 10 small Tournefortia (to 1.5 m). Its species count is 6
(1 shrub, 5 herbs), 22% of Caroline's flora. There are no
introductions.

Birds: Although during storms this motu is undoubted!)
awash, two species of seabirds were breeding in 1988: brown
noddy (one pair, on ground) and white tern (two pairs, in low
scrub).

30. MOTU NAUTONGA "Sea Cucumber Islet- (0.34 ha)
(Figs. 28.55)

We named this motu for the Gilbertese word for the black
sea cucumbers or "beche-de-mer" (Ludwigothuria sp. ) that are
densely strewn over all ofCaroline's reef shallows (PI. 10).

Physiography: Semicircular in shape. Nautonga is seventh
in the Central Leeward chain, measuring 70 m long and 80 m
wide. Situated close to the lagoon, it is one of three small islets
that barely protrude above the reef flats. Nautonga' s perimeter
beaches are all narrow (2m), and its seaward reef flats are wide
(495 mi.

Vegetation: There are nine indigenous species (three
trees, one shrub, five herbs). 33% of the atoll's flora. Though
small. Nautonga's vegetation is concentrically zoned,
comprising herb mats (10-14 m wide), and a central forest of
Tournefortia and Pisonia (84 m wide) up to 10 m high.

46

Birds: Five species of seabirds bred in 1988: red-footed
booby ( 1 1 pairs), great frigatebird (2 pairs), brown noddy
(7 pairs), black noddy (32 pairs), and white tern (10 pairs).
Lesser frigatebirds appeared to be preparing to nest in May
1 990. One pair of blue-gray noddies, flying toward Azure Isle,
was seen by the 1990 expedition in May.

31. AZURE ISLE (0.20 ha) (Figs. 28,55; PI. 53)

We named this small, wedge-shaped motu for the striking
blue-green of its nearby lagoon.

Physiography: Eighth from the north, this small, elongated
triangle of land is 30 m long and 66 m wide. Its seaward reef
flats are wide (512 m). Its surge channels are narrow and
shallow.

Vegetation: Azure has only seven species (one tree, one
shrub, five herbs), 26% of Caroline's flora. The Pisonia tree is
6 m tall. A young motu. Azure is a superb example of an early
stage of biological succession. Its plant cover consists of a
single mound of Toumefortia scrub crowned by a single
Pisonia tree, growing from rubble only one meter above sea
level. Only 45% of its surface is vegetated; the rest, primarily
on the ocean side, is coarse rubble. Azure Isle presents what
may be the minimum width of vegetation (38 m) in which
Pisonia can develop on Caroline.

Birds: This motu illustrates the speed at which seabirds
will utilize newly available habitats. Within its dozen or so
Toumefortia shrubs (to 4 m tall), three species of seabirds nest:
red-footed booby ( 7 pairs ), great frigatebird ( 2 pairs ), and white
tern (2 pairs). A pair of blue-gray noddies were seen in May
1990.

32. REEF-FLAT ISLET (0.09 ha) (Figs. 27,55)

We named this young motu for its primary characteristic —
its reef flats. Ninth in the Central Leeward chain, this curved
strip of coarse rubble lies parallel to the surge channels that
surround it. It measures about 20 m long and 60 m wide. Only
three species of plants are present (one shrub, two herbs), 1 1%
of Caroline's flora. They cover less than one-fourth of its area
and are distributed so sparsely that not one bird was present.

33. BIRD ISLET (4.05 ha) (Figs. 29,55)

This is one of the motus named on Arundel ' s chart ( Fig. 4),
probably reflecting the presence of black noddies and/or sooty
terns.

Physiography: Bird is ovoid, measuring 230 m long by
200 m wide. It sits close to the inner edge of the lagoon reef,
whereas 400 m of seaward reef flats stretch westward.

Vegetation: There are 12 species of plants (4 trees,
2 shrubs, 6 herbs), 44% of Caroline's flora. A small Cocos
grove is the only introduction. It is well wooded, with very
narrow herb mats (6% of total area). Toumefortia (to 8 m) and
Pisonia (to 14 m) each cover 42% of its surface; the rest is
rubble. The Pisonia forest is of good quality (90-95% canopy
cover), having scattered Mo rinda. Boerhavia. andAchyranthes

as an understory. One large clump of Suriana (14 x 14 m,
2.5 m high) occurred in the islet center (A. Garnett, personal
observation). Bird Islet shows very few signs of past disturbance,
having prime plant communities, rich in breeding seabirds.

Birds: Five species of seabirds nested in 1988: red-footed
booby (29 pairs), great frigatebird (6 pairs), brown noddy
(42 pairs), black noddy (329 pairs), and white tern (48 pairs).
In June 1990, many thousands of sooty terns laid on Bird and
adjacent Fishball.

34. FISHBALL ISLET (0.57 ha) (Figs. 28,55,56)

Eleventh and southernmost in the Central Leeward chain,
we named Fishball after discovering a large glass fishing float
with a broken bottom, decorously placed in the islet's center
within a square of coral slabs.

Physiography: Paramecium-shaped, Fishball lies close to
the lagoon and is separated from Bird by a shallow, rubble-
strewn channel 100 m wide. The motu is 45 m long by 144 m
wide, with seaward reef flats 595 m in extent. South of the islet,
the reef flats — wadable at very low tide — stretch 1 .4 km to the
Southern Leeward Islets.

Vegetation: The number of plant species is eight (one
seedling "tree," one shrub, six herbs), 30% of Caroline's flora.
Figure 56 depicts an east-west cross-section of Fishball, showing
a vertical profile and the relative abundance and distribution of
each species. Fishball exemplifies an emerging motu. All
plants are low and halophytic; most are herbs. The motu is half-
covered with a sparse herb mat of Heliotropium (10% cover),
with scattered Laportea. Lepturus , and Portulaca (less than 1%
cover each). Small Toumefortia shrubs (to 2 m tall) are
scattered in the central sector, while a tiny drift seedling of
Morinda, 7 cm high, struggled to gain a foothold in the
exposed, salty rubble.

This motu is a fine example of the initial stages of islet
formation and colonization. It demonstrates that sea-dispersed,
halophytic herbs first germinate on the coarse rubble, later
becoming shaded out by Toumefortia, enabling a greater plant
species diversity to establish. It is very unlikely that a water
lens is present.

Birds: Two species of seabirds bred in 1988: red-tailed
tropicbird (three pairs) and brown noddy (five pairs). In May
1990, many thousands of sooty terns covered the ground and
swirled in the air, day and night. On 23 May, no eggs were
found, but laying occurred on Fishball and adjacent Bird Islet
in June (A. Falconer, personal communication).

Southern Leeward Islets (Pis. 14, 78)

This chain of five small motus lies along the southwestern
edge of the lagoon. All are built upon piles of rubble about
3 m high, oriented in an east-west direction, and are separated
by shallow, narrow channels. They range in size from 1 .5 1 to
3.67 ha, and their topography, vegetation, and breeding seabirds
are similar. Although situated on the leeward side of the atoll,
the Southern Leeward Islets exhibit some windward
characteristics: they lie opposite and slightly north of a wide
break in the windward reef, which allows trade winds to sweep.

47

uninterrupted, across the lagoon. This promotes their 60-80%
cover of scrub or forest. Ana-Ana, the southernmost, was
periodically occupied from 1987-1991 by the Falconer family.

Of particular botanical interest are the interior forests,
composed of Pisonia mixed with more Cordia than elsewhere
on the atoll. Pure Cordia groves (mostly too small to map
accurately) typically occupy the forest periphery.

Their history (apart from the last 3 years) is unknown; all
appear to harbor virgin plant communities with occasional
drift-derived Cocos or Pandanus.

35. MOTU RAURAU "Blue-?ra\ Nodd\ Islet" (3.48 ha)
(Figs. 29,57; Pis. 14,78,79)

Northernmost of the Southern Leeward Islets, we named
this motu for the blue-gray noddies {raurau in Gilbertese)
observed there. A highly territorial blue-gray noddy was acting
as though a nest was nearby.

Physiography: Raurau is ovoid, with a small lagoonside
bay. and maximum dimensions of 1 80 m long and 23 1 m wide.
It has the most expansive rubble of all the Southern Leeward
Islets. This coarse coral clinker extends, apronlike, around the
islet, widest (40 m) closest to the lagoon and narrowest (10 m)
to seaward. The seaward reef flats extend 446 m to the ocean.

Vegetation: The number of plant species is 10(5 trees.
1 shrub. 4 herbs), 37% of the atoll's flora. Raurau's two plant
communities are simple: a very scant herb mat is sprinkled
with Tournefortia, which rises to 6-m-high scrub all around the
islet. Laportea forms a narrow band at the interface between
coral rubble and scrub. Centrally a Pisonia forest (to 13 m).
dotted with Cordia on the periphery, harbors much Morinda in
the understory. including the tallest Morinda ( 1 3 m) seen on the
atoll. A handful of drift-derived Cocos and Pandanus. the only
introductions, dot the scrub.

Birds: No seabirds were found on transect, but a perimeter
walk in 1 988 revealed that tour species bred on the west side in
the Tournefortia scrub: red-footed booby (10 pairs), great
frigatebird (31 pairs), brown noddy (1 pair), and white tern
( 2 pairs). This islet, for its size, is particularly rich in frigatebirds.

Comments: Polynesian rats are present.

36. MOTU EITEI "Frigatebird Islet "(1.41 ha) (Figs. 29.57;
Pis. 14.78)

Second in line from the north, we named this motu for its
nesting great frigatebirds, eitei in Gilbertese.

Physiography: Motu Eitei is fat-elliptical. 105 m long and
2X0 m wide. Lying perpendicular to the reef axis, it touches the
lagoon edge on its inner side. To seaward, the reel' Hats are
644 m wide.

Vegetation: There are eight species of plants (three trees,
one shrub, lour herbs), $09i of the atoll's flora, with no
introductions. Eitei is carpeted with three plant communities
in the usual concentric arrangement. However, there is a slight
difference in the species composition of the herb mats: on
transect, the southern mat (2 m wide) consisted solely of
Portulaca, while the north side contained a 3-m swath of
Heliotropium, Laportea. and scattered Suriana. Inside the mat

is a ring of Tournefortia scrub (to 5 m) and a central Pisonia-
Cordia forest (to 11m). Laportea is particularly abundant,
while Portulaca, normally confined to the edges, abounds in
small openings within the interior woodlands.

Birds: Four species of seabirds bred on Motu Eitei in 1988:
red-footed booby ( 1 7 pairs), great frigatebird ( 1 4 pairs ). brown
noddy (6 pairs), and white tern ( 18 pairs). The first blue-gray
noddy nest for Caroline was found in summer. 1990
(Subchapter 1.2. this volume).

37. PISONIA ISLET (2.45 ha) (Figs. 29.57: Pis. 14.78)

We named this motu for its fine Pisonia forest.

Physiography: Pisonia, third in the chain from the north,
is almost circular and lies closely appressed to its neighbor
islets. Its maximum dimensions are 140 m long and 220 m
wide. Like Raurau. it possesses a wide "apron" of coral rubble
and sparse herbs on the lagoon side. Its seaward reef flats are
300 m wide.

Vegetation: The number of plant species is 15 (5 trees.
2 shrubs, 8 herbs), 56% of the atoll's flora. The only introduction
is Cocos { few, scattered, north and south shores). Well wooded.
Pisonia harbors the customary three plant communities: the
herb mat is almost pure Heliotropium, dotted with Suriana.
One specimen of Lepidium bidentatum w as found in 1 990. The
Tournefortia scrub and forest, covering half of the motu's
length and width, grows to 9 m. while the Pisonia— Cordia
forest, covering 0.86 ha ( 35% of the islet's area), reached 1 0 m.

Birds: Despite the beautiful Pisonia forest, no black or
brown noddies nested. Only three species of seabirds bred in
1988: red-footed booby (26 pairs), great frigatebird! 14 pairs).
and white tern ( 10 pairs). Best represented were red-footed
boobies: a perimeter count yielded 18 tended nests, all in
Tournefortia scrub. A long-tailed cuckoo was heard in the
interior.

Comments: Rats were common: six were noted on a mid-
morning transect survey.

38. MOTU KIMOA "Rat Islet" (1.80 ha) (Figs. 29.57:
Pis. 14.78,80)

Fourth from the north, we named this motu for its single
mammalian inhabitant, the Polynesian rat, kimoa in Gilbertese.

Physiography: Kimoa, smallest of the Southern Leeward
Islets and shaped like a flared teardrop, is squeezed between its
neighbor motus. Its maximum dimensions are 92 m long and
218 m wide, almost four times the size mapped by Arundel
(Fig. 4). The southeast rubble and herb mats are wide. The
distance to the outer reef edge is 307 m. Of special note is the
emergent Tridacna—Acropora reef, which stretches completely
across the lagoon to Tridacna Islet. This reef is 1 5-20 m wide
(Fig. 48; Pis. 26. 63 land 1.023 m long, which, together with an
equal length in blind diverticulae. totals over 2 km. The
Tridacna clams aggregate in densities up to 80/ in' ( Sirenko &
Koltun, Subchapter 1.4).

Vegetation: Kimoa has II species of plants (3 trees,
2 shrubs. 6 herbs). 419? of Caroline's flora There are no
introductions. Though small and narrow. Kimoa is well

48

vegetated. Its herb mats are composed of Heliotropium on the
south side and Portulaca (plus Suriana) on the north. The
interior Tournefortia-Pisonia-Cordia forests (to 1 1 m) cover
nearly one-half of the islet's area.

Birds: Four species of seabirds bred in 1988: red-footed
booby (21 pairs), great frigatebird (3 pairs), black noddy
(2 pairs), and white tern (7 pairs). Red-footed booby nests were
located along the perimeter.

39. MOTU ANA- ANA "Anne 's Islet" ( 2. 1 6 ha) (Figs. 29,57;
Pis. 5a. 14,54,78,81)

This motu includes a small settlement with three thatched
huts (cooking, eating, sleeping), a water tank, chicken coop,
and garden. It was occupied from 1 989- 1 99 1 by Anne and Ron
Falconer, 2 small children, chickens, Muscovy ducks, and a
dog. When we discovered a wooden sign marked "Ana-Ana"
and adorned with a shell lei, we knew the islet had been named.

It is interesting to compare Pis. 5a and 8 1 , identical views
of Ana-Ana 105 years apart. The profiles are indistinguishable,
showing how little this motu has changed over the years.

Physiography: Ana-Ana is the southernmost motu in the
Southern Leeward Islets, 120 m long by 222 m wide at its
widest point. Approximately 3 m high, it is roughly oval, with
a hooked point and curved bay facing the lagoon. This point is
actively growing as more and more rubble is deposited by the
large flow of water passing through the channel (430 m wide)
that separates Ana- Ana and South Island. This channel contains
abundant giant clams that amass into an extensive Acropora-
Tridacna reef stretching approximately 900 m across the
lagoon to Tridacna Islet. The outer reef flats measured
281 m.

Vegetation: Ana- Ana has 15 species of plants (5 trees,
2 shrubs, 8 herbs), 56% of Caroline's flora. Introductions
include Cocos, vegetables, a few ornamentals and, as yet. no
weeds. Ana-Ana's vegetation is typical of the other Southern
Leeward Islets, except for the settlement. Narrow trails from
the southern channel lead to a neat clearing, approximately
40 m x 70 m, the only inhabited portion of the atoll. We have
advised the Falconers against introducing exotic plants with
spreading seeds and have requested them to destroy all
introductions when vacating the island permanently.

Ana-Ana has sparse strand vegetation: Suriana,
Heliotropium, Portulaca, Laportea, and Lepturus. The
Toumefortia scrub includes Cocos, Cordia, and Pandanus. A
quality Pisonia forest, 1 5 m high, covers 43% of the islet's area.

Birds: No breeding seabirds were found on any of the three
visits to Caroline. However, the Falconers have found a few
white terns and one great frigatebird nesting in the perimeter
scrub, as well as groups of brown noddies sitting on the beach.
Long-tailed cuckoos were seen around the huts in March,
April, and May 1990.

Comments: Rats are abundant. Although the house site
was clean and tidy, 1 2 rats were seen in a pile of coconut debris,
and others scurried amongst the forest litter. The Falconers

have trapped over 1 ,300 rats in less than 2 years. Several pale
geckos with a few spots and largish heads were seen in and
around the thatched huts (probably mourning geckos).

Conclusion

Lushly wooded Caroline Atoll, with the majority of its
39 islets (399 ha of land) either in near-pristine condition or
having recovered remarkably from past disturbance, is one of
the least spoiled atolls in the Pacific. Uninhabited except for
one family, it harbors plant ecosystems and breeding seabirds
(Subchapter 1.2, this volume) of national and international
importance. Its marine and terrestrial ecosystems are prime
outdoor ecological laboratories for research on geological
processes including groundwater, fish poisoning, and numerous
facets of ecology (especially plant succession). Caroline
boasts outstanding coral reefs thickly studded with giant clams,
substantial numbers of coconut crabs, breeding sites for green
turtles, wintering grounds for shorebirds including the rare
bristle-thighed curlew, ancient Tuamotuan marae, and a
crystalline lagoon. The variety, abundance, and quality of its
flora and fauna qualify it for status as an officially recognized
international preserve (Subchapter 1 .2. Conservation section).

An expedition of this magnitude entailed the help of many
people, and it gives us great pleasure to thank them. We are indebted
to Hal O'Connor and Randy Perry. Patuxent Wildlife Research
Center, for making possible our participation. Steve Kohl, FWS
Office of Internationa] Affairs, and Terry Whitledge aided immensely
by handling innumerable details with their Soviet colleagues. Members
of the Fish & Wildlife Service Mauna Loa Field Station, especially
Jim Jacobi. Julia Williams. Jack Jeffrey, and Martha Moore, provided
welcome logistical support in Hilo. and Paul Sykes willingly shouldered
additional responsibility that freed CBK to join the expedition.

On the Soviet side, we thank Professor Alia V. Tsyban, chief
scientist of the expedition, for extensive help and friendship during the
voyage. Captain Oleg A. Rostovtsev. Yevgeniy N. Nelepov. Yuri L.
Volodkovich, and the ship's crew for handling many ship-board
details and transporting us to and from Christmas Island and Caroline.
Contacts with our Soviet colleagues would have been far less
memorable without the translation skills of Svetlana V. Petrovskaya
and. especially, Valeriya M. Vronskaya.

We thank Greg Smith and Chuck Stafford for transporting us
within Caroline's lagoon in the inflatable Zodiac "Tigris." Katino
Teeb'aki shared in the hard work on transects, and his skills at
climbing palms and opening coconuts often energized us during the
wilting midday heat. Abureti Takaio. former Minister for the Line and
Phoenix Groups, permitted us to work on Caroline and. with the
residents of Christmas Island, arranged a memorable evening of
dancing and food, despite the fact that their last supply ship was
10 months previous: for this and their many kindnesses we are most
grateful.

Financial assistance for the 1988 expedition and for writing the
manuscript was provided by the US Fish & Wildlife Service. Patuxent
Wildlife Research Center, and the Natural Environment & Climate
Monitoring Laboratory. Goskomgidromet USSR.

Grateful thanks are extended to Derral Herbst for identifying and
preparing plant specimens (deposited in the B. P. Bishop Museum.
Honolulu. Hawaii ) and George Zug for identifying lizards (deposited

49

in the US National Museum. Washington. DC). The Royal New

Zealand Air Force supplied the aerial photos. Roger Clapp, Ray

rg, Gene Hclfman, Harry Maude. Ernst Reese, and David

Stoddart shared unpublished manuscripts and other information. We

are most grateful to Lynda Garrett and Wanda Manning of the
Patuxent Wildlife Research Center I ibrary . Laurel, Maryland, for
digging out obscure historical references. Harry Maude of the
Australian National University introduced us to plantation records
and similar "gray literature." which proved indispensable in
understanding Caroline's past and present ecology. The libraries and
herbarium at the University of Georgia were also useful. We
especialK thank Bonnie Fancher for her efficiency, enthusiasm, and
hard work, often late at night and on w eekends. on the computer and
in other clerical matters. The manuscript has benefitted from reviews
by Ron and Anne Falconer. Ray Fosberg. Pat Roscigno, Betty Ann
Schreiber, Fred Sibley. Terry Whitledge. and Stephen Zeeman.

AKK. as coleader of the ICBP 1 990 expedition, expresses much
gratitude to Christoff Imboden (International Council for Bird
Preservation) and coleader/expedition initiator Martin Garnett, for
sharing finances. Thanks also to Martin and Annie Garnett. John
Phillips, and Mark Linsley for help with field work, and to Alve
Henricson for his sailing skills. The expedition would not have been
successful without the dedication of Captain Graham Wragg.

skipper/owner of TeManu, w ho transported us 7.400 km in the central
Pacific (including two \ isils to Caroline), helped with field work,
and whose competence and consideration in many areas eased the
varied hardships associated with 3 months at sea in a 10.5-m
ketch. Thanks also to Scott Miller for providing insect \ials and for
preparing and depositing insect specimens in the Bishop Museum.
Hawaii.

On remote Caroline, the Falconers were exceptionally hospitable
hosts, developing a special interest in its wildlife and helping us with
field work during and after the expedition. Special thanks go to
7-year-old Alexandre, who discovered the first blue-gray noddy nest
for the island and 3 new plant records.

French Polynesian residents who assisted in various ways
include Jacques Florence. Les and Gloria Whiteley. Rick Steger.
Michael Poole. Jean Roudeix. and friends who supplied us with
fruit and \ egetables for the trip. We particularly thank those who have
aided us in follow-up conservation efforts: Kelvin Taketa.
Jim Maragos, and staff of The Nature Conservancy-Hawaii,
Christoff Imboden and staff at ICBP. Alex du Prel. Jean-Michael
Chazine, Philippe Siu. George Monet. Graham Wragg. the
Falconers. Customs authorities in Papeete. George Ariyoshi.
Secretary and Minister to the Line and Phoenix Islands, and
the Hon. Secretary to the Cabinet. Republic of Kiribati.

50

TABLE 1

Plants reported from Caroline Atoll, but considered to be
transient or extinct members of the flora.'

Scientific Name

English and
Gilbertese Names

Date Last
Reported

Comments

CLASS ANGIOSPERMAE
Family Graminae

Eleusine indica (L. ) Gaertn.

Eragrostis amabilis

(L.)H.&A.
Family Cyperaceae
°Kyllinga brevifolia Rottb.

Family Bromeliaceae
Ananas comosa L.
Family Liliaceae

goosegrass. te uteute 1 884

lovegrass, te uteute 1884

kyllinga

pineapple, te bainaboro 1884

Crinum sp.

lily, te kiebu

1884

Presently

cultivated

Family Moraceae

°Artocarpus altilis (Park.) Fosb.

breadfruit, te mai

Presently
cultivated

Fiats carica L.

common fig, te biku

1884

Family Basellaceae

Boussingaultia gracilis Miers

Madeira vine

1884

Family Leguminosae

Inocarpus fagifer

Tahitian or Pacific

1840

(Parkins, ex Z) Fosb.

chestnut, mape (Tahiti),

(= 1 . fagifents)

te ibi

Family Euphorbiaceae

Euphorbia hirla

garden spurge, sleeping

1884

(= E. pilulifera)

plant, te kaimatu

Family Guttiferae

Calophyllum inophyllum L.

Alexandrian laurel.

1884

Family Caricaceae
Carica papaya

Family Cucurbitaceae
Cucurbita pepo L.

lpomoea batatas L.

I.pes-caprae brasiliensis
(L.) v. Ooststr.

Family Scrophulariaceae
Russelia equisetiformis
Schlecht.

tamanu (Tahiti), te itai

Papaya, pawpaw,

te babaia, te mwemwera

Pumpkin, te baukin.

te bamakin

Sweet potato, te kumara

beach morning glory,
pohuehue (Hawaii),
te ruku

Coral plant, te kaihaun
("golden plant")

Presently
cultivated

Presently
cultivated
Presently
cultivated

1884

Introduced weed
Introduced weed

One clump by cistern. South Is.,
on recently disturbed ground

Introduced for cultivated fruit

Introduced ornamental. One
small specimen found on South
Is. by Anne Falconer, 1990.
Collection no. K-90-14

Not yet established, 2 trees on
South and Ana-Ana
Introduced for cultivated fruit

Introduced "vine climbing over
portico" (Trelease 1 884)

Unsuccessful introduction in
1834. Food plant

Introduced weed, unsuccessful

In the 1940's. a "few taller
Calophyllum and Pisonia"
(N.I.D. 1943). No other
reference; did observer confuse
Calophyllum with Cordial

Cultivated for fruit in 1884.
not seen in 1965. In garden on
Ana-Ana, one on South Is. by
cistern

Cultivated in 1884, not found

in 1965

Introduced in 1840, not reported

again until this expedition (tubers

brought in 1988). Collection

nos. K-159, 160

Found in 1965 by copra shed;

extensive searching on 3

expeditions in 1988 and 1990

failed to find it

Unsuccessful introduction in 19th
century

'Since 1988, the Falconers have added more vegetables and ornamentals to their ever-expanding garden: green
beans, lemon grass, peppermint, okra, banana, Tahitian gardenia (tiare), tomato, breadfruit, red hibiscus, etc.
°Not previously reported from Caroline Atoll.

51

TAB1.K 2

Vascular flora ol Caroline Atoll: relative abundance of each species within the major ecosystems, with data on seabird utilization

Scientific Name

(.minion & Gilbertese
Name

Seabird
Utilization

o
o

2-

Nalural Ecosystems

Coastal

Natural Beach Town.
Herb Scrub Scrub
Mat With
Suriana

Inland

Toum.

Forest

A nt h ropogen ic Ecosystems

Coconut Woodlands

Pisonia Cocos Dying Mixed

Forest Plantation Cocos/ Forest

Ipomoea With

Plantation Cocos

TREES
Pisonia grandis

Morinda citrifolia
Cocos nucifera
Cordia subcordata

Pandanus tectorius

'Hibiscus tiliai ens

'Thespesia populnea

SHRUBS

Tourneforlia argentea
Suriana maritima

'Ximenia americana
'Scaevola s< rit ea

Species A

HERBS

Heliotropium anomalum

Boerhavia repens
Ponulaca lulea

Laportea ruderalis

'At hvranthi s i anesi ens
Phymatosorus scolopendria

I 1 punas repens
Ipomoea nun ranlha

Tacca leontopelaloides

Lepidium bidentatum
Psilotum million

Phyllanlbus amarus
Sida falkn

nana sp.
Iiihiilus i istoides

pisonia, puka tree,

te buka

Indian mulberry, te mm

coconut, te ni

sea trumpet, kou

(Hawaii I. ic kanawa

pandanus, screwpine,

te aroka, te kaina

beach hibiscus, hau

(Hawaii). ic nui

milo (Hawaii ). te bingibing

tree heliotrope, te ren
bay cedar, te aroa, te
mount

monkeyplum
scaevola, saltbush,
half-flower, te mao

"sand rose", hnitiliinn
( Hau an I
pigvine, te wao
yellow portulaca,
seaside purslane.
te boimari, te hot
"nettle", te ukeuke,
te nekeneke

maile-scented fern,
Itittti'c or lawai fern,
te raukota, te keang
bunchgrass, te uteute
morning glory, wild
moon-flower, te ruku
Polynesian or island
arrowroot, pia t Hawaii
& I ahiti), n- makemake
peppergrass
upright psilotum. "reed
fern", te kimarawa

'ilimti (Hawaii), te kaura
crabgrass, te uteute
puncture vine, te kiebu

R

UC

C

C

R-C

R-VC

R-O

UC

UC

UC-C

O

C

c

X X o-uc vc

O VC-A
R

X A UC
R-UC R-C

A A
O

O-A R-UC
R-VC VC

R
R-UC UC-A

PR

VC

R-A R-A

UC-C O

C-A C
R-C C

LC LC

L.R-A

R R

LR

A

L.UC

L.UC

R-C

R-A

R-C

C

O

O

A

A

C-A

R-VC

L.UC

R

0

O

1..R
L.R

A

UC-A

UC

LA

R-A

L.O

LC
C

C-A.L

R-UC

UC-A UC C-A

R-C R

LC R-UC UC-C
O A

R-A R-UC C-A
R

UC A

I \

LR
PR
PR
(S)

Excludes transient and extinct species i | able 1 1 Species arranged according to then overall abundance on the atoll.

New records foi Caroline.

Mot seen on this expedition, lasi seen 1965 K'lapp & Sibley, 1971a).

TABLE 3

Distribution and abundance of plant species on Caroline Atoll.1 Motus are arranged geographically from
north to south (windward), then similarly on the leeward side.

Windward Motus

South Nake Motus

N

L

B

W

c

A

N

P

S

N

B

N

N

A

T

S

P D

B

C

L

K

M

a

0

0

i

r

t

0

i

k

0

r

0

o

r

r

0

a a

0

0

0

o

0

k

n

ii

e

i

r

g

u

r

0

d

r

u

i

u

n n

0

r

n

t

ii

e

IT

s

d

s

b

t

1

t

t

d

t

n

d

t

d g

b

a

e

a

a

u

w

c

u

h

1

h

h

y

h

d

a

h

a e

y

1

k

n

a
r

e
n

P

B

e
r

R

A

e
1

c
n

n r
u

P

a

e
n

B

i

d

t

i
g

r

0

s

0

c

r
u

a

s

1
m

a

r

t

k

n

d

h
e
r

s

d
e

1

TREES(7spp.)

Pisonia grandis

A

vc

A

A

A

A

A

A

VC

A

R

R

VCA

VC UC

R

R

R

Morinda citrifolia

UC

UC

UC

O

R

UC

R

O

o

UC VC

C

R R

R

R

R

R

Cocos nucifera

LA

LA

LR

LO

R

A

LR

LR

Cordia subcordata

UC

UC

c

LA

C

VC

C

VC

C

UC

R

LA

Pandanus tectorius

LA

LO

LA

"Hibiscus tiliaceus

LR

°Thespesia populnea

LR

SHRUBS (5 spp.)

Toumefortia argentea

A

A

VC

A

A

O

A

A

0

A

A

O

A

A

A

C

A A

A

A

A

C

A

Suriana maritima

LR

LR

LC

UC.
LC

SkicvoIci sericea

S

Ximenia americana

LA

°Species A

HERBS (15 spp.)

Heliotropium anomalum

C

VC

UC

C

c

UC

UC

UC

R

UC

C

0 UC VC

C

c

0

UC UC

UC UC

R

Boerhavia repens

C

LC

C

c

c

c

UC

C

A UC

c

VC

c

O

O R

R R

R

Portulaca hitea

C

C

c

C

c

UC

UC

c

c

A

A

A

A C

c

c

cue

O O

O O

R

Laportea ruderalis

C

VC

c

UC

UC

C

C

UC

UC

c

o

O R

R R

R

°Achyranthes canescens

LC

O

o

c

o

LC

VCVC

LC

R

o

0

R R

R R

R

Lepturus repens

R

R

R

R

R

R

R

C

c

C

O

R

R

R R

R R

R

Phymatosorus

VC

LA.

UC

UC

UC

o

C

LA.

R R

scolopendria

UC

UC

Ipomoea macrantha

LC

0

UC

o

A

Tacca leontopetaloides

*LC

LR

Psilotum nudum

*LR

LR

Phyllanthus amarus

LC

Tribulus cistoides

s

Sidafallax

LR

Lepidium Indentation

S

*Digitaria sp.

s

53

TABLE 3 - continued

Central

Southern

Leeward Motus

Leeward Motus

T

% O

M

B

M

E

S

S

N

A

R

B

F

R

E

P

K

A

O

u

1

a

m

h

c

a

z

e

i

i

a

i

i

i

n

o

n

r

n

a

t

e

a

a

u

u

e

r

s

u

t

s

m

a

t

F

i

n

c

a

r

r

r

1

r

f

d

h

r

e

0

o

-

a

r E

g

i

k

w

a

k

1

0

e

-

b

a

i

n

a

A

1

e a

i

k

i

1
i

a

1

d

e

t

n

CT

f

1

a
1

u

i
a

n

a

q c
u h

n

b

i)

a

a

1

e

a

C

r
a

b

t

n M
C o

y i
u :

TREES(7spp.)

Pisonia grandis

C

VC

UC

C

A

O

R

A

A

C

A

VC

A

32

82

1

Morinda citrifolia

R

R

R

R

R

R

D

UC

R

R

R

C

30

77

LAI?

< .'. os nucifera

LR

LR

LO

LR

LR

LR

LO

LC

17

44

AI.RI

Cordia subcordata

O

UC

C

O

LO

O

UC

C

C

C

C

C

23

59

I

Pandanus tectorius

R

C

C

LR

LR

LR

9

23

LAI?

"Hibiscus tiliaceus

1

3

I(AI,RI?)

"Thespesia populnea

1

3

1 i AI.RI?)

SHRUBS (5 spp.)

Tournefortia argentea

A

A

A

A

A

R

C

C

R

A

UC

A

A

A

A

A

39

100

1

Suriana maritima

R

UC

LR

LR

LR

O

10

19

1

Si ,h \ ola serii ea

1

3

1

'Ximenia mucin una

1

3

Species A

S

1

3

1

HERBS (15 spp.)

Heliotropium anomalum

C

C

C

C

C

R

C

C

R

R

UC

UC

C

C

C

UC

38

98

I

Boerhavia repens

C

UC

UC

C

UC

R

R

R

UC

R

UC UC UC

O

c

34

88

1

Portulaca lutea

c

c

uc

R

R

R

R

UC

R

UC

C

UC

C

LC

37

95

1

Laportea ruderalis

uc

uc

o

c

uc

R

R

R

R

R

UC VC L1C

O

c

32

82

1

\i In ranthes i anescens

uc

o

R

uc

LA

R

R

UC

R

o

C

o

29

72

I

Lepturus repens

R

R

R

R

UC

R

R

R

VC

27

69

1

Phymatosorus scolopendr

uO

R

uc

14

3d

1

Ipomoea macrantha

O

R

7

18

I

Idcca leontopetaloides

3

RI.AI *

PsilotUtn milium

3

I

Phyllanthus animus

3

X

Tributes cistoides

3

1

Sidafallax

3

I

I epidium bidentatum

S

2

5

I

'Digitaria sp.

1

3

•>

Species arranged according to frequency of occurrence. List excludes transient and extinct members of the flora i [able 1 1
No. motus ha\ ing a particular species. di\ ided in total no. motus \ 100' I
New records foi Caroline.
■ Not seen on three visits, bul possiblj still present.

54

TABLE 4

Sizes of Pacific atoll floras, with emphasis on the
percentages of indigenous plants.1

Total2

No.

No.

Species

%

Island Group

Atoll

Species

Indigenous

Indigenous

Source

Caroline Is.

Kapingamarangi

98

38

39

Niering, 1962

(Fed. States

of Micronesia)

Cook Is.

Aitutaki

45

50

Stoddart &

(New

(motus)

Gibbs. 1975

Zealand)

Rarotonga
(motus)

49

ca. 60

Stoddart &
Fosberg, 1972

Gilbert Is.

Onotoa

60

50

83

Moul, 1957

(Rep. of

Tarawa

109

28

26

Catala, 1957

Kiribati)

Northwest

Kure

42

23

55

Lamoureux, 1961;

Hawaiian Is.

Clay, 1961

(USA)

Laysan

38

27

71

Ely & Clapp, 1973

Line Is.

Caroline

27

23(25?)

85(93?)

This paper

(Kiribati)

Christmas
(Kiritimati)

69

19

28

Garnett, 1983

Fanning

123

23

19

Wester. 1985

Flint

37

14

38

St. John &
Fosberg, 1937

Maiden

9

9

100

Garnett, 1983

Palmyra

58

21

36

Wester, 1985

Starbuck

7

4

57

Garnett, 1983
(incomplete,
little known)

Vostok

3

3

100

Clapp & Sibley,
1971b; Kepler,
1990c

Washington

91

25

27

Wester. 1985

Marshall Is.

Ailuk

56

26

46

Fosberg. 1955

(Fed.

Arno

125

40

32

Hatheway, 1953

States of

Enewetak

128

55

43

Lamberson, 1987

Micronesia)

Jaluit

288

55

21

Fosberg &
Sachet, n.d.

Jemo

34

17

50

Fosberg, 1955

Kwajalein

89

25

28

Fosberg, 1955,
1959

Lae

61

35

57

"

Likiep

91

31

34

"

Taka

23

18

78

Fosberg, 1955

Ujae

61

32

52

Fosberg, 1955,
1959

Ujelang

50

29

58

"

Utirik

55

26

47

'*

Wotho

40

28

70

"

Phoenix Is.

Kanton

164

14

9

Degener &

(Kiribati)

(Abariringa)

Gillaspy, 1955

129

18

14

Garnett, 1983

Birnie

3

3

100

Fosberg &
Sachet, (n.d.)

Enderbury

23

18

78

"

Nikumaroro

35

17

49

"

Orona

ca. 29

19

ca. 66

"

McKean

7

7

100

"

Phoenix

6

6

100

"

Manra

ca. 18

14

ca. 77

"

55

TABLE 4

continued

TotaF

No.

No.

Species

',

Island Group

Atoll

Species

Indigenous

Indigenous

Source

Society Is.

Tetiaroa

95

47

49

Sachet &
Fosberg. 1983

Solomon Is.

Ontong Java

1 50

58

39

Bayliss-Smith.
1973

Tokelau Is.

Nukunono

55

35

64

Parham. 1971

(N.Z.)

Tuamotu Is.

Raroia

121

39

32

Stoddart &

(France)

Sachet, 1969

Rarioa

135

54

40

Doty. 1954

Takapoto

106

33

31

Sachet. 1983

Outlyers

Clipperton
(U.K.)

31

14

45

Sachet. 1962

Oeno

17

14

82

St. John &

(U.K.)

Philipson. 1960

Wake

94

20

21

Fosberg &
Sachet. 1969

An updated version of Table 1 I, p. 105. Stoddart and Gibbs ( 1975).

:Number of species of those indigenous are not always comparable. Ferns are usually included. bu(
certain ornamentals may not be. Artocarpus, Morinda, and Pandanus may be indigenous, aboriginal
introductions, or both. Without its full scientific name, a species has an unknown biogeographical

status.

56

E g
#1

s s

i 3

Is

I I

« .a

1 "
— P

I 5
1 .

a j

.a .8

5 a. i
1 I

o a.

It3
» 1 B

It*

ss

2 <

I I*

X X X X X

m ci f*%

o o o o

s y S s

© © o c>

X X X X X

i/-i w-i vi >© t—

-

X

,3

^

<

,3

<•?

3

8

5

I

8 S * K 3

© © © d ©

X x X XXX XXXXXXXXXXXXXX XX

X XXXX XXXXXXXXXXXXXXXX XX

X X XXX XXXXXX X X XX X X X

X XXXXXXXXXXXXXXXXXXXXXXXXX

XX XXXX xxxxxxxxxxxxxxxxxxx

XXXX X XX

XX XX X

xx=xxxxSxxxxxxxxxxxxxxxxxxx

•OooO-cS10™™^^^

© © © r* © — O — — M M m — m

>©(■-]_ v©vi>©"T_,'©r-oooo'/ir->0'»«v>p-'0*or-,©'0<>r-r- * 2 2

^---r)nM-NM----NN------n-H

n © * — ei

ioei<Nmwi*f«*t(*»ci*rc>w^,'ir'i*'>,r

R? 8

""> -* "t P. ""( ^
O f*i ^ ^1 (N f*i

II

| |

8 .2

t.N

< £ ill J3

3 3-33

o | « § « o ^

oazSSuSzS

i 1 § -2 1 s

« s § .a a s.
s « * s. i « 5

s 8 * I s

SSps

3-!J

I ill

i" 1 J 5 i 1 1

1 u 2 m 03 Z a.

I Z "a 1

jlll
I a £ s

v h vi

5 *

3 S

"S 1 1

3 j 3

in

s

i

I

1

57

u

1 g

1 £•

il

3 |S

3 "I

2 3

.s I

5 _

il

a

6 o

§ I £

- 9

O O —

= =

«S ^ —

6 o

Ov — w

■_-'_- x C at

HS< o

< U U (J O a: U

uoS

< O U O 3 a! O

O U O < _J

uuu-r9o i

< 3 3 3 O OC 3 -JO

o o o o

< 3 3 3 <J OS 3

<a<KDa:>o

O O -1 o

V = = U as oi

<a:a:OO0£3O_}-J

<OOa;aiacOa:

< a: a: a: x x

> 3 O O X X

CJ 3 O X X X

U O X X X

V U X X

o

X X

_ -

>■§

Ml

§ fi

•S. S- « P i

Ml 8. >

J <3

1 1

il

4 -

u <

o%

j£i

1
&

£1

5 3

s

a!

m _ •?

8> a 3

II

ill

■C «

•S s
b s

.1

& g

II
1 5

I -g I

Ji|

u a

■S*i

S 8

■a* i

II.

.5 i ! .s S )
_! . £ Q z z

58

TABLE 7

Widths of pioneer herb mats on seaward- and
lagoon-facing shores, Caroline Atoll.

Average Width of Pioneer Herb Mat (m)
Bordering Sea Bordering Lagoon

Leeward Motus 18.5 4.2

(3-81) (0-28)
Windward Motus 36.0 0.9

(24-69) (0-3)

TABLE 8

Species-area relationships of six Pacific islands
with entirely indigenous flora.

Island

Area

No. Species

Maiden

39.3 sq km

9

Starbuck

16.2 sq km

6

McKean

57 ha

7

Phoenix

49 ha

6

Vostok

24 ha

3

Birnie

20 ha

4

Islands are arranged according to decreasing area.
Data is from Garnett ( 1983), Fosberg and Sachet
(n.d.). Clapp and Sibley ( 1971b), and pers. obs.

TABLE 9

Area of plant communities on the islets of Caroline Atoll.

Total

Area (ha)

Area (ha)

Unvegetated Habitats

41.39

Coral Rubble and Sand

41.39

Natural Plant Communities

261.41

Natural Herb Mats

67.73

Beach Scrub with Suriana

1.49

Pandanus Forest1

3.38

Tournefortia Scrub and Forest

125.25

Cordia Forest:

1.39

Pisonia Forest

62.17

Anthropogenic Community

96.14

Coconut Woodlands 96. 1 4

Total Area Above High Water 398.94

1 Pure Pandanus only. Also mixed with Pisonia. Tournefortia. and Cocos.

2 Cordia. where mixed with Pisonia and Tournefortia, is included in totals
for those forest communities.

59

TABLE 10

Stature and extent of Toumefortia in the major habitats
of Caroline Atoll.

Av. Hgt. (m) Av. Width (m) Toumefortia No. No.

(in) (m) Cover Motus Transects

Natural Herb Mat

Toumefortia
Scrub & Forest
Toumefortia -
Pisonia Forest

1.4

(0.3-1.8)

6

(0.3-15)

9.5

(5-15)

49
(3-198)

55
(2-287)

98
(8-284)

25

14

20

(5-95)

81

38

71

(5-100)

47

18

27

(5-90)

TABLE 1 1

Distribution of well-developed (>I0 m height) Pisonia forests on the motus

of Caroline Atoll.1 Motus and transects are arranged according to the

decreasing height of their Pisonia groves. Capitals indicate those motus

whose forests were felled for Cocos plantations from 1916-20.

Area of

Pisonia

Pisonia

Motu Area

Motu & Transect

Height (m)

(ha)

(ha)

PIG

21*

3.36

7.21

NAKE, Transect 4

20*

20.79

107.46

Booby

20*

0.12

0.84

NORTH PIG

20*

1.83

5.44

NORTH BROTHERS

18*

0.43

1.71

NAKE, Transect 3 (central)

15*

20.79

107.46

LONG, Transect O

15*

15.00

75.98

BROTHERS

15*

0.37

4.31

Ana-Ana

15*

0.93

2.16

Danger

15*

0.39

2.71

NAKE, Transect 2

14*

20.79

107.46

Bird

14*

1.70

4.05

WINDWARD, Transect 2

14-

2.97

11.42

Raurau

14*

1.07

3.48

CRESCENT

13-

0.51

3.10

Mannikiba, Transect 1

12*

1.13

21.49

Shark

12-

2.60

7.98

NAKE, Transect 3 (west)

12-

20.79

107.46

LONG, Transect 1 2

12-

15.00

75.98

Pisonia

11*

0.86

2.45

Matawa

1 1

0.07

1.71

Nautonga

11

0.02

0.34

NAKE, Transect 3 (southwest)

II-

20.79

107.46

Kimoa

11

0.59

1.80

Emerald

1 1

3.20

8.34

Eitei

1 1

0.38

1.42

LONG. Transect B

10-

15.00

75.98

LONG. Transect 8

10

15.00

75.98

NAKH. Transect 1

10-

20.79

107.46

WINDWARD. Transect 1

10

2.97

11.42

Blackfin

10-

0.41

2.62

NORTH ARUNDEL

1 It-

0.18

0.91

Toumefortia or Cordia mas be present, but sub-dominant to Pisonia.
* 90-100% canopy cover.
50 SO', canop) cover.

60

TABLE 12

Area and Dimensions of Pisonia grandis on Vostok, Flint, and five islets of Caroline Atoll.

Island/Islet

Area of

No. trees

Mean

Range of

Mean

Range

Mean

Range of base

Pi si >n ui

or main

Heigth

Heights

cbh1

of cbh

base2

circumferences

ha

trunks

(m)

(m)

(m)

(m)

(m)

(m)

CAROLINE

62.17

North Pig

1.83

25

19

11-21

221

110-359

261

205-470

Brothers

0.37

10

15

15

140

50-219

243

154-340

Pig

3.36

5

16

12-17

338

290-660

282

230-333

North Brothers

0.43

3

18

18

314

293-332

Long

15.00

3

15

15

414

282-500

Total for above

islets at Caroline

20.99

46

18

11-21

213

50-660

293

154-500

VOSTOK

1 3.5

58

18

10-25

218

67-510

FLINT

approx. 4

(fragmented) 20

17

8-30

160

60-200

598

100-1000

cbh = 'ircumference at 1 .5 m.
base = base circumference at 0.3 m.

TABLE 13

Number of trees and areas planted in Cocos on Caroline's islets
during the major planting era ( 1916-1920). also showing remnant Cocos data for 1990.

Approx.

1916-

% 1990

Approx.

Forest

Islet

Area Forest of Scrub Usable

No.

Area

%

Planted

Islet

Area

(ha)

for Cocos in 19901 (ha)

Cocos
Planted

Cocos2
(ha)

Cocos
1990

in Cocos

Town.

Pis.

Other

Cocos

Total

70 Yrs Ago

South

104.41

4.20

0

1.10

80.00

86.10

13.006

94.90

477

100%

Nake

107.46

30.65

20.8

9.41

5.75

66.61

10,544

76.97

6

100%

Long

75.98

32.20

15.00

-

2.40

49.60

1.343

9.80

3

20%

Tridacna(Al)1

9.08

7.97

0

0.18

8.15

910

6.64

0

82%

Arundel

7.34

4.36

0.95

-

0

5.31

646

4.71

0

89%

N. Arundel (A2)

0.91

0.33

0.19

-

few trees

0.52

69

0.50

0

100%

Brothers

4.31

2.00

0.37

-

0.0 1

2.38

315

2.30

0.2

97%

N. Brothers (A3)

1.71

0.68

0.43

-

few trees

1.11

180

1.31

0

100%

Pig

7.25

1.61

3.36

-

0.03

5.00

538

3.93

0.4

79%

N. Pig(A4)

5.44

1.31

1.84

-

0

3.15

402

2.93

0

93%

Crescent (A5)

3.10

1.56

0.51

-

0

2.07

228

1.66

0

80%

Windward (A6)

11.42

5.70

2.97

-

0

8.67

1,299

9.48
215.10

(1

100%

' "Usable area" does not include unvegetaied rubble or natural herb mats.
- Based on Caroline's planting densities of 28 x 28 sq ft (Young ca. 1922).
1 The "A" series of islet names are from Youna (ca. 1922).

TABLE 14

Species diversity in Pisonia forests of decreasing maturity. Caroline Atoll. Motus are
arranged according to the degree of coverage of their constituent Pisonia trees; within

these groupings.the motu order is dependent upon the total number of plant species

within this same habitat. Note that there is an inverse relationship between the purilv <<\

the true Pisonia forest and species diversity.

Canopy

Canopy

Av.

Total Ni

jmbers of Species

No.

Cover

Hgt.

No.

Trees

Shrubs

Herbs

Transects

(mi

Spp.

(inch Pisonia)

100% (Pisonia only)

13

1

1

0

0

1

100% (co-dominant present)

15

3.4

2

0

5

9

90-95%

10

5.2

5

1

7

9

50-909f

10

6.2

5

2

7

15

25-50%

7

6.2

4

1

6

5

<25%'

9

5.63

3

2

9

6

Motus with one

6

4.0

3

1

3

2

Pisonia tree only

South Island not included, as its Pisonia is too rare and fragmented.

62

t ropic of CancT

* .

HAWAII

20

equator

KINGMAN REEF
■ PALMYRA ATOLL

^ • WASHINGTON IS.

/ • FANNING IS.

KIRITIMATI

[CHRISTMAS IS. I

10

• JARVIS IS.

PHOENIX

group

TOKELAU

AMERICAN
SAMOA

KIRIBATI

Q

MALDEN
IS.

FILIPPO REEF

STARBUCK IS

o

-0
VOSTOK IS.

FLINT IS.

CAROLINE
ATOLL

^"AS ,s

10

FRENCH
POLYNESIA

WESTERN
SAMOA

NIUE

COOK ISLANDS

=>*::».>,

'K

«.<,"*,

tf

.TAHITI *

y is

u^

<©

r<?20
o

tropic of Capricofn

Fig. I. Line Islands: geographic location in the Pacific Ocean.

63

WINDWARD
ISLETS

SOUTHERN
LEEWARD
ISLETS

CAROLINE ATOLL

Fig. 2. Caroline Atoll. Republic of Kiribati, with newly-named islets. Based
on photos by the Royal New Zealand Air Force, Aerial Plan No. 1036
9d-h ( 1986). Photogrammetric Branch, Dept. of Lands and Survey,
RNZAF. New Zealand.

Fig. 3. Main marae on Nake Island, Caroline Atoll, based on a plan published
by the Solar Eclipse Expedition ( Holden & Qualtrough, 1 884 ). Figures
in the margins show side views of the peripheral blocks shown in the
plan. The two end walls are represented in ground plan.

Fig. 4. Caroline Atoll, as surveyed by John Arundel, 1883 This is still the
standard hydrological chart for the atoll (Admiralty Chart No. 9791.
Thoughquiteaccurate.it has never been used in scientific publications

Fig. 5. Caroline Atoll, as charted bj the Solar Eclipse Party, also in 1883
(Holden & Qualtrough. IS84).

M

ISO'lS' 150 I4W

I50°I3'

CAROLINE
ISLAND

From RO. S28.

Boat Entrance
Settlement/' j'

2 nautical miles

Fig. 6. Caroline Atoll, a modified version of the Solar Eclipse Party's map
(1883) as portrayed by Bryan (1942). Though highly inaccurate,
modifications of this map have been used in all publications since
Bryan (1942).

N

t

SURVEY
TRANSECTS

0 campsites

I

CAROLINE ATOLL

SCALE 1 y* 000
$00 0 900 IOO0 2000

I — I I I I 1

matari

Caroline Atoll

_kJ

Fig. 7. This map, by Clapp & Sibley (1971a), was based on Fig. 6.

PERIMETER
SURVEYS

CAROLINE ATOLL

SCALE 134 000
500 0 900 I0O0 2000

inn I 1 1 1

Fig. 8. Caroline Atoll: survey transects. The distance covered was 13.3 km. Fig. 9. Caroline Atoll: perimetersurveys. The distance covered was 19.3km.

•

- coral
(Acropora spp.)

- mollusks
(Tridacna maxima)

- coral

(Fungia granulata)

m* - Holothurioidea
(Ludwigothuria sp.

(ww - algae

(Porolithon sp.)

# - algae

[Halimeda sp.)

iMiiafei^*"^^"*

edge lateral

central

Coral limestone
Sand

lateral I edge

ZONES

Fig 10. 1 )iagrammatic representation of a portion of the outstanding Acropora-Tridacna reels connecting the islets Ana-Ana, Kimoa and Tridacna.
(Tridacna maxima), thickly studding the reel', attain densities of 80 per square meter (Sirenko & Koltun. Subehapter 1.4).

Giant clams

RARE AND
LOCALIZED PLANTS

|^< | PANDANUS

| m | SCAEVOLA

| • | PSILOTUM

["">"] HIBISCUS. THE

j | TACCA

fT~| TRIBULUS

[~T~] SPECIES A

| a | LEPIDIUM

| 0 I PHYLLANTHU

PHYMATOSOR
SCOLOPENDR

Pig I i i nnii' distribution map "l rare and/or localized plants on Caroline

Hibisi us liliaceus, I epidium bidenlalum, Pandanus tectorius.

Phyllanihus amarus, Psilolum nudum, 8 m vola terit ea, Sida fallax,

leontopelaloides, Thespesia populnea, Tribulus cistoides,

Ximenia americana, and Species V Psilotum ma\ still exist on Nake

Fig. 1 2. Transect distribution map of the fern Phymatosorus scolopendria on

Caroline Atoll

66

LEPTURUS REPENS

COCOS NUCIFERA

tj££3 Healthy Cocos

■'.'■] Dying Cocoa- Ipomoea Fore

"....] Mixed Forest

■ Isolated groves

CAROLINE ATOLL

CAROLINE ATOLL

SCALE i J* 00'.'
900 O SOO 1040 1000

'■■■■I I I I I

Fig. 13. Transect distribution map of the grass Lepturus repens on Caroline Fig. 14. Entire distribution map of the coconut Cocos nucifera on Caroline
Atoll. Arrows indicate areas of highest density. Atoll.

LAPORTEA RUDERALIS

ACHYRANTHES CANESCENS

CAROLINE ATOLL

Fig. 15. Transect distribution map of Laporlea ruderalis on Caroline Atoll. Fig. 16. Transect distribution map of Achyranthescanescens. Arrows indicate
Arrows indicate areas of highest density. areas of highest density.

67

BOEflHAVIA REPENS

J PISONIA
I 1 PISONIACOBOIA

~2 PI SONIA • TOURNEFORTI A
[^H Ml XEO FOREST

PISONIA GRANDIS

Fig. 17. Transect distribution map of pigvine, Boerhavia repens, Caroline Fig. 18. Entire distribution map of the buka tree, Pisonia grandis, Caroline
Atoll. Arrows indicate areas of highest density. Atoll. Arrows indicate forests from 10 to 21 m tall.

POBTULACA LUTEA

SURIANA MARITIMA

CAROLINE ATOLL

•CALI i J« ooo

900 0 SOO 1MO 1040

'■■■■' I I I I

Fig 19 Transect distribution map of the succulent herb Portulaca lutea, Fig. 20. Transect and perimeter survey distribution map of Suriana maritma.
Caroline Atoll. Arrows indicate pure Portulaca Hats. Arrows indicate areas of highest densit]

68

IPOMOEA MACRANTHA

Fig. 21. Transect distribution map of Ipomoeamacrantha. Entire distribution

is shown for South and Nake Islands. Arrows indicate areas having Fig. 22. Entire distribution map of CorJiu subcordata. Arrows indicate small
significant amounts of this vine. but monolypic. stands.

m TOURNEFORTIA

] SAVANNA"- HERB MAT

'"■•1 MIXED FOREST

HELIOTROPIUM \ ^ ~.% \ ^s; TOURNEFORT I A ■ MOR IN DA

ANOMALUM

TOURNEFORTIA ARGEN

CAROLINE ATOLL

Fig. 24. Entire distribution map of Tournefortia argentea. Because this shrub
Fig. 23. Entire distribution map of Heliotropium anomalum. Arrows indicate dominates Caroline^ woodlands, there are no individual arrows to

areas of highest density. indicate areas of high density.

69

MORINDA CITRIFOLIA

CAROLINE ATOLL

SCALE 1:34000
0 0 S00 I0O0 2000

'■■■■' 1 I I I

100

in

80

C3

z

<

50

30

20 -

10

TOURNEFORTIA PISONIA
NATURAL FORESTS

MIXED COCOS HOUSE SITE
ANTHROPOGENIC FORESTS

Fig. 25. Transect distribution map of Morinda citrifolia. The outlined area on
rridacna Islet (northeast of South Island) encloses Toumefortia-
\4orinda forest. Arrows indicate areas of highest density.

Fig. 26. Evidence for the indigenous status of Morinda citrifolia on Caroline
Atoll : percentage cover on transects within natural and anthropogenic
forests. Morinda occurs on 30 ( 7796 ) motus. never in a "planted"
situation.

70

LEFT CIRCLE

MAJOR PLANT COMMUNITIES
UNVEGETATED
HERB MAT

RIGHT CIRCLE
PERCENT INDIGENOUS

^1

TOURNEFORTIA
PISONIA

cocos

SURIANA
PANDANUS

INDIGENOUS

ABORIGINAL INTRODUCTION

RECENT INTRODUCTION

Wll HOUSE SITE

MIXED FOREST WITH COCOS

NODDY ROCK
0.02 ha

0.005

0.015
(75%)

SKULL ISLET 0.005

0.02 ha (5%)

^s^,^ 0.019

40l§&i0? (95%)

MOTU ATIBU
0.02 ha

0.004
(2%)

r

0.02

I* (98%)

REEF-FLAT ISLET
0.09 ha

fc

0.02
(22%)

Fig. 27. Plant communities and amount of indigenous vegetation on motus less than 0.I ha, Caroline Atoll. The left "pie" depicts the relative amount of amotu's
total surface area covered by each plant community: numbers indicate actual area in hectares. The right "pie" depicts the numbers and percentages of
indigenous and anthropogenic species per motu. Data is based on the vegetation maps for each motu (Figs. 37-57) and Tables 2 and 9.

71

1.01 (5%

AZURE ISLE
0.20 ha

o.ii m

(55%H

0.04
20%)

MOTU NAUTONGA
0.34 ha

0.13(38%)

0.02 (6%)

(11%)

0.08 (24%)

SCARLET CRAB ISLET
0.46 ha

0.11 (24%)

0.18(39%)

FISHBALL ISLET
0.57 ha

0.16(28%)

0.17(30%)^^^

0.01

0.03

MOTU KOTA
0.64 ha

0.29
(45%)

Fig. 28. Plant communities and amount of indigenous \ egetation on motus 0.2 to 0.7 ha. Caroline Atoll. See Fig. 27 for explanation of the figure.

72

BOOBY ISLET
0.84 ha 0.12
(14%)

o.27 m

(32%) \

0.41
(49%)

BOSUN BIRD ISLET
0.86 ha

0.25

(30%)

^0.20

(23%)

NORTH ARUNDEL ISLET
0.91 ha

0.33

(36%)

0.27
(30%)

10

(100%)

Fig. 29a. Plant communities and amount of indigenous vegetation on motus 0.8 to 25.0 ha, Caroline Atoll. See Fig. 27 for explanation of the figure.

73

MOTU MOUAKENA
1.00 ha

0.26

(26%)

0.38
(38%)

0.36
(36%)

MOTU EITEI
1.41 ha

0.38
(27%)

0.38

(27%)

0.57
(40%)

CORAL ISLET
1.07 ha

0.13(8%) 0.07
(4%)

MOTU MATAWA
1.71 ha

0.68
(40%)

0.78

(46%)

0.52
(30%)

0.07 (4%)

(26%)

Fig. 29b. Plain communities and amount of indigenous vegetation on motus 0.8 to 25.0 ha, Caroline Atoll Sec Fig. 27 for explanation of the figur

74

NORTH BROTHERS ISLET
1.71 ha

0.43

(25%)

0.48
(28%)

0.68
(40%)

MOTU KIMOA

1.80 ha

0.59

(33%)

0.50
(28%)

12
(100%)

LONE PALM ISLET
1.99 ha

0.95

(48%)

0.39
(19%)

0.61
W (31%)

MOTU ANA-ANA
2.16 ha

0.93

(43%)

47
(22%)

(7%)

0.21 (10%)

Fig. 29c. Plant communities and amount of indigenous vegetation on motus 0.8 to 25.0 ha, Caroline Atoll. See Fig. 27 for explanation of the figure.

75

0.69
(28%)

1 (7%)

BLACKFIN ISLET

2.62 ha

0.78
(30%)

0.63

0.80
(30%)

1 (11%)

DANGER ISLET
2.71 ha

0.77
(28%)

10
(100%)

CRESCENT ISLET
3.10 ha

1.56
(50%)

0.61
(20%)

10
(100%)

Fig. 24d. Plant communities and amount of indigenous vegetation on tnotus 0.8 to 25.0 ha, Caroline Atoll. See Fig. 27 lor explanation of the figure.

76

MOTU RAURAU 1.71

3.48 ha /""" ~~"~~\ (49%)

(31%)%iixll^v/

0.04 0.66

(1%) (28%)

(10%)

BIRD ISLET
4.05 ha

0.22 (6%)

0.41 (10%)
0.01 (0.03%)

BROTHERS ISLET
4.31 ha

0.58

(13%)

0.37

(8%)

0.01
(0.2%)

NORTH PIG ISLET

0.97

5.44 ha

^ 1 8%)

1.31 /

(24%)/

l_^-rC\\\v

Ipttjll
x\ 1 1 ??

**X*X"XvX

•X'XvffiasS©' \£-*t /o)

1.84
(34%)

Fig. 29e. Plant communities and amount of indigenous vegetation on motus 0.8 to 25.0 ha. Caroline Atoll. See Fig. 27 for explanation of the figure.

77

PANDANUS ISLET
7.20 ha

2.30
(33%)

2.28
(32%)

1.61
PIG ISLET (23%)

7.25 ha

0.90

k(12%)

'":lpf
?*.'■&/ 1 .35

(46%) ^-j-ijg*:

0.03 (0.4%)

ARUNDEL ISLET
7.34 ha

4.36
(59%)

SHARK ISLET
7.98 ha

2.92

(37%)

94(12%)

0.12(1%)
2.60(33%) "

Fig. 291". Plant communities and amount of indigenous vegetation on motus 0.8 to 25.0 ha, Caroline Atoll. See Fig. 27 for explanation of the figure.

78

EMERALD ISLE
8.34 ha

3.20
(38%)

1.5(18%)

1 (8%)

TRIDACNA ISLET
9.08 ha

7.97
(88%)

0.18(2%)
10.21 (2%)

f 0.72 (8%)

11
(100%)

WINDWARD ISLET
1 1 .42 ha
5.7
(50%)

2.97
(26%)

1.6
(14%)

11
(100%)

MOTU MANNIKIBA
21.49 ha

15.69
(73%)

1 (8%)

2.60

(12%)

f 2.04(10%)

0.03(0.1%)
1.13(5%)

Fig. 29g. Plant communities and amount of indigenous vegetation on motus 0.8 to 25.0 ha, Caroline Atoll. See Fig. 27 for explanation of the figure.

79

LONG ISLAND
75.90 ha

32.2

(42%)

18.6(25%)

SOUTH ISLAND
104.60 ha

80.8 '
(77%)

4.2 (4%)
1.1 (1%)

13.6(13%)
gllf 4.9(5%)

1 (5%)

NAKE ISLAND

91.71 ha

^^^ 18.3

30.65 /

J'.W(20°o)

(33%) /

KjIvXvX

||W6.8(7°o)

\jx:x:x

;x3^ 3.38 (4%)

20.8^

£3=^6.03 (7%)

(23%)

5.75

(6%)

Fig Kl. Plant communities and amount of indigenous vegetation on motus over 25.0 ha. See Fig. 27 for explanation ol the figure.

80

_l 2

a

O

=> 20 -

O

2

O

I

1 1

11 1 1

he

II

^ 1

_ 1

~\ r^

1 1

-

■
■

"

■ ■
■ ■ ■ ■ *

■

"

■ ■

■ ■ ■ • X ■

■ ■■ ■

■

"

■
■ ■

■ ■ ■

-

■

■ ■ ■

*

■ ■

-

• •

•

-

•

• •

• •

• •

•
*•
• • • •
* •• • • •
• • •

•

*

•

1

•
1

J.

'

0

0 01

AREA OF MOTU ( HA )

Fig. 3 1 . Total numbers of plant communities ( upper graph ) and species ( lower graph ) in relation to islet area, demonstrating plant succession on the differentsized
islets encircling Caroline's lagoon. Roman numerals refer to size classes of the motus: I = < 0.2 ha, II = 0.2 to 0.7 ha, III = 0.8 to 25.0 ha,
IV = > 25.0 ha. The break between II and III marks a substantial increase in the diversity, area coverage, and height of the forest ecosystems. Data is
based on Table 5.

22

20 .

18

o

16

12

10

a 8

O

z

< 6

O

50 100

200

300

400

500

600

700

800

WIDTH OF MOTU (m)

Fig. 32. Maximum heights of Pisonia forests in relation to width of the motus. Stars represent forests with 90-100% canopy cover; dots represent forest or scrub
with less than 90% cover.

81

20 .

16 .

I 12 -

O

I 10

<
u

2 .

AREA OF MOTU (HA)

Fig. 33. Maximum canopy heights of Pisonia forests in relation to islet area. Stars represent forests with 90-100<7r canopy cover; dots represent forest or scrub
with less than 909c cover.

LAGOON

HERB
l_MAT I

I I TOURNEFORT

IA SCRUB

llo3m)

|TOURNEFORTIA
SCRUB
Itournefortia FORESTI |,o5ml

to9m

Itou
est!

HERB

MAT

I

WIND

OCEAN
REEF FLATS

^i is^^^^i jfisg la^^SaiaMfeaga gals gfesge SassS

^

fJjTJj COBAL RUBBLE
COARSE SAND

-* y* S RED FOOTED

GREAT FRlGATEBIRO

WHITE TERN

REEF HERON

%

LACK NODDY

BROWN NQDOr

BRISTLE -TMIGHED CURLEW

POL TNE SIAN HAT

i i i S. hematic profile through Arundel Kiel, recovering from disturbance over 60 years ago, showing natural herb mats, Tournefortia scrub and forest, and
five species of breeding seabirds. Vertical height is exaggerated,

WIND

TOURNEFORTIA

forestt —

PISONIA FORESTliolSml

Itournefortia HERB

"-| SCRUB | MAT |

OCEAN
REEF FLATS

OL VNESIAN RAT

Fig. 35. Schematic profile through Long Island. Tr. O. Although Long Island has been formed in the recent past by a merger of five smaller islets, this section
of the islet is very mature, containing natural herb mats, Toumefortia scrub and forest, and tall Pisonia forest. Seven species of seabirds breed. Vertical
height is exaggerated.

oc

REE

iCEAN |_HERb[tOURNEFORTIA
F FLATS I MAT! FOREST

WIND

SOUTH ISLAND.
TRANSECT 2

DYING COCOS-IPOMOEA FOREST
(to 18m)

BEACH SCRUB
COCOS |WITH SURJANAItoSm

— Lag6on

plantation! f

( to 21m)

I n j — l^sSg&k .»■">&

J — i

EgJsSfl CORAL RUBBLE
FINE SAND
HUMUS. GUANO. RUBBLE

BROWN " ■;■!■"

H „

SHOREBIRDS

REEF HERON

1 BRISTLE -THlGHED CURLEW ^^S LON

!«t^^H POLVN

ESIAN R«T

COCONUT CRAB

G ■ TAILED CUCKOO

Fig. 36. Schematic profile through South Island, where 77% of the land surface is covered with Cocos forests, primarily in a dying state. Vertical height is

exaggerated.

83

Pg^j RUBBLE
fm HERB MAT
i 1 COROIA

H TOURNEFORTtA
W3 PISONIA

I 1 COCOS

|..»| PANDANUS
IV | MIXED

BEACHROCK

i

Fig. 37. Nake Island: vegetation and physiography.

PLANT COMMUNITY

PLAHT SPEC

ES

Bo«i*i»via ••p«»*

C.co. »ur,..-.

Coiltil luttnidlll

.■v*

i»nw>n«» nt«e'««l«

-

'.:,. »r.. *,»,,

SUBSTRATES

ma— udo • •-*• - -.1

Mumul )„)■« rubblf

g^3 RUBBLE

j^P HERB MAT

f "| TOURNEFORTIA

m PISONIA

[^j COROIA

F*Tj COCOS

^s^— BEACHROCK

LONG ISLAND

Fig. 38. Long Island: vegetation and physiography.

AHCIBM1 r.H»NN( i

o

Oifi CHANMII

1

«, , „, ■ -,BW ,, •

*

II»r«tt 0< miMI b U

1. 1 *

01 ■ I AMC I l»|

I ig (9. Long Island: north-south transect showing division into former islets, floristic composition, relative abundance of plant species, degreeofspeciesoverlap,
and canopy heiehts. Vertical heighl is exaggerated. The exact locations oi the formerly more extensive Cocos plantations are unknown

84

Tournefortia- Piaonla

Tournaforlla rubbli

Boerhavla repens

Cocos nuclf era

Cordia subcordata

H eliot ropiuni anomalum

Lepturus repens

Morlnda citrifolla

Phymatosorus scolopendrli

Pisonla grandis
Portulaca lutea
Tournefortia argantea

I 20

O 10

z
<
o

I

LAGOON

[laal

100 200

TRANSECT WIDTH Iml

OCEAN

I windward I

Fig. 40. Long Island: east-west cross-section through Tr. C, a former inter-islet channel, showing floristic composition, relative abundance of plant species, degree
of species overlap, and canopy heights. Vertical height is exaggerated.

Opan Tournafortla Scrub

Flats sand.rubble rubble

Boerhavla repens
Cocos nucifera
Hehotropium anomalum
Laportea ruderalis
Lepturus repens
Phymatosorus scolopendria
Portulaca lutea
Suriana maritima
Tournefortia argentea

I

a

oio .

LAGOON

laa ■

100 200

TRANSECT WIDTH |m|

1

_

1

1

—

1

1 E=

1

I

1

1

1

300

OCEAN
Iwlndward I

Fig. 41. Long Island: east-west cross-section through Tr. 8. which passes through mature interior Pisonia forest of largest of Long's coalesced motus. Data
includes floristic composition, relative abundance of plant species, degree of species overlap, and canopy heights. Vertical height is exaggerated. Note
the absence of low vegetation on the leeward shore.

85

Fig. 42. Vegetation and physiography of Windward Islet no. 1: Bo'sun Bird Islet. Scale is larger than on the vegetation maps of other islets.

RUBBLE
HERB MAT

FTT^ RUBBLE

^M HERB MAT
fl CORD. A

"1 TOURNEFORTIA
Hi "'SONIA

~1 TOURNEFORTIA -CORDIA

WINDWARD ISLET

CRESCENT ISLET

MOTU ATIBU

'J )

L

l ig l ! Vegetation and physiography of Windward Islets nos. 2, ' and 4:
Windward and Crescent Islets, and Mom Atihu ("Coral Rubble
Kiel"). Atibu appears to have been severel) damaged during the
February 1990 storm.

