Pacific Science (1982), vol. 36, no. 4© 1983 by the University of Hawaii Press. All rights reserved

ABSTRACTS OF PAPERS

Seventh Annual Albert L. Tester Memorial Symposium15-16 April 19821

The Albert L. Tester Memorial Symposium is held in honor of ProfessorAlbert L. Tester who, at the time of his death in 1974, was Senior Professor ofZoology at the University ofHawaii. The faculty and students of the Departmentof Zoology proposed an annual symposium of student research papers as ameans ofhonoring, in a continuing and active way, Dr. Tester's lively encourage­ment of student research in a broad range of fields within marine biology. Papersreporting original research on any aspect of biology are solicited from students atthe University and these papers are presented at the symposium, which takesplace during the spring semester. Income from contributions to the Albert L.Tester Memorial Fund of the University of Hawaii Foundation is used toprovide two prizes for the best papers by graduate students in the Department ofZoology. Papers are judged on quality, originality and importance of researchreported, as well as the quality of the public presentation. Judges include severalmembers of the faculty of the Department of Zoology as well as winners of thesymposium from the preceding year, when possible. In addition, a distinguishedscholar from another university is invited to participate in the symposium as ajudge and to present the major symposium address. This year Dr. Howard Bernof the University of California, Berkeley participated in the symposium.

Epizootiologic Study of Mid-Cycle Disease of LarvalMacrobrachium rosenbergii 2

G. AKITA,3 R. NAKAMURA,3 J. BRocK,4 G. MIYAMOTO,3M. FUJIMOTO,S F. OISHI,s D. ONIZUKA,s and D. SUMIKAWAs

Mid-cycle disease (MCD) of larval Macro-brachium rosenbergii has reduced survival 1 Manuscript accepted May 1982.

f 1 1 1 h ~ '1' f 2This work was the result of research (Project N.o arva cu tures at testate laCl lty rom a A/R-8) sponsored in part by the University ofHawaii Seanormal 50-70% of the population at the post- Grant College program under Institutional Grant No.larval stage to 5-10%. Observations revealed NA799AA-D-00085 from N.O.A.A., Office ofSea Grant,a characteristic mortality pattern where a peak Department of Commerce; and State of Hawaii, Aqua­in daily mortality occurred near the middle culture Development Program, under Grant No. DLNR

11918.of the rearing cycle. 3 University of Hawaii, Department of Animal Sci-

Epizootiologic studies were initiated using ences, 1800 East-West Road, Honolulu, Hawaii 96822.existing facilities to define aspects of MCD 4 State ofHawaii, Aquaculture Development Program,and elucidate possible causes. Experiments 335 Merchant Street, Room 239, Honolulu, Hawaii

96813.were undertaken in duplicate culture tanks, 5State of Hawaii, Division of Aquatic Resources,each separated into 4 equal-volume compart- Department of Land & Natural Resources, 1151 Punch­ments by a nitex screen. Water, algae, and bowl Street, Room 330, Honolulu, Hawaii 96813.

507

508

some feed could pass between compartments;prawn larvae could not.

Reference larvae were introduced into acompartment of each tank on the day ofhatch. Test larvae populations of differentages and sources were introduced into theremaining compartments at various timesafter the initiation of the experiment. Deadlarvae in the bottom sediment were counteddaily and mortality curves were generated.Length and developmental stage of live anddead animals in each compartment wereassessed daily.

Results indicated that a toxic period of un-

PACIFIC SCIENCE, Volume 36, October 1982

known etiology began on the 10th day of cul­ture and probably extended for the durationof the rearing cycle. Test populations intro­duced during the toxic period displayed amortality peak within 4-6 days. Larvae at age15 days (stage 6-7) or older appeared refrac­tory to the toxicity.

MCD apparently reduced growth rates(length) but did not appear to selectively killlarger or smaller susceptible individuals. Itwas problematic that large numbers of ani­mals lost from populations were not ac­counted for. Studies on MCD will continue.

