visual biology of hawaiian coral reef fishes. ii. colors...

q 2003 by the American Society of Ichthyologists and Herpetologists

Copeia, 2003(3), pp. 455–466

Visual Biology of Hawaiian Coral Reef Fishes. II. Colors of HawaiianCoral Reef Fish

N. J. MARSHALL, K. JENNINGS, W. N. MCFARLAND, E. R. LOEW, AND G. S. LOSEY

The colors of 51 species of Hawaiian reef fish have been measured using a spec-trometer and therefore can be described in objective terms that are not influencedby the human visual experience. In common with other known reef fish populations,the colors of Hawaiian reef fish occupy spectral positions from 300–800nm; yellowor orange with blue, yellow with black, and black with white are the most frequentlycombined colors; and there is no link between possession of ultraviolet (UV) re-flectance and UV visual sensitivity or the potential for UV visual sensitivity. In con-trast to other reef systems, blue, yellow, and orange appear more frequently inHawaiian reef fish. Based on spectral quality of reflections from fish skin, trends infish colors can be seen that are indicative of both visually driven selective pressuresand chemical or physical constraints on the design of colors. UV-reflecting colorscan function as semiprivate communication signals. White or yellow with black formhighly contrasting patterns that transmit well through clear water. Labroid fishesdisplay uniquely complex colors but lack the ability to see the UV component thatis common in their pigments. Step-shaped spectral curves are usually long-wave-length colors such as yellow or red, and colors with a peak-shaped spectral curvesare green, blue, violet, and UV.

THE reasons why reef fish are so colorfulhave been the subject of speculation since

Darwin’s time (Darwin, 1859; Longley, 1917a;Lorenz, 1962), with the Hawaiian Islands andtheir reef fish having been particularly closelyscrutinized (e.g., Longley, 1918; Lorenz, 1962;Barry and Hawryshyn, 1999). When faced withthe many colors of reef fish, it is tempting todescribe them using our human eyes. This canlead to erroneous conclusions if we then try todraw behavioral or physiological conclusionsbased on this experience (Cuthill et al., 1999).Reef fish visual systems (Losey et al., 2003) andtheir light environment (Marshall et al., 2003)are quite different from ours, and the way theysee their own colors must necessarily be quitedifferent from the way we see them. Here weattempt a general description of a variety of Ha-waiian reef fish colors that is independent ofhuman perceptual experience. Marshall et al.(2003) use this and other data to begin to mod-el the design of reef fish colors and vision.These are early steps in the study of the visualecology of reef fish, and therefore we try todraw some general conclusions that we hopewill be of use in future work.

Coral reef dwellers are certainly one of themost colorful and varied assemblages of ani-mals. Reasons suggested for the ‘‘bright’’ col-oration of reef fish include territorial marking(Lorenz, 1962), sexual display (Thresher, 1984),camouflage through object or backgroundmatching (Randall and Randall, 1960; for a use-

ful discussion, see Endler, 1981, 1984), or dis-ruption (Cott, 1940), aposomatism (Ehrlich etal., 1977), temperature control (Fox and Vevers,1960). Some coloration could result from phy-logenetic constraint or a nonfunctional meta-bolic byproduct (Longley, 1917a,b; Fox and Vev-ers 1960). All of these possibilities are renderedmore complex by the ability of almost all reeffish to change color on one or more of a num-ber of time scales from subsecond to ontoge-netic (Townsend, 1929; Crook, 1997a,b). Herewe describe only the spectral distribution andfrequency of the most commonly seen colors of51 species of Hawaiian reef fish and attempt toplace this body of data in the context of previ-ous work.

Use of the human visual system to describecolor and its uses can be very misleading (Ben-nett et al., 1994; Cuthill et al., 1999; Endler,1990). This is well illustrated by a considerationof the ultraviolet (UV) part of the spectrum towhich most mammals, including humans, arerelatively insensitive ( Jacobs, 1993) but whichmay play an important role in processes such assexual selection (Andersson and Amundsen,1997; Andersson et al., 1998; Bennett et al.,1996), camouflage (Lavigne, 1976; Shashar,1994), and food choice (Church et al., 1998).UV is of primary interest in this paper. It is im-portant to realize, however, that just because wedo not see it, UV is not necessarily more im-portant to animals that do see it than is any oth-er spectral region.

456 COPEIA, 2003, NO. 3

Approximately half of the species of reef fish-es examined to this date are largely insensitiveto UV based on ocular media transmission mea-surements (Thorpe et al., 1993; Siebeck andMarshall, 2001; Losey, et al., 2003). Althoughthe remainder may allow UV to reach the retinaand be detected, the value of UV-sensitive visionto such fishes is unclear. Therefore, it is impor-tant to examine functional and ecological dif-ferences between reef fishes that do and thosethat do not detect UV. This is particularly in-triguing for body colors and color patterns be-cause it raises the possibility of a communica-tion channel open only to species with UV visualsensitivity and closed to all others.

