USF Libraries
USF Digital Collections

Characterizing sound production in nearshore rockfishes (Sebastes spp.)

MISSING IMAGE

Material Information

Title:
Characterizing sound production in nearshore rockfishes (Sebastes spp.)
Physical Description:
Book
Language:
English
Creator:
Nichols, Bryan
Publisher:
University of South Florida
Place of Publication:
Tampa, Fla
Publication Date:

Subjects

Subjects / Keywords:
Agonistic
Territorial
Gas bladder
Sebastes nebulosus
Fish
Dissertations, Academic -- Marine Science -- Masters -- USF
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Abstract:
ABSTRACT: Rockfishes of the genus Sebastes are trophically important in most nearshoreenvironments of the west coast of North America, and support important commercial and recreational fisheries. Though the 72 northeast Pacific species have drumming muscles along their swim bladders, little is known about sound production in the genus. Sounds produced by nearshore rockfish were recorded using fixed hydrophones and underwater video in aquaria, and acoustical dataloggers in the field. Sounds were analyzed from six species: Sebastes nebulosus, S. atrovirens, S.carnatus, S. chrysomelas, S. caurinus, and S.maliger. These six species are closely related, mostly bottom dwelling species with similar drumming musculature. No sounds were recorded from twelve other species inresponse to diver harassment or agonistic interactions. All the sound production observed was close range, agonistic and relatively quiet (estimated source level 122 dB re 1uPa). Sounds were recorded at all times of the day in the field, presumably from S. nebulosus. While courtship type behavior was observed and video-recorded for only two species, no sounds were produced during courtship. Analysis of the sounds (duration, number of pulses, pulse rate, peak frequency, interpulse interval) from the six species showed considerable overlap between call characteristics, although S. carnatus and S.chrysomelas produced some longer calls with more pulses than the other species. It appears that these agonistic sounds are designed for short-range communication and that they are not species-specific.
Thesis:
Thesis (M.A.)--University of South Florida, 2005.
Bibliography:
Includes bibliographical references.
System Details:
System requirements: World Wide Web browser and PDF reader.
System Details:
Mode of access: World Wide Web.
Statement of Responsibility:
by Bryan Nichols.
General Note:
Title from PDF of title page.
General Note:
Document formatted into pages; contains 48 pages.

Record Information

Source Institution:
University of South Florida Library
Holding Location:
University of South Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001912367
oclc - 173818060
usfldc doi - E14-SFE0001261
usfldc handle - e14.1261
System ID:
SFS0025582:00001


This item is only available as the following downloads:


Full Text
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001912367
003 fts
005 20071008112902.0
006 m||||e|||d||||||||
007 cr mnu|||uuuuu
008 071008s2005 flu sbm 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0001261
040
FHM
c FHM
035
(OCoLC)173818060
049
FHMM
090
GC11.2 (ONLINE)
1 100
Nichols, Bryan.
0 245
Characterizing sound production in nearshore rockfishes (Sebastes spp.)
h [electronic resource] /
by Bryan Nichols.
260
[Tampa, Fla] :
b University of South Florida,
2005.
3 520
ABSTRACT: Rockfishes of the genus Sebastes are trophically important in most nearshoreenvironments of the west coast of North America, and support important commercial and recreational fisheries. Though the 72 northeast Pacific species have drumming muscles along their swim bladders, little is known about sound production in the genus. Sounds produced by nearshore rockfish were recorded using fixed hydrophones and underwater video in aquaria, and acoustical dataloggers in the field. Sounds were analyzed from six species: Sebastes nebulosus, S. atrovirens, S.carnatus, S. chrysomelas, S. caurinus, and S.maliger. These six species are closely related, mostly bottom dwelling species with similar drumming musculature. No sounds were recorded from twelve other species inresponse to diver harassment or agonistic interactions. All the sound production observed was close range, agonistic and relatively quiet (estimated source level 122 dB re 1uPa). Sounds were recorded at all times of the day in the field, presumably from S. nebulosus. While courtship type behavior was observed and video-recorded for only two species, no sounds were produced during courtship. Analysis of the sounds (duration, number of pulses, pulse rate, peak frequency, interpulse interval) from the six species showed considerable overlap between call characteristics, although S. carnatus and S.chrysomelas produced some longer calls with more pulses than the other species. It appears that these agonistic sounds are designed for short-range communication and that they are not species-specific.
502
Thesis (M.A.)--University of South Florida, 2005.
504
Includes bibliographical references.
516
Text (Electronic thesis) in PDF format.
538
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
500
Title from PDF of title page.
Document formatted into pages; contains 48 pages.
590
Adviser: David A. Mann, Ph.D.
653
Agonistic.
Territorial.
Gas bladder.
Sebastes nebulosus.
Fish.
690
Dissertations, Academic
z USF
x Marine Science
Masters.
773
t USF Electronic Theses and Dissertations.
4 856
u http://digital.lib.usf.edu/?e14.1261



PAGE 1

Characterizing Sound Production in Nearshore Rockfishes ( Sebastes spp .) by Bryan Nichols A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science College of Marine Science University of South Florida Major Professor: David A. Mann, Ph.D. John C. Ogden, Ph.D. Jose Torres, Ph.D. Date of Approval: July 28, 2005 Keywords: agonistic, territorial, gas bladder, Sebastes nebulosus fish Copyright 2005, Bryan Nichols

PAGE 2

Acknowledgements This project could not have been comp leted without the guidance and assistance of David Mann. Thanks also to committee members John Ogden and Jose Torres, who were willing to help out with an exploratory project. Special thanks to Jeff Christensen at the Seattle Aquarium, who was extremely helpful and interested. John McInerney, formerly at the University of Victoria, supe rvised two of the earlier studies in British Columbia and dug through his paperwork to he lp this one. Val Fletcher offered advice, and his 1983 Masters study was the proof that at least some rockfish vocalize. John Yearsley studied rockfishes in my own home town 35 years ago. Extra special thanks to Lucy Littlejohn, formerly Lucy Wold, whos e 1988 experiment at Moss Landing provided the voices of three California sp ecies of rockfish, and who ha ppily loaned me the tapes, not to mention provided dinne r and a place to spend the ni ght at her ranch in Idaho. Milton Love contributed to an excelle nt 2002 book on rockfishes with the best introduction any ichthyological text is likely to have – rockfishes are cool indeed, and diving with them constantly inspires my curiosity and admiration. Finally, thanks to Amanda Linville, who engaged, enraged and ev entually entrapped the other half of my brain throughout the process.

PAGE 3

i Table of Contents List of Tables................................................................................................................. .....ii List of Figures................................................................................................................ ....iii Abstract....................................................................................................................... .......iv INTRODUCTION..............................................................................................................1 Drumming Muscle Anatomy..................................................................................2 Ocean Sound...........................................................................................................3 Passive Hydroacoustics...........................................................................................5 Rediscovering Lost Data.........................................................................................5 METHODS........................................................................................................................ .6 Moss Landing 1988.................................................................................................6 Bamfield 2004.........................................................................................................6 Hydrophone Recordings.............................................................................7 Datalogger...................................................................................................7 Captive Rockfishes.....................................................................................9 Seattle Aquarium 2005...........................................................................................9 Analysis.................................................................................................................11 Temporal Data..........................................................................................11 Individual Calls.........................................................................................12 RESULTS........................................................................................................................ .17 Hydrophone Recordings.......................................................................................17 Call Analysis.........................................................................................................17 Qualitative Observations.......................................................................................23 Datalogger.............................................................................................................33 DISCUSSION...................................................................................................................38 Swimbladder muscle type.....................................................................................38 Speciation..............................................................................................................39 Growls and Rumbles.............................................................................................40 Reasons for Calling...............................................................................................43 Temporal Data......................................................................................................44 Further Research...................................................................................................44 REFERENCES.................................................................................................................46

PAGE 4

ii List of Tables Table 1 Summary of field datalo gger deployments on the west coast of Vancouver Island...........................8 Table 2 Rockfish species known to be present in the Seattle Aquarium’s large tank at the time of this study, including their observed relationship to the hydrophone.................................................10 Table 3 Rockfish calls were placed into the followi ng groups (labeled by three letter codes) for comparative analys is........................................................................................................... ..............15 Table 4 List of rockfish species at the Seattle Aquarium showing swimbladder muscle type (Hallacher 1974) as well as success at recordi ng sounds produced in association with three different types of behavior.................................................................................................... ............22 Table 5 Species assemblage observed at th e three datalogger de ployment sites...........................................34

