Ernie Reese ereese at hawaii.edu
Fri Jun 8 14:58:55 EDT 2001

Dear Colleagues,
We have read with interest the discussion that was initiated by
the May 19,
2001 posting by Dr. Gomelyuk related to our butterflyfish
indicator species
methodology (Crosby and Reese 1996; download-able from
http://coral.aoml.noaa.gov/themes/butterfl.pdf).  Apologies for
not being
able to provide input to this discussion in a more timely manner. 
Hopefully, this interesting and useful discussion can continue
and be
expanded (either on the coral listserver, or off-line between all
interested parties).

Over the last few years, we have interacted with nearly 100
researchers, non-government environmental groups, and government
representatives from around the world who have either used our
(sometimes with interesting modifications) or were interested in
using the
method for various different end goals.  In partial response, we
conducted a number of training workshops throughout the Hawaiian
Saipan, American Samoa and Guam, with more planned in the near
future in
other regions of the Indo-Pacific (i.e., in the Middle East later
month).  During last years ICRI Pacific Regional Symposium and
Workshops in
Noumea, an ad hoc group convened an informal meeting under the
auspices of
our host, Dr. Michel Kulbicki (ORSTOM), to discuss the potential
for a)
convening an international symposium and workshop on the use of
butterflyfish as indicators of change in coral reef ecosystems,
and b)
developing a coordinated database for those who are actively
utilizing some
form of our monitoring method.  There was unanimous support for
both steps
to occur, but no source of financial support has yet been
identified to mak=
either a reality.

It is clear that there is significant interest in the utilization
butterflyfish as indicators of change in the condition (some may
read this
as "health") of coral reef habitats.  The recent discussion on
this topic
that was initiated on the coral listserver reiterates this
interest (as wel=
as the continued value of the coral listserver for encouraging
and debate of various issues related to coral reefs =AD Jim
Hendee continues
to deserve major kudos).

We would now like to add our two-cents worth to the recent
discussion =AD

We have admittedly been somewhat lax in not publishing more
widely in the
peer-reviewed literature our analyses of approximately six years
worth of
data on butterflyfish from throughout the Pacific.  We hope to
rectify this
in the next few months with several manuscripts that we are now
that we hope will provide more rigorous quantitative analyses of
a variety
of conditions in which the technique may be employed (including
as a measur=
of "recovery" of previously heavily impacted reefs near military
ranges).  Nevertheless, there is already quite a bit of
literature that has
been published related to various different aspects of
behavior and their relationship with the coral reef community,
and even mor=
formal presentations on this topic have been made at numerous
forums (i.e., Proceedings of the Hawai'i Coral Reef Monitoring
June 9-11, 1998, Honolulu, Hawai'i; International conference on
Aspects of coral Reef Assessment, Monitoring, and Restoration,
April 14-16,
1999, Ft. Lauderdale, FL; Ninth International Coral Reef
symposium, October
23-27, 2000, Bali)  by many others and ourselves.

Some useful points to remember when considering the butterflyfish
technique (which also includes more standard accepted coral and
transect methodologies):
(1) many people who have tried to use this method have not
followed the key
guideline which is to pick one or two  obligate coral feeding
species, not
the entire assemblage of chaetodontids,
(2) percent coral cover should be measured in areas where the
indicator species has established their territories, since the
relation is
directly between the fish and its territorial coral habitat. The
of these relationships can then be extrapolated over the broader
(3) feeding biology of the indicator species must be known.
(i.e., even
within the coral-feeding guild, a number of species specialize on
(4) for the method to work best and to provide an "early warning"
of visibl=
change, the behavioral components must be measured. To date, few
have done this,
(5) comparisons should be made within sites over time.  Care
should be take=
when attempting to make comparisons between sites that may differ
because o=
the stochastic processes of recruitment of coral and fish.