MATERIALS AND METHODS

All fish were caught using hand nets and bar-rier nets or, occasionally, rod and line either byresearchers or local tropical fish suppliers. Fishwere housed in marine aquaria at the Hawai’iInstitute of Marine Biology, University of Ha-wai’i, on Coconut Island, Kaneohe Bay. Fish col-ors were occasionally measured in situ or insmall aquaria but normally out of water. Effortswere made to reduce stress (and in some casesthe resulting color change) by anaesthetizingfish with MS222 or clove oil or by removingthem from the water for only a few seconds formeasurement. They were placed in cloth soakedin seawater and the skin was kept wet duringmeasurements. For most species, results of thistechnique are equivalent to those when fish arekept in water (Marshall 1996, 2000a), and mea-surements made out of water allow more exact-ing optical alignment with small spots andstripes. Clove oil occasionally results in fish tak-ing on a dark coloration after several minutes,so we made measurements before this com-menced. Some species (the rapid color chang-ers such as some balistids and pomacentrids, forexample) changed color too rapidly and fre-quently for us to be confident that we were mea-suring their ‘‘normal’’ coloration. Data fromsuch fish were discarded. Whether this was thecase for other fish was judged subjectively (byus closely examining fish in and out of water)and it remains possible that we overlookedsome colors that ‘‘disappear’’ or change withoutour knowledge or are outside of the range ofhuman vision. Attention should be paid to suchproblems in more detailed descriptions of in-dividual species in future work.

Spectrophotometric measurements.—The reflectancespectra from different areas of fish skin weremeasured using either an Ocean Optics S2000

spectrometer, or ‘‘Sub-Spec,’’ an Andor Tech-nology submersible spectrometer that is fullydescribed in Marshall (1996, 2000a). These in-struments and their use are detailed elsewhere(Endler, 1990; Marshall, 1996, 2000b).

The sampling area of the Sub-Spec spectrom-eter, which could be as small as 0.3 mm2 (de-pending on the range to the object being mea-sured) was visualized through a sighting opticthat is much like the viewfinder for a single lensreflex (SLR) camera but with a sampling fiber-optic embedded centrally in the mirror-plane ofthe optical system. The light reflected from acolored area on a fish was sampled through thefiber optic and stored by the spectrometer. WithS2000 measurements, the bare end of a fiberoptic probe attached to the spectrometer wasplaced close to the fish at a 458 angle so that itsampled from that colored region alone. Eachmeasurement was an average of three to 10 sam-ples of the same colored zone on the fish. De-tailed descriptions of spectrometers in generaland their design are available at the Ocean Op-tics website: www.Oceanoptics.com.

Sampling color patches on a fish is a subjec-tive process in the choice of color patches tosample. We strove to include all qualitatively dif-ferent hues. Toward this goal, we also used aUV-sensitive camera to image each fish to revealpossibly inconspicuous UV reflecting areas. De-tails of UV camera design are found in Marshall(1996) and Losey et al. (1999a).

Definition of color categories and spectral shapes.—The same 21 human-subjective color categoriesidentified in Marshall (2000a) are used here toprovide labels and ease description and com-parison of reef fish colors, not for quantitativepurposes. Colors discussed in relation to thesespecific categories are written in italics: colorwithout UV-Blue, Green, Yellow, Orange, Brown/Ol-ive, Red, Blue/Red, Blue/Far-Red, Magenta, Black,Labriform-Green Complex; UV containing colors—UV, White, Violet, Blue-UV, Blue-UV-Hump, Green-UV, Yellow-UV, Orange-UV, Red-UV, Blue/UV/RedLabriform-Purple Complex. These categories arebased on the appearance of the color to us andthe shape of the spectrum of the color (Fig. 1).For example, both Red and Red/UV may appearidentical to our visual system; however Red/UVcontains a second region of reflectance in theUV. Good examples of Yellow/UV and Orange/UVcan be seen in Figure 1B.

General descriptions of colors or collectionsof a number of color categories that may all ap-pear similar to human perception are written innormal font to contrast with the italicized colorsdefined above. For example, ‘‘yellow’’ is used to

457MARSHALL ET AL.—COLORS OF HAWAIIAN FISHES

Fig. 1. Examples of Hawaiian fish colors plotted as percent reflectance relative to a white ‘‘Spectralon’’standard (Endler, 1990; Marshall, 1996). The colors of the curves are approximate matches to the color in lifeas seen by humans. Note that color reflectance takes two forms: step-shaped and peak-shaped. Colors with oneof either of these shapes are classed as simple, colors with two or more of either shape (i.e., two peaks or apeak and a step or two peaks and a step as in Chlororus sordidus) are classed as complex (Marshall, 2000a). Foreach example, the numbers on the body areas in the photographs correspond to the numbered reflectioncurves for the fish. (A–B) Zanclus cornutus. (C–D) Chromis ovalis. (E–F) Chlororus sordidus. (G–H) Coris gaimard.The photographs show close-ups of body regions in this brilliantly colored fish. They are (clockwise) the dorsalfin, the base of the pectoral fin, and the tail close to the caudal peduncle.

describe both color categories ‘‘Yellow’’ and ‘‘Yel-low/UV.’’

The spectrum of a color, or regions of spec-tra, can be described as steplike shapes (e.g.,yellow or red, color 1, Fig. 1B) or peaklike (e.g.,blue, color 4, Fig. 1H, Chittka and Menzel,1992). The wavelength of a step-shaped color

can be expressed as the wavelength correspond-ing to the 50% reflection point (R50) in thestep (the half-way point from top to bottom ofthe step). This is similar to the 50% transmis-sion (T50) used to quantify ocular media(Douglas and McGuigan, 1989; Siebeck andMarshall, 2001; Losey, et al., 2003). Peaks are

458 COPEIA, 2003, NO. 3

TA

BL

E1.