PAGE 5

iii List of Figures Figure 1 Two types of Sebastes swimbla dder muscles, showing right side only............................................4 Figure 2 Oscillogram and power spectrum showing background noise in the Seattle Aquarium, taken from just after an S. nebulosus chase and call from which absolute sound levels were calculated................................................................................................................ ..............14 Figure 3 Oscillogram (top) showing a series of pop sounds followed by a short growl................................19 Figure 4 Oscillogram and power spectrum of a gr owl type call................................................................... .20 Figure 5 Oscillogram (top) and power spect rum (middle) of a ru mble type call..........................................21 Figure 6 Means plus or minus one standard deviation for each of the call characteristics measured............25 Figure 7 Means plus or minus one standard deviation for each of th e call characteristics measured............26 Figure 8 Histograms of peak frequencies (Hz) for the four rockfish species studied....................................27 Figure 9 Histograms showing the number of pulses per call in each of the four species studied..................28 Figure 10 Histograms showing the call du ration (in seconds) for each species............................................29 Figure 11 Pulse rate (pulses/s) hist ograms for the four species studied........................................................ 30 Figure 12 Histograms for each of the species studied showing coefficient of variation for the interpulse interval............................................................................................................ ..............31 Figure 13 Number of pulses pe r call plotted ag ainst du ration................................................................... ....32 Figure 14 Calling behavior from two deployments of a datalogger next to the apparent den site of a China rockfish ( S. nebulosus ) at a field site near Ohia t Islet, British Columbia.........................35 Figure 15 Percentage of 30 second recordings per hour in which boat noise significantly masked other sounds, averaged over deployme nts D1 and D2 at Ohiat Islet.............................................36 Figure 16 Sums of the 30 second blocks containi ng fish growls in each hour of the day, taken from the four datalogger deployments where the internal microphone overrode the hydrophone..................................................................................................................... ...............37 Figure 17 Pulse rate plotted against number of pulses for all Sebastes recorded (n=365)............................42

PAGE 6

iv Characterizing Sound Production In Nearshore Rockfishes ( Sebastes spp .) Bryan Nichols Abstract Rockfishes of the genus Sebastes are trophically importa nt in most nearshore environments of the west coast of North Am erica, and support important commercial and recreational fisheries. Though the 72 northeast Pacific spec ies have drumming muscles along their swim bladders, little is know n about sound production in the genus. Sounds produced by nearshore rockfish were reco rded using fixed hydrophones and underwater video in aquaria, and acoustical dataloggers in the field. Soun ds were analyzed from six species: Sebastes nebulosus, S. atrovirens, S.c arnatus, S. chrysomelas, S. caurinus and S. maliger These six species are closely relate d, mostly bottom dwelling species with similar drumming musculature. No sounds were recorded fro m twelve other species in response to diver harassment or agonistic interactions. All the sound production observed was close range, agonistic and relatively quiet (estimated source level 122 dB re 1Pa). Sounds were recorded at all times of the day in the field, presumably from S. nebulosus While courtship type behavior was observed and video-recorded for only two species, no sounds were produced during courtship. Analys is of the sounds (duration, number of pulses, pulse rate, peak frequency, interpulse interval) from the six species showed considerable overlap between call characteri stics, although S. carnatus and S. chrysomelas produced some longer calls with more pulses than the other species. It appears that these agonistic sounds are designed for short -range communication and that they are not species-specific.

PAGE 7

1 INTRODUCTION West coast rockfishes ( Sebastes spp .) are important predators over many nearshore habitats from Calif ornia to Alaska. Smaller spec ies as well as juveniles are prey for a variety of other fishes, birds and mammals. Rockfishes, members of the family Scorpaenidae, are long-lived ( S. aleutianus Jordan & Evermann has been aged to 205 years) and many species do not reach sexua l maturity for over a decade. There are approximately 102 species in the genus, a remarkable diversification. From a human perspective, they are also the target of large commercial a nd recreational fi sheries, though their numbers have dropped catastrophically ove r much of their range (Love et al. 2002). Unlike most fishes, rockfishes have inte rnal fertilization – they mate and the embryo develops inside of the female At least some and probably all are matrotrophically viviparous – that is, the embryos receive energy directly from their mother as well as a yolk sac. Rockfishes ar e long lived and fecund – successful larval settlement years may be a decade apart. All species have at least basic venom glands at the base of some fin spines. There are at l east 65 species along the west coast of North America, with the greatest diversity (56 speci es) off Southern California. Rockfishes can be found from the intertidal zone to dept hs of over 2800 meters (Love et al. 2002). Many fishes produce sound and Hallacher ( 1974) showed that numerous rockfish species have sonic muscles alongside their swim bladders. At least three Northwest species are capable of growli ng loud enough for a diver to hear (personal observation): China ( S. nebulosus Ayers), copper ( S. caurinus Richardson) and quillback rockfish ( S. maliger Jordan & Gilbert). One unpublished stud y (Fletcher 1983) recorded agonistic sounds produced by two species: S. nebulosus and black rockfish ( S. melanops Girard). Considering the economic importance of th e various rockfish fisheries and the threatened status of many populations, there is a surprising lack of information on sound production in this genus. Very little has b een published on the topic and it is largely

PAGE 8

2 unknown which species actually do produce sounds, what sort of sounds they are, and why they might do it. Reasons for this may include the facts that: many people that are familiar with ro ckfishes have no idea they produce sound as they typically don’t wh en feeding, caught or handled rockfishes do not appear to chorus loudly like some soniferous fishes most species do not produce sound even when harassed, so scuba divers do not hear them fish sound researchers are concentrated on the East Coast Drumming Muscle Anatomy Sound is important to fish, and indeed to nearly all verteb rates (Popper et al. 2003). Communication is not the only reason underwater, where sound travels much farther than light, sound pr ovides sensory information from much greater distances. Tavolga (1964) discusses the anatomy, phys iology and sound production characteristics of extrinsic swimbladder musculature in a variety of fish families. Though the exact anatomy differs between and within families, sounds are produced by contraction of muscles near the swimbladder, causing it to either expand or cont ract. The resulting pressure and volume changes cause the bla dder to pulsate, producing displacement which passes through the fish into the surrounding wa ter. He reported no cases where there was a muscular antagonist to the swimbladder musc les, observing instead that the elasticity of associated skeletal and connective tissue caused the muscles to return to their initial state. Tavolga called this a “highl y efficient system” functioning as a low frequency underwater loudspeaker. The large number of species in the Sebastes genus makes it an interesting topic for taxonomists, and there has b een considerable shuffling at the genus and species level over the years. Matsubara (1943) considered the swimbladder muscles of Japanese scorpaenids to be excellent taxonomic charac teristics. Hallacher (1974) examined the “gasbladder” muscles of North American ro ckfishes, separating them into two major categories with three or four subdivisions each. He concluded that most species were capable of producing sounds. In cases where bot h sexes of a species were examined, he

PAGE 9

3 found no sexual dimorphism in the swimbladde r musculature. Figure 1 illustrates the basic anatomy of Hallacher’s type I a-z rock fish muscles, the most common type in the genus. Every species found to produce sound in this study had the rarer type II a-v musculature, also illustrated in Figure 1. Ha llacher found this musc le group to be the largest in the genus. Ocean Sound Popper et al. (2003) reviewed studies s howing that anthropog enic ocean sound is increasing and may be affecting fish popul ations. The increasing sound comes from a variety of sources, especially shipping, pleasure craft, se ismic surveys, pile driving, sonar, and scientific explora tion, all of which occur in the continental shelf and nearshore habitats occupied by rockfish species. Though no work has been done on hearing in rockfishes, Hallacher’s (1974) anatomical studies showed th at rockfishes have large otoliths, and that the swimbladder muscles originate in the cranium “in the general proximity” of the otoliths. Until sufficient hearing studies are conducted, it can be assumed that fish are unlikely to prod uce sounds they cannot hear themselves. A growing body of knowledge indicates that noise in the oceans can affect fish adversely. Increasing background noise may “mask” sounds that are biologically important to fish (Fay and Megela Simmons 199 9). As the inner ear se nsory hair cells of fishes are remarkably similar to our own, louder noises, such as those from seismic testing, may cause permanent or short term hearing loss (Popper et al. 2005). Finally, loud noises may induce physiological stress re sponses in fish (Smith et al. 2004). All of these factors could be highly disruptive to fish behavior and ecology. In order to make the best decisions regard ing the effect of increasing ocean sound on rockfish populations, it will be necessary to have a much better understanding of how, why and when they use sound.