Larval Development and Metamorphosis of the ProsobranchMollusc, Epitonium ulu, Associated with a Solitary Coral

JANICE L. BELL6

In Hawaii Epitonium ulu has been foundonly on the solitary coral, Fungia scutaria.This degree of host specificity led to twoopposing hypotheses. The first was that Epi­tonium would have a very short larval periodduring which time the larvae would neverleave the optimal area for their host on thecoral reef. The opposing hypothesis was thatEpitonium could remain competent to meta­morphose for a long period, until it encoun­tered its host. A corollary to this secondhypothesis is that larvae would metamor­phose only in response to the coral Fungia.Epitonium ulu is extremely fecund for itssize (mean adult length: 13 mm). During thesummer each animal produces a mean of 32capsules per day, each capsule containing500-600 eggs. Embryos complete intracap­sular development in 6 days and hatch asplanktotrophic veligers.

Cultures of newly hatched larvae were setup to investigate the length of larval life, thegrowth of larval shell and tissue, and themetamorphosis of competent larvae. Larvae

6University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

were fed both Isochrysis galbana and Pavlovalutheri beginning on day 1 post-hatching, andthese were supplemented with Phaeodactylumtricornutum, beginning on day 13.

Shell length increased steadily from 1 to 27days post-hatching, after which it leveled offat 390 .um. Tissue mass (ash-free dry weight)was more variable than shell length, but alsoincreased steadily with age and leveled offafter day 24. Twenty percent of the larvaewere first capable of metamorphosis on day27; this metamorphosis was induced with seawater that had been in contact with Fungia.Larvae began to metamorphose spontane­ously (without contact with Fungia) on day29. By day 36, 50% had metamorphosed andby day 60, all larvae had metamorphosedor died. The competent period may be evenshorter than these data suggest. The 31-dayspan of spontaneous metamorphosis couldhave been caused by individual variability inattaining metamorphic size. Thus larvae mayreach competence and metamorphose shortlythereafter. Epitonium ulu did not fit eitherhypothesis and seems to have little flexibilityin where it settles. The lack of ability to re­main in the plankton may be balanced by thehigh fecundity of this snail.

Abstracts of Papers

The Impact of Eel Predation Reconsidered:The Populations of Potential Prey7

STANLEY D. BLUM8

509

From July 1980 through September 1981 astudy was conducted to assess the impact ofpredation by moray eels on a reef fish com­munity. The abundance of moray eels on anexperimental patch reef was reduced by thetrapping and removal of adults. The diurnalfish communities on this experimental and anearby control patch reef were censused priorto the treatment and at four additional timesduring the following year.

The 8 most abundant species comprised90% by numbers of the visually censused fishcommunity. Analysis of variance was per-

7 This research was supported by the Sea Grant Pro­gram, grant MR/R-I1.

8 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

formed on each of these 8 species to detectdifferences between reefs and differences overtime. Contrast coding was used to test thesignificance of any divergence between cor­responding experimental and control reefpopulations. Those species responding to theeel removal were expected to show such diver­gence. Chaetodon miliaris and Dascyllusalbisella were the only two species to showsignificant population increases on the experi­mental relative to the control. It is noted,however, that certain assumptions of the ex­perimental design and censusing techniqueare important enough to keep us from con­cluding that these species were indeed releasedfrom predator limitation. We intend to inves­tigate these assumptions and to strengthen theinference of these results by replication.

Applications of a New Radiotracer Technique forInvestigating Prawn (Macrobrachium rosenbergii) Ecology

and Digestive Physiology

BARRY A. COSTA-PIERCE9

Water stable, 14C radiolabeled, chemotac­tically active, bacteria-laden, artificial detri­tus was formulated by a rapid, simple method.It was used to follow the assimilation of bac­terial macromolecules into the foregut, mid­gut, and tail tissue of aquarium-acclimatedfreshwater prawns, Macrobrachium rosen­bergii (deMan) in time course experimentsunder conditions of starvation, heavy supple­mental feeding, and sudden crowding. Over a24-hr time course with starved prawns, theproportion of the label assimilated into lipidin the midgut and tail tissue increased 12.3%and 8.0%, respectively, with decreased pro-

9 University of Hawaii, Department ofOceanography,Honolulu, Hawaii 96822.

portions of the label in carbohydrate in themidgut and protein in the tail tissue. Starvedprawns incorporated similar proportions ofthe label in carbohydrate and protein of themidgut, as did heavily fed prawns, but shunteda larger proportion of the label into lipid.Protein was catabolized in order to synthesizelipid and appeared to be used as an energysource. Suddenly crowded prawns greatlyelevated the amount of radioactivity propor­tioned into carbohydrates and lipids of themidgut at the expense of greatly lowered pro­tein activities.

