ARE ROCKY INTERTIDAL POOLS A NURSERY HABITAT FOR JUVENILE REEF FISH? AN INVESTIGATION OF THE SPATIAL AND TEMPORAL ABUNDANCE

PATTERNS OF JUVENILE FISHES UTILIZING BASALT TIDEPOOLS ON THE ISLAND OF OAHU AND A COMPARATIVE GROWTH ANALYSIS OF THE

ENDEMIC KUHLIA XENURA

A THESIS SUBMITTED TO THE GRADUATE DIVISION OF THE UNIVERSITY OF HAWAI‘I AT MĀNOA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF

MASTER OF SCIENCE

IN

ZOOLOGY (MARINE BIOLOGY)

AUGUST 2012

By

Ilysa S. Iglesias

Thesis Committee:

Les Watling, Chairperson Alan Friedlander

Craig Smith

! ii!

ACKNOWLEDGEMENTS

First and foremost I wish to thank Hawai‘i for the opportunity to work along its

beautiful shores, as well as the fish involved in this study that gave the ultimate sacrifice

to my project. Next, I would like to give a huge thanks to my knowledgeable committee,

Les Watling (chair), Alan Friedlander and Craig Smith, whose holistic ecological input

proved invaluable. Specifically, I would like to thank Les for his perpetual open-door

policy and love of the intertidal, Craig for his breadth of knowledge and instruction in

ecology in all sorts of systems, and Alan for providing me with the best RAship a

graduate student could hope for, making me feel welcomed as part of his lab and

providing equipment and advice so as to complete this project! Additionally, I would

like to thank John Stimson for opening my eyes to the study of coral reef community

ecology. I would also like to thank my many friends that were gracious enough to help

out during the field portion of my project: Nyssa, Maya, Mary, Joey, Paolo, Kaipo, Sas

and Chris, your incredible enthusiasm and help is incredibly appreciated!! This project

benefited greatly from the expert advice of many thoughtful people: Dr. Ed DeMartini of

NOAA NMFS (all things recruitment in Hawaii), Dr. Robert Humphreys NOAA NMFS

(otolith processing and interpretation), Dr. Andy Taylor (help with statistical analysis),

Dr. Chris Bird (intertidal knowhow and getting me involved in the great intertidal

projects in the Hawaiian archipelago-all the way to Papahanaumokuakea), Dr. Greta

Aeby (for getting me started in Hawaii), and Erin Cox (for helpful input on tidepool

fishes). Further, I would like to thank our department chair Chris Womersley and the

women in the zoology office (Jan, Lynne and Audrey) for all of their support and help

over the last three years. I would also like to thank the division of aquatic resources

(DAR) for providing the necessary special activity permits to complete this work, as well

as IACUC. On a personal note, I would like to thank Chris Alreck for standing by me

throughout this whole process as well as my absolutely wonderful mom and dad, without

whose unconditional love and support for the last 27 years, I never would have believed

in following my passions (FSF!). Finally, I would like to thank my amazing cohort of

graduate students at the University of Hawaii, Manoa department of Zoology for all of

the support and inspiration over the years!!

! iii!

ABSTRACT

The abundance and community composition of post-settlement juvenile fishes

occupying basalt tidepools on Oahu was examined to determine the importance of this

habitat as a nursery area. Spatial and temporal patterns among pools and between sites

indicate significant differences in juvenile fish assemblage between sites (Makapu’u and

Ma‘ili), though this pattern was not consistent across months. The relative nursery

quality of tidepools was compared using otolith inferred growth rates of juvenile Kuhlia

xenura collected from Makapu’u tidepools and an adjacent stream. Results indicate

faster growth for juveniles (35-40mm total length) occupying tidepools, but this pattern

was not evident for other size-classes sampled. Periodicity of spawning and recruitment

for K. xenura were inferred from otoliths, and a pelagic larval duration of 32 days

calculated. Results indicate tidepools are an important nursery area for numerous

juvenile fish species on Oahu, especially K. xenura, and should be a priority for research

and management.

! iv!

TABLE OF CONTENTS

Acknowledgements……………………………………………………………..ii

Abstract…………………………………………………………………………iii

List of Tables……………………………………………………………………vi

List of Figures…………………………………………………………………..vii

Chapter 1: Introduction……………………………………………………..…..1

The importance of nursery areas in tropical ecosystems……………......1 The habitat associations of post-settlement fishes in Hawaii…………...2 Tidepools as a nursery habitat in other regions of the world……………4

Chapter 2: Are tidepools a nursery habitat in Hawaii?…………………………7 Abstract…….……………………………………………………………7 Introduction……………………………………………………………...7 Methods………………………………………………………………...10 Tidepool surveys……………………………………………….10 Otolith analysis…………………………………………………11 Statistical analysis……………………………………………...13 Results………………………………………………………………….15 Tidepool juvenile fish assemblages……………………………15 Otolith analysis………………………………………………...22 Discussion……………………………………………………………...28

Tidepool juvenile fish assemblage …………………………….28 Otolith microstructure analysis………………………………...31 Conclusions …………………………………………………....33

Appendices……………………………………………………………..34

Appendix A ……………………………………………………34 Appendix B …………………………………………………...36 Appendix C…………………………………………………….39

Appendix D ……………………………………………………42 Appendix E …………………………………………………...42 Appendix F ……………………………………………………44 Appendix G …………………………………………………...45 Appendix H ……………………………………………………46 Appendix I …………………………………………………….50

References……………………………………………………………..51

! v!

LIST OF TABLES

Table Page

1. ANOSIM test for statistical differences in juvenile fish assemblage among

sites: All months averaged……………………………………………………17 2. ANOSIM pairwise tests for all months averaged………………………………17

3. ANOSIM test for statistical differences in juvenile fish assemblage among

sites: June………….………………………………………………………….18 4. ANOSIM pairwise tests for June……………………………………………….18

5. ANOSIM test for statistical differences in juvenile fish assemblage among

sites: July……………………………………………………………………...19 6. ANOSIM pairwise tests for July…………….………………………………….20

7. SIMPER output: species responsible for explaining between site

differences at Makapu‘u and Ma‘ili for June………………………………….21 8. SIMPER output of species responsible for explaining between site

differences at Makapu‘u and Ma‘ili for July………………………………….21

! vi!

LIST OF FIGURES Figure Page 1. Map of study sites for tidepool surveys..………………………………………10 2. Map of collection sites for Kuhlia xenura……………………………………..12 3. Settlement band on a sagittal otolith of Kuhlia xenura………………………..13 4. nMDS plot of tidepool fish assemblage, all months averaged..……………….17 5. nMDS plot of similarity between Makapu‘u and Ma‘ili tidepool fish

assemblage, all months averaged ……………………………………………...18 6. nMDS plot of tidepool fish assemblage, June…..……………………………..19 7. nMDS plot of similarity between Makapu‘u and Ma‘ili tidepool fish

assemblage, June……………………..……………………………………….19 8. nMDS plot of tidepool fish assemblage, July…..………………………..…….20 9. nMDS plot of similarity between Makapu‘u and Ma‘ili tidepool fish

assemblage, July…………………….………………… …………………….. 20 10. The average abundance of Acanthurus triostegus by site…...…………………22 11. Linear regression of otolith length to fish total length…..…………………….23 12. Age and growth of Kuhlia xenura ……………………………………………23 13. Age and growth data for Kuhlia xenura by collection location……………….25 14. Kuhlia xenura average age verse location for size classes

35-40 mm and 40-45mm……………………………..……………………….25 15. Pelagic larval duration of Kuhlia xenura ……………………………………..26 16. Back calculated hatch dates for Kuhlia xenura……………………………….27 17. Back calculated settlement dates for Kuhlia xenura………………………….27 18. Back calculated settlement dates for Kuhlia xenura by site……………..……28 19. A map of study sites with respect to wind-wave exposure……………………30

! 1!

CHAPTER 1

INTRODUCTION

Coastal ecosystems connect human populations to the marine environment, and as

such are subject to some of the most profound negative impacts of human practices

(Jackson et al. 2001). In tropical regions, coastal ecosystems often provide essential

habitat for juvenile fishes in the form of nursery areas, typically seagrass beds and

mangrove forests (see Beck et al. 2001 for review). In the Hawaiian Islands, the local

seagrass, with its slight leaves and patchy distribution, does not generate the structurally

complex beds typical of other tropical regions necessary for refuge by juvenile fishes.

Further, the mangroves present in Hawaii are the result of an intentional introduction of

non-native species and thus have not been historically available as nursery areas. The

paucity of classical tropical nursery areas begs the question: which habitats do juveniles

of nearshore fishes recruit to before migrating to their eventual, distinct, adult habitat?

This study investigates the post-settlement patterns of juvenile reef fish to a potentially

overlooked nursery habitat, tidepools, with a two-fold ecological approach. First, surveys

of the community composition of juvenile fishes utilizing rocky-intertidal pools are

examined to explore spatial and temporal variability. Second, using otolith microstruture,

a comparative growth analysis of a common tidepool inhabitant, Kuhlia xenura, is

evaluated to compare the relative quality of tidepools to native stream habitat in

conferring conditions ideal for juvenile growth, an important factor in recruitment

success. Further, otoliths are examined to determine potential timing of hatching and

settlement for this species to tidepools. As a result of the inherent difficulties associated

with characterizing recruitment patterns of nearshore fishes, there is limited information,

even for important resource species, concerning the early habitat requirements of juvenile

fishes in Hawaii. In order to effectively manage important nearshore resource species, the

designation of habitats critical to the successful recruitment of fishes must be elucidated

so as to better conserve populations for the future.

The importance of nursery habitats in tropical ecosystems

The nursery role hypothesis states “the ecological processes operating in nursery

habitats, as compared with other habitats, must support greater contributions to adult

! 2!

recruitment from any combination of four factors: density, growth, survival of juveniles

and movement to adult habitats” (Beck et al. 2001). The classic nursery areas in the

Caribbean, seagrass meadows and mangrove habitats, provide abundant food resources as

well as structural complexity for juveniles seeking refuge, factors contributing to the

survivorship of juveniles to adulthood. Thus, the availability of nursery habitats adjacent

to adult areas has long been assumed to affect the recruitment success of nearshore fish

species. Research by Nagelkerken et al. (2002) in the Caribbean compared the density of

17 species of coral reef fish on island systems replete with mangrove and seagrass

habitats to islands lacking these habitat types and found that the density of 11 of the

species examined was significantly decreased on islands without seagrass and mangrove

nursery areas. Mumby et al. (2004) tested the importance of mangrove forests in

contributing to the abundance of coral reef fish on atolls with and without mangroves and

found a doubling of fish biomass on reefs located adjacent to mangroves. Further, the

authors observed a functional dependence of the parrotfish Scarus guacamaia on

mangrove areas, such that in areas lacking mangroves this species was locally extinct.

The importance of nursery habitats to the functioning of ecosystems and in turn the

successful recruitment of juveniles to adult populations cannot be understated, yet has

received little attention in Hawaii.

The habitat associations of post-settlement fishes in Hawaii

In the Hawaiian archipelago, juvenile fish recruitment is an important factor

contributing to adult fish densities (Walsh 1984) (although subject to inter-annual

variability), and is thus pertinent to successful management of nearshore resource

species. Although Hawaii lacks the putative nursery habitats so common in other tropical

systems, research exists on the association of recruiting fish to other specific habitat

types. The utilization of inshore, estuarine and soft bottom habitats by the young of

coastal resource species has been demonstrated for moi, Polydactyls sexfilis, (Friedlander

& Ziemann 2003, Leber et al. 1998), ulua, Caranx melampygus and Caranx ignobilis,

(Smith & Parish 2002) and ama’ama, Mugil cephalus (Major 1978, Leber et al. 1996) and

appears to be related in some instances to areas of freshwater input. Ten years of beach

! 3!

seine survey data on Oahu has also found the above-mentioned species utilizing sandy

surf zones (Friedlander, personal-communication).