\ CH con°"

Y | | TOURNEFORTIA

P^l PISONIA

| . | COCOS

I • I SURIANA

NORTH PIG ISLET

SKULL ISLET

NORTH BROTHERS
ISLET

BROTHERS ISLET

Fig 44. Vegetation and physiography of Windward Islets nos. 5 through 9:
North Pig, Pig. Skull. North Brothers, and Brothers Islets, Note the
reels extending westward into the lagoon.

86

Tournetortia Pisonia Tournelortla Herb Mat

Achyranthes canescens
Boerhavia repens
Heliotropium anomalum

Ipomoea macrantha
Morinda citrifolia

Pisonia grandis

Phymatosorus scolopendria

Portulaca lutea
Tournetortia argentea

x
a

z
<
u

30 _

20 -

10 -

LAGOON
Heel

100

TRANSECT WIDTH Iml

200

OCEAN
I windward

Fig. 45. Pig Islet: east- west cross-section through center of islet. Dataincludes floristic composition, relative abundance of plant species, degree of species overlap
and canopy heights. Vertical height is exaggerated. Pig's profile is especially symmetrical. It is remarkable that this islet was totally felled for coconuts
in 1920.

Boerhavia repens
Cocos nuci tera
Heliotropium anomalum
Laportea ruderalis
Lepturus repens
Morinda citrifolia
Pisonia grandis
Portulaca lutea
Sun ana maritima
Tournetortia argentea

LAGOON
Heel

100

TRANSECT WIDTH Iml

200

OCEAN

I windward I

Fig. 46. Brothers Islet: east-west cross-section through center of islet. Data includes floristic composition, relative abundance of plant species, degree of species
overlap and canopy heights. Vertical height is exaggerated. Note the central monotypic stand of Pisonia forest. This islet's forests were totally felled
in 1920. o7

f£23 rubble

f^^ HERB MAT

| ) TOURNEFORTIA

|^9 P'SONIA

n~l cocos

Acropora - Tndacna

H«*f ^/

gy^] RUBBLE

^| HERB MAT

| ) TOURNEFORTIA MORINOA

|» «) SURIANA

Acropora - Tndacna
Raaft

<V

TRIOACNA ISLET

I I I U

Fig. 47. Vegetation and physiography of Windward Islets nos. 10 through 12: Fig. 48. Vegetation and physiography of Windward Islet no. 13: Tridacna

Noddy Rock. North Arundel, and Arundel. See text for explanation of Islet. The best quality Acropora-Tridacna reels extend clear across

the relatively small amount of Pisonia cover (Description and Ecology the lagoon from this motu. See Desc. and Ecol. of the Motus Section

of the Motus Section). for explanation of unusual forest cover.

Tour net or t la

Tournefortia

Morinda Tournefortia Herb mat

Boerhavia repens
Heliotropium anomalum
Laportea ruderalis
Lepturus repens
Morinda ci t ri folia
Phymatosor us scolopendria
Pisonia grandls
Portulaca lutea
Tournefortia argentea

O

x

10 -

o
z
<
o

0

LAGOON

100

TRANSECT WIDTH Iml

1 1

1

1

1=

1
1

1

■ '

200

OCEAN
I wl ndward I

1 i: l'i 1 ridacna Islet: cist west cross-section through lower center of motu. Data includes Holistic composition, relative abundance of plain species, degree
ol species overlap and canopy heights. Vertical height is exaggerated Note the absence of well-developed interior forests, unusual for a motu of this
si/e (Description and Kcologv of the Motus Section)

88

g£3 RUBBLE

^B KERB MAT

| | TOURNEFORTIA

\TT^ COCOS

P 7J DYING COCOS IPOMOEA

| * «J SURIANA

1"»^ PANDANUS

~Z^~ BEACHROCK

Fig. 50. South Island: vegetation and physiography. Note the accepted landing route across its leeward reef flats.

E •

OuUr B«iCh Scrub X *

r

Boerhavirt repens
Cocos nucltera
He 1 1 ol ' opium anomalum
Ipomoea macrantha
Lapor tea ruder alls

Lepturus repens
Monnda citrtfolia

Phymalosorus scolopendria

Pisoma grandls

portulaca lutea

Sur lana man t ima

Tournefortia argentea

LAGOON
( lee )

240 360

TRANSECT WIDTH (m)

720

OCEAN
(pari windiwcrd )

Fig. 5 1 . South Island: distribution and abundance of plant species along Tr. 2, which runs at an angle of 60° from the lagoon to the south shore through the western
center of the islet. Data includes floristic composition, relative abundance of plant species, degree of species overlap and canopy heights. Vertical height
is exaggerated. Horizontal scale is half that of the profiles from smaller motus.

89

fTTj BUBBLE

|^ HERB MAT

"2 TOUHNEFOHT

J COHOIA

[*"3 PISONIA

|»1^ PANDANUS

f~r*1 cocos

PANDANUS ISLET

t£J.:'I RUBBLE

|B MERB MAT

I 1 TOURNEFORTIA

rT*^ PISONIA

!• • -1 cocos

^ BEACMROCK

Fig. 52. Vegetation and physiography of the 7 South Nake Islets: Pandanus,

Danger, Booby, Coral, and Lone Palm Islets, MotuKota( "Red-footed Fig. 53. Vegetation and physiography of Central Leeward Islet no. 1: Motu

Booby Islet"), and Motu Mouakena ("Masked Booby Islet").

Mannikiba ("Seabird Islet").

)"3 RUBBLE

| HERB MAT

[ ) TOURNEFORTIA

1 .J PISONIA

^■■^ PANDANUS TOURNEFORTIA

\: :'■: I coroia

|" • -1 cocos

C3 RUBBLE

■ HERB MAT

| 1 TOURNEFORTIA

003 CORDIA

[■ • «1 COCOS

!■»■! PANDANUS

[~3 PISONIA

r~7~| SURIANA

• Urt 400

i i I Vegetation and physiography ol the Central 1 eeward Islets nos 2 Fig. 55 Vegetation and physiography of the Central Leeward Islets nos. 5

through 4: Blackfm Met. Mom Matawa ("Fair} Tern Met"), and through 11: Shark and Scarlet Crab Islets, Motu Nautonga ("Sea

Emerald Isle Cucumber Islet"), Azure Isle, Reef-flat, Bud and Fishball Islets

90

Achyranthas canascana

Bonhivia rep«ns
He I iot r opium anomalum
L aportea rudaralis
Lepturus repans
Morinda citrif olia
Portulaca lutaa
Tournafortia argantaa

Hi

■

_ i

E

h-

a

X

ho

z
<
a

■
■

J i — — —i— A 1

0

OCEAN

50

100

TRANSECT WIDTH ( m )

LAGOON
Iwtftdward I

Fig. 56. Fishball Islet (no. 1 1. Centra] Leewards): east-west cross-section through the center of this young motu, which exhibits early stages of geological and
biological evolution. Data includes floristic composition, relative abundance of plant species, degree of species merlap and canopy heights. Vertical
height is exaggerated.

f-j,;^ BUBBLE

Q^| HERB MAT

~] TOURNEFORTIA

r~~i cordia

"2 PISONIA

[. . .| cocos

|~j~| HOUSE SITE

[ ■ | PANDANUS

, | • | SURIANA

MOTU RAURAU

MOTU E1TEI

PISONIA ISLET

MOTU KIMOA

MOTU ANA ANA

N

i

SOUTHERN LEEWARD ISLETS

Fig. 57. Vegetation and physiography of the 5 Southern Leeward Islets: Motus Raurau ("Blue-gray Noddy Islet"). Eitei ( "Frigaiebird Islet"). Pisonia Islet. Kimoa
("Rat Islet"), and Ana-Ana ("Anne's Islet").

_ lb

PI. 1 . A dawn view of Caroline as seen from the ocean in 1 988 but virtually identical with that seen by the atoll's Western "discoverer.'
de Quiros, in 1606.

PI. 2a. A clearing on South Island from which the Solar Eclipse Party made their observations in 1 KX3. Today the area is covered with
dense Cocos forest (from Holden & Qualtrough, 1884).

93

PI. 2b. One of the three European-style houses that have ever been built on Caroline, drawn in 1883 (ibid.).

■■■■ "

PI. 3a. An artists very lice rendering of Caroline In 1883 (from Holden & Qualtrouith. 18841.

94

I.IIUIIl' MHI.l .'I llll SOI'TBRR* Wll <>» < AKfU.IM. iKLisrm

a

..<•<"

V

7^.

,% *:

j «* **

* *

tif ;.■■■!.. r *tt'i.

PI. 3b. Map of the "settlement" on South Island, as drawn by the Solar
Eclipse Party (ibid.).

■B3*

. 4. Two lagoon views a century ago along the north coast of South
Island (from Holden & Qualtrough. 1884). Compare these
drawings with Plate 24.

95

PI. 5a.b. Motu Ana-Ana, unnamed in 1883 (Holden & Qualirough,
1884) but appearing virtually identical then and today
( Plate 8 1 ). Below is a substantial Tournefortia tree along South
Island's lagoon edge.

PI. 6. Base camp, northwest peninsula. South Island.

46

*

PI. 7. The junior author beside an entangling thicket of Ipomoea
macrantha, dying Cocos-Ipomoeaforest, interior South Island.

97

PI. 9. Relocating camps with inflatable canoe. Cameron Kepler and Katino Teeb'aki, Bo'sun Bird Islet.

PI. 10. BI.Kk tippped reel shark [Carcharhinus melanopterus), a numerous and aggressive denizen "l Caroline's lagoon. Note the
abundant set cucumbers i / udwigolhuria sp. ).

98

PI. 11. Jagged, upraised reef ifeo or champignon), leeward reef, southwest Nake Island. Note the wide reef Hats.

PI. 12. Floating a small boat across the soutwest reef Hats in calm weather from the "boat entrance" to the "landi
the wide reef flats. The Akademik Korolev drifts offshore.

' on South Island. Note

99

4fefe

PI, 1 3. Beach crest, sand) rubble, seaward moat, and narrow reef flats off northeast Long Island.

PI. 14 I Ktensive lagoon reef flats south of Arundel Island on the windward side. Note the Five southern leeward islets in the distance.

100

PI. 15. An incipient motu. barely connected to Motu Mouakena's southern shore. See also Fig. 69.

PI. 16. Windward reel and peaked beach crest on South Island, with the recently wrecked remains of a 26-foot sloop.

I

;.-^>

PI. 17. S

Successive ridges of coral rubble forming extensive gravel flats, northeast Nake Island.

■

PI. 18. Channel between the two northern islets. Long and Nake. Note the mixed forest with Coi os and Pandanus.

102

PI. 19. Noddy Rock (0.02 ha), an emergent reef platform along the windward reef flats. A northward view.

PI. 20. Windward beach. Long Island, showing wide rubble flats inland of the beach crest, rimmed by oceanic flotsam and jetsam.

103

Beachrock, wesi shore. South Island
blinding coral beach.

Russian vessel Akademik Korolev drifts offshore

PI. 22. A large coconut crab (Birgus (afro) shelters in a subterranean cavit) in the ft

104

..

PI 23. Sand> Inlet, a filled-in portion of the lagoon, extends its fishhook-shaped mudflat 300 m northward into Nake's landmass. Here
grow the healthiest and most productive Cocos on Caroline. Note the bristle-thighed curlews in the foreground.

PI. 24. South Island's pure ( '.>< OS plantation, looking west along the lagoon. This extensive grove has now obliterated all traces of the
former "settlement" (Plates 2-5).

105

PI. 25. Brilliantly-colored, crystalline lagoon waters adjacent to Emerald Isle (Central Leeward Islets) are studded with twisted reef
configurations and sandy channels.

-.« • -.

-

■

PI. 26. An impressive cross lagoon reel ol Acropora sp. coral and Tridacna maxima clam shells joins Tridacna Islet with Motu Kimoa.

106

PI. 27. Cordia forest (to 1 2.6 m tall ), Pig Islet.

PI. 28. Sand, silt, rubble, and hardpan mingle on the upper
reaches of Long Island adjacent to the lagoon.

PI. 29. Caroline's best sandy beach lines the lagoon shore of Shark Islet. The fine sand is overlain with sparkling pink granuli

J29[tee^4fe^i

I 'I JO Sheltered baj Brothers Islel Raurau Islet lies across the lagoon. Note the sparse herb mat and silt] shallow waters

108

PI. 31. Narrow lagoon beach lined with Tournefortia scrub. Blackfin Islet (Central Leewards).

PI. 32. Recent sand additions to South Island's northeast point, which is in part covered with excellent natural herb mats and healthy
Suriana scrub (right).

109

PI. 33. An old interislet channel (Transect C, Long Island), now filling in with herbs, Toumefortia scrub, and Cocas. Note the nesting
masked boobies in middle right.

1

&*'

^

T$

PI. 34. A large clearing within dying ( o< 05 /pom
Phymatosorus and Ipomoea

st, interior South Island. Note the

» v vi «%Bk...

prolific mats of Boerhavia interlaced wuh

110

PI. 35. Pandanus forest, south Nake.

if A * 'j-

PI. 37. Mixed fores! \uth Cocos, southwest Nake Island.

7 .-» -

i'l *s Orange, scarlet, and green phalanges of Paw/anus rest on a clump ol Portulat a. The ubiquitous Coenobitaperlatus forage on their
siriui;\ flesh.

PI. 39. Inner edge of lagoon. South Island, 1988. Cocos is progressively shading out the beach scrub with Suriana maritima.

PI. 40. Inner edge of lagoon. South Island, 1965, taken from approximately the same location as Plate 39. Note the greater extent of sand
and Suriana coverage above high water than today, due to less encroachment and shading by the palms.

113

PI. 41. Heavy understory of Achyranthes canescens, Boerhavia repens, and Phymatosorus scolopendria in a clearing adjacent to Pisonia
forest. Pic Islet.

I'l 42. Boerhavia Fruits on feathers and hill of a groat frigatebird.

114

PI. 43. Inside a mature Pisonia grandis forest, interior Nake Island. Note
the barren, dark aspect, virtually devoid of undergrowth except
root suckers.

*7 PI. 44. Fringe of Suriana, northeast point. South Island.

L^MflA*

_

PI. 45. Well developed natural herb mat. primarily Heliotropium anomalum and Lepturus repens. Seattered Tournefortia form
"savannah." Here the sandy soils are conducive to the growth of lush Heliotropium, northeast point. South Island.

PI. 4ft. Detail ol Heliotropium anomalum, with remains of the first evidence of tropicbirds on Caroline. Skull Islet.

116

V

!

PI. 47. Toumefortia scrub, fringed by a natural herb mat, and occupied
by a colony of sooty terns. An old interislet channel, northern
Long Island. Note the nesting red-footed boobies.

PI. 48. Toumefortia- Morinda forest, with nestingbrown noddies, interior
Tridacna Islet.

117

PI. 44. Skull Islet (0.02 ha), with brown nodd) tems, looking east to the

windward reef.

PI 50. Anartistic impression of&Paihldiiin prove. South Island, in 1 SS3
(from Holder & Qualtrough, ISS4).

118

PI. 51. Red-footed booby in Tournefortia scrub. Motu Raurau.

PI. 52. Mature Pisonia grandis canopy with incubating black noddies and a white tern. Pig Islet. With a canopy height of 2 1 m. this was
the most majestic interior forest on the atoll, although it cannot be older than 65 years.

19

Azure Isle (Central Leewards) — an example of a motu containing a single Pisonia tree. Note the narro
interislet channel. View east from Motu Nautonga. with Brothers Islet in the distance.

hut Mill shark-patrolled.

PI. 54. Caroline's sole clearing, with Tahitian-style huts. Motu Ana-Ana.

120

PI. 55. While tern "nest" balances precariously on the upper midrib of a coconut frond.

PI. 56. Average-sized coconut crab iBirgus lalro) on Caroline. Compare its size with the coconut.

121

PI. 57. Piles of fibrous shavings — coconut crab sign.

I 'I ss Beachrock al the lower end of Long Island, typicallj found al low water.

122

"• A

PI. 59. Sooty terns (Colony A) — approximate^ 128,000 pairs occupy a scrubby swath of vegetation in the north-central portion of Long
Island, the location of an old interislet channel. Note Cocos on the right and a mound of Pisonia at rear.

PI. 60. Windward Islets numbers 5-9 (left to right): North Pig. Pig. Skull (not visible). North Brothers, and Brothers.

123

PI. 61. North Pig (left) and Pig (right) Islets: a southeaster!} mow across the coral-studded lagoon from Shark Islet. Note the exp
central Pisonia forests.

PI, 62. Arundel Islet (foreground), looking south southwest across IYidacna Islet toSouth Island. Distant Mom Ana- Ana lies on the right.

124

■■ M matter ■=^i»' ■ * *>■* r- nls^%. « . r * TV

PI. 63. Detail, Tridacna maxima reefs, lagoonside of Tridacna Islet. This dense aggregation of giant clams amassed up to 80 per square
meter.

PI. 64. View of Pandanus Islet (center) west down the channel separating Nake (right) and Long (left) Islands.

125

. 65. Danger Islet (South Nake number 2), looking due west across the shallow upper lagoon from Long Island.

PI. 66. South Nake Islets numbers 3 6 1 right to left): Booby. Coral. Lone Palm, and Kota. Westerly view across the shallow upper lagoon
from Long Island

126

PI. 67. Lone Palm Islet (South Nake number 5): a southerly view from the shallow tidal flats of Coral Islet. Plate 68 provides a more
detailed view of the lower part of the chain.

PI. 68. South Nake Islets numbers 2-6 (right to left): Danger. Booby, Coral, Lone Palm, and Kota, from a boat. This view shows better
detail of the lower part of the chain than Plate 67.

127

PI. 69. Motu Mouakena (South Nake number 7), with its sandy cay. A westerly view from the upper lagoon. Compare with Plate 15.

PI. 70. North end, Motu Mannikiba "Seabird Islet" (Central I ee wards number I ), showing mounds ofPisonia and a closer Cocos grove
Note coral "mushroom" in laj n

12S

j**4MMfl&- V*

PI. 71.

of Pisonia.

, looking west along Transect

Tournefortia, then a distant patch

r ?>*''•

■

y-VV^

s£

s:ip;:x£-v

PI. 72. Motu Mannikiba, looking west along Transect 2. Low Tournefortia scrub covers a coarse, rubbly substrate which was probably
once an inter-slet channel.

129

PI. 73. Blackfin Islet (Central Leewards number 2) lies to the left of the ship. View west-southwest from the tip of Long Island.

PI. 74. A stunning "blue hole" within the lagoon ofl the southwest point ol Emerald Isle (Central Leewards number 4)

130

PI. 75. Mixed Pandanus-Tournefortia fores!, interior Emerald Isle.

PI. 76. Emerald Isle, looking west across hardpan and open lagoonside
scrub to a densely vegetated interior.

•>i.'

131

PI. 77. Shark Islet (Central Leewards number 5): view across crystalline shallow waters to Caroline's sole sandy beach.

PI. 78.

Southern I eeward Islets numbers I 5 (right to left): motus Raurau .mil Eitei,
northwest from 1 ridacna Islet.

Pisonia Islet, motus kimoa and Ana Ana View

32

PI. 79. View of Motu Raurau ("Blue-gray Noddy Islet") fromMotu Eitei ("Frigatebird Islel" >. Note the two nutrient-starved, chlorotic
Cordia trees growing in almost pure rubble (center left).

PI. 80.

View of Motu KimoaC'Ra
and Cordia.

Its central forest, typical of the Southern I.<

133

Motu Ana-Ana ("Anne's Islet"): a view with giant ray, from the shallows of the lagoon's southern end adjacent to South Island.
Note the similarity to Plate 5, dating from 1883.

L34

Appendix 1

Reef Information for Navigators

We include this section because no accurate hydrological
chart exists, and the Pacific Islands Pilot ( Hy drographer of the
Navy, 1982) section for Caroline is incomplete. Arundel's
1883 map (Admiralty Chart No. 979, Fig. 4) is still used today.

Caroline has neither a deep pass, nor navigable channels
into the lagoon, nor a ship anchorage beyond the reef. In 1 873,
a set of moorings was placed off the west coast of South Island
for the convenience of guano ships, approximately "a mile
north of the south-west point, in about 60 fathoms of water and
some distance from the shore" (Maude, ca. 1942). These are
long gone, although small boats can still anchor within the
close lee of South Island during normal trade winds. Today's
ships, however, must drift well offshore after approaching the
atoll from the west (PI. 12).

Of special note is a possible extension of the perimeter reef
south and southwest of Caroline. Arundel's map notes: "Reef
reported to extend four cables from southeast point." This
information probably originated in Findlay's South Pacific
Directory of 1884. quoted by Holden (1884). Evidently the
windward reef of South Island extends approximately 1.7 km
from its southeast point. From here "this reef sends out two
branches to a distance of 2.5 km. one toward the southeast, the
other toward the southwest and is consequently dangerous to
approach at night." Arundel's map does not include this
bifurcation which, according to Findlay ( 1 884 ). extends at least
across the width of South Island. He also states that "a landing
(not always safe) may be effected on the north side of the
southwest bifurcation, described above." No trace of these
submerged reefs is evident on the RNZAF aerial photos.

The "boat entrance" (Fig. 4). a narrow nick in the outer
leeward reef, marked by the stock and ring of an anchor
and immediately to the west of South Island's northwest point,
is not necessarily the easiest route to the lagoon. Landing
is possible across the steep-to reef at many locations along
the leeward reef; opposite the southern end of Ana-Ana is
good.

Landing is fairly straightforward by the anchor when the
seas are calm, especially when one becomes familiar with the
crooked notch that narrowly pierces the outer reef. After
negotiating a powerful backwash, one's boat is swept onto the
shallow reef flats — liberally laced with chunks of jagged
reef — which is exposed at low tide and barely covered at high
tide. A swift current passes west out of the lagoon between
South Island and Ana- Ana, sweeping over the reef at the notch.
Only small craft with virtually no draft can effect the 500-m
journey to South Island. Because the shallows are unchanneled
and not navigable even at high tide, one's small boat must be
carefully hauled through the water to a sheltered landing spot
adjacent to South Island's northwest point (PI. 12).

An alternative landing method used by yachts in calm
weather is via the "blind passage" (Structure and Topography
section. Fig. 50). adjacent to the northeast corner of South
Island. Despite the fact that the inner one-third of this narrow
diverticulum is calm ( and used for the residents' yacht mooring),
the outer two-thirds are rough and dangerous most of the time.
Its channel leading to and from the open sea is particularly
turbulent and should not be attempted without assistance from
the residents, and only at first light.

135

Appendix 2

Weather Data. Caroline Atoll. 1989- 1 WO

A. Wind Direction and Speed (mph), 1989

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

T~ NNE 15 NW 18 E25 E12 E 20 E 15 E 25 E 20 NE 10 N 20 NE 8

2 NE 10 W5 E 15 E 16 E 20 NE 18 SE 20 E 12 NE 12 N 15-17 NE 10

Variable

3 NE 10 NE20 E 12 SE 14 S 15-20 E 18 E 20 SE 25 NE 15 NE 15 NW 20

4 NE 15 NE 15 E 12 SE 15 E 15 E 15 E 20 SE 20 NE 15 NE 12 NW 20

5 NE20 NE 15 NE 15 E 12 NE 8 E 15 E 18 E 18 NE 12 E 12 big NW 18

swells

6

NE20

NE20

SE 10

NE8

E 15

E 10

E 18

-

NE 12

E 10

N 10

7

NE 12

SE 18

E 12

E 14

NE 10
E 10

NE25

-

E 12

-

NE 15

NE 12

NW 18

8

NE 12

E 15

calm

SE 1(1

E 10

NE20

SE 12
squalls

NE 18

-

NE 15

calm

NW 18

9

NE 12

E 18

SE 10

E20

E 15

NE 18

SE20

E 12

-

NE 12

N 10

NE 10

10

NE 15

E 16

NW 10

E20

E 15

E 15

SE 16

E 12

-

NE 10

NE 12

NE 10

11

NE 10

NE 16

calm

SE 25-

E 10
squalls

E 15

NE 15-
E 15

E 15

calm

-

NE 10

E 18

NE5

12

NE20

SE 14

E 10

W 14

E 15

E 15

E 12

E 12

-

NE 16

E 17

NE 12

13

NE15

E 12

E 12

E 16

NE 16

NE 16

E 12

-

-

NE 11

E 14

NE 12

14

NE 15

E 10

NE 10-18

E20

NE 16

NE8

E 12

calm

-

ENE 10

E 10

NE 14

15

NE15

NE8

E8

E 18

E 15

E 10

E 10

-

-

NE 10

calm

NE 14

16

NE 15

SE8

E 12

E 18

E22
squalls

E 10

NE 10

-

-

NE 1.2

calm

NE 15

17

NE 18

E20

NE 15

E 18

E25

E 15

E 10

-

-

NE 14

calm

NE 12

18

NE 12

E 12

E8

E 18

NE 18

E 15 big
swells

E 15

-

-

NE 15

N

NE 14

19

NE 12

E 12

E 10

E 18

NE 14

E 10

E 16

-

-

NE 13

NW 10

E 12

20

NE 18

E 12

E 10

E 18

NE 14

E 15

E 18

-

-

E 15

NW 15

E 10

21

NE 15

NE 10

SI X

E 14

E 12

NE20

E20
squalls

-

-

E 16

N 16

E 10 big
swells

22

NE20

E 15

calm

E 14

E 12

E 10

E15

-

-

NE 14

NE 14

E 12-NE 10

23

NE 18

thundei

squalls

N 18
thunder

SF. 14

SE 10

F 10

E 15

-

NE 20-30

E20

NE 16

calm

24

NE 15

E25 |
E 15V

N 12

SE 14

SE 10-25

F 10

E 12

-

E 20-30

NF 15

M i:

NE 15

25

NE 12

N 12

E 12

E 18

SE 17

E 10

-

E20

NE 15

NE 12

NE 16

26

NE 12

E 15

N 10

E 10

\l I 1

E 10

calm

-

E20

NE 15

N 10

\l I--

27

NE 12

1 25

SE 10

NE 10

E 18

E 10

calm

-

NE 17

E9

E9

NF. L5

28

NE 12

E 14

i: 12

E 1 i

E 18

E 10

E 15-35

-

E 1 5

NE 10

F8

NE 15

>9

NE 15

E 14

1. 18 big
swells

E 15

E 18

E 15

-

N 10

NE8

E7

NF 12

JO

M 12

1 I !

1 I !

-

E 18

F 15

-

calm

calm

-

F 10

51

calm

E 18

SE 18

-

N 20

E 16

136

B. Rainfall, 1989-1990

Month/Year Mean Monthly Mean Number of
Rainfall (mm) Rain Days

Jan "89

71.1

7

•90

177.8

22

Feb '89

160.0

14

•90

640.1

10

Mar '89

259.1

20

'90

215.9

14

Apr '89

190.5

16

'90

48.3

6

May '89

66.0

10

'90

325.1

8

Jun '89

48.3

11

'90

78.7

11

Jul '89

45.7

12

'90

68.6

8

Aug '89

35.6

12"

'90

109.2

14

Sep '89

50.8

3"

'90

81.3

6

Oct '89

73.7

11

'90

175.3

9'

Nov '89

78.7

7

'90

134.6

14

Dec '89

162.6

11

'90

154.9

9

Annual '89

1.242.1

134

'90

2.209.8

131

Source: Ron Falconer, Caroline Atoll (personal communication).
■"Based on 16 days' data.
hBased on 9 days' data.
'Based on 18 days' data.

137

1.2 Ecological Studies on Caroline Atoll,

Republic of Kiribati, South-central Pacific
Ocean

Part 2. Seabirds, Other Terrestrial Animals, and Conservation

CAMERON B. KEPLER . ANGELA K. KEPLER and DAVID H. ELLIS*

' US Fish & Wildlife Service, Patuxent Wildife Research Center, Southeast Research Station. Athens, Georgia, USA
I S Fish & Wildlife Service. Patuxent Wildlife Research Center, Laurel, Maryland, USA

Introduction

On 26 July 1988. the Soviet research vessel Akademik
Korolev sailed from Vladivostok enroute to Dutch Harbor,
Alaska. There, Soviet oceanographers joined their American
colleagues to invest:gate the Gulf of Alaska and the Chukchi
Sea in the Third Joint US-USSR Bering & Chukchi Seas
Expedition. When the arctic research was completed in early
September, the ship headed toward the central Pacific. A
rendezvous for a second contingent of Americans took place in
Hilo. Hawaii, on 9 September. Six Americans joined the ship,
which set sail on a cruise track of 14.N42 km that terminated

6 weeks later in Singapore. An important part of the expedition
was research in and around little-known Caroline Atoll, at the
southeastern edge of the Line Group. On Christmas Island, we
picked up Katino Teeb'aki, a conservation officer for the
Republic of Kiribati, who represented his government and
helped our land-based research efforts. After landing on
Caroline on 22 September, we camped in 2 locations for

7 nights, surveying the terrestrial plants and animals on all
39 islets. Caroline is a remarkably pristine atoll with its native
plant communities nearly intact on all but three islets, and
teeming seabird communities that, collectively, are second in
the Line Group only to Christmas Island (Kiribati) in diversity.
For several historical reasons, the natural values of this
spectacular blend of marine and terrestrial resources have been
overlooked.

Approximately 1,000. 000 seabirds of I I species bred mi
Caroline Atoll in November 1988. The most abundant species,
with over 900.000 birds in 1988. was the sooty tern (Sterna
fuscata). Two species (red-tailed tropicbird [Phaethon
rubricauda], blue-gray noddy [Procelsterna cerulea]) are
reported breeding for the first time. The known seabird fauna
now includes one tropicbird. three boobies, two frigatebirds,
and five terns.

Seabird distribution on Caroline is determined by the
distribution of plant communities, rats, coconut crabs (Birgus
latro). and the prevailing trade winds. Red-tailed tropicbirds
and ground-nesting brown noddies (Anous stolidus) nested on
small islets relatively free of rats and coconut crabs, masked
and brown boobies (Sula leucogaster) preferred exposed
windward beaches, primarily on Long and Nake. The tree-
nesting red-footed booby (Sula sula) and the frigatebirds

attained their highest nest densities in areas with reduced wind
speed. The black noddy (Anous minutus) was found in dense
colonies, generally high in Pisonia trees in the center of small
islets, while the uncommon blue-gray noddy (Procelsterna
cerulea) nested solitarily on open coral rubble. Sooty terns
nested in large colonies, generally near or under relatively open
Tournefortia scrub but also in open areas under Tournefortia
and closed-canopy P/.sy»i/« forests. Tree-nesting brown noddies
and white terns (Gygis alba) were found throughout the native
forests and were the only species that nested in anthropogenic
forests. Disturbed forests on South and Nake held the lowest
seabird population densities, and no birds nested on inhabited
Motu Ana-Ana.

About 300 bristle-thighed curlews {Numenius tahitiensis).
a rare shorebird, overwinter on Caroline, foraging in all terrestrial
habitats, including Pisonia and Cocos-Ipomoea forests. We
extended the known winter range of the long-tailed cuckoo by
discovering a small population on the atoll, the first record for
the Southern Line Islands.

The known lizard fauna was increased from three to six
species. Approximately 2.200 coconut crabs inhabited
12 islets on Caroline. Although primarily associated with
coconut plantations, we also found them in Pisonia and
Tournefortia.

We now know that the populations of seabirds and coconut
crabs on Caroline Atoll are of national and international
importance. The black noddy (17,000 birds) and white tern
(8,000 birds) populations are the largest in the Republic of
Kiribati, while the red-footed booby population (7.000 birds)
is the fifth largest in the world.

History of Ornithological Studies

"There were a great quantity of seabirds of several kinds,
and so importunate that they seemed to want to attack the men"
(Markham. 1904). So wrote the Portuguese explorer deQuiros
on 21 February 1601, the first European to see Caroline Atoll.

Precisely what seabirds were present remained a mystery
until the island was surveyed 364 years later by the Pacific
Ocean Biological Survey Program (POBSP) (Clapp & Sibley,
1971a). Prior to this expedition, accounts of the avifauna had
been incomplete and somewhat confused. Bennett (1840)
described red-footed boobies, a frigatebird (species '.'). white

139

terns, bristle-thighed curlews, tattlers (Heteroscelus incanum
and H. brevipes), and "a great number of small pigeons" with
white heads (certainly noddy terns, perhaps both A. minimis
and A. stolidus). The "shoal birds" that greeted him were
probably sooty terns. His most unusual contribution was
mention of a possible flightless rail: "The other birds of the
coast were a kind resembling a cool..." (p. 372).

The 1883 Solar Eclipse Party (Subchapter 1.1. History of
Caroline Atoll section) published a few sketchy notes, adding
lesser golden-plover (Pluvialis dominicd), reef heron (Egretta
sacra), and masked booby (Sula dactylatra) ("gannet") to the
bird list. Of dubious identity were two species of "seagull" and
a "snipe" (Dixon. 1884). Holden. one of the astronomers,
heard "the notes of a singing bird." which prompted us to add
mist nets to our equipment in the hopes of capturing an
Acrocephalus warbler. This resulted in our discovery of the
long-tailed cuckoo (Eudynamistaitensis) (Ellis et al., 1990) and
piqued our curiosity about what Holden might really have
heard.

The POBSP expedition spent 3 days on Caroline in June
1965. They found 10 species of seabirds (9 breeders).
4 migrant shorebirds, and a reef heron (Clapp& Sibley. 1971a),
providing rough population estimates for each species. This
w ork laid the foundation for later expeditions. Brief visits to
Caroline by the Kiribati government in 1974 and Roger Perry
in 1977 (Garnett, 1983) added no further information.

The 1988 expedition to Caroline was longer and more
extensive than all former visits. We found three new island
records: a breeding seabird ( red-tailed tropicbird ). a shorebird
(Sanderling [Crocethiaalba]), and a migratory land bird (long-
tailed cuckoo), and determined islet-by-islet distributions for
each species. Our population estimates, calculated from field
work, aerial photographs, and detailed vegetation analysis.
indicate that Caroline's avifauna is far more important than had
previously been suspected (King, 1973; Garnett. 1983). In
March and May 1 990. the ICBP 1990 Line and Phoenix Islands
Expedition (Subchapter 1.1. Methods section) filled in minor
gaps in our knowledge. Caroline's residents added another
breeding seabird, the blue-gray noddy, in summer 1990.

Methods

Distribution and Habitat Preference

We described se\ en major plant communities on Caroline
Atoll (Subchapter 1 . 1 i. With the use of aerial photos and the
transect data, we mapped the communities found on each islet.
Bird distribution was determined and plotted using these islet
vegetation maps. If a species nested within a particular plant
community, it was plotted on the distribution maps as occurring
throughout that community unless determined otherwise.

Population Sizes and Breeding Phenology

We measured transect distances for each islet using a hip-
chain and biodegradable cotton thread. We recorded all birds
seen within the 30-m-wide strips: transect width was estimated
visually. We assigned birds to one of several mutual I v exclusive
categories: adults present, adults on territory, adults on nests
(contents unknown ). eggs, naked chicks, dow ny chicks, chicks
w ith remiges erupting, chicks with scapular feathers, or chicks
in juvenile plumage. We created a range of possible laying
dates for each egg and chick using known growth parameters
for each species (Kepler. 1978; Kepler & Kepler. 1978). This
enabled us not only to estimate seabird populations, but also to
determine and plot a rough breeding phenology for each
species (Figs. 3.5.7.9.10.12). In these figures, the height of the
bar for each category ("downy," "scapulars." etc.) represents
the number of nests found or estimated with that development
stage in September 1988. The bar width represents the
approximate time span over which eggs could have been laid
to produce that stage, while the "no. day s" is a count back from
the survey dates to accommodate growth and development that
had occurred. Thus, while each figure shows what breeding
stages we found, we extend those nests back in time to show-
roughly when they would have begun. The number of clutches
begun per day is determined by dividing the number of nests
per stage by the time span in days over which those eggs were
laid.

Sooty terns nested in dense colonies. Each colon) was
mapped, and its total si/e (m') w as calculated. A minimum of
10 plots (3x3 mor.3 x 6 m). within which all eggs and chicks
were counted, were randomly located along a compass line in
each colony. The population size of each colon) w as estimated
from these plot densities.

From 22 2l> September I9S8, C. B. Kepler. A. K. Kepler,
D. H. Ellis, and K. Teeb'aki surveyed all of Caroline's 39 islets
except North Arundel Islet, naming most of them (Fig. 1 ; see
Subchapter 1.1, Methods section). We established 50 linear
transects, extending 13,300 x 30 m. laid out to ensure that at
least y< of each islet was sampled lor birds and plants (see
Subchapter I.I. Methods section and Fig. 8). Sampling was
increased with 19.300 m of perimeter surveys along the
windward and leeward coasts of 21 islets (Subchapter 1.1.
Fig. 9). On Noddv Rock. Skull. Atibu. Bo'sun Bird. Coral.
Reef-flat, and Fishball (Fig. 1). we made total counts of the
breeding seabirds. All sur\ cv s w ere conducted during daylight
hours. Some incidental data have been added from the 1990
ICBP expedition.

Mist Nets

We operated 4 ATX 4-shelf 36-mm mesh mist nets

(2.6 x 12 m)

43.5 net hours, according to the following

schedule: 14.5 net hours (daylight) beneath a 10-15 m Cocos
canopj on South. 27.5 net hours (day and night i in Pisonia-
Cocos interface 1 12 m tall) near Tr. 10 on Long, and 1.5 net
hours m Pisonia-Tournefortia within a 4-6 m canopj onTr.4,

Long. One cuckoo was collected (USNM 607 19 1 1.

Collecting Other Vertebrates

Lizards that were active and conspicuous were collected at
base camps on South and Long, either bv hand or with a
blowgun firing steel darts. No attempt was made to search lor
reptiles under coral, litter, or in other concealed locations

140

Rats were collected with a blowgun or snap traps baited with
coconut, the former proving far more effective because most
traps were sprung by hermit crabs. We preserved all specimens
in formalin and sent them to the US National Museum.

Seabird Species Accounts

Eleven species of seabirds occur at Caroline, most of
which breed in large numbers. They include one tropicbird,
three boobies, two frigatebirds. and five terns.

Red-tailed Tropicbird (Phaethon rubricauda) (Figs. 2.3; PI. 1 )

Red-tailed tropicbirds breed at widely scattered locations
throughout the tropical Pacific and Indian Oceans. In the Line
Group, they nest from Palmyra south to Starbuck ( Perry, 1980),
with a large population (8,500 birds) on Christmas Island
(Clapp, 1967). Prior to our expedition it was unknown from
Caroline, Vostok, or Flint.

Distribution and Habitat Preference: Our first indication
that red-tailed tropicbirds nested on Caroline was the discovery
of the skull, tail feather, and broken egg ( Subchapter 1 . 1 , PI. 46)
under a small Tournefortia bush on a previously unnamed islet
between Pig and North Brothers Islets. We named this sparsely
vegetated collection of rubble "Skull Islet" (Subchapter 1.1.
pis. 46,49). We later found 47 nests on another islet, naming
it Bo' sun Bird (Fig. 1 ) after the species' common name.

All nests were located under relatively open Tournefortia
scrub less than 3 m tall in open, windy locations, with the
majority (91%) on small islets (0.24-0.86 ha). All nests were
under shrubs with few stems within a 0.5 m2 nest space, and
most had peripheral cover on the sides of the shrubs, both
important factors in nest-site selection (Clark etal., 1983). All
nests were in areas relatively free of Polynesian rats (Rattus
exulans) and coconut crabs (Birgus Ultra): 5 nests on Long
were within 50 m of the island's south point.

There are large populations of Polynesian rats and coconut
crabs on Caroline's bigger, more wooded islets. This rat,
though basically vegetarian, is an effective seabird predator
(Kepler, 1967; Norman, 1975) that in some years has taken
65% of the red-tailed tropicbird eggs and 1 00% of the chicks on
Kure Atoll (Fleet, 1972). Coconut crabs are also known bird
predators (Clapp & Sibley, 1971a; Helfman, 1979; Reese.
1987). It may be no accident that tropicbirds on Caroline
are restricted to small, relatively open islets that harbor few, if
any. rats and crabs or occur only on the tip of Long Island,
where predator densities are low. We saw no rats on Bo'sun
Bird Islet. Although rats could swim the 1 65 m to the islet, the
nearly continuous presence of black-tipped reef sharks
(Subchapter 1.1. PI. 10) in the channels surrounding the islet
provides protection to its nesting tropicbirds.

Numbers: In September 1988. we found a total of
56 active nests on 5 islets (Fig. 2, Table 1) and estimated a
minimum population of 60 pairs. The May 1990 expedition
found 130 nests on Bo'sun Bird: our revised estimate for
Caroline is approximately 300 birds. Bo'sun Bird Islet was
surveyed by POBSP in June 1965, and no tropicbirds were
located on the ground or in the air (F. Sibley, personal

communication ). It is unlikely that red-tailed tropicbirds were
present but overlooked at that time, suggesting that they have
colonized the atoll only recently. The Caroline population is
now the second largest colony known from the Line Group, and
Caroline is only one of five islands in the archipelago where
red-tailed tropicbirds are known to breed.

Phenology: Of the 56 nests found in 1988, 54 contained
eggs or chicks (Table 2). The 33 chicks were divided into 4 age
classes (Fleet, 1974; Diamond, 1975a). which, together with
the 21 eggs, provided an indication of laying phenology for
140 days prior to our arrival (Fig. 3). Eggs in surviving nests
had been laid at a fairly even rate from early May (possibly
starting earlier) through September. The finding of only two
additional pairs on territory, and only one courtship flight,
indicated that laying was ending. On 24 May 1 990. many nests
contained eggs and downy chicks (75% nests with chicks), and
pairs were still courting.

On Christmas Island, peak laying generally occurs from
June to October ( Schreiber & Ashmole. 1 970), later than those
parts of the 1988 and 1990 breeding seasons we observed on
Caroline.

Masked Booby (Sula dactylatra) (Figs. 4,5; PI. 2)

The masked booby is widely distributed in the Atlantic,
Indian, and Pacific Oceans. Clapp (1967) estimated that
19,100 masked boobies bred in the Line and Phoenix Islands,
with about 13,000 of them in the Line Islands, mostly
(ca. 9,000) on Jarvis.

Distribution and Habitat Preference: Eighty-four percent
of masked booby nests (159) were on the windward, rubbly
shores of Long and Nake Islands, extending to the north end of
the atoll. Fifteen additional nests were scattered along the
lagoon edges of five South Nake Islets (Table 1 ). Nests
consisted of bare scrapes with exposed sand, usually within a
sparse ground cover of Portidaca and Heliotropium (PI. 2).
Over half the nests were amassed in one open colony on Nake
that extended nearly 1,000 m, beginning approximately 150 m
south of Tr. 2 and extending about 50 m north of Tr. 4
(Subchapter 1.1, Fig. 8). Here a nearly unbroken Heliotropium
mat 30-80 m wide, with patches of Tournefortia, occupied the
area between the leading edge of the Tournefortia scrub and the
beach crest. Nests were 20-30 m apart in the densest section
(nearTr. 3). All nests were exposed to the sun, unlike those of
the brown boobies. Some adults and juveniles roosted under
the scrub; guano deposits indicated regular occupancy.

A loose group of 7 breeding pairs was scattered on a broad
plain of low herbs along a partially filled old interislet channel
370 m south of the north end of Long Island (Tr. C.
Subchapter 1.1, Figs. 8,40). Four more pairs nested in coral
rubble along the channel separating Nake and Long, one pair
with a downy young only 2-3 cm above high-tide flow on an
"islet" between fingers of the channel, a precarious location
where nesting surely must fail in stormy periods. No birds were
seen there in March and May 1 990, following a severe storm in
February 1990. Four pairs nested singly along a leeward
1 ,000-m stretch of lagoon shore on the northern end of Long
(Fig. 4); hardpan was the primary substrate.

141

Numbers: In September 1988 we found 189 masked
booby pairs (Table 3), including those on territory (with or
without nest scrapes) and juveniles (with or without attending
adults). We found no "clubs" of nonbreeding birds. We
covered most of the habitat favored by this species except the
northern 300 m of Nake Island; in 1990 a few scattered pairs
nested there. Our population estimate, including pairs we
might have missed, was approximately 200 breeding pairs.
The only other population estimate was of "ca. 10" birds
(Clapp&Sibley, 1971a);in 1965 POBSP biologists (F.Sibley,
personal communication) surveyed all locations where we
found breeding pairs. Thus, 200 pairs represents a major
increase in the population on Caroline Atoll.

Phenology: In June 1965 only four masked booby nests
containing eggs were found ( Clapp & Sibley, 1971a). indicating
that nesting began in May or June. We found nests in all stages
in September 1988 (Table 3. Fig. 5 1. The large age class in
April may include some juveniles that could fly ( i.e.. were older
than 1 80 days). We may have undercounted naked chicks, not
wishing to expose them to the sun by frightening the brooding
adult. Laying began in April or earlier, peaked in June and July
(Fig. 5). and continued until our survey in late September. The
34 pairs on territory, many with nest scrapes (Table 3 ). indicated
that laying was still in progress and would continue into
October.

In March 1 990. 3 1 pairs were on territory or were attending
nests, eggs, or older chicks, indicating that a new breeding
season was under way as the previous season was ending. By
May 1 990 there were 63 nests, mostly with eggs, and there were
no older chicks. Thus, the 1990 season augments the 1988 data
and suggests an annual cycle with egg laying beginning slowly
in lebruary and March, peaking in June and July , and declining
to a low ebb from December to February.

The large number of fledgedjuveniles and nests with older
chicks, in both September 1988 and March 1990. indicated that
the 1988 and 1989 breeding seasons were very successful. It
also suggested that potential predators (rats and coconut crabs )
posed little hazard to this hardy species.

Brown Benin (Sula leucogaster) (Fig. 2)

This widely-distributed pantropical species has an
estimated population in the Line and Phoenix Islands of about
$,200 (Clapp. 1967; Perry. 1980). with over half of them
(2.000) recently found on Maiden Island, in the Southern Line
Group.

Distribution and Habitat Preference: Breeding brown
boobies on Caroline were restricted to the w mdward edges of
Tournefortia scrub and forest, generally within 15-20 m of
high water. In 19SS we found nests on four islets (Fig. 2,
Table I). Long, with 12 pairs, was the onlv islet supporting
more than a single pair. They w ere located on the northern two-
thirds of the island: four pairs formed a loose colony near the
head of Tr. A ( Subchapter I.I. Fig. Si. All nests were under
Tournefortia bushes approximately 3 m tall. In March 1990.
we found 20 pairs of brown boobies, all on windward Nake as
far as (he islet's northern extremity. There was no evidence of
nesting on Long Island. On May 22. 1990. only three nests, all
with engs. were found on Nake.

On 22 September 1988. we saw 2 birds plunge-div ing w ith
masked and red-footed boobies approximately 500 in west oi
South Island. On the atoll, flying brown boobies were observed
soaring only along the windward beaches. Two birds roosted
on the south-central beach of South, and another was found
roosting on Kota.

Numbers: We counted 1 5 pairs during perimeter surveys
in 1988, yet found none on the transects. Since we covered
virtually all the w indward beaches ( Subchapter 1.1. Fig. 9 1. we
are confident that fewer than 20 pairs nested on the atoll. Our
population estimate for 1990 was 25 pairs.

The POBSP (Clapp & Sibley, 1 97 1 a) found three nests on
Nake in June 1965, estimating a population of 15 birds. Even
though our surveys triple the known population, the brown
booby remains a rare seabird on Caroline.

Phenology: With the exception of one recently
fledged juvenile, all nests contained eggs in September 1988
(Table 3). Clapp & Sibley (1971a) found eggs in June 1965.
In March 1990, the 20 pairs were all on nests whose contents
ranged from eggs to an older juvenile. However, only
2 months later, only three nests, containing eggs, could be
found. These data from 3 years suggest that the species may
have trouble rearing young on the atoll. More juveniles should
have been encountered, especially in May 1990. Predationby
Polynesian rats or coconut crabs could limit reproduction on
the atoll.

Red-footed Booby (Sula sula) (Figs. 6.7: Subchapter 1.1.
PI. 51)

This pantropical booby numbers over 55.000 individuals
in the Line Group (Clapp. 1967; Perry, 1980), making it one of
the most important regions in the world for this species.
Caroline holds the fifth largest known red-footed booby colony
( see Nelson. 1978). The largest known colony ( 140.000 pairs)
is found on Tower Island (Galapagos): 3 of the 5 biggest
colonies occur in the Line Group.

Distribution and Habitat Preference: In 1988, the red-
footed booby bred on 28 islets, ranging in size from Nautonga
(0.34 ha) to Nake (107.46 ha) (Fig. 6). On the Windward
Islands, red-foots occurred from Nake to Tridacna. absent only
from the smallest islets ( Noddy Rock. Skull Islet. Motu Atibu ).
The species was also widespread on the leeward islets, extending
from Pandanus to Eitei. The tiny islets ( Fishball, Azure, Reef-
flat) were not occupied.

Red-foots are tree nesters w hose distribution on Caroline
closely matched that of Tournefortia scrub and forest
(Subchapter 1.1. PI. 51). Thev sometimes utilized smaller
Pisonia or Cordia trees where thev intermingled with
Tournefortia and occasionally built nests in the tallest (>15 mi
Pisonia. They nested in smaller Tournefortia patches within
the peripheral scrublands, especially those not directly exposed
(o the trade winds. They clearly av oided smaller islets because
of the lack of suitable Tournefortia in which to breed. Thev
nested inward from (he vegetated edges of the islets, generally
at 3-6 m in height, and were distributed in broken rings around
the smaller motus in areas of moderate winds. A higher
percentage of the population occurred on perimeter surveys
than on cross-island transects.

142

Red-foots were absent from South Island, which was
primarily covered with Cocos (Subchapter 1.1, Figs. 50,51).
Even though Tournefortia occurred on all its coastlines, no
boobies nested in them. Ana-Ana was also unoccupied: the
presence of a family of four people, a cat (removed in 1990),
and a dog undoubtedly discouraged nesting attempts. Red-
foots also avoided the mixed forests of south Nake, which
contained much Cocos and Pandanus (Subchapter 1.1,
Fig. 37 ). Red-footed boobies were thus found only in Caroline' s
indigenous woodlands, primarily in Tournefortia >2 m tall;
they avoided anthropogenic plant communities and man.

Red-foots used a wider range of habitats for roosting.
Nonbreeding birds were found throughout the taller indigenous
trees, even in leeward situations where Pisonia and Cordia
overhung the lagoon (as on Long Island).

Numbers: The POBSP(Clapp& Sibley, 197 la) estimated
5,000 ± 25% Red-foots on Caroline in June 1965, with about
2,000 + 25% nesting pairs. In 1988, we sampled systematically
more than 7% of the available habitat on all motus except
Crescent (4.6% sampled) and North Arundel, and estimated
that 2,22 1 pairs of red-footed boobies nested on 27 of Caroline' s
islets (Table 1). We found an additional 1,234 roosting,
nonbreeding birds. We know (Kepler, 1969; Nelson. 1978)
that fewer boobies remain in their colonies during the day than
at night. Thus, an unknown fraction of the population was at
sea when we conducted our counts. Impressive flights of red-
footed boobies returned each evening: 3^4 birds arrived for
each one that had remained behind, many undoubtedly mates
of incubating birds. To approximate the number of returning
nonbreeding birds, we doubled the number of roosting adults
to allow for an additional 1,234 adults and juveniles. Thus, our
conservative estimate was at least 7,000 individuals.

Because red-footed boobies were so dependent upon
Tournefortia, we determined the nesting population on each
islet by multiplying the number of nests found on transects by
the ratio of sampled to total Tournefortia area. Perimeter
counts (Subchapter 1.1, Fig. 9) were used if the number of red-
foots observed exceeded the number calculated from the cross-
island transects.

Long Island held the greatest number of nests (659),
mostly in the leeward Tournefortia and Tournefortia— Pisonia
edge. Bird densities were typically highest on the largest islets:
Windward and Tridacna, the largest Windward Islets, held 163
and 1 1 1 nests, respectively; and Mannikiba. the biggest leeward
islet, harbored the largest population ( 1 84) of the entire leeward
side. There were exceptions, however: Pandanus, with four
times the area of Tournefortia of any of the South Nake Islets,
held fewer birds than three much smaller islets (Table 1 ).

Tournefortia scrub and forest covered approximately
125.25 ha (Subchapter l.l,Table9). Overall, there were 1.75
red-footed booby nests/ 1 .000 nr of Tournefortia forest. Nest
densities for occupied islets by island groups (Table 4) showed
that red-foots favored areas less exposed to the trade winds:
most nests on the windward motus were protected by well-
developed Pisonia forests. The exposed Central Leeward
Islets held the lowest nest densities ( 1 . 2 nests/ 1 .000 m2), far less
than on the South Nake Islets (5.3/1,000 nr), which are

protected by the northern edge of Long. The greatest densities
(7.8 nests/ 1 ,000 nr) occurred on the South Nake Islets south of
Pandanus.

Broadly speaking, red-foots breed in well-dispersed
colonies. A record density of 600 nests/ 1 ,000 m2 on Tromelin
Island (Indian Ocean) is exceptional. Elsewhere,
53 pairs/1,000 nr on Tower Island (Galapagos), 40/1,000 nr
on Moku Manu (Oahu. Hawaii), and 27/1,000 m2 on Half
Moon Cay (Honduras ) are more consistent high-density colonies
(Nelson, 1978). Only on tiny Motu Kota (Subchapter 1.1,
Fig. 52), with 1 2 nests in 303 nr of Tournefortia (40/1 ,000 nr ).
did we find such density, and for this reason we named the islet
"Kota" (Gilbertese for red-footed booby).

Phenology: In September 1988, we located 339 nests. Of
the 152 whose contents could be determined, 87 were empty,
63 contained eggs, and 2 held downy chicks. We saw dozens
of flying juveniles along the windward coasts. Most pairs were
building or guarding their nests during a pre-laying stage that
lasts from 1 1-35 days (Nelson, 1969). Of the pairs with nests,
57.2% had yet to lay and 41.4% had laid their eggs between
mid-August and late September (Fig. 7). Applied to the total
breeding population, approximately 1,270 nests were in the
prelaying stage and would be expected to produce eggs
throughout October. An additional 9 1 9 nests had a mean laying
date in early September (Fig. 7). Red-footed boobies were
synchronous with brown boobies but delayed relative to masked
boobies.

In June 1965. nests containing prelaying adults, eggs, and
young in all stages indicated that the birds were in the midst of
a protracted breeding season extending from January to June.
Our data reveal that no successful nesting occurred in May-
June 1988. Data from March and May 1990 indicate that nest-
building began in January (or earlier), with eggs laid from
January to May. Red-footed boobies in other tropical locations
have variable, opportunistic breeding seasons that depend
upon food availability (Nelson. 1978); our data suggest that
similar pressures could be operating at Caroline.

Color Morphs: Red-footed boobies are polymorphic
(Nelson, 1978). The basic plumages are brown or white, with
brown morphs having many combinations of tail, back, scapular,
foot, and bill colors. A variety of brown forms and white forms
occurred on Caroline, with a ratio of 9:1 (337 brown to
35 white), which contrasted sharply with Nelson's (1978)
statement that "in the Line and Phoenix Islands all birds are
white morphs." Most of the dark morphs were the "white-
tailed" form (see Nelson, 1978. pp. 660-661 ). The variations
and proportions of plumage types show clinal change in the
Line and Phoenix Islands ( F. Sibley, personal communication),
and the question of plumage morphology needs much more
study in the central Pacific.

Great Frigatebird (Fregata minor) (Subchapter 1.1. Figs. 7,8
and PL 42)

The great frigatebird breeds at widely scattered locations
throughout tropical waters in the Atlantic, Pacific, and Indian
Oceans. It is known to breed on all of the Line Islands except
Starbuck (Perry. 1980).

143

Distribution and Habitat Preference: Great frigatebirds
nested on 25 islets, including Nake, Long, and most of the
larger islets (Fig. 8. Table 1 1, ranging in size from Azure
(0.20 ha) to Nake ( 107.46 ha). Every occupied islet had some
Pisonia forest, even if only a single tree (Azure). The larger
islets lacking Pisonia forest (Arundel. 7.34 ha; Tridacna.
9.08 ha) lacked frigatebirds in 1988. although frigatebird
chicks were present on Arundel in early 1989 (Anne Falconer,
personal communication).

Although great frigatebirds were similar in nest
requirements to red-footed boobies, there were significant
differences: the frigates tended to nest higher in, and closer to,
the outer edge of the canopy (although nests were found as low
as 1.3 m). Nest sites were more sheltered from the wind than
those of red-foots and in locations where the birds could take
flight easily. Such site preferences may explain the association
with Pisonia. Pisonia reaches 25 m on Caroline, taller than
other tree species, providing a windbreak on most islets. The
largest colonies (Nake. Long, Pig. Mannikiba) were found
leeward of these stands. We found nests in Tournefortia,
Pisonia. and Cordia. They were often in the Tournefortia-
Pisonia interface, generally in the taller Tournefortia. One
colony on south Long overhung the lagoon in a dense Pisonia
stand. Frigates were not found in any anthropogenic forests
and were absent from then-inhabited Ana-Ana.

Numbers: The previous population estimate for great
frigatebirds on Caroline was 10. 000 birds (Clapp & Sibley,
1971a: Perry, 1980). We calculated that 2.427 pairs bred or
attended territories. An additional 617 birds roosted, thus the
entire population was approximately 5,471 individuals. A
large but undetermined number of birds soared over the atoll
throughout the day, and an uncountable number of birds,
including Hedged juveniles that would ultimately return to the
island to nest (Diamond. 1971). were undoubtedly at sea.
Because this species is difficult to count accurately, it is unclear
if the population has changed since 1965.

Phenology: In frigatebirds. the scapulars, which first
appear at 81 days in Fregata magnificent (Diamond. 1973).
erupt before the primaries. Because we lack chick stage data
for F. minor and F. arid, we have modified ages from Diamond
( 1 973 ) for F. magnificens, using the hatching times for F. arid
and F. minor from Nelson (1976) and fledgling ages from
Diamond! 1 975b) to construct approximate development stages
for the species on Caroline. Since (hey Hedged at an earlier age
than /•'. magnificens, we have reduced the ages for chicks with
erupting primaries for F. arid and F. minor, kept the duration
of the earlier stages approximately the same, and reduced the
period in juvenile plumage.

We found 214 nests in 1988. Of the 144 in which we
determined contents, 49 contained eggs or young chicks.
27 held chicks with developing scapular feathers, and
68 contained older chicks (fable 5 1. The additional 70 adults
occupied nests oi unknown contents; mans probably held eggs
or young chicks or were empty. We saw fewer than
10 displaying males, so the breeding season was winding
dow n. This was also indicated by the high proportion 1 87' i |oi
nests with chicks, mam ol them old. A major laving effort had

begun in March-April (Fig. 9) and continued into September.
In March 1 990. an abundance of fly ing juveniles and occasional
larger chicks down to the downy stage indicated that the
previous year's breeding season was ending. A small number
of males were beginning another courtship cycle. By May
1990, courtship and egg-laying were still under way, and nests
contained eggs or small chicks up to the "remiges" stage. Peak
laying on Christmas Island ( Pacific Ocean) occurs from March
to May (Schreiber & Ashmole. 1970), the same laving cycle
observed on Caroline in 1988 and 1990.

Lesser Frigatebird (Fregata ariel) (Figs. 8.10)

The lesser frigatebird is. along with F. minor, one of the
true pantropical species. It breeds and disperses widely within
the tropical Pacific (Sibley & Clapp. 1967). One of the largest
populations in the world (30.000-85.000) breeds on McKean
Island, in the Phoenix Group (Garnett. 1983). Lesser frigatebirds
breed on four of the Line Islands, with the population on
Maiden (7.000) the largest in the archipelago (Perry. 1980).

Distribution and Habitat Preference: In June 1965. lesser
frigatebirds were found nesting in one compact colony on the
leeward north end of Long (Clapp & Sibley. 1971a). We found
a single colony in leeward Pisonia forest on western Nake
(Fig. 8). both in September 1988 and May 1990. The birds
nested high (to 18 m) in the Pisonia and Pisonia-Conlia edge
facing an open Tournefortia savannah. Although primarily
composed of F. ariel. a few F. minor were scattered along all
but the eastern edge of the colony. West of the birds, across the
open forest. F. minor and Sula sula nested in a mixed colony in
a denser stand of Tournefortia. Birds were seen soaring over
Nake. Long, and the leeward islets but were not found roosting
or nesting away from the colony on Nake. How ever, in March
1990. approximately 650 lesser frigatebirds were swarming
above, and roosting on. Motu Nautonga in a light cluster.
possibly preparing for nesting.

Numbers and Phenology: Pacific Ocean Biological Sun ev
Program biologists estimated a population of 1.000 lesser
frigatebirds on Caroline in June 1965. with 400 ± \0% breeding:
only eggs were found (Clapp & Sibley. 197 la). On Christmas
Island. F. ariel laid in May and June in 1959. 1963. 1964. and
1 967 ( Schreiber & Ashmole. 1970). Of 46 nests found in 1988.
we could inspect the contents of only 26: all contained
feathered chicks (Table 5). Laying dates ranged from March
through July (Fig. 10). with a peak from April to June. Caroline's
lesser frigates, therefore, appeared to be synchronous with
those on Christmas.

Because w e did not determine the colony limits, we cannot
provide a population estimate. There was a minimum of
20( ) birds in 1 988 ( 46 nests, plus roosting and flying indiv iduals)
and 500 pairs in 1990.

Sooty Tern (Sterna fuscata) (Figs. 11.12: Subchapter 1.1.
PI. 59)

This tern is the most widespread and abundant tropical
seabird in the world. Under favorable conditions it forms
immense colonies numbering into the millions. It is known to
breed on 7 of the Line Islands: the largest population in the

144

Pacific is found on Christmas Island (15,000,000 at highest
count), and 3,000,000 have been recorded on Starbuck (Perry,
1980).

Distribution and Habitat Preference: To date, 19 colonies
from 10 islets are known for the years 1965, 1988, 1989, and
1990 (Fig. 11). In September 1988, we found three colonies,
two on the northern half of Long and one on Bo' sun Bird Islet;
all fit the general habitat description in Clapp & Sibley ( 1 97 1 a).
Colony A, nearly square, was 210 m on a side. Eggs were
placed under a savannah-type Tournefortia scrub, from 1-4 m
tall with approximately 60% canopy cover (Subchapter 1.1,
PI. 59). The substrate was coral rubble mixed with sand,
covered by Heliotropium (5%), Portulaca (1%), Laportea
(< 1 % ), and Lepturus (< 1 %), typical of old interisland channels.
Colony 1 was located in a broad sandy corridor with two large
"groves" of Tournefortia. The northern subpopulation extended
116 m along the windward beach, but 248 m along the lagoon.
The southern subpopulation began 28 m further south along the
beach, fronted the seaward reef for 86 m, and was shaped like
a blunt triangle, its apex pointing toward the lagoon. Most
chicks were under Tournefortia, which consisted of shrubs
2-4 m high with 80% canopy cover. The substrate was also
older beach sands mixed with coral rubble and covered with
Portulaca (40% cover), Lepturus (<5%), and Heliotropium
(<5%). The Bo'sun Bird colony, a rough oval approximately
55 m wide by 70 m long, was under 2-3 m high Tournefortia
with 75% cover, on coral rubble/sand sparsely carpeted with
Portulaca and Heliotropium.

Numbers: Populations were determined by measuring
colony dimensions, then counting eggs and/or chicks in 9-m2
sample plots located at random points along a compass line.
Because juveniles moved as we approached, they were counted
6 m ahead of us in estimated 3 x 6 m plots. The Colony I
subcolonies (North, South) were treated separately.

Colony size (rounded) in 1988 ranged from
1 27,000 ± 30,000 "nests" (Colony A) to 1,500 ±750 new eggs
on Bo'sun Bird Islet (Table 6). There were an additional
6,900 + 1 ,600 nearly-fledged chicks in the Bo'sun Bird colony,
resulting from eggs laid three months earlier.

The total number of eggs and chicks was 188,000 ±21%.
Actual numbers of adults are difficult to estimate but in other
studies have exceeded the number of eggs and young by factors
of more than two because innumerable eggs and chicks were
lost, colonies often overlapped, and many nonbreeding adults
joined the prebreeding swarms or associated with breeding
birds. Schreiber & Ashmole (1970), relying on POBSP data
from Johnston Atoll (north-central Pacific), estimated that four
adults were present for each egg laid. Pacific Ocean Biological
Survey Program data from Johnston (Amerson & Shelton,
1976) indicated that about 600,000 adults were present in a
colony with 105,000 eggs, or approximately 5.7 adults/egg. If
we assume that real numbers of terns in our colonies lay
midway between 4 and 5.7 times the number of eggs and
chicks, then the number of sooty terns using Caroline Atoll
would have ranged between 720,000 and 1,100.000 birds
(91 1.800±21%). This is twice the estimate provided by Clapp
& Sibley (1971a), even though we found fewer colonies.

In March 1990, laying was just beginning in two colonies
on Long Island, (625 x 1 50-3 1 5 m wide and 1 80 x 1 60 m wide).
Enormous numbers of birds, both on the ground in densities up
to 9 or 1 0 pairs/nr and in the air, made it impossible to calculate
a reasonable population figure. According to Anne Falconer,
these two colonies were very successful. Similarly, counting
was difficult in May 1990 when six large prebreeding swirls
hovered like huge clouds of gnats over discrete islets and islet
groups (Fig. 1 1). Our 1988 estimate of approximately one
million birds is probably a conservative count for the atoll as a
whole on an annual basis.

Phenology: The incubation period in sooty terns is about
4 weeks (Dinsmore, 1972). Young fledged 7-8 weeks after
hatching, although fledgling ages, dependent upon food supply
(Schreiber & Ashmole, 1970), are highly variable.

Four separate sooty tern colonies had been started over the
12-week period prior to our study in 1988 (Table 6). On Bo'sun
Bird Islet a new wave of laying was just beginning in an open
area immediately southwest of most of the colony, while nearly
fledged chicks scurried about beneath the Tournefortia.
Undoubtedly many young had already fledged, so many more
eggs would have been laid in early July by this colony than
indicated (Fig. 12). The two colonies on Long were established
at different times: the short-tailed juveniles in Colony 1
preceded the large number of eggs, hatching eggs, and downy
chicks of Colony A by 3-4 weeks.

The July-September laying period on Caroline in 1988 is
very different from the bimodal breeding (May-June,
December-January) reported from Christmas Island, Pacific
Ocean (Schreiber & Ashmole, 1 970), and the May laying dates
noted for Caroline by the POBSP in 1965. Additional data
(Anne Falconer, personal communication) indicate that sooty
terns may lay any time (Fig. 11), certainly January through
September (1988 to 1990). Severe storms, which destroyed
large Long Island colonies in February 1990, were perhaps
responsible for reinitiating breeding activities on the leeward
side of the atoll within the next few months. A great deal more
research will be needed on Caroline before the breeding seasons
for this species are fully understood.

Brown Noddy (Anous stolidus) (Fig. 13)

This tern, primarily a tree nester, is widely distributed
throughout the warm oceans of the world. It is abundant in the
Line and Phoenix Groups, with an estimated total population
exceeding 40,000 birds. Brown noddies are most abundant on
Palmyra Island (10,000 birds).

Distribution and Habitat Preference: The brown noddy is
second only to the white tern in the number of motus (28) upon
which it is known to breed (Fig. 13). It utilized the smallest
(Noddy Rock, 0.02 ha) and largest (South, 104.41 ha) motus,
nesting upon coral rubble and in plant communities ranging
from the simplest herb mats to Tournefortia. Pisonia, Cordia,
Cocos, and the mixed anthropogenic forests of South and
Nake. Most pairs were well dispersed, nesting from the outer
edges of Tournefortia to the central, inner branches of Pisonia,
and from the ground to the crowns of 25-m Cocos. When
nesting sympatrically with black noddies in Pisonia. the brown

145

noddies typically occupied portions of branches closest to the
trunk. Brown noddies nested almost solitarily in the Cocos
canopy on South, were found within dense colonies of black
noddies and white terns in tall Pisonia forests, with red-footed
boobies and great frigatebirds in Tournefortia, and amidst
sooty tems and red-tailed tropicbirds (Bo'sun Bird Islet).
Apart from a few ground nesters on Raurau and Fishball, the
only ground-nesting colony (80 nests) was located on a
Portulaca mat on Noddy Rock — a site free of predators,
although flooded during storms.

Brown noddies often formed loose roosting "clubs" on the
atol 1 ' s beaches. Aggregations of 1 5-20 birds were found on the
west coast of South and on Sandy Inlet, south-central Nake.

Numbers: Clapp & Sibley ( 1 97 1 a) estimated a population
of 1,000 birds in June 1965, with about 800 birds breeding
(with eggs and young). We estimated a total population of
1,491 breeding pairs (Table 1 ). Because nests high in Cocos
palms were difficult to detect, we undoubtedly overlooked
many, and our estimate of approximately 3.000 birds is
conservative. Although larger than the population estimated
by POBSP (Clapp & Sibley, 1971a), uncertainties about the
1965 survey coverage (F. Sibley, personal communication)
prevent us from knowing if Caroline's population has changed
over the past 25 years.

Phenology: On Christmas Island, the timing of egg laying
varies between colonies. In general, peak laying occurs from
March to May, and from November to December. On Caroline,
mating and nest-building were found in March 1990, but by
May only a few eggs had been laid. Eggs and young were found
in June 1965 (Clapp & Sibley, 1971a) and in September 1988
(present study). We found 246 nests in September 1988 and
determined the contents of 106: 103 held eggs, 3 held downy
chicks. The incubation period is 35-37 days (Dorward &
Ashmole, 1963), so all viable eggs had been laid within the
previous 40 days (mid-August to late September). Because
many nests were being built, we feel confident that laying
continued into October. Clearly more research is needed to
determine whether laying occurs in regular cycles.

Black Noddy ( Anous minutus) (Fig. 14)

The black noddy is widely distributed in the tropical
Atlantic and Pacific. It is abundant in the Line and Phoenix
Groups, with populations of 16,000 estimated in the Phoenix
Islands (Clapp, 1967) and over 46.000 in the Line Group.
Centers of abundance are Palmyra (20,000) and Christmas
(14.500) (Perry, 1980).

Distribution and Habitat Preference: The black noddy is
a tree-nesting species that on Caroline prefers tall stands of
Pisonia. The largest colonies (61% of the population) were
found in the grand Pisonia forests (to 25 m) on Pig and North
Pig, We found breeding birds on 18 motus. with colonies
exceeding 200 pairs in the Pisonia on Nake, Long, Arundel,
and Bird (Fig. 14). The only significant colony not primarily
associated with Pisonia was found on Tridacna, where
approximately 230 pairs nested in the tallest (ca. 8 m), most
central Toumefortia-Morinda forest. Black noddies always
nested in dense colonies near islet centers and were integral

components of these plant communities: their droppings,
coating the ground with a film of guano, constantly enriched
the islet's meager soils.