The formulation prepared here, whichappears chemotactically and visually active asartificial "fish feces," may have wide appli­cability in many aquaculture pond ecologyand nutrition studies.

510 PACIFIC SCIENCE, Volume 36, October 1982

Does Glucose Uptake in Marine Fish Intestines Occur by Active Transport?

RONALDO P. FERRARIS l3

Initial entry rates (influxes) of 3H-D­glucose in the upper and lower intestines ofthe surgeonfish (Acanthurus mata) and theeel (Gymnothorax undulatus) occur via twoindependent processes operating simultane­ously: (1) a carrier-mediated transport systemexhibiting saturation kinetics of the Michaelis­Menten type and (2) a linear entry processhaving a rate that is proportional to the ex­ternal glucose concentration. This latter non­saturable process represents 20-40% of totalinflux when external glucose concentrationsare at least 0.5 mM. The direction of transportvia this nonsaturable process may be reversed(i.e., from cell to gut lumen) if intracellu­lar glucose concentration increases, therebydiminishing the ability of the intestine to effecta net movement of glucose from lumen toblood. Uptake studies showed that glucosewas not accumulated by fish intestinal epithe­lia to concentrations significantly above thoseof control or Na-free media. Radioactivity inethanol extracts of glucose-incubated tissues

13 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

was mostly volatile (e.g. tritiated water; 50­75%); whereas control studies using a non­metabolizable, actively transported analog ofglucose (3-0-methyPH-D-glucose) resultedin volatile fractions amounting to less than5% of total activity. Thin layer chromato­graphy of these ethanol extracts showed thatthe smaller nonvolatile activity fraction wasmostly (greater than 90%) glucose. Absenceof Na in incubation saline reduced glucoseuptake in the surgeonfish lower intestine andin both the eel intestinal segments, but not inthe surgeonfish upper intestine. This lattereffect was probably due to the relatively lowaffinity of this tissue for glucose. Thus, marinefishes have opted to metabolize a significantfraction of transported glucose to (1) maxi­mize energy-independent downhill uptake ofthis substance into the cell from the gutlumen, and (2) minimize its efflux into theluminal medium. This study further suggeststhat marine fish intestine may differ, in part,from that of freshwater and terrestrial verte­brates, which can actively transport glucose,because in these fishes diffusional entry of thesugar into epithelial cells from the lumen maybe of importance in net solute flow to theblood.

Conditioning in Honeybees10

PATRICIA A. COUVILLON ll

The performance of free-flying honeybees(Apis melli/era) is studied under conditionsanalogous to those used for the analysis oflearning in vertebrates. The training proce­dure, much like that developed by von Frisch

10 This research was supported by grants BNS75-21672and 79-05876 from the National Science Foundation toDr. M. E. Bitterman.

11 University of Hawaii, Department of Psychologyand Bekesy Laboratory of Neurobiology, Honolulu,Hawaii 96822.

for the study ofsensory processes, is to permitan animal to feed on a target marked with adistinctive color or odor, or both, and then torecord its responses to an unbaited target.One series of experiments, in which the prop­erties of compound stimuli are being ex­plored, provides strong support for thehypothesis that a compound functions as asum of its components (Pavlov's summationhypothesis). For example, after separate re­warded experiences with a color and an odor,

Abstracts of Papers

honeybees respond more to the color andodor in compound than to either of the com­ponents alone (summation of excitation);while after separate nonrewarded experienceswith the same components, they respond lessto the compound than to either of the com­ponents alone (summation of inhibition).Stimulus compounds do have unique proper­ties for honeybees. For example, after experi­ences with two color-odor compounds, beesrespond more to those compounds than to thetwo others that can be constructed ofthe samecomponents. Compound uniqueness, which isfound also in vertebrates, does not contradict

511

Pavlov's summation hypothesis. It is reason­able to assume that afferent interaction gen­erates a compound-specific product whichis rewarded along with the components intraining with the compound. The resultsof these and many other experiments withhoneybees are strikingly similar to those forvertebrates. The similarity may be surprisingin view of the remoteness of common ances­try, but perhaps not so surprising in view ofthe near-identity (homology?) of synapticmechanisms and the likelihood that many ofthe phenomena being studied may eventuallybe understood in simple synaptic terms.