CO

LO

RO

FH

AW

AII

AN

RE

EF

FISH

,OC

UL

AR

ME

DIA

TYP

EA

ND

VIS

UA

LPI

GM

EN

T(L

OSE

YE

TA

L.,

2003

).A

n‘‘X

’’in

dica

test

hatt

heco

lor

liste

din

the

top

row

appr

oxim

ates

the

colo

rse

en.A

llfis

h,un

less

mar

ked

‘‘j’’,

were

mat

ure

orte

rmin

alph

ases

.Col

orca

tego

ries

are

thos

eof

Mar

shal

l(20

00a)

and

are

deta

iled

inM

ater

ials

and

Met

hods

.C

ells

toth

ele

ftof

the

dark

vert

ical

line

code

colo

rsin

whic

hve

rylo

wor

noU

Vre

flect

ion

exist

s.C

ells

toth

eri

ght

ofth

islin

em

ark

colo

rswi

tha

signi

fican

tU

Vco

mpo

nent

toth

eir

refle

ctio

n.T

hat

is,th

eypo

sses

sa

peak

orpo

rtio

nof

thei

rre

flect

ance

ofat

leas

t20

%be

twee

n20

0an

d40

0nm

.The

‘‘UV

com

pone

nts’’

colu

mn

sum

mar

izes

the

UV

cont

ent

for

all

colo

rs.

UV-

refle

ctin

gfis

har

esc

ored

with

a‘‘1

’’un

less

white

isth

eon

lyU

V-re

flect

ing

colo

r,in

whic

hca

seth

eyar

esc

ored

with

a‘‘W

’’.T

hela

sttw

oco

lum

nsto

the

righ

tsu

mm

ariz

eco

nece

llvi

sual

pigm

ent

sens

itivi

tyda

taan

dvi

sion

liklih

ood

estim

ates

mad

efr

omoc

ular

med

iatr

ansm

issio

nda

taan

dec

olog

ical

cons

ider

atio

ns(L

osey

etal

.,20

03).

Ocu

lar

med

iasc

ored

with

ath

ree

orle

sspa

ssso

me

UV

(Los

eyet

al.,

2003

)an

dar

elik

ely

toha

vesh

ort-w

avle

ngth

-se

nsiti

vevi

sion

(UV

orVi

olet

).U

V1an

dV1

are

thos

efis

hwi

thU

V-an

dvi

olet

-sens

itive

cone

cells

and

UV-

fish

are

know

nto

lack

both

.Par

enth

etic

UV1

fish

are

thos

efo

rwh

ich

wesu

spec

tU

Vse

nsiti

vity

that

has

been

over

look

ed(L

osey

etal

.,20

03).

Spec

ies

Blue

Gre

enYe

llow

Ora

nge

Brow

n,O

live

Red

Blu/

Red

Blu/

Far-

Red

Ma-

gent

aBl

ack

Lab.

Gre

enC

om-

plex

Lab.

Purp

leC

om-

plex

Whi

teVi

o- let

UV-

Blue

Blue

/U

V-H

ump

Gre

en-

UV

Yello

w-U

VO

rang

e-U

VRe

d-U

V

UV

com

po-

nent

Visu

alpi

g-m

ent

Ocu

lar

Med

ia

Aca

nthu

rida

eA

cant

huru

sac

hille

sX

XX

XX

1U

V2

5A

cant

huru

sdu

ssum

ieri

XX

XX

14

Aca

nthu

rus

leuco

ster

non

XA

cant

huru

slin

eatu

sX

XX

1A

cant

huru

sni

grof

uscu

sX

XX

15

Aca

nthu

rus

oliv

aceu

sX

XX

X1

V1

Cte

noch

aetu

sst

rigo

sus

XX

X1

UV

25

Nas

olit

urat

usX

XX

XX

X1

V1

4N

aso

unic

orni

sX

1V

14

Zebr

asom

afla

vesc

ens

XU

V2

4

Aul

ostm

idae

Aul

osto

mus

chin

ensi

sX

V1

5

Bal

istid

aeR

hine

cant

hus

acul

eatu

sX

XX

XX

XX

XX

16

Suffl

amen

burs

aX

XX

XX

X1

V1

5

Bot

hida

eBo

thus

pant

heri

nus

XX

XW

5

Cha

etod

ontid

aeC

haet

odon

auri

gaX

XX

XX

15

Cha

etod

onep

hipp

ium

XX

XX

X1

5C

haet

odon

ocell

atus

XX

XW

Cha

etod

onor

natis

sim

usX

XX

XW

4C

haet

odon

trifa

scia

lisX

XX

XX

X1

4C

haet

odon

unim

acul

atus

XX

XW

UV

24

Forc

ipig

erfla

viss

imus

XX

X1

V1

4H

enio

chus

diph

reut

esX

XX

15

Cir

rhiti

dae

Cir

rhito

psfa

scia

tus

XX

W2

Das

yatid

idae

Taen

iura

lym

ma

XX

1


TA

BL

E1.

Con

tinue

d.

Spec

ies

Blue

Gre

enYe

llow

Ora

nge

Brow

n,O

live

Red

Blu/

Red

Blu/

Far-

Red

Ma-

gent

aBl

ack

Lab.

Gre

enC

om-

plex

Lab.