Numbers: Clapp & Sibley (1971a) estimated that
7,000 ± 25% birds were on Caroline. During our visit the
population was much larger: 5,122 pairs were estimated for Pig
and North Pig alone (Table 1 ). Basing our numbers primarily
on the densities of sampled colonies in Pisonia, we estimated
that nearly 8,400 pairs were nesting during our 1 988 visit. Our
population estimate approached 17.000 birds, to which an
unknown number of nonbreeding birds could be added. These
values place the Caroline population far above that for
Christmas, making it the largest known population in Kiribati.

Phenology: Black noddies were just beginning a new
breeding season. On 27 September we observed hundreds of
birds gathering Tournefortia leaves floating along the windward
shore (Long) or flying with fresh leaves to their nests (Pig,
North Pig). Of the 1,085 pairs counted on transect, 536(49%)
perched as pairs, were defending nest sites, or were building
nests. An additional 273 pairs were attending nearly-completed
nests but were not incubating. The remaining 276 pairs were
incubating. Thus, 75% of the pairs had not laid eggs. The
contents of 230 nests were unknown, although we assumed
they contained eggs because of the incubating positions of the
adults. Of 46 nests into which we could see, 45 held a single
egg, and one contained a downy chick less than 5 days old.

The breeding seasons for black noddies on Christmas
Island and Johnston Atoll peak in April and May (Schreiber&
Ashmole, 1970; Amerson & Shelton, 1976), where pairs are
highly synchronous, laying most of their eggs within a
2-3- month period. The Caroline colony, also synchronous,
but beginning egg-production in September, would be expected
to peak in October/November, six months out of phase with the
colonies further north. In 1990. however, black noddies were
just beginning to mate and nest in March, and by May some
were still sitting tightly on nests, while others had chicks in all
stages.

Blue-gray Noddy (Procelsterna cerulea)

Blue-gray noddies nest widely across the Pacific from the
Kermadec Islands to Hawaii. They are scattered throughout
the Line and Phoenix Groups. In the Line Islands, they were
formerly known to breed only on Christmas and Maiden
(Perry, 1980). Eggs are placed in nests minimally provided
with twigs and may be on coral rubble, sheltered under
vegetation, or under coral slabs to depths of I m (Rauzon
etai, 1984).

The blue-gray noddy was recorded as "present" on Caroline
by Perry ( 1 980). Clapp & Sibley ( 1 97 1 a) noted birds over the
lagoon hut saw none on land. When we approached Caroline,
we saw two from the ship and later observed three Hying across
the lagoon. We also saw three birds perched on the leeward
islets, one each on the reef flats of Nautonga and Fitei. A third
bird Hushed repeatedly from a small clearing around a pile of
bottles on Raurau. but we failed to find a nest. In March and
May 1 990, we observed blue-gray noddies on all of the Southern
Leewards, plus Azure and Nautonga in the Central Leewards.

146

In summer 1990, Alexandre Falconer found one small
blue-gray noddy chick, attended by its parents, on an open
expanse of coral rubble on Motu Eitei, the first breeding record
for Caroline. Eitei is adjacent to Raurau, which we predicted
was the most likely breeding location for this species.

Blue-gray noddies evidently breed in very small numbers
on Caroline. Nests are hard to find, given their cryptic placement,
the small number of birds present, and the extent of open
habitat (67.7 ha of herb mats and 41 .4 ha of consolidated coral
rubble).

White Tern (Gygis alba) (Fig. 15, PI. 3; Subchapter 1 . 1 . PI. 55)

The white tern is a widely-distributed pantropical species
occurring in moderate numbers throughout the Line and Phoenix
Groups. Clapp ( 1967) estimated 10,000 birds in the Phoenix
Group, and Perry (1980) estimated 17,050 birds for the Line
Islands.

Distribution and Habitat Preference: White terns, the most
widely distributed breeding bird on Caroline, nested on 32 of
the 39 motus (Fig. 15). The only islets not occupied were tiny
and sparsely vegetated.

White terns nested from 1 to 15 m above the ground,
wherever a branch or frond provided a relatively stable platform
in Tournefortia (PI. 3), Pisonia, Cordia, Pandanus , or Cocos
(Subchapter 1 . 1 , PI. 55). They did not form dense colonies but
were scattered from the edge to the center of each islet, even on
the windward sides, although they normally selected sites not
directly exposed to the prevailing trade winds. They utilized
isolated trees, scrub, or forest. An unusual departure from the
white tern's usual mode of "nesting" was an egg laid in an old
black noddy nest, 6 m up in an 8-m-tall Tournefortia on
Tridacna Islet.

White tern densities varied from islet to islet (Table 7). At
one extreme, we found only two nests on Raurau
(0.07/1,000 nr). Densities on other islets ranged from
0.75/1,000 m2 (Shark) to 6.67/1,000 nr (Nautonga) with a
mean density of 1 .38 pairs/1 ,000 nr of woodland. Overall, the
Windward Islets supported the highest densities. Although
white terns also nested in anthropogenic forests, their densities
were low: we believe that the low densities on South Island and
the Southern Leewards (Table 7) are attributable to man. Of
South' s 104.47 ha of vegetated land, only 4.2 ha (4.4%) was
native woodland (Subchapter 1.1, Fig. 50); fully 84% was
either Cocos (18.3 ha) or Cocos-lpomoea (62.5 ha) forest.
Although most of the Southern Leewards are covered in
unmodified natural forest, central Ana-Ana has been partly
cleared (0.21 ha) to accommodate thatched huts and a garden.
The activities of a family of four, with a dog and cat (until
October 1990), have apparently depressed the white tern
population on Ana-Ana and, perhaps, even on nearby islets.
We found no white terns on Ana- Ana during our visit, although
the Falconers, who vacated the atoll in summer 1991, assured
us that they occasionally nested.

Numbers: We used the total woodland area of each islet
to calculate islet populations (Table 1 ) from our transect data.
More birds were found on the largest islets except South Island.
We estimated 1,094 pairs for Nake, 751 pairs for Long, and
nearly 400 pairs for Tridacna; these 3 islets accounted for over

half the population (and over half the native woodland). We
estimated that 3,957 pairs bred on Caroline. This doubles the
numbers of Clapp & S ibley ( 1 97 1 a) and cited by Perry ( 1 980)
and exceeds by 3,000 the largest population formerly known
for the Line Islands.

Phenology: Of 569 pairs of white terns recorded on
transect, 437 were roosting without obvious signs of eggs or
chicks, 107 were incubating, and 25 had chicks (often adults
were not present). Of the 25 chicks recorded, 17 were downy,
7 retained extensive traces of down with remiges, and 1 was
almost ready to fly. Incubation takes about 36 days ( Ashmole,
1963); young may require from 40 to 96 days to fledged
(Gibson-Hill, 1950; Ashmole, 1968). Nearly all chicks were
far from Hedging and were less than 4 weeks old.

On Christmas Island, Schreiber& Ashmole (1970) found
that peak laying occurred in April-August each year, with
some laying in each month. On Caroline, Clapp & Sibley
(1971a) noted that about half of the birds had eggs, half had
young in June 1965. In March 1990, we found only a few eggs
and downy chicks, but in May a larger number of pairs were
breeding, with eggs and chicks in all stages. Although we
found that white terns on Caroline do lay during the peak period
on Christmas, it was clear that in 1 988 most eggs were laid after
mid-August.

Other Birds on Caroline Atoll

Seven species other than seabirds have now been recorded
on Caroline. Six of them are migrants (five shorebirds and a
long-tailed cuckoo). The reef heron is apparently resident,
although no nest has yet been found.

Reef Heron (Egretta sacra)

We found 15 reef herons scattered on 8 islands: Nake (1),
Long (2), Pig (1), Brothers (3), South (2), Mannikiba (2),
Matawa(l), and Emerald (2), as well as on the open reef flats
( 1 ). Although birds were found on both the seaward and
lagoonward sides of the islets, most were along the lagoon
edge, as also found by POBSP in 1965 (Clapp & Sibley,
197 la). We estimated that approximately 30 birds were using
the atoll. We found no signs of breeding. Of the 15 individuals
we observed, 5 were dark, 8 were white, and 2 were of the pied
morph.

Lesser Golden-plover (Pluvialis dominica)

This plover used the beaches and herb mats, generally to
seaward. In September 1988, we found them on Nake (1),
Long (4), Tridacna (4), and Mannikiba (1), estimating a total
population of 20-30 birds, the same number found by POBSP
(Clapp & Sibley, 1971a). In March 1990, we observed eight,
and in May, three, all in winter plumage.

Wandering/Siberian Tattler (Heteroscelus incanum or
H. brevipes)

In September 1988, we located 18 tattlers on 6 different
islets: Nake (3), Long (3), Crescent ( 1 ), Arundel (2), South (7),
and Emerald (2). All birds were either alone or in pairs and
generally remained in the intertidal zone, although they often

147

foraged on herb mats close to the beach scrub. The total
population was approximately 40 birds. Those few birds heard
were all H. incanum. We saw six tattlers in March 1990 and
several in May of the same year.

Ruddy Turnstone ( Arenaria interpres)

One turnstone w as found on the windward beach of Motu
Mannikiba in September 1988. and five on atoll beaches in
March 1990. The Caroline population probably does not
exceed 1 5 birds.

Bristle-thighed Curlew (Numenius tahitiensis)
(Subchapter 1.1, PI. 23)

The bristle-thighed curlew, common in the Line and
Phoenix Groups, is a widespread migrant to the low atolls of the
central and South Pacific during the boreal winter (Pratt el ai,
1987). One of the world's least-studied shorebirds, the species
is considered rare throughout its range (Johnsgard, 1981;
Marks et ai, 1990) and is a candidate for the US Fish &
Wildlife Service Endangered Species List (Gill. 1990). Clapp
& Sibley! 197 la) estimated 20 birds for Caroline in June 1965.

We counted 83 birds on 12 of Caroline's islets in 1988.
including the 3 large islands (Nake. Long, South) and motus in
the Windwards, Central Leewards, and Southern Leewards.
In March 1990, we saw 20 curlews on 10 islets during
incidental observations throughout the atoll, bringing the total
number of islets on which they have been recorded to 16. On
our return trip (May 1990) we only saw three curlews (only
eight islets visited). Undoubtedly, curlews occur on all islets,
utilizing essentially all plant communities. Although they are
most conspicuous on the beaches and reef flats, higher numbers
may actually forage in the forests during the day. The Falconers
(personal communication) note that small numbers of curlews
remain all year. They are least common between April and
August and most abundant after September/October. This
correlates with preliminary information from Rangiroa Atoll.
Tuamotu Archipelago (Gill. 1990; Gill & Redmond, in prep.).

Perimeter Habitats: On a complete perimeter count of
South Island in 1988, we found 29 curlews. Twenty-one were
foraging and loitering on the windward east coast, principally
above the beach crest on coral rubble interspersed with herb
mat. Similarly. 14 of 20 curlews found on Long and the
Windward Islets foraged along the windward beach crest, with
only 6 birds found on the lagoonward shores. Curlews were
equally common on windward and leeward shores in the
leeward islets, occupying habitats composed of coral rubble
and sand. While the numbers indicate that curlews showed a
preference tor windward shores, they may be biased because
most birds were there in late evening (19 birds on South).
Perhaps the\ use the relativeh open areas for roosting and
foraging at dusk. Certainly the largest concentrations ( 13. 14)
were found late in the day. We found our largest flock ( 14.
Sands Inlet, Nake) at 1 600 h, foraging on compacted, silt) sand
at the lagoonward end of the inlet, while single curlews dotted
the interislet channels and shallow tidal reef flats.

Vegetated Habitats: We found bristle-thighed curlew s on
natural herb mats, in Tournefortia scrub. Pisonia forest, and in

Cocos habitats, both in the healthy peripheral plantations and
w ithin the dying Cocos— Ipomoea woodlands ( Subchapter 1.1.
Fig. 36. PI. 34). One was captured in a mist net under a dense
Cocos canopy. Disintegrating plantations in the center of
South (54 ha) held a large population: calculated numbers
produced an estimate of 154 curlews. They foraged over the
//wwoefl-strewn ground, frequently using broken-topped
coconut trunks as lookouts. We also found 5 curlews on
transects in Pisonia forests up to 20 m tall on Nake (calculated
population, 41 ). They were foraging on the relatively open,
although dimly lit, forest floor.

Numbers: From the 1988 data we estimated a population
of ±300 curlews: 41 birds in Pisonia, 154 in Cocos-Ipomoea,
43 on the beaches of South Island (29) and the Sandy Inlet of
Nake (14), and another 62 scattered over the remainder of the
atoll. Because 154 of them were calculated from the sighting
of a flock of 7 curlews on one transect on South, there may be
a bias in our population estimate. Incidental observations made
off-transect did show, however, that curlews were common in
the Cocos-lpomoea woodlands, and we believe that the numbers
on the 104 ha that compose South Island approximated our
estimated density (about 1.5 birds/ha).

Bill Length: Bristle-thighed curlews show great variation
in bill length immediately after the breeding season. Because
birds of the year migrate south before their bills reach adult
length (R.Gill, personal communication ). the ratio of "long" to
"short" bills provides a rough estimate of juvenile survival. Of
3 1 curlews seen in September. 20 were clearly adult length.
7 were conspicuously shorter, and 4 were "intermediate"
(probably young birds). All March and May birds had long,
adult-sized bills.

Some subadults also remain on their Pacific wintering
grounds for up to 3 years, during which time they pass through
a flightless phase (Gill. 1 990: Marks el ai. 1990). Noflightless
birds were seen.

Foraging: We saw one curlew chase and capture a small
Polynesian rat at dusk on the south shore of South Island. The
bird bashed the rat on the coral rubble, then ran rapidly about
with the rat dangling from its bill. After about 5 minutes, the
bird swallowed the rat with vigorous gulps.

Polynesian rats, abundant on Caroline (especially in
Pisonia- and Cocas-dominated habitats), remain within the
forest during the day. but many move to the beach crest and tide
line at dusk. They provide abundant potential prey for curlews,
which can easily capture them on the open rubble. The
synchronous appearance of rats and curlews at the beach-
woodland interface at dusk may be part of the foraging strategy
of this large shorebird. The presence of curlews beneath the
forest canopy may also be partly associated w ith this source of
food.

Sanderling (Crocethia alba)

One sanderling in winter plumage was seen at water's
edge on the windward beach of Long Island on 27 September
1988. Although Sanderlings arc well-known fall migrants in
the Line and Phoenix Islands (Clapp & Sibley. 1967, 1968),
this is the first record for Caroline Atoll.

148

Long-tailed Cuckoo (Eudynamis taitensis) (Fig. 16)

The long-tailed cuckoo breeds in New Zealand and winters
in the southwest Pacific. The center of its winter range lies in
central Polynesia, but birds have been recorded as far as Palau
in the northwest and Pitcaim Island in the southeast. Although
occurring throughout French Polynesia and the Cook Islands,
it had not been recorded from the Line Islands prior to our
expedition (Bogert, 1937: Clapp& Sibley, 1 97 1 a,b; Pratt et al. ,
1987; Ellis etal, 1990).

We found long-tailed cuckoos on 4 of Caroline's 39 motus
(Fig. 16). We heard its distinctive monosyllabic and disyllabic
call notes on South, Long, and Pisonia, identified one on Nake,
and on 28 September collected a male in a mist net on Tr. 4,
Long Island ( USNM 607 191). Soon after our return home we
sent a description and photograph of this species to the Falconers:
they, and AKK, have since seen them several times on Motu
Ana- Ana in March, April, and May 1989-90.

All the cuckoo sightings were at canopy or subcanopy
level, and three of the four birds were found in Pisonia.
The South Island cuckoo was located in a Cocos canopy over
20 m high. The netted male flitted secretively within an
undisturbed, tangled low-canopy (4—6 m) Pisonia-Tournefortia
interface. We suspect that this elusive migrant occurs throughout
the mid-to-upper levels of Caroline's forest canopy.

These records establish the long-tailed cuckoo as a winter
visitor to Caroline Atoll. Our observations on four islets,
including the southernmost, northernmost, windwards, and
leewards, suggest that many individuals were present. A
March 1990 first sighting on Vostok (J. Phillips, personal
communication) further suggests that the species disperses
regularly to the Southern Line Group.

Other Vertebrates

Lizards

Although "small lizards" were observed on Caroline
in 1825 (Paulding, 1831), it wasn't until 1965 that the
first collections were made (Clapp & Sibley, 1971a). We
collected four additional lizard species, which increased the
known terrestrial herpetofauna from three to six (Table 8).
Although all are indigenous, the azure-tailed skink (Emoia
cyanura) is suspected of being partly dispersed by man
(Brown, 1956). All but two of the small lizard species known
from the Line Islands (Crombie, 1 990) have now been found on
Caroline.

Turtles

We found three Pacific green sea turtles ( Chelonia mydas),
a threatened species (McKeown, 1978), at Caroline in 1988.
Two were swimming over the lagoon reef flats, one west of
Arundel, the second east of Ana- Ana. The third was in the open
sea about 1 00 m west of South Island near the "boat entrance."
Ron Falconer has seen up to seven turtles in the lagoon in a
single day. In April and May 1990, AKK saw workers from
Tahiti capture and kill a minimum of four green turtles in the
lagoon; two more entered the lagoon during the following
4 months (R. Falconer, pesonal communication).

In March 1990, AKK and G. Wragg found three old nests,
presumably of this species, on the northwest coast of Nake
within 100 m of the northern tip of the islet. These are the first
known turtle breeding records for the atoll. Young (ca. 1922)
notes that the copra plantation laborers ate green turtles from
September to December each year. The February 1990 storm
added large amounts of sand to Caroline's shorelines, providing
potential new habitat for turtle nesting.

Terrestrial Mammals

None of the terriers (see Subchapter 1.1) that were
introduced to control rats on South Island in the early part of
this century (Young, ca. 1922) have survived (F. Sibley,
personal communication; R. Falconer, personal
communication). In May 1990 the Falconers kept a dog and a
cat on Motu Ana-Ana. Despite the fact that both animals
generally remained close to the settlement, the dog regularly
visited the other Southern Leeward Islets and accompanied the
family on excursions in their sailing canoe throughout the atoll.
As a result of our recommendations, the cat was removed from
Caroline in October 1990. The Falconers, with their dog,
vacated the atoll in mid- 1991.

Bennett ( 1 840) noted "rats of a red-brown color," the first
reference to rodents on Caroline. Dixon ( 1 884) found that rats
were "not numerous" and that they nested "just at the base of
the fronds" of the coconuts. Two specimens collected by the
POBSP proved to be Rattus exulans (Clapp & Sibley. 1 97 1 a).
They reported that rats were uncommon and restricted to South
Island.

The 19th and 20th century settlers found rats (presumably
R. exulans) to be extremely abundant and very destructive to
the coconut plantations. Maude (ca. 1938) states that rats
destroyed the nuts, and that they contributed greatly to the
eventual abandonment of copra enterprises on Caroline and
Flint. They voraciously devoured both growing and fallen
nuts, as well as dried copra. Being arboreal, they also lapped
the juices of the flower stalks, preventing nut development
(Young, ca. 1922). In a single year ( 1920) over 4,600 were
trapped on South Island (Maude, ca. 1938). Thousands more
were killed by terriers introduced to Caroline in a vain attempt
to control them.

We found rats on almost every islet; they were especially
abundant on South, Long, Nake, and in the vicinity of coconut
palms on smaller islets. We recorded rats during daylight hours
on most transects, especially within the Pisonia forests. At our
campsites on Long and South we noticed groups of 1 0-20 each
night, so tame as to approach within 1 m while we were eating.
The rats evidently undergo wide population fluctuations, as
they were less abundant in March and May 1990 than in
September 1988.

We suspect that rats periodically reach most motus, and
that the islets apparently lacking rats ( such as Noddy Rock ) are
too small and/or depauperate to support a resident population.
Because/?. e.v;</fl/f5isaknown seabird predator (Kepler, 1967;
Fleet, 1972; Norman, 1975), the restriction of some species
(i.e., red-tailed tropicbird) to small islets may be due to rat
populations on larger islets.

149

We found rats throughout the Southern Leeward Islets and
learned from the Faleoners that they are an abundant nuisance
on Ana-Ana. They trapped over 1 .300 animals in 2 years and.
like the pioneers before them, rely upon a dog to help keep them
at bay.

Marine Mammals

On March 14, 1990, members of the Line and Phoenix
Islands Expedition observed a minimum of 1 0 Pacific bottlenose
dolphins (Tursiops gilli) in the open sea about 500 m off the
southeast corner of South Island.

Coconut Crabs

The coconut crab (Birgus latro, Coenobitidae), the largest
terrestrial invertebrate on earth, ranges throughout the tropical
Indo-Pacific (Subchapter 1.1, PI. 56). It is highly esteemed as
a source of food throughout its range, and for this reason is rare
or absent on or near most inhabited islands. Because it is
heavily exploited by man, it is under consideration for
endangered species status (Reese, personal communication).
Since March 1990, dozens of Caroline's coconut crabs have
been killed for food and for preservation in formalin as curios
for the Tahiti tourist market. Because of the increasing numbers
of visitors to Caroline over the past 2 years, it is important that
Caroline's coconut crabs receive protection.

History: Young (ca. 1 922) was the first to mention coconut
crabs on Caroline. In 1910 he wrote that "hundreds of great
Coconut Crabs were seen: 40 large ones were caught by the
crew of the schooner in an hour" on South Island.

It is hardly credible that these enormous crabs, the dominant
terrestrial animal of the atoll environment, could have been
overlooked by all visitors prior to the 20th century. Perhaps
their populations had been reduced or extirpated by earlier
inhabitants. It is of interest in this regard that members of the
1934 Mangarevan Expedition saw no coconut crabs on nearby
Flint Island ( Fosberg, personal communication ), nor were they
mentioned in a historical summary paper on Flint by Maude
(ca. 1942b). Today Flint has perhaps the greatest density of
coconut crabs in the world (Kepler, 1990b).

Young ( ca. 1 922 ) noted that coconut crabs were considered
a great nuisance by plantation laborers, who killed them
mercilessly. Evidently the crabs dug up newly planted nuts and
snipped off emerging shoots. On the smaller islets, visited less
frequently than South, Nake, and Long, these depredations
were difficult to control. Thus the small motu plantations were
abandoned within a tew years of initial planting, resulting in a
remarkably rapid recovery of the original vegetation (see
Subchapter I.I. Ecological Succession section).

Distribution and Habitat Preference: In 1988 and 1990,
coconut crabs were abundant in the Cocos plantations of South
and Nake, and present, in varying densities, on 1 2 other motus
(Fig. 18). Although generally associated with ( 'ocos, we found
them in woodlands of Pisonia, Cordia, and Toumefortia, as
well as on rubble beaches (especiallj alter dusk). Although
capable of survi\ ing without coconut palms, these crabs appear
to seek them out. In the open understory of the tall plantations.

or in groves of only one or two palms, telltale piles of shredded
coconut husk fibers (Subchapter 1.1. PI. 57) disclosed the
crab's presence.

Because the prevalent coarse rubble substrates on Carol ine
are hard to burrow into, coconut crabs occupied a variety of
shelters: mounds of fallen coconuts and rotting palm fronds ( to
1.5 m high), piles of rubble pushed against tree roots, sand
burrows, tunnels within the/<?» (Subchapter 1.1, PI. 22), or
large cavities in the boles of mature Pisonia trees. Coconut
crabs also use a variety of shelters on the Tokelau Islands
(Yaldwyn & Wodzicki. 1979) and Flint (AKK. personal
observation).

Numbers: Though conspicuous and slow-moving, coconut
crabs are very difficult to count. Environmental variables such
as rainfall, tide, lunar cycle, and population size and age classes
all affect their activity (Reese. 1965: Helfman. 1977a,b).
Although unable to conduct mark-recapture studies, we did
make incidental observations on the numbers of individuals
seen during transect and perimeter surveys. Coconut crabs are
generally nocturnal, but we often found them during daylight,
at times exposed on coral gravel beaches close to the waterline.
Reese (personal communication) suggests that the abundance
of rats occupying the same habitat may "force" the crabs to be
more diurnal, as has been reported from the Indian Ocean. Our
estimate of the population on Caroline is approximately
2, 200 individuals, based on the number of daytime observations,
the area covered, and the fact that only one out of every three
or four individuals may be present on any given night ( Helfman,
1977b; Reese, 1987).

Foraging: Since the first detailed description of coconut
crabs in 1 705, their shy, curious habits have been the subject of
folklore, speculation, and misinformation (see Reyne. 1939).
No scientist has yet published a documented account of a
coconut crab actually opening a coconut (Helfman. 1979),
which is widely held to be their consummate foraging behavior.
Helfman is convinced that they do so, as he has found piles of
coconut fiber and observed crabs walking w ith husked, opened
nuts in places where he was the only other possible coconut
husker. We repeat Helfman's ( 1979) assertion that coconut
crabs do husk fallen coconuts. The piles of finely separated
fibers (Subchapter 1.1, PI. 57) we encountered are totalis
different from those produced by stick or machete husking, the
two methods commonly employed by Pacific peoples. The
crab tears virtually every fiber off individually, a process so
painstakingly slow it probably takes days. We did not observe
this on Caroline, but in March 1 990, AKK. on uninhabited Flint
Island, observed a large male coconut crab that had just husked
a coconut and was enlarging a small crack in the center of the
smooth nut in a manner similar to that described by Gardiner
(1907) in Reyne (1939, p. 297).

On Caroline we observed the aftermath of coconut crab/
sooty tern predation or scavenging. On Brothers Islet, several
entrances and pathways leading to coconut crab holes were
strewn with the feathered skeleta of adult sooty terns (and
possibly brown noddies ), along w ith numerous, freshly snipped
branches of Pisonia up to 0.7 m long (Subchapter 1.1.
Description and Ecology of the Motus section). This was also

150

recorded on Tridacna Islet by Clapp & Sibley (1971a) for sooty
tern eggs and chicks and by Reese (1987) and Helfman ( 1979)
on Enewetak, Micronesia.

Size: Living in a rich environment free of predators,
coconut crabs attain huge sizes on Caroline. The bodies of the
largest males were as wide as a full-sized, unhusked coconut
(Subchapter 1.1, PI. 56), giving them weights of at least 4 k
(Helfman. personal observation). Thorax widths for 10 crabs
(2 females with eggs. 8 males) averaged 129 mm. The thorax
of the largest male measured 200 mm across, making it. along
with many measured on Flint in 1990, one of the largest
recorded coconut crabs in the world (the previous record was
178 mm, Helfman. 1977a), with an age estimated to exceed
40 years (E. Reese, personal communication).

Conservation: Attributes of International Significance

When Bennett (1840) stated that "no reefs we had seen
could compete with those of Caroline for novelty and beauty,"
he was seeing an essentially pristine ecosystem through the
eyes of a well-traveled naturalist. Caroline is stunning, but its
value in today's shrinking world goes well beyond its physical
beauty. Caroline' s exceptional attributes need to be elucidated,
for the atoll has remained essentially unknown, even to some
who have evaluated its worth (King, 1973; Garnett, 1983,
1984). Man's presence anywhere, especially on pristine or
near-pristine islands, generally brings rapid, often irreversible,
changes. There are few, if any, islands remaining in the Pacific
that can claim the impressive array of natural features exhibited
by Caroline (Nicholson & Douglas, 1969). We believe that it
is imperative that this atoll, which has managed to escape large-
scale human disturbance, should remain undeveloped.

Caroline was inhabited from 1988 to 1991 by a single
family who lived a spartan, ecologically sound lifestyle. There
are no roads, vehicles, stores, jetties, or services (water, sewage,
or food), and no communication. There is no passage into the
lagoon or safe sea anchorage.

One of the most important of Caroline's attributes is its
relative lack of disturbance. Aside from obvious human
impacts on South. Nake. and Ana-Ana, the majority of its
motus are dominated by indigenous vegetation and its reefs are
basically pristine. There is no obvious pollution to alter the
chemistry of the lagoon, beyond the flotsam and jetsam that
spatter the windward beaches. It is thus an exceptionally clear
and clean ecological laboratory that presents a picture of
lagoon ecosystems "before" extensive disturbance by man. and
one that provides the marine biologist with an unparalleled
opportunity to study undisturbed natural communities. The
atoll is rich in marine vertebrates and invertebrates; the maze
of reefs and coral heads in the lower half of the lagoon has the
highest recorded density of living Tridacna (20/. 25 m2) ever
recorded ( Sirenko & Koltun. Subchapter 1 .4. this volume), one
of the few undisturbed world populations of this species
(Subchapter 1.1. PI. 26).

Caroline's many islets of different sizes provide excellent
examples of soil and vegetation development, accompanied by
variations in the diversity of bird life (Fig. 17). Many of its

disturbed islets have recovered so remarkably they are almost
indistinguishable from those which have remained pristine.
The changing shapes of the islets, bearing emerging and mature
plant communities, graphically portray a natural terrestrial
atoll ecosystem. Caroline's concentric pattern of plant
community development and the relationships of these
communities to islet size, shape, and location on the atoll rim
will continue to provide insight into evolutionary processes on
atolls if they are left undisturbed.

Caroline's insular flora, typical of central equatorial islands
in their natural state and covering 70% of the atoll's land area,
is of both national and international importance. The 27 extant
plant species are 85% indigenous (possibly up to 93%), an
extremely high figure for anywhere in the world. Six of the
seven plant communities are natural. Lushly wooded, Caroline
possesses some of the largest and grandest Pisonia (Pisonia
grandis) forests known (Subchapter 1.1, PI. 43), occurring on
29 islets. Although not as majestic as the prime forests on
Washington and Fanning (Northern Line Group), which enjoy
a heavier rainfall, those on Caroline are some of the finest
representatives of this forest community in the entire Pacific.
The 62 ha of Pisonia forest may well cover a larger area than
on any other Pacific atoll.

Caroline possesses significant stands of the hardwood kou
(Cordia subcordata), a tree that is now rare in the Pacific.
Caroline's groves (Subchapter 1.1, PI. 27), though small and
often occurring in mixed native woodlands, total 26 ha. possibly
the greatest area on any Pacific atoll. Its extensive coverage of
tree heliotrope (Tournefortia argentea) is also notable: scrub
and forests of this species form 40% of the atoll woodlands
(Subchapter 1.1. PI. 47). Caroline's groves are some of the
most unmodified in the Pacific: elsewhere Tournefortia is
typically restricted to coastal fringes surrounding anthropogenic
plantations (R. Fosberg. personal communication).

Caroline offers many opportunities for ecological research
under reasonably pristine conditions. Valuable clues as to the
nature of underground water supplies may lead to a better
understanding of the regulation of water supplies on inhabited
islands. Marine biological and biomedical research could
unearth clues as to the causes and treatment of ciguatoxicity of
fishes and crabs. Such topics are increasingly important as
more islands are subjected to disturbance and pollution. For
example, the abundant red snapper (Lutjanus vaigiensis) and
red spotted crab (Carpilius maculatus), both of which are
notorious for their potent poisons, are safe to eat on Caroline.
Associated with Caroline's plant communities are
1 1 species of breeding seabirds numbering well in excess of
1,000,000 individuals. The populations of most of these
species are of national importance (Table 9). Caroline has the
fifth largest red-footed booby colony (Subchapter 1.1, PI. 51)
in the world. Its black noddy and white tern (PI. 3) populations
are the largest in Kiribati. Under the 1975 Republic of Kiribati
Wildlife Conservation Ordinance (amended in 1979). all
known seabirds, migrant shorebirds. and endemic land birds
are "fully protected throughout the Gilbert Islands" (Garnett,
1983, p. 128). However, their protected status is in doubt on
Caroline, due to attempts to lease the island for development.

151

Caroline deserves protection similar to five closed areas on
Christmas Island and seven island sanctuaries in the Line and
Phoenix Groups (Garnett, 1983).

Caroline is an important wintering ground for the bristle-
thighed curlew, a rare shorebird and candidate for the US Fish
& Wildlife Service Endangered Species list. Some subadults
remain all year on the atoll. Adult curlews pass through a
flightless phase on Pacific islands, and Caroline provides a
predator-free environment for this vulnerable phase of the
curlew's life history.

Caroline is exceptional in harboring a robust population of
coconut crabs (Subchapter 1.1, Pis. 22, 56). These large
invertebrates are abundant in the Cocos plantations of South
and Nake and are found in good numbers in the indigenous
Pisonia forests on most of Caroline's larger motus.

Although green turtles are not abundant on the atoll,
worldwide populations of these marine reptiles have suffered
so greatly from overexploitation that remote, predator-free
islands such as Caroline provide important, though small,
sanctuaries. Since 1978 the Pacific green sea turtle has been
reclassified by the United States Department of the Interior as
threatened and the Pacific hawksbill sea turtle as endangered.

From an archaeological point of view, Caroline houses
one intact Tuamotuan marae ( ancient religious site) and another
smaller site, partly destroyed by storms. The main site
(Subchapter 1.1, Fig. 3, PI. 36), basically undisturbed since the
1 870' s, is a relic of prehistoric occupation worthy of protection,
being the only one of its kind in the Line and Phoenix Islands.

Currently Caroline Atoll is owned by the government of
the Republic of Kiribati and does not enjoy any legal protection
(Garnett, 1983; Government of Kiribati, personal
communication). Over the last 50 years it has been leased to
private individuals who have scarcely altered the atoll. The
benign management of the past is no guarantee for the future,
and from October 1989 to the present, pressures to develop the

atoll have mounted rapidly. Proposed schemes included an
airstrip, a blasted channel through the reef, a hotel, a casino,
logging, and commercial harvest offish and lobsters. In March
1990, commercial harvesting of fish, the taking of coconut
crabs, and illegal killing of green turtles began, emphasizing
that no island, however remote, is guaranteed protection through
isolation. In addition, during the past 2 years Caroline has
become more visited than ever before, mostly without the
knowledge or consent of the Kiribati government.

There are many reasons why Caroline is inappropriate for
resident tourists or development (remoteness, distance from
medical aid, no regular water supply, no passage into the
lagoon, etc.; see Kepler. 1990a). Caroline could support a
limited number of ship-based ecotourists each year.

Recommendations for an international preserve began in
January 1989. During the 1990 ICBP expedition to the Line
Islands, the team leaders discussed conservation matters with
Kiribati government officials and key scientists in French
Polynesia. Fortunately, documentation was obtained of illegal
land clearing and wildlife disturbance during two visits to
Caroline (Kepler, 1990a,b,c). The Kiribati government is
considering altering their plans for the development of Caroline
in favor of wildlife preservation. During summer 1990, French
customs officials in Tahiti temporarily banned the exploitation
of Caroline by French Polynesian nationals.

As of December 1 990, The Nature Conservancy of Hawaii
has restated its interest in establishing a triple-island preserve
on Caroline, Vostok, and Flint and has begun discussions with
the Kiribati government on Tarawa. The fate of these special

islands may rest upon the results of these negotiations.

We have a number of people to thank for their assistance in this
project; however, our acknowledgments for Parts 1 and 2 of the
manuscript. Ecological Studies of Caroline Atoll. Republic of
Kirabati. South-central Pacific Ocean are listed at the conclusion of
Part 1 . and are not repeated here. Again, it is our sincere pleasure to
(hank these individuals.

152

TABLE 1

Estimated number of breeding seabird pairs on Caroline Atoll, September 1988.

Location

Nake
Long
Windward Islets

Bo'sun Bird

Windward

Crescent

Atibu

North Pig

Pig

Skull

North Brothers

Brothers

Noddy Rock

North Arundel

Arundel

Tridacna
South
South Nake Islets

Pandanus

Danger

Booby

Coral

Lone Palm

Kota

Mouakena
Central Leeward Islets

Mannikiba

Blackfin

Matawa

Emerald

Shark

Scarlet Crab

Nautonga

Azure

Reef-flat

Bird

Fishball
Southern Leeward Islets

Raurau

Eitei

Pisonia

Kimoa

Ana-Ana

Red-tailed Masked Brown Red-footed Great Lesser Sooty

Tropicbird Booby Booby Booby Frigatebird Frigatebird Tern

5
47

105
69

25

496

522

659

808

163

207

28

5

31

17

14

118

25

9

*

*

37

*

111

-

32

26

139

-

52

-

28

2

48

-

12

1

8

-

184

287

*

4

5

1

3

230

*

118

11

2

7

2

29

10

31

17

14

26

14

21

3

56

Brown Black Blue-gray White
Noddy Noddy Noddy Tern

1,094
751

-

390

814

179,800

207

986

8,400

10

_

-

20

28

-

36

60

_

76

3.194

-

82

1,928

_

23

40

-

8

15

-

80

-

_

11

249

-

11

230

-

163

-

_

26

_

-

33

33

-

2

1

-

6

-

161

176

37

-

3

-

7

150

37

125

7

32

42

329

5

-

6
134

110
164

69

50

227
396

381

52
37
6
15
9
3

195
11
13
83
44
2

10

2

48

Total Estimated
Pairs

56

189

2,427

56 188,200 1,491 8,392

3,957

Breeding confirmed in 1989 or 1990.

153

TABLE 2

Stages in the breeding cycle of the red-tailed

tropicbird, Caroline Atoll, 27-29 September 1988

(ages after Stonehouse, 1962).

Approximate Age in
Nest Stage Days From Laying No. Nests

Juv.

90-133

18

Remiges

69-89

4

Scapulars

58-68

5

Downy

44-57

6

Egg

0-43

21

Pairs on

Territory

-

2

TABLE 3

Stages in the breeding cycle of the boobies of Caroline Atoll,
21-29 September 1988.

Nest stage'/Approximate age in days from laying

Flying Pairs on

Species Juv. Juv. Scapulars Remiges Downy Naked Eggs Territory

Masked >164 145-164 115-144 89-114

Brown > 1 64 144-164 114-144 88-114

Red-footed - >150 111-150 75-110

55-88 45-54 0-44
54-88 44-54 0-44
54-74 45-53 0-45

No. nests in each stase

Masked

Brown

Red-footed

40

many

38

4

33

34

-

-

3

8

29

-

919

1.270

For descriptions of nest stage, see C. B. Kepler ( 1978).

TABLE 4

Density of red-footed booby nests in occupied Toumefortia habitats on
islet groups, Caroline Atoll. September 1988.

Islet Group

Number Estimated Area of Nests/1.000 nr

Occupied Number Toumefortia of Available

Islets Nests (nr) Habitat

Nake

1

496

300,650

1.6

Long

1

659

322.000

2.0

Windward Islets

8

434

25 1 ,900

1.7

South Nake Islets

7

319

59,801)

5.3

Central Leewards

6

239

197,500

1.2

Southern Leewards

4

74

39.600

1.9

rotal

27

2,221

1.170,550

1.9

154

TABLE 5

Stages in the breeding cycle of frigatebirds on Caroline Atoll.
21-29 September 1988.

Nest stage/Approximate age in days from laying

Species

Juv.

Primaries Scaps. Downy Naked

Eggs

Great
Lesser

191-220
181-210

101-190 81-100 (56-80)a (56-80)3
91-180 71-90 56-70 46-55

No. nests in each stage

0-55
0-45

Great

Lesser

22

4

46 27 ( 30J )
13 4 5 0

19
0

Duration of naked and downy chick stages are lumped because it was often
impossible to see into canopy nests.

Colony Location

Long Island, A

Long Island. 1 N

Long Island. 1 S

TABLE 6

Sooty tern colonies on Caroline Atoll. 27-28 September 1988.

Calculated

Population Nest Approx. Weeks

Area (Mean Pairs ± SE) Stage From Laying

44,100 m: 127,449 ±30.429 hatching eggs, 4-5

downy chicks

24,200m2 41,382 ± 5,808 chicks with 7-10

short tails,
juv. plumage

6,400 nr 10.9441 1.536 " 7-10

Bo"sun Biid Islet,
old
new

Total

3,375 m: 6,883 ± 1,575 fledglings

3,375 m2 1,538 ± 758 new eggs

75,075 nr 188.196 ±40.106

11-12
1-9

TABLE 7

Density of white terns on occupied islets by islet group, Caroline
Atoll. September 1988.

# White

Density

#

Vegetated

tern

(pairs/

Islet Group

Islets

Area (ha)

Pairs

1,000 nr)

Nake

1

66.63

1,094

1.64

Long

1

49.60

751

1.51

South

1

86.10

381

0.43

Windward Islets

9

36.09

1,164

3.23

South Nake Islets

6

8.50

122

1.44

Central Leeward Islets

9

33.56

408

1.22

Southern Leeward Islets

4

6.47

37

0.57

All Occupied Islets

31

286.88

3,957

1.38

155

TABLE 8

Lizards collected on Caroline Atoll, 1965-1988.
Specimens:

Species

Clapp & Sibley 1 97 1 a Present Study

Mourning gecko

Lepidodactylus lugubris
Polynesian gecko

Gehyra oceanica
Snake-eyed skink

Cryptoblepharus poecilopleurus
Moth skink

Lipinia noctua'

Emoia impar
Azure-tailed skink
Emoia cyanura

USNM 158355-57
USNM 158353-54

USNM 158358

USNM 299773
USNM 299772

USNM 299768-70

USNM 299771

1 USNM 158358 has recently been reidentit'ied by R.I. Crombie as Lipinia
noctua. not Emoia nigra, as reported in Clapp & Sibley (1971a).

TABLE 9

Species

Comparative abundance of Caroline's breeding seabirds in the Line Group.

Estimated Population Comparative Abundance in the Line Group

Red-tailed tropicbird

300 '

Masked booby

400

Brow n boobv

40

Red-looted boobv

7.000

Great frigatebird

6.1(H)

lesser frigatebird

200+

Sooty tern

912.000

Brovwi noddv

3.000

Black noddy

17.000

Blue-gray nodd)

<10

White tern

8.000

Second largest population
Fourth largest population
Third largest population
Third largest population
Third largest population

Third largest population
Third largest population
Largest population (largest mi Kiribati)

Largest population (largest in Kiribati)

' Based upon nest count in 1990.

156

BROWN BOOBY
RED-TAILED TROPICBIRD

WINDWARO
ISLETS

Fig. 1. Caroline Atoll, Republic of Kiribati, with newly-named islets.
RED TAILED TROPICBIRD 1988

22 -
20 -
18 -

CAROLINE ATOLL

Fig. 2. Distribution map of breeding red-tailed tropicbirds and brown
boobies on Caroline Atoll. September 1988. In this and the following
distribution maps, arrows indicate concentrations of breeding birds.

a

z

0)

<

a.

o

z

16
14
12

10

uveniles

primaries

scap

downy

eggs

1.0

0.5

>-
<

Q

a.

LU
0.

CO
LU

X

u

o
6

160

APRIL

140

MAY

120

100 80

60

40

JUNE JULY

AUG.

NO. DAYS

20

SEPT.

Fig. 3. Approximate laying dates for red-tailed tropicbird nests found on Caroline Atoll in September 1988. In this and the following similar figures the numbers
of nests begun during a given time period (bars) were determined by tallying each nest into one of several age classes (bar labels): bar widths indicate
length in days for each class. The dotted line connects mean number of surviving clutches begun per day for each class. For example, a juvenile found
in September began its egg stage in the previous May or June. The number of days are counted backwards from field observations.

157

MASKED BOOBY

50% of population

CAROLINE ATOLL

SCALE i .14 ooo
100 O too 1000

'■••■' I I l_

MASKED BOOBY 1988

9)

40

~~"\

-

/

i

\

-

/

\

32

/

24

/

„

,' *ggs
_i 1

z

>
<
■J 16

s

/

rer

iiges

<

Q.

o .

scapulars
■ '

0

180 160

APRIL

140 120 100 80 60 40

MAY JUNE JULY AUG.

Fig. 4. Distribution map of breeding masked boobies on Caroline Atoll,
September 1988.

Fig. 5. Approximate laying dates for masked booby nests found on Caroline
Atoll in September 1988.

RED- FOOTED BOO

ROLINE ATOLL

Fig 'i Distribution map oi breeding red footed boobies on Caroline Atoll, Fig. 7. Approximate laying datesfoi red footed boob) nests foundonCaroline
September 1988 Atoll in September 1988

158

■ GREAT FRIGATEBIR
► LESSER FRIGATEBIRD

Fig. 8. Distribution map of breeding great and lesser frigatebirds on Caroline
Atoll, September 1988.

50

40

30

-i 20

</>

IT

<

a.

6 10

z

juveniles

primaries

GREAT FRIGATEBIRD 1988

scaps

naked/
downy

_l_

eggs

220 200 180 160

FEB. MARCH APRIL

140

120

100

MAY

JUNE

NO. DAYS

80 60

JULY

40

AUG.

20 o

SEPT.

Fig. 9. Approximate laying dates for great frigatebird nests found on Caroline Atoll in September 1988. See Figure 3 for explanation.

159

15

10

o

z

<

LESSER FRIGATEBIRD 1988

u veni les

primaries

scaps

down

220-
FEB

200

MARCH

180 160

APRIL

140

MAY

120 100

JUNE

80 60

JULY

40

AUG.

20 0

SEPT.

NO DAYS

Fig. 10. Approximate laying dates for lesser frigatebird nests found on Caroline Atoll in September lc

9 MAR 89

OMARS MAY 90

'SEP 86. • 89 &

AR 90. -* MAY 90 IN -31

*r SEP 88. • FEB 89 4 MAY 90

#MAR89,-^MAY89IN=3I

• * "JUN 65, O MAY 90
IN =31

juveniles
N number of colonies

BO' SUN BIRD

• •* SEP 88 IN = 2I

O AUG 894 MAY 90
O AUG 894MAY 90
-' : JUN 65 ,° AUG 89 IN 4

CAROLINE ATOLL

0 MO l«*0 1NO

J I I I '

SOOTY TERN 1988

BO SUN
BIRD

LONG IS

"a"

L2

Fig. 1 1. Distribution mapofbreeding sooty ternsonCaroline Atoll, September Fig. 12. Approximate lasing dales for sooty tern young found on Caroline
1988-July 1990.. Atoll in September 1988. See Figure 3 for explanation.

160

BROWN NODDY

BLACK NODDY

CAROLINE ATOLL

CAROLINE ATOLL

Fig. 13. Distribution map of breeding brown noddies on Caroline Atoll
September 1988.

Fig. 14. Distribution map of breeding black noddies on Caroline Atoll,
September 1988.

LONG-TAILED CUCKOO

Fig. 15. Distribution mapol breeding white terns on Caroline Atoll. September Fig. 16. Tentative distribution map of the long-tailed cuckoo on Caroline
1988. Atoll. The species most likely utilizes all well-wooded islets

161

SEABIRD BREEDING
SPECIES DIVERSITY

COCONUT CRAB

CAROLINE ATOLL

5C»l€

BO 0

900

1

1 1

,000

CAROLINE ATOLL

Fig. 17. Seabird breeding species diversity by islet. Caroline Atoll.

Fig. 18. Distribution map of coconut crabs on Caroline Atoll.

162

PI. 1 . Incubating red-tailed tropicbird, Bo'sun Bird Islet, Caroline Atoll, 25 September 1 988. The nest scrape is in fine coral rubble under
a Toumefortia shrub.

PI. 2. Masked booby adult with egg on coarse coral rubble substrate with Portulaca mat. Nake Island. Caroline Atoll,
26 September 1988.

163

PI. 3. White lern adult with egg in typical nest site, a dead Ttmrmforiia branch. South Island. Caroline Atoll. 23 September 1988.

164

1.3 First Records of the Long-tailed Cuckoo
(Eudynamis taitensis) on Caroline Atoll,
Southern Line Islands, Republic of Kiribati

DAVID H. ELLIS . CAMERON B. KEPLER, ANGELA K. KEPLER*, and KATINO TEEB'AKI
US Fish & Wildlife Service. Patuxent Wildlife Research Center, Laurel, Maryland, USA

US Fish & Wildlife Service, Patuxent Wildlife Research Center, Southeast Research Station, Athens, Georgia, USA
Wildlife Conservation Unit, Christinas Island, Republic of Kiribati

Introduction

The long-tailed cuckoo {Eudynamis taitensis) performs
what is perhaps the most remarkable overwater migration of
any land bird (Lack. 1959). It breeds in New Zealand and is
known to winter in the islands of the central Pacific Ocean, with
stragglers seen as far as Palau to the west and Pitcaim to the east
(Bogert. 1937). Although the Line Islands are along the
northeastern perimeter of this range, the cuckoo has never been
reported for the Line or Phoenix Islands (Pratt etai, 1987). In
1883. when Dixon (1884) visited Caroline Atoll (10°S. 150°W)
at the southeastern end of the Line Islands, he reported that a
colleague had heard "the notes of a singing bird." but no land
bird was collected. In June 1965, biologists from the Pacific
Ocean Biological Survey Program visited Caroline Atoll but
failed to detect land birds (Clapp & Sibley. 1971 ).

We encountered the long-tailed cuckoo on four islets
during our 22-29 September 1988 survey of all 39 islets of
Caroline Atoll as part of a research team from the Soviet
oceanographic research vessel Akademik Korolev. Harsh
monosyllabic or disyllabic call notes, presumably of this species,
were first heard on three islets: South. Pisonia. and Long (see
Kepler et a/., Subchapter 1.1. this volume, for islet locations).
Then, on 25 September, a single bird, probably of this species,
was heard and briefly seen on South Island. On 26 September,
an individual was positively identified (CBK; ca. 25 m; xlO
binoculars) on Nake Island.

In an attempt to capture this species, or any other
undiscovered land bird, we operated mist nets at three locations
on the atoll. In 14.5 net hours (daylight hours only; ATX
4-shelf nets. 2.6 x 12 m; mesh size 36 mm) beneath a 10-15 m
Cocos canopy on South Island, only a single bristle-thighed
curlew (Numenius tahitiensis) was captured. On Long Island
two nets along a Pisonia-Cocos interface (canopy at 6-8 m)
were unsuccessful in 27.5 net hours. Finally, on 28 September

we collected a male long-tailed cuckoo on Long (US National
Museum No. 607191) in 1.5 net hours along a Pisonia-
TournefortialCordia interface with a short canopy (4-6 m)
where two of us ( AKK, KT) had heard and followed a cuckoo-
sized land bird for about 20 minutes the day before.

The following measurements of the specimen were taken
immediately after collection: mass 125 g, length 411 mm
(central rectrices still growing so measurement was to the tip of
the worn rectrices adjacent to central rectrices). culmen
25.4 mm, and wing chord 179 mm. Soft part colors within
30 minutes of death, compared with Smithe's ( 1 975 ) color key,
were ridge of bill. No. 219 Sepia; lateral margin of bill, No. 86
Pale Neutral Gray; lower mandible, No. 53 Buff-yellow; iris.
No. 124 Buff; foot pad, No. 153 Trogon Yellow; and dorsal
surface of foot and tarsus. No. 150 Bunting Green.

These records establish the long-tailed cuckoo as a
winter visitor to the Line Islands. Since our 1988 visit, we
learned from correspondence and personal discussions with
the atoll's only human inhabitants and wardens, Ronald and
Anne Falconer, that cuckoos occasionally occurred in
April. 1989, near their dwelling on Motu Ana-Ana. the
southernmost leeward islet. This was confirmed by AKK with
further sightings on two subsequent trips to Caroline in March
and May 1990 (Kepler. 1990). Our original observations at
five widely scattered locations (the most distant were 9 km
apart) suggested that several individuals were present on the
atoll during our visit. Subsequent observations suggest that
this species disperses regularly to Caroline Atoll and perhaps
some others of the better-vegetated Line and Phoenix Islands
as well.

We are pleased to thank Mr. Harold J. O'Connor and Mr. Steve
Kohl (US Fish & Wildlife Service) and Professor Alia V. Tsyban
(Goskomgidromet. USSR ) for organizing the Third Joint US-USSR
Bering & Chukchi Seas Expedition that made these observations
possible.

165

1.4 A Study of the Benthic Communities of

Caroline Atoll (Line Islands, Pacific Ocean)

BORIS I. SIRENKO and VLADIMIR M. KOLTUN

Zoological Institute, USSR Academy of Sciences, Leningrad, USSR

Introduction

The study of coral atolls, which constitute one of the most
highly productive biological systems in oligotrophic tropical
waters, is of considerable theoretical and practical value. On
one hand, rapid human population growth over the past few
decades, with the concomitant shortage of protein-rich foods,
has driven a never-ending search for new protein sources. The
study of coral reefs as highly productive biological systems
would be helpful for establishing marine farming facilities. On
the other hand, burgeoning industrialization and the increasingly
intensive use of all natural systems has caused considerable
damage to coral reefs (Gomez & Yap, 1985). There is,
therefore, an urgent need for coral reef monitoring. Uninhabited
Caroline Atoll, situated far from principal sea routes, would be
considered an excellent monitoring site. With this in mind, we
spent a week studying the status of the coral reefs and surveying
the benthic communities around South Island and the Southern
Leeward Islands ( Kepler e t al. , Subchapter 1 . 1 , this volume ) in
Caroline Atoll (Fig. 1 ). Regrettably, time constraints did not
permit a study of sufficient scope and depth. The present paper
is therefore limited to a general description of benthos
distribution in the accessible portions of the reef and to an
account of a uniquely interesting reef situated within the atoll
lagoon.

At low tide, the Caroline Atoll Lagoon is linked to the
ocean by shallow passes of no greater than 0.5 m in depth.
Several narrow intralagoonal reefs subdivide the central lagoon
at low tide. The depth of the lagoon does not exceed 10 m. The
largely sandy bottom includes isolated patches of fragile coral
colonies, mostly Acropora.

With few exceptions, the outer side of the reef near the
southern islands has the classical structure of most coral reefs
(Preobrazhenskiy, 1986). The narrow shallow-water lagoon
facing the islands gradually becomes a reef flat cemented by
encrusted calcareous algae. This is rather extensive, averaging
562 m: range 396-759, N = 100 sites (Kepler et al..
Subchapter 1.1, this volume). The reef Oat is surmounted by
widely spread limestone coral knolls or "coral heads." Further
seaward, the reef flat, with the usual channels and overhanging
ledges, breaks off. At a depth of 5-6 m it becomes a buttress
zone (i.e., a radically crosscut sloping terrace consisting of
individual spurs or benches that become narrower in the
oceanward direction I. Ii is the latter zone that constitutes the
main portion of the reef, with its abundant growth of corals
(Acropora, Pocillipora, etc. ). It is this growth that accounts for
the origin and continued development of the reef and the entire
coral atoll. It was the status of the coral settlements in the

£:&)

3 / ^ /

4 / rX-* ' Acropora-Tridacna Reef

5(QA *

Fig. 1. Location of the Acropora-Tridacna reef and of hydrobiological
section I-I of Caroline Atoll.

buttress zone that afforded an indication of the "health" of the
reef. Our studies showed that not less than 50^ of the surface
area of the outer slope of the Caroline Atoll reef was covered
by living coral colonies. Assessment on this basis, using the
method of cross sections and areas (Gomez & Yap, 1985),
showed the condition of the reef to be sound. The notorious
"crown of thorns" starfish (Acanthaster planet), responsible
for the devastation of reefs in other parts of the Pacific Ocean,
was seen only once.

The littoral zone of the lagoon facing South Island was
sandy. The sand was filled with holes made by burrowing
PolychaetaandBalanglossi. Also present, sometimes in clusters,
were the mollusks Cerithium columna and Ccrithiuni sp., a few
Calappa sp. crabs, and gastropod mollusksAfe/ar/iape undulata
and Nerita plicata, the latter settling by water's edge on trees
whose branches dip all the way down to the water.

166

The northeastern side of South Island was washed by
strong oceanic surf throughout our stay at the atoll. The
prevailing wind in this area is easterly (see Kepler et <//..
Subchapter 1.1. Appendix 2, this volume), an observation that
is consistent with the presence on the eastern side of South
Island of extensive coarsely fragmented coral limestone banks
2-2.5 in high, attributable to windstorms. These storm banks
contained large numbers of fresh bivalve mollusk (Asaphis cf.
violascens) shells, providing evidence fora large population of
these burrowing animals. The shallow islandward lagoon and
reef flat, with its extensive sand lenses, is inhabited by the
aforementioned bivalves, large numbers of gastropods of the
genera Cerithium, Drupa, Cyprae, et cetera, as well as by
predatory mollusks of the genus Conus, including unusually
large specimens of Conns ebreus (up to 55 mm long).

The coral reef on the western side of South Island received
closer scrutiny. The distribution of organisms over the
hydrobiological section (I-I) is indicated in Fig. 2.

The beach of the island showed many large red hermit land
crabs Coenobitaperlata, as well as Ocypode sp. crabs. Next to
the beach lay a zone of lifeless coral limestone remarkable for
its extreme lack of living organisms during the season of the
year when we visited. This zone became very narrow to the
south. The fact that the sea had until recently extended this far
was evidenced only by small numbers of Melarhaphe undulata
clinging to the underside of sun-baked coral slabs. It appears
that when the wind direction changes to westerly, powerful surf
inundates this area during high tide. During our visit, however,
surf continued to crash against the opposite side of the atoll.
The dry, lifeless zone ended with a ledge approximately 1 m in
height. The water reached this ledge during high tides. The
dominant fauna consisted of small hermit crabs, Grapsus sp.
crabs, holothuria. and the mollusks Nerita plicata and Thais

armigera. Further seaward, the stones in the never-drying
pools of the islandward part of the lagoon harbored the following
fauna: Diadema sp. urchins; the mollusks Cypraea moneta,
Cerithium columna, Cerithium sp., Vasum tuhiferum. Conns
ebreus, and Conus sponsalis; the crabs Eriphia sp. and Actaea
sp.; three species of sea cucumbers (Holothuroidea); Linckia
sp. starfishes; and black and grey brittle stars (Ophiuroidea).
Part of the coral limestone in the same area was encrusted with
the calcareous algae Porolithon sp. and Lithothamnion sp. To
seaward, the littoral pools became deeperand somewhat larger,
interdigitating to form the islandward lagoon. The bottom in
this area consisted of coral limestone with small patches of
coral fragments and sand. The bottom was strewn with
limestone knolls that resulted from the intense buildup of
calcareous algae and corals of the genera Porites, Pocillopora,
Aeropora, and Montipora. These limestone knolls were
cavernous and often caved in under the weight of an adult
human. The fissures and caverns contained numerous small
crabs and other crustaceans, as well as sponges. Also prevalent
were algae bushes (Halimeda sp.).

The bottom sloped upward towards the tidal strip, where
it was succeeded by a highly tenacious, firmly attached fauna
of a very particular kind. Unfortunately, because our
examination was confined to the upper portion of the reef flat,
the richly populated interior portion remained unstudied. The
cavities and fissures of the top portion of the reef flat revealed
an abundance of corals of the aforementioned genera. Dead
coral colonies, as well as all of the old coral limestone formations,
were densely encrusted with calcareous algae that cemented
the reef-flat surface together. There were many gastropods
(large Turbo argyrostomus and the smaller Drupa ricina,
D. morum, D. grossularia, etc. ) as well as large orange-colored
hermit crabs. The perpetually surf-washed tidal areas at the

Q«fe f(3

Corals:

Crustaceans:

(put)

- Aeropora sp.

&

- Ocypode sp.

<st&

- Pocillopora sp.

¥%

- Grapsus sp.

tMV\

- Montipora sp.

**

- Eriphia sp.

C=0

• Porites sp.

$k

- Aklaea sp.

if

- Coenobita periata

Algae:

A=

- Paguridae sp.

<***.

- P'rolithion sp.

MBSr<

- Lithotamnion sp.

Mollusks:

0 - Malarhaphe undulata y^

© - Nerita plicata \U/a,

0 • Cerithium spp.

ff - Nassa sp.

(2) - Turbo argyrostromus

<Q - Vasum tuhiferum

0 - Conus spp.

(S - Cypraea spp.

Q - Thais armisera, and Drupa spp.

Echinodermata:
■ Heterocentrotus sp.

Diadema sp.

"7a[ - Lichia sp.

<y%, - Ophiuroidea
^^ - Holothuroidea

Fig. 2. Distribution of organisms on the reef on the western side of South Island of Caroline Atoll (Section I-I).

167

vers edge of the Hat exhibited magnificent specimens of the sea
urchin (Heterocentrotus sp.), with spines as thick as cigars.
The limestone caverns sheltered smaller crabs (Eriphia sp. and
\i taea sp. ). The edge of the reef Hat dropped abruptly some
5-6 m, in places forming an overhanging ledge that overlooked
the most abundantly populated and significant portion of the
reef, namely the buttress zone with its rich variety of coral
species.

Caroline's lagoon enclosed a unique natural structure, an
unusual Acropora-Tridacna reef ( Fig. 1 . Chapter Frontispiece).
This reef extended from the southernmost islet (Ana-Ana) in
the Southern Leeward Islands into the interior of a shallow
lagoon. In the middle of the lagoon, it divided into two
branches: a northern branch that ended in the center of the
lagoon, and an eastern branch that extended all the way to the
opposite shore. The reef, in effect, partitions the lagoon into
two parts; a wide channel had to be dug to permit a small flat-
bottomed rubber boat to pass. The width of the reef varied from
1 5 to 20 m, attaining 30 m at its fork. Most of the reef surface
stood above sea level, with many of the Tridacna maxima and
corals partially drying out during very low ebb tides.

The distribution of organisms over a typical cross section
of the reef is shown in Fig. 3. Five distinct zones (two edges,
two lateral strips, and a central strip) were clearly evident.

The edge strips, populated mostly by fragile colonies of
Acropora secale, A. palmerae, and Acropora sp., are growth
zones where vital activity keeps increasing the reefs width.
The dominant corals of the genus Acropora completely cover
the steep slopes of the reef as well as the neighboring rudimentary
reefs that do not reach the surface of the water in the lagoon.
The lush coral growth resembles a huge domed topiary.

The thickness of the living coral layer ranges from 20 to 70 cm.
Parts of the reef slope exhibited scarps that probably formed as
a result of the collapse of fragile coral colonies unable to carry
their own weight. It is at these scarps that measurements were
made of the longest coral branches, some of which were found
to attain a length of 70 cm. The spaces between the fine coral
tentacles of the reef slope were filled with beaded ( moniliform )
algae (Halimeda sp.). Also evident at the base of the reef slope
were a few, mostly large, specimens of Tridacna maxima.
Beginning at a depth of 5 m and extending to the reef base were
sparse, isolated growths of fungiform coral ( Fungia granulosa).
The colonies of Acropora were fine and fairly fragile, as is
typical in the still-water portions of many lagoons. It was risky
to approach the edge of the reef, since the loose and brittle coral
colonies tended to crumble underfoot.

Beyond the edge zones (including the reef slopes, and
extending several tens of centimeters into the reef flat) lie
lateral zones where Tridacna maxima clams are especially
abundant. These lateral zones have a width of 3-5 m on either
side of the reef (Fig. 3). Particularly striking was the very high
density of Tridacna. firmly attached to the reef surface by the
byssus. which formed veritable bunches atypical for these
large mollusks in other regions. Not uncommon were bunches
of five or six clams attached to one another by the byssus. The
average shell length for Tridacna found atop the reef was
12-13 cm, with a maximum length of 1 9-20 cm for individual
specimens. No less striking was the variety of coloration
exhibited in their mantles, probably occasioned by the presence
of symbiotic algae. We were able to identify as many as
10 shades of blue, green, and light brown pigmentation.
The average Tridacna population density in the lateral zone

^ - coral

(Acropora spp.)

| - mollusks

(Tridacna maxima)
<X> - coral

(Fungia granulata)

«* - Holothurioidea
(Ludwigothuria sp.)

mm - algae

(Porolithon sp.)

$ - algae

(Halimeda sp.)

r

edge i later;

Coral limestone

central

lateral | edge

ZONES

i ross section of the Kcropora rrirfacna reef in the lagoon of Caroline Atoll.

168

was estimated at 35 living individuals/m2, with some 0.25-m2
patches containing as many as 20 of these mollusks. In addition
to Tridacna clams, thu lateral strips included small colonies of
Acropora sp. coral and some less abundant lamellate Montipora
sp. Also present were algae {Halimeda sp. ) occupying spaces
between Tridacna shells and coral colonies, as well as calcareous
fouling algae encrusting the mollusk shells and dead portions
of coral colonies.

A central strip 7-10 m in width accounts for most of the
reef flat. This site harbored most of the dead coral colonies and
empty Tridacna shells, firmly bonded to the reef surface by
calcareous fouling algae (Porolithon sp.). Nearly 809r of the
surface area of the central strip was covered by these algae,
which gave the middle portion of the reef added strength. The
live Tridacna population density in this strip was one order of
magnitude lower (an average of 4-5 individuals/m:) than in the
lateral strips. The strip also contained widely scattered colonies
of Acropora sp. and, more commonly . bluish scales of Montipora
sp. coral colonies, including small clumps of Halimeda sp.
Another denizen of the central strip is the holothurian
Ludwigothuria sp. (approximately 1 specimen/nr ).

The central and lateral strips exhibited a few small holes
4-8 cm in diameter surrounded by empty shells of smaller,
thoroughly consumed Tridacna. These were probably the
remains of meals taken by predators, namely small octopuses
that are able to open bivalve shells without damaging them.
The typical division of the reef into the aforementioned /ones
is disrupted in some places. In these instances, the central strip
of the reef contained a shallow depression of friable structure
populated by corals and Tridacna, with small amounts of
calcareous fouling algae. This is probably an intermediate
stage in the merging of individual smaller reefs with the larger
reef traversing the entire lagoon. One such smaller and still-
growing reef is shown on the left side of the reef cross section
in Fig. 3.

It is interesting to note that the Acropora-Tridacna reef is
a natural farm producing large bivalves of commercial value.
The efficiency of the "farm" is difficult to assess without data
on its productivity. However, a count of the Tridacna present
is possible. According to the most conservative estimates, the
surface of the Acropora- Tridacna reef, extending 1 km into the

lagoon of Caroline Atoll, contains approximately 300,000
Tridacna clams, the raw weight of their flesh equaling not less
than 30 tons. It should be noted that the Acropora-Tridacna
reef actually investigated was not the only one in the lagoon
(Chapter Frontispiece: Kepler cial.. Subchapter 1.1. this volume.
Figs. 47,48,57). There were, in fact, several such reels, and if
we assume that they are of similar structure, the above figures
can be multiplied by an appropriate factor to show the actual
reserves of valuable food protein available in Caroline"s
waters.

The unusually high density of living Tridacna in Caroline's
lagoon was especially striking, exceeding any previously known
populations of both Tridacna maxima and T crocea (usually
more abundant in other parts of the World Ocean). For
example, the Palau Islands (western Pacific) were reported to
have just six T. maxima and 153 T. crocea within an area of
1 . 1 00 m- ( Hardy & Hardy, 1 969), whereas the same area on the
Acropora-Tridacna reel in Caroline Atoll lagoon contained
16,500 T. maxima. Richards ( 1 985 ) found that T. maxima in the
Tuamotus numbered 6-20/nr at Takapoto Atoll, and up to
60/rrr at Reao Atoll. Although these are the highest densities
previously reported, they do not equal the numbers found in the
densest patches on Caroline.

The very considerable effect of such an enormous mass of
large mollusks on the entire atoll is also noteworthy. This is
because in symbiosis with zooxanthellae. which are the principal
food of the Tridacna clam, the latter experience intensive
growth and in turn enrich their habitat with proteins (Ricard&
Salvat. 1977). Dataconcerning the natural Acropora-Tridacna
reef could be put to use in creating artificial reefs in other parts
of the ocean to achieve considerably enhanced productivity.

The sound condition of the coral reef around Caroline
Atoll, as well as the presence in its lagoon of a uniquely
interesting natural feature in the formof the Acropora-Tridacna
reef, may be deemed sufficient grounds for organizing a marine
reserve in the area.

The authors are grateful to US colleagues Kay and Cameron
Kepler and D. Smith, participants in the 47th voyage of the research
vessel Akademik Korolev, lor their active assistance in field work on
Caroline Atoll. They also wish to thank Yu.N.Latypovforidentifying
the corals.

169

Chapter 1 References

Admiralty Chart, No. 979. (1965). Central Pacific Ocean:

Islands between 160° East and 150° West Longitude

J. D. Potter. London.
Amerson, A. B„ Jr. & Shelton. P. S. ( 1 976). The natural history

of Johnston Atoll, central Pacific Ocean. Atoll Res. Bull.

192, 1-479.
Arundel. J. T. ( 1 890). The Phoenix Group and other islands of

the Pacific. New Zealand Herald, Auckland. July 5 and 12,

1890.
Ashmole, N. P. ( 1 963 ). The fairy tern Gygis alba on Ascension

Island. Ibis 103b. 365-378.
Ashmole, N. P. ( 1968). Breeding and molt in the white tern

(Gygis alba) on Christmas Island, Pacific Ocean. Condor

70. 35-55.
Bavliss-Smith.T. P. (1973). Ecosystem and Economic System

of Ontong Java Atoll, Solomon Islands. Cambridge University

Ph. D. Thesis.
Bennett. F. D. ( 1 840). Narrative of a whaling voyage round the

globe from the year 1833-36, Vol. I. BibliothecaAustraliana

46, 1-402.
Bogert, C. ( 1937). Birds collected during the Whitney South

Sea Expedition. XXXIV. The distribution and the migration

of the long-tailed cuckoo (Urodynamis taitensis Sparrman).

Am. Mus. Novit. 933, 1-12.
Braley. R. D. (1987). Spatial distribution and population

parameters of Tridacna gigas and T. derasa. Micronesica

20, 225-246.
Broughton. W. R. (1804). A Voyage of Discovery to the North

Pacific Ocean. . . Performed in His Majesty's Sloop

Providence. . . in 1795, 1796. 1797, 1798. T. Cadell and

W. Davies. Rep.. London.
Brown. W. C. ( 1956). The distribution of terrestrial reptiles in

the islands of the Pacific Basin. Proc. 5th Pac. Sci. Conf. 3A.
1479-1491.
Bryan, E. H. ( 1942). American Polynesia and the Hawaiian

Chain. Tongg Publ. Co.. Honolulu. 253 pp.
Butler. A. G. & Strecker. H. (1884). Memorandum of the

butterflies, etc. of Caroline Island. In Report of the Eclipse

Expedition to Caroline Island May 1883. Natl. Acad. Sci.

Mem. 2, 92-96.
Catala, R. L. A. ( 1957). Report on the Gilbert Islands: Some

aspects of human ecology. Atoll Res. Bull. 59, Oct. 3 1 .
Christopherson, E. ( 1927). Vegetation of Pacific equatorial

islands. B. P. Bishop Mus. Bull. 44. 79 pp.
Clapp, R.B. (1967). Line Islands. Unpublished report. 32 pp.
Clapp. R. B. & Sibley. F. C. ( 1967 ). New records of birds from

the Phoenix and Line Islands. Ibis 109. 122-125.
Clapp. R. B. & Sibley, F. C. ( 1968). Additional new records

of birds from the Phoenix and Line Islands. Ibis 110,
573-575.
Clapp. R. B. & Sibley. F. C. ( 1971a). Notes on the vascular
flora and terrestrial vertebrates of Caroline Atoll, Southern
Line Islands. Atoll Res. Bull. 145. Washington, D.C..
Smithsonian Institution. 20 pp.