PAGE 10

4 Figure 1 Two types of Sebastes swimbladder muscles, showing right side only. Type I a-z is the most common type in the genus; the diagra m shows a) striated muscle, b) tend on, c) attachment to the pectoral girdle, and d) Baudelot’s ligament. The muscles originate on the occipital cranium and insert on the ribs. Type II a-v is less common, but contains all the species investigated in this study. Type II a-v muscles are relatively larger and bypass the pectoral girdle. The tendon attaches to the swimbladder (e) and inserts onto the vertebral parapophyses (f). Adapted from Hallacher (1974).

PAGE 11

5 Passive Hydroacoustics Passive hydroacoustic techniques, where fi xed or drifting hydrophones are used to listen for fish, are becoming more common. Lucz kovich et al. (1999) showed that passive hydroacoustic surveys coul d delineate weakfish ( Cynoscion regalis ) spawning areas, and argued that that passive acoustic techniques could be a valuable tool for management, including the designation of marine reserves. Locasci o & Mann (2005) developed a technique using hydrophones attached to hous ings with dataloggers to show diel periodicity in fish sound produc tion off Florida. If rockfi sh sound production was better understood, the rapidly advanci ng field of passive acoustics could help supplement fisheries independent data, which is limite d or nonexistent for most rockfish species. Rediscovering Lost Data An extensive literature search turned up very l ittle on the topic of sound production in rockfishes, although three unpubl ished studies proved useful. One, a Master’s thesis (Fletcher 1983), is available as microfiche from the Library of Canada but was not otherwise published. Another (Yearsley 1970) existed as a vers ion that was cited as “in press” by a number of papers, but appa rently never made it to publication. An early version was read by this author after tracking down the original supervising professor and requesting a copy. This study was undertaken based on personal observations and the evidence of these papers, which cumulatively stat ed that at least five species of nearshore rockfishes had produced sounds in either the lab or field. Another important piece of the puzzle came from Greg Caillet, an ichthyologist at Moss Landing who remembered a graduate st udent who had recorded rockfish sounds sixteen years previous. That student, Lucy Wold (now Lucy Littlejohn), was tracked to a ranch in Idaho, and the tapes and observations she made (Wold 1991) were incorporated into this study.

PAGE 12

6 METHODS Moss Landing 1988 VHS tapes originally recorded in 1988 at the Moss Landing Marine Lab were retrieved and digitized (44.1 kHz) as AVI files using a Sony camcorder (DCR-PC110) to record them directly to a computer hard driv e. Fish sounds were then saved as individual files by watching the recordings in Adobe Premiere Pro and cutting the footage into individual events, typically chases. Sound was extracted from these ev ents as a WAV file for analysis. Audio tapes recorded in the 1988 study were also digitized directly into Syntrillium Software Corporation’s Cool Edit 2000 (6 kHz sample ra te), and individual fish calls were separated for analysis. Bo th video and audio tapes were originally recorded as part of a Masters Thesis (Wold 1991) and contain footage and hydrophone audio of three closely related specie s of California rockfishes: kelp ( S. atrovirens Jordan & Gilbert), gopher ( S. carnatus Jordan & Gilbert ) and black & yellow ( S. chrysomelas Jordan & Gilbert). Sound analysis was not publ ished at the time as the focus of the study was on larval development. Because of di fficulties involving calibration of both the hydrophone and original video system, absolute sound levels were not calculated for these species in this study. Bamfield 2004 Since so little was known about rockfi sh sound production in the field, a wide variety of data collection tec hniques were attempted at th e Bamfield Marine Sciences Centre on the west coast of Vancouver Is land, British Columbia, Canada. Data were collected in June and July, 2004. Fishes were captured under a permit from the Department of Fisheries and Oceans, Cana da and all research was approved by the Institutional Animal Care and Use Committ ee of the University of South Florida.

PAGE 13

7 Hydrophone Recordings A series of hydrophone recordings were made by hanging a hydrophone (HTI 96min; -164 dBV/Pa) from a dock, small boat or kayak. The recordings were originally made to a micro cassette record er, then digitized. These were done at nearshore locations, including sites in 100 meter deep Trevor Channel, in order to establish typical background noises and ascertain if fish sounds could be heard. Datalogger A datalogger system including a hydrophone (HTI 96min; -164 dBV/Pa) and battery pack attached to a Toshiba E755 PocketPC in an underwater housing was deployed in several locations near the field station. Sound was recorded for 30 second intervals every five minutes with a samp ling rate of 8820 Hz. The datalogger was primarily deployed just north of Ohiat Islet, where it was placed by divers next to a S. nebulosus “den” (at least one individual was seen in close proximity on every dive and it is likely that it was the same fish each time) In some of the recordings the internal microphone of the PocketPC overrode th e hydrophone input. Fish sounds were nonetheless clearly audible in these record ings, and while they were not used for individual call analysis, the data were us eful to show temporal patterns of sound production. Table 1 summarizes the datalogger deployments. The Reef Environmental Education Founda tion (REEF) roving diver fish survey technique (Schmitt et. al 1998) was used at each location where the datalogger was deployed to help characterize the fish assembla ge. At least four such surveys were done at each site. A video camera (Sony PC110) was set up in an underwater housing with a hydrophone attached (HTI 96min; -164 dBV/Pa). The camera was used by divers, mostly at the Ohiat Islet site to attempt to record rockfi sh sounds in the field. Various species were approached underwater, especially: S. nebulosus S. melanops S. caurinus and S. maliger Due to the apparently imperturbable, even cu rious nature of most rockfishes at the site, it was difficult to provoke sounds.

PAGE 14

8 Table 1 Summary of field datalogger deployments on the west coast of Vancouver Island. Dp is deployment number, Min is the total minutes recorded (30 seconds at a time). Depths are in meters. For deployments 1 and 2, the system worked properly – for deployments 4 through 8, the internal microphone overrode the hydrophone input. Dp Date In Date Out Min Location Lat N Long W Depth D1 2004-06-23 13:24 2004-06-24 16:15 180Ohiat 48 51.356 125 10.973 12 D2 2004-06-28 12:00 2004-06-29 14:51180Ohiat 48 51.356 125 10.973 12 D4 2004-07-05 12:00 2004-07-07 04:02 481Ohiat* 48 51.356 125 10.973 15 D5 2004-07-08 17:00 2004-07-10 09:02481Ohiat* 48 51.356 125 10.973 15 D6 2004-07-10 14:00 2004-07-12 16:28 433Blackfish* 48 50.80 125 09.65 13 D7 2004-07-13 14:00 2004-07-14 17:01481Gobytown*48 50.59 125 07.87 8 D8 2004-07-15 16:00 2004-07-16 20:35 291Ohiat* 48 51.356 125 10.973 15

PAGE 15

9 Captive Rockfishes Two copper ( S. caurinus ) and two China ( S. nebulosus ) rockfish were captured and brought back to the lab. The copper ro ckfish were collected by hook and line from shallow water at the edge of a kelp bed near the marine station; the China rockfish were collected by divers between 10 and 15 meters de ep near the Ohiat site, using a one meter jig. After allowing sufficient time (up to 48 hrs for S. nebulosus ) for their swim bladders to adjust, the fish were put in shallow outdoor tanks (approximately 1.5 meters square) in an area open on three sides but pr otected from the rain. After more time to acclimate to their surroundings, various attempts to elicit sounds were made. These included manually harassing the fish, placi ng a mirror into the tank, placing a small decorator crab ( Scyra sp. ) and larger red rock crab ( Cancer productus ) nearby, and placing one fish in with another. The S. caurinus individuals, both rela tively small, never acclimated to captive conditions well enough to take food. The S. nebulosus individuals only began to take food in the days before the study ended. Seattle Aquarium 2005 The Seattle Aquarium was visited in Febr uary 2005, when a number of species of rockfishes were reported to be breeding. Res earch was conducted in the aquarium’s large, walk through tank, which holds approximately fifteen rockfish species, as well as numerous other Northwest fishes. A hydr ophone (HTI 96min; -164 dBV/Pa) on a long cable was connected to a computer and/or video camera inside the viewing room, so visual observations could be made at the same time as audio recordings. Rockfishes of various species were watched at all daylight hours over the course of four days, while near continuous sound and occasional video/sound recordings were made. The computer was calibrated with a test tone of a known vo ltage so that received sound levels could be calculated. Sound data were analyzed in Cool Edit and separated into individual calls, especially the ones that were captured on video and could be associated with a certain species.