Feeding Behavior of Chaetodon unimaculatus in Kaneohe Bay

EVELYN Cox12

Coral feeding by reef fishes can exert a con­siderable influence on coral community struc­ture through impacts on growth and survivalof coral colonies. Patch reefs at the north endof Kaneohe Bay characteristically have anextensive coral cover. On Tom's Reef, thestudy site, the patch reef top consists ofapproximately 50% Porites compressa and50% Montipora verrucosa, but the slope isdominated by P. compressa. The coral feedingchaetodontid, Chaetodon unimaculatus, isabundant on Tom's Reef, often forming largeaggregations of20 or more fishes on the slope.

Observations by P. Motta at Haunama Bayindicated that Chaetodon unimaculatus pre­fers Montipora sp. and Leptastrea sp. as food.In order to begin a study of the interactionbetween selective feeding by C. unimaculatusand the distribution of M. verrucosa on patch

12 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

reefs in Kaneohe Bay, preliminary feedingpreference tests were conducted in the labo­ratory, with supplemental observations offoraging behavior on Tom's Reef.

Chaetodon unimaculatus were maintainedin flow-through seawater tanks at the HawaiiInstitute of Marine Biology. Feeding prefer­ence was tested by offering equal sized colo­nies of Montipora verrucosa and Poritescompressa and recording the number of biteson each species during a 3D-minute trial period.Five fishes were tested for preference (26 trialstotal), and a Goodness of Fit test, using theextrinsic hypothesis that equal numbers ofbites should be taken from both coral species,showed significantly more bites were takenfrom M. verrucosa (p < 0.001). Field obser­vations of 4 tagged fishes also supported thisfinding; in 17 observation periods of 15 min­utes each, significantly more bites were takenfrom M. verrucosa than from P. compressa(p < 0.001).

512 PACIFIC SCIENCE, Volume 36, October 1982

Feeding and Spatial Organization of Three Species of Pomacanthid Fishes14

THOMAS F. HOURIGAN, B. CARLSON, P. MOTTA, F. G. STANTON, and C. D. KELLEy15

This study explores the feeding behaviorand spatial organization of three sympatricspecies of Caribbean angelfishes (family:Pomacanthidae): Holacanthus tricolor, Poma­canthus paru, and Pomacanthus arcuatus.Saturation diving from the NOAA Under­water Laboratory System I, off the coast ofSt.Croix, U.S. Virgin Islands, allowed extendedperiods of field observation of these wide­ranging fishes. Seven H. tricolor were cap­tured, measured, sexed, marked by finclippingand released. Seven P. paru and six P. arcua­tus were identified individually by naturalmarkings.

Marked H. tricolor spent much of the timeconcealed under ledges. Male home rangeswere relatively large (x = 907 m2 ± 35 m2 ,

N = 2), encompassing the home ranges of upto three females, with whom he interacted.The home ranges of these females overlapped,averaging 335 m2 (± 175 m2 , N= 5) in size.These observations suggest that H. tricolor ex­hibits a haremic social system. During foraysaway from shelter, feeding was directed to­wards algal turfand several species ofsponges.

14This research was supported by NOAA ContractNo. NA8IAAA02688.

15 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

P. paru and P. arcuatus are closely relatedand morphologically very similar. Their be­havior differs greatly from that of H. tricolor.Individuals of both Pomacanthus spp. formedstable pairs with members of the same species.These pairs frequented large (2000-2500 m2),

intraspecifically exclusive home ranges. In­dividuals spend relatively long periods of timein nonforaging activities. Such periods in­clude extensive movements but very littlesheltering. Preliminary data indicate thatP. paru feeds primarily on sponges, andto a lesser extent on gorgonians and algae.P. arcuatus take approximately the samenumber of bites on each of these food types.

It seems likely that H. tricolor are haremicand perhaps sex-reversing. The two Poma­canthus species, on the other hand, appearmonogamous and possibly gonochoristic.Currently sociobiological theory predicts thatthese different social systems are determinedin part by the distribution and abundance ofcritical resources. Although all three speciesfeed on many of the same sponges, the twoPomacanthus spp. differ from H. tricolor byexploiting gorgonians as a food source. Inaddition, they show differential use of shelter.These are suggested as possible determinantsof social organization in these species, andresearch is continuing on these aspects.