Purp

leC

om-

plex

Whi

teVi

o- let

UV-

Blue

Blue

/U

V-H

ump

Gre

en-

UV

Yello

w-U

VO

rang

e-U

VRe

d-U

V

UV

com

po-

nent

Visu

alpi

g-m

ent

Ocu

lar

Med

ia

Fist

ulae

riid

aeFi

stul

aria

com

mer

soni

iX

X1

5

Hol

ocen

trid

aeSa

rgoc

entr

onsp

inife

rum

XX

XX

XW

5

Lab

rida

eC

oris

gaim

ard

XX

XX

XX

16*

Cor

isve

nust

raX

XX

XX

XX

16

Gom

phos

usva

rius

XX

X1

6H

alic

hoer

esor

natis

sim

usX

XX

X1

5La

broi

des

phth

irop

hagu

sX

XX

X5

Pseu

doch

elinu

stet

rata

enia

XX

XX

3T

hala

ssom

adu

perr

yX

XX

16

Tha

lass

oma

lute

scen

sX

XX

X1

6*

Mon

acan

thid

aePe

rvag

oras

pric

audu

sX

XV

15

Perv

agor

spilo

som

aX

XX

XX

1V

15

Mul

lidae

Mul

loid

icht

hys

vani

colen

sisX

XX

XX

13

Paru

pene

uscy

clos

tom

usX

XX

15

Paru

pene

usm

ultif

asci

atus

XX

XX

X1

(UV

1)

1

Ost

raci

idae

Ost

raci

onm

eleag

ris

XX

XX

X1

V1

4

Pom

acan

thid

aeC

entr

opyg

efla

viss

imus

XX

X1

5C

entr

opyg

elo

ricu

laX

XX

X1

4C

entr

opyg

epo

tteri

XX

XX

5

Pom

acen

trid

aeC

hrom

isov

alie

s(j)

XX

XX

1V

12

Chr

omis

vand

erbi

ltiX

XX

X1

(UV

1)

2?

Scar

idae

Chl

oror

usso

rdid

usX

XX

16

Scar

usps

ittac

usX

XX

X1

6*

Serr

anid

aePs

euda

nthi

asbi

colo

rX

X1

5

Tetr

adnt

idae

Aro

thro

nhi

spid

usX

XX

XX

X1

5A

roth

ron

mela

egri

sX

XW

UV

25

Zanc

lidae

Zanc

lus

corn

utus

XX

XX

W5

460 COPEIA, 2003, NO. 3

Fig. 2. The 50% reflection points (R50) for step-shaped colors (A) and positions of the peak in peak-shaped colors (B) in Hawaiian reef fish. The curveplotted through the data is a five-point moving aver-age with 1-nm wide points.

quantified simply as the wavelength of highestreflection. To examine the frequency distribu-tion of peak and step wavelengths, a five-pointmoving average with 1 nm in each step was plot-ted through the frequency distribution data toaid visual estimation of the distribution.

Monte Carlo test.—Visual inspection of the fre-quency distribution for the wavelengths peaksand R50 steps (Fig. 2) suggested nonrandomclumping of colors around certain wavelengths.A Monte Carlo test was performed to test thesignificance of this impression. Each cell of thisdistribution (each bar of the histogram) repre-sents a 5 nm wide sample. Two features that de-fine the degree to which clumps appear in afrequency distribution were tabulated. First wefound the average count for all nonzero cells ofthe distribution. We then tabulated ‘‘spacing’’as the mean distance between nearest-neighborcells of the frequency distribution that weregreater than the mean count. As a second fea-ture, we defined a clump to be a series of ad-jacent cells that were all greater than the meancount. We then measured mean ‘‘clump size’’as the number of adjacent cells in a clumpwhere all cells were greater than the meancount. Nonrandom clumping will result in asmall spacing statistic and a large clump size.During each iteration of the Monte Carlo test,a random frequency distribution was createdover the same range of wavelengths and totalcount as that of the observations. Spacing andclump size were measured for the random dis-tribution. This iterative step was repeated

10,000 times to create a random expected dis-tribution of spacing and clump size for compar-ison with the observations to determine wheth-er the data were, overall, nonrandomlyclumped. As a guide toward a post hoc decisionas to which of the visually apparent clumps inthe observations were likely to be nonrandom,iterations were repeated until 6000 randomlygenerated clumps of at least two adjacent cellswere generated. Results indicated that the prob-ability of obtaining a clump size of four or great-er was 0.051, so four was taken as the thresholdclump size to qualify as a clump.

RESULTS

Colors and color combinations.—The colors of 51species of Hawaiian reef fish were quantified(Fig. 1, Table 1). These data can usefully becompared to the larger sample (264 species) forthe Great Barrier Reef and Caribbean (GBR&C;Marshall, 2000a). Two of the color categoriesidentified for GBR&C fish, UV and blue/UV/red,were not found in Hawaiian fish but were foundonly in one and three fish, respectively, in theGBR&C sample (Table 1). It is unclear whetherthis has any phylogenetic and/or adaptive ex-planation and may simply reflect undersam-pling of Hawaiian reef fish.

All but six species possessed at least one colorwith a significant UV component. No cleartrend was noted with regards to body regioncontaining UV or human visible color withwhich UV was associated although there was ageneral trend in GBR&C fish for UV to befound in facial markings and fin margins. Thisis true for a few Hawaiian species, such as Nasolituratus that possesses Blue/UV fin margins andParupeneus multifasciatus that has Red/UV eyesand Violet facial markings.