PAGE 16

10 Table 2 Rockfish species known to be present in the Seattle Aquarium’s large tank at the time of this study, including their observed relationship to the hydrophone. The last column shows which species were known to be breeding around the time of the study. common species near most of the time often near often passing breeding china nebulosus Y Y canary pinniger Y 1 mo prev. tiger nigrocinctus Y ? widow entomelas Y ? brown auriculatus Y Y black melanops Y ? yelloweye ruberrimus Y Y quillback maliger Y Y copper caurinus Y Y yellowtail flavidus Y N rosy rosaceus Y Y redbanded babcocki Y ? Puget Sound emphaeus Y ? vermillion miniatus Y ? blue mystinus Y N

PAGE 17

11 A diver also entered the tank with a hydrophone (HTI 96min; -164 dBV/Pa) attached to a housed video camera, and attempte d to elicit sounds fr om various species of rockfishes, mainly by harassment but on a coup le occasions by “herdi ng” one territorial species into another’s territory. Due to logistics as well as ambient noi se levels, the fixed hydrophone stayed in one particular section of the tank, mainly in between four pilings above a rocky bottom. Table 2 describes the approximately 15 rockfi sh species in the tank, and their relationship to the hydrophone. According to aquarium bi ologist Jeff Christia nsen, it was possible (but “not likely”) there ma y have been chilipepper ( S. goodie Eigenmann) and/or dusky ( S. ciliatus Tilesius) rockfish mixed with the widow rockfish ( S. entomelas Jordan & Gilbert). Table 2 also shows which species were known to be breeding near the time of the study, also according to the aquarium biologist. Analysis Temporal Data Data from the datalogger deployments in Bamfield were grouped by time of day into hour-long blocks. Each thirty-second reco rding that contained recognizable fish sounds (defined as at least one growl of thr ee beats or more) was tallied in the hour long block in which it occurred. The sums were then divided into the tota l number of blocks in that hour to come up with a percentage. Time intervals wh ich contained only pop type vocalizations (presumably a si ngle hit on the bladder) were not included, in order not to mistake pops for sounds that were not produced by rockfishes. Boat noise was tallied from the first two Ohiat deployments by su mming the number of 30 second intervals in which boat noise (including engines, nets and anchors) significantly masked other sounds. These were then averaged and expresse d as a percentage of the total number of 30 second intervals per hour. Due to technical difficulties, only the fi rst two deployments (six hours recorded over approximately two days) contain valid recordings from the hydrophone – calls on these deployments were included in subseque nt call analysis. During the remaining five

PAGE 18

12 deployments, which included 36 hours of reco rdings taken over approximately seven days, the internal microphone of the PDA overrode the hydrophone input. Despite this, there were recognizable fish calls in the data while these were not considered as part of the call analysis, the 30 second blocks in which they occurred were tallied as simple sums in the hour in which they occurred, to add to the temporal data. Individual Calls Individual rockfish calls (saved as WA V sound files) were analyzed using Cool Edit 2000 as well as HotWav, a custom de signed MATLAB (The Math Works, Inc.) program. Background noise can be considerable a nd variable in aquaria (see Figure 2), so some calls with a poor signal to noise ratio co uld not be used. Also, sounds that appeared to be one call superimposed on another were left out of the analys es, though this raises important questions that can only be answer ed by further studies that include close observation, video and more than one hydrophone in order to isolate which fish is making which call. For the purposes of this study, a “call” wa s defined as a series of closely spaced pulses. Calls containing less th an three pulses were not c onsidered in the analysis. Rockfish calls were categorized for analysis in a number of ways (see Table 3). The species with enough recorded calls to be compared were S. nebulosus (n=207) S. atrovirens (n=64) S. carnatus (n=34) and S. chrysomelas (n=53). At the Seattle Aquarium, calls were elicited by harassment from S. caurinus (n=3) and S. maliger (n=2), and observed agonistically in S. caurinus (n=1), but not enough times for statistical analysis. Furthermore, there were reasonably clear calls on one of th e datalogger sets in an area where the region’s primary sound producer, S. nebulosus was not observed, but where several mature brown rockfish ( S. auriculatus Girard) were seen upon deployment and retrieval of the datalogger. In addition to comparing calls between species, a distinction was made between the “harassed” calls of S. nebulosus where a diver elicited a call from a single fish, and agonistic calls by the same species, where two individuals interacted and aggression was observed. Calls by S. nebulosus captured recently from the wild in Bamfield, British

PAGE 19

13 Columbia, were compared with those of a quarium residents in Seattle. Finally, captive calls of known species were compared to sounds recorded from a datalogger set on a S. nebulosus den in the wild off British Columbia. Background noise levels were particular ly high in the frequency bands that rockfishes call at during studies at the Seattle Aquarium (Fi gure 2) and for the Bamfield captives. This decreases the signal to noise ration and makes analysis more difficult.

PAGE 20

14 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -0.04 -0.02 0 0.02 0.04 Level (V)Time (s) 0 500 1000 1500 2000 2500 3000 20 40 60 80 100 120 140 Frequency (Hz)dB re 1uPa/sqrt(Hz) Figure 2 Oscillogram and power spectrum showing background noise in the Seattle Aquarium, taken from just after an S. nebulosus chase and call from which absolute sound levels were calculated.

PAGE 21

15 Table 3 Rockfish calls were placed into the following groups (labeled by three letter codes) for comparative analysis. To focus on species differences, the five S. nebulosus groups were considered together for some analyses. Code Species Description # of calls nSa China nebulosus Seattle agonistic: fish interactions observed on video 21 nSh nebulosus Seattle harassed: sounds provoked by a diver 27 nBa nebulosus Bamfield agonistic: sounds recorded between captive fish 31 nSm nebulosus? Seattle miscellaneous: sounds recorded with no accompanying video but presumed to be S. nebulosus due to hydrophone location and fish activity 49 nBw nebulosus? Bamfield wild: sounds recorded from a datalogger deployed next to a S. nebulosus den 79 aMa kelp atrovirens Moss agonistic: sounds recorded from captive fish 64 cMa gopher carnatus Moss agonistic: sounds recorded from captive fish 34 yMa black & yellow chrysomelas Moss agonistic: sounds recorded from captive fish 53

PAGE 22

16 In order to compare calls between th e different categories, a number of characteristics of each call were measured. These included: 1. call duration – the amount of time between the beginning and end of the call, estimated from the peak of the first pulse until the peak of the last recognizable pulse 2. number of pulses – the number of pu lses contained within a call, inferred to correspond with the number of strikes on the swimbladder 3. peak frequency – the frequency that contained the highest energy levels, calculated in a custom MATLAB program 4. the coefficient of variation (SD/me an) of the interpulse intervals, measured from pulse peak to pulse peak Amplitude was not considered in this anal ysis because of the potential differences in hydrophone sensitivity, as well as variable and/or unknown distances between the calling fish and the hydrophone. Rockfishes typica lly call during a chase, so the distance of the calling fish to the hydrophone often ch anged throughout the duration of the call. Means and standard deviations were calcu lated for each of the call characteristics in order to show differences between categor ies and species. However, small differences in the means of any of the call characteristics are not likely to be biologically relevant, so more complex statistical comparisons were not attempted. Source level (sound pressure level at 1 m from the hydrophone) was calculated from one call of S. nebulosus recorded at the Seattle Aquari um. This call was made in an aggressive chase approximately 1 m from the recording hydrophone. Source level was calculated as the route mean square level of the entire call and the peak-peak level of the loudest pulse.

PAGE 23

17 RESULTS Hydrophone Recordings Thirty-seven different hydrophone record ings (duration 30 to 60 seconds) were made at a variety of locations, including th e dock at the Bamfield Marine Research Station and the dive site where rockfishes of a variety of species were known to live. Numerous snaps and clicks of a likely bi ological origin were recorded, possibly generated by crustaceans or the jaws of su rfperch (Family Embiotocidae). However, no sounds positively recognizable as rockfishes were detected on any of these recordings. During the day, boat noise, even from distant vessels, was probably loud enough to often mask rockfish calls that were not relatively close to the hydrophone. Call Analysis Peak sound pressure level was calculated from a chase ( S. nebulosus ) at the Seattle Aquarium where the fish were observe d at a distance of approximately one meter from the hydrophone, and came to 139 dB re 1 Pa (peak). The root mean square sound pressure level was calculated as 111 dB re 1 Pa (RMS) after the signal was low pass filtered at 450 Hz to remove background noise. To assist with descriptions, the following terms will be used when discussing rockfish calls: 1. POP (Figure 3): a single, discrete hit on the swim bla dder. Pops were observed on their own as well as before and after longer calls, but were not included in call analysis in order to avoid confusion with sounds from other sources. 2. GROWL (Figure 4): a shor t call consisting of a seri es of rapid beats (at least three), often with descending fu ndamental frequencies and amplitudes. 3. RUMBLE (Figure 5): calls longer than two seconds (and typically containing more than 75 pulses).