Clapp, R. B. & Sibley. F.C. (1971b). The vascular flora and
terrestrial vertebrates of Vostock Island, south-central

Pacific. Atoll Res. Bull. 144. 1-10.
Clark. L., Ricklefs, R. E. & Schreiber, R. W. ( 1983). Nest-site

selection by the red-tailed tropicbird. Auk 100, 953-959.
Clay, H. (1961). Narrative report of botanical fieldwork on

Kure Island. 3 October 1959 to 9 October 1959. Atoll Res.

Bull. 78. 1-4.
Commercial Advertiser, The. ( 1873). From Caroline Island.

Honolulu. Two-line note in newspaper, March 5, 1873.
Crombie.R. I. (in prep. ). Herpetofauna of Oceania: Island lists

and species distributions.
Danielsson. B. ( 1954). Native topographical terms in Raroia,

Tuamotus. Atoll Res. Bull. 32, 92-96.
Darwin. C. R. ( 1 842 ). The Structure and Distribution of Coral

Reefs. Smith. Elder & Co.. London, xii + 214 pp.
Degener, O. & Gillaspy. E. (1955). Canton Island. South

Pacific. Atoll Res. Bull. 41. Aug. 15.
Diamond. A. W. (1971). The ecology of the sea birds of

Aldabra. Philos. Trans. R. Soc. hand. B260. 561-571.
Diamond. A. W. ( 1973). Notes on the breeding biology and

behavior of the magnificent frigatebird. Condor 75.

200-209.
Diamond, A. W. (1975a). The biology of tropicbirds at

Aldabra Atoll. Indian Ocean. Auk 92, 16-39.
Diamond. A. W.( 1975b). Biology and behavior of frigatebirds

Fregata spp. on Aldabra Atoll. Ibis 117, 302-323.
Dinsmore, J. J. ( 1972). Sooty tern behavior. Bull. Fla. State

Mus. Biol. Sci. 16(3). 129-179.
Dixon. W. S. ( 1884). Notes on the zoology of Caroline Island.

In Report of the Operations of the American Expedition

to. . . Caroline Island May 1883. Natl. Acad. Sci. Mem. 2,

90-92.
Dorward, D. F. & Ashmole. N. P. ( 1963). Notes on the biology

of the brown noddy (Anous stolidits) on Ascension Island.

Ibis 103b. 447-457.
Doty. M. S. & Morrison. J. P. E. ( 1954). Interrelationships of

the organisms on Raroia aside from man. Atoll Res. Bull. 35.

Nov. 30.
Ellis, D. H., Kepler. C. B., Kepler. A. K. & Teeb'aki, K. (1990).

Occurrence of the long-tailed cuckoo {Eudynamis taitensis)

on Caroline Atoll. Kiribati. Emu 90. 202.
Ellsworth. S. G. (ed. ) ( 1990). The Journals of Addison Pratt:

Being a Narrative of Yankee Whaling in the Eighteen

Twenties, A Mormon Mission to the Society Islands. . .. in

the Eighteen Forties and Fifties. University of Utah Press,

Salt Lake City, Utah. 606 pp.
Ely.C. A. & Clapp. R. B. ( 1 973 ). The natural history of Laysan

Island, northwestern Hawaiian Islands. Atoll Res. Bull. 171.

xi + 361.
Emory, K. P. (1947). Tuamotuan religious structures and

ceremonies. B. P. Bishop Mus. Bull. 191, 102 pp.
Findlay, A. G. (1884). Directory for the Navigation of the

South Pacific Ocean. 5th Edition. London.

171

Fleet, R. R.( 1972). Nesting success of the red-tailed tropicbird

on Kure Atoll. Auk 89. 651-659.
Fleet. R. R. ( 1974). The red-tailed tropicbird on Kure Atoll.

A. O. U. Monogr. 16, 1-64.
losberg, F. R. ( 1936). Vegetation of Vostok Island, Central

Pacific. B. P. Bishop Mas. Spec. Publ. 30, 19.
Fosberg. F. R. ( 1949). Atoll vegetation and salinity. Pac. Sci.

3. 89-92.
Fosberg. F. R. (1953). Vegetation of Central Pacific atolls.

Atoll Res. Bull. 23. 1-26.
Fosberg, F. R. ( 1955). Northern Marshalls Expedition. 1951-

1952. Land biota: Vascular plants. Atoll Res. Bull. 39, 1-22.
Fosberg, F. R. (1956). Military Geography of the Northern

Marshalls, prepared under the direction of the Chief

of Engineers. US Army, and the US Geological Survey.

320 pp.
Fosberg, F. R. ( 1959). Vegetation and flora of Wake Island.

Atoll Res. Bull. 67, 1-20.
Fosberg, F. R. (ed. ) (1965). Man's Place in the Island

Ecosystem: A Symposium. 10th Pacific Science Congress,

Honolulu, Hawaii, 1961. Bishop Museum Press. 264 pp.
Fosberg, F. R. (1968). Critical notes on Pacific island plants 2.

Micronesica 4(2), 255-259.
Fosberg, F. R. ( 1977a). Coral island vegetation. In Biology

ami Geology of Coral Reefs, Vol. III. (O. A. Jones &

R. Endean, eds.), pp. 256-277. Academic Press.
Fosberg, F. R. ( 1977b). An irresponsible scientific expedition.

Atoll Res. Bull. 219, 4-5.
Fosberg. F. R. ( 1983). The human factor in the biogeography

of oceanic islands. C. R. Soc. Biogeogr. 59(2), 147-190.
Fosberg, F. R. ( 1985). Present state of knowledge of the floras

and vegetation of emergent reef surfaces. Vol. 5. Proc. 5th

Int. Coral ReefCongr., Tahiti. 1985.
Fosberg. F.R.& Sachet, M.-H.l 1969). Wake Island vegetation

and flora. 1961-1963. Atoll Res. Bull. 123, 1-15.
Fosberg, F. R. & Sachet, M. -H. ( 1987). Flora of the Gilbert

Islands, Kiribati, checklist. Atoll Res. Bull. 295, 1-33.
Fosberg. F. R. & Sachet, M. -H. (MS). Flora of the Phoenix

Islands. Unpublished manuscripts. Pt. 1, 85 pp. ; Pt. II. 52 pp.
Fowler, H. W. (1901). Fishes from Caroline Island. Proc.

Acad Nat. Sci. Phila. 53, 324-326.
Frisbie, R. D. (1944). The Island of Desire. Garden City. New

York. 234 pp.
Garnett, M. C. (1983). A Management Plan for Nature

Conservation in the Line and Phoenix Islands. Pt. I:

Description, 318 pp.: Pts. II, III: Policy and

Recommendations, 131 pp. Unpublished reports.
Garnett, M. C. ( 1984). Conservation of seabirds in the South

Pacific region: A review. ICBP Tech. Pub. 2. 547-557.
Gibson Hill. C. A. ( 1950). Notes on the birds of the Cocos-

Keeling Islands. Bull. Raffles Mus. 22. 212-270.
Gill, R. F. (1990). Ecology of curlews in French Polynesia

during spring. In Summary of the Proceedings from the

Workshop on Bristle thighed Curlews. Anchorage, Alaska

(R. F. Gill & C. M. Handel, eds). 22 pp.
Gill.R. E& Redmond, R. L. ( in prep.). Distribution and habitat

associations of bristle thighed curlews on Rangiroa Atoll.

15 pp.

Gomez, E. D. & Yap. H. T. ( 1 985 ). Setting up an international

coral reef monitoring system. In Comprehensive Global

Monitoring of the World Ocean (A. V. Tsyban &

A. S. Sarkisyan, eds.). pp. 16-31. Proceedings of the First

International Symposium. Vol. 2. Gidrometeoizdat

Publishers. Leningrad, (in Russian)
Guppy. H. B. (1906). Observations of a Naturalist in the

Pacific Between 1896 and 1899. Vol. II. Plant-Dispersal.

Macmillan & Co., New York. 627 pp.
Hardy. J. T. & Hardy, S. A. ( 1969). Ecology of Tridacna m

Palau. Pac. Sci. 23, 467-472.
Hatheway, W. H. (1953). The land vegetation of Arno Atoll,

Marshall Islands. Atoll Res. Bull. 16. April 30.
Helfman. G. S. (No date). Ecological considerations for the

management of the coconut crab. Birgus latro (L).

Unpublished manuscript. 8 pp.
Helfman. G. S. ( 1977a). Ecology and Behavior of the Coconut

Crab, Birgus latro (L. >. M. S. Thesis. University of Hawaii,

Parti. 109 pp.
Helfman, G. S. (1977b). Temporal Distribution of Crabs.

M. S. Thesis, University of Hawaii. Part II. 158 pp.
Helfman. G. S. (1979). Coconut crabs and cannibalism. Nat.

Hist. 88(9), 76-83.
Holden, E. S. (1884). History, in Report of the Eclipse

Expedition to Caroline Island May 1883. Natl. Acad. Sci.

Mem. 2, 20-22.
Holden, E. S. & Qualtrough, E. F. ( 1 884). Description of the

island. In Report of the Eclipse Expedition to Caroline

Island Max 1883. Natl. Acad. Sci. Mem. 2. 22-26.
Hydrographer of the Navy. (1931. 1982). Pacific Islands

Pilot. Vol. III. Ministry of Defense. Taunton. Somerset,

England.
Johnsgard. P. A. ( 1981 ). The Plovers. Sandpipers, and Snipes

of the World. University of Nebraska Press. Lincoln.

Nebraska.
Kepler. A. K. ( 1990a). Caroline Atoll: Fact sheet, including

conservation attributes and recommendations. Unpublished

manuscript. 12 pp.
Kepler, A. K.( 1990b). Flint Island: Fact sheet. Unpublished

manuscript. 6 pp.
Kepler, A. K.( 1990c). Vostok Island: Fact sheet. Unpublished

manuscript. 6 pp.
Kepler. A. K. (1990d). Report: ICBP 1990 Line and Phoenix

Islands Expedition, Februarj 20-May31, 1990. Unpublished

manuscript. 14 pp.
Kepler, A. K. (in prep. ). Ecological studies of Vostok Island.

Republic of Kiribati.
Kepler. A. K.. Kepler. C. B. & Ellis. D. II. (1991). Ecological

studies of Caroline Atoll. Southern Line Islands. Republic ol

Kiribati. Part I. History, physiography, botany, and islet

descriptions (Subchapter I. I , this volume).
Kepler. C. B. (1967). Polynesian rat predation on Laysan

albatrosses and other Pacific seabirds. Auk 84. 426-429.
Kepler. C. B. (1969). Breeding biology of the blue-faced

booby on Green Island. Kure Atoll. Publ. Nuttall Ornithol.

ClubS, 1-97.
Kepler. C. B. (1978). Ihc breeding ecology of sea birds on

Monito Island. Puerto Rico. Condor80, 72-87.

172

Kepler, C. B. & Kepler, A. K. (1978). The sea birds of Culebra

and its adjacent islands, Puerto Rico. The Living Bird 16,

21-50.
Kepler. C. B. & Kepler, A. K. (1989). Trip report, Soviet-
American oceanographic expedition to the central Pacific:

Hilo. Hawaii to Singapore, 9 Sept.-31 Oct. 1988. Unpublished

manuscript. 9 pp.
King. W. B. ( 1973). Conservation status of birds of central

Pacific islands. Wilson Bull. 85. 89-103.
Lack, D. (1959). Migration across the sea. Ibis 101, 374-399.
Lamberson, J. O. ( 1 987). Natural history of terrestrial vascular

plants of Enewetak Atoll. In Vol. II. The Natural History

ofEnewetak Atoll. Chapter 3, pp. 17-35. US Department

of Energy. 348 pp.
Lamoureaux, C. ( 1961 ). Botanical observations on leeward

Hawaiian atolls. Atoll Res. Bull. 79. 1-10.
Lucett. E. ( 185 1 ). Rovings in the Pacific from 1837 to 1849.

with a Glance at California, Vol. II. By a merchant long

resident at Tahiti. 2 Vols., Longman. Brown and Green.

London, xi + 371 pp.
Mac Arthur, R. H. & Wilson, E. O. ( 1967). The Theory of Island

Biogeography. Princeton University Press. Princeton. New

Jersey. 203 pp.
Markham. Sir Clements. (1904). The Voyages of Pedro

Fernandez de Quiros. 1595-1606, 2 Vols.. Hakluyt

Society. Series II, Nos. 14-15. The Hakluyt Society.

London.
Marks, J. S.. Redmond, R. L., Hendricks, P., Clapp, R. B. &

Gill. R. E. ( 1990). Notes on longevity and flightlessness in

bristle-thighed curlews. Auk 107(4), 779-781.
Maude, H. E. (ca. 1938). Caroline Island. Unpublished

manuscript. 8 pp.
Maude, H. E. (ca. 1942a). Caroline Island: Part L. Ms. from

author, of unknown origin. 6 pp. (printed but not published).
Maude, H. E. (ca. 1942b). Flint Island: Part N. Author's

manuscript., of unknown origin. 4 pp. (printed but not

published).
Maude, H. E. (1953). The British Central Pacific Islands: A

report on land classification and utilization, Proc. 7th Pac.

Sci. Congr. 60, 89-97.
Maude, H. E. (1968). Of Islands and Men. Oxford University

Press, London and New York. 397 pp.
Maude, H. E. (No date). Bibliography of Books. Pamphlets.

and Articles on the Phoenix and Line Islands. Ms., printed

but not published, in author's possession. 4 pp.
McKeown. S. (1978). Hawaiian Reptiles and Amphibians.

The Oriental Publishing Company, Honolulu, Hawaii.

80 pp.
Moul, E. T. (1957). Preliminary report on the flora of Onotoa

Atoll. Gilbert Islands. Atoll Res. Bull. 57. 1-48.
Mueller-Dombois, D.. Bridges, K. W. & Carson. H. L. ( 198 1 ).

Island Ecosystems: Biological Organization in Selected

Hawaiian Communities. Hutchinson Ross Publishing

Company, Stroudsburg. Pennsylvania. 583 pp.
Nelson, J. B. ( 1969). The breeding ecology of the red-footed

booby in the Galapagos. J. Anim. Ecol. 38, 181-198.
Nelson, J. B. (1976). The breeding biology of frigatebirds — a

comparative review. The Living Bird 14, 1 13-155.

Nelson, J. B. (1978). The Sulidae: Gannets and Boobies.

Oxford University Press. Oxford, England. 1,012 pp.
Newell, N. D. (1956). Geological reconnaissance of Raroia

( Kon Tiki ) Atoll, Tuamotu Archipelago. Bull. Am. Mus. Nat.

Hist. 109.311-372.
Nicholson, E. M. & Douglas, G. (1969). Draft Check List of

Pacific Oceanic Islands. Micronesica 5(2), 327^63.
N. I. D. (Great Britain Naval Intelligence Division). (1943).

Pacific Islands, Vol. II, Eastern Pacific. H. M. Stationery

Office. London, xvi + 739 pp.
Niering. W. A. (1956). Bioecology of Kapingamarangi Atoll,

Caroline Islands: Terrestrial aspects. Atoll Res. Bull. 49.

1-32.
Niering, W. A. ( 1 963). Terrestrial ecology of Kapingamarangi

Atoll, Caroline Islands. Ecol. Monogr. 33, 131-160.
Norman, F. I. (1975). The Murine rodents Rattits rattus,

exulans, and non'egicus as avian predators. Atoll Res. Bull.

182, 1-13.
Parham, B. E. V. ( 197 1 ). The vegetation of the Tokelau Islands

with special reference to the plants of Nukunono Atoll. N. Z.

J. Bot. 9. 576-609.
Paulding, H. ( 1 83 1 ). Journal of a Cruise of the United States

Schooner Dolphin. ... G. & C. & H. Carvill, New York.

iv + 258 pp.
Perry, R. (1980). Wildlife conservation in the Line Islands,

Republic of Kiribati (formerly Gilbert Islands). Environ.

Consen: 7,311-318.
Perry, R. & Garnett, M. (No date). The Natural History and

Birdlife of Christmas Island. ICBP, Cambridge, England.

36 pp.
Phillips, N.J. ( 1984). Spatial relationships of Pisonia grandis

and Cocos nucifera. Cousin Island Research Station (ICBP.

Seychelles Islands. Indian Ocean) Technical Report. 9 pp.
Phillips, N. J. & Phillips, V. E. (1990). Monitoring vegetation

changes on Cousin — A comparative summary of changes

since 1976. Cousin Island Research Station (ICBP. Seychelles

Islands, Indian Ocean) Technical Report. 43 pp.
Pilsbry, H. A. & Vanatta, E. G. (1905a). Mollusca of Flint and

Caroline Islands, in the central Pacific. Proc. Acad. Nat. Sci.

Phi la.. 57,291-293.
Pilsbry, H. A. & Vanatta, E. G. (1905b). On some Pacific

Cerithiidae. Proc. Acad. Nat. Sci.. Phila., 57. 787-789.
Pratt, H. D., Bruner, P. L. & Berrett, D. G. (1987). A Field

Guide to the Birds of Hawaii and the Tropical Pacific.

Princeton University Press, Princeton, New Jersey.
Preobrazhenskiy. B. V. (1986). Modern Reefs. Nauka Press,

243 pp. (in Russian)
Rauzon, M. J., Harrison, C. S. & Clapp, R. B. (1984). Breeding

biology of the blue-gray noddy. J. Field Ornithol. 55.

309-321.
Reese, E. S. ( 1965). The ecology of the coconut crab. Birgus

latro(L. ). Bull. Ecol. Soc.America46(4). 191-192. Abstract.
Reese, E. S. ( 1987). Terrestrial environments and ecology of

Enewetak Atoll. In The Natural Histoiy of Enewetak Atoll,

Vol.1., pp. 187-202.
Reyne. A. ( 1939). On the food habits of the coconut crab

(Birgus latro L. ), with notes on its distribution. Arch. Neer.

I. Zool. 3. 283-320. (in French)

173

Ricard, M. & Salvat, B. (1977). Faeces of Tridacna maxima

( Mollusca-Bi val via), composition and coral reef importance.

In Proceedings: Third International Coral Reef Symposium,

Rosenstiel School of Marine and Atmospheric Science.

University of Miami. Florida.
Richard. G. (1985). Richness of the Great Sessile Bivalves in

Takapoto Lagoon. 5th Int. Coral Reef Congress, French

Polynesia, pp. 368-37 1 .
Ridley. H. N. ( 19301. The Dispersal of Plants Throughout the

World. L. Reeve & Co.. Ltd.. Ashford, United Kingdom.

744 pp.
RNZAF. (1986). Aerial photographs of Caroline Island

(Gilberts), No. 6569. taken by the Royal New Zealand Air

Force Plan No. 1036 9d-h, Department of Survey and Land

Information, Wellington, New Zealand.
Sachet, M. -H. ( 1 967 ). Coral Islands as ecological laboratories.

Micronesica 3. 45—49.
Sachet, M. -H. ( 1983 ). Takapoto Atoll. Tuamotu Archipelago:

Terrestrial vegetation and flora. Atoll Res. Bull. 277, 1^46.
Sachet. M. -H. & Fosberg. F. R. (1983). An ecological

reconnaissance of Tetiaroa Atoll, Society Islands. Atoll Res.

Bull. 275. 1-28.
St. John, H. ( 1973). List and summary of the flowering plants

in the Hawaiian islands. Pac. Trap. Bot. Garden Mem. 1
519 pp.
St. John, H. & Fosberg, F. R. ( 1 937 ). Vegetation of Flint Island,

Central Pacific. Bernice P. Bishop Museum, Occas. Pap..

Honolulu. Hawaii. 12(24). 3—1.
St. John. H.&Philipson,W.R.( I960). List of the flora of Oeno

Atoll. Tuamotu Archipelago, south-central Pacific Ocean.

Trans. R. Soc. N.Z. 88(3). 401-403.
Schreiber, R. W. & Ashmole. N. P. ( 1970). Seabird breeding

seasons on Christmas Island. Pacific Ocean. Ibis 112.

363-344.
Seelye. C. J. ( 1950). Rainfall and its variability over the central

and south-western Pacific. N.Z. J. Sci. Tech. [B] 32(2).
11-24.
Shaw. H. K.A.( 1952). On the distribution of Pisonia grandis

R. Br. (Nyctaginaceae), with special reference to Malaysia.

Ken Bull., pp. 87-97.
Sibley. F. C. & Clapp. R. B. (1967). Distribution and dispersal

of central Pacific lesser frigatebirds, Fregataariel. Ibis 109.
528-337.
Smith. G. J. & Volodkovich. V. 1.. ( 1988). Preliminary report

on marine plastics: Methods of collection and quantification.
Unpublished manuscript. 31 October 19SS. in Soviet-

American Joint Expedition III {1988), Appendix

pp.

Smithe. F. B. (1975). Naturalist's Color Guide. American
Museum of Natural History. New York. New York.

Spicer, R. A. & Newbery. D. McC. (1979). The terrestrial
vegetation of an Indian Ocean coral island: Wilingili, Addu
Atoll, Maldive Islands. [.Transect Analysis of the Vegetation.
Atoll Res. Bull. 231. 1-21.

Stevens, H. N. & Barwick, (i. F. I 1930). New light on the
discover) of Australia as revealed In the journal of captain
Don Diego de Prado y Tovai. Hakluyt Society. Series II.
No. 64. The Hakluyt Society, London.

Stoddart.D. R. ( 1976). Scientific Importance andConservation

of Central Pacific Islands. Unpublished manuscript. 28 pp.
Stoddart. D. R. & Fosberg. F. R. (1972). Reef islands of

Rarotonga with list of vascular plants. Atoll Res. Bull. 160.

1-22.
Stoddart. D. R. & Gibbs. P. E. (1975). Almost-atoll of

Aitutaki: Reef studies in the Cook Islands, South Pacific.

Atoll Res. Bull. 190, 1-158.
Stoddart. D. R. & Sachet. M. -H. (1969). Reconnaissance

geomorphology of Rangiroa Atoll. Tuamotu Archipelago.

with list of vascular flora of Rangiroa. Atoll Res. Bull. 125.

1-44.
Stone, B. C. ( 1968). Notes on Pandanus in the Line Islands.

Micronesica 4( 1 ). 85-93.
Stone, E. L. ( 1953). Summary of information on atoll soils.

Atoll Res. Bull. 23. 1-6.
Stonehouse. B. ( 1962). The tropic birds (genus Phaethori) of

Ascension Island. Ibis 103b. 124-161.
Taylor. R. C. (1973). An atlas of Pacific Islands rainfall.

Hawaii Inst. Geophys., Data Report 25.
Teeb' aki. K. ( 1 988 ). Report of visit to Caroline Atoll. Southern

Line Islands. Wildlife Conservation Unit, Christmas Island.

Republic of Kiribati. Unpublished manuscript. 12 pp.
Thaman,R.R.(1987). Plantsof Kiribati: A listing and analysis

of vernacular names. Atoll Res. Bull. 296. 42 pp.
Thomas, W. L. (1961 ). The variety of physical environments

among Pacific Islands. In Man's Place in the Island

Ecosystem (F. R. Fosberg. ed.), 3rd Edition, 1970. 264 pp.

Bishop Museum Press, 1965.
Tracey, J. I.. Jr.. Cloud. P. E. & Emery. K. O. (1955).

Conspicuous features of organic reefs. Atoll Res. Bull. 46.

National Research Council. Washington. D.C.
Trelease. W. (1884). Botany of Caroline Island. In Report of

the Eclipse Expedition to Caroline Island May ISS3. Natl.

Acad. Sci. Mem. 2. 87-90.
Tsyban. A. V. & Smith. G. J. (1988). Protocol of the Third

Soviet-American Joint Expedition. Pan II. The Central

Pacific Ocean and South China Sea. aboard the R/V Akademik

Korolev, 9 September - 31 October. 1988. Unpublished

report. 6 pp.
Upton. W. ( 1 884). Meteorology of Caroline Island. In Report

of the Eclipse Expedition to Caroline Island May 1883,

pp. 41-86.
US Dept. of Energy. (1987). The Natural History ofEnewetak

Atoll. Vols. I and II. 228 and 348 pp.. respectively. Office of

Scientific and Technical Information. US Deartment of

Energy.
Wagner. W. L. Herbst. D. R. & Sohmer. S. H. ( 1990). Manual

of the Flowering Plants qfHawai 'i. Vols. I and II. University

of Hawaii Press and Bishop Museum Press. 1. 853 pp.
Wester. L. (1985). Checklist of the vascular plants of the

northern Line Islands. Atoll Res. Bull. 287. 38 pp.
Whitehead. I). R. & Jones. C. E. (1969). Small islands and the

equilibrium theory of insular biogeographv. Evolution 23.
1 7 1 - 1 79. Lancaster. Pennsylvania.
Wiens. H.J. (1956). The Geography of Kapingamarangi Atoll
in the Eastern Carolines. Atoll Res. Bull. 48.

174

Wiens, H. J. (1962). Atoll Environment and Ecology. Yale Young, C. A. ( 1884). Introduction. In Report of the Eclipse

University Press, New Haven and London. 532 pp. Expedition to Caroline Island May Iti83. Natl. Acad. Sri.

Wilkes. C. USN (1845). Narrative of the United States Mem. 2. 11-20.

Exploring Expedition during the years 1838, 1839, 1840, Young. J. L. (ca. 1922). Notes on Flint, Caroline and Vostock

1841. 1842. Phila, Lea & Blanchard. 5 Vols, and an atlas. Islands. In Memoranda re Tahitian Business. S.R.Maxwell

Yaldwyn, J. C. & Wodzicki. K. (1979). Systematics and & Co.. Ltd., Pacific Manuscripts Bureau, Research School of

ecology of the land crabs (Decapoda: Coenobitidae, Pacific Studies. Australian National University. Canberra.

GrapsidaeandGecarcinidaelof the Tokelau Islands, central A. C. T. PMB Manuscript 23. 1-18.
Pacific. Atoll Res. Bull. 235. 1-53.

175

Chapter 2:

INVESTIGATIONS AND
ANTHROPOGENIC ECOLOGY

Editors:

SERGEI M. CHERNYAK,

CLIFFORD P. RICE, &

GREGORY J. SMITH

2.1 Distribution of Chlorinated Hydrocarbons
in Ecosystems of the Equatorial Pacific

SERGEI M. CHERNYAK and VALERIYA M. VRONSKAYA

Institute of Global Climate and Ecology, State Committee for Hydrometeorology and Academy of Sciences, Moscow, USSR

Introduction

It is a well-known fact that chlorinated hydrocarbon
pesticides (CHP's), which earned their developers a Nobel
Prize, have been both a great benefit to mankind by making it
possible to rescue up to one-half of the world's cereal-grain
harvests from pests and a detriment to the environment by their
trait of extreme stability causing them to build up in the
environment. In just a few years, their environmental hazard
has become apparent.

The multiplicity of pathways and the high rates of CHP
transport led to a situation whereby in the early 1 970' s they had
accumulated on a worldwide scale, being detectable in virtually
all environments, including mountain-peak glaciers and deep-
ocean depressions. Later, the world was to see a no less
extensive buildup of other classes of chlorinated hydrocarbons
(CH's). namely polychlorinatedbiphenyls(PCB's), chlorinated
terpenes, dioxins, benzofurans. et cetera, which were being
used not only in agriculture but also in various industries and
even in health care fields.

Despite the fact that most industrialized countries had
invoked total or partial bans upon the use of CH's in open-cycle
processes, their production on a global scale remains almost
undiminished, since no effective alternative for protecting
agricultural harvests in tropical countries has yet been devised.

There is now a considerable body of scientific literature
indicating that the World Ocean is the ultimate repository for
CH's. According to the most reliable estimates (Tanabe.
1 985 ), ocean waters are today the repository of as much as 70%
of all the CH's ever released into the environment. Previous
studies (Chemyak et al., 1985b) showed that the rates of CH
buildup and migration in constituents of the oceanic environment
depend first and foremost on the physicochemical and
hydrologic-geographic characteristics of various portions of
the World Ocean. Hence, predictive estimates of increases in
ocean pollution, which are bound to occur globally, require
information on the forms and amounts of CH's present in
various media and especially in little-studied regions such as
the Pacific Ocean, where only a few expeditions have concerned
themselves with CH's (Izrael & Tsyban, 1989). Our own
recent studies (Chernyak&Mikhaleva. 1985a) were the first to
include nearly in situ studies of microbial and photochemical
PCB decomposition processes, which are the only processes
presently at work to rid the oceanic environment of these
xenobiotic substances. We were particularly interested in
investigating PCB distribution at Caroline Atoll, a coral island
remarkable for its diverse flora and fauna, where these CH's
have accumulated, in markedly altered form and at a distance

of many thousands of kilometers from their sources. Their
presence at these locations are probably occurring through
atmospheric and oceanic transport.

The present paper sets out the findings and conclusions of
comprehensive studies conducted in 1988 in the equatorial
Pacific and at Caroline Atoll, which were obtained during the
First Joint US-USSR Central Pacific Expedition aboard the
R/V Akademik Korolev.

Materials and Methods

The locations of the sampling stations are shown in Fig. 1 .
The media sampled in order to establish the specifics of
hydrochemical processes involving chlorinated hydrocarbons
in background regions of the Pacific and coral-reef ecosystems
were seawater, sediment, plankton, neuston, demersal
organisms, and fishes. Specific studies conducted at Caroline
Atoll also included corals, flora and fauna, and eggs of local
bird species.

100°

110° 120° 130° 140°

150° 160° 170° 180°

170° 160° 150°

' HJ

40

y4&

Pacific

30'

J<$

Ocean

Hawaiian
Islands ^^

20'

\ South A
S China V>t

/Sea ^'jj&iS'^S8™-*-

V

10

%.-'/j

\

Christmas

*Y /£* *> r-,

"""■•«,,

Island v

0°

Vjfoic=£2C-\

<b

"V__ G a ^^^\_

ip <a_

^^s3*? Q ^r

Caroline
Aloll

10

\ ^

20

Fig. 1 . Expedition route and station locations in the Pacific Ocean and South
China Sea (9 September to 3! October 1988).

Seawater samples, 100 I in volume, were filtered through
XAD-2 resin at the rate of 20 1/h. The sorbed CH's were eluted
with 80 ml of ethanol mixed with an equal volume of 2%
sodium sulfate solution. The water-alcohol solution was
doubly extracted with n-hexane ( using 25 ml of hexane in each
of the two extractions). The extracts were concentrated to a
volume of 4-5 ml using a rotary evaporator, purified with
concentrated sulfuric acid, neutralized with a 5% NaHCO,
solution, rinsed twice, dried over sodium sulfate, and
concentrated in a stream of pure nitrogen to a volume of 1 ml.
The concentrate was injected into a Hewlett-Packard gas
chromatograph with the aid of an autosampler. The
chromatographic analysis conditions were as follows:

179

fused-quartz capillary column, length 30 m, and inside diameter
.•I 0.32 mm; chromatographic phase DB-1 (0.25 urn). The
temperature program lor the column was initial temperature.
1 20 C" ( I mini, rising to 250 C at the rate of 5°C/min. The
chromatographic analysis time was 40 min. the injector
temperature. 225°C, and the electron-capture detector
temperature. 300°C.

The sediment samples were centrifuged for 20 min at a
speed of 2,000 rpm to achieve total deposition of the silt. The
residue was extracted with acetone, then doubly extracted with
a 3:1 hexane-acetone mixture. The combined extract was
washed by mixing with an equal volume of 29? sodium sulfate
solution. The hexane layer was separated off, and the aqueous-
acetone solution reextracted with additional hexane. The
combined hexane extract was concentrated, then purified, first
with sulfuric acid (to remove extracted organic compounds),
then with tetrabutylammonium sulfate (to remove any sulfur
compounds). The purified solution was concentrated down to
I ml in a stream of pure nitrogen and chromatographically
analyzed. The biological samples were crushed to obtain a
homogeneous mass, defatted with acetone, then treated,
following the procedure used, with sediment samples.

In order to investigate the effect of photochemical processes
on the behavior of CH's in background ecosystems of the
Pacific, experiments were conducted on the decomposition of
a standard Aroclor 1232 solution in waters drawn from the
equatorial Pacific and around Caroline Atoll under the action
of sunlight.

The experiment was run in two 5-1 reactor vessels, one
exposed to sunlight, the other shielded using light-blocking
foil. The surface area involved was 400 cm2. The sterilized
seawater in the reactors were spiked with an acetone solution
of Aroclor 1232 to yield PCB concentrations in the water of
100 ng/1. The samples were extracted with n-hexane (twice.
50 ml each time), then concentrated to a volume of 2 ml in a
rotary evaporator. They were then purified with concentrated
sulfuric acid and chromatographically analyzed. Microbial
degradation of the PCB under the same conditions was
investigated for control purposes.

Results and Discussion

Data on CH levels in Pacific Ocean waters are presented
in Tables I and 2.

Analysis of these results clearly demonstrated the
dependence of the distribution of various PCB components on
their molecular structure.

The distribution of hcxachlorocvclohexanc (HCII) is
noteworthy since its total concentration was fairly high, though
it was still several times lower than in the Bering and Chukchi
Seas (Chernyak ci al., 1002). which arc just as far removed
from areas where this compound continues to be used. The
composition of the IICH mixture (containing as much as 90' -
of the (/.-isomer) indicates thai the sources of pollution are
probably equatorial countries that employ \ast amounts of
technical-grade hexachloran on their crops. The relatively low

TABLE 1

Chlorinated hydrocarbon levels in the surface waters
of the Pacific (ng/1).

Chlorinated Hydrocarbon Levels
Station Total
Number HCH DDT DDD DDE PCB

Caroline Atol

114

0.02

0.1

0.1

(1.1

(1.05

Caroline Atol

. lagoon

L-l

0.01

1.0

0.5

0.3

0.02

L-2

0.01

0.8

0.3

0.3

0.01

L-3

0.01

1.(1

0.3

0.5

0.02

L-4

(1.(11

1.1

0.3

0.3

002

Phoenix transect

1 15

2.7

0.07

0.02

0.01

0.2

1 Id

2.4

0.05

0.01

0.01

0.2

1 17

2.9

0.07

0.01

0.01

0.3

1 IS

2 "i

O.OS

0.02

0.02

0.2

1 19

2.3

0.02

0.01

0.01

0.4

12(1

2.5

0.05

0.02

0.01

0.3

Marianas transect

121

3.0

0.09

0.02

0.01

0.5

122

2.3

0.12

0.02

0.03

0.3

123

2.4

0.05

0.01

0.01

0.1

124

2.4

0.02

0.01

0.02

0.3

125

2.5

0.01

0.01

0.01

0.4

126

2 2

O.OS

0.02

0.01

0.4

South China 5

ea

127

4.2

0.18

0.04

0.02

0.3

I2S

4.5

OKI

0.02

0.03

0.4

124

4.8

0.10

0.04

0.05

0.3

130

3.7

0.22

0.10

0.03

0.3

131

3.3

0.IS

DOS

(Ids

0.3

TABLE 2

Distribution of chlorinated hydrocarbons over the water column in
Pacific Ocean waters.

Chlorinated

Depth

llv drocarbon Leve

s

inn

Total

HCH

DDT

PCB

Slain

m 1 1 5

0

2.7

0.10

0.2

III

2.5

O.OS

0.2

100

2.0

0.05

0.2

1.000

1.7

001

0.2

Slain

in 120

II

2.5

oos

0.3

10

2.3

(KIS

0.3

100

2.1

0.04

0.2

1.000

IS

0.04

0.2

ISO

hexachloran levels in the water here are likely due to their high
volatility, which appears to be driving most of the HCH present
in these hot climates into the atmospheric compartment.

The rather unusual distribution of HCH within the water
column — that is, the marked drop in its level that occurs at
considerable depths — is explained by the relatively high
solubility of this pollutant in seawater and by the fact that it is
present largely in dissolved form (in contrast to other globally-
occurring CH's. which are largely sorbed and precipitate
together with suspended matter). Hexachlorocyclohexane
isomers are practically the only common pollutants whose
behaviorin the open ocean can be explained largely on the basis
of hydrochemical factors.

Also worthy of note is the fact that there was almost a
twofold lower concentration level of HCH in the waters of the
atoll lagoon versus the ocean water surrounding the lagoon.
The only obvious explanation was the presence of a temperature
gradient, which entailed differing rates of photochemical and
microbial transformation of the cyclohexane ring, and the
accelerated evaporation that was taking place in the lagoon.

Another interesting finding was that the lagoon water of
Caroline Atoll contained roughly one-half as much PCB as did
the surrounding ocean water. Moreover, the composition of
the CH's was significantly different: the lagoon water contained
virtually no highly chlorinated PCB congeners, which was
probably due to the higher rates of photochemical processes in
the thoroughly heated shallow water of the lagoon. It should be
noted, however, that as a general rule, PCB levels in the
equatorial Pacific were only slightly lower than in the northern
Pacific, even though sample composition turned out to be
considerably different.

Whereas most of the PCB' s in the Bering Sea consisted of
di- and trichlorobiphenyls, the major constituents of PCB's
in the equatorial waters were tri- and tetrachlorobiphenyls
as well as heptachlorobiphenyls. Analysis of the CH's sorbed
by suspended matter revealed a clear dependence of pollutant
levels on latitude. For example, the content of HCH isomers in
suspensions from the equatorial Pacific was almost 50 times
lower than in the circumpolar parts of the ocean (Table 3 ). This
may have been due to the significant shift in sorption-process
equilibria associated with a 25°C rise in temperature. There
was a marked (almost tenfold) change in PCB levels in
suspended matter, whereas the levels in the water layer remained
virtually constant. It is a curious fact that the equatorial Pacific
is a unique region of the World Ocean where the PCB mixture
appears to be equally apportioned between the suspended
matter and the dissolved phase.

Of special interest is the distribution of the extensively
used pesticide DDT in Pacific Ocean ecosystems. Dichloro-
dipheny ltrichloroethane levels in water samples from the Bering
and Chukchi Seas have decreased considerably over the past
decade due to restrictions on the use of this compound imposed
by a number of industrialized countries. In fact, in some
instances these levels come close to analytical zero (Chernyak
et al.. 1989). In the equatorial Pacific, however, the DDT

TABLE 3

Chlorinated hydrocarbon levels (Ug/g dry weight) in suspended
matter in the Pacific Ocean.

Station

Date

Depth

Chi

urinated

Hydrocai

•bon

No.

(m)

a-HCH

Y-HCH

DDT

PCB

113

09/2 1

0

54

43

135

258

Caroline

09/23-29

0

5

5

174

31 1

Atoll

116

10/02

20

5

4

164

101

116

10/02

100

5

5

105

126

119

10/04

65

5

4

192

168

119

10/04

100

3

3

127

171

119

10/11

0

5

3

167

315

120

10/11

0

5

5

103

153

120

10/11

0

4

4

143

93

120

10/12

0

3

4

215

139

120

10/13

0

5

4

287

189

120

10/14

0

5

4

274

63

120

10/15

0

3

3

159

76

121

10/16

0

5

3

315

154

121

10/17

0

4

3

252

127

122

10/18

0

4

4

338

129

122

10/18

130

3

3

309

85

123

0

5

3

421

1X1

124

0

5

4

357

108

hazard remains considerable: its levels average 0. 1 ng/1, which
is typical of the areas most severely impacted by human
activities, namely the North Atlantic and the Indian Ocean
(Tanabe etal.. 1982).

Results for the microbial and photochemical degradation
of PCB's in the equatorial Pacific are presented in Figs. 2-4.
The gradual loss of several of the PCB congeners are plotted
over time in Fig. 2. On incubation with natural populations of
microbes from these central Pacific waters, some congeners
were reduced by more than 50% in 10 days (e.g., BZ#s.
Ballschmitter and Zell numbering system for PCB congeners:
Ballschmitter&Zell, 1980)7, 16. 49, 52, and 42. In Fig. 3, the
degree of microbial degradation of the PCB's is organized by
the PCB homologue group. From this presentation it is
observed that the greater the degree of chlorine substitution of
the bipheny 1 ring, the more resistant it is to microbial breakdown.
Plotted for comparison are the relative rates of breakdown of
the homologous groups for the two regions that were studied,
the Bering Sea. and the central Pacific Ocean. Between these
two locations, the rates of degradation of especially the mono-
and dichlorobiphenyl homologues are faster in the warmer
water of the Pacific. Figure 4 presents the times loss of total
PCB's for the central Pacific based on a comparison between
microbial degradation, photochemical degradation, and
photochemical processes under the influence of added PAH's.
It is apparent here that microbial degradation accounts for most
of the breakdown in this area and that PAH's may have the
capacity to inhibit photochemical breakdown of PCB's.

Percentage ot initial PCB concentration

1 2 3 4 5 6 7

Fig.2 Photochemical breakdown of individual PCB component1.
A

Fig. 4. Breakdown of PCB's (dichlorobiphenyls) through microbial and
photochemical action: 1) microbial; 2) photochemical: 3) photo-
chemical in the presence of PAH's.

Conclusions

As a result of human activities, chlorinated hydrocarbons
occur in all media constituting the Pacific ecosystems
investigated. Levels of HCH isomers in the equatorial Pacific
turned out to be lower than in polar areas. This is understandable,
given the high volatility of this pollutant.
Dichlorodiphenyltrichloroethane levels in certain parts of the
Pacific (such as the Caroline Atoll) were found to be similar to
those observed in impacted basins, although the concentration
of this compound in the equatorial Pacific taken as a whole
were not high. Polychlorinated biphenyls occurred in all the
water samples studied. It is now clear that despite the slowness
of photochemical and microbial breakdown of PCBs, these
processes nevertheless play an important role in the self-
purification of Pacific Ocean ecosystems.

Fig. 3. Microbial breakdou n of PCB components in the waters of the Bering
Set (A) and central Pacific Ocean (B).

2.2 Distribution of Polycyclic Aromatic
Hydrocarbons

EHA R. URBAS. NATALYA I. IRHA, and UVE E. KIRSO

Chemistry Institute of the Estonian Academy of Sciences, Tallinn, ESSR

Introduction

Three distinct regions (Caroline Atoll, tropical Pacific and
South China Sea) during the First Joint US-USSR Central
Pacific ( BERPAC) Expedition were sampled and analyzed for
polycyclic aromatic hydrocarbons (PAH's). Methods for
collection and analyses are described in Distribution of PAH's
( Irha et al. , 1 992) in Results of the Third Joint US-USSR Bering
& Chukchi Seas Expedition (BERPAC), Summer 1988 (Nagel,
1992). The regions that were sampled are identified in the
Frontispiece to this volume.

Polychlorinated aromatic hydrocarbons are important
natural and anthropogenic contaminants of marine ecosystems.
These compounds are associated with petroleum pollution,
including natural oil seepage, and with industrial contamination
and many exhibit carcinogenic and mutagenic properties. In
this paper, the presence and quantity of PAH's in both biotic
and abiotic components of tropical Pacific ecosystems is
reported.

Results

Caroline Atoll

Determinations were made of the composition and
distribution of PAH's (Table 1 ) present in water and marine-
organism suspensions sampled in inshore waters off the atoll,
as well as in the following media: bottom sand, island soils,
corals, and siphonales algae. The findings (Table 2) indicated
that the PAH pollution of this ecosystem was negligible. There
were no PAH's in samples of surface sand from the atoll, and
the soils taken from a palm tree forest area showed PAH levels
below background.

The total PAH content of coral samples did not exceed
9.6 |ag/kg of dry weight. The following representatives of
four- and five-ring PAH's were detected: benzo(e)pyrene
(BeP), 62.5; indeno( l,2,3-cd)pyrene (IPy), 23; benzo(a)pyrene
(BaP). 14.5 (in percent by weight).

The benthic sand of the atoll lagoons contained four- and
five-ring PAH's (Table 2); the major constituents (in percent
by weight of the total amount of PAH's present) were
benzo(b)fluoranthene (BbF), from 38.7 to 53.6%; and BaP,
from 16.8 to 36.6% (Fig. 1).

The surface waters off the atoll contained six four- and
five-ring PAH's, whose total amount did not exceed
1.55 ng/1, the dominant constituents being the five-ring
carcinogens BbF and BaP (Fig. 2). Suspended matter sampled

Table 1

List of specific PAH's identified by displacemenl-eludonal liquid

chromatog

raphy

Name

Symbol

Structural formula

Carcinogenicity
(Lee e/u/., 1981)

Pyrene

Py

£9

0

Chrysene

Chr

oS°

+

Benz( a (anthracene

BaA

ocx9

+

Benzol etpyrene

BeP

&

0/+

Benzol blfluoranthene

BbF

^3

++

Benzol k )fluoranthenc
Benzo(a)pyrene

BkF
BaP

cab
cxA

++
++

Benzo(g.h,i)perylene

BPer

cfi?

+

Dibenzla.h (anthracene

DBA

€8

+

Indenol 1,2.3-cdlpyrene

IPy

go9

+

Note:
Classification

Svmbol

Criterion: % of animals

that de\clopcJ lesions

noncarcinogenic
weakly carcinogenic
strongly carcinogenic

0

+
++

0
33
>31

at a depth of 90 m included four kinds of four- and five-ring
PAH's, whose total concentration was 0.68 ng/1, with the four-
ring IPy and benz(a)anthracene (BaA) predominating (Fig. 2).

The total PAH level in the gills of Mexican (northern red,
Pensacola) snapper {Lutjanus campechanus) specimens was
moderate, amounting to 0.08 |ig/kg; the corresponding value
for muscle tissue was 1.13 Hg/kg of fresh weight.

The following five-ring PAH's were identified as present
in the gills (in percent by weight): BbF. 62.5; BaP, 33.5;
benzo(k)fluoranfhene (BkF), 3.7; moreover, chrysene (Chr)
and benzophenanthrene (Bp) were also present. The main
constituents of the total PAH mix were four-ring compounds:
BaA, 88.5%, and Chr, 1 1.5%; also present were traces of BbF
and BkF.

Polycylic aromatic hydrocarbon levels in various
components of the Caroline Atoll ecosystems indicated that
this area was only slightly polluted by these specific compounds.

183

TABLE 2

PAH le\cls in particular media of the Caroline Atoll ecosystem
(September 1988).

Medium Units Py BaA Bnf BkF BaP Bep

Cm-

Water, surface ng/1 qualitatively tr.

>5 0.90 0.21 ir.

layer

present

Suspension

ng/1

0.6

0.8

tr.

tr.

-

(90m)

Benthic sand.

u/kg

- ..

-

-

-

-

lagoon

Lagoon I

-

-

-

-

-

Lagoon III

-

tr. 0.44

0.08

0.30

-

1 agoon V

-

0.60 0.60

0.09

0.26

-

Atoll soil

u/kg

-

-

-

0.83

-

Aeropora corals

u/kg

2 2

-

-

1.4

6.6

Siphonales algae

u/kg

-

-

-

0.07

-

(Caulerpaceae)

ir. - trace

06

0 2

(b)

n

n

BaA BbF BkF BaP

BaA BbF BkF BaP

Fig. I PAH levels lu/kgi in henlhic sand of lagoons V (a) and 111 (hi "I
Caroline Atoll (September 1988)
|1 egend:] A = PAH level, u/kg.

1.0

u 0.5
X

<

a.

0 1
0

(a)

(b)

n

n

Py BaA BbF BkF BaP BeP

Py BaA BkF BaP

i l ontentol PAH's (ng/1) and their distribution (in percent bj weight)

in the surface (0 0.5 m) layer ol watei (a) and inasuspension(b)drawn
from a depth <>i 90 m ofl Caroline \toll (September 1988)
[Legend:! A = I>AII level, ng/1

Tropical Pacific

Polycyclic aromaitc hydrocarbon levels in the surface
w aters were not high. In the open ocean, the total concentrations
ranged from 0.8 to 43.8 ng/1, with a maximum of 396.6 ng/1 in
the Marianas sector. The composition and distribution of the
PAH's identified in the surface waters of the open ocean
indicated that the main compounds present were the following
five-ring carcinogens (in percent by weight): BbF. 86.3: BkF,
1 1 .0; BaA and Chr. 1 .2; BeP. 1 .0; BaP. 0.5: and IPy was present
qualitatively.

The total PAH's present in the waters of the Marianas
sector were 1 0 to 1 00 times higher in concentration than in the
open ocean. The dominant compounds were four- and five-
ring PAH's ( in percent by weight), namely BaA and Chr. 38.7;
BeP, 35.2; andpyrene(Py), 22.7. TheBbFandBaPconstnuted
3.2% and 0.2% by weight, respectively.

In areas where PAH levels in the water were below the
sensitivity threshold of the liquid chromatography technique
employed, a more sensitive method for just BaP was employed.
Benzol a (pyrene is useful as a representative tracer of other
PAH's (analytical methods described in Irhae/o/.. 1992). The
following BaP levels were noted: up to 0.58 ng/1 in the open
ocean and up to 1.36 ng/1 in the Marianas sector.

Unlike the surface layer of water, suspended matter sampled
in the tropical Pacific was found to contain as many as 10 lour-
to six-ring PAH's. All samples taken at depths of 0, 20. 58, 80.
1 30. and 200 m contained the PAH's IPy, BaA, Chr, BbF. BkF,
and BaP. the dominant ones being IPy and BaA. which accounted
for 30-99% and 30-50%. respectively, of the total amount
present. In certain instances, suspensions from \ arums depths
also contained IPy and benzo(g,h,i)perylene ( BPen.

Bottom sediments of the tropical Pacific were found to
contain only three kinds of four- and five-ring PAH's. namely
BaA, BbF. and BaP. with the total PAH content not exceeding
4.5 Ug/kg. The dominant PAH (by weight) was BbF, which
accounted for 93% of the total amount of PAH's present.
Although the composition of PAH' s in the bottom sediments of
the Marianas sector of the Pacific was similar. IPy and BbF
were the dominant PAH's (50.7% and 47.3' - . respectively, of
the total weight of PAH's present).

South China Sea

Polycyclic aromatic hydrocarbon levels in the waters of
the South China Sea were much higher than in the open ocean
and in the Marianas sector (Table 3). The total amount of
PAH's in the water \ aried w idel) : the PAH mixture identified
in the surface microlayer consisted largely of four- and five-
ring PAH's (in percent by weight): BeP. 45.5: BaA and Chr.
28.7: Py . 24.6; BbF, 0.9; and BaP. 0.2. Their weight ratios and
composition in the surface layer were somewhat different
(Table 3 ). moving tow aids increased fractions of BaA and Chr.
and BeP. Considerably higher levels of PAH's. by weight.
were delected in the benthic portion of the water column, the
dominant I> All being BbF (Table 3).

In those sea areas where the PAH level fell below the
analytical detection limit, the BaP level did not exceed
1.58 ng/1.

184

TABLE 3

PAH levels in particular media of the South China Sea ecosystems (Station 127. September 1988).

Medium

Units

Pv

BaA BeP BbF BkF BaP IPy BPer DBA

Water, surface ng/1 qualitativel)

microlayer present

170 67

4.5

Water, surface ng/1

layer (0 to 0.5 m)

150

175

278

5.5

1.6

Benthic layer ng/1 qualitatively 290

present

290

14,000

300

Benthic sediment u/kg

0.46 0.14 0.15

Neusum

u/kg

1.4

1.7

.3 0.6 0.1

Suspension.
58 m

ns/1

4.7

0.6

0.9 0.06 0.01 0.07 tr. 0.07 tr.

Suspension,
80 m

ng/1

0.6

0.4

0.01

0.04

tr.

0.06

Hydroids

trace

u/kg

6.5

0.6

0.7

The total PAH level in the suspended matter drawn from
depths of 58 and 80 m were 6.4 and 3.9 ng/1. respectively
(Table 3). The PAH mix was found to contain ten four- to
six-ring PAH' s in the following relative amounts (in percent by
weight): IPy. 65.2; BeP. 17.4; BaA and Chr. 1 1.0; BeP. 1.0;
BPer. 1.0: BbF. 0.9; BkF.0.2; and IPy anddibenz(a.h (anthracene
(DBA) were present in trace amounts.

The sediment samples contained three five-ring
carcinogenic PAH's (Table 3), namely BbF. BkF. and BaP.
The total PAH content of the sediment did not exceed
0.75 (ig/kg of dry weight, the dominant constituent being BbF
(61.3% of the total content by weight).

The PAH mix detected in the marine biota indicated that
PAH's were fairly abundant. The total PAH content of the
neuston of the South China Sea was 6.6 ug/kg (Table 3). The
mix consisted of six PAH's. Most of the total amount was made
up of four- and five-ring compounds (i.e.. BaA. BeP. BbF. and
BaP).

It should be noted that the PAH mix detected in the marine
hydroids was similar to that found in the sediment (Table 3 ) and
consisted of strongly carcinogenic five-ring compounds, namely
BbF. BaP. and BkF. Their total content in the hydroids was
10 times that in the sediment, and the weight ratios reflected a

larger fraction of BbF. 83.3%. The relative content of BaP and
BkF was 8.9% and 7.6%, respectively. The same carcinogenic
PAH's were detected in benthic organisms (Petrosia sponges.
Table 4).

TABLE 4

Comparative PAH levels in sponge tissues (u/kg of dry weight)

and fish livers (u/kg of fresh weight) from the South China Sea ( 1 )

and Bering Sea (2).

PAH

1

2

Sponge SI

lark liver*

n = 2

Spongt

: Liver of
walleye pollack

Chr

_

6.15

_

34.00

BbF

0.70

0.27

0.90

4.50

BkF

0.09

0.22

0.16

tr.

BaP

0.40

1.00

0.60

13.00

BaA

qualitatively
present

1.00

-

qualitatively
present

BPer

2.80

Specimens caught in the tropical Pacific.

185

It may be noted that the total content and mix of PAH's
present in the tissues of South China Sea sponges was not much
different from those established for the same aquatic organisms
in the Bering Sea (Table 4). The predominant PAH's
(in percent by weight) for South China Sea and Bering Sea
sponges were BbF (54 and 59) and BaP (36 and 33.6),
respectively.

Concentrations of PAH's in livers of sharks ranged from
1.2 to 10.2 (ig/kg of fresh weight (Table 4). The predominant
compounds were the following four- and five-ring PAH's

(in percent by weight): Chr, 53.7; BPer, 24.7; BaA. 8.7; BaP,
8.7; BbF, 2.3; and BkF, 1.9.

For comparison we note that the levels and distribution of
PAH's in aquatic organisms of the Bering Sea (sponges and
walleye-pollack liver) also indicate relatively high total PAH's
in fish livers and the buildup of the strong carcinogens BbF,
BaP, and BkF as the dominant PAH's in sponges (Table 4).

Analysis of the findings of the present study points to
severe petroleum hydrocarbon pollution of the sea and to the
buildup of the carcinogenic PAH's — BbF and BaP — in some
aquatic organisms.

2.3 The Occurrence and Microbial

Transformation of Benzo(a)pyrene in the
Waters of the Tropical Pacific (Caroline
Atoll, Line Islands, Phoenix Islands
Transect, South China Sea)

YURIY L. VOLODKOVICH and OLGA L. BELYAEVA

Institute of Global Climate and Ecology, State Committee for Hydrometeorology and Academy of Sciences, Moscow; USSR

Introduction

Among the numerous organic pollutants working to
produce dangerous and undesirable changes in the chemical
and biological status of the marine environment, polycyclic
aromatic hydrocarbons (PAH's) of both natural and human
origin must be singled out as being particularly noxious.
Exhibiting considerable molecular stability, as well as
pronounced carcinogenic and mutagenic properties, compounds
of this series are a serious hazard to marine life. At present,
researchers have been choosing ben7.o(a)pyrene (BaP) as an
indicator of environmental pollution by PAH's.

Benzo(a)pyrene and other PAH's have been found to
occur in many marine ecosystem components from the arctic
latitudes (Volodkovich & Belyaeva, 1987) all the way to the
Antarctic, affecting mariculture and other activities.

The proximity of human sources of pollution has a powerful
impact on the marine environment, increasing BaP levels in
waters by factors ranging from tens to hundreds (Tsyban et ai,
1985b; Tsyban ef a/., 1986).

For example. BaP levels in some parts of Los Angeles
harbor have reached, and even exceeded. IIS ng/1. while
concentrations in the top 0.5 m of the water column in the
equatorial Pacific have been reported in the 1-6 ng/1 range
(Shilina, 1982).

In view of the fact that elevated PAH levels in seawater are
due largely to the proximity of a given sea area to pollution
sources, the study of BaP circulation and elimination processes
(including microbial transformation) in as yet unimpacted
open sea and open ocean areas is of considerable interest and
value.

Microorganisms, which are virtually ubiquitous in the
World Ocean, play a major role in the functioning of marine
ecosystems and in the biochemical cycling of various
compounds, including pollutants such as petroleum crude and
PAH's (Izrael & Tsyban, 1989). The processes involved in the
microbial transformation of aromatic hydrocarbons and
heterocyclic compounds have been investigated in sufficient
detail. On the other hand, the rates of microbial degradation of
BaP in the marine environment, as well as the local and
regional scale of these processes, have thus far been neglected
(Tsyban et ai, 1986).

Integrated studies of biogeochemical cycling of PAH's.
using BaP as an example, were performed in 1988 during the
cruise of the R/V Akademik Korolev as an extension
of work begun in 1981 in the South China Sea and the
Marianas sector. The research represented the first time that
such work has been undertaken in the tropical Pacific (in the
Caroline Atoll area and the Line Islands/Phoenix
Islands transect).

186

Methods

Sampling

The studies were performed on both rainwaterand seawater
samples taken from a variety of depths ranging from the surface
down to the benthic portion of the water column. Also
investigated were natural bacterioplankton communities in
surface and near-surface layers.

Seawater samples were taken with Niskin bottles.
Rainwater samples were collected in vessels of surface area
0.5 m: mounted near the bow of the vessel. The BaP present in
the water was extracted from 1-1 samples by triple benzene
extraction. Final processing and analysis of 1 00 ml volumes of
benzene extract of BaP were carried out ashore upon completion
of the cruise.

Seawater samples for use in the microbiological
experiments were taken using sterile sampling equipment
(a 5- 1 plastic Niskin bottle, and a 5-1 glass bottle, in the case of
Caroline Atoll lagoon). The water samples containing natural
microorganism communities were decanted into glass bottles
under sterile conditions. Further processing was carried out in
the shipboard microbiology lab.

The In Situ Experiments to Assess the Biodegradation Potential
of Seawater Microflora in Relation to BaP

The process involved in the microbial transformation of
BaP was studied under the conditions similar to the in situ
experiments. These involved placing 250-ml samples
containing natural microflora communities in dark glass bottles
of volume 0.5 1. A weighed amount of BaP dissolved in a
minimum quantity of acetone (0.05 ml ) was added to the bottles
immediately after addition of the water ( the acetone evaporated
within a few minutes). The transformation process was
simulated in runs using BaP concentrations of 2 and 10|ig/l. In
order to allow for abiotic factors, each experiment included a
control in the form of a sample of sterile water drawn from the
same depth and containing the same concentration of BaP.
Each complete pair of tests (experiment plus control) was run
three times.

In order to make the conditions of the in situ experiments
as close as possible to natural, the series of glass vessels were
placed in baths with running water from the surrounding sea,
with the baths themselves mounted on the deck of the research
vessel. In the case of the Caroline Atoll, the sample-containing
vessels were placed in a plastic cassette placed on the bottom
of the lagoon (at a depth of 8 m). Exposure equaled 5-7 days,
depending on the seawater temperature. On completion of
exposure, the microbial activities were terminated by adding
several milliliters of concentrated hydrochloric acid to the
sample. The residual BaP in the experimental and control
samples were extracted using benzene. The final BaP
concentration values were determined ashore.

Chemical Analysis

Quantitative determinations of the residual BaP levels in
the seawater samples involved concentrating the resulting
benzene extracts down to a volume of 1 ml and then analyzing
them on alumina plates using thin layer chromatography with
heptane-benzene-acetone (100:60:6.7, by volume) as the
solvent system. The BaP-containing area of the adsorbent was
eluted with acetone, after which the BaP was transferred to
n-octane.

The benzene extracts of BaP that were obtained in the
microbial experiments were treated without chromatographic
thin-layer separation. The evaporated portion of the benzene
extracts were eluted with 2-ml solutions of n-octane containing
1 x 10 7 g/ml of 1,12-benzopyrelene, which was used as an
internal standard.

Quantitative determinations of BaP (in n-octane solution)
were made by a fluorescence-spectrum analysis method, relying
on the Shpolsky effect (Shpolsky et al., 1952; Fedoseeva &
Khesina. 1968). The analyses were performed with a DFS- 12
spectrograph at a temperature of -196°C using supplementary
standards (BaP and othercompounds). The minimum sensitivity
of the method for BaP was 1 x 10 '" g/ml, with error brackets
of 10%.

The simulated rates for microbiological degradation were
determined from the difference between the initial (introduced)
and the final (remaining) mass of BaP in the separate reactors.

Results and Discussion

The first series of studies concerned the equatorial Pacific
(Fig. 1 ) in the rectangle bounded by 7° and 0°S and 150° and
1 80°W. Analysis of the findings showed that BaP levels in the
waters of the Line Islands/Phoenix Islands transect were
remarkable for their extreme purity. Benzo(a)pyrene levels in
the waters of the transect down to depths of 2.000 m
corresponded to the minimal background value of 1 ng/1 and
lower. Certain samples contained no detectable amount of BaP
whatever (Table 1 ). Thus, the average BaP level for the entire
water column at Station 1 17 was just 0.06, 0.018 ng/1. In the
majority of cases, the BaP levels did not exceed 5 ng/1. As seen
in Fig. 2, zones with elevated BaP levels were of a local
character. High levels were recorded for depths of
1,000-1,500 m (6.5 and 1 1.3 ng/1, respectively), with only one
instance (10.3 ng/1. Station 1 18) of a high concentration in the
top 0.5 m of the water column. It must be emphasized that the
peak levels recorded were a whole order of magnitude lower
than BaP levels in open ocean and other portions of the Pacific
as found in earlier studies (Izrael&Tsyban, 1989). The limited
occurrence and low levels of BaP buildup in seawater along the
Line/Phoenix transect may be attributed not only to the
remoteness of this geographic region from the principal human
sources of pollution (industrialized areas and shipping routes)
but also to the absence of any pronounced PAH flows in this
ultraoligotrophic part of the ocean. Nor does economic activity
in the waters of the Kiribati Republic archipelago appear to be
having any significant deleterious effect on this marine
environment.

187

100' 110" 120 130 140'

150c 160° 170°

180: 170 160 150c

5 sr^

\ South A
S China \2>t
("V/Sea %*'
"Vs. .-jo ^>

Pacific
Ocean

Hawaiian

Islands t*^)
i

Christmas
-..•3 Island n

if>*

Caroline

a

0 Atoll

«0

Fig. 1 . Expedition route and station locations in the Pacific Ocean and South
China Sea (9 September to 31 October 1988).

If this minimal concentration of BaP serves as an indicator
representing a whole series of carcinogenic PAH"s, then the
equatorial zone in question can be readily classified as one of
the "background" portions of the World Ocean.

The studies at Caroline Atoll (10°N, 150°W) indicated
that BaP was present in individual water samples both from the
interior portion of the lagoon and from the inshore waters at

Stations
120 119 118 117 116 115

Depth, m

i i i

J L

0.5

10

45
100

500

Kldil

15(H)

2000

BP concentration, ng/1

□ -n-'

□

->1 -2

□ ->2-4

->4-6

->6-8

Fig. 2. Occurrence of benzol a Ipyrene in the waters of the Line/Phoenix
transect between 7°S and 0°S in the equatorial Pacific ( October 1 988 1.

TABLE 1

Benzol a )pyrene concentrations (ng/1) in the waters of the central Pacific (September and October 1988).

Region Station No.

0.5

Depth, m

10 25 45 100

500 1.000 1.500 2.000

Caroline Atoll
Lagoon 0.0

0.1 0.1

0.1

Inshore waters 1 14 0.2

0.1 0.1 0.0 0.1

Line/Phoenix

115

0.4

1.0

3.7

2.5

4.7

4.1

4.4

4.7

-

transect

116

2.5

1.4

0.6

1.3

1.5

4.1

6.5

0.9

0.9

117

0.1

0.3

0.3

0.7

0.2

0.6

1.1

1.2

0.9

118

10.3

0.0

1.1

2.1

5.8

0.6

1.6

5.9

—

119

2.4

3.7

2.2

1.4

7.2

1.9

11.3

6.5

0.5

120

2.0

0.6

0.0

0.3

1.0

2.9

2.7

2.8

0.6

Marianas

121

80.0

7.0

5.9

9.2

3.6

5.5

5.3

2.5

2.7

transect

122

104.0

96.0

94.0

101.0

86.0

6.6

9.7

111.0

96.0

123

57.5

84.0

94.0

62.0

84.0

56.0

39.3

6.7

5.0

124

38.2

96.6

68.4

5.0

14.0

12.5

6.6

14.5

0.8

125

2.7

I.I

3.5

11.3

1.6

2.3

2.1

0.9

10.8

126

12.2

12.1

2.3

1.9

3.0

9.0

14.3

1.0

6.2

South China

127

2.0

3.2

0.9

16.0

173.0*

Sea

127

1.1

2.3

0.9

0.6

105.4*

129

18.5

8.2

3.2

12.4

9.3

27.5**

131

64. X

17.1

1 1.0

6.0

10.0

148.8***

Sea-bottom depths:

* - 6 1 m

** - 200 m

*** - 250 m

INS

depths of up to 100 in. However, these also corresponded to
background values not exceeding 0.2 ng/1 and averaging
0.08 ng/l (Table 1 ). Particularly noteworthy was the absence
of BaP in the top 0.5 m of the water column inside the lagoon.

Without excluding the possibility of PAH biosynthesis in
the highly productive coral ecosystem of Caroline Atoll (such
as the phenomenon reported for the Clipperton Lagoons;
Niassat el <//. . 1 968 ). it may be assumed, with some confidence,
that the high rates of photochemical oxidation and microbial
transformation of PAH's characteristic of equatorial waters are
conducive to the elimination of PAH's from the waters of
Caroline Atoll and especially from the surface layer.

Quite the opposite situation was observed in the Marianas
trough region along the Marianas transect (between 142°E and
128 K. along 1 1°N) (Fig. 3). Whereas the BaP levels at the
extremities of this sector, with one exception, ranged from
0.9 to 9.2 ng/l (Table I ). the water column in the central portion
of the transect (Stations 122 and 123) contained elevated and
even maximum concentrations (up to and over 100 ng/l). As is
evident from Fig. 3. high BaP levels from 38 to 104 ng/l were
noted not only in the surface layer but also in deeper waters.
Thus, the BaP concentration in the upper 100-m water layer at
Station 122 averaged 96.2 ng/l. while the average value at
1 ,500-2.000 m was 103.5 ng/l. The lowest BaP concentrations
occurred at the western end of the transect, where the BaP

HI', nii.i - i

Fig. 3. Distribution of benzo(a)pyrene in the waters of the Marianas transect
along 1 1°N (between 142°Eand 1 38°E) in the tropical Pacific in 1981
and 1988.

levels at Stations 1 26 and 1 25 in waters up to 2,000 m deep were
4.03 and 6.88 ng/l. respectively. The average value for eastern
Station 121, at depths from 10 to 2.000 m. was 5.17 ng/l.

Analysis of rainwater samples taken at 1 1°N along the
transect had BaP levels of 7.2, 8.8, and 9.2 ng/l; that was one
order of magnitude lower than the levels in rainwater measured
along the transect in 1981 (Izrael&Tsyban, 1989). Nevertheless,
these findings point to the possible contribution of PAH's to
the Marianas sector of the Pacific via long-distance atmospheric
transport of organic pollutants.

High BaP levels (up to over 100 ng/l) were noted in the
waters of the Marianas sector during the previous study period
in 1981 (Izrael & Tsyban, 1989). These ranged from 80 to
1 20 ng/l. However, the pollution in all instances was confined
to the top 2 m of the water column with peaks in the surface
microlayer (SML).

Since most of the waters of the Marianas sector lie in the
area of influence of the northern branch of the tradewind
current, it may be supposed that the invasion of deeper layers
by pollutants (including BaP) occurs through downwelling of
surface water due to the anticyclonic circulation of water
masses in this region.

The BaP levels in the southwestern South China Sea had
a broad range of values from 0.9 to 173 ng/l (Table 1 ). The
lowest BaP levels occurred in open waters at Station 127
(Fig. 1). Two days of work at this station yielded similar
estimates, the BaP levels over most of the water column
varying from 0.6 to 2.0 ng/l and averaging 1 .5 ng/l. The bottom
layer (at a depth of 61 m) exhibited a very high BaP concentration
and with an average value of 1 39.2 ng/l. indicating severe PAH
pollution.

The BaP concentrations in the waters of stations lying
close to the eastern coast of Singapore (Stations 1 29 and 131)
hadelevated BaP concentrations in the lOto 100-m layerof the
water column (averaging 8.2 and 13.2 ng/l, respectively). The
maximum BaP lev els were detected in the top 0.5 m layeron the
one hand and at the sea bottom on the other. These were as high
as 148. 8 ng/l (Table 1).

This peculiar pattern of BaP level distribution over the
water column can be attributed to the influx of PAH's with the
oil pollution that impacts the surface waters at the stations in
question as well as the deeper waters below. It should be noted
that the areas of the South China Sea investigated are busy
maritime thoroughfares for vessels carrying both crude oil and
refinery products as well as sites of intense offshore drilling
activity.

The potential physiological ability of microflora to
transform BaP in the central Pacific was studied in a series of
in situ experiments using natural bacterioplankton communities
from Caroline Atoll and the southwestern portion of the South
China Sea.

The results of the process simulations expressed in terms
of the degree of elimination of the artificially-introduced
amount of BaP are shown in Fig. 4. The bacterioplankton of the
marine areas investigated were able to transform the
polyaromatic hydrocarbon in question. On the whole, the level
of microbial degradation reached 57-98% of the mass of BaP
introduced into the samples (Table 2). In order to gain deeper

189

1(H)
BaP tranformation
as percentage of
initial mass (%). 80

ontrol

Days

Fig.4.

Dynamics of BaP transformation in the in situ experiments using
microflore from waters of Caroline Atoll lagoon (a) and from the
South China Sea (b) at Station 127 (1988 data). (Initial BaP
concentration: A- 2 |ig/l; O- 10 ug/l;#- control)

insight into the natural processes of BaP elimination from the
marine environment, we also undertook a study of the dynamics
involved in microbial transformation. The results obtained
from individual runs of each experiment yielded reasonably
consistent values.

As may be seen from Fig. 4, the highest microbial
transformation rates were noted in experiments using microflora
taken from Caroline Atoll lagoon water. The series of samples
placed in natural conditions at a depth of 8 m contained just
50% of the initially introduced amount of BaP (initial
concentration 10 ug/1) after only 2 days; after 5 days had
elapsed, 97% of the initial BaP had been transformed. The rate
of microbial transformation of BaP in the lagoon waters was
high, with the curve of the degradation process close to linear.