Of the 46 species for which we have visionlikelihood estimates and coloration measure-ments, only 17 are likely to have short-wave-length-sensitive vision (Table 1), and only threeof these lack UV-reflective colors. This is not,however, unexpected since only five of the 51species sampled lacked UV-reflective colors. Ifwe ignore the uneven sampling of families inour data and perform a contingency test forlikely possession of short-wavelength-sensitive vi-sion versus UV-reflective colors, we find a com-plete lack of relationship (chi-square 5 1.3, df5 1, P . 0.1). A similar lack of relationship be-tween UV colors and UV sensitivity was foundin the larger sample of GBR&C fish examinedby Siebeck and Marshall (2001).

The most frequent colors (Table 2) were yel-lows (35/51 species), blues (28/51 species),


black (24/51 species), and white (25/51 spe-cies). The most frequent combinations were yel-low with blue (24/51 species), black with white(17/51 species) and yellow with black (16/51species). Combinations found in Australian fish-es (Marshall, 2000a), but not in Hawai’i, weregreens with yellows, greens with reds, andgreens with white. Our relatively small samplesize, the depauperate Hawaiian fish fauna, and,in particular, uneven sampling of species be-tween families, preclude strong conclusions.One obvious trend is that butterflyfish are fre-quently colored black with white, yellow withwhite and yellow with black, usually with allthree colors present.

Color distribution throughout the spectrum.—Thepeaks and steps of the colors of Hawaiian reeffish are not randomly distributed over the spec-trum (P , 0.025 by Monte Carlo test, Fig. 2).There are regions where steps or peaks occurmore often. For steplike colors, R50 clumps areevident at 347, 385, 515, 575, and 730 nm, butthe Monte Carlo does not support the signifi-cance of the visually apparent 730 nm peak.Peak-shaped colors have maximal reflectionwith significant clumps at 400, 450 and 505 nm,but the Monte Carlo did not support the visuallysuggestive clump at 355 nm. On a larger scale,the region between 400 and 500 nm containsonly one step-shaped color, but contains 55 ofthe 104 peak-shaped colors that were plotted.Because we made every attempt to include anyperceptually different hue, it is unlikely that thegaps in this distribution are caused by samplingerror. In addition, Marshall (2000a) found analmost identical distribution in a much largersample of Australian reef fishes.

DISCUSSION

Cautionary remarks on color descriptions.—Describ-ing colors in a meaningful way is not trivial, andthere is some confusion of nomenclature, es-pecially where the terms ‘‘bright’’ or ‘‘colorful’’are used. Endler (1990) provides a useful lookat the quantifiable attributes of colors and, asothers have done, identifies hue, chroma (orsaturation) and brightness as words that de-scribe quantifiable aspects of colors. Hue refersto the spectral location of the step(s) or peak(s)of the color. Chroma defines the restriction ofthe reflection to only certain portions of thespectrum such as a very steep step or narrowpeak. Brightness refers to the overall number ofphotons reflected. When reef fish are describedas colorful or bright, this often means they con-tain many different hues that are both saturated

and bright (saturation, chroma, and brightnessare often impossible to completely disentangle).A saturated color, at some point in the spec-trum, contains a considerable and often suddenchange in reflectance, for example, the Yellowof Zanclus cornutus (Fig. 1). This color is alsobrighter than Z. cornutus orange (from 350—700nm at least, Fig. 1) as it reflects more light butis only slightly different in hue content. That is,the overall shapes of the curves are similar. Inany detailed study of fish coloration, it is criticalto describe the colors in these terms and not bymatching the color against some human per-ceptual standard.

What is the significance of color clusters?—Cluster-ing could result from constraints in the chemi-cal or physical nature of pigments and structur-al colors (for an excellent resume of the natureof animal colors, see Fox and Vevers, 1960). Offar more interest to us would be clustering thatresults from adaptive interaction among color-ation, environment, and visual pigments.

Our data agree with two, seemingly universal,aspects of natural biological colors found inflowers (Chittka and Menzel, 1992; Kevan et al.,1996), fish (Marshall, 2000a,b), and birds (Fin-ger and Burkhardt, 1994; Vorobyev et al., 1998).First, colors constructed from steps mostly liebeyond 500 nm, that is, yellow, orange, and red.It is rare to find a body or flower color, that is,for instance, red and has a peak-shaped spec-trum (an exception to this is the red feathersof birds of paradise and hummingbirds, Voro-byev et al. 1998). Second, the converse is trueat the other end of the spectrum. Greens, blues,violets, and UV colors tend to possess peak-shaped spectra.

Almost no steplike colors have R50 points be-tween 400 and 500 nm. A similar but less clear-cut trend is seen in the R50 points for Pacificreef fish (Marshall, 2000a; NJM, unpubl.). Theexception to the above trends, step-shapedclumps at 347 and 385 nm, all belong to colorsof white or almost white appearance to us and,therefore, reflect across much of the spectrum.As is the case for ‘‘white’’ flowers and bees (Kev-an et al., 1996), these fish whites (to humans)will not appear white (or at least chromaticallyflat) to fish with short-wavelength-biased visualsystems, especially those with UV receptors.