In addition to soft-bottomed areas, research has shown that juvenile reef fish are

often observed in association with semi-sheltered backreef habitats and coral patch-reefs,

often in protected lagoon systems (DeMartini et al. 2009, DeMartini et al. 2010,

DeMartini 2004 and Stimson 2005). Early life stages are often subject to intense

predation pressure and the ability of recruits to seek shelter in rugose coral substrate can

improve survivorship. Young of two species of Acanthuridae, the yellow tang

Zebrasoma flavescens and the kole Ctenochaetus strigosus, on the Big Island of Hawaii,

were found in association with coral-rich areas, especially with a specific coral species,

Porites compressa (DeMartini and Anderson 2007). The authors state the association

with one of the most rugose species of coral is a strategy to avoid predation. Ortiz and

Tisot (2008) further demonstrated an association of juvenile Zebrasoma flavescens to

Porities compressa, distinct from adult habitat utilization patterns within a coral reef on

the Big Island. Beyond these few surveys, little information exists linking juvenile reef

fishes to specific habitat types (see Stimson (2005) for reference to an episodic

recruitment event of Pervagor spilosoma to patch reefs in Kaneohe Bay and Booth

(1992) also in Kaneohe bay, described the association of recruiting Dascyllus albisella to

specific coral colonies). Although the research described herein represents a significant

contribution to our understanding of the habitat characteristics required by juvenile reef

fishes, research has generally been limited to a small number of species occurring in

specific locations. Walsh (1987) found recruitment in Hawaii to be highly variable by

location and year and thus what researchers have found to be true for a particular area,

may not be applicable to all regions.

There is an obvious need for additional knowledge concerning the habitat

requirements of post-settlement fishes in Hawaii. It is the aim of this study to address the

question of whether intertidal rock pools are a potentially overlooked nursery habitat for

juvenile fishes on Oahu. To date, extensive research specifically targeting the

community of juvenile fishes utilizing the intertidal rock pool habitat as a nursery is

lacking. However, studies do exist describing juvenile fishes in tidepools in Hawaii,

dating back to Hawaiian traditional ecological knowledge. Titcomb (1972) described the

! 4!

belief that juvenile fish “went into sea pools to live until the tiny fish were grown.” The

manini, Acanthurus triostegus, was found to transform from its pelagic larval stage to a

juvenile stage while within tidepools (Randall 1961, Sale 1969). The association of A.

triostegus to tidepools was also recorded in a study examining nearshore fish distribution

in Hawaii by Gosline (1965), in which it was thought A. triostegus and āholehole, Kuhlia

sandvicensis, were using tidepools “as an incubator for the young, which move back out

to sea as they grow up.” Further research on K. sandvicensis confirmed the exclusive

utilization of tidepools by this species as juveniles, while a similar species, Kuhlia xenura

could be found in tidepools, sandy beach habitats and streams (Benson & Fitzsimmons

2002, McRae et al. 2011). Beyond the information outlined above on A. triostegus, K.

xenura, and K. sanvicensis juveniles, three Hawaiian fish identification guides mention

the presence of juveniles of these species in tidepools (Gosline & Brock 1960, Hoover

2008, Randall 2010). To date, only one study has explored the assemblage of tidepool

fish on Oahu. Researchers examined the abundance of fish present in intertidal pools on

Oahu (both juvenile fish and adults of resident species), and determined the

environmental conditions responsible for the greatest amount of variability in the

distribution and abundance of fishes were related to differences in temperature, type of

substrate (basalt or limestone), as well as zonation within the intertidal bench (Cox et al.

2011). Although this study is important in describing environmental factors responsible

for the distribution of all tidepool fishes, sampling was not conducted at the same time

throughout the summer months, and pools were only sampled once.

Tidepools as nursery habitat in other regions of the world:

Intertidal habitats have long been the subject of scientific research in the Pacific

North West, historically providing scientists with a natural laboratory to develop some of

the most fundamental underpinnings of ecology. Although rocky intertidal habitats have

received ample research attention, comparatively little has been focused on tidepools

specifically (review by Metaxas & Scheibling 1993). Of the limited research exploring

tidepools, even less is directed at understanding the potential importance of this habitat as

a nursery for sub-tidal fish. The first mention of the utilization of tidepool habitat by

juveniles occurred in Southern California (Williams 1957, Norris 1963). Based on the

! 5!

temporal peaks in abundance of juvenile fishes during the summer months in rocky

intertidal pools of Northern California, it was concluded that “rocky intertidal pools have

seasonal importance as nursery areas for juveniles of offshore species,” particularly for

black rockfish (Moring 1986). Studebaker (2006) and Studebaker et al. (2009) found that

12 species of rockfish along the West coast (from Oregon to California) utilized tidepools

during juvenile stages, most likely related to the high food availability (Studebaker &

Mulligan 2008). A study into the specific movement patterns of young-of-the-year black

rockfish in intertidal areas in Northern California found that transplanted juveniles

exhibited both site fidelity and homing behavior to specific tidepools (Lomeli 2009),

lending greater support for the active selection of this habitat over random dispersal.

In the South African intertidal zone, researchers discovered high abundances of

juvenile fish species in coastal tidepool areas (Beckley 1985), even suggesting the utility

of these areas as an important nursery area (Smale & Buxton 1989) (although Bennett

(1986) suggested this association might not be critical for juvenile fish). High

abundances of juveniles were also observed in Australia, where post-settlement fish were

numerically dominant in intertidal populations, in some cases even settling directly from

the pelagic larval stage into pools (Lardner et al. 1993). However, Griffiths (2003) found

that juveniles only composed 1% of fish assemblages sampled, although among the 1%

was a vulnerable species, the black rock cod, perhaps further justifying the importance of

protecting this habitat. The variable importance of tidepools to juveniles of sub-tidally

associated adult fishes highlights the importance of location and study species as

considerations when exploring tidepool fish assemblages (Gibson & Yoshiyama 1999).

Even less information exists on tidepools in tropical regions of the world. In

Cabo Branco Beach, Brazil, monthly samples of tidepool icthyofauna were consistently

composed of a large number of juvenile fish (Rosa et al. 1997) and so were tidepools in

tropical Colombia (Castellanos-Galindo et al. 2005). The only other mention of juvenile

fishes occupying tropical tidepools comes from the Caribbean island of Barbados, where

36% of tidepool fishes surveyed were juveniles (Mahon & Mahon 1994). In conclusion,

although region specific and perhaps depending on survey methodology, researchers have

established that juvenile fish can be abundant in tidepools in distinct intertidal regions of

the world, and tidepools may function as a nursery area for sub-tidal species. Also

! 6!

highlighted in the research of tidepools in other regions is the prevalence of species of

commercial and recreational importance in these areas. Understanding the potential role

of tidepools as a nursery area in the early life stages of resource species is important in

elucidating which areas should be designated for conservation and management plans.

In the Main Hawaiian Islands, many coastal fishes adjacent to large population

centers are being extirpated at rates that may not be sustainable (Smith 1993, Friedlander

2004). The biomass of many regulated species being extracted from Hawaiian coastal

environments is currently unclear, as fishers are not required to report catches. However,

the distinction as regulated belies the continued need for a better understanding of early

life histories and habitat utilization patterns. This research explores the abundances of 15

species of juvenile fishes inhabiting the rocky-intertidal habitat of Oahu. Four of these

species are important for recreational fishing and consumption (manini-Acanthurus

triostegus, āholehole-Kuhlia xenura and Kuhlia sandvicensis and uouoa-Neomyxus

leuciscus). Each of these species have adult stages of their life cycle in sub-tidal reef

associated habitats and the first three are listed in the “Hawaii Fishing Regulations” due

to concerns over their populations status (http://hawaii.gov/dlnr/dar). In order to

successfully regulate nearshore fish species, the early habitat requirements of targeted

species must be elucidated such that we may better understand the factors contributing to

the successful perpetuation of healthy adult stocks.

! 7!

CHAPTER 2

ABSTRACT

The abundance and community composition of post-settlement juvenile fishes

occupying basalt tidepools on Oahu was examined to determine the importance of this

habitat as a nursery area. Spatial and temporal patterns among pools and between sites

indicate significant differences in juvenile fish assemblage between sites (Makapu‘u and

Ma‘ili), though this pattern was not consistent across months. The relative nursery

quality of tidepools was compared using otolith inferred growth rates of juvenile Kuhlia

xenura collected from Makapu‘u tidepools and an adjacent stream. Results indicate

faster growth for juveniles (35-40mm total length) occupying tidepools, but this pattern

was not evident for other size-classes sampled. Periodicity of spawning and recruitment

for K. xenura were inferred from otoliths, and a pelagic larval duration of 32 days

calculated. Results indicate tidepools are an important nursery area for numerous

juvenile fish species on Oahu, especially K. xenura, and should be a priority for research

and management.

INTRODUCTION

In Hawaii there is a scarcity of information, even for important resource species,

concerning the early habitat requirements of nearshore fish species. Although Hawaii

lacks the putative nursery habitats so common in other tropical systems (i.e., structurally

complex seagrass meadows and native mangrove populations), research exists on the

association of post-settlement fishes to other specific habitat types, specifically inshore

estuarine and soft bottom habitats by the young of moi, Polydactyls sexfilis, (Friedlander

& Ziemann 2003, Leber et al. 1998), ulua, Caranx melampygus and Caranx ignobilis,

(Smith & Parish 2002) and ‘ama‘ama, Mugil cephalus (Major 1978, Leber et al. 1996).

In addition to soft bottom areas, numerous surveys have revealed an association of young

coral reef fish to semi-sheltered backreef habitats and coral patch-reefs (DeMartini et al.

2009, DeMartini 2004 and Stimson 2005) usually in association with specific species of

! 8!

rugose coral (DeMartini and Anderson 2007 and Ortiz and Tisot 2008). The limited

research aimed at determining essential fish habitat leaves managers with an absence of

information concerning which areas should be a priority for conservation. Thus, any

knowledge of the abundance and distribution of juvenile fishes to rocky-intertidal pools

represents an important contribution to our understanding of the factors affecting

successful replenishment resource species in Hawaii.

Previous work in the Hawaiian intertidal habitat has described the presence of

juvenile reef fish in tidepools (Titcomb 1972, Randall 1961, Sale 1969, Gosline 1965,

Benson & Fitzsimmons 2002, Cox et al. 2011, McRae et al. 2011), but none of these

studies exclusively surveyed juvenile fish assemblages or tested the function of this area

as an important nursery site. Cox et al. (2011) conducted a comprehensive survey of

tidepool fishes, both juveniles and resident species from six sites on Oahu, in an effort to

determine which environmental drivers were most responsible for the observed spatial

heterogeneity in the tidepool fish assemblage and concluded temperature, type of bench

and location within bench explained the greatest amount of variability. Further, the

authors did not sample at a discrete time interval, rather throughout the summer months,

and only sampled tidepools on one occasion. Previous research in the coral reef habitat

by Walsh (1987) revealed that recruitment in Hawaii is extremely low compared to other

tropical ecosystems and highly site specific and temporally variable, thus it is important

to determine if patterns of recruitment to the intertidal habitat are consistent temporally or

whether the time in which one samples recruitment could determine the output of

observed assemblage patterns.

The goal of this study is to explore the spatial and temporal variability in the

assemblage of juvenile fishes at four basaltic intertidal benches on Oahu. Visual surveys

are repeated each month during a recruitment season (April-August), in order to

determine how assemblage structure changes with respect to space and time. Further,

individual species abundances are evaluated by site to explore potential widespread

recruitment patterns.

Spatial and temporal variability in populations of juvenile fishes in tidepools can

address questions concerning which species are most common and whether patterns exist

in their distribution, but does not reveal information about the quality of tidepools as a

! 9!

nursery, compared to other potential nursery areas, or the specific timing of species

recruitment to this area. In order to explore questions about the timing of settlement, age

of individual fish found in tidepools and the relative habitat quality of tidepools

compared to another potential nursery area, we specifically explored the early life history

of a common tidepool inhabitant, Kuhlia xenura.