PAGE 24

18 Yearsley (1970) and Fletcher (1983) both referred to call type two as a burp; both studies as well as this one noted such calls were usually associated with body movement. Neither Yearsley nor Fletcher reported rumble type calls, which this study found to be exclusive to S. carnatus S. chrysomelas and S. atrovirens closely related California species not found in the Northwest. Though twitching behavior that produced mechanical sounds was occasionally recorded, the following species were repeatedly harassed in the Seattle Aquarium without producing detectable growls or pops: S. pinniger S. rocaceous S. entomelas S. mystinus S. melanops and S. miniatus In the Seattle Aquarium, S. nigrocinctus and S. ruberrimus were not only harassed but also recorded while displaying agonistic and/or courtship type behavior, with no associated detectab le sounds. See Table 4 for a summary.

PAGE 25

19 0 0.5 1 1.5 2 2.5 3 3.5 4 -0.1 -0.05 0 0.05 0.1 0.15 LevelTime (s) 0 500 1000 1500 2000 2500 3000 3500 4000 4500 -140 -120 -100 -80 -60 -40 Frequency (Hz)dB Figure 3 Oscillogram (top) showing a series of pop so unds followed by a short growl. Taken from a datalogger set on a S. nebulosus den off Vancouver Island. The power spectrum below is taken from an individual pop. The dB scale is a relative scale since th ese recordings were not from a calibrated system.

PAGE 26

20 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 -0.4 -0.2 0 0.2 0.4 0.6 LevelTime (s) 0 500 1000 1500 2000 2500 3000 3500 4000 4500 -140 -120 -100 -80 -60 -40 -20 Frequency (Hz)dB Figure 4 Oscillogram and power spectrum of a growl ty pe call. Taken from a datalogger set on a S. nebulosus den off Vancouver Island.

PAGE 27

21 Figure 5 Oscillogram (top) and power spectrum (middle) of a rumble type call. The bottom oscillogram is a zoomed in view showing the distinct pulses. 7.4 7.6 7.8 8 8.2 8.4 -0.1 0 0.1 L eve l Time (s) 2 4 6 8 -0.4 -0.2 0 0.2 0.4 L eve l Time (s) 200 400 600 800 1000 -120 -100 -80 -60 -40 Frequency (Hz) dB

PAGE 28

22 Table 4 List of rockfish species at the Seattle Aquari um showing swimbladder muscle type (Hallacher 1974) as well as success at recording sounds produced in association with three different types of behavior. A “Y” means the sound was recorded in association with the behavior for the respective species. A “N” means the behavior was observed n ear a hydrophone but no sound was detected. Mechanical or muscle twitch sounds were not included. A “-“ means the behavior was not observed, or not observed near a hydrophone. Common Species Musculature Harassed Agonistic Courtship China nebulosus II a-v Y Y N copper caurinus II a-v Y Y quillback maliger II a-v Y canary pinniger I a-z N N tiger nigrocinctus I a-z N N widow entomelas I a-z N brown auriculatus I a-z black melanops I a-z N yelloweye ruberrimus I a-z N N yellowtail flavidus I a-z rosy rosaceus I a-z N redbanded babcocki I a-z Puget Sound emphaeus I a-z N vermillion miniatus I a-z N blue mystinus I a-z N

PAGE 29

23 The individual pulses were very short, and likely produced by a single muscle twitch on the swimbladder (Figur e 5). Figures 6 through 12 show the results of analyzing 359 calls from at least four different species. The gra phs make it clear there is considerable overlap in every parameter – peak frequency, call duration, number of pulses, pulse rate and the coeffi cient of variation of the inte rpulse interval. All rockfish calling is low frequency, similar to other species that utilize extrinsic swimbladder musculature. The only discernible characterist ics appear to be the duration and number of pulses, which are directly related (Figure 13). Though any given growl could have come from each of the four species, rumbles appear to be restricted to the California species, especially the S. carnatus/chrysomelas species complex. As it was impossible to discern which indivi dual fish were making the calls in the bulk of this study, it would be impossible to differentiate variance between individuals and variance between species For this reason, means and standard deviations were graphed to display the data, but more comp lex statistical comparisons were not used. Furthermore, minor differences in the means of any of these char acteristics would be unlikely to be biologically relevant, even if they were statistically significant. Figure 6 shows the means each of the call characteristics, plus or minus one standard deviation. While call durat ion and number of pulses stand out, S. nebulosus appears to make short calls whether behaving agonistically or being harassed by a diver. Calls were also short for this species whet her in the Seattle Aquarium, small tanks in Bamfield or in the wild, therefor e subsequent graphs group all the S. nebulosus categories together. Note that higher peak frequency va lues are strictly a result of the wild recordings – the considerable background noise associated with aquaria and laboratory tanks may be partly or sole ly responsible for this. Qualitative Observations In addition to quantifying sounds, some obs ervations made in the field and lab are worth noting: 1) most nearshore rockfishes rarely, if ever, make sound when harassed or handled; of the 15 species present at the Seattle Aquarium, considerable

PAGE 30

24 harassment by a diver only elicit ed calls from three species ( S. caurinus S. maliger and S. nebulosus ) and of those, only one species ( S. nebulosus ) would produce noise regularly 2) certain agonistic behaviors will sometimes be accompanied by sound production; sometimes not 3) sound production usually peaks when the agonistic fish are closest together in a charge/chase 4) though courtship behavior was observe d near a hydrophone for at least two species in Seattle ( S. nebulosus S. emphaeus ) and one in California ( S. atrovirens ), it appears to be visual s ounds were not associated with it 5) captive rockfishes from most source s are unlikely to be immediately suitable for sound or hearing studies due to swim bladder damage. It is unknown how long such injuries would ta ke to recover, or if permanent damage is possible. Exceptions may include some of the more pelagic species (e.g. S. melanops S. flavidus ) which are more likely to be caught near the surface, and/or better able to off gas. 6) the noise level in aquaria may be su fficient to cause significant masking and even hearing damage in rockfishes, making acoustical studies more challenging

PAGE 31

25 Mean SD aMacMayManBanBwnSanShnSm Category -50 0 50 100 150 200 250 300 350 400Peak Frequency (Hz) Mean SD aMacMayManBanBwnSanShnSm Category 20 30 40 50 60 70 80 90 100 110Rate (pulses/s) Mean SD aMacMayManBanBwnSanShnSm Category -20 0 20 40 60 80 100 120 140Number of Pulses Mean SD aMacMayManBanBwnSanShnSm Category 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8Interpulse Interval CV Mean SD aMacMayManBanBwnSanShnSm Category -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5Duration (s) Figure 6 Means plus or minus one standard deviation for each of the call characteristics measured. Graphs are grouped by categories on the X axis the first th ree represent California speci es recorded at the Moss Landing Marine Lab – the latter five are S. nebulosus Categories are aMa – S. atrovirens ; cMa – S.carnatus ; yMa – S.chrysomelas ; nBa – S. nebulosus Bamfield agonistic; nBw – S. nebulosus in situ datalogger; nSa – S. nebulosus Seattle agonistic; nSh – S. nebulosus Seattle harassed; nSm – S. nebulosus Seattle without video. Also see Table 3

PAGE 32

26 Mean SD atrovirens carnatus chrysomelas nebulosus caurinus maliger Species 20 40 60 80 100 120 140 160 180 200 220 240 260 280 300Peak Freqency (Hz) Mean SD atrovirens carnatus chrysomelas nebulosus caurinus maliger Species 20 30 40 50 60 70 80 90Rate (pulses/s) Mean SD atrovirens carnatus chrysomelas nebulosus caurinus maliger Species -20 0 20 40 60 80 100 120 140Number of Pulses Mean SD atrovirens carnatus chrysomelas nebulosus caurinus maliger Species -0.1 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7CV Interpulse Interval Mean SD atrovirens carnatus chrysomelas nebulosus caurinus maliger Species -0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5Duration (s) Figure 7 Means plus or minus one standard deviation for each of the call characteristics measured. Data are grouped by species along the X axis, including the two species where n<20 ( S. caurinus n=4, S. maliger n=2). Note the considerable over lap in every characteristic apart from longer, rumble-type calls.