The rate of microbial transformation of BaP in the South
China Sea was much lower than for the atoll. With an initial
BaP concentration of 2 ng/1, only one-third of its initial mass
had been transformed after 5 days. However, after 7 days of the
experiment, the microflora from the top 0.5 m of the water
column was able to transform as much as 77% of the introduced
BaP. With a higher concentration (10 ug/1). microbial
transformation in the experiment produced a consistent value,
34% of the initial mass over the third through the fifth days of
the experiment where a period of decreased microfloral activity
occurred (Fig. 4), the incremental transformation amounting to
just 0.6 ug of BaP. In these tests, the total mass of the BaP
degraded after 7 days; at the time which the experiment was
terminated, the mass did not exceed 5.68 ug (56% of the initial
concentration). On the whole, the rate of microbial
transformation of BaP turned out to be not particularly dependent
on the initial concentration (2 and 10 Ug/1). Assuming the
process curve to be more or less linear for both the equatorial
and the tropical Pacific, we were able to estimate the average
rates of microbial transformation of BaP. The resulting values
were 1.94 ug/l/day for the waters of Caroline Atoll and
0.81 ug/l/day for the South China Sea.

The abilities of marine microflora to degrade BaP are
therefore relatively high. In the case of the Caroline Atoll, the
biodegradation rates exceeded those of similar processes
investigated in impacted areas such as the Baltic Sea (Tsyban
ctal., 1985a). The results of our 1988 studies demonstrate the
need to consider PAH, and especially BaP. metabolism as an
exceedingly important process for detoxification through the
elimination of the pollutants in question from the highly
dynamic ecosystem that constitutes the tropical zones of the
World Ocean.

TABLE 2

Dynamics of benzo(a)pyrene transformation by microflora from central Pacific waters in the in situ

experiments (October 1988).

Region of

Operations and

sampling depth.

m

BaP concentration, ug/1

Exposure, Initial At end of experiment
days Cj Control Exp.

Microbial
transformation of

BaP, in % of
concentration C,

Caroline Atoll
Lagoon, S m

South China Sea.
Station 127,0.5 m

Same as above

0

id

1

10

2

10

3

10

4

10

5

10

5

10

()

2.0

3

2

5

2

7

2

7

2

0

10

3

10

5

10

7

10

7

10

9.9

9.9

2.0

2.0

10

9.9

7.8
4.9
3.?
1.8
0.3

0

1.71
1.37
0.46

6.62
6.00

4.32

0
22
51
65
82
97
0

14.5

3 1 .5

77.0

0

0
33.8

40.0

56.8

0

190

2.4 Cesium- 137 in the Surface Waters of the
Central Equatorial Pacific

VLADIMIR I. MEDINETS5 , VLADIMIR G. SOLOVEV , and BORIS V. GLEBOV*

'State Oceanographic Institute, Odessa Branch, USSR

* Institute of Global Climate and Ecology, State Committee for Hydrometeorology and Academy of Sciences, Moscow, USSR

Introduction

The study of radioactive contamination of ocean waters
throughout the world is one of the most important tasks facing
researchers concerned with the effects of human activities on
the ocean environment. Man-induced radioactivity of the
marine environment has three origins: 7. nuclear weapons
testing; 2. nuclear power plant operation; and 3. operation of
nuclear fuel treatment and reprocessing plants.

The need for more information about the present status of
radioactive contaminants of World Ocean waters prompted
our research efforts in the course of the 47th cruise of the R/V
Akademik Korolev in the central portion of the central equatorial
and western Pacific from September to November 1988.

The indicator of radioactive pollution selected was cesium-
1 37, a radionuclide with a long half-life. It is this radionuclide,
together with strontium-90 and plutonium-239, that presents
the greatest potential threat to the marine environment.

Methods

Seawater samples of large volume (0.8-1.1 m1) were
drawn from various depths (0-200 m) using NIVA and Malysh
immersion pumps. Selective extraction of the cesium- 1 37 was
performed with the aid of Milton-T fibrous sorbent impregnated
with copper ferrocyanide (Vakulovsky, 1986). Following
extraction, the sorbent was reduced to ash in a flameless muffle
furnace at a temperature not exceeding 450°C. The ash residue
was then hermetically sealed in a polyethylene capsule. In
December 1988. the samples were measured using a gamma-
spectrometry setup at one of the laboratories of the State
Oceanographic Institute, Odessa Branch. The error of the
cesium- 137 determinations did not exceed 10%. The detection
threshold was 0.01 Bq.

Results

Cesium- 137 levels in surface waters were determined in
the central equatorial portion of the Pacific, as well as in the
Philippine, South China, East China, and Japan Seas. The
results are summarized in Table 1. Figure 1 shows the location
of cesium-137 level measurements in the surface (0-3 m) layer
of Pacific Basin waters along the expedition route.

Analysis of the data showed that the spatial distribution
of cesium-137 levels was very uneven. The absolute
minimum and maximum concentrations were recorded in the
equatorial Pacific. The minimum level recorded in the course
of the entire expedition was situated at 8°N, 1 56°35'W between
Hawaii and Christmas Island. This was at a time when the
Akademik Korolev was traversing the northern boundary
of the intratropical convergence zone. Precipitation was
abundant, and the resulting dilution was what, in our view,
lowered the cesium- 1 37 concentration in the top layer down to
0.7 Bq/m\

A maximum concentration of 7.0 Bq/m3 was recorded in
the vicinity of Kusaie Island in the Carolines group.

During this period, the vessel was at its shortest distance
from the Marshall Islands that include the Bikini and Eniwetok
Atolls, sites where nuclear weapons were tested in the 1950's.
The elevated levels observed were attributable either to the
transport of the cesium-137 enriched water masses from the
Marshall Islands or to the presence of the local sources on the
Kusaie Island, lying eight miles north of the cruise track. A
more detailed survey of the region might allow identification of
sources.

Elevated (3.4-3.9 Bq/m1) cesium-137 levels in the waters
off Tarawa and Caroline Atolls and Christmas Island may have
been due to the transport of oceanic waters from the Tuamotu
Archipelago, where nuclear weapon testing is regularly
conducted by France. The considerable difference between the
cesium-137 levels in samples taken two miles off Caroline
Atoll and those measured in the waters of its lagoon was
especially noteworthy.

The cesium-137 level in the lagoon water was double the
value obtained for oceanic water taken off the atoll. This may
be attributed to cesium-137 enrichment of the inner lagoon
waters by the flushing of radionuclides deposited on the atoll
surface by atmospheric fallout. Analysis of the average levels
of cesium-137 for each region (Table 2) showed that the
'cleanest' waters of all the ocean areas studied were those of the
equatorial Pacific. In terms of radioactive pollution, this region
must be ranked between the Bering Sea (Medinets et al. . 1 992)
and the rest of the regions of the Pacific.

The findings of the present study agree with the results
obtained by Japanese researchers who worked in the central
Pacific in 1980-82 (Nagaya & Nakamura, 1985).

191

TABLE 1

Cesium- 137 levels in Pacific Ocean waters in the autumn of 1988.

TABLE 2

Mean cesium- 137 levels in seawater bv region.

Date

Cooi

dinates

Sampling

Mean levels

Latitude

Longitude

depth ( m )

Bq/m3

09-15

8°00' N

156°35' W

0-3

0.7

09-16

4°40' N

157°04'W

0-3

2.7

09-18

1°28'N

157°43' W

0-3

2.6

09-19

1°I6'N

I56°00'W

0-3

3.5

09-20

5°49' S

153°05'W

0-60

3.6

09-2 1

6°53' S

I53°21'W

0 - 7(1

2.9

09-25

9°59' S

!50°15'W

(1- 100

1.9

09-28*

9°59' S

15015' W

0-3

3.9

10-02

6°37' S

!61a44'W

0- 100

3.6

10-04

3°48' S

1 72°03' W

0- 100

-i 2

10-07

0°16' S

177°38' 1

0- 20

2.0

10-08

l°20'N

172 20' E

0- 100

1.9

10-11

2°41'N

168C17'E

0- 3

3.3

10-11

3°03' N

I61°01'E

0-3

2.3

10-12

4 20'N

162°27'E

0- 3

7.0

10-13

7°21'N

156°42' E

0-3

4.4

10-14

10 08' N

153=45' E

0-3

5.9

10-15

1016' N

147°45'E

0-3

3.4

10-16

i r 12' n

I42°55'E

0-3

2.9

10-17

11 1 5' N

1 39°35' E

0-3

3.0

10-18

noo'N

1 36°00' E

0- 120

4.4

10-18

1 1 °05' N

134°07'E

0-3

4.9

10-19

1 10()()' N

I32°21'E

0 - 3

3.7

10-20

1 1 °00' N

130 33' E

0-3

3.3

10-21

ID 59' N

128°46'E

0 - 200

5.5

10-24

5°I2'N

I14°05'E

0-3

6.2

10-25

5 50' N

I09°27'E

0-3

5.3

10-27

6°00' N

I06°54' E

0 - 50

5.2

10-27

6°00' N

106°54'E

0 - 50

5.2

10-28

5°14'N

106°27' E

0 - 55

4.2

10-28

5°14'N

106°27' E

0 - 60

4.5

10-29

4°18'N

)05°54' E

0-80

4.8

10-29

3 23' N

105°19'E

0 - 60

4.3

10 30

: 1 9' n

1 1 14 54' E

0-4(1

3.4

11-06

4 20' N

I06°42'E

(1- 70

5.9

11-07

4 4I'N

II 47' E

0 - 70

5.2

1 1-08

8°04' N

I16°13'E

0- 3

4.9

1 1-09

ll°35'N

1 1 8°42' E

0-3

6.3

1 1-10

I7°26'N

114 53' E

0- 120

4.9

nil

21 I6'N

121°43'E

0 - 10

4.8

11-12

24 58' N

12 V 39' E

0- 10

4.3

11-13

2') 00' N

1 27°00' E

0 - m

4.3

11-13

S2 17' N

I28°09'E

0- 10

4.4

11-14

36 41' N

130°52'E

0 ■ lu

5.3

11-15

41 M'N

I32°12'E

0- 10

4.7

II 15

43 Ol'N

131°59'E

0- Hi

5.6

Sample

draw n Iron

the lagoon (

1 Caroline Ah

II.

Region

Mean level
Bq/m

Bering Seal 50-60° N)
(from Medinets <■/ <//.. 1992 )

Pacific Ocean (10-40 N)
(from Nagaya& Nakamura, le)S5i

Equatorial Pacific (5 N- 1 0°S)

South China Sea

East China Sea

Sea of Japan

2.4 ±0.2

5.2 ± 0.3

3.610.3
5.0 ±0.2

4 3 ± 0. 1
5.2 ± 0.3

Fig. 1. Map of sampling locations to determine cesium- 137 levels (Bq/m1) in
the surface layer (0-3 m). (^denotes results obtained in the present
study; 0 denotes results reported by Nagaya & Nakamura. 1985)

Comparison of the overall radiation picture in the Pacific
region with our findings on the contamination status of the
Black and Baltic Seas conducted from 1 986 to 1 OSS show s that
cesium- 137 concentrations in the Pacific arc. at present. 10 to
20 times lower than in the seas of the European region impacted
bv local resources.

192

2.5 Quantity and Distribution of Plastics:
An Analysis of Chemical Hazards to
Marine Life

GREGORY J. SMITH' and CHARLES J. STAFFORD;

"Wildlife International Ltd., Easton, Maryland, USA

fUS Environmental Protection Agency, Analytical Chemistry Laboratoiy, Beltsville, Mankind. USA

Introduction

In recent years, plastics in the marine environment have
been recognized as important pollutants of marine ecosystems
(Laist, 1987; Prater, 1987; Wolfe, 1987). Medical waste,
plastic debris, and other types of refuse have washed ashore on
Atlantic Coast beaches at alarming rates. These events have
created new public awareness of the critical nature of the
plastic waste disposal. Plastics present a unique disposal
problem because the same attributes that make many types of
plastics useful also enhance their longevity and buoyancy in
the world's oceans.

The distribution of plastics has been studied in various
estuarine. coastal, and oceanic waters of the world.
Concentrations of plastics in surface waters have been associated
with oceanic convergence zones or the proximity to shipping
lanes (Colton et ai. 1974; Wong et ai. 1974; Wolfe, 1987).
Although the distribution and abundance of plastics in the
oceans have been studied almost exclusively in surface water,
it is well known that many types of plastics do not float and
must occur in marine sediments (Shaw, 1977). Gradual sinking
of those plastics that do float is also possible as bacteria,
diatoms, hydroids. and other marine life grow on the surface of
floating plastics (Carpenter et al., 1972; Colton et ai. 1974;
Winston, 1982). Morris (1980) indicated that floating plastics
gradually sink to the bottom or to a denser horizon where they
attain neutral buoyancy and remain in suspension. Because the
stratum of the pycnocline represents a marked change in water
density, this would appear to be a likely area for subsurface
accumulation of plastics with the appropriate neutral buoyancy.

Potential adverse effects of plastics in the marine
environment include aesthetic, physical, and chemical. Clearly,
floating plastic debris or litter that occurs on beaches is visually
unpleasant. Physical impacts of plastics have been well
documented. Entanglement and ingestion of floating plastics
by sea turtles have been found in several areas (Balazs, 1985;
Carr, 1987). Fish have been found to ingest plastic pellets
(Carpenter et ai. 1972), as have seabirds (Connors & Smith,
1982;Furness, 1985; Day & Shaw, 1987; Fry etai. 1987). In
several studies seabirds were found to have consumed massive
quantities of plastics (Day et al.. 1985: Furness. 1985; Fry
etai, 1987; Ryan. 1987). The physical ingestion of plastics has
been detrimental with respect to causing digestive system
impaction, ulcerative lesions, and reducing meal size

(Day etai. 1985; Fry etai. 1987; Ryan, 1988a). However,
there continues to be much speculation regarding the potential
chemical hazards of plastics to seabirds and other marine
organisms.

Raw plastic pellets have generally been considered to be
biologically inactive, although manufactured plastics often
contain additives known to be toxic (van Franeker, 1985).
Moreover, since Carpenter et al. (1972) published the first
account of polychlorinated biphenyls (PCB's) adsorbing to
plastic spherules in seawater. there has been concern for the
potential for PCB exposure to fish and wildlife that ingest
plastics. Ryan et al. ( 1988b) found evidence to suggest that
seabirds assimilated PCB's from ingested plastic particles.
However, Fry etai ( 1 987 ) suggested that plastics were unlikely
to pose a significant toxic hazard to birds compared to the
physical impaction effects that may result. Polychlorinated
biphenyls are commonly found in certain types of plastics and
the potential for plastics to adsorb PCB's from ambient water
would suggest two possible sources of PCB contamination
from marine plastics. One potential source would be intrinsic:
PCB's incorporated into the plastic during manufacturing. A
second source, extrinsic, represents PCB' s adsorbed or absorbed
from contaminated water. To date, there is still little known
about the association of environmental contaminants occurring
both in and on plastics in the marine environment and their
potential toxic hazard to marine organisms.

In this paper we report the abundance and distribution of
plastics in areas of the central Pacific Ocean and the South
China Sea. Both surface water and water at the level of the
pycnocline were sampled. Chemical extractions of the plastics
recovered were analyzed for several organic pollutants and
subsequent field and laboratory experiments were conducted
to elucidate potential toxic hazards that plastics may pose to
fish, seabirds, and other marine organisms.

Methods

Field Sampling

Sampling was conducted during the Third Joint
US-USSR Bering & Chukchi Seas Expedition to the two seas
and the central Pacific Ocean, along the segment from Hawaii
to Singapore. Trawls for plastics were made from 15 September
to 30 October 1988, along the 18,400-km cruise track of the
Soviet R/V Akademik Korolev. During the cruise, the vessel

193

spent 8 days at Caroline Atoll, Kiribati, a remote coral atoll at
9°59.09"S, 150° 14.04" W. Neuston sampling of the lagoon
water of Caroline Atoll was done using a stationary 102 u.
(0.5 x 1 ml neuston net placed in the tidal currents of the
entrance to the lagoon. The stationary net was left in position,
with the opening facing toward the lagoon, over a 7-day period,
during both flood and ebb tides. Also, one additional location
in the central lagoon was sampled during two 30-min tows
using two neuston nets ( 102 |i) towed alongside an inflatable
boat.

Three different types of sampling methods were used in
the open ocean. These included /. surface sampling using a
1 x 2-m tucker trawl (in the open position) equipped with a
94 u. net; 2. surface sampling with a 102 U. neuston net with a
0.5 x 1- m opening; and 3. subsurface water sampling at the
pycnocline horizon by deploying and hauling the tucker trawl
in the closed position and fishing at depth in the open position.
The pycnocline was determined using Soviet equipment
installed aboard the research vessel to determine temperature,
salinity, and density. Each net apparatus was equipped with a
flow meter to determine the linear distance sampled that was
later expressed on a surface area basis for surface sampling and
a volume basis for sampling at depth.

Surface water trawls from the ship averages 15 min with
half of the rectangular opening of each net below the surface.
The tucker trawl was towed off the bow of the ship at speeds
between 82.0 and 280.6 cm/s. Neuston trawls were done from
the stern of the ship while drifting at speeds from
5.4 to 55.0 cm/s. At the end of each tow the sample was
removed and passed through a series of acetone/hexane rinsed
stainless steel sieves with pore sizes of 4 mm, 1 mm, 500 |i, and
106 u to fractionate the sample. Each fraction was examined
under magnification for the presence of plastics; marine
organisms were separated from the sample for storage. Plastics
recovered from the sample were immediately characterized as
to size, type (raw pellet, fragments of plastic objects, fishing
gear), and frozen in chemically-clean glass jars.

Chemical Analysis

All plastics (fragments and pellets) were removed from
I lie fractionated sample, weighed, and extracted by shaking the
sample three times in separate 5-ml rinses of hexane. Hexane
extraction was used to remove nonpolar organics from the
surface of the plastics and to avoid chemically dissolving the
plastic samples. The extracts were combined and concentrated
to a volume of I ml using a stream of dry nitrogen. Sample
extracts were analyzed for chlorinated hydrocarbons using a
Hewlett-Packard (HP)589()A gas chromatograph equipped
with a "'Ni electron capture detector and a 30m DB1 column.
An IIP 5890A interfaced with an HP 5970 mass selective
detector, in the full scanning mode from 50 to 450 atomic mass
units, was used to detect the presence of other halogenated
hydrocarbons and petroleum hydrocarbons.

Plastic Adsorption and Chemical Release Experiment

1 o assess the possible chemical hazards of plastic to
marine life two experiments were conducted. Theseexperiments
were designed to evaluate the potential adsorption of organic

compounds from surface water and to determine if the
gastrointestinal environment of birds could cause the release of
organic compounds from raw polyethylene pellets.

To determine the adsorption potential of plastics
from surface water, 100 g of new polyethylene pellets
were placed in a 1-m-diameter brass-wire mesh enclosure in
Baltimore Harbor. Chesapeake Bay. After 24 h. the pellets
were collected, handled, and analyzed according to
the procedures described above. At the same time the plastics
were collected, a sample of the surface microlayer was collected
by contact and adhesion of surface water to a glass plate.
The plate was dipped repeatedly and rinsed with methylene
chloride to obtain a sample volume of approximately 1-1
of water from the microlayer. Water microlayer samples
were extracted with methylene chloride and analyzed using
the same methods as the plastic extracts. The purpose of
the microlayer sample was to facilitate the comparison
between compounds found in the surface microlayer and
what was potentially adsorbed to the plastic pellets.
This experiment was conducted to assess the possible
adsorption characteristics of plastics under environmental
conditions.

To determine if chlorinated or petroleum hydrocarbons
could potentially be released from polyethylene pellets in the
digestive system of birds, a simulation experiment was
conducted. The experiment was patterned after those of
Kimball and Munir ( 1 97 1 ) using a 42 C HC1 bath to simulate
the physicochemical conditions in the digestive system of
waterfowl. Hydrochloric acid was added to 1 N saline solution
to yield digestive solutions with pH"s of 1.4 and 2.8. Two
50-g samples of polyethylene pellets were added to glass jars
containing either a 1 N saline only (controls). pH 1.4 solutions,
pH 2.8 solutions, or a hexane (without saline). These solutions
were agitated in a water bath at 42°C. At the end of 8 days, the
plastics and the digestive solutions were extracted with
methylene chloride and these extracts were analyzed for
chlorinated and petroleum hydrocarbons using the methods
described above.

Results and Discussion

Over 80,000 m: of surface water and 93.000 m3 of
subsurface water were sampled during the course of the
expedition (Table 1 ). The total number of stations sampled
using either the 947 \i tucker trawl, the 102 |i neuston net. or
both at a single location was 28 and 8 for the Pacific Ocean and
South China Sea. respectively. Sampling at Caroline Atoll was
done at only two locations. Plastics were recovered from 219i
of the sample collected in the Pacific Ocean; five of six positive
samples contained opaque polyethylene pellets. The sixth
sample contained plastic fragments, not raw pellets. The
concentration of plastics at positive stations varied considerably.
with a maximum of 0. 1 8566 mg plastic/m: (Table 2 1. Although
sampling was done along the cruise track from as far north as
10°N to as tar south as 10°S. plastics were only recovered from
stations near the ION latitude area (Fig. I ). Tar balls were
found at only two of the stations sampled in the open waters of
the Pacific (Table I).

194

TABLE 1

Sampling information and recoveries of plastics from the central Pacific Ocean and South China Sea. Sampling
was done from 15 September through 30 October 1988.

Volume(m')

Total

Number of Sampling Stations

with

Geographic

Sample

Net

or Area (m:)

No.

Plastic

Plastic

Tar

Region

Type

Mesh(u)

Sampled

Stations

Fragments

Pellets

Balls

Pacific Ocean

Surface

947

39.214

23

5

5

2

102

3.627

8

1

0

2

Pycnocline

947

5 1 ,657

9

0

0

0

Caroline Atoll

Surface

102

21.571

fixed
station

0

0

0

Surface

102

2.508

2 trawls

0

0

0

South China Sea

Surface

947

15.213

8

7

2

2

102

394

1

1

0

0

Pycnocline

947

42,107

4

1

0

0

TABLE 2

Occurrence and concentrations of plastics from the central Pacific

Ocean and South China Sea. Sampling was done from

15 September through 30 October 1988.

Plastic" Concentrations (mg/m:)
Stations Sampled at Positive Stations

Geographic No. of No. w/

Range Locations11 Plastics Mean Range

Pacific Ocean 28 6

South China Sea 8 7

0.07340 0.01309-0.18566
0.19478 0.00126-0.69632

■^

>m. M ,!,:(

Plastic includes pellets and fragments.
' Several stations were represented by more than one sample.

In the South China Sea, seven of the eight different
locations sampled had plastics present in the surface water, but
only two of the seven had plastic pellets (Table 1 ). Most of
the plastics found in this area consisted of synthetic fishing
line and secondary, manufactured plastics in the process of
breaking down. Plastic concentrations in the South China
Sea samples were considerably greater than those found in
the Pacific Ocean, with a maximum concentration of
0.69632 mg plastic/m2 (Table 2). Although only two stations
sampled had tar balls present, one of these samples contained
hundreds of tar balls and 164 of these were greater than 4 mm
in size.

Sampling of the lagoon water at Caroline Atoll was done
at two locations, one in the central lagoon and the other at a
fixed station at the entrance channel to the lagoon. At the fixed
station, current flowed through the neuston net at an average

Fig. 1 . Locations sampled along the cruise track of the Akademik Korolev,
September-October 1988. Sampling locations are indicated by solid
dots. Plastic densities at positive stations are given in mg per meter
square. Major currents for October 1988 are shown by the arrows.

velocity of 4.08 cm/s. No plastics or other anthropogenic
materials were recovered from any of the samples at Caroline
Atoll (Table 1).

The concentrations of plastics in the open Pacific Ocean
stations that had plastics present in the surface water were
below mean concentrations reported for the subtropical North
Pacific (Wong el al, 1974; Shaw & Mapes, 1979) and higher
than from other studies in the North Pacific (Day & Shaw,
1987). Results from this study indicate that a much greater
concentration of raw plastic pellets occur in the central Pacific
compared to the North Pacific. There are no other data from the
central Pacific region that can be used for temporal comparisons.

The greater frequency of occurrence of raw material
plastic pellets compared to discarded plastic objects and tar
balls in the South China Sea was somewhat surprising. The
occurrence of discarded plastic may be expected from the

195

heavy shipping traffic in this region of the world. Moreover,
there is extensive offshore oil production in the South China
Sea and a high occurrence ot tar balls would also be expected.
The greater frequency of occurrence of plastic pellets compared
to tar balls is noteworthy. The longevity of plastics in the
marine environment likely contributes to the increase of this
material in the oceans.

The distribution of plastics in the Pacific Ocean was
largely a function of the major currents. All stations sampled
that contained plastics were along a l()°N latitude convergence
lying between the North Equatorial Current and the North
Equatorial Countercurrent (Fig. 1 ). The most likely source of
the plastics recovered at this convergence would either be from
ships or as industrial waste from plastic-producing Pacific Rim
countries.

All plastic sample extracts were analyzed for the organic
compounds listed in Table 3. None of the samples contained
quantifiable concentrations of organochlorine pesticides, PCB's,
aliphatic hydrocarbons, or polycyclic aromatic hydrocarbons
included in the analyses. The absence of detectable levels of
these contaminants suggests that the plastics that were collected
at sea did not adsorb any of these compounds, either because
the plastic surface did not facilitate this or because the
contaminants were not present in sufficient concentrations.

In the field experiment conducted in Chesapeake Bay's
Baltimore Harbor, plastic polyethylene pellets left floating in
the water for 24 h were extracted and analyzed using the same

TABLE 3

Organic compounds included in the analysis of extractions of
plastics

Chlorinated
Pesticides
ami PCB's

Aliphatic
Hydrocarbons

Polycyclic

Aromatic
Hvdrocarbons

Heptachlorepoxide
Oxychlordane
Trans - chlordane
Cis - chlordane
Trans - nonachlor
Cis - nonachlor
Dieldrin
I .mil i ii
p.p" - DDT
p.p" • DDD

p.p" 1)1)1

PBC(1254)

n - dodecane
n - tridecane
n - tetradecane
n - pentadecane
Nonycyclohexane
n - hexadecane
n - heptadecane
n_- octadecane
n - nonadecane
n - eicosane

Tetrameth) I-
pentadecane

Naphthalene

Fluorene

Phenanthrene

Anthracene

Fliioranthrene

Pj rene

1.2. -ben/anthracene

Chrysene

Benzo(b)fluoranthene

9,10-

diphenylanthrocene

Benzo(e)pyrene

Benzo(a)pyrene
1.2.5.6- dibenz
anthracene
Benzo(g,h,i)perylene
Pen lene

The lower limit of quantification was 0.05 ug/g for aliphatic
hydrocarbons, 0.4 ug/g lor polycyclic hydrocarbons, mis ug/g for
chlorinated pesticides, and 0.25 ug/g foi l'( !B's based on a 0.2 g

sample.

methods as those collected at sea. A sample of the microlayer
was taken the same time these plastics were placed in the water
and was analyzed for the same contaminants as the plastic
extracts. Two aromatic compounds, pyrene and fluoranthene,
were detected in the microlayer sample: however, only
fluoranthene was present in the extract of the polyethylene
pellets.

These data indicate that polyethylene pellets have the
potential to adsorb at least one organic compound, fluoranthene.
from ambient seawater. The specific mechanism of adsorption
could be either a binding of the organic compound to the plastic
surface directly or to a film of water surrounding the plastic
pellet. It is possible that other contaminants, especially those
occurring in high concentrations in the microlayer, could also
be adsorbed to floating plastic debris. Therefore, a potential for
the transfer of certain organic pollutants from plastics to
organisms that ingest them does exist. Plastics may serve as a
vehicle for pollutant transport that may enhance exposure of
organisms; however, if plastics also concentrate contaminants
from the water this would result in an even greater hazard to
marine organisms.

The other aspect of chemical-induced toxicosis resulting
from plastic ingestion is that of the potential release of chemicals
from the plastic to the organism. In the second experiment,
none of the extracts of the acidic digestive solutions, or the
hexane extraction of plastic pellets contained quantifiable
concentrations of any of the hydrocarbon contaminants. These
data suggest that even when exposed to mild heat, acids, and
hexane, polyethylene pellets did not release appreciable
quantities of chemicals. However, plastics subjected to real
avian digestive systems would also be exposed to digestive
enzymes and tor possibly much longer periods of time. Ryan
( 1988b) found that domestic chickens fed polyethv lene pellets
retained 98.3% ofthemoveran 18-day period. Moreover, there
are many types of plastics and additives to plastics that could
potentially be released during digestive processes in marine
vertebrates.

Results of this study show that raw material plastic pellets
are becoming increasingly more common in areas of the Pacific
Ocean far removed from industrial sources. Although tar balls
were once considered more common and widely distributed
than marine plastics, their occurrence in the present stud) was
less than that of plastics even in areas where oil development
and tanker traffic is hea\ \ . Plastic pellets do have the potential
to adsorb certain organic contaminants from seawater: however,
the types of compounds that can be adsorbed and possibly
concentrated is not well understood. Polyethylene pellets
subjected lo conditions simulating the avian digestive system
did not release detectable levels of chemicals However, other
types of plastics, and those that contain additives, were not
tested.

Clearly, fish, seabirds, and other marine organisms will
continue to be exposed to plastics at increasing rates. Regulations
prohibiting ocean dumping of plastics have already been
enacted: however, recycling, waste management, degradable
plastics, and other alternatives must continue to he developed
and implemented to abate the global problem of plastics in the
oceans of the world.

|0d

This project was part of the First Joint US-USSR Central
Pacific Expedition aboard the Soviet Research Vessel, Akademik
Korolev. We express our appreciation to the US Fish and
Wildlife Service. USA, and the State Committee for

Hydrometeorology. USSR, who made our participation possible.
We also thank H. O'Connor, A. Tsyban, O. Rostovsev, Y. Volodkovich,
B. Aleksandrov, L. Polishchuk, S. Kohl, A. Krynitsky, M Wong, and
P. Wills.

2.6 The Role of Solar Irradiation in the
Oxidative Transformation of
Benzo(a)pyrene

NATALYA I. IRHA, EHA R. URBAS. and UVE E. K1RSO
Institute of Chemistry, Estonian Academy of Sciences, Tallinn, ESSR

Introduction

Organic pollutants entering water bodies, including the
oceans, contain a wide range of substances. Some of these are
quite easily subjected to degradation under the action of natural
agents, while many other compounds are stable and persist in
various compartments of the ecosystem. Among the persistent
compounds are the carcinogens belonging to the polycyclic
aromatic hydrocarbon (PAH) group. It has been established
that PAH' s enter the ocean in several ways, emissions into the
atmosphere as a result of incomplete combustion of fuels,
spillage during transportation of oil, and so forth. On the other
hand, it is known that there exists a natural occurrence of
PAH's (natural background) in the oceans. It is the view of
several researchers (Tsyban, 1975; Izrael. 1984; Kirso et ai.
1988) that the main mechanism of self-purification of the
hydrosphere from carcinogenic PAH' s is biological ( bacterial )
oxidation. However, there are also other mechanisms for
removal, including photooxidation under solar irradiation ( Mill
etai, 1981; Bockris, 1982). It may be assumed that this process
plays a significant role, especially in the surface water layer
where the content of soluble oxygen is near \00c/c saturation
(i.e., 10 '■' mol/1). There are numerous factors that modify this
process, including the physicochemical characteristics of water
( turbidity , transparency, and the presence of other compounds )
and the air above the sea surface (condition of the atmosphere,
i.e., environment and weather conditions). Figure 1 illustrates
the characteristic sunlight spectra at the sea surface (Ranby &
Rabek. 1978). Photochemical reactions in the marine
environment are also moderated by the level of atmospheric
pollution with aerosol particles and smoke, as well as changes
in ozone and other impurities. Solar irradiation at wavelengths
less than 285 nm is to a large extent absorbed by ozone in the
upper layer of the atmosphere. Therefore, the active spectra for
this process are usually at wavelengths greater than 285 nm.
Variations in the characteristics of the ozone layer therefore
influence all the photochemical processes occurring on the
oceans' surface.

1000

E
o

200

3.

u

50

o

m

10

S-4

n.

UJ

0.5

0.1

ISO

320

360

400

X,,nm

Fig. 1 . Seasonal variations in solar spectra at the sea surface. I = July,

2 = December. E = Spectral Energy; A. = Wavelength (Ranby &
Rabek. 1978).

The extent of these processes and their contribution to self-
purification of the marine environment from carcinogenic
PAH's are determined by both the value and distribution of
solar irradiation energy (Mill et al, 1981; Rabek, 1985)andthe
level of pollution of seawater, as well as the concentration and
composition of pollutants. The latter may have an influence
upon the pattern and intensity of the degradation processes of
PAH's (Kirso & Gubergrits, 1971; Gubergrits et al.. 1975).
Although systematic studies have been carried out on the
photooxidation of individual PAH's in water (Kirso et ai,
1971; Gubergrits etai, 1975; Paalme etai, 1976. 1983). there
are few data on these processes under natural conditions.

197

Therefore, during the cruise of the R/V Akademik Korolev
(July-November 1 988 ). a study was undertaken on the kinetics
of the oxidative photolysis of a typical carcinogenic PAH —
benzol a tpyrene (BaP) — (Clar, 197 1 ) in the Bering Sea and the
tropical waters of the Pacific (Table 1 ).

It is known that the Bering Sea is almost wholly in the
subarctic zone, excluding its northern parts, which are in the
arctic temperature zones (Izrael & Tsyban, 1987). The main
body of its waters is characterized by a subarctic structure
whose specific feature is the existence of cold and warm
intermediate layers. The upper layer thickness average
25-50 m, the salinity being 32.8-33.4%o and the temperature,
about 5 to 7°C. According to Izrael and Tsyban ( 1 987), PAH" s
are permanent and typical components of these ecosystems.

TABLE 1

Exposure of BaP solution in seawater (the 47th cruise of the
R/V Akademik Korolev).

Month

Exper.

Coordinates

Water

Average dose of

(1988)

Number

Temp.
t=°C

solar radiation
Q MJ/m: during
the first 3 hours

Aiiyust

1

53°58'N/I76°28'W

14.1

1.48

2

53°58'N/176°28,W

15.7

1.09

September

3

09°54'S/156°23,W

26.4

3.50

4

04°02'S/154°07'W

27.0

3.09

5

09°59'S/150°15'W

27.0

4.90

6

()9°()0'S/I50°0()'W

27.0

3.67

October

7

09°00,S/150°00,W

27.0

2.70

8

09°0()'S/150°0()'W

27.0

3.00

Taking into account the low influence of this area from
human activities, the physicochemical parameters of the
atmosphere above these waters and characteristics of the surface
water layer, a study of sunlight photolysis of PAH' sin seawater

at lower temperatures and low intensity of solar irradiation was
of interest. The same experiment was carried out under tropical
conditions — that is. at higher temperatures and considerable
insolation with a salinity of 35.4%o.

Methods and Materials

The surface water samples taken at a depth of 0 to 0.5 m
were sterilized by autoclaving for 2 h. cooled, and filtered
through main layers of Sterilized cloth into a sterilized dish.
The BaP solution in seawater was prepared in 5-1 glass cylindrical
reactors whose sides were covered with black paper (water
column height 20 cm). Benzol a (pyrene was introduced as an
ethanol solution (ethanol content up to 0.01295 ) by constantly
mixing with a magnetic stirrer for 15 minutes. The BaP
concentration was varied from 0.7 to 6.6 nanomoles mm).

This solution was exposed to sunlight at the temperatures of the
surface water for each given region. Simultaneously, the total
solar irradiation dose, Q (MJ/m: per h), was measured
pyranometrically (Table I). The experiments were carried out
with the following solutions:

(a) BaP solution in seawater;

(b) BaP solution in sterilized seawater; and

(c) BaP solution in sterilized seawater protected from
light by black paper (autoxidation).

The exposures lasted for approximately 3 h in the first half
of the day. The solutions were periodically sampled and the
BaP concentration measured by chromatography techniques.
Conditions of analysis were as follows:

A 'Jasko' HPLC chromatograph (Japan), a
fluorimetric detector, Xtx = 295 nm, Xcm = 408
nm, solvent methanol-water (95:5) isocratic
regime, eluent flow-rate 0.7 ml/min using a
25 x 0.25 cm ODS column, sample volume was
100 ul.
Statistical kinetic data processing was performed using the
least-squares method.

Results and Discussion

From kinetic data (Figs. 2-5. Table 2), it follows that
during the first hour of exposure, a decrease in the BaP
concentrations in seawater is described by a formal-kinetic
equation for the first-order reaction where c0 and c, are the
initial BaP concentration at zero time, and that at a certain time
t, k is the constant of the first-order reaction, the dimensions for
this constant are per second (s ').

ln(c,/c,) = kt

The rate constant values obtained for the Bering Sea areas,
the tropical part of the Pacific and the Caroline Atoll (Table 2)
are of the same order as those found (Mill et <//.. 1981) by
photolysis of BaP in water (X = 366 nm. BaP concentration
5x 10s m, t° = 22-28°C, k = 3.86 ±0.71 x 10V). The half
decay time of BaP under these conditions was 0.69 h. In our
experiment the average rate constant value per unit Q was
(1.1 ±0.1) 10V.

Irrespective of the areas under study and the sterility
of water in those treatments that were subjected to
photolysis (Table 2). the initial stage of the process, as
mentioned above, is well described by an equation for the first
order kinetics. At the same time, according to experimental
data (Figs. 2-4), the oxidation of BaP in seawater
without sunlight (autoxidation) for the Bering Sea areas and
the tropical pari of the Pacific is described by a formal-
kinetic of a second-order type (r = 0.90-0.98). The rate
constant value of BaP autoxidation in the Bering Sea water
at a concentration of 5 x 10 s M (July, t0=21°C,
Q,,r0.68 MJ/m" per h) was (0.4 ± 0.01) 104M's!. but in
the tropical part at a concentration of 1.7 x 10s (October
t°=27°C. Qr,,~3.00 MJ/nr per hi (9.08 ± 1.90) 1()4M 's '.
respectively.

198

TABLE 2

Kinetic characteristics of photochemical transformation of
BaP in seawater under solar irradiation in the following areas:

(a) the Bering Sea

(b) the tropical part of the Pacific

(c) the Caroline Atoll

(47th cruise of the R/V Akademik Korolev, July-November 19881.

Experiment

Initial BaP

Rate

Number of

Correlation

Sterility of

No. See

concentration

constant

data points

coefficient

media

Table 1

M 10 K

10V'(k±n)

r (first-order)

(a)

1.

1.47

1.69±0.13

5

0.99

-

2.

4.20

1.60 ±0.08

7

0.99

+

4.44

1.20 + 0.16
(b)

6

0.97

-

3.

6.59

2.84 ±0.48

5

0.96

+

2.85

2.85 ±0.98

4

0.90

-

4.

2.06

1.60 ±0.49

(C)

4

0.92

+

5.

1.90

2.99 ±0.70

4

0.95

-

2.14

4.05 ±0.51

4

0.98

-

6.

0.70

4.20 ±0.11

4

0.99

+

7.

11.10

8.77 ±0.23

4

0.99

-

8.

2.00

3.67 ±0.27

6

0.99

-

C/Co

c
o

c

—

c

<u
o

c
o
o

a.

CO

1.01

s*-.-

-•-

'*

► •

"c

6

0.5

O

/

3

va.a

b *

o

0

?

t .hour

Fig. 2. Kinetics of BaP degradation under sunlight irradiation at the surface
of the Bering Sea (coordinates: 67°42'N 1 15°43'W): a) in sterilized
seawater: b) in non-sterilized seawater; and cl by autoxidation (w/o
light).

Thus, in the areas of exposure the rate constant values of
BaP autoxidation in seawater differed within one order of
magnitude. The shape of the kinetic curves (Figs. 2-4) also
gives evidence of a negligible decrease in BaP concentration
over time without sunhght.

While in laboratory experiments by photo-initiated
UV-irradiation within a relatively wide wavelength range —
around 200 nm — and a high oxygen atmosphere in a liquid
medium, the degradation of BaP did not change until the

C/Co
1.0

c
o

•5 0.5

c
o
o

(X

a

CO

Fig. 3.

p
©

0

c

4

3

2

G)"n

b©

-n -

©00

~©

OO--
a ~

1

©u

o

0

9

t .hour

Kinetics of BaP degradation under sunlight irradiation at the surface
ofthe Bering Sea (coordinates: 53°58'N 175°28'W): a) in sterilized
seawater: b) in non-sterilized seawater; and c) by autoxidation
(w/o light).

disappearance of the initial substance (zero-order reaction)
(Gubergritse/fl/., 1975: Paalme etal, 1983). Under natural
conditions in water (i.e., by sunlight photolysis), a decrease in
the reaction rate is observed after the first hour of exposure
(Figs. 2-5 ). This is especially noticeable in the Bering Sea area
( Figs. 2-3 ). From the graphs ( Figs. 2-5 ), it follows that the BaP
half-decay times (t":) in the Bering Sea waters were not less
than 1 h. but in the tropical part ofthe Pacific, they were 0.6 h.
The initial rate constant value of sunlight photolysis of BaP in
the tropical zone water averaged 8.9 x 1 O12 mole s ', exceeding

199

0 12 3 4 t, hour

F;ig. 4. Kinetics of BaP degradation under sunlight irradiation in the tropical
part of the Pacific (coordinates: 09°54'N I56°23'W): a) in sterilized
seawater; b) in non-sterilized seawater; and c) by autoxidation (w/o
light).

2 t, hour

Fig- 5. Kinetics of BaP degradation under sunlight irradiation in the lagoon
of the Caroline Atoll (coordinates: 09°54'N I5fv2.vwi: al j„
sterilized seawater; bi in nonsterilized seawater: c) by autoxidation
lu/o light).

by x3 the rate tor the Bering Sea area at similar initial
BaP concentrations (2.44 x 10 i: mole s '). For comparison,
it may he pointed oui thai at the initial concentrations of 10 " M,
the initial rate of sunlight photolysis of BaP in the Caroline
Atoll is much higher than that found for BaP autoxidation
in water under laboratory condition 1.67 x 10 i: and
0.3 x 10 -|: mole s '. respectively (Kirso et al., 1983).

Thus, experimental data, as expected, gave evidence of a
dependence of BaP degradation on the location of exposure,
intensity of solar UV-irradiation, and temperature of the
environment, and data agreed well w ith the results obtained by
Graupera and associates (Graupera et al., 1988). Obviously,
during photolysis of BaP by sunlight in seawater, the existence
ofmicroimpurities, inorganic components, salinity, and general

water composition all have a major impact on photochemical
reactions in different areas of the world oceans. A comparison
of experimental kinetic parameters and literature data (Mill
et al., 1981 ) suggests that proceeding from the values of k at
T"- of BaP under similar conditions of its photolysis in water
(X = 366 nm and concentration of 5 x 10 s M) the value of
quantum yield 0evp (the number of molecules subjected to
transformation as a result of adsorption of one energy quantum )
is almost equal to <p - 5.4 x 10"4 — that is, less than one (Mill
et al.. 1981). Consequently, the processes under study are
complex, involving competing chemical reactions. Thus, the
degradation of BaP is initiated directly or indirectly and proceeds
under solar irradiation.

Photochemical reactions were investigated with soluble
oxygen in the presence of different inorganic and organic-
components in water bodies. Zafiriou (Zafiriou et al.. 1989)
presented the following scheme for generation of free radicals
in the marine environment:

O, + hv > initiation of radicals R'

R- + R,NO > RNO\ R:NOH, R:NOR

seawater + O, + hv - - > initiation > oxygen con-
taining radicals

seawater + 0: + hv > superoxide

0.+ NO - - >00 * NO* H+ - -> * NO, - -> 02

seawater + *0; + hv — > H * O * OH product
seawater + H * O * OH + hv - - -> *02 + H:*0

According to (Mopper et al.. 1989) the concentration of
high-energy ( more than 4 kcal/mole ) oxygen-containing radicals
in seawater is low and makes for hydroxyl groups (OH) — for
example, in subtropical coastal areas 1 1.9 x 10 ls M and for
open sea 1.1 x 10 IBM, correspondingly. Consequently, it may
be assumed that BaP (and other PAH's) is subjected to
photodegradation due mainly to secondary photochemical
reactions w ith different reactive radicals formed directly under
the action of light quanta or indirectly (see the scheme).
According to our results, the rate of BaP transformation depends
primarily on the sunlight intensity. Obviously, then, the
mechanism of photochemical oxidation of organic xenobiotics
of the PAH type is not different in northern and southern areas
(i.e., the amount of oxidizing particles sufficient for
transformation is generated whose excess favors the first-order
reaction [pseudomonomolecular] relative to BaP).

A study of the influence of inorganic salts and micro-
impurities on the photochemical processes in the marine
ecosystems requires further research.

To sum up, it should be pointed out that the experiments
(under natural conditions) were carried out with only one
reference PAH. ben/o( a)pyrene, and in the marine environment
many other PAH"s are present (see Fig. 6). To estimate the
reactivities of other PAHs undergoing photooxidation in
seawater. the data obtained by Paalme et al. (1988) were
considered. It appears that the rate of the process for individual
homologs differs by factors of over 140 (Fig. 6). Anthracene
and its derivatives are easily oxidized, while the more condensed
systems — for example, coronene — remain more stable. It
may be assumed that as a result of photochemical oxidation, the
quantitative ratio of PAH's m the marine environment shifts
low, ml the heavier homologs.

200

a
□
□

20

10

EL

.-••

PI

•■■

-■■'
-•■'

/
/
/
/
/
/
/
/
/
/
/
,-■

.■■'

,-■
,-■
/

/

/

BaP N

K

Ac

JZL

R

-■•■
.••■■

Pn

Bba BaA TF Chr

^A

Y\ B B B ^

/
/
/
/
/
/

/

/
/
/
.--
.--
/
/
/
/
/

J

PI

5

CPAH

/
/

/
/

- 3

Pv

Per BacA BajA BahA BeP Cor BkF BbF

Fig. 6. Relative rates (v) of photoinitialed transformation of PAH's in water: (1 ) (Paalme etal., 1983); and their content (ng/1) in the water of the Bering Sea;
(2); and the Baltic; and (3) (Kirso el at.. 1989).

Thus, under conditions similar to natural ones (i.e., under
sunlight), it has been shown that a certain amount of PAH's
may be subjected to sunlight photolysis. The degree of BaP
transformation in the experiments was governed by the intensity

of total sunlight irradiation in a given region and by the
conditions of exposure (temperature, turbidity, etc.).
Autoxidation plays an insignificant role in self-purification of
the marine environment from BaP.

201

Chapter 2 References

Balazs, G. (1985). Impact of ocean debris on marine turtles:
entanglement and ingestion. In Proceedings of the Workshop
on the Fate and Impact of Marine Debris, 27-29 November
1984, Honolulu, Hawaii (R. S. Shomura & H. O. Yoshida,
eds.), pp. 387-429, NOAA Technical Memorandum NMFS.
NOA ATM-NMSF-SWFS-54, US Department of Commerce.

Ballschmitter, K. & Zell. M. ( 1 980). Analysis of polychlorinated
biphenyls (PCB's) by capillary gas chromatography.
Frensenius ' Z. Anal. Chem. 302, 20-3 1 .

Bockris. J. O. (1982). Environmental Chemistry: Moscow.
Khimiia Publishers, Moscow, 670 pp. (Russian translation)

Carpenter, E. J., Anderson, S. J.. Harvey, G. R.. Miklas, H. P.
& Peck. B. B. (1972). Polystyrene spherules in coastal
waters. Science 178, 749-750.

Carr, A. ( 1 987 ). Impact of nondegradable marine debris on the
ecology and survival outlook of sea turtles. Mar. Pollut.
Bull. 18, 352-356.

Chernyak. S. M„ Kolobova, T. P. & Vronskaya, V. M. ( 1985).
A study of chlorinated hydrocarbons. In Studies of the Baltic
Sea Ecosystem, No. 3. Gidrometeoizdat Publishers,
Leningrad, (in Russian)

Chernyak, S. M. & Mikbaleva. I. M. (1985). Chlorinated
hydrocarbons in various media constituting the Baltic Sea
environment. In Studies of the Baltic Sea Ecosystem, No. 2.
Gidrometeoizdat Publishers, Leningrad, (in Russian)

Chernyak, S. M., Vronskaya. V. M. & Kolobova, T. P. (1989).
Chlorinated hydrocarbons. In Studies of the Bering Sea
Ecosystem, No. 2. Gidrometeoizdat Publishers, Leningrad.
(in Russian)

Chernyak, S. M., Vronskaya. V. M. & Kolobova, T. P. (1992).
Migratory and bioaccumulative peculiarities in the
biogeochemical cycling of chlorinated hydrocarbon. In
Results of the Third Joint US-USSR Bering & Chukchi Seas
Expedition (BERPAC). Summer 1988 (P. A. Nagel, ed.). US
Fish and Wildlife Service. Washington, D.C.

Clar, E. (1971). Polycyclic Hydrocarbons, Vol. 2. Academic
Press, London.

Colton, J. B. Jr., Knapp. F. D. & Burns, B. R. ( 1974). Plastic
particles in surface waters of the northwestern Atlantic.
Science 185,491-197.

Connors, P. G. & Smith. K. G. ( 1982). Ocean plastic particle
pollution: Suspected effect on fat deposition in red phalaropes.
Mar. Pollut. Bull. 13. 1 8-20.

Day, R. H. & Shaw. D. G. (1987). Patterns in the abundance
of pelagic plastic and tar in the North Pacific Ocean. Mar.
Pollut. Bull. 18.311-316.

Day. R. H.. Wehle, D. H. S. & Coleman, F. C. (1985). Ingestion
of plastic pollutants by marine birds. In Proceedings of the
Workshop on the Fate and Impact of Marine Debris,
pp. 27-29, November 1984, Honolulu, Hawaii (R. S. Shomura
& H. O. Yoshida, eds.). pp. 387-429, NOAA Technical
Memorandum NMFS, NOAA-TM-NMFS-SWFS-54, US
Department of Commerce.

Fedoseeva, G. E. & Khesina, A. Ya. (1968). The use of

quasilinear luminescence spectra for the quantitative

determination of a series of polycyclic hydrocarbons.

J. Appl. Spectroscopy, Vol. 9 2,282-288. (in Russian)
Fry, D. M., Fefer, S. I. & Sileo, L. (1987). Ingestion of plastic

debris by Laysan albatrosses and wedge-tailed shearwaters

in the Hawaiian Islands. Mar. Pollut. Bull. 18. 339-343.
Furness, R. W. ( 1985). Plastic particle pollution: Accumulation

by procellariiform seabirds at Scottish colonies. Mar. Pollut.

Bull. 16, 103-106.
Graupera. B. M., Braunera, M. S., Gonzalez, R. C, Conde,

I. A. P. & Solionova, L. ( 1988). Studies on the degradation

of benzo(a)pyrene by solar ultraviolet light and temperature.

Rev. Cubana Hig. Epidemiol. 2691. 106-1 16.
Gubergrits, M. Ya., Kirso, U. E. & Paalme, L. P. (1975).

Transformation of Carcinogenic Compounds in the

Hydrosphere. Znanije Publishers, Moscow, (in Russian)
Irha, N. I., Urba, E. R. & Kirso. W. V. ( 1992). Distribution of

PAH's. In Results of the Third Joint US-USSR Bering &

Chukchi Seas Expedition (BERPAC), Summer 1988 (P. A.

Nagel, ed.). US Fish and Wildlife Service, Washington.

D.C.
Izrael, Yu. A. (1984). Ecology and Control of the State of

Environment. Gidrometeoizdat Publishers, Moscow, (in

Russian)
Izrael, Yu. A. & Tsyban, A. V. (1987). Comprehensive

Analysis of the Ecosystem of the Bering Sea. Gidrometeoizdat

Publishers. Leningrad, (in Russian)
Izrael, Yu. A. & Tsyban. A. V. ( 1989). Ocean Ecology under

Human Impact. Gidrometeoizdat Publishers, Leningrad,

527 pp. (in Russian)
Kimball, W. H. & Munir, Z. A. ( 1971 ). The corrosion of lead

shot in a simulated waterfowl gizzard. J. Wildl. Manage. 2,

360-365.
Kirso, U., Belyhk, L. & Stom, D. (1983). Kinetics of

benzo(a)pyrene oxidation in aqueous solution. Oxid. Comm.

4(1-4), 35-44.
Kirso, U. & Gubergrits, M. ( 197 1 ). Kinetics of oxidation of

3,4-benzpyrene. phenol and 5 methylresorcinol activated by

UV-irradiation. Proceedings of Estonian Academy of

Science, Chemistry, Geology 20(2). 134-139. (in Russian)
Kirso, U., Paalme, L., Voll. M., Urbas, E. & Irha, N. ( 1990).

Accumulation of carcinogenic hydrocarbons at the sediment

water interface. Marine Chemistry 30. 337-341 .
Kirso. U., Stom, D.. Belykh. L. & Irha, N. (1988).

Transformation of Carcinogenic and Toxic Substances in

the Hydrosphere. Valgus Publishers, Tallinn, (in Russian)
Laist. D. ( 1 987 ). Overview of the biological effects of lost and

discarded plastic debris in the marine environment. Mar.

Pollut. Bull. 18.319-326.
Lee, M. L., Milos. V., Novotny. K. & Bartle. D. (1987).

Analytical Chemistry of Polycyclic Aromatic Compounds.

Academic Press, New York. 5. 441—449.

203

Mamotov, G. & Wehry, E. (1988). Chemical reactivity of
polycyclic organic compounds absorbed on coal fly ash and
related solid surface final report. Energy Res. Abstr.

Medinets, V. I.. Solov'ev, V. G. & Glebov, B. V. (1992).
Cesium- 137 in the seawaters. In Results of the Third Joint
US-USSR Bering & Chukchi Seas Expedition (BERPAC),
Summer 1988 (P. A. Nagel. ed.). US Fish and Wildlife
Service, Washington, D.C.

Mill, T., Mabey, W. K., Lan, B. Y. & Baraze. A. (1981).
Photolysis of polycyclic aromatic hydrocarbons in water.
Chemosphere 10(11/12). 1281-1290.

Mopper. K. & Zhou, X. ( 1989). Photochemical production of
hydroxy radicals at the sea surface: implications of oceanic
carbon cycling. 32nd IUPAC Congress, Stockholm. Bookof
Abstr. p. 2.

Moms, R. J. (1980). Plastic debris in the surface waters of the
South Atlantic. Mar. Pollut. Bull. 11, 164-166.

Nagaya. Y. & Nakamura, K. (1985). A study of certain
radionuclides deposited in the deep waters of the Pacific
through radioactive precipitation. In Comprehensive Global
Monitoring of the World Ocean. Proceedings of the First
International Symposium. 2-10 October 1983. Vol. 2
( Yu. A. Izrael, ed. ). pp. 429—44 1 . Gidrometeoizdat Publishers,
Leningrad, (in Russian)

Niassat, P., Echardt, I. & Ottenwalder. I. (1968). Presence of
benzo3, 4pyrene in the waters of the Clipperton Atoll lagoon.
Comp. Rend. Acad. Sci. Paris. Series D. 267, 1771-1774.
(in French)

Paalme. L. & Gubergrits, M. (1976). Kinetics of separate and
co-photodegradation of benzo(a)pyrene and 3-
methylchloranthene. Proceedings of Estonian Academy of
Science. Chemistry, Geology 25(4), 271-275. (in Russian)

Paalme. L., Uibonuu, H.. Gubergrits, M. & Jacquignon. P.
(19831. The initiated oxidation of polycyclic arenes in
aqueous solution. Oxid. Conun. 4( 1-4). 27-34.

Pruter, A. T. ( 1987). Sources, quantities and distribution of
persistent plastics in the marine environment. Mar. Pollut.
Bull. 18.305-310.

Rabek, J. F. ( 1982). Experimental Methods in Photochemistry
and Photophysics. Vol. 1, Mir Publishers. Moscow. (Russian
translation)

Ranby. B. & Rabek. J. F. (1975). Photode gradation,
Photooxidation and Photo-stabilization of Polymers. Mir
Publishers. Moscow. (Russian translation)

Ryan. P. G. ( 1 987). The effects of ingested plastic on seabirds:
Correlations between plastic load and body condition.
Environ. Pollut. 46, 1 19 125.

Ryan, P. G. ( 1988a). Plastic ingestion and PCB's in seabirds:

Is there a relationship? Mar. Pollut. Bull. 19. 174-176.
Ryan. P. G. (1988b). Effects of ingested plastic on seabird
feeding: Evidence from chickens. Mar. Pollut. Bull. 19.
1 25- 128.
Shaw, D. G. ( 1977). Pelagic tar and plastic in the Gulf of
Alaska and Bering Sea: ll)75. Sci. Total Environ. 8. 13-20.
Shaw, D. G. & Mapes. G. A. ( 1979). Surface circulation and
the distribution of pelagic tar and plastic. Mar. Pollut. Bull.
10. 160-162.

Shilina. L. I. (1982). In Integrated Global Monitoring of
Natural Environmental Pollution, pp. 238-242.
Gidrometeoizdat Publishers, Leningrad, (in Russiani

Shpolsky. E. V.. Il'in, A. A. etal. ( 1952). Fluorescence spectra
of coronene in frozen solutions. Dokladny AN SSSR 1 .
935-939. (in Russian)

Tanabe, S., Tatsukawa. R.. Kawano. M. & Hidaka. H. ( 1982).
Global distribution and atmospheric input of chlorinated
hydrocarbons in the western Pacific, eastern Indiana and
Antarctic Oceans. J. Oceanogr. Soc. Jpn.. Vol. 38.

Tanabe, S. (1985). Distribution, behavior and fate of PCB's in
the marine environment. J. Oceanogr. Soc. Jpn.. Vol. 41.

Tsyban, A. V. (1975). Bacterioneuston and problems of
degradation in surface films of organic substances released
into the sea. Prog. Water Technol. 7(3/4). 792-799.

Tsyban, A. V., Khesina, A. Ya., Ventzel, M. V., Volodkovich,
Yu. L. & Korsak. M. N. el al. (1985a). Occurrence and
microbial degradation of the carcinogenic hydrocarbon.
BaP. in open waters of the Baltic Sea. In Studies of the Baltic
Sea Ecosystem, pp. 218-234. Vol. 2. Gidrometeoizdat
Publishers. Leningrad, (in Russian)

Tsyban. A. V., Panov. G. V.. Ventzel, M. V.. Volodkovich.
Yu. L. & Korsak, M. N. ( 1986). Long-term environmental
studies on impacted and background regions. In Integrated
Global Monitoring of the Status of the Biosphere.
pp. 45-60. Gidrometeoizdat Publishers, Leningrad, (in
Russian)

Tsyban, A. V.. Volodkovich. Yu. L. & Belyaeva. O. L. (1987).
A study of the circulation of benzo(a)pyrene in the Bering
Sea. In Comprehensive Analysis of the Bering Sea
Ecosystem, pp. 218-232. Gidrometeoizdat Publishers.
Leningrad, (in Russian)

Tsyban. A. V., Volodkovich. Yu. L., Panov, G. V.. Khesina.
A. Ya. & Yemakov. Ye. A. (1985b). Distribution and
microbial oxidation of carcinogenic hydrocarbons in some
regions of the World Ocean, with benzol a Ipyrene as the
model compound. In Environmental Consequences of 'Ocean
Pollution, pp. 88-112. Gidrometeoizdat Publishers.
Leningrad, (in Russian)

Vakulovskiy. S. M. (ed.) (1986). Methodological
Recommendations for the Determination of Radioactive
Contamination of Waters. Gidrometeoizdat Publishers.
Moscow. 78 pp. (in Russian)
Van Franeker, J. A. (1985). Plastic ingestion m the North
Atlantic fulmar. Mar. Pollut. Bull. 16. 367-369.

Winston, J. E. ( 1982). Drift plastic — An expanding niche for

a marine invertebrate' Mar. Pollut. Bull. 10. 348-351.
Wolfe. D. A. ( 1 987 ). Persistent plastics and debris in the ocean:
An international problem of ocean disposal. Mar. Pollut.
Bull. 18. 303-305.
Wong. C.S.. Green. D.R.&Cretney. W. J. (1 974 (.Quantitative
tar and plastic waste distributions in the Pacific Ocean.
Nature 247. 30-32.
Zafiriou, O. C, Dister, B.. Moffett. J. M.. Micinski, E. &
Blough, N. V. (1989). Photochemical production of free
radicals in the marine environment. 3 2nd IUPAC Congress.
Stockholm. Book ofAbstr. p. 201.

204

Chapter 3:

BIOLOGICAL INVESTIGATIONS IN
THE CENTRAL PACIFIC

Editors:

CAMERON B. KEPLER &
MIKHAIL N. KORSAK

3.1 A Description of Bacterioplankton

VASSILIY M. KUDRYATSEV, VLADIMIR O. MAMAEV. and NADEZHDA V. SUKHANOVA

Institute of Global Climate and Ecology, State Committee for Hydrometeorology and Academy of Sciences, Moscow, USSR

Introduction

An assessment of the role played by bacteria in
biodegradation processes occurring in the World Ocean requires
data concerning bacterial population counts, bacterial population
distributions, and a number of other functional characteristics.

Studies carried out in the course of the First Joint
US-USSR Central Pacific Expedition in the equatorial Pacific
and the South China Sea produced new findings characterizing
the present state of microbiocoenoses in this part of the World
Ocean.

Materials and Methods

The microbiological studies described below were
conducted using methods set out in several handbooks
(Romanenko& Kuznetsov, 1974;Tsyban, 1 980; Tsyban etal..
1988). Analyses of total counts, biomass, bacterioplankton
production, indicator group distribution, and degradation
process rates were performed at 1 8 stations. The samples were
taken from 6 to 19 depths using 5-1 Niskin bottles. To allow for
overall bacterial counts, samples of 20-50 ml were passed
through "Synpor 8" membrane filters with pore size 0.23 urn.
The filters were desalinated onto filter paper moistened with
distilled water, dried, and dyed with a 5% carbolic erythrosine
solution. The bacteria deposited on the filters were counted by
direct oil-immersion microscopy (xl350 magnification. 20
visual fields).

The average bacteria volume was assessed by measuring
bacterial cell size with the aid of an ocular ruler. The mean
bacterial cell volume was found to be 0.3 urn'.

The daily bacterial biomass production rate was calculated
on the basis of CO: assimilation in darkness. The latter was
ascertained using a radiocarbon technique (Romanenko. 1964;
Sorokin. 1971a). The determinations were made in
1 00- 1 20-ml jars. The radioactivity of the working solution of
carbon-labeled sodium carbonate introduced into the sample-
containing jars was 18 x 106 counts/min. The samples were
incubated in darkness for 24 h at the temperature of the water
where the sample was taken. Once the incubation was
completed, the samples were fixed with a 40% formaldehyde
solution, then passed through "Synpor-7" filters (pore size
0.35 (tm). The radioactivity of the bacteria deposited on the
filters was measured using an '•Intertechnique'" liquid
scintillation counter. The scintillation cocktail was ZhS-8.
The CO, assimilation rate in darkness (Tass) for the bacterial
plankton was calculated by means of the formula

T — '* x Qart

± ass '

R xt

where r the radioactivity of the bacterial cells on

the filter (counts/min);
the hydrocarbonate content of seawater
(mg/1 ) determined by direct titration of
0. IN HC1 in the presence of methyl
red;
— the radioactivity of the isotope Nal4CO,
used in the experiment (counts/min);
t the incubation time (h).

The bacterial biomass production was obtained by
calculation, setting Ph = TMS x 16.6. Bacterial plankton
respiration was determined by applying the formula

C„

k

D = 1J1,

where D

— the amount of oxygen expended on decom-
position (mg 1 ' d ');
Tass — rate ofCO: assimilation in darkness (|ig of

C perl"1 d '); and

the coefficient of 7 is the ratio of oxygen uptake to

CO, assimilation.

The studies were carried out in the tropical and equatorial

portions of the Pacific Ocean, as well as in the South China Sea.

The locations of the stations are indicated on the Frontispiece.

Results

The equatorial Pacific is characterized by upwelling. The
latter occurs mostly along the boundaries of west-to-east zonal
flows and alternates with surface-water downwelling zones.
The intrusion of deep waters rich in biogenic elements into the
euphotic layer determines the way in which biocoenoses develop
(i.e., their spatial and trophic structures, productivity, and other
functional characteristics) (Vinogradov, 1978).

Some temperature stratification was evident during the
period of our studies in the 0-8°N, 1 60- 1 80° E rectangle of the
equatorial Pacific. The water temperature and dissolved oxygen
content remained virtually constant down to a depth of 150 m.
The parameters in question declined rapidly below this level.
however, which necessarily affected the formation and
distribution of microbiocoenoses. Results of analyses
(Figs. 1-3) indicate that bacterioplankton counts and biomass
in the 0-100-m layer varied within fairly broad limits. The
highest bacterial population density was noted at Station 1 16,
where the average total count and biomass were
591 x 103 cells/ml ' and 13.29 ug C V d'1, respectively. The
lowest bacterioplankton concentration was noted at
Station 118, where the average total bacteria count was
199 x 103 cells/ml1 and the biomass 4.48 ug C V d '. The
average values of the total bacteria count and biomass in

207

the 0-100-m layer of the equatorial Pacific were
360 x 103 cells/ml1 and 8.10 ug C 1 ' d ', respectively. Such a
level of development of bacterioplankton and its biomass is
typical of the bacterial population density in the euphotic zone
with background upwelling (Sorokin, 1978). According to
data reported in the literature, the principal bacterioplankton
count and biomass maxima in stratified tropical waters are
usually situated at the upper boundary of the thermocline. The
formation of the microbiocoenosis in this zone is attributable to
the arrival of biogenic substances from lower layers, to the
stable existence of the phytoplankton population in the euphotic
layer, and to the stable existence of the phytoplankton population
in the euphotic layer. This in turn ensures that the
bacterioplankton receives a steady and readily assimilable
supply of organic matter (Sorokin, 1971b, 1982).

According to our observations, the vertical distribution of
microflora (Figs. 1-3) included several regularly observed
maxima occasioned by the delivery of labile organic matter to
the zone of maximum phytoplankton synthesis. These lay at
15-25 m for Stations 115-117, at the upper boundary of the
thermocline (100-250 m). and at a depth of 1.000-1.500 m
(Stations 118-120) (i.e., at the boundary of mixing of the
Antarctic waters).

The principal bacterioplankton concentration maximum
lay at a depth of 100 m. The microflora counts and biomass
were, on the average, 1 .3 times the corresponding values for the
euphotic zone. The level of microbiocoenosis development in
deeper waters (at depths of 1,000-2,000 m) was lower. The
total bacterioplankton counts and biomass value here turned
out to be lower than at the upper boundary of the thermocline
by a factor of 1.5. The constraints limiting microfloral
development in these deeper waters were very probably the
lower temperature and high pressure.

There was a clearcut tendency for the bacterial population
density in the 0.5-250-m layer to decrease in the east-to-west
direction. Thus, the bacterioplankton counts and biomass at
Stations 115 and 116, situated in the eastern part of the
equatorial zone, were, on the average, twice as high as at
Station 1 20. lying to the west of them. This observation was in
keeping with the declining intensity of upwelling from east to
west (Vinogradov. 1978).

Thus, total bacterioplankton counts and biomass in the
waters of the equatorial Pacific, which have suffered less of an
impact from human activities that other parts of the Ocean, are
fully comparable with those observed for oligotrophic and
mesotrophic waters.

As in the quantitative parameters, the functional
characteristics undergo considerable change as one moves
from east to west. Data obtained in the course of the present
study made possible some quantitative assessments of certain
functional characteristics of microbiocoenoses. such as
microfloral activity, bacterial biomass production rate, and the
rate of degradation of organic matter in the equatorial Pacific.

Analysis of the results (Figs. 1—1) shows that microfloral
activity in the 0- 1 00-m layer in the equatorial zone varied over
a broad range and had a tendency to increase in the east-to-west
direction despite the gradual diminution of bacterial population

1 T ■ ' ■ ' ' '

10

Li_——

."* " '- .

\

25

• 3

♦ 2

'J

.. — * — *

-x— J— « "■

£ 45

*

♦

Q

"*^i *

100

m**

■ >

* •

250

T— -—"*'"*

' ?

• •V"^

1000

■f

1500

< ,

•

~~'~"~i* .

Station 115

0

2

4

>:

e it

0

025

'

0 75 1

0.25 0 5 075

X-..'

,y

Figs. 1-3. Vertical distribution of: ( 1 ) water temperature in °C; (2) dissolved
oxygen in mg/l (3) total bacteria count in millions of cells/ml: and
(4) CO. assimilation by bacteria in darkness. ingCI ' d '. All data
refer to the equatorial Pacific.

density. Thus, the values for the amount of CO, assimilated by
bacteria per day under conditions of darkness averaged 1 .5-2.0
times higher for Stations 1 19 and 120 than for the more easterly
Stations 115-116. Both bacterial biomass production and
organic matter degradation also increased. Whereas the bacterial
production values in the 0-100-m layer at Stations 1 15 and
1 Id averaged 15.2 and 16.6 (.tg C 1 ' d"\ the corresponding
values at Stations I 19 and I 20 were 25.5 and 32.5 ng C 1 ' d '.
The respiratory uptake of oxygen by microorganisms in the
western area was virtually double. The rate of organic mailer
degradation increased from 78.7 HgC l'1 d'1 at Station 1 16 to

208

in 0-100 m layer

Station 115 Station 1 16 Siaiton H7 Siaiion 1 18 SlMion 1 19 StMion 120

.U

n h _Q

Suiiiin 1 15 Sluiion I Ift Station 1 17 Suihon I IH Suimn 1 19 SlMion 120

in 500-2.000 m layer

D

Sluiion I IK Suimn ll>)

I I BjLlcnjI priwluumn
Klug C/l/duy [v]] Orgunn nulla dcgrjduiii]

Fig. 4. Bacterial production and organic-matter degradation in the equatorial
Pacific.

104.4 jag C 1 ' d ' at Station 1 19. The average rate of bacterial
organic matter degradation in the0-100-m layer in this part of
the Pacific equaled 75.5 ug C 1 ' d '. The P/B coefficient for
microflora in the 0-100-m layer averaged 2.9, which was
somewhat higher than the value previously obtained by Sorokin
(1973). The average value of the P/D coefficient was 0.31.

Maximum microfloral activity in the 0-100-m layer
occurred at a depth of 25 m. The rate of assimilation in darkness
here attained a level typical of mesotrophic waters, averaging
1.94 ugC I'd'. The respiratory oxygen uptake by microflora
also showed peak values, averaging 0.28 mg/1 — a finding that
is in good agreement with that of Sorokin ( 1973). The rate of
bacterial degradation of organic matter in the euphotic zone
reached maximum levels on the order of 103.9 ug C 1 ' d '
(Fig. 4).

The activity of microflora in the 100-250-m layer was
somewhat lower compared with the euphotic zone. The rate of
CO, assimilation in darkness averaged 1.10 |ig C 1' d'.
Bacterial biomass production was lower than in the top layer by
an average factor of 1.2. The bacterial degradation rate
averaged 58.0 Ug C 1 ' d '. The P/B coefficient was 2.2, the
P/D coefficient, 0.3 1 .