The endemic Hawaiian āholehole, Kuhlia xenura is an important nearshore

resource species, yet little is known about the relative quality of juvenile nursery habitats

utilized post-settlement. By comparing habitat specific growth rates of K. xenura

collected from Makapu‘u tidepools and an adjacent stream habitat, we can derive a

relative measure of otolith inferred growth rate. Survival of early life stage fishes to adult

populations is influenced by the ability of recruiting fish to persist in the face of stark

predation pressure, food scarcity, and often-unfavorable abiotic conditions (see review by

Richards & Lindeman 1987). The ability of juvenile fish to attain a size refuge from

predators allows for an attenuation of mortality pressure (Houde 2002) and it has been

hypothesized that the faster a size-refuge is reached, the greater the chance of survival to

adulthood (Houde 2008). Otolith inferred growth rate comparisons have been calculated

to determine habitat-specific growth rate, and thus relative importance of nursery areas

by numerous authors (Haynes et al. 2011, Gilliers et al. 2006, Amara et al. 2007, Amara

et al. 2009, Vinagre et al. 2008, Gillers et al. 2004, Mateo et al. 2011, Ross 2003, Stocks

et al. 2011, Plaza et al. 2010, Stunz et al. 2002 Rooker et al. 2004 and Mateo et al. 2011),

but never involved the Hawaiian intertidal habitat or our study species Kuhlia xenura. It

is a goal of this study to explore the habitat quality of two potential nursery habitats for

Kuhlia xenura: nearshore stream habitats and tidepools.

In addition to habitat-specific growth rates, otolith interpretation can also reveal

important life history information, namely hatch date, settlement date and pelagic larval

duration for individual K. xenura collected. The timing of spawning events, age of

juveniles occupying tidepools and settlement dates between sites is pertinent information

in determining the utilization of these habitats by early stage K. xenura and can

ultimately benefit the successful management of this important recreational, cultural and

resource species.

! 10!

As the human population of Hawaii continues to swell, the potential impact on

tidepools and native stream environments increases. It is thus timely and pertinent to

examine critically the importance of native streams and tidepools to the early life stages

of numerous reef fishes, specifically K. xenura, so as to better protect habitats important

to the early life history of nearshore reef fish.

METHODS

Tidepools surveys

A series of visual surveys were conducted on the island of Oahu spanning a

typical Hawaiian recruitment season, April through August (see Walsh 1987 for more on

recruitment periodicity in Hawaii).

This study encompassed four locations of

intertidal basalt benches around the Island of

Oahu (Fig.1). These particular sites have

been selected on the criterion of being

permanently exposed structures (not covered

by sand during any time of the year),

similarity in physical composition (all basalt

benches), and being representative habitat.

Abundance of juvenile fishes in tidepools

In order to quantify the abundance of

juvenile fishes found in tidepools, while limiting potential confounding environmental

variables, basalt benches were exclusively used in this study with tidepools selected from

a similar zone at each site (defined as above the algae zone and below the pools

experiencing low water input). For each of the four basalt benches examined, each

tidepool of similar volume was marked using a handheld GPS device. Five tidepools

were selected per bench using a random number generator.

For the months April thru August (total of five months), each of the four locations

was visited at low tide (within an hour of a minus tide to minimize exposure to dangerous

wave conditions). For each pool a 10 minute, timed visual survey was conducted and

Fig.1 A map of Oahu depicting the four basalt benches surveyed in this study. For each site, 5 pools were randomly selected and surveyed on a monthly basis

! 11!

individual fish species, size estimation and abundance were recorded. In this study the

distinction of Abudefduf abdominalis from Abudefduf vagiensis was too difficult given

the small size of fish and is recorded as Abudefduf abdominalis based on prior references

of the recruitment behavior of A. abdominalis to tidepools (Cox et al. 2011 and Hoover

2008). It was also too difficult in some cases to distinguish Mugil cephalus from

Neomyxus leuscius, as these species are also very similar morphologically, especially

when small, and thus recorded as “mullet” when identification was not possible. The

distinction of juvenile for this study was determined by a combination of coloration and

size and may have included sub-adults, but no individuals were recorded greater than

10cm total length. This procedure was repeated for all five tidepools per location at all

four sites, for a total of 20 tidepools surveyed per month. In total, 100 ten minute timed

observations were conducted, or 1,000 total minutes of observation.

Otolith analysis

Study sites The comparative growth rate analysis was conducted at two locations

on the east side of Oahu. Benson (2002) and McRae et al. (2011) previously observed

Kuhlia xenura utilizing stream habitats as juveniles, and thus a stream was used as a

comparable juvenile habitat. Collections occurred on the East side of Oahu, at the

tidepools at Makapu‘u and a nearby stream in Waimanalo (connected at times to the

ocean), as this was one of the only locations for which these two habitat types were

located in close proximity, with N=1 for each type of habitat (Fig.2).

Collection of juveniles from Makapu‘u tidepools and Waimanalo Bay stream

Once a tidepool was selected for presence of Kuhlia xenura, a small dip-net was

used to extract individuals from the pool. A total of 37 juveniles were collected from the

Makapu‘u tidepools, ranging in size from 17.41mm standard length to 55mm. All

juveniles were stored frozen at the Hawaii Cooperative Fisheries research unit until

otolith extraction. The collection of Kuhlia xenura from the stream habitat of Waimanalo

involved the use of a beach seine net to corral fish at the lower reaches of the stream and

dip nets to extract them from this habitat. A total of 25 juvenile K. xenura were collected

from one stream habitat in Waimanalo Bay (Fig. 2) and these juveniles ranged in size

from 33mm to 56mm. The discrepancy in size of juveniles collected from these two

! 12!

habitats is most likely an artifact of sampling procedures, not necessarily post-settlement

processes, with small individuals in the stream better capable of evading capture. Despite

these differences, there was overlap in the size classes collected and only those

individuals between the range of 35-45 mm were included for habitat specific growth

analysis. The smaller juveniles were still analyzed and the data included for establishing

overall growth relationships.

Fig. 2 Above Left: map depicts the two collection sites for Kuhlia xenura. The top pin corresponds to the stream habitat, pictured on left, while the yellow pin on the right represents the Makapu‘u Tidepools, photo on far right.

Otolith dissection and analysis

Juvenile Kuhlia xenura were measured for standard length, fork length, total

length and wet weight prior to dissection. Sagittal otoliths were carefully removed with

forceps (for details on juvenile otolith dissection see Secor et al. (1991) and Sponaugle

(2009)) and cleaned of any remaining tissue with dissecting needles and a fine tip

paintbrush. Sagittal otoliths were labeled and stored in small plastic centrifuge tubes to

dry (3 otoliths were eliminated before processing). The right sagittal otolith of each

specimen was sectioned in the transverse plane by mounting otoliths (labeled with

individual fish number) in Crystalbond, then grinding and polishing with wet/dry sand

paper and lapping film (of 3, 9, and 12 micron grits) until the core was visible, or in the

event where the core could not be found, rings directly adjacent to this area.

A small amount of immersion oil was added to the section and otoliths were

viewed using a polarizing lens at 200x magnification on a compound microscope with

transmitted light. Benson and Fitzsimmons (2002) previously validated daily growth

bands for Kuhlia xenura. For each otolith, the daily bands from the core to the settlement

! 13!

band (defined for this study as the dark transition zone in which the rings go from heavily

pigmented (brown), wide bands into an area of less pigmentation and noticeably reduced

ring width (Fig. 3)), were counted under the microscope and from the settlement band to

the edge of each otolith, using the image analysis software Rincon. Each otolith was read

along a consistent growth axis, although due to the irregularity of otolith shape and

inherent difficulty of this species for detecting daily rings, reading often involved a non-

contiguous trajectory. Image analysis software allowed for real-time adjusting of focus

to reduce sub-daily rings. The same reader independently counted each otolith three

times and the average values were used in the analysis. Additionally, a coefficient of

variation (calculated as CV= Standard deviation/mean: see review by Campana et al.

(1995)) was calculated for each otolith and those with values greater than 0.10 (only 3

otoliths) were read an additional three times using photographs. Three otoliths were too

difficult to accurately read and were excluded from the analysis. Further, following

Benson et al. (2002), who state that any daily age estimates of juvenile Kuhlia xenura 6

months or greater (180+days) are not reliable, 6 additional specimens were excluded

from the analysis. Fig.3 The settlement band is shown in the photo at Left, from the right saggital otolith of a juvenile Kuhlia xenura. This region is defined as the transition in the otolith from darker colored, wider daily rings to thinner, less pigmented rings, corresponding biologically to a transitions of the fish from a pelagic stage to settled recruit. Statistical analysis

Spatial and temporal analysis of tidepool

community composition

In order to examine the spatial distribution and abundance of the assemblage of

juvenile fishes occupying tidepools on the island of Oahu, the abundance and species

composition data from the monthly visual surveys were used to determine whether an

individual tidepool was equally likely to support a similar composition of juvenile fish as

another tidepool from a different location. Further, we were interested in whether these

differences were consistent across months. To examine the spatial relationship of fish

abundance and species composition among pools and between sites, abundance data was

initially combined for all five months of survey data. The similarity of community

! 14!

composition among pools and sites was then statistically tested using non-parametric

permutation methods in the PRIMER-E statistical software package (Clarke & Warwick

2001).

The observational count data (a measure of abundance) was entered into

resemblance matrices, with individual tidepools as samples and species of juvenile fishes

as variables (with each tidepool treated as a replicate per site). In an effort to account for

the inherently patchy distribution of juvenile fish abundance data, and to temper the dis-

proportional effect of common species, the count data was square root transformed in

PRIMER-E (Clarke & Warwick 2001, Cox et al. 2011). Bray-Curtis similarities were

then computed based on rank similarities of species abundance data by pool. In this

analysis, tidepools with no fish were excluded, however, no more than two tidepools

were excluded in the analysis per month. Next, using the Bray-Curtis similarity matrices,

an analysis of similarity (ANOSIM) was used to statistically test the hypothesis that

tidepools at a given site were relatively more similar in community composition to each

other than tidepools from different sites (Clarke & Warwick 2001) (Tables 1, 3, 5). For

ANOSIM results with significant global p-values, pairwise comparisons were examined

between sites to determine specifically which of the sites were different (Tables 2, 4, 6).

For the pairwise comparison analysis, only R-values approximately 0.6 or greater were

considered meaningful. In order to determine which species contributed most to any

observed differences between sites, the data was then examined using a SIMPER analysis

(Tables 7,8).

To aid in the visualization of these data, non-metric multi-dimensional scaling

plots (nMDS) were also constructed from the Bray-Curtis resemblance matrices (Fig. 4,

6, 8). Each point was color-coded by site and individual months were examined to

determine spatial variability as well as whether spatial patterns were consistent between

months (April-August). Pairwise comparisons were also visualized in separate nMDS

plots to highlight the significant differences between individual sites (Fig. 5, 7, 9).

Individual species comparisons

In order to determine if there were patterns in the spatial distribution of specific

juvenile fish across sites examined on Oahu, the total number of individuals of a certain

! 15!

species observed per tidepool were averaged across the five-month sampling period.

This comparison assumes no interest in the monthly variation of species abundance; thus

treating surveys from each month as a replicate. Differences among sites were

statistically tested using non-parametric Kruskal-Wallis test because the predominance of

zeros violated assumptions of normality. Kruskal-Wallis tested the null hypothesis that

each site was identical in its abundance of a particular species of juvenile fish. For the

case when a significant p-value was obtained, a Dunns joint ranks all pairs test was

administered, to determine from pairwise site comparisons, which of the four sites were

statistically different.

Otolith analysis

In order to establish a relationship between the incremental otolith growth and the

somatic growth of Kuhlia xenura, a regression of the otolith total length (measured to the

closest micron using Image Pro Plus image analysis software), against the standard length

of individual Kuhia xenura was constructed (Fig. 11). Further, growth rates of juvenile

K. xenura were compared between sites by extracting specific size classes (size range of

35-40mm, and 40-45mm) and comparing the ages of juveniles collected from both sites,

for a specific size range, using Wilcoxon non-parametric tests (Fig. 13, 14).

RESULTS

Tidepool juvenile fish assemblages

In total, 15 species of juvenile fish were found to occupy tidepools at the four

basalt benches surveyed on Oahu. Of the species observed, Abudefduf sordidus,

Abudefduf abdominalis, Kuhlia sandvicensis, Kuhlia xenura, and Acanthurus triostegus

could be considered common (each making up 33%, 7%, 3%, 8% and 22% respectively).