PAGE 33

27 Peak Frequency (Hz)No of obs Species: atrovirens 553101149197245293341389437485 0 5 10 15 20 25 30 35 40 Species: carnatus 553101149197245293341389437485 0% 1% 3% 4% 6% 7% 8% 10% 11% Species: chrysomelas 553101149197245293341389437485 0 5 10 15 20 25 30 35 40 Species: nebulosus 553101149197245293341389437485 0% 1% 3% 4% 6% 7% 8% 10% 11% Figure 8 Histograms of peak frequencies (Hz) fo r the four rockfish species studied.

PAGE 34

28 Number of Pulses N Species: atrovirens 3.0 35.7 68.4 101.1 133.8 166.5 199.2 231.9 264.6 297.3 330.0 0 20 40 60 80 100 120 140 160 180 Species: carnatus 3.0 35.7 68.4 101.1 133.8 166.5 199.2 231.9 264.6 297.3 330.0 0% 6% 11% 17% 22% 28% 33% 39% 45% 50% Species: chrysomelas 3.0 35.7 68.4 101.1 133.8 166.5 199.2 231.9 264.6 297.3 330.0 0 20 40 60 80 100 120 140 160 180 Species: nebulosus 3.0 35.7 68.4 101.1 133.8 166.5 199.2 231.9 264.6 297.3 330.0 0% 6% 11% 17% 22% 28% 33% 39% 45% 50% Figure 9 Histograms showing the number of pulses per call in each of the four species studied.

PAGE 35

29 Duration (s)No of obs Species: atrovirens 0.020.811.612.403.193.984.775.566.357.157.94 0 20 40 60 80 100 120 140 160 180 Species: carnatus 0.020.811.612.403.193.984.775.566.357.157.94 0% 6% 11% 17% 22% 28% 33% 39% 45% 50% Species: chrysomelas 0.020.811.612.403.193.984.775.566.357.157.94 0 20 40 60 80 100 120 140 160 180 Species: nebulosus 0.020.811.612.403.193.984.775.566.357.157.94 0% 6% 11% 17% 22% 28% 33% 39% 45% 50% Figure 10 Histograms showing the call dura tion (in seconds) for each species.

PAGE 36

30 Rate (pulses/s)No of obs Species: atrovirens 16.940.864.788.7112.6136.5160.4 0 10 20 30 40 50 60 70 Species: carnatus 16.940.864.788.7112.6136.5160.4 0% 3% 6% 8% 11% 14% 17% 19% Species: chrysomelas 16.940.864.788.7112.6136.5160.4 0 10 20 30 40 50 60 70 Species: nebulosus 16.940.864.788.7112.6136.5160.4 0% 3% 6% 8% 11% 14% 17% 19% Figure 11 Pulse rate (pulses/s) histograms for the four species studied.

PAGE 37

31 CV Interpulse IntervalNo of obs Species: atrovirens 0.040.170.310.450.580.720.860.991.131.261.40 0 10 20 30 40 50 60 Species: carnatus 0.040.170.310.450.580.720.860.991.131.261.40 0% 3% 6% 8% 11% 14% 17% Species: chrysomelas 0.040.170.310.450.580.720.860.991.131.261.40 0 10 20 30 40 50 60 Species: nebulosus 0.040.170.310.450.580.720.860.991.131.261.40 0% 3% 6% 8% 11% 14% 17% Figure 12 Histograms for each of the species studied showing coefficient of variation for the interpulse interval.

PAGE 38

32 0123456789 Duration (s) 0 50 100 150 200 250 300 350 Number of Pulses Duration (s):Number of Pulses: r2 = 0.9459 Figure 13 Number of pulses per call pl otted against duration. All six species recorded (n=365) were included.

PAGE 39

33 Datalogger Table 5 shows the fish assemblages at each of the datalogger deployment sites, taken from dive surveys done by the primary au thor and stored in the Reef Environmental Education Foundation database. Note that ro ckfish species make up 29 percent of the total fish species observed. Also note that China rockfish ( S. nebulosus ), which were observed to be by far the most vocal species in the Seattle Aquarium, were seen on every dive at Ohiat but rarely or not at all at the other two sites. Data from the first two datalogger deployments where the hydrophone was fully functional, both conducted at Ohiat Islet, ar e shown in Figure 14. The lighter area in the center of the graph represents daylight hour s, taken from U.S. Naval Observatory’s Astronomical Applications Department ( http://aa.usno.navy.mil/data/docs/RS_OneYear.html ). The approximate beginning and end of civil twilight were used. The datalogger was depl oyed right above a S. nebulosus “den” at a depth of 12 meters. The data show that sound appears to play an important role in the life of that sp ecies – 23.6 percent of the 30-second recordings had recognizable fish growls in them. Interestingly, rockfish gr owls were recorded throughout the day and night, with the most notable lull around 0800. Figure 15 shows boat noise taken from the same two datalogger deployments at Ohiat. Both recreational and commercial fishing for salmon were occurring around the time of the study and are probably responsible for the majority of the boat noise. Masking of calls by boat noise does not appear to be directly responsible for the lows in calling activity that are shown in Figure 14. Figure 16 shows the results of the remaining five deployments, which included 36 hours of recordings taken over approximatel y seven days, when the internal microphone of the PDA overrode the hydrophone input. The results, though limited to only the loudest and/or closest calls, also show calli ng throughout the night and a similar lull in activity around 0800.

PAGE 40

34 Table 5 Species assemblage observed at the three datalogger deployment sites. The number of REEF surveys at each site is noted, and the sighting frequen cy (SF) and density index (DI) taken from the REEF database are provided for each species. REEF surveys record the species seen and an abundance category for each species. The DI is a measure of how many individuals of a species are observed based on a scale of 1-4. It is representative of the abundance category (1 -4) which was most frequently recorded for the species when it was observed. Ab undance categories are Single=1, Few=2, Many=3, and Abundant=4. SiteOhiat It Gobytown Blackfish # of surveys9 8 4 common species SF % DI SF % DI SF % DI Yellowtail Rockfish Sebastes flavidus 1003100 2.8 1003 Black Rockfish Sebastes melanops 100375 2.1 752.6 Blackeye Goby Coryphopterus nicholsi 1002.4100 3.1 1003 Longfin Sculpin Jordania zonope 100263 1.8 1002 Copper Rockfish Sebastes caurinus 1002100 2.1 1002.2 China Rockfish Sebastes nebulosus 1002 251 Kelp Greenling Hexagrammos decagrammus 892.6100 2 1002.2 Quillback Rockfish Sebastes maliger 892.3100 1.7 752 Unidentified Sculpin 782100 2 501.5 Painted Greenling Oxylebius pictus 781.8100 1.7 501.5 Vermilion Rockfish Sebastes miniatus 781.775 2 1002.7 Pile Perch Rhacochilus vacca 781.525 1.5 501.5 Striped Seaperch Embiotoca lateralis 671.888 1.7 751.6 Shiner Surfperch Cymatogaster aggregata 562.288 2.4 502.5 Lingcod Ophiodon elongates 441.5 501 Wolf-Eel Anarrhichthys ocellatus 33125 1 Grunt Sculpin Rhamphocottus richardsoni 33113 1 Kelp Surfperch Brachyistius frenatus 221.550 1.5 251 Spotted Ratfish Hydrolagus colliei 22138 1.3 751.6 Red Irish Lord Hemilepidotus hemilepidotus 221 Juv. Rockfish Sebastes sp. 11225 2 752 Rock Sole Pleuronectes bilineatus 75 2.1 Brown Rockfish Sebastes auriculatus 50 1 Tube-Snout Aulorhynchus flavidus 25 2 Unidentified Flatfish 13 2 C-O Sole Pleuronichthys coenosus 13 1 Longfin Gunnel Pholis clemensi 13 1 Buffalo Sculpin Enophrys bison 13 1 Canary Rockfish Sebastes pinniger 752 Unidentified Rockfish Sebastes sp. 502 Pacific Sandlance Ammodytes hexapterus 253 Spiny Dogfish Squalus acanthias 251

PAGE 41

35 Figure 14 Calling behavior from two deployments of a datalogger next to the apparent den site of a China rockfish ( S. nebulosus ) at a field site near Ohiat Islet, British Columbia. The X axis is hour of the day from midnight to 23:59 – the lighter area in the center represents civil twilight and daylight. The Y axis is the percentage per hour of 30s recordings that co ntained at least one recognizable fish growl.

PAGE 42

36 Figure 15 Percentage of 30 second recordings per hour in which boat noise significantly masked other sounds, averaged over deployments D1 and D2 at Ohiat Islet.

PAGE 43

37 Figure 16 Sums of the 30 second blocks containing fish growls in each hour of the day, taken from the four datalogger deployments where the internal microphone overrode the hydrophone. The graph represents a total of 36 hours worth of 30s recordings, taken from five deployments at three different locales.