The reason for the gap in steplike R50 pointsbetween 400 and 500 nm is elusive. Colors thatappear to be UV, blue or green to visual systemsmust necessarily not reflect strongly beyond 500nm or they would acquire a reddish tint. Thesensation of color in this short wavelength partof the spectrum is the result of peak-shaped

462 COPEIA, 2003, NO. 3

TA

BL

E2.

CO

LO

RC

OM

BIN

AT

ION

SO

FH

AW

AII

AN

RE

EF

FISH

.Fish

that

have

the

colo

rco

mbi

natio

nslis

ted

inth

eto

pro

war

esc

ored

with

anX

inth

eta

ble.

Col

orco

mbi

natio

nsfo

rco

lors

foun

dad

jace

ntto

blue

for

inst

ance

are

Bls

1G

s—bl

ues

(blu

e,bl

ue/U

V,bl

ueU

Vhu

mp,

blue

/red

and

blue

far

red)

with

gree

ns(g

reen

and

yello

w/U

V);B

ls1

OS—

blue

swi

thor

ange

s(o

rang

ean

dor

ange

/UV)

;Bl

s1

Rs–

blue

swi

thre

ds(r

edan

dre

d/U

V);

Bls

1B—

blue

swi

thbl

ack;

Bls

1W

—bl

ues

with

white

.T

his

codi

ngfo

rco

lor

com

bina

tions

isre

peat

edfo

rgr

eens

,yel

lows

,ora

nges

,and

reds

.Lab

rifo

rmco

lor

com

bina

tions

are

show

nin

the

last

colu

mn.


TA

BL

E2.

CO

NT

INU

ED

.

464 COPEIA, 2003, NO. 3

spectra (Fig. 1). Yellow, orange, and red colorscould also be the result of peak-shaped spectra;however, for unknown reasons (perhaps chem-ical constraints to pigments), fish skin colors atthis end of the spectrum all result from steplikespectra whose R50 points range from 500—700nm. There is no apparent reason why steplikespectra should not exist between 400 and 500nm except that they would appear spectrallybroad, whiteish or grey/greenish, to any visualsystem and may not convey useful informationfor contrast with longer or shorter wavelengthcolors.

Is there any relationship between possession of UV-re-flective colors and UV-sensitive vision?—As notedpreviously for Pacific and Caribbean species(Losey et al., 1999; Marshall 2000a; Siebeck andMarshall, 2001), not all species in possession ofUV colors transmit UV to the retina (Table 1;Losey et al., 2003). As a result, these species can-not see their own ultraviolet colors, leaving twopossibilities: the UV part of the color is non-adaptive (at least visually), or it is adaptive butonly with respect to the visual system of anotherspecies such as a predator. In fact, possession ofUV-reflective colors appears to have no value asa predictor of short-wavelength-sensitive vision.

Do colors containing a UV component have a specialfunction?—The colors of most Hawaiian reef fishcontain a UV reflective component (Table 1)and, as UV-blind humans, it is tempting to spec-ulate that UV colors combined with an abilityto see UV may open the possibility for a ‘‘se-cret’’ or limited-audience communication chan-nel. UV colors have the potential to be usefulfor close-range communication as these wave-lengths of light are rapidly attenuated in turbidmedia such as reef water (Lythgoe 1979; Sie-beck and Marshall 2000, 2001). Although obvi-ous to nearby conspecifics or competitors, sig-nals using UV will not transmit well to more dis-tant eyes of eavesdropping predators (Losey etal., 1999a; Marshall 2000a). Chromis ovalis andParupeneus multifasciatus, for example, both pos-sess colors that are particularly strongly reflec-tive in the UV, with peak reflectance over 40%at around 360 nm in both species, and both areUV sensitive (either with peak sensitivity in theUV or violet peaking sensitivity, which also con-tains UV sensitivity, Fig. 1, Table 1). Losey(2003) has described a similar system for youngDascyllus spp., damselfish that possess a veryshort-wavelength color patch that would be vis-ible to nearby conspecifics but largely escapethe attention of predators.

Is there functional significance to the combinations ofcolors and are there phylogenetic trends?—The colorcategories Blue, Blue/UV, Yellow, and Orange areapproximately twice as abundant in Hawaiianreef fish as elsewhere (compare Table 1 withappendix 1 in Marshall, 2000a). Approximatelythe same proportion (20%) of relatively abun-dant Hawaiian fish as Great Barrier Reef andCaribbean fish has been examined and bothsamples of fish were caught by similar methods.Whether such a dataset can be linked to phy-logenetic trends, especially in the relatively spe-cies-depauperate Hawaiian waters (Randall,1998) or to more functional reasons such as dif-ferences in water quality remains obscure. How-ever, if we confine the comparison to the threefamilies for which we have the most data in bothHawaii and Australia (Labridae, Chaetodonti-dae, and Acanthuridae) the same observationof increased relative abundance of Blue, Blue/UV, Yellow, and Orange in Hawaiian waters ismaintained.

As noted previously (Loew and McFarland,1990; Longley, 1917a; Lythgoe, 1979), the com-binations of colors used by reef fish are not ran-dom. The color pair combination of yellows ororanges with blues, Black with White, and yellowswith Black are frequently found (lower case, nor-mal font, denotes a generalized, human percep-tual, description of more than one color cate-gory). Possible functional reasons for thesecombinations are examined in detail in Mar-shall et al. (2003), but it is worth noting herethat all are highly contrasting. Also these colorstransmit particularly well and form strong dis-ruptive patterns in marine waters making themideal for communication and/or disruptivecamouflage in the marine environment (Lyth-goe, 1979; Marshall, 2000a).