Spatial and temporal analysis of tidepool community composition

Results of the ANOSIM test and nMDS plots indicated a significant difference

between sites (global p-value of 0.013*) in the assemblage and abundance of juvenile

fishes occupying tidepools when averaged across all five months (Table 1 and Fig. 4).

! 16!

The null hypothesis that there were no differences in the community composition of

tidepools between sites was rejected. A further pairwise comparison revealed that there

was a significant difference in the assemblage of juvenile fish at Ma‘ili when compared

to tidepools at Makapu‘u (R statistic= 0.628, p-value 0.008*). Thus, for these two sites

(Makapu‘u and Ma‘ili), the community composition of juvenile fish among pools at one

site was different from the tidepool juvenile fish assemblage in tidepools at the other site

(Table 2, Fig. 5).

Next, the spatial component of tidepool communities was examined with respect

to different months (examining community composition data from April, May, June, July

and August separately). ANOSIM tests indicated that for the months of April, May and

August, there was no significant difference in the community composition of juveniles

occupying tidepools at any of the sites surveyed (April: R=0.137, p-value 0.092, May: R=

0.049, p-vaule 0.241, and August: R=0.148, p-value 0.062). In other words, the

variability between tidepools at a site was as great if not greater than the variability

among sites. Thus, there was a lot of similarity in the assemblage of juveniles occupying

tidepools, regardless of site. This is visually demonstrated in the nMDS plots for April,

May and August (see appendix), in which there is a substantial amount of overlap of

points between sites, indicative of overlapping juvenile fish communities between

tidepools. Thus, for the months of April, May and August we can infer that any given

tidepool is equally likely to support a similar assemblage of juveniles.

In contrast, for the months of June and July, there were statistically significant

spatial differences in juvenile fish assemblage structure among sites. For both the months

of June (global p-value of 0.02*), and July (global p-value: 0.024*), tidepools grouped

by site were relatively more similar in community composition than tidepools from other

sites (June Table 3 Fig. 6, July Table 5 Fig. 8). A pairwise comparison revealed that for

both June and July, the sites exhibiting these significant differences were Ma‘ili and

Makapu‘u (June: R= 0.60, p-value 0.008*and July: R= 0.579, p-value 0.018*) (Tables 4,

5). This is visually represented in the nMDS plots (Fig.7, 9), where the proximity of

points from each of these sites were located in multi-dimensional space closer to each

other than to other locations and although a bit overlapping, are clearly different.

! 17!

Simper analysis: which species are responsible for observed differences?

For the month of June, the average dissimilarity between Makapu‘u and Ma‘ili

(~90% dissimilarity) can best be described by the contribution of Acanthurus triostegus,

Abudefduf sordidus and Kuhlia spp. (smaller than 3cm could not be identified to species).

These species contributed to the between site dissimilarity total value 20.6%, 20.4% and

11.6% respectively (Table 7). In the month of July, the greatest differences between sites

(Ma‘ili and Makapu‘u), was a result of the contribution of Acanthurus triostegus,

Abudefduf abdominalis, and Abudefduf sordidus (17.9%, 14.9% and 12.6% respectively)

(Table 8). Of note, both Acanthurus triostegus and Abudefduf sordidus were among the

top three species responsible for explaining the differences between sites for both June

and July between Ma‘ili and Makapu‘u.

Table.1 ANOSIM test for statistical differences in juvenile fish assemblage among sites: All months averaged ANOSIM Factor

Sample Statistic Global R

Significance level Global p-value

Number of permutations (Random sample from 192972780)

Number of permuted statistics greater than or equal to Global R

Site 0.203 0.013 999 12

Table. 2 ANOSIM pairwise tests for all months averaged.

SITES RStatistic p-value

Sandy's, Makapu‘u 0.28 .04

Sandy's, Ma‘ili -0.068 .66

Sandy's, Yokohama 0.132 .175

Makapu‘u, Ma‘ili 0.628 0.08*

Makapu‘u, Yokohama 0.044 .33

Ma‘ili, Yokohama 0.256 .048

Fig.4 nMDS plot of tidepool fish assemblage, all months averaged

! 18!

Table. 3 ANOSIM test for statistical differences in juvenile fish assemblage among sites: June. ANOSIM:

Factor

Sample Statistic Global

R

Significance level

Global p-value

Number of permutations

(Random sample from

102918816)

Number of permuted

statistics greater than or

equal to global R

Site 0.219 0.02* 999 19

Table. 4 ANOSIM pairwise tests for June.

SITES RStatistic p-value

Sandy's, Makapu‘u -0.015 .464

Sandy's, Ma‘ili 0.179 .179

Sandy's, Yokohama -0.067 .571

Makapu‘u, Ma‘ili 0.6 .008*

Makapu‘u, Yokohama -0.11 .722

Ma‘ili, Yokohama 0.392 .032*

All Months Averaged

Fig.5 nMDS plot of similarity between Makapu‘u and Ma‘ili tidepool fish assemblages, all months averaged

! 19!

Table. 5 ANOSIM test for statistical differences in juvenile fish assemblage among sites: July ANOSIM:

Factor

Sample Statistic Global

R

Significance level

Global p-value

Number of permutations

(Random sample from

102918816)

Number of permuted

statistics greater than or

equal to global R

Site 0.197 0.024* 999 23

June

Fig. 6 nMDS plot of tidepool fish assemblage, month of June. !

Fig.7 nMDS plot of similarity between Makapu‘u and Ma‘ili tidepool fish assemblage, month of June

! 20!

Table. 6 ANOSIM pairwise tests for July.

July

SITES RStatistic p-value

Sandy's, Makapu‘u 0.292 0.04

Sandy's, Ma‘ili 0.062 0.375

Sandy's, Yokohama 0.266 0.04

Makapu‘u, Ma‘ili 0.579 0.018*

Makapu‘u, Yokohama 0.086 0.23

Ma‘ili, Yokohama 0.036 0.411

Fig. 8 nMDS plot of tidepool fish assemblage for the month of July !!

Fig. 9 nMDS plot of similarity between Makapu‘u and Ma‘ili tidepool fish assemblage, month of July

! 21!

Table. 7 SIMPER output: species responsible for explaining between site differences at Makapu‘u & Ma‘ili for June: Species are listed in decreasing order of relative importance in contributing to the % dissimilarity between sites. Average dissimilarity (Makapu‘u and Ma‘ili) = 90.97% Species Contributes% Acanthurus triostegus 20.61 Abudefduf sordidus 20.42 Kuhlia spp. 11.56 Kuhlia xenura 10.42 Thalassoma purpureum

9.85 Neomyxus leuscius 5.35 Thalassoma spp. 4.84 Abudefduf abdominalis

4.64 Stethojulis baleata 4.57 Table. 8 SIMPER output: species responsible for explaining between site differences at Makapu‘u & Ma‘ili for July: Species are listed in decreasing order of relative importance in contributing to the % dissimilarity between sites. Average dissimilarity (Makapu‘u and Ma‘ili)= 72.80%

Species Contributes % Acanthurus

triostegus 17.93

Abudefduf

abdominalis 14.88

Abudefduf sordidus 12.60 Abudefduf spp. 9.89 Kuhlia xenura 9.61 Kuhlia spp.. 9.46 Kuhlia sanvicensis 8.96 Thalassoma spp. 6.42 Thalassoma

purpureum 6.21

Individual species comparisons

The considerable contribution of between site variability from relatively few

species of juveniles underlined the need for analysis of spatial distributions of juvenile

fishes at an individual species level. For each of the eight species considered to be

common throughout our surveys, however, results from the non-parametric Kruskal-

Wallis tests, comparing the abundance of individual species among sites, demonstrated a

significant difference in the average abundance of juveniles for only one species, the

manini Acanthurus triostegus (p-value:0.0156) (Fig. 10). Using the Dunn nonparametric

! 22!

multiple comparisons test, the only two sites that were significantly different for A.

triostegus were Makapu‘u and Ma‘ili (p-value 0.0312*). For the other species examined,

there was no statistically significant difference in abundance between sites (see appendix

for other species).

Otolith Analysis

Sagittal otoliths (the right otolith) from 59 juvenile Kuhlia xenura were

successfully extracted from post-settlement specimens collected between August and

November. The smallest recruit collected had a standard length of 17.41mm, which,

according to otolith analysis had only settled as recently as 11 days before capture. In

general terms, these data coincide with findings of Tester and Takata (1953), who

estimated the typical size at settlement to be one inch (or ~25mm).

Otolith length fish length

A significant relationship was found between the incremental otolith growth and

the somatic growth of Kuhlia xenura. The large R2 value of 0.95 and p<0.001* indicates

the utility of otolith daily count interpretation for age and growth studies of this species.

0

1

2

3

4

5

6

7

8

9

10

SB2 SB5 SB6 SB7 SB11 MB1 MB3 MB5 MB6 MB10 MP10 MP12 MP18 MB23 MB27 YB1 YB2 YB3 YB6 YB10

Sandy's Makapu'u Ma'ile Yokohama

P* 0.0312

Aver

age

# of

Man

ini o

bser

ved

per p

ool

Location

Fig. 10 The average abundance of Acanthurus triostegus by site. Bars are color coded by site and represent the average number of A. triostegus for each tidepool across months.

! 23!

Fig. 11 A linear regression of otolith length (microns) to fish length (mm). This graph represents a significant (p < 0.001) linear relationship between otolith and fish length, and establishes relevance otolith studies by demonstrating a correspondence of somatic growth for otolith growth.

Age and Growth:

A regression analysis of the standard length of individual K. xenura on average

age in days indicated that growth rates for our sample of K. xenura displayed linear

growth and was significant (R2 = 0.413 p<0.001) (Fig. 12). The slope of the regression

line relating the standard length of individual Kuhlia xenura to the average age in days

described a growth rate for juveniles of 0.1967 mm/day. The R2 value of 0.413 explains

a significant (p<0.001*) relationship between the standard length of individual fish and

age and is similar to the range of daily growth calculated by Tester and Takata (1953) of

0.186- 0.275mm/day.

Fig. 12 Age and growth of juvenile Kuhlia xenura collected at both habitats. The slope of the linear model represents the growth rate of these species (0.1967 mm/day).

! 24!

However, the data pictured above is representative of juvenile Kuhlia xenura

collected from both the tidepool and stream sites. In order to differentiate growth of

juveniles from either of these distinct habitats, individuals of similar size classes, 35-

40mm and 40-45mm, were separated by collection site (tidepool or stream habitat) and

then their average age compared to determine which site had older individuals for

similar-sized fish and would thus be an indication of growth rate for juvenile K. xenura.

Average age of juveniles collected from both habitats in the size range of 35-

40mm, n=6 from Waimanalo stream and n=4 from Makapu‘u tidepools, were compared

using Wilcoxon non-parametric test. It was assumed that for similar size individuals of

Kuhlia xenura a greater average age was an indication of a slower growth rate. K.

xenura collected from Waimanalo stream were significantly older than fish of the same

size collected from Makapu‘u tidepools (Wilcoxon p-value 0.01*) (Fig. 14). Thus, it can

be concluded that juvenile K. xenura (at least in the restricted size range of 35-40 mm),

are growing faster in tidepools than adjacent stream habitat.

This same procedure was repeated for the size range 40-45mm, with n=7

individuals collected from stream habitat and n=3 for Makapu’u tidepools, but this

comparison indicated a highly variable set of ages for individuals collected at both sites,

with no significant difference in the age for individuals of this size between the two

habitat-types (Fig. 14).

! 25!

Fig. 13 Age and growth data for Kuhlia xenura by location collected: average age (in days) on the x-axis and the SL (in mm) on the Y-axis. Colored dots refer to location collection. The bands (at 35-40mm and 40-45mm) represent the individuals being compared for growth between sites.

Fig. 14 Kuhlia xenura average age verse location for size classes 35-40mm and 40-45mm. For 35-40mm Kuhlia xenura, age was significantly greater (older) for individuals collected from Waimanalo stream, indicating slower growth (Wilcoxon p-value 0.01*) in this habitat. For 40-45mm, there was no significant differences between age and a lot of variability.