Some further thoughts for consideration:
=85 The general concept of "indicator species" is both powerful
and valid whe=
applied and interpreted correctly.
=85 "Criticisms" that have been sometimes been mentioned in the
past with
respect to our specific methodology are due to confusion between
the use of
obligate corallivore species=B9 behavioral changes over time, as
an indicator
of gradual sub-lethal and not readily visible change in coral
condition, versus the more general use of the broader approach of
all species of Chaetodontidae as an indicator of coral cover.
[The recent
discussions on this topic on the coral listserver have been more
than many in other venues for correctly noting this point].
=85 Despite the criticism by some in the literature that the more
method of correlating total chaetodontid numbers with percent
coral cover,
even this crude approach shows a significant correlation between
the two
variables.  Nevertheless, our contention is that obligate
chaetodontid behavioral patterns and population size are more
linked to the corals on which they are dependent as an energy
especially when following one or at most two obligate corallivore
over time, than is the linkage between the entire butterflyfish
and the coral community.
=85 Clearly, the selection of the candidate indicator species
must be based o=
knowledge of their behavioral ecology.  The life history
that are important are : a) obligate coral feeder, b) territorial
therefore strongly site-attached, c) long-lived (most
butterflyfish seem to
be), and d) colorful or otherwise easily identifiable.
=85 As a general rule-of-thumb, the butterflyfish indicator
approach is most
appropriate when: a) gradual change (either deterioration or
recovery) of a
reef area is suspected, b) repeated measures over time at the
same site are
possible, and c) funds and "professional scientists" (i.e.,
Ph.D.=B9s and/or
expensive consultants) are not available.
=85 As a general rule-of-thumb, the butterflyfish indicator 
approach is
inappropriate when a) sudden, catastrophic change occurs or has
(i.e., due to storms, coral bleaching, or predation by crown-of-
starfish), b) repeated measures over time at the same site are
not possible=
and c) funds and professional scientists are available.
=85 If the question to be answered is "what is the percent coral
cover?" at a
particular point in time (i.e. a one-time snap-shot), it is
clearly more
useful and efficient to directly sub-sample the coral community
standard methods (i.e., line transect, quadrate).  A one time
"snap-shot" o=
chaetodon abundance and behavior is not appropriate to answer
such a
=85 When gradual change in a coral reef community is suspected,
and an "early
warning" (i.e., before the reef is visibly dead) of such a change
desired, then the behavioral components (feeding rates, agonistic
rates, an=
territory size) of the indicator method are useful measurements
to collect
over time.
=85 The behavioral components of our method are unique in methods
for coral
ref assessment and monitoring.
=85 Of the behavioral components, agonistic behavior is the most
difficult to
interpret because it has multiple motivational causes.  Agonistic
includes both aggressive and submissive behaviors and may simply
be referre=
to as "fighting".  Animals fight for resources.  The three
resources required of all animals are food, mates and a place to
live.  For
obligate corallivore butterflyfish, all of these resources are
impacted by changes in the corals themselves.  If the food
quality or
quantity of a territorial organism declines, that organism will
tend to
expand its territory to increase its food supply and/or increase
its feedin=
rate.  When neighboring con-specifcs both follow this strategy,
encounters will increase.  Thus, changes in agonistic rates of
and/or feeding rates and/or territory size serve as a potentially
"early warning" of changes in ecological conditions on the reef.

The bottom line is that the idea of using butterflyfish as
indicators of
coral reef ecosystem "health" is not going to die, nor should
it.   Aside
from its clear value as a legitimate tool in the diverse arsenal
professional scientific research methods for monitoring coral
reefs, and
perhaps more importantly as a tool for volunteer monitoring and
education/outreach programs, there are simply too many people who
find it
enjoyable and educational to swim around on the coral reefs
counting corals
and chaetodontids, and as an excuse to have fun, to have the
method fall
into disuse.  As rather thoroughly discussed  at the 1998 Hawaii
there is no "perfect" method for monitoring coral reefs, and most
have value by themselves and collectively. Our method is like a
regression with each step adding information based on
implementers ability
to employ the additional step.  It can be modified to suit needs
and abilit=
of different situations and will provide reliable and useful
date if conducted appropriately.  We see further "fine tuning" of
the metho=
and its gradual wide use in coral reef conservation,
resource management and education/outreach efforts.

Dr. Ernst S. Reese      Dr. Michael P. Crosby
Professor                   The Senior Science Advisor for Marine
& Coastal
Dept of Zoology         USAID & NOAA
University of Hawaii    Washington, DC
ereese at hawaii.edu       mcrosby at usaid.gov
For directions on subscribing and unsubscribing to coral-list or the
digests, please visit www.coral.noaa.gov, click on Popular on the
menu bar, then click on Coral-List Listserver.

More information about the Coral-list-old mailing list