PAGE 44

38 DISCUSSION Swimbladder muscle type Though it contains only seven of the eight y two Sebastes spec ies (8.5 percent) examined by Hallacher (1974), his type II av musculature classi fication is obviously important to sound production. Of the fifteen sp ecies of rockfishes housed at the Seattle Aquarium, this study was only able to confir m sounds from three species, each of which had type II a-v musculature (Table 4). Furt hermore, no other members of the group were present in the aquarium, but the other half of them, three species with ranges that end south of Washington State, make up the tota l of Wold’s (1991) study. Not surprisingly then, the rockfish species with the largest drumming musculature appear to be the ones that make the most sound. At the time of the Seattle Aquarium study, it was estimated that well over ninety percent of the ro ckfish calls overheard spontaneously were associated with just one species, S. nebulosus though that could be subject to seasonal changes or crowding. Hallacher’s seventh type II a-v species, S. vexillaris has been subsequently grouped with the wide ranging and dive rsely patterned copper rockfish, S. caurinus All six (or seven) of Hallacher’s type II a-v species belong to the Sebastes subgenus Pteropodus (Kendall 2000). Mol ecular studies have shown S. caurinus is closely related to S. maliger (Gharret et al. 2001), while S. atrovirens is closest to S. chrysomelas and S. carnatus, and many authorities do no t agree on whether the latte r two are even separate species (Love et al. 2002). Th ese six species are closely re lated even for Sebastes, a genus known for rapid speciati on and controversial taxonomy. There are ecological similariti es as well all six type II a-v species are considered benthic (seasonally in the case of S. atrovirens) shallow dwelling species with small home ranges (Love et al. 2002). S. nebulosus S. chrysomelas and S. carnatus are all considered territorial. Though S. atrovirens presently isn’t, Littlejohn (personal

PAGE 45

39 communication) observed a seas onal shift to territorial be havior in captive specimens. This may correspond to a migration from kelp to benthic habitats observed in the wild (Van Dykhuizen 1983). The other muscle group with sound relate d references was type I a-z, which Hallacher (1974) describes as the basic mu scle structure found in the genus. It characterizes 62 of the 82 speci es examined (75.6 percent). Based on his dissections he concludes that most rockfish species ar e capable of sound production. To support this, both Yearsley (1970) and Fletcher (1983) re ported recording sounds from black rockfish ( S. melanops ), a schooling mid water fish with type I a-z musculature, while Yearsley also recorded pops (only) from tiger rockfish ( S. nigrocinctus ), a territorial bottom dweller. Combining Hallacher’s work with this study, as well as the results of the unpublished acoustical studies, it would seem likel y that most, if not all rockfish species are at least capable of producing pop sounds, though they may only rarely do so. At the Seattle Aquarium, repeated and varied attempts to record or illicit sounds from a number of S. nigrocinctus individuals were unsuccessful. Speciation There has been considerable interest in evolution and speciation within Sebastes and sound production may be a new way to ex amine relationships. Are members of the genus diverging because they can produce and discern different sounds? Internal fertilization allows for cour tship rituals that may incl ude specific sounds or sound patterns. Parmentier et al. (2005) recently described re gional differences in sound production characteristics of anemonefish ( Amphiprion akallopisos ). Based on the results of this study, sound pr oduction is unlikely to be contributing to rockfish speciation. This is primarily be cause all the sound production appeared to be associated with agonistic rather than cour tship behavior. This is in keeping with Hallacher’s (1974) findings that both male and female rockfish have sonic muscles – species such as weakfish ( Cynoscion regalis ) that have been s hown to produce sounds related to courtship and spawning, show sexual dimorphism in their sound producing

PAGE 46

40 musculature (Connaughton et al. 2002). In othe r Sciaenidae (drumfish and croakers), a family well known for sound production and mating choruses, sonic muscles are generally present in mature males only (Tavolga 1964). In addition to the lack of sexual dimor phism in musculature and our failure to observe sound production directly related to courtship or mating, the overlapping nature of the calls we observed also indicates sound does not play a role in rockfish speciation. As it was not possible to discern the species of a call through our analysis, it is unlikely (though still possible) that i ndividual fish can do so. The calls may overlap simply because of the close phylogenetic relations hips and similar anatomy of the species examined. Closely related rockfish species wi th overlapping ranges ap pear to be courting each other with visual cues, not sonic ones, though considerably more research needs to be done to verify this. The paucity of obser vations of rockfish c ourting and mating, with or without hydrophones present, makes this question difficult to resolve. Growls and Rumbles The only major difference between calls th at showed up in the analysis was the fact that S. carnatus and S. chrysomelas two very closely related species, appeared to be the only ones to regularly make long duration rumbles as well as shorter growls. Three possible reasons for this are: 1. all six Sebastes species studied can rumble but the two observed do it much more often 2. all six Sebastes species studied can rumble and the observed difference was related to external variable s such as season, environment or crowding. This is supported by evidence from video clip s that appear to show S. atrovirens rumbling in their territorial phase. 3. anatomical or physiological differences prevent some species from rumbling. Fish sonic muscles are capable of cont racting and recoveri ng at unusually quick rates (Tavolga 1964). It is po ssible that due to differences in anatomy and/or muscle physiology, only certain species are capable of sustaining a longer call. In this study, pulse rate and number of pulses showed a si gnificant but slight ne gative correlation (r =

PAGE 47

41 -0.11, n=365). This can be seen in Figure 17, which shows a minor negative trend towards slower pulse rates as calls get longer. S. atrovirens appeared to rumble in captivity, but not as often as S. carnatus or S. chrysomelas S. nebulosus didn’t seem to rumble, either in captivity or in situ. Thus, the producti on of rumbling sounds coincides with molecular evidence of how closely related these four sp ecies are. Further acoustical study on a wider range of rockfish species woul d help resolve this interesting question.

PAGE 48

42 050100150200250300350 Number of Pulses 0 20 40 60 80 100 120 140 160 180 200 Rate (pulses/s) Number of Pulses:Rate (pulses/s): r = -0.1114, p = 0.0333 Figure 17 Pulse rate plotted against number of pulses for all Sebastes recorded (n=365). There was a slight negative correlation between the two (r=-0.11); note that none of the calls with more than 80 pulses had rates higher than approximately 50 beats per second.

PAGE 49

43 Reasons for Calling The calculated root mean square amplitude of the calls confirms anecdotal evidence – though at least some rockfishes do call, they do not make loud calls. Given typical hearing thresholds in fi shes, it must be concluded that rockfishes do not use sound for long distance communication (Fay 1988). This is supported by observational data – any behavior associated with calling was a gonistic and took place between fish less than four meters apart. Based on a number of factors observed in the lab (loudness of sounds versus distance to hydrophone; single fish with mi rror; tests with di fferent species and harassment with a net) Yearsley (1970) conc luded that only dominant fish produced “pops”. This study supports the dominant pop id ea – when observed, pops appear to be issued as a challenge. Could the pops be an indicator of how big the fish is? Myrb erg et al. (1993) reported a direct relationship between the length of damselfish and the peak frequency of their calls. If rockfishes could produce and de tect aggressive sounds that indicated their swimbladder size, it might be useful in re ducing the energy expenditure of chases and potential harm of dir ect physical challenges – S. nebulosus can be very aggressive towards conspecific intruders. Pops may occur when an individual fish feels another is appr oaching too closely. As the intruder gets closer, the popping sounds of the aggravated fish become faster and perhaps louder, until the growl of a chase fo llows. This strategy would let an intruder know if it was getting too close. Conversely, an approaching fish may pop to intimidate a territory holder. Growls are more difficult to explain. Yearsley (1970) be lieved that some species never produced sound when submissive but ot hers did, and two species produced growls whether dominant or submissive but usua lly associated with sudden locomotory movement. Because of the dynamic nature of chases and their accompanying growls, it is difficult to ascertain which fish is making the sound, or even if it is just one. As the sounds made by S. nebulosus individuals harassed by a dive r were indistinguishable from those made during agonistic behavior, it may often be the submissive fish growling, not

PAGE 50

44 the aggressor. This is supported by at least tw o observations when growls were recorded from S. nebulosus individuals that were chas ed by fish (kelp greenling, Hexagrammos decagrammus ) other than rockfishes. Temporal Data Though other studies (see Love et al. 2002, p. 55-56) have indicated that most species of rockfishes are primar ily diurnal, there appear to be calls on the field datalogger throughout the night (Figures 14 and 16). While sounds recorded on the datalogger cannot be positively identified to species, th e location of the deployment, on the den site of at least one S. nebulosus the most vocal Northwest species, makes it likely that the vast majority of the calls recorded were ma de by that species, perhaps even by one or two individuals. Diel periodicity in mating choruses has been shown for sand sea trout ( Cynoscion arenarius ), a broadcast spawner, using datal oggers (Locascio and Mann 2005). In this study no choruses were detecte d, but there were more calls per hour after midnight and around 0500 and 1500. There were also lull s around 0800 and 1800 – these patterns might be coincidence, but could also represent changes in activity levels for the resident fish at the den site. The datalogger would appear to indicate that S. nebulosus is active throughout the night and perhaps more sedentar y in the morning, though it could also be that the individual left th e area during the morning. More deployments and visual observations would help resolve this issue and provide important information about the ecology of this species. Further Research One of the main purposes of this st udy was to direct further research. Considerable effort was involve d in gathering the results of unpublished work, all at least sixteen years old but relevant and helpful. Some of the research methods attempted during this study (eg. hydrophone trawls) proved to be unproductive, gi ven the nature of rockfish sound production. It would appear th at one of the main reasons why rockfish sound production is so poorly known is that most species of nearshore rockfishes don’t make sounds very often, and few (if any) make sounds that are relatively loud.