Bacterial activity diminished sharply at depths exceeding
250 m. The rate of bacterial biomass production in the
500-2, 000-m layer was. on the average, almost five
times lower compared with the upper layers of the water
column, ranging from 2.5 to 5.5 |lg C 1 ' d ' and averaging
4.2 ug C 1 ' d "'. The bacterial degradation rate decreased more
than fivefold compared with that of the euphotic zone. This
decreased rate of degradation was clearly attributable to lower
temperatures. The P/B coefficient in the 500-2,000-m layer
was 0.7. The water column at 500-2,000 m was found to
include two layers of heightened activity, one at 1,000 m, the
other at 2,000 m. However, bacterial production and degradation
rates in these maximum-activity layers did not attain the
average values characteristic of the top layer.

The next transect lay along 10°N in the western Pacific
(Frontispiece). The waters in this area exhibited lower
productivity as compared with those of the equatorial portion
of the central Pacific. The level of development of

microbiocoenoses at most of the stations corresponded to the
upper limit of productivity for oligotrophic waters. The bacterial
population density in the 0-100-m layer ranged from 129 to
545 x 101 cells/ml1 (Figs. 5-7). The highest bacterioplankton
concentration occurred at Station 123, where the average count
and biomass reached 387 x 10-' cells/ml ' and 8.7 ug C 1 ' d '.
The highest microfloral activity was observed at Station 121,
where the rate of CO, assimilation by bacterioplankton averaged
1.76 Ug C 1 ' d "', which corresponds to the upper limit for
oligotrophic waters. The lowest microfloral activity was noted
at Station 122. The daily rate of CO, assimilation in darkness
here was 3.4 times less than at the preceding station.

Bacterial biomass production in the 0- 1 00-m layer ranged
from 2.1 to 52 ug C 1 ' d ' (Fig. 8). The oxygen uptake due to
bacterioplankton respiration averaged 0.12 mg/1. The rate
of organic matter degradation by bacteria lay in the
27-94-ug C 1' d ' range, with an average value of
47 ug C 1 ' d '. The P/B coefficient was 2.3.

The portion of the water column lying below the
thermocline (150-500 m) exhibited relatively low bacterial
population levels. The total bacteria counts here ranged from
109 to 492 x 10' cells/ml', the biomass from 3.9 to
7.7 ug C 1 ' d "'. The mean bacteria count and biomass for the
150-500-m layer turned out to be 237 x 10- cells/ml'
and 5.3 Ug C 1' d ', respectively. The microfloral activity
was somewhat lower than in the supernatant 100 m of the
water column. The rate of CO, assimilation by bacteria in
darkness ranged from 0.12 to 1.36 ug C 1' d\ averaging
0.62 ug C 1 ' d ' . The most intense microfloral activity occurred
at Station 125, where the mean CO, assimilation rate reached
0.78 ug C 1 ' d ' . The lowest rate of 0.47 ug C 1 ' d ' corresponded
to Station 122.

The bacterial biomass production rate turned out to be 1 .5
times lower than in the 0-100-m layer. Oxygen uptake due to
bacterioplankton respiration in the 150-500-m layer averaged
0.08 mg/1, indicating a relatively low rate of degradation. The
degradation rate OB ranged from 25 to 42 ug C 1' d '. with an
average value of 33 ug C 1 ' d '.

A trend towards increased bacterial population densities
was noted in the deeper ( 1 .000-2,000 m) portion of the water
column. For example, the total bacteria counts at 2,000 m
for Stations 1 2 1 and 1 22 were much higher than further up
the water column. Some increase in bacterioplankton
concentrations was likewise noted at 1,500 m at Stations 123
and 125. The average total bacterial count and biomass for
the 1,000-2,000 m layer was 281 x 10' cells/ml' and
6.3 ug C I ' d '. respectively. These values were quite closely
comparable with the data for the 0-100-m layer.

The results showed that microfloral activity in the
deeper portions of the water column was suppressed by
low temperatures and high pressures. The rates of CO,
assimilation by bacteria in darkness for the 1 ,000-2,000-m
layer ranged from 0.11 to 0.42 ug C 1' d ', averaging
0.22 ug C 1 ' d '. Bacterial biomass production was, on the
average. 2.8 times lower than in the 150-500-m layer.
Oxygen uptake due to bacterioplankton respiration averaged
0.03 mg/1. The rate of degradation of organic matter was
minimal, the average value for the layer being 12 ug C 1 ' d '.
The P/B coefficient was 0.5.

209

0

2

.

6

8

11

)

025

15

0 75

1(

p

10

ZS

40

5

0

2

4

6

6 10

0

025

0 5

0 75 t

Q

0 25

05

C

"5

1 0

' 1

10

3

k

t 2

4 «

xl

25

•

t

»

45

E

•

4

••

t
n

•

*

//

100

*

*

250

*'

^

500

i D00

»

•

• /

Slanon 12:

2000

* I

.

i

24 30

2

4

6

6 U

0

0 25

05

0 7S 1

0 25 0 5

1500
2000

0 0 25

0 5 0 75 1

J !

*"■ i

•

3 .

>

/ i

.. .

'

/

.

/

■

^ * f

,'• ") )

. .■-"

A '*•

Slatwn 122

/ X

t ' '

i • ;

0 25 OS

0 25 0 5

Figs 5-7. Vertical distribution of: 1 1 1 water temperature in C';(2)dissolved
oxygen in mg/1 (3) total bacteria count in millions of cells/ml; and
(4) CO assimilation h> bacteria in darkness, in g C 1' d'1. All
daia applj to the western Pacific.

The bacterioplankton distribution over the water column
in the western Pacific exhibited some temperature stratification,
with several bacterial population and acti\ it) maxima present
(Figs. 5-7). Three bacterioplankton concentration peaks were
quite clearly in evidence: the first one lay below the /one of
maximum phytoplanklon synthesis at a depth of 45 m. the
second above the upper boundary of the thermoeline at a depth
of 100 ill. and the third at 1.500 m. Elevated microfloral
acti\ it} was noted in the euphotic /one (10 15 ml. above the

I nl ml nl

mil iimi m layei

Station 121 Station 122 Station 123 Station 124 Station 125 Station 126

JL nl n I rn I

Station 121 Station 122 Station 123 Station 124 Station 1 25

■ to in layei

in l.000-2.(HMi,n layei

Station 121 Station 122 Station 123 Station 124 Station 125

1 I Bacterial production

D.0,

Fig. 8 Bacterial production and organic-matter degradation in the western
Pacific.

thermoeline. and at a temperature drop boundary. This behax ior
of bacterioplankton distribution and activity over the water
column was closely related to water dynamics and was typical
of the oligotrophic waters of the tropical oceans.

The next series of studies was conducted in the South
China Sea. whose waters are more polluted than those of the
central Pacific. Analyses of total biomass. bacterioplankton
production, and bacterial degradation were performed at five
stations. The results of the measurements are presented
in Figs. 9-11. These show that the bacterioplankton count
in the euphotic layer (0-50 m) ranged from 176 to
61 1 x 10' cells/ml '. The highest bacterial population densities
were noted for Stations 1 30 and 131. situated in coastal waters:
the lowest values were obtained at pelagically situated
Station 127. The bacterioplankton count at certain levels of
the water column attained 611 x 10' cells ml'. The
average bacterioplankton count in this area equaled
354 x 10' cells/ml :.

The bacterioplankton biomass varied over a broad range
of values from 3.16 to 11.81 (.tgCl'd1. Data for the 0-50-m
layer averaged for individual stations yielded values from 6.67
to 9. 1 6 jug C I ' d ' . Mean data for bacterioplankton biomass and
counts for the euphotic /one place the investigated portions of
the South China Sea in the oligotrophic category.

Low bacterioplankton counts notwithstanding, microflora]
activity was relatively high. Thus, the rate of CO. assimilation
by bacteria in darkness ranged from 0.52 to 4.07 (ig C 1 d '.
averaging 1.60 p.g C 1' d ' for all the areas studied.
Bacterioplankton production for individual stations
ranged from 1 2.8 to 39.7 p.g C 1 ' d ' (Fig. 1 1 ). The value for
the entirety of the area investigated was 26.6 |ig C 1 ' d '.

The rates of bacterioplankton respiration in the 0-50-m
layer calculated as averages for particular stations ranged
from 0.11 to 0.34 mg O, 1' d '. The range of bacterial
respiration rates for individual samples ranged from 0.07 to
0.58 mg O, I'd'. The maximum bacterioplankton respiration
rate was observed at Station 1 30. the minimum at Station 1 29.
The average daily oxygen uptake by bacterioplankton in the
area was 0.23 mg/1 ' . which w as in good agreement with values
calculated from experimental data on the respiration of the
plankton community as a whole.

210

0

6

12

16

24 3C

0

2

4

6

e ic

0

0.25

05

0 75 1.

E 10

Q

Q25

45

1

0

6

12

'8 Stattor2427

30

0

2

4

6 8

10

0

025

05

075

1.0

0

2

4

6

8 IC

0

0 25

05

075 1.

"V>

,/■'

/

Figs. 9-10.

Vertical distribution of: ( 1 )watertemperaturein°C;(2)dissolved
oxygen in mg/1: (3) total bacteria count in millions of cells/ml;
and (4) CO, assimilation by bacteria in darkness, in g C 1 ' d '.
All data apply to the South China Sea.

Total bacterial degradation in the water column at individual
stations ranged from 41 to 127 (ig C 1 ' d '. The average for a
whole of the area studied was 86 U. C 1 ' d"1.

The vertical distribution of bacterioplankton over the
water column within the euphotic zone included a clearly
evident peak at a depth of 1 0 m (Figs. 9, 1 0). The mean bacterial
density and biomass in this layer were 450 x 10' cells/ml ' and
10.1 u.g C 1 ' d "', respectively. Both counts and biomass then
decreased with depth.

Microfloral activity varied considerably over the water
column. Two microfloral activity peaks were observed at
several stations: one lay at a depth of 10 m, the other at or near
the sea bottom. Enhanced production-degradation activity
was likewise noted at the corresponding depths. Bacterial
biomass production in the peak-activity layers was 2.5 times
the bacterial production obtained for the entire sea area studied.
The degradation rate OB in the bottom layer was 3.5 times that
noted for the surface layer.

Discussion

Particular portions of the Pacific Ocean exhibited differing
levels of microbiocoenosis development. The equatorial Pacific
and South China Sea areas are closely similar with respect to
bacterioplankton and biomass parameters in their euphotic
zones. On the other hand, the level of development of the
bacterial population in the western Pacific was found to be
much lower. The South China Sea exhibited relatively high
microfloral activity. Values characterizing bacterial production
and organic matter degradation in the sea were similar to those
reported for the mesotrophic waters of the tropical ocean. The
level of production-degradation processes in the western Pacific
was almost half of that in both the equatorial ocean and the
South China Sea and corresponded to values characteristic of
oligotrophic waters.

n

n

in 0-55 m layer

Station 126 Station 127

Station 128

Station 129 Station 130

Station 131

Scale:

I I Bacterial production
10 Ug c/l/day EH Organic matter degradation

Fig. 1 1 . Bacterial production and organic-matter degradation in the South China Sea.

211

3.2 A Study of Primary Phytoplankton
Production

MIKHAEL N. KORSAK

Institute of Global Climate and Ecology, State Committee for Hydrometeorology and Academy of Sciences, Moscow, USSR

Introduction

The study of the rate of formation of new organic matter
by phytoplankton photosynthesis in the tropical Pacific reported
in the present paper was carried out during the First Joint
US-USSR Central Pacific Expedition aboard the R/V Akademik
Korolev in 1988. The ocean areas investigated included
portions of the central and western Pacific, which had received
little previous attention. Primary production in the area of the
first transect, beginning near Caroline Atoll and ending at
Tarawa Island (Republic of Kiribati), ranges from values
characteristic of oligotrophic parts of the ocean
(100 mg C/nr/day and lower) to values corresponding in
mesotrophic areas of the Pacific (25 mg c/nr/day) (Sorokin,
1976).

The central portion of the tropical Pacific (stations of the
second transect) may be characterized as an oligotrophic
productivity zone where a major role in primary production
enhancement is played by synoptic phenomena such as cyclones,
tornadoes, waterspouts, etcetera (Sorokin, 1976). Situated in
the open portion of the northern tropical Pacific (i.e., within the
northern tradewind zone), the second transect studied showed
a primary rate of organic matter production by phytoplankton
of about 1 00 mgC/m7day (Sorokin, 1976). The relatively low
rales of primary production in this part of the Pacific are
attributable largely to the deficiency of biogenic elements in
the photosynthetic layer. These low concentrations are in turn
due to anticyclonic circulation, which produces downwelling
of nitrogen- and phosphorus-poor surface waters. It is the
resulting low nitrogen and phosphorus levels in the
photosynthetic layer that limit photosynthesis rates in the
phytoplankton community.

Despite the absence of significant seasonal variations in
illumination and water temperature in the tropical ocean,
considerable seasonal changes of photosynthesis rates have
been reported in spring and autumn (Sorokin, 1976).

Materials and Methods

Studies al Stations 1 14-120 in the central tropical Pacific
were conducted from 27 September to 7 October 1988. The
work al Stations 121-126 along the Marianas transect was
performed from L 6 October to 2] October. Primary production
was determined by means of a radiocarbon version of the "jars
method" proposed by Sorokin (Sorokin el al., 1983). Work al
each station included measurements of photosynthesis in a
surface-water sample (Cps) as well as determinations of
photosynthesis in the water layer as a function of phytoplankton
distribution over (he water column (the coefficients K,).

The light curves (the coefficients K,) were determined at one
station in each transect. The sample incubation was usually 8-
10 h, beginning in the morning. The radioactivity of filters with
l4C-labeled phytoplankton and of the working NaHl4CO,
solutions was measured using a Nuclear Chicago "Mark 2"
scintillation counter. Sample radioactivities were counted
using liquid scintillator cocktails of previously described
composition (Sorokin, 1976; Sorokin et al., 1983).

Primary production was calculated using the standard
formula, with a factor of 1.5 to correct for l4C loss due to
phytoplankton sample filtering (Sorokin, 1976). All
determinations of primary production were carried out in
triplicate. The extent of the photosynthesis zone was taken to
equal the white-disk transparency multiplied by three. Samples
for determining phytoplankton production were taken using
5-1 Niskin bottles at depths of 0.5: 10; 15; 25; 45; 70, and
100m.

Results

The vertical structure of phytoplankton communities in
high-transparency tropical ocean waters is characterized by
several phytoplankton growth peaks or maxima in the euphoric
zone, whose depth sometimes exceeds 100 m. Two layers with
elevated phytoplankton concentrations are usually in evidence.
The first of these occurs at a depth of 1 0-30 m and is associated
with a photosynthesis-optimal light zone; the second lies at
depths of 70-90 m and is related to heightened biogenic
element levels in (he vicinity of the pycnocline (Sorokin.
1 976). As is evident from Table 1 , the depth of the photosynthesis
layer at all stations of both transects was usualh slightly in
excess of 100 m. As a rule, only a single photosynthesis peak
was observed (the one in the 15-25-m depth range), since the
pycnocline lay below the photosynthesis layer boundary (see
Table 1 ). However, Stations 1 22 and 1 26 did exhibit a second
relatively small primary production peak at 70 m (Table 1).
Primary production values for the topmost levels of the water
column were usually markedly lower than at depths of
10-15 m. which was probably due to photic inhibition of
photosynthesis by the high-intensity incident light. Stratification
of water masses o\ er the w ater column had no significant effect
on primary organic-matter production by phytoplankton.
inasmuch as the top 100 m of the water column was
homothermal. Thus, the water temperature in the top 100 mof
the water column at stations of both transects varied within just
1-2°C. Salinity in the same laser varied within the same
narrow limits, so that the primary production level at various
depths depended largely upon light intensity, amount of
phytoplankton present, and biogenic-element availability.

212

The local primary production peaks occurring at a depth of
70 m at Stations 1 22 and 1 26 were probably due to the elevated
levels of biogenic elements present there.

The values obtained for primary production of organic
matter by phytoplankton for the whole of the photosynthetic
zone ranged from 70 to 140 mg C/nr/day (see Table 1 ). The
maximum phytoplankton production ( 140 mg C/nr/day) was
observed at Station 116; minimum production
( 70 mg C/nr/day) occurred at Station 120 (Table 1). The total
average levels of phytoplankton productivity in this part of the
Pacific during the period of our studies was in line with the
higher range of primary phytoplankton production values
typical of oligotrophic zones of the World Ocean (i.e., about
101 mg C/nr/day).

For stations of the second transect, situated to the northwest
of the first at about 8°N, the range of variation of primary
production rates (90-338 mg C/nr/day) exceeded the range of
values for Stations 1 14-120 (Table 1). The mean primary
production at Stations 121-126 corresponded to the lower
range of productivity rates for oligomesotrophic zones of the
World Ocean (i.e.. about 172 mg C/nr7day) (Table 1 ).

Comparing the primary organic-matter production values
measured during the 1988 expedition with those obtained

during the 1984 expedition, which covered roughly the same
parts of the Pacific, we note that both primary productivity and
the range of variation of primary production rates were greater
in 1984. The rates of primary production of organic matter by
phytoplankton over the Marianas transect in late July-early
August 1984 ranged from 1 00 mg C/nr/day to 1.16gC/nr7day,
the mean value being about 400 mg C/nr/day . The highest
primary production values for phytoplankton occurred at the
westernmost stations of the transect. The rather high primary
production rates in this part of the ocean in 1 984 were probably
due to the seasonal arrival of waters rich in biogenic elements
that originated in the equatorial divergence area to the south of
the study area.

Phytoplankton productivity rates for the Marianas transect
in 1988 were somewhat lower than those recorded in 1984.
The differences can probably be attributed to fluctuations in
the arrival of waters from the equatorial divergence region.
On the whole, the level of primary production of organic
matter in the central tropical Pacific was in keeping with
expectations based on previous studies ( Sorokin, 1 976 ) and on
the findings of the 1 984 expedition, corresponding more or less
to the level associated with oligomesotrophic zones of the
World Ocean.

TABLE 1

Primary production (mg C/nr/day) at different stations
in the central Pacific.

Depth

Station Number

(m) 114

1 15

1 16

119

120

121 122

123

124

125

126

0.5 2.9

1.7

2.3

1.2

0.42

3.5 1 .9

2.9

0.40

1.2

1.4

10 1.7

2.3

2.5

1.8

1.2

5.2 5.6

2.7

0.48

0.53

3.0

15 1.5

2.6

5.6

3.6

1.6

7.2 10.5

6.0

2.7

2.7

4.5

25 2.8

1.1

1.5

2.1

1.3

1.7 7.8

2.0

3.4

3.7

2.7

45 0.7

0.6

1.0

1.6

1.0

0.74 1.3

0.97

0.45

1.3

0.6

70 0.44

0.32

0.6

0.58

0.16

0.74 2.3

0.3

0.3

0.46

0.7

100 0.06

0.02

0.08

0.14

0.02

0.21 0.13

0.01

0.00

0.05

0.0

* P,„2 80.5

84.2

14(1

132

70

177 338

143

89

125

16'1

* P,„: is the primary produi

:tion for the

photosynthesis 1

ayer down to 100 m ( m

g C/m

7day).

3.3 Mesozooplankton

TATIANA A. PAVLOVA and AUDREY S. KUL1KOV

Institute of Global Climate and Ecology, State Committee for Hydrometeorology and Academy of Sciences, Moscow, USSR

Materials and Methods

The materials for this work were collected in the eastern
equatorial region of the Pacific Ocean at seven stations ( Stations
1 14- 120) whose coordinates ranged from 10°Sto 150°W,and
0°S to 178°W (Fig. 1). The zooplankton was collected in
daylight with a large Juday Net. with a mesh of 168 urn and a

throat diameter of 37 cm. from the 0-50. 50-100, and
100-200-m levels. The samples were fixed with a 40%
solution of formaldehyde and were processed by the standard
methods (Korshenko, 1988: Tsyban et al„ 1988) under an
MBS-9 binocular microscope in a Bogorov chamber. The
sample was concentrated and poured into a Petri dish. While
it was being examined under the binocular microscope, the

213

180 W

150'W

a)

Boundary of
0 Tradewind
Current

180 W 170'

150'

5'

x|x convergence

o

b)

180 W 170'

Fig. I. Character of currents in the region studied: (a) circulation of waters
at the surface; (h) circulation of waters at a depth of 100 m; and (c)
circulation of waters at a depth of 2(H) m.

largest animals ( over 2 mm long ) were washed out and classified
into systematic groups using a biolam R-7 microscope. The
residue of the sample was diluted 1 0-20-fold, depending on the
concentration, and 2-3 portions of 10-20 ml each were taken
with a 5-ml plunger pipet for counting. On the basis of this data,
the mesozooplankton numbers in 1 m1 and under 1 m: were
calculated. A micrometer eyepiece was used to determine the
body length of representatives of each species. On the basis of
these values and Chislenko's nomograms (Chislenko, 1968;
Vinogradov & Shushkina, 1987), the biomass of each species
was determined, and the total biomass of mesozooplankton in
I in' and under 1 m was calculated. Twenty samples were
analyzed.

Results and Discussion

li is well known (Korshenko, 1988) that the characteristics
of the distribution of zooplankton depend, to a considerable
degree, on the hydrological structure of the water masses.
Figure 1 shows a diagram of the currents in the studied region
on the surface and at depths of 1 00 m and 200 m. The dynamics
of the water masses at the easternmost station near Caroline

Atoll and at adjacent Station 1 14 are low; no major currents
pass through this area. The depths of 100 m and 200 m at
stations in the central portion of the region are characterized by
a water mass transport to the west. The w ater masses of the
western portion of the section (Stations 119 and 120) are
affected by the surface South Tradew ind Current. In addition,
at the westernmost point of the region (Station 120). an
anticyclonic equatorial current is observed (Gorshkov, 1974).

The heterogeneity of the regional water structures suggests
a nonuniform horizontal and vertical distribution of the
mesozooplankton as well as its qualitative diversity. As is
evident from Table 1 . the composition of the mesozooplankton
differed qualitatively at different points of the section. A total
of about 1 80 forms of mesozooplankton were determined as a
result of the analysis of the samples. The average number of
species for a station ranged from 75 to 139 and. for an
individual level, from 39 to 84 (i.e., it differed by a factor of
two) (Table 1; Fig. 2). The smallest number of species was
found at the station near Caroline Atoll: at the 0-50-m and
50-100-m levels, there were 39 and 56 species, respectively.
The number of species increased from south to north toward the
equator, reaching a maximum of 1 39 at Station 17. The vertical
structure of species composition also varied. In the majority of
the cases, the number of species increased appreciably with
depth, and it remained practically unchanged only at the
equator (Fig. 2a).

The species observed in the waters of the region studied
were encountered at different frequencies. Four gradations of
species occurrence were distinguished: rare, under 1 sp/m';

Number of species

2

3 30

40

s|l

60 70 80

90 100

Depih. m.

-.1 l!" Ins
V \

l\ll5 \

HI?

50-

!\ \ \ /

M4^\

sj?« b fl

100

1 I

\l

200 -

^

Number of species
140

120 119 118 117 116 115 114 Sun.)

I ig 2. Number ol meso/ooplankton species: (a) at an individual level; and
Ibl average for the station.

low, from I to 10 sp/m1; normal, from k) to 100 sp/m'; and
massive , > 100 sp/m3 (Table 1 ). As a rule, determined forms
of mesozooplankton were found at the same frequency.
However, it is evident from Table 1 that, in many cases, the
frequency of species occurrence increased in the western
portion of the section. The number of massive species at an
individual level ranged from 4 to 14. The maximum number of
massive species ( 1 3-14) was recorded in the western portion of
the section at the equator; the minimum number was recorded
in the eastern portion. We note that the populations of only 2 1
species of zooplankton reached numbers above 100 sp/m1; 18
consisted of the order Copepoda. Among other taxonomic
groups, massive concentrations were formed by Flassisagitta
enflata, Oikopleura sp. I, Euphausia similis var. "armata."

Some massive species reached a high density at all stations
and levels. They included Clausocalanus porgens, Oithona
spp., Oncaea venusta, Corycaeus gibbulus, Microsetella rosea,
and Oikopleura sp. I. Of ihese, the most numerous species was
the cyclopoid. O. venusta. On the other hand, 60 rare species
occurring in 5-15% of the samples were counted. On the
whole, the lists, of planktonic organisms observed in equatorial
waters of the Pacific Ocean (Geinrikh, 1960; Vinogradov &
Voronina, 1963; Arashkevich, 1972; Stepan'yants, 1977) are
similar to those that we obtained (Table 1 ).

Mesozooplankton species composition in the section from
Caroline Atoll to the equator was approximately uniform, from
east to west for species of Cyclopoida, Harpacticoida,
Appendicularia, and Siphonophora, as well as for the minimal
content of Euphausiacea and Chaetognatha(Fig. 3, Table 1). In
the eastern portion of the region, such groups as Ostracoda,
Mysidae, and Salpidae were practically absent. Of the 13
determined Polychaeta, only two species were detected in the
region near Caroline Atoll. Some species of Chaetognatha
(i.e., Ferosagitta ferox. Pterosagitta draco, Parasagitta
speticoela, Sagitta sp., Sagitta pulchra) found in the western
portion of the region were absent from Station 1 14. Although
species composition at the eastern stations was poor, relative to
those to the west, some species were found there that were not
found at any of the other stations (i.e., the Far-neritic
[Vinogradov & Voronina. 1963] copepod Undinula vulgaris
and certain siphonophores).

In the western section of the region, the number of species
of Cyclopoida, Harpacticoida, Euphausiacea. Amphiphoda,
Calanoida, and Chaetognatha increased significantly,
sometimes severalfold, in comparison with the eastern section
(Fig. 3). As one moved toward the equator, from Station 1 17.
the number of previously undetected deep-sea species of
Calanoida (Neocalanus gracilis, N. robustion, Bradycalanus
sp., Rhincalanus cornutus, Bradyidus armatus, Euchirella
amoena, Pleuromamma abdominalis, P. gracilis, Haloptilis
acutifrons, H. longicornis, Candacia longimana, and
Labidocera detruncata) increased in the samples. This may
have been due to a more intense mixing of equatorial waters in
the region of their increase or to more active vertical daily
migrations of mesozooplankton (Vinogradov & Voronina,
1963). The western section was also much richer in the species
composition of Cyclopoida: previously undetected species of

Numbers, sp./m
2000 ■

1500 -

a)

1000 -

500-

120 119 118 117 116 115 114 Stations

b)

400

300

200-

100-

0

— i 1 1 1 i

120 119 118 117 116 115 114 Stations

Fig. 3. Horizontal distribution of mesozooplankton in the 200 m surface
layer: (a) distribution of numbers; and (b) distribution of biomass.

Corycaeus lautus, C. robustus, and six species of the genus
Sapphirina, Copilia longistylis, and Pachysoma dentatum
were present in the samples. The Harpacticoida group in the
samples near Caroline Atoll was represented by the single
species, Microsetella rosea. The new species of Clytemnestra
scutellata and Miracia sp., from deeper waters, appeared in
samples from the western stations. In contrast to the eastern
section, where Amphipoda were represented by species in the
family Hyperiidae, to the west, there were more species of
Platyscelidae and Pronoidae. Isolated specimens of Ramosia
sp., Vogtia serrata, and Maresearsia sphaera; representatives
of the Gerionidae family; larvae of cephalopod mollusks; and
large forms of Appendicularia ( Oikopleura sp. II) were observed
only in the western stations. Thus, moving from Caroline Atoll
to the equator, the diversity and abundance of mesozooplankton
increased.

Differences in species composition were characteristic not
only of the extreme western and eastern points of the region: a
number of interesting characteristics were also discovered at
Station 117. located at the center of the region. All the
taxonomic groups were represented most completely in the
community of zooplankton found at this station. For example,
of 30 species of Amphipoda, 22 were found at Station 117. and
of 13 species of Polychaeta, 1 1 were found. Large forms of
Tomopteris sp.. Oikoleura sp. II, a single specimen of
Nematoscelis gracilis, and concentrations of large Euphausia
similis var. "armata" and tomopterids were also found in the
waters of Station 1 1 7.

215

TABLE 1

Qualitative composition of mesozooplankton.

(Notation: *-rare species, numbers below 1 sp/m';

**-low-number species, numbers from 1-10 sp/m';

**-regular species, numbers from 10-100 sp/m3; and

:*-massive species, numbers in excess of 100).

14

115

116

18

119

120

1. Foraminifera
Radiolaria

2. Spongotrochus

3. Collosum sp.
Hydrosa

4 Gerionidae gen. sp.

5. Liriope tetraphylla
Siphonophora

6. Agalma sp.

7. Marrus sp.

8. Ramosia sp.

9. Vogtia serrata

1 0. Maresearsia sphaera

1 1 . Sulculeolaria quadridentata

12. 5. quadrivalvis

13. Galetta australis

14. Diphyes bojani

15. D. dispar

1 6. Lensia achilles baryi

17. L. campanella

18. L. multicristata

1 9. Lensia spp.

20. Muggiaea atlantica

21. Eudoxoides mitra

22. E. spiralis

23. Chelphyes appendiculata

24. C. contorta

25. Ceratocymba leucartii

26. Abylopsis eschschottsii

27. A. tatrugona

28. Bassia bassensis

29. A/>\7« schmidti

30. Diphydae gen. sp.
Polychaeta

3 1 . Alciopa parasitica

32. Alciopidae gen. sp.

33. Krohnia sp.

34. Rhynchonerella sp.

35. Maupasia sp.

36. Lopadorhynchus appendicular

37. Pelagobia logicirrata

38. Phillodocidae gen. sp.
40. Tomopteris elegans
41 Tomopteris sp. II

42. Travisiopsis levinseni

43. T. lobifera

44. Typhloscolecidae gen. sp.
Mollusca

45. Cephalopoda larvae
Crustacea

46. Ostracoda: Conchoecia sp.
Copcpoda

47. ( 'alanus minor

48. ( . pauper

49. Neocalanus gracilis

50. /V. robustior

5 I . Neocalanus spp. cop.

52. Undinula darwinii

53. £/. vulgaris

54. Calanus spp. cop.

55. Bradycalanus sp.

56. Eucalanus attenuates

57. £. subcrassus

58. Rhincalanus cornutus

*
**

*
*

*

**

*

*

**

*
*

*
*

*
**

**

**

*

**

**

***

***

**

**

**

**

***

***

*

*

*

*

*

*

*

*

*

*

*

*

*

***

***

**

***

**

***

***

*

*

**

**#

***

**

**

***
*

***

*

**

**

**

**

**

**

**

*

*

**
*

**

*

*
*

216

59. Acrocalanus gibber

60. A. gracilis

61. A. monachus

62. Calocalanus pavo

63. Paracalanus aculeatus

64. Mecynocera calusi

65. Clausocalanus arcuicornis

66. C pergens

67. Pseudocalanus minimis

68. Pseudocalanidae gen. sp.

69. Bradyidus armatus

70. Euchirella amoena

71. Aetidae gen. sp.

72. Euchaeta marina

73. Scolecitricella orientalis

74. Scoletrix danae

75. Phaena spinifera

76. Centropages calaninus

77. C. elongatus

78. C gracilis

79. C. longicornis

80. Centropages spp. cop.

8 1 . Pleuromamma abdominalis

82. P. gracifc

83. Lucicutia flavicornis

84. Z.. oralis

85. Heterorhabdus papilliger

86. Haloptilis acutifrons

87. W. longicornis

88. Candacia catula

89. C longimana

90. C. paehydactyla

91. C truncata

92. Labidocera detruncata

93. Pontcllina plumata

94. Arcatia negligens
Cyclopoida

95. Oithona spp.

96. Oncaea venusta

97. (9. notopus

98. Corycaeus agilis

99. C asiaticus

100. C. fO/H.V

101 . C crassiuscidus

102. C.flaccus

103. C.gibbulus

104. C.japonieus

105. C. /m/mv

106. C. longistylis

107. C. robust us

108. C. speciosus

1 09. Corycaeus spp. cop.

1 10. Sapphirina auronitens

111. S. gastrica

112. 5. intestinata

1 13. 5. metallina

114. 5. nigromaculata

115. 5. opalina

116. S.stellata

117. Sapphirina spp. cop.

1 1 8. Copilia longistylis

119. C mirabilis

1 20. C. cpiadreta

121. Pachysoma dentation
I 22. Nogagus muraji
Harpacticoida

123. Microsetella rosea

124. Clytemnestra scutellata

125. Miracia sp.

1 26. Copepoda nauplii

TABLE 1 - continued

114 115 116

**

**

**

**

**

***

***

#**#

**

**
**
**

17

18

*

**

**

**

*

***

**

**#

**

**

**

**

**

**

**

***

**

***

**

***

***

***

***

***

***

***

***

***

***

***

***

**

***

***

****

*

****

****

* =t- =fc

****

***

****

*

***

**

**

119

**
*

**

*

**

**

*

**

**

*

*#*

***

**

**

**

120

**

**

***

***

**

**

***

**

***

****

****

***

**

***

****

****

***

***

***

**

***

***

**

**

**

*

**

**

**

**

*

**

**

**

**

**

**

*

*

*

***

**

*

*

**

**
*

*

*

*

*

*

*

***

**

**

***

***

***

****

***

#***

***

****

****

*##*

***

$; % % ^
*

***

****

*

**

**

**

**

**

**

**

**

**

**

*

*

**

*

*

*

217

TABLE 1 Continued

114 115 llfi

117

118

119

120

Mysidacea

127. Mysidae yen. sp. larvae

128. Gammaridae num.
Amphipoda

129.

Vibilia chuni

130.

Paraphronima graciis

131.

Lestrigonus shizongeneios

132.

L. shoemakeri

133.

Hyperia fabrei

134.

Hyperietta stephenseni

135.

Phronima atlantica

136.

P. curvipes

137.

Phronimella elongate

138.

Phrosina semulunata

139.

Lycaeopsis zamboangae

140.

Eupronoe armatus

141.

E. maculata

142.

E. minuta

143.

Paralycaea gracilis

144.

Parapronoe parva

145.

Brachyscelus crusculum

146.

Oxycephalic longipes

147.

Leptocotis tenuirostris

148.

Calamorhynchus pellucidus

149.

Platyscelus armatus

150.

PI. ovoides

151.

Hemityphis tenuimanus

152.

Paratyphis parvus

153.

Paratyphis sp.

154.

Amphithyrus muratus

155.

Tetrathyrus forcipatus

Ruphausiacea

156.

Euphausia similis var. armata

157.

Euphausia spp. furcilia

158.

Nematoscelis gracilis

159.

Stylocheiron affine

160.

Stylocheiron spp. furcilia

161.

Eupnausiacea caliptopici

Decapoda

162.

Decapoda larvae

163.

Lin ifer sp.

164.

Echinodertnata larvae (onhiopl.)

Chaetognatha

165.

Pterosagitta draco

166.

Aidanosagitta sp.

167.

Ferosagitta ferox

168.

E. robusta

169.

Flassisagitta enflata

170.

Parasigitta septicoela

171.

Sagitta pulchra

172.

Sagitta sp.

173.

Serratosagitta pacifica

174.

Chaetognatha spp. imm.

Thaliacea

175.

Heliosalpa virgula

176.

Salpa sp.

177.

lasis zonaria

178.

Weelia cylindrit a

179.

Salpidae yon. sp.

Doliolidae

ISO.

Thalia demot ratica

181.

Dolioloides rarum

1S2.

Doliolidae yon. sp

App

endicularia

183.

Oikopleura sp 1

184.

Oikopleura sp II

IS5.

Pisces Ian ae

*
**

*
**

*
**

**

**
**

*

**

**

*
**

**

**

*
***

**
**

*
***

**
**

^ * * * * *

$£$:): *** $$$4

2 IS

As noted previously, the distribution of mesozooplankton
within the region is mainly due to the heterogeneity of
hydrological conditions found there (Korshenko, 1988). Asa
result, two bands of increased zooplankton biomass are formed
in the surface layer along the equator (Vinogradov & Voronina,
1963). In the Southern Hemisphere west of 140°W, the zone
of abundant plankton narrows to 5°S^°S. At the same time,
zooplankton productivity in the equatorial region increases
from east to west. We found that zooplankton biomass ranged
from 5 1 to 200 mg/m3. The western portion of the region
(Stations 1 19 and 120) was located in the zone of abundant
plankton (Fig. 1 ). The total biomass of mesozooplankton
ranged from 89 to 340 mg/m3 and averaged 201 mg/m3
(Table 2). The total biomass of zooplankton in the eastern
portion of the section in the upper 200 m layer did not exceed
163 mg/m1 (Fig. 3). At the westernmost point, and at Station
117. the biomass was maximum and amounted to 321 and
340 mg/m3, respectively. The total numbers of mesozooplankton
at different stations ranged from 547 to 2,170 sp/m3 and
averaged 1,044 sp/m3 (Fig. 3; Table 2), which doubled the
known quantity of plankton organisms (493 sp/m3) known in
the northern portion of the equatorial region (Korshenko,
1 988). In the eastern section, the numbers of mesozooplankton
ranged from 547 to 765 sp/m3. As one moved toward the
equator, starting at 6°S, the numbers of mesozooplankton
increased to a maximum of 2,170 sp/m1 at Station 120
(Table 2). It was shown that the qualitative composition and
structure of the community at Stations 1 17 and 120 differed
appreciably from those recorded at the other stations in the
region. This is evidently related to the characteristics of the
structure and dynamics of the water masses.

The vertical distribution of the mesozooplankton was also
different at different stations. At six of seven stations (Fig. 4).
most of the mesozooplankton was concentrated in the upper
100-m layer, and the maximum was in the 0-50-m layer
(Table 3; Fig. 4), corresponding to findings cited by Korshenko
(1988). A significant increase in zooplankton numbers and
biomass with depth was tracked only at the westernmost point
(Station 120). Mesozooplankton biomass as a whole increased
with depth owing to an increase in the fraction of large-sized
animals.

In addition to analyzing the vertical distribution of total
numbers and biomass, we also examined the vertical distribution
of individual taxonomic groups. We found that the numbers of
Calanoida and Euphausiasea decreased with depth, yet the
biomass remained practically unchanged because of an increase
in the fraction of large-sized forms (Fig. 5; Table 3). The
numbers and biomass of Siphonophora, Polychaeta, Ostracoda,
Hapracticoida, Amphipoda, Chaetognatha, and Salpidae
increased with depth. The numbers and biomass of Cyclopoida,
Mysidae, Decapoda, and Echinodermata, on the other hand,
decreased with depth. The numbers of Doliolidae were constant
at all the levels, and the biomass increased with the relative
content of large-sized animals.

In order to analyze the structure of the mesozooplankton
community, we divided the determined forms into 17 major
systematic groups. The frequency of occurrence (by numbers
and biomass) of 14 of these basic taxonomic groups is shown
in Table 4. The maximum fraction of numbers, as an average
for the cross section, belonged to Calanoida — 53.8%. In 60%
of the cases, Calanoida comprised over half of the total numbers.
According to Korshenko (1988), the numbers of Calanoida in
the 0-200-m layer amounted to 45% of the total value. The
highest numerical percentage of Calanoida (66.5%) was found
on Station 1 14. In the western portion of the section, Calanoida
numbers decreased to 40%. The biomass amounted to over half
of the total value only in the region of the Caroline Atoll, owing
to the small number of other taxonomic groups composing the
community. To the west, the relative content of Doliolidae was
comparatively high, and the fraction of groups other than
Calanoida was insignificant. We noted a tendency for the
number of taxonomic groups to increase toward the equator.
On one hand, the fraction of numbers and biomass of the
dominant group (Calanoida) decreased, while the relative
content of Cyclopoida, Chaetognatha, and Siphonophora
increased. For example, the fraction of numbers and biomass
of Cyclopoida at the equator was 48 and 29%, respectively. In
addition to the general tendency of the Calanoida fraction to
decrease, in many cases the typical (Fig. 6) structure of the
community broke down as a result of concentrations of animals
of a particular species. For example, the biomass fraction of
Euphausiacea was maximum (35.5%) at Station 1 17 as a result

TABLE 2

Total numbers and biomass of mesozooplankton in the 200 m surface layer.

Stations

Numbers

sp/m"

Biomass

sp/m'

Levels, m

0-50

50-100

100-200

0-200

0-50

50-100

100-200

0-200

1 14

755

775

_

765

158

131

_

144

115

1.354

826

428

759

148

202

150

163

116

791

806

246

547

87

158

55

89

117

2,056

1,153

523

1 .063

387

290

342

340

1 IS

1.028

704

699

783

124

122

146

134

119

1,170

1.947

878

1.218

227

220

211

217

120

1 ,439

2.137

2,560

2.170

126

226

466

321

Average values

1.127

1,193

889

1 .044

180

193

228

201

Remark: For Station 1 14. the data and results were obtained for the 0-100 m layer.

Depth, m

500

Numbcrv sp/ltl

1500 2000

200

Biomass. mg/m
100 200 300 400 500

Fig. 4.

Vertical distribution ofmesozooplankton; (a) vertical distribution of
numbers: and (b) vertical distribution of biomass.

of massive development of Euphausia similis \ar. "armatu" in
this area. Korshenkol 1988) notedasimilarca.se for a population
of E. similis. The same value (35.5% of total biomass) was
shown by the biomass fraction of Chaetognatha at specific
levels of the western portion of the section. A typical portrait
of the structure of the mesozooplankton community of the
region is shown in Fig. 6.

The general pattern of variability of the vertical structure
had the following characteristics: both the numbers and biomass
of Calanoida, Decapoda. and Appendicularia and the biomass
fraction of Cyclopoida and Siphonophora decreased with depth.
The relative content of other groups (Harpacticoida. Amphipoda.
Euphausiasea, Ostracoda. Chaetognatha. Salpidae. Doliolidae.
Siphonophora. and Polychaeta) increased with depth.

Conclusion

/. The qualitative composition of the mesozooplankton
was nonuniform in the region studied. As one moved toward
the equator, species diversity of the zooplankton increased.
The smallest number of species was recorded in the region of
Caroline Atoll. In the majority of cases, the number of species
increased with depth. It was postulated that the heterogeneitv
of the horizontal and vertical distributions ofmesozooplankton
resulted from vertical daily migrations and equatorial ascents
of the water masses.

Twenty-one species ofmesozooplankton reached numbers
above 100 sp/m': of these groups, the group of the most diverse
composition was Calanoida. The number of massive species
increased from east to west.

TABLE 3

Vertical distribution of mesozooplankton

Levels

0

50

m

50-

00 m

100

200 m

Taxons

N

B

N

B

N

B

I . Syphonophora

3.2

17.2

5.3

19.4

7.7

20.5

2. Polychaeta

2.1

3.2

2.6

3.3

3.8

6.2

3. Ostracoda

0.3

0.2

1.2

1.0

4.4

3.5

4. Calanoida

687.0

60.9

588.0

65.6

424.0

59.0

5. Cyclopoida

553.0

30.0

429.0

3.8. 1

328.0

23.0

6. Harpacticoida

31.0

0.2

73.0

0.5

47.0

0.6

7. Amphipoda

1.3

5.2

2.6

9.8

3.6

9.8

8. Mysidae

1.2

0.1

2.6

0.3

0.7

0.1

9. Euphausiasea

12.7

19.7

3.4

18.3

13.3

21.8

10. Decapoda

1.7

3.0

1.6

2 2

1.1

1.8

1 1 . Echinodermata

5.9

0.0

4.7

0.0

3.2

0.0

12. Chaetognatha

1 3.5

21.4

17.8

24.7

25.3

57.6

13. Salpidae

0.8

10.4

1.5

3.1

1.6

14.2

14. Doliolidae

0.9

14.4

1.4

20.4

1.0

18.3

15. Appendicularia

154.0

4.9

69. 1

2.1

33.7

1.0

16. Pisces

0.6

2.1

1.3

3.6

0.8

2.3

All Groups

1.227

180

.193

193

889

228

N - average numbers, sp/m3

B - average values ofbiomass,

sp/m

220

N, sp./m

B. mg/m

N, sp./m

1

Siphonophora

80

40

60

30

40

20

20

20

10

III

0

n

0

B.mg/m Harpacticoida

20 119 118 117 116 115 114 C\

N. sp./m

B,

mg/m

Calanoida

101)0 -

XQ

soo -

6Q

600 -
400 -

4Q

2(1

\i

/ • \ /

200 -

II

o-

i i

ill i

120 119 118 117 116 115 114 C

N, sp./m"

120 119 118 117 116 115 114 CT

30

20

10

0

20.

A Euphausiacea

nil

80.

60

411

-

;u

-

ii

tt » — -•

-

!

1 1

— i r 1 1 ■

120 119 118 117 116 115 114 CT

N, sp./m

1000
800
600
400
200
0

hi I

B, mg/m ' Cyclopoida

120 119 | |,s 117 116 115 114 C-

N, sp./m"

10
0

1. ma/m"

Amphipoda

B. mg/m

Chaetognatha

, sp./m

9o|
70

40 -

50

30-

30

20 .

10
0

^\vv

10 -

~* \

0

1 1 1 1 1 1

120 1 19 118 117 116 115 I 14 C-

20 119 118 117 116 115 114 CT

1
N. sp./m

B, mg/m"

Appendicularia

120-

ioo-

80-

60"

5

40-

4
3

^\

A\/^"

20-

2
1.

' V

\_

0_

0

120 119 118 117 116 115 114 CT

Fig. 5. Horizontal distribution of the main taxonomic groups of mesozooplankton.

221

TABLE 4

Structure of mesozooplankton community.

Stations

114

1

5

1 16

117

118

119

1

20

Average

values

Taxa

% N

9c B

', N

', B

'. \

% B

% N

9c B

% N

%B

', N

%B

', N

', B

'. N

\ B

1 . Syphonophora

0.3

2.5

0.1

1.3

0.3

7.8

1.4

10.5

0.6

7.1

0.5

17.8

0.4

8.4

0.5

7.9

2. Polychaeta

0.0

0.0

0.3

2.5

0.4

2 i

0.5

2.3

1.3

0.3

0.4

3.8

0.1

1.5

0.6

1.8

3. Ostracoda

—

—

0.3

0.8

0.4

1.8

0.3

0.4

0.9

1.1

0.2

0.8

0.1

0.3

0.3

0.7

4. Calanoida

66.5

54.3

55.9

42. S

51.0

34.0

57.9

20.1

56.2

31.9

48.8

34.8

40.0

28.2

53.8

33.7

5. Cyclopoida

24. 8

10.2

30.2

11.3

28.2

12.2

20.6

4.5

24.2

10.5

30.0

9.6

47.9

28.9

29.4

12.4

6. Harpacticoida

1.4

0.0

4.2

0.1

6.6

0.2

4.3

0.2

8.4

0.2

5.1

0.2

2.6

0.2

4.7

0.2

7. Mysidae

—

—

—

—

0.2

0.2

0.0

0.0

0.1

0.1

0.2

0.1

—

—

0.1

0.1

8. Amphipoda

0.2

2.0

0.5

6.1

0.2

4.2

0.7

6.4

0.3

5.4

0.2

2.6

0.1

0.8

0.3

3.9

9. Euphausiasea

0.5

0.4

0.4

0.8

1.0

1.0

2 2

35.2

0.6

1.0

0.8

~> 2

1.4

3.7

1.0

6.3

10. Decapoda

0.3

0.6

0.1

0.9

0.6

3.2

0.3

1.3

0.1

2.0

0.5

0.3

0.1

04

0.4

1.2

1 1 . Chaetoanatha

1.0

9.3

3.1

20.5

1.8

13.7

2 2

8.7

1.9

18.6

2.1

19.6

1.2

16.7

1.9

15.3

12. Salpidae

0.0

1.0

0.0

8.2

0.3

7.7

0.1

1.2

—

—

0.3

4.6

0.1

2.5

0.1

3.6

13. Doliolidae

0.2

18.3

0.1

3.6

0.1

10.4

0.1

3.3

0.1

11.3

0.1

1.5

0.0

6.0

0.1

7.7

14. Appendicularia

5.3

0.8

4.1

0.9

3.9

0.7

8.4

1.5

6.7

1.3

10.9

1.9

5.7

1.7

6.4

1.3

% N - fraction of to

al numbers

' 1 B - fraction of total biomass.

"3 -D

— i

u

j=

c

rj o

OJJ

■*

j

-z

- C

§■

3

u

X

Hi

u

a.

<

1 \'l

6 T\p

ca

structure

total

m

mt

ers

; ai

'•J

DO

a

-D

o.

E

-

^T"

0

7Z

—

s:

c

Q.

E

a

y.

Q

X

u

a.
<

%Ti

ton c

tmmunits

: (a) structure

ot

and ih) structure of total biomass.

2. In numbers and biomass of mesozooplankton. this
region can be characterized as medium-productive, with the
western section more productive than the east. The station
poorest in mesozooplankton was near Caroline Atoll. The
average numbers of mesozooplankton were maximum in the
0-50-m layer and decreased w ith depth. On the other hand, the
average biomass increased with depth owing to the increase in
the fraction of large-sized forms

3. The structure o\' the mesozooplankton community as a
whole was characterized by the dominance of the fraction of
numbers and biomass of Calanoida and Cyclopoida. In the
majority of cases, half of the total consisted of Calanoida. In
the western portion of the section, the relative content of other
groups (Chaetognatha. Euphausiacea) increased, while the
fraction of Calanoida decreased.

3.4 Zooneuston of the Tropical Pacific

YUZENALY P. ZAITSEV. LEONID N. POLISHCHUK, and BORIS G. ALEXANDROV

Department of Active Marine Surface Hydrobiology, Institute of Southern Seas Biology, Odessa Branch, USSR

Introduction

Seventy-nine zooneuston samples were taken at 23 stations
in tropical areas of the Pacific Ocean (Table 1 ). The sampling
equipment consisted of a PNS-2 two-tier plankton-neuston net
and an MNT fry-neuston trawl (Zaitsev, 1 97 1 ). The PNS-2 net
(mouth area 0.1 nr; mesh size 1 50 urn) permitted simultaneous
manual skimming of the upper surface of the upper layers of the
pelagial (i.e., of neustal [0-5 cm] and subneustal [5-25 cm]).
The MNT trawl (moutn area 0.39 m2; mesh size 350 p.m) made
possible high-speed catching of mobile neustons at a vessel
trawling speed of 3 m/s. The average skim duration was
10 min.

To facilitate the study of neuston community formation,
the whole of the ocean area investigated was subdivided into
three types of zones, namely 1. atoll lagoons (shallow water;
complete or partial isolation from the ocean; developed littoral
communities); 2. semienclosed marginal seas (relatively little
exchange with the ocean; mostly shallow water; well-developed
pelagic-shelf communities); and J. open-ocean waters (pelagic
communities predominant). The sampling effort covered an
extensive portion of the tropical Pacific from Caroline Atoll in
the east (154°51'E). The latitudinal limits of station locations
were 1 1°20'N and 8°40'S.

TABLE 1

Neuston sample collection in the tropical Pacific.

Area

\

umber of

Static

ns

Sampl

es

PNS-2

MNT

Caroline Atoll

3

22

—

Caroline Atoll/Phoenix Islands

4

8

4

Phoenix Islands/Gilbert Islands

6

12

1

Marianas sector

5

10

4

South China Sea

5
70

18
9

-

Total 23

The neuston (surface plankton) of the tropical Pacific has
been described in publications of a general character (Geinrikh.
1 964 ). There are also papers on the distribution of its individual
components (e.g., of copepods of the Pontellaidae family
[Sherman. 1963; Voronina. 1964] and oceanic water striders of
the genus Halohates [Herring, 1961]). Results of pleuston
studies are reported in a major survey paper by Savilov ( 1 969).

The need for the present study was dictated by a considerable
hiatus in observations, as well as by the scarcity of quantitative
data on neuston distribution in conditions of increasingly
widespread and intense pollution of the World Ocean.

The zooneuston of the tropical Pacific found during the
present studies was represented by epineuston consisting of
oceanic water striders (Halohates) as well as by hyponeuston:
copepods of the family Pontellidae, marine snails of the genus
Janthina, larvae of benthic invertebrates, and fishes. In addition,
allowance was made for the usual neuston components in the
form of semisubmerged organisms: Physalia (Siphonophora),
Velella velella, and Porpita pacifica (Chondrophora).

Analysis of the quantitative distribution of organisms
leads to some basic conclusions regarding the distinctive
features of neuston in the types of water areas studied.

The average counts of animals in the open ocean and in the
South China Sea were comparable (2,730 and 3,160 ind/m3),
while the concentration of animals in the lagoon at Caroline
Atoll was 50 times lower. Larger organisms (medusae, salps,
fish larvae and fry , euphausids hyperiidea, et cetera) constituted
aconsiderable fraction of the total neuston biomass, accounting
for as much as 2.5 g/nV (Station 1 19) and sometimes attaining
values an entire order of magnitude higher ( Station 128), where
Porpitae were dominant.

The relative neuston content was richer in Caroline Atoll' s
lagoon, the reverse of that for animals overall (Table 2). The
low numbers of oceanic animal species in the lagoon against a
background of early developmental stages of benthic
invertebrates (Decapoda, Gastropoda, Cirripedia, et cetera)
were due to shallowness of the waters and their relative
isolation from the open ocean. The distinctiveness of the
individual sea area categories (atoll, marginal sea. open ocean)
was clearly evident from the proportion of benthic animal
larvae in the total neuston count. Their percentage content
decreased steadily from its maximum value of 99% in the
lagoon to 75% in the South China Sea and 46% in the open
ocean. The epineustonic Halohates, which breed along coasts,
were similarly distributed, while the Pontellidae distribution
behaved inversely (0.24% and 54%, respectively).

In addition to benthic-invertebrate larvae, the neustal also
contained large numbers of pelagic animals at early stages of
ontogenesis, which validates calling neuston "the ocean's
breeding ground."

Comparison of the numbers of organisms in the microlayers
investigated confirmed the existence of conditions conducive
to neuston growth in all of the water areas studied. The sole
exception to this rule was the waters of the Marianas sector
(Stations 121-126) (Table 3).

223

TABLE 2

Components (counts expressed as individuals/m') of neuston

biocoenoses in the basic water area types of the tropical

Pacific.

B, in
mg/m *

Component

Basic water area type
Caroline South Open waters
Atoll China of the
lagoon Sea Pacific

1.

Pleuston

0

1.27

0.20

2.

Epineuston

0.31

0.13

0.05

3.

Hyponeuston

a) benthic-invertebrate

37.36

188.35

164.13

and fish larvae

37.36

143.15

75.71

b) Pontellidae and Janthia

0

45.20

88.42

4.

Others*
Relative neuston

26.63

2,968.9

2,566.6

content, %

59

6

6

* This category groups together mesozooplankton components
found in the surface layer, but distinct from the enumerated
categories of organisms in being more uniformly distributed
over the water column.

TABLE 3

Dominance of organisms in neustal expressed as a
percentage.

Sea area

',

Caroline Atoll lagoon
South China Sea
Open ocean
Marianas sector

1.3 ±0.1
2.5 ±0.6
1.5 ±0.3
0.9 ± 0.4

A degree of symmetry of neuston distribution with respect
to the equator was noted (see Fig. 1 ). This was evidence of the
effect of the northern and southern tradewind currents and of
the anticyclonic current between them,

The observed pattern of organism distribution agrees with
the data of Voronina ( 1964), who established that peak counts
of the most common pontellid Pontella tenuiremis occurred at
1°N and 1CS. The meridional components of surface current

100
6
4
2

•
•

•

• \ *

• .

•

0

10
North

8

6 4 2 0 2 4

Degrees of latitude

6 8

South

Fig. 1 .

Latitudinal distribution of neuston biomass (B) in the tropical waters
of the Pacific Ocean.

point outwards from the equator, so that species that inhabit the
topmost water layer are constantly carried away from it. The
departing water is replaced by deeper-lying water free of
surface animals. The result is an acute trough in surface species
along the equator (Sverdrup et al., cited by Voronina, 1964).

Constituting an agglomeration of hydrobionts. neuston
attracts both aquatic and aerial predators. We know, for
example, that decapod crustacean and pontellid larvae are part
of the diet of tunas, bonitos, and other epipelagic fishes ( Marchal.
1959). Porpita are consumed by marine turtles, while fish
larvae and fry are food for marine snakes (Zaitsev, 1971).
Neuston animals are an important part of the diet of seabirds.
Flying fish are caught on the fly by frigate birds (Fregatidae)
and terns (Laridae). Procellariiformes birds (Puffinus) have a
variety of ways of feeding on neuston; albatrosses ( Diomedea)
and fulmars (Fulmarus) snatch their prey from the surface, and
fork-tailed storm-petrels (Oceanodroma) do so while
performing a "mincing walk" on water (Boaden& Seed, 1985).
In waters close to shore, neuston is eaten by skimmers
(Rhynchopidae) with their long cultiform gonys (Zaitsev,
1971 ). According to the latest count (Day et al., 1984). there
are at least 50 species of birds that feed on neuston.

In conformity with what is usually the case with distribution
in the ocean, neuston is most profuse in areas where currents
converge. However, recent decades have seen a proliferation
in the same areas of various kinds of plastic debris ( Day et al. ,
1984). Looking for neuston, marine birds often swallow these
foreign objects, whose effect is invariably harmful and often
fatal. With their high adsorption coefficients, oil lumps and
plastic objects tend to have a buildup of toxic substances on
their surface (Osipov & Charykov, 1987), thus inhibiting the
development of invertebrates and fishes in the neustal. This, in
turn, can only have a negative impact on the bioproductivity of
the entire ocean.

224

3.5 Observations of Seabirds along a

14,892-km Cruise Track in the Tropical
Pacific Ocean and the Bohol, Sulu, and
South China Seas

ANGELA K. KEPLER1, CAMERON B. KEPLER\ DAVID H. ELLIS5, and JEFFREY S. HATFIELD }
US Fish & Wildlife Service, Patuxent Wildlife Research Center, Southeast Research Station, Athens, Georgia, USA
*US Fish & Wildlife Service, Patuxent Wildlife Research Center, Laurel, Maryland, USA

Introduction

This study forms a small part of the research efforts of the
First Joint US-USSR Central Pacific Expedition (Line and
Phoenix Groups. Gilbert Islands. Micronesia, inland Philippine
Seas, and South China Sea). Seabird observations were made
from the Soviet research vessel (R/V) Akademik Korolev
(7.000 tons, 1 24 m in length) from Hilo, Hawaii, to Singapore
(Figs. 1,2) via Christmas Island (02°N, 157°W) and Caroline
Atoll (10°S, 150°W).

The primary objectives of the expedition were to
characterize and contrast the fundamental oceanographic,
hydrochemical, microbiological, hydrobiological, and
ecological parameters of arctic and tropical marine ecosystems.
Emphasis was placed on the primary productivity and ecological
health of these two major areas, including pollution studies
involving multidisciplinary experiments conducted jointly by
scientists of both countries.

In this paper, we report the marine distribution of seabirds
and other transoceanic migrants, such as shorebirds and ducks,
during the tropical portion of the cruise, which covered
14.892 km from 9 September to 31 October 1988. Weconducted
a total of 1 6 1 hours of observations on transects representing an
area of 3,609 km2, during which time the ship traveled
4.5 1 1 km (Table 1 ). Our observations covered 21 degrees of
latitude and 1 07 degrees of longitude. Because the cruise track
traversed many island archipelagos, we subdivided it into the
following 7 regions (Figs. 1,2), all lying between 14°N and
10°S latitudes:

Region 1 Line Islands, including waters south of Hawaii
(150°Wto 160°W);

Region II Phoenix Islands to the international dateline ( 1 60°W
to 180°);

Region III Gilbert Islands (180° to 165°E):
Region IV Caroline Islands', Micronesia ( 165°E to 136°E);
Region V Philippine Sea and Basin ( 1 36°E to 1 25°E);
Region VI Inland Philippine Seas: Bohol (Mindanao) and
Sulu Seas. Balabac Strait (125°E to 1 17°E); and

Not to be confused with Caroline Atoll (Island), Southern
Line Islands, at 10°00S latitude, 1 50° 13'W longitude.

Region VII South China Sea. Philippines to Singapore ( 1 1 7°E
to 103°E).

We interpret our results within the contexts of 1. breeding
phenology, nonbreeding dispersion, and migration (Fig. 3);

2. proximity to known breeding colonies and nearest landfalls;

3. previous at-sea records; 4. anthropogenic factors such as
population density, environmental alterations to coastal habitats,
and pollution (see Chapters 2,3); and 5. general areas of
upwelling, providing locally rich feeding areas.

Seabirds, more than any other group of living organisms,
illustrate that the world's oceans are united. For example,
parasitic jaegers (Stercorarius parasiticus) breed in Siberia
and Alaska, then migrate south to winter in south temperate
waters of the Pacific. Indian, and Atlantic Oceans. En route, in
the Pacific, they skirt all four continents, passing through 150
degrees of latitude and at least 80 degrees of longitude.

The 1988 US-USSR expedition provided an opportunity
to study assemblages of birds, highly visible indicators of the
health of marine ecosystems, over vast areas of the Pacific
Ocean. Studies of the marine environments utilized by these
birds provide the knowledge to encourage practical action
toward their conservation.

Previous Studies

Although the broad distributions of central and western
Pacific seabirds are well-known (Murphy, 1936; Mayr, 1945;
Delacour & Mayr, 1946; Baker, 1951; Clapp, 1967; Clapp &
Sibley, 1967, 1968; Amerson, 1969; King, 1970. 1973, 1974a;
Nelson, 1975, 1978; Perry, 1980; Engbring, 1983; Garnett.
1983, 1984; Gould, 1983; Harrison, 1985; Pratt, Bruner &
Berrett, 1987), much remains to be learned of their detailed
distribution patterns in the Southern Line Islands and areas
west of Micronesia. Information on seabirds in southeast Asia,
at least in English, is incomplete and often outdated (Delacour
& Mayr, 1946; Delacour, 1947; King & Dickinson, 1975;
Nelson, 1978; Harrison, 1985). We are unfamiliar with the
literature in Asian languages and have not pursued the numerous
reports and publications resulting from marine oriented trips to
islands and reefs in the Asian region through which we passed
(UNEP, 1984a,b;IUCN, 1 988 a,b), some of which may contain

225

Not thern

Mariana

Islands

Marshall Islands

^T^io/i*

.. M/t-M

Papua NewGuinea -I • ' knmwWT;

Kiribati *

PACIFIC OCEAN /./.•

i I

!

Tuvalu

■ ft/l?- k *IWT_IM*Tl

SSm #* \»

10 5 10/4

■ '• - \ ,/2e

*«Oi»„ ,s«.»«°s
Tokelau

\""

< s _$ to'i

Fig. I. Cruise track of the R/V Akademik Korolev: Regions I (Line Islands), II (Phoenix Islands). Ill (Gilbert Islands), and IV (Caroline Islands. Micronesia),
15 September-17 October 1988. Solid lines on cruise track represent daily observation hours, dotted lines, hours of darkness.

PACIFIC
OCEAN

AK AOEMIK

I- ID CRUISE TRACK

' 1 KOROLEV

La J CRUISE REGION BOUNDARIES "*«".

1 D fff)

Fig, 2. Cruise track of the WV Akademik Korole\ : Regions V (Philippine Sea and Basin), VI (Bohol and Sulu Seas), and VD (South China Seal. 18-31 October
lliss Solid lines on cruise track represenl dail) observation horns, dotted lines, hours ol darkness

lib

TABLE 1

Relative abundance of seubirds in each sector of the R/V Akademik Korolev cruise track. 1 5 September-3 1 October 1 988.

Region

I

I]

III

IV

V

VI

VII

Total

Observation Hours

29.95

21.88

18.00

42.78

13.70

14.57

20.45

161.33

Kilometers Tra\eled

Durinc Observations

772.91

656.73

554.70

1.210.87

356.67

404.14

555.38

4.5 11.41

Area Covered bv

Observations (km:)

618.33

525.38

443.76

968.70

285.34

323.31

444.31

3,609.13

No. Bird Species Seen

27

25

14

19

6

4

8

46

No. Individuals Seen

457

1.796

495

799

34

22

65

3.668

Average No. Birds Seen/hr

15.3

82.1

27.5

18.7

2.5

1.4

3.2

22.7

Average No. Birds Seen/10 km

of Observations

5.9

27.3

8.9

6.6

1.0

0.5

1.2

8.1

Average Bird Density

Per lOknr

10.20

45.72

14.6

10.38

1.76

1.63

3.95

12.61

No. Families Seen

7

7

5

9

4

4

4

11

Each hour is the combined total of two people simultaneously viewing opposite sides of the ship, and thus equals 2 hours of observation
in some other papers.

relevant seabird records. Most of the previous information on
breeding and at-sea distribution of tropical seabirds was gathered
from 1963 to 1969 (Humphrey, 1965) by the US National
Museum's (Smithsonian Institution) Pacific Ocean Biological
Survey Program (POBSP).

Methods

Daily at-sea observations were conducted from the flying
bridge of the R/V Akademik Korolev, 12 m above the sea
surface and within a viewing arc of approximately 180°.
During the first week we honed our observation skills, using
Harrison ( 1 985 ) as a major reference. Thereafter, we maintained
almost constant watch during daylight hours (29 days during
9 September-3 1 October). Methods were based on those of the
POBSP (King, 1970; Gould. 1974), modified by techniques
utilized elsewhere (Tasker <?/«/., 1984;Haney, 1985;Gould&
Forsell. 1989). The watch rotated between three observers
(AKK, CBK, and DHE), with two observers on watch at all
times. A change of one observer took place every hour on the
hour, each person alternating 2-hour watches with a 1-hour
break to reduce fatigue. Observations began 10-15 min before
sunrise and terminated 1 0-1 5 min after sunset. Because we did
not have dedicated ship time, observations were interrupted by
periodic oceanographic sampling stations, during which time
no seabird counts were made. Counts were not conducted
within 10 km of oceanic islands where we landed (Christmas
Island, Caroline Atoll, and Tarawa), but were made close to
land in the Bohol, Sulu. and South China Seas.

Our ship speed averaged 15 knots during observations,
higher than the 10 knots that Gould & Forsell ( 1989) consider
ideal. At this time of year in the tropical Pacific, our higher

speed did not cause identification problems because few species
were found in large numbers. When approaching or leaving a
sampling station, we interrupted observations if the ship was
moving less than 5 knots.

Bird counts were contained within rectangular strips
extending 400 m to each side of the ship. Observers stood left
and right of the midline of the ship, counting all birds seen
400 m or less ahead of the ship on their side, to a line
perpendicular to the ship's direction at their position. Thus
each hour of observation represents the pooled records of two
observers watching a combined strip 800 m wide. Data were
recorded on standardized field forms using local time. The
following information was included: identification (to species
or subspecies when possible), number of birds per sighting,
feeding flocks and other associations, flight direction, plumage
(adult, juvenile, sex), and weather. Due to poor lighting or
weather conditions, some birds were identified only to genus or
family. All birds sitting, flying, or flushed within the transect
were counted, and their different behaviors noted. No birds
were collected. The ship's position (latitude/longitude) and
speed were recorded at the beginning and end of each hourly
observation period. Ship-following species were noted during
position checks and intermittently during the watch and were
recorded when first seen.

Because the visibility of different species at sea varies
greatly, we subdivided the maximum 400 m transect width into
3 bands corresponding to the approximate detection distances
of each species: /. 1 00-m band — small species, ordinarily seen
relatively close to the ship, includes shorebirds, storm-petrels,
and Bulwer's petrel (Bulweria bulweri); 2. 30-m band — gadfly
petrels, shearwaters, most larids, and anatids, and 3. 400 m
band, large or conspicuous species such as boobies, frigatebirds.

227

Fig. 3. Status ol bud species observed along the cruise track of the R/V
Akademik Korolev, 15 September— 3 1 October 1988. Nonbreeding
visitors and migrants migrate north or south to breed.

tropicbirds, and white terns (Gygis alba). These different
detection distances were used as a basis for calculating the
densities and area coverages for each different species (Hunt
era/., 1 98 I;Briggs <■/<//.. 19871. Forease of comparison within
regions and with previous studies (King, 1970; Tasker et ai,
1984) and because the number of birds normally seen in
tropical seas is lower than that observed in colder waters, bird
densities were calculated in 3 ways: number/hour (per
2 observers), number/10 linear km, and number/10 km2. In this
paper, numbers/ 10 km' are used for comparison between
species and regions; the other 2 calculated values are reported
in Table 2.

Prior to the expedition, we examined skins of Pacific
seabndsat the US National Museum. Washington. DC. focusing
on variable, polymorphic, and particularly difficult species to
identify. Terminology used in this paper follows the American
Ornithologists' Union Check-list (AOU. 1983. 1985). The
order of species within families, and alternate English names,
follow Harrison ( 1985).

Results

Forty-six species (or field-recognizable subspecies) were
observed along the 14.892-km cruise track of the R/V Akademik
Korolev from 15 September to 31 October 1988 (Figs. 1,2).
These were represented by members ol 1 I families of seabirds
(including gulls, terns, and skuas), shorebirds (including
phalaropes), and ducks (Table 3).

The most abundant species was the sooty tern {Sterna
fusi ata), accounting for 62.8', of the total number of birds
(Fig. 3), but restricted to Regions I through IV. Noddy tern
{Anous sp.. primaril) brown noddy, A vtolidus), wedge-tailed
shearwatei I Puffinus pacificus), and white tern followed as the
next most frequently observed species We observed birds
belonging to the following three broad groupings (Fig. 3.
fable 4 );

a) Resident breeders. Seabirds may or may not have
annual breeding cycles, which affects the temporal patterns of
their dispersion at-sea. Within an archipelago, a species may
exhibit asynchronous egg-laying periods on different islands
or in separate colonies on the same island. For example, in the
Line Islands, red-footed boobies (Sahi sulci) laid primaril)

TABLE 2

Seabtrd densities along the cruise track of the R/V Akademik
Korolev, 15 September-31 October 1988.