Phylogenetic trends are difficult to identifywithout more extensive sampling, but twotrends are obvious. First, labroid fishes, the par-rotfishes and wrasses, possess highly complexcolors that are almost unique to this group: La-briform purple, Labriform green, and Blue/red orBlue/far red (Tables 1—2 and Fig. 1B, and fig. 1and appendix 1 in Marshall, 2000a). These fishlack short-wavelength-sensitive vision but havecomplex ocular filters (Siebeck and Marshall,2000, 2001) and striking ontogenetic colorchanges that are important in their social or-ganization and recognition of sex-changed in-dividuals (Warner, 1984). Unfortunately, theirretinas have proven to be extremely difficult toanalyze (Losey et al., 2003), but these complexcolors suggest that renewed efforts should bemade to carefully describe their visual systemsand relate them to body coloration. Second,


both chaetodontids and pomacanthids are of-ten colored in bold black, white, and yellows andblues. Such bold markings may be important inmaintaining contact between conspecifics andhave even been implicated in aposomatism (Er-lich et al., 1977).

ACKNOWLEDGMENTS

Many thanks to U. Siebeck for useful discus-sions and data, T. Yoshikawa, K. Asoh, J. Zam-zow, and P. Nelson for ideas and fish. J. Randallprovided the photographs in Figure 1C and E.N. Hart and two anonymous reviewers suppliedhelpful suggestions on an earlier draft. Thiswork was funded by the ARC in Australia, Na-tional Science Foundation (OCE9810387) inthe United States and NERC and the BBSRC inthe United Kingdom. Work was conducted inaccordance with University of Hawaii IACUCprotocol 95–012. Fishes were captured in accor-dance with the Hawaii Institute of Marine Biol-ogy collection permit. This is contribution 1160of the Hawai’i Institute of Marine Biology.

LITERATURE CITED

ANDERSSON, M., AND T. AMUNDSEN. 1997. Ultravioletcolor vision and ornamentation in bluethroats.Proc. R. Soc. Lond. B. Biol. Sci. 264:1587–1591.

———, J. ORNBORG, AND M. ANDERSSON. 1998. Ultra-violet sexual dimorphism and assortative mating inblue tits. Ibid. 265:445–450.

BARRY, K. L., AND C. W. HAWRYSHYN. 1999a. Effects ofincident light and background conditions on po-tential conspicuousness of Hawaiian coral reef fish.J. Mar. Biol. Ass. U.K. 79:1–14.

BENNETT, A. T. D., I. C. CUTHILL, AND K. J. NORRIS.1994. Sexual selection and the mismeasure of color.Am. Nat. 144:848–860.

———, ———, AND E. J. MAIER. 1996. Ultraviolet vi-sion and mate choice in zebra finches. Nature 380:433–435.

CHITTKA, L., AND R. MENZEL. 1992. The evolutionaryadaptation of flower colors and the insect pollina-tors’ color vision. J. Comp. Physiol. A 171:171–181.

CHURCH, S. C., T. D. BENNETT, I. C. CUTHILL, AND J.C. PARTRIDGE. 1998. Ultraviolet cues affect the for-aging behavior of blue tits. Proc. R. Soc. Lond. B.Biol. Sci. 265:1509–1514.

COTT, H. B. 1940. Adaptive coloration in animals. Me-thuen and Co. Ltd., London.

CROOK, A. C. 1997a. Color patterns in a coral reeffish. Is background complexity important? J. Exp.Mar. Biol. Ecol. 217:237–252.

———. 1997b. Determinants of the physiological col-or patterns of juvenile parrotfish, Chlorurus sordidus.Anim. Behav. 53:1251–1261.

CUTHILL, I. C., A. T. D. BENNETT, J. C. PARTRIDGE, ANDE. J. MAIER. 1999. Plumage reflectance and the ob-

jective assessment of avian sexual dichromatism.Am. Nat. 160:183–200.

DARWIN, C. 1859. On the origin of species by meansof natural selection. Murray, London.

DOUGLAS, R. H., AND C. M. MCGUIGAN. 1989. Thespectral transmission of freshwater teleost ocularmedia—an interspecific comparison and a guide topotential ultraviolet sensitivity. Vision Res. 29:871–879.

ENDLER, J. A. 1981. An overview of the relationshipsbetween mimicry and crypsis. Biol. J. Linn. Soc. 16:25–31.

———. 1984. Progressive background matching inmoths, and a quantitive measure of crypsis. Ibid. 22:187–231.

———. 1990. On the measurement and classificationof color in studies of animal color patterns. Ibid. 41:315–352.

FINGER, E., AND D. BURKHARDT. 1994. Biological as-pects of bird coloration and avian color vision in-cluding ultraviolet range. Vision Res. 34:1509–1514.

FOX, H. M., AND G. VEVERS. 1960. The nature of ani-mal colors. Sidgwick and Jackson, London.

JACOBS, G. H. 1993. The distribution and nature ofcolor vision among the mammals. Biol. Rev. 68:413–471.

KEVAN, P., M. GIURFA, AND L. CHITTKA. 1996. Why arethere so many and so few white flowers? TrendsPlant Sci. 1:280–284.