! 26!

Pelagic larval duration (PLD):

Another facet of the early life history of Kuhlia xenura that could be elucidated

from otolith microstructural analysis was the duration of time individuals spent as

planktonic larvae before settling in their respective nursery habitats. An average pelagic

larval duration was calculated using otolith microstructure by averaging the number of

rings for all individuals from the core to the settlement band. Back-calculated hatch-date

and settlement date information were also gleaned form otolith microstructure (counting

back from the otolith edge to the core based on date of collecting, and counting back

from edge of otolith to settlement band, respectively). On average, the pelagic larval

duration for Kuhlia xenura was 32.95 days. This value lies between previous estimates

for Kuhlia sandvicensis (before it was determined to be two species: Kuhlia sandvicensis

and Kuhlia xenura) of two months by Tester and Takata (1953) and 1-2 weeks based on

fresh water chemical signatures in otoliths by Benson (2002). The average PLD were

then compared statistically between sites using a simple t-test, but there was no

significant difference (p: 0.19) in PLD by site. Thus, juvenile Kuhlia xenura spent

similar periods of time as planktonic larvae, regardless of whether they eventually settled

in tidepools or streams.

15

20

25

30

35

40

45

50

Fig. 15 Pelagic larval duration of Kuhlia xenura calculated as the number of days from the core of an otolith to the settlement band. The average PLD was 32.95 days.

! 27!

Hatch-date periodicity and settlement dates

The back-calculated hatch-dates for juvenile Kuhlia xenura appeared to peak in

the month of July, after which there was a decrease in the number of individuals.

Although the dates calculated may be +/- a few days depending on the timing of the first

ring (either hatching or first feeding), when the frequencies were examined by months, an

estimate of the periodicity in hatch-dates was obtained. There was a peak in the month of

August for settlement dates, although values were also high for June and July. The

settlement date data was then divided by location, Waimanalo stream and Makapu‘u

tidepools, and there appears to be differences in the timing of settlement between these

two nearshore areas. Although a greater number of individuals and habitat replicates is

needed to statistically distinguish between-site differences in the timing of early life

history events, K. xenura appeared to settle with the greatest frequency in July and

August for fish collected from tidepools, while juvenile K. xenura collected from stream

habitats may be recruiting greatest in June and August.

Fig. 16 Back Calculated Hatch dates for Kuhlia xenura inferred from otolith microstructure

Fig. 17 Back calculated settlement dates for Kuhlia xenura inferred from otolith microstructure

MARCH APRIL MAY JUNE JULY AUGUST SEPTEMBER

APRIL MAY JUNE JULY AUGUST SEPTEMBER OCTOBER

! 28!

! Makapu‘u Tidepools Waimanalo Stream

Fig. 18 Back calculated settlement dates for Kuhlia xenura by site.

DISCUSSION

Tidepool juvenile fish assemblage

Visual survey data revealed temporally inconsistent patterns in assemblage

structure among tidepools and sites on Oahu. A total of 15 species of Hawaiian reef fish

were found to occupy tidepools as juveniles, but the variability in their abundance, even

on a particular basalt bench, reveals the breadth of variability both spatially and

temporally within this habitat during the summer. Previous work on recruitment patterns

of Hawaiian fishes to coral reef substrate suggested recruitment events were highly

variable spatially and temporally (Walsh 1987), and tidepools also appear to reflect this

variability in recruitment. Surveys found juvenile fish assemblage structure at two of the

sites surveyed for the months of June and July, as when averaged overall, while April,

May and August demonstrated no statistical difference in community composition.

SIMPER analysis confirmed that between site differences were the result of

relatively few species, and suggests the importance of individual species’ pulse

recruitment events in structuring the communities present in tidepools. For example, Cox

et al. (2011), explored the spatial variability in tidepool fish assemblages (both juveniles

and residents) in tidepools on limestone and basalt benches on Oahu, and did not

encounter Abudefduf sordidus during surveys at Sandy’s bench. However, findings from

this study revealed a great level of variability in the abundance of Abudefduf sordidus on

Sandy’s bench, with a total of 14 individuals observed in April and only 7 individuals

observed in May. Thus, it is important consider the temporal component when making

APRIL MAY JUNE JULY AUG SEP OCT APRIL MAY JUNE JULY AUG SEP OCT

! 29!

inferences about community composition as it can fluctuate significantly between

months. Had this survey been conducted only once, it may have missed the differences

in community composition between sites and the pulsed recruitment events of individual

species.

It is possible that the spatial and temporal variability observed in this study was a

result of ecological interactions occurring within this habitat, especially predation

pressure and food availability. Future research should include a greater number of sites

and surveys conducted over an even greater amount of time so as to encapsulate inter-

month variability and fine-scale spatial heterogeneity. Further, our methods neglected to

incorporate fine scale environmental parameters, including water quality, that may help

define the factors leading to the observed differences in assemblage structure between

sites.

The variability in recruitment events underlines a need to also examine the factors

affecting the dispersal of recruiting fishes before they enter the intertidal habitat. Spatial

variability in recruitment of reef fishes to island habitats is often assumed to be the result

of processes affecting the pelagic larval stage. In the Hawaiian archipelago, one

proposed mechanism for the distribution of recruiting fish to nearshore nursery habitats

has been a gradient of wind-wave exposure. One study in the Northwest Hawaiian

Islands examined the distribution of juvenile reef-fish on two isolated atolls, Pearl &

Hermes and Kure, to test the hypothesis that juvenile densities in rubble back-reef

habitats would be greater in regions fronting east-northeasterly trade winds and surface

currents. The hypothesized mechanism was that larvae were more likely to be advected

from this direction and therefore may be filtered out through the atoll (DeMartini et al.

2009). Overall, a greater density of recruits, specifically endemic labroids, was found on

windward versus leeward facing shores, supporting this hypothesis.

! 30!

Figure 19. A map of study sites grouped by wind-wave exposure. The arrows correspond to sites with the greatest abundance of Acanthurus triostegus, Kuhlia xenura and Kuhlia sandvicensis and clearly these are not concentrated on either wind-wave regime.

Although an atoll system in the Northwest Hawaiian Islands is quite different

from the conditions present in the Main Hawaiian islands, which are considered high

islands and represent a much larger area, this project tested whether these patterns were

also true for the sites surveyed on Oahu. Patterns of species abundance were explored to

evaluate if a windward, leeward gradient affected the distribution of juvenile fishes

inhabiting tidepools. Fig. 19 depicts the study site locations with respect to wind wave

exposure typical of the summer months when this study was conducted. The three fish

species in the diagram represent the three species currently regulated by the Hawaii state

Division of Aquatic Resources (DAR) and whose distribution would be of interest to

management agencies interested in designating specific areas for conservation of

essential habitat. The diagram represents the two areas for each species with the greatest

abundance, analyzed to observe any obvious wind-wave exposure patterns. However, for

all three species of juveniles, there was no obvious wind-wave exposure component to

the distribution of resource species, with one site hosting a great abundance on the wind-

wave exposed shore, and the other on the leeward side of the island. This does not rule

out the importance of prevailing wind and current conditions, rather it indicates that for a

large, high island like Oahu, perhaps the current patterns responsible for observed

patterns should be considered at a finer-scale.

To test whether tidepools are indeed serving as a nursery habitat for juvenile reef

fish, it is important to test whether tidepools host greater densities of juveniles than other

! 31!

potential nursery habitats. Although this study did not attempt to compare the densities

observed to other juvenile habitats, evidence from Australia for tidepool fish assemblages

has revealed three juvenile species common in tidepools in Australia, Chaetodon auriga,

Abudefduf vagiensis and Kuhlia mugil, (Griffiths 2003), that are similar to species

recorded from our study (Abudefduf abdominalis, Kuhlia xenura, Kuhlia sanvicensis

Chaetodon auriga). Similar species (Acanthurus triostegus and Abudefduf sordidus and a

species of Kuhlia) were also observed in rocky-intertidal pools in Tonga (Vava’u island

group) during the summer (personal observation). The utilization of tidepools by similar

(or identical) species as juveniles in Hawaii, in locations with abundant classical nursery

areas (such as seagrass beds and mangroves) in close proximity, further emphasizes the

importance of tidepools as a nursery habitat. Future research would benefit from a

comparison of juvenile fish densities inside and outside rocky-intertidal areas and an

expansion of the potential importance of tidepools as a nursery area to other regions of

the world.

Otoliths microstructure analysis

The significant difference in average age of similar sized (35-40mm) Kuhlia

xenura collected from a stream and adjacent tidepools, with individuals collected from

the stream habitat significantly older than juveniles of the same size collected in

tidepools, is a strong indication of the importance of tidepools as a nursery habitat during

early stages of growth. Because growth is an important factor in determining the

survivorship of post-settlement fishes to adulthood, the availability of basalt tidepools

may be critical in the successful replenishment of adult stocks. Although a greater

sample size would better clarify the habitat-specific growth effects of these two habitat

types for the full range of sizes of juveniles present in these areas, the abundance of

juveniles of K. xenura in both habitats, as well as their variable growth within these areas

lends evidence to the equal importance of both of these habitats as nursery areas. Benson

et al. (2002) proposed that stream habitats might serve as a nursery for K. xenura on

windward facing shores, while tidepools may be of greater importance to juveniles on

leeward coasts, which have limited stream habitats. In this project, which was conducted

on the Windward side, results indicate that perhaps tidepools are important for recently

! 32!

settled individuals and both are important habitats for larger juveniles. Future research

could further test this hypothesis, as well as the movement of individuals between these

habitats, by tagging juveniles and analyzing tag-recapture rates. Further, while interested

in comparing juvenile habitats in close proximity, the stream surveyed and possibly the

tidepools at Makapu‘u, showed some obvious signs of human impact (upstream

development in the stream area as well as an abundance of non-native stream fishes,

along with the presence of derelict fishing equipment in the tidepools). Perhaps the

results would be even more meaningful if this study could be repeated in habitats un-

disturbed by human activities.

The inherent variability in temperature, food and shelter availability may have

been important factors in creating the variable growth rates calculated in this study. On

the dates of collection of juveniles, the temperatures were very different between these

two sites, with Waimanalo stream exhibiting decreased temperatures and salinities

compared to tidepools. Future studies would benefit from a quantitative, fine scale

measurement of abiotic factors potentially responsible for affecting growth. Additionally,

although samples were purposefully collected over similar time frames, it is possible that

the abiotic conditions present varied between sampling periods.

Pelagic larval duration

A recent study by Feutry et al. (2012) examined the pelagic larval duration of two

species of Kuhliidae (same family as Kuhlia xenura) from Indo-Pacific waters. The

authors examined two species with different life histories, one endemic to the Indian

Ocean Kuhlia sauvagii and the other species Kuhlia rupestris widely distributed across

the Indo-Pacific. The authors discovered that the endemic species K. sauvagii had a

significantly shorter pelagic larval duration, defined by counting daily otolith rings from

the core to the settlement band, (the definition of which is consistent with the settlement

band distinction of this study). The endemic Kuhliidae had a pelagic larval duration of

32.3+/- 3.4 days compared to the species with an Indo-pacific distribution, 40.6+/-6.9.

Interestingly, the mean pelagic larval duration of K. xenura, was 32.95 days, which is

surprisingly similar to the endemic species of the Feutry study. K. xenura is endemic to

! 33!

the Hawaiian archipelago while Kuhlia sanvicensis is distributed across the Indo-Pacific.

An interesting future study could examine whether there are differences between K.

xenura and K. sanvicensis in terms of pelagic larval duration as it is often assumed that

endemic species have early life stages that limit their dispersal capabilities when

compared to similar species with widespread distributions.