The LD-50 Concentration and the Effects of Alcohol on the Fertility ofthe Nudibranch Mollusc Phestilla sihogae

CEDAR KEHOE16

The LD-50 concentration ofalcohol in sea­water was determined for the nudibranchmollusc Phestilla sibogae exposed for 24hours. In conjunction with the experimentsdesigned to establish the LD-50 value, data

16 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

were collected to determine the effects of thealcohol on the fertility of the egg masses laidby the animals. Analyzing the effects of alco­hol on P. sibogae is an important step indemonstrating the feasibility of using theseanimals to produce alcohol-tolerant andalcohol-dependent strains through direc­tional selection experiments.

Abstracts of Papers 513

A New Hypothesis to Explain Aspects of Coloration in Chaetodontid Fishes

CHRISTOPHER KELLEY! 7

Using black bars, black spots, and "poster"patterns, the 3 basic types of color patternsthat occur in chaetodontid fishes, 47 differenthypothetical combinations were generatedand evaluated for their potential effectivenessin reducing predation. Each of the 114 speciesof chaetodontids was then classified as one ofthe combinations. Over 90% of the specieshave bars that conceal their eyes and thusconceal a possible predator search image. Eyeconcealment alone fails to account for eye­spots, bars, or other posterior patterns thatoccur in 88% of the species. Combinationswhere these patterns could possibly deceivepredators regarding the orientation of the fishby making the anterior and posterior endssymmetrically equivalent or reversed occurin only 47% of the species. As pointed outby previous authors, poster coloration mayfunction as a warning to predators; however,this hypothesis fails to explain either the shapeor the widespread occurrence of conspicuouspatterns on the posterior end.

Consequently, a new hypothesis is pro­posed to explain these patterns, the VisualContact Hypothesis, which is that posterior

17 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

coloration functions to increase the opticalsignaling properties of the individual whenseen from behind, enabling following con­specifics to maintain visual contact. In chae­todontids this may occur because predation isminimal, and because conspecifics locate eachother visually. Their compressiform shapeand general body color make it very difficultto see them from behind during following,which is a frequent and important socialbehavior.

This hypothesis differs from previous onesin two major respects. First, in a large numberof chaetodontids, it accounts for both thelocation and the shape of conspicuous colorpatterns. This includes not only large patchesand bars whose shapes are determined by theposterior body curvature, but small blackpatches or spots on the caudal peduncle ornape which produce highly visible flicker sig­nals as the caudal fin sweeps back and forthduring swimming. Second, it appears to be thefirst testable hypothesis. Several neurophysi­ological studies are being considered as theinitial experimental approach, while com­parative studies will be undertaken to examinesuch aspects as depth, home ranges, orienta­tion during following, and size in chaetodon­tids as well as in other families of fishes wherethe hypothesis is applicable.

Sexual Reproduction in the Solitary Scleractinian Coral,Fungia scutaria (Lamarck)

DAVID A. KRUPP!8

A fortuitous set of circumstances led to therecognition and observations of spawningand larval development in the solitary scler­actinian coral, Fungia scutaria. Spawning

18 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

occurred monthly through summer and earlyfall. The eggs, lacking the symbiotic zooxan­thellae algae seen in the adults, measuredapproximately 84 pm in diameter. No ferti­lization membrane was discernible, but thereappeared to be a jelly coat. Cleavage was com­plete, polar, and irregularly radial. Slow

514

moving, ciliated larvae were observed within18 h of spawning. These developed mouthsand seemed capable offeeding within 39 h. Bythis time the planulae were highly mobile, buttended to be positively geotactic. Planulaefound in the aquarium tank in which the adultcorals were being maintained became infectedwith zooxanthellae within 4 to 5 days afterspawning. Attempts experimentally to infectlaboratory-reared planulae lacking the sym-

PACIFIC SCIENCE, Volume 36, October 1982

bionts yielded ambiguous results because ofhigh mortality of planulae. Settlement wassimilarly ambiguous. Although planulae werefound attached to glass slides placed in theaquarium, those planulae usually did not re­main attached. Mechanical dislodgement ofattached planulae induced mobility after ashort quiescent period. Only two developingsettled polyps were found.