Region

Species

Number Birds
Per Hour Peril) Per

Linear Km 10 Km3

Phoenix petrel
Tahiti/Phoenix petrel
Herald petrel
White-necked petrel
Cook's petrel
Stejneger's petrel
Bulwer's petrel
Unidentified petrel
Flesh-footed shearwater
Wedge-tailed shearwater
Sooty shearwatei
Christmas shearwater
Audubon's shearw ater
Unidentified shearwater
Unidentified shearwater/

petrel
Wilson's/Madeiran

storm-petrel
Leach's storm-petrel
Unidentified storm-petrel
Red-tailed tropicbird
Masked booby
Red-footed booby
Brown hoohy
Great frigatebird
Bristle-thighed curlew
Gray-backed tern
Soot) tern
Brown noddv
Black noddv

White tern
Unidentified tern

Phoenix petrel
Mottled petrel
Herald petrel
While necked petrel
Cook's petrel
Bulwer's petrel
Unidentified petrel
Flesh-footed shearwater
Short i.uled shearwatei
Wedge tailed shearwater
Soot) shearwater
Christmas shearwatei
Little shearwatei

0.17
0.03
0.07
0.10
0.20
0.06
0.10
0. 1 7
0.03
1.37
0.50
0. 1 3
0.17
0.10
0.70

0.37

0.07
0.01
0.03
0.04
0.08
0.02
0.04
0.07
0.01
0.53
0.19
0.05
0.07
0.04
0.03

0.14

0.1 1

0.03

0.04
0.07
0.13
0.06
0.20
0.11
0.03
0.88
0.32
0.09
0.11
0.07
0.04

0.72

0.03

0.01

DOS

0.17

0.07

0.32

0.13

0.05

0.07

0.80

0.31

0.49

0. 1 3

0.05

0.07

0.07

0.03

0.03

0.17

0.07

0.01

0.13

0.05

0.09

0.07

0.03

0.04

sss

3.44

5.73

0.10

0.44

0.07

0.17

0.07

0 1 1

0.10

0.04

0.05

0.03

001

0.03

0.09

0.03

0.05

Oils

0.02

0.03

0.05

0.02

0.03

0 52

0.11

0.17

0.14

0.05

DOS

0.09

0.03

0.16

000

0.03

DOS

0.05

0.02

0.03

0.05

0.02

0.03

0 52

0.1 1

0.17

0.09

0.03

0.05

0.05

0.02

0.03

0.05

0.02

0.03

228

TABLE 2 -continued

TABLE 2 - continued

Region

Species

Number Birds
Per Hour Peril) Per

Linear Km 10 Km

Region Species

Audubon's shearwater

0.05

0.02

0.03

Unidentified shearw ater

0.09

0.03

0.05

Unidentified shearwater/

0.05

0.02

0.03

petrel

Leach's storm-petrel

0.09

0.03

0. 1 6

White-throated storm-petrel

0.09

0.03

0.16

Red-tailed tropicbird

0.05

0.02

0.02

Masked boob)

0.32

0.11

0.13

Red-footed boob)

0.05

0.02

0.02

Great frigatebird

0.73

0.24

0.30

Lesser golden-plover

0.05

0.02

0.08

Gray-hacked tern

3.66

0.12

0.20

Soot) tern

77.91

25.96

43.27

Brow n noddy

0.09

0.03

0.05

White tern

0.73

0.24

0.31

III Herald petrel

0.56

0.02

0.03

Cook's petrel

0.06

0.02

0.03

Bulwer's petrel

0.1 1

0.04

0.20

Unidentified petrel

0.34

0.11

0. 1 8

Wedge-tai led shearwater

0.50

0.16

0.27

Unidentified shearwater/

0.17

0.05

0.09

petrel

Leach's storm-petrel

0.22

0.07

0.36

Masked boob)

0.28

0.09

0.11

Ruddy turnstone

0.06

0.02

DOS

South polar skua

0.06

0.02

0.03

Sooty tern

7.39

2.40

4.00

Brown noddy

13.11

4.26

7.09

Black noddy

0.06

0.02

0.03

Unidentified noddy

0.17

0.05

0.09

White tern

4 94

1.60

2.01

IV Kermadee petrel

0.05

0.02

0.03

Bulwer's petrel

0.47

0.02

0.80

Flesh-footed shearwater

0.05

0.02

0.03

Wedge-tailed shearwater

4.98

1.76

2.93

Unidentified shearwater/

0.14

0.05

0.08

petrel
Wilson's/Madeiran

storm-petrel
White-tailed tropicbird
Red-footed boob)
Brown booby
Great frigatebird
Northern shoveler
Lesser golden-plover
Sharp-tat led/pectoral

sandpiper
South polar skua
Parasitic jaeger
Unidentified skua
Black-naped tern
Sooty tern
Brown noddy
Unidentified noddy
White tern

0.02

0.01

0.04

0.91

0.32

0.40

0.84

0.30

0.37

0.05

0.02

0.02

0.02

0.01

0.01

0.02

0.01

0.01

0.05

0.02

0.08

0.12

0.04

0.20

0.09

0.03

0.05

0.05

0.02

0.03

0.14

0.05

0.08

0.02

0.01

0.01

4.70

1.66

2.77

4.21

0.15

0.03

4.68

1.65

2.75

0.87

0.3 1

0.38

V Streaked shearwater
Wedge-tailed shearwater
Unidentified shearwater
White-tailed tropicbird
Lesser golden-plover
Black-naped tern
White tern

VI Wedge-tailed shearwater
Great frigatebird
Red-necked phalarope
Pomarine jaeger
Unidentified tern
Unidentified gull/tern

VII Masked booby
Brown booby
Unidentified frigatebird
Red-necked phalarope
Pomarine jaeger
Parasitic jaeger
Unidentified skua
Caspian tern

Bridled tern
Crested tern
Unidentified tern

Number Birds

Per Hour

Per 10

Per

Linear Km

10 Km

0.44

0.17

0.28

1.53

0.59

0.99

0.07

0.03

0.05

0.15

0.06

0.07

0. 1 5

0.06

0.28

0.07

0.03

0.05

0.07

0.03

0.04

0.14

0.03

0.08

0.07

0.03

0.03

0.76

0.27

1.36

0.07

0.03

0.04

0.07

0.03

0.04

0.14

0.05

0.08

0.05

0.02

0.02

0.25

0.09

0.11

0.98

0.04

0.05

1.61

0.06

2.96

0.10

0.04

0.07

0.15

0.05

0.09

0.29

0.11

0.19

0.25

0.09

0. 1 5

0. 1 5

0.05

0.09

0.10

0.04

0.07

0.25

0.09

0. 1 5

from May to June on Christmas Island in 1965 (Clapp, 1967)
but from August to November on Caroline Atoll ( Kepler el ai.
Subchapter 1 .2. this volume) data indicated an egg peak from
August to November. Other species disperse variable distances
from their colonies during the breeding cycle. The case of
Audubon's shearwater {P. Iherminierl) was particularly
interesting. Generally seen within ISO km of its breeding
grounds, we found a bird 1,100 km away from its nearest
known colony. Christmas Island, but approximately 400 km
from Maiden, a little-known island that has unexplored potential
habitat for Audubon's shearwaters. Our data also included
range extensions of resident breeders. Forexample, the wedge-
tailed shearwater, the most abundant shearwater in the Pacific.
has rarely been recorded west of 180° in the Caroline Islands
(Micronesia). Philippine Sea. and southeast Asia, yet 849! of
our sightings (N = 245) occurred in this area.

b) Nonbreeding visitors. These seabirds tend to be highly
seasonal in their postbreeding movements. We observed birds
that breed from the arctic to the antarctic. Forexample. Cook's
petrels {Pterodroma cooki) breed on islands in temperate
waters of the South Pacific, then undergo long transequatorial
migrations into tropical and north temperate waters to spend
their austral w inter. Because of the enormous distances of such
migrations, and because the distributions of birds are still
poorly known in some regions, our data added range extensions
for Kermadee petrel [P. neglecta) and little shearwater
(P. assimilii I.

229

TABLE 3

Seabird abundance l\\ familj along the cruise track of the R/V Akademik Korolev, 15 September-31 Octobei 1988

Regions

Famik

1

II

III

IV

V

VI

VII

Total

Total No. Buds

457

1.796

495

799

34

2(1

67

3,668

No. Species

26

24

13

18

6

3

s

Procellariidae

117

35

22

243

28

2

-

447

I12.11*', i

Hydrobatidae

17

4

4

1

-

-

-

26

(0.71', )

Phaethontidae

4

1

-

39

2

-

-

46

(1.2591 I

Sulidae

30

8

5

38

-

-

6

87

(2.3791

Fregatidae

5

16

-

1

-

1

2

25

(0.6891 i

Anatidae

-

-

-

1

-

-

-

1

10.03', i

Charadriidae

-

1

-

2

2

-

-

5

iO .14', I

Scolopacidae

4

-

1

5

-

-

-

1(1

(0.279! )

Phalaropodidae

-

-

-

-

-

11

33

44

(1.2091 '

Stercorariidae

-

-

1

12

-

1

11

25

(0.6891 i

Laridae

280

1 .73 1

462

457

2

5

15

2.952

(80 489:

* For geographic

limits of each

region, see

Figs. 1 and

2.

TABLE 4

Seabird species abundance along the cruise track of the R/V Akademik Korolev,
15 September-31 October 1988.

Species

No. Birds Seen

Total Status

Region

III

IV

VI

VII

FAMILY PROCELLARIIDAE
Phoenix petrel {Pterodroma alba)
["ahiti/phoenix petrel (Pi. rostrata/alba)
Mottled petrel {Pi inexpe< tata)
Kermadec petrel (Pi. neglecta)
Herald petrel (Pi. arminjoniana)
W hite necked petrel (Pi. externa)
Cook's petrel (Pi. cooki)
Stejneger's petrel (Pi. longirostris)
Bulwer's petrel {Bulweria bulweri)
I indent 1 1 led petrel

Sue. iked shearwater {Calonectris leucomelas)

Flesh footed shearwater {Puffinus carneipes)

Wedge tailed shearwater (P. pacificus)

Soots shearwater (/'. griseus)

Short tailed shearwater (P. tenuirostris)

Christmas shearwater (P. nativitatis)

I .title shearwater ( /'. assimilis I

Audubon's shearwater (/'. Ihenninieri I

1 Inidentified shearwater

i nidentified shearwater/petrel

I Will 'i HYDROBATIDAE

Wilson s/\ laden an storm-petrel

(0i eanites oceanicus/Oceanodroma castro)

White throated storm-petrel {Nesofregetta fuliginosa)

1 each's si Dim petrel 1 0< eanodroma leucorhoa)

Unidentified storm-petrel

2
3

1
7

1

6

3

1

2

3

2

2

5

2

6

1
41

1
7

9

15

2

4

1

5

1

3

2

21

1

3

113

3 6

447

7

RB

1

RB

1

MS

-i

WMS

4

WMS

10

WMS

10

WMS

5

MS

27

RB

13

6

WMN

4

MS

293

RB

17

MS

1

MS

5

RB

1

MS

6

RB

6

31

26

1

WMS

2

RB

7

WMN

5

230

TABLE 4 - continued

Spe

No. Birds Seen

Total Status

Region

IV

VI

FAMILY PHAETHONTIDAE

Red-tailed tropicbird {Phaethon rubricauda)

White-tailed tropicbird [P. lepturus)

FAMILY SULIDAE
Masked boob) (Sula dactylatra)
Red-looted booby (S. sula)
Brown booby (5. leucogaster)

FAMILY FREGATIDAE

Great frigatebird (Fregata minor)

Unidentified frigatebird

24
4
2

16

39

36

2

46

5

RB

41

RB

87

1

37

RB

41

RB

5

9

RB

25

23

RB

T

■}

FAMILY ANATIDAE

Northern shoveler (Anas rlypeata)

1

1 WMN

FAMILY CHARADRIIDAE

Lesser golden-plover (Pluvialh dominica)

FAMILY SCOLOPACIDAE
Bristle-thighed curlew (Numenius tahitiensis)
Ruddy turnstone (Arenaria interpres)
Sharp-tailed pectoral sandpiper
[Calidris acuminata/C. melanotos)

5 WMN

10

4 WMN

I WMN

5 WMN

FAMILY PHALAROPODIDAE
Red-necked phalarope (Phalaropus lobatus)

FAMILY STERCORARIIDAE
South polar skua (Catharacta maccormicki)
Pomarine jaeger (Stercorarius pomarinus )
Parasitic jaeger (S. parasiticus)
Unidentified skua

FAMILY LARIDAE
Caspian tern {.Sterna caspia)
Black-naped tern (S. sumatrana)
Gray-backed tern (5. lunata)
Bridled tern (5. anaethetus)
Sooty tern (.V. fuscata)
Crested tern (S. bergii)
Brown noddy (Anous stolidus)
Black noddy (A. minimis)
Unidentified noddy
White tern (Gygis allui)
Unidentified tern
Unidentified gull
Unidentified lurid

44

2

8

266

1.705

133

201

3

2

236

18

5

1

3

200

3

1

16

89

37

31

44
25

WMN

5

WMS

->

3

MN

3

5

MN

6

12
2,952

5

5

MN

2

RB

10

RB

3

3

RB

2.305

RB

-i

2

RB

254

RB

6

RB

203

146

RB

5

7
2
2

65

3,668

Totals

457 1,796

495

799

34

2(1

RB = resident breeder in the tropical Pacific.

W = nonbreeding visitor

MS = migrant: south to temperate/antarctic breeding colonies.
MN = migrant: north " " /arctic

231

c) Direct migrants. This grouping includes shorebirds
( plovers, sandpipers, phalaropes ) and ducks as well as seabirds.
They nunc quickly through tropical waters from wintering
grounds further north or south en route to breeding areas in the
opposite hemisphere. Although some follow general routes
and can be predicted at certain times of year, the overall dearth
of studies in certain portions of the Pacific leaves much to be
learned of their at-sea distribution. For example, Stejneger's
petrel (P. longirostris) migrates from Chile to Japan, yet
sightings had been scant in between before this expedition.

Species Accounts

Family Procellariidae

Shearwaters and petrels provided the greatest species
diversity (17) of any family. Seven breed and disperse within
the tropical Pacific, 1 1 breed in the temperate South Pacific and
migrate to wintering grounds in the North Pacific, and 1 breeds
in the temperate North Pacific and migrates south to winter
(Table 4).

Although we did not encounter large migrating flocks, this
family ranked second in total numbers seen (447; Table 3) and
was particularly abundant in Regions I and II (Line and
Phoenix Islands), after which species richness declined markedly
to the west (Table 4). The western limit of procellariids
(wedge-tailed shearwaters) was the Sulu Sea (08°47'N,
121 28'E).

Densities of individual species ranged from the rarer
migrants at 0.03 birds/10 km2 (Table 2) to the widespread
resilient breeder, wedge-tailed shearwater, whose numbers
peaked in the Line Islands at 0.88 birds/ 1 0 km2. Several species
contributed to feeding flocks (Table 5 ). accounting for 10% of
their participants.

Phoenix Petrel (Pterodroma alba): All seven of this
rather uncommon species were sighted in Regions II and III.
within its relatively small range in the central Pacific. Phoenix
petrels sighted in the Line Islands were either within 1 .200 km
of their breeding grounds at Christmas Island or flying west
toward colonies in the Phoenix Islands. Their highest densities
were in the Line Islands (0.1 1/10 km2). None associated with
other birds or participated in feeding flocks.

Tahiti/Phoenix Petrel (Pterodroma roM rata or alba): One
Pterodroma found slightly east of Maiden Island (central Line
Islands i was either a Tahiti or a Phoenix petrel, look-alikes
difficull to distinguish in the field. Tahiti petrels breed in the
Society and Marquesas Islands. 800 km and 1.670 km.
respectively, from its observed position.

Mottled Petrel (Pterodroma inexpcciata): A single bird
was seen flying southeast near Birnie Island (Phoenix Group)
on 4 October. This species breeds in New Zealand during the
austral summer, so this individual was likely migrating south
from its winter quarters in the North Pacific. Its density in
Region II was 0.03/10 km

Kermadec Petrel (Plerodroma ne v lee la): The Kermadec
petrel breeds in several island groups just south of the Tropic
ol Capricorn from Lord Howe Island to the coast of Chile.
Formerly considered sedentary, recent records indicate thai it
ranges widely into the North Pacific (Gould iV' King. 1967:

Amerson, 1969; Harrison. 1985; Bailey et al., 1989). Several
records exist for the area extending from just south of Hawaii
to the Marshall and Phoenix Groups. An old record from Duke-
of-York Island (Bismarck Archipelago) indicates that this
species may also occasionally straggle almost to New Guinea
(King, 1970).

Kermadec petrels are not listed for the Gilbert Islands
(Amerson, 1969) nor for any of the Caroline Islands (Pratt
et al.. 1987): hence, the following observations extend the
known range for this species: 2 Kermadec petrels ( 1 light and
1 dark phase) were seen on the morning of 1 2 October at the far
eastern edge of Micronesia (04°03'N. 163°30'E). approximately
157 km south of Kosrae. The dark-phase petrel was flying
directly south, while the light-phase bird had joined a feeding
(lock of over 200 sooty terns mingled with small numbers of
wedge-tailed shearwaters. Bulwer's petrels anil south polar
skuas (Catharacta maccormicki). Kermadec petrels were seen
only in Region IV with a density of 0.03/10 km2.

Herald Petrel (Plerodroma arminioniana): The Pacific-
breeding range of this medium-sized gadfly petrel includes
Easter Island, the Pitcairn Islands, Tuamotus. Marquesas.
Gambiers, and, further west. Tonga and Chesterfield (Coral
Sea). In the nonbreeding season, herald petrels remain primarily
in the Southern Hemisphere, occasionally wandering north of
the equator (Harrison. 1985; Bailey etal.. 1989 (where they are
typically observed more than 150 km from land (King. 1970).
We observed four adults (three dark phase, one light) in the
Line, Phoenix, and Gilbert Groups. Two occurred just outside,
the others within, the known pelagic range of this species
(Harrison. 1985). The two dark-phase birds, beyond the
eastern borders of their known range (05°24'N. 156o60'W),
were flying southwest on 2 October (ca. 380 km east of
Washington Island. Line Group). Densities were highest in the
Line Islands (0.04/10 km2). Three of the birds were Hying
directly southwest.

White-necked Petrel ( Pterodroma externa) ( includes both
P. e. externa and P. e. cervicalis): We observed 10 of these
gadfly petrels: 3 occurred in September and October within
their main wintering grounds in the central Pacific between the
equator and Hawaii (0.07/10 km2), and 7 spanned the Phoenix
Group, where densities were highest (0. 1 7/10 km2). They are
known to be abundanl in (he former location from May to
November ( King. 1 967 ).

There are two subspecies that breed on opposite sides of
the temperate Pacific; the white-necked {externa) in the
Kermadec Islands northeast of New Zealand, and the Juan
Fernandez (cervicalis) in the Juan Fernandez Islands of Chile.
Seven birds were flv ing south or southwest I four Hew together),
presumably returning to breed on southern temperate islands
during the austral summer.

Cook's Petrel (Pterodroma cooki): Cook's petrel breeds
m the austral summer in New Zealand and the Juan Fernandez
Islands. Although this iransequatonal migrant ranges widelj
between the south and north temperate Pacific as tar as the
Aleutian Islands, its pelagic movements are not fully understood,
lew records exist from the central Pacific. There are two
records from the Phoenix Islands and sightings near the Hawaiian
Islands in (he northern spring (King, 1967; Harrison. 1985:

232

TABLE 5

Geographic Distribution of Flocks. "No. Birds" refers to the number of each species present in all

feeding (locks of each region. Species are arranged according to their overall relative abundance in all

feedinc flocks. No flocks were seen in Regions V-VI1.

Species

Region

111

IV

Flocks

No.
Birds

Flocks

No.
Buds

Flocks

No.
Birds

Flocks

No.
Birds

Sooty tern
Noddy sp.

Wedge-tailed shearwater
Brown noddy
White tern
Red-footed booby
Great frigatebird
White-tailed tropicbird
Masked booby
South polar skua
Stercorariid sp.
Audubon's shearwater
Bulwer's petrel
Kermadec petrel
"Shearwater/petrel"

85

6

1.074

2

108

1
1

200
200

1

1

3

189

1

2

3

III

3

6

4

20

1

18

1

1

1

30

2

7

1

4

1

3

1

2

1

1

1
1

1

1

1
5

Totals

85

1 .095

239

651

Pratt el al., 1987). They have not been previously recorded
from the Gilbert Islands (Amerson, 1969). Our records are as
follows:

Line Islands. Single birds were seen on 15 September at
10°38'N. 156°16'W; 10°13'N. 156°19'W (Hying north):
09°29'N. 156°26"W (with 2 Bulwer's petrels); 16Septemberat
06°16'N. 156°50'W (flying north): 04°55'N. 157°00'W (flying
south): and on 19 September at 00°22'S. 156°37'W.

Phoenix Islands. All single birds flying south, as follows:
2 October at 06°29'S, 162°19'W; 3 October at 05°16'S.
166o53'Wand05o12'S. 167°06'W.

Gilbert Islands. One bird flying southeast on 7 October at
00°49'S. 179°10'E.

We observed 12 adults (10 on transect) between
15 September and 7 October 1988. Two had particularly pale
plumage. Sightings occurred just north of and within the
Northern Line Group, where densities were highest
(0.13/10 km2), in the Phoenix Group, and east of the Gilbert
Islands. Of seven birds flying in a direct compass direction,
four from the Line and Phoenix Islands were flying south, and
one from the Gilbert Islands was flying southeast. Thus 71%
of individuals of this species were flying in the approximate
direction of their New Zealand breeding grounds, when such
movements are expected. Only one Cook" s petrel was associated
with other birds (two Bulwer's petrels).

Stejneger's Petrel {PterodromalonvirostrisY. We observed
two adult Stejneger's petrels in the Northern Line Group

(05°56'N, 156°53'W; 06°24'N. 156°50'W) on 16 September
1988, about 170 km east of Palmyra Atoll. There are few
central Pacific records; POBSP personnel saw some near the
Phoenix Islands (King. 1967). although this is not mapped in
Harrison ( 1985). There appear to be no records of Stejneger's
petrel between Hawaii and the Phoenix Islands.

Additional sightings were made during the ICBP 1990
Line and Phoenix Islands Expedition (ICBP. 1990; Kepler,
1990). We observed no Stejneger's petrels in the Line Islands
during March and April but saw several birds heading north in
the same area during May ( 1 1° to 15°S, 149° to 15 1°W).

Bulwer's Petrel (Bulweria hulueri): This species was
relatively common ( N = 27 ) throughout the four regions of the
Pacific covered by the cruise. Because of its small size, it was
only counted within 1 00 m of the ship ( see Methods). Since this
species breeds and ranges at widely-scattered locations in the
Central and western Pacific, all our observations fell within its
expected range. Regions I through IV. Three-quarters of our
sightings were in Micronesia, primarily south of Kosrae and
north of Ulithi Atoll. None were seen west of Yap (139°E).
The major Bulwer's petrel breeding grounds in the northwest
Pacific (Bonin, Volcano, islands of Taiwan and China) are well
north of Micronesia, and Bulwer's petrel is regarded as a
species whose numbers decrease gradually with increasing
distance from land ( King, 1970). Breeding occurs from April
to September (King, 1967). Of our early October birds, equal
numbers were heading either north or south.

233

The overall density of Bulwer's petrel in Micronesia was
0.80/10 km2, lour times the density encountered in Regions I
through III (Line. Phoenix and Gilbert Islandsi.

Streaked Shearwater (CalonectrisleucomelusY. This large
shearwater breeds on coastal islands off China, Japan and
Korea and is known to travel southward during October and
November toward its main wintering area in the New Guinea-
northern Australia area. We observed six individuals, all from
18-20 October in the far western Pacific close to the Philippines
(Region V). Two were recorded as having darker plumage.
Sightings were restricted to a very narrow band of ocean
between 130°15'and 1 35° 14'E longitude. These were probably
postbreeding migrants: three ( 50% ) were flying due south, and
all were solitary.

Flesh-footed Shearwater (Puft'iniis carncipes): Flesh-
footed shearwaters breed during the austral summer on islands
off Australia and New Zealand, and winter in the North Pacific
north of the subtropical convergence. We observed four flesh-
footed shearwaters in Regions I. II and IV. all migrating birds
returning to the south temperate Pacific to breed. Two were in
the central Pacific south of Hawaii and in the Phoenix Islands,
where the species has been reported in very small but regular
numbers during the migration months. October to April ( King,
1967). The rest occurred in a more expected sector of the
western Pacific, although in a relatively narrow swath (05° 1 2'S
to 03°59'N, 164°08'E to 163°46'E), viz. between the western
Gilbert and far eastern Caroline Islands. The densities of these
transequatorial migrants were equal throughout (0.03/ 10 knr).

Wedge-tailed Shearwater ( Pitffimts pucitlcus ): The wedge-
tailed shearwater has long been considered the most common
widespread shearwater of the southwest Pacific (Mayr, 1945;
Jenkins, 1979). This species accounted for 8% of our total
sightings (Fig. 4). It breeds on numerous islands throughout
most of the tropical and subtropical Pacific from eastern
Australia to Mexico south to the Pitcairn Islands. We observed
293 individuals, 213 (73%) of which occurred in Micronesia.
The two largest concentrations, of 150 and 22 birds, both
occurred on 13 October near Pohnpei. Our sightings ranged
from immediately south of Hawaii through the Line. Phoenix,
and Gilbert Islands, Micronesia, and the Philippine Islands to
the far western edge of its range at 121°28'E longitude.

Although the pelagic distribution of this species is well
known for the central and eastern Pacific regions and Marshall
Islands (King, 1967. 1970. 1974b; Amerson, 1969), observations
are surprisingly scarce in the Carolines-Belau-Philippine region
( King, I974b;63. 93). For this reason, we discuss our records
in waters west of 180' longitude (Table 6). where 84' i
(N = 245) of our wedge-tail sightings occurred:

a) Gilbert Islands (Region III. IS0-165°E). Although
wedge-tailed shearwaters are known from this area ( I larrison,
19X5). we have been unable to find records of specific sightings
in a detailed summary of published data on wedge-tailed
shearwater distribution (King, 1974b). They are not listed,
even as visitors, in the Gilbert Islands by Pratt el id. ( 1987) or
Amerson ( 1969).

Fig. 4. Relative abundance of species or species groups on the cruise track of
the R/V Akademik Korolev. Total number of birds seen was 3.6hS.
belonging to 1 1 different families.

Although a resident breeder on several of the northern
Marshall Islands (Amerson, 1969), the closest recorded sightings
to the Gilbert Islands are 10 dark-phase birds seen on
2 November 1960 at Jaluit Atoll, southern Marshall Islands
(Morzer Bruyns, 1965:58) and between 04°00' and 02°30'S,
169-155°E in October 1951 southwest of the Gilberts between
Niutao Atoll and the Admiralty Islands ( Mac Donald & Lawford,
1954). These latter birds were reported as possibly P. came i pes.
which has a totally different flight pattern. Our nine sightings
were close to the equator at 179°E longitude, approximately
300 km north-northeast of Arorae Island. All birds were flying
south.

b) Micronesia (Region IV, 165-136°E). Although known
to be resident breeders in the northern Marshall Islands
( Amerson, 1 969:295 ) and central Carolines ( Murphy. 1951:9),
pelagic records of wedge-tails are sparse (King. 1974b:93).
Breeding colonies are also difficult to locate. Prattefa/. ( 1987)
state that this species is rare in western Micronesia: the
westernmost record at these latitudes i s evidently 1 50°E ( District
ofTruk).

We observed wedge-tailed shearwaters (N = 213)
throughout the Caroline chain of islands east to I32°59'E
longitude: 73' ; of wedge-tails observed on our cruise were in
this region.

c) Philippine Sea and Basin (Region V. 136-125 Fl. Of
21 wedge-tailed shearwaters seen west of 150°E (east of the
Philippines and ca. 450 km north of Belau). all were dark phase
(Table 6). Those in direct flight were heading either south or
west. This species mas visit Belau more frequently than
records indicate. ( )w en ( 197 1 ) does not mention its occurrence,
although Pratt ei id. listed it as a visitor. The overall density of

234

wedge-tails in this area, generally depauperate in seabirds, was
0.99 birds/10 km2, far higher than for any of the other five
species seen there.

d) Philippine Islands (Region VI. 125-1 17°E longitude).
Our October observations confirm the presence of this species
in the Sulu Sea (Delacour & Mayr, 1946; MacDonald &
Law ford. 1954). where we saw two dark-phase adults flying
south (0.08/10 knr).

e) Color Morphs. Flight Directions, and Densities. Pacific
Ocean Biological Survey Program studies revealed much about
the migratory movements of the wedge-tailed shearwater in the
central and eastern Pacific (King. 1967. 1970. 1974b). In
general, the dark-phase population predominates south of
10°N latitude, the approximate area covered by the east-
flowing Equatorial Countercurrent and west-flowing South

Equatorial Current. Dark-phase birds move northward
following the currents during the northern summer, reversing
their movements in the fall.

Dark-phase birds observed during October would
be expected to be heading south: of 39 individuals of known
flight direction. 31 flew south and 4 southwest, toward
the nutrient-rich waters of the Equatorial Countercurrent
upwelling between 4°N and 9°N latitude (King, 1974a,b).
Our largest concentration of wedge-tails was near Pohnpei
(06°N, 157°E), contributing to the highest daily density
of this species (10.95 birds/10 km:). At this time the sea
was "boiling" with fish, which attracted a great variety of
seabirds. Such densities of Wedge-tails are comparable to
those within the countercurrent latitudes in the central Pacific
(King, 1974a).

TABLE 6

Distribution, abundance, and behavior of wedge-tailed shearwaters west of 180°

7 October-23 October 1988.

Pacific Ocean and inland Philippine Seas,

Overall

Nearest

Density

Location

Island

No.

(birds per

Flight

Reg

ion Lat.

Long.

Date

or Group

Birds

10 knr)

Phase

Behavior Direction

III

01°00'S-02°55'N

180°00'-165°00'E

10/7-10/11

Gilberts

9

0.27

00°49'S

179°10'E

10/7

Arorae

8

-

Directed flight

S

00°49'S

179°10'E

"

1

-

S

IV

02°55'N-11°10'N

165°00'-136°00E

10/11-

10/17

Micronesia

213

2.93

03°53'N

I64°08E

10/12

Kosrae

1
1

Dark

Directed flight

S

04°03'N

163°31E

"

"

1

17

1

■•

Feeding flock
Directed flight

N

04°30'N

161°57'E

..

..

7

1

-

»

S

NE

05°24'N

159°04'E

10/13

Pohnpei

2

Dark

Solo feeding

-

05°32'N

158°37'E

"

"

1

-

Flushed by ship

-

05°43'N

158°25E

22

-

Feeding flock with
I South Polar Skua

-

06°02'N

158°06E

"

"

1

-

Directed flight

NE

06 I7'N

157°51'E

(in sight)

150

-

Feeding flock: 7 spp
408 birds

, -

06°28'N

157°41E

"

"

1

2
2

Dark

Directed flight

S

sw
s

06°38'N

157°31E

"

"

2

"

"

sw

08°59'N

155° HE

10/14

Murilo
(Hall Is.)

1

s

V

ll°10'-10°00'N

136°00'-125o00'E

Yap, Belau

21

0.99

11°02'N

135°19'E

10/18

120 km NW of Yap,

2

Dark

Directed flight

s

400

1 1 00'N

1 34°36'E

-

Belau

1

2

1
1

"

"

s

w
s

1 1 :'0()'N

132°59'E

10/19

3
10

Circling

w

1 l°00'N

132'59'E

10/19

Belau

1

0.99

Dark

Directed flight

s

VI

10 OO'N

I25C00'-117°00E

Philippines
(Sulu Sea)

2

0.08

08°49'

1 2 1 °44'

10/22

Negros. Mindanao

1

Dark

Directed flight

s

08°47'

121°28'

1

s

235

Dark-phase birds accounted for 979? of sightings where
color phase was noted (57). Only two light-phase individuals
were seen (Region 1. 10°46'N and ()6°45'N), both south of
Hawaii, where the majority of the population is light phased.

Short-tailed Shearwater {Puffinus tenuimstris): A wide-
ranging transequatorial migrant, with movements and breeding
phenology similar to those of the flesh-footed and sooty
shearwaters (P. griseus), this species breeds off southeastern
Australia and winters off the west coast of North America and
the Bering Sea. We observed one adult (2 October) heading
south through the eastern extremity of its known migratory
pathway, the eastern Phoenix Islands (06°29'S, 162°19'W).

Sooty Shearwater (Puffinus griseus): Sooty shearwaters
arc wide-ranging, transequatorial migrants that breed from
October to May in the Southern Hemisphere on islands off
Australia. New Zealand, and Chile, migrating to equivalent
high latitudes in the northern Pacific during the nonbreeding
season. We observed very small numbers of birds on their
annual southward migration. All 17 observations were in the
south-central Pacific. We saw them on only 2 days:
16 September (N = 15). east of Washington Island (Line
Group) and 2 October (N = 2), between Sydney and Starbuck
Islands (east of the Phoenix Group). All were adults
Hying south, and none were feeding or associated with other
species. Densities were 0.32 birds/ 10 km: in Region I and
0.05 birds/10 km2 in Region II. Enormous flocks of migrants
have been encountered in this area ( King, 1 967: CBK. personal
observation).

Christmas Shearwater (Puffinus ncitivitatis): The Christmas
shearwater, a year-round resident in the tropical Pacific, tends
to remain fairly close to its breeding grounds all year, although
it is generally seen more than 180 km from land (King. 1970).
We observed four individuals in the Line Islands (05°N,
157°W) that were Hying west, and a single bird in the Phoenix
Group, flying north. Densities were low: 0.09 birds/10 km2
(Line Islands) and 0.03 birds/10 km2 (Phoenix Islands).

Little Shearwater (Puffinus assiniilis): This small,
distinctive "aukish" shearwater breeds and disperses within
south temperate waters, generally only occurring at-sea north
to 25°S latitude. However, its pelagic dispersal is not well
known and wanderers have been recorded near the Marquesas,
Marshall, and Hawaiian Islands (King. 1967). Clapp ( 1967)
listed a doubtful record from Christmas Island in the late
I950's. We add a single straggler, seen on 4 October in the
Phoenix Islands (03 59'S, 171°31'W) close to Birnie Island
and Hying north. Within 20 minutes, we also observed the
similar Audubon's shearwater: both flew close to the ship,
enabling us to compare si/e, bill length, and the degree of white
present on the underw ings.

Audubon's Shearwater (Puffinus Ihcnninicri): This small
shearwater is resident in the central and western Pacific: our
five sightings ( Regions I and II ) were within its known range.

Our sightings of Audubon's shearwater at 02°N. 157 W
were evidently from adjacent colonies on Christmas Island.
However, we saw two birds flying northwest in the Southern
I ine Islands (06 32'S, I52°35'W) on 21 September 1988,
approximately 1. 1 00 km from Christmas Island and more than
1,800 km from Phoenix Island, the nearest known colonics

(Clapp. 1967; Stoddart. 1976: Garnett. 1983). Since this
species usually ranges at sea within 180 km of its breeding
islands (King. 1967). we speculate that an unknown colony lies
w ithin the Southern Line Islands. Maiden Island, approximately
400 km distant from the Audubon's shearwaters in question, is
a likely possibility. Little visited by biologists, approximately
one-third of its interior is covered by a landlocked, supersaline
lagoon with subterranean connections to the sea, containing a
maze of interconnected islets and salt flats (Garnett. 1983;
RNZAF. 1986). These islets have never been surveyed by
ornithologists (R.B. Clapp. personal communication). They
provide potential habitat for Audubon's shearwaters, since
they resemble sites occupied by this species on Christmas
Island. Bloxham ( 1925) recorded two species of shearwaters
of unknown identity on Maiden: Garnett ( 1983) proposes that
one of them may have been Audubon's shearwater. Starbuck.
another stark, arid guano island with a similar interior, is
another probable source.

On 5 May 1990. a single Audubon's shearwater was
observed flying north-northwest around 10°S, 155°W,
approximately 640 km due south of Maiden (ICBP. 1990).
further suggesting the presence of a colony in the Southern
Line Islands.

In Region II, we saw one Audubon's shearwater in a
feeding flock at 03D57'S. 171°31'W. near a large breeding
colony (ca. 12.000 birds) on Phoenix Island (Garnett. 1983).

Family Hydrobatidae

The storm-petrels were represented by three (possibly
four) species (Table 4) and five unidentified individuals.
Small numbers (N = 26) were present at sea from the Line
Islands west to Micronesia. Wilson' s/Madeiran storm-petrel
in the Line Islands accounted for the greatest densities
(0.72 birds/ 10km:). Overall, storm-petrels accounted for().7'<
of the total number of birds on the cruise (Fig. 4). Typically
solitary feeders, they never participated in feeding flocks. Our
westernmost observation was of a Wilson' s/Madeiran storm-
petrel at 03°59'N, 163°45'E (south of Kosrae).

Wilson' sSiorm-PelvcliOcecinitesoceiiniats) and Madeiran
(Harcourt's) Storm-Petrel (Oceanodronui custro): Wilson's
storm-petrel breeds on subantarctic islands off South America
and in Antarctica and ranges widely throughout all the world's
oceans (Murphy. 1936: Murphy & Snyder. 1952). In the
Pacific it migrates northward to wintering grounds within
tropical and north temperate waters (Harrison. 1985). It is
rarely seen in the Pacific except in the far east (King. 1967;
Huber. 1971;Crossin, 1974). Records and sightings exist from
the Marshall Islands. Solomons. New Hebrides. New Caledonia,
waters close to Hawaii, the Phoenix Islands, and Christmas
Island.

The Madeiran storm-petrel, although not congeneric with
Wilson's, appears remarkably similar in the field. It is a
resident breeder in the tropical Pacific: therefore, its dispersal
range overlaps with the wintering areas of Wilson's, primarily
west of the international date line.

Eleven Wilson/Madeiran storm-petrels were observed at
00°28'S, 156°32'W, east of Jarvis Island. Hying southeast. A
Wilson's, flying south, was located southeast of Kosrae

236

(04°03'N, 163°31'E). Huber ( 1971 ) has shown that Wilson's
storm-petrels move through the Marshall Islands from April
through September, sometimes in considerable numbers.

White-throated or Polynesian Storm-Petrel (Ncsotrcuetta
titliaiiiosa): The white-throated storm-petrel is an uncommon
central Pacific resident. We saw two (0. 1 6 birds/ 1 0 knr ) flying
north on 5 October in the Phoenix Islands, 200 km northwest
of the nearest land. McKean Island, where the world's largest
population! 1.000 birds) breeds (King. 1973). Approximately
500 birds also breed on nearby Phoenix Island, remaining in
adjacent waters throughout the year, with limited dispersal
eastward along the South Equatorial Current ( Harrison, 1 985 ).

Leach's Storm-Petrel (Oceanodroma leucorhoa): This
species breeds at subarctic and temperate latitudes in the North
Pacific, wintering primarily north of the equator. Its at-sea
range is centered in the central Pacific, with greatest winter
densities in a broad belt along the equator (Crossin, 1974). Our
seven observations, all between 0 and 3°S latitude, fell within
the known range of the species (i.e., the Line, Phoenix, and
Gilbert Groups). Five were flying north, and two were flying
southeast.

Family Phaethontidae

Tropicbirds were represented by two species observed in
small numbers (N = 46) from the Line Islands west to the
Carolines, including the Gilbert Islands. Most sightings were
of solitary birds during the morning hours, up to 300 km
from the nearest landfall. The highest density was that of the
white-tailed tropicbird {Phaethon leptums) in Micronesia
( 0.40 birds/ 1 0 km: ), six times higher than elsewhere. Tropicbirds
accounted for 1.3% of the total birds seen (Fig. 3, Table 3).

Red-tailed Tropicbird (Phaethon rubricatuhi): The red-
tailed tropicbird ranges widely in the tropical and subtropical
Pacific, breeding on many islands. Typically solitary, it is
highly pelagic and is often observed many hundreds of
kilometers from the nearest landfall (Harrison, 1985). Long-
term studies have found that it is observed in roughly the same
density regardless of distance from land (King, 1970).

We observed five individuals in Regions I and II between
1 7 September and 5 October. In the Line Islands, two occurred
close to islands (Christmas. Maiden). The remainder were
approximately 300 km equidistant from Caroline, Maiden, and
Starbuck. all of which harbor small breeding colonies (Clapp,
1967; Gould el al., 1974). although that on Caroline is the
largest ( Kepler et al. . Subchapter 1 .2, this vol.). All birds were
characteristically solitary; one was resting on the water. This
was also found by POBSP, whose number of sightings exceeded
one thousand: 87% of sightings were of lone birds, and 14%
were sitting on the water (Gould et al.. 1974). Pacific Ocean
Biological Survey Program found that birds were most
commonly observed during morning hours: 80% of our birds
were seen before 0820 h. All flying birds were adults flying
south. One bird was observed in the Phoenix Islands
approximately 300 km northwest of McKean Island, where
some 500 breed (Gould et al.. 1974).

Densities were highest in the Line Islands
(0.07 birds/10 knr). Such at-sea abundance is low, probably
because the species was breeding. On Caroline Atoll, for
example, eggs and chicks were present in late September
(Kepler et al., Subchapter 1 .2. this vol.).

White-tailed Tropicbird (Phaethon leptums): Like
P. rubricauda, the white-tailed tropicbird is a resident breeder
throughout the tropical Pacific. It is. however, less pelagic. Of
the 41 individuals we observed, 38 were seen on 13 October,
with Pohnpei in sight most of the day: most were alone or in
groups of up to four or fewer. Maximum numbers were seen
in the morning. Those seen with directed flight were generally
traveling toward or away from Pohnpei. Ten (26%) were
sitting on the water, and four had joined a large feeding flock.
White-tailed tropicbirds favor high islands for nesting and are
known to nest in trees and rocky cliffs on Pohnpei (Baker,
1951).

A single white-tailed tropicbird was sighted 400 km north
of Belau. where it breeds on several islands (Baker, 1951). Our
last sighting was of a lone bird on 21 October in Region V. on
the far northwestern edge of the species' range (Harrison.
1985). It was flying east approximately 200 km east of Samar
Island (Philippines) at 10°49'N. 127°53'E.

Family Sulidae

Boobies were seen throughout the study area except in
Region V, the Philippine Sea and Basin. All three pantropical
species were seen in the Line Islands only. Although we saw
only 87 boobies, this family ranked third in total numbers,
contributing 1.3% of all individuals observed on the cruise
(Fig. 4, Table 3). It was most common in the Line Islands and
Micronesia (Figs. 5,6; Table 3). Many sightings were within
80 km of their nearest breeding islands, but there were several
exceptions, one being a possible new pelagic record of a
masked booby (S. dactylatra) in the central South China Sea
(see below).

All three species of boobies commonly participate in
feeding flocks. This tendency is greatest with the red-footed
booby (5. sula)but is much less overall than for more gregarious
species such as sooty terns or wedge-tailed shearwaters (King,
1970). Our data ( 14 flocks) indicate that boobies participated
in 21.4% of the flocks (Table 7), accounting for 1.6% of the
total number of flocking birds (Fig. 7). Of 893 feeding flocks
in a large study area centered on the Hawaiian Islands, boobies
participated in 12.3%. although their numbers only accounted
for 1.8% of the total (King. 1970).

Masked Booby (Sula dactylatra): Masked boobies, resident
breeders in the tropical Pacific, were seen sporadically across
the entire cruise track (N = 37). They were most common in
the Line Islands, where densities reached 0.49 birds/10 knr.
Small numbers were present in the Phoenix and Gilbert Groups
and South China Sea (Table 2). Although wide-ranging
throughout the Pacific, this species was not mapped as occurring
west of around 143°E (i.e., the Marianas chain) (Harrison,
1985) but is recorded elsewhere from the Philippines, coastal

237

IV
N 799

V

N 34

111

N 495

Ml

PROCELLARIIDAE

E3

HYDROBATIDAE

in

PHAETHONTIDAE

^

SULIDAE

■1

FREGATIDAE

ED

ANATIDAE

EH

CHARADRIIOAE

M

SCOLOPACIDAE

En

PHALAROPODIOAE

in

STERCORARIIDAE

□

LARIOAE

VI
N 22

VII
N 65

Fig. 5. Relative abundance of birds bj famil} in Regions I-III.

Fig. 6. Relative abundance ot birds b\ tanuh in Regions [\ \ II See Fig. 5
for legend.

Java. Indo-China, Malaya, and Borneo (Delacour & Mayr,
1946; King & Dickinson. 1975). There appear to be no records
from the central South China Sea, where we saw a single
juvenile, flying west, on 2? October at 05°49'N, 107°30'E,
a location equidistant (ca. 400 km I from Borneo, Vietnam, and
the Malay Peninsula.

Nonmigratory. masked boobies generally occur within the
\ icinity of their breeding islands, hut the} roam far out to sea.
They characteristically follow ships: one-quarter of the birds
circled around or followed the R/V Akademik Korolev.
Juveniles-subadults accounted for two-thirds of our overall
sightings, especially in the Line Islands, where they breed.

Red-footed Booby (Sula sulci): This pantropical booby
ranges across the entire tropical and subtropical Pacific, breeding
in many localities. We observed 41 individuals in the Line.
Phoenix, and Caroline Islands.

Although nonmigratory, the lack of subadult birds at some
colonies has given rise to the hypothesis that although adults
arc relativel) sedentary, many juveniles disperse from their
natal islands (Schreiber & Ashmole. 1970). Pacific Ocean
Biological Survey Program personnel have shown that red-
foots generally remain within the vicinity of breeding or
roosting areas, and that adults are seldom encountered more
than 80 km from land ( King. 1967). Our sightings conformed
to this pattern. For example, we passed 500 km east ol Palmyra,
which harbors the largest colony of red-loots in the world
(25.000). yet saw none. The only red-foots seen in the Line
Islands were three light morph adults, Hying east, within
I 10 km of Maiden Island, where approximate!) 2.000 birds

TABLE 7

Participation b\ species or species groups in 14 feeding flocks
alona the 14,892km cruise track.

Species

Flocks

Participation

Number

Percent

Sooty tern

10

71.4

White tern

8

57.1

Noddies

5

35.7

Wedge-tailed

shearwater

4

28.6

Other shearw;

Hers

and petrels

4

2S.6

Boobies

3

21.4

Greal frigatet

iird

2

14.3

Stercoral! uls

2

14.3

White-tailed tropic

bird

1

7.1

breed (Clapp. 1967). and 1 juvenile, also flying east, around
460 km northwest of Caroline Atoll. In the Phoenix Group, we
obsen ed one dark morph adult in a feeding flock about 100 km
equidistant from several islands.

In the central Pacific, red-foots lay from February to
November, depending on the island and food supply (Nelson.
1978). In the Southern Line Islands, the maun' egg-laying
period from 1988 to 1989 was September to October (ICBP.
1990: Kepler era/., Subchapter 1.2, this vol. i. which may have
accounted for the small number of birds seen on our cruise.
During March and May 1 990, at-sea observations throughout
the Southern Line Islands recorded much larger numbers of

238

rod-foots than in September 1988. These were primarily
juveniles dispersing after a successful breeding season.

On 1 3 October we observed 30 red-foots in the vicinity of
Pohnpei. Most of these were participants in two feeding flocks,
one of which was the second largest on the cruise. Both flocks
developed immediately after a storm. The proportion of light-
to dark-phase birds at Pohnpei was 22:9. Red-foot density was
highest on this day ( 1 .62 birds/10 km:). Overall densities were
1 0 times greater in Micronesia than elsewhere, due to the large
numbers seen feeding near Pohnpei.

Brown Boobv (Sula leuconaster): This pantropical species
ranges widely in the Pacific and breeds on almost every island
group, although typically in smaller numbers than red-foots.
The brown booby, less dependent on the tropical blue waters
preferred by the other two species, is often found in inshore
waters, including harbors and estuaries, which are more polluted
and more accessible to man. Because of this. Nelson ( 1978)
considers that the brown booby has suffered more than its
congeners, probably accounting for their small numbers in
populated areas such as Micronesia and the South China Sea.
Brown boobies are seldom encountered more than 80 km from
land (King. 1967), but Harrison (1985) noted that there is
evidence to support small-scale dispersal.

We observed only nine individuals in the Line Islands,
Caroline Islands (Micronesia), and South China Sea. We
attempted to relate these individuals to nearby colonies. In the
Northern Line Islands (two sightings at 02°02'N, 157°37'W)
the closest colony was Christmas Island (N = 100), 32 km
distant.

In Micronesia, the only brown boobies (two adults) were
at 10°03'N, 150°02'E, approximately 150 km northwest of
Magur, Namonuito Atoll, Truk (07°N, 147°E), and 220 km
from uninhabited East Fayu Island, Truk (08°N. 151°E).
Although brown boobies nest on uninhabited islands in the
Marshall Islands (Amerson, 1969;SPREP, 1989),theircolonies
are rare and little known in the Caroline Islands. Some birds
were seen at Truk in 1945 (Baker, 1951). and earlier this
century, "incredible numbers of seabirds." which Nelson ( 1978)
suspects may have included brown boobies, were reported on
West Fayu, Gaspar Rico, and Magur. Nelson further notes that
"East Fayu is also a breeding site," implying a colony of this
species. Although the ornithology of these islands is virtually
unstudied (NID, 1945: Nicholson & Douglas, 1969; Owen,
1 97 1 , 1 977a, b; Ray Fosberg, personal communication), several
colonies of brown boobies may still exist in the Caroline
Islands. It is unlikely that the brown boobies we observed were
from the Marshall Islands, the nearest known colony of which
is Enewetak. about 1.400 km to the east.

Seabird information in the more remote islets of the
Philippines and in the South China Sea is similarly scarce and
outdated (Delacour& Mayr. 1946: King & Dickinson. 1975;
Nelson, 1978; Harrison, 1985). Nelson states that "the status
of the Brown Booby here is little known. There are some, could
be many and may be few." Their only known nesting colony
near Balabac Strait is at Tubbataha Reefs (68°N, 1 20°E) in the
central Sulu Sea (Worcester. 1911, in Nelson, 1978, IUCN.
1988a).

In the South China Sea, we observed five brown boobies
in the Balabac Strait close to numerous islands off the southern
tip of Palawan (07°44'N, 1 16°58'E) and near northwest Borneo
(05°92'N, 114°03'E and 04°40'N, 113°20'E). The closest
known colony is Spratly Island (Nan-Sha Reefs), evidently the
only remaining colony of brown boobies in Malaysian waters
that has not been overexploited (Nelson, 1978) that was still
extant 16 years ago (Haile, 1964, in Nelson, 1978). Both
locations were within 30 km of coastlines. Our brown boobies,
approximately 200 km west of this reef and flying east, could
have originated there (or on several other little-known islands
in the southwest Sulu Sea).

Overall densities of brown boobies were low everywhere,
but were four to six times higher in the Sulu Sea-South China
Sea than in the Line Islands and Micronesia, respectively
(Table 2). These distributional patterns are most likely due to
the proximity of our cruise track to potential breeding islands
and nearshore waters. We saw no brown boobies in feeding
flocks.

Family Fregatidae

Frigatebirds were seen throughout our cruise track except
for the Gilbert Islands and the Philippine Sea. The only
identifiable species was the great frigatebird (Fregata minor):
two distant, unidentifiable frigatebirds were seen in the central
South China Sea. Surprisingly, no lesser frigatebirds (F. oriel)
were seen, even though they winter in the western tropical
Pacific (Sibley & Clapp, 1967). Densities (Table 2) were
highest in the Phoenix Islands (0.30 birds/10 km2), where the
largest populations in the Pacific breed, and lowest in
Micronesia, where frigates are known to be scarce. Overall,
frigates accounted for only 0.7% of all birds observed on the
cruise (Fig. 4).

Great Frigatebird (Frcvcita minor): We observed 23 great
frigatebirds in the Line, Phoenix, and Caroline Islands. Densities
were highest in the Phoenix Islands (Table 2), where over
30,000 are known to breed (Stoddart. 1976).

We saw only five great frigatebirds in the Line Islands,
where around 13,000 breed (Clapp. 1967; Garnett 1983;
Kepler et al, Subchapter 1.2, this vol.). These low numbers
may reflect the fact that birds are concentrated on and near
colonies during the breeding season (Clapp, 1967; Kepler
et al.. Subchapter 1.2, this vol.).

In Micronesia, we saw only one frigatebird, part of a
large feeding flock near Pohnpei. Evidently frigatebirds are
infrequent in the Caroline Islands (Baker. 1951; King, 1967;
Pratt et al., 1987). We located only one reference to breeding
colonies of this species in this region (Niering, 1961 (.otherwise
they "are probably resident, especially in the eastern part"
(Baker, 1951).

The juvenile frigate in the southwest Sulu Sea (08°02'N.
1 17°28'W) and the adult in the South China Sea near Vietnam
(05°48'N, 106°5 l'E) could have come from anywhere, as they
may wander thousands of kilometers (Sibley & Clapp, 1967;
Nelson, 1975). Great frigatebirds were minor participants in
feeding flocks, contributing only 0.3% of the total numbers
(Table 8).

239

Family Anatidae

Onl) one vagrant duck was observed (below), with an
overall density in Micronesia of 0.01 birds/10 km2.

Northern Sho\e\er (Anas chpeata): This Holarctic breeder
is a rare but regular vagrant to Oceania. Recorded locations in
Micronesia are the Marianas, Pohnpei east to Wake, and the
Marshalls (Baker. 195 1 : Pratt et al, 1987).

We saw an unsexed bird in eclipse plumage three times
w ithin a half hour on 1 3 October near Pohnpei. Each time it « as
Hushed by the ship. Its characteristic bill shape and bicolored
speculum made identification relatively eas) .

Family Charadriidae

We observed only one member of this family in small
numbers during its southward migration from arctic breeding
grounds.

Lesser Golden-Plover ( Plmialis dominica): We saw five
golden-plovers: Phoenix Islands. Caroline Islands (north of
the Hall Islands and Ulithi Atoll), and in the Philippine Sea.
where its density was highest (0.28 birds/10 km2).

This long-legged plover breeds in Siberia and arctic Alaska,
migrating annually to islands throughout the tropical Pacific.
Most abundant from August to April, small numbers are
present all year ( Pratt et al.. 1 987). It is also a common migrant
in southeast Asia (King & Dickinson. 1975).

During an oceanographic station on 19 October, one
individual, uncounted on the transects, circled the ship several
times and landed, remaining on board for a week, during which
time it weakened considerably. Its final attempt at flight
resulted in drowning.

Family Scolopac idae

Ten members and three species of this family were
observed, all present in small numbers scattered within the
Line. Gilbert, and Caroline Islands. All were migrants from
Holarctic breeding grounds.

Shorebirds are rarely seen in large numbers at sea. even
during migration. For example. POBSP personnel on the
Town soul Cromwell, working replicate tracks in a 1 7 1 million-
ha study area from March 1964 to June 1965. observed only
four species, each only once (King, 1970). Nonseabird migrants
(sandpipers, plovers, phalaropes, and a duck) observed from
the R/V Akademik Korolev accounted for 1.6% of the total
birds seen (Pig. 3).

Bristle-thighed Curlew (Numenius taliitiaisi.s): One of
the world's least-studied shorebirds. the bristle-thighed curlew-
is considered rare throughout its range and is a candidate for the

1 S Fish & Wildlife Service Endangered Species List (Gill,
1990; Marks et al.. 1990). During the boreal winter, it is
common in the Line Islands, northwest Hawaiian Islands, and
southeast Polynesia hut uncommon to rare elsewhere in the
Pacific (Pratt et al. I9S7). All our at-sea sightings were on
16 September in the Northern Line Islands (density 0.09
birds/10 km ) clustered in a small area north-northeast of
Fanning Island and close to Christmas Island (04 55'N to

02 02'N, 157 00'Wto 157 37'W). Half the birds were flying
south, the most expected direction for September, as curlews
are most abundant in (he Line Islands and Tuamotus between

October and April (Gill. 1990; Kepler et al.. Subchapter 1.2.
this vol.). We saw no more until arriving on Caroline Atoll one
week later, where we estimated a population of >300.

Ruddy Turnstone (Armaria interpres): Another arctic
breeder that winters in the Pacific, the ruddy turnstone is a
widespread migrant to the Pacific, mostly betw een August and
May. It is common from the Hawaiian Islands and Micronesia
south to Samoa and Fiji (Pratt et al.. 1 987). A single turnstone
circled the ship in late afternoon about 80 km east of Abemama
Atoll, Gilbert Islands, on 8 October.

Sharp-tailed (Calidris acuminata )/Pet:(ora\ Sandpiper
(C. mclanotos): These similar species are arctic breeders that
winter in the South Pacific. We observed five birds in fall
plumage flying south in a tight flock on 160ctoberat U°10'N,
143°17'E. Both species have been recorded from Micronesia,
especially C. acuminata in the western portion of the archipelago
(Pratt etal., 1987).

Family Phalaropodidae

One member of this family was found in the South China
Sea, one of its favored w inter quarters.

Red-necked Phalarope (Phalaropus lohatus): A
circumpolar breeder, this phalarope migrates south to winter at
sea. Although it can be encountered almost anv w here in the
Pacific-southeast Asian region between the Aleutians and the
equator, it favors two areas: coastal Peru and the South China
Sea (Harrison, 1985).

We first observed small flocks of this species on
22 October on the eastern edge of the Sulu Sea (08°59'N,
123C12'E). They continued to appear thereafter,
characteristically flushing from the water close to the ship's
sides. Those that were not flushed were observ ed resting on the
water surface or feeding in areas where currents or wind
produced debris or foam lines.

We saw 44 phalaropes. 1 .20^ of the total number of birds
on the cruise. Their highest density was in the South China Sea
(2.96 birds/ lOknf).

Family Stercorariidae

Three species of skuas/jaegers were seen (N = 13)
plus 12 unidentified stercorariids (Table 4). The) occurred
primarily in Micronesia and the South China Sea. where
densities were 0.16/10 km2 and 0.35/10 km. respectively.
None were seen in the Line and Phoenix Islands, or in the
Philippine Sea anil Basin (Regions I. II. V). Stercorariids
accounted for 0.795 of all birds observed on the cruise.

Although ranging w idely. movementsoi skuas and jaegers
are little known except that some migrator) pathways tend to
follow coastlines (King. 1967). Some juveniles evidently
remain on the wintering grounds all year.

Only the Pomarine jaeger has been recorded lor southeast
Asia (Tenasserim, Malaya, central Thailand, and Hong Kong)
(Kingcx; Dickinson. 1987:153). However. King A: Dickinson
noted that three other species ma) occur there. We report
possible first sightings of the parasitic jaeger in the South China
Sea. Since all siereoranid sightings are rare for the western
Pacific and southeast Asia, our data on I 2 unidentified species
follows:

240

a) 1 2 October, two in feeding flock at 04°03'N, 163°3 1 'E;

b) 13 October, five flying north, one in feeding Hock
at 06°16'N, 157°51'E. Heavy squalls and winds from the
north:

c) 1? October, one Hying east at 10°19'N, 147°39'E; and

d) 25 October, six Hushed by ship from resting position on
water at 05°44'N. !08°01'E.

South Polar or McCormick's Skua iCaiharacta
maccormicki): This thickset, gull-like seabird breeds in
Antarctica. It is an uncommon migrant in the Pacific, known
only from the Hawaiian. Line, and Phoenix Islands, possibly
west to the Gilbert Islands (Pratt etui. 1987). Harrison (1985)
notes that juveniles disperse much further north than adults,
possibly following a clockwise path across the Pacific rim past
Japan (May-July) to British Columbia and California
(September-October), then back to Antarctica to breed in the
austral summer.

We observed five south polar skuas (three light- and one
dark-phased adults and one juvenile) in Regions III and IV
(Gilbert and Caroline Islands). All Hew very close to the ship,
and four participated in feeding flocks.

Our sightings fall within the known range of this species
( Harrison, 1 985 ), although other authorities indicate that south
polar skuas do not range this far west (Mayr. 1945: Baker.
1951; Pratt. 1987). Since published records are sparse (they
were not listed in Baker. 195 1 ). our exact sightings follow:

a) one juvenile, 8 October, flying south at OTOO'N,
172°40'E, just south of Tarawa, Gilbert Islands.

b) two adults, light phase, 12 October, in a feeding flock
at 04°03'N, 1 63°3 1 'E, south of Kosrae. The feeding flock was
composed of six species: two south polar skuas, two Pomarine/
parasitic jaegers. >200 sooty terns. 17 wedge-tailed shearwaters
(dark phase). 1 Kermadec petrel. 2Bulwer'spetrels. Itdeveloped
in midmorning immediately after a series of rain squalls had
moved through the area from the south.

c) two adults (one very dark, one light), 13 October at
05°43'N, l5°25'E,just south of Pohnpei. The dark morph was
flying southeast, the light one had joined a small feeding flock
of 22 wedge-tailed shearwaters. Four hours later heavy winds
and squalls hit us from the north, after which a massive
concentration of fish attracted a large feeding flock which
included an unidentified skua.

The highest densities of south polar skuas were in
Micronesia (0.05 birds/10 km:).

Pomarine Jaeger or Pomarine Skua {Stercoruriits
poinariimsY. This Holarctic-breeding jaeger migrates south to
winter throughout the tropical, subtropical, and temperate
areas of the Pacific. Evidently in good breeding years adults
move south in August-September, followed by juveniles in
September-October (Harrison, 1985). Our entire cruise track
was within its known range.

From 22 to 25 October we observed three Pomarine
jaegers in the Sulu Sea and South China Sea ( Table 4 ). One was
a light-phase adult, the rest, juveniles. One juvenile was
flushed from a resting position on the water, while the others
flew east or south. The highest densities of this species were
in the South China Sea (0.07 birds/ 10 km2).

Parasitic Jaeger or Arctic Skud(Stercorarius parasiticus):
The Holarctic breeding and Pacific wintering ranges overlap
those of the Pomarine Jaeger (Harrison. 1985). Since this
species has not been recorded from the central Pacific and is
uncommon in the western Pacific (for example Bismarck
Archipelago and Bonin Islands), its migration routes are thought
to remain close to the continental coasts (King, 1967). King&
Dickinson ( 1975) mentioned that it might occur in southeast
Asia.

Our observations came from the Caroline Islands and
South China Sea. One juvenile was sighted on 12 October,
south of Kosrae (04°22'N, 162°2 1 'E) flying east; and one adult
was seen on 14 October north of the Hall Islands (09°46'N,
154°43E), flying southeast. We also saw parasitic jaegers on
24 October off the north coast of Borneo (04°40'N, 1 13°20'E.
one juvenile, flying southeast) and on 25 October in the central
South China Sea (05°49'N, 107°30E, one adult dark morph.
flying southwest).

Three of our five birds on transect were headed in a
southerly direction, which is appropriate for October, when
they dispersed south from their breeding grounds.
Greatest densities (Table 2) were in the South China Sea
(0.09 birds/ 10 knr). which fits the hypothesis that migratory
routes tend to follow continental coasts.

An additional parasitic jaeger was encountered during our
3-day layover near Mui Bai Bung Cape, Vietnam at 05°47'N,
107°45'E): On 27 October at 1605 ha juvenile, dark morph
parasitic jaeger pursued an adult male shikra (Accipiterbadius)
with prey, a barn swallow (Hirundo rustica), for about
4 minutes (Ellis et ai, 1990). Finally the jaeger bound to the
shikra and/or its prey. All three birds whirled about three
revolutions and fell into the sea. After about 7 seconds in the
water, the shikra flapped away. The jaeger remained in the
water with its prize, drifting near the ship's stern.
Oceanographers later found the swallow's head in a plankton
net.

Family Laridae

This family was represented by 9 species of terns, plus
4 unidentified larids. totaling 2.953 individuals. Larids
accounted for 81% of all birds sighted on the cruise. They
dominated the pelagic waters of Regions I to IV, after which
their numbers diminished considerably. Sooty terns (Sterna
fuscata) were the most abundant species overall, accounting
for 62.8% of all birds seen on the cruise (Fig. 3), with a
maximum density of 43.27/10 knr in the Phoenix Islands.

Larids. which characteristically breed on oceanic islands
(sooty tern, black and brown noddies, white tern), were present
in far greater abundance than species that typically inhabit
areas closer to continents (Caspian tern, S. caspia; black-naped
tern, S. sumatrana; and bridled tern, 5. anaethetus). We
expected to see far more coastal birds in southeast Asia.
Human population pressures and habitat alteration evidently
weigh heavily on the natural resources in this area.

Sooty, white, and noddy terns were major participants in
feeding flocks, accounting for 88.129! of all Hocking birds
(Table 5).

241

Caspian Tern (Sterna caspia): This large, cosmopolitan
tern primarily breeds inland in Europe (including the Baltic Sea
and interior USSR) and North America, along sea coasts and
inland along rivers and lakes, migrating and dispersing
southward during the nonbreeding season, remaining very
close to continental coastlines. Apart from its winter status in
Japan as a visitor, the Caspian tern is rare or uncommon in
southeast Asia (Harrison. 1985: King & Dickinson. 1987):
records are from Cambodia. Laos, extreme south Vietnam, and
Thai land, but not from Singapore (Hails, 1987). We observed
five Caspian terns. Hying east and south, on the last day of the
cruise. 31 October, in Singapore Roadstead. Absent elsewhere,
this species had a density of 0.49 birds/10 knr here.

Black-napcd Tern [Sterna sinnatrana): A small, mostly
white, rather delicate tern, the black-naped is mainly sedentary,
breeding in the warm watersof the western Pacific and southeast
Asia. It is primarily a resident of seacoasts and offshore islets
(Harrison. 1985; King & Dickinson, 1987), extending eastward
no further than 180° longitude (King, 1967).

We recorded two, one adult 80 km north of Ulithi Atoll
(District of Yap, Micronesia) and a juvenile approximately
550 km east of the Philippines.

Gray-backed or Spectacled Tern (Sterna lunata): The
gray-backed tern breeds in the Hawaiian, Phoenix, Line, and
Tuamotu Islands, plus several other island groups to the
southwest. At sea, it occurs as a vagrant or migrant w ithin these
approximate boundaries (King, 1967; Harrison, 1985). Their
range is poorly known, probably because they are similar to.
and often associated in small numbers with, the abundant sooty
tern.

We recorded eight gray-backed terns in the Line and
Phoenix Islands. Two adults w ere seen on 20 September within
30 km of Maiden (04°03'S. I55°01'W). Within the Line
Islands, gray-backed terns are known to breed on Christmas.
Jarvis. and Maiden (Clapp. 1967). thus we were probably
observing birds from the Maiden colony (500-1.000 birds),
which would have been breeding at that time (Clapp. 1967;
Grossman & Grossman. 1974; Garnett. 1983).

In the Phoenix Islands we saw two birds centrally in the
group and six ( including one juvenile ) north of McKean. Gray-
backed terns breed on Enderbury (N = 10.000 birds). Phoenix
( N = 1 .800 birds), and McKean ( N = 23.000 birds) < King. 1 973;
Garnett, 1983).

Gray-backed tern densities were greatest in the Phoenix
Islands (0.20 birds/10 km>. close to the largest breeding
colonies in the world. Because breeding was in progress, most
tems were likel) to be on or near land.

Bridled or Brown-winged Tern (Sterna anaethetus):
Frequenting offshore waters and open ocean in the tropics and
subtropics, tins species is the ecological counterpart of
S. lunata in the western Pacific. We identified three bridled
terns in Singapore Harbor; they are known to be offshore
residents of the Malay Peninsula (Hails. 1987; King &
Dickinson. 1987).

Sooty Tern (Sterna luscata): The most abundant seabird
in the tropical Pacific, the soots tern breeds colonial!} in large
numbers on almost every island group. Ii ranges widely
between the tropics of Capricorn and Cancer.

We observed this species in the Line, Phoenix, and
Gilbert Islands and Micronesia. It was the most abundant bird
observed on the cruise (N = 2.305), nearly eight times more
numerous than the secondmost common species, the wedge-
tailed shearwater (Table 4). Sooty terns comprised 62.8%
of the total number of birds observed (Fig. 4).

Sooty terns ranged throughout the Line Islands, with
heavier concentrations near Christmas and Jarvis. Major
colonies, of approximately two million each, exist on these two
islands. The Christmas populations have declined drastically
in recent years (from an estimated 15 million 10 years ago) due
to predation by rats and cats, direct human exploitation for
food, and reproductive failure due to the 1982-1983 El Nino
Southern Oscillation (King. 1973: Gould. 1974b; Garnett.
1984; Schreiber & Schreiber. 1984; E. Schreiber, personal
communication).

In Region II. sooty tern populations are estimated at
around four million (Gould, 1974b). Here we observed the
highest density of all species (43.27 birds/ 1 0 knr), eight times
greater than sooties in the Line Islands (Table 2). Sooty terns
are abundant year-round in the Phoenix Islands, with highest
densities recorded from October to December (Gould. 1974b).

Their numbers at sea depend on their breeding cycles,
which are often complex and unpredictable, even within a
single island. In 1965, and from 1988 to 1990. Caroline Atoll
had 19 colonies, each on a different cycle, giving rise to egg-
laying in virtually every month (Clapp & Sibley. 1971a:
Anne Falconer, personal communication; Kepler et <;/..
Subchapter 1.2, this vol.).

In Region III (Gilbert Islands) smaller numbers of sooty
terns were observed (N = 1 13): we found moderate densities
(4.00/10 knr). although the species is reported as uncommon
at sea (Amerson. 1969). Our cruise track passed close to
Aranuka. which is thought to harbor breeding colonies
(Amerson. 1969).

In Region IV (Micronesia), we only observed sooty terns
south of Kosrae and Pohnpei (N = 201): 200 of these were
members of a single feeding flock. This species is known to
breed in small numbers in Micronesia on Pohnpei. on several
uninhabited atolls in the northern Marshalls (Baker. 1951;
Amerson. 1969). and on Helen Reef (Engbring. 1983).

Although ranging entirely across the Pacific Ocean through
southeast Asia into the Indian Ocean (Harrison. 1985). sooty
terns have been seen only infrequently in the far western
Pacific (Baker. 1951; Gould. 1974b). The POBSP found that
the perimeter of their peripheral breeding localities is also the
normal limit of their pelagic dispersal, and records beyond this
have historically been attributable to climatic disturbances
such as typhoons (Gould. 1974b).

Sooty terns, highly gregarious feeders and breeders,
participated in 7 1 .49! of all 1 4 feeding flocks and contributed
70.8'f of all birds within the flocks (Tables 5.7; Fig. 7). Of
2,305 sooty terns observed. 1.467 (64' r) were in feeding
Hocks, as was also found (65'i ) by Gould ( 1974b). Over our
entire cruise track, only 4'i were juveniles, suggesting that few
soot) colonies had bred recently. By contrast, during March to
May 1990 in the Line Islands, a much larger proportion of
juveniles were observed (ICBP. 1990).

242

Crested Tern ( Thalasseus hcr^ii): We observed only two
of these large, coastal, yellow-billed terns. Although a resident
breeder on several island groups in the central and western
Pacific and southeast Asia, it is everywhere uncommon ( King,
1967; King & Dickinson. 1967; Amerson. 1969). Our birds
appeared on the last day of the cruise as we entered Singapore
Roadstead.

Brow nor Common Noddy (Anous stolidus): This abundant
tern breeds on almost all island groups in the Pacific, extending
westward into southeast Asia (King, 1967; Pratt el ai. 1987).
Although widespread, it generally occurs within 80 km of its
breeding colonies. Much remains to be learned of the precise
details of its dispersal patterns (Harrison, 1985).

The brown noddy was the third most abundant bird seen
from the R/V Akademik Korolev (N = 259). An additional
203 noddies, seen in Micronesia, could not be identified to
species. Since we saw no black noddies (A. minutus) in
Micronesia, it is likely that they were also A. stolidus.

Although we sighted brown noddies in the Line, Phoenix.
Gilbert, and Caroline Islands, we counted only five birds total.
All well out to sea, some were 900 km from the nearest known
breeding colony (ca. 7.700 birds at Phoenix Island. POBSP
estimate, in Garnett, 1983).

Despite the fact that the Gilbert Islands are heavily
populated, brown noddy densities were highest there
(7.09 birds/10 km2): they were the most abundant bird in the
region (Table 2). This species, which can nest in trees, is much
more tolerant of humans than most seabirds. Pacific Ocean
Biological Survey Program personnel also found that/4, stolidus
was common throughout the Gilberts up to 80 km from land
(November to December. 1962 and 1964). It breeds on six
islands in the southern Gilberts, including Abemama, near
which we observed 235 brown noddies on 8 October.