Conclusions

The adults of many nearshore fish species which utilize rocky-intertidal pools as

juveniles, specifically Kuhlia xenura, Kuhlia sandvicensis, Acanthurus triostegus,

Neomyxus leuciscus, among others, are being harvested for consumption as food. In the

case of K. xenura, K. sandvicensis, A. triostegus, in quantities high enough to necessitate

minimum size limits by the Hawaii Division of Aquatic Resources. The importance of

understanding the early life history, ecology and dynamics of these species is critical in

projecting the success of adult populations and thus the sustainability of their catch. In

addition to consumptive practices, numerous tidepool species are important culturally, as

well as in collection for the aquarium trade. Currently, no research exists which explores

the potential effects human activities have upon the tidepool habitat and the ability of

juvenile fish to survive and contribute to adult populations. Less than 1% of the main

Hawaiian Island marine habitat is protected from fishing, and none of these designations

have explicitly incorporated nursery areas within their borders. Yet, nursery habitats help

ensure the successful recruitment of future generations of marine fish species and should

be the focus of future research and conservation efforts.

! 34!

APPENDICES

APPENDIX A

A brief review of the history of otolith analysis in fishery science:

Otoliths, the calcified structures comprising the inner ear system in teleost fishes,

have historic utility in fisheries ecology research. Due to the difficulty in studying the

early stages of marine teleosts, involving challenges in observing spawning and dispersal

of larval stages and settlement behavior in situ, otoliths have played an increasingly

important role in determining the timing of important transitional events. Three pairs of

otoliths; the sagittae, lapilli and asterisci are located adjacent to the skull where they are

contained within otic organs filled with endolymph fluid. Otoliths aid in balance,

orientation and hearing for a living fish and post-mortem provide invaluable information

about the age, growth and environmental conditions experienced by the fish. The

deposition of calcium carbonate, typically in the form of aragonite, occurs incrementally

throughout the life of a fish, with concentric dark and light bands corresponding to

periods of differential growth. Due to the inert nature of otolith deposition, the structural

integrity and history recorded in the layers of the otolith is never further modified by

cellular processes and thus can be used retroactively to determine the history of the

organism.

The discovery by Pannella (1971) and verification by Brothers et al. (1976) that

incremental growth bands in young fish correspond to daily growth heralded the use of

otoliths for age determination in larval and juvenile stage fish. Prior to this finding, it

was assumed that because tropical fish were not subject to the same climatic variability as

temperate species, their otoliths were not usable for aging. In addition to daily age

determination, the microstructure of the otolith also records important events in the early

history of collected specimens. The discovery that significant transitional events in the

life history of fish are recorded in otolith structure has allowed for a range of inferences

into the timing of early life-stage events in coral-reef species. Specifically, early work by

Victor (1982, 1983) and Brothers and McFarland (1981), found that following a

planktonic larval stage, the transition of Caribbean wrasses to a benthic-associated

juvenile stage is recorded in an otolith as a “settlement band.” Thus, given a known date

! 35!

of collection, researchers can back-calculate the duration of time spent as a pelagic larva

and infer patterns and timing of spawning and recruitment. The pelagic existence of a

species is vital to understanding the distribution, periodicity and frequency of recruitment

events, and this information is critical to the successful management of fish stocks, as

well as the broader understanding of evolutionary relationships among species.

Another important inference that can be made through otolith analysis is the

estimation of growth rates. According to work by Campana & Neilson (1985) (and

recently reviewed by Vigliola & Meekan 2009), the width of increments is proportional

to the somatic growth rate of a fish. Daily growth increments can then be back calculated

to establish size-at-age estimations, as well as instantaneous growth rates. These

relationships are often species dependent and for juvenile fishes, this information can be

used to construct growth curves, often linear for juvenile growth (Jones 2002). Thus, it is

possible to calculate various important life history parameters from otoliths, including a

calculation of growth rates, which can be compared among habitat types.

In tropical ecosystems, few studies have focused on among site differences in

relative nursery quality with one notable exception: In the Caribbean, Mateo et al. (2011)

utilized otolith inferred growth rates to compare growth of French Grunts Haemulon

flavolineatum and Schoolmaster Lutjanus apodus residing in mangrove and seagrass

beds. Results indicated that at both islands sampled, mangrove habitats provided

conditions resulting in increased growth for both species when compared to comparable

seagrass habitat on both island systems.

References 1. Brothers EB and McFarland WN. 1981. Correlations between otolith microstructure,

growth, and life history transitions in newly recruited French grunts (Haemulon flavolineatum (Desmarest), Haemulidae). Rapport Procesvervaux des reunions, Counseil Permanent International pour lʻ Exploration de la Mer. 178: 369-374.

2. Brothers, Edward. 1976. Daily growth increments in otoliths from larval and adult

fishes. Fishery Bulletin. 74 (1):1-8.

3. Campana Steven and Neilson John. 1985. Microstructure of fish otoliths: Review. Canadian Journal of Fish and Aquatic Science. 42: 1014-1032.

! 36!

4. Jones Cynthia. 2002. Chapter 2: Age and Growth. Fishery Science the Unique Contributions of Early Life Stages. Edited by Fuiman, Lee and Werner, Robert. © Blackwell Science.

5. Mateo Ivan, Durbin Edward, Appledoorn Richard, Adams Aaron, Juanes Francis and

Durant Daisy. 2011. Inferred growth of juvenile French Grunts, Haemulon flavolineatum andSchoolmaster, Lutjanus apodus, in mangrove and seagrass habitats. Bulletin of Marine Science. 87 (0): 1-12.

6. Pannella Giorgio. 1971. Fish otoliths: daily growth layers and periodical patterns.

Science. 173 (4002): 1124-1127.

7. Victor Benjamin. 1982. Daily otolith increments and recruitment in two coral-reef wrasses, Thalassoma bifasciatum and Halichoeres bivttatus. Marine Biology. 71: 203-208.

8. Victor Benjamin. 1983. Settlement and larval metamorphosis produce distinct marks on

the otoliths of slippery dick Halichoeres bivittatus. The ecology of deep and shallow coral reefs. Underwater research symposium series, NOAA. 1: 47-51.

9. Vigliola Laurent, Meekan Mark. 2009. The back-calculation of fish growth from

otoliths. Tropical Fish Otoliths: information for assessment, management and ecology. Edited by Green, Bridget, Mapstone, Bruce, Carlos Gary and Begg, Gavin. © Springer.

10.

APPENDIX B

Study organism: Kuhlia xenura

A comparative growth rate investigation to determine the relative importance of

tidepools as a nursery habitat requires a study species with some known life history

characteristics. For the assemblage of juvenile fish species utilizing tidepool habitats in

Oahu, only the āholehole (there are two species of āholehole in Hawaii: Kuhlia xenura

and Kuhlia sanvicensis) has appreciable information available regarding aspects of the

recruitment, ecology, and distribution of juveniles.

The importance of K. sanvicensis and K. xenura in Hawaii dates back to its

traditional use in ceremonies by early Hawaiians as well as its desirability as a food fish,

even by chiefs (Titcomb 1952). K. sanvicensis and K. xenura have a convoluted

taxonomic history, with early biologists describing only one species K. sandvicensis from

Hawaii, though some researchers allude to the existence of another species, Kuhlia

marginatus. Recent morphometric and genetic analysis however has definitively

! 37!

revealed the presence of two distinct species of Kuhlia in Hawaiian waters: K.

sandvicensis, with an Indo-Pacific distribution and K. xenura, a Hawaiian endemic

(Benson 2002). Although a small commercial fishery has existed in the past for this

species, it is currently more valuable for recreational, consumptive purposes. Due to

concern over the declining populations of coastal resource species, the Hawaii

Department of Aquatic Resources maintains a list of species that require regulation, and

both species of Aholehole have minimum-length requirements on this list. However, the

regulation: a minimum length of 5 inches, with no bag limit, does little to protect the

early life stages of this nearshore resource species, and belies our greater lack of

understanding of the early habitat requirements (nursery habitats) of this important

resource species in Hawaii.

Although direct observation of spawning behavior of āholehole has never been

documented, it has been inferred that spawning occurs in open marine waters (or caverns

associated with coral reefs) throughout the year, with a peak during winter and spring

months (Tester & Takata 1953). Tester and Takata (1953) also assume that, due to the

initial appearance of juveniles in coastal waters approximately one inch (~22mm) in size,

following hatching, āholehole (previously used to describe both K. sanvicensis and K.

xenura) undergo a two-month pelagic larval phase. Little is known about the time these

species spend in the water column as planktonic larvae, except that they appear to be rare:

12 larvae of Kuhlia spp. (these surveys predated the recognition of two distinct species)

were collected from a total of 155,390 larvae collected as part of an extensive larval-fish

vertical distribution study (see Boehlert & Mundy 1996). Following this variable pelagic

stage, K. sanvicensis and K. xenura undergo metamorphosis and take up residence in

specific shallow habitats. Despite the presence of recruits year round, there is a strong

seasonal component to recruitment with marked increases in abundance between

February and August, and a peak in May (Tester & Takata 1953).

Due in large part to the extreme physiological tolerances of the āholehole to

fluctuations in salinity, oxygen and temperature (Takata 1953), juvenile āholehole are

found in a diversity of coastal habitats. The lack of differentiation between the two

species of āholehole in the older literature precludes conclusions about the distribution of

juvenile K. xenura specifically, but authors described “āholehole” as inhabiting

! 38!

freshwater stream habitats, rocky shores, tidepools and sandy beaches (Tester & Takata

1953, Gosline 1965). Since these early reports, research on the distribution of juvenile

āholehole has been able to distinguish K. xenura from K. sandvicensis. Both K. xenura

and K. sanvicensis have been observed as juveniles in brackish coastal wetlands

(MacKenzie & Bruland 2012), as well as rocky intertidal benches and tidepools (Cox et

al. 2011, Benson 2002) and sandy beach surge zones (Friedlander, personal

communication). Recent findings by MacRae et al. (2011), however have revealed that

although both species occur in these habitats, on a smaller ecological spatial scale, K.

xenura select distinct microhabitats. Using a principle component analysis to compare

the distribution of juveniles observed on the big island with visual surveys, MacRae et al.

(2011) found that both species were non-randomly distributed across all potential

juvenile habitats, with K. sandvicensis exclusively utilizing marine habitats and K. xenura

distributed to specific microhabitats within streams, estuaries, reef flats and rocky shores

and tidepools. Further, MacRae et al. (2011) found that within tidepool habitats, K.

sandvicensis are more commonly encountered in high-energy pools in the surge zone

while K. xenura was found in higher pools protected from strong wave action. This is

further strengthened by the report by MacKenzie & Kryss (2010) that mentioned K.

xenura as the most abundant native fish in the Wai’opae tidepools, with no observation of

K. sandvicensis. Based on previous work on the presence of K. xenura in numerous

coastal juvenile habitats, an exploration of the comparative growth rate of juvenile K.

xenura will thus be informative and relevant to determining the habitat quality of these

potential nursery areas. Both papers that found K. xenura as juveniles in tidepools (Cox et

al. 2011, MacRae 2011) conclude that there exists a need for research testing the nursery

role function of this habitat.

References 1. Benson Lori. 2002. Aspects of the behavioral ecology, life history, genetics, and

morphology of the Hawaiian Kuhliid fishes. Dissertation to Louisiana State University, Department of Biological Sciences.

2. Boehlert George and Mundy Bruce. 1996. Ichthyoplankton vertical distributions near

Oahu, Hawaii, 1985-1986: Data Report. NOAA-TM-NMFS-SWFSC-235. US Department of Commerce.

! 39!

3. Cox Traci Erin, Baumgartner Erin, Philippoff Joanna and Boyle Kelly. 2011. Spatial and vertical patterns in the tidepool fish assemblage on the island of Oahu. Environmental Biology of Fishes. 90 (4): 329-342.

4. Gosline William. Vertical zonation of inshore fishes in the upper water layers of the

Hawaiian Islands. 1965. Ecology. 46 (6): 823-831.

5. MacKenzie Richard and Bruland Gregory. 2012. Nekton communities in Hawaiian coastal wetlands: the distribution and abundance of introduced fish species. Estuaries and Coasts. 35: 212-226.

6. MacKenzie Robert and Kryss Caitlin. 2010. Tidepool fish assemblages at WaiʻOpae

marine life conservation district, Hawaii: Monitoring the effects of mangrove eradication on nearshore fish assemblages. A preliminary report prepared for Malama O Puna. Institute of Pacific Islands Forestry. 1-28.