Magnetite in the Green Turtle

ANJANETTE PERRy19

Migratory green turtles, Chelonia mydas,can accurately navigate in the open oceanover distances of several hundreds to thou­sands of kilometers. Numerous guidancemechanisms for their feeding to breeding sitejourneys have been postulated, but few havebeen examined experimentally. Preliminarydata suggest one navigational cue may be theability of the green turtle to detect the geo­magnetic field.

Biogenic magnetite, a possible transducerof the earth's magnetic field, has recentlybeen discovered in a variety of marine species,ranging from invertebrates to pelagic fish andmarine mammals. In order to determine ifChelonia also possess magnetite, hatchling,

19 University of Hawaii, Department of Oceano·graphy, Honolulu, Hawaii 96822.

juvenile, and adult specimens were tested formagnetic remanence using superconductingmagnetometers. Saturation induced magneticremanence one to two orders of magnitudegreater than background noise was found inthe heads of all turtles examined, and waslocalized to the dura mater in the adults.Samples were subsequently subjected to alter­nating field demagnetization and were foundto be magnetically stable.

Dura tissue from six juveniles and oneadult was digested with hypochlorite solu­tion (household bleach). The residue wasexamined microscopically and numerousopaque magnetic particles were removed.X-ray diffraction and microprobe analysis ofthe particles showed that they were composedof very pure magnetite crystals. Scanningelectron microscopy revealed unique spheri­cal crystals, which have not previously beenreported in biological samples.

Abstracts of Papers

The Dispersal Potential of Pocillopora damicornis Planulae,Based on Larval Physiology and Behavior'

ROBERT RICHMOND20

515

The hermatypic coral Pocillopora dami­cornis (Linnaeus) is distributed across thou­sands of kilometers of the tropical Pacific,which raises the question of larval dispersalpotentiaL

The larvae, which are approximately 70%lipid by body weight, contain an average of0.63 calories per individual and 7.5 calories

20 State University of New York at Stony Brook,Department of Ecology and Evolution, Stony Brook,New York, 11794 and University of Hawaii, HawaiiInstitute of Marine Biology, Honolulu, Hawaii 96822.

per mg ash-free dry weight. During photosyn­thesis, approximately 13-27% of the carbonfixed by the symbiotic zooxanthellae is trans­located to the animal portion of the larva.

Planulae have been found to remain com­petent for over 3 months. and they possess theability to return to the planktonic state afterhaving settled and calcified.

Considering energetic inputs of the sym­biotic zooxanthellae, stored energy reserves,and behavioral adaptations, long-range dis­persal of planulae from P. damicornis may bepossible.

Notes on the Biology and Feeding Habits of Caranx ignohilis andCaranx melampygus in the Northwestern Hawaiian Islands

ANTHONY SUDEKUM21

Caranx ignobilis and Caranx melampygus(Carangidae) are top-level predators of thecoral reef communities of the northwesternHawaiian Islands. An analysis of the dietsof these two closely related, sympatric jackfishes revealed that both are top-level carni­vores. Shallow-water reef fishes of the familiesLabridae, Scaridae, and Priacanthidae werefound to be important in the diets of bothspecies, though there were significant differ­ences in the relative importance of these andmany other diet items. Stomachs of C. ignobiliscontained large numbers of nocturnal andpelagic fishes, crustaceans, and cephalopods,

21 University of Hawaii, Department of Zoology andHawaii Cooperative Fishery Research Unit, Honolulu,Hawaii 96822.

while the prey of C. melampygus consistedalmost entirely of diurnally active reef fishes.Dietary overlap between top predator specieswas compared using the Shannon-Weaverindex of overlap. Minor overlap (.20) wasindicated between C. ignobilis and C. melam­pygus, and a relatively large overlap (.61) wasfound between C. ignobilis and the grey reefshark Carcharhinus amblyrhynchos. Bothcarangids appear to be reproductively activethroughout the year, with some evidence ofsummer spawning peaks. Skewed sex ratiosfor fish caught in large schools indicate thepossibility of monosexual prespawning aggre­gations for C. ignobilis. These two large, wide­ranging piscavores appear to be very impor­tant in determining the structure and com­position of the coral reef fish communities ofthe northwestern Hawaiian Islands.