LAVIGNE, D. M. 1976. Counting harp seals with ultra-violet photography. Polar Rec. 18:269–277.

LOEW, E. R., AND W. N. MCFARLAND. 1990. The un-derwater visual environment. p. 1–43. In: The visualsystem of fishes. R. H. Douglas and M. B. A. Djam-goz (eds.). Chapman and Hall, London.

LONGLEY, W. H. 1917a. Studies upon the biologicalsignificance of animal coloration. I. The colors andcolor changes of West Indian reef-fishes. J. Exp.Biol. 1:533–601.

———. 1917b. Studies upon the biological signifi-cance of animal coloration. II. A revised workinghypothesis of mimicry. Am. Nat. 11:257–285.

———. 1918. Habits and coloration of HawaiianBrachyura and fishes, with a note on the possibilityof submarine color-photography. Year Book Car-negie Inst. 17:158–163.

LORENZ, K. 1962. The function of color in coral reeffishes. Proc. R. Inst. Gt. Brit. 39:282–296.

LOSEY, G. S. 2003. Crypsis and communication func-tions of UV-visible coloration in two coral reef dam-selfish, Dascyllus aruanus and D. reticulatus. Anim.Behav. In press.

———, W. N. MCFARLAND, E. R. LOEW, J. P. ZAMZOW,P. A. NELSON, AND N. J. MARSHALL. 2003. Visual bi-ology of Hawaiian coral reef fishes. I. Ocular trans-mission and retinal sensitivity. Copeia 2003:433–454.

LYTHGOE, J. N. 1979. The ecology of vision. ClarendonPress, Oxford.

MARSHALL, N. J. 1996. Measuring colors around a cor-al reef. Biophoton. Int. July/Aug.:52–56.

———. 2000a. The visual ecology of reef fish colors,p. 83–120. In: Animal signals. Signalling and signaldesign in animal communication. Y. Espmark, Y.

466 COPEIA, 2003, NO. 3

Amundsen, and G. Rosenqvist (eds.). G. Tapir,Trondheim, Norway.

———. 2000b. Communication and camouflage withthe same bright colors in reef fish. Philos. Trans. R.Soc. Lond. 355:1243–1248.

———, K. JENNINGS, W. N. MCFARLAND, E. LOEW, AND

G. S. LOSEY. 2003. Visual biology of Hawaiian coralreef fishes. III. Environmental light, and an inte-grated approach to the ecology of reef fish vision.Copeia 2003:467–480.

RANDALL, J. E. 1998. Zoogeography of shore fishes ofthe Indo-Pacific region. Zool. Stud. 37:227–268.

———, AND H. A. RANDALL. 1960. Examples of mim-icry and protective resemblance in tropical marinefishes. Bull. Mar. Sci. Gulf Carib. 10:444–480.

SHASHAR, N. 1994. UV vision by marine animals: main-ly questions, p. 201–206. In: Ultraviolet radiationand coral reefs. D. Gulko and P. L. Jokiel (eds.).Hawai’i Institute of Marine Biology, Honolulu.

SIEBECK, U. E., AND N. J. MARSHALL. 2000. Transmis-sion of ocular media in labrid fishes. Philos. Trans.R. Soc. Lond. B Biol. Sci. 1257–1262.

———, AND ———. 2001. Ocular media transmissionof coral reef fish—can coral reef fish see ultravioletlight? Vision Res. 41:133–149.

THORPE, A., R. H. DOUGLAS, AND R. J. W. TRUSCOTT.1993. Spectral transmission and short-wave absorb-

ing pigments in the fish lens. I. Phylogenetic distri-bution and identity. Ibid. 33:289–300.

THRESHER, R. E. 1984. Reproduction in reef fishes.T.F.H. Inc. Ltd., Neptune City, NJ.

TOWNSEND, C. H. 1929. Records of changes in coloramong fishes. Zoologica 9:321–378.

VOROBYEV, M., D. OSORIO, A. T. D. BENNETT, N. J.MARSHALL, AND I. C. CUTHILL. 1998. Tetrachromacy,oil droplets and bird plumage colors. J. Comp. Phy-siol. A 183:621–633

WARNER, R. R. 1984. Mating behavior and hermaph-roditism in coral reef fishes. Am. Sci. 72:128–136.

(NJM) VISION, TOUCH AND HEARING RESEARCHCENTRE, UNIVERSITY QUEENSLAND, BRISBANE,QUEENSLAND, AUSTRALIA; (WNM) SCHOOL OFFISHERIES AND FRIDAY HARBOR MARINE LABO-RATORY, UNIVERSITY OF WASHINGTON, FRIDAYHARBOR, WASHINGTON 98250; (ERL) DEPART-MENT OF BIOMEDICAL SCIENCES, VETERINARYCOLLEGE, CORNELL UNIVERSITY, ITHACA, NEWYORK 14853; AND (GSL) HAWAI’I INSTITUTE OFMARINE BIOLOGY, UNIVERSITY OF HAWAI’I, P.O.BOX 1346, KANEOHE, HAWAII 96744. E-mail:[email protected]. Send re-print requests to NJM. Submitted: 13 Feb.2001. Accepted: 1 Feb. 2003. Section editor:W. L. Montgomery.

visual biology of hawaiian coral reef fishes. ii. colors...

Documents