7. McRae Mark, McRae Lori and Fitzsimons Michael. 2011. Habitats used by juvenile

flagtails (Kuhlia spp.; Perciformes: Kuhliidae) on the Island of Hawaii. Pacific Science. 65 (4): 441-450.

8. Takata Michio. 1953. Effects of variations in salinity, temperature, and oxygen

concentration upon the āholehole, Kuhlia sandvicensis (Steindachner). Thesis University of Hawaii, Manoa, Department of Zoology.

9. Tester Albert, Takata Michio. 1953. Contribution to the biology of the āholehole a

potential baitfish. Hawaii Marine Laboratory, Industrial Research Advisory Council Grant No. 29. 38: 1-54.

APPENDIX C

History of otolith research on Kuhlia xenura

The first attempt to age āholehole (again at the time considered one species,

currently Kuhlia xenura and Kuhlia sandvicensis) through otolith analysis was conducted

by Tester and Takata (1953). The authors collected Aholehole of various size classes and

tried to age fish using incremental growth bands on scales, otoliths and vertebrae.

However, despite their attempts at an accurate age assessment, “the results were too

confusing to warrant age estimates.” Despite the inability of producing reliable age

estimates from otoliths, the authors were able to surmise daily growth rates for āholehole

using data sampled in the field: The authors conducted weekly recruitment surveys of

juvenile āholehole from Diamond head, Oahu over two recruitment seasons (January-July

and January-May in 1952 and 1953, respectably). Through the persistence of recruitment

pulses in one location over several weeks, researchers were able to track the growth of

! 40!

recruits through estimation of the modal size of recruits in their surveys. By plotting

these size estimates over time, the authors were able to create a daily growth rate for

recently recruited fish. Their estimate of mean daily growth was 0.246mm/day and this

was verified to some extent by measuring āholehole in aquaria and salt-water ponds,

though the methods for these studies should be further evaluated before considered

accurate (Tester & Takata 1953). Nakamura (1968) also attempted to construct growth

curves for āholehole, but the smallest size class he included was 80mm-100mm which is

large considering the typical size at which an āholehole recruits ~16mm.

The next attempt at aging āholehole, by Benson and Fitzsimons (2002) and

Benson (2002), was the first to target Kuhlia xenura as a separate species. In order to

examine whether K. xenura have an obligatory fresh-water phase, the investigators

collected 10 K. xenura from four habitat types: stream mouth, upstream, tidepools and

ocean surge zones. Utilizing daily growth bands, the authors were able to determine the

specific salinity conditions experienced by each fish at precise ages. An analysis of the

microchemistry of an otolith can reveal great detail about the environmental conditions

experienced by an individual fish. As otoliths accrete calcium carbonate, occasionally

strontium is incorporated into the otolith matrix instead of calcium. Because freshwater

habitats have significantly less strontium available than marine habitats, the ratio of

strontium to calcium in a particular layer of an otolith gives an indication of the local

salinity experienced by the organism at a particular age. K. xenura juveniles had a high

Sr/Ca ratio for the first couple weeks, indicating they encountered fresh water stream

habitats after only a couple weeks. Because marine habitats are characterized by greater

Sr availability than Hawaiian streams, it is assumed that juvenile K. xenura undergo

marine pelagic larval stage from one to two weeks in duration. The authors concluded

that although each specimen examined had experienced fresh water conditions at some

part during their life, freshwater use was facultative and not critical to the completion of

their life cycle. Further, for the juvenile K. xenura collected from tidepools, the author

found that 21mm and 30mm juveniles were 56 and 73 days old, while in fresh water

streams, juveniles sized 23-25mm were between 75 and 93 days old. Although the

authors failed to calculate any growth rates in their study, it is interesting to note that

juveniles in tidepools appear to be growing faster. Benson (2002) also validated the

! 41!

daily growth bands of juvenile Kuhlia xenura by staining juvenile fishes with Alizarin

dye and validating the ring formation over a number of known days. Choat et al. (2009)

has stated that it is important to verify the daily nature of otolith banding on a species

level in order to have meaningful results.

References:

1. Benson Lori and Fitzsimons Michael. 2002. Life history of the Hawaiian fish Kuhlia sandvicensis as inferred from daily growth rings of otoliths. Environmental Biology of Fishes. 65: 131-137.

2. Benson Lori. 2002. Aspects of the behavioral ecology, life history, genetics and morphology of the Hawaiian Kuhliid fishes. Dissertation Louisiana State University and Agricultural and Mechanical College, Department of Biological Sciences.

3. Choat JH, Kritzer JP, Ackerman JL. Chapter 2: Ageing in coral reef fishes: do we need to validate the periodicity of increment formation for every species of fish for which we collect age-based demographic data? Tropical Fish Otoliths: Information for assessment, management and ecology. Edited by Green Bridget, Mapstone Brue, Carlos Gary and Begg, Gavin. © Springer 23-54

4. Nakamura Royden. 1968. An additional contribution to the biology of the āholehole, Kuhlia sandvicensis (Steindachner). Pacific Science. 22 (4): 493-496.

5. Tester Albert and Takata Michio. 1953. Contribution to the biology of the Aholehole a potential baitfish. Industrial research advisory council grant no. 29. Final Report: 1-54.

! 42!

APPENDIX D

Tidepool volume comparison by site

A comparative analysis of an index of volume (calculated as the maximum length

of a pool multiplied by the minimum width of the pool multiplied by the average depth

(measured three times per pool), by site demonstrated that there was no statistically

significant differences in the volume of tidepools among sites (this was accomplished

using non-parametric Kruskal-wallis test with a p-value of 0.8206).

APPENDIX E

Condition indices: Fulton’s condition K

In order to determine a crude estimate of overall fish condition between the two

sites compared: Makapu’u tidepools and Waimanalo stream, fish were measured

(standard length) and weighed to obtain Fulton’s condition factor K values (K= 100* Wet

weight (g)/Standard Length^3). This standard fishery measure is an indication of the

general condition of the fish and is useful in comparing estimates of condition for

K.xenura residing in both habitats and give an indication of general habitat quality.

Volume index was calculated for each pool surveyed and statistically compared between sites using Kruskal-Wallis (p non significant)

! 43!

A picture of juvenile Kuhlia xenura and a measurement of Standard length.

Fultonʻs K condition comparison

Fulton’s K condition index was

calculated for each fish based on

length and weight data and compared

between sites. After determining the

normality of this data, a comparison of

average K values was statistically

tested using a t-test (p-value: 0.30).

There was no significant difference in

the condition of juveniles between the

Waimanalo stream habitat (in the figure referred to as Bellows st) and Makapu’u

tidepools. We can thus conclude that there was no obvious difference in the overall

condition (or health) of juveniles collected from tidepool and native stream habitats,

which would indicate the relative importance of one habitat over the other.

Fig.5 A plot of Fulton’s condition factor, K for the two habitats compared. The average condition factors were not significant different

! 44!

APPENDIX F

Common species observed in tidepools: Pie chart for all observations

! 45!

APPENDIX G

Tidepool assemblage analysis for non-significant months

April

ANOSIM test for April: No statistical differences in juvenile fish assemblages among sites ANOSIM Factor

Sample Statistic Global R

Significance level Global p-value

Number of permutations (Random sample from 192972780)

Number of permuted statistics greater than or equal to Global R

Site 0.137 0.092 999 91

May

ANOSIM test for May: No statistical differences in juvenile fish assemblages among sites ANOSIM:

Factor

Sample Statistic Global

R

Significance level

Global p-value

Number of permutations

(Random sample from

488864376)

Number of permuted

statistics greater than or

equal to Global R

Site 0.049 0.241 999 240

An nMDS plot for May representing the similarity of individual tidepools in their juvenile fish assemblage. Colors represent sites from which survey data was collected. The relative distance between marks is an indication of relative similarity, with closer marks more similar.

An nMDS plot for April representing the similarity of individual tidepools in their juvenile fish assemblage. Colors represent sites from which survey data was collected. The relative distance between marks is an indication of relative similarity, with closer marks more similar.

! 46!

August

ANOSIM test for August: No significant difference in the assemblage structure of tidepools from sites surveyed ANOSIM:

Factor

Sample Statistic Global

R

Significance level

Global p-value

Number of permutations

(Random sample from

488864376)

Number of permuted

statistics greater than or

equal to global R

Site 0.148 0.062 999 61

APPENDIX H

Individual species of fish, patterns of abundance (non-significant) The following figures represent the average number of individuals of a particular species observed per pool, across all five months surveyed. The Y-axis represents the average number of individuals observed in a single pool, for the five months of observation. The x-axis represents the individual pools, and the sites on which they are located.

An nMDS plot for August representing the similarity of individual tidepools in their juvenile fish assemblage. Colors represent sites from which survey data was collected. The relative distance between marks is an indication of relative similarity, with closer marks more similar.

! 47!

! 48!

! 49!

! 50!

APPENDIX I

Pelagic larval duration by site

Comparison of PLD’s for Kuhlia xenura collected from Bellows stream in Waimanalo Bay and tidepools at Makapu’u

! 51!

REFERENCES Amara R, Selleslagh J, Billon G and Minier C. 2009 Growth and condition of 0-group European flounder, Platichthys flesus as an indicator of Estuarine habitat quality. Hydrobiologia. 627 (1): 87-98. Amara R, Meziane T, Gilliers C, Hermel G and Laffargue P. 2007 Growth and condition indices in juvenile sole Solea solea measured to assess the quality of essential fish habitat. Marine Ecology Progress Series. 351: 201-208. Beck Michael, Heck Kenneth, Kenneth W Able, Childers Daniel, Eggleston David, Gillanders Bronwyn M, Halpern Benjamin, Hays Cynthia, Kaho Hoshino, Minello Thomas, Orth Robert, Sheridan Peter and Weinstein Michael. 2001. The Identification, conservation, and management of estuarine and marine nurseries for fish and invertebrates. BioScience. 51(8): 633-641. Beckley Lynnath. 1985. Tide-pool fishes: recolonization after experimental elimination. Journal of Experimental Marine Biology and Ecology. 85 (3): 287-295. Bennett, BA. 1986. The rock-pool fish community of Koppie Alleen and an assessment of the importance of Cape rock-pools as nurseries for juvenile fish. South African Journal of Zoology. 22 (1): 25-32. Benson Lori. 2002. Aspects of the behavioral ecology, life history, genetics and morphology of the Hawaiian Kuhliid fishes. Dissertation Louisiana State University and Agricultural and Mechanical College, Department of Biological Sciences. Benson Lori and Fitzsimons Michael. 2002. Life history of the Hawaiian fish Kuhlia sandvicensis as inferred from daily growth rings of otoliths. Environmental Biology of Fishes. 65: 131-137. Booth David. 1992. Larval settlement patterns and preferences by domino damselfish Dasyllus albisella Gill. Ecology. 155 (1): 85-104. Campana Steven, Annand Christina and McMillan James. 1995. Graphical and statistical methods for determining the consistency of age determinations. Transactions of the American Fisheries Society. 124: 131-138. Castellanos-Galindo GA, Giraldo A, Rubio EA. 2005. Community structure of an assemblage of tidepool fishes on a tropical eastern Pacific rocky shore, Colombia. Journal of Fish Biology. 67 (2). 392-408. Clarke KR, Warwick RM. 2001. Change in marine communities: an approach to statistical analysis and interpretation, 2nd edition. PRIMER-E, Plymouth

! 52!