PAGE 51

45 Rockfishes apparently use sound for a gonistic purposes, but only up close – unlike other territorial sonifer ous fish, they do not appear to “call out” their territories. However, due to the lack of mating data from e ither wild or captive fish, it is difficult to rule out reproductively related sounds. Ce rtainly, the Moss Landing captive studies showed at least one species ( S. atrovirens) changes its behavior dramatically during mating season, resulting in more territorial a nd agonistic behavior, with an associated increase in agonist ic sound production. What should the next studies focus on ? Passive acoustical techniques will be unlikely to reveal mating activity as they do with other species, though they may help identify areas where rockfishes are and coul d possibly be helpful with monitoring size and/or abundance. Further study will benefit from more extensive use of in situ dataloggers, as the acoustic conditions in a quaria or laboratories ar e rarely conducive to studying the relatively low amplitude sounds that most rockfishes produce. It is also likely that rockfish species that live beyond safe diving depths ar e sound producers, and placement of hydrophones on deep r eefs will help address this. Perhaps the most interesting line of re search would be to see if there is a correlation between the pop sounds that rockfish es make and the size of the individual, and whether rockfishes are capable of detec ting it. If so, acoustical dataloggers may be helpful in passively monitoring size changes over time, something especially helpful in studies involving marine reserves. Because of the logistics of catching and acclimating rockfishes to tanks, which can take weeks or even months, captive sound work would be best suited to a longer term project. Very few aquaria, even those with large tanks, have had much success breeding rockfishes, so mating studies would be diffi cult. However, using captive studies to examine relationships between sound producti on, species, size, gender, and season would be most interesting. As acoustical background noise in many existing aquaria is difficult to work with, it is worth outlining an ideal setup. A relatively large tank, lit by ambient light, containing den sites, an array of fixed hydrophones, with viewing possibilities for video, dampened or isolated mechanical noi se, and sound absorbing walls would be an excellent start.

PAGE 52

46 REFERENCES Dotu, Y (1951) On the sound producing mechanis ms of a scorpaenid fish Sebasticus marmoratus. Kyushu Imp. Univ. Dept. Agri. Bull. Sci. 13: 286-288 Fay, R. R. and Megela Simmons, A. 1999. The sense of hearing in fish and amphibians. In: Comparative Hearing: Fish and Amphibians eds. R. R. Fay and A. N. Popper, pp. 269-318.New York: Springer-Verlag. Fay, RR (1988) Hearing in Vert ebrates: A Psychophysics Da tabook. Hill-Fay Associates, Winnetka, IL. Fay, RR and A Megela Simmons (1999) The sens e of hearing in fishes and amphibians. In Comparative Hearing: Fish and Am phibians (ed. RR Fay and AN Popper), pp. 269-318. New York: Springer-Verlag. Fletcher, VE (1983) A comparative analys is of the behaviou ral ecology, agonistic behaviour and sound production in two species of inshore eastern Pacific rockfish (Genus Sebastes ). Ottawa : National Library of Canada. Gharrett AS, AK Gray, and J Heifetz (2001) Identification of rockfish ( Sebastes spp. ) by restriction site analyses of the m itochondrial ND-3/ND-4 and 12S/16S r RNA gene region. Fish. Bull. 99:49-62 Hallacher, LE (1974) The comparative mor phology of extrinsic gasbladder musculature in the scorpion fish genus Sebastes (Pisces: Scorpaenidae). Proceedings of the California Academy of Sciences (four th series) Vol XL, No. 3: 59-86. Higgs, DM (2005) Auditory cues as ecological signals for marine fishes. Mar Ecol Prog Ser. Vol. 287: 278–281 Kendall AW (2000) An historical review of Sebastes taxonomy and systematics. Mar. Fish.Rev. 62(2):1–24 Ladich, F (2004) Sound production and acous tic communication. In: von der Emde G, Mogdans J, Kapoor BG (eds) The senses of fish: adaptations for the reception of natural stimuli. Narosa Publishing, New Delhi, p 210–230 Locasio, J and D Mann (2005) Effects of Hurrica ne Charley on fish chorusing. Biology Letters: 1744-9561.

PAGE 53

47 Love, MS, M Yoklavich and L Thorsteinson (2002) The Rockfishes of the Northeast Pacific. University of California Press Luczkovich, JJ, MW Sprague, SE Johnson, and RC Pullinger (1999) Delimiting spawning areas of weakfish Cynoscion re galis (family sciaenidae) in Palmico Sound, North Carolina using passive hydro acoustic surveys. Bioacoustics 10:143160.Matsubara, K (1943) Studies on the sc orpaenoid fishes of Japan. Trans. Sigenkagaku Kenkyusyo, No. 1, pp. 126-147 Myrberg, AA Jr, SJ Ha and MJ Shamblott ( 1993) The sounds of bicolor damselfish ( Pomacentrus partitus ), predictors of body size and a spectral basis for individual recognition and assessment. J. Acout. Soc. Am., 94 3067-3070. Parmentier, EJP Lagardere, P Vandewalle and ML Fine (2005) Geographical variation in sound production in the anemonefish Amphiprion akallopisos Proc. R. Soc. B doi:10.1098/rspb.2005.3146 Popper, AN, J Fewtrell, ME Smith and RD McCauley (2003) Anthropogenic Sound: Effects on the Behavior and Physiology of Fishes. Marine Technology Society Journal Vol. 37 No. 4: 35-40. Popper, AN, ME Smith, PA Cott, BW Ha nna, AO MacGillivray, ME Austin, DA Mann (2005) Effects of exposure to seismic air gun use on hearing of three fish species. J. Acoust. Soc. Am. 117 (6), June 2005, 3958–3971. Rocha-Olivares A, Rosenblatt RH, Vetter RD (1999) Molecular evolution, systematics, and zoogeography of the rockfish subgenus Sebastomus (Sebastes, Scorpaenidae) based on mitochondrial cytochrome b and control region sequences. Mol Phylogenet Evol 11:441–458 Schmitt, EF, DF Wells, and KM Sullivan-Seal ey (1998) Surveying coral reef fishes: a manual for data collection, processing, and interpre tation of fish survey information for the tropical northwest Atlantic. Coral Gables, FL: The Nature Conservancy, Marine C onservation Center. 84pp. Smith, ME, Kane, AS and Popper, AN (2004) No ise-induced stress response and hearing loss in goldfish (carassius auratus). J Exp Biol. 207:427-435. Spanier, E (1979) Aspects of species recognition by sound in four species of damselfishes, genus Eupomacentrus (Pisces: Pomacentridae),’’ Z. Tierpsychol. 51, 301–316. Tavolga, WN (1964) Sonic charact eristics and mechanisms in marine fishes. In Marine Bio-Acoustics, vol. 1 (ed. W. N. Tavolga), pp. 195–211. New York: Pergamon Press.

PAGE 54

48 Van Dykhuizen, GD (1983) Activity patterns and feeding chronology of the kelp rockfish, Sebastes atrovirens, in a centra l California kelp forest. Masters thesis, San Jose State University. Wold, L (1991) A practical approach to the description and identif ication of SEBASTES larvae. Masters thesis. Calif. State University, Hayward. Yearsley, JH (1970) Drumming muscle mor phology and sound production in rockfishes (Sebastodes). University of Victoria. Received as a personal communication with John McInerney, supervising professor.