In Micronesia, 18 brown noddies and 200 unidentified
noddies (probably A. stolidus) close to Pohnpei probably
originated on that island, where they are known to breed
(Baker. 1951).

Brown noddies were regular members of feeding flocks,
contributing 5% of their numbers (15% if the unidentified
noddies are included) and participating in 36% of flocks
(Tables 5.7). Corresponding figures from King (1970) are
7 and \\9c.

Black or White-capped Noddy (Anous minutus): This
species, smaller and darker than the brown noddy, breeds and
disperses over the same range. It is more sedentary, however,
and feeds closer inshore. Water temperature plays an important
part in the pelagic distribution of both noddy species (Murphy,
1936): both are largely absent from cold-water upwellings.

Despite the black noddy's widespread distribution in the
Pacific, we only observed six. Five were in the Line Islands
within a few kilometers of Christmas Island, where
approximately 10,000 breed (Garnett, 1983). The remaining
bird was in the Gilbert Islands, where A. minutus may still
breed. Twenty years ago there were small breeding colonies on
eight islands (Amerson, 1969), including Maiana, which lay
within 20 km of our cruise track.

Greatest densities (0.11 birds/10 km2) were within the
Line Islands, where breeding populations exceed 40,000
(Stoddart, 1976). Colonies are smaller in the Phoenix Islands,

but it is surprising that we did not see one, as we passed within
80 km of Orona ( Hull ). where 1 0.000 birds have been reported
(King, 1973). None participated in feeding flocks.

White or Fairy Tern (Gve/.v alba): Tolerant of man, the
white tern breeds throughout the tropical Pacific. As with
noddies, it is most common within 80 km of its breeding
colonies, but may wander great distances out to sea (King,
1967; Pratt et ai, 1987).

The white tern was the fourth most common bird (Fig. 4)
seen on the cruise (N = 146), occurring from the Line Islands
to the Philippines. Many sightings were of ones and twos; the
largest group (18) had been attracted to a large feeding flock
south of Pohnpei, Caroline Islands.

At-sea densities of white terns, as for many other tropical
species, reflect in part their breeding phenology, especially for
the more remote islands. Line Islands densities were low
(0.05 birds/10 km2) in September 1988 but much higher in
March to May 1990 (ICBP, 1990). Perry (1980) estimated
17,050 white terns in the Line Islands and noted that they
ranged widely and bred all year. Densities in the Phoenix
Islands were higher (0.31/10 km2), a surprise considering that
the overall population (10,000 birds; Clapp, 1967) is smaller.

White tern densities were greatest (2.01 birds/10 km2) in
the Gilbert Islands, due in large part to 89 birds seen the day
before anchoring at Tarawa. In Micronesia, 36 G. alba were
observed south of Kosrae and Pohnpei, and one was found
north of Ulithi Atoll. Baker ( 1 95 1 ) implies that it is common
in Micronesia, especially on low islands: Ulithi is listed as a
breeding site.

White terns are not listed for the Philippines (Delacour &
Mayr, 1946; King & Dickinson, 1975). The species evidently
rapidly decreases in density west of Micronesia. Our most
westerly sighting, an adult flying southwest at 1 1°00'N, 132°59'E
was approximately 480 km north-northwest of Belau, the
closest landfall, and on the extreme western Pacific limit of the
white tern's known range. This is the maximum distance that
we observed this species from land.

Feeding Flocks

A seabird feeding flock is regarded as an association of
five or more individuals acting as a unit (Gould, 1974b). In this
study we encountered both monospecific and mixed-species
flocks. We did not see any large associations of seabirds that
were not feeding flocks (i.e., flocks of direct migrants). The
following analysis deals with the size, abundance, composition
(species and family), and geographic distribution of the
14 feeding flocks we encountered.

Flock Size and Abundance

Flock sizes ranged from 8 to 420 birds, averaging
148 birds per flock. The total number of birds in feeding flocks
(N = 2,070) represented 56.5% of our bird sightings, a figure
comparable to the 69.5% found in 893 flocks by POBSP ( King,
1970).

The smallest flock (N = 8), of seven white terns and one
brown noddy, occurred in the Gilbert Islands on 8 October.
The largest flock (3 October in the Phoenix Islands) was
monospecific (420 sooty terns). The second largest (N = 408)

243

contained 7 species and was predominantly composed of
wedge-tailed shearwaters (8695 ) and unidentified noddies. It
was encountered on 13 October adjacent to Pohnpei. Three
environmental factors undoubtedly contributed to the size and
diversity of this flock:

/. The flock was near a high island at 07°N. within the
Equatorial Countercurrent boundaries (ca. 4°N and 9°N ). where
particularly rich upwellings provide feeding grounds for both
fish and birds (Ashmole & Ashmole, 1967: Gould, 1974a);

2. Flying fish, considerably more abundant compared to
the previous 3 days, indicated increased productivity; and

3. The flock was attracted to masses of predatory fish
(primarily tunas) vigorously leaping from the water, causing
the water to "boil."

Flock Composition

Five families (Laridae. Procellariidae. Fregatidae, Sulidae,
Phaethontidae ), 1 2 species, and 3 unidentifiable groupcategories
were represented in feeding flocks. Sooty terns dominated
the feeding flocks, followed numerically by noddies and wedge-
tailed shearwaters (Fig. 7; Tables 5,7). These three species
groups accounted for 95% of all birds seen in flocks.

Sooty terns participated in 71.4% of all flocks (Table 7). a
similar proportion to that found by POBSP (76.0%) over a
longer time frame (King, 1970). White terns were found in
57% of the flocks, noddies (primarily brown noddy) in 38%,
and wedge-tailed shearwaters in 29%.

White terns were also regular, though minor, members of
pelagic feeding flocks, found in 57% of all flocks and
representing 2% of total numbers (Table 7, Fig. 7). King
( 1970) found that white terns participated in 9% of flocks and
noted that the relatively high flocking tendency of white terns

lie. 7. Relative abundance >>i spe*. us ..i -,|x\ io groups loniul in .ill feeding
flocks (N = 14). Total numbei ol birds was -.(172 (\ = I4S).

was of interest since this species had been thought to be a
solitary feeder. Our data also suggest that the white tern
commonly joins mixed-species feeding flocks.

White terns are generally more solitary on land and at sea
and have not been found to exhibit such a high rate of flock
participation (57.1%) as we observed (King, 1970). This
species, however, joined flocks only in small numbers. Overall.
five white terns on our cruise joined over half the flocks, but
their total numbers within flocks accounted for only 2.1%
(Fig. 7; Table 7).

Geographic Distribution

We encountered feeding flocks in Regions I through IV
(Table 5). viz.. in the Pacific Ocean from 150°W to 158°E
longitude. We saw no flocks from Pohnpei westward to
Singapore.

We found only one flock, composed of 85 sooty terns, in
the Line Islands. More than half the flocking birds were within
the Phoenix Islands (Table 5). where the highest seabird
populations and greatest species diversity were found. Great
frigatebirds joined flocks only in the Phoenix Islands.

In Region III (Gilbert Islands), sooty terns, brown noddies,
and white terns were the only species present in feeding flocks.
Sooty terns and brown noddies accounted for 92% of flocking
birds in this area (Table 5).

Region IV (Micronesia) was relatively rich in feeding
flocks that contained several species not encountered elsewhere:
stercorariids, Kermadec and Bulwer"s petrel. Audubon's
shearwaters, and white-tailed tropicbirds.

The feeding flocks we observed were found in areas
known to be nutrient-rich. In broad terms, latitudes north of
10°N and south of 10°S are poor in nutrients. The zone in-
between, especially from 04 or 05°N to 09 or 1 0°N ( Equatorial
Countercurrent) and from 00 to 09 or 10°S. is considerably
richer in plankton and schools of small fish and tuna (Ashmole
& Ashmole. 1967). Superimposed on this general pattern are
local upwellings of plankton or "fronts" that occur close to
islands and arc particularly evident between 0 1 °30'S and 05°N
(King & Hida. 1957).

The geographic distribution of our feeding flocks fits these
general patterns (Fig. 8). Their latitudinal limits were07°N and
07°S. At 02°S. within the plankton bloom on either side of the
equator, we encountered a feeding flock of 153 birds. Rocks
were more frequent around island clusters. We saw four flocks
each close to the Phoenix and Gilbert Groups, and the day our
ship arrived at 05°N (13 October, south of Pohnpei! we
immediately observed a substantial increase in flying fish.
From 04°N to 07°N, we encountered 3 feeding flocks of 223.
22. and 408 birds, respectively. The dearth of fish. and. hence,
seabirds. north of 10°N was particularly evident. Not only
were there no feeding flocks, there was a substantial decrease
m (he number of birds compared to all other Pacific areas
(Table 1 ).

We saw no flocks in southeast Asia, where total numbers
of birds were low. Here we-encountered complicating factors
due to heavy pressures from commercial and subsistence
fisheries and other human population factors, which override
changes in oceanography that are associated with the continental

244

shelf. Southeast Asia has long been recognized as an outstanding
area for fish and. until the last few decades, was similarly rich
in seabird colonies (Nelson. 1978).

Regional Discussions

Region I (457 birds. 26 species. 7 families)

Region 1 (Fig. 1 ) begins with the waters between the Line
Islands and Hawaii north to 14°N. The Line Islands are a
scattered group of five atolls, five islands, and two submerged
reefs straddling the equator between 06°N and 12°S latitude,
and 162° and 150°W longitude. They are all low islands with
extremely varied ecology ranging from barren, tropical deserts
with scarcely any vascular plants to lush forests of coconut
palms and/or indigenous vegetation.

All except three are uninhabited and thereby serve as
suitable habitats for tropical seabirds. Together with the
Phoenix Islands, they constitute the largest assemblages of
breeding tropical seabirds in the Pacific, both in species diversity
and abundance. Christmas Island, with 18 species of breeding
seabirds. is one of the richest seabird islands in the world
(Ashmole & Ashmole, 1967; Garnett, 1983), due in part to
equatorial upw ellings and plankton associated with the seasonal
movements of the North and South Equatorial Currents and
Countercurrents (King & Hida, 1957; Ashmole & Ashmole,
1967).

Overall, we found greater seabird diversity and density in
the Line and Phoenix Islands (Tables 1,2,4). Two major groups
of birds were most abundant in Region I: shearwaters/petrels
and terns. Of minor importance numerically were boobies and
storm-petrels (Table 3). Of interest is the high proportion of
resident breeders and wintering birds (Fig. 3).

The 1982-1983 anomalous warm waters associated with
the El Nino Southern Oscillation severely disrupted seabird
breeding on Christmas Island (Schreiber & Schreiber, 1984).
Furthermore, feral cats, developmental threats, and increased
poaching associated with expanding human populations arc
ongoing problems on Christmas, and these affect the numbers
of birds seen at-sea in the area (Gilbert and Ellice Islands Gov.,
1974; Garnett, 1983; Teeb*aki, personal communication).

In the Line Islands, we extended the known range of herald
petrel and added sightings of Cook's petrel (six) and Stejneger's
petrel (two), both rarely recorded in the area.

Region 11 ( 1 ,796 birds. 24 species, 6 families)

The Phoenix Islands (Fig. 1) form a relatively compact
group of eight low islands lying from 03 to 05°S. Most are dry
and waterless. All except Canton are uninhabited and harbor,
like the Line Islands, some of the richest and largest seabird
colonies in the world. They lie within the boundaries of the
South Equatorial Current (04°N to ca. 10°S), a region rich in
plankton and associated fish schools (King & Hida, 1957;
Ashmole & Ashmole, 1967).

At-sea bird observations were dominated by sooty terns
(96Vr of total), with procellariids next in abundance (Table 3).
Small numbers of frigatebirds. boobies, and storm-petrels
occurred, as in Region I (Table 2). Resident breeders,
nonbreeding visitors and direct migrants were all well-
represented (Fig. 4). The highest density of any bird on this
cruise was in this area — sooty terns, at 32.45/10 km2.

In the Phoenix Islands we recorded a range extension of
the little shearwater, and added three sight records of Cook's
petrel, and nine of wedge-tailed shearwaters (Table 6). Because
of the remoteness and unsuitability of the Phoenix Islands for

Northern

Mariana

Islands

Marshall Islands

Papua New Guinea

|»»»| ranrse wm ■uumiumb

fl — ^ POLrTICM tiiu.c."..

[~T"1 CHUUi TB»C«.««*P<I»« KOWOltv

J,.s

Fig. 8. Geographic distribution of feeding flocks (dots).

245

human settlement (Howland and Baker are US National Wildlife

Refuges, and Birnie, McKean. and Phoenix Islands are Kiribati
Wildlife Refuges), its seabirds appear reasonably safe from
disturbance for the present.

Region III (495 birds. 13 species, 5 families)

The Gilbert Islands (Fig. 1 ) stretch in a compact arc from
03°17'N to 02°38'S latitude and from 176°49'E to 172°48'E
longitude. They compose a single archipelago of 1 1 atolls and
5 reef islands, forming a southerly extension of the Marshall
Islands.

All the Gilbert Islands are populated, some heavily. Since
the I-Kiribati have long utilized seabirds and their eggs for
food, species sensitive to human disturbance (shearwaters,
petrels, boobies, tropicbirds) are absent. The primary breeding
species are tree-nesters such as brown noddies and white terns
(Table 4).

Our observations were dominated by larids (Table 3).
Brown noddies (48% ) and sooty terns (219c) composed 929c of
(locking birds. With the exception of a few boobies, most other
birds were nonbreeding visitors (Table 4). Migrants were
absent (Fig. 3). We also provide additional sightings of such
rarities in the region as Cook's petrel ( 1 ). Kermadec petrel (2),
south polar skua (5), and wedge-tailed shearwater (213)
(Table 6).

Region IV (799 birds, 18 species, 10 families)

Our cruise track passed south of the districts of Kosrae and
Pohnpei, then north of Truk, Yap, and Belau (04°N to 10°N
latitude, 163°E to 138°E longitude). Micronesia contains
fewer people on far more islands than the Gilberts. However,
fishing fleets from several foreign countries are exploiting their
oceanic waters, and the presence of even a few people on an
islet deters many seabird species from successful breeding.

Ourobservations from the Caroline Islands were dominated
by sooty terns and noddies ( 52% of birds seen), but procellariids
and boobies were represented in fair numbers. Skuas first
appeared here, along with the only migrant duck. The wedge-
tailed shearwater attained a density of 2.93 birds/10 km-
( Tabic 2). Our observations of brown boobies suggest the
possible existence of a colony on Magur Islet, Namonuito
Atoll, in the District of Truk.

Region V ( 34 birds. 6 species. 4 families)

The Philippine Sea and Basin (Fig. 2) exhibited the lowest
biodiversity and species densities (Tables 1-4) of the five
Pacific regions. This relatively small area, little-known
ornithological lv and with no islands for hundreds of kilometers.
stretches west of the Marianas to the Philippines (at 11°N
latitude, from 136°E to 125°E longitude). The nearest landfalls
arc the Marianas to the northwest. Belau, New Guinea, and the
Moluccas (Indonesia) to the south, and the Philippines to the
w est. Plankton productivity and fish populations arc known to
be considerably poorer in tropical waters north of 09°N than
further south (Ashmole & Ashmole, 1967). Our sightings of
six streaked shearwaters suggest that the southerly migration
corridor for this species lies entirely cast of the Philippines.

Region \ I (22 birds. 4 species. 6 families)

The Bohol (Mindanao) and Sulu Seas stretch from 125°E
to 1 17°E. Seabird densities were low. Phalaropes. not previously
encountered on our cruise, accounted for over half the total bird
count (Table 3). A few larids, skuas, brown noddies, and
wedge-tailed shearwaters (our most western observations)
completed the list. The proportions of resident breeders,
nonbreeding visitors, and direct migrants were equal (Fig. 4).
Elevated human populations in the Philippines have undoubtedly
reduced seabird numbers.

Region VII (65 birds, 8 species. 5 families)

Our cruise track passed through the South China Sea
between 07°N to 01°N latitude and 117°Eto 104°E longitude.
The route passed from the Balabac Strait (southwestern
Philippine Islands) indirectly to Singapore (Fig. 2). These
waters, the most heavily polluted of the trip ( see Chapter 2. this
vol.), are heavily fished commercially, and human population
densities around their periphery are high. Seabirds were
sparse. As with Region VI. phalaropes contributed half of the
total. Migrant stercorariids formed the next most common bird
grouping; since stercorariid records are few from this area, our
observations of 1 2 birds of at least 2 species add to their known
dispersal areas.

Gulls and terns were in surprisingly few numbers,
considering that almost 20 species occur in the South
China Sea. We recorded Caspian, bridled, and crested
terns (and a few unidentified larids) in very small numbers.
Resident breeders and nonbreeding visitors numbered only
one species each, and there were four species of migrants
(Fig. 3).

During our indirect. 9-day passage across the South China
Sea (23-31 October) we encountered approximately 150 land
birds (including 40 raptors), totaling at least 20 species (Ellis
et ah, 1990 and Subchapter 3.6, this vol.). The presence of
owls, nightjars, falcons, a large crake, and small forest birds
provided a highly interesting replacement for the expected
terns, gulls, and other seabirds.

This area is little known ornithologically, at least in
English publications (Delacour&Mayr. 1946; Delacour. 1947;
Anon.. 1975; Jing-Xiam & Zi-Yu, 1975; Nelson. 1978; Hails.
1987; King & Dickinson. 1987). We added three parasitic
jaegers to records for the South China Sea and a possible range
extension of the masked booby. Unfortunately, in the last few
decades numerous large seabird colonies in this area have been
destroyed by direct human predation (Nelson, 1978) but
considerable efforts towards conservation of islands, reefs,
bays, and varied habitats are on-going commitments by all
countries concerned, assisted by international agencies (UNEP,
I984a,b; IUCN. 1988a,b).

This First US-USSR Central Pacific Expedition resulted from
the efforts of main people in the United States and the Soviet Union.
On the American side, the primary organization and financial support
wore from the US Fish & Wildlife Service. We especially appreciate
the efforts of Harold J. O'Connor, Director, Patuxent Wildlife Research
Center and Steve Kohl (Office of International Affairs). We thank
H. Randolph Perry for suggesting and encouraging our participation
in the expedition, Paul Sykes for willingly assuming the Michigan

246

responsibilities of CBK during the long voyage, James Hines, Brett
Hoover, and Lois Loges for computer programming assistance, and
Bonnie J. Fancher for help in manuscript preparation. The paper
benefitted from the comments of D. Ainley. W. King, and P. Gould.

AKK thanks the other members of the ICBP 1990 Line and
Phoenix Islands Expedition (M. and A. Garnett, G. Wragg. J. Phillips,
and M. Linsley ) for use of our collective at-sea data. She also extends
special thanks to Mr. O'Connor for financial support during preparation
of the manuscript.

On the Soviet side, it is a great pleasure to thank Professor
Alia V. Tsyban, Goskomgidromet, who as leader of the expedition

provided help, interest, friendship, and outstanding hospitality during
the voyage. Captain Oleg A. Rostovtsev and his crew (especially the
navigators) of the R/V Akademik Korolev provided an excellent
observation platform and abided our frequent imposition for position
fixes from their satellite and Loran C navigation systems and
hydrographic charts. Yevgeniy N. Nelepov and Yuri L. Volodkovich
provided much assistance during the voyage, and Boris Sirenko and
Boris Alexandrov willingly shared their knowledge of benthic and
pelagic organisms. Our contacts with our Soviet colleagues would
have been far less stimulating without the translation skills of
Valeriya M. Vronskaya and Svetlana V. Petrovskaya.

3.6 Evidence for a Major Fall Land Bird

Migration Corridor Across the South China
Sea from Indo-China to the Greater Sunda
Islands

DAVID H. ELLIS. ANGELA K. KEPLER, and CAMERON B. KEPLER

US Fish & Wildlife Senice, Patu.xent Wildlife Research Center, Laurel, Maryland, USA

Introduction

Until 1960, bird migration corridors in eastern Asia were
poorly known (Wetmore, 1926; Delacour, 1947; McClure,
1974; Medway & Wells, 1976). In southeast Asia, however,
the geography of the land masses surrounding the South China
Sea seems to create natural funnels that should concentrate
migrant land birds into three primary fall corridors. Important
flight paths along some of these routes have recently been
discovered.

It is known that migrants from Japan and eastern China
island-hop south through the Philippines (Wetmore. 1926;
McClure, 1974), with Ng ( 1978) presenting evidence that barn
swallows (Hirundo rustica) move directly from mainland
China to the Philippines. McClure ( 1974) asserted that many
migrants passing through the Philippines to Borneo fly west
from Palawan, then south to Borneo. Simpson (1983a,b)
encountered hundreds of migrant birds at the Tembungo offshore
oil drilling platform near the northeastern tip of Borneo (Fig. 1 )
during the fall migration of 1981. Although he reported these
observations as evidence of a passage directly across the South
China Sea, his location near Balabac Strait also suggests that
these migrants could have been moving south from the
Philippines.

Geography suggests that many migrant land birds in
Burma and western Thailand would move south along the
Malay Peninsula, a pathway known to be important ( McClure,
1974; Medway & Wells, 1976; Hails, 1987), and thence across
the narrow Straits of Malacca to Sumatra. However, migrants
from east Thailand as well as those from China, Laos, and

Vietnam, moving down the Indo-China Peninsula would
naturally converge south of the Mekong River Delta on
Mui Bai Bung. From this tip of the Indo-China Peninsula,
birds traveling overland must fly northwest into Thailand
before proceeding south. Those capable of a relatively short
(ca. 400 km) overwater flight can fly southwest across the Gulf
of Thailand toward the Malay Peninsula, a route portrayed by
McClure (1974) and Hails (1987) as a minor pathway for the
migrants from Indo-China.

McClure (1968, in McClure, 1974) illustrated a coastal
migration route from Taiwan to northern Vietnam, thence
south, crossing the South China Sea to Borneo, another route
suggested by geography. However, he provides scant evidence
for such a corridor and no evidence that migrants are
concentrated at Mui Bai Bung. McClure ( 1 974) discussed a fall
passage of willow warblers (Phylloscopus sp.) across the South
China Sea to Sarawak without offering details on their point of
origin north of the sea. Simpson (Wells, personal
communication and in prep.) reported a substantial fall
movement of land birds ( 36 species ) in the Terengganu oil field
(ca. 05°25'N, 105°13'E, see Fig. 1 ). Although this location is
only about 200 km east of the Malay Peninsula and west of a
direct route to Borneo, Simpson's records provide the best
evidence to date of a direct South China Sea crossing. The birds
observed by Simpson (1983a) at the Tembungo oil terminal
could have come from Vietnam, as he suggests, but the source
of these migrants is clouded by their proximity to the Philippines.
Although biologists from the Chinese Academy of Science and
the Beijing Natural History Museum (Anon., 1974) noted
44 species of land birds during 1974 surveys of islets in the

247

South China Sea

Fig.l. Geography of the southern half of the South China Sea showing bird
survey locations. Numbered segments are bird location survey
locations for R/V Akademik Korolev, 23-310ctober 1988.

northern two-thirds of the South China Sea, demonstrating the
potential for long-distance (ca. 1,000 km) migration, they did
not demonstrate that a corridor for land-bird migrants exists
further south between Vietnam and Borneo. In this paper, we
present data from the South China Sea that strongly support the
presence of such a migration route.

Study Area and Methods

We encountered migrant land birds during our
23-31 October 1988 indirect passage (Fig. 1) from Balabac
Strait to Singapore on the Soviet research vessel Akademik
Korolev. While in transit, we observed birds during dawn-to-
dusk seabird surveys from the flying bridge (12 m above sea
le\ el ). During a 3-day period while the ship was anchored or
drifting without power to conduct oceanographic research
(Fig. 1, Station 13;06°01'N, 106°55'E), we conducted periodic
v. alking inspections of the ship (usually at half-hour intervals)
and searched the ship each night by flashlight to count roosting
birds. Five raptors anil several barn swallows were captured by
hand (primarily at night) and examined for physical condition.

Results and Discussion

During our 9-day passage, we encountered about 1 50 land
buds 1 121 by conservative count, 84 minimum count, Table 1 1
representing 14 families. Almost all were migrants that winter
(at least in part) south of the South China Sea. Most of these
birds (9ft by conservative count ) arrived on the ship during the
! da) period while we were stationary (Fig. 1, Station 13)
around 350 km southeast of the southern tip of [ndo-China.
The presence oi land buds at (Ins location suggests that they
were in passage across the South China Sea from Indo-Chin.i
to the ( ireater Sunda Islands. The low bird counts seen before

arriving at and after leaving this location (Table 2) suggest that
this spot lies on a rather narrow migratory pathway although,
alternately, birds may have been reluctant to approach a moving
vessel. Simpson's 1982 observations (Wells, in prep.), made
in the Terengganu oil field (Fig. 1 ) very near our cruise track,
suggested that he was sampling the same corridor we visited;
if so, the pathway may be somewhat wider than we detected.

The number of birds we observed (Tables 1,2) is small
when compared with record counts for well-known migration
pathways. However, our visit was brief and probably too late
for detecting the bulk of migrating land birds. Simpson's
( 1983a) dates for six of nine frequently encountered land birds
near northeastern Borneo fell before the time of our visit, and.
just as important, migrating land birds most often aggregate
where land and water configurations encourage them to collect
(e.g.. on north or south projecting peninsulas). By contrast, we
were on the open sea where birds are much less likely to
concentrate. Considering these factors, it seems likely that
adequate spatial and temporal sampling will reveal many
thousands of land birds moving south from Indo-China across
the South China Sea.

Although our records and those of Simpson's (Wells, in
prep.) demonstrate that a sizable migration is probably normal
across the South China Sea. we should mention an alternate
hypothesis that may help explain the presence of these birds
where and when we observed them. First, our passage occurred
when Typhoon Ruby was ravaging the Philippine Islands
(Anon.. 1989). Although we did not encounter heavy seas or
strong winds, some of the birds we observed may have been
forced out to sea, if nonmigratory. or shunted away from their
normal migration route, if migratory, by the storm. However,
most of the birds we observed far from land (Table 3) are
known to be strong migrants. The four hawks tentatively
identified as shikras (Accipiter badius) and crested goshaw ks
(A. trivirgatus), and the dove (Streptopelia sp. ) are the only real
surprises, although a few others in Table 3 would not be
expected this far from land.

Flight direction may give some indication of the likelihood
of either hypothesis. If the birds were displaced migrants, they
would probably have been heading southwest (i.e.. away from
the storm). If in passage from a concentration zone on the Indo-
China Peninsula to Borneo, they should have been heading
southeast to encounter our vessel. If. as we observed, the
raptors (33'f of all land birds) were foraging at sea (Ellis etal.,
1 990 ) rather than migrating, there would likely be no consistent
trend in their flight direction. In Fig. 2. there is no clear east-
west trend in arriving or departing flights. However, although
the data are very few strong southward and westward
components are e\ idem. In constructing Fig. 2. we eliminated
directional readings for birds seen on cruise track segments
1-7 and 20-30 because these segments were near enough to
land (i.e., within 1 00 km ) that the birds' flight directions could
have been influenced by sight or sign of nearby land. In
addition, all Right bearings could have been influenced h\ the
presence of the ship.

Physical condition of the birds w e obsen ed mas also be an
indicator of the regularity with which this migration route is
used. If a high proportion of the know n o\ erseas migrants were

248

TABLE 1

Land bird totals for R/V Akademik Korolev cruise track segments and stationary watches in the South

China Sea, October 1988.

No. Land Birds Observed'

Date

23

24

25

Station/Segment2

Duration

All Land

Length

ofObs.

Rapti

3rs

Non-

•raptors

Birds

Number

(km)

(min.)

Min.

Cons.

Min.

Cons.

Min.

Cons.

1

38

88

0

0

0

0

0

0

2

15

45

0

0

0

0

0

0

3

6

17

0

0

0

0

0

0

4

67

158

0

0

0

0

0

0

5

14

44

0

0

0

0

0

0

6

19

48

0

0

0

0

0

0

7

49

109

0

0

1

1

1

1

8

22

50

0

0

0

0

0

0

9

3

11

0

0

0

0

0

0

10

30

67

0

0

0

0

0

0

11

17

38

0

0

0

0

0

0

12

63

143

0

0

4

4

4

4

26-28 13 ca. 0 912 22 30 32 44 54 74

28

14

ca. 0

43

1

1

0

0

1

1

15

22

10

0

0

0

0

0

0

16

26

78

0

0

0

0

0

0

17

ca. 0

195

1

4

4

8

5

12

29

18

ca. 0

105

1

1

1

4

2

5

19

ca. 0

20

0

0

0

0

0

0

20

25

62

0

0

0

1

0

1

21

ca. 0

80

0

0

0

0

0

0

30

22

ca. 0

23

1

1

3

3

4

4

23

ca. 0

IK)

0

0

2

4

2

4

24

ca. 0

25

0

0

0

1

0

1

25

ca. 0

95

0

1

3

3

3

4

26

ca. 0

54

0

0

0

0

0

0

27

ca. 0

52

0

0

3

5

3

5

28

ca. 0

30

1

1

0

0

1

1

31

28

ca. 0

15

0

0

0

0

0

0

29

ca. 0

15

0

0

0

0

0

0

30

135

280

1

1

3

3

4

4

TOTALS

3.022

28

40

56

81

84

121

1 Because accurate bird counts were sometimes difficult to obtain for stationary watches (i.e.. some
birds remained aboard or flew about the ship for extended periods), we report both the minimum
(min.) number of birds observed (based on subtractive values) and a conservative (cons.) number
based primarily on new arrivals. The actual number observed is believed to be about 20% higher than
the conservative count.

: Cruise track segment locations are illustrated in Fig. 1.

249

TABLE 2

Minimum and conservative land bird counts along cruise track of R/V Akademik Korolev in the
South China Sea, 23-31 October 1988. ' :

Species

Small juv. accipiter (Accipiter sp. )

Ad. Japanese sparrow-hawk (Accipiter gularis)

Ad. shikra (A. badius)3

Ad. crested goshawk (A. trivirgatusf

Eagle/kite (Accipitridae)

Peregrine falcon (Falco peregrinus)

Oriental scops owl (Otus siiniaf

Chinese pond heron (Ardeola bacchus)

Watercock {Gallicrex cinerea)

Dove [Streptopelia sp.)'

Grey nightjar (Caprimulgus indicus)

Fork-tailed swift (Apus pacificus)

Swift (Apodidae)

Dollarbird (Eurystomus orientalis)

Barn swallow (Hirundo rustica)

Swallow (Hirundo sp. )

Ashy minivet (Pericrocotus divaricatusf

Lanceolated Warbler (Locustella lanceolata)4

Great reed warbler (Acrocephalus arundinaceus i

Warbler (Acrocephalus sp.)

Arctic warbler (Phylloscopus boreal is)

Flycatcher (Ficedula sp.)

Brown shrike (Lanius cristatus)

Unidentified passerines or remains

TOTALS

Segments 1-7

Segments 8-19

Segments 20-30

(Oct.

23-24)

(Oct.

25-29)

(Oct

. 29-3 1 )

.Jin.

Cons.

Min.

Cons.

Min.

Cons.

0

0

14

24

3

3

0

0

2

2

0

0

0

0

1

->

0

0

0

0

2

2

0

0

0

0

1

1

0

0

0

0

3

3

0

0

1)

0

2

2

0

1

1)

0

0

0

3

3

0

0

1

1

0

0

0

0

1

1

0

0

0

0

1

1

1

1

0

0

4

4

0

0

0

0

2

2

0

0

1

1

0

0

0

0

0

0

14

29

5

10

0

0

1

1

0

0

0

0

1

1

1

1

0

0

1

1

1

1

0

0

1

1

0

0

0

0

1

2

1

1

0

0

0

0

1

1

0

0

2

2

0

0

0

0

6

6

1

1

0

0

5

8

0

1

1

1

66

96

17

24

'Cruise track segments and stations are illustrated in Fig. 1 and described in Table I .

'Abbreviations in column headings are: Min. (minimum count) and Cons, (conservative count) as explained in Table I.
Footnote 1 .

"Because these birds are considered non-migratory, these identifications should be treated as tentative. All are based on
nearby visual observations aided by lOx binoculars, but without photographic or other substantiation.

individuals of these species were deposited in the US National Museum: Oriental scops owl. USNM No. 607190; Ashj
minivet. USNM No. 607193; and Lanceolated warbler, spirit specimen (not assigned numbers at USNM l.

in good body condition this far from land, it is more tenable to
suppose that these species regularly use this route. In Table 3,
our best estimate of physical condition is compared for all
species that we encountered far from land. We know from
handling a few captives, and infer from the energetic flight of
others, that the raptors and the barn swallows at Station 13
(Fig. I ) were in good physical condition. For the other species,
too few individuals were present to draw firm conclusions, but
all birds of most species appeared to be in good condition.

A final hypothesis may explain the presence of some of the
raptors. Many were opportunistically foraging at sea. During
our 3-day layover dig. I, Station 13), we recorded raptors
perching for extended periods, roosting nightly on the ship, and

engaging in at least 21 hunting forays (Ellis et ai. 1990). Of
14 forays for which the outcome was known, 13 (93%) were
successful. Some accipiters even used the ship's deck lights to
forage at night. We gathered prey remains, totaling at least
20 kills. Two species, barn swallow and brown shrike (Lcmius
cristatus), suffered heavy mortality from predation. Of 14 barn
swallows (minimum count) observed from 25-29 October, at
least 7 turned up as prey. Even more significant, five of six
(minimum count) brown shrikes seen during the same 5-day
period were observed as prey. Simpson's (1983a; Wells, in
prep. ) observations of raptor behavior at both oil fields led him
to conclude that Japanese sparrow-hawks were hunting and
"commuting between nearby rigs." Our observations confirm

250

TABLE 3

Physical condition and migratory status of land birds arriving on R/V Akademik Korolev, 25-29 October 1988, South China Sea.1

Physical
Condition"

Mobility Classes2

Taxon [Number]

Known Known Known

Known Over-water Colonizer Straggler
Migrant Migrant5 of Islands to Islands Comments

Good

Good

Japanese sparrow-hawk +

(Accipiter gularis) [2]

Shikra (A. badius) [2] +

Common migrant

Western population is highly migratory;
eastern population migratory in Malaysia

Good Crested goshawk

{A. trivirgatus) [2]

Good Peregrine falcon

{Falco peregrinus) [3]

Non-migrant throughout range

Fatigued

Watercock

(Gallicrex cinerea) [ 1 ]

Good

Dove (Streptopelia sp.) [ 1

Emaciated

Oriental scops owl

(Otus sunia) [1]

Good

Gray nightjar

(Caprimulgus indicus) [ 1

Winters in Greater Sunda Islands and Celebes;
very few records as straggler

Strongly migratory, scatters across Malaysia in
winter

Good Fork-tailed swift

(Apus pacificus) [4]

Good Barn swallow

(Hirundo rustica) [29]

Fatigued Ashy minivet (Pericrocotus

divaricatus) [1]

Fatigued Lanceolated warbler

{Locustella lanceolata) [1 ]

A few migratory stragglers recorded as far east
as Marshall Islands

Winters throughout region and tropics
worldwide

Winters on larger islands of Indonesia, but not
on islands separated by large bodies of water

Winters in Greater Sunda Islands

Good Great reed warbler (Ac ro-

cephalus arundinaceus) [ 1 ]

Good Brown shrike

[Lanius cristatus) [6]

Common migrant in Indonesia

Common migrant to Greater Sunda Islands;
recorded in Palau

!Data are included only for that portion of the cruise track (stations 8-19) where the ship was far (>100 km) from land.

"Symbols in these columns: + = yes.- = no. Assignment to mobility class (i.e., regular migrant over land and over large bodies of water
>500 km. colonizer of distant land masses and islands as a breeding bird, straggler either on migration or as a resident) is at best tentative
for some species, but was derived from information in Brown & Amadon (1968), Clements (1978), King & Dickinson (1975), Medway &
Wells (1976), and Pratt et al. (1987).

'Physical condition was reported "Good" if bird flew well and was adept at avoiding capture by hand, "Fatigued" if readily captured by
hand, and "Emaciated" if sternum was sharply protruding upon capture.

251

N

i

A. ACCIPITERS IN =17)

B PASSERINES and SWIFTS i N = 7 I

o

Fig.

C. BARN SWALLOWS IN=4I

Arriving and departing flight directions for birds seen on cruise track
segments 8-19 of R/V Akademik Korolev, 26-29 Oclober 1988.

that the raptors were opportunistically using our stationary ship
for perching, roosting, hunting, and eating. When the ship was
moving under power, however, none of the raptors perched for
any extended period, and none roosted on the ship.

From all available evidence, it seems most likely that the
birds we encountered were a small part of what must be a
si/able wave of fall migrants on their way across the South
China Sea. The configuration of the land masses suggests that
the point of departure for these birds was the southern tip of
Indo-China; however, further land-based research is needed to
substantiate the point of origin and destination of birds crossing
the South China Sea to Borneo. Additional work at sea will also
be helpful in determining the timing and magnitude of the
migration, as well as corridor w idth. Work on islands in the
South China Sea or stationary platforms may substitute in part
for the at-sea studies, but it is also important to determine body
condition of birds arriving in the Greater Sunda Islands.
Intensive banding operations in Vietnam, at sea. and in Borneo
could reveal much about survival rates and all other aspects of
this little-known migration route.

This project was part of the First US-USSR Joint Pacific
Expedition to the Pacific Ocean and South China Sea. We express
appreciation to the agents of both governments who made our
participation possible. We especially thank Professor Alia V. Tsyban
(Chief Scientist). Harold J. O'Connor (Director. Patuxent Wildlife
Research Center and US organizer of the expedition ). Captain Oleg A.
Rostovtsev, and our many Soviet friends whoenhancedourenjo) ment
of the trip. Joe Marshall and Ralph Browning assisted in specimen
identification at the Smithsonian Institution. Linda J. Miller. Cath)
Ellis, and Bonnie Fancher assisted greatly in data handling and
manuscript preparation.

The following people commented on an earh version of this
manuscript and thereby compensated for our inexperience w ith the
birds of the South China Sea region: H. Elliott McClure. Duncan
Parish. Philip D. Round. Lucia Liu Severinghaus. and David R. Wells.
We appreciate additional editorial comments from Mark R. Fuller.
George F. Gee. and Gary H. Heinz.

252

Chapter 3 References

Ainley, D. G. & Boekelheide, R. J. ( 1983). An Ecological

Comparison of Oceanic Seabird Communities of the Souith

Pacific Ocean. Stud. Avian Biol. 8. 2-23.
Amerson. A. B., Jr. ( 1969). Ornithology of the Marshall and

Gilbert Islands. Atoll Res. Bull. No. 127. 1-348.
Anonymous ( 1974). Report of the zoological survey of the

islands of the South China Sea. ACTA Zoologica Sin. 20(2),

113-130.
Anonymous (1989). Marine weather review. Mariners

Weather Log 33(2), 40-50.
AOU. (1983). Check-list of North American Birds. 6th Ed.

American Ornithologists' Union, Washington, D.C.
AOU. (1985). Thirty-fifth supplement to the American

Ornithologists' Union Check-list of North American Birds.

Auk 102. 680-686.
Arashkevich, E. G. (1972). Vertical distribution of various

trophic groups of copepods in boreal and tropical regions of

the Pacific Ocean. Oceanology 12(2). 315. (in Russian)
Ashmole, N. P. & Ashmole, M. J. (1967). Comparative

feeding ecology of seabirds of a tropical oceanic island.

PeabodyMus. Nat. Hist. Yale Univ. Bull. 24. 1-131.
Bailey, S. F., Pyle, P. & Spear, L. B. (1989). Dark Pterodroma

petrels in the North Pacific: Identification, status, and North

American occurrence. American Birds 43. 400-415.
Baker, R. H. (1951). The avifauna of Micronesia: Its origin,

evolution, and distribution. Univ. Kans. Publ. Mus. Nat.

Hist. 3(1). 1-359.
Beklemishev. K. B. (1973). Ranges of tropical far-neritic

species. In Tropical Zone of the World Ocean and the

Global Processes Associated with It, pp. 292-299. Nauka

P. jlishers. Moscow, (in Russian)
Bloxham, A. ( 1925). Diary of Andrew Bloxham. naturalist of

the "Blonde" on her trip from the Hawaiian Islands from

England, 1824-1825. B. P. Bishop Mus. Spec. Publ. 10.
Boaden. P. J. S. & Seed, R. (1985). An Introduction to Coastal

Ecology. Blackie and Son Ltd., Glasgow and London. 218

pp.
Briggs. K. T„ Tyler, W. B.. Lewis, D. B. & Carlson, D. R.

(1987). Bird communities at sea off California: 1975 to

1983. Stud. Avian Biol. 11. 74 pp.
Brown, L. & Amadon, D. ( 1968). Eagles, Hawks and Falcons

of the World. Country Life Books, London.
Chislenko. L. L. (1968). Nomograms for Determining the

Weight of Aquatic Organisms from Body Size and Shape.

Nauka Publishers. Leningrad, (in Russian)
Clapp. R. B.( 1967). Line Islands. Unpublished report. 32 pp.

USNM, Smithsonian Institution, Washington. D.C.
Clapp. R. B. & Sibley. F. C. ( 1 967 ). New records of birds from

the Phoenix and Line Islands. Ibis 109. 122-125.
Clapp. R. B. & Sibley. F. C. (1968). Additional new records

of birds from the Phoenix and Line Islands. Ibis 110.

573-575.
Clapp. R. B. & Sibley. F. C. (1971a). Notes on the vascular

flora and terrestrial vertebrates of Caroline Atoll. Southern

Line Islands. Atoll Res. Bull. 145. 1-16.

Clapp. R. B.& Sibley. F. C. (1971b). The vascular flora and

terrestrial vertebrates of Vostok Island, south-central Pacific.

Atoll Res. Bull. 144, 1-9.
Clements, J.F.I 1978). Birds of the world: A check list. Two

Continents Publishing Group, New York.
Crossin, R. S. (1974). The storm-petrels (Hydrobatidae). In

Pelagic Studies of Seabirds in the Central and Eastern

Pacific Ocean (W. B. King, ed. ), pp. 154-205.
Day, R. A., Wehle, D. H. S. & Coleman, F. C. ( 1984). Ingestion

of plastic pollutants by marine birds. In Proceedings of

the Workshop of the Fate and Impact of Marine Debris,

26-29 November 1984, Honolulu, Hawaii, pp. 344-386.

NOAA Technical Memorandum SWEC-54, NMFS,

NOAA-TA, US Department of Commerce.
Delacour, J. (1947). Birds of Malaysia, xvi + 382 pp.

Macmillan Co., New York.
Delacour, J. & Mayr, E. (1946). Birds of the Philippines,

xv + 309 pp. Macmillan Company, New York.
Ellis, D. H., Kepler, A. K. & Kepler. C. B. (1990). Evidence

for a fall raptor migration pathway across the South China

Sea. J. Raptor Res. 24(1-2), 12-18.
Engbring, J. (1983). Avifauna of the southwest islands of

Palau. Atoll Res. Bull. 267, 1-22.
Garnett, M. C. (1983). A Management Plan for Nature

Consen'ation in the Line and Phoenix Islands. Pt. I:

Description, 318 pp.; Pts. II, III: Policy and Recom-
mendations. 131 pp.
Garnett, M. C. (1984). Conservation of seabirds in the South

Pacific region: A review. ICBP Tech. Pub. 2, 547-557.
Geinrikh, A. K. (1960). The near-surface plankton of the

central World Ocean. Oceanology 11. (in Russian)
Geinrikh, A. K. (1964). On the surface plankton of the

northeastern Pacific. Trudy Institute of Oceanology AN

SSSR 65, 77-94. (in Russian)
Geinrikh. A. K. (1968). Quantitative distribution of certain

plankton animals in the western Pacific Ocean. In Plankton

of the Pacific Ocean. Gidrometeoizdat Publishers, Moscow.

(in Russian)
Gilbert and Ellice Islands Government. (1974). Line Islands

Expedition report, 89 pp. Unpublished manuscript.
Gill, R. E. (1990). Ecology of curlews in French Polynesia

during spring. In Summary of the Proceedings from the

Workshop on Bristle-thighed Curlews. Anchorage. Alaska

(R. E. Gill & C. M. Handel, eds. ). 22 pp.
Gorshkov. S. G. (1974). Atlas of the Oceans. The Pacific

Ocean. Head Department of Navigation and Oceanography,

Ministry of Defense of the USSR.
Gould, P. J. (1974a). Introduction, pp. 1-5. In Pelagic

Studies of Seabirds in the Central and Eastern Pacific

Ocean (W. B. King, ed.), 277 pp. Smithson. Contrib.

Zool. 158.
Gould. P. J. (1974b). Sooty tern (Stemafuscata), pp. 6-52. In

Pelagic Studies of Seabirds in the Central and Eastern

Pacific Ocean ( W. B. King. ed. ). 277 pp. Smithson. Contrib.

Zool. 158.

253

Gould, P. J. (1983). Seabirds between Alaska and Hawaii.

Condor 85, 286-291.
Gould, P. J. & Forsell, D. J. ( 1989). Techniques for shipboard

surveys of marine birds. US Fish Wildl. Serv. Tech. Report

25. 1-22.
Gould, P. J. & King. W. B. ( 1967). Records of four species

of Pterodroma from the central Pacific Ocean. Auk 84.

591-594.
Gould, P. J., King, W. B. & Sanger. G. A. (1974). Red-tailed

tropicbird {Phaethon rubricauda), pp. 206-231. In

Pelagic Studies of Seabirds in the Central and Eastern

Pacific Ocean ( W. B. King. ed. ). 277 pp. Smithson. Contrib.

Zool. 158.
Grossman, H. & Grossman, H. ( 1974). Preliminary Report

(Birds). In Line Islands Expedition Report, August-October

1974, 5 pp. Gilbert and Ellice Islands Government.

Unpublished manuscript.
Haile, N. S. (1964). SabakSoc.J. 11. 135-137.
Hails. C.( 1987). Birds of Singapore, 168 pp. Times Editions,

Singapore.
Haney. J. C. (1985). Counting seabirds at sea from ships:

comments on interstudy comparisons and methodological

standardization. Auk 102(4). 897-898.
Harrison. P. (1985). Seabirds: An Identification Guide,

448 pp. Houghton Mifflin Company. Boston.
Herring, J. L. (1961). The genus Halobates (Hemiptera,

Gerridae). Pac. Insects 3(2-3). 223-305.
Huber. L. N. (1971). Notes on migration of Wilson's storm-
petrel {Oceanites oceanicus) near Eniwetok Atoll, western

Pacific Ocean. Notornis 18( 1 ). 38-42.
Humphrey. P. S. (1965). An ecological survey of the Central

Pacific. Smithson. Year 1%5. 24-30.
Hunt, G. L., Eppley. Z., Burgeson. B. & Squibb. R. (1981).

Reproductive ecology, foods and foraging areas of seabirds

nesting on the Pribilof Islands, 1975-1979. In Final Reports

of Principal Investigators, Vol. 12. pp. 1-258. NOAA.

Boulder. Colorado.
ICBP. ( 1990). Unpublished field notes. International Council

for Bird Protection (ICBP) 1990 Line and Phoenix Islands

Expedition. In author's possession.
IUCN. (1988a). Coral Reefs of the World. Vol. 2. United

Nations Environmental Program. Intemation Union for

Conservation of Nature and Natural Resources. Compiled

by C. Sheppard & S. M. Wells. 389 pp.
IUCN. ( 1988b). Coral Reefs of the World. Vol. 3. United

Nations Environmental Program. Intemation Union for

Conservation of Nature and Natural Resources. Compiled

by C. Sheppard & S. M. Wells. 329 pp.
Jenkins, 1. A. F. ( 1979). Observations on the wedge-tailed

shearwater {Puffinus pacificus) in the south-west Pacific.

Notornis 26. 331-348.
Jing-Xian, L. & Zi- Yu. W. ( 1 975 ). The red-footed booby in the

\isha i Paracel I Islands. (Translated from Chinese bj [aian

C. On'. ) Zool. Mag. (Dongwuxue Zazhi) 3. 40.
Kepler. A. K. (1490). ICBP 1990 Line and Phoenix Islands

1 \pedition. 20 February- 31 May 1990, 12pp. Unpublished

report, Cambridge. England.

Kepler. C. B., Kepler. A. K. & Ellis. D. H. ( 1992). Ecological

studies on Caroline Atoll. Republic of Kiribati, south-central

Pacific Ocean. Part II. Seabirds, other tenestrial animals.

and conservation. (Subchapter 1. 2, this volume).
King. B. F. & Dickinson. E. C. (1975). A Field Guide to the

Birds of South-East Asia, 480 pp. Houghton Mifflin Co..

Boston.
King. J. E. & Hida. T. S. ( 1957 ). Zooplankton abundance in the

Central Pacific. Pt II. US Fish Wildl. Serv. Fish. Bull.

57(118), 365-395.
King. W. B. ( 1967). Seabirds of the Tropical Pacific Ocean.

126 pp. Smithsonian Institution, Washington. D.C.
King, W. B. (1970). The Trade Wind Zone Oceanography

Pilot Study. Part VII: Observations of Sea Birds, March

1964 to June 1965, 136 pp. US Fish Wildl. Sen: Spec. Sci.

Rep. Fish. 586.
King, W. B. ( 1973). Conservation status of birds of central

Pacific islands. Wilson Bull. 85. 89-103.
King. W. B. (ed.) (1974a). Pelagic Studies of Seabirds in the

Central and Eastern Pacific Ocean. Smithson. Contrib.

Zool. 158.
King. W. B. (1974b). Wedge-tailed shearwater {Puffinus

pacificus), pp. 53-95. In Pelagic Studies of Seabirds in the

Central and Eastern Pacific Ocean, (W. B. King. ed).

277 pp. Smithson. Contrib. Zool. 158.
Korshenko, A. N. (1988). Ecological Conditions of the Eastern

Equatorial Region of the Northern Pacific Ocean.

Gidrometeoizdat Publishers, Moscow, (in Russian)
MacDonald. J. D. & Lawford, P. A. ( 1954). Sight records of

birds in the Pacific compiled from the bird log kept during the

recent cruises of the H. M. S. Challenger. Emu 54( 1 ). 7-28.
Marchal. E. ( 1 959 ). Analyse de quelques contrenus stomacaus

de Neothunnus albacora (Lowe). Bull. Inst. Afrique Noire

B(a). 1123-1136. (in French)
Marks. J. S., Redmond. R. L., Hendricks. P.. Clapp. R. B. &

Gill. R. E. (1990). Notes on longevity and flightlessness in

bristle-thighed curlews. Auk 107(4). 779-781.
Mayr. E. ( 1 945 ). Birds of the Southwest Pacific: A Field Guide

to the Birds of the Area Between Samoa, NewCaledonia, and

Micronesia. 316 pp. Macmillan Company. New York.
McClure. H. E. ( 1974). Migration and Survival of the Birds of

Asia. US Army Medical Component, SEATO Medical

Project. Bangkok. Thailand.
Medway, L. & Wells. D. R. (1976). Birds of the Malax

Peninsula, Vol. V. Witherby, London.
MorzerBruyns. W. F.J. ( 1965). Birds seen during west to east

trans-Pacific crossing along Equatorial Counter-current

around latitude 7°N, in the autumn of 1 960. Sea Swallow 17.

57-66.
Murphy, R. C. ( 1936). Oceanic Birds of South America. Am.

Mus. Nat. Hist.. New York, pp. 1-1245.
Murphy. R. C. ( 195 1 ). The populations of the wedge-tailed

shearwater {Puffinus pacificus). Am. Mus. Novit. 1512.

1-21.
Murphy. R. C. & Snyder. J. P. ( 1 952 ). The "Pealea" phenomena

and other notes on storm-petrels. Amer. Mus. Novit. 1596.

1-16.

254

Nelson. J. B. (1975). Breeding biology of frigatebirds — a

comparative review. The Living Bird 14, 1 13-156.
Nelson. J. B. (1978). The Sulidae: Gannets and Boobies.

Oxford University Press, Oxford, England. 1. 012 pp.
Ng. S. C. ( 1978). Migratory barn swallows (Hirundo rustica)

in the South China Sea. Malay. Nat. J. 32, 67-73.
Nicholson, E. M. & Douglas, G. ( 1969). Draft checklist of

Pacific oceanic islands. Micronesica 5(2), 327-463.
NID. ( 1 945 ). Pacific Islands. Vol. IV. Western Pacific, 526 pp.

Great Britain Naval Intelligence Division, B. R. 519C

Geographical Handbook Series, Butler & Tanner, Ltd.,

London.
Niering, W. A. & Sachet, M. -H. (1961). Observations on

Puluwat and Gaferut. Caroline Islands, with historical and

climatic information on Gaferut Island. Atoll Res. Bull. 76,

1-15.
Osipov. N. N. & Charykov, A. K. (1987). Concentration of

heavy metals present in sea water by films of petrolic origin.

In Geographic and Economic Problems Involved in the

Investigation and Harnessing of the Southern Seas of the

USSR. Abstracts: the Third Ail-Union Conference on Ocean

Geography and Cartography (Rostoc-na-Donu, May 1987),

Leningrad, (in Russian)
Owen, R. P. ( 1 97 1 ). The birds of the Palau Islands including

Helen Reef, and Palau bird name list, 10 pp. Trust Territory

of the Pacific Islands. Office of the Chief Conservationist,

Biology Laboratory, Koror, Palau, Caroline Islands.

Unpublished report.
Owen, R. P. (1977a). New bird records for Micronesia and

major island groups in Micronesia. Micronesica 13( 1 ).

57-63.
Owen. R. P. (1977b). A checklist of the birds of Micronesia.

Micronesica 13( 1 ), 65-81.
Perry. R. (1980). Wildlife conservation in the Line Islands,

Republic of Kiribati (Gilbert Islands). Envir. Cons. 7,

311-318.
Pratt. D. H., Bruner. P. L. & Berrett, D. G. (1987). The Birds

of Hawaii and the Tropical Pacific. Princeton University

Press. Princeton. New Jersey.
RNZAF. (1986). Aerial photos of Maiden Island, SN 657 1 .

Royal New Zealand Air Force, Dept. of Survey & Land

Information, Wellington, New Zealand.
Romanenko, V. I. & Kuznetsov, S. I. (1974). Ecology of

Freshwater Microorganisms. A Laboratory Handbook.

Nauka Publishers. Leningrad. 194 pp. (in Russian)
Savilov. A. I. ( 1969). Pleuston of the Pacific Ocean. In Tikhii-

Pacific Ocean. Part II. Deep-water Benthic Fauna. Pleuston,

pp. 264-353. Nauka Publishers, Moscow, (in Russian)
Schreiber. R. W. & Ashmole, N. P. (1970). Seabird breeding

seasons on Christmas Island, Pacific Ocean. Ibis 112,

363-394.
Schreiber. R. W. & Schreiber. E. A. ( 1984). Central Pacific

seabirds and the El Nino Southern Oscillation: 1982 to 1983

perspectives. Science 225. 7 1 3-7 1 6.
Sherman. K. (1963) Pontellidcopepod distribution in relation

to surface water types in the central North Pacific. Limnol.

Oceanogr. 8(2). 214-227.

Sibley, F. C. & Clapp, R. B. ( 1 967). Distribution and dispersal

of central Pacific lesser frigatebirds Fregata ariel. Ibis 109.

328-337.
Simpson, D. M. ( 1983a). Autumn migration of land birds off

North Borneo in 1981. Sea Swallow 32, 48-54.
Simpson, D. M. ( 1 983b). Birds seen at the Tembungo gas flare.

North Borneo during the development of Typhoon Clara.

Sea Swallow 32, 82-83.
Sorokin. Yu. I. (1971a). Assessment of the role of

bacterioplankton in the productivity of Pacific Ocean waters.

In The Functioning of Pelagic Communities in Tropical

Ocean Waters, pp. 92-122. Nauka Publishers, Moscow, (in

Russian)
Sorokin, Yu. I. ( 1 97 1 b). Quantitative assessment of the role of

bacterio-plankton in the biological productivity of the tropical

Pacific. In The Functioning of Pelagic Communities in

Tropical Ocean Waters. Nauka Publishers. Moscow, (in

Russian)
Sorokin. Yu. I. ( 1973). Productivity of bacterioplankton in the

western Pacific. Oceanology 13( 1 ), 95-106. (in Russian)
Sorokin, Yu. I. ( 1976). Primary production in seas and oceans.

In Science and Technology Findings. General Ecology.

Biocenology. Hydrobiology. Vol. 1. VINITI. Moscow, (in

Russian)
Sorokin. Yu. I. ( 1978). Description of primary production and

heterotrophic microplankton in the Peruvian upwelling

region. Oceanology 18(1 ). 97-1 10. (in Russian)
Sorokin, Yu. I. (1982). The role of microheterotrophic

organisms in the functioning of marine ecosystems. Usp.

Sovrem. Biol. 93(2), 236-252. (in Russian)
Sorokin, Yu. I., Korsak, M. N„ Mamaeva, T. I. & Kogel'shchitu,

D. D. ( 1983). Primary production and activity of microflora

in Peruvian upwelling areas. In Bioproductivity of Upwelling

Ecosystems, pp. 74-87. Moscow, (in Russian)
SPREP. ( 1 989). The Northern Marshall Islands natural diversity

and protected areas survey, 7-24 September 1988, 133 pp.

South Pacific Regional Environment Program. Noumea.

New Caledonia, and East-West Center. Honolulu, Hawaii.
Stepan'yants, S. D. (1977). Siphonophores of the central

Pacific Ocean. In Marine Plankton (Systematics and

Faunistics), pp. 54-81. Leningrad, (in Russian)
Stoddart, D. R. (MS). Scientific importance and conservation

of Central Pacific Islands, 28 pp. Dept. of Geography,

Cambridge University, England. Unpublished typescript

manuscript.
Tasker. M. L.. Jones. P. H„ Dixon, T. & Blake, B. F. ( 1984).

Counting seabirds at sea from ships: A review of methods

employed and a suggestion for a standardized approach. Auk

101(31,567-577.
Tsyban, A. V. (ed. ) (1980). Handbook of Methods for the

Biological Analysis of Sea Water and Benthic Deposits

(1980). Gidrometeoizdat. Leningrad. 191pp. (in Russian)
Tsyban, A. V.. Panov, G. V.. Subbotin, V. G.. Timoshenkova,

N. P. & Chernyak, S. M. (eds.). (1988). Methodological

Foundations of Integrated Ecological Monitoring of the

Ocean. Gidrometeoizdat Publishers, Moscow. 287 pp. (in

Russian)

255

UNEP. (1984a). Marine Environmental Centres: East Asian

Seas. United Nations Environment Programme, Regional

Seas Directories and Bibliographies, Rome, Italy. 139 pp.
UNEP. (1984b). Marine Environmental Centres: South Asian

Seas. United Nations Environment Programme, Regional

Seas Directories and Bibliographies. Rome, Italy. 31 pp.
Vinogradov, M. E. ( 1978). Physical and biological features of

equatorial upwellings. Bulletin MOIP, otd. bioL 83( 1 ).

5-15. (in Russian)
Vinogradov, M. E. & Shushkina, E. I. ( 1987). Functioning of

the plankton communities of the ocean's epipelagial. Nauka

Publishers, Moscow, (in Russian)
Vinogradov, M. E. & Voronina, N. M. ( 1963). Distribution of

plankton in waters of equatorial currents of the Pacific

Ocean. Oceanology 21, 22. (in Russian)

Voronina, N. M. ( 1964). Distribution of surface zooplankton
in equatorial- current waters of the Pacific Ocean (focusing
on copepods of the Pontellidae family). Trudy Institute of
Oceanology AN SSSR 65, 95-106. (in Russian)

Wells, D. R. (in prep. ). Malayan bird report: 1982 and 1983.
Malay. Nat. J. 43.

Wetmore, A. (1926). The Migrations of Birds. Harvard
University Press, Cambridge, Massachusetts.

Worcester, D. C. (191 1). Hybridism among boobies. Philipp.
J.Sci. Sect. D6, 179-182.

Zaitsev, Yu. P. ( 1971 ). Marine Neustonology. Published for
NMFS. NOAA, US Department of Commerce and National
Science Foundation, Washington, D.C. 207 pp. (Translated
from the Russian)

256

General Conclusions

The BERPAC expedition of 1988 instituted a series of
research projects that greatly amplified the scope of the original
program's overall research objectives. The cruise included an
8-day stay on Caroline Atoll, an uninhabited necklace of islets
enclosing a pristine lagoon. Chapter 1 deals at length with the
history, geology, reef structure, botany, and ornithology of this
little-studied ecosystem. Notable features of the atoll included
a nearly-continuous reef surmounted by 39 islets in various
stages of plant succession, many of them covered in virgin
forest (Subchapter 1.1); a community of 1 1 species of breeding
seabirds. numbering in excess of 1,000,000 individuals
(Subchapter 1.2); and a remarkable Acropora -Tridacna reef
containing the world's densest known colony of Tridacna
clams (Subchapter 1 .4). The inner reef system has developed
within a lagoon that is perched several inches above sea level
at low tide. New species of plants, lizards, land birds
(Subchapter 1.3), and seabirds were discovered for the atoll
during the 1988 expedition.

Caroline also provided an opportunity to compare basic
oceanographic parameters between the open ocean, an enclosed
lagoon, and the confined waters of the South China Sea.

In Chapter 2, one of the more recent aspects of marine
pollution — plastic contaminants — was examined in the central
Pacific Ocean and South China Sea. Surface and subsurface
water was sampled and the quantity and distribution of plastic
debris determined. Plastics collected from the water were
extracted and analyzed for organic pollutants to assess the
potential hazard of the transfer of pollutants to marine organisms
that ingest plastics. Results of sampling revealed that plastic
debris, and specifically raw polyethylene pellets used in
manufacturing, is widespread in the Pacific Ocean and South
China Sea. Plastics occurred more frequently than did tar balls.

even in the South China Sea. Organic contaminants were not
associated with plastics at any detectable levels. However,
subsequent studies demonstrated that plastics can adsorb certain
contaminants and that this could represent a potential hazard to
marine life.

In Chapter 3, primary productivity was compared between
waters close to and distant from Caroline Atoll. Waters close
to Caroline showed less productivity than waters nearer to
the equator (Subchapters 3.1 and 3.2). Not surprisingly,
mesozooplankton (Subchapter 3.3) and neuston
(Subchapter 3.4) diversity and biotnass also increased toward
the equator, relative to the waters near Caroline, due in part
to the hydrodynamics of the water columns near the equator.
The equatorial parts of the Pacific were also much richer in
seabird diversity and density than waters at higher latitudes
(Subchapter 3.5). One surprise of the expedition was
the discovery of a major land bird migration corridor over
the South China Sea from Vietnam to Borneo
(Subchapter 3.6).

The second leg of the 1988 BERPAC expedition was the
first of a planned series of similar cruises in the tropical Pacific
and, as such, should be viewed as a forerunner of expeditions
to come. Many disciplines relevant to the unique biology of the
central Pacific were not represented in 1988, particularly those
concerned with the functioning of coral reef ecosystems. Steps
have been taken to issue a broader call to marine biologists to
participate in future expeditions in order to further our
understanding of the World Ocean and the marine-derived
shallow waters and terrestrial ecosystems of which it is
composed. If the publication of this volume inspires other
biologists to follow, it has fulfilled part of its goal in presenting
the varied findings of a successful initial expedition.

257

Appendix A

Participants of the First Joint US-USSR
Pacific Expedition, Fall 1988.

Tsyban. A. V. USSR Chief Scientist

Volodkovich, Y. L. USSR Assistant Chief Scientist

Smith. G. J. US Chief Scientist

Rostovtsev, O. A. Captain of the R/V Akademik Korolev

Nelepov, Y. N. Assistant to the Captain on Scientific Affairs

Vaytekaya, Y. I. Scientific Secretary

Petrovskaya, S. V. Interpreter

USSR Participants:

Alexandrov, Boris G. (Microbiologist — Neuston)
Junior Scientist,
Institute of Biology of the South Seas &

Academy of Sciences of Ukranian SSR. Odessa Branch
37 Pushkinskaya Street, 27001 1 Odessa, UkSSR

Barinova, Svetlana (Microbiologist — Microorganism indicators)
Leading Engineer,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Belyaeva, Olga L. (Hydrobiologist — Polyaromatic hydrocarbons)
Junior Scientist,

Institute of Global Climate and Ecology
USSR State Committee for Hydrometeorology &
Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Chernyak, Sergei M. (Chemist — Chlorinated hydrocarbons)
Senior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow. USSR

Drakov, Sergei N. (Hydrobiologist — Vertical profiles)
Leading Designer,
Institute of Physics &
Academy of Sciences, Beleorousskoi SSR
Leninsky Prospekt. 70
Minsk, BSSR 220602

259

Irha, Natalya I. (Organic Chemist — Photochemical oxidation of PAH's)
Scientist,

Institute of Chemistry &
Academy of Sciences, Estonia SSR
Akademia Teye Street, 1 5
Tallin. ESSR 200108

Kolobova, Tatiana P. (Analytical Chemist — Trace metals)
Senior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Koltun, Vladimir M. (Hydrobiologist — Benthos)
Department Head and Leading Scientist,
Zoological Institute &

USSR Academy of Sciences
Universitetskaya Street, I
Leningrad, USSR 199034

Korsak, Mikhail N. (Hydrobiologist — Primary production)
Senior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Korzhikov, Igor A. (Hydrobiologist — Primary production)
Junior Scientist,

Far Eastern Regional Research Institute of
Goskomgidromet USSR
24 Dzerjinskiy Stret, 690600
Vladivostok, USSR

Kudryatsev, Vassiliy (Microbiologist — Bacterial production)
Senior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Kulikov, Audrey S. (Hydrobiologist — Mesozooplankton)
Junior Scientist.

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Kumeisha, Alexander A. (Hydrooptician — Hydrooptics)
Senior Scientist.
Institute of Physics &

Academy of Sciences, Beleorousskoi SSR
Leninsky Prospekt. 70
Minsk, BSSR 220602

260

Levina, Olga N. (Hydrobiologist — Microzooplankton)

Engineer,

Southern Division of the Oceanographic Institute &

Academy of Sciences
Gelendzhik 7,353470
Oceanologiya, USSR

Lukin, Alexander E. (Hydrooptician — Hydrooptics)
Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &
Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Mamaev. Vladimir O. (Microbiologist — Numbers and biomass of microorganisms)
Post Graduate Student,
Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Mamaeva, Nila V. (Protozoologist — Microzooplankton)
Senior Scientist,
Southern Division of the Oceanographic Institute of the USSR &

Academy of Sciences
Gelendzhik 7. 353470
Oceanologiya, USSR

Marchenko, Alexanders. (Hydrobiologist — Technician)
Senior Engineer,

Institute of Biology of the South Seas &
Academy of Sciences of Ukranian SSR. Odessa Branch
37 Pushkinskaya Street. 27001 I
Odessa, UkSSR

Medinets, Vladimir I. (Hydrobiologist — Biosedimentation)
Engineer,

Odessa Department of the State Oceanographic Institute
Goskomgidromet, USSR
Proletarski Boulevard, 89
Odessa, USSR 15, 270015

Nelepov, Yeugeniy N. (Assistant to the Captain on scientific affairs)
Far Eastern Regional Research Institute of
Goskomgidromet of USSR
24 Dzerjinskiy Street. 690600
Vladivostok. USSR

Panov, Gennadiy V. (Microbiologist — Bacterial degradation of pollutants)
Senior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

261

Pershina, Irina V. (Analytical Chemist — Dissolved organic matter)
Junior Scientist.

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Petrovskaya. Svetlana (Expedition Interpreter)
Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street. 107258
Moscow, USSR

Polishchuk, Leonid N. (Hydrobiologist — Biogeography of zooneuston)
Senior Scientist,
Institute of Biology of South Seas &

Academy of Sciences of Ukranian SSR. Odessa Branch
37 Pushkinskaya Street. 27001 1
Odessa, UkSSR

Rostovtsev, Oleg A. (Captain of the R/V Akademik Korolev)
Far Eastern Regional Research Institute of
Goskomgidromet of USSR
24 Dzerjinskiy Street, 690600
Vladivostok. USSR

Shigaev. Viktor V. (Oceanographer — Conductivity/temperature/depth)
Senior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Sirenko, Boris I. (Hydrobiologist — Macros-benthos)
Senior Scientist.
Zoological Institute &

USSR Academy of Sciences
Universitetskaya Street. I
Leningrad, USSR 199034

Tsyban, Alia V. (USSR Chief Scientist)

Professor of Ecology and Deputy Director.
Institute of Global Climate and Ecology.
USSR State Committee for Hydrometeorology &
Academy of Sciences
12. Morozo\ Street. 123376
Moscow, USSR

I Irbas, Eha R. (Organic Chemist — Photochemical oxidation of PAH's)
Scientist.

Institute of Chemistry &
Academy of Sciences. Estonia SSR
Akademia Teye Street. 15
l.illin. ESSR 200I0S

262

Vaytekaya, Yanina I. (Hydrochemist — Scientific Secretary)
Hydrometeorological Observatory of Klaipeda
Board of Hydrometeorology of Lithuania
Goskomgidromet of USSR
Taikos Street. 26, 235800
Klaipeda, USSR

Ventsel, Mikhail V. (Hydrobiologist — Phytoplankton)
Junior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &
Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Volodkovich, Yuriy L. (Hydrobiologist — USSR Assistant Chief Scientist)
Senior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &

Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

Vronskaya, Valeriya M. (Hydrobiologist — Photochemical oxidation of PAH's)
Junior Scientist,

Institute of Global Climate and Ecology,
USSR State Committee for Hydrometeorology &
Academy of Sciences
Glebovskaya Street, 107258
Moscow, USSR

US Participants:

Ellis, David H. (Research Biologist, Ph.D. — Atoll ecology and oceanic seabird surveys)
US Fish and Wildlife Service
Patuxent Wildlife Research Center
Laurel, Maryland 20708

Kepler, Angela K. (Biologist, Ph.D. — Atoll ecology and oceanic seabird surveys)
400 Snapfinger Drive
Athens, Georgia 30605

Kepler, Cameron K. (Research Biologist, Ph.D. — Atoll ecology and oceanic seabird surveys)
US Fish and Wildlife Service
Patuxent Wildlife Research Center
Southeast Research Group
University of Georgia
Athens, Georgia 30602

Sibley. Thomas H. (Aquatic Toxicologist, Ph.D. — Bioaccumulation of contaminants)
Fisheries Research Institute. WH-10
University of Washington
Seattle, Washington 98195

263

Smith, Gregory J. (Biologist, Ph.D. — Methods of collection of plastics)
Wildlife International, Ltd.
305 Commerce Drive
Euston, Maryland 21601

Stafford. Charles J. (Analytical Chemist — Methods of analysis)
US Environmental Protection Agency
Analytical Chemistry Laboratory
Beltsville, Maryland 20705

264