Cox Traci Erin, Baumgartner Erin, Philippoff Joanna and Boyle Kelly. 2011. Spatial and vertical patterns in the tidepool fish assemblage on the island of Oahu. Environmental Biology of Fishes. 90 (4): 329-342. DeMartini Edward. 2004. Habitat and endemism of recruits to shallow reef fish populations: Selection criteria for no-take MPAs in the NWHI coral reef ecosystem preserve. Bulletin of Marine Science. 74 (1): 185-205. DeMartini Edward and Anderson Todd. 2007. Habitat associations and aggregation of recruit fishes on Hawaiian coral reefs. Bulletin of Marine Science. 81 (1): 139-152 DeMartini Edward, Anderson Todd, Kenyon Jean, Beets James and Friedlander Alan. 2010. Management implications of juvenile reef fish habitat preferences and coral susceptibility to stressors. Marine and Freshwater Research. 61: 532-540. DeMartini Edward, Zgliczynski Brian, Boland Raymond and Friedlander Alan. 2009. Influences of wind-wave exposure on the distribution and density of recruit reef fishes at Kure and Pearl and Hermes Atolls, Northwester Hawaiian Islands. Environmental Biology of Fishes 85: 319-332. Feutry Pierre, Valade Pierre, Ovenden Jennifer, Lopez Pascal Jean and Keith Philippee. 2012. Marine and Freshwater Research 63: 397-402. Friedlander Alan and Ziemann David. 2003. Impact of hatchery releases on the recreational fishery for Pacific threadfin (Polydactylus sexfilis) in Hawaii. Fishery Bulletin. 2003. 101(1): 32-43. Friedlander Alan. 2004. Status of Hawaiiʻs coastal fisheries in the new millennium. Proceedings of the 2001 fisheries symposium sponsored by The American Fisheries Society Hawaii Chapter. Gibson RN and Yoshiyama RM. 1999. Intertidal Fish Communities. Intertidal Fishes: Life in Two Worlds (book). Edited by Michael Horn, Karen Martin and Michael Chotkowski. Academic press. Gilliers Camille, Amara Rachid, Bergeron Jean-Pierre, Le Pape Olivier. 2004. Comparison of growth and condition indices of juvenile flatfish in different coastal nursery grounds. Environmental Biology of Fishes 71: 189-198. Gilliers Camile, Le Pape Olivier, Desaunay Yves, Bergeron Jean-Pierre, Schreiber Nathalie, Guerault Daniel and Amara Richard. 2006. Growth and condition of juvenile sole Solea solea as indicators of habitat quality in coastal and estuarine nurseries in the Bay of Biscay with a focus on sites exposed to the Erika oil spill. Scientia Marina. 183-192.

! 53!

Gosline William, Brock Vernon. 1960. Handbook of Hawaiian fishes. University of Hawaii Press (Honolulu). Gosline William. Vertical zonation of inshore fishes in the upper water layers of the Hawaiian Islands. 1965. Ecology. 46 (6): 823-831. Griffiths Shane. 2003. Rockpool ichthyofaunas of temperate Australia: species composition, residency and biogeographic patterns. Estuarine Coastal and Shelf Science. 58: 173-186. Haynes Paula, Brophy Deirdre, McGrath David. 2011. The early life history of turbot (Psetta maxima) on nursery grounds along the West coast of Ireland: 2007-2009, as described by otolith microstructure. 110: 478-482. Hoover John. 2008. The Ultimate guide to Hawaiian reef fishes sea turtles, dolphins, whales and seals. Mutual Publishing (Honolulu). Copyright © by John Hoover. Houde, ED. 2002. Mortality. In: Fishery Science: The Unique Contributions of Early Life Stages: Edited by Fuiman, LA and Werner RG. Blackwell Science Ltd. Oxford, UK. 64-87. Houde, Edward. 2008. Emerging from Hjort’s shadow. Journal of Northwest Atlantic Fishery Science. 41: 53-70. Jackson Jeremy, Kirby Michael, Berger Wolfgang, Bjorndal Karen, Botsford Louis, Bourque Bruce, Bradbury Roger, Cooke Richard, Erlandson Jon, Estes James, Hughes Terence, Kidwell Susan, Lange Carina, Lenihan Hunter, Pandolfi John M, Peterson Charles, Steneck Robert, Tegner Mia and Wagner Robert. 2001. Historical overfishing and the recent collapse of coastal ecosystems (review). Science. 293: (629-638). Lardner R, Ivantsoff W, Crowley LELM. 1993. Recolonization by fishes of a rocky intertidal pool following repeated defaunation. Australian Zoologist. 29 (1-2): 85-92. Leber Kenneth, Arce Steve, Sterritt David, Brennan Nathan. 1996. Marine stock-enhancement potential in nursery habitats of striped mullet, Mugil cephalus, in Hawaii. Fishery Bulletin. 94 (3). 452-455. Leber Kenneth, Brennan Nathan and Arce Steve. 1998. Recruitment patterns of cultured juvenile Pacific Threadfin, Polydactylus sexfilis released along sandy marine shores in Hawaii. Bulletin of Marine Science. 62 (2): 389-408. Lomeli Mark. 2009. The movement and growth patterns of young-of-the-year black rockfish (Sebastes melanops) inhabiting two rocky intertidal areas off Northern California. Thesis (MS) Humboldt state University, Natural Resources: Fisheries Biology.

! 54!

Mahon Robin and Mahon Susan. 1994. Structure and resilience of a tidepool fish assemblage at Barbados. Environmental Biology of Fishes. 41 (1-4): 171-190. Major Peter F. 1978. Aspects of estuarine intertidal ecology of juvenile striped mullet, Mugil cephalus, in Hawaii. Fishery Bulletin. 76 (2): 299-303. Mateo Ivan, Durbin Edward, Bengtson David and Durant Daisy. 2011. Variations in growth of Tautog in nursery areas in Narragansett Bay and Rhode Island Coastal Ponds. Marine and Coastal Fisheries: Dynamics, Management and Ecosystem Science. 3: 271-278. Mateo Ivan, Durbin Edward, Appledoorn Richard, Adams Aaron, Juanes Francis and Durant Daisy. 2011. Inferred growth of juvenile French Grunts, Haemulon flavolineatum andSchoolmaster, Lutjanus apodus, in mangrove and seagrass habitats. Bulletin of Marine Science. 87 (0): 1-12. McRae Mark, McRae Lori and Fitzsimons Michael. 2011. Habitats used by juvenile flagtails (Kuhlia spp.; Perciformes: Kuhliidae) on the Island of Hawaii. Pacific Science. 65 (4): 441-450. Metaxas A, Scheibling RE. 1993. Community structure and organization of tidepools. Marine Ecology Progress Series. 98 (1-2): 187-198. Moring JR. 1986. Seasonal presence of tidepool fish species in a rocky intertidal zone of Northern California, USA. Hydrobiologia 134: 21-27. Mumby Peter, Edwards Alasdair, Arlas-Gonzalez Ernesto, Lindeman Kenyon, Blackwell Paul, Gall Angela, Gorczynska Malgosia, Harborne Alastair, Pescod Claire, Renken Henk, Wabnitz Colette and Llewellyn Ghislane. 2004. Mangroves enhance the biomass of coral reef fish communities in the Carribbean. Nature. 427: 533-536. Nagelkerken I, Roberts CM, van der Velde G, Dorenbosch M, van Riel MC, Cochertet de la Moriniere E, Nienhuis PH. 2002. How important are mangroves and seagrass beds for coral-reef fish? The nursery hypothesis tested on an island scale. Marine Ecology Progress Series. 244: 299-305. Norris K. 1963. The functions of temperature in the ecology of the percoid fish Girella nigricans (Ayers). Ecological Monographs. 33: 23-62. Ortiz Delisse and Tisot Brian. 2008. Ontogenetic patterns of habitat use by reef-fish in a marine protected area network: a multi-scaled remote sensing and in situ approach. Marine Ecology Progress Series. 365: 217-232.

! 55!

Plaza Guido, Katayama Satoshi and Omori Michio. 2010. Daily patterns of settlement and individual growth rates of young-of-the-year of the rockfish Sebastes inermis in a Sargassum bed. Fisheries Research. 103: 48-55. Randall John. 1961. A contribution to the biology of the convict sureonfish of the Hawaiian Islands, Acanthurus triostegus sandvicensis. Pacific Science. 15 (2): 215-272. Randall John. 2010. Shore fishes of Hawaii. Revised Edition © University of Hawaii Press. Richards William and Lindeman, Kenyon. 1987. Recruitment dynamics of reef fishes: planktonic processes, settlement and demersal ecologies and fishery analysis. Bulletin of Marine Science. 41 (2): 392-410. Rooker JR, Landry AM, Geary BW and Harper JA. 2004. Assessment of a shell bank and associated substrates as nursery habitat of postsettlement red snapper. Estuarine, Coastal and Shelf Science. 59: 653-661. Rosa Ricardo Rosa lerece and Rocha Luiz. 1997. Diversidade da ictiofauna de pocas de mare da praia do cabo branco, oao Pessoa, Paraiba, Brazil. Revta bras.Zool. 14 (1): 201-212. Ross Steve. 2003. The relative value of different estuarine nursery areas in North Carolina for transient juvenile marine fishes. Fishery Bulletin. 101: 384-404. Sale Peter. 1969. Pertinent stimuli for habitat selection by the juvenile Manini, Acanthurus triostegus sanvicensis. Ecology. 50 (4): 616-623. Secor David, Dean John and Laban Elisabeth. 1991. Manual for otolith removal and preparation for microstructural examination. Published by the electric power research institute and the Belle Baruch Institute for Marine Biology and Coastal Research. Technical Publication Number 1991-01. Sponaugle Su. 2009. Daily otolith increments in the early stages of tropical fish. Tropical Fish Otoliths: information for assessment, management and ecology. Edited by Green Bridget, Mapstone Bruce, Carlos Gary and Begg Gavin. © Springer. Smale MJ and Buxton CD. 1989. The subtidal gully fish community of the Eastern Cape and the role of this habitat as a nursery area. South African Journal of Zoology. 24 (1): 58-67. Smith G.C and Parrish J.D. Estuaries as nurseries for the jacks Caranx ignobilis and Caranx melampygus (Carangidae) in Hawaii. 2002. Estuarine, Coastal and Shelf Science. 55: 347-359.

! 56!

Smith Kimberly. 1993. An ecological perspective on inshore fisheries in the Main Hawaiian Islands. Marine fisheries review. 55 (2): 34-49. Stimson John. 2005. Achipelago-wide episodic recruitment of the file fish Pervagor spilosoma in the Hawaiian Islands as revealed in long-term records. Environmental Biology of Fishes. 72 (1): 19-31. Stocks J, Stewart J and Gray CA. 2011. Using otolith increment widths to infer spatial, temporal and gender variation in the growth of sand whiting Sillago ciliata. Fisheries Management and Ecology. 18: 121-131. Studebaker Rebecca. 2006. Use of rocky intertidal areas by juvenile Sebastes in Northern California. A thesis presented to the faculty at Humbolt State University. Masters of Science in Natural Resources: Fisheries Biology. Studebaker RS and Mulligan TJ. 2008. Temporal variation and feeding ecology of juvenile Sebastes in rocky intertidal tidepools of Northern Califnornia, with emphasis on Sebastes melanops Girard. Journal of Fish Biology. 72: 1393-1405. Studebaker Rebecca, Cox Karah and Mulligan Timothy. 2009. Recent and historical spatial distributions of juvenile rockfish species in rocky intertidal tidepools, with emphasis on Black Rockfish. Transactions of the American Fisheries Society. 138: 645-651. Stunz Gregory, Minello Thomas and Levin Phillip. 2002. Growth of newly settled red drum Sciaenops ocellatus in different estuarine habitat types. Marine Ecology Progress Series. 238: 227-236. Tester Albert and Takata Michio. 1953. Contribution to the biology of the Aholehole a potential baitfish. Industrial research advisory council grant no. 29. Final Report: 1-54. Titcomb Margaret. 1972. Native Use of Fish in Hawaii. University of Hawaii Press, Honolulu. © The University Press of Hawaii. Vinagre C, Fonseca V, Maia A, Amara R and Cabral H. 2008. Habitat specific growth rates and condition indices for the sympatric soles Solea solea and Solea senegalensis Kaup 1858, in the Tagus estuary, Portugal, based on otolith daily increments and RNA-DNA ratio. Journal of Applied Ichthyology. 24 (2): 163-169. Walsh, William. 1984. Aspects of nocturnal shelter, habitat space, and juvenile recruitment in Hawaiian coral reef fishes. Dissertation at the University of Hawaii at Mānoa, Department of Zoology. Walsh, William. 1987. Patterns of recruitment and spawning in Hawaiian reef fishes. Environmental Biology of Fishes. 18 (4): 257-276.