516 PACIFIC SCIENCE, Volume 36, October 1982

The Likely Site of the Magnetic Sense Organ in Yellowfin Tuna(Thunnus albacares) and Blue Marlin (Makaira nigricans)22

MICHAEL M. WALKER23

The problems we face in understandinganimal navigation are epitomized by the mi­grations of pelagic marine fishes. It is veryunlikely that fish could use the sun compass orother known orienting mechanisms to guidepelagic migrations. There is a growing body ofindirect evidence that magnetic sensitivity is avery important orienting mechanism for mi­gratory animals. This paper reports successfulconditioning of yellowfin tuna (Thunnusalbacares) to magnetic fields and an investi­gation of the possibility that the physical basisfor a magnetic sense in both the yellowfin andthe blue marlin (Makaira nigricans) is bio­genic magnetite deposited within the ethmoidbones of the skulls of both species.

Yellowfin tuna were trained using a discretetrials-fixed interval testing paradigm to dis­criminate between the normal Hawaiian mag­netic field and an altered field generated by acoil encircling their tank. Fish were trained toproduce or withhold a conditioned response,i.e., swimming through a plastic pipe framelowered into the water for 30 sec trial periods,in anticipation of positive or negative rein­forcement. Studies using the cryogenic mag­netometers at Hawaii Institute of Geophysics

22 The author gratefully acknowledges support for thiswork from the National Marine Fisheries Service Hono­lulu Laboratory and a travel grant from the GraduateStudent Organization, University of Hawaii.

23 University of Hawaii, Department of Zoology,Honolulu, Hawaii 96822.

and California Institute ofTechnology set outto identify concentrations of magnetic ma­terial within the bodies of yellowfin tuna andblue marlin. Where concentrations of mag­netic material were found, the tissues weredemagnetized in an alternating field to deter­mine the coercivity of the magnetic material.

All yellowfin tested for magnetic sensitivitylearned to discriminate between the two mag­netic fields after two 30-trial sessions. Controltrials were performed with one fish by inter­rupting the circuit between the DC powersupply and the coil around the experimentaltank. All other procedures continued as before,but discrimination between the positively andnegatively reinforced trials ceased. Discrimi­nation by the fish was reestablished when thecircuit was recompleted providing good evi­dence for the existence of a magnetic sense inthese fish.

Concentrations of magnetic material werepredictably found within the ethmoid bonesof the skull of both the yellowfin tuna and theblue marlin. Examination of the coercivityand other properties of the magnetic materialshowed that it was single-domain magnetitein both cases. Theoretical analyses imply thatthe amount of magnetite present is sufficientto provide these fish with a very sensitive mag­netoreceptor if the magnetite is linked to thenervous system. This finding of magnetite inthe same structure in species from differentfamilies of pelagic fish strongly suggests thatmagnetite deposited within the ethmoid bonesis the basis of the magnetic sense organ.

Abstracts of Papers

Conspecific Attraction in a Territorial Damselfish:Analysis of the Spatial Distribution of Stegastes fasciolatus

WILLIAM J. WALSH24

517

The spatial dispersion patterns of a per­manently territorial, algivorous damselfishwere examined before and after an unusuallysevere storm. As a result of the storm, largenumbers of individuals shifted their territoriesfrom shallow reef flat areas to deeper, moreprotected ones. Territorial movements inthese areas during the early post-storm periodresulted in a distinct clumping of territorieswhich continued throughout the first post­storm year.

Comparisons between mapped distributionpatterns and the descriptions provided bythree ecological measures of dispersionshowed marked differences in the accuracy

24 University of Hawaii, Department of Zoology,

Honolulu, Hawaii 96822.

and usefulness of these procedures. Pointnearest neighbor analysis inadequately de­scribed observed patterns. Circle nearestneighbor analysis and Morisita's index of dis­persion provided increasing amounts of infor­mation and reliability.

Post-storm aggregation was unrelated totopographic heterogeneity or to the influenceof competitors or predators. Attraction be­tween conspecifics was implicated as being theprimary driving force producing the observedterritorial aggregation. Consistent with thisinfluence were the patterns ofsubadult coloni­zation and the adult reoccupation of vacatedterritories. A review of other studies suggeststhat conspecific attraction may be an impor­tant and often underestimated factor in thedetermination ofspatial patterns among coralreef fishes.